Treating bacterial burden in chronic lower leg ulcers · Treating bacterial burden in chronic lower leg ulcers Foreword The Australian Wound Management Association has a major goal

Treating bacterial burden in chronic lower leg ulcers

A Randomised Controlled Trial Comparing Two Antimicrobial Dressings

- Cadexomer Iodine and Nanocrystalline Silver

March 2009

Charne Miller1, Nelly Newall2,3, Suzanne Kapp1, Gill Lewin2,3, Leila Karimi1, Keryln Carville2,3,

Terry Gliddon1, Nick Santamaria3

1 Royal District Nursing Service Helen Macpherson Smith Institute of Community Health 2 Silver Chain Nursing Association 3 Curtin University of Technology

A Research Project Supported by the Angior Family Foundation


Address for Correspondence:

Royal District Nursing Service Helen Macpherson Smith Institute of Community Health 31 Alma Rd St Kilda Victoria 3182 Australia

ISBN 978-0-9806380-0-4

Library: Email: [email protected] Telephone: 61 3 9536 5249 Fax: 61 3 9537 0271 RDNS Helen Macpherson Smith Institute of Community Health: Telephone: 61 3 9536 5251 Fax: 61 3 9536 5300 Published by: © Royal District Nursing Service Limited 2009 Melbourne Australia ABN 49 052 1088 717

This document is copyright. Apart from any fair dealing for the purposes of private study, research, criticism or review, as permitted under Australian copyright law, no part of this document may be reproduced or copied in any form, or scanned or stored in any type of information retrieval device or transmitted in any form or by any means, without the prior written permission of Royal District Nursing Service Limited.

© Royal District Nursing Service Limited Page 1 of 133 Version 1.0 –March 2009

mailto:[email protected]


Foreword

The Australian Wound Management Association has a major goal of promoting well constructed research into approaches to wound care and it is to my immense pleasure that such an important study was carried out here in Australia.

Such randomised trials are far too infrequent in wound management. Trials that address common clinical issues – that health workers face in their daily professional lives – are even less frequent. This study is a superb addition to this and sets the standards for evaluating wound management approaches. The authors are to be congratulated for their vision, and more importantly for completing, and writing up, this large trial – no mean feat in any area of healthcare.

It would be improper to further distil the conclusions of this trial but they do help direct health professionals facing an infected wound. This study tells us more about the value of swabs in evaluating wounds assessed by the clinicians as infected, the predictors of wound healing and the choice of a local antiseptic agent to assist wound healing. It is perhaps not surprising that so few differences were found between the effects of the two products – cadexomer iodine and a range of dressings that deliver silver. The wounds were well managed by wound experts, and it is possible that this excellent care drove the outcomes as much as or more than the choice of product applied to the wound. Similarly, it is well recognised that compression is now more important in healing venous leg ulcers than the dressing itself.

The study does raise a series of further questions, well identified by the authors and hopefully these will motivate other researchers to search for answers through further trials. We certainly need more of this research – case studies have their place, but the randomised trial provides the most resilient information to direct clinicians. Possibly the greatest next challenge is to incorporate the knowledge provided by studies such as this into our clinical practices and to measure whether this is indeed occurring.

Will this study be referred to by trade representatives talking about their silver or iodine product? Will wound management educators present this data in educational courses and material? Will clinicians treating an infected wound utilise this data in their product selection – will they feel more confident that the choice between iodine cadexomer or silver is perhaps less vital to wound healing – and that what may matter more is the overall approach to the wound, not the product choice? AWMA certainly hope that high quality research like this does inform and influence practise. We want Australians and all wound care specialists to achieve the best outcomes for people with wounds. This study certainly moves us close to that goal.

Associate Professor Michael Woodward.

President, Australian Wound Management Association





Acknowledgements

Special thanks go to: The Angior Family Foundation for providing the major funding for this project.

RDNS Foundation and Silver Chain Foundation for funding wound dressings for study participants.

Department of Human Services (Victoria) for providing the funding that enabled the Best Practice Wound Education to be rolled-out to all Royal District Nursing Service staff.

The RDNS and Silver Chain staff who recruited clients, provided the care, and stayed committed to the outcomes of this research and pursuit of best practice care for the benefit of their clients.

The clients who participated in the study.

Associate Professor Michael Woodward and staff at the Wound Management Clinic, Heidelberg Repatriation Hospital, Austin Health for assisting with the identification of study participants.

To Professor Nick Santamaria and Medseed for permitting and supporting the use of The Advanced Medical Wound Imaging System (AMWIS) at no cost during the study.

Melbourne Pathology and Clinipath from whom the trial purchased its wound swab tests. The support received in setting up the trial documentation and procedures, and design of the reporting structure is greatly appreciated.

Smith and Nephew Pty Limited who provided reduced product pricing for the RCT and provided Visitrak Depth® measures to one site at no cost.





Contents

Page

Foreword ........................................................................................................................2 Acknowledgements........................................................................................................2 Contents .........................................................................................................................6 List of Tables .................................................................................................................8 List of Figures ..............................................................................................................10 Executive Summary .....................................................................................................12 1.0 Introduction............................................................................................................16

1.1 Lower Leg Ulcer Initiative Overview................................................................16 1.2 Literature Review...............................................................................................18 1.3 RCT Aim & Objectives .....................................................................................23

2.0 Method ...................................................................................................................24 2.1 Research Design.................................................................................................24 2.2 Ethics Approval .................................................................................................25 2.3 Research Sample................................................................................................26 2.4 Client Recruitment .............................................................................................27 2.5 Protocol Amendments........................................................................................28 2.6 Participation in Angior RCT Study....................................................................28 2.7 Study Measures & Data Collection Forms ........................................................29 2.8 Statistical Analysis.............................................................................................31

3.0 Results: Sample Description & Comparison .........................................................34 3.1 The RCT Sample................................................................................................34 3.2 Segmentation Group Comparison......................................................................44 3.3 Chapter Summary ..............................................................................................47

4.0 Results: Wound Healing Outcomes ......................................................................50 4.1 Healing Rate Analysis........................................................................................50 4.2 Healing Rate Analysis: Segmentation ...............................................................54 4.3 Tissue Type Comparison by Antimicrobial.......................................................60 4.4 Chapter Summary ..............................................................................................64

5.0 Results: Time to Wound Healing..........................................................................66 5.1 Number of Wounds Healed ...............................................................................66 5.2 Time to Healing .................................................................................................67 5.4 Chapter Summary ..............................................................................................68

6.0 Results: Presence of Pain and Exudate .................................................................70 6.1 Experience of Pain Analysis ..............................................................................70 6.2 Experience of Pain Analysis: Segmentation ......................................................72 6.3 Amount of Wound Exudate Analysis ................................................................77 6.4 Chapter Summary ..............................................................................................80

7.0 Results: Wound Culture Results ...........................................................................82 7.1 The Wound Culture Findings.............................................................................82 7.2 Comparing Wound Culture Results & Signs of Critical Colonisation / Infection..................................................................................................................................85 7.3 Comparing Wound Culture Results & Wound Healing Rate ............................87 7.4 Chapter Summary ..............................................................................................89



8.0 Results: Cost Comparison.....................................................................................92 8.1 Silver Versus Iodine...........................................................................................92 8.2 Segmentation Group Comparison......................................................................95 8.3 Chapter Summary ..............................................................................................98

9.0 Discussion & Recommendations .........................................................................100 9.1 Clinical Effectiveness of the Antimicrobials ...................................................100 9.2 Comparison of Signs of Bacterial Burden and Wound Swab Results .............103 9.3 Cost Comparison of the Antimicrobials ..........................................................106 9.4 Study Strengths and Limitations......................................................................107 9.5 Conclusion .......................................................................................................109 9.6 Recommendations............................................................................................109

References..................................................................................................................112 Appendix A................................................................................................................120 Appendix B ................................................................................................................124 Appendix C ................................................................................................................126 Appendix D................................................................................................................130



List of Tables Table Page

Table 3.1 Demographic Descriptors of the Treatment Groups.................................. 37

Table 3.2 Comorbidity Index Results for the Treatment Groups ............................. 38

Table 3.3 HACC Nutritional Risk Scores for the Treatment Groups ....................... 39

Table 3.4 Activities of Daily Living (ADL) and Independent Activities of Daily Living (IADL) Results for the Treatment Groups .......................................... 39

Table 3.5 Wound Characteristics for the Treatment Groups .................................... 40

Table 3.6 Prior Wound Dressings for the Treatment Groups ................................... 41

Table 3.7 Distribution of the Eligibility Criteria for the Treatment Groups ............. 42

Table 3.8 Wound Dimensions at Baseline for the Treatment Groups ...................... 43

Table 3.9 Exudate and Pain Levels for the Treatment Groups ................................. 44

Table 4.1 Descriptive Statistics for Healing Rate by Treatment Group ................... 52

Table 4.2 LMM analysis of Healing Rate ................................................................. 53

Table 4.3 ANCOVA of Healing Rate ....................................................................... 54

Table 4.4 Descriptive Statistics for Healing Rate by Treatment Group by Healed/ Not Healed Segmentation ............................................................................ 55

Table 4.5 ANCOVA of Healing Rate by Healed/ Not Healed Segmentation .......... 56

Table 4.6 Descriptive Statistics for Healing Rate by Treatment Group by Wound Duration Segmentation ................................................................................. 58

Table 4.7 ANCOVA of Healing Rate by Wound Duration Segmentation ............... 58

Table 4.8 Descriptive Statistics for Healing Rate by Treatment Group by Wound Size Segmentation ........................................................................................ 60

Table 4.9 ANCOVA of Healing Rate by Wound Size Segmentation ...................... 61

Table 4.10 Descriptive Statistics for Changes in Granulation by Treatment Group ........................................................................................................................ 63

Table 4.11 ANCOVA of changes in Granulation by Treatment Group ................... 63

Table 4.12 Descriptive Statistics for Changes in Slough by Treatment Group ........ 63

Table 4.13 ANCOVA of changes in Slough by Treatment Group ........................... 64

Table 4.14 Descriptive Statistics for Changes in Epithelialisation by Treatment Group ....................................................................................................... 64

Table 4.15 ANCOVA of changes in Epithelialisation by Treatment Group ........... 64

Table 5.1 Descriptive Statistics for the Number of Clients who Healed .................. 67

Table 5.2 Cox Regression Analysis Predicting Days to Heal ................................... 69

Table 6.1 Descriptive Statistics for the Experience of Pain by Treatment ............... 72



Group

Table 6.2 LMM analysis of the Experience of Pain by Treatment Group ................ 72

Table 6.3 ANCOVA of Pain Intensity by Treatment Group .................................... 73

Table 6.4 Descriptive Statistics for the Experience of Pain by Treatment Group by Healed/ Not Healed Segmentation ............................................................ 74

Table 6.5 ANCOVA of Pain Intensity by Treatment Group by Healed/ Not Healed Segmentation ......................................................................................... 74

Table 6.6 Descriptive Statistics for the Experience of Pain by Treatment Group by Wound Duration Segmentation ................................................................ 76

Table 6.7 ANCOVA of Pain Intensity by Treatment Group by Wound Duration Segmentation ............................................................................................. 76

Table 6.8 Descriptive Statistics for the Experience of Pain by Treatment Group by Wound Size Segmentation ........................................................................ 77

Table 6.9 ANCOVA of Pain Intensity by Treatment Group by Wound Size Segmentation ............................................................................................................. 78

Table 6.10 Descriptive Statistics and Chi Square Results for the Wound Exudate by Treatment Group .................................................................................... 79

Table 7.1 Degree of Bacterial Burden and Leucocytes Growth at each Swab ......... 84

Table 7.2 Degree of Bacterial Burden and Leucocytes Growth by Swab Indication .................................................................................................................. 85

Table 7.3 Type of Bacterial Organism at each Swab ................................................ 85

Table 7.4 Presence of Signs of Infection or Critical Colonisation ........................... 86

Table 7.5 Number of Eligibility Criteria Identified and Degree of Bacterial Growth ...................................................................................................................... 87

Table 7.6 Comparing the Healing Rate for First Fortnight for the Treatment Groups and Swab Results ......................................................................................... 89

Table 7.7 Comparing the Healing Rate at the First Fortnight by Treatment Group and Wound Eligibility Criteria at Baseline .................................................... 90

Table 8.1 Average Costs for Treatment Groups ....................................................... 94

Table 8.2 Average Costs for Treatment Groups excluding Acticoat® 7 .................. 95

Table 8.3 Average Costs for Individual Antimicrobial Products ............................. 96

Table 8.4 Average Costs for Treatment Groups by Healed / Not Healed Segmentation ............................................................................................................. 97

Table 8.5 Average Costs for Treatment Groups by Wound Duration Segmentation.............................................................................................................. 98

Table 8.6 Average Costs for Treatment Groups by Wound Size Segmentation ............................................................................................................. 99



List of Figures Figures Page Figure 1.1. Lower leg ulcer initiative overview .......................................................18

Figure 2.1. Treatment allocation ...............................................................................26

Figure 3.1. Participant flowchart ...............................................................................35

Figure 4.1 Fortnightly Healing Rate by Treatment Group ........................................52

Figure 4.2 Fortnightly Healing Rate by Treatment Group: Not Healed Segment .....57

Figure 4.3 Fortnightly Healing Rate by Treatment Group: Wound Duration Segmentation..............................................................................................................59

Figure 5.1. Kaplan-Meier healing curves presenting the total days to heal by antimicrobial treatment group ...................................................................................68

Figure 6.1 Fortnightly Pain Ratings by Treatment Group .........................................71

Figure 6.2 Fortnightly Pain Rating by Treatment Group: Healed/ Not Healed Segmentation..............................................................................................................75

Figure 6.3 Fortnightly Exudate Rating by Treatment Group.....................................79

Figure 6.4 Fortnightly Exudate Rating by Treatment Group: Healed/ Not Healed Segmentation..............................................................................................................80

Figure 6.5 Fortnightly Exudate Rating by Treatment Group: Wound Size Segmentation..............................................................................................................81

Figure 6.6 Fortnightly Exudate Rating by Treatment Group: Compression Bandaging Adherence Segmentation.........................................................................81





Executive Summary

The impact of chronic leg ulcers on the physical and emotional wellbeing of those affected is significant. The high incidence and recurrence of ulceration and the protracted healing time of these wounds translates into considerable and costly demands on health care resources. Any advancement that can be achieved by identifying and disseminating best practice clinical care and circumventing barriers to the application of best practice care can make sizable inroads to improving the quality of life of many and optimising the use of community health care services.

Two of Australia’s largest community nursing services, Royal District Nursing Service (RDNS) in Victoria and Silver Chain Nursing Association (Silver Chain) in Western Australia, obtained funding to undertake an initiative aimed at improving the care provided to and outcomes for all people living with a lower leg ulcer who, at these services, represent a large proportion of clients attended and between 35-48 per cent of all wounds seen (Carville & Lewin, 1998; Kapp & Nunn, 2005). Consisting of three components, this initiative included a best practice education programme (Kapp, Flowers, Karimi, & Gliddon, 2008), a study to investigate the enablers and barriers to implementing best practice compression therapy (Annells, O'Neill, & Flowers, 2006), and a Randomised Controlled Trial (RCT) comparing two commonly used antimicrobial dressings for wounds showing signs of critical colonisation or infection. The results of this report reflect the latter RCT outcomes.

Antimicrobial treatment is clinically indicated when a wound appears critically colonised or infected (Principles of best practice: Wound infection in clinical practice. An international consensus, 2008). Nanocrystalline Silver (Silver) and Cadexomer Iodine (Iodine) are two of the most commonly used antimicrobials but these have never been the focus of any direct comparison in clinical studies. An RCT was conducted between March 2006 and February 2007 at RDNS and Silver Chain with clients randomised to one of these treatments. The RCT aim was stated as: To refine existing knowledge and protocols for lower leg ulcer management and develop an evidence based solution to the problem of bacterial colonisation and wound infection.

The specific objectives of the RCT were:

1. To evaluate and compare the clinical effectiveness of two antimicrobial treatments (Iodine and Silver dressings) in the control of bacterial colonisation and infection in wounds, and to demonstrate the subsequent effect on wound healing.

2. To compare the cost effectiveness, acceptability, ease of use and utility of Iodine and Silver dressings in controlling critical colonisation and infection in wounds.

3. To develop an evidence-based protocol for the management of critically colonised and infected wounds.

Data were gathered from 281 clients, 266 (95%) of whom were able to be included in an intention-to-treat analysis. Primary outcome measures included



wound healing rate assessed fortnightly and quantified by the Medseed AMWIS™ wound imaging software (Santamaria, Austin, & Clayton, 2002; Santamaria, Carville, Ellis, & Prentice, 2004; Santamaria & Clayton, 2000) and wound healing within a 12 week monitoring period. Data were collected on potential covariates while costs associated with wound product usage and nursing care were monitored. Wound swabs were attended at baseline, when ceasing or recommencing an antimicrobial treatment, and at six and 12 weeks if treatment with an antimicrobial continued.

Results are reported for the sample overall and for several segmentations (sub-groups) including comparisons of the treatment groups for the segment of clients who healed during the observation period and those that did not, for segments with short and long wound duration (defined as more or less than 12 weeks of duration) and those with small and large wound size (defined as more or less than 3.6cm2 in size).

This trial found both antimicrobial dressings had a positive effect on healing lower leg ulcers that were complicated with signs of critical colonisation and infection; 64% of ulcers healed in the 12 week observation period. Both antimicrobials had a comparable performance with respect to the number of wounds healed, healing rate, and in the resolution of pain. There were two circumstances in which Silver was associated with quicker healing times. This occurred in the first fortnight though overall healing rates were comparable after 12 weeks of study observation. Secondly, Silver was found to be more clinically effective for wounds unlikely to heal in a 12 week period. In this sample these wounds were characterised as being larger (an average size of 10cm2), more enduring (a median of 24 weeks duration or almost six months), and with moderate to high levels of wound exudate (a quarter of the sample) as compared to the wounds which did heal in the 12 week observation period.

Research may also consider the area of wound inflammation to examine the hypothesis that the effectiveness of Silver in the first fortnight reflects Silver’s anti-inflammatory mechanism (Sibbald et al., 2007) as contrasted to Iodine’s inflammatory action (Moore, Thomas, & Harding, 1997). Examination of wound inflammation may also identify situations in which one antimicrobial, based on the level of inflammation present, would be preferred over the other.

Five factors were identified as predictive of time to healing. Adherence to compression bandaging made a positive contribution to healing times. The presence of slough and exudate was associated with slower healing. Wounds that were large and of long duration prior to treatment were also slower to resolve. These findings add weight to existing evidence regarding the clinical effectiveness of compression therapy (Moffatt, 2002; O'Meara, Cullum, & Nelson, 2009), and swifter healing when wounds are smaller and younger (Margolis, Allen-Taylor, Hoffstad, & Berlin, 2004; Phillips et al., 2000; Tennvall & Hjelmgren, 2005). They also reaffirm the need to adopt appropriate dressings or treatments to reduce wound exudate and slough to optimise healing times.

A comparison of microbiology data gathered using wound swabs and clinician assessed signs of critical colonisation and infection from a list of 10 markers derived from the literature, revealed no association between the two diagnostic approaches. As such, the finding from wound culture data that Silver was more effective when there was a low level of bacterial growth has been treated with caution. Additional studies can confirm the reliability of this result and further



explore the disparity between wound swab and clinician assessment. The role of the biofilm in obstructing the wound swab from reliably sampling the bacteria on the surface of the wound is tendered as a possible reason for this lack of comparability.

The antimicrobials were generally equivalent with respect to the cost associated with products for the study period. Nurse time, the cost of products used, and a combination of these were used as an overall costing for this analysis. However, there are two instances where, in the absence of a clinical rationale indicating the use of one particular antimicrobial, one product might be favoured over the other. Firstly, if considering the cost to the client, Iodine might be a preferred product if the client’s wound is a young wound (less than 12 weeks old) given that product costs were cheaper for this group than Silver. However, from a Health Service perspective adding the nursing care costs to the product cost does not show the same benefit for young wounds, indicating that the delivery of Iodine in terms of treatment time and frequency outweighs the initial product cost saving. If treating a large wound (greater than 3.6cm2) there are cost savings possible for both client and health care service with a Silver antimicrobial. There is also merit in changing from Acticoat® Absorbant as soon as clinically indicated, and in selecting Acticoat® 7 to minimise the cost associated with home visits.

This trial offers the first data, gained from a large multi-site RCT, directly comparing two commonly used antimicrobial treatments. The following recommendations emerge directly from the results of this trial.

1. That clinical guidelines be revised to reflect the results of this trial as indicating:

a) Both Silver and Iodine treatments achieved comparable clinical effectiveness with the exception of two instances in which Silver had a quicker healing rate than Iodine including:

i. the first fortnight of treatment.

ii. wounds that did not heal in a 12 week period. In the study these wounds, in comparison to wounds that did heal, were larger (an average size of 10cm2) and of longer duration (a median of 24 weeks duration), and more likely (a quarter of this sample) to have moderate to high levels of exudate.

b) Both Silver and Iodine treatments are comparable from a cost perspective with the exception of:

i. young wounds (less than 12 weeks old) which have cheaper product costs when treated with Iodine, though this difference was not extended to the costs of nursing related to its application nor to the combined product-nursing cost.

ii. large wounds (bigger than 3.6cm2) which, though comparable for product costs, had cheaper nursing costs and combined product-nursing cost when treated with Silver.

c) Wound healing was expedited by the use of multi-layer compression bandaging therapy, resolving wound slough, managing wound exudate, and by intervening when the wound is young and small.



2. That emphasis in Australian Venous Leg Ulcer Management Guidelines would be warranted with respect to:

a) enhancing access to multi-layer compression bandaging therapy when clinically indicated and optimising adherence to its use.

b) dressings and treatments to resolve wound exudate and slough.

c) early intervention before wounds become large and chronic.

3. That further research is required to compare:

a) Silver and Iodine in managing moderate to high levels of wound exudate generally.

b) Silver and Iodine in managing wound exudate in the absence of compression bandaging therapy.

4. That research undertaken to compare these two antimicrobials in the future incorporate measures of wound inflammation to help understand the mechanism through which Silver and Iodine act and to explore whether the antimicrobials would be better suited to particular wounds based on the presence or risk of inflammation.

5. That the role of biofilms in wounds is the focus of additional research to identify:

a) how microbiology measures can reliably sample wound bacteria in the presence of a biofilm.

b) effective antimicrobial and other treatments to resolve critical colonisation and infection in a wound with a biofilm.

6. That to inform the disparity observed for clinician assessment and wound culture results:

a) the inter and intra rater reliability and validity of clinician assessment of signs of critical colonisation and infection is investigated.

b) future trials explore the use of multiple measures of critical colonisation and infection as well as other markers of wound status including inflammation to assist with the optimisation of therapies to match healing.

7. That future investigations contrast healing outcomes of antimicrobial treatment:

a) with a comparison, no antimicrobial treatment group given the presence of recommended multi-layer compression therapy and best practice wound bed preparation for both groups.

b) for individuals living with Diabetes Mellitus, arterial leg ulcers, and wounds other than lower leg ulcers.



1.0 Introduction

The chronic leg ulcer is a debilitating, unpleasant and often painful condition associated with poor health in old age. With a prevalence of 1.1 to 3.0 per thousand of the adult population (in Western countries), leg ulcers can result in or exacerbate significant morbidity and financial burden. The annual cost of treating these chronic wounds is estimated at A$3 billion with recurrence of ulceration reported to be as high as 69% (Angel, Sieunarine, Abbas, & Mwipatayi, 2005; Kapp & Sayers, 2008).

For health services, leg ulcers use a significant amount of health care providers’ time in both community and institutional settings and require enormous quantities of expensive wound pharmaceuticals. For community nursing services across Australia, wound care, and specifically leg ulcer care, consumes a large part of their total resources. Wound prevalence studies from two of Australia’s largest providers of community nursing services suggest that leg ulcers represent between 35% (Royal District Nursing Service) and 48% (Silver Chain Nursing Association) of all wounds seen by these services (Carville & Lewin, 1998; Kapp & Nunn, 2005). Given that half the nursing activities provided by these community care agencies relate to wound care, it can be extrapolated that nearly a quarter of the work of nurses in these agencies is related to leg ulcer management.

1.1 Lower Leg Ulcer Initiative Overview In 2005, Royal District Nursing Service (RDNS) in Melbourne, Victoria and Silver Chain Nursing Association (Silver Chain) in Perth, Western Australia received funding to undertake a Randomised Controlled Trial (RCT) to compare the clinical and cost effectiveness of two commonly used antimicrobial dressing preparations.

A shared agenda to establish and create evidence for best practice in the treatment of lowers leg ulcers not only spanned the RCT but preceded and follows that research. At the outset, in order to ensure a best practice context and minimise any variation in practice across and within the study sites, a best practice education program was developed and rolled out at the Victorian site. The education package was developed in an intensive two day exercise at RDNS led by Associate Professor Keryln Carville who also had overall responsibility for best practice at Silver Chain. Thus a level playing field was established so that any differences in client outcomes in the RCT would be attributed to the study intervention. While this best practice ‘lens’ was focusing on venous leg ulcer management at RDNS a qualitative study was also undertaken to explore why compression bandaging, an established best practice component of venous leg ulcer treatment, does not always take place.

An overview of this multi-faceted ‘Lower Leg Ulcer Initiative’ is provided as background to the following report which focuses on the conduct and results of the RCT. Figure 1.1 presents the three major research and evaluation components of the Lower Leg Ulcer Initiative and their relationship to the refinement of guidelines and ongoing education regarding the treatment of critically colonised and infected lower leg ulcers. Funding to undertake this




project was received from the Angior Family Foundation, a philanthropic trust, with aspects of the project receiving further monetary support from the RDNS Foundation and Silver Chain Foundation. The capacity to conduct the ‘Wound Best Practice Education’ across the entire RDNS site was enabled by funding from the Department of Human Services (Victoria).

Figure 1.1. Lower leg ulcer initiative overview

The ‘Wound Best Practice Education’ component established a standardised understanding of current best practice treatment prior to the commencement of the RCT at both RCT study sites. RDNS committed to an organisation wide mandatory education program with the goal of updating and standardising knowledge of wound best practice at all its service sites. An evaluation of the impact of the education on nurse knowledge and client outcomes was progressed as a component of the Lower Leg Ulcer Initiative. The results of this evaluation have been presented (Kapp, Flowers, Karimi et al., 2008). A detailed wound best practice education program was rolled out across Silver Chain prior to March 2005 for all registered nurses.

The ‘Compression Bandaging Qualitative Study’ investigated the enablers and barriers to implementing best practice compression therapy for clients with venous leg ulcers. This component of the Lower Leg Ulcer Initiative explored with clinicians why, when compression bandaging therapy is an established and known best practice treatment, it is not always used when and as indicated. The results of this qualitative study have been published (Annells et al., 2006; Annells, O'Neill, & Flowers, 2008).

In order to establish whether one of two commonly used antimicrobials – Silver and Cadexomer Iodine1 – offer any clinical or cost benefits compared to the

1 Referred to in the Results chapters as ‘Silver’ and ‘Iodine’. The use of capitals has been consistently applied to aid the reader swiftly identify the treatment groups.

Compression Bandaging Qualitative Study

Refine Guidelines for the treatment of colonised or infected lower leg ulcers

Antimicrobial Randomised Controlled Trial

Wound Best Practice Education and Evaluation


other, a multi-site RCT was implemented. In the absence of any clinical trial to date which directly compared these two antimicrobials, a sizable RCT was undertaken to produce the first evidence to guide clinical practice. The findings of this RCT are to be incorporated into the refinement of guidelines regarding the selection of primary dressing products for critically colonised or infected lower leg ulcers. From the outset it has been an aim that these guidelines will be published and promoted, as well as incorporated into revisions to wound best practice education programs.

The Lower Leg Ulcer Initiative has aimed to improve the care provided to and outcomes for all people living with a lower leg ulcer. It aimed to achieve this by establishing and sharing what is best practice and exploring enablers and barriers to its implementation.

The remainder of this chapter, and indeed the report, focuses solely upon the RCT component of the Lower Leg Ulcer Initiative. This chapter reviews the literature regarding leg ulceration and the two antimicrobials compared in the RCT, as well as stating the RCT aim and objectives.

1.2 Literature Review 1.2.1 The Leg Ulcer Problem

The leg ulcer is a common skin condition afflicting predominantly aged persons who have chronic health concerns. There is no consensus on the definition of a leg ulcer. Lower leg ulcers have been defined as ‘loss of skin on the leg or foot which takes more than 6 weeks to heal’ (Nelson, Bell-Syer, & Cullum, 2000). For the purpose of this study leg ulcers were defined as an ulcer on the lower leg with an aetiology of venous insufficiency. Leg ulceration has been found to have a major impact on the health-related quality of life of the people afflicted (Franks, 1998). The person living with a leg ulcer can expect to experience pain and discomfort (Chase, Melloni, & Savage, 1997; Douglas, 2001; Husband, 2001; Hyde, Ward, Horsfall, & Winder, 1999; Krasner, 1997; Rich & McLachlan, 2003) and restrictions of mobility and activity which may negatively impact the ability to socialise (Chase et al., 2000; Neil & Munjas, 2000). There may be disappointment when unable to wear aesthetically pleasing footwear, and embarrassment associated with wound odour and wound exudate, while it has been established that leg ulceration is associated with loss of control and social isolation (Douglas, 2001; Hyde et al., 1999; Neil & Munjas, 2000; Rich & McLachlan, 2003). As healing can be slow and recurrence common, pessimism concerning a future free of leg ulceration has been identified (Douglas, 2001; Hyde et al., 1999; Rich & McLachlan, 2003).

In addition to these health and social effects, leg ulcers present costs to the sufferer, health care providers, and health care systems. In 1994 the monetary cost of leg ulcers was estimated at up to A$431 million per annum (Baker & Stacey, 1994) increasing in 2005 to A$3 billion per annum (Angel et al., 2005).

Financial costs to the individual include those associated with obtaining health care from general practitioners and nursing services and the cost of wound dressing and compression therapy products. There is considerable inequity across Australian states and government funded programs with respect to the funding of health care visits and subsidisation of wound products. Leg ulcers can cost clients large sums of money in the absence of subsidisation or other



financial support. It has been suggested that when clients are burdened with the responsibility to purchase wound products, some may select cheaper wound products which can compromise the provision of best practice care (Flowers, Kapp, Lewin, Newall, & Gliddon, 2007).

1.2.2 Treating the Leg Ulcer & Bacterial Burden Best practice treatment of lower leg venous ulcers is moist wound dressings and graduated compression therapy (O'Meara et al., 2009) with multi-layer compression therapy considered the gold standard in venous leg ulcer management (Moffatt, 2002). For the uncomplicated venous ulcer, this treatment is usually successful and the ulcer will heal. Regrettably, multi-layer compression therapy has a high unit cost and even though it has been demonstrated to be cost effective (Posnett, 2006), the upfront cost is an insurmountable barrier for some clients.

Not all wounds heal with standard therapy and one factor which may complicate the healing process is bacterial colonisation. Healing is impaired when bacterial colonisation reaches a critical level, at which stage progression to infection is inevitable unless an effective intervention is instigated. Bacterial burden must be managed to prevent delayed healing, the development of infection and complications such as cellulitis, sepsis or death.

The current understanding is that the use of primary dressings with antimicrobial properties is clinically indicated when a wound becomes critically colonised (Cutting & Harding, 1994; Gardner, Frantz, & Doebbeling, 2001; Schultz et al., 2004) or infected, and bacterial burden is unlikely to be managed by drainage, debridement or cleansing alone (Principles of best practice: Wound infection in clinical practice. An international consensus, 2008). There is, however, little consensus as to what constitutes best practice in the selection and use of antimicrobial preparations to treat bacterial colonisation and wound infection. Silver and Cadexomer Iodine primary dressings are antimicrobial products commonly used for the treatment of critically colonised or infected leg ulcers. Though evidence has been accrued with respect to both of these types of antimicrobials, there has been no clinical trial research directly comparing these treatments for their clinical or cost effectiveness.

1.2.3 Evidence Pertaining to Antimicrobial Dressings While a range of Silver and Iodine products is available, the literature review which follows focuses primarily on Silver or Nanocrystalline Silver and Cadexomer Iodine dressings.

1.2.3.1 Silver and Nanocrystalline Silver Dressings

The use of silver in public health and health care already had a significant history prior to its renaissance with the development of silver applications for the treatment of burns, wounds more broadly, and even catheterisation in recent decades (Dowsett, 2003; Klasen, 2000; Russell & Hugo, 1994; White, 2001). Its first known use, to line water vessels to preserving the water, dates back to ancient Greek and Roman civilizations (White, 2001). In wound care, Silver nitrate emerged as a topical application for chronic wounds or ulcers. Its use was documented in a publication appearing in 1617, called ‘The Surgeons Mate’ written by a surgeon named John Woodall to instruct practice among ships’ surgeons (Klasen, 2000).



The 20th Century refined modern perspectives of the role of microbiological infection and the biocidal properties of silver were explored and further understood. New mechanisms to further enable the sustained delivery of silver ions which bind to bacterial walls and result in the death of the bacteria (Lansdown, 2002) through dressings, have more recently been developed.

Systematic reviews of antimicrobial dressings have been limited in their ability to comment as to the effectiveness of Silver dressings in leg ulcer management by the scarcity of robust clinical research. In 2001, a systematic review identified that Silver dressings resulted in a significant reduction in ulcer size at two and four week intervals, though not at a six week follow-up (O'Meara, Cullum, Majid, & Sheldon, 2001). However, the authors excluded Silver dressings in a list of topical preparations which aid healing (though Silver sulphadiazine was noted) perhaps because of the limited research evidence available and the failure of the trial that was considered to show a significant difference at the final follow-up.

In a more recent systematic review (Vermeulen, Van Hattem, Storm-Versloot, & Ubbink, 2007), three RCTs were identified that had investigated the effects of Silver dressings and topical silver on infected acute or chronic wounds. While the evidence for Silver dressings had some prospect, it was generally underwhelming. One study found though the initial reduction in wound size was significantly faster for a Silver foam (Contreet®) compared to a hydrocellular foam (Allevyn®), however after a four-week period there were no significant differences in wound healing rate, pain levels, or patient satisfaction and quality of life (Jorgensen, Price, Andersen, et al., 2005). In another study, a small treatment effect was found for healing rate, which was slightly higher for a Silver alginate (Silvercel®) when compared to an alginate (Algosteril®), however, there were no significant differences between the groups in terms of the number of wounds healed, or absolute and relative decrease in wound size (Meaume, Vallet, Morere, & Teot, 2005). Finally, relative wound size reduction was significantly quicker when Silver foam (Contreet®) was compared to best local practice (i.e. a range of dressing: foam and alginates, hydrocolloids, gauze, silver dressings, other antimicrobial dressings or other active dressings) (Munter et al., 2006) though the systematic review queried what data were used in the endpoint analysis for this study.

Systematic reviews of the effectiveness of Silver dressings suffer from a lack of quality trials to include in their analysis. This is an area where rigorous RCTs are required if clarity is to be provided for clinicians on antimicrobial wound dressing selection.

1.2.3.2 Comparison among Silver Dressings

In vitro studies have compared a range of Silver products. In one study, Actisorb® Silver 220 demonstrated the greatest reduction of bacteria followed by Contreet-H®, while Avance® produced no significant antimicrobial effect (Addison & Rennison, 2002, cited in Thomas & McCubbin, 2003). In another study, Acticoat® and Contreet-H® were seen to exhibit sustained antibacterial action with no or limited bacterial transfer, while Acticoat® was deemed the most effective due to its rapid release of silver ions (Thomas & McCubbin, 2003). Overall, the most positive findings for this study were detected for Acticoat® and Contreet-H®, with Actisorb® Silver 220 demonstrating moderate effectiveness and Avance® performing poorly.



A large number of studies have focused exclusively on the product Acticoat®. Acticoat® which has been showed to achieve better clinical outcomes compared to Silver nitrate (Tredget, Shankowsky, Groenveld, & Burrell, 1998; Wright, Lam, & Burrell, 1998; Yin, Langford, & Burrell, 1999), Silver sulphadiazine (Wright et al., 1998; Yin et al., 1999), gauze dressing impregnated with polyhexamethlyese biguanide (Wright et al., 1998; Wright, Lam, Olson, & Burrell, 2003), standard dressing (Paddock, Schulz, Perrin, Moldawer, Wright, Burrell, & Mozingo, 2002 as cited in Burrell et al., 2003), xeroform and eight ply gauze dressing continually moistened with a 0.01% neomycin and polymyxin solution (Demling & DeSanti, 2002), xeroform petrolatum impregnated absorbent gauze (Olson, Wright, Lam, & Burrell, 2000 as cited in Burrell et al., 2003), and mafenide acetate (Yin et al., 1999). Acticoat® has also been reported to reduce matrix metalloproteinases to assist in wound healing (Kirsner, Orsted, & Wright, 2001).

While Acticoat® received some favourable findings in dressing donor sites in animals (Olson et al., 2000 as cited in Thomas & McCubbin, 2003), it was out performed by the Allevyn® dressing which had no silver properties in a small human study on grafting sites for patients with burn wounds (Innes, Umraw, Fish, Gomez, & Cartotto, 2001). Another in vitro study found Acticoat® to be cytotoxic to cultured keratinocytes and recommended against use on cultured skin grafts (Lam, Chan, Ho, & Liew, 2004). Silver sulphadiazine and 0.2% chlorhexidine digluconate was more effective against a greater variety of burn pathogens than Acticoat® (Fraser, Bodman, Sturgess, Faoagali, & Kimble, 2004). Findings have also suggested that the effectiveness of Acticoat® is confined to the length of time the dressing is able to stay in contact with burn pathogens (Holder, Durkee, Supp, & Boyce, 2003).

An effective dressing, though with the above exceptions noted, Acticoat® has been further seen to be safe, has not to date been linked to the development of resistant bacteria, and although transient staining of the surrounding skin has been noted, hyperpigmentation was not permanent (Sibbald, Browne, Coutts, & Queen, 2001).

With respect to cost effectiveness of Silver dressings, one study was identified which provided comment on the cost benefits of using a Silver impregnated dressing, taking into account reduced dressing changes, quicker healing times and reduced complications, and a reduced resource demand (Stephens, Silverstein, Meites, Jett, & Brou, 1999, cited in Dunn & Edwards-Jones, 2004). This study observed that a 36% reduction in labour and supply costs could be obtained if Acticoat® was used once every three to four days when compared with the twice daily use of Silver sulfadiazine.

Generally speaking, the use of Silver dressings in wound care represents a promising area of clinical endeavour. There is no shortage of researchers advocating for the conduct of large and quality clinical research which takes place outside the laboratory and on people with wounds (Caruso, Foster, Hermans, & Rick, 2004; Dowsett, 2003; Drosou, Falabella, & Kirsner, 2003; Lansdown, 2002). Clearly there is considerable scope and need to explore the cost effectiveness associated with the use of Silver dressings in comparison to other dressings, and especially a commonly used alternative, Cadexomer Iodine.



1.2.3.3 Cadexomer Iodine

Though Cadexomer Iodine has been a treatment modality for some time, reflected in the age of many of the Cadexomer Iodine primary studies, both Cadexomer Iodine and Silver dressings are presented as “advanced technologies incorporated into dressings [which have] added new dimensions to infected wound therapy” (Campton-Johnston & Wilson, 2001, pg74). Like Silver dressings, Cadexomer Iodine is characterised in the available literature as offering antimicrobial properties, yet its ability in “absorbing wound exudate and assisting in the debriding process” are also presented as key advantages of this product (Campton-Johnston & Wilson, 2001, pg74). This may highlight the importance of considering the wound characteristics when comparing Cadexomer Iodine to a Silver dressing as the treatments may perform optimally in the presence of certain wound characteristics such as exudate (Falanga, 1997, pg1).

Reviews of Cadexomer Iodine with other dressings converge with the conclusion that Cadexomer Iodine promotes wound healing more than other (non-Silver) dressings as shown by both animal and human research (Bianchi, 2001; Drosou et al., 2003; Jones & Milton, 2000; Sundberg & Meller, 1997).

With respect to human comparison studies, the use of Cadexomer Iodine resulted in improved wound healing when compared to Hydrocolloid dressing or paraffin gauze (Hansson, 1998), saline wet–to-dry compressive dressings (Holloway, Johansen, Barnes, & Pierce, 1989), zinc paste dressing (Laudanska & Gustavson, 1988), saline dressings, enzyme-based debriding agents or non-adhesive dressings (Moberg, Hoffman, Grennert, & Holst, 1983), dextranomer (Tarvainen, 1988), non-adherent dressings (Hillstrom, 1988), and other standard treatment (Skog et al., 1983).

In contrast to the many clinical trials finding Cadexomer Iodine to be more effective than alternatives, Cadexomer Iodine was not found to result in superior healing when compared to gentamicin solution, streptodornase/strepto-kinase, or dry saline gauze for diabetic foot wounds (Apelqvist & Ragnarson Tennvall, 1996) and dextranomer beads (Moss, Taylor, & Shuster, 1987).

Despite these exceptions, the weight of clinical studies have found Cadexomer Iodine to be effective and beneficial to wound healing. Individual clinical studies and reviews alike generally agree that in addition to accelerating wound healing, Cadexomer Iodine aids the development of epithelial and granulation tissue while reducing undesirable wound tissue including the presence of pus, debris, and exudate, as well as reducing pain associated with the wound.

Animal studies indicate that Cadexomer Iodine is effective in significantly reducing the number of organisms in the wound without hindering wound healing (Mertz, Davis, Brewer, & Franzen, 1994) and aiding the rate of wound epithelialisation (Lamme, Gustafsson, & Middelkoop, 1998; Mertz et al., 1994).

There have been few cost effectiveness studies conducted involving Cadexomer Iodine. What investigations have been undertaken found Cadexomer Iodine to be a more cost efficient treatment when compared to hydrocolloid dressings and paraffin gauze (Hansson, 1998) and standard treatment (Apelqvist & Ragnarson Tennvall, 1996).



As a supported, safe treatment (Falanga, 1997; Zhou, Nahm, Badiavas, Yufit, & Falanga, 2002), Cadexomer Iodine is a useful treatment for critically colonised and infected wounds. How it compares both clinically and from a cost perspective to the promising alternative of Silver dressings is unknown. This study was planned with the objective of answering this question.

1.3 RCT Aim & Objectives Given the routine use of Silver dressings and Cadexomer Iodine in everyday practice, and with no research evidence to clarify which product offers either clinical or cost benefits, this RCT was established to compare the clinical and cost effectiveness of these two treatments.

The RCT aim was stated as:

“To refine existing knowledge and protocols for lower leg ulcer management and develop an evidence based solution to the problem of bacterial colonisation and wound infection.”

The specific objectives of the RCT were:

a. To evaluate and compare the clinical effectiveness of two antimicrobial treatments (Cadexomer Iodine and Silver dressings) in the control of bacterial colonisation and infection in wounds, and to demonstrate the subsequent effect on wound healing.

b. To compare the cost effectiveness, acceptability, ease of use and utility of Cadexomer Iodine and Silver dressings in controlling critical colonisation and infection in wounds.

c. To develop an evidence-based protocol for the management of critically colonised and infected wounds.

Following is a detailed description of the method adopted for this RCT. The results have been formulated into a number of chapters each with a chapter summary. The main findings and conclusions are presented in a discussion and conclusions chapter.



2.0 Method

This chapter describes the method employed to address the research aims. Outlined are the study design, sample selection and randomisation procedures, recruitment, data collection, and management of client withdrawals, adverse events, and discharges. Factors pertinent to treatment delivery are also detailed.

2.1 Research Design The research used a parallel-group, Randomised Controlled Trial (RCT) design comparing the effects of a Nanocrystalline Silver dressing and Cadexomer Iodine on clients with a lower leg ulcer who met the study criteria. The RCT was conducted at two of Australia’s largest community nursing agencies: Royal District Nursing Service (RDNS) in Victoria and Silver Chain Nursing Association (Silver Chain) in Western Australia (WA). It was an open label trial as treatments were not blinded from study participants, data collectors, or the research team. It was emphasised in all study communications that both products were commonly used dressings and as there had been no prior direct trial comparison it was unknown whether one performed better than the other.

Clients were randomised to receive one of the two antimicrobial treatments. Randomisation lists were generated using the random number function in Microsoft Excel™ and participant allocation used concealed, numbered envelopes. The products selected for inclusion in the RCT were Acticoat® (Nanocrystalline Silver) and Iodosorb® (Cadexomer Iodine) manufactured by Smith and Nephew Pty Limited. These products were selected for the trial because they were well represented and emerged favourably from the existing literature. They were also familiar to and well regarded by clinicians at both sites. It is important to note, however, that interest was in the two generic antimicrobial treatment types rather than particular commercial products. As each antimicrobial treatment offered a range of application modes, the particular dressing (within the treatment group) selected for each client was determined by the clinician according to wound characteristics such as moisture levels. The progression from eligibility to the establishment of the treatment plan for each recruit to the study is illustrated in Figure 2.1. The Nanocrystalline Silver product range included Acticoat®, Acticoat Absorbant® and Acticoat 7®. Options in the Cadexomer Iodine product range were Iodosorb® Ointment and Iodosorb® Powder.

These products are collectively referred to as Silver and Iodine throughout the remainder of the report except for discussion pertaining to the different applications. As detailed in Figure 2.1, with the exception of Acticoat 7® which required a weekly dressing change, all remaining products were changed every three days unless more frequent changes were clinical indicated.

All study participants were observed for 12 weeks from their time of entry to the study, or until their wound healed if this occurred first.



Determine which product is

most appropriate for the wound (Prior dressing change

frequency is an indicator of moisture)

Dressing change frequency

Treatment randomised to Product to receive

Acticoat

Iodosorb

Daily

2-3 days

More than 3 day

Acticoat Absorbant®

Acticoat®

Daily

2-3

More than 3 days Acticoat 7®

Iodosorb® Powder

Every three days

Once weekly

Iodosorb® Ointment

Eligible client

Figure 2.1. Treatment allocation

2.2 Ethics Approval The RCT was funded by a philanthropic foundation and was undertaken independently of the product manufacturer. Ethics approval was received from the RDNS Research Ethics Committee and the Silver Chain Human Research Ethics Committee.

Participants in both antimicrobial treatment groups were provided with a Plain Language Statement which outlined the study and were required to complete a Consent Form prior to their participation in the study (See Appendix A for the RDNS Plain Language Statement and Consent Form).

Participation was voluntary and all participants were informed that they could withdraw from the study at any time. It was emphasised that their withdrawal would not influence the treatment received from RDNS or Silver Chain.

Clients did not receive an incentive to participate in the research. However, to ensure cost was not a barrier to the selection and use of best practice wound products in Victoria where wound products are not subsidised for all clients, wound dressings and compression bandaging were provided free of charge to clients during their participation in the study. As wound products are already subsidised in WA there was no need for the study to pay for the wound dressing products in the WA site. This ensured comparability across study sites.

Six monthly updates on the progress of the RCT were provided to the RDNS Research Ethics Committee. Annual updates were provided to the Silver Chain Human Research Ethics Committee.

The RCT was further monitored by a Data and Safety Monitoring Board that was convened on a six monthly basis during data collection to assess the trial against pre-determined criteria (Appendix B) and make a judgement on the advisability of the RCT’s continuation.




2.3 Research Sample Participants were recruited from clients receiving home nursing care at selected service centres at RDNS and Silver Chain between March 2006 and May 2007 who met the inclusion criteria and provided informed consent. Both study sites (RDNS and Silver Chain) were expected to recruit half of the required sample. A total of 14 centres or bases were involved in the RCT to recruit clients. The sample was stratified by Ankle Brachial Pressure Index (ABPI) (0.6-0.79; 0.8-1.2; >1.2) within each centre or base to ensure sufficient numbers were achieved within each subgroup.

2.3.1 Eligibility Criteria The following criteria applied for a client to be eligible for this study:

1. Lower leg ulcer (not pressure ulcer) with ankle brachial pressure index (ABPI) of 0.6 or above

2. The ‘study wound’ was the major wound and was 15cms or less in diameter

3. The client was 18 years or older

4. Had not been on a course of topical antiseptic treatment in the one week prior to recruitment

5. Was not using any antibiotics (e.g. topical, intravenous, intramuscular, oral) for either wound or other condition 48 hours prior to recruitment

6. Was not using systemic steroids (e.g. delivered oral, intravenous, intramuscular, or pre-rectal). Clients using topical and inhalation steroids could be included.

7. Did not have a diagnosis of diabetes

8. Did not have a diagnosis of malignancy related to the leg ulcer

9. Was not receiving palliative care

10. No contraindication to the treatment products

11. Was an RDNS client or Silver Chain Metropolitan Region client

12. Was not a fee for service client

13. The client had no planned absences from the service (holidays/ respite/ acute care etc) during the 12 week study timeframe

and one of the following was present…. 2

One or more signs of infection:

1. Cellulitis (pain, heat, erythema, swelling of surrounding tissues)

2. Suppuration (purulence or the presence of pus)

3. Lymphangitis

2 As derived from the literature. (Cutting & Harding, 1994; Gardner, Frantz, & Doebbeling, 2001; Gardner, Frantz, Troia et al., 2001; Schultz et al., 2004; Schultz et al., 2003; Sibbald et al., 2000)



4. Sepsis (confirmed by blood test)

5. Bacteraemia (confirmed by blood test)

or

Critical colonisation or covert infection:

6. Changes in granulation tissue (change in colour of granulation tissue, which may appear dusky, darker or bright red, hypergranulation tissue, friable or fragile granulation tissue that is prone to bleeding)

7. Increased or malodourous exudate

8. New areas of slough or wound breakdown

9. Impaired or delayed wound healing (epidermis fails to migrate across the wound bed, static, rolled or undermined wound edges or bridging segments of epithelial closure and breakdown in the tissues)

10. Increased or new pain

2.3.2 Sample Size Determination

The effect size estimate for this RCT was drawn from a number of cited studies for Cadexomer Iodine (Drosou et al., 2003) and from two Silver dressing studies using comparable wound aetiology to the Cadexomer Iodine trials (Wunderlich & Orfanos, 1991 as cited in O'Meara et al., 2001; Sibbald et al, 20073}. These estimates were based on the percentage decrease in ulcer size per week averaged across all studies. It was approximated that over a 12 week period the wound size would decrease 78.2% for those wounds being treated with Cadexomer Iodine and 92.5% for those wounds being treated with Silver. Using a power estimate of .80, a significance level of .05, and a standard deviation of 15, a one tailed power analysis suggested a minimum sample of 117 per treatment group.

To account for attrition from the study and to ensure a sufficient sample size to detect the true effect size, the trial targeted the recruitment of 180 participants per treatment group or 360 clients in total. The original design, prior to the sample size amendment detailed in Section 2.6, saw each of the two participating states target an equal number i.e. 180 clients each; 90 of these would be randomised to receive Silver and 90 to receive Iodine at each site.

2.4 Client Recruitment The attending clinician at each site provided clients (or carer/ guardian if appropriate) who met the eligibility criteria with a copy of the Plain Language Statement and Consent Form to read. The client then received a phone call within one working day to determine if a study representative could visit them to discuss study participation further. If necessary, an interpreter was accessed to explain the information and material to the client.

After the client had reviewed the Plain Language Statement and clinician had responded to any questions raised by the client, the client signed the Consent 3 Effect size obtained from pre-published data from this study.


Form if they wished to participate in the trial. The client was then randomised to a treatment group and treatment commenced. Clients not wishing to participate in the study were still recommended to receive one of the two antimicrobials as this treatment was clinically indicated and routine client care and reporting took place independently of the study.

Clients presenting with multiple wounds were recruited but only the one wound which best met the eligibility criteria was included in data collection (referred to as the ‘study wound’). When multiple wounds were present and more than one was clinically indicated for antimicrobial treatment, all clinically indicated wounds were treated with the same antimicrobial, thus eliminating the possibility of systemic effects from another treatment affecting the healing rate of the study wound. A client could be recruited only once to the study, thus study participants who experienced a healed ulcer recurring or a different ulcer developing could not repeat their study participation.

Recruitment for the RCT commenced in March 2006 and was completed in February 2007.

2.5 Protocol Amendments The Study Protocol was amended on three occasions during the study.

In April 2006 the protocol was amended with respect to the exclusion criterion of Fee for Service clients. Clients who were registered as Fee for Service yet paid for all their care and stated a willingness to participate in the study were henceforth not excluded from study participation. This amendment was to enhance access for Fee for Service clients to the study and open up recruitment opportunities.

The second protocol amendment occurred in June 2006 when, because of the difficulty recruiting sufficient numbers, the minimum sample targets within each ABPI category were withdrawn. It was agreed that the analysis that would be enabled by this minimum sample size criterion was secondary to the intention of the study and was not a priority if it was likely to compromise the achievement of the overall sample target.

In November 2006 the sample size target at Silver Chain was adjusted to 100 clients or to however many had been recruited by the end of February 2007. This change was the result of a pragmatic decision based on a slower than expected recruitment rate.

2.6 Participation in Angior RCT Study 2.6.1 Clinical Care during RCT Participation

Clients received the antimicrobial dressing until all signs of critical colonisation and infection were absent for one week. The clinician was able to select the most appropriate clinically indicated non-antimicrobial primary dressing for the wound after this time. If any one or more of the signs of critical colonisation or infection which constituted the study’s eligibility criteria were observed after this time, the antimicrobial to which the client was originally randomised was recommenced.

Treatment protocols were clear in requiring adherence to compression bandaging, standardised to Profore 4 Layer® or Profore Lite®, subject to the



ABPI by hand-held Doppler assessment. Adherence to compression bandaging was monitored so that adherence could be examined as a covariate to healing.

All other clinical care was provided in accordance to established best practice principles which were refreshed at the Victorian site prior to the commencement of the RCT with the implementation of the ‘Wound Best Practice Education’ program already implemented at Silver Chain. General Practitioners and any involved specialists, were informed of their client’s commencement on the trial, any swab results, and any treatment changes throughout the study period.

2.6.2 Adverse Events, Client Withdrawal, and Discharge The protocol required that any indication of an adverse event was to be immediately discussed with clinical experts to confirm if the event was related to the use of the antimicrobial. If deemed related, the client would cease the use of the antimicrobial and commence the alternate antimicrobial dressing. If a further adverse event was observed, the client would commence another clinically indicated dressing. All changes to treatment were recorded fortnightly in the study’s data collection forms. An adverse event did not mean the client was withdrawn from the study as monitoring and data collection continued if the client consented to this.

Clients could withdraw from the study at any time. Any client wishing to withdraw was asked for their permission to utilise the data collected in the study thus far in the analysis. If the client gave their permission and sufficient data were gathered to calculate a healing rate, the data from withdrawing clients were included in the final analyses.

Any client who was discharged to another service or into self care before they completed the study observation period was followed up by the trial co-ordinating team at each site. Where possible, the study team completed essential aspects of the fortnightly data collection pertaining to adherence to the treatment protocol, antibiotic use, and whether the wound had healed, by contacting the appropriate service. If sufficient data were gathered to calculate a healing rate either before the discharge or after a discharge period, the data for these clients were included in the final analysis. Other clients for whom there was insufficient data as a result of a discharge were classified as being ‘lost to follow up’.

2.7 Study Measures & Data Collection Forms 2.7.1 Primary and Secondary Study Measures

The primary outcome measures for this study were wound healing rate and the number of wounds healed (100% closure).

The following were secondary measures for this study:

1 Wound characteristics including percentage of tissue type in wound and signs of critical colonisation and infection

2 Bacteria type and quantity in wound




3 Health status4, nutritional risk5, medication use, and demographic characteristics

4 Cost of care and resource use

Client satisfaction and primary nurse satisfaction regarding comfort, pain and impact of treatment were also explored. The results for the latter investigation have been published (Flowers et al., 2007), and the paper related to the former is still in development. Neither is addressed in this report.

2.7.2 Quantifying Wound Size Wound size was measured using the Medseed AMWIS™ wound imaging software (Santamaria et al., 2002; Santamaria et al., 2004; Santamaria & Clayton, 2000). This software quantifies the total surface area of a wound from a tracing superimposed on a photograph either by the attending clinician or by a clinician / researcher at a later time. Photographs were taken at baseline and at fortnightly intervals throughout the study. Clinicians traced the total wound area as well as tissue areas within the wound. The AMWIS software calculated the total wound area and tissue type areas (mm2). This data were used to determine a daily healing rate. At RDNS the wound was profiled by field nurses and in most instances it was the nurse who attended the wound. At Silver Chain, the project coordinator profiled all wounds as they were returned to the central office shortly after the photo was taken. Given differences in approach used by the study sites, an inter-rater reliability investigation was conducted to contrast how reliably clinicians profile the same wound. The results of the inter-rater reliability investigation which found high inter-rater reliability for the total wound area, granulation and slough tissue measures has been published (Flowers et al., 2008).

2.7.3 Quantifying Bacterial Burden Wound cultures were performed during the study: at recruitment; when ceasing or recommencing the antimicrobial; and, at six weeks and 12 weeks if still using an antimicrobial. A sample of any bacteria in the wound was gathered using a wound swab. The swabs were taken using a Z-technique (Cooper & Lawrence, 1996) by the attending nurse and analysed by Melbourne Pathology in Victoria and CliniPath in Western Australia. The pathology agencies liaised to ensure swabs were managed and tests conducted consistently and to the usual industry standards. Gram stain and semi-quantitative analysis were conducted to identify the cultures of bacteria present and provide an estimation of the degree of colonisation. Sensitivity and resistance data were also gathered. The results were reported according to an agreed, standardised format across the two sites. All wound pathology results were forwarded to the client’s general practitioner and other involved specialists.

4 As measured by the Charlson Comorbidity Index (Age-Adjusted). (Charlson, Pompei, Ales, & MacKenzie, 1987). 5 As measured by the HACC Nutritional Risk Screening and Monitoring Tool. (Department of Human Services, 2001).


2.7.4 Data Collection Methods Data were collected at recruitment (‘Initial Data Collection Tool’ see Appendix C for Silver Chain example) and every two weeks for three months or until healing if healing occurred first (‘Fortnightly Data Collection Tool’ see Appendix D for Silver Chain example).

Although the actual data collected were comparable at both study sites, the method of data collection differed by site because of slight variations in service provision processes and infrastructure. Hard copy forms were available to clinical staff at both sites to record data. These forms once complete were sent to the site-based project co-ordinator, the forms were checked, and processed ready for data entry. All data were securely stored by each site in accordance with the advice given during the consent process. In addition to hard copy data collection, data were collected electronically in Victoria using assessment forms enabled by its client record system. Data already recorded in electronic databases as part of usual client information recording were also extracted at the conclusion of data collection.

2.8 Statistical Analysis The Statistical Package for the Social Sciences (SPSS) for MS Windows Release 16.0 (SPSS Inc., Chicago, IL) was used to analyse these data. The two treatment groups (Silver and Iodine) were compared on a number of demographic, wound and medical characteristics to identify any differences at baseline which might act as a potential confounder to the main results.

Data were checked for out of range values and assumptions of the statistical tests used prior to analysis. All cases were included in analyses except when missing data made inclusion not possible. Untransformed data were used if the data were relatively normally distributed and if the impact of any deviation from the normal distribution was judged to be negligible.

An alpha level of 0.05 was used to classify findings as significant. In assessing statistical significance, when the SPSS output = .000, it was interpreted as meaning p< .0005 (Pallant, 2007).

Different methods of analysis were used according to the research question being investigated and the nature of the data. Analysis of covariance (ANCOVA), Linear Mixed Model (LMM), Survival analysis, Cox regression, and Mann-Whitney U analysis were all used for different sections of the analyses.

Linear Mixed Models (LMM) and Analysis of Covariance (ANCOVA) were used to test the main hypotheses of the study using healing rate data. LMM expands the general linear model so that the error terms and random effects are permitted to exhibit correlated and non-constant variability. LMM, therefore, provides the flexibility to model not only the mean of a response variable, but its covariance structure as well. LMM is an extension of the general linear model, in which factors and covariates are assumed to have a linear relationship to the dependent variable (Singer & Willet, 2003). ANCOVA is an extension of analysis of variance and is used to examine the mean differences between two or more groups but can account for inter-group variation associated not only with the treatment itself, but with covariates as well (Pallant, 2007).



In the current study, the LMM simultaneously considered healing rates for each fortnight in the single analysis. It identified if there was an impact of the antimicrobial treatment on the healing rate overall – considering all data at the same time. It considered the overall trend. In contrast, ANCOVA examined only at a single point in time to see if there was a significant difference in healing rate between the two antimicrobials. ANCOVA was used for each fortnight separately and for an overall study healing rate that was calculated. ANCOVA differed from LMM as the assessment of each individual point in time is not influenced by the data from other points in time, but then it cannot assess what trends or pattern of activity occurred the way LMM can. ANCOVA was also used in the cost analysis.

Survival analysis was used to assess time to wound healing between both treatment groups. This analysis is a technique which examines the time it takes for something to happen (e.g. a cure, a failure) (Tabachnick & Fidell, 2007). Kaplan Meier analysis and Cox Regression were used to examine the time to healing between two treatment groups and determine the predictors of time to heal.

Mann-Whitney U test (Mann-Whitney-Wilcoxon) is a non-parametric test which is used as an alternative for the two sample t-test to examine if a difference exists between two groups. This test was used to compare the signs of wound colonisation presented as categorical data between two treatment groups and wound swab data.


http://en.wikipedia.org/wiki/Non-parametric_statistics





140 clients allocated to Silver dressing 141 clients allocated to Iodine dressing

Client discharged/ lost to follow up (1)

Client withdrawal (insufficient data for analysis) (5)

Client died unrelated to ulcer (1)

TOTAL CLIENTS EXCLUDED (7)

Client discharged/ lost to follow up (3)

Client withdrawal (insufficient data for analysis) (5)

TOTAL CLIENTS EXCLUDED (8)

133 included in analysis 133 included in analysis

281 clients randomly allocated to treatment groups

3.0 Results: Sample Description & Comparison

This chapter profiles the sample included in the main analysis and compares the treatment groups to identify demographic or wound characteristic differences that may have influenced the primary outcome findings. A similar profiling and comparison exercise was undertaken to explore the sample segmentations pursued during this study and is also detailed in this chapter.

3.1 The RCT Sample Equivalent numbers were randomly allocated to the treatment groups; 140 participants to Silver and 141 to Iodine. Figure 3.1 depicts the random allocation of clients to treatments and follow through to analysis. All clients with at least two wound measurements, for whom a healing rate could therefore be calculated, were included in the final analysis in the treatment arm to which they were originally randomised even if a change had occurred (as per intention-to-treat6). With a total of 15 clients excluded from the analysis on these grounds (seven from the Silver treatment group and eight from the Iodine treatment group) 266 or 95% of clients were incorporated into the final analysis. In addition to the intention-to-treat analysis to explore the population effects, efficacy subset analysis7 was also undertaken to explore the specific treatment effects of the antimicrobial dressings.

Figure 3.1. Participant flowchart

6 (Referred to as ‘intention-to-treat’ or ‘intent-to-treat’): “The intent-to-treat principle refers to a set of criteria for the evaluation of the benefits and risks of a new therapy that essentially calls for the complete inclusion of all data from all patients randomized in the final analysis.” (Lachin, 2000) 7 “Patients and observable patient data are excluded from the analysis on the basis of information obtained postrandomization.” (Lachin, 2000)


Twenty-one adverse events were recorded during the study period (eight related to Iodine, 13 related to Silver). Fifteen of these adverse events related to unacceptable levels of pain or discomfort which could not be managed. Two were identified as allergic reactions (generalized) or local sensitivity (local signs of irritation only). The remaining four were classified as any other negative clinical symptoms, most typically a general wound deterioration, assessed as reflecting an adverse event attributable to the antimicrobial treatment. In instances where clients had an adverse event to both antimicrobials, both incidents are represented in the total. The nature of the adverse events did not differ from normal clinical experiences with these products.

3.1.1 Demographic and Comorbidity Profile The sample can be described as: elderly, 80 years of age on average; slightly more likely to be female (58.6%); Australian born (75.2%); English speaking (98.1%); and presented with a range of different living situation circumstances (Table 3.1).

Despite being in their senior years and in the receipt of home nursing care, the sample was defined by low scores on the Age-adjusted Charlson Comorbidity Index (ACCI). Where a maximum score 40 is indicative of high risk of mortality, the trial sample had an average index score of 4.40 (Table 3.2). The conditions most commonly reported at recruitment included peripheral vascular disease (19.5%) followed by congestive heart failure (16.5%). This sample consisted of non-smokers in the main (92.4%).

The risk of malnutrition among the sample was low with an average score of 0.9 on the HACC Nutritional Risk Screening and Monitoring Tool for which a score of zero indicates no risk and a score of 11 indicates high risk (Table 3.3). The most common nutritional concerns included needing assistance to shop for food (32.0%) and needing assistance to prepare food (18.4%). One in ten clients were scored as having ‘obvious overweight affecting life quality’ (10.9%).

Consistent with the nutritional risk results, ratings of dependence on items of Activities of Daily Living (ADL) and Independent Activities of Daily Living (IADL) were high for shopping (60.0% dependent) and moderately high for food preparation (37.8% dependent) (Table 3.4). Clients were also more likely to be dependent for housekeeping (65.7%), transportation (54.7%) and laundry (47.5%). Overall, 34.2% of clients required carer assistance. Though shopping and food preparation were both flagged as areas requiring assistance in the ADL / IADLs and HACC Nutritional Risk Screening and Monitoring Tool measures, the proportion of clients requiring assistance varied considerably between these tools. Potentially accounting for this variation are differences in the time at which these questions were administered; ADLs and IADLs are completed upon admission to the service with the HACC Nutritional Risk Screening and Monitoring Tool completed upon recruitment to the RCT, a slightly different response format and the number of categories available to the respondent, and differences in the item wording; ‘shop for food’ was explicit in the HACC Nutritional Risk Screening and Monitoring Tool while the ADL/ IADL measure was generic to ‘shopping’.

The demographic data for the two treatment groups were compared to assess their equivalence and identify any differences between the groups that could potentially confound the results. Where the treatment groups were determined to



vary significantly on any of the dimensions considered in this study, the variable was included as a covariate in the analyses to control for these differences when it was found to have a significant association with the dependent variable.

The treatment groups did not differ significantly for:

• Gender [χ2(1)=0.76, p>0.05]

• Age [t(264)=-0.45, p>0.05]

• Born in Australia [χ2(1)=0.99, p>0.05]

• English as first language [χ2(1)=0.00, p>0.05]

• Living Situation [χ2(7)=5.52, p>0.05]

• Charlson Comorbidity Index [χ2(10)=13.84, p>0.05]

• Nutritional Risk Factors [χ2(4)=2.48, p>0.05]

Furthermore, there were no significant differences on any items of Activities of Daily Living or Independent Activities of Daily Living.

Table 3.1

Demographic Descriptors of the Treatment Groups (n=266)†

†Sample includes all clients included in the intention-to-treat analysis

Silver Iodine Total (n=133) (n=133) (n=266)

Ave 79.35 79.99 79.67 Age (years)

SD 12.66 10.90 11.80 Gender (%)

Male 44.4 38.3 41.4

Female 55.6 61.7 58.6

Born in Australia (%) Yes 78.2 72.2 75.2

No 21.8 27.8 24.8

English is first language (%)

Yes 97.7 98.5 98.1

No 2.3 1.5 1.9

Living Circumstances (%) Carer not needed* 17.3 17.3 17.3

Lives alone, has carer 3.0 8.3 5.6

Lives alone, has no carer 31.6 32.2 32.0

Lives with another, has no carer 12.8 14.3 13.5

Lives with another, has residential carer 22.6 15.0 18.8

Lives with another, has non-residential carer 1.5 1.5 1.5

Lives in mutually dependant situation* 5.3 6.0 5.6

Lives in supported accommodation* 6.0 5.3 5.6

* RDNS only code



Table 3.2

Comorbidity Index Results for the Treatment Groups (n=266)†

Silver Iodine Total

(n=133)^ (n=133)^ (n=266)^

Charlson Comorbidty Index Ave 4.29 4.51 4.40

SD 1.88 1.71 1.80

Charlson Comorbidty Index items (% yes)

Cerebrovascular disease 6.8 11.3 9.0

Chronic pulmonary disease 6.8 6.8 6.8

Congestive heart failure 12.0 21.1 16.5

Connective tissue disease 2.3 0.8 1.5

Dementia 2.3 7.5 4.9

Gastric ulcer disease 3.0 3.0 3.0

Hemiplegia 0.8 0.0 0.4

Lymphoedema 0.8 3.0 1.9

Myocardial infarction 6.8 5.3 6.0

Peripheral vascular disease 16.5 22.6 19.5

Rheumatoid arthritis 7.5 3.8 5.6

HIV/ AIDS - - -

Leukaemia - - -

Malignant lymphoma - - -

Liver disease

None 99.2 99.2 99.2

Mild 0.8 0.8 0.8

Renal disease

None 93.0 94.6 93.8

Mild 5.4 4.7 5.0

Moderate 0.8 0.8 0.8

Severe 0.8 - 0.4

Malignant solid tumor

None 96.9 97.7 97.3

Non-metastatic 2.3 2.3 2.3

Metastatic 0.8 - 0.4

Smokes

No 90.1 94.7 92.4

Yes 9.9 5.3 7.6

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data



Table 3.3 HACC Nutritional Risk Scores for the Treatment Groups (n=266)†

Silver Iodine Total (n=133)^ (n=133)^ (n=266)^ Nutritional Risk score Ave 0.78 0.95 0.87 SD 0.95 1.10 1.03 Proportion with <1.0 risk item % 50.4 54.9 52.6 Nutritional Risk Screening Tool items (% yes)

Obvious underweight frailty 9.8 3.8 6.8 Unintentional weight loss 4.5 6.8 5.6 Reduced appetite/ food or fluid intake 4.5 3.8 4.1 Mouth or teeth problems 1.5 3.0 2.3 Swallowing problems 0.8 1.5 1.1 Follows a special diet 2.3 0.8 1.5 Needs assistance to shop for food 27.8 36.1 32.0 Needs assistance to prepare food 17.3 19.5 18.4 Needs assistance to feed self - 0.8 0.4 Obvious overweight affecting life quality 8.3 13.5 10.9 Unintentional weight gain 1.5 5.3 3.4


Table 3.4 Activities of Daily Living (ADL) and Independent Activities of Daily Living (IADL) Results for the Treatment Groups (n=266)†

ADLS & IADLS (% dependent) Silver Iodine Total (n=133)^ (n=133)^ (n=266)^ Mobility bed 4.6 8.5 6.6 Transferring 13.9 17.5 15.7 Toileting 5.5 6.0 5.7 Hygiene 27.9 25.6 26.7 Dressing 17.3 16.4 16.8 Eating 4.5 2.6 3.5 Telephone 9.2 10.3 9.8 Food preparation 38.4 37.3 37.8 Housekeeping 60.0 71.2 65.7 Laundry 49.1 46.0 47.5 Shopping 57.4 62.6 60.0 Medication 17.4 14.4 15.9 Transport 51.0 58.5 54.7 Finances 30.8 21.1 26.0 Carer Assistance 31.4 36.9 34.2

†Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data



3.1.2 Wound Characteristic Profile The vast proportion of wounds were located on the lower leg (97.0%) with the remainder recorded as being specifically located on the ankle or foot (Table 3.5). Wounds were diagnosed as ‘venous’ in the main (73.7%) with the remainder ‘mixed’. Three quarters had an ankle brachial pressure index (ABPI) of 0.8-1.2. With ABPI a principle (but not absolute) indicator of venous disease it is not surprising that the proportion of participants with a 'normal' ABPI, assessed in this study as 0.8-1.2, closely mirrors the proportion of participants with venous diagnosis. With respect to those participants for whom a venous diagnosis did not match the expected ABPI, it may be that adjunct assessments identified other etiological factors that influenced the final diagnosis.

Wounds at recruitment varied greatly with respect to their longevity. One wound was less than a week old when recruited to the study, whereas 31 wounds had not healed for more than one year (12.1%). The median duration for all wounds in the trial was 12.0 weeks at recruitment.

Prior to commencing the trial the use of compression bandaging was inconsistent, with a third of clients (33.2%) (Table 3.6) not having compression bandaging as part of their wound management immediately prior to entering the trial. Approximately a quarter of the sample were using moderate (26.7%) or light (27.9%) compression, with only 11.5% having high compression applied.

Table 3.5

Wound Characteristics for the Treatment Groups (n=266)†

Silver Iodine Total

(n=133) (n=133) (n=266) Wound duration (weeks) Median 12.00 12.00 12.00 Ave* 58.06 49.38 53.72 SD 260.61 165.90 218.08 Wound ‘leg’ (%)

Right leg 47.4 49.6 48.5 Left leg 52.6 50.4 51.5

Wound location (%) Lower leg 97.7 96.2 97.0 Ankle / foot 2.3 3.8 3.0

Wound diagnosis (%) Venous ulcer 75.9 71.4 73.7 Mixed ulcer 24.1 28.6 26.3

ABPI score (%) 0.6-0.79 7.5 12.8 10.2 0.8-1.2 76.7 72.9 74.8 >1.2 15.8 14.3 15.0

† Sample includes all clients included in the intention-to-treat analysis *The distribution of this data are skewed and as such the median is suggested as more reliable

indicator of central tendency



As can be seen in Table 3.6, a range of primary and secondary dressings were recorded as being in use prior to the study, although half the sample had a non-adherent dressing for their secondary dressing (50.4%). There was greater variety in the primary dressings than the secondary dressings, and these included: tulle (19.9%); non-adherent dressings (15.4%); hydrocolloids (14.7%); and, hydrogels (10.5%).

Clients became eligible for the study if their wound showed at least one of ten possible signs of critical colonisation or infection. In general the wounds recruited demonstrated multiple signs, the most common including impaired or delayed wound healing (88.0%) and new areas of slough or wound breakdown (67.7%) (Table 3.7). Changes in granulation tissue and increased or malodorous exudate were also common in these study wounds (51.1% and 45.5% respectively).

Table 3.6

Prior Wound Dressings for the Treatment Groups (n=266)†

† Sample includes all clients included in the intention-to-treat analysis

Prior Dressing (%) Silver Iodine Total

(n=133)^ (n=133)^ (n=266)^ Primary Dressing

Non-adherent dressing 12.8 18.0 15.4 Hydrogel 9.8 11.3 10.5 Hydrocolloid 15.8 13.5 14.7 Alginate 7.5 9.0 8.3 Foam 10.5 6.8 8.6 Tulle 22.6 17.3 19.9 Other 21.1 24.1 22.6

Secondary Dressing Non-adherent dressing 46.1 54.9 50.4 Foam 10.4 7.1 8.8 Zetuvit 9.6 8.0 8.8 Other 33.9 30.1 32.0

Compression Bandaging Nil 32.3 34.1 33.2 Light 27.8 27.9 27.9 Moderate 27.1 26.4 26.7 High 11.3 11.6 11.5 Unknown 1.5 - 0.8

^Sample size for question(s) vary due to missing data



Table 3.7

Distribution of the Eligibility Criteria for the Treatment Groups (n=266)†

Eligibility Criteria (% yes) Silver Iodine Total

(n=133) (n=133) (n=266) Cellulitis (pain, heat, erythema, swelling) 24.1 25.6 24.8 Changes in granulation tissue 54.1 48.1 51.1 Impaired or delayed wound healing 88.0 88.0 88.0 Increased or malodorous exudate 43.6 47.4 45.5 Increased or new pain 34.6 31.6 33.1 New areas of slough or wound breakdown 71.4 63.9 67.7 Suppuration (purulence / pus) 6.8 11.3 9.0 Lymphangitis - - - Bacteraemia 0.8 - 0.4 Sepsis 6.8 3.8 5.3


The study protocol restricted inclusion to wounds of 15cm in diameter or less; a pragmatic, cost/resource driven decision. The average wound size was 704.66(mm2) and the most extensive wound characteristics were slough (327.97mm2) and granulation tissue (281.74mm2). Wound depth was typical of venous and mixed venous arterial ulcers with an average depth of 1.40cm.



Table 3.8

Wound Dimensions at Baseline for the Treatment Groups (n=266)†

Wound Dimensions Silver Iodine Total

(n=133)^ (n=133)^ (n=266)^

Ave 596.96* 812.37* 704.66 Total wound surface area (mm2)

SD 631.78 1065.23 880.71

Granulation area (mm2) Ave 219.06* 343.47* 281.74

SD 348.90 537.31 457.11

Slough area (mm2) Ave 274.61 381.34 327.97

SD 465.55 692.08 591.09

Necrotic area (mm2) Ave 9.69 4.97 7.33

SD 61.59 25.63 47.14

Hypergranulation area (mm2) Ave 9.77 8.23 9.02

SD 47.75 42.78 45.26

Infection area (mm2) Ave 11.01 0.96 5.99

SD 114.29 11.05 81.20

Undermining area (mm2) Ave - - -

SD - - -

Epithelialisation area (mm2) Ave 38.67 47.51 43.11

SD 104.49 224.70 175.16

Wound Depth (cm) Ave 1.57 1.23 1.40

SD 1.90 1.30 1.63

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data * Significant <0.05 ** Significant <0.01 *** Significant <0.001 Wounds were classified as having low (41.0%) to moderate (41.4%) levels of wound exudate (Table 3.9). A further 14.7% were heavily exudating wounds. Pain intensity was assessed on a pain scale where zero indicated no pain and ten indicated severe pain. Three quarters of clients (73.4%) reported pain, which was most commonly described as intermittent (47.5%). An average pain score of 3.2 was reported.



Table 3.9

Exudate and Pain Levels for the Treatment Groups (n=266)†

Silver Iodine Total

(n=133)^ (n=133)^ (n=266)^

Amount of Wound Exudate (%)

Not applicable/ other/ unknown 0.8 2.3 1.5

Low 39.1 42.9 41.0

Moderate 44.4 41.4 42.9

Heavy 15.8 13.5 14.7

Pain Intensity (scale 0-10) Ave 3.27 3.18 3.23

SD 2.69 2.67 2.67

Pain Frequency (%)

None 26.0 27.3 26.6

Only at intervention/ time of dressing 9.9 7.6 8.7

Intermittent 51.1 43.9 47.5

Constant 6.1 7.6 6.8

Nocturnal 3.8 9.8 6.8

Other 3.1 3.8 3.4

Unknown 1.5 - 0.8

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data The treatment groups were compared on the number of characteristics pertaining to their lower leg ulcer. The treatment groups did not differ significantly for:

• Leg the ulcer was on [χ2(1)=0.06, p>0.05]

• Location of the ulcer on the leg [χ2(9)=7.20, p>0.05]

• Wound diagnosis [χ2(1)=0.49, p>0.05]

• Prior compression bandaging [χ2(4)=2.02, p>0.05]

• Prior primary dressing [χ2(6)=4.03, p>0.05]

• Prior secondary dressing [χ2(3)=2.03, p>0.05]

• ABPI [χ2(2)=2.04, p>0.05]

• Wound duration [t(264)=0.32, p>0.05]

• Wound exudate level [χ2(3)=1.60, p>0.05]

• Pain intensity [t(262)=0.27, p>0.05]

• Pain frequency [χ2(5)=4.99, p>0.05]

Finally, the treatment groups were compared in terms of the average wound size at baseline. Wounds for the Silver treatment group were smaller on average (596.96mm2) compared to those wounds allocated into the Iodine treatment



group (812.37mm2) [t(213)=-2.00, p<0.05]. There was no significant difference between the groups in terms of the proportion of tissue types within the wounds including slough, necrotic, infected, and hypergranulated tissue. There was, however, a significant difference in the quantity of granulated tissue [t(225)=-2.23, p<0.05], with wounds assigned to the Silver treatment group having less granulation tissue (219.06mm2) than wounds assigned to the Iodine treatment group (343.47mm2). These differences in wound characteristics and size between the treatment groups were not considered to be a major concern as healing rate, which is calculated in reference to baseline size of each wound, was the primary outcome measure used. However, a secondary measure of interest, the number of wounds healed, was considered to be potentially more susceptible to the differences between the baseline size of the wounds in the two groups. To adjust for this, baseline wound size was used as a covariate in all analyses where wound size was identified as having a strong, significant relationship with the dependent variable.

There was no significance difference between the treatment groups with respect to wound depth [t(259)=1.72, p>0.05].

3.2 Segmentation Group Comparison In addition to conducting the analysis for the sample as a whole, three key segmentations were explored in this study. This chapter presents the comparison of the segmented samples and examines any differences between treatment groups within each segment.

The segmentations explored included whether the wound healed during the 12 week observation period or not (healed and not healed); wound age at recruitment; and wound size at recruitment.

The first segmentation (healed / not healed) was in response to studies which have identified that the trajectories for clients who heal and do not heal differ significantly (Robson, Hill, Wooske, & Steed, 2008; Steed, Hill, Woodske, Payne, & Robson, 2006). As such, these groups were analysed separately to determine if any differences in outcomes would be detected.

Wound age and wound size at recruitment were both clinically informed segmentations; there was interest to examine the impact of the antimicrobial dressings separately for relatively new wounds compared to chronic wounds, and small compared to large wounds. Frequency data were examined to determine the appropriate point of segmentation. The wound age data presented with a skewed distribution and as such the median of 12 weeks of duration at recruitment was used to create two categories; one of ‘younger’ wounds and one of ‘older’ wounds. The median was also used to categorise wounds as either small (equal to or less than 3.6cm2) or large (greater than 3.6cm2).

3.2.1 Healed and Not Healed Group Comparison A series of chi-square and t-tests were conducted to directly compare the ‘healed’ and ‘not healed’ samples on the various demographic and wound characteristics dimensions measured by the trial to determine if these two groups differed significantly from each other on any of these dimensions.



The groups differed significantly at baseline for total wound size, wound duration at recruitment, wound exudate levels, the presence of cellulitis, the risk of being overweight, and the prior primary dressing.

It is perhaps not surprising that wounds that went on to heal during the 12 week observation period were smaller in size at baseline than wounds that did not [t(149)=-4.07, p<0.001]. Indeed, wounds that healed had an average wound size of 528.08mm2 compared with a wound size of 1013.69mm2 for wounds that did not heal during the observation period. Wounds that healed had a median age of eight weeks at recruitment compared with 24 weeks for wounds that did not heal [t(97)=-4.88, p<0.01].

Wounds that healed had less wound exudate than did wounds that did not heal [χ2(3)=17.07, p<0.001]. While almost one quarter of wounds (23.7%) that did not heal presented with ‘heavy’ levels of exudate at baseline, only 9.5% of the healed group were regarded as having ‘heavy’ exudate.

Interestingly, wounds that healed were slightly more likely, when compared to wounds that didn’t heal, to have cellulitis noted as a sign of infection at recruitment [χ2(1)=4.98, p<0.05]. Three in ten wounds that healed had this sign present (29.6%) compared to 16.5% of wounds that did not heal during the observation period. Though cellulitis is a sign of infection and therefore perhaps a reason why healing should be delayed in these wounds, the impact of this significant difference is perhaps less sizable when it is considered that cellulitis was one of several signs of bacterial burden used in this study to determine eligibility for recruitment and to guide continued antimicrobial use during the monitoring period, and because clients were ineligible if infection had been medically diagnosed and treated.

The not healed group was also significantly more likely be rated as having ‘obvious overweight affecting life quality’ on the HACC Nutritional Risk Screening and Monitoring Tool compared to the healed group [χ2(1)=4.05, p<0.05]. Though still a small proportion of the entire segment, 16.5% of the not healed group were rated as having this nutritional risk factor compared to 7.7% of the healed group.

The healed and not healed wound groups also differed significantly on the primary dressing that was in use prior to recruitment [χ2(6)=18.47, p<0.01]. While there were a great variety of dressings in use pre-recruitment, noteworthy variations appeared in the use of alginate and foam dressings. There was higher use of alginate dressings pre-recruitment among wounds that did not heal (15.5%) compared to wounds that did (4.1%), while there was a higher use of foam dressings among wound that healed (11.8%) compared to wounds that did not (3.1%).

3.2.2 Comparing the Treatment Groups within the Healed and Not Healed Segments The demographics and wound characteristics of the individuals within the healed and the not healed segments of the Silver and Iodine treatment groups were compared to identify any possible confounders.

Within the healed segment the treatment groups differed significantly only in terms of their ABPI scores [χ2(2)=6.84, p<0.05]. There were fewer individuals with a low ABPI of 0.6-0.79 in the Silver group (4.7%) compared to the Iodine



group (16.7%), and slightly more individuals with an ABPI greater than 1.2 in the Silver group (17.6%) compared to the Iodine group (11.9%). The majority of individuals in both antimicrobials groups had an ABPI in the mid-range of 0.8-1.2 (77.6% for Silver; 71.4% for Iodine).

Significant differences between the treatment groups within the not healed segment were identified on two variables. Reflecting the finding already described for the sample as a whole, there was a significant difference in wound size between the treatment groups. The Silver group’s wounds were smaller on average (800.21mm2) than the wounds in the Iodine group (1218.45mm2) [t(94)=-2.04, p<0.05].

Clients in the Silver group were also more likely to have been assessed as having obvious underweight frailty, with 14.6% of this group considered at nutritional risk on this basis, whereas no clients in the Iodine group were assessed in this way [χ2(1)=5.68, p<0.05].

3.2.3 Wound Duration Group Comparison The demographic and wound characteristics of individuals segmented by the age of their wound at recruitment into wounds 12 weeks or less duration (‘young’ wounds) and wounds 12 weeks or longer in duration (‘old’ wounds) were compared using chi-square or t-tests, depending on the measure and its distribution.

Older wounds tended to be larger overall [t(204)=-2.44, p<0.05] and to have more granulation tissue [t(183)=-2.09, p<0.05] than young wounds. There was a significant difference between the young and old wounds according to the presence of the eligibility criterion of impaired or delayed wound healing with 95.0% of old wounds with this characteristic compared to 82.2% of young wounds [χ2(1)=9.04, p<0.01].

The presence of pain also distinguished individuals with young and old wounds, with clients with old wounds reporting less pain, an average of 2.76 on a scale of 0-10 (where 0 means no pain), than clients with young wounds who reported an average pain level of 3.60 [t(262)=2.57, p<0.05].

There was a significant difference in the number of nutritional risk factors identified for the two wound age groups [χ2(4)=10.50, p<0.05]; a difference which favoured more risk factors for older wounds, though an inspection of the frequency data would suggest these differences were small.

Old and young wounds differed significantly in terms of the primary dressing that was being applied prior to recruitment [χ2(6)=18.46, p<0.01]. There was greater use of non-adherent dressings (21.2%) and foam dressings (11.0%) on young compared to older wounds (8.3% and 5.8% respectively). Older wounds were more likely to have been dressed with a hydrogel (13.3%) or alginate dressing (13.3%) compared to young wounds (8.2% and 4.1% respectively).

More females had young wounds (67.1%) whereas more men had old wounds (51.7%) a difference that was statistically significant [χ2(1)=8.83, p<0.01].

3.2.4 Comparing the Treatment Groups within the Wound Duration Segments The demographics and wound characteristics of the Silver and Iodine treatment groups were compared within the wound duration segments to identify any



differences between the groups by segment that could confound the differences between the groups in terms of the main outcome measures.

There were no significant differences in client demographics and wound characteristics between treatment groups for young wounds.

For old wounds on the other hand, there was a difference in wound size at baseline which approached significance for total wound size [t(118)=-1.93, p>0.05] and was significant for the amount of granulation tissue [t(118)=-2.07, p<0.05]. There was also a significant difference between the treatment groups on the ADL item of hygiene [χ2(3)=6.83, p<0.05]. Almost one in five clients (19.4%) in the Iodine treatment group were dependent on this dimension compared to 5.4% of clients in the Silver treatment group.

3.2.5 Wound Size Group Comparison Wounds were segmented into small wounds (less than or equal to 3.6cm2) and large wounds (greater than 3.6cm2) based on their wound measurements at recruitment. The treatment groups were again compared demographically and in terms of their wound characteristics.

Small and large wounds differed by client age, type of secondary dressing used prior to recruitment, and whether the wound had shown signs of increased or malodorous exudate at recruitment.

Clients with large wounds were younger (77.8 years) than clients presenting with smaller wounds (81.6 years). Larger wounds were more likely to have been treated prior to the RCT with a non-adherent secondary dressing (55.7%) compared to smaller wounds (44.1%) for whom an increased use of absorbent dry dressings (13.5%) and foam (10.8%) compensated for this difference.

Larger wounds were more likely to have been recruited to the trial with increased or malodorous exudate (52.3%) compared to smaller wounds (38.6%) [χ2(1)=4.42, p<0.05].

3.2.6 Comparing the Treatment Groups within the Wound Size Segments The Silver and Iodine treatment groups were compared on the range of demographic and wound characteristic measures within the wound size segments.

With respect to the small wound segment, clients randomised to Iodine were more likely to report Congestive Heart Failure (22.2%) than clients randomised to Silver (7.2%) [χ2(1)=4.84, p<0.05]. At baseline, there was more slough present in wounds treated with Silver than wounds being treated with Iodine [t(125)=2.17, p>0.05].

There was a significant difference in total wound size [t(107)=-2.10, p>0.05] and amount of granulation tissue [t(121)=--2.07, p<0.05] at baseline in the segment with large wounds, the Iodine group reporting a larger wound area with a greater quantity of granulation tissue than the Silver group.

3.3 Chapter Summary The sample included in the RCT was older, principally Caucasian, and tended to fare well with respect to co-morbidity status. All had a lower leg ulcer that were



primarily of venous aetiology, low to moderately exudating, and presenting with multiple signs of critical colonisation and infection, most notably impaired or delayed healing and new areas of slough or wound breakdown. Pain intensity was rated as low on average with one quarter of clients not experiencing wound pain, and the remaining clients mostly describing their pain as intermittent.

The two antimicrobial groups were comparable on the vast majority of dimensions measured by the RCT. Treatment groups were comparable for all demographic measures. The groups differed significantly on overall wound size and amount of granulation tissue at baseline, both being smaller in the Silver treatment group than the Iodine group. As the wound healing rates are calculated taking into account the initial wound size, the impact of any differences at baseline are negligible.

The sample was segmented during analysis to consider the healed / not healed, wound age at recruitment and wound size groups.

The wounds that healed were more likely than wounds that did not heal during the observation period to be smaller in size, of shorter duration at recruitment, and less exudating. They also presented with more signs of cellulitis and these clients were less likely to be rated with ‘obvious overweight affecting life quality’. There was a significant difference between the treatment groups for the healed segment for ABPI score, while there was a significant difference in the treatment groups for the not healed segment for wound size at baseline and risk of having obvious underweight frailty.

Old wounds were larger, more likely to be classified as having impaired or delayed healing, and the individuals with these wounds having more nutritional risk factors than individuals with young wounds. Younger wounds were more painful than older wounds. There were no significant differences between the treatment groups for young wounds. The treatment groups for older wounds differed significantly for wound size at baseline and the level of dependence on the ADL item of hygiene.

Larger wounds were more likely to have increased or malodorous exudate at recruitment and were more likely to be treated with a non-adherent secondary dressing prior to recruitment. Individuals with these wounds were younger than other clients. In the small wound segment, significant differences were found between the treatment groups in terms of the frequency of a diagnosis of congestive heart failure and the quantity of slough tissue at baseline. Whereas in the large wound segment the treatment groups differed significantly in total wound size and amount of granulation tissue at baseline.

Variables for which the treatment groups differed significantly were controlled for throughout the analysis as covariates whenever a relationship with the dependent variable was detected.





4.0 Results: Wound Healing Outcomes

Prior to each analysis the data were examined for compliance with the assumptions for each of the statistical methods used. Transformed variables were included in the analysis if the normal distribution assumption was violated. The need to include transformed variables applied only to covariates; the original dependent variable was used in each instance because the normal distribution assumption was not violated and to enhance the meaningfulness and interpretability of the results. A series of Pearson’s R correlations for continuous and bi-variate data, and chi-square analyses for remaining categorical variables were performed to identify variables associated with the dependent variable. To achieve a parsimonious analysis solution covariates were only included in the analysis if they were identified to have a strong and significant association with the dependent variable. Variables which were identified as being significantly different between the treatment groups either overall or for any of the planned segmentation analyses were included as a covariate if shown to be associated with the dependent variable.

The healing rate is presented as a per day healing rate to accommodate slight variations in the timing of the visit and therefore the data collection and signifies a rate of change in the wound size daily with a positive result indicating a reduction in wound size and a negative healing rate suggesting the wound area increased between data collection points. Healing rates were then calculated to represent healing for each fortnight of the study period and an overall healing rate was calculated from baseline to the client’s completion of the study either after 12 weeks of observation or when the wound healed.

Analysis of continuous dependent variables was conducted using Linear Mixed Models (LMM) to assess the overall trend of the data and Analysis of Covariance (ANCOVA) to consider the healing rates for each fortnight including the final (overall) healing rate separately, uninfluenced by the pattern of healing at other times during the observation period. Significance among covariates is presented in the tables for both the main and segmentation analyses.

As detailed in Section 3.2 of this report, segmentation analysis was pursued, sample size permitting, to investigate the healed/ not healed, wound duration and wound size comparisons. These segments were compared to determine if the action of the antimicrobial treatment groups were comparable for each and are presented alongside the main results in this chapter.

4.1 Healing Rate Analysis 4.1.1 Descriptive Statistics for Healing Rate

The means and standard deviations for healing rate for each fortnight and overall are presented in Table 4.1 for each of the treatment groups. The mean wound healing rates were similar for the Silver and Iodine groups with Silver recording a marginally higher healing rate (Ave=2.10; SD=1.89) compared to Iodine (Ave=1.69; SD=2.46). The healing rates, when plotted across the fortnightly assessments, reveal considerable fluctuation by fortnight for both



treatment groups and notably, a negative average healing rate for the Iodine group at fortnight one (Figure 4.1).

Figure 4.1 Fortnightly Healing Rate by Treatment Group

2.12

1.63

2.82

0.35

-0.22

3.15

1.94

1.321.64

2.99

1.26

1.29

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

Fortnight 1 Fortnight 2 Fortnight 3 Fortnight 4 Fortnight 5 Fortnight 6

Dai

ly h

ealin

g ra

teSilver Iodine

Table 4.1

Descriptive Statistics for Healing Rate by Treatment Group (n=266)† n= Ave SD

Silver Iodine Silver Iodine Silver Iodine

Overall 133 133 2.10 1.69 1.89 2.46

Fortnight 1 129 127 2.12 -0.22 2.94 8.18

Fortnight 2 115 117 2.99 3.15 4.53 4.97

Fortnight 3 92 91 1.63 1.94 11.34 5.18

Fortnight 4 72 70 2.82 1.32 4.84 10.19

Fortnight 5 58 51 0.35 1.64 11.68 4.91

Fortnight 6 48 41 1.26 1.29 4.79 6.09


4.1.2 Linear Mixed Models (LMM) Analysis LMM was used to compare the overall healing rate trend for the antimicrobial treatment groups. Nutritional factors, adherence to compression bandaging, and wound characteristics (wound exudate and slough) were entered as covariates. LMM identified no significant difference between the treatment groups as regards the rate of wound healing [F(1,234)=0.36, p>0.05] (Table 4.2). While there was also no significant ‘time and treatment’ interaction effect [F(5,685)=1.69, p>0.05], there was a significant effect for ‘time’ [F(5,686)=2.45, p<0.05]; the effect of ‘time’ reflecting a change in healing rate across the fortnightly data collections. These results suggest that though the pattern of healing did not differ by antimicrobial treatment, wounds generally



decreased in size over time across the study period. Though significantly associated with the dependent variable, there was no significant difference in the levels of any of the covariates for this analysis.

Table 4.2

LMM analysis of Healing Rate (n=263)†^

F value

Covariates

Wound exudate 0.07

Wound slough 2.52

Nutritional risk 0.04

Adherence to compression 3.02

Factor

Antimicrobial treatment 0.36

Time Period 2.45*

Treatment * Time 1.69

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data * Significant <0.05 ** Significant <0.01 *** Significant <0.001

4.1.3 Analysis of Covariance (ANCOVA) Analysis ANCOVA was used to examine differences between the treatment groups for each fortnight and for an overall study healing rate. The effects of covariates including nutrition factors, adherence to compression bandaging, and wound characteristics (wound exudate and slough) were controlled in the analysis. ANCOVA found no significant difference between the treatment groups in overall wound healing rate, or for any of the fortnight healing rates with the exception of the first fortnight [F(1,250)=9.16, p<0.01] (Table 4.3). At the first fortnight, the healing rate for the Silver group (Ave=2.12, SD=2.94) was significantly higher than the healing rate for the Iodine group (Ave=-0.22, SD=8.18). Table 4.3 details the significant relationships between the covariates and healing rate for each fortnight, with three significant differences detected for the overall healing rate; adherence to compression bandaging [F(1,260)=7.25, p<0.01], the amount of wound exudate [F(1,260)=11.34, p<0.01] and the amount of slough tissue in the wound [F(1,260)=9.52, p<0.01]. Therefore, these analyses provided a statistical basis to support the proposition that healing rate can be optimised if the client adheres to their recommended compression bandaging regime and slower healing times can be expected if the wound has lower levels of exudate and slough at baseline.

These results suggest that though there were no performance differences between the antimicrobials for the entire study observation period, Silver does have an immediate effect in reducing wound size compared to Iodine. These analyses may reflect a difference in the action of the two products in terms of the way they act on exudate and slough in a wound. The data do not, however,



allow for this difference to be isolated from what might be claimed to be an upfront faster healing rate for silver.

Table 4.3

ANCOVA of Healing Rate (n=263)†^

F value FN1 FN2 FN3 FN4 FN5 FN6 Overall

Covariates Adherence to compression 0.32 1.18 0.01 6.44* 0.01 2.77 7.25** Wound exudate 4.07* 0.01 0.05 0.48 1.27 0.08 11.34** Wound slough 1.30 3.74 7.77** 2.94 2.71 0.01 9.52** Nutritional risk 0.32 0.03 1.22 0.11 1.72 13.76** 0.12

Factor Antimicrobial treatment 9.16** 0.01 0.42 1.39 0.80 1.01 2.46


4.1.4 Per Protocol Analysis

A per protocol or ‘efficacy subset analysis’ was performed to complement the intention-to-treat analysis providing insight as to the actual product effect, in contrast to the population effect, of these antimicrobials (i.e. as used versus as randomised). The results were unchanged, a finding which was not surprising given that there were few adverse events experienced during the course of the trial necessitating a change in antimicrobial treatment. Given the high comparability of results, the findings are not presented in detail in this report.

4.1.5 Product Specific Analysis The healing rate associated with each product was also considered in addition to the analysis for Silver and Iodine treatments combined. For the Silver treatment group the products included Acticoat®, Acticoat® Absorbant, and Acticoat® 7. For the Iodine treatment group the products included Iodosorb® Ointment and Iodosorb® Powder.

ANCOVA was used to examine differences between the products for each fortnight and for an overall study healing rate. The effects of covariates including nutrition factors, adherence to compression bandaging, and wound characteristics (wound exudate and slough) were controlled in the analysis. ANCOVA found no significant difference between the products in overall wound healing rate [F(4,257)=1.31, p>0.05], or for any of the fortnight healing rates with the exception of the first fortnight [F(4,247)=3.31, p<0.01].

Pairwise Comparisons revealed that at the first fortnight, the healing rate for Acticoat® (Ave=2.09, SD=2.96) was significantly higher than the healing rate for Iodosorb® Ointment (Ave=-0.75, SD=9.53) [F(1,247)=3.20, p<0.05]. These analyses suggest that any of the products assessed in this RCT offer equivalent clinical effectiveness over a 12 week monitoring period; a finding of relevance to the discussion of cost treatment explored in Chapter 8 of this report.



4.2 Healing Rate Analysis: Segmentation 4.2.1 Healed and Not Healed Group Comparison

The performance of the treatment groups within the segments of clients who healed during the 12 week observation period and those that did not heal was contrasted. Table 4.4 presents the means and standard deviations for both groups. LMM analysis found no significant difference in healing rate for the treatment groups in either the healed [F(1,131)=0.55, p>0.05] or not healed [F(1,94)=0.82, p>0.05] segments. For healed wounds, ANCOVA did not find any significant difference between the treatment groups for any fortnight or for the overall study (Table 4.5). However, a significant difference between the treatment groups was identified using ANCOVA for not healed clients for the first fortnight [F(1,87)=16.88, p<0.01] and an overall study healing rate [F(1,91)=4.19, p<0.05]. The healing rate was significantly higher for the Silver group at the first fortnight (Ave=1.06, SD=2.07) and overall (Ave=0.57, SD=0.86) compared to the healing rate for the Iodine group at the first fortnight (Ave=-2.53, SD=5.61) and overall (Ave=-0.25, SD=2.57).

Table 4.4

Descriptive Statistics for Healing Rate by Treatment Group by Healed/ Not Healed Segmentation (n=266)†

n= Ave SD

Silver^ Iodine^ Silver Iodine Silver Iodine Overall Healed 85 84 2.96 2.82 1.76 1.52 Not healed 48 49 0.57 -0.25 0.86 2.57 Fortnight 1 Healed 82 81 2.73 1.08 3.19 9.10 Not healed 47 46 1.06 -2.53 2.07 5.61 Fortnight 2 Healed 70 74 4.42 5.18 5.06 2.18 Not healed 45 43 0.76 -0.33 4.19 4.24 Fortnight 3 Healed 51 48 4.20 3.70 8.64 5.57 Not healed 41 43 -1.56 -0.02 13.40 3.90 Fortnight 4 Healed 33 32 6.49 2.63 2.95 14.64 Not healed 39 38 -0.28 0.22 3.85 3.40 Fortnight 5 Healed 18 15 0.86 5.39 13.17 3.87 Not healed 40 36 0.13 0.08 11.11 4.46 Fortnight 6 Healed 7 7 7.01 8.50 0.33 2.78 Not healed 41 34 0.27 -0.19 4.49 5.50




Table 4.5

ANCOVA of Healing Rate by Healed/ Not Healed Segmentation (n=266)†^ F value FN1 FN2 FN3 FN4 FN5 FN6 Overall

Healed Segment Covariates

Adherence to compression 0.04 0.15 0.09 1.46 0.90 0.23 0.01 Wound exudate 2.41 1.17 0.01 1.04 2.32 0.01 2.32 Wound slough 0.15 4.46* 3.56 3.96 0.05 0.03 6.77* Nutritional risk 0.35 0.57 6.23* 0.75 0.23 1.39 0.01

Factor Antimicrobial treatment 2.51 0.82 0.22 2.09 1.69 1.77 0.62

Not Healed Segment Covariates

Adherence to compression 0.20 1.00 0.68 3.73 1.25 0.33 0.83 Wound exudate 0.21 3.95* 0.13 8.56** 0.72 1.26 0.02 Wound slough 1.38 0.81 4.00* 0.75 5.53 0.08 0.68 Nutritional risk 0.30 0.19 0.43 0.63 1.26 12.75** 0.66

Factor Antimicrobial treatment 16.88** 3.35 0.42 0.04 0.01 0.60 4.19*


Figure 4.2 presents the fortnightly healing rates for the not healed segment. While the Silver group was characterised by positive healing on commencement of the antimicrobial and at the end of the study period, there was a drop in healing rate in the middle of the study perhaps accounting for why this segment did not go on to be healed within the study period. The Iodine group for this segment was slow to respond to treatment as defined by initial wound size reduction, though the healing rate steadily improved during the observation period.



Figure 4.2 Fortnightly Healing Rate by Treatment Group: Not Healed Segment

1.06

-1.56

-0.28

-2.53

-0.33

0.130.76

0.27-0.02

0.22

0.08 -0.19

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

Dai

ly h

ealin

g ra

te

Silver Iodine


Wounds that were unlikely to heal within 12 weeks (as dictated by the period of observation for the study) were larger at baseline, in excess of 10cm2 compared to half that for wounds that did heal. They were of long duration with a median age of 24 weeks or almost six months compared to just eight weeks or two months for wounds that healed, and were more likely to present as moderately to heavily exudating. The results presented in Table 4.5 of a positive and significant effect of Silver on exudate, slough and healing rate suggest that Silver would be the preferred antimicrobial for wounds presenting with signs of bacterial burden and matching this description.

4.2.2 Wound Duration Group Comparison The performance of the treatment groups within the segment of clients with a short duration wound at recruitment (less than or equal to 12 weeks) and the segment of clients with chronic wounds at recruitment (greater than 12 weeks) was contrasted. Table 4.6 presents the means and standard deviations for both wound duration segments. LMM found no significant difference in healing rates for the treatment groups for either young [F(1,82)=2.19, p>0.05] or old [F(1,110)=0.04, p>0.05] wounds. ANCOVA analysis was conducted to compare both treatment groups among clients with young and old wounds for each fortnight and for an overall study healing rate (Table 4.7). The results of analysis showed a significant difference in healing rate in favour of Silver for the first fortnight for young wounds [F(1,135)=5.10, p<0.01] and old wounds [F(1,109)=4.75, p<0.01]. In both instances, the healing rate for the Silver group [young wounds (Ave=2.27, SD=2.84) / old wounds (Ave=1.93, SD=3.08)] was significantly higher in the first fortnight than the healing rate for the Iodine group [young wounds (Ave=-0.64, SD=10.22) / old wounds (Ave=0.25, SD=4.93)].



Table 4.6 Descriptive Statistics for Healing Rate by Treatment Group by Wound Duration Segmentation (n=266)†

n= Ave SD

Silver^ Iodine^ Silver Iodine Silver Iodine

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data Table 4.7 ANCOVA of Healing Rate by Wound Duration Segmentation (n=266)†^

Overall Young Wounds 76 70 2.39 2.22 1.75 2.23 Old Wounds 57 63 1.70 1.10 1.99 2.59 Fortnight 1 Young Wounds 73 68 2.27 -0.64 2.84 10.22 Old Wounds 56 59 1.93 0.25 3.08 4.93 Fortnight 2 Young Wounds 66 60 3.91 4.49 5.13 3.22 Old Wounds 49 57 1.76 1.75 4.70 4.90 Fortnight 3 Young Wounds 49 42 3.30 1.92 8.52 5.39 Old Wounds 43 49 -0.27 1.95 13.74 5.04

3.99 1.14 5.57 14.17 Fortnight 4 Young Wounds 35 34 Old Wounds 37 36 1.71 1.49 3.78 3.94 Fortnight 5 Young Wounds 26 22 1.77 3.17 10.23 4.62 Old Wounds 32 29 -0.79 0.48 12.78 4.87 Fortnight 6 Young Wounds 18 15 2.16 2.91 5.14 7.93 Old Wounds 30 26 0.72 0.35 4.57 4.65

FN1 FN2 FN3 FN4 FN5 FN6 Overall F value

Young Wound Segment Covariates

Adherence to compression 0.07 0.05 0.01 5.01 1.20 2.46 0.20 Wound exudate 2.84 0.24 1.00 0.58 2.08 1.50 1.25

0.71 7.63** 2.78 3.57 1.25 0.38 13.54** Wound slough 0.06 0.71 7.15** 0.08 3.50 15.22** 0.37 Nutritional risk

Factor 5.10* 0.34 0.18 0.94 0.16 0.18 0.30 Antimicrobial treatment

Old Wound Segment Covariates

Adherence to compression 0.39 2.37 0.21 0.89 0.41 0.33 6.86** Wound exudate 1.37 0.02 0.21 0.38 0.55 0.06 9.79** Wound slough 0.40 0.04 4.16* 0.01 6.75* 0.17 1.37 Nutritional risk 0.28 1.17 0.03 0.01 0.21 5.73* 1.91

Factor Antimicrobial treatment 4.75* 0.12 1.65 0.16 0.41 0.17 2.48

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data Significant <0.05 ** Significant <0.01 *** Significant <0.001



Figure 4.3 illustrates these results by presenting the fortnightly healing rates for both segments across both treatment groups. With the exception of the first fortnight, the healing rate lines tend to cluster together in terms of wound duration; that is, the healing rate for young wounds are high regardless of which antimicrobial was used and therefore are positioned at the top of the graph, while the healing rates for older wounds are low and therefore dominate the lower part of the graph. However, in the first fortnight the healing rates are clustered by antimicrobial with the Silver groups for young and old wounds alike presenting with higher healing rates compared to the lower healing rates for the Iodine groups, suggesting an immediate effect for the Silver antimicrobial regardless of prior wound duration.

Figure 4.3 Fortnightly Healing Rate by Treatment Group: Wound Duration Segmentation

2.27

3.3

3.99

1.77

-0.64

4.49

1.93

-0.27-0.79

0.720.25

3.91

2.16

3.17

1.14

1.92

2.91

1.711.76

0.50.48

1.49

1.951.75

-2

-1

0

1

2

3

4

5

Dai

ly h

ealin

g ra

te

Silver - young wound Iodine - young woundSilver - old wound Iodine - old wound


4.2.3 Wound Size Group Comparison The performance of the treatment groups for clients with a small wound at recruitment (less than or equal to 3.6cm2) and clients with a large wound at recruitment (greater than to 3.6cm2) was contrasted. Table 4.8 presents the means and standard deviations for both wound size segments. LMM found no significant difference in healing rates for the treatment groups for either the small [F(1,142)=1.68, p>0.05] or big [F(1,82)=0.32, p>0.05] wounds.

As presented in Table 4.9, ANCOVA identified a significant difference in healing rate for the first fortnight between the antimicrobial treatment groups for both small wounds [F(1,122)=6.02, p<0.01] and big wounds [F(1,122)=4.70, p<0.05]. At the first fortnight, the healing rate for the Silver group for small



wounds (Ave=2.20, SD=3.31) and big wounds (Ave=2.03, SD=2.50) was significantly higher compared to the Iodine group for small wounds (Ave=-1.08, SD=11.10) and big wounds (Ave=0.56, SD=3.83). There were no other significant differences between the treatment groups for the remaining fortnights or overall for either segment.

Table 4.8 Descriptive Statistics for Healing Rate by Treatment Group by Wound Size Segmentation (n=266)†

n= Ave SD


Overall Small Wounds 69 63 2.56 1.82 2.00 3.08

Big Wounds 63 69 1.63 1.57 1.63 1.75

Fortnight 1 Small Wounds 67 61 2.20 -1.08 3.31 11.10

Big Wounds 62 66 2.03 0.56 2.50 3.83

Fortnight 2 Small Wounds 59 54 3.56 3.41 5.12 5.57

Big Wounds 55 62 2.43 2.95 3.80 4.47


Big Wounds 50 52 2.65 2.04 4.12 4.60

Fortnight 4 Small Wounds 27 27 4.04 -1.25 4.90 15.60

Big Wounds 44 42 2.16 2.84 4.73 3.62


Big Wounds 36 32 0.53 2.35 8.83 4.06


Big Wounds 30 30 0.72 1.65 5.49 6.65




Table 4.9

ANCOVA of Healing Rate by Wound Size Segmentation (n=266)†^ F value FN1 FN2 FN3 FN4 FN5 FN6 Overall

Small Wound Segment Covariates

Adherence to compression 0.17 2.87 0.39 3.30 0.91 1.98 6.73** Wound exudate 1.61 0.61 0.17 1.18 0.21 3.26 1.68 Wound slough 0.64 1.52 5.60* 4.06* 1.09 0.25 2.88 Nutritional risk 0.42 0.38 1.04 0.81 1.26 3.80 0.19

Factor Antimicrobial treatment 6.02** 0.17 1.08 2.65 0.02 0.89 3.33

Big Wound Segment Covariates

Adherence to compression 0.83 0.41 2.74 2.76 1.87 7.97 0.32 Wound exudate 12.12** 0.48 0.04 0.46 1.18 3.32 14.71*** Wound slough 1.92 1.76 1.89 1.69 1.11 0.97 9.73** Nutritional risk 0.02 1.14 0.27 3.60 0.51 10.83 0.02

Factor Antimicrobial treatment 4.70* 0.49 0.36 0.42 1.11 0.65 0.07


4.3 Tissue Type Comparison by Antimicrobial In addition to collecting data on the total wound size, data were collected at baseline and fortnightly intervals assessing the presence and quantity of different tissue types in the wound. Clinicians defined wound areas using the following classifications; granulation, slough, necrotic, hypergranulation, infected, epithelialisation, and undermining. Given the limited use of classifications other than granulation, slough and epithelialisation, the analysis has been restricted to those tissue types.

The change in the percentage of the wound represented by each tissue type was calculated between fortnightly data points and for the overall study period and the treatment groups were compared using LMM and ANCOVA controlling for nutritional factors and adherence to compression bandaging as covariates. The data reflects the percentage change as a daily rate to account for slight variations in the visit for the fortnightly data collection. A positive result is indicative of a decrease between time points in the percentage of the wound represented by that tissue type. Alternatively, a negative finding suggests that the percentage of the wound represented by the tissue type had increased. While it would be hoped that slough tissue would reduce during healing, thus a positive result would be observed, it would be expected that epithelial tissue would increase as might granulation tissue initially as it replaces the slough areas, though it would perhaps ultimately disappear when the wound achieves 100% epithelialisation.

It is noted that in an inter-rater reliability assessment of the AMWIS software used to measure wounds in the present study (Flowers et al., 2008), only



granulation and slough were evaluated due to insufficient sample for other tissue types and that qualitative feedback identified confusion and variation in the definition and assessment of epithelial tissue. As such the results pertaining to epithelial tissue are difficult to interpret as to whether they consistently reflect an improvement in wound area and should therefore be treated with caution. High inter-rater reliability was, however, observed for granulation and slough tissue, as indeed it was for assessment of the total wound area.

4.3.1 Linear Mixed Models (LMM) Analysis LMM identified no significant difference between the treatment groups and the percentage change in any of the tissue types; granulation [F(1,551)=0.67, p>0.05], slough [F(1,590)=0.25, p>0.05], and epithelial tissue [F(1,661)=0.27, p>0.05]. There was a significant effect for ‘time’ and percentage changes in granulation [F(5,704)=2.74, p<0.05], slough [F(5,743)=4.23, p<0.05], and epithelial tissue [F(5,668)=2.43, p<0.05]; the effect of ‘time’ reflecting a reduction in tissue types across the fortnightly data collections for slough and granulation, though changes for epithelial tissue fluctuated by time point. There was a ‘time and treatment’ interaction for epithelial tissue between the two treatment groups [F(5,667)=2.81, p<0.05] suggesting that while there was no overall differences between the treatment groups, there was a difference at time points according to treatment, the difference most marked at the third fortnight where Silver had a reduction in epithelial tissue whereas the area increased for the Iodine group. As noted, suggested variation in the tracing of epithelial tissue and the lack of an inter-rater reliability measure highlights the need to treat this finding with caution.

4.3.1 Analysis of Covariance (ANCOVA) Analysis Descriptive data and ANCOVA findings are presented respectively for granulation (Tables 4.10 and 4.11), slough (Tables 4.12 and 4.13), and epithelial tissue (Tables 4.14 and 4.15). ANCOVA identified no significant differences between the change in tissue types for granulation [F(1,241)=0.10, p>0.05] or slough [F(1,241)=0.60, p>0.05]. There was a significant difference at fortnight three for epithelial tissue [F(1,181)=5.79, p<0.05] with a decrease of epithelial tissue in the Silver group and an increase for the Iodine group.



Table 4.10

Descriptive Statistics for Changes in Granulation by Treatment Group (n=243)† n= Ave SD


Overall 125 118 0.82 0.77 1.61 1.16

Fortnight 1 129 126 -0.25 -0.43 3.27 3.23

Fortnight 2 114 116 0.84 0.92 4.27 3.59

Fortnight 3 92 91 0.31 0.69 3.58 4.05

Fortnight 4 73 70 1.23 0.92 4.40 3.68

Fortnight 5 59 51 0.52 0.97 3.18 2.90

Fortnight 6 47 41 0.62 1.20 3.22 2.81


Table 4.11

ANCOVA of changes in Granulation by Treatment Group (n=243)†^

F value Overall FN1 FN2 FN3 FN4 FN5 FN6

Covariates

Adherence to compression 0.96 2.97 1.21 0.02 4.46* 0.16 0.72

Nutritional risk 0.24 0.02 0.81 3.19 0.01 1.75 0.39

Factor

Antimicrobial treatment 0.09 0.17 0.01 0.68 0.36 0.50 0.63


Table 4.12

Descriptive Statistics for Changes in Slough by Treatment Group (n=243)† n= Ave SD


Overall 126 117 0.80 0.90 1.00 1.22

Fortnight 1 130 125 1.01 0.97 2.64 2.86

Fortnight 2 114 116 0.72 0.75 2.77 2.97

Fortnight 3 92 89 0.51 0.58 2.49 2.30

Fortnight 4 73 68 0.54 0.67 2.31 2.68

Fortnight 5 60 50 0.26 -0.39 2.52 2.29

Fortnight 6 48 40 0.38 0.39 2.35 2.34




Table 4.13

ANCOVA of changes in Slough by Treatment Group (n=243)†^


Covariates

Adherence to compression 2.43 2.99 0.75 6.10** 1.83 0.03 1.85


Factor



Table 4.14

Descriptive Statistics for Changes in Epithelialisation by Treatment Group (n=244)†

n= Ave SD


Overall 126 118 0.10 0.12 0.47 0.73

Fortnight 1 130 126 -0.38 -0.22 1.78 1.78

Fortnight 2 114 116 -0.27 0.23 2.86 2.52

Fortnight 3 92 91 0.97 -0.25 4.43 1.76

Fortnight 4 72 70 0.10 0.09 2.05 2.52

Fortnight 5 58 51 0.43 0.79 1.94 3.16

Fortnight 6 48 41 0.30 0.22 1.77 2.82


Table 4.15

ANCOVA of changes in Epithelialisation by Treatment Group (n=244)†^


Covariates

Adherence to compression 0.17 0.01 0.73 0.07 1.32 0.33 0.12


Factor

Antimicrobial treatment 0.08 0.55 0.14 5.78* 0.01 0.57 0.03




The sum of these findings suggests limited differential impact of the treatment groups on tissue type; an area of investigation pursued for the reason that it could offer insight into and aid interpretation of the overall findings by suggesting any differential mechanisms by which the antimicrobial treatments act.

4.4 Chapter Summary The results indicate that both antimicrobials were associated with encouraging healing outcomes for these lower leg ulcers. When considering the fortnightly data simultaneously or assessing a calculated overall healing rate, there was no significant difference in the clinical effectiveness of either antimicrobial with respect to wound healing rate. There does, however, appear to be an early benefit to using a Silver antimicrobial. A significantly quicker healing in the first fortnight was observed for Silver in contrast to an average negative healing rate among wounds being treated with Iodine. However, this immediate benefit did not translate to any overall benefit for Silver in a 12 week observation period. A similar finding was observed when undertaking a per protocol (or efficacy subset analysis). When assessing changes in tissue types across time by antimicrobial treatment, limited evidence of differences by treatment group was observed.

The benefit in the first fortnight for the Silver antimicrobial translated into all segmentation groups (not healed, wound duration, and wound size) with the exception of the segment that healed during the observation period for which no significant difference between the antimicrobials was detected. The only additional significant difference between the antimicrobials was detected for wounds that did not heal during the observation period. For this segment there was significantly quicker healing rate identified for Silver overall compared to Iodine. Being able to identify those characteristics which can delineate wounds that would be expected to heal during a 12 week period and those that would not will assist clinicians to determine the situations in which the Silver antimicrobial would be the preferred product of choice.

As detailed in Chapter 3.0, wounds that did not heal were larger (in excess of 10cm2 on average) and older (a median of 24 weeks or almost six months) at recruitment to the study. They were also likely to have higher levels of wound exudate, were less likely to have cellulitis noted at baseline, and these clients had more ‘obvious overweight affecting life quality’ compared to those with wounds that did heal during the study period. It is suggested that for clients with a wound meeting these characteristics, Silver would be preferred to Cadexomer Iodine as it would provide an immediate reduction in wound size and reduce wound size more quickly overall.

It is noted that though Silver may operate in a manner which reduces wound size immediately, the lack of a difference between these antimicrobial’s clinical effectiveness for the sample overall, might suggest that Iodine operates in a different though ultimately as effective manner. Indeed, Iodine’s noted inflammatory mechanism (Moore et al., 1997) which contrasts to Silver’s anti-inflammatory action (Sibbald et al., 2007) may explain the absence of an immediate wound size reduction observed in this trial for wounds in the Iodine treatment group. This trial lacks data to comment further on the mechanism of action of both antimicrobials with particular reference to inflammation. It is



recommended that future research which undertakes to compare these dressings incorporate a measure which can monitor the presence of inflammation at baseline and throughout the trial to assist the interpretation of immediate and longer term healing rates as well as explore the potential that a sub-type of client would be better served by one antimicrobial given the presence of existing wound inflammation.



5.0 Results: Time to Wound Healing

Further to the wound healing rate analysis, the antimicrobial treatment groups were compared for the number of wounds that healed during the observation period, the time taken to achieve healing, and the contribution of antimicrobial treatment and covariates in predicting healing.

Chi square test was used to compare the treatment groups for the number of wounds healed, Survival Analysis was used for the time to healing investigation, and Cox Regression was used to model the predictors of healing. Preliminary checks were conducted to ensure that there was no violation of the assumptions associated with each of the statistical tests. Variables were included in each analysis if they exhibited a strong and positive association with the dependent variable.

5.1 Number of Wounds Healed Data from 266 clients (intention-to-treat) were entered into the analysis. As presented in Table 5.1, 64% of clients achieved wound healing defined as 100% epithelialisation within 12 weeks of the RCT observation period. Within the treatment groups, a comparable number of clients healed; 64% from the Silver group and 63% from the Iodine group. There was no significant difference found between antimicrobial treatment group and the number of wounds that healed overall for the RCT period [χ²(1)=0.02, p>0.05] nor at any of the individual fortnightly assessments.

Table 5.1

Descriptive Statistics for the Number of Clients who Healed (n=266)†

n= %

Silver Iodine TOTAL Silver Iodine TOTAL χ² Overall 85 84 169 64 63 64 0.02 Fortnight 1 10 6 16 8 5 6 1.03 Fortnight 2 20 27 47 17 21 19 0.96 Fortnight 3 17 16 33 17 17 17 0.01 Fortnight 4 16 16 32 20 21 20 0.01 Fortnight 5 13 9 22 20 16 18 0.37 Fortnight 6 9 8 17 17 17 17 0.01

† Sample includes all clients included in the intention-to-treat analysis * Significant <0.05 ** Significant <0.01 *** Significant <0.001

A comparison of the number of wounds that healed within each treatment group was explored for the wound duration and wound size segmentations. As found in the overall findings, there was a high degree of similarity between the treatment groups in the number of wounds that healed overall for young [χ²(1)=0.07, p>0.05] and old wounds [χ²(1)=0.17, p>0.05] as well as for small [χ²(1)=0.10, p>0.05] and big wounds [χ²(1)=0.02, p>0.05].



5.2 Time to Healing Using survival analysis, wound healing trajectories (plots of the percentage of wounds that healed according to the total days to wound healing) were constructed to determine if antimicrobial treatment influenced the amount of time taken to wound healing. The wound healing trajectories were highly comparable for both treatment groups (see Figure 5.1), suggesting that both antimicrobials produce a similar healing time curve for leg ulceration. There was no statistically significant difference between the treatment groups for the total number of days to heal (Log-rank P=0.7; Wilcoxon P=0.8).

Figure 5.1. Kaplan-Meier healing curves presenting the total days to heal by antimicrobial treatment group

Cox Regression was used to build a predictive model for time to healing. The predictive effects of the antimicrobial treatments and covariates including wound duration at recruitment, type of compression bandaging used, adherence to compression bandaging, presence of wound exudate, presence of slough, and initial wound size on the time to healing were assessed. These covariates were chosen on the basis of a preliminary correlation analysis which identified these as highly correlating with dependent variables included in the main analysis. The model was built by including the covariates as Step 1, with the



antimicrobial treatment group included in Step 2 to assess their unique contribution given the presence of the covariates.

As shown in Table 5.2, treatment group was not a statistically significant predictor of time to heal. However, among the covariates, adherence to compression bandaging, the absence of slough tissue or exudate at recruitment, smaller wounds, and wounds of shorter duration at recruitment, were all significant predictors of less time to healing. In other words, faster healing was achieved for clients who presented with smaller, younger, and healthier wounds and adhered to compression bandaging therapy.

Table 5.2

Cox Regression Analysis Predicting Days to Heal (n=266)†


B value Step 1 Step 2

Wound exudate -0.37* -0.38*

Wound slough -0.44** -0.43**

Adherence to compression bandaging -0.45** -0.45**

Initial wound size -0.54** -0.54**

Wound duration -0.79** -0.83**

Light/nil compression bandaging

Moderate compression bandaging -0.22 -0.21

High compression bandaging -0.26 -0.28

Antimicrobial Treatment Group -0.14

^Sample size for question(s) vary due to missing data * Significant <0.05 ** Significant <0.01 *** Significant <0.001

5.4 Chapter Summary Almost two thirds (64%) of wounds observed in this RCT healed in the 12 week observation period. Prior studies have placed the number of wounds to heal at 70% for 12 weeks (Colgan et al., 1996; Moffatt et al., 2003), 62% for 16 weeks (Partsch et al., 2001), and between 74% (McCollum et al., 1997; Polignano, Bonadeo, Gasbarro, & Allegra, 2004) and 88% (Moffatt et al., 2003) achieved healing by 24 weeks. The literature suggests that where leg ulcers are ‘simple’ and not compromised by bacterial burden, we should expect approximately 70% of wounds to heal at 12 weeks. Therefore, to achieve 64% of wounds healed after 12 weeks in the presence of bacterial burden and significant healing compromise is a good result.

The analysis found that the antimicrobial to which the client was randomised played no role in influencing the time to wound healing and had no bearing on the achievement of healing during the 12 week observation period.

The finding that adherence to compression therapy was a predictor of healing adds to a credible body of evidence regarding the clinical effectiveness of graduated compression therapy on the healing of those leg ulcers for which compression therapy is indicated (Moffatt, 2002; O'Meara et al., 2009). It adds



weight to the need to optimise the use of compression bandaging therapy when clinically indicated, addressing the barriers of client resistance, nurse resistance, and access and equity with regard to wound care products which may hamper the use of optimal compression bandaging therapy for the duration of an individual’s wound care (Annells et al., 2008; Flowers et al., 2007).

In addition, the findings revealed that the presence of wound slough and exudate was a negative factor in the process of wound healing, the implication for practice thus being the need to manage these factors effectively in the treatment of chronic venous leg ulcers. Wound size and wound age at recruitment were factors of influence in the wound healing process, thus it should be anticipated that larger wounds and wounds that have a longer duration would take longer to heal compared to their smaller and younger counterparts.

The results of this regression analysis has substantiated existing evidence in the area of chronic leg ulcer management by reinforcing long established messages regarding adherence to compression bandaging and confirming that undesirable wound characteristics such as the presence of slough, exudate, a larger wound size, and longer wound duration are likely to be associated with slower wound healing. This trial has also suggested that the influence of the two antimicrobial treatments on time to healing was negligible.



6.0 Results: Presence of Pain and Exudate

Ratings of pain intensity and level of exudate in the wound were gathered at baseline and for each fortnightly assessment throughout the trial. The ability for the antimicrobial treatments to resolve pain and exudate associated with a wound was compared in the analysis. Pain was analysed using Linear Mixed Models (LMM) to assess the overall trend of the data and Analysis of Covariance (ANCOVA) to consider the level of pain for each fortnight. Exudate was assessed using chi square test. Segmentation of these results was pursued for healed/ not healed, wound duration, and wound size groups. An additional segmentation on adherence to compression bandaging was explored in relation to wound exudate ratings.

6.1 Experience of Pain Analysis 6.1.1 Descriptive Statistics for the Experience of Pain

The means and standard deviations for the experience of pain (as measured on a scale of 0-10 with 10 being the most pain imaginable) are presented in Table 6.1. The mean pain ratings were comparable for the treatment groups throughout the study period, both declining sharply between the baseline and fortnight one assessment and steadily reducing for the remaining study period (Figure 6.1).

Figure 6.1 Fortnightly Pain Ratings by Treatment Group

3.27

1.431.29 1.37

2.4

1.71

1.19

1.731.9

1.65

3.18

1.851.6 1.7

0

0.5

1

1.5

2

2.5

3

3.5

Baseline Fortnight1

Fortnight2

Fortnight3

Fortnight4

Fortnight5

Fortnight6

Leve

l of P

ain

(0-1

0)

Silver Iodine



Table 6.1

Descriptive Statistics for the Experience of Pain by Treatment Group (n=266)† n= Ave SD

Silver Iodine Silver Iodine Silver Iodine

Baseline 133 131 3.27 3.18 2.69 2.67

Fortnight 1 128 126 1.90 2.40 2.49 2.73

Fortnight 2 104 105 1.73 1.85 2.32 2.38

Fortnight 3 86 86 1.43 1.71 2.16 2.42

Fortnight 4 68 63 1.29 1.60 2.23 2.25

Fortnight 5 52 46 1.65 1.70 2.34 2.27

Fortnight 6 43 36 1.37 1.19 2.07 1.80


6.1.2 Linear Mixed Models (LMM) Analysis of Pain Experience LMM was used to compare the experience of pain for the antimicrobial treatment groups. Adherence to compression bandaging was entered as covariate. As shown in Table 6.2, there was no significant difference in pain ratings between the treatment groups [F(1,297)=0.12, p>0.05] or the interaction of ‘treatment and time’ [F(1,725)=1.21, p>0.05]. Overall, there was a significant reduction in the level of pain intensity observed during the study period [F(6,724)=25.88, p<0.01]. The LMM analysis also detected a significant difference in the experience of pain depending on whether the client adhered to compression bandaging or not [F(1,258)=9.22, p<0.01], with the group of clients that adhered to compression bandaging experiencing significantly less pain (Ave=1.79; SD=2.35) compared to clients not adhering to their recommended compression bandaging regime (Ave=2.39; SD=2.66). There were no significant differences in the rating of pain intensity at baseline between clients who went on to adhere to compression therapy and those that did not adhere. Table 6.2

LMM analysis of the Experience of Pain by Treatment Group (n=266)†^


F value

Covariate

Adherence to compression 9.22**

Factor

Antimicrobial treatment 0.12

Time Period 25.88**

Treatment * Time 1.21




6.1.3 Analysis of Covariance (ANCOVA) Analysis of Pain Experience

ANCOVA was used to examine differences in pain between the treatment groups at baseline and for each fortnight. The effect of adherence to compression bandaging as a covariate was controlled in the analysis. ANCOVA found no significant difference between the treatments at any of the fortnightly assessments for the experience of pain (Table 6.3). A significant difference in terms of a lower level of pain intensity with adherence to compression bandaging was found in the ANCOVA analysis but only at the third fortnight [F(1,169)=6.25, p<0.05].

Table 6.3

ANCOVA of Pain Intensity by Treatment Group (n=264)†^

F value Baseline FN1 FN2 FN3 FN4 FN5 FN6

Covariate

Adherence to compression 2.63 3.50 7.26 6.25* 1.00 3.61 1.30

Factor



These results suggest that both antimicrobials performed equally well, reducing pain across a 12 week monitoring period, while the use of compression bandaging was also associated with a reduction of pain.

6.2 Experience of Pain Analysis: Segmentation 6.2.1 Healed and Not Healed Group Comparison

Table 6.4 presents the means and standard deviations for the healed and not healed segments. LMM identified there was no significant difference in the experience of pain across the antimicrobial treatment groups for healed [F(1,385)=0.14, p>0.05] and not healed clients [F(1,91)=1.64, p>0.05]. As presented in Table 6.5, however, ANCOVA identified a significant difference in pain intensity between the treatment groups at the first fortnight assessment for the not healed segment [F(1,91)=7.04, p<0.01]. At the first fortnight, the pain intensity for the Iodine group (Ave=2.87, SD=2.74) was significantly higher compared to the Silver group (Ave=-1.55, SD=2.11) among the not healed group (Figure 6.2). There were no other significant differences between the treatment groups for the remaining fortnights for either segment.



Table 6.4

Descriptive Statistics for the Experience of Pain by Treatment Group by Healed/ Not Healed Segmentation (n=264)†

n= Ave SD

Silver^ Iodine^ Silver Iodine Silver Iodine Baseline Healed 85 84 3.61 3.07 2.80 2.49 Not healed 48 47 2.67 3.38 2.39 2.97 Fortnight 1 Healed 81 79 2.10 2.11 2.68 2.70 Not healed 47 47 1.55 2.87 2.11 2.74 Fortnight 2 Healed 59 58 1.58 1.19 2.21 1.77 Not healed 45 47 1.93 2.66 2.46 2.77 Fortnight 3 Healed 42 41 0.95 0.88 1.65 1.76 Not healed 44 45 1.89 2.47 2.49 2.70 Fortnight 4 Healed 28 21 0.57 0.81 1.50 2.06 Not healed 40 42 1.80 2.00 2.52 2.25 Fortnight 5 Healed 9 8 0.56 1.25 1.67 1.75 Not healed 43 38 1.88 1.79 2.41 2.37 Fortnight 6 Healed 3 2 0.00 0.00 0.00 0.00 Not healed 40 34 1.48 1.27 2.11 1.83


Table 6.5

ANCOVA of Pain Intensity by Treatment Group by Healed/ Not Healed Segmentation (n=264)†^


F value Baseline FN1 FN2 FN3 FN4 FN5 FN 6

Healed Segment Covariate

Adherence to compression 4.14* 1.85 2.86 1.60 0.32 0.007 - Factor

Antimicrobial treatment 1.88 0.001 0.99 0.001 0.28 0.65 - Not Healed Segment

Covariate Adherence to compression 0.32 1.81 1.81 1.20 0.001 3.12 0.75

Factor Antimicrobial treatment 1.71 7.04** 1.94 1.19 0.14 0.03 0.19




Figure 6.2 Fortnightly Pain Rating by Treatment Group: Healed/ Not Healed Segmentation

3.61

1.58

0.95

0.57

0

3.07

0.81

0

1.55

1.93 1.89 1.8 1.88

1.48

3.38

2.47

1.27

0.56

2.12.11

0.88

1.251.19

2.67

2.01.79

2.87 2.66

0

0.5

1

1.5

2

2.5

3

3.5

4

Baseline Fortnight1

Fortnight2

Fortnight3

Fortnight4

Fortnight5

Fortnight6

Leve

l of P

ain

(0-1

0)

Silver - healed wound Iodine - healed woundSilver - not healed wound Iodine - not healed wound

While the level of pain experienced was reduced for both healed and not healed segments, it is suggested that the difference observed for the not healed segment reflects the varied action of the antimicrobial on the wound. That is, with the experience of pain associated with inflammation (Marieb, 2001), and with the not healed wounds representing larger and more chronic wounds which would potentially be characterised by greater inflammation, the action of Silver to reduce inflammation (Sibbald et al., 2007) would reduce pain more substantially than would Iodine whose action can generate an inflammatory response (Moore et al., 1997). The effect might not be observed as acutely for the healed segment, the wounds in which were younger and would therefore have a lower level of chronic inflammation.

6.2.2 Wound Duration Group Comparison Table 6.6 presents the means and standard deviations for both the wound duration segments. LMM identified no significant difference in the experience of pain between the antimicrobial treatment groups for the young [F(1,175)=0.54, p>0.05] and old wounds [F(1,124)=0.03, p>0.05]. As presented in Table 6.7, ANCOVA identified no significant difference in pain intensity between the treatment groups at any fortnightly assessment for either segment of young and old wounds.

Adherence to compression was a significant covariate for both segments though the effect was observed later in the young wounds at the fortnight three and four data collections compared with an earlier effect at fortnights one and two in older wounds.



Table 6.6

Descriptive Statistics for the Experience of Pain by Treatment Group by Wound Duration Segmentation (n=264)†

n= Ave SD


Baseline Young Wounds 76 70 3.62 2.59 2.55 2.63

Old Wounds 57 61 2.81 2.72 2.81 2.65

Fortnight 1 Young Wounds 72 66 1.94 2.79 2.37 2.86

Old Wounds 56 60 1.84 1.97 2.66 2.53


Old Wounds 48 57 1.60 1.75 2.09 2.58


Old Wounds 40 46 1.55 1.41 2.26 2.37


Old Wounds 35 36 1.63 1.78 2.33 2.69


Old Wounds 31 29 2.20 1.66 2.36 2.64


Old Wounds 28 24 1.71 1.38 2.23 1.97


Table 6.7

ANCOVA of Pain Intensity by Treatment Group by Wound Duration Segmentation (n=264)†^


Young Wound Segment

Covariate

Adherence to compression 0.78 0.34 2.77 7.48** 5.72* 1.94 0.15

Factor


Old Wound Segment

Covariate

Adherence to compression 2.93 4.71* 4.98* 0.63 0.23 1.67 1.33

Factor

Antimicrobial treatment 0.01 0.24 0.31 0.01 0.04 0.50. 0.29




6.2.3 Wound Size Group Comparison Table 6.8 presents the means and standard deviations for both the wound size segments. LMM identified no significant difference between the treatment groups in the experience of pain for small [F(1,166)=0.15, p>0.05] or big wounds [F(1,131)=1.28, p>0.05]. As presented in Table 6.9, ANCOVA identified a significant difference in pain intensity for the second fortnight between the antimicrobial treatment groups for big wounds [F(1,108)=4.60, p<0.05] with a greater average degree of pain reported for the Iodine group (Ave=2.24, SD=2.59) compared to the Silver group (Ave=1.33, SD=2.03). There was no significant difference between treatment groups at any other fortnightly assessment for either wound size segment. Of note, adherence to compression was not a significant covariate for small wounds at any time point but was significant at most assessment points for large wounds.

Table 6.8

Descriptive Statistics for the Experience of Pain by Treatment Group by Wound Size Segmentation (n=264)†

n= Ave SD


Baseline Small Wounds 70 63 3.46 2.95 2.75 2.67

Big Wounds 62 67 3.00 3.40 2.58 2.68


Big Wounds 60 65 1.85 2.46 2.58 2.83


Big Wounds 54 57 1.33 2.24 2.03 2.59


Big Wounds 49 49 1.10 1.79 1.86 2.28


Big Wounds 39 40 1.20 1.82 2.07 2.51


Big Wounds 32 30 1.87 1.70 2.59 2.45


Big Wounds 27 24 1.40 1.20 2.08 1.81




Table 6.9

ANCOVA of Pain Intensity by Treatment Group by Wound Size Segmentation (n=264)†^


Small Wound Segment

Covariate

Adherence to compression 0.76 0.27 1.43 2.58 0.22 0.01 0.76

Factor


Big Wound Segment

Covariate

Adherence to compression 5.04* 4.03* 4.97* 2.83 2.29 6.03* 5.04*

Factor

Antimicrobial treatment 0.19 1.73 4.60* 2.88 1.58 0.08 0.19


6.3 Amount of Wound Exudate Analysis Data were collected about the amount of wound exudate at baseline and for each fortnightly assessment. Gathered as a rating of nil, low, moderate, or high exudate as determined by the attending clinician, the data were ultimately re-classified as nil-low exudate or moderate-heavy exudate due to low cell sample sizes. As shown in Figure 6.3 more than half the wounds at baseline were classified as having moderate-heavy exudate (60% Silver; 55% Iodine). In both treatment groups the amount of moderate-heavy exudate reduced during the study period to 23% for the Silver group and 38% for the Iodine group. Of note, while the Silver group maintained a relatively steady trend of exudate reduction throughout, the level of wound exudate reduction plateaued for the Iodine group after fortnight three until the end of the study.

Chi-square test was used to compare the rating of wound exudate amount for the antimicrobial treatment groups at baseline and for each fortnight. As shown in Table 6.10, there were no significant differences between the treatments groups at any assessment point with the exception of fortnight four (χ2(1)=4.96, p<0.05); a significantly higher level of exudate was observed for the Iodine group (39% moderate-high exudate) compared with the Silver group (27% moderate-high exudate).



Figure 6.3 Fortnightly Exudate Rating by Treatment Group

60

3630

21 23

4238

35

27

3937

5544

38

0

10

20

30

40

50

60

70

Baseline Fortnight1

Fortnight2

Fortnight3

Fortnight4

Fortnight5

Fortnight6

% m

oder

ator

-hea

vy e

xuda

te

Silver Iodine

Table 6.10

Descriptive Statistics and Chi Square Results for the Wound Exudate by Treatment Group (n=266)†

Mod-heavy exudate (%)

Silver Iodine

n= χ2

Baseline 60 55 266 0.75

Fortnight 1 35 44 255 2.43

Fortnight 2 36 42 209 0.88

Fortnight 3 30 37 172 0.93

Fortnight 4 21 39 134 4.96*

Fortnight 5 27 38 97 1.30

Fortnight 6 23 38 80 2.01

† Sample includes all clients included in the intention-to-treat analysis * Significant <0.05 ** Significant <0.01 *** Significant <0.001

6.3.1 Segmentation for the Level of Exudate Additional segmentation analysis was performed comparing the amount of exudate for each fortnight for the antimicrobial treatment groups for the healed/ not healed segment, wound duration segment, wound size segment and for segments that did and did not adhere to compression bandaging therapy.

There was no significant difference in exudate levels at any assessment for wounds that healed during the study. There was a significant difference in the level of exudate at the first fortnight for clients that did not heal [χ2(1)=4.16, p<0.05]. Figure 6.4 presents the results for the healed / not healed segmentation which illustrates the comparability of exudate levels for the healed segment regardless of antimicrobial received, while highlighting the sharp reduction of



exudate in the Silver treatment group for clients that did not heal from baseline to fortnight one compared to limited change for the same time period for wounds being treated with Iodine.

Figure 6.4 Fortnightly Exudate Rating by Treatment Group: Healed/ Not Healed Segmentation

54

14

0 0

45

50

71

5255

48

3532

25

6154

40

19

11

23

24

10 0

22

67

45

7163

0

10

20

30

40

50

60

70

80

Baseline Fortnight 1 Fortnight 2 Fortnight 3 Fortnight 4 Fortnight 5 Fortnight 6

% M

oder

ate-

Hea

vy E

xuda

teSilver - healed wound Iodine - healed woundSilver - not healed wound Iodine - not healed wound

There was no significant difference between the treatment groups and the amount of exudate at baseline or for any fortnight in either the young or old wound segments.

There was also no significant difference in the level of exudate between the treatment groups at any of the assessments for the small wound segment (Figure 6.5). For clients with a big wound, there was a significant difference between the treatment groups in the amount of exudate at fortnight four [χ2(1)=8.76, p<0.01]. As noted in Section 3.2.5 larger wounds were more likely to be recruited to the trial showing signs of increased and malodorous exudate (52.3%) compared to smaller wounds (38.6%).

For clients who adhered to their recommended level of multi-layer compression bandaging (57%), there were no significant differences in the level of exudate between the treatment groups at any of the assessments. For other clients who did not adhere to compression bandaging, there was a significant difference in the amount of exudate at fortnight four [χ2(1)=6.64, p<0.05]. Figure 6.6 presents the results of this segmentation which illustrates those wounds receiving Iodine and not adhering to compression bandaging had heightened levels of exudate throughout the study period most notably at the fourth fortnight compared to all other groups.

Overall, Silver resulted in significantly reduced levels of moderate to heavy exudate at fortnight four for the RCT sample overall, for large wounds at baseline, and when the client did not adhere to compression bandaging. Silver was also more effective at reducing exudate at fortnight one for clients who did not heal during the study period.



Figure 6.5 Fortnightly Exudate Rating by Treatment Group: Wound Size Segmentation

49

2619 18

37

27

14 15

73

4743

37

22

31 30

53 5450

1316

23

17

2021

72

48

6160

0

10

20

30

40

50

60

70

80


% M

oder

ate-

Hea

vy E

xuda

te

Silver - small wound Iodine - small woundSilver - big wound Iodine - big wound

Figure 6.6 Fortnightly Exudate Rating by Treatment Group: Compression Bandaging Adherence Segmentation

59

30

19 17

54

36 35

63

33 34

23 21

28

47

55

4039

373325

2926

42

37

54 4852

40

0

10

20

30

40

50

60

70


% M

oder

ate-

Hea

vy E

xuda

te

Silver - Adherence Iodine - AdherenceSilver - Did Not Adhere Iodine - Did Not Adhere

6.4 Chapter Summary Client pain and the level of wound exudate declined for both treatment groups during the study period demonstrating both treatments to be effective in



relieving pain and resolving exudate. The reported degree of client pain was highly comparable for both antimicrobials, suggesting either antimicrobial would be clinically effective from this perspective.

It was only for the fortnight four assessment that the Silver antimicrobial was found to reduce exudate more effectively than Iodine. In the absence of a consistent and statistically validated pattern regarding the effectiveness of the Silver antimicrobial in reducing exudate levels compared to Iodine, little can be concluded other than further investigation of these antimicrobial treatments and exudate management is required.

There was no difference between the antimicrobials and the experience of pain for all segments with the exception of clients that did not heal during the study period and for large wounds; both of these segments were identified as having significantly more pain in those wounds being treated by Iodine at the first and second fortnights respectively. It was in these same segments that differences between the antimicrobials were identified for the level of wound exudate. There was more exudate identified in the Iodine group for the first fortnight for wounds that did not heal during the study period and for the fourth fortnight for large wounds. Wound exudate was significantly higher at the fourth fortnight for clients who did not adhere to their recommended compression bandaging regime.

It is difficult to drawn too many conclusions from these results regarding the impact of antimicrobial treatment on wound exudate management as the difference between the treatment groups was significant for only one of the fortnightly assessments and because the pattern of Silver to reduce exudate levels while Iodine plateaued in terms of its performance was not statistically substantiated. The capacity for Silver dressings to manage exudate more effectively than Iodine dressings requires further investigation by future research studies.



7.0 Results: Wound Culture Results

During the RCT wound swabs were obtained to confirm and quantify bacterial burden of the study wounds, given that these wounds were recruited because they demonstrated one or more signs of critical colonisation or infection. The study wound of every RCT participant was swabbed on recruitment to the RCT, providing baseline data and also when ceasing or recommencing an antimicrobial based on the absence or presence of the clinical signs of bacterial burden. Swabs were repeated at six and 12 weeks if the client was still being treated with an antimicrobial.

The data obtained from these swabs have been used in this analysis to:

• Compare wound culture findings to clinical signs of critical colonisation and infection as informed by the published literature and judged by the attending nurse; and

• To compare wound culture findings with healing rates and the performance of the antimicrobial treatments.

The principle statistical analysis appropriate for these investigations included Spearman’s correlation, Chi Square and Mann-Whiney U tests. Preliminary checks were conducted to ensure that there was no violation of the assumptions associated with each of the statistical tests.

While the RCT analyses have excluded 15 of the original 281 recruits because a healing rate could not be calculated, all clients were included in the swab culture analyses where possible, except when examining healing rate. Therefore, most of the swab analyses reported in this chapter are based on all trial participants with the exception of two clients for whom a baseline swab was not undertaken and one client for whom there were missing data regarding which signs of critical colonisation and infection were present on recruitment. Thus, for the majority of swab analyses, a base of 278 participants applies.

In assessing the degree of growth for each swab, the highest level of growth for either bacilli positive, bacilli negative, cocci positive, and cocci negative was used regardless of which organism was isolated from the wound culture. This approach acknowledges the fact that a high degree of bacterial growth could be determined even when a specific organism could not be isolated; and indeed for almost half the wound cultures a single organism was not isolated. When assessing the growth associated with a particular organism, the level of growth indicated for the corresponding bacterial colony for each organism was used. As an additional measure the leucocytes colony was considered given that increased white bloods cells as indicated by a high growth in the leucocytes gram stain may represent an immune response associated with infection.

7.1 The Wound Culture Findings Further to the wound swabs collected at recruitment, there were an additional 301 swabs obtained during the course of the study. Only 17.9% of these additional swabs were collected because the client was ceasing the antimicrobial treatment, and only one because the antimicrobial was recommenced. The vast




proportion of these additional swabs were obtained because the client was still using an antimicrobial after six weeks (41.2%) and 12 weeks (21.6%) of study participation.

Of the 278 swabs considered at baseline, 37.8% had a nil or scant level of bacterial growth as determined by the wound culture (Table 7.1). This compares to 65.5% of swabs with a nil or scant growth of leucocytes, and 33.7% of the proportion of clients with one or more signs of infection (excluding those with only signs of critical colonisation). These results are of note given that to be eligible for this RCT all wounds demonstrated one or more clinical signs of critical colonisation or infection; a result clearly inconsistent with the microscopic examination and suggestive of either questionable validity of the stated clinical signs of bacterial burden or the reliance on a wound swab to give an accurate view of the bacteria present in a wound. The remaining swabs at baseline presented with a relatively even distribution of low, moderate, and heavy growth.

Table 7.1

Degree of Bacterial Burden and Leucocytes Growth at each Swab (n=278) ^


While the first swab represents baseline data, remaining swabs were obtained for a multitude of reasons. These swabs (2nd through to 5th) were re-classified accordingly to the reason they were gathered and the level of bacterial growth. A criterion for ceasing the antimicrobial was the absence of clinical signs of critical colonisation or infection for one week. As shown in Table 7.2, over half of the swabs collected for this reason had no bacterial growth (51.4%) or growth of leucocytes (84.7%). Contrary to what might have been expected, in the absence of any clinical sign of bacterial burden or infection, wound swabs were still detecting a low to heavy bacterial load in half the wounds. On the other hand, of those wounds swabbed after six and 12 weeks as per the study protocol because signs of bacterial burden remained, over a quarter of wounds registered no bacterial burden (26.2% and 31.6% respectively) and three quarters registered no growth of leucocytes (73.1% and 70.5% respectively). The

% n= Nil Low Mod Heavy

Bacterial Burden

1st swab 278 37.8 18.7 18.0 25.5

2nd swab 171 34.9 27.3 16.3 21.5

3rd swab 85 34.9 24.4 20.9 19.8

4th swab 23 47.8 17.4 8.7 26.1

5th swab 5 20.0 60.0 20.0 -

Leucocytes growth

1st swab 278 65.5 22.7 9.4 2.5

2nd swab 189 72.0 18.0 7.9 2.1

3rd swab 97 74.2 18.6 7.2 -

4th swab 26 84.6 7.7 7.7

5th swab 5 80.0 20.0 - -

-



findings again demonstrate incongruency between the results of the two different methods of assessing bacterial burden and infection; by clinical judgment using criteria identified in the literature and wound swabbing.

Table 7.2

Degree of Bacterial Burden and Leucocytes Growth by Swab Indication (n=278) ^


Overwhelmingly, wound swab cultures detected colonisation or infection by Staphylococcus aureus with almost nine in ten wounds from which an organism could be isolated, identified as being burdened with this bacteria (Table 7.3). While Pseudomonas organisms and Streptococcus were also detected their numbers were very low. Only 16 swabs were identified with methicillin resistant staphylococcus aureus (MRSA). Of these, 15 were community derived MRSA as determined by their sensitive to Methicillin only, with one instance of MRSA classified as being of hospital origin (resistant to Methicillin and Gentamycin).

Table 7.3

Type of Bacterial Organism at each Swab (n=278)

* Staphylococcus. aureus ~ Pseudomonas aeruginosa / Pseudonitzschia spp.

% Nil Low Mod Heavy

Bacterial Burden

Ceasing antimicrobial treatment swab (n=37) 51.4 18.9 24.3 5.4

Six week swab (n=107) 26.2 29.9 15.0 29.0

12 week swab (n=57) 31.6 24.6 17.5

Leucocytes Growth

Ceasing antimicrobial treatment swab (n=46) 84.7 6.5 6.5 2.2

Six week swab (n=115) 73.1 8.7 0.9

12 week swab (n=61) 70.5 18.0 11.5 -

26.3

17.4

% n= Staph.

aureus*

Strepto-

coccus

Pseudo-

monas~

Other

1st swab 177 2.8 8.5 1.7 87.0

2nd swab 120 90.8 1.0 2.5

3rd swab 68 91.2 1.5 5.9 1.5

4th swab 21 85.7 - 14.3 -

5th swab

5.8

3 66.7 - 33.3 -



7.2 Comparing Wound Culture Results & Signs of Critical Colonisation / Infection Table 7.4 presents the frequencies of the signs of critical colonisation and infection at baseline and each fortnight of the study as well as the proportion of wounds with no signs and the average number of signs of critical colonisation and infection.

On recruitment to the study the vast majority of clients were judged to have impaired or delayed healing (88.1%) and many had new areas of slough or wound breakdown (69.1%). Changes in granulation tissue (51.8%) and malodorous exudate (45.0%) were common, and pain was also represented in a third of participants (34.2%). A third of clients (33.7%) had one or more signs of infection (considered to include Cellulitis, Suppuration, Lymphangitis, Sepsis and Bacteraemia) and could also have signs of critical colonisation, with the remaining 66.3% of clients assessed as having one or more signs of critical colonisation alone.

Almost three quarters of clients (71.8%) had multiple signs of infection or critical colonisation at baseline, with an average of 3.28 suggesting that the wounds sampled for this trial were substantially compromised by bacterial burden and that an antimicrobial dressing was an appropriate treatment response.

Table 7.4

Presence of Signs of Infection or Critical Colonisation (n=278)^


Baseline FN1 FN2 FN3 FN4 FN5 FN6

(n=278) (n=272) (n=248) (n=194) (n=158) (n=124) (n=98)

% with no signs of bacterial burden

0.0 27.6 41.1 47.4 45.6 43.5 43.9

Number of signs of bacterial burden (Ave)

3.28 1.46 1.14 1.06 1.00 1.07

% of signs of bacterial burden

Impaired / delayed healing 88.1 53.3 38.6 38.5 40.3 44.0

New areas of slough / wound breakdown

69.1 14.0 13.3 11.3 11.3 9.6

Changes in granulation tissue 51.8 33.1 25.3 22.6 18.9 22.4

Increased / malodorous exudate

45.0 12.1 12.9 12.8 13.2 8.0

Increased / new pain 34.2 12.1 8.4 6.2 5.0 4.8

Cellulitis 24.8 6.2

0.88

40.4

7.1

24.2

8.1

1.0

6.0 4.6 1.9 4.8 3.0

Suppuration 9.4 5.9 4.4 2.6 3.1 4.8 6.1

Lymphangitis 0.0 0.7 0.4 0.0 0.0 0.0 0.0

Sepsis 5.8 10.3 6.0 7.7 6.3 8.8 8.1

Bacteraemia 0.4 0.0 0.0 0.0 0.0 0.0 0.0


A comparison of the wound culture results and the presence of bacterial burden as determined by clinical signs of critical colonisation and infection was conducted using the baseline data where all clients were known to have a wound swab attended and all were showing clinical signs of bacterial burden. Table 7.5 cross-tabulates the number of eligibility criteria identified at recruitment and the level of bacterial growth identified by the wound culture. It was expected that a cluster of few signs of critical colonisation and infection would be observed for the ‘nil/ scant’ category with many signs present for swabs revealing a ‘heavy’ growth. However, Table 7.5 suggests this correlation did not exist. In fact, the number of clinical signs of bacterial burden was distributed fairly evenly across all the wound swab categories of bacterial burden and leucocytes growth. A chi square test of the association found no significant association between clinician rating and bacterial burden [χ²(6)=9.41, p>0.05] or the presence of leucocytes [χ²(3)=5.93, p>0.05]. This finding is consistent with a Spearman’s correlation which also found a very low, non-significant relationship for bacterial burden (r=-.081, p>0.05) and the presence of leucocytes (r=-.011, p>0.05). These results suggest that there is doubt as to the value of wound swabs confirming clinical judgment of the presence of bacterial burden or infection in a wound; clinical judgment being the most common basis for commencing antimicrobial treatment.

Table 7.5

Number of Eligibility Criteria Identified and Degree of Bacterial Growth (n=278) ^

Degree of Bacterial Burden Degree of Leucocytes Growth

Nil/

Scant

Low Moderate Heavy Nil/

Scant

Low Moderate Heavy

%

Number of wound characteristic criteria met (n=105) (n=52) (n=50) (n=71) (n=182) (n=63) (n=26) (n=7)

1 criterion 8.6 7.7 4.0 9.9 9.3 4.8 3.8 14.3

2 criteria 15.2 13.5 30.0 25.4 19.8 23.8 15.4 14.3

3 criteria 37.1 19.2 20.0 38.0 30.8 36.5 26.9 -

4 criteria 23.8 25.0 26.0 12.7 20.9 20.6 26.9 28.6

5 criteria 13.3 25.0 12.0 11.3 17.0 9.5 11.5 14.3

6 criteria 1.9 9.6 6.0 2.8 1.6 4.8 15.4 28.6

7 criteria - - 2.0 - 0.5 - - -


In addition to assessing the degree of growth overall with the number of eligibility criteria identified at baseline, the presence of particular organisms or bacterial colony were assessed against the type of critical colonisation or infection sign identified by clinicians, to determine if any one sign tended to suggest a type of bacteria burdening the wound. A significant association was identified between the presence of leucocytes and cellulitis [χ²(1)=5.62, p<0.05] and between Staph. aureus and increased or malodorous exudate [χ²(3)=7.77,



p<0.05] There were no other significant associations identified between the clinical observed signs of bacterial burden and the wound swab results. Thus, with the two exceptions noted above, these results suggest that either the assumption of bacterial burden in a wound that has delayed healing or other signs commonly regarded as associated with infection or critical colonisation cannot be supported; or that the use of a wound swab as a diagnostic tool to confirm bacterial burden is questionable.

7.3 Comparing Wound Culture Results & Wound Healing Rate The presence of bacterial burden - both bacterial colony and bacterial growth - as identified by the wound swab results was compared to the healing rate achieved in each antimicrobial treatment group to determine if the action of one antimicrobial was better in the presence of a particular type or amount of bacterial colonisation or infection.

Table 7.6 presents the results of Mann-Whiney U test comparing the swab results at baseline with healing rate in the first fortnight for both antimicrobial treatment groups. To optimise the sample size in this analysis, categories of nil/low and moderate/heavy were formed. Where moderate to heavy growth was identified in the wound swab, there were no differences in healing rates between the treatment groups. In contrast, where nil or low bacterial growth was identified in the wound swab, Silver had a significantly greater healing rate compared to Iodine for leucocytes [Mann-Whitney U (224)=5159.50, p<0.01], gram positive bacilli [Mann-Whitney U (178)=3237.00, p<0.05], gram positive cocci [Mann-Whitney U (143)=1935.00, p<0.01], and gram negative cocci [Mann-Whiney U (168)=2881.50, p<0.05]. These results suggest that for low level colonisation, with the exception of gram negative bacilli (which was the only bacterial colony for which no significant differences were detected between the treatment groups), Silver is more effective than Iodine, though for greater burden the performance of the antimicrobial treatments was comparable.

In reconciling these results with the findings from Chapter 4, and especially those pertaining to the effectiveness of Silver for wounds unlikely to heal in a 12 week period (that is, large, chronic, and potentially moderately to heavily exudating wounds), the degree of bacterial burden and the presence of leucocytes for wounds that healed and did not heal in the 12 week study period was contrasted. It was anticipated that there would be a lower level of bacterial burden and leucocytes growth for the group of wounds that did not heal in the 12 week observation period which would account for the greater effectiveness of Silver in this group. No significant difference in the presence of bacterial burden [χ²(1)=0.22, p>0.05] or leucocytes [χ²(1)=0.30, p>0.05] was detected between the healed and not healed groups. Therefore, the observed clinical effectiveness of Silver for the not healed group cannot be explained by the subsequent finding that Silver was more effective on lower levels of bacterial burden or leucocytes growth.

It should be further noted that the degree of bacterial burden [χ²(3)=0.93, p>0.05] and the presence of leucocytes [χ²(3)=2.24, p>0.05] was comparable at baseline between the treatment groups. Thus, the significantly quicker healing rate observed for the Silver antimicrobial in the first fortnight cannot be explained by the presence of lower levels of bacterial burden or leucocytes growth in this treatment group.



Table 7.6

Comparing the Healing Rate for First Fortnight for the Treatment Groups and Swab Results (n=259)†^

n= Mean rank Colony Degree of

Growth Silver Iodine Silver Iodine

Mann-

Whitney U

Leucocytes Nil/ Low 116 110 124.02 102.40 5159.50**

Mod/ High 16 17 18.69 15.41 109.00

Gram + Bacilli Nil/ Low 90 90 99.53 81.47 3237.00*

Mod/ High 5 1 3.80 2.00 ††

Gram - Bacilli Nil/ Low 62 64 67.97 59.17 1707.00

Mod/ High 43 32 40.60 34.50 576.00

Gram + Cocci Nil/ Low 72 73 82.62 63.51 1935.00**

Mod/ High 41 35 40.56 36.09 633.00

Gram - Cocci Nil/ Low 87 83 93.88 76.72 2881.50*

Mod/ High - - - - -

† Sample includes all clients included in the intention-to-treat analysis ^Sample size for question(s) vary due to missing data †† Not enough sample to support the result * Significant <0.05 ** Significant <0.01 *** Significant <0.001

Healing rate was also compared between treatment groups according to the wound eligibility criteria at baseline. Only six of the original ten criteria were included in the analysis due to small sample sizes in the remaining four criteria. As presented in Table 7.7, the healing rate was significantly higher for those receiving the Silver treatment when the wound had impaired or delayed healing [Mann-Whitney U (230)=5116.00, p<0.01], changes in granulation tissue [Mann-Whitney U (130)=1743.00, p<0.05], or new areas of slough / wound breakdown reported [Mann-Whitney U (175)=2939.00, p<0.01].

There were no differences in healing rate between the treatment groups when increased or malodourous exudate, increased or new pain, or cellulitis were present, but it was in the absence of these three that Silver had a quicker healing rate than Iodine: wound exudate [Mann-Whitney U (138)=1884.50, p<0.05], new pain [Mann-Whitney U (170)=2599.00, p<0.01] or cellulitis [Mann-Whitney U (192)=3788.00, p<0.05].

These results confirm the findings regarding pain as discussed in Section 6.1. That is, no differences were found in the effectiveness of the antimicrobial treatment groups in resolving wound pain. These results reiterate the need to treat with caution any suggestion that Silver was more effective in managing wound exudate compared to Iodine as discussed in Section 6.3 of this report. Silver was seen to be more effective in the absence of exudate and of comparable effectiveness to Iodine when there was increased or malodorous exudate at baseline (Table 7.7).



Table 7.7

Comparing the Healing Rate at the First Fortnight by Treatment Group and Wound Eligibility Criteria at Baseline (n=259) †^

n= Mean rank Eligibility Characteristic Present

Silver Iodine Silver Iodine

Mann-

Whitney U

Yes 116 114 128.40 102.38 5116.00** Impaired/ delayed healing

No 16 14 15.47 15.54 111.50

Yes 95 82 99.06 77.34 2939.00** New areas of slough / wound breakdown No 37 46 44.78 39.76 748.00

Yes 68 64 72.87 59.73 1743.00* Changes in granulation tissue No 64 64 70.38 58.62 1671.50

Yes 58 62 64.91 56.37 1542.00 Increased / malodorous exudate No 74 66 78.03 62.05 1884.50*

Yes 48 40 45.54 43.25 910.00 Increased/ new pain

No 84 88 99.56 74.03 2599.00**

Yes 33 33 37.03 29.97 428.00 Cellulitis

No 99 95 106.74 87.87 3788.00*


In sum, the results pertaining to clinical signs are suggestive of Silver offering benefits when the wound has signs of a low level of bacterial growth or when showing signs of slowed or impaired healing, some wound breakdown, developing areas of slough, or changes in healthy granulation tissue.

7.4 Chapter Summary Wound swabs were obtained during the trial as a supplementary diagnostic aid and were useful to compare laboratory detection of bacterial burden and infection with the clinical signs of critical colonisation and infection used as the basis for recruitment to the trial and commencement and cessation of treatment with an antimicrobial dressing. The wound swab culture results were also contrasted with healing rates for the antimicrobial treatment groups to determine the performance of the antimicrobials in the presence of certain types or degree of bacteria.

These results reveal that what clinicians judged as signs of critical colonisation and infection according to 10 markers derived from the literature had very little association with what the wound swabs detected. Given that wounds were usually recruited with multiple signs of bacterial burden (an average of 3.28 signs), it is significant that almost four in ten wounds were determined to have nil or scant levels of any bacterial growth and two thirds had no growth of leucocytes according to the wound culture results. A Spearman’s correlation confirmed a very low and non-significant relationship between clinician’s judgment of bacterial burden and infection using this study’s eligibility criteria and the wound swab findings.



These results represent an interesting consideration for researchers and practitioners alike. In addition to the incongruence between the identification of signs of bacterial burden and swab results determined in this study, the practice of obtaining wound swabs as a routine means of determining the need for an antimicrobial dressing is not standard practice for community nurses. It was for this reason that the trial decided to use agreed signs of critical colonisation and infection, not wound swab results, as the benchmark to determine suitability for antimicrobial in this research context.

There is some evidence to suggest that bacteria survive within chronic wounds beneath biofilms which protect bacteria from host defences and the effects of topically applied antimicrobial agents (Edwards & Harding, 2004). It has also been proposed that wound swabs do not penetrate the biofilm and thus the type and extent of bacteria that exist in a wound will not be identified in the presence of a biofilm. This study has identified the need for further investigations to ascertain the effectiveness of wounds swabs for identifying covert or overt infection in chronic wounds.

The results presented in this chapter are consistent with the finding that there was no difference in the clinical effectiveness of the antimicrobial treatments and the presence of wound pain. The results from this chapter also suggest that the performance of the antimicrobials in the presence of increasing or heavy exudate was comparable, with Silver more effective than Iodine in the absence of this dimension. Thus, this result is inconsistent with and that caution is warranted with respect to an earlier finding which suggested that Silver was more effective at managing wound exudate.

The limited number of organisms isolated from wound swabs in this trial restricted an extensive analysis of the antimicrobials by organism type. Nevertheless, the performance of the antimicrobial dressing was considered broadly against the presence of leucocytes, and gram positive and negative bacilli and cocci. The emergent picture is suggestive of better healing with the use of Silver when there is a low level of bacterial growth and signs of delayed or impaired healing, new areas of slough or wound breakdown, and changes in granulation tissue. The effectiveness of Silver in a low level bacterial environment was not able to explain the effectiveness of Silver in the first fortnight or for the group of clients that did not heal in the 12 week study period with the treatment groups and healed / not healed segments comparable with respect to their bacterial load.






8.0 Results: Cost Comparison

This chapter presents the cost comparison of the antimicrobial treatments used in the RCT. As noted in Chapter 4 of this report, there was no significant difference in the healing rate associated with the antimicrobial treatments or the individual products for the 12 week observation period. Thus, in the absence of clinical effectiveness for one treatment compared to the other, the issue of whether one treatment can offer costs savings becomes relevant.

The analysis considers the cost associated with the Silver and Iodine treatment overall as well as exploring the cost comparison for the different dressing types within the product ranges; Acticoat®, Acticoat® 7, Acticoat® Absorbent for Silver and Iodosorb® Powder and Iodosorb® Ointment for Iodine. The cost analysis is examined for the healed/ not healed, wound duration, and wound size segmentations.

To determine the cost associated with each treatment, all wound care products used to treat the study wound were recorded during the observation period and totalled to obtain the respective total ‘product’ costs. The prices used were current market rates upon commencement of the trial in 20068. The number of nurse visits per fortnight was monitored and multiplied by visit duration recorded at each data collection fortnightly visit. Data were only collected for nurse time involved in a client’s care as distinct from time spent in activities related to the trial. A nursing rate of $80 per hour9 was applied to represent the costs associated with the direct care time, travel time and expenses, and other infrastructure costs. The findings in this chapter are therefore presented for ‘product’, ‘nurse’, and the combined ‘product-nurse’ expenses. The total ‘number of visits’ is also tabled as there were differences between products within each of the antimicrobial treatment ranges in terms of their prescribed application frequency.

Analysis of Covariance (ANCOVA) was used to compare costs for the treatment groups. Preliminary checks were conducted to ensure that there was no violation of the assumptions of normality, linearity, homogeneity of variances, homogeneity of regression slopes, and reliable measurement of covariates. Wound size at baseline was used as a covariate in all analyses given the strong and positive association with the dependent variable.

8.1 Silver Versus Iodine 8.1.1 Comparison of the Treatment Groups

Of the clients randomised to Silver, just over half had their wound dressed with Acticoat® – a three day application (56%); a quarter (25%) with Acticoat® 7 - a 8 The contract price of the antimicrobial products was reduced slightly for the trial. The reduction was consistent across product and therefore has not introduced differences by antimicrobial treatment. 9 Derived from the Victorian HACC nursing rate to be used as a benchmark of nursing costs for both agencies. Both organisations note, however, that this figure is not necessarily reflective of the actual more inflated cost of providing the nursing care.



seven day application, and the remaining clients’ wounds were dressed with Acticoat® Absorbant. The majority of clients randomised to Iodine had the Iodosorb® Ointment (69%) used on their wound, with the remainder dressed with Iodosorb® Powder.

The average cost of wound products per client was $605.77 and the average cost associated with nursing care was $1,384.64. A total product and nurse cost was calculated as $1,987.4910 per client. This figure represents the product and nursing care costs per client for a 12 week period, or less if the wound healed. Over this time there was an average of 19.6 visits with an average duration of almost one hour (56 minutes) per visit. The Silver (Ave=18.47) and Iodine (Ave=20.62) groups had a comparable number of nurse visits during the study period [F(1,260)=2.13 , p>0.05]. Table 8.1 presents the cost and visit data for the two antimicrobial treatment groups and the results of the ANCOVA analysis.

No significant difference between the treatment groups was identified for the product cost [F(1,260)=1.41, p>0.05], nurse cost [F(1,260)=2.38, p>0.05], or product-nurse cost [F(1,260)=0.75, p>0.05]. A significant effect between baseline wound size (covariate) and each of the cost measures was identified: product cost [F(1,260)=50.22, p<0.01], nurse cost [F(1,260)=50.64, p<0.01] and product-nurse cost [F(1,260)=68.27, p<0.01]. It was determined that a larger wound at baseline was associated with increased costs; a finding that is not surprising given the expectation that a larger wound would require a longer duration of care, hence more nurse time, and larger (and more) wound products which have a higher per unit purchase price.

Table 8.1 Average Costs for Treatment Groups (n=263)†


10 This total product-nurse cost differs marginally throughout the report from a direct sum of the product and nurse averages due to missing data in either the product or nurse totals altering the average of the combined cost.

Ave SD

Silver Iodine Silver Iodine F value

(n=131)^ (n=132)^

Ave. product costs ($) 618.92 577.62 416.63 407.74 1.41

Ave. nurse costs ($) 1281.28 1480.22 873.57 1021.68 2.38

Ave. product-nurse costs ($) 1895.03 2056.70 115.91 1279.08 0.75

Ave. number of visits (n) 18.47 20.62 10.31 11.16 2.13


Table 8.2

Average Costs for Treatment Groups excluding Acticoat® 7 (n=230)†

†

Sample includes all clients included in the intention-to-treat analysis

Ave SD

Silver Iodine Silver Iodine F value

(n=98)^ (n=132)^

Ave. product costs ($) 670.83 577.62 433.85 407.74 3.40

Ave. nurse costs ($) 1380.29 1480.22 890.68 1021.68 0.60

Ave. product-nurse costs ($) 2048.28 2056.70 1117.88 1279.08 0.01

Ave. number of visits (n) 20.40 20.62 10.00 11.16 0.01

^Sample size for question(s) vary due to missing data * Significant <0.05 ** Significant <0.01 *** Significant <0.001 As Acticoat® 7 required once a week dressing changes, with remaining products indicated for a change at three day intervals, there was interest to complete the cost effectiveness comparisons inclusive and exclusive of this antimicrobial dressing sub-type. It was observed that Acticoat® 7 was used at only one of the study sites though it was available for selection at both. It would seem that familiarity and perceived effectiveness of the product played a role in whether clinicians felt this product was clinically indicated for their clients.

Table 8.2 presents the results of ANCOVA excluding the results for clients who received Acticoat® 7. There remained no significant difference between the antimicrobials for product cost [F(1,227)=3.40, p>0.01], nurse cost [F(1,227)=0.60, p>0.01], product-nurse cost [F(1,227)=0.01, p>0.01], or total number of visits [F(1,227)=0.01, p>0.01]. Once again, a significant effect between baseline wound size (covariate) and each of the cost measures was identified: product cost [F(1,227)=41.63, p<0.01], nurse cost [F(1,227)=47.45, p<0.01] and product-nurse cost [F(1,227)=62.80, p<0.01].

8.1.2 Comparison of the Individual Antimicrobial Products ANCOVA was used to compare the cost effectiveness of the different antimicrobial dressing products (Acticoat®, Acticoat® 7, Acticoat® Absorbant, Iodosorb® Powder, and Iodosorb® Ointment). Only clients who remained on the same antimicrobial product and did not change to another product in the range were included in this analysis. Only nine clients were excluded from the analysis because they used a different product within the antimicrobial range they were randomised to. Table 8.3 presents these results for the five dressing products after adjusting for baseline wound size. The results identified a significant difference between the antimicrobials for the product costs [F(4,248)=3.32, p<0.01] and total number of visits [F(4,248)=4.08, p<0.01]. However, the nurse costs [F(4,248)=1.49, p>0.01] and product-nurse costs [F(4,248)=1.73, p>0.01] did not differ significantly based on which antimicrobial product was used.



Pairwise Comparisons revealed that product cost was significantly higher for Acticoat® Absorbant (Ave=$878.56, SD=366.21) compared to the cost associated with the use of Acticoat® 7 (Ave=$474.85, SD=318.87) [F(1,248)=313.45, p<0.05] and Iodosorb® Ointment (Ave=$521.32, SD=363.16) [F(1,248)=275.57, p<0.05].

The average number of visits was significantly lower for clients receiving Acticoat® 7 (Ave=12.96, SD=9.13) compared to other antimicrobial types which each had close to or more than 20 visits during the study period; Acticoat® (Ave=19.74, SD=10.11) [F(1,248)=-6.23, p<0.05], Acticoat® Absorbent (Ave=24.12, SD=8.45) [F(1,248)=-8.58, p<0.01], Iodosorb® Powder (Ave=23.05, SD=11.04) [F(1,248)=-7.39, p<0.05], and Iodosorb® Ointment (Ave=20.11, SD=10.95) [F(1,248)=-6.90, p<0.01].

Again, baseline wound size (covariate) was significantly related with all cost indicators: product cost [F(1,248)=41.22, p<0.01], nurse cost [F(1,248)=50.95, p<0.01] and product-nurse cost [F(1,248)=66.08, p<0.01]. Overall, the larger wounds were associated with greater expense and more visits.

Table 8.3

Average Costs for Individual Antimicrobial Products (n=254)†

Acticoat® Iodosorb®

Acticoat® 7 day Absorbant Powder Ointment F value

(n=70)^ (n=32)^ (n=24)^ (n=40)^ (n=88)^

Product costs ($) Ave. 617.08 474.85 878.56 715.74 521.32 3.32**

SD 437.92 318.87 366.21 466.55 363.16

Nurse costs ($) Ave. 1369.52 1001.49 1541.47 1658.77 1441.49 1.49

SD 903.89 766.80 842.21 1177.41 938.16

Product-nurse costs ($) Ave. 1982.61 1463.87 2420.03 2359.55 1967.89 1.73

SD 1115.32 998.79 1030.94 1447.54 1173.95

Number of visits (n) Ave. 19.74 12.96 24.12 23.05 20.11 4.08**

SD 10.11 9.13 8.45 11.04 10.95


8.2 Segmentation Group Comparison 8.2.1 Healed and Not Healed Group Comparison

A cost effectiveness comparison of the treatment groups was completed for both healed and not-healed segments. Table 8.4 presents the results of ANCOVA analyses comparing the cost effectiveness of the two antimicrobials for the healed and not healed groups, after adjusting for wound size. The results



mirrored the overall results. That is, there was no significant difference between the treatment groups for any of the cost indicators for either the healed; product cost [F(1,164)=3.59, p>0.01]; nurse cost [F(1,164)=1.49, p>0.01], product-nurse cost [F(1, 164)=0.05, p>0.01], or not healed segments; product cost [F(1, 93)=0.01, p>0.01], nurse cost [F(1,93)=1.87, p>0.01], product-nurse cost [F(1,93)=1.44, p>0.01]. Similarly, there was also no significant difference in the total number of visits for the healed [F(1,164)=0.72, p>0.01] or not healed groups [F(1,93)=3.25, p>0.01]. There was a significant effect between baseline wound size (covariate) and each cost indicator for the healed (product cost [F(1,164)=22.39, p<0.01]; nurse cost [F(1,164)=27.67, p<0.01]; product-nurse cost [F(1,164)=34.02, p<0.01]) and not healed groups (product cost [F(1,93)=10.93, p<0.01], nurse cost [F(1,93)=5.05, p<0.01]; product-nurse cost [F(1,93)=10.88, p<0.01]). Table 8.4

Average Costs for Treatment Groups by Healed / Not Healed Segmentation (n=263)†

Average SD


F value

Healed Segment (n=167)^

Product costs ($) 529.62 438.61 361.48 296.82 3.59

Nurse costs ($) 900.96 1017.41 637.04 679.14 1.49

Product-nurse costs ($) 1432.01 1461.42 882.02 873.17 0.05

Number of visits (n) 14.09 15.09 8.26 8.12 0.72

Not Healed Segment (n=96)^

Product costs ($) 778.53 813.08 462.81 461.74 0.01

Nurse costs ($) 1961.01 2264.17 829.57 1031.06 1.87




8.2.2 Wound Duration Group Comparison ANCOVA was used to examine the cost effectiveness of the treatment groups for both young wounds (>12 weeks) and old wounds (<12 weeks) as determined by their wound duration at entry to the study. As shown in Table 8.5, there was a significant difference between the treatment groups for product cost for young wounds [F(1,140)=3.98, p<0.05] with the average cost of products higher for Silver (Ave=620.31, SD=405.92) compared to the Iodine group (Ave=498.98, SD=348.66). However, the higher cost for products did not translate to the overall product-nurse costs. The baseline wound size (covariate) was significantly related with all cost variables for young wounds; product cost [F(1,140)=21.73, p<0.01], nurse cost [F(1,140)=8.65, p<0.01] and product-nurse cost [F(1,140)=15.49, p<0.01], and old wounds; product cost



[F(1,117)=24.96, p<0.01], nurse cost [F(1,117)=46.60, p<0.01] and product-nurse cost [F(1,117)=55.54, p<0.01].

Table 8.5

Average Costs for Treatment Groups by Wound Duration Segmentation (n=263)†

0

Average SD


F value

Young Wound Segment (n=143)^

Product costs ($) 620.31 498.98 405.92 348.66 3.98*

Nurse costs ($) 1097.51 1318.93 740.90 970.92 2.59



Old Wound Segment (n=120)^

Product costs ($) 617.13 663.75 433.77 451.14 0.03

Nurse costs ($) 1519.86 1656.88 976.41 1054.00 0.06




8.2.3 Wound Size Group Comparison

ANCOVA was used to examine the cost effectiveness of the antimicrobial treatment groups for small wounds (≤3.61cm²) and big wounds (>3.61 cm²) as determined by their wound size at entry to the study. As shown at Table 8.6, there was a significant difference between the treatment groups in the big wound segment for nurse cost [F(1,101)=6.96, p<0.01] and the combined product-nurse cost [F(1,101)=4.45, p<0.05] with nursing costs higher for the Iodine group (Ave=2110.78, SD=884.33) compared to the Silver group (Ave=1647.19, SD=897.78). The combined product-nurse cost was also higher for Iodine (Ave=2881.83, SD=1151.46) compared to Silver (Ave=2409.75, SD=117.68).



Table 8.6

Average Costs for Treatment Groups by Wound Size Segmentation (n=263)†

0

Average SD


F value

Small Wound Segment (n=157)^

Product costs ($) 520.41 440.44 369.94 294.30 2.04

Nurse costs ($) 1037.18 1027.31 753.40 748.86 0.01



Big Wound Segment (n=103)^

Product costs ($) 763.31 873.01 398.85 447.22 0.05

Nurse costs ($) 1647.19 2110.78 897.78 884.33 6.96**

Product-nurse costs ($) 2409.75 2881.83 117.68 1151.46 4.45*



8.3 Chapter Summary The average cost of wound products per client in this trial was $605.77. The average cost of nursing care per client in this trial was $1,384.64. There was no significant difference between the antimicrobial treatment groups in terms of: the cost of wound products; the cost of providing nursing care; the combined cost of products and nursing care; or the number of visits. In the absence of any differences between the two antimicrobial treatments with respect to their overall clinical effectiveness, these antimicrobial treatments also represent a comparable cost for either the client or health care service.

Wound size at baseline was consistently associated with the indicators of cost considered in this trial. A larger wound at baseline will cost more in products, nursing care, combined products and nursing care, and will require more visits. This finding is unsurprising as larger wounds require bigger, more expensive dressings and will take longer to heal, requiring more visits and more nursing care.

The individual products were contrasted to determine if any product offered cost savings compared to another. It should be emphasised that cost alone would not be expected to be the basis on which a clinician would select one of these products over the other; clinical presentation of the wound including features such as exudate levels would be the major influence on clinical decision making. However, in environments where the client must meet the cost of wound products, cheaper products may be selected if the client cannot afford recommended practice. Yet, the selection of cheaper products can potentially create greater overall costs if delayed healing occurs, with the knock on effect of a longer period of care, more nursing costs, and the purchase of more, albeit cheaper, wound products. The analysis did demonstrate significantly higher



products costs when Acticoat® Absorbant was used, compared to Acticoat® 7 and Iodosorb® Ointment. There were also significantly fewer visits for clients receiving Acticoat® 7 compared to all other products. This finding is simply a reflection of the fact that Acticoat® 7 can remain in place for seven days compared to all other products which require changing at a maximum of every three days.

The segmentation analysis found both healed and not healed segments replicated the overall results with no significant cost differences between the treatment groups and any of the cost variables. Young wounds being treated with Silver had a higher product cost compared to those receiving Iodine, though the difference was not significant for nursing care nor when nursing care was factored into the overall product-care cost. There were no differences between the treatment groups for older wounds. Big wounds on the other hand were significantly more expensive with respect to nurse care costs and the combined product and care costs if the wound was dressed with Iodine rather than Silver reflecting the clinical benefits of the Silver dressing for this segment.

This trial has generally observed the clinical effectiveness of the two antimicrobial treatments to be comparable (with some exceptions noted in Chapter 4). In the absence of a clinical imperative to select a particular antimicrobial, this trial has also found that there was no cost justification identified to select either Silver or Iodine. Two exceptions to this statement apply: one for young wounds and the other for large wounds. If product expenses are of concern to the client or health care service, Iodine might be a preferred product if the client has a young wound and no other indications for which Silver would be preferred. It needs noting, however, that once nursing care costs are taken into consideration there is no cost difference between the treatments. This finding may be particularly relevant for services where clients are required to pay for their wound products but receive government or program funding to subsidise nursing care costs as was the case at one of the RCT sites. When presented with wounds larger than 3.6cm2 the cost associated with nursing care and with the combined product-nurse cost is significantly greater when being treated with Iodine compared to Silver. Thus, unless there are known contraindications to the use of Silver for an individual, there are cost savings possible for the client or health care service if treating a large wound with a Silver antimicrobial.



9.0 Discussion & Recommendations

This Randomised Controlled Trial (RCT) was undertaken to compare the clinical and cost effectiveness of two commonly used antimicrobial products to treat lower leg ulcers that were displaying signs of bacterial burden. The use of an antimicrobial is clinically indicated when a wound appears critically colonised or infected and the bacterial burden is unlikely to be managed by drainage, debridement or cleansing alone (Principles of best practice: Wound infection in clinical practice. An international consensus, 2008). In the absence of clinical trial comparisons of two commonly used antimicrobials – Silver and Iodine dressings - for their impact on wound healing, this trial sought to provide evidence and guidance on antimicrobial selection and in doing so to respond to the principle aim of a broader initiative; to improve the care provided to and outcomes for all people living with a lower leg ulcer.

This RCT gathered data from 281 clients located at two Australian community nursing services; Royal District Nursing Service (RDNS) in Melbourne, Victoria and Silver Chain Nursing Association (Silver Chain) in Perth, Western Australia. An intention-to-treat analysis was possible for 266 clients (95% inclusion rate) – 133 randomly allocated to Silver and 133 randomly allocated to Iodine – for whom a healing rate could be determined. Healing rate was measured using the Medseed AMWIS™ wound imaging software (Santamaria et al., 2002; Santamaria et al., 2004; Santamaria & Clayton, 2000). This trial assessed healing and healing rates fortnightly for 12 weeks. Data were collected on potential covariates, costs associated with wound product usage and nursing care were monitored, and wound swabs were attended at baseline, when ceasing or recommencing an antimicrobial treatment, and at six and 12 weeks if treatment with an antimicrobial continued.

Data were considered for the overall sample as well as for several segmentations including wounds that healed and those that did not during the twelve week period, those with short and long wound duration, and two sizes of wounds broadly divided into and described as small and large. All of these were used as sub-analyses to determine whether there were differences in the response to the antimicrobials which were not identifiable in the overall combined sample analysis.

In this chapter, the findings from the results chapters are summarised with the implications of these findings on clinical practice and future research discussed, and the strengths and limitations of this trial considered.

9.1 Clinical Effectiveness of the Antimicrobials 9.1.1 Wound Healing

This trial found both antimicrobials had a positive effect on healing lower leg ulcers that were assessed as having signs of critical colonisation or infection. Sixty-four per cent of wounds that were compromised by bacterial burden at recruitment went on to heal within the 12 week observation period; a significant outcome given that prior studies expected 70% healing after 12 weeks of



treatment for uncomplicated leg ulcers (Colgan et al., 1996; Moffatt et al., 2003).

This trial determined that the two antimicrobials had a comparable performance for healing rate and the number of wounds that healed. Based on this evidence clinicians should feel justified to select either antimicrobial without compromising the clinical care they provide to their clients with leg ulcers. There were, however, some specific circumstances where the Silver dressing achieved a quicker healing rate compared to the Iodine dressing. These circumstances included:

1. In the first fortnight. The positive result for Silver was identified for the overall sample and all groups explored in segmentation analyses except for the ‘healed’ segment (thus this finding was true for the ‘not healed’, short or long wound duration, and both wound size segments).

2. When treating a client who would be unlikely to heal in a 12 week period. That is, a client with a large (in this study, an average size of 10cm2) and chronic wound (a median of 24 weeks duration or almost six months characterised this sample), who may have moderate to high levels of wound exudate (a quarter of the sample) also fared better on the Silver antimicrobial.

With respect to exudate management, the results showed that both treatments reduced the level of wound exudate. There was a significantly lower level of exudate for the fourth fortnight assessment for the Silver group compared to Iodine. There were no other statistically substantiated differences between the treatment groups and indeed the effectiveness of the Silver treatment was not supported by other analyses which found no statistically significant difference between the two antimicrobial treatments for a segment of clients who were recruited to the trial for the reason of having increased or malodorous exudate. Rather, it was when participants did not demonstrate this criterion at recruitment that Silver was associated with swifter healing. Thus, in the absence of consistent and statistically validated results, little can be concluded regarding any differences in the effectiveness of the antimicrobials in managing wound exudate. The need for further research is recommended.

The level of pain experienced by clients also declined comparably for both antimicrobials; a significant outcome given that wound pain is a highly reliable indication of wound compromise (Gardner, Frantz, & Doebbeling, 2001).

Based upon these results, it is suggested that when wanting to achieve an immediate reduction in wound size and when attending clients with large and chronic wounds, a Silver antimicrobial should be selected given the clinical benefits observed for Silver in these circumstances. At all other times, the clinician should feel confident to select either antimicrobial without compromising the clinical care they provide to their clients.

Wound culture results identified that Silver was more effective when wounds presented with low levels of bacterial burden. This result was used to explore whether the treatment groups and the segment of chronic wounds that did not heal during the 12 week observation period differed with respect to their level of bacterial burden which, if they did, might explained the effectiveness of the Silver antimicrobial for the first fortnight and for the not healed segment. However, there was no difference in the level of bacterial burden as measured



by the wound swabs between either the antimicrobial treatment groups or the healed and not healed segments.

In attempting to understand these results, it appears that Silver reduces wound size immediately, with the lack of a difference in the overall clinical effectiveness of the treatment suggestive that Iodine operates in a different though ultimately as effective manner. Analysis of changes in tissue type during the observation period did not reveal any consistent variations in effect on tissue type by the antimicrobials. It is speculated that Iodine’s noted inflammatory mechanism (Moore et al., 1997) which contrasts to Silver’s anti-inflammatory action (Sibbald et al., 2007) may explain the absence of an immediate wound size reduction observed in this trial for wounds in the Iodine treatment group. However, this trial lacks data to comment further on the mechanism of action of both antimicrobials with particular reference to inflammation. It is recommended that future research which compares these dressings incorporate a measure which can monitor the presence of inflammation at baseline and throughout the trial to assist the interpretation of immediate and longer term clinical progression of the wound as well as explore the potential that a sub-type of client would be better served by one antimicrobial given the presence or absence of existing wound inflammation.

In reflecting back on the published literature, it is of interest to note systematic reviews also detected a significant and immediate reduction in ulcer size when treated with a silver antimicrobial, in those trials within the first four weeks, a result that was not sustained during longer term follow up (O'Meara et al., 2001). Other systematic reviews joined the call for more and rigorous randomised controlled trials to further assess Silver treatments (Meaume et al., 2005; Munter et al., 2006). The published literature generally suggested that Iodine was an effective treatment and indeed this trial has found that both treatments were used effectively to resolve bacterial burden and aid the healing of wounds. Emerging from the literature was the suggestion that Iodine was characterised as offering wound exudate absorption and assisted with the debriding process (Campton-Johnston & Wilson, 2001). It is a recommended that the capacity of these antimicrobials to manage wound exudate is the focus of future investigations given some interesting but inconsistent results obtained in this study regarding the greater effectiveness of Silver rather than Iodine in exudate management.

9.1.2 Predictors of Time to Healing The results of this trial substantiate existing evidence in the area of chronic leg ulcer management by reinforcing long established messages regarding adherence to compression bandaging when indicated, and confirming that undesirable wound characteristics such as the presence of slough, exudate, a larger wound size and longer wound duration are likely to be associated with slower wound healing.

Certainly other studies have found that faster healing is found among smaller and younger venous leg ulcers (Margolis et al., 2004; Phillips et al., 2000; Tennvall & Hjelmgren, 2005), and in this trial the group of clients that did not heal were characterised by having larger and older leg ulcers. The need for early intervention to prevent wounds becoming large and chronic is implicit; a matter which may best be achieved through client and clinician education to



ensure early action in the event of a new wound occurring or, even better, upon the development of signs of skin integrity compromise. The implications of these findings reaffirm the need to adopt appropriate dressings or treatments to reduce wound exudate and slough to achieve faster healing.

The finding that adherence to compression therapy was a predictor of healing adds to a credible body of evidence regarding the clinical effectiveness of graduated compression therapy on the healing of those leg ulcers for which compression therapy is indicated (Moffatt, 2002; O'Meara et al., 2009). It adds weight to the need to optimise the use of compression bandaging therapy when clinically indicated, addressing any barriers of client resistance, nurse resistance, and access and equity with regard to wound care products which may hamper the use of optimal compression bandaging therapy for the duration of an individual’s wound care (Annells et al., 2008; Flowers et al., 2007; Posnett, 2006). The Australian Wound Management Association which is currently developing the first Australian guidelines on the management of venous leg ulcers will be encouraged and informed by the findings of this study to adopt a strong stance on reinforcing every possible avenue and opportunity that will facilitate the ongoing use of graduated compression bandaging when indicated.

9.2 Comparison of Signs of Bacterial Burden and Wound Swab Results Comparisons of clinician assessed signs of critical colonisation and infection according to 10 markers derived from the literature with the results obtained from wound swabs identified no relationship between the two approaches.

A comparison of the treatment groups for microbiology data and signs of critical colonisation and infection suggested that better healing was achieved with the use of Silver when there was a low level of bacterial growth as well as for cases reported to have delayed or impaired healing, new areas of slough or wound breakdown, and changes in granulation tissue. Though these results would ordinarily translate into a recommendation to use Silver when wound culture results suggest a low level of bacterial burden or given these particular signs of critical colonisation, this trial has raised several questions regarding the utility of wound swab data compared to clinician assessment. As a result of this there appears to be a strong case for obtaining further clarity regarding the reliability of wound swab and clinical assessment data through formal research.

In practice, clinician assessment of bacterial burden as a basis to commence an antimicrobial treatment or, in the case of suspected infection, to recommend antibiotic treatment, is frequently the only information used to initiate treatment. That is not to say that wound swabs aren’t also pursued to further inform decision making; quite often wound swabs, and perhaps wound biopsies, are sought, though the pursuit of these results can delay the commencement of treatment. This trial has clearly identified that these two methods of assessing wound bacterial burden do not necessarily produce congruent results; a finding which has implications for both research and practice. In this trial clients were recruited and commenced on an antimicrobial dressing based upon clinical assessment using a list of 10 signs of critical colonisation and infection derived from the published literature. As the evidence obtained from the wound swabs was not comparable the question therefore arises, how can the two methods complement each other?



Firstly to consider the appropriateness of clinician assessment and the signs of bacterial burden used to identify critical colonisation and infection. The list of 10 signs of critical colonisation and infection used to assess eligibility with respect to having a wound that was indicating critical colonisation or infection was extracted from a review of the literature (see Box 9.1) and was the subject of intensive discussion and subsequent consensus within the study team. The wounds included in this study presented with a number of clinical signs which would suggest antimicrobial treatment was warranted. It might be the case that the signs which constituted the list of critical colonisation and infection are not reliably associated with or indicative of bacterial burden, or alternatively the disparity with the wound culture results may suggest that clinical signs of burden develop in advance of bacterial presence that can be accurately detected on a standard wound swab and culture. This trial did not investigate the comparability or reliability of different clinician’s assessment of these signs of critical colonisation and infection, a comparison which might be the focus of future research. However, the authors postulate that clinician assessment using the 10 signs of critical colonisation and infection was an appropriate means of assessing bacterial burden in a wound, and is, nevertheless, the most likely ongoing basis for deciding whether antimicrobial treatment is warranted given logistic constraints.

Box 9.1. Signs of Critical Colonisation & Infection

• Cellulitis (pain, heat, erythema, swelling of surrounding tissues)

• Suppuration (purulence or the presence of pus)

• Lymphangitis

• Sepsis (confirmed by blood test)

• Bacteraemia (confirmed by blood test)

• Changes in granulation tissue (change in colour of granulation tissue, which may appear dusky, darker or bright red, hypergranulation tissue, friable or fragile granulation tissue that is prone to bleeding)

• Increased or malodourous exudate

• New areas of slough or wound breakdown

• Impaired or delayed wound healing (epidermis fails to migrate across the wound bed, static, rolled or undermined wound edges or bridging segments of epithelial closure and breakdown in the tissues)

• Increased or new pain

(Cutting & Harding, 1994; Gardner, Frantz, & Doebbeling, 2001; Gardner, Frantz, Troia et al., 2001; Schultz, Barillo, Mozingo, & Chin, 2004; Schultz et al., 2003; Sibbald et al., 2000)

The swabs in this study were obtained by nurses according to a standardised swabbing technique and in accordance with recommended specimen storage and transportation guidelines. Both trial sites obtained the swabs using Z or zig-zag technique (Cooper & Lawrence, 1996) from wounds which had been irrigated



using sterile water. Two swabs were attended at each collection; one swab was placed in the Amies Transport Media, the second used to prepare a glass slide by placing the tip of the swab in the centre of the glass then rolling the side of the swab over the middle one-third of the slide. The air dried slide was then placed in a plastic slide carrier. Eskies and cold packs were provided to clinicians to store specimens if the outside temperature was greater than room temperature. Specimens were arranged for collection or dropped off to a collection centre within 24 hours of collection. Overall, it could be claimed that the measures used to ensure the wound swab was attended in a valid and reliable way resulted in integrity of the bacteria on the wound swab.

The wound swab offers a number of advantages compared to other methods. For example, aspirating fluid from a wound or obtaining a wound biopsy is invasive and can cause trauma to the wound bed, and the value of biopsy results compared to surface swabs has been questioned (Bill et al., 2001; Cooper & Lawrence, 1996). However, there are limitations which include sampling bacteria from the surface of the wound only, restrictions on its ability to isolate individual species (Cooper & Lawrence, 1996), and emerging as a foci in recent times, the postulated inability of the wound swab to penetrate biofilms on wounds.

Biofilms are complex communities of bacteria which attach to solid surfaces before forming microcolonies which finally differentiate into an exopolysacccharide-encased mature biofilm (Wysocki, 2002). A favoured environment conducive to biofilm development given the moist conditions (Wysocki, 2002), it has been estimated that 60% of chronic wounds may contain a biofilm (James et al., 2008).

Though the clinician may assess signs of critical colonisation or infection in and around the wound, if the wound has a biofilm, the swab is unlikely to penetrate the biofilm to sample the bacteria within. The wound culture results would then present a distorted picture of wound bacterial burden and would not be a reliable indicator of critical colonisation or infection.

Not only is the ability of the swab to penetrate the biofilm questioned, but with particular relevance to the principle objective of the trial at hand, clarity is required to address the incongruence of in vitro findings regarding the effectiveness of antimicrobials in resolving bacterial burden in the presence of a biofilm. In one study Silver, in the presence of a biofilm, was observed to kill all pathogens within 48 hours (Percival, Bowler, & Woods, 2008). Another study found that microorganisms were less susceptible to antimicrobials when in a biofilm state (Percival, Bowler, & Dolman, 2007).

The implications of this research include the need for further studies to: identify when a wound has a biofilm; to advance microbiology methods to reliably assess bacterial burden in the presence of a biofilm; and, to refine the treatment for wounds that have a biofilm. Effective and safe debridement techniques, particularly in the home nursing setting, presents as an obvious area of enquiry as surgical removal of the sessile population has been noted as necessary to eliminate the pathogenic bacteria and prevent the recurrence of infection (Wysocki, 2002). Until methods of sampling and assessing bacteria in the presence of a biofilm are better understood, it would appear sensible to avoid an absolute reliance on microbiology measures. Preferably, a number of methods would be used in conjunction with clinician assessment with some consideration



of how additional data could be harnessed to further assess the validity and reliability of the respective approaches.

9.3 Cost Comparison of the Antimicrobials The cost comparison of the two antimicrobial treatments principally considered the cost for the study period associated with wound product use, the provision of nursing care, and the combined product and nursing time cost. The number of nurse visits during the study period was also monitored.

There was no overall cost difference between the antimicrobial treatments identified for any of the cost outcome measures. Thus, there is no cost saving to be achieved by selecting one antimicrobial in favour of the other. The segmentation analysis found only two exceptions where the costs associated with the antimicrobial treatment differed; for young wounds and large wounds. Costs were greater for large wounds (bigger than 3.6cm2) with respect to nurse care costs and the combined product-nurse costs if the client was receiving Iodine rather than Silver reflecting the clinical benefits of the Silver dressing for this segment. Young wounds (less than 12 weeks old) being treated with Silver had a higher product cost compared to those treated with Iodine, though the difference was not significant for nursing care nor when nursing care was factored into the overall product-nurse cost.

Comparisons of the individual products within the Silver or Iodine ranges were undertaken to determine if any specific product offered cost savings compared to the others. The cost of products calculated for the study period was significantly higher for Acticoat® Absorbant compared to Acticoat® 7 and Iodosorb® Ointment. This was not the case for Acticoat® or Iodosorb® Powder which had moderate costs and for which the cost difference to Acticoat® Absorbant was not as sizable. Acticoat® Absorbant is a specialist antimicrobial designed for heavily exudating wounds. The implication of this result is to use Acticoat® Absorbant only when clinically indicated and to change from Acticoat® Absorbant to an alternate dressing when it is clinically justified, in order to minimise unnecessary expense. There were also significantly fewer visits for clients receiving Acticoat® 7 compared to all other products; this finding a reflection of the fact that Acticoat® 7 can remain in place for seven days compared to all other products which require changing at a maximum of every three days. Again, if clinically justified, there are advantages for clients as well as services in selecting Acticoat® 7 in situations where the expense associated with a visit is great, for example, remotely located clients.

Wound size at baseline was consistently associated with the trial’s cost outcome measures. A larger wound at baseline cost more in products, nursing care, combined products and nursing care, and required more visits. This finding is not surprising as larger wounds require bigger, more expensive dressings and will take longer to heal, requiring more clinical visits and nursing care (Margolis et al., 2004; Phillips et al., 2000; Tennvall & Hjelmgren, 2005).

To summarise, if product expenses are of concern to the client and, in the absence of a clinical rationale indicating the use of one particular antimicrobial, Iodine might be a preferred product if the client has a young wound as product costs are cheaper for this group than Silver. However, for a Health Service, adding the nursing care costs to the product cost does not show the same benefit for young wounds, indicating that the delivery of Iodine in terms of treatment



time and frequency outweighs the initial product cost saving. There are cost savings possible for the client or health care service if treating a large wound (greater than 3.6cm2) with a Silver antimicrobial. There is also merit in changing from Acticoat® Absorbant as soon as clinically indicated, and in selecting Acticoat® 7 to minimise the cost associated with home visits.

9.4 Study Strengths and Limitations 9.4.1 Study strengths

With a prevalence of 1.1 to 3.0 per thousand of the adult population (in Western countries), leg ulcers can result in or exacerbate significant morbidity and financial burden. The annual cost of treating these chronic wounds is estimated at A$3 billion with recurrence of ulceration reported to be as high as 69% (Angel et al., 2005; Kapp & Sayers, 2008). Any inroads that can be achieved to guide clinical practice and in doing so improve client care and outcomes, will reduce the distress of living with an ulcer for individuals, relieve resource strapped community health care providers, and reduce the cost of caring for these wounds for individuals and the community.

This trial offers the first direct comparison of two commonly used antimicrobial dressings – Silver and Iodine. It represents the only evidence available to date as to the clinical effectiveness and cost comparison of these two treatments. It provides the first data from which guidelines as to antimicrobial selection can be refined. The evidence was gathered using a multi-site, randomised controlled trial design which achieved a remarkable 95% inclusion rate of its recruited clients in an intention-to-treat analysis. Digital imaging and wound sizing software was utilised to quantify the wound area and proportions of different tissue types for which a very high inter-rater reliability was observed (Flowers et al., 2008). Wound swabs were additionally undertaken to aid analysis of healing rates and for comparison to clinical assessment of critical colonisation and infection in wounds. It is suggested that these results are a robust representation of the group of clients who met the eligibility criteria.

This trial also pursued multiple means of evaluating outcomes of lower leg ulcer care. Total ulcer healing (100% epithelialisation or wound closure), the traditional outcome measure for clinical trials (Steed et al., 2006) was accompanied by other important indicators of advancement in ulcer care including healing rate and assessment of clinical characteristics of the wound such as tissue type, wound pain, and wound exudate. It has been noted that single point measures such as 100% wound closure fail to provide detailed information of healing and efficacy of treatments over an entire continuum of the ulcer healing process (Hokanson, Hayward, Carney, Phillips, & Robson, 1991; Robson et al., 2008). Wound healing trajectories (healing time curves) have been reported to be effective outcome measures for leg ulcers and provide more information about the healing process while allowing for dynamic comparison at meaningful intervals. It was because of this approach that nuances in the clinical effectiveness of the antimicrobial treatments could be identified.

Finally, this RCT represents a significant achievement and benchmark study for community nursing research in Australia if not more broadly, providing an opportunity for community nursing to demonstrate its capacity to lead research



and innovation in the health sector. It has led to questions and recommendations on how future rigor and methods might be improved in formal investigations, including a paper which details the strengths of and difficulties associated with the conducted of trial research by community nursing services (Newall et al., 2009). It has also directly led to further research and research questions as well as informing the wound care community generally both through this report and the publication and presentation of results nationally and internationally (Carville et al., 2008a, 2008b; Kapp, Flowers et al., 2008a, 2008b).

9.4.2 Study Limitations The design of this trial could have been strengthened by participant and data collector blinding. However, presentation of the two antimicrobial treatments was quite distinctive and different. It was therefore felt that participants could not be blinded from their allocated treatment. Indeed, even if great attention had been given to shielding the client from viewing the wound during dressing changes, the composition of the treatment, one a dressing, the other a paste and powder, might have been discerned by sensation alone. For similar reasons it would be impossible to blind the data collector who, given resource constraints, was also responsible for providing the clinical care. To avoid clinicians viewing the randomisation to one antimicrobial as removing clinical judgement with regard to the most appropriate treatment, considerable effort was invested in ensuring all study communications and education to nurses identified that there was no prior comparison of these two product ranges and, therefore, how they would perform with respect to each other was unknown.

Another limitation of the study might perceivably be the lack of a “no treatment” or comparison group. How would the outcomes of clients randomised to receive one or the other antimicrobial have compared to a similar group for whom no antimicrobial was the randomised choice? This question was asked, discussed at length and discarded at the outset, when the study team decided against such a comparison group because of the clinicians’ duty of care in instigating the evidence based treatment approach when clinical signs pointed to critical bacterial load in the wound. However, the question remains outstanding to compare healing outcomes from treatment with an antimicrobial to a comparison, no antimicrobial treatment group given the presence of recommended multi-layer compression therapy and best practice wound bed preparation for both groups. This is an area which would be the welcomed focus for future research.

In hindsight, this trial would have benefited from a considered measure of wound inflammation collected consistently at the same fortnightly intervals as remaining trial data. In attempting to understand the different modes of action of each antimicrobial, attention was directed to evidence as to Iodine’s noted inflammatory mechanism (Moore et al., 1997) and Silver’s anti-inflammatory action (Sibbald et al., 2007). A measure of inflammation assessed at baseline and throughout the trial may have assisted with the interpretation of healing rate data. The presence of inflammation at baseline may also be identified by future studies as a basis for selecting one antimicrobial over the other.

There was no statistical adjustment for power used in the analyses. As such, any result which was marginally significant at an alpha of 0.05, did not reconcile



with other data, or present as a consistent pattern of effect was qualified in the text.

Finally, a principle limitation of clinical trial research, this trial included, is the representativeness of the findings to groups of clients excluded from the trial. In an attempt to enhance the power of the study given restraints of time, money, and sample, restrictions on eligibility for trials are established. In this trial the most significant exclusion was clients with a diagnosis of diabetes mellitus. Therefore, there is a need for research comparing these antimicrobials on clients with diabetes mellitus as well as arterial leg ulcers and on wounds other than leg ulcers.

9.5 Conclusion This trial has confirmed anecdotal expectations of these two familiar antimicrobials to treat critically colonised or infected lower leg ulcers; that is, both performed well at resolving bacterial burden and aiding healing. Generally speaking, clinicians would be justified to select either, knowing they were contributing positively to their client’s wound healing by doing so. It would be prudent, however, for clinicians to favour Silver if seeking an immediate wound size reduction and for clients with large and long standing wounds. There is little information from a cost comparison of these products – unless presenting with a large or young wound - that would suggest that one antimicrobial should be preferred over the other. This trial provides affirmation of the clinical effectiveness of both treatments and some insights as to a handful of circumstances where one product may be a preferred treatment over the other.

This trial has confirmed the role of and need for health care providers to manage slough and wound exudate whilst advocating strongly for the use of compression bandaging therapy. Measures are also required to facilitate early intervention to avoid wounds becoming large and chronic. All these factors were identified as significant with respect to wound healing. In relation to the accurate diagnosis of bacterial burden or how laboratory findings relate to clinical signs, the way is open for further research which might lead to greater understanding of the part played by biofilms and the clinical choices regarding their management.

9.6 Recommendations The following recommendations were derived directly from the findings of the antimicrobial RCT.

1. That clinical guidelines be revised to reflect the results of this trial as indicating:

a) Both Silver and Iodine treatments achieved comparable clinical effectiveness with the exception of two instances in which Silver had a quicker healing rate than Iodine including:

i. the first fortnight of treatment.

ii. wounds that did not heal in a 12 week period. In the study these wounds, in comparison to wounds that did heal, were larger (an average size of 10cm2) and of longer duration (a



median of 24 weeks duration), and more likely (a quarter of this sample) to have moderate to high levels of exudate.

b) Both Silver and Iodine treatments are comparable from a cost perspective with the exception of:

i. young wounds (less than 12 weeks old) which have cheaper product costs when treated with Iodine, though this difference was not extended to the costs of nursing related to its application nor to the combined product-nursing cost.

ii. large wounds (bigger than 3.6cm2) which, though comparable for product costs, had cheaper nursing costs and combined product-nursing cost when treated with Silver.

c) Wound healing was expedited by the use of multi-layer compression bandaging therapy, resolving wound slough, managing wound exudate, and by intervening when the wound is young and small.

2. That emphasis in Australian Venous Leg Ulcer Management Guidelines would be warranted with respect to:

a) enhancing access to multi-layer compression bandaging therapy when clinically indicated and optimising adherence to its use.

b) dressings and treatments to resolve wound exudate and slough.

c) early intervention before wounds become large and chronic.

3. That further research is required to compare:

a) Silver and Iodine in managing moderate to high levels of wound exudate generally.

b) Silver and Iodine in managing wound exudate in the absence of compression bandaging therapy.

4. That research undertaken to compare these two antimicrobials in the future incorporate measures of wound inflammation to help understand the mechanism through which Silver and Iodine act and to explore whether the antimicrobials would be better suited to particular wounds based on the presence or risk of inflammation.

5. That the role of biofilms in wounds is the focus of additional research to identify:

a) how microbiology measures can reliably sample wound bacteria in the presence of a biofilm.

b) effective antimicrobial and other treatments to resolve critical colonisation and infection in a wound with a biofilm.

6. That to inform the disparity observed for clinician assessment and wound culture results:

a) the inter and intra rater reliability and validity of clinician assessment of signs of critical colonisation and infection is investigated.

b) future trials explore the use of multiple measures of critical colonisation and infection as well as other markers of wound status



including inflammation to assist with the optimisation of therapies to match healing.

7. That future investigations contrast healing outcomes of antimicrobial treatment:

a) with a comparison, no antimicrobial treatment group given the presence of recommended multi-layer compression therapy and best practice wound bed preparation for both groups.

b) for individuals living with Diabetes Mellitus, arterial leg ulcers, and wounds other than lower leg ulcers.



References Angel, D., Sieunarine, K., Abbas, M., & Mwipatayi, B. (2005). The difficult leg ulcer:

A case review illustrating the problems and difficulties associated with treatment. Primary Intention, 13(1), 7-16.

Annells, M., O'Neill, J., & Flowers, C. (2006). Compression Bandaging in Venous Leg Ulcer Care: Community Nurses' Perspectives on Enablers and Constraints. Melbourne: RDNS Helen Macpherson Smith Institute of Community Health.

Annells, M., O'Neill, J., & Flowers, C. (2008). Compression bandaging for venous leg ulcers: the essentialness of a willing patient. Journal of Clinical Nursing, 17(3), 350-359.

Apelqvist, J., & Ragnarson Tennvall, G. (1996). Cavity foot ulcers in diabetic patients: a comparative study of cadexomer iodine ointment and standard treatment. An economic analysis alongside a clinical trial. Acta Dermato-Venereologica, 76(3), 231-235.

Baker, S. R., & Stacey, M. C. (1994). Epidemiology Of Chronic Leg Ulcers In Australia. ANZ Journal of Surgery, 64(4), 258–261.

Bianchi, J. (2001). Cadexomer-iodine in the treatment of venous leg ulcers: what is the evidence? Journal of Wound Care, 10(6), 225-229.

Bill, T. J., Ratliff, C. R., Donovan, A. M., Knox, L. K., Morgan, R. F., & Rodeheaver, G. T. (2001). Quantitative Swab Culture versus Tissue Biopsy: A Comparison in Chronic Wounds. Ostomy / Wound Management, 47(1), 34-37.

Burrell, R. E., Demling, R., Pancani, S., Papp, A., Strand, O., & Voinchet, V. (2003). The role of Acticoat with nanocrystalline silver in the management of burns. Paper presented at the European Burns Association meeting.

Campton-Johnston, S. M., & Wilson, J. A. (2001). Infected Wound Management: Advanced Technologies, Moisture-Retentive Dressings, and Die-Hard Methods. Critical Care Nursing Quarterly, 24(2), 64-77.

Caruso, D. M., Foster, K., N., Hermans, M. H. E., & Rick, C. (2004). Aquacel Ag in the Management of Partial-Thinkness Burns: Results of a Clinical Trial. Journal of Burn Care & Rehabilitation, 25(1), 89-97.

Carville, K., Flowers, C., Newall, N., Karimi, L., Lewin, G., Gliddon, T., et al. (2008a). Predicting covert and overt infection in venous leg ulcers: A randomised controlled trial. Paper presented at the Australian Wound Management Assoc. National Conference.

Carville, K., Flowers, C., Newall, N., Karimi, L., Lewin, G., Gliddon, T., et al. (2008b). Predicting covert and overt infection in venous leg ulcers: A randomised controlled trial Paper presented at the World Union of Wound Healing Conference.

Carville, K., & Lewin, G. (1998). Caring in the community. Primary Intention, 6(2), 54-62.

Charlson, M. E., Pompei, P., Ales, K., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Disease, 40.

Chase, S., Melloni, M., & Savage, A. (1997). A forever healing: The lived experience of venous ulcer disease. Journal of Vascular Nursing, 15(2), 73-78.



Chase, S., Whittemore, R., Crosby, N., Freney, D., Howes, P., & Phillips, T. (2000). Living with chronic venous leg ulcers: A descriptive study of patients’ experiences. Journal of Community Health Nursing, 17(1), 1-13.

Colgan, M. P., Teevan, M., Mc Bride, C., O'Sullivan, L., Moore, D., & Shanik, G. (1996). Cost comparisons in the management of venous leg ulceration Paper presented at the European Conference on Advance in Wound Management.

Cooper, R., & Lawrence, J. C. (1996). The isolation and identification of bacteria from wounds. Journal of Wound Care, 5(7), 335-340.

Cutting, K., & Harding, K. (1994). Criteria for identifying wound infection. Journal of Wound Care, 3, 198-201.

Demling, R., & DeSanti, L. (2002). The rate of re-epithelialization across meshed skin grafts is increased with exposure to silver. Burns, 28, 264-266.

Department of Human Services. (2001). Identifying and planning assistance for home-based adults who are nutritionally at risk: A resource manual. Melbourne: Aged, Community and Mental Health Division.

Douglas, V. (2001). Living with a chronic leg ulcer: An insight into patients’ experiences and feelings. Journal of Wound Care, 10(9), 355-360.

Dowsett, C. (2003). An overview of Acticoat dressing in wound management. British Journal of Nursing, 12(19), S44-S49.

Drosou, A., Falabella, A., & Kirsner, R. S. (2003). Antiseptics on Wounds: An Area of Controversy. Wounds, 15(5), 149-166.

Dunn, K., & Edwards-Jones, V. (2004). The role of Acticoat™ with nanocrystalline silver in the management of burns. Burns, 30(Suppl. 1), S1-S9.

Edwards, R., & Harding, K. G. (2004). Bacteria and wound healing. Current Opinion in Infectious Diseases, 17, 91-96.

Falanga, V. (1997). Iodine-containing pharmaceuticals: a reappraisal. Paper presented at the 6th European Conference on Advances in Wound Management.

Flowers, C., Kapp, S., Lewin, G., Newall, N. C., K. , & Gliddon, T. (2007). The Angior Trial: Community nurse perceptions of wound best practice initiatives. Primary Intention, 15(4), 149-161.

Flowers, C., Newall, N., Kapp, S., Lewin, G., Gliddon, T., Carville, K., et al. (2008). Clinician inter-rater reliability using a medical wound imaging system. Wound Practice and Research, 16(1), 22-31.

Franks, P. J., & Moffat, C. J. (1998). Who suffers most from leg ulceration? Journal of Wound Care, 7(8), 383-385.

Fraser, J. F., Bodman, J., Sturgess, R., Faoagali, J., & Kimble, R. M. (2004). An in vitro study of the anti-microbial efficacy of a 1% silver sulphadiazine and 0.2% chlorhexidine digluconate cream, 1% silver sulphadiazine cream and a silver coated dressing. Burns, 30, 35-41.

Gardner, S., Frantz, R., & Doebbeling, B. (2001). The validity of the clinical signs and symptoms used to identify localized chronic wound infection. Wound Repair and Regeneration, 9, 178-186.

Gardner, S., Frantz, R., Troia, C., Eastman, S., MacDonald, M., Buresh, K., et al. (2001). A tool to assess clinical signs and symptoms of localized infection in chronic wounds: Development and reliability. Ostomy/Wound Management, 47(1), 40-47.

Hansson, C. (1998). The effects of cadexomer iodine paste in the treatment of venous leg ulcers compared with hydrocolloid dressing and paraffin gauze dressing. International Journal of Dermatology, 37(5), 390-396.



Hillstrom, L. (1988). Iodosorb compared to standard treatment in chronic venous leg ulcers--a multicenter study. Acta Chirurgica Scandinavica - Supplementum, 544, 53-56.

Hokanson, J., Hayward, P., Carney, D., Phillips, L., & Robson, M. (1991). A mathematical model for the analysis of experimental wound healing data. Wounds, 3, 213-220.

Holder, I. A., Durkee, P., Supp, A. P., & Boyce, S. T. (2003). Assessment of a silver-coated barrier dressing for potential use with skin grafts on excised burns. Burns, 29, 445-448.

Holloway, G. A. J., Johansen, K. H., Barnes, R. W., & Pierce, G. E. (1989). Multicenter trial of cadexomer iodine to treat venous stasis ulcer. Western Journal of Medicine, 151(1), 35-38.

Husband, L. (2001). Venous ulceration: The pattern of pain and the paradox. Clinical Effectiveness in Nursing, 5(1), 35-40.

Hyde, C., Ward, B., Horsfall, J., & Winder, G. (1999). Older women’s experience of living with chronic leg ulceration. International Journal of Nursing Practice, 5(4), 189-198.

Innes, M. E., Umraw, N., Fish, J. S., Gomez, M., & Cartotto, R. C. (2001). The use of silver coated dressings on donor site wounds: a prospective, controlled matched pair study. Burns, 27(621-627).

James, G. A., Swogger, E., Wolcott, R., deLancey Pulcini, E., Secor, P., Sestrich, J., et al. (2008). Biofilms in chronic wounds. Wound Repair & Regeneration, 16, 37-44.

Jones, V., & Milton, T. (2000). When and how to use iodine dressings. Nursing Times, 96(45), 2-3.

Jorgensen, B., Price, P., Andersen, K. E., et al. (2005). The silver-releasing foam dressing, Contreet Foam, promotes faster healing of critically colonised venous leg ulcers: a randomised, controlled trial. International Wound Journal, 2(1), 64-73.

Kapp, S., Flowers, C., Karimi, L., & Gliddon, T. (2008). Evaluation of best practice education & subsidised wound products on the healing of leg ulcers Paper presented at the Australian Wound Management Assoc. National Conference.

Kapp, S., Flowers, C., Newall, N., Karimi, L., Lewin, G., Gliddon, T., et al. (2008a). The Angior Initiative. Results of a randomised controlled trial comparing the effectiveness of two antimicrobials. Paper presented at the Australian Wound Management Assoc. National Conference.

Kapp, S., Flowers, C., Newall, N., Karimi, L., Lewin, G., Gliddon, T., et al. (2008b). The Angior Initiative: Results of a randomised controlled trial comparing the effectiveness of two antimicrobials. Paper presented at the World Union of Wound Healing Conference.

Kapp, S., & Nunn, R. (2005). A profile of RDNS clients receiving wound care. Melbourne: RDNS Helen Macpherson Smith Institute of Community Health

Kapp, S., & Sayers, V. (2008). Preventing venous leg ulcer recurrence: a review. Wound Practice and Research, 16(2), 38-47.

Kirsner, R. S., Orsted, H., & Wright, J. B. (2001). Matrix Metalloproteinases in Normal and Impaired Wound Healing: A Potential Role of Nanocrystalline Silver. Wounds, 13(Supplement C), 4-14.

Klasen, H. J. (2000). A historical review of the use of silver in the treatment of burns. Burns, 26, 117-138.



Krasner, D. (1997). Painful venous ulcers: Themes and stories about living with pain and suffering. Journal of Wound, Ostomy and Continence Nursing, 25(3), 158-168.

Lachin, J. M. (2000). Statistical Considerations in the Intent-to-Treat Principle. Controlled Clinical Trials, 21(3), 167-189.

Lam, P. K., Chan, E. S. Y., Ho, W. S., & Liew, C. T. (2004). In vitro cytotoxicity testing of a nanocrystalline silver dressing (Acticoat) on cultured keratinocytes. British Journal of Biomedical Science, 61(3), 125-127.

Lamme, E. N., Gustafsson, T. O., & Middelkoop, E. (1998). Cadexomer-iodine ointment shows stimulation of epidermal regeneration in experimental full-thickness wounds. Archives of Dermatological Research, 290(1-2), 18-24.

Lansdown, A. B. G. (2002). Silver I: its antibacterial properties and mechanism of action. Journal of Wound Care, 11(4), 125-130.

Laudanska, H., & Gustavson, B. (1988). In-patient treatment of chronic varicose venous ulcers. A randomized trial of cadexomer iodine versus standard dressings. Journal of International Medical Research, 16(6), 428-435.

Margolis, D. J., Allen-Taylor, L., Hoffstad, O., & Berlin, J. A. (2004). The accuracy of venous leg ulcer prognostic models in a wound care system. Wound Repair Regen, 12(2), 163-168.

Marieb, E. (2001). Human Anatomy and Physiology (5th ed.). San Franscico Benjamin Cummings.

McCollum, C. N., Ellision, D. A., Groarke, L., Fielden, S., Connoly, M., Franks, P. J., et al. (1997). Randomised trial comparing Profore and the original four layer bandage. . Paper presented at the Conference of the European Wound Management Association.

Meaume, S., Vallet, D., Morere, M. N., & Teot, L. (2005). Evaluation of a silver-releasing hydroalginate dressing in chronic wounds with signs of local infection. Journal of Wound Care, 14, 411-419.

Mertz, P. M., Davis, S. C., Brewer, L. D., & Franzen, L. (1994). Can antimicrobials be effective without impairing wound healing? The evaluation of a cadexomer iodine ointment. Wounds: A Compendium of Clinical Research and Practice, 6(6), 184-193.

Moberg, S., Hoffman, L., Grennert, M. L., & Holst, A. (1983). A randomized trial of cadexomer iodine in decubitus ulcers. Journal of the American Geriatrics Society, 31(8), 462-465.

Moffatt, C. J. (2002). Four-Layer Bandaging: From Concept to Practice. Lower Extremity Wounds, 1(1), 13-26.

Moffatt, C. J., McCullagh, L., O'Connor, T., Doherty, D. C., Hourican, C., Stevens, J., et al. (2003). Randomized trial of four-layer and two-layer bandage systems in the management of chronic venous ulceration Wound Repair & Regeneration, 11(3), 166-171

Moore, K., Thomas, A., & Harding, K. G. (1997). Iodine Released from the Wound Dressing Iodosord Modulates the Secretion of Cytokines by Human Macrophages Responding to Bacterial Lipopolysaccharide. Int. J. Biochem. Cell Biol., 29(1), 163-171.

Moss, C., Taylor, A. E., & Shuster, S. (1987). Comparison of cadexomer iodine and dextranomer for chronic venous ulcers. Clinical & Experimental Dermatology, 12(6), 413-418.

Munter, K. C., Beele, H., Crespi, A., Gronchenig, E., Basse, P., Alikadic, N., et al. (2006). Effect of a sustained silver-releasing dressing on ulcers with delayed healing: the CONTOP study. Journal of Wound Care, 15(5), 199-206.



Neil, J. A., & Munjas, B. A. (2000). Living with a chronic wound: The voice of sufferers. Ostomy Wound Management, 46(5), 28-38.

Nelson, E. A., Bell-Syer, S. E., & Cullum, N. A. (2000). Compression for preventing recurrence of venous ulcers. Cochrane Database of Systematic Reviews(4), CD002303.

Newall, N., Miller, C., Lewin, G., Kapp, S., Gliddon, T., Carville, K., et al. (2009). Nurses' Experiences of Participating in a Randomised Controlled Trial in the Community. Wound Practice and Research, 17(1), 24-34.

O'Meara, S., Cullum, N., & Nelson, E. A. (2009). Compression for venous leg ulcers: Cochrane Database of Systematic Reviews.

O'Meara, S. M., Cullum, N. A., Majid, M., & Sheldon, T. A. (2001). Systematic review of antimicrobial agents used for chronic wounds.[see comment]. British Journal of Surgery, 88(1), 4-21.

Pallant, J. F. (2007). SPSS survival manual: A step by step guide to data analysis using SPSS. NSW: Allen & Unwin.

Partsch, H., Damstra, R. J., Tazelaar, D. J., Schuller-Petrovic, S., Velders, A. J., de Rooij, M. J., et al. (2001). Multicentre, randomised controlled trial of four-layer bandaging versus short-stretch bandaging in the treatment of venous leg ulcers. VASA, 30(2), 108-113.

Percival, S. L., Bowler, P., & Dolman, J. (2007). Antimicrobial activity of silver-containing dressings on wound microorganisms using an in vitro biolfilm model. International Wound Journal, 4(2), 186-191.

Percival, S. L., Bowler, P., & Woods, E. J. (2008). Assessing the effect of an antimicrobial wound dressing on biofilms. Wound Repair & Regeneration, 16, 52-57.

Phillips, T. J., Machado, F., Trout, R., Porter, J., Olin, J., & Falanga, V. (2000). Prognostic indicators in venous ulcers. Journal of American Academy of Dermatology, 43(4), 627-630.

Polignano, R., Bonadeo, P., Gasbarro, S., & Allegra, C. (2004). A randomised controlled study of four-layer compression versus Unna's Boot for venous ulcers. . Journal of Wound Care, 13(1), 21-24.

Posnett, J. (2006). Making cost effectiveness the basis of product selection. . Journal of Wound Care, 15(1), S14.

Principles of best practice: Wound infection in clinical practice. An international consensus (2008). London: MEP Ltd. Available from www.mepltd.co.uk

Rich, A., & McLachlan, L. (2003). How living with a leg ulcer affects peoples’ daily life: A nurse-led study. Journal of Wound Care, 12(2), 51-54.

Robson, M. C., Hill, D. P., Wooske, M. E., & Steed, D. L. (2008). Wound Healing Trajectories as Predictors of Effectiveness of Therapeutic Agents. Arch Surg, 135, 773-777.

Russell, A. D., & Hugo, W. B. (1994). Antimicorbial activity and action of silver. Progress in Medical Chemistry, 31, 351-371.

Santamaria, N., Austin, D., & Clayton, L. (2002). Multi-site trial and evaluation of the Alfred/Medseed Wound Imaging System prototype. Primary Intention, 10(3), 119-124.

Santamaria, N., Carville, K., Ellis, I., & Prentice, J. (2004). The effectiveness of digital imaging and remote wound consultation on healing rates in chronic lower leg ulcers in the Kimberley region of Western Australia. Primary Intention, 12(2), 62-70.

Santamaria, N., & Clayton, L. (2000). The Development of the Alfred/ Medseed Wound Imaging System. Collegian, 7(4), 14-18.



Schultz, G., Barillo, D., Mozingo, D., & Chin, G. (2004). Wound bed preparation and a brief history of TIME. International Wound Journal, 1(1), 19-32.

Schultz, G., Sibbald, R. G., Falanga, V., Ayello, E., Dowsett, C., Harding, K., et al. (2003). Wound bed preparation: A systematic approach to wound management. Wound Repair and Regeneration, 11(2), S1-S28.

Sibbald, G. R., Contreras-Ruiz, J., Coutts, P., Fierheller, M., Rothman, A., & Woo, K. (2007). Bacteriology, Inflammation, and Healing: A Study of Nanocrystalline Silver Dressings in Chronic Venous Leg Ulcers. Advances in Skin & Wound Care, 20(10), 549-558.

Sibbald, R. G., Browne, A. C., Coutts, P., & Queen, D. (2001). Screening Evaluation of an Ionized Nanocrystalline Silver Dressing in Chronic Wound Care. Ostomy Wound Management, 47(10), 38-43.

Sibbald, R. G., Williamson, D., Orsted, H., Campbell, K., Deast, D., Krasner, D., et al. (2000). Preparing the wound bed - debridement, bacterial balance, and moisture balance. Ostomy/Wound Management, 46(11), 14-35.

Singer, J. B., & Willet, J. B. (2003). Applied longitudinal data analysis : modelling change and event occurrence. New York: Oxford University Press.

Skog, E., Arnesjo, B., Troeng, T., Gjores, J. E., Bergljung, L., Gundersen, J., et al. (1983). A randomized trial comparing cadexomer iodine and standard treatment in the out-patient management of chronic venous ulcers. British Journal of Dermatology, 109, 77-83.

Steed, D. L., Hill, D. P., Woodske, M. E., Payne, W. G., & Robson, M. C. (2006). Wound-healing trajectories as outcome measures of venous stasis ulcer treatment. International Wound Journal, 3(1), 40-47.

Sundberg, J., & Meller, R. (1997). A Retrospective Review of the Use of Cadexomer Iodine in the Treatment of Chronic Wounds. Wounds, 9(3), 68-86.

Tabachnick, B. G., & Fidell, L. S. (2007). Using Multivariate Statistics (5th ed.). Boston: Allyn and Bacon.

Tarvainen, K. (1988). Cadexomer iodine (Iodosorb) compared with dextranomer (Debrisan) in the treatment of chronic leg ulcers. Acta Chirurgica Scandinavica - Supplementum, 544, 57-59.

Tennvall, G. R., & Hjelmgren, J. (2005). Annual costs of treatment for venous leg ulcers in Sweden and the United Kingdom. Wound Repair & Regeneration, 13(1), 13-18.

Thomas, S., & McCubbin, P. (2003). A comparison of the antimicrobial effects of four silver-containing dressings on three organisms. Journal of Wound Care, 12(3), 101-107.

Tredget, E. E., Shankowsky, H. A., Groenveld, A., & Burrell, R. (1998). A Matched-Pair, Randomized Study Evaluating the Efficacy and Safety of Acticoat Silver-Coated Dressing for the Treatment of Burn Wounds. Journal of Burn Care & Rehabilitation, 19(6), 531-537.

Vermeulen, H., Van Hattem, J. M., Storm-Versloot, M. N., & Ubbink, D. T. (2007). Topical Silver for Treating Infected Wounds (Review). The Cochrane Database of Systematic Reviews(1).

White, R. (2001). An historical overview of the use of silver in modern wound management. British Journal of Nursing, 15(silver supplement), 3-8.

Wright, J. B., Lam, K., & Burrell, R. E. (1998). Wound management in an era of increasing bacterial antibiotic resistance: A role for topical silver treatment. American Journal of Infection Control, 26(6), 572-577.



Wright, J. B., Lam, K., Olson, M. E., & Burrell, R. E. (2003). Is Antimicrobial Efficacy Sufficient? A Question Concerning the Benefits of New Dressings. Wounds, 15(5), 133-142.

Wysocki, A. B. (2002). Evaluation and managing open skin wounds: colonization versus infection. AACN Clinical Issues, 13(3), 382-397.

Yin, H. Q., Langford, R., & Burrell, R. E. (1999). Comparative Evaluation of the Antimicrobial Activity of ACTICOAT Antimicrobial Barrier Dressing. Journal of Burn Care & Rehabilitation, 20(3), 195-200.

Zhou, L. H., Nahm, W. K., Badiavas, E., Yufit, T., & Falanga, V. (2002). Slow release iodine preparation and wound healing: in vitro effects consistent with lack of in vivo toxicity in human chronic wounds. British Journal of Dermatology, 146, 365-374.





Appendix A









Appendix B

Criteria used when reviewing the data and safety for this Randomised Controlled Trial.

1. Drop out rate from one study group is four times greater than the dropout rate from the other group.

2. Less than 10 clients recruited per month (averaged over three months). The monthly target recruitment goal is 30 clients.

3. The rate of adverse health events in one group exceeds the other two fold.

4. The wound healing % is significantly greater in one group compared to the other when evaluated using a p value of 0.001 at six months of data collection.

5. The rate of adverse health events differs sharply from normal clinical experiences.

6. Any unexplained mortality or mortality which has been attributed to the antimicrobial treatment during the trial is cause for immediate review and appropriate action dependent on the recommendation of the review.





Appendix C









Appendix D








Documents

Treating bacterial burden in chronic lower leg ulcers · Treating bacterial burden in chronic lower leg ulcers Foreword The Australian Wound Management Association has a major goal