The Knee 16 (2009) 262–264

Contents lists available at ScienceDirect

The Knee

Does intraarticular steroid infiltration increase the rate of infection in subsequenttotal knee replacements?

Aravind Desai a,⁎, Sreekumar Ramankutty a, Tim Board b, Videsh Raut b

a Wrightington Hospital, Hall lane, Appley Bridge, Wigan WN6 9EP, United Kingdomb Consultant orthopaedic surgeon, Wrightington Hospital, Hall lane, Appley Bridge, Wigan WN6 9EP, United Kingdom

⁎ Corresponding author. 297, Lovely lane, WarringtoTel.: +44 01257256304, +44 07885994545; fax: +44 0

E-mail address: desaiaravind@yahoo.co.uk (A. Desai

0968-0160/$ – see front matter © 2008 Elsevier B.V. Adoi:10.1016/j.knee.2008.12.002

a b s t r a c t

a r t i c l e i n f o

Article history:

Steroid injection into the a Received 1 April 2008Received in revised form 7 November 2008Accepted 1 December 2008

Keywords:Intra-articular steroid injectionTotal knee replacement surgeryInfection

rthritic joint is a well-known treatment. Its efficacy is well documented. Anincrease in the incidence of infection secondary to steroid injection has been reported in recent literature.Based on the current literature we carried out a retrospective study to evaluate the incidence of infection inprimary total knee arthroplasty as a result of previous steroid infiltration into the knee joint.In our study, 440 patients underwent total knee replacement between 1997 and 2005. Only 90 patients hadintraarticular steroid injection prior to surgery, of which 45 patients had injection within 1 year prior tosurgery. A matched cohort of 180 patients who had total knee replacement without steroid injection wasused as control group to compare the infection rate. All patients had at least 1 year follow up. Two cases ofsuperficial infection were noted in the study group and five cases of superficial infection in the control group.No cases of deep infection were noted in either group. Statistical analysis showed no significant difference inincidence of infection in either group (P value 1.0).This study showed no increase in the incidence of infection in patients with total knee arthroplasty withprior steroid injection.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Steroid infiltration is a widely practiced procedure by rheumatol-ogists, general practitioners and orthopaedic surgeons in the manage-ment of pain in arthritic joints [1,2]. This may be for conservativemanagement of mild osteoarthritis or inflammatory arthritis, tem-porary management of pain in patients with significant medical co-morbidities, whichmake surgical intervention difficult, or for delayingsurgery in young patients [3].

A recent article by Papavasiliou et al. [4] questioned the usefulnessof the procedure and suggested that there was a high risk of infectionin subsequent total knee replacement (TKR). This article generatedmuch subsequent debate [5–7] questioning the validity of the study.

In order to address this issue further, we attempted to collate alarger patient group and compare the incidence of infection inpatients who had an intra-articular steroid infiltration prior to a kneereplacement, with those who had not.

2. Materials and methods

All total knee replacements performed by the senior author between1997 and 2005 were entered on a prospective database. There were 440

n, WA5 0AF, United Kingdom.1257256591.).

ll rights reserved.

patients who had total knee replacement in this period with a minimum1 year follow-up. Patients who had an injection prior to undergoing kneereplacement were identified by analysis of the notes. Exclusion criteriawere a history of malignancy, immuno-suppressive drugs, sero-negativeinflammatory arthropathy or a previous infection in the ipsilateral knee.Patients with diabetes mellitus and with history of smoking were notexcluded from the study. Ninety knees in 80 patients had undergoneinjection prior to knee replacement, after applying the exclusion criteria,which formed the study group. Matched cohort of 180 knees (170patients)were identified frompatientswho hadno injectionprior toTKRsurgery. The matching was done on the basis of age, sex and year ofoperation. Age selection was done by 10 yearly intervals and year ofoperation at three yearly intervals. This control groupwas established byapplying the same exclusion criteria adopted for the study group. Eachknee in the study group was matched with two knees from the non-injected group. The matching was undertaken prior to any evaluation ofpatient notes with respect to the presence of infection. The 90 knees inthe control group comprised of 26 males and 54 females. The primarydiagnosiswas osteoarthritis in 58patients and rheumatoid arthritis in 22.There were 42 right knees and 48 left knees. The mean age was 68 years(49–87). Themean followupwas33months (range1 to6 years). Only 45patients had a knee replacement within 12 months after the injection.

The control group comprised of 74 males and 96 females. Theprimary diagnosis was osteoarthritis in 133 and rheumatoid arthritis in37. There were 82 right knees and 98 left knees. The mean age was72 years (51–88). Themean followupwas 48months (range 1 to 6 years).

Table 1Details of patients with superficial wound problems

Case Age Sex Diagnosis No. of injections Date of last Injection Date of surgery Time gap between last injection and surgery (MONTHS) Out come Place of Injection

1 67 M O.A. 3 07/03/2004 01/09/2005 18 Sup.Inf O.P.D2 73 F R.A. 4 30/08/2001 01/04/2003 18 Sup.Inf O.P.D3 76 F O.A. Nil N/A 24/01/2003 N/A Sup.Inf N/A4 67 M O.A. Nil N/A 10/09/2003 N/A Sup.Inf N/A5 69 M O.A. Nil N/A 19/06/2002 N/A Sup.Inf N/A6 80 F O.A. Nil N/A 08/05/2003 N/A Sup.Inf N/A7 84 F O.A. Nil N/A 09/02/2004 N/A Sup.Inf N/A

263A. Desai et al. / The Knee 16 (2009) 262–264

All data was analysed and assessed which included review of thecase notes including haematology and microbiology reports, imagingstudies, operations notes and follow up data.

All the total knee replacements in the study were cemented PFC-Sigma (Depuy,Warsaw, Indiana, U.S.A.) performed by the senior author.

3. Injection

Intra articular steroid infiltrationwas done in the operating theatreas a day case procedure. The affected knee was exposed and preparedwith chlorhexidine solution. Under strict aseptic precautions, the kneewas infiltrated with Depomedrone 2 ml (Depomedrone 40 mg/ml,Pharmacia, Surrey, U.K.) and Chirocaine (5 mg/ml, Abbott, Maiden-head, U.K.). The patient was discharged later the same day. After theinjection, the patient was reviewed in outpatients at six weeks and theresponse to the steroid infiltration was assessed.

Some of the earlier patients (30 knees) underwent injection as anoutpatient procedure. This was performed in the outpatient treatmentroom under strict aseptic precautions.

4. Total knee replacement

The knee replacementwas done using amidvastus approach undertourniquet. All patients underwent cemented PFC knee replacement.Usage of suction was kept to a minimum and was used only duringpulsed lavage irrigation prior to cementation. Prior to closure thewound was irrigated with 0.05% Chlorhexidine. Meticulous haemos-tasis was achieved before closure over drains. The tourniquet wasreleased after wound closure. The patients were given 3 doses ofCefuroxime postoperatively in addition to the loading dose to ensurecover for 24 h after the operation.

5. Infection

Evidence of superficial wound problems/infection and deepinfection was sought in both groups of patients. Superficial woundproblems/infection was defined as cases which had discharge fromsuperficial layers within one month of surgery, patients who receivedantibiotic cover for more than 1 week for delayed wound healing andcases which had positive cultures from superficial layers which settledand did not require further surgery for the knee.

Deep infection was defined as those cases with positive swabcultures or tissue biopsy from the deep tissues, patients who under-

Table 2Clinical and microbiological details of patients with superficial wound problems

Case Age Sex Diagnosis Time between injectionand surgery (MONTHS)

Symptoms and signs

1 67 M O.A. 18 Sup.wound discharge2 73 F R.A. 18 Swelling, redness, discharge3 76 F O.A. N/A Serous discharge, redness,

swelling & pain4 67 M O.A. N/A Sup.Discharge5 69 M O.A. N/A Pain and discharge6 80 F O.A. N/A Serous discharge, redness.7 84 F O.A. N/A Discharge, redness and pain

went exploration and wash out of the wound with positive culturereport and cases which underwent revision surgery for infection.

6. Statistical analysis

Statistical analysis of the results was performed using Stata(Release 9). To assess non-inferiority of injection on risk of infection,a 95% confidence interval (CI) for the relative difference between thepercentages of patients suffering infections in the two groups wereobtained using the normal approximation method.

7. Results

There were two patients (2.2%) with a serous discharge constituting superficialwound problem/infection in the study group. Both the patients had received steroidinjection more than 1 year before surgery. Both these patients received injection as anoutpatient procedure. They were investigated and found to be clear of infection andcontinued to be infection free during follow up (Table 1).

In the control group there were five knees (2.7%) that met the criteria for superficialwound problem/infection. All patients were treated with appropriate antibiotics withoutany further intervention (Table 2).

All these seven cases have been followed up regularly both clinically andradiologically. At the last follow up none of the patients had any signs of superficial ordeep infection. All the patients were thoroughly investigated in terms of ESR and CRP atregular intervals after the initial treatment with respective antibiotics during their stayin the hospital. After discharge they were followed up in the out patient clinic morefrequently than others, and the ESR and CRP were repeated every 2 weeks, 6 weeks,3 months and 6 months till base line values were obtained. There were no knees withdeep infection in either group.

The relative risk of infection was 0.80 (95% CI 0.16–4.03). Fisher's exact test wasnon-significant (P value 1.0). The conclusion was that whilst there was no significantdifference between the two groups in terms of infection risk in total knee replacement,and the observed risks of infection were similar in the two groups, non-inferiority ofinjection could not be established.

8. Discussion

Steroid injection into the arthritic joints is a well-known treatmentof arthritic joints [8,9].

The incidence of deep infection after total knee arthroplasty isremarkably low with most studies suggesting rates of around 1%although rates between0.5% and 12%have been reported [10]. However,the consequences of deep infection are severe and further surgery isrequired often resulting in two-stage revision surgery. Consequently it isparamount to surgeons to know whether a particular intervention willincrease the risk of infection. Several risk factors for infection followingtotal knee arthroplasty have been reported in the form of diabetes,

Swab result Treatment Outcome Last follow up

S. aureus Cipro+fluclox (4 weeks) No E/O Infection 03/04/2007No growth Fluclox −6 weeks No E/O infection 05/12/2006No growth Fluclox −6 weeks No E/O infection 10/12/2006

S. aureus Fluclox−6 weeks No E/O infection 10/09/2006E. coli Ticoplanim+cipro No E/O infection 23/08/2006No growth Fluclox−4 weeks No E/O infection 23/05/2006No growth Fluclox−2 weeks No E/O infection 12/12/2006

264 A. Desai et al. / The Knee 16 (2009) 262–264

previous knee operations, rheumatoid arthritis, poor nutrition, renalfailure, hypothyroidism and poor local skin condition [11,12].

We had two patients with a serous discharge and signs of superficialwound site infection after a knee replacement following a previoussteroid infiltration. Both these cases had injection in outpatient clinicwith a time interval of 18months between the timeof injection and totalknee replacement surgery. This compared favorably with five patientswith similar superficial wound problems out of 180 knees that had notbeen injected prior to surgery. Statistical analysis showed no increase ininfection compared to those who did not have an injection.

Papavasiliou et al. [4] reported an increase in infection in kneereplacements after steroid injections. They reported on 54 knees thathad received an injection prior to knee replacement and comparedthese to a control group of 90 knees that did not receive injection. Theyreported infection in 12 patients with previous steroid infiltration, ofwhich three were deep infection. In the control group there were 10patients with superficial infection and none with deep infection. Theywere unable to isolate the exact reasons for the high infection rate inthe patientswho had steroid infection. They suggest, as did Kaspar andBeer [13] in their paper on infections in hip replacements after steroidinjections that this could be due to the fact that the steroid agent doesnot fully dissolve and remains trapped in the soft tissues of the knee.They allude to the fact that steroid injections are given with ‘token’aseptic precautions in many centers and feel that this could be one ofthe reasons for the high infection rate in knee replacements followinginjections in general. Their study does not define whether multiplesurgeons did the surgeries nor does it give details of the technique forthe injection. This paper has been debated in the literature over thesmall sample size [6], definition of infection [7] and lack of correlationbetween timing of injection and infection rate [5]. Similar studies byJoshy et al. [14] and Chitre et al. [15] on the effect of intra-articularsteroid on deep infection showed that there was no statisticallysignificant difference between injected and non-injected groups interms of its effect on deep infection.

The main strengths of our study is that it is the largest of its kind,has a well matched cohort of patients with double the number of thetreatment group, longer follow up of all patients, all cases operated bywith standard protocol and no cases lost to follow up.

Our study has the limitations of any retrospective and non-randomised study. Despite having larger number of patients in eithergroup as compared to all other studies, it is still underpowered.

However, we have no clearly established non-inferiority of theinjection in terms of infection rate, as the upper limit of the 95% CI is arelative risk of 4.03, meaning that the infection rate could be up to fourtimes higher in the injected than the non-injected group. Given thelow infection rate, sample sizes around 2000 patients per groupwould be needed to rule out (with 95% CI) a 50% increase in infectionrate amongst those injected, thus suggesting the need for a multi-centre larger study.

We conclude that our low infection rate in the present study couldbe due to our standardized technique of injection of the knee in theoperation theatre under strict aseptic conditions, minimum usage ofsuction, chlorehexidine wash of the wound during surgery and 3doses of i.v. antibiotics postoperatively after the initial loading dose[16,17]. Given our results, we see no increased incidence of infection inknee replacements after a prior steroid infiltration and would notadvise discontinuing their use.

Acknowledgment

We acknowledge Dr Chris J Sutton PhD, Senior lecturer in statistics,UCLANUniversity, Preston, UK, for his helpwith the statistical analysis.

References

[1] Hochberg MC, Altman RD, Brandt KD, Clark BM, Dieppe PA, Griffin MR. Guidelinesfor the medical management of osteoarthritis: Part II. Osteoarthritis of the knee.Arthritis Rheum 1995;38:1541–6.

[2] ScottDL.Guidelines for thediagnosis, investigationandmanagementofosteoarthritisof the hip and knee. J R Coll Physicians Lond 1993;27:391–6.

[3] Creamer P. Intra-articular corticosteroid treatment in Osteoarthritis. Curr OpinRheumatol 1993;11:417–21.

[4] Papavasiliou AV, Isaac DL, Marimuthu, Skyrme A, Armitage A. Infection in kneereplacements after previous injection of intra articular steroid. J Bone Jt Surg [Br]2006;88–B:321–3.

[5] Vashista GN. Infection in knee replacements after previous injection of intrarticularsteroid. J Bone Jt Surg [Br] Mar 2007;89(3):422–3 [author reply 423. No abstractavailable].

[6] Dodd LE. Infection in knee replacements after previous injection of intrarticularsteroid. J Bone Jt Surg [Br] Mar 2007;89(3):422–3 [author reply 422. No abstractavailable].

[7] Little NJ, Chipperfield A, Ricketts DM. Infection in knee replacements afterprevious injection of intrarticular steroid. J Bone Jt Surg [Br] Mar 2007;89(3):422–3 [author reply 422. No abstract available].

[8] Bellamy N, Campbell J, Robinson V, Gee T, Bourne R, Wells G. Intraarticularcorticosteroid for treatment of osteoarthritis of the knee. Cochrane Database SystRev 2005(2), doi:10.1002/14651858.CD005328.pub2 [Art. No.: CD005328].

[9] Kullenberg B, Runesson R, Tuvhag R, Olsson C, Resch S. Intraarticular corticosteroidinjection: pain relief in osteoarthritis of the hip? J Rheumatol 2004;31(11):2265–8.

[10] Cooper C, Kirwan JR. Risks of corticosteroid therapy. Clin Rheumatol 1990;4:305–32.

[11] Wilson MG, Kelly K, Thornhill TS. Infection as a complication of total kneearthroplasty. Risk factors and treatment in sixty-seven cases. J Bone Jt Surg [Am]1990;72(6):878–83.

[12] Yang K, Yeo SJ, Lee BP, Lo NN. Total knee arthroplasty in diabetic patients: a study of109 consecutive cases. J Arthroplast 2001;16(1):1–26.

[13] Kaspar S, De V de Beer J. Infection in hip arthroplasty after previous injection ofsteroid. J Bone Jt Surg [Br] 2005;87-B:454–7.

[14] Joshy S, Thomas B, Gogi N, Modi A, Singh BK. Effect of intra-articular steroids ondeep infections following total knee arthroplasty. Int Orthop 2006;30:91–3.

[15] Chitre AR, Fehily MJ, Bamford DJ. Total hip replacement after intra-articularinjection of local anesthetic and steroid. J Bone Jt Surg [Br] 2007;89(2):166–8.

[16] Sreekumar R, Venkiteswaran R, Raut VV. Infection in primary hip arthroplasty afterprevious steroid infiltration. Int Orthop 2007;31:125–8.

[17] Furnes O, Havelin LI, Espehaug B. The Norwegian arthroplasty register; 1997–2004.