ORIGINAL ARTICLE

The prevalence of nutrition impact symptoms and theirrelationship to quality of life and clinical outcomes in medicaloncology patients

H. Tong & E. Isenring & P. Yates

Received: 13 December 2007 /Accepted: 22 May 2008 / Published online: 13 June 2008# Springer-Verlag 2008

AbstractGoals of work The aims of this secondary analysis were todetermine the prevalence of nutrition impact symptoms inmedical oncology patients at 1, 6, and 12 months aftercommencement of chemotherapy and to investigate therelationship of these symptoms to quality of life (QoL) andperformance status.Materials and methods A prospective longitudinal surveywas conducted in 219 medical oncology patients who hadcommenced chemotherapy in the past month. The Rotter-dam Symptom Assessment scale assessed the number anddistress level of symptoms. The association betweensymptoms and global QoL and performance status asmeasured by the Life Satisfaction Scale was investigated.Main results Symptom prevalence as determined by theproportion of patients experiencing at least one nutritionimpact symptom was 79% and 72% at 1 and 6 months afterstarting chemotherapy. Even at 12 months, symptomprevalence was 46%. The most common symptomsincluded dry mouth, nausea, and constipation with the most

distressing symptoms reported as dry mouth, diarrhea, andstomach pain. A higher number of symptoms was associ-ated with lower QoL (T1: r=−0.35, n=217, P<0.05; T2:r=−0.406, n=194, P<0.001; T3: r=−0.353, n=157, P<0.001). Patients experiencing more symptoms were morelikely to have lower performance status at T2 and T3 (T2:n=189, P=0.019; T3: n=143, P=0.003).Conclusion Nutrition impact symptoms were commonlyexperienced, even 12 months following commencement ofchemotherapy, and were associated with poorer QoL andperformance status. This highlights the importance of earlyidentification and management of nutrition impact symp-toms with adequate follow-up in order to provide optimalcare for people with cancer.

Keywords Nutrition . Oncology . Cancer . Chemotherapy .

Symptoms . Patient-generated subjective global assessment

Introduction

Disease-related malnutrition occurs frequently in patientswith cancer, with prevalence ranging from 50% to 80%depending on the tools used and population studied [1, 13,22]. Patients with pancreatic, gastrointestinal, or head-and-neck cancer are at higher risk of malnutrition while breastcancer patients have comparatively lower risk [13, 19].Decreased dietary intake, cancer cachexia (characterizedmainly by weight loss and muscle wasting), and nutritionimpact symptoms may all contribute to cancer-relatedmalnutrition [16]. Nutrition impact symptoms are definedas those which are likely to negatively impact on nutritionalstatus and increase the risk of malnutrition [18]. Thesesymptoms may result from the tumor itself and/or anyresulting side effects of treatment, leading to a reduced

Support Care Cancer (2009) 17:83–90DOI 10.1007/s00520-008-0472-7

HT was the main author of the manuscript and assisted in statisticalanalysis and interpretation. EI supervised the project and assisted instatistical analysis, interpretation, and writing the manuscript. PYinitiated the study, supervised the project, and assisted in writing themanuscript.

H. TongQueensland University of Technology,Brisbane, Queensland 4059, Australia

E. Isenring (*) : P. YatesInstitute of Biomedical Health Innovation and School of PublicHealth, Queensland University of Technology,60 Musk Ave, Kelvin Grove,Brisbane, Queensland 4059, Australiae-mail: e.isenring@qut.edu.au

dietary intake compromising nutritional status and hencequality of life (QoL) [7, 16].

A number of studies have investigated the interrelation-ship between cancer and malnutrition [1, 13, 16] anddemonstrated the benefits of early and intensive nutritionsupport on improving nutritional status and QoL in cancerpatients [9, 21]. However, studies examining the relation-ship between the prevalence of nutrition impact symptomsand QoL and other clinical outcomes are limited [25]. Inaddition, the majority of published studies have docu-mented the symptom prevalence in patients at an advancedstage of disease [11, 17] but few studies have exploredthese in an early treatment phase of the cancer population.There is also a paucity of data available on longitudinalfollow-up of these symptoms after completing treatment. Inorder to provide optimal care for cancer patients, a betterunderstanding of these symptoms and their relationshipwith other clinical outcomes is essential.

The aims of this study were to: (1) identify and describethe prevalence and distress level of nutrition impactsymptoms in ambulatory medical oncology patients atthree time points, (1, 6, and 12 months after commence-ment of chemotherapy) and (2) investigate the relationshipof these nutrition impact symptoms to QoL and perfor-mance status.

Materials and methods

Subjects

An observational longitudinal study was conducted inpatients aged greater than 18 years attending two majoroncology units in Brisbane, Australia. The inclusion of bothpublic and private clinics ensured that patients from allareas of oncology and hematology, with a wide range ofsolid tumor and hematological malignancies, wererecruited. Other criteria for inclusion in the study were:outpatients with a histological proven diagnosis of cancer, adiagnosis within the last 6 months, commenced activetreatment at the center within the past 4 weeks, and beingable to speak and read English. Patients diagnosed withrecurrent disease, unstable condition, and/or those who hadcommenced chemotherapy treatment at another unit–ser-vice were excluded from the study. The sample wasobtained over 6 months of consecutive recruitment.Successive patients who met the inclusion criteria wereasked to participate in the study. Data were collected atthree time points: 1 (T1), 6 (T2), and 12 (T3) monthsfollowing commencement of chemotherapy. This projectreceived ethics approval and all participants providedinformed written consent. Data collection concluded in2000 and primary data analysis occurred in 2001. This

paper presents a secondary analysis of the dataset focusingon nutrition-related parameters.

Data collection

Demographic data including age, gender, and clinicalinformation on performance status were obtained from theself-report questionnaire. Information on diagnosis, typeand stage of cancer, and chemotherapy regimen andduration were obtained from the patient’s oncologist andmedical records using a standard data collection form. Asthis was a heterogenous population receiving treatment formixed solid tumors and hematological malignancies, therewere many different tailored chemotherapy regimens (>20).Based on known side effect profiles and in consultationwith specialist oncology clinicians, these chemotherapyregimens were classified according to their toxicity level.Chemotherapy toxicity was thus rated as low, moderate, orhigh by two independent clinicians.

The Rotterdam Symptom Checklist (RSCL) [24] wasused to measure both the prevalence and distress levelassociated with specific symptoms that may occur in cancerpatients. The original 34-item checklist was modified byadding an additional nine symptoms relevant to thepurposes of the primary study; however, these additionalsymptoms did not influence the nutrition impact symptomsof interest to this secondary analysis. The checklist includessymptoms which were categorized into subscales, includinggastrointestinal toxicities and discomforts and eating diffi-culties. Subjects were asked to indicate whether or not theyhad experienced each symptom during the past week and torate how much they had been bothered by them (1=not atall to 4=very much). A high overall global score forsymptoms subscales indicates high symptom prevalence.Life Satisfaction Scale was used to measure global QoL [4].The scale contains seven items that reflect variousdimensions of life satisfaction. Patients were asked to ratehow satisfied they were with each item on a five-pointLikert scale (1=very dissatisfied to 5=very satisfied), witha higher score reflecting a better global QoL.

Initial interviews were conducted in the Oncology andHematology clinics by two trained research officers whowere registered nurses experienced in interviewing cancerpatients. Subsequent follow-up interviews were conductedat the oncology clinics if patients had follow-up appoint-ments at the appropriate time or were conducted in thepatient’s home or by phone and mail if patients residedoutside the Brisbane metropolitan area. It took approxi-mately 45 min for the participants to complete the self-report questionnaire which incorporated demographic,medical, and use of support services information as wellas information on QoL and symptoms. Interviewers werepresent and provided assistance when necessary.

84 Support Care Cancer (2009) 17:83–90

PG-SGA

Patient-Generated Subjective Global Assessment (PG-SGA) [18] is a nutrition assessment tool specificallydeveloped for use in the cancer population and has beencommonly used in nutrition studies undertaken in oncologypatients [2, 22]. Unfortunately, as this study was notdesigned as a nutrition trial, a comprehensive nutritionassessment using the PG-SGA was not undertaken. How-ever, data were available so that the nutrition impactsymptom section of the PG-SGA could be scored and usedin this secondary analysis to investigate the relationshipbetween nutrition-related symptoms and QoL and perfor-mance status. Ten nutrition-related symptoms containedwithin the PG-SGA were identified from the RSCL andthese symptoms were given a PG-SGA score (one to threepoints) based on the relative impact these symptoms haveon nutritional status [18] (Table 1). PG-SGA scores areassigned to these symptoms based on the likelihood ofnegatively impacting on nutritional status (i.e., 0=minimalimpact on nutritional risk or status, 1=mild, 2=moderate,3=potentially severe) [18]. Ottery [18] recommends that ascore ≥9 of the full PG-SGA assessment reflects a criticalneed for symptom management and/or nutrition interven-tion. Other studies have confirmed that a PG-SGA score ofgreater ≥9 is associated with a malnourished status and,thus, requires symptom management and nutrition inter-vention [2, 8].

Statistical analysis

Data were entered into SPSS for Windows version 14.0(Statistical Package for Social Sciences, IL, USA). Nor-mality was confirmed using the Kolmogorov–Smirnov testand results were presented as means±standard deviation.Global scores for symptom subscales and life satisfactionscores at T1 were normally distributed. Independent t testswere used to compare continuous descriptive variables(e.g., age and global symptom subscales) for thoseparticipants who did and did not complete the study.Nonparametric tests were used for continuous variableswhich were not normally distributed. A range of univariateand bivariate statistical procedures were performed. Anal-ysis of variance was used to compare global scores forsymptom subscales for each diagnosis, stage of cancer, andchemotherapy toxicity category. Kruskal–Wallis analyseswere used for nonparametric data comparison (performancestatus and PG-SGA scores). For correlation analysis,Spearman and Pearson tests were used to examine theassociation between global scores for symptom subscalesand life satisfaction scores and the association between PG-SGA scores and life satisfaction scores. P value of <0.05was accepted as statistically significant.

Results

Subject characteristics

A sample of 257 patients was eligible for the study and 219agreed to participate: 36% (79) were male and 64% (140)were female. The mean age was 53±13 years. Patientcharacteristics are shown in Table 2. At T1, 79% of patientswere diagnosed with a solid tumor and 21% werediagnosed with a hematological-related cancer. Mostpatients had stage II (28%) or stage III (25%) cancer atT1. Over 90% patients rated their level of performancestatus as either fully active or ambulatory at both T1 andT2. At 6 months, there were 197 patients remaining in thestudy while at 12 months there were 159 patients.

At both T2 and T3, there were no differences in agebetween those who completed the study compared withthose who were lost to attrition (52.3±13 vs 56.3±14 years;P=0.177) and T3 (52.3±13.6 vs 53.7±12 years; P=0.483).Due to the lack of power, statistical tests could not be usedto determine whether or not there were differences in cancertype and stage between completers and noncompleters butit does appear that there were more patients with stage IVcancer who were lost to attrition (16 vs eight) across bothT2 and T3. However, there were no differences in symptomscales for completers vs noncompleters at T2 (39.7±7.4 vs41.7±9.3; P=0.243) and T3 (39.7±40.6±8.7; P=0.434)

Table 1 Comparison of nutrition impact symptoms using PG-SGAand RSCL

Symptoms in PG-SGA Symptoms in the study(RSCL)

Score

No appetite Poor appetite 3Vomiting Vomiting 3Diarrhea Diarrhea 3Paina Stomach pain and/or

low abdominal pain3

Mouth sores Sore mouth or painwhen swallowing

2

Problems swallowing Difficulty swallowing 2Nausea Nausea 1Constipation Constipation 1Dry mouth Dry mouth 1No taste Loss of taste 1Smells bother me / 1Feel full quickly / 1Other, e.g. depression–money–dental

/ 1

PG-SGA Patient-Generated Subjective Global Assessment [14], RSCLRotterdam Symptom Checklist [12]a Pain refers to that which is likely to negatively impact on nutritionalstatus, e.g., stomach or abdominal pain

Support Care Cancer (2009) 17:83–90 85

and PG-SGA score T2 (4.3±3.7 vs 5.3±4; P=0.220) andT3 (4.4±3.8 vs 4.3±3.6; P=0.861).

No statistically significant differences were found betweenprevalence of symptoms and types of cancers except at T2(P=0.006). There was no association between prevalence ofsymptoms and the stage of cancer. At T1, there appeared tobe a clinically relevant difference between chemotherapytoxicity levels and the global score for symptom subscales(mean±SD; low=38.7±7.6, moderate=40.3±7.5, high=42.3±9.1) and PG-SGA score median (low=4, moderate=4, high=6) but neither of these reached statistical signifi-cance (P<0.196 and P<0.405, respectively). While at T2there was also no difference between toxicity and symptom(P=0.546) and PG-SGA scores (P=0.347), there weresignificant differences between the duration of chemothera-py. Patients receiving 6–12 months of treatment had thehighest symptom (P=0.006) and PG-SGA scores (P=0.017).

At T3, there was no difference between toxicity andsymptom score (P=0.243) and PG-SGA score (p=0.162).At T3, there were no associations between duration ofchemotherapy and symptoms (p=0.214) and PG-SGA scores(p=0.188).

At T1, 24 (11%) patients had seen a dietitian while, atT2 and T3, the number decreased to 14 (7%) and ten (6%),respectively. There were 27 (12.3%) and 28 (14.2%)patients with PG-SGA scores ≥9 (indicating a need forsymptom management and nutrition intervention) at T1and T2 and 15 (9.4%) patients with PG-SGA scores ≥9 atT3.

Symptom prevalence and distress levels

The prevalence of nutrition impact symptoms was relativelylower at T3, with 46% patients experiencing one or more

Table 2 Medical oncology patient characteristics

Characteristics Time 1 (n=219) Time 2 (n=197) Time 3 (n=159)

Number Percentage Number Percentage Number Percentage

GenderF 140 64 128 65 108 68M 79 36 69 35 51 32Mean age (year; SD) 53 (±13)Diagnosis categorySolid tumors 173 79 156 79 125 79Head–neck 7 3 5 3 2 1Gastrointestinal–colorectal 47 21 44 22 33 21Breast 63 29 60 31 54 34Respiratory–lung 15 7 11 6 7 4Genitalia–urinary–reproductive 31 14 28 14 23 15Soft tissue–skin–brain–CNS 7 2.5 6 3 5 3Primary unknown 3 3 2 1 1 1Hematological malignancies 46 21 41 21 34 21Stage of tumora

Stage I 19 9 18 12 12 10Stage II 62 28 57 37 55 44Stage III 55 25 53 34 43 34Stage IV 24 11 16 10 8 6Unknown 10 6 9 6 5 4Missing 3 2 3 2 2 2Chemotherapy toxicityb

Low 65 30 33 17 13 8Moderate 141 64 69 35 8 5High 13 6 9 5 2 1N/A 0 0 85 43 136 86Fully active, asymptomatic 80 37 79 40 78 49Ambulatory, light work 128 58 101 51 61 38Immobile, <50% of time 7 3 9 5 4 3Immobile, >50% of time 4 2 1 0.5 0 0

Time 1Within 4 weeks of commencing chemotherapy, Time 2 6 months after commencing study, Time 3 1 year after commencing study, F female,M male, CNS central nervous system, N/A not applicable as not currently receiving chemotherapya Data taken from T1 for patients remaining in the studyb Toxicity of treatment was rated independently by two clinicians.

86 Support Care Cancer (2009) 17:83–90

nutrition impact symptoms compared with 79% at T1 and72% at T2.

As shown in Table 3, the five most commonly reportednutrition impact symptoms at T1 included dry mouth(38%), belching (36%), nausea (35%), bad taste in mouth(28%), and constipation (27%). The five most commonlyreported nutrition impact symptoms at T2 and T3 weresimilar (Table 3). In particular, dry mouth was shown to bethe most prevalent symptom affecting about one third of thecancer patients across all time points.

The most distressing nutrition impact symptoms at T1were dry mouth, belching, nausea, and constipation, withmean score±SD of 1.4±0.7. The most distressing nutritionimpact symptoms across all time points were dry mouth,diarrhea, and stomach pain. Mean scores of nutrition impactsymptom distress were less at T3 (Table 3).

Relationship between symptoms and QoL

Overall mean rating for the life satisfaction scores remainedrelatively consistent between T1 (27.6±4.6) and T2 (27.0±5.2); however, at T3, patients reported a decrease in theirsatisfaction with life (24.5±7.3; p<0.05). A mediumnegative correlation was found between total number ofsymptoms and life satisfaction scores across all time points(T1: r=−0.35, n=217, p<0.05; T2: r=−0.406, n=194,p<0.001; T3: r=−0.353, n=157, p<0.001), with higheramount of symptoms associated with lower levels of QoL.

Similarly, a small to medium negative correlation wasfound between PG-SGA scores (Table 3) and life satisfac-

tion scores across all time points (T1: r=−0.224, n=218,p=0.001; T2: r=−0.350, n=196, p<0.001; T3: r=−0.288,n=157, p<0.001).

Relationship between symptoms and performance status

No significant relationship was found between symptomprevalence using global symptom scores and the level ofperformance status across all time points. However, whenconsidering the relationship between PG-SGA scores andlevel of performance status, a statistically significantdifference was found at T2 and T3, respectively (p=0.019, p=0.003). Mean rank of PG-SGA scores differedby level of performance status, with patients experiencingmore nutrition impact symptoms being more likely to havea lower level of performance status (Kruskal–Wallis test,T2: n=189, p=0.019; T3: n=143, p=0.003).

Discussion

Symptom prevalence

This secondary analysis identified the prevalence ofnutrition impact symptoms in ambulatory medical oncologypatients at three time points over the year commencingchemotherapy. The prevalence of nutrition impact symp-toms was high with 79% and 72% of patients experiencingat least one of these symptoms 1 month followingcommencement of chemotherapy and 6 months later (T1

Table 3 Symptom prevalence and distress level experienced by medical oncology patients

Symptoms Time 1 (n=219) Time 2 (n=197) Time 3 (n=159)

Prevalence Distressa Prevalence Distressa Prevalence Distressa

No (%) Mean (SD) No (%) Mean (SD) No (%) Mean (SD)

Loss of taste 56 (26) 1.3 (±0.7) 36 (18) 1.3 (±0.7) 15 (10) 1.1 (±0.5)Sore throat 40 (18) 1.2 (±0.6) 29 (15) 1.2 (±0.5) 21 (13) 1.1 (±0.4)Difficulty swallowing 16 (7) 1.1 (±0.4) 13 (7) 1.1 (±0.4) 9 (6) 1.1 (±0.4)Sore mouth or pain when swallowing 37 (17) 1.2 (±0.5) 19 (10) 1.2 (±0.5) 8 (5) 1.1 (±0.4)Dry mouth 83 (38) 1.4 (±0.7) 73 (37) 1.5 (±0.8) 54 (34) 1.4 (±0.7)Bad taste in mouth 60 (28) 1.3 (±0.7) 43 (22) 1.3 (±0.7) 21 (13) 1.2 (±0.6)Low abdominal pain 30 (14) 1.2 (±0.5) 30 (15) 1.2 (±0.6) 23 (15) 1.2 (±0.6)Belching 78 (36) 1.4 (±0.7) 50 (26) 1.3 (±0.6) 33 (21) 1.2 (±0.6)Nausea 76 (35) 1.4 (±0.7) 49 (25) 1.3 (±0.6) 33 (21) 1.2 (±0.5)Poor appetite 57 (26) 1.3 (±0.6) 44 (22) 1.3 (±0.8) 26 (16) 1.2 (±0.6)Diarrhea 57 (26) 1.3 (±0.6) 58 (30) 1.4 (±0.7) 34 (22) 1.3 (±0.6)Stomach pain 50 (23) 1.3 (±0.7) 53 (27) 1.3 (±0.7) 35 (22) 1.4 (±0.8)Heartburn 30 (14) 1.2 (±0.5) 28 (14) 1.2 (±0.5) 14 (9) 1.1 (±0.4)Vomiting 18 (8) 1.1 (±0.4) 14 (7) 1.1 (±0.4) 7 (4) 1.1 (±0.3)Indigestion 46 (21) 1.2 (±0.6) 38 (20) 1.2 (±0.5) 20 (13) 1.2 (±0.5)Constipation 59 (27) 1.4 (±0.7) 44 (22) 1.2 (±0.6) 22 (14) 1.2 (±0.6)

a Scores ranged from 1 (not at all distressed) to 4 (very much distressed) according to Rotterdam Symptom Checklist [12]

Support Care Cancer (2009) 17:83–90 87

and T2). While our study findings may be consistent withthose observed by other researchers [6, 11, 12], directcomparisons between the studies require careful interpreta-tion as the majority of the patients in this study had anoverall stage of II or III. For example, the prevalence ofnutrition impact symptoms in this study is slightly higher tothat reported in another study by Khalid et al. [11] inoncology patients with gastrointestinal and lung cancer atcomparatively earlier stages of their disease. Khalid et al.reported that their sample of patients who had not startedchemotherapy or radiotherapy had a nutrition impactsymptom prevalence of 62% [11].

Symptom prevalence and types of cancer

No statistically significant differences were found betweennutrition impact symptom prevalence and types of cancersexcept at 6 months (T2) following commencement oftreatment, where a strong association (p=0.006) was foundbetween symptom prevalence and types of cancers. Ahigher number of nutrition impact symptoms were reportedby patients with head or neck cancer as compared to lungcancer patients. Previous studies have reported that theprevalence of nutrition impact symptoms is higher inpatients with certain types of cancer patients such asgastrointestinal, head or neck, and pancreatic cancerpatients [22]. The association was not demonstrated in ourstudy except at T2, most probably due to the relativelysmall numbers of these types of patients in the studysample.

Symptom prevalence and chemotherapy toxicity

No statistical associations were found between globalsymptom scales nor PG-SGA score and chemotherapytoxicity level despite there appearing to be clinicallyimportant differences, with those rated as receiving a hightoxicity experiencing the greatest number of symptoms.The small number of patients receiving a high toxicityrating over each of the time points may explain the lack ofpower to detect any difference. This lack of association mayalso reflect the low specificity of the toxicity classificationand does not take into account an individual’s response totreatment. At T2, duration of chemotherapy was associatedwith more symptoms. This highlights the need for ongoingmanagement of symptoms throughout treatment.

Symptom prevalence at 12 months

The findings of this study indicate that, even at 12 months(T3) following commencement of chemotherapy, theprevalence of nutrition impact symptoms was still moder-ately high with 46% of patients experiencing nutrition

impact symptoms. Taking into consideration that mostpatients had finished their treatment at this time point, thehigh prevalence of symptoms was of concern. A recentstudy by Olsson et al. [17] found that upper gastrointestinalpatients at 12 months after surgery still suffer fromdiscomforting symptoms such as reflux, diarrhea, andconstipation which can cause decreased dietary intake andhence impact on nutritional status. In view of medicaladvances made in cancer prognosis and treatment, thenumber of cancer survivors has increased substantially overthe past two decades. Numerous studies have beenundertaken with patients in these earlier phases of cancerdiagnosis and treatment; however, studies which follow uppatients during their recovery phase are limited. Thissuggests that more research needs to be done to understandthe nutrition support needs of patients following treatmentcompletion.

Common symptoms

The most commonly experienced nutrition impact symp-toms were dry mouth, belching, nausea, bad taste in mouth,and constipation. Common symptoms were also the mostdistressing symptoms and this is in agreement withprevious studies, where up to 60–80% of the most commongastrointestinal symptoms have been reported as moderate-ly or severely intense [5]. However, there are somedifferences with the most common nutrition impact symp-toms reported in other studies in this field. Khalid et al. [11]reported that the most common symptoms were loss ofappetite, early satiety, and pain among patients withgastrointestinal and lung cancer. These differences arelikely to be attributable to differences in study samples interms of the stage of disease, types of cancer treatment,drug dosage, treatment duration, and frequency [3]. Forexample, patients in our study reported symptoms to be lessdistressing at 12 months after commencing treatment. Atthis time, many patients will have completed treatment andbe experiencing fewer treatment-related side effects.

Associations between symptoms and global QoLand physical function

Importantly, our study has shown a relationship betweenthe presence of nutrition impact symptoms and lowerperformance status and lower levels of life satisfaction.Other studies have reported similar relationships. One studyby Olsson reported that patients’ QoL decreased 12 monthspost upper gastrointestinal surgery with increasing fatigueand reflux syndrome [17]. A literature review conducted byMarin Caro et al. [14] similarly concluded that a low QoLis associated with presence of nutrition-related symptoms.Schuit et al. [23] has also reported that impaired functional

88 Support Care Cancer (2009) 17:83–90

status was associated with increasing nutrition symptomscores. The PG-SGA score was found to be associated withQoL in 60 cancer patients receiving radiotherapy to thegastrointestinal or head-and-neck area [8]. Our study showsthat higher amount of symptoms were associated withlower levels of QoL. This study finding adds to the currentbody of evidence that, as part of the routine nutritionassessment, nutrition-related symptoms as assessed by thePG-SGA tool could be a useful measure in identifyingfactors that may impact on a patient’s performance statusand QoL.

Dietetic consultations

Nutrition problems are common in patients undergoingchemotherapy, with malnutrition prevalence estimates rang-ing from 50% to 80% [1, 13, 22]. A study by Read et al.[22] showed that among a group of 141 cancer patientsattending the medical oncology outpatient clinic for the firsttime, more than half of them were either malnourished (PG-SGA global rating C) or suspected to be malnourished (PG-SGA global rating B), with 55 (39%) patients having aPG-SGA score ≥9. Although the proportion of patientshaving a PG-SGA score ≥9 in our study was lower, withonly 27 (12.3%) patients at 1 month after commencingtreatment, the number of patients who had seen a dietitianwas less than anticipated. At 6 months (T2) after com-mencement of treatment, 28 (14.2%) patients had PG-SGAscore ≥9, a level at which patients would be at critical needfor nutritional intervention [2, 8]. Despite this, only half ofthis group had seen a dietitian. Considering the fact thatnutrition impact symptoms are only one of the componentsof the PG-SGA, it is highly likely that patients’ overall PG-SGA score would be higher, with additional scores awardedfrom other components including weight loss, reduced foodintake, reduced functional capacity, and muscle and fatdepletion on physical examination [18]. The low level ofnutrition intervention indicates that routine nutrition screen-ing may need to be implemented in these populations tosupport referral to a dietitian or for appropriate nutritionsupport including dietary counseling on a high-energy andprotein diet, modified texture, supplementation, or consid-eration of tube feeding if required. We have previouslyreported that 26% of 50 medical oncology patients weresuspected or moderately malnourished according to PG-SGA global rating and 20% had a PG-SGA score ≥9indicating a critical need for nutritional intervention [10].The benefits of early and intensive nutrition interventionhave been demonstrated in several studies. Oncologyoutpatients (N=60) receiving radiotherapy to the gastroin-testinal or head-and-neck area who received nutritionintervention had less weight loss, deterioration in nutrition-al status, global QoL, and physical function compared with

usual care [9]. In palliative care, controlling nutritionimpact symptoms via timely nutritional support can helpimprove or postpone the deterioration of QoL [15].

Study strengths and limitations

As there is a general lack of longitudinal follow-up of thenutrition impact symptoms during and after chemotherapy,the longitudinal and prospective nature of this study whichfollowed patients over 12 months provided valuable data.For instance, a study of breast cancer patients was conductedwith follow-up at 2 weeks and 3 months posttreatment only[25]. With increasing survival rates, it is vital to obtain abetter understanding of the impact of symptoms over time.The heterogeneous mix of patients and large sample size forthis study are also strengths, as most previous studies havefocused only on specific types of cancer populations, thuslimiting the ability to generalize study results.

The main limitation of our study is that it is a secondaryanalysis and was not designed as a nutrition trial. Some ofthe nutritional parameters included in the survey weretherefore limited. For example, weight loss was notincluded in the RSCL survey, and dietary intake and a fullnutrition assessment were also not undertaken. Previousstudies have reported that weight loss is associated with alower QoL. Weight loss has also been shown to be animportant prognostic indicator of poor survival [6], andenergy intake has been found to be correlated withsymptoms including nausea, vomiting, and diarrhea [20].This longitudinal study over 12 months included somepatients with advanced stages of disease which notsurprisingly resulted in significant attrition between T2and T3. Although there were no differences between age,number of symptoms, or PG-SGA score at the previoustime point, it does appear that more stage IV patients werelost to attrition, i.e., this study may be underreporting thenutrition impact symptom prevalence.

Recommendations for future research include conduct-ing a full nutrition assessment using a validated methodsuch as the PG-SGA and investigating the impact ofintensive and individualized dietary counseling on nutri-tion impact symptoms, nutritional status, and clinicaloutcomes. Further research could investigate why patientsexperiencing nutrition-related problems did not see adietitian and whether or not this was due to access issuesor not perceiving the nutrition problems as important. Inour study, dry mouth, nausea, and constipation appeared tobe distressing for patients at T1 yet rated a lower PG-SGAscore due to the perceived lower impact on nutritionalstatus. However, while dry mouth was still noted to beingdistressing across all time points, so were diarrhea andstomach pain which do rate higher on the PG-SGA scale.Future research could investigate the impact that specific

Support Care Cancer (2009) 17:83–90 89

symptoms and symptom clusters have on nutritionalstatus.

Conclusion

Our study findings demonstrate that, despite evolvingpharmacological advancements in cancer treatment, patientscontinue to experience high levels of nutrition impactsymptoms even at 12 months following commencement ofchemotherapy and that the presence of these symptoms areassociated with poorer QoL. This highlights the need notonly for early symptom identification but also adequatefollow-up and continuing efforts to improve the QoL forpeople with cancer. While most published studies havefocused primarily on the earlier phases of diagnosis or inpalliative patients with advanced cancer, studies conductedduring the posttreatment recovery phase of cancer treat-ments are limited. A multidisciplinary health care teamapproach including early and ongoing nutrition screening iswarranted so as to promote the optimal care for patientswith cancer.

Acknowledgements This manuscript was written during an inde-pendent study unit as part of the Bachelor of Nutrition and Dieteticsdegree, Queensland University of Technology and the projectsupervised as part of an NHMRC Australian Clinical TrainingFellowship (ID:324777).

References

1. Argiles JM (2005) Cancer-associated malnutrition. Eur J of OncolNursing 9:S39–S50, Medline. DOI 10.1016/j.ejon.2005.09.006

2. Bauer J, Capra S, Ferguson M (2002) Use of the scored patient-generated subjective global assessment as a nutrition assessmenttool in patients with cancer. Eur J Clin Nutr 56:779–785, Medline.DOI 10.1038/sj.ejcn.1601412

3. Capra S, Ferguson M, Ried K (2001) Cancer: impact of nutritionintervention outcome-nutrition issues for patients. Nutrition 17(9):769–772, Medline. DOI 10.1016/S0899-9007(01)00632-3

4. Clinton M, Lunney P, Edwards H, Weir D, Barr J (1998)Perceived social support and community adaptation in schizo-phrenia. J Adv Nurs 27:955–965, Medline. DOI 10.1046/j.1365-2648.1998.t01-1-00573.x

5. Donnelly S, Walsh D (1995) The symptoms of advanced cancer.Semin Oncol 22:67–72, Medline

6. Grosvenor M, Bulcavage L, Chlebowski R (1989) Symptomspotentially influencing weight loss in a cancer population. Cancer63:330–334, Medline. DOI 10.1002/1097-0142(19890115)63:2<330::AID-CNCR2820630221>3.0.CO;2-U

7. Gupta D, Lis CG, Granick J, Grutsch JF, Vashi PG, LammersfeldCA (2006) Malnutrition was associated with poor quality of life incolorectal cancer: a retrospective analysis. J of Clin Epidemiology59(7):704–709, Medline. DOI 10.1016/j.jclinepi.2005.08.020

8. Isenring E, Bauer J, Capra S (2003) The scored Patient-GeneratedSubjective Global Assessment (PG-SGA) and its association withquality of life in ambulatory patients receiving radiotherapy. Eur Jof Clin Nutr 57:305–309, Medline. DOI 10.1038/sj.ejcn.1601552

9. Isenring E, Capra S, Bauer J (2004) Nutrition intervention isbeneficial in oncology outpatients receiving radiotherapy to thegastrointestinal or head and neck area. Br J Cancer 91:447–452,Medline. DOI 10.1038/sj.bjc.6601962

10. Isenring E, Cross G, Daniels L, Kellett E, Koczwara B (2006)Validity of the malnutrition screening tool as an effective predictorof nutritional risk in oncology outpatients receiving chemotherapy.Support Care Cancer 14:1152–1156, Medline. DOI 10.1007/s00520-006-0070-5

11. Khalid U, Spiro A, Baldwin C, Sharma B, McGough C, NormanAR et al (2007) Symptoms and weight loss in patients withgastrointestinal and lung cancer at presentation. Support CareCancer 15:39–46, Medline. DOI 10.1007/s00520-006-0091-0

12. Komurcu S, Nelson KA, Walsh D (2000) The gastrointestinalsymptoms of advanced cancer. Support Care Cancer 9:32–39,DOI 10.1007/s005200000164

13. Laviano A, Meguid MM (1996) Nutritional issues in cancermanagement. Nutrition 12(5):358–371, Medline. DOI 10.1016/S0899-9007(96)80061-X

14. Marin Caro MM, Laviano A, Pichard C (2007) Impact ofnutrition on quality of life during cancer. Curr Opin Clin NutrMetab Care 10(4):480–487, Medline. DOI 10.1097/MCO.0b013e3281e2c983

15. Marin Caro MM, Laviano A, Pichard C (2007) Nutritionalintervention and quality of life in adult oncology patients. ClinNutr 26(3):289–301, Medline. DOI 10.1016/j.clnu.2007.01.005

16. Mercadante S (1996) Nutrition in cancer patients. Support CareCancer 4:10–20, Medline. DOI 10.1007/BF01769869

17. Olsson U, Bosaeus I, Svedlund J, Bergbom I (2007) Patients’subjective symptoms, quality of life and intake of food during therecovery period 3 and 12 months after upper gastrointestinalsurgery. Eur J of Cancer Care 16(1):74–85, Medline. DOI10.1111/j.1365-2354.2006.00709.x

18. Ottery FD (2000) Patient-generated subjective global assessment.In: McCallum P, Polisena C (eds) The clinical guide to oncologynutrition. The American Dietetic Association, Chicago, pp 11–23

19. Ravasco P, Monteiro-Grillo I, Camilo ME (2003) Does nutritioninfluence quality of life in cancer patients undergoing radiother-apy. Radiotherapy and Oncol 67(2):213–220, Medline. DOI10.1016/S0167-8140(03)00040-9

20. Ravasco P, Monteiro Grillo I, Vidal PM, Camilo ME (2004)Cancer: disease and nutrition are key determinants of patients’quality of life. Support Care Cancer 12:246–252, Medline. DOI10.1007/s00520-003-0568-z

21. Ravasco P, Monteiro-Grillo I, Vidal P, Camilo ME (2005) Impactof nutrition on outcome: a prospective randomized controlled trialin patients with head and neck cancer undergoing radiotherapy.Head Neck 27:659–668, Medline. DOI 10.1002/hed.20221

22. Read JA, Choy SB, Beale P, Clarke SJ (2006) An evaluation ofthe prevalence of malnutrition in cancer patients attending theoutpatient oncology clinic. Asia-Pacific J of Clin Oncol 2(2):80–86, DOI 10.1111/j.1743-7563.2006.00048.x

23. Schuit KW, Sleijfer DT, Meijler WJ, Otter R, Schakenraad J, Vanden Bergh FCM, Jong BM (1998) Symptoms and functionalstatus of patients with disseminated cancer visiting outpatientdepartments. Journal of Pain and Symptom Management 16:290–297, Medline. DOI 10.1016/S0885-3924(98)00091-8

24. Stein KD, Denniston M, Baker F, Dent M, Hann DM, BushhouseS, West M (2003) Validation of a modified Rotterdam symptomchecklist for use with cancer patients in the United States. Journalof Pain and Symptom Management 26(5):975–989, Medline. DOI10.1016/S0885-3924(03)00321-X

25. Wengstrom Y, Haggmark C, Strander H, Forsberg C (2000)Perceived symptoms and quality of life in women with breastcancer receiving radiation therapy. Eur J of Oncol Nursing 4(2):78–88, Medline. DOI 10.1054/ejon.1999.0052

90 Support Care Cancer (2009) 17:83–90