Cancer Survivors SLR - World Cancer Research Fund International [PDF]

A systematic review of randomized controlled trials investigating the effect of nutritional and physical activity interventions on cancer survival. UPDATED REPORT

September 2006

Review Team Dr Trudy Bekkering Rebecca Beynon George Davey Smith Anna Davies Roger Harbord Dr Jonathan Sterne Dr Steve Thomas Lesley Wood

Position Research Associate in Epidemiology Research Associate in Epidemiology Professor of Clinical Epidemiology Research Associate in Epidemiology Research Associate in Medical Statistics Professor of Medical Statistics and Epidemiology Consultant Surgeon and Senior Lecturer Research Associate in Epidemiology

In this updated report, information added since the original report is highlighted in grey. In the tables, information changed since the original report is highlighted in yellow.

1

CONTENTS EXECUTIVE SUMMARY............................................................................................................................................ 4 1. BACKGROUND ......................................................................................................................................................... 5 2. METHODS................................................................................................................................................................... 7 2.1 Search strategy ............................................................................................................................................................ 7 2.2 Inclusion and exclusion criteria ................................................................................................................................. 8 2.3 Data extraction methods ............................................................................................................................................. 10 2.4 Assessing the quality of trials..................................................................................................................................... 10 2.5 Statistical methods ...................................................................................................................................................... 11 3. RESULTS.................................................................................................................................................................... 12 3.1 Selection process......................................................................................................................................................... 12 3.2 Nutritional intervention trials in cancer survivors: general characteristics............................................................. 13 3.3 Nutritional intervention trials in cancer survivors: quality of study design ............................................................ 15 3.4 Nutritional intervention trials in cancer survivors: outcomes .................................................................................. 17 3.5 Nutritional intervention trials in pre-cancer survivors: general characteristics ...................................................... 18 3.6 Nutritional intervention trials in pre-cancer survivors: quality of study design ..................................................... 20 3.7 Nutritional intervention trials in pre-cancer survivors: outcomes ........................................................................... 21 3.8 Physical activity trials in cancer survivors: general characteristics......................................................................... 22 3.9 Physical activity trials in cancer survivors: quality of study design........................................................................ 23 3.10 Physical activity trials in cancer survivors: outcomes............................................................................................ 24 3.11 Loss to follow-up ...................................................................................................................................................... 24 4. OUTCOMES............................................................................................................................................................... 25 4.1 All-Cause Mortality and Cancer-Specific Mortality in Patients With Cancer........................................................ 25 4.2 Cancer Recurrence and Second Primary Cancers in Patients With Cancer ............................................................ 25 4.3 Development of Cancer in Patients With Pre-invasive Lesions .............................................................................. 26 Fig 4.1 Study selection process ........................................................................................................................................ 28 Fig 4.2 Meta-analyses examining the effect of dietary interventions on all-cause mortality in trials of patients with cancer ................................................................................................................................................................................. 29 Fig 4.3 Meta-analyses examining the effect of dietary interventions on cancer mortality in trials of patients with cancer ................................................................................................................................................................................. 30 Fig 4.4 Meta-analyses examining the effect of dietary interventions on cancer recurrence in trials of patients with cancer. ................................................................................................................................................................................ 31 Fig 4.5 Meta-analyses examining the effect of dietary interventions on second primary cancer occurrence in trials of patients with cancer........................................................................................................................................................... 32 Fig 4.6 Meta-analyses examining the effect of dietary interventions development from pre-invasive lesions to cancer in trials of patients with pre-invasive lesions .................................................................................................................. 33 Fig 4.7 Meta-analyses examining the effect of dietary interventions on recurrence of pre-invasive lesions to cancer in trials of patients with pre-invasive lesions. ................................................................................................................. 34 4.4 Quality of Life Data.................................................................................................................................................... 35 4.5 Side effects .................................................................................................................................................................. 36 5. DISCUSSION AND CONCLUSIONS..................................................................................................................... 37 6. REFERENCES............................................................................................................................................................ 39 7. RESULTS TABLES ................................................................................................................................................... 51 7.1 General characteristics of the 53 trials reporting nutritional interventions in cancer survivors, in alphabetical order by first author.................................................................................................................................................. 51 7.2 Break down of the 53 nutritional trials by specific type of intervention in cancer survivors.............................. 75 7.3 Table to show the quality of nutritional intervention trials in cancer survivors................................................... 76

2

7.4 Table to show the outcomes reported by nutritional intervention trials in cancer survivors. Trials are groups by intervention type and outcomes are reported as Odds Ratios (95% confidence intervals).................................. 88 7.5 Effect of nutritional interventions on quality of life outcomes in cancer survivors............................................. 92 7.6 Description of quality of life measurement tools used in randomized control trials ........................................... 99 7.7 Side-effects of nutritional interventions in cancer survivors................................................................................. 102 7.8 General characteristics of the 41 trials reporting nutritional interventions in pre-cancer survivors, in alphabetical order by first author............................................................................................................................. 108 7.9 Break down of the 41 nutritional trials by specific type of intervention in pre-cancer survivors....................... 120 7.10 Table to show the quality of nutritional intervention trials in pre-cancer survivors........................................... 121 7.11 Table to show the outcomes reported by nutritional intervention trials in pre-cancer survivors. Trials are grouped by intervention type and outcomes are reported as Odds Ratios (95% confidence intervals) ............................ 131 7.12 Effect of nutritional interventions on quality of life outcomes in pre-cancer survivors ..................................... 135 7.13 Side-effects of nutritional interventions in pre-cancer survivors ......................................................................... 137 7.14 General characteristics of the 23 trials reporting physical activity interventions in cancer survivors, in alphabetical order by first author............................................................................................................................. 141 7.15 Table to show the quality of physical activity intervention trials in cancer survivors........................................ 156 7.16 Table to show the outcomes reported by physical activity intervention trials in cancer survivors. Trials are grouped by cancer site.............................................................................................................................................. 161 7.17 Effect of physical activity interventions on quality of life in cancer survivors................................................... 163 7.18 Side effects of physical activity interventions in cancer survivors ...................................................................... 177 7.19 Loss to follow-up reported by outcome for each trial arm, in alphabetical order by first author ....................... 179

8. APPENDICES ............................................................................................................................................................. 180 APPENDIX A: Search strategies for on-line databases ................................................................................................. 180 APPENDIX B: In/Out form ............................................................................................................................................. 192 APPENDIX C: Data extraction form............................................................................................................................... 193 APPENDIX D: Randomized controlled trials excluded from the meta-analysis because they had no relevant outcome data ..................................................................................................................................................................................... 223

3

EXECUTIVE SUMMARY Objective The aim was to systematically locate and review all randomized controlled trials (RCTs) which tested the effect of a nutritional and/or physical activity interventions in cancer and pre-cancer survivors. Methods Three methods were used to locate relevant RCTs: Systematic searching of four on-line databases: MEDLINE, EMBASE, AMED and the Cochrane Library Scanning of key texts that were highly relevant to the subject field Scanning the references of relevant reviews Full papers of all relevant trials were retrieved and assessed by two researchers to determine if they met the study criteria. Data were extracted by two independent reviewers and compared for consensus. The main outcome measures of interest were: All-cause mortality, disease-free survival, cancer mortality, primary cancer, cancer recurrence, number of days in hospital, recurrence of pre-malignancies, development from pre-malignancies to cancer, quality of life and intervention side-effects. Results A total of 7061 papers were located of which 543 were considered relevant and obtained in full. 176 of these papers met the study criteria of which 78 reported relevant outcome data. Of these papers 13 included a physical activity intervention, 62 included at least one type of nutritional intervention whilst three trials included a combined nutritional /physical activity intervention. An updated search in March 2006 resulted in the inclusion of an additional 23 trials: 13 of nutritional interventions in cancer survivors; 3 of nutritional interventions in populations with pre-invasive lesions; and 3 of physical activity interventions in cancer survivors. Conclusions The major conclusion to be drawn from this systematic review is the paucity of robust evidence on the effects of nutritional or exercise interventions in the management of diagnosed cancer. Randomised trials were generally small and often reported inadequate details to formally access quality. While promotion of a generic healthy diet was associated with reduced overall mortality the effect estimate was highly imprecise (pooled odds ratio 0.89; 95% confidence interval 0.49 to 1.63). Many of the interventions – such as antioxidants, retinol, fibre – had no clear effects and in specific analyses as many findings tended to the harmful (e.g. antioxidants and cancer recurrence; fibre and progression from colorectal adenomas to cancers) as protective (calcium and progression of colorectal malignancy; generic healthy diet and cancer recurrence). Similarly data on quality of life outcomes in these trials could not be unambiguously interpreted. The main conclusion of this systematic review is that given the large investment in such potential therapies among cancer patients large-scale trials adequately powered to provide robust conclusions should be supported and conducted.

4

1. BACKGROUND Improvements in early cancer detection and advances in cancer treatment mean that individuals are now living for longer after their initial cancer diagnosis. In the United States the five-year survival rate for all cancers combined is estimated at 63%.1 Whilst rates vary with type of cancer and stage at diagnosis this is an increase of 3% from the 2001 published figures. In January 2001 the National Cancer Institute estimated that there were 9.6 million cancer survivors (anyone who has been diagnosed with cancer, from the time of diagnosis through the rest of life2) in the United Sates alone.1, 3 The continuing medical advancements and an ageing population mean that we can expect this number to continue increasing. Information on the number of cancer survivors worldwide is limited: however, the increasing prevalence of cancer in economically developing countries means it must be considered a worldwide public health concern.4 Nutrition and physical activity are frequently classified under the umbrella-term ‘complementary / alternative medicine’ (CAM). CAM, which includes a wide range of therapies, has be defined as “…diagnosis, treatment and/or prevention which complements mainstream medicine by contributing to a common whole, by satisfying a demand not met by orthodoxy or by diversifying the conceptual frameworks of medicine.5 It is well recognised that cancer patients commonly use CAM therapies. A systematic review (including 22 surveys, 12 of which were related to diet, from 13 countries) reported that the average prevalence of CAM use in adult cancer patients was 31.4%; the use of CAM therapies ranging between studies from 7 to 64%.6 There is strong evidence that general interest in CAM is growing; although data are often limited to the United States (US). Between 1990-1997 the use of CAM in the US increased from 33.8 to 42.1%.7 Furthermore, it was estimated that in 1997 out-of-pocket expenditure on complementary and alternative therapies was between $27.0 and 34.3 billion (depending on how conservatively one estimated the cost of visits to complementary practitioners).8 This was similar to the projected cost of out-of-pocket expenditure resulting from all US physician services for 1997 ($29.3 billion).9 Data on specific nutritional or physical activity therapies is scarce although the personal success stories of cancer survivors using such therapies are often the focus of popular news coverage (http://news.independent.co.uk/uk/health_medical/story.jsp?story=5865 and http://www.cnn.com/HEALTH/library/CM/00002.html). Despite the lack of data, studies from the Us have reported that between 1990 and 1997 out-of-pocket expenditure for high-dose vitamins increased from $0.9 to 3.3 billion 10; supporting a picture of growing CAM use. Despite the growth of the CAM industry and the common use of complementary medicines in cancer management the effects of many CAM therapies remain unproven. In a recent nonsystematic review of the literature the American Cancer Society found no convincing evidence for a beneficial effect of nutrition or physical activity interventions for survivors of the four main cancer types: breast, colorectal, lung or prostate.2 Whilst no evidence of harmful effects was observed the benign nature of nutritional and physical activity therapies in general cannot be assumed. Randomized controlled trials investigating the effect of dietary interventions on cancer prevention have suggested that they can have potential harmful effects.11 The increasing numbers of cancer survivors coupled with their growing interest in CAM highlights the importance of rigorous research in this field. The conflicting, and often incorrect,

5

information available means that there is great uncertainty for both health professionals and cancer survivors as to the possible role of diet and physical activity in cancer management. This problem was highlighted in a recent study investigating the CAM information relating to cancer available on the World Wide Web. The authors of this study reported that the most popular websites on CAM for cancer offered information of extremely variable quality.12 Most studies offered recommendations on a wide range of treatments (diet was one of the three most frequently discussed) which the authors found were often unsubstantiated by sound scientific evidence. Furthermore, three of the 32 web-sites reviewed in this study offered advice which had the potential for harming patients.12 While the effect of nutrition and physical activity on primary cancer prevention has been comprehensively reviewed4 we cannot assume that the resulting recommendations are relevant for cancer survivors. For these reasons we were funded by WCRF/AICR to review the scientific evidence on the effect of nutrition and physical activity in cancer survival.

6

2. METHODS 2.1 Search Strategy We conducted a systematic review of the literature using standard review methodology.13, 14 Four on-line databases were searched for articles published before or during September 2003: i. MEDLINE from 2000 onwards ii. EMBASE from 1999 onwards iii. AMED from 1985 onwards iv. The Cochrane Library [including Database of Abstract of Reviews of Effects (DARE), Cochrane Database of Systematic Review (CDSR), The Cochrane Central Register of Controlled Trials (CENTRAL) and Health Technology Assessments (HTA)] all years. MEDLINE was only searched from 2000 and EMBASE from 1999 onwards because the Cochrane Library includes all RCTs from MEDLINE and EMBASE published before or during 1999 (MEDLINE) and 1998 (EMBASE). Four individual search strategies were created according to the search specifications of each online database using Medical Subject Headings (MeSH) terms and title/abstract text words for: i. cancer (e.g. cancer$.tw., neoplasm$.tw.) ii. survivors (e.g. surviv$.tw., recur$.tw, , treatment$.tw.) iii. nutrition or physical activity (e.g. diet$.tw., supplement*, exercis$.tw.) iv. randomized controlled trials (e.g. random$.ti,ab., placebo$.tw.) Our initial search of the literature revealed a key body of scientific work which investigated the effect of nutrition and physical activity in populations with pre-malignancies. Whilst our initial protocol had focused on cancer survivors we felt that it was important to broaden the study selection criteria to include participants who had survived pre-malignant as well as malignant lesions. For this reason search terms such as: polyp*, adeno$.tw., precancer$.tw. and dysplasia$.tw.) were included in the search strategies. The complete search strategies for all four databases are listed in Appendix A. Additional papers were located from scanning the references of two relevant books 15, 16, reviews found during the on-line database search17-44 and three key non-systematic reviews published after searching was complete.2, 45, 46 Our search was more restricted than specified in the protocol. Limited time and project resources meant that we had to prioritise databases to ensure that they were searched thoroughly. In consultation with our information specialist we did not search ISI Web of Science or BIOSIS as we felt they would not give us any additional peer-reviewed trials to those already located in MEDLINE, the Cochrane Library, EMBASE or AMED. One reviewer (AD) was responsible for searching the title and abstracts of all papers located by the search strategies. Relevant studies, which met the inclusion criteria, together with those whose suitability could not be determined from the abstract or title, were retrieved. Two reviewers (AD and ST) read all retrieved papers in full and re-confirmed their suitability for inclusion using a specifically designed ‘In/Out’ Form (Appendix B). Where necessary, a third reviewer (GDS) was consulted on disagreements. Non-English papers were sent for translation

7

before deciding on their relevance. At this stage there was no exclusion based on the quality of the study. Furthermore, studies were only excluded on the basis of the outcomes they reported at the data extraction stage. Using the same search strategy, we searched The Cochrane Library, 2006 Issue 1, and MEDLINE, EMBASE, AMED to update from previous searches to March 2006. LW searched titles and abstracts, and read retrieved papers in full. Inclusion/exclusion was confirmed by ST or JS. 2.2 Inclusion and Exclusion criteria To be included in the review papers had to be published in peer-review journals as full articles or abstracts. In addition they had to meet four inclusion criteria (summarized in Table 2.1): i) The study design must be a randomized controlled trial. Studies using ‘historical’ or ‘medical literature rates’ as controls, studies in which participants were not randomized as individuals but in groups and studies in which there was no suitable control group (in order for conclusions on the effect of the nutritional/physical activity intervention to be drawn) were excluded. ii) The study population must be cancer or pre-cancer survivors. Our definition of a cancer (or pre-cancer) survivor was adapted from Brown et al. and included ‘anyone who has been diagnosed with cancer (or pre-cancer), from the time of diagnosis through the rest of life’.2 There were no restrictions according to type or stage of cancer, age, sex or ethnicity. iii) At least one intervention must include a nutritional and/or physical activity component. The definition of a ‘dietary intervention’ was one that altered the intake of foods either directly, for example by giving vitamin supplementation pills, or indirectly, for example through health education. Food was defined as ‘…beverages, confectionary, ingredients in preparation of foods and advertised dietary supplements which contained added vitamins’.47 In addition micronutrients were also included in our definition of food. This definition excludes a number of substances, including herbs, botanicals, enzymes, organ tissues and glandulars, which are included in, for example, the definition of ‘dietary supplements’ issued by the US Food and Drug Administration (http://www.cfsan.fda.gov/~dms/qa-sup5.html). Peri-operative feeding, where outcomes were related to peri-operative complications and not cancer survival, were excluded as were synthetic retinoids, vitamin analogs and polysaccharide K (PSK) interventions as they are not primary foods. PSK is a protein bound polysaccharide that is extracted from the mycelia of the mushroom Coriolus versicolor. iv) One or more of the following outcomes must be reported: All-cause mortality; Disease-free survival; Cancer mortality; Primary cancer; Cancer recurrence; Second primary cancer; Number of days in hospital; Recurrence of pre-malignancies; Development of pre-malignancies to cancer; Quality of life; Intervention side-effects.

8

Table 2.1 Review inclusion and exclusion criteria. Included Study characteristic Design

Participants

Intervention

Outcomes

• Individual patient parallel randomized controlled trials

• Study participants must be cancer (or pre-cancer) survivors (anyone who has been diagnosed with cancer (or pre-cancer), from the time of diagnosis through the rest of life) • At least one intervention must include a nutritional and/or physical activity component • We did not exclude studies where nutritional / physical activity interventions were given alongside other cancer treatments (e.g. chemotherapy and radiotherapy)

Excluded • Trials with historical or medical literature rates as controls • Trials using group, rather than individual, randomization methods • Trials with no suitable ‘control’ against which the impact of the nutritional and/or physical activity intervention could be assessed • Trials where only a sub-group of the population were cancer (or pre-cancer) survivors and the outcome data were not reported by these sub-groups • Trials in which the comparison groups were healthy controls • Nutritional interventions for cancer cachexia • Intra-venous nutritional interventions (i.e. parenteral feeding). • Interventions which only compared types of nutritional delivery (e.g. parenteral vs. enteral feeding / application or not of a jejunal feeding tube) • Nutritional interventions not classified as ‘food’* • ‘Visualization trials’ (e.g. where participants visualize their immune system combating cancer • ‘Muscle relaxation’ or physiotherapy interventions • Peri-operative feeding where outcomes are related to peri-operative complications and not cancer survival • Synthetic retinoid interventions / vitamin analogs • Protein-bound polysaccharide K (PSK) interventions (PSK is a protein bound polysaccharide that is extracted from the mycelia of Coriorus versicolor).

Studies must report one or more of the following outcome measures: All-cause mortality Disease-free survival Cancer mortality Cancer recurrence Primary cancer Second primary cancer Number of days in hospital Development from pre-cancer to cancer Recurrence of pre-malignancies Side-effects or Quality of Life

*Our definition of food included ‘…beverages, confectionary, ingredients in the preparation of foods and advertised dietary supplements which contain added vitamins’.47 In addition we also included micronutrients.

9

2.3 Data Extraction Methods Data extractions were performed in duplicate by two independent reviewers (AD and TB) using a pre-piloted data extraction form (Appendix C). Extractions were compared for agreement; where a consensus could not be made a third reviewer (GDS) was consulted. Data on study design, study participants, intervention, stage of cancer, cancer treatment (pretreatment, during-treatment or post-treatment) at time of intervention, compliance, outcome measures, time of follow-up, quality of trials and loss to follow-up were extracted. When papers reported data from multiple time points and/or post-intervention time points we extracted outcome data for the time point closest to the end of the intervention. In studies which did not report outcomes of interest only limited data (up to question 25 on the data extraction form (Appendix C)) were recorded. All data were entered into a specifically designed Access database by one researcher (RB). Given resources, we were unable to contact authors for clarification of methods or results. Data extraction for the 2006 update was performed by LW and verified by JS and ST. Data were added to the Access database by LW. 2.4 Assessing the Quality of Trials We used four measures to assess the quality of the trials: i. Adequate methods of randomization (generation of the allocation sequence) ii. Adequate concealment of allocation iii. Blinding of outcome assessors to participant treatment allocation iv. The use of intention to treat analysis Generation of allocation sequences were considered adequate if they generated unpredictable sequences (e.g. computer generated random numbers, table of random numbers, drawing lots or envelopes, coin tossing, shuffling cards or throwing dice).13 Concealment of allocation sequences was only adequate if patients, and enrolling investigators, could not foresee their assignment (e.g. use of a priori numbered or coded drug containers prepared by an independent pharmacy; central randomization; sequentially numbered, sealed, opaque envelopes).13 Allocation concealment is the dimension of trial quality that has been empirically demonstrated to have the strongest association with the size of estimated treatment effects; in contrast there is little evidence that the method of generating the sequence is important. The coding and interpretation of blinding is difficult since blinding, when reported in a trial, might refer to participants, health care providers, data collectors, judicial assessors of outcomes, data analysts or personal who wrote the article.48 Since the blinding of physicians/carers is often not feasible in nutritional/physical activity interventions we have concentrated on assessing whether the outcome assessors were blinded to participant treatment allocation. Blinding of outcome assessors was recorded as ‘yes’, ‘no’ or ‘not clear’ if the information was not available in the paper. The importance of blinding, as a measure of design quality, varies according to the outcomes of interest: it is clearly more important when measuring outcomes such as quality life; and less important when measuring all-cause mortality. We also assessed whether the participants themselves were blinded to their group allocation; although in many interventions (e.g. changing dietary habits or undertaking more physical activity) such blinding is not feasible. Intention to treat analysis was recorded as having occurred only if all individuals (including non-compliers) were included in an analysis according to the groups that they had originally

10

been allocated to at randomisation. Studies that described their methods of analysis as ‘intention to treat’ but had not actually performed it were recorded as not having used intention to treat analysis. 2.5 Statistical Methods Trials recruiting cancer patients were analysed separately from those recruiting pre-cancer survivors. Data from all cancer sites were analysed together. Odds ratios were used to quantify treatment effects. Where trials had multiple intervention arms, and the interventions were different types (e.g. one antioxidant supplement and one dietary counselling intervention) each arm was compared to the control group and analysed separately. This meant that some trials could be included in more than one analysis. When trials included more than one intervention arm, but interventions were of the same type, data from both arms were treated as a single group. This method avoided the control group being included twice in the same meta-analysis but was only performed after first checking that the results from comparing each arm to the control separately were consistent in size and direction of effect. Factorial trials were analysed assuming no interaction between interventions. Both fixed-effect meta-analysis (Mantel-Haenszel method) and random-effects meta-analysis (der Simonian and Laird method) were used. As well as deriving tests for heterogeneity by referring the heterogeneity statistic Q to the chi-squared distribution, we quantified the amount of heterogeneity in each meta-analysis using the I2 statistic49, which gives the percentage of variance in the meta-analysis that is attributable to heterogeneity and is calculated as I2 = (Q - df) / Q x100%, where the number of degrees of freedom df is one less than the number of studies in the meta-analysis. The results of the random-effects model are reported only when they differ from the fixed-effect model, which is the case when I2> 0. All analyses were performed in Stata 9.2. As anticipated in the study protocol there were too few studies in the meta-analyses of specific outcomes and interventions to justify the use of meta-regression, performing analyses of dose-response effects, or an examination of small study effects.

11

3. RESULTS 3.1 Selection process (original review) Figure 4.1 (page 28) shows the study selection process. The search strategies generated 7061 potentially relevant papers. From the 543 papers retrieved for more detailed evaluation 367 did not meet the inclusion criteria. Of the remaining 176 papers 98 contained no relevant outcome data and were excluded: the references for these papers are listed in Appendix D. Of these 98 two papers did report relevant outcomes but the data was ambiguous and could not be extracted accurately.50, 51 Seventy-eight papers are, therefore, included in the results section: 13 included a physical activity intervention, 62 nutritional interventions and 3 a physical activity and nutritional intervention component. Since some papers reported outcomes for more than one trial the number of trials in this review is greater than the number of papers: from the 78 papers located the outcomes of 91 trials were reported. These additional 13 trials are a result of the following: three papers, as detailed above52-54, tested interventions that included both nutritional and physical activity components, eight papers reported outcomes for two interventions55-62, of which two also reported outcomes for two different study populations.58, 63 The matrix below summaries the number of papers, and corresponding number of actual nutritional and physical activity trials, located for cancer and pre-cancer survivors. Cancer survivors Papers Trials

Pre-cancer survivors Papers Trials

Total Papers

Trials

65 16 81*

75 16 91

Intervention Nutritional Physical activity Total

37 16 53

39 16 55

28 0 28

36 0 36

* The total number of 81 papers arises from the fact that 3 papers reported both nutritional and physical activity interventions (78 + 3 = 81). In summary, we found 91 trials reporting relevant outcome data: • 39 nutritional intervention trials in cancer survivors from 37 papers 52-54, 60, 63-95 (one paper reported data for two nutritional interventions60; one paper reported data for two study populations63). • 16 physical activity intervention trials in cancer survivors52-54, 96-108. • 36 nutritional intervention trials in pre-cancer survivors from 28 papers55-59, 61, 62, 109128 (seven papers reported data for two nutritional interventions55-59, 61, 62; one paper reported data for two study outcomes58). • 0 physical activity interventions in pre-cancer survivors There were four nutritional trials (Calcium Polyp Prevention Study110, 111; Nutritional Prevention Cancer Study Group64, 71, 92; Huixian Intervention Study61, 127; Linxian Dysplasia Trial119, 120) and one physical activity intervention (‘Starting Again’ Programme) which reported outcomes of interest in more than one paper.52, 53

12

Results from the updated search are shown as follows: Hits 2419 2315 89 3200 82 35 196 11 8347

MEDLINE EMBASE AMED CENTRAL CDSR Reviews CDSR protocols DARE HTA Total

Duplicates 192 627 65 2613 36 8 98 2 3641

2227 1688 24 587 46 27 98 9 4706

Of the 170 papers retrieved for more detailed evaluation, 23 were included in the updated review. The matrix below summaries the number of papers, and corresponding number of actual nutritional and physical activity trials, located for cancer and pre-cancer survivors. Cancer survivors Papers Trials

Pre-cancer survivors Papers Trials

Total Papers

Trials

15 7 22

16 7 23

Intervention Nutritional Physical activity Total

12 7 19

13* 7 20

3 0 3

3 0 3

* One element of a combined intervention in one trial129 was withdrawn part way through. This was treated as two trials – (1) combined beta-carotene and vitamin E, and (2) vitamin E alone. Please note that in the following results section when we discuss ‘trials’ (rather than ‘papers’) the actual references are to the original papers and, therefore, a single reference may be given to represent the results of more than one trial. 3.2 Nutritional-based intervention trials in cancer survivors: general characteristics The general study characteristics of the 53 nutritional intervention trials in cancer survivors are reported in Table 7.1. Two papers included two different nutritional interventions; however the general characteristics of these trials were the same and are, therefore, reported within the same row.60, 130. Of the additional trials included in the 2006 update, details from one trial were extracted from two separate publications131, 132, both of which have been included in Table 7.1 and Table 7.3, but the outcomes are reported in one row in Table 7.4. For these reasons Table 7.1 contains 52 references. Of the 53 trials 9 were in breast cancer survivors63, 65, 69, 77, 84, 91, 133, 134, 3 bladder 73, 74, 90, 6 skin64, 71, 75, 79, 92, 131, 5 head or head and neck86, 129, 130, 135 (of which one was that of Bairati et al.129 which was treated as two trials after one element of the combined intervention was withdrawn as described above), 3 lung87, 88, 136, 1 cervix85, 4 liver68, 70, 80, 137, 1 haematological72, 2 colorectal83, 138, and 1 prostate139; the remaining 17 trials included study populations with more than one type of cancer survivor.52-54, 60, 66, 67, 76, 78, 81, 82, 89, 93-95, 140, 141 All trials were in adult study populations.

13

Food-based interventions The most commonly reported trial (n = 14) was a food-based intervention which counselled a change in dietary life-style. Of these 14 trials, two gave dietary advice to improve weight loss in overweight women vs. usual treatment63; nine counselled a general reduction in fat (as a percentage of total calorie intake), increase in fibre, increase in fruit and vegetable intake and/or an optimal calorie/protein intake versus general dietary guidelines82, 138 or usual treatment/diet.66, 76-78, 130, 131, 141 Three trials tested an exercise intervention combined with dietary advice vs. usual treatment.52-54 One trial gave a commercial supplement in addition to the counselling guidelines if necessary.141 Two trials had approximately 100 participants or more in each trial arm.52, 53 The numbers in the remaining trials ranged from 19 to 79 in each study arm. This data may underestimate the size of the study populations since three trials 54, 63, 82 did not report numbers randomized, but gave only numbers with baseline characteristics. Six of the food-based intervention trials lasted for ≥ 12 months (mo63, 66, 77, 78, 138; four for ≥ 3 mo but 80% of participants in the first 3 years of the trial in both groups and 79% (intervention) vs. 82% (control) in the 4y. In the comparative controlled trial109 >75% of the assigned supplement was consumed by 74% of the intervention and 84% of the control; the difference in compliance between trial arms was statistically significant at the 5% level. Six trials57, 114, 118, 121, 125 generally reported reductions in the total energy consumption derived from fat; increases in fibre and increases in fruit and vegetable consumption in the intervention compared with the control; table 7.10 reports these changes in detail. The method of recording compliance was unclear in one study142 but the level of compliance over the four years of the intervention was reported for both interventions. Again, details are reported in table 7.10. Supplement Intervention Trials In four trials randomization was performed using computer generated random numbers110, 111, 117, 144 ; one trial used seed numbers randomization by reaction time of the programme user112, and one allocated a number using a table of random numbers before the start of the trial.143 The remaining 26 trials did not clearly report randomization method. Two trials55, 96 used independent randomization centres to ensure adequate concealment of allocation. In one trial144 the assignment codes were maintained by the statistician and pharmacist, in one the “key table was sealed; blind breaking was conducted after data and evaluated subjects were fixed”143, while in three studies110, 111, 117 the method of concealment was not reported

20

although the control and placebo tablets were reported to be identical; from these descriptions we cannot conclude that these latter four trials used adequate concealment of allocation. In the remaining trials the method of concealment was not reported. Only one trial did not report survival outcome data.116 Of the remaining 31 studies nine undertook intention to treat analysis56, 61, 110, 111, 123, 127, 145; ten did not55, 57, 115, 119, 120, 128, 142, 143 while in 11 it could not be concluded from the paper.58, 62, 112, 113, 117, 124, 126, 144 In the final study122 all-cause mortality was analysed by intention to treat but recurrence of premalignancies was not. Three studies did not report whether outcome assessors were blinded.58, 96, 145 Assessors were blind in the remaining 29 trials. Four studies58, 128 did not report whether participants were blind to treatment allocation. In the remaining trials three57, 142, 145 did not blind participants. A total of twenty-five trials55, 56, 59, 61, 62, 110-113, 115-117, 119, 120, 122-124, 126, 127, 143, 144 were therefore double-blind. Compliance was not reported in five trials.58, 112, 123, 144 In one trial the method of recording compliance was not clear, although details of levels of compliance were reported.142 In the remaining 26 trials one study reported giving the supplement under strict supervision126, one ascertained compliance by interview143 while the remaining trials used two main methods to measure compliance: biochemical (blood, fecal or urine) levels and pill disappearance. As stated with food-based interventions the definitions used to determine compliance varied considerably between trials Table 7.10 gives the specific details for each trial.

3.7 Nutritional-based intervention trials in pre-cancer survivors: outcomes Table 7.11 details the outcomes for each trial. Nine trials reported all-cause mortality, 7 disease-free survival, 4 cancer mortality, 21 recurrence of pre-malignancies, 11 development of pre-malignant to malignant, 1 quality of life and 12 data on side-effects. The specific results of the ‘survival’ outcomes (all-cause mortality, disease-free survival, cancer mortality, recurrence of pre-malignancies and development from pre-malignant to malignant) will be discussed first followed by the data on quality of life and then side-effects. The point estimates (95% confidence intervals) for all outcomes are reported for each trial in table 7.11. Only those intervention/outcome combinations reported in more than one trial were analysed further. All-cause mortality Seven trials reported data on all-cause mortality: 4 food-based55, 109, 125, 142;2 calcium55, 110; 2 multivitamins119, 121; and one probiotic bacteria.142 Three food-based interventions counselled or gave increased fibre 55, 109, 142, and one counselled lower fat, higher fibre and higher fruit and vegetable intake125. Disease-free survival Eight trials reported data on disease-free survival: 2 beta-carotene59, 124; 1 calcium127; 1 green-tea61; 2 multivitamins120, 144; 1 retinol59, and 1 vitamin C.128 Cancer Mortality Four trials reported data on cancer mortality: 1 food-based intervention142, 1 beta-carotene112, 1 multivitamin119, and one probiotic bacteria.142

21

Development from pre-malignant to malignant Eleven trials reported data on the development of pre-malignancies into cancer: 3 foodbased109, 125, 142; 1 aged garlic extract143; 1 beta-carotene62; 2 calcium61, 111; 1 folate62; 1 lycopene145; 1 multivitamins119; and one probiotic bacteria.142 Three food-based interventions which aimed to examine the effect of fibre supplementation on the progression of colorectal adenomas to malignancy were combined. The beta-carotene, lycopene and multivitamin trials were combined to look at the effect of antioxidants on the development from pre-cancer to cancer. The effect of the two calcium trials on cancer development is also shown. (Figure 4.6) Recurrence of pre-malignancies Twenty-one trials reported data on the recurrence of pre-malignancies: 7 food-based55, 57, 109, 121, 125, 142 ; 2 beta-carotene56, 57; 2 calcium55, 111; 2 folate117, 123; 5 multivitamins56, 58, 115, 122; 1 N-acetylcysteine58; one probiotic bacteria.142, and 1 retinol.126 The food-based interventions were pooled (after combining MacLennan et al.’s intervention groups) to look at the effect of a healthy diet on recurrence of pre-malignancies. In this forest plot the high fibre and low-fat dietary advice arms of one trial57 were combined after checking that the effects were similar to avoid including control arm twice and overweighting the trial. Six trials were combined to look at the overall effect of antioxidant supplements on the recurrence of pre-malignancies in pre-cancer survivors. In this forest plot the multivitamin and N-acetylcysteine arms of one trial58 were combined after checking that the effects were similar to avoid including control arm twice and overweighting the trial. Finally the effect of the two calcium and folate trials on the recurrence of pre-malignancies are shown. (Figure 4.7) Quality of Life One nutritional intervention114 reported QoL outcomes in pre-cancer survivors. The details of the quality of life outcomes are reported in Table 7.12. Table 7.6 gives a description of quality of life measurement tools used in these randomized control trials. Results section 4.4 further discusses the quality of life data found in all trials included in the review. Side Effects Twelve trials reported data on the effect of nutritional-based intervention on side-effects in pre-cancer survivors: 2 food-based interventions55, 109; 1 aged garlic extract143; 5 betacarotene59, 62, 112, 116, 124; 2 calcium55, 111;1 multivitamins115, and 1 retinol59 intervention. The specific details of the side-effects reported in these trials are detailed in table 7.13. Results section 4.5 further discusses the quality of life data found in all trials included in the review.

3.8 Physical activity trials in cancer survivors: general characteristics Table 7.14 reports the general characteristics of the twenty-three physical activity intervention trials. Twelve of the trials were in breast cancer survivors96-99, 101, 103, 105, 106, 146149 ; two haematological108, 150; one lung107; one prostate104, and seven which included study populations with more than one type of cancer survivor.52-54, 100, 102, 151, 152 The majority of trial participants were either taking treatment (chemotherapy, radiotherapy or undergoing surgery, during the intervention period96-98, 100, 101, 104, 108, 146, were undergoing maintenance therapy150, or were post-treatment.52, 53, 99, 102, 103, 105, 106, 148, 149, 151, 152 In the remaining three trials study participants were classified as ‘pre-treatment’.54, 107, 147 All trials, except one,150 were in adult populations. The trials included a range of supervised and home-based interventions which varied in both intensity and length of activity. Furthermore, the comparison groups differed between studies: ten trials had a ‘usual treatment’ control

22

group.52-54, 96-98, 100, 101, 104, 146; in six trials the control groups were either sedentary 99, 105, performed no exercise102, 103, 107 or received no therapy108; one trial compared an exercise intervention with muscle relaxation151; one trial compared exercise with a psychosocial programme148; three used a control group in which participants were told to continue usual activity147, 150, 152; one trial placed control participants on a waiting list to start an exercise intervention,106, and one trial149 used a crossover design (the outcome was taken at the end of the first period) where the control group was taken from a waiting list, and was asked to engage in their usual level of activity. Eleven trials included ≤ 20 participants in each trial arm97, 99, 101, 102, 105, 106, 108, 146, 148-150, five included between 20 and 50 participants96, 98, 100, 103, 151, five between 50 and 101 participants52, 53, 104, 107, 152 and one trial had 221 participants in each trial arm.54 One trial gave only the number of participants (38) in the trial.147 Three of these trials did not give numbers randomized but only the numbers of individuals in each trial arm for baseline measurements.54, 96, 101 The longest trial lasted 6 mo.97 One trial was for 4 months.150 Eight trials lasted 12 wks or more: one 4 mo trial54, one 26 wk trial98; one 15 wk trial103; one 14 week trial152, and five trials of exactly 12 wks intervention.104, 106, 146, 148, 149 Ten trials lasted between 7d and 10 wks: three trials included 10 wks interventions101, 102, 153; two trials lasted 7 wks52, 53, two trials 6 wks105, 108; one trial was for three weeks151; one trial lasted a mean of 13.6 d (exercise) vs. 15.2 d (control)100and one trial included a 7 d intervention.107 In two trials the length of intervention was not clear.96, 147 Only one trial52 did not measure outcomes during the intervention period. In this trial all-cause mortality was measured at 12 months follow-up.

3.9 Physical activity trials in cancer survivors: quality of study design Table 7.15 summarizes the design quality of the 23 physical activity trials. Seven trials used adequate randomisation methods: five used computer-generated random number lists96, 108, 146, 151, 152 ; three a random numbers table98, 103, 104; and two used Effron’s method of biased coin.52, 53 One study reported that randomisation instructions were prepared by the hospital statistician but further details were not available.107 One reported using a non-probability sampling method in which participants were assigned randomly by a computer.147 The remaining eleven trials did not report the randomisation methods used. Method of concealment allocation was not reported in fifteen trials. In the remaining eight trials only three used adequate methods: two trials used a numbered series of sealed envelopes96, 103 whilet in one trial allocation was carried out at a statistics centre.54 One trial reported using sealed envelopes107, and one reported using envelopes but did not give details of whether these were sealed.150 Neither gave details of whether or not these were sequentially numbered and, therefore, these methods were considered inadequate. Finally, three trials reported further information about concealment of allocation - one trial reported that the study co-ordinator revealed group assignment after baseline testing98, in one the study co-ordinator was concealed from the treatment allocation until completion of baseline testing and stratification – but the specific methods were not detailed104 and one trial reported that the sequence was concealed until the assignment of interventions.151 The majority of trials (n = twenty) did not report whether the outcome assessor was blinded to treatment allocation. In the remaining three trials one did blind the outcome assessor104, while two did not.100, 151 The nature of the intervention meant that no participants were blinded to their treatment allocation; therefore no studies were double-blind.

23

Only four studies reported survival outcomes (rather than quality of life and side-effect data). In these trials two used intention to treat analysis52, 100, one did not108, and in one it was not clear.105 In thirteen of the studies the degree of compliance to the intervention was not clear or not reported. The remaining trials used a variety of methods to report compliance including session attendance rate, completion of prescribed exercises, improvement in functional capacity and self-report of physical activity level. Of the ten trials that reported compliance six stated that they obtained ≥ 80 % compliance rates.99, 100, 102, 103, 106, 107 One trial reported a 40% improvement on functional capacity in the exercise group over the course of the intervention.101 One trial reported 70% attendance at an exercise class.146 One reported 100% compliance with a request not to start further physical activity in the completing intervention group.148 The final trial reported the percentage of individuals who did not comply and found that 22% of the exercise group never initiated the brisk walking intervention whilst 13% stopped the walking program at least 1 mo before study conclusion. Furthermore, 50% of the usual-care group reported maintaining or increasing their physical activity to a moderateintensity level.96 Caution is necessary when interpreting these figures since several trials only report the compliance of those men and women who completed the intervention; rather than of all those randomized to each group.

3.10 Physical activity trials in cancer survivors: outcomes Survival Outcomes Only four physical activity intervention trials reported data on ‘survival’ outcomes: one trial of individuals with haematological cancer108, and one trial including survivors with cancer at a range of sites52 looked at all-cause mortality; one breast cancer trial reported cancer recurrence105, and one trial in participants with a range of different cancers looked at the effect of the intervention on number of days stayed in hospital.100 The odds ratios and 95% confidence intervals are reported for all cause mortality and cancer recurrence in Table 7.16. Quality of Life Twenty trials reported the effect of physical activity interventions on quality of life. Eleven of these trials were in breast cancer survivors.97-99, 101, 103, 105, 106, 146-149 Of the remaining nine trials one was in lung cancer survivors107, one in prostate cancer survivors104, one in haematological cancer survivors150 and five in study populations with more than one type of cancer survivor.53, 54, 100, 102, 151, 152 All trials except one150 were in adult populations. Table 7.17 details the findings from each trial. Table 7.6 gives a description of quality of life measurement tools used in these randomized control trials. Results section 4.4 further discusses the quality of life data found in all trials included in the review. Side-Effects Five trials reported side-effects of physical activity interventions; three in breast cancer survivors96, 103, 149, one in survivors of haematological cancer150, and one in a study population with more than one type of cancer survivor.100 Table 7.18 reports these findings. 3.11 Loss to follow up Available information on loss to follow up in each trial is presented in Table 7.19. We urge caution in interpreting this table, since loss to follow is difficult to define and often poorly

24

reported. In particular, a paper that reports no losses to follow up may reflect the fact that the authors successfully followed all participants, but it is also possible that the true scale of losses to follow up were not accurately reported. In their landmark study examining dimensions of methodological quality associated with estimates of treatment effects in controlled trials Schulz et al.154 found that, contrary to their expectations, trials that reported having excluded participants after randomization did not yield exaggerated estimates of treatment effects. One suggested explanation was that some authors had inappropriately reported an absence of loss to follow up.

4. OUTCOMES 4.1 All-Cause Mortality and Cancer-Specific Mortality in Patients With Cancer As shown in Fig. 4.2 (top), there is little evidence that a healthy diet, given as dietary advice separately or in specific combinations with supplements, weight loss, or exercise is associated with a reduction in all-cause mortality (pooled OR = 0.89 (95% CI 0.49 to 1.63).52, 63, 76, 82, 84, 131 Results from the three breast cancer trials only (results not shown in figure), the pooled odds ratio was 0.70 (95% CI 0.26 to 1.87).63, 84 Data from these three small breast cancer studies (Fig 4.3, top) also suggested a reduction in cancer-specific mortality (pooled OR = 0.53) with healthy diet interventions although the confidence interval was wide (95% CI = 0.16 to 1.79). There was no between-trial heterogeneity ( I2 = 0%). The nine trials60, 67, 71, 74, 79, 83, 89, 135, 136 that that included an antioxidant supplement found no evidence of an association between this intervention and all-cause mortality, compared with placebo or usual treatment (OR = 0.99, 95% CI = 0.87 to 1.13), with no heterogeneity ( I2 = 0%). (Fig 4.2, middle). Combined data from only the two skin cancer trials (data not shown), gave a similar result (pooled OR = 0.92, 95% CI = 0.70 to 1.22).71, 79 Although a large selenium trial64 in skin cancer patients showed a 41% (95% CI = 61% to 12%) reduction in cancer mortality in the intervention group, the other two trials reporting the effect of antioxidants67, 135 on cancer mortality found no evidence of a protective effect of β carotene on head and neck cancer mortality (OR = 1.26, 95% CI = 0.62 to 2.56, and OR 0.88, 95% CI 0.26 to 2.96). The pooled odds ratio for these three trials was 0.80 (95% CI = 0.47 to 1.35) and considerable between study heterogeneity (I2 = 42.1%). (Fig 4.3, middle). There was no evidence of an effect of retinol on all-cause mortality (Fig 4.2, four trials: OR = 0.97, 95% CI = 0.83 to 1.13),60, 72, 75, 87 cancer mortality (Fig 4.3, three trials OR = 0.92, 95% CI = 0.65 to 1.31),75, 85, 87 or disease-free survival (three trials;60, 72, 75 data not shown in figures), compared with usual treatment. 4.2 Cancer Recurrence and Second Primary Cancers in Patients With Cancer As indicated in Fig. 4.4, there was considerable heterogeneity in the results of four trials, all of which included breast cancer survivors only, that examined the effect of healthy diet compared with usual treatment on cancer recurrence. Although all these trials were small, there was evidence of a protective effect (OR = 0.20, 95% CI = 0.06 to 0.65) in the largest (110 patients, 19 events)77.

25

There was little overall evidence of an effect of antioxidant, retinol, or vitamin B6 interventions on cancer recurrence; however, the trials estimating antioxidant effects included a diverse range of cancer sites and had widely varying results ( I 2 = 70.0) (Fig 4.4). Results from two trials86, 87 of the effect of retinol on second primary cancers also suggested a reduction in the risk of second primary cancers (summary OR = 0.57), although the confidence interval was wide (95% CI = 0.31 to 1.07) (Fig 4.5). A meta-analysis of five trials of antioxidants to prevent occurrence of second primary cancer (including two of the largest trials in the review64, 79) showed considerable heterogeneity (I2 = 76.6%) (Fig 4.5). The three trials that included only patients with head and neck cancer each showed an increased risk of a second primary cancer although on each occasion the estimate was imprecisely measured. These results were reflected in the pooled estimate of the five anti-oxidant trials (OR 1.21; 95% CI = 0.81 to 1.83) (Fig 4.5) Combined data from the two large skin cancer trials64, 79 only (data not shown), provided some evidence of a small reduction in the risk of second primary cancer associated with the antioxidant intervention (pooled OR = 0.89, 95% CI = 0.62 to 1.27), although there was also evidence of considerable between trial heterogeneity ( I2 = 77.1 %). There was no consistent effect of healthy diet interventions on second primary cancer occurrence. (Fig 4.5).

4.3 Development of Cancer in Patients With Pre-invasive Lesions The three healthy diet interventions109, 125, 142 included a fibre component, and the combined effect estimate suggested an increased risk of progression from colorectal adenomas to malignancy (OR = 1.96, 95% CI = 0.66 to 5.80). (Fig 4.6, top). Data from the multivitamin, β -carotene trials and lycopene trials62, 119, 145 were combined to investigate the effect of antioxidants on the development of invasive cancer from preinvasive lesions with a pooled estimate of OR = 0.59 (95% CI = 0.23 to 1.51) in patients with preinvasive lesion in the oesophagus, stomach or prostate. (Fig 4.6, middle). Two trials in patients with colorectal110 and oesophageal61 pre-invasive lesions evaluating the effect of calcium versus placebo on the risk of developing cancer also found little evidence of effect (OR = 0.89; 95% CI = 0.38 to 2.05). (Fig 4.6, bottom). 4.4 Recurrence of Pre-invasive Lesions Six healthy diet interventions55, 57, 109, 121, 125, 142 that all included participants with colorectal pre-invasive lesions and aimed to increase fibre intake showed little evidence of an effect on the risk of recurrence of colorectal polyps (OR = 1.07, 95% CI = 0.89 to 1.28) ( I2 = 38.7%). (Fig 4.7, top) After combining data from the two β -carotene trials56, 57, four multivitamin trials56, 58, 115, 122, and the trial58 containing a multivitamin arm and an N-acetylcysteine arm, there was weak evidence of a reduction in risk of colorectal polyps with these antioxidant interventions (OR = 0.63, 95% CI = 0.36 to 1.12) (I2= 78.2%), and smaller studies reported extreme effects. (Fig 4.7, middle).

26

Fig 4.7 also shows that two calcium interventions55, 110 showed some evidence of a reduced risk of recurrence of colorectal polyps (OR = 0.74, 95% CI = 0.59 to 0.94). The two folate trials117, 123 were too small to allow treatment effects to be estimated with precision (OR = 0.58, 95% CI = 0.19 to 1.81).

27

RESULTS - FIGURES Potentially relevant papers identified and screened for retrieval (n = 7061) Papers excluded because they did not meet the study inclusion criteria (n = 6518) Papers retrieved for more detailed evaluation (n = 543)

Potentially appropriate RCTs to be included in the meta-analysis (n = 176)

Papers excluded because they did not meet the study inclusion criteria (n = 367) 6 reported interim findings or were abstracts superseded by end of trial reports 15 did not include cancer or pre-cancer survivors 121 were not randomized controlled trials 127 did not include a nutritional or physical activity intervention (21 of these trials were physiotherapy, relaxation or psychological interventions and 24 trials include PSK supplement interventions) 2 had no suitable control group 5 did not report individually-randomized data 1 included pre-malignant and healthy participants but did not analyse the data for pre-malignant groups separately 25 were in-hospital, peri-operative intervention feeding trials aimed at improve surgical outcomes (including enteral and parenteral feeding) 65 were synthetic vitamin derivatives or vitamin analogs

RCTs excluded from the meta-analysis because they had no relevant outcome data (n = 98) (Appendix D)

RCTs to be included in the metaanalysis (n = 78)

RCTs including physical activity interventions with usable outcome data (n = 13) RCTs including nutritional interventions with usable outcome data (n = 62) RCTs including combined nutritional /physical activity interventions with usable outcome data (n = 3)

Figure 4.1 Study selection process. 28

Patients with cancer: all-cause mortality

Inte rv'n S tud y

In terven tio n

Con tro l

co ntrol

S ite

S tage

Event s/N E ve nts/ N

Ef fe ct (9 5% CI)

_ __ ___ __ _ He althy d ie t _ __ ___ __ __ B erglund

DA+exercis e

UT

B r/O v/ Ur/Oth

UC

1/9 8

0 /10 1

3. 12 (0 .13 , 77 .5 9)

B la ck

DA

UT

Sk

UC

1/5 7

2 /58

0. 50 (0 .04 , 5. 67 )

E lk o rt

DA

UT

Br

I, II, III or IV

4/2 4

3 /23

1. 33 (0 .26 , 6. 74 )

E va ns

DA+LDS

UT

Col

A

19/ 21

2 7/3 3

2. 11 (0 .38 , 11 .6 1)

E va ns

DA+LDS

UT

Lu

A

24/ 30

3 5/3 6

0. 11 (0 .01 , 1. 01 )

Ove sen

DA+/-LDS

UT

B r/L u/Ov

UC

18/ 57

1 3/4 8

1. 24 (0 .53 , 2. 90 )

d e Wa ard (a)

DA

UT

Br

UC

3/3 0

3 /24

0. 78 (0 .14 , 4. 25 )

d e Wa ard (b)

DA

UT

Br

UC

2/2 9

4 /19

0. 28 (0 .05 , 1. 70 ) 0. 89 (0 .49 , 1. 63 )

S ub tota l (I-squ ared = 8 .3%) _ __ ___ __ _ Antioxida nts __ ___ __ ___ 108 /6 53

1 29/ 65 9

0. 81 (0 .61 , 1. 08 )

60/ 60

6 2/6 3

2. 90 (0 .12 , 72 .6 9)

E

79/ 91 3

7 2/8 92

1. 08 (0 .77 , 1. 51 )

Bl

E

8/3 5

8 /30

0. 81 (0 .26 , 2. 52 )

P

H&N

I or II

29/ 12 8

3 3/1 25

0. 82 (0 .46 , 1. 45 )

P

Col

A

26/ 51

2 6/4 9

0. 92 (0 .42 , 2. 02 )

Lu

UC

54/ 64

6 4/7 2

0. 68 (0 .25 , 1. 83 )

UT

H&N

UC

9/1 04

1 5/1 10

0. 60 (0 .25 , 1. 44 )

UT

H&N/Lu

I, II or III

328 /1 285 3 04/ 12 88

E

Clark

Se

P

Sk

Cre ag an

V it C

P

Col/UGI/Lu /Oth A

Gre enb erg

B -carot ene

P

Sk

L am m

MV+Zn

UT

Mayne

B -carot ene

Moe rte l

V it C

P ath ak

MV+Se,Cu, Zn UT

Tom a

B -carot ene

va n Zan dwijk

NAC

1. 11 (0 .93 , 1. 33 ) 0. 99 (0 .87 , 1. 13 )

S ub tota l (I-squ ared = 0 .0%) _ __ ___ __ ___ _ Re tin ol _ __ ___ _____ _ Meyskens

V it A

UT

L eu

UC

45/ 57

5 7/6 7

0. 66 (0 .26 , 1. 66 )

Meyskens

V it A

UT

Sk

I

40/ 11 9

4 3/1 21

0. 92 (0 .54 , 1. 56 )

P astorino

V it A

UT

Lu

I

55/ 15 0

6 4/1 57

0. 84 (0 .53 , 1. 33 )

va n Zan dwijk

V it A

UT

H&N/Lu

I, II or III

318 /1 290 3 14/ 12 83

1. 01 (0 .84 , 1. 21 ) 0. 97 (0 .83 , 1. 13 )

S ub tota l (I-squ ared = 0 .0%)

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.2. Meta-analyses examining the effect of dietary interventions on all-cause mortality in trials of patients with cancer Key to abbreviations: CI = confidence interval. DA = dietary advice toward a healthier optimal diet; LDS = liquid diet supplement; MV = multivitamins; NAC = N -acetylcysteine; Se = selenium; Zn = zinc; Cu = copper. P = placebo; UT = usual treatment. Bl = bladder; Br = breast; Col = colorectal; H&N = head and neck; Leu = leukemia; Lu = lung; Oth = other (not specified in text); Ov = ovary; Sk = skin; UGI = upper gastrointestinal; Ur = urological. A = advanced; E = early; UC = unclear. Notes: In Berglund et al.52 one-third of control subjects were offered dietary advice, and two-thirds of control subjects were offered usual treatment. In De Waard et al. (a)63, locoregional disease was present in 28% of the study population. Numbers were not reported separately for intervention arms. Also in De Waard et al. (b)63, locoregional disease was present in 94% of the study population. Numbers were not reported separately for intervention arms. In Lamm et al.74, the percentages of participants with Ta, T1, and T2 disease were 66%, 23%, 11% (intervention arm) and 67%, 20%, 13% (control arm), respectively. In Ovesen et al.76, the percentage ratio of local/extensive cancer was 56 : 44 (intervention arm) and 63 : 37 (control arm). The trial of Evans et al.82 contained the following two intervention arms: 1) dietary advice + liquid diet supplement + zinc + magnesium and 2) dietary advice + liquid diet supplement. To get the most accurate estimate of the effect of dietary advice on all-cause mortality, only intervention arm 2 was compared with usual treatment. For this trial, the liquid dietary supplement was administered enterally or parenterally depending on the calorie intake and patient tolerance. A factorial trial by van Zandwijk et al.60 included the following arms: 1) vitamin A, 2) N –acetylcysteine (NAC), 3) vitamin A + NAC, and 4) usual treatment. The effect of vitamin A was assessed by comparing arms 1 and 3 versus arms 2 and 4, and the effect of NAC was assessed by comparing arms 2 and 3 versus arms 1 and 4.

29

Patients with cancer: Cancer mortality

Interv'n Study

Intervention control Si te

Stage

Control

Events/N Events/N

Effect (95% CI)

_________ Healthy diet __________ Elkort

DA

UT

Br

I, II, III or IV 2/24

2/23

0.95 (0.12, 7.41)

de Waard (a)

DA

UT

Br

UC

1/30

2/24

0.38 (0.03, 4.46)

de Waard (b)

DA

UT

Br

UC

2/29

3/19

0.40 (0.06, 2.62)

Subtotal (I-squared = 0.0%)

0.53 (0.16, 1.79)

_________ Antioxidants __________ Duffield- Lillico

Se

40/621

66/629

0.59 (0.39, 0.88)

Mayne

B-carotene P

H&N I or II

20/128

16/125

1.26 (0.62, 2.56)

Toma

B-carotene UT

H&N UC

5/104

6/110

0.88 (0.26, 2.96)

P

Sk

E

Subtotal (I-squared = 42.1%)

0.80 (0.47, 1.35)

____________ Retinol ____________ Kucera

Vit A

UT

Cer

A

5/21

5/21

1.00 (0.24, 4.14)

Meyskens

Vit A

UT

Sk

I

35/119

34/121

1.07 (0.61, 1.86)

Pastorino

Vit A

UT

Lu

I

44/150

53/157

0.81 (0.50, 1.32)

Subtotal (I-squared = 0.0%)

0.92 (0.65, 1.31)

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.3 Meta-analyses examining the effect of dietary interventions on cancer mortality in trials of patients with cancer Key to abbreviations: CI = confidence interval. DA = dietary advice toward a healthier optimal diet; Se = selenium; Zn = zinc; Cu = copper. P = placebo; UT = usual treatment. Br = breast; Cer = cervical; H&N = head and neck; Lu = lung; Sk = skin; A = advanced; E = early; UC = unclear. Notes: In De Waard et al. (a)63), locoregional disease was present in 28% of the study population. Numbers were not reported separately for intervention arms. Also in De Waard et al. (b)63, loco-regional disease was present in 94% of the study population. Numbers were not reported separately for intervention arms.

30

Patients with cancer: cancer recurrence

Study Intervention _________ Healthy diet __________ Elkort DA Sopotsinskaya DA de Waard (a) DA de Waard (b) DA Subtotal (I-squared = 65.7%)

Interv'n Control control Site Stage Events/NEvents/N UT UT UT UT

Effect (95% CI)

Br Br Br Br

I UC UC UC

4/12 4/56 2/30 2/29

1/14 15/54 0/24 2/19

6.50 (0.61, 68.96) 0.20 (0.06, 0.65) 4.30 (0.20, 93.90) 0.63 (0.08, 4.90) 1.03 (0.18, 5.93)

_________ Antioxidants __________ Bairati B-carotene+Vit EP Bairati Vit E P Clark Se P Lamm MV+Zn UT Mayne B-carotene P Toma* B-carotene UT Subtotal (I-squared = 70.0%)

H&N H&N Sk Bl H&N H&N

UC UC E E I or II UC

18/79 35/194 218/511 14/35 16/128 9/104

16/77 28/190 190/537 24/30 21/125 14/110

1.13 (0.53, 2.41) 1.27 (0.74, 2.19) 1.36 (1.06, 1.74) 0.17 (0.05, 0.51) 0.71 (0.35, 1.43) 0.65 (0.27, 1.57) 0.85 (0.53, 1.36)

____________ Retinol ____________ Jyothirmayi Vit A P Kucera Vit A UT Meyskens Vit A UT Pastorino Vit A UT Subtotal (I-squared = 0.0%)

H&N Cer Sk Lu

UC A I I

11/56 5/21 43/119 38/150

5/50 7/21 47/121 46/157

2.20 (0.71, 6.84) 0.63 (0.16, 2.42) 0.89 (0.53, 1.50) 0.82 (0.49, 1.35) 0.91 (0.65, 1.27)

__________ Vitamin B6 ___________ Byar Vit B6 P Newling Vit B6 P Subtotal (I-squared = 45.6%)

Bl Bl

I I

15/32 29/48 82/126 77/126

0.58 (0.23, 1.43) 1.19 (0.71, 1.98) 0.92 (0.47, 1.80)

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.4 Meta-analyses examining the effect of dietary interventions on cancer recurrence in trials of patients with cancer. Key to abbreviations: Int. = intervention; CI = confidence interval. DA = dietary advice toward a healthier optimal diet; MV = multivitamins; Se = selenium; Zn = zinc. UT = usual treatment; P = placebo. Bl = bladder; Br = breast; Cer = cervix; H&N = head & neck; Lu = lung; Sk = skin. A = advanced; E = early; UC = unclear. Notes: Part of the intervention in Bairati et al.’s trial129 was withdrawn part way through the trial on ethical grounds, based on the findings of previous studies. Results are therefore given for the combined intervention of B carotene and vitamin E for those participants and the control group up until B-carotene was withdrawn. The trial continued with just Vitamin E. In De Waard et al. (a)63, loco-regional disease was present in 28% of the study population. Numbers were not reported separately for intervention arms. In De Waard et al. (b)63, loco-regional disease was present in 94% of the study population. Numbers were not reported separately for intervention arms. In Jyothirmayi et al.86 , participants were already treated and in complete remission of lesions. In Lamm et al.74 , the percentages of participants with Ta, T1, and T2 disease were 66%, 23%, 11% (intervention arm) and 67%, 20%, 13% (control arm), respectively. In Sopotsinskaya et al.77, participants had to have sufficiently high levels of blood leukocytes to suggest that they had successfully recovered from six months of chemotherapy. *Results for Toma et al.135 are from a combined category of both cancer recurrence and 2nd primary cancer

31

Patients with cancer: second primary cancer occurrence

Interv'n Control Study

Intervention

control Site Stage Events/NEvents/N

Effect (95% CI)

_________ Healthy diet __________ de Waard (a)

DA

UT

Br

UC

2/30

0/24

4.30 (0.20, 93.90)

de Waard (b)

DA

UT

Br

UC

1/29

1/19

0.64 (0.04, 10.94)

Subtotal (I-squared = 0.0%)

1.53 (0.19, 12.37)

_________ Antioxidants __________ Bairati

B-carotene+Vit EP

H&N UC

10/79

Bairati

Vit E

P

H&N UC

38/194 14/190

3.06 (1.60, 5.86)

Duffield-Lillico

Se

P

Sk

E

105/621 137/629

0.73 (0.55, 0.97)

Greenberg

B-carotene

P

Sk

E

362/878 340/850

1.05 (0.87, 1.28)

Mayne

B-carotene

P

H&N I or II 17/128 14/125

7/77

1.45 (0.52, 4.03)

1.21 (0.57, 2.58)

Subtotal (I-squared = 76.6%)

1.21 (0.81, 1.83)

____________ Retinol ____________ Jyothirmayi

Vit A

P

H&N UC

0/56

Pastorino

Vit A

UT

Lu

18/150 29/157

I

2/50

0.17 (0.01, 3.66) 0.60 (0.32, 1.14)

Subtotal (I-squared = 0.0%)

0.57 (0.31, 1.07)

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.5 Meta-analyses examining the effect of dietary interventions on second primary cancer occurrence in trials of patients with cancer. Key to abbreviations: Int. = intervention; CI = confidence interval. DA = dietary advice toward a healthier optimal diet; MV = multivitamins; Se = selenium; Zn = zinc. UT = usual treatment; P = placebo. Bl = bladder; Br = breast; Cer = cervix; H&N = head & neck; Lu = lung; Sk = skin. A = advanced; E = early; UC = unclear. Notes: Part of the intervention in Bairati et al.’s trial129 was withdrawn part way through the trial on ethical grounds, based on the findings of previous studies. Results are therefore given for the combined intervention of B carotene and vitamin E for those participants and the control group up until B-carotene was withdrawn. The trial continued with just Vitamin E. In De Waard et al. (a)63, loco-regional disease was present in 28% of the study population. Numbers were not reported separately for intervention arms. In De Waard et al. (b)63, loco-regional disease was present in 94% of the study population. Numbers were not reported separately for intervention arms. In Jyothirmayi et al.86 , participants were already treated and in complete remission of lesions. In Lamm et al.74 , the percentages of participants with Ta, T1, and T2 disease were 66%, 23%, 11% (intervention arm) and 67%, 20%, 13% (control arm), respectively. In Sopotsinskaya et al.77, participants had to have sufficiently high levels of blood leukocytes to suggest that they had successfully recovered from six months of chemotherapy. *Results for Toma et al.135 are from a combined category of both cancer recurrence and 2nd primary cancer

32

Patients with pre-invasive lesions: development of cancer

Interv'n Study

Intervention control Site

Control

Events/N Events/N

Effect (95% CI )

_________ Healthy diet __________ Alberts

Fibre

Fibre

Col

7/719

2/584

2.86 (0.59, 13.82)

Ishikawa

High fibre

UT

Col

0/191

2/189

0.20 (0.01, 4.11)

Schatzkin

DA

UT

Col

10/1037 4/1042

2.53 (0.79, 8.08)

Subtotal (I-squared = 22.5%)

1.96 (0.66, 5.80)

_________ Antioxidants __________ Li

MV

P

Os

123/1657 128/1661

0.96 (0.74, 1.24)

Mohanty

Lycopene

UT

Pro

2/20

6/20

0.26 (0.05, 1.49)

Zhu

B-carotene P

Sto

2/118

3/54

0.29 (0.05, 1.81)

Subtotal (I-squared = 45.1%)

0.59 (0.23, 1.51)

____________ Calcium ____________ Baron

Ca

P

Col

3/454

5/459

0.60 (0.14, 2.54)

W ang

Ca

P

Os

8/82

8/88

1.08 (0.39, 3.03)

Subtotal (I-squared = 0.0%)

0.89 (0.38, 2.05)

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.6. Meta-analyses examining the effect of dietary interventions development from pre-invasive lesions to cancer in trials of patients with pre-invasive lesions. Key to abbreviations: Int. = intervention; CI = confidence interval. DA = Dietary advice toward a healthier optimal diet; Ca = calcium; MV = multivitamins; NAC = N -acetylcysteine. P = placebo; UT = usual treatment. Col = colorectal; Os = Oesophagus; Pro = prostate; Sto = stomach.

33

Patients with pre-invasive lesions: recurrence of pre-invasive lesions

Intervention Study _________ Healthy diet __________ High fibre Alberts Fi bre Bonithon-Kopp High fibre Ishikawa DA and/or fibre MacLennan DA+high fibre Mc Keown-Eyssen DA Schatzkin Subtotal (I-squared = 38.7%)

control

Site

Col Low fibre Col P Col UT Col P UT + l ow fibre Col Col UT

Interv'n Control E vents/N Events/N

Effect (95% CI)

338/719 58/198 119/191 65/291 17/78 380/958

299/584 36/178 106/189 21/99 16/87 374/947

0.85 (0.68, 1.05) 1.63 (1.01, 2.63) 1.29 (0.86, 1.95) 1.07 (0.61, 1.86) 1.24 (0.58, 2.65) 1.01 (0.84, 1.21) 1.07 (0.89, 1.28)

68/187 36/51 36/192 34/67 28/78 10/34

1.05 (0.74, 1.47) 0.42 (0.18, 0.98) 1.46 (0.90, 2.38) 0.69 (0.35, 1.35) 0.11 (0.04, 0.33) 0.61 (0.24, 1.57) 0.63 (0.36, 1.12)

_________ Antioxidants __________ Vit C, E and/or B-caro.P Greenberg P MV+minerals Hofstad P B-carotene MacLennan P Vit C, E Mc Keown-Eyssen UT Vit A,C,E Ponz de Leon UT Vit A,C,E or NAC Ponz de Leon Subtotal (I-squared = 78.2%)

Col Col Col Col Col Col

211/564 21/42 50/198 29/70 4/70 14/69

____________ Calcium ____________ Ca Baron Ca Bonithon-Kopp Subtotal (I-squared = 0.0%)

P P

Col Col

196/454 232/459 28/176 36/178

____________ Folate _____________ Folate Kim Folate Paspatis Subtotal (I-squared = 5.0%)

P P

Col Col

5/8 4/31

0.74 (0.57, 0.96) 0.75 (0.43, 1.29) 0.74 (0.59, 0.94)

1.33 (0.19, 9.31) 0.39 (0.10, 1.47) 0.58 (0.19, 1.81)

5/9 8/29

0.1

0.5 1 0.5

10

Odds Ratio Fig. 4.7. Meta-analyses examining the effect of dietary interventions on recurrence of preinvasive lesions to cancer in trials of patients with pre-invasive lesions. Key to abbreviations: Int. = intervention; CI = confidence interval. DA = Dietary advice toward a healthier optimal diet; Ca = calcium; MV = multivitamins; NAC = N -acetylcysteine. P = placebo; UT = usual treatment. Col = colorectal; Es = esophagus; Sto = stomach. Notes: MacClennan et al.57used a 2 ´ 2 ´ 2 factorial trial of β -carotene, dietary advice, and high fi ber. In this comparison, arms including dietary advice or high fiber or both were combined and compared with those receiving neither. Table 4.11 gives details of numbers of events for each intervention separately. Greenberg et al.79 used a 2×2 factorial trial of vitamin C + E and β -carotene. We combined the three active interventions arms and compared with placebo. ¶ Ponz de Lyon et al.58 contained the following two antioxidant intervention arms: 1) vitamin A+ C + E; 2) N acetylcysteine, which were combined and compared with usual treatment

34

4.5 Quality of Life Data A total of thirty-three trials reported quality of life data: twelve nutritional intervention trials in cancer survivors66, 70, 76, 89, 130, 133, 134, 137, 138, 140, 141; one nutritional intervention trial in precancer survivors114, 18 trials testing physical activity interventions in cancer survivors.97-107, 146-152 , and two trials which contained a combination of physical activity and nutritional intervention components53, 54 The studies varied in their definition of outcome, the choice of instrument used to measure the outcome and the utility of the results in health decisionmaking. Despite the limited information available to assess the effect of nutrition and physical activity interventions on quality of life in cancer (and pre-cancer) survivors we feel it is an important outcome where efficacy in respect of survival is equivocal. The structure, properties, and format of the questionnaires varied and included both descriptive and preference-based questions. The instruments used included generic descriptive questionnaires, for example the Functional Assessment of Cancer Therapy (FACT) scale, the EORTC QLQ-C30 and the Medical Outcomes Short Form SF-36. Furthermore, a variety of instruments were used to assess the same outcome: for example, depression was assessed using the Back Depression Inventory, the Centre for Epidemiological Studies of Depression Scale (CESD) and the POMS scale. The instruments used are listed in Table 7.6. The trials assessed a broad range of dimensions from ‘Happiness’, ‘Hope and Power’, ‘body image’ and ‘fatigue’ to ‘anger’ ‘self-esteem’ and ‘loss of appetite’. Physiological outcomes were used in the physical activity studies which measured peak oxygen consumption and maximal physical performance. Each of the studies stated a hypothesis on the selection criterion for the quality of life scale used and reported that the scale had been validated and was reproducible. The time frame for assessing quality of life was documented in all studies. The results are shown in tables 7.5, 7.12 and 7.17. In the four trials investigating the effect of a nutrition-only intervention on quality of life in cancer survivors three found no effect66, 70, 76, 130, 133, 134 while one vitamin C trial reported higher numbers in the intervention group having appetite, strength and pain control but no difference between groups in activity levels.89, 130 One trial of high fat drink found a non-significant trend for improvement of psychological condition, physical condition, and general well-being, and a significant improvement in ‘leisure activities’ in the treatment group using the LASA instrument.140 One trial of dietary counselling reported a smaller decrease and faster recovery of QoL score in the nutrition group compared with the usual care group.141 One trial of dietary counselling found a significant improvement in ‘health action’ in the treatment group, but found no effect in the remaining eight elements or in the overall score.138 A trial of a commercial supplement consisting of branched chain amino acids and other micro-nutrients reported a significantly improved overall QoL score in the treatment group compared with the usual diet control group.137 One three-armed trial of individualised dietary counselling, or high protein supplement, compared with usual diet reported only change from baseline score to score at three-months for each group rather than between-group comparisons. In general the three groups showed similar patterns (full details are shown in Table 7.5).130 Only one trial investigated quality of life in pre-cancer survivors.114 This trial reported that an intervention aimed at a healthier dietary lifestyle had no effect on taste, ease of access to intervention foods, cost of foods, health assessment or life satisfaction but did lead to increased belief in food to improve health, health action and ‘feeling good’ compared with a control group given limited a dietary information. In contrast, the intervention appeared to reduce the social enjoyment of food compared with the control group. In the 18 physical activity-only trials the intensity and duration of the interventions varied between studies. Of the seven interventions which measured quality of life: five

35

observed improvements with physical activity (one using a linear analogue scale102, one using Functional Assessment of Chronic Illness Therapy-Fatigue Version IV (FACIT-F)147 and four using Functional Assessment of Cancer Therapy (either FACT-G or FACT-B)103, 104, 146, 149). However, one of these149 found no association when using the SF-36 scale; One trial using FACIT-F found a non-significant trend in improvement in the intervention group148, while one trial (also using FACT) showed no effect.98 Physical or functional capacity/ performance or aerobic capacity, reported in six trials, improved with physical activity in all trials.97, 98, 100103 Two trials using the Rosenberg Self-Esteem (RSE) inventory found that physical activity also improved self-esteem.99, 103 Two trials, both using the European Organization for Research and Treatment of Cancer Assessment Tool (EORTCQLQ-C30), reported only prepost intervention comparisons for each group, rather than between group comparisons. In one of these p values were given and patterns were generally similar in each group, although the treatment group (aerobic exercise) showed a significant improvement in maximal physical performance over time, which was not shown in the comparative (relaxation training) group.151 The other trial, which as well as using the EORTCQLQ-C30 scale also measured cardio-respiratory fitness, and used the HADS instrument to measure mental distress, reported mean changes in scores (with 95% confidence intervals) over time as well as modelling changes between groups over time. This trial found an improvement in cardiorespiratory fitness and a reduction in fatigue in the treatment group. No other elements of the scales showed any difference between the intervention and control groups.152 The only trial of a physical exercise intervention in children found no benefit of the intervention compared with the control group of no instruction and no physical therapy intervention.150 Finally, in the two interventions that included both physical activity and nutritional components one trial reported no effect on anxiety, depression, intrusion or avoidance54; whilst in the second trial the intervention improved physical strength, physical training, body image and fighting spirit but showed no effect on tiredness, global health, anxiety or depression.53 A number of problems were encountered when assessing quality of life; in particular, how to interpret the clinical significance of the measures used and how to aggregate the data reported. Although the instruments used to measure quality of life were valid and reproducible, there was no uniform approach to assessment and trials included both descriptive and numeric results. As a result, a measure of effect size across studies was not possible. Although it is of limited use we have, however, documented in the tables (tables 7.5, 7.12 and 7.17) the associations based on the reported probability value. 4.6 Side effects Side-effects are given in Table 7.7 (Side-effects of nutritional interventions in cancer survivors), Table 7.13 (Side-effects of nutritional interventions in pre-cancer survivors), and Table 7.18 (Side effects of physical activity interventions in cancer survivors). Data varied between detailed descriptions of all symptoms/side effects, and overall numbers of sideeffects experienced in each arm of a trial. Data presented should be interpreted with caution as it is not always clear whether all side-effects were reported or only those which the reporting authors considered relevant to the study. In addition, some trials gave numbers of participants experiencing side-effects, with others giving the number of side-event events and hence the number of events can exceed the number of participants.

36

5. DISCUSSION AND CONCLUSIONS The use of dietary modifications and dietary supplements in cancer management, generally initiated outside of conventional biomedical treatment, is widespread, as has recently been highlighted in the British (http://news.independent.co.uk/uk/health_medical/story.jsp?story=5865) and other (http://www.cnn.com/HEALTH/library/CM/00002.html) media. However the unproven nature of such therapies has also been highlighted in recent popular news coverage. This systematic review supports a sceptical view, but largely through lack of robust evidence. In general it is not possible to exclude meaningful benefit or harm of such treatments. The main conclusion of this review must be that large-scale trials with longer follow-up are required in all the areas that have been studied. With respect to the specific issues for which there were data that at least allow for an imprecise estimate of effects, trials of a generic “healthy” diet - those that generally counsel reductions of fat intake and increases of fibre, fruit and vegetable intake, sometimes with vitamin and mineral supplements - yielded an odds ratio of 0.89 (95%CI 0.49 to 1.63) when all cancers were combined and all-cause mortality was examined. Such a benefit would be useful, but considerably more data are required to obtain a reliable effect estimate. Furthermore it seems unlikely that such dietary approaches would have an equal or even same direction effect on different cancer types. When sub-dividing by cancer type, data were even sparser, and for colon cancer and breast cancer the effect estimates were too imprecise to be of use with respect to mortality. When examining more specific dietary intervention-outcome associations the data were equally limited, although two underpowered but potentially important findings emerged. First, dietary fibre supplementation interventions were associated with an apparent increase in the progression from colorectal pre-malignancies to colon cancer (odds ratio 1.96; 95% CI 0.66 to 5.80). This finding – which runs against observational epidemiological data suggesting that dietary fibre reduces the risk of colon cancer – is not definitive, given study power, and given the strength of the observational data a large trial is probably justified, although it may be difficult to mount given these findings. In the opposite direction three trials of dietary advice to reduce calorie intake in women with breast cancer provided a pooled odds ratio for cancer mortality of 0.53 (95% CI 0.016 to 1.79), but clearly data are limited. However further trials of calorie and fat reduction in women with breast cancer are clearly worthwhile, and such trials are ongoing. These results are supported by those assessing all cause mortality findings, which also show a reduction, although very imprecisely estimated. Calcium supplementation was associated with reduced recurrence of pre-malignant colorectal adenomas in two trials (OR 0.74; 95% CI 0.59 to 0.94), again indicating that further large trials would be valuable. For antioxidant vitamin supplementation the data are equally sparse. The one interpretable vitamin C trial was in patients with advanced cancer, when an effect may not reasonably be expected and none was seen (OR = 0.92; 95% CI = 0.42 to 2.02).83 For vitamin E supplementation data are equally limited. Beta-carotene supplementation had been tested in larger numbers. In three trials (one of patients with skin cancer79 and two of patients with head and neck cancer67, 135) all-cause mortality was unaltered although each estimate was imprecise. Similarly, recurrences of pre-malignancies in skin cancer were little altered, but with wide confidence intervals. Data from a small study in mouth cancer were promising but too sparse to be reliable (0/21 recurrences in treatment group, 7/33 recurrences in comparison

37

group).126 Vitamin A supplementation and multivitamin supplementation also produced no meaningful effects on all-cause mortality. We also examined physical activity, but the studies that reported mortality outcomes in this regard were too small to be informative. Similarly, quality of life outcomes in both physical activity and diet or nutrition intervention trials could not be reliably estimated. In this case, however, lack of power was not entirely the issue, rather there was great heterogeneity in the quality of life measures that were recorded, and this militated against any useful summary estimates. The conclusions of this review are, therefore, mostly negative. The trials are generally small and of low quality and impact on cancer survival and quality of life cannot be reliably estimated. Rather than conducting more small, diverse trials, the main recommendation should be that large-scale, high quality trials are carried out for the most promising interventions. Currently these would be calorie and fat restriction in breast cancer and calcium in colorectal pre-malignancies, and such trials are ongoing. A large-scale exercise trial in a cancer site for which there is some observational evidence to suggest this may have benefit (colorectal, breast or prostate in particular) would also be justified. It should not be assumed that nutritional or exercise interventions are benign. Exercise interventions can lead to fatigue, muscle strain and fractures. Nutritional interventions, noticeably antioxidant supplementation in the primary prevention setting, have yielded unexpected negative effects, particularly with respect to beta-carotene supplementation and lung cancer. Fibre interventions with colorectal pre-malignancies have also produced a worrying, though imprecisely estimated, adverse effect. Therefore the notion that nutritional interventions can be promoted because at least they will do no harm should not be maintained. Funding agencies should consider that backing the establishment of large-scale and informative trials is a priority.

38

6. References (1) Cancer Facts and Figures 2004. Atlanta, GA: American Cancer Society; 2004. (2) Brown JK, Byers T, Doyle C et al. Nutrition and physical activity during and after cancer treatment: An American Cancer Society Guide for informed choices. CA Cancer J Clin 2003;. 53(5). (3) Ries LAG, Eisner MP, Kosary CL et al. SEER Cancer Statistics Review, 19752001. National Cancer Institute Bethesda, MD 2006;Available at: URL: http://seer.cancer.gov/csr/1975_2001/, 2004. (4) WRCF/AICR. Food, Nutrition and the Prevention of Cancer: A Global Perspective. 1997. Washington, DC: American Institute for Cancer Research. Ref Type: Serial (Book,Monograph) (5) Ernst E, Resch K, Mills S. Complementary medicine - a definition. British Journal of General Practice 1995;45:506. (6) Ernst E, Cassileth BR. The prevalence of complementary/alternative medicine in cancer: A systematic review. Cancer 1998;. 83(4). (7) Eisenberg DM, Davis RB, Ettner SL et al. Trends in alternative medicine use in the United States, 1990-1997: results of a follow-up national survey. Jama 1998 November 11;280(18):1569-75. (8) Eisenberg DM, Davis RB, Ettner SL et al. Trends in alternative medicine use in the United States, 1990-1997: results of a follow-up national survey. Jama 1998 November 11;280(18):1569-75. (9) Eisenberg DM, Davis RB, Ettner SL et al. Trends in alternative medicine use in the United States, 1990-1997: results of a follow-up national survey. Jama 1998 November 11;280(18):1569-75. (10) Eisenberg DM, Davis RB, Ettner SL et al. Trends in alternative medicine use in the United States, 1990-1997: results of a follow-up national survey. Jama 1998 November 11;280(18):1569-75. (11) Albanes D, Heinonen OP, Huttunen JK et al. Effects of alpha-tocopherol and betacarotene supplements on cancer incidence in the Alpha-Tocopherol Beta-Carotene Cancer Prevention Study. Am J Clin Nutr 1995;. 62(6 SUPPL.). (12) Schmidt K, Ernst E. Assessing websites on complementary and alternative medicine for cancer. Annals of Oncology 2004;. 15(5). (13) Egger M, Davey SG, Altman D. Systematic Reviews in Health Care. Meta-analysis in context. 2nd ed. London: BMJ Publishing Group; 2001. (14) Moher D, Cook DJ, Eastwood S et al. Improving the quality of reports of metaanalyses of randomised controlled trials: the QUOROM statement. The Lancet 1999;354(9193):1896-900.

39

(15) American Institute for Cancer Research. Dietary Options for Cancer Survivors. Washington; 2002. (16) American Institute for Cancer Research. Nutrition After Cancer. Washington; 2002. (17) Asano T, McLeod RS, Asano T, McLeod RS. Dietary fibre for the prevention of colorectal adenomas and carcinomas. Cochrane Database Syst Rev 2002;(2):CD003430. (18) Asano TK, McLeod RS. Vitamins and minerals for the prevention of colorectal adenomas and carcinomas (Protocol for a Cochrane Review). The Cochrane Library, Oxford 2003;Issue 3. (19) Austin S. Recent progress in treatment and secondary prevention of breast cancer with supplements. Alternative Medicine Review 1997 Jan;2(1):4-11. (20) Batist G. Clinical evaluation of vitamin C and other micronutrients in the treatment of cancer. Journal of Orthomolecular Medicine 2000;. 15(4). (21) Courneya KS, Courneya KS. Exercise interventions during cancer treatment: biopsychosocial outcomes. Exercise & Sport Sciences Reviews 2001 April;29(2):604. (22) Demark-Wahnefried W, Morey MC, Clipp EC et al. Leading the Way in Exercise and Diet (Project LEAD): intervening to improve function among older breast and prostate cancer survivors. Control Clin Trials 2003 April;24(2):206-23. (23) Dutta A, Dutta SK. Vitamin E and its role in the prevention of atherosclerosis and carcinogenesis: A review. J Am Coll Nutr 2003;22(4):258-68. (24) Fairey AS, Courneya KS, Field CJ, Mackey JR. Physical exercise and immune system function in cancer survivors: a comprehensive review and future directions. [Review] [69 refs]. Cancer 2002 January 15;94(2):539-51. (25) Faivre J, Bonithon-Kopp C, Faivre J, Bonithon-Kopp C. Effect of fibre and calcium supplementation on adenoma recurrence and growth. IARC Sci Publ 2002;156:45761. (26) Friedenreich CM, Courneya KS. Exercise as rehabilitation for cancer patients. Clinical Journal of Sport Medicine 1996;6(4):237-44. (27) Gaynor ML. Isoflavones and the prevention and treatment of prostate disease: Is there a role? Cleve Clin J Med 2003;70(3):203-16. (28) Jacobson JS, Workman SB, Kronenberg F. Research on complementary/alternative medicine for patients with breast cancer: A review of the biomedical literature. J Clin Oncol 2000;18(3):668-83. (29) Karki A, Simonen R, Malkia E, Selfe J. Efficacy of physical therapy methods and exercise after a breast cancer operation: A systematic review. Critical Reviews in Physical & Rehabilitation Medicine 2001;. 13(2-3).

40

(30) Macrae F. Wheat bran fiber and development of adenomatous polyps: Evidence from randomized, controlled clinical trials. Am J Med 1999;106(1 A):38S-42S. (31) Messina MJ, Loprinzi CL. Soy for breast cancer survivors: a critical review of the literature. J Nutr 2001 November;131(11 Suppl):3095S-108S. (32) Moyad MA. Results and lessons from clinical trials using dietary supplements for cancer: Direct and indirect investigations. Semin Urol Oncol 2001;19(4):232-46. (33) Peehl DM, Feldman D. The role of vitamin D and retinoids in controlling prostate cancer progression. Endocr Relat Cancer 2003 June;10(2):131-40. (34) Prasad KN, Kumar A, Kochupillai V, Cole WC. High doses of multiple antioxidant vitamins: essential ingredients in improving the efficacy of standard cancer therapy. [Review] [71 refs]. J Am Coll Nutr 1999 February;18(1):13-25. (35) Rissanen A, Fogelholm M. Physical activity in the prevention and treatment of other morbid conditions and impairments associated with obesity: Current evidence and research issues. Med Sci Sports Exerc 1999;31(11 SUPPL.):S635-S645. (36) Rock CL, Demark-Wahnefried W. Can lifestyle modification increase survival in women diagnosed with breast cancer? J Nutr 2002;132(11 SUPPL.):3504S-7S. (37) Rock CL. Carotenoid update. J Am Diet Assoc 2003 April;103(4):423-5. (38) Saxe GA, Rock CL, Wicha MS, Schottenfeld D. Diet and risk for breast cancer recurrence and survival. Breast Cancer Res Treat 1999 February;53(3):241-53. (39) Sengupta S, Tjandra JJ, Gibson PR. Dietary fiber and colorectal neoplasia. Dis Colon Rectum 2001;44(7):1016-33. (40) Smorenburg SM, Vink R, Otten HM, Swaneveld F, Buller HR. The effects of vitamin K-antagonists on survival of patients with malignancy: a systematic analysis.[comment]. [Review] [10 refs]. Thromb Haemost 2001 December;86(6):1586-7. (41) Solà I, Thompson E, Subirana M, López C, Pascual A. Non-invasive interventions for improving well-being and quality of life in patients with lung cancer. Solà I, Thompson E, Subirana M, López C, Pascual A Non invasive interventions for improving well being and quality of life in patients with lung cancer The Cochrane Database of Systematic Reviews: Reviews 2004 Issue 4 Chichester, UK: John Wiley & Sons, 2004;(4). (42) Tamayo C, Richardson MA. Vitamin C as a cancer treatment: State of the science and recommendations for research. Altern Ther Health Med 2003;9(3):94-102. (43) The Scientific Program Committee. Physical activity across the cancer continuum: report of a workshop: review of existing knowledge and innovative designs for future research. Cancer 2002 September 1;95(5):1134-43. (44) Thune I, Smeland S. Is physical activity important in treatment and rehabilitation of cancer patients? Tidsskr Nor Laegeforen 2000 November 10;120(27):3302-4.

41

(45) Norman HA, Butrum RR, Feldman E et al. The role of dietary supplements during cancer therapy. J Nutr 2003 November;133(11 Suppl 1):3794S-9S. (46) Courneya KS. Exercise in cancer survivors: an overview of research. Med Sci Sports Exerc 2003 November;35(11):1846-52. (47) Applebe GE, Wingfield J. Pharmacy Law and Ethics. 6th edition ed. London: Pharmaceutical Press; 1999. (48) Devereaux PJ, Manns BJ, Ghali WA et al. Physician interpretations and textbook definitions of blinding terminology in randomized controlled trials. Jama 2001 April 18;285(15):2000-3. (49) Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002 June 15;21(11):1539-58. (50) Galantino ML, Capito L, Kane RJ, Ottey N, Switzer S, Packel L. The effects of Tai Chi and walking on fatigue and body mass index in women living with breast cancer: a pilot study. Rehabil Oncol 2003;21:17-22. (51) Israel L, Hajji O, Grefft-Alami A et al. [Vitamin A augmentation of the effects of chemotherapy in metastatic breast cancers after menopause. Randomized trial in 100 patients]. [French]. Annales de Medecine Interne 1985;136(7):551-4. (52) Berglund G, Bolund C, Gustafsson UL, Sjoden PO. One-year follow-up of the 'Starting Again' group rehabilitation programme for cancer patients. Eur J Cancer 1994;30A(12):1744-51. (53) Berglund G, Bolund C, Gustavsson U-L, Sjoden P-O. A randomized study of a rehabilitation program for cancer patients: the 'starting again' group. Psychooncology 1994;3:109-20. (54) Petersson LM, Nordin K, Glimelius B, Brekkan E, Sjoden PO, Berglund G. Differential effects of cancer rehabilitation depending on diagnosis and patients' cognitive coping style. Psychosom Med 2002 November;64(6):971-80. (55) Bonithon-Kopp C, Kronborg O, Giacosa A, Rath U, Faivre J. Calcium and fibre supplementation in prevention of colorectal adenoma recurrence: A randomised intervention trial. Lancet 2000;356(9238):1300-6. (56) Greenberg ER, Baron JA, Tosteson TD et al. A clinical trial of antioxidant vitamins to prevent colorectal adenoma. Polyp Prevention Study Group.[comment]. New England Journal of Medicine 1994 July 21;331(3):141-7. (57) MacLennan R, Macrae F, Bain C et al. Effect of fat, fibre, and beta carotene intake on colorectal adenomas: further analysis of a randomized controlled dietary intervention trial after colonoscopic polypectomy. Asia Pac J Clin Nutr 1999;8(suppl):S54-S58. (58) Ponz dL, Roncucci L. Chemoprevention of colorectal tumors: role of lactulose and of other agents. Scandinavian Journal of Gastroenterology Supplement 1997;222:72-5.

42

(59) Sankaranarayanan R, Mathew B, Varghese C et al. Chemoprevention of oral leukoplakia with vitamin A and beta carotene: an assessment. Oral Oncology 1997 July;33(4):231-6. (60) van Zandwijk N, Dalesio O, Pastorino U et al. EUROSCAN, a randomized trial of vitamin A and N-acetylcysteine in patients with head and neck cancer or lung cancer. For the EUropean Organization for Research and Treatment of Cancer Head and Neck and Lung Cancer Cooperative Groups.[comment]. Journal of the National Cancer Institute 2000 June 21;92(12):977-86. (61) Wang LD, Zhou Q, Feng CW et al. Intervention and follow-up on human esophageal precancerous lesions in Henan, northern China, a high-incidence area for esophageal cancer. Gan To Kagaku Ryoho 2002 February;29 Suppl 1:159-72. (62) Zhu S, Mason J, Shi Y et al. The interventional effect of folic acid on the development of gastric and other gastrointestinal cancers - Clinical trial and followup for seven years. Chinese Journal of Gastroenterology 2002;7(2):73-8. (63) de Waard F, Ramlau R, Mulders Y, de Vries T, van Waveren S. A feasibility study on weight reduction in obese postmenopausal breast cancer patients. Eur J Cancer Prev 1993 May;2(3):233-8. (64) Duffield-Lillico AJ, Reid ME, Turnbull BW et al. Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomized clinical trial: a summary report of the Nutritional Prevention of Cancer Trial. Cancer Epidemiology, Biomarkers & Prevention 2002 July;11(7):630-9. (65) Van Patten CL, Olivotto IA, Chambers GK et al. Effect of soy phytoestrogens on hot flashes in postmenopausal women with breast cancer: a randomized, controlled clinical trial.[comment]. J Clin Oncol 2002 March 15;20(6):1449-55. (66) Persson CR, Johansson BB, Sjoden PO, Glimelius BL. A randomized study of nutritional support in patients with colorectal and gastric cancer. Nutr Cancer 2002;42(1):48-58. (67) Mayne ST, Cartmel B, Baum M et al. Randomized trial of supplemental betacarotene to prevent second head and neck cancer. Cancer Res 2001 February 15;61(4):1457-63. (68) Meng WC, Leung KL, Ho RL, Leung TW, Lau WY. Prospective randomized control study on the effect of branched-chain amino acids in patients with liver resection for hepatocellular carcinoma. Australian & New Zealand Journal of Surgery 1999 November;69(11):811-5. (69) Barton DL, Loprinzi CL, Quella SK et al. Prospective evaluation of vitamin E for hot flashes in breast cancer survivors. J Clin Oncol 1998 February;16(2):495-500. (70) Yu-Chung C, Fukuda T, Hamazoe R et al. Long-term oral administration of branched chain amino acids after curative resection of hepatocellular carcinoma: A prospective randomized trial. Br J Surg 1997;84(11):1525-31.

43

(71) Clark LC, Combs GF, Jr., Turnbull BW et al. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group.[comment][erratum appears in JAMA 1997 May 21;277(19):1520]. Jama 1996 December 25;276(24):1957-63. (72) Meyskens FL, Jr., Kopecky KJ, Appelbaum FR, Balcerzak SP, Samlowski W, Hynes H. Effects of vitamin A on survival in patients with chronic myelogenous leukemia: a SWOG randomized trial. Leukemia Research 1995 September;19(9):605-12. (73) Newling DW, Robinson MR, Smith PH et al. Tryptophan metabolites, pyridoxine (vitamin B6) and their influence on the recurrence rate of superficial bladder cancer. Results of a prospective, randomised phase III study performed by the EORTC GU Group. EORTC Genito-Urinary Tract Cancer Cooperative Group. Eur Urol 1995;27(2):110-6. (74) Lamm DL, Riggs DR, Shriver JS, vanGilder PF, Rach JF, DeHaven JI. Megadose vitamins in bladder cancer: a double-blind clinical trial. J Urol 1994 January;151(1):21-6. (75) Meyskens FL, Jr., Liu PY, Tuthill RJ et al. Randomized trial of vitamin A versus observation as adjuvant therapy in high-risk primary malignant melanoma: a Southwest Oncology Group study. J Clin Oncol 1994 October;12(10):2060-5. (76) Ovesen L, Allingstrup L, Hannibal J, Mortensen EL, Hansen OP. Effect of dietary counseling on food intake, body weight, response rate, survival, and quality of life in cancer patients undergoing chemotherapy: a prospective, randomized study. J Clin Oncol 1993 October;11(10):2043-9. (77) Sopotsinskaia, E.B., Balitskii et al. [Experience with the use of a low-calorie diet in breast cancer patients to prevent metastasis]. [RUSSIAN]. Vopr Onkol 1992;38(5):592-9. (78) Macia E, Moran J, Santos J, Blanco M, Mahedero G, Salas J. Nutritional evaluation and dietetic care in cancer patients treated with radiotherapy: prospective study. Nutrition 1991 May;7(3):205-9. (79) Greenberg, E.R., Baron et al. A clinical trial of beta carotene to prevent basal-cell and squamous-cell cancers of the skin. The Skin Cancer Prevention Study Group [published erratum appears in N Engl J Med 1991 Oct 31;325(18):1324] [see comments]. N Engl J Med 1990;323:789-95. (80) van der Merwe CF, Booyens J, Joubert HF, van der Merwe CA. The effect of gamma-linolenic acid, an in vitro cytostatic substance contained in evening primrose oil, on primary liver cancer. A double-blind placebo controlled trial. Prostaglandins Leukotrienes & Essential Fatty Acids 1990 July;40(3):199-202. (81) Arnold C, Richter MP. The effect of oral nutritional supplements on head and neck cancer. Int J Radiat Oncol Biol Phys 1989 June;16(6):1595-9.

44

(82) Evans WK, Nixon DW, Daly JM et al. A randomized study of oral nutritional support versus ad lib nutritional intake during chemotherapy for advanced colorectal and non-small-cell lung cancer. J Clin Oncol 1987 January;5(1):113-24. (83) Moertel CG, Fleming TR, Creagan ET, Rubin J, O'Connell MJ, Ames MM. Highdose vitamin C versus placebo in the treatment of patients with advanced cancer who have had no prior chemotherapy. A randomized double-blind comparison. New England Journal of Medicine 1985 January 17;312(3):137-41. (84) Elkort RJ, Baker FL, Vitale JJ, Cordano A. Long-term nutritional support as an adjunct to chemotherapy for breast cancer. JPEN J Parenter Enteral Nutr 1981 September;5(5):385-90. (85) Kucera H. [Adjuvanticity of vitamin A in advanced irradiated cervical cancer (author's transl)]. [Review] [123 refs] [German]. Wiener Klinische Wochenschrift Supplementum 1980;118:1-20. (86) Jyothirmayi R, Ramadas K, Varghese C, Jacob R, Nair MK, Sankaranarayanan R. Efficacy of vitamin A in the prevention of loco-regional recurrence and second primaries in head and neck cancer. European Journal of Cancer Part B, Oral Oncology 1996 November;32B(6):373-6. (87) Pastorino U, Infante M, Maioli M et al. Adjuvant treatment of stage I lung cancer with high-dose vitamin A.[comment]. J Clin Oncol 1993 July;11(7):1216-22. (88) Kokron O, Alth G, Cerni C et al. [Results of a comparative therapy study for inoperable lung cancer]. [German]. Onkologie 1982 February;5(1):20-2. (89) Creagan ET, Moertel CG, O'Fallon JR et al. Failure of high-dose vitamin C (ascorbic acid) therapy to benefit patients with advanced cancer. A controlled trial. New England Journal of Medicine 1979 September 27;301(13):687-90. (90) Byar D, Blackard C. Comparisons of placebo, pyridoxine, and topical thiotepa in preventing recurrence of stage I bladder cancer. Urology 1977 December;10(6):55661. (91) Delanian S, Porcher R, Balla-Mekias S, Lefaix JL. Randomized, placebo-controlled trial of combined pentoxifylline and tocopherol for regression of superficial radiation-induced fibrosis. J Clin Oncol 2003 July 1;21(13):2545-50. (92) Duffield-Lillico AJ, Dalkin BL, Reid ME et al. Selenium supplementation, baseline plasma selenium status and incidence of prostate cancer: an analysis of the complete treatment period of the Nutritional Prevention of Cancer Trial. BJU Int 2003 May;91(7):608-12. (93) Douglass HO, Jr., Milliron S, Nava H et al. Elemental diet as an adjuvant for patients with locally advanced gastrointestinal cancer receiving radiation therapy: a prospectively randomized study. JPEN J Parenter Enteral Nutr 1978 November;2(5):682-6.

45

(94) Jebb SA, Osborne RJ, Maughan TS et al. 5-fluorouracil and folinic acid-induced mucositis: no effect of oral glutamine supplementation. Br J Cancer 1994 October;70(4):732-5. (95) Ovesen L, Allingstrup L. Different quantities of two commercial liquid diets consumed by weight-losing cancer patients. JPEN J Parenter Enteral Nutr 1992 May;16(3):275-8. (96) Pickett M, Mock V, Ropka ME, Cameron L, Coleman M, Podewils L. Adherence to moderate-intensity exercise during breast cancer therapy. Cancer Pract 2002 November;10(6):284-92. (97) Schwartz, Cunningham AL. A randomized trial to examine effects of exercise on bone density and body composition in breast cancer patients receiving adjuvant chemotherapy. Breast Cancer Res Treat 2001;69(3):305. (98) Segal R, Evans W, Johnson D et al. Structured exercise improves physical functioning in women with stages I and II breast cancer: results of a randomized controlled trial. J Clin Oncol 2001 February 1;19(3):657-65. (99) Segar ML, Katch VL, Roth RS et al. The effect of aerobic exercise on self-esteem and depressive and anxiety symptoms among breast cancer survivors.[comment]. Oncol Nurs Forum 1998 January;25(1):107-13. (100) Dimeo F, Fetscher S, Lange W, Mertelsmann R, Keul J. Effects of aerobic exercise on the physical performance and incidence of treatment-related complications after high-dose chemotherapy. Blood 1997 November 1;90(9):3390-4. (101) MacVicar MG, Winningham ML, Nickel JL. Effects of aerobic interval training on cancer patients' functional capacity. Nurs Res 1989 November;38(6):348-51. (102) Burnham TR, Wilcox A. Effects of exercise on physiological and psychological variables in cancer survivors. Med Sci Sports Exerc 2002 December;34(12):1863-7. (103) Courneya KS, Mackey JR, Bell GJ, Jones LW, Field CJ, Fairey AS. Randomized controlled trial of exercise training in postmenopausal breast cancer survivors: cardiopulmonary and quality of life outcomes.[comment]. J Clin Oncol 2003 May 1;21(9):1660-8. (104) Segal RJ, Reid RD, Courneya KS et al. Resistance exercise in men receiving androgen deprivation therapy for prostate cancer.[comment]. J Clin Oncol 2003 May 1;21(9):1653-9. (105) Nieman DC, Cook VD, Henson DA et al. Moderate exercise training and natural killer cell cytotoxic activity in breast cancer patients. International Journal of Sports Medicine 1995 July;16(5):334-7. (106) Pinto BM, Clark MM, Maruyama NC, Feder SI. Psychological and fitness changes associated with exercise participation among women with breast cancer. Psychooncology 2003 March;12(2):118-26.

46

(107) Wall LM. Changes in hope and power in lung cancer patients who exercise. Nurs Sci Q 2000 July;13(3):234-42. (108) Cunningham BA, Morris G, Cheney CL, Buergel N, Aker SN, Lenssen P. Effects of Resistive Exercise on Skeletal muscle in Marrow Transplant Recipients Receiving Total Parenteral Nutrition. Journal of Parenteral & Enteral Nutrition 1986;10:55863. (109) Alberts DS, Martinez ME, Roe DJ et al. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians' Network.[comment]. New England Journal of Medicine 2000 April 20;342(16):1156-62. (110) Baron JA, Beach M, Mandel JS et al. Calcium supplements and colorectal adenomas. Polyp Prevention Study Group. Ann N Y Acad Sci 1999;889:138-45. (111) Baron JA, Beach M, Mandel JS et al. Calcium supplements for the prevention of colorectal adenomas. Calcium Polyp Prevention Study Group.[comment]. New England Journal of Medicine 1999 January 14;340(2):101-7. (112) Bayerl C, Schwarz B, Jung EG. A three-year randomized trial in patients with dysplastic naevi treated with oral beta-carotene. Acta Dermato-Venereologica 2003;83(4):277-81. (113) Childers JM, Chu J, Voigt LF et al. Chemoprevention of cervical cancer with folic acid: a phase III Southwest Oncology Group Intergroup study. Cancer Epidemiology, Biomarkers & Prevention 1995 March;4(2):155-9. (114) Corle DK, Sharbaugh C, Mateski DJ et al. Self-rated quality of life measures: effect of change to a low-fat, high-fiber, fruit and vegetable enriched diet. Ann Behav Med 2001;23(3):198-207. (115) Hofstad B, Almendingen K, Vatn M et al. Growth and recurrence of colorectal polyps: a double-blind 3-year intervention with calcium and antioxidants. Digestion 1998;59(2):148-56. (116) Keefe KA, Schell MJ, Brewer C et al. A randomized, double blind, Phase III trial using oral beta-carotene supplementation for women with high-grade cervical intraepithelial neoplasia. Cancer Epidemiology, Biomarkers & Prevention 2001 October;10(10):1029-35. (117) Kim YI, Baik HW, Fawaz K et al. Effects of folate supplementation on two provisional molecular markers of colon cancer: a prospective, randomized trial. Am J Gastroenterol 2001 January;96(1):184-95. (118) Lanza E, Schatzkin A, Daston C et al. Implementation of a 4-y, high-fiber, highfruit-and-vegetable, low-fat dietary intervention: results of dietary changes in the Polyp Prevention Trial. Am J Clin Nutr 2001;74(3):387-401. (119) Li JY, Taylor PR, Li B et al. Nutrition Intervention Trials in Linxian, China: Multiple Vitamin/Mineral Supplementation, Cancer Incidence, and Disease-Specific

47

Mortality Among Adults with Esophageal Dysplasia. Journal of the National Cancer Institute 1993;85(18):1492-8. (120) Mark SD, Liu SF, Li JY et al. The effect of vitamin and mineral supplementation on esophageal cytology: results from the Linxian Dysplasia Trial. Int J Cancer 1994 April 15;57(2):162-6. (121) McKeown-Eyssen GE, Bright-See E, Bruce WR et al. A randomized trial of a low fat high fibre diet in the recurrence of colorectal polyps. Toronto Polyp Prevention Group.[erratum appears in J Clin Epidemiol 1995 Feb;48(2):i]. Journal of Clinical Epidemiology 1994 May;47(5):525-36. (122) McKeown-Eyssen G, Holloway C, Jazmaji V, Bright-See E, Dion P, Bruce WR. A randomized trial of vitamins C and E in the prevention of recurrence of colorectal polyps. Cancer Res 1988 August 15;48(16):4701-5. (123) Paspatis GA, Karamanolis DG. Folate supplementation and adenomatous colonic polyps. Dis Colon Rectum 1994 December;37(12):1340-1. (124) Romney SL, Ho GY, Palan PR et al. Effects of beta-carotene and other factors on outcome of cervical dysplasia and human papillomavirus infection. Gynecol Oncol 1997 June;65(3):483-92. (125) Schatzkin A, Lanza E, Corle D et al. Lack of effect of a low-fat, high-fiber diet on the recurrence of colorectal adenomas. Polyp Prevention Trial Study Group.[comment]. New England Journal of Medicine 2000 April 20;342(16):114955. (126) Stich HF, Hornby AP, Mathew B, Sankaranarayanan R, Nair MK. Response of oral leukoplakias to the administration of vitamin A. Cancer Lett 1988 May;40(1):93101. (127) Wang LD, Qiu S-L, Yang G-R, Lipkin M, Newmark HL, Yang CS. A Randomized Double-Blind Intervention Study on the Effect of Calcium Supplementation on Esophageal Precancerous Lesions in a High-Risk Population in China. Cancer Epidemiology Biomarkers and Prevention 1993;2:71-8. (128) Zullo A, Rinaldi V, Hassan C et al. Ascorbic acid and intestinal metaplasia in the stomach: a prospective, randomized study. Aliment Pharmacol Ther 2000 October;14(10):1303-9. (129) Bairati I, Meyer F, Gelinas M et al. A randomized trial of antioxidant vitamins to prevent second primary cancers in head and neck cancer patients.[see comment]. Journal of the National Cancer Institute 2005 April 6;97(7):481-8. (130) Ravasco P, Monteiro-Grillo I, Vidal PM, Camilo ME. Impact of nutrition on outcome: A prospective randomized controlled trial in patients with head and neck cancer undergoing radiotherapy. Head Neck 2005;27(8):659-68. (131) Black HS, Thornby JI, Wolf JE et al. Evidence that a low-fat diet reduces the occurrence of non-melanoma skin cancer. International journal of cancer Journal international du cancer 1995;62(2):165-9.

48

(132) Black HS. Influence of dietary factors on actinically-induced skin cancer. Mutat Res 1998;422(1):185-90. (133) Gothard L, Cornes P, Earl J et al. Double-blind placebo-controlled randomised trial of vitamin E and pentoxifylline in patients with chronic arm lymphoedema and fibrosis after surgery and radiotherapy for breast cancer. Radiother Oncol 2004 November;73(2):133-9. (134) MacGregor CA, Canney PA, Patterson G, McDonald R, Paul J. A randomised double-blind controlled trial of oral soy supplements versus placebo for treatment of menopausal symptoms in patients with early breast cancer. Eur J Cancer 2005 March;41(5):708-14. (135) Toma S, Bonelli L, Sartoris A et al. beta-carotene supplementation in patients radically treated for stage I-II head and neck cancer: results of a randomized trial. Oncol Rep 2003 November;10(6):1895-901. (136) Pathak AK, Bhutani M, Guleria R et al. Chemotherapy alone vs. chemotherapy plus high dose multiple antioxidants in patients with advanced non small cell lung cancer. J Am Coll Nutr 2005 February;24(1):16-21. (137) Poon RT, Yu WC, Fan ST, Wong J. Long-term oral branched chain amino acids in patients undergoing chemoembolization for hepatocellular carcinoma: a randomized trial. Aliment Pharmacol Ther 2004 April 1;19(7):779-88. (138) Pakiz B, Flatt SW, Mills KC, Jordan LJ, Carethers JM, Rock CL. Quality of life and diet intervention in individuals at risk for recurrence of colorectal adenomas. Psychooncology 2005 February;14(2):85-93. (139) Reid ME, Stratton MS, Lillico AJ et al. A report of high-dose selenium supplementation: response and toxicities. Journal of Trace Elements in Medicine & Biology 2004;18(1):69-74. (140) Breitkreutz R, Tesdal K, Jentschura D, Haas O, Leweling H, Holm E. Effects of a high-fat diet on body composition in cancer patients receiving chemotherapy: a randomized controlled study. Wiener Klinische Wochenschrift 2005 October;117(19-20):685-92. (141) Isenring EA, Capra S, Bauer JD. Nutrition intervention is beneficial in oncology outpatients receiving radiotherapy to the gastrointestinal or head and neck area. Br J Cancer 2004 August 2;91(3):447-52. (142) Ishikawa H, Akedo I, Otani T et al. Randomized trial of dietary fiber and Lactobacillus casei administration for prevention of colorectal tumors. Int J Cancer 2005 September 20;116(5):762-7. (143) Tanaka S, Haruma K, Kunihiro M et al. Effects of aged garlic extract (AGE) on colorectal adenomas: a double-blinded study. Hiroshima Journal of Medical Sciences 2004 December;53(3-4):39-45. (144) Heimburger DC, Alexander CB, Birch R, Butterworth CE, Jr., Bailey WC, Krumdieck CL. Improvement in bronchial squamous metaplasia in smokers treated

49

with folate and vitamin B12. Report of a preliminary randomized, double-blind intervention trial.[erratum appears in JAMA 1988 Jun 27;259(23):3410]. Jama 1988 March 11;259(10):1525-30. (145) Mohanty NK, Saxena S, Singh UP, Goyal NK, Arora RP. Lycopene as a chemopreventive agent in the treatment of high-grade prostate intraepithelial neoplasia. UROL 2005 November;23(6):383-5. (146) Campbell A, Mutrie N, White F, McGuire F, Kearney N. A pilot study of a supervised group exercise programme as a rehabilitation treatment for women with breast cancer receiving adjuvant treatment. European Journal of Oncology Nursing 2005 March;9(1):56-63. (147) Headley JA, Ownby KK, John LD. The effect of seated exercise on fatigue and quality of life in women with advanced breast cancer. Oncol Nurs Forum 2004 September;Online. 31(5):977-83. (148) Mustian KM, Katula JA, Gill DL, Roscoe JA, Lang D, Murphy K. Tai Chi Chuan, health-related quality of life and self-esteem: a randomized trial with breast cancer survivors. Support Care Cancer 2004 December;12(12):871-6. (149) Sandel SL, Judge JO, Landry N, Faria L, Ouellette R, Majczak M. Dance and movement program improves quality-of-life measures in breast cancer survivors. Cancer Nurs 2005 July;28(4):301-9. (150) Marchese VG, Chiarello LA, Lange BJ. Effects of physical therapy intervention for children with acute lymphoblastic leukemia. Pediatric Blood & Cancer 2004 February;42(2):127-33. (151) Dimeo FC, Thomas F, Raabe-Menssen C, Propper F, Mathias M. Effect of aerobic exercise and relaxation training on fatigue and physical performance of cancer patients after surgery. A randomised controlled trial. Support Care Cancer 2004 November;12(11):774-9. (152) Thorsen L, Skovlund E, Stromme SB, Hornslien K, Dahl AA, Fossa SD. Effectiveness of physical activity on cardiorespiratory fitness and health-related quality of life in young and middle-aged cancer patients shortly after chemotherapy. J Clin Oncol 2005 April 1;23(10):2378-88. (153) Courneya KS, Friedenreich CM, Sela RA, Quinney HA, Rhodes RE. Correlates of adherence and contamination in a randomized controlled trial of exercise in cancer survivors: an application of the theory of planned behavior and the five factor model of personality. Ann Behav Med 2002;24(4):257-68. (154) Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. Jama 1995 February 1;273(5):408-12.

50

7. RESULTS - TABLES Information added since the original report is highlighted in grey. Information changed since the original report is highlighted in yellow. Table 7.1 General characteristics of the 53 trials reporting nutritional interventions in cancer survivors, in alphabetical order by first author. ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

23

9/23

Median (range) 64.1 (34 to 88)

27

20/27

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Not clear but participants still have cancer

During treatment

No

All-cause mortality and disease-free survival were reported at 3 months follow up when participants were not blind to treatment allocation

No

3.5 years after randomisation

Year Location * Cancer site(s) 101591 Arnold 1989

Planning to receive potentially curative RT without CT.

960 or 1080 kcal Sustacal (tm) daily

10 weeks

Usual treatment

Median (age range) 68.3 (43 to 80)

United States Head/ Larynx/ Mouth/ Nasopharynx/ Oesophagus/ Pharynx/ Supraglottis parotid 177901 Bairati 2005 Canada Head and neck (tongue/ gum/ mouth/ oropharynx/ hypopharynx/ larynx)

Patients with stage I or II squamous cell carcinoma of head and neck scheduled to be treated by radiation therapy; aged 18 yrs or over. Exclusion criteria included patients with multiple primary head and neck cancers or a history of cancer

400 IU di-alphatocopherol (1 capsule daily)

Unclear

194

223/273 $

62.9 (10)

190

203/267 $

62.3 (9.5)

Undergoing RT

Not clear

Intervention began on 1st day of radiation treatment

Placebo capsule

51

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

79

223/273 $

62.9 (10)

77

203/267 $

62.3 (9.5)

63

0/54

>18

62

0/50

>18

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Not clear

Intervention began on 1st day of radiation treatment

No

3.5 years after randomisation

Not clear

Not clear

Yes

N/A

Year Location * Cancer site(s) 177902 Bairati 2005 Canada Head and neck (tongue/ gum/ mouth/ oropharynx/ hypopharynx/ larynx) 100741 Barton 1998 United States Breast

Patients with stage I or II squamous cell carcinoma of head and neck scheduled to be treated by radiation therapy; aged 18 yrs or over. Exclusion criteria included patients with multiple primary head and neck cancers or a history of cancer

400 IU di-alphatocopherol (1 capsule daily) and 30 micro-gm betacarotene (1 capsule daily)

Women with hot flashes, a life expectancy of 6 months or more, a ECOG performance status of 0 or 1 and a history of breast cancer. Women were excluded if they had current or planned CT, were taking agents for treating hot flashes or were taking more than 2 multivitamin tablets per day of over 60 IU of vitamin E daily.

800 IU vitamin E succinate daily

Unclear

Placebo capsules

Placebo

4 weeks

Patients on a range of planned therapies including RT and CT were excluded

52

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

98

4/98

Mean 52.5 (max 75)

101

3/101

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Not clear

During treatment

Yes

N/A

No

All-cause mortality was measured at 12 months; participants were not blind to their group allocation.

Year Location * Cancer site(s) 171141 Berglund 1994 Sweden Breast/ Ovary / Prostate/ Testicular/ Other

171081 Berglund 1994 Sweden Breast/ Ovary/ Prostate/ Testicular/ Other

Undergone curative treatment for primary tumour and within 2 months of finishing post-operative treatment with RT or CT.

Undergone curative treatment for primary tumour and within 2 months of finishing post-operative treatment with RT or CT.

Eleven 2-hour sessions: 4 sessions of physical training; 4 sessions of information (including dietary advice) and 3 sessions of coping training. 1/3 offered 1 information session; 2/3 received usual treatment Eleven 2-hour sessions: 4 sessions of physical training; 4 sessions of information (including dietary advice) and 3 sessions of coping training. 1/3 offered 1 information session; 2/3 received usual treatment

7 weeks

Mean 53.9 (max 75)

7 weeks

98

4/98

101

3/101

Mean 52.5 (max 75) Mean 53.9 (max 75)

Undergoing curative treatment for a primary cancer and within 2 months of finishing postoperative treatment with RT or CT

Not clear

During treatment Undergoing curative treatment for a primary cancer and within 2 months of finishing postoperative treatment with RT or CT

53

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

57

31/57

50.6 (9.7)

58

39/58

52.3 (13.2)

57

31/57

50.6 (9.7)

58

39/58

52.3 (13.2)

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Not clear: index carcinoma

Not clear

No

N/A

Not clear: index carcinoma

Not clear

No

All-cause mortality was reported at the end of the 2 yr follow-up. Number of new, confirmed skin cancers per patient were totalled in 8 month periods of the 2 yr study period.

Year Location * Cancer site(s) 116091 Black $$ 1998 United States Skin

Patients with nonmelanoma skin cancer (either basal or squamous cell carcinoma).and no more than two previous non-melanoma skin cancers

Weekly classes of instruction to reduce fat intake, compensated by increase in complex carbohydrates. Objective was to limit calories from total fat to 20% of total calories

8 weeks

Usual diet

117931 Black $$ 1995 United States Skin

Patients with nonmelanoma skin cancer (either basal or squamous cell carcinoma).and no more than two previous non-melanoma skin cancers

Weekly classes of instruction to reduce fat intake, compensated by increase in complex carbohydrates. Objective was to limit calories from total fat to 20% of total calories Usual diet

8 weeks

54

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

12

9/12

57.8 (1.3)

11

8/11

60.6 (3.1)

33

Not given

Not given

50

Not given

Not given

Time since diagnosis

Time since treatment

Tumours were non-resectable or advanced

During chemotherapy

Stage I. Tumours could be new or recurrent

During treatment

Year Location * Cancer site(s) 171881 Breitkreutz 2005 Germany Colorectal /stomach

114271 Byar 1977 United States Bladder

Patients with histologically proven gastrointestinal adenocarcinomas, at least 3 weeks form major surgery and no prior chemotherapy

High fat diet (fatenriched diet + professionally designed schedule about balanced diets)

8 weeks

Were outcomes measured at the point that the intervention ended?

Point of followup

Yes

N/A

Control (usual diet + professionally designed schedule about balanced diets)

Patients at 10 Veterans Administration Hospitals with Stage 1 cancer. Individuals were excluded if they had received previous RT and CT for bladder cancer, a carcinoma arising completely with a diverticulum or bladder papillomatosis (those with papillomatosis were included if the tumors could be completely removed by transurethral resection).

25mg pyridoxine daily Placebo

2 years

Participants undergoing transurethral resection

No

Cancer mortality and recurrence were measured up until 5 years follow-up; it was not clear if participants were still blind to their treatment allocation

55

ID

First Author

Target Population

Intervention

Length of intervention

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

Mean 4.5 years (intervention) mean 4.3 years (control)

653

482/653

63.4 (10.2)

659

498/659

63.0 (10.0)

Until death or until the patient could no longer take medications by mouth; mean survival of patients given vitamin C was 293 days compared with 38 days for the controls.

75

37/60

Not given

75

39/63

Not given

Range 0 to 3 years (intervention) 1.5 to 3 years (control)

Not clear

0/30

Not clear

0/24

Control

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Early stage as life expectancy at least 5 years. At least one carcinoma occurring within the prior year

Not clear

Yes

N/A

Advanced cancer

Not clear

Yes

N/A

Post-treatment

Yes

N/A

Year Location * Cancer site(s) 100961 Clark 1996 United States Skin

114121 Creagan 1979

Patients at 7 200 μg selenium dermatology clinics in daily low-selenium areas; history of ≥2 basal cell Placebo carcinomas or ≥1 squamous cell carcinoma 1 of which occurred within the last year; life expectancy of at least 5 years. Advanced cancer; 10g vitamin C unsuitable for treatment daily with systematic CT

United States Colon/ Rectum / Lung/ Pancreas/ Stomach, also 'other' (not specified)

101301 de Waard 1993 Netherlands Breast

Patients at 3 hospital in Dietary advice and the Netherlands with psychological histologically support to achieve confirmed cancer which 10kg weight loss had received primary treatment; Usual treatment postmenopausal; overweight by >10kg with no signs of distant metastases at primary diagnosis and treatment

Age range 50-69 Not clear but no distant Age range 50-69 metastases

56

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

30

0/29

19

0/19

12

0/12

57.2 (6.8)

12

0/12

54.4 (7.1)

13

Not given

Not given

17

Not given

Not given

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Post-treatment

Yes

N/A

Post-treatment

Yes

N/A

Not clear

N/A

Year Location * Cancer site(s) 101302 de Waard 1993 Poland Breast

116831 Delanian 2003 France Breast

148471 Douglass 1978 United States

Dietary advice and Patients at 2 Range 0 to 1 psychological oncological hospitals in year support to achieve (intervention) 0.5 Poland with 10kg weight loss histologically to 1 year confirmed cancer which (control) Usual treatment had received primary treatment were postmenopausal and overweight by >10kg with no signs of distant metastases at primary diagnosis and treatment Women who underwent 1,000 U vitamin E 6 months RT with clinically daily measurable radiation induced fibrosis Placebo occurring > 6 months after RT completion without evidence of recurrent or evolutionary cancer or skin pathology. 900 calories Mean 16 days Locally advanced, nonelemental dietary (intervention); 26 resectable or recurrent supplement days (control) carcinoma; no solution daily metastases outside the radiation fields in Usual treatment which RT was planned.

Age range 50-69 Not clear but no distant metastases Age range 50-69

Not clear

More than 6 months after RT

Locally advanced, nonresectable or recurrent carcinoma

During treatment Undergoing RT and CT

Colon/ Rectum/ Pancreas/ Stomach

57

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

653

460/621

63.4 (10.2)

659

472/629

63.0 (9.9)

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

History of cancer within 1 year prior to randomisation.

Not clear

Yes

N/A

Not clear

Yes

N/A

During treatment

Yes

N/A

Year Location * Cancer site(s) 100031 Duffield-Lillico 2002 United States Skin

117101 Duffield-Lillico 2003 United States Skin

101911 Elkort 1981 United States Breast

Patients at 7 dermatology clinics in low-selenium areas; history of ≥ 2 basal cell carcinomas or ≥ 1 squamous cell carcinomas 1 of which occurred within the last year who had a life expectancy of at least 5 years.

200 μg selenium daily Placebo

Maximum 12 years and 5 months (calculated from when the first person entered the trial to when the trial ended). Person years of follow-up 4694 (intervention) and 4607 (control)

Patients at 7 200 μg selenium Mean 7.6 dermatology clinics in daily (intervention); low-selenium areas; 7.3 (control). history of ≥ 2 basal cell Placebo Maximum length carcinomas or ≥ 1 in both groups 13 squamous cell years carcinomas 1 of which occurred within the last year who had a life expectancy of at least 5 years. Tissue confirmed 500cc of liquid diet 12 months diagnosis of cancer and supplement (Isocal at high risk for, or with or Sustacal) daily + objective evidence of intensive metastic disease; nutritional support undergoing standard to ensure patients CT. Stage I patients, met protein and except stage T1 N1A calorie optimal Mo, were excluded. requirements

Early stage based on life expectancy≥ 5 years

Not clear

457/457

64.9 (8.8)

Not clear

470/470

63.7 (9.4)

History of cancer within 1 year prior to randomisation. Early stage based on life expectancy≥ 5 years

24

0/12

Not given

23

0/14

Not given

Participants had cancer, Stage I.

Participants were undergoing CT

Usual treatment

58

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

Not clear

36/60 (augmented nutrition); 32/51(standard nutrition)

Age range 34-74 (augmented nutrition); 23-79 (standard nutrition)

41/69

Age range 33-78

1 male was included – group unknown.

Not given

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Advanced cancer

During treatment

Yes

N/A

Yes

N/A

Year Location * Cancer site(s) 101721 Evans 1987 United States and Canada Colon/ Rectum/ Lung (Lung and colorectal patients reported separately and shown as two rows in Table 3.4)

181631 Gothard 2004 England Breast

Patients at Emory University School of Medicine, Atlanta USA; Sloan-Kettering Cancer Centre, New York, USA and the University of Toronto Hospital, Canada with histological proof of cancer, a lifeexpectancy of more than 16 weeks and a ECOG performance status ≥ 3. Individuals who underwent prior chemotherapy were excluded.

(1) Counselling to ensure optimal calorie intake for weight for height and 25% of calories from protein + commercial enteral supplement + 150mg zinc & 266mg magnesium (2) Counselling to ensure optimal calorie intake + commercial enteral supplement

General dietary guidelines Patients with long – d1-alpha standing arm tocopheryl acetate lymphodema following 500 mg twice a day lymphatic radiotherapy orally plus for early breast cancer pentoxifylline 400 mg twice a day orally

12 weeks

Not clear

6 months

35 33

>= 21 days from surgery and undertaking CT

Early breast cancer

Following radiotherapy

The remaining 67 were female

Corresponding placebos

59

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

913

630/913

< 85

892

621/892

< 85

29

24/29

60.6 (15.6)

31

27/31

63.3 (12.5)

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

At least one carcinoma within ~3 years since start of recruitment. Stage not clear.

Not clear

Yes

N/A

Not clear

During radiotherapy (28 patients were postoperative, 2 were preoperative, and 30 had no plans for surgery)

Yes

N/A

Year Location * Cancer site(s) 101441 Greenberg 1990 United States Skin

184241 Isenring 2004 Australia Head and neck/ gastrointestinal

Patients at 4 clinical 50mg beta-carotene centres with ≥1 biopsydaily proven or squamous cell carcinoma who Placebo agreed not to take vitamin supplements containing vitamin A or beta-carotene. Patients were excluded if they had an active cancer other than skin cancer, xeroderma pigmentosum or basal cell nevus syndrome. Outpatients receiving Individualised radiotherapy to the nutrition head and neck or counselling, gastrointestinal areas following a standard nutrition protocol of the American Dietetic Assoc. Where appropriate oral nutrition supplements were provided for up to 3 mths

5 years

12 weeks

Usual care of each centre. No individual nutrition advice, less nutrition assessment, less follow-up. Those receiving RT to areas other than head and neck could request up to 2 dietetic consultations

60

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

14

Not given

Not given

14

Not given

Not given

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Advanced cancer

During treatment

Yes

N/A

No

Cancer recurrence and second primary cancer were measured at a 3 years; it was not clear if participants knew their group allocation

No

All-cause mortality was measured at 5 years; participants were not blind to their group allocation.

Year Location * Cancer site(s) 153281 Jebb 1994

Individuals with advanced metastic cancers

16g glutamine daily

8 days

Placebo

All participants were undergoing CT

United Kingdom Colon/ Rectum/ Gall bladder/ Pancreas/ Stomach/and one unknown cancer (not specified) 107121 Jyothirmayi 1996 India Head

113791 Kokron 1982 Austria Lung

Inhabitants of Kerala with complete clinical regression of lesions after therapy; had radical RT or surgery and no clinical evidence of disease.

200,000 IU oral vitamin A (retinyl palmitate) weekly

Individuals with inoperable cancer. Individuals were excluded if they had a primary tumour > 8cm or distant metastasis.

Vitamin A 1.5 Mill: E (maximum dose) plus radiation therapy at two thirds usual dose

1 year

56

37/56

Mean 56.3

50

36/50

Mean 58.2

15

15/15

Not given

11

11/11

Not given

Not clear

Already treated, resulting in complete remission of lesions

Placebo

Normal level radiotherapy

35 days

Post-treatment

Participants had inoperable and advanced cancer

Not clear

61

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

21

0/21

Not given

21

0/21

35

30/35

Mean 65.9

30

24/30

Mean 68.1

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Participants had inoperable and advanced cancer

Pre-treatment

No

Cancer mortality and cancer recurrence were measured at 12 months; it is not clear if participants were aware of their group allocation

Yes

N/A

Yes

N/A

Year Location * Cancer site(s) 101951 Kucera 1980 Austria Cervix

Women with inoperable 1.5 mill. IU per day cervical cancer (FIGO of vitamin A III) who have had no (retinyl palmitate) prior cancer treatment taken orally. The total dose over the entire intervention was 30mill. IU

4 weeks

Participants had received no prior cancer treatment

Usual treatment 101121 Lamm 1994 United States Bladder

178471 MacGregor 2005 Scotland

Biopsy confirmed transitional cell carcinoma.

Recommended daily allowance of multivitamins + 40,000 IU vitamin A + 100mg vitamin B6 + 2000mg vitamin C + 400 IU vitamin E + 90mg zinc

Multivitamins in the recommended daily allowance Patients >18 years with 2 capsules of 235 histologically mg soy extract with confirmed pre-existing 17.5mg isoflavines diagnosis sof breast twice a day (i.e. cancer and menopausal 70mg isoflavines symptoms per day)

Maximum 5 years

Had cancer, Stage not clear.

During treatment Trial included transurethral resection of recurrent tumours

12 weeks 36

0/36

Median (range) 51 (37 – 69)

36

0/36

51 (33 – 70)

Had cancer, not advanced or metastatic

During treatment

Breast Identical placebo tablets

62

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

30

Not given

Not given

62

Not given

Not given

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Had cancer, Stage not clear.

During treatment

Yes

N/A

Yes

N/A

Year Location * Cancer site(s) 101411 Macia 1991 Spain Bladder/ Breast/ Head/ Larynx/ Mouth / Nasopharynx/ Ovary/ Pharynx/ Testicular / Retroperitoneum 100241 Mayne

State of health =2. Individuals were energy excluded if they had requirements + prior chemotherapy supplements if (except adjuvant needed chemotherapy in breast cancer) and were Ad lib diet + undergoing current or nutritional support planned hormonal at the discretion of therapy the physician

Mean (range) 58 (22 to 80)

Individuals with previous therapy (other than transurethral resection or fulguration) were excluded

Not clear

During treatment Patients undergoing CT

66

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Number randomized to control

Number of males in control†

Age in control group (years)#

14

4/12

12

1/8

Age range 2976; median age 66

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Had advanced cancer

Not clear

Yes

N/A

Year Location * Cancer site(s) 171341 Ovesen 1992 Denmark Breast/ Ovary/ Lung

Advanced cancer, 500mL or more of unintentional and commercial liquid continuous weight loss diet including > 5% of body weight partially despite ≥ 2 months hydrolyzed soy efforts to increase protein as source of weight using ordinary nitrogen (Standard foods; performance Top Up, status of not more than Novonordisk, 3 on the ECOG Scale. Copenhagen, No evidence of ascites Denmark) daily or metastatic disease to the brain/life500mL or more expectancy >2 months. commercial liquid diet based on intact milk proteins (Salvimulsin MCT, Ercopharm, Copenhagen, Denmark) daily

2 months

Age range 4470; median age 61

67

ID

First Author

Target Population

Intervention

Length of intervention

Control

Number randomized to intervention

Number of males in intervention†

Age in intervention group (years)#

Time since diagnosis

Time since treatment

Were outcomes measured at the point that the intervention ended?

Point of followup

Number randomized to control

Number of males in control†

Age in control group (years)#

37

27/37

Breakdown by agegps given

Within previous 6 months

Resection within previous 6 months

Yes

N/A

40

27/40

154

140/150

90% in 105/126 (intervention) vs. 92/116 (control) in year 1 and 2/2 (intervention) vs. 4/4 (control) in year 8. Supplementation produced a 9-10 fold increase in median plasma beta-carotene levels in the intervention vs. no change in the control

Monitored by researcher

Consumption of intervention and also the intake (especially in the control group) was supervised. 2 patients complained that it was unpalatable but were still compliant (suggesting 100% compliance) Not clear

1999 101051

Meyskens 1995

No

Requires German translation

1982

2001

Requires German translation

Were survival outcomes analyzed by intention to treat?* Not clear

Not clear

Not clear Randomization list prepared by study biostatistician

Not applicable Not applicable Not clear

Yes Not clear Yes Yes No Yes Yes

Yale and University of Miami Research Pharmacies administered the randomization list Computer-generated list

No

Closed envelopes Not clear Not clear

Yes Not clear

No

Not clear Not clear

82

ID

Author Year

101181

Meyskens

101781

1994 Moertel 1995

101101

Newling

How was compliance recorded?

Not clear

Ovesen

Biochemical levels

The authors report that in the 75 control and 85 intervention group participants red cell levels of vitamin B6-dependent enzyme, glutamate oxalacetate transaminase (EGOT) and the concentration of pyridoxal 5-phosphate (P5P) at baseline and during follow-up suggested excellent compliance with the prescribed treatment. Mean (SD) change from baseline at 5 months in intervention vs. control was 0.6 (2.3) vs. -0.3 (2.0) for energy intake (MJ/d) and 8 (21) vs. -1 (22) for protein intake (g/d); both differences between groups were significant at the pBreast cancer specific

 (p=0.066)

Total

↑↑ (p