Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases

Background Animal and physiological research as well as observational studies suggest that antioxidant supplements may improve survival. Objectives To assess the effect of antioxidant supplements on mortality in primary or secondary prevention randomised clinical trials. Search strategy We searched The Cochrane Library (Issue 3, 2005), MEDLINE (1966 to October 2005), EMBASE (1985 to October 2005), and the Science Citation Index Expanded (1945 to October 2005). We scanned bibliographies of relevant publications and wrote to pharmaceutical companies for additional trials. Selection criteria We included all primary and secondary prevention randomised clinical trials on antioxidant supplements (beta-carotene, vitamin A, vitamin C, vitamin E, and selenium) versus placebo or no intervention. Included participants were either healthy (primary prevention trials) or had any disease (secondary prevention trials). Data collection and analysis Three authors extracted data. Trials with adequate randomisation, blinding, and follow-up were classified as having a low risk of bias. Random-effects and fixed-effect meta-analyses were performed. Random-effects meta-regression analyses were performed to assess sources of intertrial heterogeneity. Main results Sixty-seven randomised trials with 232,550 participants were included. Forty-seven trials including 180,938 participants had low risk of bias. Twenty-one trials included 164,439 healthy participants. Forty-six trials included 68111 participants with various diseases (gastrointestinal, cardiovascular, neurological, ocular, dermatological, rheumatoid, renal, endocrinological, or unspecified). Overall, the antioxidant supplements had no significant effect on mortality in a random-effects meta-analysis (relative risk [RR] 1.02, 95% confidence interval [CI] 0.99 to 1.06), but significantly increased mortality in a fixed-effect model (RR 1.04, 95% CI 1.02 to 1.06). In meta-regression analysis, the risk of bias and type of antioxidant supplement were the only significant predictors of intertrial heterogeneity. In the trials with a low risk of bias, the antioxidant supplements significantly increased mortality (RR 1.05, 95% CI 1.02 to 1.08). When the different antioxidants were assessed separately, analyses including trials with a low risk of bias and excluding selenium trials found significantly increased mortality by vitamin A (RR 1.16, 95% CI 1.10 to 1.24), beta-carotene (RR 1.07, 95% CI 1.02 to 1.11), and vitamin E (RR 1.04, 95% CI 1.01 to 1.07), but no significant detrimental effect of vitamin C (RR 1.06, 95% CI 0.94 to 1.20). Low-bias risk trials on selenium found no significant effect on mortality (RR 0.91, 95% CI 0.76 to 1.09). 1 Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases (Review) Copyright © 2008 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd Authors’ conclusions We found no evidence to support antioxidant supplements for primary or secondary prevention. Vitamin A, beta-carotene, and vitamin E may increase mortality. Future randomised trials could evaluate the potential effects of vitamin C and selenium for primary and secondary prevention. Such trials should be closely monitored for potential harmful effects. Antioxidant supplements need to be considered medicinal products and should undergo sufficient evaluation before marketing. P L A I N L A N G U A G E S U M M A R Y No evidence to support antioxidant supplements to prevent mortality in healthy people or patients with various diseases Previous research on animal and physiological models suggest that antioxidant supplements have beneficial effects that may prolong life. Some observational studies also suggest that antioxidant supplements may prolong life, whereas other observational studies demonstrate neutral or harmful effects. Randomised trials have largely been neutral. We need evidence from randomised trials to decide if antioxidant supplements should be used for prevention. The present systematic review includes 67 randomised clinical trials. In total, 232,550 participants were randomised to antioxidant supplements (beta-carotene, vitamin A, vitamin C, vitamin E, and selenium) versus placebo or no intervention. Twenty-one trials included 164,439 healthy participants. Forty-six trials included 68111 participants with various diseases (including gastrointestinal, cardiovascular, neurological, ocular, dermatological, rheumatoid, renal, endocrinological, or unspecified diseases). Overall, the antioxidant supplements did not seem to reduce mortality. A total of 17880 of 136,023 participants (13.1%) randomised to antioxidant supplements and 10136 of 96527 participants (10.5%) randomised to placebo or no intervention died. In the analyses of the trials with low risk of bias, beta-carotene, vitamin A, and vitamin E significantly increased mortality. There were no significant differences between the effects of antioxidant supplements in healthy participants (primary prevention trials) or participants with various diseases (secondary prevention trials). Randomised trials with adequate bias control found no significant effect of vitamin C. In some of our analyses, selenium seems to reduce mortality. The current evidence does not support the use of antioxidant supplements in the general population or in patients with certain diseases. The combined evidence suggests that additional research on antioxidant supplements is needed. The evidence on vitamin C and selenium was not conclusive. Future trials could focus on vitamin C and selenium and should assess both potential beneficial and harmful effects. Conduct of additional primary and secondary prevention trials on vitamin A, beta-carotene, and vitamin E seems questionable, at least in the dosage range examined. The present review does not assess antioxidant supplements for treatment of specific diseases (tertiary prevention), antioxidant supplements for patients with demonstrated specific needs of antioxidants, or the effects of antioxidants contained in fruits or vegetables. Further research and systematic reviews on these types of interventions are therefore warranted.


T A B L E O F C O N T E N
the antioxidant supplements had no significant effect on mortality in a random-effects meta-analysis (relative risk [RR] 1.02, 95% confidence interval [CI] 0.99 to 1.06), but significantly increased mortality in a fixed-effect model (RR 1.04, 95% CI 1.02 to 1.06).
In meta-regression analysis, the risk of bias and type of antioxidant supplement were the only significant predictors of intertrial heterogeneity. In the trials with a low risk of bias, the antioxidant supplements significantly increased mortality (RR 1.05, 95% CI 1.02 to 1.08). When the different antioxidants were assessed separately, analyses including trials with a low risk of bias and excluding selenium trials found significantly increased mortality by vitamin A (RR 1.16, 95% CI 1.10 to 1.24), beta-carotene (RR 1.07, 95% CI 1.02 to 1.11), and vitamin E (RR 1.04, 95% CI 1.01 to 1.07), but no significant detrimental effect of vitamin C (RR 1.06, 95% CI 0.94 to 1.20). Low-bias risk trials on selenium found no significant effect on mortality (RR 0.90, 95% CI 0.80 to 1.01).

Authors' conclusions
We found no evidence to support antioxidant supplements for primary or secondary prevention. Vitamin A, beta-carotene, and vitamin E may increase mortality. Future randomised trials could evaluate the potential effects of vitamin C and selenium for primary and secondary prevention. Such trials should be closely monitored for potential harmful effects. Antioxidant supplements need to be considered medicinal products and should undergo sufficient evaluation before marketing.

No evidence to support antioxidant supplements to prevent mortality in healthy people or patients with various diseases
Previous research on animal and physiological models suggest that antioxidant supplements have beneficial effects that may prolong life. Some observational studies also suggest that antioxidant supplements may prolong life, whereas other observational studies demonstrate neutral or harmful effects. Randomised trials have largely been neutral. We need evidence from randomised trials to decide if antioxidant supplements should be used for prevention.

B A C K G R O U N D
Oxidative stress may play role in the pathogenesis of cancer and cardiovascular disease, the leading causes of death in middle-and high-income countries (Sies 1985;Halliwell 1999). Diet provides numerous vitamins and trace elements that are essential for good health. Several observational studies have shown a significant positive association between higher intake of fruits and vegetables and reduced risk of chronic diseases (Block 1992;Ames 1993;Willcox 2004). However, exactly which specific dietary constituents of fruits and vegetables might be beneficial is not clear. Furthermore, causal inferences are hard to establish from observational studies. Antioxidants have attracted most attention as promising preventive agents. Fruits and vegetables are sources of numerous micronutrients and some of these, including beta-carotene, vitamin A, vitamin C, vitamin E, and selenium have antioxidant potential. Many people take antioxidant supplements believing to improve their health (Balluz 2000;Millen 2004;Radimer 2004;Nichter 2006).
Whether antioxidant supplements are beneficial or harmful is uncertain (Herbert 1997;Caraballoso 2003;Vivekananthan 2003;Bjelakovic 2004;Stanner 2004;Miller 2005;Berger 2005). Antioxidants may play dual roles, acting as double-edged swords (Bjelakovic 2007b). Excessive antioxidants can adversely affect key physiological processes. The results of our recent systematic review and meta-analyses of the role of antioxidant supplements for prevention of gastrointestinal cancers were unforeseen (Bjelakovic 2004). We found that antioxidant supplements significantly increased mortality in the antioxidant group with the fixed-effect model meta-analysis but not with the randomeffects meta-analysis (Bjelakovic 2004). The effect of antioxidant supplements on mortality has also been assessed in several large trials on primary and secondary prevention of diseases ( HPS 2002Low;ATBC 2003Low;CARET 2004Low;SUVIMAX 2004Low;HOPE TOO 2005Low;WHS 2005Low). The results of the individual trials are equivocal. Furthermore, none of the trials had sufficient statistical power to identify the effect of antioxidants on mortality. Accordingly, we performed a systematic review of randomised trials on antioxidant supplements for primary and secondary prevention.

O B J E C T I V E S
Our aim was to assess the effect of antioxidant supplements (betacarotene, vitamin A, vitamin C, vitamin E, and selenium) on overall mortality in primary or secondary prevention randomised clinical trials.

M E T H O D S
Criteria for considering studies for this review

Types of studies
All primary and secondary prevention randomised clinical trials, irrespective of trial design, blinding, publication status, publication year, or language. From cross-over trials, only the first trial period was considered.

Types of participants
Adult participants (age 18 years or over) who were • healthy participants or were recruited among the general population (primary prevention); • diagnosed with a specific disease in a stable phase (secondary prevention).
We excluded tertiary prevention trials, ie, randomised trials in which antioxidant supplements were used to treat a specific disease or nutritional defects, like trials with patients with acute, infectious, or malignant diseases (except non-melanoma skin cancer). We excluded trials including children and pregnant women since they may be in need of certain antioxidant supplements.

Types of interventions
We considered for inclusion trials that compared antioxidant supplements (ie, beta-carotene, vitamin A, vitamin C, vitamin E, and selenium) at any dose, duration, and route of administration versus placebo or no intervention. The antioxidants could have been administered • separately or in any combination among themselves; or • in combination with other vitamins; or • in combination with trace elements without antioxidant function.
Concomitant interventions were allowed if used equally in both intervention arms of the trial.

Types of outcome measures
Our sole outcome measure was all-cause mortality.

Search methods for identification of studies
We searched the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Issue 3, 2005), MEDLINE (1966( to October 2005, EMBASE (Excerpta Medica Database) (1985( to October 2005, and the Science Citation Index Expanded (1945( to October 2005 (Royle 2003). We scanned bibliographies of relevant publications for additional trials. All search strategies are given in Appendix 1. We sent letters by post or e-mail to major manufacturers of antioxidant supplements, ie, CBH in China, DSM in Switzerland, CVC4health in USA, and BASF in Germany, asking for unpublished randomised trials. No reply was received from any of the contacted manufacturers.

Data collection and analysis
The present review is based on our protocol on antioxidant supplements for preventing gastrointestinal cancers (Bjelakovic 2003) adopted to assess overall mortality. An abbreviated version of the review has previously been published (Bjelakovic 2007a).

Inclusion criteria application
Two of the three authors (GB and DN or RGS) independently assessed trial eligibility without blinding of the study authors. We listed excluded trials with the reasons for exclusion. Disagreement was resolved by discussion or in consultation with LLG or CG. We contacted authors of the trials for missing information.

Data extraction Participant characteristics, diagnosis, and interventions
From each trial we recorded first author; country of origin; country income category (low, middle, high) (World Bank 2006); number of participants; characteristics of participants: age range (mean or median) and sex ratio; participation rate; dropout rate; trial design (parallel, factorial, or crossover); type of antioxidant; dose; duration of supplementation; duration of follow-up (ie, duration of intervention plus post-intervention follow-up); and co-interventions.

Trial characteristics
We recorded the date, location, sponsor of the trial (known or unknown and type of sponsor) as well as publication status.

Assessment of methodological quality
We defined the methodological quality as the confidence that the design and report restrict bias in the intervention comparison ( Schulz 1995;Moher 1998;Kjaergard 2001). Due to the risk of overestimation of intervention effects in randomised trials with inadequate methodological quality (Schulz 1995;Moher 1998;Kjaergard 2001), we assessed the influence of methodological quality of the four components below as reported in the trials. When this information was not available, we asked the authors of the trial publications to provide it.

Generation of the allocation sequence
• Adequate, if the allocation sequence was generated by a computer or random number table. Drawing of lots, tossing of a coin, shuffling of cards, or throwing dice was considered as adequate if a person who was not otherwise involved in the recruitment of participants performed the procedure.
• Unclear, if the trial was described as randomised, but the method used for the allocation sequence generation was not described.
• Inadequate, if a system involving dates, names, or admittance numbers were used for the allocation of patients.

Allocation concealment
• Adequate, if the allocation of patients involved a central independent unit, on-site locked computer, identically appearing numbered drug bottles or containers prepared by an independent pharmacist or investigator, or sealed envelopes.
• Unclear, if the trial was described as randomised, but the method used to conceal the allocation was not described.
• Inadequate, if the allocation sequence was known to the investigators who assigned participants or if the study was quasi-randomised.

Blinding (or masking)
• Adequate, if the trial was described as double blind and the method of blinding involved identical placebo or active drugs.
• Unclear, if the trial was described as double blind, but the method of blinding was not described.
• Not performed, if the trial was not double blind.

Follow-up
• Adequate, if the numbers and reasons for dropouts and withdrawals in all intervention groups were described or if it was specified that there were no dropouts or withdrawals.
• Unclear, if the report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated.
• Inadequate, if the number or reasons for dropouts and withdrawals were not described.
Trials with adequate generation of the allocation sequence, adequate allocation concealment, adequate blinding, and adequate follow-up were considered low-bias risk trials (high methodological quality) (Kjaergard 2001;Gluud 2006a). We appended 'Low' to the references of these trials. Trials with one or more unclear or inadequate quality components were classified as high-bias risk trials (low methodological quality) (Kjaergard 2001;Gluud 2006a). We also reported on whether the investigators had performed a sample-size calculation and used intention-to-treat analysis ( Gluud 2001).

Statistical analyses
We performed the meta-analyses according to the recommendations of The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2006). For the statistical analyses, we used RevMan Analyses (RevMan 2003), STATA 8.2 (STATA Corp, College Station, Tex), Sigma Stat 3.0 (SPSS Inc, Chicago, Ill), and Stats-Direct (StatsDirect Ltd, Altrincham, England). We analysed the data with both random-effects (DerSimonian 1986) and fixed-effect (DeMets 1987) model meta-analyses. We presented the results of random-effects model analyses. When statistically significant results are obtained in either the random-or fixed-effect model, we present both analyses. Results are presented as the relative risk (RR) with 95% confidence intervals (CI). We assessed heterogeneity with I 2 , which describes the percentage of total variation across studies due to heterogeneity rather than chance (Higgins 2002). Random-effects meta-regression analyses were performed to assess potential covariates that could predict intertrial heterogeneity, ie, which covariates that were statistically associated with estimated intervention effects. The included covariates were bias risk (low or high), type and dose of supplement, single or combined antioxidant experimental supplement regimen, duration of treatment, and type of prevention (primary or secondary). We also performed subgroup analyses comparing the primary and secondary prevention trials. Furthermore, we performed sensitivity analyses excluding trials using small dose antioxidant supplements in both the experimental and control study groups. The exclusion of trials using small dose antioxidant supplements in both the experimental and control study groups was based on the fact that addition of, eg, a vitamin pill could be a confounder. We observed that selenium seemed to have a beneficial effect on gastrointestinal cancer development (Bjelakovic 2004). The sensitivity analysis removing selenium trials from our analysis to evaluate their influence on our conclusions was therefore not a post hoc decision. The influence of trials with zero events in the treatment or control group was assessed by re-calculating the random-effects meta-analyses with 0.5, 0.05, and 0.005 continuity corrections (Sweeting 2004;Bradburn 2007). We also performed additional meta-analyses including one large hypothetical trial with one event in the treatment and control group and a sample size corresponding to the total number of participants in the zero events trials. All our analyses followed the intention-to-treat principle. We accounted all of the participants for each trial and performed the analyses irrespective of how the original trialists had analysed the data. Participants lost to follow-up were considered survivors. For trials with a factorial design, we based our results on 'at the margins' analysis, comparing all groups that received antioxidant supplements with groups that did not receive antioxidant supplements (McAlister 2003). This entails a risk of interaction between the antioxidant and the other intervention(s) assessed, whether significant or not in the individual trial. Due to the risk of confounding in factorial trials assessing other interventions, we conducted posthoc sensitivity analysis including only factorial trial data, which could not be affected of such confounding (ie, 'inside the table' analysis) (McAlister 2003). To determine the effect of a single antioxidant we also performed 'inside the table' analysis (McAlister 2003) in which we compared the single antioxidant intervention with the placebo or no intervention. In the trials with parallel group design with more than two arms and additional therapy, we compared only antioxidant intervention with placebo or no intervention. For cross-over trials we included only data from the first period (Higgins 2006). Comparison of intervention effects was conducted with test of interaction (Altman 2003). We performed adjusted rank correlation (Begg 1994) and regression asymmetry test (Egger 1997) for detection of bias. A P < 0.10 was considered significant.

R E S U L T S Description of studies
See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies.

Search results
Database searches yielded 16111 references ( Figure 1). Exclusion of duplicates and irrelevant references left 1201 references describing 815 trials. To obtain additional information we wrote to authors of about 500 eligible trials. More than one hundred authors responded.

Figure 1. Flow diagram of identification of randomised trials for Inclusion
We excluded 815 references dealing with 748 studies. After further evaluation we excluded 339 studies because they were not randomised trials or did not fulfil our inclusion criteria. The remaining 409 were randomised trials examining antioxidant supplements. Four of these were still ongoing. The authors of the 405 trials did not report data on mortality (these trials are shown at http://ctu.rh.dk). The majority of these were small phase I or phase II trials with short duration of follow-up without assessment of clinical outcome measures. We contacted the authors and about one fifth of them confirmed that mortality was indeed zero. We included 386 references describing 67 randomised trials fulfilling our inclusion criteria and able to provide data for our analyses (Table 1; Table 2) (http://ctu.rh.dk). This corresponds to a median of 6 references per included trial (range 1 to 44 references per trial).   The antioxidant supplements were given in the following combinations: • beta-carotene and vitamin A • beta-carotene and vitamin C • beta-carotene and vitamin E • vitamin A and vitamin C • vitamin C and vitamin E • vitamin E and selenium • selenium and zinc • beta-carotene, vitamin C, and vitamin E • beta-carotene, vitamin C, vitamin E, and selenium • beta-carotene, vitamin C, vitamin E, selenium, and zinc • vitamin A, vitamin C, vitamin E • vitamin A, vitamin C, vitamin E, selenium, and zinc • vitamin A, vitamin C, vitamin E, selenium, methionine, and ubiquinone.

Control interventions
Sixty-three trials used placebo and four trials used no intervention in the control group (ter Riet 1995;GISSI 1999;PPP 2001;Takagi 2003).

Excluded studies
The reason for exclusion of studies is given in the table 'Characteristics of excluded studies'.

Risk of bias in included studies
Forty-seven of the 67 trials (70.1%) had low-bias risk, ie, had adequate generation of the allocation sequence, allocation concealment, blinding, and follow-up. For an overview of the included trials with low risk of bias see Table 1 and Table 4. Twenty trials had one or more unclear or inadequate methodological components. For an overview of the included trials with high risk of bias (low methodological quality) see Table 2 and Table 3. Sixty-six trials reported losses to follow-up. The exact number of participants lost to follow-up separately for each group was reported in 65 trials. There was not a substantial difference in the losses to follow-up between the intervention group and the control group (2669 out of 108,480 (2.5%) versus 2593 out of 91168 (2.8%))

Sensitivity analyses taking trials with zero events into account
We included 14 trials with zero mortality in one arm. To account for the potential influence of these trials, we calculated the RR with 0.5, 0.05, and 0.005 as empirical continuity corrections ( Sweeting 2004;Bradburn 2007). The random-effects model RR for the three continuity corrections were RR 1.02, RR 1.02, and RR 1.03. All three analyses were insignificant. The fixed-effect models with the same corrections were all showing significantly increased mortality in the antioxidant group and found the RR 1.04 in all three analyses. Overall, 405 trials had zero mortality in both the experimental and control groups. These trials are excluded from the meta-analyses using RR as association measure. The total number of participants in these trials was about 40000. Therefore we performed exploratory analyses adding an imagined trial with 1 death and 20000 participants in each intervention group. The influence of zero events trials on our final result was not noticeable.

Analyses of bias risk
Inspection of the funnel plot in Figure 2 suggests potential bias (asymmetry). The adjusted-rank correlation test (P = 0.41) and regression asymmetry test (P = 0.21) found, however, no significant evidence of bias.

Meta-regression analyses
Univariate meta-regression analyses revealed that the following covariates were significantly associated with estimated intervention effect on mortality: bias-risk (RR 1.17; 95% CI 1.05 to 1.30, P = 0.004) and dose of beta-carotene (RR 1.004, 95% CI 1.009 to 1.007; P = 0.013), dose of vitamin A (RR 1.000006, 95% CI 1.000002 to 1.000009, P = 0.003), and dose of selenium (RR 0.998, 95% CI 0.997 to 0.999, P = 0.005). None of the other covariates (dose of vitamin C; dose of vitamin E; single or combined experimental antioxidant regimen; duration of supplementation; and primary or secondary prevention) were significantly associated with estimated intervention effect on mortality.
In multivariate meta-regression analysis including all covariates, low bias risk of the trials was associated with significantly higher estimated intervention effect on mortality (RR 1.16, 95% CI 1.04 to 1.29, P = 0.007), and dose of selenium was associated with significantly lower estimated intervention effect on mortality (RR 0.999, 95% CI 0.997 to 1.000, P = 0.013). None of the other covariates was significantly associated with the estimated intervention effect on mortality.
Intervention effects according to bias risk of trials (Analysis 1.1) In trials with low-bias risk, mortality was significantly increased in the supplemented group (RR 1.05, 95% CI 1.02 to 1.08, P = 0.003) without significant heterogeneity (I 2 = 7.5%). Exploratory analysis adding an imagined trial with 1 death and 20000 participants in each study group had no noticeable effect on the results. In trials with high-bias risk (low-methodological quality) mortality was significantly decreased in the supplemented group (RR 0.92, 95% CI 0.85 to 0.99, P = 0.03) without heterogeneity (I 2 = 0%). Exploratory analysis adding an imagined trial with 1 death and 20000 participants in each study group had no noticeable effect on the results. The difference between the estimate of antioxidants on mortality in low-and high-bias risk trials was statistically significant by test of interaction (z = 3.2, P = 0.0014).

Random-effects and fixed-effect model meta-analyses
For an overview of the effect of the different antioxidant supplements on mortality in a random-effects or fixed-effect model see Table 5 and Table 6. In primary prevention trials with low-bias risk, mortality was not significantly increased in the supplemented group in random-effects analysis (RR 1.05, 95% CI 0.98 to 1.12, P = 0.18), but significantly increased in a fixed-effect analysis (RR 1.06, 95% CI 1.04 to 1.10, P < 0.0001) with significant heterogeneity (I 2 = 46.9%).
As stated above, meta-regression analysis did not find significant difference in the estimated intervention effects in primary and secondary prevention trials.
Sensitivity analyses excluding trials with co-administration of additional supplements to the experimental group (Analysis 1.3) Sensitivity analyses excluding 11 trials that used co-interventions in the form of extra vitamins or trace elements with or without antioxidant functions in the experimental group did not noticeably change our results. In the 39 low-bias risk trials antioxidant supplements significantly increased mortality (RR 1.05, 95% CI 1.02 to 1.09, P = 0.0004) without significant heterogeneity (I 2 = 5.2%).
Sensitivity analyses excluding trials with co-administration of additional antioxidant supplements to both the experimental and control groups (Analysis 1.4) Sensitivity analyses excluding nine trials that used co-interventions in the form of extra vitamins in both the experimental and control groups did not noticeably change our results. In the 43 low-bias risk trials antioxidant supplements significantly increased mortality (RR 1.05, 95% CI 1.01 to 1.08, P = 0.01) without significant heterogeneity (I 2 = 12.6%).

Sensitivity analysis excluding factorial trials testing collateral interventions (Analysis 1.5)
Sensitivity analysis excluding factorial trials testing collateral interventions which could lead to potential confounding did not noticeably change our results. In the 34 trials with low-bias risk, mortality was significantly increased in the supplemented group (RR 1.11, 95% CI 1.05 to 1.17, P < 0.0001, I 2 = 0%).
Sensitivity analysis excluding trials with co-administration of additional supplements to experimental and control groups, and factorial trials testing collateral interventions (Analysis 1.6) Sensitivity analysis excluding trials with co-administration of additional supplements to the experimental and control groups, and factorial trials testing collateral interventions did not noticeably change our results. In the 24 trials with low-bias risk, mortality was significantly increased in the supplemented group (RR 1.12, 95% CI 1.07 to 1.19, P < 0.0001, I 2 = 0%). These trials explored the influence of beta-carotene, vitamin A, vitamin C, vitamin E, and selenium.
The estimate of mortality risk in the 24 low-bias risk trials without potential confounding interventions (1.12, 95% CI 1.07 to 1.19) compared to the estimate of mortality risk in the 48 low-bias risk trials with potential confounding (1.05, 95% CI 1.02 to 1.08) was significantly increased (z = -2.09; P = 0.036).
Sensitivity analysis excluding trials with co-administration of additional supplements to the experimental and control groups, factorial trials testing collateral interventions, and selenium trials (Analysis 1.7) Sensitivity analysis excluding trials with co-administration of additional supplements to the experimental and control groups, factorial trials testing collateral interventions, and selenium trials did not noticeably change our results. In the 19 trials with low-bias risk, mortality was significantly increased in the supplemented group (RR 1.16, 95% CI 1.09 to 1.23, P < 0.00001, I 2 = 0%).
The estimate of mortality risk in the 19 low-bias risk trials without any potential confounding interventions (1.16, 95% CI 1.09 to 1.23) compared to the estimate of mortality risk in the 48 lowbias risk trials with potential confounding (1.05, 95% CI 1.02 to 1.08) was significantly increased (z = -2.92; P = 0.0035).

D I S C U S S I O N
Our systematic review contains a number of findings. Betacarotene, vitamin A, and vitamin E given singly or combined with other antioxidant supplements significantly increase mortality. There is no evidence that vitamin C may increase longevity. We lack evidence to refute a potential negative effect of vitamin C on survival. Selenium tended to reduce mortality but only when high-bias risk trials were considered. Accordingly, we need more research on vitamin C and selenium. We confirmed that trials with inadequate bias control significantly overestimate intervention effects (Schulz 1995;Moher 1998;Kjaergard 2001;Bjelakovic 2004;Bjelakovic 2006). The detrimental effect of antioxidant supplements became significantly more pronounced when we excluded all trials with potential confounding. Our findings support and extend our previous findings regarding antioxidant supplements and increased mortality (Bjelakovic 2004).

Fixed-effect model and random-effects model meta-analyses
The fixed-effect model meta-analysis assumes that the true intervention effect is the same in every randomised trial, ie, the effect is fixed across trials. The random-effects model assumes that the effects being estimated based on the different randomised trials differ, but follow some general distribution. When there is no heterogeneity (I 2 = 0%), then fixed-and random-effects meta-analyses tend to give the same result. With increasing heterogeneity, then the estimated intervention effect as well as the corresponding 95% confidence interval will differ between the two models.
The meta-analyses we conducted included a heterogeneous set of randomised trials, eg, healthy participants or patients, single antioxidant supplement or combination, short duration of followup or long duration. These aspects could argue for only employing the random-effects model. We were requested to do so in our sister publication of the present review in JAMA (Bjelakovic 2007a). The standard random-effect model used in RevMan Analysis (RevMan 2003) is the DerSimonian and Laird method, which models the known differences between trials by incorporating a variance parameter tau 2 to account for across-trial variation (DerSimonian 1986). Adoption of the random-effects model in meta-analysis permits extension of inferences to a broader population of studies than the fixed-effect model does, which excludes the parameter tau 2 from the model.
The use of the random-effects model may come at a price. If there is between trial heterogeneity, then the weight of the large trials (usually providing more realistic estimates of intervention effects) becomes less. At the same time, the weight of small trials (usually providing more unrealistic estimates of intervention effects due to 'bias' (systematic errors) and 'chance' (random errors)) increases.
We, therefore, also analysed our meta-analyses with the fixed-effect method. In all meta-analyses the pooled estimate of increased mortality in the antioxidant-supplemented group became more pronounced ( Table 6). More of the meta-analyses became more significant or changed from non-significant to significant detrimental effects of vitamin A, beta-carotene, and vitamin E or borderline effects of selenium ( Table 6).
The choice of statistical model for performing meta-analysis of sparse data is important (Sweeting 2004;Bradburn 2007 We conducted a thorough review with methodology following the recommendations of The Cochrane Collaboration (Higgins 2006) and findings of methodological studies (Schulz 1995;Moher 1998;Kjaergard 2001). More than two thirds of the included trials with more than 180,000 participants fall in the group of low-bias risk trials. This highlights the validity of our results (Schulz 1995;Moher 1998;Kjaergard 2001). Antioxidant supplements not only seem to be one of the most researched topics in the world, they also seem to be one of the most adequately researched questions. Usually, only a small proportion of trials use adequate methodologies (Gluud 2006a; Gluud 2006b).
Our meta-analyses had little trial heterogeneity. This increases the trustworthiness of our findings. Our analyses were robust to sensitivity analyses involving different imputations of mortality in the zero-event study groups. We gave full account of all 405 identified trials assessing the supplements having zero events in both intervention groups. These trials were mostly assessing short-term supplement administration and surrogate outcome measures. Our results were robust to exploratory analyses adding an imagined trial with 20000 participants and one death in each intervention group. Accordingly, the increased mortality does not seem to be an artefact created by exclusion of trials with zero events in both intervention groups (Sweeting 2004;Bradburn 2007). Furthermore, all-cause mortality should generally be connected with unbiased estimates (Wood 2008).
Our estimates of increased mortality in low-bias risk trials increased significantly when we excluded factorial trials as well as other trials with collateral interventions. These trials may all suffer from potential confounding from the collateral interventions. This highlights the potential dramatic public health consequences of our results.
A large number of unpublished trials on supplements may exist. Their results are more likely to have been either neutral or negative than to have shown beneficial effects (Dickersin 2003). Accordingly, our estimates of increased mortality are likely to be conservative.

Limitations
Our systematic review has several limitations. As with all systematic reviews, our findings and interpretations are limited by the quality and quantity of available evidence on the effects of specific supplements on mortality. The examined populations varied. The effects of supplements were assessed in the general population or in patients with gastrointestinal, cardiovascular, neurological, skin, ocular, renal, endocrinological, rheumatoid, and undefined diseases. These populations mostly came from countries without overt deficiencies of specific supplements. Accordingly, we are unable to assess how antioxidant supplements affect mortality in populations with specific needs.
We have compared antioxidants with different properties, given at different doses and duration, singly or combined. We are aware of the potential risks in assessing the effects of different types of antioxidants together with different mechanisms of action, biotransformation, and bioavailability. There are pros (Palace 1999;Vertuani 2004;Kawanishi 2005) and cons (Maxwell 1999) in the literature about vitamin A being antioxidant. We fully acknowledge this. Most trials assessed combinations of different supplements, which reflects the way supplements are marketed, sold, and taken by people (Balluz 2000;Millen 2004;Radimer 2004;Nichter 2006).
All available non-enzymatic antioxidants work differently in the human body, and most of them exert effects that are non-antioxidant. We are not able to point to the specific biochemical mechanisms behind the detrimental effects. We found that trials examining the individual supplements singly were rare. It has been suggested that antioxidant supplements may show interdependency and may have effects only if given in combination ( Hercberg 1998).
Most trials investigated the effects of supplements administered at higher doses than those commonly found in a balanced diet, and some of the trials used doses well above the recommended daily allowances and even above the tolerable upper intake levels ( Anonymous 2000a; Anonymous 2000b) (see Figure 4 for overview of recommended dietary allowance, tolerable upper intake level, and experimental doses and regimen used). Our meta-regression analyses revealed significant effects of dose of beta-carotene, vitamin A, and selenium on mortality. The duration of supplementation and follow-up differed among the trials. However, we found no significant effect of treatment duration on our results.

Figure 4. Recommended dietary allowance, tolerable upper intake level, experimental doses, and regimen used in antioxidant supplements
We only assessed all-cause mortality. We are not able to determine the cause of the increased mortality. It is likely that increased cancer and cardiovascular mortality are the main reasons for the increased all-cause mortality (Caraballoso 2003;Vivekananthan 2003;Bjelakovic 2004;Lawson 2007). Further study of causes of mortality is needed. Our results extend previous reviews ( Caraballoso 2003;Vivekananthan 2003;Bjelakovic 2004;Miller 2005;Bjelakovic 2006;Davies 2006;Huang 2006;Bjelakovic 2007a) and guidelines (Ritenbaugh 1999;Atkins 2002;McKevith 2003) suggesting that antioxidant supplements may not be beneficial.
Our overall analyses should be evaluated with care as they include data from trials in which bias is likely as well a data from trials in which bias is less likely. Furthermore, our review include several trials, in which we cannot exclude confounding by other interventions examined in these trials. By excluding data from such trials our relative mortality risk rose from 1.05 to 1.16 in low-bias risk trials. Beta-carotene, administered singly or in combination with other antioxidants had no significant effect on mortality. After exclusion of high-bias risk and selenium trials, beta-carotene singly or combined significantly increased mortality. Recent studies have suggested that beta-carotene may act as a co-carcinogen (Lee 2003;Paolini 2003). Vitamin A used singly or in combination with other antioxidants had no significant effect on mortality. After exclusion of high-bias risk and selenium trials, vitamin A singly or combined significantly increased mortality. Recent research revealed that vitamin A can cause oxidative damage to deoxyribonucleic acid (Murata 2000) We found that vitamin E given singly or combined with four other antioxidants had no significant effect on mortality. However, after exclusion of high-bias risk trials, vitamin E given singly or combined significantly increased mortality. This is in agreement with a recent meta-analysis (Miller 2005). The dose of vitamin E was without significant effect on mortality in our analysis. The chance that vitamin E may benefit seems low (Brown 2005;Devaraj 2005;Guallar 2005). The trials in which vitamin C was applied singly or in different combinations with beta-carotene, vitamin A, vitamin E, and selenium found no significant effect on mortality. According to the confidence intervals, small beneficial or large harmful effects cannot be excluded. We calculated the proportion of participants who died in the trials in which participants took vitamin C alone. In the control group it was 0.019 and in the vitamin C group it was 0.017. With alpha set to 0.05 and power to 0.90, and a minimal relevant difference of 2% the required sample size would be 186,000 participants. We are still far from having examined a sufficient sample. Studies have demonstrated that vitamin C may act as both a pro-oxidant and as an antioxidant in vivo (Podmore 1998;Duarte 2005), current and future trials should be monitored closely for harm. Selenium given singly or in combination with other supplements seemed to significantly decrease mortality, but after exclusion of high-bias risk trials, the effect disappeared. Inspection of the funnel plot of the low-bias risk trials on selenium suggested asymmetry. Results of ongoing randomised trials with selenium will likely increase our understanding of the effects of selenium (Klein  2003). Recently, a randomised trial and an observational study have shown that selenium may carry health risks (Bleys 2007a;Bleys 2007b;Stranges 2007). Therefore, current and any future trials should be monitored closely for harm. Our findings contradict the findings of observational studies, claiming that antioxidants improve health (Machlin 1987;Diplock 1994;van Poppel 1997;Diplock 1998). Considering that more than 10% to 20% of the adult population (80 million to 160 million people) in North America and Europe may consume the assessed supplements (Balluz 2000;Millen 2004;Radimer 2004;Nichter 2006) the public health consequences may be substantial. We are exposed to intense marketing. This is also reflected by the high number of publications per included randomised trial found in the present review. There are several possible explanations for the increased mortality induced by antioxidant supplements. Although oxidative stress has a hypothesised role in the pathogenesis of many chronic diseases, it may be the consequence of pathological conditions (Halliwell 2000). By eliminating free radicals from our organism, we interfere with some essential defensive mechanisms like apoptosis, phagocytosis, and detoxification (Simon 2000;Salganik 2001;Kimura 2005). Recent evidence suggest that inhibition of reactive oxygen species formation in cells decrease the life span of nematodes ( Schulz 2007). Antioxidant supplements are not subjected to the same rigorous toxicity studies as other pharmaceutical agents ( Bast 2002). Better understanding of mechanisms and actions of antioxidants in relation to a potential disease is needed (Ratnam 2006). The methodological quality of some of the trials was assessed using the published reports, which may not reflect the actual design and bias risk of the trials. Only some authors responded to our requests for further information.
Because we examined only the influence of antioxidant supplements, our findings should not be translated to potential effects of fruits and vegetables. Furthermore, we did not examine the treatment effect of antioxidant supplements (tertiary prevention) in specific patient groups or the preventive effects of antioxidant supplements for patient groups with verified specific need for antioxidant supplements. Other systematic reviews should address these issues.

A U T H O R S ' C O N C L U S I O N S Implications for practice
We found no convincing evidence that antioxidant supplements decrease mortality. Even more, beta-carotene, vitamin A, and vitamin E seem to increase mortality. Therefore, we cannot recom-mend the use of antioxidant supplements as a primary and secondary preventive measure in the population groups studied in the present review.

Implications for research
More randomised trials seem to be needed to more firmly establish the potential effects of vitamin C and selenium. Due to the risks that such antioxidant supplements may harm, such trials should have independent data monitoring and safety committees that should regularly review the accumulating data and incorporate these data with the data of our present review as well as other emerging evidence.
Our review highlights the necessity that national and international laws and regulations require that anything sold to the public claiming health benefits should be subjected to adequate assessment of benefits and harms before market release.
The significant association between unclear or inadequate methodological quality and overestimation of intervention effects has again focused on the need for more objective assessment of preventive and therapeutic interventions.
The published trials included in the systematic review lacked important information. We suggest that researchers in future trials should seriously consider adopting the CONSORT

ADCS 1 1997
Methods Alzheimer's Disease Cooperative Study (ADCS 1) Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: unclear, not reported. Allocation concealment: unclear, not reported. Blinding: adequate, identical placebo pills. Follow-up: adequate. The losses to follow-up were 7% in the placebo group, 5% in the selegiline group, 9% in the alpha-tocopherol group, and 6% in the selegiline and alpha-tocopherol group.

Participants
Country: United States of America. Number of participants randomised: 341, older than 18 years, mean age 73 years, 65% females from 23 centres participating in the Alzheimers Disease Co-operative Study Inclusion criteria: patients with probable Alzheimer's disease of moderate severity, as measured by a Clinical Dementia Rating of 2, free of other central nervous system diseases, were not taking psychoactive medications, and were residing either at home or in a supervised setting with a care giver but not in a skilled-nursing facility. Exclusion criteria: none stated.

Interventions
The patients were randomly assigned to receive: group 1: selegiline 5 mg (n = 87), group 2: alpha-tocopherol 1000 IU (n = 85), group 3: selegiline and alpha-tocopherol (n = 85), group 4: placebo (n = 84). Selegiline was given in a dose of 5 mg twice a day, and a racemic mixture of dl-alpha-tocopherol was given in a dose of 1000 IU twice a day. Both agents were given in the morning and in the afternoon for two years.

Outcomes
The primary outcome measure was: the time to the occurrence of any one of the following outcomes: death; institutionalisation; loss of the ability to perform at least two of three basic activities of daily living (ie, eating, grooming, using the toilet), as measured by part 2 of the Blessed Dementia Scale; and severe dementia, defined as a Clinical Dementia Rating of 3. Secondary outcome measures were: measures of cognition, function, behaviour, and the presence or absence of extrapyramidal signs.

Notes
Compliance was monitored in two ways. At each visit, unused medication was returned and the pills were counted. Measures to counter poor compliance included additional phone contact or review of the correct medication dosing schedule with the appropriate caregivers. Compliance was also monitored with surveillance of serum tocopherol concentrations, and the level of selegiline was monitored by measuring amphetamine, its major metabolite, in the urine. Compliance with treatment was good. Urine samples were available from 318 patients for analysis of amphetamine levels. The proportion of patients with positive tests for selegiline was 93% in the combined group, 98% in the selegiline group, 11% in the alpha-tocopherol group, and 13% in the placebo group. Serum samples were available from 332 patients. The proportion of patients with positive tests for alphatocopherol was 91% in the combined group, 93% in the alpha-tocopherol group, 9% in the selegiline group, and 12% in the placebo group. Follow-up was conducted one month after enrolment and at three-month intervals for the remainder of the two-year trial period. At each interval, every effort was made to assess primary and secondary outcomes, regardless of whether an outcome measure had been reached or the medication had been discontinued. Study agents were supplied by Somerset Pharmaceuticals, Tampa, Fla. (Selegiline) and Hoffmann-LaRoche, Nutley, N.J. (alpha tocopherol).

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Allsup 2004Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, using a computer generated list of random numbers in the order in which consent was obtained. Allocation concealment: adequate, the identities of placebo and supplement were kept with the manufacturer (Recip AB, Arsta, Sweden) and were not revealed to the researchers until all data had been analysed. Blinding: adequate, active intervention and placebo had identical appearance.

Outcomes
The primary outcome was: response to influenza vaccine.

Notes
The nursing staff at each home were responsible for the administration of tablets.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

AMDS 1996Low
Methods Age Related Macular Degeneration Study (AMDS) Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, on a computer. Group one and group two patients were randomised between capsule number 1601 (starch placebo) and capsule number 1602 (Ocugard) at each centre by the optometrist co-investigator. Allocation concealment: adequate, coded bottles. Both the capsule for the placebo and antioxidant group (Ocuguard) were formulated by independent laboratories, and an intermediary company was responsible for assigning and maintaining the identity of codes, labelling and distributing masked bottles of capsules to each Medical Centre pharmacy service, informing the Medical Centre of capsule identity in case of adverse reaction, and breaking the code. None of the optometrist and the registered dietician co-investigators and the participants knew of the identity of the capsules. Exclusion criteria: greater than 1 year prior use of vitamins, veterans who were former prisoners of war and veterans who were chronic alcoholics with tobacco/nutritional amblyopia or gastrointestinal absorption disorders.

Outcomes
The primary outcome was: non-exudative age-related macular degeneration.

Notes
Compliance with treatment was not described. Trial agents were provided by Twin Laboratories, Inc., Ronkonkoma, NY.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

AREDS 2001Low
Methods Age Related Eye Disease Study (AREDS). Randomised, double-blind, placebo-controlled trial. Generation of the allocation sequence: adequate, centrally by the Co-ordinating Centre using the on-site computers, with procedures to protect the integrity of randomisation. Multiple levels of data encryption ensured the integrity of the treatment assignment files. Allocation concealment: adequate. Each centre had two treatment assignment databases: one for patients without age-related macular degeneration (AMD); (category 1) containing approximately 100 records consisting of bottle numbers for placebo (five bottle numbers) or antioxidants (five bottle numbers), and one for patients with some AMD (categories 2, 3, or 4) containing approximately 420 records consisting of a different sequence of bottle numbers for placebo and antioxidants (ten bottle numbers each) as well as for zinc and antioxidant and zinc formulations ten bottle numbers each). Each treatment assignment database residing on the hard drives at each Clinical Center is encrypted and includes check numbers to insure tamper-free operation and proper sequential treatment assignments. The computerised randomisation system identified which of the two randomisation tables (category 1 compared to categories 2, 3, or 4) should be used for assigning a treatment, and the participant was randomly assigned a bottle number from the appropriate list. Blinding: adequate, identical placebo tablets. Follow-up: adequate. About 2.3% of participants were lost to follow-up. The rate of participants withdrawal from the trial medication was 14% by 60 months and 15% by the end of trial. This rates include participants lost to follow-up and current smokers, 24% of whom withdrew from the trial medication after the results from the clinical trials of beta-carotene and lung cancer were announced. Overall, the vital status was known for 4753 out of 4757 particcipants. Intention-to-treat analysis: yes. Sample-size calculation: yes.

Participants
Country: United States of America. Number of participants randomised: 4757; 56% female, aged 55 to 80 years, median age 68 years. Inclusion criteria: participants free of any illness or condition that would make long-term follow-up or compliance with study medications unlikely or difficult. On the basis of fundus photographs graded by a central reading centre, best corrected visual acuity, and ophthalmologic evaluations, participants were enrolled in one of several AMD categories. At least one eye of each participant was free from eye disease that could complicate assessment of AMD, lens opacity progression, or visual acuity, and that eye could not have had previous ocular surgery (other than cataract surgery). Exclusion criteria: illness or disorders (eg, history of cancer with a poor 7-year prognosis, major cardiovascular or cerebrovascular event within the last year, haemachromatosis) that would make long term followup or compliance with the study protocol unlikely or difficult. Persons bilaterally aphakic or pseudophakic were ineligible for AMD category one.

Interventions
Participants were divided into two clinical trials: AMD Trial (n = 128) and Cataract Trial (n = 4629).

AREDS 2001Low (Continued)
Two study medication tablets were to be taken each morning and two each evening, to meet the total daily dose requirement. Tablets were to be taken with food to avoid potential irritation of an empty stomach by zinc. Participants were followed up for an average of 6.3 years.
Outcomes Primary outcome measures were: an increase from baseline in nuclear, cortical, or posterior subcapsular opacity grades or cataract surgery, and at least moderate visual acuity loss from baseline (> 15 letters).

Notes
Compliance with treatment was checked by random serum assessments. Compliance with treatment was excellent. Overall adherence was estimated to be 75% or greater (participants took > 75% of their study tablets) for 70% of participants at five years. At 60 months, 20% of participants (20% both for current smokers and former or non-smokers) reported taking some multivitamin supplement containing at least one of the study medication. In 1994 and 1996, AREDS participants were informed of the results of the Alpha-Tocopherol, Beta Carotene Cancer Prevention Study and the Beta-Carotene and Retinol Efficacy Trial suggesting potential harmful effects of beta-carotene among smokers. The trial was supported from Bausch & Lomb Inc, Rochester, NY. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

ASAP 2003Low
Methods The Antioxidant Supplementation in Atherosclerosis Prevention (ASAP) Study. Randomised, partially double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, computer generated randomisation lists. Participants were randomised separately in four strata of approximately equal size. Allocation concealment: adequate, the masking was carried out by the provider of the supplements and delivered to the data centre of the Field Centre, Research Institute of Public Health, University of Kuopio. Blinding: adequate, tablets were identical in appearance, size, and colour. Follow-up: adequate. Of the 520 participants randomised, 440 (84.6%) completed the study and underwent the 6-year re-examination. overall, 55 participants in the three vitamin groups and 25 participants in the placebo group were lost to follow-up. Intention-to-treat analysis: yes. Sample-size calculation: yes.
Participants Country: Finland. Number of participants randomised: 520, 256 men and 264 postmenopausal women, smoking and non smoking, aged 45 to 69 years with serum cholesterol > 5 mmol/L (193 mg/dL). Inclusion criteria: participants with hypercholesterolemia defined as serum cholesterol levels > 5 mmol/L (193 mg/dL). Exclusion criteria: regular intake of antioxidants, acetosalicylate, or any other drug with antioxidative properties, severe obesity (body mass index >32 kg/m2), type 1 diabetes, uncontrolled hypertension (sitting diastolic blood pressure >105 mm Hg), any condition limiting mobility, or severe disease shortening life expectancy. Premenopausal women and those taking oral oestrogen therapy were also excluded.

ASAP 2003Low (Continued)
Interventions The study consisted of 8-week dietary counselling and placebo lead-in phase, a 3-year double-masked treatment period, and a 3-year open treatment period. The participants were randomly allocated to receive twice daily with meal: group 1: d-alpha tocopherol 91 mg (corresponding to 100 mg of d-alpha-tocopheryl acetate and 136 IU of vitamin E) (n = 130); group 2: 250 mg slow-release vitamin C (n = 130); group 3: both d-alpha-tocopherol and slow-release ascorbic acid in a single tablet (CellaVie), (n = 130); group 4: placebo only (n = 130); for a period of 6 years.

Outcomes
The primary outcome measure was: progression of carotid atherosclerosis.

Notes
Compliance with treatment was checked by random serum assessments. Of the 390 participants randomised to supplementation, 335 continued the study after 3 years and 256 (76.4%) took the supplements as instructed for 6 years, whereas 62 participants stopped the supplements during the first 3 study years and additional 18 participants during the last 3 study years. The mean plasma alpha-tocopherol and ascorbate concentration increased in 6 years in the group randomised to supplementation, and in the unsupplemented group decrease. Ferrosan A/S, Denmark, provided the vitamin supplements.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

ATBC 2003Low
Methods Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC). Randomised, double-blind, placebocontrolled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, centrally by computer. Randomisation performed in blocks of eight within each of the study areas.

Interventions
Participants were randomly assigned in four groups to receive: group 1: alpha-tocopherol 50 mg (n = 7286); group 2: beta-carotene 20 mg (n = 7282); group 3: alpha-tocopherol and beta-carotene, (n = 7278); ATBC 2003Low (Continued) group 4: placebo (n = 7287); daily for five to eight years (median 6.1 years). All participants took a single capsule daily. The four trial intervention groups were well balanced for all baseline characteristics evaluated. The two-by-two factorial design allowed assessment of the two intervention agents independently, with one-half of participants receiving alpha-tocopherol (n = 14,564) and the other half not (n = 14,569); similarly, half of the participants received beta-carotene (n = 14,560) and half did not (n = 14,573).
The study was conducted between 1985 and 1993 (mean 6.1 years). The active intervention continued through April 30, 1993 and postintervention follow-up until April 30, 2001. Mean follow-up time regarding incident cancers and cause-specific deaths was 12.1 years and overall mortality 14.1 years.

Outcomes
The primary outcome measure was: incidence of lung cancer. Secondary outcome measures were: incidence of other major cancers, overall and cause specific mortality and incidence of other diseases.

Notes
Compliance with treatment was assessed by counts of the remaining capsules at each visit, by measurement of serum alpha-tocopherol and beta-carotene levels after three years of supplementation, and by measurements in random serum samples throughout the study. Compliance with treatment was excellent with four out of five active participants taking more than 95% of the scheduled capsules. Dropout rate and compliance were similar between all four groups. All capsules were supplied by Hoffmann-La Roche, Basel, Switzerland. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Bonelli 1998
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported. Exclusion criteria: Familiar adenomatous polyposis, inflammatory bowel diseases, polypectomy performed more than six months before randomisation, adenoma with invasive carcinoma, previous colorectal resection, invasive cancer at any site, life-threatening chronic heart, liver or kidney diseases, current use of vitamin or calcium supplements, mental disability precluding informed consent to participate and adherence to the treatment, patients with 10 adenomas or more and those with large sessile adenomas (3 cm or more in diameter). The clean colon after polypectomy was assessed by means of total colonoscopy. When a total colonoscopy was not feasible a double contrast barium enema was performed. Colonoscopy was scheduled on years one, three and five after randomisation.

Outcomes
The primary outcome measure was: occurrence of metachronous adenomas detected at follow-up endoscopic examinations.

Notes
The overall 5 year actuarial compliance to the treatment was 51%. Of the 304 randomised patients, 233 (76.6%) underwent at least one endoscopic follow up examination: 117 in the active compound group and 116 in the placebo group. Active intervention and placebo were provided by Pharma Nord. Additional information obtained through personal communication with authors.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

CARET 2004Low
Methods The Beta-Carotene and Retinol Efficacy Trial (CARET). Randomised, double-blind, placebo-controlled trial with two-by-two factorial design in a pilot phase and then one-by-one. Generation of the allocation sequence: adequate, permuted block design with random block size chosen uniformly among 8, 10, 12, 14, and 16. Allocation concealment: adequate, locked central database with the link between study identifier and intervention assignment; all data analyses were performed centrally; the analyses that involved intervention assignment were performed only by the Co-ordinating Center's statisticians using coded intervention assignment unknown to the statisticians; analyses involving the coded intervention assignments were seen only by CARET's Data and Safety Monitoring Board, Co-ordinating Center statisticians, and a single CARET investigator who saw no participants. Blinding: adequate, identical placebo capsules provided by a sponsor. Follow-up: adequate. The losses to follow-up were less than 2% at the end of treatment. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America. Number of participants randomised: 18314; 12025 males and 6289 females. Inclusion criteria: smokers, former smokers, and workers exposed to asbestos at high risk of developing lung cancer. A total of 4060 male workers, mean age 57 years, exposed to asbestos and 14254 heavy smokers (44% of whom were women), mean age 58 years, were randomised. The participants agreed to limit their supplemental intake of vitamin A to less than 5500 IU per day and to take no supplemental beta-carotene.

Interventions
CARET builds on the experience of two pilot studies performed in Seattle (1985)(1986)(1987)(1988). The first pilot study initiated a phase III trial of the safety and efficacy of the study vitamins in 816 asbestos-exposed participants randomised to a daily combination of 15 mg 13-carotene and 25,000 IU retinol or a placebo medication. Participants were eligible up to age 74 and were not required to have a history of cigarette smoking; otherwise, the eligibility criteria were the same as for the asbestos-exposed population in CARET. The second pilot study was a phase II trial of the comparative safety of the study vitamins in heavy smokers. The eligibility criteria were identical to those for heavy smokers in CARET. Overall 539 men and 490 women were randomised to one of four intervention groups: group 1: a daily combination of 30 mg 13carotene and 25,000 IU retinol; group 2: 30 mg 13-carotene only; group 3: 25,000 IU retinol only; group 4: placebo medication. All 1845 participants in the two pilot studies continue to be followed for outcomes in CARET, together with approximately 16,000 additional participants. Participants of CARET trial were randomly assigned to receive: group 1: combination of 30 mg beta-carotene and 25,000 IU vitamin A (n = 9420); group 2: placebo, (n = 8894). Both formulations were given as capsules. Beta-carotene beadlets were combined with retinyl palmitate in a single capsule and dispensed in bottles, which were weighed and their content checked. The design projected active intervention until late 1997. The CARET active intervention was stopped 21 months earlier because of clear evidence of no benefit and substantial evidence of possible harm. The average duration of follow-up was 10.0 years

Outcomes
The primary outcome measure was: the incidence of lung cancer. Other outcomes reported are: mortality rates, and incidence of other cancers.

Notes
Compliance was assessed by weighing the returned bottles to estimate the number of capsules remaining (in 85% of the assessments), or by relying on the participants own estimates (in 15% of the assessments). Compliance with treatment was excellent. Among the active participants, the mean rate of capsule consumption was 93% through five years of follow-up, with no significant differences between the treatment groups. Participants who stopped receiving study vitamins for any reason other than death were defined as inactive participants and were still followed for outcomes and counted in the analysis. As of December 15, 1995 ascertainment of vital status for more than 98% was complete. Active agents and placebos were CARET 2004Low (Continued) purchased from Hoffmann-La Roche and formulated by Tischon Corporation. Data were extracted from the primary publication, but additional information was received through personal communication with the authors. Influenza vaccine was given four weeks before the end of the study. Participants with infection were treated appropriately with antimicrobial agents and supportive measures.

Outcomes
The primary outcome measures were: immunocompetence and occurrence of infection related illness.

Notes
Compliance was verified by interview at fortnightly visits and counting of leftover medication. There was no report about the compliance.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

CHAOS 1996Low
Methods Cambridge Heart Antioxidant Study (CHAOS

Outcomes
The primary outcome measures were: non-fatal myocardial infarction alone and combination of non-fatal myocardial infarction and cardiovascular death.

Notes
Compliance with treatment was measured as the ratio of days that study medication was requested to perprotocol days prescribed. 73.2% of all prescribed alpha-tocopherol or placebo were requested as followup medications. There was no difference between treatment groups in the proportion of participants who were 100% compliant with the trial medication (48% placebo, 49% alpha-tocopherol

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Collins 2003Low
Methods Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, using computer generated permuted blocks. Allocation concealment: adequate, for the vitamin E randomisation sequence was given to research pharmacist and the prescription bottles were coded. Experimental and control assignments were in sealed envelopes.
Blinding: adequate, identical placebo capsules. Follow-up: adequate. Six randomised participants did not complete the study; 1 participant from the exercise and vitamin E group, 3 participants from exercise plus placebo group, and 2 participants from the placebo group. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America. Number of participants randomised: 52, mean age 67, 98% males. Inclusion criteria: current diagnosis of peripheral arterial disease, a history of intermittent claudication, and an ankle-brachial index < 0.95 at rest and/or < 0.85 after exercise. Exclusion criteria: taking any of the following drugs, vitamin E, Coumadin, or pentoxifylline.

Interventions
Participants were randomly assigned in four groups to receive: group 1: PoleStriding exercise with vitamin E (n = 13); group 2: PoleStriding exercise with placebo (n = 14); group 3: vitamin E without PoleStriding exercise (n = 13); group 4: placebo without exercise (n = 12). The dose of vitamin E was 400 IU daily. Participants were supplemented 0.5 year, and followed 2.5 years.
PoleStriding is a form of walking that uses muscles of the upper and lower body in a continuous movement similar to cross country skiing.

Outcomes
The primary outcome was: walking ability and perceived quality of life.

Notes
Compliance with the study drug treatment was monitored in two ways: patient self-report and measured vitamin E levels. Vitamin E levels were obtained at baseline and 3 and 6 months. Investigators did not receive the measured vitamin E levels until the trial ended. For the first 3 months drug compliance was assessed biweekly by the study staff and monthly thereafter. Vitamin E and placebo capsules were provided by the Henkel Corporation, La-Grange, IL). Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Correa 2000Low
Methods Randomised, controlled, partially double-blind, chemoprevention trial with two-by-two-by-two factorial design. Generation of the allocation sequence: adequate, computer-generated lists. Participants were randomly assigned in a single step, using a permuted block design, to one of eight different treatment regimens. Before randomisation, participants were classified into one of three strata: atrophy (without metaplasia), intestinal metaplasia, or dysplasia, according to baseline histologic diagnosis.

Interventions
Participants were assigned to a dietary supplement of beta-carotene (30 mg once per day) and/or ascorbic acid (1 g twice a day) or their corresponding placebos, for a six-year period. The prevalence of Helicobacter pylori infection among all gastric biopsy specimens was 97%. Anti-Helicobacter pylori treatment consisting of amoxicillin (500 mg three times per day), metronidazole (375 mg three times per day), and bismuth subsalicylate (262 mg three times per day) was given for 14 days to half of the study participants assigned randomly. This treatment was not blinded or placebo controlled because an appropriate placebo was not available for bismuth subsalicylate. Participants were divided in eight treatment groups to receive: group 1: placebo (n = 117); group 2: anti-Helicobacter pylori treatment, which consisted of amoxicillin, metronidazole, and bismuth subsalicylate (n = 120); group 3: beta-carotene (n = 117); Correa 2000Low (Continued) group 4: ascorbic acid (n = 130); group 5: Helicobacter pylori treatment, which consisted of amoxicillin, metronidazole, and bismuth subsalicylate, and additionally beta-carotene (n = 126); group 6: Helicobacter pylori treatment, which consisted of amoxicillin, metronidazole, and bismuth subsalicylate, and additionally ascorbic acid (n = 111); group 7: beta-carotene and ascorbic acid (n = 121); group 8: Helicobacter pylori treatment, which consisted of amoxicillin, metronidazole, and bismuth subsalicylate, and additionally beta-carotene and ascorbic acid (n = 134). Gastric biopsy specimens taken at baseline were compared with those taken at 72 months.

Outcomes
The primary outcome measures were: progression, no change or regression of gastric precancerous lesions (preliminary histologic diagnosis of multifocal atrophic gastritis with or without intestinal metaplasia and dysplasia). For our purposes we extracted data about the incidence of gastric cancer.
We have also extracted data on overall mortality for all antioxidants as well as for beta-carotene and vitamin C.

Notes
Compliance with treatment was constantly encouraged and monitored by a social worker who interviewed the participants and recorded pill counts every three months. In addition, blood levels of beta-carotene and ascorbic acid were measured every three months in a 20% random sample of the participants. Compliance with treatment among participants who completed the study was high for all intervention modalities (mean compliance for ascorbic acid, 91.8%; for beta-carotene, 92.3%; and for anti-Helicobacter pylori treatment, 99.1%). Active medication and placebos were provided like identical coded tablets by Hoffmann-La Roche Inc.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

DATATOP 2005Low
Methods The Deprenyl and Tocopherol Antioxidative Therapy of Parkinsonism (DATATOP). Randomised, double-blind, placebo-controlled secondary prevention trial with two-by-two factorial design. Generation of the allocation sequence: adequate, centrally by computer.

Interventions
Patients were randomly assigned to receive: group 1: vitamin E (dl-alpha-tocopherol; all-racemic) 1000 IU; group 2: deprenyl 5 mg; group 3: tocopherol and deprenyl; group 4: placebo; twice daily with morning and evening meals. 401 participants were assigned to tocopherol and 399 participants to deprenyl. Participants were instructed to take one tablet and one capsule twice daily with morning and evening meals. A standard multivitamin containing vitamin E (30 IU) was provided to all participants. Median duration of vitamin E exposure during the randomised phase was 2.6 years. Preliminary analysis in the fall of 1989, after an average 1.5 years of follow-up, indicated unexpectedly striking effects of deprenyl in postponing PD disability as measured by the need for levodopa therapy. After this disclosure, all active trial participants, whether or not they required levodopa therapy, were placed on open-label deprenyl, 10 mg/day, for about 3.5 years, from fall 1989 to spring 1993. Blinded tocopherol treatment assignments were maintained for about 3 years after the initial randomisation. Participants began taking levodopa (with carbidopa, a peripheral dopa decarboxylase inhibitor) in addition to their experimental treatments at any point in the trial when they were judged clinically to require therapy for emerging disability. Investigators adjusted levodopa dosage to achieve optimal clinical benefits and avoid dopaminergic adverse effects. Because of concerns about the sustained benefit of deprenyl, a second randomisation was undertaken in spring 1993. Consenting research participants who required levodopa were randomised, independently of their original randomisation, to continue deprenyl (50%) or to switch to deprenyl placebo (50%). Further adjustments of levodopa dosage were permitted after the second randomisation. This additional placebocontrolled phase of deprenyl assignment was continued for 2 years until the last formal (face-to-face) clinical evaluation in spring 1995. The 800 participants have therefore been followed at least annually for an average of 8.2 years. Participants who underwent the second randomisation had a minimum of 3.2 years and a maximum of 7.3 years of exposure to active deprenyl. Participants were followed 13 years.

Outcomes
The primary outcome in the trial occurred when, in the judgement of the enrolling investigator, a participant reached a level of functional disability sufficient to warrant the initiation of levo-dopa therapy. Operationally, the primary response variable in the trial was defined as the time from randomisation to the end point. After the outcome, the experimental treatments were withdrawn in blinded fashion, and approximately 30 days later the participants received a final evaluation.

Notes
Monitoring of compliance was carried out in follow-up evaluations in which unused doses return by the subject were counted, the serum tocopherol levels measured, and the urinary levels of amphetamine and metamphetamine metabolites of deprenyl determined. The results of the compliance monitoring were not shared with the participants or the investigators. Compliance in taking experimental medications was excellent among all treatment groups. The overall compliance rate, as a percentage of the doses dispensed that were actually taken, ranged from 97.9 to 99.5 percent for both tocopherol and deprenyl. Tablets of ldeprenyl and matching placebos were provided by Someret Pharmaceuticals, Denville; N.J. Capsules of dl-alpha tocopherol and matching placebos were provided by Hoffman-LaRoche, Nutley, N.J. A standard vitamin (One-A-Day) by Miles Laboratories, Elkhart, Ind.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

DATOR 2004Low
Methods Randomised, single-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, block randomisation using the Geigy manual randomisation tables. Allocation concealment: adequate, sealed envelopes. Blinding: adequate, identical placebo capsules. Follow-up: adequate. During the course of the trial, 1 participant from each group dropped out -1 participant due to thyroid dysfunction and 1 participant due to an accident. Intention-to-treat analysis: no. Sample size calculations: no.

Outcomes
The primary outcome measure was: impact on lipids and peroxidation during statin treatment.

DATOR 2004Low (Continued)
Notes Omega-Pharma NV is acknowledged for the supply of alpha-tocopherol and placebo. Additional information received through personal communication with the authors.
Due to the addition of 0.25 mg tocopherol to the control group, the 'placebo' control of the trial can be discussed.

Interventions
Patients were randomly assigned to receive: group 1: vitamin E 500 mg (in the form of alpha-tocopheryl acetate, n = 37; group 2: placebo, n = 37. Participants were supplemented and followed 1 year.

Outcomes
The primary outcome measure was: the liver function, mortality, and hospitalisation rates.

Notes
Patients were seen once a month by a nurse practitioner at the liver disease clinic. Patients were asked about compliance to the treatment, which was assessed by counting the leftover tablets. Blood samples were obtained at the beginning of the study and every 3 months to measure serum levels of vitamin E. Financing was provided by Roche and Saval Laboratories. Exclusion criteria: diabetes patients, users of (multi)vitamin-, vitamin E, vitamin C, beta-carotene, garlic, or fish oil supplements, users of vitamin K antagonists (phenprocoumon, acenocoumarol), individuals with current illness interfering with participation, and unwillingness to participate.

Outcomes
The primary outcome measure was: progression of atherosclerosis in lifelong male smokers measured by 2-year change of the common carotid intima media thickness as measured by B-mode ultrasonography.

Notes
Compliance with treatment is not reported. Vitamin E or placebo capsules were provided by F Hoffman La Roche Ltd, Basel.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Desnuelle 2001Low
Methods Amyotrophic Lateral Sclerosis riluzole-tocopherol study (ALSRT). Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, randomisation schedule prepared by laboratory. Randomisation was balanced by centre. Exclusion criteria: signs of dementia and/or major psychiatric disorders, another concomitant serious disease, or handicap likely to interfere with their assessment of survival, forced vital capacity of less than 60%, monoclonal gammopathy, conduction blocks of motor nerves on electromyography, hepatic or renal disfunction, pregnancy or breast feeding, creatinine plasma concentration above 200 µM, alanine aminotransferase and/or aspartate transaminase activity greater than twice the upper limit of the normal range, known hepatic disease, taking drugs known to be hepatotoxic, enzyme-inducing or enzyme-inhibiting, taking vitamin E.

Outcomes
The primary outcome measure was: change in functional status of each patient using the modified Norris limb scale. The secondary outcome measures were: survival (defined as the time to death or tracheostomy), bulbar function assessed with the Norris bulbar scale (total possible score 39) and manual muscle testing.

Notes
Compliance with treatment was checked by measuring the plasma vitamin E levels.
Compliance with treatment good. In the vitamin E group, a highly significant increase in plasma levels of vitamin E was observed after 3 months of treatment. Study agents were supplied by: alpha-tocopherol (Toco 500R) Laboratories Rhone-Poulenc Rorer (now trading under the name Laboratories Aventis); riluzole Rhone-Poulenc Rorer. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Gillilan 1977
Methods Randomised, double-blind, placebo-controlled cross-over trial. Generation of the allocation sequence: unclear, not reported.

Interventions
Participants were randomly assigned to receive: group 1: vitamin E (d-alpha-tocopherol succinate) 1600 IU (n = 26); group 2: placebo (containing 2.5 mg of riboflavin), (n = 26); three capsules of vitamin E (400 IU) or placebo daily for a period of six months, and then cross-over. The mean duration of double-blind therapy was 189 days of vitamin E and 192 days of placebo.

Outcomes
The primary outcome measure was: any improvement of angina pectoris.

Notes
Drug adherence was followed by capsule count and a urine fluorescence test. Serum vitamin E levels were measured at baseline and at the end of the first and six months of each treatment phase. The capsule count data shows a mean consumption of 88% of the prescribed capsules during vitamin E phase, and 84% consumption during placebo therapy. The percent of urine specimens with fluorescence indicate 78% of taking placebo prescribed medication. The serum tocopherol levels were significantly higher during the supplementation. Vitamin E and placebo capsules supplied by Wilson and Wolfer Pharmaceutical Manufacturers and Distributors, Detroit, Michigen.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Girodon 1997
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported.

Interventions
Participants were randomly assigned to receive: group 1: placebo (n = 20); group 2: trace elements (zinc 20 mg in a form of zinc sulfate; selenium 100 µg in a form of selenite) (n = 20); group 3: vitamins (vitamin C 120 mg; beta-carotene 6 mg; vitamin E 15 mg) (n = 20); group 4: combination of trace elements and vitamins at equal doses (n = 21); daily (one capsule a day) for a period of 2 years. Mean duration of follow-up was 730 days.

Outcomes
The primary outcome measure was: impact of a trace element and vitamin supplementation on infectious morbidity.

Notes
Compliance with treatment was checked by measuring the plasma vitamin levels and counting of the returned capsules. Compliance was good. After 6 months of supplementation, a significant increase in vitamin and trace element serum levels was obtained in the corresponding treatment groups: a plateau was then observed for the whole study. No changes appeared in the placebo group. Study agents were provided by Produits Roche SA.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

GISSI 1999
Methods GISSI-Prevenzione trial (Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto miocardico). Randomised clinical secondary prevention trial with two-by-two factorial design. Generation of the allocation sequence: adequate, computer programme based on the biased-coin algorithm, stratified by hospital.

Outcomes
The primary outcome measures were: cumulative rate of all-cause death, non-fatal myocardial infarction, and non-fatal stroke; and the cumulative rate of cardiovascular death, non-fatal myocardial infarction, and nonfatal stroke.

Notes
Compliance with treatment was measured by refilling drug supplies every three months. Compliance with assigned treatment was excellent. At year one and at the end of the study, 11.6% and 28.5% of patients receiving n-3 PUFA and 7.3% and 26.2% of those receiving vitamin E, respectively, had permanently stopped taking the study drug. Conversely, during the whole course of the study, only two patients not assigned vitamin E and 26 patients not assigned n-3 PUFA were receiving these drugs. The trial was supported by grants from Bristol-Myers Squibb, Pharmacia-Upjohn, Società Prodotti Antibiotici, and Pfizer. Pharmacia-Upjohn and Società Prodotti Antibiotici supplied marketed capsules containing 850 to 882 mg EPA/DHA ethyl esters. Vitamin E (acetyl d, l-a-tocopherol) was supplied by Bracco.

Graat 2002Low
Methods Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, a computer-generated, 4-per-block, randomisation list. Allocation concealment: adequate, randomisation list was created by the pharmacy allocating treatment to participant number. Numbered boxes containing identical looking capsules were transported from the pharmacy to the Wageningen University. At enrolment, boxes were assigned consecutively to participants. Treatment allocation was kept at the pharmacy exclusively in sealed opaque envelopes while participant identity was known exclusively at the Wageningen University. None of the treatment codes was broken during the study period. Blinding: adequate, identical looking placebo capsules. Follow-up: adequate. In total, 16% of the participants discontinued the treatment. Overall, 26 participants assigned to receive multivitamin-mineral, 25 to vitamin E, 26 to multivitamin-mineral plus vitamin E, and 20 to placebo were lost to follow-up. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: The Netherlands. Number of participants randomised: 652, 325 men and 327 women, older than 60 years. Inclusion criteria: noninstitutionalized elderly persons older than 60 years. Exclusion criteria: used immunosuppressive treatment, anticoagulants interfering with vitamin K metabolism, or dietary supplements in the previous 2 months or if they had a history of cancer, liver disease, or fat malabsorption during the 5 years before randomisation.

Outcomes
The primary outcomes were: incidence and severity of acute respiratory tract infections.

Graat 2002Low (Continued)
Notes Compliance with treatment was checked by measuring the plasma vitamin levels and counting of the returned capsules. Baseline plasma samples were collected for determination of alpha-tocopherol, ascorbic acid, retinol, and carotenoids. To monitor compliance, these assessments were repeated in a postintervention sample of a subset (n = 300). Returned capsules were counted for all participants. After treatment, ascorbic acid, total carotenoids, alpha-tocopherol, and cholesterol-adjusted alpha-tocopherol levels increased significantly in the multivitamin-mineral and multivitamin-mineral plus vitamin E group, while gamma-tocopherol decreased significantly. In the vitamin E group, alpha-tocopherol and cholesteroladjusted alpha-tocopherol levels increased significantly, while gamma-tocopherol levels decreased significantly. In the placebo group, none of the measured vitamins changed significantly. 94% of the participants met the compliance criteria of 80% capsule intake. Trial agents were provided by Roche Vitamins, Europe, Basel, Switzerland.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Graf 2005Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, computer generated. Allocation concealment: adequate, central independent unit. Blinding: adequate, identical looking capsules. Follow-up: adequate, no losses to follow-up. Intention-to-treat analysis: yes. Sample size calculations: no.
Participants Country: Germany. Number of participants randomised: 160, 104 males and 56 females, mean age 58 years with probable or definite amyotrophic lateral sclerosis (ALS). Inclusion criteria: patients with probable or definite amyotrophic lateral sclerosis (ALS) treated with riluzole and disease duration of less than 5 years. Exclusion criteria: none stated.

Outcomes
Primary outcome measure was: survival, calculating time to death, tracheostomy, or permanent assisted ventilation. Secondary outcome measures were: the rate of deterioration of function assessed by the modified Norris limb and bulbar scales, manual muscle testing, spasticity scale, ventilatory function and the Sickness Impact Profile.

Graf 2005Low (Continued)
Additional information obtained through personal communication with authors.

Notes
Compliance with treatment was checked by measuring the plasma vitamin E levels. Vitamin plasma levels were, as expected, significantly higher in the high dose vitamin E group than in the placebo group. Trial agents were provided by Schwarzhaupt, Cologne, Germany. Additional information obtained with personal communication with authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

HATS 2001Low
Methods The HDL-Atherosclerosis Treatment Study (HATS). Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, random numbers generated by computer.

HATS 2001Low (Continued)
Simvastatin therapy began at 10 mg per day for patients with an LDL cholesterol level of 110 mg per decilitre (2.84 mmol per litre) or lower on screening and 20 mg per day for those with an LDL cholesterol level higher than 110 mg per decilitre. The dose was increased by 10 mg per day in patients whose LDL cholesterol level was higher than 90 mg per decilitre (2.33 mmol per litre) in any sample during the first year of the study and was reduced by 10 mg per day if the LDL cholesterol level fell below 40 mg per decilitre (1.03 mmol per litre) at any time during the study. During treatment, patients receiving the matching placebo were given 10 mg of simvastatin if their LDL cholesterol level was 140 mg per decilitre (3.62 mmol per litre) or higher; the target level was 130 mg per decilitre (3.37 mmol per litre) or lower. The dose of slow-release niacin was increased linearly from 250 mg twice daily to 1000 mg twice daily at four weeks. Patients whose HDL cholesterol levels had not increased by at least 5 mg per decilitre (0.13 mmol per litre) at 3 months, at least 8 mg per decilitre (0.21 mmol per litre) at 8 months, and at least 10 mg per decilitre at 12 months were switched to crystalline niacin the dose of which was gradually increased to 3 g per day or, at most, 4 g per day in order to meet the target levels. Niacin "placebo" tablets (taken at a dose of 50 mg twice daily) were active, provoking flushing without affecting lipid levels.

Outcomes
The primary outcome measures were: arteriographic evidence of a change in coronary stenosis and the occurrence of a first cardiovascular event (death, myocardial infarction, stroke, or revascularisation).

Notes
Compliance with the trail regimens, measured by means of pill counts, ranged between 80 percent and 95 percent. The mean doses of simvastatin and niacin taken by patients were 13 ± 6 mg per day and 2.4 ± 2.0 g per day, respectively. Plasma vitamin concentrations increased significantly in 75 patients who received active vitamin therapy. The active agents and placebos were provided by: Simvastatin (Zocor, Merck, West Point, Pa.) slow-release niacin (Slo-Niacin, Upsher-Smith, Minneapolis) crystalline niacin (Niacor, Upsher-Smith).

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Hogarth 1996
Methods Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: unclear, not reported.

Outcomes
The primary outcome measures were: weight, serum albumin levels, and activities of daily living, cognitive functioning and length of stay.

Notes
Compliance with the energy supplement (active or placebo) was monitored by measuring unconsumed fluid each day during admission. Following discharge, patients or carers were asked to complete a form estimating the volume of fluid (in quarters) remaining in each bottle each day. Vitamin compliance was monitored by tablet count at the end of the 1-month period at the final assessments. Compliance was poor with the liquid energy supplement with only one-third of patients consuming > 50% of offered drinks during the study period (17/55 patients, active group; 16/51 patients, placebo group). Vitamin capsule compliance was higher with approximately 90% of patients taking more than 50% of the capsules provided (48/52 patients, active group; 49/54 patients, placebo group). The energy supplement (Lucozade) and placebo preparation were provided by SmithKline and Beecham.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Methods
The Heart Outcomes Prevention Evaluation Study (HOPE). Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, by computer. Allocation concealment: adequate, randomisation is done by a telephone call to a central office.
After receipt of appropriate baseline data over the telephone, the patient is randomised.

Interventions
Patients were randomly assigned to receive either group 1: 400 IU of vitamin E (RRR-a-tocopheryl acetate) daily from natural sources (n = 4761); or group 2: matching placebo (n = 4780); or group 3: an angiotensinconverting-enzyme inhibitor (ramipril 10 mg) (n = 4645); or group 4: matching placebo (n = 4652), once a day for a four to six years, mean 4.5 years. The Heart Outcomes Prevention Evaluation [HOPE] trial is conducted between December 21, 1993, and April 15, 1999. The Heart Outcomes Prevention was extended (HOPE-The Ongoing Outcomes [HOPE-TOO]) between April 16, 1999, andMay 26, 2003. Of the initial 267 HOPE centres that had enrolled 9541 patients, 174 centres participated in the HOPE-TOO trial. Of 7030 patients enrolled at these centres, 916 were deceased at the beginning of the extension of the trial, 1382 refused participation, 3994 continued to take the study intervention, and 738 agreed to passive follow-up. The mean follow-up period was 7 years.

Outcomes
The primary outcome measures were: cancer incidence, cancer deaths, major cardiovascular events (myocardial infarction, stroke, and cardiovascular death). The secondary outcomes were: unstable angina, congestive heart failure, revascularisation or amputation, death from any cause, complications of diabetes, and cancer.

Notes
Compliance with treatment was checked by measuring the plasma vitamin E levels in randomly selected patients. The rate of compliance with the assigned regimen was high throughout the study.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

HPS 2002Low
Methods Heart Protection Study. Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, computer generated random numbers. Inclusion criteria: adults with coronary disease, other occlusive arterial disease, or diabetes, and non-fasting blood total cholesterol concentrations of at least 3.5 mmol/L. Exclusion criteria: other life-threatening conditions, such as chronic liver disease, severe renal disease, severe heart failure, severe chronic airways disease, or diagnosed cancer (other than non-melanoma skin cancer). In addition, anyone already taking high-dose vitamin E supplements, or in whom such supplements were considered indicated, was not to be randomised.

Interventions
Participants were randomly assigned to receive: group 1: 600 mg vitamin E, 250 mg vitamin C, and 20 mg beta-carotene daily (n = 10,269); or group 2: matching placebo capsules (n = 10,267), daily during the scheduled 5-year treatment period.

Outcomes
The primary outcome measures were: major coronary events (for overall analyses) and fatal or nonfatal vascular events (for subcategory analyses), with subsidiary assessments of cancer and of other major morbidity.

Notes
Compliance with treatment was assessed at each follow-up by reviewing the calendar packed tablets remaining and, for those who had stopped, the reasons for doing so were sought. An average of 83% of participants in each treatment group remained compliant during the scheduled five-year treatment period. To assess the effects of the treatment allocation on blood concentrations of the vitamins being studied, assays were performed in non-fasting samples collected from about 5% of participants at the initial screening visit and at an average of about three years of follow-up (the approximate mid-point of the study). Vitamins were provided by Roche.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

LAST 2004Low
Methods Lutein Antioxidant Supplementation Trial (LAST Inclusion criteria: diagnosis of atrophic age-related macular degeneration (ARMD) by stereo bio-ophtalmoscopy and at least one vision-degrading visual-psychophysical abnormality associated with ARMD in one or both eyes; clear non-lenticular ocular media (cornea, aqueous and vitreous), free of advanced glaucoma and diabetes or any other ocular or systemic disease that could affect central or parafoveal macular visual function. Exclusion criteria: recent (within 6 months) cataract or retinal surgery; taking photosenzing drugs (such as phenotiazines and chloroquine).

Outcomes
The primary outcome measures were: visual function and symptoms in atrophic age-related macular degeneration (ARMD).

Notes
Compliance was assessed by telephone at one week, two weeks, four weeks, six weeks, three months, and 12 months. Compliance with treatment was good. During one-year study, 96% of the participants took approximately 92% of their assigned capsules. There was no difference in compliance among the three groups. Lutein (Floraglo R) was provided by Kemin Foods International, Des Moines, Iowa); lutein in combination with additional antioxidants and nutrients (OcuPower R) and placebo were provided by Nutraceutical Sciences Institute, Boynton Beach, Florida.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Limburg 2005Low
Methods Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, the random allocation sequences by sex were generated by the US data-coordinating centre before the baseline evaluation by computer. Allocation concealment: adequate, the masked code linking agent bottles to intervention group assignments was retained by data-coordinating centre staff and concealed from all but the study statisticians until completion of the study analyses. Blinding: adequate, identical placebo capsules. Follow-up: adequate. Vital status was ascertained at the trial end in all patients in the vitamin E group and in 99.9% in the placebo group. Intention-to-treat analysis: yes. Sample size calculations: yes.
Participants Country: China. Number of participants randomised: 360, mean age 47, 42% males. Inclusion criteria: at least 1 grossly visible oesophageal lesion with biopsy-proven mild or moderate squamous dysplasia, according to histological interpretation by a single pathologist. Exclusion criteria: history of cancer (except nonmelanoma skin cancer), symptoms suggestive of an upper gastrointestinal tract malignancy, recently treated peptic ulcer disease, or contraindications to the intervention agent(s) or study-related procedures. Subjects were also excluded if a grossly visible lesion could not be confirmed during the baseline EGD or if the worst histological diagnosis at baseline was less than mild or greater than moderate dysplasia.

Outcomes
The primary outcome measure was: change in histological grade of squamous dysplasia (determined by comparing the most advanced histological diagnosis for each subject at the baseline and end-of-trial evaluations) and was categorised as regression, stable, or progression.

Notes
Compliance was assessed by pill counts and by direct observation by the village doctors who watched all participants take 2 morning pills each day throughout the intervention period. Compliance was further assessed biochemically by comparing baseline and end-of-trial serum selenium concentrations. Compliance with the single daily dose of selenomethionine (or placebo) and 1 of the 2 daily doses of celecoxib (or placebo) was in excess of 99% by both direct observation and pill counts. Pfizer, Incorporated, provided active and placebo celecoxib agent supplies.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Mineral And Vitamin Intervention Study (MAVISl)
Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, participants were randomly allocated to intervention or placebo by batch numbers generated by a password protected computer program. Allocation concealment: adequate, the computer program used for randomisation could be accessed by only the data programmer, blind to treatment allocation. The identity of the tablets was concealed in a double envelope sent by the manufacturer, which was kept locked in a cabinet during the trial. Blinding: adequate, identical placebo tablets. Follow-up: adequate, only 13% (n = 121) of the participants were lost to follow-up or reported stopping taking tablets. At least 1 diary was provided by 99% (901) of participants, 6 diaries by 93% (846), and 12 diaries by 89% (808). Losses to follow-up was equal in the active and the placebo (n = 22) groups. Fourteen participants in the active group and 18 participants in the placebo group withdrew. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: Scotland Number of participants randomised: 910, 479 men and 431 women, aged 65 or over who did not take vitamins or minerals.

MAVIS 2005 Low (Continued)
Inclusion criteria: all people aged 65 or over who were registered with the practices were eligible, irrespective of chronic illness, unless their doctors considered them too unwell. Exclusion criteria: use of vitamin, mineral, or fish oil supplements in the previous three months (one month in the case of water soluble vitamins) or vitamin B12 injection in the past three months.

Outcomes
The primary outcome measures were: number of contacts with primary care (doctor and other primary care workers, in person or by phone) for infection, number of self reported days of infection, and health related quality of life measured by the EuroQol and SF-12. The secondary outcome measures were: number of antibiotic prescriptions in primary care, number of days that antibiotics were prescribed, number of hospital admissions (including those related to infection), number of days in hospital with infection, number of infection related and all outpatient visits, adverse events reported by participants, and compliance with trial drugs (from diaries submitted monthly in all participants and tablet count at six and 12 months in a random sample of 10% of participants).

Notes
Compliance with treatment was assessed by self-report and was consistent with tablet counting. There were no differences between the groups for compliance with drug taking. Compliance in participants still taking tablets and returning information in diaries was over 91% throughout the trial.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

McKeown-Eyssen 1988
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported.

Interventions
Participants were randomly assigned to receive: group 1. vitamin C 400 mg; vitamin E 400 mg (n = 96). group 2: lactose placebos (n = 89); over a period of two years. Second colonoscopic examination was performed approximately two years after the initial examination, but could be performed earlier if judged clinically necessary. The physician assessed the presence and location of polyps, and any observed were removed.

Outcomes
The primary outcome measure was: recurrence of colorectal polyps.

Notes
Compliance with treatment was assessed by random urine samples collected at each visit from which urinary vitamin C levels were assessed, using a dipstick, as an index of compliance. The compliance to the vitamin supplements appears to be good. Of the 185 eligible participants, 137 (75%) completed the study with second colonoscopic examination conducted when most participants (81,5% of those on vitamins and 82,3% of those on placebos) had been receiving supplements for 12 to 30 months. Trial agents were supplied by H. Newmark of Roche, New Jersey and Roche, Canada.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Meydani 2004Low
Methods Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, in blocks of 4 according to lists generated by the trial's statistician, who used a computer program. Six randomisation lists were constructed. Allocation concealment: adequate, those enrolling the participants had no access to the randomisation lists. Participants were unknown to the statistician. Blinding: adequate, identical looking placebo capsules. Capsules were in 2 equal batches, soft gel and identical in colour and taste. Follow-up: adequate. Of the 617 randomised persons, 37% in the vitamin E and 36% in the placebo groups, respectively, completed the 1-year trial period. Forty-one participants in the vitamin group and 42 participants in the placebo group were lost to follow-up. The losses to follow-up were equal in both groups.
Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States. Number of participants randomised: 617, 169 men and 448 women, mean age 84 years. Inclusion criteria: aged 65 years or older; life expectancy greater than 6 months; no anticipated discharge within 3 months; not room-bound for the past 3 months; absence of active neoplastic disease; no tube feeding, no kidney dialysis; no intravenous or urethral catheters for the last 30 days; no tracheostomy or chronic ventilator; antibiotic-free for more than 2 weeks; no long-term steroid treatment greater than 10 mg/d, no use of immunosuppressive drugs, or greater than the recommended daily allowance (RDA) level of supplements of vitamins E, C, or B6, selenium, zinc, beta-carotene, or fish oil; body mass index of at least 18; serum albumin at least 3.0 g/dL; able to swallow pills; willing to receive influenza vaccine; and willing to provide informed consent (for patients with dementia, family members provided informed consent).

Interventions
Participants were randomly assigned to receive: group 1: 200 IU of vitamin E (dl-alpha-tocopherol) (n = 311); group 2: placebo 4 IU of vitamin E (n = 306); both in soybean oil, one capsule daily for a period of one year. All participants received a capsule containing half the recommended daily allowance of essential vitamins and minerals. All participants received influenza vaccine.

Outcomes
Primary outcomes of the trial were: incidence of, number of persons with, and number of days with respiratory tract infections (upper and lower), and number of new antibiotic prescriptions for respiratory tract infection. Secondary outcomes included emergency department visits, hospitalisation, and death. A post hoc subgroup analysis was performed to determine the effect of vitamin E on common colds.

Notes
Adherence to trial protocol was verified by nursing home medication records, returned pill count, and quarterly measurement of plasma vitamin E levels. Ninety-eight percent of those completing the trial consumed the capsules for at least 330 days (> 90% of the 1-year supplementation period). The number of missed supplements did not differ statistically between the vitamin E and placebo groups. Capsules were manufactured by Tishcon Corporation (Westbury, NY). The capsules were packed by Pharmasource Healthcare Inc (Marlboro, Mass).

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Mezey 2004Low
Methods Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, on a computer in blocks of 4. Allocation concealment: adequate. The vitamin E capsules and placebo capsules were prepared and labelled by the hospital pharmacies. For each patient entered into the trial the investigators opened a consecutive sealed container which had a 3-month supply of capsules. Blinding: adequate, capsules were identical in looks, smell, and taste. Follow-up: adequate. During the initial 3-month period of therapy, one patient in the treatment group withdrew from the trial. Four patients, 2 in each group, died during the initial 3 months. Five patients in the treatment group and 4 patients in the placebo group were lost between 3 and 12 months to follow up. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States and Spain. Number of participants randomised: 51, 34 men and 17 women, mean age 48 years, Inclusion criteria: age (18 years to 70 years), recent history of heavy alcohol ingestion and clinical and laboratory characteristics adopted by the International Informatics Hepatology Group for the diagnosis of alcoholic hepatitis. These criteria included moderate elevation of the serum aspartate aminotransferase AST (< 10 times above normal), an AST/alanine aminotransferase (ALT) ratio greater than 1.0 and no evidence of liver disease due to viral hepatitis, autoimmune disease, haemochromatosis, Wilsons disease or drug-induced hepatitis. Exclusion criteria: pregnancy, breast feeding, cardiovascular, pulmonary, kidney disease, pancreatitis, type I diabetes, recent (within 1 month) gastrointestinal bleeding, peptic ulcer disease, concurrent infection, history of thrombophlebitis, HIV positivity and history of ingestion of more than 100 IU vitamin E for the prior month.

Interventions
Participants were randomly assigned to receive: group 1: vitamin E (dl-alpha-tocopheryl acetate) 1000 IU ( n = 25) group 2: placebo (n = 26); one capsule daily 3 months. The patients were followed for 1 year after entry into the trial.

Outcomes
The primary outcome measures was: clinical and laboratory parameters of liver function and on markers of fibrogenesis.

Notes
Compliance with treatment was checked by serum assessments. Plasma alpha-tocopherol levels increased in patients on vitamin E. The authors published results of shorter (3 months) and longer (1-year) follow-up period.

Outcomes
The primary outcome measures were: delayed-type hypersensitivity skin response, humoral response to influenza vaccine, and infectious morbidity and mortality.

Notes
Compliance with treatment was assessed first by the nursing teams that administered the pills every morning and then at the end of each six months by counting the remaining capsules in the pillboxes, and by random serum assessments. High compliance (> 85%) was observed. The supplements and placebo were provided by Produits Roche SA, Fontenay-aux-Roses, France. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Mooney 2005Low
Methods Randomised, double-blind placebo-controlled parallel group trial Generation of the allocation sequence: adequate, sample IDs were generated sequentially for each patient at each visit, without reference to treatment group by computer. Inclusion criteria: men and women ages > 18 years who smoked at least 10 CPD, did not take vitamin supplements or use a nicotine patch in the 3 months before enrolment, had no prior history of cancer or liver disease, lived at a permanent address, owned a home telephone, were willing to comply with the 2year protocol, and completed a 1-month placebo run-in. Exclusion criteria: none stated.

Interventions
Participants were randomly assigned to receive: group 1: vitamin C 500 mg and vitamin E 400 IU (n = 142); group 2: placebo (n = 142); for a period of 1.25 years.

Outcomes
The primary outcome measure was: level of benzo(a)pyrene [B(a)P]-DNA adducts as an intermediate cancer risk marker.

Notes
Treatment compliance was assessed by serum vitamin measurements and pill counts. Compliance with treatment did not differ by gender measured by blood levels of {alpha}-tocopherol at 15 months of followup or by pill counts. At all time points after randomisation, in all participants, the treatment group had significantly higher levels of vitamin E than the placebo group. However, in women, the blood levels of vitamin E did not plateau until the 9-month time point. Trial agents were provided by Hoffman-LaRoche.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Murphy 1992Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported.

Interventions
Participants were randomly assigned to receive: group 1: vitamin A 60,000 µg retinol equivalent (200,000 IU) (n = 53); group 2: placebo (vitamin A 300 retinol equivalents (1000 IU) as retinyl palmitate in arachis oil (n = 56). All capsules contained 40 IU of vitamin E as an antioxidant. Patients receiving multivitamin preparations at the onset of the trial continued to receive them. No patient was commenced on vitamin A-containing supplements during the follow-up period. A content of a single capsule was given to participants by research assistant. Participants were followedup for 90 days.

Outcomes
The primary outcome measure was: incidence of antibiotic treated bacterial infections among elderly nursing-home residents.

Notes
Participants were given a content of a single capsule by a research assistant. Due to the administration of vitamin A in the control group, the 'no intervention' in that group can be discussed. Trial capsules were provided by Roche, Basel, Switzerland.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

NIT1 1993
Methods Nutrition Intervention Trial (NIT); The General Population Trial, in Linxian, China. Randomised, placebo-controlled trial with one-half replicate of a two-by-two-by-two-by-two factorial design. Generation of the allocation sequence: adequate, random numbers generated by independent data management centre in USA. The randomisation sequence was known only at data management centre until the intervention concluded.

Interventions
Participants were randomly assigned to receive one of eight vitamin/mineral supplement combinations in the form of individual oral tablets. The eight intervention groups (each with 3677 to 3709 participants) were derived from a one-half replicate of a two-by-two-by-two-by-two factorial design which allowed to asses four factors (ie, nutrient combinations) in a single experiment. The four factors designated by the letters A, B, C, D were: A -retinol (as palmitate) 5000 IU, zinc (as zinc oxide) 22.5 mg; B -riboflavin (vitamin B2) 3.2 mg and niacin (vitamin B3) 40 mg; C -ascorbic acid 120 mg and molybdenum (as molybdenum yeast complex) 30 µg; D -beta carotene 15 mg, selenium (as selenium yeast) 50 µg, and alpha-tocopherol 30 mg. Doses of each nutrient varied from one to two times US Recommended Daily Allowances (RDAs). The eight intervention groups were defined by the following combinations of supplements; AB, AC, AD, BC, BD, CD, ABCD, or placebo and packed in coded bottles containing a one-month supply. Bottles were distributed monthly beginning in March 1986 and continuing through May 1991, average 5.25 years.

Outcomes
The primary outcome measures were: cancer incidence, cancer mortality, and overall mortality.

Notes
Compliance with study treatment was assessed by monthly pill counts and biochemical measures. Compliance was excellent throughout the study. The overall pill disappearance rate was 93% for all participants, with no difference by treatment group (range 92% to 93%) and little change during the trial. All vitamin/mineral supplements and placebos were provided by Hoffmann-La Roche, Basel, Switzerland and Lederle Laboratories, Inc. Data were extracted from the primary publication. Additional information received through personal communication with the authors. The doses were typically two to three times the US Recommended Daily Allowances (RDAs), but ranged from 0.26 to seven times the RDA depending on the vitamin or mineral. Each participant was given three pills daily, including one capsule beta-carotene or placebos and two tablets of vitamin/mineral supplement, or placebos.

Outcomes
The primary outcome measures were: cancer incidence, cancer mortality, and overall mortality.

Notes
Compliance with treatment was assessed by counting unused pills for all trial participants and by assessing nutrient levels in blood collected from samples of individuals randomly selected without replacement every three months throughout the trial. Compliance with treatment was excellent. The overall pill disappearance rate was 94% in both groups with slight decline (from 96% in year 1 to 92% in year 6 in both groups) over the duration of the trial. Data were extracted from the primary publication.

NIT2 1993Low (Continued)
Active medications and placebos were provided: beta-carotene as Solatane by Hoffmann-La Roche, Inc., Nutley, N.Y., and vitamin/mineral supplement as Centrum Lederle Laboratories, Inc., Pearl River, N.Y. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

NPCT 1996Low
Methods Nutritional Prevention of Cancer Trial (NPCT). Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, computer generated random numbers. Allocation concealment: adequate, treatment group assignment was made centrally. The co-ordinating centre held all treatment information in blinded form. Medications were distributed using sealed pill bottles. Follow-up: adequate. At the end of the blinded period of treatment no participants were lost to vital follow-up, and only 7 participants (3 in the selenium group and 4 in the placebo group) declined to provide additional information about the illness. Blinding: adequate, identical placebo tablets. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America. Number of participants randomised: 1312; 75% males, aged 18 to 80 years, mean age 63 years. Inclusion criteria: history of two or more basal cell skin cancers (BCC) or one squamous cell skin cancer (SCC), with one of this occurring within the year prior the randomisation, life expectancy of at least five years and no internal malignancies treated within the previous five years. Exclusion criteria: history of significant liver or kidney disorders. Recruitment began on September 15, 1983 and continued each year through 1991.

Interventions
Patients were randomly assigned to receive: group 1: 200 µg of selenium supplied in a 0.5 g high-selenium bakers yeast tablet (n = 653); group 2: placebo (n = 659); The end of a blinded period of treatment was on February 1, 1996. Mean length of treatment was 4.5 years and follow-up 7.4 years.

Outcomes
The primary outcome measures were: incidences of basal cell and squamous cell carcinoma of the skin.
In 1990 secondary outcome measures were identified, which included: total mortality and cancer mortality, as well as the incidence of the lung, colorectal, and prostate cancers.

NPCT 1996Low (Continued)
Notes Compliance with treatment: excellent, 79.3% of the participants (80.3% in the placebo group and 78.4% in the selenium group) missed taking a pill less than twice a month. Trial medications were provided by Nutrition 21 (La Jolla, CA), through 1995 and by Cypress Systems (Fresno, CA) thereafter.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

NSCPT 1999Low
Methods Nambour Skin Cancer Prevention Trial (NSCPT). Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, randomisation computer program. Allocation concealment: adequate, the treatment code was known only to the investigator who generated it and the two people who packaged the tablets for distribution. None of these people had contact with participants. Blinding: adequate, identical looking placebo tablets of beta-carotene. Follow-up: adequate. At the end of the trial after 5 years, 15% participants had withdrawn without a complete skin examination by a dermatologist in the follow-up period. Fifty participants, assigned to sunscreen and beta carotene, 70 assigneed to sunscreen and placebo, 59 assigned to no sunscreen and beta carotene, and 59 assigned no sunscreen and placebo were lost to follow-up. Intention-to-treat analysis: yes. Sample size calculations: yes.
Participants Country: Australia. Number of participants randomised: 1621, 708 men and 913 women, aged between 20 and 69 years, mean age 48.8 years. Inclusion criteria: examination by a dermatologist with removal of all diagnosed skin cancers, written informed consent. Exclusion criteria: taking vitamin supplements containing beta-carotene and those who reported that they were already applying sunscreen on a strict daily basis.
Use of a placebo skin cream is considered unethical from two points of view: an oil-in-water emulsion with no active chemicals may enhance ultraviolet damage after evaporation of the water component; and people in the trial may use the placebo skin cream rather than a protective sunscreen in situations that

Penn 1991 (Continued)
Exclusion criteria: patients who were catheterised, or who had pressure sores, or who were receiving medication known to affect immune function.

Outcomes
The primary outcome measures were: nutritional status and cell-mediated immune function.

Notes
Compliance with treatment is not described.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

PHS 1996Low
Methods Physicians Health Study (PHS). Randomised, double-blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, by computer in blocks. Allocation concealment: adequate, the shipping department sent out calendar packs (which were identical whether active or placebo) to individual participants depending on this code. All of the calendar packs were in coded boxes, supplied by the drug manufacturer, so that the shippers did not know which drug they were shipping. Blinding: adequate, identical placebo capsules. Follow-up: adequate. By December 31, 1995, the scheduled end of the trial, less than 1% of the participants were lost to follow up. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America. Number of participants randomised: 22071 US male physicians at age 40 to 84 years, mean age 53 years. Inclusion criteria: US male physicians willing to take part in this trial. Exclusion criteria: chronic liver disease or evidence of abnormal liver function, severe renal disease or evidence of impaired renal function, inflammatory muscle disease or evidence of muscle problems (creatine kinase > 750 IU/L); concurrent treatment with cyclosporin, fibrates, or high-dose niacin; child-bearing potential; severe heart failure; some life-threatening condition other than vascular disease or diabetes (eg, severe chronic airways disease or any cancer other than non-melanoma skin cancer); or conditions that might limit long-term compliance (eg, severely disabling stroke, dementia, or psychiatric disorder).

PHS 1996Low (Continued)
Interventions Physicians were randomly assigned to one of the four groups including: group 1: active aspirin 325 mg on alternate days plus beta-carotene placebo; group 2: active beta-carotene 50 mg on alternate days plus aspirin placebo; group 3: both active agents; or group 4: both placebos. The randomised aspirin component of the study was terminated early, on 25 January 1988. The betacarotene component continued uninterrupted until its scheduled end in December 1995. A total of 11036 physicians were assigned at random to receive beta-carotene and 11035 to receive betacarotene placebo. Time from randomisation to the end of study averaged 12 years, and time of follow-up 12.9 years.

Outcomes
The primary outcome measures were: overall and within subgroups, incidence of malignant neoplasms (except non melanoma skin cancer), incidence of cardiovascular disease, and overall mortality.

Notes
Compliance with treatment was checked by random serum assessments obtained at unannounced visits to trial participants. Compliance with treatment excellent, the average per cent of pills taken was 97% in both the active and placebo groups. There was 85% compliance with beta-carotene treatment after five years and 78% after 12 years. The use of vitamin A supplements was reported by only 6% of the placebo group even by the end of trial. Active trial packs and matching placebos were provided by: aspirin (

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Pike 1995Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, blocks of computer generated random numbers.

Outcomes
The primary outcome measure was: immune indices in healthy elderly.

Notes
Compliance was verified by interview with the patient during trimonthly visits to the center and through morbidity forms, phone calls, and checking supplement containers when brought back to the center. Trial agents were provided by Hermes Arzneimittel GmbH. Inclusion criteria: old age (> 65 years); hypertension (systolic blood pressure > 160 mmHg or diastolic blood pressure > 95 mm Hg on at least three separate occasions); hypercholesterolaemia (total blood cholesterol > 6.4 mmol/L on at least two separate occasions); diabetes mellitus (fasting venous plasma glucose concentration > 7.8 mmol/L on at least two separate occasions (chronic drug treatment for any of the three latter conditions was also a criterion for inclusion); obesity (body mass index > 30 kg/m2); and family history of myocardial infarction before 55 years of age in at least one parent or sibling. Exclusion criteria: treatment with antiplatelet drugs (history of vascular events or diseases); chronic use of anti-inflammatory agents or anticoagulants; contraindications to aspirin; diseases with predictable poor short-term prognosis; and predictable psychological or logistical difficulties affecting compliance with the trial requirements.

Outcomes
The primary outcome measure was: the cumulative rate of cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke. Predefined analyses included cardiovascular deaths, total deaths, total cardiovascular events (cardiovascular death, nonfatal myocardial infarction, non-fatal stroke, angina pectoris, transient ischaemic attacks, peripheral artery disease, and revascularisation procedures).

Notes
At the beginning, and repeatedly during the trial, all patients received advice on compliance with background treatments. Compliance with treatments: at year 1 and at the end of the study 19.2% and 19.3% of the patients randomised to aspirin and 13.1% and 13.6% of those randomised to vitamin E had stopped taking the treatment. Side effects were the reason for discontinuation for 7.9% of the patients in the aspirin group and 1.1% in the vitamin E group. At the end of the trial, 7.2% of the patients not randomised to aspirin were taking aspirin and 0.2% of those not randomised to vitamin E were taking vitamin E.

PPP 2001 (Continued)
Bayer supplied the aspirin preparation, and vitamin E capsules were provided by Bracco SpA. Inclusion criteria: at least one adenoma diagnosed within the previous three months, patients have undergone colonoscopy with the entire large bowel seen and judged to be free of further polyps, good health, age less than 80 years. Exclusion criteria: familial polyposis, a history of invasive colorectal cancer, malabsorption syndromes, or any condition (such as a history of renal calculi or thrombophlebitis) that might be worsened by dietary supplementation with vitamin C or E. Participants agreed not to take supplemental vitamin C or E or beta carotene outside the trial. The trial protocol called for two follow-up colonoscopic examinations, the first approximately one year after the colonoscopy that qualified the patient for study (year 1), and second 36 months after the first (year 4). A colonoscopy was considered to be satisfactory for study purposes if cecum was reached, the entire mucosa was seen, and all polyps were removed. The endoscopist recorded the size and location of all raised mucosal lesions.

PPS 1994Low (Continued)
The study agents were provided in the form of soft gelatine capsules (containing placebo, beta carotene alone, vitamin E alone, or beta carotene plus vitamin E) and tablets (containing placebo or vitamin C) packaged in calendar packs, with each day's blister containing one capsule and one pill.

Outcomes
The primary trial outcome was: the occurrence of new adenomas between the colonoscopic examinations conducted at year 1 and year 4.

Notes
Compliance was checked by random serum assessments. Compliance with treatment was good, 82% of all patients reported taking the study agents at least six days per week, and further 5% took them three to five days per week. Only five patients stopped taking the medications because of their presumed toxicity. Trial agents were provided at no cost by BASF of of Wiandotte, Michigan. Additional information received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Prince 2003Low
Methods Randomised, double-blind, placebo-controlled cross-over trial. Generation of the allocation sequence: adequate, computer-generated. Allocation concealment: adequate, therapy was randomly allocated using computer-generated tables by an independent pharmacist (central randomisation). Blinding: adequate, identical looking capsules. Follow-up: adequate, 44 (72%) patients completed the trial per protocol. One patient died from previously undiagnosed ischaemic heart disease during the first (active) treatment period. Eight further patients (5 from the active group) withdrew from the trial during the first treatment period and 8 (4 from the active group) withdrew during the second treatment period. Intention-to-treat analysis: no. Sample size calculations: yes.

Participants
Country: United Kingdom. Number of participants randomised: 61 patient with primary biliary cirrhosis, 92% women, mean age 58 years. Inclusion criteria: primary biliary cirrhosis and self-reported fatigue. Exclusion criteria: change in disease (or symptom) altering medication in the 3 months prior to randomisation (e.g. ursodeoxycholic acid, colestyramine, rifampicin), current use or use within the last 3 months of nutritional supplements containing antioxidants, inability to complete symptom severity assessment documents, life-threatening intercurrent disease; presence of other uncontrolled disease with fatigue forming part of its clinical spectrum (eg, hypothyroidism, anaemia, renal failure, depression); drug dependency or addiction; women of child-bearing potential who were not practising effective contraception.

Rayman 2006Low
Methods Prevention of Cancer by Intervention with Selenium Pilot Study (PRECISEp). Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, computer-generated permuted blocks. Allocation concealment: adequate, research nurses telephoned the independent randomisation service to obtain an anonymous code for each volunteer, and then gave the volunteer their corresponding pre-coded tablets.
Blinding: adequate, identical intervention and placebo yeast. Follow-up: adequate, 34 participants (7%) withdrew from treatment within the first 6 months. There was no significant difference in treatment withdrawals between groups (7, 10, 5, and 12 in the placebo and 100, 200, and 300 µg groups respectively. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United Kingdom. Number of participants randomised: 501, 53% men at age 60 to 74 years, mean age 67 years. Inclusion criteria: volunteers from four general practices. Exclusion criteria: incapable of carrying out light housework or office work, active liver or kidney disease, prior diagnosis of cancer (excluding nonmelanoma skin cancer), diagnosed HIV infection, immunosuppressive therapy, diminished mental capacity, taking > 50 µg/day of selenium supplements in the previous six months (by patient report).

Outcomes
The primary outcome measures were: mood, quality of life, and plasma selenium level.

Notes
Compliance with randomised treatment was determined by pill count, with participants considered compliant if they took at least 80% of their allocated tablets. Reasons for participant withdrawal were noted. Four hundred fifty three of the 467 participants (97%) who completed six months were compliant according to pill count. Trial agents were provided by Pharma Nord, Vejle, Denmark.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

REACT 2002Low
Methods The Roche European American Cataract Trial (REACT). Randomised, double blind, placebo-controlled, trial with parallel groups design. Generation of the allocation sequence: adequate, by Efron's biased coin method. Allocation concealment: adequate, the generator of the assignment (the individual who generated, using a bias-free method, the listing that identified the intervention assignment for every participant) was located in Munich, Germany. The executor of the assignment (the individual who, having determined a participant's eligibility, consulted the assignment system for that participant's intervention designation) was also located in Munich, Germany. The persons who prepared the randomisation scheme were not involved in determining eligibility, administering intervention, or assessing outcomes. Blinding: adequate, identical placebo capsules contained corn oil as the major constituent provided by a sponsor. Follow-up: adequate, the pattern of drop-outs was similar in the groups, and drop-outs created no imbalances between the placebo and treatment groups. There were no differences noted between the vitamin and placebo groups regardless of the length of follow-up. Overall 59 participants in the vitamin group and 68 participants in the placebo group were lost to follow-up. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States and United Kingdom Number of patients randomised: 297, mean age 68 years, 59 % females. Inclusion criteria: at least one eye met the following ocular criteria: cataract extraction unlikely within two years, immature idiopathic 'senile' cataract present in one or both eyes, (U.S. patients) presence of minimal cataract by Lens Opacities Classification System (U.K. patients) presence of cataract of minimal Oxford grade: cortical and posterior subcapsular grades: grade I; nuclear brunescence: grade II; and white nuclear scatter: grade II. If both eyes met the inclusion criteria, and cataracts were of different types, the cataract type in the eye with the worse visual acuity determined the group for randomisation. If an eye had more than one type of cataract, the morphological type that in the clinicians opinion was the more destructive to visual acuity determined the group for randomisation, no visually significant fundus pathology, no clinical signs of glaucoma and intraocular pressure, no history of amblyopia, eye surgery, argon or YAG laser eye treatment, or major eye trauma, no history of iritis, retinal crystalline deposits, or optic nerve disease, no extended (daily for >3 months) use of ocular corticosteroid or glaucoma therapy, no participation in another clinical trial investigating an anticataract formulation within the last year. Exclusion criteria: pregnancy, insulin dependent diabetes mellitus, severe renal failure or kidney stones, fat malabsorption syndrome, history of major intestinal surgery, chronic diarrhoea, alcoholism, extended use (daily for > 3 months) of systemic corticosteroid treatment, use of anticoagulants, or regular use of any vitamin supplement.

Interventions
Patients were randomly assigned to receive: group 1: 600 mg vitamin E (all-rac alpha-tocopherol acetate), vitamin C 750 mg, and beta-carotene 18 mg (n = 149); group 2: placebo (n = 148); The actual supplementation period ranged from 2 to 51 months, 231 patients were followed for at least two years, 158 patients for at least three years and 36 patients for at least four years. Patients remained in the study for 34 months.

Outcomes
The primary outcome was: the measure of area, 'increase % pixels opaque' cataract severity documented with serial digital retroillumination imagery of the lens; progression was quantified by image analysis assessing increased area of opacity.

Notes
Compliance with treatment was checked by serum assessments. The plasma concentrations of vitamin C, vitamin E, and beta-carotene in the treated and placebo groups were maintained at consistent levels throughout the trial indicating excellent compliance with instructions about the use of the trial medication. There appeared to be little if any supplementation of placebo with other vitamins or failure to take the vitamin capsules. The work was supported by grants from F. Hoffmann-La Roche, Ltd., Basel, Switzerland, and Roche Vitamins, Inc., Parsippany, NJ.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Sasazuki 2003
Methods Randomised, double-blind placebo-controlled trial with three-by-three, then two-by-two factorial design and then parallel group design. Generation of the allocation sequence: unclear, not reported. Allocation concealment: adequate, the capsules were allocated to each participant in a central control centre. Blinding: adequate, identical looking capsules. Follow-up: inadequate, losses to follow-up were high due to modification of the protocol. Out of 439 participants randomised, 134 has dropped out before the trial was altered. Of the 397 remaining participants, 305 (77%) consented to take part in a modified trial and 244 completed the trial. Overall, 98 participants, assigned to receive 500 mg vitamin C and 97 participants, assigned to receive 50 mg vitamin C were lost to follow-up. Intention-to-treat analysis: yes. Sample size calculations: yes.
Participants Country: Japan. Number of participants randomised: 439, 35% men and 65% women, aged 40 to 69 years, mean age 57 years. Inclusion criteria: men and women living in four municipalities (3 towns and one village of Yokote Public Health Centre District in Akita prefecture, participated in annual screening programmes for circulatory diseases with chronic atrophic gastritis (determined by serum pepsinogen (PG) levels (PG I < 70 ng/ml and PG I/PG II ration < 3.0). Exclusion criteria: past history of gastric cancer or surgery, liver cancer or cirrhosis, and other cancers within 5 years; abnormal liver function (AST > 100 IU/L, ALT > 100 IU/L, or ALP > 800 IU/L), use of supplements containing beta-carotene or vitamin C, unable to follow-up for at least one year.

Sasazuki 2003 (Continued)
Interventions Participants were randomly assigned to receive: group 1: vitamin C 50 mg and beta-carotene placebo; group 2: vitamin C 500 mg and beta-carotene placebo; group 3: vitamin C 50 mg and beta-carotene 15 mg; group 4: vitamin C 500 mg and beta-carotene 15 mg. 217 participants (low-dose group) were assigned to receive 50 mg of vitamin C and 0/15 mg of betacarotene; 222 participants were assigned to receive 500 mg of vitamin C and 0/15 mg of beta-carotene. daily for 5 years.
Out of 439 persons initially participating in the study, 134 participants dropped before and on modification of the study protocol based on a National Cancer Institute report that indicated that 2 beta-carotene trials had shown no benefit or potential harm from the supplement. Of the 305 remaining participants, 244 completed this study. Participants were supplemented with beta-carotene from September 1995 to March 1996, (three to six months). After that study was continued with parallel group design. Participants were randomly assigned to receive: group 1: vitamin C 50 mg (n = 144); group 2: vitamin C 500 mg (n = 161); for five years.

Outcomes
The primary outcome measure was: the 10-year cumulative incidence of gastric cancer. The secondary outcome measure was: 5-year change in serum levels of pepsinogens. After the modification of the protocol the primary outcome measure was 5-year change in serum levels of pepsinogens and other biomarkers.

Notes
Compliance with treatment was constantly encouraged and monitored by nurses, who interviewed the participants and recorded pill counts every 3 months (compliance rate, 80%). Compliance with treatment was checked by serum assessments. Blood samples were drawn and stored three times (at baseline, and after the first, and the fifth year) in order to measure serum level of ascorbic acid. Compliance in taking the vitamin capsules was 92.9% in men and 95.4% in women. Additional information about all-cause mortality obtained through personal communication with authors. These data, which are extremely positive, are not published.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

SCPS 1990Low
Methods Skin Cancer Prevention Study (SCPS). Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, on a computer in block of 16 without stratification. Allocation concealment: adequate, central independent unit. Blinding: adequate, identical looking capsules. Follow-up: adequate. Of the 1805 patients who have been randomised, 89% completed at least three years of observation, 79% four years, and 46% five years. The patterns of follow-up were similar in the two treatment groups. Intention-to-treat analysis: no. Sample size calculations: no.

Participants
Country: United States of America. Number of participants randomised: 1805, 1251 (70%) men and 554 (30%) women. Inclusion criteria: < 85 years old, at least one biopsy-proved basal-cell or squamous-cell carcinoma, could not become pregnant, agreement not to take vitamin supplements containing vitamin A or betacarotene, not a vegan vegetarian (one who eats no animal products, including milk or eggs). Exclusion criteria: xeroderma pigmentosum, basal-cell nevus syndrome, an active nonskin cancer, known exposure to arsenic, or any other major medical problem that would limit their ability to participate in the planned five years of study.

Interventions
Patients were randomly assigned to receive: group 1: beta-carotene 50 mg (n = 913); group 2: placebo (n = 892); one capsule daily for five years. Duration of follow-up was five years.

Outcomes
The primary outcome measures were: the first occurrence of basal-cell or squamous-cell skin cancer.

Notes
Compliance with the study medication was determined by interviewing the patients and by measurement of plasma beta-carotene levels. At four months interval patients were asked to complete questionnaires concerning compliance in taking the capsules. Reported adherence to treatment did not differ appreciably between the placebo and beta-carotene groups, and during each of the first four years at least 80% of the patients reported taking half or more of their capsules. Plasma beta-carotene levels showed more than eight-fold increase in the group that received beta-carotene and almost no-change in the placebo group. The trial agents were provided by BASF, Wyendotte, Michigen. Though a study with a longer follow-up on this same trial was published in JAMA (Mortality associated with low plasma concentration of beta carotene and the effect of oral supplementation. JAMA 1996;275(9):699-703.), we could not use the mortality data from the latter because it does not report on the 85 excluded patients. However, replacing the number of deaths from Greenberg 1990 with the mortality data given in the JAMA publication (146 vs 139 (placebo) does not change the result noticeably.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

SIT 2001
Methods Shandong Intervention Trial Randomised, double-blind, placebo controlled, primary prevention trial with stratified, factorial design 2x2x2 versus 2x2. Generation of the allocation sequence: unclear, not reported. Allocation concealment: adequate, treatments were randomly assigned at Westat, Inc. and this assignment is used to distribute coded bottles of capsules from the pharmacy in the city of Weifang in Shandong Province. Blinding: adequate, using identical placebo capsules. Follow-up: adequate, overall 15 participants from placebo and 19 participants from active intervention group were lost to follow-up. Intention-to-treat analysis: no. Sample size calculations: yes.

Participants
Country: China (Linqu County, Shandong Province). Number of participants randomised: 3411, 1753 men and 1658 women aged 35 to 64 years. Inclusion criteria: participants aged 35 to 64 years willing to participate in 42-month study, baseline gastroscopy with biopsies, known Helicobacter pylori status. Exclusion criteria: illness, bleeding disorders, cancers (except nonmelanoma skin cancer), heart failure, emphysema, renal or liver diseases, other life-threatening illnesses, allergy to penicillin or related antibiotics.

Interventions
Participants were first divided on the basis of whether they showed serologic evidence of Helicobacter pylori infection at baseline (2285) or not (1126). Participants with serologic evidence of Helicobacter Pylori at baseline were eligible to receive amoxicillin (1 g twice a day) and omeprazole (20 mg twice a day) in three capsules (two 500 mg amoxicillin and one 20 mg omeprazole) to be taken twice daily (before breakfast and dinner) for 2 weeks. Look-alike placebo capsules containing lactose and starch for amoxicillin and sucrose and starch for omeprazole were given to serologically positive controls and to all seronegative participants. Approximately 3 months after initial treatment for Helicobacter Pylori, supplementation with 100 IU alpha-tocopherol, 250 mg vitamin C, and 37.5 µg selenium twice a day began its 39-month course. Participants receive this mixture in one capsule, to be taken twice daily before or after breakfast and dinner. From December 1995 to May 1996, this mixture also contained beta-carotene (7.5 mg twice a day). Look-alike placebo capsules contained cellulose, lactose, and magnesium stearate. In the garlic group, participants take two capsules twice a day before or after breakfast and dinner. Each capsule contains 200 mg Kyolic aged garlic extract and 1 mg steam-distilled garlic oil. To prepare the extract, the manufacturer slices garlic cloves and soaks them in aqueous ethanol (about 20%) for over 18 months at room temperature. The extract is then filtered, concentrated, and dried. The look-alike placebo capsules contain cellulose, granulated sugar, caramel, and magnesium stearate. Bottles holding placebo capsules contained minute quantities of garlic oil so they would smell like garlic. HP-seropositive at baseline (2258) entered 2x2x2 factorial of antibiotics, vitamins, and garlic. HP-seronegative at baseline (1126) entered 2x2 factorial trial of vitamins, and garlic. Participants were randomised in 12 groups: group 1: amoxicillin and omeprazole, garlic, vitamin and selenium (n=286); group 2: amoxicillin and omeprazole, garlic, vitamin and selenium placebo (n=285); group 3: amoxicillin and omeprazole, garlic placebo, vitamin and selenium (n=286); group 4: amoxicillin and omeprazole, garlic placebo, vitamin and selenium placebo (n=285); group 5: amoxicillin and omeprazole placebo, garlic, vitamin and selenium (n=285); group 6: amoxicillin and omeprazole placebo, garlic, vitamin and selenium placebo (n=286); group 7: amoxicillin and omeprazole placebo, garlic placebo, vitamin and selenium (n=286);

Outcomes
The primary outcome measures were: prevalence of dysplasia or gastric cancer, prevalence of severe chronic atrophic gastritis, intestinal metaplasia, dysplasia or gastric cancer, and average severity score. Secondary outcome measures were: rates of transition from baseline to final histopathologic states and the effects of treatments on these rates of transition; evidence of the effectiveness of amoxicillin and omeprazole in eradicating Helicobacter pylori, based on 13C-urea breath tests 3 months following treatment, on annual serology, and on a final pathologic examination of biopsies to look for Helicobacter pylori; and blood pressure at the time of the final examination.

Notes
Compliance with treatment was checked by measuring the plasma vitamin levels in randomly selected participants every 3 months and counting of the pills. Compliance with treatment was good. The average monthly proportion of participants taking all pills was 92.3%. Serum samples obtained from randomly selected participants demonstrate higher levels of vitamins C and E in participants assigned to vitamins 2 and higher levels of S-allylcysteine in those assigned to garlic preparation.
(Wakunaga of America, Co., Ltd, Mission Viejo, CA) provided the garlic preparation, Astra (East Asia Region) provided axomicillin and omeprazole; and Sino-American Shanghai-Squibb Pharmaceuticals, Ltd. provided vitamin and mineral supplement.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

SKICAP AK 1997Low
Methods Skin Cancer Prevention Study -actinic keratoses (SKICAP-AK). Randomised, double-blind placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, on a computer in permuted blocks of size 4. Allocation concealment: adequate, numbered coded bottles. Blinding: adequate, identical looking capsules. Follow-up: adequate. Overall, 99 participants from the intervention group and 88 from the placebo group were lost to follow-up. Intention-to-treat analysis: no. Sample size calculations: yes.

SKICAP AK 1997Low (Continued)
Participants Country: United States. Number of participants randomised: 2297, 679 (30%) women and 1618 (70%) men, aged 21 to 84 years, median age 63 years, with a history of more than 10 actinic keratoses and at most 2 squamous cell carcinoma (SCC) or basal cell carcinoma (BCC) skin cancers. Inclusion criteria: free living participants aged 21 to 84 years, ambulatory and capable of self care, with no diagnosis of life threatening diseases, an intended continual resident of Arizona for at least five years, willing to return during the five years for semi-annual follow-up clinic visits, and willing to limit nonstudy vitamin A supplementation to no more than 10,000 IU per day, clinical laboratory values within the 95% normal range for total cholesterol, liver function (AST and ALT), WBC count, haemoglobin and platelet count, and history of more than 10 actinic keratoses and at most 2 squamous cell carcinoma (SCC) or basal cell carcinoma (BCC) skin cancers. Exclusion criteria: cancer diagnosis or treatment within the year preceding the trial other than BCC or SCC, history of xeroderma pigmentosum or basal-cell nevus syndrome.

Interventions
Patients were randomly assigned to receive: group 1: vitamin A (retinol) 25,000 IU (n = 1157); group 2: placebo (n = 1140); one capsule daily for a period of 5 years (median follow-up time of 3.8 years.

Outcomes
The primary outcome measures were: the time to first new occurrence of SCC and time to first new occurrence of BCC pathologically confirmed by the study pathologist.

Notes
Compliance with the study medication was determined by counting capsules in the returned medication bottles and by measurement of plasma vitamin A levels. Participants were scheduled for a return clinic visit one month after randomisation and then every six months. They were interviewed to evaluate adherence, motivated to adhere and provided with a six-month supply of capsules. Participants were telephoned and mailed postcards between clinic visits for symptom assessment and adherence monitoring and motivation. Vitamin intake was reported by 73% of the participants, and 30% of participants reported dietary intake near or below the recommended. Calculated adherence to the intervention was almost identical between the placebo and retinol groups. During the 5-year intervention period, at least 85% of participants reported taking at least three-quarters of their capsules, and at least 95% reported taking at least half of their capsules. The results show very similar baseline retinyl palmitate levels and approximately an 8-fold increase in the median serum retynil palmitate level in the group assigned to receive retinol. Hoffmann-LaRoche provided intervention capsules.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

SPACE 2000Low
Methods Secondary prevention with antioxidants of cardiovascular disease in end stage renal disease (SPACE). Randomised, double blind, placebo-controlled, secondary prevention intervention trial with parallel group design. Generation of the allocation sequence: adequate, computer generated coin toss within each stratum. Exclusion criteria: anticoagulant therapy with warfarin sodium; known history of malignant disease (except non-melanoma skin cancer); active liver disease; treatment with hypolipaemic agents for less than eight weeks before the study started; pregnant or planning to become pregnant during duration of the study; any condition the treating physician deemed to preclude the patient on grounds of safety or study evaluation.

Interventions
Patients were randomly assigned to receive: group 1: vitamin E 800 IU/day (n = 97); group 2: matching placebo (n = 99); Vitamin E was provided as two capsules of 400 IU each. Patients were instructed to take two capsules nightly. Median follow-up time was 519 (range 10 to 763) days.

Outcomes
The primary outcome measure was: a composite variable consisting of: acute myocardial infarction (fatal and nonfatal); ischaemic stroke; peripheral vascular disease (excluding the arterio-venous fistula) in a limb not previously affected; and unstable angina. Secondary outcome measures were: fatal and non-fatal myocardial infarction, cardiovascular disease mortality (fatal myocardial infarction, ischaemic stroke or sudden death), total mortality, ischaemic stroke, peripheral vascular disease, and unstable angina.

Notes
Compliance with treatment was evaluated by measuring serum vitamin E concentrations. Throughout the study, patients continued to receive regular monthly follow-up by their unit dieticians, who instructed them to comply with dietary recommendations for maintenance haemodialysis patients. Additional vitamin supplementation was similar in the two treatment conditions. Folate (5 to 10 mg/day), vitamin B6 (10 to 250 mg/day), and vitamin B12 (250 µg/day) were prescribed to 57 (57.5%) patients in the placebo group and 55 (56.7%) patients in the vitamin E group. Only one patient (in the vitamin E group) received vitamin B12 as a monthly intramuscular injection. Vitamin C (100 to 500 mg/day) was prescribed to 42 (42.5%) of the placebo group and 42 (43.3%) of the vitamin E group. Vitamin E and placebos were provided by Solgar, Inc, New York, USA, during the first year and Henkel Corp, La Grange, IL, USA, during the second year.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Stevic 2001
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported.

Outcomes
The primary outcome measures were: survival and rate of disease progression as expressed by decline in limb-function, bulbar-function and muscle-testing scores. Secondary outcome measures were: activity of antioxidative components, and level of vitamin E in blood.

Notes
Compliance with treatment is not reported.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

SUVIMAX 2004Low
Methods The SUpplementation en VItamines et Mine´raux AntioXydants (SU.VI.MAX) Study. Randomised, double-blind, placebo-controlled, primary-prevention trial with parallel group design. Generation of the allocation sequence: adequate, random treatment allocation was performed by blocksequence generation stratified by sex and age group by computer. Allocation concealment: adequate, capsule boxes were labelled with the participant's number, using partitioned organisation to ensure total security of the blind study. Blinding: adequate, identical placebo capsules. Follow-up: adequate. Losses to follow-up; 5.4 % in the intervention group and 6.2% in the placebo group.
Overall, 739 participants in the active and 828 participants in the placebo group were lost to follow-up. Intention-to-treat analysis: yes. Sample size calculations: yes.
Participants Country: France. Number of participants randomised: 13017 French adults, 5141 men and 7876 women, aged from 35 to 60 years, mean age 48.95 years. Inclusion criteria: lack of disease likely to hinder active participation or threatened 5-year survival; acceptance of possibility to be given placebo and acceptance of the constraints of participation; lack of previous regular supplementation with any of the vitamins and minerals in the supplement provided and absence of extreme beliefs or behaviour regarding diet. Exclusion criteria: none stated.

Outcomes
The primary outcome measures were: major fatal and nonfatal ishaemic cardiovascular events and cancer of any kind, except for the basal cell carcinoma of the skin. The secondary outcome measure was: all cause mortality.

Notes
Compliance for the intervention group was confirmed by measuring the biochemical markers of supplementation after 2 years and after 7 years for beta-carotene, vitamin C and selenium. At the end of followup, 74% of participants reported having taken at least two thirds of the capsules. There were no differences between the groups mean percentage of capsules taken, ie, 79% in each group

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Takagi 2003
Methods Randomised, clinical trial with parallel group design. Generation of the allocation sequence: unclear, not reported. Allocation concealment: unclear, not reported. Blinding: inadequate, no intervention in the control arm. Follow-up: adequate. Seven patients in the vitamin E group and 3 patients from the control group dropped out from the trial. Intention-to-treat analysis: no. Sample size calculations: no.
Participants Country: Japan. Number of patients randomised: 93, 45% males and 55% females, mean age 62.5 years. Inclusion criteria: liver cirrhosis caused by hepatitis C infection. Exclusion criteria: none stated.

Outcomes
The primary outcome measures were: tumor-free survival and cumulative survival rate.

Notes
Compliance was not reported.

Item Authors' judgement Description
Allocation concealment? No C -Inadequate

Takamatsu 1995
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: unclear, not reported. Allocation concealment: adequate, centrally by pharmacy. Capsules required by each participant for a 6month period were placed in identical bottles labelled with the personal numbers of the volunteers. Blinding: adequate, capsules identical in size, shape, weight and colour. Follow-up: adequate. Losses to follow-up were 4 participants (6.75%) in the active treatment group and 10 participants (13.69%) in the placebo group during the trial. Intention-to-treat analysis: no. Sample-size calculation: no.
Participants Country: Japan. Number of participants randomised: 161, 64 men and 97 women, aged 39 to 56 years. Inclusion criteria: healthy Japanese adults free of acute and chronic illness, including hypertension.

Takamatsu 1995 (Continued)
Exclusion criteria: taking oral contraceptives, vitamins or mineral supplements, pregnant or lactating women.

Outcomes
The primary outcome measure was: any illness.

Notes
Medication compliance during the trial period was 89.6% in vitamin E group and 91.3% in the placebo group. Vitamin E (d-alpha tocopheryl acetate capsules) were provided by Eisai Co. Ltd (Tokyo, Japan).

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

Tam 2005Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, by computer. Allocation concealment: adequate, by a central independent unit (the school of pharmacy

Outcomes
The primary outcome measures were: effects on markers of oxidative stress, antioxidant defence, and endothelial function.

Notes
Compliance was assessed by tablet counting, and patients with less than 70% compliance were excluded from the analyses. Overall compliance by pill count was 95%.

Item Authors' judgement Description
Allocation concealment? Unclear B -Unclear

VEAPS 2002Low
Methods The Vitamin E Atherosclerosis Prevention Study (VEAPS). Randomised, double-blind, placebo-controlled, primary-prevention trial with parallel group design. Generation of the allocation sequence: adequate, computer-generated random numbers.

Interventions
Participants were randomly assigned to receive: group 1: vitamin E (DL-alpha-tocopherol) 400 IU (n = 177); group 2: placebo (n = 176); daily for a period of three years. Participants were instructed to take trial pills with their greatest fat-containing meal of the day. The initial trial design called for a 2-year treatment period. Based on evolving null results from other antioxidant clinical trials, the External Data and Safety Monitoring Board recommended after 2 years of initiation of the study that the treatment period be extended to 3 years. Participants were offered the opportunity to continue another year of their randomised and blinded treatment assignment. The 81% of 2-year completers, 73% of randomised elected to continue for a 3-year treatment period.

Outcomes
The primary trial outcome was: rate of change in the right distal common carotid artery intima-media thickness in computer image-processed B-mode ultrasonograms.

VEAPS 2002Low (Continued)
Notes Compliance with treatment was assessed by counting unused pills and measuring plasma vitamin levels.
Mean pill compliance was 92% in the placebo-treated group and 91% in the vitamin E group. Pill compliance for the placebo-treated versus the active vitamin E participants was maintained throughout the trial, as follows: 89% versus 87%, 90% versus 89%, 92% versus 92%, 91% versus 93%, 94% versus 91%, and 93% versus 93% at 6,12,24,30, and 36 months, respectively. There was an appropriate rise in the mean plasma vitamin E level in the active vitamin E group from a baseline level. The trial was supported by Hoffmann-La Roche, Inc.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

VECAT 2004Low
Methods itamin E, Cataract and Age-Related Maculopathy Trial (VECAT). Randomised, double blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, the randomisation schedule was prepared by a biostatistician using a permuted block allocation scheme. Allocation concealment: adequate, central, independent unit. The allocation list was stored at a remote site. Identification of which patients received vitamin E or placebo was concealed from the outcome assessors and data analysts until the analysis was finalised. Blinding: adequate, the medications were dispensed in identical containers so that neither the trial staff nor the participants were aware of the intervention in any specific patient. Follow-up: adequate. Overall, 60 participants from the placebo group and 58 participants from the vitamin group withdrew from the trial. Intention-to-treat analysis: yes. Sample size calculations: yes.
Participants Country: Australia. Number of participants randomised: 1193, 44% men and 56% women, aged 55 to 80, mean age 65.7 years. Inclusion criteria: good general health, early or no cataract. Exclusion criteria: prior cataract surgery, advanced cataract in both eyes, glaucoma, known sensitivity to vitamin E, and long-term treatment with steroids or anticoagulants.

Outcomes
The primary outcome measures were: major age-related types of cataract: nuclear, cortical cuneiform, and posterior subcapsular.

VECAT 2004Low (Continued)
Notes Compliance with the trial medication was determined by counting capsules in the returned medication bottles and by measurement of plasma vitamin E levels in a random sample of participants. Overall, 77% of the actively treated group and 79% of those participants randomised to placebo were estimated to have consumed 80% or more of their capsules. After 4 years of follow-up, 74% of the vitamin E group and 76% of the placebo group remained on their assigned medication and participated in the annual reviews. Among the remaining 25% of the participants, 12% in each group ceased taking the assigned medication but continued participating to have their eye examined. The trial was funded by Smith and Nphew, Australia, Henkel, Australia.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

WAVE 2002Low
Methods Women's Angiographic Vitamin and Estrogen Trial (WAVE). Randomised, double blind, placebo-controlled trial with two-by-two factorial design. Generation of the allocation sequence: adequate, centrally by computer. Randomisation lists were produced for each stratum using a permuted block design with block sizes of two and four being randomly selected. Allocation concealment: adequate, the clinical centres nurse co-ordinators randomised women by calling the dedicated randomisation computer at the study co-ordinating centre. Blinding: adequate, identical placebo tablets. Follow-up: adequate. Twenty-three participants from the placebo group and 29 from the HRT group, 38 from the vitamins, and 27 from the HRT and vitamin group withdrew from the trial. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America and Canada. Number of participants randomised: 423 women mean age 65 years. Inclusion criteria: postmenopausal women as defined by any one of the following criteria: (bilateral oophorectomy at any age or age 45 to 55 with FSH 40 mIU/ml or older than 55 years. Protocol angiogram within four months performed while haemodynamically stable demonstrating at least one vessel segment free of intervention, with 15 to 75% stenosis. If the angiogram was performed within two weeks of a myocardial infarction, the qualifying segment may not be the infarct segment. Exclusion criteria: oestrogen replacement therapy within the past three months. Estrogen vaginal cream permitted if used no more than 25% of the time. Concurrent use of vitamins C and E exceeding the recommended dietary allowance, history of breast cancer or mammogram suggestive of cancer without subsequent negative biopsy, history of endometrial carcinoma without subsequent hysterectomy, any abnormal uterine bleeding or endometrial hyperplasia at baseline, pap smear with dysplasia of cervical intraepithelial neoplasia grade I or greater, uncontrolled diabetes or hypertension, myocardial infarction less than four weeks prior to randomisation, planned or prior coronary artery bypass grafting, fasting triglycerides 500 mg/dl within four months of randomisation, creatinine 2.0 mg/dl, symptomatic gallstones, New York Heart association class IV congestive heart failure or known ejection fraction 25%, history of haemorrhagic stroke or bleeding diathesis, history of pulmonary embolism or idiopathic deep venous thrombosis, history of osteoporosis unless treated with nonhormonal therapy, anticipated survival three years, concurrent participation in other masked clinical trial, participation in an interventional device trial or short-term postangioplasty antithrombotic trial was permitted so long as follow-up angiography was not a requirement of that trial.

Interventions
The participants were randomly assigned to receive: group 1: vitamins (vitamin E 400 IU and vitamin C 500 mg) and hormone replacement therapy (HRT) placebo (n = 105); group 2: HRT (women with a prior hysterectomy took one tablet containing conjugated equine estrogens (0.625 mg of Premarin, while the women who had not had a hysterectomy took one tablet containing conjugated equine estrogens and medroxyprogesterone acetate (0.625 mg/2.5 mg of Prempro) and vitamins placebo daily (n = 103); group 3: vitamins C and E and HRT (n = 107); group 4: vitamin placebo and HRT placebo (n = 108); twice daily for a median of three years.

Outcomes
The primary outcome measure was: annualised mean change in minimum lumen diameter from baseline to concluding angiogram of all qualifying coronary lesions averaged for each patient. Patients with intercurrent death or myocardial infarction were imputed the worst rank of angiographic outcome.

Notes
Compliance with treatment was checked by serum assessments. Among the women with angiographic follow-up, those assigned to HRT took 67% of their prescribed medication according to pill counts, and those assigned to HRT placebo took 70%. The corresponding figures were both 84% for antioxidant vitamins and vitamin placebo. Nine women assigned to placebo estrogen crossed over to open-label estrogen, and one woman assigned to placebo vitamin supplements crossed over to open-label vitamins. Hormone replacement therapy drugs supplied by Wyeth Pharmaceuticals, Collegeville, Pa.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

WHS 2005Low
Methods Women's Health Study (WHS). Randomised, double-blind, placebo-controlled trial with two-by-twoby-two factorial design in the beginning and than two-by-two. Generation of the allocation sequence: adequate, centrally by computer in batches of blocks of size 16. Allocation concealment: adequate, the randomisation allocation is coded. Shipping department sends out calendar packs (which are identical whether active or placebo) to individual participants depending on this code. All of the calendar packs are in coded boxes, supplied by the drug manufacturer, so that the shippers do not know which drug they are shipping. Blinding: adequate, identical placebo capsules. Follow-up: adequate. Losses to followup; 0.01% in beta-carotene group and 0.005% in placebo group. Overall, 132 participants in the active group and 102 participants in the placebo group had unknown vital status. Intention-to-treat analysis: yes. Sample size calculations: yes.

Participants
Country: United States of America. Number of participants randomised: 39876 females aged 45 years or older, mean age 54.6 years. Inclusion criteria: female health professionals willing to take part in the trial. Age 45 years or older; no previous history of coronary heart disease, cerebrovascular disease, cancer (except nonmelanoma skin cancer), or other major chronic illnesses; no history of adverse effects from aspirin; no use of aspirin or nonsteroidal anti-inflammatory drugs (NSAIDs) more than once a week, or willingness to forgo their use; no use of anticoagulants or corticosteroids; and no use of individual supplements of vitamin A, E, or beta carotene for more than once a week. Exclusion criteria: history of cancer (except non-melanoma skin cancer), coronary heart disease, or cerebrovascular disease.

Interventions
Participants were randomly assigned to one of the eight treatment groups. The active agents were 100 mg of aspirin, given on alternate days; 600 IU of vitamin E, given on alternate days; and 50 mg of betacarotene, given on alternate days. group 1: aspirin 100 mg, beta carotene 50 mg, vitamin E 600 IU; group 2: aspirin 100 mg, beta carotene 50 mg, vitamin E placebo; group 3: aspirin 100 mg, beta carotene 50 mg placebo, vitamin E 600 IU; group 4: aspirin 100 mg, beta carotene placebo, vitamin E placebo; group 5: aspirin placebo, beta carotene 50 mg, vitamin E 600 IU; group 6: aspirin placebo, beta carotene 50 mg, vitamin E placebo; group 7: aspirin placebo, beta carotene placebo, vitamin E 600 IU; group 8: aspirin placebo, beta carotene placebo, vitamin E placebo; A total of 19939 women were assigned at random to receive beta-carotene and 19937 to receive placebo in the beginning of April 1993. A total of 19937 women were assigned at random to receive vitamin E and 19939 to receive placebo. The beta-carotene component of the trial was terminated early, on January 18

Outcomes
The primary outcome measures were: incidence of invasive cancer (except non-melanoma skin cancer), myocardial infarction, and stroke. The secondary outcome measures were: non-fatal myocardial infarction, non-fatal stroke, death from cardiovascular causes, and death from any cause.

Notes
Compliance with treatment was checked by random serum assessments. Compliance with treatment was excellent. At the time of termination of the beta-carotene component, 87% of the active group have taken at least two thirds of the study capsules, while 9.9% of the women in the placebo group have taken betacarotene or vitamin A supplements outside the trial. The active agents were provided as follows: aspirin by

WHS 2005Low (Continued)
Bayer AG, Leverkusen, Germany; vitamin E by Natural Source Vitamin E Association, Washington DC; and beta-carotene by Lurotin, BASF Corporation, Wiandotte, MI. Data were extracted from the primary publication, but additional information was received through personal communication with the authors.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Witte 2005Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, random numbers generated by computer. Allocation concealment: adequate, the randomisation allocation co-ordinated by a remote pharmacy with which the investigators had no contact during the study (ie, central randomisation Inclusion criteria: stable chronic heart failure due to ischaemic heart disease. Exclusion criteria: neurological or inflammatory conditions or other significant chronic morbidity affecting quality of life (eg, severe rheumatoid arthritis) or requiring long-term systemic steroid or non-steroidal anti-inflammatory drugs therapy (except low-dose aspirin). Patients in persistent atrial fibrillation were also excluded in order to optimise the reproducibility of the estimation of left ventricular function.

Outcomes
The primary outcome measures were: left ventricular function, levels of pro-inflammatory cytokines, and quality-of-life in elderly patients with chronic heart failure.

Notes
Authors did not measure blood levels of all the micronutrients to assess compliance, although the changes in the ferritin, vitamin B12, and folate levels in the patients, randomised to the micronutrient combination suggest that they took them.

Item Authors' judgement Description
Allocation concealment? Yes A -Adequate

Wluka 2002Low
Methods Randomised, double-blind, placebo-controlled trial with parallel group design. Generation of the allocation sequence: adequate, computer generating block randomisation program. University osteoarthritis index (WOMAC) pain score above 20%. Pain that was at least mild in severity (no compromise of daily activities, frequent but tolerable pain that is worsened by unusual activity and patient may take a pain reliever occasionally). Exclusion criteria: known sensitivity to vitamin E, current anticoagulation therapy, previous stroke or history of poorly controlled hypertension, major morbidities such as a cancer or life threatening illnesses, inability to co-operate with study requirements and give informed consent, dementia, other forms of arthritis, inability to walk 50 feet without the use of assistive devices, hemiparesis of either lower limb, those awaiting knee replacement, grade IV knee osteoarthritis, and any contraindication to magnetic resonance imaging (MRI) (eg, pacemaker, cerebral aneurism clip, cochlear implant, presence of shrapnel/metal in strategic locations such as in the orbit, and claustrophobia).

Outcomes
The primary outcome measure was: change in cartilage volume in patients with knee osteoarthritis.

Notes
Compliance was assessed by returned pill counts at each visit. Average compliance, assessed on the basis of residual capsule counts, was similar in the two groups; 95.7% in the vitamin E group and 97.0% in the placebo group. No side effects were attributed to vitamin E. (Continued)

Barringer 2003
Randomised, double-blind, placebo-controlled trial to assess the effect of a multivitamin and mineral supplement on infection and quality of life. Twenty-eight persons did not complete the trial. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors.

Basnayake 1983
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Bassenge 1998
Randomised clinical trial. The authors did not report any deaths during the follow-up period.
Bates 1998 This is not a randomised clinical trial. A cross-sectional analysis of survey data to characterize relationships among blood pressure, pulse rate, vitamin C status, and other protective and risk factors for older British people.

Beaton 2002
Randomised clinical trial. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors.

Beckman 2003
Randomised clinical trial. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors.

Benton 1991
Randomised clinical trial. The possibility that a subclinical deficiency of the trace element selenium might exist in a sample of the British population was examined by giving a selenium supplement for 5 weeks.
Using a double-blind cross-over design, 50 participants received either a placebo or 100 mcg selenium on a daily basis. On three occasions they filled in the Profile of Moods States. A food frequency questionnaire was used to estimate the intake of selenium in the diet. Authors did not report any deaths during the trial.
Berger 1998 This randomised, placebo-controlled trial studied clinical and immune effects of trace element supplements. Twenty patients, aged 40 +/-16 y (mean +/-SD), burned on 48 +/-17% of their body surfaces, were studied for 30 d after injury. They consumed standard trace element intakes plus supplements (40.4 micromol Cu, 2.9 micromol Se, and 406 micromol Zn; group TE) or standard trace element intakes plus placebo (20 micromol Cu, 0.4 micromol Se, and 100 micromol Zn; group C) for 8 days. Demographic data were similar for both groups. This trial did not fulfil our inclusion criteria.

Bernard 2003
Randomised clinical trial. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors.
Berson 1993 Randomised, clinical, double-masked trial with 2 x 2 factorial design and duration of 4 to 6 years to determine whether supplements of vitamin A or vitamin E alone or in combination affect the course of retinitis pigmentosa. Six-hundred-and-one patients aged 18 through 49 years with retinitis pigmentosa met preset eligibility criteria. Ninety-five per cent of the patients completed the trial (29/601). Four of 29 patients died and 25 decline to continue participation, most after the fourth year. We were not able to extract relevant data about the mortality in each arm from the published article. Authors were not able to provide these data too. (Continued)

Crimi 2004
Randomised, double-blind placebo-controlled trial to examine the effect of antioxidant supplementation in enteral feeding in critically ill patients. Baseline characteristics of patients were reason for exclusion. Eighteen patients in antioxidant group and 16 patients in placebo group had malignancy. This trial did not meet our inclusion criteria.

Crogan 2005
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Dabiri 1994
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Daga 2003
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Dakhale 2005
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Darko 2002
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Davison 2005
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Dawson 1999
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

De las Heras 2000
This is not a randomised clinical trial. The purpose of this report is to analyse the results of a 1-year clinical study of antioxidant therapy in the treatment of pain and recurrent inflammatory episodes in patients with chronic and acute recurrent pancreatitis, using a prospective, descriptive, pre-post, open design. The studied patients were with acute recurrent or chronic pancreatitis who had suffered from pain or acute inflammatory episodes the year before the beginning of treatment with a complex containing L-methionine, beta-carotene, vitamin C, vitamin E and organic selenium.
de Sanjose 1996 Randomised clinical trial. The authors did not report any deaths during the follow-up period.
de Vet 1991 Randomised clinical trial. The authors did not report any deaths during the follow-up period.
de Waart 1997 Randomised clinical trial. The authors did not report any deaths during the follow-up period.
DeCosse 1989 Over a 4-year period in a chemoprevention randomised, double-blind, placebo-controlled trial on large bowel neoplasia, patients with familial adenomatous polyposis were treated with 4 g of ascorbic acid (vitamin C)/day plus 400 mg of alpha-tocopherol (vitamin E)/day alone or with a grain fiber supplement (22.5 g/day). Of the 62 randomly assigned patients, 58 were assessable. Four patients withdrew from the study. Authors did not report any deaths during the trial.
DeMaio 1992 Randomised, double-blind placebo-controlled trial to test whether alpha-tocopherol prevents restenosis following percutaneous transluminal coronary angioplasty (PTCA). Patients were randomised after successful PTCA to receive vitamin E in the form of dl-alpha-tocopherol, 1200 IU/day, orally versus an inactive placebo for 4 months. Fifteen participants withdrew from the trial before its end. Authors did not report any deaths during the trial. This trial did not meet our inclusion criteria.
(Continued) and the number of swollen joints; (2) three measures of pain--pain in the morning, pain in the evening, and pain after chosen activity; (3) haematological and biochemical measures of inflammatory activity; (4) assays for the oxidative modification of proteins and lipids. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors.

Egan 1992
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Eiselt 2001
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

El-Bayoumy 2002
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Elkashef 1990
Randomised clinical trial. The authors did not report any deaths during the follow-up period.
Ernster 1985 Randomised, double-blind placebo-controlled trial of the effect of vitamin E on clinically palpable benign breast findings. Women were supplemented mean two months. Eleven participants did not complete the trial. Authors did not report any deaths during the trial.

Everett 2002
Randomised clinical trial. The authors did not report any deaths during the follow-up period.
Fairley 1996 A randomised double-blind placebo controlled trial was designed for 117 women with abnormal cervical morphology. Thirty milligrams of oral beta carotene were administered daily for 12 months. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Fairris 1989
Since reduced concentrations of selenium in whole blood, plasma and white cells had previously been observed in psoriasis, 69 patients were supplemented daily with either 600 micrograms of selenium-enriched yeast, 600 micrograms of selenium-enriched yeast plus 600 IU of vitamin E, or a placebo for 12 weeks. Before supplementation, the patients' mean concentrations of selenium in whole blood and plasma were reduced compared with those of matched healthy controls, but their red cell glutathione peroxidase (GSH-Px) activity was normal. During the study, 4 patients were excluded. Authors did not report any deaths during the trial.

Falsini 2003
Non-randomised, comparative clinical study to evaluate the influence of short-term antioxidant supplementation on retinal function in age-related maculopathy (ARM) patients by recording focal electroretinograms.

Fang 2002
Randomised, double-blind, placebo-controlled trial to examine the effect of vitamins C and E on progression of transplant-associated arteriosclerosis. This trial did not meet our inclusion criteria.

Farvid 2005
In a randomised, double-blind, placebo-controlled clinical trial, 69 type 2 diabetic patients were randomly divided into four groups; each group received one of the following daily supplement for 3 months: group M (n = 16), 200 mg Mg and 30 mg Zn; group V (n = 18), 200 mg vitamin C and 100 IU vitamin E; group MV (n = 17), minerals plus vitamins; and group P (n = 18), placebo. The aim of the trial was to assess the effect of magnesium plus zinc, vitamins C plus E, and a combination of these micronutrients on nephropathy (Continued) Mann 1987 A randomised, prospective, placebo-controlled trial of daily multivitamin supplementation in 101 noninstitutionalized ambulatory elderly persons (median age, 64 years). Vitamin levels were assayed at baseline, and at two and four months of supplementation. The authors did not report any deaths during the followup period.

Manzella 2001
A double-blind randomised controlled trial; 50 patients with type 2 diabetes were assigned to treatment with vitamin E (600 mg/d) or placebo for 4 months. The possible effects of vitamin E on the cardiac autonomic nervous system, as assessed by analysis of heart rate variability, in patients with type 2 diabetes and cardiac autonomic neuropathy were investigated.

Margaritis 2003
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Marotta 2003
The aim of this randomised clinical trial was to test the effect of antioxidants on enzymatic abnormalities and free radicals-modified DNA adducts associated with pre-malignant changes in HP-negative chronic atrophic gastritis patients. Sixty patients with and intestinal metaplasia underwent a GI endoscopy with biopsy samples for histology and for: alpha-tocopherol, malonyldialdehyde, xanthine oxidase, ornithine decarboxylase and 8-hydroxydeoxyguanosine. Patients were randomly allocated into three groups supplemented for 6 months with: vitamin E, 300 mg/day; multivitamin, 2 tablets/day, and a certified fermented papaya preparation 6 g/nocte (Immune-Age FPP, Osato Research Institute, Gifu, Japan). Ten dyspeptic patients without histological abnormalities served as control. Histological and biochemical parameters were blindly repeated at 3 and 6 months. The authors did not report any deaths during the follow-up period.

Martinez-Abun 2001
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Massey 2005
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Mastaloudis 2004
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

Mathews-Roth 1972
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

McAuliffe 1998
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

McDowell 1994
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

McGavin 2001
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

McKay 2000
Randomised, double-blind, placebo-controlled trial to determine whether a daily multivitamin/mineral supplement can improve micronutrient status, plasma antioxidant capacity and cytokine production in healthy, free-living older adults already consuming a fortified diet. The authors did not report any deaths during the trial. (Continued) Nieman 1997 Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Nieman 2002
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Nimmagadda 1998
This is an open-label study to assess the effect of short-term beta-carotene administration (180 mg/d with meals for 4 weeks) on the plasma human immunodeficiency virus (HIV) RNA levels and CD4+ lymphocyte counts in 21 HIV-infected patients. The trial did not meet our inclusion criteria.

Nyyssonen 1994
Randomised clinical trial. There were no deaths during the follow-up period. Additional information was obtained through personal communication with authors.

O'Byrne 2000
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Olmedilla 2002
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Olmedilla 2003
A randomised, double-blind, placebo-controlled trial investigated the effect of long-term antioxidant supplementation (lutein and alpha-tocopherol) on serum levels and visual performance in patients with cataracts. The authors did not report any deaths during the follow-up period.

Ono 1985
Randomised clinical trial. The authors did not report any deaths during the follow-up period. Orndahl 1994 A randomised, double-blind, placebo-controlled trial to study the effect of a combined selenium and vitamin E treatment in patients with myotonic dystrophy. Twenty-seven patients with myotonic dystrophy divided into an experimental (n = 13) and a control (n = 14) group. The experimental group was given increasing doses for 4 months up to a maximum of 1.6 mg selenium and 800 mg vitamin E daily and the control group a corresponding number of placebo tablets. The total treatment period was 2 years. Muscle strength (knee extension, knee flexion, hand-grip), maximal walking speed for 30 m, function in daily activities (disability), well-being and cognitive functioning. The authors did not report any deaths during the trial.
Osilesi 1991 Randomised clinical trial. The authors did not report any deaths during the follow-up period. (Continued)

Tahir 2005
Randomised clinical trial to determine the impact of the combination of vitamins C and E or vitamin C only on serum levels of cell adhesion molecules and C-reactive protein in patients with chronic degenerative AS, with or without concomitant coronary artery disease. One hundred patients with asymptomatic or mildly symptomatic moderate AS were randomised in 2:2:1 format in an open-label trial. Forty-one patients received vitamin E (400 IU) and vitamin C (1000 mg) daily, 39 patients received vitamin C (1000 mg) only, and 20 patients were followed as controls. The authors did not report any deaths during the trial.
Tardif 1997 Randomised clinical trial. One month before angioplasty, 317 patients were randomly assigned to receive one of four treatments: placebo, probucol (500 mg), multivitamins (30,000 IU of beta carotene, 500 mg of vitamin C, and 700 IU of vitamin E), or both probucol and multivitamins-all given twice daily. Patients were treated for four weeks before and six months after angioplasty. The patients received an extra 1000 mg of probucol, 2000 IU of vitamin E, both probucol and vitamin E, or placebo 12 hours before angioplasty, according to their treatment assignments. Base-line and follow-up angiograms were interpreted by blinded investigators using a quantitative approach. This trial did not meet our inclusion criteria.

Tarng 2004
Randomised clinical trial. The authors did not report any deaths during the follow-up period.
Tarp 1985 Randomised clinical trial assessing a possible antirheumatic effect of selenium. Forty patients with active RA were included in a 6-month double-blind clinical study of selenium versus placebo. The patients in the selenium group were given daily supplements of 256 micrograms selenium in selenium-enriched yeast. Although concentrations of selenium in serum and erythrocytes increased considerably, no significant antirheumatic effect of selenium could be demonstrated. All patients completed the trial.

Tauler 2002
Randomised clinical trial. The authors did not report any deaths during the follow-up period.
Tessier 1995 Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Thomson 1988
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Title 2000
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Tofler 2000
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Tousoulis 2003
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Trebble 2003
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

Trebble 2005
Randomised clinical trial. There were no deaths during the follow-up period. Additional information obtained through personal communication with authors. (Continued) Zaridze 1993 Randomised clinical trial. This intervention trial carried out in Uzbekistan (former USSR) in an area with a high incidence of oral and esophageal cancer involved random allocation of 532 men, 50 to 69 years old, with oral leukoplakia and/or chronic esophagitis to one of four arms in a double-blind, two-by-two factorial design, with active arms defined by the administration of (a) riboflavin; (b) a combination of retinol, beta-carotene, and vitamin E; or (c) both. Weekly doses were 100,000 IU of retinol, 80 mg of vitamin E, and 80 mg of riboflavin. The dose of beta-carotene was 40 mg/d. Men in the trial were followed for 20 months after randomisation. The aim of the trial was to determine whether treatment with these vitamins or their combination could affect the prevalence of oral leukoplakia and/or protect against progression of oral leukoplakia and esophagitis, conditions considered to be precursors of cancer of the mouth and esophagus. Among 519 men, 32 gave up taking supplements. The authors did not report any deaths during the trial.

Zhu 2002
Randomised, double-blind, placebo-controlled clinical trial with an aim to evaluate the roles of folic acid and beta-carotene in the chemoprevention of gastric and other gastrointestinal (GI) cancers. A total of 216 patients with atrophic gastritis were randomly assigned to one of the four groups: (1)

Zimmermann 1997
In this study the effect of antioxidant therapy with sodium selenite was investigated in patients with systemic inflammatory response syndrome and multiple organ failure. Forty patients were included and observed over a period of 28 days. This study did not meet our inclusion criteria.

Zollinger 1999
Randomised clinical trial. The authors did not report any deaths during the follow-up period.

APPOSE 2001
Trial name or title Australian Prostate Cancer Prevention Trial Using Selenium (APPOSE). Randomised double-blind, placebo-controlled trial with parallel group design.

APPOSE 2001 (Continued)
Participants Country: Australia. Number of participants randomised: 6000 men who are at increased risk because of a first-degree relative with prostate cancer.

Interventions
Participants will be supplemented with selenium.

Outcomes
The primary outcome measure is incidence of prostate cancer.

SELECT 2003
Trial name or title The selenium and vitamin E cancer prevention trial, SELECT). Randomised double-blind, placebo-controlled trial with two-by-two factorial design.

Participants
Country: United States of America. Number of participants randomised: 32,400 males, aged 50 years or older. Inclusion criteria: age > 55 years for Caucasians and > 50 years for African-Americans, digital rectal examination not suspicious for prostate cancer, total serum prostate specific antigen < 4.0 ng/ml, no prior history of prostate cancer or high-grade prostatic intraepithelial neoplasia, no anticoagulation therapy, except low-dose aspirin, normal blood pressure (systolic blood pressure < 150 mm Hg and diastolic blood pressure < 90 mm Hg), willing to restrict supplementation of selenium and vitamin E during participation.

Interventions
Participants were randomly assigned to receive either 200 µg of 1-selenomethionine, 400 mg of racaemic alpha-tocopherol, and an optional multivitamin containing no selenium or vitamin E. The racaemic mix of alpha-tocopherol will include both the d-and l-isomers. Participants will be divided in four group according to two-by-two factorial design: group 1: placebo (n = 8100); group 2: vitamin E (n = 8100); group 3: selenium (n = 8100); group 4: vitamin E and selenium (n = 8100).

Outcomes
The primary outcome measure is the clinical incidence of prostate cancer as determined by a clinical diagnostic work-up, including yearly digital rectal examination and serum prostate specific antigen level. Secondary outcome measures will include prostate cancer-free survival, all cause mortality, and the incidence and mortality of other cancers and diseases potentially impacted by the chronic use of selenium and vitamin E. Other trial objectives will include periodic quality of life assessments, assessment of serum micronutrient levels and prostate cancer risk, and studies of the evaluation of biological and genetic markers with the risk of prostate cancer.

Reply
We have read the letter from Hickey et al with interest. Overall, we disagree that our statistical analyses are inappropriate. As described in our review, our analyses were planned in advance. We do not understand why our assumptions are 'biologically unsound'. We employ methodology of reviewing systematically all randomised trials we could identify. This is the soundest scientific method incorporating biological knowledge regarding interpretation of beneficial and harmful effects of interventions. We used the Cochrane well-established procedure for selecting trials for inclusion in our review and no one can accuse us of performing a biased selection of trials. The selection of statistical analyses was not post hoc, but ac-cording to our protocol and to the standard methodology within the Cochrane Hepato-Biliary Group.

When results like our present become published, many people may react based on different backgrounds. Hickey and co-authors accuse us for not having responded to a previous critical objection published by them in JOM, ie, Journal of Orthomolecular Medicine. First, we do not have access to this journal. Second, none of the authors have sent us their objections. Third, Hickey and co-authors should know that the usual academic procedure is to submit such objections to JAMA, which in 2007 published an abbreviated version of our review. Had they done so, we would of course have responded to any of their objections. 2. We have assessed the evidence regarding the primary and secondary preventive effects of antioxidant supplements with traditional meta-analytic methodology of randomised clinical trials. Randomised clinical trials -and especially those having low risk of bias -are to be found at the top of the evidence hierarchy. We did not employ Baysian meta-analyses. That Hickey and co-authors might have had prior probabilities being more positive towards antioxidant supplements than a neutral prior could be due to them placing too much confidence in results of observational studies and in basic research findings from in vivo and in vitro studies. The literature is loaded with examples where evidence from randomised trials does not concur with evidence from lower levels of the evidence hierarchy. That antioxidant supplements could be another of these examples does not come as a surprise to us. 3. Our 2004 Cochrane Hepato-Biliary Group systematic review on antioxidant supplements including 14 randomised trials, among which only 9 trials provided data on overall mortality, led us to observe an increased mortality in participants on antioxidant supplements. At that time we were informed in an Editorial accompanying our paper in The Lancet that this conclusion could be wrong due to the fact that we had only included the randomised trials that looked on antioxidant supplement prevention of gastrointestinal cancers. Our present Cochrane Hepato-Biliary Group systematic review on antioxidant supplements is a response to the request for a broader meta-analysis, including all randomised preventive trials on antioxidant supplements that report mortality. We do, therefore, agree with Hickey and co-authors that the results of our previous review influenced our present review, now including 67 randomised trials. It is also correct that our analyses excluding selenium trials can be seen as a post-hoc decision. However, it should be seen as a post-hoc decision following our findings in our 2004 Cochrane Hepato-Biliary Group systematic review on antioxidant supplements. It is not a post-hoc decision taken following the analyses of data in our present review. We think this distinction is of central importance to the inferences drawn from our analyses. So, when we write: "The sensitivity analysis removing selenium trials from our analysis to evaluate their influence on our conclusions was therefore not a post hoc decision", we should have added perhaps "taken after the analyses of trials in the present review". Of course, we let us informed of our previous results. Isn't this the whole meaning of doing research?
tocotrienols. The following passage is taken from Sen et al. (2006): 'An expanding body of evidence support that members of the vitamin E family are functionally unique. In recognition of this fact, title claims in manuscripts should be limited to the specific form of vitamin E studied. For example, evidence for toxicity of a specific form of tocopherol in excess may not be used to conclude that high-dosage "vitamin E" supplementation may increase all-cause mortality. Such conclusion incorrectly implies that tocotrienols are toxic as well under conditions where tocotrienols were not even considered.' -We are aware of the facts mention above. However, we could not include trials with tocotrienols, gama tocopherol, or any other form of vitamin E because such randomised trials have not been published. The majority of the trials conducted tested alphatocopherol. We will in future updates try to highlight this issue. Magosso, Yuen, and Gopalan wrote: Similar arguments apply to beta-carotene, which is a single homologue of a range of about 600 carotenoids found in nature, 50 of which exert the role of vitamin A precursors and occurring as cis/trans racemic mixture at a variable ratio (Schieber and Carle, 2005;Krinsky and Johnson, 2005). Lyn (2000) in his review suggested 'the efficient uptake of synthetic all-trans beta-carotene [?] appears to make the synthetic form more desirable for effective absorption. But the tendency of synthetic beta-carotene to alter normal serum trans/cis ratios in favor of the trans-isomer may not be a beneficial effect' and that 'the consequences of using all-trans synthetic betacarotenes are at this point unknown'. It has been suggested the negative outcome in clinical trials involving beta-carotene might be ascribed to the use of the purified, synthetic form (Ben-Amotz and Levy, 1996;Lyn, 2000).
In the review entitled 'The use of antioxidant therapies during chemotherapy', Drisko and co-authors (2003) highlighted the importance of natural mixed carotenoids, suggesting that 'the use of synthetic beta-carotene as a single agent rather than natural mixed carotenoids may actually promote cancer formation'.
-The answer is as above. We included trials with beta-carotene that we were able to identify. We will in future updates try to highlight the raised issues.
Magosso, Yuen, and Gopalan wrote: In all but a handful of studies considered in this review synthetic alpha-tocopherol and synthetic beta-carotene were used.
-We included the trials that we were able to identify according to our protocol. Further trials and systematic reviews have to access whether there are certain benefits or harms connected to 'synthetic' as well as 'natural' vitamins.
Magosso, Yuen, and Gopalan wrote: In addition to many examples in which different isomers of the same compound present different level of activity and toxicity, the FDA does not register new pharmaceuticals without chiral definition (FDA, 1992;FDA, 1995;FDA 1997).
-It would be much better for the FDA and other regulatory agencies to require that dietary supplements sold to the public claiming health benefits are subjected to adequate assessment of benefits and harms before market release, similar to any other drug. Magosso, Yuen, and Gopalan wrote: We believe certain aspects of the review should consider the following to reflect objectively several facts: 1) The origin of the antioxidants, either synthetic or natural, was not mentioned with regard to the included studies. Thus assuming that there are not biological differences between the two sources. -

We mentioned the form of antioxidant used in the table 'Characteristics of included studies'. The origin of the antioxidants was mentioned in the majority of the trials, but in some of them not.
Magosso, Yuen, and Gopalan wrote: 2) Tocotrienols were not present in any of the preparations administered in the studies considered, but the authors did not differentiate them from the generic name of vitamin E or alpha-tocopherol and thus shared the same detrimental effect on survival rate.
-We were not able to identify such trials. In case that at any future review updates, we identify randomised trials with tocotrienols, we will include them in our meta-analyses. As already stated, we will address this issue in further updates. Magosso, Yuen, and Gopalan wrote: 3) Since the authors' conclusion as currently stated all but prohibit further antioxidant trials it will be helpful to perform a subgroup analysis by natural or synthetic origin of the antioxidants administered and the range of antioxidants to which the conclusion is applicable needs to be stated.
-If we identify enough number of trials with natural antioxidant supplements, we can perform subgroup analyses. However, we are of the opinion that we have fulfilled the main Cochrane criterion to look at the totality of evidence for the effects of a specific intervention. If you are aware of any randomised trials with 'natural' antioxidant supplements that is not included in our analyses, we will appreciate to receive this information from you.

A large population based randomised clinical trial (Physician Health Study II 1 ) has recently been completed. Its authors concluded that neither vitamin E nor vitamin C supplementation reduced the risk of major cardiovascular events. A second trial (SELECT 2 ) stopped supplementation after the Data and Safety Monitoring Committee reviewed the available data and found that selenium and vitamin E supplements failed to prevent prostate cancer. Participants taking vitamin E had a small increase in prostate
cancer, while participants taking only selenium were more likely to develop diabetes. These results strongly support the findings of our review.