Is there really a power shortage in clinical trials testing the "homocysteine hypothesis?".

Results from a Substudy of the VITAmins TO Prevent Stroke (VITATOPS) Trial To the Editor: Since January 1, 1998 the United States and Canada have implemented policies mandating the fortification of cereal grain flour products with folic acid (0.4 to 1.4 mg/lb) to reduce the risk of neural tube birth defects in newborns.1 The disadvantage of this measure for researchers is that randomized, controlled trials of total homocysteine (tHcy)-lowering therapy, by means of folic acid and other multivitamin therapies, to prevent cardiovascular outcomes in the US and Canada are likely to achieve only 20% to 25% of the initially projected mean treatment effects of tHcy-lowering therapy (1.0 to 2.0 mol/L versus 4.0 to 6.0 mol/L).2,3 Indeed, the recently completed Vitamins In Stroke Prevention (VISP) Study, which compared high-dose multivitamins (including folic acid 2.5 mg, vitamin B12 0.4 mg, and vitamin B6 25 mg) with low-dose multivitamins (including folic acid 0.02 mg, vitamin B12 0.006 mg, vitamin B6 0.2 mg) in 3,680 stroke survivors, only achieved a mean 2.0 mol/L reduction in tHcy.4,5 The results failed to reliably exclude a modest but clinically important effect of folic acid–based multivitamin therapy in reducing the risk of major vascular events and death by up to 20% and 30%, respectively.5 The VITAmins TO Prevent Stroke (VITATOPS) Trial is the only large randomized, double-blind, placebo-controlled trial currently testing the “homocysteine hypothesis” in patients with previous transient ischemic attack (TIA) or stroke.6,7 It is being conducted in 19 countries, most where mandatory folic acid fortification of cereal grain products is not implemented. Patients are randomized to once daily treatment with placebo or the combination of folic acid 2.0 mg plus vitamin B12 0.5 mg plus vitamin B6 25 mg. The projected average treatment effect is a 4.0 mol/L reduction in tHcy.7 We examined the effect of the active VITATOPS trial medication on mean tHcy concentrations at 6 months after randomization, compared with placebo, among 285 consecutive patients with recent stroke or TIA randomized in one center in Perth, Australia. Fasting plasma tHcy was measured at baseline and at 6 months after randomisation using high-performance liquid chromatography (HPLC). The significance of any difference in plasma tHcy between multivitaminand placebo-treated patients was examined using an independent t test. For the purposes of this analysis, only a third-party statistician (Q.Y.) not involved in any aspects of the design or day-to-day running of the trial was unblinded to treatment allocation. Patients in the two treatment groups were well matched at baseline with regard to demographic variables and cardiovascular risk factors, and 250 of the original cohort of 285 patients underwent repeat blood testing at 6 months. The mean baseline tHcy level was 13.4 mol/L in patients randomized to multivitamins and 12.8 mol/L in patients randomized to placebo. This is similar to mean baseline tHcy levels of around 13.5 mol/L in the VISP study.5 At 6 months, mean tHcy levels were 9.1 mol/L in patients randomized to multivitamins compared with 12.8 mol/L in those randomized to placebo (difference 3.7 mol/L, 95% confidence interval 2.7 to 4.7 mol/L, P 0.001) (Table). Based on these data, if elevated tHcy is causally related to atherothrombotic events of the heart, brain, and limbs, the 3.7 mol/L reduction in plasma tHcy achieved with the VITATOPS tablets in our study should reduce the risk of subsequent serious vascular events by at least 15%,8,9 which is what our study was powered to detect with a type I error of 5% and type II error of 20%, assuming an average follow-up of 2 years.7 The magnitude of homocysteine-lowering observed in our Australian population (where voluntarily fortification of foods with folic acid is increasing) may not be correlated to TIA and stroke patients in many other countries participating in the VITATOPS trial (eg, Hong Kong, India, Malaysia, Moldova, Pakistan, Philippines, Singapore, Sri Lanka) where nutritional status is likely to be lower and any effect of supplementation is also likely to be lower or nonexistent. However, our estimate of the effect of multivitamins on plasma tHcy is unlikely to be exaggerated by survival bias or loss to follow-up because a similar number died or declined repeat blood testing from each treatment group and baseline characteristics of these patients were similar to those of patients who underwent repeat blood collection at 6 months (data not presented). Furthermore, the observed treatment effect is likely to underestimate the true effect size because 9.1% of patients allocated active VITATOPS treatment ceased taking their VITATOPS tablets during the 6-month follow-up period, and some who were allocated placebo took up folic acid therapy.