The most recent findings on soy and isoflavones were presented in Washington DC on the occasion of the 9th International Symposium on the Role of Soy in Health Promotion and Chronic Disease Prevention and Treatment (October 16-19, 2010). A major part of the topics was dedicated to safety and health benefits in menopausal women, and to the issue of soy and breast cancer.
Dr. James Anderson (University of Kentucky, Lexington, USA) reported on the results of a meta-analysis on soy effects on serum lipoproteins (Anderson 2010). Soy protein intake is associated with significant decreases of LDL cholesterol, which led to a recommendation of the FDA for a daily intake of 25 g of soy protein. Anderson et al. had previously made a meta-analysis on the effects of soy protein on serum lipids (Anderson et al. 1995). The efficacy of soy protein effects has been questioned, possibly based on clinical trials with suboptimal design. The EFSA recently rejected a health claim for soy protein and cardiovascular disease (see also the contribution of Dr. Harland).
Anderson updated the meta-analysis of 1995 with 42 clinical trials published since 1995 (20 parallel group trials and 22 cross-over-studies). Doses of soy protein varied between 11 and 60 g, with a median intake of 30 g. Studies with non-soy protein formulations were not included into the study.
The maximum effect was reached after 6-8 weeks, which may explain why the effect found in rather short 4-week trials was modest. In the cross-over trials the washout period could be shown to be a crucial factor for carry-over effects, which may have been the reason the observation of weaker outcomes in cross-over trials. Of 28 cross-over trials seven studies featured no wash-out period at all, seven a 2-week washout period, 4 studies a period of 3 weeks, and 9 a washout period of four weeks. The short wash-out periods are clearly insufficient for the avoidance of carry-over effects, and will therefore have impacted the overall study outcomes.
In the parallel group trials LDL was reduced by 0.23 mmol/l (95 % CI -0.28 to -0.18), corresponding to a reduction by 5.5 %. In the cross-over trials the reduction was 0.16 mmol/l (95 % CI -0.22 to -0.11), corresponding to a reduction by 4.2 % (p < 0.0001) – the less pronounced effect observed in the cross-over trials may well be related to design issues such as the already mentioned insufficient wash-out time, which would lead to lower differences between verum and placebo in the second study phase.
In parallel group studies triglycerides were improved by 9.8 % (p < 0.008), and HLD was elevated by 3.2 % (p < 0.007). Reductions of LDL of approximately 5 % translate into a 6-10 % risk reduction of cardiovascular disease. 10 % reduction in triglycerides approximately correspond to a further 5 % risk reduction, and an increase of HDL by 3 % to another 9 % reduction, which sums up to a 20-24 % total risk reduction. In addition, the improvement of systolic blood pressure by 5 mm (12 % risk reduction) and of diastolic blood pressure by 3 mm (7.5 % risk reduction) must be taken into account when the overall potential benefit of soy is addressed.
References
Anderson JW (2010). Soy protein effects on serum lipoproteins: an updated meta-analysis. 9th International Symposium on the Role of Soy in Health Promotion and Chronic Disease Prevention and Treatment, Washington DC, 16-19 October.
Clinical effects 