Soy Protein Effects on Serum Lipoproteins: A Quality Assessment and Meta-Analysis of Randomized, Controlled Studies

OBJECTIVES: Many randomized controlled trials (RCTs) have assessed the effects of soy protein on serum lipoprotein risk factors for coronary heart disease (CHD). This review and meta-analysis assessed the quality of these RCTs and estimated the effects of soy protein consumption on serum lipoproteins. DATA SOURCES: A comprehensive search using multiple databases was conducted for the years 1996 through 2008 to identify clinical trials related to soy protein intake and serum lipoprotein changes. STUDY ELIGIBILITY: RCTs were assessed that met these requirements: soy protein intake compared with nonsoy protein, provided information on serum low-density lipoprotein (LDL)-cholesterol values, provided no more than 65 g of soy protein daily, and obtained LDL-cholesterol measurements between 4 and 18 weeks of treatment. Randomized parallel and crossover studies were evaluated. METHODS: Studies were graded for quality using 12 criteria with a possible maximum grade of 24. Net changes in lipoproteins with soy protein consumption compared with nonsoy control diets were analyzed by meta-analyses and funnel plots. Confidence intervals were constructed using inverse weighting. Analyses compared parallel to crossover studies and studies with lower and higher grades. RESULTS: Analyses included 20 parallel-design studies and 23 crossover studies. Parallel studies scored significantly higher (p < 0.001) in study quality, with a mean grade of 15.8 (95% confidence interval [CI], 14.3 to 17.3) compared with 10.1 (95% CI, 8.2 to 11.9) for crossover trials. Soy protein intake was associated with net changes in serum LDL-cholesterol values of -0.23 mmol/l (95% CI, -0.28 to -0.18 mmol/l) or a 5.5% reduction in parallel studies and -0.16 mmol/l (95% CI, -0.22 to -0.11 mmol/l) or a reduction of 4.2% with crossover studies (p < 0.001 for parallel vs crossover). In parallel studies, net serum HDL-cholesterol values were 3.2% higher (p < 0.007) with soy vs control, and fasting serum triacylglycerol values were 10.7% lower (p < 0.008) for soy vs control. CONCLUSIONS AND IMPLICATIONS: Soy protein consumption with a median of 30 g/d was associated with a significant improvement in lipoprotein risk factors for CHD. Compared with crossover RCTs, parallel RCTs had significantly higher quality grades and were associated with significantly greater improvements in serum LDL-cholesterol values. Regular consumption of 1 to 2 servings of soy protein daily (15 to 30 g) has a significant favorable impact on serum lipoprotein risk factors for CHD.