Applicability of allele/genotype frequency from documented controls for case-control studies on genotypes among...

In a case-control study, controls have to be selected from the population where cases are identified. However, there is an idea to make common controls applicable for different case-control studies, which could reduce study costs. This study compared methylenetetrahydrofolate reductase C677T among subjects from different studies, to examine the applicability of combined subjects as controls. Case-control studies and cross-sectional studies with more than 100 Japanese controls or subjects were selected from PubMed in December 2008. Between 1996 and 2008, 31 eligible studies with 14,510 subjects in total were published; the 677T allele frequency varied from 0.300 to 0.438. The genotype frequencies were all in Hardy-Weinberg equilibrium. The average weighted with the number of subjects was 0.385. The 95% confidence interval (95%CI) of 10 studies did not include the weighted average. The study whose proximal limit of 95%CI was furthest from the weighted average was removed, and then the weighted average was recalculated. Through the process, 7 studies were excluded, resulting in the remaining 24 studies having a 95%CI including the weighted average (0.391) with 10,854 subjects. Of the 7 excuded studies, one was from patients enrolled in a clinical study (0.431), two were relatively isolated (0.300 and 0.438), two were published ten years ago (0.315 and 0.334), one was from a study whose allele frequency of female subjects was 0.382 (0.436), and one had a large sample size (0.360). This example demonstrated that the allele frequency of MTHFR 677T was in common among 24 out of 31 Japanese studies, suggesting that the calculated allele frequency could be used for Japanese case-control studies, whose cases are sampled in a general population.