Genetic risk factors for complex diseases: race is not important

Article type
Authors
Trikalinos T, Ioannidis J
Abstract
Background: There is considerable controversy on whether race modifies the effect of genetic risk factors for common diseases. Objectives: To explore the variation in the frequencies and effect sizes within and between racial groups for specific genetic markers implicated in the susceptibility to various diseases.

Methods: We evaluated 41 meta-analyses of genetic association studies that provided adequate data for racial subgroup analyses on a total of 495 case-control studies. For each eligible meta-analysis, we estimated race-specific weighted frequencies of the genetic marker of interest in the control groups and race-specific summary odds ratios. Racial descent groups were categorized as European, East Asian, African and other. Pair-wise comparisons were performed with general variance models. Between-study heterogeneity was assessed by the Q and I2 statistics.

Results: The between-race variance for the weighted control frequencies was larger than the within-race variance in 22 meta-analyses. Conversely, the between-race variance for the race-specific summary odds ratios was larger than the within-race variance in only 6 meta-analyses.

When we compared the weighted control frequencies, statistically significant differences were seen in 21/32 (66% [95% confidence interval (CI) 47-81%]) European vs. East Asian descent comparisons, 7/18 (39% [95% CI 17- 64%]) European vs. African descent comparisons, and 10/16 (63% [95% CI 35-85%]) East Asian vs. African descent comparisons. In pair-wise comparisons between the race-specific odds ratios, statistically significant differences were seen in only 3/32 (9% [95% CI, 2-25%]) European vs. East Asian descent comparisons, 2/18 (11% [95% CI 1-17%]) European vs. African descent comparisons, and in none of the 16 (0% [95% CI 0-9%]) East Asian vs. African descent comparisons. Control frequencies showed large heterogeneity between the various racial descent groups in 24 cases (59%), while only in 3 cases (7%) for the genetic effects (odds ratios).

Conclusions: While genetic markers vary in frequency across populations of different ancestry, genetic effects are similar across racial subgroups. Claims about racial subgroup differences in genetic effects may often be spurious.

Acknowledgements: Support provided by the General Secretariat for Research and Technology, Greece and the European Commission (PENED grant).