Sex differences and genetic associations with myocardial infarction.

Abstract

MYOCARDIAL INFARCTION (MI) AND RELATED ATHerosclerotic cardiovascular disease (CVD) are the leading causes of death in men and women. Although much is known regarding risk factors for MI, many questions remain unanswered in the prevention of MI, including what specific genetic variants underlie susceptibility to MI and why there appears to be a protective advantage against MI in women compared with men. The article by Schuit et al in this issue of JAMA helps address some of these questions but raises others. Myocardial infarction is a complex trait to which multiple environmental and genetic factors contribute. Risks for MI and other cardiovascular events are substantially increased in adults with a known history of CVD in parents or co-twins. With completion of the human genome sequence, the race has begun among a large and growing number of genetic association studies to test all plausible hypotheses regarding susceptibility conferred by myriad genes and biological pathways. Such studies often proceed in 2 basic steps. First, common (ie, minor allele frequency 1%) genetic variants—often biallelic single-nucleotide polymorphisms (SNPs)—are identified within or near the sequence of DNA encoding a candidate gene. Because neighboring SNPs are often highly correlated (due to linkage disequilibrium), they can be combined to identify combinations of SNPs forming haplotypes that more fully define the variation within genes. Second, significant differences are sought in disease prevalence or in levels of a quantitative disease trait across the genotypes defined by the SNP variant or by the haplotypes (equivalently, genotype frequencies are compared between cases and controls). By this approach, many genes have emerged as possibly implicated in MI risk. However, evidence for sex differences in associations between genetic variants and CVD has only rarely been replicated (for example, the association of angiotensinconverting enzyme inhibition with hypertension in men but not women). Of the many hypotheses to explain the advantage in women, none has received more attention than the “estrogen hypothesis.” Although a vast effort in both basic science and clinical research has documented a role of estrogen in protection against CVD, recent evidence has indicated a lack of benefit and an early increase in CVD risk from the administration of exogenous estrogen. However, it remains plausible that genetic variation in estrogen-mediated pathways contributes to interindividual and sex-specific variation in CVD risk. For example, genetic variation in estrogen receptors appears to predict changes in high-density lipoprotein (HDL) cholesterol levels in response to exogenous estrogen. In the study by Schuit et al, investigators examined the association of 2 estrogen receptor (ESR1) variants (c.454397T>C and c.454-351A>G) previously associated with bone mineral density, onset of natural or surgical menopause, and radiographic osteoarthritis. The authors tested 3 haplotypes, resulting from the 2 markers, for association to ischemic heart disease and incident MI in 6408 men and women enrolled in the Rotterdam Study. Carriage of 1 or 2 copies of the T-A haplotype of ESR1 compared with the group without the T-A haplotype was associated with a 2-fold increased risk of MI and ischemic heart disease events in women. However, no increased risk was found in men; if anything the trend was in the opposite direction. Some of these findings apparently are at odds with recent findings reported by Shearman et al. These authors tested the c.454-397 SNP in 1739 men and women from the Framingham Heart Study offspring cohort and found that male carriers of 2 copies of the C allele, compared with those with 0 or 1 copy, have a 3-fold increase in MI risk. Since only 5 women in the study had an MI, the authors could not evaluate the role of this polymorphism in women. The T-A haplotype in the study by Schuit et al is completely correlated with the T allele of c.454-397 reported in the study by Shearman et al. Thus, the studies appear to contradict each other: Schuit et al reported that having 2 copies of the C allele of c.454-397 is protective in women and has no effect in men, whereas Shearman et al found that having 2 copies of the C allele is harmful in men. The Rotterdam and Framingham study designs are remarkable more for their underlying similarities than for differences. Both well-designed studies were conducted in prospective, population-based cohorts of middle-aged to elderly