Myocardial infarction (MI), one of the most common manifestations of cardiovascular system aging, is often fatal. The vast majority of studies on genetic susceptibility to age-dependent diseases are carried out using the case-control study design. However, its use involves a number of difficulties, most of which arise when establishing the control group of relatively healthy individuals. In this work, survival functions were analyzed for carriers of alternative polymorphic variants of 18 genes that had been tested for association with MI using the case-control approach in our previous study, and the magnitude of the shift in the age of the disease onset depending on individual variations of the genome was estimated. The following risk variants were associated with the age of MI: rs2430561*A of IFNG (HR = 1.3, P = 0.043), rs1799889*5 of PAI-1 (HR = 1.3, P = 0.039), rs1800896*GG of IL10 (HR = 1.5, P = 0.0048), rs1800471*C of TGFB1 (HR = 1.5, P = 0.043), and rs11614913*TT of MIR196A2 (HR = 1.5, P = 0.035). In carriers of these variants, the disease developed 3-6 years earlier than in carriers of alternative variants. The results of this study were compared with data on the associations with MI previously obtained on the same sample using the case-control approach. It turned out that the estimates based on the two methods mostly disagreed. However, the age-dependent approach relies on fewer assumptions that can be additionally verified. In our opinion, it makes this approach more promising than the case-control design.
Keywords: age-dependent diseases; machine learning; myocardial infarction; single nucleotide polymorphisms; survival analysis.