The interaction effect of cardiac and non-cardiac comorbidity on myocardial infarction mortality: A nationwide cohort study

Abstract

BackgroundWhether the prognostic impact of comorbidity on myocardial infarction (MI) mortality is due to comorbidity alone or/and its interaction effect is unknown. MethodsWe used Danish medical registries to conduct a nationwide cohort study of all first-time MIs during 1995–2016 (n = 179,515) and a comparison cohort matched on age, sex, and individual comorbidities (n = 880,347). We calculated age-standardized 5-year all-cause mortality rates. Interaction was examined on an additive scale by calculating interaction contrasts (difference in rate differences). ResultsAmong individuals without comorbidity, the 30-day mortality rate per 1000 person-years was 1851 (95% CI: 1818–1884) for MI patients and 22 (21–24) for comparison cohort members (rate difference = 1829). For individuals with low comorbidity, corresponding baseline mortality rates were 2498 (2436–2560) in the MI and 54 (50–57) in the comparison cohort (rate difference = 2444). The interaction contrast (616) indicated that the interaction accounted for 25% (616/2498) of the total 30-day mortality rate in MI patients with low comorbidity. This percentage increased further for moderate (35%) and severe (45%) comorbidity levels. Absolute and relative interaction effects were largest within the first 30 days and younger individuals. Dose-response patterns were also observed during 31–365 days and 1–5 years of follow-up (p-values for trends<0.002). The interaction differed substantially between individual types of cardiac and non-cardiac comorbidities. ConclusionCardiac and non-cardiac comorbidities interact with MI to increase short- and long-term mortality beyond that explained by their additive effects. The interaction had a dose-response relation with comorbidity burden and a magnitude of clinical importance.