Overall cardiac mortality has fallen dramatically in the United States over recent decades, yet that headline improvement obscures a troubling shift in the underlying drivers — one that has direct implications for whether future gains in heart disease survival can continue. Even as total ischemic heart disease deaths declined from the 1990s onward, the proportion of those deaths attributable to elevated BMI and hyperglycemia has grown, signaling that metabolic dysfunction is quietly filling the vacuum left by better-controlled traditional risk factors like smoking and hypertension.
Published in JAMA Cardiology, the study used a comparative risk assessment framework applied to U.S. mortality data spanning roughly three decades. Researchers calculated the population-attributable fraction — essentially the share of ischemic heart disease deaths that would be prevented if specific risk factors were eliminated — and tracked how those fractions changed over time. BMI-related and hyperglycemia-related attributable deaths rose as a proportion of total ischemic heart disease mortality even while absolute death counts fell. This means that gains from declining smoking rates and improved blood pressure management are increasingly offset by the metabolic consequences of obesity and insulin dysregulation.
This finding sits at a critical inflection point in cardiovascular epidemiology. For decades, public health victories against smoking and cholesterol drove the dramatic decline in heart disease mortality. The data here suggest that curve is now bending — metabolic risk is becoming the dominant residual driver. From a longevity standpoint, this is significant: BMI and hyperglycemia are modifiable, yet they track closely with dietary patterns, physical inactivity, and socioeconomic conditions that are notoriously difficult to shift at a population level. The emergence of GLP-1 receptor agonists as effective weight and glucose modulators may eventually register in mortality data, but that signal is years away. Importantly, population-attributable fraction analyses are observational and assume causal relationships from epidemiological associations, so confounding remains a limitation. Still, the trend is directionally clear and consistent with parallel findings from Europe and Asia.
