Prevention of Heart Failure in Hypertension—the Role of Coronary Heart Disease Events Treated With Versus Without Revascularization: The ALLHAT Study

Abstract

In modern clinical practice, less than half of patients with new-onset heart failure (HF) undergo ischemic evaluation and only a minority undergo revascularization. We aimed to assess the proportion of the effect of hypertension (antihypertensive treatment) on incident HF to be eliminated by prevention of coronary heart disease (CHD) event treated with or without revascularization, considering possible treatment-mediator interaction. The causal mediation analysis of Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) included 42,418 participants (age 66.9 ± 7.7, 35.6% black, 53.2% men). A new CHD event (myocardial infarction or angina) that occurred after randomization but before the incident HF outcome was the mediator. Incident symptomatic congestive HF (CHF) and hospitalized/fatal HF (HHF) were the primary and secondary outcomes, respectively. Logistic regression (for mediator) and Cox proportional hazards regression (for outcome) were adjusted for demographics, cardiovascular disease history, and risk factors. During a median 4.5-year follow-up, 2,785 patients developed CHF, including 2,216 HHF events. Participants who developed CHD events had twice the higher incidence rate of CHF than CHD-free (28.5 vs 13.9 events/1,000 person-years). The proportion of reference interaction indicating direct harm because of a CHD event for lisinopril (234% for CHF, 355% for HHF) and amlodipine (244% for CHF, 468% for HHF) was greater than for chlortalidone (143% for CHF, 269% for HHF). In patients with revascularized CHD events, chlortalidone and amlodipine eliminated 21% to 24% and lisinopril eliminated −45% of HHF. Antihypertensive treatment could not eliminate harm from CHD events treated without revascularization. In conclusion, the antihypertensive drugs (chlortalidone, lisinopril, and amlodipine) prevent HF not principally by preventing CHD events but by way of other pathways. HF is moderated but not mediated by CHD events. Revascularization of CHD events is paramount for HF prevention.