Regression of electrocardiographic left ventricular hypertrophy is associated with less hospitalization for heart failure in hypertensive patients

Background: Reduction of electrocardiographic left ventricular hypertrophy (LVH) has been associated with decreased cardiovascular death, stroke, myocardial infarction, and atrial fibrillation. However, whether reduction of electrocardiographic LVH is associated with decreased heart failure is unclear. Objective: To examine the relation of reduction of electrocardio graphic LVH to incident heart failure. Design: Multicenter cohort study derived from a randomized, con trolled trial. Setting: Losartan Intervention For Endpoint reduction in hyperten sion study. Patients: 8479 hypertensive patients without history of heart failure who were randomly assigned to losartan or atenolol treatment. Measurements: Change in Cornell product electrocardiographic LVH between baseline and in-study electrocardiograms, examined as both a continuous variable and a dichotomous variable (above or below the median decrease of 236 mm . msec) to predict heart failure hospitalization occurring after the 6-month follow-up visit. Results: During mean follow-up of 4.7 years (SD, 1.1 years), 214 patients were hospitalized for heart failure (2.5%): 77 patients with an in-treatment decrease of 236 mm . msec or more (4.4 per 1000 patient-years) and 137 patients with a reduction less than 236 mm msec during treatment (6.8 per 1000 patient-years). In a univariate Cox analysis in which change in Cornell product was treated as a time-varying continuous variable, decrease in Cornell product during treatment was associated with a decreased risk for new-onset heart failure, with a 24% lower risk for heart failure for every 817-mm . msec (1 SD of the mean) lower Cornell product (hazard ratio, 0.76 [95% Cl, 0.72 to 0.80]). In a parallel analysis in which change in Cornell product was entered as a time-varying dichotomous variable, a greater-than-median in-treatment decrease in Cornell product (236 mm . msec) was associated with a 43% lower risk for heart failure (hazard ratio, 0.57 [CI, 0.44 to 0.761). After adjustment for treatment, baseline risk factors for heart failure, baseline and in-treatment blood pressure, and baseline severity of electrocardiographic LVH, in-treatment decrease of Cornell product LVH in time-varying multivariate Cox models remained strongly associated with new heart failure hospitalization, with a 19% lower risk for every 817-mm . msec lower Cornell product treated as a continuous variable (hazard ratio, 0.81 [CI, 0.77 to 0.85]) or a 36% decreased rate of new heart failure in patients with an in-treatment reduction in Cornell product of 236 mm . msec or more (hazard ratio, 0.64 [Cl, 0.47 to 0.89]; P < 0.001 for all comparisons). Limitations: Use of electrocardiographic LVH to select patients may have increased risk compared with unselected hypertensive patients, and use of hospitalization for heart failure as the end point will underestimate the incidence of new heart failure. Conclusion: Reduction in Cornell product electrocardiographic LVH during antihypertensive therapy is associated with fewer hospitalizations for heart failure, independent of blood pressure lowering, treatment method, and other risk factors for heart failure.