Amlodipine or lisinopril was not better than chlorthalidone in lowering CHD risk in hypertension.

Although not mentioned by Dr. Psaty in his commentary (1), concerns have been raised over the heart failure results in ALLHAT (2). Asymptomatic left ventricular (LV) systolic dysfunction is common in older patients, particularly men and patients with cardiovascular risk factors (3, 4). ALLHAT participants were high-risk patients, and many may have had asymptomatic LV dysfunction that could have been unmasked at the start of the study by withdrawal of diuretic therapy. The early difference in heart failure (HF) incidence observed between the lisinopril and chlorthalidone groups, as well as the merging of the HF curves toward the end of the trial, is consistent with this possibility. Furthermore, a true difference in HF incidence over 5 years should have resulted in greater mortality in the lisinopril group; this did not occur. A second concern is the validation of the HF endpoint, the diagnosis of which was up to the local ALLHAT investigator. An ALLHAT subcommittee, which reviewed a small fraction of HF hospitalizations, only agreed with 85% of diagnoses (5). In contrast, blinded endpoint committees in STOP-Hypertension 2 (6) and in ANBP2 (7), both of which included an angiotensin-converting enzyme (ACE) inhibitor group, found no difference in HF incidence. In the latter study, men appeared to benefit more from ACE inhibitors than diuretics, consistent with the epidemiology of asymptomatic LV dysfunction (3, 4). Third, there seem to be race-based differences between the lisinopril and chlorthalidone groups with respect to the risks for stroke, and possibly combined coronary heart disease, combined cardiovascular disease, as well as HF. Though the statistical significance of these apparent interactions is not provided in the original report, they are not unexpected (2, 8). The ALLHAT results emphasize the importance of diuretics in hypertension treatment. However, data do not support Dr. Psaty’s assertion that “If BP is controlled with a nondiuretic, the patient should be switched to a low-dose diuretic.” By ignoring the arguments for and against the ALLHAT results, we may neglect to pursue potentially important racial and gender differences in hypertension treatment. George A. Heckman, MD, FRCPC Heart and Stroke Fellow Geriatrics and Internal Medicine McMaster University Hamilton, Ontario, Canada