Antihypertensive Medications

To the Editor: We have read with interest the recent report by Heckbert and colleagues.' However, we feel the suggestion that treatment with calcium channel blockers or loop diuretics would be associated with worse cognitive function in older hypertensive patients is not substantiated by the evidence presented. First, the analysis included only 1268 of 2815 potentially eligible hypertensive patients (45%). The 2815 hypertensive patients represented are already a selected group as the overall response rate was 58%. In addition, the authors did not state explicitly that all study participants underwent a magnetic resonance imaging (MRI) scan. Consenting patients are self-selected to a large extent. Furthermore, the authors included the 3073 normotensive subjects in only part of their analyses although a substantial proportion of them must have been taking cardiovascular drugs for reasons other than hypertension, e.g., angina pectoris. Why Heckbert and colleagues' chose patients taking p-blockers as the reference group (104 patients receiving monotherapy) rather than those taking thiazides remains unclear. Indeed, a total of 479 patients were taking thiazide diuretics either alone (n = 254) or in combination with other agents (n = 225). Thiazide diuretics, according to the national guidelines in the United States; remain the drug of choice to treat older hypertensive patients. Moreover, the results on dementia observed in the Systolic Hypertension in the Elderly Program (SHEP) trial3 are not referenced. In this prospective trial, diuretic-based antihypertensive treatment, compared with placebo, did not significantly change the incidence of dementia. This null result in a long-term trial with a double-blind, placebo-controlled design is another reason why selecting patients prescribed diuretics as the reference group would have been more appropriate. As in other observational studies on the adverse effects of medications, confounding by indication is a potential source of bias in Heckbert's study.' Hypertensive patients treated with calcium channel blockers and loop diuretics could have differed in some way from those treated with p-blockers. Also, the suggestion that a higher white matter grade on an MRI scan correlated with a higher dose of the calcium channel blocker is questionable. A test statistic with a P value of .163 is nonsignificant, even in evaluating a one-sided hypothesis. Furthermore, the authors' only considered the antihypertensive medications at the annual visit just preceding the date of the MRI scan and did not attcmpt to account for the duration of drug usage or the total number of patient-years on specific antihypertensive agents. On the other hand, Heckbert and colleagues' are right in stating that randomized trials of the long-term use of antihypertensive agents are needed to assess possible causal associations between brain abnormalities and the intake of these drugs. The Vascular Dementia substudy4 to the placebocontrolled, double-blind Syst-Eur trial' aims to provide information about the effects of antihypertensive treatment on the incidence of dementia and on the change in the Mini-Mental State score in older patients with isolated systolic hypertension4 Active treatment consists of the dihydropyridine calcium channel blocker nitrendipine (10-40 mg/day) with the possible addition in treatment-resistant patients of enalapril (5-20 mg/day) and/or hydrochlorothiazide (12.5-25 mg/ day).' A total of 3110 patients have been enrolled in this substudy. The principal results will become available in the course of 1998.