Are cardiovascular outcomes related to antihypertensive treatment in older women?

Postmenopausal women aged 50 to 79 years enrolled in the Women’s Health Initiative Observational Study (WHI-OS) in the United States were evaluated for this observational study over a mean of 5.9 years. Only subjects with hypertension without cardiovascular disease (CVD) receiving monotherapy with an angiotensin-converting enzyme (ACE) inhibitor, beta blocking agent, calcium channel blocker, diuretic, or combination of a diuretic plus one of the other drugs were included. Of the original 93,676 women in the WHI-OS study, 15,787 subjects met the entry criteria and 11,294 were receiving monotherapy. End points of the study included coronary artery disease, stroke, and CVD mortality assessed by the medications used at baseline. The authors noted an association between medication classes and outcomes, but no causal conclusion can be drawn from this study as discussed below.

The major weakness of this study is its observational design. The factors that determined the initial antihypertensive medication choice for the patient cannot be ascertained. The physician’s rationale for choosing the calcium channel blocker may have been based on the patient’s overall medical condition, which precluded the use of diuretics and beta blockers. This may have resulted in sicker patients being included in the calcium channel blocker group.

The patients taking diuretics, either as monotherapy or as combined therapy, had the lowest baseline blood pressure. Conversely, the group receiving monotherapy with calcium channel blockers or combined beta blocker and diuretic therapy had the highest baseline blood pressure readings. The severity of the initial blood pressure may have affected the medication choice. For example, those patients with mild hypertension and a few comorbidities may have been placed on a diuretic, and those with multiple illnesses and more severe hypertension may have been placed on a calcium channel blocker. There is no way to ascertain this. The blood pressure readings at or around the time of the end point are also unknown. Was the end point related to the inadequate blood pressure control, the medication class effect, or neither? Such conclusions cannot be derived owing to the study design. The authors state that they adjusted for some, but not all, confounding factors, and they themselves conclude that the results may have been a chance finding. The significant amount of adjustment required highlights the bias inherent in this study and places the result in more doubt as a true finding.

It is a concern that no breakdown of calcium channel blocker class or types was provided or analyzed. There are two types of calcium channel blockers, dihydropyridines (DHPs) with potent vasodilator action without cardiodepressive effects and non-DHPs with greater suppressive effects on cardiac contractility and conductivity. No such distinction was made. Further, no differentiation was made about the type of calcium channel blockers by the half-life. This trial was started in 1994, but the concerns with short-acting calcium channel blockers were not recognized by the Food and Drug Administration until January 1996; it is very likely that patients on short-acting calcium channel blockers were included. The inference made from this observational study may not be translated to any specific calcium channel blocker on the market today. For example, one of the more commonly prescribed medications, amlodipine, may have a completely different outcome. The hypertension arm of the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT), involving amlodipine and ACE inhibitors, was stopped early because of the benefit in the amlodipine arm compared with diuretics and beta blockers.

This study also makes several assumptions about medication use. Medication use was not followed for the length of the study; medication data were evaluated only at baseline and in year 3. It was simply assumed that patients remained on the medication they were prescribed at baseline. As the authors noted, there was no way of knowing whether the drugs were halted or switched during the trial. Thus, this study cannot link medication to outcome for any of the classes. Adherence issues were also not verified.

The end points in this study may not be reliable. They were ascertained from mail-in questionnaires and telephone interviews using potential outcomes based on available medical records and death certificates. The criteria for CVD death were not stringent. A definite cause of death due to CVD was not required; as defined in this study, death from “possible coronary artery disease” and “CVD of unknown type” were included in CVD mortality.

The authors note one important observation. Among the women with hypertension and diabetes, only a minority (21.1%) had blood pressures lower than 130/80 mm Hg. As the authors stated, many patients are not being treated aggressively enough to control their blood pressure. This is an important point for all to remember.