Interpretation:

The risk of the prespecified primary endpoint of fatal CHD or nonfatal MI did not differ between initial use of the diuretic chlorthalidone versus initial use of the ACE inhibitor lisinopril or the calcium antagonist amlodipine for the treatment of hypertension. However, for certain secondary outcomes, such as CHF, rates were lower among patients treated with chlorthalidone versus lisinopril or amlodipine. For amlodipine, this CHF difference persisted into long-term follow-up. Each of the three drugs reduced BP from baseline, although chlorthalidone use was associated with larger systolic BP reductions versus lisinopril or amlodipine. Chlorthalidone seemed to be associated with a lower risk of hip/pelvic fractures.

It has been speculated that potential metabolic effects of diuretics such as hypokalemia, dyslipidemia, and insulin resistance might offset the beneficial effects of BP reduction. As a result, therapies without such potential side effects, such as ACE inhibitors and CCBs, have frequently been used for the treatment of hypertension, with the belief that CHD outcomes would be improved with these therapies over diuretics. While some side effects were more common in the chlorthalidone arm, such results did not translate to an increase in CV events.

A potentially important side effect in the chlorthalidone arm was the higher fasting glucose levels versus the lisinopril or amlodipine arms in an analysis of all patients and in nondiabetics. In contrast, ACE inhibitors have been associated with a reduction in the development of diabetes and the progression of diabetic nephropathy. The impact of chlorthalidone on diabetes and CVD may not be fully manifested in the relatively short follow-up period of 4 years, the last timepoint reported in the manuscript.

One criticism of the trial is that the ACE inhibitor arm was potentially at a disadvantage since the first add-on therapy specified by the trial treatment algorithm for this arm was a beta-blocker, rather than a diuretic or CCB, both of which are more commonly used in clinical practice. The increased risk of heart failure in the lisinopril arm was unexpected and is in contrast to the benefits of ACE inhibitors observed in other trials for the treatment of heart failure, such as SOLVD.

The present analysis was an intent-to-treat analysis. Because of the large rate of crossover in both arms (13% of patients in the chlorthalidone arm were taking a diuretic with a CCB or an ACE inhibitor and 9% were taking a CCB or an ACE inhibitor without diuretic), it will be important to see the results of an "as treated" analysis.

Because BP reduction with the ACE inhibitor lisinopril was lower than with the diuretic chlorthalidone, it is unclear if the improvement in the secondary clinical outcomes with chlorthalidone over lisinopril was due to greater improvements in BP or other effects of the diuretic. In the case of amlodipine, the situation is more complicated. The reduced risk of CHF in the chlorthalidone arm versus the amlodipine arm may represent a drug-specific effect since the BP differences between amlodipine and chlorthalidone were small (<1 mm Hg) and there was lower systolic BP in chlorthalidone, but lower diastolic BP for amlodipine.

Further substudies addressing the economic impact of these medications in ALLHAT for treatment of hypertension are forthcoming. Given the fact that more than 40 million adults have hypertension and 21.9 million are treated with medication, the ALLHAT study may have a potentially large economic impact on health care delivery.

In the lay press, it has been emphasized that patients not discontinue or change current hypertensive medications without consulting their physician. Physicians can anticipate being contacted by patients about the issues raised in the ALLHAT trial.