Heckman reiterates a number of published criticisms of ALLHAT. In the Cardiovascular Health Study,¹ only 46 (1.3%) of 3579 older adults without clinical cardiovascular disease had an abnormal ejection fraction on echocardiogram. New HF events in older adults usually occur with preserved systolic function.² ACE inhibitors have not been evaluated in the setting of HF with preserved systolic function, so generalisations from the existing HF trials may well be uninformative. Except in the setting of a myocardial infarction, the case fatality for new HF is low, and the follow up time after HF events in ALLHAT would have provided little or no power to assess associations with mortality.

Trials should be judged and interpreted by the quality of their methods. In double blind trials such as ALLHAT, misclassification of an endpoint introduces a nonsystematic error that generally moves point estimates toward the null. Heckman cites the STOP-Hypertension 2 trial. In this trial, only the endpoint classification is blinded. Thus, physicians and subjects, aware of the treatment allocation, have considerable influence in “ascertaining” which events go to this blinded endpoint committee. The absence of double blinding yields effect estimates that are exaggerated by 17% on average.³

Many patients who entered ALLHAT had been receiving treatment at recruitment. At randomisation, some of them were switched to low dose diuretic therapy. Patients randomised to low dose diuretics had reduced risks of ≥1 major cardiovascular outcome. The public health rationale for using a diuretic as first line therapy is well supported. Switching patients from ACE inhibitors to low dose diuretics would save billions of dollars and prevent tens of thousands of cardiovascular events.