Reply to Editorial

Abstract

We agree with Professor Bulpitt (1) that if a study on antihypertensive treatment is not placebo controlled, clinic blood pressure and side-effect data have a limited scientific value. For this reason, we reported them briefly (to characterize our population by a commonly used set of values) and focused on ambulatory blood pressure measurements (2), which we thought were scientifically valid because (a) the number of patients in whom ambulatory blood pressure monitoring could be performed was rather large; (b) patients' age extended to the so-called "old-elderly" range, in which information on ambulatory blood pressure changes during antihypertensive treatment is still limited; and (c) ambulatory blood pressure is substantially devoid of a placebo effect. On this crucial third point, we disagree with Professor Bulpitt's considerations. There is now a large body of evidence that, unlike clinic blood pressure, 24-h, daytime, and nighttime average blood pressures are not significantly less when placebo is administered for weeks or months (3,4). It is also clear that the data from Staessen et al. (5) should not be interpreted as at variance with these results. One, in the study by Staessen et al. (5), the "placebo" patients showed a reduction in 24-h average systolic blood pressure that was not only small (2.4 mm Hg) but also about a third the concomitant reduction in clinic systolic blood pressure (6.6 mm Hg). Two, whereas in the placebo group, clinic diastolic blood pressure was significantly reduced (1.4 mm Hg), ambulatory diastolic blood pressure was not. Three, and most important, in the study by Staessen et al. (5) statistically significant blood pressure reductions were seen only after administration of placebo for 1 year. This makes the results somewhat irrelevant for antihypertensive drug studies, in which only short- or medium-term placebo administration is ethically acceptable. It also makes interpretation of the data difficult because it is conceivable that in elderly patients with systolic hypertension [i.e., the patients involved in the study by Staessen et al. (5)], a small long-term reduction in systolic blood pressure is caused by a decline in cardiac function. We have also some reservations on Professor Bulpitt's definition of a "controlled" study. To us this does not mean that a study has to answer all or most of the important points concerning an issue, but that its design allows the results to be interpreted according to the original hypothesis, even if single and simple. For us the hypothesis was whether a combination of a low-dose thiazide and an ACE inhibitor (a treatment largely used as first line in the clinical practice) effectively reduced blood pressure over the 24 h in aged patients (i.e., in patients in whom suboptimal or undesirable blood pressure effects of treatment cannot be easily excluded). We believe our study allowed this to be obtained. It did not allow, of course, determination of the role of the combination components in the ambulatory blood pressure effect, an interesting question that would have required, as Professor Bulpitt emphasizes, a parallel design involving three groups of patients or a cross-over design on a single group of patients studied with the three different treatments.