Antihypertensives and Central Aortic Pressure

Abstract & Commentary

By Michael H. Crawford, MD. Dr. Crawford is a Professor of Medicine, and Chief of Clinical Cardiology, at the University of California, San Francisco; and he is the Editor of Clinical Cardiology Alert.

Source: Williams B, et al. Differential Impact of Blood Pressure-Lowering Drugs on Central Aortic Pressure and Clinical Outcomes: Principal Results of the Conduit Artery Function Evaluation (CAFE) Study. Circulation. 2006;113:1213-1225.

The observation that beta-blockers do not reduce cardiovascular events as compared to other agents with equal blood pressure lowering in clinical trials may be due to different effects on central aortic pressure. Thus, a substudy of the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT), the Conduit Artery Functional Evaluation (CAFE) study was conducted. The main objective of the CAFE study was to test this hypothesis. A second objective of CAFE was to evaluate the relationship between central aortic pressure and cardiac events.

ASCOT randomized patients 40 to 79 years of age with untreated or inadequately treated hypertension and at least 3 additional cardiovascular disease risk factors, including male sex, to atenolol plus thiazide or amlodipine plus perindopril as required. CAFE used radial artery tonometry pulse wave analysis to derive central aortic pressure. Of the 2199 ASCOT patients recruited for CAFE, 126 were excluded for inadequate pulse waveforms for analysis. The participants were largely white men (mean age, 63) with a mean pretreatment blood pressure of 160/93 mm Hg.

At the end of the study, brachial artery cuff pressures were similar and had decreased similarly (-26/-14 atenolol + thiazide vs -29/-16 mm Hg for amlodipine + perindopril). Derived central aortic pressure was reduced more by the amlodipine-based as compared to atenolol-based regimen (difference 4.3/1.4 mm Hg, P < .001). These differences were consistent throughout the study period. As expected, heart rate was significantly lower with the atenolol-based therapy. Central aortic pulse pressure was associated with a composite end point of total cardiovascular procedures/events and the development of renal dysfunction adjusted for baseline comorbidity (P < .05), as was peripheral pulse pressure (P = .05). Williams and colleagues concluded that different antihypertensive therapies may have different effects on central aortic pressure despite similar changes in brachial cuff blood pressure. Central aortic pulse pressure is related to cardiovascular outcomes, and may explain the different outcomes between the 2 treatment arms of ASCOT.


Over the years we have seen various theories advanced for picking one antihypertensive agent over another (eg, race, left ventricular hypertrophy regression), and now we have the affect on central aortic pressure. Smaller studies of shorter duration have shown that some antihypertensive drugs have less effect on central vs peripheral pressure, but this is the largest and longest duration study to document this. Also, the ASCOT study showed reduced cardiovascular events, including death and stroke, on amlodipine plus perindopril as compared to atenolol plus a thiazide. The central aortic pressure effect hypothesis is one potential explanation for this observation. However, the differences in central vs peripheral pressure in CAFE were small (4/1 mm Hg). Are these clinically significant?

The outcomes data in the CAFE study suggested that central aortic pulse pressure was predictive of a combined cardiovascular end point that included procedures and renal function, but so was peripheral pulse pressure. The CAFE subgroup did not establish that central pressures were more predictive of outcomes than peripheral pressures. Thus, the explanation for the differences in the ASCOT study were not proven to be the effects on central aortic pressure; but this hypothesis was not disproved because the subgroup was underpowered for outcomes. In addition, the fact that central pressures did correlate with outcomes keeps this hypothesis alive.

Recently, there has been much discussion about the lack of efficacy of beta-blockers for reducing cardiovascular outcomes in hypertensive patients. Also, beta-blockers have been shown to be less effective than other agents for reducing left ventricular hypertrophy and peripheral arterial medial thickness. In addition, brain natriuretic peptide levels are increased by beta-blockers vs reduced on other antihypertensives, suggesting that beta-blockers do not reduce central aortic and, hence, left ventricular systolic pressure as well. In the CAFE study, Williams et al suggest that heart rate reduction is the likely explanation for these differences. A slower heart rate allows reflected waves from the periphery time to augment the next systolic wave in the central aorta. The problem with this hypothesis is that a recent meta-analysis has shown that only studies using atenolol failed to show improved cardiovascular outcomes, but not those using metoprolol, yet both decrease heart rates. Perhaps the answer is that there is a problem with atenolol. If heart rate is the explanation, then other classes of drugs that lower heart rate (eg, diltiazem) should also be less effective. There is no data to support this. Thus, the mechanistic explanation for these results is still unclear.

There are several limitations to the CAFE study. The most important is that central aortic pressure was not measured, but derived from peripheral pulse waves. Apparently this method is well-validated, but for the non-physicist doubt lingers. Also, this is a substudy of a trial designed for other purposes, and bias could be introduced. The groups were well-matched and seemed to reflect the larger ASCOT study, but here were some baseline differences of unknown clinical significance. In addition, this was a relatively select patient population of largely older white men with moderate risk for coronary artery disease. Perhaps in such a group there is already some arterial stiffening that would augment differences between drugs on central aortic pressure.

At this point it may be prudent to reserve beta-blockers for hypertensive patients who need them for other reasons or for those who fail other agents, but don't be surprised if this rationale for picking one antihypertensive vs another also fades with time and more information.