Conduit Artery Function Evaluation - CAFE


The goal of the trial was to evaluate the effect on central aortic pressure of treatment with an antihypertensive regimen of calcium channel blocker with or without an angiotensin-converting enzyme (ACE) inhibitor compared with a regimen of beta-blocker with or without a diuretic among hypertensive patients.

Study Design

Study Design:

Patients Enrolled: 2,073
Mean Follow Up: Median 5.5 years

Patient Populations:

Enrollment in the ASCOT - Blood Pressure Arm trial

Drug/Procedures Used:

Patients in the ASCOT trial were randomized open-label to one of the two antihypertensive treatments: amlodipine 5 mg (n=9,639) or atenolol 50 mg (n=9,618). In order to achieve target blood pressure goals of <140/90 mm Hg, study drug doses were increased and second-line drugs were added (perindopril 4 mg for the amlodipine group and bendroflumethiazide 1.25 mg for the atenolol group) as necessary. In the CAFE sub-study (n=2,073) of the ASCOT trial, central aortic pressure was assessed using the non-invasive Sphygmocor system, in which a computer program estimates the central aortic pressure by examining the shape of the pulse wave at the wrist. The CAFE sub-study started one year after randomization in the ASCOT trial.

Principal Findings:

Among the 2073 patients in the CAFE sub-study, 1042 were in the amlodipine group and 1031 were in the atenolol group. Baseline characteristics were well matched both between randomization groups within the substudy and with the overall trial. The average number of risk factors was 3.7 and 24.4% of patients were diabetics. In the calcium channel blocker group, amlodipine was used in 90.0% of patients, perindopril in 56.0% of patients, and both in 47.8% of patients. In the beta-blocker group, atenolol was used in 88.1% of patients, bendroflumethiazide in 69.8%, and both in 60.2% of patients.

Brachial systolic blood pressure (SBP) did not differ by treatment group (133.2 mmHg for amlodipine vs 133.9 mmHg for atenolol group, p=NS), but diastolic blood pressure (DBP) was slightly lower in the amlodipine group (76.9 vs 78.6 mmHg, p<0.001). The area under the curve for central SBP was lower in the amlodipine group (121.2 vs 125.5 mmHg, p<0.001). Likewise, central pulse pressure was lower with amlodipine (43.4 vs 46.4 mmHg, p<0.001), while brachial pulse pressure did not differ (56.2 vs 55.3 mmHg, p=0.06).


Among hypertensive patients, treatment with a regimen of calcium channel blocker with or without an ACE inhibitors was associated with reductions in central systolic blood pressure and pulse pressure compared with a regimen of beta-blocker with or without a diuretic, despite no difference in brachial blood pressure.

The overall results in the ASCOT trial demonstrated a reduction in many secondary endpoints, including mortality, associated with a first line regimen of calcium channel blocker therapy with or without an ACE inhibitor. While minor reductions in brachial systolic blood pressure were observed in the amlodipine/perindopril strategy, the small difference were not large enough to account for the entire clinical benefit. The present substudy suggests that reductions in blood pressure may have been greater than originally estimated when assessed using brachial artery evaluation. Systolic pressure is not constant throughout the arterial tree. Central aortic assessment, however, measures the pressure in the main arteries adjacent to the heart.


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:epub before print.

Presented by Dr. Bryan Williams at the American Heart Association Scientific Session, Dallas, Texas, November 2005.

Clinical Topics: Prevention, Statins, Hypertension

Keywords: Perindopril, Brachial Artery, Bendroflumethiazide, Diuretics, Blood Pressure, Risk Factors, Calcium Channel Blockers, Drug Combinations, Wrist, Amlodipine, Hypertension, Diabetes Mellitus

< Back to Listings