Effects of antihypertensive drugs on central blood pressure: new evidence, more challenges

βa-blocking drugs have long been used as a first-line therapy for hypertension because of their well-established efficacy in lowering brachial blood pressure (BP)¹ and their long-term favorable effects on all-cause and cardiovascular mortality, as shown in a number of prospective large-scale trials and meta-analyses.² More recently, their use as a preferred strategy for the treatment of essential hypertension has been challenged, as their protective effect on incident stroke may be suboptimal.³ In the Losartan Intervention For End point reduction in Hypertension study, a treatment based on atenolol was less effective than one based on losartan in achieving left ventricular mass reduction⁴ and in preventing cardiovascular morbidity and mortality.⁵

β-Blockers and central BP

One of the explanations why β-blockers may have a suboptimal protective effect on cerebrovascular complications and left ventricular hypertrophy is related to their effects on central BP. BP measured at peripheral arteries may not accurately reflect the level of BP closer to the heart, such as at the ascending aorta and at the carotid arteries. This is particularly true for systolic and pulse pressures, which are normally higher at the level of the brachial artery than at the level of the central arteries, a phenomenon termed as BP ‘amplification’, whereas diastolic and mean BP differ only slightly.⁶ The difference between central and peripheral BP is determined by a number of factors, including cardiac ejection, arterial impedance, heart rate, and wave reflections.⁷ In the Conduit Artery Function Evaluation substudy of the Anglo-Scandinavian Cardiac Outcomes Trial, the subjects randomized to atenolol had a higher in-treatment aortic systolic BP than those receiving amlodipine-based treatment, despite brachial BP levels not being different in the two groups.⁸ In a substudy of a double-blind randomized intervention trial comparing atenolol with perindopril/indapamide, the atenolol group showed a smaller reduction in central systolic/pulse pressure and in left ventricular mass, and the differences in left ventricular mass between the two drug regimens were significantly linked to central, but not brachial, pulse pressure change.⁹ Findings from these trials have raised the hypotheses that antihypertensive drugs may not be equal in reducing central BP regardless of similar effects on brachial BP, and that central BP is more tightly related to target organ damage than brachial BP.

A number of other clinical trials, mainly based on older generation β-blockers without ancillary vasodilating properties, have also shown that β-blockers decrease central systolic BP less than brachial systolic BP.^{10, 11, 12, 13} As a result, central-to-brachial systolic BP amplification usually decreases after treatment with β-blocking drugs. The mechanism through which β-blockers reduce BP amplification has been incompletely understood. It is commonly believed that the heart rate reduction induced by β-blockers may prolong systolic time duration and thus determine a relatively earlier arrival of the reflected pressure wave in the systolic phase, thereby increasing central systolic BP and augmentation index (AIx).^{14, 15, 16} Whereas some authors observed that β-blockers may increase total peripheral resistance at least initially,¹⁷ no evidence of altered pressure wave reflections has been provided after 1 year of treatment with atenolol.¹⁸ Furthermore, there is little or no evidence that the suboptimal reduction in central BP with β-blockers, as compared with other BP-lowering drugs, may be related to a lower impact on central arterial stiffness.¹⁹

The present study

The above effects of β-blockers on central hemodynamics might not be a universal consequence of β-blocking treatment, however. Some β-blockers with a peripheral vasodilatory effect, such as nebivolol, labetalol and carvedilol, have generally shown a more favorable effect on central hemodynamics.^{20, 21, 22} Clearly, there is a need for more trials to confirm these observations. In the present issue of Hypertension Research,²³ Kim et al.²³ provide new evidence on the widely debated issue of the impact of different antihypertensive drug classes on central BP. The effects of a vasodilating β-blocker have been investigated in a head-to-head comparison with an angiotensin-II receptor blocker, a drug class that had previously demonstrated a neutral or favorable effect on BP amplification.²⁴ In a 24-week, prospective, open-label clinical trial, 182 patients with mild-to-moderate hypertension were randomized to receive carvedilol, a β-blocker with vasodilating properties, or losartan, an angiotensin-II receptor antagonist. The primary aim was to compare the efficacy of the study drugs on central BP, AIx and aortic pulse wave velocity (PWV). As a secondary end point, the effects on metabolic and inflammatory profile were assessed. Aortic PWV was measured with a semiautomatic device (PP-1000, Hanbyul, Korea), which uses sensors attached with velcro to an elastic band around the neck and the thigh. The distance between the two recording sites was derived from age-, gender- and height-based regression equations, and transit time was measured with the foot-to-foot method. Central BP parameters were estimated from the transformation of radial waveform profile (measured by the Gaon Semiautomated System, Hanbyul, Korea) through a not well-specified and not validated generalized transfer function, after calibration of radial waveform to brachial cuff BP values. The study design included a starting daily dose of losartan 50 mg or carvedilol 12.5 mg; doses were doubled and hydrochlorothiazide was added when needed, to achieve an optimal brachial systolic/diastolic BP control (<140/90 mm Hg, or <130/80 mm Hg if diabetic). The mean daily dose at the end of the study was 81 mg for losartan, 21.8 mg for carvedilol and 17.1/17.8 for hydrochlorothiazide.

The two drugs had a similar effect on brachial systolic BP (−15.1±15 mm Hg for losartan and −14.8±15 mm Hg for carvedilol). More importantly, losartan and carvedilol did not differ in their impact on central systolic BP (−14.3±15 mm Hg for losartan and −13.6±16 mm Hg for carvedilol). Similarly, no between-drug differences were found in diastolic BP reduction, either determined at the brachial or central level. Losartan, not carvedilol, induced a small but significant reduction in aortic AIx (−5.0±16.3% vs. +0.18±12.0%, P=0.024 for between-group difference), although this difference was no longer significant (P=0.075) when AIx was adjusted for heart rate. Thus, in this study, the minor although significant losartan-induced reduction in AIx may not have a major influence on central BP.

This study suggests that carvedilol, a β-blocking agent with vasodilating properties, may differ from older generation β-blockers in its action on central hemodynamics. The combined non-selective blocking effects of carvedilol on α₁- and β-receptors has been proposed to be the primary mechanism of its vasodilating action,²⁵ and it can be hypothesized that peripheral vasodilatation may reduce pressure wave reflection and thus partially counterbalance the expected increase in aortic AIx and systolic BP due to heart rate deceleration. These data are in keeping with a previous study, in which carvedilol induced a lower heart rate-adjusted AIx and a lower reduction in pulse pressure amplification than atenolol.²¹ In another randomized study, central BP and AIx were reduced to the same extent by the vasodilating β-blocker nebivolol and the angiotensin-II receptor blocker irbesartan.²⁶ The larger reduction in central systolic BP provided by vasodilating β-blockers may potentially have significant clinical implications, given the prognostic value of aortic BP.²⁷ In a direct comparison between nebivolol and metoprolol, Kampus et al²² found that nebivolol was more effective than metoprolol in reducing central BP and left ventricular wall thickness. In that study, the change in wall thickness was significantly correlated with central BP change. This suggests that β-blockers with vasodilating properties may offer advantages over conventional β-blockers in antihypertensive therapy; however, this remains to be tested in a larger trial.

Carotid–femoral PWV was not reduced by treatment despite a significant reduction in both brachial and central BP, and it was even slightly increased in the losartan arm. This finding may appear in contrast with the well-known principle that aortic PWV depends on distending pressure and with the previous literature evaluating the impact of losartan and other angiotensin-II receptor blockers on carotid–femoral PWV.^{28, 29, 30} Some technical differences between the present and other studies may contribute to explain this discrepancy. In the present study, for the evaluation of carotid–femoral PWV, the distance was calculated according to regression equations based on height, gender and age in accordance with the data provided by the Korean Research Institute of Standard and Science, but no references about validation of this method are provided in the text. The method may also not be easily applied to non-Asian populations. Moreover, we are not aware of any validation study of carotid–femoral PWV as determined by PP-1000 against either invasive standards or other established noninvasive devices.

Central BP or BP amplification: which target for treatment?

It is worth remembering that the effects of antihypertensive drugs on central (systolic or pulse) pressure can be expressed either as absolute central BP values (in mm Hg) or as relative BP changes between the brachial and the aortic level, that is, changes in central-to-peripheral BP amplification. In turn, amplification of systolic or pulse pressure from the aorta to the peripheral arteries can be calculated in absolute (peripheral–central BP, in mm Hg) or relative (peripheral–central BP) terms. Similarly, when trying to compare results from different studies evaluating the effect of drugs on central BP, it should be considered that several methodological errors may limit the ability of noninvasive devices to precisely estimate central BP (Table 1). The prospective evaluation of the relative changes in BP amplification during treatment may represent a reasonable alternative to compensate for some of the above these errors (when the same device is used during a prospective evaluation), and could be potentially useful to compare results from different studies to observe the possible different effect on central BP between antihypertensive drug classes. Although Kim et al.²³ do not express the effects of treatment in terms of BP amplification, it can be estimated from their data that systolic BP amplification was affected only minimally and to a remarkably similar extent by losartan (−0.8 mm Hg) and carvedilol (−1.2 mm Hg). This further reinforces the suggestion that carvedilol might not share the same adverse effects on central hemodynamics as other older generation β-blockers. Although the impact of a given treatment on central BP and on BP amplification are obviously correlated with each other, it remains to be clarified which of the above approaches better reflects the potential effect of central hemodynamics on cardiovascular risk and the differential impact of drugs.

Table 1 Methodological errors inherent in the noninvasive estimation of central BP

Results from clinical trials reporting the effects of different antihypertensive regimens on peripheral and central hemodynamics may represent a meaningful surrogate end point for their long-term consequences in terms of morbidity and mortality, provided that well-established and validated methods for the noninvasive estimation are used, and that variability linked to measurement errors is reduced as much as possible. The time seems ripe to seek robust answers from large-scale prospective clinical studies centered on central BP with the aim of reinforcing the current knowledge and providing answers to open questions.