Clinical Trials. Pierre Boutouyrie, Assya Achouba, Patrick Trunet, Stéphane Laurent, for the EXPLOR Trialist Group

Transcription

1 Clinical Trials Amlodipine-Valsartan Combination Decreases Central Systolic Blood Pressure More Effectively Than the Amlodipine-Atenolol Combination The EXPLOR Study Pierre Boutouyrie, Assya Achouba, Patrick Trunet, Stéphane Laurent, for the EXPLOR Trialist Group Abstract The -blocker atenolol is less effective than angiotensin-receptor blockers and calcium-channel blockers for reducing central blood pressure (BP). The trial was designed to determine whether the advantages of angiotensinreceptor blockers over atenolol remained significant when both were combined with the calcium-channel blocker amlodipine. A prospective, randomized, blinded endpoint (PROBE design) parallel group, multicenter trial including 393 patients with essential hypertension resistant to 4 weeks of 5 mg of amlodipine was set out. Central systolic BP, augmentation index (AIx; either rough or adjusted on heart rate), and carotid-to-femoral pulse wave velocity were measured with applanation tonometry (SphygmoCor) at inclusion and after 8 and 24 weeks of active treatment with an amlodipine-valsartan combination (5/80 mg and then 10/160 mg) or an amlodipine-atenolol combination (5/50 mg and then 10/100 mg). From baseline to week 24, central systolic BP decreased significantly more in the amlodipine-valsartan group ( mm Hg; P ) than in the amlodipine-atenolol group ( mm Hg; P ; difference: 4.00 mm Hg [95% CI: 7.10 to 0.90]; P 0.013), despite similar changes in brachial systolic BP. The difference in rough AIx reduction was 6.5% (95% CI: 8.3 to 4.7; P ) in favor of amlodipine-valsartan. AIx adjusted on heart rate was significantly reduced in favor of amlodipine-valsartan ( 2.8% [95% CI: 4.92 to 0.68]; P 0.01). Heart rate decreased significantly more with amlodipine-atenolol (difference: 11 bpm [95% CI: 14 to 8 bpm]; P 0.001). Pulse wave velocity decreased by 0.95 m/s in both groups with no significant difference. Differences in central systolic BP and rough AIx remained significant after adjustment to the changes in heart rate. The amlodipine-valsartan combination decreased central (systolic and pulse) pressure and AIx more than the amlodipineatenolol combination. (Hypertension. 2010;55: ) Key Words: hypertension blood pressure aorta arteries antihypertensive agents compliance randomized, controlled trial Central aortic systolic and pulse pressures are independent predictors of cardiovascular events in various populations. 1 Antihypertensive agents differ in their ability to lower central aortic blood pressure (BP), despite similar reduction in brachial BP. 1,2 Angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), and calcium channel blockers (CCBs), which are powerful vasodilators, have been shown to improve wave reflection and central aortic pressure. 3 5 By contrast, the -blocker atenolol, prescribed alone 6,7 or in combination with thiazides, 8 is less effective than ACEIs, ARBs, and CCBs for lowering central pressure and wave reflection. It is now well accepted that antihypertensive therapy based on brachial artery recordings may overestimate the effect of -blocking drugs on central aortic systolic BP (SBP) and underestimate those of ACEIs, ARBs, and CCBs. 9,10 A common explanation for the lesser effect of -blockers on central BP is the bradycardia-induced dyssynchrony or uncoupling between outgoing and reflected waves, increasing central systolic and pulse pressures. 11 In addition, atenolol does not reduce total peripheral resistance and sympathetic drive and fails to induce a long-term remodeling of large 6,7 and small arteries, 12 which is required for the functional and structural improvement of arterial stiffness and resistance and the consequent reduction in augmentation index (AIx) and central aortic BP. 10,13 Although the effects of antihypertensive agents on central BP have been studied in several clinical trials as monothera- Received December 11, 2009; first decision December 26, 2009; revision accepted March 31, From the Université Paris-Descartes (P.B., S.L.), Paris, France; Institut National de la Santé et de la Recherche Médicale U970 (P.B., S.L.), Paris, France; Assistance Publique-Hôpitaux de Paris (P.B., S.L.), Hôpital Européen Georges Pompidou, Paris, France; Novartis-Pharma (A.A., P.T.), Paris, France. Correspondence to Pierre Boutouyrie, Department of Pharmacology, INSERM U970 (team 7), Hôpital Européen Georges Pompidou, Assistance Publique-Hôpitaux de Paris, Université Paris Descartes, 20 rue Leblanc, Paris, France. pierre.boutouyrie@egp.aphp.fr 2010 American Heart Association, Inc. Hypertension is available at DOI: /HYPERTENSIONAHA

2 Boutouyrie et al Central BP After Amlodipine-Valsartan 1315 Screened patients n = 473 Randomized patients N = 393 Excluded patients n = 80 Not fulfilling inclusion criterions N = 55 Withdrawal of consent: N = 16 Adverse event or lost to follow-up N = 13 Amlodipine-Valsartan N = 193 Amlodipine-Atenolol N = 200 Early termination: n=24 Adverse event/side effect: n=18 Treatment failure: n=2 Consent withdrawal: n=3 other: n=1 Early termination: n=38 Adverse event/side effect: n=28 Treatment failure: n=1 Consent withdrawal: n=3 Other: n=6 Figure 1. Study flowchart. pies, 2 5,11,14,15 very few long-term randomized, controlled clinical trials have studied drug strategies and combination therapies. The REASON Study compared an ACEI/diuretic combination with atenolol as monotherapy but not with atenolol included in a combination. 6,7 In the Conduit Artery Functional Endpoint Study, 8 an ancillary study of the Anglo- Scandinavian Cardiac Outcomes Trial, 16 an ACEI/CCB combination was compared with atenolol combined with a diuretic. An important unanswered question is whether the lesser effect of atenolol on central aortic BP is dampened or reversed when combined with CCBs or whether it remains significant despite associated vasodilatation. Different mechanisms with opposite effects may be surmised, and the relative weight of each is unknown. Atenolol-mediated bradycardia may persist despite CCB-induced baroreflex activation. On the other hand, baroreflex-induced vasoconstriction in response to atenolol might not be attenuated by CCBs. Thus, both bradycardia and increased wave reflection may increase central aortic BP in response to atenolol despite its combination with a CCB. By contrast, associating an ARB with a CCB can potentiate their vasodilating effects and further reduce central BP. Our working hypothesis was that the ARB valsartan would be more effective than atenolol on central aortic BP, when both drugs are combined with amlodipine. Thus, the EXPLOR Trial aimed at comparing the effects of amlodipine-valsartan with those of amlodipine-atenolol on aortic SBP in hypertensive patients under prospective, randomized, blinded endpoint (PROBE), parallel group conditions. Patients were treated for 6 months to increase the possibilities of unmasking a difference between the 2 drug regimens. In addition, to better analyze the pharmacodynamics of both combinations, we measured central aortic SBP, diastolic BP (DBP), pulse pressure (PP), AIx, and carotidfemoral pulse wave velocity (PWV). Methods Study Design This study was a national, multicenter, randomized, probe-type (prospective, randomized, open-label, blinded end point) trial (NCT ) with parallel groups versus the reference medication. It was investigator initiated and driven and sponsored by Novartis Pharma. Patients fulfilling inclusion criteria underwent a 2-week washout period before entering a 4-week open-label run-in period with amlodipine 5 mg (once daily). At the end of this period, patients whose SBP and/or DBP were not adequately controlled (defined by SBP 140 mm Hg and/or DBP 90 mm Hg and SBP 130 mm Hg and/or DBP 80 mm Hg in the case of diabetes mellitus or renal insufficiency) were randomized to amlodipine-valsartan 5/80 mg or amlodipine-atenolol 5/50 mg (once daily) treatment groups for a period of 8 weeks. After 8 weeks, patients were force titrated to either amlodipine-valsartan 10/160 mg or amlodipine-atenolol 10/ 100 mg for an additional 16-week period and up to completion of the study. Patients not able to tolerate forced titration were withdrawn from the study (amlodipine valsartan, n 18; amlodipine-atenolol, n 28; Figure 1). All of the patients took the scheduled dosage. All of the patients who terminated the study early were invited to undergo an investigation with applanation tonometry within 48 hours of study termination. Patients Between January 11, 2008, and January 15, 2009, general practitioners from 100 general practitioner centers connected with 10 tonometry centers across France were asked to recruit all of their consenting consecutive patients aged 18 to 75 years with essential hypertension resistant to 2 drugs belonging to 2 different pharmacological classes. After a 2-week washout period followed by a 4-week open-label run-in period with amlodipine 5 mg, patients still presenting an uncontrolled office BP (defined as SBP 140 mm Hg or DPB 90 mm Hg) were selected. Exclusion criteria included the following: BP controlled by 5 mg of amlodipine, contraindication to one of the drugs used in the protocol, women of childbearing

3 1316 Hypertension June 2010 Table 1. Baseline Population Characteristics Parameters Amlodipine-Valsartan Amlodipine-Atenolol P Sex, male/female, n (%) 111 (57)/82 (43) 99 (50)/101 (50) 0.11 Age, mean SD, y Weight, mean SD, kg Height, mean SD, cm Body mass index, mean SD, kg/m Duration of hypertension, mean SD, y Smoking, current/past or never, n (%) 42 (22)/151 (78) 56 (28)/143 (72) 0.14 Dyslipidemia, yes/no, n (%) 54 (28)/139 (72) 57 (29)/142 (71) 0.88 Diabetes mellitus, yes/no, n (%) 24 (12)/169 (88) 32 (16)/167 (84) 0.30 Family history of premature CVD, yes/no, n (%) 21 (11)/172 (89) 21 (11)/178 (89) 0.91 Abdominal obesity, yes/no, n (%) 72 (37)/121 (63) 86 (43)/113 (57) 0.23 Previous cardiovascular or renal disease, yes/no, n (%) 17 (9)/176 (91) 21 (11)/178 (89) 0.47 High cardiovascular risk, yes/no, n (%) 89 (46)/104 (54) 99 (50)/100 (50) 0.83 Previous antihypertensive treatments, yes/no, n (%) 107 (55)/86 (45) 113 (56)/87 (44) 0.85 ACEI or ARB, n (%) 77 (40) 76 (38) 0.72 CA, n (%) 16 (8) 21 (11) Blockers, n (%) 12 (6) 12 (6) 0.92 Diuretics, n (%) 43 (22) 29 (15) 0.05 Others, n (%) 1 (1) 2 (1) Office blood pressure at screening, mean SD, mm Hg SBP DBP PP Heart rate, mean SD, bpm CVD indicates cardiovascular disease; CA, calcium channel blockers. potential not using effective contraception, any active chronic disease, and SBP 180 mm Hg or DBP 110 mm Hg after the run-in period. The protocol was approved by the ethics committee of Saint-Germain-en-Laye Hospital in France. All of the patients gave their informed written consent. Central Pressure and Arterial Stiffness Measurement BP was measured using an Omron 705 C oscillometric device, both at the general practitioner s office and the tonometry center. Three measurements were performed, and the average of measures 2 and 3 was retained. Applanation tonometry was performed using a Sphygmo- Cor device (Atcor), as described previously 5 and recommended. 9,17 Briefly, the applanation probe is positioned on the radial artery (right arm), and optimal applanation is obtained using visual inspection and following built-in quality control indices. Radial waveforms are calibrated using brachial SBP and DBP measured before and after applanation (average). The central aortic waveform is calculated by the device software using the generalized transfer function. BP values are derived from the curve. AIx is measured, and AIx at heart rate 75 (AIx@75) is calculated through the software. The procedure is repeated at the level of the carotid artery, and the calibration is made using the DBP and the mean BP obtained from the radial tracings. 5,9 Applanation is then performed immediately afterward on the femoral artery, pulse transit times from concomitant ECG are calculated using the intersecting tangent, and PWV is calculated from the direct (carotid-to-femoral) path length. Certification of Centers and Quality Control and Blinding Procedures The description of procedures can be found in the online Data Supplement (available at Statistical Methods The number of patients to be included in the study was calculated on the basis of a 13-mm Hg SD for the fall in central SBP between inclusion and week 24 according to a previous study. 5 The power was set at 80% using a 2-sided risk of 5%. We estimated that 334 patients were necessary to demonstrate a difference of 4 mm Hg between the 2 groups for the reduction in central SBP and a difference in the reduction in brachial SBP 2 mm Hg. The estimated dropout rate after randomization (16%) lead us to include 398 patients. Statistics were performed using the SAS 9.01 package and NCSS 2004 software. All of the analyses were performed on the intent-totreat population. Data are presented as mean SD, unless otherwise specified. Changes are expressed as W24 minus baseline (negative means decrease) and their 95% CIs. Baseline (W0) characteristics were compared using unpaired Student t tests and 2 tests, as appropriate. The main comparisons were performed using a mixed model including time as the within factor and drug as the between factor. A center effect was systematically included. We used multivariate robust models to determine the changes in arterial parameters, after adjustment to initial values (W0), potential confounders (eg, changes in heart rate, age, and sex), and drug treatment as dummy variables. All of the tests were bilateral using Results We initially screened 473 patients because the rate of dropout was lower than expected (Figure 1). Eighty patients (16%) were excluded at the end of the washout and 5-mg amlodipine run-in period, including 55 patients (11%) adequately controlled with 5 mg of amlodipine. After randomization, the 2 populations were of comparable age and BP. Forty-six percent of patients were at high cardiovascular risk, as predicted from the frequency of associated risk factors (Table 1). The 2 groups were matched for all of the important cardiovascular risk factors.

4 Boutouyrie et al Central BP After Amlodipine-Valsartan 1317 Office blood pressure (mmhg) Amlodipine 5 Amlodipine 5/Valsartan 80 or Amlodipine 5/Atenolol 50 Amlodipine 10/Valsartan 160 or Amlodipine 10/Atenolol Time (weeks) Figure 2. Brachial BP in response to amlodipine-valsartan (red) and amlodipine-atenolol (blue). Values were obtained at a general practitioner office using an Omron 705C device. At week 24, the decreases in brachial BP either at office (Figure 2) or at tonometry center (Figure 3 and Table 2) were not significantly different between the groups. Drug tolerance was good (please see the online Data Supplement). At the tonometry center, brachial SBP decreased by and Systolic pressure (mmhg) Brachial SBP Mean difference [-4.28 to 1.99] p=ns mm Hg in the amlodipine-atenolol and amlodipinevalsartan groups, respectively, with no significant difference between the groups (1.14 mm Hg [95% CI: 4.28 to 1.99 mm Hg]; P 0.47; Table 2 and Figures 2 and 3). This value is lower than the predetermined level of 2-mm Hg equivalence. Brachial DBP was reduced to a lesser extent, down by 7.9 mm Hg with both treatments. The difference between the drug regimens was close to 0 (Table 2). The fall in brachial BP was time and dose dependent (Figure 2). Most of the decrease in brachial BP was achieved with the low dosage, but the higher dosage added a further significant decrease of 2.6 mm Hg for brachial SBP and 2.0 mm Hg for brachial DBP (P 0.05). The rates of BP control after the valsartan-amlodipine and amlodipine-atenolol combinations were similar after 8 weeks (47% and 43%, respectively) and remained similar at 24 weeks (56% and 59%, respectively). As expected, heart rate was more reduced in the amlodipineatenolol group than in the amlodipine-valsartan group 9.4 bpm (95% CI: [ 11.6 to 7.1 bpm]; P 0.001; Table 2). By contrast to brachial SBP, aortic SBP decreased more after amlodipine-valsartan than after amlodipine-atenolol (Table 2). The difference between the groups reached the prespecified threshold of 4 mm Hg ( 3.95 mm Hg [95% CI: 7.08 to 0.83 mm Hg]) and was significant (P 0.02). Aortic PP decreased more after amlodipine-valsartan than after amlodipine-atenolol (Table 2), with a significant difference between the groups ( 3.74 mm Hg [95% CI: 5.33 to 2.15 mm Hg]; P 0.001). By contrast, no significant difference in the fall in aortic DBP was observed between the groups (Table 2). The AIx significantly decreased after amlodipine-valsartan, whereas it significantly increased after amlodipine-atenolol (Table 3 and Figure 4), with a significant difference between the groups ( 6.50% [95% CI: 8.28% to 4.72%]; P 0.001). After adjustment of the AIx to Pulse pressure (mmhg) Brachial PP Mean difference [-3.33 to 0.6] p=ns Aortic SBP Mean difference [-7.08 to -0.83], p=0.02 Baseline W8 W Aortic PP Mean Difference -3.74[-5.33 to -2.15] p<0.001 Baseline Figure 3. Brachial (dotted lines) and aortic BPs (continuous lines) in response to amlodipine-valsartan (red) and amlodipine-atenolol (blue). Values were obtained at a tonometry center using an Omron 705C device. W8 W24

5 1318 Hypertension June 2010 Table 2. BP Response to Amlodipine-Valsartan or Atenolol-Amlodipine Baseline Change vs Baseline Amlodipine-Valsartan, Amlodipine-Atenolol, Amlodipine-Valsartan, Amlodipine-Atenolol, Difference Parameters Mean SD Mean SD Mean SD Mean SD Mean 95% CI P Aortic parameters SBP, mm Hg to DBP, mm Hg to PP, mm Hg to AIx, % to AIx@75, % to PP amplification to CF-PWV, m/s to Brachial BP SBP, mm Hg to Mean BP, mm Hg to DBP, mm Hg to PP, mm Hg to HR, bpm to Differences are calculated as amlodipine-valsartan minus amlodipine-atenolol. CF indicates carotid-femoral; HR, heart rate. heart rate, both drug regimens significantly reduced this parameter, and the difference in AIx@75 between the groups remained significant ( 2.80% [95% CI: 4.92% to 0.68%]; P 0.01). Amplification of SBP was enhanced after amlodipine-valsartan (2.03% [95% CI: 1.42% to 2.65%]; P 0.001). Results obtained at the common carotid level paralleled those observed at the aortic level but with lower probability values (except for the AIx; please see Table S1 in the online Data Supplement). Both amlodipinevalsartan and amlodipine-atenolol regimens decreased aortic PWV by 1 m/s; however, no significant difference between drugs was observed ( 0.02 m/s [95% CI: 0.46 to 0.41]; P 0.92; Figure 4). Changes were slightly different after 8 weeks of treatment only. Indeed, central aortic SBP was nonsignificantly decreased ( 2.62 mm Hg [95% CI: 5.81 to 0.58]; P 0.11), whereas the difference in central aortic PP was already significant ( 3.98 mm Hg [95% CI: 5.76 to 2.21 mm Hg]; P ). The raw AIx was significantly decreased with amlodipinevalsartan compared with amlodipine-atenolol ( 5.83% [95% CI: 7.79% to 3.87%] P ), although this difference was no longer significant after correction for heart rate changes. At week 24, the between-group differences in the reduction of aortic SBP, PP, and AIx remained significant after adjustment to baseline values, changes in heart rate, and other significant covariates in a multivariate regression analysis (Table 3). Indeed, although the reduction in aortic SBP, PP, and AIx were negatively correlated with the changes in heart rate (the larger the reduction in heart rate, the lower the reduction in aortic PP and AIx), the reduction in aortic SBP, PP, and AIx remained significantly larger with amlodipinevalsartan than with amlodipine-atenolol. Patients previously treated had similar changes as patients previously untreated. Additional adjustment on changes in mean BP did not alter these results. After adjustment on baseline value, the reduction in carotid-femoral PWV was positively associated with the reduction in aortic SBP and heart rate and was larger in women; treatment was not significantly associated with changes in carotid-femoral PWV. Results were similar if the final value of parameter was used instead of changes in parameter (Table S2). Discussion The major finding of the present study is that, despite a similar reduction in brachial BP in both groups, aortic SBP, PP, AIx, and AIx@75 were more reduced in the amlodipinevalsartan group than in the amlodipine-atenolol group. These results were observed with similar and considerable reductions in brachial BP with both combinations and remained significant after adjustment to heart rate reduction. ARBs, Atenolol, and Reduction of Central Aortic Pressure The amlodipine-valsartan combination lowered central aortic SBP and PP and the AIx more than the amlodipine-atenolol combination. By contrast, no significant difference between the groups was observed with regard to aortic stiffness and brachial BP. The ability of amlodipine-atenolol to lower PWV to a similar extent as amlodipine-valsartan is not surprising, because both regimens had similar effects on mean BP, which represents the main load exerted on the stiff material of the aortic wall, which, in turn, determines aortic stiffness. In a multivariate model predicting the regression of PWV in the present study, we observed that baseline PWV value, mean BP reduction, and heart rate reduction were the 3 major determinants, explaining 19.0%, 8.0%, and 4.3% of variance, respectively. This confirms the prominent role of mean BP reduction on aortic stiffness changes. The similar reductions in aortic stiffness by both drug regimens are comparable to previous observations in the REASON 6,7 and Conduit Artery Functional Endpoint 8 studies.

6 Boutouyrie et al Central BP After Amlodipine-Valsartan 1319 Table 3. Determinants of Changes in BP and Carotid-Femoral PWV Between Week 24 and Inclusion Parameters In/Out R 2 increment, % -Coefficient Lower CI Upper CI P Change in aortic systolic pressure (R , RMSE 11.4) Baseline aortic systolic In pressure, mm Hg Age at inclusion, y In Valsartan/atenolol In Change in heart rate, bpm Out 0.90 Sex Out 0.95 BMI Out 0.85 Change in aortic PP (R , RMSE 6.6) Baseline aortic PP, mm Hg In Change in heart rate, bpm In Age at inclusion, y In Valsartan/atenolol In Sex Out 0.81 BMI Out 0.89 Change in aortic AIx (R , RMSE 6.4) Baseline AIx, % In Change in heart rate, bpm In Sex In Valsartan/atenolol In Age at inclusion, y In BMI Out 0.75 Change in aortic PWV (R , RMSE 1.87) Baseline PWV, m/s In Change in aortic SBP, mm Hg In Change in heart rate, bpm In Sex In Valsartan/atenolol Out Age at inclusion, y Out In indicates entered the model; Out, did not enter the model; RMSE, root mean square error. Changes are calculated as final minus baseline. A negative -coefficient for baseline SBP means a larger decrease in the more hypertensive subjects; atenolol and valsartan are coded 0 and 1, men and women 0 and 1, respectively. There was no significant difference in the reduction of brachial SBP or PP between the 2 regimens despite a significant difference in central aortic SBP or PP. This was also expected. Central pressure is lower than peripheral pressure, which is consistent with the amplification phenomenon according to which the amplitude of the pressure wave is higher in peripheral arteries than in central arteries 18 (Table 2). After treatment, amplification was enhanced by valsartan compared with atenolol. Drugs exert differential effects on the PP amplification between central and peripheral arteries. 7,8 Altogether these findings indicate that the addition of CCBs to atenolol did not abolish the adverse effect of atenolol on central BP 10,11,13 and suggest that it occurred either through bradycardia or increased wave reflection. The bradycardia-induced dyssynchrony or uncoupling between outgoing and reflected waves can by itself increase central SBP and PP. 11,19 Indeed, the reduction in heart rate allows more time during systole for early return of the reflected wave. Bradycardia is also associated with an increased ejection volume, which translates into an increased peak SBP. A recent analysis of the Conduit Artery Functional Endpoint Trial suggests that most of the differences between the treatment groups were attributable to changes in heart rate. 20 This is indeed not the case in our study, because adjustment to the change in heart rate only marginally lessened the improvement in central BP and AIx with amlodipine-valsartan. In the present study, heart rate was 10 bpm lower with amlodipine-atenolol than with amlodipinevalsartan. We checked whether differences in central aortic BP and AIx between the drug regimens could be attenuated by adjustment to heart rate. This was done by building a multivariate model with changes in aortic SBP, PP, or AIx as

7 1320 Hypertension June 2010 Carotid to femoral pulse wave velocity (m/s) Aortic augmentation index (%) Mean difference at W24 : m/s [-0.46 to 0.41], p=ns 10 Mean difference at W % [-8.28 to -4.72], p< Baseline W8 W24 Baseline W8 W24 Figure 4. Changes in PWV and AIx in response to amlodipine-valsartan (red) and amlodipine-atenolol (blue). dependent variables and changes in mean BP, heart rate, baseline values of central pressure parameters, and treatment as independent variables. We showed that part of the difference in aortic PP reduction between the 2 treatment arms was significantly independent of the changes in heart rate. Indeed, in a fully adjusted model for aortic PP changes (Table 3), treatment with amlodipine-valsartan (versus amlodipineatenolol) accounted for 2.65 mm Hg ( -coefficient, P 0.001) of the total 3.74 mm Hg (Table 2), whereas a 10-bpm reduction in heart rate explained only 1.50 mm Hg ( -coefficient, P 0.001). In the present study, AIx significantly decreased with amlodipine-valsartan, whereas it significantly increased after amlodipine-atenolol, with a significant difference between groups (Table 3 and Figure 3). The influence of heart rate was analyzed in 2 ways. First, we adjusted AIx to heart rate and calculated AIx@75. The difference in AIx@75 reduction between the groups remained significant, although it was half that of unadjusted AIx ( 2.80% versus 6.50%, respectively; Table 2) Second, in a fully adjusted model for aortic AIx (Table 3), treatment with amlodipine-valsartan (versus amlodipine-atenolol) explained a large part of the difference, that is, 4.40% ( -coefficient; P 0.001) out of 6.50% (Table 2), whereas a 10-bpm reduction in heart rate only explained 2.50% ( -coefficient, P 0.001). Because bradycardia explained only part of the difference in central pressures and AIx between the 2 drug regimens, other mechanisms need to be discussed, particularly changes in reflection sites. Vasodilatation, Arterial Remodeling, and Wave Reflection Reduction in reflection sites intensity can be obtained pharmacologically either through vasodilatation or long-lasting structural remodeling, that is, decreased wall:lumen ratio of small arteries. Although atenolol is known to prevent the deleterious effects of catecholamines on the heart, it does not reduce total peripheral resistance and sympathetic drive 21 ;it is less effective than blockers of the renin-angiotensin system to reduce small artery damage, that is, vasoconstriction and increased media:lumen ratio 12 ; and it is less effective than vasodilators in reducing aortic and carotid stiffness, carotid intima-media thickness, and cerebrovascular resistance in hypertensive patients. 22 That a significant difference in aortic PP and AIx between the 2 regimens was already observed after 8 weeks in the present study suggests a difference in vasomotor tone rather than in structural changes. Indeed, the pharmacological remodeling of small arteries leading to a reduction in wall:lumen ratio is a long-lasting process requiring several months Thus, the difference in central pressure and AIx between the 2 regimens suggests that the amplitude of vasodilatation and, therefore, the reduction of wave reflection were higher when an ARB was combined with a CCB than when the -blocker atenolol was combined with a CCB. A recent report mentioned a better improvement of PWV and central pressure with CCB combined with ARB than with diuretics 26 ; however, the difference in central pressure was of the same magnitude as changes in peripheral pressure. By contrast, in the EXPLOR Trial, differences in central SBP were larger than those in brachial SBP, DBP, and mean BP. Furthermore, they were independent of baseline value, changes in mean BP, and heart rate. The same holds true for AIx. Both bradycardia and increased wave reflection may have increased central aortic BP and AIx in response to atenolol despite its combination with a CCB. This may reflect the persistence of a relative vasoconstriction and bradycardia with atenolol despite chronic treatment and association with potent vasodilatators, such as amlodipine. These results are unlikely to apply to vasodilating -blockers. Indeed, vasodi-

8 Boutouyrie et al Central BP After Amlodipine-Valsartan 1321 lating -blockers, including celiprolol, 27 dilevalol, 28 and nebivolol, 11,29 have been reported to reduce central PP and AIx, and it is very likely that combining them with a CCB will reinforce their beneficial effects. Increased heart rate has been associated with increased arterial stiffness, mainly because of arterial wall viscosity phenomenon. 30 Therefore, bradycardia and reduction in cardiac output with -blockers may explain the similar reduction in aortic PWV as with vasodilating drugs and may also explain part of their beneficial effect. Reduction of Central BP and Outcomes Meta-analyses of hypertension treatment trials have confirmed the lower impact of -blocker based therapies in preventing stroke, particularly with atenolol, and showed that the risk of myocardial infarction was not significantly different. 31 In particular, the Losartan Intervention for Endpoint Reduction in Hypertension Study 32 and Anglo-Scandinavian Cardiac Outcomes Trial 16 showed that losartan- and amlodipine-based treatments, respectively, proved to be more effective than atenolol-based treatments in reducing the incidence of stroke. Because brachial BP was lowered to a similar extent on both arms, and because central aortic BP is an independent predictor of cardiovascular events in various populations, 1 it has been suggested that the difference in outcomes could be attributed to the difference in the reduction of central aortic BP. 1,8,10 Although a differential effect on central BP has been reported in the Conduit Artery Function Evaluation Study, in favor of the amlodipine-perindopril combination, the lack of measurement of central pressure at the initiation of the Anglo-Scandinavian Cardiac Outcomes Trial limits the validity of conclusions. Methodological Issues The present study is, to our knowledge, the first one to provide data on the long-term effects of -blockers in combination therapies on central BP. Previous studies could not address this issue, because the drugs to be compared were not combined with the same antihypertensive agent. For instance, the Conduit Artery Function Evaluation Study 8 compared an ACEI/CCB combination with atenolol combined with a diuretic. The REASON study 7 compared an ACEI/diuretic combination with atenolol as monotherapy but not with atenolol included in a combination. This is why the present study was designed with both parallel groups, including amlodipine. Although the PROBE design has been successfully used in the past, this was mainly in clinical trials with hard end points like in the Hypertension Optimal Treatment Study. 33 The true double-blind nature of trials involving -blockers has been questioned before because of the evident reduction in heart rate. In the present study, tonometry centers in charge of measuring the central BP were not responsible for delivering the drug to the patients and were asked (by procedures) not to enquire about drugs that patients were receiving. We also ensured that the core laboratory was never in direct contact with any of the investigators (general practitioners or tonometry centers). All of the feedback information from the core laboratory to the investigator transited through the contract research organization executive. As detailed above, tracings were blinded as to the center, the period, and the identity of the patient, as well as to all of the queries to the core laboratory from investigators or the CRO. Thus, the end point was truly blinded, according to the PROBE design. AIx is not a pure index of wave reflection, and proving that the differential effect of atenolol and valsartan, added to amlodipine, is attributed to a differential effect on wave reflection would need a concomitant measurement of aortic blood flow to separate forward and backward waves with more precision. 34 Radial artery tonometry calibrated on brachial pressure leads to imperfect calibration. 35 However, this was similarly done in both groups, and it is unlikely to influence the differential effect of drugs on central pressure. In conclusion, the amlodipine-valsartan combination improves central systolic and PP, together with AIx (a surrogate measure of wave reflection), more than the amlodipine-atenolol combination. Clinical Perspectives Because central aortic SBPs and PPs are independent predictors of cardiovascular events in various populations, it is important to determine to what extent antihypertensive agents differ in their ability to lower central aortic BP, despite similar reduction in brachial BP. The present study is, to our knowledge, the first one to provide data on the long-term effects of -blockers in combination therapies on central BP. We have demonstrated that a 24-week treatment with an amlodipine-valsartan combination improved central SBP and PPs, together with wave reflection, more than the amlodipineatenolol combination and that differences were independent of the reduction in heart rate. The lesser effects of -blockers on outcomes in hypertension have been put in evidence by recent meta-analyses; the present study strongly suggests that, even when combined with a CCB, atenolol might not reduce central BP enough to effectively protect against cardiovascular events. An outcome trial is required to determine whether the differences in central BP between the treatment groups translate into clinical benefit. Appendix The EXPLOR Trialist Group consists of the following individuals: P. Gosse, Bordeaux; G. London and B. Pannier, Fleury-Mérogis; P. Poncelet, Henin-Beaumont; P. Lantelme and L. Legedz, Lyon; P. Fesler, Montpellier; B. Levy, Paris; C. Thuillez and R. Joannides, Rouen; D. Stephan, Strasbourg; P. Laurent, Toulon; and G. Doll, Tours. Acknowledgments We warmly thank Fatima Hamza and Julie Perucca for their excellent technical work, as well as Erwan Bozec, Kim Thanh Ong, and Cédric Collin for their help during the training sessions. Sources of Funding This investigator-initiated study was funded by Novartis Pharma. Disclosures A.A. and P.T. are employees of Novartis-Pharma France. P.B. has received honorariums from Novartis Pharma for scientific consultation.

9 1322 Hypertension June 2010 References 1. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, Pannier B, Vlachopoulos C, Wilkinson I, Struijker-Boudier H. Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J. 2006;27: Van Bortel LM, Struijker-Boudier HA, Safar ME. Pulse pressure, arterial stiffness, and drug treatment of hypertension. Hypertension. 2001;38: Mahmud A, Feely J. Favourable effects on arterial wave reflection and pulse pressure amplification of adding angiotensin II receptor blockade in resistant hypertension. J Hum Hypertens. 2000;14: Mahmud A. Reducing arterial stiffness and wave reflection: quest for the Holy Grail. Artery Research. 2007;1: Tropeano AI, Boutouyrie P, Pannier B, Joannides R, Balkestein E, Katsahian S, Laloux B, Thuillez C, Struijker-Boudier H, Laurent S. Brachial pressureindependent reduction in carotid stiffness after long-term angiotensinconverting enzyme inhibition in diabetic hypertensives. Hypertension. 2006; 48: Asmar RG, London GM, O Rourke ME, Safar ME. Improvement in blood pressure, arterial stiffness and wave reflections with a verylow-dose perindopril/indapamide combination in hypertensive patient: a comparison with atenolol. Hypertension. 2001;38: London GM, Asmar RG, O Rourke MF, Safar ME. Mechanism(s) of selective systolic blood pressure reduction after a low-dose combination of perindopril/indapamide in hypertensive subjects: comparison with atenolol. J Am Coll Cardiol. 2004;43: Williams B, Lacy PS, Thom SM, Cruickshank K, Stanton A, Collier D, Hughes AD, Thurston H, O Rourke M. 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: Laurent S, Tropeano AI, Boutouyrie P. Pulse pressure reduction and cardiovascular protection. J Hypertens. 2006;24(suppl 3):S13 S Laurent S, Briet M, Boutouyrie P. Large and small artery cross-talk and recent morbidity-mortality trials in hypertension. Hypertension. 2009;54: Dhakam Z, Yasmin, McEniery CM, Burton T, Brown MJ, Wilkinson IB. A comparison of atenolol and nebivolol in isolated systolic hypertension. J Hypertens. 2008;26: Schiffrin EL. Remodeling of resistance arteries in essential hypertension and effects of antihypertensive treatment. Am J Hypertens. 2004;17: Laurent S, Boutouyrie P. -Blocker-induced heart rate reduction: too simplistic to explain the deleterious effects of -blockers. J Am Coll Cardiol. 2009;53: Mahmud A, Feely J. Reduction in arterial stiffness with angiotensin II antagonist is comparable with and additive to ACE inhibition. Am J Hypertens. 2002;15: Morgan T, Lauri J, Bertram D, Anderson A. Effect of different antihypertensive drug classes on central aortic pressure. Am J Hypertens. 2004; 17: Dahlof B, Sever PS, Poulter NR, Wedel H, Beevers DG, Caulfield M, Collins R, Kjeldsen SE, Kristinsson A, McInnes GT, Mehlsen J, Nieminen M, O Brien E, Ostergren J. Prevention of cardiovascular events with an antihypertensive regimen of amlodipine adding perindopril as required versus atenolol adding bendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomised controlled trial. Lancet. 2005;366: Agabiti-Rosei E, Mancia G, O Rourke MF, Roman MJ, Safar ME, Smulyan H, Wang JG, Wilkinson IB, Williams B, Vlachopoulos C. Central blood pressure measurements and antihypertensive therapy: a consensus document. Hypertension. 2007;50: Avolio AP, Van Bortel LM, Boutouyrie P, Cockcroft JR, McEniery CM, Protogerou AD, Roman MJ, Safar ME, Segers P, Smulyan H. Role of pulse pressure amplification in arterial hypertension: experts opinion and review of the data. Hypertension. 2009;54: Nichols WW, McDonald DA. Wave-velocity in the proximal aorta. Med Biol Eng. 1972;10: Williams B, Lacy PS. Impact of heart rate on central aortic pressures and hemodynamics: analysis from the CAFE (Conduit Artery Function Evaluation) Study: CAFE-Heart Rate. J Am Coll Cardiol. 2009;54: Burns J, Mary DA, Mackintosh AF, Ball SG, Greenwood JP. Arterial pressure lowering effect of chronic atenolol therapy in hypertension and vasoconstrictor sympathetic drive. Hypertension. 2004;44: Ariff B, Zambanini A, Vamadeva S, Barratt D, Xu Y, Sever P, Stanton A, Hughes A, Thom S. Candesartan- and atenolol-based treatments induce different patterns of carotid artery and left ventricular remodeling in hypertension. Stroke. 2006;37: Thybo NK, Stephens N, Cooper A, Aalkjaer C, Heagerty AM, Mulvany MJ. Effect of antihypertensive treatment on small arteries of patients with previously untreated essential hypertension. Hypertension. 1995;25: Savoia C, Touyz RM, Endemann DH, Pu Q, Ko EA, De CC, Schiffrin EL. Angiotensin receptor blocker added to previous antihypertensive agents on arteries of diabetic hypertensive patients. Hypertension. 2006;48: Agabiti-Rosei E, Heagerty AM, Rizzoni D. Effects of antihypertensive treatment on small artery remodelling. J Hypertens. 2009;27: Matsui Y, Eguchi K, O Rourke MF, Ishikawa J, Miyashita H, Shimada K, Kario K. Differential effects between a calcium channel blocker and a diuretic when used in combination with angiotensin II receptor blocker on central aortic pressure in hypertensive patients. Hypertension. 2009;54: Boutouyrie P, Bussy C, Hayoz D, Hengstler J, Dartois N, Laloux B, Brunner H, Laurent S. Local pulse pressure and regression of arterial wall hypertrophy during long-term antihypertensive treatment. Circulation. 2000;101: Kelly R, Daley J, Avolio A, O Rourke M. Arterial dilation and reduced wave reflection: benefit of dilevalol in hypertension. Hypertension. 1989; 14: Mahmud A, Feely J. -Blockers reduce aortic stiffness in hypertension but nebivolol, not atenolol, reduces wave reflection. Am J Hypertens. 2008;21: Albaladejo P, Copie X, Boutouyrie P, Laloux B, Declere AD, Smulyan H, Benetos A. Heart rate, arterial stiffness, and wave reflections in paced patients. Hypertension. 2001;38: Lindholm LH, Carlberg B, Samuelsson O. Should blockers remain first choice in the treatment of primary hypertension? A meta-analysis. Lancet. 2005;366: Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U, Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H. Cardiovascular morbidity and mortality in the Losartan Intervention for Endpoint Reduction in Hypertension Study (LIFE): a randomised trial against atenolol. Lancet. 2002;359: Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S, Menard J, Rahn KH, Wedel H, Westerling S. Effects of intensive bloodpressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) Randomised Trial HOT Study Group. Lancet. 1998;351: Manisty CH, Zambanini A, Parker KH, Davies JE, Francis DP, Mayet J, McG Thom SA, Hughes AD. Differences in the magnitude of wave reflection account for differential effects of amlodipine- versus atenolol-based regimens on central blood pressure: an Anglo- Scandinavian Cardiac Outcome Trial substudy. Hypertension. 2009;54: Verbeke F, Segers P, Heireman S, Vanholder R, Verdonck P, Van Bortel LM. Noninvasive assessment of local pulse pressure: importance of brachial-to-radial pressure amplification. Hypertension. 2005;46: