Journal of Hypertension 2006, 24: a Centro Interuniversitario di Fisiologia Clinica e Ipertensione, University of Milan,

Transcription

1 Original article 2163 Outcomes in subgroups of hypertensive patients treated with regimens based on valsartan and amlodipine: an analysis of findings from the VALUE trial Alberto Zanchetti a, Stevo Julius b, Sverre Kjeldsen b,c, Gordon T. McInnes d, Tsushung Hua e, Michael Weber f, John H. Laragh g, Francis Plat h, Edouard Battegay i, Cesar Calvo-Vargas j, Andrzej Cieśliński k, Jean Paul Degaute l, Nicolaas J. Holwerda m, Janna Kobalava n, Ole Lederballe Pedersen o, Faustinus P. Rudyatmoko p, Kostas C. Siamopoulos q and Öyvind Störset r Background In the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial the primary outcome (cardiac morbidity and mortality) did not differ between valsartan and amlodipine-based treatment groups, although systolic blood pressure (SBP) and diastolic blood pressure reductions were significantly more pronounced with amlodipine. Stroke incidence was non-significantly, and myocardial infarction was significantly lower in the amlodipine-based regimen, whereas cardiac failure was non-significantly lower on valsartan. Objectives The study protocol specified additional analyses of the primary endpoint according to: sex; age; race; geographical region; smoking status; type 2 diabetes; total cholesterol; left ventricular hypertrophy; proteinuria; serum creatinine; a history of coronary heart disease; a history of stroke or transient ischemic attack; and a history of peripheral artery disease. Additional subgroups were isolated systolic hypertension and classes of antihypertensive agents used immediately before randomization. Methods The hypertensive patients participating in VALUE were divided into subgroups according to baseline characteristics. Treatment by subgroup interaction analyses were carried out by a Cox proportional hazard model. Within each subgroup, treatment effects were assessed by hazard ratios and 95 confidence intervals. Results For cardiac mortality and morbidity, the only significant subgroup by treatment interaction was of sex (P U 0.016), with the hazard ratio indicating a relative excess of cardiac events with valsartan treatment in women but not in men, but SBP differences in favour of amlodipine were distinctly greater in women. No other subgroup showed a significant difference in the composite cardiac outcome between valsartan and amlodipine-based treatments. For secondary endpoints, a sex-related significant interaction was found for heart failure (P < ), with men but not women having a lower incidence of heart failure with valsartan. Conclusion As in the whole VALUE cohort, in no subgroup of patients were there differences in the incidence of the composite cardiac endpoint with valsartan and amlodipinebased treatments, despite a greater blood pressure decrease in the amlodipine group. The only exception was sex, in which the amlodipine-based regimen was more effective than valsartan in women, but not in men, whereas the valsartan regimen was more effective in preventing cardiac failure in men than in women. J Hypertens 24: Q 2006 Lippincott Williams & Wilkins. Journal of Hypertension 2006, 24: Keywords: amlodipine, cardiac events, heart failure, sex, valsartan a Centro Interuniversitario di Fisiologia Clinica e Ipertensione, University of Milan, Istituto Auxologico Italiano and Ospedale Maggiore, Milan, Italy, b University of Michigan, Ann Arbor, Michigan, USA, c Ullevaal University Hospital, Oslo, Norway, d University of Glasgow, Glasgow, UK, e Novartis Pharma, East Hanover, New Jersey, f State University of New York, New York, g Cornell Medical Center, New York, New York, USA, h Novartis Pharma AG, Basel, i University Hospital, Basel, Switzerland, j Hospital Civil Juan I, Menchaca University of Guadalajara, Guadalajara, Mexico, k Karol Marcinkowski University, Poznaň, Poland, l Hôpital Erasme, Brussels, Belgium, m Sint Elisabeth Ziekenhuis, Tilburg, The Netherlands, n Russian University of People Friendship, Moscow, Russia, o Viborg Hospital, Viborg, Denmark, p J.I. Sutorejo-Utara F. 24, Surabaya, Indonesia, q University of Ioannina, Ioannina, Greece and r Akershus University Hospital, Nordbyhagen, Norway Correspondence and requests for reprints to Prof. Alberto Zanchetti, Centro di Fisiologia Clinica e Ipertensione, Via F. Sforza, 35, Milano, Italy alberto.zanchetti@unimi.it Received 30 March 2006 Accepted 27 June 2006 Introduction The Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial was designed to test the hypothesis that, for the same blood pressure control, the angiotensin receptor blocker, valsartan, would reduce cardiac morbidity and mortality (primary endpoint) more than the calcium antagonist, amlodipine, in hypertensive patients at high cardiovascular risk. The principal results of the trial have recently been reported [1]. In brief, after a mean follow-up of 4.2 years with randomized treatment, 1599 out of hypertensive patients at high risk of cardiovascular events experienced a primary endpoint ß 2006 Lippincott Williams & Wilkins

2 2164 Journal of Hypertension 2006, Vol 24 No 11 The primary outcome of cardiac mortality and morbidity did not differ between the two treatment groups, although systolic blood pressure (SBP) and diastolic blood pressure reduction was significantly more pronounced with the amlodipine-based regimen (by 2.2/1.6 mmhg throughout the trial, 4.0/2.1 mmhg after the first month). Among secondary outcomes, rates of stroke (non-significantly) and myocardial infarction (significantly) were lower in the amlodipine-based regimen group, whereas cardiac failure incidence was lower (non-significantly) in the valsartan-based regimen group. The study protocol [2] specified that additional analyses of the primary endpoint were to be performed for certain prespecified subgroups: male and female sexes; age less than 65 years and 65 years and older; race: black, Caucasian, oriental, and other; geographical region: Asia, Europe, Latin America, North America, and rest of the world; current smokers and never or ex-smokers; the presence and absence of type 2 diabetes mellitus; total cholesterol greater than 240 mg/dl (6.2 mmol/l), and 240 mg/dl or less; the presence and absence of left ventricular hypertrophy by electrocardiogram with or without strain pattern; the presence and absence of proteinuria (dipstick 1þ or higher); serum creatinine greater than 1.7 mg/dl (150 mmol/l) and 1.7 mg/dl or less; history and no history of coronary heart disease; history and no history of stroke or transient ischemic attack; and a history and no history of peripheral artery disease. Additional subgroup analyses, although not specifically indicated in the protocol, concerned isolated systolic hypertension (defined as SBP 160 mmhg and diastolic blood pressure < 90 mmhg) and classes of antihypertensive agents used immediately before randomization (beta-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, diuretics). Methods Study population and treatment A total of hypertensive patients participated in this randomized, double-blind, parallel-group comparison of therapy based on valsartan or amlodipine. The baseline characteristics of the randomized patients were described in the report of the principal results of VALUE [1], in which details can also be found about treatment procedures and follow-up (4 6 years, mean 4.2). The subgroups considered were those specified in the Introduction. Outcome measures The primary endpoint was time to first cardiac event (a composite of sudden cardiac death, fatal myocardial infarction, death during or after percutaneous coronary intervention or coronary artery bypass graft, death as a result of heart failure, and death associated with recent myocardial infarction at autopsy, heart failure requiring hospital management, non-fatal myocardial infarction, or emergency procedures to prevent myocardial infarction). An endpoint committee, blinded to therapy allocation, reviewed the clinical records of cardiovascular events reported by clinical centres, and adjudicated them according to protocol criteria [1]. As the occurrence of new heart failure is a relatively soft endpoint [2], only hospitalized heart failure was considered as an outcome, and in no case was the diagnosis of heart failure validated solely on the basis of ankle oedema [1]. The VALUE protocol also prespecified a number of subgroup analyses. As subgroup analyses have decreased power to detect treatment-related differences in outcome, these subgroup analyses focus on the primary endpoint, as indicated in the protocol [3], and the three key secondary endpoints of myocardial infarction, heart failure and stroke. Statistical methods The treatment (valsartan-initiated versus amlodipineinitiated) by subgroup interaction analyses included the treatment by subgroup interaction term in a Cox proportional hazard model, with age, the presence of coronary heart disease and the presence of left ventricular hypertrophy at baseline as covariates. Within each subgroup, treatment effects were assessed by hazard ratios (HR) and 95 confidence intervals (CI) based on the Cox regression model. For an assessment of statistical significance, two-tailed tests were used and significance was expressed as P values. A P value less than 0.05 was accepted as statistically significant for all analyses, except treatment by subgroup interactions, for which a P value less than 0.10 was accepted. Results Treatment subgroup interaction analyses for the primary endpoint The baseline characteristics of the patients in each subgroup will be presented in a separate paper, investigating the role of individual risk factors or diseases on the primary outcome in the VALUE cohort, independent of treatment assignment. For the primary endpoint (composite cardiac mortality and morbidity), the only significant subgroup by treatment interaction was for sex (P ¼ 0.016), with the HR indicating a relative excess of cardiac events with valsartan-based treatment in women but not in men (HR 1.21, 95 CI 1.03, 1.42 for women compared with 0.94, 95 CI 0.83, 1.07 for men). Blood pressure reduction, particularly SBP reduction, was larger throughout the duration of the trial, and especially during the first 3 months (treatment titration), in amlodipinetreated subjects compared with valsartan-treated subjects, but the between-treatment difference was greater in women than in men (Fig. 1). At month 1 the SBP difference in favour of amlodipine was 5.1 mmhg in women and 3.2 mmhg in men, and the average SBP difference in favour of amlodipine throughout

3 VALUE trial Zanchetti et al Fig. 1 6 SBP the trial was 2.8 mmhg for women and 1.8 mmhg for men. BP difference (mmhg) BP difference (mmhg) DBP Time (months) Females Males Females Males Between-treatment differences in systolic blood pressure (SBP) and diastolic blood pressure (DBP) throughout the VALUE study in men and women. Positive values indicate a lower blood pressure achieved with amlodipine-based than with valsartan-based treatment. BP, Blood pressure. 60 Effects of randomized treatment on the primary endpoint in various subgroups These are summarized in Figs 2 4, in which treatmentdependent blood pressure differences, incidences of composite cardiac events and HR are given for the various subgroups considered. With exception of female sex (Fig. 2), no subgroup showed a significant difference in the composite cardiac outcome between valsartanbased and amlodipine-based treatments. A tendency in favour of amlodipine in patients younger than 65 years did not reach a level of statistical significance (HR 1.17, 95 CI 0.98, 1.4, P ¼ 0.09). There were no significant effects of ethnicity, but the population was predominantly Caucasian (n ¼ ). As shown in Fig. 3 in no subgroup based on the presence or absence of the risk factors or disease used for recruiting patients into the VALUE trial was the incidence of cardiac outcomes significantly different between valsartan and amlodipine-treated patients, and in all subgroups the HR were very close to unity. The VALUE trial included a very large number of patients with type 2 diabetes mellitus (n ¼ 4823) and with proteinuria (n ¼ 3435), but the presence or absence of diabetes mellitus or proteinuria had no influence on the relative risk of cardiac events in valsartan compared with amlodipine-treated patients. Fig. 2 variable SBP/DBP (V-A) mmhg Valsartan Amlodipine HR (95 CI) P All 2.23/ Age < 65 (n = 5679) 65 (n = 9566) 1.57/ / Sex Female (n = 6468) Male (n = 8777) 2.83/ / Race Black (n = 639) Caucasian (n = 13617) Oriental (n = 533) Other (n = 456) 4.54/ / / / Region Asia (n = 441) Europe (n = 9347) Latin Amer. (n = 685) North Amer.(n = 4383) Rest of World (n = 389) 2.22/ / / / / Favours valsartan Favours amlodipine Primary endpoint (composite cardiac mortality and morbidity) in various subgroups of patients subdivided by demographic characteristics. Number of patients in each subgroup in parentheses. The other columns indicate (from left to right): systolic blood pressure (SBP) and diastolic blood pressure (DBP) mean differences between valsartan-based and amlodipine-based treatments [DSBP/DBP (V A)] (positive values indicate a greater blood pressure reduction by amlodipine-based treatment); percentages of patients with a first event in the valsartan and amlodipine groups; hazard ratios (HR) with 95 confidence intervals (CI), and P values. Cox regression models.

4 2166 Journal of Hypertension 2006, Vol 24 No 11 Fig. 3 variable SBP/DBP (V-A) mmhg Valsartan Amlodipine HR (95 CI) P Smoking status No (n = 11581) Yes (n = 3664) 2.32/ / Diabetes No (n = 10422) Yes (n = 4823) 2.02/ / Cholesterol > 240 mg/dl No (n = 10168) Yes (n = 5077) 2.17/ / LVH No (n = 12473) Yes (n = 2772) 1.93/ / Proteinuria No (n = 11810) Yes (n = 3435) 2.09/ / S. Creatinine > 1.7 mg/dl No (n = 14695) Yes (n = 530) 2.19/ / CHD No (n = 8264) Yes (n = 6981) 2.18/ / Stroke No (n = 12231) Yes (n = 3014) 2.14/ / PAD No (n = 13131) Yes (n = 2114) 2.23/ / ISH No (n = 13217) Yes (n = 2025) 1.91/ / Favours valsartan Favours amlodipine Primary endpoint in various subgroups of patients subdivided by risk factor or disease. CHD, Coronary heart disease; ISH, isolated systolic hypertension; LVH, left ventricular hypertension; PAD, peripheral artery disease. All other explanations as in Fig. 2. Fig. 4 variable SBP (V-A) mmhg Valsartan Amlodipine HR (95 CI) P Beta-blockers No (n = 10189) 1.85/1.43 Yes (n = 5056) 2.94/ Calcium antagonists No (n = 8988) 1.69/1.37 Yes (n = 6257) 2.49/ ACEI No (n = 8937) 2.04/1.44 Yes (n = 6308) 2.38/ Diuretics No (n = 9957) 1.63/1.32 Yes (n = 5288) 2.87/ Favours valsartan Favours amlodipine Primary endpoint in various subgroups of patients subdivided according to antihypertensive therapy immediately before randomization. ACEI, Angiotensin-converting enzyme inhibitor. All other explanations as in Fig. 2.

5 VALUE trial Zanchetti et al Figure 3 also shows that a history of various types of clinical event at the time of randomization (coronary heart disease, n ¼ 6981; stroke or transient ischemic attack, n ¼ 3014; peripheral arterial disease, n ¼ 2114) had no effect on the treatment-dependent HR for cardiac outcomes, which in all subgroups was very close to unity. In patients with isolated systolic hypertension (n ¼ 2025) the HR for cardiac events was Figure 4 shows that in 92 of the VALUE patients who were under antihypertensive treatment at the time of randomization, the nature of this previous treatment had no influence on the treatment-related HR for the primary cardiac endpoint. Treatment subgroup interaction analyses for secondary endpoints No significant treatment subgroup interaction was found for most demographic characteristics, risk factors or disease, when secondary endpoints such as myocardial infarction, heart failure and stroke were considered. For heart failure, however, there was a sex-related significant interaction (P < 0.001); male patients, but not female patients, had a significantly lower incidence of heart failure with the valsartan-based than with the amlodipine-based regimen (men: valsartan 4.1, amlodipine 5.8, HR 0.73, 95 CI ; women: valsartan 5.3, amlodipine 4.6, HR 1.18, 95 CI 0.95, 1.47). For stroke, only a history of stroke by treatment interaction was significant (P ¼ 0.027), with patients with a history of stroke having a tendency to a reduced risk of stroke with the valsartan-based regimen (valsartan 7.5, amlodipine 8.3, HR 0.92, 95 CI 0.71, 1.18), whereas those without a stroke history had a reduced risk of stroke with amlodipine (valsartan 3.4, amlodipine 2.6, HR 1.34, 95 CI 1.09, 1.65). Discussion The VALUE trial was designed to test the hypothesis that for the same blood pressure control, valsartan would more effectively reduce the primary endpoint of cardiac morbidity and mortality than amlodipine, but unequal achieved blood pressures confounded this comparison. The results of the prespecified subgroup analyses are consistent with the overall findings previously reported for the whole VALUE cohort [1]. There were no differences in the incidence of the composite cardiac event endpoint with valsartan-based and amlodipine-based treatments despite a significantly greater blood pressure decrease in the amlodipine group, independently of demographic or ethnic characteristics, concomitant risk factors and history of diseases, and the type of previous antihypertensive therapy. The only exception was sex, in which the amlodipinebased regimen was significantly more effective than the valsartan-based regimen for the primary endpoint in women, but not in men. The difference in outcome between sexes may be a chance finding. This is always a risk inherent in subgroup analyses, especially when a large number of subgroups is considered [4]. An alternative explanation for the outcome difference between sexes may be the greater unintended blood pressure difference in favour of the amlodipine-based regimen in women than in men (2.8 versus 1.8 mmhg throughout the trial, and even greater during the first few months). In the high-risk VALUE population even small blood pressure differences between treatments had important influences on outcomes [1,5]. It cannot be excluded that there is a genuine sex difference in the cardiac protection afforded by amlodipinebased and valsartan-based treatments. This possibility is consistent with another finding of the present analysis, namely that the trend towards less cardiac failure in valsartan-treated patients was significant in men, but not in women, and with findings of the Australian Blood Pressure Study 2, which showed a beneficial effect of enalapril limited to the male sex [6]. Also in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), amlodipine-based therapy, compared with lisinopril-based therapy, had a slightly greater effect on blood pressure in women than in men, and this was associated with a greater stroke protection of amlodipine in women [7]. However, two previous studies comparing an angiotensin receptor blocker with other antihypertensive treatment did not show a sex-related response to either losartan [8] or candesartan [9]. Likewise, other studies comparing regimens based on different antihypertensive agents have not reported a sex-related difference in outcome reduction for ACE inhibitors or calcium antagonists [10 12]. Contrary to what might have been expected [13], there were no greater differences in blood pressure reduction between the treatment groups in black compared with white patients, perhaps because of the addition of diuretics and other agents in most subjects of both treatment groups. Findings from the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study have suggested that compared with atenolol-based therapy, treatment based on the angiotensin receptor blocker losartan is not effective in preventing outcomes in black individuals [14]. The VALUE subgroup analysis showed a trend towards a slightly greater outcome reduction with amlodipine in black individuals, but the number of black individuals was small, the CI were very wide, and the difference was far from significant. VALUE included a considerable number of diabetic patients. The finding that the primary endpoint was equally affected by valsartan-based and amlodipinebased regimens in diabetic and non-diabetic individuals,

6 2168 Journal of Hypertension 2006, Vol 24 No 11 with HR very close to unity, and a narrow CI, gives strong support to a recent meta-analysis of the Blood Pressure Lowering Treatment Trialists Collaboration that antihypertensive regimens based on different agents have the same beneficial effects on cardiovascular outcomes in diabetic patients [15]. A similar conclusion can be drawn from our subanalysis of patients with baseline proteinuria, 42 of whom had diabetes. The increased cardiovascular risk associated with diabetes may increase the effects of blood pressure differences and would mask the possible blood pressure-independent effects of a particular treatment [2]. For stroke the only significant treatment by subgroup interaction was found in patients with or without a history of stroke, in whom a previous stroke or transient ischemic attack was associated with a trend towards a reduced HR for stroke recurrence after valsartan treatment. This is consistent with subgroup analyses in the Study on Cognition and Prognosis in the Elderly (SCOPE) [9], showing that a reduction in major cardiovascular events with candesartan was greater in patients who had had a previous stroke. 8 Os I, Oparil S, Kjeldsen S, Gerdts E, Hille DA, Lyle PA, et al. Female sex and effect of losartan versus atenolol in patients with hypertension and left ventricular hypertrophy. The LIFE Study [Abstract]. J Am Coll Cardiol 2005; 45:373A. 9 Trenkwalder P, Elmfeldt D, Hofman A, Lithell H, Olofsson B, Papademetriou V, et al. The study on COgnition and Prognosis in the Elderly (SCOPE): major cardiovascular events and stroke in subgroups of patients. Blood Pressure 2005; 14: Kjeldsen SE, Hedner T, Syvertsen JO, Lund-Johansen P, Hansson L, Lanke J, et al. Influence of age, gender and blood pressure levels on the principal endpoints of the Nordic Diltiazem (NORDIL) study. J Hypertens 2002; 20: Black HR, Elliot WJ, Grandit G, Grambsch P, Lucente T, White WB, et al. Principal results of the Controlled Onset Verapamil Investigation of Cardiovascular Events (CONVINCE) trial. JAMA 2003; 289: Pepine CJ, Handberg EM, Cooper-DeHoff RM, Marks RG, Kowey P, Messerli FH, et al. A calcium antagonist vs a non-calcium antagonist hypertension treatment strategy for patients with coronary artery disease. The International Verapamil-Trandolapril Study (INVEST): a randomized controlled trial. JAMA 2003; 290: Brewster LM, van Montfrans GA, Kleijnen J. Systematic review: antihypertensive drug therapy in black patients. Ann Intern Med 2004; 141: Julius S, Alderman MH, Beevers G, Dahlöf G, Devereux RB, Douglas JG, et al. Cardiovascular risk reduction in hypertensive black patients with left ventricular hypertrophy. J Am Coll Cardiol 2004; 43: Blood Pressure Lowering Treatment Trialists Collaboration. Effects of different blood pressure lowering regimens on major cardiovascular events in individuals with and without diabetes. Results of prospectively-designed overviews of randomized trials. Arch Intern Med 2005; 165: Precise information on the type of antihypertensive therapy received by VALUE patients before randomization showed that the type of previous therapy did not influence the response to randomized treatment, and justified the roll-over technique used in VALUE. This is an important observation, as the evidence from VALUE that high-risk hypertensive patients should have their blood pressure decreased promptly will make placebo run-in periods ethically difficult to justify in future trials involving high-risk patients. However, the price to pay of a roll-over design may be an unintended difference in blood pressure during the first months after randomization. References 1 Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, et al. Outcomes in hypertensive patients at high cardiovascular risk treated with regimens based on valsartan or amlodipine: the VALUE randomised trial. Lancet 2004; 363: Zanchetti A. Evidence-based medicine in hypertension: what type of evidence? J Hypertens 2005; 23: Mann J, Julius S, for the VALUE Trial Group. The Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial of cardiovascular events in hypertension. Rationale and design. Blood Pressure 1998; 7: Oxman AD, Guyatt GH. A consumer s guide to subgroup analyses. Ann Intern Med 1992; 116: Weber M, Julius S, Kjeldsen SE, Brunner HR, Ekman S, Hansson L, et al. Blood pressure dependent and independent effects of antihypertensive treatment on clinical events in the VALUE trial. Lancet 2004; 363: Wing LMH, Reid CM, Ryan P, Beilin LJ, Brown MA, Jennings GLR, et al. A comparison of outcomes with angiotensin-converting-enzyme inhibitor and diuretics for hypertension in the elderly. N Engl J Med 2003; 348: Leenen FHH, Nwachuku CE, Black HR, Cushman WC, Davis BR, Simpson LM, et al. Clinical events in high-risk hypertensive patients randomized to calcium channel blocker vs. ACE-inhibitor in ALLHAT. Hypertension 2006; 48: