Revisão Acesso aberto Revisado por pares

The Anglo-Scandinavian Cardiac Outcomes Trial

2012; Lippincott Williams & Wilkins; Volume: 60; Issue: 2 Linguagem: Inglês

10.1161/hypertensionaha.111.187070

ISSN

1524-4563

Autores

Peter Sever,

Tópico(s)

Hormonal Regulation and Hypertension

Resumo

HomeHypertensionVol. 60, No. 2The Anglo-Scandinavian Cardiac Outcomes Trial Free AccessReview ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessReview ArticlePDF/EPUBThe Anglo-Scandinavian Cardiac Outcomes TrialImplications and Further Outcomes Peter S. Sever Peter S. SeverPeter S. Sever From the International Centre for Circulatory Health, Imperial College London, London, United Kingdom. Originally published2 Jul 2012https://doi.org/10.1161/HYPERTENSIONAHA.111.187070Hypertension. 2012;60:248–259Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2012: Previous Version 1 At a meeting of the European Blood Pressure Group held in Madonna di Campiglio, Italy, in the spring of 1988, concern was expressed that there were no plans to evaluate the long-term efficacy of newer classes of blood pressure–lowering drugs, including the calcium channel blockers and the angiotensin-converting enzyme inhibitors, in morbidity and mortality outcome trials. These drugs were being increasingly used in clinical practice worldwide, and there was a view that industry perceived there was no need for long-term studies. Much had been made of the potential benefits of these newer drugs in short-term studies on surrogate end points compared with older therapies, but there was little pressure for investment in outcome trials in hypertension.At a further meeting of the European Blood Pressure group the following year, in 1989, a steering group outlined a proposal for a factorial-designed study to investigate not only whether newer treatments were better than old but also whether cholesterol lowering in a hypertensive population would confer benefits on coronary heart disease (CHD) events. At approximately the same time in the United States, proposals along similar lines were being discussed by the National Heart, Lung, and Blood Institute. However, further progress was delayed until September 1995, when it was proposed that if a group of United Kingdom trialists was to collaborate with the Gothenburg Trial Centre in Sweden, Pfizer would fund a major European outcome study.Preliminary discussions between United Kingdom and Swedish colleagues reviewed ongoing trials, including the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial, and focused on the design of a 2-way comparison trial comparing older treatment strategies with a newer treatment strategy based a calcium channel blocker and an angiotensin-converting enzyme inhibitor. The concept of a "combination treatment" trial germinated, particularly because it was recognized that most patients required >1 drug to control blood pressure and that no other trial was designed to compare specific treatment regimens. In addition, there was further discussion on a factorial design, involving a comparison of a lipid-lowering agent with placebo in a subgroup of patients with normal or modestly raised cholesterol levels for whom, at the time, there was no indication for lipid lowering. Further details on the background to Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) are given in Appendix 1 (available in the online-only Data Supplement).An independent steering committee was set up in 1996. The trial protocol was agreed and finalized, named the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT), and recruitment commenced in February 1998.Summary of Trial DesignThe ASCOT protocol has been published,1 and further information is available at www.ascotstudy.org. In summary, patients were recruited between February 1998 and May 2000, largely from family practices in United Kingdom, Ireland, and the Nordic countries. Hypertensive patients, on or off antihypertensive treatment, with no previous history of myocardial infarction or clinical CHD but with ≥3 risk factors for cardiovascular disease were eligible for the ASCOT-Blood Pressure–Lowering Arm (BPLA). These risk factors included male sex, age >55 years, a history of smoking, left ventricular hypertrophy or other specified ECG abnormalities, history of early CHD in a first-degree relative, microalbuminuria or proteinuria, noninsulin-dependent diabetes mellitus, peripheral vascular disease, previous stroke or transient ischemic attack, or ratio of plasma total cholesterol:high-density lipoprotein cholesterol of ≥6. Exclusion criteria included previous myocardial infarction, currently treated angina, cerebrovascular event within the previous 3 months, fasting serum triglycerides >4.5 mmol/L, heart failure, uncontrolled arrhythmias, or any clinically important hematologic or biochemical abnormalities.In ASCOT-BPLA, patients were randomized to 1 of the 2 blood pressure–lowering strategies, either amlodipine with or without perindopril (amlodipine based) or atenolol with or without bendroflumethiazide (atenolol based). In the ASCOT-Lipid-Lowering Arm (LLA), those with a fasting total cholesterol of ≤6.5 mmol/L (250 mg/dL) who were currently untreated with a statin or fibrate were randomized, using a factorial design, to either 10 mg of atorvastatin daily or matching placebo. Overall, 19342 patients were assigned either amlodipine-based treatment or atenolol-based treatment, and 10305 of these subjects were assigned atorvastatin or placebo. In the BPLA, at each follow-up visit, antihypertensive drug therapy was titrated and additional drugs added (perindopril to amlodipine and bendroflumethiazide K to atenolol) to achieve target blood pressure levels of <140/90 mm Hg for nondiabetic patients and 5.6 mmol/L). The HRs were 0.65 (P=0.015) and 0.63 (P=0.012), respectively, in these 2 groups. Similarly, in a further post hoc analysis, HRs for patients with baseline total cholesterol concentrations 80% of patients were currently on antihypertensive medication. The study was stopped prematurely after 5.5 patient-years of follow-up on the recommendation of the Data Safety Monitoring Board because of highly significant benefits of the amlodipine-based regimen on all-cause mortality and stroke outcomes.At the final BPLA visits, complete information was obtained on 99.5% of the 19257 patients originally randomized. Of the remainder, vital status was obtained on all but 122 patients (61 withdrew consent and 61 were lost to follow-up).Mean blood pressure reductions were substantial in both arms. Compared with those allocated to the atenolol-based treatment, blood pressures were lower throughout the trial among those allocated to amlodipine-based treatment. These differences were largest (5.9/2.4 mm Hg) at 3 months, but the average difference throughout the trial was 2.7/1.9 mm Hg. At the final visit, mean BPs had fallen to 137.7/79.2 mm Hg and 136.1/77.4 mm Hg in the atenolol-based and amlodipine-based limbs, respectively, representing mean falls of 25.7/15.6 mm Hg and 27.5/17.7 mm Hg.As intended by design, the majority of patients, 71%, were taking ≥2 antihypertensive agents, and 51% and 52% in the atenolol and amlodipine-based limbs, respectively, were taking one or the other of step 1 or step 2 agents of the regimen to which they were allocated.The primary end point of nonfatal myocardial infarction, including silent myocardial infarction and fatal CHD, was nonsignificantly lower by 10% (HR, 0.90 [95% CI, 0.79–1.02]; P=0.105) in the amlodipine-based group than in the atenolol-based group (Figure 3). There were significant reductions in 6 of the 7 secondary end points of the trial among those allocated to the amlodipine-based regimen compared with the atenolol-based regimen, including all-cause mortality, which was reduced by 11%, nonfatal myocardial infarction (excluding silent myocardial infarction), and fatal CHD (13%); total coronary events (13%); total cardiovascular events and procedures (16%); cardiovascular mortality (24%); and fatal and nonfatal stroke (23%). The results for 4 of the 7 tertiary end points were significantly in favor of those allocated to the amlodipine-based regimen with a significant 30% reduction (P≤0.00001) in new-onset diabetes mellitus. Only 1 end point (silent MI) was not affected favorably among those allocated the amlodipine-based regimen.Download figureDownload PowerPointFigure 3. Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT)-Blood Pressure–Lowering Arm (BPLA) summary of all end points. Modified from Dahlöf et al6 with permission of the publisher. Copyright © 2005, Elsevier.To facilitate comparisons with other major cardiovascular trials and because of the increasingly used diagnosis and treatment of acute coronary syndromes, 2 post hoc analyses were carried out; combined end points (cardiovascular death plus nonfatal myocardial infarction and stroke) and coronary revascularization plus the primary end point were evaluated. Both of these post hoc end points were significantly reduced among those allocated the amlodipine-based regimen. The proportional effect of allocation to the amlodipine-based regimen on the primary end point and on total cardiovascular events and procedures did not differ significantly in any prespecified subgroup, with benefits being significant in all of the subgroups for the latter composite end point.Slightly less than 25% of patients stopped therapy because of adverse events, with no significant difference between groups. There was, however, a significant difference in favor of the amlodipine-based regimen in the proportion stopping trial therapy because of serious adverse events (P=0.024).ASCOT-BPLA Implication of FindingsThe results of ASCOT-BPLA clearly demonstrated that, in hypertensive patients at increased risk of developing cardiovascular disease because of associated risk factors, the regimen based on amlodipine and perindopril was superior to one based on atenolol and bendroflumethiazide in reducing major cardiovascular events and all-cause mortality and also in inducing less new-onset diabetes mellitus.Although statistically significant reductions in the primary end point were not achieved, there was a strong trend in this direction, and given that the trial was powered to detect differences in this end point based on reaching 1150 primary events, the likelihood of achieving such a result was thereby reduced. However, combining all of the coronary events resulted in a significant benefit of amlodipine-based treatment, and on most other end points the "newer" treatment strategy was clearly associated with a better cardiovascular outcome, and a justifiable post hoc analysis, where coronary revascularization procedures were added to the primary event (thereby achieving a total number of 1284 events), was again associated with a significant reduction in the amlodipine-based treatment arm.Perhaps the most often asked question has been to what extent the blood pressure differences, which occurred predominantly early in the trial, could be implicated in the mechanisms underlying the differences in the 2 blood pressure arms of the study. This question was particularly relevant in the light of recent meta-analyses that have not found any major differences in coronary and cardiovascular outcomes when different antihypertensive treatment strategies are compared.Correcting for blood pressure differences in randomized trials is problematic, and there is no perfect way of achieving this. We undertook further analyses in an attempt to explain the magnitude of the beneficial effect of the amlodipine-based arm that could have been explained by the differences in blood pressure.7 For further details see Appendix 4.In a search for other possible factors that could have contributed to the outcome differences in ASCOT-BPLA, further analyses of the ASCOT-LLA were undertaken after the unblinding of the blood pressure arms. It appeared that assignment to one or other of the blood pressure arms influenced the risk reductions, particularly for coronary events when atorvastatin was compared with placebo. Additionally, differences in cardiovascular outcome comparing the amlodipine- based limb with the atenolol-based limb were influenced by whether patients were assigned atorvastatin. These analyses were further extended as summarized below.Synergy Between Atorvastatin- and Amlodipine-Based TreatmentIn those allocated amlodipine, the primary end point of nonfatal myocardial infarction plus fatal CHD was significantly reduced by 53% (HR, 0.47; 95% CI, 0.33–0.69; P 20 substudies were proposed and ratified by the Substudies Committee.13 Several additional substudies and preplanned subanalyses were incorporated subsequent to the onset of the trial. By March 2012, of 60 full journal publications from ASCOT, 39 were reports of substudies and a further 11 from subanalyses of the ASCOT database. A full list of the substudies is available on the ASCOT web site and listed in the supplementary online ASCOT bibliography (Appendix 5). For the purpose of this review, the author has selected 4 key hemodynamic substudies with which he has been associated, which have been highly cited, and which have important clinical implications.Ambulatory Blood Pressure Monitoring Substudy of ASCOT14In this important substudy, 1905 patients from 4 ASCOT centers underwent repeated ambulatory blood pressure monitoring (ABPM) over a median follow-up period of 5.5 years. An average of 3.5 recordings was obtained from each participant. As expected from the whole ASCOT population, in this substudy clinic blood pressures were lower in the amlodipine/perindopril-treated patients compared with those treated with atenolol/thiazide (−1.5/−1.2 mm Hg).On ABPM, daytime blood pressures during follow-up were higher in patients treated with amlodipine/perindopril (+1/+1.6 mm Hg). On the other hand, nighttime systolic but not diastolic blood pressure was lower in patients treated with amlodipine/perindopril therapy (−2.2/3.8 mm Hg; Figure S4). It is important to note that, when comparing these blood pressure differences with those reported in the original ASCOT-BPLA article, the ABPM recordings and related clinic blood pressures were obtained from visits ≈1 year into the trial and on ≥2 occasions thereafter. Therefore, the early differences in blood pressure, for the first few months after randomization, in favor of amlodipine/perindopril, would not have been apparent in this ABPM substudy. ABPM values were significantly associated with the rates of cardiovascular events, and nocturnal pressures provided complementary and incremental use over clinic blood pressures in the prediction of cardiovascular risk in these treated hypertensive patients.Blood Pressure Variability as A predictor of Cardiovascular Outcome in ASCOT and the Differential Effect of Amlodipine/Perindopril and Atenolol/Thiazide RegimensThe hypothesis by Rothwell et al,15,16 based on observations in patients with recurrent transient ischemic attacks or strokes, that blood pressure variability rather than blood pressure alone is an important predictor of future events, was explored in the ASCOT population. Several measures of blood pressure variability were obtained throughout the trial. Visit-to-visit variability was expressed as the SD or coefficient of variation of between visit measures. In addition, a transformation of blood pressure variability independent of in-trial blood pressure was also determined. Three measures of blood pressure were obtained at each visit throughout the trial, which allowed an estimate of within-visit variability. Finally, blood pressure variability was determined from the subgroup of patients who underwent 24-hour ABPM.In the first set of analyses, data were analyzed from 6 months after randomization to the end of the trial. The rationale for this was that, during the first few months of the trial, blood pressures were rapidly falling in both treatment arms as a result of dosage adjustment and the introduction of second- and third-line drugs. By 6 months in-trial, blood pressures had largely stabilized. Mean blood pressures within trial and measures of visit-to-visit variability throughout the trial were expressed as deciles.In keeping with earlier analyses on the initial ASCOT database, in-trial mean blood pressure was a poor predictor of stroke outcome. Excess stroke risk was only seen in the highest decile of in-trial mean systolic blood pressure, and in-trial mean blood pressure did not predict coronary risk (Figure 5). On the other hand, all of the measures of visit-to-visit systolic blood pressure variability (SD, coefficient of variation, and variation independent of mean) were powerful predictors of both stroke and coronary outcome in both treatment groups (Figure 6). Comparing the top with the bottom decile of systolic blood pressure visit-to-visit variability, there was an ≈4-fold excess risk of stroke outcomes and a 2- to 3-fold increase in the risk of coronary events. Visit-to-visit blood pressure variability (by all measures) was consistently lower in the amlodipine/perindopril group than in the atenolol/thiazide group.16Download figureDownload PowerPointFigure 5. Visit-to-visit mean systolic blood pressure expressed in deciles, hazard ratios (95% CI), and number of stroke and coronary events in each decile. Modified from Rothwell et al16 with permission of the publisher. Copyright © 2010, Elsevier.Download figureDownload PowerPointFigure 6. Visit-to-visit systolic blood pressure variability expressed in deciles, hazard ratios (90% CI), and number of strokes and coronary events in each decile. Modified from Rothwell et al16 with permission of the publisher. Copyright © 2010, Elsevier.Both stroke risk and coronary risk were also predicted by measures of within-visit variability in blood pressure, although the relationship was much weaker than observed with between-visit variability. Likewise, intra-ABPM blood pressure variability and daytime systolic blood pressure also predicted both stroke and coronary events but less so than visit-to-visit variability. No measures of variability in heart rate were found to predict any cardiovascular outcomes.We reported in the main trial results that risk reductions in stroke (23%) and coronary events (13%) were associated with assignment to amlodipine/perindopril treatment compared with atenolol/thiazide treatment. In a multiple regression model, incorporation of usual systolic blood pressure throughout the trial has been reported previously to have a minimal effect on the HRs for stroke in favor of amlodipine/perindopril treatment and to have no effect on coronary outcome. However, incorporation of measures of visit-to-visit variability and within-visit variability of blood pressure throughout the trial into the model totally eliminated the differences observed in both stroke and coronary outcome in favor of amlodipine/perindopril treatment.16These data strongly suggest that blood pressure variability rather than blood pressure, per se, was an important determinant of cardiovascular outcomes in this large hypertensive population. In addition there were important differences between the 2 treatment arms in ASCOT that were explained almost exclusively by differences in blood pressure variability.Conduit Artery Functional Evaluation SubstudySee Appendix 5.17Studies on Wave Reflection, Differential Effects of Amlodipine/Perindopril Regimen Versus Atenolol/Thiazide Regimen on Central Blood Pressure, and Predictions of Cardiovascular EventsSee Appendix 6.18,19ASCOT Biomarker ProgramAn important initiative was undertaken before the commencement of ASCOT to support a repository of blood samples from all of the participants to be used for the subsequent identification of potential biomarkers for the future development of cardiovascular end points. A similar DNA resource was established from >9000 ASCOT patients.Of 10 potential b

Referência(s)
Altmetric
PlumX