Calcium Channel Blockers Versus Other Classes of Drugs for Hypertension (Review)

Calcium channel blockers versus other classes of drugs for hypertension (Review)

Chen N, Zhou M, Yang M, Guo J, Zhu C, Yang J, Wang Y, Yang X, He L

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2010, Issue 8 http://www.thecochranelibrary.com

Calcium channel blockers versus other classes of drugs for hypertension (Review) Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. TABLE OF CONTENTS HEADER...... 1 ABSTRACT ...... 1 PLAINLANGUAGESUMMARY ...... 2 BACKGROUND ...... 2 OBJECTIVES ...... 2 METHODS ...... 3 RESULTS...... 3 Figure1...... 5 Figure2...... 7 Figure3...... 9 Figure4...... 10 Figure5...... 11 Figure6...... 12 Figure7...... 14 Figure8...... 15 Figure9...... 16 Figure10...... 16 Figure11...... 17 Figure12...... 18 Figure13...... 19 DISCUSSION ...... 19 AUTHORS’CONCLUSIONS ...... 21 ACKNOWLEDGEMENTS ...... 21 REFERENCES ...... 21 CHARACTERISTICSOFSTUDIES ...... 25 DATAANDANALYSES...... 45 Analysis 1.1. Comparison 1 All cause mortality, Outcome 1 CCBs vs other classes of antihypertensive agents. . . . 48 Analysis 2.1. Comparison 2 Myocardial infarction, Outcome 1 CCBs vs other classes of antihypertensive agents. . . 50 Analysis 2.2. Comparison 2 Myocardial infarction, Outcome 2 Amlodipine vs. ACE inhibitor...... 51 Analysis 3.1. Comparison 3 Stroke, Outcome 1 CCBs vs other classes of antihypertensive agents...... 52 Analysis 4.1. Comparison 4 Congestive heart failure, Outcome 1 CCBs vs other classes of antihypertensive agents. . 54 Analysis 5.1. Comparison 5 Cardiovascular mortality, Outcome 1 CCBs vs other classes of antihypertensive agents. . 55 Analysis 5.2. Comparison 5 Cardiovascular mortality, Outcome 2 DHP vs. β-blocker...... 57 Analysis 6.1. Comparison 6 Major cardiovascular events, Outcome 1 CCBs vs other classes of antihypertensive agents. 58 Analysis 6.2. Comparison 6 Major cardiovascular events, Outcome 2 Sensitivity analysis: CCBs vs. ACE inhibitors. . 59 Analysis 7.1. Comparison 7 Blood pressure reduction, Outcome 1 Systolic blood pressure reduction...... 60 Analysis 7.2. Comparison 7 Blood pressure reduction, Outcome 2 Diastolic blood pressure reduction...... 61 Analysis 7.3. Comparison 7 Blood pressure reduction, Outcome 3 Sensitivity analysis: CCBs vs. ACE inhibitors. . . 63 APPENDICES ...... 63 HISTORY...... 65 CONTRIBUTIONSOFAUTHORS ...... 65 DECLARATIONSOFINTEREST ...... 65 SOURCESOFSUPPORT ...... 66 DIFFERENCES BETWEEN PROTOCOL AND REVIEW ...... 66 NOTES...... 66 INDEXTERMS ...... 66

Calcium channel blockers versus other classes of drugs for hypertension (Review) i Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. [Intervention Review] Calcium channel blockers versus other classes of drugs for hypertension

Ning Chen1, Muke Zhou1, Mi Yang1, Jian Guo1, Cairong Zhu2, Jie Yang1, Yucai Wang3, Xue Yang1, Li He1 1Department of Neurology, West China Hospital, Sichuan University, Chengdu, China. 2Epidemic Disease & Health Statistics De- partment, School of Public Health, Sichuan University, Sichuan, China. 3West China Hospital, Sichuan University, Sichuan, China Contact address: Li He, Department of Neurology, West China Hospital, Sichuan University, Wai Nan Guo Xue Xiang #37, Chengdu, Sichuan, 610041, China. [email protected]. Editorial group: Cochrane Hypertension Group. Publication status and date: New, published in Issue 8, 2010. Review content assessed as up-to-date: 6 November 2009. Citation: Chen N, Zhou M, Yang M, Guo J, Zhu C, Yang J, Wang Y, Yang X, He L. Calcium channel blockers versus other classes of drugs for hypertension. Cochrane Database of Systematic Reviews 2010, Issue 8. Art. No.: CD003654. DOI: 10.1002/14651858.CD003654.pub4.

ABSTRACT Background Calcium channel blockers (CCBs) are a relatively new antihypertensive class. The effect of first-line CCBs on the prevention of cardiovascular events, as compared with other antihypertensive drug classes, is unknown. Objectives To determine whether CCBs used as first-line therapy for hypertension are different from other first-line drug classes in reducing the incidence of major adverse cardiovascular events. Search methods Electronic searches of the Cochrane Central Register of Controlled Trials, MEDLINE, EMBASEand the WHO-ISH Collaboration Register (up to May 2009) were performed. We also checked the references of published studies to identify additional trials. Selection criteria Randomized controlled trial (RCT) comparing first-line CCBs with other antihypertensive classes, with at least 100 randomized hypertensive participants and with a follow-up of at least two years. Data collection and analysis Two authors independently selected the included trials, evaluated the risk of bias and entered the data for analysis. Main results Eighteen RCTs (14 dihydropyridines, 4 non-dihydropyridines) with a total of 141,807 participants were included. All-cause mortality was not different between first-line CCBs and any other first-line antihypertensive classes. CCBs reduced the following outcomes as compared to β-blockers: total cardiovascular events (RR 0.84, 95% CI [0.77, 0.92]), stroke (RR 0.77, 95% CI [0.67, 0.88]) and cardiovascular mortality (RR 0.90, 95% CI [0.81, 0.99]). CCBs increased total cardiovascular events (RR 1.05 , 95% CI [1.00, 1.09], p = 0.03) and congestive heart failure events (RR 1.37, 95% CI [1.25, 1.51]) as compared to diuretics. CCBs reduced stroke (RR 0.89, 95% CI [0.80, 0.98]) as compared to ACE inhibitors and reduced stroke (RR 0.85, 95% CI [0.73, 0.99]) and MI (RR 0.83, 95% CI [0.72, 0.96]) as compared to ARBs. CCBs also increased congestive heart failure events as compared to ACE inhibitors (RR 1.16, 95% CI [1.06, 1.27]) and ARBs (RR 1.20, 95% CI [1.06, 1.36]). The other evaluated outcomes were not significantly different.

Diuretics are preferred ﬁrst-line over CCBs to optimize reduction of cardiovascular events. The review does not distinguish between CCBs, ACE inhibitors or ARBs, but does provide evidence supporting the use of CCBs over β-blockers. Many of the differences found in the current review are not robust and further trials might change the conclusions. More well-designed RCTs studying the mortality and morbidity of patients taking CCBs as compared with other antihypertensive drug classes are needed for patients with different stages of hypertension, different ages, and with different co-morbidities such as diabetes.

PLAIN LANGUAGE SUMMARY

Calcium channel blockers versus other classes of drugs for hypertension

Calcium channel blockers (CCBs) are used as a ﬁrst-line antihypertensive drug for patients with hypertension, but whether this is the best way to reduce adverse cardiovascular events is unknown. We aimed to evaluate the effect of ﬁrst-line CCBs on preventing adverse cardiovascular events as compared to other antihypertensive drugs. We found 18 trials in 141,807 participants to answer the question. All-cause mortality was not different between CCBs and any other antihypertensive drug classes. Diuretics were found to be better at reducing total cardiovascular events than CCBs and CCBs were found to be better at reducing total cardiovascular events than β- blockers. This information will be helpful for health professionals and patients to assist them in choosing the best drug for initial treatment of hypertension.

BACKGROUND prospectively-designed overview of randomised trials showed that It is established in placebo-controlled trials that the lowering of there were no significant differences in total major cardiovascu- elevated blood pressure (BP) with thiazide diuretics and β-block- lar events between regimens based on angiotensin-converting en- ers as first-line agents reduces the incidence of major cardiovas- zyme (ACE) inhibitors, CCBs, or diuretics or β-blockers, although cular complications of hypertension (Psaty 1997; Wright 1999; ACE-inhibitor-based regimens reduced blood pressure less than Wright 2009). Observational studies have suggested that calcium the other classes (Turnbull 2003). channel blockers (CCBs) or calcium antagonists might be inferior to diuretics and β-blockers in reducing these complications Currently available individual trials, in isolation, do not have suf- (Psaty 1995; Pahor 1995). Two large trials have reported find- ficient power to allow ranking of the various available classes of ings suggesting that it may matter how elevated BP is lowered antihypertensive drugs based on their ability to prevent morbidity (HOPE 2000; ALLHAT). The investigators of the Heart Out- and mortality. The issue of first-line drug selection is highly rele- comes Prevention Evaluation (HOPE) trial suggested that much vant for millions of subjects receiving drug therapy for hyperten- of the benefit of ramipril appears to be mediated through non- sion. Since the documentation of moderate treatment differences blood pressure lowering mechanisms (HOPE 2000). In ALLHAT, requires a very large number of events, the most reliable informa- first-line doxazosin was shown to be inferior to first-line chlorthali- tion will come from controlled clinical trials of sufficient size and done for several cardiovascular outcomes (ALLHAT). Moreover, duration. As MacMahon & Neal pointed out in a recent editorial, recent controlled trials of CCBs, compared to alternative ther- approximately 1,000 events are required in order to detect moder- apies, have shown similar BP lowering potential, but trends to- ate relative treatment group differences of 10 to 15% (MacMahon wards high event rates for the CCBs (ABCD; FACET). A prospec- 2000). They also drew attention to the value of meta-analysis to tive, randomised, double-blind trial in Europe and Israel indicated increase the power of detecting differences. that nifedipine once daily and co-amilozide were equally effective in preventing overall cardiovascular or cerebrovascular complications (INSIGHT). Another prospective, randomised, open, OBJECTIVES blinded endpoint study showed that diltiazem is as effective as treatment with diuretics, β-blockers, or both in preventing the To determine whether CCBs used as first-line therapy for hyper- combined primary endpoint of all stroke, myocardial infarction tension were different from other first-line treatment alternatives (MI), and other cardiovascular death (NORDIL). Results of a in reducing the incidence of adverse major cardiovascular events.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 2 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. METHODS Electronic databases were searched using a strategy combining the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity-maximizing version Criteria for considering studies for this review (2008 revision) with selected MeSH terms and free text terms re- lating to calcium channel blockers and hypertension. No language restriction was used. The MEDLINE search strategy (Appendix Types of studies 1) was translated into CENTRAL (Appendix 2) and EMBASE (Appendix 3) using the appropriate controlled vocabulary as ap- Each included study of this present review (a) was a randomized plicable. controlled trial (RCT), (b) followed the participants for at least two years, and (c) randomized 100 or more participants. Searching other resources Types of participants 1. Clinical Trials.gov and Current Controlled Trials 2. Hand searching of those high-yield journals and conference Participants all had a baseline BP of at least 140 mm Hg systolic proceedings which have not already been hand searched on or 90 mm Hg diastolic, measured in a standard way on at least 2 behalf of the Cochrane Collaboration occasions. If a trial was not limited to patients with elevated BP it 3. Reference lists of all papers and relevant reviews identified must separately reported outcome data on patients with elevated 4. Authors of relevant papers were contacted regarding any BP as defined above. further published or unpublished work 5. Authors of trials reporting incomplete information were Types of interventions contacted to provide the missing information Trials comparing first-line CCBs with other first-line antihyper- 6. ISI Web of Science was searched for papers which cited tensive classes were included. The majority (> 70%) of the patients studies included in the review in all study groups must be taking the assigned drug class after 1 year. Supplemental drugs other than CCBs from other drug classes were allowed as stepped therapy. Data collection and analysis Eligible studies were identified according to whether a clearly identified first-line CCB was compared to one or more drugs from Types of outcome measures other first-line antihypertensive classes in patients with elevated Main outcomes of the study were: BP. All data were independently abstracted by two investigators 1. all cause mortality with the use of standardized abstract forms. Disagreements from 2. MI (non-fatal and fatal MI plus sudden or rapid death) abstraction were resolved with discussion. Data on the outcomes 3. stroke (non-fatal and fatal stroke) of acute myocardial infarction (AMI), stroke, congestive heart fail- 4. congestive heart failure ure, combined major cardiovascular events, and mortality were 5. cardiovascular mortality abstracted. Review Manager (RevMan) 5 software was used for 6. major cardiovascular events (MI, congestive heart failure, all data analyses. The overall risk ratios (RRs) and 95% confi- stroke and cardiovascular mortality) dence intervals (CIs) of each outcome were calculated by means of 7. systolic and diastolic BP the Mantel-Haenstzel method using fixed effects models (Mantel 1959). The Breslow-Day test for heterogeneity was used to assess the extent to which the differences among the trial results due to Search methods for identification of studies random fluctuations. BP change was calculated by subtracting the baseline value at randomization from the value reported at the end of the trial. Data for Electronic searches BP reduction were combined using the weighted mean-difference The Database of Abstracts of Reviews of Effectiveness (DARE) method. We also used the fixed effects model. When significant were searched for related reviews. heterogeneity was present, a random effects model was used. Risks The following electronic databases were searched for primary stud- of bias were assessed using the Risk of Bias Tool and Tables of ies: RevMan 5. 1. The Cochrane Central Register of Controlled Trials (CENTRAL) to May 2009. 2. Bibliographic databases, including MEDLINE (2000- May 2009) and EMBASE (1999- May 2009). RESULTS

Calcium channel blockers versus other classes of drugs for hypertension (Review) 3 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Description of studies cular disease (CVD) (ALLHAT;ASCOT-BPLA;CONVINCE; INSIGHT); with coronary artery disease (CAD) (INVEST); with See: Characteristics of included studies; Characteristics of excluded cardiovascular risks or CVD (VALUE); with type 2 diabetes mel- studies; Characteristics of studies awaiting classification. litus (NIDDM) (ABCD; FACET) or type 2 diabetes mellitus and See Tables: Characteristics of included studies and Characteristics nephropathy (IDNT); African Americans with hypertensive kid- of excluded studies. ney disease (AASK). According to the aforementioned methods we identified 37 po- All the 18 RCTs recruited participants of both sex, but require- tentially eligible trials, from which we excluded 18 after screening ments for age differed among trials: > 40 years (MIDAS) > 50 the full texts. The reasons for exclusion included: non-random- years (INVEST; VALUE) > 55 years (ALLHAT; CONVINCE) ized design (Abascal 1998; DHCCP;Pahor 1995; Psaty 1995); or > 60 years (NICS-EH; SHELL), 18-70 years (AASK), 30-70 the controlled group used placebo, instead of other classes of an- years (IDNT), 40-65 years (VHAS), 40-74 years (ABCD), 40- tihypertensive drugs (STONE; Syst-China; Syst-Eur); compari- 79 years (ASCOT-BPLA), 45-69 years (TOMHS), 45-75 years son was performed between different kinds of CCBs, without any (ELSA), 50-74 years (NORDIL), 55-80 years (INSIGHT), or 70- other classes of antihypertensive drugs (Kes 2003); the follow- 84 years (STOP-Hypertension-2). up was shorter than 2 years (Gottdiener 1997; Papademetriou Most trials followed a goal BP in their protocols, mostly less 1997; PRESERVE; Schneider 1991); small sample of partic- than 140/90 mm Hg (ALLHAT;ASCOT-BPLA; CONVINCE; ipants (less than 100 were randomized) (Bakris 1996;Bakris FACET INVEST; INSIGHT; VALUE), or less than 130/85 mm 1997;FACTS;Maharaj 1992;Radevski 1999); study groups were Hg for patients with diabetes or renal impairment (ASCOT- different in target BPs instead of drug classes (HOT). BPLA;INVEST), or ≤135/85 mm Hg or a decrease ≥ 10 mm Finally, 18 RCTs (AASK; ABCD;ALLHAT; ASCOT-BPLA; Hg systolic for diabetic patients (IDNT), or ≤ 160/95 mm CONVINCE; ELSA; FACET; IDNT; INSIGHT; INVEST; Hg (STOP-Hypertension-2), or less than 90 mm Hg diastolic MIDAS; NICS-EH; NORDIL; SHELL; STOP-Hypertension-2; (NORDIL) or 95 mm Hg (TOMHS), or less than 95 mm Hg TOMHS; VALUE; VHAS) with a total of 141,807 participants with a fall of at least 5 mm Hg (ELSA), or less than 90 mm Hg remained to be reviewed at length (the total number is different with a fall of at least 10 mm Hg (MIDAS), reduction more than from the summation of numbers of participants given in the char- 20 mmHg or systolic BP ≤160 mmHg (SHELL), or ≤ 90 mm acteristic form, due to a subtraction of placebo groups from two Hg or ≤ 95 mm Hg with a reduction of at least 10% from baseline trials (IDNT;TOMHS)). value (VHAS), or 75 mm Hg or less diastolic in the intensive- All the included RCTs supplied explicit inclusion and exclu- treatment group while 80-89 mm Hg diastolic in the moderate- sion criteria. Seventeen of these trials included only hyperten- treatment group (ABCD), or 102-107 mm Hg of mean arterial sive patients, but were defined differently as: 140/90 mm Hg or pressure (MAP) in the usual goal group while 92 mm Hg or less more (FACET; INVEST); more than 160/95 mm Hg (VHAS); in the lower goal group (AASK), or a decrease ≥20 mm Hg of BP more than 135/85 mm Hg for patients with diabetes mellitus if systolic BP was more than 160 mm Hg or diastolic BP more (IDNT); 140-179 mm Hg systolic and/or 90-109 mm Hg dias- than 110 mm Hg (FACET). tolic (ALLHAT); 150-210 mm Hg systolic and 95-115 mm Hg Of CCBs for hypertension, dihydropyridines (DHPs) were the diastolic (ELSA); systolic BP ≥180 mm Hg and/or diastolic BP most commonly studied, especially amlodipine (AASK; ALLHAT; ≥ 105 mm Hg (STOP-Hypertension-2); diastolic BP of 100 mm ASCOT-BPLA; FACET; IDNT; TOMHS; VALUE). Other Hg or more (NORDIL) or of 90-99 mm Hg (TOMHS); treated DHPs also studied in the present review were nifedipine ( hypertension with a upper limit of 175/100 mm Hg or untreated INSIGHT), felodipine (STOP-Hypertension-2), nisoldipine ( hypertension of 140-190 mm Hg systolic or 90-110 mm Hg di- ABCD), nicardipine (NICS-EH), lacidipine (ELSA; SHELL), and astolic (CONVINCE); BP ≥160/100 mmHg for subjects with isradipine (MIDAS). Non-DHPs were studied in other trials, such untreated hypertension or BP ≥140/90 mmHg for subjects on an- as an aralkylamine derivative verapamil (CONVINCE; INVEST; tihypertensive treatment (ASCOT-BPLA); systolic BP ≥150 mm VHAS), and a benzothiazepine derivative diltiazem (NORDIL). Hg and diastolic BP ≥ 95 mm Hg, or only systolic BP ≥ 160 The included RCTs compared one of the above CCBs to other mm Hg (INSIGHT); only diastolic BP ≥ 95 mm Hg (AASK) classes of antihypertensive drugs, including: a diuretic (ALLHAT; or of 90-115 mm Hg (MIDAS); 160-210/220 mm Hg systolic INSIGHT; MIDAS; NICS-EH; SHELL; TOMHS; VHAS), a β- and less than 115 mm Hg diastolic (VALUE/NICS-EH); ≥160 blocker (AASK; ASCOT-BPLA; ELSA; INVEST; TOMHS), a di- mmHg systolic, and ≤ 95 mmHg diastolic (SHELL). Only one uretic or a β-blocker or both, data of which could not be separated trial (ABCD) did not limit patients to elevated BP (diastolic BP ≥ for each drug (CONVINCE; NORDIL; STOP-Hypertension-2), 80 mm Hg), but it separately reported outcomes on patients with an α -antagonist (TOMHS), an ACE inhibitor (AASK; ABCD; elevated BP (diastolic BP ≥ 90 mm Hg), so data of hypertensive 1 ALLHAT; FACET; STOP-Hypertension-2; TOMHS), or an an- patients could be extracted. giotensin receptor blocker (ARB) (IDNT; VALUE). Additional inclusion criteria varied for each study: with other Supplemental antihypertensive agents other than study drugs were risk factor(s) for coronary heart disease (CHD) or cardiovas-

Calcium channel blockers versus other classes of drugs for hypertension (Review) 4 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. allowed in most included trials, often administered sequentially VALUE). However, when trials did not supply SDs for the base- to achieve BP goals (AASK; ABCD; ALLHAT; ASCOT-BPLA; line and final BPs (ASCOT-BPLA; CONVINCE; INSIGHT; CONVINCE; ELSA; IDNT; INSIGHT; INVEST; MIDAS; STOP-Hypertension-2; ABCD; IDNT; ELSA; MIDAS; VHAS; NORDIL; SHELL; STOP-Hypertension-2; VALUE; VHAS). SHELL), the BP results were not entered into the meta-analysis. The FACET trial added the study drug of the other group to Mean duration of follow-up ranged from 2 to 5 years. One trial patients whose BPs were not controlled well. The TOMHS trial (STOP-Hypertension-2) stated that no patient was lost to follow- studied 5 classes of first-line antihypertensive drugs, and it added up and no patient refused to continue in the study, while loss to chlorthalidone or enalapril, both of which were study drugs, to follow-up and withdrawal were reported in the other 16 trials. patients to control BPs. NICS-EH prohibited any other antihy- All trials, except the NICS-EH trial, stated the performance of an pertensive drugs. intention-to-treat analysis. Outcomes differed among studies, but results of cardiovascular events and BP changes, which we planed to evaluate, were reported in most trials. However, fatal MI, stroke and heart failure Risk of bias in included studies were sometimes contained in death events, and components of cardiovascular events were not separately reported in some trials. Since trials with a small sample were excluded from the present re- As a result, not every trial supplied data to each meta-analysis for view, most of included trials were large and multicenter trials with outcomes of this review. Only two trials explicitly presented the standardized protocols. The methodological quality of included mean BP changes with standard deviations (SDs) (INVEST) or trials was evaluated in several ways. According to the summary as- standard errors (SEs) (TOMHS) which could be directly inputted sessment of the risk of bias for each important outcome (Haggins in the RevMan for analysis. In some other trials, mean BP change 2008 - Chap. 8), 4 of the trials (ALLHAT; ASCOT-BPLA; IDNT; could be calculated by subtracting the baseline value at random- INVEST) were rated as a low risk of bias, 2 trials (FACET; ization from the value reported at the end of the trial, but SDs for NICS-EH) as a high risk of bias and the other 12 as an unclear changes were not reported. We calculated change-from-baseline risk of bias. The risk of bias graph (Figure 1) shows judgements SDs when baseline and final values were known (Haggins 2008 of the review authors about each domain presented as percentages - Chap. 16)(ALLHAT; NORDIL; FACET; NICS-EH; AASK; across all included studies.

Figure 1. Methodological quality graph: review authors’ judgements about each methodological quality item presented as percentages across all included studies.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 5 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. All these studies were stated as randomized controlled trials. A computer-generated code for randomization was often used, but 6 on the whole equal among the treatment groups, and an inten- trials did not report the methods of allocation (ABCD; INSIGHT; tion-to-treat analysis, which meant data were analyzed accord- NICS-EH; NORDIL; STOP-Hypertension-2; VHAS). Alloca- ing to randomized treatment assignments regardless of the sub- tion concealment was seldom described, only 4 trials (ALLHAT; sequent medications (Fergusson 2002), was performed in most ASCOT-BPLA; IDNT; INVEST) stated their randomization included trials, except the STOP-Hypertension-2 trial (with neg- codes were concealed at the clinical trials center; while in the ligible loss) and the NICS-EH trial. Some sites and their pa- CONVINCE trial, an Interactive Voice Response System (IVRS) tients were excluded after randomization because of poor docu- for randomizing, assigning, and tracking blinded medication was mentation of informed consent, data integrity concerns, or mis- used. Information of other trials was insufficient to assess this risk conduct (ASCOT-BPLA;ALLHAT; CONVINCE; INSIGHT), of bias domain. which might lead to attrition bias; fortunately sufficient informa- All the trials compared two first-line antihypertensive drug classes tion of reports helped us to restore those participants to the correct to each other. With exception of the FACET trial, which was groups. open-label, the others were all stated as blinded. Active drugs were Another potential risk of bias came from the requirement that described as of indistinguishable appearance in some trials, but to control elevated BP, additional antihypertensive agents were it was still impossible to know the extent of blinding (Haggins usually allowed. If this was not balanced between groups it would 2008 - Chap. 8). A “prospective, randomised, open trial with lead to a risk of performance bias. A high risk of this type of bias blinded-endpoint” (PROBE) design was performed in 4 trials ( was judged for the FACET and the NORDIL trial. ASCOT-BPLA; INVEST; NORDIL; STOP-Hypertension-2); it Effects of interventions differed from the classical double-blind method. In a PROBE study, outcomes are evaluated by a blinded end-point committee The diuretic and β-blocker subgroup included data of studies in (Hansson 1992) to avoid detection bias. In that way allocation of which a diuretic, a β-blocker or both were used but could not be treatment might be open to risk of bias from patients and doctors. separately analyzed. Missing data caused by loss to follow-up or withdrawals were All cause mortality (Analysis 1.1; Figure 2)

Calcium channel blockers versus other classes of drugs for hypertension (Review) 6 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 2. Forest plot of comparison: 1 All cause mortality, outcome: 1.1 CCBs visus other classes of antihypertensive agents.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 7 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. The effect of CCBs on death from any cause was not significantly different from that of any other evaluated agents: diuretics (5 trials that of diuretics (5 trials with 34,072 participants: RR 1.00, 95% with 35,057 participants: RR 0.98, 95% CI [0.92, 1.04], I2= CI [0.92, 1.08], I2=0%), β-blockers (3 trials with 22,249 par- 0%), β-blockers (4 trials with 44,825 participants: RR 0.94, 95% ticipants: RR 0.90, 95% CI [0.79, 1.02], I2=0%), diuretics and CI [0.88, 1.00], I2=0%), diuretics and β-blockers (3 trials with β-blockers (3 trials with 31,892 participants: RR 1.05, 95% CI 31,892 participants: RR 1.03, 95% CI [0.94, 1.12], I2=0%), ACE [0.93, 1.19], I2=72%), ACE inhibitors (5 trials with 24,006 par- inhibitors (5 trials with 24,006 participants: RR 0.96, 95% CI ticipants: RR 1.06, 95% CI [0.98, 1.15], I2=72%). Incidence of [0.91, 1.03], I2=0%), and ARBs (2 trials with 16,391 participants: MI was statistically significantly lower (p = 0.009) for CCBs com- RR 0.98, 95% CI [0.90, 1.07], I2=0%). pare to ARBs (2 trials with 16,391 participants: RR 0.83, 95% MI (non-fatal and fatal MI plus sudden or rapid death) CI [0.72, 0.96], I2=35%), with a risk difference (RD) of -0.01. The effect of CCBs on MI was not significantly different from (Analysis 2.1; Figure 3)

Calcium channel blockers versus other classes of drugs for hypertension (Review) 8 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 3. Forest plot of comparison: 2 Myocardial infarction, outcome: 2.1 CCBs visus other classes of antihypertensive agents.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 9 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Significant statistical heterogeneity was shown between trials comparing CCBs to diuretics and β-blockers (I2=72%, p = 0.03) groups (RR 0.81, 95% CI [0.65, 1.02]); while the other two trials and CCBs to ACE inhibitors (I2=72%, p = 0.006). The pos- showed both diltiazem and felodipine led to a higher risk of MI sible reason might be that the CCB studied was of different (RR 1.18, 95% CI [0.96, 1.45]; RR 1.17, 95% CI [0.95, 1.44], kind in each trial. The three trials involving diuretics and β- respectively); but all the differences were insignificant. The five blockers respectively studied an aralkylamine derivative (vera- trials involving ACE inhibitors both studied DHPs, but three of pamil, CONVINCE), a benzothiazepine derivative (diltiazem, them gave patients amlodipine (AASK; ALLHAT; FACET) and NORDIL) and a DHP (felodipine, STOP-Hypertension-2). The the other two gave felodipine (STOP-Hypertension-2) or nisol- CONVINCE trial showed some benefits of verapamil for reduc- dipine (ABCD). The pooled RR for the trials camparint amlodip- ing the incidence of MI comparing to diuretics and β-blockers ine to and ACE inhibitor was 1.00 (95% CI [0.91, 1.10], I2=0%) (Analysis 2.2; Figure 4).

Figure 4. Forest plot of comparison: 2 Myocardial infarction, outcome: 2.2 Amlodipine vs. ACE inhibitor.

Stroke (non-fatal and fatal stroke) (Analysis 3.1; Figure 5)

Calcium channel blockers versus other classes of drugs for hypertension (Review) 10 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 5. Forest plot of comparison: 3 Stroke, outcome: 3.1 CCBs visus other classes of antihypertensive agents.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 11 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Incidence of stroke was not significantly different between CCB and diuretic groups (5 trials with 34,072 participants: RR 0.94, The reason for significant statistical heterogeneity between tri- 95% CI [0.84, 1.05], I2=0%), or between CCB and diuretic and als comparing CCBs to diuretics and β-blockers (I2=55%, p = β-blocker group (3 trials with 31,892 participants: RR 0.92, 95% 0.11) might be related to the type of CCBs, similar to what was CI [0.81, 1.03], I2=55%). Hypertensive patients treated with CCB explained in the MI results. It showed an insignificantly higher had a significantly lower risk of developing a stroke than those with stroke incidence in verapamil group compared to diuretics and a β-blocker (3 trials with 22,249 participants: RD -0.01; RR 0.77, β-blockers group (RR 1.14, 95% CI [0.89, 1.46]); with an in- 95% CI [0.67, 0.88], I2=0%), or an ACE inhibitor (5 trials with significantly lower stroke incidence in diltiazem (RR 0.82, 95% 24,006 participants: RD -0.01; RR 0.89, 95% CI [0.80, 0.98], I CI [0.67, 1.01]) and felodipine groups (RR 0.88, 95% CI [0.74, 2=40%), or an ARB (2 trials with 16,391 participants: RD -0.01; 1.05]) respectively. RR 0.85, 95% CI [0.73, 0.99], I2=53%). Congestive heart failure (Analysis 4.1; Figure 6)

Figure 6. Forest plot of comparison: 4 Congestive heart failure, outcome: 4.1 CCBs visus other classes of antihypertensive agents.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 12 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. There was no significantly difference in development of congestive heart failure between patients on CCBs and β-blockers (2 trials only required participants to have hypertension and cardiovascu- with 19,915 participants: RR 0.83, 95% CI [0.67, 1.04], I2=0%), lar risks. There was a significant increase of congestive heart fail- and between CCB and diuretic and β-blocker groups (3 trials ure events among the diabetic nephropathic patients treated with with 31,892 participants: RR 1.15, 95% CI [0.99, 1.33], I2= a CCB compared to an ARB (RR 1.58, 95% CI [1.17, 2.14]) 0%). However, the risk of developing congestive heart failure was (IDNT). markedly higher in patients on CCBs than those on diuretics (5 Cardiovascular mortality trials with 34,072 participants: RD 0.02; RR 1.37, 95% CI [1.25, Death caused by cardiovascular disease was added as a supplemen- 1.51], I2=17%), ACE inhibitors (4 trials with 23,626 participants: tal outcome different from the published protocol, since it was RD 0.01; RR 1.16, 95% CI [1.06, 1.27], I2=0%), and ARBs (2 judged to be important and was reported in most of the included trials with 16,391 participants: RD 0.01; RR1.20, 95% CI [1.06, trials. 1.36], I2=73%). Only a marginally significantly lower cardiovascular mortality was The lack of homogeneity between the two included trials com- shown in the CCBs group than in the β-blocker group (4 trials paring a CCB to an ARB may be due to the different inclusion with 44,825 participants: RR 0.90, 95% CI [0.81, 0.99], I2=62%, criteria of participants: the IDNT trial included hypertensive pa- RD -0.003. The other comparisons for this outcome were not tients with type 2 diabetic nephropathy, while the VALUE trial different: (Analysis 5.1; Figure 7)

Calcium channel blockers versus other classes of drugs for hypertension (Review) 13 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 7. Forest plot of comparison: 5 Cardiovascular mortality, outcome: 5.1 CCBs visus other classes of antihypertensive agents.

The heterogeneity among trials involving β-blockers (I2=62%, p= 0.05) might be caused by different types of study CCBs, which was a non-DHP in the INVEST trial (verapamil) but a DHP in the other three trials (amlodipine (AASK; ASCOT-BPLA) or lacidipine (ELSA)). After deselecting the INVESTtrial, the pooled RR was 0.77 (95% CI [0.66, 0.90], I2=0%) and the RD was -0.01, still showing a signiﬁcant decrease of cardiovascular mortality in the CCB group (Analysis 5.2; Figure 8).

Calcium channel blockers versus other classes of drugs for hypertension (Review) 14 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 8. Forest plot of comparison: 5 Cardiovascular mortality, outcome: 5.2 DHP vs. β-blocker.

Major cardiovascular events (MI, congestive heart failure, stroke and cardiovascular mortality) Compared to β-blockers, the CCBs significantly reduced major cardiovascular events (3 trials with 22,249 participants: RD -0.01; RR 0.84, 95% CI [0.77, 0.92], I2=0%). In contrast compared to the diuretics, CCBs significantly increased major cardiovascular events (4 trials with 33,642 participants: RD 0.01; RR 1.05 , 95% CI [1.00, 1.09], I2=0% p = 0.03). There was no significant difference on the effect of total major cardiovascular events comparing CCBs to diuretics or β-blockers (2 trials with 21,011 participants: RR 1.02, 95% CI [0.95, 1.10], I2=0%), or to ACE inhibitors (4 trials with 23,536 participants: RR 0.98, 95% CI [0.94, 1.02], I2= 56%).(Analysis 6.1; Figure 9). Major cardiovascular events could not be extracted from the trials involving ARBs.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 15 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Figure 9. Forest plot of comparison: 6 Major cardiovascular events, outcome: 6.1 CCBs visus other classes of antihypertensive agents.

Poor methodological quality of the FACET trial might contribute to the heterogeneity among the 4 trials comparing CCBs with ACE inhibitors. To undertake a sensitivity analysis on this effect, we deselected the FACET trial the results were unchanged, RR 0.97 (3 trials with 23,156 participants: 95% CI [0.91, 1.02], I2= 11%) (Analysis 6.2; Figure 10).

Figure 10. Forest plot of comparison: 6 Major cardiovascular events, outcome: 6.2 Sensitivity analysis: CCBs vs. ACE inhibitors.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 16 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Systolic and diastolic BP reduction Using the weighted mean-difference method and the ﬁxed effects model, we found that the mean systolic BP reduction of the CCB group was 0.81 mmHg, 95% CI [0.56, 1.06], less than that of the diuretic-based regimen group and 3.00 mmHg, 95% CI [2.59, 3.41] less than the diuretic-and-β-blocker-based regimen group: . Systolic BP reduction was 1.11 mm Hg 95% CI [-1.40, -0.82] more with CCB than with ACE inhibitor and 2.10 mmHg 95% CI [-2.46, -1.74] more than the ARB-based regimen group. There was no signiﬁcant difference between the CCB group and β-blocker group (p=0.38), or between the CCB group and α1-antagonist group (p=0.27). (Analysis 7.1; Figure 11)

Figure 11. Forest plot of comparison: 7 Blood pressure reduction, outcome: 7.1 Systolic blood pressure reduction.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 17 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. For diastolic BP,the mean reduction of the CCB group was -0.68 mmHg 95% CI [-0.84, -0.52] more than the diuretic group; - 0.63 mmHg 95% CI [-0.81, -0.44] more than the ACE inhibitor group, -1.70 mmHg 95% CI [-1.91, -1.49] more than the ARB group, and -1.20 mmHg 95% CI [-2.39, -0.01] more than the α1- antagonist group. Mean diastolic changes between CCB and the β-blocker groups were not signiﬁcantly different. (Analysis 7.2; Figure 12)

Figure 12. Forest plot of comparison: 7 Blood pressure reduction, outcome: 7.2 Diastolic blood pressure reduction.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 18 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. There was heterogeneity, I2 value 85%, for the 4 trials comparing systolic BP reduction with CCBs vs ACE inhibitors, however, there was no heterogeneity for the same comparison for diastolic BP reduction. The heterogeneity was most likely due to the poor methodological quality of the FACET trial. Sensitivity analyses, without the FACET trial, resulted in a homogeneous signiﬁcant mean differences for both systolic and diastolic BP,mean difference -1.00, 95% CI [-1.29, -0.70] and -0.62, 95% CI [-0.81, -0.44], respectively). (Analysis 7.3; Figure 13)

Figure 13. Forest plot of comparison: 7 Blood pressure reduction, outcome: 7.3 Sensitivity analysis: CCBs vs. ACE inhibitors.

line CCBs as compared to diuretics is explained by increased congestive heart failure events with CCBs. The RD for heart failure for the comparison of CCB and diuretics was 0.02 and is thus DISCUSSION clinically important and consistent with either a protective effect of diuretics or a harmful effect of CCBs for this outcome. The Summary of main results further finding that first-line CCBs also increased congestive heart After a systematic search and selection process according to the failure as compared to ACE inhibitors and ARBs suggests that this protocol, we included 18 RCTs with 141,807 participants, assess- is most likely due to a harmful effect of the CCBs with regard to ing cardiovascular outcomes and/or BP change among hyperten- this outcome. The only other significant differences found were sive patients. The two most important outcomes from the patients that first-line CCBs reduced stroke more than ACE inhibitors, perspective were total all cause mortality and total cardiovascular and reduced stroke and MI more than ARBs. events. This latter is important as it is a composite of the indi- This review was not designed to assess the effect of CCBs versus vidual outcomes stroke, MI and congestive heart failure. First-line placebo or no treatment, but other meta-analyses have reviewed CCBs were not significantly different from any of the other classes this and demonstrated that first-line CCBs reduce stroke and total of antihypertensive drugs for total mortality. First-line CCBs re- cardiovascular events, but have not been demonstrated to reduce duced total cardiovascular events as compared to β-blockers and total mortality or MI (BPLTTC 2000; Wright 2009). increased total cardiovascular events as compared to diuretics. The The results of the current review are consistent with findings reduction in total cardiovascular events as compared to β-blockers of the comparison between first-line β-blockers and first-line is explained by a significant reduction in stroke and cardiovascular CCBs in another Cochrane systematic review (Wiysonge 2007). mortality. The increased in total cardiovascular events for first- That review concluded that β-blockers reduced total cardiovas-

Calcium channel blockers versus other classes of drugs for hypertension (Review) 19 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. cular events significantly less than CCBs. A similar meta-analy- the evidence for non-DHPs inadequate. sis (Opie 2002) including 6 of our included trials (INSIGHT; Several trials and reviews mentioned that CCBs were a first-line MIDAS; NICS-EH; NORDIL; STOP-Hypertension-2; VHAS) choice for patients with hypertension and diabetes (Triggle 2007). concluded that mortality and major cardiovascular events with The ASCOT-BPLA study concluded that the amlodipine-based CCBs were similar to those seen with conventional first-line ther- regimen was better than atenolol-based regimen for patients with apy (diuretics or β-blockers). Other authors have claimed that new onset of diabetes. On the other hand Opie and colleagues CCBs are more effective than other treatments in decreasing the (Opie 2002) made the point that CCBs may be less safe than ACE risk of stroke for hypertensive patients (Angeli 2004; Verdecchia inhibitors in patients with diabetes, although with insufficient ev- 2005). This is consistent with our results which showed that stroke idence. Due to the impossibility to extract data in order to sep- events are significantly reduced by CCBs as compared to β-block- arately evaluate effects on hypertensive patients with diabetes in ers, ACE inhibitors and ARBs. It is possible that this is due to the our review, it is impossible to say whether CCBs have different greater BP lowering effect of CCBs as compared to ACE inhibitors effects in diabetic hypertensive patients. and ARBs as was found in this review, but it does not explain the difference for β-blockers, which did not have a different BP lowering effect. It has been hypothesized that CCBs might have anti- Potential biases in the review process atherosclerotic actions that could be helpful in reducing stroke as The included trials have differences in designs and methods, base- well (Angeli 2004). line and goal BPs, study populations and drugs, so putting their With regard to congestive heart failure, CCBs significantly increase data together to get a conclusion may have some limitations. For the risk, as compared to diuretics, ACE inhibitors and ARBs. This example, CCBs are a heterogeneous group of drugs that can be finding is consistent with those of other reviews (Opie 2000; Black subclassified into the DHPs and the non-DHPs. The different 2004). Since CCBs did not have any other advantages as compared classes have different in binding sites on the calcium channel pores to diuretics this would suggest that diuretics are preferred first-line and thus could have different effects (Opie 2000; Triggle 2007). drugs over CCBs for most patients with hypertension. In the current review, we did not focus on the comparison among BPs decreased in all treatment arms of all the included trials, different types of CCBs, but it might not be appropriate to com- but mean BP reduction differed. First-line CCB-based regimens bine them in a meta-analysis. The high I2 values for pooled tri- lowered systolic BP less than first-line diuretic-based regimens als involving both DHPs and non-DHPs, 72% for three trials and conventional treatment based regimens. In contrast first-line (CONVINCE; NORDIL; STOP-Hypertension-2) assessing MI CCBs lowered diastolic BP better than diuretic-based regimens. events) are consistent with this possibility. However, in this case First-line CCB based regimens also lowered both systolic and di- dividing the trials into DHPs and non-DHPs does not explain astolic BPs more than ACE inhibitors and ARBs. This could be a the heterogeneity. Likewise, heterogenous populations in the in- partial explanation for the differences in stroke outcomes as dis- cluded trials might heterogeneity of the effect. In the present re- cussed above. view, patients enroled included those with diabetes, cardiovascular disease, kidney disease or other conditions. It was not possible to investigate the effect of these subgroup populations on the effect Overall completeness and applicability of size. In general there was excellent homogeneity of most effects evidence as shown by an I2 value of 0%, and only a few outcomes were Most of the included trials with the exception of TOMHS reported associated with I2 >50% causing us to believe that the conclusions relevant hypertension outcomes, but not all the desired outcomes of our review are overall valid. were available from each trial. Furthermore, supplemental inclu- Although the benefits of BP lowering for prevention of CVD sion criteria were required in several trials, and most trials were are well established (Ezzati 2002; BPLTTC 2003; Wright 2009), event-driven hypertension studies, which meant that the included which antihypertensive drug class should be prescribed first is still participants tended to have more complicated hypertension or somewhat controversial. However, more than one antihyperten- advanced disease (Zanchetti 2005). Patients at the two extremes, sive agents need to be given to many patients to achieve the BP goal those with uncomplicated hypertension at one extreme and those (Chobanian 2003; Mancia 2007; Haller 2008). This fact leads with severe or acute hypertension, and secondary hypertension at to another limitation in the review and perhaps its major weak- the other extreme were not included in the present analysis. ness. Since additional antihypertensive agents other than first-line Although 18 studies with a large number of participants com- drugs were administered sequentially to reach BP goals in most paring several classes of first-line antihypertensive drugs were in- original trials, results may have been confounded, although they cluded, the number of trials for each of the subgroups was limited. were presumed to reflect effect of the first drug. Only one small Because of this data were insufficient for some comparisons. This included trial (NICS-EH) in our review prohibited any other an- was particularly the case for ARBs and α1-antagonists. Further- tihypertensive drugs, and it concluded that the CCB and diuretic more, most of the CCBs included were dihydropyridines, leaving groups had a similar decrease of BPs and cardiovascular events. BP

Calcium channel blockers versus other classes of drugs for hypertension (Review) 20 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. differences between different classes of drugs could have an impact whether CCBs, ACE inhibitors or ARBs should be prescribed sec- on outcomes (Staessen 2003; Wright 2009), which is a further ond, but does provide evidence supporting choosing CCBs over limitation of this type of review. β-blockers. It should be noted that many of the differences found We have tried as much as possible to reduce the risk of attrition in the current review are not robust and further trials might change bias by reporting on the intent-to-treat population as much as the conclusions. Therefore it will be important to follow the re- possible. We do not think publication bias is likely as we have done search in this field closely and update this review when new data an extensive search of the pertinent literature, including published becomes available. and unpublished studies, without any language restrictions. Implications for research More well-designed RCTs studying the morbidity and mortality of first-line CCBs, compared with ACE inhibitors or ARBs are needed, especially for patients with co-morbidities such as dia- AUTHORS’ CONCLUSIONS betes. These trials must avoid confounding factors as much as possible, such as by making sure the secondary drugs added to each Implications for practice arm of the trial are the same. It is important that all relevant outcomes are well defined and reported. New RCTs comparing first- First-line CCBs do not affect total mortality as compared to the line DHP CCBs with non-DHP CCBs would be useful to deter- other first-line antihypertensive drug classes. First-line CCBs re- mine whether it is appropriate to combine subclasses of CCBs. duce total cardiovascular events, stroke and cardiovascular mortality as compared to first-line β-blockers. First-line CCBs increase total cardiovascular events and congestive heart failure events as compared to diuretics. First-line CCBs reduce stroke as compared ACKNOWLEDGEMENTS to ACE inhibitors and reduce stroke and MI as compared to ARBs, but they increase congestive heart failure events as compared to We would like to acknowledge the original authors of this both ACE inhibitors and ARBs. Therefore based on this review Cochrane protocol (Onder G, Furberg CD, Moore A, Psaty BM, diuretics are the preferred first-line drug class to optimize the re- Pahor M) who identified the topic and contributed extensively to duction in cardiovascular events. The review does not distinguish the background.

REFERENCES

References to studies included in this review JAMA 2002;288:2421–31.

ABCD {published data only} AASK {published data only} Estacio R, Jeffers B, Hiatt W, et al.The effect of nisoldipine Agodoa L, Appel L, Bakris G, et al.Effect of Ramipril as compared with enalapril on cardiovascular outcomes vs Amlodipine on RenalOutcomes in Hypertensive in patients with non-insulin-dependent diabetes and Nephrosclerosis: A Randomized Controlled Trial. JAMA hypertension. N Engl J Med 1998;338:645–52. 2001;285(21):2719–28. Savage S, Johnson Nagel N, Estacio RO, et al.The ABCD Kusek J, Lee J, Smith D, et al.Effect of Blood Pressure (AppropriateBlood Pressure Control in Diabetes) trial: Control andAntihypertensive Drug Regimen on Qualityof rationale and design of a trial of hypertension control Life: The African American Study of KidneyDisease and (moderate or intensive) in type II diabetes. Online J Curr Hypertension (AASK) Pilot Study. Control Clin Trials 1996; Clin Trials 1993 Nov 24;Doc No 104:6250 words; 128 16:40S–46S. paragraphs. Norris K, Bourgoigne J, Gassman J, et al.Cardiovascular Schrier R, Estacio R. Additional Follow-Up from the ABCD Outcomes in the African American Study of Kidney Disease Trial in Patients with Type 2 Diabetes and Hypertension. N and Hypertension (AASK) Trial. Am J Kidney Dis 2006;48 Engl J Med 2000;343:1969. (5):739–51. Wrght J, Kusek J, Toto R, et al.Design and baseline ALLHAT {published data only} characteristics of participants in the African American Study Davis B, Cutler J, Gordon D, et al.Rationaleand design of Kidney disease and hypertension (AASK) pilot study. for the Antihypertensive and Lipid-Lowering Treatment Control Clin Trials 1996;17(4 Suppl):3S–16S. to Prevent Heart Attack Trial (ALLHAT). Am J Hypertens Wright J, Bakris G, Greene T, et al.Effect of blood pressure 1996;9:342–60. lowering and antihypertensive drug class on progression of The ALLHAT Ofﬁcers and Coordinators for theALLHAT hypertensive kidney disease: results from the AASK trial. Collaborative Research Group. Major outcomes

Calcium channel blockers versus other classes of drugs for hypertension (Review) 21 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. in moderately hypercholesterolemic, hypertensive patient characteristics. Nephro Dial Transplant 2000;15: patients randomized to pravastatin vs usual care: the 487–97. Antihypertensive and Lipid-Lowering Treatment to Prevent INSIGHT {published data only} Heart Attack Trial. JAMA 2002;288:2998–3007. Brown M, Palmer C, Castaigne A, et al.Morbidity and The ALLHAT Ofﬁcers and Coordinators for the ALLHAT mortality in patients randomised to double-blind treatment Collaborative Research Group. Major outcomes in high-risk with a long-acting calcium-channel blocker or diuretic in hypertensive patients randomized to angiotensin-converting the International Nifedipine GITS study: Intervention as a enzyme inhibitor or calcium channel blocker vs diuretic:The Goal inHypertension Treatment (INSIGHT). Lancet 2000; Antihypertensive and Lipid-Lowering Treatment to Prevent 356:366–72. Heart Attack Trial (ALLHAT). JAMA 2002;288:2981–97. INVEST {published data only} ASCOT-BPLA {published data only} Pepine C, Handberg E, Cooper-DeHoff R, et al.A calcium Dahlöf B, Sever P, Poulter N, et al.Prevention of antagonist vs a non-calcium antagonist hypertension cardiovascular events with an antihypertensive regimen of treatment strategy for patients with coronary artery disease. amlodipine adding perindopril as required versus atenolol The international Verapamil-Trandolapril Study (INVEST): adding bendroﬂumethiazide as required, in the Anglo- a randomized controlled trial. JAMA 2003;290:2805–16. ScandinavianCardiac Outcomes Trial-Blood Pressure Pepine C, Handberg-Thurmond E, Marks R, et al.Rationale Lowering Arm (ASCOTBPLA): a multicentre randomised and design of the International Verapamil SR/Trandolapril controlled trial. Lancet 2005;366:895–906. Study (INVEST). J Am Coll Cardiol 1998;32:1228–37. Sever P, Dahlöf B, Poulter N, et al.Rationale, design, MIDAS {published data only} methods and baseline demography of participants of the Borhani N, Mercuri M, Borhani P, et al.Final outcome Anglo-ScandinavianCardiac Outcomes Trial. J Hypertens results of the Multicenter Isradipine Diuretic Atherosclerosis 2001;6:1139–47. Study (MIDAS): a randomized controlled trial. JAMA CONVINCE {published data only} 1996;276:785–91. Black H, Elliott W, Grandits G, et al.Principal Results Borhani NO, Miller ST, BruggerSB, et al.MIDAS: of the Controlled OnsetVerapamil Investigation of hypertension and atherosclerosis, a trial of the effect of CardiovascularEnd Points (CONVINCE) Trial. JAMA antihypertensive drug treatment on atherosclerosis. J 2003;289:2073–82. Cardiovasc Pharmacol 1992;19(suppl 3):16–20. Black HR, Elliott W, Neaton J, et al.Baseline characteristics Furberg C, Byington R, Borhani N. Designfeatures: and early blood pressure control in the CONVINCE trial. Multicenter Isradipine Diuretic AtherosclerosisStudy Hypertension 2001;37:12–8. (MIDAS). Am J Med 1989;86(suppl 4A):37–9. Black HR, Elliott WJ, Neaton JD, et al.Rationale and NICS-EH {published data only} design for the Controlled Onset Verapamil Investigation Kuramoto K, and the National Intervention Cooperative of Cardiovascular Endpoints (CONVINCE) Trial. Study Group. Treatment of elderly hypertensive in Japan: ControlledClinical Trials 1998;19:370–390. National Intervention Cooperative Study in Elderly ELSA {published data only} Hypertensives. J Hypertens 1994;12(suppl6):S35–S40.. Zanchetti A, Bond M, Hennig M, et al.Calcium antagonist National Intervention Cooperative Study in Elderly lacidipine slows down progression of asymptomatic carotid Hypertensives Study Group. Randomized double-blind atherosclerosis: principal results of the European Lacidipine comparison of a calcium antagonist and a diuretic in elderly Study on Atherosclerosis(ELSA), a randomized, double- hypertensives. Hypertension 1999;34:1129–33. blind, long-term Trial. Circulation 2002;106:2422–7. NORDIL {published data only} FACET {published data only} Hansson L, Hedner T, Lund-Johansen P, et al.Randomised Tatti R, Guarisco R, Pahor M, et al.Outcome results of trial of effects of calcium antagonists compared with the Fosinopril Versus Amlodipine Cardiovascular Events diuretics and β-blockers on cardiovascular morbidity Randomized Trial (FACET) in patients with hypertension and mortality in hypertension: the Nordic Diltiazem and NIDDM. Diabetes Care 1998;21:597–603. (NORDIL) study. Lancet 2000;356:359–65. IDNT {published data only} Hedner T. Progress report on the Nordic diltiazem study Berl T, Hunsicker L, Lewis J, et al.Cardiovascular outcomes (NORDIL): an outcome study in hypertensive patients. in the Irbesartan Diabetic Nephropathy Trial of patients Blood press 1999;8(5-6):296–9. with type 2 diabetes and overt nephropathy. Ann Intern NORDIL Study Group. The Nordic Diltiazem Study Med 2003;138:542–9. (NORDIL). A Prospective Intervention Trial of of Calcium Lewis E, Hunsicker L, Clarke W, et al.Renoprotective effect Antagonist Therapy in Hypertension. Blood Press 1993;2: of angiotensin-receptor antagonist irbesartan in patients 312–21. with nephropathy due to type 2 diabetes. N Engl J Med SHELL {published data only} 2001;245:851–60. Malacco E, Mancia G, Rappelli A, et al.Treatment of Rodby R, Rohde R, Clarke W, et al.The Irbesartan Type Isolated Systolic Hypertension: The SHELL Study Results. II Diabetic Nephropathy Trial: study design and baseline Blood Pres 2003;12:160–7.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 22 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. STOP-Hypertension-2 {published data only} Bakris 1996 {published data only} Dahlöf B, Hansson L, Lindholm L, et al.STOP- Bakris G, Copley J, Vicknair N, et al.Calcium channel Hypertension-2: a prospective intervention trial of “newer” blockers versus other antihypertensive therapies on versus“older” treatment alternatives in old patients with progression of NIDDM associated nephropathy. Kidney Int hypertension. Blood Press 1993;2:136–41. 1996;50(5):1641–50. Hansson L, Lindholm L, Ekbom T, et al.Randomised trial Bakris 1997 {published data only} of old and new antihypertensive drugs in elderly patients: Bakris G, Mangrum A, CoPLEY J, et al.Effect of calcium cardiovascular mortality and morbidity the Swedish Trial in channel or β-blockade on the progression of diabetic Old Patients with Hypertension-2 study. Lancet 1999;354: nephropathy in African Americans. Hypertension 1997;29: 1751–56. 744–50. Lindholm L, Hansson L, Dahlöf B, et al.The Swedish trial DHCCP {published data only} in Old Patients with Hypertension (STOP-Hypertension- Bulpitt C, Palmer A, Beevers D, et al.Calcium channel 2): a progress report. Blood Press 1996;5:300–4. blockers and cardiac mortality in the treatment of TOMHS {published data only} hypertension: a report from the Department of Health Mascioli S, Grimm R, Neaton J, et al.TheTreatment of Hypertension Care Computing Project (DHCCP). J Hum Mild Hypertension Study (TOMHS): characteristics Hypertens 1997;11:205–11. of participants at baseline. Am J Cardiol 1990;66(9): FACTS {published data only} 32C–35C. Pahor M, Franse L, Deithcher S, et al.Fosinopril versus Neaton J, Grimm R, Prineas R, et al.Treatment of Mild Amlodipine Comparative Treatments Study: a randomized Hypertension Study: ﬁnal results. JAMA 1993;270: trial to assess effects on plasminogen activator inhibitor-1. 713–24. Circulation 2002;105:457–61. Stamler J, Prineas R, Neaton J, et al.Background and Gottdiener 1997 {published data only} design of the new US trial on diet and drug treatment of Gottdiener J, Reda D, Massie B, et al.Effect of single-drug ’mild’ hypertension (TOMHS). Am J Cardiol 1987;59(14): therapy on reduction of left ventricular mass in mild to 51G–60G. moderate hypertension. Circulation 1997;95:2007–14. The Treatment of Mild Hypertension Research Group. The Treatment of Mild Hypertension Study. Arandomized, HOT {published data only} placebo-controlled trial of a nutritional-hygienic regimen Hansson L, Zanchetti A, Carruthers S, et al.Effects of along with various drug monotherapies. Arch Intern Med intensive blood-pressure lowering and low-dose aspirin 1991;151:1413–23. inpatients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. VALUE {published data only} Lancet 1998;351:1755–62. Julius S, Kjeldsen S, Weber M, et al.Outcomes in hypertensive patients at high cardiovascular risk treated with Kes 2003 {published data only} regimens based on valsartan or amlodipine: the VALUE Kes S, Caglar N, Canberk A, et al.Treatment of mild-to- randomised trial. Lancet 2004;363(9426):2022–31. moderate hypertension with calcium channel blockers: a Mann J, Julius S. The Valsartan Antihypertensive Long- multicentre comparison of once-daily nifedipine GITS term Use Evaluation (VALUE) trial of cardiovascular events with once-daily amlodipine. Curr Med Res Opin 2003;19: in hypertension. Rationale and design. Blood Press 1998;7 226–37. (3):176–83. Maharaj 1992 {published data only} Maharaj B, van der Byl K. A comparison of the acute VHAS {published data only} hypotensive effects of two different doses of nifedipine. Am Rosei E, Palù C, Leonetti G, et al.Clinical results of the Heart J 1992;124(3):720–5. Verapamil in Hypertension and Atherosclerosis Study. J Hypertens 1997;15:1337–44. Pahor 1995 {published data only} Zanchetti A, Rosei E, Palù C, et al.The Verapamil in Pahor M, Guralnik J, Corti C, et al.Long-term survival and Hypertension and Atherosclerosis Study (VHAS): results use of antihypertensive medications in older persons. J Am of long-term randomized treatment with either verapamil Geriatr Soc 1995;43(11):1191–7. or chlorthalidone on carotid intima-media thickness. J Papademetriou 1997 {published data only} Hypetens 1998;16:1667–76. Papademetriou V, Gottdiener F, Narayan P, et al.Hydrochlorothiazide is superior to Isradipine for References to studies excluded from this review reduction of left ventricular mass: results of a multicenter trial. J Am Coll Cardiol 1997;30:1802–8. Abascal 1998 {published data only} PRESERVE {published data only} Abascal V, Larson M, Evans J, et al.Calcium antagonists and Devereux R, Palmieri V, Sharpe N, et al.Effects of once- mortality risk in men and women with hypertension in the daily angiotensin-converting enzyme inhibition and Framingham Heart Study. Arch Intern Med 1998;158(17): calcium channel blockade-based antihypertensive treatment 1882–6. regimens on left ventricular hypertrophy and diastolic ﬁlling

Calcium channel blockers versus other classes of drugs for hypertension (Review) 23 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. in hypertension. The Prospective Randomized Enalapril BPLTTC 2000 Study Evaluating Regression of Ventricular Enlargement Blood Pressure Lowering Treatment Trialists’ Collaboration. (PRESERVE) trial. Circulation 2001;104:1248–54. Effects of ACE inhibitors, calcium antagonists, and other Psaty 1995 {published data only} blood pressure-lowering drugs: Results of prospectively Psaty BM, Heckbert SR, Koepsell TD, et al.The risk of designed overviews of randomized trials. Lancet 2000;355: myocardial infarction associated with antihypertensive drug 1955–64. therapies. JAMA 1995;274:620–5. BPLTTC 2003 Radevski 1999 {published data only} Blood Pressure Lowering Treatment Trialists’ Collaboration. Radevski I, Skudicky D, Candy G. Antihypertensive Effectsof different blood pressure lowering regimens monotherapy with nisoldipine CC is superior to enalapril on major cardiovascular events: results of prospectively in black patients with severe hypertension. Am J Hypertens designed overviews of randomised trials. Lancet 2003;362: 1999;12:194–203. 1527–35. Schneider 1991 {published data only} Chobanian 2003 Schneider E, Jennings A, Opie L. Captopril, nifedipine and Chobanian AV, Bakris GL, Black HR, et al.Seventh report their combination for therapy of hypertensive urgencies. S of the Joint National Committee on Prevention, Detection, Afr Med J 1991;80(6):265–70. Evaluation, and Treatment of High Blood Pressure. Hypertension 2003;42:1206–52. STONE {published data only} Gong L, Zhang W, Zhu Y, et al.Shanghai trial of nifedipine Ezzati 2002 in the elderly (STONE). J Hypertens 1996;14:1237–45. Ezzati M, Lopez AD, Rodgers A, et al.Selected major risk factors and global and regional burden of disease. Lancet Syst-China {published data only} 2002;360:1347–60. Liu L, Wang J, Gong L, et al.Comparison of active treatment and placebo in older Chinese patients with isolated systolic Fergusson 2002 hypertension. Systolic Hypertension in China (Syst-China) Fergusson D, Aaron S, Guyatt G, Hébert P. Post- Collaborative Group. J Hypertens 1998;16:1823–9. randomisation exclusions: the intention to treat principle and excluding patients from analysis. BMJ 2002;325: Syst-Eur {published data only} 652–4. Amery A, Birkenhäger W, Bulpitt C, et al.Syst-Eur: a Haggins 2008 - Chap. 16 multicentre trial on the treatment of isolated systolic Higgins JPT, Deeks JJ, Altman DG (editors). Chapter hypertension in the elderly:objectives, protocol, and 16: Special topics in statistics. In: Higgins JPT, Green organization. Aging Clin Exp Res 1991;3:287-302. S editor(s). Cochrane Handbook for Systematic Reviews of Staessen J, Fagard R, Thijs L, et al.Randomised double- Interventions Version 5.0.1 (updated September 2008). The blind comparison of placebo and active treatment for older Cochrane Collaboration, 2008. patients with isolated systolic hypertension. Lancet 1997; 350:757–64. Haggins 2008 - Chap. 8 Stanessen J, Thijs L, Fagard R, et al.Calcium channel Higgins JPT, Altman DG (editors). Chapter 8: Assessing blockade and cardiovascular prognosis in the European trial risk of bias in included studies. In: Higgins JPT, Green on isolated systolic hypertension. Hypertension 1998;32: S editor(s). Cochrane Handbook for Systematic Reviews of 410–6. Interventions Version 5.0.1 (updated September 2008).. The Cochrane Collaboration, 2008. References to studies awaiting assessment Haller 2008 J-MIC(B) {published data only} Haller H. Effective management of hypertension Yui Y. Long-term effects of nifedipine retard vs ACE with dihydropyridine calcium channel blocker-based inhibitors in hypertension with coronary artery disease: combination therapy in patients at high cardiovascular risk. ﬁnal reports of J-MIC(B). Circ J 2002;66(suppl 1):357. Int J Clin Pract 2008;62:781–90. Hansson 1992 Additional references Hansson L, Hedner T, Dahlöf B. Prospective Randomized Open Blinded End-point (PROBE) study. Blood Press 1992; Angeli 2004 1:113–9. Angeli F, Verdecchia P, Reboldi GP, et al.Calcium channel blockade to prevent stroke in hypertension: A meta-analysis HOPE 2000 of 13 studies with 103,793 subjects. Am J Hypertens 2004; The Heart Outcomes Prevention Evaluation Study 17:817–22. Investigators. Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk Black 2004 patients. N.Engl.J.Med 2000;342:145–53. Black HR. Calcium Channel Blockers in theTreatment of Hypertension andPrevention of Cardiovascular Disease: MacMahon 2000 Results from Major Clinical Trials. Clinical Cornerstone MacMahon S, Neal B. Differences between blood-pressure- 2004;6(4):53–66. lowering drugs. Lancet 2000;356:352–3.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 24 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Mancia 2007 regimens on major cardiovascular events: results of Mancia G, De Backer G, Dominiczak A, et al.2007 prospectively-designed overviews of randomised trials. Guidelines for the management of arterial hypertension: Lancet 2003;362(9395):1527–35. The Task Force for the Management of Arterial Verdecchia 2005 Hypertension of the European Society of Hypertension Verdecchia P, Reboldi G, Angeli F, et al.Angiotensin- (ESH) and of the European Society of Cardiology (ESC). Converting Enzyme Inhibitors and Calcium Eur Heart J 2007;28(12):1462–536. ChannelBlockers for Coronary Heart Disease and Stroke Mantel 1959 Prevention. Hypertension 2005;46:386–92. Mantel N, Haenszel W. Statistical aspects of the analysis of Wiysonge 2007 data from retrospective studies of disease.. J.Natl.Cancer Inst Wiysonge CSU, Bradley HA, Mayosi BM, et al.Beta- 1959;22:719–48. blockers for hypertension. Cochrane Database of Opie 2000 Systematic Reviews 2007, Issue 1. [DOI: 10.1002/ Opie LH, Yusuf S, Kübler W. Current status of safety 14651858.CD002003.pub2] and efficacy of calcium channel blockers in cardiovascular diseases: a critical analysis based on 100 studies. Progress in Wright 1999 Cardiovascular Diseases 2000;43:171–96. Wright JM, Lee CH, Chambers GK. Systematic review of anti-hypertensive therapies: Does the evidence assist in Opie 2002 choosing a first-line drug?. Can Med Assoc J 1999;161: Opie LH, Schall R. Evidence-based evaluation of calcium 25–32. channel blockers for hypertension: Equality of mortality and cardiovascular risk relative to conventional therapy. J Wright 2009 Am Coll Cardiol 2002;39:315–22. Wright JM, Musini VM. First-line drugs for hypertension. Cochrane Database of Systematic Reviews 2009, Issue 3. Psaty 1997 [DOI: 10.1002/14651858.CD001841.pub2] Psaty B, Smith N, Siscovick D, et al.Health outcomes associated with antihypertensive therapies used as first-line Zanchetti 2005 agents. A systematic review and meta-analysis. JAMA 1997; Zanchetti A. Evidence-based medicine in hypertension: 277:739–45. what type of evidence?. J Hypertens 2005;23:1113–20. Staessen 2003 References to other published versions of this review Staessen JA, Wang JG, Thijs L. Cardiovascular prevention and blood pressure reduction: a quantitative overview. J Hypertens 2003;21:1055–76. Pahor 2000 Pahor M, Psaty BM, Alderman MH, Applegate WB, Triggle 2007 Williamson JD, Cavazzini C, Furberg CD. Health outcomes Triggle DJ. Calcium channel antagonists: Clinical uses-Past, associated with calcium antagonists compared with other present and future. Biochemical Pharmacology 2007;74:1–9. first-line antihypertensive therapies: a meta-analysis of Turnbull 2003 randomised controlled trials. Lancet 2000;356(9246): Turnbull F, Blood Pressure Lowering Treatment Trialists’ 1949–54. [MEDLINE: 11130522] Collaboration. Effects of different blood-pressure-lowering ∗ Indicates the major publication for the study

Characteristics of included studies [ordered by study ID]

AASK

Methods Long-term, multicenter, randomized controlled, double-blind trial, using a 3 × 2 factorial design. Because of acute changes in glomerular ﬁltration rate (GFR), the amlodipine arm was halted about one year ahead of the protocol, while most of patients in other groups were followed to the planned end, giving a median follow-up of 4.3 years to the cardiovascular composite outcome. All analyses were intent-to-treat

Participants Participants (N=1094) were self-identiﬁed African Americans with hypertension (diastolic BP was 95mm Hg or higher), aged 18 to 70 years, with GFR between 20 to 65 mL/min per 1.73 m2.

Interventions Participants were equally randomized to one of two blood pressure goals (usual MAP goal of 102 to 107 mm Hg (N=554), or a lower MAP goal of 92 mm Hg (N=540) ), and to treatment with 1 of 3 antihypertensive drugs (metoprolol, 50 to 200 mg/d (N=441); ramipril, 2.5 to 10 mg/d (N=436); or a dihydropyridine CCB amlodipine, 5 to 10 mg/d (N=217), using a 2:2:1 randomization ratio). And additional open-labeled antihypertensives were added if the BP goal could not be achieved by the randomized drug

Outcomes Rate of change in GFR and other renal outcomes, and all cardiovascular events including cardiovascular deaths and hospitalizations for MI, strokes, heart failure, revascularization procedures, and other hospitalized cardiovascular events were reviewed

Notes It was carried out at 11 clinical centers in the USA.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes The randomization was stratiﬁed by city using randomly permuted blocks and the Data Coordinating Center (DCC) performed randomization centrally

Allocation concealment? Unclear Method of allocation concealment was not described.

Blinding? Yes Participants and investigators were masked All outcomes to randomized drug but not BP goal

Incomplete outcome data addressed? Yes Missing data were equal among the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

ABCD

Methods A prospective, controlled, randomized, double-blinded trial with a long-term follow-up of more than 5 years. Cardiovascular end points were analyzed using the intention-to- treat principle

Participants Enrolled patients were diagnosed with NIDDM, and had diastolic BP of 80 mm Hg or higher and were receiving no antihypertensive medications at the time of randomization. The current systematic review only focuses on the hypertensive cohort (N=470) (mean base-line diastolic BP ≥90 mm Hg)

Interventions Patients randomized to active study medication received either nisoldipine (N=235) (10 mg per day, with increases to 20, 40, and 60 mg per day, plus placebo for enalapril) or enalapril (N=235) (5 mg per day, with increases to 10, 20, and 40 mg per day, plus placebo for nisoldipine). Open-label antihypertensive medications except the study drugs were added when necessary

Outcomes Cardiovascular outcomes including death due to cardiovascular events, nonfatal MI, nonfatal cerebrovascular accident, heart failure requiring hospital admission and pul- monary infarction were reviewed

Notes Results at the end of the planned 5-year follow-up were reported in 1998, and additional follow-up results were described in 2000, both in the New England Journal of Medicine

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes Study drug in each group plus placebo for All outcomes the other study drug were administered in a double-blind manner

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

ALLHAT

Methods A randomized, double-blind, multicenter clinical trial with a large sample size and long follow-up (with a mean length of 4.9 years (SD 1.4 years)). Visit adherence decreased over time from about 92% at 1 year to 84% to 87% at 5 years in all 3 treatment groups. Intent-to-treat analysis was used

Participants Participants (N=33357) were men and women aged 55 years or older who had stage 1 or stage 2 hypertension with at least 1 additional risk factor for CHD events

Interventions Treatment with the study drug was initiated the day after randomization. Participants were randomly assigned to chlorthalidone, amlodipine, or lisinopril in a ratio of 1.7:1:1, which meant 15255, 9048 and 9054 patients were enrolled in the 3 groups respectively. Goal BP was less than 140/90 mm Hg achieved by titrating the assigned study drug, and additional open-label agents were allowed if necessary

Outcomes The primary outcome was fatal CHD or nonfatal MI combined, and secondary outcomes included all-cause mortality, stroke, combined CHD and combined CVD

Notes It sponsored by the National Heart, Lung, and Blood Institute and carried in 623 centers in the USA, Canada, Puerto Rico, and the US Virgin Islands

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Randomization scheme was generated by computer.

Allocation concealment? Yes It speciﬁed that the concealed randomization scheme was implemented at the clinical trials center and stratiﬁed by center

Blinding? Yes Study drugs were encapsulated and identi- All outcomes cal in appearance.

Incomplete outcome data addressed? Yes Two sites and their patients originally re- All outcomes ported were excluded, which might impact on the results, although it was because of their poor documentation of informed consent. But an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

ASCOT-BPLA

Methods An independent, investigator-initiated, investigator-led, multicentre, prospective, randomised controlled trial, with 5.5 years’ median follow-up. It compared the time to ﬁrst event on an intention-to-treat basis

Participants A total of 19257 patients aged 40-79 years were recruited; all of them had either untreated hypertension (systolic BP ≥160 mm Hg, diastolic ≥100 mm Hg, or both) or treated hypertension with systolic BP≥ 140 mm Hg, diastolic ≥90 mm Hg, or both, and they had to have at least three other cardiovascular risk factors

Interventions Participants were randomised to CCB-based regimen (amlodipine 5-10 mg; N=9639) or β-blocker-based regimen (atenolol 50-100 mg; N=9618). Additional antihypertensive agents were given in both groups according to a prespeciﬁed algorithm: perindopril 4- 8 mg was added to amlodipine-based group as required; bendroﬂumethiazide 1.25-2.5 mg was added to atenolol-based group as required

Outcomes Primary endpoints: non-fatal MI + fatal CHD; Secondary endpoints: all-cause mortality, total stroke, primary endpoint minus silent MI, all coronary events, total cardiovascular events and procedures, cardiovascular mortality, and non-fatal and fatal heart failure Tertiary endpoints: silent MI, unstable angina, chronic stable angina, peripheral arterial disease, life-threatening arrhythmias, development of diabetes, development of renal impairment

Notes Patients were recruited between February, 1998, and May, 2000, in the UK, Ireland and the Nordic countries

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes The randomisation was a computer generated optimum allocation

Allocation concealment? Yes The allocation was blinded for any person involved in the undertaking of the study

Blinding? Yes A PROBE design was used. All outcomes

Incomplete outcome data addressed? Yes Two centres with 85 patients were excluded All outcomes after randomisation, but missing data were equal between the treatment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

CONVINCE

Methods Prospective, double-blind, randomized, active-controlled, multicenter, international clinical trial with a mean follow-up of 3 years

Participants A total of 16602 participants with hypertension and 1 or more additional risk factors for CVD were enrolled, but 126 of them were excluded because of data integrity, so ﬁndings from 16476 subjects were reported

Interventions Patients were administered standard-of-care drug (β-blocker or diuretic) chosen by the investigator prior to randomization. Then they were randomized to verapamil (N=8241) (begin with 180 mg daily, increased dose or added other drugs when necessary) group or standard-of-care regimen group (N=8361) (β-blocker or diuretic)

Outcomes Effect in preventing AMI, stroke, or cardiovascular disease related death, and all-cause mortality

Notes It was conducted in 661 clinical sites in 15 countries; the sponsor closed the study 2 years shorter than the planned 5-year follow-up because of commercial reasons

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes A simple automated system, the Interactive Voice Response System (IVRS), for randomizing, assigning, and tracking blinded medication was used

Allocation concealment? Yes The same to the above reason.

Blinding? Yes Labeled bottles with active drug or placebo All outcomes were given to participants, and the content was blinding

Incomplete outcome data addressed? Yes Participants from 2 sites (n=126; 62 ran- All outcomes domized to COER verapamil) were ex-

cluded because of data integrity concerns, and its impact on results was unclear, but intention-to-treat analysis was performed in this systematic review

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Unclear Withdraw between groups was imbal- anced (115 with verapamil, while 207 with atenolol or hydrochlorothiazide)

ELSA

Methods A randomized, double-blind, multicenter trial with a mean follow-up of 3.75 years. Fourty nine lacidipine and 43 atenolol patients lost to follow-up, and the intention-to- treat analysis was performed

Participants All enrolled participants (N=2334) were aged 45 to 75 years with sitting systolic BP of 150 to 210 mm Hg and diastolic BP of 95 to 115 mm Hg

Interventions Patients were randomized to receive either lacidipine 4 mg once daily (N=1177) or atenolol 50 mg once daily (N=1157). If diastolic BP goal was not achieved, the dose of lacidipine could be increased to 6 mg, and atenolol could be increased to 100 mg (month 1), with open-label hydrochlorothiazide added (12.5 mg daily month 3 and 25 mg daily month 6)

Outcomes Change in mean maximum intima-media thickness (IMT), proportion of patients with an increase or decrease in plaque number, incidence of cardiovascular events and total mortality

Notes It was conducted in 410 clinical units in France, Germany, Greece, Italy, Spain, Sweden, and the United Kingdom

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Randomization was computer-generated.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes Patients and study personnel, excluding the All outcomes Safety Committee, were blinded to treatment assignment for the study duration

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

FACET

Methods An open-label, randomized prospective trial and participants were followed for up to 3.5 years. All analyses were intention-to-treat, unless otherwise stated

Participants Patents (N=380) with a diagnosis of NIDDM and hypertension (systolic BP more than 140 mmHg or diastolic BP more than 90 mmHg)

Interventions Participants were randomly assigned to open-label fosinopril (20 mg/day) (N=189) or amlodipine (10 mg/day) (N=191). The other study drug was added when necessary

Outcomes Serum lipids and diabetes control, Cardiovascular events, BP control and renal function status

Notes Patients were recruited from an outpatient diabetes clinic in Marino, Italy

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Random number sequence were generated in a computer.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? No Study drugs were administered open-label. All outcomes

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? No When BP was not controlled well on monotherapy, the other study drug was added. It could impact on the evaluation of

effect of each study drug

IDNT

Methods An international, prospective, randomized, double-blind, placebo-controlled trial conducted in multicenter. The mean duration of follow-up was 2.6 years. Sixteen enrolled patients never received study medication and follow-up was incomplete in 11 patients, but reasons were not speciﬁed. All analyses were based on the intention-to-treat principle

Participants A total of 1715 hypertensive patients (a systolic BP of more than 135 mmHg while sitting, a diastolic BP of more than 85 mmHg while sitting, or documented treatment with antihypertensive agents) with diabetic nephropathy due to type II diabetes mellitus underwent randomization

Interventions Eligible patients were randomized into one of three groups treated with irbesartan (300 mg daily) (N=579), amlodipine (10 mg daily) (N=567), or placebo (N=569). The target BP was 135/85 mm Hg or less in all groups, and other classes of antihypertensive agents were allowed as needed in each group

Outcomes The primary end point was renal outcomes. The secondary end point was the composite of fatal or non-fatal cardiovascular events, which were not statistically different in the three groups Adverse events were recorded at quarterly visits.

Notes Conducted in 209 centers in the America, Europe, Israel and Australasia by the clinical coordinating center and the various committees of the Collaborative Study Group

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Adequate sequence was generated by computer.

Allocation concealment? Yes To minimize any center effect, randomization was blocked by center

Blinding? Yes Probably done because it was stated as All outcomes a “double-blind clinical trial”, “matched placebo” were given in the controlled group, and the blinded clinical data base was provided to the center for statistical analyses

Incomplete outcome data addressed? Yes Missing data were equal among the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

INSIGHT

Methods A prospective, randomised, double-blind trial. Analysis was done by intention to treat. 254 patients were excluded after randomization from centres withdrawn for misconduct, and they were not included in the analysis

Participants A total of 6575 patients aged 55-80 years with hypertension (BP≥150/95 mm Hg, or ≥ 160 mm Hg systolic) were enrolled. They all also had at least one additional cardiovascular risk factor

Interventions They randomly assigned patients nifedipine 30 mg in a long-acting gastrointestinal- transport system (GITS) formulation (n=3289), or co-amilozide (hydrochlorothiazide 25 g plus amiloride 2·5 mg; n=3286). Dose titration was by dose doubling, and addition of atenolol 25-50 mg or enalapril 5-10 mg

Outcomes Cardiovascular death, MI, heart failure, or stroke.

Notes It was conducted in 703 centres in 8 countries in western Europe and Israel

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes Identical placebo was administered at the All outcomes same time of day.

Incomplete outcome data addressed? Unclear 254 patients (132 and 122 patients in each All outcomes group) were excluded after randomization from centres withdrawn for misconduct, and they were not included in the analysis

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

Methods An international, multicenter study with a prospective, randomized, open blinded endpoint evaluation design. Mean follow-up was 2.7 years. Intention-to-treat principle was used in analyses

Participants A total of 22576 hypertensive CAD patients aged 50 years or older were enrolled

Interventions Patients were randomly assigned to either verapamil sustained release (240 mg/d) (N= 11267) or atenolol (50 mg/d) (N=11309). Additional antihypertensive agents were allowed to administer to achieve BP goals

Outcomes Primary: all-cause mortality, nonfatal MI, or nonfatal stroke. Additional outcomes: time to most serious event, cardiovascular death, angina, cardiovascular hospitalizations, BP control, and so on

Notes It recruited patients at 862 site in 14 countries.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes An Internet-based management system au- tomatically randomized each patient to a treatment strategy

Allocation concealment? Yes The randomization result was stored in the central database, but drugs also might be open-label because of the PROBE design. Risk of bias in this aspect could be limited by strict randomization and blinded end point assess

Blinding? Yes It used the blinded end point design. All outcomes

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

Methods A multicenter, randomized, double-blind, controlled clinical trial with a follow-up of 3 years. All analyses were performed using the intention-to-treat approach

Participants Enrolled patients (N=883) all had hypertension (average diastolic BP from 90 to 115 mm Hg)

Interventions Participants were randomized into 2 treatment groups: hydrochlorothiazide, 12.5 to 25 mg twice a day (n=441) or isradipine, 2.5 to 5.0 mg twice a day (n=442). If the diastolic BP had not reached the planned goal with the highest dose of study drug, open-label enalapril were added

Outcomes Mean maximum IMT, and some other ﬁndings of carotid artery, and vascular events/ procedures

Notes It was conducted in 9 medical centre clinics.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes The randomization process was stratiﬁed and blocked by clinic

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes It was stated to be double-blind, but All outcomes method of blinding was not described

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

NICS-EH

Methods A randomized, double-blind trial with follow-up of 5 years. In total, 15 participants were withdrawn, but intention-to-treat analysis was not used

Participants Patients ≥60 years of age with systolic BP of 160 to 220 mm Hg and diastolic BP ,115 mm Hg were enrolled (N=429), and they were without history of cardiovascular complications

Interventions Patients were randomly assigned to 20 mg of sustained-release nicardipine hydrochloride twice daily (N=215) or 2 mg of trichlormethiazide once daily (N=214). Doubling of the dose was permitted if BP response was insufﬁcient, but any other antihypertensive drugs were prohibited

Outcomes Cardiovascular complications.

Notes It was conducted in Japan.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes It was stated as double-dummy, but details All outcomes were not described

Incomplete outcome data addressed? No Data of withdrawn patients were not in- All outcomes cluded using intention-to-treat analysis, but the missing data could be obtained to include in our systematic review

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

NORDIL

Methods A prospective, randomised, open, blinded-endpoint multicenter, parallel-group study. The mean follow-up was 4.5 years, and 52 (0,5%) were lost to follow-up. Analysis was done by intention-to-treat

Participants A total of 10881 patients, aged 50-74 years, with diastolic BP of 100 mm Hg or more on two occasions, were enrolled

Interventions Participants were randomised to a diltiazem-based regimen (180-360 mg daily, N=5410) or conventional antihypertensive treatment (N=5471) with diuretics, β-blockers, or both. Additional antihypertensive treatment could be given to any patient to lower diastolic BP to less than 90 mm Hg

Outcomes Stroke, MI, and other cardiovascular death.

Notes Recruitment of patients was from Oct 9, 1992, to Oct 31, 1999 in 1032 health centres in Norway and Sweden

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Yes Method of concealment was not described. But risk of bias in this aspect could be limited by strict randomization and blinded end point assess

Blinding? Yes It was a blinded endpoint study that all end- All outcomes points were blinding before evaluation by the separate end-point committee

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Unclear Many other antihypertensive agents were used, but whether the distributions of these drugs were different between groups was not clariﬁed, only stating that additional treatment did not affect the results

SHELL

Methods A randomized, controlled, multicenter trial conducted in outpatient clinics. Follow-up visits were made at monthly intervals during the ﬁrst 3 months after randomization and thereafter after 6 months and every year for a maximum of 5 years (median 32 months) . Data were analyzed on an intention-to-treat basis by BETA Trial Center

Participants Patients (N=1882) were recruited if sitting systolic BP was 160 mmHg with a diastolic BP equal or lower than 95 mmHg

Interventions Participants were randomly assigned to the administration of chlorthalidone 12.5 mg/ d (N=940) or lacidipine 4 mg/d (N=942). Increased dose of study drug and additional antihypertensive agents could be administered to help control blood pressure

Outcomes Composite of cardiovascular and cerebrovascular events were the primary outcome, including stroke, sudden death, MI, congestive heart failure, etc

Notes Participants were recruited from 134 units located in northern, central and southern Italy

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Randomization was made by BETA Trial Center, Genoa (Italy), using a sequentially based criterion

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Unclear Method of blinding was not described. All outcomes

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

STOP-Hypertension-2

Methods A prospective, randomised, open blinded endpoint trial. No patient was lost to follow- up and no patient refused to continue in the study. Analysis was by intention to treat and of only the ﬁrst occurrence of each event in question

Participants Participants must had hypertension (BP ≥180 mm Hg systolic, ≥105 mm Hg diastolic, or both), and were aged 70-84 years

Interventions There were three groups given different classes of drugs respectively: conventional antihypertensive drugs (N=2213) (atenolol 50 mg, metoprolol 100 mg, pindolol 5 mg, or ﬁxed-ratio hydrochlorothiazide 25 mg plus amiloride 2·5 mg daily), ACE inhibitors (N=2205) (enalapril 10 mg or lisinopril 10 mg daily), or calcium antagonists (N=2196) (felodipine 2·5 mg or isradipine 2·5 mg daily). Additional antihypertensive drugs were allowed if necessary

Outcomes Cardiovascular death, other cardiovascular event and blood pressure change were the major outcomes

Notes It was conducted in 212 health centres in Sweden.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Yes Method of concealment was not described. But risk of bias in this aspect could be limited by strict randomization and blinded end point assess

Blinding? Yes It was a blinded endpoint design. All outcomes

Incomplete outcome data addressed? Yes Missing data were equal among the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

TOMHS

Methods A randomised, double-blind, placebo-controlled clinical trial. Participants were seen at least every 3 months for a median follow-up of 4.4 years. All analyses were by treatment allocation (intention to treat)

Participants A total of 902 subjects aged 45 to 69 years, with mild hypertension (diastolic BP was 90 to 99 mm Hg at both of the ﬁrst two eligibility visits and averaged 90 to 99 mm Hg over the three eligibility visits) were included

Interventions Sustained nutritional-hygienic advice to all participants and increase physical activity. Participants were randomly allocated to take (1) placebo (n=234); (2) chlorthalidone (15 mg/d, n=136); (3) acebutolol (400 mg/d, n=132); (4) doxazosin mesylate (1 mg/d for 1 month, then 2 mg/d, n=134); (5) amlodipine maleate (5 mg/d, n=131); or (6) enalapril maleate (5 mg/d, n=135). If BP was not controlled well, drug dose was doubled, followed by use of additional drugs when necessary

Outcomes Systolic and diastolic BP. Data on morbidity and mortality outcomes was not provided for the different treatment arms

Notes It was conducted in 4 hypertension screening and treatment centres in the United States

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes A block randomization scheme with strat- iﬁcation was used by clinical center

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes Active drugs and placebo administered to All outcomes participants were prepared in identical cap- sule form

Incomplete outcome data addressed? Yes Missing data were equal among the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

VALUE

Methods A prospective, multinational, double-blind, randomised, active-controlled, parallel- group trial with a mean follow-up time of 4.2 years. All endpoints and BP values were analysed using the intention-to-treat approach

Participants Enrolled patients were 50 years or older, with treated or untreated (mean sitting systolic BP between 160 and 210 mm Hg, and a mean sitting diastolic BP of less than 115 mm Hg) hypertension at baseline and combinations of cardiovascular risk factors and cardiovascular disease. Further antihypertensive drugs excluding ARB could be given to achieve BP goal

Interventions Partipants were randomised to a group treated with either valsartan 80 mg (N=7649) or amlodipine 5 mg (N=7596)

Outcomes Time to ﬁrst cardiac event, incidence of MI, heart failure and stroke, all-cause mortality and new-onset diabetes

Notes It was carried out in 31 countries.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Yes Randomisationa list was generated by computer, and was prepared centrally by the sponsor

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Yes The study medication was provided in ex- All outcomes ternally indistinguishable capsules

Incomplete outcome data addressed? Unclear 68 patients in 9 centers were excluded af- All outcomes ter randomisation because of good clinical practice deﬁciencies, and they were not included in intention-to-treat analyses, which might lead some bias

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

VHAS

Methods A multi-centre randomized double-blind (for the ﬁrst 6 months, open subsequently) parallel-group trial lasting 2 years (prolonged to 4 years for the subgroup of patients evaluated by carotid ultrasonography). In total 215 participants dropped out (21.6% and 22.9% of each group respectively) due to poor compliance or to adverse events. Analyses of outcomes were performed in the intention-to-treat population

Participants Criteria for inclusion of patients including essential hypertension (a systolic BP when seated ≥160 mmHg and a diastolic BP ≥95 mmHg measured at the end of a placebo run-in period of 3 weeks), age of 40-65 years, either sex. A total of 1414 hypertensive patients were enrolled

Interventions The study included a run-in period (3 weeks), a double-blind-treatment period (6 months, either 240 mg sustained release verapamil (n = 707) or 25 mg chlorthalidone (n = 707) once a day), and an open-treatment period (18 months); captopril were added to the treatment of non-responding patients; free therapy of other drugs follow up was allowed when necessary

Outcomes BP reduction, heart rate, clinical safety, cardiovascular events, deaths and IMT

Notes It was a multicenter trial carried out in Italy.

Risk of bias

Item Authors’ judgement Description

Adequate sequence generation? Unclear Method of sequence generation was not described.

Allocation concealment? Unclear Method of concealment was not described.

Blinding? Unclear Method of blinding was not described. All outcomes

Incomplete outcome data addressed? Yes Missing data were equal between the treat- All outcomes ment groups, and an intention-to-treat analysis was performed

Free of selective reporting? Yes Outcomes listed in the methods section were all reported.

Free of other bias? Yes No other potential bias was found.

Characteristics of excluded studies [ordered by study ID]

Study Reason for exclusion

Abascal 1998 It was a non-randomized trial.

Bakris 1996 Only 52 patients were included in this trial.

Bakris 1997 Only 34 subjects were included in this trial.

DHCCP Its design was combination of two case control studies and two longitudinal studies

FACTS Only 96 patients were randomized to treatment in this study; it was less than 100 randomized participants which was speciﬁed in the protocol

Gottdiener 1997 It primarily evaluated the reduction of left ventricular mass, and its follow-up lasted only 1 year

HOT Participants were randomly assigned to groups with different target diastolic blood pressure instead of groups with different study drug

Kes 2003 It compared two kinds of CCBs, nifedipine and amlodipine; no other classes of antihypertensive drugs were studied

Maharaj 1992 Only 30 patients were included in this trial.

Pahor 1995 It was a prospective cohort study rather than a randomized controlled trial

Papademetriou 1997 It primarily evaluated the reduction of left ventricular mass, and its follow-up lasted only 6 months

PRESERVE It primarily evaluated the reduction of left ventricular mass, and its follow-up lasted only 12 months

Psaty 1995 It was a population-based case-control study instead of a randomized design

Radevski 1999 Only 96 patients were included in this trial.

Schneider 1991 It aimed to evaluate effect of therapy for hypertensive urgencies and the follow-up was short

STONE It was a non-randomized placebo-controlled trial.

Syst-China It was a non-randomized placebo-controlled trial.

Syst-Eur It was a placebo-controlled trial, which did not compare CCBs with any other classes of drugs for hypertension

Characteristics of studies awaiting assessment [ordered by study ID]

J-MIC(B)

Methods A prospective, randomized, controlled clinical trial comparing the effect of nifedipine retard versus ACE inhibitors on the incidence and mortality due to cardiovascular disease. The follow-up was 3-year long, and analysis was done on an intention-to-treat basis

Participants A total of 1836 hypertensive patients with coronary artery disease were enrolled

Interventions Participants were randomly assigned to nifedipine retard group or the ACE inhibitor group

Outcomes Cardiac death, AMI, hospitalization for angina pectoris or heart failure, and coronary revascularization

Notes It was performed in Japan. But only an abstract was found, instead of a full-text containing details of methods and results

Comparison 1. All cause mortality

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 16 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 5 35057 Risk Ratio (M-H, Fixed, 95% CI) 0.98 [0.92, 1.04] 1.2 CCB vs. β-blocker 4 44825 Risk Ratio (M-H, Fixed, 95% CI) 0.94 [0.88, 1.00] 1.3 CCB vs. diuretic or 3 31892 Risk Ratio (M-H, Fixed, 95% CI) 1.03 [0.94, 1.12] β-blocker 1.4 CCB vs. ACE inhibitor 5 24006 Risk Ratio (M-H, Fixed, 95% CI) 0.96 [0.91, 1.03] 1.5 CCB vs. ARB 2 16391 Risk Ratio (M-H, Fixed, 95% CI) 0.98 [0.90, 1.07]

Comparison 2. Myocardial infarction

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 15 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 5 34072 Risk Ratio (M-H, Fixed, 95% CI) 1.00 [0.92, 1.08] 1.2 CCB vs. β-blocker 3 22249 Risk Ratio (M-H, Fixed, 95% CI) 0.90 [0.79, 1.02] 1.3 CCB vs. diuretic and 3 31892 Risk Ratio (M-H, Fixed, 95% CI) 1.05 [0.93, 1.19] β-blocker 1.4 CCB vs. ACE inhibitor 5 24006 Risk Ratio (M-H, Fixed, 95% CI) 1.06 [0.98, 1.15] 1.5 CCB vs. ARB 2 16391 Risk Ratio (M-H, Fixed, 95% CI) 0.83 [0.72, 0.96] 2 Amlodipine vs. ACE inhibitor 3 19135 Risk Ratio (M-H, Fixed, 95% CI) 1.00 [0.91, 1.10]

Comparison 3. Stroke

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 15 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 5 34072 Risk Ratio (M-H, Fixed, 95% CI) 0.94 [0.84, 1.05] 1.2 CCB vs. β-blocker 3 22249 Risk Ratio (M-H, Fixed, 95% CI) 0.77 [0.67, 0.88] 1.3 CCB vs. diuretic or 3 31892 Risk Ratio (M-H, Fixed, 95% CI) 0.92 [0.81, 1.03] β-blocker 1.4 CCB vs. ACE inhibitor 5 24006 Risk Ratio (M-H, Fixed, 95% CI) 0.89 [0.80, 0.98] 1.5 CCB vs. ARB 2 16391 Risk Ratio (M-H, Fixed, 95% CI) 0.85 [0.73, 0.99]

Calcium channel blockers versus other classes of drugs for hypertension (Review) 45 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Comparison 4. Congestive heart failure

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 13 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 5 34072 Risk Ratio (M-H, Fixed, 95% CI) 1.37 [1.25, 1.51] 1.2 CCB vs. β-blocker 2 19915 Risk Ratio (M-H, Fixed, 95% CI) 0.83 [0.67, 1.04] 1.3 CCB vs. diuretic and 3 31892 Risk Ratio (M-H, Fixed, 95% CI) 1.15 [0.99, 1.33] β-blocker 1.4 CCB vs. ACE inhibitor 4 23626 Risk Ratio (M-H, Fixed, 95% CI) 1.16 [1.06, 1.27] 1.5 CCB vs. ARB 2 16391 Risk Ratio (M-H, Fixed, 95% CI) 1.20 [1.06, 1.36]

Comparison 5. Cardiovascular mortality

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 13 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 4 32721 Risk Ratio (M-H, Fixed, 95% CI) 1.02 [0.93, 1.12] 1.2 CCB vs. β-blocker 4 44825 Risk Ratio (M-H, Fixed, 95% CI) 0.90 [0.81, 0.99] 1.3 CCB vs. diuretic or 3 31892 Risk Ratio (M-H, Fixed, 95% CI) 1.04 [0.92, 1.18] β-blocker 1.4 CCB vs. ACE inhibitor 4 23626 Risk Ratio (M-H, Fixed, 95% CI) 0.97 [0.89, 1.07] 1.5 CCB vs. ARB 1 1146 Risk Ratio (M-H, Fixed, 95% CI) 0.73 [0.48, 1.09] 2 DHP vs. β-blocker 3 22249 Risk Ratio (M-H, Fixed, 95% CI) 0.77 [0.66, 0.90]

Comparison 6. Major cardiovascular events

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 CCBs vs other classes of 10 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only antihypertensive agents 1.1 CCB vs. diuretic 4 33643 Risk Ratio (M-H, Fixed, 95% CI) 1.05 [1.00, 1.09] 1.2 CCB vs. β-blocker 3 22249 Risk Ratio (M-H, Fixed, 95% CI) 0.84 [0.77, 0.92] 1.3 CCB vs. diuretic and 2 21011 Risk Ratio (M-H, Fixed, 95% CI) 1.02 [0.95, 1.10] β-blocker 1.4 CCB vs. ACE inhibitor 4 23536 Risk Ratio (M-H, Fixed, 95% CI) 0.98 [0.94, 1.02] 2 Sensitivity analysis: CCBs vs. 3 23156 Risk Ratio (M-H, Fixed, 95% CI) 0.97 [0.93, 1.02] ACE inhibitors

Calcium channel blockers versus other classes of drugs for hypertension (Review) 46 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Comparison 7. Blood pressure reduction

No. of No. of Outcome or subgroup title studies participants Statistical method Effect size

1 Systolic blood pressure reduction 8 Mean Difference (IV, Fixed, 95% CI) Subtotals only 1.1 CCB vs. diuretic 3 24963 Mean Difference (IV, Fixed, 95% CI) 0.81 [0.56, 1.06] 1.2 CCB vs. β-blocker 3 23474 Mean Difference (IV, Fixed, 95% CI) 0.25 [-0.31, 0.81] 1.3 CCB vs. diuretic or 1 10881 Mean Difference (IV, Fixed, 95% CI) 3.0 [2.59, 3.41] β-blocker 1.4 CCB vs. ACE inhibitor 4 19368 Mean Difference (IV, Fixed, 95% CI) -1.11 [-1.40, -0.82] 1.5 CCB vs. ARB 1 15245 Mean Difference (IV, Fixed, 95% CI) -2.10 [-2.46, -1.74] 1.6 CCB vs. α1-antagonist 1 235 Mean Difference (IV, Fixed, 95% CI) -1.40 [-3.89, 1.09] 2 Diastolic blood pressure 8 Mean Difference (IV, Fixed, 95% CI) Subtotals only reduction 2.1 CCB vs. diuretic 3 24963 Mean Difference (IV, Fixed, 95% CI) -0.68 [-0.84, -0.52] 2.2 CCB vs. β-blocker 3 23474 Mean Difference (IV, Fixed, 95% CI) 0.15 [-0.16, 0.45] 2.3 CCB vs. diuretic or 1 10881 Mean Difference (IV, Fixed, 95% CI) 0.10 [-0.07, 0.27] β-blocker 2.4 CCB vs. ACE inhibitor 4 19368 Mean Difference (IV, Fixed, 95% CI) -0.63 [-0.81, -0.44] 2.5 CCB vs. ARB 1 15245 Mean Difference (IV, Fixed, 95% CI) -1.70 [-1.91, -1.49] 2.6 CCB vs. α1-antagonist 1 235 Mean Difference (IV, Fixed, 95% CI) -1.20 [-2.39, -0.01] 3 Sensitivity analysis: CCBs vs. 3 Mean Difference (IV, Fixed, 95% CI) Subtotals only ACE inhibitors 3.1 Systolic blood pressure 3 18988 Mean Difference (IV, Fixed, 95% CI) 1.00 [-1.29, -0.70] reduction 3.2 Diastolic blood pressure 3 18988 Mean Difference (IV, Fixed, 95% CI) -0.62 [-0.81, -0.44] reduction

Calcium channel blockers versus other classes of drugs for hypertension (Review) 47 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 1.1. Comparison 1 All cause mortality, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 1 All cause mortality

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup CCBs Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 1256/9048 2203/15255 85.1 % 0.96 [ 0.90, 1.03 ]

INSIGHT 153/3289 152/3286 7.9 % 1.01 [ 0.81, 1.25 ]

MIDAS 8/442 9/441 0.5 % 0.89 [ 0.35, 2.28 ]

SHELL 145/942 122/940 6.3 % 1.19 [ 0.95, 1.48 ]

VHAS 5/707 4/707 0.2 % 1.25 [ 0.34, 4.64 ] Subtotal (95% CI) 14428 20629 100.0 % 0.98 [ 0.92, 1.04 ] Total events: 1567 (CCBs), 2490 (Other agents) Heterogeneity: Chi2 = 3.38, df = 4 (P = 0.50); I2 =0.0% Test for overall effect: Z = 0.69 (P = 0.49) 2 CCB vs. -blocker AASK 22/217 49/441 1.8 % 0.91 [ 0.57, 1.47 ]

ASCOT-BPLA 738/9639 820/9618 46.6 % 0.90 [ 0.82, 0.99 ]

ELSA 13/1177 17/1157 1.0 % 0.75 [ 0.37, 1.54 ]

INVEST 873/11267 893/11309 50.6 % 0.98 [ 0.90, 1.07 ] Subtotal (95% CI) 22300 22525 100.0 % 0.94 [ 0.88, 1.00 ] Total events: 1646 (CCBs), 1779 (Other agents) Heterogeneity: Chi2 = 2.15, df = 3 (P = 0.54); I2 =0.0% Test for overall effect: Z = 1.92 (P = 0.055) 3 CCB vs. diuretic or -blocker CONVINCE 337/8241 319/8361 34.8 % 1.07 [ 0.92, 1.25 ]

NORDIL 231/5410 228/5471 24.9 % 1.02 [ 0.86, 1.23 ]

STOP-Hypertension-2 362/2196 369/2213 40.3 % 0.99 [ 0.87, 1.13 ] Subtotal (95% CI) 15847 16045 100.0 % 1.03 [ 0.94, 1.12 ] Total events: 930 (CCBs), 916 (Other agents) Heterogeneity: Chi2 = 0.63, df = 2 (P = 0.73); I2 =0.0% Test for overall effect: Z = 0.59 (P = 0.56) 4 CCB vs. ACE inhibitor AASK 22/217 34/436 1.3 % 1.30 [ 0.78, 2.17 ]

ABCD 18/235 14/235 0.8 % 1.29 [ 0.65, 2.52 ]

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 48 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup CCBs Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI ALLHAT 1256/9048 1314/9054 75.8 % 0.96 [ 0.89, 1.03 ]

FACET 5/191 4/189 0.2 % 1.24 [ 0.34, 4.54 ]

STOP-Hypertension-2 362/2196 380/2205 21.9 % 0.96 [ 0.84, 1.09 ] Subtotal (95% CI) 11887 12119 100.0 % 0.96 [ 0.91, 1.03 ] Total events: 1663 (CCBs), 1746 (Other agents) Heterogeneity: Chi2 = 2.22, df = 4 (P = 0.70); I2 =0.0% Test for overall effect: Z = 1.15 (P = 0.25) 5 CCB vs. ARB IDNT 83/567 87/579 9.3 % 0.97 [ 0.74, 1.29 ]

VALUE 818/7596 841/7649 90.7 % 0.98 [ 0.89, 1.07 ] Subtotal (95% CI) 8163 8228 100.0 % 0.98 [ 0.90, 1.07 ] Total events: 901 (CCBs), 928 (Other agents) Heterogeneity: Chi2 = 0.00, df = 1 (P = 0.97); I2 =0.0% Test for overall effect: Z = 0.48 (P = 0.63)

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 49 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 2.1. Comparison 2 Myocardial infarction, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 2 Myocardial infarction

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 798/9048 1362/15255 89.4 % 0.99 [ 0.91, 1.07 ]

INSIGHT 94/3289 84/3286 7.4 % 1.12 [ 0.84, 1.49 ]

MIDAS 8/442 7/441 0.6 % 1.14 [ 0.42, 3.12 ]

NICS-EH 2/215 2/214 0.2 % 1.00 [ 0.14, 7.00 ]

SHELL 28/942 27/940 2.4 % 1.03 [ 0.61, 1.74 ] Subtotal (95% CI) 13936 20136 100.0 % 1.00 [ 0.92, 1.08 ] Total events: 930 (Favours experimental), 1482 (Other agents) Heterogeneity: Chi2 = 0.73, df = 4 (P = 0.95); I2 =0.0% Test for overall effect: Z = 0.01 (P = 0.99) 2 CCB vs. -blocker AASK 5/217 19/441 2.5 % 0.53 [ 0.20, 1.41 ]

ASCOT-BPLA 429/9639 474/9618 94.1 % 0.90 [ 0.79, 1.03 ]

ELSA 18/1177 17/1157 3.4 % 1.04 [ 0.54, 2.01 ] Subtotal (95% CI) 11033 11216 100.0 % 0.90 [ 0.79, 1.02 ] Total events: 452 (Favours experimental), 510 (Other agents) Heterogeneity: Chi2 = 1.29, df = 2 (P = 0.52); I2 =0.0% Test for overall effect: Z = 1.69 (P = 0.091) 3 CCB vs. diuretic and -blocker CONVINCE 133/8241 166/8361 34.7 % 0.81 [ 0.65, 1.02 ]

NORDIL 183/5410 157/5471 32.9 % 1.18 [ 0.96, 1.45 ]

STOP-Hypertension-2 179/2196 154/2213 32.3 % 1.17 [ 0.95, 1.44 ] Subtotal (95% CI) 15847 16045 100.0 % 1.05 [ 0.93, 1.19 ] Total events: 495 (Favours experimental), 477 (Other agents) Heterogeneity: Chi2 = 7.16, df = 2 (P = 0.03); I2 =72% Test for overall effect: Z = 0.77 (P = 0.44) 4 CCB vs. ACE inhibitor AASK 5/217 18/436 1.2 % 0.56 [ 0.21, 1.48 ]

ABCD 27/235 9/235 0.9 % 3.00 [ 1.44, 6.24 ]

ALLHAT 798/9048 796/9054 82.4 % 1.00 [ 0.91, 1.10 ]

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 50 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI FACET 13/191 10/189 1.0 % 1.29 [ 0.58, 2.86 ]

STOP-Hypertension-2 179/2196 139/2205 14.4 % 1.29 [ 1.04, 1.60 ] Subtotal (95% CI) 11887 12119 100.0 % 1.06 [ 0.98, 1.15 ] Total events: 1022 (Favours experimental), 972 (Other agents) Heterogeneity: Chi2 = 14.29, df = 4 (P = 0.01); I2 =72% Test for overall effect: Z = 1.38 (P = 0.17) 5 CCB vs. ARB IDNT 27/567 44/579 10.6 % 0.63 [ 0.39, 1.00 ]

VALUE 313/7596 369/7649 89.4 % 0.85 [ 0.74, 0.99 ] Subtotal (95% CI) 8163 8228 100.0 % 0.83 [ 0.72, 0.96 ] Total events: 340 (Favours experimental), 413 (Other agents) Heterogeneity: Chi2 = 1.55, df = 1 (P = 0.21); I2 =35% Test for overall effect: Z = 2.60 (P = 0.0092)

0.2 0.5 1 2 5 Favours experimental Favours control

Analysis 2.2. Comparison 2 Myocardial infarction, Outcome 2 Amlodipine vs. ACE inhibitor.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 2 Myocardial infarction

Outcome: 2 Amlodipine vs. ACE inhibitor

Studyorsubgroup Amlodipine ACEinhibitors RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI AASK 5/217 18/436 1.5 % 0.56 [ 0.21, 1.48 ]

ALLHAT 798/9048 796/9054 97.3 % 1.00 [ 0.91, 1.10 ]

FACET 13/191 10/189 1.2 % 1.29 [ 0.58, 2.86 ] Total (95% CI) 9456 9679 100.0 % 1.00 [ 0.91, 1.10 ] Total events: 816 (Amlodipine), 824 (ACE inhibitors) Heterogeneity: Chi2 = 1.75, df = 2 (P = 0.42); I2 =0.0% Test for overall effect: Z = 0.00 (P = 1.0) Test for subgroup differences: Not applicable

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 51 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 3.1. Comparison 3 Stroke, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 3 Stroke

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 377/9048 675/15255 80.3 % 0.94 [ 0.83, 1.07 ]

INSIGHT 67/3289 74/3286 11.8 % 0.90 [ 0.65, 1.25 ]

MIDAS 6/442 3/441 0.5 % 2.00 [ 0.50, 7.93 ]

NICS-EH 6/215 8/214 1.3 % 0.75 [ 0.26, 2.12 ]

SHELL 37/942 38/940 6.1 % 0.97 [ 0.62, 1.51 ] Subtotal (95% CI) 13936 20136 100.0 % 0.94 [ 0.84, 1.05 ] Total events: 493 (Favours experimental), 798 (Other agents) Heterogeneity: Chi2 = 1.41, df = 4 (P = 0.84); I2 =0.0% Test for overall effect: Z = 1.07 (P = 0.29) 2 CCB vs. -blocker AASK 9/217 23/441 3.4 % 0.80 [ 0.37, 1.69 ]

ASCOT-BPLA 327/9639 422/9618 93.5 % 0.77 [ 0.67, 0.89 ]

ELSA 9/1177 14/1157 3.1 % 0.63 [ 0.27, 1.45 ] Subtotal (95% CI) 11033 11216 100.0 % 0.77 [ 0.67, 0.88 ] Total events: 345 (Favours experimental), 459 (Other agents) Heterogeneity: Chi2 = 0.23, df = 2 (P = 0.89); I2 =0.0% Test for overall effect: Z = 3.74 (P = 0.00018) 3 CCB vs. diuretic or -blocker CONVINCE 133/8241 118/8361 21.4 % 1.14 [ 0.89, 1.46 ]

NORDIL 159/5410 196/5471 35.6 % 0.82 [ 0.67, 1.01 ]

STOP-Hypertension-2 207/2196 237/2213 43.1 % 0.88 [ 0.74, 1.05 ] Subtotal (95% CI) 15847 16045 100.0 % 0.92 [ 0.81, 1.03 ] Total events: 499 (Favours experimental), 551 (Other agents) Heterogeneity: Chi2 = 4.42, df = 2 (P = 0.11); I2 =55% Test for overall effect: Z = 1.48 (P = 0.14) 4 CCB vs. ACE inhibitor AASK 9/217 23/436 2.2 % 0.79 [ 0.37, 1.67 ]

ABCD 11/235 7/235 1.0 % 1.57 [ 0.62, 3.98 ]

ALLHAT 377/9048 457/9054 65.5 % 0.83 [ 0.72, 0.94 ]

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 52 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI FACET 10/191 4/189 0.6 % 2.47 [ 0.79, 7.75 ]

STOP-Hypertension-2 207/2196 215/2205 30.8 % 0.97 [ 0.81, 1.16 ] Subtotal (95% CI) 11887 12119 100.0 % 0.89 [ 0.80, 0.98 ] Total events: 614 (Favours experimental), 706 (Other agents) Heterogeneity: Chi2 = 6.63, df = 4 (P = 0.16); I2 =40% Test for overall effect: Z = 2.28 (P = 0.023) 5 CCB vs. ARB IDNT 15/567 28/579 7.9 % 0.55 [ 0.30, 1.01 ]

VALUE 281/7596 322/7649 92.1 % 0.88 [ 0.75, 1.03 ] Subtotal (95% CI) 8163 8228 100.0 % 0.85 [ 0.73, 0.99 ] Total events: 296 (Favours experimental), 350 (Other agents) Heterogeneity: Chi2 = 2.13, df = 1 (P = 0.14); I2 =53% Test for overall effect: Z = 2.06 (P = 0.039)

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 53 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 4.1. Comparison 4 Congestive heart failure, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 4 Congestive heart failure

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 706/9048 870/15255 94.9 % 1.37 [ 1.24, 1.51 ]

INSIGHT 26/3289 12/3286 1.8 % 2.16 [ 1.09, 4.28 ]

MIDAS 2/442 0/441 0.1 % 4.99 [ 0.24, 103.61 ]

NICS-EH 0/215 3/214 0.5 % 0.14 [ 0.01, 2.74 ]

SHELL 23/942 19/940 2.8 % 1.21 [ 0.66, 2.20 ] Subtotal (95% CI) 13936 20136 100.0 % 1.37 [ 1.25, 1.51 ] Total events: 757 (Favours experimental), 904 (Other agents) Heterogeneity: Chi2 = 4.84, df = 4 (P = 0.30); I2 =17% Test for overall effect: Z = 6.66 (P < 0.00001) 2 CCB vs. -blocker AASK 8/217 22/441 8.4 % 0.74 [ 0.33, 1.63 ]

ASCOT-BPLA 134/9639 159/9618 91.6 % 0.84 [ 0.67, 1.06 ] Subtotal (95% CI) 9856 10059 100.0 % 0.83 [ 0.67, 1.04 ] Total events: 142 (Favours experimental), 181 (Other agents) Heterogeneity: Chi2 = 0.09, df = 1 (P = 0.76); I2 =0.0% Test for overall effect: Z = 1.64 (P = 0.10) 3 CCB vs. diuretic and -blocker CONVINCE 126/8241 100/8361 30.2 % 1.28 [ 0.98, 1.66 ]

NORDIL 63/5410 53/5471 16.1 % 1.20 [ 0.84, 1.73 ]

STOP-Hypertension-2 186/2196 177/2213 53.7 % 1.06 [ 0.87, 1.29 ] Subtotal (95% CI) 15847 16045 100.0 % 1.15 [ 0.99, 1.33 ] Total events: 375 (Favours experimental), 330 (Other agents) Heterogeneity: Chi2 = 1.36, df = 2 (P = 0.51); I2 =0.0% Test for overall effect: Z = 1.88 (P = 0.060) 4 CCB vs. ACE inhibitor AASK 8/217 20/436 1.7 % 0.80 [ 0.36, 1.80 ]

ABCD 8/235 10/235 1.3 % 0.80 [ 0.32, 1.99 ]

ALLHAT 706/9048 612/9054 78.1 % 1.15 [ 1.04, 1.28 ]

STOP-Hypertension-2 186/2196 149/2205 19.0 % 1.25 [ 1.02, 1.54 ]

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 54 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Subtotal (95% CI) 11696 11930 100.0 % 1.16 [ 1.06, 1.27 ] Total events: 908 (Favours experimental), 791 (Other agents) Heterogeneity: Chi2 = 1.98, df = 3 (P = 0.58); I2 =0.0% Test for overall effect: Z = 3.21 (P = 0.0013) 5 CCB vs. ARB IDNT 93/567 60/579 14.4 % 1.58 [ 1.17, 2.14 ]

VALUE 400/7596 354/7649 85.6 % 1.14 [ 0.99, 1.31 ] Subtotal (95% CI) 8163 8228 100.0 % 1.20 [ 1.06, 1.36 ] Total events: 493 (Favours experimental), 414 (Other agents) Heterogeneity: Chi2 = 3.76, df = 1 (P = 0.05); I2 =73% Test for overall effect: Z = 2.85 (P = 0.0044)

0.2 0.5 1 2 5 Favours experimental Favours control

Analysis 5.1. Comparison 5 Cardiovascular mortality, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 5 Cardiovascular mortality

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 592/9048 992/15255 92.9 % 1.01 [ 0.91, 1.11 ]

INSIGHT 60/3289 52/3286 6.5 % 1.15 [ 0.80, 1.67 ]

NICS-EH 2/215 0/214 0.1 % 4.98 [ 0.24, 103.06 ]

VHAS 5/707 4/707 0.5 % 1.25 [ 0.34, 4.64 ] Subtotal (95% CI) 13259 19462 100.0 % 1.02 [ 0.93, 1.12 ] Total events: 659 (Favours experimental), 1048 (Other agents) Heterogeneity: Chi2 = 1.64, df = 3 (P = 0.65); I2 =0.0% Test for overall effect: Z = 0.40 (P = 0.69) 2 CCB vs. -blocker

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 55 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI AASK 7/217 12/441 1.0 % 1.19 [ 0.47, 2.97 ]

ASCOT-BPLA 263/9639 342/9618 43.4 % 0.77 [ 0.66, 0.90 ]

ELSA 4/1177 8/1157 1.0 % 0.49 [ 0.15, 1.63 ]

INVEST 431/11267 431/11309 54.6 % 1.00 [ 0.88, 1.14 ] Subtotal (95% CI) 22300 22525 100.0 % 0.90 [ 0.81, 0.99 ] Total events: 705 (Favours experimental), 793 (Other agents) Heterogeneity: Chi2 = 7.89, df = 3 (P = 0.05); I2 =62% Test for overall effect: Z = 2.12 (P = 0.034) 3 CCB vs. diuretic or -blocker CONVINCE 152/8241 143/8361 29.8 % 1.08 [ 0.86, 1.35 ]

NORDIL 131/5410 115/5471 24.0 % 1.15 [ 0.90, 1.48 ]

STOP-Hypertension-2 212/2196 221/2213 46.2 % 0.97 [ 0.81, 1.16 ] Subtotal (95% CI) 15847 16045 100.0 % 1.04 [ 0.92, 1.18 ] Total events: 495 (Favours experimental), 479 (Other agents) Heterogeneity: Chi2 = 1.40, df = 2 (P = 0.50); I2 =0.0% Test for overall effect: Z = 0.70 (P = 0.49) 4 CCB vs. ACE inhibitor AASK 7/217 12/436 0.9 % 1.17 [ 0.47, 2.93 ]

ABCD 11/235 6/235 0.7 % 1.83 [ 0.69, 4.88 ]

ALLHAT 592/9048 609/9054 71.8 % 0.97 [ 0.87, 1.09 ]

STOP-Hypertension-2 212/2196 226/2205 26.6 % 0.94 [ 0.79, 1.13 ] Subtotal (95% CI) 11696 11930 100.0 % 0.97 [ 0.89, 1.07 ] Total events: 822 (Favours experimental), 853 (Other agents) Heterogeneity: Chi2 = 1.90, df = 3 (P = 0.59); I2 =0.0% Test for overall effect: Z = 0.59 (P = 0.55) 5 CCB vs. ARB IDNT 37/567 52/579 100.0 % 0.73 [ 0.48, 1.09 ] Subtotal (95% CI) 567 579 100.0 % 0.73 [ 0.48, 1.09 ] Total events: 37 (Favours experimental), 52 (Other agents) Heterogeneity: not applicable Test for overall effect: Z = 1.54 (P = 0.12)

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 56 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 5.2. Comparison 5 Cardiovascular mortality, Outcome 2 DHP vs. β-blocker.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 5 Cardiovascular mortality

Outcome: 2 DHP vs. -blocker

Studyorsubgroup Favoursexperimental -blockers RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI AASK 7/217 12/441 2.2 % 1.19 [ 0.47, 2.97 ]

ASCOT-BPLA 263/9639 342/9618 95.5 % 0.77 [ 0.66, 0.90 ]

ELSA 4/1177 8/1157 2.3 % 0.49 [ 0.15, 1.63 ] Total (95% CI) 11033 11216 100.0 % 0.77 [ 0.66, 0.90 ] Total events: 274 (Favours experimental), 362 (-blockers) Heterogeneity: Chi2 = 1.39, df = 2 (P = 0.50); I2 =0.0% Test for overall effect: Z = 3.31 (P = 0.00093) Test for subgroup differences: Not applicable

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 57 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 6.1. Comparison 6 Major cardiovascular events, Outcome 1 CCBs vs other classes of antihypertensive agents.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 6 Major cardiovascular events

Outcome: 1 CCBs vs other classes of antihypertensive agents

Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 2432/9048 3941/15255 91.2 % 1.04 [ 1.00, 1.09 ]

INSIGHT 200/3289 182/3286 5.7 % 1.10 [ 0.90, 1.33 ]

MIDAS 25/442 14/441 0.4 % 1.78 [ 0.94, 3.38 ]

SHELL 90/942 88/940 2.7 % 1.02 [ 0.77, 1.35 ] Subtotal (95% CI) 13721 19922 100.0 % 1.05 [ 1.00, 1.09 ] Total events: 2747 (Favours experimental), 4225 (Other agents) Heterogeneity: Chi2 = 2.98, df = 3 (P = 0.39); I2 =0.0% Test for overall effect: Z = 2.12 (P = 0.034) 2 CCB vs. -blocker AASK 23/217 65/441 4.2 % 0.72 [ 0.46, 1.12 ]

ASCOT-BPLA 796/9639 937/9618 92.5 % 0.85 [ 0.77, 0.93 ]

ELSA 27/1177 33/1157 3.3 % 0.80 [ 0.49, 1.33 ] Subtotal (95% CI) 11033 11216 100.0 % 0.84 [ 0.77, 0.92 ] Total events: 846 (Favours experimental), 1035 (Other agents) Heterogeneity: Chi2 = 0.53, df = 2 (P = 0.77); I2 =0.0% Test for overall effect: Z = 3.90 (P = 0.000094) 3 CCB vs. diuretic and -blocker CONVINCE 793/8241 775/8361 62.7 % 1.04 [ 0.94, 1.14 ]

STOP-Hypertension-2 450/2196 460/2213 37.3 % 0.99 [ 0.88, 1.11 ] Subtotal (95% CI) 10437 10574 100.0 % 1.02 [ 0.95, 1.10 ] Total events: 1243 (Favours experimental), 1235 (Other agents) Heterogeneity: Chi2 = 0.46, df = 1 (P = 0.50); I2 =0.0% Test for overall effect: Z = 0.49 (P = 0.62) 4 CCB vs. ACE inhibitor AASK 23/217 61/436 1.3 % 0.76 [ 0.48, 1.19 ]

ALLHAT 2432/9048 2514/9054 83.7 % 0.97 [ 0.92, 1.02 ]

FACET 27/191 14/189 0.5 % 1.91 [ 1.03, 3.52 ]

STOP-Hypertension-2 450/2196 437/2205 14.5 % 1.03 [ 0.92, 1.16 ] Subtotal (95% CI) 11652 11884 100.0 % 0.98 [ 0.94, 1.02 ]

0.2 0.5 1 2 5 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 58 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup Favoursexperimental Otheragents RiskRatio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI Total events: 2932 (Favours experimental), 3026 (Other agents) Heterogeneity: Chi2 = 6.84, df = 3 (P = 0.08); I2 =56% Test for overall effect: Z = 0.94 (P = 0.35)

0.2 0.5 1 2 5 Favours experimental Favours control

Analysis 6.2. Comparison 6 Major cardiovascular events, Outcome 2 Sensitivity analysis: CCBs vs. ACE inhibitors.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 6 Major cardiovascular events

Outcome: 2 Sensitivity analysis: CCBs vs. ACE inhibitors

Studyorsubgroup Favoursexperimental ACEinhibitors Risk Ratio Weight RiskRatio n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI AASK 23/217 61/436 1.4 % 0.76 [ 0.48, 1.19 ]

ALLHAT 2432/9048 2514/9054 84.1 % 0.97 [ 0.92, 1.02 ]

STOP-Hypertension-2 450/2196 437/2205 14.6 % 1.03 [ 0.92, 1.16 ] Total (95% CI) 11461 11695 100.0 % 0.97 [ 0.93, 1.02 ] Total events: 2905 (Favours experimental), 3012 (ACE inhibitors) Heterogeneity: Chi2 = 2.25, df = 2 (P = 0.33); I2 =11% Test for overall effect: Z = 1.14 (P = 0.25) Test for subgroup differences: Not applicable

0.2 0.5 1 2 5 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 59 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 7.1. Comparison 7 Blood pressure reduction, Outcome 1 Systolic blood pressure reduction.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 7 Blood pressure reduction

Outcome: 1 Systolic blood pressure reduction

Studyorsubgroup CCBs Otheragents MeanDifference Weight Mean Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 9048 -11.5 (9.7) 15255 -12.3 (9.8) 97.1 % 0.80 [ 0.55, 1.05 ]

NICS-EH 215 -24.9 (9) 214 -25.6 (9.8) 2.0 % 0.70 [ -1.08, 2.48 ]

TOMHS 114 -15.6 (9.6) 117 -17.7 (10.8) 0.9 % 2.10 [ -0.53, 4.73 ] Subtotal (95% CI) 9377 15586 100.0 % 0.81 [ 0.56, 1.06 ] Heterogeneity: Chi2 = 0.94, df = 2 (P = 0.62); I2 =0.0% Test for overall effect: Z = 6.36 (P < 0.00001) 2 CCB vs. -blocker AASK 217 -17 (17) 441 -15 (15.67) 4.3 % -2.00 [ -4.69, 0.69 ]

INVEST 11267 -18.7 (22.2) 11309 -19 (22.6) 91.2 % 0.30 [ -0.28, 0.88 ]

TOMHS 114 -15.6 (9.6) 126 -17 (11.2) 4.5 % 1.40 [ -1.23, 4.03 ] Subtotal (95% CI) 11598 11876 100.0 % 0.25 [ -0.31, 0.81 ] Heterogeneity: Chi2 = 3.44, df = 2 (P = 0.18); I2 =42% Test for overall effect: Z = 0.88 (P = 0.38) 3 CCB vs. diuretic or -blocker NORDIL 5410 -21.3 (10.9) 5471 -24.3 (10.8) 100.0 % 3.00 [ 2.59, 3.41 ] Subtotal (95% CI) 5410 5471 100.0 % 3.00 [ 2.59, 3.41 ] Heterogeneity: not applicable Test for overall effect: Z = 14.42 (P < 0.00001) 4 CCB vs. ACE inhibitor AASK 217 -17 (17) 436 -16 (14.48) 1.2 % -1.00 [ -3.64, 1.64 ]

ALLHAT 9048 -11.5 (9.7) 9054 -10.5 (10.8) 94.6 % -1.00 [ -1.30, -0.70 ]

FACET 191 -18 (8.6) 189 -13 (8.6) 2.8 % -5.00 [ -6.73, -3.27 ]

TOMHS 114 -15.6 (9.6) 119 -14.7 (9.8) 1.4 % -0.90 [ -3.39, 1.59 ] Subtotal (95% CI) 9570 9798 100.0 % -1.11 [ -1.40, -0.82 ] Heterogeneity: Chi2 = 19.99, df = 3 (P = 0.00017); I2 =85% Test for overall effect: Z = 7.49 (P < 0.00001) 5 CCB vs. ARB VALUE 7596 -17.3 (11.4) 7649 -15.2 (11.4) 100.0 % -2.10 [ -2.46, -1.74 ] Subtotal (95% CI) 7596 7649 100.0 % -2.10 [ -2.46, -1.74 ] Heterogeneity: not applicable

-4 -2 0 2 4 Favours CCB Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 60 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup CCBs Otheragents MeanDifference Weight Mean Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI Test for overall effect: Z = 11.37 (P < 0.00001) 6 CCB vs. 1-antagonist TOMHS 114 -15.6 (9.6) 121 -14.2 (9.9) 100.0 % -1.40 [ -3.89, 1.09 ] Subtotal (95% CI) 114 121 100.0 % -1.40 [ -3.89, 1.09 ] Heterogeneity: not applicable Test for overall effect: Z = 1.10 (P = 0.27) Test for subgroup differences: Chi2 = 435.30, df = 5 (P = 0.00), I2 =99%

-4 -2 0 2 4 Favours CCB Favours control

Analysis 7.2. Comparison 7 Blood pressure reduction, Outcome 2 Diastolic blood pressure reduction.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 7 Blood pressure reduction

Outcome: 2 Diastolic blood pressure reduction

Studyorsubgroup CCBs Otheragents MeanDifference Weight Mean Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

1 CCB vs. diuretic ALLHAT 9048 -9.3 (6.4) 15255 -8.6 (6.3) 96.4 % -0.70 [ -0.87, -0.53 ]

NICS-EH 215 -13.2 (6.1) 214 -13.5 (6.2) 1.9 % 0.30 [ -0.86, 1.46 ]

TOMHS 114 -12.9 (4.3) 117 -12.3 (5.4) 1.7 % -0.60 [ -1.86, 0.66 ] Subtotal (95% CI) 9377 15586 100.0 % -0.68 [ -0.84, -0.52 ] Heterogeneity: Chi2 = 2.79, df = 2 (P = 0.25); I2 =28% Test for overall effect: Z = 8.19 (P < 0.00001) 2 CCB vs. -blocker AASK 217 -15 (8.98) 441 -14 (8.68) 4.5 % -1.00 [ -2.44, 0.44 ]

INVEST 11267 -10 (12.4) 11309 -10.2 (12.4) 89.6 % 0.20 [ -0.12, 0.52 ]

TOMHS 114 -12.9 (4.3) 126 -13.1 (5.6) 5.9 % 0.20 [ -1.06, 1.46 ] Subtotal (95% CI) 11598 11876 100.0 % 0.15 [ -0.16, 0.45 ] Heterogeneity: Chi2 = 2.54, df = 2 (P = 0.28); I2 =21% Test for overall effect: Z = 0.93 (P = 0.35) 3 CCB vs. diuretic or -blocker

-4 -2 0 2 4 Favours experimental Favours control (Continued ... )

Calcium channel blockers versus other classes of drugs for hypertension (Review) 61 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (... Continued) Studyorsubgroup CCBs Otheragents MeanDifference Weight Mean Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI NORDIL 5410 -18.2 (4.6) 5471 -18.3 (4.7) 100.0 % 0.10 [ -0.07, 0.27 ] Subtotal (95% CI) 5410 5471 100.0 % 0.10 [ -0.07, 0.27 ] Heterogeneity: not applicable Test for overall effect: Z = 1.12 (P = 0.26) 4 CCB vs. ACE inhibitor AASK 217 -15 (8.98) 436 -14 (9.48) 1.5 % -1.00 [ -2.49, 0.49 ]

ALLHAT 9048 -9.3 (6.4) 9054 -8.7 (6.6) 95.2 % -0.60 [ -0.79, -0.41 ]

FACET 191 -8 (8.6) 189 -7 (8.6) 1.1 % -1.00 [ -2.73, 0.73 ]

TOMHS 114 -12.9 (4.3) 119 -11.5 (5.5) 2.1 % -1.40 [ -2.66, -0.14 ] Subtotal (95% CI) 9570 9798 100.0 % -0.63 [ -0.81, -0.44 ] Heterogeneity: Chi2 = 1.93, df = 3 (P = 0.59); I2 =0.0% Test for overall effect: Z = 6.66 (P < 0.00001) 5 CCB vs. ARB VALUE 7596 -9.9 (6.6) 7649 -8.2 (6.6) 100.0 % -1.70 [ -1.91, -1.49 ] Subtotal (95% CI) 7596 7649 100.0 % -1.70 [ -1.91, -1.49 ] Heterogeneity: not applicable Test for overall effect: Z = 15.90 (P < 0.00001) 6 CCB vs. 1-antagonist TOMHS 114 -12.9 (4.3) 121 -11.7 (5) 100.0 % -1.20 [ -2.39, -0.01 ] Subtotal (95% CI) 114 121 100.0 % -1.20 [ -2.39, -0.01 ] Heterogeneity: not applicable Test for overall effect: Z = 1.98 (P = 0.048) Test for subgroup differences: Chi2 = 192.21, df = 5 (P = 0.00), I2 =97%

-4 -2 0 2 4 Favours experimental Favours control

Calcium channel blockers versus other classes of drugs for hypertension (Review) 62 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Analysis 7.3. Comparison 7 Blood pressure reduction, Outcome 3 Sensitivity analysis: CCBs vs. ACE inhibitors.

Review: Calcium channel blockers versus other classes of drugs for hypertension

Comparison: 7 Blood pressure reduction

Outcome: 3 Sensitivity analysis: CCBs vs. ACE inhibitors

Studyorsubgroup CCBs ACEinhibitors MeanDifference Weight Mean Difference N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

1 Systolic blood pressure reduction AASK 217 -17 (17) 436 -16 (14.48) 1.3 % -1.00 [ -3.64, 1.64 ]

ALLHAT 9048 -11.5 (9.7) 9054 -10.5 (10.8) 97.3 % -1.00 [ -1.30, -0.70 ]

TOMHS 114 -15.6 (9.6) 119 -14.7 (9.8) 1.4 % -0.90 [ -3.39, 1.59 ] Subtotal (95% CI) 9379 9609 100.0 % -1.00 [ -1.29, -0.70 ] Heterogeneity: Chi2 = 0.01, df = 2 (P = 1.00); I2 =0.0% Test for overall effect: Z = 6.63 (P < 0.00001) 2 Diastolic blood pressure reduction AASK 217 -15 (8.98) 436 -14 (9.48) 1.6 % -1.00 [ -2.49, 0.49 ]

ALLHAT 9048 -9.3 (6.4) 9054 -8.7 (6.6) 96.3 % -0.60 [ -0.79, -0.41 ]

TOMHS 114 -12.9 (4.3) 119 -11.5 (5.5) 2.2 % -1.40 [ -2.66, -0.14 ] Subtotal (95% CI) 9379 9609 100.0 % -0.62 [ -0.81, -0.44 ] Heterogeneity: Chi2 = 1.75, df = 2 (P = 0.42); I2 =0.0% Test for overall effect: Z = 6.58 (P < 0.00001) Test for subgroup differences: Chi2 = 4.44, df = 1 (P = 0.04), I2 =78%

-4 -2 0 2 4 Favours experimental Favours control

APPENDICES Appendix 1. MEDLINE search strategy Database: Ovid MEDLINE(R) <2000 to current> Search Strategy: ------1. (calcium channel blockers or amlodipine or amrinone or bencyclane or bepridil or cinnarizine or conotoxins or diltiazem or felodipine or fendiline or ﬂunarizine or gallopamil or isradipine or lidoﬂazine or magnesium sulfate or mibefradil or nicardipine or nifedipine or nimodipine or nisoldipine or nitrendipine or perhexiline or prenylamine or verapamil or omega-agatoxin iva or omega- conotoxin gvia or omega-conotoxins).mp. 2. calcium adj2 (inhibit$ or agonist? or exogenous or blockader?).tw. 3. 1 or 2 4. hypertension/ 5. hypertens$.tw. 6. (blood adj pressure).tw.

Calcium channel blockers versus other classes of drugs for hypertension (Review) 63 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 7. or/4-6 8. 3 and 7 9. randomized controlled trial.pt. 10. controlled clinical trial.pt. 11. randomized.ab. 12. placebo.ab. 13. drug therapy.fs. 14. randomly.ab. 15. trial.ab. 16. groups.ab. 17. or/9-16 18. animals/ not (humans/ and animals/) 19. 17 not 18 20. 8 and 19 21. limit 20 to ed=20000101-20090430

Appendix 2. CENTRAL search strategy 1. (calcium channel blockers or amlodipine or amrinone or bencyclane or bepridil or cinnarizine or conotoxins or diltiazem or felodipine or fendiline or ﬂunarizine or gallopamil or isradipine or lidoﬂazine or magnesium sulfate or mibefradil or nicardipine or nifedipine or nimodipine or nisoldipine or nitrendipine or perhexiline or prenylamine or verapamil or omega-agatoxin iva or omega- conotoxin gvia or omega-conotoxins).mp. 2. calcium adj2 (inhibit$ or agonist? or exogenous or blockader?).tw. 3. 1 or 2 4. hypertension/ 5. hypertens$.tw. 6. (blood adj pressure).tw. 7. or/4-6 8. 3 and 7

Appendix 3. EMBASE search strategy 1. Randomized Controlled Trial/ 2. Clinical Trial/ 3. Multicenter Study/ 4. Controlled Study/ 5. Crossover Procedure/ 6. Double Blind Procedure/ 7. Single Blind Procedure/ 8. exp RANDOMIZATION/ 9. Major Clinical Study/ 10. PLACEBO/ 11. Meta Analysis/ 12. phase 2 clinical trial/ or phase 3 clinical trial/ or phase 4 clinical trial/ 13. (clin$ adj25 trial$).tw. 14. ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).tw. 15. placebo$.tw. 16. random$.tw. 17. control$.tw. 18. (meta?analys$ or systematic review$).tw. 19. (cross?over or factorial or sham? or dummy).tw. 20. ABAB design$.tw. 21. or/1-20

Calcium channel blockers versus other classes of drugs for hypertension (Review) 64 Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 22. human/ 23. nonhuman/ 24. 22 or 23 25. 21 not 24 26. 21 and 22 27. 25 or 26 28. hypertension/ 29. hypertens$.tw. 30. (blood adj pressure).tw. 31. or/28-30 32. exp calcium channel blockers/ 33. (calcium channel blockers or amlodipine or amrinone or bencyclane or bepridil or cinnarizine or conotoxins or diltiazem or felodipine or fendiline or ﬂunarizine or gallopamil or isradipine or lidoﬂazine or magnesium sulfate or mibefradil or nicardipine or nifedipine or nimodipine or nisoldipine or nitrendipine or perhexiline or prenylamine or verapamil or omega-agatoxin iva or omega- conotoxin gvia or omega-conotoxins).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name] 34. calcium adj2 (inhibit$ or agonist? or exogenous or blockader?).tw. 35. or/32-34 36. 27 and 31 and 35

HISTORY Protocol ﬁrst published: Issue 2, 2002 Review ﬁrst published: Issue 8, 2010

Date Event Description

6 October 2009 New search has been performed The ﬁst draft of full review was ﬁnished by authors

1 May 2009 New citation required and major changes Protocol re-published with new authors and amended methods.

23 August 2006 New citation required and major changes Protocol withdrawn by authors

CONTRIBUTIONSOFAUTHORS Ning Chen and Muke Zhou revised the protocol and drafted the review. Li He offered expert advice, reviewed the protocol, and was responsible for developing the review. Ning Chen and Mi Yang were responsible for including or excluding trials, data extraction and analysis. Cai Rongzhu offered expert advice. Jie Yang, Yuecai Wang and Xue Yang were responsible for searching for trials.

SOURCES OF SUPPORT Internal sources • West China Hospital, Sichuan University, China.

External sources • No sources of support supplied

DIFFERENCESBETWEENPROTOCOLANDREVIEW We added an outcome - cardiovascular mortality - in the review.

NOTES This protocol was ﬁrst published in The Cochrane Library in issue 2, 2002, by Onder G, Furberg CD, Moore A, Psaty BM, Pahor M. Subsequently, the protocol was withdrawn by the original authors in June 2006, because they were not able to continue working on it.

INDEX TERMS

Medical Subject Headings (MeSH) Adrenergic beta-Antagonists [therapeutic use]; Antihypertensive Agents [∗therapeutic use]; Calcium Channel Blockers [∗therapeutic use]; Cardiovascular Diseases [prevention & control]; Cause of Death; Hypertension [∗drug therapy]; Randomized Controlled Trials as Topic; Sodium Chloride Symporter Inhibitors [therapeutic use]

MeSH check words Humans