CD007187 Standard

Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension (Review)
Chen JMH, Heran BS, Wright JM
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2009, Issue 4 http://www.thecochranelibrary.com
Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension (Review) Copyright 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS HEADER . . . . . . . . . . ABSTRACT . . . . . . . . . PLAIN LANGUAGE SUMMARY . BACKGROUND . . . . . . . OBJECTIVES . . . . . . . . METHODS . . . . . . . . . RESULTS . . . . . . . . . . Figure 1. . . . . . . . . Figure 2. . . . . . . . . Figure 3. . . . . . . . . Figure 4. . . . . . . . . Figure 5. . . . . . . . . DISCUSSION . . . . . . . . AUTHORS CONCLUSIONS . . ACKNOWLEDGEMENTS . . . REFERENCES . . . . . . . . CHARACTERISTICS OF STUDIES DATA AND ANALYSES . . . . . HISTORY . . . . . . . . . . CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST . SOURCES OF SUPPORT . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 2 2 3 3 5 6 8 12 13 13 17 20 20 21 28 86 91 91 91 91 91
[Intervention Review]
Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension
Jenny MH Chen1 , Balraj S Heran2 , James M Wright1
1 Department
of Anesthesiology, Pharmacology and Therapeutics, University of British Columbia, Vancouver, Canada. 2 Peninsula Technology Assessment Group (PenTAG), Peninsula College of Medicine & Dentistry, University of Exeter, Exeter, UK Contact address: Jenny MH Chen, Department of Anesthesiology, Pharmacology and Therapeutics, University of British Columbia, 2176 Health Sciences Mall, Vancouver, BC, V6T 1Z3, Canada. koko_jesse@hotmail.com.
Editorial group: Cochrane Hypertension Group. Publication status and date: New, published in Issue 4, 2009. Review content assessed as up-to-date: 31 December 2008. Citation: Chen JMH, Heran BS, Wright JM. Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD007187. DOI: 10.1002/14651858.CD007187.pub2. Copyright 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
ABSTRACT Background Diuretics are widely prescribed for hypertension not only as a rst-line drug but also as a second-line drug. Therefore, it is essential to determine the effects of diuretics on blood pressure (BP), heart rate and withdrawals due to adverse effects (WDAEs) when given as a second-line drug. Objectives To quantify the additional reduction in systolic blood pressure (SBP) and diastolic blood pressure (DBP) of diuretic therapy as a secondline drug in patients with primary hypertension Search strategy CENTRAL (The Cochrane Library 2008, Issue 2), MEDLINE (1966-July 2008), EMBASE (1988-July 2008) and bibliographic citations of articles and reviews were searched. Selection criteria Double-blind, randomized, controlled trials evaluating the BP lowering efcacy of a diuretic in combination therapy with another class of anti-hypertensive drugs compared with the respective monotherapy (without a diuretic) for a duration of 3 to 12 weeks in patients with primary hypertension. Data collection and analysis Two review authors independently extracted the data and assessed trial quality. Main results Fifty-three double-blind RCTs evaluating a thiazide in 15129 hypertensive patients (baseline BP of 156/101 mmHg) were included. Hydrochlorothiazide was the thiazide used in 49/53 (92%) of the included studies. The additional BP reduction caused by the thiazide as a second drug was estimated by comparing the difference in BP reduction between the combination and monotherapy groups. Thiazides as a second-line drug reduced BP by 6/3 and 8/4 mmHg at doses of 1 and 2 times the manufacturers recommended starting dose respectively. The BP lowering effect was dose related. The effect was similar to that obtained when thiazides are used as a single agent. Only 3 double-blind RCTs evaluating loop diuretics were identied. These RCTs showed a BP lowering effect of a starting dose of about 6/3 mmHg.
Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension (Review) Copyright 2009 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 1
Authors conclusions Thiazides when given as a second-line drug have a dose related effect to lower blood pressure that is similar to when they are added as a rst-line drug. This means that the BP lowering effect of thiazides is additive. Loop diuretics appear to have a similar blood pressure lowering effect as thiazides at 1 times the recommended starting dose. Because of the short duration of the trials and lack of reporting of adverse events, this review does not provide a good estimate of the incidence of adverse effects of diuretics given as a second-line drug.
PLAIN LANGUAGE SUMMARY Diuretics effective in lowering blood pressure when given as a second drug Diuretics are a good rst-line treatment for high blood pressure (BP). Diuretics include loop diuretics and thiazides. We asked how much do diuretics reduce BP when used as the second drug to treat hypertension. We performed a search of the available scientic literature to nd all trial evidence to assess this question. Fifty-six trials were found. Fifty-three trials involved thiazide diuretics (92% with the drug hydrochlorothiazide) and included a total of 15310 participants. Adding a thiazide to another anti-hypertensive drug further reduces the BP by an additional 6/3 mmHg when given at the starting dose and reduces BP by 8/4 mmHg at 2 times the starting dose. This is approximately the same effect as when the drugs are used alone. A good estimate of the harms associated with diuretics cannot be estimated in this review because of the lack of reporting and the short duration of the trials.
BACKGROUND
Elevated blood pressure (hypertension) is a condition which contributes to the development of cerebrovascular disease, ischaemic heart disease, cardiac and renal failure. The main goal of treatment of hypertension is to reduce the risk of stroke, cardiovascular disease and death that are associated with elevated blood pressure. Reductions in blood pressure with the use of antihypertensive drugs have been associated with reductions in long-term outcomes and low-dose thiazides remain the preferred rst-line therapy in the management of hypertension (Wright 1999, Psaty 2003, Wright 2009). Based on the results of randomized controlled trials , it is clear that it is a challenge to reach and maintain blood pressure targets in many patients. Even with individualized monotherapy BP targets are not likely to be achieved with the rst drug used, even when titrated to high doses. For example the standard target of 140/ 90mmHg is only achieved in 50% of a population with a baseline diastolic blood pressure of 95-109 mmHg (Materson 1993). The pharmacologic rationale is that hypertension is a multifactorial condition with more than one mechanism, each of which contributes to a variable extent to its pathogenesis. Modication of one physiologic system by monotherapy usually triggers a compensatory response from another system, limiting the fall in blood pressure. Multiple inhibitory mechanisms are therefore likely to be more effective than a single one. The pharmacological rationale for combining two drugs of different classes is that by working at a separate site, interference with different effector pathways can be achieved. Also, the antihypertensive effectiveness of administration of a single drug might be lessened by counter-regulatory mechanisms in the body which tends to return BP values towards pre-treatment values. By combining antihypertensive agents that possess different mechanisms of action, each component drug can potentially neutralize or minimize counter-regulatory mechanisms triggered by the other, and thus help to further lower the blood pressure further. Antihypertensive therapeutics decisions are usually based upon the level of BP in a patient. Although other factors independent of BP lowering effect may contribute to the reduction in mortality and morbidity associated with antihypertensive drugs, BP lowering ability remains an important factor. A systematic review of the dose-related BP lowering efcacy of diuretics has been previously performed (Musini 2002). However, a systematic review of the additional reduction in BP with diuretics, when given in combination with other antihypertensive drugs, has not been previously conducted. It is possible for the effect of a diuretic as a second-line drug to be additive, sub-additive or synergistic. The aims of this systematic review are: 1) to quantify the additional BP reduction
2
achieved with a diuretic as a second drug versus placebo; and 2) to compare withdrawals due to adverse effects between a diuretic added as a second drug and placebo. This review should be able to provide clinicians with better information about the magnitude of BP lowering when a diuretic is given as second-line therapy in the management of elevated blood pressure.
Types of interventions Combination therapy with a diuretic plus other non-diuretic antihypertensive drug(s) versus other non-diuretic antihypertensive drug(s) alone. The addition of a diuretic must have been the only difference between the combination and monotherapy groups. In the case of xed-dose combination, the pill should be identical in appearance and taste to the individual components; in other cases where drugs are administered separately in the combination group, the monotherapy group should receive a matching placebo (double-dummy design) The diuretics that were included in this review were loop diuretics and thiazide or thiazide-like drugs. Potassium-sparing diuretics and aldosterone antagonists were not included in this review. The non-diuretic component included pharmacological agents in the following drug classes: Angiotensin-converting enzyme inhibitor (ACEI); calcium channel blocker (CCB); beta-blocker (BB); angiotensin receptor blocker (ARB); renin inhibitor (RI); and centrally-acting drugs (CAD) (but limited to guanabenz, rilmenidine, clonidine, moxonidine, methyldopa and guanfacine). All dosages and combinations of these drugs were considered. Trials in which titration to a higher dose was based on BP response were excluded. For forced titration trials, data from the lowest dose given within 3 to 12 weeks period were extracted.
OBJECTIVES
Primary objective
To quantify the additional reduction in systolic blood pressure (SBP) and diastolic blood pressure (DBP) of diuretics as second-line therapy in patients with primary hypertension.
Secondary objectives
To determine the effects of a second-line diuretic on variability of blood pressure. To determine the effects of a second-line diuretic on pulse pressure (PP). To quantify the effects of a second-line diuretic on heart rate (HR). To quantify the effects of a second-line diuretic on withdrawals due to adverse effects (WDAE).
Types of outcome measures
Primary outcomes
METHODS
Criteria for considering studies for this review
Types of studies Only randomized controlled trials (RCTs) were included and the study must have met the following criteria: Double-blind Parallel design with random allocation to treatment groups Washout period of at least 2 weeks prior to randomization Types of participants Men and non-pregnant women at least 18 years old with an ofce baseline SBP of at least 140 mmHg systolic and/or a DBP of at least 90 mmHg. Participants with signicant renal failure or creatinine level greater than 1.5 times the normal value were excluded. Participants were not restricted other baseline risks or co-morbid conditions.
Additional reduction in SBP and DBP with second-line diuretics. This was dened as the change from baseline (following washout) in ofce trough SBP and DBP at 3 to 12 weeks between the combination and monotherapy groups. If BP measurements were available at more than one time during the 3 to 12 week treatment period, the weighted mean of the BP data were used in this review. Trough BP was dened as the BP measurement taken before the next dosing schedule. In cases where timing of measurement was not provided, BP was assumed to have been taken at trough.
Secondary outcomes
Change in standard deviation of BP with combination therapy as compared to monotherapy Incidence of withdrawals due to adverse effects with combination therapy as compared to monotherapy Change in heart rate with combination therapy as compared to monotherapy Change in pulse pressure with combination therapy as compared to monotherapy
3
Search methods for identication of studies

The Database of Abstracts of Reviews of Effectiveness (DARE) and the Cochrane Database of Systematic Reviews were searched for related reviews. The following electronic databases were searched for primary studies: a) The Cochrane Central Register of Controlled Trials (CENTRAL) (2008, Issue 2) b) English language databases, including MEDLINE (1966-July 2008) and EMBASE (1988-July 2008) Electronic databases were searched using a strategy combining the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE with selected MeSH terms and free text terms to identify RCTs assessing a diuretic combined with other antihypertensive drug classes for the treatment of primary hypertension. No language restrictions were used. The MEDLINE search strategy was translated into the other databases using the appropriate controlled vocabulary as applicable. Full electronic database strategies for MEDLINE and EMBASE are included in Appendix 1 and Appendix 2. Other sources: a) Reference lists of all papers and relevant reviews identied b) Authors of trials reporting incomplete information were contacted to provide the missing information
In the case of missing information in the included studies, investigators were contacted (by email, letter and/or fax) to obtain the missing information. In the case of missing standard deviation of the change in BP or heart rate, the standard deviation was imputed based on the information in the same trial or from other trials using the same dose. The following hierarchy (listed from high to low preference) was used to impute standard deviation values: 1. Standard deviation of change from a different position than that of the BP/HR data used. 2. Standard deviation at the end of treatment. 3. Standard deviation at the end of treatment measured from a different position than that of the BP/HR data used. 4. Standard deviation at baseline (except if this measure was used for entry criteria). 5. Mean standard deviation of change from other trials using the same drug and dose. Risk of Bias Assessment The risk of bias in included studies was assessed using the Cochrane Collaborations recommended tool, which is a domain-based critical evaluation of the following domains: sequence generation; allocation concealment; blinding; incomplete outcome data; selective outcome reporting; and other sources of bias (Higgins 2008). Data Analysis and Statistical Considerations Data synthesis and analyses were done using the Cochrane Review Manager software, RevMan 5.0.20. Data for changes in BP and HR were combined using a weighted mean difference method. The withdrawals due to adverse effects was analyzed using relative risk, risk difference, and number needed to harm. If possible, subgroup analyses were used to assess the results for specic categories of participants: 1) Age: adults (18-69 yrs), older people (70 years and older) 2) Race: White; Black; other 3) Baseline severity of hypertension: mild; moderate; severe 4) Drug classes: The different classes of drug used in combination with a diuretic
Direct and indirect comparisons
Data collection and analysis
Study Selection The titles and/or abstracts obtained from the search strategies were screened by one reviewer (JMHC). During the initial abstract screening, those studies that were irrelevant to the review or clearly did not meet the inclusion criteria were rejected. The remaining trials were obtained in full text to assess whether they met the prespecied inclusion criteria. Trials were then assessed for inclusion eligibility by two reviewers independently (JMHC and BSH). Any discrepancies were resolved by a third reviewer (JMW). Trials with multiple publications were counted only once.
Data Extraction Data were extracted independently by two reviewers (JMHC and BSH) using a standardized form, and then cross-checked. All numeric calculations and graphic interpolations were conrmed by a second reviewer (BSH). The position of the patient during BP measurement may affect the BP lowering affect. When measurements were reported in more than one position, the order of preference was: 1) sitting; 2) standing; and 3) supine.
When possible, direct and indirect comparisons of effect sizes were performed between different doses of diuretics. In the direct method, only trials that randomized patients to various doses of diuretics in combination therapy were included in the analysis. In the indirect method, an adjusted indirect comparison and the associated standard error were calculated using the methods previously described by Bucher 1997 and Song 2003. A p-value less than 0.05 was considered statistically signicant for all comparisons. Tests for heterogeneity of the treatment effect
4
between trials were made using a standard chi-square statistic for heterogeneity. The xed effect model was applied to obtain summary statistics of pooled trials, unless signicant between-study heterogeneity was present, in which case the random effects model was used. This model provides a more conservative statistical comparison of the difference between combination therapy group and monotherapy group because a condence interval around the effect estimate is wider than a condence interval around a xed effect estimate. If a statistically signicant difference was still present using the random effects model, the xed effect pooled estimate and condence interval were used as the best estimate because of the tendency of smaller trials, which are more susceptible to publication bias, to be over-weighted with a random effects analysis.
See: Characteristics of included studies; Characteristics of excluded studies. See: Characteristics of included studies; Characteristics of excluded studies
Search ndings The search strategy was developed to identify all double-blind RCTs that assessed the BP lowering efcacy of combination therapy versus their individual components in hypertensive patients. A search was performed to identify relevant trials for this review, and also for 3 other reviews which addressed similar research questions for drugs inhibiting the renin-angiotensin system (ACE inhibitors, angiotensin receptor blockers and renin inhibitors) (Chen 2008a), beta blockers (Chen 2008b) and calcium channel blockers (Chen 2008c) as second-line drugs in combination therapy. In total, this search strategy identied 25 084 citations and only 56 (0.2%) trials met the inclusion criteria and had data suitable for analysis in this systematic review (Figure 1). Sixty other studies that met the inclusion criteria of this review were excluded because they did not provide extractable data (see below).
RESULTS
Description of studies
Figure 1. QUOROM ow diagram.*The protocols of the three other reviews have been published recently ( Chen 2008a; Chen 2008b; Chen 2008c)
Characteristics of Included Studies The characteristics of each included study are summarized in the Characteristics of included studies table. Of the 56 included studies, 53 (95%) were published in English, 1 in German, 1 in French, and 1 in Spanish. Funding sources were only reported in 28 (50%) of the included studies. All the studies that reported funding source were industry funded. Fifty-three of the included studies assessed thiazides (49 with hydrochlorothiazide, 2 with indapamide, 1 with clopamide, and 1 with chlorthalidone) and 3 studies assessed loop diuretics (2 with piretanide and 1 with frusemide). All the included studies were of parallel design and assessed the effect of a second-line diuretic. There were no studies found that assessed the effect of a diuretic as a third-line drug. There were also 7 studies that assessed the effect of a thiazide and a potassium-sparing diuretic combination as second-line therapy but these studies were excluded from this review (see Characteristics of excluded studies). Characteristics of Excluded Studies Sixty of the studies that met the preliminary inclusion criteria were excluded from this review. A majority of the excluded studies were crossover trials in which pre-crossover data were not provided. The reasons for exclusion of each trial are provided in the Characteristics of excluded studies table. Imputation of missing variance data The weighted mean standard deviation (SD) of both SBP and DBP changes were calculated from studies that provided the SD of SBP and DBP changes, respectively. Thirty-two (57%) of the included studies reported the SD of SBP change and thirty-four (61%) included studies reported the SD of DBP change. Eleven studies included in the other three reviews mentioned previously (Scholze 1998; Viskoper 1997; Messerli 1998; Levine 1995; Scholze 1999; Chan 1997; Pittrow 1997; Azizi 2000; Kuschnir 2004; Farsang 2001; Lessem 1989) also provided the SD of the changes in SBP and DBP. The values from all these studies were pooled in order to calculate the weighted mean estimates of the SD of the change in SBP and DBP for the combination and monotherapy groups. Table 1. Starting doses of Diuretics analyzed in the review Drug Furosemide Piretanide Hydrochlorothiazide Indapamide Type of Diuretic Loop Loop Thiazide Thiazide
This was based on the assumption that the effect on BP variability is similar across drug classes. Eight studies (Messerli 1998; Chan 1997; Kellaway 1993; Drayer 1995; Frishman 1994; Manning 1996; Farsang 2001; Frei 1994) were ultimately excluded from the calculation because their SD values were not within three standard deviations of the estimated weighted mean SD of SBP change. For the same reason, ve studies (Chan 1997; Farsang 2001; Drayer 1995; Lacourciere 2005; Frishman 1994) were excluded from the calculation of the weighted mean SD of DBP change. After these adjustments, the weighted mean SD of SBP and DBP change values for the combination group were 13.2 (SD 1.5) mmHg and 8.1 (SD 0.8) mmHg, respectively. For the monotherapy group, the weighted mean SD of SBP and DBP change values were 13.6 (SD 1.8) mmHg and 8.3 (SD 1.1) mmHg, respectively. There were no statistically signicant differences in the SD of SBP change or the SD of DBP change between the combination and monotherapy groups. These values were used according to the imputation hierarchy for trials that did not report the SD of the BP change value and in the nine included trials that reported outlier SD of the BP change values. The SD of BP change was imputed for 32/56 (57%) of the included studies. Of these studies, 4 (7%) were imputed using the SD of baseline SBP, 2 (4%) imputed using the SD of endpoint SBP, 5 (9%) imputed using the SD of endpoint DBP, 21 (38%) imputed using the weighted mean SD of SBP change from other trials, and 24 (43%) using the weighted mean SD of DBP change from other trials. Pooling of trials The diuretics were classied into two groups: 1) thiazide and thiazide-like diuretics; and 2) loop diuretics. The thiazide group was analyzed as a sub-class by pooling all trials that reported trough BP measurements. The loop diuretic group was analyzed separately. The doses of the individual diuretics were categorized as proportions of the manufacturers recommended starting dose (Table 1). This assumes that the starting dose recommended is effective in reducing BP and that all starting doses have a similar BP lowering efcacy. In the case where a range of starting doses is recommended by the manufacturer, the lowest dose is considered to be the starting dose (1x).
Starting dose/day for hypertension 40-80mg/day 6-12mg/day 12.5mg/day 1.25mg/day
Available in Canada? Yes No Yes Yes

7
Table 1. Starting doses of Diuretics analyzed in the review
(Continued)
Chlorthalidone Clopamide
Thiazide Thiazide
12.5mg/day 5-10mg/day
Yes No
Risk of bias in included studies

See Figure 2 Figure 2. Methodological quality graph: review authors judgments about each methodological quality item presented as percentages across all included studies.
Sequence generation and allocation concealment bias There was an unclear risk of bias in all the included studies in terms of sequence generation and allocation concealment due to poor reporting. The authors merely stated randomly assigned or using a randomized design without dening the process or the approach used. Authors should report their methods of sequence generation and allocation concealment clearly. Blinding bias Only double-blind, randomized, controlled trials were considered eligible for this review. Nearly all the trials merely stated that the trial was double-blind without providing further details about the blinding methods employed. It is assumed that the monotherapy and combination therapy tablets were identical in appearance. Moreover, none of the studies tested whether the double-blind procedure was successful at the end of the study. Blinding of the patient to either diuretic treatment or non-diuretic treatment could have been broken as patients on diuretics might have noticed in-
creased urine output in the rst few days of active treatment. Incomplete outcome data There was inconsistency in the methods of analysis and reporting of results. In a majority of the trials, not all patients randomized were included in the BP efcacy analysis as only those patients who did not violate the protocols or who completed the entire trial (i.e. per-protocol analysis and complete-patient analysis) were included. However, the number of randomized patients included in the efcacy analysis was greater than 80% in all except 6 studies (Genthon 1994; Hart 1991; Kellaway 1993; Oparil 1980; Prisant 2000; Vaicaitis 1980). In Hart 1991, 153/299 (51%) of the patients were included in the per-protocol analysis. In the other 5 studies, about 25-38% of the randomized patients were missing in the treatment arms in the efcacy analysis. In this review, exclusion of more than 20% of the randomized patients was judged to have a high risk of bias. The majority of the other included studies were judged to have a low risk of bias in terms of incomplete outcome
8
data (see Risk of bias in included studies). Selective Outcome Reporting All the included studies provided data on the change in BP, which was the primary outcome of this review. However, there were 4 studies that provided DBP without SBP data. There is a possibility of selective reporting bias for heart rate and withdrawals due to adverse effects since only 11% and 66% of the trials reported these outcomes, respectively. Publication Bias Publication bias, dened in this review as the selective publication of studies with positive results, is another source of bias that may have skewed the results of this review. The most common way to investigate whether or not an effect estimate is subject to publication bias is to examine for funnel plot asymmetry. The funnel plots appeared reasonably symmetrical upon visual examination of the funnel plots. However, funnel plots cannot be investigated adequately in this review because each comparison in this review was set up according to treatment arms (i.e. the unit of analysis is treatment arm and not the entire study as a whole). Selection Bias The method of recruitment of the participants for the trials may serve as another source of bias. Studies may have selected participants who are previously known to be responders to diuretics, either as a rst-line drug or second-line drug. This may result in an overestimation of the effect size of the BP lowering efcacy of the diuretics, as compared to the typical BP response observed in the general population taking diuretics. However, the degree of selection bias could not be assessed because the types of patients recruited were not described adequately. Also, in most studies, participants who were known to have allergic reactions to diuretics or any of the drugs used in the particular trial were excluded. Thus, these trials would have underestimated the incidence of adverse effects or withdrawals due to adverse effects associated with diuretics. Funding 28/56 (50%) of the included studies were industry sponsored. The other 28 (50%) studies did not report any funding source. Therefore, it was not possible to compare the results between industry funded and non-industry funded trials because there were no trials that were reported as being non-industry funded, Other factors Most BP measurements were taken just before the next dosing schedule (i.e. trough) in this review. However, 12/56 (21%) studies were included that did not mention the timing of the measurement.. If these studies all measured peak BP, and diuretics have a greater BP lowering effect at peak as compared to trough, then including these 12 studies in the overall effect estimate may result in an overestimation of the trough BP lowering efcacy of diuretics. However, a sensitivity analysis excluding these 12 trials did not result in statistically signicant difference in the effect estimate so including these trials and assuming they were taken at trough seems reasonable.
Effects of interventions
Blood Pressure Lowering Efcacy The additional BP lowering efcacy of adding a diuretic as secondline therapy to a non-diuretic (other) antihypertensive drug is summarized below. The outcome assessed is the difference in BP reduction between the combination (diuretic + other drug) and monotherapy (other drug alone) groups in parallel, double-blind RCTs. Using this approach, the difference is specied as the additional BP reduction induced by adding a diuretic as the second drug.
Thiazide plus other drug vs other drug alone
By comparing the difference in BP reduction between combination therapy (thiazide + monotherapy) and monotherapy groups, the additional BP reduction resulting from adding a thiazide as the second drug was estimated.
Thiazide plus ACEI vs ACEI alone Twenty-four included studies assessed the BP lowering efcacy of thiazide plus ACEI combination versus ACEI alone. DBP data, which was the primary outcome assessed in all these studies, was provided in all 24 studies, whereas SBP data was provided in only 20 of these studies. HCTZ was the thiazide that was given in all except one study (Safar 1994), which assessed indapamide at 0.625mg/day to 2.5mg/day. Although the dose of HCTZ was studied over a wide range (5mg/day to 45mg/day), a majority of the trials evaluated doses of 12.5mg/day and 25mg/day, which correspond to 1x and 2x the manufacturers recommended starting dose. In these studies, the drugs were given once daily except for Weinberger 1982, which administered the drugs three times daily (TID). Two studies provided SBP data for the addition of a thiazide 0.5x to an ACEI and the pooled result showed an additional reduction of -4.8 (95% CI -8.4, -1.2) mmHg. There were 4 studies with DBP data that showed the average additional effect on DBP reduction was not statistically signicant [-1.2 (95% CI -2.7, +0.3) mmHg]. The combination of thiazide plus ACEI was superior to ACEI alone in lowering both mean SBP and DBP when the dose of thiazide was 1x the manufacturers recommended starting dose or greater. Compared with ACEI alone, adding thiazide 1x lowered SBP by -5.2 (95% CI -6.3, -4.1) mmHg and DBP by -3.1 (95% CI -3.7, -2.5) mmHg. The addition of thiazide 2x to ACEI lowered SBP and DBP by -7.5 (95% CI -9.4, -5.7) mmHg and -3.8 (95% CI -4.7, -2.9) mmHg, respectively, as compared to ACEI alone. Based on an indirect comparison, the additional SBP reduction achieved by thiazide 2x was statistically greater than by thiazide 1x. However, when 3 of the studies (Genthon 1994; Kellaway 1993; Hart 1991) that were judged to have a high risk of incomplete outcome data bias were removed, the difference was no longer sta9
tistically signicant. A direct comparison of the 2 doses was performed with 3 studies (Chrysant 1994; Parati 2006; Yodfat 1994) which studied HCTZ at both 12.5mg/day and 25mg/day. From the pooled analysis of these 3 trials, adding HCTZ at 25mg/day also did not show a statistically signicant difference in the magnitude of BP reduction compared to adding HCTZ at 12.5mg/ day [SBP -1.6 (95% CI -4.3, +1.2) mmHg; DBP -1.1 (95%CI 2.4, +0.3) mmHg]. There was only one study assessing thiazide 3x (Pool 1997). This study showed an additional SBP and DBP reduction of -9.0 (95% CI -13.6, -4.4) mmHg and -3.5 (95% CI -5.9, -1.1) mmHg, respectively, with the addition of HCTZ 37.5mg/day. The magnitude of additional BP reduction achieved with this dose was not signicantly greater than that achieved with HCTZ 12.5mg/day or HCTZ 25mg/day. There was also only one study assessing thiazide 3.6x (Weinberger 1982). In this study, HCTZ was given at 15mg TID (45mg total daily dose). The addition of HCTZ 15mg TID resulted in a further SBP and DBP reduction of -18.5 (95% CI -23.1, -13.9) mmHg and -9.2 (95% CI -12.0, -6.4) mmHg, respectively. These reductions were statistically greater than those achieved with HCTZ at lower doses. The data showed that there is a possibility of greater BP reduction with higher HCTZ doses and with TID dosing. However, more studies are needed before conclusions can be made about doses greater than 2x the starting dose and different dosing schedules. A sensitivity analysis, where trials that measured BP in standing or supine positions were removed, was performed and showed that our effect estimate was not affected by the position of BP measurement. Removing the single study that assessed indapamide (Safar 1994) also did not statistically change the overall result. BP was measured before the next dosing schedule in all but 4 studies ( Kellaway 1993; Pool 1987; Safar 1994; Weinberger 1982), which did not mention the timing of BP measurement. Removing these 4 studies also did not statistically change our BP lowering effect estimate. Eight of the 24 included studies were industry sponsored; the other 16 trials did not report the source of funding. Therefore, it was not possible to investigate if funding source affected the results. The mean age of patients in the included studies ranged from 47 to 58 years and only one study had a study population with a mean age greater than 70 years (Hart 1991). A subgroup analysis based on age and gender was not performed due to insufcient data.
CI -2.5, -0.7) mmHg when HCTZ 6.25mg/day was added to an ARB. A pooled analysis of 12 studies showed an additional reduction in SBP of -7.1 (95% CI -8.0, -6.3) mmHg and DBP of -3.3 (95% CI -3.8, -2.8) mmHg with the addition of HCTZ 12.5mg/day to an ARB. A dose-response was observed, with a statistically signicantly greater reduction in BP with the addition of HCTZ 12.5mg/day as compared to HCTZ 6.25mg/day. Furthermore, there was a signicantly greater reduction in BP with HCTZ 25mg/day compared with HCTZ 12.5mg/day, based on an indirect comparison. This observation was conrmed by performing a direct comparison between the doses from 7 of the studies. The pooled analysis showed that the combination of HCTZ 25mg/day plus ARB resulted in a signicantly greater reduction in both SBP and DBP compared with the combination of HCTZ 12.5mg/day plus ARB [SBP -1.6 (95% CI -2.6, -0.5) mmHg; DBP -1.2 (95% CI -1.8, -0.5) mmHg]. A sensitivity analysis showed that the position of BP measurement did not signicantly effect the results. Two of the trials (Philipp 1997; Pool 2007a) did not record the timing of BP measurement. Removal of these trials also did not change the results. Ten of the studies were industry sponsored and the source of funding was not reported for the other 3 studies so a sensitivity analysis could not be conducted.
Thiazide plus renin inhibitor vs renin inhibitor alone Only one included study (Villamil 2007) was identied assessing the BP lowering efcacy of HCTZ plus renin inhibitor versus a renin inhibitor alone. Three different doses of HCTZ were assessed (6.25mg/day, 12.5mg/day and 25mg/day). Addition of HCTZ at each dose resulted in a signicantly greater reduction in both SBP and DBP. However, due to the lack of studies, there is insufcient data to conclude if there was a dose-response relationship.
Thiazide plus BB vs BB alone Nine included studies assessing the BP lowering efcacy of thiazide plus beta-blocker versus beta-blocker alone were identied. HCTZ 6.25 to 50mg/day was assessed in 7 studies, chlorthalidone 20 mg/day in 1 study (Bermudez 1982) and clopamide 10mg/day in another study (Safar 1973). Based on the available evidence, adding a thiazide at doses as low as 0.5x to 1x the manufacturers recommended starting dose to a beta-blocker signicantly reduced both SBP and DBP. Addition of thiazide 1.6x did not show a statistically signicant additional BP reduction. However, this observation was based on only one small trial evaluating the BP lowering efcacy of chlorthalidone 20mg/day added to a beta-blocker, which included only 24 patients in the combination group for the efcacy analysis (Bermudez 1982). Increasing the thiazide dose reduced BP further. There were 5 studies evaluating the addition of a thiazide at 2x the starting dose.
10
Thiazide plus ARB vs ARB alone Thirteen studies assessed the combination of thiazide plus ARB versus ARB alone. HCTZ administered once daily was the only thiazide studied in all 13 trials. Addition of HCTZ (6.25mg/day to 25mg/day) to an ARB signicantly reduced both SBP and DBP. A pooled analysis of 4 studies showed an additional reduction in SBP of -3.4 (95% CI -4.9, -1.8) mmHg and DBP of -1.6 (95%
Pooling the data from these 5 studies showed an additional reduction of -8.2 (95% CI -10.3, -6.2) mmHg in SBP and -4.0 (95% CI -5.3, -2.8) mmHg in DBP. Three of the trials were measured at trough and the other two trials did not report when BP was measured. Removing these 2 studies did not signicantly alter the overall BP effect size. The drugs were given once daily except in Asplund 1981, in which HCTZ 12.5mg was given twice daily (BID). Four of the 5 studies evaluated HCTZ whereas the other study (Safar 1973) evaluated clopamide 5mg given twice daily. A sensitivity analysis removing either of these studies did not change the overall additional BP effect size of thiazides added as a second drug to a beta-blocker. A direct comparison of the BP lowering efcacy of adding HCTZ 12.5mg/day and 25mg/day to a beta-blocker could be performed in 2 of the included studies (Lacourciere 1994; Papademetriou 2006). These studies were pooled and the data showed that adding HCTZ 25mg/day resulted in a numerically greater systolic BP reduction than adding HCTZ 12.5mg/day, but did not reach statistical signicance. The additional BP reduction resulting from the addition of thiazide 4x to a BB was assessed in 2 included studies. The studies were fairly small in size with a total number of 31 patients in the combination group . The pooled results showed an additional SBP reduction of -16.1 (95% CI -23.1, -9.2) mmHg and an additional DBP reduction of -6.3 (95% CI -10.6, -2.1) mmHg. As reected by the wide condence intervals, the estimate of the BP lowering efcacy at this dosage is imprecise. Thiazide plus CCB vs CCB alone There were 5 included studies assessing the BP lowering efcacy of thiazide plus CCB versus CCB alone. Two of the studies assessed HCTZ given once daily (Manning 1996; von Manteuffel 1995), another two studies assessed HCTZ given twice daily (Weir 1992; Pool 1993), and one study assessed indapamide given once daily (Prisant 2000). Due to a lack of studies for each of the different regimens, it was not possible to sufciently analyze the effect of each regimen on BP.
Three of the studies evaluated thiazides at 1x the manufacturers recommended starting dose and the other two studies at 2x the starting dose. The pooled results showed an additional reduction of -3.7 (95% CI -6.7, -0.6) mmHg in SBP and -2.6 (95% CI 4.5, -9.7) mmHg in DBP by adding thiazide 1x to a CCB. The additional reduction with thiazides 2x was -10.3 (95% CI -12.8, -7.8) mmHg in SBP and -7.0 (95% CI -8.5, -5.6) mmHg DBP. Based on an indirect comparison, the addition of a thiazide at 2x the manufacturers starting dose to CCB resulted in a signicantly greater reduction in BP compared with adding a thiazide at 1x the starting dose.
Thiazide plus centrally acting drug vs centrally acting drug alone There was only one included study assessing the BP lowering efcacy of a thiazide plus a centrally acting drug versus a centrally acting drug alone. This was a fairly small trial with only 42 patients in the HCTZ 25mg/day plus moxonidine combination group and 37 patients in the moxonidine monotherapy group (Frei 1994). This trial showed that adding HCTZ 25mg/day resulted in an additional SBP reduction of -7.0 (95% CI -12.9, -1.1) mmHg and a DBP reduction of -4.0 (95% CI -7.8, -0.3) mmHg. However, as reected by the wide condence intervals, the precision of our estimate of the additional BP lowering efcacy of adding a thiazide to a centrally-acting drug is low.
Summary of the additional BP lowering efcacy of thiazides when added to other antihypertensive drug classes Figure 3 provides an overview of the additional BP reduction that was observed when a thiazide was given in combination with another class of antihypertensive drug. The dose of thiazides ranged from 0.4x to 4x the manufacturers recommended starting dose. It is evident that the addition of a thiazide 0.4x to 4x as the second drug resulted in a statistically signicant additional BP reduction. The magnitude of the additional BP reduction ranged from 4/2 to 14/6 mmHg.
11
Figure 3. Blood pressure lowering efcacy of thiazides (as multiples of the starting dose) in combination with other antihypertensive drug classes.
A dose-response relationship was established in which higher doses of thiazides resulted in a signicantly greater reduction in BP. Thiazide 2x was superior to thiazide 1x in lowering both SBP and DBP, with a 2/1 mmHg difference that was statistically signicant. However, data were limited for thiazides greater than 2x the starting dose. There was one study assessing thiazide 3x (od dosing), one study assessing thiazide 3.6x (tid dosing), and 2 studies assessing thiazide 4x (bid dosing). Pooled analysis of these 4 studies showed that the additional reduction in SBP and DBP with thiazide 3-4x was -14.2 (95% CI -17.2, -11.3) mmHg and -6.0 (95% CI -7.7, -4.3) mmHg, respectively. Although indirect comparisons showed that thiazide 3-4x resulted in an additional BP reduction that was statistically higher than those achieved with thiazides 2x, no solid conclusions could be made about this difference because of the lack of studies and the difference in dosing schedules. Moreover, both of the studies assessing thiazide 4x were judged to have a high risk of incomplete outcome data bias (Oparil 1980; Vaicaitis 1980).
Subgroup analysis based on the class of drug used in combination with HCTZ
A majority of the available data were for HCTZ 12.5mg/day and HCTZ 25mg/day. A sub-group analysis was done to determine if the rst drug has a signicant effect on the BP lowering of a thiazide given as a second drug. Most of the data were for the combination of HCTZ with an ACEI or an ARB. As shown in Figure 4, the additional DBP reductions resulting from adding HCTZ 12.5mg/day to another class of antihypertensive drugs were similar with overlapping 95% condence intervals. However, based on indirect comparisons, the additional systolic BP reduction with HCTZ 12.5mg/day was signicantly greater when added to an ARB as compared to an ACEI or a renin inhibitor. Since these 3 classes of drugs are believed to lower BP by inhibiting the renin-angiotensin-aldosterone system (RAAS), albeit at different levels, one would not expect that adding HCTZ to one class would be different from adding it to the other in terms of magnitude of BP reduction. Furthermore, when HCTZ 25mg/day was added to the other antihypertensive drugs, the class of drugs to which HCTZ was added to did not have a statistically signicant effect on the magnitude of the additional BP reduction ( Figure 5). Therefore, the difference that was found between adding HCTZ 12.5mg/day to ARB and to the other drug classes is likely to be due to chance.
12
Figure 4. Blood pressure lowering efcacy of HCTZ 12.5mg/day in combination with other antihypertensive drug classes.
Figure 5. Blood pressure lowering efcacy of HCTZ 25mg/day in combination with other antihypertensive drug classes.
13
Loop diuretic plus other drug vs other drug alone
By comparing the difference in BP reduction between combination therapy (a loop diuretic + monotherapy) and monotherapy groups, the additional BP reduction resulting from adding a loop diuretic as the second drug was estimated. Loop diuretic plus ACEI vs ACEI alone There were two included studies assessing the BP lowering efcacy of the combination of piretanide and ACEI versus ACEI alone. The rst study (Homuth 1993) assessed two different doses of piretanide (3mg/day and 6mg/day) whereas the other study (Thijs 1995) assessed only one dose (6mg/day). No included studies assessed piretanide at any other doses in combination with an ACEI. Homuth 1993 showed that the addition of piretanide 3mg/day, which corresponds to 0.5x the manufacturers recommended starting dose, to ACEI did not result in a signicant additional BP reduction. At 1x the starting dose, a statistically signicant additional BP reduction was observed when the results from both studies were pooled [SBP -6.5 (95% CI -9.0, -4.0) mmHg; DBP -3.1 (95% CI -4.5, -1.7) mmHg]. Loop diuretic plus BB vs BB alone Only one included study (Chadha 1983) compared the BP lowering efcacy of a loop diuretic plus a beta-blocker versus a beta-
blocker alone. In this study, the combination group was given frusemide 40mg/day plus a beta blocker. There were only 16 patients in the combination group and 11 patients in the betablocker monotherapy group. Therefore, due to the lack of data for this combination, our effect estimates have extremely wide condence intervals. This study showed that the addition of frusemide 40mg/day (2x the manufacturers recommended starting dose) resulted in a statistically signicant additional reduction of -13.0 (95% CI -33.0, -7.0) mmHg in SBP, but a non-signicant -8.0 (95% CI -19.0, +3.0) mmHg reduction in DBP. Pulse pressure
Thiazides as second-line
Pulse Pressure (PP) was not reported as an outcome in any of the included studies. Therefore, the value of change in PP was calculated by subtracting DBP change from SBP change for each treatment arm in the trial. Using this approach, the change in PP can only be calculated from trials that provided data for both SBP and DBP. Both SBP and DBP data were provided in 47/49 (96%) included studies assessing HCTZ. Table 2 summarizes the change in PP in the combination and monotherapy groups. The data demonstrate a trend towards a greater additional PP reduction with increasing doses of HCTZ in the combination group (Table 2).
Table 2. Pulse Pressure Reduction at end of treatment. Combination therapy versus monotherapy: Hydrochlorothiazide as a second drug Hydrochlorothiazide # of studies (multiples of starting dose) # of patients (combo/mono) Weighted mean change in pulse pressure (95% CI) - combination group (with HCTZ) 1563/1720 3974/4072 2913/2886 185/182 -3.1 (-4.4, -1.9) -5.9 (-7.6, -3.8) -8.6 (-11.4, -5.9) -6.8 (-12.2, -1.4) Weighted mean change in pulse pressure (95% CI) - monotherapy group (no HCTZ) -1.3 (-2.3, -0.3) -3.0 (-4.4, -1.6) -4.0 (-6.2, -1.8) +0.7 (-0.7, +2.2)
0.5x 1x 2x 3x-4x
10 30 22 4
14
An estimate of the reduction in PP achieved by indapamide, clopamide or chlorthalidone (as a second drug) cannot be estimated separately because data were obtained from only 1 study each. However, pooling of data for all thiazides did not signicantly alter the results. To determine the additional effect of all thiazides as a second drug on PP, the data for indapamide, clopamide and chlorthalidone were combined with HCTZ (Table 3). There was a trend towards a greater PP reduction with higher doses of thiazides with an additional reduction in PP of -7.5 (95% CI 11.9, -3.2) mmHg with thiazide 3-4x. Table 3. Difference in Pulse Pressure Reduction at end of treatment. Combination therapy versus monotherapy: Thiazides as a second drug class Thiazides (multiples of starting dose) 0.5x 1x 2x 3x-4x
Loop diuretics as second-line
# studies 10 30 25 4
Difference in PP between combination and monotherapy (95% CI) -1.7 (-3.2, -0.1) -2.8 (-5.0, -0.7) -4.7 (-6.9, -2.4) -7.5 (-11.9, -3.2)
Due to the lack of studies with the loop diuretics, the reduction in pulse pressure (PP) could not be estimated. BP variability
Baseline variability The standard deviation of BP at baseline was reported in 27/53 (51%) included studies. As shown in Table 4, there was no statistically signicant difference between the variability of SBP (p=0.6) or DBP (P=0.5) at baseline between the combination and monotherapy groups. Table 4. Variability of SBP and DBP at baseline (thiazides) Combo group SBP Weighted mean SD SD of weighted mean SD 14.1 2.8 Mono group 13.8 1.8
15
Table 4. Variability of SBP and DBP at baseline (thiazides)
(Continued)
DBP
Weighted mean SD SD of weighted mean SD
4.7 1.1 NS
4.5 0.9 NS
t-test
Combo group vs mono group
Baseline vs endpoint variability The standard deviations of BP at baseline (after the run-in period) and the standard deviations of BP at endpoint were compared in 9 included studies. As shown in Table 5, there was no statistically signicant difference between the SBP variability at baseline and endpoint in the combination group or in the monotherapy group. For DBP variability, the baseline SDs were statistically signicantly lower than the endpoint values in both the combination and monotherapy groups, an effect likely due to the fact that all studies had DBP entry criteria (Musini 2009). Table 5. Standard deviations of BP at baseline vs. endpoint in trials with DBP entry criteria combination group Weighted mean SD of SBP At baseline (SD) At endpoint (SD) t-test baseline vs. Endpoint Weighted mean SD of DBP At baseline (SD) At endpoint (SD) t-test baseline vs. endpoint 14.1 (2.2) 15.7 (1.9) p = 0.1 5.0 (0.9) 8.6 (1.1) P<0.0001 Monotherapy group 14.3 (2.5) 16.0 (2.2) p = 0.3 4.9 (0.9) 8.4 (1.0) P<0.0001
Combination vs monotherapy To determine if the addition of thiazides affects BP variability, the standard deviations of BP at the end of treatment were compared between the combination therapy and monotherapy groups. The standard deviation of BP at endpoint was reported in 9 included studies. As shown in Table 5, the weighted mean SD of SBP at the end of treatment was 15.7 mm Hg for the combination group and 16.0 mm Hg for the monotherapy group. The weighted mean SD of DBP at the end of treatment was 8.6 mm Hg for the combination group and 8.4 mm Hg for the monotherapy group. Variability of BP was not signicantly different between the combination
group and monotherapy group for both SBP (p=0.8) and DBP (p=0.7).
Loop diuretic as second-line
Due to the lack of studies, the effects of loop diuretics on BP variability could not be estimated.
Heart rate
16
Heart rate data were extracted from 6 (11%) included studies, of which 4 studies assessed the combination with HCTZ (Vaicaitis 1980, Asplund 1981, Frishman 1995, Manning 1996). The addition of HCTZ to another antihypertensive drug class did not show a statistically signicant change in heart rate [+0.5 (95% CI -1.1, 2.1) beats/min]. There was only one trial each for clopamide and chlorthalidone so no conclusions could be drawn about their effect on heart rate.
None of the included studies assessing loop diuretics provided heart rate data. Withdrawals due to adverse effects
Data on WDAEs during the 3 to 12 weeks of treatment were extracted from 35/53 (66%) of the included studies for analysis. Of the 35 trials that reported WDAE, 4 studies reported no WDAE in both combination and monotherapy groups. Overall, 178/5944 (3%) of the patients in the combination group withdrew due to adverse effects compared with 84/3314 (2.5%) of the patients in the monotherapy group. There was no statistically signicant difference in WDAE between the combination and monotherapy groups [RR 1.09 (95% CI 0.84, 1.42)].
By comparing the difference in BP reduction between the combination (thiazide + 1 other drug) and monotherapy (placebo or no treatment + 1 other drug) groups, the additional BP reduction with a thiazide as second-line therapy was estimated. Hydrochlorothiazide was assessed in 49/53 (92%) of the included studies. The contribution of the thiazide component was statistically signicant in terms of the magnitude of BP reduction. Participants who received combination therapy (with a thiazide) experienced greater reductions in both systolic and diastolic BP than participants receiving monotherapy (without thiazide). This is evident at all doses that were assessed in the review. The decrease in SBP/DBP ranged from an additional 4/2 mmHg for lower dose thiazides to 14/6 mmHg for higher dose thiazides. In this review, the doses of thiazides have been categorized according to multiples of the manufacturers recommended starting dose ranging from 0.4x to 4x and data were pooled based on this categorization. A dose-response relationship was demonstrated at each increment of thiazide dose. This represents a very robust demonstration that the BP lowering effect of thiazides is dose-dependent. It is worth noting that the data are based primarily on white patients. Although non-white (blacks and others) patients were included in the efcacy analysis, they consisted of only a minority of the patients and most studies did not provide separate data for this subgroup. Therefore, our results are mostly generalizable to white hypertensive patients.
What is the additional BP reduction of loop diuretics when given in addition to another class of antihypertensive drugs?
There are limited data available for the BP lowering efcacy of loop diuretic as a second drug. Only three included studies provided these data. Adding a loop diuretic as a second drug was effective starting at 1x the manufacturers recommended dose, resulting in an additional reduction of -6.5 (95% CI -9.0, -4.0) mmHg in SBP and -3.1 (95% CI -4.5, -1.7) mmHg in DBP. Based on a small study with only 16 patients in the combination group, adding loop diuretic 2x showed an additional reduction of -13.0 (95% CI -33.0, -7.0) mmHg in SBP and -8.0 (95% CI -19.0, +3.0) mmHg in DBP. Due to a paucity of data at each dose, a dose-response relationship could not be determined. Unfortunately, due to the wide condence intervals, it is not possible to assess how loop diuretics compare with thiazides.
Based on 2 included studies, adding a loop diuretic as the second drug was not shown to have any signicant impact on the number of withdrawals due to adverse effects.
DISCUSSION What is the additional BP reduction of thiazide and thiazide-like diuretics when added to other classes of antihypertensive drugs?
Blood pressure efcacy data for the thiazide class was extracted from 53 included studies. Since most of the studies included multiple comparison treatment arms, not all data were extracted for this review. Only data that addressed the comparisons relevant to this review were included. These data were taken from a total of 15 129 hypertensive patients, 9483 treated with combination therapy and 5646 treated with monotherapy. These patients had a BP at baseline averaging 156/101 mmHg and a mean age of 54 years. These studies ranged from 3 to 12 weeks with an average 6-week double-blind treatment period.
Is there a difference in the reduction of blood pressure between adding thiazides to different classes of drugs?
As described above, the available data have demonstrated a doserelated BP lowering effect of thiazides as a second-line drug over the range of 0.4x to 4x the manufacturers recommended starting
17
dose. In order to determine if there were a difference in the additional BP reduction achieved between adding thiazides to different classes of rst-line drugs, comparisons were made for the doses of HCTZ that had the most available data, 12.5mg/day and 25mg/day. Based on indirect comparisons, the choice of the rst drug did not have a signicant effect on the additional BP lowering of a thiazide when given as a second-line drug. A more direct way of assessing if the addition of a thiazide to different drug classes leads to differences in BP lowering would be by meta-analyzing head-to-head combination trials where the same dose of HCTZ was added to two different classes of drugs. The data here give a pretty good indication that the effect is independent of the rstline drug, a nding that was not expected. Based on their proposed mechanisms of action, most clinicians believe that adding a thiazide to an ACE inhibitor or ARB would have a greater effect than adding a thiazide to a CCB.
therapy?
The BP lowering efcacy of HCTZ monotherapy versus placebo in patients with mild to moderate primary hypertension has been determined in a systematic review by Musini 2002. By performing an indirect comparison of the results between this review and that of Musini 2002 , we can investigate whether the additional BP achieved by adding HCTZ as the second drug in combination therapy differs from that achieved by administering HCTZ alone as monotherapy. Based on the best available evidence, the estimated additional BP that can be expected when HCTZ 12.5mg/day is added as a second drug is -6.0 (95% CI -6.5, -5.4) mmHg for SBP and -3.1 (95% CI 3.4, -2.8 ) mmHg for DBP. The estimated additional BP reduction that can be expected with the addition of HCTZ 25mg/day as a second-line drug is -8.0 (95% CI -8.7, -7.3) mmHg for SBP and 4.0 (95% CI -4.4, -3.6) mmHg for DBP. Musini 2002 showed that BP lowering efcacy of HCTZ 12.5mg/day as a rst-line agent versus placebo was -5.7 (95% CI -7.0, -4.5) mmHg for SBP and 3.9 (95% CI -4.7, -3.0) mmHg for DBP. This review also showed that HCTZ 25mg/day lowered SBP by -8.5 (95% CI -10.4, -6.6) mmHg and DBP by -4.7 (95% CI -5.8, -3.5) mmHg (see Table 6 and Table 7).
Is there a difference in the blood pressure lowering efcacy of diuretics given as initial therapy or as a second line drug in combination
Table 6. SBP reduction achieved by addition of HCTZ either in combination with other drugs or as a single drug. Fixed effect model (95%CI). Drug Dose HCTZ in combination SBP reduction (95%CI) # patients (combination group) -5.4) 4190 HCTZ as monotherapy* SBP reduction (95%CI) # patients (HCTZ group) Signicance
HCTZ mg/day
12.5 -6.0 (-6.5, mmHg -8.0 (-8.7, mmHg
-5.7 (-7.0, mmHg
-4.5) 579
NS
HCTZ 25 mg/day
-7.3) 2913
-8.5 (-10.4, -6.6) 368 mmHg
NS
*adopted from Musini 2002
Table 7. DBP reduction achieved by addition of HCTZ either in combination with other drugs or as a single drug. Fixed effect model (95%CI) Drug Dose HCTZ in combination DBP reduction (95%CI) # patients (combination group) HCTZ as monotherapy* DBP reduction (95%CI) # patients (HCTZ group) Signicance
18
Table 7. DBP reduction achieved by addition of HCTZ either in combination with other drugs or as a single drug. Fixed effect model (95%CI) (Continued) HCTZ 12.5mg/day -3.1 (-3.4, mmHg HCTZ 25mg/day -4.0 (-4.4, mmHg -2.8) 4279 -3.9 (-4.7, mmHg -4.7 (-5.8, mmHg -3.0) 579 NS
-3.6) 3093
-3.5) 368
NS
*adopted from Musini 2002
Indirect comparisons showed that the additional BP reduction (both systolic and diastolic) that resulted from the addition of HCTZ 12.5mg/day or HCTZ 25mg/day to another class of drugs was not statistically different from the BP reduction with HCTZ 12.5mg/day or HCTZ 25mg/day monotherapy, respectively (Table 6). Therefore, our review has demonstrated that HCTZ has a similar BP lowering efcacy as rst-line and second-line therapy, and that the BP lowering effect of HCTZ as second-line therapy is additive.
Does age have an effect on BP lowering of diuretics?

The age inclusion criteria for most studies ranged from 18 to 80 with an average of 54 years. There was only one study that included only elderly patients (Hart 1991) with a mean age of 71 yrs. Due to the lack of reporting and limited data, a subgroup analysis of older versus younger patients could not be performed.
termine endpoint variability were based on endpoint SDs. Analysis of the available data showed that adding a diuretic as second-line therapy did not alter resting SBP and DBP variability since there were no statistically differences between the combination and the monotherapy groups. This conrms the results of others for thiazides (Musini 2009) and is similar to what has been shown for ACE inhibitors (Heran 2009a), ARBs (Heran 2009b) and CCBs (Wong 2007). Because mean values were used, both inter- and intra-individual variabilities were accounted for. To determine the effects on intra-individual variability, individual patient BP data are needed.
What is the effect of second-line diuretics on pulse pressure?

Pulse pressure (PP) has become increasingly recognized as an independent risk factor for cardiovascular events (Franklin 1999; Haider 2003). Although PP has not been reported as one of the primary or secondary endpoints in any of the included studies, we were able to calculate it from trials that provided both SBP and DBP data. Fifty-one (96%) of the included studies assessing a thiazide provided both SBP and DBP data, whereas the other 2 studies only reported DBP data. By subtracting the change in DBP from the change in SBP for each of the 51 trials, we found that PP was signicantly reduced by adding a thiazide as a second drug at the dose range studied (0.4x to 4x the manufacturers recommended starting dose). There is a possibility of a dose-response relationship because there was a trend towards a greater reduction in PP with higher doses of thiazides (see Table 2). Hydrochlorothiazide was the thiazide used in all except 3 studies (Prisant 2000; Bermudez 1982; Safar 1973). Removal of these 3 studies did not signicantly alter the overall results. Since there were only 3 studies assessing loop diuretics at various doses, their effect on pulse pressure could not be assessed in this review.
19
Does co-morbidity have an effect on BP lowering of diuretics?

It was not possible to perform a subgroup analysis of hypertensive patients with other co-morbid diseases. None of the trials specically selected for patients with co-morbid conditions and the majority of the studies excluded patients with signicant major diseases, including renal, cardiovascular, hepatic and neurologic problems. Furthermore, data for these subgroups of patients, if included, have not been reported separately.
What is the effect of second-line diuretics in combination therapy on BP variability?

The standard deviations (SD) of the BP at baseline and/or endpoint were reported in 27/53 (51%) of the included studies. The endpoint variabilities of the combination and monotherapy groups were compared in order to determine the effect of adding diuretics as a second drug on BP variability. The values used to de-
What is the effect of second-line diuretics on heart rate?

The data for heart rate were reported incompletely in many of the studies, where only p-values were given or changes were described as not signicantly different. For the loop diuretics, there were no included studies providing heart rate data for analysis. Quantitative heart rate data could only be extracted from 6 (11%) of the included studies assessing thiazides. Pooled analysis of these 6 trials showed that addition of a thiazide to another antihypertensive drug did not signicantly affect heart rate [+0.4 (95% CI 1.1, +2.0) beats/min].
the addition of a thiazide by 4/2, 6/3, 8/4, 14/6 mmHg at doses 0.5x, 1x, 2x and 3x-4x respectively. 3. There was a trend towards a dose related pulse pressure reduction with thiazides. 4. Adding a thiazide diuretic as second-line therapy does not affect resting BP variability 5. Adding a thiazide diuretic as second-line therapy does not affect resting heart rate 6. Adding a thiazide diuretic as second-line therapy did not change the rate of withdrawals due to adverse events within 3-12 weeks treatment period. However, only 35/53 (66%) of the studies reported this outcome and there is a possibility for selective outcome reporting bias. 7. The magnitude of the BP reduction achieved by HCTZ given as a second-line drug is similar to the magnitude of BP lowering for HCTZ alone. Thus the BP lowering effect is additive. 8. The drug class to which thiazide was added did not appear to affect the additional BP lowering of the thiazide. 9. The evidence for loop diuretics is weak but they appear to reduce BP by 6/3 mmHg at the recommended starting dose.
What is the effect of second-line add-on diuretics on withdrawals due to adverse effects?
For the thiazides, the number of withdrawals due to adverse effects during the 3-12 week treatment period was reported in 35/53 (66%) of the included studies. Consistent with the data for thiazides given as monotherapy (Musini 2002), adding a thiazide as a second-line drug did not result in a signicant increase or reduction in withdrawals due to adverse effects [RR 1.09 (95% CI 0.84, 1.42)]. The type of drug class to which thiazides was added also did not seem to signicantly affect the results. Not all studies reported the reasons for withdrawals, and the total adverse events in the studies were, on average, not statistically different between combination and monotherapy groups. It is worth noting that most studies excluded patients with previous known allergic reactions to diuretics or any of the drugs used in these studies. Therefore, this review is not a good assessment of adverse effects in a general population taking these drugs for long-term therapy. Due to the lack of included trials, WDAE data for loop diuretics are very limited. Meta-analysis of the 2 studies where a loop diuretic was added to ACEI shows that WDAE were not significantly changed [RR 0.94 (95% CI 0.25, 3.5)] as compared to ACEI monotherapy.
Implications of ndings
Thiazides produce a reproducible dose related additive blood pressure lowering effect when given as the second drug for hypertension.
Implications for research

1. No trials were found that assessed the additional BP lowering efcacy of thiazides as a third line drug. Such trials are needed to determine the magnitude of BP reduction in that setting. 2. Although peak BP was not one of the primary outcomes in this review, it was seldom reported in the included trials. Trials should measure and report BP data for both trough and peak effects. 3. Trials should always report all withdrawals due to adverse effects and serious adverse events in each of the treatment arms of RCTs. 4. The included trials were conducted in a predominantly white population. Further research of the BP lowering efcacy of thiazides in black populations is needed.
AUTHORS CONCLUSIONS Implications for practice
Findings of this review

1. Adding a thiazide diuretic as second-line agent in combination therapy resulted in a greater reduction in BP as compared to monotherapy (without thiazide). 2. The additional BP reduction induced by thiazides was dosedependent. The SBP/DBP decreased further from baseline with
ACKNOWLEDGEMENTS
20
The authors would like to acknowledge the assistance provided by Mr. Stephen Adams, who retrieved the papers for this review, and the Cochrane Hypertension Group.
REFERENCES
References to studies included in this review

Asplund 1981 {published data only} Asplund J. A xed-ratio combination of metoprolol and hydrochlorothiazide (Co-betaloc(TM)) in essential hypertension: A comparison between the individual drugs. Current Therapeutic Research, Clinical & Experimental 1981;29(3):387394. [: EMBASE1981130462] Benz 1998 {published data only} Benz JR, Black HR, Graff A, Reed A, Fitzsimmons S, Shi Y. Valsartan and hydrochlorothiazide in patients with essential hypertension. A multiple dose, double-blind, placebo controlled trial comparing combination therapy with monotherapy. Journal of Human Hypertension 1998;12:8616. [MEDLINE: MEDLINE 99098179; : CN00158693] Bermudez 1982 {published data only} Bermudez J.A. Kornhauser Araujo C. Parra Carrillo J.Z. Paz Barahona M. Management of essential arterial hypertension with a single daily dose of a xed combination of slow-release oxprenolol and chlorthalidone [MANEJO DE LA HIPERTENSION ARTERIAL ESENCIAL CON LA COMBINACION FIJA DE OXPRENOLOL DE LIBERACION LENTA MAS CLORFTALIDONA, EN UNA DOSIS AL DIA. ENSAYO COMPARATIVO Y MULTICENTRICO]. Investigacion Medica Internacional 1982;9(2):156163. Brown 1990 {published data only} Brown CL, Backhouse CI, Grippat JC, Santoni JP. The effect of perindopril and hydrochlorothiazide alone and in combination on blood pressure and on the renin-angiotensin system in hypertensive subjects.. European Journal of Clinical Pharmacology 1990;39(4): 32732. [MEDLINE: MEDLINE 91168986; : CN00073926] Camera 1995 {published data only} Camera MI, Waisman GD, Galarza CR, Ale J, Arahabety A, Saggese O, Magi MI, Mayorga LM. Multi-centre study of the antihypertensive effect of lisinopril (20 mg) and a xed combination of lisinopril (20 mg) and hydrochlorothiazide (12.5 mg) once daily in mild to moderate essential hypertension.. Drug Development Research. 1995;34(suppl. 2):2023. [: EMBASE 1995095667; CN00169283] Waisman GD, Galarza CR, Ale J, Mayorga LM, et al.Comparison of the Haemodynamic Effects of Lisinopril and Lisinopril Plus Hydrochlorothiazide During Rest and Orthostatic Stress. Drug Development Research 1995;Suppl 2:2429. Chadha 1983 {published data only} Chadha DR, Houtzagers JJ. A comparison of a combination of penbutolol and frusemide with the two compounds individually in the treatment of hypertension. British Journal of Clinical Pharmacology 1983;16(5):5614. [MEDLINE: MEDLINE6357258; : CN00032826]
Chrysant 1994 {published data only} Chrysant SG. Antihypertensive effectiveness of low-dose lisinopril-hydrochlorothiazide combination. A large multicenter study. Archives of Internal Medicine 1994;154(7):73743. [MEDLINE: MEDLINE 94197508; : EMBASE 1994117160, CN00100251] Chrysant 1996 {published data only} Chrysant SG, Fagan T, Glazer R, Kriegman A. Effects of benazepril and hydrochlorothiazide, given alone and in low- and high-dose combinations, on blood pressure in patients with hypertension. Archives of Family Medicine 1996;5(1):17-24; discussion 25. [MEDLINE: MEDLINE8542050; : CN00122483] Chrysant 2004 {published data only} Chrysant SG, Weber MA, Wang AC, Hinman DJ. Evaluation of antihypertensive therapy with the combination of olmesartan medoxomil and hydrochlorothiazide.. American journal of hypertension : journal of the American Society of Hypertension 2004; 17(3):2529. Drayer 1995 {published data only} Drayer JI, Stimpel M, Fox A, Weber M. The Antihypertensive Properties of the Angiotensin-Converting Enzyme Inhibitor Moexipril Given Alone or in Combination with a Low Dose of a Diuretic [Am J Ther]. 1995 2;8:525531. Fernandez 1994 {published data only} Fernandez M, Madero R, Gonzalez D, Camacho P, Villalpando J, Arriaga J. Combined versus single effect of fosinopril and hydrochlorothiazide in hypertensive patients. Hypertension 1994;23 (1 Suppl):I20710. [MEDLINE: MEDLINE 94109871; : CN00098271, EMBASE 1994025969] Frei 1994 {published data only} Frei M, Kster L, Gardosch von Krosigk PP, Koch HF, Kppers H. Moxonidine and hydrochlorothiazide in combination: a synergistic antihypertensive effect.. Journal of cardiovascular pharmacology 1994;24 Suppl 1:S258. Frishman 1994 {published data only} Frishman WH, Bryzinski BS, Coulson LR, DeQuattro VL, Vlachakis ND, Mroczek WJ, Dukart G, Goldberg JD, Alemayehu D, Koury K. A multifactorial trial design to assess combination therapy in hypertension. Treatment with bisoprolol and hydrochlorothiazide. Archives of Internal Medicine 1994;154(13): 14618. [MEDLINE: MEDLINE 94288713; : CN00102520, EMBASE 1994210883] Frishman 1995 {published data only} Frishman WH, Burris JF, Mroczek WJ, Weir MR, Alemayehu D, Simon JS, Chen SY, Bryzinski BS. First-line therapy option with low-dose bisoprolol fumarate and low-dose hydrochlorothiazide in patients with stage I and stage II systemic hypertension. Journal of
21
Clinical Pharmacology 1995;35(2):1828. [MEDLINE: MEDLINE 95270751; : EMBASE 1995055004, CN00114120] Genthon 1994 {published data only} Genthon R, Andrieux LA, Arnaud J, Ballade M, Billou J, Bosredon A, Boutges B, Dubon T, Goigoux B, Lalanne G, Laurentjoye F, Lion A, Lorans P, Maurat X, Moulinet P, Ricard L, Roche A, Sdrigotti D, Rangoonwala B, et al.Study of the efcacy and safety of the combination ramipril 2.5 mg plus hydrochlorothiazide 12.5 mg in patients with mild-to-moderate hypertension. INT J CLIN PHARMACOL RES 1994;14(1):19. [MEDLINE: MEDLINE7927956; : EMBASE 1994205899; CN00170369] Hart 1991 {published data only} Hart W. Lisinopril-hydrochlorothiazide combination compared with the monocomponents in elderly hypertensive patients. Journal of Human Hypertension 1991;5 Suppl 2:859. [MEDLINE: MEDLINE1665184; : CN00081965; EMBASE1992032801] Homuth 1993 {published data only} Homuth V, Faulhaber HD, Loose U, Lofer K, Luft FC. Usefulness of piretanide plus ramipril for systemic hypertension: a multicenter trial. American Journal of Cardiology 1993;72(9): 66671. [MEDLINE: MEDLINE8249842; : CN00097410; EMBASE1993273569] Kayanakis 1987 {published data only} Kayanakis JG, Baulac L. Comparative study of once-daily administration of captopril 50 mg, hydrochlorothiazide 25 mg and their combination in mild to moderate hypertension. British Journal of Clinical Pharmacology 1987;23 Suppl 1:89S92S. [MEDLINE: MEDLINE 87213979; : CN00048048] Kellaway 1993 {published data only} Kellaway GS. A comparison of the efcacy of cilazapril versus cilazapril plus hydrochlorothiazide in patients with mild to moderate essential hypertension. Inhibace General Practice Study Group.. European journal of clinical pharmacology 1993;44(4): 3779. Kochar 1999 {published data only} Kochar M, Guthrie R, Triscari J, Kassler-Taub K, Reeves RA. Matrix study of irbesartan with hydrochlorothiazide in mild-tomoderate hypertension. American Journal of Hypertension 1999;12 (8 Pt 1):797805. [MEDLINE: MEDLINE 99408350; : CN00167270] Lacourciere 1994 {published data only} Lacourcire Y, Arnott W. Placebo-controlled comparison of the effects of nebivolol and low-dose hydrochlorothiazide as monotherapies and in combination on blood pressure and lipid prole in hypertensive patients.. Journal of human hypertension 1994 Apr;8(4):2838. Lacourciere 2005 {published data only} Lacourcire Y, Poirier L, Hebert D, Assouline L, Stolt P, Rehel B, Khder Y. Antihypertensive efcacy and tolerability of two xeddose combinations of valsartan and hydrochlorothiazide compared with valsartan monotherapy in patients with stage 2 or 3 systolic hypertension: an 8-week, randomized, double-blind, parallel-group trial.. Clinical therapeutics 2005;27(7):101321.
Lenz 1994 {published data only} Lenz T, Schulte KL, Wagner B, Lilienthal J, Gotzen R. Quinapril, hydrochlorothiazide, and combination in patients with moderate to severe hypertension.. European Heart Journal 1994;15(7):9406. [: CN00104892; EMBASE 1994237184] Li 2003 {published data only} Li Y, Liu G, Jiang B, Gao R, Chen L, Su L, Li J. A comparison of initial treatment with losartan/HCTZ versus losartan monotherapy in chinese patients with mild to moderate essential hypertension.. International Journal of Clinical Practice 2003;58(7):6737. [: CN00459020] MacKay 1996 {published data only} MacKay JH, Arcuri KE, Goldberg AI, Snapinn SM, Sweet CS. Losartan and low-dose hydrochlorothiazide in patients with essential hypertension. A double-blind, placebo-controlled trial of concomitant administration compared with individual components.. Archives of Internal Medicine 1996;156(3):27885. [: EMBASE 1996062529; CN00123252] Schoenberger JA, Bauer J, Barden LP, Brown R, Byyny R, Cohen J, Davis A, Davis P, Popli S, Guthrie G, Herman T, Hotltzman J, Kelly D, Marbury TC, Lewis G, Littlejohn T, MacKay JH, Mulrow P, O CD, et al.Losartan with hydrochlorothiazide in the treatment of hypertension.. Journal of Hypertension, Supplement. 1995;13(1): S43S47. [: EMBASE 1995266978] Manning 1996 {published data only} Manning G, Joy A, Mathias CJ, McDonald CJ, Millar-Craig MW. Double-blind, parallel, comparative multicentre study of a new combination of diltiazem and hydrochlorothiazide with individual components in patients with mild or moderate hypertension. Journal of human hypertension 1996 Jul;10(7):4438. McGill 2001 {published data only} McGill JB. Angiotensin II receptor antagonist plus a thiazide diuretic is more efcacious for treating hypertension than either drug alone.. Blood Pressure Monitoring 2001;6(SUPPL. 1):S3S13. McGill JB, Reilly PA. Telmisartan plus hydrochlorothiazide versus telmisartan or hydrochlorothiazide monotherapy in patients with mild to moderate hypertension: a multicenter, randomized, double-blind, placebo-controlled, parallel-group trial. Clinical Therapeutics 2001;23(6):83350. [MEDLINE: MEDLINE 21333338; : CN00373202] Merrill 1987 {published data only} Merrill DD, Byyny RL, Carr A, Dauer AD, Kazilionis JE, Lester FM, Miller K, Gibson TP, Whipple J. Lisinopril/HCTZ in essential hypertension [abstract]. Clinical Pharmacology and Therapeutics 1987; Vol. 41, issue 2:227. Mersey 1993 {published data only} Mersey J, DHemecourt P, Blaze K. Once-daily xed combination of captopril and hydrochlorothiazide as rst line therapy for mild to moderate hypertension. CURR. THER. RES. CLIN. EXP. 1993;53 (5):502512. [MEDLINE: EMBASE 1993242231; : CN00182021] Meyer 1994 {published data only} Meyer BH, Pauly NC. Double-blind comparison of the efcacy and safety of trandolapril 2 mg and hydrochlorothiazide 25 mg in patients with mild-to-moderate essential hypertension. J CARDIOVASC PHARMACOL 1994;23(SUPPL. 4):S77S80.
22
[MEDLINE: MEDLINE7527107; : EMBASE 1994264217; CN00172078] Neutel 2008 {published data only} Neutel J.M, Franklin S.S, Lapuerta P, Bhaumik A, Ptaszynska A. A comparison of the efcacy and safety of irbesartan/HCTZ combination therapy with irbesartan and HCTZ monotherapy in the treatment of moderate hypertension.. Journal of Human Hypertension 2008;22(4):266274. Oparil 1980 {published data only} Oparil S. Multiclinic double-blind evaluation of timolol combined with hydrochlorothiazide in essential hypertension. Current Therapeutic Research, Clinical & Experimental 1980;27(4): 527537. [: CN00333696; EMBASE 1980132951] Papademetriou 2000 {published data only} Papademetriou V, Reif M, Henry D, et al.Combination Therapy with Candesartan Cilexetil and Hydrochlorothiazide in Patients with Systemic Hypertension. Journal of Clinical Hypertension 2000; 2:3728. Papademetriou 2006 {published data only} Papademetriou V, Hainer JW, Sugg J, Munzer D, ATTACH Study Group. Factorial antihypertensive study of an extendedrelease metoprolol and hydrochlorothiazide combination.. American journal of hypertension : journal of the American Society of Hypertension 2006 Dec;19(12):121725. Parati 2006 {published data only} Parati G, Omboni S, Castiglioni G, Cereda U, Corradi L, Fogari R, Garavelli G, Malacco E, Pisani A, Villa G. Antihypertensive efcacy of zofenopril and hydrochlorothiazide combination on ambulatory blood pressure. Blood Pressure 2006;15(Suppl 1):717. Philipp 1997 {published data only} Philipp T, Letzel H, Arens HJ. Dose-nding study of candesartan cilexetil plus hydrochlorothiazide in patients with mild to moderate hypertension.. Journal of human hypertension 1997 Sep;11 Suppl 2: S678. Pool 1987 {published data only} Pool JL, Gennari J, Goldstein R, Kochar MS, Lewin AJ, Maxwell MH, McChesney JA, Mehta J, Nash DT, Nelson EB, et al.Controlled multicenter study of the antihypertensive effects of lisinopril, hydrochlorothiazide, and lisinopril plus hydrochlorothiazide in the treatment of 394 patients with mild to moderate essential hypertension. Journal of Cardiovascular Pharmacology 1987;9 Suppl 3:S3642. [MEDLINE: MEDLINE2442550; : CN00254084; EMBASE1987181938] Pool 1993 {published data only} Pool PE, Applegate WB, Woehler T, Sandall P, Cady WJ. A randomized, controlled trial comparing diltiazem, hydrochlorothiazide, and their combination in the therapy of essential hypertension. Pharmacotherapy 1993;13(5):48793. [MEDLINE: MEDLINE 94068188; : CN00097378] Pool 1997 {published data only} Pool JL, Cushman WC, Saini RK, Nwachuku CE, Battikha JP. Use of the factorial design and quadratic response surface models to evaluate the fosinopril and hydrochlorothiazide combination therapy in hypertension. American Journal of Hypertension 1997;10 (1):11723. [MEDLINE: MEDLINE 97160960; : CN00136040]
Pool 2007a {published data only} Pool JL, Schmieder RE, Azizi M, Aldigier JC, Januszewicz A, Zidek W, Chiang Y, Satlin A. Aliskiren, an orally effective renin inhibitor, provides antihypertensive efcacy alone and in combination with valsartan.. American journal of hypertension : journal of the American Society of Hypertension 2007 Jan;20(1): 1120. Pool 2007b {published data only} Pool JL, Glazer R, Weinberger M, Alvarado R, Huang J, Graff A. Comparison of valsartan/hydrochlorothiazide combination therapy at doses up to 320/25 mg versus monotherapy: a double-blind, placebo-controlled study followed by long-term combination therapy in hypertensive adults.. Clinical therapeutics 2007 Jan;29 (1):6173. Prisant 2000 {published data only} Prisant LM. Ambulatory Blood Pressure Proles in Patients Treated with Once-Daily Diltiazem Extended-Release or Indapamide Alone or in Combination. American Journal of Therapeutics 2000;7:177184. Romero 1995 {published data only} Romero R, Castellote E, Ocon J, Wagner B. Controlled multicenter study with quinapril, hydrochlorothiazide, and combination in patients with moderate to severe hypertension.. Journal of Cardiovascular Pharmacology 1995;26(1):1148. [MEDLINE: MEDLINE 96057138; : CN00120025] Rosenthal 1990 {published data only} Rosenthal T, Grossman E, Rathaus M, Bernheim J, Zevin D, Levi J, Weinstein J, Rogel S. Treatment of hypertension by enalapril and hydrochlorothiazide separately and together: a multicenter study. Israel Journal of Medical Sciences 1990;26(2):636. [MEDLINE: MEDLINE2180849; : CN00066622; EMBASE1990104829] Safar 1973 {published data only} Safar M, Maiz HB, Weiss Y, Lagrue G, Milliez P. [Antihypertensive action of a beta-blocker: value of a controlled therapeutic trial] [Action antihypertensive dun btabloqueur: intrt dun essai thrapeutique contrl.]. Therapeutique (La Semaine des hpitaux) 1973 SepOct;49(7):45962. Safar M, Weiss Y, Sobel A, Lagrue G, Milliez P. [Antihypertensive properties of a beta blocking agent, pindolol] [Action antihypertensive dun btabloqueur, le pindolol.]. La Nouvelle presse mdicale 1973 Nov;2(40):26858. Safar 1994 {published data only} Safar M, Zanchetti A, Sever PS, et al.Perindopril and Indapamide as a Combination in the Treatment of Mild to Moderate Hypertension - A Double-Blind Randomized Placebo-Controlled European Multicenter Study [POSTER]. AJH 1994; Vol. 7, issue 4:43A. Saruta 2007 {published data only} Saruta T, Ogihara T, Matsuoka H, et al.Antihypertensive Efcacy and Safety of Fixied-Dose Combination Therapy with Losartan Plus Hydrochlorothiazide in Japanese Patients with Essential Hypertension. Hypertens Res 2007;30:729739. Saul 1995 {published data only} Saul P. Efcacy and tolerability of a once-daily low-dose lisinopril /hydrochlorothiazide preparation: Comparison with its monocomponents. Drug Development Research. 1995;34(suppl. 2): 813. [: EMBASE 1995095665; CN00173176]
23
Schoenberger 1986 {published data only} Schoenberger JA, Wilson DJ. Once-daily treatment of essential hypertension with captopril.. Journal of Clinical Hypertension 1986; 2(4):37987. [: CN00046328] Thijs 1995 {published data only} Thijs L, Celis H, Kiowski W, Lofer K, Middeke M, Schulz W, Staessen J, Amery A. Double-blind comparison of antihypertensive treatment with ramipril and piretanide, given alone or in combination. Journal of Cardiovascular Pharmacology 1995;26(1): 338. [MEDLINE: MEDLINE7564362; : EMBASE 1995198201; CN00120021] Vaicaitis 1980 {published data only} Vaicaitis JS. VVaicaitis JS. Evaluation of a beta-blocker, timolol maleate, combined with hydrochlorothiazide in essential hypertension.. Current Therapeutic Research, Clinical & Experimental 1980;27(3):365373. [: EMBASE1980107978 EMBASE1980107978] Villamil 2007 {published data only} Villamil A, Chrysant Sg, Calhoun D, et al.Renin Inhibition with Aliskiren Provides Additive Antihypertensive Efcacy When Used in Combination with Hydrochlorothiazide. Journal of Hypertension 2007;25:217226. von Manteuffel 1995 {published data only} von Manteuffel GE, Rakette S, Woll EM, Reinfrank J, Schiemann J. Effectiveness and tolerance of combined verapamil retard and hydrochlorothiazide. Results of a double-blind, randomized study [Wirksamkeit und Vertraglichkeit der Kombination Verapamil retard und Hydrochlorothiazid. Ergebnisse einer doppelblinden, randomisierten Studie.]. Fortschritte der Medizin 1995;113(26): 3748. Weinberger 1982 {published data only} Weinberger MH. Comparison of captopril and hydrochlorothiazide alone and in combination in mild to moderate essential hypertension.. British journal of clinical pharmacology 1982;14 Suppl 2:127S131S. Weir 1992 {published data only} Weir MR, Weber MA, Punzi HA, Serfer HM, Rosenblatt S, Cady WJ. A dose escalation trial comparing the combination of diltiazem SR and hydrochlorothiazide with the monotherapies in patients with essential hypertension. Journal of Human Hypertension 1992;6 (2):1338. [MEDLINE: MEDLINE1597846; : CN00084591; EMBASE1992151099] Yodfat 1994 {published data only} Yodfat Y, Zimilchman R. Dose-nding and dose justication of once-daily cilazapril in combination with hydrochlorothiazide in non-obese patients with mild-to-moderate essential hypertension. J. DRUG DEV. 1993;6(3):117121. [: EMBASE 1994092035; CN00179498]
Agrawal 1987 {published data only} Agrawal RL. Double-blind comparison of Inderal LA (160mg), Half-Inderal LA (80 mg), and Half-Inderal LA plus bendrouazide (2.5mg) in the treatment of elderly hypertensive patients. The British Journal of Clinical Practice 1987;41:916920. Anderton 1988 {published data only} Anderton JL, Vallance BD, Stanley NN, Crowe PF, Mittra B, Perks WH. Atenolol and Sustained Release Nifedipine Alone and in Combination in Hypertension. Drugs 1988;35 (suppl. 4):2226. Bainbridge 1993 {published data only} Bainbridge AD, Macfadyen RJ, Stark S, Lees KR, Reid JL. The antihypertensive efcacy and tolerability of a low dose combination of ramipril and felodipine ER in mild to moderate essential hypertension.. British Journal of Clinical Pharmacology 1993;36(4): 32330. Bateman 1979 {published data only} Bateman DN, Dean CR, Mucklow JC. Atenolol and Chlorthalidone in Combination for Hypertension. Br. J. Clin. Pharmac 1979;7:357363. Bauer 1984 {published data only} Bauer JH, Gaddy P. Effects of Enalapril Alone, and in Combination with Hydrochlorothiazide, on Renin-Angiotensin-Aldosterone, Renal Function, Salt and Water Excretion, and Body Fluid Composition. American Journal of Kidney Diseases 1985;6(4): 22232. Comparative studies: enalapril versus hydrochlorothiazide as rststep therapy for the treatment of primary hypertension. Bauer JH, Jones LB. American Journal of Kidney Diseases 1984;4(1):5562. Bertrand 1982 {published data only} Bertrand Ed, Dienot B, Ravinet L. Study of treatment of arterial hypertension with oxprenolol (compared with placebo) and chlortalidone [Etude du traitement de lhypertension artrielle par Oxprnolol (compar un placebo) et par Chlortalidone]. Mdecine dAfrique Noire 1982;29(4):271274. Boike1982 {published data only} Boike SC, Durley Y, Cubberley RB. Atenolol and Chlorthalidone Administered Alone and in Combination for Essential Hypertension. Clinical Pharmacy 1982;1:44953. Bolzano 1984 {published data only} Bolzano K, Krempler F, Sandhofer F. [A combination of methyldopa, hydrochlorothiazide and amiloride in the treatment of essential hypertension] [Eine Kombination von Methyldopa, Hydrochlorothiazid und Amilorid bei der Behandlung der essentiellen Hypertonie.]. Acta medica Austriaca 1984;11(2):4953. Cajochen 1984 {published data only} Cajochen C, Krauchi K, Von AMA, Mori D, Graw P, Wirz-Justice A. A multicenter comparison of the antihypertensive effects of atenolol and chlorthalidone given alone and in combination. CURR THER RES, CLIN EXP 1984;35(1):3139. Canter 1994 {published data only} Canter D, Frank GJ, Knapp LE, Phelps M, Quade M, Texter M. Quinapril and hydrochlorothiazide combination for control of hypertension:assessment by factorial design. Quinapril InvestigatorGroup [see comments]. Journal of Human Hypertension 1994;8(3):155-62.
24
References to studies excluded from this review

Agrawal 1979 {published data only} The treatment of hypertension with propranolol and bendrouazide. Agrawal RL, Alliott RJ, George M, Gomez G, Trafford JA, Jequier PW, Lishman JD, Turner JR, Baber NS, Dawes PM. Journal of the Royal College of General Practitioners 1979;29 (207):6026.
Cappuccio 1993 {published data only} Cappuccio FP, Markandu ND, Singer DRJ, MacGregor GA. Amlodipine and Lisinopril in Combination for the Treatment of Essential Hypertension: Efcacy and Predictors of Response. Journal of Hypertension 1993;11:83947. Chalmers 1982 {published data only} Chalmers JP, Wing LMH, Grygiel JJ, West MJ, Graham JR, Bune AJ. Effects of Once Daily Indapamide and Pindolol on Blood Pressure, Plasma Aldosterone Concentration and Plasma Renin Activity in a General Practice Setting. Eur J Clin Pharmacol 1982; 22:1916. Chalmers 1986 {published data only} Chalmers JP, Wing LMH, West MJ, Bune AJC, Elliott JM. Effects of Enalpril and Hydrochlorothiazide on Blood Pressure, Reninangiotensin System, and Atrial Natriuretic Factor in Essential Hypertension: A Double-blind Factorial Cross-over Study. Aust NZ J Med 1986;16:475480. Chrysant 1992 {published data only} Chrysant SG, Chappel C, Farnham DJ, Levin B, Lueg M, McCluskey D, Steiner C. Antihypertensive and metabolic effects of single and combined atenolol regimens. Journal of Clinical Pharmacology 1992;32(1):615. [MEDLINE: MEDLINE 1740538; : CN00081715; EMBASE1992169096] Crowe 1987 {published data only} Crowe PF, Finnegan OC, Rusell CJMcL, Varma MPS. A Doubleblind Study of INderex in the Treatment of Essential Hypertension. The British Journal of Clinical Practice 1987;41(10):967970. Dahlof 1993 {published data only} Dahlof B, Jonsson L, Borgholst O, Ekblad G, Engstrand C, Grundestam I, Lindh A. Improved antihypertensive efcacy of the felodipine-metoprolol extended-release tablet compared with each drug alone.. Blood Pressure Supplement 1993;1:3745. Dargie 1986 {published data only} Dargie H, Cleland J, Findlay I, et al.Combination of Verapamil and Beta Blockers in Systemic Hypertension. Am J Cardiol 1986;57: 80D82D. Dargie HJ. Verapamil and Beta-Blockers in Hypertension: A Benecial Drug Interaction?. British Journal of Clinical Practice 1985;39(6 SUPPL. 42):4447. De Divitiis 1981 {published data only} De Divitiis O, Petitto M, Di Somma S, et al.Atenolol, chlorthalidone, and reserpine in mild-moderate hypertension: Double-blind comparison. Drugs Under Experimental & Clinical Research 1981;7(6):773779. De Divitiis 1983 {published data only} De Divitiis O, Di Somma S, Petitto M, Fazio S, Ligouri V. Indapamide and Atenolol in the Treatment of Hypertension: Double-blind Comparative and Combination Study. Current Medical Research Opinion 1983;8(7):493500. Durel 1992 {published data only} Durel LA, Hayashi PJ, Weidler DJ, Schneiderman N. Effectiveness of Antihypertensive Medications in Ofce and Ambulatory Settings: A Placebo-Controlled Comparison of Atenolol, Metroprolol, Chlorthalidone, Verapamil, and an AtenololChlorthalidone Combination. J Clin Pharmacol 1992;32:564570.
Erwteman 1984 {published data only} Erwteman TM, Nagelkerke N, Lubsen J, Koster M, Dunning AJ. Beta blockade, diuretics, and salt restriction for the management of mild hypertension: a randomised double blind trial. British medical journal (Clinical research ed.) 1984 Aug;289(6442):4069. Fernandez 1980 {published data only} Fernandez PG, Zachariah PK, Bryant DG, Missan SS. Anithypertensive efcacy of alpah-methyldopa, chlorothiazide and Supres-150 (alpha-methyldopa-chlorothiazide). CMA Journal 1980;123:284287. Forette 1979 {published data only} Forette F, Henry JF, Hervy MP, Forette B, Berthaux P. The treatment of hypertension in the elderly using a beta-blocker: acebutolol [Traitement de lhypertension artrielle du sujet g par un btabloquant: lacbutolol.]. La Nouvelle presse mdicale 1979; 8(36):28814. Frishman 1987 {published data only} Frishman WH, Goldberger J, Sherman D. Enalapril, hydrochlorothiazide, and combination therapy in patients with moderate hypertension.. Journal of Clinical Hypertension 1987;3(4): 5207. Hart 1985 {published data only} Hart HC, Van dLDL, Lustermans FAT, et al.Double-blind comparison of dytenzide, atenolol and the combination of dytenzide and atenolol in patients with essential hypertension.. CURR-THER-RES,-CLIN-EXP 1985;38(5):702709. [: CN00175975; EMBASE 1986040698] Hunter 1999 {published data only} Hunter SJ, Wiggam MI, Ennis CN, Whitehead HM, Sheridan B, Atkinson AB, Bell PM. Comparison of effects of captopril used either alone or in combination with a thiazide diuretic on insulin action in hypertensive Type 2 diabetic patients: a double-blind crossover study.[see comment].. Diabetic Medicine 1999;16(6): 4827. Jaattela 1979 {published data only} Jaattela A. The Fixed Combination of Propanolol and Bendrouazide in the Treatment of Hypertension. Annals of Clinical Research 1979;11:8082. Jackson 1986 {published data only} Jackson G, Rowland M, Adam G, MacFarlane E, Jackson PG. Placebo Controlled Double-blind Randomised Cross-over Trial of Atenolol, Hydrochlorothiazide and Amiloride, and the Combination (Kalten) in Patients Over 60 Years of Age. The British Journal of Clinical Practice 1986;40(6):230234. Khalil 1982 {published data only} Khalil SI, El Zein O, El Mahadi Bella M. A double-blind, crossover study of acebutolol and hydrochlorothiazide/amiloride diuretic in Sudanese patients with essential hypertension.. Current medical research and opinion 1982;1:3943. Kieso 1983 {published data only} Kieso HA, Gould BA, Mann S, Hornung RS, Altman DG, Raftery EB. Effect on Intra-aterial Blood Pressure of Slow Release Metoprolol Combined with Placebo or Chlorthalidone. British Medical Journal 1983;287:717720.
25
Koh 2007 {published data only} Koh KK, Quon MJ, Lee Y, et al.Additive Benecial Cardiovascular and Metabolic Effects of Combination Therapy with Ramipril and Candesartan in Hypertensive Patients. European Heart Journal 2007;28:14407. Kubik 1979 {published data only} Kubik M, Kendall M, Ebbutt A, John V. Metoprolol with and without chlorthalidone in hypertension. Clinical pharmacology and therapeutics 1979;25(1):2532. Lang 1991 {published data only} Lang H. The results of a large multicentre study comparing lowdose lisinopril-hydrochlorothiazide with the monocomponents.. Journal of Human Hypertension 1991;5 Suppl 2:736. Lechi 1982 {published data only} Lechi A, Pomari S, Berto R, et al.Clinical Evaluation of Labetolol Alone and Combined with Chlorthalidone in Essential Hypertension: A Double-blind Multicentre Controlled Study. European Journal of Clinical Pharmacology 1982;22:28993. Leonetti 1986 {published data only} Leonetti G, Pasotti C, Capra A. Low-dose atenolol-chlorthalidone combination for treatment of mild hypertension.. International journal of clinical pharmacology, therapy, and toxicology 1986 Jan;24 (1):437. Magee 1986 {published data only} Magee PFA, Freis ED. Is Low-Dose Hydrochlorothiazide Effective?. Hypertension 1986;8 (suppl II):135139. Mehta 1988 {published data only} Mehta J, Lopez LM, Thorman AD. Lisinopril Versus Lisinopril plus Phydrochlorothiazide in Essential Hypertension. Am J Cardiol 1988;61:8036. Middlemost 1994 {published data only} Middlemost SJ, Tager R, Davis J, Sareli P. Effectiveness of enalapril in combination with low-dose hydrochlorothiazide versus enalapril alone for mild to moderate systemic hypertension in black patients.. American Journal of Cardiology 1994;73(15):10927. Moncloa 1980 {published data only} Moncloa F, Hwang IK, Muccilli AC. Multiclinic evaluation of the antihypertensive effect of a methyldopa, hydrochlorothiazide, and amiloride combination.. Clinical Therapeutics 1980;3(3):16875. [MEDLINE: MEDLINE7006811; : CN00024331; EMBASE1981086995] Morgan 1992 {published data only} Morgan TO, Anderson A. Hemodynamic Comparisons of Enalapril and Felodipine and Their Combination. Kidney International 1992;41 (Suppl 36):S7881. Morgan TO, Anderson A, Jones E. Comparison and Interaction of Low Dose Felodipine and Enalapril in the Treatmetn of Essential Hypertension in Elderly Subjects. American Journal of Hypertension 1992;5:23843. Morgan 2002 {published data only} Morgan T, Anderson A. A comparison of candesartan, felodipine, and their combination in the treatment of elderly patients with systolic hypertension.. American Journal of Hypertension 2002;15 (6):5449.
Muiesan 1976 {published data only} Muiesan G, Magnani B, Agabiti-Rosei E, Alicandri C, Ambrosioni E, Miele N. Evaluation of the effect of timolol alone and in combination with hydrochlorothiazide and amiloride in the treatment of mild to moderate arterial hypertension: a doubleblind, controlled study. Clinical Science & Molecular Medicine Supplement 1976;3:529s531s. [MEDLINE: MEDLINE799563; : CN00015916] Neutel 2006 {published data only} Neutel JM, Smith DHG, Weber MA, et al.Efcacy of Combination Therapy With Amlodipine Besylate/Benazepril Hydrochloride for Lowering Systolic Blood Pressure in Stage 2 Hypertension. The American Journal of Geriatric Cardiology 2006;15(3):14250. Petrie 1975 {published data only} Petrie JC, Galloway DB, Webster J, Simpson WT, Lewis JA. Atenolol and Bendrouazide in Hypertension. British Medical Journal 1975;4:1335. Ricciardelli 1985 {published data only} Hemodynamic effects of the antihypertensive treatment with atenolol plus chlortalidone in xed combination: A double-blind crossover comparison with atenolol. Ricciardelli B, Cuocolo A, De LN, et al. Current Therapeutic Resesarch 1985;37(5):90111. Rosenthal 1989 {published data only} Rosenthal J, Letzel H, Blumner E, Volger K-D. Antihypertensive Therapy - Dose-Effect Relationship Between a Combination of Triamterene/Hydrochlorothiazide and Verapamil. Munch. med. Wschr 1989;131(16):31922. Salako 1990 {published data only} Salako LA, Falase AO, Aderounmu AF, Walker O. Assessment of a Fixed-dosage Combination of Atenolol and Chlorthalidone (Tenoretic) in Hypertensive Nigerians. Afr. J. Med. Sci 1990;19: 5761. Salvetti 1987 {published data only} Salvetti A, Innocenti PF, Iardella M, Pambianco F, Saba GC, Rossetti M, Botta GF. Captopril and nifedipine interactions in the treatment of essential hypertensives: a crossover study.. Journal of Hypertension - Supplement 1987;5(4):S13942. Salvetti 1989 {published data only} Salvetti A, Magagna A, Innocenti, et al.Chlorthalidone does not increase the hypotensive effect of nifedipine in essential hypertensives: a crossover multicentre study. Journal of hypertension 1989;7 (suppl 6):S2501. Salvetti A, Magagna A, Innocenti, et al.The Combination of Chlorthalidone with Nifedipine Does Not Exert an Additive Antiypertensive Effect in Essential Hypertensives: A Crossover Multicenter Study. Journal of Cardiovascular Pharmacology 1991; 17:332335. Scholze 1993 {published data only} Scholze J, Breitstadt A, Cairns V, Bauer B, Bender N, Priestley C, Moreadith C, Phillips J, Vander Elst E, Koch G. Short report: ramipril and hydrochlorothiazide combination therapy in hypertension: a clinical trial of factorial design. The East Germany Collaborative Trial Group. Journal of hypertension 1993;11(2): 21721.
26
Singer 1987 {published data only} Singer DRJ, Markandu ND, Shore AC, MacGregor GA. Nifeidipine and Acebutolol in COmbination for the Treatment of Moderate to Severe Essential Hypertension. Journal of Human Hypertension 1987;1:317. Tonkin 1990 {published data only} Tonkin AL, Wing LMH, Russell AE, et al.Diltiazem and atenolol in essential hypertension: additivity of effect on blood pressure and cardiac conduction with combination therapy. Journal of Hypertension 1990;8:10151019. Topouchian 1999 {published data only} Topouchian J, Asmar R, Sayegh F, Rudnicki A, Benetos A, Bacri AM, Safar ME. Changes in arterial structure and function under trandolapril-verapamil combination in hypertension.. Stroke 1999; 5:105664. VACSGAA 1977 {published data only} Veterans Administration Cooperative Study Group on Antihypertensive Aggents. Propanolol in the Treatment of Essential Hypertension. JAMA 1977;237(21):230310. Van Staden 1983 {published data only} Van Staden DA, Van Gelder LW, Van Der Lingen RE. Trial of Bendrouazide Compared to Propranolol plus Bendrouazide (Inderetic) for the Treatment of Essential Hypertension in Blacks. Current Therapeutic Research 1983;34(4):620626. Viskoper 1997 {published data only} Viskoper RJ, Compagnone D, Dies R, Zilles P. Verapamil and trandolapril alone and in xed combination on 24-hour ambulatory blood pressure proles of patients with moderate essential hypertension. Current Therapeutic Research - Clinical and Experimental 1997;58(6):34351. Weinberger 1983 {published data only} Weinberger MH. Inuence of an angiotensin converting-enzyme inhibitor on diuretic-induced metabolic effects in hypertension.. Hypertension 1983;5(5 Pt 2):III1328. Wing 1988 {published data only} Wing LMH, Chalmers JP, West MJ, et al.Enalapril and Atenolol in Essential Hypertension: Attenuation of Hypotensive Effects in Combination. Clin. and Exper. Hyper. -Theory and Practice 1988; 10(1):119133. Wing 1994 {published data only} Wing LM, Russell AE, Tonkin AL, Bune AJ, West MJ, Chalmers JP. Felodipine, metoprolol and their combination compared with placebo in isolated systolic hypertension in the elderly.. Blood Pressure 1994;3(1-2):829.
randomized controlled trials. Journal of Clinical Epidemiology 1997; 50(6):683691. Chan 1997 Chan P, Lin CN, Tomlinson B, Lin TH, Lee YS. Additive effects of diltiazem and lisinopril in the treatment of elderly patients with mild-to-moderate hypertension. American Journal of Hypertension 1997;10(7 Pt 1):7439. Chen 2008a Chen JMH, Heran BS, Perez MI, Wright JM. Blood pressure lowering efcacy of drugs inhibiting the renin-angiotensin system as second-line therapy for primary hypertension. Cochrane Database of Systematic Reviews 2008, Issue 2. [DOI: 10.1002/ 14651858.CD007188] Chen 2008b Chen JMH, Heran BS, Perez MI, Wright JM. Blood pressure lowering efcacy of beta-blockers as second-line therapy for primary hypertension. Cochrane Database of Systematic Reviews 2008, Issue 2. [DOI: DOI: 10.1002/14651858.CD007185] Chen 2008c Chen JMH, Heran BS, Perez MI, Wright JM. Blood pressure lowering efcacy of calcium-channel blockers as second-line therapy for primary hypertension. Cochrane Database of Systematic Reviews 2008, Issue 2. [DOI: 10.1002/14651858.CD007186] Farsang 2001 Farsang, C. Kawecka-Jaszcz K. Langan J. Maritz F. Zannad F. Antihypertensive effects and tolerability of candesartan cilexetil alone and in combination with amlodipine. Clinical Drug Investigation 2001;21(1):1723. Franklin 1999 Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is Pulse Pressure Useful in Predicting Risk for Coronary Heart Disease?: The Framingham Heart Study. Circulation 1999;100:354360. Haider 2003 Haider AW, Larson MG, FranklinSS, Levy D, FraminghamHeart Study. Systolic bloodpressure, diastolic blood pressure, and pulse pressure as predictors of risk for congestiveheart failure in the Framingham Heart Study. Ann Intern Med 2003;138(1):1016. Heran 2009a Heran BS, Wong MMY, Heran IK, Wright JM. Blood pressure lowering efcacy of angiotensin converting enzyme (ACE) inhibitors for primary hypertension. Cochrane Database of Systematic Reviews 2008, Issue 4. [DOI: 10.1002/ 14651858.CD003823.pub2] Heran 2009b Heran BS, Wong MMY, Heran IK, Wright JM. Blood pressure lowering efcacy of angiotensin receptor blockers for primary hypertension. Cochrane Database of Systematic Reviews 20098, Issue 4. [DOI: 10.1002/14651858.CD003822.pub2] Higgins 2008 Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.1 [updated September 2008]. The Cochrane Collaboration 2008; Vol. Available from www.cochranehandbook.org.
27
Additional references
Azizi 2000 Azizi M, Linhart A, Alexander J, Goldberg A, Menten J, Sweet C, Menard J. Pilot study of combined blockade of the reninangiotensin system in essential hypertensive patients.. Journal of Hypertension 2000;18(8):113947. Bucher 1997 Bucher HC, Guyatt GH, Grifth LE, Walker SD. The results of direct and indirect treatment comparisons in meta-analysis of
Kuschnir 2004 Kuschnir E, Bendersky M, Resk J, Panart MS, Guzman L, Plotquin Y, Grassi G, Mancia G, Wagener G. Effects of the combination of low-dose Nifedipine GITS 20mg and Losartan 50mg in patients with mild to moderate hypertension. Journal of cardiovascular pharmacology 2004;43:300305. Lessem 1989 Lessem JN, Barone EJ, Berl T, Detwiler J, Lewin AT, Lubash GD, Margolis R, McGowan RL, Ram CV, Vlachakis N, et al.Nicardipine and propranolol in the treatment of essential hypertension.. American Journal of Hypertension 1989;2(3 Pt 1):14653. Levine 1995 Levine JH, Ferdinand KC, Cargo P, Laine H, Lefkowitz M. Additive effects of verapamil and enalapril in the treatment of mild to moderate hypertension. American Journal of Hypertension 1995;8 (5 Pt 1):4949. Materson 1993 Materson BJ, Reda DJ, Cushman WC, et al.Single-drug therapy for hypertension in men: comparison of six antihypertensive agents with placebo. N Engl J Med 1993;328:91421. Messerli 1998 Messerli F, Frishman WH, Elliott WJ. Effects of verapamil and trandolapril in the treatment of hypertension. American Journal of Hypertension 1998;11(3 Pt 1):3227. Musini 2002 Musini VM. A systematic review of the blood pressure lowering efcacy of thiazide diuretics in the treatment of adult patients with primary hypertension. MSc Thesis. The University of British Columbia. 2000. Musini 2009 Musini VM, Wright JM. Factors Affecting Blood Pressure Variability: Lessons Learned from Two Systematic Reviews of Randomized ControlledTrials. PLoS ONE 2009;4(5):e5673. [DOI: 10.1371/journal.pone.0005673] Pittrow 1997 Pittrow DB. Antlsperger A. Welzel D. Wambach G. Schardt W. Weidinger G. Evaluation of the efcacy and tolerability of a low-
dose combination of isradipine and spirapril in the rst-line treatment of mild to moderate essential hypertension. Cardiovascular Drugs & Therapy 1997;11(5):619627. Psaty 2003 Psaty BM, Lumley T, Furberg CD, Schellenbaum G, Pahor M, Alderman MH, Weiss NS. Health outcomes associated with various antihypertensive therapies used as rst-line agents: A network meta-analysis. JAMA 2003;289:25342544. Scholze 1998 Scholze J, Zilles P, Compagnone D. Verapamil SR and trandolapril combination therapy in hypertension - A clinical trial of factorial design. British Journal of Clinical Pharmacology 1998;45(5): 491495. Scholze 1999 Scholze J, Bauer B, Massaro J. Antihypertensive proles with ascending dose combinations of ramipril and felodipine ER. Clinical & Experimental Hypertension (New York) 1999;21(8): 144762. Song 2003 Song F, Altman DG, Glenny A-M, Deeks JJ. Validity of indirect comparison for estimating efcacy of competing interventions: empirical evidence from published meta-analyses. BMJ 2003;325: 472476. Wong 2007 Wong MMY. A systematic review of the blood pressure lowering efcacy of calcium channel blockers for primary hypertension. MSc Thesis. The Univeristy of British Columbia. 2007. Wright 1999 Wright JM. Cheng-Han Lee, Chambers KC. Systematic review of antihypertensive therapies: does the evidence assist in choosing a rst-line drug?. CMAJ 1999;161(1):2532. [MEDLINE: 99349345] Wright 2009 Wright JM, Musini VM. First-line drugs for hypertension. Cochrane Database of Systematic Reviews 2009, Issue 3. [DOI: 10.1002/14651858.CD001841.pub2] Indicates the major publication for the study
28
CHARACTERISTICS OF STUDIES
Characteristics of included studies [ordered by study ID]

Asplund 1981 Methods 12 weeks double-blind, parallel study after 4 weeks run-in period with placebo. Randomized (methods not described) 75 patients with supine BP >160/95 mmHg at end of run-in period. Age range: 21-71 (mean: 48) yrs 3 treatment groups (BID): Monotherapy: Metoprolol 100mg (M100); HCTZ 12.5mg (H12.5) Combination: Fixed-ratio M100/H12.5 The dosage was 1 tablet morning and evening in all cases. Standing SBP and DBP and HR (timing of measurement not reported), WDAE Funding: Not reported
Participants
Interventions
Outcomes Notes Risk of bias Item Adequate sequence generation?
Authors judgement Unclear
Description patients were then allocated at random to treatment not described the tablets were identical in appearance and the study was designed as a doubleblind comparison of three parallel groups 75 patients were randomized. One of these patients failed to attend the check-ups and was excluded from the study and another was excluded owing to side effects in the form of leg fatigue and gastric discomfort during treatment with metoprolol Efcacy analysis on 73 patients who completed the study. Comment: low risk of bias
Allocation concealment? Blinding? All outcomes
Unclear Yes
Incomplete outcome data addressed? All outcomes
Yes
Free of selective reporting?
Yes
29
Benz 1998 Methods 8-week, double-blind, placebo-controlled parallel study after washout period of at least 2 weeks and single-blind placebo treatment for 2-4 weeks; randomized (methods not described) 871 adult out-patients with sitting diastolic BP 95-115 mm Hg. Difference between enrolment and randomization <=10mm Hg. Age: >18 (range 22-26, mean 52); 58% men. 75% white, 14%black, 11%other. 9 treatment groups o.d.: -placebo monotherapy: valsartan 80 or 160mg, Hydrochlorothiazide 12.5 or 25mg -combination therapy: valsartan 80mg/ hydrochlorothiazide 12.5mg (V80/H12.5), V160/H12.5, V80/H25, V160/H25. Sitting trough SBP and DBP; HR (no quantitative data), WDAE Efcacy analyses: ITT (865 patients with post-randomization measurements) 867 patients included in safety analysis 792/871 patients completed the trial 41 WDAE ( the number of WDAE in each group is not reported) No statistically or clinically signicant differences between groups in sitting or standing pulse rate Funding: Novartis
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomised not described Study drugs were packaged in a double-dummy fashion to maintain blinding; Valsartan 80mg and 160mg were supplied as capsules of identical appearance...Placebo capsules matched the valsartan and HCTZ capsules in appearance the number of patients included in the primary efcacy analysis at end-point was 865 (six prematurely-discontinued patients had no post-randomisation measurements). Comment: low risk of bias (<1% patients missing)
Yes
Yes
30
Bermudez 1982 Methods 8 week comparative, double-blind, multi-center study after a 2-week placebo run-in period. Data extracted at week 4. Dose is doubled after week 4 if DBP > 90 mmHg. 76 ambulatory patients with DBP of 95 to 130 mmHg after 2 weeks of placebo run-in. Mean age: 47 years. 24% male. 3 treatment groups (od): Monotherapy: Oxprenolol 160mg (O160); Chlorthalidone 20mg (C20) Combination: O160/C20 Trough supine SBP and DBP and HR; WDAE Efcacy analysis in 75 patients Safety analysis in all 76 patients Language: Spanish Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description randomized - see English abstract on page 163 not described double-blind
Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Free of selective reporting?
Unclear Unclear
Yes
Efcacy analysis in 75/76 (99%) patients.
Yes
Brown 1990 Methods 4-week double-blind study after a 2 week placebo period. Randomized (methods not described) 40 patients, aged between 18 and 70 yrs with supine diastolic BP between 95 and 115mmHg after placebo phase. Mean age 58yrs. 19 male and 21 female. 4 treatment groups (od): Placebo Monotherapy: Perindopril 4mg/day (P4); HCTZ 25mg/day (H25) Combination: P4/H25 (tablets were identical in appearance)
Participants
Interventions
31
Brown 1990
(Continued)
Outcomes Notes
Trough (24 hours post-dose) standing SBP and DBP; WDAE; HR (no quantitative data) Heart rate was not inuenced by any of the drug treatments used Funding: Not reported
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description patients selected were then randomised to treatment not described In order to ensure double-blind conditions, each patient received once daily, two tablets which were identical in appearance. At the end of the study, the double-blind code was then broken by a nurse who had no part in the clinical study No dropouts. All patients in the group of interest was included in the efcacy analysis.
Unclear Yes
Yes
Yes
Camera 1995 Methods 8-week double-blind active treatment following a 4-week single-blind placebo treatment. Randomized (methods not described). 149 patients, aged between 30 and 70 years, with uncomplicated mild to moderate hypertension (DBP 90-114mmHg). 42% men. Mean age: 53 years. 2 treatment groups (o.d.): Monotherapy: Lisinopril 20mg (L20) Combination: L20 plus HCTZ 12.5 Trough (24-28 hours post-dose) sitting SBP and DBP; WDAE 143/149 patients completed the study. Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Authors judgement Description

32
Camera 1995
(Continued)
Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Unclear Unclear Unclear
Patients were randomly assigned not described the trial took the form of a randomised double-blind study 8 weeks: 3/75 patients withdrew from study in combination group (2 WDAE); 3/74 patients withdrew from study in monotherapy group (2 WDAE, 1 lost to follow-up) Efcacy analysis was assumed to have included all randomized patients.
Yes
Yes
Chadha 1983 Methods 6 week double-blind trial after a 4-week placebo washout period. Randomized (methods not described). After 4 weeks of therapy, a second dose in the evening was added if DBP >95mmHg. On admission a low salt diet of 3-6g NaCl was prescribed. Data up to 4 weeks will be used in the review. 41 adult outpatients, aged 30-70 years, with DBP between 95 and 120mmHg at the end of placebo run-in. 3 treatment groups (o.d.): Monotherapy: Penbutolol 40mg (P40); Frusemide 40mg (F40) Combination: P40/F40 Double-dummy technique Erect SBP and DBP and HR (timing of measurement not reported) Funding: Not reported
Participants
Interventions
Outcomes Notes Risk of bias Item Adequate sequence generation?
Authors judgement Unclear
Description randomly allocated; the randomization process failed to produce groups which were strictly comparable in terms of the level of hypertension not described
Allocation concealment?
Unclear
33
Chadha 1983
(Continued)
Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Yes
double-blind; a double dummy technique was used Number of patients included in efcacy analysis was not reported. It was assumed in this review that all patients were included in the efcacy analysis
Unclear
Yes
Chrysant 1994 Methods 8-week multicenter, double-blind, parallel, placebo-controlled study after a 4-week, single-blind placebo period. 505 patients whose sitting diastolic BP was 100 to 114mmHg after placebo period. 311 (62%) men. Mean age: 53 yrs. 67% white. 24% black. 6 treatment groups (od):Placebo Monotherapy: Lisinopril 10mg (L10); HCTZ 12.5 mg (H12.5); H25 Combination: L10/H12.5; L10/H25 Trough (242 hours post-dose) sitting SBP and DBP 467 patients completed the study. Efcacy analysis: ITT (467 patients) Funding: ICI Pharmaceuticals Group
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomized into the double-blind phase not described This was a blind...study multicenter, double-
Yes
Ecacy analysis performed on patients who completed the study. 8/168 missing from combination group. 8/87 missing from monotherapy group. Comment: Percentage of missing patients was low. Risk of bias was judged to be low.
34
Chrysant 1994
(Continued)
Yes
Chrysant 1996 Methods 6-week, double-blind, parallel multicenter study after a 1-4 week placebo run-in period. Randomized (methods not described). 334 outpatients, aged 18 years or older, with sitting diastolic BP between 95 and 114mmHg at two consecutive visits during the placebo phase, with a difference of 10mmHg or less at the two visits. 210 (63%) men. 26% black. 70% white. Mean age: 53.5yrs. 8 treatment groups (o.d.): Placebo Monotherapy: Benazapril HCl 20mg (B20); HCTZ 25mg (H25) Combination: B5/H6.25; B10/H12.5; B20/H25; B20/H6.25; B5/H25 Study medication was provided as identical appearance and capsules of identical appearance. Trough (22 to 26 hrs post-dose) sitting SBP and DBP; HR (no quantitative data) Safety analysis in 334 patients No clinically important changes from baseline were reported in mean pulse for any treatment group Funding: Ciba Pharmaceuticals
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Yes Description according to a computer-generated randomization table study medication was provided as tablets of identical appearance and capsules of identical appearance, which concealed treatment allocation Comment: methods used to assigned treatment not described (e.g. central allocation or opaque sealed envelopes) double-blind...trial; study medication was provided as tablets of identical appearance and capsules of identical appearance, which concealed treatment allocation
Unclear
Blinding? All outcomes
Yes
35
Chrysant 1996
(Continued)
Yes
An intent-to-treat method was used wherein statistical analysis of data from all patients randomized to receive doubleblind treatment was performed (n=328) last post-randomization measurement carried forward Week 6: 41/43 patients in benazepril group and 83/85 patients in combination group not included in efcacy analysis Comment: Missing patients probably will not make a signicant difference. 90% of patients completed the trial.
Yes
Chrysant 2004 Methods Participants 8-week double-blind factorial design study after a 4 week placebo run-in period. 502 patients with average sitting DBP >=100mmHg and <=115mmHg at both week 3 and week 4 placebo run-in visits and a difference of <=7mmHg between the two measurements. 55.6% male and 74.1% white. Mean age: 53years. 12 treatment groups (od): Placebo Monotherapy: Olmesartan 10, 20 or 40mg; Hydrochlorothiazide 12.5 or 25mg Combination: All possible combinations Trough sitting SBP and DBP; percentage of WDAE Percentage of WDAE was low (2%) last observation carried forward for patients who did not complete the protocol ; 451/502 (90%) patients completed the study Funding: Sankyo Pharma Inc.
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomized not described double-blind treatment
36
Chrysant 2004
(Continued)
Yes
2 patients in the combination groups were missing in the intent-to-treat analysis. Comment: Risk of bias was judged to be low.
Yes
Drayer 1995 Methods 8-week double-blind, multicenter, placebo-controlled, parallel group study after a 4week placebo run-in phase. Randomized (methods not described) 413 ambulatory outpatients with sitting DBP between 95 and 114 mmHg after placebo period. 9 treatment groups (od): Placebo Monotherapy: Hydrochlorothiazide 12.5mg (H12.5); Moexipril 3.75mg (M3.75); M7.5; M15; M30 Combination: H12.5/M3.75; H12.5/M7.5; H12.5/M15 Trough sitting SBP and DBP; WDAE; HR (no quantitative data) no signicant changes in HR within any of the combination or monotherapy treatment groups Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomly allocated not described double-blind treatment period
Yes
all patients included in efcacy analysis 391 completed the 8-week double-blind period, whereas 22 discontinued prematurely (14 due to adverse experiences, 2 because of lack of therapeutic response, and the others for such reasons as protocol violations, withdrawal of consent, or loss to follow-up). These 22 patients were included in efcacy analysis, and their last37
Drayer 1995
(Continued)
available trough blood pressure values during the double-blind period were pooled with the 8-week treatment values of the other patients to obtain an endpoint for all patients Comment: not clear how many patients withdrew from each group and if the rate is similar between gruops. Free of selective reporting? Yes
Fernandez 1994 Methods Participants 8-week double-blind study. Randomized (method not described) 67 patients of either sex 18 to 75 yrs of age with seated diastolic BP >=95 and <=110 mm Hg. Mean age: 53 yrs. 23 (34%) male. 4 treatment groups (o.d): -placebo -monotherapy: fosinopil 20mg (F20), hydrochlorothiazide 12.5mg (H12.5) -combination: F20/H12.5 sitting trough SBP and DBP (24+-3hrs after previous dose) Funding: Not reported
Interventions
Outcomes Notes Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes
Authors judgement Unclear Unclear Unclear
Description Patients were randomized not described Combination therapy consisted of individual distinct tablets of each drug; double-blind procedure was not tested at the end of the study Comment: No mentioning of the appearance of placebo drugs. Efcacy Analyses: (BP data at baseline and at least one follow-up visit) performed in all randomized patients. At 6 weeks of therapy two patients were withdrawn from the study (1 in fosinopril, 1 in placebo). Another patient on placebo abandoned the
38
Yes
Fernandez 1994
(Continued)
study at week 4 for unknown reasons Free of selective reporting? Yes
Frei 1994 Methods 8-week multicenter, double-blind, placebo-controlled, parallel-group trial after a 4-week placebo run-in phase. Randomized (methods not described) 161/177 patients with DBP >=95 and <=114mmHg, and SBP <=240mmHg. Mean age 55.1yrs. 4 treatment groups (od): Placebo Monotherapy: Moxonidine 0.4mg (M0.4); HCTZ 25mg (H25) Combination: M0.4/H25 Sitting BP (timing of measurement not reported); WDAE Funding: not reported
Participants
Interventions
Outcomes Notes Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Description randomized not described double-blind
Yes
Because one patient had no post day 0 efcacy assessment, only 160 patients could be analyzed in the intent-to-treat efcacy sample Comment: 160/161 patients included in efcacy analysis. Risk of bias was judged to be low.
Yes
39
Frishman 1994 Methods 12-week double-blind, placebo-controlled, multicenter 3x4 factorial trial after a 4-6 week single-blind placebo phase. Randomized (methods not described). 512 patients aged 21 years or older with mild to moderate essential hypertension whose weight was within 35% of the ideal for height and frame were eligible for randomization. Mean sitting diastolic BP was stable and between 95 and 115 mmHg (inclusive). 364 (71%) male. 366 (71%) non-black. 12 treatment groups (od): Placebo Monotherapy: HCTZ 6.25mg (H6.25); H25; Bisoprolol 2.5mg (B2.5); B10; B40 Combination: H6.25/B2.5; H6.25/B10; H6.25/B40; H25/B2.5; H25/B10; H25/B40 Trough (24 hours post-dose) sitting SBP and DBP; WDAE 403/512 patients completed the study. Funding: American Cyanamd Co.
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Yes Description randomized not described To maintain blinding, matching placebo tablets were provided for both bisoprolol and hydrochlorothiazide the primary efcacy variable was dened a priori as the change from baseline sitting diastolic blood pressure in patients evaluable at weeks 3-4 (n=465) week 12: 17/190 in combination group and in 17/197 monotherapy group missing in efcacy analysis Comment: An equal percentage (low) of patients missing in the groups.
Yes
Yes
Frishman 1995 Methods 4-week multicenter, double-blind, placebo-controlled, parallel-group study after 4- to 6-week single-blind placebo treatment.
40
Frishman 1995
(Continued)
Participants
547 patients, 21 years or older with sitting DBP between 95 and 115mmHg on 3 consecutive weekly visitng during placebo period. 4 treatment groups (od): Placebo Monotherapy: Bisoprolol 5mg (B5); HCTZ 6.25mg (H6.25) Combination: B5/H25 Trough (24hours post-dose) sitting SBP and DBP and HR; WDAE 509/547 patients for efcacy analysis [310 (61%) male. 82% non-black, 18%black] Funding: American Cyanamid Company.
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomized not described double-blind treatment
Yes
The primary efcacy analysis was based on the change from baseline sitting DBP in patients with a valid baseline and valid sitting DBP reading on the week 3 and 4 visits week 4: 10/160 patients in combination therapy and 7/158 patients in monotherapy group missing in efcacy analysis. Comment: Percentage of missing patients was low. Risk of bias was judged to be low.
Yes
Genthon 1994 Methods 8-week double-blind, multicentre, parallel-group trial following a 4-week, single-blind, placebo run-in period. Randomized (methods not described) 660 patients of either sex, aged between 18 and 75 years, with supine DBP of more than 95mmHg but less than 115mmHg after placebo run-in. Systolic BP <=200mmHg. 3 treatment groups (o.d.): Monotherapy: Ramipril 2.5mg (R2.5); HCTZ 12.5mg (H12.5)
41
Participants
Interventions
Genthon 1994
(Continued)
Combination: R2.5/H12.5 Outcomes Trough (22-26 hours post dose) supine SBP and DBP; WDAE; HR (no quantitative data) 624/660 patients completed the study. Whilst heart rate was stable in all groups throughout the study, there was a slight decrease in all groups Funding: Laboratoires Hoechst
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomized not described double-blind treatment
No
Efcacy Analysis: per-protocol analysis which excluded certain protocol violators. A total of 125 patients did not strictly fulll the inclusion criteria for supine DBP during the placebo run-in and so were exlcuded, thus giving a per-protocol population of 535 patients. Data for ITT analysis not reported but the actual reductions were smaller than those seen in the per-protocol analysis, because patients excluded from the per-protocol analysis were often less hypertensive Comment: High risk of bias. 15-25% of patients missing in the monotherapy and combination groups.
Yes
Hart 1991 Methods 8-week double-blind, parallel-group multicentre study after a 2-week single-blind placebo run-in period. Randomized (methods not described) 299 patients aged 65-80 years with a sitting DBP of 100-120mmHg (inclusive) during the placebo run-in period. 123 (41%) male. Mean age: 71 yrs
Participants
42
Hart 1991
(Continued)
Interventions
3 treatment groups (od): Monotherapy: Lisinopril 20mg (L20); HCTZ 12.5mg (H12.5) Combination: L20/H12.5 Trough (24-28 hours) sitting SBP and DBP; WDAE 278/299 patients completed the study There was no statistically signicant difference between the groups for HR 2 patient died during the study for reasons unrelated to the trial medication Funding: Not reported
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomly assigned not described double-blind active treatment
No
Efcacy was based on per-protocol analysis: patients were considered to be protocol deviator, for example if the number of tablets returned was less or more than expected or most importantly, if there was any deviation from the protocol during the interval betwen the last dose of medication and BP measurement; 153/299 (51%) patients complied with protocol. Comment: almost half of the patients were excluded.
Yes
Homuth 1993 Methods 6-week multiclinic, double-blind, placebo-controlled trial after a 2-week placebo run-in period. Randomized (methods not described) 480 patients, aged 21 to 65, with a diastolic BP between 100 and 115mmHg after runin period. 295 (74%)men. Mean age: 47 years. 12 groups (od): Placebo Monotherapy: Ramipril 2.5mg (R2.5); R5; R10; Piretanide 3mg (P3); P6 Combination: R2.5/P3; R2.5/P6; R5/P3; R5/P6; R10/P3; R10/P6
43
Participants
Interventions
Homuth 1993
(Continued)
Outcomes Notes
Trough (24hrs post-dose) SBP and DBP; WDAE BP data obtained from g 1 and g 2 on page 669. The position of measurement was not mentioned. Funding: Cassella AG
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Yes Description randomized not described The regimens had an identical apperance and the physicians, nurses and patient0care personnel, as well as the patients, were unaware of the regimens; double-blind ITT analysis: Any randomized patient meeting the inclusion criteria who had any postrandomization data availbale at any visit during the double-blind treatment phase was included in the analyses 452/480 (94%) patients completed the study Comment: it was assumed that all patients were included
Yes
Yes
Kayanakis 1987 Methods 8-week multicenter, double-blind, placebo-controlled parallel study after a 2 week placebo run-in period; randomized (method not described) 211 men and women aged 20-70 yrs with diastolic BP between 95 and 120 mm Hg and systolic BP between 160 and 200mmHg. Mean age: 53 yrs. 56% male 4 groups (od): Placebo Monotherapy: Captopril 50mg (C50) ; HCTZ 25mg (H25) Combination: C50/H25 Supine trough SBP and DBP (20-24 h after the last medication); HR (data not shown); WDAE
Participants
Interventions
Outcomes
44
Kayanakis 1987
(Continued)
Notes
205 completed the study Heart rate did not change signicantly with either treatment Funding: Not reported
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description Patients were randomised not described A double-blind design was used
Yes
all patients were included in the efcacy analysis No patients in the combination group withdrew from the study. 1 patient in the captopril withdrew due to inefcacy.
Yes
Kellaway 1993 Methods 8-week multicenter, double-blind, parallel group study after a 2-week single blind placebo run-in period. Randomized (method not described). Non responders at week 4 received double dose of cilazapril. Data up to week 4 were used. 69 patients, aged 18 to 75 yrs, with sitting DBP 95-115 mmHg at end of placebo runin. 2 treatment groups (od): Monotherapy: Cilazapril 2.5mg (C2.5) Combination: C2.5/HCTZ12.5 Sitting SBP and DBP (timing of measurement not reported); WDAE; heart rate (data not shown) 87 patients entered run-in phase (51%male; mean age: 54.6yrs) Heart rate did not differ signicantly at week 4 and 8 Funding: Not reported
Participants
Interventions
Outcomes
Notes
Risk of bias Item Authors judgement Description
45
Kellaway 1993
(Continued)
Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Unclear Unclear Unclear
Patients were randomised not described double-blind...study
No
57/69 patients were valid for efcacy analysis. Type of analysis employed not described. Comment: 7/21 (23%) of patients missing in the combination group. High risk of bias.
Yes
Kochar 1999 Methods 8-week double-blind, placebo-controlled parallel study after a 4-5 week single-blind placebo phase (conducted at 46 sites in the US); randomized (method not described) 683 men and women (>=18yrs) with seated diastolic BP between 95 and 110 mm Hg at weeks 3 and 4 or optional week 4 and 5 of placebo lead in phase. Difference between weeks of <=8mm Hg. Mean age: 55+-10.5. 65% male, 85% white and 48 (14%) black. 4x4 factorial design (od) placebo Monotherapy: Irbesartan 37.5mg (I 37.5); I 100; I 300; Hydrochlorothiazide 6.25 (H 6.25); H 12.5; H 25 Combination: I 37.5/H 6.25; I 37.5/H 12.5; I 37.5/H 25; I 100/H 6.25; I 100/H 12.5; I 100/H 25; I 300/H 6.25; I 300/H 12.5; I 300/H 25 trough (24+-2h after the last dose) sitting SBP and DBP; WDAE Safety analysis: patients with at least one dose of randomized therapy = 683 631 completed the study Funding: Bristol-Myers Squibb Pharmaceutical Research Institute
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Authors judgement Unclear Unclear Description patients were then randomized not described
46
Kochar 1999
(Continued)
Unclear
double-blind treatment; All patients were insturcted to take three capsules (irbesartan or matching placebo) and one table (HCTZ or matching placebo) comment: are they identical in size, shape, taste? Analyses of efcacy assessments included data for all patients who had a baseline evaluation and at least one on-therapy evaluation Of the 683 patients randomized to double-blind therapy, 52 patients were discontinued. Of the patients who completed the study, six were not inluded in the efcacy analyses because of incomplete source documentation at one investigators site. In addition, ve discontinued patients had data collected at week 8 and so were included in the efcacy analysis. Thus 630 patients were included in the efcacy analysis Comment: 8% of patients missing in efcacy analysis (low risk)
Yes
Yes
Lacourciere 1994 Methods 12-week double-blind, placebo-controlled, parallel 3 X 4 factorial design study after a 4-week single-blind placebo run-in period. Randomized (methods not described) 240 outpatients of both sexes, aged 18-70 yrs, with sitting DBP between 95 and 110mmHg, on the last visit during the run-in period. On average, the patient population was aged in the early fties and predominantly male. 12 treatment groups (od): Placebo Monotherapy: Nebivolol 1mg (N1); N5; N10; HCTZ 12.5mg (H12.5); H25 Combination: N1H12.5; N1/H25; N5/H12.5; N5/H25; N10/H12.5; N10/H25 Trough (24hrs post dose) sitting SBP and DBP; HR (no extractable data provided for analysis) No signicant changes in HR in N1 or H monotherapies and combination groups. N5 and N10 monotherapies and in combination with H induced a signicant reduction from baseline in seated HR (range 10.4-12.4bpm and 2.0-11.5bpm respectively) Funding: Janssen Research Foundation and Le Centre Hospitalier de lUniversite Laval, Research Centre
47
Participants
Interventions
Outcomes
Notes
Lacourciere 1994
(Continued)
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomly assigned not described double-blind treatment
Yes
All 240 randomized patients were included in the intention-to-treat efcacy analysis 226/240 (94%) patients completed the study.
Yes
Lacourciere 2005 Methods 8 week double-blind, parallel group study after a 2-week placebo run-in period. Randomized by the next availbale sequential number to receive double-blind treatment. 774 patients with SBP >=160 and <=200mmHg after washout period. Mean age: 60years. 3 treatment groups (od): Monotherapy: Valsartan 160mg Combination: Valsartan 160mg plus Hydrochlorothiazide 12.5mg or 25mg Drugs were forced titrated at week 4: Group 1: from valsartan 80mg to valsartan 160mg Group 2: from valsartan 160mg to valsartan 160mg plus H12.5 Group 3: from valsartan 160mg to valsartan 160mg plus H25 Trough sitting SBP and DBP; WDAEs; ITT population: 767 patients; 411 (54%) men. Funding: Novartis Pharmaceuticals
Participants Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Yes Description patients were randomized by the next available sequential treatment number to receive double-blind reatment with V80 ( monotherapy group) or V160 (combina48
Lacourciere 2005
(Continued)
tion therapy groups) Allocation concealment? Yes patients were randomized by the next available sequential treatment number to receive double-blind reatment with V80 ( monotherapy group) or V160 (combination therapy groups) double-blind treatment
Unclear
Yes
The ITT population consisted of all enrolled patients who received study medication and for whom at least 1 measurement was obtained after baseline; The ITT population consisted of 767/774 patients. The 7 missing patients were from the combination group 18/261 (7%) patients in valsartan monotherapy group were withdrawn from study. (12 to adverse events, 4 to lack of efcacy, and 2 to investigators decision) 13/513 (3%) patients in combination groups were withdrawn from study. (13 to adverse events, 8 to lack of efcacy and 13 to investigators decision). Comment: Risk of bias was judged to be low.
Yes
Lenz 1994 Methods 8 week multicentre, double-blind, forced-titration parallel group study after 4-week placebo-baseline period. Randomized (methods not described). At week 4, doses are doubled. Only data up to 4 weeks are used. 368 men and women, at least 18yrs, with supine DBP >=105mmHg and <=120mmHg at two consecutive visits at end of palcebo phase. 165 (45%) male. 364 (99%) white. 3 treatment groups (od): Monotherapy: Quinapril 10mg (Q10); HCTZ 12.5mg (H12.5) Combination: Q10/H12.5 Trough standing BP; WDAE 346/368 patients completed the study. All patients included in safety analysis.
49
Participants
Interventions
Outcomes Notes
Lenz 1994
(Continued)
Funding: Parke-Davis GmbH Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description Patients were randomized not described double-blind treatment
Yes
An intent-to-treat analysis of efcacy data was performed on all patients having data from the placebo-baseline phase and from the double-blind phase Patients with only DB data within fewer than 19 days are excluded. (n=318) 4weeks: 19/120 missing in quinapril group and 16/124 patients missing in combination group for efcacy analysis. Reasons for exclusion from efcacy evaluation were inadequate baseline blood pressure, concomitant medication that had not been approved, late discontinuation of prior antihypertensive medication, blood pressure measurements taken outside the 20-28 h postodse limit, or within fewer than 19 days on the respective dose Comment: About 13-15% of patients missing. The risk of bias was judged to be low.
Yes
Li 2003 Methods 8-week double-blind, parallel-group study after a 2-week single-blind placebo phase. Randomized (methods not described). After 4 weeks, patients with uncontrolled sitting DBP (>=90mmHg) received double doses. Only data up to 4 weeks of treatment was used in this review. 179 men and women aged 18-65 with sitting DBP 95-115mmHg and sitting SBP <180mmHg after placebo washout. Mean age: 46.6 yrs. 65%male. 2 treatment groups (od): Monotherapy: Losartan 50mg Combination: Losartan 50mg plus HCTZ 12.5mg
50
Participants
Interventions
Li 2003
(Continued)
Outcomes Notes
Trough (22-26 hours post dose) sitting SBP and DBP; WDAE Safety analysis: (n=179) Funding: Merck & Co Inc.
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description patients were randomised to active treatment not described double-blind
Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Unclear Unclear
Yes
the primary analysis used an all patients treated approach, i.e. inlcuded all patients who received active treatment and who had valid blood pressure measurements at baseline and on treatment; four did not have any treatment period data. 2 patients missing from each group. Comment: Not likely to have a high risk of bias
Yes
MacKay 1996 Methods 12-week double-blind, parallel group placebo-controlled multicenter study after a 4-week single blind placebo run-in period. Stratied randomization (methods not described). (same study as Schoenberger 1995) 703 men and women at least 18 years of age with sitting DBp >=95mmHg after rst 2 weeks of placebo run-in, and sitting DBP of 95 to 115mmHg (not differed by more than 7mmHg) after last 2 weeks of placebo run-in. 420 (60%) male. Mean age: 53 yrs. 86% white, 12% black, 2% other. 5 treatment groups (od): placebo monotherapy: losartan 50mg (L50), HCTZ 12.5mg (H 12.5) combination: L50/H6.25; L50/H12.5 Trough (22-26hr) sitting SBP and DBP; heart rate (data not shown), WDAE
Participants
Interventions
Outcomes
51
MacKay 1996
(Continued)
Notes
No clinically signicant mean changes from baseline for heart rate in any groups 604/703 patients completed the study. Safety analysis: all patients for whom safety data were available Funding: Merck Research Laboratories, Clinical Research.
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomly allocated not described double-blind
Yes
The primary analysis of efcacy in this study included all randomized patients with at least one treatment period measurement. The last double-blind measurements of the withdrawn patients were carried forward to subsequent time points. Week 12: 1/139 patients in monotherapy group and 9/282 patients in combination therapy group missing in efcacy analysis. Comment: Percentage of missing patients was low. RIsk of bias was judged to be low.
Yes
Manning 1996 Methods 6-week multicentre, double-blind, placebo-controlled study after a 4-6 week run-in period. After the 6-week treatment, another 4-week treatment followed where patients could be titrated to 2x dosage depending on the response (data not retrieved). Randomized (methods not described). 63 patients, aged between 18 and 80 years, with supine resting diastolic BP between 95 and 115 mmHg on two occasions and the lower reading within 10% of the higher reading during the run-in period. Mean age: 53.3yrs. 64% male. 3 treatment groups (od): Monotherapy: Controlled release diltiazem 150mg (D150); Normal release HCTZ 12.5mg (H12.5) Combination: D150/H12.5 Supine SBP and DBP (timing of measurement not reported); HR; WDAE
52
Participants
Interventions
Outcomes
Manning 1996
(Continued)
Notes
61/63 patients completed the study Funding: Not reported
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description were randomised not described double-blind
Yes
2/63 patients missing in efcacy analysis two patients were withdrawn during the treatment period due to adverse events. Comment: low risk of bias with 3% of patients missing.
Yes
McGill 2001 Methods 8-week multicenter, double-blind, double-dummy, placebo-controlled, parallel group, 4x5 factorial study after a 4-week, single-blind, placebo run-in period. Randomization was according to enrolment order and a computer-generated list, and was stratied by race (black/non-black) 818men and women aged between 18 and 80 years with supine DBP between 95114mmHg during the last 2 weeks of placebo run-in and SBP between 140 and 200mmHg immediately before randomization. Mean supine DBP could not vary by >7mmHg over last 2 weeks of run-in. Mean age: 53 yrs. 60% men. 27.1black and 72.9% non-black. 4x5 treatment groups (o.d.): Placebo Monotherapy: Telmisartan 20mg (T20); T40; T80; T160; HCTZ 6.25mg (H6.25); H12.5; H25 Combination: T20/H6.25; T40/H6.25; T80/H6.25; T160/H6.25; T20/H12.5; T40/ H612.5; T80/H12.5; T160/H12.5; T20/H25; T40/H25; T80/H25; T160/H25 Trough (24 hours post dose) supine SBP and DBP; HR (no extractable data); WDAE Of the 1293 patients screened, 818 were enrolled. 749/818 patients completed the trial. No signicant changes from baseline were seen in supine trough heart rate with any of the active treatments
53
Participants
Interventions
Outcomes Notes
McGill 2001
(Continued)
Funding: Not reported Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description a computer-generated list was used for the allocation to groups, with allocation by enrolment order a computer-generated list was used for the allocation to groups, with allocation by enrolment order double-blind, double-dummy
Unclear
Yes
Yes
ITT: randomized patients with 1 postrandomization BP measurement (last observation carried forward) the ITT population included 807 (98.7%) of the original 818 patients (11 patients were excluded for lack of evaluable post-randomization measurements of vital signs). Comment: Risk of bias was judged to be low.
Yes
Merrill 1987 Methods Participants Interventions 8-week multiclinic double-blind trial. Randomized (methods not described) 207 patients with a sitting DBP of 90-115mmHg after a a 4week placebo baseline. 5 treatment groups (od): Monotherapy: Lisinopril 20mg (L20); HCTZ 12.5mg (H12.5) Combination: L20/H6.25; L20/H12.5; L20/H25 Trough SBP and DBP (timing of measurement not reported); WDAE seventeen patients (8.2%) were discontinued due to adverse effects Source: Paper abstract Funding: Not reported
Outcomes Notes
Risk of bias
54
Merrill 1987
(Continued)
Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Free of selective reporting?
Description randomly allocated not described double-blind trial
Yes
all patients were included in the efcacy analysis No systolic BP data was provided. It is not mentioned in the paper abstract whether SBP was one of the primary outcome.
Unclear
Mersey 1993 Methods 8-week double-blind placebo-controlled multicenter study after 4-6 week placebo period. Randomzed (methods not described) 345 white men and women with sitting DBP of 92 to 109 on two occasions during the placebo period. Efcacy data in 322 patients (58% men; mean age: 50.8yrs) 5 treatment groups (o.d.): Placebo Monotherapy: Captopril 25mg (C25); HCTZ 12.5mg (H12.5) Combination: C25/H12.5; C50/H25 Trough (24+-3 hrs post dose) sitting SBP and DBP; HR (no extractable data); WDAE 296/345 (85.6%) patients completed the study the only statistically signicant change in HR was in C50/H25 group (3beats/min) Funding: Bristol-Myers Squibb Company
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomly assigned not described in a double-blind manner
55
Mersey 1993
(Continued)
Yes
week8: 5/68 (7%) patients missing in monotherapy and 7/69 (10%) patients missing in combination therapy for efcacy analysis. Possible exclusion include: loss to follow-up, adverse drug experiences, patient request; concomitant illness, noncompliance, prohibited medication, or seated DBP<92mmhg after the lead in period Of 345 patients who were randomized 296 (85.8%) completed the 8-week treatment period Comment: The number of patients in each group who did not complete the study and had data carried forward were not given.
Yes
Meyer 1994 Methods 16-week multicenter trial after 4-week single-blind placebo run-in. Randomized (methods not described) 205 patients with DBP of 95-115mmHg after placebo run-in. 63% male. Mean age: 51 yrs. 3 treatment groups (od): Monotherapy: Trandolapril 2mg (T2); HCTZ 25mg (H25) Combination: T2/H25 Trough (24 hours post-dose) supine SBP and DBP; WDAE (treatment-related only); HR (no quantitative data) No clinically signicant changes in supine or standing heart rate Funding: Not reported
Participants
Interventions
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomly assigned not described double-blind fashion
56
Meyer 1994
(Continued)
Incomplete outcome data addressed? All outcomes Free of selective reporting?
Yes
Efcacy analysis performed in overall population
Yes
Neutel 2008 Methods 12-week double-blind, parallel-grop study after a 21-day single-blind placebo wash-out period. Randomized (methods not described) 538 patients randomized. Patients were aged 18 years or older with moderate hypertension that at enrolment was either untreated for at least 4 weeks or uncontrolled by monotherapy. Untreated patients were enrolled if seated SBP 160-179mmHg or seated DBP 100-109mmHg. Pateints on monohterapy could be enrolled if seated SBP 150179mmHg or DBP 95-109mmHg. Mean age 55yrs. 292 (55%) male. 84% white and 14% black/African-American. 228 (42.4%) had hyperlipidemai,. 74 (13.8%) had diabetes mellitus. 3 treatment groups (od): Monotherapy: Hydrochlorothiazide - H12.5 for 2 weeks followed by forced titration to H25 for 10 weeks Irbesartan - I150 for 2 weeks followed by forced titration to I300 for 10 weeks Combination: I150/H12.5 for 2 weeks followed by forced titration to I300/H25 for 10 weeks Trough sitting SBP and DBP; WDAE 472 patients (87.7%) completed double-blind treatment. Safety analyses in all randomized patients who took at least 1 dose of study medication (n=538) Funding: Bristol-Myers Squibb and Sano-Aventis
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients randomized not described double-blind
Yes
All efcacy analyses were made on an intent-to-treat basis, using the data set of all randomized subjects The single most common reason for pre57
Neutel 2008
(Continued)
mature withdrawal was AEs. One patient in each group withdrew due to lack of efcacy similar numbers discontinued prematurely in combination (41; 12.5%) and irbesartan group (12; 11.3%). Free of selective reporting? Yes
Oparil 1980 Methods 8-week double-blind study after 6 week placebo period; randomized study (methods not described) 97 outpatients of either sex, 21 to 65 yrs of age, with supine diastolic BP between 100 and 120 mm Hg. Difference between weeks 2 or 4 and week 6 of placebo phase <=10mm Hg. Mean age: 51 yrs. 34 (35%) male; 62 (64%) white and 35 (36%) black. 3 treatment groups (BID): Monotherapy: Timolol maleate 10mg (T10); HCTZ 25mg (H25) Combination: T10/H25 Erect SBP and DBP; WDAE; supine HR (only range was given; not extracted for analysis) 90/97 patients completed the study Efcacy analysis not described (n=78 at week 8) For HR, the decrease (beats per minute) ranged from 16 to 18 in the timolol/HCTZ group, 14 to 17 in the timolol group, and 5 to 6 in the HCTZ group Funding: Merck Sharp & Dohme
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description Patients at each clinic were assigned at random; randomized block design Comment: size of block not mentioned not described double-blind; all medication were administered as identically appearing tablet 5, 10, 8 of the 34 patients in combination group missing for efcacy analysis at week 4,6,8 respectively. 5, 10, 5 of the 33 patients in monotherapy group missing for efcacy analysis at week
58
Unclear Yes
No
Oparil 1980
(Continued)
4,6,8 respectively. Comment: 15-30% of patients missing. No reasons given. Because a large percentage of randomized patients are excluded, there is a high risk of bias. Free of selective reporting? Yes
Papademetriou 2000 Methods 8-week multcenter, double-blind, placebo controlled study after a 4-5week placebo runin phase. Randomized (methods not described). 275 patients with DBP between 95 and 114 mmHg on 2 separate occasions during qualifying clinic visists. 56% male, 21% black. Mean age: 52 years. 4 treatment groups (od): Placebo Monotherapy: Hydrochlorothiazide 12.5mg (H12.5); Candesartan 32mg (C32) Combination: H12.5/C32 Trough sitting SBP and DBP; WDAE; HR (no extractable data) No clinically signicant changes in HR Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomized not described double-blind treatment phase
Unclear
Primary efcacy analysis was conducted on the intent-to-treat population, with the last observation carried forward Comment: it was assumed that all patients were included in the ITT.
Yes
59
Papademetriou 2006 Methods 8-week multicenter, double-nliond, parallel group, unbalanced factorial study after a 4to 5-week single-blind placebo run-in period. 1571 patients aged 18 to 80 years with sitting DBP 95 to 114mmHg and SBP <180mmHg after placebo run-in. Mean age: 53 years. About half were men. 26% were African American. 10% had diabetes mellitus type 2. 17 treatment groups (o.d.): Placebo Monotherapy: ER Metoprolol (M25, M50, M100, M200); Hydrochlorothiazide (H6.25, H12.5, H25) Combination: M25/H6.25; M25/H12.5; M50/H6.25; M50/H12.5; M100/H6.25; M100/H12.5, M100/H25; M200/H12.5; M200/H25 Trough sitting SBP and DBP; WDAE (not used; only total percentage of WDAE reported); HR (not extractable data) 2.9% of patients had WDAE. There were no deaths. 1395 (88.8%) completed the study. Heart rate decreased with increasing doses of ER-metoprolol (max decrease of 10beats/ min at 200mg) and did so independent of hydrochlorothiazide Funding: AstraZeneca LP
Participants
Interventions
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Authors judgement Yes Description A central, computer-gerenated randomization schedule using an interactive voice response system allocated patients to treatment groups within study center A central, computer-gerenated randomization schedule using an interactive voice response system allocated patients to treatment groups within study center To blind study treatmet, hydrochlorothiazide in identically appearing tablets of 6.25, 12.5, or 25mg, or matching placebo and ER-metoprolol tablets of 25 or 100mg or matching placebo, were blister-card packaged according to a double-dummy design Primary efcacy anlayses used the intentto-treat (ITT) population (all randomized patients taking at least one dose of study drug, and with at least one postbaseline BP) and imputed missign values from week 8
60
Yes
Yes
Yes
Papademetriou 2006
(Continued)
by carrying the last observation forward; of the 1571 randomized pateints, 1559 (99.2%) were included in the ITT) Comment: Less than 1% of patients missing. The risk of bias was judged to be low. Free of selective reporting? Yes
Parati 2006 Methods 12-week multicenter, double-blind, parallel-group trial after a 2-week placebo run-in period. 353 patients aged between 18 and 75 years and a sitting DBP between 95 and 110mmHg placebo wash-out.. Mean age: 57 years. 56% male. 10 treatment groups (o.d.): Monotherapy: Zofenopril (Z15, Z30, Z60); Hydrochlorothiazide (H12.5, H25) Combination: Z15/H12.5; Z30/H12.5; Z60/H12.5; Z15/H25; Z30/H25 Trough sitting SBP and DBP; WDAE (numbers of WDAE in each group not reported) ; HR (no extractable data) Total of 6 (1.7%) patients with WDAE. 330 patients completed the trial. Safety analyses: all randomized patients. No signicant differences in heart rate between groups or from baseline Funding: Menarini Industrie Farmaceutiche Riunite and Istituto Lusofarmaco dItalia
Participants
Interventions
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description recruited subjects were randomized not described double-blind...trial
Yes
ITT: all randomized patients who received at least one dose of the assigned treatment and who had at least one visit after baseline all the 353 randomized atients were valid for the intention-to-treat analysis
Yes
61
Philipp 1997 Methods 8-week double-blind, factorial design, multicentre study after a 4-week placebo run-in period. Randomized (methods not described) 1096/1306 patients aged 18-75 years with sitting diastolic BP of 95-110mmHg. Mean age: 55.1 years. 15 parallel groups (od): Placebo Candesartan Cilexetil 2, 4, 8 or 16mg, HCTZ 12.5 or 25mg Combination therapy iwth both agents at these respective doses Sitting SBP and DBP (timing of measurement not reported); HR (data not shown); WDAE (given as percentages; number of WDAE in each group not reported) No effects on heart rate 2.4% of patients withdrew from the study due to adverse occurrences Funding: Not reported
Participants
Interventions
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomised not described double-blind
Yes
efcacy results were analysed on an intentto-treat (ITT) basis Comment: Low percentage of missing patients. The risk of bias was judged to be low.
Yes
Pool 1987 Methods 12-week multicenter study after a placebo meication for 2-4 weeks. Randomized (methods not described). After week 4 or 8, the dosage could be double to achieve the target BP. Therefore, only data up to week 4 will be used in this review. 394 patients of either sex, aged >=18 years, with sitting DBP of 90-120mmHg inclusive after placebo period. 75% male. Mean age: 73 years. 73% white; 21% black.
Participants
62
Pool 1987
(Continued)
Interventions
3 treatment groups (o.d.)in 2:2:1 ratio: Monotherapy: lisinopril 20mg (L20); HCTZ 12.5mg (H12.5) Combination: L20/H12.5 Double-dummy technique Sitting SBP and DBP (timing of measurement not reported); HR Funding: Not reported
Outcomes Notes Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes
Description randomly assigned not described double-blind active treatment
Yes
The per-protocol method analyzed all completed patients but excluded protocol deviations and dropouts if dropout occurred prior to the time point being analyzed (n=368) The treatment groups were similar with regard to reasons discontinued week 4: 8/154 (5%) patients missing in monotherapy and 4/70 (6%) patients missing in combination group
Yes
Pool 1993 Methods 6-week multicenter, double-blind, placebo-controlled, parallel group trial after a 4-6 week single-blind placebo run-in phase. Randomized (methods not described) 298 patients, aged 18-70 years, with 2 consecutive weekly mean supine DBP reading of >=95 and <=110mmHg, that varied <=7mmHg after baseline phase. 4 treatment groups (BID): Placebo Monotherapy: Diltiazem SR 120mg (D120); HCTZ 12.5mg (H12.5) Combination: D120/H12.5
Participants
Interventions
63
Pool 1993
(Continued)
Outcomes
Trough (+-2hours prior ot the next scheduled dose) standing SBP and DBP; HR (not shown) 254/298 patients completed the study. 79% non-black. Mean age 54.4yrs. 66% men. no signicant changes in supine HR from baseline to end of study for any of the treatment groups Funding: Marion Merrell Dow Inc.
Notes
Risk of bias Item Adequate sequence generation? Authors judgement Yes Description blinded, blocked randomization schedule not described double-blind; all drugs and placebo were identical in appearance efcacy data set included only participants who were randomized and completed the entire study protocol wtihout being discontinued or having a protocol violation (254/298 patients) 34 of the 298 did not complete the entire protocol for a variety of reasons including adverse reactions. Ten others completed the protocol were not included in the efcacy data set due to protocol violations (primarily poor adherence). Comment: 15% of the patients were not included in the efcacy analysis. The risk of bias is judged to be low.
Unclear Yes
Yes
Yes
Pool 1997 Methods 8-week multicenter 4 X 4 factorial, double blind, parallel group trial after a single-blind, placebo lead-in of 4-5 weeks. Randomized (methods not described). 550 white, Asian or black outpatients aged 1 to 75 years inclusive with seated DBP >=95mmHg and <=110mmHg at consecutive visits (3rd or 4th weeks, or 4th and 5th weeks) during placebo lead-in. Mean age: 51.5 yrs. 335 (61%) male. 461 (84%) white and 82 (15%) black.
Participants
64
Pool 1997
(Continued)
Interventions
16 treatment groups (o.d.): Placebo Monotherapy: Fosinopril 2.5mg (F2.5); F10; F40; HCTZ 5mg (H5); H12.5; H37.5 Combination: F2.5/H5; F2.5/H12.5; F2.5/H37.5; F10/H5; F10/H12.5; F10/H37.5; F20/H12.5; F40/H5; F40/H12.5; F40/H37.5; Trough (24+-3hrs post dose) sitting SBP and DBP; HR (data not shown); WDAE Efcacy analysis: 1) using ITT population: patients having BP data at baseline and at least one follow-up visit and 2) using efcacy population: subset of the ITT population, made up of all patients who did not violate any of the terms in the protocol that might affect efcacy outcome. The results from both anlayses were similar. Only data on ITT population were presented (n=516). 506/550 patients completed the study Funding: Bristol-Myers Squibb Company
Outcomes Notes
Yes
efcacy analysis based on intent-to-treat population, dened as patients having blood pressure data at baseline and at least one follow-up vsit. week8: 12/293 (4%) missing in combination and 9/95 (9%) missing in fosinopril group for efcacy analysis Comment: Low percentage of missing patients. Risk of bias was judged to be low.
Yes
Pool 2007a Methods 8-week multicenter, double-blind, multifactorial study after a 3- to 4-week single-blind placebo run-in period. 1123 men and nonpregnant women >= 18 years of age with mean sitting DBP >=95mmHg after a placebo run-in. Patients with mean sitting DBP >=110mHg or SBP >= 180mmHg were excluded. Mean age 56.1yrs. 56% male. 92% while; 7% African
65
Participants
Pool 2007a
(Continued)
American. Interventions 11 treatment arms (o.d.): Placebo Monotherapy: Aliskiren (A75, 150 or 300mg); Valsartan (V80, 160, 320mg) Combination: A75/V80; A150/V160; A300/V320; V160/HCTZ 12.5 Sitting SBP and DBP (timing of measurement not reported); WDAE Safety analyses: all randomized patients who received at least one dose of study treatment (n=1123) Funding: Novartis Pharma AG
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description randomization by region was performed by the interactive voice response system provider using a validated system that automates the random assignment of treatment groups to randomization numbers randomization codes were kept strictly condential until the database was locked double-blind
Yes
Unclear
Yes
efcacy analysis were performed on the intent-to-treat (ITT) population (all randomized patients with baseline and at least one post baseline measurement during the double-blind treatment period). The last available BP measurement was used for patients who dropped out (n=1117) 56 (95%) of patients in combination group and 161 (91%) of patients in valsartan monotherapy group completed the trial. week8: 1/59 patient in combination group and 1/59 patient in monotherapy group missing in efcacy analysis. Comment: not likely to bias the result
Yes
66
Pool 2007b Methods 8-week multicenter, double-blind parallel-group trial after a 2- to 4-week, single-blind, placebo run-in period. Randomized (methods not described). 1346 patients aged >=18years with mean sitting DBP >=90 and <110mmHg after washout, and mean sitting DBP >=95 and <110mmHg at randomization. Mean age: 52.7years. 55%male; 69% white; 22% black.; 2% asian. 8 treatment groups (o.d.): placebo Monotherapy: Valsartan (V160 or V320); Hydrochlorothiazide (H12.5 or H25) Combination: V160/H12.5; V320/H12.5; V320/H25 Trough stting SBP and DBP; WDAE Safey analyses: all randomized patients who received trial medication in a double-blind manner. Sex distribution in baseline characteristics (P=0.025) Funding: Novartis Pharmaceuticals Corporation
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description were randomized not described double-blind fashion
Yes
ITT analysis: all randomized patients who had at a baseline and at least one postbaseline efcacy measurement (n=1329) week 8: a total of 15/1346 patients missing in efcacy analysis. Comment: 1% of patients missing in efcacy analysis. The risk of bias was judged to be low.
Yes
Prisant 2000 Methods 6-week double-blind, placebo-controlled, parallel-group, multicenter study after a 4week single-blind placebo treatment. Randomized (methods not described). 425 men or women aged 18-80 years with supine DBP within the range of 95 to 114mmHg after 3 and 4 weeks of placebo treatment with a difference between measurements of 7mmHg or less.
67
Participants
Prisant 2000
(Continued)
Interventions
Placebo Monotherapy: Diltiazem XR (120, 180, 240, or 360mg) o.d.; Indapamide (1.25 or 2.5mg) o.d. Combination: each of the combination of Diltiazem plus indapamide (excluding Diltiazem 360mg) Trough (before next morning dose) supine SBP and DBP; WDAE (only total was given; number of WDAE for each group not reported) Baseline demographics of all treated patients were similar (n=329): 60%male. Age: 5054yrs. 29 subjects (7.3%) with WDAE. no pattern was observed in the rate of discontinuations to adverse events or incidence of adverse events among patients who received any treatment Funding: Not reported
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description patients were randomly assigned to double-blind medication not described double-blind; blinding was achieved by placing diltiazem XR tablets and indapamide powder in capsules, which were identical in appearnace, size, shape, and color to those of their respective placebos. These capsules were blister-packed into medication cards and dispensed at each weekly visit 425 patients received double-blind medication. The variability relative to the effect for the lower dose of 1.25mg indapamide was too large to demonstrate consistent effects and will not be further discusses; Twenty-nine subjects withdrew from the study for adverse clinical events All these 125 patients were not included in the efcacy analysis. Data for I1.25mg in monotherapy and in combination were not reported. Comment: The analysis is subject to high risk of bias.
68
Unclear Yes
No
Prisant 2000
(Continued)
Yes
Romero 1995 Methods 8-week multicenter, double-blind, active-controlled parallel study after 2-4 week placebo phase; randomized (method not described) 323 men and women, >==18yrs, with supine diastolic BP of >=105 and <=120 mm Hg at 2 consecutive visits during the placebo period. Mean age: 53yrs. 9 black, 1 Arabian, 313 white. 57% male. 3 parallel treatment groups (o.d.) with placebo matching drugs: monotherapy: Quinapril 20mg (Q20); HCTZ 12.5mg (H12.5) combination: Q20/H12.5 Trough (24hr) supine DBP and SBP; HR (no quantitative data); WDAE Heart rate was not signicantly modied in any of the groups Funding: Parke Davis GmbH
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomized not described double-blind
Yes
Efcacy analysis using evaluable data analysis which included all patients wthout protocol deviations who had received at least 26 days of double-blind treatment. with the data of the last evaluable double-blind visit as endpoint. (n=291) 9/108 patients missing in combination group and 10/106 missing patients in quinapril monotherapy group for efcacy analysis. the most frequent reasons for exclusion were inadequate baseline DBP and time treated with study drug <26days Comment: 8-9% of patients missing in efcacy analysis. Risk of bias was judged to be low.
69
Romero 1995
(Continued)
Yes
Rosenthal 1990 Methods 4-week double-blind four-center study after a 2-week baseline placebo period. Randomized (methods not described). After 4 weeks, normotensives with DBP of 90-95mmHg continued the same dosage for 4 more weeks; the others took double dosages (data will not be used in this review) 81 patients, aged 24 to 70, with supine diastolic BP of 100-120 after placebo period. Mean age: 52 years. 58 (72%) men. 3 treatment groups (od): Monotherapy: Enalapril 20mg (E20); Hydrochlorothiazide 12.5 (H12.5) Combination: E20/H12.5 Trough supine SBP and DBP 69/81 patients completed the 8-week study. Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients...were randomly distributed not described double-blind
Yes
all patients were included in the efcacy analysis at week 2. There were 3 and 2 dropouts in the combination and enalapril group at week 8. It was assumed that all patients were included in the efcacy analysis at week 4. Comment: the small percentage of dropout is not likely to bias the result.
Yes
70
Safar 1973 Methods 45 days double-blind study after a 15 days placebo period. Randomized (methods not described). 30 men with DBP >=100mmHg. Mean age: 40years. 2 treatment groups: Monotherapy: Pindolol 5mg TID Combination: Pindolol 5mg TID plus Clopamide 5mg BID Standing SBP and DBP (timing of measurement not reported); WDAE; HR Language: French Funding: Not reported
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Authors judgement Unclear Unclear Yes Description randomized not described double-blind; The presentation of the medicaments is identical for all the patients and throughout the essay The asks and the labels always are identical all patients included in efcacy analysis.
Incomplete outcome data addressed? All outcomes Free of selective reporting?
Yes
Yes
Safar 1994 Methods 8-week double-blind study after 4-week single-blind placebo run-in period. (Randomized: methods not described). 465 patients, 19-72yrs, with supine DBP between 95 and 114mmHg. 6 treatment groups (od) Placebo Monotherapy: Perindopril 4mg (P4); Indapamide 1.25mg (I1.25) Combination: P4/I0.625; P4/I1.25; P4/I2.5 Supine DBP (timing of measurement not reported); WDAEs (number of WDAE in each group not reported)
Outcomes
71
Safar 1994
(Continued)
Notes
Source: Poster In total, 25 patients discontinued from the study (14 due to AEs) Safety data: ITT analysis. Funding: Not reported
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients...were randomized not described double-blind treatment
Yes
efcacy results were based on data obtained from completers Comment: 25 (5%) patients discontinued. Risk of bias was judged to be low.
Yes
Saruta 2007 Methods 8-week double-blind, parallel group study after a 4 to 6 weeks of placebo run-in period. Randomized (methods not described). 961 Japanese patients between 25 and 74 years of age with mean trough sitting DBP of 95 to 115mmHg and SBP <210mmHg at each visit throughout the placebo run-in period up to the day of randomization. 6 treatment arms (o.d.): Placebo Monotherapy: Losartan 50mg (L50); hydrochlorothiazide 12.5mg (H12.5) Combinattion: L25/H6.25; L50/H6.25; L50/H12.5 Trough sitting SBP and DBP and HR; WDAE Safety analyses: patients who had taken the study medication at least once (n=954) Funding: Banyu Pharmaceutical Co. Ltd.
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Unclear Description randomized
72
Saruta 2007
(Continued)
Unclear Unclear
not described double-blind
Yes
Efcacy analyses in 942 patients who took at at least one dose of study medication and have data of trough DBP postrandomization The last measurements of the withdrawn patients in the double-blind period were carried forward to subsequent time points Seven patients in whom previous medication for essential hypertension was discontinued or tapered before obtaining written consent were excluded from all analyses. Eleven patients who did not meet the major entry criteria and one patient who lacked any postrandomization data were excluded week 8: 4/317 patients in combination group and 3/160 patient in monotherapy group missing in efcacy analysis Comment: Low percentages (1-2%) of patients missing in efcacy analysis.
Yes
Saul 1995 Methods 8-week double-blind treatment following a 4-week, single-blind, placebo run-in period. Randomized (methods not described). 256 male or female patients, aged 18 to 80 yeasrs, with sitting DBP of 100-114mmHg (inclusive) during placebo run-in. 142 (55%) men. Mean age: 57.6yrs. 3 treatment groups (od): Monotherapy: Lisinopril 10mg (L10); HCTZ 25mg (H25) Combination: L10/H25 Ratio of 2:1:1 for L/H:L:H Trough (22-26 hours post-dose) sitting SBP and DBP; HR (no quantitative data); WDAE Only small mean changes in heart rate from baseline and there were no signicant treatment differences Funding: Not reported
Participants
Interventions
Outcomes
Notes
73
Saul 1995
(Continued)
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomised to treatment not described double-blind
Yes
Efcacy analysis using a completed patients analysis which included all randomised patients who completed the study week 8: 5/121 patients in combination group withdrew from study due to adverse events. 1/70 from monotherapy group withdrew due to poor compliance. Comment: Percentage of missing patients was low. Risk of bias was judged to be low.
Yes
Schoenberger 1986 Methods 16-week double-blind multicenter study after a 4-6week placebo lead-in period. Randomized (methods not described). Captorpil doses were double at the end of rst 4 weeks of active treatment. Only data up to 4 weeks will be used in this review. 382 caucasion or hispanic patients, aged >=18 yrs, with DBP 92-110mmHg on the last two visits of placebo period. Approximately 65% male. Mean age: 52 yrs. 4 treatment groups: Placebo BID Monotherapy: Captopril 50mg (C50) o.d.; C50 BID Combination: C50 o.d. plus HCTZ 25mg o.d. Trough (24+-3 hours post dose) sitting SBP and DBP Funding: Not reported
Participants
Interventions
Outcomes Notes Risk of bias Item
Authors judgement
Description
74
Schoenberger 1986
(Continued)
Adequate sequence generation?
Unclear
patients were allocated in a double-blind randomized fashion not described patients were allocated in a double-blind randomized fashion Efcacy analysis was performed in 358/382 patients. Comment: Percentage of missing patients was low. Risk of bias was judged to be low.
Unclear Unclear
Yes
Yes
Thijs 1995 Methods 4-week double-blind multicenter trial after a 2-4 week single-blind placebo run-in phase. Randomized (methods not described). Treatment was discontinued if DBP <80 or >115mmHg. 611 men and women aged 21-70 years with supine DBP averaged 100-114mmHg. Supine SBP <220mmHg. 51% men. Mean age: 55 years. 3 parallel groups (o.d.): Monotherapy: Ramipril 5mg (R5); Piretanide 6mg (P6) Combination: R5/P6 Trough (24 hours post-dose) standing BP; WDAE One patient died of myocardial infarction in the combination group Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description stratied per center and randomized not mentioned double-blind
Yes
per-protocol analysis in the 583 patients who completed the 1-month follow-up week4: 8/201 (1 due to AE, 2 not fullling inclusion criteria, 1 due to doctor with75
Thijs 1995
(Continued)
drawing collaboration, 3 bad compliance) patients in combination group and 11/209 (2 due to AE, 3 nonresponse, 1 not fullling inclusion criteria, 1 due to doctor withdrawing collboration, 4 bad compliance) patients in monotherapy group withdrew from study. Comment: 3-5% patients missing in the groups. Risk of bias was judged to be low. Free of selective reporting? Yes
Vaicaitis 1980 Methods 8-week multiclinic, double-blind study after 6 week baseline phase with placebo; randomized (method not described) 27 outpatients 21 to 65 years of age with supine diastolic BP between 100 and 120 mm Hg at end of baseline phase. Difference between week 2 or 4 and week 6 <=10mm Hg. Sitting systolic BP <160 mm Hg. Heart rate>=56beats/min. Mean age: 54yrs. 17 (63%) male. 3 treatment groups (one tablet BID): Monotherapy: Timolol maleate 10mg (T10); HCTZ 25mg (H25) Combination: T10/H25 Erect SBP and DBP (timing of measurement not reported); supine pulse rate; WDAE 25/27 patients completed the study Funding: Merck Sharp & Dohme
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Yes Description assigned at random not described all medications were given as identical-appearing tablets; double-blind Efcacy analysis not described (n=22) 1,3,3 out of 10 patients in combination group missing in efcacy analysis at week 4,6,8 respectively 3,3,2 out of 8 patients in monotherapy group missing in efcacy analysis at week
76
No
Vaicaitis 1980
(Continued)
4,6,8 respectively Comment: The percentage of missing patient in each group is high. The reasons for exclusion were not mentioned. The risk of bias was judged to be high. Free of selective reporting? Yes
Villamil 2007 Methods 8-week, multicenter, double-blind, multifactorial study trial after a 2-week single-blind placebo run-in period. Randomized (methods not described) 2776 patients aged 18 years or older with mean sitting DBP >=95mmHg after placebo run-in. patients with mean sitting DBP>=110mmHg and/or SBP>=180mmHg were excluded . Mean age: 55yrs. 55% male. 85% Caucasian; 5% blacks. 15 treatment groups (o.d.): placebo Monotherapy: Aliskiren (A75, A150 or A300mg); Hydrochlorothiazide (H6.25, H12.5 or H25mg) Combination: A75/H6.25, A75/H12.5, A75/H25, A150/H6.25, A150/H12.5, A150/H25, A300/H12.5, A300/H25 Trough sitting SBP and DBP; WDAE Safety analyses: on all patients who received at least one dose of double-blind study medication = 2762 patients Funding: Not reported
Participants
Interventions
Outcomes Notes
Yes
ITT: all randomized patients with a baseline measurement and at least one postbaseline measurement; ITT consists of 2752/2776 patients. 8week: 12/1471 patients in combination group and 6/552 patients in monotherapy group missing in efcacy analysis.
77
Villamil 2007
(Continued)
Comment: Percentage of missing patients was low. Risk of bias was judged to be low. Free of selective reporting? Yes
von Manteuffel 1995 Methods 6-week double-blind parallel group study after a 4 week placebo washout period. Block randomization was used. 173 patients with DBP of at least 100mgHg and not more than 114mmHg after placebo run-in. 4 treatment groups (od): Placebo Monotherapy: Verapamil SR 240mg od (V240); Hydrochlorothiazide 12.5mg od (H12.5) Combination: V240/H12.5 Seating SBP and DBP (timing of measurement not reported) Language: German Funding: Not reported
Participants
Interventions
Outcomes Notes
Risk of bias Item Adequate sequence generation? Authors judgement Yes Description block randomization with edge balance after 8 patients medication delivered respectively at a test place correspond to the sequentially numbered medication Comment: identical appearance? double-blind
Unclear
Unclear
Yes
2 patients each missing in efcacy analysis in the combination and monotherapy group at week 6.
Yes
78
Weinberger 1982 Methods 6-week double-blind multicentre trial after a 2-week placebo treatment. Randomized (methods not described). 207/255 patients with supine DBP of >=92mmHg and <110mmHg on 2 biweekly visits of placebo treatment. 58% men. Mean age: 49 yrs. 3 treatment groups: Monotherapy: Captopril 25mg TID (C25); HCTZ 15mg TID (H15) Combination: C25/H15 TID Supine SBP and DBP (examined 3-8 hours after the last dose of drug); HR (no extractable data for analysis); WDAE 198/207 patients completed the study. Heart rate, both supine and standing, increased slightly (1.4-4.7%) but signicantly in both groups receiving H but not the group receiving C alone, where slight (-1.7%0.7%) but non-signicant decreases in heart rate were observed Funding: Not reported
Participants
Interventions
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomly assigned not described a double-blind design was used
Yes
6weeks: 4/66 patients in combination and 2/71 patients in monotherapy group were exlcuded in efcacy analysis. (reasons: did not complete study). 3 and 2 patients in the groups were withdrew due to adverse events respectively. Comment: not likely bias the results
Yes
Weir 1992 Methods 12-week double-blind, placebo controlled, multicentre study after a 4-6 week singleblind placebo run-in period. The 12 week period consists of three 4-week evaluation period with increasing xed dose at each period. Randomized study (methods not described). Data on the rst 4-week evaluation period will be used in this review. 298 volunteers between 18 and 70 yrs of age with supine DBP >=95 and <=110 mmHg.
79
Participants
Weir 1992
(Continued)
Interventions
4 treatment groups (1st evaluation period) Placebo Monotherapy: Diltiazem SR 60mg BID (D60); HCTZ 6.25mg BID (H6.25) Combination: D60/H6.25 BID Trough (12+-2hrs post dose) supine SBP and DBP; supine HR (no quantitative data); WDAE Efcacy analysis: patients who completed a 4-week evaluation period without protocol violations (used) with n = 274 for period 1 (65%male; 86%non-black; mean age = 53.5 yrs) 14 patients discontinued period 1. No signicant differences between groups in HR in period 1 Funding: Marion Merrell Dow Inc.
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description patients were randomised not described double-blind study
Unclear
The primary analysis (efcacy) used data sets for each evaluation period which included measurements only from patients who completed that particular evaluation period without a protocol violation. This included 274/298 patients. The number of patients missing in each treatment group was not reported.
Yes
Yodfat 1994 Methods 4-week parallel-group placebo-controlled multicentre study after 4-week single-blind placebo period. Randomized (methods not described). Efcacy data up to 4 weeks of treatment were recorded. 377 patients of both sexes, between 20 and 68 yrs, with average sitting DBP >100mmHg and those who demonstrated at least 80% compliance after placebo period. 244 (65%) men. Mean age: 53 yrs.
Participants
80
Yodfat 1994
(Continued)
Interventions
8 treatment groups (o.d.): Placebo Monotherapy: Cilazapril 2.5mg (C2.5); C5; HCTZ 12.5mg (H12.5); H25 Combination: C1.25/H6.25; C2.5/H12.5; C5/H25 Trough (22-24 hours post dose) sitting DBP; WDAE (number of WDAE for each group not reported) 363/377 patients completed the study Assume N for trough BP analysis is same as N in Table 2 which report n patients with normalized BP. (N=373) (see page 120) Non-signicant changes from baseline in mean sitting and standing, trough and peak pulse rates, were found in all treatment groups when compared to placebo Funding: Not reported
Outcomes
Notes
Risk of bias Item Adequate sequence generation? Allocation concealment? Blinding? All outcomes Incomplete outcome data addressed? All outcomes Authors judgement Unclear Unclear Unclear Description randomly assigned not described double-blind
Unclear
number of total patients in efcacy analysis is not explicitly reported. Assume the total number of patients for trough BP analysis is same as those for DBP normalization. (373/377 patients) (see page 120)
Yes
Characteristics of excluded studies [ordered by study ID]
Agrawal 1979
Crossover trial with no pre-crossover data for rst 4 or 6 weeks of treatment (Propranolol 80mg BID vs bendrouazide 2.5mg BID or their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (propanolol 80mg od vs bendrouazide 2.5mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Nifedipine 20mg BID vs Atenolol 50mg od vs their combination)
81
Agrawal 1987
Anderton 1988
(Continued)
Bainbridge 1993
Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Felodipine 5mg od vs ramipril 2.5mg od vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (atenolol 100mg od vs chlorthalidone 25mg od vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment. 8 week treatment but titration in non-responders after 4 week treat ment. (Enalapril 10mg BID vs Hydrochlorothiazide 25mg BID vs their combination) Chlorthalidone given 3 times per week. Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 100mg od vs chlorthalidone 25mg od vs their combination vs placebo) Parallel group trial with 8-week treatment period. Hydrocholrothiazide was co-administered with amiloride. ( Methyldopa 250mg (M250) vs Hydrochlorothiazide 25mg plus Amiloride 5mg (H25/A5) vs their combination) Crossover trial with BP data for the rst 4 weeks of treatment. Number of patients per treatment arm for the rst phase of study not mentioned. (Atenolol 100mg od vs Chlorthalidone 50mg od vs their combination) Parallel group trial with 8-week treatment period. Number of patients per treatment arm not reported (16 parallel arms: placebo; Quinapril 2.5, 10, 40mm/day; Hydrochlorothiazide 6.25, 12.5, 25mg/day; all possible combinations). Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Amlodipine 5mg od vs Lisinopril 10mg od vs their combination) Crossover trial with no pre-crossover data for rst 8 weeks of treatment. (Indapamide 2.5mg od vs Pindolol 10mg od vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Enalapril 10mg BID vs Hydrochlorothiazide 25mg BID vs their combination vs placebo) Parallel group trial with 4-week treatment period. Hydrochlorothiazide was co-administered with triamterene. (Atenolol 25mg vs HCTZ 25mg plus triamterene 50mg vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Bendrouazide 5mg od vs propanolol 160mg od vs their combination) Crossover trial with no data for rst 6 weeks of treatment (12-week treatment but dose doubled in non-responders after 6 weeks). (Felodipine 5mg od vs Metoloprolol 50mg od vs their combination). Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Verapamil 120mg TID vs propranolol 80mg TID vs their combination vs placebo)
Bateman 1979
Bauer 1984
Bertrand 1982 Boike1982
Bolzano 1984
Cajochen 1984
Canter 1994
Cappuccio 1993
Chalmers 1982
Chalmers 1986
Chrysant 1992
Crowe 1987
Dahlof 1993
Dargie 1986
82
(Continued)
De Divitiis 1981
Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 100mg od vs chlorthalidone 50mg od vs their combination vs chlorthalidone 50mg od plus reserpine 0.25mg od) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (indapamide 2.5mg od vs atenolol 100mg vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 100mg od vs Chlorthalidone 50mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (chlorthalidone 25mg od vs metoprolol 200mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (alpha-methyldopa 750mg/day vs chlorothiazide 340mg/day vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 8 weeks of treatment (acebutolol 400mg/day vs clorexolone 3mg/day+canrenone 25mg/day vs combo) Parallel group trial with 48 week treatment period, titration based on response starting at week 4. Pre-titration data not reported. (enalapril 10mg BID vs hydrochlorothiazide 25mg BID vs combination) Parallel group trial with 4-week treatment period. Hydrochlorothiazide was co-administered with triamterene. (Atenolol 100mg vs HCTZ 25mg plus Triamterene 50mg vs their combination) Crossover trial with no pre-crossover data for rst 12 weeks of treatment (captopril 50mg BID vs bendrouazide 2.5mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Propanolol 80mg BID vs bendrouazide 2.5mg BID vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 50mg vs hydrochlorothiazide 25mg/amiloride 2.5mg od and their combination) Crossover trial with pre-crossover data for the rst 6 weeks of treatment. However, hydrochlorothiazide was coadministered with amiloride. (Acebutolol 400mg vs Acebutolol 400mg plus HCTZ50mg plus Amiloride 5mg) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Metoprolol 200mg od vs Metoprolol 200mg/chlorthalidone 25mg od) Crossover trial with no pre-crossover data for rst 8 weeks of treatment (ramipril 10mg od vs candesartan 16mg od vs their combination) Crossover trial with pre-crossover data for rst 8 weeks of treatment. However, number of patients in each group not reported. [Metoprolol 100mg BID (M100) vs M100 BID plus 2 x chlorthalidone-K 25mg od ( Chlorthalidone-K tablets each containing 6.7mmol potassium)]
De Divitiis 1983
Durel 1992
Erwteman 1984
Fernandez 1980
Forette 1979
Frishman 1987
Hart 1985
Hunter 1999
Jaattela 1979
Jackson 1986
Khalil 1982
Kieso 1983
Koh 2007
Kubik 1979
83
(Continued)
Lang 1991
Parallel group with 8 week treatment . Number of patients per treatment arm for analysis not included. (Lisinopril 10mg od vs hydrochlorothiazide 12.5mg od vs their combination) Crossover trial with no pre-crossover data for rst 3 weeks of treatment (Labetalol 300mg od vs chlorthalidone 30mg od vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment. (Atenolol 50mg od vs Chlorthalidone 12.5mg od vs their combination) Crossover trial with no pre-crossover data for rst 3 weeks of treatment (Nadolol 80mg od and the combination of Nadolol 80mg od with hydrochlorothiazide 12.5, 25 or 50mg od) Parallel group trial with 24 week treatment period, titration based on response starting at week 4. Pre-titration data not reported (Lisinopril 20mg od vs Lisinopril 20mg/Hydrochlorothiazide 12.5mg od) Parallel group trial with 8 week treatment period. Only ambulatory blood pressure measurement provided. No clinic blood pressure recorded. (Enalapril 20mg od vs enalapril 20mg plus hydrochlorothiazide 12.5mg od) Parallel group treatment with 4 week treatment period. Hydrochlorothiazde was co-administered with amiloride. (Methyldopa 250mg BID vs HCTZ 25mg/Amiloride 2.5mg BID vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment. 6 week treatment but dose doubled in monotherapy group after 4 weeks. (Felodipine 5mg od vs Enalapril 5mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment. (felodipine 5mg od vs candesartan 16mg od vs their combination vs placebo) Parallel group treatment with 6 week treatment period. Hydrochlorothiazide was co-administered with amiloride. (Timolol 10mg BID vs HCTZ/Amiloride 25/2.5mg BID vs their combination) Parallel group with 8 week treatment period. Only ambulatory blood pressure data provided (Amlodipine 5mg od vs benazepril 20mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 100 BID vs bendrouazide 2.5mg BID vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (atenolol 100mg od vs chlorthalidone 25mg plus atenolol 100mg od) Parallel group trial after a 3 week treatment period. Hydrochlorothiazide was co-administered with triamterene. [Verapamil 80mg or 160mg vs Trimaterene 25mg/HCTZ 12.5mg vs Triamterene 50mg/HCTZ 25mg vs all possible combinations) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Atenolol 100mg od vs Chlorthalidone 25mg od vs their combination)
Lechi 1982
Leonetti 1986
Magee 1986
Mehta 1988
Middlemost 1994
Moncloa 1980
Morgan 1992
Morgan 2002
Muiesan 1976
Neutel 2006
Petrie 1975
Ricciardelli 1985
Rosenthal 1989
Salako 1990
84
(Continued)
Salvetti 1987
Crossover trial with no pre-crossover data for rst 4 weeks of treatment (captopril 50mg BID vs nifedipine 20mg BID vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Nifedipine 20mg BID vs chlorthalidone 25mg od vs their combination) Parallel group trial with 6 weeks treatment period. Number of patient per treatment arm is not reported (placebo vs ramipril 2.5, 5 or10mg vs HCTZ 12.5 or 25mg vs all possible combinations) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Nifediipine 20mg BID vs acebutolol 200mg BID vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Diltiazem 120mg BID vs Atenolol 50mg od vs their combination vs placebo) Parallel group trial with 6 months treatment. No data for 3-12 weeks treatment period. (trandolapril 2mg od vs verapamil 180mg plus trandolapril 2mg od) Parallel group trial with 6 months treatment, titration based on response starting at week 4. Pre-titration data not reported. (Propanolol 40mg TID vs Propanolol in combination with hydrochlorothiazide 35mg or hydralazine 35mg TID vs the combination of 3 drugs) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Propanolol 80mg BID vs Bendrouazide 2.5mg BID vs their combination) Parallel group with 8 weeks treatment. Only provided ambulatory BP measurements. (verapamil 180mg od vs trandolapril 2mg od vs their combination) Study B: Parallel group with 12 week treatment period. BP measurements at 3-8hrs after the last drug dose. No properly dened peak or trough BP measurement. (captopril 52mg BID vs hydrochlorothiazide 25mg BID vs their combination vs placebo) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Enalapril 20mg od vs Atenolol 50mg od vs their combination) Crossover trial with no pre-crossover data for rst 4 weeks of treatment (Felodipine 5mg BID vs metoprolol 100mg BID vs their combination)
Salvetti 1989
Scholze 1993
Singer 1987
Tonkin 1990
Topouchian 1999
VACSGAA 1977
Van Staden 1983
Viskoper 1997
Weinberger 1983
Wing 1988
Wing 1994
85
DATA AND ANALYSES
Comparison 1. Combination therapy (with thiazide 0.5x start dose) vs monotherapy (without thiazide)
Outcome or subgroup title 1 Change in SBP 1.1 ACEI + HCTZ 5mg vs ACEI 1.2 ACEI + HCTZ 6.25mg vs ACEI 1.3 ARB + HCTZ 6.25mg vs ARB 1.4 BB + HCTZ 6.25mg vs BB 1.5 RI + HCTZ 6.25mg vs RI 2 Change in DBP 2.1 ACEI + HCTZ 5mg vs ACEI 2.2 ACEI + HCTZ 6.25mg vs ACEI 2.3 ARB + HCTZ 6.25mg vs ARB 2.4 BB + HCTZ 6.25mg vs BB 2.5 RI + HCTZ 6.25mg vs RI 2.6 ACEI + Indapamide 0.625mg vs ACEI
No. of studies 10 1 1 4 3 1 12 1 2 4 3 1 1
No. of participants 3283 179 83 1130 1165 726 3509 179 164 1130 1165 726 145
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size -3.70 [-4.60, -2.80] -4.77 [-9.37, -0.16] -4.9 [-10.75, 0.95] -3.35 [-4.93, -1.77] -3.61 [-5.07, -2.14] -4.02 [-5.88, -2.16] -1.67 [-2.19, -1.15] -0.80 [-3.20, 1.61] -1.52 [-3.93, 0.88] -1.62 [-2.54, -0.71] -1.82 [-2.71, -0.94] -1.80 [-2.96, -0.63] -1.30 [-4.20, 1.60]
Comparison 2. Combination therapy (with thiazide 1x start dose) vs monotherapy (without thiazide)
Outcome or subgroup title 1 Change in SBP 1.1 ACEI + HCTZ 12.5mg vs ACEI 1.2 ARB + HCTZ 12.5mg vs ARB 1.3 BB + HCTZ 12.5mg vs BB 1.4 CCB + HCTZ 12.5mg vs CCB 1.5 RI + HCTZ 12.5mg vs RI 2 Change in DBP
No. of studies 32 14 12 2 3 1 35
No. of participants 8482 2326 3930 873 254 1099 8804
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size -5.97 [-6.54, -5.41] -5.18 [-6.27, -4.10] -7.12 [-7.95, -6.28] -4.66 [-6.45, -2.87] -3.67 [-6.72, -0.62] -5.24 [-6.76, -3.73] -3.09 [-3.42, -2.75]
86
2.1 ACEI + HCTZ 12.5mg vs ACEI 2.2 ARB + HCTZ 12.5mg vs ARB 2.3 BB + HCTZ 12.5mg vs BB 2.4 CCB + HCTZ 12.5mg vs CCB 2.5 RI + HCTZ 12.5mg vs RI 2.6 ACEI + Indapamide 1.25mg vs ACEI
16 12 2 3 1 1
2503 3930 873 254 1099 145
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
-3.12 [-3.73, -2.51] -3.31 [-3.81, -2.81] -2.28 [-3.39, -1.17] -2.58 [-4.49, -0.67] -1.00 [-3.95, -2.05] -2.60 [-5.50, 0.30]
Outcome or subgroup title 1 Change in SBP 1.1 ACEI + HCTZ 25mg vs ACEI 1.2 ARB + HCTZ 25mg vs ARB 1.3 BB + HCTZ 25mg vs BB 1.4 CCB + HCTZ 25mg vs CCB 1.5 RI + HCTZ 25mg vs RI 1.6 CAD + HCTZ 25mg vs CAD 1.7 BB + Chlorthalidone 20mg vs BB 1.8 BB + Clopamide 10mg vs BB 1.9 CCB + Indapamide 2.5mg vs CCB 2 Change in DBP 2.1 ACEI + HCTZ 25mg vs ACEI 2.2 ARB + HCTZ 25mg vs ARB 2.3 BB + HCTZ 25mg vs BB 2.4 CCB + HCTZ 25mg vs CCB 2.5 RI + HCTZ 25mg vs RI 2.6 CAD + HCTZ 25mg vs CAD 2.7 BB + Chlorthalidone 20mg vs BB 2.8 BB + Clopamide 10mg vs BB
No. of studies 25 7 8 4 1 1 1 1 1 1 29 10 8 4 1 1 1 1 1
No. of participants 6079 815 3003 680 130 1092 79 49 30 201 6572 1169 3003 680 130 1092 79 49 30
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size -8.01 [-8.69, -7.34] -7.53 [-9.40, -5.66] -8.43 [-9.41, -7.46] -8.60 [-10.68, -6.52] -7.4 [-11.56, -3.24] -6.92 [-8.44, -5.40] -7.0 [-12.93, -1.07] -6.0 [-13.50, 1.50] -0.40 [-9.99, 9.19] -11.22 [-15.09, 7.35] -4.05 [-4.45, -3.65] -3.85 [-4.78, -2.92] -4.17 [-4.76, -3.57] -4.01 [-5.30, -2.72] -5.9 [-8.39, -3.41] -3.53 [-4.49, -2.58] -4.0 [-7.76, -0.24] -4.0 [-8.59, 0.59] -4.1 [-9.97, 1.77]
87
2.9 ACEI + Indapamide 1.25mg vs ACEI 2.10 CCB + Indapamide 2.5mg vs CCB
1 1
139 201
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
-3.20 [-6.24, -0.16] -5.83 [-8.20, -3.46]
Outcome or subgroup title 1 Change in SBP 1.1 ACEI + HCTZ 37.5mg vs ACEI 1.2 ACEI + HCTZ 45mg vs ACEI 1.3 BB + HCTZ 50mg vs BB 2 Change in DBP 2.1 ACEI + HCTZ 37.5mg vs ACEI 2.2 ACEI + HCTZ 15mg vs ACEI 2.3 BB + HCTZ 50mg vs BB
No. of studies 4 1 1 2 4 1 1 2
No. of participants 367 178 131 58 367 178 131 58
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size -14.20 [-17.15, 11.25] -9.01 [-13.63, -4.40] -18.5 [-23.09, 13.91] -16.12 [-23.08, 9.17] -4.00 [-7.69, -4.31] -3.51 [-5.94, -1.09] -9.20 [-12.01, -6.39] -6.31 [-10.56, -2.05]
Comparison 5. Combination therapy (with thiazide - all doses) vs monotherapy (without thiazide)
Outcome or subgroup title 1 Change in SBP 1.1 HCTZ 6.25mg/day 1.2 HCTZ 12.5mg/day 1.3 HCTZ 25mg/day 1.4 HCTZ 37.5mg/day 1.5 HCTZ 45-50mg/day 1.6 Indapamide 2.5 mg/day 1.7 Chlorthalidone 20mg/day 1.8 Clopamide 10mg/day 2 Change in DBP 2.1 HCTZ 6.25mg/day 2.2 HCTZ 12.5mg/day 2.3 HCTZ 25mg/day 2.4 HCTZ 37.5mg/day 2.5 HCTZ 45-50mg/day
No. of studies 49 10 32 22 1 3 1 1 1 53 11 34 25 1 3
No. of participants
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size Subtotals only -3.70 [-4.60, -2.80] -5.97 [-6.54, -5.41] -7.97 [-8.66, -7.28] -9.01 [-13.63, -4.40] -17.78 [-21.61, 13.95] -11.22 [-15.09, 7.35] -6.0 [-13.50, 1.50] -0.40 [-9.99, 9.19] Subtotals only -1.69 [-2.22, -1.16] -3.09 [-3.43, -2.76] -4.02 [-4.43, -3.60] -3.51 [-5.94, -1.09] -8.32 [-10.68, -5.96]
88
3283 8482 5799 178 189 201 49 30 3364 8659 6153 178 189
2.6 Indapamide 0.625 mg/day 2.7 Indapamide 1.25 mg/day 2.8 Indapamide 2.5 mg/day 2.9 Chlorthalidone 20mg/day 2.10 Clopamide 10mg/day 3 Change in Heart rate 3.1 Bisoprolol/HCTZ vs Bisoprolol 3.2 Timolol/HCTZ vs Timolol 3.3 Metoprolol/HCTZ vs Metoprolol 3.4 Diltiazem/HCTZ vs Diltiazem 3.5 Oxprenolol/ Chlorthalidone vs Oxprenolol 3.6 Pindolol/Clopamide vs Pindolol 4 Withdrawals due to adverse events 4.1 Benazepril/HCTZ vs Benazepril 4.2 Captopril/HCTZ vs Captopril 4.3 Cilazapril/HCTZ vs Cilazapril 4.4 Fosinopril/HCTZ vs Fosinopril 4.5 Lisinopril/HCTZ vs Lisinopril 4.6 Moexipril/HCTZ vs Moexipril 4.7 Perindopril/HCTZ vs Perindopril 4.8 Ramipril/HCTZ vs Ramipril 4.9 Quinapril/HCTZ vs Quinapril 4.10 Candesartan/HCTZ vs Candesartan 4.11 Irbesatan/HCTZ vs Irbesartan 4.12 Losartan/HCTZ vs Losartan 4.13 Telmisartan/HCTZ vs Telmisartan 4.14 Valsartan/HCTZ vs Valsartan 4.15 Aliskiren/HCTZ vs Aliskiren
1 1 2 1 1 6 1 1 1 1 1 1 35 1 3 1 2 4 1 1 1 2 1 2 3 1 2 1
145 145 340 49 30 483 301 11 52 40 49 30 9258 128 361 69 453 703 275 20 440 458 136 950 1077 623 892 2005
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI)
-1.30 [-4.20, 1.60] -2.60 [-5.50, 0.30] -4.83 [-6.70, -2.96] -4.0 [-8.59, 0.59] -4.1 [-9.97, 1.77] 0.43 [-1.10, 1.97] 0.5 [-1.25, 2.25] -3.40 [-14.18, 7.38] -2.70 [-9.56, 4.16] 2.70 [-2.36, 7.76] -1.00 [-9.66, 3.66] 7.80 [-2.68, 18.28] 1.09 [0.84, 1.42] 4.45 [0.25, 80.75] 1.83 [0.63, 5.32] 0.82 [0.15, 4.59] 1.10 [0.37, 3.33] 1.01 [0.45, 2.29] 2.01 [0.51, 7.89] Not estimable 0.20 [0.02, 1.67] 0.15 [0.03, 0.85] 4.5 [0.52, 39.23] 2.70 [1.03, 7.07] 1.06 [0.45, 2.48] 1.01 [0.38, 2.65] 0.50 [0.24, 1.06] 1.50 [0.73, 3.09]
89
4.16 Metoprolol/HCTZ vs Metoprolol 4.17 Nebivolol/HCTZ vs Nebivolol 4.18 Timolol/HCTZ vs Timolol 4.19 Diltiazem/HCTZ vs Diltiazem 4.20 Moxonidine/HCTZ vs Moxonidine 4.21 Oxprenolol/ Chlorthalidone vs Oxprenolol 4.22 Pindolol/Clopamide vs Pindolol
1 1 2 2 1 1 1
53 180 85 191 79 50 30
Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI) Risk Ratio (M-H, Fixed, 95% CI)
0.35 [0.01, 8.12] 0.17 [0.01, 4.06] 2.12 [0.33, 13.71] 1.63 [0.22, 11.98] 0.44 [0.04, 4.66] 3.0 [0.13, 70.30] Not estimable
Comparison 6. Combination therapy (with loop diuretic - all dose) vs monotherapy (without loop diuretic)
Outcome or subgroup title 1 Change in SBP 1.1 0.5x start dose 1.2 1x start dose 1.3 2x start dose 2 Change in DBP 2.1 0.5x start dose 2.2 1x start dose 2.3 2x start dose 3 Withdrawals due to adverse events 3.1 Ramipril/Pirentanide vs Ramipril
No. of studies 3 1 2 1 3 1 2 1 2 2
No. of participants
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI)
Effect size Subtotals only -1.77 [-5.16, 1.62] -6.47 [-8.70, -4.24] -13.0 [-23.32, -2.68] Subtotals only -0.40 [-2.48, 1.68] -3.09 [-4.53, -1.66] -8.0 [-14.31, -1.69] Subtotals only 0.94 [0.25, 3.50]
240 629 27 240 629 27
770
Comparison 7. Direct Comparison
Outcome or subgroup title 1 Change in SBP 1.1 HCTZ 25mg/day + ACEI vs HCTZ 12.5mg/day + ACEI 1.2 HCTZ 25mg/day + ARB vs HCTZ 12.5mg/day + ARB 1.3 HCTZ 25mg/day + BB vs HCTZ 12.5mg/day + BB 2 Change in DBP
No. of studies 11 2 7 2 12
No. of participants
Statistical method Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size Subtotals only -1.57 [-4.30, 1.16] -1.55 [-2.59, -0.51] -1.88 [-4.59, 0.83] Subtotals only
90
303 2478 398
2.1 HCTZ 25mg/day + ACEI vs HCTZ 12.5mg/day + ACEI 2.2 HCTZ 25mg/day + ARB vs HCTZ 12.5mg/day + ARB 2.3 HCTZ 25mg/day + BB vs HCTZ 12.5mg/day + BB
3 7 2
393 2478 398
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
-1.06 [-2.42, 0.31] -1.18 [-1.81, -0.54] -1.90 [-3.57, -0.23]
HISTORY
Protocol rst published: Issue 2, 2008 Review rst published: Issue 4, 2009
CONTRIBUTIONS OF AUTHORS
James Wright formulated the idea for the review and developed the basis for the protocol with Jenny Chen. Jenny Chen took the lead roles in searching, identifying and assessing studies, in data extraction and analyses, and in writing up the review. Balraj Heran assisted with data extraction and the preparation of the manuscript.
DECLARATIONS OF INTEREST
None
SOURCES OF SUPPORT Internal sources

Department of Anesthesiology, Pharmacology and Therapeutics, University of British Columbia, Canada.
External sources
Canadian Institutes of Health Research (CIHR), Canada.
INDEX TERMS
91
Medical Subject Headings (MeSH)

Antihypertensive Agents [ therapeutic use]; Blood Pressure [drug effects]; Diuretics [ therapeutic use]; Drug Therapy, Combination [methods]; Heart Rate [drug effects]; Hypertension [ drug therapy]; Randomized Controlled Trials as Topic; Sodium Chloride Symporter Inhibitors [ therapeutic use]; Sodium Potassium Chloride Symporter Inhibitors [ therapeutic use]
MeSH check words

Humans
92

CD007187 Standard

Cargado por

Información del documento

Descripción original:

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

CD007187 Standard

Cargado por

Copyright:

Formatos disponibles

Blood pressure lowering efcacy of diuretics as second-line therapy for primary hypertension (Review)

Chen JMH, Heran BS, Wright JM

Types of outcome measures

Criteria for considering studies for this review

Search methods for identication of studies

Data collection and analysis

Starting dose/day for hypertension 40-80mg/day 6-12mg/day 12.5mg/day 1.25mg/day

Available in Canada? Yes No Yes Yes

Table 1. Starting doses of Diuretics analyzed in the review

Risk of bias in included studies

Thiazide plus other drug vs other drug alone

Loop diuretic plus other drug vs other drug alone

Table 4. Variability of SBP and DBP at baseline (thiazides)

Weighted mean SD SD of weighted mean SD

Combo group vs mono group

Loop diuretic as second-line

Loop diuretic as second-line

Loop diuretic as second-line

12.5 -6.0 (-6.5, mmHg -8.0 (-8.7, mmHg

-5.7 (-7.0, mmHg

-8.5 (-10.4, -6.6) 368 mmHg

*adopted from Musini 2002

*adopted from Musini 2002

Does age have an effect on BP lowering of diuretics?

What is the effect of second-line diuretics on pulse pressure?

Does co-morbidity have an effect on BP lowering of diuretics?

What is the effect of second-line diuretics in combination therapy on BP variability?

What is the effect of second-line diuretics on heart rate?

Implications for research

AUTHORS CONCLUSIONS Implications for practice

Findings of this review

References to studies included in this review

References to studies excluded from this review

Characteristics of included studies [ordered by study ID]

Outcomes Notes Risk of bias Item Adequate sequence generation?

Authors judgement Unclear

Allocation concealment? Blinding? All outcomes

Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Efcacy analysis in 75/76 (99%) patients.

Allocation concealment? Blinding? All outcomes

Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Risk of bias Item Authors judgement Description

Unclear Unclear Unclear

Free of selective reporting?

Outcomes Notes Risk of bias Item Adequate sequence generation?

Authors judgement Unclear

Blinding? All outcomes Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Free of selective reporting?

Blinding? All outcomes

Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Incomplete outcome data addressed? All outcomes

Free of selective reporting?

Authors judgement Unclear Unclear Unclear

Incomplete outcome data addressed? All outcomes

study at week 4 for unknown reasons Free of selective reporting? Yes

Authors judgement Unclear Unclear Unclear

Description randomized not described double-blind