|Home | About | Journals | Submit | Contact Us | Français|
A 2007 systematic review compared angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) in patients with hypertension. Direct renin inhibitors (DRIs) have since been introduced, and significant new research has been published. We sought to update and expand the 2007 review.
We searched MEDLINE and EMBASE (through December 2010) and selected other sources for relevant English-language trials.
We included studies that directly compared ACE inhibitors, ARBs, and/or DRIs in at least 20 total adults with essential hypertension; had at least 12 weeks of follow-up; and reported at least one outcome of interest. Ninety-seven (97) studies (36 new since 2007) directly comparing ACE inhibitors versus ARBs and three studies directly comparing DRIs to ACE inhibitor inhibitors or ARBs were included.
A standard protocol was used to extract data on study design, interventions, population characteristics, and outcomes; evaluate study quality; and summarize the evidence.
In spite of substantial new evidence, none of the conclusions from the 2007 review changed. The level of evidence remains high for equivalence between ACE inhibitors and ARBs for blood pressure lowering and use as single antihypertensive agents, as well as for superiority of ARBs for short-term adverse events (primarily cough). However, the new evidence was insufficient on long-term cardiovascular outcomes, quality of life, progression of renal disease, medication adherence or persistence, rates of angioedema, and differences in key patient subgroups.
Included studies were limited by follow-up duration, protocol heterogeneity, and infrequent reporting on patient subgroups.
Evidence does not support a meaningful difference between ACE inhibitors and ARBs for any outcome except medication side effects. Few, if any, of the questions that were not answered in the 2007 report have been addressed by the 36 new studies. Future research in this area should consider areas of uncertainty and be prioritized accordingly.
The online version of this article (doi:10.1007/s11606-011-1938-8) contains supplementary material, which is available to authorized users.
A 54-year-old woman with a history of hypertension is seen by her doctor for persistently elevated blood pressure in spite of adherence to hydrochlorothiazide 25 mg daily. She is overweight and has a strong family history of coronary artery disease. To control her blood pressure, she and her doctor discuss adding an angiotensin-converting enzyme (ACE) inhibitor, an angiotensin II receptor blocker (ARB), or a direct renin inhibitor (DRI) to her regimen. She is primarily interested in avoiding the cardiovascular complications of hypertension, but does not want to take medication more than once a day, and she is concerned about side effects and the cost of her medication. What information is available to help guide her decision?
Almost 75 million American adults have hypertension. Advances in antihypertensive therapy have dramatically reduced cardiovascular, cerebrovascular, and renal events.1–3 Among the effective pharmacotherapies are inhibitors of the renin-angiotensin-aldosterone (renin) system. In 2007 the Agency for Healthcare Research and Quality (AHRQ) sponsored a comparative effectiveness review of the two most common renin system inhibitors, ACE inhibitors and ARBs, to answer the following three key questions for adults with essential hypertension: Do ACE inhibitors and ARBs differ in the following: 1) blood pressure control, cardiovascular events, quality of life, and other outcomes; 2) safety, tolerability, persistence with therapy, or treatment adherence; and 3) effects within important subgroups of patients? We reported high-level evidence demonstrating that ACE inhibitors and ARBs had similar effects on blood pressure control, and that ACE inhibitors had higher rates of cough than ARBs; however, data regarding long-term cardiovascular outcomes, quality of life, progression of renal disease, medication adherence or persistence, rates of angioedema, and differences in key patient subgroups were limited.4,5
Since the 2007 review, several original research studies have directly compared ACE inhibitors and ARBs in patients with hypertension, and direct renin inhibitors (DRIs) have been introduced as a new class of medication targeting the renin system. In the present review, we sought to update the 2007 report on the comparative effectiveness of ACE inhibitors and ARBs, expand the review to include DRIs, and determine whether the conclusions of the initial review have changed in light of new evidence.
The present manuscript is derived from a new comparative effectiveness review commissioned by AHRQ. In that review, the protocol used for the 2007 report, including the three key questions listed above, was adapted to include DRIs and applied to the direct comparison literature published since the 2007 report. Further details of our methods, results, and conclusions are available in the full AHRQ report.6
To identify relevant studies, we updated and expanded (to include DRIs) the original search, conducted through May 2006, using search terms for drug interventions, hypertension, and applicable study designs. We searched MEDLINE and EMBASE (the latter not included in the original search) through December 23, 2010; the Cochrane Central Register of Controlled Trials (Issue 2, 2006); a register of systematic reviews underway in the Cochrane Hypertension Review Group (December 1, 2010); and grey literature sources (e.g. regulatory data, clinical trial registries, and conference abstracts) identified by AHRQ’s Effective Health Care Program (Appendix Table A available online).
Title and abstract screening was performed by two independent reviewers. Articles that were included by either reviewer moved forward for full-text screening. Full-text screening was also performed by two independent reviewers; however, the reviewers worked together to reconcile most differences, with any remaining disagreement adjudicated by a third reviewer.
We included all clinical studies directly comparing ACE inhibitors, ARBs, and/or DRIs in at least 20 total adults with essential hypertension, provided they had at least 12 weeks of follow-up and reported at least one outcome of interest. Our inclusion criteria were identical to those in the 2007 report,4,5 with the addition of DRIs as a potential comparator. Fixed-dose combination medications were included if the non-ACE inhibitor/ARB/DRI medication was identical across treatment arms (e.g., studies with enalapril/hydrochlorothiazide compared to losartan/hydrochlorothiazide would be included if the hydrochlorothiazide dose was the same in both treatment arms). Because of the number of direct comparison studies, we did not include indirect comparisons. Studies not conducted solely in patients with hypertension had to report subgroup results for those with hypertension.
Data were extracted using a standardized template (Appendix Table B available online). For each article, one investigator abstracted data, and a second over-read the abstraction for accuracy and completeness. Disagreements were resolved by consensus, or when necessary, by a third reviewer’s adjudication.
To assess the quality of clinical trials and cohort studies, we adapted criteria developed by the U.S. Preventive Services Task Force and the Centre for Reviews and Dissemination7,8 and categorized studies as “good,” “fair,” or “poor” in quality. We assessed the strength of the body of evidence for each key question using the approach recommended in AHRQ’s Methods Guide for Effectiveness and Comparative Effectiveness Reviews.9 This approach is conceptually similar to the GRADE framework10 used in the 2007 report. Table 1 summarizes the results of this grading for both the 2007 report and the current update.
When evaluating groups of studies reporting the same or similar outcomes for potential data synthesis, we tended to be inclusive of individual studies unless their populations were clearly dissimilar. When calculating summary effect sizes, we stratified by study design, separating randomized controlled trials (RCTs) from observational studies. We used random-effects models to allow for statistical heterogeneity and calculated the Q statistic as a measure of this. Even in the presence of statistical heterogeneity, we performed meta-analysis if studies appeared to be clinically and methodologically similar. We examined the potential for publication bias, but found no statistical evidence for this. 6 We combined dichotomous events using odds ratios (ORs) and continuous measures using differences in means. We used Comprehensive Meta-Analysis Version 2 (Comprehensive Meta-Analysis Version 2, Biostat, Englewood, NJ) for all analyses.
A total of 2090 citations were identified by the literature search (905 new since the 2007 report), of which 100 distinct studies described in 110 articles were included in the updated review (36 new ACE inhibitor versus ARB comparisons; three DRI studies; Fig. 1). The new studies contributed 176,308 additional patients. Table 2 summarizes the total number of studies, number of new studies, study design, quality ratings, and number of participants for each outcome. The specific agents compared are summarized in Appendix Table C (available online).
Two good-quality RCTs compared the DRI aliskiren at a maximum dose of 300 mg to the ACE inhibitor ramipril at a maximum dose of 10 mg.11–13 In both studies, aliskiren produced a greater reduction in blood pressure at 12 weeks, with between-group blood pressure (SBP/DBP) differences of −2.7/-1.611,12 and −2.3/-1.5 mmHg.13 These studies also reported safety, adverse events, persistence, and renal function, but did not find any differences in these outcomes. One good-quality RCT compared aliskiren up to 300 mg to the ARB losartan at a maximum dose of 100 mg and did not find any significant between-group difference for any outcome.14 Results from these studies are noted under the relevant specific outcomes below.
The updated literature search identified an additional 30 studies (three evaluating a DRI) that reported a blood pressure outcome, for a total of 80 studies.11–97 Study durations ranged from 12 weeks to 5 years, with a median of 24 weeks. Only nine studies reported enrollment of at least 10% black patients.11,12,24,47,51,61,63,74,86,90 The funding source was reported in 47 studies (59%),11–15,17–24,26,27,30,32,42–44,47,49,51–55,57–62,66,67,69,70,72–76,78–80,82–84,86–88,92,97 with the majority of these (32 studies) funded by the manufacturer of one of the study medications.
For the overall comparison of blood pressure lowering between ACE inhibitors and ARBs, 58 studies (74%) reported no difference,15,17–31,33,35,37,40,41,44–49,51,52,56–63,66,67,69,70,73,75–93,95,97 two studies (3%) favored ACE inhibitors,64,65,74 11 studies (14%) favored ARBs,16,34,36,43,50,53–55,68,72,94 and 6 studies (8%) did not report a statistical comparison between the two agents.32,38,39,42,71,96 Because of substantial heterogeneity in study protocols, quantitative meta-analysis was not performed for blood pressure lowering.
We identified seven new studies in the update (three evaluating a DRI) that reported the outcome of successful blood pressure control with a single antihypertensive agent, for a total of 29 studies.11–14,17,27,31,34,40,42,44–47,49,51–53,58,60,63,71,73–76,89–92,98 Sample sizes ranged from 30 to 13,303 patients. The rates of successful blood pressure control with a single agent ranged from 6% to 93.3% (median 55%).
We performed a meta-analysis of data from the 26 studies directly comparing ACE inhibitors and ARBs. Data from RCTs gave an estimated OR of 1.08 (95% CI 0.94 to 1.25), suggesting that the odds of successful blood pressure control are not statistically significantly different with an ARB alone versus an ACE inhibitor alone (Fig. 2). The relationship between the summary estimated OR, number of randomized patients, and number of comparative studies is shown in Appendix Figure 1 (available online) and demonstrates minimal change in the estimated effect in spite of multiple new studies over the last several years.
An additional 12 studies with 31,144 patients that reported how many patients died or had myocardial infarction or stroke were published since the 2007 review, for a total of 21 studies (three evaluating DRIs).11–14,17–21,23,43,46,47,49,52,58,75,80,84,87,94,95,99–102 Most of these studies excluded high risk patients and reported only one or two clinical events. The study by Barnett et al.23 provided the largest sample and had the longest duration of follow-up. This study evaluated telmisartan versus enalapril in 250 patients with hypertension, type 2 diabetes and early nephropathy over a 5-year treatment period. Cardiovascular events occurred at a similar rate in both treatment groups: there were six strokes in each group; nine nonfatal MIs in the telmisartan group and six in the enalapril group; and nine patients with heart failure in the telmisartan group and six in the enalapril group. This study also reported 12 deaths, six in the telmisartan group and six in the enalapril group.
Large direct comparison trials in high-risk patients, such as the Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET), did not meet our inclusion criteria. Despite a 10-fold increase in the number of patients since the initial review, only 38 deaths were reported, 22 of which were new since the original report.
Twenty studies (eight new, none involving a DRI) met our inclusion criteria and evaluated serum lipid changes,19,29,33,35,40,43,46,49,62,65,73,80,81,85,89,91,92,95,96 although none addressed the use of lipid-lowering agents during the study period. Of 12 studies reporting direct statistical comparisons,29,33,35,43,62,65,73,81,85,87,92,95 only three found different effects between the medications compared, with results favoring ACE inhibitors in one study and ARBs in the other two.
Twenty-three studies (10 new) measured serum glucose or hemoglobin A1c.19,22,29,33,35,40,43,46,49,60,62,65,77,80,81,85,87,91,92,95,96,99,102 None addressed hypoglycemic therapy during the study. No change occurred in hemoglobin A1c either between or within groups in the 16 studies reporting this outcome.
Thirteen studies (five new, including one evaluating a DRI) presented original results on left ventricular mass or function.14,22,25,27,28,41,69,78,79,84,88,89,92 Apart from one study in which there was a greater reduction of left ventricular hypertrophy with an ARB,89 there were no differences between ACE inhibitor and ARB groups. Only one study evaluated a DRI; it found similar effects between aliskiren and an ARB.14
Thirty studies (11 new, including two evaluating a DRI) presented original results on either serum creatinine/GFR,13,19,22,28,33,43,59,72,82,87,91,92 proteinuria,11,29,58,60,80,103 or both.23,30,46,49,56,62,65,73,81,97,99,102 The ten new studies that described changes in creatinine or glomerular filtration rate did not consistently demonstrate differential effects with use of ACE inhibitors versus ARBs, nor did the two studies comparing a DRI with the ACE inhibitor ramipril;11,13 these findings were similar to those from the 2007 report. The analysis of all studies gave an estimated standardized mean difference of 0.11 (95% CI −0.05 to 0.27), suggesting that mean posttreatment creatinine levels were non-significantly higher for the ARB studies.
As in the 2007 report, the six new studies that quantitatively described changes in urine albumin or protein excretion also consistently demonstrated no differential effects with use of ACE inhibitors versus ARBs.30,58,62,81,97,102 None evaluated DRIs, all were conducted in non-U.S. sites, and most had between 6 and 12 months follow-up.30,58,62,81
The updated search identified no new studies that reported quality-of-life outcomes.
Fourteen studies (seven new, including two evaluating a DRI) met our inclusion criteria and reported overall rates of serious adverse events.11,14,15,43,45,50,52,53,58,59,72,73,87,94 The nature of serious adverse event reporting was inconsistent, and overall rates were low (0 to 6%); thus, data on these events are insufficient to quantitatively assess differential effects.
One serious adverse event, angioedema, has been reported to occur in ACE inhibitor-treated patients with much greater frequency than in ARB-treated patients.101 However, this outcome was reported in only four studies (one new).11,45,57,63 One of the reported cases occurred in a patient treated with a DRI; the other five cases were in patients treated with an ACE inhibitor.
Forty-eight studies (18 new, including two evaluating a DRI) reported rates of one or more specific adverse events.11,13–16,19,24,26–30,34,36,37,43,45,47–55,57–59,61–63,65,68,72,74,75,78,81,87,89,90,94,95,97,103–105 Given the large number of commonly reported specific adverse events, we focused on three specific events with the largest difference in absolute rates across studies: dizziness, headache, and cough.
Rates of dizziness in 31 studies reporting this event ranged from 1% to 20% in ARB-treated groups (mean 3.7%), 0% to 18% in ACE inhibitor-treated groups (mean 4.4%), and 3% to 8% in DRI-treated groups (mean 6.0%).
Rates of headache in the 30 included studies ranged from 1% to 22% in ARB-treated groups (mean 5.8%), 0% to 34% in ACE inhibitor-treated groups (mean 7.0%), and 9% to 11% in DRI-treated groups (mean 10.0%). This compares to mean rates of 6.3% in ARB-treated groups and 7.9% in ACE inhibitor-treated groups reported in the 2007 review.
Forty-two studies (13 new, including two evaluating a DRI) reported cough as an adverse event; of these, 40 studies compared rates of cough in subjects treated with ACE inhibitors or ARBs,15,16,19,24,26–30,34,36,37,45,47–53,55,57,59,61–63,65,68,72,74,75,87,89,90,94,95,97,103–105 and two in subjects treated with an ACE inhibitor or a DRI.11,13 Rates of cough ranged from 0 to 13% for ARB-treated groups (mean 2.2%) and 0 to 23% in ACE inhibitor-treated groups (mean 8.7%), and were 4% in DRI-treated patients. Based on our meta-analysis of RCTs, ACE inhibitors have consistently been shown to be associated with higher risk of cough than ARBs (OR 4.74; 95% CI 3.56 to 6.31) (Fig. 3); the new evidence did not significantly improve precision of the 2007 estimate (Appendix Figure 2 available online).
Forty-one studies (17 new, including 3 evaluating a DRI) reported withdrawals due to adverse events.11,13,14,16,19,22–24,26–28,30,36,37,43,45,47–50,52–54,57–63,65,72,78,80,82,88–90,92,95,97 Thirty-six trials reported withdrawals due to adverse events for both ACE inhibitors and ARBs; in 28 of these trials (78%) there were more withdrawals in the ACE inhibitor-treated groups.
Meta-analysis of RCTs reporting this outcome yielded an estimated OR of 1.77 (95% CI 1.42 to 2.21) (Fig. 4); however, absolute rates of withdrawal due to an adverse event were low: 5.4% for ACE inhibitors and ~3% for ARBs. In spite of increasing evidence, this estimated OR has been remarkably stable over the last decade (Appendix Figure 3 available online).
Two studies (both RCTs) compared a DRI with an ACE inhibitor for this outcome. Combining these studies yielded an estimated OR of 0.89 (95% CI 0.46 to 1.71) for DRI-treated patients relative to ACE inhibitor-treated patients.
Thirty-nine studies (22 new, including three evaluating a DRI) reported at least some quantitative information on persistence or adherence.11,13,14,16,26,30,36,37,42–44,47,48,50,55,57,58,62,73,76,81,88,89,93–95,97,98,106–118 Adherence to study protocol, in terms of pill counts, was universally high: at least 97% in five of the nine studies assessed and above 90% in all nine, without significant between group differences. Persistence with ARBs was modestly better than with ACE inhibitors, primarily due to adverse effects.
There was limited reporting of the comparative effectiveness of ACE inhibitors, ARBs, and DRIs in particular subgroups of patients. Among the 80 studies reporting blood pressure outcomes, four (none new) reported data for women,35,45,54,55 nine studies (three new) reported results in patients ≥65 years old,13,17,34,38,45,53,54,75,90 five studies (one new) reported data in a subgroup of black patients,13,17–21,61,74,90 and 10 studies (two new) reported results for patients with diabetes and hypertension.23,29,32,33,46,47,49,60,73,80 None of the results for these subgroups differed appreciably from the combined results for all patients.
Since the 2007 report on the comparative effectiveness of ACE inhibitors and ARBs in the treatment of hypertension, substantially more original research directly comparing these agents has been published (59% more studies). However, data from these new studies have not significantly changed the conclusions or the strength of evidence from the original report. As demonstrated in Appendix Figures 1, 2, and 3 (available online), the number of new studies is increasing, but with little added precision for our estimates of effects on blood pressure lowering, withdrawals due to adverse events, and cough, all of which are known with a high level of certainty. Conversely, the new evidence reported here did not significantly add to our understanding of long-term outcomes, quality of life, renal outcomes, medication adherence, or differences in key patient subgroups. New evidence on the comparative effectiveness of DRIs versus either ACE inhibitors or ARBs was limited to three studies with 2049 patients and did not allow definitive conclusions for any of our included outcomes. More long-term studies directly comparing DRIs with ACE inhibitors and ARBs are needed before this new medication class can be recommended.
The lack of clinically useful new information, in spite of many more new publications, is likely explained by several factors. There could be a mismatch between the areas of greatest research value and the actual research being funded. Formal analysis of the value of new research Information may be useful for guiding decisions about the use of future research funds.119 Even if value-of-information methods were applied to this area, the incentives for comparative effectiveness research investment differ between public funders and private, for-profit corporations. Therefore, the highest value research may not be the most likely to be funded.
The primary outcomes of many of the new studies were biochemical, and the clinical information abstracted from their reports was sparse. Even among our clinically measured outcomes, many are only a proxy for the clinically meaningful outcomes. The importance of directly comparing these medications’ effects on clinical outcomes is particularly important considering the mixed results of other placebo-controlled and direct comparison studies of ACE inhibitors and ARBs for these outcomes.120–122
Our review has some limitations. While our results are restricted to patients with essential hypertension, the agents studied here have been compared in large studies and in systematic reviews for related conditions such as congestive heart failure, ischemic heart disease, and chronic kidney disease.120,123,124 These systematic reviews, like ours, have limited inclusion to studies conducted in patients with the target condition; however, these reviews have examined an overlapping set of efficacy and safety outcomes. As a result, important direct comparison trials are often excluded from reviews because they do not meet the target condition inclusion criteria. Such was the case with ONTARGET, which was excluded from this review because no results were reported exclusively for patients with hypertension. It is likely that combining studies reporting identical outcomes, but in different target populations, may yield important new information, particularly for rarer events.125
The ability to evaluate heterogeneity in treatment response is limited by infrequent reporting of subgroup results. This limitation could be overcome if individual patient-level data were available for meta-analyses. This analysis would be logistically challenging given the proprietary nature of the data and a lack of uniform data quality standards; however, there are successful examples demonstrating such broad collaboration.126
There continue to be significant efforts to differentiate the comparative benefits and harms of the vast array of available pharmacotherapies for reducing vascular risk. Many of the differences between medication classes are likely small but nevertheless important when applied to the large population of patients. Detecting these differences may depend on the ability to maximize the information available across completed research studies and prioritization of future research around remaining areas of uncertainty. Coordinating these efforts across the broad spectrum of stakeholders would require a significant commitment from private and public sponsors of research, but one that may better serve a public interest in research efficiency.
Based on the available evidence you first recommend against a DRI because there is not enough evidence to support its use. You advise the patient that an ACE inhibitor or an ARB would likely lower her blood pressure to a similar degree. You cannot be sure which one will provide greater risk reduction for a heart attack, stroke, kidney failure, or development of diabetes; however, when these have been directly compared in research studies, there was no significant difference. You advise her that approximately 9% of people who start taking an ACE inhibitor develop a dry cough which goes away only with stopping the medication, compared to only 2% of those who start an ARB. There is also a 1 in 1000 risk for developing angioedema from an ACE inhibitor, and while the risk is not known for sure for ARBs, it is likely much lower (approximately 1 in 10,000). You also point out the significantly lower cost of ACE inhibitors; however this cost difference will likely decrease in the future with generic ARBs such as losartan. Based on the proven efficacy, cost difference, and acceptably low risk for side effects, she decides to try the ACE inhibitor lisinopril first, with a plan to switch to an ARB if she notices any side effects from this medication.
This project was funded under Contract No. Contract No. 290-02-0025 from the Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsements by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
This project was funded under Contract No. Contract No. 290-02-0025 from the Agency for Healthcare Research and Quality (AHRQ), U.S. Department of Health and Human Services. AHRQ commented on the planned methods and draft report, but did not participate in the search of the literature, data abstraction, or evaluation of individual studies. The first author is funded by a HSR&D career development award from the Department of Veterans Affairs.
This work was presented at the Academy Health Annual Research Meeting in Seattle, WA on June 14th, 2011.