RESEARCH ARTICLE
Cardiovascular and Renal Outcomes of
Renin–Angiotensin System Blockade in Adult
Patients with Diabetes Mellitus: A Systematic
Review with Network Meta-Analyses
Ferrán Catalá-López1,2,3,4*, Diego Macías Saint-Gerons2, Diana González-Bermejo2,
Giuseppe M. Rosano5, Barry R. Davis6, Manuel Ridao7,8, Abel Zaragoza2,
Dolores Montero-Corominas2, Aurelio Tobías9, César de la Fuente-Honrubia2,10,
Rafael Tabarés-Seisdedos1,4, Brian Hutton3,11
1 Department of Medicine, University of Valencia/INCLIVA Health Research Institute, Valencia, Spain,
2 Division of Pharmacoepidemiology and Pharmacovigilance, Spanish Agency of Medicines and Medical
Devices (AEMPS), Madrid, Spain, 3 Clinical Epidemiology Program, Ottawa Hospital Research Institute,
Ottawa, Ontario, Canada, 4 Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM),
Instituto de Salud Carlos III, Madrid, Spain, 5 Centre for Clinical and Basic Research, Department of Medical
Sciences, Istituto di Ricovero e Cura a Carattere Scientifico San Raffaele Pisana, Rome, Italy, 6 The
University of Texas School of Public Health, Houston, Texas, United States of America, 7 Instituto Aragonés
de Ciencias de la Salud, Red de Investigación en Servicios de Salud en Enfermedades Crónicas
(REDISSEC), Zaragoza, Spain, 8 Fundación para el Fomento de la Investigación Sanitaria y Biomédica de
la Comunitat Valenciana (FISABIO–Salud Pública), Valencia, Spain, 9 Spanish Council for Scientific
Research (CSIC), Barcelona, Spain, 10 Area of Budgetary Stability, Ministry of Finance and Public
Administrations, Madrid, Spain, 11 School of Epidemiology, Public Health and Preventive Medicine, Faculty
of Medicine, University of Ottawa, Ottawa, Ontario, Canada
* ferran_catala@hotmail.com
Abstract
Background
Medications aimed at inhibiting the renin–angiotensin system (RAS) have been used exten-
sively for preventing cardiovascular and renal complications in patients with diabetes, but
data that compare their clinical effectiveness are limited. We aimed to compare the effects
of classes of RAS blockers on cardiovascular and renal outcomes in adults with diabetes.
Methods and Findings
Eligible trials were identified by electronic searches in PubMed/MEDLINE and the Cochrane
Database of Systematic Reviews (1 January 2004 to 17 July 2014). Interventions of interest
were angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers
(ARBs), and direct renin (DR) inhibitors. The primary endpoints were cardiovascular mortal-
ity, myocardial infarction, and stroke—singly and as a composite endpoint, major cardiovas-
cular outcome—and end-stage renal disease [ESRD], doubling of serum creatinine, and all-
cause mortality—singly and as a composite endpoint, progression of renal disease. Sec-
ondary endpoints were angina pectoris and hospitalization for heart failure. In all, 71 trials
(103,120 participants), with a total of 14 different regimens, were pooled using network
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 1 / 30
OPEN ACCESS
Citation: Catalá-López F, Macías Saint-Gerons D,
González-Bermejo D, Rosano GM, Davis BR, Ridao
M, et al. (2016) Cardiovascular and Renal Outcomes
of Renin–Angiotensin System Blockade in Adult
Patients with Diabetes Mellitus: A Systematic Review
with Network Meta-Analyses. PLoS Med 13(3):
e1001971. doi:10.1371/journal.pmed.1001971
Academic Editor: Maarten W. Taal, Royal Derby
Hospital, UNITED KINGDOM
Received: September 24, 2015
Accepted: January 26, 2016
Published: March 8, 2016
Copyright: © 2016 Catalá-López et al. This is an
open access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information files
(S6 Table and S7 Table).
Funding: FCL and RTS are partially funded by
Generalitat Valenciana (PROMETEOII/2015/021).
RTS is also funded by INCLIVA and Institute of
Health Carlos III (PI14/00894)/CIBERSAM. BH is
supported by a New Investigator Award from the
Canadian Institutes of Health Research and the Drug
Safety and Effectiveness Network. The funders had
meta-analyses. When compared with ACE inhibitor, no other RAS blocker used in mono-
therapy and/or combination was associated with a significant reduction in major cardiovas-
cular outcomes: ARB (odds ratio [OR] 1.02; 95% credible interval [CrI] 0.90–1.18), ACE
inhibitor plus ARB (0.97; 95% CrI 0.79–1.19), DR inhibitor plus ACE inhibitor (1.32; 95% CrI
0.96–1.81), and DR inhibitor plus ARB (1.00; 95% CrI 0.73–1.38). For the risk of progres-
sion of renal disease, no significant differences were detected between ACE inhibitor and
each of the remaining therapies: ARB (OR 1.10; 95% CrI 0.90–1.40), ACE inhibitor plus
ARB (0.97; 95% CrI 0.72–1.29), DR inhibitor plus ACE inhibitor (0.99; 95% CrI 0.65–1.57),
and DR inhibitor plus ARB (1.18; 95% CrI 0.78–1.84). No significant differences were
showed between ACE inhibitors and ARBs with respect to all-cause mortality, cardiovascu-
lar mortality, myocardial infarction, stroke, angina pectoris, hospitalization for heart failure,
ESRD, or doubling serum creatinine. Findings were limited by the clinical and methodologi-
cal heterogeneity of the included studies. Potential inconsistency was identified in network
meta-analyses of stroke and angina pectoris, limiting the conclusiveness of findings for
these single endpoints.
Conclusions
In adults with diabetes, comparisons of different RAS blockers showed similar effects of
ACE inhibitors and ARBs on major cardiovascular and renal outcomes. Compared with
monotherapies, the combination of an ACE inhibitor and an ARB failed to provide significant
benefits on major outcomes. Clinicians should discuss the balance between benefits, costs,
and potential harms with individual diabetes patients before starting treatment.
Review registration
PROSPERO CRD42014014404
Introduction
Diabetes mellitus has become one of the most challenging public health problems worldwide,
affecting approximately 410 million people [1] and accounting for 1.3 million deaths in 2013,
twice as many as in 1990 [2]. Complications of diabetes mellitus, especially cardiovascular and
renal sequelae, cause substantial premature death and disability [1–4].
Medications aimed at inhibiting the renin–angiotensin system (RAS) have been used exten-
sively for preventing cardiovascular and renal outcomes in patients with diabetes. Blockade of the
RAS is a key therapeutic target because RAS controls circulatory volume and electrolyte balance
and is an important regulator of hemodynamic stability. Currently, three classes of drugs that
interact with the RAS are used to inhibit the effects of angiotensin II: angiotensin-converting
enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), and direct renin (DR) inhibitors.
ACE inhibitors block the conversion of angiotensin I into angiotensin II, ARBs selectively inhibit
angiotensin II from activating the angiotensin-specific receptor AT1, and DR inhibitors block the
conversion of angiotensinogen into angiotensin I. Although all RAS blockers are intended to
inhibit the effects of angiotensin II, there are differences that may distinguish them [5].
Most evidence-based guidelines for the management of hypertension and diabetes have gener-
ally recommended the use of ACE inhibitors and ARBs in preference to other antihypertensive
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 2 / 30
no role in study design, data collection and analysis,
decision to publish, or preparation of the manuscript.
Competing Interests: BH has previously received
funds for methodologic advice from Amgen Canada
and Cornerstone Research Group for the provision of
advice related to systematic reviews and meta-
analysis. None of the advice is related to the content
of this manuscript. All the other authors have
declared that no competing interests exist.
Abbreviations: ACE, angiotensin-converting
enzyme; ARB, angiotensin receptor blocker; BB, beta
blocker; CCB, calcium channel blocker; CI,
confidence interval; CrI, credible interval; DIC,
deviance information criterion; DR, direct renin;
ESRD, end-stage renal disease; OR, odds ratio;
RAS, renin–angiotensin system; SUCRA, surface
under the cumulative ranking curve.
agents [6–9]. In these guidelines, any particular RAS blocker (ACE inhibitor or ARB) is preferen-
tially recommended as the treatment of choice. However, current guidelines are based on only a
small number of randomized trials comparing the effects of RAS blockade specifically in patients
with diabetes. Cardiovascular and renal outcomes with RAS blockers for adults with diabetes have
been evaluated in large multicenter randomized controlled trials [10–22] and meta-analyses [23–
28]. The task of establishing the comparative effectiveness of RAS blockers has been limited by the
very complex array of trials that compare treatments. Results of recent meta-analyses have
highlighted potential differences in treatment effects between ACE inhibitors and ARBs [23,24].
Traditionally, meta-analyses of RAS blockers have been limited by not including all the valuable
information on the most common serious cardiovascular and renal outcomes [23–28], not explor-
ing effects in the subgroup of patients with diabetes [29–37], and, importantly, omitting large trials
with direct comparisons of RAS blockers and competing agents in clinically important subgroups
[22,38–41]. Determining whether RAS blockers may be different in terms of their relative benefits
and safety is a topic of great interest to patients, clinicians, scientists, guideline developers, and
policy-makers. Unlike for previous analyses [23–27,42], many more trials, patients, and outcome
data are now available for a comprehensive study to address this clinical question.
Given this knowledge gap, we aimed to examine the comparative effects of classes of RAS
blockers in terms of cardiovascular and renal outcomes in the treatment of adult patients with
diabetes mellitus. We used network meta-analyses to integrate direct and indirect evidence
comparing multiple interventions of interest into unified analyses of all available randomized
trials that can serve to guide evidence-based decision-making.
Methods
This systematic review was conducted and reported in accordance with the reporting guidance
provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) statement (S1 Checklist) [43,44]. We developed a systematic review protocol and
registered it with PROSPERO (CRD42014014404). Our methods are briefly described here.
Study Eligibility Criteria
Studies meeting the following selection criteria were included: randomized parallel-group con-
trolled trials of a minimum 1-y duration; participants adults aged 18 y or older with type 1 or 2
diabetes mellitus; reporting at least one of the cardiovascular or renal outcomes of interest as
well as the number of participants enrolled in each treatment arm and the number of partici-
pants with events in each treatment arm; comparison of RAS blockers in monotherapy and/or
combination therapy regimens, including ACE inhibitors, ARBs, and DR inhibitors versus
each other or versus placebo or other active antihypertensive treatments (e.g., calcium channel
blockers [CCBs], beta blockers [BBs], or diuretics).
Electronic Literature Search
Based on awareness of a large number of existing reviews and meta-analyses, an unlimited pri-
mary search for randomized trials was not conducted. In its place, we used a staged approach
to study identification, beginning with a systematic search of relevant trials included in system-
atic reviews available in PubMed/MEDLINE and the Cochrane Database of Systematic Reviews
(1 January 2004 to 17 July 2014) and existing meta-analyses of which we were aware (see S1
Text for details of search terms and S2 Text for references of previous reviews). PubMed/MED-
LINE was next searched to identify other additional relevant trials published outside the time
frames of previous reviews (1 January 2010 to 5 September 2014; reflects content from PubMed
searches last conducted on 1 October 2014). We compiled a list of the unique PubMed/
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 3 / 30
MEDLINE identification numbers of all relevant articles, and performed a related articles
search. This technique has been shown to be effective in identifying relevant studies [45],
increases efficiency in study identification in the presence of an already large evidence base,
and is being used as part of an ongoing network meta-analysis research program [46]. Searches
were supplemented by manual searches of clinical trial registers (including www.ClinicalTrials.
gov) and review of references of relevant papers. Finally, we contacted authors of primary pub-
lications and sponsors of trials for missing outcome data or unclear information. Six drug man-
ufacturers (Boehringer Ingelheim for ONTARGET [Ongoing Telmisartan Alone and in
Combination with Ramipril Global Endpoint Trial] [40,41], PRoFESS [Prevention Regimen
for Effectively Avoiding Second Strokes] [47], and TRANSCEND [Telmisartan Randomized
Assessment Study in ACE-Intolerant Subjects with Cardiovascular Disease] [48,49]; Daiichi
Sankio for ROADMAP [Randomized Olmesartan and Diabetes Microalbuminuria Prevention]
[50]; Sanofi for IDNT [Irbesartan Diabetic Nephropathy Trial] [16] and IRMA 2 [Irbesartan in
Patients with Type 2Diabetes and Microalbuminuria] [17]; Servier for EUROPA/PERSUADE
[Perindopril Substudy in Coronary Artery Disease and Diabetes] [51]; Takeda for DIRECT-
Prevent 1 [52,53], DIRECT-Protect 1 [52,53], and DIRECT Protect 2 [53,54]; and Novartis for
ALTITUDE [Aliskiren Trial in Type 2 Diabetes Using Cardiovascular and Renal Disease End-
points] [22], VALUE [Valsartan Antihypertensive Long-Term Use Evaluation] [38], VAL-
IANT [Valsartan in Acute Myocardial Infarction Trial] [39], Val-HeFT [Valsartan Heart
Failure Trial] [55], and ASTRONAUT [Aliskiren Trial on Acute Heart Failure Outcomes]
[56]) and four independent investigators (for ALLHAT [Antihypertensive and Lipid-Lowering
Treatment to Prevent Heart Attack Trial] [20], CASE-J [Candesartan Antihypertensive Sur-
vival Evaluation in Japan] [57], COLM [Combination of Olmesartan and CCB or Low Dose
Diuretics in High Risk Elderly Hypertensive Patients] [58], and OSCAR [Olmesartan and Cal-
cium Antagonists Randomized] [59]) provided additional information about outcome data.
Prespecified Outcome Measures
The prespecified primary endpoints were major cardiovascular outcome (composite endpoint
including death from cardiovascular causes, myocardial infarction, and stroke) and progres-
sion of renal disease (composite endpoint including doubling of baseline serum creatinine
level, end-stage renal disease [ESRD] defined as the need for any dialysis or renal transplanta-
tion, and all-cause mortality). We also analyzed all of the component endpoints of these two
composite endpoints separately. Secondary endpoints were angina pectoris and hospitalization
for heart failure. All outcomes were based on the longest follow-up period available for each
included study.
Screening, Data Extraction, and Risk of Bias Assessment
Eligible trials identified from our searching efforts were screened by one reviewer (F. Catalá-
López, qualified clinical epidemiologist) and verified independently by two trained reviewers
(D. Macías Saint-Gerons and C. de la Fuente-Honrubia, senior pharmacoepidemiologists).
Using a predesigned form that was piloted on a small sample (14%) of studies, the same review-
ers were also responsible for extraction and verification of data on general participant charac-
teristics (e.g., average age, gender, type and duration of diabetes, level of albuminuria, presence
of hypertension, presence of coronary artery disease, mean or median follow-up) and outcome
data. The Cochrane risk of bias scale [60]—which considers sequence generation, allocation
concealment, blinding, and other aspects of bias—was used to assess each study’s risk of bias.
The overall rating of risk of bias for each study was the worst rating for any of the criteria (e.g.,
if any domain is scored high risk of bias, the study was considered high risk of bias). We used
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 4 / 30
the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) meth-
odology [61,62] to evaluate the quality of evidence for each outcome (high, moderate, low, or
very low quality). The approach we used was consistent with previous reporting of network
meta-analyses [63]. Risk of bias assessments and application of GRADE methods were per-
formed by one reviewer (F. Catalá-López) and validated by a second reviewer (D. Macías Saint-
Gerons, C. de la Fuente-Honrubia. or M. Ridao). Any discrepancies between reviewers for any
of the above steps were discussed until consensus was achieved.
Data Synthesis and Analyses
Our analysis classified RAS blockers used in monotherapy and/or combination as separate
treatment nodes irrespective of their doses: ACE inhibitor, ARB, ACE inhibitor combination
(e.g., ACE inhibitor plus ARB, ACE inhibitor plus DR inhibitor, ACE inhibitor plus diuretic,
ACE inhibitor plus CCB), ARB combination (e.g., ARB plus DR inhibitor, ARB plus diuretic,
ARB plus CCB), DR inhibitor combination (e.g., DR inhibitor plus diuretic), and placebo or
control (conventional therapy/usual care). In the ASTRONAUT trial [56] (aliskiren versus pla-
cebo), more than 95% of the patients with diabetes were receiving a diuretic in both groups;
therefore, the diabetes subgroup analysis from this study was regarded as a trial of DR inhibitor
plus diuretic versus diuretic. Similarly, the ALTITUDE trial [22] (aliskiren versus placebo)
evaluated a strategy of aliskiren in patients with diabetes who were also receiving an ACE
inhibitor or an ARB and was therefore regarded as a trial of DR inhibitor plus ACE inhibitor
versus DR inhibitor plus ARB versus ACE inhibitor versus ARB.
Whenever possible we used results from intention-to-treat analyses. To calculate direct esti-
mates of treatment effect for each pair of medications, we conducted pairwise random effects
meta-analyses. We report the results as odds ratios (ORs) and corresponding 95% confidence
intervals (CIs). We evaluated statistical heterogeneity by estimating the variance between stud-
ies with Cochran’s Q test and I2 statistic [64,65]. The I2 statistic is the proportion of total varia-
tion observed between the trials attributable to differences between trials rather than to
sampling error. We considered I2< 30% as representing low statistical heterogeneity and I2>
75% as representing high statistical heterogeneity. Patient and study characteristics were also
empirically assessed by members of the research team to establish that these were comparable
across studies and comparisons. For each outcome, we present graphically the geometry of the
treatment network of all comparisons using a network graph [66]. Using a Bayesian frame-
work, we did network meta-analyses for each prespecified outcome and treatment. Network
meta-analysis [67,68] allows the integration of data from both direct and indirect evidence,
increasing precision while randomization is preserved, but can also be used to estimate com-
parisons between pairs of treatments that have not been compared in individual studies. When
performing a network meta-analysis, we relied on the assumptions of transitivity [69] (i.e., if
drug B is superior to drug A, and drug C is superior to drug B, it is assumed that drug C is supe-
rior to drug A) and consistency (equivalency of treatment effects from direct and indirect evi-
dence). We used random effects network meta-analysis models as recommended elsewhere,
which account for correlations in multi-arm trials [70] and which use vague (noninformative)
prior distributions for all treatment effects as well as the between-study variance parameter (S3
Text). We report the results as posterior median ORs with corresponding 95% credibility inter-
vals (CrIs), which are the Bayesian analogue of 95% CIs. The ORs reported are relative effects
of multiple RAS blocker regimens. We express these using ACE inhibitor as the reference treat-
ment, because ACE inhibitors were the first class to be developed (with the arrival of captopril
and enalapril in the early 1980s) and because these agents are frequently prescribed worldwide.
All network meta-analyses were based on a total of 100,000 iterations or more, with a burn-in
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 5 / 30
of 50,000 iterations. We assessed convergence on the basis of Brooks–Gelman–Rubin plots
[71]. The consistency of results was qualitatively examined by comparing the results obtained
via pairwise meta-analyses versus network meta-analyses. Consistency was also examined by
fitting both consistency and inconsistency models [72] for network meta-analysis and compar-
ing the deviance information criterion (DIC) between models, with smaller values indicative of
a better fit, and a difference of five or more being considered as important. In general, when
both models had a similar fit to the data as indicated by their DIC values, we concluded that
there was no evidence of inconsistency. We also used the surface under the cumulative ranking
curve (SUCRA) to potentially rank the treatments [73].
A priori sensitivity analyses were conducted to explore potentially important effect modifi-
ers. These included separate analyses that involved exclusion of the following: studies with high
risk of bias, small studies with fewer than 100 patients, studies in which>50% patients pre-
sented type 1 diabetes, and studies in non-hypertensive patients. Other preplanned analyses
were the extension of the primary unadjusted network meta-analysis model to include covari-
ates in meta-regression models [74] that considered adjustments for the following: year of pub-
lication, mean age of trial participants, percent of male participants, and control group risk
rate. Considering that follow-up times varied between the included trials, we also estimated the
rate ratios based on patient-years and corresponding 95% CrIs. All pairwise meta-analyses
were conducted using Stata 13 (StataCorp, College Station, Texas, US), while Bayesian network
meta-analyses were performed using WinBUGS 1.4.3 (MRC Biostatistics Unit, Cambridge,
UK).
Results
Overview of Study Selection and Study Characteristics
We included 71 studies (described in 88 publications) that satisfied our inclusion criteria
(PRISMA flowchart is shown in S1 Fig; details of the included studies are shown in S4 Text and
S1–S7 Tables), with 103,120 individuals randomly assigned to one of the study treatments. The
median duration of follow-up of the set of included studies was 3.2 y (range 1.0–9.0), and the
median sample size was 436 individuals (range 30–13,168), with 55 trials having at least 100
participants and 24 trials having more than 1,000. Fifty-four studies (76%) were two-arm trials,
12 (17%) were three-arm trials involving three active agents or two active agents and placebo,
and four (6%) were analyzed as multi-arm trials with four different active comparisons or three
active agents and placebo. In terms of methodological quality and potential risk of bias (S2 and
S3 Tables), 37 trials (52%) had an unclear risk of bias for at least one criterion, 25 trials (35%)
had a low risk of bias, and nine trials (13%) had a high risk of bias. Fifty trials (70%) were
funded by drug companies.
The characteristics of patients amongst the included studies are described in S4 Table. At
baseline, the median age was 60 y, the median percentage of male participants was 61%, and
the median duration of diabetes was 10.3 y. Forty-seven studies (63%) targeted type 2 diabe-
tes, (5%) enrolled patients with type 1 diabetes, and ten studies (32%) contained mixed popu-
lations with type 1 or type 2 diabetes. In terms of level of albuminuria, 15 trials enrolled
patients with macroalbuminuria, 13 enrolled patients with microalbuminuria, seven enrolled
patients with normoalbuminuria, and 25 trials enrolled mixed populations. Approximately
half of all studies (37 trials) included patients with hypertension at baseline, 15 trials enrolled
normotensive participants, and 19 trials enrolled mixed populations. In terms of history of
coronary disease, 58% of all studies (41 trials) were rated as unclear or not reported, and 27
trials were rated as having a mixed population (e.g., 5%–75% of participants with coronary
disease at baseline).
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 6 / 30
The structure of the underlying evidence base for each analysis is described in Table 1 and
Fig 1. The studies reporting data for each clinical outcome of interest are summarized in
Table 1 (and S4–S7 Tables). Figs 1 and S2 show the evidence networks of eligible treatment
comparisons for the full network meta-analysis. ACE inhibitors, placebo use or control (con-
ventional therapy/usual care), and ARBs were most investigated (49, 36, and 25 trials, respec-
tively). Treatments of interest in the main analyses were any RAS blocker compared with
another RAS blocker or placebo or control. Other antihypertensive control arms (e.g., CCBs,
BBs, or diuretics) were included in the evidence networks in order to incorporate additional
indirect evidence for the analyses while preserving randomization. The results for these alter-
native antihypertensive treatments are presented in S8 Table.
Findings from Network Meta-Analysis
Summaries from all network meta-analyses are provided below for each outcome.
Major cardiovascular outcome (composite) was reported in 12,328 of 92,469 patients from a
total of 33 studies. With placebo use as the reference, ACE inhibitors (OR 0.87; 95% CrI 0.75–
0.99; moderate confidence) significantly reduced the risk of major cardiovascular outcome. Com-
pared with ACE inhibitor as the reference treatment, none of the remaining RAS blockers used
in monotherapy or combination were associated with significant risk reductions: ARB (1.02; 95%
CrI 0.90–1.18; moderate confidence), ACE inhibitor plus ARB (0.97; 95% CrI 0.79–1.19; moder-
ate confidence), DR inhibitor plus ARB (1.00; 95% CrI 0.73–1.38; low confidence), and DR inhib-
itor plus ACE inhibitor (1.32; 95% CrI 0.96–1.81; low confidence) (Tables 2, 3, and S8).
Progression of renal disease (composite endpoint of ESRD, doubling of serum creatinine,
and all-cause mortality) occurred in 9,267 of 69,380 patients from 18 studies with available
data. No RAS blocker was found to reduce progression of renal disease compared to placebo.
Compared with ACE inhibitor as the reference treatment, no RAS blocker used in monother-
apy and/or combination was associated with any significant reduction of progression of renal
disease. Compared with ACE inhibitor as the reference, ORs were 1.10 (95% CrI 0.90–1.40;
moderate confidence) for ARB, 0.97 (95% CrI 0.72–1.29; moderate confidence) for ACE inhibi-
tor plus ARB, 0.99 (95% CrI 0.65–1.57; low confidence) for DR inhibitor plus ACE inhibitor,
and 1.18 (95% CrI 0.78–1.84; low confidence) for DR inhibitor plus ARB (Tables 4, 5, and S8).
Cardiovascular mortality occurred in 6,166 of 95,060 patients from 41 studies that provided
data. With placebo as the reference treatment, ACE inhibitor plus CCB (OR 0.14; 95% CrI
0.01–0.83) appeared to reduce the risk of cardiovascular mortality, though with very low confi-
dence. Compared with ACE inhibitor as the reference treatment (Tables 2 and S8), no RAS
blocker (except ACE inhibitor plus CCB) was associated with a significant reduction in cardio-
vascular mortality: ARB (1.07; 95% CrI 0.88–1.33; moderate confidence), ACE inhibitor plus
ARB (1.06; 95% CrI 0.79–1.49; moderate confidence), DR inhibitor plus ARB (0.98; 95% CrI
0.61–1.60; low confidence), and DR inhibitor plus ACE inhibitor (1.45; 95% CrI 0.91–2.35; low
confidence).
Myocardial infarction occurred in 4,593 of 84,792 patients from 48 studies with available
data. Both ACE inhibitors (OR 0.77; 95% CrI 0.62–0.92; moderate confidence) and ARBs
(0.82; 95% CrI 0.67–0.98; moderate confidence) were associated with reduced risk of myocar-
dial infarction compared with placebo. Compared with ACE inhibitor as the reference treat-
ment (Tables 2 and S8), no RAS blocker used in monotherapy and/or combination was
associated with a significant reduction in myocardial infarction: ARB (1.07; 95% CrI 0.89–1.28;
moderate confidence), ACE inhibitor plus ARB (1.00; 95% CrI 0.78–1.33; moderate confi-
dence), DR inhibitor plus ARB (0.83; 95% CrI 0.52–1.31; low confidence), and DR inhibitor
plus ACE inhibitor (1.35; 95% CrI 0.85–2.12; low confidence).
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 7 / 30
Table 1. Availability of data for principal analyses.
Treatment Data Available by Outcome
Major
Cardiovascular
Outcome
Cardiovascular
Mortality
Myocardial
Infarction
Stroke Angina
Pectoris
Hospitalization
for Heart Failure
Progression of
Renal Disease
ESRD Doubling
of CrS
All-Cause
Mortality
ACEi
Number of
trials
17 25 30 23 16 18 7 14 13 39
Number of
participants
16,692 17,601 17,730 17,446 14,307 17,382 11,933 13,382 11,818 19,165
Number of
events
2,607 1,394 1,153 835 1,340 1,211 1,900 278 429 2,522
ARB
Number of
trials
18 18 21 22 13 18 12 9 11 22
Number of
participants
21,792 21,788 23,111 23,422 15,738 21,125 14,726 12,276 14,002 23,249
Number of
events
2,772 1,335 1,005 1,041 1,125 1,274 1,712 455 633 2,277
ACEi + ARB
Number of
trials
3 3 5 5 4 6 3 3 2 6
Number of
participants
4,571 4,571 5,322 5,322 5,180 5,930 4,149 4,149 3,425 5,931
Number of
events
913 568 413 229 612 679 797 137 166 999
DRi + ACEi
Number of
trials
1 1 1 1 0 1 1 1 1 1
Number of
participants
1,926 1,926 1,926 1,926 0 1,926 1,926 1,926 1,926 1,926
Number of
events
234 136 84 65 0 106 108 53 82 202
DRi + ARB
Number of
trials
1 1 1 1 0 1 1 1 1 1
Number of
participants
2,353 2,353 2,353 2,353 0 2,353 2,353 2,353 2,353 2,353
Number of
events
224 115 64 86 0 103 155 69 134 179
DRi
+ diuretic
Number of
trials
1 1 1 1 1 1 0 0 1 1
Number of
participants
319 319 319 319 319 319 0 0 319 319
Number of
events
69 62 11 6 16 108 0 0 65 72
ACEi
+ diuretic
Number of
trials
1 1 1 2 0 1 1 1 1 3
Number of
participants
5,569 5,569 27 5,595 0 27 5,569 5,569 5,569 5,761
Number of
events
480 211 0 288 0 2 488 25 55 411
ARB
+ diuretic
Number of
trials
1 1 2 1 1 1 1 1 1 1
(Continued)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 8 / 30
Stroke was observed in 4,591 of 95,155 patients from 42 studies. The combination of ACE
inhibitor plus ARB was associated with a reduced risk of stroke compared with placebo (OR
0.80; 95% CrI 0.62–0.98; moderate confidence). Compared with ACE inhibitor as the reference
treatment (Tables 2 and S8), no RAS blocker used in monotherapy and/or combination was
associated with a significant reduction in the incidence of stroke: ARB (1.01; 95% CrI 0.88–
1.16; moderate confidence), ACE inhibitor plus ARB (0.88; 95% CrI 0.71–1.08; moderate confi-
dence). Combinations of DR inhibitor with ACE inhibitor/ARB were potentially associated
with an increased risk of stroke: DR inhibitor plus ARB compared with ACE inhibitor plus
ARB (1.44; 95% CrI 1.00–2.13; low confidence; S8 Table).
ESRD was reported in 1,786 of 67,316 patients from the 22 studies with available data. ACE
inhibitors (OR 0.68; 95% CrI 0.51–0.91; moderate confidence), ARBs (0.74; 95% CrI 0.57–0.97;
moderate confidence), and combinations of ACE inhibitor plus ARB (0.62; 95% CrI 0.42–0.90;
moderate confidence) were associated with a reduced risk of ESRD compared with placebo.
Compared with ACE inhibitor as the reference treatment (Tables 4 and S8), no RAS blocker
used in monotherapy and/or combination was associated with a significant reduction in ESRD:
ARB (1.09; 95% CrI 0.84–1.42; moderate confidence), ACE inhibitor plus ARB (0.91; 95% CrI
0.63–1.27; moderate confidence), DR inhibitor plus ACE inhibitor (1.09; 95% CrI 0.66–1.85;
low confidence), and DR inhibitor plus ARB (1.17; 95% CrI 0.72–1.93; low confidence).
Table 1. (Continued)
Treatment Data Available by Outcome
Major
Cardiovascular
Outcome
Cardiovascular
Mortality
Myocardial
Infarction
Stroke Angina
Pectoris
Hospitalization
for Heart Failure
Progression of
Renal Disease
ESRD Doubling
of CrS
All-Cause
Mortality
Number of
participants
678 678 734 678 678 678 678 678 678 678
Number of
events
26 7 9 16 2 5 8 3 6 29
ACEi + CCB
Number of
trials
1 3 2 1 0 0 0 0 0 4
Number of
participants
104 530 158 104 0 0 0 0 0 696
Number of
events
4 1 2 1 0 0 0 0 0 5
ARB + CCB
Number of
trials
1 2 2 2 2 2 1 1 1 2
Number of
participants
684 1,003 1,003 1,003 1,003 1,003 684 684 684 1,003
Number of
events
26 5 6 34 3 10 8 0 8 29
Placebo/
control
Number of
trials
17 19 19 18 13 14 11 10 17 31
Number of
participants
21,798 21,998 16,097 21,813 11,848 14,987 16,118 12,507 16,533 22,785
Number of
events
2,580 1,094 747 1,138 806 607 1,880 408 629 1,976
There were also trials involving comparisons with other control arms (diuretics, CCBs, and/or BBs in monotherapy), which were included in the evidence
networks to preserve randomization. ACEi, ACE inhibitor; CrS, serum creatinine; DRi, DR inhibitor.
doi:10.1371/journal.pmed.1001971.t001
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 9 / 30
Doubling of serum creatinine level was reported in 2,645 of 67,505 patients from 24 studies.
With placebo as the reference treatment, ACE inhibitors (OR 0.70; 95% CrI 0.52–0.91; moder-
ate confidence) were associated with important reductions in the risk of doubling of serum cre-
atinine. Compared with ACE inhibitor as the reference treatment (Tables 4 and S8), no RAS
blocker used in monotherapy and/or combination was associated with a significant reduction
of doubling of serum creatinine: ARB (1.26; 95% CrI 0.97–1.79; moderate confidence), ACE
inhibitor plus ARB (0.99; 95% CrI 0.65–1.56; moderate confidence), DR inhibitor plus ARB
(1.27; 95% CrI 0.73–2.35; low confidence), and DR inhibitor plus ACE inhibitor (0.93; 95% CrI
0.53–1.74; low confidence).
Death from any cause was reported in 11,199 of 101,369 patients from 59 studies. Com-
pared with placebo or with ACE inhibitor as the reference treatment (S8 Table), no RAS
blocker used in monotherapy and/or combination was associated with a significant reduction
in all-cause mortality. Compared with ACE inhibitor as the reference treatment (Tables 4 and
S8), the ORs were 1.04 (95% CrI 0.92–1.18; moderate confidence) for ARB, 1.03 (95% CrI
0.88–1.21; moderate confidence) for ACE inhibitor plus ARB, 1.29 (95% CrI 0.96–1.72; low
confidence) for DR inhibitor plus ACE inhibitor, and 0.90 (95% CrI 0.67–1.21; low confidence)
for DR inhibitor plus ARB.
Angina pectoris was reported in 5,026 of 65,656 patients from 30 studies. Both ACE inhibi-
tors (0.81; 95% CrI 0.65–0.96; moderate confidence) and ARBs (0.76; 95% CrI 0.61–0.98; mod-
erate confidence) were associated with a reduced risk of angina compared with placebo.
Compared with ACE inhibitor as the reference treatment (Tables 2 and S8), no RAS blocker
Fig 1. Evidence network of all treatment comparisons for all studies. Lines represent direct comparisons within randomized controlled trials. The size of
nodes is proportional to the number of randomized participants (sample size), and the width of the lines is proportional to the number of trials comparing each
pair of treatments. Nodes in green represent RAS blockers (in monotherapy and/or combination therapies). Nodes in blue represent other control arms
included in the evidence networks to preserve randomization. ACEi, ACE inhibitor; DRi, DR inhibitor.
doi:10.1371/journal.pmed.1001971.g001
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 10 / 30
Table 2. Cardiovascular outcomes: comparisons of random effects pairwisemeta-analysis with Bayesian network meta-analysis, including confi-
dence of assessments.
Outcome Treatment Number of Studies with
Direct Comparison to ACEi
(Participants)
Pairwise Meta-analysis
OR (95% CI); Quality
of Evidence
I2 (95%
CI)
Cochran’s Q
p-Value
Network Meta-analysis
OR (95% CrI); Quality
of Evidence
Major
cardiovascular
outcome
ACEi
(reference)
ARB 5 (13,480) 0.93 (0.77–1.13);
Moderate
62.3%
(0.0–
82.0)
0.03 1.02 (0.90–1.18);
Moderate
ACEi + ARB 3 (9,046) 0.95 (0.85–1.05);
Moderate
0.0%
(0.0–
90.0)
0.85 0.97 (0.79–1.19);
Moderate
DRi + ACEi 1 (3,790) 1.09 (0.90–1.34); Low — — 1.32 (0.96–1.81); Low
DRi + ARB 1 (4,217) 0.83 (0.68–1.02); Low — — 1.00 (0.73–1.38); Low
DRi
+ diuretic
0 (0) — — — 1.07 (0.63–1.91); Very
low
ACEi
+ diuretic
0 (0) — — — 1.05 (0.75–1.52); Very
low
ACEi + CCB 1 (206) 0.54 (0.15–1.91); Very
low
— — 0.47 (0.12–1.41); Very
low
Cardiovascular
mortality
ACEi
(reference)
ARB 5 (13,480) 0.97 (0.78–1.23):
Moderate
60.7%
(0.0–
85.0)
0.04 1.00 (0.82–1.23);
Moderate
ACEi + ARB 3 (9,046) 1.03 (0.86–1.22);
Moderate
29.6%
(0.0–
93.0)
0.24 1.06 (0.79–1.49);
Moderate
DRi + ACEi 1 (3,790) 1.18 (0.91–1.52); Low — — 1.45 (0.91–2.35); Low
DRi + ARB 1 (4,217) 0.80 (0.61–1.04); Low — — 0.98 (0.61–1.60); Low
DRi
+ diuretic
0 (0) — — — 1.32 (0.63–2.84); Very
low
ACEi
+ diuretic
0 (0) — — — 0.81 (0.49–1.36); Very
low
ARB
+ diuretic
0 (0) — — — 0.71 (0.02–14.19); Very
low
ACEi + CCB 1 (206) 0.49 (0.04–5.44); Very
low
— — 0.15 (0.01–0.88); Very
low
ARB + CCB 0 (0) — — — 0.41 (0.02–5.75); Very
low
Myocardial
infarction
ACEi
(reference)
ARB 5 (13,480) 0.98 (0.78–1.23);
Moderate
48.3%
(0.0–
81.0)
0.10 1.07 (0.89–1.28);
Moderate
ACEi + ARB 3 (9,046) 0.94 (0.81–1.10);
Moderate
0.0%
(0.0–
90.0)
0.69 1.00 (0.78–1.33);
Moderate
DRi + ACEi 1 (3,790) 1.10 (0.80–1.52); Low — — 1.35 (0.85–2.12); Low
DRi + ARB 1 (4,217) 0.68 (0.48–0.95); Low — — 0.83 (0.52–1.31); Low
DRi
+ diuretic
0 (0) — — — 0.61 (0.24–1.60); Very
low
ARB
+ diuretic
0 (0) — — — 3.06 (0.37–39.80); Very
low
(Continued)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 11 / 30
Table 2. (Continued)
Outcome Treatment Number of Studies with
Direct Comparison to ACEi
(Participants)
Pairwise Meta-analysis
OR (95% CI); Quality
of Evidence
I2 (95%
CI)
Cochran’s Q
p-Value
Network Meta-analysis
OR (95% CrI); Quality
of Evidence
ACEi + CCB 1 (206) 0.32 (0.03–3.13); Very
low
— — 0.45 (0.06–2.18); Very
low
ARB + CCB 0 (0) — — — 1.45 (0.21–14.40); Very
low
Stroke ACEi
(reference)
ARB 6 (13,522) 0.93 (0.79–1.09);
Moderate
0.0%
(0.0–
75.0)
0.96 1.01 (0.88–1.16);
Moderate
ACEi + ARB 4 (9,100) 0.86 (0.71–1.04);
Moderate
0.0%
(0.0–
85.0)
0.58 0.88 (0.71–1.08);
Moderate
DRi + ACEi 1 (3,790) 1.09 (0.76–1.56); Low — — 1.19 (0.82–1.67); Low
DRi + ARB 1 (4,217) 1.18 (0.84–1.66); Low — — 1.28 (0.91–1.78); Low
DRi
+ diuretic
0 (0) — — — 0.45 (0.16–1.20); Very
low
ACEi
+ diuretic
0 (0) — — — 1.05 (0.81–1.41); Very
low
ARB
+ diuretic
0 (0) — — — 0.75 (0.19–2.90); Very
low
ACEi + CCB 1 (206) 0.32 (0.03–3.13); Very
low
— — 0.31 (0.01–2.30); Very
low
ARB + CCB 0 (0) — — — 1.21 (0.40–3.67); Very
low
Angina pectoris ACEi
(reference)
ARB 3 (9,046) 1.03 (0.80–1.31);
Moderate
57.0%
(0.0–
88.0)
0.10 1.14 (0.98–1.37);
Moderate
ACEi + ARB 4 (10,231) 0.98 (0.79–1.21);
Moderate
48.5%
(0.0–
83.0)
0.12 1.00 (0.82–1.23);
Moderate
DRi + ACEi 0 (0) — — — NA
DRi + ARB 0 (0) — — — NA
DRi
+ diuretic
0 (0) — — — 1.32 (0.56–3.17); Very
low
ACEi
+ diuretic
0 (0) — — — NA
ARB
+ diuretic
0 (0) — — — 9.81 (0.27–564.0); Very
low
ACEi + CCB 0 (0) — — — NA
ARB + CCB 0 (0) — — — 3.25 (0.28–89.45); Very
low
Hospitalization for
heart failure
ACEi
(reference)
ARB 5 (13,480) 0.99 (0.81–1.21);
Moderate
48.1%
(0.0–
81.0)
0.10 0.99 (0.86–1.12);
Moderate
ACEi + ARB 5 (10,284) 0.91 (0.80–1.03);
Moderate
0.0%
(0.0–
79.0)
0.71 0.86 (0.73–1.01);
Moderate
(Continued)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 12 / 30
used in monotherapy and/or combination was associated with a significant reduction in angina
pectoris: ARB (1.14; 95% CrI 0.98–1.37; moderate confidence), ACE inhibitor plus ARB (1.00;
95% CrI 0.82–1.23; moderate confidence), and DR inhibitor plus diuretic (1.32; 95% CrI 0.56–
3.17; very low confidence).
Hospitalization for heart failure was reported in 5,272 of 81,373 patients from 33 studies.
With placebo as reference treatment, ARB alone or in combination with ACE inhibitor seemed
to be associated with a reduced risk of hospitalization for heart failure: 0.86 (95% CrI 0.72–
0.99; moderate confidence) for ARB and 0.75 (95% CrI 0.60–0.91; moderate confidence) for
ACE inhibitor plus ARB. Because of borderline estimates, ACE inhibitors were potentially
associated with a decreased risk of heart failure hospitalization compared with placebo: 0.86
(95% CrI 0.73–1.01; moderate confidence). Compared with ACE inhibitor as the reference
treatment (Tables 2 and S8), no RAS blocker used in monotherapy and/or combination was
associated with a significant reduction in heart failure hospitalization: ARB (0.99; 95% CrI
0.86–1.12; moderate confidence), ACE inhibitor plus ARB (0.86; 95% CrI 0.73–1.01; moderate
confidence), DR inhibitor plus ACE inhibitor (1.06; 95% CrI 0.75–1.45; low confidence), and
DR inhibitor plus ARB (0.83; 95% CrI 0.59–1.14; low confidence).
Additional Analyses and Evaluation of Models
A summary of SUCRA values with 95% CrIs by treatment and outcome is reported in S9
Table. Overall, many of these estimates were imprecise and do not allow for firm conclusions
to be drawn. The full details of the sensitivity analyses are reported in S10 Table. For each out-
come, our findings were robust when we removed studies with high risk of bias, small studies,
type 1 diabetes mellitus studies, and studies in normotensive patients. We note that the results
were not materially different when we examined the distribution of potential effect modifiers
by including covariates in network meta-regression models or when estimating the rate ratios
based on patient-years. Because of the small number of studies with available data, we were
unable to perform preplanned meta-regression analysis to evaluate the impact of history of cor-
onary disease. There was generally a better trade-off between model fit and complexity when
consistency was assumed than when it was not (S11 Table). Potential significant inconsistency
was identified in a small number of cases: for stroke (consistency model DIC = 581.58 versus
Table 2. (Continued)
Outcome Treatment Number of Studies with
Direct Comparison to ACEi
(Participants)
Pairwise Meta-analysis
OR (95% CI); Quality
of Evidence
I2 (95%
CI)
Cochran’s Q
p-Value
Network Meta-analysis
OR (95% CrI); Quality
of Evidence
DRi + ACEi 1 (3,790) 0.91 (0.69–1.20); Low — — 1.06 (0.75–1.45); Low
DRi + ARB 1 (4,217) 0.72 (0.54–0.94); Low — — 0.83 (0.59–1.14); Low
DRi
+ diuretic
0 (0) - — — 1.05 (0.65–1.70); Very
low
ACEi
+ diuretic
1 (54) 5.39 (0.25–117.77);
Very low
— — 2.02 (0.22–22.89); Very
low
ARB
+ diuretic
0 (0) — — — 0.37 (0.05–2.43); Very
low
ACEi + CCB 0 (0) — — — NA
ARB + CCB 0 (0) — — — 0.55 (0.10–2.23); Very
low
An OR > 1 favors the ACE inhibitor (i.e., fewer events occur with the ACE inhibitor than with the other active agent).
ACEi, ACE inhibitor; DRi, DR inhibitor; NA, not applicable.
doi:10.1371/journal.pmed.1001971.t002
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 13 / 30
inconsistency model DIC = 589.80) and for angina pectoris (407.85 versus 414.21). Data were
double-checked, and we could not identify any important effect modifier that differed across
comparisons.
Discussion
In our main analyses, we found no significant differences in the risk of major cardiovascular
outcome (composite of cardiovascular death, myocardial infarction, and stroke) between ACE
Table 3. Major cardiovascular outcome (composite endpoint) and all possible treatment comparisons.
Pbo/
Control
0.87
(0.75–
0.99)
ACEi
1.12
(0.81–
1.54)
1.29
(0.95–
1.76)
BB
0.81
(0.67–
0.97)
0.93
(0.79–
1.12)
0.72
(0.52–
1.02)
CCB
0.89
(0.78–
1.01)
1.02
(0.90–
1.18)
0.80
(0.59–
1.09)
1.10
(0.93–
1.30)
ARB
0.41
(0.11–
1.22)
0.47
(0.12–
1.41)
0.36
(0.09–
1.14)
0.50
(0.13–
1.51)
0.46
(0.12–
1.37)
ACEi
+ CCB
0.92
(0.69–
1.28)
1.06
(0.82–
1.47)
0.82
(0.56–
1.28)
1.14
(0.87–
1.55)
1.03
(0.79–
1.43)
2.27
(0.74–
8.95)
Diuretic
0.91
(0.66–
1.26)
1.05
(0.75–
1.52)
0.82
(0.52–
1.30)
1.13
(0.78–
1.64)
1.03
(0.73–
1.46)
2.26
(0.72–
8.94)
1.00
(0.62–
1.51)
ACEi
+ Diuretic
NA NA NA NA NA NA NA NA ARB
+ Diuretic
0.84
(0.67–
1.05)
0.97
(0.79–
1.19)
0.75
(0.53–
1.08)
1.04
(0.80–
1.33)
0.94
(0.77–
1.16)
2.07
(0.67–
7.94)
0.91
(0.63–
1.26)
0.92
(0.62–
1.35)
NA ACEi
+ ARB
1.15
(0.82–
1.58)
1.32
(0.96–
1.81)
1.02
(0.67–
1.57)
1.41
(0.99–
1.98)
1.29
(0.93–
1.75)
2.81
(0.90–
11.22)
1.25
(0.79–
1.84)
1.25
(0.78–
1.96)
NA 1.37
(0.94–
1.95)
DRi
+ ACEi
0.87
(0.62–
1.20)
1.00
(0.73–
1.38)
0.78
(0.51–
1.20)
1.08
(0.75–
1.51)
0.98
(0.71–
1.33)
2.15
(0.69–
8.45)
0.95
(0.60–
1.40)
0.95
(0.59–
1.50)
NA 1.04
(0.72–
1.48)
0.76
(0.54–
1.08)
DRi
+ ARB
0.93
(0.54–
1.66)
1.07
(0.63–
1.91)
0.83
(0.45–
1.60)
1.14
(0.67–
2.03)
1.04
(0.60–
1.86)
2.29
(0.67–
9.73)
1.00
(0.63–
1.62)
1.01
(0.54–
1.97)
NA 1.10
(0.62–
2.02)
0.81
(0.44–
1.56)
0.58
(0.00–
1.06)
DRi
+ Diuretic
NA NA NA NA NA NA NA NA 0.99
(0.52–
1.84)
NA NA NA NA ARB
+ CCB
Summary ORs and CrIs from network meta-analysis. Comparisons between treatments should be read from left to right. The treatment at the top of each
column is the reference group for comparisons in that column, while the comparator is the lower treatment in the stepladder. For major cardiovascular
outcome, OR < 1 suggests fewer outcomes with the comparator than with the reference group. Effect sizes for treatment comparisons including ARB
+ diuretic and/or ARB + CCB were generally unstable and uninterpretable owing to small sample sizes, rare events, and lack of direct evidence.
ACEi, ACE inhibitor; DRi, DR inhibitor; NA, not available; Pbo/Control, placebo or control.
doi:10.1371/journal.pmed.1001971.t003
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 14 / 30
Table 4. Renal outcomes and all-causemortality: comparisons of random effects pairwise meta-analysis with Bayesian network meta-analysis,
including confidence of assessments.
Outcome Treatment Number of Studies with
Direct Comparison to ACEi
(Participants)
Pairwise Meta-analysis
OR (95% CI); Quality of
Evidence
I2 (95%
CI)
Cochran’s Q
p-Value
Network Meta-analysis
OR (95% CrI); Quality
of Evidence
Progression of
renal disease
ACEi
(reference)
ARB 3 (10,976) 1.13 (0.89–1.42);
Moderate
56.7%
(0.0–
88.0)
0.10 1.10 (0.90–1.40);
Moderate
ACEi + ARB 2 (6,780) 1.04 (0.92–1.18);
Moderate
0.0%
(0.0–
90.0)
0.43 0.97 (0.72–1.29);
Moderate
DRi + ACEi 1 (3,790) 1.11 (0.83–1.47); Low — — 0.99 (0.65–1.57); Low
DRi + ARB 1 (4,217) 1.31 (1.01–1.71); Low — — 1.18 (0.78–1.84); Low
DRi
+ diuretic
0 (0) — — — NA
ACEi
+ diuretic
0 (0) — — — 0.98 (0.61–1.62); Very
low
ARB
+ diuretic
0 (0) — — — NA
ACEi + CCB 0 (0) — — — NA
ARB + CCB 0 (0) — — — NA
ESRD ACEi
(reference)
ARB 3 (10,976) 1.14 (0.76–1.71);
Moderate
62.1%
(0.0–
0.89)
0.07 1.09 (0.84–1.42);
Moderate
ACEi + ARB 2 (6,780) 0.99 (0.75–1.32);
Moderate
0.0%
(0.0–
90.0)
0.98 0.91 (0.63–1.27);
Moderate
DRi + ACEi 1 (3,790) 1.20 (0.80–1.80); Low — — 1.09 (0.66–1.85); Low
DRi + ARB 1 (4,217) 1.28 (0.87–1.88); Low — — 1.17 (0.72–1.93); Low
DRi
+ diuretic
0 (0) — — — NA
ACEi
+ diuretic
0 (0) — — — 1.75 (0.81–3.80); Very
low
ARB
+ diuretic
0 (0) — — — NA
ACEi + CCB 0 (0) — — — NA
ARB + CCB 0 (0) — — — NA
Doubling of
serum creatinine
ACEi
(reference)
ARB 3 (10,976) 1.20 (0.99–1.45);
Moderate
0.0%
(0.0–
90.0)
0.39 1.26 (0.97–1.79);
Moderate
ACEi + ARB 2 (6,780) 0.99 (0.78–1.25);
Moderate
0.0%
(0.0–
90.0)
0.45 0.99 (0.65–1.56);
Moderate
DRi + ACEi 1 (3,790) 0.98 (0.71–1.34); Low — — 0.93 (0.53–1.74); Low
DRi + ARB 1 (4,217) 1.33 (1.00–1.76); Low — — 1.27 (0.73–2.35); Low
DRi
+ diuretic
0 (0) — — — 1.81 (0.72–4.62); Very
low
ACEi
+ diuretic
0 (0) — — — 1.75 (0.84–3.84); Very
low
(Continued)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 15 / 30
inhibitor and either ARB or the combination of ACE inhibitor plus ARB. For the risk of pro-
gression of renal disease (composite of ESRD, doubling of serum creatinine, and all-cause mor-
tality), no significant differences were detected between ACE inhibitor and any of the
remaining therapies, such as ARB or the combination of ACE inhibitor plus ARB. Our findings
also suggest that no RAS blocker strategy was superior to ACE inhibitor with respect to all-
cause mortality, cardiovascular mortality, myocardial infarction, ESRD, or doubling of serum
creatinine. Evidence can now be drawn from data for over 103,000 adults with diabetes from
71 randomized trials with at least 1 y of follow-up, and this increased sample size can provide
more precise estimates of the cardiovascular and renal effects. We believe it is unlikely that
future trials will show clinically relevant advantages of ARBs over ACE inhibitors (and vice
versa) in preventing cardiovascular outcomes in adults with diabetes mellitus. For renal out-
comes, some may consider, however, the necessity for additional industry-independent trials
providing specific comparative data for the renoprotective effects of RAS blockade in patients
with diabetic kidney disease. Despite many systematic reviews evaluating RAS blockers [23–
37], to the best of our knowledge, this is the most comprehensive published systematic review
(and network meta-analysis) of comparative treatment effects associated with the use of RAS
blockers. It fully integrates all major cardiorenal outcomes and studies assessing ACE inhibi-
tors, ARBs, and the DR inhibitor aliskiren in adults with diabetes, thereby providing clinicians
and patients with an overall appraisal of these therapies. Compared with previous reviews in
Table 4. (Continued)
Outcome Treatment Number of Studies with
Direct Comparison to ACEi
(Participants)
Pairwise Meta-analysis
OR (95% CI); Quality of
Evidence
I2 (95%
CI)
Cochran’s Q
p-Value
Network Meta-analysis
OR (95% CrI); Quality
of Evidence
ARB
+ diuretic
0 (0) — — — NA
ACEi + CCB 0 (0) — — — NA
ARB + CCB 0 (0) — — — NA
All-cause
mortality
ACEi
(reference)
ARB 6 (13,670) 0.96 (0.86–1.08);
Moderate
12.1%
(0.0–
78.0)
0.34 1.04 (0.92–1.18);
Moderate
ACEi + ARB 4 (10,231) 1.03 (0.93–1.14);
Moderate
0.0%
(0.0–
85.0)
0.93 1.03 (0.88–1.21);
Moderate
DRi + ACEi 1 (3,790) 1.12 (0.91–1.39); Low — — 1.29 (0.96–1.72); Low
DRi + ARB 1 (4,217) 0.79 (0.63–0.98); Low — — 0.90 (0.67–1.21); Low
DRi
+ diuretic
0 (0) — — — 1.53 (0.91–2.76); Very
low
ACEi
+ diuretic
1 (53) 3.24 (0.13–83.03); Very
low
— — 0.90 (0.67–1.26); Very
low
ARB
+ diuretic
0 (0) — — — 1.19 (0.35–3.97); Very
low
ACEi + CCB 0 (0) — — — 0.47 (0.13–1.14); Very
low
ARB + CCB 0 (0) — — — 0.88 (0.31–2.47); Very
low
An OR > 1 favors the ACE inhibitor (i.e., fewer events occur with the ACE inhibitor than with the other active agent).
ACEi, ACE inhibitor; DRi, DR inhibitor; NA, not applicable.
doi:10.1371/journal.pmed.1001971.t004
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 16 / 30
diabetes mellitus [23–26] (S12–S14 Tables), our analyses incorporate between more than two
and five times the number of study participants and events. Our review was prospectively regis-
tered in PROSPERO and also adheres to the recently developed PRISMA reporting standards
for network meta-analyses (S1 Checklist) [44]. Our main conclusions differed from those
reported in recent meta-analyses in adults with diabetes, which claim different treatment
effects for classes of RAS blockers [23,24].
Unlike previous reviews, which were limited to analyzing data only from published random-
ized trials, we were able to include unpublished data from large trials and/or specific data for
subgroups of patients with diabetes [20,39–41,47–49]. Our network meta-analyses considered
Table 5. Progression of renal disease (composite outcome) and all possible treatment comparisons.
Pbo/
Control
0.92
(0.73–
1.14)
ACEi
9.00
(2.15–
41.83)
9.80
(3.37–
45.05)
BB
1.08
(0.81–
1.50)
1.17
(0.90–
1.64)
0.12
(0.03–
0.50)
CCB
1.01
(0.84–
1.24)
1.10
(0.90–
1.40)
0.11
(0.02–
0.47)
0.93
(0.69–
1.24)
ARB
NA NA NA NA NA ACEi
+ CCB
1.05
(0.69–
1.63)
1.14
(0.78–
1.74)
0.12
(0.02–
0.50)
0.98
(0.64–
1.43)
1.04
(0.67–
1.59)
NA Diuretic
0.90
(0.58–
1.39)
0.98
(0.61–
1.62)
0.10
(0.02–
0.44)
0.83
(0.48–
1.39)
0.89
(0.54–
1.42)
NA 0.85
(0.46–
1.57)
ACEi
+ Diuretic
NA NA NA NA NA NA NA NA ARB
+ Diuretic
0.90
(0.65–
1.19)
0.97
(0.72–
1.29)
0.10
(0.02–
0.42)
0.83
(0.54–
1.17)
0.89
(0.66–
1.13)
NA 0.85
(0.51–
1.34)
1.00(0.57–
1.66)
NA ACEi
+ ARB
0.91
(0.58–
1.45)
0.99
(0.65–
1.57)
0.10
(0.02–
0.45)
0.84
(0.50–
1.38)
0.90
(0.58–
1.39)
NA 0.86
(0.48–
1.56)
1.01(0.54–
1.92)
NA 1.02
(0.63–
1.71)
DRi
+ ACEi
1.08
(0.69–
1.70)
1.18
(0.78–
1.84)
0.12
(0.02–
0.53)
1.00
(0.60–
1.62)
1.07
(0.70–
1.63)
NA 1.03
(0.58–
1.84)
1.20(0.65–
2.27)
NA 1.21
(0.76–
2.02)
1.19
(0.73–
1.92)
DRi
+ ARB
NA NA NA NA NA NA NA NA NA NA NA NA DRi
+ Diuretic
NA NA NA NA NA NA NA NA 0.99(0.33–
2.98)
NA NA NA NA ARB
+ CCB
Summary ORs and CrIs from network meta-analysis. Comparisons between treatments should be read from left to right. The treatment at the top of each
column is the reference group for comparisons in that column, while the comparator is the lower treatment in the stepladder. For progression of renal
disease, OR < 1 suggests fewer outcomes with the comparator than with the reference group. Effect sizes for treatment comparisons including ARB
+ diuretic and/or ARB + CCB were generally unstable and uninterpretable owing to small sample sizes, rare events, and lack of direct evidence. There
were no studies evaluating ACE inhibitor + CCB or DR inhibitor + diuretic.
ACEi, ACE inhibitor; DRi, DR inhibitor; NA, not available; Pbo/Control, placebo or control.
doi:10.1371/journal.pmed.1001971.t005
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 17 / 30
both direct and indirect comparisons of multiple antihypertensive regimens, by including rele-
vant evidence on the comparative effects of RAS blockade. Specifically, the direct comparisons
of ACE inhibitors, ARBs, the DR inhibitor aliskiren, and their combinations are largely driven
by head-to-head comparative data, including a number of large randomized trials [39–
41,55,75]. In 2003, results of the Valsartan in Acute Myocardial Infarction Trial (VALIANT)
[39] found no significant differences between an ARB and an ACE inhibitor in all-cause mor-
tality or major cardiovascular events among patients with myocardial infarction complicated
by congestive heart failure and/or evidence of left ventricular systolic dysfunction. However,
combining an ARB and an ACE inhibitor increased the rate of adverse events without improv-
ing survival. Similarly, in 2008, the Ongoing Telmisartan Alone and in Combination with
Ramipril Global Endpoint Trial (ONTARGET) [40,41] showed no differences between an ACE
inhibitor and an ARB, alone or in combination, for major cardiovascular and renal events, but
highlighted the danger of dual blockade of RAS, reporting an increased risk of acute dialysis
and hyperkalemia in patients with vascular disease or high-risk diabetes and who were pre-
scribed an ACE inhibitor and an ARB together. To specifically assess the renal effects of dual
therapy, the Veterans Affairs Nephropathy in Diabetes (VA NEPHRON-D) trial [75] was con-
ducted in patients with diabetes mellitus and diabetic nephropathy. In line with previous stud-
ies [39–41], the VA NEPHRON-D trial showed that dual therapy with an ACE inhibitor and
an ARB yielded no significant benefit with respect to the endpoints of renal disease progres-
sion, cardiovascular disease, or all-cause mortality, but was associated with an increased risk of
adverse events compared with ARB alone.
Our findings reinforce the recommendations of current guidelines in North America and
Europe suggesting that ACE inhibitors and ARBs, as preferred antihypertensive therapies, have
generally similar effects on cardiovascular and renal outcomes [6–9]. Several points also
deserve to be mentioned. Our findings show that dual therapy with an ACE inhibitor and an
ARB had no additional beneficial effect on major cardiovascular and renal outcomes when
compared to an ACE inhibitor or ARB alone in adults with diabetes. Accordingly, given the
lack of consistent benefits on major clinical outcomes and considering the existent evidence
[22,30,39–41,75] of adverse effects in terms of hypotension, hyperkalemia, and acute kidney
injury, overall recommendations cautioning against the use of combination therapy with an
ACE inhibitor and an ARB are still valid. In addition, this review provided relevant evidence to
refute the claim raised during the last decade that ARBs may increase the risk of cardiovascular
outcomes [76–79]. For example, results of the Randomized Olmesartan and Diabetes Microal-
buminuria Prevention (ROADMAP) trial [50] and the Olmesartan Reducing Incidence of End
Stage Renal Disease in Diabetic Nephropathy Trial (ORIENT) [80] showed an increased num-
ber of cardiovascular deaths among diabetic patients randomized to the single ARB olmesar-
tan. Regarding cardiovascular safety, the US Food and Drug Administration and the European
Medicines Agency conducted safety reviews with inconclusive findings [81,82]. Overall, our
network meta-analysis found that ARBs as a class were not associated with an increased risk of
cardiovascular risk (major cardiovascular outcome, cardiovascular death, myocardial infarc-
tion, stroke, angina pectoris, or hospitalization for heart failure) when compared with ACE
inhibitors or placebo. Our evaluation showed favorable effects of ARBs as a class compared
with placebo for the risk of myocardial infarction, stroke, and hospitalization for heart failure;
potential benefits for decreasing the risk of major cardiovascular outcome and angina pectoris;
and neutral effects on cardiovascular mortality and all-cause mortality.
For the DR inhibitor aliskiren, our review identified limited evidence and did not allow us
to reach definitive conclusions for any of the cardiovascular and renal outcomes of interest.
Most trials currently available in the literature comparing therapy with DR inhibitor have
focused on surrogate outcomes and have been underpowered to provide robust estimates of
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 18 / 30
major outcomes and adverse events [5,30]. The Aliskiren Trial in Type 2 Diabetes Using Car-
diovascular and Renal Disease Endpoints (ALTITUDE) [22] trial was stopped prematurely
because the addition of the DR inhibitor aliskiren to standard therapy with RAS blockade did
not result in a decrease of cardiovascular and renal events and increased the risk of adverse out-
comes such as stroke, hyperkalemia, and hypotension. Similarly, the Aliskiren Trial on Acute
Heart Failure Outcomes (ASTRONAUT) [56] failed to show that the DR inhibitor was supe-
rior to placebo in addition to standard therapy in reducing cardiovascular death or heart failure
rehospitalization among patients hospitalized for heart failure with reduced left ventricular
ejection fraction. ASTRONAUT was stopped early, and event rates of hyperkalemia, hypoten-
sion, and renal impairment/renal failure were found to be higher in the aliskiren group. In
ASTRONAUT, all-cause mortality was significantly increased with aliskiren in patients with
diabetes but not in those without diabetes. As a consequence of these findings, regulatory
authorities recommended new contraindications and warnings for aliskiren-containing medi-
cines in patients with diabetes or kidney problems [83] and required aliskiren investigational
treatment to be withdrawn from patients with diabetes in Aliskiren Trial to Minimize Out-
comes in Patients with Heart Failure (ATMOSPHERE) [84,85], an ongoing trial of more than
7,000 patients with heart failure comparing aliskiren and the ACE inhibitor enalapril, alone
and in combination.
During the performance and reporting of our review, an additional network meta-analysis
of antihypertensive medications in diabetes and kidney disease was published and identified.
Palmer et al. [42] evaluated blood-pressure-lowering agents in 157 studies comprising 43,256
participants and concluded that no treatment prolonged survival, but ACE inhibitor and ARB
treatment, alone or in combination, were the most effective strategies against ESRD. Compared
with placebo, they found that the combination of an ACE inhibitor and an ARB seemed to pre-
vent ESRD, did not increase doubling of serum creatinine, and improved albuminuria, at the
expense of an increased risk of hyperkalemia and acute kidney injury. Although our review
indicates a fairly consistent pattern of findings with Palmer et al. [42] for ESRD and all-cause
mortality for ACE inhibitors and ARBs against placebo, our network meta-analysis provided
greater insight into the comparative effects of classes of RAS blockers on major cardiovascular
and renal outcomes, largely based on expanded long-term trial datasets (considering both
unpublished outcomes and events).
There are limitations to be noted regarding our review. First, as in other meta-analyses,
there is clinical and methodological heterogeneity in the included trials in terms of different
study characteristics and broad group populations according to the original trial designs. These
data should be viewed as reflecting real world practice treatment with different RAS blockers,
reflecting more closely the heterogeneous case mix of diabetic patients encountered in clinical
practice. Second, we used study-level data instead of individual patient data, so the small num-
ber of studies limited the sensitivity analyses that could be conducted to account for heteroge-
neity in the absence of patient-level data. Although meta-regression analyses were performed
to evaluate the effect of potential effect modifiers, the results of these analyses may be under-
powered. Given the lack of consistently reported data in each trial, we did not adjust our analy-
ses for treatment dose or adherence to therapy. Third, we followed the clinical guidelines
approach of formulating our treatment comparisons at the drug-class level. Caution should be
used when interpreting findings, considering the fact that not all single ACE inhibitor/ARB
agents are represented in the randomized trials that formed the evidence base for our network
meta-analyses. Fourth, our study does not consider costs or patient preferences for the different
treatment strategies and outcomes. Future research needs to be conducted to explore the com-
parative cost-effectiveness of RAS blockade. Fifth, we evaluated major cardiovascular and renal
outcomes using composite endpoints for a comparative effectiveness assessment. In past
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 19 / 30
decades, the composite endpoints of major cardiovascular outcome and progression of renal
disease have become standard outcomes of large randomized trials [13,40,41,75]. The power of
analyses may be increased when such composite measures are used as compared with the indi-
vidual components of these measures, since by grouping numerous types of events into a larger
category, the composite endpoint will occur more frequently than any of the individual compo-
nents. Confidence in composite endpoints depends partly on a belief that similar risk reduc-
tions apply to all the components of the composite endpoint. However, it is acknowledged that
the use of composite endpoints is frequently complicated because of poor reporting and uncer-
tain clinical relevance in many trials [86,87]. For example, our results showed no benefits of
RAS blockers over placebo in the composite endpoint of progression of renal disease (a com-
posite of the onset of ESRD, doubling of the baseline serum creatinine concentration, or death
from any cause). ESRD was the only renal outcome for which RAS blockers (ACE inhibitor or
ARB alone or in combination) were convincingly beneficial. Indeed, in this instance, ACE
inhibitors reduced the incidence of both ESRD and doubling of creatinine. No RAS blocker
reduced overall mortality in adults with diabetes. These examples highlight the challenges that
clinicians, trialists, and guideline developers face when making decisions on the basis of com-
posite endpoints, particularly when results for each single component of a composite endpoint
are not presented separately in published reports. Although we adopted reproducible defini-
tions that will enhance cross-study comparisons, this review was challenged by the heteroge-
neous reporting of events in studies. None of the trials examined and reported all of the
outcomes of interest. It is also important to emphasize the limited ability of randomized con-
trolled trials with a median duration of about 3 y to inform lifelong treatment decisions.
Finally, publication bias was not quantitatively assessed, as there were inadequate numbers of
included trials with direct comparisons to properly assess a funnel plot or use more advanced
regression-based methods.
Conclusions
Using randomized trial data and a novel evidence synthesis approach, our analyses indicate
that comparisons of different RAS blockers showed similar effects of ACE inhibitors and ARBs
on major cardiovascular and renal outcomes in adults with diabetes mellitus. Compared with
monotherapy, the combination of an ACE inhibitor and an ARB failed to provide significant
benefits in terms of major outcomes. Clinicians should discuss the balance between benefits,
costs, and potential harms with the individual patient with diabetes before starting treatment.
These findings are important for helping clinicians, health-care providers, policy-makers, and
patients with diabetes make informed decisions regarding treatment selection.
Supporting Information
S1 Checklist. PRISMA network meta-analysis checklist of items to include when reporting
a systematic review involving a network meta-analysis.
(DOCX)
S1 Fig. PRISMA flow diagram for study selection process.
(DOCX)
S2 Fig. Network geometry of all treatment comparisons included in the analyses.
(DOCX)
S1 Table. Baseline characteristics of included studies.
(DOCX)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 20 / 30
S2 Table. Study-specific outcome-level assessment of the quality of evidence (high, moder-
ate, low, very low, and not reported).
(DOCX)
S3 Table. Risk of bias and sponsorship of included studies.
(DOCX)
S4 Table. Summary characteristics of included studies.
(DOCX)
S5 Table. Reporting of outcomes in included studies.
(DOCX)
S6 Table. Number of cardiovascular events per trial and treatment comparison.
(DOCX)
S7 Table. Number of deaths and renal events per trial and treatment comparison.
(DOCX)
S8 Table. Main results for all possible treatment comparisons. Summary ORs and CrIs.
(DOCX)
S9 Table. Summary of SUCRA values with 95% CrIs, by outcome and treatment.
(DOCX)
S10 Table. Sensitivity analyses.
(DOCX)
S11 Table. Summary of model fit statistics from network meta-analysis by outcome.
(DOCX)
S12 Table. Methodological differences from previous reviews of cardiovascular and/or
renal outcomes of RAS blockade in patients with diabetes.
(DOCX)
S13 Table. Randomized controlled trials included in our systematic review versus previous
reviews.
(DOCX)
S14 Table. Randomized controlled trials excluded in our systematic review that were
included in previous reviews.
(DOCX)
S1 Text. PubMed search terms.
(DOCX)
S2 Text. List of screened systematic reviews and meta-analyses.
(DOCX)
S3 Text. Example of WinBUGS code for main analyses.
(DOCX)
S4 Text. List of included clinical trials.
(DOCX)
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 21 / 30
Acknowledgments
We thank all the trialists and sponsors/manufacturers providing unpublished outcome data or
clarifications for our network meta-analysis.
Author Contributions
Conceived and designed the experiments: FCL. Performed the experiments: FCL DMS-G GMR
DGB BRD DMCMR AZ CFH AT RTS BH. Analyzed the data: FCL BH. Contributed reagents/
materials/analysis tools: FCL DMS-G DGB BRD DMCMR AZ CFH. Wrote the first draft of
the manuscript: FCL. Contributed to the writing of the manuscript: DMS-G GMR DGB BRD
DMCMR AZ CFH AT RTS BH. Agree with the manuscript’s results and conclusions: FCL
DMS-G GMR DGB BRD DMCMR AZ CFH AT RTS BH. Responsible for study concept and
design, drafting of the manuscript, statistical analyses, and study supervision: FCL. Acquired
data: FCL DMS-G DGB BRD DMC AZMR CFH. Clinical interpretation of the data: FCL
DMS-G RTS GMR. All authors have read, and confirm that they meet, ICMJE criteria for
authorship.
References
1. Global Burden of Disease Study 2013 Collaborators. Global, regional, and national incidence, preva-
lence, and years lived with disability for 301 acute and chronic diseases and injuries for 188 countries,
1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;
386:743–800. doi: 10.1016/S0140-6736(15)60692-4 PMID: 26063472
2. GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and national age-sex specific
mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease
Study 2013. Lancet. 2015; 385:117–171. doi: 10.1016/S0140-6736(14)61682-2 PMID: 25530442
3. GBD 2013 DALYs and HALE Collaborators, Murray CJ, Barber RM, Foreman KJ, Abbasoglu Ozgoren
A, Abd-Allah F, et al. Global, regional, and national disability-adjusted life years (DALYs) for 306 dis-
eases and injuries and healthy life expectancy (HALE) for 188 countries, 1990–2013: quantifying the
epidemiological transition. Lancet. 2015; 386:2145–2191. doi: 10.1016/S0140-6736(15)61340-X
PMID: 26321261
4. Clarke PM, Glasziou P, Patel A, Chalmers J, Woodward M, Harrap SB, et al. Event rates, hospital utili-
zation, and costs associated with major complications of diabetes: a multicountry comparative analysis.
PLoS Med. 2010; 7:e1000236. doi: 10.1371/journal.pmed.1000236 PMID: 20186272
5. Sanders GD, Coeytaux R, Dolor RJ, Hasselblad V, Patel UD, Powers B, et al. Angiotensin-converting
enzyme inhibitors (ACEIs), angiotensin II receptor antagonists (ARBs), and direct renin inhibitors for
treating essential hypertension: an update. 2011 Jun [cited 25 Jul 2015]. Rockville (Maryland): Agency
for Healthcare Research and Quality. Available: http://www.ncbi.nlm.nih.gov/books/NBK61789/.
6. American Diabetes Association. Standards of medical care in diabetes—2015. Diabetes Care. 2015;
38(Suppl 1):S1–S94.
7. James PA, Oparil S, Carter BL, CushmanWC, Dennison-Himmelfarb C, Handler J, et al. 2014 evi-
dence-based guideline for the management of high blood pressure in adults: report from the panel
members appointed to the Eighth Joint National Committee (JNC 8). JAMA. 2014; 311:507–520. doi:
10.1001/jama.2013.284427 PMID: 24352797
8. Rydén L, Grant PJ, Anker SD, Berne C, Cosentino F, Danchin N, et al. ESC guidelines on diabetes,
pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD: the task force on
diabetes, pre-diabetes, and cardiovascular diseases of the European Society of Cardiology (ESC) and
developed in collaboration with the European Association for the Study of Diabetes (EASD). Eur Heart
J. 2013; 34:3035–3087. doi: 10.1093/eurheartj/eht108 PMID: 23996285
9. Dasgupta K, Quinn RR, Zarnke KB, Rabi DM, Ravani P, Daskalopoulou SS, et al. The 2014 Canadian
Hypertension Education Program recommendations for blood pressure measurement, diagnosis,
assessment of risk, prevention, and treatment of hypertension. Can J Cardiol. 2014; 30:485–501. doi:
10.1016/j.cjca.2014.02.002 PMID: 24786438
10. UK Prospective Diabetes Study Group. Efficacy of atenolol and captopril in reducing risk of macrovas-
cular and microvascular complications in type 2 diabetes: UKPDS 39. BMJ. 1998; 317:713–720. PMID:
9732338
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 22 / 30
11. Tatti P, Pahor M, Byington RP, Di Mauro P, Guarisco R, Strollo G, et al. Outcome results of the Fosino-
pril Versus Amlodipine Cardiovascular Events Randomized Trial (FACET) in patients with hypertension
and NIDDM. Diabetes Care. 1998; 21:597–603. PMID: 9571349
12. Estacio RO, Jeffers BW, Hiatt WR, Biggerstaff SL, Gifford N, Schrier RW. The effect of nisoldipine as
compared with enalapril on cardiovascular outcomes in patients with non-insulin-dependent diabetes
and hypertension. N Engl J Med. 1998; 338:645–652. PMID: 9486993
13. Heart Outcomes Prevention Evaluation Study Investigators. Effects of ramipril on cardiovascular and
microvascular outcomes in people with diabetes mellitus: results of the HOPE study and MICRO-
HOPE substudy. Lancet. 2000; 355:253–259. PMID: 10675071
14. Lindholm LH, Hansson L, Ekbom T, Dahlöf B, Lanke J, Linjer E, et al. Comparison of antihypertensive
treatments in preventing cardiovascular events in elderly diabetic patients: results from the Swedish
Trial in Old Patients with Hypertension-2. STOP Hypertension-2 Study Group. J Hypertens. 2000;
18:1671–1675.
15. Niskanen L, Hedner T, Hansson L, Lanke J, Niklason A, CAPPP Study Group. Reduced cardiovascular
morbidity and mortality in hypertensive diabetic patients on first-line therapy with an ACE inhibitor com-
pared with a diuretic/beta-blocker-based treatment regimen: a subanalysis of the Captopril Prevention
Project. Diabetes Care. 2001; 24:2091–2096. PMID: 11723089
16. Lewis EJ, Hunsicker LG, ClarkeWR, Berl T, Pohl MA, Lewis JB, et al. Renoprotective effect of the
angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes. N Engl
J Med. 2001; 345:851–860. PMID: 11565517
17. Parving HH, Lehnert H, Bröchner-Mortensen J, Gomis R, Andersen S, Arner P, et al. The effect of irbe-
sartan on the development of diabetic nephropathy in patients with type 2 diabetes. N Engl J Med.
2001; 345:870–878. PMID: 11565519
18. Brenner BM, Cooper ME, de Zeeuw D, KeaneWF, Mitch WE, Parving HH, et al. Effects of losartan on
renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy. N Engl J Med.
2001; 345:861–869. PMID: 11565518
19. Lindholm LH, Ibsen H, Dahlöf B, Devereux RB, Beevers G, de Faire U, et al. Cardiovascular morbidity
and mortality in patients with diabetes in the Losartan Intervention For Endpoint Reduction in Hyperten-
sion Study (LIFE): a randomised trial against atenolol. Lancet. 2002; 359:1004–1010. PMID: 11937179
20. Whelton PK, Barzilay J, CushmanWC, Davis BR, Iamathi E, Kostis JB, et al. Clinical outcomes in anti-
hypertensive treatment of type 2 diabetes, impaired fasting glucose concentration, and normoglycemia:
Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Arch Intern
Med. 2005; 165:1401–1409. PMID: 15983290
21. Patel A, ADVANCE Collaborative Group, MacMahon S, Chalmers J, Neal B, Woodward M, et al.
Effects of a fixed combination of perindopril and indapamide on macrovascular and microvascular out-
comes in patients with type 2 diabetes mellitus (the ADVANCE trial): a randomised controlled trial. Lan-
cet. 2007; 370:829–840. PMID: 17765963
22. Parving HH, Brenner BM, McMurray JJ, de Zeeuw D, Haffner SM, Solomon SD, et al. Cardiorenal end
points in a trial of aliskiren for type 2 diabetes. N Engl J Med. 2012; 367:2204–2213. doi: 10.1056/
NEJMoa1208799 PMID: 23121378
23. Cheng J, ZhangW, Zhang X, Han F, Li X, He X, et al. Effect of angiotensin-converting enzyme inhibitors
and angiotensin II receptor blockers on all-cause mortality, cardiovascular deaths, and cardiovascular
events in patients with diabetes mellitus: a meta-analysis. JAMA Intern Med. 2014; 174:773–785. doi:
10.1001/jamainternmed.2014.348 PMID: 24687000
24. WuHY, Huang JW, Lin HJ, LiaoWC, Peng YS, Jung KY, et al. Comparative effectiveness of renin-
angiotensin system blockers and other antihypertensive drugs in patients with diabetes: systematic
review and Bayesian network meta-analysis. BMJ. 2013; 347:f6008. doi: 10.1136/bmj.f6008 PMID:
24157497
25. Vejakama P, Thakkinstian A, Lertrattananon D, Ingsathit A, Ngarmukos C, Attia J. Reno-protective
effects of renin-angiotensin system blockade in type 2 diabetic patients: a systematic review and net-
work meta-analysis. Diabetologia. 2012; 55:566–578. doi: 10.1007/s00125-011-2398-8 PMID:
22189484
26. Nakao YM, Teramukai S, Tanaka S, Yasuno S, Fujimoto A, Kasahara M, et al. Effects of renin-angio-
tensin system blockades on cardiovascular outcomes in patients with diabetes mellitus: a systematic
review and meta-analysis. Diabetes Res Clin Pract. 2012; 96:68–75. doi: 10.1016/j.diabres.2011.11.
025 PMID: 22197527
27. Casas JP, ChuaW, Loukogeorgakis S, Vallance P, Smeeth L, Hingorani AD, et al. Effect of inhibitors of
the renin-angiotensin system and other antihypertensive drugs on renal outcomes: systematic review
and meta-analysis. Lancet. 2005; 366:2026–2033. PMID: 16338452
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 23 / 30
28. Strippoli GF, Craig M, Deeks JJ, Schena FP, Craig JC. Effects of angiotensin converting enzyme inhibi-
tors and angiotensin II receptor antagonists on mortality and renal outcomes in diabetic nephropathy:
systematic review. BMJ. 2004; 329:828. PMID: 15459003
29. Xue H, Lu Z, TangWL, Pang LW,Wang GM,Wong GW, et al. First-line drugs inhibiting the renin angio-
tensin system versus other first-line antihypertensive drug classes for hypertension. Cochrane Data-
base Syst Rev. 2015; 1:CD008170. doi: 10.1002/14651858.CD008170.pub2 PMID: 25577154
30. Makani H, Bangalore S, Desouza KA, Shah A, Messerli FH. Efficacy and safety of dual blockade of the
renin-angiotensin system: meta-analysis of randomized trials. BMJ. 2013; 346:f360. doi: 10.1136/bmj.
f360 PMID: 23358488
31. Savarese G, Costanzo P, Cleland JG, Vassallo E, Ruggiero D, Rosano G, et al. A meta-analysis report-
ing effects of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in patients
without heart failure. J Am Coll Cardiol. 2013; 61:131–142. doi: 10.1016/j.jacc.2012.10.011 PMID:
23219304
32. van Vark LC, Bertrand M, Akkerhuis KM, Brugts JJ, Fox K, Mourad JJ, et al. Angiotensin-converting
enzyme inhibitors reduce mortality in hypertension: a meta-analysis of randomized clinical trials of
renin-angiotensin-aldosterone system inhibitors involving 158,998 patients. Eur Heart J. 2012;
33:2088–2097. doi: 10.1093/eurheartj/ehs075 PMID: 22511654
33. Sciarretta S, Palano F, Tocci G, Baldini R, Volpe M. Antihypertensive treatment and development of
heart failure in hypertension: a Bayesian network meta-analysis of studies in patients with hypertension
and high cardiovascular risk. Arch Intern Med. 2011; 171:384–394. doi: 10.1001/archinternmed.2010.
427 PMID: 21059964
34. Bangalore S, Kumar S, Wetterslev J, Messerli FH. Angiotensin receptor blockers and risk of myocardial
infarction: meta-analyses and trial sequential analyses of 147 020 patients from randomised trials.
BMJ. 2011; 342:d2234. doi: 10.1136/bmj.d2234 PMID: 21521728
35. Kuenzli A, Bucher HC, Anand I, Arutiunov G, Kum LC, McKelvie R, et al. Meta-analysis of combined
therapy with angiotensin receptor antagonists versus ACE inhibitors alone in patients with heart failure.
PLoS ONE. 2010; 5:e9946. doi: 10.1371/journal.pone.0009946 PMID: 20376345
36. Baker WL, Coleman CI, Kluger J, Reinhart KM, Talati R, Quercia R, et al. Systematic review: compara-
tive effectiveness of angiotensin-converting enzyme inhibitors or angiotensin II-receptor blockers for
ischemic heart disease. Ann Intern Med. 2009; 151:861–871. doi: 10.7326/0003-4819-151-12-
200912150-00162 PMID: 20008762
37. Psaty BM, Lumley T, Furberg CD, Schellenbaum G, Pahor M, Alderman MH, et al. Health outcomes
associated with various antihypertensive therapies used as first-line agents: a network meta-analysis.
JAMA. 2003; 289:2534–2544. PMID: 12759325
38. Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, et al. Outcomes in hypertensive
patients at high cardiovascular risk treated with regimens based on valsartan or amlodipine: the
VALUE randomised trial. Lancet. 2004; 363:2022–2031. PMID: 15207952
39. Pfeffer MA, McMurray JJ, Velazquez EJ, Rouleau JL, Køber L, Maggioni AP, et al. Valsartan, captopril,
or both in myocardial infarction complicated by heart failure, left ventricular dysfunction, or both. N Engl
J Med. 2003; 349:1893–1906. PMID: 14610160
40. ONTARGET Investigators, Yusuf S, Teo KK, Pogue J, Dyal L, Copland I, et al. Telmisartan, ramipril, or
both in patients at high risk for vascular events. N Engl J Med. 2008; 358:1547–1559. doi: 10.1056/
NEJMoa0801317 PMID: 18378520
41. Mann JF, Schmieder RE, McQueenM, Dyal L, Schumacher H, Pogue J, et al. Renal outcomes with tel-
misartan, ramipril, or both, in people at high vascular risk (the ONTARGET study): a multicentre, rando-
mised, double-blind, controlled trial. Lancet. 2008; 372:547–553. doi: 10.1016/S0140-6736(08)61236-
2 PMID: 18707986
42. Palmer SC, Mavridis D, Navarese E, Craig JC, Tonelli M, Salanti G, et al. Comparative efficacy and
safety of blood pressure-lowering agents in adults with diabetes and kidney disease: a network meta-
analysis. Lancet. 2015; 385:2047–2056. doi: 10.1016/S0140-6736(14)62459-4 PMID: 26009228
43. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMAGroup. Preferred reporting items for systematic
reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009; 6:e1000097. doi: 10.1371/
journal.pmed.1000097 PMID: 19621072
44. Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension
statement for reporting of systematic reviews incorporating network meta-analyses of health care inter-
ventions: checklist and explanations. Ann Intern Med. 2015; 162:777–784. doi: 10.7326/M14-2385
PMID: 26030634
45. Sampson M, Shojania KG, McGowan J, Daniel R, Rader T, Iansavichene AE, et al. Surveillance search
techniques identified the need to update systematic reviews. J Clin Epidemiol. 2008; 61:755–762. doi:
10.1016/j.jclinepi.2007.10.003 PMID: 18586179
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 24 / 30
46. Hutton B, Tetzlaff J, Yazdi F, Thielman J, Kanji S, Fergusson D, et al. Comparative effectiveness of
monotherapies and combination therapies for patients with hypertension: protocol for a systematic
review with network meta-analyses. Syst Rev. 2013; 2:44. doi: 10.1186/2046-4053-2-44 PMID:
23809864
47. Yusuf S, Diener HC, Sacco RL, Cotton D, Ounpuu S, LawtonWA, et al. Telmisartan to prevent recur-
rent stroke and cardiovascular events. N Engl J Med. 2008; 359:1225–1237. doi: 10.1056/
NEJMoa0804593 PMID: 18753639
48. Telmisartan Randomised Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease
(TRANSCEND) Investigators, Yusuf S, Teo K, Anderson C, Pogue J, Dyal L, et al. Effects of the angio-
tensin-receptor blocker telmisartan on cardiovascular events in high-risk patients intolerant to angioten-
sin-converting enzyme inhibitors: a randomised controlled trial. Lancet. 2008; 372:1174–1183. doi: 10.
1016/S0140-6736(08)61242-8 PMID: 18757085
49. Mann JF, Schmieder RE, Dyal L, McQueen MJ, Schumacher H, Pogue J, et al. Effect of telmisartan on
renal outcomes: a randomized trial. Ann Intern Med. 2009; 151:1–10,W1-W2. PMID: 19451556
50. Haller H, Ito S, Izzo JL Jr, Januszewicz A, Katayama S, Menne J, et al. Olmesartan for the delay or pre-
vention of microalbuminuria in type 2 diabetes. N Engl J Med. 2011; 364:907–917. doi: 10.1056/
NEJMoa1007994 PMID: 21388309
51. Daly CA, Fox KM, RemmeWJ, Bertrand ME, Ferrari R, Simoons ML, et al. The effect of perindopril on
cardiovascular morbidity and mortality in patients with diabetes in the EUROPA study: results from the
PERSUADE substudy. Eur Heart J. 2005; 26:1369–1378. PMID: 15860521
52. Chaturvedi N, Porta M, Klein R, Orchard T, Fuller J, Parving HH, et al. Effect of candesartan on preven-
tion (DIRECT-Prevent 1) and progression (DIRECT-Protect 1) of retinopathy in type 1 diabetes: rando-
mised, placebo-controlled trials. Lancet. 2008; 372:1394–1402. doi: 10.1016/S0140-6736(08)61412-9
PMID: 18823656
53. Bilous R, Chaturvedi N, Sjølie AK, Fuller J, Klein R, Orchard T, et al. Effect of candesartan on microal-
buminuria and albumin excretion rate in diabetes: three randomized trials. Ann Intern Med. 2009;
151:11–20,W3-W4. PMID: 19451554
54. Sjølie AK, Klein R, Porta M, Orchard T, Fuller J, Parving HH, et al. Effect of candesartan on progression
and regression of retinopathy in type 2 diabetes (DIRECT-Protect 2): a randomized placebo-controlled
trial. Lancet. 2008; 372:1385–1393. doi: 10.1016/S0140-6736(08)61411-7 PMID: 18823658
55. Cohn JN, Tognoni G, Valsartan Heart Failure Trial Investigators. A randomized trial of the angiotensin-
receptor blocker valsartan in chronic heart failure. N Engl J Med. 2001; 345:1667–1675. PMID:
11759645
56. Gheorghiade M, BöhmM, Greene SJ, Fonarow GC, Lewis EF, Zannad F, et al. Effect of aliskiren on
postdischarge mortality and heart failure readmissions among patients hospitalized for heart failure: the
ASTRONAUT randomized trial. JAMA. 2013; 309:1125–1135. doi: 10.1001/jama.2013.1954 PMID:
23478743
57. Nakao K, Hirata M, Oba K, Yasuno S, Ueshima K, Fujimoto A, et al. Role of diabetes and obesity in out-
comes of the candesartan antihypertensive survival evaluation in Japan (CASE-J) trial. Hypertens Res.
2010; 33:600–606. doi: 10.1038/hr.2010.38 PMID: 20379187
58. Ogihara T, Saruta T, Rakugi H, Saito I, Shimamoto K, Matsuoka H, et al. Combinations of olmesartan
and a calcium channel blocker or a diuretic in elderly hypertensive patients: a randomized, controlled
trial. J Hypertens. 2014; 32:2054–2063. doi: 10.1097/HJH.0000000000000281 PMID: 24999799
59. Matsui K, Kim-Mitsuyama S, Ogawa H, Jinnouchi T, Jinnouchi H, Arakawa K, et al. Sex differences in
response to angiotensin II receptor blocker-based therapy in elderly, high-risk, hypertensive Japanese
patients: a subanalysis of the OSCAR study. Hypertens Res. 2014; 37:526–532. doi: 10.1038/hr.2014.
23 PMID: 24599010
60. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al. The Cochrane Collabora-
tion’s tool for assessing risk of bias in randomised trials. BMJ. 2011; 343:d5928. doi: 10.1136/bmj.
d5928 PMID: 22008217
61. Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, et al. GRADE guidelines: 3.
Rating the quality of evidence. J Clin Epidemiol. 2011; 64:401–406. doi: 10.1016/j.jclinepi.2010.07.015
PMID: 21208779
62. Puhan MA, Schünemann HJ, Murad MH, Li T, Brignardello-Petersen R, Singh JA, et al. A GRADE
Working Group approach for rating the quality of treatment effect estimates from network meta-analy-
sis. BMJ. 2014; 349:g5630. doi: 10.1136/bmj.g5630 PMID: 25252733
63. Rochwerg B, Alhazzani W, Sindi A, Heels-Ansdell D, Thabane L, Fox-Robichaud A, et al. Fluid resusci-
tation in sepsis: a systematic review and network meta-analysis. Ann Intern Med. 2014; 161:347–355.
doi: 10.7326/M14-0178 PMID: 25047428
64. CochranWG. The combination of estimates from different experiments. Biometrics. 1954; 10:101–129.
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 25 / 30
65. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ.
2003; 327:557–560. PMID: 12958120
66. Ioannidis JP. Integration of evidence frommultiple meta-analyses: a primer on umbrella reviews, treat-
ment networks and multiple treatments meta-analyses. CMAJ. 2009; 181:488–493. doi: 10.1503/cmaj.
081086 PMID: 19654195
67. Catalá-López F, Tobías A, Cameron C, Moher D, Hutton B. Network meta-analysis for comparing treat-
ment effects of multiple interventions: an introduction. Rheumatol Int. 2014; 34:1489–1496. doi: 10.
1007/s00296-014-2994-2 PMID: 24691560
68. Cipriani A, Higgins JP, Geddes JR, Salanti G. Conceptual and technical challenges in network meta-
analysis. Ann Intern Med. 2013; 159:130–137. doi: 10.7326/0003-4819-159-2-201307160-00008
PMID: 23856683
69. Catalá-López F, Hutton B, Moher D. The transitive property across randomized controlled trials: if B is
better than A, and C is better than B, will C be better than A? Rev Esp Cardiol. 2014; 67:597–602. doi:
10.1016/j.rec.2013.11.016 PMID: 25037537
70. Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making 2: a generalized linear
modeling framework for pairwise and network meta-analysis of randomized controlled trials. Med Decis
Making. 2013; 33:607–617. doi: 10.1177/0272989X12458724 PMID: 23104435
71. Gelman A, Rubin DB. Inference from iterative simulation using multiple sequences. Stat Sci. 1992;
7:457–472.
72. Dias S, Welton NJ, Sutton AJ, Caldwell DM, Lu G, Ades AE. Evidence synthesis for decision making 4:
inconsistency in networks of evidence based on randomized controlled trials. Med Decis Making. 2013;
33:641–656. doi: 10.1177/0272989X12455847 PMID: 23804508
73. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results
frommultiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011; 64:163–171.
doi: 10.1016/j.jclinepi.2010.03.016 PMID: 20688472
74. Dias S, Sutton AJ, Welton NJ, Ades AE. Evidence synthesis for decision making 3: heterogeneity—
subgroups, meta-regression, bias, and bias-adjustment. Med Decis Making. 2013; 33:618–640. doi:
10.1177/0272989X13485157 PMID: 23804507
75. Fried LF, Emanuele N, Zhang JH, Brophy M, Conner TA, Duckworth W, et al. Combined angiotensin
inhibition for the treatment of diabetic nephropathy. N Engl J Med. 2013; 369:1892–1903. doi: 10.1056/
NEJMoa1303154 PMID: 24206457
76. Verma S, Strauss M. Angiotensin receptor blockers and myocardial infarction. BMJ. 2004; 329:1248–
1249. PMID: 15564232
77. Strauss MH, Hall AS. Angiotensin receptor blockers may increase risk of myocardial infarction: unravel-
ing the ARB-MI paradox. Circulation. 2006; 114:838–854. PMID: 16923768
78. Hall AS, Strauss MH. More about the “ARBMI paradox”. Heart. 2007; 93:1011–1014. PMID: 17699164
79. Hobbs FD. Angiotensin receptor blockers and cardiovascular outcomes. BMJ. 2011; 342:d2193. doi:
10.1136/bmj.d2193 PMID: 21527455
80. Imai E, Chan JC, Ito S, Yamasaki T, Kobayashi F, Haneda M, et al. Effects of olmesartan on renal and
cardiovascular outcomes in type 2 diabetes with overt nephropathy: a multicentre, randomised, pla-
cebo-controlled study. Diabetologia. 2011; 54:2978–2986. doi: 10.1007/s00125-011-2325-z PMID:
21993710
81. US Food and Drug Safety Administration. FDA drug safety communication: FDA review of cardiovascu-
lar risks for diabetics taking hypertension drug olmesartan not conclusive; label updates required. 2014
Jun 6 [cited 12 Jun 2015]. Available: http://www.fda.gov/Drugs/DrugSafety/ucm402323.htm.
82. European Medicines Agency. Pharmacovigilance Risk Assessment Committee (PRAC). Minutes of the
Meeting—29–31 October 2012. 2012 Nov 29 [cited 12 Jun 2015]. Available: http://www.ema.europa.
eu/docs/en_GB/document_library/Minutes/2012/12/WC500135712.pdf.
83. European Medicines Agency. European Medicines Agency recommends new contraindications and
warnings for aliskiren-containing medicines. 2012 Feb 17 [cited 2 Jun 2015]. Available: http://www.
ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2012/02/news_detail_001446.
jsp&mid=WC0b01ac058004d5c1.
84. Krum H, Massie B, AbrahamWT, Dickstein K, Kober L, McMurray JJ, et al. Direct renin inhibition in
addition to or as an alternative to angiotensin converting enzyme inhibition in patients with chronic sys-
tolic heart failure: rationale and design of the Aliskiren Trial to Minimize OutcomeS in Patients with
HEart failuRE (ATMOSPHERE) study. Eur J Heart Fail. 2011; 13:107–114. doi: 10.1093/eurjhf/hfq212
PMID: 21169387
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 26 / 30
85. Cleland JG, Clark AL, Costanzo P, Francis DP. Diabetes, aliskiren, and heart failure: let’s bring
ASTRONAUT down to earth. Eur Heart J. 2013; 34:3097–3099. doi: 10.1093/eurheartj/eht366 PMID:
23999453
86. Montori VM, Permanyer-Miralda G, Ferreira-González I, Busse JW, Pacheco-Huergo V, Bryant D, et al.
Validity of composite end points in clinical trials. BMJ. 2005; 330:594–596. PMID: 15761002
87. Ferreira-González I, Busse JW, Heels-Ansdell D, Montori VM, Akl EA, Bryant DM, et al. Problems with
use of composite end points in cardiovascular trials: systematic review of randomised controlled trials.
BMJ. 2007; 334:786. PMID: 17403713
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 27 / 30
Editors' Summary
Background
Chronic high blood pressure can damage blood vessels, heart, and kidneys, and cause car-
diovascular and kidney (or renal) disease. Diabetes increases the risk for high blood pres-
sure. An estimated two-thirds of adults with diabetes have high blood pressure or take
blood-pressure-reducing drugs (also called antihypertensives). Because diabetes itself
increases the risk for heart and kidney diseases, controlling blood pressure is important.
Several of the drugs commonly prescribed to patients with diabetes target the renin–angio-
tensin system (RAS), a hormone system that regulates blood pressure and fluid balance.
When renal blood flow is low, kidney cells produce renin and secrete it into the blood
stream. Renin in the blood generates a protein called angiotensin I. The angiotensin-con-
verting enzyme (ACE) in the lungs subsequently converts angiotensin I to angiotensin II,
and angiotensin II docks with its partner, the angiotensin receptor. This then leads to a sig-
naling cascade that causes blood vessels to constrict and the kidneys to secrete water and
salt into the blood, resulting in increased blood pressure.
Drugs that interrupt different steps in the cascade can lower blood pressure and thereby
prevent cardiovascular and renal disease. In addition to their blood-pressure-lowering
effects, some RAS blockers have also been shown to protect heart and kidney function
through other mechanisms. RAS blockers fall into three main classes: ACE inhibitors
block the conversion of angiotensin I into angiotensin II, angiotensin receptor blockers
(ARBs) prevent angiotensin II from activating the angiotensin receptor, and direct renin
inhibitors inhibit the production of angiotensin I.
WhyWas This Study Done?
As many adults with diabetes are prescribed RAS blockers and likely need to take them or
related drugs for the rest of their lives, the question of which drug (or drug combination)
has the best results and the least side effects is important. Data from many randomized
controlled trials (RCTs, which are the most rigorous clinical tests) have generated results
that address this question. Their combined analysis, however, is complicated because the
trials often don’t compare all treatments directly, don’t use the exact same treatments in
patients with the exact same conditions, or don’t measure the exact same outcomes. None-
theless, scientists have developed methods to analyze such complex data, including a
method called network meta-analysis, which helps to integrate and synthesize diverse
results to determine the relative merits of multiple treatments. Here the researchers apply
network meta-analysis to all available data from RCTs in which adults with diabetes were
given RAS blockers, alone or in combination with other high blood pressure treatments,
and that measured cardiovascular or renal outcomes.
What Did the Researchers Do and Find?
The researchers started with a systematic search of the medical literature for RCTs that
tested RAS blockers alone or in combination in adults with diabetes. They also contacted
the sponsors of many studies to find out whether they had additional data that were not
included in the published reports. They included all trials that had at least 100 participants,
followed the participants for at least 12 months, and reported cardiovascular or renal out-
comes. In all, 71 trials that tested a total of 14 different treatment regimens in 103,120
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 28 / 30
participants met these criteria. With network meta-analysis, the researchers were able to
summarize the results of all these trials in a meaningful way. They did this by integrating
direct comparisons of RAS blockers within the same trial (where those were available)
with indirect comparisons across all trials. The primary endpoints in their analysis were
major cardiovascular outcome (a composite of cardiovascular mortality, nonfatal heart
attack, and nonfatal stroke) and progression of renal disease (a composite of end-stage
renal disease, doubling of serum creatinine [a marker of reduced kidney function], and all-
cause mortality). They also analyzed the individual cardiovascular or renal outcomes sepa-
rately, but these results carried more uncertainty because fewer patients had the individual,
compared with the combined, outcomes.
The researchers found no significant differences in the risk of major cardiovascular out-
come between ACE inhibitors and either ARBs or a combination of an ACE inhibitor plus
an ARB. Similarly, for the risk of progression of renal disease, no significant differences
were detected between ACE inhibitors and any of the remaining therapies, such as ARBs
or a combination of an ACE inhibitor plus an ARB. They also found that no RAS blocker
strategy was superior to ACE inhibitors with respect to all-cause mortality, cardiovascular
mortality, heart attacks, strokes, end-stage renal disease, or doubling of serum creatinine.
Overall, any single ACE inhibitor or ARB was equally effective as any other, and just as
effective as any drug combination.
What Do These Findings Mean?
The findings reinforce North American and European guidelines that recommend ACE
inhibitors and ARBs for antihypertensive therapy in patients with diabetes. The fact that
combinations of two drugs had outcomes no better than those of a single ACE inhibitor or
ARB—together with results from other studies that reported increased adverse effects in
patients who took drug combinations—cautions against the use of combination therapy.
The findings also contradict earlier reports suggesting that ARBs may increase the risk of
cardiovascular outcomes. There were not enough data comparing direct renin inhibitors
(a drug class that was developed more recently and therefore has been studied less) with
ACE inhibitors or ARBs to allow strong conclusions; additional research might therefore
be warranted.
The analyses here did not consider the costs of any particular drug, or any side effects that
were not relevant to the outcomes measured. These factors, together with the results here
and a patient’s specific situation, should be taken into account when doctors and their
patients discuss and decide appropriate treatments.
Additional Information
This list of resources contains links that can be accessed when viewing the PDF on a device
or via the online version of the article at http://dx.doi.org/10.1371/journal.pmed.1001971.
• The MedlinePlus has links to information on diabetes
• The American Diabetes Association provides information on kidney disease, heart
disease, and high blood pressure
• The European Society of Cardiology has guidelines on diabetes, heart disease, and high
blood pressure
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 29 / 30
• The UK National Institute for Health and Care Excellence has guidelines on renin–
angiotensin system drugs
Renin-Angiotensin System Blockade in Diabetes: Systematic Review with Network Meta-Analyses
PLOSMedicine | DOI:10.1371/journal.pmed.1001971 March 8, 2016 30 / 30