RES E A R C H A R T I C L E Open Access
Effect of mineralocorticoid receptor
antagonists on proteinuria and progression
of chronic kidney disease: a systematic
review and meta-analysis
Gemma Currie
1*†
, Alison H. M. Taylor
1†
, Toshiro Fujita
2
, Hiroshi Ohtsu
3
, Morten Lindhardt
4
, Peter Rossing
4,5,6
,
Lene Boesby
7
, Nicola C. Edwards
8
, Charles J. Ferro
8
, Jonathan N. Townend
8
, Anton H. van den Meiracker
9
,
Mohammad G. Saklayen
10
, Sonia Oveisi
11
, Alan G. Jardine
1
, Christian Delles
1
, David J. Preiss
12
and Patrick B. Mark
1
Abstract
Background: Hypertension and proteinuria are c ritically i nvolved in the progression of chronic kidney disease.
Despite treatment with renin angioten sin system inhibition, kidne y f unction declines in many patients.
Aldosterone excess is a risk factor for progression of kidney disease. Hyperkalaemia is a concern with the use
of mineralocorticoid receptor antagonists. We aime d to determine whether the renal protective benefits of
mineralocorticoid antagon ists outweigh the risk of hyperkalaemia associated with this treatment in patients
with chronic kid ney disease.
Methods: We conducted a meta-analysis investigati ng ren oprotective effects and risk of hyperkalaemia in
trials of mineralocorticoi d receptor antagonists in chron ic kidn ey disease. Trials were ident ified fr om MEDLINE
(1966–2014), EMBASE (1947–2014) and the Cochrane Clinical Trials Database. Unpublished summary data were
obtained from investigators. We included randomi sed control led trial s, and the first period of randomised
cross over trials lasting ≥4 weeks in adults.
Results: Nineteen trials (21 study groups, 1 646 patients) were included. In random effects meta-analysis,
addition of mineralocorticoid receptor antagonists to renin angiotensin system inhibition resulted in a reduction from
baseline in systolic blood pressure (−5.7 [−9.0, −2.3] mmHg), diastolic blood pressure (−1.7 [−3.4, −0.1] mmHg) and
glomerular filtration rate (−3.2 [−5.4, −1.0] mL/min/1.73 m
2
). Mineralocorticoid receptor antagonism reduced weighted
mean protein/albumin excretion by 38.7 % but with a threefold higher relative risk of withdrawing from the trial due to
hyperkalaemia (3.21, [1.19, 8.71]). Death, cardiovascular events and hard renal end points were not reported in sufficient
numbers to analyse.
Conclusions: Mineralocorticoid receptor antagonism reduces blood pressure and urinary protein/albumin excretion
with a quantifiable risk of hyperkalaemia above predefined study upper limit.
* Correspondence: gemma.currie@glasgow.ac.uk
†
Equal contributors
1
Institute of Cardiovascular and Medical Sciences, British Heart Foundation
Glasgow Cardiovascular Research Center, 126 University Place, Glasgow, UK
Full list of author information is available at the end of the article
© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Currie et al. BMC Nephrology (2016) 17:127
DOI 10.1186/s12882-016-0337-0
Background
Chronic kidney disease (CKD) is associated with risk of
premature cardiovascular (CV) disease and death [1–6].
Hypertension (HTN) is the major modifiable risk factor
for CKD progression and is associated with development
of left ventricular hypertrophy (LVH) and proteinuria,
both predictors of CV mortality [3, 7]. In CKD patients
with proteinuria and/or HTN renin angiotensin system
(RAS) inhibitors are commonly prescribed as these
agents have been shown to reduce proteinuria and delay
CKD progression through a combination of BP dependent
and independent mechanisms. Despite this, patients still
progress to end stage renal disease (ESRD) or die from
CV events [8–14].
There is renewed interest in aldosterone as a medi-
ator of C V and renal disea se, beyond its BP effe ct
resulting in an enthusiasm for usin g mineralocorticoid
rece ptor antagonists (MRA) to minimised prot einuria
and delay CKD progression. A 2009 meta-analysis,
updated in 2014 demonstrated that addition of MRA
to RAS blockade reduced BP and prote inuria in CKD
[15, 16]. The beneficial effects on outcomes w ere con-
founded by increased risk of hyperkalaemia, a fac tor
limiting MRA prescribing in CKD [17–19]. Similar
findings were described in a more re cent meta-
analysis on the cardiovascular actions of MRAs in
CKD [20]. Howe ver, the conclusions of these analyses
are drawn from mainly published proteinuria data
only, derived by consolidating disparate urinary pro-
tein excretion measures used in different trials (variably
reported a s protein or albumin excretion in spot
samples or 24 h colle ctions). Furthermore, in some
studies in these meta-analyse s, the MRA effect is
impossible to d issociat e from that of additional anti-
hypertensives c o-administered with MRA in the inter-
vention arm.
In the past 6 years severa l studies of effe cts o f se-
lective and non-selective MR A s in CKD have been
published [21–34]. We performed an updated meta-
analysis of this treatment strategy using summarised
unpublished data where possible, as well as including
data from 3 studies which were not considered in the
previous publication [28, 30,34].Thisisparticularly
relevant as one of these studies focu ssed on patients
with CKD stage 3–4 [30] an area where evidence for
use of this strategy is lacking, and also be cause the
resultant number of participants included exceeds that
of the previous publications. We focused on change
in urinary protein/albumin excretion, progression of
CKD and risk of hyperkalaemia whilst additionally
collecting hard clinical endpoint s where these data
were available. Our aim was to determine whether
renoprotective benefits of MRAs outweigh risk of
hyperkalaemia associated with this treatment.
Methods
Literature search was performed independently by two
authors (GC, AT) using PubMed (1966 - 1st Dec 2014),
EMBASE (1947 - 1st Dec 2014) and the Cochrane
Clinical Trials Database. Search strategy is shown in
Appendix 1 (see Additional file 1).
Trial type
We analysed randomised controlled trials in humans
published in English of both selective and non-selective
MRAs performed in CKD stage 1–5 where MRA was
compared to placebo or open label trials where MRA
was additional therapy compared to the non-MRA arm.
Trials were eligible if MRA was used alone or combined
with ACE-I alone, ARB alone, or both ACE-I and ARB,
performed in CKD patients or for prevention of CKD
progression. The first period of randomised crossover
trials was also considered eligible. Trials directly com-
paring MRA to other antihypertensive agents were
excluded.
Participants
Trials enrolling patients with CKD stages 1–5 not re-
quiring RRT, with albuminuria or proteinuria were
included [11]. Our search included haemodialysis, peri-
toneal dialysis and renal transplantation ensure all ap-
propriate trials were identified but RRT studies were
excluded from the main analysis to minimise the
confounding effect of dialysis on blood pressure and
potassium.
Interventions
Trials using selective and non-selective MRAs with pla-
cebo, ACE-I, ARB or both were included. Minimum trial
duration was 4 weeks.
Outcome measures
Analysis plan included effects of MRAs on the following
measures:
a) End of treatment urinary albumin or protein
excretion (24-h proteinuria or albuminuria ,
or urinary protein ratio or albumin creatinine
ratio)
b) End of treatment renal function: serum creatinine
(μmol/L); eGFR (ml/min/1.73 m
2
); creatinine
clearance (ml/min); doubling of serum creatinine;
need for RRT. When several mea sures of kidney
function were available this wa s meta-analysed
following the hierarchy- isot opic GFR, crea tinine
clearance from 24 h urine collection, eGFR (MDRD
or CKD-EPI formulae), formula estimated creatinine
clearance (Cockcroft-Gault)
c) End of treatment SBP and DBP (mmHg)
Currie et al. BMC Nephrology (2016) 17:127 Page 2 of 14
d) Hyperkalaemia (serum potassium above trial
protocol limit)
e) Death, need for RRT, cardiovascular events.
Data collecti on
The search strategy in Additional file 1: Appendix 1
identified titles and abstract s. Both reviewers a ssessed
titles and abstracts independently, discarding those
not meeting inclusion criteria. Full text s of remaining
trials were independently a ssessed. A third author
(PM) settled discrepa ncies. Data extraction was per-
formed using spe cific extraction forms (Appendix 2,
see Additional file 1). Original authors were contacted
tor further information.
Assessment of risk of bias
Two independent reviewers (GC, AT) assessed trial
quality using the Cochrane Collaboration risk of bias as-
sessment tool [35]. Items assessed were: adequate se-
quence generation; allocation concealment; blinding of
participants, trial personnel and outcome assessors;
reporting of incomplete outcome data; suggestion of se-
lective outcome reporting and intention to treat analysis
(Additional file 1: Table S1).
Statistical analysis
Random effe cts meta-analysis was performed for con-
tinuous and categorical outcomes. For continuous out-
comes, weighted mean differences were performed using
two approaches: in analyses of continuous data, final
visit results were compared for treatment and control
arms after verifying by meta-analysis that baseline data
for the relevant outcome were no different between trial
arms; second, where sufficient data were available ,
change in weighted mean difference from ba seline to
final visit was calculated by meta-analysis. For categor-
ical outcomes, we calculated risk ratios (RRs) as ratio of
cumulative incidence and 95 % CIs from available data
for trial participants at baseline and for those developing
the outcome of interest. Random effects models were
used as preferable approach to manage between-trial
heterogeneity introduced by analysing differing trial pop-
ulations. A s standard deviations were unavailable for all
measures of change in albuminuria and proteinuria, per-
centage change was analysed using weighted means and
weighted standard deviations across trials in exploratory
analyses. Statistical heterogeneity across studies was
quantified using I
2
statistics, providing measure of pro-
portion of overall variation attributable to between-trial
heterogeneity, with p < 0.10 considered significant. We
assessed publication bias with funnel plots and Egger
tests , for the most commonly reported outcome -
SBP. Analyses were conducte d using Stata version 13
(StataCorp, College Station, Texas).
Results
Literature search and trial characteristics
Search results
Search of PubMed, EMBASE and the Cochrane database
identified 299 citations. 243 were excluded (overlapping
searches; non-randomised trials; trials e valuating inter-
ventions not included in this review) (Fig. 1). Full text
assessment of 56 articles resulted in selection of 19 eli-
gible trials including 1646 patients [22–31, 34, 36–43],
more than were included in the previously published
meta-analyses [15, 16, 20]. Authors were contacted for
unpublished data; we obtained supplemental sum-
marised results for ten trials [23, 26–30, 39, 40, 42, 43].
These were whole group mean and standard deviations
pre- and post- intervention for the outcomes of interest,
allowing a more complete and precise analysis.
One trial compared ACE-I and ARB (or placebo) and
Spironolactone (or placebo), (four groups in total) [37]
and another compared two doses of Eplerenone against
placebo, (three groups in total) [38]. All comparable
arms of both trials were included for analysis. Therefore,
19 trials of 21 study groups were included for analysis.
Fourteen trials (889 patients) compared Spironolactone
plus ACE-I or ARB with ACE-I or ARB alone; and 5 tri-
als (757 patients) compared Eplerenone plus ACE-I or
ARB to ACE-I or ARB alone. One trial focused on triple
RAS blockade vs dual RAS blockade [41]. We did not
identify any trials comparing MRA to placebo in the
absence of ACE-I/ARB treatment.
Trial characteristics
Included were five randomised placebo controlled trials,
seven randomised controlled trials (treatment compared
to standard care) and seven randomised crossover trials.
Six trials included patients with non-diabetic CKD, eight
trials focused on diabetic nephropathy, 3 included both
diabetic and non-diabetic CKD and 2 included patients
with CKD and HTN.
Study duration was between 8 and 52 weeks. Sample
size was small (n = 18 to 359). No trials were powered to
measure mortality or long-term renal outcomes. Dose of
Spironolactone was 25 mg/day in most trials, while two
trials used 25–50 mg [24, 42]. Eplerenone dose ranged
from 25 to 100 mg.
The primary endpoint in the majority of trials was
change in urinary protein/albumin excretion, although
there was significant heterogeneity in measures used.
Trials reported change in urine protein:cre atinine ratio
(PCR); albumin:creatinine ratio (ACR) or 24-h urine
protein/albumin excretion. In three trials change in
protein/albumin excretion wa s a se condary outcome
measure where BP, pulse wave velocity and left ven-
tricular ma ss index (LVMI) respe ctively were p rimary
endpoints [23, 24, 30].
Currie et al. BMC Nephrology (2016) 17:127 Page 3 of 14
In trials reporting estimated glomerular filtration rate
(eGFR) calculation method included Modification of
Diet in Renal Disease (MDRD); Cockcroft Gault cre-
atinine clearance (CrCl) and Chronic Kidney Disease
Epidemiology Collaborat ion (CKD-EPI) too ls. In three
trial s GFR was measur ed by 5 1Cr-EDTA [26, 39, 40].
Characteristics of participants and inte rventions of in-
cluded trials are listed in Table 1.
Trial quality
Trial quality assessed using the Cochrane Collaboration
tool was variable (Additional file 1: Table S1). Sequence
generation was adequately described in 9 (50 %) trials.
Allocation concealment wa s adequate in 8 (44 %) trials.
Participants and investigators were blinded in 12 (67 %)
trials and intention to treat analysis was performed in 2
(11 %) trials. Dropout s were adequately accounted for in
16 (84 %) trials and did not differ between treatment
and control groups.
Trial outcomes
Baseline data from included studies is shown in Additional
file 1: Table S2. Meta-analysis of baseline data for all
outcomes of interest showed these were balanced,
confirming that end-of-trial meta-analysis was appro-
priate (Additional file 1: Table S3).
Effect of treatment on blood pressure
There was a significant change in systolic blood pressure
(SBP) at final visit (−5.7 mmHg) and in the nine trials
where change from baseline (−3.3 mmHg) was available,
with addition of MRA to ACE-I and/or ARB in compari-
son to ACE-I and/or ARB alone (Fig. 2a). Addition of
MRA to ACE-I and/or ARB also led to a significant
change in end of trial diastolic blood pressure (DBP)
(−1.7 mmHg) and change from baseline to final visit
(−2.8 mmHg) in comparison to ACE-I and/or ARB alone
(Table 2).
Effect of treatment on renal excretory function
There was a small, non-significant increase in end-of-
trial serum creatinine (3.8 μmol/L) with addition of
MRA to ACE-I and/or ARB compared to ACE-I and/or
ARB alone (Table 3). Addition of MRA to ACE-I and/or
ARB led to similar change in eGFR (−2.7 mL/min/
1.73 m
2
) and CrCl (−2.5 mL/min) compared to ACE-I
and/or ARB alone with little heterogeneity between in-
cluded study groups respectively (I
2
=0 %, P = 0.696 and
I
2
=0%,P = 0.727) (Fig. 2b).
Effect of treatment on urinary albumin/protein excretion
In trials reporting ACR, there was a non-significant
change (−10.91 mg/mmol creatinine) with addition of
Fig. 1 Study flow chart
Currie et al. BMC Nephrology (2016) 17:127 Page 4 of 14
Table 1 Summary of populations and interventions in included studies. Data are mean ± SD or median (IQR)
Study Kidney disease No. of patients
included
Intervention group Control group Co-intervention Study
duration
Baseline eGFR
(ml/min/1.73 m
2
)
Endpoints
Abolghasmi 2011 [
24] CKD with resistant
hypertension
41 Spironolactone
25–50 mg
Placebo multi-drug regime
including ACE-I+/−ARB
12 weeks Not available BP, potassium, creatinine,
urinary sodium
Ando 2014 [
28] CKD with hypertension 314 Eplerenone
50 mg
Placebo ACE-I+/−ARB of at least
8 weeks duration
1 year Treatment 67.7 ± 14.3
Control 68.6 ± 13.6
UACR, creatinine, eGFR,
urinary L-FABP, 24 h
urinary sodium, incidence
of cerebrovascular and
cardiovascular events
Bianchi 2006 [
36] Non-diabetic CKD
(idiopathic GN)
165 Spironolactone
25 mg
ACE-I+/−ARB ACE-I+/−ARB 1 year Treatment 62.4 ± 21.9
Control 62.2 ± 19.0
24 h urinary protein, BP,
creatinine, eGFR potassium
Boesby 2011 [
29]
(XO)
Non-diabetic CKD 40 Eplerenone
25–50 mg
multi-drug regime
including ACE-I+/−ARB
multi-drug regime
including ACE-I+/−ARB
8 weeks 59 ± 26 24 h urinary albumin, BP,
potassium, creatinine
clearance
Boesby 2013 [
30] Diabetic and
non-diabetic CKD
26 Eplerenone
25–50 mg
ACE-I+/−ARB ACE-I+/−ARB 24 weeks 36 ± 10 cfPWV, AIx, AASI, 24 h
urinary albumin
Chrysostomou 2006
a
[37]
Diabetic and
non-diabetic CKD
41 Spironolactone
25 mg
Placebo as ARB;
Placebo as
Spironolactone
ACE-I alone; ACE-I +
ARB
3 months Not available 24 h urinary protein, BP,
creatinine, creatinine
clearance, potassium
Edwards 2009 [
23] Non-diabetic CKD with
no renovascular
diagnosis
112 Spironolactone
25 mg
Placebo ACE-I/ARB 36 weeks Treatment 49 ± 12
Control 53 ± 11
LVMI, cfPWV, aortic
distensibility, AIx, BP
Epstein 2006+ [
38] Diabetic nephropathy 359 Eplerenone
50 mg or 100 mg
Placebo ACE-I 12 weeks ACE ± EPL 50
73 (62.1–83.6)
ACE ± EPL 100
75 (62.8–85.9)
Control
74 (60.5–82.2)
UACR, potassium, BP, eGFR
Guney 2009 [
25] Non-diabetic CKD 24 Spironolactone
25 mg
ACE-I+/−ARB ACE-I+/−ARB 6 months Treatment 63.0 ± 22.71
Control 56.3 ± 35.6
UPCR, urinary TGF-β1,
eGFR, creatinine,
potassium, BP, aldosterone
Mehdi 2009 [
22] Diabetic nephropathy 81 Spironolactone
25 mg
Placebo or ARB ACE-I 48 weeks Not available UACR, BP, creatinine
clearance, potassium
Nielsen 2012 [
26]
(XO)
Diabetes with
microalbuminuria
21 Spironolactone
25 mg
Placebo ACE-I/ARB 60 days Not available 24 h urinary albumin, BP,
GFR, urinary L-FABP,
urinary NGAL, urinary
KIM-1
Rossing 2005 [
39]
(XO)
Diabetic nephropathy 20 Spironolactone
25 mg
Placebo ACE-I+/−ARB 8 weeks Not available 24 h urinary albumin,
BP, GFR
Saklayen 2008 [
43]
(XO)
Diabetic nephropathy 24 Spironolactone
25–50 mg
Placebo ACE-I/ARB 3 months Treatment 61.9 ± 23.4
Control 54.4 ± 20.1
BP, creatinine, potassium,
UPCR
Schjoedt 2005 [
40]
(XO)
Diabetic nephropathy 20 Spironolactone
25 mg
Placebo ACE-I+/
−ARB 2 months Not available 24 h urinary albumin,
BP, GFR
Currie et al. BMC Nephrology (2016) 17:127 Page 5 of 14