this article is protected by copyright. all rights reserved. · we searched medline, embase, the...
TRANSCRIPT
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/dom.14108
Title page
Manuscript title: Glucagon-like peptide 1 receptor agonists and sodium glucose co-transporter 2
inhibitors as combination therapy for type 2 diabetes. A systematic review and meta-analysis
Short running title: GLP-1 RA and SGLT2i combination therapy for type 2 diabetes
Authors
Chrysanthi Mantsiou MD1; Thomas Karagiannis MD, MSc, PhD1,2; Panagiota Kakotrichi MD1;
Konstantinos Malandris MD, MSc1; Ioannis Avgerinos, MD, MSc1,2; Aris Liakos MD, MSc, PhD1,2;
Apostolos Tsapas MD, PhD, MSc(Oxon)1,2,3; Eleni Bekiari MD, MSc, PhD1,2
Author affiliations
1 Clinical Research and Evidence-Based Medicine Unit, Second Medical Department, Aristotle
University of Thessaloniki, Thessaloniki, Greece
2 Diabetes Centre, Second Medical Department, Aristotle University of Thessaloniki, Thessaloniki,
Greece
3 Harris Manchester College, University of Oxford, Oxford, United Kingdom
Corresponding author
Chrysanthi Mantsiou, Clinical Research and Evidence-Based Medicine Unit, Second Medical
Department, Aristotle University of Thessaloniki, Konstantinoupoleos 49, 54642 Thessaloniki,
Greece. Email: [email protected]
This article is protected by copyright. All rights reserved.
Abstract
Aim To assess the efficacy and safety of combination therapy with a glucagon-like peptide 1 receptor
agonist (GLP-1 RA) and a sodium glucose co-transporter 2 inhibitor (SGLT2i) in patients with type 2
diabetes.
Methods We searched Medline, Embase, the Cochrane Library, and grey literature sources until
December 2, 2019 for randomized controlled trials in adults with type 2 diabetes assessing the
combination of GLP-1 RA and SGLT2i, either as co-initiation therapy or as add-on to each other,
against placebo or an active comparator. The primary outcome was change in HbA1c. Secondary
outcomes included change in body weight, blood pressure, and eGFR, and incidence of severe
hypoglycemia, all-cause mortality, cardiovascular mortality, myocardial infarction, stroke and
hospitalization for heart failure. We pooled data using random effects meta-analyses.
Results Seven trials (1913 patients) were eligible. Compared to GLP-1 RA, GLP-1 RA/SGLT2i
combination therapy was associated with greater reduction in HbA1c (weighted mean difference -
0.61%, 95% CI -1.09 to -0.14%, four studies), body weight (-2.59 kg, -3.68 to -1.51 kg, three studies),
and systolic blood pressure (-4.13 mmHg, -7.28 to -0.99 mmHg, four studies). Compared to SGLT2i,
GLP-1 RA/SGLT2i combination therapy reduced HbA1c (-0.85%, -1.19 to -0.52%, six studies) and
systolic blood pressure (-2.66 mmHg, -5.26 to -0.06 mmHg, six studies), but not body weight (-1.46
kg, -2.94 to 0.03 kg, five studies). After excluding data for one trial that had a considerably longer
duration than the remaining studies, body weight was also reduced versus SGLT2i (-1.79 kg, -2.99 to
-0.59 kg, five studies). Combination therapy did not increase the incidence of severe hypoglycemia.
Data for mortality and cardiovascular outcomes were scarce.
Conclusions GLP-1 RA/SGLT2i combination therapy seems to reduce HbA1c, body weight and
systolic blood pressure without increasing the risk for severe hypoglycemia compared to either GLP-1
This article is protected by copyright. All rights reserved.
RA or SGLT2i. No robust conclusions can be made regarding long-term effectiveness or effect on
cardiovascular outcomes.
This article is protected by copyright. All rights reserved.
Introduction
Most patients with type 2 diabetes will eventually require a combination of at least two antidiabetic
agents with different mechanisms of action to achieve glycemic control. Glucagon-like receptor
agonists (GLP-1 RAs) enhance insulin secretion and inhibit glucagon secretion, while sodium-glucose
cotransporter-2 inhibitors (SGLT2is) promote urinary glucose excretion.1,2 In addition, both drug
classes have a favorable effect on body weight and blood pressure, while cardiovascular outcome
trials have demonstrated that specific agents within each class can reduce mortality or incidence of
cardiovascular events.3,4 As a result, both GLP-1 RAs and SGLT2is have a prominent place in the
American Diabetes Association and the European Association for the Study of Diabetes
(ADA/EASD) consensus report and the guidelines issued by the European Society of Cardiology
(ESC) in collaboration with EASD.5,6 However, it is unclear whether these beneficial effects of
individual GLP-1 RAs and SGLT2is are retained or even enhanced in patients who are treated with a
combination of a GLP-1 RA and an SGLT2i.
Previous meta-analyses have assessed the effect of combination therapy with a GLP-1 RA and an
SGLT2i.7–10 However, either they focused solely on comparisons against SGLT2is and not versus
GLP-1 RAs and did not incorporate data from all relevant randomized controlled trials (RCTs)
published to date, or they did not take into consideration the pattern of combination therapy in each
trial (sequential administration or co-initiation) in their analysis.7–10 Sequential administration aims to
better patient adherence and minimization of adverse events, while co-initiation therapy can lead to
earlier achievement of therapeutic goals by targeting multiple pathogenic mechanisms
simultaneously, as demonstrated in the VERIFY (Vildagliptin Efficacy in combination with
metfoRmIn For earlY treatment of type 2 diabetes) study.11,12
We aimed to systematically retrieve, critically appraise and synthesize all currently available evidence
evaluating the efficacy and safety of combination therapy with a GLP-1 RA and an SGLT2i in
This article is protected by copyright. All rights reserved.
patients with type 2 diabetes, by means of an up-to-date systematic review and meta-analysis of
RCTs.
Methods
This systematic review is reported according to the PRISMA (Preferred Reporting Items for
Systematic reviews and Meta-Analyses) guidelines13 and was conducted according to a registered
protocol (PROSPERO identifier: CRD42020152355).
Data Sources
We identified eligible studies by searching MEDLINE (via PubMed), EMBASE (via Ovid) and the
Cochrane library without any language restrictions until December 2, 2019 (Supplementary Table
S1). We also searched grey literature sources, including clinical trial registries (clinicalTrials.gov,
clinicaltrialsregister.eu), manufacturers’ trial registries, and the abstracts of the annual ADA and
EASD meetings from 2013 to 2019.
Study Selection
We included RCTs in adults with type 2 diabetes that compared the combination of a GLP-1 RA and
an SGLT2i (GLP-1 RA/SGLT2i) against placebo or an active control (including individual GLP-1
RAs or SGLT2is) and reported data for at least one of the prespecified outcomes of interest. Records
retrieved from electronic databases were imported in an online reference management software. After
deduplication, three independent reviewers (CM, PK, KM) screened all records at title and abstract
level and then assessed the full text of eligible records. Any disagreements were resolved by
consultation of a fourth reviewer (TK). Two reviewers (CM and PK) juxtaposed the results from the
grey literature search against the search results from the three electronic databases.
Data Extraction
This article is protected by copyright. All rights reserved.
Data extraction from eligible studies was performed independently by two reviewers (CM and PK)
using a predesigned form and any conflicts were resolved by a third reviewer (TK). Multiple reports
of the same study were collated to maximize information yield. In case that numerical data were
presented only in graphs, they were extracted using relevant software
(http://plotdigitizer.sourceforge.net).14 If a study comprised multiple arms with different doses, we
combined data from all approved doses using appropriate methodology.15 Missing data were
requested by contacting corresponding authors.
For each eligible trial, we extracted data for study characteristics, participants’ baseline characteristics
and prespecified outcomes. Our primary outcome was the change in HbA1c from baseline. Secondary
efficacy outcomes included the change from baseline in body weight, systolic blood pressure,
diastolic blood pressure, and estimated Glomerular Filtration Rate (eGFR). We also extracted data for
all-cause mortality and cardiovascular mortality, and number of patients who experienced at least one
event of severe hypoglycemia (as defined in each study), myocardial infarction, stroke and
hospitalization for heart failure.
Risk of Bias Assessment and Strength of Evidence
Two independent reviewers (CM and PK) used the revised Cochrane Collaboration’s Risk of Bias
Tool (ROB) version 2.0 to assess risk of bias within studies for change in HbA1c, body weight, and
systolic blood pressure.16 Any disagreements were resolved in consensus with a third reviewer (TK).
We used the Grading of Recommendations Assessment, Development and Evaluation (GRADE)
approach to rate the certainty of evidence in effect estimates for the three aforementioned outcomes.17
We planned to assess publication bias by means of visual interpretation of a funnel plot for the
primary outcome and with Egger’s test.18
Data Synthesis
This article is protected by copyright. All rights reserved.
For continuous outcomes we calculated weighted mean differences (WMDs) and 95% confidence
intervals (CIs) applying an inverse variance random effects model. For dichotomous outcomes we
calculated odds ratios (ORs) along with 95% CIs using the random effects Mantel-Haenzel approach.
In cases of comparisons with zero events in one or both arms, we applied a treatment arm continuity
correction based on the inverse of the opposite arm’s size.19 We assessed statistical heterogeneity with
the I2 statistic, with I2 values above 60% representing high heterogeneity.20 We performed separate
analyses based on type of comparator (either GLP-1 RA or SGLT2i alone). In addition, we performed
prespecified subgroup analyses, based on whether the two drugs were given as simultaneous initiation
(co-initiation) or as sequential administration (add-on to each other).
If a study reported outcome at different time points, in our main analyses we utilized data from the
report with the longest duration of intervention. Of note, we identified only one study (DURATION-8
trial) reporting data at different time points (at 28, 52 and 104 weeks),21–23 whereas the remaining
studies had a short duration (ranging between 12 and 30 weeks). Therefore, for the change in HbA1c,
body weight and systolic blood pressure, we conducted a post-hoc secondary analysis utilizing short-
term (at 28 weeks)23 instead of long-term (at 104 weeks)21 data from the DURATION-8 trial. In
addition, we conducted a post-hoc sensitivity analysis for the change in HbA1c excluding the
DURATION-8 trial21, in which patients’ baseline HbA1c was considerably higher than that of the
remaining studies. Moreover, in an additional sensitivity analysis we excluded one trial,24 for which
we extracted continuous outcome data from graphs using appropriate software. Finally, based on the
findings from individual three-arm trials assessing co-initiation combination therapy, we speculated
whether the effect of combination therapy on HbA1c, body weight and systolic blood pressure, was
likely to be additive, more than additive or less than additive, by comparing it to the sum of the effect
in the two trial arms allocated to individual GLP-1 RA and SGLT2i. All analyses were done using
RStudio version 1.2.5001 and the R package meta (version 4.9-7).
This article is protected by copyright. All rights reserved.
Results
Search Results and Study Characteristics
Seven RCTs with 1913 patients were included in the systematic review and meta-analysis (Figure
1).21,22,31–36,23–30 Of these, two studies were retrieved as conference abstracts and one study was
identified in clinicaltrials.gov (NCT02964247).25,28 Of note, for one of these conference abstracts and
for the trial from clinicaltrials.gov we utilized post-hoc data from their respective journal article
which were published shortly after our search date.33,36 The characteristics of included studies and
participants’ baseline characteristics are shown in Table 1. All studies had a parallel-group design.
Five studies were double-blind,22,24,33–35 one was open-label36, while blinding status was not reported
in one study28. Duration of intervention ranged between 12 and 30 weeks in all studies, except for the
DURATION-8 trial which reported outcome data at 2823, 5222 and 10421 weeks.
Three studies22,28,36 evaluated the combination of GLP-1 RA/SGLT2i as simultaneous initiation
therapy versus isolated GLP-1 RA and SGLT2i. Three studies33–35 compared a GLP-1 RA versus
placebo as add-on therapy in patients already treated with an SGLT2i, while one study was a post-hoc
subgroup analysis24 of the CANVAS (CANagliflozin cardioVascular Assessment Study) trial
programme37 comparing an SGLT2i versus placebo as add-on therapy in patients already treated with
a GLP-1 RA. In all RCTs, patients continued receiving their background anti-diabetic treatment
which was mostly metformin or metformin plus a sulphonylurea. Across included trials, mean
baseline HbA1c ranged between 8.0% and 8.2% in all studies, except for the DURATION-8 trial in
which mean HbA1c at baseline was 9.3%. Mean body weight, body mass index and systolic blood
pressure at baseline ranged from 90.9 kg to 91.7 kg, 31.9 kg/m2 to 37.4 kg/m2 and 127.9 mmHg to
136.7 mmHg, respectively. Across all trials, mean participant’s age ranged between 52.3 and 61.0
years (Table 1).
This article is protected by copyright. All rights reserved.
Risk of Bias Assessment
Overall risk of bias for change in HbA1c was deemed at some concerns in three studies24,28,36 and low
in four studies.22,33–35 A detailed risk of bias assessment for HbA1c, body weight and systolic blood
pressure is available in Supplementary table S2. We did not explore the presence of publication bias
due to the small number of included studies (less than 10 studies).38
Change in HbA1c
All studies assessed the change in HbA1c. Combination therapy with GLP-1 RA/SGLT2i led to a
greater reduction in HbA1c compared to GLP-1 RA alone (WMD -0.61%, 95% CI -1.09 to -0.14%,
I2=66%, four studies). The combination therapy was also more efficacious in a subgroup analysis of
three studies evaluating simultaneous initiation therapy versus GLP-1 RA (WMD -0.39%, 95% CI -
0.69 to -0.09%, I2=0%) and in one study evaluating the sequential addition of SGLT2i to GLP-1 RA
(WMD -1.03%, 95% CI -1.34 to -0.72%) (Figure 2A). Similarly, combination therapy with GLP-1
RA/SGLT2i reduced HbA1c compared to SGLT2i alone (WMD -0.85%, 95% CI -1.19 to -0.52%,
I2=88%, six studies). Subgroup analyses suggested similar results both for simultaneous initiation
(WMD -0.66%, 95% CI -0.95 to -0.37%, I2=0%, three studies) and for sequential addition of GLP-1
RA to SGLT2i (WMD -0.94%, 95% CI -1.39 to -0.49%, I2=95%, three studies) (Figure 2B). The
GRADE assessment for all estimates was very low, downgraded mainly due to inconsistency,
indirectness, or risk of bias assessment (Supplementary Table S3).
A secondary analysis using short-term (at 28 weeks)23 instead of long-term (at 104 weeks)21 data from
the DURATION-8 trial also demonstrated that combination therapy was more efficacious in reducing
HbA1c compared to either GLP-1 RA (WMD -0.60%, 95% CI -1.09 to -0.12%, I2=70%, four studies)
or SGLT2i (WMD -0.85%, 95% CI -1.18 to -0.51%, I2=88%, six studies). Individual trial data from
studies assessing the simultaneous initiation of GLP-1 RA/SGLT2i 21–23,28,36 suggested that the effect
This article is protected by copyright. All rights reserved.
of the combination therapy on the reduction of HbA1c was likely to be less than additive
(Supplementary Table S4).
Change in Body Weight
Six studies21,24,33–36 assessed the change in body weight. Body weight loss was greater with GLP-1
RA/SGLT2i combination therapy compared to GLP-1 RA (WMD -2.59 kg, 95% CI -3.68 to -1.51 kg,
I2=47%, three studies) (Figure 3A), whereas no difference was found when compared to SGLT2i
(WMD -1.46 kg, 95% CI -2.94 to 0.03 kg, I2=91%, five studies) (Figure 3B). These estimates were
consistent regardless of type of combination therapy (simultaneous initiation or sequential
administration). The GRADE assessment for these findings was very low (Supplementary Table
S3).
In a secondary analysis using short-term data from the DURATION-8 trial23, GLP-1 RA/SGLT2i
therapy decreased body weight not only versus GLP-1 RA (WMD -2.56 kg, 95% CI -3.46 to -1.66 kg,
I2=45%, three studies), but also versus SGLT2i (WMD -1.79 kg, 95% CI -2.99 to -0.59 kg, I2=88%,
five studies). Moreover, based on data from individual reports of the DURATION-8 trial at different
time points, the reduction in body weight after simultaneous initiation of GLP-1 RA/SGLT2i was
likely to be additive at 28 weeks23 but less than additive after 104 weeks21 of treatment
(Supplementary Table S4).
Change in Systolic Blood Pressure
The change in systolic blood pressure was assessed in all studies. GLP-1 RA/SGLT2i combination
therapy reduced systolic blood pressure more than GLP-1 RA (WMD -4.13 mmHg, 95% CI -7.28 to -
0.99 mmHg, I2=30%, four studies) (Figure 4A) or SGLT2i (WMD -2.66%, 95% CI -5.26 to -0.06
mmHg, I2=71%, six studies) (Figure 4B). The GRADE assessment for these estimates was very low
(Supplementary Table S3).
This article is protected by copyright. All rights reserved.
Results were similar both against GLP-1 RA (WMD -4.06 mmHg, 95% CI -7.00 to -1.12 mmHg,
I2=31%, four studies) and against SGLT2i (WMD -2.74 mmHg, 95% CI -5.21 to -0.28 mmHg,
I2=71%, six studies) in a secondary analysis utilizing short-term data from the DURATION-8 trial23.
Individual trial data from studies assessing the simultaneous initiation of GLP-1 RA/SGLT2i21–23,36
suggested that the reduction in systolic blood pressure was likely to be more than additive
(Supplementary Table S4).
Change in Diastolic Blood Pressure
Five studies reported data for the change in diastolic blood pressure.22,24,33–35 Combination therapy
with GLP-1 RA/SGLT2i had no effect on diastolic blood pressure when compared with either GLP-1
RA (WMD -1.09 mmHg, 95% CI -2.70 to 0.53 mmHg, I2=0%, two studies) (Supplementary Figure
S1A) or with SGLT2i (WMD -0.20 mmHg, 95% CI -1.06 to 0.66 mmHg, I2=0%, four studies)
(Supplementary Figure S1B).
Change in eGFR
Only four trials assessed change in eGFR.22,24,34,35 However, these trials did not report results in a way
that they could be used in a meta-analysis.
Severe Hypoglycemia
Six studies were included in the analysis for severe hypoglycemia.21,24,33–36 Five studies 21,24,33–35
defined severe hypoglycemia as an event requiring the assistance of another person. In the remaining
study, albeit a detailed definition for severe hypoglycemia was not reported, based on data from
clinicaltrials.gov (NCT02324842) all hypoglycemic episodes were reported as being non-serious.36
Combination treatment with GLP-1 RA/SGLT2i did not increase the incidence of severe
hypoglycemia compared with GLP-1 RA (OR=1.38, 95% CI 0.14 to 13.14, I2=0%, three studies)
(Supplementary Figure S2A) or SGLT2i (OR=2.39, 95% CI 0.47 to 12.27, I2=0%, five studies)
(Supplementary Figure S2B).
This article is protected by copyright. All rights reserved.
Mortality and Cardiovascular Outcomes
Events of death, myocardial infarction and stroke were very rare across all arms in all studies. Odds
ratios for these outcomes were not significant in the comparisons of GLP-1 RA/SGLT2i combination
with either GLP-1 RA or SGLT2i (Supplementary Table S5). We did not identify any trial reporting
data for hospitalization for heart failure.
Sensitivity Analyses
In a sensitivity analysis excluding one study with considerably higher baseline HbA1c values
(DURATION-8 trial),21 GLP-1 RA/SGLT2i combination reduced HbA1c compared to either GLP-1
RA (WMD -0.76%, 95% CI -1.34 to -0.18%, I2=34%, three studies) or SGLT2i (WMD -0.91%, 95%
CI -1.30 to -0.52%, I2=90%, five studies), consistently with the findings of the main analysis.
We also performed sensitivity analyses excluding one trial24 for which we extracted standard
deviation values for HbA1c, body weight and systolic blood pressure from graphs using relevant
software. Results from these analyses were similar to those of the main analyses comparing
simultaneous combination therapy versus GLP-1 RA for the change in HbA1c, (WMD -0.39%, 95% -
0.69 to -0.09%, I2 0%, three studies), body weight (WMD -2.72 kg, 95% CI -5.04 to -0.39 kg,
I2=71%, two studies) and systolic blood pressure (WMD -3.02 mmHg, 95% CI -5.56 to -0.48 mmHg,
I2=1%, three studies).
Discussion
In this systematic review and meta-analysis we assessed the efficacy and safety of the combination
treatment with a GLP-1 RA and an SGLT2i in patients with type 2 diabetes. Our systematic review
highlights that the current evidence base on the GLP-1 RA/SGLT2i combination therapy is limited
due to the small number of available RCTs and the presence of clinical variability among eligible
trials in terms of type of individual GLP-1 RA or SGLT2i used in each combination. In addition, six
This article is protected by copyright. All rights reserved.
out of the seven RCTs reported only short-term results, whereas only one trial had a long-term
extension period, with outcome data at 104 weeks. Based on our meta-analyses utilizing trials with a
short treatment duration (ranging between 12 and 30 weeks), combination therapy seemed to be
associated with a greater reduction of HbA1c, body weight and systolic blood pressure compared to
either GLP-1 RA or SGLT2i. However, when we incorporated long-term data (at 104 weeks) from the
DURATION-8 trial in our meta-analyses, we did not find any difference in body weight reduction
between GLP-1 RA/SGLT2i combination therapy and SGLT2i alone. Finally, GLP-1 RA/SGLT2i
treatment had no effect on diastolic blood pressure and did not increase odds of severe hypoglycemia
compared to either GLP-1 RA or SGLT2i, while mortality and cardiovascular outcomes were very
rare among all arms, yielding non-significant and imprecise ORs.
The interaction between GLP-1 RAs and SGLT2is when administered as combination therapy has not
been yet clarified, as multiple pathophysiologic mechanisms may be implicated. Speculations
regarding the effect of combination therapy on metabolic parameters is likely to be additive, more
than additive, or less than additive could be made based on individual trial arm data from three-arm
RCTs comparing the simultaneous initiation GLP-1 RA/SGLT2i combination both versus GLP-1 RA
and versus SGLT-2i. We identified three such trials21–23,28,36, for which we compared the effect on
HbA1c, body weight and systolic blood pressure in the arm allocated to the GLP-1 RA/SGLT2i
combination with the sum of the respective effects in the two remaining trial arms allocated to
individual GLP-1 RA and SGLT2i (Supplementary Table S4). Based on these data from three
RCTs21–23,28,36, combination therapy was likely to have a less than additive effect on HbA1c. One
possible explanation of this sub-additive effect could be the mitigation of the actual effect of GLP-1
RAs owing to the indirect increase of glucagon and consequently glucose production caused by
SGLT2is.39 Notably, based on data from the DURATION-8 trial, which reported both short-term and
long-term results, the overall body weight reduction effect in patients randomized to once weekly
This article is protected by copyright. All rights reserved.
exenatide/dapagliflozin combination seemed to decrease over time (being additive after 28 weeks but
less than additive after 104 weeks of treatment). This time-dependent diminishing effect should
probably be attributed to the GLP-1 RA, rather than the SGLT2i component of the combination
therapy. In particular, in patients randomized to the individual GLP-1 RA arm (once weekly
exenatide), body weight reduction at 104 weeks was almost half of that achieved at 28 weeks.
Conversely, patients randomized to the SGLT2i arm (dapagliflozin) achieved greater body weight
loss at 104 weeks compared to 28 weeks. Finally, based on data from the DURATION-8 trial, effect
of combination therapy with GLP-1 RA/SGLT2i on systolic blood pressure seemed to be more than
additive both at 28 weeks and at 104 weeks.
The effect of combination therapy with a GLP-1 RA and an SGLT2i has been addressed in four
previous meta-analyses.7–10 However, these meta-analyses synthesized data only from comparisons
against an SGLT2i. Instead, we also analyzed studies and study arms that included a GLP-1 RA as a
comparator.22,24,28,36 Moreover, we conducted an up-to-date and extensive literature search, thus
retrieving data for three additional studies that had not been included in previous meta-analyses.28,33,36
Of note, we excluded one study40 that assessed the effect of insulin Degludec/Liraglutide (iDegLira)
vs. Insulin Glargine (IGlar U100) as add-on to SGLT2i, because mean liraglutide dose in the iDegLira
arm was lower than the approved therapeutic dose for liraglutide and also because type of basal
insulin was not identical between intervention and comparator arms. Finally, we conducted subgroup
analyses based on type of combination (simultaneous initiation or sequential addition), as well as
secondary analyses based on trial duration.
Our study has certain limitations. The type of GLP-1 RA and SGLT2i in each combination regimen
was not identical between eligible trials, which introduces clinical heterogeneity in our analyses and
attenuates the generalizability of our findings. Due to the limited number of included trials we could
not conduct subgroup analyses based on type of GLP-1 RA or SGLT2i used in each combination. In
This article is protected by copyright. All rights reserved.
fact, each one of the seven eligible RCTs assessed a different GLP-1 RA/SGLT2i combination,
namely liraglutide/canagliflozin36, liraglutide/empagliflozin28, once weekly exenatide/dapagliflozin21–
23, liraglutide/any SGLT2i33, dulaglutide/any SGLT2i34, semaglutide/any SGLT2i35 and exenatide or
liraglutide/canagliflozin24. Therefore, we analyzed GLP-1 RAs and SGLT2is solely as drug classes
and our meta-analyses did not account for any potential differences between individual agents of the
same class. Of note, at least four pertinent studies assessing the effect of combination therapy are
ongoing, thus subgroup analyses focusing on specific GLP-1 RA/SGLT2i combinations may be
feasible upon completion of these trials (Supplementary Table S6). Furthermore, duration of
treatment was short in almost all trials, ranging between 12 and 30 weeks, whereas long-term 104-
week data were available only for the DURATION-8 trial. Therefore, our meta-analysis cannot
provide robust conclusions regarding the long-term effects of GLP-1 RA/SGLT2i combinations, since
it is unknown whether the findings of the DURATION-8 trial can be extrapolated to GLP-1
RA/SGLT2i combinations other than once weekly exenatide/dapagliflozin. Apart from having a short
duration, none of the included studies was designed to assess cardiovascular outcomes, hence our
meta-analyses effect estimates for mortality and vascular outcomes are imprecise and of limited
clinical applicability. Finally, even though we conducted subgroup analyses based on the pattern of
combination therapy (simultaneous initiation or sequential administration), our meta-analysis does not
allow for robust conclusions regarding differences between these two approaches. Ideally, this would
require three-arm trials comparing an arm allocated to simultaneous initiation with GLP-1
RA/SGLT2i versus an arm allocated to initiation with GLP-1 RA and subsequent addition of SGLT2i
versus an arm allocated to initiation with SGLT2i and subsequent addition of GLP-1 RA. However,
we did not identify any such trial; in fact, all included RCTs evaluating sequential administration were
placebo-controlled trials assessing the addition of a GLP-1 RA on preexisting background therapy
with an SGLT2i and vice versa, rather than being active-controlled trials assessing initiation of GLP-1
This article is protected by copyright. All rights reserved.
RA/SGLT2i treatment in a sequential manner versus initiation treatment with a GLP-1 RA alone and
versus initiation treatment with an SGLT2i alone.
In their latest treatment algorithms, both the ADA/EASD and the ESC recommend specific GLP-1
RAs in patients with established or at high risk for atherosclerotic cardiovascular disease, and specific
SGLT2is in patients with heart failure or chronic kidney disease.5,6 Notably, the ADA/EASD in their
consensus report suggest that this decision should be made independently of baseline HbA1c.
Subsequently, if the HbA1c target is not achieved, addition of an SGLT2i is recommended for patients
on a GLP-1 RA, and vice versa for patients on an SGLT2i. They do highlight, however, that current
evidence from available RCTs does not address the impact of the combination of the two drug classes
on long-term cardiorenal outcomes.6 Our systematic review and meta-analysis corroborates that no
robust conclusions can be made whether treatment with specific GLP-1 RA/SGLT2i combinations
can confer complementary cardiovascular benefits compared to treatment with either of their
individual components alone. The GLP-1 RA/SGLT2i combination is also prominent in the
ADA/EASD consensus report when factors other than baseline cardiovascular risk profile are taken
into account, including glycemic efficacy, body weight reduction and hypoglycemic risk.6 Our results
suggest that, in the short term, patients are likely to benefit from GLP-1 RA/SGLT2i combination
therapy in terms of reducing HbA1c, body weight and systolic blood pressure compared to treatment
with either GLP-1 RA or SGLT2i alone. However, due to paucity of relevant RCTs, no implications
can be made regarding the comparative effects between different types of GLP-1 RA/SGLT2i
combinations. Moreover, it is unknown whether long-term data from the DURATION-8 trial
suggesting a possible time-dependent diminishing effect of once weekly exenatide/dapagliflozin on
body weight reduction21, can be extrapolated to other types of GLP-1 RA/SGLT2i combinations.
Finally, given the higher cost of GLP-1 RAs and SGLT2is compared with other drug classes, robust
data from well-conducted country specific cost-effectiveness analyses should also be considered
This article is protected by copyright. All rights reserved.
alongside clinical effectiveness and safety data, when choosing among antidiabetic classes and
individual agents.
Conclusion
Based on data from a limited number of RCTs, combination therapy with a GLP-1 RA and an
SGLT2i seems to reduce HbA1c and systolic blood pressure without increasing the risk for severe
hypoglycemia compared to either GLP-1 RA or SGLT2i alone. Combination therapy with GLP-1
RA/SGLT2i can also reduce body weight compared to either GLP-1 RA or SGLT2i in the short term,
while long-term data from one trial suggest that combination therapy was similar to SGLT2i in terms
of body weight reduction. Currently available research evidence does not allow for valid assessment
of the long-term effectiveness, effect on cardiovascular outcomes or differences between types of
GLP-1 RA/SGLT2i combinations.
This article is protected by copyright. All rights reserved.
Author Contributions
CM, TK, IA, AT and EB conceived and designed the study. CM and TK wrote the study protocol.
CM registered the study protocol. CM did the literature search and CM, PK, TK and KM did the
literature screening, data extraction and quality appraisal of included studies. CM, TK, IA and AL did
the statistical analyses. CM and TK wrote the first draft of the manuscript. All authors interpreted the
data and contributed to the writing of the final version of the manuscript. All authors agreed on the
results and conclusions of this Article.
Conflicts of Interest
EB has received research support from Novo Nordisk. AT has received research support from
Boehringer Ingelheim. All other authors report no conflict of interest.
Funding: None
This article is protected by copyright. All rights reserved.
References
1. Drucker DJ. The Cardiovascular Biology of Glucagon-like Peptide-1. Cell Metab.
2016;24(1):15-30.
2. Heerspink HJL, Perkins BA, Fitchett DH, Husain M, Cherney DZI. Sodium Glucose
Cotransporter 2 Inhibitors in the Treatment of Diabetes Mellitus: Cardiovascular and Kidney
Effects, Potential Mechanisms, and Clinical Applications. Circulation. 2016;134(10):752-772.
3. Zelniker TA, Wiviott SD, Raz I, et al. Comparison of the Effects of Glucagon-Like Peptide
Receptor Agonists and Sodium-Glucose Cotransporter 2 Inhibitors for Prevention of Major
Adverse Cardiovascular and Renal Outcomes in Type 2 Diabetes Mellitus. Circulation.
2019;139(17):2022-2031.
4. Kristensen SL, Rorth R, Jhund PS, et al. Cardiovascular, mortality, and kidney outcomes with
GLP-1 receptor agonists in patients with type 2 diabetes: a systematic review and meta-
analysis of cardiovascular outcome trials. Lancet Diabetes Endocrinol. 2019;7(10):776-785.
5. Cosentino F, Grant PJ, Aboyans V, et al. 2019 ESC Guidelines on diabetes, pre-diabetes, and
cardiovascular diseases developed in collaboration with the EASD. Eur Heart J. 2019:255-
323.
6. Buse JB, Wexler DJ, Tsapas A, et al. 2019 Update to: Management of Hyperglycemia in Type
2 Diabetes, 2018. A Consensus Report by the American Diabetes Association (ADA) and the
European Association for the Study of Diabetes (EASD). Diabetes Care. 2020;43:487-93.
7. Zhou Y, Geng Z, Wang X, Huang Y, Shen L, Wang Y. Meta-analysis on the efficacy and
safety of SGLT2 inhibitors and incretin based agents combination therapy vs. SGLT2i alone or
add-on to metformin in type 2 diabetes. Diabetes Metab Res Rev. 2020;36(2):e3223.
8. Patoulias D, Stavropoulos K, Imprialos K, et al. Glycemic efficacy and safety of GLP-1
receptor agonist on top of SGLT-2 inhibitor treatment compared to sodium-glucose co-
This article is protected by copyright. All rights reserved.
transporter-2 inhibitor alone: a systematic review and meta-analysis of randomized controlled
trials. Diabetes Res Clin Pract. 2019;158:107927.
9. Castellana M, Cignarelli A, Brescia F, et al. Efficacy and safety of GLP-1 receptor agonists as
add-on to SGLT2 inhibitors in type 2 diabetes mellitus: A meta-analysis. Sci Rep. 2019;9(1):1-
8.
10. Guo M, Gu J, Teng F, et al. The efficacy and safety of combinations of SGLT2 inhibitors and
GLP-1 receptor agonists in the treatment of type 2 diabetes or obese adults: a systematic
review and meta-analysis. Endocrine. 2020;67:294-304.
11. Cersosimo E, Johnson EL, Chovanes C, Skolnik N. Initiating therapy in patients newly
diagnosed with type 2 diabetes: Combination therapy vs a stepwise approach. Diabetes, Obes
Metab. 2018;20(3):497-507.
12. Matthews DR, Paldánius PM, Proot P, Chiang YT, Stumvoll M, Del Prato S. Glycaemic
durability of an early combination therapy with vildagliptin and metformin versus sequential
metformin monotherapy in newly diagnosed type 2 diabetes (VERIFY): a 5-year, multicentre,
randomised, double-blind trial. Lancet. 2019;394(10208):1519-1529.
13. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic
reviews and meta-analyses of studies that evaluate health care interventions: explanation and
elaboration. J Clin Epidemiol. 2009;339:b2700.
14. Jelicic Kadic A, Vucic K, Dosenovic S, Sapunar D, Puljak L. Extracting data from figures
with software was faster, with higher interrater reliability than manual extraction. J Clin
Epidemiol. 2016;74:119-123.
15. Rücker G, Cates CJ, Schwarzer G. Methods for including information from multi-arm trials in
pairwise meta-analysis. Res Synth Methods. 2017;8(4):392-403.
16. Sterne JAC, Savović J, Page MJ, et al. RoB 2: A revised tool for assessing risk of bias in
This article is protected by copyright. All rights reserved.
randomised trials. BMJ. 2019;366:1-8.
17. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: An emerging consensus on rating quality of
evidence and strength of recommendations. BMJ. 2008;336(7650):924-926.
18. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple,
graphical test. BMJ. 1997;315:629-634.
19. Diamond GA, Bax L, Kaul S. Uncertain effects of rosiglitazone on the risk for myocardial
infarction and cardiovascular death. Ann Intern Med. 2007;147(8):578-581.
20. Deeks J, Higgins J, Altman D, eds. Chapter 9: Analysing data and undertaking meta-analyses.
In: Higgins JPT, Green S (editors). Cochrane Handbook of Systematic Reviews of
Interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, 2011.
21. Jabbour S, Guja C, Hardy E, Bhattacharya S, Ohman PK, Frias JP. DURATION-8
Randomized Controlled Trial 104-Week Results—Once-Weekly Exenatide (ExQW) plus
Once-Daily Dapagliflozin (DAPA) vs. ExQW or DAPA Alone. Diabetes. 2018;67(Suppl
1):104-LB.
22. Jabbour SA, Frías JP, Hardy E, et al. Safety and efficacy of exenatide once weekly plus
dapagliflozin once daily versus exenatide or dapagliflozin alone in patients
withtype2diabetesinadequately controlled with metformin monotherapy: 52-week results of
the DURATION-8 randomized controlled tri. Diabetes Care. 2018;41(10):2136-2146.
23. Frías JP, Guja C, Hardy E, et al. Exenatide once weekly plus dapagliflozin once daily versus
exenatide or dapagliflozin alone in patients with type 2 diabetes inadequately controlled with
metformin monotherapy (DURATION-8): a 28 week, multicentre, double-blind, phase 3,
randomised controlled trial. Lancet Diabetes Endocrinol. 2016;4(12):1004-1016.
24. Fulcher G, Matthews DR, Perkovic V, et al. Efficacy and safety of canagliflozin when used in
conjunction with incretin-mimetic therapy in patients with type 2 diabetes. Diabetes, Obes
This article is protected by copyright. All rights reserved.
Metab. 2016;18(1):82-91.
25. Abdul-Ghani M, Ali A, Martinez R, et al. Effect of combination therapy with liraglutide plus
canagliflozin on HGP, plasma hormones and HbA1c versus each therapy alone in type 2
diabetes [Abstract 641]. Diabetologia. 2018;61(Suppl 1):S312.
26. Abdul-Ghani M, Ali AM, Martinez RA, et al. Effect of Combination Therapy with Liraglutide
plus Canagliflozin on HGP and HbA1c vs. Each Therapy Alone in T2DM. Diabetes.
2018;67(Suppl 1):109-OR.
27. Martinez R, Al-Jobory H, Adams J, et al. Glycaemic control and beta cell function in type 2
diabetes on SGLT2i and GLP-1RA combination therapy [Abstract 882]. Diabetologia.
2017;60(Suppl 1):S408.
28. Ikonomidis I, Kalogeris A, Kostelli G, et al. The combined treatment with glucagon like
peptide-1 analogues and sodium-glucose co-transporter 2 causes a greater improvement of
arterial stiffness than each treatment alone in type 2 diabetes. Eur Heart J. 2018;39(Suppl 1).
29. Frías JP, Hardy E, Ahmed A, et al. Effects of exenatide once weekly plus dapagliflozin,
exenatide once weekly alone, or dapagliflozin alone added to metformin monotherapy in
subgroups of patients with type 2 diabetes in the DURATION-8 randomized controlled trial.
Diabetes, Obes Metab. 2018;20(6):1520-1525.
30. Jabbour SA, Frías JP, Guja C, Hardy E, Ahmed A, Öhman P. Effects of exenatide once weekly
plus dapagliflozin, exenatide once weekly, or dapagliflozin, added to metformin monotherapy,
on body weight, systolic blood pressure, and triglycerides in patients with type 2 diabetes in
the DURATION-8 study. Diabetes, Obes Metab. 2018;20(6):1515-1519.
31. Woo V, Wysham C, Mathieu C, Vercruysse F, Capuano G, Fulcher G. Canagliflozin reduces
both HbA1c and body weight in patients with type 2 diabetes on background dipeptidyl
peptidase-4 inhibitors or glucagon-like peptide-1 agonists [Abstract 186]. Diabetologia.
This article is protected by copyright. All rights reserved.
2015;58(Suppl 1):S93
32. Woo V, Wysham C, Mathieu C, et al. Canagliflozin (CANA) Added on to Dipeptidyl
Peptidase-4 Inhibitors (DPP-4i) or Glucagon-like Peptide-1 (GLP-1) Agonists with or Without
Other Antihyperglycemic Agents (AHAs) in Type 2 Diabetes Mellitus (T2DM). Can J
Diabetes. 2013;37:S38.
33. Blonde L, Belousova L, Fainberg U, et al. Liraglutide as add‐on to SGLT2 inhibitors in
patients with inadequately controlled type 2 diabetes: LIRA‐ADD2SGLT2i, a 26‐ week,
randomized, double‐blind, placebo‐ controlled trial. Diabetes, Obes Metab. 2020;22(6):929-
937.
34. Ludvik B, Frías JP, Tinahones FJ, et al. Dulaglutide as add-on therapy to SGLT2 inhibitors in
patients with inadequately controlled type 2 diabetes (AWARD-10): a 24-week, randomised,
double-blind, placebo-controlled trial. Lancet Diabetes Endocrinol. 2018;6(5):370-381.
35. Zinman B, Bhosekar V, Busch R, et al. Semaglutide once weekly as add-on to SGLT-2
inhibitor therapy in type 2 diabetes (SUSTAIN 9): a randomised, placebo-controlled trial.
Lancet Diabetes Endocrinol. 2019;7(5):356-367.
36. Ali AM, Martinez R, Al-jobori H, et al. Combination Therapy With Canagliflozin Plus
Liraglutide Exerts Additive Effect on Weight Loss, but Not on HbA1c, in Patients With Type 2
Diabetes. [published online ahead of print March 27, 2020]. Diabetes Care. doi:10.2337/dc18-
2460.
37. Neal B, Perkovic V, Mahaffey KW, et al. Canagliflozin and Cardiovascular and Renal Events
in Type 2 Diabetes. N Engl J Med. 2017;377(7):644-657.
38. Higgins J, Green S. Cochrane Handbook for systematic reviews of interventions: Addressing
reporting biases. Recommendations on testing for funnel plot asymmetry. https://handbook-5-
1.cochrane.org/chapter_10/10_4_3_1_recommendations_on_testing_for_funnel_plot_asymme
This article is protected by copyright. All rights reserved.
try.htm. Accessed: 24 January, 2020
39. De Block C. SGLT2 inhibitors and GLP-1 receptor agonists: a sound combination. Lancet
Diabetes Endocrinol. 2018;6(5):349-352.
40. Philis-Tsimikas A, Billings LK, Busch R, et al. Superior efficacy of insulin
degludec/liraglutide versus insulin glargine U100 as add-on to sodium-glucose co-transporter-
2 inhibitor therapy: A randomized clinical trial in people with uncontrolled type 2 diabetes.
Diabetes, Obes Metab. 2019;21(6):1399-1408.
This article is protected by copyright. All rights reserved.
Figure titles
Figure 1. Flow diagram of the study selection process.
Figure 2. Effect of GLP-1 RA plus SGLT2i on change from baseline in HbA1c (%). A. Compared
with GLP-1 RA. B. Compared with SGLT2i.
Abbreviations: CI, confidence interval; GLP-1 RA, glucagon-like peptide 1 receptor agonist; IV,
inverse variance; SGLT2i, sodium glucose co-transporter 2 inhibitor; SD, standard deviation.
Figure 3. Effect of GLP-1 RA plus SGLT2i on change from baseline in body weight (kg) A.
Compared with GLP-1 RA. B. Compared with SGLT2i.
Abbreviations: CI, confidence interval; GLP-1 RA, glucagon-like peptide 1 receptor agonist; IV,
inverse variance; SGLT2i, sodium glucose co-transporter 2 inhibitor; SD, standard deviation.
Figure 4. Effect of GLP-1 RA plus SGLT2i on change from baseline in systolic blood pressure
(mmHg). A. Compared with GLP-1 RA. B. Compared with SGLT2i.
Abbreviations: CI, confidence interval; GLP-1 RA, glucagon-like peptide 1 receptor agonist; IV,
inverse variance; SGLT2i, sodium glucose co-transporter 2 inhibitor; SD, standard deviation.
This article is protected by copyright. All rights reserved.
Table 1. Baseline characteristics of trials included in the systematic review and meta-analysis.
Study; NCT Blinding
Study duration
(extension period), weeks
Background therapy Study arms
Number of participants randomized,
n
HbA1c, %
BMI, kg/m2
Body weight,
kg
Systolic blood
pressure, mmHg
Diabetes duration,
years
Age, years
Simultaneous initiation of GLP-1 RA plus SGLT2i
Ali 202025–27,36; NCT02324842 Open-label 16 metformin ±
sulphonylurea
liraglutide 1.8mg QD/canagliflozin 300mg
QD 15 8.10 34.80 NA 136.00 8.30 51.00
liraglutide 1.8mg QD 15 8.40 35.10 NA 131.00 7.10 53.00 canagliflozin 300mg QD 15 8.20 34.80 NA 134.00 7.30 53.00
Ikonomidis 201828 NA 12 NA
liraglutide/empagliflozin QD 20 8.20 NA NA 139.00 NA NA
liraglutide QD 20 8.00 NA NA 137.00 NA NA empagliflozin QD 20 7.80 NA NA 134.00 NA NA
Jabbour 2018 (DURATION
8)21–23,29,30; NCT02229396
Double-blind
28 (+24+52)
metformin ≥1500mg/day
exenatide 2mg QW/dapagliflozin 10mg
QD 228 9.34 33.20 91.79 130.10 7.60 53.80
exenatide 2mg QW 227 9.30 32.00 89.77 129.10 7.40 54.20 dapagliflozin 10mg QD 230 9.30 33.00 91.06 130.00 7.10 54.50
Sequential addition of GLP-1 RA to SGLT2i
Blonde 2020 (LIRA-
ADD2SGLT2i)33; NCT02964247
Double-blind 26
SGLT2i (dapagliflozin, canagliflozin, empagliflozin) ± metformin
liraglutide 1.8mg QD 203 8.00 32.00 91.00 127.50 10.10 54.70
placebo 100 8.00 32.60 91.40 128.50 9.60 56.00
Ludvik 2018 Double- 24 metformin dulaglutide 1.5mg QW 142 8.04 32.87 92.87 129.70 9.21 56.17
This article is protected by copyright. All rights reserved.
(AWARD 10)34; NCT02597049
blind n=136, canagliflozin
n=20, dapagliflozin
n=66, empagliflozin
n=56 metformin
n=135, canagliflozin
n=19, dapagliflozin
n=61, empagliflozin
n=61
dulaglutide 0.75mg QW 141 8.04 32.77 91.07 130.35 10.05 58.55
metformin n=135,
canagliflozin n=19,
dapagliflozin n=72,
empagliflozin n=49
placebo 140 8.05 32.39 90.50 130.57 8.87 57.10
Zinman 2019 (SUSTAIN 9)35; NCT03086330
Double-blind 30
SGLT2i n=150,
metformin n=106,
sulphonylurea
semaglutide 1.0mg QW 151 8.00 31.10 89.60 127.20 9.80 57.50
This article is protected by copyright. All rights reserved.
n=19 SGLT2i n=151,
metformin n=110,
sulphonylurea n=20
placebo 151 8.10 32.70 93.80 128.60 9.60 56.60
Sequential addition of SGLT2i to GLP-1 RA
Fulcher 201624,31,32;
NCT01032629
Double-blind 18
exenatide n=70,
liraglutide n=25†
canagliflozin 300mg QD 30 8.30 37.70 NA 136.40 15.30 61.50 canagliflozin 100mg QD 35 8.20 37.20 NA 136.40 14.60 60.70
placebo 30 8.00 37.30 NA 136.90 14.70 60.90
Data are mean values. Abbreviations: GLP-1 RA, glucagon-like peptide 1 receptor agonist; NA, not available; NCT number, clinicaltrilas.gov identifier; QW, once weekly; QD, once daily; SGLT2i, sodium glucose co-transporter 2 inhibitor. †Patients receiving GLP-1 RA as background therapy were a subpopulation of the CANVAS (CANagliflozin cardioVascular Assessment Study) trial programme. Patients were receiving insulin, sulphonylurea or metformin also as background therapy but the exact percentages are not available.
This article is protected by copyright. All rights reserved.
Figure 1
12436 Records identified through database searching:
2035 MEDLINE 6097 EMBASE 4304 Cochrane Central Register of Randomized Trials
9159 Records screened by title and abstract
9111 Records excluded
48 Full-text records assessed for eligibility
38 Excluded 5 ineligible study design 3 ongoing trials 2 ineligible outcomes 28 ineligible intervention
16 Records of 7 studies included in the systematic review and meta-analysis
3277 Duplicate records removed
6 Records of 3 studies identified through manual search of conference meetings and clinicaltrials.gov. Two of these studies were subsequently published as journal articles.
This article is protected by copyright. All rights reserved.
Figure 3
A
B
This article is protected by copyright. All rights reserved.
Figure 4
A
B
This article is protected by copyright. All rights reserved.