this article is protected by copyright. all rights reserved. · we searched medline, embase, the...

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.

http://dx.doi.org/10.1111/dom.14108




http://crossmark.crossref.org/dialog/?doi=10.1111%2Fdom.14108&domain=pdf&date_stamp=2020-06-01

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


RA or SGLT2i. No robust conclusions can be made regarding long-term effectiveness or effect on

cardiovascular outcomes.


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


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


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


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).


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).


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


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


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



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


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).


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


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


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


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


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


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.


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


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-


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


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


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.


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


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.


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.



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.




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


(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


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.


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.


Figure 3

A

B


Figure 4

A

B


this article is protected by copyright. all rights reserved. · we searched medline, embase, the...

Documents