catherine only

hsj

16.pdf

Home >Chemistry homework help >Pharmacology homework help >catherine only

A cc

ep te

d A

rt ic

le Efficacy and Safety of Glucagon-like peptide-1 receptor agonists

in type 2 diabetes

Systematic review and mixed-treatment comparison analysis

GLP-1RAs treatments in diabetes

Zin Z Htike , MBBS

Francesco Zaccardi , MD

Dimitris Papamargaritis, PhD

David R Webb, PhD

Kamlesh Khunti, PhD

Melanie J Davies, MD

Diabetes Research Centre, University of Leicester, Leicester, UK Leicester Diabetes Centre, Leicester General Hospital, University Hospitals of Leicester NHS Trust

Corresponding Author and reprint requests Dr Zin Zin Htike Leicester Diabetes Centre, Leicester General Hospital, Leicester, UK Phone: +44 0116 258 8602 – Fax: +44 0116 258 4499 Email: [email protected] ABSTRACT

Aims

To compare efficacy and safety of glucagon-like peptide-1 receptor agonists (GLP-1RAs) in

subjects with type 2 diabetes

Materials and Methods

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.12849

A cc

ep te

d A

rt ic

le We electronically searched, up to June 3rd, 2016, published randomised clinical trials lasting

between 24 and 32 weeks and comparing a GLP-1RA (albiglutide, dulaglutide, twice-daily

(EBID) and once-weekly exenatide, liraglutide, lixisenatide, semaglutide, and taspoglutide)

with placebo or another GLP-1RA. Data on cardiometabolic and safety outcomes were

analysed using a mixed-treatment comparison meta-analysis

Results

34 trials (14464 participants) met the inclusion criteria; no published data for semaglutide

were available. Compared to placebo, all GLP-1RAs reduced HbA1c and fasting plasma

glucose (FPG) (from -0.55% and -0.73mmol/L for lixisenatide to -1.21% and -1.97mmol/L

for dulaglutide). There were no differences within short-acting (EBID and lixisenatide) or

long-acting (albiglutide, dulaglutide, once-weekly exenatide, liraglutide, and taspoglutide)

groups. Compared to EBID, dulaglutide treatment was associated with the greatest HbA1c

and FPG reduction (0.51% and 1.04mmol/L), followed by liraglutide (0.45%, 0.93mmol/L)

and once-weekly exenatide (0.38% and 0.85mmol/L); similar reductions were found when

these three agents were compared to lixisenatide. Compared to placebo, all GLP-1RAs except

albiglutide reduced weight and increased the risk of hypoglycaemia and gastrointestinal (GI)

side effects and all agents except dulaglutide and taspoglutide reduced systolic blood pressure.

When all GLP-1RAs were compared against each other, no clinically meaningful differences

were observed in weight loss, blood pressure reduction or hypoglycaemia risk. Albiglutide

had the lowest risk of nausea and diarrhoea and once-weekly exenatide the lowest risk of

vomiting.

Conclusions

A cc

ep te

d A

rt ic

le RCTs demonstrate that all GLP-1RAs improve glycaemic control, reduce body weight, and

increase the risk of adverse gastrointestinal symptoms vs placebo. Although there are no

differences when short-acting agents are compared against each-other or when long-acting

agents are compared against each-other, dulaglutide, liraglutide, and once-weekly exenatide

are superior to EBID and lixisenatide at lowering HbA1c and FPG. There are no differences

in hypoglycaemia between these three agents whilst once-weekly exenatide has a lowest risk

of vomiting. These results, along with patient’s preferences and individualised targets, should

be considered in selecting a GLP-1RA.

A cc

ep te

d A

rt ic

le Introduction

Type 2 diabetes (T2DM) is a chronic metabolic disorder characterised by complex

pathophysiology of progressive beta cell dysfunction and a varying degree of insulin

resistance. As the condition progresses, achieving and maintaining glycaemic control

becomes a challenge despite the availability and use of a number of classes of glucose

lowering therapies[1]. Current guidelines recommend a patient-centred approach when

choosing appropriate glucose lowering treatments, with a primary goal of achieving

individualised glycaemic target whilst minimising adverse effects, particularly weight gain

and hypoglycaemia [2].

Glucagon like peptide-1 receptor agonists (GLP-1RAs) are a class of therapeutic agent,

which provide significant improvement in HbA1c with an added benefit of promoting weight

loss and low risk of hypoglycaemia [3-6]. GLP-1 is a gut hormone produced by the small

intestine in response to oral ingestion of glucose, which promotes a glucoregulatory effect by

increasing insulin and suppressing glucagon secretion [7]. It also facilitates weight loss by

delaying gastric emptying and acting centrally on the satiety centre to reduce the food intake

[7]. Manipulating the molecular structure of GLP-1 alters its pharmacological properties and

produces biological effects that can be exploited clinically. GLP-1RAs are increasingly

classified by duration of action into long-acting (albiglutide, dulaglutide, once-weekly

exenatide, liraglutide, semaglutide, and taspoglutide) and short-acting (twice-daily exenatide

(EBID) and lixisenatide) agents [8].

To date, EBID, lixisenatide, liraglutide, albiglutide, dulaglutide, and once-weekly exenatide

are licensed by the US Food and Drug Administration (FDA) to be used in the management

of T2DM. The clinical trials and development of taspoglutide were discontinued in 2010 and

therefore it is not available in clinical practice while semaglutide in a once-weekly

subcutaneous formulation is in phase III clinical trials.

A cc

ep te

d A

rt ic

le The American Diabetes Association (ADA)/European Association for the study of Diabetes

(EASD) guidelines recommend the use of GLP-1RAs as an adjunctive therapy to lifestyle

modification and metformin [9]. However, there are no specific recommendations on which

GLP-1RA to choose in clinical practice possibly due to limited availability of head to head

studies comparing the efficacy and safety of GLP-1RAs [10]. When direct comparisons are

limited, mixed-treatment comparison analysis (also known as network meta-analysis) is

regarded as the methodology of choice to compare several treatments [11, 12]. Using this

approach and available data, we therefore aimed to compare the clinical profiles of GLP-

1RAs.

A cc

ep te

d A

rt ic

le Methods

Data Sources and Searches

This study was performed following a pre-specified protocol and PRISMA guidelines for

conducting and reporting of systematic reviews and meta-analysis [11, 13, 14]

(Supplementary Material). We searched PubMed, ISI Web of Science, and the Cochrane

Library for randomised clinical trials (RCTs) published from inception to June 3rd, 2016 and

comparing a GLP-1RA with placebo or another GLP-1RA (EBID, lixisenatide, liraglutide,

albiglutide, dulaglutide, once-weekly exenatide, semaglutide, and taspoglutide) in adults with

type 2 diabetes [15]. Only fully published RCTs were included (abstracts were excluded). We

sought additional studies by manually scanning reference lists of eligible studies and previous

systematic reviews and meta-analysis. Although the clinical trials and the development of

taspoglutide were discontinued in 2010, we included taspoglutide RCTs data as they

contribute to indirect estimations. No language restrictions were applied. Details on the

search strategy are provided in supplementary Figure S1.

Study Selection

Fully published RCTs in subjects with T2DM lasting between 24 and 32 weeks with data on

EBID 10µg, lixisenatide 20µg, liraglutide 1.8mg, albiglutide 50mg, dulaglutide 1.5mg, once-

weekly exenatide 2mg, semaglutide 1mg, and taspoglutide 20mg were included if they

reported data on HbA1c. RCTs including patients with chronic kidney disease were excluded.

Data extraction and Quality Assessment

Three authors (ZZH, DP, FZ) independently performed the literature search and extracted

study data using pre-defined forms. Extracted information included: first author name, trial

name/acronym, year of journal publication, background glucose-lowering therapies, GLP-

A cc

ep te

d A

rt ic

le 1RA treatments, study duration, sample size, gender, age, diabetes duration, baseline HbA1c.

We collected data on arm-specific number of participants, mean difference and standard error

(or standard deviation) for continuous data (cardiometabolic outcomes: HbA1c, fasting

plasma glucose (FPG), body weight, total, low- and high-density lipoprotein cholesterol,

triglycerides, systolic and diastolic blood pressure, and heart rate) and on total number of

participants and participants with events for dichotomous data (safety outcomes: nausea,

vomiting, diarrhoea, all (total) hypoglycaemic events, and injection site reactions). Data were

extracted according to the intention to treat principle. When it was not possible to obtain

efficacy and safety information from the published report, we retrieved data from

ClinicalTrials.gov. In case of disagreement between the three reviewers on the eligibility of

an article or on extracted data, consensus was reached by re-evaluation of the article and

consultation with an independent reviewer. Study quality was assessed using the Cochrane

risk of bias tool [16].

Data Synthesis and Analysis

We performed a mixed-treatment comparison within a frequentist framework based on the

method of multivariate meta-analysis [17-19]. We used treatment-specific mean difference

from baseline and odds ratio (OR) as effect measure for continuous and dichotomous data,

respectively; we added 0.5 when studies reported zero events for safety outcomes. For each

outcome, we graphically summarised the evidence by using a network diagram [20] and

performed random-effects network meta-analyses assuming that all treatment contrast have

the same heterogeneity variance. We presented results against placebo and comparisons

across all GLP-1RAs in graphs and tables. For each outcome, we assessed consistency

between direct and indirect evidence by using the ‘design by treatment’ interaction model

[21]. Stata 14.1 (Stata Corp, College Station, TX, USA) was used for all analyses and results

A cc

ep te

d A

rt ic

le are reported with 95% confidence intervals (CIs); a p-value <0.05 was deemed statistically

significant.

Results

Study characteristics

We identified 7697 records, of which thirty-four fulfilled the inclusion criteria

(Supplementary Figure S1). Included RCTs were published between 2004 and 2016; no

published data for semaglutide fulfilling inclusion criteria were available (Table 1) [22-55];

ten were ‘head-to-head’ comparisons. A total of 14464 participants (47.9% female; range

35.5-63.5) with T2DM were included in our analysis with a mean baseline age of 56.2 (range,

53.0-63.6) years, HbA1c of 8.2% (range, 7.6-9.0), and duration of diabetes of 8.1 (range, 1.5-

17.2) years. Background therapies were metformin, sulphonylurea, thiazolidinedione, and

insulin alone or any combination of these medications. Overall risk of bias for the individual

studies was variable, either low (69.6% for the six domains), high (9.8%) or unclear (20.6%)

(Table S1).

Mixed-treatment comparison

Estimates for cardiometabolic outcomes are shown in Figure 1, Table 2 and Table S2, and

results for safety outcomes in Figure 2, Table 3 and Table S3. Networks of comparisons are

depicted in Figure S2. For both cardiometabolic and safety outcomes, results are first

presented for against placebo. Then, when comparing GLP-1RAs with each other, results are

first reported comparing short-acting (EBID and lixisenatide) vs long-acting; then comparing

short-acting against each other; and lastly long-acting GLP-1RAs agents against each other.

Comparison vs Placebo for cardiometabolic outcomes

Glycaemic control

A cc

ep te

d A

rt ic

le Compared to placebo, all GLP-1RA treatments improved both HbA1c and FPG. HbA1c

reduction ranged from 0.55% to 1.21%, the greatest being with dulaglutide (1.21%; 1.05,

1.36), followed by liraglutide (1.15%; 1.03, 1.27), once-weekly exenatide (1.08%; 0.89, 1.27),

taspoglutide (0.99%; 0.85, 1.14), albiglutide (0.94%; 0.64, 1.24), EBID (0.70%; 0.59, 0.81)

and lixisenatide (0.55%; 0.42, 0.68) (Figure 1, Table S2). Improvement in FPG followed the

same pattern as that of HbA1c, ranging from a maximum of 1.97mmol/L (1.49, 2.44) with

dulaglutide to a minimum of 0.73mmol/L (0.38, 1.08) with lixisenatide (Figure 1, Table S2).

Body weight

Data on body weight were available for 14054 participants. Compared to placebo, treatment

with GLP-1RAs was associated with significant weight loss, the greatest being with

liraglutide (1.96kg; 1.25, 2.67) followed by EBID (1.67kg; 1.05, 2.29), dulaglutide (1.57;

0.66, 2.48), taspoglutide (1.54kg; 0.66, 2.41), once-weekly exenatide (1.49kg; 0.40, 2.58),

and lixisenatide (0.78kg; 0.09, 1.48) (Figure 1, Table S2).

Blood pressure and lipid profile

Data for other cardiometabolic outcomes ranged from 4955 subjects for triglycerides to 7375

subjects for systolic blood pressure (SBP). Compared to placebo, treatment with liraglutide,

once-weekly exenatide, EBID and dulaglutide resulted in significant reduction in SBP of

4.04mmHg (2.90, 5.19), 3.64mmHg (2.13, 5.15), 3.43mmHg (2.17, 4.69) and 3.35mmHg

(2.10, 4.61), respectively; and in diastolic blood pressure (DBP) of 1.80 mmHg (1.00, 2.60)

for EBID, 1.16mmHg (0.17, 2.15) for once-weekly exenatide, 0.92mmHg (0.13, 1.71) for

liraglutide (Figure 1, Table S2).

A cc

ep te

d A

rt ic

le With the exception of lixisenatide, treatment with all GLP-1RAs significantly raised heart

rate (HR) compared to placebo, ranging from a minimum of 1.07bpm (0.00, 2.15) with EBID

to a maximum of 3.28bpm (2.45, 4.11) with liraglutide (Figure 1, Table S2).

Compared to placebo, dulaglutide and liraglutide lowered triglycerides while albiglutide,

dulaglutide, once-weekly exenatide, and liraglutide lowered both total cholesterol and LDL-C

(Figure 1, Table S2). No clinically meaningful changes in HDL-C level were noted.

Comparison among GLP-1RAs for cardiometabolic outcomes

Glycaemic control

Compared to EBID, treatment with dulaglutide, liraglutide, once-weekly exenatide and

taspoglutide showed greater HbA1c reduction, with differences of 0.51% (95% CI: 0.34,

0.68), 0.45% (0.31, 0.59), 0.38% (0.21, 0.55), and 0.30% (0.13, 0.46) respectively while

albiglutide showed no difference. Corresponding FPG reductions were 1.04mmol/L (0.54,

1.53) for dulaglutide, 0.93mmol/L (0.56, 1.31) for liraglutide, 0.85mmol/L (0.41,1.28) for

once-weekly exenatide, and 0.75mmol/L (0.32,1.18) for taspoglutide; no significant

difference was found between albiglutide and EBID (Table 2).

When compared to lixisenatide, HbA1c differences were 0.66% (0.46, 0.86) with dulaglutide,

0.60% (0.43, 0.77) with liraglutide, 0.53% (0.31, 0.75) with once-weekly exenatide, 0.45%

(0.26, 0.64) with taspoglutide, and 0.39% (0.07, 0.72) with albiglutide. Corresponding FPG

reductions were 1.23mmol/L (0.65, 1.81) for dulaglutide, 1.13mmol/L (0.66, 1.60) for

liraglutide, 1.05mmol/L (0.47, 1.63) for once-weekly exenatide, and 0.95mmol/L (0.45, 1.45)

for taspoglutide; there was no significant difference between albiglutide and lixisenatide

(Table 2).

Amongst all GLP-1RAs comparisons, the largest reductions in HbA1c (0.66%; 0.46, 0.86)

and FPG (1.23mmol/L; 0.31, 1.87) were observed with dulaglutide vs lixisenatide. Within

A cc

ep te

d A

rt ic

le short-acting agents, no significant differences were noted for both HbA1c and FPG; similarly,

no differences were observed for the same outcomes when long-acting agents were compared

against each other.

Body weight

Among all possible comparisons of GLP-1RAs, the only differences noted were reduction of

0.89kg (0.01, 1.76) with EBID vs lixisenatide and 1.17kg (0.19, 2.15) with liraglutide vs

lixisenatide (Table 2).

Blood pressure and lipid profile

For SBP, among all possible GLP-1RAs comparisons the only difference was between

liraglutide and taspoglutide (2.40mmHg; 0.21, 4.59 reduction in favour of liraglutide).

Regarding DBP, EBID lowered it to a significantly greater extent than taspoglutide

(1.41mmHg; 0.03, 2.78), dulaglutide (1.07mmHg; 0.15, 1.99) and liraglutide (0.88mmHg;

0.94, 1.72) while no differences were noted for DBP among long-acting GLP-1RAs (Table 2).

Compared to EBID, HR was 2.21bmp (1.05, 3.37), 2.18 bpm (0.92, 3.44), and 1.51bpm (0.41,

2.61) higher with liraglutide, once-weekly exenatide, and dulaglutide, respectively (Table 2).

Corresponding values vs lixisenatide were 3.48 (1.96, 5.01) increase for liraglutide, 3.45

(1.36, 5.55) for once-weekly exenatide, and 2.79 (1.25, 4.32) for dulaglutide. Among all

GLP-1RAs, the treatment with liraglutide resulted in the greatest increase in heart rate of

3.48bpm (1.96, 5.01) compared to lixisenatide. Within short-acting agents, no significant

differences were noted for HR; similarly, no differences were observed when long-acting

agents were compared against each other.

Only marginal differences were found for the lipid profile among all GLP-1RAs (Table 2).

A cc

ep te

d A

rt ic

le The network meta-analytical comparisons between lixisenatide and other GLP-1RAs for SBP,

DBP, and lipid parameters was not possible from available published RCTs. Statistical

inconsistencies of the networks were not significant for all cardiometabolic outcomes.

Comparison vs Placebo for Safety outcomes

Hypoglycaemia

Total hypoglycaemic events were reported in 12761 subjects; all GLP-1RAs except

albiglutide significantly increased the risk of hypoglycaemia compared to placebo from a

minimum OR of 1.59 (1.10, 2.31) for lixisenatide to a maximum OR of 3.74 (1.51, 2.24) for

taspoglutide (Figure 2, Table S3); results were consistent after excluding studies reporting

background sulphonylurea and insulin therapy (Table S3).

Gastrointestinal side-effects

Data on treatment-associated nausea, vomiting and diarrhoea were available in 12258, 11835,

and 12064 subjects, respectively. Compared to placebo, all GLP-1RAs had significantly

higher risk of nausea (except albiglutide), vomiting, and diarrhoea (except taspoglutide and

lixisenatide). The highest risks were with taspoglutide for nausea (OR 8.28; 4.16, 16.13) and

vomiting (10.33; 4.42, 24.15) and with albiglutide for diarrhoea (3.25; 1.59, 6.64) (Table S3).

Other side effects

Imprecise estimates were found in the injection site reactions analyses, possibly due to

heterogeneous outcomes definitions across studies. Compared to placebo, available data

suggested an increased risk with lixisenatide and, particularly, taspoglutide (Table S3).

Comparison among GLP-1RAs for safety outcomes

A cc

ep te

d A

rt ic

le Hypoglycaemia

Among all GLP-1RAs comparisons, the risk of hypo was not significantly different (Table 3).

Statistical inconsistency was found for hypoglycaemia only in the main analysis (p=0.044;

analysis excluding background sulphonylurea and insulin, p=0.118).

Gastrointestinal side-effects

Both EBID and lixisenatide increased the risk of nausea compared to albiglutide (OR 3.35

and 3.07, respectively) and once-weekly exenatide (OR 2.38 and 2.17, respectively) (Table 3).

No difference was found between EBID and lixisenatide. Among long-acting agents,

taspoglutide (OR 3.67), dulaglutide (OR 2.94), and liraglutide (OR 2.64) increased the risk of

nausea compared to once-weekly exenatide; a similarly pattern was found when these three

drugs were compared to albiglutide. The largest differences amongst all comparisons was

between taspoglutide vs albiglutide (OR 5.20; 1.71, 15.80).

The risk of vomiting was lower comparing once-weekly exenatide vs EBID (OR 0.52) and

higher comparing taspoglutide vs lixisenatide (OR 2.66). No difference was found between

EBID and lixisenatide. Among long-acting agents, taspoglutide (OR 4.64), dulaglutide (2.70),

and liraglutide (OR 2.39) had an increased risk of vomiting compared to once-weekly

exenatide and taspoglutide vs albiglutide (OR 3.83). The largest differences amongst all

comparisons was between taspoglutide vs once-weekly exenatide (OR 4.64; 1.68. 12.77).

The risk of diarrhoea was lower with EBID vs liraglutide (OR 0.67) and with lixisenatide vs

liraglutide (OR 0.45), dulaglutide (OR 0.42) and albiglutide (OR 0.40; 0.19, 0.87); no

significant differences were within short-acting and within long-acting agents (Table 3). The

largest differences amongst all comparisons was between albiglutide and lixisenatide (OR

2.50; 1.15, 5.26).

A cc

ep te

d A

rt ic

le Discussion

GLP-1RAs are a class of therapeutic agents available for the management of hyperglycaemia

in people with type 2 diabetes. ADA/EASD treatment algorithm recommends the use of

GLP-1RAs as a second line therapy when glucose control on metformin alone is inadequate

[9]. However, no specific recommendations are made as to which GLP-1RAs should be

chosen. In individual RCTs, GLP-1RAs have been shown to improve glycaemic control with

weight loss compared to other conventional glucose lowering agents. However, there is

limited evidence from ‘head-to-head’ studies comparing the efficacy and safety of GLP-

1RAs. As a result, decision-makers have to rely mainly on anecdotal evidence and clinical

judgement in choosing a particular agent among all available GLP-RAs. To help clarify the

evidence, we have conducted a systematic review and mixed-treatment comparison meta-

analysis aiming to compare GLP-1RAs across a wide range of clinically relevant outcomes.

These results could provide unified hierarchies of evidence for GLP-1RAs in the holistic

management of hyperglycaemia in T2DM.

We presented efficacy and treatment-related adverse event results against placebo and

compared GLP-1RAs against each other to aid decision makers in selecting the appropriate

agent. When compared to placebo, all GLP-1RAs significantly improved both HbA1c and

FPG, the greatest reduction being with dulaglutide (1.2% and 1.97 mmol/L, respectively).

Compared to short-acting agents (EBID and lixisenatide), our results showed that the long-

acting agents dulaglutide, liraglutide, and once-weekly exenatide all resulted in significantly

greater HbA1c reduction. Treatment with dulaglutide was associated with the greatest HbA1c

reduction of 0.51% followed by liraglutide (0.45%) and once-weekly exenatide (0.38%) vs

EBID; the pattern of reduction for these three drugs was similar when compared against

lixisenatide. Likewise, there was a significant reduction in FPG with dulaglutide, liraglutide,

and once-weekly exenatide compared to EBID or lixisenatide. However, short-acting agents

A cc

ep te

d A

rt ic

le (EBID and lixisenatide) were similar in terms of HbA1c and FPG lowering efficacy when

compared with each other; similarly, there were no differences for the same outcomes across

long-acting agents.

Previous studies have shown that GLP-1RAs are effective in reducing body weight and blood

pressure compared to placebo [56, 57]. Our results are in line with these observations as they

show a reduction of body weight and particularly of systolic blood pressure when compared

to placebo; however, we did not find clinically meaningful differences across all GLP-1RAs

for these two outcomes. Regarding lipid profiles, differences between each GLP-1RAs vs

placebo and across all GLP-1RAs were clinically marginal.

Notably, available data showed that twice-daily and once-weekly exenatide, liraglutide, and

dulaglutide significantly raised heart rate vs placebo from a minimum of 1.1bpm for EBID to

a maximum of 3.3bpm for liraglutide. The highest difference was found between liraglutide

and lixisenatide (3.5bpm) although no differences were noted within short-acting agents or

within long-acting agents.

Compared to placebo, taspoglutide was associated with the highest (4-times) and lixisenatide

with the lowest (1.6-times) risk of hypoglycaemia. However, no difference were noted among

all GLP-1RAs when compared against each other. Treatment-associated gastrointestinal

symptoms are well-recognised adverse effects of GLP-1RAs and are demonstrated to be

higher when compared to placebo or other common glucose lowering therapies [58]. While

our results have confirmed an increased risk of gastrointestinal side effects compared to

placebo, we also noted important differences among GLP-1RAs for nausea, vomiting and

diarrhoea. Albiglutide had the lowest risk of nausea (5-times lower than taspoglutide) and

diarrhoea (2.5-times lower than lixisenatide) and once-weekly exenatide of vomiting (5-times

lower than taspoglutide).

A cc

ep te

d A

rt ic

le Overall, our findings showed that treatments with the long acting GLP-1RAs dulaglutide,

liraglutide, and once-weekly exenatide resulted in greater glycaemic efficacy but not

clinically meaningful differences in blood pressure, body weight, lipid profile, or risk of

hypoglycaemia when compared to short-acting agents. Across theses three long-acting agents,

however, once-weekly exenatide had the lowest risk of nausea and vomiting.

To the best of our knowledge, this is the first attempt to systematically compare all available

GLP-1RAs combining several clinically-meaningful efficacy and safety outcomes

simultaneously and comprehensively. We have pooled data from published RCTs with a

mixed-treatment comparison meta-analytical approach aiming to fulfil the gap in providing

decision-makers with evidence to choose the appropriate agents considering specific

advantages and disadvantages of each agent. We therefore collated data for multiple

outcomes relevant to clinical practice (i.e., HbA1c, FPG, body weight, blood pressure, lipid

profile, hypoglycaemia, and common gastrointestinal side-effects). However, we should also

acknowledge some shortcomings of this study. First, it should be noted that postprandial

glucose, along with FPG, contributes to the overall glycaemic control [59]. Although several

studies suggest short acting GLP-1RAs, lixisenatide and EBID, exert a greater effect of on

postprandial glucose [60-62], we could not assess the comparative efficacy of GLP-1RAs on

this outcome given its variable definition and very limited availability of these data across all

GLP-1RAs studies. Similarly, injection site reactions are common to these treatments. Direct

comparisons would suggest more injection site reactions with once-weekly exenatide,

albiglutide, and taspoglutide compared to exenatide and liraglutide due to the differences in

formulation and delivery technology [42-45, 54] but data were inconsistently reported in the

included RCTs limiting the possibility to combine and compare them analytically. Second,

some characteristics (population included, duration of follow-up, background therapies and

quality of studies) differ across included RCTs. Differences in background therapies are

A cc

ep te

d A

rt ic

le particularly relevant as cardiometabolic effects of GLP-1RAs may differ when these agents

are added to metformin, insulin, thiazolidinedione, or sulphonylurea. Third, the search was

based only on published papers. Although the risk of publication bias should be lower for

RCTs compared to other study types, yet it is possible that some studies have been conducted

and not registered and/or reported (e.g. fully published SUSTAIN studies for semaglutide

fulfilling inclusion criteria were not reported at the time of literature search). Fourth, we used

“intermediate” surrogate biomarkers of vascular disease; future long-term head-to-head RCTs

would clarify whether these differences translate to different ‘hard’ outcomes. As of

September 30th 2016, three placebo-controlled studies assessing cardiovascular outcomes

have been published. Short acting GLP-1RA lixisenatide showed neutral effect on major

adverse cardiovascular events (cardiovascular death, nonfatal MI or nonfatal stroke) while

the long-acting GLP-1RAs, liraglutide and semaglutide, demonstrated 13% and 26% relative

risk reduction in the LEADER and SUSTAIN-6 study, respectively [63-65]. Other ongoing

studies EXSCEL (once-weekly exenatide) and REWIND (dulaglutide) are awaiting to be

reported. Fifth, although a number of potential sensitivity analyses are possible, we assessed

the risk of hypoglycaemia excluding studies with background SU and insulin as we deemed it

more clinically relevant. Lastly, this meta-analysis shares the same drawbacks common to

other network analyses [66].

In conclusion, although there are no differences when short-acting agents are compared

against each-other and when long-acting agents are compared against each-other, available

data indicate dulaglutide, liraglutide, and once-weekly exenatide are superior to EBID and

lixisenatide at lowering HbA1c and FPG. There are no differences in hypoglycaemia between

these three agents whilst once-weekly exenatide has a lowest risk of nausea and vomiting.

The choice among these agents should be tailored taking into account their differences in

safety and efficacy along with patient’s targets and needs.

A cc

ep te

d A

rt ic

le AUTHOR CONTRIBUTION ZZH, FZ, MJD study idea and design

ZZH, DP, FZ literature search and data extraction

FZ data analysis

ZZH, DP, FZ, DW, KK, MJD study critical revision and manuscript draft

All authors provided final approval of the version to publish. ZZH is the study guarantor.

Statistical codes and datasets are available from the corresponding Author (ZZH).

ACKNOWLEDGMENTS

We acknowledge the support of the following institutes in this work: the National Institute

for Health Research, Collaboration for Leadership in Applied Health Research and Care -

East Midlands (NIHR CLARHC East Midlands); the National Institute for Health Research

(NIHR) Diet, Lifestyle & Physical Activity Biomedical Research Unit based at University

Hospitals of Leicester and Loughborough University. The views expressed are those of the

authors and not necessarily those of the NHS, the NIHR or the Department of Health.

CONFLICT OF INTEREST

KK has acted as a consultant and speaker for Novartis, Novo Nordisk, Sanofi-Aventis, Lilly

and Merck Sharp & Dohme. He has received grants in support of investigator and

investigator initiated trials from Novartis, Novo Nordisk, Sanofi-Aventis, Lilly, Pfizer,

Boehringer Ingelheim and Merck Sharp & Dohme. KK has received funds for research,

honoraria for speaking at meetings and has served on advisory boards for Lilly, Sanofi-

Aventis, Merck Sharp & Dohme and Novo Nordisk.

MJD has acted as consultant, advisory board member and speaker for Novo Nordisk, Sanofi-

Aventis, Lilly, Merck Sharp & Dohme, Boehringer Ingelheim, AstraZeneca and Janssen and

A cc

ep te

d A

rt ic

le as a speaker for Mitsubishi Tanabe Pharma Corporation. She has received grants in support

of investigator and investigator initiated trials from Novo Nordisk, Sanofi-Aventis and Lilly.

DRW has received grant in support of investigator initiated studies and honoraria from

Sanofi-Aventis and Novo Nordisk. All other Authors have no conflict of interests to disclose.

Funding Source: FZ is a Clinical Research Fellow funded with an Educational Grant from

Sanofi-Aventis to the University of Leicester. The research was supported by the National

Institute for Health Research (NIHR) Diet, Lifestyle & Physical Activity Biomedical

Research Unit based at University Hospitals of Leicester and Loughborough University; and

the NIHR CLARHC East Midlands. The views expressed are those of the authors and not

necessarily those of the Sanofi-Aventis, NHS, the NIHR or the Department of Health.

A cc

ep te

d A

rt ic

le References

[1] Wysham CH. New perspectives in type 2 diabetes, cardiovascular risk, and treatment goals. Postgrad Med. 2010; 122: 52-60 [2] Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycaemia in type 2 diabetes, 2015: a patient-centred approach. Update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetologia. 2015; 58: 429-442 [3] Madsbad S. Exenatide and liraglutide: different approaches to develop GLP-1 receptor agonists (incretin mimetics) - preclinical and clinical results. Best Practice and Research: Clinical Endocrinology and Metabolism. 2009; 23: 463-477 [4] Monami M, Dicembrini I, Marchionni N, Rotella CM, Mannucci E. Effects of glucagon-like peptide-1 receptor agonists on body weight: a meta-analysis. Experimental diabetes research. 2012; 2012: 672658 [5] Vilsboll T, Christensen M, Junker AE, Knop FK, Gluud LL. Effects of glucagon-like peptide-1 receptor agonists on weight loss: systematic review and meta-analyses of randomised controlled trials. BMJ. 2012; 344: d7771 [6] Sun F, Chai S, Li L, et al. Effects of Glucagon-Like Peptide-1 Receptor Agonists on Weight Loss in Patients with Type 2 Diabetes: A Systematic Review and Network Meta- Analysis. Journal of Diabetes Research. 2015; 2015: 157201 [7] Ahren B. Gut peptides and type 2 diabetes mellitus treatment. Curr Diab Rep. 2003; 3: 365-372 [8] Meier JJ. GLP-1 receptor agonists for individualized treatment of type 2 diabetes mellitus. Nature reviews Endocrinology. 2012; 8: 728-742 [9] Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia. 2012; 55: 1577-1596 [10] Zaccardi F, Htike ZZ, Webb DR, Khunti K, Davies MJ. Benefits and Harms of Once- Weekly Glucagon-like Peptide-1 Receptor Agonist Treatments: A Systematic Review and Network Meta-analysis. Ann Intern Med. 2016; 164: 102-113 [11] Hoaglin DC, Hawkins N, Jansen JP, et al. Conducting Indirect-Treatment- Comparison and Network-Meta-Analysis Studies: Report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: Part 2 Value in Health. 2011; 14: 429-437 [12] Zaccardi F, Webb DR, Htike ZZ, Youssef D, Khunti K, Davies MJ. Efficacy and safety of sodium-glucose co-transporter-2 inhibitors in type 2 diabetes mellitus: systematic review and network meta-analysis. Diabetes, obesity & metabolism. 2016; 18: 783-794 [13] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol. 2009; 62: 1006-1012 [14] Hutton B, Salanti G, Caldwell DM, et al. The PRISMA Extension Statement for Reporting of Systematic Reviews Incorporating Network Meta-analyses of Health Care Interventions: Checklist and ExplanationsPRISMA Extension for Network Meta-analysis. Annals of Internal Medicine. 2015; 162: 777-784 [15] Hawkins N, Scott DA, Woods B. How Far Do You Go? Efficient Searching for Indirect Evidence. Medical Decision Making. 2009: [16] Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ. 2011; 343: d5928 [17] White IR. Multivariate random-effects meta-regression: Updates to mvmeta. Stata Journal. 2011; 11: 255-270

A cc

ep te

d A

rt ic

le [18] White IR, Barrett JK, Jackson D, Higgins JPT. Consistency and inconsistency in network meta-analysis: model estimation using multivariate meta-regression. Research Synthesis Methods. 2012; 3: 111-125 [19] White IR. Network meta-analysis. Stata Journal. 2015; 15: 951-985 [20] Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PloS one. 2013; 8: e76654 [21] Jackson D, Barrett JK, Rice S, White IR, Higgins JP. A design-by-treatment interaction model for network meta-analysis with random inconsistency effects. Stat Med. 2014; 33: 3639-3654 [22] Buse JB, Henry RR, Han J, Kim DD, Fineman MS, Baron AD. Effects of exenatide (exendin-4) on glycemic control over 30 weeks in sulfonylurea-treated patients with type 2 diabetes. Diabetes Care. 2004; 27: 2628-2635 [23] DeFronzo RA, Ratner RE, Han J, Kim DD, Fineman MS, Baron AD. Effects of exenatide (exendin-4) on glycemic control and weight over 30 weeks in metformin-treated patients with type 2 diabetes. Diabetes Care. 2005; 28: 1092-1100 [24] Kendall DM, Riddle MC, Rosenstock J, et al. Effects of exenatide (exendin-4) on glycemic control over 30 weeks in patients with type 2 diabetes treated with metformin and a sulfonylurea. Diabetes Care. 2005; 28: 1083-1091 [25] Apovian CM, Bergenstal RM, Cuddihy RM, et al. Effects of exenatide combined with lifestyle modification in patients with type 2 diabetes. The American journal of medicine. 2010; 123: 468 e469-417 [26] Liutkus J, Rosas Guzman J, PNorwood P, et al. A placebo-controlled trial of exenatide twice daily added to thiazolidinediones alone or combination with metformin. Diabetes, obesity and metabolism. 2010; 12 1058-1065 [27] Ji L, Onishi Y, Ahn CW, et al. Efficacy and safety of exenatide once-weekly vs exenatide twice-daily in Asian patients with type 2 diabetes mellitus. Journal of Diabetes Investigation. 2013; 4: 53-61 [28] Marre M, Shaw J, Brandle M, et al. Liraglutide, a once-daily human GLP-1 analogue, added to a sulphonylurea over 26 weeks produces greater improvements in glycaemic and weight control compared with adding rosiglitazone or placebo in subjects with Type 2 diabetes (LEAD-1 SU). Diabetic medicine : a journal of the British Diabetic Association. 2009; 26: 268-278 [29] Nauck M, Frid A, Hermansen K, et al. Efficacy and safety comparison of liraglutide, glimepiride, and placebo, all in combination with metformin, in type 2 diabetes: the LEAD (liraglutide effect and action in diabetes)-2 study. Diabetes Care. 2009; 32: 84-90 [30] Zinman B, Gerich J, Buse JB, et al. Efficacy and safety of the human glucagon-like peptide-1 analog liraglutide in combination with metformin and thiazolidinedione in patients with type 2 diabetes (LEAD-4 Met+TZD). Diabetes Care. 2009; 32: 1224-1230 [31] Russell-Jones D, Vaag A, Schmitz O, et al. Liraglutide vs insulin glargine and placebo in combination with metformin and sulfonylurea therapy in type 2 diabetes mellitus (LEAD-5 met+SU): a randomised controlled trial. Diabetologia. 2009; 52: 2046-2055 [32] Buse JB, Rosenstock J, Sesti G, et al. Liraglutide once a day versus exenatide twice a day for type 2 diabetes: a 26-week randomised, parallel-group, multinational, open-label trial (LEAD-6). Lancet. 2009; 374: 39-47 [33] Lind M, Hirsch IB, Tuomilehto J, et al. Liraglutide in people treated for type 2 diabetes with multiple daily insulin injections: randomised clinical trial (MDI Liraglutide trial). BMJ. 2015; 351: [34] Ahmann A, Rodbard HW, Rosenstock J, et al. Efficacy and safety of liraglutide versus placebo added to basal insulin analogues (with or without metformin) in patients with

A cc

ep te

d A

rt ic

le type 2 diabetes: a randomized, placebo-controlled trial. Diabetes, obesity & metabolism. 2015; 17: 1056-1064 [35] Riddle MC, Forst T, Aronson R, et al. Adding once-daily lixisenatide for type 2 diabetes inadequately controlled with newly initiated and continuously titrated basal insulin glargine: a 24-week, randomized, placebo-controlled study (GetGoal-Duo 1). Diabetes Care. 2013; 36: 2497-2503 [36] Bolli GB, Munteanu M, Dotsenko S, et al. Efficacy and safety of lixisenatide once daily vs. placebo in people with Type 2 diabetes insufficiently controlled on metformin (GetGoal-F1). Diabetic medicine : a journal of the British Diabetic Association. 2014; 31: 176-184 [37] Riddle MC, Aronson R, Home P, et al. Adding once-daily lixisenatide for type 2 diabetes inadequately controlled by established basal insulin: a 24-week, randomized, placebo-controlled comparison (GetGoal-L). Diabetes Care. 2013; 36: 2489-2496 [38] Seino Y, Min KW, Niemoeller E, Takami A. Randomized, double-blind, placebo- controlled trial of the once-daily GLP-1 receptor agonist lixisenatide in Asian patients with type 2 diabetes insufficiently controlled on basal insulin with or without a sulfonylurea (GetGoal-L-Asia). Diabetes, obesity & metabolism. 2012; 14: 910-917 [39] Pinget M, Goldenberg R, Niemoeller E, Muehlen-Bartmer I, Guo H, Aronson R. Efficacy and safety of lixisenatide once daily versus placebo in type 2 diabetes insufficiently controlled on pioglitazone (GetGoal-P). Diabetes, obesity & metabolism. 2013; 15: 1000- 1007 [40] Rosenstock J, Hanefeld M, Shamanna P, et al. Beneficial effects of once-daily lixisenatide on overall and postprandial glycemic levels without significant excess of hypoglycemia in type 2 diabetes inadequately controlled on a sulfonylurea with or without metformin (GetGoal-S). J Diabetes Complications. 2014; 28: 386-392 [41] Rosenstock J, Raccah D, Koranyi L, et al. Efficacy and safety of lixisenatide once daily versus exenatide twice daily in type 2 diabetes inadequately controlled on metformin: a 24-week, randomized, open-label, active-controlled study (GetGoal-X). Diabetes Care. 2013; 36: 2945-2951 [42] Drucker DJ, Buse JB, Taylor K, et al. Exenatide once weekly versus twice daily for the treatment of type 2 diabetes: a randomised, open-label, non-inferiority study. The Lancet. 2008; 372: 1240-1250 [43] Blevins T, Pullman J, Malloy J, et al. DURATION-5: exenatide once weekly resulted in greater improvements in glycemic control compared with exenatide twice daily in patients with type 2 diabetes. The Journal of clinical endocrinology and metabolism. 2011; 96: 1301- 1310 [44] Buse JB, Nauck M, Forst T, et al. Exenatide once weekly versus liraglutide once daily in patients with type 2 diabetes (DURATION-6): a randomised, open-label study. Lancet. 2013; 381: 117-124 [45] Pratley RE, Nauck MA, Barnett AH, et al. Once-weekly albiglutide versus once-daily liraglutide in patients with type 2 diabetes inadequately controlled on oral drugs (HARMONY 7): a randomised, open-label, multicentre, non-inferiority phase 3 study. The lancet Diabetes & endocrinology. 2014; 2: 289-297 [46] Wysham C, Blevins T, Arakaki R, et al. Efficacy and safety of dulaglutide added onto pioglitazone and metformin versus exenatide in type 2 diabetes in a randomized controlled trial (AWARD-1). Diabetes Care. 2014; 37: 2159-2167 [47] Nauck M, Weinstock RS, Umpierrez GE, Guerci B, Skrivanek Z, Milicevic Z. Efficacy and safety of dulaglutide versus sitagliptin after 52 weeks in type 2 diabetes in a randomized controlled trial (AWARD-5). Diabetes Care. 2014; 37: 2149-2158

A cc

ep te

d A

rt ic

le [48] Dungan KM, Povedano ST, Forst T, et al. Once-weekly dulaglutide versus once-daily liraglutide in metformin-treated patients with type 2 diabetes (AWARD-6): a randomised, open-label, phase 3, non-inferiority trial. Lancet. 2014; 384: 1349-1357 [49] Dungan KM, Weitgasser R, Perez Manghi F, et al. A 24-week study to evaluate the efficacy and safety of once-weekly dulaglutide added on to glimepiride in type 2 diabetes (AWARD-8). Diabetes, obesity & metabolism. 2016; 18: 475-482 [50] Raz I, Fonseca V, Kipnes M, et al. Efficacy and safety of taspoglutide monotherapy in drug-naive type 2 diabetic patients after 24 weeks of treatment: results of a randomized, double-blind, placebo-controlled phase 3 study (T-emerge 1). Diabetes Care. 2012; 35: 485- 487 [51] Henry RR, Mudaliar S, Kanitra L, Woloschak M, Balena R. Efficacy and safety of taspoglutide in patients with type 2 diabetes inadequately controlled with metformin plus pioglitazone over 24 weeks: T-emerge 3 trial. Journal of Clinical Endocrinology and Metabolism. 2012; 97: 2370-2379 [52] Bergenstal RM, Forti A, Chiasson JL, Woloschak M, Boldrin M, Balena R. Efficacy and safety of taspoglutide versus sitagliptin for type 2 diabetes mellitus (T-emerge 4 trial). Diabetes Therapy Research, Treatment and Education of Diabetes and Related Disorders. 2012; 3: 13 [53] Hollander P, Lasko B, Barnett AH, et al. Effects of taspoglutide on glycemic control and body weight in obese patients with type 2 diabetes (T-emerge 7 study). Obesity. 2013; 21: 238-247 [54] Rosenstock J, Balas B, Charbonnel B, et al. The fate of taspoglutide, a weekly GLP-1 receptor agonist, versus twice-daily exenatide for type 2 diabetes: the T-emerge 2 trial. Diabetes Care. 2013; 36: 498-504 [55] Moretto TJ, Milton DR, Ridge TD, et al. Efficacy and tolerability of exenatide monotherapy over 24 weeks in antidiabetic drug-naive patients with type 2 diabetes: a randomized, double-blind, placebo-controlled, parallel-group study. Clinical therapeutics. 2008; 30: 1448-1460 [56] Potts JE, Gray LJ, Brady EM, Khunti K, Davies MJ, Bodicoat DH. The Effect of Glucagon-Like Peptide 1 Receptor Agonists on Weight Loss in Type 2 Diabetes: A Systematic Review and Mixed Treatment Comparison Meta-Analysis. PloS one. 2015; 10: e0126769 [57] Sun F, Wu S, Guo S, et al. Impact of GLP-1 receptor agonists on blood pressure, heart rate and hypertension among patients with type 2 diabetes: A systematic review and network meta-analysis. Diabetes research and clinical practice. 2015; 110: 26-37 [58] Sun F, Chai S, Yu K, et al. Gastrointestinal Adverse Events of Glucagon-Like Peptide-1 Receptor Agonists in Patients with Type 2 Diabetes: A Systematic Review and Network Meta-Analysis Diabetes Technology & Therapeutics. 2015; 17: 35-42 [59] Zaccardi F, Pitocco D, Ghirlanda G. Glycemic risk factors of diabetic vascular complications: the role of glycemic variability. Diabetes/Metabolism Research and Reviews. 2009; 25: 199-207 [60] Arnolds S, Dellweg S, Clair J, et al. Further improvement in postprandial glucose control with addition of exenatide or sitagliptin to combination therapy with insulin glargine and metformin: a proof-of-concept study. Diabetes Care. 2010; 33: 1509-1515 [61] Kapitza C, Forst T, Coester HV, Poitiers F, Ruus P, Hincelin-Mery A. Pharmacodynamic characteristics of lixisenatide once daily versus liraglutide once daily in patients with type 2 diabetes insufficiently controlled on metformin. Diabetes, obesity & metabolism. 2013; 15: 642-649 [62] Meier JJ, Rosenstock J, Hincelin-Mery A, et al. Contrasting Effects of Lixisenatide and Liraglutide on Postprandial Glycemic Control, Gastric Emptying, and Safety Parameters

A cc

ep te

d A

rt ic

le in Patie A Rand [63] and Acu 2257 [64] Outcom [65] Patients [66] network

Figure

Figure Differen HbA1c, diastolic ALB, al LIR, lira

ents With Ty domized, Op Pfeffer MA ute Coronar

Marso SP, D mes in Type

Marso SP, s with Type Cipriani A, k meta-anal

e Legends

1: Differen nces estimate glycated ha

c blood press lbiglutide; D aglutide; LIX

ype 2 Diabe pen-Label T

A, Claggett B ry Syndrom

Daniels GH 2 Diabetes Bain SC, C

e 2 Diabetes , Higgins JP lysis. Ann In

nces vs place ed with netwo aemoglobin; sure; HR, hea DUL, dulaglu XI, lixisenati

etes on Opti Trial. Diabet B, Diaz R, e

me. The New

H, Brown-Fr . The New E onsoli A, et

s. New Engl P, Geddes JR ntern Med.

ebo for card ork meta-ana FPG, fastin

art rate utide; EBID ide; TAS, ta

imized Insu tes Care. 20 et al. Lixise

w England jo

randsen K, England jou t al. Semagl land Journa R, Salanti G 2013; 159:

diometaboli alysis vs Plac ng plasma g

D, twice-dail spoglutide

ulin Glargine 015; 38: 126 enatide in Pa ournal of m

et al. Liragl urnal of med lutide and C l of Medicin

G. Conceptu 130-137

ic outcomes cebo.

glucose; SBP

y exenatide;

e With or W 63-1273 atients with

medicine. 20

lutide and C dicine. 2016

Cardiovascu ne. 2016; 0

ual and techn

P, systolic b

; EQW, onc

Without Met

h Type 2 Dia 15; 373: 22

Cardiovascu 6; 375: 311 ular Outcom : null nical challe

lood pressur

ce-weekly ex

tformin:

abetes 247-

ular -322

mes in

enges in

re; DBP,

xenatide;

A cc

ep te

d A

rt ic

Figure Differen ALB, al LIR, lira

2: Differen nces estimate lbiglutide; D aglutide; LIX

nces vs place ed with netwo DUL, dulaglu XI, lixisenati

ebo for safe ork meta-ana utide; EBID ide; TAS, ta

ety outcome alysis vs Plac

D, twice-dail spoglutide.

es cebo. y exenatide;; EQW, oncce-weekly exxenatide;

A cc

ep te

d A

rt ic

le Table 1: Baseline characteristics of the included studies

Drugs Year Study First Author Background Therapy Study duration

(weeks) Total

participants (N) Females

(%) Age^ (years)

HbA1c^ (%)

Diabetes duration^ (years)

Short‐acting GLP1‐RAs vs Placebo EBID, PLA 2004 NCT 00765817 Buse SU 30 259 42.9 59.0 8.4 12.0 EBID, PLA 2005 PMID 15855571 Kendall MET+SU 30 488 42.4 55.5 8.5 9.1 EBID, PLA 2005 PMID 15855572 DeFronzo MET 30 226 40.3 53.0 8.2 5.8 EBID, PLA 2008 NCT 00381342 Moretto Diet + PA 24 155 42.0 54.0 7.8 1.5 EBID, PLA 2010 PMID 20399326 Apovian Diet + PA 24 194 62.5 54.8 7.6 5.5 EBID, PLA 2010 PMID 20977576 Liutkus TZD±MET 26 165 40.6 54.7 8.2 6.3 LIXI, PLA 2012 GetGoal‐L‐Asia Seino Basal ± SU 24 311 52.1 58.3 8.5 13.9 LIXI, PLA 2013 GetGoal‐Duo 1 Riddle Basal + MET ± TZD 24 446 50.0 56.0 7.6 9.2 LIXI, PLA 2013 GetGoal‐L Riddle Basal ± MET 24 495 46.3 57.0 8.4 12.5 LIXI, PLA 2013 GetGoal‐P Pinget PIO ± MET 24 484 47.7 55.8 8.1 8.1 LIXI, PLA 2014 GetGoal‐F1 Bolli MET 24 482 55.3 56.1 8.1 6.0 LIXI, PLA 2014 GetGoal‐S Rosenstock SU ± MET 24 859 49.5 57.3 8.3 9.3 Long‐acting GLP1‐RAs vs Placebo LIR, PLA 2009 LEAD 1 SU Marre SU 26 348 48.8 55.3 8.5 6.5# LIR, PLA 2009 LEAD 2 Nauck MET 26 363 40.7 56.5 8.4 8.0 LIR, PLA 2009 LEAD 4 Met TZD Zinman MET ± ROS 26 355 43.5 55.0 8.5 9.0 LIR, PLA 2009 LEAD 5 Met + SU Russell‐Jones MET + SU 26 346 46.6 57.6 8.3 9.3 LIR, PLA 2015 MDI Liraglutide trial Lind MDI ± MET 24 124 35.5 63.6 9.0 17.2 LIR, PLA 2015 NCT 01617434 Ahmann Basal ± MET 26 450 43.2 58.4 8.3 ‐ DUL, PLA 2014 AWARD 5 Nauck MET 26 481 50.9 54.4 8.1 7.0 DUL, PLA 2016 AWARD 8 Dungan SU 24 299 55.9 57.8 8.4 7.6 TAS, PLA 2012 T‐Emerge 1 Raz Diet + PA 24 252 63.5 55.4 7.7 2.2 TAS, PLA 2012 T‐Emerge 3 Henry MET+TZD 24 207 48.3 55.0 8.1 7.9 TAS, PLA 2012 T‐Emerge 4 Bergenstal MET 24 277 48.0 56.3 8.0 5.6 TAS, PLA 2013 T‐Emerge 7 Hollander MET 24 292 59.2 53.5 7.6 5.1 Comparisons between GLP‐1RAs ALB*, LIR 2014 HARMONY 7 Pratley MET ± SU ± TZD 32 812 50.0 55.6 8.2 8.4 DUL, EBID, PLA 2014 AWARD 1 Wysham MET + PIO 26, 52~ 696 42.6 55.4 8.1 9.0 DUL, LIR 2014 AWARD 6 Dungan MET 26 599 52.0 56.7 8.1 7.2 EBID, EQW 2008 DURATION 1 Drucker Diet/PA/MET/SU/TZD 30 295 46.8 55.0 8.3 6.5 EBID, EQW 2011 DURATION 5 Belvins MET ± SU ± TZD 24 252 42.4 55.5 8.5 7.0 EBID, EQW 2013 NCT 00917267 Ji MET ± SU ± TZD 26 678 45.9 55.5 8.7 8.2 LIR, EBID 2009 LEAD 6 Buse MET ± SU 26 464 48.0 56.7 8.2 8.2 LIR, EQW 2013 DURATION 6 Buse MET ± SU ± PIO 26 911 45.5 57.0 8.5 8.5 LIXI, EBID 2013 GetGoal‐X Rosenstock MET 24 634 46.7 57.4 8.0 6.8 TAS, EBID 2013 T‐Emerge 2 Rosenstock MET ± TZD 24, 52~ 765 49.5 55.5 8.1 6.8

A cc

ep te

d A

rt ic

le ^ When not reported for the overall population, values have been estimated as weighted means; ~ Shorter follow‐up data used; * 30 to 50mg; # median

ALB, albiglutide; DUL, dulaglutide; EBID, twice‐daily exenatide; EQW, once‐weekly exenatide; LIR, liraglutide; LIXI, lixisenatide; TAS, taspoglutide; PLA, placebo. OADs, oral glucose–lowering drugs; PA, physical activity; PIO, pioglitazone; ROS, rosiglitazone; INS, insulin therapy; Basal, basal insulin; MET, metformin; TZD, thiazolidinedione; SU, sulphonylurea; MDI, multiple daily injection; PMID, PubMed identification number

A cc

ep te

d A

rt ic

le Table 2: Comparisons of GLP‐1 RAs for cardiometabolic outcomes

Total Cholesterol (mmol/l) LDL Cholesterol (mmol/l)

TAS TAS

LIXI ‐ LIXI ‐

LIR ‐ 0.09

(‐0.11,0.29) LIR ‐

0.04 (‐0.11,0.19)

EQW 0.02

(‐0.12,0.16) ‐

0.11 (‐0.10,0.32)

EQW 0.02

(‐0.08,0.12) ‐

0.05 (‐0.11,0.22)

EBID ‐0.14

(‐0.26,‐0.03) ‐0.13

(‐0.25,‐0.00) ‐

‐0.04 (‐0.22,0.15)

EBID ‐0.07

(‐0.16,0.01) ‐0.05

(‐0.15,0.04) ‐

‐0.02 (‐0.16,0.13)

DUL 0.15

(0.00,0.30) 0.01

(‐0.17,0.18) 0.03

(‐0.12,0.17) ‐

0.12 (‐0.10,0.33)

DUL 0.10

(‐0.01,0.21) 0.03

(‐0.10,0.15) 0.04

(‐0.06,0.15) ‐

0.08 (‐0.08,0.24)

ALB 0.06

(‐0.18,0.29) 0.21

(‐0.01,0.43) 0.07

(‐0.16,0.29) 0.08

(‐0.10,0.27) ‐

0.18 (‐0.09,0.44)

ALB 0.12

(‐0.05,0.28) 0.21

(0.06,0.37) 0.14

(‐0.02,0.30) 0.16

(0.03,0.29) ‐

0.20 (‐0.00,0.40)

HDL Cholesterol (mmol/l) Triglycerides (mmol/l)

TAS TAS

LIXI ‐ LIXI ‐

LIR ‐

‐0.02 (‐0.06,0.02)

LIR ‐ 0.17

(‐0.06,0.40)

EQW

0.01 (‐0.02,0.03)

‐ ‐0.01

(‐0.06,0.03) EQW

‐0.18 (‐0.35,‐0.00)

‐ ‐0.01

(‐0.21,0.20)

EBID

0.01 (‐0.01,0.02)

0.01 (‐0.01,0.03)

‐ ‐0.01

(‐0.05,0.03) EBID

‐0.00 (‐0.00,0.00)

‐0.18 (‐0.35,‐0.00)

‐ ‐0.01

(‐0.21,0.20)

DUL

‐0.01 (‐0.03,0.01)

‐0.01 (‐0.03,0.02)

‐0.00 (‐0.02,0.02)

‐ ‐0.02

(‐0.07,0.03) DUL

0.18 (0.03,0.34)

0.01 (‐0.16,0.18)

‐ 0.18

(‐0.06,0.41)

ALB ‐0.02

(‐0.05,0.01) ‐0.03

(‐0.06,0.00) ‐0.02

(‐0.06,0.01) ‐0.02

(‐0.04,0.00) ‐

‐0.04 (‐0.09,0.01)

ALB ‐0.08

(‐0.32,0.16) 0.10

(‐0.13,0.34) 0.10

(‐0.13,0.34) ‐0.07

(‐0.24,0.09) ‐

0.10 (‐0.19,0.38)

Systolic Blood Pressure (mmHg) Diastolic Blood Pressure (mmHg)

TAS TAS

LIXI ‐ LIXI ‐

LIR ‐ 2.40

(0.21,4.59) LIR ‐

0.53 (‐0.85,1.90)

EQW ‐0.40

(‐1.69,0.88) ‐

2.00 (‐0.40,4.39)

EQW 0.24

(‐0.60,1.08) ‐

0.77 (‐0.73,2.26)

EBID ‐0.21

(‐1.51,1.09) ‐0.61

(‐1.92,0.69) ‐

1.79 (‐0.47,4.04)

EBID 0.64

(‐0.20,1.48) 0.88

(0.04,1.72) ‐

1.41 (0.03,2.78)

DUL ‐0.08

(‐1.54,1.39) ‐0.29

(‐1.95,1.38) ‐0.69

(‐2.02,0.64) ‐

1.71 (‐0.54,3.96)

DUL ‐1.07

(‐1.99,‐0.15) ‐0.43

(‐1.48,0.63) ‐0.18

(‐1.01,0.64) ‐

0.34 (‐1.04,1.72)

ALB ‐ ‐ ‐ ‐ ‐ ‐ ALB ‐ ‐ ‐ ‐ ‐ ‐

A cc

ep te

d A

rt ic

le Table 2 (cont’d): Comparisons of GLP‐1 RAs for cardiometabolic outcomes

Heart Rate (bpm) HbA1c (%) TAS TAS

LIXI ‐

LIXI ‐0.45

(‐0.64,‐0.26)

LIR ‐3.48

(‐5.01,‐1.96) ‐

LIR

0.60 (0.43,0.77)

0.15 (‐0.03,0.34)

EQW 0.03

(‐1.68,1.74) ‐3.45

(‐5.55,‐1.36) ‐

EQW

‐0.07 (‐0.26,0.12)

0.53 (0.31,0.75)

0.09 (‐0.14,0.32)

EBID 2.18

(0.92,3.44) 2.21

(1.05,3.37) ‐1.27

(‐2.95,0.40) ‐

EBID

‐0.38 (‐0.55,‐0.21)

‐0.45 (‐0.59,‐0.31)

0.15 (‐0.00,0.30)

‐0.30 (‐0.46,‐0.13)

DUL ‐1.51

(‐2.61,‐0.41) 0.67

(‐1.00,2.34) 0.70

(‐0.16,1.55) ‐2.79

(‐4.32,‐1.25) ‐ DUL

0.51 (0.34,0.68)

0.13 (‐0.10,0.35)

0.06 (‐0.11,0.23)

0.66 (0.46,0.86)

0.21 (0.00,0.42)

ALB ‐ ‐ ‐ ‐ ‐ ‐ ALB ‐0.27

(‐0.59,0.06) 0.24

(‐0.07,0.55) ‐0.14

(‐0.48,0.19) ‐0.21

(‐0.49,0.07) 0.39

(0.07,0.72) ‐0.06

(‐0.39,0.28)

Fasting Plasma Glucose (mmol/l) Body Weight (kg) TAS TAS

LIXI ‐0.95

(‐1.45,‐0.45)

LIXI

‐0.75 (‐1.86,0.36)

LIR 1.13

(0.66,1.60) 0.18

(‐0.30,0.66)

LIR

1.17 (0.19,2.15)

0.42 (‐0.69,1.53)

EQW ‐0.08

(‐0.58,0.41) 1.05

(0.47,1.63) 0.10

(‐0.49,0.68)

EQW

‐0.47 (‐1.59,0.65)

0.71 (‐0.55,1.97)

‐0.04 (‐1.40,1.32)

EBID ‐0.85

(‐1.28,‐0.41) ‐0.93

(‐1.31,‐0.56) 0.20

(‐0.22,0.61) ‐0.75

(‐1.18,‐0.32)

EBID

0.18 (‐0.80,1.16)

‐0.29 (‐1.12,0.54)

0.89 (0.01,1.76)

0.14 (‐0.87,1.14)

DUL 1.04

(0.54,1.53) 0.19

(‐0.44,0.81) 0.10

(‐0.39,0.60) 1.23

(0.65,1.81) 0.28

(‐0.31,0.87) DUL

‐0.10 (‐1.13,0.92)

0.07 (‐1.27,1.42)

‐0.39 (‐1.42,0.64)

0.78 (‐0.35,1.92)

0.03 (‐1.22,1.28)

ALB ‐0.56

(‐1.46,0.33) 0.47

(‐0.36,1.31) ‐0.38

(‐1.27,0.52) ‐0.46

(‐1.20,0.28) 0.67

(‐0.21,1.55) ‐0.28

(‐1.17,0.61) ALB

‐1.16 (‐3.21,0.89)

‐1.26 (‐3.22,0.69)

‐1.08 (‐3.18,1.01)

‐1.55 (‐3.32,0.22)

‐0.38 (‐2.40,1.65)

‐1.13 (‐3.22,0.96)

Data are reported as mean difference (95% confidence interval) and indicate column‐to‐row differences (i.e. compared with lixisenatide, taspoglutide reduces HbA1c by 0.45%). Statistically significant differences are in bold. HbA1c, glycated haemoglobin; ALB, albiglutide; DUL, dulaglutide; EBID, twice‐daily exenatide; EQW, once‐weekly exenatide; LIR, liraglutide; LIXI, lixisenatide; TAS, taspoglutide. P‐values indicate statistical inconsistency for the network

A cc

ep te

d A

rt ic

le Table 3: Comparisons of GLP‐1 RAs for safety outcomes

Hypoglycaemia Nausea

TAS TAS

LIXI 2.35

(0.88,6.24) LIXI

1.70 (0.78,3.67)

LIR 0.93

(0.54,1.59) 2.19

(0.81,5.89) LIR

0.82 (0.48,1.39)

1.39 (0.61,3.14)

EQW 0.91

(0.52,1.58) 0.85

(0.43,1.66) 1.98

(0.68,5.83) EQW

2.64 (1.60,4.36)

2.17 (1.22,3.85)

3.67 (1.56,8.62)

EBID 0.75

(0.45,1.24) 0.68

(0.44,1.04) 0.63

(0.38,1.03) 1.48

(0.56,3.94) EBID

0.42 (0.28,0.64)

1.12 (0.74,1.70)

0.92 (0.60,1.40)

1.55 (0.73,3.33)

DUL 0.92

(0.50,1.67) 0.69

(0.33,1.42) 0.62

(0.34,1.13) 0.58

(0.29,1.16) 1.36

(0.46,4.02) DUL

0.80 (0.46,1.39)

0.34 (0.18,0.65)

0.90 (0.52,1.56)

0.74 (0.39,1.40)

1.25 (0.51,3.06)

ALB 2.07

(0.76,5.66) 1.90

(0.76,4.76) 1.42

(0.53,3.78) 1.29

(0.57,2.89) 1.20

(0.45,3.17) 2.81

(0.78,10.11) ALB

4.17 (1.64,10.60)

3.35 (1.41,7.93)

1.42 (0.57,3.50)

3.75 (1.76,7.96)

3.07 (1.22,7.69)

5.20 (1.71,15.80)

Vomiting Diarrhoea

TAS TAS

LIXI 2.66

(1.01,7.01) LIXI

1.03 (0.46,2.31)

LIR 0.73

(0.40,1.34) 1.94

(0.72,5.24) LIR

0.45 (0.28,0.73)

0.46 (0.20,1.07)

EQW 2.39

(1.44,3.95) 1.74

(0.92,3.31) 4.64

(1.68,12.77) EQW

1.49 (0.93,2.39)

0.67 (0.38,1.16)

0.69 (0.28,1.67)

EBID 0.52

(0.34,0.79) 1.23

(0.81,1.88) 0.90

(0.55,1.48) 2.39

(0.94,6.07) EBID

1.00 (0.66,1.52)

1.49 (1.01,2.20)

0.67 (0.45,1.01)

0.69 (0.31,1.56)

DUL 0.71

(0.41,1.21) 0.37

(0.19,0.70) 0.87

(0.50,1.52) 0.64

(0.31,1.29) 1.70

(0.59,4.87) DUL

0.62 (0.37,1.05)

0.62 (0.34,1.16)

0.93 (0.56,1.54)

0.42 (0.23,0.76)

0.43 (0.17,1.07)

ALB 2.26

(0.90,5.68) 1.60

(0.68,3.74) 0.83

(0.34,2.02) 1.97

(0.94,4.12) 1.44

(0.55,3.74) 3.83

(1.11,13.19) ALB

0.96 (0.44,2.10)

0.60 (0.29,1.22)

0.60 (0.28,1.29)

0.89 (0.49,1.63)

0.40 (0.19,0.87)

0.41 (0.15,1.16)

Injection Site Reaction

TAS

LIXI 9.59

(0.44,211.15)

LIR 8.52

(1.20,60.30) 81.64

(2.14,3109.93)

EQW 0.17

(0.04,0.68) 1.43

(0.32,6.39) 13.67

(0.44,420.21)

EBID 3.80

(1.81,7.99) 0.64

(0.13,3.01) 5.43

(1.43,20.61) 52.01

(1.81,1495.96)

DUL 2.52

(0.21,30.87) 9.60

(0.82,112.97) 1.61

(0.17,15.51) 13.68

(1.01,184.93) 131.18

(2.41,7148.37)

ALB 0.10

(0.01,1.47) 0.26

(0.03,2.08) 1.00

(0.14,7.07) 0.17

(0.04,0.66) 1.42

(0.13,15.54) 13.60

(0.28,666.59)

A cc

ep te

d A

rt ic

Data are reported as odds ratio (95% confidence interval) and indicate column‐to‐row ratios [i.e. compared with albiglutide, lixisenatide is associated with an odds ratio of 3.07 of nausea; compared to albiglutide, lixisenatide is associated with an odds ratio of 0.40 of diarrhoea, or equivalently albiglutide increases the risk by an odds ratio of 2.50 (=1/0.40)]. Statistically significant differences are in bold. ALB, albiglutide; DUL, dulaglutide; EBID, twice‐daily exenatide; EQW, once‐weekly exenatide; LIR, liraglutide; LIXI, lixisenatide; TAS, taspoglutide P‐values indicate statistical inconsistency for the network