Cardiovascular and renal outcomes of GLP-1 receptor agonists vs. DPP-4 inhibitors and basal insulin in type 2 diabetes mellitus: A systematic review and meta-analysis

Abstract

Objective

To compare the cardiovascular and renal outcomes of GLP-1 RA versus DPP4i and basal insulin in the management of T2DM.

Methods

Data from 22 studies involving over 200,000 participants were pooled using the inverse variance method and random-effects meta-analysis. The review was reported in accordance with PRISMA.

Results

Compared with DPP4i, treatment with GLP-1 RA was associated with a greater benefit on composite cardiovascular outcomes (HR:0.77, 95% CI:0.69–0.87), myocardial infarction (HR:0.82, 95% CI:0.69–0.97), stroke (HR:0.83, 95% CI: 0.74–0.93), cardiovascular mortality (HR:0.76, 95% CI:0.68–0.85) and all-cause mortality (HR:0.65, 95% CI:0.48–0.90). There was no difference in effect on heart failure (HR:0.97, 95% CI:0.82–1.15). Compared with basal insulin, GLP-1 RA was associated with better effects on composite cardiovascular outcomes (HR:0.62, 95% CI:0.48–0.79), heart failure (HR:0.57, 95% CI:0.35–0.92), myocardial infarction (HR:0.70, 95% CI:0.58–0.85), stroke (HR:0.50, 95% CI:0.40–0.63) and all-cause mortality (HR:0.31, 95% CI:0.20–0.48). Evidence from a small number of studies suggests that GLP-1 RA had better effects on composite and individual renal outcomes, such as eGFR, compared with either DPP4i and basal insulin.

Conclusion

Available evidence suggests that treating T2DM with GLP-1 RA can yield better benefits on composite and specific cardiorenal outcomes than with DPP4i and basal insulin.

PROSPERO Registration Number

CRD42022335504.

Keywords

GLP-1 receptor agonists DPP-4 inhibitors basal insulin diabetes mellitus cardiovascular outcomes renal safety

Key messages

What is already known on this topic

There is a dearth of evidence on the relative benefits of GLP-1 RA compared to DPP4i and basal insulin in the management of T2DM.

What this study adds

Our study aims to fill this knowledge gap and provides a comprehensive synthesis of the available evidence on this topic, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

How this study might affect research, practice or policy

Available evidence suggests that treating T2DM with GLP-1 RA can yield better benefits on composite and specific cardiorenal outcomes than with DPP4i and basal insulin.

Introduction

Diabetes mellitus (DM) poses a growing burden of morbidity and death globally. Prevalence is estimated to be 9.8% (537 million people) worldwide and is expected to rise to 10.8% (643 million) by 2030.¹ It is among the top 10 causes of death in adults, causing 6.7 million deaths in 2021.^1,2 It is well established that people with DM carry a significantly high burden of cardiovascular diseases (CVD).³ In particular, type 2 diabetes mellitus (T2DM) is associated with a 2–3 times increased risk of developing cardiovascular abnormalities, while CVD is the leading cause of mortality in people with T2DM.⁴ Given the proven risks of CVD among people with diabetes, the role of glucose-lowering agents in reducing cardiovascular risks in this population is increasingly being explored. People with T2DM also carry a high burden of adverse renal outcomes, with diabetic nephropathy being the leading cause of chronic kidney disease (CKD) in this population.⁵

Since 2008, evidence from randomised controlled trials (RCTs) and observational studies of people with DM has demonstrated the beneficial cardiovascular effects of glucose-lowering drugs, including newer ones like glucagon-like peptide-1 receptor agonists (GLP-1 RA).^6–9 However, the current body of evidence is inconclusive regarding which drug class offers superior cardiovascular benefits. A previous meta-analysis reported that GLP-1 RA was more beneficial in reducing major adverse cardiovascular events (MACE) including non-fatal myocardial infarction, non-fatal stroke and cardiovascular death; and other endpoints such as all-cause mortality, peripheral artery disease, and heart failure compared with other glucose-lowering drugs, with the exception of sodium-glucose transporter-2 inhibitors (SGLT2i).¹⁰ Nonetheless, a large cohort study found no superiority of GLP-1 RA over comparators such as dipeptidyl peptidase-4 inhibitors (DPP4i), sulphonylureas (SU), and insulin.⁹ The evidence of insulin is less certain, with some recent evidence suggesting that insulin therapy increases cardiovascular risk in T2DM.¹¹

Available evidence also suggests that glucose-lowering drugs might offer renal protection in the context of T2DM. Studies investigating renal outcomes have reported significant benefits of therapies such as GLP-1 RA and DPP4i against the risks of new-onset micro- or macroalbuminuria, and estimated glomerular filtration rate (eGFR) reduction, dialysis, renal-replacement therapy, hospitalisation, and death due to renal causes.^12–14 Nonetheless, the extent to which these glucose-lowering therapies can offer renal outcome benefits in individuals with T2DM remains unclear.

Given these uncertainties, and owing to the availability of new data from recent observational studies and availability of head-to-head clinical trial data, this meta-analysis seeks to update the current evidence base by considering evidence across both observational studies and clinical trials. This will contribute to a better understanding of cardiovascular and renal outcomes of antidiabetic medications and will help inform future guideline recommendations on the use of glucose-lowering therapies to improve cardiorenal outcomes in adults with T2DM.

Therefore, the primary objective of the review was to compare the cardiovascular and renal outcomes of GLP-1 RA versus DPP4i and basal insulin in the management of T2DM. The secondary objectives were to: (i) compare the cardiovascular and renal efficacy/effectiveness of GLP-1 RA, DPP4i and basal insulin in the management of T2DM; and (ii) compare the therapeutic safety of GLP-1 RA, DPP4i and basal insulin in people with T2DM. This review’s head-to-head comparison of GLP-1 RA and DPP4i was informed by several considerations. When T2DM patients are not tolerating or have contraindications to SGLT2i, and are faced with choosing an alternative therapy from the newer drugs, it is likely a choice between GLP-1 RA and DPP4i. Moreover, since it is not recommended that GLP-1 RA and DPP4i be used in combination, and either of them can be used in combination with SGLT2i; it is reasonable to compare them head-to-head without adding SGLT2i as a comparator in this review. The comparison of GLP-1 RA and basal insulin was informed by the need to evaluate how basal insulin compares to a newer injectable treatment option (GLP-1 RA) in terms of cardiorenal benefits.

Methods

Study design

This protocol was developed and reported in accordance with the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement.¹⁵ The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO, registration number: CRD42022335504).¹⁶

Eligibility criteria

Eligible participants were adults aged 18 years and older with T2DM. Intervention was glucose control using GLP-1 RA to achieve guideline-recommended glycaemic targets, while comparators included DPP4i or basal insulin treatment to achieve guideline-recommended glycaemic targets. The primary cardiovascular outcomes of interest were composite cardiovascular outcomes such as major adverse cardiovascular events (MACEs); and single endpoints, such as non-fatal stroke, non-fatal myocardial infarction, peripheral artery disease, heart failure, cardiovascular death, and all-cause mortality. Renal outcomes included composite kidney endpoints, comprising new-onset macroalbuminuria, doubling of serum creatinine, reduction in eGFR, dialysis, renal-replacement therapy, hospitalisation, and death due to renal causes. Secondary outcomes included reported short- and long-term safety and tolerability outcomes.

Study design

Eligible studies included observational studies reporting head-to-head comparisons of therapeutic effectiveness and/or safety of antidiabetic drugs belonging to the GLP-1 RA, DPP4i or basal insulin classes. RCTs, including cardiovascular outcome trials (CVOTs) with head-to-head comparisons between GLP-1 RA vs DPP4i or basal insulin, were also considered for inclusion. Eligibility was restricted to studies with at least 300 participants and at least 3 months follow-up period.

Search strategy

The search strategy was developed with the guidance of a health sciences subject matter librarian. Search strings were designed to be sensitive to the tautological array of alternative terminologies and keywords related to T2DM, GLP-1 RA, DPP4i, basal insulin and cardiorenal outcomes. The search was restricted by applying filters to retrieve cardiovascular outcomes trials (CVOTs) and studies published from 2008 (when the first formulation of FDA guidance called for evaluating the cardiovascular safety of glucose-lowering therapies) to date.^17,18 Only studies published in English were eligible for inclusion.

Electronic databases searched included Embase, MEDLINE (via PubMed), Cochrane Library (including the Cochrane Central Register of Controlled Trials (CENTRAL) and the Database of Abstracts of Reviews of Effects (DARE)), https://ClinicalTrials.gov and the International Clinical Trials Registry Platform (ICTRP). A pilot search strategy was tested and optimised in PubMed and adapted for other databases using appropriate search syntaxes. The Open Grey literature database was also searched to identify grey literature. Additional databases searched for grey literature included websites and archives of the American Diabetes Association (ADA), European Association for the Study of Diabetes (EASD) and Diabetologia. Conference abstracts were considered for inclusion, provided they reported sufficient data and sample size. Hand searches of reference lists of relevant studies were conducted to identify additional potentially eligible records. See supplementary file for search strategies as applied across databases.

Study selection

Search outputs retrieved from the databases were collated using EndNote software for duplicate records removal. Two reviewers (PK and CA) independently screened titles and abstracts of every unique record retrieved from the search for potentially eligible studies. Following title and abstract screening, full texts of potentially eligible studies were retrieved for full-text assessment. Two reviewers (PK and CA) again independently reviewed full-text records to assess eligibility for inclusion in the review using pre-defined inclusion criteria. Disagreements were resolved through discussion or, if required, a third reviewer (CN) was consulted to arbitrate. A PRISMA flow chart¹⁹ was used to describe the study selection process.

Data extraction

We created a pre-designed data extraction tool to capture study characteristics and outcome data. Two reviewers (PK and CA) extracted the relevant data from the included studies, and a third reviewer (CN) checked the extracted data for accuracy. The following study-level data were extracted: (i) Study identification including (citation details and author contact details); (ii) Methods: study design, number of study centres and location, study setting, withdrawals, date of study, follow-up, confounding factors considered, and the methods used to control for confounding, aspects of risk of bias and how missing data were handled; (iii) Participants: number, mean/median age, age range, gender, severity of condition, diagnostic criteria, inclusion criteria, exclusion criteria, screening criteria, socio-demographics; (iv) Interventions: intervention components, comparison; (v) Outcomes: primary and secondary; and (vi) Miscellaneous: funding source, notable conflicts of interest of study authors, ethical approval, key conclusions of the study authors, miscellaneous comments from the study authors, references to other relevant studies, correspondence required.

Risk of bias assessment

Two reviewers (PK and CA) independently assessed the risk of bias for each study using study design-appropriate tools: including the Risk of Bias in Non-randomised Studies of Interventions (ROBINS-I) tool for non-randomised, cohort-type interventional studies and Cochrane Risk of Bias Tool for randomised controlled trials.^20–22 Disagreements were resolved by discussion and consensus or by involving a third reviewer (CN).

Data analysis

Meta-analysis was conducted to pool primary and secondary outcomes reported by studies meeting the review's inclusion criteria. The meta-analysis was conducted for outcomes of interest with reported hazard ratios (HRs) that are both statistically and clinically comparable across studies. Hazard ratios and their corresponding 95% confidence intervals (CIs) were pooled using the inverse variance method. Where a study did not report HRs, they were computed from reported summary statistics reports using the method proposed by Tierney et al (2007). Estimates were pooled using a random-effects meta-analysis due to anticipated between-study heterogeneity.²³ The random-effects model assumes that the effect estimates follow a normal distribution, considering both within-study and between-study variation. Forest plots were generated to illustrate primary outcome findings graphically. Meta-analysis was performed using the Review Manager (RevMan) version 5.4 and Stata version 16.^24,25

Between-study heterogeneity was assessed through visualisation with the aid of forest plots. It was quantified by computing the variance between studies using I² statistic.²⁶ The I² statistic estimates the proportion of variation in effect estimates resulting from true variation rather than random error, with values of 25%, 50% and 75% correlating to low, medium and high heterogeneity, respectively.²⁷ In addition, a Chi-squared test was performed to assess statistically significant heterogeneity. Where the computed p-value was <0.1, heterogeneity was deemed significant.²⁷

Subgroup analyses were conducted to compare treatment effects on the cardiovascular and renal outcomes of interest. Where feasible (if at least 10 citations were included in the meta-analysis), Funnel plots would have been used to assess the presence/absence of likely publication bias in the reviewed evidence.²⁸ Sensitivity analyses were conducted to explore the effect of different variables (such as restricting the analysis to studies with a low risk of bias or excluding studies without propensity score matching) on the effect size of the primary outcomes. Where the relevant outcome data in the main full texts of included studies were unclear or missing, additional data were sourced from supplementary data or files archived with full texts. In cases where data remained missing despite this, the impact of such missing data was explored in the risk of bias assessment and sensitivity analysis.

Assessment of certainty of evidence

The certainty of the evidence for each study (graded as high, moderate, low and very low) was determined using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) criteria.²⁹ Any disagreements in the evidence certainty assessment were resolved through consensus.

Patient and public involvement

This is a review of publicly available literature. Patients and the public were not directly involved in the design, conduct, reporting or dissemination this study.

Results

Search and study selection results

A total of 14,376 unique records were identified from literature database searches. After screening titles and abstracts, 14,306 clearly ineligible publications were excluded. The full texts of the remaining 70 potentially eligible studies were assessed against pre-specified eligibility criteria, from which 22 were included in the review while 48 were excluded for various reasons. 16 studies involving a total of >200,000 participants with T2DM were included in the meta-analysis. Figure 1 presents the PRISMA flow chart of the study selection process and reasons for exclusion.

Figure 1.

PRISMA flow chart illustrating the study search, identification and selection processes.

Characteristics of included studies

Table 1 summarises the characteristics of included studies. Studies were conducted between 2014 and 2022 in 21 countries; Australia, Brazil, Denmark, Hungary, Italy, Japan, Mexico, Norway, Poland, Puerto Rico, Romania, Russia, Slovenia, South Africa, South Korea, Spain, Sweden, Taiwan, Ukraine, the United Kingdom and United States. In terms of study design, 20 studies were longitudinal (cohort-type) observational studies (19 of which were retrospective while one was prospective) and two were RCTs. Most (16/20) cohort studies employed a propensity score matching design. 16 of the included studies reported on cardiovascular but no renal outcomes (eight comparing GLP-1 RA with DPP4i, five comparing GLP-1 RA with basal insulin and three comparing GLP-1 RA with both DPP4i and basal insulin), while four studies reported on renal but no cardiovascular outcomes. Two studies reported on both cardiovascular and renal outcomes. Observational studies were mostly based on patient data from hospital-based medical and health administrative databases.

Table 1.

Characteristics of included studies.

Studies citation	Country	Study design	Study setting, time frame and participants	Intervention	Comparator	Outcomes
Anyanwagu 2016⁴⁶	UK	Retrospective cohort study (propensity score matched)	This study uses the UK primary care database with information systematically entered by primary care physicians. 2003 people with T2DM aged 18 and above, whose MET + SU dual therapy was intensified with either insulin or GLP-1 RA from January 2006 to May 2014	Metformin (MET)+sulphonylurea (SU) + insulin	Metformin+ SU +GLP-1 RA (exenatide, liraglutide or lixisenatide)	Time to the risk of a composite major adverse cardiovascular event (MACE) (including non-fatal myocardial infarction (MI), non-fatal stroke and cardiovascular death)
Chen 2022⁴⁷	Taiwan	Retrospective cohort study (propensity score matched)	The study was built on data from the national health insurance research database (NHIRD) in Taiwan A total of 27,279 participants, from January 1, 2012, to July 2021. Data for people with T2DM and stage 5 CKD or end-stage kidney disease (ESKD) (male)	GLP-1 RA	DPP4i	All-cause mortality and MACE (including MI, cardiogenic shock, heart failure (HF), coronary revascularisation, coronary artery bypass surgery, thrombolysis therapy, malignant arrhythmia, and stroke)
Dawwas 2018⁴⁸	USA	Retrospective cohort study (propensity score matched)	Truven health analytic MarketScan commercial and Medicare supplemental databases 321,606 patients participated from January 2006 to December 2013. Individuals were aged ≥18 years and new users of DPP4i or GLP-1 RA	DPP4i	GLP-1 RA	First hospitalisation due to HF defined as an inpatient hospital claim with a primary or secondary diagnosis of HF
Diamant 2014³⁶	USA and Puerto Rico, Europe, Russia, Australia, South Korea, Taiwan, and Mexico	Open label randomised trial	Patients were recruited at 72 health facilities and followed up between May 13, 2008, and January 30, 2012 456 eligible participants aged 18 years or older, with T2DM and suboptimum glycaemic control at screening despite maximum tolerated doses of metformin alone or with an SU for 3 months or longer	GLP-1 RA (exenatide)	Insulin glargine	Urinary albumin-to-creatinine ratio (ACR) (mg/mmol)
Elharram 2020⁴⁹	USA	Retrospective cohort study (unmatched)	Hospital – Truven health MarketScan research database A cohort of 118,341 adults (aged 18 years or older) with T2DM, followed up from 2011 to 2015 with ongoing metformin use, initiating second-line therapy	DPP4i (sitagliptin, saxagliptin, alogliptin, or linagliptin)	GLP-1 RA (exenatide, liraglutide, albiglutide, and dulaglutide)	Hospitalisation for HF was the primary outcome
Huang 2021³⁹	Taiwan	Retrospective cohort study (propensity score matched)	Hospital – Taiwan's health and Welfare database (the Kanagawa medical association, Japan) A total of 5657 participants with T2DM were enrolled, from October 1, 2012, to December 31, 2016	GLP-1 RA (liraglutide)	Basal insulin	The primary outcome was a composite CVD outcome, consisting of MI, stroke, HF, and cardiovascular mortality
Kobayashi 2021³³	Japan	Retrospective cohort (propensity score matched)	Hospital data from the Kanagawa medical association, Japan 763 Japanese adults with T2DM and CKD who received SGLT2is for >1 year were enrolled from October – December 2018	DPP4i+SGLT2i	GLP-1 RA+SGLT2i	The primary outcome was the change in the eGFR (ΔeGFR) and the logarithmic value of ACR
Lin 2021⁵⁰	Taiwan	Retrospective cohort study (propensity score matched)	Hospital data from the national health insurance database A total of 948,342 participants who were treated for T2DM with a GLP-1 RA or DDP4i were recruited between 1 May 2011 and 31 December 2017 (follow-up: 3 years)	GLP-1 RA	DPP4i	A composite of major adverse limb events, a composite of cardiovascular death, non-fatal MI and non-fatal ischaemic stroke
Longato 2020⁵¹	Italy	Retrospective cohort study (propensity score matched)	The administrative data repository of the Veneto region, a database in northeast Italy 330,193 individuals were enrolled between January 1, 2011, and September 30-, 2018	GLP-1 RA	DPP4i	3-Point major adverse cardiovascular events (3P-MACE, a composite of death, MI, or stroke). 3P-MACE components and hospitalisation for heart failure were secondary outcomes
Longato 2021⁵²	Italy	Retrospective cohort study (propensity score matched)	The administrative data repository of the Veneto region, a database in northeast Italy 330,193 individuals were enrolled between January 1, 2011, and September 30, 2018	GLP-1 RA	Basal insulin	3-Point major adverse cardiovascular events (3P-MACE; a composite of death, MI, or stroke). 3P-MACE components and hospitalisation for HF were secondary outcomes
Luo 2020⁵³	USA	Retrospective cohort study (propensity score matched)	Two large United States (US) commercial claims databases, optum and truven MarketScan 26,308 patients with T2DM enrolled from January 2012 to December 2017	Injectable GLP-1 RAs	Basal insulin	The primary outcome was MI or stroke
O'Brien 2018⁵⁴	USA	Retrospective cohort study (unmatched)	A nationwide database of US national administrative claims data 132,737 participants, adults 18 years or older with T2DM were enrolled from January 2017 to October 2018 (follow-up: 1.3 years)	DPP4i	GLP-1 RA, SGLT2i, thiazolidinediones (TZDs), basal insulin, SUs	The primary outcome was time to first MACE after starting the index antidiabetic medicines
Pasternak 2020¹³	Scandinavia region (Sweden, Denmark, and Norway)	Retrospective cohort study (propensity score matched)	Nationwide register data from Sweden, Denmark, and Norway 38,731 adults aged 35–84 years, who filled their first prescription for either a GLP-1 RA or - DPP4i - during the study period from 2010 to 2016	GLP-1 RA	DPP4i	The primary outcome was serious renal events, a composite including renal replacement therapy, death from renal causes, and hospitalisation
Patorno 2016⁵⁵	USA	Retrospective cohort study (propensity score matched)	A large US commercial insurance database covering more than 14 million persons annually from all 50 states 715,549 people with T2DM aged ≥18 years treated with metformin, who initiated a GLP-1 RA or a comparator patient between April 2005 to December 2013	GLP-1 RA	DPP4i, a second-generation SU and insulin	A composite CVD comprising acute MI, unstable angina, ischaemic
Paul 2015⁵⁶	USA	Retrospective cohort study (propensity score matched)	General electric centricity database from centricity electronic medical record 39,225 adults aged ≥18 years with T2DM who received the first prescription of insulin or EBID between June 2005 and May 2009 (follow-up: 5 years)	GLP 1-RA exenatide twice daily (EBID)	Insulin	Macrovascular outcomes in people with T2DM
Pendergrass 2012³⁴	USA	Retrospective claims analysis (unmatched)	The data source was the medco national integrated database 491,539 adults with diabetes who received at least one prescription for either metformin, an SU or a TZD before starting a new antidiabetic medication (exenatide or sitagliptin) during 1 July 2008 to 31 December 2009	GLP-1 RA (exenatide)	DPP4i (sitagliptin)	The primary outcome was the first occurrence of acute renal failure after the index date
Svanstorm 2019⁵⁷	Denmark and Sweden	Retrospective cohort study (propensity score matched)	Data from the nationwide register 577 adults aged 35 to 84 who initiated liraglutide or DPP4i during the study period from January 2010 to December 2016 (median follow-up: 3.3 years)	GLP-1 RA as a class (>70% exenatide)	DPP4i, SU, and insulin	A composite outcome that included MI, unstable angina, stroke, or coronary revascularisation
Tuttle 2018³⁵	Brazil, Hungary, Mexico, Poland, Romania, South Africa, Spain, Ukraine and the USA	Multicentre, open-label, randomised trial	AWARD-7 was a multicentre, open-label trial done at 99 sites in nine countries 577 participants were adults with T2DM and moderate-to-severe CKD between August 15, 2012, and November 30, 2015	GLP-1 RA Dulaglutide 1.5 mg and 0.75 mg	Insulin glargine	The primary outcome was HbA_1c at 26 weeks, with a 0·4% non-inferiority margin. Secondary outcomes included estimated glomerular filtration rate (eGFR) and urine ACR
Xie 2020³¹	USA	Retrospective cohort study (propensity score matched)	The VA databases 143 hospitals and 1241 outpatient clinics 216,558 individuals with T2DM (US veterans who were mostly older, white, and male). Patients were enrolled from 1 October 2016 to 30 November 2019	GLP-1 RA	DPP4i and SU	The primary study outcome was time until the occurrence of the composite outcome of eGFR decline <50%, ESKD, or all-cause mortality
Xu 2022³⁰	Sweden	Retrospective cohort study (propensity score matched)	Individuals included in the Stockholm Creatinine measurements project, a healthcare utilisation cohort including all adult residents from the region of Stockholm A total of 19,776 individuals with T2DM participants between January 1, 2008, and December 31, 2018	GLP-1 RA	DPP4i	The primary outcome was a composite of sustained doubling of creatinine, kidney failure or kidney death
Yang 2020⁴³	Taiwan	Retrospective cohort study (propensity score matched)	Taiwan’s national health insurance research database (NHIRD) 3195 participants with T2DM, between 2011 and 2014 (follow-up: 2 years)	GLP-1 RA	DPP4i, SU and insulin	A composite of cardiovascular disease (CVD) events and assessed up to the end of 2015
Zerovnik 2021⁵⁸	Slovenia	Retrospective cohort study (Unmatched)	Slovenian outpatient prescription and healthcare databases 3817 people with T2DM with newly introduced DPP4i (n = 3,817), GLP-1 RA (n = 855) or SGLT2i (n = 2851) add-on therapy from January 2010 – December 2016 (follow-up: 3 years)	GLP-1 RA	DPP4i	The primary outcome was a major adverse CV event (MACE), while the secondary outcomes were cardiovascular death

Overall, the majority of the studies were judged to be of low risk of bias, with possible confounding and missing data being the most common sources of bias. Figure 2 illustrates the risk of bias assessment outcomes of all studies included in the meta-analysis while Figure 3 shows the risk of bias assessment outcomes for individual studies.

Figure 2.

Outcomes of risk of bias assessment of studies included in the meta-analysis.

Figure 3.

Meta-analysis of the cardiovascular outcomes comparing GLP-1 RA vs DPP4i on: (a). Composite cardiovascular outcome; (b). Heart failure (c). Myocardial infarction; (d). Stroke (e). Cardiovascular mortality; (f). All-cause mortality.

Cardiovascular outcomes (GLP-1 RA vs DPP4i)

Nine studies compared the effect of GLP-1 RA and DPP4i on composite cardiovascular outcomes as the primary endpoint. However, the definition of composite cardiovascular outcomes and their component outcomes varied across studies (See Table 1). Overall, compared with DPP4i, treatment with GLP-1 RA was associated with a greater benefit on composite cardiovascular outcomes (HR: 0.77, 95% CI: 0.69–0.87, I² = 73%) (Figure 3(a)). In terms of myocardial infarction, GLP-1 RA yielded a slightly favourable result compared with DPP4i (HR: 0.82, 95% CI: 0.69–0.97, I² = 56%) (Figure 3(c)). Similarly, GLP-1 RA yielded a slightly favourable effect on stroke compared with DPP4i (HR: 0.83, 95% CI: 0.74–0.93, I² = 6%) (Figure 3(d)). There was a reduction in cardiovascular mortality in people treated with GLP-1 RA compared with those treated with DPP4i (HR: 0.76, 95% CI: 0.68–0.85, I² = 68%) (Figure 3(e)). A similar reduction in all-cause mortality was found in people who received GLP-1 RA relative to those who received DPP4i treatment (HR: 0.65, 95% CI: 0.48–0.90, I² = 86%) (Figure 3(f)). However, there was no difference in effect between GLP-1 RA and DPP4i regarding heart failure (HR: 0.97, 95% CI: 0.82–1.15, I² = 76%) (Figure 3(b)).

Cardiovascular outcomes (GLP-1 RA vs basal insulin)

Eight studies compared composite cardiovascular outcomes between GLP-1 RA and basal insulin. They, however, differed in their definition of their composite cardiovascular outcomes and component endpoints (See Table 1). Compared with basal insulin, treatment with GLP-1 RA was associated with a greater benefit on composite cardiovascular outcomes (HR: 0.62, 95% CI: 0.48–0.79, I² = 82%) (Figure 4(a)). It also showed superior effects on heart failure (HR: 0.57, 95% CI: 0.35–0.92, I² = 76%) (Figure 4(b)). GLP-1 RA yielded a better reduction in myocardial infarction compared with basal insulin (HR: 0.70, 95% CI: 0.58–0.85, I² = 0%) (Figure 4(C)). GLP-1 RA yielded a more substantial reduction in stroke relative to basal insulin (HR: 0.50, 95% CI: 0.40–0.63, I² = 0%) (Figure 4(d)). Likewise, GLP-1 RA was substantially more beneficial than basal insulin on all-cause mortality (HR: 0.31, 95% CI: 0.20–0.48, I² = 69%) (Figure 4(e)).

Figure 4.

Meta-analysis of the cardiovascular outcomes comparing GLP-1 RA vs basal insulin on: (a). Composite cardiovascular outcome; (b). Heart failure (c). Myocardial infarction; (d). Stroke; (e). All-cause mortality.

Renal outcomes

Only a few studies reported on renal outcomes, the estimates of which are unsuitable for pooling in a meta-analysis. In a cohort study comparing renal effects between people with diabetes who started GLP-1 RA and those who received DPP4i treatment, GLP-1 RA was found to be superior to DPP4i in terms of composite renal outcomes (a composite of kidney outcome (composite of sustained doubling of creatinine, kidney failure or kidney death) (HR: 0.72, 95% CI: 0.53–0.98).³⁰ In another cohort study, treatment GLP-1 RA was associated with a reduced risk of the composite outcome in eGFR <90 to ≥60, <60 to ≥45, and <45 mL/min/1.73 m² compared with DPP4i.³¹ Similarly, a large retrospective cohort study found that the use of GLP-1 RA was associated with a significantly lower risk of renal replacement therapy (HR 0.73 [0.62–0.87]) and hospitalisation for renal events (HR 0.73 [0.65–0.83]) but not death from renal causes (HR 0.72 [0.48–1.10]).³² From another retrospective cohort study, the incidence of a >6.4% decrease in the eGFR was significantly lower in the GLP-1 RA group than in the DPP4i group (35% vs 52%, respectively).³³ However, the authors found no significant difference between GLP-1 RA and DPP4i in terms of the decrease in ACR (−0.12 ± 0.48 vs −0.13 ± 0.45). A study comparing the incidence of acute renal failure between people with diabetes treated with either GLP-1 RA or DPP4i and controls without diabetes found that neither GLP-1 RA (HR 0.77, CI 0.42-1.41) nor DPP4i (HR 1.17, CI 0.82-1.65) was more beneficial in reducing the risk of acute renal failure.³⁴

Available data on the renal effects of GLP-1 RA in comparison with those of basal insulin come from two RCTs. One of the RCTs found better effects on eGFR with GLP-1 RA than with Insulin glargine (34 mL/min per 1·73 m² vs 31·3 mL/min per 1·73 m²). However, the effect of GLP-1 RA on urinary albumin-to-creatinine ratio (ACR) reduction was not significantly different from that of insulin glargine (−22·5% [95% CI -35·1 to −7·5] vs −13·0% [-27·1 to 3·9 mg/mmol].³⁵ Likewise, the second RCT found no significant difference between GLP-1 RA and basal insulin in terms of their effects on ACR reduction (mean difference: −0·83 (−3·28 to 1·62) mg/mmol).³⁶

Sensitivity analyses

Pooled estimates were robust to sensitivity analyses when meta-analyses were restricted to only propensity score-matched studies (excluding studies that did not use propensity score matching). Estimates were also robust to the various definitions of composite cardiovascular outcome across studies.

GRADE assessment of the certainty of the evidence

The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) criteria²⁹ were used to assess the certainty of the evidence for each relevant cardiovascular outcome. The GRADE assessment outcomes are presented in the summary of findings in Table 2:

Table 2.

GRADE Summary of findings for the cardiovascular effects of GLP-1 RA vs DPP4i and basal insulin.

Discussion

This systematic review with meta-analysis highlights the current evidence on the cardiovascular and renal outcomes of GLP-1 RA compared with DPP4i and basal insulin in people with T2DM from 22 studies conducted across 21 countries. Overall, the results show that treatment with GLP-1 RA yielded better benefits on composite and individual cardiovascular outcomes, compared with DPP4i. Notably, GLP-1 RA had a better reduction in the risk of composite cardiovascular outcomes, including 3-point MACE, than DPP4i. This finding is consistent with those of a previous meta-analysis of real-world studies,¹⁰ while supporting the evidence of the cardiovascular benefits of GLP-1 RA demonstrated by previous meta-analyses of placebo-controlled clinical trials.¹²

Compared with DPP4i, treatment with GLP-1 RA was also associated with a greater reduction in individual cardiovascular endpoints, such as myocardial infarction, stroke, cardiovascular mortality and all-cause mortality. Nevertheless, we did not find a significant difference in effect between GLP-1 RA and DPP4i on heart failure. The finding is in agreement with previous review findings that GLP-1 RA had the least benefits on heart failure compared with other cardiovascular outcomes.¹⁰ These comparisons need to be interpreted cautiously, nonetheless, given the differences in the patient populations, study design, active comparators, follow-up period and definition of clinical endpoints (particularly the varied definition of composite outcomes and MACE) across studies included in the meta-analysis. These explained the moderate to high heterogeneity (I² >50%) seen in some of the Forest plots.

To a greater extent, GLP-1 RA compared favourably against basal insulin, with better effects on both composite and individual cardiovascular outcomes, including myocardial infarction, stroke, cardiovascular mortality and all-cause mortality. These findings lend an important addition to the current evidence base on the cardiovascular safety of basal insulin. The emerging evidence points to the cardiovascular superiority of newer glucose-lowering therapies such as GLP-1 RA over older antidiabetic treatments, for example, sulfonylureas and insulin. Therefore, it may be reasonable to prioritise these new therapies for people with T2DM at high risk of cardiovascular adverse outcomes.³⁷ However, due to the between-study heterogeneity mentioned earlier, these comparisons must be interpreted with caution. Another consideration for cautious interpretation of this finding is the question of the extent to which the greater cardiovascular benefits of GLP-1 RA over basal insulin are confounded by baseline cardiovascular risks, if people with higher cardiovascular risk are more likely to have required insulin. While there is likely to be confounding in this regard, recent evidence indicates that insulin therapy has a poorer short- and long-term safety profile than that found in many other anti-T2DM therapies suggesting insulin therapy should be weighed against the cardioprotective properties of other therapies for T2DM. The causal and dose-dependent mechanisms underlying this have been described, including insulin’s predisposition to dyslipidemia, atherosclerosis, hypertension, heart failure and arrhythmias.¹¹

Evidence from a small number of studies investigating renal outcomes was less convincing. Overall, the available evidence is limited, suggesting that GLP-1 RA is probably better on composite and individual renal outcomes such as eGFR, than either DPP4i or basal insulin. Notably, a large cohort study found a significant reduction in other renal outcomes, such as renal-replacement therapy and hospitalisation for renal events, but not death due to renal causes, compared with DPP4i.³² Available evidence from two head-to-head randomised trials comparing renal outcomes between GLP-1 RA and basal insulin shows no difference in effects between GLP-1 RA and basal insulin on urinary albumin-to-creatinine ratio (ACR) reduction.^35,36

When taken together, the evidence from our review supports the promising evidence of cardiovascular and renal protection with GLP-1 RA treatment from CVOTs. The evidence-trend in this direction has led to a paradigm shift in diabetes care resulting in a focus on using new glucose-lowering drugs in people at high cardiorenal risk, irrespective of glucose control.³⁸ The cost of treatment is often a key consideration for clinical decision-making. As such, patients and providers are often faced with weighing the relatively higher cost of GLP-1 RA against its clinical benefits. Available evidence from health economic evaluation studies however suggests that the higher cost is offset by better clinical outcomes and that the use GLP-1 RA is ultimately cost-saving from both patient and healthcare system perspectives.^39,40

Our evidence, which derives largely from observational, real-world studies, helps to overcome the practical limitations of evidence from RCTs. As such, our study has an important strength worth noting. For instance, the real-world setting of the studies included in the meta-analysis helps to account for the lower adherence levels compared with those typically seen in experimental settings of RCTs. For this reason, our review helps to provide evidence that is more generalisable to the general population and the realities of routine clinical practice.

The current review is not without limitations. As indicated earlier, the interpretation of the findings is limited due to substantial heterogeneity between studies in terms of differences in the patient populations, study design, active comparators, follow-up period and definition of clinical endpoints.

For instance, while some studies used propensity score-matching designs, others reported estimates based on unmatched data. On sensitivity analysis, our pooled estimates were, however, robust to such differences, as estimated effect sizes remained consistent when meta-analyses were restricted to only propensity score-matched studies. Pooled estimates were also robust to the various definitions of composite cardiovascular outcome across studies. Most of the studies included in the review used health administrative data. Prior studies have indicated that the predictive value of identifying cardiovascular endpoints from health administrative data is high; 95% for myocardial infarction and stroke, for instance.^41,42 Interpretation of findings is further complicated by the likely differences in patients requiring GLP-1 RA treatment and those on insulin or DPP4i therapy, particularly in terms of prognostic factors such as baseline body weight. For example, individuals selected for GLP-1 RA treatment are more likely to be overweight or obese at baseline for whom weight loss may be a therapeutic goal, whereas insulin is more likely to be selected for individuals with more severe hyperglycaemia to achieve optimal glyacaemic control possibly at the expense of weight gain. Admittedly, our pooled estimates of effect do not account for the extent to which weight loss or weight gain contributes to observed cardiorenal effects. Lastly, most real-world studies are limited by the lack of information regarding diabetes duration and variables such as HbA_1c levels, other co-existing metabolic conditions and anthropometric features at baseline and during follow-up, which may be significant predictors of cardiovascular risk.^9,37,43 For instance, CVOTs have suggested that HbA_1c reduction was a relevant mediator of the GLP-1 RA beneficial effect on MACE, and this association seemed to be driven by that between HbA_1c lowering and stroke.^44,45 Hence, the unavailability of data regarding baseline levels and changes in cardiovascular risk factors represents a limitation of most real-world studies investigating cardiovascular outcomes that might have hindered the interpretation of results.

While the current review adds to the evidence bases, its findings have implications for future research. It will be valuable for future research using real-world data to explore the differences in effects between different types of GLP-1 RA and DPP4i. Further clinical and epidemiological studies are also needed to clarify the role of baseline variables such as HbA_1c as a predictor of cardiorenal effects of newer antidiabetic medications such as GLP-1 RA and DPP4i on both compositive and specific outcomes.

Conclusions

Available evidence from studies investigating cardiovascular outcomes suggests that treating T2DM people with GLP-1 RA can yield better benefits on composite and individual cardiovascular outcomes such as myocardial infarction, stroke, cardiovascular mortality and all-cause mortality, compared with DPP4i. However, there was no difference in effect between GLP-1 RA and DPP4i on heart failure. GLP-1 RA compares better against basal insulin, with better effects on both composite cardiovascular and specific cardiovascular outcomes such as myocardial infarction, stroke, cardiovascular mortality and all-cause mortality. Evidence from a small number of studies investigating renal outcomes suggests that GLP-1 RA had better effects on composite and individual renal outcomes such as eGFR, than either DPP4i and basal insulin.

Footnotes

Acknowledgements

Last Mile P/S provided research and writing assistance.

Author contributions

Conceptualisation: HR and NM. Protocol development: CN, PK, and NM. Database search: PK and CN. Search results screening, study selection, quality assessment and data extraction: PK, CA and CN. Data analysis and interpretation: CN, and PK. Preparation of the first draft of the manuscript: CN. Provision of critical insights and refinement of the manuscript: CA, PK, NM, HR, ME and MD. All authors have read and approved the final version of the manuscript.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: ME has received honoraria from AstraZeneca, Boehringer Ingelheim, and Novo Nordisk. HR and MD are employees of Novo Nordisk Denmark A/S. NM is employed by Last Mile P/S. Last Mile receives consultancy fees from Novo Nordisk. CN, PO, CA have no competing interests to declare.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this review was provided by Novo Nordisk. Novo Nordisk reviewed the publication for medical and scientific accuracy at the first outline stage and ahead of final author approval for submission and reserves the right to share scientific comments with the Author. Final decision to submit lies exclusively with the Author. Novo Nordisk may provide statistical, editorial and medical writing assistance to the Author as agreed upon by the Parties, through a third party.

ORCID iD

Nicole McCreedy

Chukwudi A Nnaji

Data Availability Statement

Data are available upon reasonable request

Appendix

References

International Diabetes Federation (IDF) . IDF diabetes atlas. 10th ed. Brussels, Belgium: IDF, 2021.

Saeedi

Petersohn

Salpea

, et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: results from the international diabetes federation diabetes atlas, 9^th edition. Diabetes Res Clin Pract 2019; 157: 107843.

Liu

Simon

Shi

, et al. Impact of diabetes mellitus on risk of cardiovascular disease and all-cause mortality: evidence on health outcomes and antidiabetic treatment in United States adults. World J Diabetes 2016; 7(18): 449–461.

Almourani

Chinnakotla

Patel

, et al. Diabetes and cardiovascular disease: an update. Curr Diabetes Rep 2019; 19(12): 113–116.

Pálsson

Patel

. Cardiovascular complications of diabetic kidney disease. Adv Chron Kidney Dis 2014; 21(3): 273–280. DOI: 10.1053/j.ackd.2014.03.003.

White

Cannon

Heller

, et al. Alogliptin after acute coronary syndrome in patients with type 2 diabetes. N Engl J Med 2013; 369: 1327–1335.

Baksh

McAdams‐DeMarco

Segal

, et al. Cardiovascular safety signals with dipeptidyl peptidase‐4 inhibitors: a disproportionality analysis among high‐risk patients. Pharmacoepidemiol Drug Saf 2018; 27(6): 660–667.

DeRemer

Vouri

Guo

, et al. Comparing cardiovascular benefits between GLP-1 receptor agonists and SGLT2 inhibitors as an add-on to metformin among patients with type 2 diabetes: a retrospective cohort study. J Diabet Complicat 2021; 35(9): 107972.

Patorno

Htoo

Glynn

, et al. Sodium-glucose cotransporter-2 inhibitors versus glucagon-like peptide-1 receptor agonists and the risk for cardiovascular outcomes in routine care patients with diabetes across categories of cardiovascular disease. Ann Intern Med 2021; 174(11): 1528–1541. DOI: 10.7326/m21-0893.

10.

Caruso

Cignarelli

Sorice

, et al. Cardiovascular and renal effectiveness of GLP-1 receptor agonists vs. Other glucose-lowering drugs in type 2 diabetes: a systematic review and meta-analysis of real-world studies. Metabolites 2022; 12(2): 183.

11.

Herman

O'Keefe

Bell

, et al. Insulin therapy increases cardiovascular risk in type 2 diabetes. Prog Cardiovasc Dis 2017; 60(3): 422–434.

12.

Sattar

Lee

MMY

Kristensen

, 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 randomised trials. Lancet Diabetes Endocrinol 2021; 9(10): 653–662. DOI: 10.1016/s2213-8587(21)00203-5.

13.

Pasternak

Wintzell

Eliasson

, et al. Use of glucagon-like peptide 1 receptor agonists and risk of serious renal events: scandinavian cohort study. Diabetes Care 2020; 43(6): 1326–1335. DOI: 10.2337/dc19-2088.

14.

Gerstein

Colhoun

Dagenais

, et al. Dulaglutide and renal outcomes in type 2 diabetes: an exploratory analysis of the REWIND randomised, placebo-controlled trial. Lancet 2019; 394(10193): 131. DOI: 10.1016/s0140-6736(19)31150-x.

15.

Shamseer

Moher

Clarke

, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ 2015; 349.

16.

Booth

Clarke

Dooley

, et al. The nuts and bolts of PROSPERO: an international prospective register of systematic reviews. Syst Rev 2012; 1(1): 2–9.

17.

Sharma

Pagidipati

Califf

, et al. Impact of regulatory guidance on evaluating cardiovascular risk of new glucose-lowering therapies to treat type 2 diabetes mellitus: lessons learned and future directions. Circ 2020; 141(10): 843–862.

18.

US Food and Drug Administration (USFDA) . Guidance for industry: diabetes mellitus–evaluating cardiovascular risk in new antidiabetic therapies to treat type 2 diabetes 2008. https://www.fda.gov/media/71297/download

19.

Liberati

Altman

Tetzlaff

, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 2009; 6(10): e1000100.

20.

Jørgensen

Paludan-Müller

Laursen

, et al. Evaluation of the Cochrane tool for assessing risk of bias in randomized clinical trials: overview of published comments and analysis of user practice in Cochrane and non-Cochrane reviews. Syst Rev 2016; 5(1): 1–13.

21.

Jüni

Loke

Pigott

, et al. Risk of bias in non-randomized studies of interventions (ROBINS-I): detailed guidance. BMJ 2016; 355.

22.

Peterson

Welch

Losos

, et al. The Newcastle-Ottawa scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute 2011; 2(1): 1–12.

23.

Tierney

Stewart

Ghersi

, et al. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007; 8: 16.

24.

Wang

Leeflang

. Recommended software/packages for meta-analysis of diagnostic accuracy. J Lab Precis Med 2019; 4: 22.

25.

StataCorp L. Stata Statistical Software: Release 2019; 11: 2001–2015.

26.

Borenstein

Higgins

Hedges

, et al. Basics of meta‐analysis: I2 is not an absolute measure of heterogeneity. Res Synth Methods 2017; 8(1): 5–18.

27.

Ruppar

. Meta-analysis: how to quantify and explain heterogeneity? Eur J Cardiovasc Nurs 2020; 19(7): 646–652.

28.

Sterne

Sutton

Ioannidis

, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 2011; 343: d4002.

29.

Guyatt

Oxman

AD,

Vist

et al. GRADE Working Group . Rating quality of evidence and strength of recommendations GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008; 336: 924–926.

30.

Clase

, et al. GLP-1 receptor agonist versus DPP-4 inhibitor and kidney and cardiovascular outcomes in clinical practice in type-2 diabetes. Kidney Int 2022; 101(2): 360–368.

31.

Xie

Bowe

Gibson

, et al. Comparative effectiveness of SGLT2 inhibitors, GLP-1 receptor agonists, DPP-4 inhibitors, and sulfonylureas on risk of kidney outcomes: emulation of a target trial using health care databases. Diabetes Care 2020; 43(11): 2859–2869.

32.

Pasternak

Wintzell

Eliasson

, et al. Use of glucagon-like peptide 1 receptor agonists and risk of serious renal events: scandinavian cohort study. Diabetes Care 2020; 43(6): 1326–1335.

33.

Kobayashi

Toyoda

Hatori

, et al. The comparison of the kidney effects of dipeptidyl peptidase 4 inhibitors and glucagon-like peptide 1 agonist-administered concomitant with sodium-glucose cotransporter 2 inhibitors in Japanese patients with type diabetes mellitus and chronic kidney disease. J Diabetes Res 2021; 2021.

34.

Pendergrass

Fenton

Haffner

, et al. Exenatide and sitagliptin are not associated with increased risk of acute renal failure: a retrospective claims analysis. Diabetes Obes Metabol 2012; 14(7): 596–600.

35.

Tuttle

Lakshmanan

Rayner

, et al. Dulaglutide versus insulin glargine in patients with type 2 diabetes and moderate-to-severe chronic kidney disease (AWARD-7): a multicentre, open-label, randomised trial. Lancet Diabetes Endocrinol 2018; 6(8): 605–617.

36.

Diamant

Van Gaal

Guerci

, et al. Exenatide once weekly versus insulin glargine for type 2 diabetes (DURATION-3): 3-year results of an open-label randomised trial. Lancet Diabetes Endocrinol 2014; 2(6): 464–473.

37.

O’Brien

Karam

Wallia

, et al. Association of second-line antidiabetic medications with cardiovascular events among insured adults with type 2 diabetes. JAMA Netw Open 2018; 1(8): e186125.

38.

Cefalu

Kaul

Gerstein

, et al. Cardiovascular outcomes trials in type 2 diabetes: where do we go from here? Reflections from a diabetes care editors’ expert forum. Diabetes Care 2018; 41(1): 14–31.

39.

Huang

Chen

, et al. Cardiovascular outcomes and healthcare costs of liraglutide versus basal insulin for type 2 diabetes patients at high cardiovascular risk. Sci Rep 2021; 11(1): 1430. DOI: 10.1038/s41598-020-80753-9.

40.

Yang

Chen

, et al. Cost-effectiveness of GLP-1 receptor agonists versus insulin for the treatment of type 2 diabetes: a real-world study and systematic review. Cardiovasc Diabetol 2021; 20(1): 21. DOI: 10.1186/s12933-020-01211-4.

41.

Choma

Griffin

Huang

, et al. An algorithm to identify incident myocardial infarction using Medicaid data. Pharmacoepidemiol Drug Saf 2009; 18(11): 1064–1071.

42.

Roumie

Mitchel

Gideon

, et al. Validation of ICD‐9 codes with a high positive predictive value for incident strokes resulting in hospitalization using Medicaid health data. Pharmacoepidemiol Drug Saf 2008; 17(1): 20–26.

43.

Yang

C-T

Yang

C-Y

H-T

, et al. Comparative cardiovascular safety of GLP-1 receptor agonists versus other glucose-lowering agents in real-world patients with type 2 diabetes: a nationwide population-based cohort study. Cardiovasc Diabetol 2020; 19(1): 83.

44.

Wang

J-S

Liu

W-J

Lee

C-L

. HbA1c trajectory and cardiovascular outcomes: an analysis of data from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study. Ther Adv Chronic Dis 2021; 12: 20406223211026391.

45.

NAVIGATOR Study Group, Holman

Haffner

McMurray

, et al. Effect of nateglinide on the incidence of diabetes and cardiovascular events. N Engl J Med 2010; 362(16): 1463–1476. DOI: 10.1056/NEJMoa1001122.

46.

Anyanwagu

Mamza

Mehta

, et al. Cardiovascular events and all-cause mortality with insulin versus glucagon-like peptide-1 analogue in type 2 diabetes. Heart (British Cardiac Society) 2016; 102(19): 1581–1587. DOI: 10.1136/heartjnl-2015-309164.

47.

Chen

Jenq

, et al. Association of glucagon-like peptide-1 receptor agonist vs dipeptidyl peptidase-4 inhibitor use with mortality among patients with type 2 diabetes and advanced chronic kidney disease. JAMA Netw Open 2022; 5(3): e221169. DOI: 10.1001/jamanetworkopen.2022.1169.

48.

Dawwas

Smith

Park

. Risk of heart failure hospitalization among users of dipeptidyl peptidase-4 inhibitors compared to glucagon-like peptide-1 receptor agonists. Cardiovasc Diabetol 2018; 17(1): 102. DOI: 10.1186/s12933-018-0746-4.

49.

Elharram

Moura

Abrahamowicz

, et al. Novel glucose lowering agents are associated with a lower risk of cardiovascular and adverse events in type 2 diabetes: a population based analysis. Int J Cardiol 2020; 310: 147–154. DOI: 10.1016/j.ijcard.2020.03.025.

50.

Lin

Lee

Chen

. Major adverse cardiovascular and limb events in patients with diabetes treated with GLP-1 receptor agonists vs DPP-4 inhibitors. Diabetologia 2021; 64(9): 1949–1962. DOI: 10.1007/s00125-021-05497-1.

51.

Longato

Di Camillo

Sparacino

, et al. Better cardiovascular outcomes of type 2 diabetic patients treated with GLP-1 receptor agonists versus DPP-4 inhibitors in clinical practice. Cardiovasc Diabetol 2020; 19(1): 74. DOI: 10.1186/s12933-020-01049-w.

52.

Longato

Di Camillo

Sparacino

, et al. Cardiovascular outcomes after initiating GLP-1 receptor agonist or basal insulin for the routine treatment of type 2 diabetes: a region-wide retrospective study. Cardiovasc Diabetol 2021; 20(1): 222. DOI: 10.1186/s12933-021-01414-3.

53.

Luo

Zhou

Liu

, et al. 925-P: risk of cardiovascular events and serious hypoglycemia comparing patients with type 2 diabetes initiating a GLP-1 receptor agonist or basal insulin: a propensity-score matched cohort study. Diabetes 2020; 20(Supplement_1): 69. DOI: 10.2337/db20-925-P.

54.

O'Brien

Karam

Wallia

, et al. Association of second-line antidiabetic medications with cardiovascular events among insured adults with type 2 diabetes. JAMA Netw Open 2018; 1(8): e186125. DOI: 10.1001/jamanetworkopen.2018.6125.

55.

Patorno

Everett

Goldfine

, et al. Comparative cardiovascular safety of glucagon-like peptide-1 receptor agonists versus other antidiabetic drugs in routine care: a cohort study. Diabetes Obes Metabol 2016; 18(8): 755–765. DOI: 10.1111/dom.12665.

56.

Paul

Klein

Maggs

, et al. The association of the treatment with glucagon-like peptide-1 receptor agonist exenatide or insulin with cardiovascular outcomes in patients with type 2 diabetes: a retrospective observational study. Cardiovasc Diabetol 2015; 14: 10. DOI: 10.1186/s12933-015-0178-3.

57.

Svanström

Ueda

Melbye

, et al. Use of liraglutide and risk of major cardiovascular events: a register-based cohort study in Denmark and Sweden. Lancet Diabetes Endocrinol 2019; 7(2): 106–114. DOI: 10.1016/s2213-8587(18)30320-6.

58.

Zerovnik

Kos

Locatelli

. Cardiovascular morbidity and mortality in patients with type 2 diabetes using novel antidiabetic medicines as add-on therapy: an observational real-world study. BMJ Open 2021; 11(9): e051549. DOI: 10.1136/bmjopen-2021-051549.