What this is
- This study evaluates the effects of adding glucagon-like peptide-1 receptor agonists (GLP-1 RAs) to insulin therapy in patients with type 2 diabetes (T2D).
- Using a nationwide database from Taiwan, researchers compared outcomes of patients receiving GLP-1 RAs versus dipeptidyl peptidase-4 (DPP-4) inhibitors and sulfonylureas.
- The study focuses on long-term cardiovascular and microvascular outcomes, addressing a gap in existing evidence.
Essence
- Adding GLP-1 RAs to insulin therapy in T2D patients is associated with significantly lower risks of cardiovascular events, major microvascular complications, and all-cause mortality compared to DPP-4 inhibitors and sulfonylureas.
Key takeaways
- GLP-1 RA use significantly reduces the risk of major adverse cardiovascular events (MACE) by 48% compared to DPP-4 inhibitors, indicating a strong protective effect.
- Patients on GLP-1 RAs experience a 58% lower risk of major microvascular complications, including end-stage kidney disease and leg amputation, compared to those on DPP-4 inhibitors.
- The study shows that GLP-1 RA therapy is linked to a 62% reduction in all-cause mortality compared to DPP-4 inhibitors, underscoring its potential for improving patient survival.
Caveats
- The study's observational design limits causal inference, and residual confounding may affect the results. Further randomized controlled trials are needed for validation.
- The findings may not be generalizable beyond the Taiwanese population due to the study's specific demographic and healthcare context.
Simplified
1. Introduction
At the time of diagnosis, individuals with type 2 diabetes (T2D) have typically lost approximately 50% of their β-cell secretory capacity, with subsequent annual declines ranging from about 2% in older adults to as high as 40% in younger patients or those with rapidly progressive disease [1,2,3,4,5]. Consequently, insulin therapy eventually becomes inevitable. Previous randomized trials have demonstrated that intensive insulin therapy significantly reduces the risk of both microvascular and macrovascular complications, underscoring the importance of early and rigorous glycemic control in improving long-term outcomes among individuals with T2D [6,7,8]. Nevertheless, nearly 70% of patients with T2D fail to achieve adequate glycemic control even while receiving insulin therapy [9]. Systematic reviews have shown that combination insulin therapy provides better glycemic control than insulin monotherapy [10,11]. However, even with basal insulin combined with oral hypoglycemic agents, approximately 40–50% of patients still fail to achieve adequate glycemic control after 24 weeks of treatment [12].
Selecting appropriate non-insulin hypoglycemic agents to complement insulin therapy is therefore crucial for optimizing both glycemic control and long-term outcomes [13]. Major classes of non-insulin agents include sulfonylureas, metformin, dipeptidyl peptidase-4 (DPP-4) inhibitors, sodium–glucose cotransporter-2 (SGLT-2) inhibitors, and glucagon-like peptide-1 receptor agonists (GLP-1 RAs), each with distinct mechanisms of action and clinical profiles [13,14]. While sulfonylureas may increase the risk of hypoglycemia and mortality when combined with insulin, metformin enhances insulin sensitivity, and DPP-4 inhibitors have shown protective effects against stroke [15,16]. GLP-1 RAs offer multiple advantages, including improved glycemic control, weight reduction, and a lower risk of hypoglycemia, though their use may be limited by gastrointestinal side effects, cost, and the requirement for injections [10,17,18]. Current international guidelines increasingly recommend GLP-1 RAs and SGLT-2 inhibitors for patients at elevated cardiovascular risk [19].
Despite these recognized benefits, evidence regarding the long-term cardiovascular and microvascular effects of adding GLP-1 RAs to insulin therapy remains limited [10,11]. To address this knowledge gap, we conducted a retrospective cohort study using Taiwan's National Health Insurance Research Database. The primary objective was to evaluate the impact of adding GLP-1 RAs to insulin therapy on major cardiovascular and microvascular outcomes. A secondary objective was to compare the individual macrovascular and microvascular outcomes of adding GLP-1 RAs with those achieved by adding either DPP-4 inhibitors or sulfonylureas.
2. Results
2.1. Study Subjects
From the NHIRD, a total of 893,990 patients with T2D receiving insulin therapy were initially identified. After applying exclusion criteria and propensity score matching, 6779 matched pairs of GLP-1 RA users versus DPP-4 inhibitor users and 5242 matched pairs of GLP-1 RA users versus sulfonylurea users were obtained (Figures S1 and S2). Baseline demographics, comorbidities, concomitant medications, and diabetes duration were well balanced across groups, with all standardized mean differences (SMDs) < 0.1 (Table 1 and Table S1). In the matched cohorts (GLP-1 RA vs. DPP-4 inhibitors and GLP-1 RA vs. sulfonylureas), the proportion of men was 53.32% and 51.50%, the mean (SD) age was 51.24 (12.67) and 52.66 (13.33) years, the mean (SD) diabetes duration was 7.17 (3.71) and 7.14 (3.76) years, and the mean (SD) follow-up was 3.45 (1.99) and 3.43 (2.00) years, respectively.
2.2. Key Findings
2.2.1. GLP-1 RA Versus DPP-4 Inhibitor Use
In the propensity score-matched cohort comparing GLP-1 RAs with DPP-4 inhibitors, GLP-1 RA treatment was significantly associated with a lower risk of MACE [aHR 0.52, 95% CI 0.46–0.58; RR 0.54, 95% CI 0.49–0.61, corresponding to a 46% risk reduction], as well as hospitalizations for coronary artery disease (aHR 0.64, 95% CI 0.54–0.75), stroke (aHR 0.48, 95% CI 0.40–0.56), and heart failure (aHR 0.33, 95% CI 0.25–0.42). GLP-1 RA use was also associated with reduced risks of major microvascular complications [aHR 0.42, 95% CI 0.35–0.50; RR 0.39, 95% CI 0.33–0.47, a 61% reduction], end-stage kidney disease (aHR 0.08, 95% CI 0.04–0.14), sight-threatening retinopathy (aHR 0.62, 95% CI 0.50–0.76), leg amputation (aHR 0.16, 95% CI 0.05–0.57), and all-cause mortality (aHR 0.38, 95% CI 0.32–0.44) (Table 2). Kaplan–Meier analyses further showed significantly lower cumulative incidences of MACE, major microvascular complications, all-cause mortality, and hospitalizations for coronary artery disease, stroke, heart failure, end-stage kidney disease, and sight-threatening retinopathy in GLP-1 RA users compared with DPP-4 inhibitor users, with all log-rank tests exhibiting p < 0.001 (Figure 1, Figures S3 and S4). Sensitivity analysis excluding SGLT2 inhibitor users confirmed that the observed benefits of GLP-1 RAs were independent of concomitant SGLT2 inhibition (Table S2).
2.2.2. GLP-1 RA Versus Sulfonylurea Use
In the matched cohort comparing GLP-1 RAs with sulfonylureas, GLP-1 RA treatment was significantly associated with reduced risks of MACE (aHR 0.66, 95% CI 0.58–0.75), hospitalization for coronary artery disease (aHR 0.74, 95% CI 0.61–0.88), stroke (aHR 0.64, 95% CI 0.54–0.76), heart failure (aHR 0.54, 95% CI 0.42–0.70), major microvascular complications (aHR 0.68, 95% CI 0.56–0.82), end-stage kidney disease (aHR 0.39, 95% CI 0.27–0.57), leg amputation (aHR 0.29, 95% CI 0.09–0.91), and all-cause mortality (aHR 0.46, 95% CI 0.39–0.54). By contrast, the reduction in sight-threatening retinopathy risk did not reach statistical significance (aHR 0.85, 95% CI 0.67–1.08) (Table 3). Kaplan–Meier analyses showed significantly lower cumulative incidences of MACE (log-rank p < 0.001), major microvascular complications (log-rank p < 0.001), all-cause mortality (log-rank p < 0.001), hospitalization for coronary artery disease (log-rank p = 0.003), stroke (log-rank p < 0.001), heart failure (log-rank p < 0.001), and ESKD (log-rank p < 0.001) among GLP-1 RA users compared with sulfonylurea users (Figure 2, Figures S5 and S6).
2.3. Subgroup Analyses
2.3.1. GLP-1 RA Versus DPP-4 Inhibitor Use
Across multiple subgroups, GLP-1 RA use was consistently associated with reduced risks of hospitalization for coronary artery disease, stroke, heart failure, end-stage kidney disease, sight-threatening retinopathy, and all-cause mortality compared with DPP-4 inhibitor use (). Tables S3–S8
2.3.2. GLP-1 RA Versus Sulfonylurea Use
In subgroup analyses, GLP-1 RA therapy was generally associated with lower risks of hospitalization for coronary artery disease, stroke, heart failure, ESKD, and all-cause mortality compared with sulfonylurea use (). Tables S9–S13
3. Discussion
This nationwide cohort study in Taiwan demonstrated that, among patients with type 2 diabetes receiving insulin therapy, the addition of GLP-1 RAs was associated with reduced risks of MACE, hospitalizations for stroke, coronary artery disease, and heart failure, as well as major microvascular complications, end-stage kidney disease, sight-threatening retinopathy, leg amputation, and all-cause mortality, compared with the addition of DPP-4 inhibitors or sulfonylureas.
Long-term follow-up from the United Kingdom Prospective Diabetes Study (UKPDS) and the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) suggested that intensive glycemic control with insulin may reduce the risk of myocardial infarction [6,7]. In contrast, observational studies have indicated that insulin therapy may be associated with an increased risk of cardiovascular events and mortality [13,14,18]. A previous meta-analysis of randomized controlled trials of GLP-1 RAs demonstrated significant reductions in the risk of MACE and myocardial infarction in patients with T2D [20]. On this basis, the American Diabetes Association recommends combining GLP-1 RAs with insulin to improve treatment efficacy, sustain long-term benefits, and reduce cardiovascular risk [19]. Consistent with these recommendations, our study showed that in insulin-treated patients with T2D, GLP-1 RA use significantly reduced the risk of hospitalization for coronary artery disease compared with DPP-4 inhibitor and sulfonylurea use.
T2D is also associated with an increased risk, severity, and recurrence of stroke [21]. Meta-analyses of GLP-1 RAs have shown significantly lower risks of both total and non-fatal stroke compared with placebo [20,21]. Our previous study further indicated that DPP-4 inhibitor use in insulin-treated patients with T2D was linked to a lower stroke risk compared with non-DPP-4 inhibitor use [15]. In the present study, GLP-1 RA therapy was associated with a significantly reduced risk of hospitalization for stroke compared to both DPP-4 inhibitors and sulfonylureas in patients with T2D on insulin.
Compared with individuals without T2D, patients with T2D have a two- to four-fold higher prevalence of heart failure (HF) [22]. Among patients with both T2D and HF, approximately 30% are treated with insulin [23]. However, the weight gain and sodium retention induced by insulin may accelerate HF progression [14,17]. Therefore, selecting appropriate non-insulin adjunctive agents is critical in insulin-treated patients to mitigate worsening HF [10]. A meta-analysis reported that GLP-1 RA therapy significantly reduced the risk of HF hospitalization compared with placebo [20]. Similarly, our study demonstrated that GLP-1 RA use in insulin-treated patients was associated with a significantly lower risk of HF hospitalization than use of DPP-4 inhibitors or sulfonylureas. Collectively, the evidence suggests that GLP-1 RAs may help attenuate HF progression in insulin-treated patients with T2D.
Randomized controlled trials have shown that intensive insulin therapy reduces microvascular complications, particularly diabetic kidney disease [6,7,8]. However, insulin-treated patients generally have more severe and long-standing T2D, placing them at higher risk of kidney disease progression [22]. A meta-analysis has shown that GLP-1 RAs significantly reduce composite kidney outcomes, primarily by slowing progression to macroalbuminuria [19]. More recently, the FLOW trial demonstrated that subcutaneous semaglutide significantly reduced the risk of worsening renal function in patients with T2D and chronic kidney disease [24]. In line with these findings, our study showed that GLP-1 RA use in insulin-treated patients was associated with a lower risk of end-stage kidney disease compared to DPP-4 inhibitors or sulfonylureas.
Earlier studies suggested that insulin therapy may increase the risk of diabetic retinopathy, possibly due to rapid glucose lowering [25]. One randomized trial even reported a potential association between GLP-1 RA use and severe retinopathy [26]. By contrast, our previous cohort study found no significant increase in the risk of sight-threatening retinopathy with GLP-1 RA use compared with non-use, and in fact demonstrated a significantly lower risk compared with DPP-4 inhibitor use [27]. The present study further confirmed that, among insulin-treated patients with T2D, GLP-1 RA therapy was associated with a significantly lower risk of sight-threatening retinopathy compared with DPP-4 inhibitors.
Patients on insulin therapy often have poorer glycemic control and a higher burden of complications [14,17]. Amputation is one of the most disabling complications, leading to substantial disability and impaired quality of life [28]. Observational data suggest that GLP-1 RA use is linked to a significantly lower risk of amputation compared with DPP-4 inhibitor use [29]. Consistently, our study showed that among insulin-treated patients with T2D, GLP-1 RA therapy was associated with a markedly lower risk of amputation than either DPP-4 inhibitors or sulfonylureas.
Our previous research demonstrated that the combination of DPP-4 inhibitors with insulin was associated with reduced mortality risk in insulin-treated patients with T2D [30]. In addition, meta-analyses of cardiovascular outcome trials have shown that GLP-1 RA therapy significantly reduces cardiovascular death and all-cause mortality compared with placebo [19]. Furthermore, meta-analyses have indicated that combining GLP-1 RAs with insulin improves glycemic control, body weight, and blood pressure more effectively than insulin alone [10,11]. Together, these benefits may contribute to lowering both macrovascular and microvascular complications and reducing mortality. Our present study demonstrated that in insulin-treated patients with T2D, GLP-1 RA therapy was associated with a significantly lower risk of all-cause mortality compared with DPP-4 inhibitors and sulfonylureas.
The mechanisms underlying the protective effects of GLP-1 RAs against macrovascular and microvascular complications in T2D are likely multifactorial: (i) promotion of weight loss, reduction in systolic blood pressure, and improvement in lipid levels—key cardiovascular risk factors [17,20]; (ii) direct cardioprotective effects, including reduced ischemia-reperfusion injury, improved contractility, decreased fibrosis, and enhanced endothelial function, along with neuroprotection that may reduce risks of HF, myocardial infarction, and stroke [21,31]; (iii) lowering of blood glucose, oxidative stress, and inflammation, thereby slowing the progression of retinopathy, neuropathy, and atherosclerosis—drivers of both microvascular and macrovascular disease [11,31]; and (iv) renoprotection through reduced glomerular hyperfiltration and enhanced natriuresis, which ease renal stress and preserve kidney function [31].
3.1. Perspectives for Clinical Practice
Our findings indicate that, among patients with T2D receiving insulin therapy, the addition of GLP-1 RAs may confer substantial long-term benefits compared with DPP-4 inhibitors or sulfonylureas. GLP-1 RA use was associated with significant reductions in major cardiovascular events, critical microvascular complications (including kidney failure, sight-threatening retinopathy, and amputation), and all-cause mortality. These results underscore the role of GLP-1 RAs not only as effective glucose-lowering agents but also as therapies that improve prognosis and reduce overall disease burden.
In practice, GLP-1 RAs may be particularly advantageous for insulin-treated patients with elevated cardiovascular or microvascular risk. When clinically appropriate, they can be combined with metformin or SGLT2 inhibitors to further enhance cardiometabolic protection. Careful monitoring for adverse events is essential, and practical considerations such as treatment cost and accessibility should be taken into account to support equitable and effective implementation.
3.2. Limitations
This study has several limitations. First, the NHIRD lacks information on family history, physical activity, and dietary habits. Clinical data such as blood cholesterol, glucose levels, hemoglobin A1C, renal function, and inflammatory biomarkers are also unavailable. To mitigate these limitations, we used propensity score matching to balance important covariates between groups, including the DCSI, CCI, use of oral antidiabetic drugs, and diabetes duration, which may serve as proxies for disease severity. Second, medication adherence, including prescribed insulin doses, could not be reliably assessed from claims data. We also did not analyze hypoglycemia events, as outpatient hypoglycemia is often underreported in the database. Third, the study was conducted primarily in a Taiwanese population, which may limit generalizability to other ethnic groups. Finally, as with any cohort study, residual confounding and unmeasured biases cannot be excluded, and causal inference remains limited. Further confirmation from randomized controlled trials is warranted.
4. Materials and Methods
4.1. Study Population and Data Source
This retrospective cohort study used data from Taiwan's National Health Insurance Research Database (NHIRD), which covers nearly 99% of the national population. The database contains comprehensive healthcare information, including demographics, clinical diagnoses, medical procedures, and prescription records [32]. Diagnoses were coded using the International Classification of Diseases, Ninth and Tenth Revisions, Clinical Modification (ICD-9-CM and ICD-10-CM). Mortality data were validated through linkage with the National Death Registry to improve accuracy. Ethical approval was obtained from the Research Ethics Committee of China Medical University Hospital (CMUH110-REC1-038 [CR-3]). As all data were de-identified before analysis, the need for informed consent was waived.
This study was designed and reported in accordance with the STROBE guidelines and checklist [33].
4.2. Study Design and Procedures
Patients with T2D were identified in the NHIRD between 1 January 2008, and 31 December 2021 (Figure S1). T2D was defined as having at least two outpatient visits or one hospitalization with a T2D diagnosis within a single year, based on outpatient and/or inpatient records (Table S14) [34]. The validity of using ICD codes for identifying T2D in Taiwan has been confirmed, with an accuracy of 93.3% [34]. From this cohort, patients with newly diagnosed T2D who initiated insulin therapy were included.
Exclusion criteria were as follows: (1) age < 20 or >100 years; (2) missing data on age or sex; (3) history of type 1 diabetes, dialysis, sight-threatening retinopathy (defined as ≥2 outpatient visits or ≥1 hospitalization with surgical intervention, laser photocoagulation within 90 days of diagnosis, vision loss, or anti-VEGF injection), or leg amputation before the index date; (4) use of GLP-1 RAs or DPP-4 inhibitors within 3 months before insulin initiation; (5) concurrent use of GLP-1 RAs and DPP-4 inhibitors; and (6) not receiving either of these agents.
Patients who initiated GLP-1 RAs after their T2D diagnosis were classified as GLP-1 RA users, while those who began DPP-4 inhibitors or sulfonylureas were categorized accordingly. The index date for each patient was defined as the date of initiating the respective medication. Since GLP-1 RAs and DPP-4 inhibitors became available in Taiwan only after 2011, the index date for all groups was set on or after 1 January 2011. Patients were followed from the index date until 31 December 2021.
4.3. Demographics and Related Variables of the Participants
We adjusted for a range of clinically relevant variables that could influence the likelihood of receiving GLP-1 RA therapy and thereby affect outcomes (Table 1 and Table S1). These included demographic factors (age, sex), as well as comorbidities present before the index date, such as overweight/obesity, smoking, alcohol-related disorders, hypertension, dyslipidemia, coronary artery disease, stroke, heart failure, atrial fibrillation, peripheral arterial disease, chronic obstructive pulmonary disease, liver cirrhosis, chronic kidney disease, retinopathy, and cancer.
Medication use was also considered, including the number and classes of oral antidiabetic agents, antihypertensives, aspirin, and statins. In addition, we included the Charlson Comorbidity Index (CCI) [35], the Diabetes Complications Severity Index (DCSI) [36], and diabetes duration.
4.4. Main Outcomes
The primary outcomes were: (1) major adverse cardiovascular events (MACEs), defined as a composite of hospitalizations for coronary artery disease (CAD), stroke, and heart failure, and (2) major microvascular complications, defined as a composite of end-stage kidney disease (ESKD), sight-threatening retinopathy, and leg amputation [34]. Secondary outcomes included each of these components individually—hospitalizations for CAD, stroke, heart failure, ESKD, sight-threatening retinopathy, non-traumatic lower limb amputation—as well as all-cause mortality. Mortality and causes of death were confirmed using death certificates and data from the National Death Registry. Patients were followed until the occurrence of any outcome, death, or 31 December 2021, whichever came first.
4.5. Statistical Analysis
To ensure comparability between GLP-1 RA users and non-users, we performed 1:1 propensity score matching. Propensity scores were estimated using a non-parsimonious multivariable logistic regression model, with GLP-1 RA initiation as the dependent variable. Covariates included demographic characteristics (age, sex), smoking status, obesity, comorbidities, CCI and DCSI scores (≤1 vs. >1), index year, medication use, and duration of T2D (Table 1 and Table S1). Differences between groups were evaluated using Student's t-tests for continuous variables and chi-squared tests for categorical variables. Matching was conducted using the nearest-neighbor method, with balance considered acceptable when the standardized mean difference (SMD) was <0.1 or the p-value > 0.05.
Outcomes were compared between GLP-1 RA users and DPP-4 inhibitor/sulfonylurea users using Cox proportional hazards models with robust sandwich variance estimators. The proportional hazards assumption was assessed using Schoenfeld residuals. Results are expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). Cumulative incidence of outcomes was estimated using Kaplan–Meier curves and compared with log-rank tests.
Subgroup analyses were performed to assess effect modification by sex, age, hypertension, dyslipidemia, CCI, DCSI, number of oral antidiabetic agents, statin use, and diabetes duration. A sensitivity analysis was conducted excluding SGLT2 inhibitor users from the GLP-1 RA and DPP-4 inhibitor cohorts to confirm that the observed benefits of GLP-1 RAs were not attributable to concomitant SGLT2 inhibition.
A two-sided p-value < 0.05 was considered statistically significant. All analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC, USA).
5. Conclusions
In this nationwide cohort of patients with type 2 diabetes receiving insulin therapy, the addition of GLP-1 RAs was associated with significantly lower risks of cardiovascular events, major microvascular complications, and all-cause mortality compared with the addition of DPP-4 inhibitors or sulfonylureas. These findings suggest that incorporating GLP-1 RAs into insulin regimens may help optimize treatment, reduce disease burden, and improve survival.
Nevertheless, as this was an observational study, the possibility of residual confounding must be acknowledged, and the generalizability of our findings beyond the Taiwanese healthcare setting may be limited. Further randomized controlled trials and investigations in more diverse populations are warranted to confirm these benefits and inform clinical practice.