Efficacy and safety of combination therapy using SGLT2 and DPP4 inhibitors to treat type 2 diabetes: An updated systematic review and meta‐analysis with focus on an Asian subpopulation

Abstract

Aims

This updated meta‐analysis investigates the efficacy and safety of combining sodium‐glucose cotransporter 2 inhibitors (SGLT2is) with dipeptidyl peptidase‐4 inhibitors (DPP4is) in treating type 2 diabetes (T2D), especially in Asian subpopulations.

Materials and Methods

A systematic review was conducted on randomized controlled trials (RCTs) published through January 2024 that compared SGLT2i/DPP4i combination therapy with DPP4i or SGLT2i monotherapy. The primary outcome was haemoglobin A1c (HbA1c) changes. Subgroup analyses were conducted based on the baseline HbA1c level (using 8.0%–8.5% as the cut‐off) and racial groups (Asian vs. non‐Asian).

Results

This analysis included 17 RCTs with 7588 participants. Compared to DPP4i, the SGLT2i/DPP4i combination significantly reduced HbA1c (mean difference [MD] −0.57%, 95% confidence interval [CI] −0.67 to −0.46%), while promoting modest weight loss (MD −1.57 kg, 95% CI −1.93 to −1.20 kg). When compared to SGLT2i, SGLT2i/DPP4i promoted further reductions in HbA1c (MD −0.46%, 95% CI −0.55 to −0.38%) with no significant effects on body weight. Subgroup analyses revealed that the efficacy of adding DPP4i in reducing HbA1c was more pronounced in Asian participants (MD −0.55%, 95% CI −0.71 to −0.39) than in non‐Asian participants (MD −0.38%, 95% CI −0.46 to −0.31). The combination therapy was associated with a similar risk of hypoglycaemia compared to both monotherapy groups, with no statistically significant differences observed.

Conclusions

Combination therapy of SGLT2i and DPP4i improves glycaemic control in T2D, with enhanced DPP4i efficacy noted in Asian populations, highlighting its role in personalized diabetes management for these patients.

1. INTRODUCTION

Over the past decade, dipeptidyl peptidase‐4 inhibitors (DPP4is) and sodium‐glucose cotransporter 2 inhibitors (SGLT2is) have emerged as pivotal options for glycaemic control in type 2 diabetes (T2D), offering complementary mechanisms of action. ¹ , ² DPP4i enhances insulin secretion and suppresses glucagon release, which reduces endogenous glucose production. DPP4is have been associated with modest reductions in HbA1c (0.6%–0.7%), weight neutrality and a low risk of hypoglycaemia. ¹ , ³ SGLT2i lowers plasma glucose by promoting urinary glucose excretion, leading to HbA1c reductions of 0.5%–1.0% and weight loss of approximately 2–3 kg. Both drug classes are well tolerated, with low risks of severe adverse events. ¹ , ⁴ However, the glucose‐lowering effects of this combination were not synergistic in previous RCTs, ⁵ , ⁶ explained by the opposing effects of the drugs on endogenous glucose production. ⁷

Several meta‐analyses and systematic reviews have investigated the efficacy of the combination of SGLT2i and DPP4i, reporting consistent efficacy and safety of this combination. ⁸ , ⁹ , ¹⁰ These prior studies have included variable background medications, ranging from drug‐naïve individuals to those on metformin monotherapy, dual oral agents or insulins. With triple oral therapy becoming more common in clinical practice, there is a growing need for evidence that identifies the most effective agent combinations. ¹¹ In 2018, our research group published a systematic review and meta‐analysis on the efficacy and safety of combination therapy using SGLT2i and DPP4i in treating T2D. ¹² We found that SGLT2i/DPP4i combination therapy significantly improved glycaemic control compared to monotherapy with either agent, reducing HbA1c by approximately 0.6% relative to DPP4i alone and 0.3% relative to SGLT2i alone. ¹²

SGLT2i and DPP4i combination therapy has gained widespread use recently, particularly in Asia, where both agents potentially provide heightened efficacy compared to non‐Asian populations. ¹³ , ¹⁴ , ¹⁵ Therefore, to address current developments and incorporate recent randomized controlled trials (RCTs), we conducted an updated meta‐analysis focusing on the effectiveness and safety of SGLT2i/DPP4i combination therapy in triple combination setting. This updated analysis aimed to provide refined insights into the therapeutic potential of this combination treatment, especially in Asian subpopulations, where diabetes prevalence is rapidly increasing.

2. MATERIALS AND METHODS

2.1. Search strategy and study selection

The Preferred Reporting Items for Systemic Reviews and Meta‐Analyses (PRISMA) checklist statement was implemented in preparation for this meta‐analysis. To identify eligible RCTs, we searched PubMed, EMBASE, Cochrane Library and Web of Science databases comprehensively from inception to 4 January 2024. The eligibility criteria included RCTs that investigated the efficacy and safety of the SGLT2i/DPP4i combination treatment. The search terms are shown in Table S1.

We included RCTs that compared SGLT2i/DPP4i with DPP4i ± placebo or SGLT2i ± placebo. RCTs with information on HbA1c change from baseline and results published in English were eligible for inclusion. Studies in the extended phase were excluded. We investigated study titles, abstracts and full texts to confirm whether the studies met the inclusion criteria. Any disagreements between the authors (MJK and YKC) were resolved by consensus. Figure 1 presents a flowchart of the study selection process.

FIGURE 1.

Flowchart of the selection of eligible trials.

2.2. Data extraction

Changes in HbA1c from baseline to the initial endpoint time point of each study were applied as the primary outcome. The secondary outcomes were the changes in fasting plasma glucose (FPG) levels and body weights, the proportion of subjects achieving an HbA1c level <7.0% and the risk of hypoglycaemia at the same initial endpoint applied for the primary outcome. For studies where baseline changes were not reported, the change from baseline was calculated as the difference in the baseline and end‐of‐treatment values. FPG values presented in mmol/L were converted to mg/dL using the formula: 1 mmol/L = 18 mg/dL. The definitions of hypoglycaemia are shown in Table S2. In addition to outcome measures, two authors (YKC and CHJ) extracted data on the authors and publication years of each study alongside background data, including the anti‐diabetic medications used in addition to SGLT2is or DPP4is, treatment duration, number of randomized subjects, age, percentage of men, body mass index (BMI) and baseline HbA1c. For the continuous outcomes, we extracted the mean difference (MD) between baseline and final measure values in each group alongside the relevant standard deviations. When the precision measures were reported as standard error or confidence intervals, we calculated standard deviations using the reported data. The number of events and randomized subjects for treatment and placebo groups were extracted for the dichotomous outcomes. Two (MJK and YKC) individuals independently extracted data using the pre‐specified protocol. Any discrepancy was resolved by consensus.

2.3. Assessment of methodological quality

We evaluated the quality of the included RCTs according to Cochrane's collaboration tool for assessing the risk of bias. ¹⁶ Two independent reviewers (MJK and YKC) evaluated the risk of bias, and any disagreement was discussed until a consensus was reached. We assessed the risk of bias in random sequence generation and allocation concealment (selection bias), anonymizing participants and personnel (performance bias), anonymizing outcome assessments (detection bias), incomplete outcome data (attrition bias) and selective reporting (reporting bias). The risks of bias were categorized as high, low and unclear. The risk of bias summary assessment is shown in Figure S1.

2.4. Statistical analysis

We calculated the pooled estimates of weighted MD and 95% confidence intervals (CIs) for continuous outcomes, including changes in HbA1c, FPG and body weight, and pooled odds ratio (OR) and their 95% CIs for dichotomous outcomes including the proportion of subjects achieving target HbA1c levels and the risk of hypoglycaemia. We have further stratified the analysis by initial combination or sequential add‐on approaches. Studies were combined using the random effects model, and forest plots were applied to present the summary results. Statistical heterogeneity between the studies was evaluated using I² statistics from the random effect models. Funnel plots with Egger's test were used to assess potential publication bias and any association between the treatment estimates and sample size. To explore heterogeneity, we conducted subgroup analyses of HbA1c reduction based on the baseline HbA1c level (using 8.0%–8.5% as the cut‐off) and racial groups (Asian vs. non‐Asian). Given the prior evidence suggesting baseline HbA1c and ethnicity may influence the efficacy of DPP4 inhibitors, we selected these factors for subgroup analysis to assess treatment heterogeneity. We used R version 4.1.3 for all analyses.

3. RESULTS

3.1. Search characteristics and results

A flowchart of the study selection process is shown in Figure 1, and the characteristics of the included studies are presented in Table 1. A total of 17 citations for SGLT2i plus DPP4i were identified via our electronic literature search, of which 13 eligible RCTs involving 4282 subjects with T2D investigated the combined efficacy and safety of SGLT2i/DPP4i versus DPP4i; nine eligible RCTs involving 3306 subjects with T2D investigated the combined efficacy and safety of SGLT2i/DPP4i versus SGLT2i. ⁵ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² Five RCTs were included in both meta‐analyses, including both cases (i.e., SGLT2i/DPP4i v. DPP4i and SGLT2i/DPP4i vs. SGLT2i). In these five RCTs, the combined efficacy and safety of SGLT2i/DPP4i were evaluated as the simultaneous combination treatment compared to DPP4i or SGLT2i alone in the metformin‐treated or treatment‐naïve patients with T2D. In the other 12 RCTs, the added efficacy and safety of SGLT2i or DPP4i was compared with the equivalent dose of placebo in the metformin‐treated or single component‐treated patients (i.e., DPP4i for the SGLT2i add‐on trials and SGLT2i for the DPP4i add‐on trials) with T2D.

TABLE 1.

Baseline characteristics of included studies.

Author, year	Study duration (weeks)	Background therapy	Interventions	N	Men (%)	Asian (%)	Baseline data
							Age (y)	Weight (kg)	BMI (kg/m2)	HbA1c (%)	HbA1c (mmol/mol)	FPG (mg/dL)
Jabbour, 2014 17	24	Metformin + sitagliptin b	Placebo	224	52.7	1.8	55.0	89.2	NA	8.0	64.0	163.0
			Dapagliflozin d	223	57.0	0.9	54.8	91.0	NA	7.9	63.0	162.2
Defronzo, 2015 18	24	Metformin	Linagliptin	128	50	10.9	56.2	85	30.6	8.0	64.2	156.3
			Empagliflozin f	140	46.4	14.3	55.5	87.7	31.8	8.0	64.2	159.9
			Linagliptin e +Empagliflozin f	134	53.7	16.4	57.1	85.5	30.6	7.9	62.8	154.6
Mathieu, 2015 19	24	Metformin + saxagliptin c	Placebo	160	47.5	0.6	55	NA	32.2	8.2	65.8	177.0
			Dapagliflozin d	160	43.7	0.6	55.2	NA	31.2	8.2	66.6	179.0
Matthaei, 2015 20	24	Metformin + dapagliflozin d	Placebo	162	46.9	4.9	54.5	NA	31.4	7.9	62.4	158.0
			Saxagliptin c	153	47.7	3.3	54.7	NA	31.4	8.0	63.6	164.0
Rosenstock, 2015 5	24	Metformin	Saxagliptin c	176	50	6.0	55	NA	31.5	8.9	73.4	185.0
			Dapagliflozin d	179	53	6.0	54	NA	31.8	9.0	75.2	192.0
			Saxagliptin c + Dapagliflozin d	179	47	7.0	53	NA	31.8	8.9	74.0	180.0
Rodbard, 2016 21	26	Metformin + sitagliptin b	Placebo	106	51.9	11.3	57.5	90	31.7	8.4	68.3	180.4
			Canagliflozin g	107	61.7	18.7	57.4	94.1	32.3	8.5	69.4	185.5
Pratley, 2017 22	26	Metformin	Sitagliptin b	247	62.3	11.7	54.8	89.8	31.7	8.5	69.4	177.4
			Ertugliflozin h	248	54	8.9	55.3	88	31.5	8.6	70.2	179.5
			Sitagliptin b + Ertugliflozin h	244	51.6	14.8	55.1	87.5	31.8	8.6	70.1	177.2
Dagogo‐Jack, 2017 23	26	Metformin + sitagliptin b	Placebo	152	65.4	21.6	58.3	86.4	30.3	8.0	64.3	169.6
			Ertugliflozin h	152	53.6	18.3	59.7	86.6	30.9	8.0	64.0	171.7
Søftland, 2017 24	24	Metformin + linagliptin e	Placebo	108	55.6	29.6	55.9	82.3	29.6	8.0	63.6	163.8
			Empagliflozin f	110	64.5	27.3	55.4	84.4	29.9	8.0	63.6	169.2
Tinahones (a), 2017 25 , a	24	Metformin + empagliflozin 10 mg	Placebo	125	56	0	56.8	85.6	30.8	8.0	64.3	157.1
			Linagliptin e	122	56.6	0.8	56.6	88.4	31.3	8.0	64.4	159.5
Tinahones (b), 2017 25 , a	24	Metformin + empagliflozin f	Placebo	110	57.3	0	56.1	89.9	32.0	7.9	62.6	155.4
			Linagliptin e	110	47.3	0	56.6	85.7	30.8	7.8	61.9	152.1
Han, 2018 26	24	Metformin + sitagliptin b	Placebo	66	48.5	100	57.4	67.9	26.1	7.9	63.0	159.5
			Ipragliflozin 50 mg	73	50.7	100	57.6	67.5	25.5	7.9	63.0	158.0
Handelsman, 2018 27	26	Metformin	Sitagliptin b	229	48	5.2	55.8	NA	32.8	8.9	74.0	175.0
			Saxagliptin c + Dapagliflozin d	232	43.1	3.4	55.9	NA	33.3	8.8	73.0	171.8
Rosenstock, 2019 28	24	Metformin	Saxagliptin c	291	54	2.1	57	92.3	32.4	83.	67.0	180.2
			Dapagliflozin 5 mg	289	52.6	3.1	55.9	89.5	31.8	8.2	66.0	176.6
			Saxagliptin c + Dapagliflozin 5 mg	290	49	3.1	57.2	87.2	31.5	8.1	65.0	171.1
Sahay, 2023 29	16	Metformin	Sitagliptin b	139	57.6	100	48.6	67.9	26.4	9.0	75.0	168.9
			Dapagliflozin d	139	62.6	100	49.1	68.8	26.6	8.9	74.0	167.4
			Sitagliptin b + Dapagliflozin d	137	59.9	100	49.1	66.5	26.0	9.1	75.0	171.7
Lee, 2023 30	24	Metformin + dapagliflozin d	Placebo	154	52.6	100	55	73.6	27.1	7.9	63.0	136.6
			Gemigliptin 50 mg	158	62.7	100	55.7	71.1	26.0	7.8	62.0	139.8
Moon, 2023 31	35	Metformin + dapagliflozin d	Placebo	142	64.1	100	58.2	70.2	25.7	NA	NA	144.0
			Evogliptin 5 mg	141	55.3	100	55.1	71.6	26.1	NA	NA	146.4
Jain, 2024 32	16	Metformin	Linagliptin e	116	NA	100	49.8	65.7	25.3	8.8	73.0	159.2
			Linagliptin e + Dapagliflozin d	116	NA	100	49.5	66	25.2	8.7	72.0	155.5

Note: Unless otherwise indicated, data are presented as the mean (continuous variables) or percentages (dichotomous variables).

Abbreviations: BMI, body mass index; FPG, fasting plasma glucose; HbA1c, haemoglobin A1c; NA, not available.

^a Tinahones et al. comprised two separate trials of linagliptin (5 mg)/empagliflozin (10 mg) or placebo/empagliflozin (10 mg) plus metformin (Tinahones [a]) or linagliptin (5 mg)/empagliflozin (25 mg) or placebo/empagliflozin (25 mg) plus metformin (Tinahones [b]).

^b Sitagliptin 100 mg.

^c Saxagliptin 5 mg.

^d Dapagliflozin 10 mg.

^e Linagliptin 5 mg.

^f Empagliflozin 25 mg.

^g Canagliflozin 100 mg or 300 mg; 6 weeks after starting canagliflozin 100 mg, the dose was increased to 300 mg (or from placebo to matching placebo) if all of the following criteria were met: baseline estimated glomerular filtration rate ≥ 70 mL/min/1.73 m²; fasting self‐monitored blood glucose ≥5.6 mmol/L (≥ 100 mg/dL); no volume‐depletion‐related adverse events within 2 weeks of dose increase.

^h Ertugliflozin 15 mg.

Table 1 provides an overview of the baseline characteristics of the 20 studies included in this systematic review and meta‐analysis, encompassing a study duration ranging from 16 to 35 weeks, with the majority conducted over 24 weeks. The trials uniformly employed metformin as background therapy. The intervention arms evaluated various SGLT2is (including dapagliflozin, empagliflozin, canagliflozin, ertugliflozin and ipragliflozin), DPP4is (including linagliptin, saxagliptin, sitagliptin and gemigliptin) or a combination. Sample sizes per study group ranged from 66 to 291 participants, with the proportion of male participants varying between 43.1% and 65.4%. The mean age of participants spanned from 48.6 to 58.3 years, while the baseline body weight ranged from 54.5 to 89.9 kg. BMI values indicated that most participants were overweight or obese, with baseline measurements ranging from 25.2 to 33.3 kg/m². Glycaemic parameters at baseline were consistent with poor glycaemic control, with mean HbA1c levels between 7.8% and 9.1% (62.0–75.0 mmol/mol) and FPG levels ranging from 136.6 to 192.0 mg/dL.

3.2. Glycaemic efficacy

In the meta‐analysis comprising 13 RCTs that compared SGLT2i/DPP4i combined therapy with DPP4i alone, the combined therapy significantly reduced HbA1c compared with DPP4i alone (MD −0.57%, 95% CI −0.67% to −0.46%; Figure 2A). When the reduction in HbA1c was further analysed according to the combined SGLT2i and DPP4i (i.e., an initial combination of SGLT2i and DPP4i or SGLT2i added to DPP4i), the reduction was slightly greater when SGLT2i was added to DPP4i (MD −0.73%, 95% CI −0.83% to −0.62%; Figure S2A), compared with an initial combination of SGLT2i and DPP4i (MD −0.45%, 95% CI −0.53% to −0.37%; Figure S2A). In the meta‐analysis of nine RCTs that compared SGLT2i/DPP4i with SGLT2i, SGLT2i/DPP4i was associated with a significant reduction in HbA1c compared with SGLT2i alone (MD −0.46%, 95% CI −0.55 to −0.38%; Figure 2B). When the decrease in HbA1c was further analysed according to the combined SGLT2i and DPP4i treatment, the initial combination of SGLT2i and DPP4i showed similar HbA1c reductions (MD −0.42%, 95% CI −0.50% to −0.34%) when compared with DPP4i supplementation to the initial SGLT2i treatment (MD −0.50%, 95% CI −0.64% to −0.36%; Figure S2B).

FIGURE 2.

The weighted mean difference of changes in HbA1c values from baseline. (A) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus DPP4i. (B) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus SGLT2i. The squares indicate an individual study's effect, while the size of the squares reflects the study's weight, with the horizontal lines extending from the symbols representing 95% confidence intervals (CIs). The squares indicate the pooled estimates.

FPG changes from baseline were assessed in 11 studies of SGLT2i/DPP4i versus DPP4i (n = 4282) and eight for SGLT2i/DPP4i versus SGLT2i (n = 2977). Compared with DPP4i, SGLT2i/DPP4i significantly lowered FPG (MD −22.91 mg/dL, 95% CI −27.76 to −18.05 mg/dL; Figure S3A). The FPG‐mediated decrease was more pronounced in studies utilizing a sequential add‐on approach (MD −30.52 mg/dL, 95% CI −34.41 to −26.64 mg/dL) compared to those using an initial combination strategy (MD −17.29 mg/dL, 95% CI −22.51 to −12.07 mg/dL; Figure S4A). When compared with SGLT2i alone, SGLT2i/DPP4i significantly reduced FPG (MD −10.08 mg/dL, 95% CI −12.92 to −7.23 mg/dL; Figure S3B), showing consistent efficacy across different combination approaches (Figure S4B).

Next, we analysed studies that reported the proportion of participants attaining the target HbA1c levels of <7.0% (Figure S5). The SGLT2i/DPP4i group demonstrated a significantly greater proportion of participants achieving the HbA1c target of <7.0% compared to the DPP4i group (OR 3.26, 95% CI 2.49–4.26) or the SGLT2i group (OR 4.57, 95% CI 3.08–6.77). This significant effect was observed regardless of the combined treatment approaches (Figure S6).

3.3. Glycaemic efficacy according to subgroups

The subgroup analysis of HbA1c reduction based on baseline HbA1c levels (<8.5% versus ≥8.5%) revealed significant differences in the additive effect of SGLT2i (Figure 3A). For participants with a baseline HbA1c <8.5%, the combination therapy with the SGLT2i resulted in a MD of −0.40% (95% CI −0.54 to −0.26), indicating a modest but significant reduction in HbA1c. In contrast, for participants with a baseline HbA1c ≥8.5%, the HbA1c reduction by SGLT2i was more pronounced, with an MD of −0.76% (95% CI −0.95 to −0.57). This indicates that the glycaemic effect of SGLT2i is substantially greater in patients with higher baseline HbA1c levels. The heterogeneity was minimal in both subgroups, underscoring the robustness of these findings. According to the baseline HbA1c values, the additive effect of DPP4i, when combined with SGLT2i, was consistent across subgroups (MD; −0.40% vs. −0.48%) (Figure 3B). This indicates that DPP4i provides consistent glycaemic benefits regardless of initial HbA1c levels.

FIGURE 3.

The weighted mean difference of changes in HbA1c from baseline according to the baseline HbA1c. The cut‐off for the baseline HbA1c is 8.5%. (A) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus DPP4i according to the HbA1c baseline. (B) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus SGLT2i according to the HbA1c baseline. The squares indicate an individual study's effect, and the size of the squares reflects the study's weight, with the horizontal lines extending from the symbols representing 95% confidence intervals (CIs). The squares indicate the pooled estimates.

The subgroup analysis of the Asian and non‐Asian populations demonstrated that the additive effect of SGLT2i, when combined with DPP4i, was consistent across groups. For Asian participants, the SGLT2i/DPP4i combination resulted in an MD in HbA1c levels of −0.57% (95% CI −0.76 to −0.37) compared to DPP4i alone. Similarly, for non‐Asian participants, the MD was −0.59% (95% CI −0.71 to −0.46). In contrast, the MD in HbA1c reduction when comparing SGLT2i/DPP4i to SGLT2i, representative of the glycaemic efficacy of DPP4i, was −0.55% (95% CI −0.71 to −0.39) in the Asian subgroup and − 0.38% (95% CI −0.46 to −0.31) in the non‐Asian subgroup. These findings suggest that the glycaemic efficacy of DPP4i may be more effective in Asian individuals when combined with SGLT2i (Figure 4).

FIGURE 4.

The weighted mean difference of changes in HbA1c from baseline according to the Asian versus non‐Asian studies. The upper studies mainly include Asians, and the lower studies include non‐Asians. (A) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus DPP4i according to the HbA1c baseline. (B) The change in HbA1c (%) from baseline using SGLT2i/DPP4i versus SGLT2i according to the HbA1c baseline. The squares indicate an individual study's effect, and the size of the squares reflects the study's weight, with the horizontal lines extending from the symbols representing 95% confidence intervals (CIs). The squares indicate the pooled estimates.

3.4. Clinical efficacy other than glycaemic control

Significant body weight reductions were observed from baseline between SGLT2i/DPP4i and DPP4i (MD −1.57 kg, 95% CI −1.93 to −1.20 kg), indicating that SGLT2i promotes a notable weight‐lowering effect. However, when compared with SGLT2i alone, the SGLT2i/DPP4i combination showed a neutral impact on body weight (MD 0.36 kg, 95% CI 0.12 to 0.60 kg). Substantial reductions in systolic blood pressure (SBP) were observed when comparing SGLT2i/DPP4i and DPP4i monotherapy (MD −3.07 mmHg, 95% CI −4.14 to −1.99 mmHg), highlighting the significant effect on lowering BP by SGLT2i. In contrast, when SGLT2i/DPP4i was compared to SGLT2i alone, no significant difference was found in the SBP (MD 0.17 mmHg, 95% CI −0.77 to 1.10 mmHg) (Figures S7 and S8).

3.5. Safety

All RCTs included in our meta‐analyses reported the number of hypoglycaemic events; however, none demonstrated significant differences in hypoglycaemia risk between treatment groups. As shown in Figure 5A, the risk of hypoglycaemia was low and comparable between the SGLT2i/DPP4i and DPP4i groups (relative risk [RR] 1.12, 95% CI 0.81 to 1.55). Similarly, when comparing SGLT2i/DPP4i to SGLT2i alone, the RR was 1.21 (95% CI 0.72–2.04; Figure 5B).

FIGURE 5.

The relative risk of hypoglycaemia. (A) Relative risk of hypoglycaemia using SGLT2i/DPP4i versus DPP4i. (B) Relative risk of hypoglycaemia using SGLT2i/DPP4i versus SGLT2i. The squares indicate an individual study's effect, and the size of the squares reflects the study's weight, with the horizontal lines extending from the symbols representing 95% confidence intervals (CIs). The squares indicate the pooled estimates.

3.6. Exploration of publication bias

Funnel plots for key efficacy and safety outcomes are shown in Figures S9–S14. Egger's test indicated possible publication bias in some analyses (Figures S9B, S11A and S13B).

4. DISCUSSION

This updated meta‐analysis provides robust evidence of the efficacy and safety of combination therapy using SGLT2i and DPP4i to manage T2D. Our findings demonstrate that the combined use of these agents significantly improves glycaemic control, with an MD in HbA1c reduction of 0.57% for SGLT2i and 0.46% for DPP4i within the combined treatment. These results are consistent with our previous study, ¹² which reported HbA1c reductions of 0.62% and 0.35% for SGLT2i and DPP4i, respectively. Furthermore, subgroup analyses in the present updated meta‐analysis revealed that the additional effect of DPP4i on lowering HbA1c was more pronounced in Asian participants (−0.55%) than non‐Asian participants (−0.38%). This finding suggests that this combination treatment potentially has greater therapeutic efficacy in Asian populations.

Combined DPP4i and SGLT2i therapy has become increasingly popular for managing T2D. Several meta‐analyses, including our 2018 study, have previously examined the efficacy and safety of this combination therapy. ⁹ , ¹⁰ , ¹² Indeed, our previous study reported significant HbA1c reductions using combined SGLT2i and DPP4i therapy of −0.62% and −0.35%, based on analyses of eight RCTs comparing the combination to DPP4i and five RCTs comparing it to SGLT2i, respectively. ¹² Similarly, Li et al. (14 RCTs, 4828 patients) reported reductions of −0.71% and −0.31%, ¹⁰ while Min et al. (7 RCTs, 2082 patients) observed a decrease of −0.6% when SGLT2i was added to DPP4i. ⁹ These meta‐analyses varied in scope and methodology. Li et al. emphasized subgroup analyses based on the SGLT2i dosage, ¹⁰ while Min et al. focused on differences between sequential and simultaneous drug combination strategies. ⁹ Our prior research provided a detailed stratification according to baseline HbA1c levels. ¹² Despite these differences, all three studies confirmed that combination therapy offers substantial glycaemic improvement with a favourable safety profile, underscoring its value for patients with inadequate glycaemic control.

Since the publication of these meta‐analyses, data from additional RCTs have provided combination therapy using SGLT2i and DPP4i further prominence in clinical practice. ²² , ²³ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² This updated meta‐analysis incorporates these newer studies, allowing for expanded analyses, including examining ethnic differences. A systematic review and meta‐analysis were conducted in 2013, ¹³ which observed that DPP4i reduced HbA1c levels by a weighted MD of −0.92% in studies with predominantly Asian participants, compared to an MD of −0.65% in non‐Asian studies. A study by Cai et al. also found that Asian patients experienced a significant reduction in HbA1c (MD −0.81%; 95% CI −0.95 to −0.68), which was greater than that observed in Caucasian patients. ³³ Notably, our results align with these observations, with DPP4i showing a more pronounced additive effect in Asians (MD −0.55%, 95% CI −0.71 to −0.39) compared to non‐Asians (MD −0.38%, 95% CI −0.46 to −0.31) (Figure 4). Despite notable heterogeneity in the Asian subgroup (I ² = 65%), which suggests that additional factors may influence therapeutic outcomes, our findings reaffirm that DPP4i promotes a superior glucose‐lowering efficacy in Asian patients relative to other ethnic groups.

A meta‐analysis reported that glucagon‐like peptide‐1 receptor agonists (GLP‐1RAs), another incretin‐based therapy, also show greater HbA1c reduction in Asian populations than in non‐Asian. ³⁴ Some explanations have been proposed for these ethnic differences. ³⁵ Asian patients with T2D show a more prominent early‐phase β‐cell dysfunction, with relatively preserved insulin sensitivity and lower obesity rates. ³⁶ This could be one reason why incretin‐based therapies, which enhance glucose‐dependent insulin secretion and effectively address β‐cell impairment, act more effectively in this population. Additionally, lower GLP‐1 secretion in East Asians compared to Caucasians may heighten the response to incretin‐based drugs. ³⁷ , ³⁸ Beyond these physiological differences, emerging evidence suggests that genetic factors may also contribute to ethnic variability in drug response. For instance, the GLP1R gene variant rs3765467, which may enhance GLP‐1 receptor signalling, is more frequently observed in Asian populations. ³⁹ Similarly, the rs7903146 risk allele of the TCF7L2 gene, which may reduce DPP4i responsiveness, is less prevalent in East Asians than in Europeans. ⁴⁰ , ⁴¹ Dietary factors may also play a role. Higher fish intake has been associated with improved glycaemic response to DPP4i, but direct evidence supporting dietary habits influencing ethnic differences in drug response remains limited. ⁴² , ⁴³

Regarding the efficacy of SGLT2is, our analysis found no significant difference between studies using Asian and non‐Asian populations. However, we observed that the efficacy of SGLT2is in combination therapy remained consistent across baseline HbA1c levels, with more pronounced reductions in patients with HbA1c levels ≥8.5% (MD −0.76%, 95% CI −0.95 to −0.57) than those with HbA1c values <8.5% (MD −0.40%, 95% CI −0.54 to −0.26; Figure 3). These results are consistent with the well‐established notion that the glucose‐lowering effect of SGLT2i is proportional to baseline glycaemic levels. ⁴⁴ At higher HbA1c levels, increased renal glucose filtration leads to enhanced glucose excretion following SGLT2 inhibition, thus amplifying its therapeutic benefit. ⁴⁵ These findings indicate the importance of individualized treatment strategies, particularly for patients with varying levels of baseline glycaemic control.

This study is an updated meta‐analysis that builds on previous evaluations of SGLT2i/DPP4i combination therapy by including recently published RCTs and is the first to assess ethnic differences in treatment efficacy for both DPP4i and SGLT2i. The strengths of our study also include adherence to PRISMA guidelines, the inclusion of high‐quality RCTs and comprehensive subgroup analyses that provide clinically relevant insights. In particular, the focus on ethnic differences, including the unique response patterns observed in Asian populations, represents a significant advancement over prior meta‐analyses. Additionally, including recent RCTs has enabled a more contemporary evaluation of combination therapy efficacy, ensuring relevance to current clinical practices. This systematic exploration of both glycaemic and non‐glycaemic outcomes, such as body weight and blood pressure, further adds to the robustness of our findings.

Limitations such as the short duration of most trials and the lack of cardiovascular outcome data warrant caution when interpreting our findings to longer‐term implications. Furthermore, the observed heterogeneity in some subgroups, particularly in the Asian populations (I ² = 65%), suggests that regional differences in dietary habits, genetic predispositions and healthcare access may influence outcomes. Further sensitivity analyses based on these potential contributing factors were unable due to lack of information on baseline characteristics and small numbers of studies included. Moreover, our findings could have been influenced by small‐study effects or selective publication, as indicated by Egger's test. However, this potential bias should be interpreted cautiously due to the limited number of studies included in the analysis. To confirm these results and minimize the impact of publication bias, larger prospective studies are warranted in the future. Excluding non‐English studies in our data collection may have omitted relevant data from certain regions. Lastly, the heterogeneity in definitions of hypoglycaemia may have influenced the reliability of pooled estimates for safety outcomes.

In conclusion, this updated meta‐analysis highlights the efficacy and safety of combined SGLT2i/DPP4i therapy, with a trend towards improved effectiveness of DPP4i in Asian populations. This combination therapy showed additional benefits on weight reduction and blood pressure when compared to DPP4i monotherapy, making it a valuable option for patients with high metabolic burden. Future research should focus on long‐term outcomes, including the durability of this combination therapy and cardiovascular benefits, to further optimize T2D management strategies and validate these findings in diverse patient groups.

AUTHOR CONTRIBUTIONS

M.J.K. and Y.K.C. contributed equally to this study as first authors. M.J.K. and Y.K.C. contributed to data collection, interpreted the data and drafted the initial manuscript. S.H.K. contributed to statistical analysis. J.Y.M. and C.H.J. investigated and revised the manuscript. W.J.L. designed the research, interpreted the results, revised the manuscript and took primary responsibility for the final content. All the authors approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interests.

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.16550.

Supporting information

Data S1. Supporting information.

Kim MJ, Cho YK, Kim S, Moon JY, Jung CH, Lee WJ. Efficacy and safety of combination therapy using SGLT2 and DPP4 inhibitors to treat type 2 diabetes: An updated systematic review and meta‐analysis with focus on an Asian subpopulation. Diabetes Obes Metab. 2025;27(9):5019‐5031. doi: 10.1111/dom.16550

DATA AVAILABILITY STATEMENT

The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.