Combined Semaglutide and Sodium–Glucose Cotransporter-2 Inhibitor Use and Metabolic Outcomes in Chinese Adults With Diabetes: Post Hoc Analysis

Abstract

Introduction

Real-world use of once-weekly semaglutide in adults with type 2 diabetes in China has been shown in SCHOLAR. This post hoc analysis investigated the use of combination semaglutide–sodium–glucose cotransporter-2 inhibitor (SGLT2i) treatment, glycated hemoglobin (HbA1c) changes, and safety in a subset of SCHOLAR participants.

Methods

Participants from SCHOLAR were included in this post hoc, single-arm analysis if they were prescribed once-weekly subcutaneous semaglutide and had ≥ 1 SGLT2i prescription within 6 months of initiating semaglutide. Participants were grouped according to whether semaglutide use was continuous or not and by baseline HbA1c (≥ 7%). We evaluated the proportion of SCHOLAR participants prescribed combined semaglutide–SGLT2i. Key endpoints were: change in HbA1c from baseline to month 6; proportion of participants achieving a HbA1c level  < 7% at month 6; change in lipid profile from baseline to month 6; and safety profile of combined semaglutide–SGLT2i.

Results

A total of 11,361 (42.3%) SCHOLAR participants were treated with combined semaglutide–SGLT2i. Mean change in HbA1c (95% confidence interval [CI]) from baseline to month 6 was − 0.64%-points (− 0.71, − 0.56) (p < 0.0001) for the overall combined semaglutide–SGLT2i group. The greatest change was seen in the continuous semaglutide group with baseline HbA1c ≥ 7% (− 1.20%-points [95% CI − 1.36, − 1.04]). The proportion of participants with HbA1c levels  < 7% increased from 37.8% at baseline to 55.1% at month 6 for the overall combined treatment group. There were significant reductions in mean total cholesterol (p < 0.0001), low-density lipoprotein cholesterol (p < 0.0001), and triglycerides (p = 0.0043) as well as a significant increase in high-density lipoprotein cholesterol (p < 0.0001). Gastrointestinal adverse events were most frequent.

Conclusions

Combined semaglutide–SGLT2i treatment was prevalent in this subset of the SCHOLAR population and associated with significant reduction in HbA1c and changes in lipid profile. The safety profile of combined semaglutide–SGLT2i was similar to the overall SCHOLAR population.

Trial Registration Number

NCT06351748.

Key Summary Points

Why carry out this study?

There are limited data regarding the use of combined semaglutide (a glucagon-like peptide 1 [GLP-1] analogue) and sodium–glucose cotransporter-2 inhibitor (SGLT2i) treatment, as well as its effectiveness and safety profile in China.

What was learned from the study ?

Combined semaglutide–SGLT2i treatment was associated with significant reduction in glycated hemoglobin (HbA1c) and changes in lipid profile.

The safety profile was similar to that observed for the overall SCHOLAR population.

Introduction

Management of type 2 diabetes (T2D) is multifaceted, frequently necessitating combination therapy to achieve and maintain glycemic goals [1]. Comorbidity management is another consideration, with people with T2D at higher risk than the general population of developing cardiovascular disease (CVD) [2] and chronic kidney disease (CKD) [3].

Results of large, randomized controlled studies investigating glucagon-like peptide 1 receptor agonists (GLP-1RAs) and sodium–glucose cotransporter-2 inhibitors (SGLT2is), revealed beneficial and/or protective effects in participants with T2D with respect to glucose reduction, as well as cardiovascular (CV) and renal outcomes [4–8]. In light of this evidence, GLP-1RAs and SGLT2is with proven CV or renal benefits are recommended in people with T2D who have or are at risk of developing CVD or CKD [9].

Given the different mechanisms of action for GLP-1RAs and SGLT2is, potential additive effects of using them in combination have been investigated [10, 11]. A number of studies reported a decreased risk of negative CV outcomes, such as major adverse CV events (MACE), CV-related mortality, and hospitalization for heart failure, with combined GLP-1RA–SGLT2i therapy versus either therapy alone [11, 12]. There is also evidence that combined GLP-1RA–SGLT2i therapy results in greater improvement in glycemic control indicators, such as glycated hemoglobin (HbA1c) and fasting plasma glucose, as well as greater reduction in body weight, in people with T2D compared with either therapy alone [12, 13]. However, knowledge regarding the efficacy and utilization of combined GLP-1RA–SGLT2i outside of a clinical setting in China is limited.

Semaglutide is a GLP-1 analogue developed as an adjunct to diet and exercise to improve glycemic control in adults with T2D. The efficacy and safety of once-weekly subcutaneous (s.c.) semaglutide for the treatment of T2D has been investigated in the SUSTAIN program [12, 14–22] and the SURE program [23–31], which studied semaglutide in clinical and real-world settings, respectively. In China, real-world use of semaglutide was shown in the Study of CHina Ozempic cLinicAl pRactice (SCHOLAR, NCT06351748), which investigated the impact and safety of once-weekly s.c. semaglutide in adults with T2D using data from the regional electronic health records database in Tianjin, China [32].

Guidelines from the Chinese Diabetes Society support combination therapy to improve glycemic control in people with uncontrolled T2D. For people with T2D and CVD or CKD, combination therapy with GLP-1RA and SGLT2i proven to have CV and renal benefit is recommended, but there is a scarcity of real-world evidence in China regarding how frequently combination therapy is prescribed and its impact on glycemic control and its safety profile in routine clinical practice [9].

We performed a post hoc analysis in a subset of the SCHOLAR population to evaluate the proportion of participants prescribed combination therapy and its impact on HbA1c changes and its safety profile.

Methods

Study Design

The design of SCHOLAR (ClinicalTrials.gov; NCT06351748) has been described previously [32]. Briefly, SCHOLAR included eligible participants extracted from the Tianjin Healthcare and Medical Big Data Platform, a city-level electronic health records database, which includes electronic medical records from 82 hospitals (43 tertiary hospitals and 39 secondary hospitals) in Tianjin. Eligible participants were those who initiated once-weekly s.c. semaglutide for the first time between January 1, 2022, and February 28, 2023 (index date), as documented in the database. In this post hoc analysis, participants from SCHOLAR were included if they were prescribed once-weekly s.c. semaglutide and had ≥1 SGLT2i prescription, including empagliflozin, dapagliflozin, and canagliflozin, within 6 months of initiating semaglutide. This post hoc analysis was a single-arm study that evaluated and further analyzed the proportion of combination semaglutide–SGLT2i therapy use in the SCHOLAR population. Analyses were stratified by semaglutide persistence and baseline HbA1c.

Key Endpoints

The key endpoints were: change in HbA1c from baseline to month 6; proportion of participants achieving an HbA1c level < 7% at month 6; change in lipid profile from baseline to month 6; and the safety profile of combined semaglutide–SGLT2i. A ± 3-month assessment window was applicable, as assessments were not scheduled at exactly 6 months as a result of the real-world nature of the study.

Statistical Analysis

Participants were grouped according to whether semaglutide use was continuous or not and by baseline HbA1c (≥ 7%). Continuous semaglutide treatment was defined as a lack of discontinuation of once-weekly semaglutide for 6 months, during the observation period (post-index date). Discontinuation was defined as failure to refill once-weekly semaglutide within 45 days of the depletion of the previous prescription.

Descriptive statistics were performed for baseline characteristics by group. The change in HbA1c was calculated as the mean change from baseline to month 6. Only participants with HbA1c values at baseline and month 6 were included in the analysis. P-values were calculated using paired t tests. The lipid levels analyzed in this analysis were total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, and triglycerides. The change in lipid levels was calculated as the mean change from baseline to month 6. As with HbA1c analysis, only participants with lipid values at baseline and month 6 were included in each corresponding analysis, and p-values were calculated using paired t tests.

Ethical Approval

SCHOLAR was conducted in accordance with the Declaration of Helsinki and the Guidelines for Good Pharmacoepidemiology Practices. Patient consent was not required as the analyses used retrospective de-identified data without access to personal identifying information. The study protocol and the application for a waiver to the informed consent were approved by the independent ethics committee and institutional review board.

Results

Patients and Baseline Characteristics

Of the full analysis set of 26,859 participants included in SCHOLAR, 11,361 (42.3%) participants were treated with combined semaglutide–SGLT2i and thus included in this post hoc analysis; 3607 (31.7%) of the included participants were continuously treated with semaglutide. Of these continuously treated participants, 685 (19.0%) had HbA1c values at baseline and month 6, 419 (61.2%) of whom had a baseline HbA1c ≥ 7%. Of the participants included in the overall combined semaglutide–SGLT2i group, 1855 (16.3%) had HbA1c values at baseline and month 6, 1154 (62.2%) of whom had a baseline HbA1c ≥ 7%. Full details of the population disposition are presented in Fig. 1. For all participants treated with combined semaglutide–SGLT2i (n = 11,361), the mean (standard deviation [SD]) age, duration of T2D, body mass index (BMI), and HbA1c at baseline was 57.2 (12.4) years, 4.6 (3.2) years, 26.5 (3.9) kg/m², and 8.0% (1.8), respectively. Overall, 701 (37.8%) participants had a baseline HbA1c < 7%. There were more male participants than female participants (6860 [60.4%] versus 4501 [39.6%], respectively). CVD was the most common baseline comorbidity in the overall combined semaglutide–SGLT2i group (80.2%), followed by coronary heart disease (65.9%) and hypertension (63.4%). Baseline characteristics and comorbidities are presented in Tables 1 and 2, respectively.

Fig. 1.

Participant disposition flow chart. Percentages at each level were calculated based on the participant number of the previous level. HbA1c glycated hemoglobin, SGLT2i sodium–glucose cotransporter-2 inhibitor, T1D type 1 diabetes, T2D type 2 diabetes

Table 1.

Baseline participant demographics and characteristics

Characteristic	Combined semaglutide–SGLT2i treatment (overall) n = 11,361	Combined semaglutide–SGLT2i treatment (continuous semaglutide) n = 3607
Age, years	57.2 (12.4)	58.1 (11.6)
Sex, n (%)
Female	4501 (39.6)	1461 (40.5)
Male	6860 (60.4)	2146 (59.5)
Duration of T2D, years	4.6 (3.2)	4.7 (3.1)
BMI, kg/m2a	26.5 (3.9)	26.5 (3.7)
HbA1c, %b	8.0 (1.7)	7.9 (1.7)
HbA1c < 7%, n (%)c	701 (37.8)	266 (38.8)
Total cholesterol, mmol/Ld	5.0 (1.3)	5.0 (1.3)
HDL cholesterol, mmol/Le	1.1 (0.3)	1.1 (0.3)
LDL cholesterol, mmol/Lf	3.3 (1.0)	3.2 (1.0)
Triglycerides, mmol/Lg	2.4 (2.5)	2.3 (2.1)
UACR (mg/g)h	134.0 (439.9)	134.2 (446.0)
eGFR, mL/min/1.73m2i	101.1 (21.3)	100.1 (21.2)

Data are mean (SD) unless otherwise stated

BMI body mass index, eGFR estimated glomerular filtration rate, HbA1c glycated hemoglobin, HDL high-density lipoprotein, LDL low-density lipoprotein, SD standard deviation, SGLT2i sodium–glucose cotransporter-2 inhibitor, T2D type 2 diabetes, UACR urine albumin to creatinine ratio

^aOverall n = 824, continuous semaglutide n = 248

^bOverall n = 3973, continuous semaglutide n = 1273

^cOverall n = 1855, continuous semaglutide n = 685 with an available HbA1c measurement at baseline and 6 months post-index

^dOverall n = 3845, continuous semaglutide n = 1192

^eOverall n = 3629, continuous semaglutide n = 1159

^fOverall n = 3641, continuous semaglutide n = 1159

^gOverall n = 4211, continuous semaglutide n = 1336

^hOverall n = 805, continuous semaglutide n = 231

ⁱOverall n = 3410, continuous semaglutide n = 1014

Table 2.

Baseline comorbidities for overall combined semaglutide–SGLT2i treatment group

Comorbidity, n (%)	Combined semaglutide–SGLT2i treatment (overall) n = 11,361
CVD	9106 (80.2)
Coronary heart disease	7482 (65.9)
Hypertension	7206 (63.4)
Dyslipidemia	7098 (62.5)
Diabetic neuropathy	7077 (62.3)
Peripheral vascular disease	6400 (56.3)
Cerebrovascular disease	3868 (34.0)
Diabetic retinopathy	3788 (33.3)
Obesity	3480 (30.6)
Hepatic steatosis	3265 (28.7)
Arrhythmia	2570 (22.6)
CKD	2010 (17.7)
Renal cyst	767 (6.8)
Myocardial infarction	650 (5.7)
Congestive heart failure	637 (5.6)
Kidney stones	448 (3.9)
Sleep apnea	124 (1.1)

CKD chronic kidney disease, CVD cardiovascular disease, SGLT2i, sodium–glucose cotransporter-2 inhibitor

Endpoints

Mean HbA1c reduction was −0.73%-points [95% CI − 0.84, − 0.62] for combined semaglutide–SGLT2i with continuous semaglutide use, and − 0.64%-points [95% CI − 0.71, − 0.56] in the overall combined semaglutide–SGLT2i population (Fig. 2). The proportion of participants with an HbA1c level  < 7% increased from 37.8% at baseline to 55.1% at month 6 for the overall combined semaglutide–SGLT2i group, and from 38.8% to 58.2% in the continuous semaglutide group.

Fig. 2.

Mean change in HbA1c (%) from baseline to month 6 post-index in participants with available HbA1c values. *p < 0.0001. HbA1c glycated hemoglobin, SGLT2i sodium–glucose cotransporter-2 inhibitor, NA not applicable

In the population with HbA1c ≥ 7% at baseline, HbA1c reductions were − 1.20%-points (95% CI − 1.36, − 1.04; n = 419) in the continuous semaglutide group and − 1.10%-points (95% CI − 1.20, − 0.99; n = 1154) in the overall combined semaglutide–SGLT2i population. Additionally, 39.5% of the overall combined semaglutide–SGLT2i group and 43.0% of the continuous semaglutide group had achieved HbA1c  < 7% at month 6 (Fig. 2).

For the overall combined semaglutide–SGLT2i group, the mean total cholesterol, LDL cholesterol, and triglycerides were significantly reduced versus baseline at month 6 (− 0.16 mmol/L [95% CI − 0.23, − 0.10; n = 1,633; p < 0.0001], − 0.17 mmol/L [95% CI − 0.23, − 0.13; n = 1,471; p < 0.0001], and − 0.16 mmol/L [95% CI − 0.27, − 0.05; n = 1,822; p = 0.0043], respectively), while mean HDL cholesterol levels were significantly increased (0.03 mmol/L; 95% CI 0.02, 0.04; n = 1,460; p < 0.0001) (Fig. 3).

Fig. 3.

Mean changes in lipid levels from baseline to month 6. *p < 0.01. HDL high-density lipoprotein, LDL low-density lipoprotein

In this analysis, the most common adverse events were gastrointestinal in nature, including constipation, abdominal discomfort, and diarrhea (experienced by 1.51%, 0.91%, and 0.80% of participants, respectively) (Fig. 4).

Fig. 4.

Selected adverse events for overall combined semaglutide–SGLT2i therapy (SCHOLAR post hoc analysis). ^aLaboratory test results were used to determine occurrence of hypoglycemia, lipase increase, and amylase increase. If hypoglycemia was not stated in the medical records, hypoglycemia was inferred if any glucose value of < 56 mg/dL was recorded post-index. ^bDiabetic retinopathy complications included vitreous hemorrhage, blindness, or conditions requiring treatment with an intravitreal agent or photocoagulation. Adverse events were selected according to product labeling in China and if they were adverse events of special interest. SGLT2i sodium–glucose cotransporter-2 inhibitor

Discussion

SCHOLAR was a real-world study in Tianjin, China, that provided insight on the use and impact of once-weekly semaglutide in adults with T2D who initiated semaglutide [32]. This post hoc analysis of SCHOLAR included 42.3% of the SCHOLAR participants who were treated with combined semaglutide–SGLT2i treatment. Of these, 60.4% were male and 39.6% were female. The proportion of male participants was slightly higher than that observed in the full SCHOLAR analysis population (57.3%) [32]. Female sex is one of the strongest predictors of genitourinary infection with SGLT2i treatment, and clinicians are advised to identify high-risk populations before initiating SGLT2i therapy [33, 34]. Thus, the greater percentage of male participants may reflect clinicians’ cautious prescribing of SGLT2i inhibitors to women. Clinically significant improvements in glycemic control were observed, as well as improvements in lipid levels. In a pooled analysis of the SURE study program, which investigated real-world clinical use of once-weekly semaglutide across different countries, 27.5% of participants were treated with an SGLT2i at baseline [35]. However, individual SURE studies indicate that the prevalence of SGLT2i at the point of semaglutide initiation can vary greatly, possibly reflecting regional prescribing guidelines, regional medical insurance, or medical policy. For example, the proportion of participants treated with SGLT2is at baseline in the SURE Denmark/Sweden multicenter study was 27.7%, similar to the aforementioned SURE pooled analysis [28, 35]. In SURE UK, the proportion of baseline SGLT2i use was higher at 38.6%, but lower in SURE Italy at 9.8% and lower still in SURE France at 0.2% [24, 26, 27].

At baseline, patient demographics, such as HbA1c and lipid levels, were generally similar between the total SCHOLAR population (full analysis set) [32] and those patients receiving combined semaglutide–SGLT2i treatment. Although baseline demographics were generally similar, cardiovascular comorbidities appeared to be more prevalent in the population receiving combined semaglutide–SGLT2i treatment than in the total population, in particular CVD (80.2%; 72.4%), coronary heart disease (65.9%; 59.1%), hypertension (63.4%; 58.1%), and peripheral vascular disease (56.3%; 45.1%), suggesting that these comorbidities rather than glycemic control or BMI may have been more of a driver for treatment with this combination. Indeed, almost a third of patients had HbA1c < 7% at baseline. Regardless of the underlying driver for combined semaglutide–SGLT2i treatment in this post hoc analysis, combined semaglutide–SGLT2i treatment was associated with a significant reduction in HbA1c. A greater reduction in HbA1c was seen in participants who had a baseline HbA1c ≥ 7% (a population that would better reflect that used in clinical trials) compared with the full analysis set, with an even greater reduction in participants who were treated with semaglutide continuously. These data are consistent with the global SUSTAIN program, where HbA1c reductions with semaglutide were noted to be greater in individuals with higher versus lower baseline HbA1c [36]. The phase 3 trial of semaglutide in Chinese patients with T2D (SUSTAIN China) has demonstrated the efficacy of semaglutide for glycemic control, with HbA1c reductions of 1.5% with semaglutide 0.5 mg and 1.8% with semaglutide 1.0 mg at week 30 [18]. However, the combination of semaglutide and SGLT2i was not assessed in this study. In SUSTAIN 9, a multicenter phase 3b trial, semaglutide was added to SGLT2i treatment in participants with T2D, which was associated with an HbA1c reduction of 1.5% at week 30 [12]. A meta-analysis of eight randomized controlled trials that investigated GLP-1RA and SGLT2i combination therapy reported a reduction in HbA1c of 0.77%, which was greater than treatment with either drug class individually [13]. Prospective real-world evidence studies [37] and retrospective studies [38] of patients prescribed SGLT2i and GLP-1RA combination therapy in Spain have reported HbA1c reductions of 1.1%. A retrospective study in the USA has also reported that a combination of SGLT2i and a GLP-1RA could reduce HbA1c by 1.14%, a significant difference of 0.51% (p < 0.001) compared with SGLT2i treatment alone [10]. The higher reductions in HbA1c in the SUSTAIN trials than observed in our study may be a result of the clinical trial setting and its greater oversight of treatment adherence and dosage. In SCHOLAR, dose information was not available; therefore, the average dose may have been lower than the doses used in SUSTAIN 9 and SUSTAIN China. The greater HbA1c reductions observed in our post hoc study among individuals with continuous semaglutide treatment, compared with those who did not have continuous treatment, together with the greater HbA1c reductions reported in the randomized controlled trials, suggest that consistent adherence to semaglutide might be a key driver of glycemic control in real-world clinical practice.

Much of the discourse surrounding GLP-1RA and SGLT2i combination treatment examines the cardiometabolic effects of these two drug classes, with many studies, including real-world studies, reporting a decreased risk of negative CV outcomes, such as MACE, CV mortality, and hospitalization for heart failure, compared with monotherapy of either drug class [10, 11, 39–41]. Clinical trials have also suggested a potential benefit of the GLP-1RA on major renal events and blunting decline in estimated glomerular filtration rate in patients with T2D and CKD [42], and these benefits appear to be consistent in patients with or without SGLT2i use at baseline [43]. Our analysis reported significant decreases in total cholesterol, LDL cholesterol, and triglycerides and a significant increase in HDL cholesterol associated with combined semaglutide–SGLT2i treatment, which may support the benefit on cardiometabolic risk that has been previously reported with GLP-1RA and SGLT2i combination treatment. The more frequent cardiovascular-related comorbidities observed in our post hoc study align with the prescribing guidelines for the use of GLP-1RAs and SGLT2is in patients with established or high-risk CVD. Our analysis does not provide a comprehensive investigation of CVD and CKD outcomes associated with combined GLP-1RA–SGLT2i treatment, although this has been investigated elsewhere [44]. However, there remains a dearth of information regarding CVD and CKD outcomes in a Chinese population. Changes in body weight and blood pressure were also not analyzed in this post hoc analysis because of low collection rate of this information and inaccuracy in medical records. There is evidence to suggest that combining GLP-1RA and SGLT2i may have an additive effect on weight loss, so this may be an area of further research [45].

The safety profile of combined semaglutide–SGLT2i reflected that of s.c. semaglutide reported in SCHOLAR [32] and in other real-world studies, such as the SURE study program [35], as well as clinical trials such as SUSTAIN 9 and SUSTAIN China [12, 18], with gastrointestinal-related adverse events being the most prevalent.

Strengths and limitations of SCHOLAR have been discussed previously [32]. In this post hoc analysis in a subset of the SCHOLAR population, SGLT2i treatment was defined as ≥ 1 SGLT2i prescription within 6 months of initiating semaglutide, so it is not known whether the participant was treated with SGLT2i throughout the follow-up period. The retrospective nature of SCHOLAR and the post hoc methodology of this analysis mean that the conclusions should be interpreted with caution. Additionally, a relatively small proportion of participants had available outcome measures, which may limit the generalizability of findings.

Conclusions

In this post hoc analysis conducted in a subset of the SCHOLAR population, a large proportion of participants were treated with combined semaglutide–SGLT2i therapy, which was associated with a significant reduction in HbA1c and significant changes in the lipid profile. The safety profile of combined semaglutide–SGLT2i was similar to that seen in the overall SCHOLAR study.

Author Contributions

Data were provided and analyzed by the IQVIA real-world evidence team. Shanshan Wang, Bixuan Dong, and Zewei Shen contributed to the design of the study, participated in interpretation of the data, drafting and revising the manuscript, and reviewed and approved the final submitted version.

Funding

This analysis was sponsored by Novo Nordisk. The journal’s Rapid Service Fee was funded by Novo Nordisk.

Data Availability

Data for this analysis were sourced from the Tianjin Healthcare and Medical Big Data Platform, which is not available online. On-site access is available, upon reasonable request to Tianjin Healthcare and Medical Big Data Co., Ltd.

Declarations

Conflict of Interest

Bixuan Dong and Zewei Shen are employees of and shareholders in Novo Nordisk A/S, the sponsors of this study. Shanshan Wang reports no conflicts of interest.

Ethical Approval

SCHOLAR was conducted in accordance with the Declaration of Helsinki and the Guidelines for Good Pharmacoepidemiology Practices. Patient consent was not required as the analyses used retrospective de-identified data without access to personal identifying information. The study protocol and the application for a waiver to the informed consent were approved by the independent ethics committee and institutional review board.