Comparative effectiveness of semaglutide and dulaglutide combined with hypocaloric diet in newly diagnosed type 2 diabetes: a retrospective real-world study

Abstract

Background

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are increasingly used in early T2D management due to their glycemic and weight-reducing effects. However, real-world comparative data in newly diagnosed treatment-naïve individuals remain limited.

Methods

This retrospective real-world study evaluated 106 adults aged 18–75 years with newly diagnosed T2D who initiated semaglutide (1.0 mg) or dulaglutide (1.5 mg) between December 2023 and December 2024. All patients received concurrent hypocaloric dietary intervention (1,200–1,400 kcal/day). Clinical and biochemical parameters were assessed at baseline and 6 months, with additional weight assessments at 3 months. Multivariable analyses were performed to adjust for baseline differences.

Results

Both treatment groups demonstrated significant reductions in HbA1c, fasting blood glucose, liver enzymes, and lipid parameters at 6 months. After adjustment for baseline covariates, semaglutide was associated with greater weight reduction and LDL cholesterol improvement, whereas HbA1c reduction was comparable between groups. Mild increases in pancreatic enzyme levels were observed in both groups, with higher elevations in the dulaglutide group. Gastrointestinal adverse events were common but generally mild.

Conclusions

In this retrospective real-world cohort receiving structured dietary intervention, both semaglutide and dulaglutide improved glycemic control and metabolic parameters in newly diagnosed T2D. Semaglutide demonstrated superior weight reduction, while glycemic effects were similar. Baseline differences and the retrospective design should be considered when interpreting these findings.

Clinical trial number

Not applicable.

Introduction

Despite the availability of multiple treatment options for type 2 diabetes (T2D), optimal glycemic control remains difficult to achieve in a substantial proportion of patients. Current international guidelines emphasize individualized glycemic targets while minimizing hypoglycemia risk and weight gain when selecting antidiabetic therapies [1]. In recent years, glucagon-like peptide-1 receptor agonists (GLP-1 RAs) have emerged as effective therapeutic agents, as they stimulate insulin secretion and suppress glucagon release in a glucose-dependent manner, thereby improving glycemic control with a low risk of hypoglycemia [2].

Glucagon-like peptide-1 (GLP-1) is synthesized by enteroendocrine L cells in the intestine and by specialized neurons in the nucleus tractus solitarius [3]. It exerts a wide range of physiological effects through GLP-1 receptors expressed in multiple tissues, including the brain, kidneys, cardiovascular system, gastrointestinal tract, and lungs [4–7]. Following nutrient intake, GLP-1 enhances glucose-dependent insulin secretion and suppresses glucagon release, contributing to postprandial glucose regulation. In addition, GLP-1 delays gastric emptying, promotes satiety via central nervous system pathways, and protects pancreatic β cells from apoptosis [3, 4, 8, 9]. Collectively, these effects facilitate glucose uptake in peripheral tissues, reduce energy intake, and support weight loss [3, 4].

Insulin secretion in response to GLP-1 occurs primarily under conditions of elevated plasma glucose, which explains the low risk of hypoglycemia associated with GLP-1–based therapies. Experimental studies have demonstrated that chronic hyperglycemia may reduce β-cell expression of GLP-1 receptors, leading to diminished responsiveness to endogenous GLP-1 [7, 10]. Although circulating GLP-1 levels are generally preserved in individuals with T2D, impaired insulin secretion and relative resistance to GLP-1 action are frequently observed [7].

Native GLP-1 is rapidly degraded by dipeptidyl peptidase-4 (DPP-4), resulting in a short plasma half-life of approximately 1–2 min [11]. Consequently, long-acting GLP-1 RAs resistant to DPP-4 degradation, such as dulaglutide and semaglutide, have been developed for the treatment of T2D and obesity.

Dulaglutide is a long-acting GLP-1 RA consisting of two identical disulfide-linked chains, each containing a GLP-1 analog covalently fused to the Fc fragment of a modified human immunoglobulin G4 (IgG4). It shares approximately 90% sequence homology with native human GLP-1 and incorporates structural modifications that confer resistance to DPP-4–mediated degradation [12, 13]. Its large molecular size (~ 63 kDa) reduces renal clearance and prolongs its half-life to approximately 4–5 days, allowing once-weekly subcutaneous administration. Dulaglutide has been approved for the treatment of adults with T2D since 2014 [12].

Semaglutide is another long-acting GLP-1 RA with 94% structural homology to native GLP-1 [14]. Specific amino acid substitutions and a C18 fatty-acid side chain enhance albumin binding, reduce renal elimination, and confer resistance to DPP-4 degradation, extending its half-life to approximately one week and enabling once-weekly administration [14]. Semaglutide 1.0 mg administered subcutaneously once weekly was approved by the U.S. Food and Drug Administration (FDA) in 2017 for the treatment of T2D, with clinical use beginning in 2018 [15]. An oral formulation was subsequently approved in the United States in 2019 and in Europe in 2020 [15].

Although both agents are long-acting GLP-1 RAs, differences in molecular structure and pharmacokinetic properties may translate into clinically meaningful variations in glycemic and weight-related outcomes. In the SUSTAIN-7 randomized head-to-head trial, semaglutide demonstrated greater reductions in HbA1c and body weight compared with dulaglutide at comparable maintenance doses in patients with T2D inadequately controlled on metformin monotherapy [16]. However, because participants in that trial were receiving background oral therapy and had established diabetes, treatment responses may differ in newly diagnosed, treatment-naïve individuals.

In contrast, the present study evaluated semaglutide (1.0 mg) and dulaglutide (1.5 mg) in patients with newly diagnosed T2D and obesity (or overweight with comorbidities) who had not previously received any antidiabetic therapy. By focusing on a treatment-naïve population, this study aimed to provide a comprehensive real-world assessment of the comparative efficacy and tolerability of these two GLP-1 receptor agonists. To our knowledge, this is the first study conducted in Turkey addressing this clinical question.

Methods

This study represents a retrospective secondary analysis of routinely collected clinical data. The study was conducted retrospectively in a single-center private internal medicine clinic in Istanbul, Turkey. Patients who provided consent for the use of their medical data and who were treated with dulaglutide or semaglutide were included in the study. After obtaining informed consent from all participants, data regarding diagnoses, medications used, demographic characteristics, laboratory findings, and abdominal ultrasonography between December 2023 and December 2024 were retrieved from electronic medical records.

Participants

The study included 106 individuals aged 18–75 years with newly diagnosed type 2 diabetes (T2D), defined as fasting blood glucose (FBG) ≥ 125 mg/dL or hemoglobin A1c (HbA1c) ≥ 6.5% at first clinical presentation. Newly diagnosed T2D was defined as patients who had no prior history of diabetes diagnosis and had not previously received any oral antidiabetic therapy or GLP-1 receptor agonists. As HbA1c reflects glycemic status over the preceding 2–3 months, the diagnosis was considered to represent recent-onset diabetes.

Eligible participants had a body mass index (BMI) ≥ 30 kg/m² (obesity) or a BMI ≥ 27 kg/m² with obesity-related comorbidities, such as hypertension or hyperlipidemia (overweight with comorbidities). Individuals with prior use of oral antidiabetic drugs or GLP-1 receptor agonists, a history of bariatric surgery, or conditions associated with unintentional weight loss (including malignancy or chronic infection) were excluded.

Baseline demographic and clinical characteristics—including age, sex, body weight (kg), height (cm), BMI (kg/m²), waist circumference (WC, cm), and hip circumference (HC, cm) were recorded. Laboratory parameters assessed at baseline and at 6 months included liver function tests [aspartate aminotransferase (AST), alanine aminotransferase (ALT), and γ-glutamyltransferase (GGT)], lipid profile [total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides], HbA1c, and FBG. Pancreatic enzyme levels (amylase and lipase) were measured at baseline and at 3 and 6 months.

Liver ultrasonography was performed in all participants at baseline using a Toshiba Aplio 500 ultrasound system (Toshiba Medical Systems, Minato, Tokyo, Japan).

All included patients had completed at least 6 months of treatment. Changes in body weight were recorded at 3 and 6 months. Liver function tests, lipid parameters, FBG, and HbA1c were reassessed at 6 months, while pancreatic enzyme levels were evaluated at baseline, 3 months, and 6 months. Body weight measurements were obtained at each clinic visit using the same calibrated digital scale (Tanita F1BC-601PRO; Tanita Co., Tokyo, Japan).

Laboratory and biochemical measurements

All biochemical analyses were performed in an accredited clinical laboratory using a Cobas Integra 400 automated clinical chemistry analyzer (Roche Diagnostics GmbH, Mannheim, Germany).

HbA1c levels were measured using an immunoturbidimetric method based on Turbidimetric Inhibition Immunoassay (TINIA).

AST and ALT were measured using enzymatic UV kinetic methods in accordance with International Federation of Clinical Chemistry (IFCC) standards, with NADH consumption monitored at 340 nm.

Total cholesterol was measured using an enzymatic colorimetric method (CHOD-PAP). Triglycerides were measured using an enzymatic colorimetric method (GPO-PAP). HDL cholesterol was determined using a homogeneous enzymatic colorimetric assay with selective inhibition of non-HDL lipoproteins.

LDL cholesterol was primarily calculated using the Friedewald formula (LDL = Total Cholesterol – HDL – [Triglycerides/5], mg/dL). When triglyceride levels exceeded 400 mg/dL, LDL cholesterol was measured directly using a homogeneous enzymatic colorimetric method.

All laboratory measurements were performed in accordance with manufacturer instructions and internal quality control procedures.

Procedures

Study medications were administered subcutaneously into the abdominal region, typically by the patients themselves, independent of meal timing and at any time of day. Injections were performed on the same day each week throughout the study period.

For semaglutide, treatment was initiated at 0.25 mg once weekly and titrated every 4 weeks (0.25 mg → 0.5 mg → 1.0 mg) until the approved maintenance dose of 1.0 mg was achieved. Patients receiving dulaglutide were treated with a fixed dose of 1.5 mg once weekly, corresponding to the approved maintenance dose for glycemic control in T2D according to international guidelines and product labeling. This regimen reflected routine clinical practice during the study period. As this was a retrospective study, detailed information regarding potential initial dose adjustments or titration strategies was not consistently available in the medical records.

In addition to pharmacological therapy, all participants received standardized dietary counseling and were advised to follow a hypocaloric diet providing approximately 1,200–1,400 kcal per day for 24 weeks. Dietary recommendations were individualized according to baseline body weight and metabolic status and were reinforced during routine follow-up visits. However, adherence to the dietary intervention was not objectively monitored using food diaries or validated nutritional assessment tools, and therefore compliance could not be quantitatively assessed.

Statistical analysis

All statistical analyses were performed using SPSS version 27.0 (IBM Corp., Armonk, NY, USA).

Continuous variables were presented as mean ± standard deviation (SD) for normally distributed data and as median (minimum–maximum) for non-normally distributed variables. Categorical variables were expressed as frequencies and percentages.

Normality of distribution was assessed using skewness and kurtosis statistics. For between-group comparisons, the Student’s t-test was used for normally distributed variables and the Mann–Whitney U test for non-normally distributed variables. Within-group comparisons were conducted using the paired samples t-test or Wilcoxon signed-rank test as appropriate. For repeated measurements involving more than two time points, the Friedman test was applied. Categorical variables were compared using the chi-square test.

To quantify baseline imbalance between groups, standardized mean differences (SMD) were calculated. An SMD < 0.1 was considered negligible, 0.1–0.5 moderate, and > 0.5 large imbalance.

Multivariable linear regression analyses were performed to evaluate independent associations between treatment group and primary outcomes (% weight loss and HbA1c reduction at 6 months), adjusting for age, sex, baseline BMI, baseline HbA1c, and hepatosteatosis grade. Logistic regression analysis was conducted to assess the association between treatment group and achievement of ≥ 10% weight loss. Linear models were estimated using robust (HC3) standard errors.

All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant.

Effect sizes were calculated where appropriate. Cohen’s d was used to interpret standardized mean differences between groups (0.2 small, 0.5 moderate, 0.8 large). For repeated-measures analyses, eta-squared (η²) values were interpreted according to Cohen’s benchmarks (0.01 small, 0.06 moderate, 0.14 large) (Cohen, 1988; Richardson, 2011).

Results

Baseline characteristics

A total of 106 patients were included in the study, of whom 51 (48.1%) received dulaglutide and 55 (51.9%) received semaglutide. Baseline demographic and clinical characteristics are presented in Table 1.

Table 1.

Baseline Characteristics of the Study Population

Variable	Dulaglutide (n = 51)	Semaglutide (n = 55)	SMD
Age (years), mean ± SD	49.6 ± 11.2	42.7 ± 11.9	0.60
Male sex, n (%)	28 (54.9)	21 (38.2)	0.33
BMI (kg/m²), mean ± SD	35.6 ± 5.4	33.2 ± 4.5	0.49
HbA1c (%), mean ± SD	7.34 ± 0.89	6.86 ± 0.39	0.70
Hepatosteatosis, n (%)
– None	4 (7.8)	16 (29.1)	0.55
– Grade 1	14 (27.5)	17 (30.9)	0.08
– Grade 2	22 (43.1)	20 (36.4)	0.14
– Grade 3	11 (21.6)	2 (3.6)	0.54

BMI: body mass index; SMD: standardized mean difference

Baseline balance between groups was evaluated using SMD values. SMD < 0.1 was considered negligible, 0.1–0.3 small, 0.3–0.5 moderate, and > 0.5 indicative of substantial imbalance

Baseline balance between groups was evaluated using standardized mean differences (SMD). SMD values < 0.1 were considered negligible, 0.1–0.3 small, 0.3–0.5 moderate, and > 0.5 indicative of substantial imbalance. Given the observational design, SMDs were preferred over p-values for assessing baseline balance.

Patients in the dulaglutide group were older (mean 49.6 ± 11.2 vs. 42.7 ± 11.9 years; SMD = 0.60) and had a higher baseline BMI (35.6 ± 5.4 vs. 33.2 ± 4.5 kg/m²; SMD = 0.49). Baseline HbA1c levels were also higher in the dulaglutide group (7.34 ± 0.89 vs. 6.86 ± 0.39; SMD = 0.70), indicating a moderate-to-large imbalance in glycemic severity.

A moderate imbalance was observed in sex distribution (SMD = 0.33). Regarding hepatosteatosis grading, larger differences were observed in the “None” (SMD = 0.55) and “Grade 3” (SMD = 0.54) categories, whereas Grade 1 (SMD = 0.08) and Grade 2 (SMD = 0.14) were relatively balanced between groups.

These baseline imbalances were addressed in subsequent multivariable regression analyses. Although baseline heterogeneity was observed, the independent association of treatment with outcomes was confirmed in adjusted analyses.

Comparison of waist circumference, hip circumference, and weight loss

Waist circumference

Changes in waist circumference are presented in Table 2.

Table 2.

Changes in Waist Circumference, Hip Circumference, and Body Weight by Treatment Group

Variable	Dulaglutide (n = 51)	Semaglutide (n = 55)	p-value (between-group)		Cohen’s d
Waist circumference (cm)
Baseline, mean ± SD	107.0 ± 13.7	97.9 ± 10.8	0.001		0.74
6 months, mean ± SD	99.5 ± 12.9	85.8 ± 11.5	0.001		1.12
Change (Δ), mean ± SD	−7.5 ± 3.9	−12.1 ± 4.9	0.001		1.02
Hip circumference (cm)
Baseline, mean ± SD	113.7 ± 12.2	115.3 ± 10.8	0.464		−0.14
6 months, mean ± SD	107.0 ± 11.1	104.6 ± 9.9	0.249		0.23
Change (Δ), mean ± SD	−6.7 ± 3.2	−10.7 ± 4.7	0.001		0.99
Weight loss (kg)
3 months, median (range)	5 (0–10)	8 (4–20)	0.001		1.08
6 months, median (range)	7.5 (1–15)	12 (6–28)	0.001		1.31
Change (3–6 mo), median (range)	2 (0–5)	4 (1–13)	0.001		1.17

Values are presented as mean ± SD or median (min–max) as appropriate. Between-group comparisons were performed using Student’s t-test or Mann–Whitney U test, and within-group comparisons using paired samples t-test or Wilcoxon signed-rank test. Effect sizes for between-group comparisons were calculated using Cohen’s d and interpreted as 0.20 (small), 0.50 (moderate), and 0.80 (large)

At baseline, waist circumference was higher in the dulaglutide group (107.0 ± 13.7 cm) compared with the semaglutide group (97.9 ± 10.8 cm; p = 0.001).

After 6 months, waist circumference decreased significantly in both groups. The mean reduction was 7.53 ± 3.94 cm in the dulaglutide group and 12.05 ± 4.88 cm in the semaglutide group (both p < 0.001 vs. baseline).

Between-group comparison of change demonstrated a significantly greater reduction in the semaglutide group (p = 0.001), corresponding to a large effect size (Cohen’s d = 1.02).

Hip circumference

Changes in hip circumference are presented in Table 2.

Hip circumference decreased significantly from baseline to month 6 in both groups (p < 0.001 for both). The mean reduction was 6.71 ± 3.18 cm in the dulaglutide group and 10.71 ± 4.69 cm in the semaglutide group. Although baseline and month-6 between-group differences were small, the magnitude of reduction over time was significantly greater in the semaglutide group (p = 0.001), with a large effect size (d = 0.99) (Table 2).

Weight loss

Weight loss outcomes are shown in Table 2; Fig. 1.

Fig. 1.

Primary and secondary adjusted outcomes at 6 Months. Adjusted treatment effects of semaglutide versus dulaglutide at 6 months. Forest plot showing adjusted beta coefficients (β) with 95% confidence intervals for percentage weight loss and HbA1c reduction. Models were adjusted for age, sex, baseline BMI, baseline HbA1c, and hepatosteatosis grade. Odds ratio (OR) for achieving ≥ 10% weight loss at 6 months is presented as annotation

At 3 months, median weight loss was 5 kg (0–10) in the dulaglutide group and 8 kg (4–20) in the semaglutide group (p = 0.001).

At 6 months, weight loss increased to 7.5 kg (1–15) and 12 kg (6–28), respectively (p = 0.001). The additional weight reduction between months 3 and 6 was also greater in the semaglutide group (p = 0.001).

The magnitude of between-group differences corresponded to large effect sizes at both 3 and 6 months (d = 1.08 and d = 1.31, respectively), indicating a clinically meaningful advantage for semaglutide.

Adjusted analyses

To account for baseline imbalances between treatment groups, multivariable regression analyses were performed (Table 3). After adjustment for age, sex, baseline BMI, baseline HbA1c, and hepatosteatosis grade, semaglutide remained independently associated with greater percentage weight loss at 6 months (adjusted β = 6.16% points; 95% CI 4.51–7.82; p < 0.001; adjusted R² = 0.453).

Table 3.

Multivariable regression analyses for primary and secondary endpoints

Outcome	Model Type	Independent Variable	Adjusted Effect (95% CI)	p-value	Model Fit
% Weight Loss (6 months)	Linear regression	Semaglutide vs. Dulaglutide	β = 6.16 (4.51 to 7.82)	< 0.001	Adj. R² = 0.453
HbA1c Reduction (baseline–6 months)	Linear regression	Semaglutide vs. Dulaglutide	β = 0.29 (0.13 to 0.46)	< 0.001	Adj. R² = 0.382
≥ 10% Weight Loss (6 months)	Logistic regression	Semaglutide vs. Dulaglutide	OR = 14.84 (4.91 to 44.87)	< 0.001	Pseudo R² = 0.297

All models were adjusted for age, sex, baseline BMI, baseline HbA1c, and hepatosteatosis grade. Linear models used robust (HC3) standard errors

Semaglutide was also associated with a modest but statistically significant greater reduction in HbA1c (adjusted β = 0.29%-point; 95% CI 0.13–0.46; p < 0.001; adjusted R² = 0.382).

In supportive logistic regression analysis, semaglutide was associated with substantially higher odds of achieving ≥ 10% weight loss at 6 months (OR 14.84; 95% CI 4.91–44.87; p < 0.001) (Table 3; Fig. 1).

Comparison of blood parameters

Liver and lipid parameters

Changes in liver enzymes and lipid parameters are presented in Table 4.

Table 4.

Liver and Lipid Parameters

Variable	Dulaglutide (n = 51)	Semaglutide (n = 55)	p-value (between-group)
AST (U/L)
Baseline	22 (8–61)	20 (11–55)	0.766
6 months	17 (6.6–50)	17 (8–32)	0.590
Change (Δ)	−2 (− 27–5)	−4.5 (− 29–5)	0.209
ALT (U/L)
Baseline	27 (6.1–102)	22 (9–90)	0.224
6 months	19 (4.3–96.7)	16.2 (7–57)	0.076
Change (Δ)	−6 (− 72–23.7)	−5 (− 52.2–15)	0.617
GGT (U/L)
Baseline	37 (11–153)	24 (6–185)	0.024
6 months	25.2 (8–93)	18 (4–332)	0.012
Change (Δ)	−11 (− 62–21)	−5.7 (− 119–282)	0.157
Total Cholesterol (mg/dL)
Baseline	206.6 ± 49.9	210.1 ± 40.1	0.687
6 months	192.0 ± 42.5	187.5 ± 34.6	0.547
Change (Δ)	−14.6 ± 21.6	−21.4 ± 24.4	0.134
LDL (mg/dL)
Baseline	125.0 ± 36.2	139.4 ± 34.7	0.039
6 months	118.6 ± 35.5	123.2 ± 30.0	0.471
Change (Δ)	−6.4 ± 15.8	−16.1 ± 19.6	0.006
HDL (mg/dL)
Baseline	45.0 ± 14.8	49.4 ± 12.2	0.096
6 months	44.8 ± 14.4	48.3 ± 11.8	0.169
Change (Δ)	−0.3 ± 5.7	−1.2 ± 6.2	0.441
Triglycerides (mg/dL)
Baseline	173 (65–552)	131 (51–477)	0.001
6 months	153 (57–364)	92 (35–239)	0.001
Change (Δ)	−23 (− 265–68)	−36 (− 276–40)	0.112

Data are presented as mean ± SD or median (min–max), as appropriate. Between-group comparisons were performed using Student’s t-test or Mann–Whitney U test. AST: aspartate aminotransferase; ALT: alanine aminotransferase; GGT: gamma-glutamyl transferase; LDL: low-density lipoprotein; HDL: high-density lipoprotein

At 6 months, significant reductions in AST, ALT, and GGT levels were observed in both treatment groups. Although baseline GGT levels differed between groups, the between-group difference in change over time was not statistically significant.

Total cholesterol and LDL cholesterol levels decreased in both groups. The reduction in LDL cholesterol was significantly greater in the semaglutide group (p = 0.006). Triglyceride levels also declined significantly at 6 months in both groups, with a greater absolute reduction observed in the semaglutide group.

No significant changes in HDL cholesterol levels were observed in either group.

Glycemic and pancreatic enzyme parameters

Changes in glycemic and pancreatic enzyme parameters are presented in Table 5.

Table 5.

Glycemic and Pancreatic Enzyme Parameters

Variable	Dulaglutide	Semaglutide	p-value
HbA1c (%)
Baseline	7.0 (6.5–9.8)	6.8 (6.5–8.5)	0.020
6 months	6.1 (5.1–7.8)	5.7 (4.9–6.3)	0.001
Change (Δ)	−1.0 (− 2.6–−0.1)	−1.1 (− 2.6–−0.3)	0.074
Fasting Glucose (mg/dL)
Baseline	145 (105–291)	145 (110–204)	0.825
6 months	105 (85–171)	100 (80–133)	0.002
Change (Δ)	−32 (− 153–−9)	−46.7 (− 117–−13)	0.023
Amylase (U/L)
Baseline	65.8 ± 25.0	61.5 ± 23.8	0.373
3 months	70 (33–183)	65 (17–111)	0.045
6 months	76 (34–540)	65 (19.4–135)	0.005
Change (baseline–6 mo)	17 (− 33–474)	2 (− 58–56)	0.001
Lipase (U/L)
Baseline	38 (18–157)	32 (13–161)	0.020
3 months	45.6 (14–145)	35.2 (18–110)	0.015
6 months	50 (16–400)	35.6 (19–122)	0.005
Change (baseline–6 mo)	9.8 (− 46–364)	5 (− 39–54)	0.060

Data are presented as mean ± SD or median (min–max), as appropriate. Between-group comparisons were performed using Student’s t-test or Mann–Whitney U test. HbA1c: hemoglobin A1c

At 6 months, HbA1c levels decreased significantly in both groups, with mean reductions of approximately 1.0–1.5% points from baseline. Although baseline and 6-month HbA1c values differed between groups, the magnitude of HbA1c reduction was not significantly different (Table 5). The temporal pattern of HbA1c change is illustrated in Fig. 2.

Fig. 2.

HbA1c trajectory from baseline to 6 months in the dulaglutide and semaglutide groups. Change in HbA1c levels from baseline to 6 months by treatment group. Mean HbA1c values with standard error (SE) bars for dulaglutide and semaglutide groups at baseline and 6 months. Both groups demonstrated significant reductions over time, with modest between-group differences after adjustment

Fasting plasma glucose levels also declined significantly in both groups, with a greater absolute reduction observed in the semaglutide group.

During follow-up, modest increases in amylase and lipase levels were observed, with higher values detected in the dulaglutide group at 3 and 6 months. However, clinically significant pancreatic enzyme elevations were rare.

Adverse events

Adverse events are summarized in Table 6. The proportion of patients reporting no adverse events was higher in the semaglutide group (50.9%) compared with the dulaglutide group (25.5%).

Table 6.

Frequency of adverse events by treatment group

Adverse Event	Dulaglutide (n = 51)	Semaglutide (n = 55)
None	13 (25.5%)	28 (50.9%)
Diarrhea	3 (5.9%)	2 (3.6%)
Nausea	12 (23.5%)	15 (27.3%)
Constipation	15 (29.4%)	6 (10.9%)
Heartburn	4 (7.8%)	3 (5.5%)
Reflux	1 (2.0%)	1 (1.8%)
Abdominal pain	3 (5.9%)	4 (7.3%)
Pancreatitis	1 (2.0%)	0 (0%)

Values are presented as n (%)

Gastrointestinal symptoms were the most commonly reported adverse events in both groups. Nausea was observed in 23.5% of patients receiving dulaglutide and 27.3% of those receiving semaglutide. Constipation was more frequent in the dulaglutide group (29.4%) than in the semaglutide group (10.9%). Diarrhea occurred in 5.9% and 3.6% of patients in the dulaglutide and semaglutide groups, respectively. Other gastrointestinal complaints, including heartburn, reflux, and abdominal pain, were reported at relatively low and comparable frequencies between groups. One case of pancreatitis (2%) was observed in the dulaglutide group, while no cases were reported in the semaglutide group. The episode was clinically managed and was considered likely related to gallstone disease rather than direct drug toxicity. Overall, adverse events were predominantly mild to moderate and did not lead to treatment discontinuation in most cases.

Discussion

In this retrospective secondary analysis of routinely collected clinical data, we compared the efficacy and safety of semaglutide and dulaglutide in newly diagnosed, treatment-naïve patients with type 2 diabetes and excess body weight. To our knowledge, this is the first real-world study conducted in Turkey evaluating these two GLP-1 receptor agonists in a treatment-naïve population over a 6-month period.

Both semaglutide and dulaglutide achieved clinically meaningful reductions in HbA1c levels. Although baseline HbA1c values were higher in the dulaglutide group, the magnitude of HbA1c reduction over 6 months did not differ significantly between groups in unadjusted analyses. However, after adjustment for age, sex, baseline BMI, baseline HbA1c, and hepatosteatosis grade, semaglutide was independently associated with a modest but statistically additional reduction in HbA1c. These findings suggest that while glycemic efficacy is broadly comparable, small differences may persist after accounting for baseline heterogeneity. In contrast to the SUSTAIN-7 trial, which demonstrated superior HbA1c reduction with semaglutide compared to dulaglutide in patients receiving background metformin therapy [16], glycemic improvement in our cohort was more similar between groups. This discrepancy may relate to the inclusion of newly diagnosed, treatment-naïve individuals in our study, who may exhibit greater initial responsiveness to incretin-based therapy. Similarly, prior exposure to GLP-1 receptor agonists in other comparative studies may partly explain previously reported differences in glycemic efficacy [17].

Weight-related outcomes demonstrated clearer separation between treatments. Semaglutide was associated with significantly greater reductions in body weight, waist circumference, and hip circumference at both 3 and 6 months. Importantly, this association remained robust in multivariable analyses, with semaglutide independently associated with an additional 6.16%-point reduction in body weight at 6 months and markedly higher odds of achieving ≥ 10% weight loss. These findings are consistent with previously described pharmacodynamic differences between the two agents and may reflect differences in central appetite regulation and receptor binding characteristics. Nonetheless, given the non-randomized design, residual confounding cannot be entirely excluded.

Both treatments were associated with significant improvements in liver enzyme levels. Decreases in AST, ALT, and GGT were observed in both groups, consistent with previous evidence suggesting beneficial hepatic effects of GLP-1 receptor agonists beyond weight reduction [18, 19]. However, as follow-up ultrasonographic assessments were not systematically performed, hepatic outcomes were primarily inferred from biochemical markers.

Regarding lipid parameters, total cholesterol and triglyceride levels decreased in both groups. LDL cholesterol reduction was greater in the semaglutide group, although mean LDL levels remained above recommended targets for high-risk individuals with T2D. Current guidelines recommend LDL cholesterol levels below 70 mg/dL in high-risk patients and below 55 mg/dL in very high-risk individuals [1, 20]. In our cohort, mean LDL cholesterol levels remained above recommended target values, and triglyceride levels did not reach optimal therapeutic thresholds in either group. These findings are consistent with previous evidence suggesting that although GLP-1 receptor agonists improve lipid parameters, they are often insufficient to achieve guideline-recommended lipid targets without additional lipid-lowering therapy [21, 22]. Furthermore, the cardiovascular benefits of semaglutide demonstrated in the SUSTAIN-6 trial support its role as part of a broader cardiometabolic risk reduction strategy [23]. These observations emphasize the importance of integrating GLP-1 receptor agonist therapy into comprehensive cardiovascular risk management, including statin treatment when indicated. Pancreatic enzyme elevations were modest and generally less than two-fold. Although amylase and lipase levels were numerically higher in the dulaglutide group during follow-up, clinically significant pancreatitis was rare, with only one case observed, likely related to gallstone disease rather than direct pharmacologic toxicity. These findings are consistent with prior incretin-based therapy studies demonstrating small, transient enzyme elevations without clear clinical significance [24–26]. The fixed maintenance-dose regimen used in routine practice may have influenced tolerability; however, detailed titration data were not consistently available in the retrospective records.

Adverse events were predominantly gastrointestinal and generally mild to moderate. While constipation was more frequent in the dulaglutide group and nausea slightly more common with semaglutide, overall tolerability profiles were comparable. The higher proportion of patients without reported adverse events in the semaglutide group should be interpreted cautiously given the non-randomized and observational nature of the study.

Taken together, our findings suggest that in newly diagnosed, treatment-naïve patients with T2D, both semaglutide and dulaglutide provide effective glycemic control and metabolic benefits when combined with lifestyle intervention. Semaglutide appears to confer a greater effect on weight-related outcomes and LDL cholesterol reduction, even after adjustment for baseline differences. However, given the retrospective design and baseline heterogeneity between groups, these results should be interpreted as associations rather than definitive evidence of superiority.

Longer-term prospective and randomized studies are needed to confirm the durability of weight loss, clarify differential metabolic effects, and evaluate long-term cardiovascular and hepatic outcomes in treatment-naïve populations.

Limitations

Although this study provides clinically meaningful findings, several limitations should be acknowledged. First, the retrospective design precludes causal inference and is subject to inherent selection bias. Treatment allocation was not randomized and may have been influenced by physician judgment and baseline patient characteristics. Baseline differences between treatment groups, including demographic and metabolic parameters, may have influenced outcomes despite statistical adjustment, and residual confounding cannot be entirely excluded. The absence of randomization further limits the ability to attribute observed differences solely to treatment effects.

Second, comprehensive comorbidity indices (such as the Charlson comorbidity index), as well as structured data on smoking and alcohol consumption, were not systematically documented in the retrospective records and therefore could not be incorporated into the analysis. Dietary adherence was also not objectively quantified using validated assessment tools, and differential compliance between groups may have contributed to variations in weight-related outcomes. Furthermore, detailed information regarding dose titration strategies, particularly for dulaglutide, was not consistently available in the medical records, which may have influenced tolerability and safety findings.

Third, the study was conducted at a single center with a relatively modest sample size, potentially limiting generalizability. Although multiple metabolic parameters were assessed, statistical adjustment for multiple comparisons was applied to reduce the risk of type I error. Additionally, follow-up duration was limited to six months, and longer-term data are required to evaluate the durability of metabolic effects. While baseline liver ultrasonography was available for all participants, follow-up imaging assessments were not systematically performed, and hepatic outcomes were therefore primarily inferred from biochemical markers.

Future prospective, randomized studies with larger cohorts, longer follow-up periods, structured comorbidity assessment, and objective monitoring of lifestyle adherence are warranted to confirm and extend these findings.

Conclusions

In conclusion, both semaglutide and dulaglutide improved glycemic control and multiple metabolic parameters in newly diagnosed, treatment-naïve patients with type 2 diabetes accompanied by obesity or overweight with comorbidities. After adjustment for baseline differences between groups, semaglutide was associated with greater weight reduction and more pronounced LDL cholesterol lowering, whereas improvements in HbA1c were broadly comparable between treatments.

Both agents demonstrated favorable effects on liver enzyme levels and lipid profiles. Although pancreatic enzyme elevations were more frequently observed in the dulaglutide group, no clinically significant pancreatitis occurred during the study period. Differences in safety outcomes should be interpreted cautiously, as dosing regimens reflected routine clinical practice and detailed titration data were not consistently available in this retrospective analysis.

When combined with structured lifestyle intervention, GLP-1 receptor agonists represent effective therapeutic options in early T2D management. Early initiation in newly diagnosed patients may facilitate improved metabolic control and weight management, potentially contributing to long-term cardiometabolic risk reduction. However, given the retrospective design and 6-month follow-up, larger prospective studies with longer observation periods are required to confirm the durability and long-term clinical implications of these findings.

Author contributions

S.C. conceived and designed the study, collected and analyzed the data, and wrote and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All data generated or analyzed during this study are included in this published article.

Declarations

Institutional review board statement

The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Bakırköy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey (Approval number: 2024/374; Date: 25 December 2024)

Consent to publish

Not applicable.

Informed consent

Informed consent forms were obtained from all patients.

Competing interests

The authors declare no competing interests.