Effects of Evogliptin and Dapagliflozin on Bone and Glucose Metabolism in Postmenopausal Women with Type 2 Diabetes: A Multicenter Randomized Exploratory Clinical Trial

Abstract

Background

Postmenopausal women with type 2 diabetes mellitus (T2DM) are at increased risk of fragility fractures, emphasizing the importance of selecting antidiabetic agents that support glycemic control and bone health. Most dipeptidyl peptidase-4 inhibitors have shown neutral or favorable effects on bone metabolism; however, data specifically on evogliptin remain limited. This study evaluated the effects of evogliptin and dapagliflozin on bone metabolism and glycemic parameters in postmenopausal women with T2DM.

Methods

This multicenter, randomized clinical trial enrolled postmenopausal women with T2DM assigned to receive evogliptin or dapagliflozin for 48 weeks. The primary outcome was change in bone mineral density (BMD). Glycemic indices were the secondary endpoints.

Results

Sixty participants were included in the final analysis, with 30 in each group. After 48 weeks, the change in BMD at the L1–L4 spine and total hip did not differ significantly between groups. Femoral neck BMD decreased significantly in the dapagliflozin group compared with the evogliptin group, although the absolute between-group difference was small (0.00 ± 0.02 vs −0.02 ± 0.04 g/cm², p = 0.0402). After adjustment for baseline femoral neck BMD and 25OHD, the between-group difference was attenuated (p = 0.0763). The body mass index (BMI) decreased more in the dapagliflozin group than in the evogliptin group. In analyses of factors associated with femoral neck BMD change, no statistically significant predictors were identified, although decreasing BMI showed a trend toward association. Glycemic indices did not differ significantly between the groups.

Conclusion

In postmenopausal women with T2DM, BMD remained stable with evogliptin, whereas dapagliflozin was associated with a small reduction in femoral neck BMD, although this finding was weakened after adjustment. These findings provide preliminary evidence regarding the skeletal effects of evogliptin and dapagliflozin and warrant confirmation in larger studies.

Clinical Trial Registration

ClinicalTrials.gov (NCT04706637).

Introduction

Diabetes mellitus is a representative chronic disease worldwide, with 90% of cases classified as type 2 diabetes mellitus (T2DM), most of which are treated with oral hypoglycemic agents. Diabetes is associated with an increased risk of fractures, independent of bone mineral density (BMD), and is considered a secondary cause of osteoporosis, thereby raising concerns about bone health in patients with diabetes.^1–3 Hyperglycemia, oxidative stress, and advanced glycation end products (AGEs) are pathophysiological mechanisms underlying bone fragility in diabetes mellitus.^1,2 Additionally, the increased risk of falls due to hypoglycemia induced by antidiabetic medications, including the effects of drugs such as thiazolidinediones on bone metabolism,^4,5 is a factor contributing to the elevated fracture risk in patients with diabetes.

Menopause is accompanied by estrogen deficiency, which accelerates bone resorption and increases insulin resistance. Consequently, postmenopausal women with type 2 diabetes represent a particularly vulnerable population with an increased risk of fracture despite having normal bone mineral density, due to complex mechanisms such as deterioration of bone quality, low bone turnover, and poor glycemic control.^6,7 Therefore, in patients with diabetes with a high risk of osteoporosis, such as those experiencing menopause,⁸ treatment plans should consider not only proper glycemic control but also bone metabolism for comprehensive management.

Dipeptidyl peptidase-4 inhibitors (DPP-4i) are representative medications for T2DM management and are widely used because of their neutral effects on body weight and cardiovascular outcomes, including their low risk of hypoglycemia.⁹ Various preclinical^10–13 and clinical^14–16 studies have investigated the relationship between DPP-4i and bone metabolism. While many of these studies suggest positive^10–14,17 effects on bone metabolism by increasing the bioavailability of incretins, which stimulate osteoblast-mediated bone formation and suppress bone resorption, or neutral^15,16,18 effects by maintaining BMD without the adverse skeletal effects often observed with other glucose-lowering agents, further investigation is needed, particularly for individual agents.

Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are increasingly used as recommended medications owing to their cardiovascular and renal benefits.⁹ SGLT2i regulate blood glucose by acting on the proximal tubules of the kidney to increase urinary glucose excretion. This process may influence calcium and phosphate homeostasis, with the potential to affect bone metabolism.¹⁹ Therefore, various studies have investigated the relationship between SGLT2i and bone metabolism. However, findings remain inconclusive regarding whether SGLT2i have a negative^20–22 or neutral^23–27 effect on fracture risk.

To our knowledge, no prospective randomized clinical study has examined the effects of evogliptin on BMD and bone turnover markers (BTMs) in postmenopausal women. Therefore, we conducted this study to evaluate the effects of evogliptin and dapagliflozin on bone metabolism, with BMD as the primary outcome, and to compare their glucose-lowering effects using diabetes-related markers as the secondary outcomes in postmenopausal women with T2DM, a population particularly vulnerable to impaired bone health.

Subjects and Methods

Study Design and Participants

This multicenter, randomized, exploratory clinical trial was conducted in South Korea and reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. Participant enrollment was conducted from September 2021 through May 2024, and each participant was followed for 48 weeks. The study included postmenopausal women aged 40 to 80 years with T2DM who were either not taking antidiabetic medications, taking metformin alone, or on a combination therapy of metformin and sulfonylureas, with HbA1c levels between 7.0% and 9.0%, indicating inadequate glycemic control. Participants were required to have BMD measurements at the lumbar spine, femoral neck, and total hip with a T-score ≥ −2.4. Patients who had taken any antidiabetic medication other than metformin or metformin-plus-sulfonylurea therapy within 4 weeks, Glucagon-like Peptide-1 Receptor Agonist (GLP-1 RA) within 12 weeks, or thiazolidinedione within 24 weeks before screening were excluded. Patients with hypersensitivity to DPP4 or SGLT2i were excluded from the study. Patients who had used bisphosphonates or denosumab within 5 years, female hormones, or Selective Estrogen Receptor Modulator (SERM) agents within 12 months were also ineligible. Treatments affecting bone and mineral metabolism, such as long-term corticosteroid or diuretic use, also led to exclusion. Additionally, patients with type 1 diabetes, diabetic ketoacidosis, a history of malignancy within 5 years, or genetic disorders affecting glucose metabolism were excluded. Those with recent lab results showing aspartate aminotransferase (AST)/alanine aminotransferase (ALT) levels exceeding three times the upper normal limit or an estimated glomerular filtration rate (eGFR) below 60 mL/min/1.73 m² were also ineligible.

The trial was conducted in accordance with the Declaration of Helsinki, Korean Good Clinical Practice, and relevant regulations, and was approved by the local institutional review boards of all participating centers. The first institutional review board approval was obtained on September 2, 2020, before study initiation. The study was prospectively registered at ClinicalTrials.gov (NCT04706637) on January 10, 2021, before enrollment of the first participant. All the participants provided written informed consent before participating in the study.

Procedures

Once the participants voluntarily agreed to participate in the clinical trial, they were evaluated to confirm their eligibility based on the inclusion and exclusion criteria. Participants deemed suitable were randomly assigned to one of two groups: the experimental group receiving evogliptin 5 mg or the control group receiving dapagliflozin 10 mg, each administered at its approved standard therapeutic dose for the treatment of type 2 diabetes mellitus. Randomization was performed by an independent statistician not involved in participant recruitment or allocation, using block randomization with a block size of 4, stratified by clinical site, at a 1:1 ratio. The randomization list was provided to the drug packaging manager to ensure that investigators had no access to the allocation sequence prior to enrollment. As this was an open-label trial, no blinding was applied to participants, treating physicians, or outcome analysts.

Among the study participants, those who had been taking calcium or vitamin D supplements discontinued these agents upon enrollment. At the time of randomization, all participants were initiated on a standardized calcium–vitamin D supplement (Dicamax-D, containing 1000 IU cholecalciferol and 250 mg calcium carbonate per tablet), which was administered concomitantly with the study medication.

Study Outcomes

The outcome measures in this study were categorized into bone metabolism- and glycemic control-related parameters. The primary endpoint was the between-group difference in change from baseline to 48 weeks in BMD at the lumbar spine, femoral neck, and total hip. BTMs (CTX, P1NP), bone metabolism parameters (25OHD, PTH-Intact, FGF23, 24-hour urine Ca, P, and Cr) at 12 and 48 weeks, and thoracolumbar spine X-ray (T-L spine X-ray) performed at 48 weeks were included as subsidiary variables. The other secondary endpoints included a comparison of changes in diabetes-related markers such as HbA1c, fasting blood sugar (FBS), insulin, C-peptide, AGEs, and continuous glucose monitoring (CGM) data at 12 and 48 weeks. Body impedance analysis (BIA) parameters, including body mass index (BMI), weight, fat mass, intracellular water, extracellular water, and total body water, were measured at week 48. Safety assessments included the monitoring of adverse events, vital signs, and laboratory parameters in accordance with the protocol.

Statistical Analyses

The sample size calculation was based on the percentage change in total hip BMD at 12 months. Based on the variability reported by Ha et al²⁸, the standard deviation was assumed to be 1.8 percentage points, and a target between-group difference of 1.5 percentage points was selected for this exploratory study. At a two-sided significance level of 0.05 and 80% power, 24 participants per group were required. Allowing for an anticipated dropout rate of 20%, the target enrollment was set at 30 participants per group, for a total of 60 participants.

When stratification issues occurred in the clinical trial center, additional analysis was performed using the Mantel–Haenszel test, considering the clinical trial center as a stratification factor.

The primary analysis was based on the full analysis set (FAS), defined as all participants who received at least one dose of study medication and completed at least one post-baseline efficacy assessment. One participant in the evogliptin group who did not initiate study medication was therefore excluded from all analyses. Missing data for the primary and secondary efficacy outcomes were handled using the last observation carried forward (LOCF) method in the FAS population.

Continuous variables were summarized as means and standard deviations, and categorical variables as frequencies and percentages. Normality was assessed using the Shapiro–Wilk test. Between-group comparisons of continuous variables, including changes from baseline, were performed using two-sample t-tests or Wilcoxon rank-sum tests, as appropriate. Within-group changes were evaluated using paired t-tests or Wilcoxon signed-rank tests. Categorical variables were compared using chi-squared tests or Fisher’s exact tests, as appropriate.

Changes in BMD, the primary endpoint, were evaluated at each measurement site using the methods described above. BTMs and other bone metabolism parameters were analyzed using the same methods, and secondary outcomes, including diabetes-related markers, were evaluated in a similar manner.

As a post hoc supportive analysis, femoral neck BMD at week 48 was compared between the treatment groups using ANCOVA, with treatment group as the main effect and baseline femoral neck BMD and baseline 25OHD as covariates. Separately, rank-based robust regression analysis was performed as an exploratory supplementary analysis to assess the factors associated with changes in femoral neck BMD. The model included treatment group, change in BMI, baseline 25OHD, and age as covariates.

For safety assessment, descriptive statistics, including the number of participants, incidence rates, and number of adverse events, were presented separately for each treatment group based on the adverse events that occurred during the clinical trial. Adverse events were coded using the Medical Dictionary for Regulatory Activities (MedDRA) according to the System Organ Class (SOC) and Preferred Term (PT). The number of participants experiencing coded adverse events, incidence rates, and the total number of events are presented. Vital signs and basic laboratory test results observed at each time point were summarized using descriptive statistics, and the incidence rates and counts of abnormal findings were reported.

A two-sided P value <0.05 was considered statistically significant. No formal adjustment for multiple comparisons was performed. All statistical procedures were conducted using R software version 4.2.3.

Results

Sixty-eight participants underwent screening, of whom seven were excluded, leaving 61 participants enrolled in the study (enrollment period: September 2021 – May 2024). Of these, 31 were assigned to the evogliptin group and 30 to the dapagliflozin group. However, one participant in the evogliptin group did not initiate the study medication and was excluded from all analyses. Therefore, all evaluations, from baseline characteristics to safety assessments, were conducted on the 30 participants in the evogliptin group. Among the 60 participants included in the analysis, 50 (83.3%) completed the study without protocol violations, whereas 10 (16.7%) discontinued the study early (Figure 1).

Figure 1.

Patient flow chart.

All participants in this study were female. The mean age was 66.40 ± 6.09 years in the evogliptin group and 63.70 ± 6.58 years in the dapagliflozin group. The mean age at menopause was 49.27 ± 6.07 years and 49.00 ± 4.67 years in each group, respectively (Table 1). There were no significant differences in the BMD between the groups for the lumbar spine, femoral neck, or total hip. Abnormal T–L spine findings at baseline were observed in 26 participants (89.7%) in the evogliptin group and 24 participants (80.0%) in the dapagliflozin group. Most of these abnormalities were attributable to degenerative changes on T–L spine radiographs.

Table 1.

Baseline Characteristics of the Study Participants

	Evogliptin (n = 30)	Dapagliflozin (n = 30)	P value
Age (years)	66.40 ± 6.09	63.70 ± 6.58	0.1043
Menopause age (years)	49.27 ± 6.07	49.00 ± 4.67	0.8495
BMD (g/cm2)
Lumbar spine	1.05 ± 0.15	1.04 ± 0.14	0.7862
Femoral neck	0.82 ± 0.11	0.83 ± 0.09	0.6391
Total hip	0.92 ± 0.12	0.92 ± 0.13	0.9916
T-score
Lumbar spine	−0.67 ± 1.21	−0.76 ± 1.13	0.7856
Femoral neck	−1.05 ± 0.92	−0.92 ± 0.74	0.5589
Total hip	−0.13 ± 0.97	−0.19 ± 0.89	0.8041
BTMs (ng/mL)
P1NP (ng/mL)	40.05 ± 12.23	44.70 ± 16.45	0.2183
CTX (ng/mL)	0.37 ± 0.16	0.38 ± 0.14	0.6720
FGF23 (pg/mL)	43.73 ± 12.92	43.90 ± 17.10	0.9662
25OHD (ng/mL)	40.70 ± 17.15	29.33 ± 12.03	0.0043
PTH-Intact (pg/mL)	29.13 ± 11.55	30.90 ± 11.40	0.5533
24-hour urine
Ca (g/day)	0.18 ± 0.11	0.18 ± 0.13	0.9981
P (mg/day)	535.53 ± 242.7	480.19 ± 284.4	0.4386
Cr (mg/day)	757.64 ± 244.9	713.23 ± 288.9	0.5391
HbA1c (%)	7.46 ± 0.47	7.34 ± 0.37	0.3080
FBS (mg/dL)	157.72 ± 30.70	139.83 ± 16.85	0.0180
Insulin (μU/mL)	9.55 ± 7.28	10.39 ± 6.83	0.6517
c-peptide (ng/mL)	3.22 ± 2.55	2.63 ± 1.01	0.2466
CGM
Average glucose (mg/dL)	181.40 ± 27.02	179.20 ± 15.22	0.8779
Standard Deviation	49.20 ± 15.01	77.20 ± 60.21	0.3642
GMI (%)	7.65 ± 0.65	7.60 ± 0.36	0.8810
BIA
BMI (kg/m2)	25.21 ± 3.66	25.71 ± 3.15	0.5809
Lean body mass (kg)	39.27 ± 5.18	37.55 ± 4.30	0.1693
Fat mass (kg)	22.76 ± 6.44	23.60 ± 5.54	0.5923
Intracellular water (L)	17.36 ± 2.11	16.89 ± 2.00	0.3855
Extracellular water (L)	11.06 ± 1.26	10.73 ± 1.19	0.3164
Total body water (L)	28.42 ± 3.35	27.63 ± 3.17	0.3556
T-L spine			0.4716
Normal	3 (10.34)	6 (20.00)
Abnormal	26 (89.66)	24 (80.00)
History of fractures	2 (6.67)	3 (10.00)	1.0000

Notes: Variables are presented as the mean ± SD or n (%).

Abbreviations: BMD, bone mineral density; BTMs, Bone Turnover Markers; P1NP, amino-terminal propeptide of procollagen type 1; CTX, carboxy-terminal telopeptide of type 1 collagen; FGF23, fibroblast growth factor 23; 25OHD, 25-hydroxy-vitamin D; PTH-Intact, intact parathyroid hormone ; HbA1c, glycated hemoglobin; FBS, fasting blood sugar; CGM, continuous glucose monitoring; GMI, glucose management indicator; BIA, bioelectrical impedance analysis; BMI, body mass index; T-L spine, thoracolumbar spine.

The analysis of the primary study outcomes, examining between-group differences in BMD from baseline to 48 weeks, showed no significant differences in the lumbar spine and total hip measurements. However, at the femoral neck, the evogliptin group demonstrated a change of 0.00 ± 0.02 g/cm² (p = 0.9620), while the dapagliflozin group showed a change of −0.02 ± 0.04 g/cm² (p = 0.0275), revealing a statistically significant between-group difference (p = 0.0402) (Table 2 and Figure 2A). After adjustment for baseline femoral neck BMD and baseline 25OHD in the supportive ANCOVA, the between-group difference remained in favor of evogliptin, although it did not reach statistical significance (adjusted difference, 0.0188 g/cm²; 95% CI, −0.0021 to 0.0396; p = 0.0763).

Table 2.

Change from Baseline in BMD

			Evogliptin (n = 30)		Dapagliflozin (n = 30)		P value2
			mean ± SD	P value1	mean ± SD	P value1
BMD (g/cm2)	Lumbar spine	Baseline	1.05 ± 0.15		1.04 ± 0.14		0.7862
		48 weeks	1.05 ± 0.16		1.02 ± 0.17		0.4994
		Difference	0.00 ± 0.05	0.6653	−0.01 ± 0.06	0.1806	0.2043
	Femoral neck	Baseline	0.82 ± 0.11		0.83 ± 0.09		0.6391
		48 weeks	0.82 ± 0.11		0.81 ± 0.09		0.8503
		Difference	0.00 ± 0.02	0.9620	−0.02 ± 0.04	0.0275	0.0402
	Total hip	Baseline	0.92 ± 0.12		0.92 ± 0.13		0.9916
		48 weeks	0.92 ± 0.12		0.93 ± 0.11		0.9223
		Difference	0.00 ± 0.05	0.9718	0.00 ± 0.07	0.8210	0.8721

Notes: Variables are presented as the mean ± SD. P value¹ indicates the within-group comparison of change from baseline, and P value² indicates the between-group comparison. Within-group changes were evaluated using paired t-tests or signed-rank tests, as appropriate, and between-group comparisons were performed using two-sample t-tests or Wilcoxon rank-sum tests, as appropriate. A two-sided P value <0.05 was considered statistically significant.

Figure 2.

Comparison of percent changes from baseline in (A) BMD, (B) BTMs, and (C) BIA after 48 weeks of treatment. Statistical significance is indicated as follows: *p < 0.05, **p < 0.01.

Abbreviation: ns, not significant.

In the exploratory regression analysis performed to assess factors associated with changes in femoral neck BMD, no statistically significant associations were identified (Supplementary Table 2).

For BTMs and other bone metabolism parameters, between-group comparisons were conducted at baseline, 12 weeks, and 48 weeks, including for changes from baseline to 12 weeks and baseline to 48 weeks. No significant between-group differences were observed for CTX or P1NP either at individual time points or in changes from baseline values. The percentage change in P1NP levels from baseline to 48 weeks was 4.8% in the evogliptin group and 3.6% in the dapagliflozin group. For CTX, the percentage change from baseline to 48 weeks was 8.1% in the evogliptin group and 23.7% in the dapagliflozin group, showing some numerical difference, but without reaching statistical significance (Table 3 and Figure 2B). Similarly, no significant between-group differences were detected in PTH-intact, FGF23, or 24-hour urinary calcium, phosphorus, and creatinine levels.

Table 3.

Change from Baseline in BTMs and Bone Metabolic Parameters

			Evogliptin (n = 30)		Dapagliflozin (n = 30)
			mean ± SD	P value1	mean ± SD	P value1	P value2
BTMs	P1NP (ng/mL)	Baseline	40.05 ± 12.23		44.70 ± 16.45		0.2183
		12 weeks	38.46 ± 15.09		45.41 ± 18.99		0.1216
		48 weeks	41.94 ± 16.95		46.29 ± 19.42		0.3595
		Δ (Baseline to 12 weeks)	−1.59 ± 9.75	0.3793	0.71 ± 9.42	0.6828	0.3568
		Δ (Baseline to 48 weeks)	1.90 ± 12.59	0.4160	1.59 ± 19.99	0.6670	0.9430
	CTX (ng/mL)	Baseline	0.37 ± 0.16		0.38 ± 0.14		0.6720
		12 weeks	0.39 ± 0.21		0.46 ± 0.22		0.2097
		48 weeks	0.40 ± 0.19		0.47 ± 0.22		0.1705
		Δ (Baseline to 12 weeks)	0.02 ± 0.12	0.3466	0.08 ± 0.16	0.0170	0.1589
		Δ (Baseline to 48 weeks)	0.03 ± 0.16	0.3145	0.09 ± 0.18	0.0105	0.1952
Bone Metabolic Parameter	FGF23 (pg/mL)	Baseline	43.73 ± 12.92		43.90 ± 17.10		0.9662
		12 weeks	46.07 ± 12.81		43.43 ± 13.35		0.4389
		48 weeks	48.87 ± 19.46		45.27 ± 19.57		0.4778
		Δ (Baseline to 12 weeks)	2.33 ± 12.76	0.3249	−0.47 ± 17.21	0.8829	0.4769
		Δ (Baseline to 48 weeks)	5.13 ± 17.62	0.1214	1.37 ± 23.41	0.7515	0.4842
	25OHD (ng/mL)	Baseline	40.70 ± 17.15		29.33 ± 12.03		0.0043
		12 weeks	42.35 ± 10.80		38.13 ± 14.38		0.2038
		48 weeks	47.65 ± 14.85		39.98 ± 12.88		0.0368
		Δ (Baseline to 12 weeks)	1.66 ± 13.71	0.5132	8.81 ± 12.01	0.0004	0.0358
		Δ (Baseline to 48 weeks)	6.96 ± 16.82	0.0311	10.66 ± 11.65	<0.0001	0.3261
	PTH-Intact (pg/mL)	Baseline	29.13 ± 11.55		30.90 ± 11.40		0.5533
		12 weeks	30.49 ± 11.73		30.23 ± 8.32		0.9215
		48 weeks	29.00 ± 11.17		32.50 ± 10.59		0.2180
		Δ (Baseline to 12 weeks)	1.36 ± 9.93	0.4594	−0.67 ± 10.69	0.7352	0.4500
		Δ (Baseline to 48 weeks)	−0.13 ± 8.39	0.9312	1.60 ± 9.59	0.3684	0.4591
24-hour urine	Ca (g/day)	Baseline	0.18 ± 0.11		0.18 ± 0.13		0.9981
		12 weeks	0.19 ± 0.10		0.19 ± 0.13		0.9453
		Δ (Baseline to 12 weeks)	0.02 ± 0.06	0.1740	0.02 ± 0.09	0.2335	0.9109
	P (mg/day)	Baseline	535.53 ± 242.76		480.19 ± 284.42		0.4386
		12 weeks	514.11 ± 210.67		505.42 ± 266.95		0.8935
		Δ (Baseline to 12 weeks)	−21.42 ± 203.30	0.5887	25.23 ± 186.06	0.4714	0.3740
	Cr (mg/day)	Baseline	757.64 ± 244.92		713.23 ± 288.97		0.5391
		12 weeks	779.89 ± 197.47		689.74 ± 254.10		0.1461
		Δ (Baseline to 12 weeks)	22.25 ± 173.36	0.5108	−23.48 ± 188.80	0.5085	0.3504

Notes: Variables are presented as the mean ± SD or n (%). P value¹ indicates the within-group comparison of change from baseline, and P value² indicates the between-group comparison. Within-group changes were evaluated using paired t-tests or signed-rank tests, as appropriate, and between-group comparisons were performed using two-sample t-tests or Wilcoxon rank-sum tests, as appropriate. A two-sided P value <0.05 was considered statistically significant.

Abbreviations: BTMs, Bone Turnover Markers; P1NP, amino-terminal propeptide of procollagen type 1; CTX, carboxy-terminal telopeptide of type 1 collagen; FGF23, fibroblast growth factor 23; 25OHD, 25-hydroxy-vitamin D; PTH-Intact, intact parathyroid hormone ; T-L spine, thoracolumbar spine.

However, baseline 25-hydroxyvitamin D (25OHD) levels were significantly lower in the dapagliflozin group than in the evogliptin group (evogliptin group: 40.70 ± 17.15 ng/mL, dapagliflozin group: 29.33 ± 12.03 ng/mL; p = 0.0043). Although 25OHD levels increased in both groups, the change from baseline did not differ significantly between the groups, despite a numerically greater increase in the dapagliflozin group. At week 48, 25OHD levels remained significantly lower in the dapagliflozin group compared with the evogliptin group (evogliptin group: 47.65 ± 14.85 ng/mL, dapagliflozin group: 39.98 ± 12.88 ng/mL; p = 0.0368). Nevertheless, the mean 25OHD concentrations in both groups remained within a sufficient range throughout the study period (Table 3).

In the analysis of secondary outcomes, diabetes-related parameters (HbA1c, FBS, insulin, C-peptide, and CGM) were compared between the groups at baseline, 12 weeks, and 48 weeks, including for changes from baseline to weeks 12 and 48. No statistically significant between-group differences were observed for most diabetes-related parameters, either at individual time points or in the changes from baseline values (Table 4).

Table 4.

Change from Baseline in Diabetes Mellitus-Related Parameters

		Evogliptin (n = 30)		Dapagliflozin (n = 29)		P value2
		mean ± SD	P value1	mean ± SD	P value1
HbA1c (%)	Baseline	7.46 ± 0.47		7.34 ± 0.37		0.3080
	12 weeks	7.24 ± 0.98		7.28 ± 1.02		0.8698
	24 weeks	6.98 ± 0.75		7.23 ± 0.80		0.2175
	36 weeks	7.00 ± 0.79		7.11 ± 0.70		0.5669
	48 weeks	6.99 ± 0.73		7.17 ± 0.68		0.3226
	Δ (Baseline to 12 weeks)	−0.22 ± 0.85	0.1809	−0.06 ± 1.00	0.7396	0.5197
	Δ (Baseline to 24 weeks)	−0.48 ± 0.65	0.0005	−0.11 ± 0.88	0.5183	0.0754
	Δ (Baseline to 36 weeks)	−0.46 ± 0.75	0.0030	−0.23 ± 0.77	0.1161	0.2573
	Δ (Baseline to 48 weeks)	−0.47 ± 0.71	0.0015	−0.17 ± 0.72	0.2156	0.1146
FBS (mg/dL)	Baseline	157.72 ± 30.70		139.83 ± 16.85		0.0086
	12 weeks	143.10 ± 30.32		133.76 ± 24.59		0.2026
	24 weeks	138.48 ± 27.22		128.62 ± 18.52		0.1131
	36 weeks	140.03 ± 30.74		130.14 ± 23.55		0.1742
	48 weeks	137.10 ± 20.35		132.07 ± 22.69		0.3775
	Δ (Baseline to 12 weeks)	−14.62 ± 38.50	0.0504	−6.07 ± 27.75	0.2488	0.3360
	Δ (Baseline to 24 weeks)	−19.24 ± 29.95	0.0017	−11.21 ± 24.32	0.0193	0.2668
	Δ (Baseline to 36 weeks)	−17.69 ± 38.78	0.0205	−9.69 ± 26.39	0.0579	0.3629
	Δ (Baseline to 48 weeks)	−20.62 ± 30.69	0.0012	−7.76 ± 26.00	0.1193	0.0906
Insulin (μU/mL)	Baseline	9.55 ± 7.28		10.39 ± 6.83		0.6517
	12 weeks	8.21 ± 4.08		7.13 ± 3.88		0.2991
	48 weeks	7.94 ± 3.31		7.61 ± 3.50		0.7134
	Δ (Baseline to 12 weeks)	−1.34 ± 5.42	0.1861	−3.26 ± 7.00	0.0181	0.2423
	Δ (Baseline to 48 weeks)	−1.62 ± 6.42	0.1787	−2.78 ± 6.81	0.0366	0.5027
c-peptide (ng/mL)	Baseline	3.22 ± 2.55		2.63 ± 1.01		0.2466
	12 weeks	2.39 ± 0.80		2.28 ± 0.96		0.6335
	48 weeks	2.28 ± 0.76		2.17 ± 0.70		0.5665
	Δ (Baseline to 12 weeks)	−0.83 ± 2.32	0.0598	−0.35 ± 0.98	0.0653	0.3033
	Δ (Baseline to 48 weeks)	−0.95 ± 2.35	0.0355	−0.46 ± 0.87	0.0077	0.2997
AGE (ng/mL)	Baseline	115.97 ± 31.39		113.24 ± 26.07		0.7187
	12 weeks	100.58 ± 38.29		94.04 ± 29.96		0.4692
	48 weeks	69.54 ± 35.38		73.06 ± 39.05		0.7182
	Δ (Baseline to 12 weeks)	−15.39 ± 33.32	0.0171	−19.21 ± 25.34	0.0003	0.6236
	Δ (Baseline to 48 weeks)	−46.43 ± 33.34	<0.0001	−40.19 ± 30.28	<0.0001	0.4552
CGM
Average glucose (mg/dL)	Baseline	181.40 ± 27.02		179.20 ± 15.22		0.8779
	12 weeks	176.60 ± 35.47		180.40 ± 24.00		0.8477
	Δ (Baseline to 12 weeks)	−4.80 ± 9.42	0.3181	1.20 ± 13.48	0.8519	0.4382
Standard Deviation	Baseline	49.20 ± 15.01		77.20 ± 60.21		0.3642
	12 weeks	42.40 ± 15.71		63.00 ± 24.88		0.1561
	Δ (Baseline to 12 weeks)	−6.80 ± 8.76	0.1575	−14.20 ± 39.35	0.4649	0.7007
GMI (%)	Baseline	7.65 ± 0.65		7.60 ± 0.36		0.8810
	12 weeks	7.53 ± 0.85		7.63 ± 0.57		0.8479
	Δ (Baseline to 12 weeks)	−0.11 ± 0.23	0.3285	0.03 ± 0.32	0.8506	0.4444

Notes: Variables are presented as the mean ± SD. P value¹ indicates the within-group comparison of change from baseline, and P value² indicates the between-group comparison. Within-group changes were evaluated using paired t-tests or signed-rank tests, as appropriate, and between-group comparisons were performed using two-sample t-tests or Wilcoxon rank-sum tests, as appropriate. A two-sided P value <0.05 was considered statistically significant.

Abbreviations: HbA1c, glycated hemoglobin; FBS, fasting blood sugar; AGE, advanced glycation end products; CGM, continuous glucose monitoring; GMI, glucose management indicator.

In the BIA, the evogliptin group showed a significant change in BMI with a mean difference of −0.42 ± 0.95 kg/m² (p = 0.0258). The dapagliflozin group demonstrated a more pronounced BMI reduction of −1.14 ± 1.07 kg/m² (p < 0.0001), with a statistically significant between-group difference in the magnitude of change (p = 0.0085). Fat mass measurements also revealed divergent patterns between groups: the evogliptin group showed minimal change (−0.18 ± 3.87 kg, p = 0.8087), while the dapagliflozin group demonstrated a significant reduction (−1.97 ± 2.47 kg, p = 0.0001). The between-group difference in fat mass change was statistically significant (p = 0.0396). Other parameters, including lean body mass and total body water, showed no significant differences at baseline, week 48, or in the magnitude of change from baseline to week 48 (Table 5 and Figure 2C).

Table 5.

Change from Baseline in BIA

			Evogliptin (n = 29)		Dapagliflozin (n = 30)		P value2
			mean ± SD	P value1	mean ± SD	P value1
BIA	BMI (kg/m2)	Baseline	25.21 ± 3.66		25.71 ± 3.15		0.5809
		48 weeks	24.80 ± 3.59		24.57 ± 2.95		0.7917
		Difference	−0.42 ± 0.95	0.0258	−1.14 ± 1.07	<0.0001	0.0085
	Weight (kg)	Baseline	62.68 ± 11.72		61.1 ± 7.11		0.5291
		48 weeks	61.95 ± 10.41		57.78 ± 7.33		0.0781
		Difference	−0.74 ± 4.46	0.3731	−3.32 ± 1.68	<0.0001	0.0052
	Lean body mass (kg)	Baseline	39.27 ± 5.18		37.55 ± 4.30		0.1693
		48 weeks	39.48 ± 7.68		36.73 ± 4.17		0.0971
		Difference	0.20 ± 4.76	0.8198	−0.82 ± 0.93	<0.0001	0.2547
	Fat mass (kg)	Baseline	22.76 ± 6.44		23.60 ± 5.54		0.5923
		48 weeks	22.58 ± 7.40		21.63 ± 4.98		0.5642
		Difference	−0.18 ± 3.87	0.8087	−1.97 ± 2.47	0.0001	0.0396
	Total body water(L)	baseline	28.42 ± 3.35		27.63 ± 3.17		0.3556
		48 weeks	28.15 ± 3.17		27.01 ± 3.07		0.1651
		Difference	−0.27 ± 0.82	0.0880	−0.62 ± 0.70	<0.0001	0.0823

Notes: Variables are presented as the mean ± SD. P value¹ indicates the within-group comparison of change from baseline, and P value² indicates the between-group comparison. Within-group changes were evaluated using paired t-tests or signed-rank tests, as appropriate, and between-group comparisons were performed using two-sample t-tests or Wilcoxon rank-sum tests, as appropriate. A two-sided P value <0.05 was considered statistically significant.

Abbreviation: BIA, body impedance analysis.

In the safety assessment, of the 60 participants evaluated for adverse events, 12 (40.00%) in the evogliptin group and 11 (36.7%) in the dapagliflozin group experienced adverse events. Twenty-six adverse events were reported in the evogliptin group, and 15 adverse events were reported in the dapagliflozin group (Supplementary Table 1). Most adverse events were mild in severity (25 cases [96.15%] in the evogliptin group and 14 cases [93.33%] in the dapagliflozin group), and none were determined to be definitely related to the study intervention. Only one probable-related adverse event was reported in each group: dizziness in the evogliptin group and urinary tract infection in the dapagliflozin group.

One newly developed fracture was detected in the evogliptin group and was identified as an L2 compression fracture.

Discussion

In this trial comparing evogliptin and dapagliflozin in postmenopausal patients with T2DM, a small between-group difference in femoral neck BMD was observed. The evogliptin group showed no significant changes in BMD, whereas the dapagliflozin group showed a modest reduction. After adjustment, the direction of the between-group difference remained in favor of evogliptin, although the evidence for a statistically significant difference was less robust. Given the small absolute difference and the variability in BMD changes, the clinical significance of this finding should be interpreted with caution.

Regarding the relationship between DPP-4i and bone metabolism, several studies have revealed an inverse correlation between DPP-4 activity levels and BMD, suggesting that DPP-4 plays a significant role in maintaining bone homeostasis.^29–32 A substantial body of evidence, including randomized controlled trials, cohort studies, meta-analyses, and systematic reviews, has demonstrated that DPP-4i exhibit either neutral^28,33 or beneficial^17,34,35 effects on bone metabolism. However, within the DPP-4i class, individual agents have varying effects on bone metabolism. Studies have shown that sitagliptin either exhibits protective³⁶ effects on bone metabolism or maintains neutral³⁷ effects. Linagliptin has shown a favorable outcome³⁸ regarding fracture risk, and alogliptin has shown positive effects³⁹ on bone fracture prevention. Vildagliptin exhibits a safety profile comparable to that of gliclazide-modified release (MR) in terms of bone metabolic effects.¹⁶ In contrast, saxagliptin has yielded mixed results, with reports of both neutral effects,^15,40 fracture risk, and adverse effects⁴¹ on bone metabolism. Additionally, trelagliptin is associated with an increased risk of fracture.⁴² Among the numerous studies, none have investigated the association between evogliptin and bone metabolism. Evogliptin exhibits a neutral effect on bone metabolism in the present study.

In the present study, dapagliflozin was associated with a small reduction in femoral neck BMD, in contrast to some previous studies^{23,24,43–45} reporting a neutral effect of dapagliflozin on bone metabolism. However, these findings are consistent with the results of randomized controlled trials on canagliflozin.²⁰ Similar to findings from canagliflozin trials, a modest reduction in femoral neck BMD was observed in the dapagliflozin group. Significant weight loss was observed in the dapagliflozin group, and although the increase in serum CTX levels was not statistically significant, CTX increased numerically by 23.7% in the dapagliflozin group compared with an 8.1% increase in the evogliptin group. This finding may be compatible with enhanced bone remodeling during dapagliflozin treatment, potentially implicating an effect on bone health and raising the possibility that treatment-induced weight loss may have contributed to the observed changes in CTX⁴⁶ (Supplementary Fig. 1). However, any mechanistic interpretation should be considered hypothesis-generating.

Similar studies, although not primarily focused on analyzing the effects of SGLT2 inhibitors on bone metabolism, have reported relevant findings. Kohan et al⁴⁷ observed an increased incidence of fractures accompanied by weight loss in a dapagliflozin-treated group. Additionally, Bolinder et al⁴⁸ reported significant weight loss in the dapagliflozin-treated group, with non-vertebral BMD showing a greater reduction than that in the placebo group, although the overall BMD differences were not statistically significant.

In the exploratory multivariable regression analysis, no statistically significant factors associated with changes in femoral neck BMD were identified, including change in BMI, baseline 25OHD, age, and treatment group (Supplementary Table 2). However, visual inspection of the scatter plot in the dapagliflozin group suggested that greater BMI reduction might accompany greater femoral neck BMD loss, although this relationship was not statistically significant (Supplementary Fig. 1). This observation should therefore be considered hypothesis-generating and warrants further investigation of treatment-related weight loss as a potential contributor to BMD change. The lack of statistical significance may in part reflect the relatively small sample size and the generally preserved BMD of the study population. Larger studies are needed to determine whether changes in femoral neck BMD differ according to the magnitude of treatment-related weight reduction.

A Lancet report⁴⁹ suggested that vitamin D supplementation may confer modest benefits to the femoral neck. In this study, the baseline 25OHD levels were significantly lower in the dapagliflozin group than in the evogliptin group. Although vitamin D supplementation led to increases in both groups without a significant between-group difference in change from baseline, 25OHD levels remained significantly lower in the dapagliflozin group at week 48. Baseline 25OHD was not significantly associated with femoral neck BMD change in the exploratory regression analysis. In the supportive ANCOVA adjusted for baseline femoral neck BMD and 25OHD, the between-group difference in femoral neck BMD did not reach statistical significance. Nonetheless, the baseline imbalance in 25OHD should be considered when interpreting these findings. In addition, the initiation of a standardized calcium–vitamin D supplement in all participants at randomization may have attenuated potential between-group differences in bone-related outcomes.

This study’s limitations are as follows: First, although a formal sample size calculation was performed, this study was exploratory with a relatively small sample size and was not designed as a confirmatory non-inferiority or equivalence trial. Therefore, the study may have limited statistical power to detect subtle between-group differences in BMD outcomes, and non-significant findings should not be interpreted as evidence of equivalence. Furthermore, while the overall reduction in BMD was not substantial, its high variability made it difficult to identify robust predictors. Regarding the factors associated with FN BMD reduction, a larger cohort might have yielded more conclusive results. Consequently, these findings should be interpreted as exploratory, offering meaningful preliminary insights rather than definitive conclusions. Second, despite randomization, significant baseline imbalances in 25OHD and FBS were observed between the groups, likely occurring by chance due to the relatively small sample size. Although the supportive analysis adjusted for baseline 25OHD, the findings should still be interpreted in light of these baseline differences. Third, as multiple skeletal sites were assessed as primary endpoints and numerous secondary endpoints were evaluated without multiplicity adjustment, the risk of type I error is increased. The significant finding at the femoral neck should therefore be interpreted with caution. Fourth, the mechanisms by which evogliptin and dapagliflozin affect BMD were not investigated in this study. Fifth, in this relatively older study population, degenerative spinal changes may have limited the accuracy of the lumbar spine BMD measurements. Additionally, this study was not designed or powered to assess fracture outcomes, and only one incident fracture was observed in the evogliptin group. Therefore, any conclusions regarding fracture risk should be interpreted with caution. Lastly, although adjusted analyses were performed for femoral neck BMD, residual confounding from unmeasured factors cannot be fully excluded. Future studies incorporating detailed, systematic evaluations of these variables would provide more comprehensive insights.

Nevertheless, this study has notable strengths, including its design as a randomized clinical trial that minimizes bias, controls for external factors, and enhances the reliability of the statistical analyses. To our knowledge, this is the first study to report the safety profile of evogliptin concerning bone health, which represents a significant strength. Future large-scale clinical trials based on these findings are warranted to evaluate the fracture risk and long-term effects on the bone.

Conclusion

In conclusion, compared with dapagliflozin, evogliptin was associated with preservation of femoral neck BMD, although the between-group difference should be interpreted cautiously after adjustment for baseline differences. To our knowledge, this is the first study to evaluate the effect of evogliptin on bone metabolism. These findings provide preliminary evidence to inform the selection of glucose-lowering therapy in patients at increased risk of impaired bone health. Further large-scale studies with longer follow-up are warranted to confirm these findings and determine their implications for fracture risk.

Funding Statement

This study was an investigator-initiated trial (IIT) funded by Dong-A ST. The sponsor had no role in study design, data analysis, interpretation, or writing of the report.

Declaration of Competing Interest

The authors declare no competing interests.

Data Sharing Statement

The deidentified individual participant data used and analyzed during the current study are available from the corresponding author upon reasonable request.

Ethics Approval

This multicenter study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Boards of the participating institutions: Kyung Hee University Hospital at Gangdong (KHNMC 2020-08-007-001), Gangneung Asan Hospital (GNAH 2020-10-012-001), Kyung Hee University Hospital (KHUH 2020-10-035-005), Yeouido St. Mary’s Hospital (SC20MIDV0118), Soonchunhyang University Seoul Hospital (SCHUH 2020-10-019-001), Konkuk University Hospital (KUMC 2020-10-013-002), and Inha University Hospital (2020-08-025). Written informed consent was obtained from all participants prior to enrollment.

Author Contributions

Ho Yeon Chung: Conceptualization, Methodology, Supervision.

So Young Park, Ha-Young Kim, Dong-Won Byun, Kee-Ho Song, Seongbin Hong, Ki-Hyun Baek, Ho Yeon Chung: Investigation.

Jungbok Lee: Formal analysis.

Hyun Jin Ryu, Jungbok Lee, Ki-Hyun Baek, Ho Yeon Chung: Data curation.

Hyun Jin Ryu: Writing – original draft.

Hyun Jin Ryu, So Young Park, Ha-Young Kim, Dong-Won Byun, Kee-Ho Song, Seongbin Hong, Jungbok Lee, Ki-Hyun Baek, Ho Yeon Chung: Writing – review & editing.

All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

Hyun Jin Ryu, So Young Park, Ha-Young Kim, Dong-Won Byun, Kee-Ho Song, Seongbin Hong, and Jungbok Lee have nothing to disclose. The abstract of this paper was presented at the American Diabetes Association (ADA) 85^th Scientific Sessions as a poster presentation with interim findings. The poster’s abstract was published in “Poster Abstracts” in Diabetes: https://doi.org/10.2337/db25-792-P.