A Randomized Controlled Trial on the Effect of Luseogliflozin on Bone Microarchitecture Evaluated Using HR-pQCT in Elderly Type 2 Diabetes

Abstract

Introduction

Bone fragility is a critical issue in the treatment of elderly people with type 2 diabetes (T2D). In the Canagliflozin Cardiovascular Assessment Study, the subjects with T2D who were treated with canagliflozin showed a significant increase in fracture events compared to a placebo group as early as 12 weeks post-initiation. In addition, it has been unclear whether sodium-glucose co-transporter 2 (SGLT2) inhibitors promote bone fragility. We used high-resolution peripheral quantitative computed tomography (HR-pQCT) to prospectively evaluate the short-term effect of the SGLT2 inhibitor luseogliflozin on bone strength and microarchitecture in elderly people with T2D.

Methods

This was a single-center, randomized, open-label, active-controlled pilot trial for ≥ 60-year-old Japanese individuals with T2D without osteoporosis. A total of 22 subjects (seven women and 15 men) were randomly assigned to a Lusefi group (added luseogliflozin 2.5 mg) or a control group (added metformin 500 mg) and treated for 48 weeks. We used the second-generation HR-pQCT (Xtreme CT II®, Scanco Medical, Brüttisellen, Switzerland) before and 48 weeks after the treatment to evaluate the subjects' bone microarchitecture and estimate their bone strength.

Results

Twenty subjects (Lusefi group, n = 9; control group, n = 11) completed the study, with no fracture events. As the primary outcome, the 48-week changes in the bone strength (stiffness and failure load) estimated by micro-finite element analysis were not significantly different between the groups. As the secondary outcome, the changes in all of the cortical/trabecular microarchitectural parameters at the radius and tibia from baseline to 48 weeks were not significantly different between the groups.

Conclusions

In the pilot trial, we observed no negative effect of 48-week luseogliflozin treatment on bone microarchitecture or bone strength in elderly people with T2D.

Trial Registration

UMIN-CTR no. 000036202 and jRCT 071180061.

Key Summary Points

Why carry out this study?

The mega-trial Canagliflozin Cardiovascular Assessment Study described increased fracture events in canagliflozin-treated patients as early as 12 weeks post-initiation.

However, it remains unclear whether sodium-glucose co-transporter 2 (SGLT2) inhibitors deteriorate bone microarchitecture and decrease bone strength in elderly people with type 2 diabetes.

This study aimed to prospectively investigate the short-term effect of an SGLT2 inhibitor luseogliflozin on bone microarchitecture in elderly individuals with type 2 diabetes, with the use of second-generation high-resolution peripheral quantitative computed tomography (HR-pQCT).

What was learned from this study?

In the short-term (48 weeks) pilot trial in elderly people with type 2 diabetes, luseogliflozin did not deteriorate bone microarchitecture or decrease bone strength compared to metformin treatment when determined by second-generation HR-pQCT.

It may not be necessary to avoid the use of SGLT2 inhibitors for elderly people with type 2 diabetes who potentially have bone fragility.

Introduction

Diabetic bone disease is widely known as a secondary form of osteoporosis induced by diabetes mellitus. Compared to healthy individuals, people with type 2 diabetes have an increased risk of bone fractures despite having normal to increased bone mineral density (BMD) [1]. The causes of the elevated fracture risk in individuals with type 2 diabetes are thus suspected to be bone quality that has deteriorated due to microvascular complications, the accumulation of advanced-glycation end-products (AGEs) in the bone collagen fibers, and decreased bone turnover from exposure to hyperglycemia and oxidase stress [2, 3]. Many developed countries are facing an aging population, and the increasing prevalence of osteoporosis among elderly individuals with diabetes has become a substantial public health issue.

The influences of glucose-lowering agents on fracture risks have been investigated in clinical studies of people with type 2 diabetes. Treatment with a thiazolidinedione is known to be an independent risk factor for bone fracture [4, 5] because its activation of peroxisome proliferator-activated receptor-γ decreases the differentiation of osteoblasts [6]. In contrast, metformin, alpha-glucosidase inhibitors (α-GIs), and dipeptidyl peptidase-4 (DPP4) inhibitors have been shown to be neutral to bones [7, 8–9]. The U.S. Food and Drug Administration strengthened the warning for the sodium-glucose co-transporter 2 (SGLT2) inhibitor canagliflozin related to an increased risk of bone fracture, based on the outcomes of the Canagliflozin Cardiovascular Assessment Study (CANVAS), in which fractures increased as early as 12 weeks after the initiation of canagliflozin compared to placebo [10, 11].

However, other SGLT2 inhibitors did not increase fracture events in randomized controlled trials (RCTs) [12, 13–14], and canagliflozin also did not increase fractures in the Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation trial, which is a more recent RCT of individuals with type 2 diabetes complicated by chronic kidney disease [15]. A recent meta-analysis demonstrated that SGLT2 inhibitors exhibited no significant effects on BMD changes at the lumbar spine, femur, or radius when determined by dual-energy X-ray absorptiometry (DXA) in subjects with type 2 diabetes [16]. The question of whether SGLT2 inhibitors increase the risk of bone fractures in clinical practice among individuals with type 2 diabetes remains controversial. It also remains unknown which drug(s) among the SGLT2 inhibitors affect bone fragility. It is possible that the selectivity of SGLT2 against sodium-glucose co-transporter 1 (SGLT1) may affect bone metabolism and the fracture risk, since the selectivity of canagliflozin (the study drug in CANVAS) is the lowest among the SGLT2 inhibitors under clinical development [17].

The fact is that for clinicians, the algorithm for predicting fracture risk is based on interview, clinical examination, and the fracture risk assessment tool (FRAX), as well as two-dimensional areal BMD assessed using DXA. However, the combination of those measures cannot allow us enough to predict the risk of a subsequent fracture, since the nature of the microarchitecture of bone tissue, which is not assessed using FRAX and DXA, plays a key role in the development of a fracture. As additional DXA-based analyses, trabecular bone score (TBS) [18] and hip structural analysis (HSA) [19] have been developed to derive lumbar spine structural quality and femoral geometry, respectively. These DXA-derived two-dimensional images are unable to reflect the microarchitectural aspects of bone and do not allow for independent assessment of cortical and trabecular bone.

High-resolution peripheral quantitative computed tomography (HR-pQCT) is an innovative clinical research tool providing noninvasive three-dimensional imaging (Fig. 1). HR-pQCT enables an analysis of bone quality based on an evaluation of the bone microarchitecture, the three-dimensional volumetric BMD, and geometry separately for the cortical and trabecular compartments of the distal radius and tibia [20]. HR-pQCT images can be used in micro-finite element (µFE) analyses to calculate bone-strength indices. This technique can provide a fracture risk that is independent of the areal BMD determined by DXA [21] and FRAX scores [22]. The most recently developed second-generation HR-pQCT provides enhanced spatial resolution (voxel size 61 mm) and allows more detail analyses of cortical and trabecular bones compared to the first-generation HR-pQCT (voxel size 82 mm) [23].

[See PDF for image]

Fig. 1

Image samples of distal radius (A–F) of a patient with type 2 diabetes allocated in the Lusefi group scanned by second-generation HR-pQCT (Xtreme CT II^®, Scanco Medical, Brüttisellen, Switzerland). A Three-dimensional imaging of distal radius, B the cortical compartments, C the trabecular compartments, D distribution of cortical porosity, E cross-sectional image of distal radius, F magnification of cortical bone with porous and construction of trabecule

Luseogliflozin, an oral SGLT2 inhibitor with high selectivity for inhibition of SGLT2 over SGLT1, was licensed by the regulatory authority of Japan in 2014 [24]. As well as other SGLT2 inhibitors, luseogliflozin showed reductions of glycemia, body weight, serum lipid levels, and markers of renal and hepatic functions in a meta-analysis [25]. However, to date, luseogliflozin has been approved by only several Asian countries including Japan, Thailand, Malaysia, and the Philippines, and there is less evidence on its efficacy and safety (including fracture risk) owing to the absence of large-scale RCTs of luseogliflozin.

Investigations of bone microstructural changes using HR-pQCT when patients were treated with SGLT2 inhibitors have not been reported to date. We conducted the present study to prospectively investigate the effect of an SGLT2 inhibitor, luseogliflozin, on bone microarchitecture in elderly individuals with type 2 diabetes, with the use of second-generation HR-pQCT.

Methods

Subjects

We recruited Japanese individuals who fulfilled the following inclusion criteria at the screening visit: (i) age ≥ 60 years; (ii) diagnosed with type 2 diabetes; (iii) an outpatient; (iv) treated with diet therapy alone or with an orally administered antidiabetic agent including metformin (≤ 1000 mg/day) and/or an α-GI and/or a DPP4 inhibitor, and whose treatment had not been changed within 6 months before enrollment; and (v) providing written informed consent.

Individuals were enrolled when they also fulfilled the criteria at the baseline, i.e., week 0: (i) an HbA1c (NGSP) value ≥ 7.0% (53 mmol/mol) and < 9.0% (75 mmol/mol); and (ii) a T-score of the areal BMD in both lumbar vertebrae and the femoral neck determined by DXA >  − 2.5 standard deviation (SD) [26].

Individuals were excluded if they met any of the following criteria: a history of treatment with any SGLT2 inhibitor, complicated with a bone metabolic disorder, a history of any treatment for osteoporosis (including calcium and/or vitamin D supplementations) within 12 months before enrollment, an estimated glomerular filtration rate (eGFR) < 30 ml/min/1.73 m², anemia (hemoglobin < 10 g/dl), malignancy, alcohol consumption ≥ 20 g/day, tobacco smoking habit within 12 months before enrollment, chronic liver disease with a Child–Pugh score ≥ 6 points, and a body mass index (BMI) < 18.5 kg/m².

Study design

This study was a single-center, randomized, open-label, parallel-group, active-controlled trial carried out at Nagasaki University Hospital from April 2019 to August 2023. The study was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR, no. 000036202) and with the Japan Registry of Clinical Trials (jRCT, no. 071180061). The study design is depicted in Fig. 2, and the study's details have been published as the study protocol paper [27].

[See PDF for image]

Fig. 2

Summary of the study design and the randomized subjects. AE adverse event, α-GI alpha glucosidase inhibitor, BMD bone mineral density, DPP4i dipeptidyl peptidase-4 inhibitor, DXA dual-energy X-ray absorptiometry, HbA1c glycosylated hemoglobin, SD standard deviation

We allocated the enrolled subjects in a 1:1 ratio to either the Lusefi group or the control group through a randomization process. The randomization was conducted using a stratified block method based on sex (with a block size of 4) and was managed using the Research Electronic Data Capture (REDCap) system.

The subjects in the Lusefi group were administered luseogliflozin (Lusefi®, Taisho Pharmaceutical Co., Tokyo) 2.5 mg once a day from the baseline (week 0) in addition to their previous treatment. They visited our hospital every 12 weeks during the 48-week treatment period of the study. The dose of luseogliflozin increased to 5 mg unless a subject achieved an HbA1c level < 7% (53 mmol/mol) or a decrease in the HbA1c of ≥ 0.5% compared to the previous visit. If a subject in the Lusefi group did not maintain an HbA1c level as described above, metformin 500 mg/day (250 mg twice a day) was added at each 12-week scheduled visit.

The subjects allocated to the control group were newly or additively administered metformin 250 mg twice a day (500 mg/day) from the baseline (week 0) in addition to their previous treatment. The additional metformin was administered at each 12-week scheduled visit by 500 mg increments unless the subject achieved an HbA1c level < 7% (53 mmol/mol) or a decrease in HbA1c ≥ 0.5% compared to the previous visit. In both the control and Lusefi groups, a maximum metformin dose at 2250 mg per day was permitted.

The dosage regimens of luseogliflozin and metformin were in accordance with the guidance by Japan Pharmaceutical and Medical Device Agency. In Japan, metformin is recommended to be initiated at 500 mg/day and is approved to be gradually increased its dosage up to 2250 mg/day if patient’s glycemic control is unsatisfied. Luseogliflozin is recommended to be initiated at 2.5 mg/day and is approved for administration up to 5 mg/day.

Participants were prohibited from receiving any treatment for osteoporosis, including calcium and/or vitamin D supplementations, during the study period. We monitored the subjects with blood sampling at each visit to determine whether any adverse events had occurred. All subjects were evaluated by DXA (iDXA®, GE Healthcare Medical Systems, Waukesha, WI, USA) and second-generation HR-pQCT (Xtreme CT II®, Scanco Medical, Brüttisellen, Switzerland) at baseline (week 0) and at the end of the treatment period (week 48).

The primary outcome of the study was the 48-week change from baseline in the predictive value of bone strength estimated by the HR-pQCT with a µFE analysis (Scanco Medical FE software ver. 1.13, Scanco Medical) in the Lusefi group compared with those in the control group. The predictive bone strength values include the µFE-estimated stiffness (N/mm) and the failure load (N) at both the distal radius and the distal tibia of the nondominant body side. The secondary outcome of the study was the changes in all parameters of bone morphology measured by HR-pQCT, the areal BMDs at the lumbar vertebrae, femur, and radius measured by DXA, laboratory data including HbA1c and bone biological markers, and clinical parameters including body weight from baseline (week 0) to the end of the treatment period (week 48) in the Lusefi group compared to those in the control group. All adverse events (including bone fractures) that occurred in either group during the trial were recorded.

Statistical Analyses

This study was a first pilot trial aimed at evaluating changes in bone microstructure through HR-pQCT following treatment with luseogliflozin in comparison with metformin. There were no analogous previous studies comparing bone strength before and after an intervention, and no available information to estimate an optimal sample size. We therefore used a sample size of 12 subjects in each group in reference to Julious et al. [28]. The study's results are presented as the mean ± SD or median (interquartile range). The primary purpose of this study was to evaluate the effect of luseogliflozin on bone-quality changes, and we thus used a per-protocol set for the primary analysis. The bone-deterioration effects of the treatment are expressed as the mean changes from baseline (week 0) to week 48 calculated by linear regression models adjusted for the baseline values of the respective measures and a stratification factor (sex) with corresponding 95% confidence intervals (CIs). Since this was a pilot study, the p values should be interpreted in an exploratory context. All analyses were performed with the use of R ver. 4.2.2.

Ethical Considerations

The study was approved by the Certified Review Board of Nagasaki University Hospital (no. CRB18-0006). We conducted the study in accordance with the Helsinki Declaration of 1964 and its later amendments. Written informed consent was obtained from all subjects.

Results

Subjects' Characteristics and Composition of the study

Figure 2 shows the consort diagram of the study. A total of 27 subjects (nine women and 18 men) were enrolled. One man withdrew his informed consent before enrollment. Two women and two men did not meet the run-in criteria of the study, and we were able to enroll two fewer eligible subjects than the planned 24 subjects during the enrollment period. A total of 22 subjects (seven women and 15 men) were thus randomly assigned to the Lusefi group (n = 10) or the control group (n = 12) as intention-to-treat subjects. At randomization, the baseline characteristics of the Lusefi group and the control group were as follows: age (years), 69.4 ± 5.7 and 67.7 ± 5.5; BMI (kg/m²), 26.1 ± 4.3 and 23.4 ± 1.6; and HbA1c (%), 7.6 ± 0.3 and 7.4 ± 0.3, respectively. Table 1 summarizes the 22 subjects' characteristics. Since the randomization was conducted based on only sex because of the small sample size, the inhomogeneity between the groups was observed in weight and hematocrit levels at baseline.

Table 1. Baseline characteristics of the subjects at randomization

	Lusefi group, n = 10	Control group, n = 12	p value
Age, years	69.4 ± 5.7	67.7 ± 5.5	0.56
Female, n (%)	3 (30%)	4 (33%)	0.86
Height, cm	167.1 ± 10.7	159.2 ± 9.4	0.075
Weight, kg	73.0 ± 14.8	59.5 ± 8.5	0.043
BMI, kg/m2	26.1 ± 4.3	23.4 ± 1.6	0.31
Duration of diabetes, years	6.1 ± 4.9	10.8 ± 7.4	0.20
Family history of type 2 diabetes, n (%)	2 (20%)	5 (42%)	0.27
Family history of osteoporosis, n (%)	3 (30%)	1 (8.3%)	0.97
Complications, n (%)
Hypertension	4 (40%)	6 (50%)	0.49
Dyslipidemia	6 (60%)	9 (75%)	0.73
Diabetic retinopathy	0 (0%)	1 (3.3%)	0.35
Diabetic nephropathya	2 (20%)	2 (16.7%)	0.55
Diabetic neuropathy	1 (10%)	1 (8.3%)	0.81
Cardiovascular disease	2 (20%)	1 (8.3%)	0.43
Smoking habits, n (%)
Never	9 (90%)	10 (83%)	0.65
Former	1 (10%)	2 (17%)	0.65
Laboratory data
Hematocrit, %	43.8 ± 2.3	40.4 ± 3.3	0.011
Creatinine, mg/dl	0.84 ± 0.20	0.77 ± 0.17	0.17
Calcium, mg/dl	9.2 ± 0.3	9.4 ± 0.3	0.23
Phosphate, mg/dl	3.0 ± 0.4	2.9 ± 0.4	0.28
Intact PTH, pg/ml	52.3 ± 32.3	37.4 ± 9.06	0.60
25OHD, ng/ml	18.7 ± 7.4	16.9 ± 5.1	0.55
Uric acid, mg/dl	5.4 ± 1.3	5.4 ± 1.8	0.97
HbA1c, %	7.6 ± 0.3	7.4 ± 0.3	0.063
HbA1c, mmol/mol	58.5 ± 3.1	56.7 ± 3.8	0.063
Areal BMD measured by DXA
Lumbar spine (L1–4), g/cm2	1.281 ± 0.315	1.096 ± 0.173	0.18
Lumbar spine (L1–4), T-score	0.98 ± 2.22	− 0.24 ± 1.28	0.22
Total hip, g/cm2	1.037 ± 0.142	0.939 ± 0.086	0.25
Total hip, T-score	− 0.05 ± 1.11	− 0.57 ± 0.70	0.23
Femoral neck, g/cm2	0.925 ± 0.142	0.828 ± 0.094	0.15
Femoral neck, T-score	− 0.65 ± 1.13	− 1.31 ± 0.78	0.15
1/3 radius, g/cm2	0.844 ± 0.132	0.803 ± 0.166	0.77
1/3 radius, T-score	− 0.78 ± 1.00	− 0.98 ± 1.14	0.81

Data are mean ± standard deviation

PTH parathyroid hormone, 25OHD 25-hydoroxyvitamin D, BMD bone mineral density, BMI body mass index, DXA dual energy X-ray absorptiometry, HbA1c glycated hemoglobin

^aDefined by albuminuria ≥ 30 mg/gCr or/and proteinuria ≥ 0.5 g/gCr, or/and estimated glomerular filtration rate < 30 ml/min/1.73 m²

One woman allocated to the control group and one man allocated to the Lusefi group left the study at 22 and 43 weeks after the randomization, respectively, due to adverse events. Finally, 20 subjects (nine in the Lusefi group and 11 in the control group) completed the study and were analyzed as the per-protocol set. The study drugs (luseogliflozin and metformin) were administered according to the study protocol. The final dose of luseogliflozin in the Lusefi group was 3.9 ± 1.3 mg. The final additional dose of metformin in the Lusefi and control groups were 444 ± 464 mg and 1432 ± 561 mg, respectively.

Primary Outcome

The µFE-Estimated Bone Strength

As primary outcomes, the stiffness (N/mm) and the failure load (N) at both the radius and the tibia were calculated using μFE analyses based on each voxel within the segmented HR-pQCT images [29]. Figure 3A and Supplemental Table S1 explain the stiffness and the failure load at the baseline (week 0) and week 48, and the changes in these parameters from baseline to week 48.

[See PDF for image]

Fig. 3

Results of HR-pQCT values in the Lusefi group and the control group. A The µFE-estimated bone strength (stiffness and failure load) at baseline (week 0) and week 48, and the comparisons between the 48-week changes in the Lusefi group and those in the control group. B The cortical/trabecular microarchitectural values at baseline (week 0) and week 48, and the comparisons between the 48-week changes in the Lusefi group and those in the control group. Red marks: the Lusefi group (L). Blue marks: the control group (C). Red (blue) bars and bold horizontal lines indicate interquartile ranges and medians. Black diamonds within each bar indicate means. Ct.Ar cortical area, Ct.Pm cortical perimeter, Ct.Po cortical porosity, Ct.PoDm cortical pore diameter, Ct.Th cortical thickness, Ct.vBMD cortical volumetric bone mineral density, NS not significant, Tb.1/N.SD inhomogeneity of trabecular network, Tb.Ar trabecular area, Tb.BV/TV trabecular bone volume fraction, Tb.N trabecular number, Tb.Th trabecular thickness, Tb.Sp trabecular separation, Tb.vBMD trabecular volumetric bone mineral density, Wk week. §p < 0.05 vs. baseline (week 0) in each group

In the radius, the stiffness and the failure load did not significantly change from baseline to week 48 in the Lusefi group, or in the control group. In linear regression models adjusted for the baseline values and sex, the 48-week changes in the radial stiffness and failure load showed no significant differences between the groups (β 466, 95% CI − 3452 to 4384 in stiffness, and β 35, 95% CI − 215 to 284 in failure load respectively for the Lusefi group compared to the control group).

In the tibia, the stiffness and the failure load significantly increased from baseline to week 48 in the control group, but not in the Lusefi group. However, no significant differences in the 48-week changes in either the stiffness or the failure load were observed between the groups in the linear regression models (β − 314, 95% CI − 7178 to 884 in stiffness, and β − 147, 95% CI - 330 to 36 in failure load, respectively, for the Lusefi group compared to the control group).

In the primary outcomes, changes in the predictive bone strength from baseline to week 48 were not apparent differences between the Lusefi group and the control group.

Secondary Outcomes

To further evaluate the effect of treatment with luseogliflozin on bone, we compared the differences in the 48-week changes in the HR-pQCT parameters of microarchitecture of cortical/trabecular bone between the Lusefi group and the control group. Additionally, the differences in the changes in clinical/laboratory data during 48 weeks between the groups were also evaluated.

Cortical and Trabecular Bone Microarchitecture

The bone microarchitectural values determined by HR-pQCT at the baseline (week 0) and week 48 and their changes from baseline to week 48 are presented in Fig. 3B and Supplemental Tables S2 and S3. In the Lusefi group, none of the cortical and trabecular values of radius and tibia changed significantly from the baseline to week 48 in the Lusefi group. In the control group, the cortical porosity and cortical thickness of the tibia increased from the baseline to week 48, but the other parameters of the tibia and all of the parameters of the radius did not change significantly. No significant between-group differences in the 48-week changes in any of the parameters were observed.

BMD Measured by DXA

Supplemental Table S4 presents the results of the comparison between the baseline (week 0) and week 48 in the areal BMD values of the lumbar spine, femur, and radius determined by DXA in the Lusefi and control groups. There were no significant changes in the BMD values between baseline and week 48 in either group; nor were there significant between-group differences in the 48-week changes in the BMD values.

Changes in Clinical Parameters and Laboratory Data

Figure 4 A shows the changes in clinical and biochemical values from baseline (week 0) over each visit during the 48-week treatment period of the study in the Lusefi and control groups. In the Lusefi group, the systolic/diastolic blood pressure values showed reductions (18.6/7.3 mmHg, respectively) at week 48 compared to baseline, but these parameters showed no significant changes (0.2/3.0 mmHg, respectively) in the control group. The subjects' body weight decreased from baseline to week 48 by 2.18 kg in the Lusefi group and 0.82 kg in the control group, and the difference in weight changes between the groups was − 1.1 kg (95% CI − 2.5 to 0.22, p = 0.094).

[See PDF for image]

Fig. 4

Changes in physical/biochemical data and bone metabolic markers from baseline (week 0) in the Lusefi and control groups. A Changes in physical and biochemical data from baseline (week 0) to 12, 24, 36, and 48 weeks after treatment with the study drugs. B Changes in bone metabolic markers from baseline (week 0) to week 48. Red marks: the Lusefi group (L). Blue marks: the control group (C). Red (blue) bars and bold horizontal lines indicate interquartile ranges and medians. 1,25OHD 1,25-dihydroxyvitamin D, 25OHD 25-hydoroxyvitamin D, ALT alanine aminotransferase, AST aspartate aminotransferase, BAP bone alkaline phosphatase, BMI body mass index, BUN blood urea nitrogen, Ca calcium, Cl chloride, Cr creatinine, DBP diastolic blood pressure, DKK1 dickkopf-related protein 1, FECa fractional excretion of calcium, FENa fractional excretion of sodium, FEP fractional excretion of phosphate, GGT gamma-glutamyl transpeptidase, HbA1c glycosylated hemoglobin, HDL-C high-density lipoprotein cholesterol, K potassium, LDL-C low-density lipoprotein cholesterol, Mg magnesium, Na sodium, P phosphate, P1NP procollagen type 1 N-terminal propeptide, PTH parathyroid hormone, RBC red blood cells, SBP systolic blood pressure, TG triglyceride, TRACP-5b tartrate-resistant acid phosphatase 5b, ucOC undercarboxylated osteocalcin. *p < 0.05 vs. baseline (week 0) in each group; †p < 0.05 vs. the control group

The HbA1c levels decreased from baseline to week 48 by 0.59% in the Lusefi group and 0.32% in the control group. In the Lusefi group, the levels of hematocrit, hemoglobin, red blood cells (RBC), high-density lipoprotein-cholesterol, blood urea nitrogen, creatinine, and sodium increased, but the levels of gamma-glutamyl transferase (GGT) and uric acid were decreased at week 48 compared to the baseline. In the control group, the levels of GGT, triglyceride, and low-density lipoprotein-cholesterol were decreased at week 48 compared to the baseline. The changes from baseline to week 48 in hematocrit, hemoglobin, RBC, and UA differed between the Lusefi and control groups.

The urinary analyses revealed that in the Lusefi group, the levels of glucose increased from baseline to week 12 and maintained this tendency until week 48. The values of the fractional urinary excretion of sodium (FENa) increased from baseline to week 48 in the Lusefi group, and the changes were different compared to those in the control group. In both groups, the values of fractional urinary excretions of calcium and phosphate and the urinary albumin/creatinine ratio were not significantly different at week 48 compared to the baseline.

Changes in Bone Metabolic Markers

Figure 4B illustrates the changes in bone metabolic markers from baseline to week 48 in the Lusefi and control groups. The levels of intact parathyroid hormone (PTH) increased from baseline to week 48 in the Lusefi group but decreased in the control group. However, the between-group difference in the 48-week change was not significant (β: 6.9, 95% CI − 0.17 to 14; p = 0.055). In the Lusefi group, there were no apparent changes from baseline to week 48 in the levels of 25-hydoroxyvitamin D, 1,25-dihydroxyvitamin D, undercarboxylated osteocalcin, bone alkaline phosphatase (BAP), tartrate-resistant acid phosphatase 5b, total procollagen type 1 N-terminal propeptide (P1NP), osteocalcin, antagonists of the Wnt/β-catenin signaling pathway including sclerostin and dickkopf-related protein 1 (DKK1), or pentosidine. In the control group the levels of BAP and total P1NP decreased from baseline to week 48. There were no significant between-group differences in the 48-week changes of any markers.

Associations Between the Changes in HR-pQCT Parameters and the Changes in Clinical/Laboratory Data

We investigated the associations between the 48-week changes in the HR-pQCT parameters (including stiffness and failure load) and those in the clinical/laboratory data, and we detected no significant correlations between the HR-pQCT parameters, and any factors examined (including HbA1c and weight reduction) in either the Lusefi group or the control group.

In the secondary outcomes, there were no significant differences in the 48-week changes in both HR-pQCT-derived bone microarchitecture and clinical/laboratory data between the Lusefi group and the control group.

Adverse Events

No fracture events occurred in the Lusefi and control groups during the study. Ketoacidosis, hyperglycemic coma, and severe hypoglycemia did not occur in either group. One woman in the control group hospitalized due to the coronavirus disease 2019 at 22 weeks and one man in the Lusefi group developed acute cardioembolic stroke at 43 weeks during the study. These subjects' adverse events were judged by clinicians as unrelated to the study treatments.

Discussion

To the best of our knowledge, the present study is the first prospective trial investigating the impact of SGLT2 inhibitors on bone strength and microarchitecture using second-generation HR-pQCT. Individuals with type 2 diabetes are known to have an increased risk of bone fragility, with a 40–70% higher risk of hip fracture compared to their healthy counterparts [1, 30]. It is important to determine the precise effects of anti-diabetes medications on the bone health of elderly individuals with type 2 diabetes. In the efficacy and safety of luseogliflozin in people with type 2 diabetes, there were no negative reports that address the effect of luseogliflozin on bone health [25]. Our present analyses revealed no adverse effects on bone strength or microarchitecture during 48 weeks of treatment with luseogliflozin in elderly individuals with type 2 diabetes compared to those treated with metformin as the active control. Metformin is used worldwide to treat type 2 diabetes and has been reported to have neutral or at least not detrimental effects on bone [7, 31]. By ensuring similar glycemic control between the present Lusefi and control groups during the treatment period, we aimed to avoid overestimating the effectiveness of luseogliflozin on bone metabolism through improved glycemic control. The outcomes of decreases in weight, blood pressure, HbA1c, GGT and UA, and increases in hematocrit, HDL-C and glycosuria observed in the Lusefi group were comparable with those from phase 3 trial of luseogliflozin for type 2 diabetes [32].

In earlier studies using HR-pQCT, bone microarchitectural analyses in subjects with type 2 diabetes demonstrated the subjects' skeletal fragility, which was contradictory to their relatively higher BMD determined by DXA. For example, with the use of HR-pQCT, Burghardt et al. were the first to report that individuals with type 2 diabetes had increased cortical porosity, suggesting a potential link to bone fragility in type 2 diabetes [33]. Waard et al. observed that in people with type 2 diabetes, inadequate glucose control was negatively correlated with cortical bone deterioration, with decreased thickness and volumetric BMD and increased porosity [34]. However, the findings of increased cortical porosity were not necessarily consistent across the studies using HR-pQCT in type 2 diabetes. Shanbhogue et al. stated that increased cortical porosity was not a characteristic of all subjects with type 2 diabetes but rather was a characteristic of a subgroup characterized by the presence of microvascular complications [35]. Samakkarnthai et al. reported that an accumulation of AGEs and reduced microvascular blood flow contributed to the increase of cortical porosity in patients with type 2 diabetes [36]. These results suggest that additional factors might be needed to explain the cortical bone deterioration in individuals with type 2 diabetes.

A recent meta-analysis of 16 studies using first-generation HR-pQCT in type 2 diabetes revealed increased cortical porosity in the radius but not in the tibia compared to controls [37]. Both the radius and tibia in the type 2 diabetes groups showed significantly increased cortical thickness compared to the controls, whereas the cortical volumetric BMD values were similar in the two groups. The individuals with type 2 diabetes also exhibited higher trabecular bone parameters, including an increased trabecular number and increased volumetric BMD, and reduced inhomogeneity of the trabecular network compared to controls [37]. The enhanced trabecular features might suggest a compensatory response to cortical microarchitectural weakness in type 2 diabetes.

Weight loss was reported to be a significant factor associated with deteriorations of HR-pQCT indices (cortical porosity, cortical thickness, cortical volumetric BMD (vBMD), trabecular number and trabecular vBMD, and failure load), especially in the tibia rather than the radius, in elderly people [26]. In the present study, the body weight reduction was larger in the Lusefi group than the control group (Fig. 4A). The differences in the 48-week changes in the stiffness and the failure load of the tibia tended to increase (but not significantly) from the baseline to week 48 in the control group compared to the Lusefi group, but the changes in the indices of the radius were comparable between the groups (Fig. 3A). Although we investigated a small number of cases, this difference in weight reduction might have affected the changes of estimated tibial strength between the groups. In the Lusefi group, there were no amelioration/deteriorations observed in the HR-pQCT indices of cortical/trabecular microarchitectures in the radius and tibia during the study, whereas in the control group both positive (increased cortical thickness) and negative (increased cortical porosity) effects on tibial microarchitecture were observed (Fig. 3B). No negative effects on bone microarchitecture were observed after 48 weeks of luseogliflozin treatment.

The results of basic research suggested a mechanism whereby SGLT2 inhibitors exert adverse effects on bone [38]. SGLT2 inhibitors decrease sodium reabsorption in the proximal tubules, which increases the availability of sodium to drive the cotransport of phosphate and sodium. Increased levels of serum phosphate may stimulate chronic hypersecretions of PTH from the parathyroid glands and fibroblast growth factor 23 from osteocytes, resulting in an enhancement of bone resorption and a decrease in the activation of vitamin D. We observed a tendency of an increase in PTH secretion (Fig. 4B) but did not observe a significant increase in serum phosphate levels in the Lusefi group compared to those of the control group (Fig. 4A). We also did not observe significant changes in activated vitamin D or bone metabolic markers after treatment with luseogliflozin (Fig. 4B). In clinical settings, the putative adverse effect of SGLT2 inhibitors on bone metabolism might thus be marginal.

In diabetic bone disease, the harmful effect of AGEs including pentosidine on bone fragility has been widely recognized [39]. SGLT2 inhibitors are known to have a pleiotropic effect of decreasing the accumulation of AGEs [40, 41]. However, we did not observe significant changes in serum pentosidine levels in the Lusefi group compared to the control group.

Recent research revealed that metformin contributes to the steering of anabolic activity in osteocytes by causing lower expressions in osteocytes of sclerostin and DKK1, which are the negative regulatory signals of the Wnt pathway [42]. In the present study, decreases in the serum levels of sclerostin or DKK1 were not observed in the control group with additional metformin treatment, and these changes were comparable to those observed in the Lusefi group.

Our study has several limitations. It was a pilot trial carried out at single center with a small sample size. We could not perfectly match the baseline characteristics of the subjects between the Lusefi and control groups because of the small sample size. The 48-week study period might be too short to lead to conclusions. Our findings may not apply to other ethnic groups. We are unable to conclude that our results were a class effect of the SGLT2 inhibitors.

Conclusions

Compared to metformin treatment, a 48-week treatment with luseogliflozin exerted no negative effect on bone strength or bone microarchitecture when evaluated with second-generation HR-pQCT. Although further investigations with larger populations and longer study periods are necessary before any conclusions can be made, our results support the tolerability of SGLT2 inhibitors regarding bone fragility in elderly people with type 2 diabetes.

Acknowledgements

We thank the participants of the study.

Medical Writing/Editorial Assistance

We thank Ms. Rachel Feldman at KN International, Inc. for her assistance in revising/editing English writing. The funding for her assistance was provided by Taisho Pharmaceutical Co., Ltd.

Author Contributions

Ichiro Horie, Ko Chiba, Makoto Osaki, Atsushi Kawakami and Norio Abiru contributed to the study conception and design. Riyoko Shigeno, Ichiro Horie, and Ai Haraguchi, Norio Abiru accomplished the study. Riyoko Shigeno and Shigeki Tashiro performed data management. Riyoko Shigeno, Ryuji Niimi, Yurika Kawazoe, and Shuntaro Sato analyzed the data for the study. Riyoko Shigeno, Ichiro Horie, and Norio Abiru drafted the manuscript.

Funding

This study was funded by Taisho Pharmaceutical Co., Ltd. (Tokyo). The funder played no part in the study design; the collection, management, analysis, or interpretation of data; writing the report; or the decision to submit the report for publication. The journal’s Rapid Service Fee was funded by Taisho Pharmaceutical Co., Ltd.

Date Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Riyoko Shigeno, Ichiro Horie, Ai Haraguchi, Atsushi Kawakami, and Norio Abiru have received research grants from Taisho Pharmaceutical Co., Ltd. Ryuji Niimi, Ko Chiba, Shigeki Tashiro, Yurika Kawazoe, Shuntaro Sato, and Makoto Osaki have nothing to disclose.

Ethical Approval

The study was approved by the Certified Review Board of Nagasaki University Hospital (no-CRB18-0006). We conducted the study in accordance with the Helsinki Declaration of 1964 and its later amendments. Written informed consent was obtained from all participants.

References

1. Vestergaard, P. Discrepancies in bone mineral density and fracture risk in patients with type 1 and type 2 diabetes–a meta-analysis. Osteoporos Int; 2007; 18, 4 pp. 427-444.1:STN:280:DC%2BD2s7jtVCltw%3D%3D [DOI: https://dx.doi.org/10.1007/s00198-006-0253-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17068657]

2. Napoli, N; Chandran, M; Pierroz, DD; Abrahamsen, B; Schwartz, AV; Ferrari, SL. Mechanisms of diabetes mellitus-induced bone fragility. Nat Rev Endocrinol; 2017; 13, 4 pp. 208-219.1:CAS:528:DC%2BC28XhsF2qsrjE [DOI: https://dx.doi.org/10.1038/nrendo.2016.153] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27658727]

3. de Waard, EA; van Geel, TA; Savelberg, HH; Koster, A; Geusens, PP; van den Bergh, JP. Increased fracture risk in patients with type 2 diabetes mellitus: an overview of the underlying mechanisms and the usefulness of imaging modalities and fracture risk assessment tools. Maturitas; 2014; 79, 3 pp. 265-274. [DOI: https://dx.doi.org/10.1016/j.maturitas.2014.08.003] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25192916]

4. Kahn, SE; Zinman, B; Lachin, JM et al. Rosiglitazone-associated fractures in type 2 diabetes: an Analysis from A Diabetes Outcome Progression Trial (ADOPT). Diabetes Care; 2008; 31, 5 pp. 845-851.1:CAS:528:DC%2BD1cXmsFGrt7o%3D [DOI: https://dx.doi.org/10.2337/dc07-2270] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18223031]

5. Aubert, RE; Herrera, V; Chen, W; Haffner, SM; Pendergrass, M. Rosiglitazone and pioglitazone increase fracture risk in women and men with type 2 diabetes. Diabetes Obes Metab; 2010; 12, 8 pp. 716-721.1:CAS:528:DC%2BC3cXhtVOmu7fK [DOI: https://dx.doi.org/10.1111/j.1463-1326.2010.01225.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20590749]

6. Adil, M; Khan, RA; Kalam, A et al. Effect of anti-diabetic drugs on bone metabolism: evidence from preclinical and clinical studies. Pharmacol Rep; 2017; 69, 6 pp. 1328-1340.1:CAS:528:DC%2BC2sXhsl2qtLvJ [DOI: https://dx.doi.org/10.1016/j.pharep.2017.05.008] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29132091]

7. Wolverton, D; Blair, MM. Fracture risk associated with common medications used in treating type 2 diabetes mellitus. Am J Health Syst Pharm; 2017; 74, 15 pp. 1143-1151.1:CAS:528:DC%2BC1MXps1OmsQ%3D%3D [DOI: https://dx.doi.org/10.2146/ajhp160319] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28743778]

8. Kalaitzoglou, E; Fowlkes, JL; Popescu, I; Thrailkill, KM. Diabetes pharmacotherapy and effects on the musculoskeletal system. Diabetes Metab Res Rev; 2019; 35, 2 [DOI: https://dx.doi.org/10.1002/dmrr.3100] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30467957]e3100.

9. Mamza, J; Marlin, C; Wang, C; Chokkalingam, K; Idris, I. DPP-4 inhibitor therapy and bone fractures in people with type 2 diabetes—a systematic review and meta-analysis. Diabetes Res Clin Pract; 2016; 116, pp. 288-298.1:CAS:528:DC%2BC28XnsVWlur4%3D [DOI: https://dx.doi.org/10.1016/j.diabres.2016.04.029] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27321347]

10. Neal, B; Perkovic, V; de Zeeuw, D et al. Rationale, design, and baseline characteristics of the Canagliflozin Cardiovascular Assessment Study (CANVAS)—a randomized placebo-controlled trial. Am Heart J; 2013; 166, 2 pp. 217-23e11.1:CAS:528:DC%2BC3sXhtVSmtbfF [DOI: https://dx.doi.org/10.1016/j.ahj.2013.05.007] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23895803]

11. Watts, NB; Bilezikian, JP; Usiskin, K et al. Effects of canagliflozin on fracture risk in patients with type 2 diabetes mellitus. J Clin Endocrinol Metab; 2016; 101, 1 pp. 157-166.1:CAS:528:DC%2BC28XotlWqu7c%3D [DOI: https://dx.doi.org/10.1210/jc.2015-3167] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26580237]

12. Zinman, B; Wanner, C; Lachin, JM et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med; 2015; 373, 22 pp. 2117-2128.1:CAS:528:DC%2BC28XntVSkurs%3D [DOI: https://dx.doi.org/10.1056/NEJMoa1504720] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26378978]

13. Wiviott, SD; Raz, I; Bonaca, MP et al. Dapagliflozin and cardiovascular outcomes in type 2 diabetes. N Engl J Med; 2019; 380, 4 pp. 347-357.1:CAS:528:DC%2BC1MXit12rsLY%3D [DOI: https://dx.doi.org/10.1056/NEJMoa1812389] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30415602]

14. Azharuddin, M; Adil, M; Ghosh, P; Sharma, M. Sodium-glucose cotransporter 2 inhibitors and fracture risk in patients with type 2 diabetes mellitus: a systematic literature review and Bayesian network meta-analysis of randomized controlled trials. Diabetes Res Clin Pract; 2018; 146, pp. 180-190.1:CAS:528:DC%2BC1cXitVKhsLzP [DOI: https://dx.doi.org/10.1016/j.diabres.2018.10.019] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30389620]

15. Perkovic, V; Jardine, MJ; Neal, B et al. Canagliflozin and renal outcomes in type 2 diabetes and nephropathy. N Engl J Med; 2019; 380, 24 pp. 2295-2306.1:CAS:528:DC%2BC1MXhtF2ru77O [DOI: https://dx.doi.org/10.1056/NEJMoa1811744] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30990260]

16. Wang, X; Zhang, F; Zhang, Y et al. Effect of SGLT2 inhibitors on fractures, BMD, and bone metabolism markers in patients with type 2 diabetes mellitus: a systematic review and meta-analysis. Osteoporos Int; 2023; 34, 12 pp. 2013-2025.1:CAS:528:DC%2BB3sXhvFSntLbI [DOI: https://dx.doi.org/10.1007/s00198-023-06908-2] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37695339]

17. Suzuki, M; Honda, K; Fukazawa, M et al. Tofogliflozin, a potent and highly specific sodium/glucose cotransporter 2 inhibitor, improves glycemic control in diabetic rats and mice. J Pharmacol Exp Ther; 2012; 341, 3 pp. 692-701.1:CAS:528:DC%2BC38XotlWlt7g%3D [DOI: https://dx.doi.org/10.1124/jpet.112.191593] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22410641]

18. Hans, D; Goertzen, AL; Krieg, MA; Leslie, WD. Bone microarchitecture assessed by TBS predicts osteoporotic fractures independent of bone density: the Manitoba study. J Bone Miner Res; 2011; 26, 11 pp. 2762-2769. [DOI: https://dx.doi.org/10.1002/jbmr.499] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21887701]

19. Beck, TJ. Extending DXA beyond bone mineral density: understanding hip structure analysis. Curr Osteoporos Rep; 2007; 5, 2 pp. 49-55. [DOI: https://dx.doi.org/10.1007/s11914-007-0002-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17521505]

20. MacNeil, JA; Boyd, SK. Accuracy of high-resolution peripheral quantitative computed tomography for measurement of bone quality. Med Eng Phys; 2007; 29, 10 pp. 1096-1105. [DOI: https://dx.doi.org/10.1016/j.medengphy.2006.11.002] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17229586]

21. Sornay-Rendu, E; Boutroy, S; Duboeuf, F; Chapurlat, RD. Bone Microarchitecture Assessed by HR-pQCT as Predictor of Fracture Risk in Postmenopausal Women: The OFELY Study. J Bone Miner Res; 2017; 32, 6 pp. 1243-1251.1:CAS:528:DC%2BC2sXhtVWjs7bJ [DOI: https://dx.doi.org/10.1002/jbmr.3105] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28276092]

22. Mikolajewicz, N; Bishop, N; Burghardt, AJ et al. HR-pQCT measures of bone microarchitecture predict fracture: systematic review and meta-analysis. J Bone Miner Res; 2020; 35, 3 pp. 446-459. [DOI: https://dx.doi.org/10.1002/jbmr.3901] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31643098]

23. Manske, SL; Zhu, Y; Sandino, C; Boyd, SK. Human trabecular bone microarchitecture can be assessed independently of density with second generation HR-pQCT. Bone; 2015; 79, pp. 213-221.1:CAS:528:DC%2BC2MXht1KrsLzO [DOI: https://dx.doi.org/10.1016/j.bone.2015.06.006] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26079995]

24. Yamamoto, K; Uchida, S; Kitano, K et al. TS-071 is a novel, potent and selective renal sodium-glucose cotransporter 2 (SGLT2) inhibitor with anti-hyperglycaemic activity. Br J Pharmacol; 2011; 164, 1 pp. 181-191.1:CAS:528:DC%2BC3MXhtVOlu7%2FM [DOI: https://dx.doi.org/10.1111/j.1476-5381.2011.01340.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21410690][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3171870]

25. Galigutta RR, Hasik PN, Thomas C, Undela K. Efficacy and safety of luseogliflozin in patients with type 2 diabetes mellitus: a systematic review and meta-analysis. Endocrine. 2024.

26. Johannesdottir, F; Putman, MS; Burnett-Bowie, SM; Finkelstein, JS; Yu, EW; Bouxsein, ML. Age-related changes in bone density, microarchitecture, and strength in postmenopausal black and white women: the SWAN Longitudinal HR-pQCT Study. J Bone Miner Res; 2022; 37, 1 pp. 41-51.1:CAS:528:DC%2BB3MXis1OrsbjI [DOI: https://dx.doi.org/10.1002/jbmr.4460] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34647644]

27. Haraguchi, A; Shigeno, R; Horie, I et al. The effect of luseogliflozin on bone microarchitecture in older patients with type 2 diabetes: study protocol for a randomized controlled pilot trial using second-generation, high-resolution, peripheral quantitative computed tomography (HR-pQCT). Trials; 2020; 21, 1 379.1:CAS:528:DC%2BB3cXos1Gnu7g%3D [DOI: https://dx.doi.org/10.1186/s13063-020-04276-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32370806][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7201752]

28. Julious, SA. Sample size of 12 per group rule of thumb for a pilot study. Pharm Stat; 2005; 4, 4 pp. 287-291. [DOI: https://dx.doi.org/10.1002/pst.185]

29. Pistoia, W; van Rietbergen, B; Lochmüller, EM; Lill, CA; Eckstein, F; Rüegsegger, P. Estimation of distal radius failure load with micro-finite element analysis models based on three-dimensional peripheral quantitative computed tomography images. Bone; 2002; 30, 6 pp. 842-848.1:STN:280:DC%2BD38zntVyitg%3D%3D [DOI: https://dx.doi.org/10.1016/S8756-3282(02)00736-6] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/12052451]

30. Janghorbani, M; Van Dam, RM; Willett, WC; Hu, FB. Systematic review of type 1 and type 2 diabetes mellitus and risk of fracture. Am J Epidemiol; 2007; 166, 5 pp. 495-505. [DOI: https://dx.doi.org/10.1093/aje/kwm106] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17575306]

31. Martinez-Montoro JI, Garcia-Fontana B, Garcia-Fontana C, Munoz-Torres M. Evaluation of quality and bone microstructure alterations in patients with type 2 diabetes: a narrative review. J Clin Med. 2022;11(8).

32. Sakai, S; Kaku, K; Seino, Y et al. Efficacy and safety of the SGLT2 inhibitor luseogliflozin in Japanese patients with type 2 diabetes mellitus stratified according to baseline body mass index: pooled analysis of data from 52-week phase III trials. Clin Ther; 2016; 38, 4 pp. 843-62.e9.1:CAS:528:DC%2BC28XitlSgtLs%3D [DOI: https://dx.doi.org/10.1016/j.clinthera.2016.01.017] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27021608]

33. Burghardt, AJ; Issever, AS; Schwartz, AV et al. High-resolution peripheral quantitative computed tomographic imaging of cortical and trabecular bone microarchitecture in patients with type 2 diabetes mellitus. J Clin Endocrinol Metab; 2010; 95, 11 pp. 5045-5055.1:CAS:528:DC%2BC3cXhsV2nsbnO [DOI: https://dx.doi.org/10.1210/jc.2010-0226] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20719835][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2968722]

34. de Waard, EAC; de Jong, JJA; Koster, A et al. The association between diabetes status, HbA1c, diabetes duration, microvascular disease, and bone quality of the distal radius and tibia as measured with high-resolution peripheral quantitative computed tomography—the Maastricht Study. Osteoporos Int; 2018; 29, 12 pp. 2725-2738. [DOI: https://dx.doi.org/10.1007/s00198-018-4678-3] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30209523][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6267131]

35. Shanbhogue, VV; Hansen, S; Frost, M et al. Compromised cortical bone compartment in type 2 diabetes mellitus patients with microvascular disease. Eur J Endocrinol; 2016; 174, 2 pp. 115-124.1:CAS:528:DC%2BC28Xos1emtr0%3D [DOI: https://dx.doi.org/10.1530/EJE-15-0860] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26537860]

36. Samakkarnthai, P; Sfeir, JG; Atkinson, EJ et al. Determinants of bone material strength and cortical porosity in patients with type 2 diabetes mellitus. J Clin Endocrinol Metab; 2020; 105, 10 pp. e3718-e3729. [DOI: https://dx.doi.org/10.1210/clinem/dgaa388] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32556277][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7458544]

37. Walle, M; Whittier, DE; Frost, M; Muller, R; Collins, CJ. Meta-analysis of diabetes mellitus-associated differences in bone structure assessed by high-resolution peripheral quantitative computed tomography. Curr Osteoporos Rep; 2022; 20, 6 pp. 398-409. [DOI: https://dx.doi.org/10.1007/s11914-022-00755-6] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36190648][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9718715]

38. Taylor, SI; Blau, JE; Rother, KI. Possible adverse effects of SGLT2 inhibitors on bone. Lancet Diabetes Endocrinol; 2015; 3, 1 pp. 8-10.1:CAS:528:DC%2BC2MXksF2ltLo%3D [DOI: https://dx.doi.org/10.1016/S2213-8587(14)70227-X] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25523498]

39. Brandt, IAG; Jessen, MH; Rimestad, DE; Hojgaard, MKF; Vestergaard, P. Advanced glycation end products and bone—how do we measure them and how do they correlate with bone mineral density and fractures? A systematic review and evaluation of precision of measures. Bone; 2022; 165, 1:CAS:528:DC%2BB38Xis1aju7fM [DOI: https://dx.doi.org/10.1016/j.bone.2022.116569] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36174927]116569.

40. Xie, K; Sugimoto, K; Tanaka, M et al. Effects of luseogliflozin treatment on hyperglycemia-induced muscle atrophy in rats. J Clin Biochem Nutr; 2023; 72, 3 pp. 248-255.1:CAS:528:DC%2BB3sXht1yhs7rE [DOI: https://dx.doi.org/10.3164/jcbn.22-58] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37251965][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10209601]

41. Shigiyama, F; Kumashiro, N; Miyagi, M et al. Effectiveness of dapagliflozin on vascular endothelial function and glycemic control in patients with early-stage type 2 diabetes mellitus: DEFENCE study. Cardiovasc Diabetol; 2017; 16, 1 84. [DOI: https://dx.doi.org/10.1186/s12933-017-0564-0] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28683796][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5500953]

42. de Vries, TJ; Kleemann, AS; Jin, J; Schoenmaker, T. The differential effect of metformin on osteocytes, osteoblasts, and osteoclasts. Curr Osteoporos Rep; 2023; 21, 6 pp. 743-749. [DOI: https://dx.doi.org/10.1007/s11914-023-00828-0] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37796390][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10724308]