Changes in glycemic control and skeletal muscle mass indices after dapagliflozin treatment in individuals with type 1 diabetes mellitus

Abstract Aims/Introduction Dapagliflozin is used for individuals with type 1 diabetes, although the effect of this medication on skeletal muscle mass is not well established. In addition, there are few studies examining the effects of good glycemic control on skeletal muscle mass in type 1 diabetes patients. We investigated changes in glycemic control and skeletal muscle mass with dapagliflozin in individuals with type 1 diabetes, and the association between these changes. Materials and Methods This was a post‐hoc analysis of a multicenter, open‐label, non‐randomized, prospective, interventional study in individuals with type 1 diabetes. The participants received dapagliflozin at 5 mg/day for 4 weeks, and were reviewed before and after treatment. Weight‐ and height‐corrected appendicular skeletal muscle mass (ASM) were calculated as indices of skeletal muscle mass using bioelectrical impedance analysis. Results A total of 36 individuals were included in the analysis. After the 4 weeks of dapagliflozin treatment, ASM/height2 decreased in the body mass index <23 group (P = 0.004). ASM / weight decreased in all men aged >60 years. The change in ASM / weight (%) was negatively correlated with the change in glycated hemoglobin (%;P = 0.023). The change in ASM / height2 (kg/m2) was also positively correlated with the change in time within the glucose range of 70–180 mg/dL (P = 0.036). Conclusion Dapagliflozin treatment of individuals with type 1 diabetes, particularly non‐obese individuals and older men, might result in loss of skeletal muscle mass. However, good glycemic control during treatment might prevent the onset and progression of sarcopenia.


INTRODUCTION
Sarcopenia is the age-related loss of skeletal muscle mass and strength, which is associated with impaired mobility, falls and fractures, a decline in activities of daily living, and increased risk of death 1 .Skeletal muscle also plays a major role in glucose metabolism, and sarcopenia is a recognized complication of diabetes 2 .Individuals with diabetes mellitus, regardless of disease type, are at high risk for the incidence and progression of sarcopenia [3][4][5][6] .With the recent increase in the number of individuals with diabetes globally, there is a need for effective measures to prevent functional impairment and physical disability in these individuals.
The relationship between glycemic control and sarcopenia in individuals with diabetes has been previously noted.For example, hyperglycemia is associated with muscle weakness 7 , and treatment with oral hypoglycemic agents might improve muscle function and inhibit muscle mass loss 8 .In a longitudinal observational study examining the relationship between glycemic control and sarcopenia in type 2 diabetes patients, height-adjusted appendicular skeletal muscle mass (ASM) and walking speed increased in the group whose glycated hemoglobin (HbA1c) level decreased by ≥1% over 1 year, independent of exercise habits 9 .In contrast, few studies have examined whether good glycemic control has a positive effect on sarcopenia in individuals with type 1 diabetes.
Currently, dapagliflozin, a sodium-glucose cotransporter 2 inhibitor (SGLT2i), is recognized as a novel treatment for both type 1 and type 2 diabetes, and is used as an additional therapy to insulin to improve glycemic control and prevent complications of diabetes.However, muscle mass loss in individuals with type 2 diabetes using SGLT2i has also been observed 10,11 .Although SGLT2i medication and accompanying insulin reduction might cause muscle mass loss, there is, in comparison, limited evidence of skeletal muscle mass changes during SGLT2i medication in type 1 diabetes.
In the present study, our group investigated the changes in glycemic control and body composition in individuals with type 1 diabetes after the use of dapagliflozin.We also examined the relationship between glycemic control changes and skeletal muscle changes.

Study design and individuals
We undertook a post-hoc analysis of a multicenter, openlabel, non-randomized, prospective, interventional study, which is named as the RISING-STAR study 12,13 .The study design and patient population have been described previously 12,13 .Briefly, 60 participants with type 1 diabetes between the ages of 20 and 80 years, HbA1c <10.5%, and body mass index (BMI) >18.5 kg/m 2 were enrolled.The individuals were started on a dapagliflozin dose of 5 mg/day administered for 4 weeks.Each received a 10% reduction in insulin dose before dapagliflozin administration, and was able to titrate both basal and bolus insulin according to the given algorithm throughout the study.All participants were instructed to carry out home selfketone measurements every morning, and were instructed to stop taking dapagliflozin and consult the investigators if the measurements exceeded 600 lm/L 12 .The primary measured outcome was the daily frequency of hypoglycemia during the intervention period.Severe hypoglycemia is defined as the presence of hypoglycemic symptoms that cannot be managed by self alone and a venous plasma glucose level <60 mg/dL or capillary whole blood <50 mg/dL at onset, detection or visit 14 .Excluding the four individuals who withdrew from the study, of the remaining 56 participants, 36 whose body composition could be measured before and after the intervention were included in the post-hoc analysis.The 36 participants were all able to take dapagliflozin continuously for the 4-week trial period.The RISING-STAR study was registered with the Japan Registry of Clinical Trials (jRCTs051190114), and approved by the ethics committees of the Kyoto Prefectural University of Medicine (CRB5200001) and complied with the Declaration of Helsinki.All participants gave written informed consent.

Measurement
Blood tests were carried out at the baseline and at week 4 (referring to the last observation day, 4 weeks after the intervention) to measure blood glucose and HbA1c.BMI and body composition were measured using Inbody 720 (InBody Co., Ltd., Seoul, Korea), a non-invasive body composition analyzer, at the baseline and at week 4. Other than body fat percentage (total fat mass / body weight [%]) and ASM (kg), weightadjusted ASM (ASM / bodyweight [%]) and height-adjusted ASM (ASM / height 2 [kg/m 2 ]) were calculated as indicators of skeletal muscle mass.The study participants measured their blood glucose levels continuously at home using the Freestyle Libre system (Abbott Japan Co. Ltd.).The frequency of time (%) spent above range (TAR; blood glucose level >180 mg/dL), below range (TBR; <70 mg/dL) and within range (TIR; 70-180 mg/dL) was evaluated for 4 weeks before and after the intervention.The average insulin dose for the 3 days before the baseline and week 4 was used to determine the daily insulin use at each time point.

Statistical analysis
Baseline characteristics were summarized as frequency (%) for categorical variables and as meanstandard deviation (SD) for continuous variables.The associations between baseline skeletal muscle mass and other measured variables were examined using unpaired t-tests or Pearson's correlation coefficients.Changes in glycemic control and body composition from the baseline to week 4 were examined using paired t-tests.Participants were stratified by age, and body composition changes between groups were examined using the Mann-Whitney test.Correlations between variables were examined by calculating Pearson's correlation coefficient or Spearman's rank correlation coefficients.We also carried out the stratified analysis based on BMI ≥23.0 kg/m 2 , which has been proposed as a cut-off for the diagnosis of overweight in Asian people 15 .This definition of overweight has often been used in the Japanese population 16,17 .All statistical analyses were carried out using JMP software version 14.0.0 (SAS Institute Inc., Cary, NC, USA).

DISCUSSION
ASM / height 2 has been used as an index to evaluate skeletal muscle mass in sarcopenia 18 ; however, this index is highly correlated with BMI, the standard for obesity.The index mainly identifies thinner people as sarcopenic, and might underestimate sarcopenia in overweight people 19 .Therefore, many studies use ASM / weight to assess relative skeletal muscle mass.In the present results, ASM / height 2 was positively correlated with BMI, as previously reported, thus we used both ASM / weight and ASM / height 2 as skeletal muscle mass indices.
We found that ASM / height 2 , but not ASM / weight decreased after starting dapagliflozin medication.The present findings also suggest that the addition of dapagliflozin to individuals with type 1 diabetes, but without obesity, might reduce skeletal muscle mass.In contrast, we could not clarify how dapagliflozin affects muscle mass in individuals with a BMI above a certain level.It has been reported that plasma glucagon concentrations and hepatic glycogenesis increase with energy loss due to urinary glucose excretion in individuals on SGLT2i 20 .Substrates for glycogenesis include glycerol from adipose tissue and glycogenic amino acids from skeletal muscle.In non-obese individuals, body fat as an energy source might be limited and, consequently, skeletal muscle might be prone to catabolism.The present study shows that SGLT2i administration might have an unfavorable effect on skeletal muscle mass in older men with type 1 diabetes.It is widely known that the absolute amount of skeletal muscle mass is higher in men than in women 21 , and this is thought to be due to the influence of sex hormones.Testosterone is related to changes in skeletal muscle mass in men, and plays an important role in muscle hypertrophy and muscle loss prevention by regulating insulin-like growth factor-1, and activating protein synthesis and satellite cells [22][23][24] .It has also been reported that age-related muscle mass loss is greater in men than in women 21,25 .Long-term administration of SGLT2i to older men with type 1 diabetes might result in synergistic loss of muscle mass due to aging during treatment, although the development and progression of sarcopenia during SGLT2i administration requires future research.It is known that the prevalence of sarcopenia is higher in older individuals with type 2 diabetes than in younger people 26 , but such effects are not clear in individuals with type 1 diabetes.In the present study, we found no association between age and skeletal muscle mass in either men or women.
Good glycemic control in individuals with type 1 diabetes has been established to prevent the development of microvascular diseases 27 .In contrast, we have not found any studies that have longitudinally examined the relationship between glycemic control and muscle mass index in type 1 diabetes patients.In the current study, we found that changes in ASM / weight (%) were negatively correlated with changes in TAR and HbA1c, and changes in ASM / height 2 (kg/m 2 ) were further positively correlated with changes in TIR.These results suggest that good glycemic control in type 1 diabetes patients might contribute to the prevention and further progression of sarcopenia.No association was found between sarcopenia and glucose control in a cross-sectional study of individuals with diabetes 28 , and the present results also show no clear association between baseline glycemic control and muscle mass index.Skeletal muscle function is impaired in individuals with type 1 diabetes 4 , and factors linking diabetes and sarcopenia include decreased insulin signals, chronic inflammation, mitochondrial dysfunction and peripheral neuropathy [29][30][31][32] .Furthermore, insulin use contributes to the preservation of muscle mass in type 1 diabetes patients and is thought to prevent sarcopenia 33 .Changes in total insulin (unit/day/kg), basal insulin (unit/day/kg) and bolus insulin (unit/day/kg) were also not associated with changes in skeletal muscle mass.As the number of individuals with neuropathy was only six, a detailed study of the relationship between the presence of neuropathy and skeletal muscle mass is a future priority.
Decreases in bodyweight, BMI and body fat, and increases in TBR, were observed with the start of SGLT2i treatment, and these changes were correlated.The decrease in body fat might represent energy production from adipose tissue due to hypoglycemia 34 .Importantly, when rapid weight loss is observed after SGLT2i medication administration, as well as a positive effect related to improvements in insulin resistance, attention must also be paid to an increased frequency of hypoglycemia.
The key limitations of the present study were that it was a post-hoc analysis based on short-term results with no control participants.Furthermore, we did not collect indicators of muscle function, such as grip strength and walking speed.In contrast, our analysis provides valuable evidence of changes in the two skeletal muscle mass parameters.The study sample size of 36 participants was not enough to generalize the findings from this study.In contrast, it is necessary to accumulate the effects of dapagliflozin in type 1 diabetes patients.Indeed, Mart ınez- Montoro et al. 35 reported the benefit of dapagliflozin combination therapy in a retrospective analysis of 38 participants with type 1 diabetes treated with dapagliflozin at two centers.Both nutritional statuses, including protein intake and the amount of exercise, were not considered.Participants receive no exercise or dietary advice/instruction during the intervention period.Future research should consider such variables to further enhance our understanding of dapagliflozin treatment effects on type 1 diabetes individuals.Overall, the present results show that the use of dapagliflozin might result in the loss of skeletal muscle mass in non-obese individuals and in older men with type 1 diabetes.In contrast, good glycemic control during treatment might prevent the onset and progression of sarcopenia.

Figure 2 |
Figure 2 | Relationships between age and skeletal muscle mass change in men (n = 12).(a) Correlation between age and changes in skeletal muscle mass using Spearman's rank correlation coefficient.(b) Difference in skeletal muscle mass changes between the age <60 years group (n = 7) and the age ≥60 years group (n = 5) using the Mann-Whitney test.The line in the middle of the box indicates the median value; the box extends from the 25th to the 75th percentiles.q, Spearman's rank correlation coefficient.ASM, appendicular skeletal muscle mass; IQR, interquartile range.

Table 1 |
Summary characteristics of the study cohort Data are expressed as frequencies (percentages) for categorical variables or mean (standard deviation) for continuous variables.ASM, Appendicular skeletal muscle mass; BMI, body mass index; HbA1c, glycated hemoglobin; SD, standard deviation; TAR, time spent above range; TBR, time spent below range; TIR, time spent within range.

Table 2 |
Changes (paired t-test) in glycemic control parameters and body composition from the baseline to week 4

Table 3 |
Correlation (Pearson's correlation coefficient) between changes in glycemic control parameters and changes in body composition ASM, appendicular skeletal muscle mass; BMI, body mass index; HbA1c, glycated hemoglobin; Is, insulin; r, Pearson's correlation coefficient; Tar, time spent above range; TBR, time spent below range; TIR, time spent within range.