Improved insulin sensitivity with sodium-glucose cotransporter 2 inhibitor treatment in a patient with slowly progressive type 1 diabetes mellitus with metabolic syndrome: a case report

Abstract

We herein report the clinical course of a 56-year-old Japanese patient with slowly progressive type 1 diabetes mellitus, metabolic syndrome, non-alcoholic fatty liver disease, and severe insulin resistance. The patient’s intravenous glucose tolerance test indicated marked reductions in insulin sensitivity and endogenous insulin secretion. Accordingly, administration of ipragliflozin l-proline, a sodium-glucose cotransporter 2 inhibitor, promoted improvements in insulin sensitivity and blood glucose levels, as well as a decrease in visceral fat, improvement in dyslipidemia, and decrease in hepatic lipid content, suggesting the potential efficacy of sodium-glucose cotransporter 2 inhibitors for obese patients with type 1 diabetes mellitus exhibiting insulin resistance.

Introduction

Slowly progressive type 1 diabetes mellitus (SPT1D), also known as latent autoimmune diabetes in adults (LADA), is characterized by the absence of insulin dependence upon diabetes mellitus (DM) onset and persistence of islet cell autoantibodies, such as glutamic acid decarboxylase autoantibodies (GADA) and insulinoma-associated antigen-2 autoantibodies (IA-2A) [1, 2]. Most patients with SPT1D exhibit a gradual decline in the number of pancreatic β cells and become insulin dependent after a mean period of 3 years [3, 4].

Sodium-glucose cotransporter 2 inhibitors (SGLT2is), a recently developed oral antidiabetic medication that promotes urinary glucose excretion by inhibiting sodium-glucose cotransporter 2, demonstrated efficacy in reducing hemoglobin A1c (HbA1c) levels, glycemic variability, blood pressure, and body weight without intrinsic properties that cause hypoglycemia among individuals with type 1 DM [5].

We herein report a case involving a patient with SPT1D and GADA accompanied by metabolic syndrome (MetS), non-alcoholic fatty liver disease (NAFLD), and severe insulin resistance who harbored type 1 diabetes susceptibility haplotypes of human leukocyte antigens (HLA) DRB1 and DQB1. Accordingly, insulin secretion and sensitivity, body fat distribution, and hepatic lipid content were evaluated before and after SGLT2i administration.

Case report

We evaluated a 56-year-old Japanese woman [height 157 cm, body weight 87.0 kg, and body mass index (BMI) 35.3 kg/m²] who, at 32 years of age, had been determined to have hyperglycemia during a regular medical checkup, was diagnosed with DM through an oral glucose tolerance test, and was determined to be positive for GADA before admission to our institution. Despite treatment with oral antidiabetic agents, her HbA1c (NGSP) levels gradually increased, for which insulin treatment was initiated at 41 years of age. She subsequently visited the Chiba Central Medical Center for further follow-up at 42 years of age.

The patient had a smoking history of 15 pack-years, no history of alcohol consumption, and no family history of DM. Upon admission, her plasma glucose and C-peptide immunoreactivity (CPR) levels were 244 mg/dL and 3.56 ng/mL, respectively. Meanwhile, her GADA titers determined by radioimmunoassay (RIA) or enzyme-linked immunosorbent assay (ELISA) at 42, 49, and 56 years of age were 2.8, 3.4 (normal range <1.5 U/mL on RIA), and 124.6 U/mL (normal range < 5.0 U/mL on ELISA), respectively. We could not confirm an increase in GADA levels due to different assay values. Her IA-2A titer at 42 years of age was <0.399 U/mL. The HLA-DR and HLA-DQ alleles were DRB1*04:05/09:01 and DQB1*03:03/04:01, respectively, while the haplotypes were DRB1*04:05-DQB1*04:01 and DRB1*09:01-DQB1*03:03, respectively, which were consistent with the haplotypes and type 1 DM susceptibility profile of the Japanese population [6, 7]. The patient was then diagnosed with SPT1D based on the diagnostic criteria for the same [8]. Her urinary albumin/creatinine ratio (ACR) was 460 μg/mg. Although a renal biopsy at 38 years of age showed no diabetic nephropathy, nephrosclerosis had been observed. At 55 years of age, her estimated glomerular filtration ratio (eGFR) was >70 mL/min/1.73 m² despite increasing ACR (670 μg/mg), with no diabetic retinopathy having been observed. During her first visit, the patient had been taking once-daily doses of candesartan cilexetil (8 mg) and nifedipine (80 mg) for hypertension and atorvastatin calcium hydrate (10 mg) for hypercholesterolemia.

Since her first visit, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyl transpeptidase (γ-GTP) levels had continued to elevate, together with low-density lipoprotein cholesterol (LDL-C) and triglycerides (TG) levels (Table 1).

Table 1.

Laboratory data before and after ipragliflozin L-proline administration

Variables	Before	After
AST (IU/mL)	36 (33–39)	32 (28–42)
ALT (IU/mL)	55 (51–61)	43 (38–53)
γ-GTP (IU/mL)	78 (63–91)	63 (48–74)
LDL-C (mg/dL)	135 (127–144)	132 (117–140)
HDL-C (mg/dL)	60 (58–63)	72 (67–74)
TG (mg/dL)	264 (181–317)	197 (147–238)
Total ketone bodies (μmol/L)	< 154	< 154
eGFR (mL/min/1.73 m2)	73 (69–76)	72 (67–75)
ACR (mg/g creatinine)	670 (590–730)	870 (840–920)

Data are presented as median (interquartile range) during 15 months

AST aspartate aminotransferase, ALT alanine aminotransferase, γ-GTP gamma-glutamyl transpeptidase, LDL-C low density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, TG triglycerides, eGFR estimated glomerular filtration ratio, ACR albumin/creatinine ratio

At 53 years of age, the patient’s cross-sectional abdominal visceral fat area (V) and subcutaneous fat area (S) were measured through computed tomography (CT). Accordingly, her V and S were 220 and 351 cm², respectively, while the ratio of V to S (V/S ratio) was 0.63, suggesting visceral fat accumulation. Thus, the patient was subsequently diagnosed with MetS [9]. Moreover, given her liver to spleen CT value ratio (L/S ratio) of 0.84, the patient was diagnosed with NAFLD based on the diagnostic criteria for the same [10].

Her HbA1c levels remained above 8.5%, requiring gradual increases in daily basal, bolus, and total insulin doses, which reached 140 IU/day at 22 years of age after the initial diagnosis (Fig. 1). Casual CPR levels gradually decreased, with her fasting CPR reaching 0.6 ng/mL at 55 years of age. The patient was unsuccessful with lifestyle modifications and remained morbidly obese (BMI >35 kg/m²). At 55 years of age, an intravenous glucose tolerance test (IVGTT) with 25 g of glucose, followed by a regular intravenous insulin bolus of 5 IU (0.05 IU/kg body weight) added 20 min after the glucose load, was performed to determine pancreatic β cells’ insulin secretory capacity and assess whole-body insulin sensitivity. Insulin secretory capacity was estimated according to the two-compartment model analysis of CPR kinetics [11–13], whereas insulin sensitivity was estimated using insulin-modified minimal-model parameters [14]. A software program developed based on the report by Eaton et al. [15] was used to calculate the CPR secretion rate (CSR), with the sum of CSR from 0 to 5 min normal range 6.8–18.5 ng/mL/min) being used to determine the first phase of CSR (CS1). A minimal-model software program was used to calculate whole-body insulin sensitivity index (Si) based on the algorithm described by Bergman et al. (normal range 2.6–7.6 × 10⁻⁴/min/µU/mL) [14]. Accordingly, our calculations showed markedly decreased fasting CPR levels (0.9 ng/mL) and CS1 (0.24 ng/mL/min), as well as Si (<0.1 × 10⁻⁴/min/µU/mL).

Fig. 1.

Changes in (a) hemoglobin A1c (HbA1c) levels, (b) basal (blue) and bolus (brown) insulin doses, (c) C-peptide immunoreactivity (CPR) levels, and (d) body mass index (BMI) starting from the initial visit. The vertical line indicates the initiation of treatment with a sodium-glucose cotransporter 2 inhibitor (SGLT2i)

Considering that she had been receiving treatment for oligomenorrhea from a gynecologist, a transvaginal ultrasound was performed to assess for potential polycystic ovary syndrome (PCOS) as a concomitant diagnosis. Although polycystic ovaries had been observed, she had a serum testosterone level of 53 ng/dL (the upper limit for females, 45–60 ng/dL) and no history of hirsutism. Thus, a concurrent PCOS diagnosis had not been established. Given that she was negative for insulin receptor autoantibodies, type B insulin resistance syndrome was ruled out. Counter-regulatory hormones, including glucagon, adrenaline, cortisol, and growth hormone, were all within their normal ranges.

At 55 years of age, the SGLT2i ipragliflozin l-proline (100 mg/day orally) was initiated to lower her blood glucose level. This promoted a decline in HbA1c levels, which prompted a reduction in her insulin dose from 140 to 90 IU/day (Fig. 1). Additionally, her BMI started to decline a year after initiating ipragliflozin l-proline, although no increase in blood ketone bodies was observed (Table 1). The patient continued lifestyle modifications (dietary prescription of 1600 kcal/day, including walking after eating) before and after SGLT2i administration.

Interstitial glucose levels immediately before and 1 year after SGLT2i administration were compared using continuous glucose monitoring with a FreeStyle Libre Pro flash glucose monitoring system (Abbott Diabetes Care, Alameda, CA, USA). Accordingly, the patient’s median glucose level markedly decreased after SGLT2i administration despite the decrease in total insulin dose from 138 to 98 IU/day. Based on the international consensus indices [16], her average glucose level and glucose monitoring index decreased from 230 to 167 mg/dL and from 8.8 to 7.3%, respectively, while glucose variability increased from 25 to 34%. Moreover, the percentage of time in range (70–180 mg/dL) increased from 21 to 57%, the percentage of time above range (>180 mg/dL) decreased from 79 to 42%, and the percentage of time below range (<70 mg/dL) slightly increased from 0 to 1%. Such results were consistent with the decrease in HbA1c levels. The IVGTT performed 6 months after initiating SGLT2i treatment to evaluate its effect on insulin secretion and sensitivity revealed that although CS1 remained low (0.32 ng/mL/min), Si (4.7 × 10⁻⁴/min/µU/mL) increased to within normal range, suggesting that insulin sensitivity improved significantly.

To determine the mechanism for the improvement in insulin sensitivity, laboratory data before and after SGLT2i administration (Table 1) were compared. After 15 months of the administration, the patient’s median ALT, γ-GTP, and TG levels tended to be lower, while the median high-density lipoprotein cholesterol (HDL-C) level tended to be higher, although differences were not statistically significant (p > 0.05). Despite receiving atorvastatin, the patient’s LDL-C levels did not change. The total ketone bodies and median eGFR levels did not change, while the median ACR levels tended to be higher (p > 0.05).

After SGLT2i administration at 57 years of age, the patient’s V and S decreased to 172 and 337 cm², respectively, while her V/S ratio decreased to 0.51. Moreover, the patient’s waist circumference decreased from 122 to 117 cm, while her L/S ratio increased to 1.22. These results suggested that the improvement in dyslipidemia, reduction in visceral fat, and decrease in hepatic lipid content after SGLT2i administration contributed to the recovery of insulin sensitivity.

Discussion

The current patient with SPT1D, MetS, and NAFLD had experienced progressive worsening of glycemic control, for which her daily insulin dose was increased together with positive GADA levels. She exhibited morbid obesity (BMI >35 kg/m²), as well as visceral and hepatic fat accumulation. The patient’s IVGTT results suggested severe insulin resistance in addition to decreased insulin secretion. After SGLT2i administration, we observed changes in HbA1c and continuous glucose monitoring test indices, which indicated marked improvements in glycemic control, improvement in dyslipidemia, and decrease in visceral and hepatic lipid contents. Insulin sensitivity returned to normal despite no improvement in insulin secretion.

Most Japanese patients with SPTID experience a gradual decrease in insulin secretion over several years and progression to a deficient state, resulting in insulin dependence [3, 4], while usually maintaining insulin sensitivity. We had previously reported a case wherein a patient with SPT1D exhibited normal insulin sensitivity for more than 10 years [17]. The present case, however, displayed markedly low insulin sensitivity.

Most patients with LADA in Western countries are obese and display hyperinsulinemia [18, 19]. Moreover, reports have shown that patients with LADA have comparable insulin resistance to patients with type 2 DM [20]. Patients with LADA are at significant risk for the metabolic consequences of insulin resistance other than glucose metabolism, which include those associated with MetS. Furthermore, excessive body weight has been considered a strong contributor to the development of LADA. Insulin resistance has been suggested to promote autoimmune DM by increasing insulin demand [21]. Moreover, the association between excessive weight and insulin resistance has been well established [22], with ectopic lipid accumulation and lipotoxicity and proinflammatory cytokine release from visceral fat tissue being the proposed pathways mediating this association [23]. PCOS, type B insulin resistance syndrome, and excessive production of counter-regulatory hormones had been ruled out as underlying causes for insulin resistance in the present case, a result consistent with that observed in patients with LADA in Western countries.

A number of studies across different populations have suggested that intra-abdominal fat increases the risk for insulin resistance and diabetes, independent of the effects of total body obesity in type 2 DM [24, 25]. One study showed that the amount of visceral fat is closely correlated with the amount of fat in the liver [26], indicating that abdominal obesity is a marker of NAFLD and MetS. Empagliflozin has been shown to reduce liver fat and improve ALT levels among patients with type 2 DM and NAFLD [27], dapagliflozin to suppresses potent atherogenic small dense LDL-C and increased HDL2-C, a favorable cardiometabolic marker [28], and luseogliflozin to decrease the visceral fat area 24 weeks after the administration among patients with type 2 DM [29]. These results suggest that SGLT2is are useful for treating type 2 diabetes with MetS. Similarly, our patient with SPT1D showed a decrease in visceral fat and hepatic lipid content, as well as improvement in dyslipidemia, insulin resistance, and glycemic control, after SGLT2i administration. Thus, SGLT2is might be effective in improving insulin resistance and glycemic control among obese patients with SPT1D. This has been the first case report evaluating insulin sensitivity after SGLT2i administration in a patient with SPT1D.

Despite an increase in the absolute risk for diabetic ketoacidosis was demonstrated in type 1 DM, consensus recommendations have indicated the safety of SGLT2i therapy in type 1 DM [30]. Nonetheless, careful attention is necessary during SGLT2i administration among patients with SPT1D.

All procedures conducted herein were in accordance with the ethical standards of the institutional and national committees on human experimentation, as well as with the 1964 Helsinki Declaration and later versions. Informed consent or a substitute thereof was obtained from the patient included in this study.

Funding

Toshiharu Ishizuka, none; Yoshiharu Tokuyama, none; Atsuya Horie, none; Azuma Kanatsuka, none.

Compliance with ethical standards

Conflict of interest

Support: There was no financial sponsor for this study.