Study design

This post-hoc sub-analysis was based on data obtained from the UTOPIA trial. The study design of the original UTOPIA trial has been described previously [11]. Briefly, the UTOPIA trial was a 2-year prospective, randomized, open-label trial to evaluate the efficacy of tofogliflozin in preventing the progression of atherosclerosis in patients with T2DM. All randomized patients were followed up until the scheduled end of the study, irrespective of adherence to or discontinuation of study medication for any reason. Clinical and biochemical data were collected at 0, 26, 52, 78, and 104 weeks after randomization. The primary outcomes of the original study were changes in the mean IMT of the common carotid artery.

The protocol of the original study was approved by the Ethical Review Board of Osaka University Hospital (approval number: N18007; date of approval: 8/7/2019), and written informed consent was obtained from all patients after a complete explanation of the study. The protocol of this sub-analysis was also approved by the Ethical Review Board of Osaka University Hospital (approval number: 23382; date of approval: 3/5/2024). As this was a post-hoc analysis using only existing materials, it was considered exempt from the requirement for written informed consent from study participants in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan (partially amended version issued on 2023/3/27). Instead, relevant information regarding the study was open to the public, and opportunities for refusal were ensured. The study was conducted in accordance with the Declaration of Helsinki, Ethical Guidelines for Medical and Health Research Involving Human Subjects, Clinical Trials Act, and other current legal regulations in Japan.

The UTOPIA study was registered with the University Hospital Medical Information Network Clinical Trials Registry, a nonprofit organization in Japan that meets the requirements of the International Committee of Medical Journal Editors (registration number: UMIN000017607).

To avoid bias, data management and monitoring were conducted by a third-party entity (Soiken Inc., Toyonaka, Osaka, Japan).

Study population

The original inclusion criteria were age between 30 and 74 years, presence of T2DM, and a glycated hemoglobin (HbA1c) level of ≥ 6% but < 9%. The exclusion criteria included presence of type 1 or secondary diabetes mellitus; participant being in the perioperative period; presence of a serious infection or injury; history of myocardial infarction, angina, stroke, or cerebral infarction; estimated glomerular filtration rate (eGFR) of < 30 mL/min/1.73 m2 or end-stage renal failure; serious liver function impairment; moderate to severe heart failure; urinary tract or genital infection; pregnancy, possibility of pregnancy, or nursing status; history of hypersensitivity to tofogliflozin; presence or history of a malignant tumor; prohibition from using tofogliflozin; and other reasons for ineligibility determined by an investigator.

Originally, 340 patients who met the eligibility criteria were enrolled at 23 institutions across Japan (Supplementary Material 1) and randomly allocated to either the tofogliflozin group (20 mg of tofogliflozin once daily, n = 169) or the conventional treatment group (those using drugs other than the SGLT2 inhibitor, n = 171). After excluding one patient from further analysis because of lack of data regarding the primary endpoint, 169 and 170 patients in the tofogliflozin and conventional treatment groups, respectively, were included in the full analysis set. Finally, those with a BMI of ≥ 25.0 kg/m2 were included in the present analysis.

Randomization and study intervention

The enrolled participants were randomly assigned to either the tofogliflozin treatment group or the conventional treatment group without SGLT2 inhibitors. In the tofogliflozin group, 20 mg of tofogliflozin once daily was started in addition to ongoing therapy. However, the addition of an alternative antidiabetic agent (excluding another SGLT2 inhibitor) was permitted 12 weeks after randomization. In the conventional treatment group, either the dosage of the ongoing therapy was increased, or a concomitant oral glucose-lowering drug (excluding any SGLT2 inhibitor) was added 12 weeks after randomization. Treatment was continued to achieve the target value specified in the Japanese Treatment Guide for Diabetes [17] (HbA1c < 7.0%) in both groups. In case of hypoglycemia, the dosage of the concomitant oral glucose-lowering drug was titrated.

Study outcomes

Data changes over time in the tofogliflozin and conventional treatment groups were compared within and between groups. The primary outcome was the percentage of patients in each treatment group achieving a weight loss of least 3% from week 0 to 26 and 104 weeks, and the difference between treatment groups. It was reported that, after 6 months of weight loss guidance for individuals with obesity, the greater the weight loss from pre-intervention, the greater the improvement in blood pressure, blood glucose levels, lipid levels, uric acid levels, and liver function, and that a 3% weight loss from the pre-intervention weight was enough to show significant improvement [18]. Hence, the Japan Society for the Study of Obesity (JASSO) Obesity Treatment Guidelines set the weight loss goal for obesity at 3% of current body weight over a 3–6 month period [19]. Therefore, the primary endpoint of this study was also set at 3% weight loss at 6 months (26 weeks) after the start of the intervention.

Secondary outcomes included achievement rate of HbA1c levels < 7%, systolic blood pressure < 130 mmHg, diastolic blood pressure < 80 mmHg, low-density lipoprotein cholesterol (LDL-C) levels < 120 mg/dL, triglyceride levels < 150 mg/dL, HDL-C levels of ≥ 40 mg/dL, and uric acid levels of 7 mg/dL or less at 0, 26, 52, 78, and 104 weeks in each treatment group; and intra- and inter-group comparisons for changes over time regarding each item (HbA1c, weight, BMI, abdominal circumference, blood pressure, etc.) at 0, 26, 52, 78, and 104 weeks. The incidence of cardiovascular events (ischemic heart disease, cerebrovascular disease, and arteriosclerosis obliterans) and adverse events were also compared between groups.

Clinical and biochemical assessment

Height, weight, and waist circumference were measured, and BMI was calculated. The determination of obesity (BMI ≥ 25.0 kg/m2) was based on the JASSO criteria [16]. Blood pressure was measured at rest with a mercury sphygmomanometer. Blood samples were collected after an overnight fast. HbA1c, glucose, serum lipids (total cholesterol, HDL-C, LDL-C, and triglycerides), creatinine, aspartate transaminase (AST), alanine transaminase (ALT), gamma-glutamyl transpeptidase (γGTP), uric acid, red blood cells (RBCs), Hb, hematocrit (Ht), white blood cells (WBC), and total platelet counts (PLT) were measured using standard techniques. Urinary albumin excretion was measured by the improved bromocresol purple method using a spot urine sample. The eGFR was calculated using the following formula: eGFR (mL/min per 1.73 m2) = 194 × age − 0.287 × serum creatinine − 0.1094 (× 0.739 for females).

Assessment of subclinical atherosclerosis

Carotid IMT, brachial-ankle pulse wave velocity (baPWV), and ankle-brachial index (ABI) were evaluated as indicators of subclinical atherosclerosis.

Ultrasonography scans of the carotid artery were performed by an expert (specifically-trained sonographers) based on the Japan Society of Ultrasonics in Medicine guidelines [20]. Scanning of the extracranial carotid artery was performed bilaterally in different projections, and the site of greatest thickness, including plaque lesions, was sought along the arterial walls. IMT was measured as the distance between two parallel echogenic lines corresponding to the vascular lumen and the adventitial layer. To avoid inter-reader variability, all scans were stored electronically and sent to the IMT evaluation committee. The sent images were inspected in random order by experienced investigators who were not informed of the clinical characteristics of the participants, using automated digital edge detection software (Intimascope; MediaCross, Tokyo, Japan) [21]. The software system averaged approximately 200 points of the IMT values in the segment 2 cm proximal to the dilation of the carotid sinus (mean-IMT-CCA). Additionally, the maximum thicknesses of the intima and media layers, including the plaque lesions, in the common carotid arteries (max-IMT-CCA) were captured separately.

The baPWV and ABI were measured using the same volume plethysmography apparatus (BP-203RPE II form PWV/ABI, Omron Healthcare Co., Ltd., Kyoto, Japan), with patients in the supine position after at least 5 min of rest, as previously reported [22]. Specifically, four oscillometric cuffs, each connected to a plethysmographic sensor that determined the volume of the pulse from and to an oscillometric pressure sensor that measured blood pressure, were wrapped on both the brachia and ankle; two electrocardiogram electrodes were placed on each wrist. The cuffs were simultaneously pressurized to the approximate value of the patient’s diastolic pressure such that the pulse volume waveforms could be recorded using semiconductor pressure sensors. The distance between the sampling points of the baPWV was calculated automatically according to the height of the participant. The path length from the suprasternal notch to the ankle (La) was calculated as follows: La = 0.8129 × height (in cm) + 12.328. The path length from the suprasternal notch to the brachium (Lb) was calculated as follows: Lb = 0.2195 × height − 2.0734. The baPWV was calculated according to the following formula: baPWV = (La − Lb)/Tba, where Tba was the time interval between the wavefront of the brachial waveform and that of the ankle waveform [22]. Two simultaneous measurements of baPWV were recorded on the right and left sides.

Assessment of diabetes therapy-related QOL

In the current study, we used the Diabetes Therapy-Related Quality Of Life questionnaire (DTR-QOL) 7, a short version of the original DTR-QOL, which comprised seven questions selected from the original 29 items. The total score, after simple addition of the item scores except the Q2 score, was converted to 0–100 (best-case response = 100; worst-case response = 0), with a higher total score after conversion reflecting better treatment satisfaction. The Q2 score, which reflected weight gain with treatment, was separately evaluated, with a higher score indicating better treatment satisfaction. Details are provided in Supplemental Material 2.

Safety evaluation

All adverse events were recorded during the study and are described in Supplemental Material 3.

Statistical analysis

All enrolled patients were included in the analysis. Differences between treatment groups regarding the percentage of patients achieving at least 3% weight loss from week 0 to 26 and 104 weeks and in the percentage achieving each treatment goal were evaluated with the χ-squared test, and tests of significance (intra-group comparison) in the percentage achieving goals at each observation point from baseline were evaluated with the McNemar test. Similarly, differences between treatment groups in the percentage of achievement of each secondary outcome (HbA1c levels, blood pressure, serum lipid levels, etc.) at each observation time point were evaluated by the χ-squared test. Moreover, regarding secondary outcomes, the Student’s t test was used to test the null hypothesis that the differences in the measured values and changes in each endpoint in the two groups were equal. The 95% confidence intervals (CIs), based on the t-distribution of the difference between the measured values and the amount of change for each endpoint, were calculated. If the distribution of the measured values deviated significantly from the normal distribution, a t-test or Wilcoxon rank-sum test was performed after log transformation. The significance of change over time (change from baseline) within a treatment group was tested with a one-sample t test. The significance level was set at a two-sided p-level of 0.05, and no multiplicity adjustment was performed. All analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC, USA).