FormalPara Key Summary Points

Why carry out this study?

 No study has compared the effects of sodium–glucose cotransporter 2 inhibitors (SGLT2is) and dipeptidyl peptidase 4 inhibitors (DPP4is) on patients’ quality of life (QOL).

 This study enrolled drug-naïve patients with type 2 diabetes mellitus and compared the effects of dapagliflozin, an SGLT2i, and DPP4is on patients’ QOL.

What was learned from the study?

 Dapagliflozin showed a comparable or more favorable impact on patients’ QOL than DPP4is.

 The reduction in body weight was significantly correlated with improvement in patients’ QOL by dapagliflozin.

 The results of this study demonstrate that dapagliflozin can be used as a first-line drug for T2DM, with a beneficial impact on patients’ QOL.

Digital Features

This article is published with digital features, including a summary slide, to facilitate understanding of the article. To view digital features for this article go to https://doi.org/10.6084/m9.figshare.13042037.

Introduction

The American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) recommend metformin as first-line therapy for type 2 diabetes mellitus (T2DM) with comprehensive lifestyle management [1, 2]. Other oral hypoglycemic agents (OHAs), such as sodium–glucose cotransporter 2 inhibitors (SGLT2is) and glucagon-like peptide 1 (GLP-1) receptor agonists, are to be used as second-line or further lines of therapy on the basis of cardiovascular or renal risks and glycemic control [1, 2]. In contrast, the Japanese Clinical Practice Guideline for Diabetes does not define any first-line drug; it recommends the selection of antidiabetic agents on the basis of the individual’s pathological condition and risk of hypoglycemia [3, 4]. Because of their effectiveness and low risk of hypoglycemia [5], especially in East Asian populations [6], dipeptidyl peptidase 4 inhibitors (DPP4is) have become the most commonly used OHAs in Japan [7]. However, studies have reported benefits of SGLT2is such as weight reduction [8, 9] as well as cardiovascular and renal protection [10,11,12,13], with low risk of hypoglycemia [14, 15]. A clinical trial comparing SGLT2is and DPP4is reported greater reduction in fasting plasma glucose and body weight in the SGLT2i group than in the DPP4i group, although the frequency of hypoglycemic events was similar in both groups [16].

A “patient-centered approach” [1, 2, 17], as recommended by the ADA/EASD, is important for developing a treatment strategy and should consider medication preference and quality-of-life (QOL) of patients. Although both DPP4i [18, 19] and SGLT2i [20,21,22,23] improve treatment satisfaction and QOL, no study has compared the effects of DPP4is and SGLT2is on patients’ QOL. Thus, this Japanese T2D study for evaluating benefits of dapagliflozin, as new first-line therapy, to improve QOL (J-BOND study) enrolled drug-naïve patients with T2DM who had been newly initiated on OHA therapy and compared the effects of dapagliflozin, an SGLT2i, and DPP4is on their QOL. Since many kinds of DPP4is are used in real-world situations, all once-daily DPP4is were used as the control agents in this study, and the effects of dapagliflozin and all kinds of once-daily DPP4is on the QOL were compared.

Methods

Study Design

The J-BOND study was a prospective, randomized, open-label, blinded-endpoint trial, conducted across 42 medical institutions (Supplementary Table 1) in Japan between February 2018 and January 2020. This study and its protocols were first approved by the institutional review board of each participating institution (Supplementary Material) according to the Ethical Guidelines for Medical and Health Research Involving Human Subjects issued by the Ministry of Health, Labour and Welfare in Japan. Following enforcement of the Clinical Trials Act in April 2018, this study and its protocols were again inspected and approved by the Nara Medical University Certified Review Board, which had obtained certification from the Minister of Health, Labour and Welfare in Japan. This study was first registered in the University Hospital Medical Information Network Clinical Trial Registry (UMIN-CTR) (registration number UMIN000030514) and then in the Japan Registry of Clinical Trials (jRCT) (registration number jRCTs051180165) after approval from the certified review board according to the Clinical Trials Act. The study was conducted in accordance with the Declaration of Helsinki, Ethical Guidelines for Medical and Health Research Involving Human Subject, the Clinical Trials Act, and other current legal regulations in Japan. Written informed consent was obtained prior to treatment from all enrolled patients who met the eligibility criteria. All authors had access to the study data and reviewed and approved the final manuscript.

Patient Population

Patients with T2DM who were initiating therapy with OHAs were included in this study. The inclusion criteria were as follows: (1) patients with T2DM who had newly started therapy with OHAs, (2) HbA1c ≥ 6.5% at the time of consent, (3) body mass index (BMI) ≥ 23 kg/m2, (4) 20 years ≤ age < 75 years, and (5) written consent. The exclusion criteria were as follows: (1) history of any antidiabetic medication within the last 3 months, (2) history of severe hypoglycemia within a year, (3) type 1 diabetes mellitus or secondary diabetes, (4) perioperative period or with severe infection or severe physical injury, (5) moderate to severe heart failure (New York Heart Association at class III or higher), (6) moderate-to-severe renal disease (estimated glomerular filtration rate < 45 mL/min/1.73 m2), (7) severe liver disease (aspartate aminotransferase ≥ 100 IU/L), (8) addiction to alcohol or a drug, (9) possibly pregnant, pregnant, or planning to be pregnant during the study period, (10) dementia, (11) contraindication of the study drug, and (12) other conditions that the investigator/researcher considered inappropriate for the study.

Randomization and Study Intervention

After obtaining informed consent, eligible patients were randomly assigned to either the dapagliflozin group or the DPP4i group at a ratio of approximately 1:1. Randomization was performed using a computer-based dynamic allocation method with a minimization procedure to balance two allocation factors (HbA1c and BMI) across the groups. Subjects assigned to the dapagliflozin and DPP4i groups started to use 5 mg of dapagliflozin per day or DPP4i (sitagliptin, vildagliptin, alogliptin, linagliptin, teneligliptin, anagliptin, or saxagliptin), respectively, at an initial dose defined in the package insert of each product. Since many kinds of DPP4is are used in real-world situation, all once-daily DPP4is were used as the control agents in this study. To compare medication adherence and preference, once-weekly DPP4is and twice-daily DPP4is were not included in this study. Subjects were followed up for 24 weeks, with observation time points at baseline, week 12, and week 24.

Study Outcomes

The primary endpoint was the proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24. SHIELD-WQ-9 (Study to Help Improve Early evaluation and management of risk factors Leading to Diabetes Weight Questionnaire-9) is a specific questionnaire to assess health-related QOL associated with weight change to determine how weight change has affected nine aspects of daily life: physical health, interaction with family, work performance, interaction with coworkers/friends, social activities, self-esteem, daily activities, emotional health, and overall QOL [24]. Responses to each domain include four options: worsened, improved, stayed the same, and not applicable. The secondary endpoints included the proportion of patients indicating improvement in other domains (physical health, interactions with family, work performance, interactions with coworkers/friends, social activities, daily activities, self-esteem, and emotional health) of SHIELD-WQ-9 at week 24, change in Diabetes Therapy-Related QOL (DTR-QOL) and EuroQol five-dimensional five level version (EQ-5D-5L) scores, medication preference, medication adherence, diet therapy adherence, body weight, BMI, abdominal circumference, HbA1c, and frequency of adverse events. DTR-QOL is a diabetes-specific questionnaire with four domains: domain 1, burden on social activities and daily activities; domain 2, anxiety and dissatisfaction with treatment; domain 3, hypoglycemia; and domain 4, treatment satisfaction [25]. The each item in the DTR-QOL is a 7-point Likert scale (1, strongly agree; 7, strongly disagree). The score of each item was reversed so that 7 represented the highest QOL. The total score and each domain score was calculated by addition scores of the attribute items and the conversion of scoring range to 0–100, so that 0 and 100 represents the lowest and highest QOL, respectively. EQ-5D-5L is a generic questionnaire to assess health-related QOL in five dimensions—mobility, self-care, usual activities, pain/discomfort, and anxiety/depression—with five levels of severity: no problems, slight problems, moderate problems, severe problems, and extreme problems [26]. The answer to each dimension is converted to a 5-point Likert scale (1, no problems; 5, extreme problems), and EQ-5D-5L score is calculated using standard value sets for Japanese EQ-5D-5L based on the composite time trade-off valuation method [27]. Higher EQ-5D-5L score indicates better QOL. Medication preference was investigated by a questionnaire asking “prefer SGLT2i”, “prefer DPP4i”, or “neither” at each visit after full explanation of advantages and disadvantages of SGLT2i and DPP4i from the investigators. Medication adherence for the study agent throughout the study was calculated as the frequency of actual administration relative to the total planned administration. Diet therapy adherence was investigated by a questionnaire asking “1. not conducted at all”, “2. rarely conducted”, “3. conducted about half of planned”, “4. almost conducted”, or “5. perfectly conducted”. The numbering of each option was used as the score for diet therapy adherence. The patient-reported outcomes (PROs) (SHIELD-WQ-9, DTR-QOL, EQ-5D-5L, medication preference, medication adherence, and diet therapy adherence) were answered by the study subjects by themselves on the paper questionnaire. Other clinical outcomes were collected by the paper case report forms from the investigators.

Sample Size Calculation and Statistical Analysis

On the basis of the results of a previous study [20], we assumed the proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 to be 29.3% in the dapagliflozin group in this study, which was comparable with that in the dapagliflozin group in the previous study, and 14.0% in the DPP4i groups in this study, which was comparable with that in the placebo group in the previous study. Under this assumption, 113 subjects per group would provide a power of over 80% to detect an intergroup difference using a two-sided t test at 5% significance. The dropout rate was estimated to be approximately 10%. Thus, 126 subjects were required per group, yielding a total sample size of 252.

Analyses for the primary and secondary endpoints were performed on the full analysis set (FAS), which included all subjects assigned to a study intervention. However, subjects with a significant study protocol violation (e.g., registration without consent or registration out of the enrollment period) were excluded from the FAS. Sensitivity analysis was performed with the per-protocol set by excluding subjects with a protocol violation such as violation of eligibility criteria, use of prohibited drugs, or poor medication adherence to the study agent or control agent (< 75% or > 120%). Safety analysis included all treated subjects. All tests were two-sided, and a p < 0.05 was considered statistically significant. The number and proportion of subjects who met the criteria were calculated for analysis of the primary endpoint, and the chi-square test was performed for intergroup comparison. Summary statistics for measurement and change from baseline were calculated for the analysis of secondary endpoints, and one-sample t test for intragroup comparison and two-sample t test for intergroup comparison were performed for continuous variables. The chi-square test or Wilcoxon signed-rank test was performed for categorical variables. Medication adherence for the study agent throughout the study was calculated as the frequency of actual administration relative to the total planned administration. Summary statistics were calculated, and Fisher’s exact test was used for intergroup comparisons of frequency of adverse events. Correlation analysis was performed with Pearson’s correlation coefficient and Spearman's rank correlation coefficient. The SAS statistical software package (version 9.4, SAS, Cary, NC, USA) was used to perform all statistical analyses. To avoid bias and ensure quality, data collection, data management, and statistical analyses were performed by third-party entities (Soiken Inc., Osaka, Japan).

Results

Baseline Characteristics of Study Participants

A total of 4247 patients were screened, and 3994 patients were excluded from the study. Of these 3994 patients, 3842 did not meet the eligibility criteria, 100 did not provide consent, and 52 were excluded for other reasons. Thus, 253 subjects were registered and randomly allocated to the treatment groups. Although the target sample size was set to 252, as described in the “Methods”, because the last two eligible candidates were registered at the same time, both were included in this study. This issue was reported to and approved by the certified review board. Because of consent withdrawal or visit discontinuation before the start of the study intervention, two and three patients were excluded from the dapagliflozin and DPP4i groups, respectively. As a result, 124 subjects were included in the safety analysis set, and both groups were considered as the FAS (Fig. 1). Of these, because one withdrew consent, and eight in the dapagliflozin group and four in the DPP4i group discontinued the study, 235 subjects (115 subjects in the dapagliflozin group and 120 subjects in the DPP4i group) completed the study.

Fig. 1
figure 1

Study flow chart showing subject enrollment, allocation, and analysis. DPP4i dipeptidyl peptidase inhibitor

Full size image

The baseline characteristics of the subjects are summarized in Table 1. No significant intergroup differences in baseline characteristics were found, except in the proportion of smokers.

Table 1 Baseline characteristics of subjects
Full size table

Patient-Reported Outcomes

SHIELD-WQ-9

SHIELD-WQ-9 is a specific questionnaire to assess health-related QOL associated with weight change to determine how weight change has affected nine aspects of daily life: physical health, interaction with family, work performance, interaction with coworkers/friends, social activities, daily activities, self-esteem, emotional health, and overall QOL [24]. The number of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24 was 31 (28.4%) and 22 (18.6%) in the dapagliflozin and DPP4i groups, respectively. Although the proportion was higher in the dapagliflozin group than in the DPP4i group, no significant intergroup difference was observed (p = 0.08) (Table 2). Additionally, the proportion of subjects indicating improvement in the “physical health” domain at week 24 was significantly higher in the dapagliflozin group (46, 42.2%) than in the DPP4i group (28, 23.7%) (p = 0.004). Other domains did not show any significant intergroup difference.

Table 2 Proportion of subjects indicating improvement in each domain of SHIELD-WQ-9 at week 24
Full size table

DTR-QOL

DTR-QOL is a diabetes-specific questionnaire with four domains: domain 1, burden on social activities and daily activities; domain 2, anxiety and dissatisfaction with treatment; domain 3, hypoglycemia; and domain 4, treatment satisfaction [25]. DTR-QOL total scores significantly increased from baseline to week 24 in both groups: changes from baseline were 14.3 ± 15.6 (p < 0.001) and 10.2 ± 15.6 (p < 0.001) in the dapagliflozin and DPP4i groups, respectively (Table 3). The degree of increase tended to be higher in the dapagliflozin group than in the DPP4i group (p = 0.05), although this was not statistically significant. All scores of domains 1 to 4 significantly increased from baseline to week 24 in both groups (all p values for intragroup comparisons < 0.001 in all domains in both groups). The degree of increase in domain 1 scores tended to be higher in the dapagliflozin group than in the DPP4i group (p = 0.05); however, this did not reach statistical significance. No other significant differences or tendency of difference were observed in intergroup comparisons of the domain scores.

Table 3 DTR-QOL scores
Full size table

EQ-5D-5L

EQ-5D-5L is a generic questionnaire to assess health-related QOL in five dimensions—mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [26]. EQ-5D-5L scores increased significantly from baseline to week 24 in the DPP4i group, but not in the dapagliflozin group: changes from baseline were 0.023 ± 0.088 (p = 0.005) and 0.005 ± 0.096 (p = 0.57) in the DPP4i and dapagliflozin groups, respectively (Table 4). However, no significant intergroup difference was observed (p = 0.14). EQ-5D-5L visual analogue scale (VAS) increased significantly from baseline to week 24 in both groups: changes from baseline were 4.9 ± 16.1 (p = 0.002) and 5.4 ± 14.1 (p < 0.001) in the dapagliflozin and DPP4i groups, respectively. No significant intergroup difference was observed (p = 0.81).

Table 4 EQ-5D-5L
Full size table

Medication Preference, Medication Adherence, and Diet Therapy Adherence

The medication preference differed between the groups at every visit (p = 0.007 and < 0.001 at baseline and week 24, respectively) (Table 5). The proportion of subjects indicating “prefer assigned study agent” increased in both groups. Medication adherence throughout the study was comparable between the groups: 91.9 ± 14.8% and 94.1 ± 9.6% in the dapagliflozin and DPP4i groups, respectively (p = 0.16). Both at baseline and week 24, the median of diet therapy adherence was 3, and no significant intergroup difference was observed (p = 0.74 and 0.56 at baseline and week 24, respectively).

Table 5 Medication preference, medication adherence, and diet therapy adherence
Full size table

Clinical Outcomes

Body weight significantly decreased in the dapagliflozin group, but not in the DPP4i group from baseline to week 24: change from baseline was −2.6 ± 2.5 kg (p < 0.001) and 0.3 ± 3.3 kg (p = 0.39) in the dapagliflozin and DPP4i groups, respectively (Table 6). The decrease in body weight was significantly greater in the dapagliflozin group than in the DPP4i group (p < 0.001). Similarly, BMI decreased significantly in the dapagliflozin group, but not in the DPP4i group from baseline to week 24: change from baseline was − 1.0 ± 0.9 kg/m2 (p < 0.001) and −0.1 ± 1.2 kg/m2 (p = 0.42) in the dapagliflozin and DPP4i groups, respectively. The decrease in BMI was significantly greater in the dapagliflozin group (p < 0.001). Abdominal circumference decreased significantly in both groups from baseline to week 24: changes from baseline were − 2.7 ± 4.3 cm (p < 0.001) and −1.2 ± 4.8 cm (p = 0.008) in the dapagliflozin and DPP4i groups, respectively. The decrease in abdominal circumference was significantly greater in the dapagliflozin group (p = 0.019).

Table 6 Clinical outcomes
Full size table

HbA1c decreased significantly in both groups from baseline to week 24: changes from baseline were − 0.8 ± 1.1% (p < 0.001) and − 0.9 ± 1.4% (p < 0.001) in the dapagliflozin and DPP4i groups, respectively. No significant intergroup difference was observed in the change in HbA1c (p = 0.39).

Correlation Analysis

The proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24 was significantly correlated with a decrease in body weight from baseline to week 24 (Spearman's rank correlation coefficient ρ = − 0.29, 95% confidence interval (CI) − 0.40 to − 0.16, p < 0.001) (Table 7). The proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24 was also significantly correlated with a decrease in HbA1c level from baseline to week 24 (ρ = − 0.20, 95% CI − 0.33 to − 0.07, p = 0.003). Similarly, the proportion of subjects indicating improvement in the “physical health” domain was significantly correlated with a decrease in body weight from baseline to week 24 (ρ = − 0.39, 95% CI − 0.50 to − 0.27, p < 0.001) and with a decrease in HbA1c from baseline to week 24 (ρ = − 0.20, 95% CI − 0.32 to − 0.06, p = 0.004). The change in DTR-QOL total score from baseline to week 24 was significantly correlated with a decrease in body weight from baseline to week 24 (ρ = − 0.17, 95% CI − 0.29 to − 0.04, p = 0.012) and with a decrease in HbA1c level from baseline to week 24 (ρ = − 0.21, 95% CI − 0.34 to − 0.08, p = 0.001). In contrast, the EQ-5D-5L score did not show any significant correlation with the decrease in body weight (p = 0.10) or HbA1c level from baseline to week 24 (p = 0.64).

Table 7 Correlation analysis
Full size table

Safety Outcomes

During the study, 16 of 124 subjects in the dapagliflozin group (12.9%) and 13 of 124 subjects in the DPP4i group (10.5%) reported adverse events (Table 8). Death was not reported during the study, and serious adverse events were reported in 2 (1.6%) and 1 (0.8%) subjects in the dapagliflozin and DPP4i groups, respectively. One subject in the dapagliflozin group reported hypoglycemia-like symptoms, but the symptom was mild, and recovered without any treatment intervention. No severe hypoglycemia was reported in either group throughout the study. No significant difference was observed in the frequency of adverse events or serious adverse events between the groups. Known side effects of the study agents, such as dehydration, urinary tract infection, genital infection, and cystitis in the dapagliflozin group, and bullous pemphigoid in the DPP4i group, were not reported throughout the study.

Table 8 Adverse events
Full size table

Discussion

The J-BOND study was conducted to compare the effects of an SGLT2i, dapagliflozin, and DPP4is on patients’ QOL in Japanese patients with T2DM who had newly started therapy with OHAs. Since all subjects did not use insulin or GLP-1 receptor agonist at baseline, all subjects in this study were drug-naïve patients with T2DM. This was the first prospective randomized controlled trial to estimate the effects of SGLT2is and DPP4is on patients’ QOL. Dapagliflozin showed comparable or even more favorable benefits on patients’ QOL in this study, was well tolerated, and significantly reduced the body weight in the subjects. The reduction in body weight significantly correlated with the improvement of patients’ QOL.

Although both DPP4i [18, 19] and SGLT2i [20,21,22,23] improve treatment satisfaction and QOL, no study has directly compared the effects of DPP4is and SGLT2is on patients’ QOL. This was the first report to directly compare the effects of SGLT2i and DPP4is on patients’ QOL, as measured by SHIELD-WQ-9. In this study, the proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24 was higher in the dapagliflozin group than in the DPP4i group. Body weight significantly decreased in the dapagliflozin group compared with the DPP4i group, and the proportion of subjects indicating improvement in the “overall quality of life” domain of SHIELD-WQ-9 at week 24 significantly correlated with the decrease in body weight from baseline to week 24. Similarly, the proportion of subjects indicating improvement in the “physical health” domain of SHIELD-WQ-9 at week 24 was significantly higher in the dapagliflozin group, and it significantly correlated with the decrease in body weight from baseline to week 24. As SHIELD-WQ-9 measures the perceived benefit of weight control [24], these results suggest the benefit of dapagliflozin on patients’ QOL via the body weight reduction. Previous studies have also demonstrated a greater proportion of subjects treated with dapagliflozin reporting improvement in overall QOL [20] or physical health [28] in SHIELD-WQ-9 than those treated with the placebo. There are no reports on the effects of DPP4is on weight-change-related QOL using SHIELD-WQ-9. As DPP4is had a neutral effect on body weight in this study, we presumed that DPP4is did not have much impact on SHIELD-WQ-9 owing to any change in body weight.

The total scores and domain 1 scores of DTR-QOL improved in the dapagliflozin group compared with those in the DPP4i group. The change in DTR-QOL total scores from baseline to week 24 significantly correlated with the decrease in body weight from baseline to week 24 in all subjects. These results also suggest that the body weight reduction by dapagliflozin contributed to the improvement in the patients’ QOL. DTR-QOL is a diabetes-specific QOL questionnaire [25]. Previous studies have demonstrated that both SGLT2is [29, 30] and DPP4is [31, 32] improve scores in DTR-QOL. The results of this study showing a significant improvement in DTR-QOL scores in both groups were consistent with the findings of previous reports. However, this is the first report to directly compare the effects of SGLT2i and DPP4is on DTR-QOL and show the more favorable impact of dapagliflozin on diabetes therapy-related QOL compared with the DPP4is.

The proportion of subjects indicating improvement in the “overall quality of life” and “physical health” domains of SHIELD-WQ-9 at week 24 and change in total scores of DTR-QOL from baseline to week 24 significantly correlated with the decrease in HbA1c from baseline to week 24. Both dapagliflozin and DPP4is significantly and comparably improved the HbA1c in the study subjects. These results suggest that the beneficial effects of dapagliflozin and DPP4is on the patients’ QOL by improving glycemic control were comparable.

Although the DPP4i group showed a significant increase in the EQ-5D-5L scores, no significant intergroup difference was found in the change in EQ-5D-5L scores and VAS. Although previous studies have shown that neither dapagliflozin [21] nor DPP4is [33, 34] improved EQ-5D scores compared with the placebo, only one study using sitagliptin showed a significant increase in EQ-5D scores [35]. Although the results for EQ-5D-5L scores in this study were consistent with those of previous reports, this was the first report to directly compare the effects of SGLT2i and DPP4is on EQ-5D-5L scores and demonstrate that dapagliflozin and DPP4is had comparable impact on patients’ QOL, as measured by EQ-5D-5L scores.

The administration of SGLT2i reduces body weight [20, 36,37,38], whereas DPP4i has a neutral effect on body weight [39,40,41]. The results of this study were consistent with those of previous reports: body weight and BMI significantly decreased in the dapagliflozin group, but not in the DPP4i group. Since no rebound of body weight was observed in this study (dapagliflozin group: 76.6 ± 16.1, 74.1 ± 16.1, and 73.8 ± 16.7 kg at baseline, week 12, and 24. DPP4i group: 76.4 ± 15.3, 75.6 ± 13.4, and 75.6 ± 13.4 kg at baseline, week 12, and 24), it may not affect the PROs. Although a significant decrease in abdominal circumference was observed in both groups, it was significantly greater in the dapagliflozin group. As significant correlations were found between the decrease in body weight and SHIELD-WQ-9 or DTR-QOL scores, the favorable effects of dapagliflozin on body weight may have contributed to the improvement in patients’ QOL.

No remarkable safety concern was observed in either group throughout the study. Only one hypoglycemia-like symptom was reported in the dapagliflozin group, but the symptom was mild, and recovered without any treatment intervention. Known side effects of the study agents, such as dehydration, urinary tract infection, genital infection, and cystitis in the dapagliflozin group, and bullous pemphigoid in the DPP4i group, were not reported throughout the study. This may be the reason why the domain 2 score of DTR-QOL for “anxiety and dissatisfaction with treatment” improved in both groups, and no significant intergroup difference was observed in the change in domain 2 scores of DTR-QOL.

Probably because both dapagliflozin and the DPP4is improved patients’ QOL, the subjects’ responses in both groups indicated that they preferred the assigned study agent. Medication adherence for the study agent was comparable in both groups. A previous study comparing treatment satisfaction and medication adherence among DPP4is, α-glucosidase inhibitors, biguanides, and sulfonylureas showed that the DPP4i group had the highest treatment satisfaction and adherence, comparable HbA1c-lowering effects, and fewer adverse events [19]. DPP4is were the most preferable option in terms of treatment satisfaction. Because the previous study was conducted before the launch of SGLT2is in Japan, it did not compare medication adherence for SGLT2is. In this study, medication adherence in the dapagliflozin group was relatively high (91.9 ± 14.8%) and comparable to that in the DPP4i group in this study (94.1 ± 9.6%) and to that in the DPP4i group in the previous study (92.9%) [19].

Altogether, dapagliflozin showed a comparable or even more favorable benefit on patients’ QOL compared with that of DPP4is, with comparable medication adherence, HbA1c reduction, and frequency of adverse events including hypoglycemia. These results suggest that dapagliflozin can be used as a comparable or superior antidiabetic agent compared with DPP4is.

This study has several limitations. First, this was an open-label trial, and the primary endpoint and the major secondary endpoints were PROs. Therefore, the study design as the open-label trial could be a cause of bias to the PROs. Second, all the subjects in this study were Japanese. Although SGLT2is and DPP4is are available worldwide, the results of this study cannot be generalized to patients with T2DM in other countries. In addition, DPP4is are more effective in East Asian patient populations compared with non-Asian patients [6]. As the change in HbA1c was significantly correlated with the improvement in SHIELD-WQ-9 and DTR-QOL scores in this study, the beneficial impact of DPP4is on patients’ QOL may differ between Asian and non-Asian patients. Further study is required to compare the benefit of SGLT2is and DPP4is on patients’ QOL in other Asian and non-Asian countries. Third, this study enrolled drug-naïve patients with T2DM who did not know the effects of other OHAs, and whose responses regarding medication preference were mainly based on the explanation of mode of action, effects, advantages, and possible side effects by the investigators. The effects of SGLT2is and DPP4is on patients’ QOL and medication preference should be further investigated upon the switch or add-on of other OHAs.

Conclusion

Dapagliflozin showed comparable or even more favorable benefit on patients’ QOL compared with the DPP4is. Dapagliflozin was well tolerated, and it significantly reduced body weight. The reduction in body weight significantly correlated with the improvement in patients’ QOL. In concordance with the recommendation of a “patient-centered approach” by the ADA/EASD [1, 2, 17] for developing a treatment strategy, our results demonstrate that dapagliflozin can be used as a first-line drug for T2DM in Japan with a beneficial impact on patients’ QOL.