PIONEER REAL Japan: Primary results from a multicenter, prospective, real‐world study of oral semaglutide in adults with type 2 diabetes in Japanese clinical practice

ABSTRACT

Aims/Introduction

PIONEER REAL Japan was a non‐interventional prospective study of oral semaglutide in adults with type 2 diabetes in Japanese clinical practice.

Materials and Methods

Adults naïve to injectable glucose‐lowering therapies initiated oral semaglutide in routine clinical practice and were followed for 34–44 weeks. The primary endpoint was change in glycated hemoglobin (HbA_1c) from baseline to end of study; the co‐primary endpoint was number of adverse events (AEs). Secondary endpoints included change in bodyweight from baseline to end of study. Analyses were also carried out for subgroups aged <75 and ≥75 years.

Results

A total of 624 participants initiated oral semaglutide; 578 completed the study. Mean baseline HbA_1c and bodyweight were 7.7% and 72.4 kg, respectively. At end of study, estimated change (95% confidence interval [CI]) in HbA_1c from baseline was −0.7 percentage points (−0.77, −0.61) overall, −0.8 percentage points (−0.86, −0.67) in the <75 years subgroup and −0.5 percentage points (−0.68, −0.41) in the ≥75 years subgroup (all P < 0.0001). Estimated change (95% CI) in bodyweight was −2.8 (−3.19, −2.50) kg overall, −2.9 (−3.38, −2.49) kg in the <75 years subgroup and − 2.7 (−3.18, −2.14) kg in the ≥75 years subgroup (all P < 0.0001). AEs occurred in 161 (25.8%) participants: 99 of 423 (23.4%) and 62 of 201 (30.8%) participants in the <75 and ≥75 years subgroups, respectively. Gastrointestinal AEs were the AEs most frequently leading to oral semaglutide discontinuation.

Conclusions

In routine clinical practice, HbA_1c and bodyweight were significantly reduced from baseline in adults initiating oral semaglutide, including those aged ≥75 years, with no new safety concerns.

PIONEER REAL Japan was a non‐interventional study in which oral semaglutide was investigated in participants with type 2 diabetes in routine clinical practice. Study participants had significant reductions in glycated hemoglobin and bodyweight after oral semaglutide initiation, including among participants aged ≥75 years, with no new safety findings identified. This study provides insights into oral semaglutide use in a diverse real‐world population of Japanese adults with type 2 diabetes.

INTRODUCTION

It is estimated that >11 million adults in Japan (~11.8%) are living with diabetes ¹ . Japan has an aging population; a study from the Japan Diabetes Clinical Data Management Study Group (JDDM 71) showed that the proportion of people living with type 2 diabetes aged ≥65 years increased from 49.8% in 2012 to 61.2% in 2019 ² .

Japanese individuals with type 2 diabetes generally have lower insulin secretory capacity and are less likely to have obesity‐related type 2 diabetes than Western individuals ³ , ⁴ , ⁵ . Consequently, prescribing patterns differ; in Japan, dipeptidyl peptidase‐4 inhibitors (DPP‐4is) are the most prescribed treatment, particularly in individuals aged ≥65 years ² , ⁶ . Conversely, metformin is recommended in European and US type 2 diabetes populations, with increasing acceptance of glucagon‐like peptide‐1 (GLP‐1) receptor agonists (GLP‐1RAs) for some individuals ⁷ .

An oral formulation of the GLP‐1 analog, semaglutide, was approved in Japan in 2020 for treatment of individuals with type 2 diabetes and insufficient glycemic control following diet and exercise therapy ⁸ , ⁹ . Previously, GLP‐1RAs were relatively underutilized in Japan, likely due to their availability as subcutaneous injections only ² , ⁶ , although use has gradually increased since first approval of a GLP‐1RA in 2010 ² , ¹⁰ . Furthermore, Japanese Diabetes Society treatment algorithms for type 2 diabetes include GLP‐1RAs as first‐line therapy for non–insulin‐dependent individuals not achieving glycemic control with medical nutrition therapy and exercise ³ , ⁴ , ¹¹ , ¹² , ¹³ . However, their use as first‐line therapy depends on the severity of the metabolic disorder and comorbidities. Nevertheless, GLP‐1RAs can provide additional benefit in aiding weight loss in individuals with type 2 diabetes and obesity ³ , ⁴ , ¹² , ¹³ .

Real‐world evidence‐generating studies are required to understand and inform oral semaglutide use in Japanese clinical practice. In addition, insufficient clinical trial data on semaglutide use in people with type 2 diabetes aged ≥75 years in Japan means insights into semaglutide efficacy and safety in these individuals are limited ¹⁴ .

PIONEER REAL is an ongoing program of 13 non‐interventional phase IV studies, each carried out in a different country, investigating oral semaglutide use in routine clinical practice among adults with type 2 diabetes naïve to injectable glucose‐lowering therapies. Results from PIONEER REAL studies in Canada, the Netherlands, Sweden and Switzerland are being published ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ . The PIONEER REAL study carried out in Japan (PIONEER REAL Japan) was carried out to investigate glycemic control, bodyweight changes and safety in adults with type 2 diabetes in Japan who had not previously received injectable glucose‐lowering medication and initiated oral semaglutide according to local practice. The study design and participant baseline characteristics of PIONEER REAL Japan have been published ¹⁹ ; here, we present the primary clinical outcome and safety data. To address specific interest from regulatory authorities in investigating oral semaglutide in participants aged ≥75 years and to fill data gaps, we present data split into <75 years and ≥75 years subgroups.

PIONEER REAL Japan is a Pharmaceuticals and Medical Devices Agency–required safety study for oral semaglutide (Rybelsus®), registered with ClinicalTrials.gov (NCT04878393).

MATERIALS AND METHODS

Study design and participants

PIONEER REAL Japan was a non‐interventional, prospective, multicenter, single‐arm, phase IV study. Study design and participant eligibility have been described ¹⁹ . Briefly, participants provided informed consent before or at visit 1 (week 0), which was the baseline and oral semaglutide initiation visit. Participants were followed for 34–44 weeks, when they could attend any number of intermediate visits (visit 2.x), before an end‐of‐study (EoS) visit at week 34–44 (visit 3) in accordance with routine clinical practice. Due to COVID‐19 pandemic restrictions, intermediate visits and visit 3 could be conducted by telephone, and the interval to visit 3 was extended to 52 weeks. The treating physician's decision to include a participant in the study was independent of their decision to initiate oral semaglutide. Enrolled participants were aged ≥20 years with a type 2 diabetes diagnosis and naïve to injectable glucose‐lowering medication (aside from ≤14 days of insulin treatment for acute illness).

Regulatory compliance, informed consent and study approvals have been described ¹⁹ .

Endpoints

The primary endpoint was change in glycated hemoglobin (HbA_1c) from baseline to EoS. The co‐primary safety endpoint was the number of adverse events (AEs) from baseline to EoS.

Secondary endpoints measured from baseline to EoS included relative (%) and absolute (kg) change in bodyweight, participants reaching HbA_1c <7%, participants reaching the composite endpoint of HbA_1c reduction ≥1 percentage points and bodyweight reduction of ≥5%, participants reaching the composite endpoint of HbA_1c reduction ≥1 percentage points and bodyweight reduction of ≥3%, and the number of adverse drug reactions (ADRs), serious AEs, and serious ADRs. AEs were coded using the Medical Dictionary for Regulatory Activities, version 25.1. ADRs were defined as AEs for which a causal relationship between oral semaglutide and the AE was suspected and judged as possible or probable by the sponsor or investigator.

Exploratory and other analyses included change from baseline in waist circumference at EoS, oral semaglutide dose at EoS, participants treated with oral semaglutide at EoS, participants with self‐reported severe hypoglycemia and participants with gastrointestinal AEs. Participants were asked to complete a dosing conditions questionnaire at visit 3 (or on treatment discontinuation) on how oral semaglutide was consumed.

The primary and co‐primary safety endpoints were also evaluated in participants with mild, moderate and severe renal impairment, defined by estimated glomerular filtration rates of 60–<90, 30–<60 and 15–<30 mL/min/1.73 m², respectively, calculated based on creatinine values using the Chronic Kidney Disease Epidemiology Collaboration equation ²⁰ .

All clinical data were recorded in an electronic case report form.

Statistical analysis

Sample size calculations have been described ¹⁹ .

The full analysis set (FAS) included all eligible participants who provided informed consent and initiated oral semaglutide treatment. Data points from the FAS were selected based on two observation periods. The in‐study observation period represented the time participants were in the study, irrespective of whether they discontinued oral semaglutide. This started at visit 1 and ended at visit 3, withdrawal of informed consent, last participant–physician contact if the participant was lost to follow‐up or study withdrawal for any other reason (e.g., death). The on‐treatment observation period represented the time when participants were treated with oral semaglutide until the last dose of oral semaglutide according to the treating physician (plus 2 weeks after discontinuation date for measurements recorded 2 weeks after discontinuation), visit 3, withdrawal of informed consent, last participant–physician contact if the participant was lost to follow‐up or study withdrawal for any other reason (e.g., death).

The primary endpoint analysis was carried out on the FAS with at least one post‐baseline HbA_1c measurement in the in‐study observation period. A mixed model for repeated measurements with random intercept and time (slope) for each participant was used, with baseline value (HbA_1c), time and time squared as covariates; the adjusted model also used age and body mass index (BMI) as covariates, and sex, number of baseline glucose‐lowering therapies, type 2 diabetes duration and study sites as fixed effects. A secondary analysis of the primary endpoint was carried out for the on‐treatment observation period to evaluate effects of treatment discontinuation using the same statistical methods.

For safety‐related endpoints, descriptive statistics were used to summarize the number of AEs and participants with AEs; the event rate (events per 100 participant‐years) was calculated. Safety‐related endpoints were assessed in the FAS.

For secondary and exploratory endpoints, changes from baseline in bodyweight and waist circumference were analyzed as per the primary endpoint. Categorical endpoints were summarized as percentages.

Missing data were not imputed.

Statistical analyses were carried out separately for subgroups of participants aged <75 years and ≥75 years. A cap was set on the number of participants enrolled aged <75 years, so that proportionally enough participants aged ≥75 years were enrolled to ensure sufficient power in the primary subgroup analysis (described previously) ¹⁹ .

RESULTS

Participants

PIONEER REAL Japan took place across 62 sites (clinics and hospitals) between 30 April 2021 and 9 February 2023. Overall, 650 individuals signed consent forms; 624 were enrolled and initiated oral semaglutide treatment, with 578 (92.6%) completing the study (Figure 1). Of the 578 participants who attended visit 3, 10 (1.7%) did so by telephone or video contact and six (1.0%) attended after the week 34–44 window. Overall, 134 participants (21.5%) discontinued treatment, most commonly due to safety concerns related to oral semaglutide (Figure 1). In the ≥75 years subgroup, 32.3% of participants discontinued, compared with only 16.3% in the <75 years subgroup (Figure 1). Treatment status over time is presented in Figure S1.

Figure 1.

Participant disposition. Percentage values are based on the full analysis set. ^†Participants who initiated oral semaglutide treatment and attended the EoS visit. ^‡As recorded in the ‘Discontinuation of oral semaglutide form’ in the electronic case report form. ^§Participants who were receiving oral semaglutide treatment and attended the EoS visit. EoS, end of study.

Participant demographics and baseline characteristics have been described previously ¹⁹ , and are summarized in Table 1. At baseline, 379 participants (67.0%) had obesity, defined as a BMI of ≥25 kg/m² based on Japanese obesity guidelines ²¹ . Overall, 364 participants (58.3%) switched from a DPP‐4i within a maximum of 180 days of oral semaglutide initiation.

Table 1.

Participant demographics and baseline characteristics in the overall population and in the <75 years and ≥75 years age subgroups (full analysis set) ^†

Characteristic	Total (N = 624)	<75 years subgroup (n = 423)	≥75 years subgroup (n = 201)
Age (years)	64.1 ± 14.1	57.1 ± 11.7	78.9 ± 3.3
HbA1c (%)	7.7 ± 1.1	7.7 ± 1.2	7.8 ± 1.0
Bodyweight (kg)	n = 614 72.4 ± 16.1	n = 415 76.8 ± 16.6	n = 199 63.3 ± 9.9
BMI (kg/m2)	n = 566 27.5 ± 5.0	n = 384 28.6 ± 5.2	n = 182 25.1 ± 3.4
Type 2 diabetes duration (years)	n = 404 10.7 ± 8.5	n = 284 9.1 ± 7.3	n = 120 14.6 ± 9.8
Kidney function, n (%) ‡	n = 557	n = 379	n = 178
Normal (eGFR ≥90 mL/min/1.73 m2)	292 (52.4)	269 (71.0)	23 (12.9)
Mild impairment (eGFR 60–<90 mL/min/1.73 m2)	208 (37.3)	86 (22.7)	122 (68.5)
Moderate impairment (eGFR 30–<60 mL/min/1.73 m2)	55 (9.9)	23 (6.1)	32 (18.0)
Severe impairment (eGFR 15–<30 mL/min/1.73 m2)	2 (0.4)	1 (0.3)	1 (0.6)

All values are mean ± standard deviation, except where stated. BMI, body mass index; eGFR, estimated glomerular filtration rate; HbA_1c, glycated hemoglobin; n, number of participants with given characteristic; N, number of participants evaluated for a given characteristic, where this differed from the full analysis set.

^† More participant demographics and baseline characteristics are provided in Suzuki et al. ¹⁹

^‡ eGFR calculated based on the creatinine value using the Chronic Kidney Disease Epidemiology Collaboration equation ²⁰ .

For HbA_1c with respect to renal function, the mean (standard deviation [SD]) HbA_1c at baseline was 7.8% (1.3%), 7.5% (0.8%), 7.8% (0.8%), 8.2% (0.4%) and 7.9% (0.9%) for participants with normal kidney function (n = 264), and mild (n = 188), moderate (n = 48), severe (n = 2) and unknown (n = 56) renal impairment, respectively.

Study drug exposure

The mean (SD) in‐study observation period was 36.4 (5.3) weeks (range 1.7–50.1 weeks) overall, and 36.7 (5.3) weeks and 35.9 (5.3) weeks for the <75 years and ≥75 years subgroups, respectively. The mean (SD) on‐treatment observation period was 31.4 (10.7) weeks (range 0.1–44.1 weeks) overall, and 33.0 (9.4) weeks and 28.2 (12.5) weeks in the <75 years and ≥75 years subgroups, respectively. The mean (SD) oral semaglutide exposure time was 31.0 (11.3) weeks (range 0.1–44.1 weeks) overall, and 32.7 (9.9) weeks and 27.6 (13.3) weeks in the <75 years and ≥75 years subgroups, respectively.

All participants initiated oral semaglutide at a dose of 3 mg. At EoS, 124 participants (26.6%) remained on the 3 mg dose: 71 (21.1%) in the <75 years subgroup and 53 (40.8%) in the ≥75 years subgroup. At EoS, 288 (61.8%) were receiving 7 mg: 219 (65.2%) and 69 (53.1%) in the <75 years and ≥75 years subgroups, respectively; 52 (11.2%) were receiving 14 mg: 45 (13.4%) and seven (5.4%) in the <75 years and ≥75 years subgroups, respectively. Only two participants (0.4%) had their oral semaglutide dose decreased to <3 mg (calculated as 1.5 mg and 2.33 mg per day).

Clinical outcomes

At EoS, the estimated change from baseline in HbA_1c was −0.7 percentage points (95% confidence interval [CI] −0.77, −0.61; P < 0.0001); significant reductions in HbA_1c were observed in both age subgroups (P < 0.0001; Figure 2). In a secondary analysis of the primary endpoint for the on‐treatment observation period (539 participants), the mean (SD) HbA_1c at baseline was 7.7% (1.1%), the estimated change from baseline was −0.7 percentage points (95% CI −0.81, −0.62; P < 0.0001) and the estimated mean HbA_1c at EoS was 7.0%.

Figure 2.

Estimated change from baseline in glycated hemoglobin (HbA_1c; % points) at end of study (mixed model for repeated measurements adjusted, in‐study observation period, full analysis set). CI, confidence interval; N, number of participants evaluated for endpoint; SD, standard deviation.

Among participants with renal impairment, estimated changes from baseline in HbA_1c ranged from −0.6 percentage points to −0.9 percentage points, and reached significance across the renal impairment categories, except for the severe impairment group, in which statistical analysis was not carried out due to the low sample size (Figure S2).

The estimated change in bodyweight at EoS was −2.8 kg in the total population, −2.9 kg in the <75 years subgroup and −2.7 kg in the ≥75 years subgroup (all P < 0.0001; Figure 3a). For the on‐treatment observation period (539 participants), the mean (SD) bodyweight was 72.8 kg (16.1 kg) at baseline; the estimated mean at EoS was 69.9 kg, and estimated changes from baseline were −3.0 kg (95% CI −3.41, −2.64; P < 0.0001) and −4.2% (95% CI −4.67, −3.69; P < 0.0001).

Figure 3.

Secondary efficacy endpoints in the total population and <75 years and ≥75 years subgroups. (a) Estimated change in body weight (kg) at end of study (EoS); (b) change in bodyweight (%) at EoS; (c) proportion of participants with glycated hemoglobin (HbA_1c) <7% at baseline and EoS; (d) proportion of participants with reductions from baseline in HbA_1c ≥1 percentage points and bodyweight ≥5% at EoS; (e) proportion of participants with reductions from baseline in HbA_1c ≥1 percentage points and bodyweight ≥3% at EoS (full analysis set). CI, confidence interval; n, number of participants reaching a given endpoint; N, total number of participants evaluated for a given endpoint; SD, standard deviation.

Proportionally more participants had HbA_1c <7% at EoS versus baseline, especially among participants aged <75 years (Figure 3c); the proportions of participants reaching HbA_1c targets of <7% (among participants with baseline HbA_1c ≥7%), <7.5% and <8% are shown in Figure S3. Across multiple individual HbA_1c targets, proportionally more participants reached their targets at EoS versus baseline (Figure S4).

In total, 88 participants (16.5%) and 119 participants (22.3%) reached the composite endpoints of HbA_1c reduction of ≥1 percentage points and bodyweight reduction of ≥5%, and HbA_1c reduction of ≥1 percentage points and bodyweight reduction of ≥3%, respectively, at EoS (Figure 3d,e). Higher proportions of participants aged <75 years reached these composite endpoints versus those aged ≥75 years. Results for each component of these endpoints are shown in Figure S5.

Mean HbA_1c, fasting plasma glucose, bodyweight, BMI, waist circumference, total cholesterol, low‐density lipoprotein cholesterol and blood pressure all decreased from baseline to EoS (Table S1). No changes were observed in serum creatinine levels; estimated glomerular filtration rate levels and the relative proportions of participants with mild, moderate, severe and no renal impairment were similar at EoS to baseline (Table S1).

Safety

Overall, 268 AEs were reported in 161 participants; most were mild or moderate and non‐serious (Table 2). The most common AEs, ADRs and AEs leading to oral semaglutide discontinuation by system organ class were gastrointestinal AEs (Tables 2 and S2). Participants in the ≥75 years subgroup experienced a higher rate of AEs than the <75 years subgroup.

Table 2.

Summary of safety – in‐study observation period (full analysis set)

	Total (N = 624) †			<75 year subgroup (n = 423) †			≥75 year subgroup (n = 201) †
	n (%)	No. events	Event rate ‡	n (%)	No. events	Event rate ‡	n (%)	No. events	Event rate ‡
AEs	161 (25.8)	268	61.5	99 (23.4)	149	50.1	62 (30.8)	119	86.1
AE severity
Mild	136 (21.8)	220	50.5	85 (20.1)	129	43.4	51 (25.4)	91	65.8
Moderate	35 (5.6)	42	9.6	17 (4.0)	19	6.4	18 (9.0)	23	16.6
Severe	4 (0.6)	6	1.4	1 (0.2)	1	0.3	3 (1.5)	5	3.6
Serious AEs	21 (3.4)	25	5.7	10 (2.4)	12	4.0	11 (5.5)	13	9.4
AEs with fatal outcome	2 (0.3)	3 §	0.7	0	0	0	2 (1.0)	3 §	2.2
Causality due to study drug
Probable	95 (15.2)	119	27.3	59 (13.9)	72	24.2	36 (17.9)	47	34.0
Possible	37 (5.9)	40	9.2	17 (4.0)	17	5.7	20 (10.0)	23	16.6
Unlikely	62 (9.9)	107	24.6	41 (9.7)	60	20.2	21 (10.4)	47	34.0
ADRs	128 (20.5)	159	36.5	75 (17.7)	89	30.0	53 (26.4)	70	50.6
Serious ADRs	2 (0.3)	2	0.5	0	0	0	2 (1.0)	2	1.4
AEs leading to oral semaglutide discontinuation in ≥1% of participants in any group by preferred term ¶
All AEs	68 (10.9)	88	20.2	33 (7.8)	39	13.1	35 (17.4)	49	35.4
Gastrointestinal AEs	53 (8.5)	59	13.6	30 (7.1)	33	11.1	23 (11.4)	26	18.8
Nausea	21 (3.4)	21	4.8	13 (3.1)	13	4.4	8 (4.0)	8	5.8
Vomiting	9 (1.4)	9	2.1	5 (1.2)	5	1.7	4 (2.0)	4	2.9
Diarrhea	8 (1.3)	8	1.8	3 (0.7)	3	1.0	5 (2.5)	5	3.6
Constipation	7 (1.1)	7	1.6	5 (1.2)	5	1.7	2 (1.0)	2	1.4
Abdominal discomfort	5 (0.8)	5	1.1	2 (0.5)	2	0.7	3 (1.5)	3	2.2
Metabolism and nutrition AEs	9 (1.4)	9	2.1	2 (0.5)	2	0.7	7 (3.5)	7	5.1
Decreased appetite	8 (1.3)	8	1.8	2 (0.5)	2	0.7	6 (3.0)	6	4.3
General AEs and administration site conditions	3 (0.5)	3	0.7	0	0	0	3 (1.5)	3	2.2
Malaise	2 (0.3)	2	0.5	0	0	0	2 (1.0)	2	1.4

ADR, adverse drug reaction; AE, adverse event; n, number of participants with event; NIS, non‐interventional study.

^† Total observation time: 435.4 years (total), 297.1 years (<75 years age subgroup) and 138.3 years (≥75 years age subgroup).

^‡ Event rate per 100 years of observation time.

^§ Two fatal events were recorded for one participant; numbers are presented as per the data captured.

^¶ Discontinuations due to adverse events were reported using the ‘NIS Safety Form’ in the electronic case report form.

Some AEs of note were reported. Hypoglycemia occurred in three participants in the ≥75 years subgroup (event rate 2.2/100 participant‐years in this subgroup). Although no cases were serious, one was severe (requiring another person to take corrective action), and all three were considered ADRs. All participants with hypoglycemia had switched from DPP‐4is within 180 days of oral semaglutide initiation; one was receiving concomitant gliclazide, luseogliflozin, metformin and pioglitazone; another was receiving imeglimin, insulin, ipragliflozin, metformin, mitiglinide and voglibose; the third participant received no concomitant glucose‐lowering medications.

One 76‐year‐old female participant was diagnosed with serious autoimmune pancreatitis of moderate intensity 138 days after discontinuing oral semaglutide, which resolved. A 51‐year‐old male participant had moderate acute cholecystitis 227 days after oral semaglutide initiation, which was resolving. Both AEs were reported by the investigator as unlikely to be related to oral semaglutide.

A total of 159 ADRs were reported in 128 participants (20.5%); only one ADR was severe (nausea). All ADRs were considered probably (119 events in 95 participants [15.2%]) or possibly (40 events in 37 participants [5.9%]) related to the study drug by the investigator, and 79 ADRs resulted in withdrawal of oral semaglutide in 64 participants (10.3%). Serious ADRs were reported in two participants: intestinal obstruction and drug eruption (event rate for each: 0.2/100 participant‐years). Both were moderate in severity, were considered possibly related to study drug and resolved. Oral semaglutide was withdrawn in both cases.

Two deaths occurred during the in‐study observation period. A 78‐year‐old male participant with a BMI of 24.9 kg/m² died 169 days after initiating and 114 days after discontinuing oral semaglutide; the cause of death was myocardial infarction or arrhythmia, and both AEs were considered unlikely to be study drug‐related. An 83‐year‐old male participant with a BMI of 18.9 kg/m² died 16 days after starting and 5 days after discontinuing oral semaglutide due to anorexia; the cause of death was infectious pneumonia that was considered unlikely to be related to study drug.

Among participants with mild renal impairment, 52 participants (25.0%) experienced 99 AEs, including 11 serious AEs in nine participants (Table S3). A total of 17 participants (30.9%) with moderate renal impairment experienced 29 AEs, and the two participants with severe renal impairment experienced three AEs in total. Most AEs were mild or moderate in severity, with gastrointestinal AEs most frequently reported.

Other assessments

Overall, 422 participants completed the dosing conditions questionnaire. Of these, 159 participants (37.7%) found the medication very easy to consume; nearly all (97.2%) took the medication in the morning, and among 407 participants, the mean (SD) time that the tablet was taken after the last meal was 10.0 h (2.6 h; Table S4).

DISCUSSION

In the present non‐interventional study of adults with type 2 diabetes in Japanese clinical practice, 34–44 weeks of oral semaglutide treatment was associated with reduction in HbA_1c and a proportional increase in participants reaching HbA_1c targets. HbA_1c was reduced to a similar extent during the on‐treatment observation period, indicating that discontinuation or switching treatment did not impact HbA_1c reduction, as other glucose‐lowering therapies could be added. Bodyweight also reduced during the study. The safety profile of oral semaglutide was consistent with that of the PIONEER 9 and 10 pivotal trials of oral semaglutide in Japanese individuals with type 2 diabetes ²² , ²³ , with no new safety concerns, although the ≥75 years subgroup experienced more AEs.

The estimated change from baseline in HbA_1c was smaller in the ≥75 years subgroup than in the <75 years subgroup, and a greater proportion of participants aged ≥75 years than <75 years discontinued treatment. These findings were potentially due to participants aged ≥75 years having less stringent individual HbA_1c targets, as per Japanese Diabetes Society guidelines ¹² , ¹³ . More discontinuations due to investigators' or participants' safety concerns could also explain the smaller HbA_1c change. The higher incidence of gastrointestinal AEs in the ≥75 years subgroup was possibly related to lower bodyweight in these participants, who received a higher dose of semaglutide per bodyweight. Furthermore, the ≥75 years subgroup had longer type 2 diabetes duration and, thus, potentially reduced β‐cell function, also suggested by the lower bodyweight. As HbA_1c‐lowering effects of GLP‐1RAs are affected by β‐cell function ²⁴ , ²⁵ , these participants might have experienced a weaker response to oral semaglutide than those aged <75 years.

Although not directly comparable due to study design and eligibility criteria differences, these findings generally complement those from PIONEER 9 and 10, in which oral semaglutide showed notable glycemic control and weight loss in Japanese individuals with type 2 diabetes ²² , ²³ . However, the estimated change from baseline in HbA_1c described (−0.7 percentage points) was smaller than that reported in PIONEER 9 and 10 (−1.1 percentage points to −2.0 percentage points at week 26 with oral semaglutide 3, 7 and 14 mg) ²² , ²³ . Although participants in the present study had lower baseline HbA_1c than those in PIONEER 9 and 10 (7.7% vs 8.0–8.4%), a subgroup analysis of PIONEER 9 and 10 participants with baseline HbA_1c <8.0% reported HbA_1c reductions of −0.7 percentage points to −1.3 percentage points at week 26 ²⁶ , so this difference in HbA_1c reduction might not be explained by baseline HbA_1c, but by differences in assessment time points, study designs and populations. Other factors might also account for this difference. Participants in PIONEER 9 and 10 were excluded if they had received DPP‐4i ≤90 days before screening, whereas 58.3% of participants in this study had switched from DPP‐4i within 180 days of starting oral semaglutide, so could have had DPP‐4i–associated reductions in HbA_1c before initiating oral semaglutide. Also, the mean participant age was higher in this study than in PIONEER 9 and 10, and we observed smaller HbA_1c reductions in the ≥75 years subgroup, suggesting that age differences between studies contributed to the differences in HbA_1c reduction. Conversely, participants in PIONEER REAL Japan experienced greater bodyweight reductions than those in PIONEER 9 and 10 (−0.1 kg to −2.4 kg at week 26) ²² , ²³ , despite similar bodyweights at baseline between studies (except in the ≥75 years subgroup). This could partially be due to concomitant sodium–glucose cotransporter 2 inhibitor use during this study ¹⁹ .

The HbA_1c and bodyweight reductions observed in PIONEER REAL Japan are generally similar to those in other non‐interventional studies of oral semaglutide in Japanese individuals with type 2 diabetes. A retrospective study in 47 individuals reported 6‐month decreases from baseline in mean HbA_1c from 8.6% to 7.9%, and mean body weight from 77.1 kg to 75.0 kg in individuals naïve to GLP‐1RAs ²⁷ . Another retrospective analysis of 88 individuals reported a greater reduction in HbA_1c (−1.2 percentage points), but a smaller reduction in bodyweight (−1.4 kg) and a slightly lower proportion of individuals reaching HbA_1c <7% (48%) at 6 months than the present study ²⁸ . Individuals in the prior study had considerably higher HbA_1c at baseline than those here; furthermore, the HbA_1c reduction with oral semaglutide was greater when baseline HbA_1c was higher, which would explain the difference in HbA_1c reductions between studies ²⁸ . Although we performed no statistical comparisons for laboratory parameters, trends were similar to those reported previously. The study in 47 individuals observed significant decreases in systolic blood pressure, triglycerides, low‐density lipoprotein cholesterol and non–high‐density lipoprotein cholesterol after 6 months of oral semaglutide ²⁷ . The analysis in 88 individuals observed decreases in triglycerides and total cholesterol, and no change in estimated glomerular filtration rate after 6 months of oral semaglutide; the most common AE was gastrointestinal (nausea), consistent with the present study ²⁸ .

PIONEER REAL Japan used data collected from a real‐world setting, and oral semaglutide was prescribed by treating physicians according to clinical practice and independent of the decision to enroll participants; inclusion and exclusion criteria were also relatively few. Therefore, the generalizability of the results to the real‐world type 2 diabetes population naïve to injectable glucose‐lowering therapies in Japan can be considered relatively high.

A main limitation of the present study was the lack of a comparator arm, preventing the establishment of causality for the study findings. Additionally, although conventional confounders were accounted for in the primary analysis, additional unmeasured cofounding factors cannot be ruled out. Data were collected from routine clinical practice and not through mandatory assessments at prespecified time points, potentially impacting the robustness and completeness of the data and conclusions. Finally, the changes in HbA_1c might have been influenced by the clinical reason to initiate semaglutide (confounding by indication).

In conclusion, PIONEER REAL Japan provides insights into oral semaglutide utilization in routine clinical practice in a diverse real‐world population of adults with type 2 diabetes. Statistically significant reductions in HbA_1c and bodyweight were observed after oral semaglutide initiation, including among participants aged <75 years and ≥75 years. AEs were generally mild, with no new safety findings identified.

DISCLOSURE

Daisuke Yabe received consulting/lecture fees from Eli Lilly Japan K.K., Kyowa Kirin Co., Ltd., Nippon Boehringer Ingelheim Co., Ltd., Novo Nordisk Pharma Ltd., Sanofi K.K. and Sumitomo Pharma Co., Ltd.; and research funding/grants from Arkray Inc., Nippon Boehringer Ingelheim Co., Ltd., Novo Nordisk Pharma Ltd., Taisho Pharmaceutical Co., Ltd. and Terumo Corporation. Professor Yabe is an Editorial Board member of Journal of Diabetes Investigation and a co‐author of this article. To minimize bias, they were excluded from all editorial decision‐making related to the acceptance of this article for publication. Yoshiyuki Hamamoto received consulting/lecture fees from Novo Nordisk Pharma Ltd. and Sumitomo Pharma Co., Ltd.; and research funding from Nippon Boehringer Ingelheim Co., Ltd. and Sumitomo Pharma Co., Ltd. Dr Hamamoto is an Editorial Board member of Journal of Diabetes Investigation and a co‐author of this article. To minimize bias, they were excluded from all editorial decision‐making related to the acceptance of this article for publication. Daiji Kawanami received consulting/lecture fees from Bayer Yakuhin Ltd., Mitsubishi Tanabe Pharma Corporation, Novo Nordisk Pharma Ltd., Sanofi K.K. and Sumitomo Pharma Co., Ltd.; and grants from Bayer Yakuhin Ltd., Nippon Boehringer Ingelheim Co., Ltd., Nipro Corporation and Sumitomo Pharma Co., Ltd. Rimei Nishimura received consulting/lecture fees from Abbott Japan LLC, Astellas Pharma Inc., AstraZeneca K.K., Eli Lilly Japan K.K., Kissei Pharmaceutical Co., Ltd., Kowa Company Ltd., Medtronic Japan Co., Ltd., Nippon Boehringer Ingelheim Co., Ltd., Novo Nordisk Pharma Ltd., Ono Pharmaceutical Co., Ltd., Sanofi K.K., Sumitomo Pharma Co., Ltd. and Teijin Pharma Ltd.; and research funding/grants from Abbott Japan LLC, Mitsubishi Electric Corporation, Nippon Boehringer Ingelheim Co., Ltd., Ono Pharmaceutical Co., Ltd., Sumitomo Pharma Co., Ltd. and Taisho Pharmaceutical Co., Ltd. Yasuo Terauchi received consulting/lecture fees from Novo Nordisk Pharma Ltd., Astellas Pharma Inc., AstraZeneca K.K., Eli Lilly Japan K.K., Kowa Company Ltd., Mitsubishi Tanabe Pharma Corporation, MSD K.K., Nippon Boehringer Ingelheim Co., Ltd., Sanofi K.K. and Sumitomo Pharma Co., Ltd.; and grants from Nippon Boehringer Ingelheim Co., Ltd. and Sumitomo Pharma Co., Ltd. Hanan Amadid is employed by, and a shareholder in, Novo Nordisk. Uffe Christian Braae is employed by, and a shareholder in, Novo Nordisk. Atheline Major‐Pedersen is employed by, and a shareholder in, Novo Nordisk. Ryo Suzuki received lecture fees from Astellas Pharma Inc., Eli Lilly Japan K.K., Kowa Company Ltd., Mitsubishi Tanabe Pharma Corporation, MSD K.K., Novo Nordisk Pharma Ltd., Sanofi K.K., Sumitomo Pharma Co., Ltd. and Teijin Healthcare Ltd; and grants from Nippon Boehringer Ingelheim Co., Ltd.

Approval of the research protocol: The protocol for this research project has been approved by suitably constituted ethics committees or institutional review boards of the institutions, and it conforms to the provisions of the Declaration of Helsinki. Ethics committees, institutional review boards and approval numbers are listed in the Supporting Information.

Informed consent: All informed consent was obtained from the participants and/or their guardians.

Approval date of registry and the registration no. of the study/trial: This study is registered with ClinicalTrials.gov (NCT04878393; first submitted 5 May 2021).

Animal studies: N/A.

Supporting information

Table S1. Observed laboratory parameters and clinical measurements at baseline and end of study, and proportion of participants with eGFR levels within each category at baseline and end of study (in‐study observation period, FAS).

Table S2. Most frequent AEs and ADRs by MedDRA preferred term (≥1% of participants in any group, in‐study observation period, FAS).

Table S3. Summary of safety in participants according to kidney function – in‐study observation period (FAS).

Table S4. Participant responses to the dosing conditions questionnaire at end of study (FAS).

Figure S1. Treatment status over time in the total study population (FAS).

Figure S2. Estimated change from baseline in HbA_1c at end of study, by kidney function (MMRM adjusted, in‐study observation period, FAS).

Figure S3. Participants reaching individual HbA_1c targets at baseline and end of study in the total population (FAS).

Figure S4. (a) Proportions of participants reaching their individual HbA_1c targets set by the treating physician at baseline, and (b–f) the proportion of participants in each HbA_1c category according to the individual HbA_1c target set by the treating physician at baseline (FAS).

Figure S5. Proportion of participants at end of study reaching components of the composite secondary endpoints (FAS).

DATA AVAILABILITY STATEMENT

Data are available on reasonable request. Data will be shared with bona fide researchers submitting a research proposal approved by the independent review board. Access request proposals can be found at novonordisk‐trials.com. Data will be made available after research completion and approval of the product and product use in the European Union and the USA. Individual participant data will be shared in datasets in a de‐identified/anonymized format.