Real-world use of oral semaglutide in adults with type 2 diabetes

In 2024, an estimated 330,000 individuals in Denmark were living with type 2 diabetes (T2D) [1]. More than 80% of individuals with T2D in Denmark are managed in primary care, whereas specialist care is restricted to those who experience complications or do not reach treatment targets [2]. Semaglutide, a glucagon-like peptide-1 receptor agonist (GLP-1RA), is available as a once-weekly subcutaneous injection (0.5, 1.0 and 2.0 mg) [3, 4] and a once-daily oral formulation (3, 7 and 14 mg), which serve as an adjunct to diet and exercise in adults with insufficiently controlled T2D [5]. Changes to the oral formulation to improve bioavailability have introduced new doses (1.5, 4 and 9 mg) [5, 6]. The phase 3 PIONEER programme demonstrated the superiority of oral semaglutide over placebo and most active comparators for glycaemic control and weight loss, with a safety profile consistent with the GLP-1RA class. PIONEER REAL comprises 13 non-interventional studies, supporting the PIONEER clinical trial programme by investigating oral semaglutide in a real-world setting in adults with T2D. Here, we assessed changes in HbA_1c, body weight and treatment satisfaction associated with the use of once-daily oral semaglutide within routine clinical practice in Denmark.

Methods

Study design

PIONEER REAL Denmark (ClinicalTrials.gov NCT04537637) was a 34-44-week, prospective, open-label, non-interventional, single-arm study across 21 sites in Denmark. Clinical data and patient-reported outcomes (PROs) were collected in accordance with local clinical practice. The study design, procedure and endpoints are similar across the PIONEER REAL studies [7], cited therein.

Participants

The study included adults aged ≥ 18 years with T2D, treatment-naïve to injectable glucose-lowering medication (except for insulin use ≤ 14 days for acute illness) and with an HbA_1c value within 90 days prior to visit (V) 1 or at V1 in accordance with local clinical practice. See Supplementary table S1 for full eligibility criteria. According to local regulations, institutional review board/ethics committee approval for this non-interventional study was not required. This study was conducted in accordance with the Declaration of Helsinki (2013) and the Guidelines for Good Pharmacoepidemiology Practices (2015). Participants provided written informed consent.

Study procedures and visits

Participants received once-daily oral semaglutide (3, 7 and 14 mg) per local label and local clinical practice, which recommended GLP-1RAs as the first choice for those with known cardiovascular (CV) or kidney disease, and as the second choice for those without known CV or kidney disease after treatment with metformin [8, 9]. Participants attended an initiation visit (V1, week 0), several intermediate visits depending on the local clinical practice (V2.X, week 1-33) and an end-of-study (EOS) visit (V3, week 34-44) (Supplementary figure S1). Baseline (BL) (week 0) assessments were recorded < 90 days prior to the informed consent and treatment initiation visit (V1). The first visit between week 34 and 44 was considered the EOS visit.

Endpoints and assessments

The primary endpoint was change in HbA_1c (percentage points or mmol/mol) from BL to EOS. Key secondary endpoints included relative (%) and absolute (kg) change in body weight from BL to EOS, percentage of participants achieving HbA_1c < 7% (< 53 mmol/mol) at EOS, and composite endpoints of achieving an HbA_1c reduction ≥ 1 percentage points (≥ 11 mmol/mol) and body weight reduction of ≥ 3% or ≥ 5% from BL to EOS. PROs were measured using the Diabetes Treatment Satisfaction Questionnaire status/change (DTSQs/DTSQc). DTSQs and DTSQc scores correspond to changes in absolute treatment satisfaction from BL to EOS and relative treatment satisfaction at EOS, respectively [10, 11] (Supplementary table S2). DTSQs was completed at V1 and V3, and DTSQc at V3 only. Exploratory endpoints at EOS included oral semaglutide dose, treatment with oral semaglutide, addition of a new/removal or a dose increase/reduction of a glucose-lowering medication from BL (other than oral semaglutide), and physician-reported clinical success. Change in waist circumference and self-reported severe hypoglycaemia were assessed (BL to EOS). All adverse events (AEs) that occurred from informed consent until the EOS visit were monitored and recorded by the treating physician, who asked participants whether they experienced AEs at the treatment initiation visit, the EOS visit and all intermediate clinic visits, which were scheduled per local clinical practice.

Statistical analysis

This was described previously [7]. Clinical endpoints and safety evaluations were based on the full analysis set (FAS). Primary analyses of primary and secondary endpoints were based on the FAS during the in-study observation period, regardless of discontinuation of oral semaglutide. Secondary analyses of the primary and secondary endpoints were based on the FAS in the on-treatment observation period (i.e. during treatment with oral semaglutide). Additional sensitivity analyses were performed, excluding participants with an EOS visit outside of week 34-44 due to COVID-19. See the Supplement for details.

Data availability

The datasets generated and/or analysed during the present study are available from Novo Nordisk upon 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 completion of the research and approval of the product and its use in the European Union and the United States. Individual participant data will be shared as datasets in a de-identified/
anonymised format.

Trial registration: ClinicalTrials.gov (NCT04537637).

Results

Participant disposition and characteristics

Between 28 August 2020 and 8 May 2023, 99 participants provided informed consent; 97 enrolled and 96 initiated oral semaglutide. A total of 93 participants (96.9%) completed the study, with 76 (79.2%) remaining on semaglutide treatment at EOS (Supplementary figure S2).

Most participants were male (65.6%), with a median BMI of 31.8 kg/m², an HbA_1c of 7.6% (60.0 mmol/mol) and a diabetes duration of 3.5 years. At BL, 28.1% of participants had HbA_1c < 7.0% (53 mmol/mol) (Table 1). Among the 96 participants, 39 (40.6%) had a CV-related medical history; 33 (34.4%) had hypertension and 16 (16.7%) had dyslipidaemia (Supplementary table S3). Metformin and sodium-glucose co-transporter-2 inhibitors (SGLT2is) were the most frequent concomitant medications at BL (61.5% and 13.5%, respectively) and EOS (65.6% and 13.5%, respectively). At BL, 35.4% of participants were not receiving any concomitant glucose-lowering medications (Supplementary table S4).

All participants were treated in a primary care setting, with one participant (1.0%) treated by an independent diabetes specialist (endocrinologist) and the remaining 95 (99.0%) by primary care physicians. All participants were initiated on 3 mg oral semaglutide. Median oral semaglutide exposure was 38.7 (IQR: 33.1; 42.1) weeks.

Glycaemic parameters and body weight

For the primary endpoint, a reduction in HbA_1c from 7.9% (63 mmol/mol) to 6.9% (52 mmol/mol) was observed from BL to EOS, with an estimated mean change of −0.9 (95% CI: −1.2; −0.6) percentage points (p < 0.0001) or −9.9 (95% CI: −13.1; −6.6) mmol/mol (Figure 1, Supplementary table S5). Secondary and sensitivity analyses were consistent with the primary analysis (Figure 1). A total of 86 participants had complete covariates and dependent variable information and were included in the primary analysis of mean change in body weight. Mean body weight decreased from BL to EOS (estimated mean change: −4.4 (95% CI: −5.4; −3.4) kg, −4.5 (95% CI: −5.5; −3.5)%; Supplementary table S5). At EOS, 66.7% and 50.7% of participants had HbA_1c < 7.0% (53 mmol/mol) or < 6.5% (48 mmol/mol), respectively, with 32.4% and 20.6% of participants achieving HbA_1c reduction ≥ 1.0 percentage points plus a body weight reduction of ≥ 3% or ≥ 5%, respectively (Figure 2).

Treatment satisfaction

Treatment satisfaction improved with oral semaglutide by EOS in DTSQs (absolute treatment satisfaction; estimated mean change from BL: 1.9 (95% CI: 0.22; 3.52) points) and DTSQc (relative treatment satisfaction; estimated mean change from BL: 10.5 (95% CI: 8.86; 12.23) points) scores (Supplementary table S5).

Exploratory endpoints

At EOS, 76 participants (79.2%) were treated with oral semaglutide: 13 (17.1%), 23 (30.3%) and 40 (52.6%) received 3, 7 and 14 mg, respectively. The mean dose was 10.0 (SD = 4.5) mg. A total of 12 (12.5%) participants had an additional/increased dose of glucose-lowering medications, and six (6.3%) ceased medication/decreased dose. The mean change in waist circumference from BL to EOS was −3.4 (SEM = 0.96; 95% CI: −5.33; −1.43) cm. Among treating physicians, oral semaglutide was considered a clinical success for 69 (73.4%) participants. There were no issues with self-reported severe hypoglycaemia.

Safety

Overall, 20 AEs were reported in 13 participants (13.5%); all except one AE were reported as mild or moderate in severity (Table 2). One serious AE, reported in a single participant (benign, malignant or unspecified neoplasm), resolved by EOS and was considered unrelated to oral semaglutide. Consistent with the GLP-1RA class safety profile, gastrointestinal disorders were most frequently reported, observed in 12 participants (12.5%). AEs led to oral semaglutide withdrawal in ten (10.4%) participants and a dose reduction in one participant (1.0%). No deaths were reported.

Discussion

Use of oral semaglutide in Denmark was associated with clinically significant improvements in glycaemic control and body weight reduction in adults with T2D, the majority of whom were treated in primary care.

A reduction in HbA_1c of 0.9 percentage points was observed from BL to EOS. HbA_1c reductions for oral semaglutide 14 mg from the PIONEER programme ranged from 1.0 to 2.0 percentage points (treatment policy estimand) [12], and −1.0 (95% CI: −1.08; −0.97) percentage point from the PIONEER REAL pooled analysis [7]. The 0.9 percentage point HbA_1c reduction reported here (BL: 7.9%) was greater than the 0.7 percentage point reduction (BL: ≥ 7% to < 8.0%) from the pooled analysis [7]. A similar 0.9 percentage point HbA_1c reduction was reported in the IGNITE study [13]. Similarly, the 4.4 kg weight loss observed here was in line with the PIONEER programme (~ 1.6-4.4 kg; treatment policy estimand) [12]. The −4.5% body weight reduction lies within the range in the PIONEER REAL studies (−4.2% to −7.2%) and is consistent with the pooled analysis [7]. The weight loss noted here may be partly due to lifestyle changes (diet optimisation and increased exercise) [14-16]. Weight loss related to uncontrolled T2D may have contributed. However, the degree of weight loss observed here exceeds that reported in lifestyle intervention studies in T2D (ranging from a weight reduction of 1.7 kg to 3.2 kg) [14, 16]. Additionally, age may impact weight loss, with higher age associated with greater and sustained weight loss following diabetes diagnosis [17]. As such, the weight loss seen here may result from a combination of factors, but it was most likely driven by the initiation of semaglutide. Treatment was considered a clinical success by the treating physician in almost three-quarters of participants; however, this is a subjective assessment that depends on individual glycaemic targets. The DTSQs and DTSQc supported physician evaluation, with treatment satisfaction improved at EOS.

Danish participants were almost exclusively from primary care, in contrast to previous PIONEER REAL studies that included individuals from secondary/specialist care. Approximately 35% were not receiving concomitant medication at BL. This does not reflect local labelling or local clinical practice at the time of the study. The reasons for this remain unknown. The relatively low proportion of participants not receiving concomitant medication at BL and the high proportion (67%) of individuals who reached HbA_1c < 7.0% by EOS indicate that the population was not highly dysregulated regarding T2D. Although 41% of participants had a medical history of CV-related conditions, hypertension (34%) and dyslipidaemia (17%) were most common. Therefore, this population presents limited CV disease risk factors as opposed to markers of established disease (e.g. heart failure, stroke or chronic kidney disease).

At EOS, 53% of participants were receiving the maximum approved oral semaglutide dose (14 mg). This proportion was similar to the PIONEER 7 trial (59%) at week 52, which, as a treat-to-target study, reflects clinical practice [18]. Here, the duration of diabetes was reported as ≥ 1 year in 78% of participants; however, 35% were not taking any glucose-lowering agents at BL. Metformin (62%) and SGLT2is (14%) were most commonly prescribed.

The tolerability profile of oral semaglutide was consistent with the phase 3 PIONEER trials, with no new safety concerns. Consistent with the GLP-1RA class, gastrointestinal AEs were most commonly reported (13% of participants), with the majority being mild to moderate. Treatment-emergent AEs led to oral semaglutide withdrawal in 10% of participants; within the range of 2-15% for the maintenance doses of 7 mg and 14 mg in phase 3 PIONEER trials [12]. A dose reduction occurred in one participant, indicating that oral semaglutide efficacy was high.

Study limitations include the observational, single-arm design, which cannot exclude alternative explanations for changes from BL in HbA_1c and other evaluated endpoints. The observed changes in HbA_1c may have been influenced by the clinical reason for initiating semaglutide (confounding by indication). Data collected through routine clinical practice versus compulsory prespecified assessments may impact the study’s robustness and completeness. Challenges in enrolment at secondary sites and in secondary care settings resulted in fewer participants than planned, despite extending the enrolment period. Consequently, the study was not sufficiently powered to allow meaningful statistical significance analysis. However, assessment of the impact of a reduced participant number indicated that the data remain robust and acceptable for further interpretation.

Conclusions

The PIONEER REAL programme provides insights into the efficacy of oral semaglutide and its use in clinical practice in adults with T2D. Despite low participant numbers, PIONEER REAL Denmark demonstrated improved glycaemic control and body weight loss in participants treated with oral semaglutide in primary care, reflecting real-world practice.