Once-Weekly Semaglutide Use in Patients with Type 2 Diabetes: Results from the SURE Spain Multicentre, Prospective, Observational Study

SURE Spain was a multicentre, prospective open-label, single-arm, non-interventional study assessing the use of OW semaglutide in adult patients with T2D in routine clinical practice in Spain. Informed consent and treatment initiation took place on the first visit (week 0), followed by an anticipated exposure period of ~30 weeks (range: 28–38 weeks). Intermediate visits scheduled according to local practice and data collection were performed throughout the entire study.

The decision to initiate semaglutide treatment was at the discretion of the treating physician, following requirements stated in the Summary of Product Characteristics (SmPC), therapeutic positioning report and local/regional guidelines, and clearly separated from the decision to include the patient in the SURE Spain study. All parameters collected in the study (except the patient-reported outcomes) were part of routine clinical practice. Patients were treated OW with commercially available s.c. semaglutide (Ozempic^®; Novo Nordisk A/S, Bagsværd, Denmark), available in a pre-filled, multidose, pen injector. The treating physician determined the maintenance dose and any subsequent changes to it. Diet and physical activity counselling could be offered in line with routine clinical practice, with modifications to prescribed antihyperglycaemic treatment at the physician’s discretion.

This study was conducted in accordance with the Declaration of Helsinki [25], the Guidelines for Pharmacovigilance Practices Module VI [26], and Good Pharmacoepidemiology Practices [27]. Prior to study initiation, the protocol, protocol amendment, patient information/informed consent form, together with any other written information to be provided to the patient and patient enrolment procedures, were reviewed and approved by the independent ethics committee/institutional review board at each study site (first approved in 2019 by the Ethics Committee of CEIm de EUSkadi, project identifier: NN9535-4368). Written informed consent was obtained from all patients prior to any study-related activities. This study is registered on ClinicalTrials.gov (NCT04067999↗).

Adult patients (age ≥ 18 years) diagnosed with T2D were included from 34 sites in Spain, with the first participant’s first visit on 5 August 2019, and the last participant’s last visit on 19 July 2021. Inclusion criteria included diagnosis of T2D and availability of one or more documented values of HbA_1c within 12 weeks prior to semaglutide treatment initiation. Exclusion criteria included previous participation in a SURE study, mental incapacity, unwillingness, or language barriers precluding adequate understanding or cooperation, prior treatment with any investigational drug (90 days before enrolment), and hypersensitivity to semaglutide or any of the excipients. The study duration of 30 weeks was considered sufficient to initiate and optimise the study treatment regimen and to obtain real-world data for the evaluation of the primary endpoint.

The primary endpoint was a change from baseline to end of study (EOS) in HbA_1c (%-point and mmol/mol). Supportive secondary endpoints included: change from baseline to EOS in BW (kg and %) and waist circumference (cm); proportion of patients achieving HbA_1c < 8.0% (64 mmol/mol), <7.5% (59 mmol/mol) and <7.0% (53 mmol/mol) [28]; reduction in HbA_1c from baseline to EOS of ≥1.0%-point; weight reduction from baseline to EOS of ≥3.0% [29] and ≥5.0%; HbA_1c reduction from baseline to EOS of ≥1.0% and weight reduction from baseline to EOS of ≥3.0% [29]; patient-reported severe or documented hypoglycaemia between baseline and EOS; and change from baseline to EOS in scores for patient-reported outcomes of: the Diabetes Treatment Satisfaction Questionnaire–status (DTSQs; absolute treatment satisfaction) comprising eight questions, of which six questions are combined into a total Treatment Satisfaction score (scale: 0 to 36); the Diabetes Treatment Satisfaction Questionnaire–change (DTSQc; relative treatment satisfaction), total treatment satisfaction (scale: −18.0 to 18.0); and the 36-item Short-Form Health Survey version 2 (SF-36^®v2), physical and mental summary component. The proportion of patients who completed the study under treatment with semaglutide was also investigated.

Exploratory assessments included: weekly dose of semaglutide at EOS; proportion of patients who had not added new antihyperglycaemic drug(s) to semaglutide treatment at any time during the study, evaluated at EOS; proportion of patients who had achieved clinical success, in relation to the reason to initiate semaglutide treatment, as assessed by the physician at EOS; patient-reported 8-Item Morisky Medication Adherence Scale (MMAS-8) score at EOS (low, medium, high) [30,31,32]; and the number of severe or documented hypoglycaemic episodes. Post hoc assessments included change from baseline to EOS in BMI (kg/m²). Permission for use of the MMAS-8 was granted prior to the study.

Only information on serious adverse drug reactions (SADRs), fatal events, pregnancies in female patients, and adverse events (AEs) in foetuses or newborns were systematically collected during the study. Voluntary reporting of other safety information by the physician followed the same process as for the systematic safety reporting. All episodes of patient-reported documented and/or severe hypoglycaemia were to be recorded.

Power calculations showed that a sample size of 130 patients was required, based on the criterion of 90% probability of obtaining a 95% confidence interval (CI) for mean change from baseline in HbA_1c whose half-width was at most 0.30. The half-width of 0.30 was chosen as a reasonable uncertainty allowing for a robust evaluation of glycaemic efficacy, in line with diabetes guidelines [33]. To ensure sufficient statistical power to evaluate the efficacy of semaglutide on glycaemic control (on the basis of evidence from previous observational studies with GLP-1RA treatment), it was necessary to include at least 217 enrolled patients initiating semaglutide, to ensure that 130 patients completed the study on-treatment [34,35].

The Full Analysis Set (FAS), which included all patients in the study who initiated semaglutide treatment, was used for characterising baseline demographics, analysis of the secondary endpoint related to study completion on-treatment, the selected exploratory assessments, description of AEs, and the sensitivity analyses of the primary and secondary endpoints.

The Effectiveness Analysis Set (EAS) included all patients in the FAS who completed the study and were receiving semaglutide treatment at EOS. The EAS was used for characterising baseline demographics at EOS, the description of antihyperglycaemic medications at baseline and EOS, and the primary, secondary and exploratory endpoint analyses.

Baseline demographic data are summarised using descriptive statistics (mean ± standard deviation [SD] or median and interquartile range for continuous variables and number and proportion for categorical variables). Change in the continuous variables of the primary and secondary endpoints from baseline to EOS were analysed using the Analysis of Covariance (ANCOVA) model. Categorical endpoints were analysed using descriptive statistics.

Sensitivity analyses investigated the robustness of the conclusions from the main analyses and explored the impact of missing data in the primary analysis, for which patients were excluded if they did not complete the study or discontinued treatment, or if HbA_1c data were missing at EOS. The prespecified in-study sensitivity analysis of the primary endpoint included all patients in the FAS with at least one post-baseline HbA_1c measurement in the in-study period. For this analysis, the primary endpoint was analysed using a Mixed Model for Repeated Measures (MMRM) including all HbA_1c assessments in the in-study period. The on-treatment sensitivity analysis included patients in the FAS with at least one post-baseline HbA_1c assessment, but it only included HbA_1c assessments in the on-treatment period and used the same statistical approach as the in-study sensitivity analysis.

Because of the COVID-19 pandemic, the EOS visit (V6) window was extended beyond 38 weeks to allow participants to complete their EOS assessments. Consequently, an additional post hoc sensitivity analysis was performed to explore the impact of extending the EOS visit (V6) window on the primary endpoint. The sensitivity analysis of the primary endpoint was the same as the primary analysis of the primary endpoint but included only those patients who had an EOS visit between weeks 28 and 38 (the original visit window). An additional post hoc sensitivity analysis was performed to explore the impact of extending the EOS visit on the secondary endpoint of change from baseline to EOS in BW. This sensitivity analysis was the same as the main analysis of this endpoint but included only those patients who had an EOS visit between weeks 28 and 38.

Of the 228 patients who signed the consent form, one did not meet the eligibility criteria. Therefore, the FAS comprised the 227 patients who were enrolled in the study and who had initiated semaglutide treatment (Figure 1). A total of 210 patients (92.5%) completed the study, and the mean treatment duration was 33.7 weeks. The reasons for non-completion were: death (n = 1; 0.4%), lost to follow-up (n = 3; 1.3%), withdrawal by patient (n = 3; 1.3%), and missed EOS visit within the visit window (n = 10; 4.4%) (Figure 1). The EAS comprised 196 patients (86.3%) who had completed the study on semaglutide treatment (Figure 1). Twelve patients (5.2% of the FAS) had an unknown treatment status at EOS. With regard to discontinuations, 16 patients (7.0%) discontinued treatment due to unacceptable gastrointestinal (GI) intolerability, and a further three patients (1.3%) had ‘other’ recorded as the reason (Figure 1).

Baseline characteristics of patients are summarised in Table 1. Hypertension and dyslipidaemia were the most frequent CV comorbidities at baseline, affecting 75.8% and 76.2% of patients, respectively.

Most patients initiated semaglutide at a dose of 0.25 mg (83.3%); 13.7% initiated at 0.5 mg and 3.1% at 1.0 mg. The most common reasons for initiating semaglutide as part of T2D treatment were weight reduction (94.3%) and to improve glycaemic control (88.5%).

The most frequent antihyperglycaemic drugs used by patients in the EAS at baseline were metformin (75.5% of patients), sodium–glucose cotransporter-2 inhibitors (SGLT-2is) (42.3%), basal insulin (32.7%), and dipeptidyl peptidase-4 inhibitors (DPP-4is) (20.4%) (). Supplementary Table S1

For patients in the EAS receiving semaglutide, statistically significant reductions were observed at EOS for mean HbA_1c, BW, waist circumference and BMI (Table 2). The mean HbA_1c at EOS was 7.1%, and the estimated mean change from baseline was −1.3%-points [95% CI −1.51;−1.18%-points; p < 0.0001] (Table 2, Supplementary Figure S1); mean BW at EOS was 93.2 kg, and the estimated mean change from baseline was −5.7 kg [95% CI −6.36; −4.98 kg; p < 0.0001] (Table 2, Supplementary Figure S1); mean BMI at EOS was 34.4 kg/m²_, and the estimated mean change from baseline was −2.1 kg/m² [95% CI −2.37; −1.86 kg/m²; p < 0.0001]; and mean waist circumference at EOS was 113.4 cm, and the estimated mean change from baseline to EOS was −5.3 cm [95% CI −6.29; −4.41 cm; p < 0.0001] (Table 2, Supplementary Figure S1).

At EOS, 81.0%, 67.7% and 54.0% of patients in the EAS had an HbA_1c of < 8.0%, <7.5% and <7.0%, respectively (Figure 2). The proportion of patients achieving an HbA_1c reduction ≥1%-point was 56.6% and the proportions achieving weight reduction of ≥ 3.0% and ≥5.0% were, respectively, 69.2% and 49.7% (Figure 2). The proportion of patients in the EAS achieving the composite endpoint of an HbA_1c reduction of ≥ 1.0% and weight reduction ≥3.0% at EOS was 44.3% (Figure 2). In the FAS, 86.3% of patients completed the study on-treatment with semaglutide (Figure 1).

Prespecified sensitivity analyses were used to explore the impact of missing data in the main analysis. The on-treatment sensitivity analysis of the FAS showed that the mean HbA_1c decreased over time from initiation of semaglutide to week 30, with an estimated change of −1.4%-points [95% CI −1.59; −1.27%-points] (Supplementary Figure S2). The estimated mean changes from baseline to EOS and associated 95% CIs were similar across sensitivity analyses and showed that the mean changes in HbA_1c were statistically significantly different from having no mean change in HbA_1c (Supplementary Table S2, Supplementary Figure S2). Moreover, the estimated mean HbA_1c and estimated change in HbA_1c were similar over the course of the study for both the in-study and on-treatment period.

Additional post hoc sensitivity analyses were performed in patients who had their EOS visit within the original visit window (week 28–38). The post hoc analyses of the mean changes from baseline to EOS for HbA_1c and for BW showed similar results to those seen in the primary analysis (Supplementary Table S3).

Collectively, the sensitivity analyses supported the conclusions from the primary analysis, which included assessments for patients who were on-treatment at the EOS visit (V6), also including those completing the study after week 38.

The mean ± SD weekly dose of semaglutide at EOS was 0.85 ± 0.24 mg. At EOS, five (2.6%) patients were receiving 0.25 mg OW semaglutide, 50 (25.5%) were receiving 0.5 mg, two (1.0%) were receiving between >0.5 mg and <1.0 mg, and 139 (70.9%) were receiving 1.0 mg.

In patients receiving semaglutide at EOS, DTSQs score increased by 4.4 [95% CI 3.66; 5.07; p < 0.0001] from baseline to EOS, representing a significant increase in absolute treatment satisfaction (Figure 3). Patients receiving semaglutide also reported a DTSQc score at EOS of 13.1 (95% CI 12.36; 13.85) out of a maximum score of 18, indicating a significant relative improvement in treatment satisfaction (p < 0.0001) (Figure 3).

At EOS, significant increases were observed in both the SF-36^®v2 health-related quality-of life (HRQoL) questionnaire physical component score (p < 0.0001) and the mental component score (p = 0.0013), indicating an improvement in quality of life from baseline to EOS (Figure 3).

Mean MMAS-8 score was 7.0 at baseline and 7.4 at EOS, indicating a medium level of treatment adherence. The proportion of participants with medium and high adherence was, respectively, 49.5% and 36.6% at baseline and 34.1% and 57.3% at EOS.

AEs and severe or documented hypoglycaemic episodes in patients receiving semaglutide are summarised in Table 3. In the FAS, 15 patients (6.6%) reported 26 treatment-emergent AEs: 88.5% of AEs were non-serious (13 [5.7%] patients; 23 events) and 46.2% were moderate in intensity (7 [3.1%] patients; 12 events). A total of 13 patients reported 22 GI AEs, which accounted for the highest number of AEs by system organ class. Three serious AEs (SAEs) were reported (Medical Dictionary for Regulatory Activities preferred terms: atrial fibrillation, left ventricular failure, and myocardial infarction) by two patients (0.9%), which were all judged as unlikely to be related to semaglutide treatment by the investigators. One severe SAE (preferred term: myocardial infarction) was reported, which had a fatal outcome.

Eight patients (4.1%) in the EAS reported severe or documented hypoglycaemia episodes between baseline and EOS, with similar results in the FAS (12 patients; 5.3%). At EOS, 20 severe or documented hypoglycaemic episodes were reported in the EAS, and 26 were reported in the FAS. Of these 26 events, 22 were reported by patients while using insulin and 3 occurred while using sulphonylureas. The date of one hypoglycaemic episode was unrecorded, which prevented an assessment of concurrent medication use. No severe hypoglycaemic episodes were reported during the study.

The SURE Spain study was non-interventional and single-armed, so the potential impact of other predictive factors cannot be excluded. The fundamental limitation of such a study design is the absence of a randomised comparator, which would otherwise have enabled differentiation of the changes caused by treatment, and the impact of other factors. Data in this study were collected during routine clinical practice, rather than through mandated examinations at predetermined time points, which may have impacted the robustness and completeness of the dataset.

The primary analysis was based on data from patients who completed the study on semaglutide treatment and with their HbAlevels recorded at EOS. This could have resulted in selection bias, because patients who benefit from the study treatment are more likely to continue than those who do not. To account for this, sensitivity analyses of the primary endpoint included all post-baseline HbAassessments as well as evaluations from intermediate visits also including patients who did not complete the study or discontinued semaglutide during the study. In addition, secondary supportive analyses assessed the percentage of patients who had started semaglutide treatment and were receiving it at the EOS. 1c 1c

The inclusion criteria were purposely designed to be broad and reflect a real-world T2D population, which is rarely the case in a standard RCT. However, it is likely that physicians who were highly motivated would have been overrepresented among the participating centres, and that the centres included either highly motivated patients or patients who were difficult to treat with the other therapies available. As a result, the enrolled group may only represent subsets of individuals who are eligible for semaglutide therapy. Nevertheless, study participants were profiled in terms of demographics and clinical data, which allowed for the assessment of the representativeness of the recruited population. Details of medical history (including T2D diagnosis) and concurrent diseases were obtained without further confirmation as provided by the investigators.

A potential limitation of SURE Spain is the study’s geographical location and time of initiation. The study was conducted soon after the launch of OW semaglutide, in a real-world setting, in a diverse T2D population recruited by investigators at 34 sites in Spain. However, the 34 sites that enrolled patients account for approximately half of the communities/regions within Spain, so may not be representative of the entire population. Furthermore, in Spain, GLP-1RAs are only reimbursed for patients who have a BMI ≥ 30 kg/m, and only approximately 8% of Spanish patients with T2D are prescribed GLP-1RAs [,]. Therefore, none of the patients enrolled in the study had a ‘normal’ BMI (≥18.5–<25 kg/m). These country-specific factors may have influenced the study results; in the future, however, semaglutide will likely be prescribed to a broader range of patients with T2D, including those with less severe disease progression. 2 2 ^{11 47}

The COVID-19 pandemic impacted intermediate and EOS visits in SURE Spain. Because of accessibility issues, several of these visits were instead conducted by telephone, rather than in-person. To further mitigate the challenges raised by the pandemic, changes were made to the study design that allowed patients to postpone their last visit (after the 38-week timepoint). An additional post hoc sensitivity analysis was performed to assess how the primary and secondary endpoints were affected by extending the EOS visit window. Extending the EOS visit window had no impact on the study outcomes.

Evidence has been reported that patients with T2D in Spain may have gained weight during the COVID-19 lockdown, due to their substantial lifestyle changes []. Sánchez et al. noted that if another lockdown were to be imposed, there should be greater emphasis on avoiding weight gain, in which case GLP-1RAs might be an effective therapy for these patients. Despite the influence of COVID-19, the data from this study are regarded as robust, and are suitable for further interpretation. ⁴⁸