Early uptake of semaglutide for type 2 diabetes in Scandinavia and characteristics of initiators in Denmark: A register-based drug utilization study

Abstract

The use of glucose-lowering drugs has changed considerably during the last decade.^{1, 2} To understand the role of semaglutide in the treatment of type 2 diabetes, we aimed firstly to describe the uptake in adults in the early phases after launch in three Scandinavian countries, including the impact of semaglutide on the use of other glucagon-like peptide-1 receptor agonists (GLP-1RAs) indicated for treatment of type 2 diabetes. Secondly, we aimed to compare the characteristics of semaglutide initiators with those of initiators of liraglutide and other glucose-lowering drugs and to compare early versus late initiators of subcutaneous semaglutide in Denmark.

The total dispensed quantity of GLP-1RAs in Scandinavia reached 21 million DDDs, with semaglutide accounting for 74% by the end of September 2022 (Figure 1 and Figure S1). The incidence rate and prevalence proportion of semaglutide use reached 19 per 10 000 person-years and 109 per 10 000 persons in the third quarter of 2022, mainly driven by Ozempic (Figure 1) and use in individuals aged 40–79 years in all countries (Figures S2A–C and S3A–C). Uptake of Rybelsus during the study period was limited.

Semaglutide initiators were younger compared to initiators of other non-GLP-1RA glucose-lowering drugs (Table S1). Initiators of semaglutide less frequently had previous and concomitant use of DPP-4 inhibitors, SUs and insulin compared to initiators of liraglutide. They did, however, use slightly more SGLT2 inhibitors compared to initiators of liraglutide. There were only minor differences between semaglutide and liraglutide initiators in use of comedication. Semaglutide initiators had slightly less cardiovascular and renal disease compared to initiators of liraglutide. General practitioners (GPs) prescribed most initial prescriptions in both semaglutide and liraglutide users, although hospital physicians played a greater role in the prescribing of liraglutide. There were differences in characteristics between users of Ozempic and Rybelsus (Table S1). Users of Rybelsus were older, they more frequently had previous use of other glucose-lowering drugs except for insulin and SUs, they more often used statins and antihypertensives, and almost exclusively had their first prescription issued by a GP.

‘Late’ initiators of Ozempic were younger and had fewer comorbidities compared to ‘early’ initiators (Table 1). There were more women among the late versus early (50% vs. 42%) initiators. Late initiators less frequently had previous and concomitant use of other glucose-lowering drugs, they used statins (63% vs. 77%) and antihypertensives (67% vs. 80%) less frequently, and had less cardiovascular and renal disease compared to early initiators. More prescriptions were issued by GPs among late versus early initiators (84% vs. 64%). One-fifth of Ozempic initiators had at least one condition specified as an exclusion criterion in the SUSTAIN trials, with 15% having impaired renal function.

The early uptake of semaglutide in Scandinavia has been considerable, mainly at the expense of liraglutide. Initiators of semaglutide and liraglutide differed only slightly in patient characteristics. The small differences in characteristics may stem from the use of Ozempic for indications other than type 2 diabetes.

Late initiators of Ozempic had less cardiovascular drug use and comorbidity compared to early initiators, in line with previous findings from Denmark showing less cardiovascular comorbidity over time among initiators of GLP-1RAs, and also included fewer individuals with a diabetes hospital diagnosis.¹ This finding might support use of Ozempic outside approved indication late in the period, in line with the growing public interest in off-label use of Ozempic for weight loss.⁹

Real-world users of Ozempic were older compared to study participants in SUSTAIN 1 and 2.^{7, 8} One-fifth of real-world Ozempic users would have been ineligible for trials after SUSTAIN 1, primarily due to having an eGFR <60 mL/min/1,73 m² (13% of users) and presence of malignant neoplasms (6% of users). Although 6% of real-world users had a hospital diagnosis of heart failure (NYHA Class IV was used as an exclusion criterion in trials), we did not have access to data on NYHA class.

The primary strength of the study was the use of nationwide data on dispensed drugs, with no risk of selection bias. Limitations include lack of data on indications for use, inclusion of hospital recorded diagnoses only, and use of filled prescriptions as a proxy for drug use.

None.

Anton Pottegård reports participation in research projects funded by Alcon, Almirall, Astellas, Astra-Zeneca, Boehringer-Ingelheim, Novo Nordisk, Servier and LEO Pharma (all regulator-mandated Phase IV studies), and an unrestricted research grant from Novo Nordisk, all with funds paid to the institution where he was employed (no personal fees) and with no relation to the work reported in this paper. Lotte Rasmussen, Jacob Harbo Andersen, Øystein Karlstad, Diego Hernan Giunta, Marie Linder and Kari Furu report participation research projects funded by pharmaceutical companies, including Novo Nordisk, all regulator-mandated Phase IV studies and all with funds paid to the institution where they were employed (no personal fees) and with no relation to the work reported in this paper.