Discussion

This study, reflecting the impact of the COVID-19 pandemic and its associated restrictions on antidiabetic drug utilization, showed that antidiabetic consumption increased with the onset of the pandemic over a period of nearly five years, accompanied by a parallel increase in pharmaceutical expenditure. Following the relaxation of restrictions, oral antidiabetics demonstrated a distinct upward trajectory compared with overall trends and insulin-dominated injectable therapies, largely driven by increased use of SGLT-2 inhibitors. While antidiabetic drug consumption increased, prescribing levels declined during the period of intensified pandemic-related disruption and reduced healthcare utilization, possibly reflecting temporary measures that enabled continued medication dispensing without a new prescription. Although prescribing partially recovered after restrictions were eased, it remained below pre-pandemic levels. Prescribing for new users remained largely unchanged; however, GLP-1 receptor agonists among injectable therapies showed an increasing trend, especially after restrictions were eased.

Previous studies from countries such as Bulgaria and France reported no significant changes in overall antidiabetic consumption during the COVID-19 pandemic (Mathieu et al., Reference Mathieu, Pambrun, Bénard, Pernelle and Faillie2022; Mitkova et al., Reference Mitkova, Stanimirova, Manova, Gerasimov, Mitov and Petrova2025). In contrast to these findings, our study demonstrated an increase in consumption in DuR, which was also accompanied by rising expenditure in that period. This finding appears particularly concerning given that Turkey has the highest diabetes prevalence in Europe, with approximately one-sixth of the population having a confirmed diagnosis of diabetes mellitus, a burden that has remained largely unchanged (International Diabetes Federation, 2021; International Diabetes Federation, 2025). Several behavioural and healthcare-related factors may help explain this observation, including panic stockpiling. Diabetes mellitus is a well-established risk factor for adverse COVID-19 outcomes (Sonmez et al., Reference Sonmez, Demirci, Haymana, Tasci, Dagdelen, Salman, Ata, Sahin, Emral, Cakal, Atmaca, Sahin, Celik, Demir, Ertugrul, Unluturk, Caglayan and Satman2021), leading to specific clinical recommendations for this vulnerable population (Bornstein et al., Reference Bornstein, Rubino, Ludwig, Rietzsch, Schwarz, Rodionov, Khunti, Hopkins, Birkenfeld, Boehm, Amiel, Holt, Skyler, DeVries, Renard, Eckel, Zimmet, Alberti, Geloneze, Chan, Mbanya, Onyegbutulem, Ramachandran, Basit, Hassanein, Spinas, Beuschlein and Mingrone2021). In parallel, substantial disruptions in routine follow-up and outpatient care during the pandemic, coupled with efforts to compensate for unmet healthcare needs through telehealth services (Adhikari et al., Reference Adhikari, Titus, Baum, Lopez, Kanchi, Orstad, Elbel, Lee, Thorpe and Schwartz2023), may have contributed to patients adopting a more cautious approach to glycemic control and medication adherence. Accordingly, a study from Malaysia has reported improvements in both parameters during the COVID-19 lockdown (Sim et al., Reference Sim, Chong, Loganadan, Hussein, Adam and Lee2023).

The parallel increase in expenditure further emphasizes the policy relevance of the observed period-related utilization trends. In this context, emergency medication-access policies for chronic diseases requiring continuous pharmacotherapy should consider not only treatment continuity and total utilization, but also expenditure monitoring during and after periods of healthcare disruption. Differences in the contribution of antidiabetic classes to pharmaceutical expenditure, particularly insulins, further suggest that such monitoring might benefit from a drug-class-specific perspective.

Notably, the increase in consumption in DuR was accompanied by a concomitant decline in prescribing. Access to routine healthcare services and physician consultations necessary for prescribing was substantially reduced during this extraordinary period, both in Turkey and internationally (Uğur and Durak, Reference Uğur and Durak2024; Zhong et al., Reference Zhong, Ma, Ionova, Bhatt, Vargas, Banh and Wilson2025). Moreover, prescribing levels in AfR remained below pre-pandemic levels, potentially reflecting heightened risk perception among this vulnerable population, which might have discouraged patients from attending healthcare facilities or seeking physician consultations for prescription-related care. The divergence in consumption and prescribing trends might be partly explained by the nationwide implementation of prescription-free dispensing practices aimed at facilitating patient access to medications during the pandemic (Turkish Medicines and Medical Devices Agency, 2020), a strategy that has also been associated with similar outcomes in other countries (Mathieu et al., Reference Mathieu, Pambrun, Bénard, Pernelle and Faillie2022). Comparable patterns associated with this practice were also identified in other components of the parent study examining the utilization of different drug groups over the same timeframe (Vizdiklar et al., Reference Vizdiklar, Aydin, Tazegul, Kaskal and Akici2024; Reference Zhong, Ma, Ionova, Bhatt, Vargas, Banh and Wilson2025a; Reference Vizdiklar, Aydin, Yilmaz and Akici2025b; Reference Vizdiklar, Aydin, Yilmaz and Akici2026). The results indicate that this practice might have played a role in maintaining treatment continuity and improving access to chronic medications. Conversely, it may also have carried a potential risk of inadvertently enabling patients to obtain antidiabetic drugs in quantities exceeding short-term therapeutic needs. From a pharmaceutical policy perspective, such practices might require safeguards to preserve continuity of care while avoiding unnecessary accumulation, wastage, and avoidable excess expenditure. The distinct consumption trends in AfR between injectable medications, predominantly insulins, and oral antidiabetic drugs might further support this interpretation. Overall antidiabetic consumption levels were slightly lower in AfR than those observed in DuR, with injectable medications returning to pre-pandemic levels, whereas oral antidiabetic drugs continued to exhibit a modest, statistically nonsignificant increase. Compared with oral antidiabetic drugs, insulin requires stricter storage conditions and is more vulnerable to stability issues related to environmental factors, rendering it less amenable to stockpiling (Heinemann et al., Reference Heinemann, Braune, Carter, Zayani and Krämer2020). Consequently, the utilization of previously accumulated insulin supplies may have contributed to the discontinuation of stockpiling behaviour and a subsequent reduction in demand after the most critical phase of the pandemic.

Given the central role of primary care in the management of type 2 diabetes (Seidu et al., Reference Seidu, Cos, Brunton, Harris, Jansson, Mata-Cases, Neijens, Topsever and Khunti2022), the observed changes in antidiabetic utilization have direct implications for primary care practice. The disruptions in routine healthcare services during the COVID-19 pandemic (Uğur and Durak, Reference Uğur and Durak2024; Zhong et al., Reference Zhong, Ma, Ionova, Bhatt, Vargas, Banh and Wilson2025), along with temporary medication-access policies (Turkish Medicines and Medical Devices Agency, 2020), emphasize the importance of coordination between primary care physicians and community pharmacists in supporting continuity of chronic disease management (Angibaud et al., Reference Angibaud, Jourdain, Girard, Rouxel, Mouhib, Nogueira, Rat and Huon2024). Such collaborative approaches might help preserve appropriate medication access, reduce unnecessary stockpiling, and support rational medicine use during future public health emergencies.

Among oral antidiabetic drugs, metformin and DPP-4 inhibitors accounted for the majority of consumption and are regarded as widely used agents in the pharmacological management of diabetes; notably, metformin is also among the top-utilized medicines in Turkey (Bajaj et al., Reference Bajaj, McCoy, Balapattabi, Bannuru, Bellini, Bennett, Beverly, Briggs Early, ChallaSivaKanaka, Echouffo-Tcheugui, Everett, Garg, Laffel, Lal, Matfin, Nelson, Neumiller, Pandya, Pekas, Peters, Pilla, Romeo, Rosas, Sadhu, Segal, Simmons, Szmuilowicz and ElSayed2026; Bayram et al., Reference Bayram, Vızdıklar, Aydın, İşli and Akıcı2020). However, the increase that persisted into AfR was largely driven by SGLT-2 inhibitors, a more recently introduced class of antidiabetic therapies (Bajaj et al., Reference Bajaj, McCoy, Balapattabi, Bannuru, Bellini, Bennett, Beverly, Briggs Early, ChallaSivaKanaka, Echouffo-Tcheugui, Everett, Garg, Laffel, Lal, Matfin, Nelson, Neumiller, Pandya, Pekas, Peters, Pilla, Romeo, Rosas, Sadhu, Segal, Simmons, Szmuilowicz and ElSayed2026). Consumption of these drugs increased to approximately 3-fold in DuR and reached nearly 4.5-fold pre-pandemic levels thereafter. This increase may partly reflect consensus recommendations issued by the American Diabetes Association and the European Association for the Study of Diabetes shortly before the start of the pandemic, emphasizing earlier use of SGLT-2 inhibitors in selected patient populations, particularly those with established cardiovascular disease, heart failure, or chronic kidney disease, and potentially contributing to switches from other antidiabetic agents and increased consumption of these therapies (Buse et al., Reference Buse, Wexler, Tsapas, Rossing, Mingrone, Mathieu, D.’Alessio and Davies2020). In addition, the introduction of SGLT-2 inhibitors in heart failure treatment algorithms in major cardiology guidelines during the pandemic period might have substantially contributed to the increase in consumption (Heidenreich et al., Reference Heidenreich, Bozkurt, Aguilar, Allen, Byun, Colvin, Deswal, Drazner, Dunlay, Evers, Fang, Fedson, Fonarow and Hayek2022; McDonagh et al., Reference McDonagh, Metra, Adamo, Gardner, Baumbach, Böhm, Burri, Butler, Čelutkienė, Chioncel, Cleland, Coats, Crespo-Leiro, Farmakis, Gilard, Heymans, Hoes, Jaarsma, Jankowska, Lainscak, Lam, Lyon, McMurray, Mebazaa, Mindham, Muneretto, Francesco Piepoli, Price, Rosano, Ruschitzka, Kathrine Skibelund, Christian Schulze, Abdelhamid, Aboyans, Adamopoulos, Anker, Arbelo, Asteggiano, Bauersachs, Bayes-Genis, Borger, Budts, Cikes, Damman, Delgado, Dendale, Dilaveris, Drexel, Ezekowitz, Falk, Fauchier, Filippatos, Fraser, Frey, Gale, Gustafsson, Harris, Iung, Janssens, Jessup, Konradi, Kotecha, Lambrinou, Lancellotti, Landmesser, Leclercq, Lewis, Leyva, Linhart, Løchen, Lund, Mancini, Masip, Milicic, Mueller, Nef, Nielsen, Neubeck, Noutsias, Petersen, Sonia Petronio, Ponikowski, Prescott, Rakisheva, Richter, Schlyakhto, Seferovic, Senni, Sitges, Sousa-Uva, Tocchetti, Touyz, Tschoepe, Waltenberger, Adamo, Baumbach, Böhm, Burri, Čelutkienė, Chioncel, Cleland, Coats, Crespo-Leiro, Farmakis, Gardner, Gilard, Heymans, Hoes, Jaarsma, Jankowska, Lainscak, Lam, Lyon, McMurray, Mebazaa, Mindham, Muneretto, Piepoli, Price, Rosano, Ruschitzka and Skibelund2021). Accordingly, while these agents may have been preferentially used in diabetic patients with COVID-19-related renal complications, the upward trend in their utilization appears to be only partly related to the pandemic itself (Bailey et al., Reference Bailey, Day and Bellary2022). In contrast to other oral antidiabetic agents, this pronounced peak in consumption was accompanied by stable prescribing levels in DuR and a subsequent significant increase in AfR, limited to ongoing users. As our prescribing data included only prescriptions for diabetes mellitus, SGLT-2 inhibitor use for non-diabetic indications was not reflected in prescribing trends, potentially explaining the stable trend in new users. Conversely, the rise in prescribing for ongoing users might indicate a preference for these agents in patients with diabetes and relevant comorbidities. From an expenditure perspective, the shift toward SGLT-2 inhibitors might also reflect broader changes in antidiabetic treatment patterns, underscoring the need to interpret short-term pharmaceutical spending alongside their potential long-term clinical value.

Injectable antidiabetic agents displayed a distinct utilization pattern, with insulin accounting for the majority of this category and following a fluctuating trend throughout the periods as discussed earlier, while GLP-1 receptor agonists exhibited a rise persisting into AfR. In addition to increased consumption, GLP-1 receptor agonists also showed rising prescribing trends in AfR among both ongoing and new users, a pattern that may be partly explained by their relatively recent introduction and their potential role in weight management among patients with diabetes. In particular, the increase in prescribing to first-time users in AfR might have been influenced by the demand for weight management associated with reduced physical activity and changes in dietary habits resulting from pandemic-related confinement (Maestre et al., Reference Maestre, Sospedra, Martínez-Sanz, Gutierrez-Hervas, Fernández-Saez, Hurtado-Sánchez and Norte2021; Sánchez-Sánchez et al., Reference Sánchez-Sánchez, Ramírez-Vargas, Avellaneda-López, Orellana-Pecino, García-Marín and Díaz-Jimenez2020; Shiba et al., Reference Shiba, Cowden, Counted, VanderWeele and Fancourt2024; Tazegül et al., Reference Tazegül, Aydın, Kıraç Aksuna, Dik, Aksoy, Koç, Tiryaki, Kurt, Görk, Erdoğan, Aysevinç, Çeçen Düzel, Topçu, Güven, Akıcı and Fak2026), which might have contributed to the preference for this drug class. Nevertheless, it should be noted that newer GLP-1-based therapies, including GLP-1 receptor agonists such as semaglutide and dual GLP-1/gastric inhibitory polypeptide receptor agonists such as tirzepatide, which are also approved for indications such as weight management, were not available in the Turkish pharmaceutical market during the study period. Consequently, the observed trends may not fully capture the extent of potential subsequent increases in GLP-1-based therapy use, which might become increasingly relevant for future evaluations of antidiabetic pharmaceutical expenditure and budget impact.

Several limitations should be considered when interpreting the findings of the study. The nationwide drug consumption dataset was generated from pharmaceutical wholesaler-to-pharmacy distribution records, whereas the prescribing dataset was based on a nationally representative projection of a sample of prescriptions issued for diabetes mellitus diagnoses excluding type 1 diabetes. Accordingly, the methodological limitations previously described in publications from the parent project, arising from the nature of the data, are also applicable to the present analysis (Vizdiklar et al., Reference Vizdiklar, Aydin, Tazegul, Kaskal and Akici2024; Reference Zhong, Ma, Ionova, Bhatt, Vargas, Banh and Wilson2025a; Reference Vizdiklar, Aydin, Yilmaz and Akici2025b; Reference Vizdiklar, Aydin, Yilmaz and Akici2026). The use of these drugs for different indications could not be distinguished in the consumption data, which constitutes an inherent limitation of the study. Although prescribing data were restricted to prescriptions issued for non-type 1 diabetes mellitus, which partly mitigated this limitation, this discordance should be considered when interpreting the findings in a broader context. Expenditure calculations were based on the euro/Turkish lira exchange rate set by the national health authority, which is updated annually or at multi-month intervals; therefore, short-term fluctuations in real-world exchange rates may not have been fully reflected in the results. Finally, to minimize potential confounding related to the profound disruptions in healthcare access after the February 2023 earthquake in Turkey (Manirambona et al., Reference Manirambona, Obnial, Musa, Miranda, Haruna and Shomuyiwa2023), AfR was deliberately limited to nine months.

In conclusion, this study demonstrated a sustained increase in antidiabetic drug consumption after the implementation of COVID-19-related public health restrictions, with elevated levels persisting even after the subsequent easing of most measures. Policies implemented to facilitate access to medications during extraordinary circumstances might have contributed to the maintenance of these utilization levels, alongside other potentially reinforcing factors such as panic stockpiling, improved adherence to glycemic control, and the introduction of newer pharmacological options into routine clinical practice. Findings from a country with a high burden of diabetes, such as Turkey, might therefore provide valuable insights for strengthening primary care-based diabetes management, enhancing coordination between family physicians and community pharmacists, and supporting medication access, pharmaceutical expenditure monitoring, and preparedness for future public health emergencies.