Studies on the use of oral isotretinoin during the COVID-19 pandemic remain limited, and none have been conducted in France. They reported early treatment interruptions and an increased use of teleconsultations to ensure continuity of acne-related care [7,8,9,10,11]. A Turkish study found no difference between the pandemic and nonpandemic periods regarding the initiation of isotretinoin treatment [12]. However, most were single-center, had small sample sizes, and were conducted over short study periods. The present study spans the entire French COVID-19 period and is based on SNDS data, which cover 99% of the French population, ensuring a representative dataset with a large sample size.

Furthermore, the management of acne with oral isotretinoin and its organizational constraints vary across countries. During the COVID-19 pandemic, country-specific measures reshaped healthcare systems, potentially affecting isotretinoin use and compliance with recommendations. Analyzing the impact of the pandemic on isotretinoin use in the French context was therefore particularly relevant.

On the basis of previous European and French studies, we expected an increase in oral isotretinoin dispensations during the first lockdown due to stockpiling, followed by a decrease in treatment initiations [5, 6, 15]. However, we observed a decrease in isotretinoin dispensations during the first lockdown and at the end of 2022, and an increase over the remainder of the period. This upward trend may be explained by: (i) a catch-up phenomenon following the decline in initiations during the first lockdown, (ii) the increase of teleconsultation use, (iii) the mandatory wearing of masks, which can exacerbate acne, and (iv) individuals’ heightened attention to their health and well-being during the pandemic [16,17,18,19].

Furthermore, we expected a more marked and prolonged worsening of recommendations due to patients’ fear of contamination and the increased involvement of healthcare professionals and laboratories in managing COVID-19 patients and screening. However, noncompliance with prescription renewal and PT worsened mainly during the first lockdown and tended to stabilize thereafter, suggesting the adaptation of patients and healthcare professionals to a prolonged pandemic context, likely facilitated by the use of teleconsultation. The rise of this type of consultation during pandemic was observed in other countries [7, 10, 20]. It offers a number of advantages, such as being well-suited for dermatology, improving access to care, ensuring high patient satisfaction, and saving time for both the patient and physician. However, it also has disadvantages, including concerns about security and privacy, its applicability to all patients, and the need for appropriate equipment and internet access [20]. Conversely, noncompliance with the recommendation for treatment initiation by dermatologists showed a sustained decline, not only during the first lockdown but throughout the COVID-19 period.

Overall, compliance with recommendations remained suboptimal in France during the COVID-19 pandemic, which justifies the actions recently implemented by the French Medicines Agency (ANSM) in 2023: educational videos for patients and a QR code on medication boxes, patient brochures, and letters to healthcare professionals. The aim of the videos is to popularize the information, making it concise and easy to understand. The QR code enables faster and broader dissemination among both patients and healthcare professionals, while centralizing the information. This approach seems particularly relevant in today’s digital age, but further study is needed to assess the efficacy of these new measures.

Other interventions, such as electronic transmission of prescriptions and automated alert systems in prescribing and dispensing software, could improve the practices of prescribers and pharmacists [21, 22]. Home pregnancy tests in the American IPLEDGE program appear to be an effective model of care by limiting logistical obstacles [23], although the risk of falsification of results remains a concern [24].

Our study has several limitations. First, the assessment of drug exposure based on records of dispensed drug supplies may have led to its overestimation. However, 82% of treatments were dispensed at least twice, suggesting that patients were using isotretinoin. Second, owing to the lack of biological test results in the SNDS database, we assumed that pregnancy tests were negative at the time the treatment was prescribed and dispensed. Third, the prescriber specialty is misinformed for hospital practitioners in the SNDS database, and we considered all of them dermatology specialists as soon as they prescribed isotretinoin, inducing a potential classification bias.