In this ITS analysis of French outpatient pharmacy sales data, we found an increase in overall antidepressant sales and a decrease in overall anxiolytic sales over the period from 2012 to 2019. With onset of the COVID-19 pandemic, we observed a relevant and persisting increase of antidepressant sales trends compared with the prepandemic trend. We estimated that an excess of about 113 DDD/TID of antidepressants was sold from May 2020 until December 2022 compared with what would have been expected. Meanwhile, we found an immediate increase in anxiolytic sales in February 2020, but no long-lasting change in trend compared with before the pandemic. While anxiolytic sales were elevated from February 2020 until December 2022 with an estimated excess of 34 DDD/TID, they returned to what would have been expected based on the prepandemic trend by December 2022. We observed no clear association between antidepressant and anxiolytic sales and stringency of pandemic mitigation measures.

The observed increases in the antidepressant sales trend from 2012 to 2019 can likely be explained by an increasing prevalence of depression.2 3 This may be driven by increased diagnosis and treatment of depressive disorders and by changes in the population structure, as depression is more prevalent in the growing elderly and migrant populations.35 Antidepressants demonstrating an increasing prepandemic trend were duloxetine, fluoxetine, mianserin, mirtazapine, paroxetine, sertraline, venlafaxine and vortioxetine, which are commonly prescribed as first-line treatment in depressive or anxiety disorders.36 Meanwhile, sales of tricyclic antidepressants, which were not recommended as first-line treatment,37 and certain selective serotonin reuptake inhibitors with cardiac adverse effects such as escitalopram decreased.38

Anxiolytic sales demonstrated a declining trend from 2012 to 2019 in line with other studies, as their utilisation—especially the prescription of benzodiazepines—was increasingly discouraged in clinical practice guidelines due to their risk for adverse effects and dependence.6 7 This decline was observed for most anxiolytics, with the exception of alprazolam, clotiazepam, diazepam and oxazepam. Those drugs are recommended first-line drugs for short-term treatment of acute crises of panic disorders and generalised anxiety disorders, while antidepressants such as selective serotonin reuptake inhibitors are recommended for their long-term treatment.37 39

Our findings of an increase in trend of antidepressant sales with the COVID-19 pandemic are in accordance with several other studies conducted earlier during the pandemic and based on shorter observation timeframes, which have reported increases in the prevalence of depressive and anxiety disorders alongside increased sales or consumption of psychotropic drugs.8 10–18 One early study from the United Kingdom argued that temporary effects of the pandemic could be responsible for these increases and suggested that early increases of antidepressants and anxiolytics sales may be due to stockpiling effects because of the lockdown.10 Potential stockpiling was also reported in several Scandinavian countries.16 While sales of both drug classes spiked in March 2020 likely for this reason in our study (figure 2), increases in sales were sustained for almost 3 years suggesting a long-term impact. One study from Portugal found that there was a sustained decrease in antidepressant prescriptions over the first year after pandemic onset.40 Another study from Spain found a larger relative increase in anxiolytic prescriptions compared with antidepressants, which contrasts with our findings of a larger relative increase in antidepressant sales.11 Finally, a study from Croatia reported that observed increases were not statistically significantly different from prepandemic trends based on an ITS of annual data (likely resulting in lower precision than monthly data).18 Methodological differences and shorter observation periods may explain at least some of the differences with our study, but specific healthcare systems, provision of mental health services, and cultures also need to be considered.

There is evidence that quarantine and isolation, social distancing measures and the uncertainties of the pandemic contributed to increases in depressive and anxiety disorders, which may be especially pronounced for individuals who already had a medical condition prior to the pandemic.41 42 Several studies have also reported increases in social anxiety in the population with the pandemic.43 This may be partly reflected by our observation of a temporary increase in anxiolytic sales which returned to expected levels by the end of 2022. However, we observed no association between antidepressant and anxiolytic sales and stringency, and the increase in trend of antidepressant sales persisted until the end of 2022 when stringency was greatly reduced. The longer timeframe of this study, thus, provides arguments for a real and sustained impact of the pandemic on antidepressant sales, making explanations other than short-term changes more likely.

In our study, we observed sales trend increases primarily for recommended first-line antidepressants including fluoxetine, mianserin, paroxetine, sertraline, venlafaxine and citalopram.36 37 The results suggest an overall increasing prevalence of depressive and anxiety disorders due to the pandemic, in line with other studies.8 However, antidepressant sales are only a proxy for the prevalence and several reasons may have contributed to this observation: First, the pandemic and its consequences on the healthcare workforce have severely affected the provision of health services in general practice and outpatient mental healthcare in France.20 21 Before the pandemic, these sectors had already experienced important challenges by shortages of medical personnel.19–21 As non-pharmacological supportive treatments are time-consuming, require personal contact and consultations with psychologists were not fully covered by the French general social security system,19 newly diagnosed (incident) cases of depression and anxiety may consequently have been treated more frequently using medication than prior to the pandemic. Additionally, pre-existing conditions previously managed non-pharmacologically may have had to be managed with antidepressants or anxiolytics. Second, pre-existing psychiatric conditions may have exacerbated with the pandemic, leading to an increase in antidepressant prescriptions for individuals already suffering from mental health issues prior to pandemic onset. Third, the sustained increases in the trend of antidepressant sales may also be partially due to an increased prevalence of anxiety disorders, for which long-term treatment with antidepressants is recommended.37 39 Fourth, potential long-term effects of SARS-CoV-2 infection, also termed post-COVID-19 conditions or Long COVID, may have additionally contributed to an increased prevalence of depressive and anxiety disorders.44 45 And last, other concurrent events such as global warming, economic crises and international geopolitical events could have increased the prevalence of depression and anxiety in France. All these reasons can explain the sustained increase in trend of antidepressant sales until the end of 2022 and the lack of an association between antidepressant and anxiolytic sales and stringency.

This study presents a detailed analysis of routine healthcare data on antidepressant and anxiolytic consumption over the last decade. The used ITS approach is a strong quasi-experimental design allowing to assess temporal changes caused by impactful events while controlling for prior trends.32 Meanwhile, several limitations need to be considered. First, ITS analyses assume that there are no relevant changes in the underlying population. While it is likely that migration and changes in age distributions have occurred over the observed timeframe, we consider it unlikely that such would have been responsible for the observed trends after pandemic onset. Second, the data set did not include drugs prescribed in the inpatient setting. Despite limitations in inpatient healthcare provision during the pandemic, we did not consider this a relevant issue since the majority of antidepressant and anxiolytic drugs are expected to be prescribed in outpatient settings. Third, the study used DDD/TID as a primary outcome measure, which may not necessarily represent changes in the prevalence of individuals being prescribed the drugs (ie, the prevalence of depression or anxiety disorders) or the number of prescriptions per individual. However, we consider it a strong proxy for the number of prescriptions and the prevalence of corresponding mental health conditions. Fourth, the applied ITS approaches bear several assumptions regarding seasonality, autocorrelation and trends over time. However, we performed various sensitivity analyses, which resulted in similar estimates and conclusions. Finally, our analysis covered the entire population of France and was not designed to evaluate regional or urban–rural differences. While in line with several other studies, it is unclear to what extent the findings are generalisable to other countries in Europe and beyond.