Background and importance In response to the COVID-19 pandemic, scientific societies and regulatory agencies quickly reviewed any available evidence to fill the therapeutic gap. In this context, many drugs were used with an uncertain benefit–risk profile that needs to be evaluated.

Aim and objectives To analyse the safety profile of experimental therapies that were used at the beginning of the COVID-19 pandemic.

Material and methods A retrospective observational study was conducted to analyse the safety profile of anti-COVID therapy accessible according to the protocols that were approved. Patients admitted with a COVID-19 diagnosis between March and May 2020 who had an adverse event (AE) coded in the discharge/death medical report were obtained from the National Minimum Data Set. The suspected drug was identified based on previous information. Those with AEs attributed to anti-COVID therapy were selected. The causal relationship was evaluated using Naranjo’s algorithm (NA).

Results 141 AEs were coded in 105 patients admitted with a diagnosis of COVID-19. 60.3% were attributed to anti-COVID therapy in 66 patients with a median age of 72 years (95% CI 68 to 76), 62.1% men (37.9% women). AE severity was: 63.5% mild, 29.4% moderate and 7.1% severe. 23.5% of AEs did not require intervention, 37.6% required pharmacological treatment, 35.3% suspension of the drug, 2.4% close monitoring and 1.2% dose reduction.

Conclusion and relevance NA established a probable drug–AE causal relationship for most events. Most AEs were moderate to mild in severity but 75% required medical intervention. Consequently, it is important to know the AE–drug relationship to ensure a favourable benefit–risk profile, especially for experimental therapies.

Conflict of interest No conflict of interest