Background Antimicrobials are widely used in hospitals and are often associated with adverse drug reactions (ADRs). The objective of this study was to determine the incidence of ADRs caused by antimicrobials and classify them according to the type of reaction, the class of antimicrobials used, causality, severity and avoidability.

Methods A prospective cohort study was carried out with paediatric patients for 6 months. Causality was verified using the Naranjo and Liverpool algorithms, the severity was verified with the adapted scale of Hartwig and the avoidability was verified with the Liverpool Avoidability Assessment Tool.

Results A total of 303 patients were followed, and 18.2% (55/303) of them had one or more ADRs during the hospital stay. Just over half of the patients (28/55) had diarrhea. The most used antimicrobials were beta-lactams and second-generation cephalosporins. Suspicions were classified mainly as possible 78.6% (55/70) according to the Naranjo algorithm, and as probable 48.6% (34/70) according to the Liverpool algorithm. The antimicrobial most involved with ADRs was cefepime. The risk of manifesting ADR was greater with the use of some antimicrobials such as clindamycin (relative risk (RR) 3.0, CI 1.67 to 5.4), as well as with the increase in hospitalisation days (OR 1.022, CI 1.008 to 1.036) and in the number of antimicrobials prescribed (OR 1.649, CI 1.360 to 2.001).

Conclusion ADRs were observed in approximately one-fifth of patients and were mostly gastrointestinal, moderate, unavoidable and with variable causality, depending on the algorithm used.