RESUMEN
INTRODUCTION: Adverse drug reactions (ADRs) to antiseizure therapy can worsen the quality of life, reduce adherence, and potentially lead to treatment discontinuation and uncontrolled seizures. OBJECTIVES: The aim of the study was to develop a prognostic model for ADRs to antiseizure therapy in adult patients with epilepsy from Colombia. METHODS: This case-control study included adult patients with epilepsy, who were separated into two groups: one group with ADRs to antiseizure therapy (cases), as determined by a complete evaluation conducted by an epileptologist, and another group without ADRs (controls). Variables were analyzed to identify statistical differences between the two groups and were then selected to construct a prognostic model using logistic regression. The Bonferroni method was applied for multiple comparisons. RESULTS: Three hundred fifty-four patients with epilepsy were studied. One hundred and fifty (42%) patients had ADRs and 204 (57%) patients did not have ADs. A total of 362 ADRs were reported, with a third of them being general symptoms and most frequently occurring with older-generation antiseizure drugs (58%). Female sex, drug-resistant epilepsy, LEV, and CZP were risk factors, whereras the presence of tumoral etiology, absence of seizure triggers, and VPA were identified as protective factors. A prognostic model was constructed using previously reported risk factors for ADRs to antiseizure therapy and other variables available in this population study. In the multivariable analysis, the number of previously used antiseizure drugs (1, 2, or ≥3), TPM, CZP, LEV, PHT, and female sex were predictors of ADRs. The corrected p-values were estimated by the Bonferroni method; however, not all the variables achieved statistical significance with this adjustment. CONCLUSIONS: In adult patients with epilepsy from Colombia, we found that the number of previously used antiseizure drugs, TPM, CZP, LEV, PHT, and female sex were predictive factors for ADRs to antiseizure therapy.