RESUMO
STUDY OBJECTIVES: To determine the frequency with which reported antibiotic allergies alter drug selection and to assess the validity of these allergies. DESIGN: Retrospective medical record review, with concurrent interviews conducted in a selected subgroup of patients. SETTING: Tertiary care academic medical center. PATIENTS: Three hundred patients with at least one documented antibiotic allergy and who received an antibiotic while hospitalized. MEASUREMENTS AND MAIN RESULTS: Data were collected to determine the patients' allergies documented in the medical record. The first antibiotic regimen that each patient received while hospitalized was evaluated for deviation from the standard of care as determined from institutional protocols, recommendations in the literature, and expert opinion. A total of 416 allergies to antibiotics were reported. Penicillins were the agents most commonly reported (198 reports), followed by sulfonamides, cephalosporins, macrolides, and fluoroquinolones. The reported allergies altered antibiotic therapy in 91 (30.3%) patients. Report of a penicillin or cephalosporin allergy and use of antibiotics for prophylaxis were strong predictors of altered therapy. The subgroup consisted of 100 patients who were interviewed to determine the specific details of their reported allergic reactions. For 22 of the 100 patients, major discrepancies were found between their verbal reports and medical record documentation. The Naranjo adverse drug reaction probability scale was used to determine the validity of their reactions. Among these 100 patients, 109 (78.4%) of 139 reported reactions to antibiotics were deemed to be allergic in nature. For 55 (50.5%) of the 109 allergic reactions, the Naranjo score was 5 or greater, which correlates with probable to definite validity. CONCLUSION: Discrepancies between the medical record and in-depth allergy histories are common, and the validity of reported allergic reactions is frequently questionable. Because documentation of an antibiotic allergy frequently alters therapy, increased effort to verify these reactions may be beneficial.
