Objectives: To quantify the risk of skin reactions to antibacterial drugs under everyday circumstances in a large population with automated data from general practitioners (GP). Design: A retrospective cohort study in a dynamic population. Setting: Data came from the Integrated Primary Care Information (IPCI) database. The IPCI database consists of all data on consultations, morbidity, and prescriptions and other interventions, as registered by the GP in a source population of approximately 150,000 persons. Methods: The study period started on April 1, 1994, and ended on September 30, 1995. All patients who were treated with an antibacterial drug were enrolled on the first day of starting treatment until the end of the study period or until the occurrence of one or more of the following diagnoses within the risk period: allergic reaction, rash, erythema, pruritus and urticaria, or a notification of a skin reaction in the free text. Subsequently, patient profiles were assessed by two authors who were blinded to exposure. The risk period was defined as the legend duration of the antibacterial drug plus 14 days to control for carry-over of drug effects and delay in patient presentation. Age, gender, and comedication were examined as potential confounders. Results: In the study period 13,679 patients received 19,961 prescriptions of an antibacterial drug. It concerned 5330 men (39.0%) and 8349 (61.0%) women with a mean age of 41 and 42 years, respectively. One hundred thirty-five patients developed a skin reaction in the risk period. Rash, pruritus, urticaria, and miscellaneous skin reactions were encountered in 76 (56.3%), 18 (13.3%), 19 (14.1%), and 22 (16.3%) patients, respectively. The three most frequently reported causes of skin reactions were combinations of trimethoprim with sulfonamides (2.1% of users; incidence density [ID]: 2.1/1000 exposed days), fluoroquinolones (1.6% of users; ID: 1.5/1000 person days), and penicillins (1.1% of users; ID: 1.3/1000 person days). Compared to tetracyclines, the broad spectrum penicillins showed an incidence density ratio (IDR) of 3.7, the combination of amoxicillin with clavulanic acid of 3.3, the fluoroquinolones of 2.8, and the combination of trimethoprim with sulphonamides of 4.4. The presence of infectious mononucleosis increased the risk of rash in amoxicillin users with a factor of 58. Conclusions: We found that the frequency of skin reactions to antibacterial drugs in general practice is around 1% and highest for the combination of trimethoprim with sulphonamides, penicillins, and fluoroquinolones. The outpatient incidence for skin reactions is probably lower than the incidence in hospitalized patients. Although this may be partly explained by negative misclassification, it is also likely that the actual incidence is lower as some hospitalized patient groups may be more prone to develop a skin reaction.

Additional Metadata
Keywords Antibacterial agents, Cohort studies, Pharmacoepidemiology, Skin reactions
Persistent URL dx.doi.org/10.1016/S0895-4356(98)00041-9, hdl.handle.net/1765/61094
Journal Journal of Clinical Epidemiology
van der Linden, P.D, van der Lei, J, Vlug, A.E, & Stricker, B.H.Ch. (1998). Skin reactions to antibacterial agents in general practice. Journal of Clinical Epidemiology, 51(8), 703–708. doi:10.1016/S0895-4356(98)00041-9