A Comparative Analysis Between Antibiotic- and Nonantibiotic-Associated Delayed Cutaneous Adverse Drug Reactions

Abstract

The difference in clinical presentation, causality assessments, and outcomes of patients with delayed antibiotic-associated cutaneous adverse drug reactions (AA-cADR) and nonantibiotic-associated (NA)-cADR is ill defined. We examined the etiology of AA-cADR, with regard to the type of antibiotic exposure, allergy labeling, and patient outcomes, in comparison with NA-cADR. A retrospective observational inpatient cohort study of cADR was performed from January 2004 to August 2014. Patients were divided into AA-cADR and NA-cADR groups for analysis. cADR was defined as erythema multiforme, fixed drug eruption, acute generalized erythematous pustulosis, drug reaction with eosinophilia and systemic symptoms (DRESS), drug-associated linear IgA disease, Stevens-Johnson syndrome, and toxic epidermal necrolysis. Of the 84 patients with cADR, 48% were AA-cADR. Male sex (60% vs 32%, P= .004), median length of stay (14.5 vs 11 days, P= .05), median Charlson comorbidity index (3 vs 1, P= .03), and inpatient mortality (20% vs 5%, P= .04) were higher in AA-cADR compared with NA-cADR. The median drug latency was lower in AA-cADR (6 vs 20 days, P= .001). Sulfonamide antibiotics and glycopeptides were implicated in 20% of AA-cADR. DRESS was more frequently reported in AA-cADR. After cADR diagnosis, further antibiotic therapy was administered in 64% of patients, higher in AA-cADR (75%, 30 of 40) compared with NA-cADR (55%, 24 of 44) (P= .06). Fluoroquinolones (53% vs 21%, P= .02), glycopeptides (vancomycin and teicoplanin; 70% vs 38%, P= .05), and carbapenems (33% vs 13%, P= .11) were used more commonly in AA-cADR. Antibiotics were the cause of cADR requiring hospital admission in 48% of episodes, and were associated with longer length of stay, higher age-adjusted Charlson comorbidity index, shorter drug latency, and mortality. In AA-cADR, glycopeptide and sulfonamide antibiotic exposure predominated.