Utility of IFN-γ Enzyme-Linked Immunospot as Adjunctive Diagnostic for Severe T Cell–Mediated Reactions to Cotrimoxazole in High HIV and Tuberculosis Settings

Background: Cotrimoxazole is a broad-spectrum antibiotic associated with severe cutaneous adverse drug reactions (SCARs). In persons living with HIV, cotrimoxazole is typically used in combination with other SCAR-associated medications, creating challenges for culprit drug ascertainment. Objective: To investigate the utility of IFN-γ release enzyme-linked immunospot (ELISpot) in identifying cotrimoxazole as the culprit drug in SCAR. Methods: Registry of Severe Cutaneous Adverse Reactions–validated SCAR cases from the prospective AFRISCAR patient registry with Naranjo and/or Algorithm of Drug Causality for Epidermal Necrolysis causality scores completed were eligible. Diagnostic assessments included cotrimoxazole patch testing, IFN-γ ELISpot, and oral challenge. ELISpot was performed on cryopreserved peripheral blood mononuclear cells using validated concentrations of cotrimoxazole (trimethoprim/sulfamethoxazole), its metabolite 4-nitroso-sulfamethoxazole, and sulfamethoxazole (sulfamethoxazole). Results: Cotrimoxazole was the probable culprit in 79 SCAR cases (55 drug reactions with eosinophilia and systemic symptoms, 10 cases of Stevens-Johnson syndrome/toxic epidermal necrolysis, and 14 generalized bullous fixed drug eruptions) based on clinical drug causality assessment. Mean age was 39 years (SD, 11 years), and 90% were persons living with HIV (median CD4 count, 77; interquartile range, 25-142). Eighteen of 73 IFN-γ ELISpots (25%) were positive, nine of which reacted only to 4-nitroso-sulfamethoxazole (50%). Patch tests were all negative, including three of 13 with IFN-γ ELISpot and four of 13 with oral challenge positivity. IFN-γ ELISpot positivity was similar in Stevens-Johnson syndrome/toxic epidermal necrolysis (25%) and drug reactions with eosinophilia and systemic symptoms (26%), and no test was positive for samples older than 90 days from SCAR onset. IFN-γ ELISpot sensitivity and specificity were 38% (95% CI, 9-76) and 100% (95% CI, 40-100), respectively, using a cutoff of spot-forming units/million of 50 or greater. Conclusions: Cotrimoxazole IFN-γ ELISpot had overall low but superior sensitivity to patch testing, offering good rule-in utility. Ongoing optimization of in vivo and in vitro diagnostic tools for cotrimoxazole SCAR is still required.