Is hydroxylamine-induced cytotoxicity a valid marker for hypersensitivity reactions to sulfamethoxazole in human immunodeficiency virus-infected individuals?

Hypersensitivity (HS) reactions to sulfonamides and sulfones continue to limit their use in human immunodeficiency virus (HIV)-infected individuals. In vitro cytotoxicity of hydroxylamine metabolites toward peripheral blood mononuclear cells (PBMCs) has been proposed as a marker for these HS reactions. To test the validity of this in vitro system, we determined the selective susceptibility of PBMCs from HIV-infected patients to the cytotoxic effects of hydroxylamine metabolites of sulfamethoxazole (SMX) and dapsone (DDS). Concentration-cytotoxic response data were collected using PBMCs from 12 sulfa-HS (10 SMX-HS and 2 SMX/DDS-HS) and 10 sulfa-tolerant HIV-infected individuals. Although sulfamethoxazole hydroxylamine (SMX-NOH) and dapsone hydroxylamine (DDS-NOH) both caused concentration-dependent increases in cell death, DDS-NOH was significantly more potent in each subject (P <.0001). A comparison of a variety of mean data for sulfa-HS and -tolerant patient populations failed to demonstrate the increased susceptibility of PBMCs from HS patients, noted by others, to either SMX-NOH or DDS-NOH. Moreover, any trend toward an increased susceptibility of PBMCs from HS patients was eliminated when adjusted for control cell death. PBMCs from sulfa-HS patients showed significantly greater susceptibility to the stress of short term in vitro incubation (P <. 02). Mean (S.D.) vehicle control cell death values were 24.1% (7.6%) for HS patients and 17.1% (4.4%) for tolerant patients. No significant correlation was observed between hydroxylamine-induced or control cell death and any of the recorded clinical parameters. Although several potential reasons are proposed to explain the disparity with past investigations, the data suggest that in vitro cytotoxicity is not a valid marker for HS reactions in HIV-infected individuals using currently accepted experimental procedures.