The fate of dapsone hydroxylamine has been investigated in diabetic and normal human erythrocytes. In erythrocytes from four type 1 (insulin dependent) diabetic subjects, there was a significant decrease in dapsone hydroxylamine-mediated methaemoglobin formation compared with cells drawn from normal individuals (P < 0.01). However, the ability of the diabetic cells to detoxify the hydroxylamine to dapsone was not correspondingly reduced and was not different to normal cells. The initial;ate of the accelerating effect of diethyl dithiocarbamate (DDC) on hydroxplamine-mediated methaemoglobin and dapsone formation was significantly reduced in diabetic compared with normal cells. There was no significant difference in hydroxylamine-dependent methaemoglobin formation between diabetic erythrocytes pretreated with either statil or sorbinil and untreated diabetic cells. Dapsone recovery in diabetic erythrocytes incubated with statil was not significantly different from statil-free incubations. However, in the presence of sorbinil, there was a marked reduction in dapsone formation at all four time points, (P < 0.001 at 15 min). Mean measured levels of glutathione did not differ significantly between the normal (380 +/- 30.9 mg/L; N = 8) and diabetic (349 +/- 58.7 mg/L; N = 8) volunteers. In summary, although diabetic erythrocytes were less sensitive to the effect of dapsone hydroxylamine-mediated methaemoglobin formation in comparison with normal cells, glutathione-dependent hydroxylamine reduction to dapsone was unaffected.