African sleeping sickness (SS) is a severe, potentially lethal parasitic disease. The treatments of choice are the antiparasitic agents suramin, which is adrenotoxic, and/or melarsoprol. We evaluated the functional integrity of the hypothalamic-pituitary-adrenal (HPA) axis of patient with SS before, during, and after therapy with suramin and/or melarsoprol, in two sequential stages. First, we employed the standard adrenocorticotropic hormone (ACTH) 1-24 stimulation test (250 μg i.v.) to assess the maximal adrenocortical responsiveness of 69 patients with SS and 38 normal controls. We demonstrated paradoxically subnormal Cortisol responses before suramin therapy [net Cortisol response 60 min after stimulation: 10.5 ± 2.9 (mean ± SE) vs. 17.5 ± 1.0 μg/dl for controls, p = 0.004], with 27% of the patients falling within the adrenal insufficiency range (stimulated cortisol concentration <20 μg/dl). These responses subsequently and unexpectedly improved with suramin and/or melarsoprol therapy. Second, we performed a human corticotropin-releasing hormone (hCRH) test (100 μg i.v.) in 68 additional patients with SS and 14 control subjects to examine whether the glucocorticoid deficiency observed was primary and/or secondary. Compared to controls, the ACTH and Cortisol responses to hCRH were blunted (ACTH after 60 min: 29 ± 7 vs. 58 ± 8 pg/ml in controls, p = 0.014; cortisol: 15.2 ± 1.5 vs. 19.6 ± 0.7 μg/dl, p = 0.018), suggesting the presence of secondary adrenal insufficiency. There was improvement of both ACTH and cortisol responsiveness to hCRH with therapy, with cortisol recovery occurring before ACTH, suggesting an additional primary component of adrenal dysfunction in these patients. Plasma concentrations of tumor necrosis factor (TNF)-α (16.0 ± 4.1 vs. 2.9 ± 1.4 pg/ml in controls, p = 0.003) and interleukin (IL)-6 (19.2 ± 7.3 vs. 1.3 ± 0.2 pg/ml, p = 0.0001), but not IL-1β (2.0 ± 0.2 vs. 0.9 ± 0.2, p = NS), were elevated when adrenocrotical function impairment and disease activity were at their maximum, but gradually decreased into the normal range with therapy. We found a negative correlation between baseline cytokine concentrations and maximal Cortisol concentrations during hCRH testing (TNF-α: r = –0.31, p = 0.003; IL-6: r = –0.34, p = 0.002). We conclude that unmedicated SS is associated with significant impairment of adrenocortical function which is reversed with suramin and/or melarsoprol therapy in the majority of patients. This impairment may be due to the elevated plasma cytokine concentrations, and may represent a natural adaptation of the HPA axis in inflammatory states. A controlled therapeutic trial is necessary to demonstrate whether supplemental glucocorticoids could be beneficial in SS.