Patients suffering from chronic mountain sickness (CMS) have excessive erythrocytosis. Low level cobalt toxicity as a likely contributor has been demonstrated in some subjects. We performed a randomized, placebo controlled clinical trial in Cerro de Pasco, Peru (4380 m), where 84 participants with a hematocrit (HCT) >= 65% and CMS score > 6, were assigned to four treatment groups of placebo, acetazolamide (ACZ, which stimulates respiration), N-acetylcysteine (NAC, an antioxidant that chelates cobalt) and combination of ACZ and NAC for 6 weeks. The primary outcome was change in hematocrit and secondary outcomes were changes in PaO2, PaCO2, CMS score, and serum and urine cobalt concentrations. The mean ( SD) hematocrit, CMS score and serum cobalt concentrations were 69 +/- 4%, 9.8 +/- 2.4 and 0.24 +/- 0.15 mu g/l, respectively for the 66 participants. The ACZ arm had a relative reduction in HCT of 6.6% vs. 2.7% (p = 0.048) and the CMS score fell by 34.9% vs. 14.8% (p = 0.014) compared to placebo, while the reduction in PaCO2 was 10.5% vs. an increase of 0.6% = 0.003), with a relative increase in PaO2 of 13.6% vs. 3.0%. NAC reduced CMS score compared to placebo (relative reduction of 34.0% vs. 14.8%, p = 0.017), while changes in other parameters failed to reach statistical significance. The combination of ACZ and NAC was no better than ACZ alone. No changes in serum and urine cobalt concentrations were seen within any treatment arms. ACZ reduced polycythemia and CMS score, while NAC improved CMS score without significantly lowering hematocrit. Only a small proportion of subjects had cobalt toxicity, which may relate to the closing of contaminated water sources and several other environmental protection measures.