The effects of diazepam and the incidence of hypoxaemia on the course of acute chloroquine poisoning were studied prospectively in 21 patients. Were excluded patients who had ingested more than one drug or who had major symptoms on admission (systolic blood pressure less than 80 mmHg; QRS greater than 0.12 s; cardiac dysrhythmias, respiratory disturbances). Arterial blood gases were measured on admission (T0) and 15 min after 0.5 mg.kg-1 of diazepam had been given (T1). Gastric lavage was carried out as soon as the results of the blood gases had been obtained, and after treatment of hypoxaemia (PaO2 less than 90 mmHg). An infusion of diazepam (1 mg.kg-1.day-1) was then given. Arterial blood gases were measured after 1 (T2), 6 (T3), 12 (T4) and 24 h (T5). Hypoxaemia was present on admission in four patients who had a PaO2 = 75 +/- 10 mmHg (Pa(sys) = 130 +/- 19 mmHg; blood chloroquine concentration = 8.2 +/- 5.2 mumol.l-1; kaliemia = 3.1 +/- 0.3 mmol.l-1; PaCO2 = 35 +/- 1 mmHg). In two patients, hypoxaemia decreased after the initial dose of diazepam (T1); however, oxygen was still required by the other two at that time. Oxygen was no longer needed by any patient at T2, as all the blood gas values had returned to normal.(ABSTRACT TRUNCATED AT 250 WORDS)