Fifty-two adult patients with cerebral malaria were randomly categorized into two groups to receive either quinine dihydrochloride (Qn) alone or a combination of Qn and pentoxifylline (Px). Thirty-two of them received intravenous (i.v.) Qn (group I), and 20 patients (group II) received i.v. Qn along with parenteral Px support (10 mg/kg/day) for the initial 3 days. There was significant improvement in coma resolution time in group II (21.6 +/- 13.9 h) in comparison with group I (63.5 +/- 19.7 h) (P < 0.001), and mortality was 25% of patients in group I against 10% patients receiving Px adjunct (P > 0.05). Three days post-therapy, serum tumour necrosis factor-alpha (TNF-alpha) levels decreased significantly in patients on Px support (day 0 TNF = 415.62 +/- 477.80 pg/ml; day 3 TNF = 47.92 +/- 27.9 pg/ml; P = 0.0029). There was no significant change in TNF levels in those on quinine alone (day 0 TNF = 477.08 +/- 933.90 pg/ml; day 3 TNF = 589 +/- 602.3 pg/ml; P > 0.05). There were no serious side-effects necessitating withdrawal of patients receiving Px therapy.