In an open, randomized, controlled therapeutic trial, 56 children with cerebral malaria (CM) were randomly assigned to receive standard quinine regimen with or without pentoxifylline (10 mg/kg/day by continuous intravenous infusion). Pentoxifylline exerted an inhibitory effect on the synthesis of tumor necrosis factor (TNF), a possible mediator of CM. The 26 children who received pentoxifylline had significantly shorter comas than controls (median, 6 vs. 46 h; P < .001) Pentoxifylline recipients showed a trend toward a lower mortality, with a borderline significant difference (P = .055). The better outcome in the pentoxifylline group was associated with a decline in TNF serum levels on the third day of treatment in a few subjects that was not seen in controls. While alternative or concurrent mechanisms of action may be of some relevance, larger double-blind trials are needed to determine whether pentoxifylline has a therapeutic role in CM.