Chloramphenicol pharmacokinetics were studied in 29 Nepalese adults diagnosed with uncomplicated enteric fever and randomized to receive succinate ester 30 mg/kg i.v. or i.m. Serial plasma concentrations of chloramphenicol, and iothalamate (to estimate glomerular filtration rate), antipyrine (hepatocellular function) and Indocyanine Green (liver blood flow) were measured by HPLC and kinetic parameters estimated by non-compartmental analysis. In culture-positive patients (n = 16), mean residence times (MRTs) and steady-state volumes of distribution (V(d)ss) for i.v. chloramphenicol (mean +/- S.D.; 4.9 +/- 0.9 h and 1.9 +/- 0.8 L/kg; n = 7) were less than after i.m. chloramphenicol (12.3 +/- 7.3 h and 3.7 +/- 2.5 L/kg; n = 9; P < 0.05), with a higher peak plasma concentration after i.v. (16.2 +/- 9.1 versus 7.8 +/- 3.6 mg/L; P < 0.05); plasma clearance (Cl(p)) was similar in the two groups (368 +/- 172 and 310 +/- 224 mL/kg/min after i.v. and i.m. respectively). In 17 patients examined during convalescence, MRT and Vdss were less than in acute illness regardless of route chloramphenicol administration. There were similar changes in chloramphenicol kinetic parameters in culture-negative patients. Antipyrine Cl(p) and liver blood flow correlated weakly with chloramphenicol Cl(p) in culture-positive patients (P < 0.1) and were higher in convalescence; no such associations were seen for iothalamate Cl(p). These data indicate that i.v. chloramphenicol produces peak plasma concentrations which are on average twice those after i.m. injection of the same dose, due principally to a smaller V(d)ss. Cl(p) is uninfluenced by route of administration and is determined more by hepatic metabolism than renal excretion. Intramuscular treatment may result in sub-therapeutic chloramphenicol concentrations initially, but continued regular i.v. dosing is more likely to produce levels at which bone marrow toxicity occurs.