The currently recommended protocol for treatment of kala-azar (KA) necessitates repeated bone marrow/splenic aspiration to monitor the response and duration of therapy as well as to detect resistance and change to alternative drugs. These procedures being invasive, there is a pressing need for less invasive diagnostic tools for this purpose. We have evaluated the role of C-reactive protein (CRP) in 201 children with visceral leishmaniasis at different stages of the disease to work out the relationship, if any, between CRP levels and disease activity, including response to therapy. The subjects belonged to the 2-12 year age group in whom CRP estimation was done on admission, every 5th day during treatment, and repeated on follow-up at 2 and 6 months. The levels were compared with those of 50 randomly chosen age-matched healthy children who served as controls. The mean serum CRP value in the study group before the commencement of treatment was 62.96 +/- 1.03 mg/l, which was significantly higher (p < 0.001) in comparison to the control group. Commencement of treatment resulted in a simultaneous decline in serum CRP. Parasitic clearance from the spleen was faster in patients with an initial low serum CRP level (< 60 mg/l) in comparison to patients with high levels (> 60 mg/l). During treatment, mean serum CRP levels were significantly higher in late responders than in early responders (p < 0.001). Correlation of CRP levels to indicate the presence or absence of parasites suggested a cut-off level of 12 mg/l by day 10, with a sensitivity of 82.5 per cent, specificity of 78.5 per cent, positive predictive value of 92 per cent, and negative predictive value of 60.2 per cent. Our observations suggest a promising role for CRP estimation every 5-10 days during therapy in visceral leishmaniasis for monitoring the response to therapy and to detect possible resistance.