A prospective study was conducted with Bangladeshi children with rotavirus (RV) diarrhea to assess whether nutritional and clinical parameters, RV serotypes, levels of interleukin-10 (IL-10), tumor necrosis factor alpha (TNF-alpha), and gamma interferon (IFN-gamma), and RV-specific antibody titers in plasma and stool were associated with the development of persistent diarrhea. Children with watery diarrhea for 6 to 8 days, selected from the Dhaka Hospital of the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), were enrolled in the study and monitored until diarrhea improved. Children were classified as having acute diarrhea (AD) if diarrhea resolved within 14 days of onset and as having persistent diarrhea (PD) if diarrhea persisted for more than 14 days after onset. Uninfected, control children (n = 13) from the Nutrition Follow-Up Unit of ICDDR,B were also enrolled. Of the 149 children with diarrhea enrolled, 29 had diarrhea with RV alone, of which 19 had AD and 10 developed PD. Samples of stool and blood were collected from all children on enrollment. Stool samples were collected again from children when they developed PD. Of the 10 children who had an initial RV infection and then developed PD, only one had persistent RV infection. Plasma levels of IL-10 and TNF-alpha were higher in children with diarrhea compared to uninfected controls but were similar in children with AD and PD. Plasma IFN-gamma levels were higher in children who developed PD than in those with AD (P = 0.008) or uninfected controls (P = 0.001). In stools, the levels of TNF-alpha, the only cytokine detected, were similar in the three groups of children. RV-specific immunoglobulin G (IgG) titers in plasma were higher in uninfected children than in those with AD (P < 0.001) or PD (P = 0.024) but titers were similar in children with AD and PD. RV-specific IgA titers in plasma and stool were similar in the three groups of children. From all observed parameters, only elevated plasma IFN-gamma levels were associated with subsequent development of PD. However, a larger sample size is necessary to substantiate this observation.