We studied a community cohort of 193 individuals exposed to endemic Schistosoma japonicum infection in the Dongting Lake region of China to assess subclinical morbidity and the 2-year benefit of curative therapy (praziquantel) administered in 1996. Prevalence and intensity of S. japonicum infection before treatment were 28% and 192 eggs per gram faeces (epg), respectively. Two years after cure, 22% of the cohort were re-infected, but with a lighter intensity (67 epg). Sixty-four subjects (37%) showed significant improvement in ultrasound parenchyma images after treatment and 51 subjects (54%) showed significant improvement of periportal fibrosis. Left-lobe enlargement also reversed (P < 0.05) and splenomegaly reversed in 6 of 8 cases and developed in only 1. Two years post-treatment a dilated portal vein became less frequent, but the decline was not significant (16% vs 11%, P < 0.05). The serum levels of laminin and collagen IV associated with re-infection and intensity and hyaluronic acid levels correlated with ultrasound findings (P < 0.01). Overall, treatment induced a marked decrease in subclinical hepatosplenic morbidity attributable to S. japonicum although low-intensity re-infection after treatment remained relatively frequent. Stratified analysis and logistic models evaluated potential confounding factors for assessment of treatment effects on hepatic fibrosis. S. japonicum infection and moderate-heavy alcohol intake interacted: improvement in parenchymal morbidity was impeded among drinkers (P < 0.05). Chemotherapy focused on at-risk residents controls prevalent subclinical hepatic fibrosis but re-infection indicates the need for complementary control strategies.