Despite the advent of screening of blood products for anti-hepatitis C virus (HCV), the incidence of HCV infection among haemodialysis (HD) patients is alarmingly high and suggest transmission within the HD unit. To analyse trends in the prevalence and incidence of HCV infection, and evaluate the impact of dialysis room and reuse policies on the incidence of HCV infection, a hospital survey instrument was sent out to medical directors of all 71 HD units in Portugal in August 1994. Information for the years 1991, 1992 and 1993 was requested with respect to HCV infection, defined as positive anti-HCV test. Sixty-two of 71 units (87%) treating 4232 patients in 1993 responded. Overall, data from 5774 patient-years were available for analyses. Observations over multiple intervals were pooled into a single sample, and pooled logistic regression was used to evaluate the relationship between risk factors/strategies and incidence of HCV infection. By 1993, regular anti-HCV testing of patients and staff was practised by 98% and 82% of units, respectively. There was a significant decline in the incidence of HCV infection from 9.9% in 1991 to 5.7% in 1992 and 5.1% in 1993. The incidence was directly related to the prevalence in the dialysis unit. Units with a prevalence of less than 19% had an annual incidence of 2.5% compared to a 35.3% incidence in units with a prevalence greater than 60%. There was wide variation in the incidence of HCV infection in HD units across the country, with geographical location, unit ownership and socioeconomic factors playing a significant role. The incidence was lowest among units that: (i) were located in the northern regions of the country; (ii) were private hospital-based units; and (iii) used dedicated machines or separate rooms for anti-HCV-positive patients. The incidence among units that reprocessed dialysers (6.1%) was not significantly different from that among units that did not reprocess dialysers (7.4%). However, among units that did reprocess dialysers, the incidence of HCV infection was lowest in: (i) units that used separate rooms for reprocessing dialysers from anti-HCV-positive patients or did not reprocess these dialysers; and (ii) units that used Renalin as the sterilant. These results suggest the transmission of HCV infection in HD units and that use of dedicated machines and isolation of anti-HCV-positive patients and their dialysers may reduce the incidence of HCV infection.