Objective: To determine the clinical course of hepatitis C virus in the first decade of infection in a group of patients who acquired their infections on a known date.
Design: Cohort study.
Setting: Clinical centres throughout the United Kingdom.
Participants: 924 transfusion recipients infected with the hepatitis C virus (HCV) traced during the HCV lookback programme and 475 transfusion recipients who tested negative for antibodies to HCV (controls).
Main outcome measures: Clinical evidence of liver disease and survival after 10 years of infection.
Results: All cause mortality was not significantly different between patients and controls (Cox's hazards ratio 1.41, 95% confidence interval 0.95 to 2.08). Patients were more likely to be certified with a death related to liver disease than were controls (12.84, 1.73 to 95.44), but although the risk of death directly from liver disease was higher in patients than controls this difference was not significant (5.78, 0.72 to 46.70). Forty per cent of the patients who died directly from liver disease were known to have consumed excess alcohol. Clinical follow up of 826 patients showed that liver function was abnormal in 307 (37.2%), and 115 (13.9%) reported physical signs or symptoms of liver disease. Factors associated with developing liver disease were testing positive for HCV ribonucleic acid (odds ratio 6.44, 2.67 to 15.48), having acquired infection when older (at age > or = 40 years; 1.80, 1.14 to 2.85), and years since transfusion (odds ratio 1.096 per year, 1.00 to 1.20). For patients with severe disease, sex was also significant (odds ratio for women 0.38, 0.17 to 0.88). Of the 362 patients who had undergone liver biopsy, 328 (91%) had abnormal histological results and 35 (10%) of these were cirrhotic.
Conclusions: Hepatitis C virus infection did not have a great impact on all cause mortality in the first decade of infection. Infected patients were at increased risk of dying directly from liver disease, particularly if they consumed excess alcohol, but this difference was not statistically significant.