International Journal of Epidemiology 2000;29:922-927
© International Epidemiological Association 2000
Hepatitis C virus (HCV) infection and liver-related mortality: a population-based cohort study in southern Italy
Laboratories of Epidemiology and Biostatistics,
a Experimental Pathophysiology
b and Biochemistry;
d Department of Medicine,
c Medical Research Institute IRCCS ‘Saverio De Bellis’ Castellana Grotte, Italy.
Reprint requests to: Alberto R Osella, Laboratory of Epidemiology and Biostatistics, Via F Valente 4, 70013 Castellana G (BA), Italy. E-mail: osellar{at}libero.it
Background Hepatitis C virus (HCV) is a common cause of chronic liver diseases but the degree to which these diseases contribute to liver-related mortality is not well established. The aim of this study was to estimate the absolute and relative effects of HCV infection on liver-related mortality.
Methods A population random sample of 2472 subjects aged 
30 years was enrolled and followed up from 1985 to 1996. At enrolment, a structured interview and a clinical evaluation were performed. Serum samples were tested using HCV ELISA and RIBA HCV. Outcomes were overall and liver-related mortality and tracing procedures included review of office and hospital records, death certificates, and interviews with general practitioners, attending hospital and next of kin. Statistical analysis was performed using Poisson and binomial prospective data regression.
Results Crude overall and liver-related mortality rates were 7.66 (95% CI : 6.68–8.79) and 0.9 (95% CI : 0.3–2.2) per 103 person-years, respectively. For HCV infection effect, incidence rate ratio and difference (per 103 person-year), risk ratio and difference were 27.5 (95% CI : 6.5–115.6), 4 (95% CI : 3–7), 33.1 (95% CI : 7.8– 139.3) and 0.06 (95% CI : 0.04–0.08), respectively; all measures were adjusted for age at death, sex and daily alcohol intake.
Conclusions The results show a strong relative but weak absolute effect of HCV infection on liver-related mortality in the 10-year period considered. Poisson and binomial models are virtually equivalent, but the choice of the summarizing measure of effect may have a different impact on health policy.
Keywords Binomial models, cohort study, HCV infection, liver-related mortality, Poisson models
Accepted 17 March 2000
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. P. Vandenbroucke, E. v. Elm, D. G. Altman, P. C. Gotzsche, C. D. Mulrow, S. J. Pocock, C. Poole, J. J. Schlesselman, M. Egger, and for the STROBE initiative Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration Ann Intern Med, October 16, 2007; 147(8): W-163 - W-194. [Abstract] [Full Text] [PDF]
|
|