The contribution of statistics in the forensic appraisal of post-transfusion contamination by hepatitis C virus

Many patients were contaminated by blood products in the 1980s and early 1990s, a period during which there was no means of detecting hepatitis C virus (HCV) and measures of vigilance in French blood transfusion centers were lacking in rigor, and it has become a real public health issue by now. The blood transfusion centers which produce and distribute the products are liable for such contamination, as long as the expert appraisal has established a link of causality between the transfusion and the contamination. When the viral serologic status (i.e. the contaminating potential) of the donors is not known, experts often resort to indirect evidence, including use of probability estimates calculated according to the following method. The probability (P) that there was at least one contaminating donor among those having provided the blood given to a patient, is calculated using the following formula:P=1.(1.pHCV)(N)where pHCV is the prevalence of subjects who have tested positive for anti-C virus antibodies in a population of blood donors, and N is the number of donors having contributed to the transfusion in question. Although the polymerase chain reaction (PCR) is very sensitive, its threshold, below which weak replication is undetectable, varies depending on the technique used. One study has demonstrated that below a dilution of 10(-4), RNA is no longer detected by the majority of laboratories. The specificity of PCR is not flawless either, since false-positive results are found, probably resulting from contamination of the samples. Considering that 0.10% is a reasonable approximation of the prevalence of subjects who had anti-HCV antibodies that were detectable by recombinant immunoblot assay during that period, one can calculate the probability that there is a contaminating donor among the blood donors who contributed to the product that was administered and determine a curve.