BACKGROUND: To examine the feasibility and to perform a cost-benefit analysis of a pooling protocol of enzyme immunoassay (EIA) screening for antibody to hepatitis C virus (anti-HCV) under real conditions, a prospective study was carried out using sera from 1875 local blood donors. STUDY DESIGN AND METHODS: In the absence of knowledge of the anti-HCV reactions, the donor's sera were pooled into groups of five consecutive samples for testing by EIA. The dilution and final volume of the serum pool were adjusted to equal those recommended for single-serum EIA by the manufacturer of the test kit. The results obtained were compared with those of single-serum EIA to assess the feasibility of the pooling protocol. By applying probability theory, the percentage of reduction in the number of tests performed (L value) when the serum-pooling strategy was used was calculated for several anti-HCV seroprevalences and for varied sizes of pool. The calculations were performed on a computer using a program compiled by the authors. RESULTS: The results showed that seroprevalence was 2.24 percent (95% CI, 1.57-2.91%); the rate of false negativity was 0 (95% CI, 0-8.4%), the sensitivity of the pooling protocol was 100 percent (95% CI, 91.6-100.0%), the rate of false positivity was 0.8 percent (95% CI, 0-1.8%), and the specificity of the pooling protocol was 99.2 percent (95% CI, 98.2-100.0%). Cost-benefit analysis showed that the pooling protocol could save 69.3 percent of the cost. A table of L values can be used conveniently by serologists to determine the optimum pool size if estimates of seroprevalence are available. CONCLUSION: The pool EIA did not perform worse than individual EIAs, and the pooling strategy was markedly less expansive. The pooling protocol was recommended for screening of anti-HCV-positive subjects from large populations with low seroprevalence.
BACKGROUND: To examine the feasibility and to perform a cost-benefit analysis of a pooling protocol of enzyme immunoassay (EIA) screening for antibody to hepatitis C virus (anti-HCV) under real conditions, a prospective study was carried out using sera from 1875 local blood donors. STUDY DESIGN AND METHODS: In the absence of knowledge of the anti-HCV reactions, the donor's sera were pooled into groups of five consecutive samples for testing by EIA. The dilution and final volume of the serum pool were adjusted to equal those recommended for single-serum EIA by the manufacturer of the test kit. The results obtained were compared with those of single-serum EIA to assess the feasibility of the pooling protocol. By applying probability theory, the percentage of reduction in the number of tests performed (L value) when the serum-pooling strategy was used was calculated for several anti-HCV seroprevalences and for varied sizes of pool. The calculations were performed on a computer using a program compiled by the authors. RESULTS: The results showed that seroprevalence was 2.24 percent (95% CI, 1.57-2.91%); the rate of false negativity was 0 (95% CI, 0-8.4%), the sensitivity of the pooling protocol was 100 percent (95% CI, 91.6-100.0%), the rate of false positivity was 0.8 percent (95% CI, 0-1.8%), and the specificity of the pooling protocol was 99.2 percent (95% CI, 98.2-100.0%). Cost-benefit analysis showed that the pooling protocol could save 69.3 percent of the cost. A table of L values can be used conveniently by serologists to determine the optimum pool size if estimates of seroprevalence are available. CONCLUSION: The pool EIA did not perform worse than individual EIAs, and the pooling strategy was markedly less expansive. The pooling protocol was recommended for screening of anti-HCV-positive subjects from large populations with low seroprevalence.
Authors: D Thorburn; D Dundas; E A McCruden; S O Cameron; D J Goldberg; I S Symington; A Kirk; P R Mills Journal: Gut Date: 2001-01 Impact factor: 23.059
Authors: Baligh R Yehia; Ramin S Herati; John A Fleishman; Joel E Gallant; Allison L Agwu; Stephen A Berry; P Todd Korthuis; Richard D Moore; Joshua P Metlay; Kelly A Gebo Journal: PLoS One Date: 2014-07-17 Impact factor: 3.240