Refining the risk estimate for transfusion-transmission of occult hepatitis B virus.

BACKGROUND AND OBJECTIVES: We previously published a model to estimate the residual risk (RR) for occult hepatitis B infection (OBI) in the absence of universal anti-HBc testing. To incorporate new information on the epidemiology of OBI, we describe model refinements and estimate a more accurate HBV RR due to OBI in Australia.

MATERIALS AND METHODS: In our original model, the OBI risk, p(OBI), was defined by the rate of 'non-detection' by the HBV DNA screening test in use, p(NAT non-detection), and the average infectivity of blood components from OBI donors, p(transmission). We revised the model by integrating three refinements: that donations with anti-HBs levels of >10 IU/l, or donations solely for manufactured plasma products, be excluded from the risk calculation, and an updated estimate of p(transmission).

RESULTS: Refining our OBI RR model resulted in a more than 10-fold reduction in the reported RR risk to recipients from OBI in our donor population. Based on the use of a common data set, the mean OBI RR risk decreased from 1 in 374 354 donations (95% CI: 1 in 191 940-1 072 681) to 1 in 3 984 033 (95% CI: 1 in 1 146 188-65 268 257) for the refined model.

CONCLUSION: Our model refinements provide a more realistic measure of the HBV RR in the donor population. Unlike the previous model, the new model demonstrates that the risk of HBV due to OBI in the Australian blood donor population is negligible, and further potentially cost-ineffective risk management strategies are not currently warranted.