FAQ: Residual risk estimates for viral transfusion-transmissible infections

We developed the following frequently asked questions and answers to provide health professionals with a better understanding of how the Blood Service residual risk estimates for viral transfusion-transmissible infections (TTI) are calculated, including the rationale for using these calculations.
 

We publish estimates of the residual risks of transfusion-transmissible infections (TTI) as a service to clinicians to guide transfusion decision-making and informed consent processes.

We publish the TTI residual risk estimates in the Blood Component Information (BCI) booklet, and on this web site. The viral risk estimates are reviewed and updated periodically. Our estimates are based on published methods.

The 'window period' (WP) is defined as the interval between infection and first positive test marker in the bloodstream.

WP-based models assess the rate of incident infection (i.e. positive donors who have previously tested negative at the Blood Service for the same viral marker) in the repeat donor (RD) population as a measure of viral incidence (i.e. the rate of newly acquired infection).
 

The average inter-donation interval for all incident donors between the positive result and previous negative result is also incorporated. The longer this interval for an individual donor, the lower the probability that the donor was in the WP at the time of donation. In other words, the inter-donation interval is inversely proportional to the risk.

For infections subject to NAT testing (HIV, HBV and HCV), the Weusten model [1] uses incidence to estimate the risk of infection in a recipient of a tested blood component based on the lower limit of detection of the applied NAT test and the probability of transmission based on a number of factors, including the volume of transfused plasma and the presumed ‘infectious dose’ of the infectious agent.

The WP model used for HTLV [2] estimates only the probability of failing to detect a WP donation in a given time period based on either the rate of incident infection and inter-donation interval, or the prevalence in first-time donors. 
 

The final model, applied only to HBV, estimates the risk specifically for occult HBV infection (OBI). The method is based on assessing the probability of 'non-detection' by HBV NAT and the average probability of HBV transmission from NAT non-reactive donations. NAT non detection is derived by examining HBV NAT data and assessing the frequency of prior NAT non-detectable donations from donors identified as OBI by NAT. The transmission function is based on investigation of the outcome of transfusions from blood components (termed lookback) sourced from donors with OBI. The full method is available in reference [3].

The WP-based models assume that the risk of collecting blood from an infectious donor predominantly relates to them being in the WP (i.e. incident infection) and the best estimate of incidence in the donor population is the rate of incident donors in the repeat donor population. 

The Weusten model assumes a concentration-dependent probability that the virus is not detected in the log-linear ramp-up phase of plasma viraemia in acute infection, and a dose-dependent probability that an infection develops in the recipient of the contaminated blood product. Both these probabilities contribute to the overall residual risk.
 

While the assumption that WP donors account for the majority of risk seems to hold true for HIV, HCV and HTLV, HBV is problematic because of 'chronic' infection (i.e. HBsAg negative/anti-HBc positive with low levels of HBV DNA). WP-based models do not satisfactorily address the risk of long-term OBI. Therefore the Blood Service developed a specific model to estimate the OBI risk which is summed with the WP risk to derive the overall HBV residual risk estimate. Importantly, HBV NAT will incrementally identify OBI donors since the vast majority can be detected using the highly sensitive ID NAT employed by the Blood Service.

When considering the significance of specific risks, it is often useful to compare them to the risks associated with everyday living.

The risks of transfusion transmitted infections are very small compared to risks of everyday living. (see Calman Chart).

The Calman Chart for Explaining Risk (UK risk per 1 year)
Classification Risk range Example
Negligible <1:1,000,000 Death from a lightning strike
Minimal 1:100,000–1:1,000,000 Death from a train accident
Very low 1:10,000–1:100,000 Death from an accident at work
Low 1:1,000–1:10,000 Death from a road accident
Moderate 1:100–1:1,000 Death from smoking 10 cigarettes per day
High >1:100 Transmission of chickenpox to susceptible household contacts

Source: Calman K. Cancer: science and society and the communication of risk. BMJ 1996;313:801.

The chance of dying in a road accident, for example, is about 1 in 10,000 per year which is considered a ‘low’ risk. Comparatively, all the viral risk estimates are well below this level, being considered as ‘negligible’.

 
Reference
  1. Weusten J, Vermeulen M, van Drimmelen H, et al. Refinement of a viral transmission risk model for blood donations in seroconversion window phase screened by nucleic acid testing in different pool sizes and repeat test algorithms. Transfusion 2011;51:203-215.
  2. Seed CR, Kiely P, Keller AJ. Residual risk of transfusion transmitted human immunodeficiency virus, hepatitis B virus, hepatitis C Virus and human T lymphotrophic virus. Intern Med J 2005;35(10):592–598.
  3. Seed CR, Kiely P, Hoad VC, Keller AJ: Refining the risk estimate for transfusion-transmission of occult hepatitis B virus. Vox Sang 2017;112(1):3-8.