Mathematical modelling of potential risks of any changes to the eligibility criteria for men who have sex with men (MSM) to give blood and plasma is needed, both in the current blood safety context and if alternative risk reduction strategies (e.g. pathogen inactivation (PI)) are applied in the future. Mathematical modelling can be used to estimate the level of risk that can be expected under different conditions. Blood operators in Canada are considering implementing PI as an additional safeguard against transfusion-transmitted infections. Applied after blood collection, PI can inactivate a broad spectrum of pathogens, including HIV, hepatitis B virus (HVB) and hepatitis C virus (HVC). PI can also reduce risk from emerging transfusion-transmitted pathogens before they are recognized as a threat to the blood supply. Implementing PI could change the estimates of risk that behavior-based blood donor deferrals are based on. This study uses mathematical risk modeling to evaluate the impact of changing donor deferral criteria in the event that PI is implemented and applied to all blood components. The models calculate “residual risk”, which is defined as the amount of risk remaining after efforts to identify and control risks have been made. The project is aligned with the international MSM modelling initiative led by Dr. O’Brien, working towards a common goal of using mathematical models to understand risk with changes to policy.
What was done?
The initial risk simulation estimates residual risk of having an HIV infectious donation contaminating a plasma pool destined for fractionation following a change in time-based deferral periods for MSM (from a 3-month MSM deferral to no MSM deferral). In this instance, plasma destined for fractionation would be donated by apheresis (source plasma donation) and would be treated with a pathogen inactivation process that is already being used by fractionators.
As part of the international MSM modelling team (the Surveillance, Risk Assessment and Policy (SRAP) subgroup of the International Society of Blood Transfusion (ISBT)), this project has been aligned with the common MSM modelling initiative. The initial focus was on the risk simulation to estimate HIV risk with plasma apheresis for fractionation from MSM donors with different time deferrals and the use of PIs. The probability of infection (based on type of donation, proportion of new donors, interdonation interval, HIV prevalence and HIV incidence among donors, viral load of infected donation, etc.) and impact of safety measures (such as viral load reduction, process failure using PIs and nucleic acid testing (NAT – the test used to detect HIV in donated blood), etc.) was simulated based on Canadian Blood Services and Héma-Québec data. A total of 300,000 pools were simulated to estimate the Canadian residual risk of HIV-infected source plasma. The models were based on most likely, optimistic, and pessimistic scenarios.
What was found?
A two stage Monte Carlo simulation using Bayesian networks was used to simulate 300,000 pools (around 2 billion donations) for each scenario. For the most likely scenario of the three-month deferral model, there were 2.79 HIV positive donations per 1,000,000 donations with an average number of HIV copies per HIV positive pool of 9.7606 × 10-8. For the most likely scenario with no time-based deferral, there were 3.01 HIV positive donations per 1,000,000 donations with an average number of HIV copies per HIV positive pool of 9.7856 × 10-8. For both models, if the infectious threshold is one copy of the virus per pool, none of the simulated pools were considered as potentially infectious after the use of PI. Data so far indicates that the HIV residual risk in source plasma after the use of PI is extremely low (0.00; 95% CI [0.00; 1.23 × 10-5]). Based on the simulation results, there would be no significant HIV-risk associated with the removal of time-based deferral without quarantine for source plasma from MSM donors.
Opportunities for change
The next step is to model risk for PI-treated blood derived products for transfusion and include HVC and HVB in the simulation. Our models will give residual risk estimates that will indicate whether the safety of blood could be assured through the use of PIs to eliminate pathogens in blood donations. This approach could offer an alternative to the population-based eligibility criteria currently applied to MSM donors.
The most important limitation of this project is the availability and specificity of data used for to build the model.
Research publications and resources
Aubé et al. 2021: HIV residual risk in Canada for apheresis source plasma donation without deferral for men who have sex with men. Vox Sang 2021.