Progress in monitoring blood safety

Abstract

In this issue of TRANSFUSION, Zou and colleagues1 of the American Red Cross (ARC) report the marker prevalence and derive the incidence of human immunodeficiency virus (HIV) and hepatitis C virus (HCV) for donations of approximately 66 million units of allogeneic whole blood collected between January 1999 and December 2008, encompassing the 10-year period since introduction of nucleic acid tests (NATs). The authors estimated the residual risk of these infections in donated blood as the product of incidence multiplied by the NAT window period for HIV and HCV. They provide an analysis of the temporal trends in risk and the underlying donor demographics. This article complements a previous publication by investigators at ARC and Abbott Laboratories in which the incidence and residual risk of hepatitis B virus (HBV) infection in whole blood donors was estimated by two independent methods for November 2006 through July 2008 and compared with a similarly derived result for November 1997 through July 1999.2 In a separate article in this issue, Zou and colleagues3 report on the marker rates, incidence, and residual risk of HIV, HCV, HBV, and human T-lymphotropic (HTLV) for apheresis donations of blood components from January 2004 through December 2008. They provide comparisons among these variables by apheresis component compared with whole blood collections. This report updates the findings of a previous investigation that was conducted under the first Retrovirus Epidemiology Donor Study (REDS-I) at five U.S. blood collection centers between 1991 and 1994.4 These studies represent the most current estimates of incidence and residual risk for the major blood-transmissible viruses affecting the U.S. blood supply. Moreover, the analyses of temporal trends and demographics related to marker-positive donations provide insights into the sources of risk in the blood system and how they are changing over time. A few limitations warrant mention, however. First, it should be noted that the incidence estimates depend strongly on the number of NAT yield (RNA-positive, but antibody-negative) cases, which could have been affected by changes over time in the sensitivity of the NAT and serologic assays, as well as by current estimates for the marker-negative infectious window periods. Second, inferences of trends based on small numerators (number of incident cases) with very large denominators (total number of collections) can be misleading as to level of both statistical significance and public health importance. Third, it is possible that the data obtained solely within the ARC system might not be nationally representative. Nevertheless, these data provide us with a unique and very rich source of information that directs our attention appropriately to a number of issues and concerns. The role of NAT in enabling accurate assessments of donor incidence (number per 100,000 person-years) of HIV and HCV infection is central to these analyses. NAT-yield cases represent recently acquired and thus incident cases of infection, although estimation of the incidence rate in first-time donors still requires an extrapolation from the directly observed rate in repeat donors. In an examination of transfusion risk in the NAT era, it is perhaps also worth remembering that the introduction of NAT as a donor screening method for HIV resolved an otherwise vexing concern about disparate blood safety due to differences in sensitivity (and hence window period risk) of the available licensed screening tests for antibodies.5 The authors' results are summarized in Table 1 as rounded values of the reported point estimates for incidence and residual risk of HIV, HCV, and HBV. The infectious window periods for HIV and HCV were based on donor testing by minipool NAT, while that for HBV was based on estimated time to detection of hepatitis B surface antigen (HBsAg). The authors make several important observations based on more detailed analyses of the multiyear data. These include the following: HCV prevalence appears to have decreased steadily over 10 years in first-time donors. While significant as evidence of a favorable change in the donor base (new donors have less cumulative HCV risk than in the past), this finding may reflect a cohort effect (higher risk in a historic group of donors) or an overall change in the epidemiology of HCV in the United States rather than any temporal improvement in donor selection. Cases of acute hepatitis reported to the Centers for Disease Control and Prevention (CDC) under the CDC's National Notifiable Disease Surveillance System indicate that annual incidence of HCV has declined markedly from 5.2 cases per 100,000 population in 1995 to 0.5 per 100,000 population in 2007 among persons aged 25 to 39 years, the age group with the current and historically highest incidence.6 This general decline may result from a decreased number of HCV incidence in injection drug users. The authors estimated that residual risk of HBV in blood donation has decreased approximately threefold comparing the periods of 2006 to 2008 to 1997 to 1999.2 Regardless of cause (such as a general increase in vaccination against HBV), this represents good progress in blood safety. However, the risk for transmission of HBV is approximately three to five times higher than for HIV or HCV. As discussed recently at a meeting of the FDA's Blood Products Advisory Committee, widespread implementation of licensed donor screening tests for HBV DNA by NAT, and the further development of more sensitive NAT assays should further reduce this risk.7 In contrast to HBV, and of concern, is a progressive increase in the incidence of HIV and HCV over a 10-year period. An apparent decrease in the incidence of HIV and HCV in 2005 to 2006 compared with 2003 to 2004 is an unexplained exception to the general trend. Comparing 2007-2008 to 2005-2006, increased incidence of HCV in repeat donors was due to an increased rate of seroconversions rather than any observed increased number of NAT-yield cases, perhaps suggesting a bias against donations by repeat donors who have had more recent high-risk exposures. If a bias against donation soon after risk exposure actually exists, then the true HCV risk might be lower than has been estimated. In the case of HIV, the increased incidence was attributed primarily to young (age 16-19 years) male African American donors. More generally, HIV incidence and prevalence are higher in African American males and females compared to Caucasians of comparable age.8 Donations by African Americans are critical to the availability of red blood cells (RBCs) with antigenic characteristics compatible with those of African American recipients, especially for chronically transfused patients with sickle cell disease.9 The authors noted a consistently higher incidence of HIV among donors in the South compared with other geographic areas. In the case of HCV, the increased incidence in repeat donors was attributed primarily to Caucasian males and females over age 50 years. The authors note that HCV transmission from nonhospital procedures, and specifically endoscopies, could be a risk factor in this age group based on recognized outbreaks in the United States. However, other possible medical and nonmedical parenteral exposures in this age group cannot be ruled out. The authors did not comment on an apparent sharp increase in HCV incidence for the Western ARC regions comparing 2007-2008 to all previous years that can be seen in Fig. 5B of the article.1 The increased incidence of HIV and HCV in donors should prompt renewed efforts to investigate incident cases of HIV, HCV, and HBV to determine the donors' underlying risk factors. This is especially important for NAT-yield cases because such donors might well have transmitted their infections to recipients had they presented to donate even just a few days earlier. Detection of HIV or HCV in such cases can be thought of as “near-miss” events in questionnaire screening. Risk factor investigations in these cases might point to areas where donor questions could be improved. As noted above, ARC investigators have also reported on the relative risks of apheresis compared with whole blood donations.3 The importance of this study lies in the fact that use of apheresis collections including at mobile sites has increased rapidly in recent years, especially for RBCs, with likely changes in donor demographics compared with collections of whole blood. The most recent data from the 2007 National Blood Collection and Utilization Survey indicated that apheresis RBCs now constitute 10% of all transfused allogeneic whole blood or RBCs and 88% of all transfused platelet (PLT) equivalent units (one apheresis PLT collection was considered equivalent to six whole blood–derived PLT units).10 Although marker rates for HIV, HCV, HBV (as indicated by HBsAg screening), and HTLV in apheresis donors as a whole were much lower than for whole blood donors, the authors noted that this observation largely reflects the much higher frequency of apheresis donations (other than RBCs) compared with whole blood, preventing any direct implication for relative incidence. However, of more basic significance was the finding that first-time male donors of double RBCs, who provided only 1% to 5% of apheresis collections, accounted for 25% to 100% of the various marker test–positive units. In particular, first-time male donors aged 30+ years who donated double RBCs represented only 3% of donors, but accounted for more than 50% of all HCV positives. A severalfold increase in such male donations between 2004 and 2008 has resulted in an increase in marker prevalence rates for apheresis collections overall and especially for HCV. The marker rates (and interdonation intervals) in this subgroup were comparable to those of whole blood donors, but higher than those for donors of non–double-RBC units (mainly PLT donors). Once again, the finding of a higher risk cohort of donors should underline the importance of further studies to examine the risk factors in this group of donors as a basis to develop improved donor selection procedures. This is especially important in light of the movement toward predominant use of male donors to make large-volume plasma products (including apheresis PLTs and plasma) as a preventive measure against transfusion-related acute lung injury.11 Table 2 provides a summary of the authors' study on incidence of major transfusion-transmissible viral infections in apheresis donors compared with whole blood donors.3 The incidence of HTLV was not estimated due to the absence of any seroconverting donor in the period of study, consistent with the very low rate of HTLV positivity that was found in apheresis donors. Unlike the findings summarized in Table 1, which are based on data including first-time and repeat donors, those summarized in Table 2 pertain only to repeat donors. An overall comparison of residual risks for apheresis versus whole blood collections would require inclusion of data on infection incidence in first-time donations. The authors' main conclusion is that, within the limits of statistical significance, all the donor groups have similar incidence for HIV, HCV, and HBV resulting in comparable per unit risks. Biovigilance endeavors are expected to shape blood safety monitoring in the future. A robust discussion of biovigilance is beyond the scope of this editorial. However, the following brief discussion outlines the direction that national hemovigilance is taking in the United States. Hemovigilance, a subset of biovigilance, is the state or attitude of watchfulness (monitoring) of the entire blood system from donor recruitment, donor health, and infectious disease markers to transfusion practices and recipient adverse events through the lens of a quality system. In some countries such as those in the European Union, the European Directive requires quality systems as well as hemovigilance “from vein to vein” and also requires its member states to develop periodic reports of the progress.12,13 In 2009, the Department of Health and Human Services (HHS) and its public health agencies released a white paper, “Biovigilance in the United States: Efforts to Bridge a Critical Gap in Patient Safety and Donor Health,” that was presented to the Secretary's Advisory Committee for Blood Safety and Availability.14 The gaps that were identified relevant to blood safety include: Gap 1: Patchwork and sometimes fragmented system of various adverse event reporting; Gap 2: Likely underreporting of transfusion adverse events; Gap 3: Challenges with FDA-required reporting; Gap 4: Need for accurate recipient denominator data, precise definitions, and training; Gap 5: No national surveillance of donor serious adverse events other than fatalities; Gap 6: Need for accurate donor denominator data, precise definitions, and training; Gap 7: Need for accurate tracking of all donor infectious disease test data; Gap 8: Need for timely analysis of reported data. Pertinent to the articles by Zou and colleagues, Gap 7, the need for accurate tracking of all donor infectious disease test data, is extremely important to identify trends and identify risk factors in donor groups. While NAT of small pools of donor specimens (“minipools”) has narrowed the window period to an estimated 9.0 ± 0.6 days for HIV, breakthrough cases still can occur.15-17 To address this risk, proactive national donor surveillance is needed to understand the circumstances, behaviors, or mode of transmission underlying transfusion-transmitted disease (TTD) infections in donors, especially those with recent infections. A standardized post-TTD conversion interview presently under development by several major blood collecting organizations could provide information useful to improving the donor selection process. The authors make a compelling case for proactive monitoring of TTD markers by large blood establishments. It is important to note that surveillance of TTD markers is critical not only for large establishments but for all establishments that collect and process blood. This point is further emphasized by the review of the 2008 FDA Biological Product Deviation (BPD) reports, which also encouraged establishments to evaluate TTD seroconversion rates among their donor populations.181, 2 illustrate BPD reports from 2007 and 2008 for which a unit of blood or plasma was distributed from a marker-negative donor who subsequently tested confirmed positive for a viral marker. BPD reports are individually filed with the FDA and are not associated with a denominator. However, it is noteworthy that the number of BPD reports rose 18% for blood establishments and 60% for plasma establishments between 2007 and 2008. These observations, which encompass reports from all blood and plasma establishments, may reflect more complete reporting under a recently established Web-based system. However, they are consistent with the apparent increases in HIV and HCV incidence in ARC donors reported by Zou and colleagues. Blood establishment BPD reports of a unit of blood distributed from a marker-negative donor who subsequently tested confirmed positive. Plasma establishment BPD reports of a unit of blood distributed from a marker-negative donor who subsequently tested confirmed positive. The increase in FDA BPD reports of donors who subsequently confirm positive for TTD markers suggests an ongoing need for postinterview follow-up to investigate underlying risk factors including high-risk behavior. Assessment and analysis of prevalence and incidence data (Gap 8) along with understanding of the characteristics of those donors could help in policy decisions, recruitment strategies, and screening practices to strengthen donor selection and maintain the safety of the blood supply. As the authors point out, understanding of independent associations with demographic characteristics such as age, sex, high-risk behavior, geographic location, and medical procedures can help to improve donor selection and screening. One example of the use of hemovigilance data has been reported from Germany. Germany with its federal system has 29 competent authorities to ensure compliance with European Directives such as hemovigilance. Through analysis of hemovigilance data, the Paul-Ehrlich-Institut made a decision to mandate minipool NAT screening for TTD markers of not only blood products for transfusion but also therapeutic plasma.19 After several years of planning, nationwide tools to collect donor and recipient adverse events have been developed in the United States to improve patient safety and donor health.20 This year the CDC National Healthcare Safety Network has released its hemovigilance module for national participation, which allows health care facilities to report both infectious and noninfectious transfusion-related adverse events among recipients.21 Enrollment is voluntary, and aggregated data will be analyzed periodically. Additionally, AABB has developed a blood donor hemovigilance tool through support from HHS. This system has been piloted and now is ready to be launched nationwide for tracking adverse donor events. The studies by Zou and colleagues demonstrate the power of large-scale data collection to permit analysis of trends in blood safety. However, routine, national hemovigilance systems to collect and analyze data from all blood establishments would provide the most robust strategy for blood safety monitoring. Ideally, this system should include standardized protocol and definitions, reporting of denominators (number of collections, released units, and transfusions), adverse events in donors and recipients, and product deviations. Routine ongoing monitoring of TTD marker rates coupled with post-TTD conversion interviews of donors would help to form a composite picture of the current safety of the blood supply and the sources of potential threats from known agents. Comprehensive hemovigilance is an important goal both nationally and globally. It is clear that hemovigilance data and its appropriate analysis can be a vital tool to improve blood safety by identifying opportunities for advancements in prevention and control. Zou and colleagues at ARC are commended for their illuminating work in this area of transfusion-transmissible disease. None. The opinions expressed in this editorial are not necessarily those of the Department of Health and Human Services or its Operating Divisions.