The rapid, pandemic spread of Zika virus (ZIKV) spurred an international public health emergency in 2015. Cases of ZIKV infection have now been reported in 85 countries or territories1; in 72 of those locations, no such cases had been reported previously. The association of ZIKV with severe teratogenic effects,2 its persistence in whole blood,3 and four possible cases of transmission of ZIKV by blood transfusion in Brazil4-6 have raised questions about a potential risk to the blood supply. In February 2016, the Food and Drug Administration (FDA) recommended that collection of donated blood cease in areas in the United States in which ZIKV was active, unless blood testing or pathogen-reduction methods could be implemented.7 At the time, no vectorborne cases of ZIKV infection had been reported in the continental United States; however, an epidemic was under way in the U.S. territories, most notably in Puerto Rico and the U.S. Virgin Islands. An unprecedented response from both blood-collection centers and industry (i.e., commercial manufacturers of donor screening tests) followed; within 6 weeks after the FDA guidance was issued, blood-donation screening began in Puerto Rico with the use of a newly developed, high-performance nucleic acid test for ZIKV under an investigational new drug application. In August 2016, the FDA expanded its earlier guidance to include blood-donation screening throughout the continental United States. Consequently, all donated blood in the United States is now screened for ZIKV. In this issue of the Journal, results of the screening of more than 4,000,000 U.S. blood donations for ZIKV during a 15-month period in 2016 and 2017 are reported.8
The current U.S. strategy for ZIKV screening comes at a cost with an unclear gain. Although concerns about blood safety are vital, is it appropriate and possible to reverse blood-donation screening for ZIKV in the United States in light of emerging data? Addressing this question requires that we understand the variables that contributed to the decision to implement blood-donation screening along with the reasons to consider whether those variables remain relevant.
Several important considerations suggested the need for blood-donation screening.9,10The first is the obligation to prevent undesirable outcomes. ZIKV infection is associated with severe neurologic sequelae in babies born to ZIKV-infected women,2 which raises particular concern for transfusions that occur during pregnancy. The risk that transfusion recipients might infect their sexual partners, who could be pregnant or trying to conceive, is also of concern. The persistence of ZIKV RNA in semen is well established (a recent estimate of the median time to clearance is 34 days).11,12 Pertinent to blood donation, 5 of 14 men (35.7%) who were identified during routine blood-donation screening in Puerto Rico and Florida as being positive for ZIKV RNA in plasma had detectable ZIKV RNA in their semen. Although subsequent viral cultures were negative, this finding echoes earlier concerns about possible viral propagation.13
The second consideration that suggested the need for screening is the demographic distribution of infection in both the general population and the blood-donor population that relates to the risk of collecting ZIKV-infected blood. ZIKV infection is now widely distributed throughout the Americas and the Caribbean. In the continental United States alone, more than 200 locally acquired cases of mosquito-borne ZIKV infection and more than 5300 cases of travel-associated ZIKV infection have been reported.14Competent vectors (notably Aedes aegypti and A. albopictus) are widely distributed across the southern and southeastern United States,14 further amplifying the risk of the spread of ZIKV infection. Although there is a paucity of data regarding the rates of ZIKV infection among blood donors before routine testing began, a molecular survey of blood donations during a 2013–2014 outbreak in French Polynesia showed that 42 of 1505 blood donors (2.8%) were positive for ZIKV RNA at the time of the donation, and 11 of the 42 donors subsequently became symptomatic.15 In 2016, nucleic acid testing performed in blood donors in Puerto Rico revealed an increasing incidence of ZIKV RNA-positive blood donations that peaked at 1.1% during the final week of reporting.16
The third consideration that suggested the need for screening is whether blood collected from infected donors confers transfusion-related risk. Making this determination necessitates evaluating the ability of ZIKV to survive standard storage and processing conditions as well as the transmissibility of ZIKV through transfusion. To date, four cases of transfusion-transmitted ZIKV infection have been reported.4-6 All four cases were from Brazil; three occurred after the transfusion of platelets and one after the transfusion of packed red cells.
The fourth consideration is that, because the vast majority of ZIKV infections are subclinical, history taking is an unreliable means of identifying infected donors. The fifth is the successful development of high-performance tests that were amenable to rapid implementation.17
Preservation of public trust in the safety of the blood supply is another imperative that prompts the implementation of blood-donation screening. However, does screening for ZIKV take that imperative too far? When data are scant, as was the case early in the ZIKV pandemic, the effect of fear on decision making should not be underestimated, particularly amid media reports regarding infection of pregnant women and the possibility of devastating neurologic effects in their offspring. The U.S. public has a low threshold for risk to the blood supply, in part an aftermath of historical failures surrounding the human immunodeficiency virus (HIV) pandemic and the regulation of blood safety.18 As a result, many of the current risks pertaining to blood safety are so low that, at least in the United States, they are now based on modeled estimates rather than on actual reporting.18 These efforts to greatly minimize risk are reflected in the cost effectiveness of blood-donation screening measures, many of which have exceeded $1 million per quality-adjusted life-year gained by the intervention, which is approximately 10 times as high as costs deemed appropriate in clinical medicine.19
The precautionary principle maintains that when confronted by a potential public health risk, one should act to safeguard against such a risk, even if data are incomplete.20 When the principle is invoked, there are core elements that should not be overlooked. These include proportionality, such that a given intervention matches that of the threat, consistency with respect to previous adopted measures, and review in the emergence of new scientific data.21 Acting in the face of uncertainty is not new to public health. The elements that are lacking in the context of blood-transfusion safety are communication of risks to the public, auditing practices, and implementation of appropriate modifications to policy as new data emerge. Such is consequent to the unavoidable challenges associated with aligning various stakeholders, including blood donors, blood-collection centers, industry, patients, professional organizations, and regulatory agencies alike, each of which is critical to the process. For example, blood-collection centers have to reconcile the ideal response to a potential public health risk with their ability to ensure a safe and adequate blood supply to meet demand. Industry invests in a mercurial market, and failure to deliver on such investment may compromise the ability to attract industry partnership when the next transfusion risk emerges. Regulatory agencies (e.g., the FDA) are blamed either for being too vigilant, thus wasting resources, or for being too lax in a society with little appetite for risk. Timing is also critical. In the face of an epidemic, the collection of useful information is timed to media and societal attention. There is urgency to implement an intervention while funding is available and before the inevitable complacency arises.22
Nevertheless, the question is no longer about whether to initiate blood-donation screening for ZIKV in the United States, but whether to continue, modify, or stop screening. Many factors need to be considered. First, we are unaware of any case of clinical ZIKV infection that has been ascribed to blood transfusion. A distinction between laboratory evidence (e.g., detectable ZIKV RNA, antibodies, or both) and symptoms, signs, or clinical sequelae of infection is critical in characterizing the risk posed to transfusion recipients.10 The four reported cases of transfusion-transmitted ZIKV infection were based solely on laboratory evidence; of the three living transfusion recipients, none displayed any clinical sign of ZIKV infection. The ability to detect evidence of transfusion-transmitted disease without clinical manifestations is not unique to ZIKV. For instance, during a dengue virus (another flavivirus) epidemic in Brazil, more than a third of the units of blood that were positive for dengue virus RNA were shown to transmit the virus to transfusion recipients,23 yet clinical cases of transfusion-transmitted dengue virus infection are extraordinarily rare. Second, even if ZIKV did pose a clinical risk to transfusion recipients, much of that risk pertains to fetuses. At least in the United States, receiving a transfusion while pregnant is uncommon (0.24% to 0.46% of pregnancies24), which suggests that national screening is unnecessary or alternatively, that compiling a limited inventory of ZIKV-screened blood for use in high-risk recipients (e.g., women receiving intrauterine transfusion) could be considered.10 Third, ZIKV remains rare among U.S. blood donors, and most cases identified during screening would otherwise have been identified through inquiry regarding travel to a country in which ZIKV is highly endemic.25 Between April 2016 and April 2017, the screening of 4,065,045 blood donations across the United States resulted in the identification of 30 cases (less than 0.001%) that met criteria for true positive results for ZIKV on the basis of repeat testing results.26,27
The cost and economics of screening also warrant consideration. Industry partners are critical to the development of new tests, since a substantial investment is needed to carry an assay through the regulatory process, including validating the test, conducting investigational trials, obtaining licensure, and managing surveillance after implementation. Under exceptional circumstances, such as those that occurred with ZIKV, this process may be expedited. Otherwise, the process may be measured in years. Although ZIKV testing has been favorable from a commercial standpoint, it has incurred an unanticipated increase in screening cost (of $7 to $13 per donation in U.S. dollars) that has been passed down to blood centers, hospitals, and patients. The current strategy of nucleic acid testing for ZIKV in individual donors is projected to incur an annual cost of $137 million.28 The cost could well be justified if testing were shown to prevent adverse effects in transfusion recipients, but this has arguably not been the case.
To place ZIKV in context, there are established infectious risks to the U.S. blood supply that are not being addressed. For example, more than 200 cases of transfusion-transmitted babesiosis with an associated mortality have been reported in the United States.29 Despite long-term recognition of this risk, screening for Babesia microti has not yet been mandated. Likewise, bacterial contamination, notably of platelets, remains the foremost infectious risk to the U.S. blood supply30; such contamination has resulted in septic reactions and fatalities, despite extant safeguards (e.g., bacterial culture). Collectively, these risks arouse ethical concerns regarding the appropriate allocation of resources to enhance blood safety whereby competing measures that are probably of greater overall benefit may be foregone in the interest of mitigating perceptions of risk. However, the assumption that resources allocated for one priority will necessarily be reallocated appropriately to another may be naive.
Timing is another major consideration. Specifically, it may be premature to cease testing donated blood for ZIKV when risk has not yet been fully characterized. A growing body of published and operational data that have been generated during the past year give pause. For example, the initial rate of blood donations in the continental United States that were positive for ZIKV RNA but negative for ZIKV antibody exceeds the rates of donations that were positive for HIV RNA, hepatitis C virus RNA, and hepatitis B virus RNA combined.25,31 In addition, a study conducted in Martinique showed that more than half (54.7%) of blood donors who tested positive according to nucleic acid testing reported ZIKV symptoms at a follow-up visit 2 weeks later.32 Studies are also under way to address the natural history and epidemiology of ZIKV in blood donors, the ability of ZIKV to survive storage and processing conditions in blood banks, and transfusion transmissibility of ZIKV in animal models.33 Data from these efforts, which are likely to be available in the near future, would strengthen any policy decisions, thus lending support to deferring decisions surrounding any immediate change.
Although ZIKV has not been eliminated, the epidemic has waned substantially.34 In the continental United States, the reported number of locally acquired, mosquito-borne cases of ZIKV infection decreased from 226 in 2016 to 2 in 201714; at the time of this writing, there are no active areas of ZIKV transmission, and donations that are positive for ZIKV RNA are rare.35 Now that ZIKV has made its way through much of the Americas, the change in the burden of infection may be swaying the argument to reassess blood-donation screening in the United States. A meeting of the Blood Products Advisory Committee of the FDA was convened on December 1, 2017, to consider all available options. These include continuation of donation screening, restriction of testing to states designated as being at high risk of ZIKV, restriction of testing to areas with ZIKV activity and to donors who identify themselves as having a risk of exposure to ZIKV, maintenance of a limited inventory of screened blood for high-risk recipients, and finally, cessation of all measures until a new, substantial outbreak occurs.35 Each option attempts to weigh the risk with the cost and logistic complexity of implementation.
Cessation of blood-donation screening may actually prove to be far more challenging than the decision to start. Indeed, there is no historical precedent for the termination of a blood-donation testing program for a given pathogen. The few examples of changes in testing practices have represented either a substitution in cases in which an improvement in testing became available (e.g., measurement of alanine aminotransferase levels for hepatitis C was suspended in favor of tests for hepatitis C virus antibody and tests for HIV type 1 p24 antigen were ceased in favor of enhanced testing for HIV) or a modification in practice (e.g., testing for Trypanosoma cruzitransitioned from testing every donation to testing a donor’s first donation only). The reluctance to withdraw intervention measures related to blood safety was recently evaluated.36 The authors examined the underlying assumptions associated with this reluctance, including the notion that omission of a protective measure is viewed more favorably than active withdrawal of care or protection and a belief that a given intervention should be continued once it is initiated. The authors found fault with both assumptions; they concluded that adverse events are rare in the case of withdrawing blood safety interventions and that the merits of sustaining low-yield interventions are highly questionable.
The ZIKV blood-donation screening debate highlights the complexity of decision making and the need for objective and explicit characterization of risk. To this end, risk-based decision frameworks that are tailored specifically to the blood-transfusion setting have been developed.37 When such frameworks are applied effectively, the variables that affect transfusion-associated risk are assessed systematically and weighed against the currently available evidence. Rather than simply confining the assessment to medical and scientific evidence alone, these frameworks are broad-based and examine the interplay of health economics and outcomes assessments, social perspectives, and ethical and legal considerations.37
In conclusion, the available medical and scientific evidence suggests that the risk of transmission of ZIKV to transfusion recipients remains low and is likely to diminish further as the pandemic continues to wane. There is no doubt that ZIKV poses a major public health threat. However, the actual and perceived risks to the blood supply seem to be conflated. The initial decision to test for ZIKV should not be condemned; a similar proactive response to HIV could have resulted in a reality very different from the pandemic that unfolded in the 1980s. The precautionary principle and risk-based approaches encourage continuous review, with reassessment as new data emerge.18 The hurdle now seems to be the willingness of key stakeholders to conduct such a review and to lead a graceful retreat from a policy decision that has thus far shown limited utility.
