Parvoviruses and blood transfusion

Abstract

In recent years many of our assumptions about human parvoviruses and blood transfusion have been challenged. It was thought that human parvovirus B19 (B19V), the only parvovirus that we needed to be concerned about, had a single genotype, was highly resistant to heat inactivation, and caused a transient viremia. Concerns about this viremic stage were abrogated by the fact that pooled plasma would contain large amounts of neutralizing antibody, rendering the virus noninfectious. Recent studies, including three articles in this issue of TRANSFUSION and the discovery of two new viruses infecting humans, show that this complacency is unjustified. This editorial reviews the potential ongoing threat of parvoviruses to blood and blood product safety in the light of these new discoveries. Parvoviruses are small, 20- to 25-nm icosahedral nonenveloped viruses, encapsidating a linear single-stranded DNA genome of approximately 5000 nucleotides. They are found ubiquitously in nature, infecting a wide range of vertebrates and insects. The viruses encode no DNA polymerase and are dependent on either cells undergoing division or infection with a helper virus for viral replication. The absence of a lipid membrane and the relatively small DNA genome renders the viruses relatively stable and resistant to many procedures used for inactivating blood-borne viruses. Several different parvoviruses have now been identified in humans, including B19V,1 adenoassociated viruses (AAVs2), PARV4 virus,3,4 and the human bocavirus (HBoV).5 These viruses are all distinctly different in terms of their pathogenicity, epidemiology, frequency of detection in blood and blood products, and therefore in their potential threat to blood transfusion safety. B19V was first identified in 1975, and the consequences of infection with this virus are now well described, with the specific disease manifestation depending on the host.6 In healthy children and adults it causes erythema infectiosum (fifth disease, or slapped-cheek disease), an erythematous disease that can be accompanied by varying degrees of arthralgia and arthropathy. B19V infection in patients with low erythropoietic reserve leads to transient aplastic crisis, and in patients who are immunocompromised chronic infection can lead to pure RBC aplasia or chronic anemia. Infection during pregnancy can result in fetal loss or the development of hydrops. Rarely B19V infection may be associated with myocarditis, hepatitis, and a variety of neurologic manifestations. Thus, the consequences of B19V virus infection range from asymptomatic to serious, and potentially fatal, conditions in a minority of the population that is particularly predisposed, through immunosuppression, decreased erythropoietic reserve, or both. B19V infection is a common illness of childhood, and by the age of 15 years approximately 60 percent of children have detectable IgG.7 Transmission is predominantly by the respiratory route probably by droplet spread and is highest at the time of viremia, before the onset of rash or arthralgia. It has been estimated based on the data from studies of susceptible women that approximately 0.5 to 1 percent of adult blood donors will seroconvert to B19V each year.7-9 In temperate climates most infections occur in the spring, with miniepidemics occurring at regular intervals several years apart.7 B19V is highly erythrotropic and replicates to high titer in erythroid progenitor cells (BFU-E and CFU-E) in peripheral blood and marrow.10 In healthy individuals at the height of the transient viremia viral titers as high as 1014 IU per mL are detectable. It is this high viral load that is transient, and in our own unpublished data (Brown and Yu) B19 DNA levels drop to about 106 IU per mL in the first week as the antibody response develops and then more slowly drops to about 103 IU per mL over the next 1 to 2 months. Levels of between 102 and 103 IU per mL can then persist for many months and even for years. In tissues there is good evidence that B19V DNA detectable by sensitive PCR may remain detectable for life,11 and detection of viral sequences in tissues constitutes a “Bioportfolio” of viruses an individual has been infected with in life. This much slower resolution of low levels of B19 DNA in serum than previously anticipated is reflected in the data of Schmidt and colleagues12 where in the donors with high-titer DNA detected the levels dropped, within the first 3 months, but were detectable in most cases 6 months after the original donation. It is also reflected in the different figures given for the prevalence of B19V DNA in blood donors; in studies with a relatively insensitive method such as CIE (detection limit, approx. 109 IU/mL) used in the earliest studies of B19 in donors, rates of approximately 0.001 to 0.003 percent were observed,13,14 compared to PCR-based studies in this issue where there are prevalence rates of 0.26 percent (Schmidt and coworkers, PCR sensitivity, approx. 103 IU/mL) and 0.88 percent (Kleinman and coworkers,15 PCR sensitivity, 20 IU/mL). Thus, depending on the sensitivity of the assay used and whether there is a local outbreak of B19V infection, this will give very different estimates of the prevalence of B19V DNA in blood donors. The studies in this issue of TRANSFUSION are of considerable value in assessing the transfusion risk of parvoviruses. Because of differences in the use of inactivation procedures and pooling strategies and the use of specific screening methods for B19 DNA, the transfusion risk of B19V for recipients of fractionated plasma derivatives and cellular components, such as RBCs and PLTs, and plasma units for transfusion (e.g., fresh-frozen plasma) should be considered separately. For the latter group, neither screening for B19V nor viral inactivation is normally performed. The frequent occurrence of low-level viremia of B19V and the lack of blood component inactivation potentially expose a measurable proportion of susceptible individuals to B19 infection through transfusion. Whether these low levels of B19V DNA in blood and blood products are actually infectious, however, remains uncertain. Although the B19V that persists has been shown to contain full-length coding sequences,16,17 the reactivation of a potentially infectious virus may be prevented by cytotoxic T-cell responses that persist for months or years after acute infection.18 In addition, the low levels of virus observed in some blood donors may be effectively neutralized if the donors are seropositive for the presence of anti-B19. Furthermore, from the study from Kleinman and coworkers, frequent transmissions resulting from the observed frequency of B19V viremia of approximately 1 percent in donors might be expected to have been recognized clinically if transfusion-acquired infection of B19V occurred with any regularity. Several factors, however, may hinder its identification. First, a considerable proportion of blood recipients (>80% given the age distribution of recipients of blood and blood products) have been previously exposed to B19V, have neutralizing antibodies, and would therefore likely be resistant to reinfection. Second, the symptoms and clinical presentation of B19V infections acquired through transfusion may be unlike those from respiratory routes, particularly among the immunosuppressed, or when given in the presence of B19V antibody from other transfused blood components or products administered around the same time. Finally, although there has been extensive surveillance of recipients of plasma derivatives for iatrogenically acquired B19 transmissions, the authors are not aware of any large-scale equivalent investigation of blood component recipients to specifically address the issue of B19 transfusion risk. For these reasons, it is difficult to establish the true ongoing frequency of transfusion-acquired B19 infections or its clinical consequences; the findings from Kleinman and colleagues15 and Schmidt and colleagues12 reported here are of value in refocussing attention on this issue. For recipients of plasma derivatives, considerable progress has been recently made to reduce the transfusion risk of B19V. A study of B19V in factor (F)VIII concentrates suggested a minimum infectious dose of 2 × 104 IU in the absence of anti-B19 IgG.19 In plasma pools, it is almost certainly different due to the presence of B19 antibody in the preparation. In a postmarket surveillance study of solvent/detergent (S/D)-treated plasma in B19V-seronegative plasma pool recipients, it was shown that only those recipients who received 200 mL of plasma containing more than 107 IU per mL became infected or seroconverted, whereas there was no seroconversion in those that received products containing less than 104 IU per mL.20,21 It is on the basis of this data that the FDA in the US and the European Pharmocopoeia (5.0) derived their current recommendation and requirements limiting the levels of B19V DNA in plasma pools. Although B19V clearly is resistant to many procedures used to inactivate viruses, and transmission by heat-treated blood products is well documented,22 it has long been known that it is not quite so resistant to heat as other parvoviruses.23 The heat stability varies widely depending on the environment,24,25 especially the diluent used, and the residual amount of moisture.26 The structure of the B19 capsid differs from that of other parvoviruses at the molecular level,27 and it is this difference in structure that contributes to the increased instability of B19V DNA reported in this issue.28 Although this is good news for inactivation studies with animal parvoviruses as models for B19V inactivation, extrapolation of this increased instability to other human parvoviruses cannot be ascertained without experimental studies. The strategy of screening plasma for B19 DNA to eliminate high-titer units, and reduce residual viral load to levels that can be effectively inactivated, depends on assays capable of detecting all genetic variants of B19. Although the nucleotide sequence of the B19V genome is relatively well conserved, at least compared to RNA viruses, it is now recognized that there are three distinct genotypes.29 Genotype 1 isolates show less than 5 percent sequence variation and have a worldwide circulation, with isolates from different parts of the world and from different periods. In contrast the other genotypes appear to be much rarer, especially in most parts of Europe and America, although Genotype 3 infection may be common in some parts of Africa,30 and studies based on the detection of B19 DNA in tissues suggest that Genotype 2 viruses were more prevalent in the 1960s. Many primers pairs used for detection of B19V DNA in blood products have not been optimized to detect all three genotypes and unless specifically designed will miss some of these variant sequences.31,32 Even with these caveats in place, the current screening program has led to a marked decrease in the viral loads in plasma pools and plasma pool fractionation products.33 B19V DNA, however, is still detectable in some of these products, necessitating viral inactivation and the need for continuing surveillance of B19V levels in plasma pools. Although there is a considerable amount of information about the transmission risk of B19 by transfusion and the development of strategies for its prevention, the recent discovery of two new human parvoviruses poses a much more tenuous and uncertain risk for transfusion safety.3-5 HBoV is now recognized as a primarily respiratory virus, with evidence for a significant etiologic role in infant and young child respiratory disease.5,34-36 Similar to B19V, infections with HBoV are systemic, with viremia detectable during acute infection,37 with indirect evidence for HBoV spread and replication in the gastrointestinal tract (sequences can be detected in feces from individuals with gastroenteritis37). HBoV differs from B19V, however, in not establishing long-term persistence of viral DNA sequences; all studies of autopsy and biopsy specimens of marrow, lymphoid tissue, and lung have proven negative for HBoV by PCR,38 whereas B19 remains detectable in such tissues lifelong.11 Large-scale screening of pooled plasma samples has failed to detect HBoV DNA sequences,39 in marked contrast to the near ubiquity of B19 contamination of plasma used for blood product manufacturing. The absence of detectable HBoV viremia in blood or plasma donors is, however, perhaps an indication of the restriction of HBoV infections to young children observed in previous studies of respiratory samples. Its absence in the donor population may therefore be the direct result of much higher frequencies of past infection and immunity in adults compared to B19V. The research community awaits a reliable serologic assay for HBoV antibody detection to resolve these uncertainties about its real epidemiology, prevalence, and age of acquisition. PARV4 was originally cloned from an individual with an acute febrile syndrome who was at risk for infection with human immunodeficiency virus (HIV) through injecting drug use (IDU) or high-risk sexual behavior and enrolled in the San Francisco–based Options Project cohort.3 In marked contrast to both B19V and HBoV, infections with PARV4 appear to specifically target individuals with histories of parenteral exposure, in whom it establishes persistent infections. In our own studies,38,40 PARV4 DNA sequences were detected in autopsy marrow and lymphoid tissue only in IDUs and persons with hemophilia and absent in study subjects without a history of parenteral exposure. Higher frequencies of infection were found among IDUs coinfected with HIV (70%), but PARV4 infection was absent in other risk groups for HIV infection, such as male homosexuals. Our conclusions from these two studies are that parenteral exposure represents the principal transmission route of PARV4 and that its association with HIV in IDUs is a simple reflection of a shared relative inefficiency of transmission by this route (compared to viruses such as hepatitis C virus infections that are close to universal in IDUs who have shared nonsterilized needles). The limited transmission of PARV4 may result from the nature of the persistent infection it establishes. Although PARV4 DNA sequences remain detectable in marrow and lymphoid tissue lifelong, this is not associated with a persistent viremia. With a highly sensitive PCR (10 DNA copies/mL), plasma samples from HIV-infected IDUs were uniformly negative for PARV4 DNA sequences despite the likely existence of past exposure and persistent infection of PARV4 in more than 50 percent of this study group.38,40 In this respect, the low frequencies of chronic low-level viremia reported in B19-seropositive study subjects (≤1%), despite its lifelong presence in marrow and lymphoid tissue, may indeed be recapitulated for PARV4. Infections of PARV4 in IDUs therefore may be largely restricted to serial transmissions between previously unexposed IDUs undergoing preseroconversion viremia during primary infections. This type of transmission network would be analogous to that of HIV in this risk group. In Edinburgh, for example, the remarkable prevalence of HIV in Edinburgh IDUs in the 1980s41 is thought to have arisen originally through a chain of acute infections originating from a single infected individual from Spain, an occurrence that earned their city the (temporary) dubious honor as the “AIDS capital of Europe.” Genetic analysis of PARV4 sequences among Edinburgh IDUs supports the hypothesis for its rapid, recent epidemic spread. PARV4 comprises two genotypes that differ from each other by approximately 13 to 14 percent in nucleotide sequence (similar to the divergence between B19 genotypes 1, 2, and 3).29,42 IDUs born before 1956 (and who were likely first infected with HIV and PARV4 in the 1980s) were invariably infected with Genotype 2,4 whereas those infected later (dates of birth from the 1960s) invariably carried Genotype 1.38 Nucleotide sequences of Genotype 1 amplified from the more recently infected Edinburgh IDUs were almost invariant and frequently identical to the Genotype 1 sequence in the individual from the US cohort in whom PARV4 was first discovered. Because we now know that the rate of sequence change of parvoviruses is relatively rapid,43 this extremely limited diversity predicts a common ancestor for Genotype 1 within the past 15 years. IDUs in needle-sharing transmission networks may be subject to ongoing, epidemic transmission cycles of PARV4, an occurrence with considerable consequences for blood transfusion practice. As with HBoV, much of the current uncertainty about the epidemiology and transmission of PARV4 would be resolved with the development of serologic assays for PARV4-specific IgM and IgG for diagnosis of acute infection and as a marker for past exposure, respectively. (The current reliance on autopsy tissue testing for necessarily retrospective diagnosis is clearly a limitation in current epidemiologic studies.) Accurate diagnosis would be of considerable value in determining the pathogenicity of PARV4, for which absolutely no data have been obtained to date. It is established that parvoviruses generically target dividing cells, however, and their replication during acute infection leads to varying degrees of impairment of marrow function, particularly erythropoiesis in the case of parvovirus B19, as well as fetal damage and frequently pathogenic infections of the respiratory and gastrointestinal tracts (feline and canine parvoviruses, HBoV). Although some of these disease outcomes may occur during the acute phase of PARV4 infection, much less is known about the effect of chronic infections. In assessing the risk of transmission of PARV4 to recipients of blood and blood products, information on its prevalence in acutely and chronically infected blood and plasma donors is required, the existence of identifiable risk factors (such as past IDU) that would lead to donor deferral, its transmissibility and partitioning into cellular and plasma components, and in the case of plasma derivatives, its resistance to currently used viral inactivation procedures. Recent investigations at the National Institute for Biological Standards and Controls provide data on prevalence highly relevant to this assessment.4,39 Frequencies of PARV4-contaminated archived and more recently collected source plasma pools from several blood product manufacturers were 21 percent (24 positive from 115 tested) and 4 percent (14 from 351),39 respectively. Although this frequency, and measured viral loads in positive pools (<200 × 105-7.6 × 105 DNA copies/mL) were both substantially lower than recorded for B19V, they nevertheless represent a level of contamination that could lead to transmission of PARV4 to recipients of blood products. Indeed, our finding of PARV4 infection in autopsy samples from persons with hemophilia40 provides direct evidence for its actual occurrence. A contributory factor to its transmission by fractionated plasma derivatives is the possibility that PARV4 will be resistant to currently used viral inactivation procedures applied to plasma-derived blood products. Although different genera of parvoviruses vary in their resistance to viral inactivation procedures, the observation that even the relatively sensitive parvovirus B19 is regularly transmitted by virally inactivated plasma-derived FVIII and F IX in the absence of specific screening to remove high titer donations strongly suggests the likelihood that PARV4 may be similarly transmitted. The data obtained so far on the epidemiology and presence in donations of HBoV and PARV4 are very much part of the blood industry's proactive response and close surveillance advocated for newly emergent parenterally transmitted virus infections.44 Although we have no information on the ultimate origins of these new parvoviruses or the time scale of their emergence, findings of a predominantly parenteral route of transmission of PARV4 as well as its lack of genetic diversity (that demonstrate its very recent origin in IDUs) point toward recent large-scale epidemic infection with essentially unknown consequences for those exposed to virus infections through parenteral routes, therapeutic or illegal. In summary, our knowledge of the role of parvoviruses in blood and blood products has changed markedly in recent years, and the assumption that this is a group of innocuous viruses with limited risks to transfusion is no longer tenable. There are still large areas, however, where more information is needed, not least in the pathogenicity of PARV4, inactivation kinetics of PARV4 and HBoV, and the infectivity of low levels of any of the viruses. We might also highlight the lack of any research or assessment of transfusion risk from AAVs in blood and blood products. Although these viruses are generally considered to be nonpathogenic, AAV DNA sequences can be detected in monocytes,45 and in one study in more that 30 percent of marrow samples examined.46 The consequences of transmission to immunosuppressed patients, for example, are entirely unknown. Finally, there is nothing to suggest that we have identified the full range of parvoviruses that can infect humans and hence the requirements for ongoing surveillance and robust inactivation methods to tackle this group of challenging viruses.