Hepatitis B virus traceback and lookback: factors to consider


  • F. Blaine Hollinger,

    1. From the Departments of Medicine, Molecular Virology & Epidemiology, Baylor College of Medicine, Houston, Texas; and the American Red Cross Holland Laboratory, Rockville, Maryland.
    Search for more papers by this author
  • Roger Y. Dodd

    1. From the Departments of Medicine, Molecular Virology & Epidemiology, Baylor College of Medicine, Houston, Texas; and the American Red Cross Holland Laboratory, Rockville, Maryland.
    Search for more papers by this author

  • This work was supported in part by the Eugene B. Casey Foundation and the William and Sonya Carpenter Fund, Baylor College of Medicine.

F. Blaine Hollinger, MD, Baylor College of Medicine, One Baylor Plaza, BCM-385, Houston, TX 77030; e-mail: blaineh@bcm.edu.


American Red Cross


hepatocellular carcinoma


individual-donation nucleic acid testing

Numerous studies suggest that the residual risk of transfusion-transmitted hepatitis B virus (HBV) is significantly greater than that for hepatitis C virus (HCV) or human immunodeficiency virus (HIV),1,2 yet there is no requirement in the United States to perform lookback for HBV. Why is this and is it time to reconsider? The practice of lookback was formalized as a response to the emergence of HIV/acquired immune deficiency syndrome (AIDS) as a transfusion-transmissible disease. Shortly after the initiation of testing for HIV, it became apparent that recipients of prior donations from newly tested, seropositive donors were at risk of infection, because the majority of the detected donors had prevalent infections. Subsequently, it was recognized that prior seronegative donations from incident cases also conferred risk of infection. Indeed, quantitation of this observation led to the first estimate of the length of the infectious window period.3 Although no effective AIDS treatment was available at the time, there were powerful arguments to support the introduction of lookback, including the need to provide medically important information to exposed recipients and to provide a means to prevent secondary infections from those that were infected. Similarly, lookback was implemented after the initiation of donor testing for antibodies to HCV but only after considerable public debate.4 It should be noted that lookback is not routinely conducted for other transfusion-transmitted infections at this time.

There are several issues that need to be reviewed and resolved before lookback for HBV can be recommended. One of these relates to the value of routine HBV nucleic acid testing (NAT) in pools or in individual donations by blood banks to identify donors who might be more likely to have transmitted HBV previously. Currently in the United States, this approach has not been recommended or mandated by the Food and Drug Administration (FDA), although several organizations have unilaterally initiated minipool testing for HBV DNA. Regardless, consideration of a credible policy for HBV lookback will require careful balancing of benefits against cost and effort for hospital transfusion services, community blood collection centers, and the national blood organizations. Should lookback be recommended, the following discussion represents suggestions for conducting such an investigation. The information presented herein provides relevant data supporting lookback for HBV along with additional information that should be considered in making a decision on this issue. This discussion on lookback evolved from a recent review in TRANSFUSION on the vagaries of occult hepatitis B in transfusion medicine5 and from deliberations conducted by the Transfusion-Transmitted Diseases Committee of the AABB. The views presented, however, are those of the authors and not of any specific group.


A distinction is made between the notification and testing of donors that have been implicated in a case of possible posttransfusion hepatitis B (traceback) and the notification and testing of recipients after the administration of potentially infectious HBV-containing blood (lookback). The traceback strategy relates primarily to donor evaluation and consists of elements that are both forward-looking (follow-up samples) and back-tracking. The donor assessment ideally should be executed regardless of any other history obtained from a recipient (such as injection drug use or high-risk sex) who presents as a confirmed case of acute hepatitis B (i.e., hepatitis B surface antigen [HBsAg] and immunoglobulin M [IgM] hepatitis B core antigen [anti-HBc] positive) even though the probability of transfusion-associated infection may be low. These investigations should include examination of appropriate follow-up specimens from the donor or the results from recipients who were transfused with subsequent donations from each definitively implicated donor employing licensed tests of optimal sensitivity and specificity. This might include individual-donation NAT (ID-NAT) with a limit of detection of 10 IU per mL or less in addition to HBsAg, total anti-HBc, and IgM-specific anti-HBc. Antibody to hepatitis B e antigen (anti-HBe) may also be helpful in the evaluation. The IgM-specific anti-HBc can remain positive for up to 2 years in patients whose disease has resolved and its evaluation would assist in substantiating a recent infection. Obviously, if an aliquot from the original donor sample is available, as it is in Japan and some other countries, retesting of that specimen would be optimal. Alternatively, at least a 3-month interval is recommended between the original donor sample and a follow-up specimen to allow for serology to develop, although a positive ID-NAT could shorten this time interval. Current practice is to clear and reinstate implicated donors who show no evidence of hepatitis infection on a subsequent donation or after follow-up. Assessment of a hepatitis B clinical or serologic event in other recipients who received blood components from a definitively implicated donor during a comparable time period should also be undertaken and include HBsAg and IgM anti-HBc testing. Unfortunately, in many instances the number of alleged cases of posttransfusion hepatitis B is often much higher than the number of actual transmissions, placing an added burden on the transfusion services and collection agencies.6

Because a definitively implicated donor may have been infectious before the index donation, it also is essential to review the HBV status of the donor's recipients over the prior 12-month period. The proposed time interval of 12 months is based on an extensive analysis of a large number of clinical trials7-14 in which the incubation period of hepatitis B occurred within 3 to 6 months after infection except in two instances7,14 where the time interval increased to 7 to 9 months presumably due to modulation of the disease by exogenous specific antibody (Table 1). Thus, a 12-month back-tracking time interval should capture all potential implicated donations associated with an acute case of hepatitis B. For a recipient who is discovered serendipitously to have had a recent hepatitis B infection (IgM anti-HBc–reactive) or to have chronic disease, donor traceback might have to proceed backward for at least 3 years or indefinitely for as long as records are available to determine whether an association exists between a previous specific donor and the recipient. Because many of these implicated donors would have donated subsequently, records would exist to render an opinion regarding their potential involvement in the transmission of HBV.

Table 1. Longest incubation periods in humans for HBV infection under various conditions
ReferenceConditionIncubation period (days)
Sawyer et al.7HBV contaminated yellow fever vaccine<210
Havens8Serum hepatitis (SH) virus≤70
Neefe et al.9Serum hepatitis (SH) virus<135
Hoofnagle et al.10 (1951-1954 HBV infectivity studies)Single implicated donor infectivity studies≤77
Barker and Murray11 (1951-1954 HBV infectivity studies)HBV dilution series<130
After UV irradiation≤134
After heating at 60°C for 2-4 hr≤110
After storage at 25°C for 90 days≤85
After treatment with β-propiolactone<130
Krugman and Giles12Intramuscular injection of HBV (MS-2)≤91
Oral inoculation of HBV (MS-2)≤108
Krugman and Giles13Diluted HBV (MS-2)≤91
MS-2 followed by IG 4 hr later≤112
MS-2 followed by HBIG 4 hr later≤162
Grady et al.14After normal or intermediate immune globulin≤180
After high-titer immune globulin≤270


Targeted lookback of recipients represents another strategy for identifying recipients at some risk of transfusion-transmitted infection. This relates to recipient evaluation when HBV screening serology from a repeat donor registers a documented seroconversion for HBsAg and/or total anti-HBc or is newly identified as HBV DNA–positive. One approach would be to restrict recipient evaluation to those who received blood components from repeat donors who are found to have developed a seronegative NAT-positive result or an HBsAg-positive result in the absence of anti-HBc (provided that the HBsAg reactivity was not due to a recent HBV vaccination or a misinterpreted neutralization inhibition result; reviewed by Hollinger5). In these situations, recipient lookback could be limited to only those donations made by the donor of interest during the previous 6 months, since any donor acquiring acute HBV infection before that time should also be anti-HBc–positive in the current seroreactive unit of blood, which should trigger a different notification strategy as discussed below. Recipients should be evaluated for a period of at least 6 months after their transfusion or until infection is documented if this occurs at an earlier time period. At a minimum, recipient blood samples should be tested for HBV DNA, HBsAg, and anti-HBc as discussed above. The addition of anti-HBs testing in an unvaccinated recipient also might provide further verification if only anti-HBc is found to be reactive.

For repeat donors who have developed a verified de novo anti-HBc conversion, it becomes next to impossible to determine when that response might have occurred so that lookback should continue backward to include those donations administered up to 6 months before the most recent transfusible donation for which records exist. Confidence in the authenticity (i.e., specificity) of the anti-HBc reaction can come from several directions, such as its cooccurrence with HBV DNA, HBsAg, IgM anti-HBc, anti-HBs, and/or anti-HBe. Anti-HBs results may be confounded by hepatitis B vaccination in the recipient, although this information is usually discoverable. In addition, the prevalence of vaccination in the adult population is currently limited, although this will change in the next generation. The validity of a positive test result for anti-HBc may also be supported by a strong signal level (a high signal-to-cutoff, or cutoff-to-signal ratio, depending on the test); appropriate levels have been defined for some, but not all, available tests (S. Stramer, personal communication, 2008). Additional approaches to verify a repeatedly reactive anti-HBc would be to use newer-generation assays with enhanced specificity (e.g., PRISM HBcore chemiluminescent assay, Abbott Laboratories, Abbott Park, IL) or to require repeat reactivity in two assays from different manufacturers. Similarly, follow-up of the donor to document the subsequent development of immunity (e.g., anti-HBs) can provide confirmation. Regardless, as previously stipulated, recipient lookback in this category should proceed backward for any units transfused from the implicated donor over a finite 6-month interval before and including the most recent seronegative donation, depending on donor and recipient record availability. Any investigation beyond that time span would have a trivial yield. In the absence of a pretransfusion sample from a recipient who has elected to participate in the lookback process, the probability of detecting an unassociated HBV serologic marker in that individual will depend on the prevalence of hepatitis B in a similar cohort of patients. Therefore, without a pretransfusion sample to establish susceptibility or prior infection, any conclusions regarding transmission remains suspect unless linkage can be established with certainty through sequencing of the virus if available. The fact that only one of the components from a donor may have been implicated in transmission should not divert attention from that donor since a low-component plasma volume, immunity in a recipient, or the administration of anti-HBs in additional blood components given concurrently to another potentially at-risk recipient could confound the outcome.


In the United States, at least four studies have predicted that the risk of receiving an otherwise suitable unit of blood from a repeat donor who is in the hepatitis B preseroconversion window period ranges from 1:29,000 to 1:63,000.6,15-18 Recently, however, downward trends in key infectious disease markers among presenting donors have lowered the estimated residual rate of disease in repeat donors based on the relationship between window period and incidence rate. For the American Red Cross (ARC), the estimated rate of infectious donations using current screening procedures was calculated to be 1 in 205,000.19 However, subsequent estimates based on 24-minipool NAT or ID-NAT and the use of more sensitive and specific assays for HBsAg and anti-HBc have cited an HBV yield of 1 in 352,451 donations15 to 1 in 610,488 donations.20 As these newer technologies are applied, we suspect that only a few more donors will be identified as potential sources of HBV infection from previously donated blood, thereby triggering lookback. To assess risk and to determine whether lookback should be conducted, assumptions regarding the potential for transmission and recipient outcome after an HBV infection need to be considered. Therefore, for the purpose of the following analysis, an HBV residual risk of 1 in 205,000 will be employed since it reflects current practice among first-time and repeat donors while recognizing that this is not the best- or worst-case scenario that could have been selected.

There are 14 million donations processed annually in the United States that could yield approximately 68 HBV DNA–positive donors. The number of components collected and transfused per donation is 1.4521 potentially resulting in the transmission of HBV to approximately 99 recipients. Not everyone who is transfused with HBV DNA–containing blood or blood components, however, will become infected for a number of reasons, such as immunity of the recipient due to immunization or prior infection, the presence of defective virions or low viral load in the blood, or mitigation of disease by specific anti-HBs in concurrent or sequentially transfused units. Therefore, assuming that the rate of transmission in the United States could be 75 percent,22-26 this reduces the number of recipients who will become infected to approximately 74 per year. Clinical data suggest that 20 to 35 percent of these recipient cases (mean, 28%) will be symptomatic (approx. 21 cases) with approximately 70 percent of these (approx. 15 cases) being hospitalized annually.27 As a result, these 21 HBV-infected patients will likely experience considerable morbidity for a finite time interval that can range from 2 to 6 months. These numbers for clinically relevant disease also are close to what the relatively inefficient passive surveillance monitoring system has observed in the United States for transfusion-associated hepatitis B.28 Of the 15 estimated cases hospitalized with acute, icteric disease, death from fulminant hepatitis is predicted to occur within 2 to 4 months after exposure in approximately 1.0 percent of the cases (range, 0.5%-1.5%) or in approximately 0.15 recipients per year, although there have been circumstances when the case-fatality rate has approached 50 percent.29,30 This translates to about one death every 7 years from fulminant disease after transfusion of HBV-infected blood, assuming that none of these patients will receive a transplant (which presents its own set of problems and costs). The above estimate implies that approximately 53 of the 74 infected recipients will develop inapparent or anicteric hepatitis and remain unrecognized in the community, although they would still be capable of transmitting disease while infectious.

Long-term outcome is difficult to predict since morbidity and mortality from hepatitis B following progression to chronic hepatitis is time-dependent and is subject to recipients surviving long enough to develop significant end-stage liver disease such as cirrhosis or hepatocellular carcinoma (HCC). Before these more dour events, however, most patients are asymptomatic so their lifestyle is generally unaffected, although their health needs may be compromised. Vamvakas31 recently reviewed long-term survival after blood transfusion in the 1990s and reported 5- and 10-year probabilities of survival of 47 and 40 percent, respectively. Survival was found to depend on a number of factors, however, such as volume and type of blood administered, reasons for transfusion (surgical or nonsurgical), age and sex of the recipient, and whether the recipient was immunocompetent or immunocompromised. In contrast to this analysis, Seeff and coworkers32 found survival in uninfected transfused recipients enrolled in a multicenter, collaborative non-A, non-B (i.e., HCV) hepatitis study to be more than 80 percent at 5 years posttransfusion and to have declined to only 35 to 45 percent after 25 years. The application of a number of recruitment exclusions before anticipated transfusion may have biased the overall posttransfusion outcome resulting in lower mortality rates in this population. For the purpose of this analysis of outcome, therefore, we will assume a minimum median survival rate of 50 percent after 5 years of follow-up, 40 percent after 10 years, and 25 percent after 25 years.

Based on an estimated 4 percent rate of chronicity that can occur among newly infected, immunocompetent recipients (range, 3%-5%), it is predicted that approximately 3 cases of chronic hepatitis B will develop each year from the original infected cohort of 74 recipients. (Among immunocompromised recipients, the rate of chronicity would be substantially higher but this would be offset by an expected lower rate of long-term survival.) Taking into account the annual increase of 3 cases of chronic hepatitis B and factoring in the survival rates over time, approximately 10 chronically infected recipients might survive for 15 to 25 years. Of these, approximately 25 percent (2.5 recipients) will have developed cirrhosis or HCC within that time interval.32,33 This analysis does not take into consideration current therapy or future breakthroughs in the management of hepatitis B that will clearly delay or eliminate the progression of disease in up to 40 percent of these individuals, nor does it consider liver transplantation as an option. However, significant costs are associated with each of these strategies. To this low rate of complications from chronic HBV should be added the 3 to 4 deaths from fulminant hepatitis that are predicted to occur in that 25-year time span since such events are immediate. It is unlikely that any of these fulminant hepatitis episodes could be prevented even with a successful surveillance program that identifies seroconverting donors, although some of these cases could be rescued by liver transplantation. Regardless, looking for other recipients who were administered blood components from a donor implicated in transmission of HBV to a recipient developing fulminant hepatitis should be undertaken. In each of these scenarios, ID-NAT remains the most sensitive approach to prevention of transmission of HBV and fulminant hepatitis.

Table 2 places these estimates into their proper perspective. In addition to morbidity and mortality from acute or fulminant hepatitis or from the development of end-stage liver disease, the risk related to the potential for transmitting HBV to others via the percutaneous or permucosal routes of transmission after the development of acute or chronic hepatitis B in a recipient must be considered. The impact of current routine vaccination programs, however, will mitigate some of the risk among recipients and their contacts. As discussed, most of these acute or chronically infected recipients will have inapparent or subclinical infections and will remain unrecognized for an indefinite time interval. The potential to transmit HBV from an often unrecognized, but infected, recipient to a susceptible contact remains an unknown entity as the number of chronically infected recipients gradually expands from approximately 3 cases the first year to 32 persistently infected survivors 25 years later. These estimates assume a 10 percent loss of infected recipients per year due to death for the first 5 years, a 2 percent loss per year for the next 5 years, and a 1 percent loss per year thereafter. In addition to the cirrhosis/HCC and fulminant hepatitis concern, this is the issue that merits the most attention since prevention of hepatitis B by vaccination of contacts in addition to education of the cases can be effective deterrents of disease transmission to others and treatment can be considered.

Table 2. Estimated number of posttransfusion hepatitis B events occurring annually and over 25 years
EventAnnualOver 25 years
  • * 

    Based on survival statistics for a transfused cohort followed over 25 years (see discussion in text).

  • † 

    Based on number of chronically infected recipients who survive 15 to 25 years and who are estimated to develop cirrhosis or HCC during that time period.

Acute symptomatic hepatitis B (hospitalized)21 (15)525 (375)
Chronic hepatitis331.5*
Death from fulminant hepatitis0.153.75
Cirrhosis/HCC 2.5


Having concluded that a quantifiable risk of transmission of HBV exists among transfused recipients, what published documentation do we have that traceback or lookback might provide a meaningful public health benefit? In one situation, the investigation of acute hepatitis B in a patient might implicate a donor who is subsequently found to have underlying liver disease. Removal of this donor from the donor pool could prevent disease from developing in other recipients although in almost all cases the next donation would likely be HBV DNA–reactive and/or -seropositive. In addition, it may be established that the donor is in the early stages of acute hepatitis B34-36 and therefore likely to transmit HBV to others or that the donor has underlying liver disease that could benefit from treatment to prevent progression to end-stage liver disease or the development of HCC.37,38 Should the donor be a person who abuses parenteral drugs or be sexually active or of child-bearing age, knowledge of their HBV infection could be instrumental in preventing transmission to other unsuspecting contacts. For this effort to be successful, however, it entails educating health care professionals to ask the appropriate questions of their HBV-infected referrals to include defined social, sexual, and risk histories when assessing the source of infection.

Once a potentially infected donor has been identified, lookback can be initiated based on the criteria that have been outlined above. One of the reasons for lookback is because cases of posttransfusion hepatitis B in recipients are often not identified. Indeed, in one study,37 three of four cases of acute hepatitis B and two suspected seroconversions went unrecognized by local physicians as being related to recent transfusions. Soldan and coworkers35 were able to identify up to 20 cases of hepatitis B associated with presumably infectious donors over the past 15 years, none of whom were recognized by local physicians in a timely fashion.

Perhaps the best reason for advocating lookback among recipients of blood or blood components are the results obtained from a systematic lookback study conducted by the Japanese Red Cross from 1997 to 2004.5,39 In this report, they evaluated 15,721 archived donations from repeat donors before their becoming positive for the presence of HBsAg, anti-HBc, or HBV DNA in a 50-minipool NAT for the first time. Among these stored samples, they found 158 donors that were HBV DNA–positive (± low-titer anti-HBc) in an ID-NAT that had a lower limit of detection of 100 copies per mL. Even in this well-designed study, however, only 58 percent of the recipients who had been transfused with ID-NAT–positive blood elected to be evaluated for infectivity, and 104 of 181 recipients (57%) had already expired without having been tested. This left 77 living HBV-susceptible recipients of this potentially infectious blood, 12 (16%) of whom were found to be infected with HBV after lookback, with 9 of these still HBV DNA–positive and presumably capable of transmitting their disease to others. Seven additional recipients without a pretransfusion blood sample were positive for the presence of anti-HBc and/or anti-HBs. Eleven of the 12 HBV infections (92%) were associated with a blood donation that was HBV DNA–positive, anti-HBc–negative (preseroconversion window period). In 51 other recipients of ID-NAT–positive blood, no HBV DNA or HBsAg was detected but knowledge of their pretransfusion anti-HBc or anti-HBs status confounds the issue of susceptibility. In addition, the role of passively transfused anti-HBs in concurrently administered blood components in protecting these recipients was not assessed. Among those donors whose blood was found to be HBV DNA–positive for the first time by minipool NAT, the probability of detecting HBV DNA in a previous donation by ID-NAT was 8.5 percent. If HBsAg was the first serologic test to become reactive, HBV DNA was detected in only 2.2 percent of prior donations whereas the rate was 0.78 percent among those archived samples examined from donors who subsequently were found to be anti-HBc seroconverters.

In contrast to these persuasive studies, the potential benefits and resource requirements of HBV lookback demands a comparison with those of other programs. Lookback for HIV and HCV in the United States clearly identified a large number of infected blood recipients after the initiation of donor testing, but its value has declined significantly since that time. Even 8 to 10 years ago, analyses suggested that little tangible benefit accrued once the initial group of prevalent cases had been identified.4,40,41 Although the four known cases of transfusion-transmitted HIV that occurred since the initiation of NAT were all identified as a result of lookback, there have been no cases of transfusion-transmitted HIV identified by lookback since 2002 and no HCV cases since 1999 (S. Stramer, personal communication, 2008). An excellent argument for doing HBV lookback in the early 1970s could have been made, but current yields are expected to be low as previously discussed. Certainly, the tangible benefits of HBV lookback are unlikely to meet the levels suggested above as a result of the imperfect and incomplete nature of lookback itself. For example, infections due to one-time donors or those donors who do not return after a window-period donation will not be identified. Donations in different blood centers may not be located, and not every donation will be traceable to the recipient(s). The magnitude of these effects is not clear, but is likely to reduce the potential total by at least 25 percent. For hepatitis C, the efficacy of previous lookback efforts relating to anti-HCV seroconversion has been relatively low in recent years, although 6 percent or more of living, notified, and tested recipients who received a transfusion after 1992 were newly identified as anti-HCV–positive.4

HBV lookback involves relatively little effort when HBsAg or HBV DNA is newly found in a repeat donor and, at a minimum, could be extended to the entire blood banking community. In contrast, the use of anti-HBc seroconversion as a lookback trigger is problematic, in part because of the historically poor specificity of the EIA anti-HBc assay and because of the sheer magnitude of the effort required. However, newer test modalities, such as the Abbott PRISM HBcore chemiluminescent assay have improved specificity (e.g., 99.88%) and sensitivity (e.g., 99.49%) (Abbott product insert) and this aspect may no longer be a major issue. Validation with another anti-HBc assay, cooccurrence with other HBV seromarkers, or a strong signal can further corroborate the results. Lookback for de novo HBsAg and HBV DNA seroconversions (in the absence of anti-HBc) would add about 50 to 55 cases for investigation in the ARC system (approx. 100-110 cases estimated for all blood collection agencies) and these would require lookback only for 6 months. This effort is certainly achievable. The annual number of verifiable repeatedly reactive anti-HBc seroconversions among repeat donors in the ARC system is not precisely known, but could be on the order of 5000. Since incident cases of specific anti-HBc seroconversion also imply the prior occurrence of an HBV infection, all of these would have to be evaluated in some way, and many of these (perhaps half or more) would require lookback to 6 months before the last seronegative donation with the proposition that a 5-year time limit might be considered. This would involve a large and continuing incremental effort. It is unclear how effective such an effort would be, but it would certainly not identify all of the potentially infected recipients. Because lookback relates to prior donations, those recipients that were identified would probably not be notified until after the occurrence of clinically apparent acute disease, were it to occur, so the actual benefit would be confined to those chronically infected recipients who would develop severe chronic disease and/or transmit secondary infection—estimated to represent a relatively small number each year in the United States. One alternative approach that might be more operable and acceptable would be to limit HBV lookback only to those anti-HBc–reactive donors who also are IgM anti-HBc–positive and to limit the investigation to the 6 months before the last transfusible donation over the prior 2 years.


In the United States, there is an existing standard of practice and ethical position favoring lookback for HIV and HCV, which was adopted after implementation of tests for these infections. Unequivocally, lookback for HBV will identify some individuals infected through transfusion as evidenced by the Japanese studies described above; thus, there is a precedent to justify a program for HBV lookback. In addition, blood centers already have an obligation to recover in-date products when a subsequent seroconversion is identified in that donor.42 This suggests that the FDA already places some credence to the possibility that such blood has a potential to be infectious.

The major benefit of existing HIV and HCV lookback programs has been the identification of recipients of blood from donors who were already infected before testing was implemented. This benefit is largely absent for HBV because sensitive testing of donors for this agent has been in place for almost 35 years. Nevertheless, the incidence of new HBV infections among repeat donors to the ARC has remained relatively constant (range, 1.267-1.974 per 100,000 person-years).19 Relative to the resources expended, analysis of available data shows that the actual yield of lookback in terms of identifying new infection in recipients will be low. However, this should be offset by the value of new treatment modalities and the availability of a safe and effective vaccine that can prevent transmission of the disease to others.

Is the ethical posture as simple as it appears? Certainly, lookback assures autonomy for the potentially exposed recipient and the absence of lookback will deny an individual the knowledge of that potential exposure and the opportunity to obtain a test and perhaps treatment. From that perspective, denying lookback could be seen as maleficence inasmuch as it could result in some number of instances of preventable disease, although perhaps at the cost of giving drugs to those who may never get sick. Other ethical principles are beneficence—production of benefits that outweigh risks, and justice—achievement of an equitable distribution of risks and benefits. So the question is the extent to which risks and resource usage accrue from HBV lookback. As with other lookback programs, the direct risk of notification is low and is almost entirely based upon the generation of some level of concern and anxiety among those recipients who are notified, but who will not have been infected—certainly a significant majority of those notified.39,40 It is really not possible to determine whether, in aggregate, this offsets the benefits, but the point is that there are additional perspectives that need to be considered when making the decision.

A broader ethical question, however, is whether a lookback program would actually shift risks from one group to another. Aspects of commentary about the invocation of the precautionary principle are worthy of discussion in this context. Such commentary indicates that measures that are taken should not exceed those used in other known circumstances and that it may also be appropriate to consider cost-benefit considerations.43 There is little question that lookback for HBV would have been equivalent to that for other programs if it had been introduced at the time of initial testing for HBV. Clearly, however, implementing a full-scale HBV program at this time will not have an equivalent level of benefit although the data do indicate that some cases will be identified or possibly prevented as a result of this process. The amount of effort involved, however, will be significant, most particularly for anti-HBc seroconversions which may be mitigated to some extent by limiting the investigation to those anti-HBc positive donors who also are IgM anti-HBc positive.

Success of a lookback policy will depend on how the leadership of the various community blood collection centers, hospital transfusion services, and national blood organizations views their responsibility to the recipients of transfused products. To be successful, they will need to be more aggressive, conscientious, and persuasive in their approach to enrolling recipients in the lookback program and to monitoring donors. Clearly there are many roadblocks to this process including restricted access to protected health information, mortality after transfusion, and numerous layers of intermediaries that must be bridged before a recipient is identified. Much of this process depends on the level of participation of the recipient's physician since he or she will ultimately decide whether to notify the recipients and to determine what laboratory tests or referrals are needed. However, we believe that once the donor and recipient are identified, the initial contact with the patient or his or her physician should be assigned to a committed hepatologist or other expert in HBV disease and not to ancillary health care professionals. After enrollment with further testing by the collection agency, follow-up should be the responsibility of the recipient and their physician. This approach should improve the relatively ineffective results that often have accompanied other lookback programs.

In summary, traceback and lookback for HBV would certainly benefit some donors and blood recipients and could have the potential to interdict secondary infections. Many might assert that these are sufficient grounds to recommend implementation of the practice despite the limited impact that it would have on public health. However, because of the long-standing policy of testing blood donors for HBV markers, the question remains whether the benefits would justify the necessary investment of resources. This is the quandary facing the blood banking community as it ponders this topic.