Risk of HIV & Hepatitis Infections From Blood & Tissue Donors
SUMMARIES OF TWO STUDIES REPORTED IN DETAIL FOLLOWING SUMMARIES
Undetected Viremia among Tissue Donors
In this study, on the basis of data on 11,391 tissue donors, the probability of undetected viremia with the human immunodeficiency virus, hepatitis B virus, hepatitis C virus, and human T-lymphotropic virus was estimated to be 1 in 55,000 donors, 1 in 34,000, 1 in 42,000, and 1 in 128,000, respectively.
The prevalence of these four viral infections is higher among tissue donors than among first-time blood donors. The use of more sensitive screening methods can reduce the risks of infection among tissue recipients.
Editorial AT END OF THIS REPORT Detection of HIV-1 and HCV RNA among Antibody-Negative Blood Donors
Since 1999, nucleic acid--amplification testing has been used in the United States to identify units of blood from donors with viremia in the window period before seroconversion. This approach identifies approximately 1 unit infected with human immunodeficiency virus type 1 (HIV-1) among 3.1 million units screened and 1 infected with hepatitis C virus (HCV) among 230,000 units screened.
The addition of nucleic acid--amplification testing to blood-donor screening prevents about 5 cases of transfusion-transmitted HIV-1 infection and 56 of HCV infection per year, resulting in a residual risk of these infections of approximately 1 per 2 million units of blood transfused.
Detection of HIV-1 and HCV Infections among Antibody-Negative Blood Donors by Nucleic Acid--Amplification Testing
New England Jnl of Medicine August 19, 2004, Volume 351, Number 8
Susan L. Stramer, Ph.D., Simone A. Glynn, M.D., M.P.H., Steven H. Kleinman, M.D., D. Michael Strong, Ph.D., Sally Caglioti, M.T. (A.S.C.P.), M.T. (A.S.C.P.), S.B.B., David J. Wright, Ph.D., Roger Y. Dodd, Ph.D., Michael P. Busch, M.D., Ph.D., for the National Heart, Lung, and Blood Institute Nucleic Acid Test Study Group
Screening of potential blood donors has historically relied on the use of immunoassays to detect viral antibodies or antigens. In 1999, new screening methods involving nucleic acid amplification to detect human immunodeficiency virus type 1 (HIV-1) and hepatitis C virus (HCV) RNA were implemented in the United States under an investigational new drug protocol approved by the Food and Drug Administration (FDA). This new technique was used to test multiple samples in small pools, referred to as "minipools." The decision to implement this technique was based on its ability to identify HIV-1-- and HCV-infected donors early in the infectious window period, before seroconversion, and the experience of plasma-derivative manufacturers showing the practicality of this approach for pooled specimens.1 Finally, it was recognized that the availability of nucleic acid--based tests would support future testing of emerging agents.
The advent of nucleic acid--amplification testing has led to the discontinuation of two less effective screening tests. HIV-1 p24 antigen screening was recommended by the FDA in 1996 for the early detection of HIV-1 infection, and the FDA allowed this approach to be discontinued on the licensure of the HIV-1 nucleic acid--amplification test. Elevated levels of alanine aminotransferase have been used as a surrogate (nonspecific) marker for HCV infection since 1986. The use of this screening approach was never an FDA requirement, so blood centers have voluntarily discontinued this test. RNA-based donor screening has afforded an opportunity to study events occurring early in HIV-1 and HCV infection. To quantify the relative risk of transmission of HIV-1 and HCV from first-time blood donors and those who donated blood repeatedly, we analyzed the number of RNA-positive, antibody-nonreactive allogeneic blood donations from donors infected with HIV-1, HCV, or both that were identified in the first three years after the implementation of nucleic acid--amplification testing in the United States.
Testing of blood donors for human immunodeficiency virus type 1 (HIV-1) and hepatitis C virus (HCV) RNA by means of nucleic acid amplification was introduced in the United States as an investigational screening test in mid-1999 to identify donations made during the window period before seroconversion.
We analyzed all antibody-nonreactive donations that were confirmed to be positive for HIV-1 and HCV RNA on nucleic acid--amplification testing of "minipools" (pools of 16 to 24 donations) by the main blood-collection programs in the United States during the first three years of nucleic acid screening.
Among 37,164,054 units screened, 12 were confirmed to be positive for HIV-1 RNA — or 1 in 3.1 million donations — only 2 of which were detected by HIV-1 p24 antigen testing. For HCV, of 39,721,404 units screened, 170 were confirmed to be positive for HCV RNA, or 1 in 230,000 donations (or 1 in 270,000 on the basis of 139 donations confirmed to be positive for HCV RNA with the use of a more sensitive HCV-antibody test). The respective rates of positive HCV and HIV-1 nucleic acid--amplification tests were 3.3 and 4.1 times as high among first-time donors as among donors who gave blood repeatedly. Follow-up studies of 67 HCV RNA--positive donors demonstrated that seroconversion occurred a median of 35 days after the index donation, followed by a low rate of resolution of viremia; three cases of long-term immunologically silent HCV infection were documented.
Minipool nucleic acid--amplification testing has helped prevent the transmission of approximately 5 HIV-1 infections and 56 HCV infections annually and has reduced the residual risk of transfusion-transmitted HIV-1 and HCV to approximately 1 in 2 million blood units.
Since 1999, allogeneic blood donations in the United States have been screened for HIV-1 and HCV RNA in a minipool format with the use of one of two nucleic acid--amplification tests. The Gen-Probe Transcription-Mediated Amplification system uses a multiplex HIV-1 and HCV assay and minipools of 16 donor samples. All donation samples within a reactive minipool are tested individually to identify both the sample that was reactive and the viral cause of the reaction. The Roche Molecular Systems Cobas AmpliScreen HIV-1 and HCV tests separately detect HIV-1 and HCV RNA in minipools of 24 donor samples. Both assays are highly specific and sensitive, with 50 percent detection limits (i.e., the level at which 50 percent of test results would be expected to be reactive) of 14 or fewer copies of HIV-1 per milliliter and 12 or fewer copies of HCV per milliliter on the basis of probit analyses. The 95 percent detection limits as defined in the package inserts for both tests range from 30 to 60 copies per milliliter for HIV-1 and HCV. Both systems have received FDA approval for routine screening of blood donors.
All major laboratories in the United States participating in nucleic acid--amplification screening (accounting for over 98 percent of tested blood donations) participated in this study and reported data collected on cases identified between March 1999 and January 2002, and in some instances from March 1999 through April 2002. A case was defined as an allogeneic donation that was nonreactive to antibody against HIV-1, HCV, or both but that was reactive on minipool nucleic acid--amplification screening and confirmed to be positive for HIV-1 or HCV RNA. Five testing programs used the Gen-Probe assay and reported cases of HIV-1 and HCV viremia identified on screening of 27,956,758 donations. The Roche Cobas AmpliScreen was used in 13 laboratories, which tested a total of 9,207,296 donations for HIV-1 RNA and 11,764,646 donations for HCV RNA. All participating sites received approval of this study from their institutional review board. Data were contributed by the blood-collection organizations and the manufacturers of the nucleic acid assays (Roche Molecular Systems, Gen-Probe, and Chiron).
The date of donation, the donor's status as a first-time or repeat donor, and whether the unit would have qualified for transfusion if not for the result of the nucleic acid--amplification test (i.e., whether the unit was transfusable) were collected for each case. Furthermore, the results of HIV-1 p24 antigen testing were compiled for cases of HIV-1 viremia, whereas data on alanine aminotransferase levels and the presence or absence of antibody against hepatitis B core antigen (anti-HBc) were collected for cases of HCV viremia. When applicable, we also compiled the results of repeated serologic analyses, repeated nucleic acid--amplification testing of the index sample with the use of a different type of RNA method (e.g., different techniques, primers, or probes), nucleic acid--amplification testing of an independent sample from the index donation, and serologic and nucleic acid--amplification testing of samples collected from donors participating in the follow-up analysis. For HIV-1, antibody was detected with the use of enzyme immunoassays and confirmed by Western blotting; for HCV, antibodies were detected by either second- or third-generation enzyme immunoassays and confirmed by recombinant immunoblot assay (RIBA, Chiron). Laboratories that routinely used second-generation HCV-antibody tests to screen donations were also asked to report the results of third-generation HCV-antibody tests performed on the HCV RNA--positive donations. This allowed categorization of cases of HCV viremia into those in which antibodies were detectable only by the more sensitive third-generation test and those with no detectable HCV antibody on both second- and third-generation HCV-antibody tests. A case was considered confirmed if the index donation was reactive to HIV-1 or HCV RNA with the use of a second type of nucleic acid--amplification test, if another sample from the index donation was reactive on the nucleic acid assay, or if at least one follow-up sample was reactive on nucleic acid--amplification testing or antibody testing.
An expanded data set was developed by the largest participating program (the American Red Cross) to study the dynamics of HCV infection. This data set included follow-up of HCV RNA--positive donors identified from March 1999 through mid-June 2003, thus providing an additional 15 months of follow-up on a well-characterized group of donors with acute HCV infection. For this program, a standardized prospective protocol was used to enroll donors, with specimens collected at approximately four-week intervals through the time of seroconversion, as confirmed by third-generation HCV-antibody tests, and beyond.
To determine whether trends observed for HIV-1--positive and HCV RNA--positive donors were constant beyond this three-year study, an additional two years of data from the American Red Cross were analyzed. Data on HIV-1--positive and HCV RNA--positive donors from March 1999 through March 2002 were compared with those for the subsequent two-year period from April 2002 to April 2004.
To evaluate rates of positive nucleic acid--amplification tests for HIV-1 and HCV RNA in specific subgroups (first-time and repeat donors and donors with otherwise transfusable donations), data were included only from laboratories that routinely reported this information. These subgroups represented about 37.0 million of the 39.7 million total donations. On the basis of data from the American Red Cross for 1999 through 2002, it was estimated that 23 percent of allogeneic donations were collected from first-time donors and 1 percent of all donations were discarded owing to reactivity to another routine serologic screening test in addition to nucleic acid--amplification testing.
Rates of positive nucleic acid--amplification tests per million donations were calculated by dividing the number of cases by the number of known donations screened (or for samples from first-time or repeat donors and samples that were otherwise transfusable, by the estimated number of donations) and multiplying by 106. When the number of donations was known, the associated 95 percent confidence interval for the rate was computed. When the number in a subgroup of donations was estimated, an approximate 95 percent confidence interval was computed incorporating the uncertainty around the estimated number of donations. Fisher's exact tests and Wilcoxon's tests were used to compare categorical and continuous variables, respectively. All reported P values are two-sided.
Assuming that each of the 13.6 million allogeneic units of blood donated annually in the United States is converted on average to 1.45 transfusable components, our data indicate that the implementation of minipool nucleic acid screening likely prevented about 5 cases of transfusion-transmitted HIV-1 infection and 56 cases of HCV infection annually. The documented findings are consistent with the those predicted from mathematical models. Despite the fact that these rates are relatively low and have remained stable for five years, implementation of these tests was consistent with the goal of maximizing blood safety. It has been estimated that nucleic acid screening has reduced the residual risk of transfusion-associated infection for both HIV-1 and HCV to about 1 in 2 million blood units from repeated donors. This is a reduction from rates of 1 in 276,000 for HCV and 1 in 1.5 million for HIV-1 with the use of serologic testing alone. The residual risk after the implementation of nucleic acid--amplification testing results from the presence of virus below the limit of detection of minipool testing; individual nucleic acid screening of each sample, rather than screening of small pools of multiple samples, would further decrease the residual risk but at a substantially greater cost.
With the licensure of nucleic acid--amplification tests, the FDA has permitted the discontinuation of HIV-1 p24 antigen testing on the basis of data showing that HIV-1 RNA screening is better able to detect infection in the window period shortly after infection and that all p24 antigen--positive donations are also RNA-positive. This policy is supported by our data, in which HIV-1 nucleic acid screening identified 12 infected donors, only 2 of whom were identified by p24 antigen testing; in contrast, there were no RNA-negative donations from HIV-1--infected donors that were identified as positive by p24 antigen testing. The detection of p24 antigen in the absence of antibody corresponds to the peak viremic period when blood donors are likely to defer donations owing to influenza-like symptoms.
Approximately one third of the units detected by HCV nucleic acid--amplification testing would have been discarded anyway owing to elevated alanine aminotransferase levels. Because of the relative nonspecificity of this surrogate marker, the absence of evidence of additional transfusion-transmissible hepatitis agents, and the implementation of a sensitive screening method for the detection of HCV RNA, the continued use of alanine aminotransferase screening for preventing transfusion-associated hepatitis is no longer justified; consequently, many blood centers have stopped using this test. In addition, the presence of circulating HCV RNA is a direct marker of viral replication and indicates a diagnosis of HCV infection with greater sensitivity and specificity than does the presence of elevated liver enzymes.
Our data show that new HCV and HIV-1 infections occur three to four times as often among first-time donors as among repeat donors, substantiating previous observations. This finding supports the general principle that retention of repeat donors enhances both the adequacy and safety of the blood supply. Possible reasons for higher rates among first-time donors include inappropriate use of blood donation to obtain the results of viral tests; failure to understand the questions for donors and, hence, the donor-selection criteria; and self-deferral of the donor after the first donation owing to the realization that his or her donation was unsuitable.
The routine use of nucleic acid--amplification tests and serologic assays for donor screening has made possible the identification of persons in the very early stages of HIV-1 and HCV infection; this information can provide insights into risk factors associated with viral infection and potentially contribute to studies of the natural history, pathogenesis, and treatment of these infections. For example, an analysis of recent risk-related behavior among HCV-infected donors identified by nucleic acid--amplification testing may identify behavioral and demographic characteristics that could be used to improve donor-qualification criteria, provided effective questions could be designed. The addition of HCV RNA testing to routine HCV-antibody screening has also allowed seropositive donors to be subdivided into those with active infection (plasma RNA--positive) and those with either resolved HCV infection or intermittent viremia (plasma RNA--negative at the time of donation). Enrollment of these donors into natural-history and early-treatment trials could enhance our understanding of the pathogenesis of HCV infection, including the factors underlying the spontaneous resolution of HCV viremia.
Several reports have suggested that serologic testing may miss a substantial proportion of infected persons. We found that only three seronegative donors with persistent hepatitis C viremia did not seroconvert during the expected time frame. During this same time at the American Red Cross, more than 800 HIV-seropositive donors and more than 16,000 HCV-seropositive donors were identified. Thus, persistent immunologically silent infections are extremely rare, reinforcing the continued reliance on serologic analyses for HIV-1 and HCV as the primary tools for diagnostic testing.
Because blood centers had already implemented nucleic acid--amplification testing for HIV-1 and HCV, it was feasible in 2003, in collaboration with the Centers for Disease Control and Prevention and the FDA and with the rapid development of nucleic acid--amplification tests by manufacturers, to implement screening for West Nile virus in less than nine months. Results indicate that close to 1000 donors with West Nile virus infection were identified by nucleic acid--amplification testing in 2003 and their donations discarded, probably preventing more than 1000 transfusion-related infections.
The relatively low yield and poor cost effectiveness of HIV-1 and HCV minipool nucleic acid--amplification testing have led some to question the value of such screening. Using somewhat different analyses and assumptions, two independent groups studying the cost-effectiveness of HIV-1 and HCV minipool nucleic acid--amplification testing, both in the context of eliminating p24 antigen screening, estimated costs of $1.5 million to $4.3 million per quality-adjusted year of life. Costs increase further if each donated blood unit is to be tested rather than combined in minipools, with yet further increases in cost for the automation required to perform large numbers of individual screening tests. Therefore, the cost of HIV-1 and HCV nucleic acid--amplification testing would need to decrease substantially to bring it in line with that of most other accepted medical practices. However, the aggregate cost-effectiveness of nucleic acid--amplification testing may have substantially improved with the implementation of such screening for West Nile virus. The rapid development and introduction of nucleic acid screening for West Nile virus and the ability to expand nucleic acid--amplification testing to include other emerging infections in the future further serve to support the adoption of this important tool for the screening of blood donations.
Supported by the individual blood programs represented as well as by contracts (N01-HB-97077 [superseded by N01-HB-47114], N01-HB-97078, N01-HB-97079, N01-HB-97080, N01-HB-97081, and N01-HB-97082) with the National Heart, Lung, and Blood Institute.
Probability of Viremia with HBV, HCV, HIV, and HTLV among Tissue Donors in the United States
New England Jnl of Medicine Volume 351, Number 8
Shimian Zou, Ph.D., Roger Y. Dodd, Ph.D., Susan L. Stramer, Ph.D., D. Michael Strong, Ph.D., for the Tissue Safety Study Group
Tissue-banking organizations in the United States have introduced various review and testing procedures to reduce the risk of the transmission of viral infections from tissue grafts. We estimated the current probability of undetected viremia with hepatitis B virus (HBV), hepatitis C virus (HCV), human immunodeficiency virus (HIV), and human T-lymphotropic virus (HTLV) among tissue donors.
Rates of prevalence of hepatitis B surface antigen (HBsAg) and antibodies against HIV (anti-HIV), HCV (anti-HCV), and HTLV (anti-HTLV) were determined among 11,391 donors to five tissue banks in the United States. The data were compared with those of first-time blood donors in order to generate estimated incidence rates among tissue donors. The probability of viremia undetected by screening at the time of tissue donation was estimated on the basis of the incidence estimates and the window periods for these infections.
The prevalence of confirmed positive tests among tissue donors was 0.093 percent for anti-HIV, 0.229 percent for HBsAg, 1.091 percent for anti-HCV, and 0.068 percent for anti-HTLV. The incidence rates were estimated to be 30.118, 18.325, 12.380, and 5.586 per 100,000 person-years, respectively. The estimated probability of viremia at the time of donation was 1 in 55,000, 1 in 34,000, 1 in 42,000, and 1 in 128,000, respectively.
The prevalence rates of HBV, HCV, HIV, and HTLV infections are lower among tissue donors than in the general population. However, the estimated probability of undetected viremia at the time of tissue donation is higher among tissue donors than among first-time blood donors. The addition of nucleic acid--amplification testing to the screening of tissue donors should reduce the risk of these infections among recipients of donated tissues.
Hepatitis B virus (HBV), hepatitis C virus (HCV), human immunodeficiency virus (HIV), and human T-lymphotropic virus (HTLV) have all been transmitted by tissue transplantation.These viruses have also been transmitted by blood transfusion, almost always as a result of the collection of blood during the so-called viremic window period, before infection can be detected by laboratory testing. The probability of collecting blood during this window period has been extensively evaluated. However, similar estimates have not been made for tissue donors, even though such estimates would be helpful in evaluating the efficiency of current and future measures designed to ensure the safety of tissue transplantation.
Tissue banks in the United States obtain, process, and distribute a variety of tissues, including heart valves, venous tissue, bone, bone-derived products (such as powders used for dental work), and connective tissue. The vast majority of these tissues come from cadavers, and all are essentially avascular and can be stored for long periods. Although tissue donors may also provide organs for transplantation, the converse is not necessarily true. The infectivity of different tissues varies, in part as a reflection of their anatomical origin and nature, but also as a result of processing after collection. For example, a highly processed bone powder would be much less likely to transmit a viral infection than would a fresh-frozen bone segment. Currently, the measures used to assess tissue donors include a retrospective review of the donor's medical history and testing of cadaveric blood samples for hepatitis B surface antigen (HBsAg) and antibodies against HIV (anti-HIV), HCV (anti-HCV), and HTLV (anti-HTLV).
We estimated the probability of viremia at the time of tissue donation by using the incidence--window-period model developed to estimate the residual risk of viremia among blood donors. In order to do this, we estimated the incidence rates of HIV, HBV, HCV, and HTLV infection on the basis of measured prevalence rates among tissue donors and available data from other sources. Information on the duration of the window periods of viremia, before seroconversion, for these infections was obtained from the peer-reviewed literature.
Determination of Prevalence Rates among Tissue Donors
Data on the prevalence of anti-HIV, HBsAg, anti-HCV, and anti-HTLV in tissue donors were obtained from existing databases of the Northwest Tissue Center (for 2001 through 2002), the American Red Cross Tissue Services (for 2000 through 2002), the Musculoskeletal Transplant Foundation (for 2002), the Community Blood Center/Community Tissue Services (for 2001), and LifeNet (for 2002). The data did not include any donor identifiers. During the periods covered, all five centers followed the review and testing standards of the American Association of Tissue Banks. Four of the centers reported confirmed positive results; one reported only the results of the screening tests. For donors at this center, we estimated the rates of confirmed positive results by subtracting the number of false positive results (determined on the basis of specificity analyses of data from the other sites) from the number of reactive screening results. Pooled data were used to determine age- and sex-specific prevalence rates for the markers; prevalence was defined as the number of donors with confirmed positive tests divided by the total number of donors tested.
Estimation of Incidence Rates among Tissue Donors
The incidence rate of new infections among tissue donors was estimated by applying age- and sex-specific incidence rates for first-time blood donors to the tissue-donor population. Prevalence and incidence rates among voluntary donors and donors of directed whole blood were obtained from a research database of blood donors to the American Red Cross Blood Services. Incidence was defined as the number of donors who seroconverted per 100,000 person-years among a group who repeatedly donated blood. Dodd et al. and Janssen et al. reported incidence ratios among first-time donors as compared with those who made repeated donations of 2.42 for HCV infection and 2.43 for HIV infection. No such data were available for HBV and HTLV infections. On the basis of the ratios for HIV and HCV, a ratio of 2.5 was assumed for HBV and HTLV. The ratios were applied to the incidence rates among persons who donate blood repeatedly to estimate incidence rates for first-time blood donations. These incidence rates were adjusted to reflect the difference in prevalence rates between blood and tissue donors by multiplying by the ratios of prevalence rates in the two groups. Prevalence and incidence rates for corresponding groups in the general population were also obtained through a search for published epidemiologic data and unpublished data from the Centers for Disease Control and Prevention (CDC) (Alter M: personal communication) and were used in a similar manner to derive alternative estimates of incidence rates among tissue donors.
Estimation of the Probability of Viremia
We estimated the risk of infectivity — the probability that any tissue donor was in the viremic window period with an infection that was undetected by means of serologic screening methods at the time of donation — by the method developed by Petersen et al., Busch et al., Lackritz et al., and Schreiber et al. The estimated probability is obtained from the product of the incidence rate and the length of the window period for each infection.
Unless otherwise specified, frequencies were compared with the use of the chi-square test; all reported P values are two-sided. Possible ranges of the estimated risks of infectivity resulting from the collection of tissues during the window periods for these infections were determined by means of Monte Carlo simulation with the use of Crystal Ball software. Basically, possible variations in the prevalence rates among tissue donors and first-time blood donors according to sex and age, incidence rates among those who repeatedly donated blood according to sex and age, overall prevalence and incidence estimates and their assumed sex- and age-based distributions in the general population, incidence ratios for first-time donors as compared with those who repeatedly donated blood, and window periods were incorporated into the incidence- and risk-determination models to derive the 2.5 and 97.5 percentiles of the risk estimates. For prevalence and incidence rates, 95 percent confidence intervals were incorporated into all models except for those for the prevalence of HIV and for the incidence of the three markers in the general population; these models used a 50 percent variation owing to the lack of data on confidence intervals. A variation of 50 percent was also applied to the incidence ratios for HIV, HBV, HCV, and HTLV infections between first-time donors as compared with those who repeatedly donated on the basis of the variations in the incidence rates for HIV and HCV. All the ratios were assumed to follow triangular distributions. The window periods were assumed to follow triangular distributions with different degrees of variation, as reported by Schreiber et al.
To check the estimated frequency of confirmed positive results among the unconfirmed reactive results from a single center, we used a recombinant immunoblot assay (RIBA 3.0 SIA test, Chiron) to test 50 serum samples obtained post mortem that were initially reactive for anti-HCV. Thirty-six (72 percent) were positive, seven (14 percent) were indeterminate, and seven (14 percent) were negative. Similarly, we used Western blotting (HIV Western Blot Kit, Cambridge Biotech) to test nine serum samples that were initially reactive for HIV. Seven were negative, and two were indeterminate. Among tissue donors from other tissue centers, 74 percent of samples that were reactive to anti-HCV on initial screening were confirmed to be positive (81 of 110) and 11 percent of samples that were reactive to anti-HIV on initial screening were confirmed to be positive (2 of 19). The differences between these values and values found by evaluation testing were not significant (2=0.006, P=0.98 for anti-HCV and P=1.00 for anti-HIV by Fisher's exact test), indicating that the approach used to extrapolate the rates of confirmed positive results was appropriate.
Prevalence and Incidence of Viral Infections in the General Population
According to U.S. data from the CDC (and from the AIDS [Acquired Immunodeficiency Syndrome] Public Information Data Set at www.cdc.gov/hiv/software/apids.htm), the current prevalence of HIV infection (excluding AIDS) is approximately 0.20 percent. The incidence of HIV infection is estimated to be 40,000 cases per year, with approximately 70 percent of cases in males and 30 percent in females; the age distribution of incident HIV infections is not available. The age distribution of patients with AIDS — 18.30 percent of whom are less than 30 years of age, 70.85 percent 30 to 49 years of age, and 10.85 percent 50 years of age or older — was assumed for HIV infections.
For viral hepatitis, the CDC estimates that 78,000 HBV infections and 25,000 HCV infections occurred in 2001 (from the Division of Viral Hepatitis, at www.cdc.gov). The age distribution of incident HBV infections for 2000 — 37.09 percent younger than 30 years of age, 46.80 percent 30 to 49 years of age, and 16.11 percent 50 years of age or older — was assumed for cases of HBV. The age distribution for incident HCV infections for 2001 was 29 percent younger than 30 years of age, 64 percent 30 to 49 years of age, and 7 percent 50 years of age or older, and the male:female ratio was 1.7:1 (Alter M: personal communication).
No current prevalence data are available for HBV or HCV. On the basis of testing of serum samples from persons who participated in the Third National Health and Nutrition Examination Survey from 1988 through 1994, McQuillan et al. reported a prevalence rate of HBsAg of 0.42 percent, and Alter et al. reported a prevalence of anti-HCV of 1.8 percent. Furthermore, the study by McQuillan et al. showed a male:female ratio of 1.4:1 with respect to the prevalence of total HBV infections. These data are assumed to represent the current status and were used in this assessment. No data are available on HTLV infection in the general population.
Estimated Incidence Rates among Tissue Donors
By extrapolating from the rates among first-time blood donors, we estimated that the incidence rates among tissue donors were 30.118 per 100,000 person-years for HIV, 18.325 per 100,000 person-years for HBsAg, 12.380 per 100,000 person-years for HCV, and 5.586 per 100,000 person-years for HTLV. The prevalence ratios for tissue donors relative to those in the general population were 0.46 for HIV, 0.54 for HBsAg, and 0.61 for HCV; the corresponding estimated incidence rates per 100,000 person-years for tissue donors were 7.099, 15.100, and 4.910, respectively. The estimates derived from the blood-donor approach were higher than those derived from the general-population approach. Prevalence and incidence data from blood donors are less likely to be underestimates, owing to the systematic testing of each donation.
Our prevalence results were based on data from five tissue banks across the United States. A survey of tissue banks accredited by the American Association of Tissue Banks, conducted in June 2000 for calendar year 1999, showed rates of reactivity on screening of 0.35 percent for HIV (66 of 19,091 donations), 0.94 percent for HBsAg (179 of 19,090 donations), 1.49 percent for HCV (285 of 19,130 donations), and 0.53 percent for HTLV (101 of 19,072 donations). Our results — 0.34 percent, 0.71 percent, 1.51 percent, and 0.60 percent, respectively — are close to those of the survey. Such consistency suggests that our data are representative of the tissue-donor population in the United States.
The measured prevalence rates among tissue donors fall between those found among first-time blood donors and those attributed to the general population. This is not surprising, since tissue donors, although more representative of the general population than are blood donors, are carefully selected on the basis of medical history, physical examination, and interviews with the next of kin. Such a process, however, is not as effective as the face-to-face interview that is conducted with blood donors.
By imputing rates from first-time blood donors and, separately, from the general population, we used an indirect approach to assign incidence rates to tissue-donor populations. For our primary estimates, we adjusted these rates to reflect the different prevalence rates among the tissue donors and the populations used for comparison. We used the resulting incidence rates with estimated window periods to estimate the probability of viremia at the time of tissue donation that would have gone undetected on screening with the use of current serologic tests.
Our data are based on information from 11,391 tissue donors. Donations from approximately 20,000 tissue donors are processed annually in the United States, generating roughly 1 million separate products. According to our estimates, the probability that a donor is viremic at the time of donation is 1 in 55,000 in the case of HIV infection, 1 in 34,000 in the case of HBV infection, 1 in 42,000 in the case of HCV infection, and 1 in 128,000 in the case of HTLV infection. We suggest that the respective upper bounds of these figures would be 1 in 22,000, 1 in 19,000, 1 in 17,000, and 1 in 41,000; in other words, 1 or fewer donors would be viremic per year. These figures clearly indicate that the risk of infectivity is low, and in fact, most transplanted products are treated to reduce or eliminate the risk of infectivity. However, since tissues from a single donor may be used in an average of 50 patients, a single donor has the potential to infect an unknown, although probably small, number of recipients.
The implementation of nucleic acid--amplification testing of "minipools" (pools of 16 to 24 blood donations) has markedly reduced the residual risk of viremia and transfusion-transmitted infection; the reduction in risk is directly proportional to the decrease in the length of the window period achieved by the use of this approach, by 5 days for HIV and by 60 days for HCV. Studies have shown that nucleic acid--amplification testing of individual donations would reduce the window period to 7 days for HIV and HCV and to 20 days for HBV. If individual testing were to be used for tissue donors, the probability of donor viremia would be reduced to 1 in 173,000 for HIV, 1 in 421,000 for HCV, and 1 in 100,000 for HBV. Assuming that it would cost approximately $150 ($50 per virus on the basis of current charges) to test each donor for the three viruses, the overall cost of eliminating one potentially infectious donor would be $4.0 million in the case of HIV infection, $2.3 million in the case of HCV infection, and $2.6 million in the case of HBV infection. Presumably, that cost would be spread over 1 million or more tissue products each year. Currently, efforts are under way to implement nucleic acid--amplification testing of cadaveric samples.
Overall, we believe that current measures used to evaluate tissue donors are effective and that the probability of collecting products from a viremic donor is low, but not negligible. On the basis of the model used for donated blood, this probability could be further reduced by the addition of nucleic acid--amplification testing at an approximate cost of less than $5 per product.
The Safety and Availability of Blood and Tissues — Progress and Challenges
Jesse L. Goodman, M.D., M.P.H.
The availability of a safe blood supply is critical for both medical progress and national security. Safety has been increased by nucleic acid--amplification testing, as documented by Stramer et al. in this issue of the Journal.1 As health care providers, public health officials, and providers and users of donated blood and tissues we strive to improve the safety of the blood supply and to consider future threats, including threats to the safety of the donated tissue supply.
Twenty years ago — with tragic consequences — up to 1 in 100 blood units in the United States transmitted the human immunodeficiency virus (HIV) or hepatitis C virus (HCV), as did plasma that did not undergo what is now recognized as effective viral inactivation. Careful screening of donors for risk factors and technological innovations, from immunoassays to nucleic acid--amplification testing, has since prevented thousands of transfusion-transmitted infections. As described by Stramer et al., the use of nucleic acid--amplification testing to detect HIV and HCV during the initial "window period" of seronegativity after infection has further reduced the risk of transmission and, together with the advent of modern serologic testing, has improved the safety of the blood supply by a factor of more than 1000. Although this accomplishment is remarkable, we must now consider how to prevent future potential catastrophes.
As residual risks decrease, the costs of addressing them often rise. Effective serologic testing reduced the risk of transmission of HIV and HCV to 1 in 1.5 million and 1 in 276,000, respectively. The addition of nucleic acid--amplification testing has reduced the risks of both to approximately 1 in 2 million blood units. This additional prevention of the transmission of approximately 5 HIV infections and 56 HCV infections per year costs more than $100 million annually, or about $2 million per infection prevented.2 Some costs may be offset by the discontinuation of less effective tests. Noneconomic factors must also be considered, but in some cases the cost-effectiveness of such testing may not compare favorably with that of other preventive health measures.2,3 The use of nucleic acid--amplification testing for hepatitis B virus is likely to raise similar issues, as will the potential use of nucleic acid--amplification testing for individual units of blood, as opposed to pooled units, to achieve even greater sensitivity.
Although it is likely that unknown threats to the safety of blood and tissue pose the most danger, what threats are known? Bacterial contamination, which is usually due to skin organisms and less often to occult bacteremia in the donor, occurs in approximately 1 in 2000 platelet transfusions, largely because of the need to store platelets at room temperature, and such contamination can have serious consequences.4 Recently developed, rapid culture methods offer the potential to reduce this risk, though their clinical effectiveness is not yet known. Mosquito-borne infections, including malaria and possibly dengue, both acquired during travel and, potentially, from domestic transmission, pose additional threats. As with West Nile virus infections, any epidemic will predominantly be mosquito-borne rather than blood-borne, making vector control the primary intervention. But such control may be difficult to achieve, necessitating donor testing.
Tick-borne babesial infection can, like malaria, result in an asymptomatic carrier state, as can infection with agents normally considered "tropical," including Trypanosoma cruzi, the agent of Chagas' disease, of concern among Central and South American immigrants,5 and leishmania, of special concern in returnees from the Middle East.6 Variant Creutzfeldt--Jakob disease has been transmitted by transfusion in animal models, and two cases of transfusion-transmitted disease have been reported in patients in the United Kingdom.7,8 In the absence of a useful screening test for this disease in donors, the Food and Drug Administration (FDA) has recommended deferring any potential donor with substantial risk factors for exposure.9 Although the FDA used modeling to minimize donor losses, the use of deferrals based on risk factors such as travel is inefficient and diminishes already short blood supplies. Finally, the use of biologic and other types of agents by terrorists may threaten blood and tissue safety, whether through product contamination or infection of donors. We have entered a time when both our capacity to identify and respond to new threats and the expectation and need to do so effectively have increased.
New forms of technology have revolutionized our response capacity. It took several years to implement nationwide nucleic acid--amplification testing of blood donors for HIV and HCV. In contrast, by building on that technique when faced with a growing epidemic of West Nile virus infection and its transmissibility through antibody-negative blood,10 prompt cooperative action by government and industry led to nationwide implementation of investigational nucleic acid--amplification testing of donors within nine months, preventing the transfusion of more than 1000 units infected with the virus in 2003. Related techniques have made it possible to identify new infectious threats rapidly; it took over three years to identify HIV but less than a month to pinpoint the viral cause of severe acute respiratory syndrome. We need further innovation. For example, the availability of mass-produced nucleic acid chips or nanoassays for proteins or nucleic acids could make possible the cost-effective detection of all known pathogens, including biologic agents used by terrorists, and even newly identified members of families of pathogens. Although these advances are promising, much work remains to develop and field-test practical, sensitive methods.
The lessons learned from blood apply to human tissues, such as skin, bone, and ligaments, particularly in view of their increasing use (in approximately 1 million medical procedures a year). These tissues can transmit the same viral infections as blood, and the products of a single tissue donation may be transplanted to as many as 100 recipients. As discussed by Zou et al. in this issue of the Journal,11 tissue donors are screened for risk factors and have lower rates of infection than the general population, though not as low as blood donors. Nucleic acid--amplification testing of tissue donors for HIV and HCV should therefore have yields similar to those for blood donors, and with similar cost--benefit issues.
In contrast to blood, transplanted tissues are commonly of cadaveric origin and undergo processing that may increase their inherent risk of bacterial or fungal contamination. Serious outcomes may result12 but have been reported uncommonly, owing in part to underrecognition but also to the good health of most recipients, the common perioperative use of antibiotics, and the adoption of voluntary standards by many in the tissue-processing industry. The FDA has stepped up its inspection and enforcement activities and is implementing a regulatory framework for tissues that includes thorough screening and testing of donors for known and emerging biologic threats, the use of validated, good tissue-manufacturing practices, and routine reporting of adverse events.13,14 These steps should enhance tissue safety but, like those for blood, will not be fail-safe.
Several actions are needed to promote the safety and availability of blood and tissue. First, we must recognize donor testing as an integral component of disease surveillance and prevention. Screening of donors for West Nile virus has helped track the epidemic in real time, finding early cases and defining geographic spread, often in advance of clinical case reporting. Donor screening also provides valuable information on disease trends in healthy populations. Better screening tools for identifying multiple pathogens may also prove useful in the early detection of emerging pathogens or bioterrorism events.
Second, we need to establish proactive and collaborative ways of communicating about emerging infectious diseases and build response capabilities into programs devised to monitor the safety of blood and tissue. We know what can be accomplished through rapid, collaborative communication and action — for example, with respect to West Nile virus. Current activities to identify and prepare for future risks to the blood supply, such as periodic scientific assessments conducted by the Public Health Service and frequent consultation with the American Association of Blood Banks task forces, have been extremely important. Similar approaches should be developed for tissues. Continuing education regarding infections that can potentially be transmitted by blood and tissues is also needed to promote the reporting of adverse events, an important frontline component of surveillance.
Third, we need to identify and prioritize an agenda of applied scientific development and related implementation activities. Market incentives for these activities are limited, and carefully targeted scientific and technical support, often collaborative, can have substantial effects. For example, the study by Stramer et al. involved academic transfusion-medicine experts and industry experts and was supported by long-term funding from the National Institutes of Health. Further scientific collaboration to identify and address public health priorities in blood and tissue safety is essential. Activities should be linked to public health preparedness needs in an anticipatory manner. An example would be the development, in advance of the need for widespread use, of assays to detect agents that represent potential threats, with phased-in formatting for and evaluation of the assays in actual screening settings. The FDA has recently described a "critical-path" initiative that seeks to identify needs and encourage scientific progress in order to facilitate the development and availability of safe and effective products.15 Such needs include the development of methods for testing for multiple pathogens; methods for better and longer-term preservation and storage of blood, blood cells, and tissues; and more effective ways of removing or inactivating pathogens or achieving sterility while maintaining function.
Fourth, we must promote public discussion of the implementation of measures to enhance the safety of blood and tissue. Evidence-based public discussion, including consumer perspectives, should help set priorities. Benefits and costs should be viewed broadly. Changing procedures may have unintended effects in other areas or on product availability. Both public workshops and the Public Health Service's Advisory Committee on Blood Safety and Availability have provided useful forums for such discussions. Finally, we have to recognize and support the donation, availability, and safety of blood and tissue as part of our critical national infrastructure. Our health care system and medical progress depend on the safety and availability of blood, tissues, and organs for transplantation. We need to improve our understanding of donors' motivations and to support efforts to increase donations. The availability of blood and tissues such as skin and hematopoietic stem cells may be of special importance in wartime or after terrorist events, and in fact, their supply and integrity may be threatened during such times. The best preventive is a sound infrastructure. Industry and government are working cooperatively to anticipate and be prepared for such challenges. These efforts must be expanded and sustained.