Skip Navigation


American Journal of Epidemiology Advance Access originally published online on March 22, 2007
American Journal of Epidemiology 2007 165(12):1443-1453; doi:10.1093/aje/kwm030
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
165/12/1443    most recent
kwm030v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Goedert, J. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Goedert, J. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Copyright © 2007 by the Johns Hopkins Bloomberg School of Public Health All rights reserved; printed in U.S.A.

ORIGINAL CONTRIBUTIONS

Reconstruction of the Hepatitis C Virus Epidemic in the US Hemophilia Population, 1940–1990

James J. Goedert1, Bingshu E. Chen1, Liliana Preiss2, Louis M. Aledort3, Philip S. Rosenberg1 and for the Second Multicenter Hemophilia Cohort Study

1 Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Rockville, MD
2 RTI International, Rockville, MD
3 Mount Sinai School of Medicine, New York, NY

Correspondence to Dr. James J. Goedert, Viral Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, 6120 Executive Boulevard, Room 7066, Rockville, MD 20852 (e-mail: goedertj{at}mail.nih.gov).

Received for publication August 25, 2006. Accepted for publication November 30, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
Hepatitis C virus (HCV) is a blood-borne infection readily transmitted by transfusion. Persons with hemophilia were at very high risk of acquiring HCV, but the chronology and correlates of HCV incidence in the US hemophilia population remain unknown. The authors used multiple data sources and new statistical methods to reconstruct HCV incidence in White males with hemophilia A from 1940 through 1990. HCV incidence was ~1%/year until 1950 but 2–3%/year by 1955. With mild hemophilia, HCV incidence increased in the 1960s, reaching a plateau of ~8%/year from 1969 to 1980. With moderate and severe hemophilia, HCV incidence increased steeply to peaks of 11.7%/year in 1970 and 17.2%/year in 1968, respectively. Overall, HCV incidence declined after 1970, steeply after 1984, to nearly zero by 1990. With improving and increasing use of plasma derivatives, the size of the hemophilia population increased 86% during these 50 years. Study results imply that these life-saving treatments also carried an increasing risk of HCV, particularly before clotting factor concentrates were licensed in the 1970s. They also suggest that multiple synergistic interventions since 1970, particularly donor deferral, screening for hepatitis B and human immunodeficiency virus, and viral inactivation of clotting factor concentrates, were needed to reduce transfusion of HCV prior to its discovery.

blood component transfusion; factor VIII; hemophilia A; Hepacivirus; hepatitis B virus; HIV; models, statistical; plasma

Abbreviations: AIDS, acquired immunodeficiency syndrome; CI, confidence interval; HBsAg, hepatitis B virus surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; MHCS-II, Second Multicenter Hemophilia Cohort Study; NHANES III, Third National Health and Nutrition Examination Survey


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
During the last 40 years, three infections of major clinical importance were discovered and largely eliminated from the blood transfusion system (1). In 1965, Blumberg et al. (2) discovered hepatitis B virus (HBV) surface antigen in 29 percent of sera from patients with hemophilia. In 1983–1984, groups at the Institut Pasteur and the US National Cancer Institute discovered human immunodeficiency virus (HIV) in patients with the acquired immunodeficiency syndrome (AIDS) (3, 4). In 1989, hepatitis C virus (HCV) was discovered by Choo et al. (5) in the serum of patients with non-A, non-B hepatitis. As recently reviewed (6), HCV is a heterogeneously distributed global infection and a major cause of liver disease.

The origins and dissemination of HCV in the US population have not been well characterized. "Molecular clock" analyses of rates of genetic divergence suggest that the HCV genotype 1 appeared in Japan around 1882 and in the United States around 1910, with dissemination in these populations in the 1930s and 1960s, respectively (7, 8). Few specimens from those years exist to confirm this evolutionary analysis. During 1948–1954, HCV serum antibodies were detected in 17 (0.2 percent) of 8,568 young adult recruits to the US military (9). Results from the Third National Health and Nutrition Examination Survey (NHANES III) suggested that, by 1988–1994, HCV seroprevalence among young adult men might have increased eightfold (10). The pattern of HCV incidence during the intervening 40 years is unsettled.

To begin addressing the historical dissemination of HCV, Armstrong et al. (11) modeled data from the United States on rates of acute non-A, non-B hepatitis, seroprevalence in NHANES III, and assumptions about loss of HCV antibodies. According to models with little or no loss of HCV antibodies, which had the best fit to the NHANES III data, HCV incidence was low from 1950 to the mid-1960s and then increased steeply during the 1970s to a plateau from 1980 to 1990. Using additional sources of data, Salomon et al. (12) developed a statistical model showing a low incidence of HCV in the United States in the 1950s, a steep increase after 1965 that peaked in 1985, and a gradual decline to 1997. In France, Deuffic et al. (13) back-calculated HCV incidence from a Markov natural history progression model derived from three cohorts of HCV-infected patients, as well as rates of death from hepatocellular carcinoma and from all causes. Constrained to an exponential shape, the model suggested that the HCV epidemic in France started about 1940, with rapidly increasing incidence until 1970, after which it leveled off until blood screening started in 1990. With a similar but less constrained model, Sypsa et al. (14) estimated that HCV infections in Greece started in the late 1940s and increased steadily to the late 1980s. Other than the marked effect of HCV screening in 1990, factors related to historic increases and decreases in HCV incidence have not been identified in any population.

HCV seroprevalence is very high among people with hemophilia who were treated with plasma or plasma derivatives before 1985, and most of these infections have been attributed to clotting factor concentrates made from large pools of donors (15, 16). However, the risk of HCV prior to licensure of concentrates in the early 1970s has not been examined. The objective of the current study was to clarify the incidence of HCV from 1940 to 1990 among people with hemophilia in the United States. Given the data available, the analysis focused exclusively on males in the United States with hemophilia A. Because of their need for frequent infusions of plasma and plasma derivatives to treat or prevent bleeding, the experience of this population provides information about the changes in transfusion-transmitted infections as they evolved in the United States during those 50 years.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 References
 
The Second Multicenter Hemophilia Cohort Study population
As reported previously (17), the Second Multicenter Hemophilia Cohort Study (MHCS-II) was established to identify, quantify, and develop markers for complications of HCV, as well as later complications of HIV and its treatment. All patients at 42 US and 12 non-US collaborating comprehensive hemophilia treatment centers were invited to participate in the MHCS-II if they had reached age 13 years and had a congenital coagulation disorder, including hemophilia A or B (congenital factor VIII or IX deficiency, respectively), von Willebrand disease, or other types. The current analysis is limited to those with hemophilia A. MHCS-II eligibility also required a positive result on a locally performed, licensed assay for HCV antibodies (anti-HCV), HIV antibodies, or HIV RNA, performed on or after January 1, 1993. Earlier HCV and HIV results were not considered to avoid erroneous results with first generation assays. Supplemental HCV serology was not performed, because of the very high HCV prevalence and related very high positive predictive value of locally performed anti-HCV assay results in this population. As expected, HCV RNA was subsequently detected in 73 percent of participants reported by the centers to have been anti-HCV positive (18). All 1,858 anti-HCV-positive participants with hemophilia A who enrolled in the MHCS-II at US centers from 2001 to 2005 were eligible for the current analysis. There were 1,196 with severe hemophilia (defined as <1 percent of normal factor VIII activity), 291 with moderate hemophilia (1–5 percent activity), and 371 with mild hemophilia (>5 percent activity). Additional information is available at https://mhcs-ii.rti.org/.

Types of plasma exposures
Development and clinical use of fresh-frozen plasma started in the late 1940s, cryoprecipitate in the mid-1960s, and clotting factor concentrate in the early 1970s. Fresh-frozen plasma was obtained from individual donors. Cryoprecipitate was pooled from a few to a maximum of 20 donors. Fresh-frozen plasma and cryoprecipitate are subsequently referred to as "cryo/plasma." Factor concentrate was derived from large pools of plasma from 20,000 to 50,000 donors. Use of concentrates that had been heat treated to inactivate enveloped viruses became standard practice in the early to mid-1980s. Earlier versions of clotting factor concentrates are termed "non-heat treated." Based on the types and the order of the treatment received, each participant in the current study was classified as having one of three types of exposure: 1) cryo/plasma only; 2) cryo/plasma before non-heat-treated factor concentrate; and 3) non-heat-treated concentrate before or without cryo/plasma.

Statistical methods
Determining annual incidence required estimating the number at risk and the number infected each year. Detailed methods are described in the Appendix. Briefly, we calculated the size of the hemophilia A population by obtaining natality and mortality data for White male births from the US Census, to which we applied hemophilia A-specific data, including a prevalence of 1:7,500 male livebirths, mortality data from the MHCS-I cohort (19), and the relative risk of death before MHCS-I from the report by Jones and Ratnoff (20). To obtain the number at risk for HCV infection in each year, we subtracted the number previously infected from the number alive in each year.

To calculate the number infected in each year, we estimated the distribution of infection dates in the MHCS-II cohort, adjusted these for deaths among infected patients before initiation of MHCS-II, and finally scaled the annual infection dates to the national US hemophilia population. We used two methods, termed "EMS2D" (expectation-maximization algorithm with smoothing in two dimensions) and "method B," respectively, to estimate individual dates of HCV infection. The EMS2D method has been reported and applied recently (18). For both methods, we assumed that the concentrate used after December 31, 1986, was noninfectious, but that the concentrate used prior to December 31, 1986, resulted in infection with any single exposure. Therefore, for participants in group 3 above, we assumed that HCV infection occurred as a consequence of the first exposure. For participants in groups 1 and 2 above, the precise dates of infection were unknown but occurred in the interval between the first and last potentially infectious exposures. For EMS2D, we grouped participants by severity and birth dates within ±8.6 years for severe hemophilia, ±8.5 years for moderate hemophilia, and ±8.1 years for mild hemophilia and then smoothed the age-specific exposure probabilities for these cohorts using an expectation-maximization algorithm (21, 22).

For participants in groups 1 and 2 above, method B assumed that the hazard of HCV infection increased as HCV seroprevalence increased in the population (9, 10) and that it was equivalent to that for non-A, non-B hepatitis described in donor-recipient studies (23, 24), specifically, 3–6 non-A, non-B hepatitis cases/1,000 units transfused (0.003–0.006 cases per unit). Method B assigned HCV infection to the age at which the participant reached 50 percent infection-free survival. If the participant never reached 50 percent infection-free survival, then the HCV infection age was assigned to the age of last exposure. With both methods, the imputed dates of infection were calculated by adding the imputed ages to the corresponding dates of birth.

To account for mortality prior to MHCS-II, we obtained the number of HCV infections in each calendar year by scaling up the "seen sample" in MHCS-II to account for the "unseen sample" (25). The MHCS-II data were scaled to the national hemophilia population by using national AIDS surveillance and MHCS-I survivorship by HIV status (19, 26). As described in the Appendix, a global test of heterogeneity was used to determine whether HCV incidence curves differed among participants with mild, moderate, and severe hemophilia A.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX
 References
 
From 1940 to 1985, the US hemophilia A population increased from approximately 9,400 to 17,500. As shown in figure 1, there were increases from 4,400 to 5,800 (32 percent) with mild, from 2,500 to 4,000 (60 percent) with moderate, and from 2,500 to 7,700 (208 percent) with severe hemophilia. The population declined to 16,000 by 1995 during the HIV/AIDS epidemic.


Figure 1
View larger version (11K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 1. Number of White males with hemophilia A (factor VIII deficiency) alive in the United States from 1940 to 2000 (dashed line) and the corresponding 95% confidence interval from 1980 to 2000 (solid lines), by severity of hemophilia.

 
Two methods (EMS2D and method B) were used to estimate HCV infection dates from the individual treatment history data of participants in MHCS-II. Among all hemophilia A participants with sufficient treatment history data (n = 1,620), the mean age at infection was 9.3 years with EMS2D compared with 10.2 years with method B. Between the two methods, the imputed ages were highly correlated (R = 0.972). Excluding participants with inconsistent or missing data, we found the mean age at infection to be 8.0 years with EMS2D compared with 8.4 years with method B (R = 0.98, n = 1,131). In the subgroup that used cryo/plasma exclusively or before non-heat-treated clotting factor concentrate, the mean age at infection was 10.9 years with EMS2D compared with 12.3 years with method B (R = 0.96, n = 1,046). Excluding from this subgroup participants with inconsistent or missing data, we found the mean age at infection to be 9.7 years with EMS2D compared with 10.6 years with method B (R = 0.96, n = 557). With inclusion or exclusion of participants with inconsistent or missing data, agreement between the two methods was homogeneous across the range of ages (data not presented).

By severity of hemophilia A, figure 2 presents the distribution of HCV infection dates, estimated by EMS2D for MHCS-II participants before (upper panels) and after (lower panels) accounting for the "unseen sample" of those who died prior to MHCS-II. Each panel shows a few infections before 1960, many around 1970, and fewer in the late 1980s. The large number of unseen infections during the 1970s, especially for those with severe hemophilia, reflects high mortality prior to MHCS-II. Dates estimated by method B were similar and are available upon request.


Figure 2
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 2. Distribution of hepatitis C virus (HCV) infection dates among males with hemophilia A (factor VIII deficiency) enrolled from 2001 to 2005 in the Second Multicenter Hemophilia Cohort Study (MHCS-II), by severity of hemophilia. Upper panels are histograms of infection dates in MHCS-II imputed by use of the "EMS2D" method (expectation-maximization algorithm with smoothing in two dimensions). Lower panels are dates adjusted for the unseen cohort of people who died prior to initiation of MHCS-II.

 
Figure 3 presents annual HCV incidence rates (and 95 percent confidence intervals) by severity of hemophilia A and the date estimation method. As with the imputed infection ages, the incidence rate curves with EMS2D (figure 3, upper panels) and method B (figure 3, lower panels) were very similar, particularly with a broad plateau in the incidence among people with mild hemophilia (left panels). By EMS2D, the peak years of incidence were 1971 (95 percent confidence interval (CI): 1969, 1980) with mild, 1970 (95 percent CI: 1967, 1973) with moderate, and 1968 (95 percent CI: 1967, 1970) with severe hemophilia. The peak years were less well defined with method B: 1976 (95 percent CI: 1968, 1983) with mild, 1970 (95 percent CI: 1962, 1974) with moderate, and 1969 (95 percent CI: 1968, 1977) with severe hemophilia. After the peak intervals, HCV incidence declined to nearly zero by 1990 in all severity groups. With mild hemophilia, HCV incidence (EMS2D method) increased from 3.4 (95 percent CI: 1.4, 6.0) percent/year in 1965 to its peak of 8.5 (95 percent CI: 4.4, 13.2) percent/year in 1971. With moderate and severe hemophilia, both methods showed a steady increase, starting in 1949. By EMS2D, the peak incidence rate was 11.7 (95 percent CI: 4.7, 20.4) percent/year with moderate hemophilia in 1970, and it was 17.2 (95 percent CI: 12.4, 22.6) percent/year with severe hemophilia in 1968.


Figure 3
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 3. Annual hepatitis C virus (HCV) incidence from 1940 to 1990 among White males with hemophilia A (factor VIII deficiency) in the United States by severity of hemophilia. Maximum likelihood estimates (with 95% confidence intervals) are presented. Infection dates were estimated with the "EMS2D" method (expectation- maximization algorithm with smoothing in two dimensions) (upper panels) or "method B" (lower panels).

 
HCV incidence differed significantly by severity (two-sided p = 0.002), and it increased earliest with severe hemophilia, later with moderate hemophilia, and even later with mild hemophilia (figure 4, upper panel). With the severity groups combined, the HCV incidence curve by EMS2D, annotated with relevant transfusion developments, is presented (figure 4, lower panel). From 1950 to 1970, during introduction of plasma and cryoprecipitate but prior to initiation of hepatitis B virus surface antigen (HBsAg) screening, HCV incidence increased from approximately 1 percent/year to 13 percent/year. Despite introduction of factor VIII concentrate in 1972, HCV incidence declined steadily to 1984 and then steeply to 1990 following recommendations to avoid use of factor concentrate, deferral of donors with AIDS risk factors, screening for HIV antibodies, and licensure of factor concentrates that had been treated with heat and other procedures to inactivate viruses (27).


Figure 4
View larger version (14K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 4. Annual hepatitis C virus (HCV) incidence from 1940 to 1990 among White males with hemophilia A (factor VIII deficiency) in the United States. The upper panel shows maximum likelihood estimates by the "EMS2D" method (expectation-maximization algorithm with smoothing in two dimensions) and p value to test the difference in incidence by severity of hemophilia. The lower panel shows the maximum likelihood estimate and 95% confidence interval for all severities of hemophilia A, annotated by events postulated to affect the risk of HCV transmission by plasma and plasma products. HBsAg, hepatitis B virus surface antigen; AIDS, acquired immunodeficiency syndrome; HT, heat-treated; HIV, human immunodeficiency virus.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX
 References
 
Our reconstruction of the HCV epidemic among people with hemophilia has implications for understanding and minimizing the risk of transfusion-associated infections (1, 28). Development and subsequent widespread use of fresh-frozen plasma in the late 1940s, cryoprecipitate in the mid-1960s, and clotting factor concentrate in the early 1970s enabled moderately effective treatment and ultimately prophylaxis of hemophilic bleeding. Surgery, previously unthinkable for hemophilic patients, became possible. Hemorrhagic morbidity and mortality fell steadily. Our study is a summation of the experience of people born in each year and with varied severity of hemophilia. We found that, prior to its discovery, the risk of HCV increased with the introduction of plasma infusions, especially with cryoprecipitate, and decreased with the steps taken to stem the transmission of HBV and HIV.

Hepatitis was a recognized complication of transfusion before 1950. The risk of posttransfusion hepatitis declined in the 1970s with the gradual elimination of payments for whole blood donations, and it fell markedly with HBsAg testing of donated blood and plasma that started in 1970 and was mandated throughout the United States in 1972 (24). Although the prevalence of HBsAg in factor VIII concentrate fell to virtually zero (27), clotting factor concentrates continued to transmit HBV occasionally and non-A, non-B hepatitis frequently. This is not surprising. People with severe hemophilia, as well as those with moderate hemophilia who have traumatic or surgical bleeding, require intensive clotting factor replacement therapy and thus are highly susceptible to infectious and other complications of plasma replacement therapy (29). Our reconstruction of HCV incidence with severe or moderate hemophilia revealed a steeply increasing risk of HCV as fresh-frozen plasma and cryoprecipitate were introduced to control bleeding. The peak incidence of HCV prior to nationwide HBsAg screening implies that HCV already had reached very high prevalence in the moderate and severe hemophilia populations prior to the availability of clotting factor concentrates. Because children continued to be born with severe and moderate hemophilia after 1970, saturation would not account for attenuation of the risk in the later years. Rather, the declining risk parallels the declining prevalence of paid and HBsAg-positive donors.

People with mild hemophilia are more akin to the general population of transfusion recipients, because many of them seldom if ever require clotting factor concentrate to control bleeding. With mild hemophilia, HCV incidence did not increase noticeably until the mid-1960s, an observation that is consistent with the statistical model that Salomon et al. (12) developed for HCV from all sources. In that model, HCV incidence did not peak until 1985 and declined slowly thereafter. Nonetheless, that model had a dip in HCV incidence in 1972–1973, interpreted by the authors to reflect the initiation of HBsAg screening of blood donors (12). Support for a substantial risk with cryo/plasma prior to screening is provided by the observation that antibodies to the hepatitis B core antigen were found in nearly all hemophilic patients who had received at least 100 units of plasma (30). Our observation that there was no discrete peak in HCV incidence but rather a high steady rate from 1968 to 1983 among people with mild hemophilia is consistent with intermittent use of cryo/plasma for surgery or trauma episodes. Comparison of this flat incidence rate for mild hemophilia with the model of Salomon et al. (12) suggests that the second, larger wave of HCV infections in the United States from 1973 to 1985 was primarily a consequence of nontransfusion exposure, such as injection drug use.

In all people with hemophilia, the HCV risk fell sharply after 1984. Compared with HCV, the HIV epidemic for people with hemophilia was much more acute, running from 1978 to 1987 with a threefold higher peak risk (60 percent per year) in 1982, contemporaneous with the first recognized cases of AIDS (31). The reduction and eventual elimination of HIV transmission, and perhaps HCV transmission, to people with hemophilia can be attributed primarily to three increasingly effective interventions—deferring donors deemed to be at high risk for AIDS, screening donated plasma for HIV antibodies, and licensure of clotting factor concentrates that had been treated with heat or other procedures to inactivate viruses (31). As additional responses to the HIV contamination of the blood supply, the intensified surveillance in the United States for suspicious health complications possibly related to transfusion, stimulation of research, and implementation of new technologies already have improved transfusion safety (1, 32).

One cannot conclude from our study that changes in plasma products or interventions caused changes in HCV incidence. Nonetheless, we postulate that HCV incidence increased in the 1950s and 1960s as a direct result of increasing likelihood of exposure to and use of newer products. Compared with fresh-frozen plasma from single donors, cryoprecipitate was pooled from up to 20 donors. Factor concentrates, pooled from 20,000–50,000 donors, were developed in the 1970s. Despite a massive exposure to potentially infected donors in factor concentrates, it is plausible that the declining HCV incidence during the 1970s and 1980s resulted from interventions directed against HBV and HIV because of their shared risk factors.

Our study has other limitations. Our mathematical imputation of HCV infection dates and extrapolation to the US hemophilia population required the use of data and parameters that are not known with certainty. These included recall of initial exposures to plasma products by participants in MHCS-II, changes in mortality rates (19, 20), and the size of the US hemophilia population by HIV status and severity (26). Imprecision in the population size would primarily affect the scale (vertical axis) rather than the shape of the incidence curves. Different estimates of mortality with hemophilia would affect the number at risk in each year and consequently the shape of the incidence curves. Validation in another population with good mortality data, such as the United Kingdom (16), would be helpful.

Method B for imputing HCV infection dates depended on the ability of MHCS-II participants to recall the frequency of cryo/plasma use and on the per-unit hazard of non-A, non-B hepatitis (23). Compared with EMS2D, HCV incidence by method B had a sharper peak around 1970, perhaps because we assumed that the risk of HCV increased fourfold, from 0.0015 to 0.006 per unit, in that year (23, 24). The actual increase in risk certainly would have been more gradual. After 1970, imputed infection dates would be slightly too early if the actual probability of HCV infection with initial use of non-heat-treated clotting factor concentrate was less than the 100 percent that we assumed.

Because unbiased specimen collections do not exist to directly examine the spread of HCV, we used historical data to reconstruct the HCV epidemic in the US hemophilia population. Morbidity and mortality from bleeding were reduced by plasma products that were pooled from increasing numbers of donors and introduced between 1950 and 1975. HCV incidence was temporally increased with introduction and increasing use of fresh-frozen plasma and cryoprecipitate, and it was temporally decreased with conversion to a volunteer whole blood supply, specific measures to reduce the risk of HBV and HIV, and introduction of heat-treated clotting factor concentrate. In addition to its historical interest and possible lessons for other blood-borne infections, imputation of HCV infection dates may help to clarify factors associated with HCV chronicity or progression (18).


   APPENDIX
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
References
 
Statistical Methods
Overview.
We used four steps to calculate annual HCV incidence in the hemophilia population. First, the distribution of HCV infections by calendar year for current hemophilia patients was estimated from histories of exposure to plasma clotting factors. Second, to obtain the numerators for annual HCV incidence, we adjusted the distribution of these infections for mortality in earlier years ("the unseen cohort"). Third, to obtain the denominators for annual HCV incidence, we calculated the number of susceptible (HCV uninfected) hemophilia patients alive in each year by use of data from several sources. Finally, the HCV incidence curves were smoothed.

Estimation of HCV infections by calendar year.
Before the early to mid-1980s, when heat inactivation of enveloped viruses in clotting factor concentrate became standard practice, people with hemophilia were treated with non-heat-treated clotting factor concentrate that was derived from large pools of plasma (20,000–50,000 donors). They also were treated with single-donor units of fresh-frozen plasma and with cryoprecipitate that pooled a few to a maximum of 20 single-donor units (subsequently referred to as "cryo/plasma"). For the current analysis, each anti-HCV-positive participant in the Second Multicenter Hemophilia Cohort Study was classified into one of three groups, on the basis of the types and the order of treatment received: 1) only cryo/plasma; 2) cryo/plasma before non-heat-treated clotting factor concentrate; and 3) non-heat-treated clotting factor concentrate before or without cryo/plasma.

We used two methods, termed "EMS2D" and "method B," respectively, to estimate individual dates of HCV infection. For both methods, we assumed that concentrate used after December 31, 1986, was noninfectious but that concentrate used on or before December 31, 1986, resulted in infection with any single exposure. This date was well after use of non-heat-treated clotting factor concentrate had ceased in the mid-1980s. However, sporadic cases of HCV transmission attributed to heat-treated clotting factor concentrates occurred over the next few years, until development and implementation of more stringent virucidal treatments and exclusion of anti-HCV-positive donors. Therefore, for participants in group 3 above, both EMS2D and method B assumed that HCV infection occurred as a consequence of the first exposure to pre-1987 concentrate.

For participants in groups 1 and 2 above, the precise dates of infection were unknown but occurred in the interval (window) between the first and last exposure to potentially infectious plasma. The EMS2D method used an extension of Turnbull's maximum likelihood approach (33) to estimate an age-at-exposure distribution appropriate for each subject using observed exposure windows for a "similar" group of subjects. To define groups of subjects with similar exposure histories, we noted that naïve estimates of age-at-exposure defined as the midpoints of the exposure windows were strongly dependent on the severity of hemophilia and the year of birth but not on the HIV status (data not shown). Therefore, for each subject i with window [Li, Ri], we assembled all ni subjects including subject i with the same severity and dates of birth (occurring within ±8.6 years for severe, 8.5 years for moderate, and 8.1 years for mild severity), yielding a sample [Lk, Rk], k = 1, ..., ni for estimation of the age-at-exposure distribution Fi(a) for subject i. These data were discretized to the nearest month, and Turnbull's expectation-maximization algorithm was applied with the modifications described by Becker and Marschner (21) and by Silverman et al. (22), which incorporate a smoothing step after each maximization step, yielding an expectation-maximization-smoothing estimator Formula . Specifically, after each maximization step, the probability of exposure at each age in months was smoothed by use of locally weighted regression (34). After convergence, we imputed subject i's most likely age-at-infection Formula as the expected value of Formula (a) conditional on a's falling within the observed window [Li, Ri]. We called the entire procedure the "EMS2D" method because smoothing was carried out in two dimensions, over birth cohort (using the severity-specific ±8.1- to 8.6-year span intervals noted above) and age (using a ±7.5-year span (for severe) or a ±8.6-year span (for moderate and mild) severity). These spans were chosen empirically to incorporate only a modest amount of smoothing, in the sense that the smoothed estimates Formula (a) captured the main features apparent in the unsmoothed estimates Formula (a) obtained using Turnbull's approach. The latter estimators are less suitable for imputation, however, because the estimated distribution functions generally have "gaps" (33).

As with EMS2D, for method B, we assumed that participants who first received or only received non-heat-treated clotting factor concentrates (group 3) had become infected at the first exposure. For the remaining participants (groups 1 and 2), method B assumed that the risk of HCV infection was the product of HCV in the donor population times the number of cryo/plasma units received. HCV risk per unit was assumed to be equivalent to that for non-A, non-B hepatitis described in donor-recipient studies (23, 24), specifically, 3–6 non-A, non-B hepatitis cases/1,000 units transfused (0.003–0.006 cases per unit). Method B further assumed that the risk increased as HCV seroprevalence increased in the population (9, 10). To estimate the number of cryo/plasma units received, we assumed that the number of units in each treatment varied by age, specifically, 5, 10, 15, and 30 units for ages 0–5, 6–12, 13–20, and over 20 years of age, respectively. MHCS-II questionnaires collected frequency of cryo/plasma use during each age interval as one of the following: regularly (several or many times (estimated as eight) per year); occasionally (up to three times (estimated as two) per year); and never. A fourth level, rarely (0.5 time per year), was created for intervals between first and last exposure checked as "never" and for participants (n = 34, 2.4 percent) who checked "never" despite some use during the age interval. Infection-free survival then was computed as 1 – {(1 – r1)t1 x (1 – r2)t2 x (1 – r3)t3}, where r1 = 0.0006 is the risk/unit before 1960, r2 = 0.0015 is the risk/unit during 1960–1969, and r3 = 0.006 is the risk/unit in 1970 and later; t1, t2, and t3 are the number of cryo/plasma units received before 1960, in 1960–1969, and in 1970 and later, respectively. Method B assigned HCV infection to the age at which the participant reached 50 percent infection-free survival. For group 2 (non-heat-treated clotting factor after starting cryo/plasma), infection-free survival fell to 0 percent with initial exposure to non-heat-treated clotting factor concentrate. For participants in group 1 (cryo/plasma only) who never reached 50 percent infection-free survival, the HCV infection age was assigned to the age of last exposure.

For both EMS2D and method B, the imputed dates of infection were calculated by adding the imputed ages to the corresponding dates of birth.

Population size, adjustment for mortality.
We estimated the number of HCV infections in each single calendar year by taking the unseen sample into account (25). Let j = 1, 2, and 3 denote the group indicator for mild, moderate, and severe hemophilia patients. Suppose that there were Nj participants (MHCS-II seen sample) in group j. For each participant i, let Bi be the birth year, Gi be the year of infection, Ei be the enrollment year of MHCS-II, and Di be the year of death. The seen sample comes from the population with Di > Ei. Any participant that dies before the enrollment of the MHCS-II will be in the unseen sample.

Using natality and mortality data for White males from the US National Center for Health Statistics, a hemophilia A prevalence of 1:7,500 male livebirths (http://www.hemophilia.org/bleeding_disorders/hemophilia_a.htm), mortality data from the MHCS-I cohort (19), and the relative risk of death for people with hemophilia A from the study by Jones and Ratnoff (20), we estimated the mortality rates of hemophilic subjects at single year of age t = 0, ···, 100 and birth cohort b = 1919–2000 by

Formula
For the years 1919–1979, the hazard was considered as a fixed function. For the years 1980–2000, variation of the hazard function hj(y, t) was calculated by the bootstrap method using resampling with replacement for patients in each severity group of the MHCS-I cohort.

For each subject i that was born at year Bi and infected with HCV at year Gi, the probability that he was alive and enrolled in MHCS-II at year Ei is

Formula
Therefore, for each participant in the seen sample that was born in year Bi, infected with HCV in year Gi, and survived to enroll in the MHCS-II in year Ei, there existed other Formula subjects that also were born in year Bi and were infected with HCV in year Gi, but that did not survive long enough to be enrolled in the MHCS-II. This gives the total number of unseen subjects in the MHSC-II cohort as Formula

For each single year x from 1940 to 1990, the total number of subjects that were infected with HCV in the jth group from both the seen and unseen sample can be estimated as follows:

Formula
where I{·} is the indicator function equal to 1 if the expression in the bracket is true and 0 otherwise. Similarly, the total number of patients who were infected with HCV before year y and still alive in year y was estimated by the expression that follows:

Formula
This expression assumes that HCV infection did not change the subsequent mortality rates. The quantity mj(y) was subtracted from the total hemophilia population to obtain the denominator for the incidence of HCV in year y.

HCV incidence curves adjusted for US hemophilia population size.
To obtain the incidence curve of HCV infection, we adjusted for the hemophilia population size at each calendar year from 1940 to 1990. We assumed that one of 7,500 White males was born with hemophilia A and that 25, 20, and 55 percent of them had mild, moderate, and severe disease, respectively. We denote the number of US White male births at year y from 1840 to 1990 as W(y). The number of mild, moderate, and severe hemophilia births at year y is V1(y) = 0.25 x V(y), V2(y) = 0.20 x V(y), and V3(y) = 0.55 x V(y), respectively.

Then we used the mortality rates of hemophilia patients to obtain the total number with hemophilia in year x:

Formula
This gives the estimated US White male hemophilia population by severity group. In figure 1, uncertainty in population sizes is implied by dashed lines for the years 1940–1980. For the years 1981–2000, 95 percent confidence intervals were calculated for figure 1 by the bootstrap method by use of resampling with replacement for patients in each severity group of the MHCS-I cohort, which gave an estimate of the hemophilia population size for group j in year x: Qj(x), j = 1, 2, 3 and x = 1940 ... 2000. Therefore, the incidence of HCV infection in year x is

Formula
Here, fj is the sample fraction of the MHCS-II over the total number of the hemophilia population with HCV infection, estimated from published data (19, 26). The sampling fractions by severity were f1 = 6.47 percent, f2 = 7.16 percent, and f3 = 14.5 percent.

The incidence curves were smoothed by use of a kernel smoother with a bandwidth of 3 years. This smoother is similar to a three-point moving average and thus largely nonparametric. To reflect uncertainty, pointwise 95 percent confidence bands were estimated for each incidence curve by bootstrap resampling.

The 95 percent confidence intervals of the HCV incidence curves were estimated by bootstrap resampling with replacement for patients in each severity group in both the MHCS-I cohort (for variation in the estimated mortality hazard rates and population sizes between 1980 and 1990) and the MHCS-II cohort (for variation of the numbers of HCV infections).

Finally, we constructed a global test for the heterogeneity of the HCV incidence curves among mild, moderate, and severe hemophilia subjects, obtaining a single p value for the following null hypothesis:

Formula
The permutation test gave p values of 0.002 (EMS2D) and 0.016 (method B).


   ACKNOWLEDGMENTS
 
Funded in part by the Intramural Research Program, National Cancer Institute, National Institutes of Health, and by the National Cancer Institute and the National Heart, Lung, and Blood Institute, National Institutes of Health, through contract N01-CP-01004 with RTI International.

The authors are grateful to Dr. Charles Rabkin for helpful suggestions on method B.

The following collaborators and institutions participated in MHCS-II. North American hemophilia centers: Cardeza Foundation Hemophilia Center, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania (Dr. Jamie Siegel, Kay Miller); Case Western Reserve University, University Hospitals Center for AIDS Research, Cleveland, Ohio (Dr. Michael M. Lederman); Comprehensive Hemophilia Program, Chicago Children's Memorial Hospital, Chicago, Illinois (Dr. Alexis Thompson, Jennifer Gamerman, Susan Gamerman); Children's Center for Cancer and Blood Disorders, Palmetto Richland Memorial Hospital, Columbia, South Carolina (Dr. Kevin McRedmond, Janice Withycombe); Children's Hospital Medical Center, Cincinnati, Ohio (Dr. Ralph Gruppo, Gina Stack); Hemophilia Treatment Center, Children's Hospital of Michigan, Detroit, Michigan (Dr. Jeanne Lusher, Linda Percy); Department of Hematology/Oncology, Children's Hospital of Orange County, Orange, California (Dr. Diane Nugent, Marianne McDaniel); Division of Hematology, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania (Dr. Catherine Manno, Regina Butler, Amanda Wade); Children's National Medical Center, Washington, DC (Dr. Anne Angiolillo, Dr. Naomi L.C. Luban, Christine Guelcher); Comprehensive Bleeding Disorders Center, Peoria, Illinois (Dr. Michael Tarantino, Mary Brooks); Comprehensive Center for Bleeding Disorders, Wauwatosa, Wisconsin (Dr. Joan Gill, Jodie Nelson); Comprehensive Hemophilia Diagnostic and Treatment Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina (Dr. Gilbert White, Dr. Alice Ma, Dr. Michael Fried, Aime L. Grimsley); Hemophilia Treatment Center, Cornell Medical Center, New York, New York (Dr. Donna DiMichele, Ilene Goldberg); Dayton Children's Medical Center, Dayton, Ohio (Dr. James French, Sandra Hibner); Emory University School of Medicine, Atlanta, Georgia (Dr. James Steinberg, Steven Faust, Francie Lassiter); Fairview University Medical Center, Minneapolis, Minnesota (Dr. Nigel Key, Vicky Hannemann); Lombardi Cancer Center, Georgetown University Medical Center, Washington, DC (Dr. Craig M. Kessler, Anastasia E. Lee); Hemophilia Center of Central Pennsylvania, Penn State College of Medicine at Hershey, Hershey, Pennsylvania (Dr. M. Elaine Eyster, Kathryn Galli, Gillian Jenkins); Hemophilia Center of Western New York, Inc., Buffalo, New York (Dr. Zale P. Bernstein, Linda Belling); Indiana Hemophilia and Thrombosis Center, Inc., Indianapolis, Indiana (Dr. Amy Shapiro, Patti Noblet); London Health Sciences Center, South Western Ontario Regional Hemophilia Program, London, Ontario, Canada (Dr. Lawrence Jardine, Lori Laudenbach); Hemophilia Treatment Center, Long Island Jewish Medical Center, New Hyde Park, New York (Dr. Richard Lipton, Christine Pece); Louisiana Comprehensive Hemophilia Care Center, Tulane University Medical School, New Orleans, Louisiana (Dr. Cindy A. Leissinger, Cecilia Schmidt); Richmond Hemophilia Treatment Center, Medical College of Virginia, Richmond, Virginia (Dr. Marcus Carr, Melinda Nolte); Mountain States Regional Hemophilia and Thrombosis Center, Aurora, Colorado (Dr. Marilyn Manco-Johnson, Ruth Ann Kirschman); Moncton Hemophilia Clinic, South-East Health Care Corp., Moncton, New Brunswick, Canada (Dr. Sheldon H. Rubin, Dorine Belliveau); Mount Sinai School of Medicine, New York, New York (Dr. Louis Aledort, Johanna McCarthy); Mountain States Regional Hemophilia Center-Utah, Primary Children's Hospital, Salt Lake City, Utah (Dr. Richard Lemons, Shirley Bleak); Ohio State University Hemophilia Center, Columbus, Ohio (Dr. Eric Kraut, Leslie Witkoff); Oklahoma Center for Bleeding Disorders, Children's Hospital, Oklahoma City, Oklahoma (Dr. Charles Sexauer, Felicia Kiplinger); Puget Sound Blood Center, Seattle, Washington (Dr. Arthur Thompson, Charles Cooper); South Texas Comprehensive Hemophilia Center, Santa Rosa Children's Hospital, San Antonio, Texas (Dr. Howard Britton, Karen Aufdemorte); Missouri/Illinois Regional Hemophilia Center-Adult, St. Louis University Cancer Center, St. Louis, Missouri (Dr. Hans Joachim Reimers, Judy A. Bagato); University of California, Davis, Northern Central California Hemophilia Program, Sacramento, California (Dr. Jerry S. Powell, Muriel Herr); University of Arizona Health Science Center, Tucson, Arizona (Dr. John Hutter, Mary Lou Damiano); University of California, San Francisco, Medical Center, San Francisco, California (Dr. Willis Navarro, Susan Roth); University of Cincinnati Medical Center, Cincinnati, Ohio (Drs. Joseph Palascak, Kenneth Sherman, Madeline Heffner); University of Iowa Hospitals and Clinics, Iowa City, Iowa (Dr. Jorge DiPaolo, Michael Lammer); Pediatric Hematology/Oncology, University of Mississippi Medical Center, Jackson, Mississippi (Dr. Rathi Iyer, April Morris); Pediatric Hematology/Oncology Program, University of New Mexico Health Sciences Center, Albuquerque, New Mexico (Dr. Prasad Matthew, Marcia Schwartz); University of Pennsylvania Health System, Presbyterian Medical Center, Philadelphia, Pennsylvania (Dr. Barbara Konkle, Angelique Wallace); Gulf States Hemophilia and Thrombophilia Center, University of Texas, Houston, Health Sciences Center, Houston, Texas (Dr. Keith Hoots, Dr. Deborah Brown, Megan Ullman); Hemostatis and Thrombosis Clinic, Vanderbilt University Medical Center, Nashville, Tennessee (Dr. Anne Neff, Marney Green); Department of Pediatrics, Wake Forest University School of Medicine, Winston-Salem, North Carolina (Dr. Hernan Sabio, Anita Smith). European hemophilia centers: Department of Hygiene and Epidemiology, Athens University Medical School, Athens, Greece (Dr. Anastasia Karafoulidou, Dr. Angelos Hatzakis, Vana Milona); Hemophilia and Thrombosis Center, Instituto di Medicina Interna dell Universita di Milano, Milan, Italy (Dr. Alessandro Gringeri, Dr. Augusto Federici, Dr. Antonella Saladini); Coagulation Unit, Karolinska Hospital, Stockholm, Sweden (Dr. Sam Schulman, Doris Naslin); Department of Coagulation Disorders, Malmo University Hospital, Malmo, Sweden (Dr. Erik Berntorp, Karin Lindvall); Department of Medicine, Medizinische Einrichtungen der Universitaet Bonn, Bonn, Germany (Dr. Jurgen Rockstroh, Dr. Esther Voigt, Anja Nixdorf); Hematology Department, San Bartolo Hospital, Vicenza, Italy (Dr. Francesco Rodeghiero, Dr. Giancarlo Castaman). South American hemophilia centers: Laboratorio de Coagulacao, Fundacao Faculdade de Medicina, Sao Paulo, Brazil (Dr. Paula Ribeiro Villaca, Dr. Elbio D'Amico, Eliane Sandoval); Instituto de Investigaciones Hematologicas "Mariano R. Castex," Academia Nacional de Medicina de Buenos Aires, Buenos Aires, Argentina (Dr. Raul Perez-Bianco, Dr. Patricia Bare, Dr. Daniela Neme); Laboratorio de Hemostasia-Hemocentro de UNICAMP, State University of Campinas, Campinas, Brazil (Dr. Margareth Ozelo, Dr. Erich V. de Paula). RTI International, Rockville, Maryland: Dr. Barbara Kroner, Maryanne Ardini, Sylvia Cohn, Violet Devairakkam, Kim Doeden, Tiki Firdu, Tabitha Hendershot, Kellie Kinsey, Jennifer Martin, Emily Moser, Monica Pecha, Liliana Preiss, Tracy Wills, Susan Wilson, Patty Yost, Danny Ringer. Computer Sciences Corporation, Rockville, Maryland: Dr. Michael Plankey, Dr. Frances Yellin, Myhang Dotrang, Chuck Prorok, Phillip Virgo. Science Applications International Corporation (SAIC), Frederick, Maryland: Dr. Denise Whitby, Dr. Betty Conde. National Heart, Lung, and Blood Institute, Bethesda, Maryland: Dr. Luiz Barbosa. National Cancer Institute, Bethesda, Maryland: Dr. James J. Goedert, Dr. Eric Engels, Dr. Greg Kirk, Dr. Thomas O'Brien, Dr. Charles Rabkin, Dr. Bingshu Eric Chen, Dr. Philip S. Rosenberg, Dr. Fan-chen Tseng, Dr. Tania Welzel, Dr. Mingdong Zhang.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 

  1. Busch MP, Kleinman SH, Nemo GJ. Current and emerging infectious risks of blood transfusions. JAMA (2003) 289:959–62.[Free Full Text]
  2. Blumberg BS, Alter HJ, Visnich SA. "New" antigen in leukemia sera. JAMA (1965) 191:541–6.[Medline]
  3. Popovic M, Sarngadharan MG, Read E, et al. Detection, isolation, and continuous production of cytopathic retroviruses (HTLV-III) from patients with AIDS and pre-AIDS. Science (1984) 224:497–500.[Abstract/Free Full Text]
  4. Barré-Sinoussi F, Chermann JC, Rey F, et al. Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science (1983) 220:868–71.[Abstract/Free Full Text]
  5. Choo QL, Kuo G, Weiner AJ, et al. Isolation of a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science (1989) 244:359–62.[Abstract/Free Full Text]
  6. Shepard CW, Finelli L, Alter MJ. Global epidemiology of hepatitis C virus infection. Lancet Infect Dis (2005) 5:558–67.[CrossRef][ISI][Medline]
  7. Tanaka Y, Hanada K, Mizokami M, et al. Inaugural article: a comparison of the molecular clock of hepatitis C virus in the United States and Japan predicts that hepatocellular carcinoma incidence in the United States will increase over the next two decades. Proc Natl Acad Sci U S A (2002) 99:15584–9.[Abstract/Free Full Text]
  8. Pybus OG, Charleston MA, Gupta S, et al. The epidemic behavior of the hepatitis C virus. Science (2001) 292:2323–5.[Abstract/Free Full Text]
  9. Seeff LB, Miller RN, Rabkin CS, et al. 45-year follow-up of hepatitis C virus infection in healthy young adults. Ann Intern Med (2000) 132:105–11.[Abstract/Free Full Text]
  10. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med (1999) 341:556–62.[Abstract/Free Full Text]
  11. Armstrong GL, Alter MJ, McQuillan GM, et al. The past incidence of hepatitis C virus infection: implications for the future burden of chronic liver disease in the United States. Hepatology (2000) 31:777–82.[CrossRef][ISI][Medline]
  12. Salomon JA, Weinstein MC, Hammitt JK, et al. Empirically calibrated model of hepatitis C virus infection in the United States. Am J Epidemiol (2002) 156:761–73.[Abstract/Free Full Text]
  13. Deuffic S, Buffat L, Poynard T, et al. Modeling the hepatitis C virus epidemic in France. Hepatology (1999) 29:1596–601.[CrossRef][ISI][Medline]
  14. Sypsa V, Touloumi G, Tassopoulos NC, et al. Reconstructing and predicting the hepatitis C virus epidemic in Greece: increasing trends of cirrhosis and hepatocellular carcinoma despite the decline in incidence of HCV infection. J Viral Hepat (2004) 11:366–74.[CrossRef][ISI][Medline]
  15. Eyster ME, Diamondstone LS, Lien JM, et al. Natural history of hepatitis C virus infection in multitransfused hemophiliacs: effect of coinfection with human immunodeficiency virus. The Multicenter Hemophilia Cohort Study. J Acquir Immune Defic Syndr (1993) 6:602–10.[ISI][Medline]
  16. Darby SC, Ewart DW, Giangrande PL, et al. Mortality before and after HIV infection in the complete UK population of haemophiliacs. UK Haemophilia Centre Directors' Organisation. Nature (1995) 377:79–82.[CrossRef][Medline]
  17. Goedert JJ. Prevalence of conditions associated with human immunodeficiency hepatitis virus infections among persons with haemophilia, 2001 –2003. Haemophilia (2005) 11:516–28.[CrossRef][ISI][Medline]
  18. Zhang M, Rosenberg PS, Brown DL, et al. Correlates of spontaneous clearance of hepatitis C virus among people with hemophilia. Blood (2006) 107:892–7.[Abstract/Free Full Text]
  19. Goedert JJ, Eyster ME, Lederman MM, et al. End-stage liver disease in persons with hemophilia and transfusion-associated infections. Blood (2002) 100:1584–9.[Abstract/Free Full Text]
  20. Jones PK, Ratnoff OD. The changing prognosis of classic hemophilia (factor VIII "deficiency"). Ann Intern Med (1991) 114:641–8.[ISI][Medline]
  21. Becker NG, Marschner IC. A method for estimating the age-specific relative risk of HIV infection from AIDS incidence data. Biometrika (2003) 80:165–78.[CrossRef]
  22. Silverman BW, Jones MC, Wilson JD, et al. A smoothed EM approach to indirect estimation problems, with particular reference to stereology and emission tomography. J R Stat Soc [Ser B] (1990) 55:271–324.
  23. Aach RD, Szmuness W, Mosley JW, et al. Serum alanine aminotransferase of donors in relation to the risk of non-A, non-B hepatitis in recipients: the Transfusion-transmitted Viruses Study. N Engl J Med (1981) 304:989–94.[Abstract]
  24. Alter HJ, Holland PV, Purcell RH, et al. Posttransfusion hepatitis after exclusion of commercial and hepatitis-B antigen-positive donors. Ann Intern Med (1972) 77:691–9.[ISI][Medline]
  25. Hoover DR, Munoz A, Carey V, et al. The unseen sample in cohort studies: estimation of its size and effect. Multicenter AIDS Cohort Study. Stat Med (1991) 10:1993–2003.[ISI][Medline]
  26. Rosenberg PS, Goedert JJ. Estimating the cumulative incidence of HIV infection among persons with haemophilia in the United States of America. Stat Med (1998) 17:155–68.[CrossRef][ISI][Medline]
  27. Institute of Medicine Committee to Study HIV Transmission through Blood and Blood Products. HIV and the blood supply: an analysis of crisis decisionmaking (1995) Washington, DC: National Academy Press.
  28. Hoots WK. History of plasma-product safety. Transfus Med Rev (2001) 15:3–10.[CrossRef][ISI][Medline]
  29. Pneumocystis carinii pneumonia among persons with hemophilia A. MMWR Morb Mortal Wkly Rep (1982) 31:365–7.[Medline]
  30. Hasiba U, Eyster ME, Gill FM, et al. Liver dysfunction in Pennsylvania's multitransfused hemophiliacs. Dig Dis Sci (1980) 25:776–82.[CrossRef][ISI][Medline]
  31. Kroner BL, Rosenberg PS, Aledort LM, et al. HIV-1 infection incidence among persons with hemophilia in the United States and western Europe, 1978 –1990. Multicenter Hemophilia Cohort Study. J Acquir Immune Defic Syndr (1994) 7:279–86.[ISI][Medline]
  32. Busch M, Chamberland M, Epstein J, et al. Oversight and monitoring of blood safety in the United States. Vox Sang (1999) 77:67–76.[CrossRef][ISI][Medline]
  33. Turnbull BW. The empirical distribution function with arbitrarily grouped, censored and truncated data. J R Stat Soc [Ser B] (1976) 38:290–5.
  34. Cleveland WS. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc (1979) 74:829–36.[CrossRef][ISI]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
165/12/1443    most recent
kwm030v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Goedert, J. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Goedert, J. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?