Survival after transfusion in the Netherlands.

BACKGROUND: Cost-effectiveness analyses of blood safety interventions require estimates of the life expectancy after blood product transfusion. These are best derived from survival after blood transfusion, per age group and blood component type. STUDY DESIGN AND METHODS: In the PROTON (PROfiles of TransfusiON recipients) study transfusion recipient data was collected from a hospital sample covering 28% of the total blood use between 1996 and 2006 in the Netherlands. The dataset includes date of transfusion, blood component type transfused and recipient identification details. PROTON data were individually matched to mortality data of the Netherlands. Survival after first transfusion and after any transfusion was calculated, per blood component type and age group. PROTON mortality rates were compared to mortality rates in the general population. The results were used to estimate survival beyond the study period and to estimate life expectancy after transfusion. RESULTS: Of all 2,405,012 blood product transfusions in the PROTON dataset, 92% was matched to the national Dutch Municipal Population Register, which registers all deaths. After 1 year, survival after any transfusion was 65·4%, 70·4% and 53·9% for RBC, FFP and PLT respectively. After 5 years, this was 46·6%, 58·8% and 39·3% for RBC, FFP and PLT, respectively. Ten years after transfusion, mortality rates of recipients are still elevated in comparison with the general population. CONCLUSION: Mortality rates of transfusion recipients are higher than those of the general population, but the increase diminishes over time. The mortality rates found for the Netherlands are lower than those found in comparable studies for other countries.

The PROTON study: profiles of blood product transfusion recipients in the Netherlands.

BACKGROUND: Transfusion recipient data are needed for correct estimation of cost-effectiveness in terms of recipient outcomes after transfusion. Also, such data are essential for monitoring blood use, estimation of future blood use and benchmarking. STUDY DESIGN AND METHODS: A sample of 20 of 93 Dutch hospitals was selected. Datasets containing all blood product transfusions between 1996 and 2006 were extracted from hospital blood bank computer systems, containing transfusion date, blood product type and recipient characteristics such as gender, address, date of birth. The datasets were appended and matched to national hospitalization datasets including primary discharge diagnoses (ICD-9). Using these data, we estimated distributions of blood recipient characteristics in the Netherlands. RESULTS: The dataset contains information on 290,043 patients who received 2,405,012 blood products (1,720,075 RBC, 443,697 FFP, 241,240 PLT) from 1996 to 2006. This is 28% of total blood use in the Netherlands during this period. Comparable diagnosis and age distributions of all hospitalizations indicate included hospitals to be representative, per hospital category, for the Netherlands. Of all red blood cells (RBC), fresh-frozen plasma (FFP) and platelets (PLT), respectively 1.7%, 2.5% and 4.5% were transfused to neonates. Recipients of 65 years or older received 57.6% of RBC, 41.4% of FFP and 29.0% of PLT. Most of the blood products were transfused to patients with diseases of the circulary system (25.1%) or neoplasms (22.0%). CONCLUSION: Transfusion data from a limited sample of hospitals can be used to estimate national distributions of blood recipient characteristics.

Monitoring viral incidence rates: tools for the implementation of European Union regulations.

BACKGROUND AND OBJECTIVES: European legislation requires manufacturers of plasma products to report epidemiological data on human immunodeficiency virus, hepatitis B virus and hepatitis C virus in donor populations. The incidence rates of such infections are directly related to the risk of infection transmission. We propose two statistical tests to evaluate these incidence rates. MATERIALS AND METHODS: Infection data of the four Dutch blood collection centres from 2003 through 2006 were analysed. For transversal comparison of centres and detection of increased incidence rates, a new statistical test was developed (outlier test). For longitudinal detection of trends in incidence rates, a generic test for trend is proposed. The power and risk of non-detection are evaluated for both tests. RESULTS: Application of the outlier test did not reveal any significantly increased incidence rates among centres in The Netherlands. The test for trend showed no significant increase in incidence rates in individual centres, but on national level a statistically significant increase in hepatitis C virus incidence was observed (P-value of 0.01). CONCLUSION: The proposed tests allow signalling of outlier centres and trends in incidence rates both at individual centre and at national levels. Graphical support and the use of as much relevant historical data as possible is recommended. The statistical tests described are generic and can be applied by any blood establishment and plasma fractionation institute.

Paying for blood donations: still a risk?

It is presently disputed whether studies indicating a higher risk of infectious diseases among paid blood donors are lessons of the past, or still hold relevance. Comparative studies published between 1968 and 2001 were assessed for a possible trend of change in the relative risk for infectious disease markers between paid and unpaid blood or plasma donors. Studies reporting that paid donors had lower risk were found, but most studies, including recent ones, continued to report that paid donors have higher rates of infectious disease markers than unpaid donors. By log-linear regression analysis of the relative risk estimates for infectious disease markers among paid and unpaid donors from 28 published data sets, evidence was not found to indicate that the difference in risk for infectious disease markers between paid donors and unpaid donors had diminished over time (P = 0.128, not significant). Paid donors are still more likely than unpaid donors to donate blood in the period during which infectious donations escape detection by blood-screening tests (the "window-period"). Therefore, paid donations have a higher risk that labile blood components (such as red blood cells and platelets) are infected. Additional safety measures for handling plasma donations, and the preparation, purification and viral-inactivation steps employed for the production of plasma derivatives, may render the difference in infectious disease marker rates in donors irrelevant for plasma products. However, not all viruses are inactivated and paid donors were repeatedly found to have higher frequencies of markers for emerging agents. In a quality system, critical steps of the process should be addressed, and selection of the donor population is one of the first steps in this process. It is advised that blood establishments present yearly reports (with complete and raw data) to authorities on the incidence and prevalence of infectious disease markers among their donors as an ongoing surveillance on the "quality" of their donor populations. Paid blood or plasma donors still have higher rates for infectious disease markers than unpaid donors.

International collaborative study on the second EUROHEP HCV-RNA reference panel.

Eighty-six laboratories participated in a collaborative study and tested the second EUROHEP HCV-RNA reference panel. The coded panel comprised 4 HCV-RNA positive plasma samples (one weak positive), 6 HCV-RNA negative plasma samples and two dilution series of HCV-RNA genotype 1 and 3 plasma standards. The 86 laboratories submitted 136 coded data forms for evaluation. Of these data sets 99 were tested using a PCR assay developed in-house, 28 using a commercially available HCV-PCR test (AMPLICOR, Roche Diagnostic Systems) and 9 using other amplification methods. Twenty-two data forms (16%) had faultless results, 39 (29%) missed the weak positive sample only and 75 data sets (55%) had false positive and/or false negative results. Participants using the commercial HCV-PCR test tended to reach a sufficient quality score more often than investigators using assays developed in-house (64% versus 45%, P = 0.11). The UNG system in the commercial HCV-PCR test did not prevent five laboratories generating false-positive results in the 6 HCV-RNA negative samples. Among the laboratories with satisfactory results, up to 10000-fold differences in sensitivity were observed in the dilution series. The 50% and 90% laboratories detection endpoints in the dilution series of the HCV genotype 1 plasma standard were approximately 600 genome equivalents per ml (geq/ml) and 7750 geq/ml according to a standard applied in a signal amplification assay (bDNA, Chiron). Our results suggest that the detection efficiency for genotype 3 by commercial HCV-RNA assays is lower than by the in-house assays. Internationally characterized HCV-RNA plasma standards should be made available for validation and standardization of HCV-RNA assays for HCV diagnosis and virological safety testing of blood products.

Subsets of CD34+ cells and rapid hematopoietic recovery after peripheral-blood stem-cell transplantation.

PURPOSE: To study whether there is a relationship between transplanted cell dose and rate of hematopoietic recovery after peripheral-blood stem-cell (PBSC) transplantation, and to obtain an indication whether specific subsets of CD34+ cell populations contribute to rapid recovery of neutrophils or platelets. PATIENTS AND METHODS: Based on data from 59 patients, we calculated for each day after PBSC transplantation the dose of CD34+ cells that resulted in rapid recovery of either neutrophils or platelets in the majority (> 70%) of patients. Using dual-color flow cytometry, subsets of peripheral-blood CD34+ cells were quantified and the numbers of CD34+ cells belonging to each of the reinfused subsets correlated with hematopoietic recovery following high-dose chemotherapy. RESULTS: The calculated threshold values with a high probability of engraftment showed a steep dose-effect relationship between CD34+ cell dose and time to recovery of both neutrophils or platelets. Predominantly CD34+ cells with the phenotype of myeloid precursors were mobilized. A minority of CD34+ cells expressed the erythroid and megakaryocytic lineage-associated antigens and a low but distinct population of CD34+ cells expressed antigens associated with multipotent stem cells. Analysis showed that the number of CD34+CD33- cells (r = -.74, P < .05), as well as the number of CD34+CD41+ cells (r = -.81, P < .005), correlated significantly better with time to neutrophil and platelet recovery, respectively, than with the total number of CD34+ cells (r = -.55 and r = -.56, respectively). CONCLUSION: The numbers of CD34+CD33- cells and CD34+CD41+ cells may help to predict short-term repopulation capacity of PBSCs, especially when relatively low numbers of CD34+ cells per kilogram are reinfused.

Expression of adhesion molecules on CD34+ cells: CD34+ L-selectin+ cells predict a rapid platelet recovery after peripheral blood stem cell transplantation.

Adhesion molecules play a role in the migration of hematopoietic progenitor cells and regulation of hematopoiesis. To study whether the mobilization process is associated with changes in expression of adhesion molecules, the expression of CD31, CD44, L-selectin, sialyl Lewisx, beta 1 integrins very late antigen 4 (VLA-4) and VLA-5, and beta 2 integrins lymphocyte function-associated 1 and Mac-1 was measured on either bone marrow (BM) CD34+ cells or on peripheral blood CD34+ cells mobilized with a combination of granulocyte colony-stimulating factor (G-CSF) and chemotherapy. beta 1 integrin VLA-4 was expressed at a significantly lower concentration on peripheral blood progenitor cells than on BM CD34+ cells, procured either during steady-state hematopoiesis or at the time of leukocytapheresis. No differences in the level of expression were found for the other adhesion molecules. To obtain insight in which adhesion molecules may participate in the homing of peripheral blood stem cells (PBSCs), the number of CD34+ cells expressing these adhesion molecules present in leukocytapheresis material was quantified and correlated with hematopoietic recovery after intensive chemotherapy in 27 patients. The number of CD34+ cells in the subset defined by L-selectin expression correlated significantly better with time to platelet recovery after PBSC transplantation (r = -.86) than did the total number of CD34+ cells (r = -.55). Statistical analysis of the relationship between the number of CD34+L-selectin+ cells and platelet recovery resulted in a threshold value for rapid platelet recovery of 2.1 x 10(6) CD34+ L-selectin+ cells/kg. A rapid platelet recovery (< or = 14 days) was observed in 13 of 15 patients who received > or = 2.1 x 10(6) CD34+ L-selectin+ cells/kg (median, 11 days; range, 7 to 16 days), whereas 10 of 12 patients who received less double positive cells had a relative slow platelet recovery (median, 20 days; range, 13 to 37 days). The L-selectin+ subpopulation of CD34+ cells also correlated better with time to neutrophil recovery (r = -.70) than did the total number of reinfused CD34+ cells (r = -.51). However, this latter difference failed to reach statistical significance. This study suggests that L-selectin is involved in the homing of CD34+ cells after PBSC transplantation.

T cell function in vitro is an independent progression marker for AIDS in human immunodeficiency virus-infected asymptomatic subjects.

The predictive value of low T cell reactivity to CD3 monoclonal antibodies for development of AIDS was evaluated and compared with low CD4+ cell numbers and the presence of syncytium-inducing human immunodeficiency virus (HIV) variants in 122 seropositive asymptomatic homosexual men for 4.5 years. Low T cell reactivity was a strong predictor for progression to AIDS in a multivariate proportional hazards analysis using these markers as covariates at entry and as time-dependent covariates. The combination of the three markers was associated with development of AIDS in 6 of 7 men within 15 months. In contrast, the group that lacked any of these markers had a very low risk (11%) for developing AIDS. In groups with one or two of these three markers, progression rates were 33% and 66%, respectively. These data demonstrate that measurement of T cell function in vitro is of value for staging of HIV infection and may be useful for monitoring therapy.

Comparison of quantitative cDNA-PCR with the branched DNA hybridization assay for monitoring plasma hepatitis C virus RNA levels in haemophilia patients participating in a controlled interferon trial.

The branched DNA (bDNA) assay was compared with a semi-quantitative cDNA-polymerase chain reaction (cDNA-PCR) assay for monitoring HCV RNA levels in plasma in 17 haemophilia patients participating in a controlled alpha-interferon trial. Good correlation between the HCV RNA levels as detected by the two assays was observed, with a correlation co-efficient of 0.83 (P < 0.0001) and 0.90 (P < 0.0001) at week 0 and 24, respectively. Hepatitis C virus RNA (HCV RNA) levels could be assessed with the bDNA assay in 14/17 (82 percent) HCV cDNA-PCR positive pretreatment samples. The bDNA assay apparently failed to detect low viral titres. Interferon treated patients (n = 11) showed either a complete response, being a large reduction in HCV RNA level to below the detection limit of the HCV cDNA-PCR assay (6/11) or no significant reduction in HCV RNA level (5/11). A "partial" virological response was not observed. The changes in HCV RNA plasma levels in non-responders during interferon (IFN) treatment were similar to the (small) natural fluctuations in viral load observed in controls (untreated patients). Although the bDNA assay was not as sensitive as cDNA-PCR, given its user friendliness and quantitative results, it is concluded that it is a useful test for monitoring HCV RNA levels in patients treated with interferon. However, patients who are non-reactive in the bDNA assay have to be retested by cDNA-PCR because low viral titres are not detected by the bDNA assay.

Enhanced sensitivity of a second generation ELISA for antibody to hepatitis C virus.

A second generation ELISA for combined detection of antibodies to three hepatitis C virus (HCV) recombinant proteins, i.e. C100, C33c and core, was compared with a first generation anti-HCV ELISA in which only antibodies to C100 are detected. The results of the ELISAs were evaluated in 225 haemophilia patients (panel A) and 44 patients with non-A, non-B (NANB) hepatitis (panel B). HCV infection was established by cDNA-polymerase chain reaction (PCR) (in panel B only) and by studying the anti-HCV reaction patterns in 4 separate ELISAs for detection of antibodies to the recombinant proteins C100, C33c, core and a combination of two synthetic peptides sp67/65 derived from the C100 region. The sensitivity for the detection of HCV infection had increased from 0.92[95% confidence interval (CI): 0.87-0.95] to 1.00 (95% CI: 0.89-1.00) in haemophiliacs and from 0.84 (95% CI: 0.66-0.95) to 1.00 (95% CI: 0.89-1.00) in NANB hepatitis patients when the second generation ELISA was used instead of the first generation ELISA. Concurrently the chance of a false negative result was reduced in panel A and B from 0.37 to 0 and from 0.28 to 0, respectively. Analysis of anti-HCV reaction patterns revealed that 172 of 206 (83.5%) anti-HCV ELISA-reactive haemophilia patients had antibodies to all 4 antigens tested. In the NANB hepatitis patients 18 of 31 (58.1%) anti-HCV ELISA-reactive subjects reacted with 4 antigens. In the PCR tested panel of NANB hepatitis patients 2 subjects who showed antibody reactivity to only one antigen and 5 patients with reactivity to 2 antigens were PCR-positive.(ABSTRACT TRUNCATED AT 250 WORDS)

Variations in the T-cell repertoire against HLA antigens in humans.

Allospecific anti-HLA class I antigen cytotoxic T-lymphocyte precursor frequencies (CTLpf) have been estimated in peripheral blood of healthy blood donors with responder stimulator combinations mismatched for one HLA-A,B antigen. The CTLpf ranged from 1 in 400 to 1 in 10,000, with most frequent values of 1 in 600 to 4000. The following observations were made: (1) CTLpf against the same HLA antigen vary among different responders; (2) CTLpf of one responder against various HLA antigens may be different; (3) "narrow" responders produce cytotoxic T lymphocytes that recognize only the private (stimulator) alloantigen, while "broad" responders produce mainly broadly cross-reactive cytotoxic T lymphocytes with public specificity. Split-well analysis shows that very few cytotoxic T lymphocytes of "broad" responders recognize the private alloantigen only. These individual variations are not dependent on the HLA phenotype, because they also occurred in unrelated HLA-identical responders stimulated against the same mismatched stimulator cells.