Emeritus Faculty-Med Ctr Line, Pathology
MD, Harvard Medical School, Medicine (1979)
BS, University of Maryland, Biochemistry (1975)
Transfusion-transmitted infections and donor screening for infectious diseases. National policies for blood banks. Enhancement of transfusion safety and effectiveness, with a focus on quality assurance in blood banking and transfusion therapy; transfusion medicine education; pediatric and adult transfusion therapy.
Hemoglobinuria was observed after packed red blood cell transfusion in a series of patients at our pediatric treatment center. Laboratory testing was suggestive of intravascular hemolysis with no support for an immunohematologic process.We investigated these adverse events to define a quality improvement plan and to prevent future hemolytic adverse events. Multiple factors were investigated, and the only change identified was the implementation of a new infusion pump (Pump A) that replaced a previous model (Pump B).In vitro pump analyses, a retrospective review of urinalyses, and prospective urinalysis and nursing surveillances were also performed.In in vitro analysis of the pumps, irradiated units with higher hematocrit at a low flow rate through Pump A had a greater than thirty-fold increase in free hemoglobin from baseline compared to minimal free hemoglobin changes seen with Pump B. Irradiated units with a lower hematocrit had a minimal change in free hemoglobin from baseline with both Pumps A and B at either low or high flow rate. Subsequently, only units with lower hematocrits were issued for transfusion of pediatric patients, and Pump A was replaced by Pump B in the outpatient unit. Retrospective and prospective surveillances found no additional unexplained cases of gross hemoglobinuria associated with transfusion.The investigation determined that infusion of higher hematocrit units using a specific commercial pump was associated with mechanical hemolysis. The change to units with lower hematocrit through an alternative pump has been an effective corrective action to date.
View details for PubMedID 25887866
BACKGROUND: The Gerbich (Ge) blood group system consists of 11 antigens carried on red blood cell (RBC) membrane glycophorins C and D; of these, Ge:3 antigen is of high prevalence, and the anti-Ge3 is found to be clinically significant. CASE REPORT: A 34-week neonate born to a Hispanic mother with anti-Ge3 developed late-onset hemolysis with hyperbilirubinemia and was successfully treated with transfusions from her mother. Relevant clinical findings and laboratory results for this case are summarized and compared to three other previously reported cases; all babies were born from a mother of Hispanic ethnicity. CONCLUSION: Hemolytic disease of the fetus and new born associated with anti-Ge3 is rare but should be considered when working up a broadly reactive RBC antibody screen in women of Hispanic ethnicity. Early identification of pregnant women with anti-Ge3 is recommended for prenatal transfusion planning and close monitoring of the newborn infant for evidence of late-onset anemia.
View details for DOI 10.1111/trf.12027
View details for Web of Science ID 000325374300011
The Trima Accel displays a "verify WBCs" message if the plateletpheresis product (PLT) may not be leukoreduced (LR). Most blood banks require sensitive white blood cell (WBC) testing of these PLTs by flow or Nageotte. We evaluated how often these PLTs were non-LR by European or US Food and Drug Administration (FDA) criteria and whether sensitive WBC testing is necessary.Phase?1 reviewed the frequency of this message with various procedure types and the flow WBC results for PLTs with or without the message. Phase?2 assessed how many FDA LR failures were detectable by a hematology analyzer. In Phase?3, PLTs were managed by hematology analyzer results.In Phase?1, 3.8% of PLT-only and 11.1% of PLT-plasma collections had the "verify WBCs" message. Only 1% of "verify" PLTs contained more than 1?×?10(6) WBCs and only 0.5% were FDA LR failures. In Phase?2, 10 of 670 "verify" PLTs and one nonflagged PLT were FDA LR failures. Six of 11 LR failures had hematology analyzer WBC concentrations of 0.4?×?10(9) /L or higher. In Phase?3, "verify" PLTs were allowed in inventory if hematology analyzer WBC concentration was below 0.4?×?10(9) /L; inventory quality control showed no FDA LR failures by flow. Trima Version 6.0 software lowered the "verify" message frequency in PLT-plasma procedures but not in PLT-only procedures.Four percent of Trima PLT collections have the "verify WBCs" message but almost all of these are LR by European and FDA criteria. Fifty percent of FDA LR failures were detectable by a hematology analyzer. Sensitive WBC testing of all "verify WBCs" PLTs may not be necessary to satisfy LR quality assurance requirements.
View details for DOI 10.1111/j.1537-2995.2011.03398.x
View details for Web of Science ID 000303876900011
View details for PubMedID 22023335
Immune refractoriness to platelet (PLT) transfusion is primarily due to HLA antibody. Patients at our institution are identified as refractory due to HLA by a Luminex-based immunoglobulin (Ig)G-single-antigen-bead (SAB) assay, but in highly sensitized patients, antigen-negative compatible donors cannot be found due to the high sensitivity of the IgG-SAB method. We developed an assay that detects only HLA antibodies binding the first complement component (C1q). We hypothesized that the C1q-SAB method might be more relevant than the IgG-SAB method because the antibodies identified may activate the complement cascade causing PLT destruction.Thirteen highly sensitized refractory patients received 177 PLT units incompatible by the IgG-SAB method. They were retrospectively retested by the C1q-SAB method. Calculated percent reactive antibody (CPRA) and HLA antibody specificities were compared between the two methods and corrected count increment (CCI) values were analyzed. Additionally the impact of ABO compatibility on CCI responses was evaluated.The mean CPRA value was significantly lower by C1q-SAB (60%) than by IgG-SAB (94%; p < 0.05). Patients showed significantly better CCI (10.6 × 10(9) ± 0.8 × 10(9) /L) with C1q-compatible (n = 134) than with C1q-incompatible PLTs (n = 43) (2.5 × 10(9) ± 0.9 × 10(9) /L/m(2) ; p < 0.0001). ABO compatibility did not significantly impact the CCI values (p < 0.0001). Our results show that 75% of PLT units previously considered incompatible were actually compatible.For highly refractory patients to PLT transfusion, the C1q-based SAB binding assay may be a better method for identifying clinically relevant HLA antibodies and selecting PLT units that will result in acceptable CCI.
View details for DOI 10.1111/j.1537-2995.2011.03194.x
View details for Web of Science ID 000298340300015
View details for PubMedID 21615749
Blood centers and hospital transfusion services are challenged with maintaining an adequate platelet (PLT) inventory to minimize the number of outdated units without risking a major shortage. A novel approach to inventory management was established at our institution through a collaboration between the Stanford University Medical Center (SUMC) Transfusion Service, the Stanford Blood Center (SBC), and the Department of Management Science and Engineering.An analysis of the supply chain performance between SBC and SUMC Transfusion Service was performed. First, the interaction between processes, such as blood collection, rotation, and inventory management, was studied. Second, changes were implemented based on the recommendations from the analysis team. Finally, a postanalysis was performed reflecting on the improvement of the operations between SUMC and SBC.A comprehensive data analysis of the PLT supply chain allowed the identification of three series of improvements to be implemented: 1) on SBC's PLT collection, 2) on SBC's rotation process, and 3) on the PLT inventory management policy at SUMC. A postimplementation analysis showed a reduction in the overall PLT outdate rate from 19% in the first quarter of 2006, down to 9% in the third quarter of 2008.A multidisciplinary effort among SUMC Transfusion Service, SBC, and experts in supply chain management resulted in a process improvement, which reduced the rate of PLT outdate at both SBC and SUMC Transfusion Service down to 9%, with a significant cost reduction of more than half a million dollars per year.
View details for DOI 10.1111/j.1537-2995.2009.02236.x
View details for Web of Science ID 000270430100008
View details for PubMedID 19538430
Minipool (MP) screening for West Nile virus (WNV) RNA may fail to detect presumptive viremic donations (PVDs) detectable by individual donation screening (IDS). Most blood centers switch collection regions to IDS when PVD detection by MP screening reaches a certain frequency. Use of IDS for all donations during WNV season was assessed during a clinical trial of the Roche cobas TaqScreen WNV test. Also evaluated was whether PVD detection reliably identifies regions that should be targeted for IDS.Test results, deviation reports, and service records were reviewed for 13.5 weeks of IDS in 2006 and 11.5 weeks of IDS in 2007. Numbers of PVDs and clinical WNV cases were obtained from public health and AABB Web sites and regional donor centers.Approximately 1000 donations were tested per week divided in six test runs. Each run required 1.2 shifts of technologists plus volunteers. A total of 7.2 percent of samples were initially unreportable in 2006 and 4.8 percent in 2007. Of 26,952 donations screened by IDS, none were reactive for WNV. A comparison of PVD and clinical case reports indicates that PVD detection in areas with intermediate or high clinical case prevalence may not reach commonly used criteria for triggering testing to IDS.Seasonal IDS was feasible using the cobas TaqScreen WNV test on the s 201, although staffing was impacted and a relatively high number of samples required retesting because of error messages. Seasonal IDS utilizing this highly specific assay may be a reasonable alternative to IDS triggered by regional PVD detection.
View details for DOI 10.1111/j.1537-2995.2008.01715.x
View details for Web of Science ID 000257386700029
View details for PubMedID 18466179
Two HCV antibody tests (EIA 2.0 [EIA2], Abbott; and the Version 3.0 ELISA [EIA3], Ortho) are currently licensed for screening of US blood donors. Testing of donors for HCV RNA allows comparison of the sensitivities of the two antibody-screening assays.All allogeneic blood donations at 13 US test sites were screened for HCV RNA by testing plasma minipools using an investigational assay (COBAS AmpliScreen HCV test, v2.0, Roche Molecular Systems). Some sites screened for HCV antibody by EIA2 and some used EIA3. The frequency of RNA-positive and antibody-negative (RNA-pos and Ab-neg) donations among donors screened by each antibody assay was compared. Antibody appearance was assessed in a donor follow-up study.A total of 5.51 x 10(6) donations were screened for HCV RNA. Of these, 2.27 million were screened for antibody by EIA2, and 3.24 million by EIA3. Twenty-three donations were HCV RNA-pos and Ab-neg. The frequency of RNA-pos and Ab-neg donations was higher among donations screened by EIA2 (1 in 134,000), compared to those screened by EIA3 (1 in 540,000) (p = 0.001). Of the 17 RNA-pos and Ab-neg donations identified by test sites that used EIA2, 14 were retested by EIA3 and 10 (71%) were reactive. Most RNA-pos and Ab-neg donors appear to be in the process of seroconversion. Donors that were initially EIA2-negative and EIA3-reactive showed a more prolonged pattern of seroconversion compared to those that were initially nonreactive by both antibody assays. Four donors were EIA2-negative, EIA3-reactive, and RIBA-indeterminate (c33c) for at least 90 days, 1 for more than 317 days.EIA3 would have detected the majority of RNA-positive donations missed by EIA2. Some RNA-positive donors are EIA2-negative and EIA3-reactive for a prolonged period of time.
View details for Web of Science ID 000179161000018
View details for PubMedID 12421226
In 1998, the FDA recommended look-back for HCV. The recommendation was initially limited, however, to donors who reacted on a multiantigen HCV screening test and to components collected since January 1, 1988. A lookback program was extended to include donors who reacted on the first-generation (single-antigen) HCV screening test and who were positive on a supplemental assay (RIBA-1 or -2) and all components for which transfusion records could be found (back to 1978).The yield of the incremental lookback programs was compared to that originally recommended by the FDA by comparing the number of newly identified HCV-positive recipients in each program. The results of lookbacks were reviewed on 385 blood components for which 314 transfusion recipients were identified.Of the 135 recipients in the FDA program, 70 percent were dead, 28 percent were living and notified, and 2 percent could not be located. In the incremental programs, there were 179 recipients, of whom 80.4 percent were dead, 16.2 percent were living and notified, and 3.4 percent could not be located. Most adult recipients were dead (81%), but the majority of pediatric recipients were alive (57%); 76 percent of tested recipients were HCV seropositive, with no significant difference between programs. One-half of test-positive recipients in each program were newly identified through the lookback program. Seven of the 20 newly identified HCV-positive recipients were found through the incremental programs. The yield, defined as newly detected HCV cases per total number of recipients, was 9.6 percent for the FDA and 3.9 percent for the incremental programs. This difference was significant (p = 0.04).The yield of both programs was limited by the high percentage of recipients who had died. Pediatric recipients were more likely to be living at the time of notification. The incremental program was less efficacious than the FDA program in identifying newly HCV-positive recipients, but one-third of the newly detected HCV cases were identified through the incremental program.
View details for Web of Science ID 000165492600018
View details for PubMedID 11099671
Laboratory quality can be defined as "doing the right thing right." The chance of doing the right thing right is greater if basic process control systems are implemented in laboratories. The elements of process control were reviewed, and practical suggestions for organizing procedures, training, determining competency, maintaining equipment, and implementing self-assessment and process improvement activities were offered. Although process control can improve the reliability of internal operations, the quality of patient care can be further improved by extending this level of control outward to include interaction with patients. Finally, to expand the concept of quality, quality activities must be extended toward a multidisciplinary model of coordinated care.
View details for Web of Science ID A1997WQ90000008
View details for PubMedID 9124229
Trypanosoma cruzi, the protozoan parasite that causes Chagas'disease, is endemic in Central and South America and in Mexico. Risk of infection is related to exposure to insects harboring T. cruzi or to the transfusion of blood from an infected donor. Large numbers of immigrants from endemic areas reside in California, but the frequency with which persons at risk for T. cruzi contribute to the blood supply there is not known.A questionnaire was used to survey donors in 18 California donor centers for risk factors for T. cruzi infection.Otherwise eligible allogeneic blood donors (n = 17,521) completed questionnaires. Of this group, 427 (2.4%) had lived in endemic areas for more than 1 year, and 39 of these donors had lived in dwellings with mud walls or thatched roofs. Sixteen donors had received transfusions in endemic areas. Six donors gave a history of Chagas' disease. Fifty-seven donors (0.33% of total) had at least one risk factor for T. cruzi infection. Donors at risk for T. cruzi were found in all 18 centers studied, at a median prevalence of 1 per 340 donors.Donors at risk for T. cruzi are contributing to the blood supply throughout California. Further consideration should be given to donor screening for this transfusion-transmissible infection.
View details for Web of Science ID A1996UE23400008
View details for PubMedID 8604507
Shortly after the first cases of AIDS were reported in 1981, it became apparent that this disease was caused by a blood-borne infectious agent that could be transmitted by blood transfusion. Early reports documented a reduced ratio of CD4+ to CD8+ T cells not only in AIDS patients but also in likely carriers of the AIDS pathogen. On this basis, from July 1983 to June 1985, our blood center utilized flow cytometry to test each donated unit for the ratio of CD4 to CD8 T cells; we did not transfuse blood from donors with CD4/CD8 < 0.85. Despite available data supporting the utility of this or other surrogate blood tests to screen blood donors, the vast majority of blood banks did not initiate blood donor testing for AIDS until 1985, following the discovery of HIV and development of commercial HIV antibody tests. Retrospective analysis indicates that donor screening with surrogate markers would have removed the majority of AIDS-contaminated units from the blood supply and prevented a substantial fraction of the estimated 10,000 cases of transfusion-transmitted AIDS in the United States. In this report, we review the events that led to our decision to initiate blood donor testing prior to the identification of HIV, the results of such testing, the consequences of the decision by most blood banks not to initiate such testing, the current status of the blood supply with respect to HIV, and steps that can be taken in the future to protect the blood supply from new pathogens.
View details for Web of Science ID A1995QV29000008
View details for PubMedID 7612221