T R A N SO FRUI SG II ON NA LP RAARCTTIICCLEE( C M E ) CME
Provision of KEL1-negative blood to obstetric patients: a 3-year single-institution retrospective review Kerry L. O’Brien,1,2* Yeowon A. Kim,3* Richard L. Haspel,1,2 and Lynne Uhl1,2
BACKGROUND: KEL1 alloimmunization is a major cause of hemolytic disease of the fetus and newborn (HDFN). While select countries have guidelines for preventing transfusion-associated KEL1 alloimmunization, the United States does not. Beth Israel Deaconess Medical Center instituted a policy in April 2009 whereby women not more than 50 years of age on the obstetric service were transfused KEL1-negative red blood cells (RBCs). We sought to determine compliance and impact for prevention of KEL1 alloimmunization and HDFN. STUDY DESIGN AND METHODS: All women not more than 50 years of age without anti-K transfused RBCs during an obstetric admission from April 9, 2009, to April 9, 2012, were identified (227). Adherence to policy, factors contributing to nonadherence, and subsequent impact were evaluated. For comparison, all cases of anti-K detection in women not more than 50 years of age admitted to nonobstetric services and all cases of transfusion-associated KEL1 alloimmunization in women not more than 50 years of age during the 10 years prior were identified. RESULTS: Eighty-four percent received only KEL1negative units. Three (1.3%) women not more than 50 years of age on the obstetric service were identified with anti-K, while 17 (1.5%) women not more than 50 years of age on nonobstetric services had anti-K detected; only five of 20 had a prior RBC transfusion. In the 10 years prior, there were 27 cases of transfusionassociated KEL1 alloimmunization in women not more than 50 years of age. There were no cases of KEL1 HDFN in either period. CONCLUSION: Although the findings demonstrate feasibility of providing KEL1-negative RBCs to women of childbearing potential, evidence for clinical benefit is lacking. The low prevalence of KEL1 in blood donors, the lack of significant differences in alloimmunization rates, and no cases of HDFN during the study period questions the clinical benefit of such a policy.
T
he Kell blood group system is the third most polymorphic blood group system and its antigens are highly immunogenic.1 Antibodies to the KEL1 antigen (also known as anti-K) are the most common red blood cell (RBC) antibodies outside the ABO and Rh systems.2 Anti-K is clinically significant and can cause severe transfusion reactions and hemolytic disease of the fetus and newborn (HDFN). While the rate of anti-D–related maternal alloimmunization and HDFN has decreased with RhIG prophylaxis, the incidence and frequency of KEL1 alloimmunization has increased and surpassed that of anti-D in some populations studied.3,4 Most recent data show the frequency of anti-KEL1 to be one per 1000 pregnant females5 and the incidence of KEL1-related HDFN to be about one in 40,000 births.6 The management of pregnancies affected by anti-K differs from that of anti-D, as antibody titers may not correlate
ABBREVIATIONS: BIDMC = Beth Israel Deaconess Medical Center; HDFN = hemolytic disease of the fetus and newborn; IQR = interquartile range. From the 1Department of Pathology, Beth Israel Deaconess Medical Center; 2Department of Pathology, Harvard Medical School; and 3Joint Program in Transfusion Medicine, Harvard Medical School, Boston, Massachusetts. Address reprint requests to: Kerry L. O’Brien, MD, Blood Bank, Beth Israel Deaconess Medical Center, YA-309, 330 Brookline Avenue, Boston, MA 02215; e-mail: [email protected]. *Both authors contributed equally to this manuscript. This work was conducted with support from Harvard Catalyst | The Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health Award 8UL1TR000170-05) and financial contributions from Harvard University and its affiliated academic health care centers. Received for publication April 17, 2014; revision received July 9, 2014, and accepted July 9, 2014. doi: 10.1111/trf.12814 © 2014 AABB TRANSFUSION **;**:**-**. 2015;55:599–604.
Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION599 1
O’BRIEN ET AL.
with the extent of fetal disease and severe disease can occur in the setting of low titers.7,8 This has been explained by the differences in the suggested mechanisms of anti-D– and anti-K–related HDFN: while hemolysis is the predominant process in anti-D alloimmunization, anti-K cause fetal anemia primarily by suppression of erythropoiesis.9-11 The etiology of anti-KEL1 identified during pregnancy is predominantly transfusion associated and not a consequence of immunization after transplacental hemorrhage from the fetus,6 and while there is a wide range (30% to >50%) in the estimated percentage of maternal Kell alloimmunization cases attributed to transfusion,8,12-16 even the most conservative estimates suggest that a substantial fraction of maternal KEL1 alloimmunization can be prevented by the provision of KEL1 antigen–negative blood. Such policies have been adopted by several countries, excluding the United States, and have reduced the proportion of alloimmunized women with a history of transfusion.17 However, there are no published data regarding the impact of such policies on the rate of anti-Kell–related HDFN. In April 2009 our institution implemented a policy to provide women aged not more than 50 years admitted to the inpatient obstetric service KEL1 antigen–negative RBCs if available. The policy’s goal was to avoid transfusion-related maternal KEL1 alloimmunization and HDFN. Since only 9% of the Caucasian (donor) population is positive for the KEL1 antigen, implementation of the policy at this time required modest changes in the RBC component inventory management system with minimal financial and operational impact. However, in light of the current constraints on health care expenditure and operational laboratory support, we were obliged to retrospectively review the clinical impact of this policy. Specifically, 3 years after implementation, we sought to retrospectively evaluate policy compliance, nonadherence, and the impact on prevention of transfusion-related maternal KEL1 alloimmunization and HDFN. It was considered that the data derived from this study could be used to inform the feasibility and effectiveness of broadly providing KEL1 antigen–negative RBCs to all women of childbearing potential in our institution and possibly in similar institutions in the United States.
MATERIALS AND METHODS The study was approved by the institutional review board of the Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts. BIDMC is a 650-bed tertiary care training institution offering a full range of services including an active obstetric service with more than 5000 births per year, a Level 3 nursery, allogeneic and autologous hematopoietic stem cell and organ (kidney and liver) transplant services, a cardiothoracic surgery practice, 77 600 TRANSFUSION Volume Volume March 2015 2 TRANSFUSION **,55, ** **
critical care beds, and a Level 1 trauma service. The hospital does not have a pediatric service, and all patients are at least 15 years of age except neonates. A computerized query of the Center for Clinical Computing (Boston, MA) blood bank database was performed to identify all women admitted to the inpatient obstetric service between April 9, 2009, and April 9, 2012, and who were transfused RBCs. For all patients, data on age at transfusion(s), underlying diagnosis, pregnancy history, the occurrence of KEL1-related HDFN, dates of transfusions, indication for transfusion, number of KEL1untested and KEL1 antigen–negative RBC units transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted via medical chart review and transfusion record review. Patients who had the following were excluded from analysis: those who had a history of anti-K or an inconclusive antibody screen upon presentation and age more than 50 years. As a comparison, all females not more than 50 years of age in whom anti-K was detected during this same time frame were identified. Pregnancy and RBC transfusion history were evaluated to determine possible mode of anti-K alloimmunization. We then performed a search for all women not more than 50 years of age who were transfused RBCs during this same 3-year period and who were not admitted to the obstetric service. Those women who had already made anti-K or who had inconclusive antibody panels on presentation were excluded from analysis. For all these patients, data on age at transfusion(s), underlying diagnosis, pregnancy history, the occurrence of KEL1-related HDFN, dates of transfusions, indication for transfusion, number of KEL1-untested and KEL1 antigen–negative RBC units transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted via medical chart review and transfusion record review. To determine the potential impact of the policy in preventing cases of maternal KEL1 alloimmunization and HDFN resulting from transfusion, a second computerized query was performed to evaluate for the number of transfusion-associated KEL1 alloimmunization and HDFN cases in women of childbearing age in the 10 years before the policy change. All patients who had an anti-K on antibody screen and were transfused RBCs between April 8, 1999, and April 8, 2009, were identified. For all patients, data on sex, age at transfusion(s), underlying diagnosis, pregnancy history, occurrence of KEL1-related HDFN, admission status at time of transfusion (i.e., inpatient versus outpatient, if the patient was admitted as an inpatient, the primary admitting service), dates of transfusions, indication for transfusion, number of KEL1-untested and KEL1 antigen–negative RBC units
KEL1-NEGATIVE RBCs RBCS TO TO OBSTETRIC OBSTETRIC PATIENTS PATIENTS
transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted as above. For the cases of transfusionassociated KEL1 alloimmunization, the source of alloimmunization was determined by assessing the temporal relationship between transfusion, results of antibody screens, and pregnancies. Transfusion was determined as the source of alloimmunization if the patient developed an anti-K after transfusion and did not have any documented pregnancies during the time interval between transfusion and anti-KEL1 positivity. For patients who developed anti-K after transfusion but also were pregnant during the time interval between change in anti-KEL1 status, the source of alloimmunization was designated as likely transfusion related, although pregnancy could not be ruled out. Antibody screens were performed using polyethylene glycol indirect antiglobulin test (Gamma Biologicals, Inc., Houston, TX) and a three-cell screen (R1R1, R2R2, rr; Immucor, Inc., Norcross, GA) between 1999 and 2007 according to standard institutional procedure, with agglutination reactions read macroscopically in tube and graded from 0 to 4+. Beginning 2008, solid-phase technology (Capture-R, Immucor, Inc.) and a two-cell screen (R1R1, R2R2) was used for antibody screens. Positive screening tests were followed by serologic investigation using a panel of 14 phenotypically defined RBCs (Immucor, Inc.) selected to identify the presence of anti-K. KEL1 antigen phenotyping of donors and recipients were performed by incubating donor or recipient RBCs with monoclonal IgM antibodies against KEL1 antigen (Immucor, Inc.) according to standard institutional procedure, with agglutination reactions read macroscopically in tube. Agglutination reactions 1+ and greater were classified as positive for the KEL1 antigen.
Statistical analysis Transfusion data are presented as medians with interquartile range (IQR) and their between-period comparisons were evaluated by the Wilcoxon rank-sum test. Dichotomous variables were reported as proportions (%) and were compared using the Fisher exact test. A twotailed p value of less than 0.05 was considered significant. All statistical analyses were performed using computer software (GraphPad Prism 6 for Windows, GraphPad Software, Inc., La Jolla, CA).
RESULTS Between April 9, 2009, and April 9, 2012, a total of 227 women aged not more than 50 years historically negative for anti-K were admitted to the inpatient obstetric service and were transfused RBCs. The vast majority of the
227 women aged ≤50 years and without history of anti-K were transfused RBCs between 4/9/09 and 4/9/12 2 with inconclusive antibody screen 225 women
188 women received only KEL1-negative RBCs
37 women received at least one KEL1-untested RBC
9 had repeat antibody screen ≥4 weeks posttransfusion
28 did not have repeat antibody screen ≥4 weeks post-transfusion
None of the women had anti-K antibodies on post-transfusion screen
Fig. 1. Flow chart showing disposition of women aged not more than 50 years admitted to the BIDMC inpatient obstetric service between April 9, 2009, and April 9, 2012, who were transfused RBCs.
women were transfused in the setting of peripartum hemorrhage. Six (2%) of the 227 women were not pregnant and were transfused in the setting of total abdominal hysterectomy (three), menorrhagia (two), and drainage of a tuboovarian abscess (one). Of this entire population, two women had inconclusive antibody screens on presentation and were excluded from the analysis. Of the 225 females included (Fig. 1), 188 (84%) females were transfused only phenotypically KEL1-negative units while 37 (16%) women received at least one KEL1-untested unit. Of the 37 women who were transfused at least one KEL1untested unit, 25 (68%) women also received at least one KEL1-negative unit. Nine (24%) women who received at least one KEL1-untested unit (median, 2 units; IQR, 1-2) had a repeat antibody screen at least 4 weeks posttransfusion. For these nine women, the median time between the last KEL1-untested RBC transfusion and the last antibody screen was 109 weeks (IQR, 35-115 weeks). None of the nine women demonstrated anti-K on repeat screening. Women who received at least one KEL1untested unit (n = 37) had a significantly greater number of total RBC units transfused (median, 5.0; IQR, 2-13) compared to women who received only KEL1-negative units (median, 2.0; IQR, 2-3; p < 0.0001). The indications for transfusion for women receiving at least one KEL1untested unit are as follows: peripartum hemorrhage (70%), nonobstetric surgical blood loss (16%), miscellaneous (14%). Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION601 3
O’BRIEN ET AL.
TABLE 1. The number of women not more than 50 years of age admitted to the obstetric and nonobstetric services from April 9, 2009, to April 9, 2012 and April 8, 1999, to April 8, 2009, who were transfused at least 1 RBC unit, the frequency of transfusion associated anti-K antibody detection, and the incidence of transfusion associated anti-K per 1000 women transfused Population of interest Women ≤ 50 years of age admitted to the obstetric service Transfused Frequency of transfusion-associated anti-K Incidence of transfusion-associated anti-K (per 1000 women transfused) Women ≤ 50 years of age admitted to the nonobstetric service Transfused Frequency of transfusion-associated anti-K Incidence of transfusion-associated anti-K (per 1000 women transfused)
During this same time frame, 20 total women not more than 50 years of age had anti-K detected during pretransfusion evaluation; three (15%) of the women, all admitted to nonobstetric services, were transfused KEL1untested RBCs between April 9, 2009, and April 9, 2012; pregnancy was ruled out as a source of alloimmunization in all three women after chart review. Two (10%) of the women, none of whom were admitted to the obstetric service at the time of their transfusions, were transfused at BIDMC before the study period; pregnancy could not be ruled out as a source of alloimmunization in these women. The remaining 15 (75%) women had no record of transfusion at our institution; they did all have a history of prior pregnancy documented in their medical record. During the 10-year period before policy implementation (April 8, 1999 to April 8, 2009), there were 27 documented cases of probable transfusion-associated KEL1 alloimmunization in women aged ≤50 years: 22 (81%) cases were in women transfused outside of the inpatient obstetric setting while five (19%) cases occurred in women transfused during an inpatient obstetric admission. The median number of KEL1-untested units transfused was 6 (2-8). There were no documented cases of KEL1-related HDFN. Comparison of alloimmunization rates before and after policy implementation showed no significant difference in anti-K sensitization after transfusion in either women not more than 50 years of age admitted to the obstetric inpatient service (p = 0.59) or in women not more than 50 years of age admitted to non-obstetric services (p = 0.46; Table 1).
DISCUSSION KEL1 alloimmunization is the second major cause of HDFN overall as well as severe HDFN requiring intrauterine transfusion.17 The management of pregnancies affected by anti-K differs from that of anti-D, as antibody titers may not correlate with the extent of fetal disease and severe disease can occur in the setting of low titers. While select countries have instituted guidelines for preventing 602 TRANSFUSION Volume Volume March 2015 4 TRANSFUSION **,55, ** **
April 9, 2009-April 9, 2012
April 8, 1999-April 8, 2009
p value
225 0 0.0
940 5 5.3
0.59
1125 3 2.7
4305 22 5.1
0.46
transfusion-related maternal KEL1 alloimmunization through the provision of KEL1-negative or Kell antigen– matched RBCs, no such policy has been standardized in the United States. Moreover, there is a lack of published data on the impact of provision of KEL1-negative or Kellmatched RBCs to women of childbearing potential on the incidence of Kell HDFN. In response to the 2007 TRANSFUSION editorial18 that strongly advocated for the provision of KEL1 antigen– matched blood for female recipients under the age of 50 years and the perceived ease of implementation from an operational perspective, our institution implemented a policy in April 2009 in which all women aged not more than 50 years and admitted to the inpatient obstetric service were to be transfused KEL1 antigen–negative RBCs if available. The present study was devised to determine the impact of this policy and inform our consideration for extending the policy to all women not more than 50 years of age. Thus, we examined policy compliance and impact on prevention of transfusion-related maternal KEL1 alloimmunization and HDFN through a retrospective review of all women who met policy criteria during the period between April 9, 2009, and April 9, 2012. Furthermore, all cases of potentially preventable transfusionassociated KEL1 alloimmunization in women not more than 50 years admitted to nonobstetric services during the 3 years of this policy practice as well as during the 10 years prior were identified and investigated. We found policy compliance to be high, with 84% of females who met criteria for transfusion with KEL1negative RBCs during the 3-year period having received only KEL1-negative units. The major reason for policy nonconformance was high transfusion requirement in an acute setting, which likely surpassed the blood bank’s ability to provide only phenotypically KEL1 antigen– negative units. This is demonstrated by the fact that the women who received at least one KEL1-untested unit were more heavily transfused than those who received only KEL1-negative units. Furthermore, 68% of patients who received at least one KEL1-untested unit also received KEL1-negative units. Although the number of
KEL1-NEGATIVE RBCs RBCS TO TO OBSTETRIC OBSTETRIC PATIENTS PATIENTS
1000 women without history of antiK each needing 1 RBC unit
No policy/restriction
Yes policy/restriction
910 (91%) KEL1-negative women at risk for KEL1 alloimmunization
910 (91%) KEL1-negative women at risk for KEL1 alloimmunization
82 (9%) women receive KEL1-positive RBC unit
455 (50%) women receive KEL1-untested RBC unit
41 (9%) women receive KEL1-positive RBC unit
4 (5%) KEL1-negative women develop anti-K antibodies
2 (5%) KEL1-negative women develop anti-K antibodies
Number need to treat (NNT) = 1/(proportion expected to be alloimmunized without policy – proportion expected to be alloimmunized under policy) = 1/(4/1000 – 2/1000) = 1/0.002 = 500 women, each requiring 1 unit of RBCs, need to be transfused KEL1-negative red cells to prevent one case of KEL1alloimmunization
Fig. 2. Diagram showing number of women who would have to be transfused KEL1-negative RBCs to prevent one case of transfusion-related KEL1 alloimmunization.
verified KEL1-negative RBC units in the inventory and the time it takes the blood bank to phenotype additional KEL1-negative units influences the blood bank’s ability to adhere to such a policy, we found that in most cases we were able to fully support women meeting policy criteria with KEL1-negative RBCs. Before committing to any policy change affecting transfusion practice, the potential costs and benefits must be considered. It may be argued that the cost–benefit ratio of restricting women of childbearing potential to KEL1negative RBCs is unfavorable due to the low frequency of the KEL1 antigen and incidence of anti-K–related HDFN; however, the potentially devastating sequelae of KEL1 alloimmunization vis-à-vis HDFN and the costs of managing Kell-sensitized pregnancies must be weighed. As previously mentioned, restricting women of childbearing potential to KEL1-negative RBCs necessitates only subtle changes in the RBC component inventory management system. To assess the cost–benefit ratio of providing KEL1negative RBCs to women of childbearing potential, we calculated the number needed to treat. Specifically, we calculated the number of women who would have to be transfused KEL1-negative RBCs to prevent one case of transfusion-related KEL1 alloimmunization (Fig. 2). For simplicity’s sake, we used a hypothetical population of
1000 women each needing 1 unit of RBCs and made the following assumptions: the frequency of the KEL1 antigen in both the donor and the hypothetical patient population is 9%, the rate of KEL1 alloimmunization for KEL1negative individuals after exposure to 1 unit of KEL1positive RBCs is 5%,16 and the policy compliance rate is 50%. We found that 500 women, each needing 1 unit of RBCs, would have to be transfused KEL1-negative RBCs to prevent one case of KEL1 alloimmunization. Considering that the reagent costs associated with phenotyping 1 unit of RBCs for the KEL1 antigen is $20.00 (personal communication, blood bank manager) and the cost of managing a Kell-sensitized pregnancy is substantial, the cost–benefit ratio may be in favor of preventing transfusion-related Kell alloimmunization in women of childbearing potential. For example, 1678 women not more than 50 years of age were transfused at least 1 unit of RBCs in 2012 at our institution. Only 401 of those 1678 patients (24%) were admitted to the obstetrics and gynecology service. Extending our current policy to all women not more than 50 years of age would involve KEL1 phenotyping an additional 1277 RBCs for an estimated annual increase of 0.3% to the blood bank budget. However, the low prevalence of KEL1-positive RBCs in the donor population serving our patients limits the possible benefits of our policy change, as evidenced by the lack of a significant difference in observed alloimmunization between the patient population restricted to KEL1-negative units compared to those who were not (Table 1, p = 0.59). Furthermore, the chance of a woman with KEL1 alloimmunization becoming pregnant by a male who is at least heterozygous for the KEL1 antigen is extremely low making the risk for anti-K–related HDFN remote as was evidenced by our data in which we observed no cases of HDFN. As the financial pressures on our health care system continue to build, there is an increasing realization that there are limits as to what can be reasonably spent to improve health.19 In this situation, prospectively phenotyping units for KEL1 antigen to create a KEL1 antigen–negative inventory is obviously more expensive than not doing so. This study has several limitations inherent in a retrospective review. Incomplete follow-up data make it difficult to accurately assess alloimmunization rates after transfusion and, therefore, to determine policy effectiveness in preventing transfusion-related maternal KEL1 alloimmunization and HDFN. Also, it is challenging to obtain accurate and complete pregnancy history through retrospective chart review. Of the 37 women who were transfused at least one KEL1-untested unit during the 3-year policy period, only nine had a repeat antibody screen at least 4 weeks posttransfusion. In addition, we were only able to find 27 documented cases of transfusion-associated KEL1 alloimmunization in women aged not more than 50 years during the 10 years before Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION603 5
O’BRIEN ET AL.
policy implementation, and five potential cases during the 3-year period of the policy; this likely represents an underestimate of alloimmunization due to limited availability of posttransfusion antibody screens. The short time period examined postpolicy (3 years) may also not be enough time to assess the impact of such a policy. In conclusion, we observed no clinical benefit of a policy directed at limiting KEL1 antigen exposure through provision of KEL1 antigen–negative units to women of childbearing potential with respect to prevention of alloimmunization or HDFN. Thus, despite the ease of implementation and good adherence, we do not endorse continuation of the policy for our obstetric population, much less its extension to all women of childbearing potential and caution against adoption of this transfusion practice by US transfusion services. ACKNOWLEDGMENTS The authors acknowledge Michele Hacker, ScD, MSPH, and the Harvard Clinical and Translational Science Center for statistical expertise and discussions. CONFLICT OF INTEREST KLO, RLH, and LU have declared no conflict of interest. YAK reports employment affiliation with Biogen Idec, Inc.
REFERENCES 1. Westhoff CM, Reid ME. Review: the Kell, Duffy, and Kidd blood group systems. Immunohematology 2004;20:37-49. 2. Daniels G. Other blood groups. In: Roback JD, Grossman BJ, Harris T, et al., editors. Technical manual. 17th ed. Bethesda (MD): American Association of Blood Banks; 2011. p. 411-36. 3. Geifman-Holtzman O, Wojtowycz M, Kosmas E, et al. Female alloimmunization with antibodies known to cause hemolytic disease. Obstet Gynecol 1997;89:272-5. 4. van der Schoot CE, Tax GH, Rijnders RJ, et al. Prenatal typing of Rh and Kell blood group system antigens: the edge of a watershed. Transfus Med Rev 2003;17:31-44. 5. Caine ME, Mueller-Heubach E. Kell sensitization in pregnancy. Am J Obstet Gynecol 1986;154:85-90.
604 TRANSFUSION Volume Volume March 2015 6 TRANSFUSION **,55, ** **
6. Klein HG, Anstee DJ. Haemolytic disease of the fetus and newborn. In: Klein HG, Anstee DJ, editors. Mollison’s blood transfusion in clinical medicine. 11th ed. Malden (MA): Blackwell Publishing Ltd; 2005. p. 526-7. 7. Bowman JM, Pollock JM, Manning FA, et al. Maternal Kell blood group alloimmunization. Obstet Gynecol 1992;79: 239-44. 8. van Wamelen DJ, Klumper FJ, de Haas M, et al. Obstetric history and antibody titer in estimating severity of Kell alloimmunization in pregnancy. Obstet Gynecol 2007;109: 1093-8. 9. Vaughan JI, Manning M, Warwick RM, et al. Inhibition of erythroid progenitor cells by anti-Kell antibodies in fetal alloimmune anemia. N Engl J Med 1998;338:798-803. 10. Vaughan JI, Warwick R, Letsky E, et al. Erythropoietic suppression in fetal anemia because of Kell alloimmunization. Am J Obstet Gynecol 1994;171:247-52. 11. Weiner CP, Widness JA. Decreased fetal erythropoiesis and hemolysis in Kell hemolytic anemia. Am J Obstet Gynecol 1996;174:547-51. 12. Santiago JC, Ramos-Corpas D, Oyonarte S, et al. Current clinical management of anti-Kell alloimmunization in pregnancy. Eur J Obstet Gynecol Reprod Biol 2008;136: 151-4. 13. Farr V, Gray E. Pregnancy outcome in mothers who develop Kell antibodies. Scott Med J 1988;33:300-3. 14. Grant SR, Kilby MD, Meer L, et al. The outcome of pregnancy in Kell alloimmunisation. BJOG 2000;107:481-5. 15. McKenna DS, Nagaraja HN, O’Shaughnessy R. Management of pregnancies complicated by anti-Kell isoimmunization. Obstet Gynecol 1999;93(5 Pt 1):667-73. 16. Mayne KM, Bowell PJ, Pratt GA. The significance of antiKell sensitization in pregnancy. Clin Lab Haematol 1990; 12:379-85. 17. Van Kamp IL, Klumper FJ, Oepkes D, et al. Complications of intrauterine intravascular transfusion for fetal anemia due to maternal red-cell alloimmunization. Am J Obstet Gynecol 2005;192:171-7. 18. Westhoff CM. Molecular genotyping for RHD: what (not) to do? Transfusion 2007;47:1337-9. 19. Kacker S, Frick KD, Tobian AA. Data and interpretation: economic evaluations in transfusion medicine, Part 4. Transfusion 2014;53:2130-3.
Provision of KEL1-negative blood to obstetric patients: a 3-year single-institution retrospective review Kerry L. O’Brien,1,2* Yeowon A. Kim,3* Richard L. Haspel,1,2 and Lynne Uhl1,2
BACKGROUND: KEL1 alloimmunization is a major cause of hemolytic disease of the fetus and newborn (HDFN). While select countries have guidelines for preventing transfusion-associated KEL1 alloimmunization, the United States does not. Beth Israel Deaconess Medical Center instituted a policy in April 2009 whereby women not more than 50 years of age on the obstetric service were transfused KEL1-negative red blood cells (RBCs). We sought to determine compliance and impact for prevention of KEL1 alloimmunization and HDFN. STUDY DESIGN AND METHODS: All women not more than 50 years of age without anti-K transfused RBCs during an obstetric admission from April 9, 2009, to April 9, 2012, were identified (227). Adherence to policy, factors contributing to nonadherence, and subsequent impact were evaluated. For comparison, all cases of anti-K detection in women not more than 50 years of age admitted to nonobstetric services and all cases of transfusion-associated KEL1 alloimmunization in women not more than 50 years of age during the 10 years prior were identified. RESULTS: Eighty-four percent received only KEL1negative units. Three (1.3%) women not more than 50 years of age on the obstetric service were identified with anti-K, while 17 (1.5%) women not more than 50 years of age on nonobstetric services had anti-K detected; only five of 20 had a prior RBC transfusion. In the 10 years prior, there were 27 cases of transfusionassociated KEL1 alloimmunization in women not more than 50 years of age. There were no cases of KEL1 HDFN in either period. CONCLUSION: Although the findings demonstrate feasibility of providing KEL1-negative RBCs to women of childbearing potential, evidence for clinical benefit is lacking. The low prevalence of KEL1 in blood donors, the lack of significant differences in alloimmunization rates, and no cases of HDFN during the study period questions the clinical benefit of such a policy.
T
he Kell blood group system is the third most polymorphic blood group system and its antigens are highly immunogenic.1 Antibodies to the KEL1 antigen (also known as anti-K) are the most common red blood cell (RBC) antibodies outside the ABO and Rh systems.2 Anti-K is clinically significant and can cause severe transfusion reactions and hemolytic disease of the fetus and newborn (HDFN). While the rate of anti-D–related maternal alloimmunization and HDFN has decreased with RhIG prophylaxis, the incidence and frequency of KEL1 alloimmunization has increased and surpassed that of anti-D in some populations studied.3,4 Most recent data show the frequency of anti-KEL1 to be one per 1000 pregnant females5 and the incidence of KEL1-related HDFN to be about one in 40,000 births.6 The management of pregnancies affected by anti-K differs from that of anti-D, as antibody titers may not correlate
ABBREVIATIONS: BIDMC = Beth Israel Deaconess Medical Center; HDFN = hemolytic disease of the fetus and newborn; IQR = interquartile range. From the 1Department of Pathology, Beth Israel Deaconess Medical Center; 2Department of Pathology, Harvard Medical School; and 3Joint Program in Transfusion Medicine, Harvard Medical School, Boston, Massachusetts. Address reprint requests to: Kerry L. O’Brien, MD, Blood Bank, Beth Israel Deaconess Medical Center, YA-309, 330 Brookline Avenue, Boston, MA 02215; e-mail: [email protected]. *Both authors contributed equally to this manuscript. This work was conducted with support from Harvard Catalyst | The Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health Award 8UL1TR000170-05) and financial contributions from Harvard University and its affiliated academic health care centers. Received for publication April 17, 2014; revision received July 9, 2014, and accepted July 9, 2014. doi: 10.1111/trf.12814 © 2014 AABB TRANSFUSION **;**:**-**. 2015;55:599–604.
Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION599 1
O’BRIEN ET AL.
with the extent of fetal disease and severe disease can occur in the setting of low titers.7,8 This has been explained by the differences in the suggested mechanisms of anti-D– and anti-K–related HDFN: while hemolysis is the predominant process in anti-D alloimmunization, anti-K cause fetal anemia primarily by suppression of erythropoiesis.9-11 The etiology of anti-KEL1 identified during pregnancy is predominantly transfusion associated and not a consequence of immunization after transplacental hemorrhage from the fetus,6 and while there is a wide range (30% to >50%) in the estimated percentage of maternal Kell alloimmunization cases attributed to transfusion,8,12-16 even the most conservative estimates suggest that a substantial fraction of maternal KEL1 alloimmunization can be prevented by the provision of KEL1 antigen–negative blood. Such policies have been adopted by several countries, excluding the United States, and have reduced the proportion of alloimmunized women with a history of transfusion.17 However, there are no published data regarding the impact of such policies on the rate of anti-Kell–related HDFN. In April 2009 our institution implemented a policy to provide women aged not more than 50 years admitted to the inpatient obstetric service KEL1 antigen–negative RBCs if available. The policy’s goal was to avoid transfusion-related maternal KEL1 alloimmunization and HDFN. Since only 9% of the Caucasian (donor) population is positive for the KEL1 antigen, implementation of the policy at this time required modest changes in the RBC component inventory management system with minimal financial and operational impact. However, in light of the current constraints on health care expenditure and operational laboratory support, we were obliged to retrospectively review the clinical impact of this policy. Specifically, 3 years after implementation, we sought to retrospectively evaluate policy compliance, nonadherence, and the impact on prevention of transfusion-related maternal KEL1 alloimmunization and HDFN. It was considered that the data derived from this study could be used to inform the feasibility and effectiveness of broadly providing KEL1 antigen–negative RBCs to all women of childbearing potential in our institution and possibly in similar institutions in the United States.
MATERIALS AND METHODS The study was approved by the institutional review board of the Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts. BIDMC is a 650-bed tertiary care training institution offering a full range of services including an active obstetric service with more than 5000 births per year, a Level 3 nursery, allogeneic and autologous hematopoietic stem cell and organ (kidney and liver) transplant services, a cardiothoracic surgery practice, 77 600 TRANSFUSION Volume Volume March 2015 2 TRANSFUSION **,55, ** **
critical care beds, and a Level 1 trauma service. The hospital does not have a pediatric service, and all patients are at least 15 years of age except neonates. A computerized query of the Center for Clinical Computing (Boston, MA) blood bank database was performed to identify all women admitted to the inpatient obstetric service between April 9, 2009, and April 9, 2012, and who were transfused RBCs. For all patients, data on age at transfusion(s), underlying diagnosis, pregnancy history, the occurrence of KEL1-related HDFN, dates of transfusions, indication for transfusion, number of KEL1untested and KEL1 antigen–negative RBC units transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted via medical chart review and transfusion record review. Patients who had the following were excluded from analysis: those who had a history of anti-K or an inconclusive antibody screen upon presentation and age more than 50 years. As a comparison, all females not more than 50 years of age in whom anti-K was detected during this same time frame were identified. Pregnancy and RBC transfusion history were evaluated to determine possible mode of anti-K alloimmunization. We then performed a search for all women not more than 50 years of age who were transfused RBCs during this same 3-year period and who were not admitted to the obstetric service. Those women who had already made anti-K or who had inconclusive antibody panels on presentation were excluded from analysis. For all these patients, data on age at transfusion(s), underlying diagnosis, pregnancy history, the occurrence of KEL1-related HDFN, dates of transfusions, indication for transfusion, number of KEL1-untested and KEL1 antigen–negative RBC units transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted via medical chart review and transfusion record review. To determine the potential impact of the policy in preventing cases of maternal KEL1 alloimmunization and HDFN resulting from transfusion, a second computerized query was performed to evaluate for the number of transfusion-associated KEL1 alloimmunization and HDFN cases in women of childbearing age in the 10 years before the policy change. All patients who had an anti-K on antibody screen and were transfused RBCs between April 8, 1999, and April 8, 2009, were identified. For all patients, data on sex, age at transfusion(s), underlying diagnosis, pregnancy history, occurrence of KEL1-related HDFN, admission status at time of transfusion (i.e., inpatient versus outpatient, if the patient was admitted as an inpatient, the primary admitting service), dates of transfusions, indication for transfusion, number of KEL1-untested and KEL1 antigen–negative RBC units
KEL1-NEGATIVE RBCs RBCS TO TO OBSTETRIC OBSTETRIC PATIENTS PATIENTS
transfused, ABO type and RhD status of patient, KEL1 antigen status of patient, and availability and results of antibody screens performed at least 4 weeks after transfusion were extracted as above. For the cases of transfusionassociated KEL1 alloimmunization, the source of alloimmunization was determined by assessing the temporal relationship between transfusion, results of antibody screens, and pregnancies. Transfusion was determined as the source of alloimmunization if the patient developed an anti-K after transfusion and did not have any documented pregnancies during the time interval between transfusion and anti-KEL1 positivity. For patients who developed anti-K after transfusion but also were pregnant during the time interval between change in anti-KEL1 status, the source of alloimmunization was designated as likely transfusion related, although pregnancy could not be ruled out. Antibody screens were performed using polyethylene glycol indirect antiglobulin test (Gamma Biologicals, Inc., Houston, TX) and a three-cell screen (R1R1, R2R2, rr; Immucor, Inc., Norcross, GA) between 1999 and 2007 according to standard institutional procedure, with agglutination reactions read macroscopically in tube and graded from 0 to 4+. Beginning 2008, solid-phase technology (Capture-R, Immucor, Inc.) and a two-cell screen (R1R1, R2R2) was used for antibody screens. Positive screening tests were followed by serologic investigation using a panel of 14 phenotypically defined RBCs (Immucor, Inc.) selected to identify the presence of anti-K. KEL1 antigen phenotyping of donors and recipients were performed by incubating donor or recipient RBCs with monoclonal IgM antibodies against KEL1 antigen (Immucor, Inc.) according to standard institutional procedure, with agglutination reactions read macroscopically in tube. Agglutination reactions 1+ and greater were classified as positive for the KEL1 antigen.
Statistical analysis Transfusion data are presented as medians with interquartile range (IQR) and their between-period comparisons were evaluated by the Wilcoxon rank-sum test. Dichotomous variables were reported as proportions (%) and were compared using the Fisher exact test. A twotailed p value of less than 0.05 was considered significant. All statistical analyses were performed using computer software (GraphPad Prism 6 for Windows, GraphPad Software, Inc., La Jolla, CA).
RESULTS Between April 9, 2009, and April 9, 2012, a total of 227 women aged not more than 50 years historically negative for anti-K were admitted to the inpatient obstetric service and were transfused RBCs. The vast majority of the
227 women aged ≤50 years and without history of anti-K were transfused RBCs between 4/9/09 and 4/9/12 2 with inconclusive antibody screen 225 women
188 women received only KEL1-negative RBCs
37 women received at least one KEL1-untested RBC
9 had repeat antibody screen ≥4 weeks posttransfusion
28 did not have repeat antibody screen ≥4 weeks post-transfusion
None of the women had anti-K antibodies on post-transfusion screen
Fig. 1. Flow chart showing disposition of women aged not more than 50 years admitted to the BIDMC inpatient obstetric service between April 9, 2009, and April 9, 2012, who were transfused RBCs.
women were transfused in the setting of peripartum hemorrhage. Six (2%) of the 227 women were not pregnant and were transfused in the setting of total abdominal hysterectomy (three), menorrhagia (two), and drainage of a tuboovarian abscess (one). Of this entire population, two women had inconclusive antibody screens on presentation and were excluded from the analysis. Of the 225 females included (Fig. 1), 188 (84%) females were transfused only phenotypically KEL1-negative units while 37 (16%) women received at least one KEL1-untested unit. Of the 37 women who were transfused at least one KEL1untested unit, 25 (68%) women also received at least one KEL1-negative unit. Nine (24%) women who received at least one KEL1-untested unit (median, 2 units; IQR, 1-2) had a repeat antibody screen at least 4 weeks posttransfusion. For these nine women, the median time between the last KEL1-untested RBC transfusion and the last antibody screen was 109 weeks (IQR, 35-115 weeks). None of the nine women demonstrated anti-K on repeat screening. Women who received at least one KEL1untested unit (n = 37) had a significantly greater number of total RBC units transfused (median, 5.0; IQR, 2-13) compared to women who received only KEL1-negative units (median, 2.0; IQR, 2-3; p < 0.0001). The indications for transfusion for women receiving at least one KEL1untested unit are as follows: peripartum hemorrhage (70%), nonobstetric surgical blood loss (16%), miscellaneous (14%). Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION601 3
O’BRIEN ET AL.
TABLE 1. The number of women not more than 50 years of age admitted to the obstetric and nonobstetric services from April 9, 2009, to April 9, 2012 and April 8, 1999, to April 8, 2009, who were transfused at least 1 RBC unit, the frequency of transfusion associated anti-K antibody detection, and the incidence of transfusion associated anti-K per 1000 women transfused Population of interest Women ≤ 50 years of age admitted to the obstetric service Transfused Frequency of transfusion-associated anti-K Incidence of transfusion-associated anti-K (per 1000 women transfused) Women ≤ 50 years of age admitted to the nonobstetric service Transfused Frequency of transfusion-associated anti-K Incidence of transfusion-associated anti-K (per 1000 women transfused)
During this same time frame, 20 total women not more than 50 years of age had anti-K detected during pretransfusion evaluation; three (15%) of the women, all admitted to nonobstetric services, were transfused KEL1untested RBCs between April 9, 2009, and April 9, 2012; pregnancy was ruled out as a source of alloimmunization in all three women after chart review. Two (10%) of the women, none of whom were admitted to the obstetric service at the time of their transfusions, were transfused at BIDMC before the study period; pregnancy could not be ruled out as a source of alloimmunization in these women. The remaining 15 (75%) women had no record of transfusion at our institution; they did all have a history of prior pregnancy documented in their medical record. During the 10-year period before policy implementation (April 8, 1999 to April 8, 2009), there were 27 documented cases of probable transfusion-associated KEL1 alloimmunization in women aged ≤50 years: 22 (81%) cases were in women transfused outside of the inpatient obstetric setting while five (19%) cases occurred in women transfused during an inpatient obstetric admission. The median number of KEL1-untested units transfused was 6 (2-8). There were no documented cases of KEL1-related HDFN. Comparison of alloimmunization rates before and after policy implementation showed no significant difference in anti-K sensitization after transfusion in either women not more than 50 years of age admitted to the obstetric inpatient service (p = 0.59) or in women not more than 50 years of age admitted to non-obstetric services (p = 0.46; Table 1).
DISCUSSION KEL1 alloimmunization is the second major cause of HDFN overall as well as severe HDFN requiring intrauterine transfusion.17 The management of pregnancies affected by anti-K differs from that of anti-D, as antibody titers may not correlate with the extent of fetal disease and severe disease can occur in the setting of low titers. While select countries have instituted guidelines for preventing 602 TRANSFUSION Volume Volume March 2015 4 TRANSFUSION **,55, ** **
April 9, 2009-April 9, 2012
April 8, 1999-April 8, 2009
p value
225 0 0.0
940 5 5.3
0.59
1125 3 2.7
4305 22 5.1
0.46
transfusion-related maternal KEL1 alloimmunization through the provision of KEL1-negative or Kell antigen– matched RBCs, no such policy has been standardized in the United States. Moreover, there is a lack of published data on the impact of provision of KEL1-negative or Kellmatched RBCs to women of childbearing potential on the incidence of Kell HDFN. In response to the 2007 TRANSFUSION editorial18 that strongly advocated for the provision of KEL1 antigen– matched blood for female recipients under the age of 50 years and the perceived ease of implementation from an operational perspective, our institution implemented a policy in April 2009 in which all women aged not more than 50 years and admitted to the inpatient obstetric service were to be transfused KEL1 antigen–negative RBCs if available. The present study was devised to determine the impact of this policy and inform our consideration for extending the policy to all women not more than 50 years of age. Thus, we examined policy compliance and impact on prevention of transfusion-related maternal KEL1 alloimmunization and HDFN through a retrospective review of all women who met policy criteria during the period between April 9, 2009, and April 9, 2012. Furthermore, all cases of potentially preventable transfusionassociated KEL1 alloimmunization in women not more than 50 years admitted to nonobstetric services during the 3 years of this policy practice as well as during the 10 years prior were identified and investigated. We found policy compliance to be high, with 84% of females who met criteria for transfusion with KEL1negative RBCs during the 3-year period having received only KEL1-negative units. The major reason for policy nonconformance was high transfusion requirement in an acute setting, which likely surpassed the blood bank’s ability to provide only phenotypically KEL1 antigen– negative units. This is demonstrated by the fact that the women who received at least one KEL1-untested unit were more heavily transfused than those who received only KEL1-negative units. Furthermore, 68% of patients who received at least one KEL1-untested unit also received KEL1-negative units. Although the number of
KEL1-NEGATIVE RBCs RBCS TO TO OBSTETRIC OBSTETRIC PATIENTS PATIENTS
1000 women without history of antiK each needing 1 RBC unit
No policy/restriction
Yes policy/restriction
910 (91%) KEL1-negative women at risk for KEL1 alloimmunization
910 (91%) KEL1-negative women at risk for KEL1 alloimmunization
82 (9%) women receive KEL1-positive RBC unit
455 (50%) women receive KEL1-untested RBC unit
41 (9%) women receive KEL1-positive RBC unit
4 (5%) KEL1-negative women develop anti-K antibodies
2 (5%) KEL1-negative women develop anti-K antibodies
Number need to treat (NNT) = 1/(proportion expected to be alloimmunized without policy – proportion expected to be alloimmunized under policy) = 1/(4/1000 – 2/1000) = 1/0.002 = 500 women, each requiring 1 unit of RBCs, need to be transfused KEL1-negative red cells to prevent one case of KEL1alloimmunization
Fig. 2. Diagram showing number of women who would have to be transfused KEL1-negative RBCs to prevent one case of transfusion-related KEL1 alloimmunization.
verified KEL1-negative RBC units in the inventory and the time it takes the blood bank to phenotype additional KEL1-negative units influences the blood bank’s ability to adhere to such a policy, we found that in most cases we were able to fully support women meeting policy criteria with KEL1-negative RBCs. Before committing to any policy change affecting transfusion practice, the potential costs and benefits must be considered. It may be argued that the cost–benefit ratio of restricting women of childbearing potential to KEL1negative RBCs is unfavorable due to the low frequency of the KEL1 antigen and incidence of anti-K–related HDFN; however, the potentially devastating sequelae of KEL1 alloimmunization vis-à-vis HDFN and the costs of managing Kell-sensitized pregnancies must be weighed. As previously mentioned, restricting women of childbearing potential to KEL1-negative RBCs necessitates only subtle changes in the RBC component inventory management system. To assess the cost–benefit ratio of providing KEL1negative RBCs to women of childbearing potential, we calculated the number needed to treat. Specifically, we calculated the number of women who would have to be transfused KEL1-negative RBCs to prevent one case of transfusion-related KEL1 alloimmunization (Fig. 2). For simplicity’s sake, we used a hypothetical population of
1000 women each needing 1 unit of RBCs and made the following assumptions: the frequency of the KEL1 antigen in both the donor and the hypothetical patient population is 9%, the rate of KEL1 alloimmunization for KEL1negative individuals after exposure to 1 unit of KEL1positive RBCs is 5%,16 and the policy compliance rate is 50%. We found that 500 women, each needing 1 unit of RBCs, would have to be transfused KEL1-negative RBCs to prevent one case of KEL1 alloimmunization. Considering that the reagent costs associated with phenotyping 1 unit of RBCs for the KEL1 antigen is $20.00 (personal communication, blood bank manager) and the cost of managing a Kell-sensitized pregnancy is substantial, the cost–benefit ratio may be in favor of preventing transfusion-related Kell alloimmunization in women of childbearing potential. For example, 1678 women not more than 50 years of age were transfused at least 1 unit of RBCs in 2012 at our institution. Only 401 of those 1678 patients (24%) were admitted to the obstetrics and gynecology service. Extending our current policy to all women not more than 50 years of age would involve KEL1 phenotyping an additional 1277 RBCs for an estimated annual increase of 0.3% to the blood bank budget. However, the low prevalence of KEL1-positive RBCs in the donor population serving our patients limits the possible benefits of our policy change, as evidenced by the lack of a significant difference in observed alloimmunization between the patient population restricted to KEL1-negative units compared to those who were not (Table 1, p = 0.59). Furthermore, the chance of a woman with KEL1 alloimmunization becoming pregnant by a male who is at least heterozygous for the KEL1 antigen is extremely low making the risk for anti-K–related HDFN remote as was evidenced by our data in which we observed no cases of HDFN. As the financial pressures on our health care system continue to build, there is an increasing realization that there are limits as to what can be reasonably spent to improve health.19 In this situation, prospectively phenotyping units for KEL1 antigen to create a KEL1 antigen–negative inventory is obviously more expensive than not doing so. This study has several limitations inherent in a retrospective review. Incomplete follow-up data make it difficult to accurately assess alloimmunization rates after transfusion and, therefore, to determine policy effectiveness in preventing transfusion-related maternal KEL1 alloimmunization and HDFN. Also, it is challenging to obtain accurate and complete pregnancy history through retrospective chart review. Of the 37 women who were transfused at least one KEL1-untested unit during the 3-year policy period, only nine had a repeat antibody screen at least 4 weeks posttransfusion. In addition, we were only able to find 27 documented cases of transfusion-associated KEL1 alloimmunization in women aged not more than 50 years during the 10 years before Volume 55,Volume March 2015 **, ** **TRANSFUSION TRANSFUSION603 5
O’BRIEN ET AL.
policy implementation, and five potential cases during the 3-year period of the policy; this likely represents an underestimate of alloimmunization due to limited availability of posttransfusion antibody screens. The short time period examined postpolicy (3 years) may also not be enough time to assess the impact of such a policy. In conclusion, we observed no clinical benefit of a policy directed at limiting KEL1 antigen exposure through provision of KEL1 antigen–negative units to women of childbearing potential with respect to prevention of alloimmunization or HDFN. Thus, despite the ease of implementation and good adherence, we do not endorse continuation of the policy for our obstetric population, much less its extension to all women of childbearing potential and caution against adoption of this transfusion practice by US transfusion services. ACKNOWLEDGMENTS The authors acknowledge Michele Hacker, ScD, MSPH, and the Harvard Clinical and Translational Science Center for statistical expertise and discussions. CONFLICT OF INTEREST KLO, RLH, and LU have declared no conflict of interest. YAK reports employment affiliation with Biogen Idec, Inc.
REFERENCES 1. Westhoff CM, Reid ME. Review: the Kell, Duffy, and Kidd blood group systems. Immunohematology 2004;20:37-49. 2. Daniels G. Other blood groups. In: Roback JD, Grossman BJ, Harris T, et al., editors. Technical manual. 17th ed. Bethesda (MD): American Association of Blood Banks; 2011. p. 411-36. 3. Geifman-Holtzman O, Wojtowycz M, Kosmas E, et al. Female alloimmunization with antibodies known to cause hemolytic disease. Obstet Gynecol 1997;89:272-5. 4. van der Schoot CE, Tax GH, Rijnders RJ, et al. Prenatal typing of Rh and Kell blood group system antigens: the edge of a watershed. Transfus Med Rev 2003;17:31-44. 5. Caine ME, Mueller-Heubach E. Kell sensitization in pregnancy. Am J Obstet Gynecol 1986;154:85-90.
604 TRANSFUSION Volume Volume March 2015 6 TRANSFUSION **,55, ** **
6. Klein HG, Anstee DJ. Haemolytic disease of the fetus and newborn. In: Klein HG, Anstee DJ, editors. Mollison’s blood transfusion in clinical medicine. 11th ed. Malden (MA): Blackwell Publishing Ltd; 2005. p. 526-7. 7. Bowman JM, Pollock JM, Manning FA, et al. Maternal Kell blood group alloimmunization. Obstet Gynecol 1992;79: 239-44. 8. van Wamelen DJ, Klumper FJ, de Haas M, et al. Obstetric history and antibody titer in estimating severity of Kell alloimmunization in pregnancy. Obstet Gynecol 2007;109: 1093-8. 9. Vaughan JI, Manning M, Warwick RM, et al. Inhibition of erythroid progenitor cells by anti-Kell antibodies in fetal alloimmune anemia. N Engl J Med 1998;338:798-803. 10. Vaughan JI, Warwick R, Letsky E, et al. Erythropoietic suppression in fetal anemia because of Kell alloimmunization. Am J Obstet Gynecol 1994;171:247-52. 11. Weiner CP, Widness JA. Decreased fetal erythropoiesis and hemolysis in Kell hemolytic anemia. Am J Obstet Gynecol 1996;174:547-51. 12. Santiago JC, Ramos-Corpas D, Oyonarte S, et al. Current clinical management of anti-Kell alloimmunization in pregnancy. Eur J Obstet Gynecol Reprod Biol 2008;136: 151-4. 13. Farr V, Gray E. Pregnancy outcome in mothers who develop Kell antibodies. Scott Med J 1988;33:300-3. 14. Grant SR, Kilby MD, Meer L, et al. The outcome of pregnancy in Kell alloimmunisation. BJOG 2000;107:481-5. 15. McKenna DS, Nagaraja HN, O’Shaughnessy R. Management of pregnancies complicated by anti-Kell isoimmunization. Obstet Gynecol 1999;93(5 Pt 1):667-73. 16. Mayne KM, Bowell PJ, Pratt GA. The significance of antiKell sensitization in pregnancy. Clin Lab Haematol 1990; 12:379-85. 17. Van Kamp IL, Klumper FJ, Oepkes D, et al. Complications of intrauterine intravascular transfusion for fetal anemia due to maternal red-cell alloimmunization. Am J Obstet Gynecol 2005;192:171-7. 18. Westhoff CM. Molecular genotyping for RHD: what (not) to do? Transfusion 2007;47:1337-9. 19. Kacker S, Frick KD, Tobian AA. Data and interpretation: economic evaluations in transfusion medicine, Part 4. Transfusion 2014;53:2130-3.
