Association of ABO Incompatibility With Red Blood Cell Indices of Cord Blood Unit

Background:Maternalefetal ABO incompatibility is one of the causes of neonatal hyperbilirubinemia. We postulate that hemoglobin (Hb), hematocrit (Hct), and red blood cell (RBC) valuesfor cord blood units (CBUs) are lower and erythroblast values higher for maternalefetal ABOincompatible dyads than for compatible dyads.

Objective:We investigated the relationship between Hb, Hct, RBC, and erythroblast CBUvalues and maternalefetal ABO blood type compatibility.

Methods:Mothers having blood group O who gave birth to infants with blood group A, B, or ABwere classified as Group I. According to baby’s blood group, the members of Group I werefurther divided into AO (baby group A, mother group O), BO (baby group B, mother groupO), and ABO (baby group AB, mother group O) subgroups. Mothers having blood group A whogave birth to infants with blood group B or AB and mothers having blood group B who gave birthto infants with blood group A or AB were classified as Group II. All other maternalefetal bloodtype pairs were considered ABO compatible and were classified as Group III. We comparedmean Hb, Hct, RBC, and erythroblast values for the infants’ CBUs among these three groupsincluding the subgroups of Group I.

Results:Group I had lower mean Hb, Hct, and RBC values than Group II and Group III (bothp<0.001). Although the mean Hb, Hct, and RBC values for Group II were lower than for Group III, the difference was not statistically significant. Mean Hb and RBC for the AO groupwere higher and nucleated RBC (nRBC) ratios were lower than for the BO group; however,these differences were also not statistically significant. Interestingly, the mean Hct value ofthe BO group was significantly lower than that of the AO group (pZ0.04).

Conclusion:Group A or B neonates with a group O mother have lower mean Hb, Hct, and RBCvalues for CBUs than other neonates. The role of RBC indices in predicting neonatal hemolytichyperbilirubinemia remains unclear and further studies are needed to identify the possibleclinical association.

Copyright©2012, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. All rightsreserved.

1. Introduction

Hemolytic disease of the newborns (HDN) due to maternalantibodies is a situation in which the lifespan of theneonate’s red cells is shortened due to the activity oftransplacental maternal antibodies. More than 99.5% ofTaiwanese are D Rh positive and, therefore, HDN due toanti-D antibodies is rarely encountered.1In addition,hemolytic disease resulting from ABO incompatibility isclinically milder than anti-D disease. However, HDN andkernicterus do occasionally occur, and hydrops fetalis hasbeen suggested as a result.2,3Thus, it is desirable to havea simple and reliable test that is able to predict thedevelopment of neonatal hyperbilirubinemia due to ABOincompatibility, after which preventive phototherapy canbe used. Hemoglobin (Hb) levels, hematocrit (Hct) counts,reticulocyte counts, direct Coombs test results, bilirubinlevels, and immunoglobulin G (IgG) titers that have beenobtained from cord blood together with maternal anti-A/anti-B titers have been suggested as approaches for predicting the severity of hyperbilirubinemia in ABO HDN.

A complete blood count (CBC) test is an absoluterequirement before any CBU cryopreservation. CBC testingof CBUs is noninvasive, convenient, simple, and rapid. Wehypothesized that ABO incompatibility influences Hb, Hct,red blood cell (RBC), and erythroblast values obtained fromCBU. The veracity of this hypothesis could be tested usingreference data and then used to help predict ABO incompatibility that is related to HDN. Most studies reported inthe literature, when obtaining CBC samples, have used cordblood obtained directly from the umbilical veins at deliveryrather than cord blood units. To our best knowledge, nosimilar study has been reported to date involving a sampleof this great size. Using 3688 CBUs, we studied the impactof differences in the mothers’ and fetuses’ blood typecombinations on Hb, Hct, RBC, and erythroblast values forCBUs.

2. Materials and Methods

2.1. Cord blood collection

Between September 2001 and November 2006, donatedcord blood samples from healthy Taiwanese singletonneonates with a gestational age more than 36 weeks born tomarried mothers were collected by the Tzu Chi Cord Blood Bank. CBUs with a net weight of more than 90 g wereaccepted. CBUs where one parent carried thalassemia werenot accepted, and collected CBUs were discarded if thebaby was diagnosed as having glucose-6-phosphate dehydrogenase deficiency (G-6-PD) deficiency. Written informedconsent was obtained from the mother donating the CBUbefore collection. All CBUs were collectedin uterousinga standard procedure. After delivery, the cord was sterilized and a 16-gauze needle was inserted into the umbilicalvein. The cord blood was collected by gravity into a collecting bag containing 28 mL anticoagulant phosphatecitrate-dextrose. The bag was stored at 4℃ to 10℃andsent to Tzu Chi Cord Blood Bank within 24 hours. Between1 mL and 2 mL of the aspirated cord blood from cord bloodbag was infused into an EDTA tube. Subsequently, the cordblood CBC, white blood cell differential count (WBC DC), Rhtyping, and ABO typing were analyzed in the central laboratory of Hualien Buddhist Tzu Chi General Hospital byexperienced technicians. Blood group typing was performed routinely using standard blood bank techniques.

2.2. Analysis of cord blood CBC and WBC DC

The CBC testing included RBC, Hb, Hct, WBC, and plateletcounts, which were measured using a Sysmex XE2100 automated hematology analyzer (Sysmex Corporation,Kobe, Japan). Nucleated RBCs (nRBCs) were reported as thenumber of nucleated RBCs per 100 WBCs. According tothe quality control chart and Westgard rules 13S/22S/R4S,the analyzer was calibrated twice daily using a commercialassayed control cell.

2.3. Materials

Among the 5602 CBUs available, 1913 units lacked somedata, and these were excluded. Furthermore, there wasonly one unit where the mother was AB blood group andbaby was O group, therefore we excluded this CBU also. Intotal, 3688 healthy neonates were included in this study,and these were divided into three groups. Mothers havingblood group O who had given birth to infants having bloodgroup A, B, or AB were classified as Group I. Group I consisted of 555 CBUs (A, B, and AB; 305, 247, and 3, respectively) where the newborn’s mother’s blood group was O.Mothers having blood group A who had given birth to infantshaving blood group B or AB together with mothers havingblood group B who gave birth to infants having blood group A or AB were classified as Group II. Group II consisted of 326newborns. Groups I and II were both considered to bematernalefetal ABO-incompatible pairs. All other maternalefetal blood group pairs were considered ABO compatible and were classified as Group III. Group IIIconsisted of 2807 newborns. According to the baby’s bloodgroup, either A, B, or AB, Group I was further divided to AO(mother O group with baby A), BO (mother O group withbaby B), and ABO (mother O group with baby AB) subgroupsand there were 305, 247, and 3 in the AO, BO, and ABsubgroups, respectively. Maternal, neonatal, and cordblood data was then obtained from medical records foreach CBU sample.

2.4.Statistical analysis

The gender of neonate and delivery route for the studiedgroups were compared by Chi-square test. The significanceof differences for the CBU mean values for Hb, Hct, nRBCratios, and RBC together with maternal age, gestationalage, neonate’s birth body weight, and cord blood weightacross the three groups were evaluated by ANOVA. We usedStudentttest to evaluate the differences in mean Hb, Hct,nRBC ratios, and RBC across the AO and BO subgroups ofGroup I. Apvalue<0.05 was considered to be statisticallysignificant.

3. Results

Totally, 3688 units of cord blood were eligible for analysis,and none of the mothers/infants were found to be DRh positive blood type. There were 555, 326, and 2807 infantsin Groups I, II, and III, respectively. In Group I, there were305 infants in the AO subgroup and 247 infants in the BOsubgroup. The mean, median, minimum, and maximumvalues for Hb, Hct, nRBC ratio, RBC, maternal age, gestational age, baby’s birth body weight, and cord blood weightacross the three groups are listed inTable 1. This showsthat there was no statistical difference for gender (male/female ratio) or delivery route (vaginal delivery vs.cesarean section) across the groups. However, the meansfor Hb, Hct, and RBC values for Group I were significantlylower than those for Groups II and III (Hb: 10.6 g/dL, 10.8 g/dL, and 10.9 g/dL, respectively,p<0.001; Ht: 35.1%,35.7%, and 36.2%,p<0.001; RBC: 3.01,×10⁶/μL,3.11×10⁶/μL, and 3.14×10⁶/μL,p<0.001). Although themean Hb, Hct, and RBC values for Group II were lower thanfor Group III, these differences were not statisticallysignificant. Furthermore, no significant differences werefound among Groups I, II, and III when their nRBC ratios,maternal age, gestational age, infant’s birth body weight,and cord blood weight were analyzed. These findings arepresented inTable 2. In Group I, the mean Hct value of theBO subgroup was significantly lower than that of the AOgroup (34.7% for the BO group vs. 35.4% for the AO group,pZ0.04) (Table 3). Although there were no statisticaldifferences in mean Hb, nRBC ratios, and RBC between theAO and BO subgroups, the mean Hb and RBC values for theAO group were higher and the nRBC ratios lower than forthe BO group (HbZ10.7 g/dL, 10.4 g/dL, respectively,pZ0.06; RBCZ3.04×10⁶/μL, 2.98×10⁶/μL, respectively,

p=0.10; nRBC=3.78, 3.84, respectively, p=0.93)

(Table 3).

4. Discussion

Many etiological factors are involved in the developmentof neonatal hyperbilirubinemia. For example, UDP glucuronosyl transferase 1A1 (UGT1A1) is the key enzymein bilirubin conjugation, and Huang et al’s study on variation at nucleotide 211 of theUGT1A1gene shows that this isa risk factor for the development of neonatal hyperbilirubinemia among Taiwanese. The variant rate within thecoding region of theUGT1A1gene in Taiwanese was foundto be 29.3%.12In addition, ABO incompatibility is considered to be one of the most common causes of neonatalhyperbilirubinemia. The major anti-A and anti-B antibodiesin blood group B and A individuals are immunoglobulin M(IgM); however, these antibodies in group O individuals areusually mostly IgG. The only immunoglobulin transferredfrom mother to fetus via the placenta is IgG. Lin et alreported that mothers and infants had the same anti-A andanti-B IgG titers when group O mothers and infants werecompared.13ABO HDN is considered to occur relativelyfrequently and can be a significant cause of neonatalmorbidity. In 14.3% of all pregnancies in Taiwanese, the

mother is O and her neonate is either A or B.14However,ABO incompatibility-induced clinically severe hemolyticdisease is comparatively rare. There are three main reasonsfor this. First, A and B blood type antigens are expressed ata low level on fetal RBCs. Second, IgG ABO antibodies areusually IgG2, which does not initiate RBC destruction.Thirdly, ABO antigens are present in many tissues, and anyIgG antibodies crossing the placenta are likely to becomebound to placental tissue.15e18As a result, this disease isusually mild; nevertheless, severe hemolysis may occasionally occur. Taiwanese neonatal jaundice has been foundto be more severe than that found in Black and Caucasianpopulations.9Desjardins et al reported that among 1704infants of blood group O mothers, infants with blood groupsA or B had significantly lower cord blood Hb concentrationsthan did newborns with blood group O; they concluded thatmost ABO-incompatible infants showed some degree ofhemolytic disease even though antibodies were notdemonstrated by either the Coombs or eluate test.19In ourstudy, Group I infants have a lower mean Hb, Hct, and RBCthan the other two groups, and this supports Desjardinset al’s findings. Chen and Ling demonstrated that a high IgGtiter among group O mothers was associated with thedevelopment of ABO HDN in Taiwan.10However, in Taiwan,the cause of neonatal hyperbilirubinemia is quite complicated because close to 30% of the population have theUGT1A1gene mutation associated with hyperbilirubinemia.

In group A and B individuals, naturally occurring anti-Band anti-A antibodies are mainly IgM, which cannot crossthe placenta. By contrast, in the group O individuals, anti-Aand anti-B antibodies are predominantly IgG, which cancross the placenta. Dufour and Monoghan reported that37.9% of group O mothers having blood group A or B infantsshowed laboratory evidence of ABO HDN, whereas only 0.8%of mothers of group A with infants having blood group B orAB and mothers of group B with infants having blood groupA or AB showed laboratory evidence of HDN.20In our study,the values for Hb, Hct, and RBC in Group I were significantlylower than those for Groups II and III (Hb: 10.6 g/dL, 10.8 g/dL, and 10.9 g/dL, respectively,p<0.001; Ht: 35.1%,35.7%, and 36.2%,p<0.001; RBC: 3.01×10⁶/μL,3.11×10⁶/μL and 3.14×10⁶/μL,p<0.001). Furthermore,Group II had lower Hb, Hct, and RBC values than Group III(Hb 10.8 g/dL, 10.9 g/dL, respectively; Hct 35.7%, 36.2;RBC 3.11×10⁶/μL, 3.14×10⁶/μL), but these differencesdid not reach statistical significance. Thus, our results wereconsistent with Dufour and Monoghan’s findings.

nRBCs are immature RBCs. Many pathological conditions,including hemolytic disease and bleeding, can cause anincrease in the number of nRBCs. Hanlon-Lundberg and Kirbyevaluated 1661 neonates and reported that cord blood nRBCcounts were lower in infants with ABO compatibility than ininfants with ABO incompatibility and that group B infantswho were borne by group O mothers had the highest nRBCcounts.21However, the differences in nRBC ratios in ourthree groups did not reach statistical significance; furthermore, there was also no significant difference between theAO and BO subgroups for the nRBC ratios. Their studyobtained the CBC samples directly from the umbilical vein.Our study obtained the CBC samples from a cord blood bagthat contained 28 mL of anticoagulant, and this might havecaused a dilution effect. This difference in CBC samplecollection may explain the different results.

HDN in Group B infants due to blood group antibodieswas claimed to be more severe than HDN hemolyticdisease in Group A infants in some studies.4,22,23However,some other studies have reported no difference inseverity between AO and BO incompatibility.9,11In ourstudy, mean Hb and RBC were 10.7 g/dL, 3.04×10⁶/μL,and 10.4 g/dL, 2.98×10⁶/μL in AO and BO group,respectively. Although the mean Hb and RBC values forthe BO group were lower than those for the AO group, thedifference did not reach statistical significance. Furthermore, although the nRBC ratio in the BO group was higherthan in AO group, again, the difference was not statistically significant (3.78 and 3.84 in AO and BO groups,respectively). Interestingly, the Hct for the AO group washigher than for the BO group (35.4% and 34.7% in AO andBO subgroups, respectively) and this difference wasstatistically significant (p=0.04).

CBU Hb, Hct, and RBC values are part of the routinedata collected by all cord blood banks. We have established Taiwanese cord blood CBC normal reference valuesand the mean values for Hb, Hct, RBC, and nRBC are11.2 g/dL, 36.9 (%), 3.22×10⁶/mL, and 3.4/100 WBC),respectively.24We can probably use these RBC indices asa reference for predicting ABO HDN. In this study, therewas no personal identification data available because ofconfidentiality regulations of the public cord blood bank,and therefore, a clinical analysis of the incidence ofhyperbilirubinemia was impossible. Nonetheless, it seemslikely that CBU Hb, Hct, reticulocyte count, direct Coombstest, eluate test, and bilirubin result may help with predicting the severity of hyperbilirubinemia in ABOHDN.4e10,13,25However, these laboratory results do notprovide a reliable early diagnosis for ABO HDN becauseneonatal hyperbilirubinemia involves multiple etiologies.These etiologies includeUGT1A1gene variations, maturityof the neonatal liver, conditions that increase bilirubinproduction, conditions that increase enterohepatic circulation and breast feeding failure. Further study needs tofocus on the association between CBU laboratory values and clinical outcome in order to clarify the clinical role ofthe CBU RBC indices.