HEMOLYSIS AND HYPERBILIRUBINEMIA IN ABO BLOOD GROUP HETEROSPECIFIC NEONATES

Abstract

Objective

We quantified hemolysis and determined the incidence of hyperbilirubinemia in direct antiglobulin titer (DAT) positive, ABO heterospecific neonates and compared variables among O-A and O-B subgroups. Study design Plasma total bilirubin (PTB) was determined predischarge and more frequently if clinically warranted, in DAT positive, blood group A or B neonates of group O mothers. Heme catabolism (and therefore bilirubin production) was indexed by blood carboxyhemoglobin corrected for inspired carbon monoxide (COHbc). Hyperbilirubinemia was defined as any PTB concentration >95th percentile on the hour-of-life-specific bilirubin nomogram.

Results

Of 164 neonates, 111 were O-A and 53 O-B. Overall, 85 (51.8%) developed hyperbilirubinemia, which tended to be more prevalent in the O-B than O-A neonates (62.3% vs. 46.8% respectively, p=0.053). Importantly, more O-B than O-A newborns developed hyperbilirubinemia at <24 hours (93.9% vs. 48.1%, p<0.0001). COHbc values were globally higher than our previously published newborn values. Babies who developed hyperbilirubinemia had higher COHbc values than the already high values of those non-hyperbilirubinemic, and O-B newborns tended to have higher values than O-A counterparts.

Conclusions

DAT positive, ABO heterospecificity is associated with increased hemolysis and a high incidence of neonatal hyperbilirubinemia. O-B heterospecificity tends to confer even higher risk than O-A counterparts.

ABO blood group heterospecific (mother group O, newborn A or B) newborns are at risk for hyperbilirubinemia due to immune based hemolysis (1). The hemolysis occurs when maternal immunoglobulin G anti-A or anti-B antibodies cross the placenta and attach to the apposite antigen site on the neonatal red cell. Resultant heme catabolism increases bilirubin production. For each molecule of bilirubin produced, equimolecular quantities of carbon monoxide (CO) are produced. Carboxyhemoglobin (COHb) quantification may index the rate of bilirubin production (2).

The direct antiglobulin titer (DAT) test is regarded as the cornerstone of diagnosis of immune hemolytic disease of the newborn (1). Some reports suggest that this test may be only a weak predictor of severe hyperbilirubinemia (3,4) and encountered only infrequently among infants readmitted for jaundice (5). However, these reports contrast with the prominence of cases of ABO heterospecificity in recent communications on severe hyperbilirubinemia and/or bilirubin encephalopathy (6–9). Furthermore, blood group incompatibility with a positive DAT is listed by the Subcommittee on Hyperbilirubinemia of the American Academy of Pediatrics (AAP) as a major risk factor for the development of severe hyperbilirubinemia and also as a risk factor for neurotoxicity (10,11). ABO heterospecificity may therefore have serious consequences.

To capture bilirubin dynamics during the first days of life, the 95^th percentile on the hour of life specific nomogram has been used to define hyperbilirubinemia (12–15). Furthermore, total bilirubin concentrations greater than this percentile may be predictive of severe hyperbilirubinemia. Our objective was to reevaluate the contribution of DAT positive ABO heterospecificity to neonatal hyperbilirubinemia by assessing the incidence of jaundice using this new definition of hyperbilirubinemia. We further assessed the incidence of hyperbilirubinemia occurring <24 hours, which may be indicative of hemolysis (10), and compared the risk of hemolysis and hyperbilirubinemia between O-A and O-B subgroups. We quantified the degree of hemolysis by measuring blood COHb corrected for inhaled (ambient) CO (COHbc) (2).

METHODS

The study was approved by the Institutional Review Board of the Shaare Zedek Medical Center. Because of the benign nature of the study which did not involve randomization or administration of a study drug, oral parental consent only was required. The clinical wing of the study was conducted in the well-infant nurseries of the Shaare Zedek Medical Center from January 2006 to April 2007. A sample of consecutive (except for the conditions mentioned below) DAT-positive blood group A or B infants who were born at ≥37 weeks gestation to mothers with blood group O was selected for enrollment in the study shortly after birth and. To avoid any inclusion bias, patients were included before any plasma total bilirubin (PTB) results were available. Neonates with any obvious condition likely to increase jaundice, other than ABO incompatibility, such severe bruising, sepsis, Down syndrome, glucose-6-phosphate dehydrogenase (G-6-PD) deficiency, or a positive DAT from any cause other than ABO isoimmunization, were excluded. Similarly, DAT positive, ABO heterospecific newborns who were also Rh-positive and born to Rh-negative mothers were excluded because of the difficulty in differentiating a positive DAT caused by Rh isoimmunization from that caused by ABO heterospecificity. Newborns were not enrolled into the study on weekends, or secular or religious holidays.

Routine management of infants born to blood group O mothers at the Shaare Zedek Medical Center was previously described (16). Blood type and DAT tests were routinely performed on cord blood on all infants born to blood group O mothers. Results were available within 24 hours of delivery, and frequently sooner.

Newborns were assessed visually for jaundice at the time of admission to the nursery and subsequently at least once per nursing shift. PTB testing was performed on any newborn with jaundice appearing within the first 24 hours, and after that time period as clinically warranted. During routine, predischarge metabolic screening, a PTB determination was performed on DAT positive infants.

All PTB results were plotted on the hour of life specific bilirubin nomogram, and the percentile and risk category determined. Phototherapy for DAT-positive infants was instituted in accordance with the 2004 AAP guidelines for neonates with risk factors (10). After discharge, follow-up PTB determination was performed when necessary on an outpatient basis at our hospital. Indications for follow up were based on the risk category designated by the predischarge PTB concentration, according to the guidelines of the Israel Neonatal Society for the management of neonatal hyperbilirubinemia and prevention of kernicterus (17).

Blood sampling for COHb determination was performed prior to discharge at the time of routine metabolic screening. Simultaneous with the COHb sampling, a sample of air from the nursery in which the infant was being cared for was collected for CO analysis. The timing of metabolic screening was suited to COHb sampling of neonates of smoking mothers, because, by 48 hours, there should no longer be any effect of the smoking on the newborns’ COHb concentrations (18).

COHb

Blood for COHb determination (150 μL) was collected in custom-prepared capillary tubes containing heparin and saponin, supplied by Stanford University. The filled tubes were sealed and the contents mixed, stored at −18°C, and sent on wet ice to Stanford University. COHb was determined as a percentage of total hemoglobin (tHb) in one batch by a gas chromatographic method (18). The tHb latter was measured using the same blood sample by a cyanmethemoglobin method (18). The within-day and between-day coefficients of variation for this method for reference blood samples are 3% and 8%, respectively. CO content of the sampled ambient air was measured at Shaare Zedek Medical Center by using a CO analyzer supplied for this purpose by Stanford University and measured COHb values corrected for inspired CO to derive COHbc, as described (18).

Previously published COHbc and tHb values for neonates from this nursery that were analyzed using the same methodology and Stanford University laboratory, are supplied for comparison. The reference newborns were not DAT positive or G-6-PD deficient (19,20).

DAT and Blood Type

DAT testing was performed routinely in the blood bank of the Shaare Zedek Medical Center on umbilical cord blood using an agglutination technique and reported on a scale of ± to ++++ [DiaMed-IDMicroTyping System, ID-Card “LISS/Coombs” (DiaMed AG, Cressier s/Morat, Switzerland)]. Blood group typing was performed routinely on umbilical cord blood using standard blood bank techniques.

Plasma Total Bilirubin

Routine PTB testing was measured on heparinized, centrifuged, capillary tube samples by absorbance of bilirubin at 455 nm (NEO BIL Model A2 [Digital and Analog Systems, Rome, Italy]).

Data Analysis

Hyperbilirubinemia was defined as any PTB value >95^th percentile on the hour-of-life-specific nomogram (12).

COHbc, tHb values and factors relating to hyperbilirubinemia were compared between the O-A and O-B subgroups, and also, within the subgroups, between hyper- and non-hyperbilirubinemic neonates. Categorical variables were compared using χ² analysis. Continuous variables with a normal distribution were compared using Student t test, and in others a Mann-Whitney Rank Sum test was used. The incidence of hyperbilirubinemia was compared by calculating the Relative Risk (95% confidence interval), in which case significance was defined as a 95% confidence interval that did not include 1. For other comparisons, significance was defined as a p value <0.05.

RESULTS

One hundred sixty-four DAT positive, blood group A or B newborns born to blood group O mothers were enrolled between January 2006 and April 2007 (Table I). Overall, 85 (51.8%) developed a PTB value >95th percentile for hour of life at any point. Age at first PTB >95th percentile was 19 ± 11 hr (range 1–48 hr, one additional newborn was readmitted at age 80 hr), with corresponding PTB 9.9 ± 2.5 mg/dL (range 5.1–17.8 mg/dL). Early hyperbilirubinemia (PTB >95th percentile during the first 24 hours) was noted in 56 (34.1% of the cohort, 66.7% of those with hyperbilirubinemia) and in 27/56 (48.2%) PTB >95th percentile was recorded within the first 12 hours. Mean (SD) PTB for the first reading >95^th percentile for the ≤24 hr group was 8.9 ± 1.9 mg/dL, and that for the ≤12 hr subgroup was 8.4 ± 1.9 mg/dL. Of the remaining 29 hyperbilirubinemic neonates, all but 1 became hyperbilirubinemic between 25–48 hours (31 ± 6 hrs). Phototherapy was administered to 80 neonates of the entire group at an average PTB concentration of 10.6 ± 3.0 mg/dL at mean age 22 ± 17 hr. Of the 53 presenting with hyperbilirubinemia at ≤24 hr of age phototherapy was administered to 51 (mean PTB 9.2 ± 2.0 mg/dL at 13 ± 8 hr). All neonates responded to phototherapy. None required intravenous immune globulin infusion or exchange transfusion.

Table 1.

Demographic details of the newborns reported for the entire group and O-A and O-B subgroups.

	Total	O-A	O-B	Significance O-A vs. O-B
Number of neonates (n)	164	111 (67.7%)	53 (32.3%)
Birth weight (gm. Mean ± SD)	3401 ± 425	3434 ± 450	3342 ± 359	p=0.2
Gestational age (wks)	39 ± 1	39 ± 1	39 ± 1	p=1
Males (%)	44	47	41	p=0.6
Cesarian section (%)	10	9	11	p=0.9
Breast feeding (exclusive or partial) (%)	88	89	85	p=0.6
Maternal smoking (%)	4	4	6	p=0.8
Jewish:Arab (n)	150 (92%): 14 (8%)	104 (93%): 7 (7%)	46 (87%): 7 (13%)	p=0.7

A comparison between the O-A and O-B subgroups is summarized in Table II. Although there was a trend in the O-B subgroup towards higher risk of developing hyperbilirubinemia in general, significantly more O-B newborns did develop hyperbilirubinemia within the 1^st 24 hours than O-A counterparts,. The age at which the first PTB value >95^th percentile was noted was earlier in the O-B neonates. Similarly, there was a trend for phototherapy to be commenced earlier in the OB neonates.

Table 2.

Comparison of hyperbilirubinemia and need for phototherapy between neonates with O-A and O-B heterospecificity.

Category	O-A	O-B	Significance
Number of infants	111	53
Hyperbilirubinemia (n) (Relative Risk, 95% Confidence Interval)	52 (46.8%)	33 (62.3%)	RR 1.34, 95% CI 0.99–1.77, p=0.053
Age at 1st PTB >95th centile (h) (mean ± SD)	20.6 ± 11.5	15.3 ± 9.8*	p=0.035
Hyperbilirubinemia within 1st 24 hr (n)	25/52 (48.1%)	31/33 (93.9%)*	p<0.0001
1st PTB >95th centile (mg/dL)	10.2 ± 2.6	9.3 ± 1.9	p=0.17
Phototherapy (n)	48 (43.2%)	33 (62.3%)	p=0.14
Age at which phototherapy commenced (h) [median (interquartile range)]	21.5 (12–32)	17 (7–27)	p=0.07
PTB at commencement of phototherapy (mg/dL)	11.2 ± 3.2	9.8 ± 2.4*	p=0.04

Hyperbilirubinemia: Any PTB value >95^th percentile; RR: Relative Risk; CI: Confidence Interval

COHbc studies for the entire cohort, as well as for the O-A and O-B subgroups, are presented in Table III. Overall, COHbc values were significantly higher than those of a previously reported newborn cohort (19). Both in the entire cohort, as well as within the O-A and O-B subgroups, those who developed hyperbilirubinemia had higher COHbc values than the already high values of those who did not have any documented PTB value >95^th percentile. Furthermore, the percentage of newborns developing hyperbilirubinemia increased in tandem with increasing COHbc percentile values (Figure). There was a trend for the O-B subgroup to have COHbc values which were higher than the already high levels of the O-A subgroup.

Table 3.

Blood carboxyhemoglobin values, corrected for inspired CO (COHbc) for the entire ABO heterospecific group, and for the O-A and O-B subgroups individually. For reference, COHbc values for previously reported, healthy, non-hemolyzing neonates are included (19).

COHbc (% tHb)	Entire ABO Group	O-A	O-B	Reference Group	Significance
Overall
n =	163	111	53	131
COHbc (mean ± SD)	1.24 ± 0.40*	1.20 ± 0.38	1.32 ± 0.441	0.77 ± 0.19	*p<0.0001, entire group vs. reference group; 1p=0.07, O-B vs. O-A
Hyperbilirubinemia
n =	85	52	33
COHbc	1.42 ± 0.39	1.40 ± 0.36	1.45 ± 0.45
Non-hyperbilirubinemia
n =	78	59	20
COHbc	1.00 ± 0.25	0.99 ± 0.26	1.04 ± 0.24
Significance	p<0.001	p<0.001	p=0.002

Figure.

Incidence of hyperbilirubinemia, defined as any plasma total bilirubin value >95^th percentile on the hour of life specific bilirubin nomogram, graded by corrected carboxyhemoglobin (COHbc) percentile value.

COHbc percentile ranges (% tHb): <50^th percentile, 0.54–1.19; 50–74^th percentile, 1.20–1.44; 75–90^th percentile, 1.45–1.76; >90^th percentile, 1.79–2.62

tHb values were 17.4 ± 2.9 g/dL for the cohort. These were lower than our previously reported reference values (19.0 ± 2.4 g/dL, p<0.001) (20). Within the ABO heterospecific cohort, tHb values for the hyperbilirubinemic newborns were lower than for those non-hyperbilirubinemic (16.9 ± 2.9 g/dL vs. 18.2 ± 2.7 g/dL, p=0.007). Even in those ABO heterospecific neonates who did not become hyperbilirubinemic, tHb values were lower than those of the reference group. Even though the tHb values were somewhat lower for the O-B neonates, the difference between these and the O-A group was not significant (17.0 ± 3.1 g/dL vs. 17.7 ± 2.8 g/dL, p=0.2).

DISCUSSION

Of the neonatal population delivered at the Shaare Zedek Medical Center, 21% comprise blood group A or B newborns born to group O mothers of which 15% are DAT positive (16). In the present study we documented a 52% incidence of hyperbilirubinemia in the DAT positive subgroup. This incidence is clearly many-fold that of other population groups studied using the identical definition of hyperbilirubinemia. For example, in a multicenter, multinational study of 1370 newborns, 8.8% developed hyperbilirubinemia (13). In an African American male cohort, hyperbilirubinemia was documented in 6.7% of 436 G-6-PD normal, control infants and 21.9% of 64 G-6-PD deficient newborns (14). Further adding to the high risk nature of these infants is the high incidence of hyperbilirubinemia occurring within the first 24 hours. However, almost all babies who developed hyperbilirubinemia did so within the first 48 hours, implying little risk of subsequent hyperbilirubinemia in those discharged after that timeframe. Many of the babies had combinations of conditions listed by the AAP as major risk factors for the development of severe hyperbilirubinemia: predischarge bilirubin value in the high risk zone; hemolysis due to blood group incompatibility with positive DAT; jaundice appearing within the first 24 hours, and exclusive breast feeding (10).

A correspondingly high number of newborns met the requirement for phototherapy. We cannot foretell the bilirubin dynamics had phototherapy not been instituted at low concentrations of PTB in accordance with AAP guidelines. It is possible that in some neonates the serum bilirubin concentrations would have leveled off and not met the criteria for phototherapy at a later age (3,21).

Considering that kernicterus is rarely encountered and exchange transfusion nowadays is unusual, it is not surprising that, despite the high incidence of hyperbilirubinemia, we encountered no cases of this nature. Although our findings are consistent with previous reports demonstrating that few DAT positive, ABO heterospecific neonates will meet the criteria for exchange transfusion (3,5), newborns of these blood group combinations do appear prominently in recent series of severe hyperbilirubinemia and kernicterus (6–9). The apparent mild nature of the disease should not cause complacency regarding ABO hemolytic disease.

There may be an increased effect of hemolytic conditions on the development of bilirubin induced neurologic dysfunction (22,23). Our elevated COHbc results, comptatible with previously reported data (13, 24–26), confirm the hemolytic nature of DAT positive ABO heterospecificity in general and emphasize the high risk nature of these neonates. Despite the overall high rate of hemolysis, not all infants in our series who developed hyperbilrubinemia had high levels of COHbc. The percentage of hyperbilirubinemic neonates increased along with increasing COHbc percentile values. We speculate that, in the lower percentile groups, varying degrees of immaturity of the bilirubin conjugating system, or presence of the (TA)₇ promoter variant of the UGT1A1 gene associated with Gilbert syndrome, along with moderately increased heme catabolism, may have contributed to the hyperbilirubinemia (27). However, in those neonates with the highest COHbc levels, all babies developed hyperbilirubinemia. In this latter group, the degree of hemolysis must have been sufficiently high to overwhelm even the most efficient conjugation process.

The literature is inconsistent with regard to the degree of hemolysis and the incidence and severity of hyperbilirubinemia between O-A and O-B subgroups. Several investigators were unable to show any difference in clinical severity between O-A and O-B hemolytic disease of the newborn, although in the former report there was a trend towards performing exchange transfusion during the first 24 hours more frequently in O-B compared with O-A infants (28–30). Similarly, a retrospective analysis of ABO hemolytic disease did not find significant relationships between the infants’ blood type and clinical outcome (31). Sisson (32) and Kaplan (33) reported no significant differences in severity or response to therapy between the two blood types. An infant whose blood group was A was as likely to be affected by ABO hemolytic disease as a blood group B infant (34). However, Bakkeheim et al found a significantly increased rate of invasive treatments, including intravenous immune globulin therapy and exchange transfusion, in O-B infants compared with O-A (35). Two studies documented a higher need for exchange transfusion in O-B neonates than in O-A (36, 37).

The reasons for the apparently increased severity of jaundice in the O-B subgroup are not clear. Maternal factors including differing levels of IgG anti-A or anti-B may affect the degree of hemolysis, and variations in IgG subgroup distribution may alter macrophage induced RBC clearance. The number of fully developed A or B antigen sites on fetal RBCs may be different, as may be the dilutional effect of other tissues bearing these surface antigens (38).

Comparisons of these studies are difficult as inclusion criteria, definitions, therapeutic indications, and time epochs differed from study to study. In contrast, we categorized our patients using an up to date definition of hyperbilirubinemia which takes into account the bilirubin dynamics of the first days of life (12). Standardized criteria for treatment were employed (10), and the rate of heme catabolism was assessed by a state of the art method (2). Selection bias was avoided by including all DAT positive ABO heterospecific newborns at birth, and prior to their becoming jaundiced. Although not all our comparisons achieved statistical significance, O-B heterospecific neonates did appear to be at higher risk than the O-A subgroup. Hyperbilirubinemia did occur earlier in the O-B infants and more O-B newborns developed hyperbilirubinemia within the first 24 hours. This difference was attributable to moderately, although not statistically significantly, increased hemolysis. Lack of statistical significance, however, does not preclude an effect of moderately increased hemolysis resulting in increased bilirubin production. Coupled with immaturity of the bilirubin conjugating system, moderately increased hemolysis may have resulted in a higher incidence of hyperbilirubinemia. A shortcoming of our study was the dependence on visual recognition of early jaundice, and lack of a fixed protocol for PTB determinations during the first 24 hours.

In summary, our study provides information about the incidence, severity and mechanism of jaundice in ABO heterospecific neonates. A high rate of hemolysis is a hallmark of DAT positivity with a resultant high incidence of hyperbilirubinemia and especially early hyperbilirubinemia. These factors have the potential of increasing the risk of severe hyperbilirubinemia and bilirubin neurotoxicity. Furthermore, O-B heterospecific neonates appear to be at higher risk than their O-A counterparts

ABBREVIATIONS

AAP

American Academy of Pediatrics

CO

carbon monoxide

COHb

carboxyhemoglobin

COHbc

carboxyhemoglobin corrected for inspired CO

DAT

direct antiglobulin titer

G-6-PD

glucose-6-phosphate dehydrogenase

PTB

plasma total bilirubin

RBC

red blood cell

tHb

total hemoglobin