Objectives. We assessed the incidence of hyperbilirubinemia, defined as serum total bilirubin ≥15 mg/dL (256 μmol/L), in a cohort of Sephardic Jewish female neonates at risk for glucose-6-phosphate dehydrogenase (G-6-PD) deficiency with especial emphasis on the heterozygotes. We studied the roles of hemolysis by blood carboxyhemoglobin (COHb) determinations and of the variant promoter of the gene for the bilirubin-conjugating enzyme uridine 5′-diphosphate glucuronosyltransferase 1 (UGT1A1) seen in Gilbert's syndrome in the pathogenesis of the hyperbilirubinemia.
Methods. Consecutively born, healthy, term, female neonates were screened for G-6-PD deficiency and observed clinically with serum bilirubin evaluations as indicated for hyperbilirubinemia. On day 3, blood was sampled for COHb, total hemoglobin (tHb), and a mandatory serum bilirubin determination. COHb, determined by gas chromatography, was expressed as percentage of tHb and corrected for inspired carbon monoxide (COHbc). DNA was analyzed for the G-6-PD Mediterranean563T mutation and for the variant UGT1A1 gene.
Results. The cohort included 54 G-6-PD-deficient heterozygotes, 19 deficient homozygotes, and 112 normal homozygotes. More heterozygotes (12/54, 22%; relative risk: 2.26; 95% CI: 1.07–4.80) and deficient homozygotes (5/19, 26.3%; relative risk: 2.68; 95% CI: 1.05–6.90) developed hyperbilirubinemia, than did normal homozygotes (11/112, 9.8%). Third-day serum bilirubin values that were obtained from 144 neonates were significantly higher in both heterozygotes (11.2 ± 3.7 mg/dL [192 ± 64 μmol/L]) and G-6-PD-deficient homozygotes (12.0 ± 3.0 mg/dL [206 ± 52 μmol/L]) than in the G-6-PD normal homozygotes (9.4 ± 3.4 mg/dL [160 ± 58 μmol/L). In contrast, COHbc values were higher only in G-6-PD-deficient homozygotes (0.74% ± 0.14%) and not in heterozygotes (0.69% ± 0.19%, not statistically significant), compared with control values (0.63% ± 0.19%). High COHbc values were not a prerequisite for the development of hyperbilirubinemia in any of the G-6-PD genotypes. A greater incidence of hyperbilirubinemia was found among the G-6-PD-deficient heterozygotes, who also had the variant UGT1A1 gene, in both heterozygous (6/20, 30%) and homozygous (4/8, 50%) forms, than was found in their counterparts with the normal UGT1A1 gene (2/26, 7.7%). This effect was not seen in the G-6-PD normal homozygote group. A color reduction screening test for G-6-PD deficiency identified only 20.4% (11/54) of the heterozygotes.
Conclusions. We showed that G-6-PD-deficient heterozygotes, categorically defined by DNA analysis, are at increased risk for neonatal hyperbilirubinemia. The screening test that was used was unable to detect most heterozygotes. Increased bilirubin production was not crucial to the development of hyperbilirubinemia, but presence of the variant UGT1A1 gene did confer increased risk. bilirubin, carbon monoxide, carboxyhemoglobin, females, gas chromatography, Gilbert's syndrome, glucose-6-phosphate dehydrogenase deficiency, hemolysis, hyperbilirubinemia, neonates, polymerase chain reaction, Sephardic Jews, screening test, uridine 5′-diphosphate glucuronosyltransferase.
Glucose-6-phosphate dehydrogenase (G-6-PD) deficiency is a common genetic disorder with worldwide distribution.1Because it is an X-linked condition, the female genotype may be normal homozygote, G-6-PD-deficient homozygote, or heterozygote. The latter group is unique to female subjects because only female subjects have the potential of inheriting both a normal and a deficient G-6-PD gene, because they possess two X-chromosomes.2,,3 Instead of heterozygotes having a homogenous red blood cell (RBC) population with intermediate G-6-PD enzyme activity, X-chromosome inactivation results in two RBC populations. One population consists of RBCs with normal G-6-PD activity, the other consists of G-6-PD-deficient cells. Usually approximately half of the cells are deficient and half are normal. However, because X-inactivation may be nonrandom, or one or the other clone may be selected preferentially, there may be varying phenotypes, and the RBCs of heterozygotes may exhibit normal, intermediate, or grossly deficient G-6-PD activity.
Because of this apparent discrepancy between phenotype and genotype, diagnosis of the G-6-PD genotype in female subjects is particularly problematic. Commonly used screening tests, quantitative tests of enzymatic activity, and cytochemical methods are all incapable of accurately distinguishing heterozygotes from homozygous normal or homozygous-deficient individuals.1 Although they are not available for mass screening, advances in molecular biological techniques have made accurate detection of heterozygotes possible by identification of the mutation in genomic DNA. Nonrandom X-chromosome inactivation does not alter the aberrant nucleotide sequence of the gene, and if the mutation sought for is known, heterozygote identification now can be accomplished reliably.1,,4
Severe neonatal hyperbilirubinemia resulting in extreme cases to kernicterus and death is a potentially serious complication of G-6-PD deficiency.1,,5 Because of the previously mentioned difficulties in determining the female G-6-PD genotype, very few studies of G-6-PD deficiency-associated neonatal jaundice have included female subjects. The aim of this study was to evaluate prospectively the risk of neonatal hyperbilirubinemia in a cohort of female subjects at high risk for G-6-PD deficiency with special emphasis on the heterozygotes. The three forms of the G-6-PD genotype were categorized by molecular DNA techniques. The rate of bilirubin production, reflecting hemolysis, was assessed by accurate determination of carboxyhemoglobin (COHb) corrected for inspired carbon monoxide (CO) (COHbc),6,,7 and the role of the variant promoter of the gene encoding the bilirubin-conjugating enzyme, uridine 5′-diphosphate glucuronosyltransferase 1A1 (UGT1A1),8 seen in Gilbert's syndrome, was evaluated.
Patient Population and Clinical Protocol
The population studied was drawn from subsets of the Sephardic Jewish population living in Israel and who were at high risk for G-6-PD deficiency. This is not a homogenous community but rather consists of several subgroups whose ancestors, exiled from their land over 2 millennia ago, had lived in Asia Minor in relative isolation.9 After establishment of the State of Israel in 1948, many immigrated to that country. Those who emigrated from Kurdistan, Iraq, Iran, Turkey, and Syria have an especially high incidence G-6-PD deficiency,10,,11 which has been found to be of the Mediterranean 563T mutation.8,,12
The study was approved by the Shaare Zedek Medical Center's Institutional Review Board. A cohort of healthy, term female neonates, born at the Shaare Medical Center to Sephardic Jewish mothers at high risk for G-6-PD deficiency, formed the patient group for this study. Most of these neonates had been included as part of a larger cohort in a previous study of G-6-PD deficiency and the variant UGT1A1 gene promoter.8 Neonates with any condition other than G-6-PD deficiency that was known to exacerbate the incidence or severity of neonatal hyperbilirubinemia, such as direct Coombs' positive hemolytic anemia, maternal diabetes, cephalhematoma, or sepsis, were excluded from the study. Umbilical cord blood was collected into a EDTA-containing tubes for extraction of DNA. A screening test for G-6-PD deficiency was performed within 24 hours of birth. On the third day of life, blood was drawn for serum total bilirubin, total hemoglobin (tHb) and COHb. The only exception was if the third day occurred on the Jewish Sabbath or other holy days when, owing to the religious restrictions of this medical center, only tests with an immediate clinical bearing may be performed. In such an event, the latter two tests were not obtained, and serum total bilirubin values were determined according to clinical indications. Serum total bilirubin determinations before the third day or subsequent to that day were performed as clinically indicated. G-6-PD-deficient neonates were observed in hospital for a minimum of 72 hours and were followed by us as outpatients until serum total bilirubin values were stabilized. Phototherapy was commenced in hospital if the serum total bilirubin values exceeded 15 mg/dL (256 μmol/L), and exchange transfusion was performed if bilirubin values remained >20 mg/dL (342 μmol/L) despite phototherapy for more than several hours. The compliance rate for obtaining follow-up bilirubin testing in the population served by this hospital is excellent.
The G-6-PD-deficient heterozygotes and homozygotes comprised neonates who were heterozygous or homozygous for the G-6-PD Mediterranean563T mutation, respectively, whereas the control group (normal homozygotes) comprised those from the same population group who did not have this mutation. Hyperbilirubinemia was defined as a serum total bilirubin ≥15 mg/dL (256 μmol/L) during the first week of life. Patients were classified according to the promoter sequence of the gene encoding UGT1A1 as normal homozygotes, bearing the sequence (TA)6TAA in the TATAA element of the promoter of both alleles, variant homozygotes with the sequence (TA)7TAA in both alleles, and heterozygotes with one of each in the respective alleles.13,,14
Data were analyzed by comparing the G-6-PD-deficient heterozygotes and deficient homozygotes with the G-6-PD normal homozygote control neonates. Categorical variables were compared using χ2 analysis or Fisher's exact test, as appropriate, whereas continuous variables were analyzed using the Student's t test. Significance was defined asP < .05. Relative risk was calculated to estimate the risk of developing hyperbilirubinemia in any of the G-6-PD-deficient groups relative to that of the control group. With hyperbilirubinemia as a dependent variable and G-6-PD status as an independent variable, 95% CIs were used as a measure of the statistical precision of each relative risk. These were regarded as significant when the 95% CIs were either entirely above or entirely below 1. If the 95% CI straddled 1, the results were regarded as not significant. Data for UGT1A1 polymorphism, taken from the previous study,8 were included in the current analysis, because it was believed that these data were essential to assess comprehensively the pathogenesis of the hyperbilirubinemia. The variant UGT1A1 allele frequency was determined by calculating the percentage of variant promoters of the total number of UGT1A1 alleles in the population studied. Incidence of hyperbilirubinemia was calculated for infants with the various combinations of UGT1A1 and G-6-PD genotypes.
G-6-PD screening was performed using a commercial visual qualitative color reduction kit (Kit No. 400; Sigma Diagnostics, St Louis, MO). The principle of the test involves the oxidation of glucose-6-phosphate to 6-phosphogluconate and the concomitant reduction of NADP+ to NADPH. NADPH reduces NADPH diaphorase, which, in turn, reduces the blue dye, dichlorophenol indophenol, to a colorless state. Serum total bilirubin values were determined by reflectance spectrophotometry using a Kodak Ektachem analyzer with Clinical Chemistry Slides (Ektachem 750 XRC Analyzer; Eastman Kodak, Rochester, NY). Blood groups and direct Coombs' testing were performed by routine laboratory techniques. DNA was extracted using a high salt extraction procedure.15
Molecular Classification of the G-6-PD Genotype
This procedure was performed at the Scripps Research Institute. A 127-bp fragment of DNA containing nt 563, the nucleotide mutated in G-6-PD Mediterranean,16 was amplified by polymerase chain reaction17 using oligonucleotides 7 and 818as primers. The presence or absence of the 563T mutation was determined by allele-specific oligonucleotide hybridization. Details of the procedure have been published elsewhere.8
The infants were sampled for COHb on the third day of life, by which point there should have been no effect of maternal carboxyhemoglobin.19,,20 Blood for COHb and tHb determinations was collected into specially prepared capillary tubes containing heparin and saponin. A steel rod was inserted, the tubes were stoppered, and the blood hemolyzed by shaking. The samples were stored at 4°C and transferred on ice to Stanford University for analysis. COHb levels have been shown to remain stable for up to several months when stored or transported at this temperature.21 COHb was determined by gas chromatography and tHb was determined by a cyanmethemoglobin method, both previously described.21,,22 COHb values were expressed as a percentage of tHb. A sample of room air that each infant was breathing was analyzed for CO content, using an electrochemical CO analyzer supplied by Stanford University.23 Measured COHb values were corrected for the contribution of inhaled CO to obtain COHbc by the following formula:23
Molecular Classification of the UGT1A1 Genotype
Identification of neonates with the variant promoter for the gene bilirubin UGT1A1 was performed by amplification of the promoter region of the gene using polymerase chain reaction with primers Bili x, 5′-ATTAACTTGGTGTCGATTGG-3′, and Bili z, 5′-AGCCATGGCGGCCTTTGCTC-3′, as previously described.8 The reaction products were separated by electrophoresis on a 10% polyacrylamide gel and stained with ethidium bromide.
A total of 185 female neonates was enrolled in the study between January and December in 1996. G-6-PD Mediterranean was absent (normal homozygotes) in 112 neonates, present (G-6-PD deficient) in the heterozygous form in 54 neonates, and present in the homozygous form in 19 neonates. The mean birth weight of the cohort was 3182 ± 417 g, the mean gestational age was 39.7 ± 1.2 weeks, and the percentage of infants delivered vaginally was 98%. None of these parameters varied significantly among the three genotypes.
The risk of developing serum total bilirubin ≥15 mg/dL (256 μmol/L) was greater in both G-6-PD heterozygotes (12/54 patients, 22%; relative risk: 2.26; 95% CI: 1.07–4.80) and deficient homozygotes (5/19 patients, 26.3%; relative risk: 2.68; 95% CI: 1.05–6.90), than in the normal homozygotes (11/112 patients, 9.8%; relative risk: 1). (Fig 1). The risk of developing hyperbilirubinemia progressed in a stepwise fashion through the G-6-PD-deficient genotypes, although the risk of the heterozygotes was closer to the risk of the deficient homozygotes than to the normal controls. The additional risk of deficient homozygotes developing hyperbilirubinemia relative to heterozygotes was not significant (relative risk: 1.23; 5% CI: 0.49–3). Phototherapy, administered to 10 (18.5%) heterozygotes, 5 (26.3%) deficient homozygotes, and 8 (7.1%) normal neonates at serum bilirubin levels (mean ± SD) of 286 ± 10 μmol/L (16.7 ± 0.6 mg/dL), 266 ± 8 μmol/L (15.6 ± 0.5 mg/dL), and 290 ± 16 μmol/L (17.0 ± 0.9 mg/dL), respectively, probably did not allow the serum total bilirubin values to reach their natural peak. No infant in any of the groups required exchange transfusion or developed kernicterus.
Third-day laboratory evaluations, including, tHb, COHbc, and serum total bilirubin, were performed on 144 neonates and are summarized inTable 1. The distribution of hyperbilirubinemia in this subgroup was similar to that of the primary cohort (13.6%, 24.4%, and 27.7% for the G-6-PD normal homozygotes, heterozygotes, and deficient homozygotes, respectively). It should be noted from Table 1 that although the time of sampling was similar in all three groups, both G-6-PD heterozygotes and deficient homozygotes had serum total bilirubin values that were significantly higher than those of the controls. Furthermore, whereas 7 (15.6%) of the heterozygotes and 5 (27.8%) of the deficient homozygotes had developed a serum total bilirubin value ≥15 mg/dL (256 μmol/L) by the time of sampling, only 4 (4.9%) of the normal homozygotes had done so, indicating earlier onset of hyperbilirubinemia in the G-6-PD-deficient homozygotes (relative risk: 5.63; 95% CI: 1.67–18.90;P = .009) with a trend to the heterozygotes also becoming hyperbilirubinemic earlier (relative risk: 3.15; 95% CI: 0.97–10.18; P = .054).
In contrast to the unequivocally elevated third-day serum total bilirubin values, simultaneously drawn COHbc values were increased significantly only in the deficient homozygotes (Table 1). The heterozygotes had intermediate COHbc values that were not significantly different from the values for either normal homozygotes or deficient homozygotes. Furthermore, COHbc values of neonates who became hyperbilirubinemic were not greater than were the values of neonates whose serum total bilirubin values did not exceed 15 mg/dL (256 μmol/L) (0.72% ± 0.25% vs 0.68% ± 0.17%, NS, for the heterozygotes; 0.79% ± 0.09% vs 0.73% ± 0.16%, NS, for the deficient homozygotes, hyperbilirubinemic vs nonhyperbilirubinemic neonates, respectively). tHb values were similar in all three G-6-PD genotypes.
Data for the UGT1A1 genotypes were available for 18 G-6-PD-deficient homozygotes, 54 heterozygotes, and 110 G-6-PD normal homozygotes. Allele frequency for the variant UGT1A1 gene was 0.40 in the G-6-PD homozygous normal population and 0.31 in the G-6-PD-deficient populations, not statistically significant (NS). The effect of the variant UGT1A1 gene in the pathogenesis of hyperbilirubinemia is summarized in Table 2. Among the G-6-PD-deficient heterozygotes, presence of the variant UGT1A1 gene in both heterozygous and homozygous forms, conferred a greater risk of hyperbilirubinemia than to their counterparts in the UGT1A1 normal homozygous group. This effect was not seen in the G-6-PD normal homozygote group. The number of infants with variant UGT1A1 genes in the G-6-PD-deficient homozygote group was too small to allow for meaningful analysis.
The biochemical screening test was able to identify the normal state in 95.5% (107/112) of the G-6-PD homozygous normal subjects and the G-6-PD-deficient state in 84% (16/19) of the deficient homozygotes but only in 20.4% (11/54) of the heterozygotes. To determine the part played by RBC G-6-PD enzyme activity in the pathogenesis of the hyperbilirubinemia, those heterozygotes with an abnormal screening test, presumably with lower enzyme activity in the RBCs, were compared with those heterozygotes who had a normal reading test, in all probability with higher RBC enzyme activity. Of these subgroups, 36% of the former developed hyperbilirubinemia, compared with 18% of the latter. Those with the abnormal screening test had an increased risk of hyperbilirubinemia relative to those with a normal test (relative risk: 1.95; 95% CI: 0.71–5.30), and those heterozygotes with a normal test had a greater risk than the normal homozygotes, but, because of the smaller patient size in these subgroups, these risks did not reach statistical significance. However, values for third-day bilirubin levels (194 ± 74 vs 192 ± 72 μmol/L [11.4 ± 4.3 vs 11.2 ± 4.2 mg/dL]), COHbc determinations (0.68 ± 0.27 vs 0.69 ± 0.15%) and tHb values (18.7 ± 2.1 vs 19.7 ± 2.3 g/dL) were not significantly different between these two groups.
Because a high incidence of neonatal hyperbilirubinemia in G-6-PD-deficient Sephardic Jewish males has been well documented,8,,25,26 a similarly high incidence in the G-6-PD-deficient female homozygotes was not surprising. However, the high risk of heterozygotes developing hyperbilirubinemia was unanticipated. This occurred especially in those females who also had the variant UGT1A1 promoter. According to the World Health Organization Working Group, G-6-PD-deficient heterozygotes have sufficient enzymatic activity to protect them from the dangers of the enzyme deficiency.27 The group recommended screening to identify only enzyme-deficient male hemizygotes and deficient female homozygotes to predict neonatal jaundice. It was suggested that it is necessary to identify heterozygotes to advise them regarding the care of their potentially G-6-PD-deficient male offspring. However, female heterozygotes are not untouched by acute hemolytic crises attributable to favism,28–31 although these patients may have been those patients with RBC enzyme activity at the lower end of the spectrum. Now, as shown by the results of the current study, the risk of heterozygotes, both with higher and lower RBC enzyme activity, developing neonatal hyperbilirubinemia, is close to the risk of their G-6-PD-deficient homozygous counterparts.
Before the current investigation, no study of hyperbilirubinemia in female neonates with the frequently encountered G-6-PD Mediterranean variant has used molecular DNA technologies for the accurate categorization of the G-6-PD genotype. Indeed, the few reported studies of G-6-PD deficiency-associated neonatal jaundice in female subjects have had conflicting results. Using erythrocyte staining methods to detect the enzyme-deficient cells and the normal cells, Sanna et al32 in Sardinia found that 2.2% of 46 female heterozygotes developed hyperbilirubinemia, compared with 10.2% of hemizygous G-6-PD-deficient males, and 5.1% of G-6-PD normal males. The surprisingly low incidence of jaundice in the heterozygotes was not explained fully, although the authors did suggest a role of differing genetic constitutions and environmental influences in tempering the incidence of hyperbilirubinemia. Meloni et al,33 also in Sardinia, used G-6-PD enzymatic activity to determine the G-6-PD-deficient subgroups. They found that 14.5% of 346 G-6-PD-deficient heterozygotes, defined as those with intermediate enzymatic activity, developed hyperbilirubinemia, compared with 30% of male hemizygotes and 25% of female deficient homozygotes. Also diagnosing G-6-PD deficiency by quantitative enzymatic activity, Al-Naamah et al34 found that 51% of 186 Iraqi neonates with serum bilirubin values ≥15 mg/dL (256 μmol/L) were G-6-D deficient and of these neonates, 53% were female subjects with intermediate G-6-PD activity. Yu et al35 compared a cohort of G-6-PD-deficient Taiwanese neonates, diagnosed by a quantitative enzyme assay, to a cohort of G-6-PD normal control neonates. Although the incidence of a serum total bilirubin value ≥256 μmol/(15 mg/dL) was significantly higher in the male G-6-PD-deficient neonates, compared with male control neonates, the incidence of hyperbilirubinemia was identical in the female G-6-PD-deficient neonates and the female control neonates (6.2%). The only study to date to use molecular techniques to determine heterozygosity is that of Huang et al,36 who identified 43 heterozygotes of 50 G-6-PD-deficient female Chinese neonates in Taiwan, in whom the nucleotide 1376 (G→T) mutation was predominant. In contrast to G-6-PD-deficient males of that population group, the G-6-PD- deficient females did not have a greater incidence of serum total bilirubin ≥15 mg/dL (256 μmol/L) than controls. These authors attributed the lack of increased incidence of neonatal hyperbilirubinemia to the moderate G-6-PD activity still available in heterozygotes. However, it is possible that the difference between the present findings and their findings is that the frequency of the UGT1A1 mutation is only 0.16 among Asians,37 compared with 0.31 in the G-6-PD-deficient patients in the current study.
We have now determined that both heterozygous and homozygous G-6-PD-deficient female neonates, definitively categorized for the commonly occurring Mediterranean genotype, are at increased risk of developing hyperbilirubinemia. In addition, the onset of the jaundice was earlier in both groups. Furthermore, by the time of sampling on the third day of life, all the deficient homozygous neonates who were to develop serum total bilirubin values ≥256 μmol/L (15 mg/dL) had already done so. Of the remainder of the infants who developed hyperbilirubinemia, 58% of the heterozygotes, but only 36% of the normal homozygotes had done so by this age. For comparison, in a simultaneously studied cohort of male neonates from the same ethnic subgroups, 13 of 59 (22%) G-6-PD-deficient infants versus 11 of 130 (8.5%) nondeficient infants developed serum total bilirubin values ≥15 mg/dL (256 μmol/L) (relative risk: 2.60; 95% CI: 1.24–5.47;P = .02).8
It is not surprising that the actual number of G-6-PD-deficient heterozygotes and homozygous deviates slightly from the expected values according to the Hardy-Weinberg equilibrium. As explained above, the population that we studied was not a homogenous randomly interbreeding population, but actually a mixture of population subgroups.9 Intermarriage between the groups is a relatively recent phenomenon.
In light of the high incidence of neonatal hyperbilirubinemia in the heterozygotes, it is of special concern that the screening test used was unable to detect the majority of these neonates. Neither did the screening test function ideally in the homozygotes. An optimally performing screening test should not misclassify a G-6-PD-deficient patient as normal (false negative), and should misclassify a G-6-PD normal patient as deficient in only a few cases (false positive).3 Most likely, those heterozygotes with a normal reacting test had sufficient G-6-PD enzymatic activity in their RBCs to result in color reduction. It is possible that prolongation of the reaction time, use of more sensitive screening tests, or use of enzyme assays may have been able to detect a greater number of heterozygotes and to identify correctly a greater number of the homozygotes. However, because proportions of normal and mutant cells vary, with considerable overlap in enzyme activity between the groups, no test other than molecular analysis is capable of detecting all heterozygotes.38,,39 This may explain the discrepancies between the two studies performed in Sardinia which, stemming from the same population group, and under similar environmental conditions, should have yielded similar results.32,,33 Therefore, data from studies that used methods other than molecular DNA techniques may not be definitive, because various proportions of the heterozygotes have been misclassified.
To determine whether the increased incidence of hyperbilirubinemia is limited to those heterozygote neonates with a more severe deficiency of enzyme activity in their RBCs, we determined if the incidence of jaundice was higher in those who were detected, compared with those who eluded detection by the screening test. Although the incidence of severe jaundice was higher in those heterozygotes who were detected by the screening test, compared with those who were not detected, and those who were not detected had a higher incidence of hyperbilirubinemia than did the normal homozygotes, these differences did not reach statistical significance. Furthermore, values for third-day serum total bilirubin, COHbc, and tHb were not significantly different. Therefore, it seems that the level of enzyme activity in the RBC does play a role in the pathogenesis of the hyperbilirubinemia, but this role is relatively minor.
Increased bilirubin production is often implied to be a major component of the pathogenesis of neonatal hyperbilirubinemia.7Although G-6-PD deficiency is associated with increased hemolysis, overproduction of bilirubin is not always a major factor in the pathogenesis of the associated hyperbilirubinemia, as recently reviewed.40 Although COHbc values in the G-6-PD-deficient heterozygotes were not significantly elevated relative to the homozygous normals (P = .09), serum total bilirubin values, sampled simultaneously, were unequivocally higher (P = .007). Mean COHbc values of G-6-PD-deficient neonates, who developed hyperbilirubinemia were not significantly higher than those whose serum total bilirubin values remained ≤15 mg/dL (256 μmol/L). It is noteworthy that heterozygotes, who can be expected to have mean levels of G-6-PD activity in the intermediate range, and whose blood COHbc values were, accordingly, midway between the normal and deficient homozygotes, had an incidence of hyperbilirubinemia closer to that of the deficient homozygotes. Thus, we were unable to demonstrate a direct relationship between increased bilirubin production and the development of hyperbilirubinemia in these G-6-PD-deficient female subjects.
Because at any point in time, serum bilirubin values reflect a balance between bilirubin production and bilirubin elimination, if increased bilirubin production is not the major factor in the pathogenesis of the hyperbilirubinemia, decreased bilirubin elimination must be the major factor, per definition. Serum-conjugated bilirubin fraction studies in males have reflected decreased bilirubin conjugation in G-6-PD-deficient neonates who developed hyperbilirubinemia.41,,42 The conjugated bilirubin fraction patterns were reminiscent of the patterns seen in adults with Gilbert's syndrome, a condition of benign bilirubinemia attributable to partial deficiency of the bilirubin-conjugating enzyme UGT1A1. Kaplan et al8 have demonstrated an interaction between the G-6-PD deficiency state and the gene for UGT1A1 bearing the variant promoter seen in Gilbert's syndrome.13,,14 In this female cohort, presence of the variant UGT1A1 gene increased the incidence of hyperbilirubinemia significantly only when in combination with the G-6-PD deficiency gene. The effect of the variant UGT1A1 gene could not be attributed to an increased frequency of that allele in the G-6-PD-deficient heterozygotes. Neither G-6-PD deficiency in the absence of the variant UGT1A1 promoter, nor the presence of the variant UGT1A1 promoter in the absence of G-6-PD deficiency placed the infants at higher risk for hyperbilirubinemia. The latter finding was also recently demonstrated by Bancroft et al.43 A combination of both the variant UGT1A1 gene and G-6-PD deficiency seemed crucial to the development of hyperbilirubinemia. Because neither of these factors individually was a risk factor for hyperbilirubinemia, the effect of the combination of the two genes implies an interaction rather than an additive effect of these two independent risk factors.
Hundreds of millions of individuals are estimated to be affected by G-6-PD deficiency worldwide1 and a large proportion of these individuals will necessarily be heterozygotes. Only a fraction will be detected by simple screening methods or even by enzyme assay. The situation is complicated further by the lack of a rapid means of assessing UGT1A1 promoter polymorphism to detect those at especially high risk of hyperbilirubinemia. World Health Organization recommendations,27 based on nonmolecular diagnosis, clearly underestimate the risk of neonatal hyperbilirubinemia in this population group. Newborn health caretakers and parents should be warned that female neonates may be G-6-PD-deficient heterozygotes despite a normal reading screening test. Especial vigilance must be used to prevent hyperbilirubinemia with its potential of kernicterus from occurring.
This study was supported in part by grants from the General Research Fund at the Shaare Zedek Medical Center, by the National Institutes of Health Grants RR00833 and HL 25552 at the Scripps Research Institute, and the National Institutes of Health Grants 14426 and RR00070, and by the Mary L Johnson Research Fund, Stanford University Medical Center.
We thank the delivery room nurses of the Shaare Zedek Medical Center for collecting umbilical cord blood samples, and Dr Nina Lonshakova, Mrs Beryl Westwood, and Mr Ronald J Wong, BA, for technical assistance.
- Received July 1, 1998.
- Accepted November 24, 1998.
Reprint requests to (M.K.) Department of Neonatology, Shaare Zedek Medical Center, PO Box 3235, Jerusalem 91031, Israel. E-mail:
Presented in part at the Pediatric AcademicSocieties/Society for Pediatric Research; May 1–5, 1998, New Orleans, LA.
- G-6-PD =
- glucose-6-phosphate dehydrogenase •
- RBC =
- red blood cell •
- COHb =
- carboxyhemoglobin •
- CO =
- carbon monoxide •
- COHbc =
- COHb corrected for inspired CO •
- UGT1A1 =
- uridine 5′-diphosphate glucuronosyltransferase 1 •
- tHb =
- total hemoglobin •
- NS =
- not statistically significant
- Beutler E
- ↵Luzzatto L. G6PD deficiency and hemolytic anemia. In: Nathan DC, Oski FA, eds. Hematology of Infancy and Childhood. Philadelphia, PA: WB Saunders Co; 1993:674–695
- Stevenson DK,
- Vreman HJ
- Kaplan M,
- Renbaum P,
- Levi-Lahad E,
- Hammerman, Lahad A, Beutler E
- Miller SA,
- Dykes DD,
- Polesky HF
- Vulliamy TJ,
- D'Urso M,
- Battistuzzi G,
- et al.
- Vreman HJ,
- Kwong LK,
- Stevenson DK
- Vreman HK,
- Stevenson DK,
- Zwart A
- Vreman HJ,
- Stevenson DK,
- Oh W,
- et al.
- Kaplan M,
- Abramov A
- World Health Organization
- Kattamis CA,
- Kyriazakou M,
- Chaidas S
- Russo G,
- Mollica F,
- Pavone L,
- Schiliro G
- Yu M-W,
- Hsiao K-J,
- Wuu K-D,
- Chen C-J
- Beutler E,
- Gelbart T,
- Demina A
- ↵Beutler E. G-6-PD activity of individual erythrocytes and X-chromosomal inactivation. In: Yunis JJ, ed. Biochemical Methods in Red Cell Genetics. New York, NY: Academic Press, Inc; 1969:95–113
- ↵Kaplan M, Muraca M, Hammerman C, et al. Bilirubin conjugation, reflected by serum conjugated bilirubin fractions, in glucose-6-phosphate dehydrogenase deficient neonates: a determining factor in the pathogenesis of hyperbilirubinemia. Pediatrics. 1998;102(3). URL: http://www.pediatrics.org/cgi/content/full/102/3/e37
- Copyright © 1999 American Academy of Pediatrics