Introduction

Unconjugated hyperbilirubinemia in neonates is a common physiological phenomenon. Bilirubin production is increased in the neonate because of the larger erythrocyte volume, the shortened erythrocyte lifespan, the degradation of heme and heme precursors from fetal extramedullary hematopoietic tissue, and possibly, the increased turnover of cytochromes.1 The ability to conjugate bilirubin is also extremely low in neonates at term, ~1% of adult values.2 Neonatal hyperbilirubinemia is probably associated with other factors such as immature hepatic bilirubin uptake and intracellular bilirubin transport and the increased enterohepatic circulation of bilirubin. The imbalance between the production and elimination of bilirubin leads to varying levels of hyperbilirubinemia in every neonate. The propensity toward neonatal hyperbilirubinemia in East Asians is well known.1 The association of UGT1A1 (TA)7 with severe jaundice in newborns with glucose-6-phosphate dehydrogenase deficiency has been reported.3 In 1998, we found that the 211G >A genotype of the UGT1A1 gene is very common among Japanese, Koreans and Chinese, and this mutation is associated with neonatal hyperbilirubinemia.4, 5

Recently, the association of the UGT1A1 211G>A genotype with neonatal hyperbilirubinemia was also confirmed in Asians6 and was especially significant in breast-fed neonates.7, 8 Breastfeeding jaundice, generally known as breast non-feeding jaundice, has been considered to result from inadequate calorie intake.9 In 2013, we found that maximal body weight loss during the neonatal period is an independent risk factor for the development of neonatal hyperbilirubinemia, and the UGT1A1 211G>A genotype is a significant risk factor only in neonates with 5.0% or greater maximal body weight loss.10

However, these data do not explain all instances of severe hyperbilirubinemia in Japanese neonates. In 2004, a significant association of polymorphisms of organic anion-transporting polypeptides (OATPs, genes: solute-carrier organic anion transporters (SLCOs)) with hyperbilirubinemia in breast-fed neonates was reported.11 OATPs mainly transport conjugated bilirubin, but the serum level of unconjugated bilirubin is twofold higher in OATP1a/1b knockout mice than in their wild-type counterparts, suggesting that OATPs may contribute to unconjugated bilirubin uptake by hepatocytes.12, 13 However, some reports did not find an association between SLCOs polymorphisms and neonatal hyperbilirubinemia.8, 14 These conflicting reports may be owing to studies of subjects in different nutritional conditions, with adequate or inadequate feeding. To clarify the relationship, specific analysis including the genetic and nutritional condition was required.

In the present paper, we studied the association of unconjugated hyperbilirubinemia with genetic (UGT1A1 and SLCOs polymorphisms) and epigenetic factors (maximal body weight loss) in breast-fed Japanese neonates.

Materials and methods

Subjects

The subjects were recruited from among the Japanese full-term and breast-fed neonates born at Yamagata Prefectural Central Hospital. We excluded the neonates who had factors that would affect the level of serum bilirubin, such as hemolytic anemia, neonatal asphyxia, maternal diabetes, congenital heart or intestine malformation, infections, and parenteral fluid therapies. We studied a total of 401 neonates who were exclusively breast-fed without formula supplementation before developing hyperbilirubinemia. The main characteristics of the 401 neonates were as follows: male-to-female ratio, 221:180; gestational age, 39.5±1.2 (mean±s.d.) weeks; and birth weight, 3080±325 (mean±s.d.) gr. Neonates with severe hyperbilirubinemia were treated with phototherapy according to the criteria as described in the methods. Medical records including gestational age, birth weight, body weight loss [(birth body weight – daily body weight)/birth body weight × 100%], total serum bilirubin levels, and peak bilirubin levels before phototherapy were reviewed. The Ethics Committee of the Yamagata University School of Medicine approved this study. After informed consent was obtained from the parents, genomic DNA was isolated from spare dried blood spots from neonatal screening cards as described previously.15

The 401 neonates were classified into group A (neonates showing <10% body weight loss during the neonatal period) and group B (neonates showing 10% or greater body weight loss during the neonatal period) (Table 1).

Table 1 Demographic characteristics of the 401 infants enrolled in this studya

Assessment of serum bilirubin level, criteria for phototherapy and definition of hyperbilirubinemia

The serum bilirubin of the neonates was assessed with a Jaundice Meter (model 103; Minolta, Osaka, Japan) once a day, at the same time, during their hospitalization or at least during the first week of life. When the reading of the transcutaneous bilirubinometer reached the criterion for further evaluation, the total serum bilirubin concentration was measured. Phototherapy was initiated if the measured bilirubin level exceeded the criterion as follows: 10.0 mg dl−1 at day 1; 14.0 mg dl−1 at day 2; 16.0 mg dl−1 at day 3; 17.0 mg dl−1 at day 4; 18.0 mg dl−1 at day 5; and 20.0 mg dl−1 at day 6. Hyperbilirubinemia was defined as a peak serum bilirubin level over the criterion of phototherapy.

UGT1A1 and SLCOs genes analysis

The (TA)7 genotype in the UGT1A1 promoter was analyzed by fragment analysis,16 and 211G>A genotype in exon 1 of the UGT1A1 and polymorphisms of the SLCOs were analyzed by TaqMan Drug Metabolism Genotyping Assay (PE Applied Biosystems, Foster City, CA, USA). The polymorphisms were UGT1A1 211G>A: G71R (rs 4148323): c_559715_20, SLCO1B1 388A>G: N130D (rs 2306283): c_1901697_20, SLCO1B1 521T>C: V174A (rs 4149056): c_30633906_10, SLCO1B3 699G>A: M233I (rs 7311358): c_25765587_40 and SLCO2B1 1457C>T: S486F (rs 2306168): c_16193013_20.

Statistical analysis

The R statistical package 2.11.1 (R Core Team, Vienna, Austria) was used for statistical analysis. Differences in the distribution of several categorical and continuous variables between the A and B groups were analyzed using the χ2 test and Mann–Whitney U test where appropriate (Tables 1 and 2). The tests were also used to assess whether the peak serum bilirubin level and incidence of hyperbilirubinemia increased with the numbers of non-wild-type alleles in the genotypes, respectively (Tables 3A and 3B). Multivariate-adjusted odds ratios (ORs) and 95% confidence intervals of hyperbilirubinemia were estimated for genetic polymorphisms of the UGT1A1 and SLCOs genes and various risk factors selected by univariate analysis using multiple logistic regression analysis (Table 4). Table 5 shows ORs for neonatal hyperbilirubinemia in group A versus group B with different UGT1A1 and SLCOs genotypes by multiple logistic regression analysis.

Table 2 UGT1A1 and SLCO genetic polymorphisms in A and B groups
Table 3A Correlation of the peak serum bilirubin level with different UGT1A1 and SLCO genotypes in group A versus group Ba
Table 3B Correlation of the incidence of hyperbilirubinemia with different UGT1A1 and SLCO genotypes in group A versus group Ba
Table 4 Multivariate regression analysis of neonatal hyperbilirubinemia
Table 5 Odds ratio for neonatal hyperbilirubinemia with different UGT1A1 and SLCO genotypes in group A versus group B

All tests were two sided. A P-value <0.05 was considered significant.

Results

A total of 401 neonates were enrolled in the study and classified into two groups. Groups A and B included neonates showing a less than 10% body weight loss or a 10% or greater body weight loss during the neonatal period, respectively. As shown in Table 1, 56 neonates (14%) showed hyperbilirubinemia and required phototherapy. The neonates in group B had a significantly higher peak bilirubin level and incidence of hyperbilirubinemia, higher frequency of cesarean delivery and shorter gestational period. Sex and body weight at birth were not significantly different between the two groups.

We compared the gene frequencies of UGT1A1 and SLCOs polymorphisms between two groups and found significant differences in the gene frequencies of the UGT1A1 211G>A genotype and SLCO1B1 521T>C genotype (Table 2). The gene frequencies of the UGT1A1 211G>A genotype in groups A and B were 0.19 and 0.12, respectively, and the gene frequencies of SLCO1B1 521T>C genotype of groups A and B were 0.13 and 0.21, respectively.

Then we studied the correlation of the peak serum bilirubin level in neonates in two groups with different UGT1A1 and SLCOs genotypes. The peak bilirubin levels were significantly higher in group B neonates heterozygous for the UGT1A1 211G>A genotype, homozygous for the UGT1A1 (TA)6 genotype (wild type), hetrozygous for the SLCO1B3 699G genotype (wild type) and homozygous for the SLCO2B1 1457C>T genotype (Table 3A). We could not analyze the neonates homozygous for the UGT1A1 211G>A genotype because of the small size of samples.

We also studied the correlation of the incidence of hyperbilirubinemia in neonates in two groups with different UGT1A1 and SLCO genotypes. The incidence of hyperbilirubinemia was significantly higher in group B neonates with heterozygous UGT1A1 211G>A genotype, homozygous UGT1A1 (TA)6 genotype (wild type) and homozygous SLCO1B3 699G genotype (wild type) (Table 3B). We could not detect a significant increase in hyperbilirubinemia in group B neonates homozygous for the SLCO2B1 1457C>T genotype, who showed higher peak serum bilirubin levels compared with those in group A neonates.

Multivariate regression analysis for neonatal hyperbilirubinemia revealed that only maximal body weight loss was an independent risk factor for the development of neonatal hyperbilirubinemia with adjusted OR (95% confidence interval) of 1.25 (1.11–1.41) (Table 4). Cesarean delivery rather significantly decreased the risk of neonatal hyperbilirubinemia.

In Table 5, we showed the OR for neonatal hyperbilirubinemia in two groups with different UGT1A1 and SLCOs genotypes. We could not detect any significant effect of gene polymorphisms on neonatal hyperbilirubinemia in group A. In group B, we detected a significant increase in the OR for hyperbilirubinemia in neonates heterozygous for the UGT1A1 211G>A, UGT1A1(TA)7, SLCO1B1 388A>G and SLCO1B1 521T>C genotypes and heterozygous and homozygous for the SLCO1B3 699G>A genotype. We did not find a significant increase in the OR for hyperbilirubinemia in neonates homozygous for those genotypes except for the SLCO1B3 699G>A genotype. We also did not find a significant increase in the OR for hyperbilirubinemia in neonates carrying SLCO2B1 1457C>T.

Discussion

Our study showed that maximal body weight loss is the only independent risk factor for the development of neonatal hyperbilirubinemia, and UGT1A1, SLCO1B1 and SLCO1B3 polymorphisms become risk factors in neonates showing 10% or greater body weight loss during the neonatal period. Inadequate feeding may increase the bilirubin burden and cause apparent hyperbilirubinemia in neonates who have polymorphisms in the genes involved in the transport and or metabolism of bilirubin.

Neonatal hyperbilirubinemia is a physiological phenomenon, but its severity is also affected by genetic factors. We previously reported that the UGT1A1 211G>A genotype was prevalent in Japanese, Koreans, and Chinese and was associated with neonatal hyperbilirubinemia.4, 5 Yamamoto et al.17 showed that homozygous and heterozygous expression of the UGT1A1 211G>A genotypes showed 32.2±1.6 and 60.2±2.5% activity of the level of the normal UGT1A1 gene, respectively. However, all Japanese neonates with severe hyperbilirubinemia did not carry the UGT1A1 211G>A genotype. To identify other predisposing factors for neonatal hyperbilirubinemia, we studied the effect of polymorphisms in the enhancer sequence (–3483/–3194) (phenobarbital response enhancer module) of the UGT1A1 gene, polymorphic (GT)n repeats in the promoter region of the hemeoxygenase-1 (a rate-limiting enzyme in heme metabolism) gene and the composition of fetal hemoglobin (a probable major source of bilirubin in neonates).18, 19 However, we could not find any predisposing factors. There has been an increase in reports that UGT1A1 211G>A genotype is associated with unconjugated hyperbilirubinemia in breast-fed neonates. However, we noticed that maximal body weight loss during the neonatal period is the only independent risk factor for the development of neonatal hyperbilirubinemia in breast-fed infants, and the UGT1A1 211G>A genotype becomes a risk factor for neonatal hyperbilirubinemia only in infants with inadequate feeding. UGT1A1 (TA)7 was reported to be associated with neonatal hyperbilirubinemia in infants with glucose-6-phosphate dehydrogenase deficiency.3 The promoter with (TA)7 genotype has 30% less transcriptional activity than that with the (TA)6 genotype (wild type).20, 21 However, UGT1A1 (TA)7 is not a prevalent polymorphism in Japanese and Asian people and could not be identified as a predisposing factor for neonatal hyperbilirubinemia.4, 5, 7, 8

There are also several reports that OATPs polymorphisms are associated with neonatal hyperbilirubinemia, but some reports denied the association.11, 15, 22, 23 OATPs, also called solute-carrier organic anion transporters (SLCOs), are encoded by SLCOs genes. OATPs are uptake transporters for a broad range of endogenous compounds and xenobiotics.24 The liver-specific expression of OATP1B1 and OATP1B3 contributes to the hepatic uptake of drugs from portal vein, and OATP2B1 is expressed in the small intestine, as well as other tissues, such as the liver, lungs and ovaries.25 Steeg et al.26 reported that complete OATP1B1 and OATP1B3 deficiency cause human Rotor syndrome. OATPs mainly transport conjugated bilirubin, but the serum level of unconjugated bilirubin is twofold higher in OATP1a/1b knockout mice than in their wild-type counterparts, suggesting that OATPs may contribute to unconjugated bilirubin uptake by hepatocytes.12, 13

As shown in Table 5, we found a significant increase in the OR for hyperbilirubinemia in the group B neonates, who were heterozygous for the UGT1A1 211G>A, UGT1A1 (TA) 7, SLCO1B1 388A>G and SLCO1B1 521T>C genotypes and heterozygous and homozygous for SLCO1B3 699G>A genotype. It is interesting to note that UGT1A1 conjugates bilirubin and SLCO1B1 and SLCO1B3 may be involved in the transport of unconjugated bilirubin in hepatocytes. The results indicated that inadequate feeding is the only independent risk factor for neonatal hyperbilirubinemia and polymorphisms of genes associated with bilirubin metabolism or transport are additional modulating risk factors for neonatal hyperbilirubinemia.

We could not confirm statistically significant difference in several points including the gene dose effect of the polymorphisms on hyperbilirubinemia probably because of the small size of samples. The neonates homozygous for the SLCO2B1 1457C>T genotype in group B showed a significant increase in peak serum bilirubin level and a high incidence of hyperbilirubinemia (Tables 3A and 3B). However, multivariate regression analysis did not reveal a significant association with neonatal hyperbilirubinemia (Tables 4 and 5). We also could not find a significant increase in peak serum bilirubin level and incidence of hyperbilirubinemia in the neonates heterozygous for the SLCO2B1 1457C>T genotype (Tables 3A and 3B). Although there appears to be a slight increase in peak serum bilirubin level in group B neonates homozygous for the SLCO2B1 1457C>T genotype, this did not affect our findings. The gene frequency of SLCO1B1 521T>C genotype was higher in group B (Table 2) and could be partly associated with a higher peak bilirubin level and incidence of hyperbilirubinemia in group B (Tables 3A and 3B). However, the full model multivariate regression analysis of group B neonates (data not shown) revealed that the UGT1A1 211G>A genotype is only significantly associated with hyperbilirubinemia. These data indicated that a higher gene frequency of SLCO1B1 521T>C genotype in group B did not affect our study findings and suggested that the UGT1A1 211G>A genotype is the strongest risk factor for hyperbilirubinemia among those genotypes.

As previously reported, cesarean delivery may cause delayed onset of lactation, leading to an increase in maximal body weight loss.27 However, multivariate regression analysis revealed that cesarean delivery instead decreased the risk of neonatal hyperbilirubinemia. Chang et al.8 reported that vaginal delivered breast-fed neonates have a high risk of developing hyperbilirubinemia. They speculated that vaginal delivery might be associated with oxytocin usage, vacuum extraction and cephalohematoma. Oxytocin exposure is a risk factor for hyperbilirubinemia, which may have a direct effect on neonatal bilirubin metabolism.28, 29

Inadequate feeding of breast-fed neonates may increase intestinal bilirubin absorption, impair hepatic conjugation and/or excretion of bilirubin because of energy deficiency, and cause hyperbilirubinemia in neonates carrying UGT1A1 and SLCOs polymorphisms. Our data suggest that adequate feeding could overcome genetic predisposition (UGT1A1 and SLCOs) for neonatal hyperbilirubinemia even in breast-fed infants.