Risk factors for neonatal hyperbilirubinemia: a systematic review and meta-analysis.

Background: Hyperbilirubinemia is the most common cause of neonatal hospitalization and, although it generally has a good prognosis, a significant percentage of neonatal patients maintain a high bilirubin level, which can lead to severe complications, including lifelong disability such as growth retardation, encephalopathy, autism and hearing impairment. The study of risk factors for neonatal hyperbilirubinemia has been controversial. Therefore, we evaluated the risk factors of neonatal hyperbilirubinemia using a meta-analysis.
Methods: Relevant English and Chinese studies that discussed risk factors for neonatal hyperbilirubinemia were retrieved from the PubMed, EMBASE, Medline, Central, China National Knowledge Infrastructure (CNKI), Wanfang and China Science Digital Library (CSDL). The literature took newborns as the research object, set up a control group, and observed the relationship between exposure factors and neonatal hyperbilirubinemia. The combined effect size was expressed by odds ratio (OR) and 95% confidence interval (CI). The Chi-square test was used to test heterogeneity of the studies, and if it existed, subgroup analyses were used to explore the source of heterogeneity, and the random-effects model was selected for the combined analysis. The fixed-effects model was chosen for the combined analysis if there was no heterogeneity. Publication bias was assessed using Egger's test and funnel plot.
Results: Risk factors for neonatal hyperbilirubinemia were exclusive breastfeeding (BF: OR =1.74, 95% CI: 1.42, 2.12, Z=5.43, P<0.00001); glucose-6-phosphate dehydrogenase deficiency (G6PD: OR =1.62, 95% CI: 1.44, 1.81, Z=8.39, P<0.00001); maternal-fetal ABO blood group incompatibility (OR =1.64, 95% CI: 1.42, 1.89, Z=6.75, P<0.00001); and preterm birth (PTB: OR =1.31, 95% CI: 1.17, 1.47, Z=4.60, P<0.00001); there was no heterogeneity or publication bias among the studies (BF: χ2=5.34, P=0.25, I2=25%; G6PD: χ2=4.40, P=0.49, I2=0%; ABO: χ2=1.91, P=0.75, I2=0%; PTB: χ2=0.81, P=0.67, I2=0%). Conclusions: Exclusive breastfeeding, G6PD deficiency, ABO incompatibility and premature birth were confirmed as risk factors for neonatal hyperbilirubinemia. Pregnant women with risk factors should be monitored more closely and clinical intervention should be given in a timely manner. 2022 Translational Pediatrics. All rights reserved.

Introduction

Hyperbilirubinemia is the most common cause of neonatal hospitalization (1), and although most newborns with hyperbilirubinemia have a good prognosis (2), approximately 8–11% maintain a high bilirubin level, which may lead to severe complications, including lifelong disability such as growth retardation, encephalopathy, autism and hearing impairment (3-5). The incidence of pathological jaundice is 1% without risk factors of hyperbilirubinemia, whereas it is 59% with risk factors (3).

The risk factors of neonatal hyperbilirubinemia are diverse, complex and interconnected. For example, the age and gestational age of the pregnant woman will affect or change the feeding mode, and perinatal diseases may affect the delivery method, causing a series of chain reactions. The research on the risk factors of neonatal hyperbilirubinemia has always been controversial; for example, Chen et al. (6) believed that exclusive breastfeeding was a risk factor. Scrafford et al. (7) believed that exclusive breastfeeding may be a protective factor for specific neonatal hyperbilirubinemia. Chen et al. (8) concluded that high breastfeeding frequency can reduce the incidence of hyperbilirubinemia. Huang et al. (9) considered that preterm birth, exclusive breastfeeding, blood group incompatibility, and glucose-6-phosphate dehydrogenase (G6PD) deficiency are risk factors for the disease. Kaplan et al. (10) considered that only G6PD, blood group inappropriate disease risk factors. The reason for these differences in our analysis is the presence of bias in the subjects studied in separate studies. The socioeconomic conditions and cultural background of different regions may have some influence on the results. Therefore, it is of great clinical significance to determine the risk factors of neonatal jaundice to effectively reduce the risk of neonatal hyperbilirubinemia and the incidence of related complications. Most of the studies are retrospective analysis, the sample size is small, and the evidence level is low, which cannot support the conclusion. In the present study we performed a necessary meta-analysis to clarify the risk factors of neonatal hyperbilirubinemia and provide clues for clinical early intervention and prevention. We present the following article in accordance with the MOOSE reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-22-229/rc).

Methods

Literature retrieval

A literature search was conducted in the PubMed, EMBASE, Medline, Central, China National Knowledge Infrastructure (CNKI), Wanfang and China Science Digital Library (CSDL) databases, using the keywords “neonatal hyperbilirubinemia” or “jaundice” and “risk factors” and selecting English and Chinese studies. The retrieval date was April 1, 2022.

Literature screening

Inclusion criteria: (I) newborns; (II) subjects allocated to experimental and control groups; (III) investigation of exposure factors of newborns or mothers, including at least one of exclusive breastfeeding, G6PD deficiency, maternal and infant ABO incompatibility, and preterm birth; (IV) outcome was neonatal hyperbilirubinemia; (V) observational studies: cohort, case-control or cross-sectional study; (VI) results included the odds ratio (OR) and 95% confidence interval (CI) of exposure factors or could be calculated from the data.

Exclusion criteria: (I) repeat reports; (II) adult subjects; (III) no control group; (IV) incomplete data and unable to be supplemented through contact with the author(s).

Data extraction

Two researchers jointly extracted the data required for the analysis, including the author(s), title, date of publication, research type, number of researchers, number of cases of neonatal jaundice, number of exclusively breastfed newborns, number of ABO-incompatible newborns, of premature babies. An attempt was made to contact the author(s) for missing data. Differences in opinions were resolved by consensus of the two researchers.

Literature quality evaluation

The Newcastle-Ottawa scale (NOS) was used to evaluate literature quality, comprising the selectivity and comparability of research methods, exposure factors, and outcomes. NOS score ≥6 was classified as low risk of bias, otherwise high risk of bias. The risk of bias in cross-sectional studies was assessed using the Joanna Briggs Institute (JBI) cross-sectional study volume risk of bias assessment criteria, with a total score of 20 points. A JBI score >14 was considered low risk of bias. Differences in opinions were resolved by consensus of the two researchers.

Statistical analysis

We used Cochrane RevMan5.3 software for statistical analysis of the data. Adjusting for confounding factors, OR values were calculated by multivariate analysis, or obtained directly from the literature. The OR value and 95% CI were used to describe the effect quantity. The Chi-square test was used and when the heterogeneity statistic I2 corrected by degrees of freedom was >50% or P<0.1, it was considered there was heterogeneity. Subgroup analyses was used to explore the causes of heterogeneity. When heterogeneity could not be eliminated, only reviews were performed without pooling results. When the I2 corrected by degrees of freedom was ≤50% and P≥0.1, it was considered that there was no heterogeneity, and the fixed-effects model was used. Publication bias was assessed using funnel plots and Egger’s test. In this study, OR values were calculated or obtained from cohort studies, case-control studies and cross-sectional studies, and in the absence of heterogeneity, pooled analyses could be performed. Two-way P<0.05 indicated statistically significant.

Results

Characteristics of the included studies

A total of 1,876 studies were retrieved from the databases and a total of 14 articles were screened according to the inclusion criteria (6-19). The flow chart of literature screening is shown in . Among the 14 studies, 5 were cohort studies, 4 were case-control studies, and 5 were cross-sectional studies. There were 2 studies from the Chinese literature, and 12 from the English literature. The baseline information of the studies and NOS/JBI scores are shown in .

Figure 1

Flow chart of literature screening.

Table 1

Characteristics of the included studies and NOS/JBI scores

Author	Year	Language	Study type	n	NOS/JBI
Chen et al. (6)	2011	English	Cohort	313	7
Scrafford et al. (7)	2013	English	Cross-sectional	18,985	14
Chen et al. (8)	2015	English	Cohort	98	7
Huang et al. (9)	2009	English	Cohort	1,034	8
Kaplan et al. (10)	1998	English	Case-control	98	8
Ketsuwan et al. (11)	2017	English	Case-control	176	6
Mojtahedi et al. (12)	2018	English	Cross-sectional	200	12
Najib et al. (13)	2013	English	Cohort	1,134	6
Owa (14)	1989	English	Case-control	234	8
Thielemans et al. (15)	2021	English	Cross-sectional	1,710	12
Thielemans et al. (16)	2018	English	Cross-sectional	2,980	12
Yu et al. (17)	1992	English	Cohort	12,379	6
Xiong et al. (18)	2019	Chinese	Case-control	200	5
Dai & Cheng (19)	2011	Chinese	Cross-sectional	206	10

NOS, Newcastle-Ottawa scale; JBI, Joanna Briggs Institute.

Table 2

NOS scores of included literatures

Study	Selection				Comparability control for important factor	Exposure			NOS
	Adequate definition of case	Representativeness of the case	Selection of controls	Definition of controls		Ascertainment of exposure	Same method of ascertain for cases and controls	Non-response rate
Chen (6)	*	*	*	*	*	*	*	–	7
Chen (8)	*	*	*	*	*	–	*	*	7
Huang (9)	*	*	*	–	**	*	*	*	8
Kaplan (10)	*	*	–	*	**	*	*	*	8
Ketsuwan (11)	*	*	*	–	*	*	–	*	6
Najib (13)	*	*	*	–	–	*	*	*	6
Owa (14)	*	*	*	–	**	*	*	*	8
Yu (17)	*	*	*	*	*	–	*	–	6
Xiong (18)	*	*	*	*	–	*	–	–	5

NOS, Newcastle-Ottawa scale.

Identified risk factors for neonatal hyperbilirubinemia

Exclusive breastfeeding

A total of 5 studies of the association between breastfeeding and neonatal hyperbilirubinemia were included in our meta-analysis. There was no heterogeneity among them (χ2=5.34, P=0.25, I2=25%), so the fixed-effects model was used. Exclusive breastfeeding was identified as a risk factor (OR =1.74, 95% CI: 1.42, 2.12, Z=5.43, P<0.00001). The Egger test and funnel diagram in shows the scatter points distributed within the CI, roughly symmetrically, and there was no publication bias (P>0.05) ().

Figure 2

Forest diagram of the correlation between exclusive breastfeeding and neonatal hyperbilirubinemia. SE, standard error; CI, confidence interval.

Figure 3

Funnel plot of the correlation between exclusive breastfeeding and neonatal hyperbilirubinemia. OR, odds ratio; SE, standard error.

Glucose-6-phosphate dehydrogenase (G6PD) deficiency

A total of 6 studies of the association between G6PD deficiency and neonatal hyperbilirubinemia were included in our meta-analysis. There was no heterogeneity among them (χ2=4.40, P=0.49, I2=0%), so the fixed-effects model was used for consolidation. G6PD deficiency was identified as a risk factor (OR =1.62, 95% CI: 1.44, 1.81, Z=8.39, P<0.00001). The Egger test and funnel diagram in shows the scatter points distributed within the CI, roughly symmetrically, and there was no publication bias (P>0.05) ().

Figure 4

Forest plot of the correlation between G6PD deficiency and neonatal hyperbilirubinemia. SE, standard error; CI, confidence interval; G6PD, glucose-6-phosphate dehydrogenase.

Figure 5

Funnel plot of the correlation between G6PD deficiency and neonatal hyperbilirubinemia. OR, odds ratio; SE, standard error; G6PD, glucose-6-phosphate dehydrogenase.

Maternal-fetal ABO blood group incompatibility

A total of 5 studies of the correlation between ABO blood group incompatibility and neonatal hyperbilirubinemia were included in our meta-analysis. There was no heterogeneity among them (χ2=1.91, P=0.75, I2=0%), so the fixed-effects model is used for consolidation. Maternal-fetal ABO blood group incompatibility was identified as a risk factor (OR =1.64, 95% CI: 1.42, 1.89, Z=6.75, P<0.00001). The Egger test and funnel diagram in shows the scatter points distributed within the CI, roughly symmetrically, and there was no publication bias (P>0.05) ().

Figure 6

Forest plot of the correlation between maternal-fetal ABO blood group incompatibility and neonatal hyperbilirubinemia. SE, standard error; CI, confidence interval.

Figure 7

Funnel plot of the correlation between maternal-fetal ABO blood group incompatibility and neonatal hyperbilirubinemia. OR, odds ratio; SE, standard error.

Premature delivery

A total of 3 studies of the association between preterm birth and neonatal hyperbilirubinemia were included in our meta-analysis. There was no heterogeneity among them (χ2=0.81, P=0.67, I2=0%), so the fixed-effects model was used for consolidation. Preterm birth was identified as a risk factor (OR =1.31, 95% CI: 1.17, 1.47, Z=4.60, P<0.00001). The Egger test and funnel diagram in shows the scatter points distributed within the CI, roughly symmetrically, and there was no publication bias (P>0.05) ().

Figure 8

Forest plot of the correlation between preterm birth and neonatal hyperbilirubinemia. SE, standard error; CI, confidence interval.

Figure 9

Funnel plot of the correlation between preterm birth and neonatal hyperbilirubinemia. OR, odds ratio; SE, standard error.

Discussion

The results of our meta-analysis revealed exclusive breastfeeding as a risk factor for neonatal hyperbilirubinemia. Infants with breast milk hyperbilirubinemia have a good prognosis (20), but 58–81.4% of infants with severe hyperbilirubinemia are exclusively or mainly breastfed (21). With the popularity of breastfeeding education, the breastfeeding rate has increased, and the incidence rate of breast milk jaundice has increased year by year (22,23). Breast milk jaundice may be related to the presence of glucuronosyltransferase inhibitors in the mother’s colostrum and the lack of bilirubin reuptake inhibitors in the infant’s small intestine. A partial reason for increased unconjugated bilirubin in breast milk jaundice is that breast milk contains more highly active glucuronic acid, which increases intestinal and liver circulation. The amount of breastfeeding after birth may be related to the severity of jaundice. It should be emphasized that the relationship between breastfeeding and neonatal hyperbilirubinemia is complex. First of all, whether to choose breastfeeding and the method of breastfeeding (including exclusive breastfeeding and mixed feeding) are related to maternal age and mode of delivery. These factors combined influence the occurrence of neonatal hyperbilirubinemia. Secondly, a study showed that exclusive breastfeeding is a risk factor for neonatal hyperbilirubinemia in infants with nutritional problems (7). On the other hand, for infants without nutritional problems, exclusive breastfeeding is protective for neonatal hyperbilirubinemia (7). Another study emphasized that insufficient breastfeeding, rather than breastfeeding per se, is the risk factor for neonatal hyperbilirubinemia (24). Finally, others have reported that increasing breastfeeding frequency can reduce the risk of hyperbilirubinemia (25). The incidence of hyperbilirubinemia was significantly lower in infants who breastfed ≥8 times/day than in infants who breastfed less than that frequency (8). The American Academy of Pediatrics recommends breastfeeding between 8 and 12 times/day in the first few weeks of life (3). The time of first feeding is also associated with neonatal hyperbilirubinemia (11).

The results of our meta-analysis showed that premature delivery was a risk factor for neonatal hyperbilirubinemia, which was consistent with the results of a previous meta-analysis suggesting that premature delivery is a maternal factor (3) affecting the incidence rate of neonatal jaundice. Neonatal hyperbilirubinemia caused by premature birth accounts for 30% of children (3) and may be related to the maturity of uridine diphosphate glucuronosyltransferase. Gestational age is positively correlated with the activity of glucuronosyltransferase diphosphate. Premature delivery may lead to delayed first feeding, insufficient breast milk, infant nutritional disorders, reduced feeding frequency and other problems, which will lead to increased concentration of bilirubin in the newborn’s blood.

Maternal-fetal ABO blood group incompatibility and G6PD deficiency were also risk factors for neonatal hyperbilirubinemia, because they both cause hemolysis, a conclusion supported by a previous study (16). ABO blood group incompatibility, G6PD enzyme deficiency, premature delivery, scalp hematoma and Rh blood group incompatibility are the most common risk factors for early neonatal jaundice (16). A previous study showed that the cases of neonatal hyperbilirubinemia caused by ABO blood group incompatibility accounted for 13.3 (3). In two Chinese studies, maternal–fetal blood group incompatibility was considered an independent risk factor for neonatal jaundice (18,19).

There are some limitations to our research. Few of the studies included in the analysis originated in less developed countries, which may affect the results. Secondly, there were relatively few risk factors included in the analysis, which therefore cannot be considered as comprehensively showing neonatal hyperbilirubinemia’s risk factors.

There are some limitations to this study. The first is that the included literature is small, the sample size is small, and the coverage is narrow. Second, we analyzed cross-sectional studies, cohort studies, and case-control studies together due to the lack of literature. Low-quality evidence from cross-sectional studies. Finally, exposure variables are few and cannot fully describe risk factors for disease. Multicenter high-quality studies are still needed to confirm our conclusions.

Exclusive breastfeeding, G6PD deficiency, maternal-fetal ABO blood group incompatibility and premature delivery were confirmed as risk factors for neonatal hyperbilirubinemia. Pregnant women with risk factors should be monitored more closely and clinical intervention should be given in a timely manner.

The article’s supplementary files as