Serum Albumin Assessment in Neonatal Jaundice: Impact on Phototherapy Decision-Making

Neonatal jaundice is a common physiological condition in the newborn period. The imbalance between bilirubin synthesis and elimination causes bilirubin to accumulate in the newborn.¹ The most concerning consequence of hyperbilirubinemia is neurological dysfunction, bilirubin encephalopathy, or kernicterus, which may cause disability or death.² To prevent complications, it is crucial to regularly monitor newborns for clinical jaundice, promptly assess the total serum bilirubin (TSB) level, and offer treatment with phototherapy or exchange transfusion (ET) when clinically indicated.^3,4 The American Academy of Pediatrics (AAP) has established standard guidelines for managing hyperbilirubinemia in newborn infants of 35 weeks’ gestation or more. These guidelines, revised in 2022, include specific treatment thresholds based on the infant’s age and gestational age and the presence of risk factors for hyperbilirubinemia neurotoxicity. The treatment threshold is lower for infants with a lower gestational age, those with a younger postnatal age, and those with neurotoxicity risk factors.⁴ 

Albumin, a water-soluble protein, plays a significant role in neonatal jaundice by binding with bilirubin in the blood, reducing free bilirubin levels that can cross the blood-brain barrier and potentially injure the brain.⁵ The AAP guidelines define low serum albumin levels (ie, <3.0 g/dL) as one of the risk factors for bilirubin neurotoxicity, along with isoimmune hemolytic disease; glucose-6-phosphate dehydrogenase (G6PD) deficiency; or other hemolytic conditions, sepsis, and significant clinical instability in the previous 24 hours. However, the existing literature on serum albumin in newborns is limited, with most research in the 1980s and 1990s focusing primarily on preterm or neonatal intensive care unit (NICU) infants.^6–11 Currently, there is insufficient information regarding the prevalence of low serum albumin levels in newborns with jaundice, the potential impact of serum albumin testing on phototherapy treatment decisions, and the cost-effectiveness of routine testing. For these reasons, the guidelines do not recommend routine assessment of serum albumin in newborn infants and instead suggest measuring serum albumin only when the TSB level exceeds the “escalation-of-care” zone (ie, 2 mg/dL less than the exchange transfusion threshold).⁴ Therefore, in routine medical practice, if serum albumin levels are not assessed, physicians may assume that albumin levels are within the normal range (ie, ≥3.0 g/dL). This could lead to some infants not receiving the necessary phototherapy because they are assigned higher phototherapy thresholds than they should have been.

To address this ambiguity, our neonatology group at the Faculty of Medicine Ramathibodi Hospital, Mahidol University, formulated an institution-wide policy for evaluating neonatal jaundice within our affiliated hospitals in late 2022. Although predominantly following the AAP guidelines, this policy includes serum albumin levels as part of laboratory evaluation, with the goal to accurately determine the bilirubin neurotoxicity risks and precisely assign individualized phototherapy treatment thresholds for each infant in accordance with the AAP recommendations.

Using retrospective patient data and laboratory test results in our centers, this study aims to investigate the influence of serum albumin assessment on early neonatal jaundice treatment decisions.

A retrospective chart review was conducted on infants of 35 weeks’ gestation or more who were evaluated for early neonatal jaundice across 3 hospitals affiliated with the Faculty of Medicine Ramathibodi Hospital, Mahidol University, in Thailand from January 1 to December 31, 2023. The three hospitals included Ramathibodi Hospital, Phayathai, Bangkok (RamaPYT); Somdech Phra Deparatana Medical Center, Phayathai, Bangkok (SDMC); and Chakri Naruebodindra Medical Institute, Bang Phli, Samut Prakan (CNMI). The same neonatologist team operated all 3 hospital neonatal services, which included a newborn nursery and a level 3 NICU at each location, using similar treatment protocols and guidelines. During this period, serum albumin was included in the jaundice assessment as per the guidelines of all units.

Infants younger than age 1 month who had serum albumin and TSB levels were identified from each hospital’s laboratory database, and their electronic medical records were reviewed. Infants were included in the study if they had serum albumin and TSB tests done within 2 weeks after birth as part of early neonatal jaundice evaluation (as indicated by physician documentation or in conjunction with other jaundice-related laboratory tests). The jaundice assessment protocol involved daily screening of all newborns in the hospital during the first week of life, as well as screening of all other infants with clinical jaundice for transcutaneous bilirubin levels (TcB) using a transcutaneous bilirubinometer (Dräger Jaundice Meter JM-105, Draeger Medical Systems, Inc., Telford, Pennsylvania). If the TcB level exceeds or is within 3.0 mg/dL of the phototherapy treatment threshold or if the TcB level is at least 15 mg/dL, a TSB will be performed along with other laboratory evaluations to assess the neurotoxicity risks, determine the need for phototherapy or ET based on AAP 2022 guidelines, and identify the cause of jaundice. Standard laboratory evaluations for infants with jaundice include complete blood count, peripheral blood smear, reticulocyte count, TSB, albumin, G6PD screening using the fluorescent spot test, blood group (ABO, Rh), and direct antibody test. Preterm infants less than 35 weeks’ gestation, infants without laboratory evaluation for jaundice, infants whose laboratory evaluations were performed after age 2 weeks, and infants whose serum albumin and TSB results were obtained for reasons other than early neonatal jaundice evaluations were excluded from the study. If the infant did not receive phototherapy, the serum albumin performed simultaneously or closest to the first TSB to evaluate for jaundice will be used as a reference. Otherwise, serum albumin measured simultaneously or closest to the TSB that qualified the infant for phototherapy will be used.

This study was approved by the Human Research Ethics Committee of our institution. Waiver of individual patient informed consent was granted.

Data collection included infants’ demographics, laboratory results, clinical and neurotoxicity risk factors, cause of jaundice, hour-specific phototherapy threshold based on gestational age and presence of a known hyperbilirubinemia neurotoxicity risk factor according to AAP 2022 guidelines, and the treatment received. The rates of hyperbilirubinemia treatment using phototherapy or ET (initiation and cessation) based on known versus presumed normal albumin levels were evaluated and compared. Only infants who received treatment according to the TSB cut point in the guideline were included in final analysis. The specific terms used in this study were defined as follows: “other hemolytic conditions” included red blood cell membrane defects, enzymatic defects, or hemoglobinopathy, which can lead to the breakdown of red blood cells but does not include hemolysis from bleeding in tissue or organs such as cephalhematoma or subgaleal hemorrhage; “sepsis” refers to infants who exhibited clinical instability suspected or proven to be caused by infection and/or received treatment with intravenous antibiotics; and “significant clinical instability” encompasses infants who required respiratory support such as continuous positive airway pressure, heated humidified high-flow nasal cannula, or mechanically ventilated, required inotropic drugs to maintain blood pressure, altered consciousness, or had a seizure.

In all centers, the albumin BCP2 assay was used to quantify albumin in serum or plasma on the Alinity c system (Abbott Laboratories) using the colorimetric method (bromocresol purple). Internal quality control runs were conducted per manufacturer guidelines, and external quality assessment was carried out per International Organization for Standardization (ISO) 17043:2010 accredited programs.

Descriptive statistical analyses were conducted. Parametric continuous variables are presented as means ± standard deviations, whereas nonparametric continuous variables are presented as medians (interquartile range). Categorical variables are presented as numbers (percentage). All statistical analyses were performed using Stata version 18.0 (StataCorp, LLC).

During the study period, approximately 3500 neonates received neonatal care in our hospital system. A total of 1233 newborns had their serum albumin and total bilirubin levels measured during the first month of life, as recorded in the hospital’s laboratory database from 3 centers, ie, RamaPYT (519), SDMC (374), and CNMI (340). After applying exclusion criteria (ie, 241 infants with gestational age <35 weeks, 11 infants with laboratory tests performed at age >14 days, and 46 infants’ laboratory tests not intended for jaundice evaluation), 935 infants remained for analysis. Among these, 728 infants had these tests performed during the birth hospitalization and 207 after discharge.

The infants’ baseline characteristics are outlined in Table 1. Of these 935 infants, 26.7% had serum albumin less than 3.0 g/dL, 18.3% had hemolytic disease, 9.9% had sepsis, and 9.8% had significant clinical instability in the previous 24 hours, all as bilirubin neurotoxicity risk factors (Table 1). In the study, 402 (43%) infants had at least 1 risk factor, whereas 533 (57%) infants had no risk factors (Figure 1).

Phototherapy was initiated in 401 infants (42.9%). Among these, 110 infants (27.4%) underwent phototherapy when their TSB levels exceeded 2 mg/dL less than the phototherapy threshold, 170 infants (42.4%) when the TSB levels were less than or equal to 2 mg/dL less than the phototherapy threshold, and 121 infants (30.2%) when the TSB levels equaled or exceeded the phototherapy threshold. All infants with TSB levels greater than the phototherapy threshold received phototherapy according to the AAP 2022 guidelines.

Of the 121 infants who received phototherapy with TSB levels meeting the guideline-designated threshold, 49 infants (40.5%) had serum albumin levels less than 3.0 g/dL, which is considered a neurotoxicity risk factor. Among these, 29 infants had additional neurotoxicity risk factors; therefore, the serum albumin results would not have influenced the phototherapy decision. This leaves 20 infants for whom the serum albumin results might have impacted the phototherapy decision. This would be the case if their TSB levels at initiation were equal to or greater than the phototherapy threshold for infants with neurotoxicity risk but less than the threshold for those without neurotoxicity risk, or if their TSB levels after phototherapy were equal to or greater than the discontinuation threshold for infants with risk but less than the discontinuation threshold for those without risk (indicating a need to continue phototherapy). As a result, low serum albumin levels affected the decision of phototherapy initiation in 13 infants, the cessation of phototherapy in 1 infant, and both the initiation and cessation of phototherapy in 2 infants (as illustrated in the boldly outlined boxes in Figure 2). Consequently, the identification of serum albumin less than 3.0 g/dL influenced the phototherapy decision in 16 infants, representing 1.7% of all infants evaluated for jaundice. The individual characteristics of these 16 infants are detailed in supplementary table.

During the study period, 3 infants had TSB levels exceeding the ET threshold. All of them had bilirubin neurotoxicity risk factors. However, none received ET because their TSB levels decreased shortly after intensive phototherapy treatment. Two infants had TSB levels that exceeded the escalation-of-care threshold but were lower than the ET threshold. Both were successfully treated with intensive phototherapy. In contrast, if the recommendation for ET based on the bilirubin to albumin ratio (B/A) in the AAP 2022 guidelines were used, 14 infants would qualify for ET, whereas only 1 infant whose TSB exceeded the bilirubin threshold for ET and 1 infant whose TSB fell in the “escalation-of-care” zone would be included.

Our study is one of the first to investigate how serum albumin assessment in infants with early neonatal jaundice influences the decision-making process for phototherapy using the phototherapy threshold designated by the new AAP 2022 guidelines. In our infant cohort whose serum albumin were obtained as part of early neonatal jaundice evaluation, despite albumin levels less than 3.0 g/dL being common findings, its impact on the decision to start or stop phototherapy is small.

Limited data are available on the serum albumin levels of newborns evaluated for jaundice, with early studies primarily reporting normal albumin values in preterm and term infants.^6,7,9,10 A small study by Cartlidge et al reported that the 50th percentile value for serum albumin in newborns 35 to 42 weeks’ gestation is around 3.0 g/dL.⁸ Similarly, Watchko et al used a large cross-sectional data set of neonates 23 to 41 weeks’ gestation in the NICU and found that up to 50% of infants of 35 weeks’ gestation or more had serum albumin less than or equal to 3.0 g/dL.¹¹ Ikuta et al developed a gestational age-dependent reference range for albumin levels in cord blood of infants 22 to 42 weeks’ gestation, finding a steady increase with gestational age and the 50th percentile for infants at 35 weeks’ gestation being 3.0 g/dL.¹² These findings indicate that as many as 50% of newborns born after 35 weeks’ gestation may have serum albumin levels less than or equal to 3.0 g/dL, which contributes to the risk of bilirubin neurotoxicity. Our study found that 26.7% of all newborns over 35 weeks’ gestation assessed for early neonatal jaundice had serum albumin levels less than 3.0 g/dL. Although this figure is lower than has been previously estimated, possibly due to variations in the populations studied (such as NICU infants vs all infants or cord blood vs serum after birth), our result is still consistent with prior findings.

The following are the incidence of other risk factors for bilirubin neurotoxicity in our study: 18.3% for isoimmune hemolytic disease (7.9% had ABO isoimmunization, 0.2% had Rh isoimmunization, 9.8% had G6PD deficiency, and 0.9% had other hemolytic conditions); 9.9% for sepsis; and 9.8% for significant clinical instability in the previous 24 hours (Table 1). Prior research on the incidence of different risk factors for bilirubin neurotoxicity in infants has been sparse, and the reported incidence may vary depending on the population. For G6PD deficiency, previous studies have shown prevalence in Thai male infants ranging from 3% to 18%, depending on the geographic region.¹³ A recent study in northeastern Thailand reported that 23.7% of infants greater than 35 weeks’ gestation had G6PD deficiency as a cause of early neonatal jaundice, which is much higher than observed in our study. However, the reported incidence of ABO isoimmunization in their cohort was 10.9%, comparable to our data.¹⁴ Yang et al retrospectively assessed 435 term newborns with hyperbilirubinemia over 4 years in southeast China. They found that 9.7% had G6PD deficiency, 3.5% had ABO isoimmunization, and 16.1% had infection as causes of early neonatal jaundice. However, in their analysis, 22.3% were classified as having a mixed etiology, indicating that there may be more infants with each of the etiologies listed earlier.¹⁵ Although the incidence of ABO isoimmunization in this Chinese study was lower, and the incidence of infection (including sepsis) was higher than in our report, the incidences of G6PD deficiency were similar.

Only one-third of newborns in our centers received phototherapy at or greater than the AAP-designated threshold. Most (42.4%) patients in our centers received phototherapy when TSB was within 2 mg/dL less than the phototherapy threshold. This practice remains consistent with the AAP 2022 guidelines, which propose initiating phototherapy at a lower threshold as an alternative approach to mitigate the risk of readmission. Despite that, almost 30% of phototherapy was given for the TSB far less than the recommended threshold (ie, >2 mg/dL less than the threshold). Our results were comparable with the recent study of Sarathy et al, which analyzed the effect of the new AAP 2022 guidelines in a large cohort of greater than 22 000 infants from 8 birthing hospitals. In their study, more than 60% of their newborns received phototherapy at TSB levels lower than the recommended thresholds, with more than 20% of those receiving phototherapy at TSB greater than 2 mg/dL lower than the threshold.¹⁶ In our centers, the cost of readmission for phototherapy after discharge, which includes both hospital and family expenses, is significantly higher than the cost of administering phototherapy in the nursery after birth. This cost discrepancy has likely led many physicians to initiate phototherapy prophylactically at levels far less than the threshold.

Because the decision to provide phototherapy for infants with TSB less than the phototherapy threshold is at the physician’s discretion, using the total number of all infants who received phototherapy for analysis would not be accurate. For this reason, only infants who received phototherapy according to the TSB cut point designated by the AAP guidelines were used for this report to determine the utility of serum albumin assessment.

Routine measurement of albumin as part of the workup for diagnosing the cause of jaundice does not appear to be justified. Our study showed that identifying serum albumin less than 3.0 g/dL affected the decision to undergo phototherapy in 16 infants or 1.7% of all 935 infants evaluated for jaundice. The low yield of serum albumin to affect any treatment decisions is mostly due to the fact that nearly one-half of infants with serum albumin less than 3.0 g/dL already had 1 or more bilirubin neurotoxicity risk factors, especially clinical instability or sepsis (Figure 1). Prior research has shown a correlation between reduced levels of serum albumin and neonatal morbidity, including respiratory distress syndrome,^17,18 infections,¹⁹ necrotizing enterocolitis,²⁰ and even mortality.^11,21 These findings align with our observations. On the other hand, had serum albumin not been performed in these 16 infants, they would not have been placed on phototherapy (or have their phototherapy discontinued earlier), as the phototherapy threshold would have been higher because they were presumed to have normal serum albumin (≥3.0 g/dL). The clinical consequences of misclassifying infants with low albumin as having no neurotoxicity risk factor, thereby subjecting them to a higher phototherapy threshold (ie, not administering phototherapy), are unclear. However, our study provides some reassurance that only a few infants may have been missed. The serum albumin level of these newborns in our study ranged from 2.77 to 2.99 g/dL (supplementary table), which were not likely to cause a clinically significant risk of neurotoxicity, when compared with those whose levels were at least 3.0 g/dL. In addition, although routine evaluation of serum albumin is not routinely performed in most places, no deleterious consequences have been reported in the literature thus far. Nevertheless, from an economic standpoint, it might be more cost-effective not to perform routine testing of serum albumin as suggested by the AAP. Serum albumin testing in our laboratory is priced at $1.5/specimen. If routine albumin testing had not been performed, we could have saved $1395 from 930 infants (excluding the 5 infants with TSB in the escalation-of-care and ET zone).

Interestingly, if the recommendation for ET was based solely on the bilirubin cut point in AAP 2022 guidelines, only 3 infants would qualify for ET, and 2 infants would fall into the escalation-of-care zone described earlier. In contrast, if the recommendation for ET was based on the B/A ratio mentioned in the guidelines, 14 infants in our cohort would qualify for ET, whereas only 1 infant whose TSB exceeded the bilirubin threshold for ET and 1 whose TSB fell in the “escalation-of-care” zone would be included. The criteria for ET in infants with jaundice using the B/A ratio in the AAP 2022 guidelines was adapted from the work of Ahlfors in 1994. The B/A ratio was created as a surrogate for bilirubin and albumin binding, correlating with free (unbound) bilirubin concentration in newborns with jaundice. The thresholds for the B/A ratio suggested by Ahlfors were defined to be correlated with the TSB threshold for ET that was widely accepted at that time, which was lower than the current recommendation (eg, TSB cut point for ET in full-term infant weight ≥2500 g without risk factors was 20 mg/dL).²² Therefore, the discrepancy between the B/A ratio and the TSB for ET threshold in the AAP 2022 guidelines could generate inconsistency in clinical practice. It might necessitate a comprehensive reassessment of the guideline to ensure coherence and effectiveness.

Our study has several limitations. First, we did not include hemolysis caused by bleeding (such as cephalhematoma or subgaleal hemorrhage) as neurotoxicity risk factors because it is unclear whether hemolysis caused by these conditions increases the risk of bilirubin neurotoxicity.⁴ Thus, our reported numbers of newborns receiving phototherapy based on the threshold recommendation may be underestimated, and the influence of low serum albumin in establishing the threshold may be even smaller if hemolysis from bleeding were included as neurotoxicity risk factors. Second, because almost 70% of phototherapy treatments were administered to infants with TSB levels less than the phototherapy threshold, it is unclear whether albumin levels would have significantly affected phototherapy decisions in these cases. Finally, although this study included infants from 3 tertiary care hospitals affiliated with a single academic institution, our sample size was limited. Hospital-specific factors may have affected patients’ characteristics and jaundice workup results. Therefore, more extensive population-based studies are required to validate our findings.

Serum albumin levels less than 3.0 g/dL are common, impacting 1 in 4 infants evaluated for jaundice. However, its impact on phototherapy decisions is minimal.

We greatly appreciate the help of Prof Pornpen Srisawasdi and Ms Sirirat Promnuch at the Department of Pathology at Ramathibodi Hospital for providing laboratory data of patients at SDMC and PYT, as well as Prof Sinitdhorn Rujirabanjerd and Mr Atthapong Sararat at the Laboratory Department at CNMI for providing laboratory data of patients at CNMI. We thank Prof Pracha Nuntnarumit for reviewing our manuscript and providing valuable suggestions.