1. Background

Neonatal hyperbilirubinemia remains a significant global health concern, particularly in low- and middle-income countries. According to the Global Burden of Disease 2016 report, neonatal jaundice contributes to approximately 1,008 deaths per 100,000 live births, with the highest burden observed in South Asia and sub-Saharan Africa, where an estimated 1.1 million infants are affected annually ¹ [1]. Clinically, neonatal hyperbilirubinemia is characterized by elevated serum bilirubin levels exceeding 85 µmol/L (5 mg/dL), although some studies define significant hyperbilirubinemia as total serum bilirubin levels exceeding 342 µmol/L in term and late preterm infants [2], typically presenting as jaundice, and a yellow discoloration of the skin, conjunctival icterus, and mucous membranes resulting from bilirubin deposition [3].

The pathophysiology of neonatal hyperbilirubinemia is multifactorial and involves increased bilirubin production from the breakdown of fetal erythrocytes, impaired hepatic conjugation due to immature enzyme systems, and enhanced enterohepatic circulation [4,5].

A wide array of maternal, neonatal, and genetic factors influences the risk and severity of neonatal hyperbilirubinemia. Preterm neonates are particularly susceptible because of underdeveloped hepatic function, which compromises bilirubin metabolism [5]. Hemolytic conditions such as Rh and ABO blood group incompatibilities contribute to elevated bilirubin levels through immune-mediated hemolysis [4]. Feeding practices also play a critical role; inadequate breastfeeding can result in dehydration and increased enterohepatic circulation, leading to early-onset breastfeeding jaundice within the first week of life [6]. Furthermore, genetic predispositions, such as polymorphisms affecting bilirubin conjugation, have been linked to higher rates of hyperbilirubinemia in East Asian and Mediterranean populations [7]. Additional risk factors include maternal infections of glucose-6-phosphate dehydrogenase (G6PD), gestational diabetes, and delayed meconium passage, all of which can contribute to elevated bilirubin levels [8].

Our study design therefore captures both early- and late-onset cases, providing a more comprehensive estimate of incidence. By determining the incidence and identifying key maternal and neonatal predictors of neonatal hyperbilirubinemia, this study contrasts with previous studies conducted in urban, better-resourced settings in Uganda and provides a clearer picture of the burden of hyperbilirubinemia at the regional referral hospital level. A study at Nsambya hospital showed the prevalence of neonatal hyperbilirubinemia was at 7.75% [9] and another conducted at Kawempe-Mulago hospital the prevalence was at 13.6% [10]. These studies were conducted in Kampala; Uganda’s capital and there’s no similar study conducted outside Kampala. The study conducted at Nsambya was retrospective and relied on chart reviews, which may have underestimated prevalence due to missing or incomplete data. The Mulago study focused exclusively on well neonates due for discharge between 24–72 hours of life, potentially missing cases of delayed-onset hyperbilirubinemia.

Findings from this study contribute to evidence-based screening and management strategies, particularly in resource-constrained settings where timely diagnosis and treatment remain a challenge.

2. Methods and materials

2.4. Study variables

2.4.1. Primary-outcome.

The primary outcome was neonatal hyperbilirubinemia requiring phototherapy, defined according to the American Academy of Pediatrics (AAP) guidelines. The cut-offs for initiating phototherapy were determined based on either gestational age and postnatal age as shown below:

Download:

• PNG
  larger image
• TIFF
  original image

https://doi.org/10.1371/journal.pone.0349861.t004

2.4.2 Exposure variables.

Maternal sociodemographic characteristics (age, education level, occupation, and household income) were collected using a structured questionnaire administered to mothers. Obstetric and clinical characteristics, including history of urinary tract infection, diabetes mellitus, hypertension, preeclampsia, premature rupture of membranes, antenatal care attendance, and duration of labor, were abstracted from antenatal cards and medical records. Diagnoses of hypertension, diabetes, and preeclampsia were based on documented clinical assessments by attending clinicians.

Neonatal characteristics, including sex, gestational age, birth weight, multiple gestation, mode of delivery, birth trauma, Apgar score at 5 minutes, and feeding practices, were obtained from delivery records and direct observation. Gestational age was primarily determined using LMP, with the New Ballard Score used when LMP was unknown or uncertain. Neonatal sepsis was diagnosed clinically using Integrated Management of Neonatal and Childhood Illness (IMNCI) criteria without laboratory confirmation. Maternal and neonatal blood group and Rhesus status were determined using direct slide agglutination testing.

3 Results

3.1 Baseline characteristics of the study participants

A total of 539 mother–infant pairs were enrolled in this study. Of these, nearly two-thirds (60.3%) of the neonates were female, and less than one-third (27.3%) were preterm. The majority were exclusively breastfed (79.0%), while 64.4% initiated feeding within the first hour of life. An Apgar score below seven at 5 minutes was observed in 29.7% of neonates. ABO incompatibility was identified in 16.9%. The baseline characteristics of the study participants are presented in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline maternal and neonatal characteristics of study participants (N = 539).

https://doi.org/10.1371/journal.pone.0349861.t001

3.2. Incidence of hyperbilirubinemia requiring phototherapy among neonates at Mubende Regional Referral Hospital

Of the 539 mother–infant pairs enrolled, five neonates were lost to follow-up after discharge; therefore, 534 neonates completed follow-up. Overall, 92 of the 534 neonates developed neonatal hyperbilirubinemia requiring phototherapy, giving an incidence of 172 per 1,000 neonates Tables 2 and 3 (17.2%; 95% CI: 14.0–20.6) Fig 1.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Proportion of neonates with and without hyperbilirubinemia requiring phototherapy among neonates enrolled at Mubende Regional Referral Hospital, Uganda (N  = 534).

https://doi.org/10.1371/journal.pone.0349861.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Bivariable analysis of risk factors for neonatal hyperbilirubinemia requiring phototherapy at Mubende Regional Referral Hospital.

https://doi.org/10.1371/journal.pone.0349861.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Multivariable analysis of risk factors for neonatal hyperbilirubinemia requiring phototherapy at Mubende Regional Referral Hospital.

https://doi.org/10.1371/journal.pone.0349861.t003

Following the back ward stepwise regression of the variables that were considered for multivariable analysis, only BMI, monthly income, gestational age, time of feeding initiation, mode of delivery, PROM and ABO incompatibility contributed to the final model and hence presented in the multivariable table (Table 3).

The risk of hyperbilirubinemia requiring phototherapy was increased by close to two times among neonates born to mothers who were underweight, overweight, or obese (based on maternal BMI during pregnancy) (aIRR = 1.703, CI = 1.082–2.680, P = 0.021; aIRR = 1.916, CI = 1.447–3.183, P < 0.001; aIRR = 1.930, CI = 1.233–3.022, P = 0.004), those from households with monthly income <200,000 UGX (aIRR = 1.798, CI = 1.306–2.477, P < 0.001), those delivered by cesarean section (aIRR = 1.835, CI = 1.291–2.610, P = 0.001), and those with premature rupture of membranes (aIRR = 1.538, CI = 1.109–2.132, P = 0.010). The risk of hyperbilirubinemia requiring phototherapy was increased by over two times for the preterm (aIR = 2.776, CI = 1.978–3.895, P < 0.001), those in whom feeding was initiated after 1 hour (aIRR = 2.145, CI = 1.411–3.261, P < 0.001), instrumental delivery (aIRR = 2.005, CI = 1.109–3.625, P = 0.021) and ABO incompatibility (aIRR = 2.985, CI = 2.060–4.325, P < 0.001).

4. Discussion

This study determined the incidence and predictors of neonatal hyperbilirubinemia requiring phototherapy at a tertiary healthcare facility in Uganda. Among the 534 neonates who completed follow-up, 92 developed neonatal hyperbilirubinemia requiring phototherapy, giving an incidence of 172 per 1,000 neonates (17.2%; 95% CI: 14.0–20.6%). This demonstrates a substantial burden of neonatal hyperbilirubinemia in this setting.

The observed incidence is lower than that reported at Nsambya Hospital, Kampala, where 227 per 1,000 neonates (22.7%) required phototherapy [9], but higher than the incidence reported at Mulago National Referral Hospital, Kampala, of 136 per 1,000 neonates (13.6%) [10]. These differences are likely due to variations in study design, timing of bilirubin measurement, and study populations. The Mulago study assessed bilirubin only at discharge between 24 and 72 hours of life, which may have missed cases of delayed-onset hyperbilirubinemia [10]. Differences in inclusion criteria and health-care infrastructure may also have influenced the reported incidences. These findings indicate that neonatal hyperbilirubinemia remains a significant clinical and public health concern in Uganda. The variation in incidence between facilities emphasizes the need for standardized monitoring protocols, early identification of at-risk neonates, and timely interventions to reduce morbidity associated with hyperbilirubinemia [10,11].

Studies from other low- and middle-income countries have reported variable incidences of neonatal hyperbilirubinemia. In Ethiopia, incidences of 205 and 246 cases per 1,000 neonates have been reported, while in Rwanda an incidence of 443 cases per 1,000 neonates was observed [11,12,13]. Most of these studies included neonates admitted to neonatal intensive care units, who are more likely to develop hyperbilirubinemia because of comorbid illnesses and complications.

In Cameroon, a lower incidence of 152 cases per 1,000 neonates was reported [14], which is lower than in our study. However, that study used retrospective data, which may have missed some cases. In Nigeria, a much higher incidence of 402 cases per 1,000 neonates was reported [15], likely because the study included hospitalized newborns who were more vulnerable. Similarly, a study from Iraq reported an incidence of 350 cases per 1,000 neonates [16], but this also involved hospitalized neonates, many of whom had other medical conditions. Differences in study design, inclusion criteria, and health-care settings likely explain the wide variation in reported incidences.

In this study, multivariable analysis identified seven independent predictors of neonatal hyperbilirubinemia requiring phototherapy: abnormal maternal BMI, low household income, prematurity, delayed initiation of feeding, cesarean or instrumental delivery, premature rupture of membranes, and ABO incompatibility.

Abnormal maternal BMI was significantly associated with increased risk of neonatal hyperbilirubinemia. This finding suggests that both maternal undernutrition and overnutrition may adversely influence neonatal bilirubin metabolism and early neonatal adaptation. Similar associations have been reported in studies from Ethiopia [11,12]. Maternal undernutrition has been linked to placental insufficiency and preterm birth, which predispose neonates to immature hepatic enzyme systems and reduced bilirubin conjugation [11]. In contrast, maternal overnutrition is associated with metabolic and inflammatory changes that may impair fetal liver maturation and bilirubin metabolism [12].

Low household income status was also associated with an increased risk of neonatal hyperbilirubinemia requiring phototherapy. Neonates born to mothers from low-income households were nearly twice as likely to develop hyperbilirubinemia compared with those from higher-income households, consistent with findings from Nigeria [15]. Low household income is often associated with inadequate maternal nutrition, limited access to quality antenatal care, and delayed health-seeking behavior, all of which increase the risk of adverse perinatal outcomes. Financial constraints may also limit access to early postnatal follow-up and delay the recognition and management of neonatal jaundice, allowing bilirubin levels to rise to levels requiring phototherapy [15].

Neonates delivered by cesarean section were nearly twice as likely to have hyperbilirubinemia compared with those born via spontaneous vaginal delivery. A study conducted in Ethiopia and Uganda showed similar associations [10,11]. Cesarean delivery is often associated with delayed initiation of breastfeeding due to postoperative maternal pain, sedation, and separation of the mother–infant dyad. Inadequate early feeding may lead to dehydration and increased enterohepatic circulation of bilirubin, resulting in elevated serum bilirubin levels. Additionally, neonates born by cesarean section may experience delayed physiological adaptation and altered hepatic enzyme activity, which can further impair bilirubin conjugation and clearance. In emergency cesarean sections, underlying obstetric complications such as fetal distress or prolonged labor may also contribute to increased bilirubin production [10,11].

Premature rupture of membranes was significantly associated with an increased risk of neonatal hyperbilirubinemia requiring phototherapy. Similar associations have been reported in Rwanda [13] and Ethiopia [11]. This finding suggests that intrapartum and perinatal infectious and inflammatory processes contribute to the development of neonatal hyperbilirubinemia. PROM predisposes neonates to early-onset sepsis, which can impair hepatic bilirubin conjugation through inflammation-mediated cholestasis and reduced hepatocellular function. In addition, subclinical infections and systemic inflammation may increase hemolysis and bilirubin production, further elevating serum bilirubin levels. Neonates born after PROM are also more likely to experience delayed initiation of breastfeeding or require closer clinical observation, which may indirectly contribute to bilirubin accumulation [11,17].

Prematurity was a strong independent predictor of neonatal hyperbilirubinemia requiring phototherapy, with preterm neonates being nearly three times more likely to develop clinically significant hyperbilirubinemia compared with term neonates. Similar findings were reported at Mulago National Referral Hospital in Kampala [10]. Preterm infants have immature hepatic enzyme systems, particularly reduced activity of uridine diphosphate-glucuronosyltransferase (UGT1A1), which limits bilirubin conjugation and excretion [10,11].

This study found that neonates whose mothers-initiated breastfeeding after the first hour of birth were more than twice as likely to develop neonatal hyperbilirubinemia requiring phototherapy, consistent with findings from Ethiopia [11]. Early breastfeeding promotes the passage of meconium, which facilitates bilirubin excretion and reduces its accumulation [11]. This finding emphasizes the importance of promoting immediate or early initiation of breastfeeding within the first hour of life as a practical and cost-effective intervention to prevent severe neonatal hyperbilirubinemia.

ABO incompatibility showed the strongest association in this study, with nearly a threefold increased risk of neonatal hyperbilirubinemia requiring phototherapy. This condition, particularly when mothers with blood group O develop antibodies against a neonate with A or B blood group, leads to hemolysis and subsequent hyperbilirubinemia. Our findings are consistent with data from Turkey, Egypt, Nepal, India, Nigeria, and Pakistan [18–20]. This association was also confirmed in a previous study conducted at Mulago [10].

5. Strengths and limitations

The strengths of this study include its prospective cohort design, which minimized recall bias, and the use of total serum bilirubin measurement, which improved diagnostic accuracy. The relatively large sample size and inclusion of both early- and late-onset cases also strengthened the reliability of the incidence estimate.

However, some limitations should be considered. Glucose-6-Phosphate Dehydrogenase (G6PD) deficiency, an important contributor to neonatal hyperbilirubinemia, was not assessed, which may have led to underestimation of the true burden of the condition in this setting where Glucose-6-Phosphate Dehydrogenase (G6PD) deficiency, is common. Neonatal sepsis was diagnosed using Integrated Management of Neonatal and Childhood Illnesses (IMNCI) clinical criteria without laboratory confirmation such as C-Reactive Protein (CRP) or blood cultures, which may have resulted in misclassification. Gestational age estimation based on the last menstrual period may have been affected by recall bias, particularly among women with irregular menstrual cycles or early pregnancy bleeding. Furthermore, neonatal assessments were conducted at discrete time points (Day 1 and Day 7) rather than through daily bilirubin monitoring, which may have limited the precision in capturing the timing and incidence of hyperbilirubinemia.

6. Conclusion

This study found a moderately high incidence of neonatal hyperbilirubinemia, affecting almost 2 in 10 neonates in a tertiary hospital in Uganda. Abnormal maternal BMI, low household income, prematurity, delayed initiation of feeding, cesarean or instrumental delivery, premature rupture of membranes, and ABO incompatibility were significant independent predictors of hyperbilirubinemia.

Routine screening for hyperbilirubinemia within health-care facilities, especially during the first week of life, should be strengthened. Screening for ABO incompatibility should be integrated into routine maternal and neonatal care. Public health efforts should emphasize early initiation of breastfeeding, maternal nutritional support, and caregiver education on early signs of hyperbilirubinemia and timely care-seeking. Expanding access to phototherapy in peripheral health facilities and strengthening referral systems will be critical in reducing delays in treatment and improving neonatal outcomes.

Supporting information

Acknowledgments

We would like to thank all the participants for their valuable contributions to this study. We also extend our sincere appreciation to the special midwives, nurses in the NICU, delivery, and maternity wards at Mubende Regional Referral Hospital for their support, cooperation, and dedication throughout the study.