 |
|
Original Articles Indian Pediatrics 2002; 39:724-730 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Early Neonatal Hyperbilirubinemia Using First Day Serum Bilirubin Level |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Ramesh Agarwal, Monica Kaushal, Rajiv Aggarwal, V.K. Paul and A.K. Deorari
NEONATAL hyperbilirubinemia is a cause of concern for the parents as well as for the pediatricians. It occurs in 5-10% of healthy term infants(1,2) and is the most common reason for readmission after early hospital discharge(3). Concerns regarding jaundice have increased after reports of bilirubin induced brain damage occurring in healthy term infants, even without hemolysis(4). Total serum bilirubin (TSB) in infants discharged within 48 hr of age generally shows an increasing trend and some of these infants later develop hyper-bilirubinemia. In a cohort of 500 healthy term infants, Alpay, et al found that hyper-bilirubinemia (serum bilirubin ³17mg/dl) occurred only after 72 hr of age(5). The American Academy of Pediatrics recommends that newborns discharged within 48 hours should have a follow-up visit after 2-3 days to detect significant jaundice and other problems(6). This recommendation is not possible in our country due to limited follow up facilities in the community. The concept of prediction of jaundice offers an attractive option to pick up babies at risk of neonatal hyperbilirubinemia. An association between cord bilirubin levels and the subsequent risk of hyperbilirubinemia has been reported(7,8). Infants who are clinically jaundiced in the first few days are more likely to develop hyperbilirubinemia later(9). Hour specific percentile charts based on serum bilirubin at different postnatal ages have been developed(10). They show that subsequent hyperbilirubinemia can be predicted with reasonable accuracy by plotting hr specific bilirubin on these charts. A TSB level ³6 mg/dl within 21 hrs of birth has a high predictive value for hyperbilirubinemia later(5). In a study, from India, a TSB level of 3.99 mg/dl at 18-24 hr was found to predict subsequent hyperbilirubinemia (>15 mg/dl) with sensitivity and specificity of 67% each(11). Similarly prediction of hyper-bilirubinemia in patients with ABO incompatibility was attempted(12). The present study was conducted to evaluate the predictive value of TSB level ³6 mg/dl at 24 ± 6 hr postnatal age for identifying near term and term infants at low risk for subsequent hyperbilirubinemia. Additionally, bilirubin values were plotted on nomograms developed by Bhutani, et al. to find out their usefulness in Indian infants(10). Subjects and Methods This was a prospective study conducted at All India Institute of Medical Sciences, New Delhi between May and September 2001. All infants with (1) gestation ³ 35 weeks (based on last menstrual period and neonatal assessment by expanded New Ballard score) (2) absence of significant illness requiring NICU admission for > 12 hr (3) absence of major congenital malformations and (4) residing near the hospital and whose parents agreed to come for follow-up were included. Infants who later developed significant illness requiring NICU admission and infants with Rh hemolysis were excluded. Infants of Rh-negative mother were included if either they were also Rh negative or had no evidence of hemolysis. All the infants were discharged as per unit policy (after 72 hr). For sake of convenience infants born between 4 AM to 2 PM were included. Bilirubin estimation Serum bilirubin estimation was performed at 24 ± 6 hr after birth, in triplicate, by a blinded observer and average of two closest values was taken. Venous whole blood was taken in three micro-capillaries and centrifuged at the rate of 10000 rpm for 5 minutes. Bilirubin estimation was done spectrophotometrically using twin beam method (540 and 465 nm wavelengths) (Bil Micrometer - Auto 2 Type, Kohsoku Denki, Tokyo, Japan). A single observer, who was not involved in follow up, did all measurements. Calibration of bilimeter was done daily using standard solution. Coefficient of variation was less than 5%. Follow up The infants were followed up clinically every 12 hr till discharge by two DM fellows, who were unaware of the 24 ± 6 hr TSB value of the baby. The infants were recalled after discharge at 5 days of age (Fig.1). If clinical assessment of bilirubin level was more than 10 mg/dl by any observer, bilirubin estimation was repeated as above. Hyperbilirubinemia was defined as TSB level of ³ 17 mg/dl. The infants were investigated and managed as per unit policy. Informed and written consent was taken from the parents. Sample Size Assuming the incidence of hyperbilirubinemia to be 1-3% in infants with TSB £ 6 mg/dl at 24 ± 6 hours(12) with alpha value of 0.05 and beta value of 0.1, it was estimated that a sample size of 150 infants would be sufficient. To account for dropouts, we enrolled 220 infants. Statistical Analysis Maternal and neonatal data were collected in predesigned and pretested proforma. Sensitivity, specificity, negative and positive predictive values and likelihood ratio of the test were calculated. Bilirubin values were plotted on previously published nomograms(10). Results Of 252 eligible infants, 220 were enrolled. The baseline characteristics of infants not enrolled were similar to those of enrolled infants. Seven infants were excluded later; three were previously enrolled in some other study, one developed hypoglycemic seizures on second day of life and three were lost to follow up. Complete information was available on 213 (96.8%) infants. Their baseline data is shown in Table I. All infants were exclusively breastfed. Mean age at bilirubin estimation was 24.7 ± 1.9 hr (range 20-30 hr) with mean TSB of 5.9 ± 1.8 (range 2-15) mg/dl. Table I- Baseline characteristics of enrolled infants (N = 213)
# Mean ± SD Clinically detectable jaundice was present in 164 (77%) infants. Study outcome, as defined by TSB of ³17mg/dl, occurred in 22 (10.3%) infants. Peak serum bilirubin of these 22 babies, at 94.4 ± 24.5 hr of age, was 19.3 ± 1.9 mg/dl. Five infants were less than 37 weeks and all developed hyperbilirubinemia after 72 hr of age. Twenty infants required phototherapy and none required exchange transfusion. In two infants TSB was below phototherapy range. ABO incompatibility was present in one and G6PD deficiency in three infants. As is evident in the scatter chart of TSB at 24 ± 6 hr with respect of age (Fig.2), no baby except one who had TSB £ 6mg/dl developed hyperbilirubinemia. To identify a cut-off predicting a large number of babies at low risk of hyperbilirubinemia, a receiver operating characteristic (ROC) curve was drawn (Fig.3). Ideal cut-off was £5 mg/dl as no baby below this level developed hyperbilirubinemia. This level, however, identified lesser number of babies at low risk of hyperbilirubinemia. A cut-off TSB level £6 mg/dl had high specificity and satisfactory sensitivity for identifying babies at low risk of subsequent hyperbilirubinemia. A TSB of £6 mg/dl at 24 ± 6 hours was present in 136 (63.8%) infants and only one infant developed hyperbilirubinemia subsequently (Table II). In remaining 77 (36.2%), with TSB > 6mg/dl, subsequent hyperbilirubinemia developed in 21 (27.2%) infants (sensitivity 95%, specificity 70.6%, positive predictive value 27.2%, negative predictive value 99.3%, likelihood ratio of positive test 3.23 and likelihood ratio of negative test 0.07). Babies with TSB levels higher than 6 mg/dl had a significant risk of developing hyperbilirubinemia (relative risk 38; 95% confidence internal 6 - 1575). Table II- Predictive value of serum bilirubin (N = 213)
Sensitivity 95%; specificity 70.6%; positive predictive value 27.2%; negative predictive value 99.3%
When classified in risk zones as per standard nomograms, 67% infants were in either low risk or low intermediate risk zone with very low probability of developing hyperbilirubinemia. Subsequently hyper-bilirubinemia developed in 20 (28.6%) infants in the high intermediate or high-risk zones. Discussion The present study found that a TSB level £6 mg/dl at 24 ± 6 hr can be used to predict the decreased risk for subsequent hyperbilirubinemia (TSB > 17 mg/dl). Such infants could thus be discharged early without need to follow up for hyperbilinubinemia later. Bhutani et al(10) showed in a large cohort that infants who develop hyperbilirubinemia have serum bilirubin levels, which are in higher percentiles soon after birth. The authors created percentile charts of serum bilirubin level at different postnatal ages in near-term and term infants who were direct Coombs test negative. They found that 6.1% of neonates had pre-discharge serum bilirubin ³ 95th percentile; 32.1% of these infants showed hyperbilirubinemia sub-sequently(10). Neonates with pre-discharge TSB levels in the low risk zone (< 40th percentile) did not show hyperbilirubinemia subsequently. However, there was an important source of bias in this study. Out of around 13,000 neonates, subsequent bilirubin estimation could be done in only around 25%. The infants included those who came for follow up or were referred by their primary physicians. A large number of infants were thus not included. It is likely that infants without significant problems were not included while developing these percentile charts. Later Alpay, et al reported that TSB levels of ³6 mg/dl in the first 24 hr predicted jaundice in all newborns subsequently(5). Awasthi, et al. showed that TSB level of 3.99 mg/dl at 18-24 hr was able to predict subsequent hyperbilirubinemia (>15 mg/dl) with sensitivity and specificity of 67% each(11). However, this study had major flaws. Complete follow up was present in infants who stayed in the hospital either for neonatal illness or some maternal reason, such as cesarean section. More than 50% of infants, who were healthy thus discharged early, were not followed up. In a study at our center, prediction of hyperbilirubinemia in patients with ABO incompatibility was attempted using healthy term and borderline term infants as controls(12). A TSB level of 5 mg/dl at postnatal age of 24 ± 6 hr was found to predict subsequent hyperbilirubinemia with high sensitivity, specificity, and negative and positive predictive values. Results from the present study are consistent with previous studies. We found that it was possible to identify infants early who are at high risk and more importantly, those who are at low risk of subsequent hyperbilirubinemia. Our study population was heterogeneous (Table I). We included infants of gestational diabetic mothers, infants with ABO incompatibility and with G6PD deficiency. Oxytocin use was high, which is a known risk factor for hyperbilirubinemia. Nearly all infants were exclusively breastfed. Complete follow up for atleast five days was recorded for 96.8% infants. We believe that data from this study could be generalized. The incidence of hyperbilirubinemia was higher in present study (10.3%) compared to others(1,2). This could be explained on basis that borderline term infants with less than 37 weeks gestation were included. A significant proportion of babies were born to mothers who received oxytocin. Our study has few limitations. The infants were in the hospital till 72 hr of life. If they did not have significant clinical jaundice (clinical levels 8-10 mg/dl), they were discharged to be followed up after 48 hr. It is possible that a small number of infants had transient elevation reaching the study definition of outcome and came down again at low level on day five when they came for follow up. However this seems more of theoretical risk. Even if this did happen, it should not be a cause for concern in real life situation also. We followed up infants for five days. If there was no significant clinical jaundice ( < 8-10 mg/dl) on day 5, the infants were not followed subsequently. Some of these infants might have developed late jaundice and were missed. In conclusion, we showed that TSB level of £6 mg/dl at 24 ± 6 hr of life predicts absence of subsequent hyperbilirubinemia with high probability. Prediction of neonatal hyperbilirubinemia should have widespread implication especially in our country where there are limited resources and few hospital beds. Infants at low risk for hyperbilirubinemia can be discharged early from the hospital. The hour specific percentile curves developed for North American population(10) can be used for Indian babies. Acknowledgement We thank Dr. N.K. Arora, Additional Professor, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi for his guidance in planning and analysis of this study. Contributors: All were involved in designing of the study. RA coordinated the study and drafted manuscript. AKD, VKP and RAG critically reviewed the manuscript. MK participated in data collection. AKD will act as guarantor for the paper. Funding: None. Competing interests: None stated.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
References | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
