Page 69 - 16Neonatal Jaundice_compressed

Neonatal jaundice

predictive accuracy of a risk index model. The cohort consisted of 51 387 babies with
birthweight ≥ 2000 g and gestational age ≥ 36 weeks born at these hospitals during a 2-year
period. Babies with peak serum bilirubin levels ≥ 427 micromol/litre within the first 30 days
after birth were defined as cases (n = 73), while controls were a random sample of babies from
the cohort with maximum serum bilirubin levels below this level (n = 423). Information on the
risk factors was collected by reviewing hospital records and interviewing parents. Using
bivariate analysis, various clinical and demographic factors were found to be associated with an
increased risk of hyperbilirubinaemia. The maternal factors considered included race, maternal
age, history of jaundice in a previous sibling, and vacuum delivery. Neonatal factors considered
included male sex, lower gestational age, early jaundice (defined either as bilirubin levels
exceeding age-specific phototherapy thresholds, or phototherapy during birth hospitalisation, or
jaundice noted in first 20 hours and bilirubin levels were not taken within 6 hours of that time),
cephalohaematoma, bruising, and exclusively breastfed at time of discharge. These factors were
then entered into multiple regression analysis to find independent predictors of
hyperbilirubinaemia. When all cases were included, the presence of early jaundice (adjusted
odds ratio (OR) 7.3, 95% CI 2.8 to 19.0), gestational age (in weeks) at birth (adjusted OR 0.6,
95% CI 0.4 to 0.7), exclusive breastfeeding at discharge (adjusted OR 6.9, 95% CI 2.7 to 17.5),
Asian race (adjusted OR 3.1, 95% CI 1.5 to 6.3), the presence of bruising (adjusted OR 3.5,
95% CI 1.7 to 7.4) , cephalohaematoma (adjusted OR 3.2, 95% CI 1.1 to 9.2), and maternal age
≥ 25 years (adjusted OR 2.6, 95% CI 1.1 to 9.2) were all independently associated with
hyperbilirubinaemia. When cases with early jaundice were excluded, the results were similar
except that family history of jaundice showed evidence of statistically significant association
with later hyperbilirubinaemia (adjusted OR 6.0, 95% CI 1.0 to 36.0). [EL II]
The above study was expanded in order to examine the association between jaundice noted in
10
the first 24 hours of life and the risk of later hyperbilirubinaemia and the need for phototherapy.
This study included babies born during a period of 4 years (compared to 2 years in the first
9
study ) and the baseline cohort population included 105 384 newborn babies. The criteria for
study selection and definitions of cases (n = 140) and controls (n = 631) were unchanged.
Information on the timing of the appearance of jaundice was extracted by medical records
analysts and this process was reliably assessed by a second analyst blindly re-abstracting data
from a random sample of 25 medical records (κ statistic for agreement = 0.75). Data on the use
of phototherapy and development of hyperbilirubinaemia (maximum serum bilirubin levels
≥ 427 micromol/litre) were also obtained from hospital records. Among the controls, the
cumulative probability of jaundice being noticed within 18 hours of birth was 2.8% and within
24 hours of birth it was 6.7% (these proportions were estimated using Kaplan–Meier survival
analysis after correcting for age of discharge). On adding the number of newborns who had
serum bilirubin measured within 24 hours (as a proxy measure of jaundice noticed in first
24 hours) to the above data, the proportions increased to 3.8% by 18 hours and 7.9% at
24 hours. There was no statistically significant association between jaundice noticed within
24 hours and risk factors such as ethnicity, sex, gestational age, breastfeeding or
cephalohaematoma. Although most of the babies did not require any intervention, these babies
were 10 times more likely to be treated with phototherapy compared with newborns noted not
to have jaundice in the first 24 hours (18.9% versus 1.7%; Mantel–Haenszel OR 10.1, 95% CI
4.2 to 24.4). Moreover, the early jaundiced babies were found to have a statistically significant
increase in the risk of developing hyperbilirubinaemia above 427 micromol/litre (14.3% versus
5.9%; Mantel–Haenszel OR 2.9, 95% CI 1.6 to 5.2). [EL II]
11
Another nested case control study from the USA estimated the effect of phototherapy and
other factors on the risk of developing severe hyperbilirubinaemia (defined as serum bilirubin
levels ≥ 427 micromol/litre) in babies who had serum bilirubin levels close to the American
19
Academy of Pediatrics (AAP) phototherapy threshold levels. The cohort included 285 295
babies with gestational age ≥ 34 weeks and birthweight ≥ 2000 g born between 1995 and
2004 in a health maintenance organisation. Babies with resolving jaundice, those whose serum
bilirubin levels were not fully documented, and those with conjugated bilirubin level
≥ 34 micromol/litre were excluded. A subset of babies (n = 13 843) with a serum bilirubin level
between 291 and 392 micromol/litre at ≥ 48 hours of age was identified. Babies with serum
bilirubin concentration ≥ 427 micromol/litre were selected as cases (n = 62), and four controls
were selected randomly for each case (n = 248). Cases and controls were matched for risk status

42