Page 103 - 16Neonatal Jaundice_compressed

Neonatal jaundice

Table 5.1 Negative predictive value (NPV) of low degree of jaundice assessed by visual inspection
Study details Sample characteristics Timing of Indicator of absent Definition of severe Results
assessment jaundice or low-grade hyperbilirubinaemia
jaundice
Riskin et al. Healthy full-term and Mean: Clinical icterus assessed Serum bilirubin levels NPV: 91.9%
61
(2008) late preterm babies with 62 ± 24 hours to be in zone A/low-risk in zone B, C and D or (2627/2857)
GA≥ 35 weeks before (median zone on Bhutani’s in intermediate (low, Negative LR:
discharge (n = 1129) 55 hours; nomogram high) and high-risk 0.45
range 9– (< 40th centile) zones on Bhutani’s
252 hours) nomogram
(> 40th centile)
Moyer et al. Full-term healthy babies Mean age Presence of icterus in Serum bilirubin levels NPV: 94.3%
(2000) with BW > 2000 g and 2 days (range lower chest (nipple line > 205 micromol/litre (33/35)
60
GA > 36 weeks 8 hours to to umbilicus) Negative LR:
(n = 122) 7 days) 0.15
Szabo et al. Healthy full-term babies Data not Kramer zone 2 assessed Serum bilirubin levels NPV: 100%
56
(2004) with BW > 2000 g and given by nurses (data not given > 250 micromol/litre
no older than 6 days for zone 0 or 1)
(n = 140)
Excluded: jaundice
within 36 hours
BW = birthweight; GA = gestational age; LR = likelihood ratio; NPV = negative predictive value

When parents or healthcare professionals consider that a baby is not visibly jaundiced, this
assessment is generally reliable in ruling out hyperbilirubinaemia. The NPV of absence of
jaundice ranged from 91% to 100% in the studies used in the meta-analysis.

Whenever parents or healthcare professionals consider that a baby is visibly jaundiced, the
bilirubin level needs to be measured within hours so the depth of jaundice can be accurately
assessed and appropriate care initiated.
Recommendations

See the end of Section 5.1.
5.1.2 Examining urine/stool

Description of included studies
62
One study of EL III from the UK was identified.
Review findings
A single non-diagnostic study (project report) from the UK was identified to provide evidence
62
for this test. This study reported the results of a community programme conducted in three
phases in which stool colour charts were used to determine liver disease during the neonatal
period. In the first phase, parents were asked to record the colour of their baby’s stools during
the first 28 days of age. The first phase recruited 109 parent–baby pairs and 5053 stool
observations were made. The six most commonly selected stool colours were then combined
with three pale colours to develop a simplified stool colour chart during the second phase. In
the third phase, acceptability and specificity of this chart was evaluated among 3629 mothers at
the time of the first health visitor visit (usually around 10–14 days). During the second visit (at
28 days), the health visitors collected the information and examined the babies. Any baby
thought to be jaundiced or with a history of passing ‘pale stools’ was referred, investigated for
the presence of cholestatic jaundice and followed up for 6 months. In total, 127 babies were
jaundiced at 28 days of age, with the incidence of jaundice in breastfed babies being 9.2%
(95% CI 7.8–11.0%). Many of these babies had abnormal liver function tests but none had
abnormal stool/urine colour and none was found to have liver disease. Four non-jaundiced
babies were reported to pass pale stools (fewer than three occasions in all), but they were not
investigated as stools returned to normal colour and all were thriving at the 6 months follow-up.

74