Page 105 - 16Neonatal Jaundice_compressed

Neonatal jaundice

serum bilirubin estimation. The Ingram icterometer showed a poor linear correlation with serum
bilirubin values (r = 0.48; P < 0.01), and had a sensitivity of 75% and specificity of 72% in
predicting serum bilirubin > 205 micromol/litre with a threshold reading ≥ 2.5. [EL II]
Another study, from Turkey, compared the correlation of both the Ingram icterometer and the
66
Minolta JM-102 bilirubinometer with serum bilirubin levels. The study sample comprised 96
full-term jaundiced babies between 1 and 5 days of age with a mean birthweight of
3380 ± 419 g. Within 30 minutes of blood sampling for serum bilirubin levels, and without
knowledge of the result, transcutaneous bilirubin levels were obtained from the forehead with
the Minolta JM-102 and from the nose with the Ingram icterometer. The results showed a
significant positive linear correlation between serum bilirubin values and the readings from both
the Minolta JM-102 (r = 0.83) and the Ingram icterometer (r = 0.78). The diagnostic accuracy of
the instruments was also assessed in predicting serum bilirubin > 220 micromol/litre. The
Minolta JM-102 showed a sensitivity of 100% with 56% specificity, while the icterometer had
the same value for sensitivity but with 48% specificity. [EL II]
65
In the third study, from the USA, varying degrees of jaundice were evaluated using the Gosset
icterometer on 90 preterm babies in a hospital setting, and compared with serum bilirubin
values obtained within 30 minutes of the icterometer reading. The instrument was used by three
observers – two with experience in its use and one with no experience. The mean birthweight
of the sample population was 1676 g and the mean gestational age was 31.7 weeks; the sample
was predominately white (95%). The linear correlation between the serum bilirubin levels and
icterometer readings by the two experienced observers was moderately positive (r = 0.71 and
r = 0.75, respectively; P < 0.001), while for the inexperienced observer the correlation
coefficient was 0.63. [EL II]
Two studies 63;64 with EL III measured the correlation of icterometer readings with serum bilirubin
64
values in black newborn babies. In the first study from Tanzania, icterometer gradings were
recorded in 70 babies (gestational age 30–42 weeks) with jaundice who were admitted to the
neonatal unit. No exclusion criteria were defined. Icterometer grading was done by blanching
the gum, and at the same time venous blood was drawn for serum bilirubin estimation. Results
showed a significant positive correlation (r = 0.91; P < 0.001) between the icterometer readings
63
and serum bilirubin levels. The second study, from Rhodesia, investigated the usefulness of the
icterometer as a screening test in 55 babies with jaundice. The birthweight of the study sample
ranged from 1050 to 3925 g, and age at testing varied from 2 to 24 days. Icterometer gradings
were done by a single person who was unaware of the serum bilirubin levels. The results
showed a highly significant positive linear correlation between the icterometer gradings and
serum bilirubin levels, with a correlation coefficient of 0.96 (P < 0.001).
Evidence summary

Results on the diagnostic accuracy of icterometers in term babies from two studies with EL II
were variable. While one study reported a correlation of 0.48 with 75% sensitivity and 72%
specificity in detecting serum bilirubin levels > 205 micromol/litre, the other study showed
correlation of 0.78 and 100% sensitivity with 48% specificity in detecting high serum bilirubin
levels (> 220 micromol/litre). In preterm babies a value of ‘r’ was reported as 0.71 and 0.75 by
two experienced observers. There is lack of good-quality evidence in babies with dark skin
tones. Results from the two studies with EL III indicate high correlation between icterometer
gradings and serum bilirubin values.

GDG translation from evidence
An icterometer can be used to confirm the clinical suspicion of jaundice in term babies but it
does not provide a reliable measure of severity. For preterm babies, good-quality evidence
shows a moderately positive association with serum bilirubin levels. Findings from poor-quality
studies suggest that icterometer readings in babies with dark skin tones correlate well with
serum bilirubin levels, but the GDG opinion is that better-quality evidence is needed before
icterometer use can be recommended in either preterm babies or babies with dark skin tones.
Overall, the GDG concluded that icterometers should not be used.

76