Page 97 - 16Neonatal Jaundice_compressed

Neonatal jaundice

Recommendations
See the end of Section 4.2.

Effectiveness of DAT
Description of included studies

For the effectiveness analysis, two studies 50;51 comparing selective versus universal DAT and a
49
third study comparing readmission rates and phototherapy rates for tested and untested babies
were included.
Review findings

49
A retrospective observational study from the USA studied the effectiveness of DAT testing in a
sample of births within a 1 year period (January to December 2000). Mean gestational age,
mean birthweight and gender were not specified; 46% of babies studied were Asian and 36.8%
were white. Cord blood DAT was performed on 2443 babies of mothers with blood group O or
Rhesus negative while 2097 babies of mothers with groups A, B, AB or Rhesus positive were
not tested. The records of all DAT-positive babies were reviewed for information relating to the
presence of jaundice and serum bilirubin results if measured in the first 24–48 hours. DAT was
positive in 193 (7.9%) of tested babies. Phototherapy was used in 36 (18.6%) of DAT-positive
babies. Data for use of phototherapy in DAT-negative babies were not provided. Readmission
for phototherapy was needed for 26 (1.1%) of all DAT-tested babies, and for 19 (0.9%) of
untested babies. This difference was not statistically significant (OR 1.17, 95% CI 0.65 to 2.13).
[EL III]

51
A cohort study from a tertiary centre in the USA compared universal with selective newborn
cord blood testing. In the retrospective cohort group, all cord blood specimens received by the
blood bank in 1989 were tested while in the prospective cohort group selective testing (all
babies in intensive care, babies with clinical jaundice, babies of Rhesus-negative mothers and/or
positive maternal antibody screening, maternal blood group unknown) was carried out on
admissions between July 1990 and June 1991. Of the retrospective cohort, 2253 of 4003
eligible babies (56.3%) were tested. Of the prospective cohort, 1048 of 4498 babies (23.3%)
were tested selectively. Cord blood collection difficulties and specimen handling problems were
given as reasons for the 1750 missing test results in the retrospective sample. Fifteen babies
were readmitted for hyperbilirubinaemia in both study periods. The prevalence of DAT positive
tests was not specified. The rate of readmission for hyperbilirubinaemia was 0.4% (15 of 4003)
among universally tested babies and 0.3% (15 of 4498) among selectively tested babies. This
difference was not statistically significant (OR 1.12, 95% CI 0.56 to 2.30). [EL III]
50
A third study from the USA also examined the effectiveness of universal versus selective DAT
testing. A retrospective analysis of all records for 1990 and 1991 was carried out to identify
babies of group O, Rhesus-positive mothers. Altogether, 301 babies with a mean gestational age
of 39.4 weeks and mean birthweight of 3343.6 g were included; 50.5% were male, 44.5% were
white and 16.3% were black. Of 113 babies tested, 29 (26%) were ABO incompatible and 14
(12%) were DAT positive. A total of 188 babies were not tested routinely. Of these, 34 (18%)
had DAT tests requested by their treating doctor; 18 (9.6% were ABO incompatible and 13
(6.9%) were DAT positive. The overall prevalence of DAT positivity was 9.0% (27 of 301
babies). Phototherapy was used in four of 113 universally tested babies (3.5%) and eight of 188
selectively tested babies (4.3%). The OR was 0.83 (95% CI: 0.24 to 2.81). The rate of
readmission for phototherapy was 1.8% (two of 113) among universally tested babies and 0.5%
(one of 188). Again, this difference was not statistically significant (OR 1.12, 95% CI 0.56 to
2.30). [EL III]
Evidence summary

Three EL III studies using undefined criteria for readmission for hyperbilirubinaemia were
included. Two studies compared universal versus selective DAT testing and one compared
DAT-tested and DAT-untested cohorts. No statistically significant difference was found in the
readmission rates or phototherapy rates between those undergoing universal testing and those

68