Page 94 - 16Neonatal Jaundice_compressed
P. 94

Early prediction of serious hyperbilirubinaemia





                         GDG translation from evidence
                         Current evidence does not support measuring umbilical cord  blood  bilirubin levels for the
                         prediction of subsequent hyperbilirubinaemia in healthy babies.

                         Recommendations

                         See the end of Section 4.2.
                         End-tidal carbon monoxide measurement (ETCOc)

                         Description of included studies
                                                                                               32
                         Two studies 32;33   with  EL II  have been included in this section. The first study   had a large
                                                                                             33
                         sample size of both term and near-term babies while the second, smaller, study  included only
                         term babies. In both studies, ROC curves were developed to evaluate the accuracy of end-tidal
                         carbon monoxide (CO)  production corrected for ambient CO (ETCOc) in predicting
                         hyperbilirubinaemia.

                         Review findings
                                                              32
                         The first study  was an international study  carried out at nine sites (four in the  USA, two in
                         China, two in Israel and one in Japan). All newborn babies with gestational age ≥ 35 weeks
                         were enrolled in the first 36 hours of life. Of the 1895 babies enrolled, 1370 (72%) completed
                         the study.  All babies had measurements of  ETCOc and serum bilirubin performed at
                         30 ± 6 hours, and serum bilirubin only at 96 ± 12 hours. Between these times, serum bilirubin
                         could be measured for clinical reasons. ETCOc  was  measured using a breath analyser with
                         single-use disposable nasal sampler. Hyperbilirubinaemia was defined as laboratory serum
                         bilirubin  ≥ 95th centile at any time during the study period. Threshold centile  values  were
                                                            34
                                                                                    19
                         taken as those defined by Bhutani et al.  and adopted by the AAP.  Inclusion and exclusion
                         criteria were well defined.  Babies with age-specific serum bilirubin  ≥ 95th centile at up to
                         96 ± 12 hours were withdrawn from the study. About 9% (120 of 1370) of babies had serum
                         bilirubin levels ≥ 95th centile at 30 ± 6 hours or at 96 ± 12 hours. The mean ETCOc levels in
                         this  group were  statistically  significantly higher  than the mean  levels in the non-
                         hyperbilirubinaemic group (P < 0.001). Logistic regression analysis was conducted with
                         variables  found  to  be  associated  with  hyperbilirubinaemia  (serum  bilirubin  percentile  at
                         30 hours,  bruising,  maternal  blood  type,  race,  maternal  diabetes,  feeding  type,  gravidity  and
                         ETCOc). Models to evaluate diagnostic accuracy of ETCOc, laboratory serum bilirubin and their
                         combination in predicting hyperbilirubinaemia were developed. ETCOc at 30 ± 6 hours with a
                         threshold value  above the  population  mean  (1.48 ± 0.49 ppm)  predicted  hyperbilirubinaemia
                         with 13% positive predictive value (PPV) and 96%  negative predictive value (NPV), while
                         laboratory serum bilirubin levels > 75th centile showed 17% PPV and 98% NPV. When both
                         tests were combined, NPV increased to 99% but PPV decreased to only 6%. It was concluded
                         that  serum  bilirubin  measurement  before  discharge  (at  30 ± 6 hours)  may  provide  some
                         assistance in predicting risk of hyperbilirubinaemia,  but the addition of ETCOc does not
                         improve its predictive accuracy. [EL II]
                                                      33
                         In the second study, from Japan,  ETCOc levels were measured every 6 hours during the first
                         3 days  of  life  in  51  healthy,  full-term  babies.  The  Minolta  JM-102  was  used to  record
                         transcutaneous bilirubin  measurements every  12 hours  during  the first  5 days  and  serum
                         bilirubin levels were measured if the JM-102 index was ≥ 22 reflectance units. An ROC curve
                         was developed to evaluate the accuracy of ETCOc at different ages in predicting
                         hyperbilirubinaemia,  which  was  defined  as  serum bilirubin  ≥ 257 micromol/litre.
                         Hyperbilirubinaemia occurred in  seven  babies,  while 44 babies had serum bilirubin levels
                         < 257 micromol/litre.  There  were  no  statistically  significant  differences  between  the
                         hyperbilirubinaemic and non-hyperbilirubinaemic babies in terms of sex, gestational age, mode
                         of delivery, Apgar score at 1 minute, age at peak transcutaneous bilirubin, or mode of feeding.
                         Moreover, the mean levels of ETCOc were similar for the two groups from 6 to 36 hours of age,
                         but the hyperbilirubinaemic group had higher mean levels at 42, 48, 54 and 66 hours. The ROC
                         curve indicated that  ETCOc at 42 hours of  age showed the best  accuracy in predicting




                                                                                                         65
   89   90   91   92   93   94   95   96   97   98   99