Page 110 - 16Neonatal Jaundice_compressed
P. 110

Recognition





                         Evidence summary – Minolta JM-103
                         The EL II evidence on diagnostic accuracy of the Minolta JM-103 shows variation in term babies.
                         The correlation coefficients between the transcutaneous bilirubin readings and serum bilirubin
                         levels ranged from a moderate positive 0.77 to a significantly positive 0.93. While three studies
                         found evidence of underestimation of serum bilirubin by the device in term babies, the fourth
                         study indicated overestimation that was much higher in babies with dark skin tones compared
                         with those with light skin tones. The AROC was 0.87 but the summary ROC curve did not show
                         a curvilinear pattern, indicating heterogeneity in the study results. The pooled sensitivity and
                         specificity were 85% and 80%, respectively, with strong evidence of statistical heterogeneity for
                         both results.
                         In preterm babies the correlation was positive, with values of 0.79 in one study and ranging
                         from 0.82 to 0.92 in the second study. The JM-103 consistently underestimated bilirubin levels
                         by a mean of 19 ± 32 micromol/litre.

                         GDG translation from evidence – Minolta JM-103
                         Forehead or sternum measurement of transcutaneous bilirubin using the Minolta JM-103 is more
                         accurate than visual assessment for the recognition of jaundice in term babies.
                         The evidence from two good-quality studies showed a positive correlation between the JM-103
                         and  serum  bilirubin  estimations.  The  GDG  was concerned that  the  JM-103  consistently
                         underestimated serum bilirubin by up to 50 micromol/litre.
                         Results from one study (EL II) showed that the reliability of the JM-103 was lower when used on
                         babies with dark skin tones when compared with those with light skin tones.
                         Recommendations

                         See the end of Section 5.2.
                         Description of included studies – BiliChek

                         Seven studies 44;45;77-81  have been included in this section – three 77-79  with EL Ib, two 44;80  with EL II
                                                                                                 78
                         and two 45;81  with EL III. The study population comprised term babies in one study,  term and
                                                                                               79
                         preterm babies in five  studies, 44;45;77;80;81  and preterm babies only in one study.  One study
                                                                                                          80
                         included African babies. It was not possible to combine the studies in a meta-analysis as there
                         were different study populations, different threshold  values of transcutaneous bilirubin for
                         calculating diagnostic accuracy, and different levels of laboratory serum bilirubin used as the
                         reference standard. Hence these studies have been described in a narrative manner.
                         Review findings – BiliChek

                                                                 77
                         The first study is a multicentre European study  conducted in six hospitals across five countries
                         – the UK, France, Germany, Italy and Switzerland. A total of 210 term and preterm babies (35
                         from each centre) who underwent serum bilirubin measurements as part of normal care at these
                         hospitals were recruited as the sample population. White babies made up 66% of the sample
                         population and about 20% had a gestational age of 36 weeks or less. A single transcutaneous
                         bilirubin measurement was made from the forehead and sternum of each baby using BiliChek
                         30 minutes before or after  blood was drawn. The laboratory estimation of serum  bilirubin in
                         each of the participating centres was done using the laboratory’s routine equipment. A portion
                         of the blood sample  was also sent to a central laboratory for bilirubin assay using HPLC-B
                         without disclosure of the hospital laboratory results. There was a significant correlation between
                         the forehead and sternum  transcutaneous bilirubin  measurements and laboratory serum
                         bilirubin levels (r = 0.87 and 0.85, respectively; P < 0.001 for both). The correlation between
                         the laboratory serum bilirubin and HPLC-B levels  was also  statistically  significant (r = 0.93;
                         P < 0.001).

                         The difference between the mean serum bilirubin values and the mean transcutaneous bilirubin
                         measurements was not statistically significant at either the forehead (MD = 2.4 micromol/litre,
                         95% CI −2.4 to 7.1 micromol/litre) or the sternum (MD = −14.8 micromol/litre, 95% CI −19.9
                         to 9.5 micromol/litre).  An  AROC curve  was plotted to calculate the diagnostic  accuracy of


                                                                                                         81
   105   106   107   108   109   110   111   112   113   114   115