656 - Diagnostic performance of visual assessment of bilirubin and transcutaneous bilirubinometry in detecting rebound hyperbilirubinemia requiring phototherapy re-initiation in neonates

Publication Number: 656.4500

Priyanka Singh, All India Institute of Medical Sciences, New Delhi, India, New Delhi, Delhi, India; Satya Prakash, All India Institute of Medical Sciences, New Delhi, New Delhi, Delhi, India; Ankit Verma, AIIMS, Delhi, Delhi, India; Anu Sachdeva, All India Institute of Medical Sciences, New Delhi, Delhi, India; Ramesh Agarwal, All India Institute of Medical Sciences, New Delhi, New Delhi, Delhi, India; M Jeeva Sankar, All India Institute of Medical Sciences, New Delhi, New Delhi, Delhi, India

Background: The American Academy of Pediatrics recommends total serum bilirubin (TSB) to assess rebound hyperbilirubinemia (RH) after stopping phototherapy in neonates. Visual assessment (VA) of jaundice and transcutaneous bilirubinometry (TcB) can reduce the need for blood sampling for RH.

Objective: To evaluate the diagnostic performance of (a) VA for jaundice and (b) TcB at 12 hours (for hemolytic jaundice) or 24 hours (for non-hemolytic jaundice) after stopping phototherapy, in detecting RH that requires phototherapy re-initiation in neonates born ≥35 weeks gestation.

Design/Methods: This diagnostic test evaluation study was conducted in a tertiary neonatal unit in New Delhi. Eligible neonates underwent VA using Modified Kramer’s method by a fellow with at least 3 years of training in pediatrics, followed by TcB (Dräger JM-105™, Germany) and were finally sampled for TSB by point-of-care spectrophotometry (One beam, Ginevri, Italy). The outcome included sensitivity, specificity, positive and negative likelihood ratios, the agreement of VA and TcB with TSB, positive and negative predictive values, and the proportion of neonates for whom sampling could be avoided using VA or TcB.

Results: 160 neonates were enrolled with a mean (SD) gestation and birth weight of 36 ± 3 weeks and 2743 ± 483 g, respectively. Of these, 35 (21.9%) had hemolytic jaundice. Diagnostic analysis demonstrated high sensitivity for both VA and TcB (100%); however, VA had a higher specificity (88.2%) compared to TcB (71.7%). The positive predictive values were moderate, with VA at 30.8% and TcB at 15.7%. Both methods had high negative predictive value (100%). The positive likelihood ratio was higher for VA (8.4) than for TcB (3.5). The area under the receiver operating characteristic (ROC) curve was 0.94 for VA and 0.86 for TcB. Bland-Altman analysis showed a mean difference of -1.8 mg/dL (95% limits of agreement [LOA] -6.7 to 2.9 mg/dL) between VA and TSB, and a mean difference of 1.1 mg/dL (95% LOA -3.3 to 5.5 mg/dL) between TcB and TSB. RH was observed in 8 (5%) newborns.VA could have reduced the need for TSB sampling in 83.8% (95% CI 77.0 - 88.7), and TcB could have reduced it in 68% of cases (95% CI 60.4 - 70.9).

Conclusion(s): Given the excellent sensitivity and negative predictive values, VA and TcB may be used to screen for identifying rebound hyperbilirubinemia in neonates ≥35 weeks gestation. This may help reduce the need for blood sampling in these neonates.

Baseline characteristics of the study population
Table 1.pdf

Comparison of visual assessment and transcutaneous bilirubin with serum bilirubin using Bland-Altman plots
Plot shows agreement between visual (top) and transcutaneous (bottom) bilirubin methods with serum bilirubin. Green line = mean difference; red lines = 95% LOA.

Diagnostic accuracy measures of visual assessment and transcutaneous bilirubinometry
Table 2.pdf