Transcutaneous Bilirubinometry in Healthy Term and Near-Term Neonates

Introduction/objectives: As the bilirubin concentration in the serum increases, it is deposited in the skin and subcutaneous tissues, producing the physical sign of jaundice.1 There is a well-established relationship between the serum total bilirubin (STB) concentration and the intensity of jaundice in neonates.2 But variation in color perception by the human eye, differences in neonatal skin pigmentation, and variation in both the intensity and color of the available light affect the ability of an individual to estimate the STB by clinical assessment alone.3-10 Transcutaneous bilirubin (TcB) measured by spectrophotometry has positive and a linear correlation with STB, and therefore, is a useful screening tool to measure clinically significant jaundice.11-14 Newer devices such as BiliCheck® can even correct for the effect of skin pigmentation on TcB measurement and therefore are reported to have closer agreement with STB.15-17 American Academy of Pediatrics in its guidelines on management of hyperbilirubinemia in term and near-term neonates, cautions against relying on clinical assessment of severity of jaundice.18 Therefore, blood sampling for STB measurement is needed in a significant proportion of healthy clinically jaundiced neonates. As compared to clinical assessment TcB is thought to be a more accurate and objective estimate of STB, its use may decrease the need for blood sampling to measure STB. Observational and retrospective analyses have estimated that routine use of Tc bilirubinometry for jaundice assessment results in 34% to 80% reduction in need for blood sampling for STB measurement.19-21 Objective of this randomized clinical trial was to determine if routine use of transcutaneous bilirubinometry decrease the need for blood sampling for confirmation of STB in healthy term and near term neonates? In addition, we also assessed the correlation and agreement of TcB and CaB with measured STB.

Methodology: Study was conducted as a randomized controlled trial in a tertiary care hospital. Neonates born at 35 or more completed weeks of gestation were eligible for enrolment if they developed clinical jaundice from 25 to 168 h of life. Neonates with Rh hemolytic disease, those requiring NICU admission for more than 24 h, having major congenital malformation or having received phototherapy were excluded. Study was carried out from November 1, 2006 to July 7, 2007 except in month of March 2007 when the primary investigator was not posted in clinical service area.

A log book was maintained in delivery areas of the hospital to identify neonates born at 35 or more weeks of gestation. Gestation was assigned based on 1st trimester ultrasound. In the event of non-availability of ultrasound report, gestation was assigned by date of last menstrual period and confirmed by Expanded New Ballard Score22 within 24 h of age. After receiving written informed consent from parents, the identified neonates were assessed every 8-12 h for clinical jaundice. On identification of clinical jaundice, the primary investigator assigned a clinically assessed level of bilirubin (CaB). This assessment was performed in an adequately illuminated room. Skin was blanched by digital pressure, revealing underlying color of skin and subcutaneous tissue in five dermal zones (as described by Kramer). Both extent and intensity of yellowish discoloration was used to assign the CaB. Each clinical assessment was immediately followed by transcutaneous measurement of bilirubin (TcB). TcB was average of five sequential measurements at forehead obtained using BiliCheck® (SpectRx Inc, Norcross, GA). Clinical assessment always preceded the transcutaneous measurement to prevent primary investigator getting biased by the TcB value.

We conducted a pilot study in our unit and found out that 15% neonates (≥35 wk gestation) undergo blood sampling for measurement of STB in first week of life (unpublished data). Sample size for a relative reduction of 40% in need for blood sampling (absolute reduction (from 15% to 9%), with a power of 80% and two-sided significance of 5% was 492 assessments in each group.

By random allocation method, either CaB or TcB was used for deciding the need for blood sampling to measure STB. Need for blood sampling was defined to be present if the bilirubin assessed by the allocated method exceeded 80% of age-specific cut-off for phototherapy as per American Academy of Pediatrics 2004 guidelines.7 STB was measured using twin-beam spectrophometry (Gineveri, Italy). Neonates discharged before completing 72 h of age were brought to hospital for assessment of jaundice 24-72 h after discharge as per AAP recommendations. Up to three assessments conducted at least 12 h apart in a neonate could be included in the study. Such repetitive assessments were independently randomized irrespective of previous assignment. Randomized sequence was generated in fixed block size of six each using a web-based random number generator.The random codes were kept in serially numbered, opaque and sealed identical envelops. The envelope was opened after obtaining both CaB and TcB values. Chief investigator was unaware of the process of randomization.

We were aware of the possibility of erroneously missing significant jaundice if baby did not qualify for sampling by the randomized method of assessment. To safeguard against such an occurrence, clinical team had an authority to take an over-riding decision to measure STB. Incidence of such 'overriding decisions' and their outcome in terms of need for phototherapy were recorded.

We analyzed relationship between two methods of assessment of bilirubin: transcutaneous bilirubinometry (TcB) and clinical assessment of bilirubin (CaB) with serum total bilirubin (STB). This relationship is described as correlation and agreement of the two with STB.

All base line and outcome data were recorded prospectively in a pre-tested proforma. The data were then entered in Epi infoTM version 3.3.2. The data were checked for completion, consistency and accuracy. For quality control ten percent of records were checked by an independent investigator on a random basis. Data were analyzed using Epi infoTM version 3.3.2. and Stata software version 9.1. Group characteristics were compared with χ2 test and two-sample t-test for discrete and continuous variables respectively. P value of <0.05 was taken as statistically significant. Analysis was intention-to-treat. Correlation of STB values to TcB and CaB was determined by Pearson correlation analysis. We determined the limits of agreement that could be applied to the whole population by the statistical analysis described by Bland and Altman.

The protocol was cleared by institutional Ethics Committee. Informed consent was taken from one of parents of enrolled infants.

Keywords: jaundice, neonates, transcutaneous bilirubinometry, clinical assessment of bilirubin