OBJECTIVES: To compare transcutaneous bilirubin (TcB) readings with total serum bilirubin (TSB) after phototherapy, estimating the range of TcB where confirmation through blood sampling can be avoided.
METHODS: Preterm and term neonates receiving in-hospital phototherapy underwent TcB measurements (device JM-103, TcB) alongside routine TSB before and after treatment. We calculated time-dependent safety margins for transcutaneous readings to correctly assign 99% of infants not to receive phototherapy.
RESULTS: Between August 2011 and December 2012, 86 newborn infants (47 preterm, 39 term) underwent a total of 189 parallel measurements. Mean difference (TcB − TSB) before treatment was −0.6 mg/dL (SD, 1.9 mg/dL). Within the first 8 hours after phototherapy, TcB levels were −2.4 mg/dL (SD, 2.1 mg/dL) below TSB. Thereafter the difference gradually returned to pretreatment values (−1.8 mg/dL in 8–16 hours, −1.1 mg/dL in 16–24 hours, and −0.8 mg/dL after 24 hours), while variations remained stable over time (SD, 1.4–1.8 mg/dL). In the first 8 hours after treatment, TcB levels of −7.3 mg/dL below the individual phototherapy threshold allowed safe rejection of confirmatory blood sampling. After 8 hours, that safety margin was reduced to approximately −5.0 mg/dL.
CONCLUSIONS: TcB measurements remain a valuable tool after phototherapy when time-dependent underestimation of TcB is being accounted for.
- TcB —
- transcutaneous bilirubin
- TSB —
- total serum bilirubin
What’s Known on This Subject:
Phototherapy decreases bilirubin concentration in skin more rapidly than in blood. During and after phototherapy, transcutaneous bilirubin measurements are considered unreliable and therefore discouraged.
What This Study Adds:
Transcutaneous bilirubin underestimates total serum bilirubin by 2.4 mg/dL (SD, 2.1 mg/dL) during the first 8 hours after phototherapy. This gives a safety margin of ∼7 mg/dL below the treatment threshold to omit confirmatory blood sampling.
In neonatal jaundice, as commonly seen in term and preterm newborns alike, gauging the severity of hyperbilirubinemia and guiding treatment decisions is based on concentration of bilirubin in the blood. The current gold standard is determination of bilirubin in serum samples (total serum bilirubin [TSB]), for example, through high performance liquid chromatography.1 Obtaining venous or heel stick blood samples is an invasive and painful procedure. Transcutaneous bilirubin (TcB) estimation is being evaluated as an easy, time-saving, and painless alternative,2,3 and transcutaneous bilirubinometry has been established as an essential part of a stepwise screening process to decrease the need for TSB.4–10
Before phototherapy, there is a good association between bilirubin concentrations in the blood and the skin, as assessed through TcB optical spectroscopy.11 With the initiation of phototherapy, a rapid decrement in dermal bilirubin is caused by photoisomerization of albumin-bound bilirubin in interstitial places and subcutaneous capillaries into lumirubin and other photoisomers.12 Two theoretical kinetics for post-phototherapy TcB have been contrasted recently (plateau versus rebound).13 In short, TcB levels were found significantly lower than TSB 6 hours after cessation of light exposure,14 but TcB may resume pre-exposure validity 18 to 24 hours after phototherapy.15 Because of this delayed equalization of bilirubin between blood and skin tissue, the validity of TcB has been questioned during and after phototherapy.16 On the other hand, serum bilirubin levels decline with a lagging kinetic, and were found to rebound after cessation of light.17 The specific timing, however, when TcB determinations become reliable again after phototherapy is unknown. All children receiving phototherapy need monitoring of bilirubin to guide continuation of treatment.
Currently, TcB assessment is used before initiating phototherapy only. Within this prospective observational study of term and preterm neonates, we assessed TcB concurrently with TSB measurements in venous and capillary samples drawn at various time points after phototherapy. We aimed to define the range of post-phototherapy TcB levels considered safe to replace blood sampling.
This was a prospective observational study conducted from August 2011 to December 2012 at the Department of Neonatology, Charité University Medical Center Berlin, Germany. The institutional review board (ethics committee of Charité University Medical Center Berlin) approved this study.
Term and preterm neonates who had a birth weight ≥1500 g who were cared for in the outpatient clinic or being admitted to the NICU after delivery and requiring phototherapy were eligible to participate in the study. Infants who had previous phototherapy or exchange transfusion were excluded. Written informed consent was obtained from parents or legal guardians who were only approached when showing a sufficient level of German language skills to understand the study information.
The need for phototherapy was based on the 2010 edition of the guideline of the German-speaking Society of Neonatology and Pediatric Intensive Care, as published by the German Association of Medical Scientific Societies.18 In short, the baseline phototherapy threshold is set at 20 mg/dL. For corrected gestational age below 38 weeks, the threshold was calculated as gestational age in weeks minus 20 (mg/dL). The threshold was further reduced by 2 mg/dL for a positive antiglobulin test and for each day below age 3 days. All neonates received 1 or more complete cycles of intermittent phototherapy, consisting of 3 consecutive 4-hour treatment/4-hour pause periods, giving a total light exposure of 12 hours. The need for an additional cycle was based on serum bilirubin levels, by using the decision criteria described above.
Transcutaneous bilirubinometry was performed by trained nurses in ambient daylight by using the JM-103 transcutaneous bilirubinometer (Drägerwerk, Lübeck, Germany), alongside serum-based measurements (TSB) before and after phototherapy, with a maximum interval of 10 minutes. Time points for measurement were scheduled according to clinical appraisal. Raw values from both methods were immediately recorded on a specific study form in the infant’s file. TcB measurements were taken on the sternum and recorded as a device-calculated mean of 3 independent determinations. TSB was determined by capillary blood gas analysis (ABL800, Radiometer Medical, Copenhagen, Denmark). When that was not available, laboratory measures in blood serum within 15 to 20 minutes after drawing were used instead (DPD reagent; Amresco Inc, Solon, OH). Background data on infant and mother as well as exact timing of phototherapy were retrieved from clinical charts.
Comparison of bilirubin levels before phototherapy was based on the last simultaneous pair of TcB and TSB before treatment. Measures after cessation of light exposure were categorized in groups of 0–8, 8–16, 16–24, and more than 24 hours after phototherapy. Only TcB/TSB pairs with TSB values below the upper detection limit of the TcB method of 20 mg/dL (device specific) were used for comparative illustration and calculations, as the device does not discriminate between data points above this threshold.
Gestational age was reported in 3 groups: early (<35 weeks), late-early (35 to <37 weeks), and term (≥37 weeks). Caucasian origin included both parents Caucasian as well as 1 parent Caucasian and the other parent’s origin unknown.
Data cleaning and evaluation was performed by using SAS 9.3 (SAS Institute, Inc, Cary, NC). Only cases with information on date of birth, start and duration of phototherapy, and both TSB and TcB measurement before phototherapy were included in this analysis. Only complete bilirubin TSB/TcB pairs after cessation of phototherapy with a valid time stamp were accounted for. Capillary blood gas-based TSB was preferred, and substituted by a serum-based measure where missing.
Guiding visual appraisal of the difference between detection methods, linear regression was used for values before treatment (versus mean levels, Bland-Altman plot19). After phototherapy, commonly used parametric models (eg, linear, exponential) were assessed, but only penalized B-spline curves yielded appropriate fit (3rd degree, using PROC SGPLOT/PBSPLINE). Ninety-five-percent confidence intervals were calculated and plotted. The interval mean ± 1.96 SD covers 95% of sample values. To assure that at least 99% of infants who had TcB values below the phototherapy threshold actually represent cases with no need for phototherapy when confirmed through TSB, safety margins were calculated as the mean difference between both methods minus 2.33 SD.
Baseline Patient Characteristics
A total of 86 neonates fulfilled inclusion criteria and the minimum data required were available. The study population consisted mainly of healthy neonates who had a gestational age of 35 or more weeks (62/86, 72.1%), of spontaneous delivery (52/86, 60.5%). Most children were entirely or partly breastfed (84/86, 97.7%); Coombs test was positive in 9/86 (10.5%, Table 1).
Inherent to the device, TcB cannot fall above the upper detection limit of 20 mg/dL. In term Coombs-negative neonates who had a phototherapy threshold at (or close to) 20 mg/dL, TcB was expectedly lower than TSB in almost all cases (arrows pointing upward, Supplemental Fig 3). Far from TcB detection limit, neonates who had TcB levels close to the individual phototherapy threshold corrected upward in TSB were selected to receive phototherapy more often than those who had a downward correction (more cases in upper left quadrant than in lower right, Supplemental Fig 4). In our sample of neonates who had a given phototherapy indication, both aspects led to higher average TSB than TcB already, before phototherapy.
Mean TSB before phototherapy was 16.2 mg/dL, and lower in early (14.5 mg/dL) and late-early preterm (16.6 mg/dL) than in term infants (17.6 mg/dL). TcB readings were on average 0.6 mg/dL (SD, 1.9 mg/dL) lower than corresponding TSB values. The mean difference and SD displayed little variation with the average magnitude (calculated as the mean of the 2 methods), as shown by the horizontal slope of a linear fit of bilirubin difference against its magnitude (Fig 1). Variation of the difference between TcB and TSB was highest in term (SD, 2.6 mg/dL) compared with early and late-early preterm infants.
At the start of phototherapy, infants were between 1 and 8 days old, with a mean (SD) of 4.2 (1.6) days. A single cycle of intermittent phototherapy was sufficient in 50 children, 36 underwent 2 or more cycles, and none received intravenous immunoglobulin treatment or exchange transfusions.
Comparison of Techniques After Phototherapy
Early after phototherapy, the difference between TcB and TSB was larger than before light exposure (Table 2). Locally smoothed approximation (penalized B-spline) of bilirubin level difference returned to values seen before treatment in the intervals usually covering the second follow-up assessment (8–16/16–24 h). Thereafter, the difference between detection methods remained relatively stable over time (Fig 2). The mean difference between measurement techniques was not dependent on its magnitude (Supplemental Fig 5).
Potentially Avoidable Blood Sampling
Based on mean difference and SD in each post-phototherapy interval, we calculated safety margins to assure <1% false-negative TcB readings. Within the first 8 hours after phototherapy, the safety margin was −7.3 mg/dL, and close to 5 mg/dL thereafter. Applying these safety margins retrospectively to our sample, 82 of 189 (43%) post-phototherapy TcB assessments were below this safety margin, and venous or capillary puncture could have been omitted (Table 2). A total of 45 of 189 (23.8%) measurement pairs used TSB from the laboratory assay instead of blood gas analysis. Restricted evaluation of these cases gave similar outcomes (−7.0 mg/dL within 8 hours).
TcB Levels After Phototherapy
Within the first 8 hours after cessation of phototherapy, TcB underestimated TSB by −2.4 mg/dL. Thereafter, the difference between both methods returned to pretreatment levels. On a biochemical level, this is readily explained by skin bilirubin levels being decreased by phototherapy and subsequent flow of bilirubin from blood to skin.17 Our study presents a rather predictable kinetic for this phenomenon, given that the variations of TcB-TSB difference were similar before and in the various post-phototherapy intervals. After all, TSB and TcB before, during, and after phototherapy are basically different but strongly related variables.13
Other studies that investigated the relation of post-phototherapy bilirubin detection methods reported only descriptive group means or measures of association (correlation) but failed to derive applicable estimates from their data.14–16,20 Answering the need for measures applicable in clinical decision-making, we calculated safety margins to assure that <1% of TcB-screened neonates will be falsely labeled negative. Up to 8 hours after finishing phototherapy, we consider TcB measures safe below levels of the individual therapy threshold −7.3 mg/dL. Later, the safety margin can be set to −5.1 mg/dL. To our knowledge, comparable numbers have not been published. This adds a noninvasive alternative to TcB measurement in covered skin areas.20
To estimate the impact on clinical procedures (ie, venous or capillary punctures) in post-therapy assessment of neonates, we applied these safety margins retrospectively to our study sample. By using a 2-step approach, calling for an initial TcB measure followed by TSB confirmation only when TcB values were above the safety margins, we could have omitted 43% of post-phototherapy punctures. Even for the first assessment after therapy, an interval usually considered most uncertain, an equally high proportion of invasive blood sampling could have been omitted. We suggest adapting current clinical guidelines, incorporating post-phototherapy TcB screening with safety margins in neonates. Before defining a definite generic approach, proposed safety margins should be externally validated in other settings. In a similar way, we recommend estimating safety margins for pre-phototherapy TcB.
Strengths and Weaknesses
Despite calibrated instruments, pre-phototherapy TcB levels were on average −0.6 mg/dL lower than TSB levels. Samples were taken during the normal rise of TSB over the first days of life, and equilibration between blood and skin tissue may lag. When TcB before phototherapy was lower than TSB (upward correction), neonates were more likely to receive treatment compared with those who had a downward correction to a similar TSB level (Supplemental Fig 4). This could be attributable to physician’s interpretation of (random) variation between detection methods as an indicator of increasing or decreasing bilirubin over time, another explanation for higher average TSB than TcB levels before phototherapy. In consecutive studies, we suggest including a pre-indication sample allowing the quantification of this phenomenon.
We did not assess the TcB-TSB difference between phototherapy sessions for neonates receiving more than 1 cycle. In this interval, bilirubin levels are presumably higher and thus closer to the individual phototherapy threshold than those assessed after the final cycle. The true proportion of potentially saved punctures is probably smaller than what we estimated from our sample of neonates who had no further phototherapy requirement. We are aware that the proportion of saved punctures in other settings will depend on criteria applied to select neonates for bilirubin assessment (ie, clinical judgment of accompanying symptoms). Bilirubin measurements were not related to levels before or during phototherapy and were analyzed as individual cross-sectional data pairs.
The pragmatic approach, sampling neonates from various sources, including preterm and term, and relying mainly on non-study personnel, led to figures readily applicable in a wide spectrum of clinical settings.
Although TcB systematically underestimates TSB levels after phototherapy, especially during the first 8 hours, it can still be used to reduce the number of blood samples with appropriate safety margins.
The JM-103 transcutaneous bilirubinometer used in this study was loaned by Draeger Air-Shields. We thank Petra Blank and Regina Nagel (secretaries), and the nursing staff and physicians at our NICU for their helpful cooperation in this study. Jana Grabenhenrich and Linus Grabenhenrich contributed equally to this work. Above all, thanks to all families for their help in realizing the study.
- Accepted August 25, 2014.
- Address correspondence to Jana Grabenhenrich, MD, Klinik für Neonatologie, Charité Universitätsmedizin Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany. E-mail:
Dr J. Grabenhenrich recruited participating families, collected the original data, cleaned the data, and wrote and revised the manuscript; Dr L. Grabenhenrich carried out statistical analyses, plotted graphs, and wrote and revised the manuscript; Dr Bührer conceived the study and revised the manuscript; Dr Berns supervised recruitment and data collection and revised the manuscript; and all authors approved the manuscript.
FINANCIAL DISCLOSURE: The hospital was given transcutaneous bilirubinometers for onsite testing by Drägerwerk, Lübeck, Germany. No money was ever paid.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Maisels MJ,
- Kring E
- Keren R,
- Luan X,
- Friedman S,
- Saddlemire S,
- Cnaan A,
- Bhutani VK
- Keren R,
- Bhutani VK,
- Luan X,
- Nihtianova S,
- Cnaan A,
- Schwartz JS
- American Academy of Pediatrics Subcommittee on Hyperbilirubinemia
- Maisels MJ,
- Ostrea EM Jr,
- Touch S,
- et al
- Nagar G,
- Vandermeer B,
- Campbell S,
- Kumar M
- Rubaltelli FF,
- Gourley GR,
- Loskamp N,
- et al
- De Luca D,
- Engle W,
- Jackson G
- ↵Hyperbilirubinämie des Neugeborenen - Diagnostik und Therapie. Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften e.V. (AWMF); 2010. Available at: www.awmf.org/uploads/tx_szleitlinien/024-007l_S2k_Hyperbilirubinaemie_des_Neugeborenen.pdf. Accessed September 1, 2014
- American Academy of Pediatrics. Provisional Committee for Quality Improvement and Subcommittee on Hyperbilirubinemia
- Bhutani VK,
- Johnson L,
- Sivieri EM
- Copyright © 2014 by the American Academy of Pediatrics