Evaluation of the BiliChek Noninvasive Bilirubin Analyzer for Prediction of Serum Bilirubin and Risk of Hyperbilirubinemia

Transcription

1 Clinical Chemistry / Evaluation of Transcutaneous Bilirubin Evaluation of the BiliChek Noninvasive Bilirubin Analyzer for Prediction of Serum Bilirubin and Risk of Hyperbilirubinemia Brad S. Karon, MD, PhD, 1 Ann Teske, 2 Paula J. Santrach, MD, 1 and Walter J. Cook, MD 2 Key Words: Bilirubin; Serum bilirubin; Transcutaneous bilirubin DOI: /AJCPRX1E3NWCXHMZ Abstract We identified clinical and laboratory variables affecting the relationship between transcutaneous and serum bilirubin levels and determined whether transcutaneous bilirubin values could be used to predict the risk of hyperbilirubinemia. Median bias between transcutaneous and diazo serum bilirubin was 2.0 mg/dl (34.2 µmol/l), while median bias between transcutaneous and the Vitros (Ortho Clinical Diagnostics, Rochester, NY) serum bilirubin values was 1.3 mg/dl (22.2 µmol/l). The mother s ethnicity, the gestational age, and postnatal age did not impact the relationship between transcutaneous and serum bilirubin values. In contrast, the serum bilirubin method (diazo vs Vitros) and collection container (clear vs amber tube) significantly impacted the relationship between transcutaneous and serum bilirubin values. Transcutaneous bilirubin was a sensitive but not specific predictor of the risk of hyperbilirubinemia using a conventional risk nomogram. Because systematic differences between serum bilirubin methods and local laboratory practices impact the relationship between transcutaneous and serum bilirubin values, the effectiveness of transcutaneous prediction of the serum bilirubin risk zone will vary by institution. Jaundice is common in the neonatal period owing to the shortened lifespan of neonatal blood cells and other factors. 1-3 Extremely high levels of bilirubin are toxic to the brain and may cause an irreversible bilirubin-induced encephalopathy known as kernicterus. 4 To encourage increased awareness and detection of severe hyperbilirubinemia, the American Academy of Pediatrics (AAP) issued revised guidelines for the management of hyperbilirubinemia in term or near-term newborn infants in The revised guidelines call for transcutaneous (TcB) or serum (TsB) bilirubin measurement in all jaundiced infants in the first 24 hours of life and for infants with jaundice appearing excessive for the infant s age. To move toward more systematic assessment of jaundice, several studies have examined the usefulness of TcB monitors as a means to predict the TsB concentration. One study found that the BiliChek transcutaneous bilirubin monitor (Respironics, Marietta, GA) overestimated the TsB on average by 1.1 mg/dl (18.1 µmol/l). 6 In contrast, 4 other studies found that the BiliChek monitor slightly underestimated the TsB level across a wide range of bilirubin concentrations Two other studies found that BiliChek TcB slightly overestimated the TsB at lower serum bilirubin concentrations but underestimated the TsB at concentrations of more than 11.7 mg/dl (200.1 µmol/l). 11,12 It is not clear what factors explain the variability in the relationship between BiliChek TcB and TsB. Despite the AAP recommendations to interpret all bilirubin measurements with respect to an infant s postnatal age, 5 most studies of BiliChek accuracy have evaluated the ability of TcB to predict a TsB level above or below a fixed cutoff. 6,8-12 Nomograms based on TcB values have recently been 976 Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ

2 Clinical Chemistry / Original Article published, 13,14 but clinical experience with these nomograms is limited, and it is not clear whether nomograms based on TcB are specific to the device and infant population studied. In our practice, the optimal use of TcB in screening infants would be to use the TcB value and the infant s postnatal age to predict the serum bilirubin risk zone based on a conventional serum bilirubin nomogram currently in use. One study evaluated the accuracy of the TcB value and the infant s age in hours to predict the risk zone for hyperbilirubinemia compared with a matching serum value. BiliChek TcB, used with postnatal age, could predict a high risk (>95th percentile for age) TsB with 100% sensitivity and 66% specificity. 7 However, this study did not examine the sensitivity or specificity of BiliChek TcB for predicting high-intermediate (>75th percentile for age) TsB levels, which limits the usefulness of these findings for many practices. The goals of the present study were to identify clinical and laboratory variables that impact the relationship between TcB and TsB and to define the sensitivity and specificity of the BiliChek TcB for predicting high-intermediate (>75th percentile for age) and/or high (>95th percentile for age) TsB values in a population of term and near-term infants in a wellinfant nursery. Materials and Methods Patient Selection Infants admitted to the well-infant nursery at the Mayo Clinic (Rochester, MN) were eligible for prospective enrollment in the study if a participating physician ordered a serum bilirubin level to assess risk for hyperbilirubinemia. The study period was August 2006 through July Recruitment of infants was not consecutive as not all physicians practicing in the nursery were involved in the study. If parents consented to involvement in the study, a TcB measurement was obtained within 30 minutes of blood collection for TsB measurement. Only the first bilirubin measurement for any infant was used for purposes of the study, consistent with the study s intent to test the ability of the TcB to screen for risk of hyperbilirubinemia. Infant gestational age at birth, postnatal age at the time of transcutaneous measurement, and the mother s ethnicity were recorded for each patient participating. There were 146 Caucasian, 19 Asian, 9 Hispanic, and 3 African American infants included in the study. The study design was approved by the Mayo Clinic Institutional Review Board. TcB Measurement TcB was measured on the forehead by using 1 of 2 identical BiliChek devices. The BiliChek devices were calibrated with a disposable tip (BiliCal) before each measurement; the device displays the average of 5 readings. Nurses performed BiliChek measurements following a training session that described use of the device according to the manufacturer s instructions. TcB precision was assessed by repeated measurement on 40 infants, with an SD of 1.1 mg/dl (18.8 µmol/l) at an average TcB value of 12.0 mg/dl (205.2 µmol/l). TsB Measurement Serum samples were obtained by capillary puncture or venipuncture. For the study, 88 specimens were collected into plain (no additive) clear serum tubes (Terumo CapiJect, Somerset, NJ), and 89 serum samples were collected using an amber-colored gel barrier serum tube (Terumo CapiJect). The TsB is routinely measured using a modification of the diazo method. 15 The diazo method used was the Amaresco DPD reagent (Amaresco, Solon, OH) or the Roche Total Bilirubin reagent (Roche Diagnostics, Indianapolis, IN) run on a Roche Modular Analytics system. Correlation between the 2 diazo methods based on 100 samples covering the reportable range of the assay was as follows: Roche T Bili = Amaresco DPD 0.09 mg/dl (1.54 µmol/l), with an r 2 of Precision of the diazo method was assessed by replicate (n = 20) measurement of serum with a mean bilirubin value of 21.0 mg/dl (359.2 µmol/l), demonstrating an SD of 0.2 mg/dl (3.4 µmol/l) at this level. The Vitros method (Ortho Clinical Diagnostics, Rochester, NY) demonstrated similar precision. The level of hemolysis (serum free hemoglobin) in each sample was determined by converting the hemolysis index measured by the Roche Modular Analytics system into a free hemoglobin level as described previously. 16 When sample volume remained after routine serum bilirubin analysis (131 of 177 samples), samples were analyzed using the Vitros BuBc slide method on a Vitros 250 analyzer (Ortho Clinical Diagnostics). Unlike the diazo methods, which measure bilirubin by colorimetric reaction with 2,5-dichlorophenyl diazonium tetrafluoroborate dye, the Vitros BuBc slide uses a mordant to partially separate the spectra of unconjugated and conjugated bilirubin, allowing measurement of both species by reflectance spectrophotometry on a single slide. Statistical Analysis Median bias (TcB minus TsB) was calculated for the diazo and Vitros TsB data sets, along with 95% confidence interval (CI) of the median bias. Bias was assessed by testing the hypothesis that the slope of the regression of TsB on TcB was equal to 1 (indicating that the values were identical). A significant P value of less than.05 would mean that there was significant bias between the TcB and TsB values. Generalized estimating equations were also used to determine the impact of gestational age, postnatal age in hours, the mother s ethnicity (Caucasian vs non-caucasian), blood collection technique (capillary puncture vs venipuncture), hemolysis level, and collection container (clear vs amber) on the relationship between TcB and TsB values. Significance of any change in median Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ 977

3 Karon et al / Evaluation of Transcutaneous Bilirubin bias (ie, Caucasian vs non-caucasian, capillary puncture vs venipuncture) was defined as a P value less than.05. Because most data sets were not normally distributed, continuous data are summarized as median values, with interquartile range and minimum and maximum values observed. Bland-Altman plots were used to display the relationship of TcB to TsB across the range of bilirubin values observed. CIs for sensitivity and specificity were calculated by using the Fisher exact test. The clinical significance of differences between TcB and TsB was defined by risk level determination according to the nomogram produced by Bhutani et al, 17 which plots the bilirubin level as a function of postnatal age in hours, from 0 to 144 hours of life. For each TcB and TsB measurement, the risk zone (low, low-intermediate, high-intermediate, high) was determined by using the Web site bilitool.org, which allows the user to determine risk category according to Bhutani et al 17 by entering the bilirubin level and the infant s age in hours. Bilirubin levels exceeding the 75th percentile for age in hours are considered high-intermediate risk, and bilirubin levels exceeding the 95th percentile for age are considered high risk. Results Relationship Between TcB and TsB TcB consistently overestimated concentrations of TsB by the diazo and Vitros methods. The median TcB concentration was 12.2 mg/dl (208.6 µmol/l), while median diazo TsB bilirubin level was 10.1 mg/dl (172.7 µmol/l) ztable 1z. For the 131 samples analyzed by Vitros, the median Vitros TsB bilirubin level was 10.9 mg/dl (186.4 µmol/l; Table 1). The median bias between the TcB and diazo TsB was 2.0 mg/dl (34.2 µmol/l; 95% CI, mg/dl [ µmol/l]), and the median bias between the TcB and Vitros TsB was 1.3 mg/dl (22.2 µmol/l; 95% CI, mg/dl [ µmol/l]). Both bias measurements were statistically significant (P <.001). Correlation between methods was calculated as follows: y(tcb) = 0.89x (diazo TsB) mg/dl (56.4 µmol/l); r 2 = 0.65; and y(tcb) = 0.90x (Vitros TsB) mg/dl (42.8 µmol/l); r 2 = Bland-Altmann plots of TcB vs diazo TsB zfigure 1z and TcB vs Vitros TsB zfigure 2z are shown. There is a positive bias between TcB and TsB that is relatively constant over the range of bilirubin values measured (Figures 1 and 2). TcB Diazo TsB TcB and Diazo TsB Mean zfigure 1z Bland-Altman plot of differences between BiliChek transcutaneous (TcB) and laboratory serum bilirubin measurements by diazo method (diazo TsB) vs mean of TcB and diazo TsB methods. The best-fit line between bias (TcB minus TsB) and mean value (TcB + TsB/2) is shown. The best-fit line was given by y(bias) = 0.10x(mean value) mg/dl; r 2 = 0.02, indicating a very weak relationship between TcB bias and bilirubin value. TcB Vitros TsB TcB and Vitros TsB Mean zfigure 2z Bland-Altman plot of differences between BiliChek transcutaneous (TcB) and laboratory serum bilirubin measurements by the Vitros method (Vitros TsB) vs mean of TcB and Vitros TsB methods. The best-fit line between bias (TcB minus TsB) and mean value (TcB + TsB/2) is shown. The best-fit line was given by y(bias) = 0.12x(mean value) mg/dl; r 2 = 0.03, indicating a very weak relationship between TcB bias and bilirubin value. ztable 1z Median, Interquartile Range, Minimum, and Maximum Values for Transcutaneous Bilirubin, Serum Bilirubin (Diazo and Vitros Methods), Gestational Age, Postnatal Age, and Serum Free Hemoglobin Variable Median Interquartile Range Minimum Maximum Transcutaneous bilirubin, mg/dl (µmol/l) 12.2 (208.6) ( ) 6.4 (109.4) 18.9 (323.2) Serum bilirubin (diazo), mg/dl (µmol/l) 10.1 (172.7) ( ) 5.3 (90.6) 16.6 (283.9) Serum bilirubin (Vitros), mg/dl (µmol/l) 10.9 (186.4) ( ) 5.8 (99.2) 17.9 (306.1) Gestational age (wk) 39 0 / / / / / 7 Postnatal age (h) Serum free hemoglobin (g/l) Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ

4 Clinical Chemistry / Original Article Relationship of Clinical Variables to TcB Minus TsB Bias Generalized estimating equations were used to determine whether the mother s ethnicity, the gestational age, or postnatal age in hours impacted the relationship between TcB and TsB (ie, TcB minus TsB bias). For purposes of data analysis, the median bias (TcB minus TsB) was analyzed for Caucasian (n = 146) and all non-caucasian (n = 31) participants. The median TcB minus diazo TsB bias was not significantly different between Caucasian and non-caucasian participants (P =.2207). For the subset having Vitros TsB measured, the median bias (TcB minus Vitros TsB) was not significantly different between Caucasian and non-caucasian participants (P =.1003). Thus, the mother s ethnicity did not contribute to variability between observed TcB and TsB values. The median gestational age was 39.0 weeks (Table 1). There was no relationship between median bias (TcB minus diazo TsB) and gestational age (P =.6026) or between TcB minus Vitros TsB bias and gestational age (P =.7570). The median postnatal age at the time of TcB measurement was 48 hours. The median bias (TcB minus diazo or Vitros TsB) was not related to the patient s age of life in hours (diazo, P =.4461; Vitros, P =.9133). Relationship of Laboratory Variables to TcB Minus TsB Bias There were 126 collections by capillary heel stick, 20 collections by venipuncture, and 31 instances in which the collection method was not specified. The collection method (capillary heel stick vs venipuncture) was not related to TcB minus diazo bias (P =.6173) or to TcB minus Vitros TsB bias (P =.7382). The median free hemoglobin level in the 177 serum samples was 1.76 g/l (Table 1). There was no ztable 2z Concordance Between Transcutaneous and Serum (Diazo) Bilirubin * relationship between the level of free hemoglobin (hemolysis) in the serum sample and TcB minus diazo TsB bias (P =.1151). There was a weak relationship that did not approach statistical significance between free hemoglobin level and TcB minus Vitros TsB bias (P =.0657). There were 88 samples collected into clear serum tubes and 89 samples collected into amber serum tubes, which protected samples from light. All samples were analyzed by the diazo method, and a subset (n = 131) was analyzed on the Vitros. The median TcB minus diazo TsB bias was 2.2 mg/ dl (37.6 µmol/l) for the 88 samples collected into clear tubes and 2.0 mg/dl (34.2 µmol/l) for the 89 samples collected into amber tubes (P =.7437; no significant difference between clear and amber tubes). In contrast, the median bias for TcB minus Vitros TsB was 1.7 mg/dl (29.1 µmol/l) for the 60 samples collected into clear tubes and 0.9 mg/dl (15.4 µmol/l) for the 71 samples collected into amber containers (P =.0119; significant difference between clear and amber tubes). Using TcB to Predict Risk of Severe Hyperbilirubinemia One of the goals of the study was to determine whether the TcB value and the infant s age in hours could be used to predict whether the serum bilirubin result fell into high or high-intermediate risk zones (demanding serum bilirubin measurement and follow-up) or whether the serum bilirubin result fell into low-intermediate or low risk zones (potentially avoiding serum measurement). TcB risk was in a higher risk zone in 128 (72.3%) of 177 neonates compared with the diazo TsB, while the TcB was in a higher risk zone for 68 (51.9%) of 131 neonates compared with the Vitros TsB. The concordance between risk category determined by TcB and TsB is shown in ztable 2z and ztable 3z. Sensitivity, Transcutaneous Bilirubin Serum Bilirubin (Diazo) Low or Low-Intermediate Risk High-Intermediate or High Risk Total Low or low-intermediate risk High-intermediate or high risk Total * All results were classified as low, low-intermediate, high-intermediate, or high risk based on infant s age of life in hours and bilirubin level using the Bhutani nomogram. ztable 3z Concordance Between Transcutaneous and Serum (Vitros) Bilirubin * Transcutaneous Bilirubin Serum Bilirubin (Vitros) Low or Low-Intermediate Risk High-Intermediate or High Risk Total Low or low-intermediate risk High-intermediate or high risk Total * All results were classified as low, low-intermediate, high-intermediate, or high risk based on infant s age of life in hours and bilirubin level using the Bhutani nomogram. Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ 979

5 Karon et al / Evaluation of Transcutaneous Bilirubin specificity, and the 95% CI are given in the following text. The sensitivity of high or high-intermediate TcB for predicting a high or high-intermediate diazo TsB was 98% (CI, 90%-100%). The specificity of TcB for predicting high or high-intermediate diazo TsB was only 40% (CI, 31%-49%). If high-intermediate or high TcB was used to predict only high risk zone (>95th percentile for age) diazo TsB, the sensitivity went up to 100% (CI, 75%-100%), but the specificity dropped to 30% (CI, 23%-37%). Assuming that infants with low-intermediate or low-risk TcB results would have been spared a blood draw, then 49 (27.7%) of 177 blood draws would have been avoided by using TcB to screen for serum bilirubin risk zone. The sensitivity of high or high-intermediate TcB predicting high or high-intermediate Vitros TsB was 94% (CI, 85%-98%). The specificity of TcB for predicting high or high-intermediate Vitros TsB was 55% (CI, 42%-67%). If high-intermediate or high risk TcB was used to predict only high risk zone (>95th percentile for age) Vitros TsB, the sensitivity went up to 100% (CI, 79%-100%), but the specificity fell to 34% (CI, 25%-43%). Assuming that a low or low-intermediate TcB value would have spared a blood draw, then 39 (29.8%) of 131 blood draws would have been avoided by using TcB screening. To determine the optimal relationship between TcB and TsB, 1.0 mg/dl (17.1 µmol/l) was subtracted from each of the 177 TcB values obtained, such that the median bias between TcB and diazo TsB was 1.0 mg/dl (17.1 µmol/l). Recalculation of sensitivity and specificity as above resulted in 95% sensitivity (CI, 85%-99%) and 63% specificity (CI, 54%-72%) for high-intermediate or high risk TcB predicting high-intermediate or high risk diazo TsB. The sensitivity for detection of high risk diazo TsB was still 100% after subtraction of 1 mg/dl (17.1 µmol/l) from each TcB value. In this case, 45% of blood draws could be safely avoided by using the same assumptions as defined above. Subtraction of 1.5 mg/dl (25.7 µmol/l) or more from each TcB value decreased the sensitivity for the detection of high risk (>95th percentile for age) diazo TsB to less than 100%. Therefore, the optimal relationship between TcB and diazo TsB for risk category prediction within our institution occurs when median TcB values are approximately 1.0 mg/dl (17.1 µmol/l) higher than diazo serum values. Discussion Factors Affecting the Relationship Between TcB and TsB One previous study found that the BiliChek device overestimated serum bilirubin as measured by a direct photometric method, 6 with a mean bias between TcB and TsB very similar to that observed in our study. At least 4 studies have concluded that the BiliChek TcB slightly underestimates the TsB across a wide range of serum bilirubin levels These studies have used diazo methods, 7,8 multiple laboratory methods and high-pressure liquid chromatography, 9 or direct photometric methods. 10 Populations studied included healthy Hispanic newborns, 7 healthy preterm infants, 10 and full-term infants determined to be at risk for hyperbilirubinemia. 8 It is not clear what factors (eg, population studied, serum bilirubin method, gestational age, or postnatal age) account for variability in the relationship between TcB and TsB observed between studies. Two other studies found that the BiliChek TcB slightly overestimated the TsB at lower serum bilirubin concentrations but underestimated serum bilirubin values at concentrations of more than 11.7 mg/dl (200.1 µmol/l). 11,12 We found that the TcB systematically overestimated serum bilirubin values by the diazo and Vitros methods over a wide range of bilirubin concentrations, which has not been observed in previous studies using similar serum bilirubin methods. Local differences in the performance of similar serum bilirubin methods are observed in proficiency testing surveys. In 1 study, a sample with a reference bilirubin value (high-performance liquid chromatography) of mg/dl ( µmol/l) was sent to 1,743 laboratories for analysis by the local method used. The range of results reported by laboratories using the Vitros BuBc technology ranged from 18.7 to 21.5 mg/dl ( µmol/l), while values from laboratories using the Roche Cobas Integra ranged from 18.7 to 24.3 mg/dl ( µmol/l), and differences among laboratories using the same assay technology were almost as great as systematic differences between methods. 18 Blood sample transport (clear vs amber tubes) has not been studied previously. We found that for the Vitros method the bias between TcB and TsB was significantly smaller when samples were collected into amber tubes that protected them from light. No relationship was found between transport condition and TcB minus diazo TsB bias. In addition, the relationship between the hemolysis level and TcB minus TsB bias was more significant for the Vitros method compared with the diazo method (although neither reached statistical significance). Local practices related to blood collection and transport affect the relationship between TcB and TsB values in manners that will also vary by bilirubin method. Together, systematic differences in serum bilirubin methods (diazo vs Vitros in our study) and local differences in the performance of serum bilirubin assays (eg, calibration issues, specimen collection and transport issues) account for much of the variability in the relationship between TcB and TsB values observed in previous studies. Factors Not Affecting the Relationship Between TcB and TsB One previous study found that the difference between TcB and TsB as measured by high-performance liquid chromatography was independent of race, gestational age, and 980 Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ

6 Clinical Chemistry / Original Article postnatal age. 9 Our study extends these findings by observing that the difference between TcB and TsB values as measured by diazo or Vitros is not affected by ethnicity, gestational age, or postnatal age. A limitation of our study and the previous investigation 9 was the inclusion of a majority of Caucasian infants. One previous study also found that TcB significantly underestimated TsB in infants with a postnatal age of more than 80 hours, 11 but our study did not include enough infants in this category to address this population. The method of specimen collection (venipuncture vs heel stick) does not seem to influence the relationship between TcB and TsB values. TcB to Predict the Risk of Severe Hyperbilirubinemia Only 1 previous study used the BiliChek to predict serum bilirubin risk zone as determined by the nomogram of Bhutani et al. 17 This study found that the sensitivity of highintermediate or high risk TcB values for predicting high risk (>95th percentile for age) TsB was 100%, with a specificity of 66%. 7 These authors compared BiliChek TcB with a diazo TsB method and found that the TcB slightly underestimated the diazo TsB by 0.6 mg/dl (10.26 µmol/l). 7 In contrast, we found that the specificity of high-intermediate or high risk TcB for predicting high risk TsB was only 30% (diazo) or 34% (Vitros) owing to systematic overestimation of the serum bilirubin level as performed in our institution. The previous study 7 did not address use of TcB values to predict highintermediate or high risk (>75th percentile) serum bilirubin values. In our institution, the optimal relationship between TcB and TsB for this type of risk prediction occurs when TcB values are approximately 1.0 mg/dl (17.1 µmol/l) higher than TsB results. This allows for 95% sensitivity and more than 60% specificity for prediction of high-intermediate or high risk TsB values and would allow for a 45% reduction in the number of blood draws for serum bilirubin measurement. Because the relationship between TcB and TsB values will vary by institution and method, this relationship would need to be established in every institution and for each serum bilirubin method used for risk assessment. Use of TcB in the TsB Nomogram Several previous studies 6,8-12 have used TcB to predict whether serum bilirubin was above or below a fixed cutoff value, an approach that is limited because it does not allow for bilirubin levels to be evaluated in the context of postnatal age, as recommended. 5 Several institutions have created independent nomograms specifically for TcB values. 13,14 However, TcB measurement is difficult if not impossible to standardize, whereas a reference method exists for TsB measurement. 18 In addition, transcutaneous nomograms may not be as readily available for everyday use as is the Bhutani nomogram via tools such as bilitool.org. The use of TcB as a screening method for predicting the serum bilirubin risk zone will allow screening by a noninvasive method at any postnatal age, with confirmation in high risk infants and clinical decisions based on a measurement that can be standardized across institutions to a reference method. Conclusion We have shown that in our institution, BiliChek TcB measurement is a sensitive but not specific screening method for hyperbilirubinemia when TcB values are used in place of serum values to predict risk category based on the Bhutani nomogram. Variables associated with serum bilirubin transport and measurement, method of serum bilirubin measurement used in the clinical laboratory, and local performance characteristics of laboratory bilirubin assays are the primary determinants of the relationship between TcB and TsB values. Until laboratory methods for serum bilirubin are better standardized, the safety and efficacy (percentage of blood draws eliminated) of TcB screening for prediction of TsB risk category will likely vary within each institution as a function of local practices in the measurement of serum bilirubin. From the Departments of 1 Laboratory Medicine and Pathology and 2 Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, MN. Address reprint requests to Dr Cook: Dept of Pediatric and Adolescent Medicine, Mayo Clinic, 200 First St SW, Rochester, MN References 1. Jain S. The neonatal erythrocyte and its oxidative susceptibility. Semin Hematol. 1989;26: Askin D, Diehl-Jones W. The neonatal liver, part III: pathophysiology of liver dysfunction. Neonatal Netw. 2003;22: Maisels M. Neonatal jaundice. Pediatr Rev. 2006;27: Van Praagh R. Diagnosis of kernicterus in the neonatal period. Pediatrics. 1961;28: American Academy of Pediatrics Subcommittee on Hyperbilirubinemia. Management of hyperbilirubinemia in the newborn infant 35 or more weeks of gestation [published correction appears in Pediatrics. 2004;114:1138]. Pediatrics. 2004;114: De Luca D, Zecca E, de Turris P, et al. Using BiliCheck for preterm neonates in a sub-intensive unit: diagnostic usefulness and suitability. Early Hum Dev. 2007;83: Kolman K, Mathieson K, Frias C. A comparison of transcutaneous and total serum bilirubin in newborn Hispanic infants at 35 or more weeks of gestation. J Am Board Fam Med. 2007;20: Samanta S, Tan M, Kissack C, et al. The value of BiliCheck as a screening tool for neonatal jaundice in term and near-term babies. Acta Paediatr. 2004;93: Rubaltelli F, Gourley G, Loskamp N, et al. Transcutaneous bilirubin measurement: a multicenter evaluation of a new device. Pediatrics. 2001;107: Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ 981

7 Karon et al / Evaluation of Transcutaneous Bilirubin 10. Szabo P, Wolf M, Bucher H, et al. Assessment of jaundice in preterm neonates: comparison between clinical assessment, two transcutaneous bilirubinometers and serum bilirubin values. Acta Paediatr. 2004;93: Boo N, Ishak S. Prediction of severe hyperbilirubinemia using the BiliCheck transcutaneous bilirubinometer. J Paediatr Child Health. 2007;43: Grohmann K, Roser M, Rolinski B, et al. Bilirubin measurement for neonates: comparison of 9 frequently used methods. Pediatrics. 2006;117: Maisels MJ, Kring E. Transcutaneous bilirubin levels in the first 96 hours in a normal newborn population 35 weeks' gestation. Pediatrics. 2006;117: De Luca D, Romagnoli C, Tiberi E, et al. Skin bilirubin nomogram for the first 96 h of life in a European normal healthy newborn population, obtained with multiwavelength transcutaneous bilirubinometry. Acta Paediatr. 2008;97: Malloy H, Evelyn K. The determination of bilirubin with the photoelectric colorimeter. J Biol Chem. 1937;119: Algeciras-Schimnich A, Cook W, Milz T, et al. Evaluation of hemoglobin interference in capillary heel-stick samples collected for determination of neonatal bilirubin. Clin Biochem. 2007;40: Bhutani V, Johnson L, Sivieri E. Predictive ability of a predischarge hour-specific serum bilirubin for subsequent significant hyperbilirubinemia in healthy term and near-term infants. Pediatrics. 1999;103: Lo S, Doumas B, Ashwood E. Bilirubin proficiency testing using specimens containing unconjugated bilirubin and human serum. Arch Pathol Lab Med. 2004;128: Am J Clin Pathol 2008;130: DOI: /AJCPRX1E3NWCXHMZ