| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
Departments of1 Pathology and 5 Pediatrics, Oregon Health and Science University, Portland, OR; Departments of2 Pediatrics and Pathology and 3 Laboratory Medicine, The Brody School of Medicine at East Carolina University, Greenville, NC;4 Waisman Center, University of Wisconsin, Madison, WI
aaddress correspondence to this author at: Department of Pathology, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd., Mailcode L-471, Portland, OR 97239; e-mail kazmierc{at}ohsu.edu
Measurement of bilirubin in sera of newborn infants is one of the most frequently performed laboratory tests in this patient population (1)(2)(3)(4). Chemical methods for measurement of bilirubin are the standard of care in the assessment of neonatal jaundice. However, transcutaneous techniques for assessing in vivo bilirubin concentrations have been advocated as a more cost-effective and less traumatic alternative to the collection of blood by heel puncture (2)(3).
Studies have varied in their assessment of the accuracy of transcutaneous bilirubin assessment; poor to excellent agreement between transcutaneous bilirubin measurements and bilirubin measured in plasma has been reported (1)(2)(3)(5)(6)(7)(8)(9). The best agreement between transcutaneous bilirubin measurements and measurement of plasma bilirubin concentrations has been reported for studies using homogeneous populations of newborns with comparable basal skin color (9). Unfortunately, many factors, such as hemoglobin concentration, the pH and bilirubin-binding capacity of albumin, the technique used to obtain the blood sample, skin pigmentation, and requirement by some methods for baseline measurement of skin color immediately after birth, can significantly effect the agreement between transcutaneous and chemical methods for bilirubin (6)(10)(11).
We evaluated the BiliCheck point-of-care device (Respironics, Marietta, GA), which performs transcutaneous measurement of bilirubin by multiwavelength spectral analysis. We compared results obtained with the BiliCheck with bilirubin concentrations in blood specimens measured in a hospital laboratory. In addition, we also measured bilirubin by HPLC. Although HPLC is labor-intensive and not practical for routine use, this method is not subject to interference from hemoglobin or lipemia. Our goals were to evaluate the accuracy of the BiliCheck method compared with a laboratory-based bilirubin method, using HPLC as the comparison method. In addition, we assessed the effect of gestational age, birth weight, postnatal age, and skin color score on the performance of the BiliCheck method.
This study was performed between January and December 2000 after approval by the hospital Institutional Review Board. Neonates <4 days of age were eligible for participation if they had not undergone phototherapy or exchange transfusion. None of the study participants had clinical manifestations of respiratory distress, sepsis, or cardiac or circulatory disease. We obtained parental consent to prospectively evaluate 108 newborns.
Blood and transcutaneous bilirubin measurements were obtained on all study participants, and none were excluded. For HPLC analysis, sufficient blood was obtained for analysis from 95 newborns.
Transcutaneous measurements were performed in duplicate immediately before heel puncture. This protocol was followed to ensure that the transcutaneous measurements were done under quiet conditions when the infant was not crying. A location on the forehead was chosen that was free of any bruising, local nevus, hemangioma, or melanotic patch.
We used skin color score rather than race to investigate the effect of skin pigmentation on the accuracy of the BiliCheck because of the wide range of skin color that can be observed among individuals of the same race. We assessed skin color by evaluating skin pigmentation of the forehead with use of a skin tone chart provided by SpectRx; this chart provides four gradations of skin color. We added intermediate gradations, so that each infant was classified with a skin color score ranging from 1 (lightest) to 8 (darkest).
All blood sampling from newborns was performed by heel stick within 5 min after measurement with the BiliCheck instrument. After warming of the heel and lancet puncture, blood was collected by the drip method into heparin-containing, amber-colored containers (12). A single operator with previous experience in the collection of blood by heel puncture and use of the BiliCheck device performed all determinations.
Bilirubin was measured in the clinical laboratory by a diazo method in an Olympus AU640 analyzer (Olympus Corp.). A second aliquot of plasma was placed in an amber-colored container and frozen at -70 °C for subsequent analysis of bilirubin by HPLC at the University of Wisconsin-Madison. Bilirubin concentrations were measured by HPLC according to a previously described method (13).
The median bilirubin concentration in samples measured by HPLC (n = 95) was 185.5 µmol/L (range, 29.1–354.0 µmol/L). Identical medians were obtained for specimens measured with the Olympus method (median, 196.7 µmol/L; range, 54.7–283.9 µmol/L; n = 108) and BiliCheck device (median, 196.7 µmol/L; range, 58.1–306.1 µmol/L; n = 108).
Bland–Altman (14) difference plots were constructed by plotting the bilirubin concentrations obtained with each of the methods against one another (Fig. 1
). These plots revealed significant differences between HPLC and the Olympus and BiliCheck methods as a function of bilirubin concentration. However, the mean differences between HPLC and the Olympus method (-5.0 µmol/L) and between HPLC and the BiliCheck (6.0 µmol/L) were not considered clinically significant. At low bilirubin concentrations, the results obtained by HPLC were lower than those obtained by the two other methods, whereas at higher bilirubin concentrations, HPLC-measured bilirubin tended to be higher than the results obtained with the two other methods (Fig. 1
, A and B). For bilirubin concentrations measured with the Olympus analyzer and HPLC, regression analysis of the data from the difference plots revealed a slope of 0.24 and y-intercept of -48.9 µmol/L (Fig. 1A
). Analysis of bilirubin concentrations measured with the BiliCheck and HPLC revealed a slope of 0.26 and y-intercept of -54.5 µmol/L (Fig. 1B
). Bland–Altman analysis of bilirubin concentrations measured by the Olympus and BiliCheck showed no concentration-dependent changes in agreement between the two methods (Fig. 1C
).
|
Skin color scores were used to evaluate the effect of skin pigmentation on the BiliCheck device. Individuals with higher (e.g., darker) skin color scores were less numerous. Thus, some skin color score groups were combined. We compared the measurement error between bilirubin measured with the BiliCheck vs HPLC in individuals with a skin color score of 1 (n = 33) with those with skin color scores of 2 (n = 13), 3 (n = 21), 4 and 5 (n = 14), and 6, 7, and 8 (n = 14). There was no relationship between measurement error and skin color score between any of the skin color score groups (ANOVA, P = 0.48).
The effect of birth weight on the accuracy of BiliCheck measurements was evaluated by grouping newborns into one of three categories based on birth weight: <2500 g (n = 25), 2500–3499 g (n = 39), and 
3500 g (n = 31). We found no difference in measurement error as a function of birth weight (ANOVA, P = 0.51).
The effect of gestational age on the accuracy of the BiliCheck was evaluated by comparing newborns with gestational age 
36 weeks (n = 25) to those with gestational ages of >36 weeks (n = 83). The mean difference in bilirubin concentrations measured with HPLC and the BiliCheck device was identical (-3.4 µmol/L) in both groups.
Finally, the effect of postnatal age on BiliCheck accuracy was investigated by performing regression analysis. Each patient’s postnatal age, in hours, when testing was performed was used as an independent variable to predict the difference in bilirubin concentrations measured by HPLC and the BiliCheck. We found no significant effect of postnatal age on BiliCheck accuracy (r = 0.16; t-test for non-zero slope, t = 1.53; P 0.05).
We did not evaluate the precision of the BiliCheck instrument. Intradevice precision (SD) has previously been determined to be 10.1 µmol/L, whereas interdevice precision has been stated to be 11.6 µmol/L (10). The imprecision (CV) for total bilirubin measurements on the Olympus analyzer ranged from 4.2% at a concentration of 123.1 µmol/L to 2.8% at a concentration of 372.8 µmol/L.
The ability to measure bilirubin concentrations in newborns simply, rapidly, accurately, cost-effectively, and with minimal risk or discomfort has taken on increased importance in the current environment of managed care, capitation, litigation, and brief postpartum hospitalization. Measurement of bilirubin by various transcutaneous techniques has been reported from several studies with mixed results (5)(7)(8)(9)(10)(15)(16)(17)(18)(19). Race, as assessed by skin color score, did not have an effect on the performance of the BiliCheck method. In addition, other potential confounding factors, such as birth weight, gestational age, and postnatal age, did not affect the BiliCheck device.
The results of our study may have been influenced by several factors. Freezing and storage of samples for HPLC analysis might have led to some degradation of bilirubin. However, the data shown in the Bland–Altman difference plots suggest that bilirubin degradation was not a significant factor. Other factors that might have influenced our results were the demographics of newborns evaluated in our study. All infants were >32 weeks of gestational age, none had undergone phototherapy, and all were <4 days of age. In addition, all transcutaneous measurements were performed in duplicate by a single individual, whereas singleton determinations by multiple caregivers would most likely be performed in routine practice.
Currently, measurement of total bilirubin in serum or plasma of newborns is the standard of care for the assessment of neonatal jaundice. Our results indicate that the transcutaneous measurement of bilirubin is as accurate as bilirubin measured in plasma in a hospital laboratory when HPLC is used as the comparison method. In conclusion, measurement of bilirubin by the BiliCheck device offers an accurate, rapid, and noninvasive means of assessing plasma bilirubin concentrations in neonates.
Acknowledgments
We thank Respironics (Marietta, GA) for supplying the reagents used in this study. We also acknowledge the cooperation of the newborn nursing staff at Pitt County Memorial Hospital.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. R. Petersen, A. O. Okorodudu, A. A. Mohammad, A. Fernando, and K. E. Shattuck Association of Transcutaneous Bilirubin Testing in Hospital with Decreased Readmission Rate for Hyperbilirubinemia Clin. Chem., March 1, 2005; 51(3): 540 - 544. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
