Professional Documents
Culture Documents
● Context.—In 2003 the Chemistry Resource Committee of variation for the 4 major instrument groups (Dimension,
the College of American Pathologists introduced a com- Olympus, Synchron, and Vitros), which report 65% of all
mutable specimen in the Neonatal Bilirubin Surveys. This results, varied from 2% to 3%. College of American Pa-
specimen was intended to help evaluate all bilirubin meth- thologists All Data mean bilirubin values were within 0.46
ods. mg/dL (7.8 mol/L) of the reference method mean in
Objective.—To evaluate the effect of commutable spec- 2003; in subsequent years these differences became larger,
imens on the performance of selected clinical analyzers in peaking at 1.87 mg/dL (32 mol/L) in 2005.
measuring neonatal bilirubin from 2003 through 2006. Conclusions.—The large systematic error of bilirubin
Design.—A human serum–based specimen enriched measurements is due primarily to failure of instrument
with unconjugated bilirubin in human serum has been in- manufacturers to produce reliable bilirubin calibrators.
cluded since 2003 in the Neonatal Bilirubin Surveys. The Primary calibrators should consist of human serum en-
bilirubin values of these specimens were determined by the riched with unconjugated bilirubin. Bilirubin values must
reference method and used to evaluate results reported by be assigned by the reference method, the performance and
various chemistry analyzers. robustness of which are reported in this article. Secondary
Results.—Coefficients of variation for College of Amer- calibrators distributed to users must be traceable to pri-
ican Pathologists All Data ranged from 4.9% to 6.2% for mary calibrators.
the Neonatal Bilirubin Survey. However, coefficients of (Arch Pathol Lab Med. 2008;132:1781–1785)
pensed in airtight plastic tubes and kept at ⫺70⬚C until dL [85.5 mol/L], and 10 mg/dL [171 mol/L]) are an-
use. Each time the reference method is used, the stock alyzed to construct a calibration curve, the regression
solution is analyzed to verify that its value is within the equation of which is used to calculate the results of the
accepted limits (mean value ⫾ 3 SD). Next, the stock so- specimens analyzed. Shown in Table 3 are molar absorp-
lution and 3 dilutions (2.5 mg/dL [42.7 mol/L], 5.0 mg/ tivities, slopes, and intercepts of calibration curves ob-
Arch Pathol Lab Med—Vol 132, November 2008 Laboratory Performance in Neonatal Bilirubin Testing—Lo et al 1783
Table 2. Molar Absorptivity of Bilirubin Alkaline Table 3. Long-term Calibration Stability and
Azopigment Results From 2 Round Robins and the Precision of the Reference Method for Total Bilirubin
Reference Laboratory of Children’s Hospital of
Calibration Curve
Wisconsin (CHW)*
(September 15, 2003–April 23, 2007)
n (L· mol⫺1 · cm⫺1)† SD
(L· mol⫺1 · cm⫺1)* Slope Intercept
Round Robin 1988 (5 labs) 10 76 490 610
Mean 76 554 0.0770 ⫺0.0003
Round Robin 1998 (6 labs) 12 76 640 601
SD† 377 0.0004 0.0018
Reference Lab CHW 11 76 869 589
N 45 46 46
* n indicates number of determinations; , molar absorptivity at 598 Min 75 579 0.0759 ⫺0.0038
nm. Max 77 257 0.0782 0.0043
† Corrected for purity (98.3%) of bilirubin.
* indicates molar absorptivity.
† One standard deviation for a single measurement.
Arch Pathol Lab Med—Vol 132, November 2008 Laboratory Performance in Neonatal Bilirubin Testing—Lo et al 1785