| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Pathology and Immunology, Division of Laboratory Medicine, Washington University School of Medicine, 660 South Euclid Ave., Box 8118, St. Louis, MO 63110
aauthor for correspondence: fax 314-362-1461, e-mail gronowski{at}pathology.wustl.edu
Bilirubin, a breakdown product of lysed red blood cells, is present in amniotic fluid in very small concentrations relative to serum (1). In an uncomplicated pregnancy, bilirubin in amniotic fluid peaks at 
19–22 weeks of gestation at concentrations of 1.6–1.8 mg/L (1). Rh isoimmunization and its severe manifestation, erythroblastosis fetalis, are associated with intrauterine hemolysis, which leads to increases in amniotic fluid bilirubin concentrations (2) up to 9.6 mg/L (3)(4). The assessment of fetal lung maturity (FLM) in these pregnancies is sometimes necessary, but the ability to utilize current methods in the presence of increased bilirubin concentrations is unclear. The commercial TDx FLM II assay (Abbott Laboratories), which is based on fluorescence polarization technology and the dye PC-16 (5), is the most commonly used method to assess FLM (6). Although the manufacturer does not recommend testing icteric amniotic fluid samples, nothing has been published about the nature or quantification of their interference with the TDx FLM II assay. Our objective was to examine the effect of bilirubin on TDx FLM II concentrations.
Frozen leftover amniotic fluid samples sent to the laboratory for physician-ordered FLM testing were sorted and combined posttesting to obtain six pools with FLM values of 23, 31, and 36 mg/g (immature, <39 mg/g); 44 mg/g (intermediate); and 60 and 77 mg/g (mature, >55 mg/g). Each pool contained amniotic fluid from at least four women. Amniotic fluid samples with even minimal visual evidence of hemolysis or meconium were excluded from the study. Each pool was analyzed for total bilirubin concentration on the Hitachi 747 analyzer. The analytical sensitivity for bilirubin measurements in amniotic fluid is poor on this analyzer at concentrations <1.0 mg/L (Table 1 in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol49/issue6/); therefore, this assay is not used routinely to measure bilirubin in amniotic fluid. This method was suitable, however, for the purpose of confirming that none of the pooled amniotic fluid samples had pre-addition bilirubin concentrations >1.0 mg/L. Human studies committee approval was obtained for this study.
A stock solution of unconjugated bilirubin (2500 mg/L) was prepared by dissolving bilirubin (cat. no. 101018; ICN Biomedicals Inc.) in laboratory-grade dimethyl sulfoxide (DMSO; cat. no. D-5879; Sigma Chemical Company; final concentration, 40 mL/L) at alkaline pH, and then neutralizing the solution with acid. The DMSO concentration in the sample with the highest concentration of bilirubin in our experiments (15 mg/L) was 1 mL/L. Addition of 2 mL/L DMSO to amniotic fluid samples caused no change in FLM concentrations.
The stock solution of unconjugated bilirubin was added to an aliquot of pooled sample, vortex-mixed, and serially diluted to create samples with decreasing concentrations of bilirubin. The FLM concentrations of these samples and the pre-addition pool were measured in duplicate in the TDx FLM II assay. Experiments were performed on 2 separate days (three pools each day). Validation studies were performed with serum to establish that there was no matrix effect when measuring bilirubin at this concentration in amniotic fluid (Table 1 of the Data Supplement). The highest bilirubin concentration (15 mg/L) was quantified with the Hitachi 747. Because of the inability of the Hitachi 747 to measure amniotic fluid bilirubin at low concentrations, subsequent concentrations are reported as "calculated" based on the serial dilutions.
A mixed-model ANOVA using the percentage change from baseline as the dependent variable was performed. Baseline was calculated as the average of the two replicates for a given pool before bilirubin was added. The percentage change from baseline as a function of pool (six pools), the baseline FLM concentration (treated as a continuous variable), replicates within pools (two replicates), and bilirubin concentrations (treated as a categorical variable, not a continuous variable) were used in this model. The baseline FLM concentration and bilirubin concentration were modeled as fixed effects, and pool and replicates were modeled as random effects. The effect of baseline FLM concentration on FLM concentration after the addition of bilirubin was not significant (P = 0.91), so the analysis was conducted excluding baseline FLM as a factor.
Our results indicate that bilirubin does not produce a clinically important or consistent effect on TDx FLM II results (Fig. 1
). The results of a mixed-model ANOVA statistical analysis indicated that only at a bilirubin concentration of 7.5 mg/L did the mean difference in TDx FLM II concentration from baseline (-3.3%) achieve statistical significance (P = 0.04). However, despite the statistical significance, the 95% confidence interval indicated that samples would, at most, be 6% different from baseline. We do not consider 6% to be clinically significant. Furthermore, a higher concentration of bilirubin (15 mg/L) did not produce a statistically significant difference in TDx FLM II from baseline.
|
The highest concentration of bilirubin tested in our experiments was 15 mg/L, which is 50% greater than the highest concentration likely to occur in the amniotic fluid of abnormal pregnancies, including erythroblastosis fetalis (4). This study did not find a clinically significant interference at this concentration or below it. For this reason, we conclude that bilirubin does not affect the assessment of FLM by the TDx FLM II assay of amniotic fluid.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, 77.0 mg/g; 
, 60.4 mg/g; 
, 44.3 mg/g; 
, 36.0 mg/g; 
, 30.9 mg/g; 
, 23.2 mg/g. A mixed-model ANOVA demonstrated no statistical difference attributable to baseline TDx FLM II concentrations.