| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters |
a Author for correspondence. Fax 32 9 240 49 85; e-mail alain.verstraete{at}rug.ac.be
Department of Clinical Chemistry, Immunology and Microbiology, University Hospital Gent, De Pintelaan 185, 9000 Gent, Belgium
To the Editor:
Test strips are widely used to estimate pH and to detect nitrites, protein, glucose, ketones, urobilinogen, bilirubin, leukocytes, and hemoglobin in urine. However, little is known about the bleeding of substances from the strip and interference with subsequent tests performed on the same aliquot. Falsely increased iodine and glycine values in urine contaminated by test strips have been reported (1)(2). We describe here how urine contamination by test strips can lead to erroneous conclusions such as caffeine poisoning or even caffeine doping and how contaminants from test strips are detected by toxicological gas chromatography/mass spectrometry (GC/MS) analysis.
A urine sample was sent to our laboratory for the exclusion of a
possible toxicological problem in a 9-year-old girl with a hypoglycemic
coma that did not improve despite adequate therapy. We performed a
toxicological screening on GC/MS (Saturn II; Varian) after
liquid/liquid extraction (ToxiTube®A; Ansys
Diagnostics). The chromatogram (Fig. 1
) showed a very high caffeine peak and many peaks with
"unknown" mass spectra. Initially, possible caffeine
poisoning was diagnosed, especially in light of the underlying diabetes
problem. However, after comparing the unknown mass spectra to the
NIST-62k database, we identified one of the unknown spectra as
tetramethylbenzidine (TMB; molecular mass, 240 Da; formula,
C16H20N2;
fit, 916 out of 1000; Fig. 1
, bottom inset). This compound is used in
the detection of hemoglobin. TMB reacts with cumene hydroperoxide by
the peroxidase action of hemoglobin to produce a blue color
(3). We suspected contamination by a urine test strip. As a
confirmation of our assumption, we dipped a strip
(N-Multistix-SG®; Bayer) for 1 s in an 8-mL
blank urine. TMB, a high caf-feine peak, and many unknown mass spectra
were observed in this sample, a pattern completely comparable to that
in the child's urine. In addition, we identified two other strip
contaminants with the NIST-62 k database: methylquinoline (molecular
mass, 143 Da; formula, C10H9N; fit, 919 out of 1000;
Fig. 1
, top inset), used for the detection of nitrite, and
diethyl-aminobenzaldehyde (molecular mass, 177 Da; formula,
C11H15NO; fit; 939 out of 1000; Fig. 1
, middle inset),
used for the detection of urobilinogen.
|
The presence of caffeine in the Bayer strip was unexpected because it was not mentioned on the package insert. However, caffeine is widely used as a reaction accelerator in the detection of bilirubin. Agents such as caffeine promote the diazo coupling of free bilirubin in aqueous solutions (4)(5). We dipped a Bayer strip in 8 mL of a test subject's urine for 1 s (n = 6) and 5 s (n = 6), respectively. For the measurement of caffeine concentrations by HPLC, 8-chlorotheophylline (internal standard) was added to the samples, and proteins in the samples were precipitated with acetonitrile. The evaporated extracts were resuspended in mobile phase (50 mmol/L potassium phosphate, pH 3.8, containing 80 mL/L acetonitrile) and injected onto a RP-C18 column (150 x 4.6 mm; Waters Symmetry) at 25 °C and a flow rate of 1 mL/min. We used ultraviolet detection at 273 nm. In the uncontaminated sample, the caffeine concentration was 10 µmol/L. The caffeine concentration increased in the contaminated samples to 51 ± 10 µmol/L (mean ± SD) in the 1-s and 113 ± 21 µmol/L in the 5-s samples. For a healthy adult, a urine concentration of 75 µmol/L is approximately equivalent to drinking 10 cups of coffee or >20 coladrinks. Furthermore, because of its stimulant activity, the concentration of caffeine in the urine is measured in doping analysis. According to the regulations of the International Olympic Committee, the caffeine concentration in urine should not exceed 62 µmol/L (6). Placing a test strip in urine for a few seconds can lead to falsely increased concentrations, with far-reaching consequences for the athlete.
We also performed GC/MS analysis on 8 mL of blank urine into which two other reagent strips had been dipped for 1 s: Rapignost® Total Screen (Behring Diagnostics) and Combur10 Test® M (Boehringer Mannheim), respectively. The chromatograms and spectra obtained with these two strips were very comparable but totally different from the Bayer strip. Neither of these two strips contained any caffeine. However, TMB was present in both strips. None of the unknown spectra obtained with these strips has been fully identified.
This report shows that dipping a test strip into urine for as little as 1 s can cause interference in subsequent analyses, such as GC/MS screening. When the Bayer N-Multistix-SG strip is used, this interference can lead to faulty assumptions of caffeine poisoning or caffeine doping. TMB, because it is present in all of the strips tested, can be used as a marker of test strip contamination. We recommend that a urine test strip be dipped in a separate urine aliquot. Unfortunately, current trends to send smaller urine sample volumes to the laboratory (e.g., 8–10 mL in tubes) will make this recommendation increasingly difficult to follow.
Footnotes
Editor's Note: The interference described in this Letter is a result of immersing a test strip into a sample. This is not good laboratory practice, for samples or reagents. This Letter should be a reminder to all of us not to place any object other than transfer pipettes into the original specimen or text sample. Although the phrase "dip stick" implies use of this technique, the sample should be transferred to the test strip by another mechanism. When it is acceptable to immerse the strip, a portion of the specimen should be taken to be used only for that purpose. In this age of point-of-care testing, it would seem wise for laboratories to advise non-laboratory staff of this basic principle of testing
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
