| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Clinical Chemistry Service, Clinical Pathology Department, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, MD 20892.
a Address for correspondence: Clinical Pathology Department, National Institutes of Health, Bldg. 10, Rm. 2C-407, 10 Center Dr., Bethesda, MD 20892-1508. Fax 301-402-1885; e-mail msampson{at}nih.gov
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials, Methods, and ResultsDiscussionReferences |
|---|
Key Words: indexing terms: analytical error • electrolytes • ion-selective electrodes
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials, Methods, and ResultsDiscussionReferences |
|---|
We encountered a false lithium concentration in a patient not receiving Li who had been admitted in a double-blind protocol evaluating therapy with lithium, carbamazepine, and (or) valproic acid. A Becton Dickinson (BD; Franklin Lakes, NJ) 10-mL plain red-top, blue-labeled plastic Vacutainer Plus Tube® (BDP) that was received about one-third filled with this patient's blood gave results for lithium of 0.13 and 0.11 mmol/L on one Lytening 2Z analyzer (Dade Lytening Systems, Danvers, MA) and 0.12, 0.12, 0.11, and 0.11 mmol/L on a second identical instrument. By flame atomic absorption (flame AA), however, lithium was undetectable.
Although carbamazepine was present in the sample, our previous study had documented that carbamazepine did not interfere with the determination of lithium by the Lytening 2Z analyzer (3). We considered that the collection tube might be a possible source of interference because we had recently switched serum collection tubes from the standard BD glass plain red-topped Vacutainer Tube® (BDG) to the new plastic BDP, which contains a silica clot activator with a silicone surfactant (SCA).
|
Materials, Methods, and Results |
|---|
|
TopAbstractIntroductionMaterials, Methods, and ResultsDiscussionReferences |
|---|
We used a bovine protein-based liquid control sera (QCLytes; Dade) at three concentrations as a daily check of performance. We also used AA spectroscopy (Model 5100PC instrument; Perkin-Elmer, Norwalk, CT) for determining both Li (flame AA; back-up method) and Si (graphite furnace AA). We used the following Vacutainer Tubes for this study: BDP with silica clot activator (prod. no. 36-7820, lot no. 4K802); the newer version, which contains less silicone surfactant (BDP-2, striped-label; prod. no. 36-7820, lot no. 6J800, released October 1996); and pink-top BDG, with no additives (prod. no. 36-7611, lot no. 3M853). We also evaluated the following Si compounds: silica powder (supplied by BD as the product used in the BDP), AA Si standards (Sigma Chemical Co., St. Louis, MO; and Perkin-Elmer), and a siliconizing agent [Sigmacote (a solution of Si in hexane); Sigma]. Although we requested a sample of the silicone surfactant used to fix the silica powder to the BDP, this proprietary material was not supplied by BD. For recovery studies we added 10 µL of a 1 g/L Li reference standard (VWR Scientific, New York, NY) to 990 µL of a Li-free serum pool (added Li concentration 1.44 mmol/L).
effect of sample volume
Because we frequently receive samples in our laboratory in which
the BDP is only partially filled, we investigated the effect of
the volume and incubation time of the sample in the BDP on the observed
interference. We added 1, 2, 3, or 4 mL of isotonic saline or a Li-free
serum pool to BDPs and determined Li by ISE—both immediately, and
after incubating each tube for 10, 30, and 120 min. We drew whole blood
into a syringe from a normal volunteer taking no medication and
aliquoted the sample into four groups of six tubes each; five BDPs with
volumes increasing from 1 to 5 mL in 1-mL increments and one BDG with 1
mL as a control. After centrifuging the samples, we harvested the serum
at 30 min and at 1, 2, and 4 h and then determined Li by ISE in
all 24 samples.
The volume of sample in the BDP was inversely related to the apparent
Li concentration (Fig. 1
). The shape of the curve was similar for all three fluids and
essentially independent of the time of incubation. The serum pool and
serum from whole-blood specimens collected into BDP tubes had no
detectable Li determined by flame AA. We also tested the BDP-2, which
had been reformulated to reduce the amount of silicone surfactant
(4). We initially recorded a Li concentration of 1.41
mmol/L in 1 mL of Li-free serum added to a BDP, but the Li
concentration measured was 0.57 mmol/L in 1 mL of the same serum pool
added to a BDP-2. This suggests that the interferent is the silicone
surfactant, present in a lower concentration in the BDP-2.
|
We drew blood from five Li-free normal volunteers to nominally fill BDP and BDP-2 tubes. The mean (±SD) Li concentration from the BDPs was 0.18 ± 0.01 mmol/L, whereas that from the BDP-2s was 0.07 ± 0.01 mmol/L. Thus, even if the specimen is collected to the full blood volume in the BDP or BDP-2, a small but consistent false increase in Li concentration persists.
interference studies
We evaluated the analyzer for interference effects (altered ISE
response after intermittent exposure to SCA). When we repeated
experiments on the same analyzer over time, we saw less interference
from the same concentration of SCA; i.e., the apparent (falsely
increased) Li concentration decreased with time. We therefore recognize
three categories of interference: immediate interference, the false Li
result obtained with a Li-free specimen exposed to SCA; short-term
interference, a change in the ISE response with repetitive analyses
(within-run) of the same specimen or identical specimens; and long-term
interference, a change in the ISE response in response to intermittent
exposure to SCA over time.
We investigated these interference effects the same way we perform precision studies. We evaluated short-term interference both from within one tube and between multiple identical tubes. We added 2 mL of Li-free serum to a BDP and a BDG. At this volume, we received no error messages from the analyzer and were able to perform eight consecutive analyses. The mean value for the BDP was 0.11 mmol/L (CV 10.4%), but the BDG tube consistently gave a result of 0.00 mmol/L. Routine precision studies performed with quality-control specimens gave a within-day CV of 1.07% at a mean Li content of 1.12 mmol/L.
After performing this study, we installed a new membrane on the analyzer to determine how quickly the results would start to decrease. We performed one experiment to measure both within-day variation between five identical BDPs and day-to-day variation with one BDP. We added 1.25 mL of Li-free serum to five BDPs and measured Li from each tube once daily for five consecutive days, starting with a new Li membrane on day 1. Within-day variation showed CVs of 3.6–6.3% per day. There was no short-term interference (trending) from tube-to-tube during within-day variation studies. We could, however, document long-term interference (day-to-day trending) because the mean Li concentration decreased on each of the 5 days that the membrane was exposed to SCA (sequential results: 1.38, 1.32, 1.22, 1.17, and 1.04 mmol/L). The among-day CVs ranged from 9.2% to 13.4%. Routine day-to-day precision studies with quality-control samples gave a CV of 1.79% at a Li concentration of 1.18 mmol/L, which is more representative of the overall performance for the instrument. Thus, the long-term interference on the ISE with serum samples from a BDP decreases with time.
We further investigated long-term interference by comparing the results from a Li ISE membrane cap that had been intermittently exposed to SCA for 6 months with those obtained with a new membrane cap. We repeated this evaluation four times at intervals of ~2 weeks, using the same old membrane (kept wet with Li ISE filling solution between trials) and the same new membrane that was intermittently exposed to the SCA in the experimental specimens throughout the 8 weeks of the study. Using both membranes, we determined Li in the following specimens: a pool of Li-positive samples (collected in BDG), a Li-supplemented serum pool in a plain glass tube, 1 mL of Li-free serum in a BDP, 1 mL of Li-supplemented serum in a BDP, and a bovine albumin-based control material (QCLytes). We noted a large difference in results between a Li membrane that had 6 months of intermittent exposure to BDP and the fresh membrane placed on the Li ISE.
Table 1
gives the Li results (mmol/L) and recovery percentages for the
differences between the two membranes. Prolonged intermittent exposure
to SCA effected a decrease in the Li concentration for both the
positive immediate interference (specimens 1 and 4) and for specimens
not exposed to SCA (specimens 2 and 3). Serum samples with added Li
showed a 20% reduction in recovery with the old membrane and
progressively lower recovery with the new membrane every 2 weeks for 8
weeks. Performing the same recovery study with our second analyzer,
which was not exposed to samples from a BDP, showed 103% recovery of
added Li with the old membrane (4 months of regular use) and 97%
recovery with the new membrane (data not shown). Quality-control
samples (QCLytes) showed no downward trend over the 6 months that the
old membrane was in place, and the values did not change when the new
membrane was put in service. Thus, change in the response of the ISE
for serum specimens, as related to exposure to SCA over time, was not
detected with control materials.
|
experiments with si
In an effort to identify the interfering substance, we performed
several experiments with various Si compounds in glass tubes. We used
graphite AA (251.6 nm, 1400 °C char, 2650 °C atomization) to
determine the Si concentration obtained from 1 mL of saline (1.89
mmol/L Si, 0.51 mmol/L Li) or 1 mL of serum (0.54 mmol/L Si, 0.38
mmol/L Li) added to a BDP. The stoichiometric relationship between Si
interference and the determined Li concentration is problematic, given
the downward trend of results seen with increased exposure of the Li
ISE to SCA. We analyzed two AA Si standards on the Lytening, both
diluted to the same concentration of Si (10-fold dilution, 3.6 mmol/L);
the Sigma standard (pH 12.52) gave a Li concentration of 0.00 mmol/L
and the Perkin-Elmer standard (pH 3.23) 2.18 mmol/L. We then buffered
the Sigma standard with K2HPO4 (1.667-fold
dilution, 21.4 mmol/L Si) to pH 6.73. This specimen gave a Li
concentration of 0.61 mmol/L on the Lytening.
Noting this effect of pH on the result and realizing that we did not
have an optimal matrix, we then evaluated the effect of pH on serum. We
adjusted the pH of six different 5-mL aliquots from a Li-free serum
pool by adding 35 µL of various dilutions of 6 mol/L HCl and then
added increasing amounts of these aliquots to BDPs. The resulting serum
pools, ranging in pH from 6.0 to 8.5, were analyzed for Li on the
Lytening. Fig. 2
shows that, in a serum matrix in a more physiological range for
pH and ionic strength, the magnitude of the interference is inversely
related to pH. To investigate the response of the ISE to pH alone, we
also performed a pH titration with (Li-free) Dulbecco's Modified
Eagles Medium (Sigma), adjusting the pH of the Eagles Medium with
increasing amounts of 6 mol/L HCl (up to 50 µL added to 5 mL) to
obtain a range of pH from 7.68 to 1.91. Measuring each aliquot with the
Li ISE yielded Li results of 0.00 mmol/L from pH 7.68 to pH 5.07; at pH
4.63 and below, however, the ISE showed positive interference from pH.
The false concentrations for Li then increased exponentially until, at
the lowest pH (1.91), the reading exceeded the upper limit of linearity
of the analyzer for Li (4.5 mmol/L).
|
In several experiments, we attempted to recreate the conditions in the BDP. We added silicon dioxide powder (supplied by BD) to Li-free serum and to saline, both in glass tubes, and measured Li both immediately and after mixing on a rotator overnight. The powder was insoluble, and the Li concentration measured by the ISE was 0.00 mmol/L. We measured the pH of the two samples and found 6.31 for the saline and 8.56 for the serum. The pH of 1 mL of saline in a BDP was 6.70 (apparent Li 0.49 mmol/L), whereas the pH of 1 mL of a Li-free serum pool in a BDP was 8.45 (apparent Li 0.30 mmol/L). The pH of the serum harvested from 2 mL of fresh Li-free whole blood drawn into a BDP was 7.80 (apparent Li 1.03 mmol/L). We also evaporated Sigmacote siliconizing agent onto plain glass tubes and then added 1 mL of Li-free serum to the coated tube. The pH of this sample was 8.36 and the Li concentration determined with the ISE was 0.00 mmol/L.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials, Methods, and ResultsDiscussionReferences |
|---|
The effect of SCA on the determination of lithium with a Lytening 2Z ISE analyzer is inversely related to the volume and pH of the blood sample in the BDP. We documented an inverse relationship between the amount of Si in the serum specimen and the magnitude of the positive interference for lithium with the ISE. Even a nominally filled BDP or BDP-2 with Li-free blood can yield a reportable Li concentration. We were unable to demonstrate a stoichiometric relation for the interference because the response of the electrode changes from day to day with intermittent exposure to SCA. A comparison between the BDP and BDP-2 shows that the newer tube with less silicone surfactant does generate less interference, and that, when full, a BDP-2 can generate a reportable result.
Si preparations have already been shown to interfere with ISE determinations of ionized magnesium: Blood specimens obtained through tubing or syringes coated with silicone spuriously increased the concentration of ionized magnesium measured with an AVL 688-4 instrument, as documented by an increase in the millivolt reading from the electrode (5). Further studies documented that the Si preparation used as a sealant for the stopper in a BDP also caused interference (6). The increase in the ionized magnesium concentration was directly related to how long the serum remained in contact with the Si preparation. Exposing the ISE to samples containing Si preparations demonstrated a persistent residual effect on subsequent determinations of ionized magnesium. Thus, these observations for a magnesium ISE are similar to those observed for this lithium ISE.
The chemical configuration of Si that interferes with the ISE for Li requires further study. We documented that silica added to saline or serum had no effect on the Li ISE. However, not having access to the surfactant used to fix the silicone to the BDP or BDP-2, we are unable to test this component of the SCA. Tests with the BDP-2, which contains less silicone surfactant than the original tube (BDP), showed that the amount of interference decreased by more that half. This suggests that the silicone surfactant is the interferent. Although the ISE begins to respond to pH alone at values <5, we documented that AA standards for Si buffered to pH 6.73 generated a positive Li interference with the ISE. We speculate that the surface area for contact between the Si preparation and blood, and the pH, are important for the molecular configuration of Si to act as an interferent. Further studies are needed to determine the precise compound and mechanism for interference with the ISE determination of Li.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials, Methods, and ResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  R. A.R. Bowen, Y. Chan, M. E. Ruddel, G. L. Hortin, G. Csako, S. J. Demosky Jr, and A. T. Remaley Immunoassay Interference by a Commonly Used Blood Collection Tube Additive, the Organosilicone Surfactant Silwet L-720 Clin. Chem., October 1, 2005; 51(10): 1874 - 1882. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A.R. Bowen, Y. Chan, J. Cohen, N. N. Rehak, G. L. Hortin, G. Csako, and A. T. Remaley Effect of Blood Collection Tubes on Total Triiodothyronine and Other Laboratory Assays Clin. Chem., February 1, 2005; 51(2): 424 - 433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Drake, R. A.R. Bowen, A. T. Remaley, and G. L. Hortin Potential Interferences from Blood Collection Tubes in Mass Spectrometric Analyses of Serum Polypeptides Clin. Chem., December 1, 2004; 50(12): 2398 - 2401. [Full Text] [PDF]
|
|
|
|
|
|
M. W. Linder and P. E. Keck Jr. Standards of laboratory practice: antidepressant drug monitoring Clin. Chem., May 1, 1998; 44(5): 1073 - 1084. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
