Determination of serum uric acid by isotope dilution mass

measurements agree with the principal measurements, we have strong evidence for the absence of measurement bias. Uric acid was determined In three ...
1 downloads 0 Views 759KB Size
Download PDF
Anal. Chem. 1990, 62, 2173-2177

2173

Determination of Serum Uric Acid by Isotope Dilution Mass Spectrometry as a New Candidate Definitive Method Polly Ellerbe,* Alex Cohen, Michael J. Welch, and E d w a r d White V National Institute of Standards a n d Technology, Gaithersburg, Maryland 20899

A new Isotope dllutlon mass spectrometrlc method for urlc acld Is described. A known weight of [1,3-15N,]url~acld Is added to a known welght of serum, and the mixture ls allowed to equlilbrate. The serum Is put through an anlon-exchange resln, and the Isolated urlc acld Is converted to the tetrakls(tert-butyldlmethylsllyl) derivative of urlc acld. For measurement, the derlvatlve ls Injected Into a gas chromatograph Interfaced with a low-resolution, magnetic sector mass spectrometer. Isotope ratlo measurements are made from the abundances of the [M tert-butyl]+ Ions at m / r 567 and 569. Blas Is Investigated by measurlng the urlc acld level In the same samples under dlfferent chromatographiccondltlons and wlth dlfferent lonlzatlontechnlques. I f these confirmatory measurements agree with the prlnclpal measurements, we have strong evidence for the absence of measurement bias. Uric acld was determlned In three lyophlllred human serum pools by this method. For Standard Reference Materlal (SRM)909, four sets of Six samples each were prepared. For Candklate SRM 909a, which consisted of two pools, each wlth a dlfferent level of urlc acld, six sets of two samples of each level were prepared. The coefflclent of variation for a single measurement ranged from 0.34% to 0.42%, while the relative standard error of the mean ranged from 0.08% to 0.14%. The results from the Confirmatory measurements demonstrated that there was no slgnlflcant blas In the measurements. The combination of hlgh preclslon and absence of slgnlflcant blas In the results quallfles this method as a candldate deflnltlve method as deflned by the Natlonal Committee for Cllnlcal Laboratory Standards.

-

INTRODUCTION Uric acid is an end product of the catabolism of certain nucleosides. Hyperuricemia (serum uric acid levels above 7 mg/dL in men and above 6 mg/dL in women) can occur in association with a number of conditions and, if untreated, can lead to gout and renal disease. Therefore, serum uric acid concentration is one of the most frequently performed clinical measurements. An accuracy base in the form of serum-based materials with known concentrations of uric acid can provide clinical laboratories with an important component of ensuring that patients’ samples are accurately measured. To this end, the need exists for a method of demonstrated accuracy and high precision, i.e., a definitive method, for the value assignment of standards. There is no universally accepted set of requirements that a method must meet to be called definitive. However, the National Committee for Clinical Laboratory Standards has published tentative guidelines for definitive methods ( I ) , which give defined rules for the acceptance or rejection of a given method as being definitive. Isotope dilution mass spectrometry (IDMS) has been the technique of choice for definitive methods a t the National Institute of Standards and Technology (NIST, formerly the National Bureau of Standards), since it does not depend on sample recovery, shows high precision, and can be tested for

bias and unknown interferences. IDMS methods for organic analytes involve spiking a sample with a labeled version of the analyte as an internal standard, processing the sample, and then measuring the ratio of unlabeled to labeled analyte by use of gas chromatography/mass spectrometry (GC/MS). Assuming complete equilibration, less than complete recovery after spiking does not affect the measured concentrations, unless there is a significant isotope effect, since it is the ratio of unlabeled to labeled analyte that is measured. Generally, deuterium is the only stable isotope routinely used in labeling organic compounds for which significant isotope effects can be seen. For this work, our labeling is with 15N. Although the probability of a significant measurement interference is low when an isolation procedure followed by capillary column GC/MS is used, it still would be possible for a substance to coelute with the uric acid, contribute to the ion intensity measured for either the unlabeled or labeled form, and thus interfere with the measurement. T o test for such an interference, we selected a subset of samples from those already measured a t the principal ion and remeasured them in two ways: with a different polarity GC column and at a different pair of ions generated by use of a different ionization method. These sets of measurements are called the confirmatory measurements. If an interference were present that biased the principal measurements for a given sample, it should cause one or both of the confirmatory results to be different. Therefore, an interference would go undetected only if it had the same retention time as uric acid on each GC column and yielded the same ions in the same proportions as uric acid in each method of ionization. Such a situation is unlikely. Organic serum analytes for which definitive methods have been developed a t NIST include cholesterol ( Z ) , glucose (3), urea ( 4 ) ,and creatinine (5). Other laboratories have published methods that they describe as definitive for cortisol ( 6 4 , cholesterol (9), creatinine (IO), and glucose (11). We have previously presented another uric acid candidate definitive method ( I Z ) , which was used to certify uric acid in Standard Reference Material (SRM) 909. This method demanded a very complicated sample preparation procedure. Also, since that time, we have made many instrument modifications and have replaced packed columns with capillary columns in the gas chromatograph. Thus, when the need to recertify uric acid arose, we took the opportunity to develop a new, simpler definitive method for uric acid. A proposed definitive method for uric acid has been reported by Siekmann (13). This method used ammonium hydroxide a t 0.015 M with a 1 2 0 1 mole ratio of ammonium hydroxide to uric acid to dissolve the uric acid. Uric acid has been shown to degrade in solutions at this mole ratio with this concentrat,ion of ammonium hydroxide (14), resulting in the possibility of significant bias. No degradation was observed when 0.001 M ammonium hydroxide with a mole ratio of 1.7 ammonium hydroxide to uric acid was used, as in this report. The method described by Siekmann also involved the use of the trimethylsilyl derivative of uric acid for measurement. No column memory test was reported, even though column memory is a possible source of significant error for compounds

0003-2700/90/0362-2173$02.50/0 0 1990 American Chemical Society

2174

ANALYTICAL CHEMISTRY, VOL. 62, NO. 20, OCTOBER 15, 1990

with silyl groups attached t o nitrogen. We have found a column memory effect with this uric acid derivative. Therefore, we chose t o use the tert-butyldimethylsilyl derivative of uric acid, for which we found no memory effect. Finally, for a method t o be definitive, an absence of bias in the results must be demonstrated. In the method as described by Siekmann, no measurements were reported that specifically demonstrated the absence of bias. Our candidate definitive method for uric acid fulfills the stringent requirements of a definitive method ( I ) . A known weight of [ 1,3-15N2]uricacid dissolved in 0.001 M ammonium hydroxide is added to a known weight of serum and is allowed to equilibrate with the naturally occurring material. The serum is passed through an anion-exchange resin, and the isolated uric acid is converted to the tetra[tert-butyldimethylsilyl] derivative of uric acid by reaction with N-(tertbutyldimethylsily1)-N-methyltrifluoroacetamide(MTBSTFA) containing 1% imidazole, plus an equal volume of acetonitrile. The derivative is injected into a gas chromatograph equipped with a nonpolar capillary column directly connected t o the ion source of a low-resolution, magnetic sector mass spectrometer. Isotope ratio measurements are made from the abundances of the [M - tert-butyl]+ ions at m/z 567 and 569. Standards are made by combining and derivatizing known amounts of unlabeled uric acid and [1,3-15N2]uric acid. Standards with isotope ratios slightly lower and slightly higher than that of each sample are measured immediately before and after the sample. Use of this measurement technique (bracketing) produces results of high precision (2,5,12).When the uric acid level was measured in the same samples under different chromatographic conditions and with a different ionization technique, we found no evidence of significant bias in the measurement process.

EXPERIMENTAL SECTION Materials. Samples of SRM 909 and Candidate SRM 909a, both of which are lyophilized human sera, were obtained from the Office of Standard Reference Materials (OSRM), NIST. The material for SRM W9a is being analyzed and will be issued when SRM 909 is out-of-stock. SRM 913, uric acid with a certified purity of 99.7 O . l % , was obtained from OSRM. [1,3-15N,]Uric acid with an isotopic purity of 99 atom % 15Nwas obtained from Merck and Co. (St Louis, MO). The labeled uric acid was purified by dissolving the material in aqueous lithium carbonate, followed by precipitation with acetic acid and drying for 24 h at 100 "C and 4 h at 200 "C. Ammonium hydroxide was reagent grade, and its molarity was calculated from assay (NH3 at 29.4%) and the specific gravity of 0.90 to be 15.6 M. For chromatography, the acetate form of a strong anion-exchangeresin (200-400mesh) with an 0.8 mequiv/mL capacity (AG1-X2) was obtained from Bic-Rad (Richmond, CA). Gold weighing boats were prepared by cutting and folding gold sheet. Other chemicals were reagent grade. Standards Preparation. For the measurement of uric acid in SRM 909, two independent sets of standards were prepared from mixtures of known quantities of SRM uric acid and labeled uric acid, each dissolved in ammonium hydroxide. Solutions of 0.001 M ammonium hydroxide were prepared 24 h in advance by dilution of 0.015 M ammonium hydroxide, which itself was prepared from the concentrated reagent and boded distilled water. After 5-7 min, the pH of all preparations was between 9.82 and 9.87. The flasks used for dissolving the uric acid were equipped with a micro stir bar (13 X 3 mm) and were sealed with a hollow polyethylene stopper. SRM uric acid was dissolved in 0.001 M ammonium hydroxide, such that the mole ratio of ammonium hydroxide to uric acid was 1.7 and the concentration of uric acid in these standard solutions was about 100 pg/mL. Labeled uric acid was dissolved in the same manner. The uric acid solutions were stirred for 100 min to ensure dissolution. Sampling of these solutions was done by weighing aliquots in capped 3-mL plastic syringes that were conditioned three times with the solution before an aliquot was taken for measurement, according to a sampling procedure previously described (2). About 2 mL (weighed to 0.01 mg) of the labeled uric acid was dispensed

*

into a series of 18 mm X 15 cm borosilicate glass test tubes. Various amounts of the SRM uric acid solution were added (accurately weighed) such that the weight ratio (unlabeled/labeled) varied from about 0.8 to 1.12. The solutions were mixed, diluted with 10 mL of water, mixed again, then divided in half. Half of the solution was stored at -10 "C for future use; the other half was lyophilized at 133 Pa (1Torr). The solid material was first transferred with a spatula to a 0.3-mL vial that had a conical bottom and then was derivatized. For the measurement of SRM 909a, the frozen half of the sets of standards was thawed and used. In addition, two new sets of standards were prepared in the same manner as above. All sets of standards were cross-checked with each other. Sample Preparation. Four sets of six samples each of SRM 909 human serum were prepared, with duplicate aliquots taken from three bottles for each set. The SRM 909 samples, which are lyophilized pellets, were reconstituted by weight; i.e. according to procedure A of the SRM 909 Certificate of Analysis, except that when a vial of serum was weighed and 10 mL of water was added, the amount of water added was determined by weight (15) rather than by volume. A fresh solution of labeled uric acid, consisting of about 5 mg (accurately weighed) of labeled uric acid in 50 mL (accurately weighed) of ammonium hydroxide (1.7 mole ratio of 0.001 M ammonium hydroxide to uric acid) was prepared and stirred 100 min until all solids were dissolved. Weighed aliquots of the ammonium hydroxide solution of labeled uric acid were added to weighed aliquots of about 1mL of the reconstituted sera, which contained 77-94 pg of uric acid, such that the unlabeled/labeled uric acid weight ratios were in the range 0.9-1.1. Each serum sample was mixed by gentle swirling with a vortex mixer immediately after addition of the labeled uric acid spike. About 0.9 mL (calculated to maintain the 1.7:1 ammonium hydroxide to uric acid mole ratio) of 0.001 M ammonium hydroxide was added to convert the natural uric acid in the serum to the ammonium salt. The walls of the tubes were washed down with 10 mL of water, and the solutions were mixed. Samples were left to equilibrate overnight at room temperature. To recover the uric acid from the sera, the mixtures were passed through columns of the anion-exchangeresin. Columns, prepared from Pasteur pipets without the top constriction, were filled to a height of about 3.3 cm. This was the equivalent of about 670 pmol, or 0.83 mL, of the resin. The column was first washed with 25 mL of water; the last 5 mL of water had a conductivity of 2.6 ppm or less as NaCl. The serum samples were then passed through the column, followed by the rinsings from the sample tube (5 X 2 mL portions of water). The column was then washed with 25 mL of water, and finally the uric acid was eluted with 15 mL of 1 M acetic acid. The eluate was collected in a 100-mL flask, and the solution was lyophilized at 133 Pa (1 Torr). The solid product was then transferred with a spatula to one of the 0.3-mL vials described above. Recovery was about 85%, estimated by comparing the peak area of labeled uric acid in samples to that in standards. At the same time that the serum sample was spiked, a standard solution of SRM uric acid was prepared. The same procedure w a used ~ for the SRM uric acid as for the labeled uric acid solution. Two aliquots were taken from this solution and were spiked with the same labeled uric acid solution used to spike the serum samples. These samples, used as additional check standards, were processed in the same manner as standards. Derivatization. The derivatizing reagent, MTBSTFA with 1% imidazole, was added to each vial containing uric acid at a mole ratio of about 3W.1 reagent to the calculated total of labeled and unlabeled uric acid, followed by the addition of an equal volume of acetonitrile. The vial was then capped and heated for 48 h at about 60 O C . The concentration of uric acid in the derivatizing solution was about 100 mg/L. Equilibration Study. To test for complete equilibration of labeled uric acid and endogenous uric acid in serum, we spiked a serum sample with a solution of labeled uric acid. After mixing, the solution was divided into nine aliquots, which were allowed to equilibrate for different periods (2,4,8,20,26,31, 44, 50, and 68 h) and were then processed as described above. GC/MS Instrumentation. The instrumentation consisted of a gas chromatograph combined with a single focusing magnetic mass spectrometer, controlled by a data acquisition system de-

ANALYTICAL CHEMISTRY, VOL. 62, NO. 20, OCTOBER 15, 1990 2175 signed for isotope ratio measurements (5). Electric switching was used to switch between the masses (16). GC/MS Conditions. The principal measurements were made at the masses of the [M - tert-butyl]+fragment ions at m / z 567 and 569 from electron ionization (EI)on a nonpolar GC column. Confirmatory measurements were made by using the same ion pair but using a moderate polarity GC column, and by using the [M + H]+ ions at m / z 625 and 627 from ammonia chemical ionization (CI) on the nonpolar column. For measurement under E1 conditions, the mass spectrometer was operated at 72 eV with an emission current of 0.5 mA and an ion source temperature of 200 "C. For measurement under ammonia CI conditions, the emission current was 1 mA, the source manifold pressure (ionization gauge) was 8 X Pa (6 X lod Torr), the analyzer pressure was 1 x Pa (8 X Torr), and the source temperature was 200 "C. An adjustable splitter was located at the front of the column, and the end of the column was placed directly into the source. For the principal measurements, and the confirmatory ammonia CI measurements, the gas chromatograph was equipped with a 15 m long, 0.32 mm i.d., nonpolar [(95%)-dimethyl(5%)-diphenyl-polysiloxane] fused silica capillary column of 1-pm film thickness (J & W Scientific Inc., Folsom, CA). The splitter was set to a vent to column ratio of 40:1, and the gas chromatograph was operated at a temperature of 255 "C with a helium flow rate of 3 mL/min. The injection port and the interface to the mass spectrometer were maintained at 250-270 "C. The usual injection was 1 p L of sample or standard. Under these conditions, the retention time for the uric acid derivative was about 5 min, and the gas chromatographic peak was usually about 10 s wide. Confirmatory measurements were also done on masses at m / z 567 and 569 on a 30 m long, 0.32 mm i.d., intermediate polarity [ (50%)-methylphenyl-plysiloxane]fused silica column of 0.5-pm film thickness (J & W Scientific Inc., Folsom, CA). The column temperature was 275 "C, with other conditions remaining unchanged; the retention time of the uric acid derivative under these conditions was about 7 min. The number of sweeps per measurement cycle was set at 6, and the number of cycles across each chromatographic peak was 30-40 (16, 17). Measurement Protocol. For the measurement of each sample, two standards were chosen: one whose ion intensity ratio was slightly lower than that of the sample and one whose ion intensity ratio was slightly higher. Each standard and sample was measured twice in succession. The two observed intensity ratios were acceptable only if they agreed within 0.5%; if not, a third measurement was made, which had to agree with one of the other two and the three were averaged. The average conwithin 0.570, stituted one valid measurement. The calculated areas for the six peaks of each standard-sample-standard group all had to be within a factor of 2 of each other, or the measurement was discarded. If a standard was used again in any given half-day, only a single measurement was made at each use if the new ion intensity ratio was within 0.5% of the previous value for that standard. Measurements were made in the following order: lower weight-ratio standard, sample, higher weight-ratiostandard. Thus, each measurement of sample was immediately bracketed both in time and ratio by measurements of standards. On a second day the order of standards was reversed, and the measurement process repeated. The weight ratios for each sample from both days had to agree within 0.5%, or a third day's measurement was done, which had to agree with one of the other day's measurements within 0.5%, or all measurements were discarded. The quantity of analyte in the sample was calculated by linear interpolation of the measured ratio of the sample between the measured ratios of the standards, whose weight ratios were known. The unlabeled material has an isotope peak at m / z 569 that overlaps the labeled material measurement peak at m/z 569, but because our standards bracket our samples so closely in weight ratio, the maximum error associated with the overlap is under 0.2% (18).

RESULTS A N D D I S C U S S I O N Equilibration Study. The addition of an isotopically labeled material to a serum matrix may not immediately result in the complete equilibration of the labeled form with the endogenous form. This equilibration is necessary for accurate measurement. The time required for equilibration depends

567

100; 1

-.-+-.-

OL.-L-4iL__f_i_L___

100

150

200

250

300

350

400

4.-.

450

mh

Figure 1. Electron impact spectra of the uric acid derivative. The mass spectrum is of the unlabeled material: tetrakis(tert-butyldi-

methylsily1)uric acid. on the matrix and on the nature of the particular analyte and affects the results obtained for that analyte. We studied the equilibration of endogenous uric acid with labeled uric acid and found the equilibration to be complete in 2 h, and the ratio remained unchanged for at least 68 h. Samples were equilibrated overnight. Choice of Derivative. T o introduce uric acid into the gas chromatograph, it is necessary to convert it into a suitable derivative. In the original uric acid definitive method, isomers of the tetraethyl derivatives (12) were formed, and were separated by following a tedious chromatographic procedure. Unfortunately, measurement of different isomers gave slightly different results. For this work, other derivatives were investigated. The trimethylsilyl derivative was tried, but a GC column memory effect was found. Furthermore, when measurements of ion intensity ratios were made for repeated injections of this uric acid derivative from one sample, the results showed poor precision (>I%CV). We then tried the tert-butyldimethylsilyl derivative. This derivative has satisfactory GC properties, with no column memory effect found, and it has intense high-mass ions, an important factor in making inteference-free measurements. We were able to obtain excellent precision in ion intensity ratios with this derivative and thus chose it for our method. Electron Ionization Mass Spectrum. The E1 mass spectrum for the uric acid derivative, tetrakis(tert-butyldimethylsily1)uric acid, is shown in Figure 1. The base peak, a t mlz 567, results from the loss of a tert-butyl group. This ion and the corresponding ion for the labeled uric acid a t m / z 569 were chosen for the ratio measurements because of their intensity and the lack of interferences at these masses. An attempt was made to use the molecular ion a t m / z 624 and 626 for confirmatory measurements, but the precision of ratio measurements a t this ion pair was poor (about 0.5%), so the attempt was abandoned. The ion a t m / z 609 results from the loss of a methyl group, and the ion a t mlz 73 is the trimethylsilyl cation, which results from rearrangement (19). Memory Effects. We tested the derivative for column memory effects. If a memory effect is present, injections of a sample or standard of one unlabeledllabeled weight ratio will affect the intensity ratio measured for subsequent injections of sample or standard. We injected the unlabeled uric acid derivative, then the labeled derivative, and then the unlabeled derivative and measured each intensity ratio. The intensity ratios for the injections of the unlabeled material were not significantly different. The lack of effect on the measured intensity ratios, even when the weight ratio differences between consecutive measurements were the most extreme and thus any memory effect should have been the most evident, provides strong evidence of the absence of column memory effects. S t a n d a r d s Cross-Check. The accuracy of results for serum samples is limited by the accuracy of the standard mixtures for calibration. For definitive methods we prepare a t least two independent sets of standards and test each standard by bracketing with standards from another set. The

2176

ANALYTICAL CHEMISTRY, VOL. 62, NO. 20, OCTOBER 15, 1990

Table I. Test of the Consistency of Two Independent Sets of Standards (SRM Uric Acid and [ 1,3-15N,]UricAcid)

std"

bracketed byb 1, 2 2, 4 4, 6 4, 6 6, 7 6, 7 6, 7 7, 9

weight ratio (unlabeled/labeled) measured by IDMS weighed-in 0.8955 0.9402 1.0127 1.0358 1.1050 1.1275 1.1175 1.1594

diff,' %

0.8959 0.9390 1.0128 1.0398 1.1074 1.1265 1.1132 1.1552

-0.04 +0.13 -0.01 -0.38 -0.22 +0.09 +0.38 +0.36

mean

Table 111. Determination of Uric Acid in Human Serum Candidate SRM 909a (mmol L-I g-')" material/ vial'

set 1

set 2

low/l low/2 high/3 high14 909/5 909/6

0.3430 0.3426 0.5592 0.5592 0.5675 0.5667

0.3420 0.3443 0.5624 0.5593 0.5668 0.5694

111 2/ 1 3i2 412 513 61 3

mean

cv, 70

set 1

concentrationb set 2' set 3

0.5618 0.5701 0.5657 0.5690 0.5694 0.5706

0.5647 0.5672 0.5672

0.5678 0.59

set 4

0.5648

0.5657 0.5685 0.5660 0.5679 0.5678 0.5661

0.5667 0.5677 0.5681 0.5680 0.5668 0.5672

0.5660 0.25

0.5670 0.21

0.5674 0.10

Statistical Summary overall mean = 0.5670 CV of single measurementd = 0.42% relative standard error of the mean = 0.084% Four independently prepared sets, 3 vials per set, 2 samples per vial. One valid measurement, as defined in the text. Samples 2 and 5 from set 2 did not have enough signal to measure. dCV of a single measurement is calculated as (s2(sets)+ s2(via1s) s2(aliquots) + s2(measurement))'i2.

+

weight ratio determined by the IDMS measurements is then compared with the weighed-in weight ratio for that standard. The agreement between these values for each member of a set allows determination of the presence or absence of a bias between sets. For the measurement of uric acid in SRM 909, two independent sets were prepared and tested by bracketing. For the measurement of SRM 909a, two additional sets of standards were prepared, and all four sets were used for the measurements. All sets were cross-checked; set one was checked against set two, set two against set three, and set three against set four. Biases were -0.03%, 0.04%, and 0.15%. Table I shows the results of one such cross-checking. Check Standards. We have previously shown that uric acid is destroyed in the presence of 0.015 M ammonium hydroxide (14). The uric acid is not appreciably destroyed in solutions where the ammonium hydroxide to uric acid mole ratio is kept below 3.4 and the ammonium hydroxide concentration is 0,001 M. At the same time that sample sets of SRM 909 were prepared, additional check standards were also prepared to determine whether any decomposition had occurred during preparation of the samples. If decomposition had occurred, the experimental weight ratios of the additional check standards would not agree with the weighed-in weight ratios. The experimental and weighed-in weight ratios of the additional check standards did agree (two aliquots of one

set 6

0.3427 0.3429 0.3423 0.3422 0.5565 0.5571 0.5585 0.5582 0.5651 0.5644 0.5660 0.5657

0.3406 0.3405 0.5563 0.5542 0.5642 0.5650

Low Level overall mean = 0.3423 CV of single measurementd = 0.34% relative standard error of the mean = 0.12%

"Standards treated as samples were from standard set 4. bStandards used as standards were from standard set 3. ePercent difference is calculated as [(measured value - weighed value) x 1001/weighed value.

sample/vial

0.3412 0.3422 0.5582 0.5588 0.5645 0.5648

set 5

Statistical Summary

+0.04

Table 11. Determination of Uric Acid in Human Serum SRM 909 (mmol L-I g

concentrationb set 3 set 4

High Level overall mean = 0.5582 CV of single measurementd = 0.37% relative standard error of the mean = 0.14%

" Six independently prepared sets. One valid measurement, as defined in the text. c * L ~ wrepresents n samples of the low level of uric acid in SRM 909a. "High" represents samples of the high level of uric acid in SRM 909a. "909" represents the control samples of SRM 909. dCV of a single measurement is calculated as (s2(sets) + s2(vials) + s2(aliquots) + s2(measurement))1/2. Table IV. Confirmatory Measurements on Samples from SRM 909 Sera (mmol L-' g-l) concentrationn EI'

cId

set

sample

principalb

1 2 2 3 3 4

6 3

0.5706 0.5672 0.5647 0.5679 0.5661 0.5668

0.5695 0.5696 0.5664 0.5684 0.5689 0.5673

0.5689 0.5675 0.5644 0.5667 0.5684 0.5659

0.5672 0.35

0.5683 0.22 +0.19%

0.5670 0.29 - 0.04%

1 4

6 5 mean

cv, % % difference from principal measurement

aOne valid measurement, as defined in the text. bPrincipal measurement using E1 ions generated at m / z 567 and 569 and a nonpolar fused silica GC column. Confirmatory measurements using E1 ions generated a t m / z 567 and 569 and an intermediate polarity fused silica GC column. Confirmatory measurements using ammonia CI ions generated a t m / t 625 and 627 and a nonpolar fused silica GC column. -___

check standard per set for a total of eight values for four sets, mean difference 0.18%, SD = 0.14%) thus indicating no decomposition of uric acid. For the measurement of Candidate SRM 909a, it was decided that the precaution of preparing check standards was no longer necessary. Serum Results. The results of the principal measurements for uric acid in SRM 909 are shown in Table 11. Four sets of six samples each (three vials per set and two samples per vial) were prepared and analyzed. The CV for a single measurement is 0.42%. The previous certified value of 0.570 f 0.003 mmol L-' g-' (15) is comfirmed. The results for Candidate SRM 909a are shown in Table 111. Six sets of two vials of each level were prepared. The CV for a single measurement for the low level is 0.34% and for the high level is 0.37%. These results demonstrate the excellent within- and between-set precision obtainable with this method. The grand mean of SRM 909 samples run as controls with the Candidate SRM 909a sets is 0.5658 mmol L-' g-', with a relative standard error of the mean of 0.094%, demonstrating that these sets

ANALYTICAL CHEMISTRY, VOL. 62, NO. 20, OCTOBER 15, 1990

Table V. Confirmatory Measurements on Samples from Candidate SRM 909a Sera (mmol L-*g -') concentrationa set

vial

1 4 6

1

2 2 mean

cv, 70

principal*

Low 0.3430 0.3423 0.3405 0.3419 0.38

70 difference from principal measurement

2 3 5

3 4 4 mean

cv, o/o

High 0.5624 0.5588 0.5582 0.5598 0.40

EI' 0.3438 0.3424 0.3414 0.3425 0.35

CId

+0.18

0.3434 0.3444 0.3404 0.3427 0.61 +0.23

0.5582 0.5574 0.5576 0.5577 0.07 -0.37

0.5625 0.5596 0.5570 0.5597 0.49 -0.02

70 difference from principal measurement aOne valid measurement, as defined in the text. *Principal measurement using E1 ions generated a t m / z 567 and 569 and a nonpolar fused silica GC column. Confirmatory measurements using E1 ions generated a t m / z 567 and 569 and an intermediate polarity fused silica GC column. Confirmatory measurements using ammonia CI ions generated a t m / z 625 and 627 and a nonpolar fused silica GC column.

are in control. The results from the confirmatory measurements are shown in Tables IV and V and demonstrate that there is no significant measurement bias. The results in Tables 11-IV are in millimoles of uric acid per liter of reconstituted serum per gram of lyophilized serum. These concentrations are based on the addition of 10.00 mL of water a t 22 "C to each vial of lyophilized serum for reconstitution. Measurements of other analytes in this SRM have shown that the dry material is homogeneous but that the fill weights of lyophilized serum vary from vial to vial (fill-weight CV = 0.5%) (15). Therefore, expressing the concentration per gram of lyophilized serum corrects the results for fill-weight variation. Error Analysis. While the imprecision of the method is small and no evidence of significant bias in the measurement process was found, an analysis of possible sources of bias and imprecision was made. Errors in the standards would contribute to errors in the determination of uric acid levels in serum. The cross-checking of standard sets ensured the absence of significant error. Furthermore, we expect that the large numbers of measurements made on each serum pool, using standards from different sets, would reduce the effect of random error. Isotope effects in the derivatization reaction could lead to bias and imprecision. With [ 1,3-15N2]uricacid as the labeled material, the small differences in results observed between independently prepared serum sets provide evidence for the absence of significant isotope effects in the derivatization, and the coelution from the GC column of the unlabeled and labeled derivatives provides evidence for the absence of a significant

2177

isotope effect on the column. Since standards and samples are derivatized and measured under as nearly identical conditions as we can achieve, residual isotope effects, if any, should cancel out.

CONCLUSIONS The combination of high precision and evidence for the lack of significant bias qualifies this method as a candidate definitive method for uric acid. Use of this method in certifying uric acid levels in reference materials will permit evaluation of the accuracy of both reference and routine clinical methods. This method has been used to certify the concentration of uric acid in a human serum pool that will be issued as SRM 909a and to confirm the certified value of uric acid in SRM 909 by a second definitive method.

ACKNOWLEDGMENT We gratefully acknowledge the assistance of Robert Watters and Susannah Schiller with the statistical analysis and that of Jean-Marc Rodier with the measurement of the mass spectra.

LITERATURE CITED

-

Tentative Guklelines for the Development of Definnive Methods in C/in ical Chemistry for the National Reference System in Clinical Chemist ~NRSCC1-T; : National Committee for Clinical Laboratory Standards: Villanova, PA 1982. Cohen, A.; Hertz, H. S.; Mandel, J.; Paule, R. C.; Schaffer, R.; Sniegoski, L. T.; Sun, T.; Welch, M. J.; White, E., V. Clin. Chem. 1980, 26, 854-860. White, E., V; Welch, M. J.; Sun, T.; Sniegoski, L. T.; Schaffer, R.; Hertz, H. S.; Cohen, A. Biomed. Mass Spectrom. 1982, 9 , 395-405. Welch, M. J.; Cohen, A.; Hertz, H.; Ruegg, F. C.; Schaffer, R.; Sniegoski, L. T.; White, E., V. Anal. Chem. 1984, 56, 713-719. Welch, M. J.; Cohen, A.; Hertz, H.; Ng, K. J.; Schaffer, R.; Van Der Lijn, P.; White, E., V. Anal. Chem. 1986, 58, 1681-1685. Siekmann, L.; Breuer, H. J. Clin. Chem. Clin. Biochem. 1982, 20, 883-892. . . ~ Jonckheere, J. A.; De Leenheer, A. P. Biomed. Mass Spectrom. 1983. 10. 197-202. Patterson; D. 0.; Patterson, M. 6.; Culbreth, P. H.; Fast, D. M.; Holler, J. S.; Sampson, E. J.; Bayse, D. D. Clin. Chem. 1984, 30, 619-626. Wright, L. A.; Breckenridge, W. C. Clln. Chem. 1987, 33, Pelletier, 0.; 1403- 1411. Siekmann, L. J. Clin. Chem. Clin. Biochem. 1985, 23, 137-144. Pelletier, 0.;Arratoon, C. Clin. Chem. 1987, 33, 1397-1402. Cohen, A.; Hertz, H. S.; Schaffer, R.; Sniegoski, L. T.; Welch, M. J.; White, E., V. Presented at the 27th Annual Conference on Mass Spectrometry and Allied Topics, Seattle, WA, June 3-8, 1979. Siekmann, L. J. Clln. Chem. Clin. Blochem. 1985, 23, 129-135. Ellerbe, P.; Cohen, A.; Welch, M. J.; White. E., V. Clin. Chem. 1988, 34, 2280-2282. NIST Certificate of Analysis for SRM 909; Office of Standard Reference Materials, N I S T Gaithersburg. MD 20899. Ellerbe, P.; Meiselman, S.; Sniegoski, L. T.; Welch, M. J.; White, E., V. Anal. Chem. 1989, 61, 1718-23. Matthews, D. E.; Hayes, J. M. Anal. Chem. 1976, 48, 1375-1382. Yap, W. T.; Schaffer, R.; Hertz, H. S.; White, E., V.; Welch, M. J. Btomed. Mass Spectrom. 1983, 10, 262-264. Phillipou, G. Org. Mass Spectrom. 1977, 12, 261.

RECEIVED for review March 21,1990. Accepted July 20, 1990. We gratefully acknowledge the support of P.E. by the College of American Pathologists. Certain commercial equipment, instruments, or materials are identified in this report to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the equipment, instruments, or materials are necessarily the best available for the purpose.