Determination of Lead in Urine and Whole Blood by ... - CiteSeerX

CUN. CHEM. 40/8,

1494-1502

(1994)

#{149} Drug

Monitoring

and Toxicology

Determination of Lead in Urine and Whole Blood by Stable Isotope Dilution Gas Chromatography-Mass Spectrometry Suresh K. Aggarwal,’

Michael Kinter,3 and David A. Herold”4’5

A stable isotope dilutiongas chromatography-mass spectrometry (GC-MS) method is described for the determination of lead (Pb) in urine and whole blood. The use of lithium bis(trifiuoroethyl)dithiocarbamate

Pb(FDEDTC)

as

a chelating agent showed strong memory effect, restricting the range of Pb isotope ratios that can be measured in unknown samples. To overcome this carryover problem, we further derivatized the Pb(FDEDTC)2 chelate with 4-fluorophenyl magnesium bromide to form Pb(FC6H4)4. The sequential analyses of solutions of natural Pb and enriched 204Pb with Pb(FC6H4)4 chelate by GC-MS demonstrated no observable memory effect. Precision and accuracy of Pb isotope ratio measurements with Pb(FC6H4)4 were established, and the isotope dilution GC-MS method was validated by determining Pb concentrations in urine standards from the National Institute of Standards and Technology, urine and blood reference materials from the New York State Department of Health, and blood Pb survey samples from the College of American Pathologists. IndexingTerms: toxicology/chelatingagents/chelationtherapy Lead (Pb) poisoning is a significant health problem. The historical widespread use of Pb in paint and gasoline has distributed Pb throughout the environment. Pb from the air may be inhaled directly or deposited by aerosol into food and water and enter the body by ingestion. Additionally, individuals can be overexposed either occupationally or accidentally, leading to Pb toxicity. The adverse effects of chronic long-term exposure to Pb have been well documented (1) and include gastrointestinal disturbances, anemia, insomnia, weight loss, motor weakness, muscle paralysis, and nephropathy. Pb may also have long-term effects on cognitive and motor functions at blood concentrations significantly below those associated with the above symptoms (2-4). A recent study by Needleman et al. (5) demonstrated that children who had moderately increased Pb concentrations in early childhood later had a 740% increase in school drop-out rates, a 580% increase in reading disability, and lower high school class standing. The federal government estimates that >3 million preschool ‘VA Medical Center, Laboratory Services, 3350 La Jolla Village Dr., San Diego, CA 92161. 2Fuel Chemistry Division, Bhabha Atomic Research Center, Trombay, Bombay 400 085, India. 3Department of Pathology, University of Virginia, Charlottesville, VA 22908. ‘Department of Pathology, University of California-San Diego, San Diego, CA 92037. for correspondence. Fax 619-552-7479. Received August 9, 1993; accepted March 25, 1994. 1494

CLINICALCHEMISTRY,Vol. 40, No. 8, 1994

children have dangerously increased Pb concentrations. Many health experts now recognize that low-level Pb poisoning (100-150 jtgfL) causes neurotoxicity in children. In December 1989, the Environmental Protection Agency’s Science Advisory Board concluded (6) “there is likely to be no threshold for lead neurotoxicity, at least within the contemporary range of blood lead levels (10100 gfL).” Further, the 1990 advisory group pointed out that “the value of 100 tg/L refers to the maximum blood-lead level permissible for all members of their groups, and not mean or median values” (7). Known biochemical properties of Pb have convinced many exports that Pb has no threshold, but instead has a continuum of toxicity. In addition to neurotoxicity, Pb has been classified as a “probable human carcinogen” (7). There are also reports on the adverse effects of Pb on the reproductive system (8) and a link between low-level Pb exposure and an increase in blood pressure (9). The problem of Pb toxicity is complicated because Pb affects different people unequally. One of the chief variables appears to be the different intake of essential trace elements such as Ca,, Fe, Zn, or P (10). Although the determination of Pb in whole blood is important for monitoring Pb exposure, urine determinations may provide a noninvasive method of screening and monitoring Pb. Pb measurements in urine are also used to assess the effect of chelation therapy and for Pb metabolic studies. The average urine Pb concentrations in Pb workers are reported to be 120 ig/L and 140 p.gfL during exposure of 0.15 mg/m3 and 0.20 mglm3, respectively. In comparison, an average urine Pb concentration of 40 gIL (range 10-190 JLgIL) in 60 adult males in urban northwestern England has been reported (11). For analysis of Pb at concentrations encountered in biological samples (tg/L), the two most commonly used methods are electrothermal atomic absorption spectroscopy (EAAS) (12) and anodic stripping voltammetry (13).6 However, isotope dilution mass spectrometry can be used as a definitive analytical method because it minimizes matrix effects and obviates the need for quantitative recovery of analyte. Thermal ionization mass spectrometry (TIMS), which is commonly used in geochronology (14), has been used

6Nonstandard abbreviations: EAAS, electrothermal atomic absorption spectroscopy; TIMS, thermal ionization mass spectrometry; ICP-MS, inductively coupled plasma mass spectrometry; GCMS, gas chromatography-mass spectrometry; Em,electron ionization; SIM, selected-ion monitoring NIST, National Institute of Standards and Technology; Li(FDEDTC), lithium bis(trifluoroethyl)dithiocarbaniate; 4-FPMB, 4-fluorophenylmagnesium bromide; SRM, Standard Reference Material; and CAP, College of American

Pathologists.

in nuclear technology (15) for the determination of isotopic composition and concentration of Pb. TIMS has also been used for the isotopic analysis of Pb in airborne particulates (16) and as a definitive method for comparing interlaboratory results for blood Pb (17). Additionally, inductively coupled plasma source mass spectrometry (ICP-MS) appears promising for the determination of trace elements in biological specimens (18). The availability of organic mass spectrometers in clinical, biomedical, and environmental laboratories has led to the investigation of fast atom bombardment mass spectrometry (19) and gas chromatography-mass spectrometry (GC-MS) (20) for trace-element determination in biological samples. In our laboratory, we have developed GC-MS for the determination of trace metals in biological samples. We have reported methods for the determination of Ni (21), Cr (22), Co (23), Pt (24, 25), Cu (26), Se (27), and Cd (28) by isotope dilution GC-MS. Here, we report an isotope dilution GC-MS method for the measurement of Pb in urine and whole blood that includes the use of a derivatized chelating agent, with no resulting memory effect. We also validated our method with various Reference Materials.

Materials and Methods Instrumentation The GC-MS system consisted of a double-focusing, reverse-geometry mass spectrometer (Model 8230; Finnigan MAT, San Jose, CA) coupled to a gas chromatograph (Varian 3700; Varian Associates, Walnut Creek, CA). The mass spectrometer was equipped with a SpectroSystem 300 data system for on-line data acquisition and processing. The instrument was operated in the electron ionization (El) mode with 70 eV electrons, a source temperature of 200#{176}C, the conversion dynode at -5000 V, and the secondary electron multiplier at 2400 V. Data were acquired in the selected-ion monitoring (SIM) mode with voltage peak switching. The gas chromatograph was equipped with a DB-1 (J. W. Scientific, Rancho Cordova, CA) poly(dimethylsioxane) bondedphase fused silica capillary column, 10 m x 0.32 mm, 0.25-pm film thickness. Samples were injected with an on-column ixjector (OCI-3; Scientific Glass Engineering, Austin, TX) at an oven temperature of 100#{176}C, followed by a 35#{176}C/mm ramp to 300#{176}C. The GC-MS interface was at 280#{176}C. High-purity He was used as the carrier gas.

Reagents The 204Pb-enriched lead nitrate (70.94 atom% 204Pb) used as an internal standard for isotope dilution was obtained from Oak Ridge National Laboratory (Oak Ridge, TN). Certified Atomic Absorption Standard (lead nitrate in dilute HNO3) purchased from Fisher Scientific (Fair Lawn, NJ) was used as the primary standard. Double subboiling quartz-distilled HNO3 in Teflon bottles was obtained from the National Institute of Standards and Technology (NIST, Gaithersburg, MD). Ultrex-grade ammonium hydroxide solution (30%) was purchased from J. T. Baker Chemical Co. (Phillipsburg,

NJ), and stabilized H202 (50%) was obtained from Fisher Scientific. Lithium bis(trifluoroethyl)dithiocarbamate [Li(FDEDTC)] was synthesized by reacting bis(trifluoroethyl)amine (PCR, Gainesville, FL) and n-butyl lithium (Aldrich Chemical Co., Milwaukee, WI) in an inert atmosphere at -70#{176}C and then adding carbon disulfide (Aldrich) (20, 29). 4-Fluorophenylmagnesium bromide (4-FPMB), the Grignard reagent (2.0 molfL solution in diethyl ether), and anhydrous diethyl ether were also obtained from Aldrich. Several precautions were necessary to minimize the potential for Pb contamination from the apparatus, reagents, personnel, and the laboratory environment. Prior to sample preparation, the acetate buffer was treated with a 0.1 mol/L solution of Na(DEDTC) and extracted with CH2C12 to remove any Pb contamination. The organic extract was discarded. The concentrations of Pb, determined by EAAS, in various reagents were as follows: H202, 0.6 j.tg/L; 40 mL of 30% solution of ammonium hydroxide per liter, 0.03 igfL; pH 3 acetate buffer, 0.3 .tgfL; CH2C12, 0.5 j.igfL. Because Pb contamination would not be a problem after the formation of Pb(FDEDTC)2, the 4-FPMB and anhydrous diethyl ether were not specifically analyzed for Pb. The overall blank was -1 ng/sample, because of the volumes of these different reagents used in the procedure for digestion and chelate formation. This overall blank limits the applicability of this technique at extremely low concentrations. We selected Li(FDEDTC) as the first step to derivatize Pb, with the objective of carrying out GC-MS analysis by using Pb(FDEDTC)2. This proved unsuccessful and therefore we reacted these preformed derivatives with Grignard reagent to produce Pb(FC6H4)4. Pb(FC6H4)4 can also be prepared by using the cominercially available sodium diethyldithiocarbamate or ammonium pyrrolidine dithiocarbamate.

Specimens Standard Reference Material (SRM) 2670, freezedried urine, was purchased from NIST and prepared according to their directions. Urine and blood reference materials were obtained from the New York State Department of Health (Patrick J. Parsons), and wholeblood Pb survey samples were obtained from the College of American Pathologists (CAP).

Preparationand Calibrationof Internal Standard A 204Pb solution was prepared by dissolving lead nitrate in deionized water and adding a few drops of concentrated HNO3. Diluted solutions were prepared from this stock solution by weight. The isotopic composition of Pb in this solution was determined experimentally by GC-MS analysis of Pb(FC6H4)4 chelate. The internal standard solution was calibrated by reverse-isotope dilution GC-MS by using the natural Pb primary standard. Weighed amounts of the primary standard solution were mixed with weighed amounts of 204Pb internal standard. Chelates were prepared from the supplemented samples (as described below) and were used for mass spectrometric determination of m/z 489:49 1, 489:

CLINICALCHEMISTRY, Vol. 40, No. 8, 1994 1495

492, and 489:493 isotope 2#{176}6Pb, 2#{176}4Pb/207Pb, and concentration of Pb in was calculated from the tope ratios, the weights 204Pb internal standard abundances of Pb, and primary standard.

ratios, corresponding to 2#{176}Pb/ 204Pb/2#{176}8Pb, respectively. The the internal standard solution experimentally determined isoof the Pb primary standard and solutions, the measured ion the concentration of Pb in the

Evaluation of Memory Effect Memory effect for Pb with Li(FDEDTC) as a chelating agent was evaluated by using two samples with 10 jig/L and 100 pg/L Pb. The geometric mean of these two concentrations is 32 gfL. Ideally, the ratio of the two m/z peaks being compared should be unity to provide optimal results. The use of the geometric mean ensures that the ratios of the naturally occurring isotopes to the added 204Pb are equally distant from unity at both the lower and upper extremes. This provides optimal results over the range of analysis. Thus, the optimum concentration of internal standard was calculated as 10 gfL 204Pb when mixing equal volumes of the sample and the internal standard solutions. The two Pb samples were, therefore, mixed with an equal volume of 204Pb internal standard having a total Pb concentration of 10 pg/L. The samples were derivatized to Pb(FDEDTC)2 and analyzed for m/z 460:464 isotope ratio. Based on the results of this experiment, the Pb(FDEDTC)2 chelates were further derivatized to Pb(FC6H4)4 and analyzed for the m/z 489:493 isotope ratio. This was followed with a more rigorous approach for Pb(FC6H4)4, which involved the replicate sequential analyses of solutions of natural Pb, enriched 204Pb, and again the natural Pb solution, measuring the mlz 489:493 ratio. Unne Pb Standard Addition Known amounts of Pb standard solution (128-605 gfL) were added to NIST SRM 2670 urine (low concentration), which has a noncertifled Pb value of 10 gfL. Pb concentration was determined by using the peak areas of mlz 491, 492, and 493 relative to the m/z 489 internal standard peak area. Digestion of Urine Samples and Chelate Formation A known volume (1 mL) of the reconstituted urine Reference Material was mixed with a weighed amount of 204Pb solution in a Teflon beaker. The amount of 204Pb in the internal standard solution added to the urine sample was optimized to obtain an isotope ratio in the mixture corresponding to the geometric mean of isotope ratio m/z 489:493 in the sample and the internal standard. The optimum addition of the internal standard is not mandatory but is preferred so as to obtain the best concentration determination by limiting random errors, discussed below, in the isotope ratio measurements. The supplemented urine samples were treated with 1 mL of concentrated HNO3 and allowed to stand at room temperature for some time (-1 h). The sample solution was subsequently heated at 50#{176}C on a hot plate to reduce the 1496

CUNICAL CHEMISTRY, Vol. 40, No. 8, 1994

volume to -50 ,L. Then 100 tL of 50% H202 was added. The solution was again heated gently at -50#{176}C on the hot plate and inspected periodically. The contents were mixed and the beaker was tapped gently to disperse frothing. The digestion with H202 was performed four to five times until a white residue remained when the solution was completely evaporated. The total time required was 3-4 h. The dry residue was dissolved in 2 mL of deionized water and the solution transferred to a clean polypropylene centrifuge tube with a conical bottom. The solution was extracted with 2 mL of CH2C12, discarding the organic phase to remove undigested lipids. Pb(FDEDTC)2 chelate was formed at pH 3, from 1 mL of an acetic acid-sodium acetate buffer (1.2 mL of acetic acid + 100 mL of 10 mmoJJL sodium acetate solution) and 100 L of a 20 mmolIL solution of Li(FDEDTC) in deionized water. The samples were vortexmixed for 2 rain and the Pb(FDEDTC)2 chelate formed was extracted with 1 mL of toluene. The organic extract containing the Pb chelate was evaporated to dryness at 60#{176}C under a stream of argon gas in the laminar-flow hood. The dried residue was redissolved in 200 jiL of diethyl ether and -200 L of 4-FPMB was added. The solution in the tube was shaken to allow complete mixing. Excess Grignard reagent was destroyed by adding 100 L of 100 mLIL isopropyl alcohol in toluene and then adding 1 mL of 1 moIJL HNO3. The Pb(FC6H4)4 chelate formed was extracted with 1 mL of toluene, evaporated to dryness, and reconstituted in 10-50 L of CH2C12 for GC-MS analysis. Digestion of Whole-Blood Samples A known volume (1 mL) of each of the blood samples was mixed with a known amount of 2#{176}4Pb internal standard solution. Then 2 mL of 4 mol/L HNO3 (250 mLJL) was added to each of the samples drop-wise with constant vortex-mixing. The solutions were centrifuged for 30 min, and the supernates were digested with HNO3 + H202 and taken for chelate formation as described above for the urine samples.

GC-MS Before Pb isotope ratio measurements, the focusing conditions of the mass spectrometer were optimized and mass calibration was established by using perfluorokerosene. The Pb isotope ratios were measured in duplicate by injecting 1 pL of the chelate solution and monitoring the isotopic cluster corresponding to the fragment ion formed by the loss of one ligand (M LY. Data were acquired in an SIM experiment and quantified by using the integrated chromatographic peak areas. The software parameters for the mass spectrometer operation were defined to perform voltage peak switching for SIM at a rate of 2 Hz, yielding 20 measurement cycles across the 10-s-wide chromatographic peaks. In the SIM experiments, the source and the collector slits were adjusted to obtain trapezoidal peak shape with flat peak tops at a mass resolution of 1000. Because the measurements were performed without a lock mass, we found it necessary to experimen-

-

tally determine each day the m/z value of the peak maximum by obtaining a histogram of the ion current at various m/z values adjacent to the calculated m/z value of the most abundant ion. This was done by measuring the ion current at intervals of 0.1 amu, up to 0.5 amu on either side of the calculated mlz for the ion being monitored. Subsequently, the mlz values given for various ions of interest in the data acquisition corresponded to these experimentally determined mlz values. Further, during the measurements, the ion current was also monitored at m/z values ± 0.05 amu from the experimentally determined position of the peak maximum. This served as a check for any change in the mass calibration.

Calculationof Concentration The following two equations were used to calculate the concentration of Pb in the 204Pb internal standard solution and the urine or blood sample, respectively.

C

-

C814 [W1, [W (1(R R,./R81,) R) (at.wt.)81, (at.wt.) (% (%intensity intensity ofj)1 of i)81,] -

(1)

C

-

C

[Wp (at.wt.)8 (% (%intensity intensity [W8 (1(R,.- Rm/Lp) R8) (at.wt.),

of i),] ofj)5] (2)

As can be clearly seen from Eqs. 1 and 2, the calibration of the internal standard solution (C) and the determination of Pb concentration (C8) in the unknown sample both involve the Pb atomic weights and intensities of various Pb isotopes present as ions in the chelate used. These factors are canceled, since the internal standard calibration is also performed by reverse-isotope dilution mass spectrometry with a primary standard of natural Pb. Also, any mass discrimination factor among the different isotopes, which may occur in the ion source, analyzer, or detector of a mass spectrometer, is canceled. Pb determination in any unknown sample involves only the experimental measurement of Rm in Eq. 2; all other factors are predetermined. Hence, one can determine C8 with high accuracy by using isotope dilution technique, since it is not affected by nonquantitative recovery and the presence of matrix material. Equations 1 and 2 were also used for the intensity ratios m/z 489:491 and 489:492, corresponding to 204Pb: 206Pb and 204Pb:207Pb, respectively.

Results and Discussion Pb(FDEDTC)2 The El mass spectrum of Pb(FDEDTC)2 in Fig. 1 shows two groups of ions containing Pb isotopic information. These ions are the molecular ion, Pb(FDEDTC)2, and a fragment ion, Pb(FDEDTC), designated M and (M LY, respectively. The most abundant isotopic group is the (M LY ion at nominal mlz 464. This ion group consists of nominal m/z values 460, 462, 463, and 464, corresponding to 2#{176}’Pb, 206Pb, 207Pb, and 2#{176}8Pb isotopes, respectively, in the ion Pb(FDEDTC). The results of the memory effect study for specimens that span the range of the low-concentration and increased-concentration NIST urine specimens (10-109 j.g/L), determined by using Pb(FDEDTC)2 chelate, are shown in Fig. 2. The memory effect observed with these two samples, with m/z 460:464 differing by a factor of 10, puts restrictions on the difference in the isotope ratios that can be measured accurately by GC-MS during sequential analyses of various samples. This effectively limits use of the Li(FDEDTC) chelating agent to a nar-

In these equations, C denotes the concentration (g/L or ng/g) of Pb in the sample, W is the sample size (volume or weight), R specifies the measured intensity ratio of ion i to that of ionj, at.wt. is atomic weight of Pb, and the subscripts s, std, sp, and m denote the unknown sample, the standard used for calibration, the 204Pb supplement, and the supplemented mixture, respectively.

Thus the equations include the measured isotope ratio 489:493 in the supplemented mixture (Rm), the amount of 204Pb as m/z 489 in the internal standard, the amount of 208Pb as m/z in the sample, and the differences in the atomic weights of 204Pb internal standard and the natural Pb in the sample. m/z

100

-

464

80 60

462

40 20 460

388 (M) 720

100

600

200 m/z

700 Fig. 1. Electron ionization mass spectrum of Pb(FDEDTC)2.

CLINICALCHEMISTRY,Vol. 40, No. 8, 1994 1497

1.0

Pb(FC6H4)4

Pb(FDEDTC)2 Mixture6

The El mass spectrum of Pb(FC6H4)4 (Fig. 3) exhibits an intense fragment ion peak, designated (M LY, containing Pb isotopic information, which we used in these experiments. Fig. 4 is a SIM trace showing the peaks at m/z 489, 491, 492, and 493, corresponding respectively to #{176}Pb, 206Pb, 207Pb, and 208Pb isotopes in the Pb(FC6H4)3 ion in the mass spectrum. Tables 1 and 2 present the results obtained for isotope ratio measurements of natural Pb and the 204Pb-enriched internal standard, respectively, with Pb(FC6H4)4 chelate. For these measurements, 1 j.&Lof the chelate solution containing -10 ng of Pb was used for GC-MS analysis. The atom percent abundances shown in Table 1 for the different isotopes in natural Pb are the recommended values based on the values measured experimentally by different laboratories (30). The atom percent abundances shown in Table 2 for the different isotopes in enriched 204Pb are the values provided by the Oak Ridge National Laboratory. The calculated values for the abundances of Pb-containing isotopic cluster fragment ion, Pb(FC6H4)3, are also given in these Tables. These values have been obtained by calculating the contributions of the isotopes of Pb and C in the fragment ion. As can be seen from the data in Tables 1 and 2, there is -

0.6 0

Ce 0.4 .0

a-

Mixture 5

Ce

Iliiiiu

0

N

a-

0.2

0

N

0.0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21

Analysis Number Fig. 2. Evaluation of cross-contamination in consecutive analyses of t Pb(FDEDTC)2 samples with isotope ratio differing by a factorof -8.

row range of Pb concentrations or else requires extensive resupplementing of samples on the basis of their initial measurements. To avoid these problems, we investigated the reaction of Pb(FDEDTC)2 with a Grignard reagent, 4-FPMB, to form a more stable organolead compound, Pb(FC6H4)4, for GC-MS analysis. 125

100 113

149

100 80

80

60 40 20

60

40

208

248

266

(M-3t.)

*

20

4i8

100

150

200

250

300

350

400

450

500

m/z 100.000

im

493

m.=

492

Fig. 3. Electron ionization mass spectrum of Pb(FC6H4)4.

50.000

ooo 20.000 40,000

m./z= 491

20,000

0

500

4:30

Retention 1498

Time (mm)

CLINICAL CHEMISTRY, Vol. 40, No. 8, 1994

5:30

6:00

Fig. 4. SIM trace showing the intensities of Pb-containing ions in Pb(FC6H4)4.

Table 1. Calculated and measured abundances of different ions in natural Pb(FC6H4)4. Fragment Ion Pb(FC,H,)31 Calculated Atom % Isotope

abundance,

Measured abundance,b

Ion (m/z

abundance

1.31 1.27 1.4 489 23.26 491 21.96 206Pb 24.1 492 24.57 24.04 22.1 207Pb 51.38 493 52.20 208Pb 52.4 #{149} Including the contributionsof the isotopes of Pb and carbon in the ion Pb(FC9H4)3. b

Not corrected for any mass discrimination factor.

Table 2. Calculated and measured abundances of different ions In enriched 204Pb(FC0H4)4. Fragment Ion Pb(FC8H4)3 Calculated Atom % Isotope

abundance,b

abundance8

204pb

70.94

±

Measured

abundance,c

Ion (im’z)

0.10

206Pb 12.68±0.10 207Pb 6.42 ± 0.10 208Pb 9.96 ± 0.10 a Values given by Oak Ridge

489

67.19

491 492

13.32 8.59

10.90 493 National Laboratory.

67.08 13.51 8.61 10.80

‘ Including the contributions of the isotopes of Pb and carbon in the ion Pb(FC6H4)3. #{176} corrected for any massdiscrimination factor.

good agreement among the calculated and measured abundances of Pb isotopes in the fragment ion in both natural Pb and the enriched 204Pb. No correction has been applied to the data shown in these Tables for any mass discrimination among the different isotopes. Any mass discrimination factor would be canceled in an isotope dilution experiment in which the internal standard solution is calibrated in the same experiment. Precision in Isotope Ratio Measurements

Precision in the determination of Pb isotope ratios by using Pb(FC6H4)4 was evaluated by performing measurements of chelated natural Pb. Five replicate analyses of 10 ng of Pb were made on each of 3 days. The results are summarized in Table 3. The mean values for

each day were used to calculate the standard deviation of the mean of means; this value is referred to as between-run precision in Table 3. The within-run precision was calculated by using the standard deviation values obtained on the individual days. Overall precision was calculated by combining the within-run precision and between-run precision values. This was done to include the effects of any variations in the mass spectrometer operating parameters that might affect the quality of isotope ratio data from one day to another. Overall precision values of 1.5% to 8.7% were obtained. As expected, precision is better for the isotope ratios mlz 491:493 and 492:493 than for the isotope ratio mlz 489: 493 because of the low abundance (1.4 atom%) of 204Pb in natural Pb. Calibration of 2#{176}4Pb Internal Standard

The 204Pb solution was calibrated by reverse-isotope dilution with a primary standard of natural Pb. Replicate samples were prepared by mixing weighed aliquots of primary standard and 2#{176}4Pb solution to achieve an optimum isotope ratio m/z 489:493 in the supplemented mixtures. The three sets of isotope ratios, m/z 489:491, 489:492, and 489:492, corresponding respectively to 2#{176}4Pb:206Pb, 204Pb:207Pb, and 2#{176}4Pb:200Pb in the fragment ion Pb(FC6H4)3, were used for calculation of the Pb concentration in the 204Pb internal standard solution. There was good agreement in the Pb concentration values obtained with all the isotope ratios (Table 4). To support these data, we investigated the use of atomic absorption; however, the Pb concentration in the 204Pb internal standard solution could not be accurately measured by atomic absorption because of the isotopic effect (31).

Evaluationof Memory Effect One of the problems with the GC-MS of metal chelates is the carryover between sequential analyses of samples of different isotopic ratios. This carryover, referred to as memory effect, can adversely affect the accuracy of data on samples with altered isotopic ratios and must be investigated for each individual metal, chelating agent, and GC-MS system employed. Figure 5 presents the results of the memory effect

Table 3. Precision in isotope ratio determination of natural Pb as Pb(FC6H4)4. Isotope ratlo,

mean ± SD

489/493 491/493 492/493 5) 0.0232 ± 0.0005 0.4506 ± 0.012 0.4652 ± 0.016 = 5) 0.0264 ± 0.0011 0.4532 ± 0.010 0.4672 ± 0.003 = 5) 0.0271 ± 0.0021 0.4546 ± 0.016 0.4713 ± 0.007 Mean of means 0.0256 0.4528 0.4679 Within-run precision,6 % 3.2 1.6 1.3 Between-run precision, % 8.1 0.4 0.7 Overall precision,c % 8.7 1.7 1.5 Notcorrectedfor any mass discrininationfactor. b Calculated with the formula S, = (,Sj2)h/2/n, where s1representsthe standarddeviation obtained on each Individual day and n Is the total numberof days. Overall precision (S,) was calculated by combiningwithin-runprecision(S and between-runprecision(S8)accordingto the formula S = (8? + S.2).

Day 1 (n Day 2 (n Day 3 (n

=

CLINICAL CHEMISTRY, Vol. 40, No. 8, 1994

1499

Table 4. Determination of Pb in 204Pbsupplement solution by reverse-Isotope dIlution GC-MS

6.5

Spe

Pb conc, &g/g of solution

489/491

0.497

±

SDb 0.010

CV, % 2.1

489/492 489/493

0.502

±

0.005

0.9

0.488-0.512 0.497-0.507

0.497

±

0.018

3.6

0.478-0.521

Ratio

Mean ±

Range

Ilni

#{149} Expected Pb (based on weight of supplement dissolved and diluted) = 0.49 g/g of solution. b n = 4; each sample was measured by duplicate injections into GC-M8. Natural

Natural

lull

0.03

! 11111 123456789101112131415

Analysis Number Fig. 6. Evaluation of cross-contamination in consecutiveanalysesof natural Pb and enriched 204Pb internal standard by using Pb(FC6H4)4 with m/z 489:493 differing by a factor of -300.

of the results by using measured ion ratios as close to unity as practical for the measurement range. Comparison

of Isotope Ratios Determined

by GC-MS

and

ICP-MS 1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16

Analysis Number Fig. 5. Evaluation of cross-contamination in consecutive analyses of

two Pb(FC6H4)4sampleswithisotope ratio differing by a factor of -8.

study for specimens that span the range of the lowconcentration and increased-concentration NIST urine specimens (10-109 j.tgfL) with 4-FPMB as derivatizing reagent. As can be seen, no appreciable memory effect was observed for either of the analysis sequences of these mixtures. Further, the constancy of the isotope ratios determined by replicate injections of each mixture also demonstrates the absence of any significant memory effect under these conditions. This should be compared with Fig. 2, in which Pb(FDEDTC)2 with mlz 460:464 differing by a factor of 10 restricted the difference in the isotope ratios that could be measured accurately by GC-MS during sequential analyses of various samples. The results for sequential analyses of a solution of natural Pb and the enriched 204Pb are shown in Fig. 6. Even in these two samples with ion ratio m/z 489:493 differing by a factor of 300, no appreciable memory effect was observed. This is, however, an extreme case that would never occur in the analyses of urine and wholeblood specimens. Good experimental design would maintain the mlz 489:493 isotope ratio in different sampies in a range that included the expected limits of Pb concentrations. Ideally, if an approximate idea of the Pb concentration is available, the optimum addition of the internal standard solution would increase the accuracy 1500

CLINICALCHEMISTRY,Vol. 40, No. 8, 1994

The GC-MS method with 4-FPMB as a derivatizing reagent was also validated by comparison of the isotope ratios obtained by ICP-MS. For this purpose, four synthetic mixtures were prepared by mixing primary standard solution and enriched 204Pb solution as previously described. A comparison of the results for the m/z 489: 493 isotope ratio calculated from the data obtained by ICP-MS and those determined by GC-MS is shown in Table 5. The calculated ratios for the m/z 489:493 in Pb(FC6H4)4 chelate were obtained by using the experimentally determined 204Pb, 206Pb, 207Pb, and 208Pb abundances in synthetic mixtures by ICP-MS and including the contributions of C isotopes. The mean measured/calculated ratio was 1.011 ± 0.023, demonstrating the close agreement between the two methods. This enhances confidence in the accuracy of isotope ratio data obtained by GC-MS. Results on Urine and Whole-Blood

Samples

The results of a standard addition experiment are shown in Table 6. The expected concentrations were

Table 5. Comparison of Pb Isotope ratios in synthetic samples by ICP-MS and GC-MS. m/z 489:493 In Pb(FC5K4)4 Synthetic mIxture

20Pb208Pb by ICP-MS

Calculated

Measured by GC-MS

Measured/ calculated 1.042

1

0.3177

0.290

0.302 ± 0.008

2

0.4884

0.445

0.452

±

3

0.8269

0.751

± 0.019

4

1.1511

0.752 1.045

1.032

±

ICP-MSvaluesdeterminedby NIST.

0.008 0.034

1.015 0.999

0.988

Table 6. Determination of added Pb (1&g/L)In urine. Conc based on Ion ratios

Table 8. Determination of Pb (ig/L) in blood by isotope dilutIon GC-MS.

Expected Sample

1 2 3 4 5

ConcentratIon based on Ion ratios

conc 128 203 357 476

489:491

489:492

489:493

123

121

128

605

342 456

203 349 458

189 349 459

568

580

596

196

Mean ± SD EAAS

196 ± 7

Sample BL-06

347

BL-07

232

124±4 4 458 ± 2 581±14 ±

Known amountsof Pb standard solution were addedto NIST SAM 2670 urine (noncertified value, 10 oJL).

calculated on the basis of the Pb added to the lowconcentration NIST urine sample. The concentrations determined by averaging the three ion ratios compared well with the expected concentrations. Regression analysis of expected concentration (x), which includes added Pb and the Pb originally present, and the measured mean concentration from the different ion ratios (y), gave the following equation: y = 0.96 (± 0.01)x + 0.21 (± 0.18) (n = 5, r = 1.00). A regression analysis of the expected values and the concentrations based on only the ion ratio m/z 489:493 gave a comparable equation: y = 0.99 (± 0.02)x + 0.44 (± 0.68) (n = 5, r = 0.999). Table 7 shows results for Pb determination in reference materials from NIST and from the New York State Department of Health. Human Urine-17 and Urine-20 were samples obtained from Pb-exposed children who were subjected to chelation with calcium disodium EDTA. The analyses were run in duplicate on all three samples. The values determined by isotope dilution GC-MS were in good agreement with the expected concentrations provided by MIST and New York State. Blood samples from the CAP and goat blood samples from the New York State Department of Health were also analyzed to determine the Pb concentrations by isotope dilution GC-MS. The results for Pb determination by using various ion ratios (Table 8) were compared with values obtained by EAAS and the results of the CAP survey. Isotope dilution GC-MS provided good precision and results compared well with the EAAS and CAP data. The signal-to-noise ratio observed in the reconstructed ion chromatogram indicated a potential limit of detection of 0.1 gfL. However, contamination from re-

Tabie 7. DeterminatIon of Pb (pgIL) in urine by isotope dIlution GC-MS. Concentration Reference materlal

based on Ion ratios

Expected

conc

489:491

489:492 489:493

Mean ± SD

SRM 2670,

117 107

99 119

105 119

107 115

Human Urine-17 150 ± 50

174

162 154

164 ±

158

155 156

Human Unne-20 400

443

442

445

156 ± 443 ±

437

447

454

446 ±

increased

a

109±4

±

60

± ±

9 7 10 2 2 9

Two independent samples of each were taken for complete preparation

and analysis.

95

489:491 489:492 489:493 Mean ± SD 91 221

BL-08

264

293

BL-09

419 330

431 336 156 423

BL-10

GB-26

150 ±

GB-27

430

40

± 40

108 233 293 432 338

165 441

111 226 292 439 328 159 428

11 6

103

±

227

±

293

± 1

434 ± 4

334 160 431

± 5 ± ±

CAP 93.7

206.0 281.1 427.2 341.0

5 9

Target values. Samples BL-06-BL-10 were blood Pb survey samplesfrom CAP (1992); samples GB-26 and 27 were goat blood samplesfrom the New York Dept. of Health;EAAS values were providedby the Dept. of Pathology,Univ. of Virginia Health Science Center, Charlottesville VA. a

agents and laboratory makes quantitation with reasonable

1 pg/L a practical precautions.

limit

of

In conclusion, isotope dilution GC-MS with 4-FPMB as a chelating agent can be used for precise and accurate determination of Pb concentrations in urine and blood. The method has the potential of providing reference values free of matrix effects. This was clearly established by precision studies, memory effect evaluation, and isotope ratio comparison with NIST ICP-MS values. The biological measurements were validated with NIST urine, New York State Department of Health Pb reference materials for urine and blood, and CAP blood Pb survey samples. We thank P. J. Paulsen, J. D. Fassett, and J. R. Moody of NIST for the ICP-MS results; Kathy Linkswiler for the EAAS results; and James

Nicholson

and Patrick

K. Anomck

for reagent

and

sample preparation. The authors also thank J. Savory and M. R. Wills for their interest in the present work and allowing use of the facilities in the trace metals laboratory. Funding for the purchase of the mass spectrometer was obtained from the National Institutes of Health, Division of Research Resources Shared Instrumentation Grant Program, Grant no. 1-S10-RRO-2418-01. Additional funding from the John Lee Pratt Fund of the University of Virginia is gratefully acknowledged. S. K A. thanks the Division of Experimental Pathology, Department of Pathology, University of Virginia Health Sciences Center, for support, and the authorities at Bhabha Atomic Research Center, Bombay, for granting leave. References 1. Putnam RD. Review of toxicology of inorganic lead. Am md Hyg Assoc J 1986;47:700-.3. 2. Needleman I-IL, Gunnoe C, Leviton A, Reed R, Peresie H, Cornelius M, Barrett P. Deficits in psychologic and classroom performance of children with elevated dentine lead levels. N Engl J Med 1979;300:689-95. 3. Charney E. Sub-encephalopathic lead poisoning central nervous system effects in children. In: Chisolm JJ Jr, O’Hara DM, eds.

Lead absorption in children. Baltimore: Urban and Schwarzenberg, 1982:35-42. 4. Boeckx RL. Lead poisoning in children. Anal Chem 1986;58: 274A-86A.

5. Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The long-term effects of exposure to low doses of lead in childhood. N Engi J Med 1990;263:83-8. 6. Environmental Protection Agency, Scientific Advisory Board. Report of the clear air scientific advisory committee: review of the OAQPS lead staff paper and the ECAO air quality criteria document supplement. Doc. no. EPA-SAB-CASAC-90-002, 1990:1.

CUNICALCHEMISTRY,Vol.40, No.8, 1994 1501

7. Environmental Protection Agency, Scientific Advisory Board. Report of the joint study group on lead: review of lead carinogenicity and EPA scientific policy on lead. Doe. no. EPA-SAB-EHC90-001, 1989:1. 8. Needleman HL, Landrigan PJ. The health effects of low level exposure to lead. Ann Rev Publ Health 1981;1981:277-98. 9. Victery W, Tyroler HA, Volpe R, Grant LD. Summary of discussion sessions: symposium on lead-blood pressure relationships. Environ Health Perspect 1988;78:139-55. 10. Centers for Disease Control. Preventing lead poisoning in children. At1anta US Department of Health and Human Services! Public Health Service, 1991:1-80. 11. Baselt RC, Cravey RH, eds. Disposition of toxic drugs and chemicals in man. Chicago: Year Book Medical Publishers, 1989: 450-3. 12. Boeckx RL. Lead in whole blood by fiameless atomic absorption spectrophotometiy. In: Fringe CS, Faulkner WR, eds. Selected methods of clinical chemistry. Washington, DC: AACC Press,

1986:75-8. 13. Roda SM, Greenland 1W, Bornschein RL, Hammond PB. Anodic stripping voltammetry procedure modified for improved accuracy in blood lead analysis. Clin Chem 1988;34:563-7. 14. Russel 1W, Farquhar RM. Lead isotopes in geology. New York: Interscience, l960:243pp. 15. Mathews CK, Jam HC, Kavimandan VD, Aggarwal SK Tracer techniques in the input accountability of plutonium in reprocessing plants. Nucl Technol 1979;42:297-303. 16. Boone 5, Hearn T, Lewis S. Comparison of interlaboratory results for blood lead with results from a definitive method. Clin Chem 1979;25:389-93. 17. Facehetti 5, Geiss F. Isotopic lead experiment: status report, Luxembourg Commission of European Communities; l982:ll4pp. 18. Olivares JA. Inductively coupled plasma-mass spectrometry. Methods Enzymol 1988;158A.205-22. 19. Jiang X, Smith DL. Quantitation of stable isotopic tracers for

calcium by fast atom bombardment mass spectrometry. Anal Chem 1987;59:2570-4. 20. Aggarwal SK, Kinter M, Wills MB, Savory J, Herold DA. Determination of isotope ratios of chromium, nickel, zinc and copper by gas chromatography-mass spectrometry by using volatile metal chelates. Anal Chim Acta 1989;224:83-95.

1502

CLINICALCHEMISTRY,Vol.40, No.8, 1994

21. Aggarwal SK, Kinter M, Wills MR, Savory J, Herold DA. Isotope dilution gas chromatography/mass spectrometry for the determination of nickel in biological materials. Anal Chem 1989; 61:1099-103. 22. Aggarwal SIC, Kinter M, Wills MR, Savory J, Herald DA. Determination of chromium in urine by stable isotope dilution gas chromatography/mass spectrometiy using lithium bis(trifluoroethyl)dithiocarbamate as a chelating agent. Anal Chem 1990;62: 111-5. 23. Aggarwal SK, Kinter M, Herold DA. Determination of cobalt in urine by gas chromatography-mass spectrometry employing nickel as an internal standard. J Chromatogr Biomed Appl 1992; 576:297-304. 24. Aggarwal SIC, Kinter M, Herald DA. Isotope dilution-gas chromatography-mass spectrometry for platinum determination in urine. J Am Soc Mass Spectrom 1991;2:85-90. 25. Aggarwal SK, Gemma NW, Kinter M, Nicholson J, Shipe JR Jr, Herald DA. Determination of platinum in urine, ultrafiltrate and whole plasma by isotope dilution gas chromatography-mass spectrometry compared to electrothermal atomic absorption spectrometry. Anal Biochem 1993;210:113-8. 26. Aggarwal SIC, Kinter M, Herald DA. Determination of copper in urine and serum by gas chromatography-mass spectrometry.

Anal Biochem 1991;194:140-5. 27. Aggarwal SIC, Kinter M, Herald DA. Determination

of selenium in urine by isotope dilution gas chromatography-mass spectrometry using 4-nitro-o-phenylenediamine, 3,5-dibromo-o-phenylenediammne, and 4-trifluoro-methyl-o-phenylenediammne as derivatizing reagents. Anal Biochem 1992;202:367-74. 28. Aggarwal SK, Orth RG, Wending J, Kinter M, Herald DA. Isotope dilution gas chromatography/mass spectrometry for cadmium determination in urine. J Anal Tox 1993;17:5-10. 29. Sucre L, Jennings W. Lithium di(trifluoroethyl)dithiocarbamate: an alternative reagent for the preparation of di(trifluoroethyl)dithiocarbamate metal chelates. Anal Lett 1980;13:497-501. 30. DeBievre P, Barnes IL. Table of the isotopic composition of the elements as determined by mass spectrometry. mt j Mass Spectrom Ion Proc 1985;65:211-30. 31. Brimhall GH. Measurement of lead isotopes by differential atomic absorption. Anal Chem 1969;41:1349-51.