Analysis of Perchlorate in Dried Blood Spots Using Ion

Dried blood spots (DBS) offer a matrix for assessing neonatal perchlorate exposure. DBS are collected from 98% of newborns in the United States as par...
0 downloads 0 Views 92KB Size
Download PDF
Anal. Chem. 2009, 81, 1931–1936

Analysis of Perchlorate in Dried Blood Spots Using Ion Chromatography and Tandem Mass Spectrometry Samaret M. Otero-Santos, Amy D. Delinsky, Liza Valentin-Blasini, Jarad Schiffer,† and Benjamin C. Blount* Centers for Disease Control and Prevention, National Center for Environmental Health, Division of Laboratory Sciences, Atlanta, Georgia 30341 Because of health concerns surrounding in utero and neonatal exposure to perchlorate, we developed a method for analyzing perchlorate in the dried blood spots (DBS) of newborns. Ion chromatography was interfaced with electrospray ionization tandem mass spectrometry to measure blood perchlorate levels in DBS samples as low as 0.10 ng/mL. Measurement of perchlorate in DBS indicated good precision (relative standard deviations ranging from 5.8% to 16.2%) and accuracy (% difference values ranging from -11.3% to -12.1%). Perchlorate was detectable in 100% of the DBS collected from 100 newborns. These samples had a median blank-adjusted concentration of 1.88 ng/mL. Such data support the utility of this method to quantify perchlorate levels in DBS samples. Applying this method to analyze neonatal DBS will improve perchlorate exposure assessments of this susceptible population. Perchlorate (ClO4-) is a small anionic compound used in solid rocket fuel, fireworks, explosives, and propellants. It can form naturally in the atmosphere and accumulate over time in arid regions such as northern Chile.1 By this mechanism, the nitrate fertilizer made from the ore called Chilean caliche can contain perchlorate, leading to human exposure.2-4 As a result of both anthropogenic and nonanthropogenic processes, perchlorate is distributed widely in the environmentsit can be found in drinking water in 36 U.S. states.5 For example, perchlorate has been consistently found at parts-per-billion (ppb) levels in the * To whom correspondence should be addressed. E-mail: [email protected]. Phone: 770-488-7894. Fax: 770-488-0181. † Currently at Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Atlanta, Georgia 30333. (1) Dasgupta, P. K.; Martinelango, P. K.; Jackson, W. A.; Anderson, T. A.; Tian, K.; Tock, R. W.; Rajagopalan, S. Environ. Sci. Technol. 2005, 39, 1569– 1575. (2) Urbansky, E. T.; Brown, S. K.; Magnuson, M. L.; Kelty, C. A. Environ. Pollut. 2001, 112, 299–302. (3) Collette, T. W.; Williams, T. L.; Urbansky, E. T.; Magnuson, M. L.; Hebert, G. N.; Strauss, S. H. Analyst 2003, 128, 88–97. (4) Dasgupta, P. K.; Dyke, J. V.; Kirk, A. B.; Jackson, W. A. Environ. Sci. Technol. 2006, 40, 6608–6614. (5) Unregulated Contaminant Monitoring Regulation (UCMR) data from public water systems. U.S. Environmental Protection Agency, 2004. http:// www.epa.gov/safewater/standard/ucmr/main.html (accessed September 14, 2006). 10.1021/ac802419n Not subject to U.S. Copyright. Publ. 2009 Am. Chem. Soc. Published on Web 01/27/2009

lower Colorado River.6 This contamination affects both drinking water and the food supply because the river is used for consumption and for irrigating crops in California and Arizona.6,7 Food crops can absorb and concentrate perchlorate.8,9 Perchlorate contamination of foods and beverages is not isolated to the southwestern United States; a recent survey of nearly three dozen countries found widespread contamination, with parts-per-billion levels of perchlorate.8,10 Milk can also contain perchlorate, possibly because cows ingest perchlorate-contaminated water or feed.11,12 A recent survey of U.S. residents found measurable perchlorate levels in the urine of all 2820 who were tested.13 Perchlorate exposure likely comes from a variety of sources, including drinking water and food.14 Human exposure to perchlorate is of concern because high doses can inhibit the thyroid’s uptake of iodide.15,16 In turn, inadequate thyroid uptake of iodide can lead to impaired biosynthesis of thyroid hormones and developmental abnormalities in fetuses and newborns.17 A recent study associated perchlorate exposure with changes in thyroid hormone levels among females with low iodine status.18 Additionally, perchlorate is actively transported across cell membranes by the sodium-iodide symporter.19,20 By this mechanism, perchlorate may cross the placenta and lactating breast tissues, exposing the fetus and the (6) Hogue, C. Chem. Eng. News 2003, 81, 37–46. (7) Sanchez, C. A.; Barraj, L.; Blount, B. C.; Scrafford, C.; Valentin-Blasini, L.; Smith, K. M.; Kreiger, R. I. J. Exp. Sci. Environ. Epidemiol., in press. (8) Yu, L.; Canas, J. E.; Cobb, G. P.; Jackson William, A.; Anderson, T. A. Ecotoxicol. Environ. Saf. 2004, 58, 44–49. (9) Jackson, W. A.; Joseph, P.; Laxman, P.; Tan, K.; Smith, P. N.; Yu, L.; Anderson, T. A. J. Agric. Food Chem. 2005, 53, 369–373. (10) El Aribi, H.; Le Blanc, Y. J. C.; Antonsen, S.; Sakuma, T. Anal. Chim. Acta 2006, 567, 39–47. (11) Kirk, A. B.; Smith, E. E.; Tian, K.; Anderson, T. A.; Dasgupta, P. K. Environ. Sci. Technol. 2003, 37, 4979–4981. (12) Sanchez, C. A.; Blount, B. C.; Valentin-Blasini, L.; Lesch, S. M.; Krieger, R. I. J. Agric. Food Chem. 2008, 56, 5443–5450. (13) Blount, B. C.; Valentin-Blasini, L.; Mauldin, J. P.; Pirkle, J. L.; Osterloh, J. D. J. Exp. Sci. Environ. Epidemiol. 2007, 17, 400–407. (14) Blount, B. C.; Valentin-Blasini, L.; Ashley, D. L. J. ASTM Int. 2006, 3, 3004– 3010. (15) Greer, M. A.; Goodman, G.; Pleus, R. C.; Greer, S. E. Environ. Health Perspect. 2002, 110, 927–937. (16) Wyngaarden, J. B.; Stanbury, J. B.; Rapp, B. Endocrinology 1953, 52, 568– 574. (17) Braverman, L. E.; Utiger, R. D. Introduction to hypothyroidism. In Werner and Ingbar’s The Thyroid: A Fundamental and Clinical Text; Braverman, L. E., Utiger, R. D., Ingbar, S. H., Werner, S. C., Eds.; Lippincott Williams & Wilkins: Philadelphia, PA, 2000; pp 719-720. (18) Blount, B. C.; Pirkle, J. L.; Osterloh, J.; Valentin-Blasini, L.; Caldwell, K. L. Environ. Health Perspect. 2006, 114, 1867–1871.

Analytical Chemistry, Vol. 81, No. 5, March 1, 2009

1931

breastfed neonate. The National Academy of Sciences lists the fetal and neonatal life stages as most susceptible to potential perchlorate toxicity because they may be more sensitive to thyroid perturbations and may have higher perchlorate exposure.21 Therefore, methods for assessing perchlorate exposure in newborns would be particularly useful. Dried blood spots (DBS) offer a matrix for assessing neonatal perchlorate exposure. DBS are collected from 98% of newborns in the United States as part of the Centers for Disease Control and Prevention’s Newborn Screening Program to test for certain treatable diseases.22 After screening is completed, DBS can be stored for years, allowing investigators to analyze residual samples.23,24 The results of analyzing perchlorate in newborn DBS, in addition to providing information about perinatal perchlorate exposure, could also be compared with thyroid hormone measurements performed on the same sample. Several analytical methods have been used for the analysis of perchlorate including spectrophotometry, electrochemistry, capillary electrophoresis, ion chromatography (IC), and mass spectrometry.25 Among these methods, IC with conductivity detection has been used the most frequently.26-30 Conductivity detection works relatively well for simple matrixes such as drinking water. However, in more complicated matrixes, conductivity traces have a high background, and as a result, perchlorate detection suffers.3,22 Therefore, extensive sample cleanup or preconcentration and pre-elution steps are often necessary when using conductivity detection for complex sample matrixes.26,29,31 Interfacing an IC system with a quadrupole mass spectrometer improves the reliability of quantitative perchlorate analysis in more complex matrixes, but the technique has the drawback of potential isobaric interferences.28-30 Tandem mass spectrometry (MS/MS) provides the needed combination of sensitivity and selectivity to quantify perchlorate in complex biological matrixes, such as soil, fruit, milk, urine, and amniotic fluid.32-35 Quantifying picogram amounts of perchlorate in the microliter volumes of blood from (19) Dohan, O.; Portulano, C.; Basquin, C.; Reyna-Neyra, A.; Amzel, L. M.; Carrasco, N. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 20250–20255. (20) Tran, N.; Valentin-Blasini, L.; Blount, B. C.; McCuistion, C. G.; Fenton, M. S.; Gin, E.; Salem, A.; Hershman, J. M. Am. J. Physiol. Endocrinol. Metab. 2008, 294, E802–E806. (21) National Research Council. Health Implications of Perchlorate Ingestion; The National Academy Press: Washington, DC, 2005. (22) Quality Assurance and Proficiency Testing for Newborn Screening; Centers for Disease Control and Prevention, 2006. www.cdc.gov/labstandards/ nsqap.htm (accessed October 17, 2008). (23) Chace, D.; Adam, B. W.; Smith, S. J.; Alexander, J. R.; Hillman, S. L.; Hannon, W. H. Clin. Chem. 1999, 45, 1269–1277. (24) Therrell, B.; Hannon, W. H.; Pass, K. A.; Lorey, F. W.; Brokopp, C.; Eckman, J.; Glass, M.; Heidenreich, R.; Kinney, S.; Kling, S.; Landenburger, G.; Meaney, F. J.; McCabe, E. R. B.; Panny, S.; Schwartz, M.; Shapira, E. Biochem. Mol. Med. 1996, 57, 116–124. (25) Urbansky, E. T. Crit. Rev. Anal. Chem. 2000, 30, 311–343. (26) Anderson, T. A.; Wu, T. H. Bull. Environ. Contam. Toxicol. 2002, 68, 684– 691. (27) Narayanan, L.; Buttler, G. W.; Yu, K. O.; Mattie, D. R.; Fisher, J. W. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2003, 788, 393–399. (28) Batjoens, P.; De Brabander, H. F.; T’Kindt, L. Anal. Chim. Acta 1993, 275, 335–340. (29) Ellington, J. J.; Evans, J. J. J. Chromatogr., A 2000, 898, 193–199. (30) Dourson, M. Standard Operating Procedure for Analysis of Perchlorate and Nitrate in Drinking Water, 2003. Toxicology Excellence for Risk Assessment. (31) Canas, J. E.; Cheng, Q. Q.; Tian, K.; Anderson, T. A. J. Chromatogr., A 2006, 1103, 102–109. (32) Valentin-Blasini, L.; Mauldin, J. P.; Maple, D.; Blount, B. C. Anal. Chem. 2005, 77, 2475–2481.

1932

Analytical Chemistry, Vol. 81, No. 5, March 1, 2009

DBS samples requires excellent analytical selectivity and sensitivity. This paper describes an IC-MS/MS method for measuring perchlorate in DBS, a method that meets such requirements. Analysis of perchlorate in newborn DBS will improve perchlorate exposure assessment in this susceptible population. METHODS Reagents and Chemicals. High-purity ammonium perchlorate was purchased from Sigma-Aldrich (St. Louis, MO). Certified reference sodium perchlorate solutions were purchased from AccuStandard (New Haven, CT). Stable-isotope-labeled sodium perchlorate (18O4 >90%) was acquired from Isotech (Miamisburg, OH). HPLC-grade, J. T. Baker (Phillipsburg, NJ) methanol was used in this study. Deionized (DI) water with a resistivity of 18.2 MΩ · cm or greater was obtained from an Aqua Solutions (Jasper, GA) water purification system. The filter paper cards used in this study were manufactured by Schleicher and Schuell (Keene, NH). Instrumentation. Samples were injected on a Dionex (Sunnyvale, CA) ICS-2500 IC system consisting of a GS 50 gradient pump, an EG 50 eluent generator, an AS 50 autosampler, an LC 30 chromatography oven, a CD 25 conductivity detector, and a 2 mm anion self-regenerating suppressor (ASRS Ultra II). We used a Sciex API-4000 triple quadrupole mass spectrometer with a Turbo-V source (MDS/Sciex, Concord, ON, Canada) for analyte detection. Stock Solutions and Calibration Curves. To prepare standards, we used an Accustandard certified 1000 ppm (sodium) perchlorate certified stock solution. Intermediate stock solutions were prepared by diluting the stock solution in deionized water and then frozen at -20 °C until use. We made daily standard solutions by adding dilutions of the intermediate stock solutions to 100 µL of 2 ng/mL aqueous labeled perchlorate internal standard. Perchlorate stock solutions were used to spike whole blood pools that were subsequently spotted onto filter paper cards and used in method development and quality control (QC) characterization. A concentrated labeled perchlorate internal standard stock solution was made by dissolving solid 18O4-labeled sodium perchlorate in deionized water. Intermediate and working perchlorate internal standard solutions were made by dilution of the concentrated stock with deionized water. Diluted internal standard stock solution was used to spike whole blood pools spotted on filter paper cards to determine perchlorate recovery. Blood Processing. Large volumes of blood were needed for perchlorate blood spot method development and QC materials. Obtaining such amounts of blood from infants is not feasible. Instead, blood from adult volunteers from the American Red Cross Blood Bank (Nashville, TN) was used to make DBS for method development, validation, and QC characterization. The blood was prepared by applying a method previously described.36 Before spotting on filter paper cards (Schleicher and Schuell grade 903, lot W001), the donor blood needed to be processed to (33) Krynitsky, A. J.; Niemann, R. A.; Nortrup, D. A. Anal. Chem. 2004, 76, 5518–5522. (34) Winkler, P.; Minteer, M.; Willey, J. Anal. Chem. 2004, 76, 469–473. (35) Blount, B. C.; Valentin-Blasini, L. Anal. Chim. Acta 2006, 567, 87–93. (36) Hannon, W. H.; Boyle, J.; Davin, B.; Marsden, A.; McCabe, E. R. B.; Schwartz, M.; Scholl, G.; Therrell, B. L.; Wolfson, M.; Yoder, F. Blood Collection on Filter Paper for Neonatal Screening Programs, 3rd ed.; Approved Standard, National Committee for Clinical Laboratory Standards Document A4A3; National Committee for Clinical Laboratory Standards: Wayne, PA, 1997.

remove clotting factors and buffy coat and to adjust the hematocrit of the blood from that of an adult to that of a newborn. To meet these objectives, isotonic blood bank saline (Fisher Scientific, Suwanee, GA) was added to the blood, which was subsequently centrifuged. The plasma and buffy coat (top) layer of the centrifuged blood was removed by vacuum suction. This process was repeated twice so that the red blood cells were rinsed with saline three times. An appropriate amount of hormone-free serum (Sera Care, Oceanside, CA) was added to the red blood cells to obtain a hematocrit of 55% ± 1%, the average hematocrit for a newborn infant. Fresh, unlysed blood was used each time dried blood samples were made, and blood volumes per punch were hematocrit-dependent. Spotting Blood Samples. Processed blood was divided into several pools, one for each targeted perchlorate concentration. An appropriate amount of perchlorate was spiked into each pool except for the blank pool. Pools were then stirred for 15-20 min prior to spotting the blood on filter paper cards. While the pools were being stirred, blood was drawn up into an electronic pipettor, and aliquots (100 µL) were dispensed into each of the 15 circles on the filter paper card. Blood spot samples were then dried and stored as described elsewhere.36 Cards were placed horizontally onto drying racks and kept overnight. Filter paper cards containing the DBS were stacked on top of one another, with a 6 in. × 6 in. piece of weighing paper between each card. Up to 20 cards were placed in a Zip Closure bag, along with desiccant packs and a humidity indicator card. Dried blood spot samples were stored at -20 °C under less than 30% humidity. Blood Spot Sample Preparation. Perchlorate was measured in DBS material excised from the initial DBS by using a 3.2 mm paper hole puncher with a chute attachment (see the Supporting Information for illustration of an unpunched and punched DBS sample). Each perchlorate sample consisted of 10 3.2 mm DBS punches. The chute attachment ensured that the punches were properly directed into the sample vessel (a 1.5 mL microcentrifuge tube). The 10 punches in each sample were taken from areas around the edges and center of the DBS to minimize sample-tosample volume variability. Between samples of varying concentration levels, five punches containing no blood were taken from a clean area on the filter paper card to cleanse the hole puncher and to minimize carryover. Methanol (100 µL) was added to the 10 discs of dried blood to precipitate proteins so that they adhere to the filter paper, and each sample was then vortexed for 30 s. Samples were then concentrated by evaporating methanol in a TurboVap at 25 °C and 2 psi for 60 min. After the transfer of the 10 blood spot punches to a microcentrifuge filtration tube (0.2 µm, nylon membrane, CoStar), aqueous labeled perchlorate internal standard solution (100 µL, 2 ng/mL) was added to each tube, and the samples were vortexed again for 30 s. Subsequently, the perchlorate-containing aqueous solution was separated from the blood residue discs via spin-filtration in a Biofuge Pico microcentrifuge (Hereaus, Germany) at 16 060g (13 000 rpm) for 5 min. Each filtered sample was placed in an autosampler vial for IC-MS/ MS analysis. IonChromatography-TandemMassSpectrometry(IC-MS/ MS). Perchlorate was analyzed by using the IC-MS/MS method previously published by Valentin-Blasini et al.32 In brief, a Dionex

AS-20 column (2 mm × 250 mm) was used for separation, and a 50 mM of sodium hydroxide (NaOH) eluant was used under isocratic conditions at a flow rate of 0.50 mL/min. An injection volume of 24 µL was obtained by using a 25 µL injection loop in partial-loop injection mode. The total chromatographic run time was of 10 min. An in-line suppressor (ASRS Ultra II, Dionex) removed potassium from the eluent before the eluent flow entered the mass spectrometer. Electrospray negative ionization was used at a source temperature of 600 °C. The mass spectrometer was operated in multiple reaction monitoring (MRM) mode, which tracked the following mass transitions: 99f 83, 101 f 85, and 107 f 89. Global source parameters were optimized for the 99 f 83 transition. The mass spectrometer was tuned in negative mode with a poly(propylene glycol) (PPG) solution (Applied Biosystems). In addition, 3-methylhippuric acid was spiked into the PPG solution to add mass (m/z 192) in an area of the mass range where there was inadequate mass coverage (between m/z 59 and 525). Data Analysis. To identify perchlorate peak, we matched the analyte peak with that of the labeled internal standard. Sciex Analyst software was used to perform data analysis and peak integration. Peaks were inspected for proper integration and manually reintegrated if needed. The instrument response was measured as the ratio of analyte area to labeled internal standard area. A set of 11 calibrators (0.1-75 ng/mL) was analyzed in duplicate with each batch of unknown samples and weighted 1/x to prepare a daily calibration curve. We determined the perchlorate concentration in each sample on the basis of the calculated amount of perchlorate per estimated initial blood volume in the DBS punches. Several variables influence blood spot sample volume, and such variables must be considered in data analysis. Blood hematocrit and the volume of blood in each DBS are each positively correlated with the volume of blood in a DBS punch.37 Lot-to-lot variation in manufacturing filter paper also contributes to variable DBS punch volume.37,38 The lot of filter paper used in this study (Schleicher and Schuell W001) had average serum and whole blood volumes of 1.40 and 3.11 µL, respectively, per 3.2 mm punch for DBS made from 100 µL of 55% hematocrit, unlysed blood.38 Blood used to make method validation DBS (55.5% hematocrit) had a volume of 3.14 µL per punch, whereas blood used to make QC DBS (54.5% hematocrit) had a volume of 3.08 µL per punch. Daily Operating Procedure. Before performing daily instrumental analyses, high-purity water used in the eluent was filtered and degassed to minimize the potential clogging of particulates in the system. An in-line filter (0.45 µm, replaced daily) was positioned before the column to prevent particulates from entering. The mass spectrometer curtain plate was cleaned daily to minimize mass spectrometer contamination. An equilibration standard was injected three times before daily instrumental analyses to ensure adequate chromatography and mass spectrometer detection. Quality Control Pools. Quality control pools were made from donor blood obtained from the American Red Cross. The blood was separated into three pools. One pool had no perchlorate added (37) Adam, B. W.; Alexander, J. R.; Smith, S. J.; Chace, D. H.; Loeber, J. G.; Elvers, L. H.; Hannon, W. H. Clin. Chem. 2000, 46, 126–128. (38) Centers for Disease Control and Prevention and Association of Public Health Laboratories. Newborn Screening Quality Assurance Program 2004 Annual Summary Report, 2005.

Analytical Chemistry, Vol. 81, No. 5, March 1, 2009

1933

Table 1. Method Precision and Accuracy DBS samplea spiked 0.33 ng/mL spiked 3.3 ng/mL spiked 33 ng/mL QC lowc QC highc a

Figure 1. Chromatogram of a dried blood sample prepared from blood containing perchlorate at 0.10 ng/mL. Quantitation ion trace is shown in bold print (S/N ∼ 18); confirmation ion trace is shown in lighter print (S/N ∼ 9).

(background), and the other two pools were spiked with perchlorate to yield concentrations of 2.27 (QC low) and 27.3 ng/mL (QC high). Blood from each pool was spotted on filter paper cards (100 µL/spot) and allowed to dry overnight. Blood spot QC samples were then placed in Zip Closure bags with desiccant and stored at -20 °C until analysis. Two analysts characterized perchlorate levels in these QC materials by completing a total of 20 independent measurements. Mean perchlorate concentrations and standard deviations were calculated and used for evaluating the precision of subsequent analysis. Precision and Accuracy. Precision and accuracy were determined by analyzing DBS samples prepared from blood containing perchlorate at three different levels (0.33, 3.3, and 33 ng/mL). Samples at each level were run in triplicate. Precision was calculated as the percent relative standard deviation (% RSD) of triplicate samples at each level. Accuracy was determined as % difference of triplicate samples from the calculated spike concentrations at each level. Additional precision and accuracy data were also obtained from 20 independent analyses of each pool of QC material. RESULTS AND DISCUSSION Reliable quantification of trace levels of perchlorate in dried blood samples requires excellent analytical sensitivity and selectivity. We achieved this objective by coupling IC and MS/MS. Figure 1 shows the signal produced by analyzing a dried blood sample prepared from blood containing perchlorate at the lowest reportable level for the method (0.10 ng/mL). The perchlorate quantification ion transition is free from significant interferences, and it produces an excellent signal-to-noise (S/N) ratio (∼18) for 3 pg of perchlorate in the dried residue of 30 µL of blood. Calibration curves were prepared in both DBS matrix and water. The average slope of duplicate injections for the calibration curves was compared, and it showed a percent difference of 11.7%. On the basis of this finding, water calibration curves were used to determine perchlorate in blood spot samples. To further evaluate this calibration model, we used a water calibration curve to determine the precision and accuracy of analyzing blood spot samples containing perchlorate at three concentration levels (0.33, 1934

Analytical Chemistry, Vol. 81, No. 5, March 1, 2009

N

theor concn (ng/mL)

mean concn (ng/mL)

% RSDb

% difference

3 3 3 20 20

0.33 3.3 33 2.27 27.3

0.37 3.1 31.4 2.3 24.2

16.2 13.0 7.7 7.4 5.8

12.1 -6.1 -4.8 1.3 -11.3

Dried blood spot. b Relative standard deviation. c Quality control.

3.3, and 33.3 ng/mL, n ) 3). Precision (% RSD) and accuracy (% difference) values ranged from 7.7% to 16.2% and -4.8% to 12.1% (Table 1), respectively, confirming that a water calibration curve can be used to quantitatively determine perchlorate in blood spot samples. Subsequent analysis of QC material at two levels (n ) 20) against a water calibration curve showed % RSD and absolute % difference values of less than 11.3% (Table 1). In addition, the QC material was used to evaluate storage stability. The perchlorate in the QC material was repeatedly analyzed for a period of 24 months, and the analyses showed that its measured amount deviates minimally (7.9%) from the amount expected based on the spiked amounts (data not shown). This result indicates that perchlorate is stable in DBS samples stored at -20 °C for a prolonged period. The limit of detection (LOD) for perchlorate in DBS was based on calculating the standard deviation at zero concentration (S0), as described by Taylor.39 We analyzed DBS (n ) 6) spiked with perchlorate concentrations equal to the four lowest calibrators. Standard deviations for each concentration level were plotted against the concentration and used to extrapolate the standard deviation at zero concentration (0.02 ng/mL). The LOD was defined as 3S0 (0.06 ng/mL). The lowest reportable perchlorate concentration was assigned as the perchlorate level in the lowest calibrator (0.10 ng/mL). Small sample volumes associated with DBS make it difficult to achieve the desired sensitivity. The use of a larger number of punches per sample can increase the sensitivity of a method, as was the case in the current study. However, the number of available DBS punches per sample is limited in most cases. Infusing an organic solvent such as acetonitrile postcolumn can enhance sensitivity by increasing ionization in the mass spectrometer.10 Thus, future applications of this method will likely include postcolumn infusion of acetonitrile. Recovery was determined by spiking the internal standard into samples at different points in sample preparation. Recovery data was based on triplicate analysis of DBS and water samples containing perchlorate at 0.33 ng/mL, 3.3 ng/mL, and 33 ng/ mL. Calculated concentrations were compared to theoretical concentrations to determine both relative and absolute recovery. Values for relative recovery can be found in Table 2; they were generally between 97% and 112%. Some perchlorate was lost in the filtration step, leading to absolute recoveries of approximately 77% (data not shown). However, using stable isotope dilution corrects for variable perchlorate loss during sample preparation and leads to relative recoveries near 100%. The internal standard (39) Taylor, J. K. Quality Assurance of Chemical Measurements; Lewis Publishers: New York, 1987.

Table 2. Recovery Values for Perchlorate in Dried Blood Spots when IS spikeda

calcd concn (ng/mL)

relative recovery %

in blood, prior to spotting in blood, prior to spotting in blood, prior to spotting after methanol fixing after methanol fixing after methanol fixing

0.36 3.3 33 0.37 3.2 33.2

109 100 100 112 97 101

concn (ng/mL) 0.33 3.3 33 0.33 3.3 33 a

Table 3. Perchlorate Concentrations in Adult Blood and Dried Blood Spots

sample

estimated concn (ng/mL)

% RSDa

% diffb

unspiked donor blood heparinized 10 DBS punches nonheparinized 10 DBS punches heparinized whole spot

3.08 3.29 3.33 2.79

2.3 2.1 2.1 1.8

6.6 7.8 -9.9

a

Relative standard deviation. b Percent difference.

Internal standard.

also corrected for ionization suppression in the mass spectrometer, as indicated by nearly identical calculated concentrations of DBS and water samples (data not shown). Several types of sample cleanup and extraction procedures were tested during method development. The following combinations of protein precipitation (PP) and C18 solid-phase extraction (SPE) were investigated: no PP or SPE, PP only, SPE only, and both PP and SPE. We performed protein precipitation by adding methanol to samples (methanol fixation). Performing methanol fixation during sample preparation resulted in samples that were transparent and that ranged in color from clear to light brown; samples that did not undergo protein precipitation were dark red and opaque. Consequently, we concluded that protein precipitation is necessary for sample preparation. Samples treated with PP only and PP with subsequent SPE were compared, and they showed no significant differences in chromatography, precision, and accuracy. Therefore, samples were prepared by use of PP only. We evaluated perchlorate contamination by lot-screening consumables used for sample processing and storage. No measurable traces of contamination were found, with the exception of the filter paper card. We tested filter paper cards by taking 10 punches from each paper blank sample from areas that contained no blood. Punches were taken from areas between circles printed on the card, from within the circles printed on the card, and from areas on the filter paper card containing ink. Perchlorate was consistently found in the filter paper, but typically at levels less than the method detection limit (0.10 ng/mL). The bleach used in the process of making filter paper cards may be a source of low-level perchlorate contamination. Perchlorate can be used as a component of ink; however, no quantifiable levels were found in the printed areas of our newborns’ screening filter paper. Additional steps need to be taken with blood spot samples to ensure that the data is analyzed as accurately as possible. The endogenous levels of perchlorate in donor blood (0.10-3.1 ng/ mL) and perchlorate contamination of filter paper (