Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Surround optical-fiber immunoassay wikipedia , lookup
Fatty acid metabolism wikipedia , lookup
Community fingerprinting wikipedia , lookup
Genetic code wikipedia , lookup
Matrix-assisted laser desorption/ionization wikipedia , lookup
Biosynthesis wikipedia , lookup
Metalloprotein wikipedia , lookup
Amino acid synthesis wikipedia , lookup
Biochemistry wikipedia , lookup
Mass spectrometry wikipedia , lookup
CLIN. CHEM. 39/1, 66-71 (1993) Rapid Diagnosis of Phenylketonuria by Quantitative Analysis for Phenylalanine and Tyrosine in Neonatal Blood Spots by Tandem Mass Spectrometry H. Chace,’ David S. Mlllington,’ Hofman3 Donald Naoto Terada,2 Stephen G. Kahier,’ Charles A new method for quantifying specific amino acids in small volumes of plasma and whole blood has been developed. Based on isotope-dilution tandem mass spectrometry, the method takes only a few minutes to perform and requires minimal sample preparation. The accurate assay of both phenylalanine and tyrosine in dried blood spots used for neonatal screening for phenylketonuria in North Carolina successfully differentiated infants who had been classified as normal, affected, and falsely positiveby current fluorometric methods. Because the mass-spectro- metric method also recognizes other aminoacidemias simultaneously and is capable of automation, it represents a useful development toward a broad-spectrum neonatal screening method. R. Roe,’ and Lindsay F. referred to as liquid secondary ionization (LSI)-is suitable for the rapid assay of specific components in complex biological samples (5). The basis for this type of analysis is the production of mostly intact molecular species from a complex mixture by use of a “soft” ionization technique, after which molecular fragments are identified from specific components induced by collisions with neutral gas molecules. The use of two mass analyzers separated by a region in which the collisionally induced fragmentation takes place allows highly selective and specific analysis for compounds of various structural classes. Time-consuming chromatographic separations are not needed because separation and analysis take place simultaneously and entirely within the - mass spectrometer. AddItIonal Keyphrases: heritable disorders metabolism . amino acids tyrosinemia Screening neonates for phenylketonuria (PKU), an inborn error of metabolism resulting from phenylalanine hydroxylase deficiency (McKusick 261600), is beneficial because it provides for early diagnosis and treatment (1, 2). The collecting of specimens as blood spets on filter paper and the subsequent quantitative analysis for phenylalanine (Phe) is used in neonatal screening laboratories worldwide (3). Most of the analytical methods, including the fluorometric assay currently used in the North Carolina State Screening Laboratory (4), generate a relatively high rate of false-positive results (typically about 0.1% of those screened), which require follow-up analysis for amino acids. In North Carolina during 1991, >100 samples were followed up for initial screening results of above-normal Phe concentrations, of which only 2 were subsequently confirmed as positive for PKU. Typically, five or six cases are detected annually, but the false-positive rate is still high. These false-positive results are generally assumed to be due to initially high Phe concentrations that later decrease to within the normal range. Tandem mass spectrometry (MS/MS) coupled with fast atom bombardment ionization-more of LSIMS/MS to the quantitative analysis for Phe and tyrosine (Tyr) in plasma and neonatal blood spots for the rapid diagnosis of PKU and related disorders of metabolism of aromatic amino acids. Materials and Methods Materials Solvents were high-purity grade (Burdick and Jackson, Muskegon, MI). The esterifying agent, 3 moltL HC1 in n-butanol, was obtained from Regis (Morton Grove, IL). [2H5]L-Phenylalanine (98 atom%), [2H41L-tyrosine (98 atom%), and [2H3-methyl]L-leucine (99 atom%) were obtained from MSD Isotopes (Montreal, Canada). The unlabeled Phe and Tyr and amino acid standard kits were from Aldrich (Milwaukee, WI) and Sigma (St. Louis, MO). Stock solutions of the analytical standards were prepared in 0.1 mol/L HC1 and stored at 4#{176}C. Filter paper for the preparation of blood spots was from Schleicher and Schuell (Keene, NH; cat. no. 903). accurately 1Department of Pediatrics, Division of Genetics and Metabolism, Duke University Medical Center, Box 3028, Durham, NC 27710. ‘Department of Pediatrics, Kyoto Prefectural University of Medicine, Kawaramachi Hirokoji Kamikyo-ku, Kyoto 605, Japan. 3State of North Carolina Department of Environment, Health, and Natural Resources, Division of Laboratory Services, 306 N. Wilmington St., Raleigh, NC 27601. 4Nonstandard abbreviations: PKU, phenylketonuria; MS/MS, tandem mass spectroscopy; and LSI, liquid secondary ionization. Received April 20, 1992; accepted August 19, 1992. 66 CLINICALCHEMISTRY,Vol. 39, No. 1, 1993 In this laboratory, we have developed and applied LSIMSIMS for the detection of abnormal urine and plasma metabolites in patients with inborn errors of metabolism (6, 7) and have described a general LSIMSfMS method to analyze for amino acids in blood and plasma (8). Here we report the specific application Samples Newborn screening cards containing blood spots collected on the third day postpartum were made available by the State of North Carolina Division of Laboratory Services. Among the samples were 10 from infants with normal concentrations of Phe, 8 from infants with increased concentrations of Phe and subsequently confirmed as positive for PKU, and 5 from infants with initially increased concentrations of Phe that were normal on follow-up and shown not to have PKU. In addition, we examined 45 cards from the New England Regional Newborn Screening Program, of which 30 were normal control subjects and 15 were patient control subjects (other than PKU). The original screening cards and plasma samples from later hospital admissions were also available from one patient with the oculocutaneous type of tyrosinemia, caused by tyro- sine transaminase deficiency (tyrosinemia type I, McKusick 276600), and from another with fumarylacetoacetate hydrolase deficiency (tyrosinemia type II, McKusick 276700). The first 206 samples from a pilot study involving the analysis of fresh neonatal samples from North Carolina by MSIMS were also included as a set of normal controls. Preparation Dried of Samples blood spots on filter paper were prepared by punching out a 6.35-mm (0.25-in.)-diameter circle into a 1-niL vial with a standard paper punch. This corresponds to --11 L of whole blood (9). About 2 nmol each of [2H3]leucine, [2H5]phenylalanine, and [2H4]tyrosine in 0.5 mL of methanol were added to the vial. The contents were vortex-mixed and allowed to stand for 30 mm at room temperature. The supernate was transferred to a 1-mL conical vial, evaporated to dryness under nitrogen, and incubated with 50 1L of 3 mol/L HC1 in n-butanol at 65#{176}C for 20 mm in a capped 1-niL glass vial. The solvent was evaporated under nitrogen, and the derivatized sample was reconstituted in 50 pL of an equivolume mixture of methanol and glycerol. To assay plasma samples, we added -5 nmol each of the isotopically labeled internal standards-[2H3]Leu, [2H5]Phe, and [2H4]Tyr-in methanol (800 L) to a 100-j.L aliquot of plasma. After vortex-mixing, we transferred the supernate to a 5-niL vial, evaporated the solvent, and reconstituted the residue in methanol (2 X 100 ML), transferring this to a 1-mL vial. The same procedure for derivatization and preparation was followed as described above. The recoveries of Phe and Tyr from blood spots were determined by standard addition at three different concentrations (25, 200, and 400 pmol/L), then experimentally measuring the added amounts and quantifying the results from the standard curves after subtracting the endogenous signals. In these experiments, we accurately pipetted 12 L of the fortified blood samples onto the paper and used the entire spots for the analysis. Mass Spectrometry We used a VG Quattro triple-quadrupele tandem mass spectrometer with a Lab-base data system (Fisons Instruments, Danvers, MA). This computer-controlled instrument incorporates a cesium ion gun operated at 10 keV, a conventional LSIMS insertion probe, and two off-axis detectors. One detector was placed in an intermediate position after the first quadrupele (Q1) but before the collision region (also the second quadrupole, Q2);the other was positioned after the third quadrupole (Q3). Tuning was optimized by using a solution containing one of the analytical standards, 2 mmol/L, deriva- tized as the butyl ester and dissolved in the LSIMS matrix. The settings for the ion source were optimized first, to achieve maximum intensity of the EM + H] ion at the intermediate detector, with unit mass resolution in the first quadrupole. During the monitoring of the protonated molecular ion at the final detector, argon was introduced into the intermediate collision cell until the intensity of the signal decreased by 50%. At this point, the daughter ion corresponding to EM + H 1021 was brought into focus and the collision energy was optimized to provide maximum sensitivity. The optimum value for all the amino acids studied was about 30 eV. The second analyzer (Q3) was also set to unit mass resolution and all lens voltages in the instrument were optimized systematically to provide maximum intensity at unit mass resolution for the daughter ion. These tuning settings were saved and recalled automatically by the computer when this type of analysis was to be performed. In general, the resolution settings and lens voltages were not altered unless maintenance in the analyzer region of the instrument was required. Source tuning and collision gas pressure were optimized when necessary, usually once daily, with use of a tuning solution containing the isotopically labeled standards for Phe and Tyr at 2 mmol/L each. Samples were analyzed with the static LSIMS probe without further tuning of the instrument. A blank solution containing only internal standards was analyzed occasionally to ensure the absence of carryover between samples. All MSIMS spectra (product and neutral loss scans) were acquired in the continuum mode, in which every data point was stored. Sixty consecutive scans of 1 a each were accumulated into a single raw spectrum, which was then automatically processed by five-point smoothing, baseline subtraction, and centroid calculation to produce the final stored mass spectrum. The result of this was a mass spectrum with much better signal-to-noise ratio and reproducibility of ion ratios than the average of 60 mass spectra recorded with use of the standard real-time centroid calculation algorithm. To produce the scan for “neutral loss of 102 Da,” we scanned Q1from mlz 125 to 300 while simultaneously scanning Q3from m/z 23 to 198. The result was a spectrum of product ions, mostly corresponding to EM + H 102]. For convenience, the mass scale was adjusted to show the precursor (EM+ H]) ion masses on the recorded mass spectra as used later for Figures 3 and 4. The mass spectra shown there and in Figure 1 are the accumulated raw spectra after smoothing and baseline subtraction (as described above). For quantitative analysis for Phe and Tyr, we determined the abundance ratios of the ions corresponding to Phe and E2H5]Phe (mlz 222:227) and Tyr and [2HJPyr (mlz 238:242). These values were obtained directly from the absolute ion abundances stored afcer processing the raw mass spectra as described above. The concentrations of Phe and Tyr were then calculated by reference to the appropriate standard curves. These calibration curves were derived by standard addition of Phe and Tyr - - CLINICAL CHEMISTRY, Vol.39, No. 1, 1993 67 Phenylalanine lXI (-102/ (-102/ 136 I I Tyrosin. 57 57 100 150 100 200 2H5-Phenylalanine 100 < 150 200 2H4-Tyroslne 242 1.1021 1-102/ 57 57 100 150 140 100 ( 150 200 1* FIg. 1. Product ion mass spectra from the collision-induced dissociation of [M + H] ions of Phe, Tyr, and their stable isotope-labeled in the tandem mass spectrometer The most abundant fragment in eachcase corresponds to the lossof 102 Da to aliquots of a control plasma having fixed concentrations of the internal standards. The ion abundance ratios (mlz 222:227 and mlz 238:242), corrected for natural isotopic abundance, were plotted against the added concentrations of Phe and Tyr, respectively. Results Analysis for Amino Acids by MS/MS As reported previously (8), the butyl ester derivatives excellent sensitivity and specificity for the detection of amino acids in blood or plasma by MS/MS. The fragmentation of the protonated molecular ions (EM + HIl of the Phe and Tyr butyl esters and of their isotopically labeled analogs under collision-induced disprovided sociation in the tandem mass spectrometer is shown in Figure 1. A common loss of the elements of butyl formate (102 Da) was observed according to the product ions at m/z 120 and 136 for Phe and Tyr and at m/z 125 analogs The aromatic amino acids, Phe and Tyr, have unique molecular masses and were detected with good sensitivity and selectivity in blood at physiologically normal ranges of concentration. This is clearly illustrated in Figure 3A, which shows the amino acid profile by MS/MS from a typical sample of normal human plasma. Signals for several amino acids were clearly discernible. Signals corresponding to the isotope-labeled internal standards, [2H3]Leu, E2H5JPhe, and E2HJI’yr, were detected at m/z 191, 227, and 242, respectively. The spectrum was essentially free of interfering signals derived from other endogenous compounds in plasma. The profile in Figure 3B was from the methanol extract of a 12-L blood spot from an original neonatal screening card corresponding to one of the control subjects. The profile of amino acids was essentially the same as that of the normal plasma (Figure 3A), and R and 140 for [2H5]Phe and E2H4]Tyr, respectively. This process, shown schematically for Phe and Tyr in Figure 2, involves the transfer of a proton to generate a resonance-stabilized carbonium ion, which is apparently specific to a-amino acids. The loss of 102 Da was not observed with either fatty acid or acylcarnitine butyl esters (data not shown). Other minor fragments were also observed in the product ion spectra, including a common ion at m/z 57 derived from the butyl group (Figure 1). 0 Phe: Tyr. A-H A-OH CH2 H3i-CH-COOC4H8 A neutral-loss scan of 102 Da in the MS/MS mode, whereby both mass analyzers are scanned simultaneously with a constant mass difference of 102 Da, generated a spectrum showing the molecular ions of the CID amino acids (as their butyl esters) present in a mixture (8). The sensitivity for each amino acid varied widely, H2it-CH-CH2-- HCOOC4H8 A and some signals were composites of more than one amino acid that either shared the same molecular mass (such as leucine, isoleucine, and hydroxyproline) or a Fig.2. Schematic representation of the specific fragmentationthat characterizes the amino acid butyl ester derivatives Loss of the elementsof butyl formate(102 Da) fromthe protonated molecule occurs under collision-induced dissociation and results in the formation of a common fragment ion (such as glutamine resonance-stabilized 68 CLINICALCHEMISTRY,Vol.39, No. 1, 1993 and lysine). carbonium ion 100 Assay Accuracy A I L 100 nv B ) n;1z Fig.3. MS/MS aminoacid profilesof normalhumanplasma(A) and a blood spot froma normalneonatalscreeningcard(B) obtained by applying the ‘neutral loss of 102 Da” scan to the denvatized methanol extracts The signals detectedcorrespond to the fragments resulting from the loss of 102 Da from the [M + H1 ion of each amino acid, but for convenience the mass scale has been adjusted to represent the precursor (protonated molecular) ionmasses.Giyclne (Gly, m/z 132), alanine + sarcoslne (Ala + Sar, m/z 146), senne (Ser, m/z 162), proline + asparagine (Pro + Asn, mhz172), valine (Val,mhz174),threonine(Thr, mhz 176), glutamine + lysine (Glu + Lys, m/z 186), leucine + isoleucine + hydroxyproline (Leu + lie + Hypro, m/z 188), methionine (Met, m/z 206), phenylalanlne (Phe, mhz 222), tyroslne (Tyr, mhz 238), aspartic acid (Asp, mhz 246), glutamic acid (Glu, m/z 260), and tiyptophan(Trp, m/z261) there was no evidence of interference from signals not derived from the amino acids and their added internal and Precision The accuracy of this method was previously determined by comparing the individual values for Phe and Tyr in selected plasma samples with those obtained by a standard method involving ion-exchange chromatography (8). The analytical recoveries of added Phe and Tyr from blood were determined in triplicate at concentrations of 25, 200, and 400 moI/L. The respective mean (SD) values obtained were 98 (7)% and 101 (8)% at 25 tmol/L; 98 (6)% and 93 (6)% at 200 MmoJJL; and 100 (5)% and 102 (6)% at 400 molJL. The methods of calibration and analysis on which isotope-dilution MS is based are independent of the recovery of analytes from the biological matrix (10,11). In principle, MS/MS greatly reduces the chance of chemical interference because of its high molecular specificity, which greatly exceeds that of conventional ultraviolet absorbance detectors. This is exemplified by the lack of interfering signals in the mass spectral profiles not derived from amino acids (Figure 3, A and B). Precision of the assay was calculated by replicate analysis of the same blood sample by the complete analytical procedure for blood spots described in Mate- Table 1. QuantitatIve Analysis s- Cenc by MS/MS, ganottL Sample Ph. Tyr 38 (<242)b 192 (<242) 211 2 45 (<242) 51 4 5 6 38 (<242) 42 (<242) 62 (<242) 48 78 Assay Calibrationand Limitsof Detection 3 51(242) 4 5 224(557) 149 (279) Phe and Tyr (Table 1). 48 3 standards. 0.99). The signal-to-noise ratios for endogenous signals for Phe and Tyr in typical blood samples were 50:1 and 20:1, respectively. The estimated detection limits-the signals projected to give a signal-to-noise ratio of 3:1correspond to concentrations of 3 and 10 pmoIJL, respectively, well below the physiologically normal ranges for Ph.yr Control group 1a Controlgroup2” 1 2 We used standard isotope-dilution assay techniques (10, 11) to derive calibration curves for the analytes of interest in human plasma, i.e., standard addition of the Phe and Tyr over the calibration range of interest with fixed amounts of the internal standards. The ratios of signals corresponding to the analytes and their internal standards were plotted as a function of concentration. The curve for added Phe showed excellent linearity over the concentration range 2-80 mol/L (r2 = 0.99). For Tyr, the range of linearity was 10-400 Mmol/L (r2 = for Phe and Tyr In Blood 150(448) 83(260) PKU positives 1 2 3 4 5 6 7 0.88 0.80 0.53 70 0.89 93 157 109 1.6 0.53 0.47 220 1.0 75 2.0 166(297) 51 3.3 490 (624) 64 77 61 53 7.7 2.5 67 67 70 307 283 21.6 3.0 8.5 195(242) 237(370) 512(430) 1454(1545) 254 (273) 598 (576) 8 Tyrosinernia 1d 0.80 0.91 37 (n.a.) 3.9 9.7 0.12 Tyrosinemia 1l 142 (n.a.) 0.45 a Selected from 55 normal samples (as assessed by the State Laboratory). b Results of fluorometric assay used by the State Screening Laboratory tsted in parentheses. C False PKUpositives (as assessedbythe State Laboratory). dTyne transaminasedeficiency. Fumaiylacetoacetate n.a., not applicable. hydrolase deficiency. CLINICALCHEMISTRY,Vol.39, No. 1, 1993 69 These values, representing the within-assay variation, were 6.2% for Phe, 8.4% for Tyr, and 6.6% for the Phe/Tyr ratio (n = 10). The CVs for Phe, Tyr, and Phe/Tyr ratio determined for the same sample on different occasions within 1 month, representing the interassay variation, were 8.8%, 9.7%, and 11.5%, respectively (n = 8). The absolute concentrations of Phe and Tyr in this sample were 115 and 85 mo1/L, respectively. rials and Methods. 100 A cJ 100 00 Analysisof Original Neonatal Screening Cards we used 206 fresh samples, from the State Screening Laboratory, analyzed of collection. The samples had already been analyzed by the State Laboratory and were considered normal. In this control group, the mean ± SD (and range) for Phe and Tyr were 55 ± 17 (12-172) mol/L and 77 ± 30 (29-236) molIL, respectively; the PhelTyr ratio in each sample was 0.78 ± 0.28(0.21-1.74). These Phe and Tyr concentrations compare well with values from the literature for neonates: 38-137 and 55-147 junol/L (12). Premature infants can have Phe and Tyr valuesof 213 and 420 moIIL, respectively (12). Figure 4 shows MS/MS analysis of unused blood spots from a second group of older samples (stored 2 years) from the State of North Carolina Screening Laboratory that had been categorized as either normal (control group 1), confirmed PKU, or falsely positive (control group 2). Representative results from each category, showing the signals for Phe, Tyr, and their respective internal standards are presented in Figure 4A-C. For comparison, the profile of a patient with tyrosinemia caused by tyrosine transaminase deficiency is presented in Figure 4D. For each profile, both Phe and Tyr were readily quantified from the ion abundances and the standard curves. Results are summarized in Table 1, including results for 6 of the 55 control samples from group 1. The mean ± SD concentrations of Phe and Tyr in all 55 controls were 85 ± 31 mol/L (range 30-192 tmol/L) and 135 ± 66 prnol/L (range 46-368 mol/L), respectively. For the true PKU positives, the abovenormal Phe values measured by the State Laboratory with a fluorometric assay were in good agreement with those determined by isotope-dilution MS, whereas the fluorometric Phe values for the false positives (control group 2) were more than twice as high as those measured by mass spectrometry (Table 1). The Phe/Tyr ratio was calculated for each sample (Table 1). For control group 1, the mean ± SD Phe/Tyr ratio was 0.71 ± 0.22 (range 0.2-1.4, n = 55)-very similar to that obtained from the analysis of the 206 fresh control samples previously described. For all cases of PKU, the mean ratio was 7.8 ± 6.5 (range 2.6-22, n = 8); for the false positives in control group 2, it was 1.1 ± 0.67 (range 0.4-2.0, n = 5). Amino acids were also analyzed from the original PKU cards that corresponded to two cases of tyrosinemia, retrieved from the State Screening Laboratory after a positive diagnosis of tyrosine transarninase deficiency in one case and fuinarylacetoacetate hydrolase B For controls, North Carolina within 2 weeks 70 CLINICAL CHEMISTRY, Vol. 39, No. 1, 1993 C I C 04 100 D 240 nVz Fig. 4. Detection of the aromaticamino acids PheandTyr,with their respective internal standards ([2H5]Phe and [2HJryr), by MS/MSof extracts from selected neonatalblood spots: (A) normal;(B) classical PKU; (C) falsely positiveby fluorometricassay and normal by MSMS; (D) tyrosinemladue to tyrosinetransaminase deficiency The Internal standards enableaccuratequantificationof Phe and Tyr deficiency in the other, The neonatal tyrosine concentrations were 307 and 283 mol/L, respectively (controls: 135 ± 66 pmoJ/L, n = 55). DIscussIon MS/MS can be a in neonatal blood samples. A larger sampling will be needed to accurately determine the sensitivity and specificity of the assay. Samples are prepared in batches of 60 by a simple solvent extraction and derivatization procedure that takes -2.5 h. The total instrument time required for the These reliable results method demonstrate for detecting that PKU analysis of each sample is -2-3 mm. The advantages of MS/MS over alternative methods of analysis are its high specificity and accuracy of quantification through use of the isotope-dilution technique, plus its speed and amenability to automation. The lack of chemical interference from other components in the sample matrix is striking in view of the simplicity of the method of sample preparation. Most of the samples identified as false positives by the fluorometric suggesting assay that the for Phe were normal fluorometric assay by MS/MS, is subject to chemical interference that leads to overestimation of the Phe concentration in at least some cases. The MS data also indicate that the Phe/Tyr ratio is an excellent discriminant for PKU-affected and normal cases. We expect the number of false positives obtained by using this ratio to be very low. In addition to PKIJ, the new method might be expected to detect at least some forms of tyrosinemia in neonates, as indicated by the results for the two cases presented here. Other disorders of amino acid metabolism such as the glycinemias, methioninemias, and maple syrup urine disease should also be detectable by the same test. When combined with the acylcarnitine profile test for disorders of fatty acid and branched-chain amino acid catabolism currently performed on the same samples in this laboratory, at least 15 metabolic disorders are potentially detectable from the same blood spot with a single analysis (13). The suitability of a tandem quadrupole mass spectrometer in the clinical laboratory can be judged by our experience, in which such an instrument has been used routinely in a clinical diagnostic setting for >2 years to assay acylcarnitines in urine and plasma. During this time, >5000 patients’ results were obtained. The equipment is compact in size and controlled by programmed computer instructions. It is actually easier and less troublesome to operate than a gas-chromatographic MS system, and in our hands downtime has been restricted to scheduled ion-source cleaning (4 h every 2 weeks) and maintenance (2 days every 6 months), with breakdowns accounting for <5 days per year. The instrument is operated daily by B.S.-degreed technicians under the remote supervision of an experienced mass spectroscopist. The current sample load of -200 samples per week is easily handled by manual methods of sample analysis. Automation of the MS analysis and data processing is currently under development and will ultimately facilitate large-scale neonatal screening of 400 samples per day by this method. The capacity for high throughput and accurate, multiple testing combined with miiiimal labor and reagent costs can offset the high capital cost of the equipment. We thankS. HAilmanforhis expert technical assistance. Financial support was from NICHD Division of the National Institutes of Health (Bethesda, MD; grant no. HD-24908) and the State of North Carolina, Division of Maternal and Child Health, Department of Environmental Health and Natural Resources (Raleigh, NC; grant no. C-05070). We also thank Harvey Levy of the New England Regional Newborn Screening Program (Jamaica Plain, MA) for a set of 45 original Guthrie cards. References 1. Kirkman kIN, Carroll CL, Moore EG, Matheson MS. Fifteenyear experience with screening for phenylketonuria with an automated fluorometric method. Am J Hum Genet 1982;34:743-52. 2. Snyderman SE. Newborn metabolic screening: follow-up and treatment results. In: Carter TP, Wiley AM, eds. Genetic disease: screening and management. New York: Alan R Lisa, 1986:195.209. 3. Guthrie R. Newborn screening past, present and future. In: Carter TP, Wiley AM, eds., Genetic disease: screening and management. New York: Alan R Lisa, 1986:319-39. 4. Hoffman GL, Laeasig RH, Hassemer DJ, Makowski ER. Dual channel continuous-flow system for determination of phenylalanine and galactose: application to newborn screening. Clin Chem 1984;30:287-90. 5. McLaffertyFW. Tandem mass spectrometry. Science 1981;214: 280-7. 6. Millington DS, Norwood DL, Kodo N, Roe CR, Inoue F. Application of fast atom bombardment with tandem mass spectrometry and liquid chromatography/mass spectrometry to the analysis of acylcarnitines in human urine, blood, and tissue. Anal Biochem 1989;180:331-9. 7. Millington DS, Kodo N, Norwood DL, Roe CR. Tandem mass spectrometry: a new method for acylcarrntine profiling with potential for neonatal screening for inborn errors of metabolism. J Inher Metab Dis 1990;13:321-4. 8. Millington DS, Kodo N, Terada N, Roe D, Chace DH. The analysis of diagnostic markers of genetic disorders in human blood and urine using tandem mass spectrometry with liquid secondary ion mass spectrometry. hit J Mass Spectrom Ion Processes 1991; 111:211-28. 9. National Committee for Clinical Laboratory Standards. Blood collection on filter paper for neonatal programs: approved standard. NCCLS Publication LA4-A. Villanova, PA: NCCJ.S, 1988. 10. Falkner F. Comments on some common aspects of quantitative mass spectrometry. Biomed Mass Spectrom 1981;8:43-6. 11. Markey SF. Quantitative mass spectrometry. Biomed Mass Spectrom 1981;8:426-.30. 12. Shapira E, Blitzer MG, Miller JB, Affrick DK. Biochemical genetics: a laboratory manual. New York: Oxford University Press, 1989. 13. Millington DS, Terada N, Chace DH, Chen Y-T, Ding J-H, Kodo N, Roe CR. The role of tandem mass apectrometry in the diagnosis of fatty acid oxidation disorders. In: Coates PM, Tanaka K, eds. New developments in fatty acid oxidation. New York: Wiley-Lisa, 1991:339-54. CLINICAL CHEMISTRY, Vol.39, No. 1, 1993 71