Determination of free and total carnitine with a random-access chemistry analyzer

Department of Pathology & Laboratory Medicine, School of Medicine, Loma Linda University, Clinical Laboratory, Loma Linda University Medical Center, Loma Linda, CA 92350.
^a Author for correspondence. Fax 909-824-4832; e-mail Hhubba{at}msn.com.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Carnitine deficiency presents as a major problem in fatty acid oxidation. The use of a plasma carnitine assay can rapidly help to describe this deficiency. The method we describe here requires two simple steps of sample preparation, followed by automated analysis with the Beckman Synchron CX4 random-access chemistry analyzer. The goal of this method development was to reduce the cost of analysis and to allow a greater number of laboratories to perform this assay on demand within 1 h for both free and total carnitine. The method has a linearity of 0–150 µmol/L and a detection limit of 5 µmol/L. The inter- and intraday CVs are <20%. The method agreed closely with both the widely used RIA and spectrophotometric methods.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
L-Carnitine (3-hydroxy-4-N-trimethylammonium butyrate) is the carrier molecule for long-chain fatty acids to cross the mitochondrial membrane (1)(2)(3). It is synthesized in the liver, requiring lysine and methionine as precursors and vitamin C as cofactor (4)(5). There are two isoforms of the enzyme (E.C.2.3.1.21)—the carnitine palmitoyl transferase (CPT)¹ I, which is located at the inner side of the outer mitochondrial membrane, transesterifies the long-chain fatty acid and transports the acyl CoA across the inner mitochondrial membrane, where CPT II is located, to break the acyl bond releasing the fatty acid, and recycling CoASH (2)(6)(7)(8). The consequences of primary carnitine deficiency and secondary deficiency caused by carnitine acyl dehydrogenase deficiency range from mild forms of muscle weakness (3) to severe forms of hypoglycemia and cardiomyopathy (9)(10)(11), from Reye-like syndrome (10)(12) to lipidosis, myopathy, and abnormal organic acid production due to the inability to utilize long-chain fatty acids as an energy source (1)(2)(10). Because long-chain fatty acids are involved in the synthesis of phospholipids such as lung surfactants, it is possible that carnitine deficiency contributes to the aggravation of cystic fibrosis and to some 10% to 15% of sudden infant death syndromes with a genetic deficiency of carnitine acyl dehydrogenase (1)(13). Therefore, the major and acute role of carnitine deficiency in a variety of abnormalities makes the provision of a widely available and easily performed carnitine assay highly desirable. This is particularly true for premature infants and infants under treatment for seizures with valproate and related anticonvulsants that suppress carnitine. The suppression of carnitine is also related to some chronic diseases associated with the aging process (6).

There are a number of methods to measure carnitine, such as radioisotopic (1)(14)(15), radioenzymatic (3)(13)(16)(17)(18), spectrophotometric (4)(18)(19)(20)(21)(22)(23)(24)(25), radioisotopic exchange with HPLC (13)(16)(26)(27)(28)(29), and tandem mass spectrometry (11)(30). Sample preparation steps for these methods include filtration and concentration (19)(22)(26)(31), solid-phase extraction (3)(12)(13)(14)(15)(16)(17)(18)(26)(27)(28)(29), and dialysis chamber extraction (32). Cederblad et al. (22) described an automated spectrophotometric method for carnitine determination with high precision, low reagent cost, and a short analysis time on a Cobas Bio centrifugal analyzer (22); it correlates well with the standard carnitine radioenzymatic assay (REA). Our method is compared with both the spectrophotometric method of Cederblad et al. and the standard radioenzymatic method (14). Compared with the method of Cederblad et al. (22), our method has greater sensitivity and linearity, random accessibility, stat capability, appreciably lower reagent costs, and no requirement for special heating equipment or expensive filtration apparatus; moreover, it provides exactly the same sample matrices for both free and total carnitine, enabling both values to be determined with one calibration curve, one dilution factor, and one reagent cartridge. With slight modifications in the procedure as used on the Beckman Synchron CX4, this method is easily adaptable on almost any of today's random access chemistry analyzers, and can be used to determine carnitine in most tissues and biological fluids (3)(18)(31).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
principle
L-Carnitine reacts with acetyl CoA catalyzed by carnitine acetyltransferase (CAT) to form acetyl L-carnitine and CoASH. CoASH reacts nonenzymatically with 5,5'-dithiobis-2-nitrobenzoate (DTNB) to form 5-thio-2-nitrobenzoate (TNB). The concentration of TNB is measured spectrophotometrically at 410 nm.

instrument
The instrument is a Beckman Synchron CX4 Chemistry Analyzer. An Eppendorf pipet is used for acid and base dispensing.

reagents
Three reagents were stored on board the instrument in a single user-defined reagent (UDR) cartridge. We dissolved 20 mg of DTNB (Sigma D8130) in 100 mL of 50 mmol/L HEPES buffer and placed it in compartment A of the UDR cartridge. It was sufficient for 500 tests. To make the HEPES buffer, 5 mL of 1 mol/L HEPES (Sigma H-7523) was diluted to 100 mL with phosphate buffer, which was made from dissolving 1.19 g of potassium phosphate monobasic (KH₂PO₄, JT Baker 1–3246) and 2.83 g of potassium phosphate dibasic (K₂HPO₄, JT Baker 3252–1) in 100 mL of type I water. The pH was adjusted to 7.5 with 5 mol/L NaOH before adding HEPES. Twenty-five milligrams of acetyl CoA (Sigma A2181) were dissolved in 10 mL of type I water and placed it in compartment B. It was sufficient for 450 tests. Fifty microliters of carnitine acetyltransferase (EC 2.3.1.7, Sigma C 8757) was diluted 1:100 by volume with type I water and stored in compartment C. It was only sufficient for 125 tests; therefore, three more refills of CAT could be made before the entire cartridge was discarded. All three reagents were stable for 3 months in the instrument reagent compartment, which was maintained at 2–6 °C. We made the protein-precipitating reagent by diluting 18 mL of 70% perchloric acid (HClO₄, Malinkrodt 2766) to 100 mL with type I water. It was stable at room temperature for 12 months. The 2 mol/L KOH at 11.2 g/L solution (JT Baker 3140–01, does not compensate for K₂CO₃ impurities) was used for both hydrolyzing the sample for total carnitine and neutralizing the HClO₄. It was stable for 6 months at room temperature. Biocell serum (Biocell Laboratory) was used to produce low and high controls.

calibrators and controls
Stock L-carnitine standard (5.0 mmol/L; Sigma C7518) was made by dissolving 98.7 mg of L-carnitine in 100 mL of type I water. Five working calibrators of 0.0, 10.0, 35.0, 75.0, and 150.0 µmol/L were made from stock L-carnitine calibrator with type I water. Once established, the calibration curve was verified by a two-point calibration comparing a zero blank (type I water) and a 150 µmol/L calibrator, which was performed biweekly or monthly, since the calibration curve was stable for at least 1 month. Every 6 months, a full linearity curve with all five calibrators was performed as required by CAP and CLIA regulations. Serum-based controls were made by addition of stock calibrator and 1 mmol/L palmitoyl carnitine aqueous calibrator (Sigma P4509) to Biocell serum to produce a low control and a high control. Both controls were stored at -17 °C, and the ranges were established by running triplicates everyday for at least 7 days. Controls were run with each batch of patient samples.

sample collection and patient preparation
Adult patients fasted for at least 4 h before venipuncture; children and pediatric patients fasted for at least 2 h. A lipemic specimen does not itself interfere with the assay, yet it can cause redistribution of the carnitine fractions in vivo (10)(34). Many foods, such as meat, dairy products, asparagus, and avocados (6)(34) contain carnitine, thus mandating the 4-h fasting for adults and 2 h for children. Prolonged fasting (>24 h), on the other hand, will cause an increase of acyl carnitine (10)(17)(22). Whole blood was collected in a 4.5-mL K₃EDTA Vacutainer Tube (Beckton Dickinson, 366536). For pediatric samples, we collected blood in a microtainer (Becton Dickenson, 5974) containing K₃EDTA. Blood was centrifuged at 3000g for 10 min, and plasma was separated and stored at -10 °C to -20 °C until assayed.

assay procedure
Two sets of 12 x 100 mm test tubes were added for each control or patient sample. Two hundred microliters of control or patient sample were added to both sets of the test tubes. To the first set, 10 µL of 2 mol/L KOH was added with an Eppendorf pipette, mixed gently, covered, and incubated at room temperature for 45 min to hydrolyze the ester bond. At the end of the incubation, 40 µL of 180 mL/L diluted HClO₄ was added to precipitate the protein; it was vortexed immediately for 10 s; another 30 µL of 2 mol/L KOH was added to neutralize the acidity; it was vortexed immediately for 10 s and centrifuged at 3000g for 5 min. The supernatant was transferred into a sample cup without causing foam and run on the analyzer for total carnitine. While the first set of tubes was incubating, the second set was processed by adding 40 µL of the 180 mL/L diluted HClO₄ to precipitate the protein; vortexed immediately for 10 s. After 40 µL of 2 mol/L KOH was added to neutralize the acidity, it was vortexed immediately again for 10 s, and centrifuged at 3000g for 5 min. The supernatant was carefully transferred into a sample cup of the CX4 and free carnitine was measured. Free and total carnitine shared the same user-defined chemistry, reagent cartridge, calibration curve, and dilution factor. The parameters for the CX4 user-defined chemistries are shown in Table 1 .

Because the controls and patient samples have a dilution factor of 1.4, whereas the calibrators are not diluted, the concentrations of the calibrators can be entered in the user-defined chemistry with this 1.4 factor, i.e., 35 µmol/L will be 49 µmol/L, and 75 µmol/L will be 105 µmol/L; thus the final instrument printout will give values corrected for the dilution.

statistics
Statistical analysis was performed by simple linear regression with SYSTAT software version 6.0.1 published by SPSS, 1996.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
linearity
The linearity of this method was between 0 and 150 µmol/L (Fig. 1 ). Five aqueous calibrators of 0, 10, 35, 75, and 150 µmol/L gave values of 0.1, 9.8, 35, 75, and 150 µmol/L, respectively (r = 1.000). The linear range encompassed the reference range of carnitine values of 20 to 100 µmol/L. A serial dilution of serum-based low control is presented in Fig. 2 .

View larger version (14K): [in this window] [in a new window]   Figure 1. Linearity of five aqueous calibrators.

View larger version (14K): [in this window] [in a new window]   Figure 2. Serial dilution of serum-based control (n = 4 for each concentration; graph represents mean ± SD).

precision
The intraday and interday (1 month) CVs for three concentrations were <20% (Table 2 ).

accuracy
Recovery of the method was 99.8% ± 0.7% for aqueous solution (n = 10) and 97.6% ± 3.9% for serum-based materials (n = 20). The method agreed well with the standard REA (x) for both free (n = 29) and total carnitine (n = 28): for free carnitine, y = 1.059x - 15.97, r = 0.942, SE = 0.072; for total carnitine, y = 1.09x - 12.357, r = 0.997, SE = 0.017 (Figs. 3 and 4). Consequently, the correlation for esterified carnitine was y = 1.077x 5.71, r = 0.996, n = 27. The method agreed also with spectrophotometric method of Cederblad et al (x): for free carnitine, y = 0.881x 0.998, r = 0.989, SE = 0.032, n = 29; for total carnitine, y = 0.861x - 0.029, r = 0.995, SE = 0.026, n = 23 (Figs. 5 and 6). Consequently, the regression for esterified carnitine was y = 0.824x 0.233, r = 0.983, n = 23. The filtrates were stable at room temperature for at least 1 h.

View larger version (15K): [in this window] [in a new window]   Figure 3. Comparison of free carnitine values measured by the proposed method and the radioenzymatic method.

View larger version (16K): [in this window] [in a new window]   Figure 5. Comparison of free carnitine values measured by the proposed method and the spectrophotometric method of Cederblad et al. (20).

detection limit
We established the detection limit of the method by assaying 15 aqueous calibrators of 3.0 µmol/L, which yielded a mean of 2.48 and SD of 1.69. The 3.0 µmol/L (rather than 0) was used because the CX4 does not accept readings less than zero. Taking the mean plus 3 SD gives a value of 7.55 µmol/L. We then subtracted the mean of 2.48 from 7.55 to give the estimated detection limit of 5.07 µmol/L.

stability of control materials
The serum-based controls were assessed for their stability when stored at -17 °C over a period of 6 months. The CVs of free and total carnitine in both the low and high controls are presented in Table 3 .

estimates of mean reference values
Our preliminary ranges of reference values are free carnitine 25–80 µmol/L and total carnitine 31–100 µmol/L, with females being slightly lower and newborns <3 months giving significantly lower values.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Method validation against the standard radioenzymatic method (14) run at Los Angeles Children's Hospital gave excellent correlation as shown in the results. In addition, it correlated well with the spectrophotometric method of Cederblad et al. (22), which is used at Corning Nichols Institute. Because that method has a positive bias compared with that of the radioenzymatic method with a 1.12 average slope (22), the current method apparently has offset this bias and improved the agreement with the radioenzymatic method, although there is still a positive bias with an average slope of 1.06. There were only a few specimens used in both method comparisons that belong to the "abnormal" range, which bear the potential to leverage the linear regression line generally used to interpret the acceptance of the comparison. However, when those few abnormal specimens were not used in constructing the linear regression curve and a new linear regression line was constructed solely on the basis of the group of normal values, the slopes changed slightly within the standard error, and extrapolation indicates that these few abnormal specimens did not alter the linear regression line or possible clinical significance with our assay values. We verified the ability of the assay to detect carnitine <20 µmol/L for infants older than 3 months, and <10 µmol/L for infants younger than 3 months (Fig. 2 ). The results indicate that we can distinguish carnitine deficiencies of <10 µmol/L and between 10 to 20 µmol/L.

The CX4 takes <8 min to measure the first sample and 30 s each sample thereafter; therefore, from the beginning of the operation, through incubation, until the reporting of both free and total carnitine is ~1 h, with labor involvement of approximately ~1 min per sample. The reagents are very stable compared with some published methods, which call for freshly made reagents (22)(23). Because we use rate measurement instead of end point, there are <2 min from the addition of CAT until the end of the measurement of absorbance, and the reported inactivation of CAT by DTNB is negligible (23). The CX4 chemistry analyzer and a common laboratory centrifuge are required. To perform the analysis with a manual spectrophotometer method requires considerable skill and expertise, which greatly increase the test cost and running time.

Our method requires a reasonably small amount of sample, with 400 µL of plasma for the determination of both free and total carnitine. Plasma separated from red cells is stable for 24 h at 2–8 °C; however, -17 °C or below in a glass vial is recommended for prolonged storage. The unique design of using acid and then base (40 µL each) in free carnitine, and using base, acid, base (10, 40, and 30 µL, respectively) in total carnitine makes the final reaction matrices the same for both free and total carnitine, which allows the use of the same UDR reagent cartridge, same calibration curve, and same dilution factor (x1.4 vs calibrators). Plasma specimens containing a high amount of protein consistently agglutinated with 20 µL of 2 mol/L KOH. We reduced the initial amount of KOH from 20 µL to 10 µL for the hydrolysis of the ester bonds and lengthened the incubation time from 30 min to 45 min to equate the hydrolysis of ester bond. Because the protein precipitate begins to dissociate at pH 4 and completely dissolves at pH 6 (35), repeated experiments were carried out to determine the types and the minimum amount of acids to use. Sulfosalicylic acid, metaphosphoric acid, and perchloric acid were compared. Perchloric acid was chosen because metaphosphoric acid was too weak and very unstable, and residual sulfosalicylate remaining after neutralization by alkali tended to hinder the following enzymatic process. Perchloric acid, however, is a strong acid with a small molecular structure, with its protein precipitation property derived solely from its acidity and ionic strength. Hence the residual acidity in the filtrate to prevent the dissolution of the precipitate is buffered by the first reagent in compartment A and, therefore, does not denature the enzyme when CAT is added.

We conclude that this carnitine assay is suitable for a large tertiary-care hospital, and also is usable in small clinical and research laboratories possessing automated chemistry analyzers. It achieves our goals for a rapid, low-cost method to measure free and total plasma carnitine.

View larger version (15K): [in this window] [in a new window]   Figure 4. Comparison of total carnitine values measured by the proposed method and the radioenzymatic method.

View larger version (15K): [in this window] [in a new window]   Figure 6. Comparison of total carnitine values measured by the proposed method and the spectrophotometric method of Cederblad et al. (22).

	Acknowledgments

 
We graciously thank Lawrence Sweetman and the Biochemistry Laboratory of Children's Hospital of Los Angeles for their continuous support in providing samples for method validation, and the Clinical Trial Department of Corning Nichols Institute for providing method validation samples. We also appreciate James Westerngard for his statistical expertise and advice. Financial support from the Department of Pathology, School of Medicine is greatfully acknowledged.

	Footnotes

 
¹ Nonstandard abbreviations: CPT, carnitine palmitoyl transferase; REA, radioenzymatic assay; CAT, carnitine acetyltransferase; DTNB, 5,5'-dithiobis-2-nitrobenzoate; TNB, 5-thio-2-nitrobenzoate; and UDR, user-defined reagent.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Barns RJ, Bowling FG, Brown G, Clague AE, Thompson A. Carnitine in dried blood spots: a method suitable for neonatal screening. Clin Chim Acta 1991;197:27-34. [ISI][Medline] [Order article via Infotrieve]
2. Coates PM, Tanaka K. Molecular basis of mitochondrial fatty acid oxidation defects. J Lipid Res 1992;33:1099-1110. [Abstract]
3. Harper P, Wadstrom C, Cederblad G. Carnitine measurements in liver, muscle tissue, and blood in normal subjects. Clin Chem 1993;39:592-599. [Abstract/Free Full Text]
4. Marquis NR, Fritz IB. Enzymological determination of free carnitine concentrations in rat tissues. J Lipid Res 1961;5:184-187.
5. Penn D, Ludwigs B, Schmidt-Sommerfeld E, Filon P. Effect of nutrition on tissue carnitine concentrations in infants of different gestational ages. Biol Neonate 1985;47:130-135. [ISI][Medline] [Order article via Infotrieve]
6. Rebouche CJ. Carnitine function and requirements during the life cycle. FASEB J 1992;6:3379-3386. [Abstract]
7. Vladutiu GD, Hogan K, Saponara I, Tassini L, Conroy J. Carnitine palmitoyl transferase deficiency in malignant hyperthermia. Muscle Nerve 1993;16:485-491. [ISI][Medline] [Order article via Infotrieve]
8. Vladutiu GD, Saponara Conroy, JM Grier, RE Brady L, Brady P. Immunoquantitation of carnitine palmitoyl transferase in skeletal muscle of 31 patients. Neuromusc Disord 1992;2:249-259. [Medline] [Order article via Infotrieve]
9. Stanley CA, et al. Chronic cardiomyopathy and weakness or acute coma in children with a defect in carnitine uptake. Ann Neurol 1991;30:709-716. [ISI][Medline] [Order article via Infotrieve]
10. Lee C. The clinical significance of dietary carnitine. J Med Technol 1985;2:528-531.
11. Van Hove JLK, Zhang W, Kahler SG, Roe CR, Chen Y-T, Terada N, et al. Medium-chain acyl-CoA dehydrogenase (MCAD) deficiency: diagnosis by acylcarnitine analysis in blood. Am J Hum Genet 1993;52:958-966. [ISI][Medline] [Order article via Infotrieve]
12. Schmidt-Sommerfeld E, Werner D, Penn D. Carnitine plasma concentrations in 353 metabolically healthy children. Eur J Pediatr 1988;147:356-360. [ISI][Medline] [Order article via Infotrieve]
13. Lloyd-Still JD, Powers CA, Wessel HU. Carnitine metabolites in infants with cystic fibrosis: a prospective study. Acta Padiatr 1993;82:145-149.
14. McGarry JD, Foster DW. An improved and simplified radioisotopic assay for the determination of free and esterified carnitine. J Lipid Res 1976;17:277-281. [Abstract]
15. Parvin R, Pande SV. Microdetermination of (-) carnitine and carnitine acetyltransferase activity. Anal Biochem 1977;79:190-201. [ISI][Medline] [Order article via Infotrieve]
16. Kerner J, Bieber LL. A radioisotopic-exchange method for quantitation of short-chain (acid-soluble) acylcarnitines. Anal Biochem 1983;134:459-466. [ISI][Medline] [Order article via Infotrieve]
17. Bhuiyan AKMJ, Jackson S, Turnbull DM, Aynsley-Green A, Leonard JV, Bartlett K. The measurement of carnitine and acyl-carnitines: application to the investigation of patients with suspected inherited disorders of mitochondrial fatty acid oxidation. Clin Chim Acta 1992;207:185-204. [ISI][Medline] [Order article via Infotrieve]
18. Sousa CD, English NR, Stacey TE, Chalmers RA. Measurement of L-carnitine and acylcarnitines in body fluids and tissues in children and in adults. Clin Chim Acta 1990;187:317-328. [ISI][Medline] [Order article via Infotrieve]
19. Cejka J, Kithier K. Serum carnitine quantification [Letter]. Clin Chem 1992;38:304-305. [Free Full Text]
20. Schafer J, Reichmann H. A spectrophotometric method for the determination of free and esterified carnitine. Clin Chim Acta 1989;182:87-94. [ISI][Medline] [Order article via Infotrieve]
21. Wan L, Hubbard RW. Rapid assay for free carnitine measurement in plasma [Abstract]. Clin Chem 1995;41:S159.
22. Cederblad G, Harper P, Lindgren K. Spectrophotometry of carnitine in biological fluids and tissue with a Cobas bio centrifugal analyzer. Clin Chem 1986;32:342-346. [Abstract/Free Full Text]
23. Maeda J, Dudrick S. Rapid spectrophotometric determination of plasma carnitine concentrations. J Parenter Enter Nutr 1990;14:527-532. [Abstract]
24. Rodriguez-Segade S, Alonso de la Pena C, Paz JM, Del Rio R. Determination of L-carnitine in serum, and implementation on the ABA-100 and CentrifiChem 600. Clin Chem 1985;31:754-757. [Abstract]
25. Seccombe DW, Dodek P, Frohlich J, Hahn P, Skala Campbell DJ. Automated method for L-carnitine determination. Clin Chem 1976;22:1589-1592. [Abstract/Free Full Text]
26. Minkler PE, Hoppel CL. Quantification of free carnitine, individual short- and medium-chain acylcarnitines, and total carnitine in plasma by high-performance liquid chromatography. Anal Biochem 1993;212:510-518. [ISI][Medline] [Order article via Infotrieve]
27. Poorthuis BJHM, Jille-Vlckova T, Onkenhout W. Determination of acylcarnitines in urine of patients with inborn errors of metabolism using high-performance liquid chromatography after derivatization with 4'-bromophenacylbromide. Clin Chim Acta 1993;216:53-61. [ISI][Medline] [Order article via Infotrieve]
28. Schmidt-Sommerfeld E, Penn D, Marinus D, Rinaldo P, Bennett MJ, Santer R, Stanley CA. Detection and quantitation of acylcarnitines in plasma and blood spots from patients with inborn errors of fatty acid oxidation. New Develop Fatty Acid Oxidation 1992;:355-362. [ISI][Medline] [Order article via Infotrieve]
29. Schmidt-Sommerfeld E, Penn D, Duran M, Bennett MJ, Santer R, Stanley CA. Detection of inborn errors of fatty acid oxidation from acylcarnitine analysis of plasma and blood spots with the radioisotopic exchange–high performance liquid chromatographic method. J Pediatr 1993;122:708-714.
30. Gaskell SJ, Guenat C, Millington DS, Maltby DA, Roe CR. Differentiation of isomeric acyl carnitines using tandem mass spectrometry. Anal Chem 1986;58:2801-2805. [Medline] [Order article via Infotrieve]
31. Roe DS, Terada N, Millington DS. Automated analysis for free and short-chain acylcarnitine in plasma with a centrifugal analyzer. Clin Chem 1992;38:2215-2220. [Abstract/Free Full Text]
32. Shihabi ZK, Oles KS, McCormick CP, Penry JK. Serum and tissue carnitine assay based on dialysis. Clin Chem 1992;38:1414-1417. [Abstract/Free Full Text]
33. Reichmann H, Lindeneiner NV. Carnitine analysis in normal human red blood cells, plasma, and muscle tissue. Eur Neurol 1994;34:40-43. [ISI][Medline] [Order article via Infotrieve]
34. Schmidt-Sommerfeld E, Penn D. Role of carnitine in children receiving total parenteral nutrition. Ferrari R Di Mauro S Sherwood G eds. L-Carnitine and its role in medicine: from function to therapy 1992:117-136 Academic Press San Diego. .
35. Xia LJ, Folker K. Improved methodology to assay carnitine and levels of free and total carnitine in human plasma. Biochem Biophy Res Comm 1991;176:1617-1623. [ISI][Medline] [Order article via Infotrieve]