<< back to Articles
Free and total serum carnitine: Shall the analyses begin?
Arthur F. Rosenthal, PhD
(This article appeared in the August 2002 issue of Dialysis & Transplantation.)
A greatly improved chromatographic method, which uses readily available derivatization reagents, is described for the analysis of free and total carnitine in serum. Using this method the approximate incidence of low free carnitine in an unselected group of dialysis patients was found to be 7.9%, but carnitine-supplemented patients often show enormously high levels both of free and total carnitine. Possible implications of these findings is briefly discussed.
The Centers for Medicare and Medicaid Services (CMS) is reportedly considering the requirement of monitoring blood carnitine levels in dialysis patients to whom carnitine is being administered as a drug. Actually, this type of monitoring has long been a standard recommendation of the manufacturer of the drug, as indicated in its PDR package insert submission.1
One result of CMS's consideration of making this analysis a requirement is the necessity for evaluating laboratory capabilities to accommodate the expected large increase in demand for carnitine testing. A study of these capabilities has, apparently, already been commissioned.
In this context, it seemed appropriate to describe the method for measurement of free and total carnitine that we have used in our laboratory successfully for over five years. Even though the present venue for presenting this information may seem unusual, the rapidity with which a dialysis-specific audience could thereby be addressed appeared to make this effort appropriate.
In common with all laboratory methods, ours has both advantages and disadvantages. The methodology utilizes HPLC (high-performance liquid chromatography), a standard analytical technique used both in analytical and clinical laboratories. In fact, this technique has already been developed for use as an established method for the analysis of carnitine in serum2-5, but, as will be discussed below, the labeling reagent to be used has traditionally been a problem.
HPLC methodologies usually have two phases: a preparatory one, in which the sample is being converted to a form in which its analyze of interest can readily be measured, and a measurement phase, in which the quantitative determination is actually performed.
The major drawback of the method, characteristic of many HPLC methods, is that it is somewhat labor-intensive, especially in its preparatory phase. Despite this, it is still capable of running fairly large batches. Potentially, the employment of modern sophisticated, automated preparative instruments could make the sample preparation much simpler. Furthermore, the measurement phase itself is inherently adapted to a considerable degree of automation.
The major advantage of the method is the low cost and simplicity of the technology. Rather than expensive immunological reagents being used, only inexpensive and simple chemicals are needed. Thus, once the method is set up, its running costs related to reagent consumption are very low.
Development of Methodology
High-performance liquid chromatography of carnitine after derivatization to form its UV-absorbing p-bromophenacyl ester has always been a very attractive, potentially low-cost way to analyze carnitine in serum. Presumably the positively charged quaternary ammonium group confers sufficient chromatographic uniqueness, so that the carnitine derivative is well separated from other bromophenacyl derivatives formed simultaneously, and quantitation becomes quite easy.
The disadvantage of this approach, however, is that the reagent used to form the ester presents a problem. The inexpensive and readily available p-bromophenacyl bromide has not gained acceptance as a derivatizing reagent for the analysis of carnitine in serum, because with current procedures it is not reactive enough to form the ester in high enough yield to be analytically useful. There is a report6 of its use for analyzing acylcarnitines in urine in the presence of diisopropylethylamine, but in our hands this approach failed completely for serum carnitine itself, giving only small amounts of an impure derivative.
To obtain a significant improvement in reactivity to allow the analysis to proceed, p-bromophenacyl trifluoromethylsulfonate (or, -triflate) has been used. When pure, this substance readily yields a good derivative within a few minutes in the presence of magnesium oxide2-5. Commercial grade bromophenacyl triflate, however, is not only quite expensive, but is so impure as to make the whole analysis of limited utility.
The synthesis of the reagent is also highly inconvenient, requiring the use of rather large amounts of diazomethane. Although one investigator has improved the facility of this synthesis,7 it was still necessary to recrystallize the reagent to achieve useful purity, which is also hardly a practical approach with the expensive and highly reactive commercial substance.
Thus, we were led back to attempting to improve the p-bromophenacyl bromide procedure so that it could become feasible for routine serum carnitine analysis. Many possible catalysts and bases were tried, until finally it was found that tetrabutylammonium hydroxide gave a greatly improved reaction, affording a high yield of a very pure bromophenacyl ester under analytically reasonable conditions. Despite its formal basicity, the tetrabutylammonium hydroxide/bromophenacyl bromide reagent fortuitously proved to be very ineffective in hydrolyzing acylcarnitines during the derivatization process. Therefore, the reagent could be used to determine both free carnitine in the presence of acylcarnitines (the normal mixture in serum), as well as total carnitine after a preliminary hydrolysis with conventional base.
The reagent cost for this analysis is, thus, so minimal as to be almost negligible, in contrast with the enzymatic methodology,8,9 for which reagent cost is sizable. When performing both a free and total carnitine analysis by our procedure, only the cost of the centrifugal dialyzer (ca. $1.50) need be considered a significant expense.
The method described has been used both for small workloads and for as many as 100 specimens in a 24-hour period on a single chromatograph. The analysis of large numbers of specimens is made possible by the completely automated nature of the autosampler-actuated chromatography, which, if need be, allows completely unattended operation throughout the night.
In summary, our method for the analysis of free and total carnitine appears to fulfill the criteria of low cost, specificity, and adaptability both to large and to small sample loads.
Materials and Methods
Carnitine (zwitterion and chloride), acetylcarnitine chloride, octanoylcarnitine chloride, stearoylcarnitine chloride, p-bromophenacyl bromide, and tetrabutylammonium hydroxide were all purchased from Sigma. Acetonitrile and methanol were obtained from Fisher Scientific, and other chemicals were purchased from standard sources. Centrifree® dialysis filters were obtained directly from Amicon.
Analysis of Serum Carnitine
Principle
When only free carnitine is desired, serum proteins are precipitated by acetonitrile: methanol (9:1) and an aliquot of the supernatant is dried and freed of chloride by shaking with a mixture of anhydrous disodium phosphate and silver oxide. When both free and total carnitine levels are required, the serum proteins are removed by the method of Cejka and Kithier,10 using an Amicon Centrifree centrifugal dialyzer. This allows hydrolysis of acylcarnitines on an aliquot of the dialyzed sample in the absence of serum proteins, which would interfere at this stage. Using this technique, both "free" and "total" aliquots are treated with acetonitrile-methanol as before, and shaken with disodium phosphate, silver oxide, and, in addition, powdered potassium dihydrogen phosphate to remove excess sodium hydroxide. The insoluble salts are then removed by centrifugation. An aliquot from any of the purification methods described is then treated with the derivatizing reagent and analyzed by HPLC.
Method
The method used is extensively modified from Lever, et al.,2 following the earlier work of Ingalls.3,4 A standard containing 100 nmol/ml of carnitine in water was prepared. Controls were prepared by spiking 7.5% albumin with carnitine as follows: To 97 µl of albumin is added 3 µl 1 nmol/µl carnitine; this is a 30 nmol/ml control. Similarly, a 70 nmol/ml addition control is made by adding 7 µl of spiking solution to 93 µl of albumin.
For samples requiring free carnitine alone, to 100 µl of standard or patient serum in a polypropylene tube was added, in order: 1.0 ml acetonitrile:methanol 9:1, and about 300 mg (visual estimate) of a mixture of 9 parts Na2HPO4 (i.e., anhydrous), 1 part Ag2O, and about 300 mg of KH2PO4 mixture was shaken gently for 1 hour and centrifuged for 5 minutes to give a clear supernatant. For samples requiring both free and total carnitine, 1 ml of the serum was first spun in an Amicon Centrifree centrifugal dialyzer until at least 300 µl of dialysate was collected. A 100-µl aliquot was removed for total carnitine analysis; 10µl of 10M NaOH was added, and the mixture heated for 30 min at 60°C. A second 100-µl aliquot of the filtrate, with the hydrolytic step omitted, was taken for free carnitine analysis. To each aliquot was added 1 ml of acetonitrile-methanol, and then the mixture of phosphates and silver oxide as before. Out of this, a 400-µl aliquot was taken and transferred to a chromatographic vial. The derivatizing reagent was prepared as follows: p-bromophenacyl bromide (40 mg) was dissolved in 1.0 ml anhydrous acetonitrile, and to this solution was added 50 µl of 40% aqueous tetrabutylammonium hydroxide. The reagent appears to be stable for at least 24 hours. To each tube containing the 400 µl organic extract was added 50 µl of the reagent; the tube was then capped, and the solution was warmed to 60°C for 120 minutes. The analysis was then started, using an appropriate chromatographic command sequence. Chromatography was performed using a Perkin Elmer 3x3 (3.3 cm x 4.6 mm; 3 µm spherical particles) silica column fitted with an appropriate guard cartridge, a Perkin Elmer Turbochrom® HPLC system consisting of a 235C detector (set at 260 nm), and a Series 200 Autosampler and pump controlled by a 600 LINK and Pentium® 100 computer equipped with Turbochrom 4.1 software. The mobile phase was 530 µl (4 mmol) of triethanolamine and 336 mg (1.6 mmol) anhydrous citric acid in 250 ml water in isopropanol; flow rate was 1.0 ml/min. The excess reagent eluted in about 1 min, and the carnitine derivative in ca. 3.10 min. Few additional peaks of any significant size were seen. A total isocratic run time of about 6 min was optimal, using an additional 4 min of equilibration to clean out the column. Carnitine was quantitated by using a single 100 nmol/ml aqueous standard, run in duplicate.
Figure 1. Typical chromatogram of carnitine (as p-bromophenacyl ester) in serum. the quantity of carnitine
(Rt 3.10) in this sample was 64.3 nmol/ml. Note the absence of nearby possible interfering peaks.
Results
Chomatography
Figure 1 is a typical chromatogram from a serum sample, showing the symmetry and clear separation of carnitine (Rt in different runs averages 3.10 min) from possible interfering substances. No blank was necessary, since neither reagent alone nor 7.5% albumin extracted and taken through the procedure gave any peak in the carnitine region.
The linear range of the method is shown in Figure 2. Linearity over at least two decades, between 10 nmol/ml and 1000 nmol/ml, is evident. Higher levels were not attempted because they are never encountered, but since there was no deviation from linearity even at 1000 nmol/ml it seems likely that the curve could be extended well beyond that value.
Figure 2. Computer generated standard curve of carnitine analyses. Values above 1000 nmol/ml are projected. Correlation coefficient was 0.9998.
Effect of Derivatization Procedure Itself on Acylcarnitines
Proof that the derivatization procedure does not itself result in appreciable hydrolysis of acylcarnitine is shown in a set of short experiments in which residual carnitine was analyzed in each of the acylcarnitines. Acetylcarnitine, octanoylcarnitine, and octadecanoylcarnitine at 100 ng/ml were taken through the standard derivatization procedure, as well as a shortened version in which the incubation was performed for 30 minutes. In addition, a separate hydrolytic run (as in the total carnitine determination above) was also performed for each acylcarnitine; data from the latter runs were taken as the respective 100% maximal hydrolysis values. The values given represent the carnitine found in each acylcarnitine as a % of the carnitine found after hydrolysis of the respective acylcarnitine.
The results obtained are shown in Table I. The data indicate that, rather than hydrolysis occurring during derivatization, the acylcarnitines as purchased are actually contaminated with a small amount of carnitine. The indicated lack of hydrolysis is also supported by the data below, which indicates that in every case the patient samples show a significantly higher total than free carnitine value, as of course it should be.
These experiments, incidentally, indicated that the acylcarnitines themselves were also derivatizable and showed the following retention times (min): acetylcarnitine, 5.71; octanoylcarnitine, 2.94; octadecanoylcarnitine, 2.63. No attempt was made to use these data analytically at this point, however.
Table I. Effect of analytical procedure on acylcarnitines. Shown is the percent of carnitine found in three acylcarnitines after carrying out the derivatization procedure for 30 and 120 min.
% of Total Carnintine
|
Carnitine Derivative |
30 minutes |
2 hours |
|
Acetylcarnitine |
3.1 |
3.8 |
|
Octanoylcarnitine |
2.2 |
2.2 |
|
Octadecanoylcarnitine |
4.4 |
4.7 |
Recovery of Carnitine Added to Serum
Twenty-eight serum specimens were analyzed for free carnitine, and 100-µl aliquots were spiked with carnitine in known amounts ranging from 20 to 100 nmol/ml, contained in 2- to 10-µl aliquots. These were again analyzed and the differences between the expected values and the levels actually found were then compared. Mean recovery of carnitine was 97.0% ?± 8.2% (c.v., 8.49%).
Serum Free and Total Carnitine Levels in Dialysis Patients
In the subsequent discussion, all specimens were obtained just prior to dialysis. An otherwise unselected group of 166 dialysis patients not on carnitine supplementation were analyzed for free carnitine by the direct solvent precipitation method. The mean carnitine in these patients was found to be 47.1 nmol/ml; range 26.8 - 67.4.
Another series of 85 specimens was also run by the dialysis filter method; these gave a mean of 47.5; range 26.5 - 68.5. Of these, 44 were run by both methods; the centrifugal dialysis method gave results that averaged 105.2% ±? 5.4% of the values for the direct (non-filtered) method. One factor in this minor difference may be the small concentrating effect of removing the serum proteins prior to analysis.
A series of 93 specimens analyzed for total carnitine gave a mean value for this parameter of 71.0 nmol/ml; range 40.6 - 101.4.
To differentiate male from female dialysis values, another series of patient samples was analyzed; results are given in Table II.
The approximate incidence of carnitine levels below 20 nmol/ml in our series of unsupplemented dialysis patients was 7.9%, and each new series run is usually found to show about the same incidence. This suggests that low carnitine levels in dialysis patients is probably a significant potential problem, and its correction may be warranted in those cases where it is uncovered.
Patients on carnitine supplementation often show very high blood levels of carnitine. A group of 46 such patients was found to give free carnitine levels ranging from 90 to 558 nmol/ml. Acyl pools are apparently not depleted by such levels, since the total carnitine levels were proportionately high, ranging from 126 to as much as 840 nmol/ml, or over 10 times the normal values. Whether or not such levels are benign is an open question, as discussed below.
Table II. "Normal" values for carnitine in dialysis patients
|
Constituent, Gender |
N |
Mean |
Range |
|
Free carnitine, F |
73 |
38.0 |
23.1 - 52.9 |
|
Total Carnitine, F |
35 |
63.0 |
40.2 - 85.8 |
|
Free Carnitine, M |
96 |
46.0 |
30.9 - 61.1 |
|
Total Carnitine, M |
73 |
74.2 |
46.2 - 102.2 |
All these values are quite close to those reported in standard sources.21
Discussion
There is now ample evidence to support the view that carnitine levels may be depressed in some dialysis patients, primarily in the free carnitine fraction.11 This may be reflected in lower muscle carnitine levels, with consequent decrease in exercise performance.12-14 It seems clear that the dialytic process itself results in loss of carnitine into the dialysate,15 but many patients are able to compensate with endogenous synthesis or with carnitine from exogenous sources. One factor limiting the former mechanism is the fact that the normal kidney is itself a major site of carnitine biosynthesis. Carnitine monitoring is becoming more common as its usefulness becomes more apparent, both for supplemented and non-supplemented patients. There is, first of all, no other good way to determine which patients are deficient in carnitine and which are not. Moreover, free carnitine values tend to decrease in dialysis patients more than do total carnitine values, so that both values together will allow a fairly complete picture of carnitine status. Judicious monitoring of the free/total ratio should, thus, allow an indication of whether muscle weakness or cramps experienced by any given dialysis patient might have carnitine deficiency as an important component. While the majority of studies have failed to show any clear effect of carnitine supplementation alone on lipid profiles,11,16 combination therapy of carnitine with simvastatin is claimed to show a synergistic beneficial effect on lipid parameters.17 The question of deleterious effects of long-term elevated carnitine in continuously supplemented patients remains open. Although the substance has a low order of acute toxicity according to the usual criteria,1 there are reports that long-term high-dose administration tends to promote abnormalities in platelet aggregation.18,19 A paradoxical myasthenia in some dialysis patients on high-dose carnitine supplementation has even been reported, but since DL-carnitine was used as the drug, the effect may well have been due to its content of D(+)-carnitine.20 The potential utility of chronic monitoring of free/total carnitine levels is again underscored by these studies.
The author wishes to express his appreciation to Ms. Jonette Jackson for her skilled technical assistance in the development of the methodology
References:
1. Medical Economics Co. Inc. Carnitor®. Physicians' Desk Reference 2002, 56th Ed., 3242-3245.
2. Lever M, Bason L, Leaver C, Hayman CM, Chambers ST. Same-day batch measurement of glycine betaine, carnitine, and other betaines in biological material. Anal Biochem 1992; 205:14-21.
3. Ingalls ST, Minkler PE, Hoppel CI, Norlander JE. Derivatization of carboxylic acids by reaction with 4'-bromophenacyl trifluoromethylsulfonate. J Chrom 1984; 299: 365-376.
4. Minkler PE, Ingalls ST, Hoppel CL. High-performance liquid chromatographic separation of acylcarnitines following derivatization with 4'-bromophenacyl trifluoromethylsulfonate. Anal Biochem 1990; 185: 29-35.
5. Minkler PE, Ingalls ST, Kormos LS, Weir DE, Hoppel CL. Determination of total carnitine in human urine by high-performance liquid chromatography. J Chrom 1984; 336: 271-283.
6. Poorthuis BJ, 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'-phenacylbromide. Clin Chim Acta 1993; 216: 53-61.
7. Rosenthal AF. unpublished work.
8. Bieber LL, Lewin L.M. Measurement of carnitine and O-acylcarnitines. Methods in Enzymol 1981; 72: 276-279.
9. Pearson DJ, Chase JFA, Tubbs PK. The assay of (-)carnitine and its O-acyl derivatives. Methods in Enzymol 1969; 14: 612-615.
10. Cejka J, Kithier K. Serum carnitine quantification. Clin Chem 1992; 38: 304-305.
11. Golper TA, Wolfson M, Ahmed S, Hirschberg R, Kurtin P, Katz LA, Nicora R, Ashbrook DW, Kopple JD. Multicenter trial of L-carnitine in maintenance hemodialysis patients. I: Carnitine concentration and lipid effects. Kidney Int 1990; 38: 904-911.
12. Guarnieri G, Situlin R, Biolo G. Carnitine metabolism in uremia. Am J Kidney Dis 2001; 38(4 Suppl 1): S63-S67.
13. Ahmed S. L-Carnitine in dialysis patients. Semin Dial 2001; 14(3): 209-217.
14. Brass EP, Adler S, Sietsema KE, Hiatt WR, Orlando AM, Amato A. Intravenous L-carnitine increases plasma carnitine, reduces fatigue, and may preserve exercise capacity in hemodialysis patients. Am J Kidney Dis 2001; 37:1018-1028.
15. Alhomida AS, Sobki SH, al-Sulaiman MH, al-Khader AA. Influence of sex and chronic hemodialysis treatment on total, free, and acyl carnitine concentrations in human serum. Int Urol Nephrol 1997; 29: 479-487.
16. Savica V, Bellinghieri G. Carnitine and lipid profile in uremia. Clin Ter 1997; 148: 229-236.
17. Savica V, Bellinghieri G, Lamanna F. The hypotriglyceridemic action of the combination of L-Carnitine + simvastatin vs. L-carnitine and vs. simvastatin. Clin Ter 1992; 140: 17-22.
18. Kalinowski M, Popawski A, Mazerska M, Daniluk A. Effects of L-carnitine on erythropoiesis and blood platelet aggregation in patients with chronic renal failure treated with hemodialysis. Pol Merkuriusz Lek 1999; 6: 76-78.
19. Wechsler A, Avram M, Levin M, Better OS, Brook JG. High dose of L-carnitine increases platelet aggregation and plasma triglyceride levels in uremic patients on hemodialysis. Nephron 1984; 38: 120-124.
20. De Grandis D, Mezzina C, Fiaschi A, Pinelli P, Bazzato G, Morachiello M. Myasthenia due to carnitine treatment. J Neurol Sci 1980; 46: 365-371.
21. Leavelle DE (ed.). Carnitine. In: Mayo Medical Laboratories Interpretive Handbook. Rochester, MN: Mayo Medical Laboratories, 1994, pp 98-99.
<< back to Articles