Perturbation of serum carnitine levels in human adults by chronic renal disease and dialysis therapy14 Lavon L. Bartel, Ph.D., John L. Hussey, M.D., and Earl Shrago, M.D. ABSTRACT Serum camnitine levels in nondialyed and dialyed patients ith chronic renal disease ere compared against a group of normal control subjects. The concentration of serum carnitine as directly correlated ith that of serum creatinine (r = ±0.734; p < 0.00 1). In nondiahyed uremic patients the serum free camnitine levels in males rose 218% (p <0.001) and in females rose 186% (p < 0.001) above normal control values. During dialysis there as a sharp decline in serum camnitine to levels reaching 20% of the ero time control value (p <0.001). The decrease in serum camnitine could be accounted for by an almost quantitative accumulation of camnitine in the dialysate fluid. After termination of dialysis there as a hyperbolic rise in serum carnitine hich reached the high values again ithin 44 to 48 h. It is postulated that frequent perturbations in serum carnitine as a result of chronic dialysis therapy over a prolonged time period could potentially head to a tissue deficiency in camnitine ith its resultant complications. Am. J. Clin. Nutr. 34: 1314-1320, 1981. KEY WORDS Camnitine, renal insufficiency, renal dialysis It is generally accepted by most nutritionists that in mammals carnitine (fi hydroxy- -y-trimethylamino butyrate) is neither an essential nutrient nor a vitamin since it can be synthesied endogenously from the amino acids!ysine and methionine (1). Recent studies suggest, hoever, that in some subjects the ability of the tissues to utilie or retain carnitine may be inadequate (2). This condition is no a recognied clinical syndrome resulting from defective oxidation of long chain fatty acids, inefficient energy production and a derangement of intermediary metabolism (3, 4). Hemodialysis hich removes small molecular eight substances besides urea from the plasma has the potential of creating a carnitine deficiency state. This might be particularly evident in patients undergoing chronic dialysis over a prolonged time period ho either self-limit their dietary intake or are instructed not to eat the normal amount of meat protein hich is rich in carnitine. The possibility of a carnitine deficiency in this population ith end-stage renal disease deserves special consideration since it has recently been demonstrated that the kidney is a major site of carrntine biosynthesis in man (5). The present communication hich addresses itself to this problem is a study of the perturbation of serum carnitine by chronic hemodia!ysis in patients ith end-stage renal disease. Experimental Subjects methods All human subjects signed advised consent forms approved by the University of Wisconsin Medical School Human Subjects Committee. Normals Blood samples ere obtained by venipuncture from 64 normal, healthy, adult subjects. The 30 male subjects 1 From the Departments of Nutritional Sciences, Medicine, Surgery and The Clinical Nutrition Center, University of Wisconsin, Madison, Wisconsin 53706. 2 Supported in part by USPH Service Grant, AM- 15893 and the Wisconsin Division of the American Diabetes Association. An abstract of this ork as presented at the Federation of American Societies for Experimental Biology Meeting in Atlantic City, Ne Jersey, April, 1978. Address reprint requests to: Dr. Earl Shrago, 491 Medical Science Building, 470 North Charter Street, Madison, Wisconsin 53706. 1314 The American Journal of Clinical Nutrition 34: iuly 1981, pp. 13 14-1320. Printed in U.S.A. 1981 American Society for Clinical Nutrition
EFFECT OF DIALYSIS ON SERUM CARNITINE 1315 ranged in age from 18 to 62 yr, average age 30, and the 34 female subjects ranged in age from 19 to 45 yr, average age 28. Subjects ere employees and students of the University of Wisconsin-Madison. Chronic renal disease-nondialyed Eight subjects ere selected from outpatients attending the University Hospitals Renal Clinic. The four male subjects ranged in age from 19 to 33 yr, hile the four female subjects ranged in age from 21 to 42. All subjects ere diagnosed as having chronic kidney disease, and their blood urea nitrogens ranged from 47 to 129 mg/dl (normal, 6 to 23 mg/dh). Patients serum samples ere obtained from the University of Wisconsin Clinical Laboratory after their permission as given at a Renal Clinic visit. Chronic renal disease-hemodialyed A total of 34 subjects ere studied over a 2#{189}-yr period. Patients ere under a program of chronic intermittent hemodialysis at the University Hospitals. Generally the subjects ere outpatients coming in to to three times eekly for hemodialysis. In experimental studies involving samples obtained beteen dialysis, subjects ere inpatients generally hospitalied for minor surgical procedures involving fistuha replacement. Except for the inpatients, subjects ere studied under their usual patterns of life. Tenty of the subjects ere males ranging in age from 17 to 68 yr, average age 33, and 14 of the subjects ere females ranging in age from 26 to 59 yr. average age 38. All individuals ere undergoing chronic dialysis, and the minimum time since initiation of dialysis as 6 months. All patients ere functionally or surgically anephric. Blood samples ere taken by hospital staff nurses, generally from shunts. Chronic renal disease-peritoneal dialysis Three patients undergoing peritoneal dialysis alloed venous samples of their blood to be taken by a staff nurse before and after the peritoneal dialysis procedure. The male subject as 45 yr old, and the to female subjects ere 19 and 31 yr old. Renal disease-postrenal transplantation To patients in the original hemodiahysis group ere kidney transplant recipients at a later date. Blood samples ere dran at the Renal Transplant Outpatient Clinic 6 months after successful transplantation. The male patient as 23 yr old, and the female subject as 36 yr old at time of transplant. Experimental conditions Treatment of blood samples In normal controls and nondialyed subjects blood as dran about 4 h postprandiah. Both hemodiahysis and peritoneal dialysis as begun about 1 h after breakfast. During dialysis there as no ingestion of food. Blood samples of patients on hemodiahysis ere obtained from their dialying shunt or fistula. This blood can be considered to be an arterial or an arterial/venous mixed sample. All other samples ere obtained from arm veins. Blood samples ere immediately placed on ice and ithin one hour the clotted blood as centrifuged at 2000 rpm for 5 mm on a Dynac Clinical Centrifuge (Clay-Adams, ith angle head no. Cl416-5). Serum samples ere then analyed or ere froen at -20#{176}C. In dialysis patients, many samples ere hepamin-plasma instead of serum due to the nature of collection on a systemically heparinied subject. We as ell as others (6) found that heparin-plasma camnitine levels do not differ from serum camnitine levels. Hemodialysis procedure In determination of serum and dialysate levels of camnitine, blood, and diahysate samples ere taken hile the subjects ere undergoing dialysis on an Extracorpreah Diahyer ith a multiple pass system. This method involves repeated contact of the patient s blood ith a specific tank volume of diahysate fluid until the entire batch of dialysate is replaced. Dialysate tank volume of 100 1 as changed every 2 h during dialysis. The pump rate for the dialysis solution as 300 mi/mm. For all other determinations (diahysate not required) patients ere on a Gambro diahyer hich utilied a single pass-continuous flo system. This method involves a single contact of patient blood ith a continuously floing and, therefore, ne dialysate. The dialysis solution flo rate as 500 mh/min hile blood flo rate as maintained at 200 mh/min. Patients ere dialyed for either 4 h, three times per eek or 5 to 6 h to times per eek. The diahysate fluid contained 140 mm sodium, 2 mm potassium, 1.8 mm calcium, 0.76 mm magnesium, 108 mm chloride, 35 mm acetate, and 2 g/loo ml glucose. Patients received either an intravenous bohus (load dose) of 1500 to 2000 units of hepanin at the beginning of the dialysis procedure folloed by about 2000 units every 2 h of ongoing dialysis or ere on continuous systemic hepariniation. Peritoneal dialysis procedure Flo rate of the peritoneal dialysis solution as approximately 30 mh/min, and the solution contained 130 mm sodium, 1.8 mm calcium, 0.5 mm magnesium, 100 mm chloride, 34.5 mm acetate, and 2.5 g/100 ml dextrose. Patients ere dialyed three times a eek for 8 to 10 h per session. Analytical Carnitine methods The method used in this study for the determination of free carnitine as the radioisotopic enymatic assay of McGarry and Foster (7) hich is recognied for its simplicity and high sensitivity. For determination of total carnitine the sample to be analyed as added to 0.1 ml of 1.0 M Tris base and 0.05 ml of 0.4 N KOH and as alloed to stand at 30#{176}Cfor I h to hydrolye the acylcamnitines present. After neutraliation the assay as described for free camnitine (7) as carried out. Six standard solutions of camnitine falling ithin the range 12.5 to 300 mm, a substrate blank and an enyme blank ere analyed concurrently ith each series of samples. The radioactivity of the product, Il- 4C]acetyhcarnitine,
1316 BARTEL ET AL. obtained from the assay as corrected to dpm for each standard. These data ere entered into a formula developed for the Monroe 1860 Calculator to establish a linear regression equation. The ra values of each triplicate sample analysis ere then entered. The program reported the SD and SEM for each sample in the series. This procedure for determination of carnitine as satisfactory for serum, dialysate, and liquid foods, and also for tissue and solid foods that ere pretreated ith a perchloric acid extraction (8). Carnitine as determined in a selected group of foods eaten as meals by patients on hemodiahysis. Tissue as prepared as follos: Tissue previously froen in liquid nitrogen as placed in mortar and finely podered. The poder as transferred to a cold homogeniing tube and as homogenied to a slurry in 2 ml of 2.5% KC1O4. The slurry as transferred to a plastic Sorvah centrifuge tube together ith to ninsings each of 1 ml 2.5% KC1O4 and as centrifuged at 0 to 4#{176}Cfor 10 mm at 5000 X g. The supernatant as removed and the solid residue ashed to times ith 1 ml distilled ater, mixed, and centrifuged as above. All of the supernatants ere pooled together and the residue as dried in a 60#{176}Cdrying oven until a constant eight as reached to determine dry eight of the tissue sample. Approximately 1 ml of 2.2 M KOH as added to a sample of the pooled supernatant and the solution neutralied to ph 7.0 to 7.2. The tubes ere mixed, alloed to stand for 30 mm at 0 to 4#{176}Cand the KC1O4 precipitate removed by centrifugation. The clear supernatant as used for assay of free and total carnitine as described previously. The recovery of free, total, and acetylcarnitine (total minus free) as determined by the addition of L-carnitine chloride and acetylcarnitine to serum samples of a normal and a uremic dialyed subject (before dialysis). Recovery as highest for free carmitine but in all cases the recovery as greater than 86%. Creatinine Results All creatinine determinations ere carried out in the University Hospitals Clinical Laboratory as part of the subject s normal blood chemistry panel. Data ere obtained from the patients interim reports and correlated ith carnitine levels measured on a portion of the sam- A The levels of serum free carnitine in a normal subject population ere found to be 53.9 ± 12.2 mol/! for males and 45.5 ± 9.9 j.tmol/l for females (Table 1). These values hich are consistent ith those published in the literature (6, 10, 11) sho differences (p <0.00!) hich indicate that males maintain a somehat higher serum carnitine level than females (12). In nondialyed patients ith severe renal insufficiency the serum free carnitine levels in males rose 218% (p < 0.001) and in females 186% (p < 0.00!) above normal control values. Chronic hemodialysis as associated ith a slight reduction of the elevated serum carnitine levels observed in nondialyed patients, but they did not reach normal control values. The dialysis treatment appeared to eliminate the differences in serum carnitine levels noted in normal male and female subjects. In a male patient ho received a kidney transplant, serum carnitine measured 6 months later as still above normal hereas in a female transplant patient the serum carnitine as in the norma! range. In both patients serum creatinine levels ere normal. About 75% of the total carnitine in human serum is in the free form, 15% is acetylcarnitine, and 10% in other forms (13). In this study, analyses of free and total serum carnitine in 26 normal male subjects gave values of 53.8 ± 6.9 and 67.0 ± 14.3 jmol/l, respec- le Serum free carnitine values of normal subjects, patients ith chronic renal disease, and Statistical analysis patients ith chronic renal disease undergoing hemodialysis* The Student s t test (9) as used to determine the significance of the correlation coefficient of serum carnitine Carnitine and creatinine. Differences beteen group means Experimental group ere also analyed using this test. Male pmol/l Materials L-Carnitine chloride as obtained from P-L Biochemicals, Inc., Milaukee, WI as as acetyl coenyme A- lithium salt. Coenyme A [acetyh-l-t4c]-acetyh (40 to 60 mci/mmoh) as obtained from Ne England Nuclear, Boston, MA, Doex 1 X 10-400 anion exchange resin in its chloride form, carnitine acetyltransferase from pigeon breast muscle, and acetyicarnitine ere obtained from Sigma Chemical Co., St. Louis, MO. Sodium tetrathionate as obtained from ICN Pharmaceuticals, Inc., Plainvie, NY. Female p.mol/l Normal controls 53.9 ± 12.2 45.5 ± 9.9f n=30 n=35 Chronic renal disease Nondiahyed 117.7 ± 6.7t 84.4 ± loot n=4 n=4 Dialyed 72.8 ± l7.4t 72.6 ± 1 l.6f n=16 n=7 * Data are expressed as the average ± SEM. t Statistical analyses compare male versus female controls and the controls ith their respective experimental group; p < 0.001.
E tively. Calculation of acetyl + acy!carnitine by difference gave a value of 13.2 ± 12.! jimol/l. In four uremic patients both free and total carnitine ere increased proportionately (101.9 ± 6.0 versus 116.6 ± 2.6 mol/l) and the acetyl + acy!carnitine of 14.7 ± 5.3 mol/ 1 as thus similar to the normal control group. There also as no significant difference in the ratio of free to total carnitine during the dialysis procedure itself indicating that the distribution of carnitine beteen free and acy! forms did not fluctuate idely. Significant amounts of carnitine have been detected in urine (14) but the mechanism of its clearance by the kidneys has not been carefully studied. Diet, fasting, and activity may have dramatic effects on rena! excretion of carnitine causing difficulty in interpretation of urinary values (15). From the results shon in Table lit seems likely that renal insufficiency impairs the clearance of carnitine from the serum. This is verified by the results in Figure 1 hich sho a positive correlation (r = +0.734; p < 0.00 1) of carnitine versus creatinine levels in the serum of a number of subjects. Since creatinine, a metabolic aste product, is removed from the plasma by g!omerular filtration and is excreted in the urine ithout significant tubular reabsorption, it provides a standard method for determination of rena! clearance. The results hich sho a positive correlation be- EFFECT OF DIALYSIS ON SERUM CARNITINE 1317 I- lx 0 cx C,) r 0.734.. O CORRESPOND TO NORMALS 4 5 6 7 8 9 0 II 2 3 4 5 SERUM CREATININE mg/ioo ml FIG. 1. Serum carnitine versus serum creatinine levels correlation coefficient r = ±0.734. teen serum carnitine and serum creatinine provide evidence that carnitine is normally cleared by the kidneys and that the increase in the serum of uremic patients is due to the impaired renal function. The changes in serum carnitine ere next studied in a series of male and female patients during the hemodialysis procedure. As noted in Table 1 the patients ho ere maintained on dialysis every 44 or 68 h had serum free carnitine values of 135 to 160% above control levels (p < 0.001). The results in Table 2 sho. TABLE 2 Decline in serum free carnitine concentrations during dialysis* H Male Carnitine tmol/l % of 0 time Carnitine tmo1/l 4 of 0 time Female Hemodiahysis 0 72.8 ± 17.4 n=l6 2 43.2 ± 18.8j n=16 4 32.6 ± 14.4t n=l6 6 20.5 ± 6.3t n=6 Penitoneal dialysis 0 59.9 n=i 6 31.5 n=l * Data are expressed as the average ± SEM. 100 72.6 ± 11.6 n=7 59 43.3 ± 9.8t n=7 45 26.9 ± 10.8t n=7 28 16.7 ± IO.6t n=3 100 67.5 n=2 53 29.0 n=2 100 60 37 23 100 43 t Statistical analyses compare male and female 0 time against their respective experimental group; p < 0.001.
1318 BARTEL ET AL. that ithin 2 h into the dialysis procedure the serum carnitine fell to 60% of the 0 h level (p < 0.001) and continued to fall until at 5 to 6 h the values ere beteen 20 and 30% of the respective 0 h levels (p < 0.00!). In three patients ho underent chronic peritoneal dialysis for 8 to 10 h there as a marked decrease in serum carnitine at the end of the dialysis procedure. Hoever, the number of subjects as too fe to carry out a statistical analysis of the data. The loss of carnitine from the extracel!ular space appears to be nearly equal to its accumulation in the dialysate (Table 3). In this experimental group a!00-! system is replaced at the end of 2 h of dialysis ith a second batch of fresh dialysate for individual patients. The first four values of 300, 180, 110, and 346 mo! of extracellular fluid free carnitine represent the first 2 h of dialysis for each patient and the second group of four values, 144, 45, 105, and 65 jlmo! of free carnitine represent the second 2-h period of dialysis. Figure 2 follos the carnitine kinetics during the dialysis procedure in to patients undergoing hemodialysis. There is a linear drop in serum carnitine hile dialysis is continuing and a hyperbolic rise during the interdialysis period. On the ordinate of Figure 2 a plot of normal carnitine values indicates that serum free carnitine levels are belo normal controls for approximately 2 h during dialysis and 6 h postdia!ysis. By 10 h the serum free carnitine has reached levels above those of norma! subjects, and by the begin- TABLE 3 Comparative values in individual patients of carnitine lost from extracellular fluid and gained in diahysate fluid during hemodiahysis Subject 1 21 4 1 21 4J Dialysis O-2h 2-4h Extracellular fluid loss#{176} 300 180 110 346 144 45 105 * Extracehhuhar fluid space, 15 1. t Dialysate, 100 1. 65 Dialysate gino! of carniline 263-297 164-196 116-130 270-390 74-124 37-68 118-122 76-82 gaint fling of the next dialysis (44 to 64 h) the levels are at their peak. The sample menu of the dialysis patients folloed in Figure 2 as analyed for carnitine content (Table 4). The calculated protein, a) 0 E I- 0 (I, Diolysis-il-Post Diolysis-t. HOURS #{163}9 PATIENT* I #{149}c PATIENT $2 FIG. 2. Changes in serum carnitine levels in patients during and postdiahysis. #{149}, patient 1 is a 32-yr-old female; A, patient 2 is a 30-yr-old male; 4, average and SEM of triplicate analyses. Normal serum carnitine values are 45.5 j.smol/l in females and 53.9 j.lmol/l in males. TABLE 4 Carnitine values in sample menu of dialysis patients Meal Total carnitine Breakfast 1 cup hole milk 34.2 2 slices bread 1.4 I tbsp butter 1.7 2 eggs 2.6 6 o orange juice 4.1 Lunch 3 o patty chopped chuck 452.0 1 bun = 40 g bread 1.1 1 o cheese 5.0 /4 head lettuce 0 #{189} cup green beans 2.5 1 medium peach 12.8 1 cup hole milk 34.2 Dinner 3 o chicken breast 32.3 1 medium baked potato 0.6 2 tbsp butter 3.4 #{189}cupcorn 9.1 /4 head lettuce 0 #{189}cupgelatin 0.1 1 cup hole milk 34.2 pmo!/serv:ng 44.0 imol/meal 507.6 imol/meal 79.7 ycmoh/meah
EFFECT OF DIALYSIS ON SERUM CARNITINE 1319 fat, carbohydrate, and calorie content of the diet are approximately 1 13 g, 88 g, 180 g, and 1962 kcal, respectively, hoever, the exact intake of the patients as not recorded. It is apparent that the menu contains a considerable amount of carnitine, the content of hich as analyed for each dietary constituent. In general, carnitine is lo in foods of plant origin and high in animal foods. The data for carnitine in foods are scarce and unsatisfactory due, until recently, to poor assay conditions (15). Depending on the quantity of food ingested, the dietary intake of carnitine may, in addition to de novo biosynthesis account for the higher than normal serum values reached in the later part of the postdialysis period. Discussion The present study as devised to determine sequential changes in serum carnitine levels during and after hemodia!ysis. In patients ith chronic renal disease, either nondia!yed or maintained on chronic dialysis, e did not note differences from normal in the distribution of free and acylated forms of serum carnitine. This finding is in contrast to the results obtained ith diabetic subjects in ketoacidosis in hom a decrease in serum free carnitine as associated ith an increase in short-chain and long-chain acylcarnitines (13). What appears to be particularly important is that the sharp decline in serum carnitine to approximately 20% of the 0 time value at the end of dialysis is gradually reversed so that by 44 to 48 h it again reaches the 0 time value hich is higher than that of the normal population. In one patient studied on five different occasions throughout a year s time the same cycle repeated itself ith the drop in serum carnitine during dialysis folloed by the return to the control value, and at no time as the serum carnitine found to be lo other than during the dialysis procedure. Bohmer et al. (16) reported a significant drop in plasma carnitine and, like ourselves, an accompanying increase in dialysate carnitine in a series of patients ith chronic rena! disease. They did not, hoever, follo the carnitine kinetics after dialysis to determine hether or at hat time interval the serum carnitine returned to the predialysis level. From our results, it can be seen that the recovery of serum carnitine after hemodialysis occurs in a hyperbolic fashion ithin 44 h. This return to above normal levels may be due to both dietary intake and de novo biosynthesis by the liver. Both of the patients e studied ere reported to have eaten ell after dialysis, and it may be assumed that their dietary intake of carnitine over the next 48 h as adequate. The critical question is hether, over a prolonged period of dialysis therapy, the frequent perturbations in serum carnitine might be reflected in overall tissue carnitine changes. It is evident that the patients serum carnitine levels over a eeks interval ould be considerably belo normal from 20 to 30 h depending on hether they ere dialyed every 2 or 3 days. In the study of Bohmer et a!. (16) muscle biopsies ere obtained, and it as reported that in eight of nine patients examined muscle carnitine concentrations ere reduced to one-tenth of those found in control subjects. Since the biopsy samples ere hydrolyed in 2 M KOH before assay the values represent total and not just free carnitine. The results obtained in their patients on hemodialysis indicate an extreme deficiency in muscle carnitine and not merely a redistribution beteen the free and acylated forms. We did not do muscle biopsies on our subjects, and there have been no reports by other orkers of muscle carnitine levels in patients on chronic hemodialysis ith hich to compare the results of Bohmer et al. (16). Hoever, our on study hich seems to agree very ell ith that of Bohmer et a!. (16) in other respects, is consistent ith the possibility that patients on long-term hemodialysis therapy may suffer from a carnitine deficiency hich ould not be apparent from an analysis of serum samples alone. In support of Bohmer et a!. (16) it might be interpreted that the elevated serum carnitine level noted during the interdialysis period as in itself inadequate to promote efficient uptake and/or retention of carnitine by muscle. Recent biochemical studies on the organ site of carnitine biosynthesis has provided information of important differences in various animals. Whereas the specific enymes for carnitine biosynthesis are very likely similar in all animals, including man, the location of the individual reactions are considerably different (5, 17, 18). In the rat, the species
1320 BARTEL ET AL. most extensively studied, many tissues catalye the entire sequence of reactions from #{128}- N-trimethyl-L-!ysine to -y butyro-betaine (17-19). Hoever, only liver is able to catalye the final step from -y butyro-betaine to carnitine hich is then rapid!y transported and distributed to other organs of the body (17-19). More recently, the finding that butyrobetaine hydroxylase activity in kidneys of rhesus monkeys and man equals or exceeds that present in the corresponding livers, confirms the finding of species differences in tissue localiation of carnitine biosynthesis (5). This may be a particularly important observation as related to the present studies since it is apparent that in man chronic renal disease ould severely impair carnitine biosynthesis. The elevated levels of serum carnitine in nondialyed uremic patients may be due to a combination of dietary intake and de novo biosynthesis in the liver. The problem of a carnitine deficiency ould become exacerbated in subjects ho may not be eating adequately during chronic dialysis hich, in turn, removes a number of nutrients of lo molecular eight such as carnitine. In a small group of patients ith a genetic deficiency of carnitine palmitoyl transferase, muscular problems occur hich are similar to those noted in McArdle s disease (3). The syndrome results from the inability of muscle tissue to oxidie long-chain fatty acids for energy requirements, particularly during periods of fasting or chronic exercise. Bohmer et a!. (!6) have suggested that some aspects of neuromuscular asthenia and fatigue noted in patients on chronic hemodia!ysis might be attributed to a lack of muscle carnitine. We no intend to determine hether any abnormalities in lipid metabolism knon to occur in these patients might be associated ith a carnitine deficiency. El The authors appreciate the helpful discussions of Dr. Charles Elson. References I. Mitchell ME. Carnitine metabolism in human subjects. I. Normal metabolism. Am J Clin Nutr 1978;3 1:293-306. 2. Mitchell ME. Carnitine metabolism in human subjects. 3. Metabolism in disease. Am J Chin Nutr l978;3 1:645-59. 3. Dimauro 5, Dimauro P. Muscle carnitine palmitoyh transferase deficiency in myoglobinuria. Science 1973; 182:929-31. 4. Angehini C. Carnitine deficiency. Lancet 1975;2:554. 5. England 5, Carnicero HH. y-butyro betaine hydroxylation to cannitine in mammalian kidney. Arch Biochem Biophys 1978;l9O:361-4. 6. Bohmer T, Rydning A, Sohberg HE. Cannitine levels in human serum in health and disease. Chin Chim Acta l974;57:55-61. 7. McGarry id, Foster DW. An improved and simplified radioisotopic assay for the determination of free and estenified carnitine. i Lipid Res l976;17:277-81. 8. Pearson DJ, Chase JFA, Tubbs PK. The assay of (-)-carnitine and its O-acyh derivatives. Meth Enymol 1969;24:6 12-22. 9. Steel RGD, Torre ih. Principles and procedures of statistics. Ne York: McGra-Hill, 1960. 10. Cedenbald G, Lindstedt S. A method for determination of carnitine in the picomohe range. Clin Chim Acta 1972;37:235-43. 11. Dimauro 5, Scott C, Penn AS, Roland LP. Serum carnitine: index of muscle destruction in man. Arch Neurol 1973;28: 186-90. 12. Cederbald G. Plasma carnitine in body composition. Chin Chim Acta h976;67:207-l2. 13. Genuth SM, Hoppel CL. Plasma and urine carnitine in diabetic ketosis. Diabetes h979;28: 1083-7. 14. Cedenbhad G, Lindstedt S. Excretion of L-carnitine in man. Chin Chim Acta 197 l;33:l 17-23. 15. Mitchell ME. Camitine metabolism in human subjects. 2. Values of carnitine in biological fluids and tissues of normal subjects. Am J Chin Nutr l978;3 1: 481-91. 16. Bohmer T, Bengrem H, Eikhid K. Carnitine deficiency induced during intermittent hemodiahysis for renal failure. Lancet 1978;3: 126-8. 17. Cox RA, Hoppel CL. Biosynthesis of cannitine and N-trimethylaminobutynate from 6-N-trimethyh lysine. Biochem J 1973;l36:1O83-90. 18. Tamphaichitr V. Broquist HP. Site of carnitine biosynthesis in the rat. J Nutr 1974;104:1669-73. 19. Carter AL, Frenkel R. The role of the kidney in the biosynthesis of carnitine in the rat. J Biol Chem l979;254: 10670-4.