Carnitine level in Chinese epileptic patients taking sodium valproate

Introduction 

The relationship between valproate treatment and blood carnitine levels in epileptic patients is controversial. Several studies have reported that blood carnitine concentrations were lower in patients taking valproate than in healthy control subjects [1], [2], [3], [4], [5], [6]. However it has also been reported that carnitine concentrations were within the normal range in patients taking valproate but free of abnormal neurologic findings and nutritional problems [7]. It has been speculated that valproate causes low carnitine levels by interference with carnitine uptake [8], [9] or impairment of renal tubular absorption of carnitine [10]. The clinical significance of “low” carnitine levels in patients taking valproate is unknown. Coulter proposed that carnitine deficiency has a part to play in the causation of valproate-associated hepatotoxicity, a potentially fatal complication [11]. Carnitine deficiency was found in several cases of valproate-induced hepatotoxicity and in some cases of Reye’s syndrome-like illness associated with valproate therapy [10]. There was also a recent report of improved survival associated with intravenous supplementation of carnitine in patients with severe and symptomatic valproate-induced hepatotoxicity [12].

Risk factors that have been associated with carnitine deficiency include young age (less than 10 years of age), treatment with multiple antiepileptic drugs, presence of multiple neurologic disabilities (mental retardation, cerebral palsy, blindness, microcephaly), nonambulatory status, and underweight [13].

Because it is not our routine practice to monitor the carnitine levels of patients taking valproate, the relationship between valproate intake and carnitine deficiency in our patients is unknown. The aim of this study is to evaluate the carnitine status of epileptic patients taking valproate to determine the prevalence and risk factors of carnitine deficiency in these patients.

Materials and methods 

All epileptic patients being treated with sodium valproate, either as monotherapy or polytherapy, were enrolled from the general pediatric clinics or pediatric neurology clinics of the Prince of Wales Hospital, Shatin, Hong Kong, if they fulfilled the following criteria: (1) ethnic Chinese, (2) 19 years of age or less, and (3) having received regular valproate treatment for more than 2 months. Their growth parameters, demographic data, presence of possible risk factors for low carnitine level including the number of antiepileptic drugs received, presence of multiple neurologic disabilities (mental retardation, cerebral palsy, blindness, microcephaly), ambulatory status, and history of epilepsy were recorded. Blood was collected for the evaluation of liver function and estimation of free and acylcarnitine level, which was determined by tandem mass spectrometry (Quattro II; Micromass, Manchester, UK). The associations between blood carnitine levels and risk factors (presence of multiple neurologic disabilities or feeding problems, being nonambulatory, number of antiepileptic drugs, age, body mass index, dosage of valproate) were analyzed by univariate analysis, followed by multiple logistic regression when significant associations with any of the risk factors were identified. The study period was from June 2000 to July 2001.

Results 

A total of 43 patients were recruited during the study period. The demographic data are presented in Table 1.

Table 1. Demographic characteristics of the study group

	Age	0.9-18.7 year (mean 9.1 year, S.D. 4.3 years)
	Sex	F:M = 21:22
	Number of drugs	One vs > one = 26:17
	Dosage of valproate (mg/kg/day)	9.8-48.95 (median 23.66, twenty-fifth percentile 16.81, seventy-fifth percentile 31.28)
	Mental development	Normal vs mental retarded = 22:21
	Mobility	Ambulatory vs nonambulatory = 32:11
	Multiple neurologic disabilities	Present vs absent = 12 vs 31
	Feeding problem	Present vs absent = 10:33
	Body mass index	10.8-17.7 (median 15.48, twenty-fifth percentile 13.63, seventy-fifth percentile 16.96)
	Age of first seizure (months)	0.5-161 (mean 39.6, S.D. 38)
	Duration of seizure (months)	4.2-219 (mean 69, S.D. 53.4)

Abbreviation:S.D.=Standard deviation

Of the patients, 26 (60.4%) received sodium valproate as the only antiepileptic drug; 15 patients (34.9%) were taking two antiepileptic drugs, and two patients (4.7%) were taking three antiepileptic drugs. The mean (S.D.) dosage of sodium valproate was 25.7 (10.4) mg/kg/day (range = 9.8–48.9 mg/kg/day). Thirty-six patients (83.7%) had one or more risk factors and 10 patients (23.2%) had three or more risk factors, as mentioned above. All patients had normal serum liver enzymes and bilirubin level. Two patients had plasma free carnitine levels below our reference range. No significant associations were identified between the carnitine levels and any of the risk factors by using Mann-Whitney U test (Table 2).

Table 2. Association of carnitine level with the possible risk factors

	Risk Factors	Category	Carnitine level (μmol/L) median 25-75 percentile)	Test Used	P*
	Age (yrs)	< / = 10 (n = 28)	38.8 (33.8-45.1)	Mann-Whitney U	0.178
		> 10 (n = 15)	43.75 (38.35-47.15)
	Body mass index	< / = 13.63† (n = 11)	43.1 (27.7-56.15)	Mann-Whitney U	0.671
		> 13.63 (n = 32)	39.4 (35.1-45.1)
	Neurologic disabilities	Yes (n = 12)	41.8 (37.6-45.4)	Mann-Whitney U	0.801
		No (n = 31)	39.2 (35.3-47.1)
	Ambulatory	Yes (n = 31)	38.9 (35.3-46.3)	Mann-Whitney U	0.359
		No (n = 12)	42.7 (37.6-45.9)
	Feeding problem	Yes (n = 10)	44.05 (38.8-47.1)	Mann-Whitney U	0.127
		No (n = 33)	39.0 (33.8-45.1)
	Number of drugs	1 (n = 26)	39.5 (35.3-46.3)	Mann-Whitney U	0.82
		> 1 (n = 17)	41.8 (36.1-45.9)
	Valproate (mg/kg/day)	< / = 23.66‡ (n = 24)	39.8 (36.17-47.2)	Mann-Whitney U	0.588
		> 23.66 (n = 19)	39.3 (33.37-45.4)

* Correlation is significant at the 0.05 level (two-tailed). † Twenty-fifth percentile. ‡ Median.

Discussion 

Carnitine is an amino acid derivative that is present in most human tissues. It is evident in highest concentration in products derived from red meat and milk. It plays an important role in fatty acid transport and metabolism. Carnitine deficiency has been observed in metabolic diseases such as fatty acid oxidation disorders, glutaric aciduria I, and mitochondrial disorders [13]. There were also numerous reports on carnitine deficiency in patients taking valproate [1], [2], [3], [4], [5].

The range of plasma free carnitine in our local patients was 19.3–53.9 μmol/L. The definition of carnitine deficiency as quoted by De Vivo et al. [14] was free carnitine less than 20 μM, which was similar to the lower limit of our reference range. The reported incidence of carnitine deficiency in patients taking valproate was 4% to 76% [1], [2], [3], [4], [5]. The reason for such wide variation is unknown. It can be partly attributed to different patient selection criteria in different studies. Although a reduction in plasma carnitine after valproate intake has been widely reported the carnitine levels rarely fell below the normal range, and there was no clear correlation between the level and symptoms of carnitine deficiency [1], [6], [13]. It is therefore important that “low” carnitine levels should be interpreted with caution. Another issue complicating the interpretation of plasma carnitine level is the distribution of carnitine in the body. Approximately 90% of total body carnitine is deposited in the muscle tissue, in which the concentration of carnitine in muscle is as much as 10 times higher than that in the blood. This results in a lag time between the decline in plasma carnitine and carnitine depletion in the body. Therefore, whereas low plasma carnitine concentrations reflect low muscle tissue concentrations, normal plasma carnitine concentrations do not necessarily represent normal body carnitine status. However, estimation of muscle carnitine concentrations require muscle biopsy, which is too invasive to be used for the monitoring for carnitine deficiency [15].

Coulter [13] suggested a number of risk factors for carnitine deficiency, which included age less than 10 years, multiple neurologic disabilities, including mental retardation, cerebral palsy, blindness, and microcephaly, nonambulatory status, underweight, and intake of multiple antiepileptic drugs. We failed to identify any significant associations between carnitine levels and any of these risk factors in our patients (i.e., age, body mass index, additional antiepileptic drugs used, presence of mental retardation, cerebral palsy or feeding problem, or nonambulatory status). Previous studies demonstrated that free carnitine levels were further decreased in patients taking multiple antiepileptic drugs containing valproate. The exact underlying mechanism is unknown and remains speculative. Whereas polytherapy might merely reflect the severity of the underlying neurologic disease, it has been suggested that polytherapy could enhance the production of toxic metabolites of valproate [3], [4], [5]. In our patients, however, although 39% (17/43) of the patients were taking polytherapy, the majority of them did not have low carnitine levels. There were also no associations between carnitine levels and the daily dosage of valproate. These observations were similar to those previously reported by other investigators [7]. Our results might be partly limited by the sample size but might also reflect that carnitine deficiency is not as common in our patients as quoted in previous studies on mainly white patients [1], [2], [3], [4], [5]. Hirose et al. [7] also demonstrated that valproate therapy did not deplete carnitine in epileptic children who were otherwise healthy. Only two of our patients (4.7% of total study patients and 20% of those with three or more risk factors) had carnitine level slightly below the normal range (17.2 and 17.5 μmol/L). Both patients suffered from perinatal hypoxia, resulting in spastic quadriplegia, mental retardation, and epilepsy. Neither of the patients had any symptoms or signs of carnitine deficiency, which were known to be nonspecific and vague. These include muscle weakness, hypotonia, nausea and vomiting, fatigue, poor concentration, apathy, headache, and poor appetite [13], which might be difficult to evaluate objectively. Carnitine supplementation in these two patients was followed by an increase of the carnitine levels to “normal,” but there was no noticeable change in their clinical status. The significance of their carnitine deficiency was therefore uncertain. Carnitine supplementation is potentially beneficial to patients with valproate-induced hepatotoxicity [12]. Its role in patients receiving valproate therapy is unclear. Coulter conducted an unblinded and uncontrolled study of carnitine supplementation in 20 children with epilepsy who had two or more risk factors for carnitine deficiency and symptoms of carnitine deficiency. Carnitine supplementation was followed by significant symptomatic improvement in these children [13]. However, a placebo-controlled study carried by Freeman et al. [16] failed to demonstrate any significant improvement in the well-being scores of children receiving carnitine supplementation when compared with those in the placebo group. In this study, however, patients were epileptic patients taking either valproate or carbamazepine who might not demonstrate laboratory or clinical evidence of carnitine deficiency. The results might have underscored the importance of potential benefit of carnitine supplementation in symptomatic patients. Carnitine is thought to be a safe and well-tolerated drug. The main side effects include nausea, gastritis, diarrhea, and a fishy body odor. The cost of carnitine supplementation is not low, being at least U.S. $2000-4,000 per patient per year [17].

We concluded from our observations that in our patients, routine measurement of plasma carnitine levels in patients taking valproate is not necessary. However, pediatricians should be aware of the potential risk factors of carnitine deficiency, especially for young patients with poor nutrition. Carnitine levels should be estimated in those who develop symptoms of carnitine deficiency or when hepatotoxicity is suspected, especially when multiple risk factors for carnitine deficiency are present. There is no evidence supporting routine carnitine supplementation in asymptomatic patients. However, if patients develop hyperammonemia, anorexia, constipation, nausea, vomiting, weakness, or hypotonia, carnitine treatment may be considered. Use of antiepileptic medications other than valproate should also be considered in these high-risk patients.