A Reply to “Carnitine” and “Role of Carnitine in Valproic Acid Toxicity”

To the Editor:

I want to thank Dr. Perez and Drs. Caraccio and Mofenson both for their insightful comments and for highlighting the continuing controversy regarding the use of L‐carnitine therapy for valproic acid‐associated hyperammonemia.

As Dr. Perez notes, the retrospective cohort study by Bohan et al. does indeed suggest a startling survival benefit from the early administration of L‐carnitine to individuals suffering from valproic acid (VPA)‐induced severe, symptomatic hepatotoxicity [1]. Fortunately, fulminant hepatotoxicity is rare with acute VPA intoxication. However, isolated hyperammonemia is a common finding in both therapeutic use and in overdose. I must clarify that my comments regarding lack of demonstrated outcome benefit from L‐carnitine therapy referred to patients with valproic acid‐associated hyperammonemia or valproate‐related hyperammonemic encephalopathy (VHE) in the setting of normal liver function.

In addition, because L‐carnitine therapy is hypothesized to act by reversing a metabolic blockade caused in part by acquired carnitine deficiency, I remain hesitant to treat individuals empirically based solely upon serum VPA concentration, without evidence of concomitant hyperammonemia or encephalopathy. In the abstract by Beauchamp and Olson, the purpose of the study was outcome risk stratification based on serum VPA levels [2]. Although patients with serum VPA concentrations greater than 450 µg/mL suffered more significant toxicity, no data regarding serum ammonia or presence of VHE were provided. As such, no further conclusions regarding the utility of L‐carnitine treatment based on serum VPA level can be made based on these data.

I want to thank Drs. Caraccio and Mofenson for readdressing the issue of L‐carnitine safety. On July 23, 1998, a “Dear Health Care Professional” letter was forwarded by the Food and Drug Administration (FDA) [3]. This letter stated, “Seizures have been reported to occur in patients with or without pre‐existing seizure activity receiving either oral or intravenous levocarnitine. In patients with pre‐existing seizure activity, an increase in seizure frequency and/or severity has been reported.” What the letter failed to address was the patient population in which this was noted, the circumstances under which these patients received L‐carnitine (i.e., acute therapy or chronic supplementation), or the frequency of such adverse events.

Although the FDA letter was circulated in 1998, no reference to it appears in the Physicians' Desk Reference (PDR) [4]. Moreover, seizures are not listed in the PDR table of adverse events, and no warnings or contraindications to use of the drug are listed. L‐carnitine is well tolerated historically, and seizures related to L‐carnitine therapy in acute VPA overdose have not been reported to date. Forty‐two patients in the paper by Bohan et al. received L‐carnitine, and no comment of seizure activity was made.

There is biological plausibility to support the use of L‐carnitine treatment in cases of VHE. Despite this, several key issues remain undetermined. First and foremost is whether L‐carnitine therapy actually makes a difference in clinical outcome. No study published to date has demonstrated a statistically significant change in ammonia kinetics in L‐carnitine‐treated patients compared with controls. It must also be noted that demonstrating an improvement in a serum biomarker may not correlate with outcome. Serum ammonia concentrations are notorious for lack of correlation with symptoms of hepatic encephalopathy. Would a more rapid decline in serum ammonia concentration result in less time on the ventilator, shortened intensive care unit length of stay, or improved outcome? Drs. Caraccio and Mofenson note that their group has successfully managed 38 patients without L‐carnitine supplementation and all have recovered. Vossler et al. reported a case series of seven patients with VHE, in whom only one received supplemental L‐carnitine [5]. They reported that serum ammonia levels declined more rapidly than resolution of VHE, and that glutamine levels or electroencephalogram appeared to be better correlated with resolution of VHE than ammonia level. In case reports by Murakami et al. and Ishikura et al., the patients remained comatose for 3 to 4 days despite early initiation of L‐carnitine treatment [6][7].

If L‐carnitine truly is beneficial in acute VPA intoxication, who might benefit from therapy? As previously noted, asymptomatic hyperammonemia is common with VPA therapy [8]. Does the presence of hyperammonemia in the setting of acute VPA overdose merit the use of L‐carnitine, or should therapy be reserved for patients demonstrating VHE?

If L‐carnitine therapy is indicated, what dose and route of administration should be used? Ishikura et al. used 100 mg/kg as a first dose, followed by 250 mg every 8 hr, a dose our poison center has recommended in the past [7]. Using this dosing regimen certainly exceeds the 3 g maximum daily dose, although adverse effects have not been noted to date. Closer review of the case report would suggest that the appropriate dose is actually 100 mg/kg, followed by 25 mg/kg every 8 hr. However, no data exist to support the use of this dosing strategy over other dosing regimens. With respect to route of therapy, the results of Bohan et al. make a strong argument for the use of intravenous rather than oral L‐carnitine supplementation.

It is clear that much more work is needed to resolve issues regarding the use of L‐carnitine therapy in acute VPA intoxication. Until these issues are resolved, it must be acknowledged that arguments both for and against the use of the drug remain educated opinion rather than driven by evidence‐based medicine.