What do you think about when you hear the words “drug toxicity”? Do you think about a patient who overdosed or someone who is experiencing a toxic side effect from a medication? A lot of times, the answer might depend on what we remember about a drug when we hear the name. For example, if I said a patient had acetaminophen toxicity you might first think the effects were related to an overdose, but if I said the patient had drug toxicity from cyclophosphamide, you might presume, since it is a chemotherapeutic agent, that it was related to a drug side effect. The point is, patients can experience drug toxicity from some exposures with both therapeutic and supratherapeutic dosing. Unfortunately, sometimes we only think about drug toxicity based on the situations we commonly see. We often forget toxicity can occur with some drugs in multiple situations. Case in point being valproic acid (aka valproate).
When most physicians think about valproic acid toxicity, they primarily think about a situation where the patient's valproate level is elevated (normal 50 to 100 mg/dL). However, valproic acid is one of several drugs (e.g., lithium and digoxin) where life-threatening drug toxicity can occur at therapeutic levels.
Valproic acid is FDA approved for use as an antiepileptic, for the prevention of migraines, and for manic or mixed episodes associated with bipolar disorder. Some additional off label uses include treatment of aggression, impulsivity, agitation, schizophrenia, and alcohol abuse.1–3 Valproic acid is 90% bioavailable via the oral route and at therapeutic doses is typically 90% protein bound. In situations where there is a saturation of protein binding sites (e.g. acute overdose), this binding decreases to closer to 15%. Traditionally, valproic acid's clinical effects have been attributed to sodium channel blockade, potentiating GABA, and effects on glutamate/NMDA inhibition.4,5 The reality is, the mechanism of action is likely much more complex than we currently appreciate, involving multiple channels and even alterations in gene regulation.6–8 Some of these effects contribute to the toxicity we see with typical overdoses: CNS depression, lethargy, coma, and seizures. However, understanding valproic acid metabolism is necessary to fully appreciate how both overdose and therapeutic exposures cause toxicity.
Valproic Acid Metabolism
The complex metabolism of valproic acid leads to much of its associated toxicity. Metabolism occurs primarily in the liver via glucuronidation (50%), β-oxidation (40%), and ⍵-oxidation (10%) (Figure 1).9 One of the most important mediators in the metabolism of valproic acid is carnitine. During the process of mitochondrial β-oxidation, VPA depletes carnitine stores. This happens in a few different ways (Figure 1).
- Valproic acid metabolism leads to the formation of valproylcarnitine (VPA-carnitine) which is renally eliminated, so you have renal losses of carnitine.
- Valproylcarnitine inhibits the ATP dependent carnitine transporter, so carnitine can’t get in the cell.
- Valproyl metabolites trap mitochondrial CoA which leads to decreased ATP production, further depleting carnitine since ATP is needed for the carnitine transporter to work.
- Decreased mitochondrial CoA results in decreased N-acetylglutamate which is a needed cofactor for carbamoyl phosphate synthetase 1 (CPS 1), the primary enzyme responsible for incorporating ammonia into the urea cycle, where 90% of ammonia is converted to urea (see Figure 2).
Cellular depletion of carnitine shifts valproic acid metabolism from β-oxidation to ⍵-oxidation. The metabolite 4-en-valproate is formed by ⍵-oxidation. This metabolite is hepatotoxic and interferes with CPS I, that enzyme involved in the initial step of the urea cycle, and hyperammonemia occurs (see Figures 1 and 2). Cumulatively, the depletion of carnitine, acetyl-CoA, and inhibition of β-oxidation lead to impaired free fatty acid metabolism and can rarely cause microvesicular steatosis similar to Reye syndrome.9–11
In the overdose setting, valproic acid-induced encephalopathy typically occurs from high serum levels without evidence of hyperammonemia; however, valproic acid-induced hyperammonemia independently can lead to encephalopathy, lethargy, and seizures.12,13
Hyperammonemia can occur in the setting of overdose, with therapeutic or subtherapeutic valproic acid levels, and regardless of the presence of liver dysfunction.5 There are many theories as to why excess ammonia causes encephalopathy.14 When ammonia exceeds the capabilities of the normal hepatic routes for metabolism (primarily the urea cycle), ammonia concentrations increase in the bloodstream and ultimately increase in the brain. In the CNS, ammonia is theorized to cause inflammation, oxidative injury, astrocyte swelling, increased NMDA activity, and osmotic stress which can all lead to brain injury and cerebral edema.12,15–17
Following an overdose, hyperammonemia usually occurs due to the depletion of carnitine stores during the body’s attempt to metabolize an excessive amount of valproic acid. But here’s the thing, some patients can develop hyperammonemia with routine use. I've seen this happen in two general scenarios:
Scenario 1: A patient is on chronic valproic acid, he/she may have mildly elevated ammonia levels from time to time and then these levels are noted to be increased from the patient’s baseline.
Scenario 2: A patient develops acute hyperammonemia shortly after the initiation of the drug.
In Scenario 1, the patient has chronic depletion of carnitine stores. This may be treated by taking L-carnitine supplements or eating foods rich in L-carnitine (e.g., meat, poultry, fish, and dairy products), or the patient may need to stop taking valproic acid. Scenario 2 is a little different and a bit more complex. Let’s look at a case to get a better understanding of how these types of patients may present.
A 17-year-old female with a history of epilepsy on carbamazepine, zonisamide, and levetiracetam is hospitalized for breakthrough seizures. She was started on valproic acid (2000 mg daily) 48 hours prior and is now obtunded. Her valproic acid level is 46 mg/L and her ammonia level is 525 mcg/dL. Electrolytes were normal with the exception of a calcium level of 8.2 mg/dL. The patient was intubated, started on L-carnitine, and dialyzed. Her ammonia level improved to 87 mcg/dL. Unfortunately, she continued to have subclinical status and ultimately died due to cerebral herniation.
So, why do some patients get hyperammonemia shortly after the initiation of the drug? (if you don’t want to go down this rabbit hole, skip to the end for treatment recs)
Inborn Errors of Metabolism
Simply put, sometimes, initiation of valproic acid therapy leads to the unmasking of previously undiagnosed inborn errors of metabolism like urea cycle disorders, fatty acid oxidation defects, and mitochondrial disorders. Remember how we discussed there were two ways normal valproic acid metabolism could ultimately interfere with the urea cycle? (Figures 1 & 2) Imagine you already have a deficiency in your urea cycle activity but you have just enough activity to keep things going, then valproic acid is introduced and diminishes the ability to clear ammonia through that cycle even more. Now you have multiple hits to the system and the effects of valproic acid on the urea cycle become more pronounced. Let’s look at ornithine transcarbamylase (OTC) deficiency, the most common urea cycle disorder and one of the more common inborn errors we see associated with the development of hyperammonemia and encephalopathy after starting valproic acid. There are several examples in the literature, in patients of all ages, where initiation of valproic acid therapy leads to the unmasking of a previously quiescent OTC deficiency.18–20 OTC deficiency is x-linked and usually detected early in life in males, with most of the cases describing detection later in life involving female patients (heterozygotes for OTC deficiency). However, there are multiple cases of valproic acid-induced hyperammonemia reported in older male patients with both chronic valproic acid use and recent initiation.21,22 In one case, a 15-year-old male on valproic acid with a history of migraines, ADD, learning and behavioral problems acutely developed hyperammonemia and fatal cerebral edema – OTC deficiency was diagnosed at autopsy.21 The patient in the scenario above also was determined to have a previously undiagnosed OTC deficiency on her post-mortem exam.
There are several hundred OTC gene mutations, which likely is why patients are affected differently. Deficiency can mean there is absent or reduced enzyme activity, with variability in the degree of reduction in enzyme activity. Patients with OTC deficiency can have unexplained seizures, vomiting, headaches, abnormal behavior, and lethargy. These symptoms can be episodic, triggered by unrecognized periods of hyperammonemia. Stressors like infection, surgery or a high protein diet may precipitate these episodes. We don’t want this to go unidentified, and you should consider obtaining blood and urine samples to test patients with acute valproic acid-induced hyperammonemia for inborn errors of metabolism prior to starting treatment, if possible – ultimately treatment shouldn’t be delayed for testing. Diagnosis can be made with a combination of studies (send out labs for most folks): plasma free & total carnitine, plasma amino acids, urine orotic acid, plasma acylcarnitines, urine organic acids, comprehensive metabolic panel, creatinine kinase, and a lactate level. Some heterozygous patients can have normal biochemistry, but patients with OTC deficiency will often have elevated glutamine, elevated alanine, decreased or absent citrulline, elevated urine orotic acid. The allopurinol challenge test is an alternative method for diagnosing OTC deficiency when it isn’t clear from the lab studies. Heterozygous patients undergoing allopurinol testing will have excessive urinary orotidine and orotic acid.23 Genetic testing can be done to rule-in or rule-out a disease, but the gene mutation isn’t always identified with OTC deficiency. [Interesting fact – treatment of OTC deficiency made headlines in 1999 when Jamie Gelsinger, an 18-year-old with mild OTC deficiency died during a gene therapy trial. This was the impetus for significant debate regarding medical ethics and research.24]
As a side note, when it comes to inborn errors of metabolism and valproic acid, you have to worry about more than just hyperammonemia. There are several case reports of valproic acid initiation leading to increased seizures and encephalopathy (without hyperammonemia), resulting in cautions against using valproic acid in patients with known mitochondrial disorders like mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes (MELAS).25–27 In patients with MELAS, they have defective oxidative phosphorylation mechanisms. Valproic acid and its metabolites can affect structures in the mitochondrial membrane28,29, and in rodent models the valproic acid impaired activity of cytochrome oxidase (cytochrome C) in the electron transport chain (Figure 3).30 The effects on cytochrome oxidase C are hypothesized to contribute to the seizure activity noted in these patients.25 Additionally, there are case reports where valproic acid can precipitate hepatic failure in patients with mitochondrial disorders.31 Given this, it makes sense that patients with fragile mitochondria would be at higher risk for decompensation with a standard dose of valproic acid tipping them over the edge.
How Much Ammonia is Too Much?
Here’s the thing, too much ammonia is bad for the human body, especially the brain. There are several studies that demonstrate elevated ammonia levels are associated with increased morbidity and mortality in both children and adults. There is no clear threshold of when ammonia levels cause problems. In fact, there is a wide range of data out there with discussions of increased mortality occurring in patients with levels anywhere from > 280 μg/dL (200 μmol/L) to 700 μg/dL (500 μmol/L).32–34 In one paper, authors found increased 28-day mortality in children for those with hyperammonemia with either liver failure (8-fold increase) or urea cycle disorders (5-fold increase) with levels above 280 μg/dL.35
To treat hyperammonemia, you need to stop the process that is creating it and you need to get rid of the ammonia you have. In the setting of an acute overdose, stop the valproic acid and give the patient L-carnitine. If you are concerned that there is an inborn error of metabolism, after discontinuing valproic acid, talk to a geneticist. They may recommend the initiation of sodium phenylacetate/sodium benzoate (ammonia scavengers) and L-arginine in an attempt to provide alternative methods for nitrogen excretion and override the OTC defect at the initial step of the urea cycle or to treat undifferentiated inborn errors of metabolism (Figure 4).12,25,36 If the ammonia levels remain high for a few hours (> 250 μmol/L) or are > 500 μmol/L, you should strongly consider hemodialysis to remove the ammonia.36
Valproic acid can cause a variety of effects in overdose: ataxia, somnolence, lethargy, hepatotoxicity, hyperammonemia, acidosis and rarely pancytopenia.37,38 However, as we have discussed, the drug itself can cause toxicity without the presence of elevated drug levels. Sometimes this is due to carnitine deficiency caused by valproic acid and other times this is related to the unmasking of a previously unrecognized inborn error of metabolism. Regardless of whether you are caring for a patient with an intentional overdose or isolated hyperammonemia, stopping the drug is the most important initial step in treatment. After the drug is stopped, there are several treatment options depending on the patient’s clinical picture.
General Treatment Recommendations
- Supportive care – this may mean intubation, fluids, etc.
- STOP the drug.
- Consider giving activated charcoal – many preps are extended-release.
- Hemodialysis to remove valproic acid should be considered when valproic acid levels are >= 900 mg/L, the patient is comatose or has significant hyperammonemia or acidosis (pH< 7.1) See the EXTRIP recommendations.
- Give L-carnitine to replete what is being used by valproic acid metabolism.4
- 100 mg/kg (max dose 6 g) as a loading dose over 30 min IV
- Maintenance dose of 15 mg/kg every 3-4 hours after the loading dose (higher dosing may be used in patients with inborn errors of metabolism (IEMs).
- Oral supplementation can be used in patients who can tolerate PO but oral bioavailability of pharmaceutical-grade carnitine is quite poor (5-18%).
- Can cause the patient to have a fishy odor.
- For anyone who develops acute hyperammonemia within days to weeks of starting valproic acid, consider an underlying IEM may exist.
- Consider obtaining plasma free & total carnitine levels, plasma amino acids, urine orotic acid, plasma acylcarnitines, urine organic acids, comprehensive metabolic panel, creatinine kinase, and a lactate level if you are concerned about underlying inborn errors.
- Talk with genetics early – they will have specific nutrition recommendations and many times will stop all protein intake and start patients on 10% dextrose infusions.36
- For the hyperammonemia
- Make sure the sample was run correctly – placed on ice immediately and run STAT.
- Consider hemodialysis to remove ammonia hemodialysis is recommended with levels > 500 μmol/L, when ammonia levels are > 250 μmol/L with encephalopathy or if no rapid drop in ammonia in 3-6 hours.36
- If high enough suspicion of IEM, consider starting IV sodium benzoacetate/sodium phenylacetate (Ammonul®) and L-arginine to help with nitrogen excretion.12
- Sodium benzoacetate/sodium phenylacetate (Ammunol®) dosing (consult pharmacist if you haven’t ordered before). Patient weight matters. There is a difference in dosing with patients weighing < 20 kg getting mL/kg dosing and > 20 kg getting ml/m2 dosing.
- < 20 kg: 2.5 mL/kg (provides 250 mg/kg of sodium phenylacetate and 250 mg/kg of sodium benzoate) IV loading dose through a central line over 90 to 120 minutes. You can then follow with the same dose given over 24 hours. Sometimes there is the resolution of hyperammonemia after just the loading dose the maintenance dose isn’t needed.
- > 20 kg: 55 mL/m2 (provides 5.5 g/m2 of sodium phenylacetate and 5.5 g/m2 of sodium benzoate) IV loading dose, then can follow with maintenance dose, as above.
- There are also dosing calculators available for sodium benzoacetate and sodium phenylacetate and for calculating BSA.
- L-arginine dosing is 200 mg/kg of L-arginine HCl IV daily.12,36,39
- There is no evidence that lactulose works as an ammonia scavenger and the electrolyte shifts that can occur with associated diarrhea may be harmful in children.40,41
A special shout out to Dr. Katherine Dempsey for her genetics knowledge and perspective!Cleveland Clinic’s Lou Ruvo Center for Brain Health by Nick Fewings
- 1.Huband N, Ferriter M, Nathan R, Jones H. Antiepileptics for aggression and associated impulsivity. Cochrane Database Syst Rev. 2010;(2):CD003499. doi:10.1002/14651858.CD003499.pub3
- 2.Reoux J, Saxon A, Malte C, Baer J, Sloan K. Divalproex sodium in alcohol withdrawal: a randomized double-blind placebo-controlled clinical trial. Alcohol Clin Exp Res. 2001;25(9):1324-1329. https://www.ncbi.nlm.nih.gov/pubmed/11584152.
- 3.Horowitz E, Bergman L, Ashkenazy C, Moscona-Hurvitz I, Grinvald-Fogel H, Magnezi R. Off-label use of sodium valproate for schizophrenia. PLoS One. 2014;9(3):e92573. doi:10.1371/journal.pone.0092573
- 4.Doyon S. Antiepileptics. In: Nelson L, Howland M, Lewin N, Smith S, Goldfrank L, Hoffman R, eds. Goldfrank’s Toxicologic Emergencies. 11th ed. New York: McGraw-Hill Educationn; 2020:1.
- 5.Gerstner T, Buesing D, Longin E, et al. Valproic acid induced encephalopathy–19 new cases in Germany from 1994 to 2003–a side effect associated to VPA-therapy not only in young children. Seizure. 2006;15(6):443-448. doi:10.1016/j.seizure.2006.05.007
- 6.Löscher W. Basic pharmacology of valproate: a review after 35 years of clinical use for the treatment of epilepsy. CNS Drugs. 2002;16(10):669-694. doi:10.2165/00023210-200216100-00003
- 7.Hoffmann J, Akerman S, Goadsby P. Efficacy and mechanism of anticonvulsant drugs in migraine. Expert Rev Clin Pharmacol. 2014;7(2):191-201. doi:10.1586/17512433.2014.885835
- 8.Geng Y, Xue J, Su J, Li H. Molecular mechanism of action of valproate acid alone or in combination with chlorpromazine in the epigenetic regulation of schizophrenia. J Biol Regul Homeost Agents. 2018;32(6):1443-1450. https://www.ncbi.nlm.nih.gov/pubmed/30574748.
- 9.Silva M, Aires C, Luis P, et al. Valproic acid metabolism and its effects on mitochondrial fatty acid oxidation: a review. J Inherit Metab Dis. 2008;31(2):205-216. doi:10.1007/s10545-008-0841-x
- 10.Pourahmad J, Eskandari M, Kaghazi A, Shaki F, Shahraki J, Fard J. A new approach on valproic acid induced hepatotoxicity: involvement of lysosomal membrane leakiness and cellular proteolysis. Toxicol In Vitro. 2012;26(4):545-551. doi:10.1016/j.tiv.2012.01.020
- 11.Sugimoto T, Nishida N, Yasuhara A, Ono A, Sakane Y, Matsumura T. Reye-like syndrome associated with valproic acid. Brain Dev. 1983;5(3):334-337. doi:10.1016/s0387-7604(83)80029-1
- 12.Savy N, Brossier D, Brunel-Guitton C, Ducharme-Crevier L, Du P-T, Jouvet P. Acute pediatric hyperammonemia: current diagnosis and management strategies. Hepat Med. 2018;10:105-115. doi:10.2147/HMER.S140711
- 13.Clay A, Hainline B. Hyperammonemia in the ICU. Chest. 2007;132(4):1368-1378. doi:10.1378/chest.06-2940
- 14.Paprocka J, Jamroz E. Hyperammonemia in children: on the crossroad of different disorders. Neurologist. 2012;18(5):261-265. doi:10.1097/NRL.0b013e318266f58a
- 15.Bosoi C, Rose C. Identifying the direct effects of ammonia on the brain. Metab Brain Dis. 2009;24(1):95-102. doi:10.1007/s11011-008-9112-7
- 16.Norenberg M, Rao K, Jayakumar A. Mechanisms of ammonia-induced astrocyte swelling. Metab Brain Dis. 2005;20(4):303-318. doi:10.1007/s11011-005-7911-7
- 17.Rodrigo R, Cauli O, Boix J, ElMlili N, Agusti A, Felipo V. Role of NMDA receptors in acute liver failure and ammonia toxicity: therapeutical implications. Neurochem Int. 2009;55(1-3):113-118. doi:10.1016/j.neuint.2009.01.007
- 18.Bega D, Vaitkevicius H, Boland T, Murray M, Chou S. Fatal hyperammonemic brain injury from valproic Acid exposure. Case Rep Neurol. 2012;4(3):224-230. doi:10.1159/000345226
- 19.Honeycutt D, Callahan K, Rutledge L, Evans B. Heterozygote ornithine transcarbamylase deficiency presenting as symptomatic hyperammonemia during initiation of valproate therapy. Neurology. 1992;42(3 Pt 1):666-668. doi:10.1212/wnl.42.3.666
- 20.Kay J, Hilton-Jones D, Hyman N. Valproate toxicity and ornithine carbamoyltransferase deficiency. Lancet. 1986;2(8518):1283-1284. doi:10.1016/s0140-6736(86)92714-5
- 21.Thakur V, Rupar C, Ramsay D, Singh R, Fraser D. Fatal cerebral edema from late-onset ornithine transcarbamylase deficiency in a juvenile male patient receiving valproic acid. Pediatr Crit Care Med. 2006;7(3):273-276. doi:10.1097/01.PCC.0000216682.56067.23
- 22.Mehta S, Tayabali S, Lachmann R. Valproate-induced hyperammonemia – uncovering an underlying inherited metabolic disorder: a case report. J Med Case Rep. 2018;12(1):134. doi:10.1186/s13256-018-1666-3
- 23.Horwich A, Fenton W. Precarious balance of nitrogen metabolism in women with a urea-cycle defect. N Engl J Med. 1990;322(23):1668-1670. doi:10.1056/NEJM199006073222310
- 24.Yarborough M, Sharp R. Public trust and research a decade later: what have we learned since Jesse Gelsinger’s death? Mol Genet Metab. 2009;97(1):4-5. doi:10.1016/j.ymgme.2009.02.002
- 25.Lam C, Lau C, Williams J, Chan Y, Wong L. Mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) triggered by valproate therapy. Eur J Pediatr. 1997;156(7):562-564. doi:10.1007/s004310050663
- 26.Lin C, Thajeb P. Valproic acid aggravates epilepsy due to MELAS in a patient with an A3243G mutation of mitochondrial DNA. Metab Brain Dis. 2007;22(1):105-109. doi:10.1007/s11011-006-9039-9
- 27.Windpessl M, Müller P. Valproic acid and MELAS: a word of warning. Intern Med. 2013;52(9):1011. doi:10.2169/internalmedicine.52.0053
- 28.Altrup U, Gerlach G, Reith H, Said M, Speckmann E. Effects of valproate in a model nervous system (buccal ganglia of Helix pomatia): I. Antiepileptic actions. Epilepsia. 1992;33(4):743-752. doi:10.1111/j.1528-1157.1992.tb02356.x
- 29.Altrup U, Reith H, Speckmann E. Effects of valproate in a model nervous system (buccal ganglia of Helix pomatia): II. Epileptogenic actions. Epilepsia. 1992;33(4):753-759. doi:10.1111/j.1528-1157.1992.tb02357.x
- 30.Ponchaut S, Van H, Veitch K. Valproate and cytochrome c oxidase deficiency. Eur J Pediatr. 1995;154(1):79. doi:10.1007/bf01972980
- 31.Chabrol B, Mancini J, Chretien D, Rustin P, Munnich A, Pinsard N. Valproate-induced hepatic failure in a case of cytochrome c oxidase deficiency. Eur J Pediatr. 1994;153(2):133-135. doi:10.1007/bf01959226
- 32.Uchino T, Endo F, Matsuda I. Neurodevelopmental outcome of long-term therapy of urea cycle disorders in Japan. J Inherit Metab Dis. 1998;21 Suppl 1:151-159. doi:10.1023/a:1005374027693
- 33.Bachmann C. Outcome and survival of 88 patients with urea cycle disorders: a retrospective evaluation. Eur J Pediatr. 2003;162(6):410-416. doi:10.1007/s00431-003-1188-9
- 34.Enns G, Berry S, Berry G, Rhead W, Brusilow S, Hamosh A. Survival after treatment with phenylacetate and benzoate for urea-cycle disorders. N Engl J Med. 2007;356(22):2282-2292. doi:10.1056/NEJMoa066596
- 35.Ozanne B, Nelson J, Cousineau J, et al. Threshold for toxicity from hyperammonemia in critically ill children. J Hepatol. 2012;56(1):123-128. doi:10.1016/j.jhep.2011.03.021
- 36.Häberle J, Boddaert N, Burlina A, et al. Suggested guidelines for the diagnosis and management of urea cycle disorders. Orphanet J Rare Dis. 2012;7:32. doi:10.1186/1750-1172-7-32
- 37.Andersen G, Ritland S. Life threatening intoxication with sodium valproate. J Toxicol Clin Toxicol. 1995;33(3):279-284. doi:10.3109/15563659509018000
- 38.Spiller H, Krenzelok E, Klein-Schwartz W, et al. Multicenter case series of valproic acid ingestion: serum concentrations and toxicity. J Toxicol Clin Toxicol. 2000;38(7):755-760. doi:10.1081/clt-100102388
- 39.Dempsey K. Personal Communication. In: ; 2020.
- 40.Als-Nielsen B, Gluud L, Gluud C. Non-absorbable disaccharides for hepatic encephalopathy: systematic review of randomised trials. BMJ. 2004;328(7447):1046. doi:10.1136/bmj.38048.506134.EE
- 41.Matoori S, Leroux J. Recent advances in the treatment of hyperammonemia. Adv Drug Deliv Rev. 2015;90:55-68. doi:10.1016/j.addr.2015.04.009