Abstracts

Rationale:
Valproic acid (VPA) and carbamazepine (CBZ) are frequently used for treating seizures in children. Despite their widespread use, recommendations on required monitoring still remain unclear for this population.
Method:
We selected 200 patients taking VPA or CBZ (100 per drug) using laboratory results for therapeutic drug monitoring between January 1st, 2011 and June 1st, 2017. The subjects were outpatients under 18 years-old.  They were included in reverse chronological order. We retrospectively reviewed their medical files. Incidence and severity of hematologic and biochemical toxicities were assessed.
Results:
Children taking VPA (56 boys, 44 girls, 1.4-17.9 years-old) had been on the drug for 8 ± 34 months. Children taking CBZ (66 boys, 34 girls, 3 months-17.8 years-old) had been on the drug for 21 ± 48.5 months. Incidence of neutropenia (neutrophils below 1.5 x 109/L) was 9% with VPA and 10% with CBZ (none below 0.5 x 109/L, one child stopped VPA following results). Incidence of thrombocytopenia (platelets below 150 x 109/L) was 8% with VPA and 2% with CBZ (one in VPA group below 50 x 109/L, normalized at follow-up 2 weeks later). Incidence of anemia (hemoglobin below 110 g/L) was 3% with VPA and 4% with CBZ. Incidence of increased levels of liver enzymes (above 3 times upper limit of normal (ULN)) was 4% with VPA and 0% with CBZ (26% between 1 and 3 times ULN with VPA, 19% with CBZ). There was no hyponatremia (sodium below 135 mmol/L) in children taking CBZ (Table 1). Hyperammonemia (ammonia above 50 mcmol/L) was seen in 26% of children taking VPA (4% were above 70 mcmol/L, one stopped VPA, two required an increased dose of levocarnitine). Five patients in the VPA group were on levocarnitine. Three of these patients had hyperammonemia. Overall, neutropenia, anemia and hyperammonemia, each independently, was seen more frequently with VPA in our study than stated in the Canadian monograph. Anemia with CBZ was also more frequent (Table 2). As we conducted a retrospective study, we were not able to verify if hyperammonemia was symptomatic. Due to the elevated number of false-positive results, we may have overestimated the incidence. Neutropenia, thrombocytopenia or hyperammonemia were more likely to be found in children with a history of these abnormalities. Non-white children were overly represented in patients with hyperammonemia as 5/7 vs 21/93 patients (p=0.02) displayed this toxicity. Incidence of hyperammonemia was also higher in children taking topiramate (6/7 vs 20/93 p=0.001).
Conclusion:
Data from the Canadian monograph seems to underestimate incidence of toxicity associated with VPA and CBZ, especially in pediatric populations. Guidelines on follow up need to be established. They should include hematologic follow-up for VPA and CBZ. Liver enzyme monitoring should be added for patients on VPA and ammonia level surveillance required when VPA is combined to topiramate. Our small sample seems to show an association between hyperammonemia and non-white patients; further prospective studies are needed to assess this.
Funding:
:No funding was received in support of this abstract.