Oxcarbazepine adjunctive therapy in infants and young children with partial seizures

Drs. Kruszewski and Klotz recognize the difficulties in conducting a clinical trial in pediatric patients, particularly one which includes a seriously ill population at risk for increased seizure activity and potential complications. [1] This trial enrolled very young patients (1 month to <_4 years="years" old="old" with="with" inadequately="inadequately" controlled="controlled" partial="partial" seizures="seizures" already="already" receiving="receiving" up="up" to="to" two="two" concomitant="concomitant" antiepileptic="antiepileptic" drugs.="drugs." we="we" used="used" a="a" unique="unique" and="and" creative="creative" study="study" design="design" comparing="comparing" low="low" high="high" doses="doses" of="of" oxcarbazepine="oxcarbazepine" not="not" placebo="placebo" in="in" an="an" attempt="attempt" endanger="endanger" these="these" patients="patients" severe="severe" epileptic="epileptic" syndromes.="syndromes." p="p"/>The adverse events as presented in Table 4 are not specific in that the events are presented by primary organ class and not by individual preferred terms aside from somnolence and pyrexia which were the only two adverse events to occur in >10% of the patients in either dose group. The higher incidence of adverse events reported for the high-dose versus the low-dose group is likely a reflection of the 3.5-fold longer duration of exposure to study drug of the high-dose group (35 days) compared to the low-dose group (9 days). For the organ classes, infections and infestations and respiratory, thoracic, and mediastinal disorders, none of the individual adverse events occurred in more than 10% of the patients in either group.

As acknowledged in our paper, the study was limited by the short duration and absence of a placebo group. Thus, a low- and high-dose group design was used in an attempt to allow meaningful comparison without placing the children at undue risk. In order to accomplish this, the high-dose group required a 26-day titration period to reach the planned dose, which the low-dose group did not need. To assess efficacy, both the baseline EEG and the EEG obtained over the last 72 hours of treatment were required. In a classic ITT population analysis, any patient missing one value would be excluded as well. The modified ITT statistical analysis employed was equivalent to a classic ITT analysis minus the patients with missing EEG values.

The efficacy results as presented reflect the planned primary assessment with type 1 seizures to demonstrate that the high-dose group is more effective than the low-dose group (p=0.043). Type 1 and type 2 seizures were combined for only one planned secondary assessment and these data are presented separately as appropriate.

We acknowledge the stated concern that seven of the eight authors have disclosed a relationship to the trial sponsor, either as an employee or a recipient of honoraria or grant funding. The disclosure statements on the front page of Neurology articles are specifically intended to report relevant information about investigators and authors to allow objectivity on the readers' behalf and enable the scientific community to draw their own conclusions about any potential conflict of interest. We feel that the information about the authors was adequately disclosed.

Disclosure: J.E.P.-G. serves as a consultant and speaker for Novartis Pharmaceuticals Inc. D.N. serves as a consultant for and has received honoraria in excess of $10,000 from Novartis Pharmaceuticals Inc. D.B., S.S., T.S., and D.T. are employees of Novartis Pharmaceuticals Inc. and D.T. and D.B. are stockholders in Novartis Pharmaceuticals Inc. Y.S. is an employee and stockholder in excess of $10,000 in Novartis Pharma AG. R.E. has no financial interests to disclose. This study was sponsored by Novartis Pharmaceuticals Inc.

We appreciate that authors and researchers are challenged by significant methodological difficulties in the implementation of clinical trials involving a pediatric population. However, we do not understand misleading or problematic interpretation and presentation of the clinical trial results. In the article by Pina-Garza et al, we believe the authors failed to straightforwardly report their findings in a number of significant ways. [1]

An initial concern is the objectivity of a group where only one of eight authors is not significantly financially linked to Novartis.

In the abstract section, the authors state that the most frequent adverse events (less than or equal to 10%) were somnolence and pyrexia. However, Table 4 demonstrates that the single most common adverse event in either treatment group was infections and infestations. Furthermore, the high dose oxcarbazepine group showed that 39.1% of the population suffered from infections and infestations, a 2.8-fold increase over the comparison low dose group. Additionally, the most alarming percentage adverse event suffered by 15.6% of the high dose group, one that represented a 9.8-fold increase over the comparison low dose group, was related to respiratory, thoracic and mediastinal disorders. Finally, the high-dose group was the only group to suffer clinically significant EKG abnormalities.

The authors statistical analysis was compromised by a lack of comparable study groups. The high-dose group received treatment for a duration approximately four times longer than the low-dose group (nine days versus 35 days). Furthermore, the authors appeared to reduce statistical validity by altering their randomization during the course of the study using a modified intent-to-treat (ITT) population for subsequent statistical analysis of baseline and treatment phase and, alternatively, using unmodified ITT for statistical evaluation of efficacy. This introduces bias and thereby reduces the power of the conclusions regarding efficacy.

Finally, the stated primary objective was to evaluate the efficacy and safety of oxcarbazepine in children ages 1 month to less than 4 years with Type 1 seizures. However, the efficacy results additionally included type 2 seizure events, inflating the efficacy significance from p=0 .043 to p=0.020. Separately, the authors fail to adequately detail the safety issues of this drug in their abstracted conclusions. The final paragraph of the paper merely summarizes the safety and side-effects suggesting that oxcarbazepine was generally well tolerated, a conclusion that is not consistent with their adverse-event data.

References 1. Piña-Garza JE, Espinoza R, Nordli D, et al. Oxcarbazepine adjunctive therapy in infants and young children with partial seizures Neurology 2005; 65: 1370 – 1375.

Disclosure: The authors report no conflicts of interest.