To the Editors:
Although randomized clinical trials provide useful information about efficacy and tolerability of new antiepileptic drugs (AEDs), the external validity or generalizability of their results can be limited as they are usually conducted in protocol-restricted patient populations using fixed study designs and dosing schedules for short treatment periods (Walker & Sander, 1997; Ben-Menachem, 2005; Gilliam, 2005).
We aimed to assess how many patients with epilepsy in clinical practice would qualify for a standard efficacy AED trial after application of common exclusion criteria. We examined all efficacy studies with AEDs published in the period from 2002 to May 2007 and found the following common exclusion criteria: (1) pregnancy or breast-feeding; (2) insufficient seizure frequency (<1 seizure per month); (3) clinically relevant systemic illness (e.g., heart, renal, or hepatic disease); (4) concomitant use of three or more AEDs; (5) progressive neurological or cerebral disease, neoplasia, or structural lesion; (6) presence of a psychiatric disorder; and (7) alcohol or drug abuse.
A total of 432 adult patients with epilepsy attending two neurology clinics in the Czech Republic represented the clinical practice cohort. Information was retrieved from medical records. Three-quarters (76%) of patients had a too low seizure frequency (less than one seizure per month) to be eligible for inclusion in a clinical trial. Half of the study population (51%) was evaluated as having clinically relevant comorbidity for exclusion from a clinical trial; 21% of patients had a neurological comorbidity, 19% had a psychiatric disorder, and 23% suffered from relevant systemic illnesses (many patients had a combination of comorbidities). Twenty-three percent of the patients used a combination of three or more AEDs. When the exclusion criteria were subsequently applied to the whole group of patients, only 9% of patients would qualify for a standard AED trial. When we applied the criteria to the subgroup of patients with one or more seizures per month (n = 106), still only 36% of these patients would be eligible for entering a clinical efficacy study.
As the prevalence of many common psychiatric and somatic conditions is high in adults with epilepsy (Gaitatzis et al., 2004; Strine et al., 2005; Tellez-Zenteno et al., 2005), exclusion of patients with comorbidity limits the possibility to extrapolate the results of clinical trials to real-world clinical practice. A history or presence of comorbidity alone would already render half of the patients in our total cohort not eligible for inclusion in a trial. Furthermore, this criterion would exclude 65% of patients with recurrent partial seizures.
In conclusion, our findings show that the efficacy trials in epilepsy tend to evaluate only a small subset of patients with a specific clinical profile that is not representative for day-to-day clinical practice. Evidence from clinical trials provides only partial information on the effectiveness of AEDs in real clinical practice and therefore observational research within the setting of daily clinical practice is necessary to complement the results of clinical trials.