Ambiguous results of an attempt to withdraw barbiturates in epilepsy patients with intellectual disability

Phenobarbital and primidone frequently have adverse effects on mental functions. Therefore, an attempt was made to taper barbiturates in 85 patients out of a resident population with epilepsy and intellectual disability who were selected according to clinical criteria. The objectives were to reduce the use of barbiturates, to improve the patients' cognitive and psychological state, and to reduce polypharmacy while avoiding seizure exacerbation. Four months after complete withdrawal changes in seizure frequency were assessed as well as changes in cognitive abilities, psychological state and behaviour (using the clinical global impression scale). In 13 patients the tapering failed due to complications (seizure increase in 11 patients). In 72 patients the barbiturate was completely withdrawn (mean duration of tapering: 393 days). Cognitive improvement was achieved in 17 patients (23.6%), 5 patients (6.9%) deteriorated. Seizure frequency remained unchanged in 33 patients (45.8%), in another 15 patients (20.8%) the seizure frequency decreased. Reduction in polypharmacy was obtained in 61 patients (84.7%). In an overall judgement (clinical global impression scale) of cognitive abilities AND seizure control, 25 patients (34.7%) were improved. 31 patients (43.1%) remained unchanged while 12 patients deteriorated (4 patients: impossible to judge). For statistical analysis three outcome groups were defined: the improved group (N=25), the unchanged group (N=31), and the deteriorated/failed group (N=25) consisting of the 12 deteriorated patients plus the 13 patients in whom tapering failed. Stepwise logistic regression revealed a history of an attempt to withdraw phenobarbital/primidone (p=0.017; OR 3.8), age (p=0.012) and seizure frequency (marginally significant: p=0.097) as outcome predictors. Older age was associated with better outcome. A high seizure frequency before tapering was related to good outcome, while seizure freedom and a history of failed withdrawal were associated with deterioration/failure. Outcome did not depend on duration of barbiturate therapy, dosage or serum concentration, co-medication, reduction rate, degree of intellectual disability, or epilepsy syndrome. In summary, the number of barbiturate medications has been considerably reduced, but the principal aim of the project, to relieve patients from assumed barbiturate side effects, has been achieved only in one out of four patients.