Diagnostic delay in psychogenic nonepileptic seizures

Methods and patients.

We identified 422 in- and outpatients of the Department of Epileptology at the University of Bonn, Germany, for whom a diagnosis of PNES was made between April 1991 and April 2001. We only included patients for whom a firm diagnosis of PNES was made. The diagnosis was considered firm if spontaneous seizures or attacks provoked by suggestive injection of 0.9% NaCl had been documented at our center and if these attacks were considered typical of habitual events by patients and seizure witnesses. EEG, video-EEG, or the direct observation of a seizure by a physician experienced in the diagnosis of epilepsy and PNES were accepted as methods of documentation. We excluded patients for whom psychogenic seizures were not proven by documentation at our center (n = 85), and patients for whom the time of onset of PNES was unclear (n = 24). An additional diagnosis of epilepsy was based on ictal EEG or video-EEG recordings or the clinical assessment of an experienced epileptologist. Biographic information, details of medical and seizure history, diagnosis, and treatment were retrieved from patient records. EEG (n = 313), video-EEG (n = 257), and MRI reports (n = 169) were obtained if available. PNES semiology was categorized into attacks with positive motor features (including tonic-clonic–like and tonic-like), negative motor features (flaccid weakness, atonic collapse), and purely sensory symptoms. Mean diagnostic delays were calculated for categories of the variables sex, coexisting epilepsy, semiologic type, history of pseudostatus, and presence of EEG or MRI abnormalities. Differences of mean delays were examined for statistical significance using Student’s t-tests for independent samples after Levene testing of variance for equality. Pseudostatus was defined as PNES lasting for more than 30 minutes. The relationship between continuous variables (age and PNES frequency) and diagnostic delay was examined using a Pearson correlation.

Results.

Of 422 patients, 313 fulfilled all inclusion and no exclusion criterion; 212 of these patients only had PNES, and 101 had psychogenic and epileptic seizures. PNES had predominantly positive motor features in 71.9%, negative motor features in 23.0%, and purely sensory features in 5.1% of patients. Patients were referred with a diagnosis of epilepsy or seizures of unclear etiology; 83.3% of all patients and 75.5% of patients with only psychogenic seizures had received anticonvulsants, 20.1% had a history of pseudostatus. The diagnosis of PNES was based on the documentation of typical seizures by video-EEG (n = 255), ictal EEG (n = 99), or observation and examination during a seizure (n = 99). Of the attacks documented on video-EEG, 215 were induced by suggestive intravenous injection of 0.9% NaCl. Additional epileptic attacks were documented on EEG or video-EEG in 66 of 101 patients. Mean age at onset of PNES in the whole patient group was 26.1 years (SD 12.7). The mean age at diagnosis of psychogenic seizures was 33.3 years (SD 13.0), mean delay of the diagnosis of PNES therefore was 7.2 years (SD 9.3). In patients with additional epilepsy, the mean age of manifestation of epileptic seizures was 14.5 years (SD 11.6), mean age at onset of PNES was 27.5 years (SD 12.1), and mean age of diagnosis of PNES was 33.9 years (SD 11.4). Patient characteristics with significant effects on diagnostic delay are presented in the table. Sex, the presence of additional epileptic seizures, semiologic category, a history of pseudostatus, MRI or nonspecific EEG abnormalities did not influence diagnostic delay. The presence of epileptiform discharges in the interictal EEG (focal sharp waves or spikes in 13 patients, generalized spike wave paroxysms in 4 patients) showed a trend toward an effect in the whole patient group (p = 0.073). Age was negatively correlated with long delay (see figure). There was no significant correlation between PNES frequency and diagnostic delay. Mean diagnostic delay after first presentation to our specialist center was 0.4 year (SD 0.97) for the whole patient group and 0.2 year (SD 0.75) for patients with exclusively psychogenic seizures.

• Download figure
• Open in new tab
• Download powerpoint

Figure. There was a negative correlation between age at onset of patients with psychogenic nonepileptic seizures (PNES) and diagnostic delay (r = −0.286, p = 0.01).

Discussion.

PNES were diagnosed with a mean delay of 7.2 years. Because the average wait for an appointment at our epilepsy center is less than 3 months, and because we were able to make a firm diagnosis within 5 months, this seems rather long and warrants further analysis. In juvenile myoclonic epilepsy, diagnostic delays of 5.9 to 14.5 years have in part been attributed to the fact that “subtle” seizures (absence attacks and myoclonic jerks) were missed.3,4⇓ However, the fact that 83.3% of our patients had received anticonvulsants before referral, and that 20.1% had a history of pseudostatus, shows that PNES were not missed but were misclassified and inappropriately treated as epilepsy. It is not difficult to see why interictal epileptiform EEG changes may have caused further delays in the diagnosis of PNES, although such EEG changes have been described in up to 2.4% of healthy individuals.5 That other patient characteristics reported to be associated with PNES (female sex, younger age) did not shorten the time to diagnosis and that patient factors confounding the differentiation from epilepsy (additional epileptic seizures, MRI abnormalities) did not increase delay, illustrate that despite considerable published knowledge about PNES, this diagnosis continues to pose a particular diagnostic challenge to nonexpert physicians. Physician factors also may explain why a history of pseudostatus or a high frequency of seizures, which should allow direct seizure observation and ictal examination, were not associated with significantly shorter delays.

Our study has a possible limitation: we included 99 patients for whom the diagnosis was based on seizure observation and ictal examination. At times it can be difficult to make a firm diagnosis with direct observation alone, and maybe our diagnosis was wrong in some cases. However, it is quite characteristic of PNES to occur in the presence of physicians in the consulting room, and the seizure semiology can be so unlike epilepsy that the diagnosis is clear and further investigations are unnecessary.6 Our results could have been biased if we had excluded such patients. In fact, the mean diagnostic delay in patients for whom our diagnosis was based on seizure observation was 7.6 years, and our results would have been much the same if we had not included these patients.

It is important for physicians involved in the diagnosis and management of seizure disorders to realize the high prevalence and to appreciate the importance of making a positive diagnosis of PNES. Once the diagnosis has been made and communicated, the patient can be offered specific treatment for any underlying psychiatric disorder,7 and should be at less risk for inappropriate and potentially harmful medical intervention.8 Patients with PNES can respond to psychotherapeutic treatment,9 which may be more successful if it is started early.10 To shorten the time from manifestation to diagnosis of psychogenic seizures we suggest that: 1) patients with atypical seizures are investigated at an early point using home video, ambulatory EEG recording, video-EEG, or suggestive seizure provocation under video-EEG monitoring; 2) therapeutic trials of anticonvulsants are avoided; 3) the diagnosis of epilepsy is critically reviewed in patients who fail to respond to anticonvulsant drugs or who develop apparent status epilepticus; and 4) a previous diagnostic label of epilepsy is not accepted without question.