SUMMARYPurpose: To evaluate the relative contribution of demographic and epilepsy-related variables, depressive symptoms, and adverse effects (AEs) of antiepileptic drugs (AEDs) to health-related quality of life (HRQOL) in adults with pharmacoresistant epilepsy. Methods: Individuals with epilepsy whose seizures failed to respond to at least one AED were enrolled consecutively at 11 tertiary referral centers. HRQOL was assessed by the Quality of Life in Epilepsy Inventory-31 (QOLIE-31), AEs by the Adverse Event Profile (AEP), and depressive symptoms by the Beck Depression Inventory-II (BDI-II). Multivariate linear regression models were used to identify variables associated with QOLIE-31 total score and subscale scores. Key Findings: Of 933 enrolled individuals aged 16 years or older, 809 (87%) were able to complete the selfassessment instruments and were included in the analysis. Overall, 61% of the variance in QOLIE-31 scores was explained by the final model. The strongest predictors of HRQOL were AEP total scores (b = )0.451, p < 0.001) and BDI-II scores (b = )0.398, p < 0.001). These factors were also the strongest predictors of scores in each of the seven QOLIE-31 subscales. Other predictors of HRQOL were age (b = )0.060, p = 0.008), lack of a driving license (b = )0.053, p = 0.018), pharmacoresistance grade, with higher HRQOL in individuals who had failed only one AED (b = 0.066, p = 0.004), and location of the enrolling center. Epilepsy-related variables (seizure frequency, occurrence of tonic-clonic seizures, age of epilepsy onset, disease duration) and number of AEDs had no significant predictive value on HRQOL. The AEP total score was the strongest negative predictor of HRQOL in the subgroup of 362 patients without depressive symptoms (BDI-II score <10), but even in this subgroup the BDI-II score was retained as a significant predictor. Significance: In individuals with pharmacoresistant epilepsy, AEs of medication and depressive symptoms are far more important determinants of HRQOL than seizures themselves. When seizure freedom cannot be achieved, addressing depressive comorbidity and reducing the burden of AED toxicity is likely to be far more beneficial than interventions aimed at reducing the frequency of seizures. KEY WORDS: Antiepileptic drugs, Pharmacoresistance, Depression, Adverse event profile.The primary objective of the management of epilepsy is to restore a normal health-related quality of life (HRQOL). To achieve this goal, complete seizure control without adverse effects (AEs) is the primary prerequisite, but other factors such as comorbidities and psychosocial constraints also need to be addressed (Perucca & Tomson, 2011;Taylor et al., 2011 To date, most studies of HRQOL in epilepsy have been conducted in relatively small or heterogenous populations of patients with controlled and uncontrolled seizures, and have focused on specific factors, most notably the impact of seizures and AEs of treatment (Baker et al., 1997;Gilliam et al., 2004a;Cramer et al., 2007) or seizures and comorbidities...

SUMMARYPurpose: To evaluate the adverse effects (AEs) of antiepileptic drugs (AEDs) in adults with refractory epilepsy and their relationship with number of coprescribed AEDs and AED load. Methods: Patients with refractory epilepsy were enrolled consecutively at 11 tertiary referral centers. AEs were assessed through unstructured interview and the Adverse Event Profile (AEP) questionnaire. AED loads were calculated as the sum of prescribed daily dose (PDD)/defined daily dose (DDD) ratios for each coprescribed AED. Results: Of 809 patients enrolled, 709 had localizationrelated epilepsy and 627 were on polytherapy. AED loads increased with increasing number of AEDs in the treatment regimen, from 1.2 ± 0.5 for patients on monotherapy to 2.5 ± 1, 3.7 ± 1.1, and 4.7 ± 1.1 for those on two, three, and ‡4 AEDs, respectively. The number of spontaneously reported AEs correlated with the number of AEs identified by the AEP (r = 0.27, p < 0.0001). AEP scores did not differ between patients with monotherapy and patients with polytherapy (42.8 ± 11.7 vs. 42.6 ± 11.2), and there was no correlation between AEP scores and AED load (r = )0.05, p = 0.16). Conclusions: AEs did not differ between monotherapy and polytherapy patients, and did not correlate with AED load, possibly as a result of physicians' intervention in individualizing treatment regimens. Taking into account the limitations of a cross-sectional survey, these findings are consistent with the hypothesis that AEs are determined more by individual susceptibility, type of AEDs used, and physicians' skills, than number of coprescribed AEDs and AED load.

LEV monotherapy was not superior to standard AEDs for the global outcome, namely time to treatment withdrawal, in patients with newly diagnosed focal or generalised seizures.

No abstract

No abstract

OBJECTIVEAlthough rates of postoperative morbidity and mortality have become relatively low in patients undergoing transnasal transsphenoidal surgery (TSS) for pituitary adenoma, cerebrospinal fluid (CSF) fistulas remain a major driver of postoperative morbidity. Persistent CSF fistulas harbor the potential for headache and meningitis. The aim of this study was to investigate whether neural network–based models can reliably identify patients at high risk for intraoperative CSF leakage.METHODSFrom a prospective registry, patients who underwent endoscopic TSS for pituitary adenoma were identified. Risk factors for intraoperative CSF leaks were identified using conventional statistical methods. Subsequently, the authors built a prediction model for intraoperative CSF leaks based on deep learning.RESULTSIntraoperative CSF leaks occurred in 45 (29%) of 154 patients. No risk factors for CSF leaks were identified using conventional statistical methods. The deep neural network–based prediction model classified 88% of patients in the test set correctly, with an area under the curve of 0.84. Sensitivity (83%) and specificity (89%) were high. The positive predictive value was 71%, negative predictive value was 94%, and F1 score was 0.77. High suprasellar Hardy grade, prior surgery, and older age contributed most to the predictions.CONCLUSIONSThe authors trained and internally validated a robust deep neural network–based prediction model that identifies patients at high risk for intraoperative CSF. Machine learning algorithms may predict outcomes and adverse events that were previously nearly unpredictable, thus enabling safer and improved patient care and better patient counseling.

Clinical, neuropsychological and neuropsychophysiological data (Q-EEG, ERPs and CNV/RT activity) were obtained from 24 patients who had more or less severe presenile primary cognitive decline without depression, and compared with similar data from 10 age-matched healthy volunteers (mean age, 59.4 years). All of the patients (15 M and 9 F; mean age 59.6 years) were selected according to the DSM III-R, ICD-10 and NINCDS-ADRDA criteria and underwent CT and MRI scanning, in addition to a standard clinical examination, a battery of psychometric tests, spectral EEG, and bit-mapped CNV complex and RT to S2 analyses. Twelve of the 24 patients presented an initial presenile idiopathic cognitive decline (PICD) but did not wholly fulfil the clinical and neuropsychological criteria for primary dementia or for a diagnosis of probable AD; the remaining 12 patients showed characteristic clinical signs and symptoms of a very probable early stage of presenile Alzheimer-type dementia (PAD). ANOVA, correlational and discriminant analyses of the neuropsychological test scores, and the neurophysiological and RT to S2 data revealed 22 highest-ranked between-group discriminant factors (all with a significance level of p < 0.01). The conclusive discriminant analysis retained 13 of these factors as final canonical functions, and these showed a 97% grouping accuracy (33 of the 34 subjects examined); the same percentage of correct classifications was also achieved using only the 15 best indicators in the group of CNV/RT findings. Using both of these sets of highest-ranked discriminators, all of the normal subjects and all of the PAD patients were correctly classified; only 1 PICD patient was misclassified as normal when the first group of 13 factors was used, and another PICD patient was misclassified as PAD using the second group of 15 factors. Our findings suggest that, providing they are correctly performed and interpreted, these non-invasive techniques may be an important tool for identifying incipient stages of presenile Alzheimer-type dementia.

OBJECTIVEGross-total resection (GTR) is often the primary surgical goal in transsphenoidal surgery for pituitary adenoma. Existing classifications are effective at predicting GTR but are often hampered by limited discriminatory ability in moderate cases and by poor interrater agreement. Deep learning, a subset of machine learning, has recently established itself as highly effective in forecasting medical outcomes. In this pilot study, the authors aimed to evaluate the utility of using deep learning to predict GTR after transsphenoidal surgery for pituitary adenoma.METHODSData from a prospective registry were used. The authors trained a deep neural network to predict GTR from 16 preoperatively available radiological and procedural variables. Class imbalance adjustment, cross-validation, and random dropout were applied to prevent overfitting and ensure robustness of the predictive model. The authors subsequently compared the deep learning model to a conventional logistic regression model and to the Knosp classification as a gold standard.RESULTSOverall, 140 patients who underwent endoscopic transsphenoidal surgery were included. GTR was achieved in 95 patients (68%), with a mean extent of resection of 96.8% ± 10.6%. Intraoperative high-field MRI was used in 116 (83%) procedures. The deep learning model achieved excellent area under the curve (AUC; 0.96), accuracy (91%), sensitivity (94%), and specificity (89%). This represents an improvement in comparison with the Knosp classification (AUC: 0.87, accuracy: 81%, sensitivity: 92%, specificity: 70%) and a statistically significant improvement in comparison with logistic regression (AUC: 0.86, accuracy: 82%, sensitivity: 81%, specificity: 83%) (all p < 0.001).CONCLUSIONSIn this pilot study, the authors demonstrated the utility of applying deep learning to preoperatively predict the likelihood of GTR with excellent performance. Further training and validation in a prospective multicentric cohort will enable the development of an easy-to-use interface for use in clinical practice.