Rationale:
Hemispherectomy is an effective treatment to improve seizure outcomes in pediatric refractory epilepsy. However, the recent Epilepsy Research Benchmarks from the American Epilepsy Society state that
cognitive outcomes for epilepsy treatment deserve further study. Given the plasticity of the child brain, one may expect relatively positive cognitive outcomes following large pediatric cortical resection. However, questions remain as to whether age at disease onset or surgery predicts cognitive outcome, and whether a single left hemisphere (LH) or right hemisphere (RH) developing in isolation can assume cognitive function that typically emerges across the two.
Methods:
Hemispherectomy patients' parents completed surveys via the Pediatric Epilepsy Surgery Alliance's Global Pediatric Epilepsy Surgery Registry (n=142; 66 with LH disease/surgery, 76 in RH; median/median absolute deviation of seizure onset age 0.21/0.31 years; median/MAD of most recent surgery age 3.29/3.55 yr; median/MAD of current age 9.45/6.57 yr). Parents rated patients' post-surgical cognition on a five-level scale. Separate multinomial logistic regression models were fit to predict cognition from hemisphere affected, seizure onset age, and surgery age. Parents also noted patients' psychological diagnoses. Logistic regression models were fit to predict likelihood of each diagnosis with hemisphere affected, seizure onset age, and surgery age as predictors in separate models.
p-values were corrected across models with the same outcome variable.
Results:
Post-surgery, 12, 36, and 35% of patients were reported to have significant, moderate, and mild impairment, respectively, while 17% had average and 2% had above average cognition. Cognition could not be predicted by hemisphere resected (χ2=4.40, p=0.92), seizure onset age (χ2=6.66, p=0.46), or surgery age (χ2=0.92, p=0.92). Moreover, 39% of patients had aphasia, 35% specific learning disability, 30% intellectual impairment, 26% dysgraphia, 25% attention deficit disorder (ADD), 23% speech apraxia, and 22% autism. Noteably, patients with LH surgery were more likely than patients with RH surgery to have aphasia (z=2.51, p=0.04) and ADD (z=2.50, p=0.04), while hemisphere affected was not predictive of other diagnoses (p>0.05). Seizure onset or surgery age could not predict likelihood of diagnosis for any condition (p>0.05). Additionally, no more than 20% of patients were reported to have anxiety/generalized anxiety disorder, dysarthria, or dyscalculia. No more than 10% had obsessive-compulsive disorder, post-traumatic stress disorder, an eating disorder, agnosia, amnesia, depression, echolalia, palilalia, or selective mutism.
Conclusions:
Neither age at seizure onset nor surgery predicts cognitive outcome/comorbidity following hemispherectomy. LH resection confers a greater risk of aphasia and of ADD than does RH resection. In healthy children, language typically emerges in the LH. Here, following LH resection, engagement of the RH for language and executive function is constrained and deficits in these functions become evident.
Funding:
NIGMS T32GM144300, NSF GRFP DGE2140739, NEI R01EY207018, NEI P30EY08098, Research to Prevent Blindness Inc, NY, Eye & Ear Foundation Pittsburgh.