Localization of Basal Ganglia and Thalamic Damage in Dyskinetic Cerebral Palsy

Background Dyskinetic cerebral palsy affects 15%-20% of patients with cerebral palsy. Basal ganglia injury is associated with dyskinetic cerebral palsy, but the patterns of injury within the basal ganglia predisposing to dyskinetic cerebral palsy are unknown, making treatment difficult. For example, deep brain stimulation of the globus pallidus interna improves dystonia in only 40% of patients with dyskinetic cerebral palsy. Basal ganglia injury heterogeneity may explain this variability. Methods To investigate this, we conducted a qualitative systematic review of basal ganglia and thalamic damage in dyskinetic cerebral palsy. Reviews and articles primarily addressing genetic or toxic causes of cerebral palsy were excluded yielding 22 studies (304 subjects). Results Thirteen studies specified the involved basal ganglia nuclei (subthalamic nucleus, caudate, putamen, globus pallidus, or lentiform nuclei, comprised by the putamen and globus pallidus). Studies investigating the lentiform nuclei (without distinguishing between the putamen and globus pallidus) showed that all subjects (19 of 19) had lentiform nuclei damage. Studies simultaneously but independently investigating the putamen and globus pallidus also showed that all subjects (35 of 35) had lentiform nuclei damage (i.e., putamen or globus pallidus damage); this was followed in frequency by damage to the putamen alone (70 of 101, 69%), the subthalamic nucleus (17 of 25, 68%), the thalamus (88 of 142, 62%), the globus pallidus (7/35, 20%), and the caudate (6 of 47, 13%). Globus pallidus damage was almost always coincident with putaminal damage. Conclusions Noting consistent involvement of the lentiform nuclei in dyskinetic cerebral palsy, these results could suggest two groups of patients with dyskinetic cerebral palsy: those with putamen-predominant damage and those with panlenticular damage involving both the putamen and the globus pallidus. Differentiating between these groups could help predict response to therapies such as deep brain stimulation.