Abstract
The effect of subthalamic nucleus deep brain stimulation (STN-DBS) on balance function in patients with Parkinson's disease (PD) and the potential outcome predictive factors remains unclear. We retrospectively included 261 PD patients who underwent STN-DBS and finished the 1-month follow-up (M1) assessment in the explorative set for identifying postoperative balance change predictors, and 111 patients who finished both the M1 and 12-month follow-up (M12) assessment in the validation set for verifying the identified factors. Motor and balance improvement were evaluated through the UPDRS-III and the Berg balance scale (BBS) and pull test (PT), respectively. Candidate predictors of balance improvement included age, disease duration, motor subtypes, baseline severity of PD, cognitive status, motor and balance response to levodopa, and stimulation parameters. In the off-medication condition, STN-DBS significantly improved BBS and PT performance in both the M1 and M12, in both datasets. While in the on-medication condition, no significant balance improvement was observed. Higher preoperative BBS response to levodopa was significantly associated with larger postoperative off-medication, but not on-medication, BBS (p < 0.001) and PT (p < 0.001) improvement in both the M1 and M12. BBS subitems 8, 9, 11, 13, and 14 were the major contributors to the prediction of balance improvement after STN-DBS. STN-DBS improves short-term off-medication, but not on-medication, balance function assessed through BBS and PT. Preoperative BBS response to levodopa best predicts postoperative off-medication balance improvement. For patients who manifested severe balance problems, a levodopa challenge test on BBS or the short version of BBS is recommended.