The pathophysiology of Pisa Syndrome (PS), lateral trunk tilt occurring in the advanced stages of Parkinson's disease (PD), remains to be solved. It might be the consequence of biased spatial referentials. We comprehensively investigated the representation of the vertical (visual-VV and postural-PV), together with that of the long body axis (LBA) in PD patients with PS.

Double-blind controlled trial in 18 PD patients showing PS and treated with bilateral deep brain stimulation (DBS) of the subthalamic nuclei (STN): 63.5 [5.5] years, 11 females, disease duration 17.5 [5.2] years. Spatial estimates were tested off medication under four conditions of stimulation (bilateral on, bilateral off, unilateral on/off, unilateral off/on). Motor state (UPDRS), scoliosis and lateral trunk tilt were also assessed. Twenty healthy age- and gender-matched participants were also included.

The median lateral trunk tilt was 7.5° (IQD=5) under the instruction to stand upright. Scoliosis was frequent (67%) reaching 82% in participants with severe PS, of moderate magnitude (16° [12.5]), and side congruent with that of the PS (67%). The correlation between PS and scoliosis (Cobb) magnitude was strong (r=0.71, P=0.008). In contrast no concordance was found between the side of the PS and disease sides (10/18 tilted to the side of the most denervated striatum; 11/18 tilted toward the most affected hemibody). Spatial perceptions were more frequently biased in patients than in controls (P<0.05): VV 7/18; PV 6/18; LBA 6/18; but not more severe (P>0.05). No statistical rule was found between any of biased referential (side or magnitude) and the trunk tilt (side or magnitude). No spatial referential was modulated by DBS.

Spatial biases are frequently in PS but are not the cause of the lateral trunk tilt. In contrast scoliosis was not only strongly associated to severe PS. We found a troubling concordance in sides and magnitude.