Pournajaf Sanaz (IRCCS San Raffaele, Roma)
Morone Giovanni (San Raffaele Sulmona, Sulmona; Università degli Studi dell’Aquila, Aquila)
Straudi Sofia (AUSL Ferrara University Hospital, Ferrara)
Leo Maria Rosaria (Villa Bellombra Hospital, Bologna)
Russo Emanuele (Fondazione Gli Angeli di Padre Pio, San Giovanni Rotondo)
De Martino Alex (IRCCS Santa Lucia, Roma); Tuoant Carrie Louise (IRCCS San Raffaele, Roma)
Santamato Andrea (Ospedali Riunuiti University Hospital, Foggia)
Calabrò Rocco Salvatore (IRCCS Centro Neurolesi Bonino Pulejo, Messina)
Franceschini Marco (IRCCS San Raffaele, Rome)
On behalf of the Italian PowerUPS-REHAB Study Group
The promising effects of Robot-Assisted Therapy (RT) for Upper Limbs (UL) have been demonstrated in the literature [1], and its utilization in clinical practice is steadily increasing. Task-specific, repetitive, and high-intensity exercises are defined as key points to facilitate motor relearning in neurorehabilitation [2], which RT can provide through an “assisted-as-needed” approach [3].
The primary aim of this study was to assess the effectiveness of a robotic exoskeleton system for upper limb rehabilitation compared to conventional rehabilitation without robotic assistance in terms of improvement in motor performance assessed with the Fugl-Meyer Assessment Upper Limb (FM-UL) – motor part. The secondary objective was to evaluate the effectiveness of robotic treatment in terms of central neurophysiological changes relating to the upper limb in a subgroup of patients.
Individuals with subacute stroke (time from acute event ≤3 months); hemiparesis and moderate to severe motor deficit (FM-UL ≤ 44); sufficient cognitive conditions to understand instructions; and stable general clinical conditions were recruited and randomly assigned to the Experimental Group (EG) receiving upper limb rehabilitation using the Armeo®Power exoskeleton (Hocoma, Zurich) for 25 sessions – 5 days/week – 45 minutes each; or the Control Group (CG) receiving conventional rehabilitation with the same dose and frequency as the EG, in addition to regular rehabilitation program. All participants were clinically evaluated at baseline (T0) and at the end of treatment (T1). Clinical assessments were based on the International Classification of Function, Disability, and Health (ICF). A satisfaction questionnaire was administered to EG participants at T1. The brain electrical activity of a subgroup of EG patients (N=5) was evaluated at T0 and T1 using resting-state electroencephalography (EEG) with eyes closed and eyes open (5 minutes each). The primary outcome was the motor part of FM-UL (Score 0-66). The Wilcoxon test (p<0.05) was used to detect significant changes between the two groups.
A total of 84 participants were recruited (42 in EG, 42 in CG), with a mean age (+SD) of 63+13 years, including 32 females (38%) and 37 with right hemiparesis (43%), mainly following subacute stroke (mean+SD: 37+27) of PACI type (40%) based on Bamford classification. Of these, 67 completed treatments, with a dropout rate of 20% (EG=11; CG=6) during the experimental phase for reasons unrelated to treatment. Both groups showed clinical improvements over time in most outcomes as demonstrated in table 1. Only EG achieved significant improvement between T0 and T1 in FM-UL Part A (p=0.037) related to proximal performance, while CG improved in FM-UL wrist part (p=0.016) and hand part (p=0.023). The modified Barthel Index (mBI) showed statistically more significant improvement in favor of EG (p=0.042). A remodulation in alpha1 and beta 2 bands (figure 1), although not significant due to the reduced number of acquisitions, was found indicating a more regular organization of the electroencephalographic rhythm of brain areas responsible for upper limb motor control. Upper limb rehabilitation using the Armeo®Power exoskeleton was well accepted by EG patients and no adverse events were noted during treatment.
This study highlights the positive impact of RT using the exoskeleton for upper limb rehabilitation on motor recovery post-subacute stroke. EG showed significant improvements in proximal motor performances, specifically in FM-UL Part A, while CG showed improvements in wrist and hand parts of FM-UL. The inclusion of mBI further underscores broader functional benefits in favor of EG. Despite preliminary non-significant results in modulating brain electrical activity, the study recognizes the potential for more organized electroencephalographic rhythms related to upper limb motor control. This aspect requires further confirmation through studies conducted on a larger number of individuals with homogeneous etiology and clinical characteristics. Overall, the results support the promising role of robot-assisted therapy in promoting task-specific motor relearning in individuals with subacute stroke undergoing neurorehabilitation.
