About 70% of hemiplegic patients suffer from hemiplegic shoulder pain after stroke. This common occurrence is a cause for concern in the rehabilitation setting as it leads to impairment of functional outcomes induced by discomfort and delays in rehabilitation, important psycho-emotional repercussions as there is a correlation between upper arm pain and depression, a longer hospital stay. Poor management of hemiplegic shoulder pain can ultimately give rise to type 1 "complex regional pain syndrome" (CRPS I). There is no consensus on treatment, care pathways or useful devices for positioning acute stroke patients in the literature. Our objective is to compare the effectiveness of a new positioning procedure of the hemiplegic arm with conventional positioning (pillow and "shoulder-immobilisation" sling) in acute stroke patients.

Patient's actions are focus in 4 points:

• Positioning in bed, wheelchair and standing with specific materials (holds, sling in external rotation and abduction)
• Be careful of shoulder's coaptation in care and functional activities (toilet, dressing, transfers)
• A training of the patient and members of family on pathology, risk, use of material
• Daily passive mobilisation of the upper member by a therapist.

Shoulder pain is one of the four most frequently encountered after stroke complications. The importance of prevention isn't more to justify but it is not consensus on common conduct to follow.

Propose: a paramedical research protocol to improve the positioning of the hemiplegic shoulder in stroke unit and prevent shoulder pain acute stroke.

Objective: show the superiority of an approach to prevention of shoulder pain of patients post-stroke, in stroke unit.

A clinical trial, a comparative of superiority, monocentric, in two parallel groups is in included phase.

Investigators are initiating a study on new positioning devices providing optimal positioning of the hemiplegic shoulder according to the criteria given in the literature: elbow flexed at 40°, hand semi-prone, fingers abducted and in extension, and thumb in abduction. The medical devices under study are the SYSTAM'® positioning device for the upper-arm (spine position) and Ultrasling ER 15° DONJOY® (sitting or standing-up position). The study will include 30 acute stroke patients (ischemic or haemorrhagic) with no alertness problems and with a shoulder motor function score between 0 and 2 on the Held-scale. A visual analog scale (VAS) will be used, thus excluding patients with aphasia and dementia. The new positioning will be maintained for a maximum of 1 month or until the Held-scale score reaches 3. The primary end-point is pain (VAS) on Day 2, Day 7 and at 1 month. The secondary end-point is the National Institute of Health Stroke Score (NIHSS) at these times. Other variables are time (in hours) between patient arrival on the stroke ward and positioning, protocol compliance by staff and patients, and the percentage of patients with a diastasis on arrival and on discharge.

Inclusion Criteria:

• NIHSS : level of consciousness ≤ 1
• Patient was admitted in a stroke unit until 48 h post stroke
• Locomotion :Medical Research Council scale ≤ 2 shoulder motility
• Spasticity : Ashworth scale ≤ 1 shoulder's adductor muscle
• Simple commands understanding: subtest 04 - Executive orders of Boston Diagnostic Aphasia Examination (BDAE)

Exclusion Criteria:

• Consciousness disorder (NIHSS>1)
• Shoulder disease background (stroke damage side)
• Depression background under 6 months as diagnosed according to DSM-IV
• Upper limb anesthesia
• Severe aphasia