Lien vers les tableaux de données probantes et la liste des références
The risk of falling is increased following stroke due to leg weakness, impaired balance, visual disturbances, cognitive impairment and sensory loss. During inpatient rehabilitation the reported incidence of falls has been reported to range from 25%-39%. Upon return to the community, the risk increases further. Forster & Young (1995) reported that up to 73% of persons had fallen within 6 months of discharge from hospital following stroke, although serious injuries were not reported frequently. Although observational studies by Maeda et al. (2009) and Said et al. (2013) suggest that patients of an older age are at higher risk of falls (p<0.05 and p=0.039, respectively), Aizen et al. (2007) found that the presence of vertigo was the only significant predictor of falling (OR=9.67, 95% CI 1.15 to 81.85) with age, use of anti-depressants and use anti-hypertensives found to be insignificant. In regards to screening for the potential risk of falls, Nystrom and Hellstrom (2013) reported that the Predict FIRST assessment tool (OR=5.21, 95% CI 1.10 to 24.78, p=0.038) and the Modified Motor Assessment Scale (OR=0.65, 95% CI 0.44 to 0.95, p=0.026) significantly predicted the risk of falling. Additional research from Pinto et al. (2014) suggests that the Timed Up and Go (TUG) test, a tool that measures the patient’s ability to stand from a seated position, walk 3 metres then sit back down, was a significant predictor of falling (OR=1.035, 95% CI 1.196 to 5.740, p=0.016).
Teasell et al. (2002) reported that one third of patients on a stroke rehabilitation unit sustained at least one fall during their stay. Of 238 patients, 88 (37%) experienced at least 1 fall, and almost half of these (45 patients [19%]) experienced at least 2 falls, over the 5-year study period. Injuries were reported in 22% of the falls. There were no differences in stroke type (P =0.393), stroke location (P =0.926), or gender (P =0.741) between fallers and nonfallers; however, there were differences in the scores of all functional measurement scores between the groups. The arm, leg, and foot components of the admission Chedoke McMaster scores were significantly lower for fallers compared with nonfallers (P <0.05). Admission Berg Balance Scale scores were significantly lower in fallers when compared with nonfallers (19.0 ± 13.9 vs. 30.7 ± 16.6, P <.0001). FIM® scores of nonfallers were higher than fallers (P <0.001) and there was an inverse relationship between admission FIM® scores and the number of falls. The average admission FIM® score for one-time fallers was 72.4 ±19.1 but declined to 43.6 ± 22.9 for those who had experienced four or more falls (P <.0001). When functional deficits between the two groups were compared fallers were more likely to be apraxic (P =0.014) and have cognitive deficits (P =0.010).
Czernuszenko & Czlonkowska (2009) assessed the incidence and circumstances of falls in patients during inpatient stroke rehabilitation, the frequency of fall-related fractures and identified the risk factors for single and repeated falls. Two hundred fifty-two falls were reported in 189 (16.3%) patients during the observation period. The incidence rate for any fall was 7.6 per 1000 patient-days (95% CI 6.6–8.5). Almost two-thirds (65%; n=163) of falls occurred in the first two weeks after admission. Most falls (n=207; 82%) occurred during the day between the hours of 6 am and 8 pm with a peak incidence between 11 am–1 pm. Patients fell during activities that included transfers (34%; n=85), while sitting (21%; n=54) and during position changes such as going from a sitting to standing or standing to sitting position (13%; n=32). Falls from bed accounted for 10 percent (n=24) of the events n=24) of the events. In 24 cases, falls resulted from inadequate or insufficient staff assistance (5 falls from bed, 19 falls from a wheelchair or toilet bowl). In three cases, patients slid on a wet floor, and falls occurred in three cases due to inadequate assistance by visitors. Seventy-two per cent (n=182) of falls resulted in no injury; 27 percent (n=67) resulted in bruises grazes or lacerations; and 1.2 percent (n=3) resulted in fractures (proximal femur, humeral bone and pelvis). Other observational studies have found incidence rates varying from 14.7% to 56.3% with mixed stroke (Baetens et al. 2013), lower functional mobility at admission (Mansfield et al. 2013), and 6 and 10-metre walking tests (Morone et al. 2014) being significant predictors of falling. Patients prone to falling have also been shown to have a greater likelihood of being older (p=0.05), Caucasian (p=0.02) and having lower diastolic blood pressure (p=0.01) (Schmid et al. 2013).
There have been very few RCTs conducted evaluating therapies to specifically reduce the occurrence of falls following stroke. Batchelor et al. (2010) conducted a systematic review and meta-analysis to examine the effectiveness of interventions that reduce falls following stroke. The results from 13 RCTs were included. The intervention types examined were classified as: physical therapy, modifying the environment or increasing knowledge, models of stroke care and medications designed to improve bone density. It should be noted that the incidence of falls was often a secondary outcome in the majority of these trials (i.e., they were not designed specifically to reduce falls). Pooling of results was limited to two treatment contrasts (exercise vs. usual care and bisphosphonate use vs. placebo) in three studies. There was no significant effect of exercise on fall rate (rate ratio=1.22; 95% CI, 0.76 –1.98) or proportion of fallers (Relative Risk= 0.77: 95% CI, 0.24- 2.43). Bisphosphonate usage was also associated with a non-significant reduction in the proportion of fallers (Relative risk=0.95; 95% CI, 0.73–1.22).
More recently, the results from two RCTs, designed specifically as therapy to reduce the incidence of falls suggest that falls prevention programs are not effective. Dean et al. (2012) randomized 151 community- based stroke patients to an intervention group that received exercise and task related training or control group that performed an upper-extremity strength training program and cognitive tasks. At 12 month follow up, although patients in the experimental group showed significantly improvement in gait speed, there was no significant difference between groups in the number of patients who fell (n=129, experimental group vs. n=133, control group). Batchelor et al. (2012) randomized 156 patients at high risk of falls into a tailored multifactorial falls prevention group or the control group which consisted of usual care. The falls prevention program consisted of an individualized home-based exercise program, falls risk strategies, education, and injury risk minimization strategies. Patients in the control group received usual care. There was no difference in the falls rate between groups. The intervention group had 1.89 falls/person-year, and the control group had 1.76 falls/person-year, incidence rate ratio=1.10, P=0.74). The proportion of fallers did not differ significantly between groups (risk ratio=0.83, 95% CI, 0.6-1.14), nor was the injurious fall rate (intervention group 0.74 vs. control group 0.49 injurious falls/person-year, incidence rate ratio=1.57, P=0.25). Further, Verheyden et al. (2013) reviewed 10 studies and revealed that although there were no significant reduction in number of falls for both acute and subacute stages post-stroke after exercise interventions, medication interventions revealed promising results with reductions for patients prescribed alendronate (95% CI 25% to 72%, p=0.0021) and Vitamin D supplements (95% CI 28% to 82%, p=0.003).
However, Taylor-Piliae et al. (2014) randomized 145 community-based patients into three exercise programs; a Tai Chi group, a strength and range of motion exercise group, and a usual care group. Patients in the Tai Chi exercise group demonstrated significantly fewer falls than the usual care group (p=0.04). Furthermore, both the Tai Chi and strength and range of motion groups displayed significant improvements in aerobic endurance whereas usual care patients did not (p=0.02 and p<0.01 respectively). All three groups significantly improved in Short Physical Performance Battery (SPPB) scores (p<0.01), SF-36 perceived physical health (p=0.04) and SF-36 perceived mental health (p<0.01). In addition, Van Swigchem et al. (2014) revealed who adopted a long-step strategy in a treadmill obstacle-avoidance intervention demonstrated a 62.9% success rate whereas short-steps resulted in a 29.1% success rate.