The Efficacy and Safety of Physiotherapy in People with Dementia: A Systematic Review

Literature search

We searched CENTRAL, MEDLINE, and PEDro from their inception to July 2022 using “Lewy-body”, “Dementia”, “Alzheimer disease”, “Frontotemporal-dementia”, “Motor Function”, “clinical exercise”, “Physiotherapy”, “Physical-therapy”, “Exercise” as keywords (Supplementary Material). Reference lists from the identified articles were cross-checked to identify any further potentially eligible studies.

Study selection

We included experimental studies (randomized controlled trials (RCT), clinical controlled trials (CCT), non-controlled clinical trials (CT)) and post-hoc analysis, testing a physiotherapy intervention in people with any type of dementia or MCI.

Physiotherapy aims to optimize independence, safety, and well-being, through movement rehabilitation, the maximization of functionality, and the minimization of secondary complications. Interventions are focused on physical capacity training, gait, mobility, balance, sensorimotor coordination, and development, as well as teaching the patient and the usual caregivers adaptive strategies to enhance functionality. Exercise-based interventions are considered part of the physiotherapy armamentarium and therefore were included in this review.

We excluded observational studies, reviews, meta-analyses, comments or editorials, and letters to the editor, written in languages other than English, French, Spanish, and Portuguese. All retrieved abstracts were independently screened by two authors (AS, ML). The full texts of potentially relevant articles were retrieved for further assessment. Disagreements were resolved by consensus.

Data extraction

Six pre-defined domains were analyzed: general information (title, year and journal of publication, aim of the study, study duration, type of intervention), methods (type of study design, method of randomization, achievement of allocation concealment and type of blinding), sample (total number of randomized patients, patients per group, dropouts rate, and sample size calculation), intervention (type, duration, type of control), outcome measures (pre-defined outcomes, time point measures), adverse events (number, type), and study results.

The quality of reporting of the included studies was assessed using the Cochrane Risk of Bias (RoB) tool. This tool quantifies the association between certain design features and estimates of treatment effects. The RoB tool is a two-part instrument, the first part describes what was reported in the trial with enough detail for a judgement to be made based on this information, and the second appraises the risk of bias for each analyzed area and classifies them as having a low, high, or unclear risk of bias.

Two authors (AS, ML) independently extracted data. Discrepancies were resolved through discussion or consultation with a third reviewer (RBM).

Data analysis

The primary outcome was the efficacy of physiotherapy interventions in improving motor functions in people with dementia. Secondary outcomes explored how physiotherapy has been prescribed for dementia, the safety of the interventions and their effect on cognitive function. We summarized the publication characteristics using frequencies and percentages.

General features

Of the 194 included studies, 159 (82%) were published between 2012 and 2022. Studies were published in 97 different journals with the majority published in the Journal of Alzheimer’s Disease (n = 19, 10%; Impact Factor (IF): 3.909) and the Journal of the American Geriatrics Society (n = 11, 6%; IF: 4.18).

The most common study designs were RCT (n = 131, 67.5%), followed by CT (n = 37, 19.1%) and CCT (n = 25, 12.9%). Usual care (n = 51, 32.7%) was the most common control intervention (Table 1).

Type of interventions

Regarding the type of intervention, 76 (39.2%) studies focused on programs composed of a set of interventions. The most frequently used interventions, either alone or in combination, were aerobic training (n = 82, 42%), strength training (n = 79, 41%), balance training (n = 48, 25%), and stretching (n = 22, 11%) (Table 2).

The sessions were most frequently held three times per week (n = 64, 33%) and twice per week (n = 61, 31%). Each session had a median duration of 50 minutes [5; 390] and lasted for a median period of 3.6 months [0–72].

Due to the lack of data and degree of heterogeneity, data regarding the intensity of the interventions could not be analyzed.

Aerobic training

Number of included studies: 81 studies (42%).

Most common frequency (n = 43, 22%): 3 times per week.

Most common duration: 50 minutes.

Reported motor benefits (as a primary outcome, with p≤0.05): improvement in functional mobility (n = 10, 5%), cardiovascular function (n = 6, 3%), balance (n = 6, 3%), muscle strength (n = 5, 3%), ability to perform ADL (n = 4, 2%), gait (n = 3, 2%), flexibility (n = 2, 1%) and dual-task ability (n = 1, 1%).

Reported non-motor benefits (p≤0.05): improvement in neuropsychiatric symptoms (n = 5, 3%), global cognition (n = 5, 3%), executive functions, and memory (n = 2, 1%).

Strength training

Number of included studies: 79 studies (41%).

Most common frequency (n = 19, 10%): 4 times per week.

Most common duration: 70 minutes.

Reported motor benefits (as a primary outcome, with p≤0.05): improvement in functional mobility (n = 10, 5%), ADL (n = 5, 3%), balance (n = 6, 3%), muscle strength (n = 5, 3%), gait (n = 3, 2%), reduction in the number of falls (n = 3, 2%), improvement in flexibility (n = 1, 1%) and dual-task ability (n = 1, 1%).

Reported non-motor benefits (p≤0.05): improvement in neuropsychiatric symptoms (n = 5, 3%), global cognition (n = 6, 3%), and executive functions (n = 2, 1%).

Balance training

Number of included studies: 49 studies (25%).

Most common frequency (n = 3, 2%): twice per week

Most common duration: 45 minutes.

Reported motor benefits (as a primary outcome, with p≤0.05): improvement in functional mobility (n = 8, 4%), ADL (n = 3, 2%) and balance (n = 5, 3%), reduced number of falls (n = 4, 2%), improvement in dual-task ability (n = 2, 1%), and gait (n = 2,1%).

Reported non-motor benefits (p≤0.05): improvement in global cognition (n = 4, 2%) and neuropsychiatric symptoms (n = 3, 2%).

Stretching

Number of included studies: 22 studies (11%).

Most common frequency (n = 4, 2%): 3 times per week

Most common duration: 60 minutes.

Reported motor benefits (as a primary outcome, with p≤0.05): improvement in ADL (n = 2, 1%), stretching (n = 2, 1%), cardiovascular function (n = 1, 1%), muscle strength (n = 1, 1%) and functional mobility (n = 1, 1%).

Reported cognitive benefits (p≤0.05): Unknown

Safety aspects

A total of 1119 adverse events were reported in 24 (12.4%) of the 194 studies that were included. In 129 studies (66.5%) safety-related aspects were not mentioned.

The most common adverse events were discomfort/pain (n = 463, 41.7%), dizziness (n = 181, 16.2%), falls (n = 92, 8.2%), stress, anxiety, and depression (n = 76, 6.8%) occurred during and after exercise.

The intervention associated with a higher percentage of studies reporting adverse events were strength training (n = 14, 17.7%), aerobic training (n = 11, 13.6%), and balance training (n = 6, 12.2%). Table 3 summarizes the frequency of adverse events per type of intervention.

Sample characterization

The median sample size was 61 [4–6104] participants, all of whom were over the age of 59. Thirty-eight (19.6%) of the included studies focused on AD patients. The most common disease stages included were mild to moderate dementia (n = 51, 26.3%), mild dementia (n = 39, 20.1%), and MCI (n = 29, 14.9%).

Quality of reporting analysis

Twenty-seven (13%) of the included studies had a high risk of bias in all domains and 80 (52%) had a low risk of bias in at least half (4/7) of the domains (Fig. 2). In 109 (56%) of the studies, the randomization process was correctly reported. Half of the studies (n = 99, 51%) used a single-blinding approach.

OPEN IN VIEWER

Fig. 2

Risk of bias assessment.

The primary outcome was specified in 100 (52%) of the studies. The mean dropout frequency was 14%.

Are physiotherapy interventions effective for dementia patients?

Although the included studies have poor methodological quality, they consistently reported motor and cognitive benefits (p≤0.05) associated with physiotherapy interventions, specifically to clinical exercise.

Although dementia is characterized by the interference of cognitive impairments in patients’ daily lives, a variety of motor symptoms (e.g., impaired coordination, abnormalities of gait, postural instability) are usually present over the course of the disease [14–16]. Our findings indicate positive results of clinical exercise on functional mobility, ADL performance, dual-task ability, and reduction in the number of falls, which appear to be useful in reversing the physical decline that patients experience. Additionally, physiotherapy has demonstrated several cognitive benefits, such as improvement of global cognition, memory, and execution functions. According to previous studies, these can be directly attributed to the slowing of cognitive degeneration and, indirectly, to the patient’s capacity for routine maintenance and social interaction [14, 15].

What are the most suitable physiotherapy interventions?

Due to the high heterogeneity of our findings, it is impossible to define a physiotherapy prescription protocol for people with dementia. However, there seem to be some trends that may help define this protocol in the future. A high percentage of the studies (n = 76, 39.2%) used a multimodal intervention (i.e., a program composed of more than one type of intervention), most of which were based on clinical exercise (aerobic, strength, balance, and stretching). The most common frequency and duration were 3 times per week, with a median duration of 50 min.

Clinical exercise is the most extensively researched intervention— in the current review more than half of the included studies focused on patients with MCI or mild to moderate dementia, and in another review with the same objective that focused on patients with severe dementia [14]— supporting the notion that physiotherapy in dementia is an exercise-based intervention.

This does not imply that other intervention strategies, like transfer training or teaching compensatory techniques, should be ignored. These can help patients cope with ADL limitations but may be difficult to implement in those with dementia because they require patients to focus attention, be aware, describe movements verbally, reflect on performance, and compare outcomes to previous performances [17]. They are generally easier to implement in the early stages of dementia and become more challenging as the disease progresses. However, physiotherapists should always evaluate how useful such interventions are because the relationship between dementia severity and procedural learning abilities is still unclear [17].

A previous review [14] highlighted how functional exercise, such as walking and dancing, potentiates the cognitive benefits of physiotherapy. Such exercises can, however, be challenging for patients with cognitive impairment. This challenge can be overcome by using exercises that mimic tasks with which the patient is familiar.

A high percentage of studies used multimodal programs. These offer a compelling option because, in addition to taking a more thorough approach and addressing a higher range of motor issues, they also increase cognitive stimulation by incorporating a greater variety and more frequent exercise switches.

Such methods may reduce or delay the long-term risk of non-compliance associated with non-pharmacological interventions such as exercise and physiotherapy.

Is physiotherapy a safe intervention in dementia?

Only 65 (33.5%) of the included studies reported on adverse events.

As physical therapy and exercise are interventions that do not require the use of drugs or invasive procedures, they are mistakenly assumed to be risk-free. This poor perception is usually associated with less attention to safety issues in clinical practice and research [18].

A total number of 1,119 adverse events were reported, with discomfort/pain, dizziness, and falls the most frequent. Although none can be considered serious adverse events (i.e., an event requiring hospitalization or causing persistent or significant incapacity), they represent a risk that should be minimized. It is, therefore, critical to increase awareness among physiotherapists of the importance of conducting a thorough assessment before sessions begin, to identify all potential risks, and, based on that information, to develop a risk minimization plan. The tasks performed in each session should be documented in detail and the physiotherapist should seek to determine, without causing alarm, whether adverse events occurred and if they are related to the intervention [18].

Implications for clinical practice and research

Physiotherapy has several motor and cognitive benefits in dementia, and it seems to be an important adjunct to pharmacological interventions. Although this type of intervention cannot follow the “one size fits all” rule, it is important to establish a physiotherapy prescription protocol for MCI and each stage of dementia, specifying the most suitable type of intervention, frequency, duration, and intensity [5].

As with medications, it is also important to understand how long physiotherapy-related benefits last. The mean duration of studies in this review was 24 weeks, and the mean follow-up duration was 20 weeks. Studies with a longer follow-up show that positive effects remain after eight weeks. However, due to few studies reporting data on follow-up assessments and the heterogeneity of the outcome tools used, long-term efficacy needs to be determined. For this purpose, a suitable motor assessment protocol for dementia patients is crucial, as are longer, larger, and better-designed RCTs. Since most clinical tests require patients to comprehend, memorize, and put into practice the given instructions, assessing patients with dementia can be challenging. A recent systematic review [19] on this topic recommended the most suitable assessment tools while emphasizing the importance of continuing to develop technology-based objective measures that provide accurate data on motor impairments with little or no cooperation from the patient [19].

Safety concerns need more attention. They can be operationalized in research by reporting the frequency and describing the adverse events (type of event, place, type of intervention being performed, time to occurrence, seriousness, severity, functional impact, and causality) [18]. Careful patient assessment, the creation of risk-reduction plans, and an efficient registry of the session’s content and patients’ complaints during or after the session can all help increase patient safety in clinical practice [18]. Therefore, educating and communicating with family members and caregivers is important especially as patients are unlikely to provide sufficient information.

Limitations

Due to the high level of heterogeneity in the interventions and study protocol (i.e., outcomes, measures instruments and time point measures) of the included studied, it was not possible to conduct a more in-depth analysis that would have allowed us to come to a more robust conclusion regarding the effectiveness of physiotherapy intervention in dementia. However, due to the scarce evidence on this topic, we believe that the data reported in this review is important not only to help design future studies on this topic, but also to help guide healthcare professionals in their clinical practice. To overcome this limitation in the future, we recommended establishing the most suitable way how to evaluate physiotherapy intervention in dementia.

Conclusions

Our research indicates that physiotherapy has in dementia several motor (e.g., improvement of functional mobility, ADL performance, dual-task ability, and the reduction in the number of falls) and cognitive benefits (e.g., improvement of global cognition, memory, and execution functions). Our results indicate that the best strategy appears to be to use a multimodal approach that includes the four types of physical exercise (aerobic, strength, balance, and flexibility), though more substantial data are required. They also emphasize the importance to pay attention to safety aspects in order to reduce the likelihood of adverse events. More research is required to develop a protocol for prescribing and assessing physiotherapy interventions in dementia. The most effective way to prescribe these interventions for the various forms and stages of dementia (including MCI) should also be explored.