In order to carry out a comprehensive, objective and reproducible systematic review, two investigators (ET and SS) independently proceeded to perform an advanced search for studies in different health-related bibliographic databases (Pubmed, PEDro, The Cochrane Library Plus , Biblioteca virtual de la Salud (VHL) and PyscoINFO) based on the Preferred Reporting Items for Systematic and Meta-Analyses-PRISMA recommendations 10, and using Mesh descriptors and keywords, together with the Boolean operators AND and OR . to expand the search. The Descriptors in Health Sciences (DeCS) page was used to choose the keywords used in the search: dementia , Alzheimer disease , exercise , physical activity , Physical Therapy modalities , non-pharmacological interventions , usual care , sleep quality , sleep disorders , sleep fragmentation , total sleep time , insomnia , circadian . The search was carried out between November 2021 and February 2022.

The eligibility criteria that should be included in the studies belonging to this review were described based on the PICO system (Patient/problem, intervention, comparison, outcomes) 11. Therefore, the studies show the following characteristics: a) Sample: Patients diagnosed with dementia with moderate cognitive impairment (DCM) or higher, that is, they must have obtained a 4 in the classification made by the Reisberg Global Deterioration Scale (GDS) and < 10 points in the Mini-Mental State Examination (MMSE); b) Intervention: Any type of treatment modality to improve disrupted sleep, except pharmacological treatment; c) Comparison: the effect of the interventions will be with some type of treatment for dementia or its usual care that is carried out in both groups (control-treatment group); d) Result: Improvement of disrupted sleep in the patient with dementia.; (e) Type of studies: Randomized clinical trials and controlled trials. The search included studies published up to February 2022, accepting English and Spanish as languages.

The included studies, being randomized clinical trials, underwent an assessment of the quality of the methodology using the Physiotherapy Evidence Database (PEDro) scale 12. Scale that allows observing internal and external validity, identifying those that are internally valid and have sufficient statistical information to guide clinical decision-making. It is made up of 11 items with a Yes or No answer, although its final score is in a range of zero to ten because criterion one is not used to calculate the total score: Score of nine-ten: excellent quality; Score six-eight: good quality; Score of four-five: poor quality; Score

Any disagreement between the authors was resolved by seeking the third author's opinion.

Three authors (ET, SS and AG) separately extracted the articles information, including the first author's name, publication year, study design, country, the number of participants, age, gender percentages, outcomes, instruments, quality score, and results. All the collected information was recorded in a checklist.

The studies included in the systematic review have been published between 1999 and 2012. With a total sample of 449 subjects analyzed, of which 163 participants were part of the control groups and 286 participants of the experimental groups (considering that we have a crossover study). The sample size of the studies included in our review varies from the 15 participants of Lyketsos et al. 17 to the 132 participants of McCurry et al. 20 representing the largest set. All the subjects comprised an age range between 60 and 86.15 years, except for the study by McCurry et al. 19 which increases the range to 101 years. The mean of the range of all ages is 82 years.

The methodological quality of the eight randomized clinical trials studied using the Physiotherapy Evidence Database (PEDro) scale is between six-seven points out of 10, except for the study by Harris et al. 12, which scores three/ten. In none of the studies were the subjects blinded or blinded. (Table 1)

The analysis of the studies is carried out by subject matter. (Table 2)

Lyketsos et al. 17 perform a crossover clinical trial where 15 subjects diagnosed with dementia were randomized into an experimental group, where the treatment was bright light therapy (BLT) administered for 1 hour every morning using a full-spectrum lamp of 10,000 lux at 3 feet, and a control group, where a dim, low-frequency light was used. Patients who had been treated with medications for problem behaviors or sleep disturbance were only enrolled in the study if they had been on a stable dose of these medicines for 1 week prior to enrollment and continued to meet inclusion criteria. Subjects spent 4 weeks in one condition and a second baseline was taken. They were then treated in another condition for an additional 4 weeks. Regarding the results measured by actigraphy and different scales, the subjects who received BLT slept an average of 6.4 hours per night during the week prior to treatment. At two weeks of treatment, they were sleeping an average of 7.6 hours each night, and at four weeks of treatment they were sleeping an average of 8.1 hours per night. The improvement in nightly sleep between baseline and week two was not statistically significant. The improvement between baseline and four weeks was statistically significant (p<0.05). The difference between the control and active treatment groups at two weeks was borderline significant, and at week four it was not statistically significant, although all study patients, including those in the control group, showed slight improvements in nighttime sleep.

The two studies by Dowling et al. 1415 tested the effect of morning BLT on nighttime sleep, daytime awakening time, and rest activity rate in 46 subjects diagnosed with dementia. This study was conducted in two phases, the first phase compares morning light exposure, between 8:00AM-20:00PM, with normal light. Chart reviews were conducted to confirm that potential subjects met the criteria for inclusion, like a stable medication regimen. Potential subjects were excluded if they were regularly taking valerian, melatonin or sleeping pills. No subjects were excluded based on other medication usage. In the first phase of the Dowling et al. 15, the experimental group, with 29 subjects, received a BLT program >2500lux in the direction of gaze, Monday through Friday from 9:30-10:30 A.M. for 10 weeks. The control group, with 17 subjects, received usual light (150-200lux). Data were recorded by means of an actigraph the first week and week 12 (post-intervention). Regarding the results, it revealed no significant changes in sleep efficiency, sleep time, wake time, or number of awakenings in the experimental group compared to the control, nor were there any significant differences between the groups at daytime wake time. half. Regarding the circadian rhythm results, there were no significant differences in acrophase between the groups, but there was a trend towards significance in amplitude with improvement in the experimental and worsening in the control (p<0.09). There were also no significant differences in the non-parametric analysis variables of the circadian rhythm.

In the second phase of the study by Dowling et al 15, morning BLT (experimental group one with 29 subjects) is compared with afternoon BLT (experimental group two with 24 subjects) with a mean age 84 years diagnosed with dementia. Experimental group two receive BLT >2500lux in the direction of gaze, from Monday to Friday from 15:30-16:30pm. for 10 weeks. The control group consisting of 17 subjects receives normal light (150-250 lux) and participates in regularly scheduled activities. With respect to the results, the analysis of covariance of repeated measures did not reveal significant differences in the variables of activity or daytime or nighttime sleep between the groups. However, there was a significant (p<0.04) main effect of treatment for acrophase with relative stability in the morning and afternoon light experimental groups. The mean peak of the rest activity rate was 131 minutes later in the control group and advanced in the morning and afternoon experimental groups, 17 minutes and 1 minute, respectively.

Burns et al. 13 evaluate the benefit of four-week BLT in a study of 48 subjects recruited from two nursing homes with dementia and behavioral disorders. The 22 subjects in the experimental group underwent a BLT program that consisted of exposing themselves to full-spectrum 10,000 lux for two hours a day between 10:00 and 12:00 in the presence of a nurse. The control group, with 26 subjects, received standard fluorescent tube light at 100 lux. Data were recorded using an actigraphy and scales such as: MMSE, Cornell Scale for Depression in Dementia (CSDD), Manchester and Oxford Universities Scale for the Psychological Assessment of Dementia (MOUSEPAD) and Crichton Royal Behavior Rating Scale (CRBRS), the first week, the fourth week (post-treatment) and after a follow-up in week eight. Any changes in psychotropic medication were noted. Regarding the results on psychological and cognitive symptoms, there were no significant differences between groups in the different scales, but improvements were obtained in both groups. Only one change to medication was recorded – a person in the treatment group who had her trazodone increased (antidepressant drug). Agitation improved in the immediate post-treatment phase, week four, in both the BLT and placebo groups, but the comparison between groups was not significant at either week four or week eight. In relation to the results, mean duration of night sleep measured by the actigraphy was 8.3 hours and 8.6 hours before and after exposure to BLT, respectively. There were no significant differences between the groups before or after therapy as measured by the actigraphy.

In the study by McCurry et al. 18 a total of 36 subjects are randomly divided into an experimental group (n=17) with an education program is carried out, combining sleep hygiene, daily walking, and BLT (greater than 100lux) three sessions a week for 2 months, and a control group (n=19) where education about dementia and support for the caregiver is carried out. Both groups also carried out 6 sessions of 1h at home for behavioral intervention in patients with dementia. Caregivers also recorded daytime naps (frequency and duration), use of sleeping medications, frequency and duration of nighttime disturbances, and subjective sleep quality ratings. After finishing the treatment, significant differences were obtained measured by the actigraph to assess sleep. Subjects in the experimental group spent an average of 36 minutes less awake time at night and had 5.3 fewer nighttime awakenings than subjects in the control group. Also, the patients in the experimental group exercised significantly more days per week and had significantly lower levels of depression. After follow-up, significant differences remained at six months favoring the experimental group in actigraphy estimates of patient awake time, exercise days, and depression.

These same authors, in a subsequent study 20 carried out a sleep education program to improve sleep in older adults with dementia. A total of 31 subjects in the experimental group undergo four sessions of sleep education program, and a control group with 16 subjects receive usual medical care for four weeks. An eligibility criterion was agreement to make no change in sedative medication use (type or dose) during the 2-month active treatment period. Patients were stratified by baseline sleep medication use. The random allocation sequence was obtained from a computer program that blocked in groups of 12 subjects. The variables were recorded with actigraphy at the beginning, at one month (post-treatment) and at six months of follow-up. Posttreatment, there were no significant differences in residents' actigraphy measures of sleep. However, there was a tendency to improve the percentage of night sleep for subjects in the experimental group. After 6 months of follow-up, they showed significant differences in the percentage of sleep (p = 0.040) and total sleep time (p = 0.013). At six months, subjects in the control group were awake an average of 24 minutes longer each night than subjects in the sleep education program condition. Regarding secondary outcomes, there was a significant increase over time in caregivers' ability to identify triggers and consequences of nighttime behaviors, develop behavior plans, and implement daytime events with residents. When comparing the two treatment groups, sleep education program group were significantly less depressed after treatment than control subjects, (p = 0.0036), but there were no differences between groups in daytime sleepiness or behaviors.

McCurry et al. 19 with 132 subjects diagnosed with dementia, developed a four-week program with the objective of testing the effects of walking, BLT, and a combined intervention (walking, light, and SEP) on sleep in people with Alzheimer's. For this, the participants were randomly assigned to one of the four treatment conditions: Walking group with 32 subjects where they had to walk at their own pace with a goal of 30 continuous minutes a day. BLT group with 34 subjects where they sat in front of the light box for one hour a day. The Nighttime Insomnia Treatment and Education in Alzheimer’s Disease (NITE-AD) group with 33 subjects where the caregivers had an individualized sleep plan session for the patient, a daily walking program, daily light exposure. The control group, made up of 33 subjects, was offered support for dementia care, but did not receive training or tasks related to sleep hygiene, walking, or exposure to light. Data on sleep variables were recorded at the end of treatment (four weeks) and after a six-month follow-up. Regarding sleep outcomes: patients in each active treatment had reduced actigraphy total wake time after four weeks of treatment compared to control subjects, participants in the Walking condition were awake 33.1 minutes less/ night (p = 0.05); Participants with BLT were awake 39.0 fewer minutes; and NITE-AD subjects were awake 39.8 minutes less. There were no significant differences in total wake time change scores between the experimental groups, except for a slightly smaller difference in wake time for participants in the BLT group (p = 0.06). compared to controls. Although not significant, there were moderate improvements in percentage sleep for all three active treatment groups (NITE-AD, 5% improvement; walking and light exposure, 3% and 4.3% improvement, respectively). In addition, there was a trend for subjects in the walking group to have fewer nocturnal awakenings than those in controls. At six-month follow-up, there were no significant differences in any treatment group regarding control for actigraphy or subjective sleep outcomes.

Harris et al.16 try to show the sleep benefit and life quality of people with dementia through a back massage characterized by long, slow, sliding, and repetitive strokes from lumbar to cervical region. We are counting on a total of 40 subjects with dementia distributed in two groups, the experimental with 20 patients where a three-minute back massage protocol is used at the time of toing to bed for two nights. And the control group with other 20 subjects that receive usual control. All participants received usual bedtime care No changes in the physical setting, nursing services, staffing patterns, medications, or patient care protocols were instituted as part of the study. Demographic data, medical diagnoses, medical history, and medications were collected via chart abstracting. This study used seven hours of night sleep as a cut reference for people who sleep “badly”. Data were registered through actigraphy, and the results did not reveal any significant differences between intervention and control over sleep latency, sleep efficiency, awakening after the beginning of the sleep and daytime inactivity. There was insufficient power to establish effectiveness.