Sleep is crucial to health, since during this process a number of reactions take place to maintain body homeostasis (1). Sleep-related problems are highly prevalent worldwide. Insomnia is defined as a sleep disorder characterized by dissatisfaction with qualitative or quantitative sleep (2), especially, is between 12 and 20% prevalent in older adults (2) and a challenge for public health, given the risks it poses for healthy aging. Sleep problems and disorders are associated with higher risks of mortality, cardiovascular diseases, depression and other conditions (3–5).

Meanwhile, another challenge to healthy aging is frailty, a multidimensional condition that leads to a decline in the ability to resist internal and external stressors, thus increasing both harmful effects on various organs and individual vulnerability (6) and associated with higher risk of hospitalization, falls and mortality (7, 8).

There has been growing interest in investigating the association between sleep quality and duration and frailty in older adults, because both conditions are precursors of disability and morbimortality (9). However, studies of this relationship have returned inconsistent results. Some studies find an association between only short (< 5 h) sleep duration and frailty (10, 11), while others find an association with only long (> 9 h) sleep duration (12) and others observe frailty to be associated with both short and long sleep duration (11).

In addition, methods of measuring both sleep and frailty vary considerably, and this may be one of the reasons for inconsistent results. Sleep can be evaluated by using subjective and objective measures (13), and studies that combine both can provide more comprehensive information about sleep quality and duration. Objective measurement is particularly important in studies of populations of older adults, because of the cognitive decline, including difficulties with memory, common in this age group (14). Similarly, various tools are used to assess frailty, chiefly the frailty phenotype (7), frailty index (15) and frailty scale (16).

Accordingly, this systematic review investigated the association between sleep quality and duration and frailty in non-institutionalized older adults, with emphasis on the methods employed in the studies to evaluate these events.

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (17) and registered in the International Prospective Register of Systematic Reviews (PROSPERO) under No. CD42023441262.

The research question, “Is there an association between sleep quality and frailty in older adults?,” applied the following PICOS strategy:

P (population): community-dwelling older adults over 60 years of age.

I (exposure): sleep quality as diagnosed by the criteria used in the original studies.

C (control): does not apply.

O (outcome): frailty as diagnosed by the criteria used in the original studies.

S (study design): observational studies.

The 17 studies included (Figure 1) evaluated a total of 58,447 older adults, whose average age ranged from 64 to 84 years. Most studies were cross-sectional and only four were prospective cohorts (20–23). In 10 studies, the population was mostly female. Most of the studies were conducted in China (12, 23–26), followed by Mexico (10, 11, 27) and the United States (20, 28, 29) (Table 1). No studies were retrieved by the reverse search.

Sleep was evaluated by different assessment measures, most of them subjective – Pittsburgh sleep quality index (n = 9), Epworth sleepiness scale (n = 4), STOP-Bang questionnaires (1), insomnia severity index (n = 1), Athens insomnia scale (n = 4), sleep scale (n = 1), Jenkins sleep scale (JSS-4) (n = 1), Berlin questionnaire (n = 1) and questions (n = 3), such as “Have you had problems with sleep recently?” “Have you had any trouble sleeping recently?” and “Normally, how many hours do you sleep?” Four studies used objective measurements: actigraphy (n = 2) and polysomnography (n = 2). Three studies used a single self-reported question to assess sleep (11, 23, 27) and only two made combined use of objective and subjective measures: the Pittsburgh sleep quality index (PSQI), Epworth sleepiness scale, actigraphy and polysomnography (20, 29) (Table 1).

Prevalence of insomnia ranged from 0.8 to 46.5% (21, 26); of poor sleep quality (IQSP >5), from 40.7 to 65.9% (30); of short sleep duration (5 h), from 11 to 25.5% (20, 23); and of long sleep duration (9 h), from 4.7 to 38.5% (10, 11). Another study, using a different cutoff point for sleep duration, revealed 13.4% prevalence of short sleep (< 6 h) and 12.5% of long sleep (> 8 h) (31). A single study estimated a prevalence of 8.9% for severe sleep apnoea, 16.6% for poor sleep efficiency and 9.2% for prolonged latency (20) at baseline (Table 1). Also, 32% of older men and 40% of women had sleep disorders (9).

Frailty was identified using mainly the Fried frailty phenotype (n = 12), frailty index (n = 2), Tilburg frailty indicator (n = 2), frail scale (1), Groningen frailty indicator (n = 1), Kihon checklist and comprehensive frailty assessment tool (n = 1). Note, however, that more than half the studies (66.7%) used the Fried frailty phenotype (9–12, 20, 22, 24, 27–31) (Table 1).

Studies using the Fried frailty phenotype found prevalence of pre-frailty ranging from 24.1% (10) to 53.8% (11), and frailty prevalence of 9.2% (11) to 41.2% (28). A single study, which evaluated frailty by other methods, pointed to prevalence of frailty of 19.6% (frailty index), 31.2% (Tilburg frailty indicator), 30.6% (Groninger frailty indicator) (30) (Table 1). The only two studies assessing incidence of pre-frailty found 35.8% (21) and 50.5% (20). Incidence of frailty was found to be 14.4% (20), 30.0% (21), 30.4% (22) and 33.2% (23).

After adjustments, one cross-sectional study carried out with 1,643 older adults Mexicans revealed that the high risk of OSA associated with frailty (OR = 1.59; 95% CI = 1.07–2.38) in the global sample and among women (OR = 1.84; 95% CI = 1.05–3.22), on the contrary, among men there was no association (OR = 1.65; 95% CI = 0.65–2.19) (11).

Excessive sleepiness was evaluated in five studies evaluated excessive sleepiness and only one cross-sectional study observed its association with frailty (OR = 3.67, 95% CI = 2.03–6.61) (28).

One cross-sectional study revealed that reduced sleep efficiency (OR = 1.37; 95% CI = 1.12–1.67) and sleep latency (OR = 1.42; 95% CI = 1.10–1.82) were associated with frailty (29). On the other hand, a longitudinal analysis of that same population found no association between sleep latency and frailty (OR = 1.20; 95% CI 0.84–1.71), while nighttime wakefulness of 90 min or longer was the only parameter of objective sleep–wake patterns associated whit frailty/dead or dead (OR = 1.31; 95% CI = 1.04–1.66 and 1.57 95% CI = 1.12–2.20 respectively) (20). Poor sleep quality (HEI > 5) was associated with pre-frailty (12, 30) and frailty (9, 24, 29–32), but only in cross-sectional studies. Sleep complaints in older adult women living in rural areas were associated with frailty in a cross-sectional analysis (OR = 3.20; 95% CI = 1.33–7.68) (27), but not in a longitudinal analysis (RR = 1.41; 95% CI = 0.94–2.12) (11).

In the 80-year-old age group, poor sleep quality was independently associated with frailty (OR = 1.78; 95% CI = [1.02, 3.11]) (31). Older adults with a sleep disorder were 0.66 times more likely (OR = 1.66; 95% CI = [1.47—1.87]) to be frail than those with no sleep disorder (p < 0.001) (9).

Insomnia was associated with frailty in both cross-sectional analyses (10, 26) and longitudinal analyses (21, 22). Short (< 5 h) sleep duration was associated with frailty in a cross-sectional analysis (OR = 2.67; 95% CI = 1.01–7.08) (10), but not in a longitudinal study (OR: 1.15; 95% CI = 0.82–1.62) (20). Long (9 h) sleep duration was associated with frailty, but only in cross-sectional analyses (12, 23, 32).

Methodological quality evaluation of 14 cross-sectional studies classified only one as being of low quality (Supplementary Table S1), while none of the four cohort studies was found to be of low quality (Supplementary Table S2).

The 17 studies included in this systematic review found that insomnia, short and long sleep duration, poor efficiency, prolonged latency, obstructive sleep apnoea and poor sleep quality are associated with pre-frailty and frailty in community-dwelling older adults in several different countries.

A variety of methods for assessing exposure and outcome were identified. On the one hand, this reflects the complexity of evaluating both sleep and frailty and, on the other, it indicates the need to standardize methods for evaluating these events. Use of a single self-reported question to assess sleep may fail to consider other sleep problems.

In that light, use of a combination of objective and subjective measures of sleep can yield more robust results. Nonetheless, this review found that studies using objective measures of sleep are scarce and, in longitudinal analysis, no association was observed between the sleep parameters evaluated and frailty. This reinforces the need for further longitudinal studies using objective sleep assessment measures.

Methods used assess frailty differed in the dimensions evaluated, reflecting the multidimensionality of this condition. The frailty phenotype, for example, evaluates five criteria in the physical dimension (fatigue, involuntary weight loss, muscle weakness, gait speed and physical activity) and individuals can be classified on these criteria as robust, pre-frail and frail (7). Other approaches address non-physical (e.g., cognitive and clinical) dimensions. However, these approaches are not always easy to apply in clinical practice and epidemiological research, as they can be time-consuming and entail broader evaluation.

Of the 17 studies evaluated, 16 found independent associations between sleep problems and frailty. As regards mechanisms for this association, one hypothesis is that alterations in sleep may cause hormonal changes, such as decreases in growth hormone, insulin-like growth factor and testosterone production, with impacts including increased muscle protein breakdown, which favors the emergence of geriatric syndromes (sarcopoenia and frailty) (33). This occurs with sleep disorders that shorten the deep sleep phase (34), which also results in increased cortisol levels, culminating in impaired physical function (33).

Given that inflammation is known to be one of the pathophysiological bases of frailty (7), the association between sleep quality and frailty stages could also be explained on the basis that sleep disorders, such as insomnia and long sleep duration, favor changes in inflammation levels involving increased C-reactive protein, interleukin-6 and tumor necrosis factor (TNF-α) (32, 35, 36). In addition, long-duration sleep may result from sleep disorders such as insomnia, obstructive sleep apnoea and restless leg syndrome (37). Also, short-duration sleep can also result in lowered testosterone levels and increased oxidative stress (33), thus contributing to the development of frailty.

Among the studies that showed significant associations between sleep disorders and frailty (n = 17), almost all (n = 13) performed adjustments for confounding variables and were adjusted for sex and age.

As age progresses, hormonal changes, such as a gradual reduction of free testosterone, occur in both sexes, culminating in greater susceptibility to frailty in men and sarcopenia in women (35, 38, 39). Sarcopenia, in turn, is part of the pathophysiological basis of frailty. Also, unlike men, women with over 5 years of menopause or early menopause, and thus falling estrogen levels, are at higher risk of musculoskeletal disorders, such as sarcopenia, which can progress to frailty (40). Common changes in aging include reduced resting energy expenditure in skeletal muscle. This decline occurs faster in females, suggesting that women are more prone to developing frailty (38, 41, 42). Since the evidence above indicates important differences between the sexes, studies that stratify by sex are important in order to understand better how the association between sleep and frailty behaves in men and women.

It is known that sleep problems may occur in older adults due to the presence of comorbidities, thus increasing their susceptibility to frailty (20). This underlines the need to consider these conditions in the process of adjustment for confounding variables in statistical analysis. Ten of the studies included in this review performed adjustments for comorbidities.

However, no pattern was observed among the studies in the selection of confounding variables for adjustment, which partly undermines the degree to which their findings approximate to reality.

Furthermore, a study included in this review evaluated the bidirectional relationship between frailty and insomnia and concluded that the association between frailty and insomnia was significant, especially in men (21), which highlights the need for future studies on the causal relationship between frailty and sleep quality and duration.

Our results also point to the need for more comprehensive assessment of sleep in older adults, including screening, diagnosis and intervention measures to improve sleep quality, mainly because several aspects of sleep quality are associated with physical frailty in this age group. Most health care services for frail older adults do not currently screen for and treat sleep disorders. Given the importance of frailty in current geriatric medicine, it is important to perform a comprehensive geriatric evaluation in older patients with sleep disorders.

We observed that poor sleep quality (insomnia, sleep deprivation, obstructive sleep apnea, excessive daytime sleepiness) and short/long sleep duration was associated with pre-frailty and frailty in older adults. This combination may result in other adverse health outcomes compromising the quality of life of this population.

The results of this review also suggest a need for further research on the relationship between sleep and frailty, including longitudinal studies and surveys stratified by sex.