Relationships between self-reported dyspnea, health conditions and frailty among Brazilian community-dwelling older adults: a cross-sectional study

ABSTRACT CONTEXT: Dyspnea is a symptom present in several chronic diseases commonly seen among older adults. Since individuals with dyspnea tend to stay at rest, with consequently reduced levels of physical activity, they are likely to be at greater risk of developing frailty, especially at older ages. DESIGN AND SETTING: Cross-sectional study at community level, Brazil. OBJECTIVE: To analyze the relationships between self-reported dyspnea, health conditions and frailty status in a sample of community-dwelling older adults. METHOD: Secondary data from the follow-up of the Frailty in Brazilian Elderly (FIBRA) study, involving 415 community-dwelling older adults (mean age: 80.3 ± 4.68 years), were used. The variables analyzed were sociodemographic characteristics, reported dyspnea, clinical data and frailty phenotype. Associations between dyspnea and other variables (age, sex, education and body mass index) were verified through the crude (c) and adjusted (a) odds ratios. RESULTS: The prevalence of dyspnea in the entire sample was 21.0%. Dyspnea was more present in individuals with pulmonary diseases, heart disease, cancer and depression. Older adults with multimorbidities (adjusted odds ratio, ORa = 2.91; 95% confidence interval, CI = 1.41-5.99) and polypharmacy (ORa = 2.02; 95% CI = 1.15-3.54) were more likely to have dyspnea. Those who reported dyspnea were 2.54 times more likely to be frail (ORa = 2.54; 95% CI = 1.08-5.97), and fatigue was their most prevalent phenotype component. CONCLUSION: Dyspnea was associated with different diseases, multimorbidities, polypharmacy and frailty. Recognizing the factors associated with dyspnea may contribute to its early identification and prevention of its negative outcomes among older adults.


INTRODUCTION
Dyspnea is a symptom that is present in different chronic diseases, with high prevalence among older adults. 1,2 The American Thoracic Society defines dyspnea as "a subjective experience of breathing discomfort that consists of qualitatively distinct sensations that vary in intensity". 3 Since dyspnea is a frequent symptom among community-dwelling older adults, 2,4 it is important that healthcare professionals should routinely evaluate and document the presence of dyspnea, just like they do for pain. 3,5 In clinical practice, dyspnea is initially assessed by a single question (e.g. "Are you short of breath? Yes or No). 6 Single-item questionnaires have clinical importance, as they are simple and quick to use, but they fail to cover all aspects of dyspnea. 7 A large number of scales for measuring dyspnea have been described in the literature. Scales based on reports of breathlessness during activities of daily living (e.g. the modified Medical Research Council Dyspnea Scale and the Baseline Dyspnea Index) and on the perception of breathlessness while making exertions (e.g. the modified Borg scale and visual analogue scale) are the ones most frequently used. 8 Because multidimensional tools such as the Multidimensional Dyspnea Profile have the ability to assess different aspects of dyspnea, they are promising tools for use in both laboratory and clinical settings. 9 Nonetheless, most breathlessness scales were developed and validated for patients with respiratory diseases and, consequently, they need to be properly tested before being applied to other populations. 7 A systematic review showed that the estimated prevalence of dyspnea among community-dwelling older adults was 36%, and that dyspnea can have an impact on different activities of daily living. 2 In Brazil, a cross-sectional study among community-dwelling older adults observed that 30.9% of the participants had dyspnea symptoms. Moreover, most of the older adults who reported dyspnea had poor physical performance, were women, had multimorbidity and were frail. 4 In addition to presenting an association between dyspnea and cardiorespiratory diseases, older adults with moderate to severe dyspnea are more likely to be unable to perform a single chair stand, have depressive symptoms, use psychoactive drugs and be underweight and/or obese, than those with no or mild dyspnea. 1,10 Because presence of dyspnea is associated with several chronic diseases and health issues, some authors have argued that dyspnea in older adults should be considered to be a multifactorial geriatric condition. 10,11 Common adaptive responses to the presence of dyspnea include avoidance of physical effort and, consequently, adoption of a sedentary lifestyle. 3 In patients with chronic obstructive pulmonary disease (COPD), this can precipitate a downward spiral of dyspnea-induced inactivity, thus resulting in physical deconditioning, progression of disease and disability. 12 Thus, some studies have shown that COPD patients were more likely to be frail, and that this was associated with airflow limitations, reduced physical function, dyspnea, disability, anxiety and depression. 13 Frailty is a medical syndrome with multiple causes and contributors that confers vulnerability to stressors and increases the risk of different negative outcomes at older age, such as dependency and/or death. 14 Regarding physical frailty, physical inactivity plays an important role in the progression of this syndrome, as it directly influences some determinant components for development of frailty. 15 Although the overlap between dyspnea-induced inactivity and vicious cycles of frailty seems obvious, the relationships between dyspnea, health conditions and frailty among community-dwelling older adults are not well established. 16 Vaz Fragoso et al., 16 for example, found that older adults with respiratory impairment had higher chances of becoming frail, and vice versa. In another study, 17 the same authors observed that poor physical-functional performance and frailty status were associated with moderate to severe exertional dyspnea in community-dwelling older adults.
Considering that the older population is growing and that dyspnea is a highly prevalent symptom in this phase of life, it would be useful to identify the factors associated with this important symptom, especially with a view to adequate healthcare management and/or prevention of negative outcomes among older adults.
In addition, in the presence of dyspnea, the person tends to stay at rest, which can cause a series of negative consequences, such as physical inactivity, deconditioning and disability, which are commonly associated with the onset and progression of frailty syndrome.

OBJECTIVE
In order to collaborate in healthcare practices for older adults, the aim of the present study was to analyze the relationships between self-reported dyspnea, health conditions and the frailty phenotype in a sample of community-dwelling older adults.

Study design and participants
This was a cross-sectional study that used the records of community-dwelling older adults who participated in the second wave

Procedures
The participants in the first wave of the FIBRA study were contacted and invited to participate in the follow-up study. At the beginning of the single session of data collection, which was conducted at the participants' homes, these older adults were informed about study procedures and voluntarily signed an informed consent statement, after agreeing to participate.
Older adults who fulfilled the inclusion and exclusion criteria completed the interview. The inclusion criteria were as follows: the participants needed to have participated in the first wave and to have remained as permanent residents in the geographical location.
Older adults with cognitive impairment suggestive of dementia, severe disability (wheelchair-bound or bedridden), motor and/or cognitive sequelae from stroke, moderate/severe Parkinson's disease, severe visual and/or hearing impairments, or decompensated diseases or terminal illness, were excluded. Considering that the number of chronic diseases (cardiovascular, respiratory, cancerous, etc.) increases with age, 19 older adults with these diseases were not excluded from the study. In addition, given that information on comorbidities was accessed through self-reporting (detailed below) and access to medical files was not feasible, it was not possible to classify the diseases reported with exactness (e.g. the type of respiratory disease reported). Current smokers were also included in the study since their prevalence in the total sample was low (i.e. 3.0%) ( Table 1).
From the original cohort, 549 older adults (≥ 65 years of age) agreed to participate in the second wave. Another 192 had died, and 543 could not located or refused to participate. Among those who completed the initial part of the interview (sociodemographic information, anthropometric measurements, physical frailty assessment and Mini-Mental State Examination, MMSE), 419 had MMSE scores above the cutoff point, adjusted for years of schooling, and participated in the second block of measurements. Out of these, 415 participants answered the question about the presence of dyspnea (i.e. the dependent variable) and thus were included in the present study.
It is important to note that a sample size ≥ 329 participants was found to be adequate for the present study, based on an estimated prevalence of dyspnea of 30.9%, 4 a 95% confidence level and an error of 0.5%.
The following variables/scales were extracted from the FIBRA study database: Health Organization (PAHO). In this way, older adults are categorized as follows: underweight with BMI ≤ 23 kg/m²; normal weight with BMI between > 23 kg/m² and < 28 kg/m²; overweight with BMI between > 28 kg/m² and < 30 kg/m²; and obese ≥ 30 kg/m². 20 • Smoking habit: Participants were considered to be current smokers if they answered "yes" to the following question: "Do you currently smoke?".
• Self-reported chronic diseases: Information was obtained from dichotomous items (yes or no) in which it was investigated whether a doctor had diagnosed heart diseases, hypertension, diabetes mellitus, depression, pulmonary diseases, osteoporosis, stroke, cancer and/or osteoarthritis at any time during the 12 months prior to the interview. Multimorbidity was considered to be the coexistence of two or more of these diseases.
• Frailty: This was assessed based on the phenotype criteria proposed by Fried et al. 15 Older adults were qualified as frail (three or more criteria), pre-frail (one or two criteria) or non-frail (no criteria), considering the following five criteria: • Self-reported unintentional weight loss: this was considered to be present if, over the last year, it was greater than or equal to 4.5 kg or 5% of body weight.
• Fatigue: this was considered to be present if the older adults answered "occasionally" or "most of the time" for either of the following two self-report items regarding how they felt during the last week: "I felt that everything I did was an effort" and "I could not get going". These items were obtained from the Center for Epidemiologic Studies Depression Scale (CESD). 21,22 • Low grip strength of the dominant hand: this was considered to be present, as measured using a Jamar dynamometer

Statistical analysis
The sample was characterized by calculating the absolute and relative frequencies and 95% confidence interval (95% CI) of the variables considered. Associations between dyspnea and the other variables were verified using Pearson's chi-square test or Fisher's exact test. After that, simple and multiple logistic regression analyses were also used to estimate crude and adjusted odds ratios. The adjusted analysis was performed independently according to age, sex, education and BMI classifications for each variable. For all analyses, the significance level was set at P < 0.05.

RESULTS
Four hundred and fifteen older people (mean age 80.3 ± 4.68 years old) answered the question about shortness of breath. The majority of the participants were very old (56.4% were aged 80 years and over), were females (69.8%), had had one to four years of education (59.7%), had normal weight (i.e. 43.8% had BMI between > 23 kg/m² and < 28 kg/m²) and were not currently smokers (97.0%). In addition, these older adults mostly had two or more self-reported diseases (69.3%), used less than five medications (62.6%) and were classified as pre-frail (64.3%) ( Table 1).
The prevalence of dyspnea in the entire sample was 21% (95% CI = 17.3 to 25.2). The frequencies and associations between dyspnea and sociodemographic variables, anthropometric data, health conditions and frailty are presented in Table 1. No significant difference between the sexes was found with regard to the prevalence of dyspnea, but the prevalence was greater (P = 0.014) in the younger age group (26.7% for 72-79 years of age) than in the older group (16.7% for 80 years and over). However, presence of dyspnea was not associated with BMI (P = 0.075) or with the smoking habit (P = 0.985).
In relation to health conditions, dyspnea was also more prevalent among participants with polypharmacy (29.8% versus 16.0%; P = 0.002) and multimorbidities (24.0% versus 10.2%; P = 0.002). Figure   1 shows the prevalence of dyspnea according to the self-reported diseases evaluated in the present study. Dyspnea was more frequent among older adults with heart diseases (31.5% versus 17.0%; P = 0.002), cancer (41.2% versus 18.1%; P = 0.001), pulmonary diseases (35.5% versus 17.6%; P = 0.004) and depression (31.3% versus 17.8%; P = 0.012). In addition, the prevalence of dyspnea was higher (P = 0.022) in older adults who were frail (35.2%) than in those classified as pre-frail (19.6%) or non-frail (17.2%) ( Table 1). The analysis on frailty components showed that dyspnea was more present in older adults who presented fatigue (34.7% versus 15.4%; P = 0.000) and gait slowness (32.8% versus 18.8%; P = 0.012) (Figure 2 Figure 3 shows the overlaps between the presences of dyspnea, multimorbidity and frailty/pre-frailty. The main overlapping detected was between multimorbidity and frailty (44.8%), followed by dyspnea versus multimorbidity versus frailty/pre-frailty (14.5%). The rates of overlapping of dyspnea with multimorbidity and frailty/pre-frailty were 3.1% and 2.5%, respectively. In addition, dyspnea only appeared separately from multimorbidity and frailty in fewer than 1% of the participants. It is important to note that 7.3% of the participants did not have any of these conditions, either alone or in combination.

DISCUSSION
In the present study, the prevalence of dyspnea among community-dwelling older adults was 21%, which was within the range reported by previous studies, even those differing in assessment method used (i.e. single question versus the Medical Research Council Dyspnea Scale). 2,4 Older adults who reported having   breathlessness were more likely to have pulmonary diseases, heart diseases, cancer, depression, multimorbidity and polypharmacy. In addition, dyspnea occurred more frequently among participants who were classified as frail (i.e. 35.2%), and fatigue was the phenotype component that was most strongly associated with dyspnea.
There is evidence in the literature that the prevalence of dyspnea increases with advancing age. 2,26,27 In a representative sample of North American older adults, 1 it was observed that one in four adults aged 70 and over experienced clinically significant breathlessness, evaluated via a single question (i.e. "How often do you become short of breath while awake?"). Greater prevalence of multimorbidities 10 and polypharmacy, 28 reduced physical fitness 3 and decreased efficiency of the respiratory system 11 make older adults more susceptible to dyspnea, which corroborate the increasing prevalence of dyspnea among older age ranges.
Although our results were in agreement with the literature in terms of the overall prevalence of dyspnea among older adults, it was surprisingly higher among those aged 72 to 79 years than among those aged 80 and over. Nonetheless, a previously published geriatric dyspnea model supports the notion of paradoxical reduction in the perception and reporting of dyspnea, explained in terms of the influence of physiological, neurological, psychological and social changes with aging. 29 Thus, it is plausible that the self-reported dyspnea rate is reduced in the older age ranges. Additionally, another study with the same cohort showed that a considerable portion of longer-living older adults (i.e. 80 years of age and over) were in good health, 30 which contradicted the results from investigations in other countries. [31][32][33] One hypothesis is that people with better health and socioeconomic conditions are more likely to reach advanced ages.
So, it is possible that the lower prevalence of dyspnea in the oldest-old group can also be partly due to this (i.e. older adults with worse conditions, including those associated with the presence of dyspnea, would die earlier). 34,35 Given that cohort studies involving very old community-dwelling adults are scarce in Brazil, the possibility for discussion of the present results is limited.
Regarding the factors associated with dyspnea in older adults, the present results partly corroborate the findings of Smith et al. 1 Among community-dwelling older adults, those authors found that individuals with respiratory diseases, multimorbidity, heart diseases, obesity and low educational level were more likely to have dyspnea. In addition, dyspnea was associated with depression, anxiety, fatigue, pain, dependence with regard to activities of daily living and high use of healthcare services. Interestingly, Johnson et al. 36 showed that age was inversely related to restrictive breathlessness during the last year of life. They argued that individuals who lived to old age achieved that age because they probably did not have the medical conditions associated with restrictive dyspnea, which thus supports the hypothesis raised above.
In old age, the presence of multimorbidity may be more complex due to overlapping of other conditions such as physical impairment, mental disorders, frailty and polypharmacy. 30,37,38 These conditions impose a considerable burden at an individual level, and on healthcare and social services too. 33 Previous studies demonstrating the complexity of healthcare management for older people showed that multimorbidity and polypharmacy were also associated with dyspnea, 1,28  and moderate-to-severe dyspnea, respectively. In addition to associations between dyspnea and some chronic self-reported diseases, multimorbidity and polypharmacy, it was also observed in the present study that dyspnea could occur in combination with multimorbidity and/or frailty/pre-frailty conditions. These, in turn, may bring more complexity to healthcare for community-dwelling older adults.
The relationship between respiratory impairment/diseases and frailty among older adults has been demonstrated in different studies. 13,16,39 In a meta-analysis, Marengoni et al. 39 found that older adults with COPD had twofold increased odds for presenting frailty, compared with those without COPD. Among community-dwelling older adults, Vaz Fragoso et al. 16 showed that respiratory impairment was associated with frailty at the baseline and that there was an increased likelihood of developing frailty features after three years of follow-up. Although studies involving COPD patients have supported the existence of a relationship between dyspnea and frailty, 13,39,40 it is not known whether this is true for community-dwelling older adults.
In the present study, older adults who reported dyspnea were more likely to be classified as frail, even after adjustment for age,   In summary, it was observed that a considerable proportion of the community-dwelling older adults of our study reported shortness of breath. Unlike other age groups, this symptom may be related not only to specific conditions (such as respiratory and cardiovascular diseases) but also to the presence of multimorbidities, polypharmacy and frailty. Thus, self-reported dyspnea should be considered to be an important symptom among older adults and should not be attributed to age per se. In addition to single-item questionnaires, it is important that shortness of breath should be investigated in a broader manner in this population, especially taking into account its impact on activities of daily living. However, it needs to be borne in mind that the scales most used for assessing dyspnea were developed before publication of the Consensus of the American Thoracic Society. 3 Thus, these scales cannot cover the entirety of the definition of dyspnea in terms of functional, sensory-perceptual and affective aspects.
In this way, multidimensional scales to evaluate dyspnea should be also considered when assessing the geriatric population. It is noteworthy that, in addition to adequate clinical management of chronic diseases and usage of medications, care for older adults with dyspnea should be carried out in an interdisciplinary and integrated manner, as much as possible.
The present study had some limitations that should be addressed.
First of all, the cross-sectional nature of the study did not allow us to draw causal associations between dyspnea, health conditions and frailty. Therefore, further longitudinal studies should include objective respiratory measurements that would help clarify such relationships and the effect of respiratory impairment on the progression of frailty, especially among community-dwelling older adults. Second, as the present sample was not representative of the community-dwelling older population and only included individuals aged 72 years and over, the results need to be interpreted with caution. Third, as the presence of dyspnea was evaluated just by a single question, the impact of its severity, commonly assessed using the Medical Research Council Dyspnea Scale, was unknown in our sample. In addition, methodological bias cannot be ruled out, given that different approaches were used to measure dyspnea (a single screening question) and frailty (a well-established tool). However, it is important to note that no reliable and valid dyspnea scales for Brazilian community-dwelling older adults were available at the time of design and application of the research protocol.

CONCLUSIONS
Among community-dwelling older adults, dyspnea was associated with different diseases, multimorbidities, polypharmacy and frailty, even after controlling for age, sex, schooling and body mass index. Among those reporting dyspnea, fatigue and gait slowness were more likely to be present. Recognizing the factors associated with reports of dyspnea may contribute to prevention of negative outcomes and to management of frailty in this population.