G Model MAT-6186; No. of Pages 6
ARTICLE IN PRESS Maturitas xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study Marisa de la Rica-Escuín, Julia González-Vaca, Rosana Varela-Pérez, María Dolores Arjonilla-García, Marta Silva-Iglesias, José Luis Oliver-Carbonell, Pedro Abizanda ∗ Geriatrics Department, Complejo Hospitalario Universitario de Albacete, Albacete, Spain
a r t i c l e
i n f o
Article history: Received 26 February 2014 Received in revised form 17 May 2014 Accepted 23 May 2014 Available online xxx Keywords: Frail elderly Disability Mortality Nursing home Institutionalization
a b s t r a c t Background: Little is known about frailty in institutionalized older adults, and there are few longitudinal studies on this topic. Objectives: To determine the association between frailty and mortality or incident disability in basic activities of daily living (BADL) in institutionalized Spanish older adults. Design: Concurrent cohort study. ˜ de Balboa and Paseo de la Cuba, in Albacete, Spain. Setting: Two nursing homes, Vasco Núnez Participants: Of the 324 institutionalized adults older than 65 years enrolled at baseline, 21 (5.5%) were lost during the one-year follow-up. Of the 303 remaining, 63 (20.8%) died, 91 (30.0%) developed incident disability, and 140 (49.2%) were free of both events. 16 participants were not suitable for analysis due to incomplete data. Measurements: Frailty was defined by the presence of three or more Fried criteria: unintentional weight loss, low energy, exhaustion, slowness, and low physical activity. Incident disability in BADL was considered when new onset disability in bathing, grooming, toileting, dressing, eating or transferring was detected with the Barthel index. Logistic regression models were constructed adjusted for age, sex, body mass index (BMI), previous Barthel index and Minimental State Examination (MMSE), and high comorbidity (Charlson index ≥3). Results: 287 participants with valid data. Mean age 84.2 (SD 6.8), with 187 (65.2%) women. 199 (69.3%) were frail, and 72 (25.1%) had high comorbidity. Mean BMI 27.6 (SD 5.2), Barthel index 53.4 (SD 37.1), and MMSE 14.2 (SD 9.7). At follow-up, 43 (21.6%) frail participants and 15 (17.0%) non-frail ones died. 73 (46.8%) frail participants and 16 (21.9%) non-frail ones developed incident disability in BADL (p < 0.001). Frailty was associated with incident disability or mortality (OR 3.3; 95% CI 1.7–6.6) adjusted for all study covariables. Conclusion: In a cohort of institutionalized older adults, frailty was associated with mortality or incident disability in BADL. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Physical frailty is a medical syndrome with multiple causes and contributors that is characterized by diminished strength, endurance, and reduced physiologic function that increases an individual’s vulnerability for developing increased dependency and/or death [1,2]. Different cohort studies have found prevalences
∗ Corresponding author at: Geriatrics Department, Complejo Hospitalario Universitario de Albacete, C/Seminario 4, 02006 Albacete, Spain. Tel.: +34 967597651; fax: +34 967597635. E-mail address: [email protected] (P. Abizanda).
between 4% and 59.1% in different settings and countries [3,4], but most of them have only included community subjects, excluding older adults at institutions. Only three studies in Spain [5–7], two in Canada [8–10], and one in Poland [11] have included institutionalized older adults, with frailty prevalence between of 29.2% and 53.7%. Frailty is an important predictor of adverse outcomes in older adults, such as death, institutionalization, falls, mobility decline, increased disability in basic (BADL) and instrumental (IADL) activities of daily living and hospitalization [12–15]. However, it is not yet well known if frailty is a valid construct for institutionalized older adults and if the pattern of association between frailty and health geriatric adverse outcomes is similar in institutionalized
http://dx.doi.org/10.1016/j.maturitas.2014.05.022 0378-5122/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model MAT-6186; No. of Pages 6
ARTICLE IN PRESS M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
2
and community subjects, because all these associations have been described in community populations. Until now, only two longitudinal studies in Canada [8,10] and one in Spain [5] have analyzed the association between frailty and health outcomes in institutionalized or assisted living older adults. Institutionalized older adults are a heterogeneous population in disability rates, multimorbidity, quality of life and vulnerability. Interventions on this population should be individualized, and it is still not well known if the detection and treatment of frailty could be of use to prevent disability, mobility decline, falls and mortality [9]. Different studies have demonstrated that clinical interventions can be effective in treating or preventing frailty [16], although any of them have been conducted in institutions. However, longitudinal studies addressing the association between frailty and adverse events are needed in institutionalized subjects, before conducting clinical trials with valid interventions as exercise or nutrition support in this population. Due to this lack of knowledge, we designed the FINAL Study to analyze the association between frailty and mortality or incident disability in a cohort of institutionalized Spanish older adults. 2. Methods 2.1. Design Concurrent cohort study in subjects older than 65 years, residents in two nursing homes from Albacete city, Spain. 2.2. Objective The objective of this study was to analyze the association between frailty and mortality or incident disability in BADL. 2.3. Study subjects Men and women older than 65 years, institutionalized in Vasco Nú˜ nez de Balboa or Paseo de la Cuba, both public nursing homes in Albacete city, Spain. Vasco Nú˜ nez de Balboa and Paseo de la Cuba have 227 and 213 residents, respectively, with different degrees of disability. Each institution has a day center and nursing beds for clinical stabilization of acute diseases. The multidisciplinary team in both centers is composed of one geriatrician, one general practitioner, nurses, social worker, physiotherapists, and occupational therapists. Medication is centrally controlled at the Complejo Hospitalario Universitario de Albacete, with the involvement of pharmacists and nutritionists. Participants had to live in the nursing home at the beginning of the study, and have to sign informed consent previous to the inclusion. In non-capable subjects, the legal tutor was informed and had to sign the informed consent. The only exclusion criteria were the refusal to participate or sign the informed consent.
and cut-offs. Low physical activity level, determined by calculating the number of kilocalories expended weekly from information given by the patient using the Calcumed® instrument, within the lowest quintile for each gender, with Fried’s original cut-off points. To construct the frailty phenotype variable, participants had to have valid values in at least 3 of the 5 criteria. Subjects were considered frail if three or more criteria were present and pre-frail if 1 or 2 were present.
2.5. Study covariables Age, gender, and body mass index (BMI) in kg/m2 were determined in the basal visit, and chronic diseases were identified from the medical records of participants. Diseases were codified following the CIE-10 classification, and grouped in homogeneous groups for analysis. Comorbidity was analyzed with the Charlson index [18]. This index contemplates 17 categories of comorbidity recorded via anamnesis, the review of patients’ clinical histories or both. Each category has a weighting based on the risk of mortality within one year. The score for each patient was obtained by adding the weighting of each of the comorbid conditions contemplated in the index. High comorbidity was considered when Charlson index score was equal or greater than 3 points, using the cut-off points validated in previous studies [19,20]. Basal disability was determined with the Barthel index that assesses the ability to independently realize 10 BADL: eating, bathing, dressing, grooming, toileting, urinary and fecal continence, transferring, walking and climbing stairs. Barthel index scores range from 0 (total disability in all 10 activities) to 100 (no disability) [21]. The Barthel index is proposed as the standard for measuring disability in BADL, for clinical and research purposes, due to its validity, reliability, sensitivity, and utility [22]. Cognitive status was determined with the Folstein’s Minimental State Examination (MMSE).
2.6. Outcome The main outcome variable was the presence of incident disability in any BADL or mortality at one-year follow-up. Mortality was obtained from the institution medical records. For the purpose of incident disability, only eating, bathing, grooming, toileting, dressing, and transferring were determined. Incident disability was considered when either new cases of disability were detected, or when Barthel scores on any of these 6 activities were lower in the follow-up visit than those in the basal visit. Urinary and fecal continence were not considered because of their high prevalence in institutions, and because medical and social factors could be involved in their presentation. Walking and climbing stairs were also not considered because they represent mobility and endurance, but not disability in BADL, and also could interact with the frailty construct in the statistical analysis.
2.4. Frailty criteria 2.7. Information sources We used the Fried frailty criteria [17], with some slight modifications. Unintentional weight loss equal to or greater than 4.600 kg or equal to or greater than 5% of body weight in the last year. Weakness as measured by grip strength, using a JAMAR® digital hand dynamometer, in the lowest 20%, adjusted for gender and BMI, according to the Fried’s original data and cut-offs. Poor energy and endurance, as indicated by self-reported exhaustion determined by two questions from the Center of Epidemiologic Studies Depression Scale (CES-D), according to Fried’s criteria. Slowness, measured as the time taken to walk 4.0 m, within the lowest 20th percentile and adjusted for gender and height, according to Fried’s original data
After the informed consent sign, information was collected through a single, one-to-one interview with the participant at the institution. Five trained geriatric nurses conducted the interviews. The information was provided by the participant him/herself or by the legal tutor if the participant was unable to do so. The performance tests were conducted on the same day as the interview by the same nurses. The information on the participants’ chronic diseases was collected from the institution medical records and nurse book. Data were anonymized, codified and included in a database for further analysis.
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model MAT-6186; No. of Pages 6
ARTICLE IN PRESS M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
In the one-year follow-up, medical records and nurse book were reviewed again, and another interview was conducted by the same five nurses to determine incident disability in BADL. 2.8. Ethical aspects This study complies with the Declaration of Helsinki and with the Organic Personal Data Protection Spanish Law 15/1999. The study was approved by the Institutional Review Board of the Albacete Health Area and the Clinical Research Committee of the Complejo Hospitalario Universitario de Albacete. All participants or legal tutors gave their written signed consent before being included in the study. When the study team found a new clinical condition, this was put in consideration of the institution geriatrician.
3
Residents from both nursing homes n=440
Agreed to parcipate
Refuse to parcipate
n=331 (75.2%)
n=109 (24.8%)
Valid frailty criteria n=324 (97.9%)
Lost at follow-up
Connue
n=21 (5.5%)
n=303 (94.5%)
2.9. Statistical analysis A descriptive analysis of the subjects’ characteristics was performed using proportions and measures of central tendency and dispersion according to the nature of the variables. Non-frail and prefrail participants were considered together as one group. A bivariate analysis was performed using the Chi-squared (when frequency of cells was less than five, Fisher’s exact test was used) to determine the association of the different control variables with the state of frailty, and between this state and the presence of incident disability in BADL or mortality. Last, we underwent a multivariant analysis with logistic regression models to describe the adjusted association between frailty and incident disability in BADL or mortality. In the models we included age, sex, MMSE score, Barthel index score, BMI and the presence of high comorbidity. We also realized logistic regression analysis to determine the independent association between every frailty component and incident disability in BADL or mortality. All data were stored and analyzed using the SPSS 17.0 software program. 3. Results Fig. 1 presents the study flow chart. Of the 440 residents from both nursing homes, 331 (75.2%) agreed to participate and 109 (24.8%) refused. Of these 109, in 88 cases the residents did not give consent to participate and in 21 the legal tutors were responsible. Of the 331 participants, 324 (97.8%) had 3 or more valid Fried’s criteria to determine frailty status, and were considered the study population. There were no significant differences between respondents and non-respondents, and between those who continued in the study or those lost at follow-up in age, sex, Barthel index and Charlson index. 21 participants were lost during follow-up, and in 16 cases data were not valid for analysis (Fig. 1). From the 287 remaining, 199 residents (69.3%) were frail, 81 (28.2%) prefrail, and 7 (2.4%) robust. 268 (93.4%) participants met the low strength criteria, 207 (72.1%) the low gait speed criteria, 209 (72.8%) the low physical activity criteria, 77 (26.8%) de exhaustion criteria, and 39 (13.6%) the weight loss criteria. Table 1 presents the basal characteristics of the complete sample, and in frail and non-frail participants. Frailty was associated with older age, female gender, and lower scores in both Barthel index and MMSE. Frail participants presented more frequently higher comorbidity, without reaching statistical significance (p = 0.07). Table 2 shows data of incident disability in BADL and mortality in frail and non-frail participants, and also for every frailty criteria. Frail participants presented incident disability and the combined event more frequently than non-frail ones. This was especially relevant for toileting, eating and transferring. Slowness, low physical
Exitus
Alive
n=63 (20.8%)
n=240 (79.2%)
Incident disability
No incident disability
n=91 (30.0%)
n=149 (49.2%)
Valid data for analysis
Valid data for analysis
Valid data for analysis
n=58
n=89
n=140
Fig. 1. Study flow-chart.
activity and exhaustion were the only frailty criteria associated with incident disability or mortality. Table 3 presents data of multivariate models between frailty, each frailty criteria, and the adverse events recorded. Frail participants had an adjusted higher risk of incident disability (OR 4.8; 95% CI 2.2–10.5) and the combined event of incident disability or mortality (OR 3.3; 95% CI 1.7–6.6). Previous disability was an interaction factor between frailty and mortality or incident disability. Of the different frailty criteria, only slowness and physical activity were associated with incident disability or the combined event. Participants with slow gait speed had an adjusted higher risk of incident disability (OR 5.2; 95% CI 2.3–11.6) and the combined event of incident disability or mortality (OR 3.9; 95% CI 2.0–7.8), and those with low physical activity had an adjusted higher risk of incident disability (OR 3.3; 95% CI 1.5–7.3) and the combined event of incident disability or mortality (OR 3.4; 95% CI 1.7–6.8). 4. Discussion Our main result is that frailty is independently associated with incident disability in BADL or mortality in institutionalized older adults. Moreover, slowness and low physical activity are the only frailty criteria independently associated with these adverse events. The greater relationship between frailty criteria with incident disability than with mortality may be a temporary matter, and it is possible that with a longer follow-up period both associations would tend to match. Since LP Fried described the frailty phenotype in 2001 [17], it has been analyzed in different international cohorts [3,15]. In Spain, five longitudinal cohorts have studied the prevalence and attributes ˜ cohort [23], The of frailty, the FRADEA Study [5], the Penagrande Toledo Study (ETES) [24], the Cuenca Study [6], and the FRALLE Study [25]. However, only the FRADEA Study in Spain, and the Canadian Study of Health and Aging (CSHA) [8] and the Alberta Continuing Care Epidemiological Studies (ACCES) [10], both in Canada,
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model
ARTICLE IN PRESS
MAT-6186; No. of Pages 6
M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
4 Table 1 Basal characteristics.
Age Gender Male Female Body mass index (kg/m2 ) Barthel index Disability in: - Bathing - Grooming - Eating - Dressing - Toleting - Transferring MMSE Charlson index High comorbidity Gait speed (m/s) Grip strength (kg) Frailty criteria Weight loss Slowness Low strength Low physical activity Exhaustion
Complete sample (n = 287)
Frail (n = 199)
Non-frail (n = 88)
p value
84.2 (6.8)
85.7 (6.3)†
82.3 (7.5)†
=0.003
100 (34.8) 187 (65.2) 27.6 (5.2) 53.4 (37.1)
46 (46.0) ‡ 153 (81.8) ‡ 27.4 (5.4) 38.6 (34.3) ‡
54 (54.0) ‡ 34 (18.2) ‡ 28.1 (4.6) 86.7 (15.1) ‡
ARTICLE IN PRESS Maturitas xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study Marisa de la Rica-Escuín, Julia González-Vaca, Rosana Varela-Pérez, María Dolores Arjonilla-García, Marta Silva-Iglesias, José Luis Oliver-Carbonell, Pedro Abizanda ∗ Geriatrics Department, Complejo Hospitalario Universitario de Albacete, Albacete, Spain
a r t i c l e
i n f o
Article history: Received 26 February 2014 Received in revised form 17 May 2014 Accepted 23 May 2014 Available online xxx Keywords: Frail elderly Disability Mortality Nursing home Institutionalization
a b s t r a c t Background: Little is known about frailty in institutionalized older adults, and there are few longitudinal studies on this topic. Objectives: To determine the association between frailty and mortality or incident disability in basic activities of daily living (BADL) in institutionalized Spanish older adults. Design: Concurrent cohort study. ˜ de Balboa and Paseo de la Cuba, in Albacete, Spain. Setting: Two nursing homes, Vasco Núnez Participants: Of the 324 institutionalized adults older than 65 years enrolled at baseline, 21 (5.5%) were lost during the one-year follow-up. Of the 303 remaining, 63 (20.8%) died, 91 (30.0%) developed incident disability, and 140 (49.2%) were free of both events. 16 participants were not suitable for analysis due to incomplete data. Measurements: Frailty was defined by the presence of three or more Fried criteria: unintentional weight loss, low energy, exhaustion, slowness, and low physical activity. Incident disability in BADL was considered when new onset disability in bathing, grooming, toileting, dressing, eating or transferring was detected with the Barthel index. Logistic regression models were constructed adjusted for age, sex, body mass index (BMI), previous Barthel index and Minimental State Examination (MMSE), and high comorbidity (Charlson index ≥3). Results: 287 participants with valid data. Mean age 84.2 (SD 6.8), with 187 (65.2%) women. 199 (69.3%) were frail, and 72 (25.1%) had high comorbidity. Mean BMI 27.6 (SD 5.2), Barthel index 53.4 (SD 37.1), and MMSE 14.2 (SD 9.7). At follow-up, 43 (21.6%) frail participants and 15 (17.0%) non-frail ones died. 73 (46.8%) frail participants and 16 (21.9%) non-frail ones developed incident disability in BADL (p < 0.001). Frailty was associated with incident disability or mortality (OR 3.3; 95% CI 1.7–6.6) adjusted for all study covariables. Conclusion: In a cohort of institutionalized older adults, frailty was associated with mortality or incident disability in BADL. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Physical frailty is a medical syndrome with multiple causes and contributors that is characterized by diminished strength, endurance, and reduced physiologic function that increases an individual’s vulnerability for developing increased dependency and/or death [1,2]. Different cohort studies have found prevalences
∗ Corresponding author at: Geriatrics Department, Complejo Hospitalario Universitario de Albacete, C/Seminario 4, 02006 Albacete, Spain. Tel.: +34 967597651; fax: +34 967597635. E-mail address: [email protected] (P. Abizanda).
between 4% and 59.1% in different settings and countries [3,4], but most of them have only included community subjects, excluding older adults at institutions. Only three studies in Spain [5–7], two in Canada [8–10], and one in Poland [11] have included institutionalized older adults, with frailty prevalence between of 29.2% and 53.7%. Frailty is an important predictor of adverse outcomes in older adults, such as death, institutionalization, falls, mobility decline, increased disability in basic (BADL) and instrumental (IADL) activities of daily living and hospitalization [12–15]. However, it is not yet well known if frailty is a valid construct for institutionalized older adults and if the pattern of association between frailty and health geriatric adverse outcomes is similar in institutionalized
http://dx.doi.org/10.1016/j.maturitas.2014.05.022 0378-5122/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model MAT-6186; No. of Pages 6
ARTICLE IN PRESS M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
2
and community subjects, because all these associations have been described in community populations. Until now, only two longitudinal studies in Canada [8,10] and one in Spain [5] have analyzed the association between frailty and health outcomes in institutionalized or assisted living older adults. Institutionalized older adults are a heterogeneous population in disability rates, multimorbidity, quality of life and vulnerability. Interventions on this population should be individualized, and it is still not well known if the detection and treatment of frailty could be of use to prevent disability, mobility decline, falls and mortality [9]. Different studies have demonstrated that clinical interventions can be effective in treating or preventing frailty [16], although any of them have been conducted in institutions. However, longitudinal studies addressing the association between frailty and adverse events are needed in institutionalized subjects, before conducting clinical trials with valid interventions as exercise or nutrition support in this population. Due to this lack of knowledge, we designed the FINAL Study to analyze the association between frailty and mortality or incident disability in a cohort of institutionalized Spanish older adults. 2. Methods 2.1. Design Concurrent cohort study in subjects older than 65 years, residents in two nursing homes from Albacete city, Spain. 2.2. Objective The objective of this study was to analyze the association between frailty and mortality or incident disability in BADL. 2.3. Study subjects Men and women older than 65 years, institutionalized in Vasco Nú˜ nez de Balboa or Paseo de la Cuba, both public nursing homes in Albacete city, Spain. Vasco Nú˜ nez de Balboa and Paseo de la Cuba have 227 and 213 residents, respectively, with different degrees of disability. Each institution has a day center and nursing beds for clinical stabilization of acute diseases. The multidisciplinary team in both centers is composed of one geriatrician, one general practitioner, nurses, social worker, physiotherapists, and occupational therapists. Medication is centrally controlled at the Complejo Hospitalario Universitario de Albacete, with the involvement of pharmacists and nutritionists. Participants had to live in the nursing home at the beginning of the study, and have to sign informed consent previous to the inclusion. In non-capable subjects, the legal tutor was informed and had to sign the informed consent. The only exclusion criteria were the refusal to participate or sign the informed consent.
and cut-offs. Low physical activity level, determined by calculating the number of kilocalories expended weekly from information given by the patient using the Calcumed® instrument, within the lowest quintile for each gender, with Fried’s original cut-off points. To construct the frailty phenotype variable, participants had to have valid values in at least 3 of the 5 criteria. Subjects were considered frail if three or more criteria were present and pre-frail if 1 or 2 were present.
2.5. Study covariables Age, gender, and body mass index (BMI) in kg/m2 were determined in the basal visit, and chronic diseases were identified from the medical records of participants. Diseases were codified following the CIE-10 classification, and grouped in homogeneous groups for analysis. Comorbidity was analyzed with the Charlson index [18]. This index contemplates 17 categories of comorbidity recorded via anamnesis, the review of patients’ clinical histories or both. Each category has a weighting based on the risk of mortality within one year. The score for each patient was obtained by adding the weighting of each of the comorbid conditions contemplated in the index. High comorbidity was considered when Charlson index score was equal or greater than 3 points, using the cut-off points validated in previous studies [19,20]. Basal disability was determined with the Barthel index that assesses the ability to independently realize 10 BADL: eating, bathing, dressing, grooming, toileting, urinary and fecal continence, transferring, walking and climbing stairs. Barthel index scores range from 0 (total disability in all 10 activities) to 100 (no disability) [21]. The Barthel index is proposed as the standard for measuring disability in BADL, for clinical and research purposes, due to its validity, reliability, sensitivity, and utility [22]. Cognitive status was determined with the Folstein’s Minimental State Examination (MMSE).
2.6. Outcome The main outcome variable was the presence of incident disability in any BADL or mortality at one-year follow-up. Mortality was obtained from the institution medical records. For the purpose of incident disability, only eating, bathing, grooming, toileting, dressing, and transferring were determined. Incident disability was considered when either new cases of disability were detected, or when Barthel scores on any of these 6 activities were lower in the follow-up visit than those in the basal visit. Urinary and fecal continence were not considered because of their high prevalence in institutions, and because medical and social factors could be involved in their presentation. Walking and climbing stairs were also not considered because they represent mobility and endurance, but not disability in BADL, and also could interact with the frailty construct in the statistical analysis.
2.4. Frailty criteria 2.7. Information sources We used the Fried frailty criteria [17], with some slight modifications. Unintentional weight loss equal to or greater than 4.600 kg or equal to or greater than 5% of body weight in the last year. Weakness as measured by grip strength, using a JAMAR® digital hand dynamometer, in the lowest 20%, adjusted for gender and BMI, according to the Fried’s original data and cut-offs. Poor energy and endurance, as indicated by self-reported exhaustion determined by two questions from the Center of Epidemiologic Studies Depression Scale (CES-D), according to Fried’s criteria. Slowness, measured as the time taken to walk 4.0 m, within the lowest 20th percentile and adjusted for gender and height, according to Fried’s original data
After the informed consent sign, information was collected through a single, one-to-one interview with the participant at the institution. Five trained geriatric nurses conducted the interviews. The information was provided by the participant him/herself or by the legal tutor if the participant was unable to do so. The performance tests were conducted on the same day as the interview by the same nurses. The information on the participants’ chronic diseases was collected from the institution medical records and nurse book. Data were anonymized, codified and included in a database for further analysis.
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model MAT-6186; No. of Pages 6
ARTICLE IN PRESS M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
In the one-year follow-up, medical records and nurse book were reviewed again, and another interview was conducted by the same five nurses to determine incident disability in BADL. 2.8. Ethical aspects This study complies with the Declaration of Helsinki and with the Organic Personal Data Protection Spanish Law 15/1999. The study was approved by the Institutional Review Board of the Albacete Health Area and the Clinical Research Committee of the Complejo Hospitalario Universitario de Albacete. All participants or legal tutors gave their written signed consent before being included in the study. When the study team found a new clinical condition, this was put in consideration of the institution geriatrician.
3
Residents from both nursing homes n=440
Agreed to parcipate
Refuse to parcipate
n=331 (75.2%)
n=109 (24.8%)
Valid frailty criteria n=324 (97.9%)
Lost at follow-up
Connue
n=21 (5.5%)
n=303 (94.5%)
2.9. Statistical analysis A descriptive analysis of the subjects’ characteristics was performed using proportions and measures of central tendency and dispersion according to the nature of the variables. Non-frail and prefrail participants were considered together as one group. A bivariate analysis was performed using the Chi-squared (when frequency of cells was less than five, Fisher’s exact test was used) to determine the association of the different control variables with the state of frailty, and between this state and the presence of incident disability in BADL or mortality. Last, we underwent a multivariant analysis with logistic regression models to describe the adjusted association between frailty and incident disability in BADL or mortality. In the models we included age, sex, MMSE score, Barthel index score, BMI and the presence of high comorbidity. We also realized logistic regression analysis to determine the independent association between every frailty component and incident disability in BADL or mortality. All data were stored and analyzed using the SPSS 17.0 software program. 3. Results Fig. 1 presents the study flow chart. Of the 440 residents from both nursing homes, 331 (75.2%) agreed to participate and 109 (24.8%) refused. Of these 109, in 88 cases the residents did not give consent to participate and in 21 the legal tutors were responsible. Of the 331 participants, 324 (97.8%) had 3 or more valid Fried’s criteria to determine frailty status, and were considered the study population. There were no significant differences between respondents and non-respondents, and between those who continued in the study or those lost at follow-up in age, sex, Barthel index and Charlson index. 21 participants were lost during follow-up, and in 16 cases data were not valid for analysis (Fig. 1). From the 287 remaining, 199 residents (69.3%) were frail, 81 (28.2%) prefrail, and 7 (2.4%) robust. 268 (93.4%) participants met the low strength criteria, 207 (72.1%) the low gait speed criteria, 209 (72.8%) the low physical activity criteria, 77 (26.8%) de exhaustion criteria, and 39 (13.6%) the weight loss criteria. Table 1 presents the basal characteristics of the complete sample, and in frail and non-frail participants. Frailty was associated with older age, female gender, and lower scores in both Barthel index and MMSE. Frail participants presented more frequently higher comorbidity, without reaching statistical significance (p = 0.07). Table 2 shows data of incident disability in BADL and mortality in frail and non-frail participants, and also for every frailty criteria. Frail participants presented incident disability and the combined event more frequently than non-frail ones. This was especially relevant for toileting, eating and transferring. Slowness, low physical
Exitus
Alive
n=63 (20.8%)
n=240 (79.2%)
Incident disability
No incident disability
n=91 (30.0%)
n=149 (49.2%)
Valid data for analysis
Valid data for analysis
Valid data for analysis
n=58
n=89
n=140
Fig. 1. Study flow-chart.
activity and exhaustion were the only frailty criteria associated with incident disability or mortality. Table 3 presents data of multivariate models between frailty, each frailty criteria, and the adverse events recorded. Frail participants had an adjusted higher risk of incident disability (OR 4.8; 95% CI 2.2–10.5) and the combined event of incident disability or mortality (OR 3.3; 95% CI 1.7–6.6). Previous disability was an interaction factor between frailty and mortality or incident disability. Of the different frailty criteria, only slowness and physical activity were associated with incident disability or the combined event. Participants with slow gait speed had an adjusted higher risk of incident disability (OR 5.2; 95% CI 2.3–11.6) and the combined event of incident disability or mortality (OR 3.9; 95% CI 2.0–7.8), and those with low physical activity had an adjusted higher risk of incident disability (OR 3.3; 95% CI 1.5–7.3) and the combined event of incident disability or mortality (OR 3.4; 95% CI 1.7–6.8). 4. Discussion Our main result is that frailty is independently associated with incident disability in BADL or mortality in institutionalized older adults. Moreover, slowness and low physical activity are the only frailty criteria independently associated with these adverse events. The greater relationship between frailty criteria with incident disability than with mortality may be a temporary matter, and it is possible that with a longer follow-up period both associations would tend to match. Since LP Fried described the frailty phenotype in 2001 [17], it has been analyzed in different international cohorts [3,15]. In Spain, five longitudinal cohorts have studied the prevalence and attributes ˜ cohort [23], The of frailty, the FRADEA Study [5], the Penagrande Toledo Study (ETES) [24], the Cuenca Study [6], and the FRALLE Study [25]. However, only the FRADEA Study in Spain, and the Canadian Study of Health and Aging (CSHA) [8] and the Alberta Continuing Care Epidemiological Studies (ACCES) [10], both in Canada,
Please cite this article in press as: de la Rica-Escuín M, et al. Frailty and mortality or incident disability in institutionalized older adults: The FINAL Study. Maturitas (2014), http://dx.doi.org/10.1016/j.maturitas.2014.05.022
G Model
ARTICLE IN PRESS
MAT-6186; No. of Pages 6
M. de la Rica-Escuín et al. / Maturitas xxx (2014) xxx–xxx
4 Table 1 Basal characteristics.
Age Gender Male Female Body mass index (kg/m2 ) Barthel index Disability in: - Bathing - Grooming - Eating - Dressing - Toleting - Transferring MMSE Charlson index High comorbidity Gait speed (m/s) Grip strength (kg) Frailty criteria Weight loss Slowness Low strength Low physical activity Exhaustion
Complete sample (n = 287)
Frail (n = 199)
Non-frail (n = 88)
p value
84.2 (6.8)
85.7 (6.3)†
82.3 (7.5)†
=0.003
100 (34.8) 187 (65.2) 27.6 (5.2) 53.4 (37.1)
46 (46.0) ‡ 153 (81.8) ‡ 27.4 (5.4) 38.6 (34.3) ‡
54 (54.0) ‡ 34 (18.2) ‡ 28.1 (4.6) 86.7 (15.1) ‡
