CLINICAL INVESTIGATIONS
Muscle Strength and Physical Performance as Predictors of Mortality, Hospitalization, and Disability in the Oldest Old Delphine Legrand, MSc,* Bert Vaes, PhD,*† Catharina Mathe€ı, PhD,*† Wim Adriaensen, MSc,*† Gijs Van Pottelbergh, PhD,*† and Jean-Marie Degryse, PhD*†
OBJECTIVES: To evaluate the predictive value of muscle strength and physical performance in the oldest old for allcause mortality; hospitalization; and the onset of disability, defined as a decline in activities of daily living (ADLs), independent of muscle mass, inflammatory markers, and comorbidities. DESIGN: A prospective, observational, population-based follow-up study. SETTING: Three well-circumscribed areas of Belgium. PARTICIPANTS: Five hundred sixty participants aged 80 and older were followed for 33.5 months (interquartile range 31.1–35.6 months). MEASUREMENTS: Grip strength, Short Physical Performance Battery (SPPB) score, and muscle mass were measured at baseline; ADLs at baseline and after 20 months; and all-cause mortality and time to first hospitalization from inclusion onward. Kaplan-Meier curves and Cox proportional hazards models were calculated for all-cause mortality and hospitalization. Logistic regression analysis was used to determine predictors of decline in ADLs. RESULTS: Kaplan–Meier curves showed significantly higher all-cause mortality and hospitalization in subjects in the lowest tertile of grip strength and SPPB score. The adjusted Cox proportional hazards model showed that participants with high grip strength or a high SPPB score had a lower risk of mortality and hospitalization, independent of muscle mass, inflammatory markers, and comorbidity. A relationship was found between SPPB score and decline in ADLs, independent of muscle mass, inflammation, and comorbidity. CONCLUSION: In people aged 80 and older, physical performance is a strong predictor of mortality, hospitalization, and disability, and muscle strength is a strong predictor of mortality and hospitalization. All of these
From the*Institut de Recherche Sante et Societe, Universite Catholique de Louvain, Brussels; and †Department of Public Health and Primary Care, Katholieke Universiteit Leuven, Leuven, Belgium. Address correspondence to Delphine Legrand, Clos Chapelle-aux-Champs, 30 bte 30.05, 1200 Woluwe-Saint-Lambert, UCL, Brussels, Belgium. E-mail: [email protected] DOI: 10.1111/jgs.12840
JAGS 2014 © 2014, Copyright the Authors Journal compilation © 2014, The American Geriatrics Society
relationships were independent of muscle mass, inflammatory markers, and comorbidity. J Am Geriatr Soc 2014.
Key words: muscle strength; physical performance; inflammatory markers; functional decline; mortality
A
ging is associated with a loss of muscle mass and strength and with functional decline.1–3 The main consequence of the loss of muscle mass and strength is the limitation of physical performance in older people, which increases the risk of falls, hospitalizations, dependency, disability, frailty, and mortality.3–6 Moreover, it has been suggested that in very old subjects, muscle strength and physical performance capacity may be more relevant indicators of sarcopenia than the muscle mass index.7–11 Previous studies have shown that grip strength and physical performance are predictors of mortality,12–16 hospitalization,17 and disability;12,18–21 higher levels of serum inflammatory markers have been found to have a strong correlation with lower muscle strength and mass;22–25 and higher circulating levels of interleukin (IL)-6, tumor necrosis factor alpha (TNF-a), and C-reactive protein (CRP) have been shown to be associated with mortality and disability in older people.23–31 Although several studies have documented the predictive value of muscle strength and physical performance for mortality, hospitalization, and disability and investigated the association with covariates such as muscle mass and inflammatory markers, no single study has been conducted that encompasses all outcomes and all important confounders. Furthermore, no previous study investigated these relationships in a representative population-based sample of subjects aged 80 and older that was able to cover the full spectre of heterogeneity in this age segment. Therefore, the predictive value of muscle strength and physical performance for all-cause mortality, hospitalization, and onset of disability, defined as a decline in activities of daily living (ADLs), independent of muscle mass, inflammation, and comorbidity, was evaluated in the BELFRAIL study (BFC80+), a large population-based cohort of subjects aged 80 and older.
0002-8614/14/$15.00
2
LEGRAND ET AL.
METHODS Study Population The BFC80+ was designed as a prospective, observational, population-based cohort study to evaluate subjects aged 80 and older living in Belgium. All participants gave informed consent, and the biomedical ethics committee of the Medical School of the Universite Catholique de Louvain of Brussels approved the study. The study protocol and sampling methods have been previously described in detail.32 Briefly, between November 2008 and September 2009, 567 individuals were included in the BFC80+.32 Only three exclusion criteria (known severe dementia, palliative situations, and medical urgency) were used. At baseline, a general practitioner (GP) recorded sociodemographic data and medical history, and a clinical research assistant (CRA) performed an extensive examination that included performance testing, questionnaires, and technical examinations. A blood sample was collected in the morning after fasting.
Clinical Characteristics The GP recorded morbidities, including arthritis, osteoarthritis, osteoporosis, chronic obstructive pulmonary disease (COPD), cancer, knee and hip prostheses, renal insufficiency, hypertension, hyperlipidemia, a history of angina pectoris or myocardial infarction, known cardiomyopathy, a history of transient ischemic attack, diabetes mellitus, cerebrovascular accident, peripheral arterial disease, a history of decompensated heart failure, atrial fibrillation, valvular disease, and a history of edema of the lower extremities. The cumulative disease count was used for further analyses.
2014
JAGS
Sammons Preston Rolyan, Bolingbrook, IL). The maximum value was used in all further calculations. Sex-adjusted tertiles of the maximum value were used for further analysis. For women, the values were 15.0 kg or less, 15.1 to 20.0 kg, and 20.1 kg or more. For men, the values were 25.3 kg or less, 25.4 to 33.2 kg, and 33.3 kg or more. Sexadjusted tertiles were subsequently determined by combining the tertiles of women with the tertiles of men, leading to three levels of grip strength: low, medium, and high. The Short Physical Performance Battery (SPPB) consisted of timed measures of walking speed, rising from a chair, and maintaining balance in a tandem stand.15 Participants who could not complete a task were assigned a score of 0, and those who completed the tasks were assigned scores of 1 to 4, with scores corresponded to quartiles of time required to complete the task, with the best time receiving a score of 4.15 The total performance score was calculated by summing the scores (ranged 0–12). Sex-adjusted tertiles of the total SPPB score were used for further analysis. For women, the cutoff values were 5 or less, 6 to 8, and 9 or greater. For men, these values were 7 or less, 8 to 10, and 11 or more. Sex-adjusted tertiles were subsequently determined by combining the tertiles of women with the tertiles of men, leading to three levels of SPPB scores: low, medium, and high. Skeletal muscle mass (SMM) was estimated using a formula developed previously:33 SMM (kg) = height (0.244 9 weight) + (7.8 9 height) + (6.6 9 sex) (0.098 9 age) + (ethnicity 3.3). Sex-specific tertiles were used to classify participants based on three levels of muscle mass. For women, the cutoff values were less than 23.5 kg, 23.5 to 27.4 kg, and more than 27.4 kg. For men, these values were less than 35.8 kg, 35.8 to 41.4 kg, and more than 41.4 kg.
Outcome Measurements Serum Inflammatory Markers Serum ultrasensitive CRP (usCRP) levels were measured using a clinical chemistry analyzer system (UniCel DxC 800 Sychron; Beckman-Coulter, Brea, CA), and serum IL6 and TNF-a were assessed using Biochip arrays (Evidence Investigator analyzer, Randox Laboratories Limited, Crumlin, UK).32 Sex-specific tertiles were used to classify participants based on three levels of inflammatory biomarkers: low, medium, and high tertiles. For women, usCRP levels were 0.10 mg/dL or less, 0.11 to 0.25 mg/dL, and 0.26 mg/dL or more, respectively; IL-6 levels were 1.40 pg/mL or less, 1.41 to 2.59 pg/mL, ≥ 2.60 pg/mL or more, respectively; and TNF-a levels were 3.01 pg/mL or less, 3.02 to 4.15 pg/mL, and 4.16 pg/mL or more, respectively. For men, usCRP levels were 0.10 mg/dL or less, 0.11 to 0.36 mg/dL, and 0.37 mg/dL or more, respectively; IL-6 levels were 1.83 pg/mL or less, 1.84 to 3.50 pg/mL, and 3.51 pg/mL or more, respectively; and TNF-a levels were 3.32 pg/mL or less, 3.33 to 4.56 pg/mL, and ≥4.57 pg/mL or more, respectively.
Muscle Strength and Muscle mass Grip strength was measured three times in the dominant hand using a digital handheld dynamometer (JAMAR Plus,
Two detailed follow-up questionnaires were received from the participating GPs 1.4 0.26 years and 3.0 0.25 years after inclusion. The questionnaire included questions on mortality and hospitalization. All-cause mortality and time to first hospitalization after the baseline CRA visit were used as outcome measurements. Functional limitations were assessed by asking respondents about the degree of difficulty that they had with six ADLs: climbing stairs, walking 5 minutes outdoors without resting, sitting down in and standing up from a chair, dressing and undressing oneself, using own or public transportation, and cutting one’s own toenails. Response categories ranged from “No, I cannot” (1) to “Yes, without difficulty” (5).34 The total score was calculated by summing the scores of all activities (range 6–30). At the second visit (19.6 2.5 months after baseline), a CRA readministered the questionnaire about ADLs. A relevant decline in ADLs was determined using the Edwards-Nunnally index for men and women separately.35 This index determines the probability of substantial individual change and avoids the problem of regression to the mean. Based on the scale reliability and the 95% confidence interval (CI) of the mean score at baseline, the index computes whether a significant change has occurred between baseline and the second visit.
JAGS
2014
PREDICTIVE VALUE OF MUSCLE STRENGTH IN THE OLDEST OLD
T0: 1st CRA visit n = 560
Data Analysis Continuous data are presented as the means standard deviations or medians and interquartile ranges (IQRs). The characteristics of the study participants were compared according to their survival and hospitalization status using the Student t-test or Mann-Whitney U-test for continuous variables and the chi-square test for categorical variables. Sex-adjusted tertiles of grip strength and SPPB scores at baseline were plotted on Kaplan-Meier curves representing all-cause mortality and first hospitalization during follow-up and compared using a log-rank test. Bivariate hazard ratios (HRs) with corresponding 95% CIs were calculated using the Cox proportional hazards model. Two strategies were subsequently used to examine the associations between muscle strength and mortality or hospitalization. The first strategy used grip strength at baseline, and the second used SPPB scores at baseline to calculate the HR for all-cause mortality and hospitalization, adjusting for age, sex, body mass index (BMI), muscle mass, all individual morbidities, and inflammatory markers in three consecutive models. The relationship between muscle strength or physical performance at baseline and decline in ADLs (between baseline and the second visit) was examined using logistic regression analysis and corrected for age, sex, BMI, muscle mass, all individual morbidities, and inflammation at baseline in three consecutive models. Variables were first checked for multicollinearity. SPSS 20.0 (SPSS, Inc., Chicago, IL) was used for the data analysis, and P < .05 was considered statistically significant.
RESULTS A CRA did not observe seven of the 567 people included in the BFC80+, leaving 560 participants for analysis (351 women (62.7%), 209 men (37.3%)). Five hundred four participants lived at home (90%), and 56 were institutionalized (10%). At baseline, the women had a mean age of 84.9 3.8, a mean weight of 66.5 13.7 kg, and a median of five morbidities (IQR 3–6). The men had a mean age of 84.4 3.3, a mean weight of 75.5 11.8 kg, and a median of five morbidities (IQR 3–6). Data were available for 546 (97.5%) subjects on grip strength, 540 (96.4%) subjects on SPPB score, and 557 subjects (99.5%) on muscle mass. At baseline, information was obtained for 546 study participants (97.5%) on usCRP and on IL-6 and TNF-a levels for 414 (73.9%). In the remaining participants, cytokine markers could not be determined because of technical difficulties.
All-Cause Mortality There was no loss to follow-up for mortality. The median length of the follow-up was 33.5 months (IQR 31.1– 35.6 months). During the follow-up period, 23% (129) of the participants died (Figure 1). Participant characteristics are shown in Table 1 according to outcomes. Participants who died were significantly older and had more morbidities. Serum inflammatory marker levels were higher in subjects who died than in those who survived. Kaplan-Meier curves showed significantly higher all-cause
3
Mortality n = 72
Hospitalization n= 190
T1: eligible for 2nd CRA visit n = 488
Underwent 2nd CRA visit n = 428
Refused 2nd CRA visit n = 60
ADL decline estimated n = 417
Mortality n = 47
Hospitalization n = 94
T2 n = 431
Figure 1. Flowchart detailing the data available for survival analysis and the data available for activity of daily living (ADL) decline analysis from the BELFRAIL cohort. a 19.6 2.5 months after baseline. b33.5 2.5 months after baseline. CRA = clinical research assistant.
mortality in subjects in the lowest tertile of grip strength and SPPB scores (Figure 2). Table 1 shows that the highest tertiles of grip strength, SPPB score, and muscle mass were associated with less risk of death than the lowest tertiles. The adjusted Cox proportional hazards model confirmed that participants with high grip strength or a high SPPB score had a lower risk of mortality, independent of muscle mass, comorbidity, and inflammatory markers (Table 2). Other factors independently associated with mortality were chronic obstructive pulmonary disease (COPD) and arthritis. In the model with SPPB scores, participants with high IL-6 levels had greater risk of mortality. Muscle mass was not related to mortality in the multivariate model.
Hospitalization There was no loss to follow-up for hospitalization. During follow-up, 51% (n = 284) of participants were hospitalized (Figure 1). Participant characteristics are shown in Table 1 according to hospitalization. Participants who were hospitalized during follow-up had more morbidities than those who were not. The Kaplan–Meier curves indicated significantly more hospitalization in participants with low grip strength and low SPPB scores (Figure 2). The highest tertiles of grip strength, SPPB score, and muscle mass were associated with less risk of hospitalization than the lowest tertiles. The unadjusted Cox proportional hazards model showed a relationship between all-cause hospitalization and grip strength (HR = 0.76, 95% CI = 0.65– 0.88) and between all-cause hospitalization and SPPB level (HR = 0.74. 95% CI = 0.63–0.86). The multivariate
117 (37) 104 (32) 100 (31) 150 140 129 23
138 (33) 137 (33) 139 (34)
180 180 186 23 123 (29) 142 (34) 156 (37) 8 3.2 118 (28) 137 (32) 168 (40) 31 8.6 127 (30) 155 (36) 147 (34)
180 (33) 182 (33) 184 (34) 8 3.4
176 (33) 170 (31) 194 (36) 31 8.7
185 (33) 185 (33) 187 (34)
(36) (33) (31) 9.8
125 (39) 106 (33) 90 (28)
139 (34) 136 (32) 139 (34)
(33) (33) (34) 9.5
0.17 (0.08–0.36) 1.9 (1.2–3.3) 3.6 (2.9–4.5)
4 (3–6)
4 (3–6)
0.19 (0.08–0.41) 2.1 (1.3–3.8) 3.7 (2.9–4.8)
158 (37) 273 (63) 84.4 3.5 70 13 28 4.6
Survived, n = 431
209 (37) 351 (63) 84.7 3.7 70 14 27 4.9
All Participants
b
1 1.6 (0.93–2.8) 2.0 (1.2–3.3)a 1 2.04 (1.3–3.2)b 1.4 (0.85–2.2)
1 0.43 (0.27–0.68)b 0.62 (0.42–0.94)a
1 0.35 (0.22–0.55)b 0.53 (0.34–0.81)a
1 0.63 (0.42–0.95)a 0.49 (0.32–0.76)b
30 (24)b 40 (31)b 57 (45)b 20 8.0b 57 (46)b 40 (32)b 28 (22)b 6 3.7b 58 (50)b 33 (28)b 26 (22)b 30 9.1b 58 (45)a 31 (24)a 40 (31)a
b
21 (23)a 33 (35)a 39 (42)a
b
1 2.2 (1.2–4.2) 4.0 (2.2–7.2)
b
0.98 (0.94–1.0)
1 1.1 (0.77–1.6) 1.1 (1.0–1.1)
HR (95% CI)
14 (15) b 30 (32)b 49 (53) b
0.25 (0.11–0.75) 2.8 (1.8–3.8) b 3.9 (3.2–5.5)a
5 (3–7)
51 (40) 78 (61) 85.8 3.9b 68 16 27 5.7
Deceased, n = 129
(35) (35) (30) (10.1)
78 (29) 96 (35) 100 (36)
69 (26) 96 (36) 101 (38) 31 8.9
71 (26) 95 (35) 105 (39) 8 3.2
95 95 80 24
74 (35) 78 (36) 61 (29)
79 (37) 70 (33) 64 (30)
0.17 (0.08–0.35) 1.9 (1.2–3.2) 3.6 (2.9–4.4)
4 (3–5)
93 (34) 190 (65) 84.4 3.7 71 14 28 5.0
No Hospitalization, n = 276
(31) (31) (38) (8.8)a
107 (38)a 90 (32)a 87 (30)a
107 (39)a 74 (27)a 93 (34)a 31 8.6a
109 (40)a 87 (31)a 79 (29)a 7 3.5b
85 86 105 22
64 (32)a 59 (29)a 78 (39)
60 (30) 66 (30) 75 (40)
0.20 (0.08–0.5)a 2.4 (1.4–4.3)a 3.8 (2.9–5.2)
5 (3–7)b
116 (41) 168 (59) 85.0 3.7 69 13 27 4.6
Hospitalization, n = 284
1 0.8 (0.6–1.0)a 0.75 (0.56–0.99)a
1 0.5 (0.4–0.7)b 0.6 (0.5–0.8)b
1 0.7 (0.5–0.9)b 0.6 (0.4–0.8)b
1 1.1 (0.8–1.4) 1.4 (1.1–1.9)a
1 0.92 (0.7–1.3) 1.4 (0.97–2.1)
1 1.2 (0.82–1.6) 1.5 (1.1–2.1)a
0.98 (0.95–1.0)
1 1.2 (0.94–1.5) 1.0 (1.0–1.1)a
HR (95% CI)
LEGRAND ET AL. 2014
Muscle mass, grip strength, and Short Physical Performance Battery (SPPB) score tertiles were adjusted for sex. a P < .05; bP ≤ .001. HR = hazard ratio; CI = confidence interval; SD = standard deviation; IQR = interquartile range; IL-6 = interleukin-6; TNF-a = tumor necrosis factor alpha; usCRP = ultrasensitive C-reactive protein.
Sex, n (%) Male Female Age, meanSD Weight, kg, mean SD Body mass index, kg/m2, meanSD Number of morbidities, median (IQR) Inflammatory markers usCRP, median (IQR) IL-6, median (IQR) TNF-a, median (IQR) IL-6 tertile, n (%) Low Medium High TNF-a tertile, n (%) Low Medium High usCRP tertile, n (%) Low Medium High Grip strength, median (IQR) Grip strength tertile, n (%) Low Medium High SPPB score, median (IQR) SPPB score (tertiles) Low Medium High Muscle mass, mean SD Muscle mass tertile, n (%) Low Medium High
Characteristic
Table 1. Subject Characteristics at Baseline According Vital Status and Hospitalization
4 JAGS
JAGS
2014
PREDICTIVE VALUE OF MUSCLE STRENGTH IN THE OLDEST OLD
All-cause mortality
Hospitalization Grip strength 100
Log rank p
Muscle Strength and Physical Performance as Predictors of Mortality, Hospitalization, and Disability in the Oldest Old Delphine Legrand, MSc,* Bert Vaes, PhD,*† Catharina Mathe€ı, PhD,*† Wim Adriaensen, MSc,*† Gijs Van Pottelbergh, PhD,*† and Jean-Marie Degryse, PhD*†
OBJECTIVES: To evaluate the predictive value of muscle strength and physical performance in the oldest old for allcause mortality; hospitalization; and the onset of disability, defined as a decline in activities of daily living (ADLs), independent of muscle mass, inflammatory markers, and comorbidities. DESIGN: A prospective, observational, population-based follow-up study. SETTING: Three well-circumscribed areas of Belgium. PARTICIPANTS: Five hundred sixty participants aged 80 and older were followed for 33.5 months (interquartile range 31.1–35.6 months). MEASUREMENTS: Grip strength, Short Physical Performance Battery (SPPB) score, and muscle mass were measured at baseline; ADLs at baseline and after 20 months; and all-cause mortality and time to first hospitalization from inclusion onward. Kaplan-Meier curves and Cox proportional hazards models were calculated for all-cause mortality and hospitalization. Logistic regression analysis was used to determine predictors of decline in ADLs. RESULTS: Kaplan–Meier curves showed significantly higher all-cause mortality and hospitalization in subjects in the lowest tertile of grip strength and SPPB score. The adjusted Cox proportional hazards model showed that participants with high grip strength or a high SPPB score had a lower risk of mortality and hospitalization, independent of muscle mass, inflammatory markers, and comorbidity. A relationship was found between SPPB score and decline in ADLs, independent of muscle mass, inflammation, and comorbidity. CONCLUSION: In people aged 80 and older, physical performance is a strong predictor of mortality, hospitalization, and disability, and muscle strength is a strong predictor of mortality and hospitalization. All of these
From the*Institut de Recherche Sante et Societe, Universite Catholique de Louvain, Brussels; and †Department of Public Health and Primary Care, Katholieke Universiteit Leuven, Leuven, Belgium. Address correspondence to Delphine Legrand, Clos Chapelle-aux-Champs, 30 bte 30.05, 1200 Woluwe-Saint-Lambert, UCL, Brussels, Belgium. E-mail: [email protected] DOI: 10.1111/jgs.12840
JAGS 2014 © 2014, Copyright the Authors Journal compilation © 2014, The American Geriatrics Society
relationships were independent of muscle mass, inflammatory markers, and comorbidity. J Am Geriatr Soc 2014.
Key words: muscle strength; physical performance; inflammatory markers; functional decline; mortality
A
ging is associated with a loss of muscle mass and strength and with functional decline.1–3 The main consequence of the loss of muscle mass and strength is the limitation of physical performance in older people, which increases the risk of falls, hospitalizations, dependency, disability, frailty, and mortality.3–6 Moreover, it has been suggested that in very old subjects, muscle strength and physical performance capacity may be more relevant indicators of sarcopenia than the muscle mass index.7–11 Previous studies have shown that grip strength and physical performance are predictors of mortality,12–16 hospitalization,17 and disability;12,18–21 higher levels of serum inflammatory markers have been found to have a strong correlation with lower muscle strength and mass;22–25 and higher circulating levels of interleukin (IL)-6, tumor necrosis factor alpha (TNF-a), and C-reactive protein (CRP) have been shown to be associated with mortality and disability in older people.23–31 Although several studies have documented the predictive value of muscle strength and physical performance for mortality, hospitalization, and disability and investigated the association with covariates such as muscle mass and inflammatory markers, no single study has been conducted that encompasses all outcomes and all important confounders. Furthermore, no previous study investigated these relationships in a representative population-based sample of subjects aged 80 and older that was able to cover the full spectre of heterogeneity in this age segment. Therefore, the predictive value of muscle strength and physical performance for all-cause mortality, hospitalization, and onset of disability, defined as a decline in activities of daily living (ADLs), independent of muscle mass, inflammation, and comorbidity, was evaluated in the BELFRAIL study (BFC80+), a large population-based cohort of subjects aged 80 and older.
0002-8614/14/$15.00
2
LEGRAND ET AL.
METHODS Study Population The BFC80+ was designed as a prospective, observational, population-based cohort study to evaluate subjects aged 80 and older living in Belgium. All participants gave informed consent, and the biomedical ethics committee of the Medical School of the Universite Catholique de Louvain of Brussels approved the study. The study protocol and sampling methods have been previously described in detail.32 Briefly, between November 2008 and September 2009, 567 individuals were included in the BFC80+.32 Only three exclusion criteria (known severe dementia, palliative situations, and medical urgency) were used. At baseline, a general practitioner (GP) recorded sociodemographic data and medical history, and a clinical research assistant (CRA) performed an extensive examination that included performance testing, questionnaires, and technical examinations. A blood sample was collected in the morning after fasting.
Clinical Characteristics The GP recorded morbidities, including arthritis, osteoarthritis, osteoporosis, chronic obstructive pulmonary disease (COPD), cancer, knee and hip prostheses, renal insufficiency, hypertension, hyperlipidemia, a history of angina pectoris or myocardial infarction, known cardiomyopathy, a history of transient ischemic attack, diabetes mellitus, cerebrovascular accident, peripheral arterial disease, a history of decompensated heart failure, atrial fibrillation, valvular disease, and a history of edema of the lower extremities. The cumulative disease count was used for further analyses.
2014
JAGS
Sammons Preston Rolyan, Bolingbrook, IL). The maximum value was used in all further calculations. Sex-adjusted tertiles of the maximum value were used for further analysis. For women, the values were 15.0 kg or less, 15.1 to 20.0 kg, and 20.1 kg or more. For men, the values were 25.3 kg or less, 25.4 to 33.2 kg, and 33.3 kg or more. Sexadjusted tertiles were subsequently determined by combining the tertiles of women with the tertiles of men, leading to three levels of grip strength: low, medium, and high. The Short Physical Performance Battery (SPPB) consisted of timed measures of walking speed, rising from a chair, and maintaining balance in a tandem stand.15 Participants who could not complete a task were assigned a score of 0, and those who completed the tasks were assigned scores of 1 to 4, with scores corresponded to quartiles of time required to complete the task, with the best time receiving a score of 4.15 The total performance score was calculated by summing the scores (ranged 0–12). Sex-adjusted tertiles of the total SPPB score were used for further analysis. For women, the cutoff values were 5 or less, 6 to 8, and 9 or greater. For men, these values were 7 or less, 8 to 10, and 11 or more. Sex-adjusted tertiles were subsequently determined by combining the tertiles of women with the tertiles of men, leading to three levels of SPPB scores: low, medium, and high. Skeletal muscle mass (SMM) was estimated using a formula developed previously:33 SMM (kg) = height (0.244 9 weight) + (7.8 9 height) + (6.6 9 sex) (0.098 9 age) + (ethnicity 3.3). Sex-specific tertiles were used to classify participants based on three levels of muscle mass. For women, the cutoff values were less than 23.5 kg, 23.5 to 27.4 kg, and more than 27.4 kg. For men, these values were less than 35.8 kg, 35.8 to 41.4 kg, and more than 41.4 kg.
Outcome Measurements Serum Inflammatory Markers Serum ultrasensitive CRP (usCRP) levels were measured using a clinical chemistry analyzer system (UniCel DxC 800 Sychron; Beckman-Coulter, Brea, CA), and serum IL6 and TNF-a were assessed using Biochip arrays (Evidence Investigator analyzer, Randox Laboratories Limited, Crumlin, UK).32 Sex-specific tertiles were used to classify participants based on three levels of inflammatory biomarkers: low, medium, and high tertiles. For women, usCRP levels were 0.10 mg/dL or less, 0.11 to 0.25 mg/dL, and 0.26 mg/dL or more, respectively; IL-6 levels were 1.40 pg/mL or less, 1.41 to 2.59 pg/mL, ≥ 2.60 pg/mL or more, respectively; and TNF-a levels were 3.01 pg/mL or less, 3.02 to 4.15 pg/mL, and 4.16 pg/mL or more, respectively. For men, usCRP levels were 0.10 mg/dL or less, 0.11 to 0.36 mg/dL, and 0.37 mg/dL or more, respectively; IL-6 levels were 1.83 pg/mL or less, 1.84 to 3.50 pg/mL, and 3.51 pg/mL or more, respectively; and TNF-a levels were 3.32 pg/mL or less, 3.33 to 4.56 pg/mL, and ≥4.57 pg/mL or more, respectively.
Muscle Strength and Muscle mass Grip strength was measured three times in the dominant hand using a digital handheld dynamometer (JAMAR Plus,
Two detailed follow-up questionnaires were received from the participating GPs 1.4 0.26 years and 3.0 0.25 years after inclusion. The questionnaire included questions on mortality and hospitalization. All-cause mortality and time to first hospitalization after the baseline CRA visit were used as outcome measurements. Functional limitations were assessed by asking respondents about the degree of difficulty that they had with six ADLs: climbing stairs, walking 5 minutes outdoors without resting, sitting down in and standing up from a chair, dressing and undressing oneself, using own or public transportation, and cutting one’s own toenails. Response categories ranged from “No, I cannot” (1) to “Yes, without difficulty” (5).34 The total score was calculated by summing the scores of all activities (range 6–30). At the second visit (19.6 2.5 months after baseline), a CRA readministered the questionnaire about ADLs. A relevant decline in ADLs was determined using the Edwards-Nunnally index for men and women separately.35 This index determines the probability of substantial individual change and avoids the problem of regression to the mean. Based on the scale reliability and the 95% confidence interval (CI) of the mean score at baseline, the index computes whether a significant change has occurred between baseline and the second visit.
JAGS
2014
PREDICTIVE VALUE OF MUSCLE STRENGTH IN THE OLDEST OLD
T0: 1st CRA visit n = 560
Data Analysis Continuous data are presented as the means standard deviations or medians and interquartile ranges (IQRs). The characteristics of the study participants were compared according to their survival and hospitalization status using the Student t-test or Mann-Whitney U-test for continuous variables and the chi-square test for categorical variables. Sex-adjusted tertiles of grip strength and SPPB scores at baseline were plotted on Kaplan-Meier curves representing all-cause mortality and first hospitalization during follow-up and compared using a log-rank test. Bivariate hazard ratios (HRs) with corresponding 95% CIs were calculated using the Cox proportional hazards model. Two strategies were subsequently used to examine the associations between muscle strength and mortality or hospitalization. The first strategy used grip strength at baseline, and the second used SPPB scores at baseline to calculate the HR for all-cause mortality and hospitalization, adjusting for age, sex, body mass index (BMI), muscle mass, all individual morbidities, and inflammatory markers in three consecutive models. The relationship between muscle strength or physical performance at baseline and decline in ADLs (between baseline and the second visit) was examined using logistic regression analysis and corrected for age, sex, BMI, muscle mass, all individual morbidities, and inflammation at baseline in three consecutive models. Variables were first checked for multicollinearity. SPSS 20.0 (SPSS, Inc., Chicago, IL) was used for the data analysis, and P < .05 was considered statistically significant.
RESULTS A CRA did not observe seven of the 567 people included in the BFC80+, leaving 560 participants for analysis (351 women (62.7%), 209 men (37.3%)). Five hundred four participants lived at home (90%), and 56 were institutionalized (10%). At baseline, the women had a mean age of 84.9 3.8, a mean weight of 66.5 13.7 kg, and a median of five morbidities (IQR 3–6). The men had a mean age of 84.4 3.3, a mean weight of 75.5 11.8 kg, and a median of five morbidities (IQR 3–6). Data were available for 546 (97.5%) subjects on grip strength, 540 (96.4%) subjects on SPPB score, and 557 subjects (99.5%) on muscle mass. At baseline, information was obtained for 546 study participants (97.5%) on usCRP and on IL-6 and TNF-a levels for 414 (73.9%). In the remaining participants, cytokine markers could not be determined because of technical difficulties.
All-Cause Mortality There was no loss to follow-up for mortality. The median length of the follow-up was 33.5 months (IQR 31.1– 35.6 months). During the follow-up period, 23% (129) of the participants died (Figure 1). Participant characteristics are shown in Table 1 according to outcomes. Participants who died were significantly older and had more morbidities. Serum inflammatory marker levels were higher in subjects who died than in those who survived. Kaplan-Meier curves showed significantly higher all-cause
3
Mortality n = 72
Hospitalization n= 190
T1: eligible for 2nd CRA visit n = 488
Underwent 2nd CRA visit n = 428
Refused 2nd CRA visit n = 60
ADL decline estimated n = 417
Mortality n = 47
Hospitalization n = 94
T2 n = 431
Figure 1. Flowchart detailing the data available for survival analysis and the data available for activity of daily living (ADL) decline analysis from the BELFRAIL cohort. a 19.6 2.5 months after baseline. b33.5 2.5 months after baseline. CRA = clinical research assistant.
mortality in subjects in the lowest tertile of grip strength and SPPB scores (Figure 2). Table 1 shows that the highest tertiles of grip strength, SPPB score, and muscle mass were associated with less risk of death than the lowest tertiles. The adjusted Cox proportional hazards model confirmed that participants with high grip strength or a high SPPB score had a lower risk of mortality, independent of muscle mass, comorbidity, and inflammatory markers (Table 2). Other factors independently associated with mortality were chronic obstructive pulmonary disease (COPD) and arthritis. In the model with SPPB scores, participants with high IL-6 levels had greater risk of mortality. Muscle mass was not related to mortality in the multivariate model.
Hospitalization There was no loss to follow-up for hospitalization. During follow-up, 51% (n = 284) of participants were hospitalized (Figure 1). Participant characteristics are shown in Table 1 according to hospitalization. Participants who were hospitalized during follow-up had more morbidities than those who were not. The Kaplan–Meier curves indicated significantly more hospitalization in participants with low grip strength and low SPPB scores (Figure 2). The highest tertiles of grip strength, SPPB score, and muscle mass were associated with less risk of hospitalization than the lowest tertiles. The unadjusted Cox proportional hazards model showed a relationship between all-cause hospitalization and grip strength (HR = 0.76, 95% CI = 0.65– 0.88) and between all-cause hospitalization and SPPB level (HR = 0.74. 95% CI = 0.63–0.86). The multivariate
117 (37) 104 (32) 100 (31) 150 140 129 23
138 (33) 137 (33) 139 (34)
180 180 186 23 123 (29) 142 (34) 156 (37) 8 3.2 118 (28) 137 (32) 168 (40) 31 8.6 127 (30) 155 (36) 147 (34)
180 (33) 182 (33) 184 (34) 8 3.4
176 (33) 170 (31) 194 (36) 31 8.7
185 (33) 185 (33) 187 (34)
(36) (33) (31) 9.8
125 (39) 106 (33) 90 (28)
139 (34) 136 (32) 139 (34)
(33) (33) (34) 9.5
0.17 (0.08–0.36) 1.9 (1.2–3.3) 3.6 (2.9–4.5)
4 (3–6)
4 (3–6)
0.19 (0.08–0.41) 2.1 (1.3–3.8) 3.7 (2.9–4.8)
158 (37) 273 (63) 84.4 3.5 70 13 28 4.6
Survived, n = 431
209 (37) 351 (63) 84.7 3.7 70 14 27 4.9
All Participants
b
1 1.6 (0.93–2.8) 2.0 (1.2–3.3)a 1 2.04 (1.3–3.2)b 1.4 (0.85–2.2)
1 0.43 (0.27–0.68)b 0.62 (0.42–0.94)a
1 0.35 (0.22–0.55)b 0.53 (0.34–0.81)a
1 0.63 (0.42–0.95)a 0.49 (0.32–0.76)b
30 (24)b 40 (31)b 57 (45)b 20 8.0b 57 (46)b 40 (32)b 28 (22)b 6 3.7b 58 (50)b 33 (28)b 26 (22)b 30 9.1b 58 (45)a 31 (24)a 40 (31)a
b
21 (23)a 33 (35)a 39 (42)a
b
1 2.2 (1.2–4.2) 4.0 (2.2–7.2)
b
0.98 (0.94–1.0)
1 1.1 (0.77–1.6) 1.1 (1.0–1.1)
HR (95% CI)
14 (15) b 30 (32)b 49 (53) b
0.25 (0.11–0.75) 2.8 (1.8–3.8) b 3.9 (3.2–5.5)a
5 (3–7)
51 (40) 78 (61) 85.8 3.9b 68 16 27 5.7
Deceased, n = 129
(35) (35) (30) (10.1)
78 (29) 96 (35) 100 (36)
69 (26) 96 (36) 101 (38) 31 8.9
71 (26) 95 (35) 105 (39) 8 3.2
95 95 80 24
74 (35) 78 (36) 61 (29)
79 (37) 70 (33) 64 (30)
0.17 (0.08–0.35) 1.9 (1.2–3.2) 3.6 (2.9–4.4)
4 (3–5)
93 (34) 190 (65) 84.4 3.7 71 14 28 5.0
No Hospitalization, n = 276
(31) (31) (38) (8.8)a
107 (38)a 90 (32)a 87 (30)a
107 (39)a 74 (27)a 93 (34)a 31 8.6a
109 (40)a 87 (31)a 79 (29)a 7 3.5b
85 86 105 22
64 (32)a 59 (29)a 78 (39)
60 (30) 66 (30) 75 (40)
0.20 (0.08–0.5)a 2.4 (1.4–4.3)a 3.8 (2.9–5.2)
5 (3–7)b
116 (41) 168 (59) 85.0 3.7 69 13 27 4.6
Hospitalization, n = 284
1 0.8 (0.6–1.0)a 0.75 (0.56–0.99)a
1 0.5 (0.4–0.7)b 0.6 (0.5–0.8)b
1 0.7 (0.5–0.9)b 0.6 (0.4–0.8)b
1 1.1 (0.8–1.4) 1.4 (1.1–1.9)a
1 0.92 (0.7–1.3) 1.4 (0.97–2.1)
1 1.2 (0.82–1.6) 1.5 (1.1–2.1)a
0.98 (0.95–1.0)
1 1.2 (0.94–1.5) 1.0 (1.0–1.1)a
HR (95% CI)
LEGRAND ET AL. 2014
Muscle mass, grip strength, and Short Physical Performance Battery (SPPB) score tertiles were adjusted for sex. a P < .05; bP ≤ .001. HR = hazard ratio; CI = confidence interval; SD = standard deviation; IQR = interquartile range; IL-6 = interleukin-6; TNF-a = tumor necrosis factor alpha; usCRP = ultrasensitive C-reactive protein.
Sex, n (%) Male Female Age, meanSD Weight, kg, mean SD Body mass index, kg/m2, meanSD Number of morbidities, median (IQR) Inflammatory markers usCRP, median (IQR) IL-6, median (IQR) TNF-a, median (IQR) IL-6 tertile, n (%) Low Medium High TNF-a tertile, n (%) Low Medium High usCRP tertile, n (%) Low Medium High Grip strength, median (IQR) Grip strength tertile, n (%) Low Medium High SPPB score, median (IQR) SPPB score (tertiles) Low Medium High Muscle mass, mean SD Muscle mass tertile, n (%) Low Medium High
Characteristic
Table 1. Subject Characteristics at Baseline According Vital Status and Hospitalization
4 JAGS
JAGS
2014
PREDICTIVE VALUE OF MUSCLE STRENGTH IN THE OLDEST OLD
All-cause mortality
Hospitalization Grip strength 100
Log rank p
