Frailty is a commonly used term which denotes a clinical entity in older persons(Fried et al., 2001). It is an important geriatric syndrome, with a global prevalence of 4.9%–27.3% (Choi et al., 2015) and is defined as an age-related decrease in the physiological reserve required to maintain biological homeostasis and increased vulnerability to stressors. Recognition of physical frailty is important for clinicians because it poses a greater risk of adverse health outcomes including falls, increased morbidity, physical dependence, hospitalization, and death (Fried et al., 2001). Thus, frailty has become one of the most significant clinical entities to afflict the elderly population. The most widely used diagnostic criteria are those initially proposed by Fried et al. and the Cardiovascular Health Study Research Group (Fried et al., 2001). These criteria include three or more of the following five: 1) Weight loss (10 lb in the past year), 2) self-reported exhaustion, 3) weakness (decreased HGS), 4) slow walking speed (> 6 to 7 s for 15 ft), or 5) decreased physical activity (males < 383 kcal, females < 270 kcal)(Fried et al., 2001). However, a number of other definitions of frailty exist, and no consensus, literature based definition has yet been reached. The presence of frailty has been associated with increased all-cause mortality(Cooper et al., 2010; Masel et al., 2010) and increased incident CVD (Newman et al., 2006), as well as poor survival after cardiac and surgical procedures (Chainani et al., 2016). The objective components of frailty syndrome (HGS and gait speed) have also been linked with an increased risk of all cause-mortality and adverse
15 health outcomes.
In elderly patients diagnosed with diabetes, frailty can be more common than in peers without diabetes (Group, 2003). Many studies report that frail individuals with diabetes have a higher mortality than non-frail individuals with diabetes. In addition, elderly patients with diabetes may have an increased risk for functional dependency and frailty (Araki and Ito, 2009). The frailty prevalence of 32% to 48% in persons >
65 years with diabetes is much higher than the 5% to 10% seen in the general population of the same age (Wang et al., 2014), thus, elderly patients with diabetes require screening for frailty(Won and Kim, 2016). In general, the association between higher serum cholesterol levels and increased CVD risk is attenuated in old age and may be reversed(Schatz et al., 2001; Sanford and Morley, 2014). An apparent increase in mortality associated with low cholesterol level in older people can be related to malnutrition, frailty, and chronic diseases, which simultaneously decreases the cholesterol level and increases the mortality risk(Kronmal et al., 1993; Corti et al., 1997).
E. Association between handgrip strength (HGS) and cardiovascular disease (CVD)
There is some evidence that HGS is associated with chronic disease; in the
16
Prospective Urban-Rural Epidemiology (PURE) study, a large, longitudinal population study, during a median follow-up period of 4 years, and among 142,861 participants, after adjustment, the association between grip strength and each outcome, with the exceptions of cancer and hospital admission due to respiratory illness, was similar across country-income strata. Grip strength was inversely associated with all-cause mortality (hazard ratio per 5 kg reduction in grip strength, 1.16; 95% CI, 1.13–
1.20; p < 0.0001), cardiovascular mortality (1.17; 1.11–1.24; p < 0.0001), non-cardiovascular mortality (1.17; 1.12–1.21; p < 0.0001), myocardial infarction (1.07;
1.02–1.11; p = 0.002), and stroke (1.09; 1.05–1.15; p < 0.0001). In addition, HGS was a stronger predictor of all-cause and cardiovascular mortality than systolic blood pressure (Leong et al., 2015)
Meta-analysis of a prospective cohort (Wu et al., 2017), using data obtained from 42 studies, including 3,002,203 participants, for the lowest versus highest category of grip strength, the HRs (95% CIs) were 1.41 (1.30–1.52) for all-cause mortality, 1.63 (1.36–1.96) for CVDs, and 0.89 (0.66–1.20) for cancer. The HRs (95% CIs) with a per-5 kg decrease in grip strength were 1.16 (1.12–1.20) for all-cause mortality, 1.21 (1.14–1.29) for CVDs, 1.09 (1.05–1.14) for stroke, 1.07 (1.03–1.11) for coronary heart disease, and 1.01 (0.98–1.05) for cancer. The observed associations did not differ by sex, and remained after excluding participants with either CVDs or cancer at baseline
17
A recent study reported that HGS is strongly and inversely related to all-cause mortality and the incidence of CVD, respiratory disease, chronic obstructive pulmonary disease, all cancer and subtypes of cancer, including colorectal, lung, and breast cancer, and with the associations being modestly stronger in the younger age groups. In women and men, hazard ratios per 5-kg lower grip strength were higher (all at p < 0.05) for all-cause mortality (1.20, 95% confidence interval, 1.17–1.23, and 1.16, 1.15–1.17, respectively) and cause specific mortality from CVD (1.19, 1.13–
1.25, and 1.22, 1.18–1.26, respectively), all respiratory disease (1.31, 1.22–1.40, and 1.24, 1.20–1.28, respectively), chronic obstructive pulmonary disease (1.24, 1.05–
1.47, and 1.19, 1.09–1.30, respectively), all cancer (1.17, 1.13–1.21, and 1.10, 1.07–
1.13, respectively), colorectal cancer (1.17, 1.04–1.32, and 1.18, 1.09–1.27, respectively), lung cancer (1.17, 1.07–1.27, and 1.08, 1.03–1.13, respectively), and breast cancer (1.24, 1.10–1.39, women only) (Celis-Morales et al., 2018).
In a cross sectional study (Lee et al., 2018) using the Korea National Health and Nutrition Examination Survey (2014–2016), that included 8,576 participants aged 40 to 79 years (men: 3807 and women: 4769), the individual CVD risk was evaluated by calculating the atherosclerotic cardiovascular disease (ASCVD) risk score and the Framingham risk score (FRS) in subjects aged 40 to 79 years without prior CVD.
Multivariate linear regression analysis revealed a significant inverse association (in both men and women) between relative HGS and cardiovascular risk factors, including blood pressure, levels of fasting glucose and triglycerides, waist
18
circumstance, FRS, high sensitivity C-reactive protein levels, and ASCVD risk. A significant positive association between relative handgrip and a low level of high density cholesterol levels in both men and women was identified. In both men and women, subjects in the lowest quartile of HGS had an increased risk of CVD compared to those in the highest quartile (OR range, 2.05–3.03). Furthermore, a systematic review assessing the relationship of grip strength and cardiovascular mortality reported that decreased HGS was associated with increased mortality in the majority of studies (8/12) (Table 2)(Chainani et al., 2016)
A study that reported the relationship between HGS and CV events, and all-cause mortality based on an analysis of the Outcomes Reduction with an Initial Glargine Intervention (ORIGIN) clinical trial, including 12,516 individuals (35% women) of mean (SD) age 63.6 (7.8) years, had a baseline HGS and were followed for a median of 6.2 years. This study showed that higher HGS is significantly associated with a lower incidence of death from cardiovascular events such as myocardial infarction, stroke or CV death myocardial infarction, stroke and CV death (CV death, per 1 kg increased HGS: men, 0.88 (0.86–0.90), p < 0.001; women, 0.70 (0.66–0.75), p <
0.001; stroke, per 1 kg increased HGS: men, 0.90 (0.87–0.93), p < 0.001; women, 0.84 (0.77–0.90), p < 0.001; MI, per 1 kg increased HGS: men, 0.97 (0.94–1.00), p = 0.04; women, 0.79 (0.72–0.86), p < 0.001; heart failure, per 1 kg increased HGS: men, 0.89 (0.86–0.91), p < 0.001; women, 0.70 (0.64–0.76), p < 0.001)(Lopez-Jaramillo et al., 2014)
19
The Hisayama study, which investigated the association between HGS and risk of cause-specific death in the middle-aged and older Japanese population, followed a total of 2,527 community-dwelling Japanese (1,064 men and 1,463 women) aged ≥ 40 years for a period of 19 years. During the follow-up period, 783 participants died, of whom 235 died of CVD, 249 of cancer, 154 of respiratory disease, and 145 of other causes. They demonstrated that higher levels of HGS were significantly associated with a decreased risk of cardiovascular death, respiratory death, and death from other causes in the middle-aged and elderly (CV death, men, HR = 0.52; 95% CI, 0.41–
0.66; p < 0.05, and women, HR = 0.77; 95% CI, 0.64–0.93; p < 0.05)(Kishimoto et al., 2014).
Several studies have reported no association between HGS and cardiovascular death. A prospective cohort study in Taiwan (Chen et al., 2012), that aimed to determine the prognostic value of HGS and walking speed in predicting the cause-specific mortality for older men, recruited all subjects 75 years and older in northern Taipei in March of 2008 and followed to December of 2010.During the study period, 99 participants died and the baseline HGS and walking speed were significantly lower than those of the survivors (p both < 0.001). Cox survival analysis showed that the lowest quartile of HGS did not significantly predict cardiovascular death when they fully adjusted for covariates (vs Q1, Q2: HR = 0.90; 95% CI, 0.28–2.87; Q3: HR = 0.63; 95% CI, 0.14–2.24; Q4: HR = 0.70; 95% CI, 0.19–2.59). Another longitudinal
20
study (McDermott et al., 2012), which analyzed whether lower calf muscle density and poorer upper and lower extremity strength were associated with higher mortality rates in men and women with peripheral arterial disease, included 238 participants with peripheral arterial disease attending their fourth annual follow-up visit in the original Walking and Leg Circulation Study (WALCS), and 240 peripheral arterial disease participants newly identified for WALCS II. At baseline, participants underwent measurement of calf muscle density with computed tomography, in addition to knee extension power, isometric knee extension, plantar flexion, and HGS measures. They reported that lower calf muscle density was associated with higher all-cause and CVD mortality among participants with PAD, but that there was no significant association of HGS with CVD mortality in men (below median vs above median HGS: HR = 1.63; 95% CI, 0.47–5.68).
21
Table 2 Relationship between grip strength and cardiovascular death in previous studies
Author(year) N(Age±SD) Main finding Adjusted HR(HGS and CVD
12,516(64±8) Higher HGS associated with lower CVD mortality and events
2,527(59±2) Higher HGS associated with lower CVD mortality in both
Age, SBP, antihypertensive use, diabetes, total cholesterol, BMI, EKG abnormalities, smoking, Alcohol and leisure-time physical activity
22
Age, height, BMI, waist circumference, current smoking, exercise, hypertension, diabetes mellitus, proteinuria, walking speed, hemoglobin, platelet,
neutrophil count,
lymphocytecount,HbA1C,alaninetransaminase, creatinine, albumin, uric acid, triglycerides, total cholesterol and HDL
2,313(78±10) HGS predicted CVD mortality in females in fully adjusted model
Per SD increased in HGS in females:
0.60(inferred from text)
Age, gender, HbA1c, SBP, total cholesterol, HDL, DHEAS, CRP, walking distance, pulmonary peak flow, waist circumference, disability scale, cancer diagnosis, involuntary weight loss, smoking.
23
III. Method
A. Study Population
The study population was derived from a nationwide panel survey, the Korean Longitudinal Study of Aging (KLoSA), on individuals over the age of 45 years. The KLoSA data was gathered for the purpose of preparing for the aged society in terms of system reform and policy decision. The data was composed of seven categories including population, family, health, employment, income, wealth, subjective expectation, and life expectation. The KLoSA is a biennial survey of nationally-representative Koreans aged 45 years or older, excluding institutionalized people and residents of Che-Ju Island, used to build-up the basic data needed to devise effective social and economic policies in order to address the trends in the population ageing process. Participants were selected randomly using a multistage, stratified probability sampling design to create a nationally representative sample. Sampling was conducted by sorting the population surveyed in a given area and 15 residential types according to the order of the administrative codes, and then extracting the assigned number by applying a systematic extraction method (the multistage and stratified sampling method). In the first baseline survey in 2006, 10,254 individuals in 6,171 households (1.7 per household) were interviewed, and results demonstrated that there were 292 individuals with cancer. The second survey, in 2008, followed up with
24
8,675 subjects, who represented 86.6% of the original panel. The third survey, in 2010, followed up with 8,229 subjects, who represented 81.7% of the original panel, the fourth survey, in 2012, followed up with 7,813 subjects, who represented 80.1%
of the original panel, and the fifth survey, in 2014, followed up with 8,387 subjects (including 920 who newly participated in the sample), who represented 80.4% of the original panel. The sixth survey, in 2016, followed up with 9,913 subjects (including 878 who newly participated in the sample), who represented 79.6% of the original panel. Out of all the public data available in Korea, the KLoSA was considered to be the most suitable data for the analysis involved in the current study. In this study, 8,424 participants were included in the analysis at baseline, excluding those diagnosed with heart diseases and stroke, and those with missing values for the variables of interest (Figure 2). The design and protocol of this study were approved by the Institutional Review Board (IRB) of Ajou University Hospital (IRB No.
AJIRB-SBR-EXP-18-474).
25
Figure 2 Flow diagram of the study population at baseline (2006)
B. Dependent variable
Cardiovascular disease, which was a dependent variable in this study, was defined as the prevalence of CVD in patients with heart disease (angina, myocardial infarction,
26
congestive heart failure), and stroke, which could be extracted from the KLoSA data by referring to previous studies(Leong et al., 2015; Byung-Taek Oh, 2017). The KLoSA data has a structure whereby it is possible to skip the questionnaire in the next wave if the previous wave has confirmed chronic diseases. Therefore, it is unlikely that a disease was recorded incorrectly (Institute, 2007). The presence of CVD, such as heart diseases and stroke, is assessed in response to the questionnaire, and the answer depends on whether the subjects are diagnosed by a doctor, not by the individual’s own judgment.
C. Measurement of handgrip strength (HGS)
Handgrip strength was measured by a handgrip dynamometer (Model number:
NO6103, Manufacturer: TANITA, Japan). The test was performed in a sitting position with the elbow flexed at 90˚ on both the right and the left sides. The grip strength measurement was performed in order to determine whether or not a respondent was in a state in which the measurement could be made, and was not performed when the user refused, if one of the hands was injured, or if the user was ill. In this study, we defined dominant HGS as the maximal HGS of the dominant hand, whereas absolute HGS was calculated as the summation of the maximal reading from each hand using measured HGS values, and was expressed in kilograms. The relative HGS was defined as the absolute HGS divided by the BMI(Lawman et al., 2016; Yi et al., 2018). We applied various cut-off points proposed by the sarcopenia
27
guidelines (Chen et al., 2014; Studenski et al., 2014; Yoo et al., 2017; Cruz-Jentoft et al., 2019) as a criterion of weak muscle strength, using HGS to compare to the optimal cut-off value of this study.
D. Covariates
We selected the covariates of this study by referring to previous studies (Byung-Taek Oh, 2017; Kim et al., 2019) that reported the factors related to decreased HGS in a Korean elderly population. Age, sex, marital status, education level, residential region, and economic activity were considered as sociodemographic variables. The levels of education were categorized as ‘less than elementary school’, ‘middle school graduate’, ‘high school graduate’, or ‘college graduate or beyond’. Two age group categories were used as follows: < 64 and ≥ 65 years.Marital status was classified as married or unmarried (including divorce and separation). The residential regions were categorized into metropolitan (Seoul), urban (administrative divisions of a city:
Daejeon, Daegu, Busan, Incheon, Kwangju, or Ulsan), or rural (not classified as administrative divisions of a city). Current economic activity was categorized as either ‘yes’ or ‘no’. In terms of health behaviors, we selected smoking, alcohol consumption, and regular exercise; smoking was classified into ‘non-smokers’, who never smoked, ‘former smoker’, and ‘smoker’. Alcohol consumption was divided into two groups according to the current drinking status, and regular exercise was divided into two groups based on more than one exercise per week.
28
The BMI, ADL, and Mini-Mental State Examination (MMSE) were included in the health-related variables. Body mass index was calculated from self-reported data on weight and height during the first wave; the study population was divides into three groups as follows: Underweight (BMI of less than 18.5 kg/m2) , normal (BMI of 18.5–
24.9 kg/m2), and overweight (BMI of 25 kg/m2 or over). Activities of daily living is a measure of basic daily living skills, including changing clothes, watering/brushing/hair-drying, bathing/showering, eating, going out of the room, using the toilet, and is measured according to whether the individual requires help to perform the activities. If they need some help in their everyday life, or if they need help altogether, they are set to '1'. If they do not need any help, they are set to '0'(Institute, 2007).TheADL index is classified as normal at 0 point, and it is accepted that more than 1 point represents an individual who is limited in daily life practice.
Cognitive impairment was determined by the K-MMSE score on the first wave. The K-MMSE score is used to determine cognitive function from questions designed to assess various categories of cognitive function, such as time and place, orientation, registration, attention and calculation, memory recall, language, and visual construction. The MMSE score is calculated based on the summation of the score of each category. A score of 17 or less indicates a ‘suspicion of dementia’, a value of 18 or more and 23 or less indicates ‘cognitive decline’, and 24 or more is classified as
‘normal’. In addition, we included diabetes and hypertension as a health condition variable that could affect CVD.
29
Figure 3 Framework of research
E. Statistical analysis
In this study, using 2006 as the baseline year and considering the follow-up period until 2016. Descriptive statistics were performed, using the chi-squared test for categorical variables and the t-test for continuous variables, to examine the demographic characteristics and constitution scores of the participants. The generalized estimating equation (GEE) regression model was used to determine whether the probability of CVD changed over time. In GEE, proc genmod was used, with link logit, distribution normal. In the GEE model, the clustered approaches to analysis, controls for the characteristics of individuals that change over time, such as
30
confounding variables. The clustering can be expressed in terms of the correlation among the measurements in units within the same cluster. The Quasi-Akaike Information Criterion (QIC) was calculated in order to compare the best fit among the three HGS indices. Pearson’s correlation coefficients were calculated to relate each HGS index versus age at follow-up, and ROC curves were used to HGS indices in sensitivity and specificity to predict CVD. All analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA), and statistical significance was considered at the level of p < 0.05.
31
IV. Results
A. Characteristics of Participants
The general characteristics of the participants according to the muscle weakness (lowest 20%) at baseline are shown in Table 3. Among a total of 8,494 subjects, 3,802 (44.8%) were men, and 4,692 (55.2%) were women. The average age of participants was 60.7 ± 10.6 years (range, 45–98 years; men, 60.6 ± 10.3 years; women, 60.7 ± 10.9 years). Among the total subjects, 1,382 (16.3%) had weak HGS, 562 (14.8%) men and 820 (17.5%) women. In both men and women, the proportion of low muscle strength in those 65 years and over was significantly higher than those under 65 years of age (men, 31.5% vs 5.6%; p < 0.0001; and women: 36.4% vs 7.4%; p < 0.0001).
In both men and women, The means of the dominant HGS, absolute HGS, and relative HGS with normal muscle strength were significantly higher than those of the low muscle strength group (men: DHGS 35.9 ± 5.5 kg vs 23.3 ± 5.0 kg, p < 0.0001; AHGS 70.4 ± 10.6 kg vs 43.2 ± 8.21 kg, p < 0.0001; RHGS 3.0 ± 0.5 vs 2.0 ± 0.4, p = 0.001;
women: DHGS 22.7 ± 4.0 kg vs 13.7 ± 3.2 kg, p < 0.0001; AHGS 44.1 ± 7.5 kg vs 25.1 ± 5.2 kg, p < 0.0001; RHGS 1.9 ± 0.4 vs 1.1 ± 0.3, p < 0.001, respectively). In terms of chronic diseases that affect CVD, the distribution of low HGS with hypertension and diabetes was significantly higher in men than in the undiagnosed group (hypertension: 19.3% vs 13.5% p< 0.0001; diabetes: 22.1% vs 13.8%, p <
32
0.0001), and the trend was similar in women (hypertension: 24.4% vs 15.0%, p<
0.0001; diabetes: 27.1% vs 16.4, p< 0.0001). In addition, there was a significance difference in marital status, education level, area of residence, economic activity, regular exercise, smoking status, alcohol consumption, the ADL score, and the MMSE score among the weak and normal in both sexes (all p values < 0.05).
33
Table 3 General characteristics of participants for analysis according to low handgrip strength by sex at baseline (2006)
N / mean % / SD N / mean % / SD N / mean % / SD N / mean % / SD
Age <.0001 <.0001
<65 137 5.6 2,316 94.4 225 7.4 2,814 92.6
≥65 425 31.5 924 68.5 595 36.0 1,058 64.0
Dominant HGS, kg 23.3 4.92 35.9 5.51 <.0001 13.7 3.23 22.7 4.01 <.0001
Absolute HGS, kg 43.2 8.21 70.4 10.59 <.0001 25.1 5.20 44.1 7.48 <.0001
Relative HGS, kg/BMI 2.0 0.44 3.0 0.53 <.0001 1.1 0.28 1.9 0.38 <.0001
Marital status <.0001 <.0001
Married 486 13.8 3,040 86.2 394 11.6 2,992 88.4
Single 76 27.5 200 72.5 426 32.6 880 67.4
Region of residence <.0001 0.001
Metropolitan 229 13.8 1,434 86.2 367 17.2 1,773 82.9
Urban 164 12.8 1,117 87.2 237 15.4 1,300 84.6
Rural 169 19.7 689 80.3 216 21.3 799 78.7
Education level <.0001 <.0001
Elementary school or less 311 28.1 797 71.9 668 26.6 1,845 73.4
Middle school graduate 87 13.4 561 86.6 67 8.4 731 91.6
High school graduate 122 9.0 1,228 91.0 71 6.3 1,054 93.7
College graduate or higher 42 6.0 654 94.0 14 5.5 242 94.5
Economic activity <.0001 <.0001
Yes 157 6.7 2,204 93.4 110 8.7 1,157 91.3
No 405 28.1 1,036 71.9 710 20.7 2,715 79.3
Alcohol consumption <.0001 <.0001
Yes 299 11.9 2,215 88.1 105 11.4 819 88.6
No 263 20.4 1,025 79.6 715 19.0 3,053 81.0
Smoking 0.024 0.022
Yes 210 13.3 1,375 86.8 35 24.7 107 75.4
No 352 15.9 1,865 84.1 785 17.3 3,765 82.8
Regular exercise <.0001 <.0001
Yes 166 10.1 1,483 89.9 198 11.4 1,542 88.6
No 396 18.4 1,757 81.6 622 21.1 2,330 78.9
ADL <.0001 <.0001
0 525 14.0 3,217 86.0 763 16.6 3,827 83.4
≥1 37 61.7 23 38.3 57 55.9 45 44.1
BMI, kg/m2 <.0001 <.0001
Underweight (<18.5) 52 39.4 80 60.6 88 43.4 115 56.7
Normal (18.5-24.9) 312 19.5 1,289 80.5 388 18.4 1,727 81.7
Overweight(≥25) 198 9.6 1,871 90.4 344 14.5 2,030 85.5
MMSE <.0001 <.0001
Suspicion of dementia (0-17) 53 63.9 30 36.1 235 57.3 175 42.7
Cognitive decline (18-23) 142 39.3 219 60.7 232 28.9 570 71.1
Normal (≥24) 367 10.9 2,991 89.1 353 10.1 3,127 89.9
Hypertension <.0001 <.0001
diagnosed 162 19.3 676 80.7 306 24.4 948 75.6
undiagnosed 400 13.5 2,564 86.5 514 15.0 2,924 85.1
Diabetes melitus <.0001 <.0001
diagnosed 97 22.1 343 78.0 132 27.1 355 72.9
undiagnosed 465 13.8 2,897 86.2 688 16.4 3,517 83.6
Total 562 14.8 3,240 85.2 820 17.5 3,872 82.5
Numeric parameters are expressed as mean and standard deviation in parentheses Categorical parameters are expressed as counts and percentages in parentheses
Cut-off value for low handgrip strength was defined as the lowest 20% of HGS of the study population. ( <26.1 kg for men and <16.0 kg for women, respectively).
Acronyms: BMI, body mass index; MMSE, Mini-mental state examination; SD, standard deviation; ADL activities of daily living
p -value Women (n=4,692)
p -value Women (n=4,692)