A Cross-sectional Survey of Prevalence and Related Factors for Frail Status Among Middle-aged and Elderly People in Chinese Communities

Background With the aggravation of social competition and work burden pressure, the health condition of the middle-aged and above population in China has declined signicantly. Frailty can be used as a criterion for evaluating a person's unhealthy state. However, there is limited data on the prevalence and related factors of frail status in Chinese middle-aged and older people.The objective of this study was to explore the correlation between frail status, chronic diseases, abnormal physical examination indicators among middle-aged and older populations. Methods Participants were 9,985 community-dwelling adults over the age of 40 years living in China. Data were from the 2015 China Health and Retirement Longitudinal Survey which was a nationally representative sample and frailty phenotype was based on Fried frailty criteria. We analyzed the demographics of participants and multivariate-adjusted related factors for frail and pre-frail population. Results The overall prevalence of frailty and pre-frailty was 3.1% (95% CI: 2.8, 3.5) and 53.66% (95% CI: 52.7, 54.6) among the general Chinese population aged 40 years or older. Hypertension (OR: 1.8, 95% CI: 1.3, 2.593), pain (OR: 1.8, 95% CI: 1.3, 2.5), and hip fracture (OR: 2.2, 95% CI: 1.1, 4.4) were associated with prevalent frailty. Relative factors for frailty also included increased cystatin C (OR: 4.5, 95% CI: 3.0, 6.7) and glycated hemoglobin (OR: 1.2, 95% CI: 1.1, 1.4), as well as decreased peak expiratory ow (OR: 0.993, 95% CI: 0.991, 0.994). Conclusions Fried-dened frailty and pre-frailty are highly prevalent in the Chinese population over 40 years older. Hypertension, pain, hip fracture, low education, and underweight are major related factors for frailty. The decrease in peak expiratory ow and the increase in cystatin C and glycated hemoglobin are good indicators for detecting frailty. Thus, frailty is an 6.86 , 95% CI: 4.63, 10.15; P < 0.001).In multivariable-adjusted analyses, hypertension was still associated signicantly with risk of frailty (OR: 1.82, 95% CI: 1.27, 2.59; P < 0.001) and pre-frailty (OR: 1.16, 95% CI: 1.02, 1.32; P= 0.02). Compared to normal blood pressure, person in grade 3 hypertension had a two-fold greater odds of frailty (OR: 2.14, 95% CI: 1.10, 4.17; P = 0.024).Multivariate analysis showed that gender (OR:0.52, 95% CI: 0.47, 0.59; P < 0.001),pain (OR: 1.80, 95% CI: 1.31, 2.46; P < 0.001), hip fracture (OR: 2.18, 95% CI: 1.07, 4.43; P = 0.032), cystatin C (OR: 4.50, 95% CI: 3.01, 6.74; P < 0.001), glycated hemoglobin (OR: 1.07, 95% CI: 1.07, 1.40; P = 0.003), and peak expiratory ow (OR: 0.993, 95% CI: 0.991, 0.994; P < 0.001) were also relative factors with frailty.


Background
With the rapid development of China's social modernization and population aging, the aggravation of social competition and work burden has led to the obvious decline in the healthycondition of the middle-aged and above Chinese people.As anunhealthy status of individual body and mind, frailty has attracted more and more attention from researchers. There is a common agreement that frailty is a medical syndrome in which strength and physiological functions are reduced by multiple factors to increase individual vulnerability [1].Fried phenotype of frailty is the most frequently used serve as a criterion for assessing a person's frail state in the literature [2].The causes of frailty are multifaceted, including age, lower-income, polypharmacy, sarcopenia, depression, malnutrition, falls [3][4][5][6][7][8][9][10][11][12][13]. Several studies have also shown that some chronic diseases, such as hypertension, anemia, chronic obstructive pulmonary disease and diabetes are related to the occurrence of frailty [14][15][16][17].
China has the largest population base and the health of expenditure is a heavy economic burden for personal and societal cost is enormous [18]. Compared with developed countries, China's population has great regional differences in diet and living habits, unbalanced economic status and medical services, and the number of patients with chronic diseases continues to grow rapidly [19][20][21]. Therefore, the related factors and chronic disease characteristics of the frail population in China may be very different from those in Europe and the United States [22]. However, most of the study evidence comes from developed countries, and the research on frailty in China has just started [23].Based on 2011 China Health and Retirement Longitudinal Survey (CHARLS) data, Wu and his colleagues found that the prevalence of frailty among the Chinese community population aged 60 and older was 7% [24]. But their study did not include the middleaged population, and did not calculate the prevalence of pre-frailty in our country. Somestudies indicate that frailty and pre-frailty are identi ed in middle age (40 to 59 years old), which may have implications for prognosis and planned intervention, especially in individuals with multimorbidity [25]. This study aimed to identify the prevalence of frailty and pre-frailty in middle-aged and elderly Chinese,and to explore related factors.

Data source and study design
In our study, we used the survey date from the China Health and Retirement Longitudinal Survey (CHARLS), conducted in 2013 and 2015 (available at http://charls.pku.edu.cn/zh-CN). CHARLS is a nationally representative longitudinal cohort study including communitydwelling mid-aged and elderly population collectively in China, and they were followed up every two years.It used a multistage strati ed cluster sampling procedure to select the respondents. A large-scale baseline survey was conducted from 2011 to 2012, covering all countylevel units in mainland China (excluding Tibet). The sample includes 150 counties/districts (spread across 28 provinces and regions) and 450 villages / urban communities with a total of 17,708 people. The survey gathered information about socioeconomic characteristics, living habits, health status, and other demographic messages by face-to-face interviews and also presented biomarkers, blood date, anthropometric and other physical measurements. By analyzing the date from the Cardiovascular Health Study, Fried and his colleagues theorized the frailty phenotype model, which evaluated ve aspects of pace, grip strength, exhaustion, physical activity and weight loss [2].As speci c measurements and questions for the CHARLS differed from the Cardiovascular Health Study, which is we de ned frailty status in terms of Fried frailty criteria and made minor modi cations based on data availability for the Chinese version. Table 1 shows the de nition of Fried and colleagues alongside our de nition.
Biomarker and follow-up questionnaire data In CHARLS, a trained interviewer measured blood pressure and pulse 3 times on the participants' left arm with an electronic blood pressure monitor (Omron HEM-7200 Monitor). We rst calculated the mean value of the 3 blood pressure and pulse readings for every participant.
According to the values of systolic blood pressure (SBP) and diastolic blood pressure (DBP), the population was divided into ve categories (Footnote to Table 5) [26].
The interviewer measured twice on each hand of the participants by using an electronic grip dynamometer ( Yuejian WL-1000 mechanical dynamometer), and the average value of the dominant hand was calculated as the nal grip strength to be included in the analysis.
Similarly, the interviewer used a stopwatch to record the time at which the participant walked at a usual pace of 2.5 meters twice, taking the average time to calculate the pace. walking speed = 2.5 (m)/time (s).
height and weight were measured by trained interviewers with a stadiometer and scale. Waist circumference was measured at navel level with soft type. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.
Trained interviewers used the peak ow meter to conduct the peak expiratory ow (PEF) for three times. Participants opened their mouth and closed lips rmly around the outside of the mouthpiece, and then blew as hard and as fast as they can into the mouthpiece. We recorded the measurement readings and calculated the mean value.
According to answer these questions, such as "have you fallen down?" and "have you ever fractured hip", and "are you troubled with body pains?", participants were divided with "yes or no". The de nition of chronic diseases was based on participates self-reported diagnosis by a doctor. The protocol of the blood-based biomarker sample collection study was approved by the ethical review committee. Written informed consent was obtained from all study participants.

Statistical analysis
In our study, sociodemographic characteristics, self-reported health status, physical tests, and other relevant information (a little of date were missing) were summarized for frail, pre-frail and non-frail participants. We provided reliable estimates of the prevalence of frailty status for both men and women in ve age groups. The prevalence of frailty status among men and women is presented as percentages with a 95% con dence interval (CI). Quantitative data conforming to normal distribution and homogeneity of variance tests were expressed as mean values and standard deviation and compared using one-way analysis of variance. Qualitative data were expressed as numbers and percentages and compared with the Chi-squared test.
We assessed the association between risk factors and frailty status with univariate and multivariable logistic regression models. The different frailty status of chronic diseases was calculated odds ratios (OR), 95% CI, and P-value. All the important variables in the descriptive analysis were included in the nal model. We used SPSS 23.0 (IBM Corp., Armonk, NY, USA) software and Excel 2016 (Microsoft Corp., Redmond Washington, USA) to conduct statistical analyses, and the signi cance level was set at 0.05 (2 tail test).

Discussion
Our estimated prevalence of frailty was 3.1% among the Chinese population aged 40 years or older, which is very similar to the reported by Hanlon and colleagues that around 3% of people aged 37 to 73 were considered frail in the United Kingdom [25]. To our knowledge, our study is the rst survey reporting the middle-aged of frailty in China, showing prevalence of 2.4% aged 40-49 and 3.6% aged 50-59, which was lower than the middle-age of frailty in community-dwelling Europeans was 4.1% and was much lower than reported the prevalence of frailty was 7.6% and 13.4% among adult white Americans aged 35-44 and 45-54, respectively [27,28]. It can be seen that frailty not only has a higher incidence in the elderly, but also occurs in the middle-aged population. Research has shown that the frailty of the body may lead to disability in the future, but in the early stages of intervention, this outcome can be changed [29]. So many studies published about the prevalence of pre-frailty, such as United States (45.0%), UK(38.0%), Japan (48.1%), Israel (57.4%), similarly, the proportion of pre-frail Chinese in our research was also very high (53.7%) [3,25,30,31].
These distinctions might be mainly due to differences in socioeconomic background, ethnicity, and lifestyle of people living in different countries and regions [31]. In addition, using different frailty measurement tools for the same population, such as Fried frailty phenotype, Frail index, and FRAIL scales, would have an impact on the outcome of the prevalence of frailty and pre-frailty [32].
We found in the multivariate analysis, people who reported hypertension had 1.8 times the risk of developing frailty compared to those who did not, and the risk of pre-frailty is 1.1 times.After adjusting for potential confounding factors, the higher the blood pressure level, the greater the chance of suffering from frailty. Also in previous studies, there was a strong association between hypertension and frailty, and it was con rmed that patients with frailty and pre-frailty had signi cant subclinical vascular and cardiacchanges [33,34]. However, the conclusions of blood pressure and frailty research are divergent, and some authors have shown that measured mean blood pressure was lower in frail compared to non-frail individuals [35]. They thought low blood pressure may reduce blood perfusion and oxygenation of vital organs, resulting in damage, loss of function, and a frail state [36].This difference in blood pressure was explained that high blood pressure is harmful to people's health, whether the population was frail or not, it was possible to use antihypertensive treatment. Studies have shown that hypertension can cause frailty, and as a variable, frailty may affect changes in blood pressure. These ndings raise questions about statistical differences regarding the blood pressure values in different frail states of the population require additional evidence.
We observed there was a strong association between frail status and elevated serum cystatin C.Multivariate analysis showed that subjects with higher cystatin C had 4.5-fold increased risk of frailty and 2.4-fold increased risk of pre-frailty. Hart et al found that higher serum cystatin C was associated with increased risks of progression of frailty status and death [37].Cystatin C is a low molecular weight protein that has a stable production rate and can be freely ltered by the glomeruli[38], which is more appropriate for the group who are susceptible to be frail because it seems to be less affected by muscle mass or dietary protein intake [39].Our ndings, like those of other studies,suggest that cystatin C is more sensitive to the detection of frailty than other blood biomarkers such as creatinine [40]. These results may partially explain the increased sensitivity of cystatin C in identifying older people with less muscle. Because cystatin C was not affected by muscle mass or dietary protein intake, it was more re ective of changes in physical performance than other kidney function markers such as creatinine [39].In ammation may play a role as a potential mechanism linking cystatin C to frailty [41]. Due to the limitations of current frailty knowledge, the mechanism linking higher cystatin C to frailty was unclear and future research is needed to further explore this phenomenon [37].
In this study, we also identi ed other relative factors associated with frailty and pre-frailty in the Chinese community population.Gender, education level, BMI,glycated hemoglobin (HbA1C), PEF and pain are major relative factors for frailty. The prevalence of frailty and prefrailty in women was higher than in men, which is consistent with previously published studies [5,42].We also found that higher education has a strong protective effect on the prevalence of frailty. The education level of individuals is one of the important indicators to measure their socioeconomic status [43]. Lower economic income is associated with an increased risk of frailty [5,28]. The prevalence of frailty increased with low BMI. Our ndings corroborate the views that a lower BMI may indicate insu cient reserve capacity and weight loss, which was a key factor in the formation of frailty [44]. Underweight may be due to chronic illness or malnutrition and sarcopenia, which were also associated with increased risk of frailty [45].
Our study indicated that the decrease in PEF and the increase in HbA1C may be good indicators for detecting frailty. The result is similar to previous studies on respiratory function and frailty, in which PEF values were lower in frail elderly people than in non-frail people [46]. This mechanism may be related to decreased function of the respiratory muscles or chronic in ammation of the lungs that led to a decline in the body's energy reserves and physical strength [47]. Similarly, our results showed that for every 1 percent increased in serum HbA1C, the risk of frailty had higher odds. Several studies had shown that diabetes and HbA1C were risk factors for frailty and patients with poor blood sugar control had a higher risk of developing frailty [17,48]. Thus, frailty is an increasingly common condition and will become an increasingly important health issue for people over middle age. Next, how to prevent the frailty of the population should become one of the major medical concerns.
Our study has several limitations. First, the chronic diseases and symptoms were self-reported, therefore, it would cause reporting bias and recall bias. Second, because the biological and clinical characteristics of middle-aged people were different from those of the elderly, the concept of frailty, especially the de nition of Fried frailty phenotype, when applied to non-elderly, the applicable conditions and principles of de nition here need to be further studied in the future. Third, due to the survey of CHARLS was only included community-dwellers, not including those in hospitals and nursing homes, and some people who may have serious illnesses that cannot be interviewed and measured were excluded. In addition, the proportion of the elderly population in the surveyed population was low (age ≥ 60 years had 48.1% and age ≥ 80 years had 4.1%), while the elderly population was a high-risk group with frailty. Therefore, the actual incidence of frailty may be underestimated.Fourth, We found that in the above studies on the frailty and pre-frailty, due to the large or minor modi cations to the Fried frailty phenotype or other frailty assessment instruments in order toaccommodate the data, would have a signi cant impact on the estimation of frailty and pre-frailty population, and we should be cautious about the publication of this conclusion. The last, this research was a retrospective cross-sectional study with insu cient strength as a level of clinical evidence and were unable to evaluate the role of chronic diseases at differing stages of the frailty process. Due to the limitations of cross-sectional study design, we could not understand the causal and temporal relationship between frailty and chronic diseases. The next step is to conduct prospective cohort studies and randomized controlled trials to serve clinical guidelines.
Abbreviations BMI: Body mass index; CHARLS: China Health and Retirement LongitudinalStudy; CI: con dence interval; DBP: diastolic blood pressure; HbA1C: glycated hemoglobin; OR: odds ratios; SBP: systolic blood pressure; PEF: peak expiratory ow; Declarations Ethics approval and consent to participate The biomarker sample collection study protocol was approved by the ethical review committee (institutional review board) of Peking University (IRB 00001052-11014). Written informed consent was obtained from all study participants.

Availability of data and materials
All data used in the study and questionnaire can be accessed through after registering publicly online website (http://charls.pku.edu.cn/en).

Consent for publication
Not applicable.
The authors declare that they have no competing interests. The funder played no role in study design, data collection, and analysis, the decision to publish, or preparation of the manuscript.

Authors' contributions
The individual contributions of each author are as follows: CS, YH, JY, LH, and XW conceived and designed the study. XW supervised the study. XL, CH, JX and RX did the statistical analysis. CS, YH JX and YL contributed to data collection, analysis, and interpretation. CS, YH and XW drafted the report. All authors have read and approved the manuscript.  Tables   Table 1. Frailty criteria a Cardiovascular Health study China Health and Retirement Longitudinal Survey Weight loss Self-reported: "In the last year, have you lost more than 10 pounds unintentionally?" (not due to dieting or exercise)" (unintentional weight loss of at least 5% previous year's body weight).
Measured weight in 2015 lost more than 4.5 kilograms or weight loss of at least 5% previous year's body weight. (weight loss > 4.5 kg or 5% in two years) = 1, other = 0.
Exhaustion Self-reported (CES-D Depression Scale), The question is asked: "How often in the last week (1) did you feel that everything was an effort, or (2)   Same as the frailty phenotype indicators originally described by Fried and colleagues.
Abbreviations: BMI, body mass index; a Note: Participants met 3 or more of the above 5 criteria were de ned as frail, met 1 or 2 criteria were de ned as pre-frail, and met none of the frailty criteria were de ned as not frail or robust.  a Data were weighted and expressed as mean (SD) and compared with ANOVA. d Heart diseases included heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems e Memory-related disease includes Alzheimer's disease, Parkinson's disease, and brain atrophy.