Prevalence and influencing factors of cognitive frailty in elderly with diabetes mellitus in China: A meta-analysis

LIU Ting, SONG Mi, DIAO Zi-yan, PI Hong-ying

1. Medical School of Chinese PLA, Beijing 100853, China

2. Department of Gastroenterology, the Second Medical Center, Chinese PLA General Hospital, Beijing 100853, China

3. Air Force Hospital of Northern Theater Command, Shenyang 110042, China

4. Medical Service Training Center, Chinese PLA General Hospital, Beijing 100853, China

Keywords:

ABSTRACT Objective: To systematically review the epidemiological characteristics and influencing factors of cognitive frailty in Chinese elderly with diabetes mellitus.Methods: PubMed, Embase, Web of Science, CINAHL, Cochrane Library, China National Knowledge Infrastructure(CNKI),Wanfang and Chinese biomedical literature database(CBM) were electronically searched to collect cross-sectional studies on the prevalence and influencing factors of cognitive frailty in Chinese elderly with diabetes mellitus from inception to November 1st,2022.After quality evaluation and data extraction of included studies, R4.2.2 software was used to perform metaanalysis.Results: A total of 7 Chinese and one English literatures involving 2208 participants were included.The results of meta-analysis showed that the prevalence rate of cognitive frailty in elderly diabetic patients was 17.1%(95%CI 8.7% to 30.9%).Subgroup analysis showed that the prevalence rate of cognitive frailty was 10.2%(95%CI 5.0% to 19.7%) in old diabetic adults in community, 35.7%(95%CI 20.8% to 53.9%) in hospital, and the difference was statistically significant(P=0.01).The prevalence rate of cognitive frailty was 21.7%(95%CI 10.3% to 40.2%) in male diabetic old adults, 24.2%(95%CI 13.3% to 40.1%) in female diabetic old adults, and the difference was not statistically significant(P=0.81).The rate was 12.6%(95%CI 6.6% to 22.5%) in 60～69 years old group, 25.8%(95%CI 12.6% to 45.5%) in 70～79 years old group, 53.0%(95%CI 23.9% to 80.2%) in diabetic old adults aged 80 years and above, and the difference was statistically significant(P=0.02).The influencing factors for cognitive frailty in diabetic old adults included educational level[OR=0.230, 95%CI(0.117～0.454), P<0.000 1], regular exercise[OR=0.357, 95%CI(0.217～0.588), P<0.01], malnutrition[OR=2.372,95%CI(1.472～3.822), P=0.0004], depression[OR=3.207, 95%CI(2.156～4.768), P<0.0001],and HbA1c 7.0%[OR=3.112, 95%CI(1.880～5.152), P<0.0001].Conclusion: The prevalence of cognitive frailty in Chinese elderly patients with diabetes is high, and there are differences in cognitive frailty prevalence among different sources and different ages.Educational level,exercise habits, nutritional status, depression and HbA1c are the factors affecting cognitive frailty in elderly patients with diabetes.Depression, malnutrition, HbA1c 7.0% were risk factors, and high educational level, regular exercise were protective factors.Attention should be paid to early and accurate identification of cognitive frailty and intervention targeting influencing factors to delay or reverse its progression to adverse health outcomes.

1.Introduction

Diabetes is a common chronic disease, and its prevalence rate continues to increase globally.China has the largest number of diabetic patients, reaching 140 million in 2021, and the elderly are the main group affected by the disease[1, 2].Cognitive frailty(CF),namely the coexistence of cognitive function impairment and physical weakness, has become a new complication of senile diabetes[3], which has gradually attracted the attention of researchers in recent years.CF increases the risk of adverse events such as falls, disability, hospitalization, and death in patients with diabetes,and has a significant cumulative effect on the occurrence of adverse events compared with patients with only cognitive impairment and only physical weakness[4-7].Therefore, attention should be paid to the disease status of CF in elderly diabetic patients.Although there have been a certain number of studies on the prevalence of CF in elderly diabetic patients and its influencing factors in China,the results of these studies are quite different.This study intends to comprehensively evaluate the prevalence of CF in elderly diabetic patients and analyze its influencing factors by comprehensively searching relevant articles published by Chinese scholars, so as to provide reference for subsequent relevant studies.

2.Materials and methods

2.1 Literature sources and search strategies

A comprehensive search of PubMed, Embase, Web of Science,CINAHL, Cochrane Library, CNKI, Wanfang, and Chinese biomedical database on the prevalence of CF in elderly diabetic patients and its influencing factors was conducted.The retrieval period was from the establishment of the database to November 1,2022.The search method combining subject words and free words is adopted, and the references of the included documents are retrieved twice, and the dissertation is included as grey documents.Chinese keywords included: diabetes, cognition, frailty, cognitive asthenia;English search words include diabetes, diabetic, diabet*, cognition,cogniti*, frailty, frail*, cognitive frailty.

2.2 Criteria for inclusion and exclusion of literature

Inclusion criteria:(1) The subjects met the 1999 WHO diagnostic criteria for diabetes[8];(2) Participants were older adults aged over 60 years or studies that can extract the data of people over 60 years old;(3)Studies have a clear definition of CF and diagnostic criteria[9,10];(4) The study was carried out in mainland China;(5) Crosssectional study;(6) The language of publication shall be Chinese or English.Exclusion criteria: (1) Duplicate literatures are excluded except those with the most detailed data; (2) Incomplete data or the required data cannot be extracted from the original text; (3) The full text cannot be obtained; (4) The sample size was less than 50 cases;(5) Doubtful data.

2.3 Literature screening and data extraction

The literature will be imported into Endnote X9 and cross-checked by two people trained in evidence-based medicine according to inclusion and exclusion criteria.In case of any disagreement, a third person will decide.First read the title and the abstract for the preliminary screening, the preliminary screening of the literature reading the full text.Data extraction included: first author, year of publication, region, source of research objects, CF assessment tools,CF diagnostic criteria, sample size and number of cases, and CF influencing factors.

2.4 Literature quality assessment

The Quality of the included literature was evaluated using the cross-sectional literature quality evaluation tool recommended by the Agency for Healthcare Research and Quality (AHRQ).There were 11 items in the scale, each item was 1 point, and the total score was up to 11 points.0 to 3 is low quality, 4 to 7 is medium quality, and 8 to 11 is high quality[11].The evaluation was carried out by two researchers separately and then cross-checked.In case of disagreement, the third researcher was asked to decide.

2.5 Statistical analysis

Meta analysis was performed using R4.2.2 software.The I2 and Q tests were used to evaluate the heterogeneity.When I2> 50%, the random effects model was selected; when I2< 50%, the fixed effects model was selected.Sensitivity analysis and subgroup analysis were used to explore the sources of heterogeneity.Funnel plot was drawn and Egger test was used to check publication bias, P > 0.05 indicated no publication bias.

3.Results

3.1 Literature search results

A total of 1470 literatures were retrieved, 601 duplicate literatures were removed, 80 literatures were included after reading and abstract, 8 literatures were included after reading the full text[12-19],7 Chinese literatures and 1 English literatures, and a total of 2208 elderly diabetic patients were included.The literature screening process is shown in Figure 1.

Fig 1 Literature Screening Process

3.2 Characteristics and quality assessment results of the included literature

A total of 2208 elderly patients with diabetes were investigated in the included literature, among which 411 patients were diagnosed with CF.5 studies[12-15, 19] investigated people from the community,and 3 studies[16-18] investigated people from inpatients.The survey areas included Hangzhou in Zhejiang province, Xianning in Hubei province, Zhengzhou in Henan province, Shenyang in Liaoning province, Rugao in Jiangsu province, Jinan in Shandong province,Urumqi in Xinjiang Province and Nanning in Guangxi Province.All articles were published between 2020 and 2022.AHRQ scale was used to evaluate the literature quality, and the evaluation results were all between 6 and 9 points, indicating the overall quality of the literature was high.The results are shown in Table 1.

3.3 Meta-analysis of CF prevalence in elderly patients with diabetes

The prevalence of CF in the included studies ranged from 3.125% to 56.5%.Heterogeneity analysis showed that there was heterogeneity among the included elderly diabetic patients (I2=97%, Q=233.18, P< 0.01).Random-effects model was used.The prevalence rate of CF in elderly diabetic patients was 17.1%(8.7%～30.9%), as shown in Figure 2.

Fig 2 Forest plot of the prevalence of CF in elderly patients with diabetes

Tab 1 Characteristics and quality evaluation of included literature

3.4 Subgroup analysis

Results of subgroup analysis of the prevalence of CF are shown in Table 2 Subgroup analysis by setting showed heterogeneity in the prevalence of CF in both community and hospital elderly patients with diabetes (community: I2=95%, Q=85.94, P < 0.01;Hospitalization: I2=97%, Q=68.62, P < 0.01), choosing random effects model.The results showed that the prevalence of CF was 10.2%(95%CI :5.0%～19.7%) in elderly diabetic patients in the community and 35.7%(95%CI :20.8%～53.9%) in hospitalized elderly diabetic patients, and the difference was statistically significant (P=0.0037).Shown in Figure 3.

Subgroup analysis by gender showed heterogeneity in the prevalence of CF in both male and female elderly patients with diabetes (male: I2=95%, Q=96.3, P < 0.01; Female: I2=93%, Q=70.5,P < 0.01), choosing random effects model.The results showed that the prevalence of CF in elderly male patients with diabetes mellitus was 21.7%(95%CI :10.3%～40.2%) and that in elderly female patients with diabetes mellitus was 24.2%(95%CI :13.3%～40.1%),and the difference was not statistically significant (P=0.81).Shown in Figure 4.

Subgroup analysis by age showed heterogeneity in the prevalence of CF in elderly diabetic patients aged 60～69, 70～79 and over 80 years old in the literature included (60～69 years old :I2=89%,Q=28.21, P < 0.01; 70～79 years old: I2=89%, Q=26.79, P < 0.01;80 years old: I2=87%, Q=23.71, P < 0.01), choosing random effects model.The results showed that the prevalence of CF was 12.6%(95%CI :6.6%～22.5%) in elderly patients aged 60～69,25.8%(95%CI :12.6%～45.5%) in elderly patients aged 70～79, and 53.0%(95%CI :23.9%～80.2%) in elderly patients aged over 80.The difference was statistically significant (P=0.02).Shown in Figure 5.

3.5 Sensitivity analysis and publication bias

Sensitivity analysis showed that the meta-analysis results had great stability, as shown in Figure 6.Draw the funnel diagram, there is no obvious asymmetry was observed, as shown in Figure 7.The Egger test (t = -2.31, P= 0.0605) showed that there was no publication bias in the included literature of this study.

3.6 Meta-analysis of influencing factors of CF in elderly patients with diabetes

Six literatures were included for meta-analysis of influencing factors, with a total of 9 independent influencing factors.There were 3[14-16] articles that involved age; In two[12, 15] articles, education level was involved; Three articles[12, 14, 16] involved regular exercise.In 5[12-15, 17] articles, depression was involved; Two articles[12, 17]involved malnutrition.In two[13, 17]papers, the influencing factors involved HbA1c; One article[17] was related to the duration of diabetes.One article[13] involved personal monthly income level and insufficient sleep time at night.Meta-analysis was performed on the influencing factors involved in two or more articles, while only descriptive analysis was performed for one article.Results of metaanalysis showed that malnutrition [OR=2.372, 95%CI(1.472-3.822),P=0.000 4], depression [OR=3.207, 95%CI(2.156-4.768), P < 0.000 1], HbA1c 7.0% [OR=3.112, 95%CI(1.880-5.152), P < 0.000 1] is a risk factor for CF in elderly patients with diabetes, regular exercise[OR=0.357, 95%CI(0.217-0.588), P < 0.000 1], high education[OR=0.230, 95%CI(0.117～0.454), P < 0.0001] was a protective factor for CF in elderly patients with diabetes.Shown Table 3.

Fig 3 Forest plot of prevalence of CF in diabetic elderly patients from different sources

Fig 4 Forest plot of prevalence of CF in diabetic elderly patients with different gender

Fig 5 Forest plot of prevalence of CF in diabetic elderly patients of different ages

Fig 6 Results of sensitivity analysis of CF prevalence in Chinese elderly patients with diabetes

Tab 2 Subgroup meta-analysis of prevalence of CF in elderly patients with diabetes

Fig 7 Funnel plot of the included literature

Tab 3 Meta-analysis of influencing factors of CF in elderly diabetic patients

Only one[17]paper showed that the diabetes duration over 10 years was associated with a higher risk of CF: 10～20 years[OR=4.598, 95%CI(2.051～10.309), P < 0.01], 21～29 years[OR=3.429, 95%CI(1.355～8.678), P < 0.01], 30 years [OR=6.898,95%CI(2.172～21.909), P < 0.01].One article[13]concluded that monthly income 1 000 yuan is associated with low risk of CF, and night sleep duration < 5 h is associated with high risk of CF:Individual monthly income 1 000～2 000 yuan [OR=0.199,95%CI(0.045～0.887), P=0.034], individual monthly income > 2 000 yuan [OR=0.216, 95%CI(0.059～0.785) , P=0.020]; Night sleep duration < 5 h: [OR=2.818, 95%CI(1.022～7.774) , P=0.045].

4.Discussion

4.1 Prevalence of CF in elderly patients with diabetes

In this study, the prevalence rate of CF in Chinese elderly diabetic patients was 17.1%, and the prevalence rate of CF in hospitalized elderly diabetic patients (35.7%) was higher than that in community elderly diabetic patients (10.2%).There was no difference in the prevalence rate of CF in elderly diabetic patients of different genders,but there were differences in the prevalence rate of CF in elderly diabetic patients of different ages.It shows a tendency to increase gradually with age.The meta-analysis of 16 articles included by Lin Yuzhu et al[20] showed that the incidence of CF among the elderly in China was 5.0%, and the risk of CF among the elderly with a history of diabetes was 1.73 times that of the elderly without a history of diabetes.Qu et al[21] conducted a meta-analysis of factors affecting CF in the elderly, which also showed that diabetes was a risk factor for CF in the elderly, and the risk was 1.47 times that of non-diabetic elderly.The prevalence rate of CF in middle-aged and elderly diabetic patients in this study was higher than that of Lin Yuzhu et al[22] and Qu Qian et al[21].The reason may be that the studies included in this study were all cross-sectional studies with elderly diabetic patients as research objects, and the sample size was relatively sufficient.However, in the meta-analysis of Lin Yuzhu et al[22] and Qu Qian et al[21], the research objects included in the literature were the elderly, and diabetes patients were only part of the population, so the sample size was relatively small, which may lead to the underestimated prevalence of CF.In addition, as the diagnostic criteria for CF have not been unified at present, there are differences in the diagnostic criteria adopted by various studies, so the difference in the prevalence of CF may be caused by the selection of different diagnostic criteria.Lyu et al[23] included a total of 15 studies in a meta-analysis and concluded that the prevalence of CF in elderly diabetic patients in the community was 11%, which was close to the prevalence of CF in elderly diabetic patients in the Chinese community in this study.

4.2 Influencing factors of CF in elderly patients with diabetes

4.2.1 Age

Several studies have shown that the risk of CF increases with age[21,22, 24-26].In this study, age is not an independent influencing factor of CF, which may be due to the small number of literatures included in this study and only 3 influential factors involving age.Therefore,the influence of age on CF could not be analyzed.Moreover, there is heterogeneity among the articles included in the analysis of ageinfluencing factors, and publication bias exists according to Egger test.

4.2.2 Regular exercise

The results of this study showed that regular exercise was a protective factor for CF in elderly diabetic patients, reducing the risk of CF development by 64.3%.Exercise can improve the physical and cognitive functions of the elderly[27].Zhang Shuang et al[28]used comprehensive exercise intervention measures to intervene the community elderly with diabetes and frailty for 12 weeks, and the results showed that the frailty state and physical function of the elderly were significantly improved.Ye Ming et al[29] used multicomponent exercise to intervene 90 elderly patients with CF, and the patients’ cognitive function and frailty were significantly improved after 3 and 6 months.Otago exercise[30], resistance exercise[31] and structured moderate intensity physical exercise[32] have also been proved to be effective in CF intervention.Therefore, regular exercise should be encouraged in elderly diabetic patients to reduce the risk of CF.

4.2.3 Malnutrition

In this study, the prevalence of CF in elderly diabetic patients with malnutrition was 2.372 times that of those with normal nutritional status, which was consistent with the results of Kwan et al[33] and Seesen et al[34].Chye et al[35] conducted a study on the association between malnutrition and CF in 5414 cases of people over 55 years old in longitudinal aging research in Singapore.The study showed that the risk of malnutrition in CF patients was 8.16 times that of non-CF patients, indicating that malnutrition was closely related to CF.In 2013, an expert panel composed of the International Institute of Nutrition and Aging and the International Geriatrics Society pointed out that a healthy diet (such as the Mediterranean diet) can be used as an intervention for CF in the elderly[9].A study on the correlation between food diversity and CF conducted by scholars in Taiwan indicated that the intake of more protein-rich foods, such as dairy products, fish and other seafood, was associated with a lower prevalence of CF[36].At present, the number of CF intervention studies is still small, and there are few studies related to CF nutrition intervention in diabetic patients.However, in epidemiological studies, it has been confirmed that reduced and excessive energy intake, insufficient intake of protein, B vitamins, vitamin D and omega-3 fatty acids are associated with frailty, cognitive decline or increased mortality[37].

4.2.4 Depression

Depression is relatively common in patients with diabetes.A meta-analysis involving 48 cross-sectional studies showed that the prevalence of depression in Chinese patients with type 2 diabetes was 25.9%%[38].In this study, the risk of CF in patients complicated with depression was 3.207 times that of non-depressed patients,which was consistent with the conclusions of Kwan et al[33] and Zou et al[39].Depression, physical frailty and cognitive impairment may have common pathophysiological mechanisms such as oxidative stress, chronic inflammation and mitochondrial dysfunction.For example, chronic inflammation may cause the decline of muscle mass, strength and dopamine function, thus leading to body weakness, depression and cognitive impairment[26, 33].It is therefore possible that interventions for depression may reduce the risk of CF.

4.2.5 HbA1c

This study showed that HbA1c was an independent factor affecting CF in elderly patients with diabetes, and the risk of CF increased by 2.112 times when HbA1c 7.0%.Two meta-analyses have suggested that HbA1c is a factor affecting cognitive impairment and frailty in elderly diabetic patients[40, 41].It may be related to the damage of vascular endothelium caused by the fluctuation of blood glucose level, the increase of oxidative stress response, and the accumulation of excessive glycation products caused by long-term hyperglycemia in the body, thus inducing a series of inflammatory reactions[42].

4.2.6 Other factors

In this study, elderly diabetic patients with high education level had a 77% lower risk of CF than those with low education level, which may be because the elderly with high education level have better medical conditions, better control of diabetes and other chronic diseases, and stronger awareness of health care.

In conclusion, the prevalence of CF in elderly patients with diabetes is high, malnutrition, depression and high HbA1c level are risk factors for the incidence of CF in elderly patients with diabetes, and regular exercise and high education are protective factors.Screening of CF in elderly patients with diabetes should be carried out early,and targeted intervention measures should be developed according to the influential factors that can be interfered with.Special attention should be paid to nutritional and exercise interventions to delay the onset of CF and delay or prevent its progression to adverse outcomes such as dementia.

This paper mainly has the following limitations and deficiencies:(1) Due to the late research in related fields, the number of included literatures and sample size are still insufficient; (2) There was a large heterogeneity among articles, and there was still a large heterogeneity in subgroup analysis; (3) Only 2 articles were included in the meta-analysis of some influencing factors, and the extrapolation of the results remains to be verified.

Authors’ contribution

Liu Ting: Responsible for topic selection, literature retrieval and screening, data extraction, statistical analysis and article writing;Song Mi: Literature screening, data extraction; Diao Ziyan: Revised article; PI Hongying: Feasibility assessment of topic selection, article revision.

Conflict of interest

There is no conflict of interest.