Abstract

Background

Obesity has been arousing a critical public health issue, and posting threats to senior population. We aimed to explore gender differences in associations between general/central obesity (body mass index/waist circumference) and hypertension, diabetes, dyslipidemia based on electronic health records of 3.5听million Chinese senior population over 65 years.

Methods

3571189 electronic health records of Chinese senior population were collected from platform of Zhejiang provincial Basic Public Health Services Project. Sociodemographic characteristics, behavioral lifestyle, physical data, and biochemical indices were included in the research. Multivariate logistic regression models and restricted cubic spline models were used to explore associations between obesity and diseases.

Results

7.7% (5.3% for male, 9.7% for female) senior population were having general obesity, and 31.8% (25.2% for male, 37.4% for female) of them had central obesity. 48.0% (46.0% for male, 49.7% for female), 14.0% (12.3% for male, 15.5% for female), and 58.9% (51.6% for male, 65.2% for female) senior population were having hypertension, diabetes, and dyslipidemia, respectively. 37.9% (29.5% for male, 45.1% for female) and 33.7% (28.1% for male, 38.5% for female) senior population were having abnormal TC and TG, respectively. General obesity and central obesity strongly and negatively associated with hypertension [OR鈥�=鈥�2.61 (95%CI: 2.58鈥�2.63), and 2.20 (95%CI: 2.18鈥�2.21)], diabetes [OR鈥�=鈥�1.33 (95%CI: 1.31鈥�1.35), and 1.56 (95%CI: 1.54鈥�1.57)], and dyslipidemia [OR鈥�=鈥�1.66 (95%CI: 1.64鈥�1.68), and 1.84 (95%CI: 1.83鈥�1.85)] based on existing obesity standards. Male population with BMI higher than 28.7听kg/m², 30.1听kg/m², 22.7听kg/m², and with WC higher than 99.0听cm, 95.9听cm, 82.1听cm, while female population with BMI higher than 26.9听kg/m², 23.3听kg/m², 18.1听kg/m², and with WC higher than 92.1听cm, 83.1听cm, 65.7听cm, the ORs were over 1.0 for having hypertension, diabetes, and dyslipidemia, respectively.

Conclusions

Senior population were more likely to have central obesity over general obesity, and nearly half of them were having hypertension and dyslipidemia. Obesity negatively and strongly associated with chronic diseases in senior population, yet general obesity exerted more impact on hypertension, whereas central obesity exerted more impact on diabetes and dyslipidemia. Female population with obesity were in higher risk than male having hypertension, diabetes, and dyslipidemia. We recommended senior population control BMI lower than 28.7听kg/m² and 23.3听kg/m², as well as WC lower than 95.9听cm and 83.1听cm for male and female, respectively. Optimal BMI and WC in senior population may be around the overweight or mild obesity range. There were risks for having dyslipidemia or abnormal lipid-related indices even in senior population without obesity. TC and TG were major indicators of discovering disease and preventing senior population from dyslipidemia.

Introduction

Overweight and obesity are defined as abnormal or excessive fat accumulation [1], which have been arousing a pandemic as a critical worldwide public health issue. According to some World Health Organization (WHO) global estimates in recent years, the prevalence of obesity almost tripled from 1975 to 2016 worldwide, inducing 39% and 13% of adults aged 18 years and over were overweight and obese respectively in 2016 [1]. Nearly the same results were reported in line with the data from Chinese Residents鈥� Nutrition and Chronic Disease Status Report, people with overweight and obesity accounted for 34.3% and 16.4% individually in adults aged 18 years and older in 2020 [2]. In addition, overweight and obesity post threats and impair people鈥檚 health, senior population and people with cardio-metabolic risk factors in particular, leading to severer hypertension, diabetes, dyslipidemia, cardiovascular diseases (e.g., heart disease and stroke), musculoskeletal disorders (e.g., osteoarthritis), some cancer, and even death [1].

Body mass index (BMI), the most classic and practical index measuring overweight and general obesity, was proposed by Lambert Adolphe Francois Quetelet, the 19th century Belgian polymath, statistician and astronomer, who validated a mathematical way to estimate human body size independent of height [3, 4]. However, its limitations prevent BMI from distinguishing the volume and distribution of body fat. When body fat accumulates much in a specific part of body, viscus and belly in particular, waist circumference (WC) is being referred to measure central obesity in this case. Excessive accumulations and abnormal distribution of fat associate much stronger with obesity related diseases.

Obesity intimately associates with cardio-metabolic risk factors - hypertension, diabetes, and dyslipidemia - in biological mechanism, as well as in population epidemiology. Obesity is mainly driven by the effects of insulin resistance in peripheral tissues and pro-inflammatory adipokines in adipose tissue, elevating fasting plasma total cholesterol (TC), triglycerides (TG), low density lipoprotein cholesterol (LDL-C), fasting blood glucose (FBG) and insulin levels, and high systolic blood pressure (SBP) and/or diastolic blood pressure (DBP), meanwhile decreasing high density lipoprotein cholesterol (HDL-C) [5, 6].

Negative and statistically significant associations were observed in some epidemiology researches. Liu L.J. found that obese subjects had a 3.9-fold higher risk of being hypertensive (relative risk 4.9; 95%CI: 3.4 to 7.3) compared with those subjects who had a BMI less than 25听kg/m² in Chinese people aged 48 to 56 years [7]. Dang S.N. & Yan H. found that the BMI 鈥楽ubstantial gain鈥� pattern was associated with the hazard of type 2 diabetes after adjusting for potential covariates with HR varying from 1.36 (95%CI: 1.01 to 1.84, p鈥�=鈥�0.047) to 1.50 (95%CI: 1.10 to 2.02, p鈥�=鈥�0.010) from their 20-year long-term survey in Chinese adults [8]. Besides, Jiang Y.G. & Zhao G.M. found that the association between obesity and high TC, high TG, high LDL-C, and low HDL-C, with the adjusted odds ratios ranging from 1.11 to 2.00 in Chinese adults aged 20 to 74 years [9].

Electronic health records are sources with high quality for collecting health data, such as physiological and biochemical indices. China National Basic Public Health Services Project (BPHSP) is the most important health project with highest expenditure from Ministry of Health China and health department of local government, which ask all primary level medical and healthcare centers to establish electronic health records for each resident to manage people鈥檚 health since 2009. The statistics and rates based on official platform and huge population are relatively objective, which are included in the assessment of local government and primary level medical and healthcare centers.

Take Zhejiang province for instance, a district with strong economy locates in the east of China, electronic health records (EHRs) of 54.4听million residents have been established by the end of 2022, including 16.5听million major population, i.e., 0鈥�6 years children, pregnant and parturient women, and 鈮モ��65 years senior population, as well as 7.4听million major patients, i.e., 鈮モ��35 years patients with hypertension, type 2 diabetes mellitus, severe mental disorders, and pulmonary tuberculosis. Statistically, 67.4% senior population have been under health management, which involves lifestyle and health status evaluation, physical examination and auxiliary examination, and health guidance. Plus, 70.8% and 70.4% patients with hypertension and diabetes respectively have been standardized managed, consisting of disease screening, follow-up evaluation and classified intervention, and medical examination. Zhejiang province is managing EHRs in the leading place among other provinces in China.

Actually, there were abundant of previous researches concerning relationship between obesity and chronic disease in all kinds of degrees and angles. However, most of them studied general obesity with only one kind of disease in the research with a relatively small general sample. Consequently, we are going to explore the gender difference in associations between general/central obesity and hypertension, diabetes, dyslipidemia in the research at a time with millions of specific population samples-senior population, and to provide scientific evidences for preventing chronic diseases for senior population.

Methods

Study design and data

A cross-sectional study was conducted with Zhejiang provincial EHRs of 3.5听million鈥夆墺鈥�65 years senior population in 2022. Sociodemographic characteristics (e.g., gender, age, education, region), behavioral lifestyle (e.g., smoking, drinking, physical exercise), obesity measuring physical data (e.g., height, weight, WC, SBP, DBP), and biochemical indices (e.g., FBG, TC, TG, LDL-C, HDL-C) of senior population were collected from EHRs platform of Zhejiang provincial BPHSP.

The study merely utilized and analyzed the data of EHRs. Doctors at primary level medical and healthcare centers take the responsibilities to conduct the questionnaire and to organize local senior population to attend the physical examination collectively, or senior population themselves went to the closest local center in their spare time to attend the physical examination. For each senior adult older than 65 years, having medical examination at least once a year is required. The questionnaire and examination procedures were consistently regulated by Ministry of Health China and health department of provincial government. The collection and test of biological samples complied to the national norms of medical practice.

Inclusion and Exclusion Criteria

Standardized EHRs, which followed the standards of quality control for EHR of Zhejiang province [10], of 鈮モ��65 years senior population were included in the study. We excluded EHRs without medical examination data, and EHRs that too many values of important research variables were missing, special characters instead of numbers, numbers beyond normal range, or the like.

Definitions of variables

BMI was calculated as body weight (kg) divided by the square of height (m²), and was categorized as lean (<鈥�18.5听kg/m²), normal (18.5鈥墌鈥�24.0听kg/m²), overweight (24.0鈥墌鈥�28.0听kg/m²), and general obesity (鈮モ��28.0听kg/m²) [11]. WC was categorized as normal (<鈥�85.0听cm for male, <鈥�80.0听cm for female), pre-obesity (85.0鈥墌鈥�90.0听cm for male, 80.0鈥墌鈥�85.0听cm for female), and central obesity (鈮モ��90.0听cm for male, 鈮モ��85.0听cm for female) based on Overweight and Obesity Prevention and Control Guidelines in Chinese Adults (2021) [11].

Hypertension and type 2 diabetes had been labeled in the EHRs of senior population who were under health management in primary medical and healthcare centers. The criteria for hypertension management were SBP鈥夆墺鈥�140 mmHg and/or DBP鈥夆墺鈥�90 mmHg according to Chinese Guidelines for Hypertension Prevention and Treatment (2018) [12], and the criterion for type 2 diabetes were FBG鈥夆墺鈥�7.0 mmol/L according to Chinese Guidelines for type 2 diabetes Prevention and Treatment (2020) [13]. Dyslipidemia was defined as any of the following abnormalities, TC鈥夆墺鈥�5.2 mmol/L, TG鈥夆墺鈥�1.7 mmol/L, LDL-C鈥夆墺鈥�3.4 mmol/L, HDL-C鈥�<鈥�1.0 mmol/L, according to the Chinese Guidelines for Lipid Management (2023) [14].

As for covariates, age was presented as groups of 66鈥墌鈥�70 years, 71鈥墌鈥�75 years, 76鈥墌鈥�80 years, and 鈮モ��81 years; gender was presented as male and female; education was presented as primary school or below, middle school, and high school or above; region was presented as urban and rural area; smoking was presented as non-smoker, previous smoker and current smoker; dringking and physical exercise were presented as never, sometimes, and often.

Statistical analysis

Research variables derived from EHRs provincial platform were stored in Microsoft Excel software, and statistical analysis was performed with R software. Descriptive statistics were presented as Means鈥壜扁�塖tandard Deviation (SD) for continuous variables, and frequencies (percentages) for categorical variables. T-tests were used to compare differences in means of normal distributed values of BMI, WC, SBP, DBP, FBG, TC, TG, LDL-C, and HDL-C between groups of gender. Chi-square tests were used to compare differences in percentages of BMI categories, WC categories, hypertension, diabetes, dyslipidemia, as well as abnormal TC, TG, LDL-C, and HDL-C between groups of gender.

Multivariate logistic regression analyses were used to explore the adjusted Odds Ratios (ORs), representing risks of hypertension, diabetes, and dyslipidemia in corresponding level of BMI and WC, after adjusted by covariables, i.e., gender, age, education, area, smoking, drinking, physical exercise, as well as hypertension, diabetes, and dyslipidemia for corresponding diseases. Restricted cubic spline models fitted for multivariate logistic regression models with 7 knots between the 2.5th to 97.5th percentiles of BMI and WC, which were used to explore the continuous ORs trends of diseases with rising BMI and WC for male and female. All tests were two-tailed, and p鈥�<鈥�0.05 was considered statistically significant.

Results

Characteristics of senior population

3,571,189 senior population were included in the study. As shown in Table听1, their mean age was 73.75鈥壜扁��6.11 years. 54.3% (n鈥�=鈥�1,937,899) of them were female. 76.6% (n鈥�=鈥�2,735,114) of them were at primary school education level or below. 53.5% (n鈥�=鈥�1,917,649) of them came from rural area. 14.8% (n鈥�=鈥�2,858,412) of them were current smoker. 12.2% (n鈥�=鈥�437,171) of them claimed 鈥渙ften鈥� drinking. 18.3% (n鈥�=鈥�652,145) of them claimed 鈥渙ften鈥� keeping physical exercise.

Table 1 Background characteristics of senior population included in the study (N鈥�=鈥�3,571,189)

The prevalence of diseases and abnormal lipid-related indices between genders

As shown in Table听2, the mean BMI of senior population was 23.44鈥壜扁��3.11听kg/m²; 32.9% (n鈥�=鈥�1,174,717) of them were 鈥渙verweight鈥�, and 7.7% (n鈥�=鈥�274,704) of them were 鈥済eneral obesity鈥�. Their mean WC were 83.98鈥壜扁��8.51听cm for male and 82.23鈥壜扁��8.68听cm for female; 24.5% (n鈥�=鈥�873,491) of them were 鈥減re-obesity鈥�, and 31.8% (n鈥�=鈥�1,135,881) of them were 鈥渃entral obesity鈥�. Their mean SBP was 138.91鈥壜扁��16.94 mmHg, and mean DBP was 79.11鈥壜扁��9.78 mmHg; 48.0% (n鈥�=鈥�1,713,969) of them were having hypertension. Their mean FBG was 5.51鈥壜扁��0.76 mmol/L; 14.0% of them were having diabetes (n鈥�=鈥�500,980). Their mean TC was 4.93鈥壜扁��1.03 mmol/L, mean TG was 1.46鈥壜扁��0.68 mmol/L, mean LDL-C was 2.81鈥壜扁��0.87 mmol/L, and mean HDL-C was 1.45鈥壜扁��0.39 mmol/L; 37.9% (n鈥�=鈥�1,354,767), 33.7% (n鈥�=鈥�1,204,375), 24.3% (n鈥�=鈥�867,363), and 8.7% (n鈥�=鈥�309,663) of them were having abnormal TC, TG, LDL-C, and HDL-C, respectively; 58.9% (n鈥�=鈥�2,103,926) of them were having dyslipidemia.

Male population were significantly higher (p鈥�<鈥�0.001) in means of indices of WC (83.98鈥壜扁��8.51听cm) and DBP (79.73鈥壜扁��9.81 mmHg), as well as in prevalences of overweight (33.6%) and abnormal HDL-C (10.8%) than female.

Female population were significantly higher (p鈥�<鈥�0.001) in means of indices of BMI (23.60鈥壜扁��3.28听kg/m²), SBP (140.01鈥壜扁��17.04 mmHg), FBG (5.53鈥壜扁��0.76 mmol/L), TC (5.11鈥壜扁��1.03 mmol/L), TG (1.54鈥壜扁��0.68 mmol/L), LDL-C (2.91鈥壜扁��0.88 mmol/L), and HDL-C (1.48鈥壜扁��0.38 mmol/L) than male, as well as in prevalences of general obesity (9.7%), central obesity (37.4%), hypertension (49.7%), diabetes (15.5%), abnormal TC (45.1%), abnormal TG (38.5%), abnormal LDL-C (28.6%), and dyslipidemia (65.2%).

Table 2 The prevalence of diseases and abnormal lipid-related indices of senior population

The prevalence of diseases and abnormal lipid-related indices in units of BMI and WC

As shown in Fig.听1, Tables听3 and 4, prevalences of diseases and abnormal lipid-related indices ascended from the beginning of BMI鈥�<鈥�18听kg/m², as well as WC鈥�<鈥�81听cm for male, and WC鈥�<鈥�76听cm for female, and ended at BMI鈥�<鈥�32听kg/m², as well as WC鈥�<鈥�95听cm for male, and WC鈥�<鈥�90听cm for female, where the corresponding prevalences of diseases and abnormal lipid-related indices close to whole male or female subjects.

Among those diseases, the prevalences of dyslipidemia, from 30.5 to 51.3% for male BMI, from 48.6 to 65.2% for female BMI, from 40.6 to 49.8% for male WC, and from 57.2 to 63.9% for female WC respectively, were significantly higher (p鈥�<鈥�0.001) than those of hypertension and diabetes as the units of BMI and WC increased.

The prevalences of abnormal TC, from 21.9 to 29.5% for male BMI, from 39.7 to 45.3% for female BMI, from 26.9 to 29.3% for male WC, and from 44.1 to 45.3% for female WC respectively, were highest (p鈥�<鈥�0.001) among those abnormal lipid-related indices. The prevalences of abnormal TG ascended prominently from 10.7 to 27.7% for male BMI, from 18.4 to 38.4% for female BMI, from 17.2 to 26.1% for male WC, and from 26.3 to 36.0% for female WC respectively, comparing to those of abnormal TC, LDL-C, and HDL-C as the units of BMI and WC increased.

Female population were significantly higher (p鈥�<鈥�0.001) in the prevalences of hypertension, diabetes, dyslipidemia, abnormal TC, TG, and LDL-C than male as the units of BMI and WC increased, while the prevalences of abnormal HDL-C for male population were significantly higher (p鈥�<鈥�0.001) than those for female as the units of BMI and WC increased.

In the 鈥渓ean鈥� and 鈥渘ormal鈥� category levels of BMI, the prevalences of diseases and abnormal lipid-related indices ascended remarkably as the units of BMI increased, whereas those ascended steadily in all three category levels of WC as the units of WC increased.

Fig. 1

Prevalences of hypertension, diabetes, dyslipidemia, as well as abnormal TC, TG, LDL-C, and HDL-C as the units of BMI and WC increased for male and female of Chinese senior population. Dotted lines were prevalence of diseases and abnormal biomedical indices for whole male or female population. Light grey and dark grey background denoted different category levels of BMI and WC

Table 3 Prevalences of hypertension, diabetes, dyslipidemia, as well as abnormal TC, TG, LDL-C, and HDL-C as the units of BMI increased for male and female of Chinese senior population

Table 4 Prevalences of hypertension, diabetes, dyslipidemia, as well as abnormal TC, TG, LDL-C, and HDL-C as the units of WC increased for male and female of Chinese senior population

The prevalence of diseases and abnormal lipid-related indices in category levels of BMI and WC

As shown in Fig.听2, the prevalences of dyslipidemia were 31.3%, 46.6%, 59.8%, and 65.8% in respective category levels of male BMI (p鈥�<鈥�0.001 for differences), as well as 49.7%, 63.1%, 69.4%, and 70.5% in respective category levels of female BMI (p鈥�<鈥�0.001 for differences). The prevalences of dyslipidemia were 44.0%, 56.1%, and 63.5% in respective category levels of male WC (p鈥�<鈥�0.001 for differences), as well as 59.9%, 66.2%, and 70.0% in respective category levels of female WC (p鈥�<鈥�0.001 for differences). The prevalences of dyslipidemia were significantly higher (p鈥�<鈥�0.001) than those of hypertension and diabetes in category levels of BMI and WC.

The prevalences of hypertension ascended notably by 37.6% for male, and 34.6% for female as category levels of BMI increased, as well as ascended notably by 22.9% for male, and 19.8% for female as category levels of WC increased.

The prevalences of abnormal TG ascended prominently, from 11.2 to 44.3% for male BMI, and from 20.3 to 41.2% for male WC, as well as 19.5鈥�49.1% for female BMI, and from 29.8 to 46.9% for female WC, as category levels of BMI and WC increased.

Fig. 2

Prevalences of hypertension, diabetes, dyslipidemia, abnormal TC, abnormal TG, abnormal LDL-C, and abnormal HDL-C as the category levels of BMI and WC rose for male and female of Chinese senior population. Light grey and dark grey background denoted different category levels of BMI and WC

Risks of hypertension, diabetes, and dyslipidemia based on existing obesity standards

For all senior population as shown in Tables听 and 6, the risks of obese population having hypertension were 2.6081 (95%CI: 2.5816鈥�2.6349, p鈥�<鈥�0.001) and 2.1978 (95%CI: 2.1844鈥�2.2112, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having hypertension for per unit increased of BMI and WC were 1.1384 (95%CI: 1.1374鈥�1.1394, p鈥�<鈥�0.001) and 1.0448 (95%CI: 1.0445鈥�1.0451, p鈥�<鈥�0.001), respectively. The risks of obese population having diabetes were 1.3314 (95%CI: 1.3147鈥�1.3483, p鈥�<鈥�0.001) and 1.5582 (95%CI: 1.5445鈥�1.5721, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having diabetes for per unit increased of BMI and WC were 1.0478 (95%CI: 1.0466鈥�1.0491, p鈥�<鈥�0.001) and 1.0262 (95%CI: 1.0257鈥�1.0266, p鈥�<鈥�0.001), respectively. The risks of obese population having dyslipidemia were 1.6584 (95%CI: 1.6417鈥�1.6752, p鈥�<鈥�0.001) and 1.8320 (95%CI: 1.8310鈥�1.8330), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having diabetes for per unit increased of BMI and WC were 1.0975 (95%CI: 1.0966鈥�1.0985, p鈥�<鈥�0.001) and 1.0366 (95%CI: 1.0363鈥�1.0369, p鈥�<鈥�0.001), respectively.

For male population, the risks of obese population having hypertension were 2.6632 (95%CI: 2.6160鈥�2.7112, p鈥�<鈥�0.001) and 2.3045 (95%CI: 2.2830鈥�2.3261, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having hypertension for per unit increased of BMI and WC were 1.1522 (96%CI: 1.1506鈥�1.1538, p鈥�<鈥�0.001) and 1.0483 (95%CI: 1.0478鈥�1.0488, p鈥�<鈥�0.001), respectively. The risks of obese population having diabetes were 1.4182 (95%CI: 1.3861鈥�1.4510) and 1.5939 (95%CI: 1.5721鈥�1.6159, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having diabetes for per unit increased of BMI and WC were 1.0644 (95%CI: 1.0621鈥�1.0666, p鈥�<鈥�0.001) and 1.0298 (95%CI: 1.0291鈥�1.0306, p鈥�<鈥�0.001), respectively. The risks of obese population having dyslipidemia were 2.1880 (95%CI: 2.1506鈥�2.2260, p鈥�<鈥�0.001) and 2.1772 (95%CI: 2.1577鈥�2.1969, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having dyslipidemia for per unit increased of BMI and WC were 1.1402 (95%CI: 1.1386鈥�1.1417, p鈥�<鈥�0.001) and 1.0478 (95%CI: 1.0474鈥�1.0483, p鈥�<鈥�0.001), respectively.

For female population, the risks of obese population having hypertension were 2.5585 (95%CI: 2.5267鈥�2.5907, p鈥�<鈥�0.001) and 2.0844 (95%CI: 2.0675鈥�2.1014, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having hypertension for per unit increased of BMI and WC were 1.1286 (95%CI: 1.1273鈥�1.1298, p鈥�<鈥�0.001) and 1.0413 (95%CI: 1.0407鈥�1.0419, p鈥�<鈥�0.001), respectively. The risks of obese population having diabetes were 1.2703 (95%CI: 1.2512鈥�1.2897, p鈥�<鈥�0.001) and 1.4867 (95%CI: 1.4694鈥�1.5041, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having diabetes for per unit increased of BMI and WC were 1.0378 (95%CI: 1.0363鈥�1.0393, p鈥�<鈥�0.001) and 1.0224 (95%CI: 1.0219鈥�1.0230, p鈥�<鈥�0.001), respectively. The risks of obese population having dyslipidemia were 1.4090 (95%CI: 1.3916鈥�1.4267, p鈥�<鈥�0.001) and 1.5917 (95%CI: 1.5788鈥�1.6046, p鈥�<鈥�0.001), comparing with whom were having 鈥渘ormal鈥� BMI and WC respectively; the risks of having dyslipidemia for per unit increased of BMI and WC were 1.0690 (95%CI: 1.0679鈥�1.0702, p鈥�<鈥�0.001) and 1.0269 (95%CI: 1.0264鈥�1.0273), respectively.

Table 5 Adjusted ORs of hypertension, diabetes, and dyslipidemia in corresponding level of BMI in Chinese senior population

Table 6 Adjusted OR of hypertension, diabetes, and dyslipidemia in corresponding level of WC in Chinese senior population

Risks of hypertension, diabetes, and dyslipidemia in continuous rising trends

As shown in Fig.听3, relationships were non-linear (p鈥�<鈥�0.001) between BMI/WC and continuous ORs, which ascended with the rising of BMI/WC. With the BMI ascended higher than the point of 28.7听kg/m², 30.1听kg/m², and 22.7听kg/m² for male, as well as 26.9听kg/m², 23.3听kg/m², and 18.1听kg/m² for female, the ORs of hypertension, diabetes, and dyslipidemia were higher than 1.0. With the WC ascended higher than the point of 99.0听cm, 95.9听cm, and 82.1听cm for male, as well as 92.1听cm, 83.1听cm, and 65.7听cm for female, the ORs of hypertension, diabetes, and dyslipidemia were higher than 1.0.

Fig. 3

Adjusted ORs with the increased units of BMI and WC for male and female subjects with hypertension, diabetes, and dyslipidemia (ORs were adjusted by gender, age, education, area, smoking, drinking, physical exercise, as well as hypertension, diabetes, and dyslipidemia for corresponding diseases)

Discussion

The study explored gender differences in associations between obesity and hypertension, diabetes, dyslipidemia based on electronic health records of 3.5听million Chinese senior population older than 65 years. Sociodemographic characteristics, behavioral lifestyle, physical data and biochemical indices of subjects were collected from EHRs for further model analyses. We used multivariate logistic regression models and restricted cubic spline models to measure the risks with the values of adjusted OR, and to fit non-linear relationships between BMI/WC and ORs. We then offered recommendations for diseases prevention and health promotion based on model analyses and research evidences.

Senior population were more likely to accumulate body fat in waist and belly area leading to central obesity instead of general obesity, and they were vulnerable for having hypertension and dyslipidemia. The prevalences of obesity and chronic diseases presented great gender differences. We found that the 31.8% senior population having central obesity, which higher than having general obesity (7.7%). 37.4% female population having central obesity, whereas 25.2% for male. Besides that, we found that the 48.0% and 58.9% senior population having hypertension and dyslipidemia, respectively. 49.7% and 65.2% female population having hypertension and dyslipidemia respectively, while 46.0% and 51.6% for male respectively.

General obesity and central obesity strongly and negatively associated with hypertension, diabetes, and dyslipidemia. In particular, general obesity associated more negatively with hypertension, yet central obesity associated more negatively with diabetes and dyslipidemia. We found that the risks of those diseases increased respectively by 160.8%, 33.1%, and 65.8% for obese population, comparing with whom were having normal BMI. In addition to that, the risks of those diseases increased respectively by 119.8%, 55.8%, and 84.2% for obese population, comparing with whom were having normal WC. Here we should point out, the associations between obesity and chronic diseases we studied in the research based on existing obesity standards of BMI and WC for general population, in which young and middle-aged populations were in majority.

Abnormal TC and TG were the main reasons of dyslipidemia, and could be used as major indicators for discovering and preventing disease. We found that 37.9% (29.5% for male, 45.1% for female) senior population were having abnormal TC, which was the highest among lipid-related indices. Besides TC, we found that the increase of the prevalence of abnormal TG was in accordance with that of dyslipidemia, while the prevalence of TC, LDL-C, and HDL-C increased with few percentages. With the rising of the units of BMI and WC, the prevalences of abnormal TG increased by 17.0% and 8.9% respectively for male, as well as 20.0% and 9.7% respectively for female, which were close to the increase of the prevalences of dyslipidemia by 20.8% and 9.2% respectively for male, as well as 16.6% and 6.7% respectively for female. Moreover, with the rising of the category levels of BMI and WC, the prevalences of abnormal TG increased by 33.1% and 21.0% respectively for male, as well as 29.6% and 17.1% respectively for female, which were also close to the increase of the prevalences of dyslipidemia by 34.5% and 19.5% respectively for male, as well as 20.8% and 10.1% respectively for female.

Comparing with whom were having normal BMI or WC, male population with general obesity or central obesity were riskier for having hypertension, diabetes, and dyslipidemia than female based on existing obesity standard. We found that the adjusted ORs for male with general obesity were 2.6632, 1.4182, and 2.1880 higher than those for female, i.e., 2.5585, 1.2703, and 1.4090, of having hypertension, diabetes, and dyslipidemia, respectively. Plus, the adjusted ORs for male with central obesity were 2.3045, 1.5939, and 2.1772 also higher than those for female, i.e., 2.0844, 1.4867, and 1.5917, of having those diseases respectively.

In real situation however, female senior population were riskier than male for having hypertension, diabetes, and dyslipidemia in the same level of BMI or WC. We found that when BMI were 28.7听kg/m², 30.1听kg/m², and 22.7听kg/m² respectively, the ORs for male all were 1.0, whereas those ORs were higher than 1.0 for female in the risk of having hypertension, diabetes, and dyslipidemia. Similarly, when WC were 99.0听cm, 95.9听cm, and 82.1听cm respectively, the ORs were 1.0 for male, while the ORs were higher than 1.0 for female in the risk of having diseases. For this paradox, we considered that an insignificant proportion of male population could have been categorized to obesity based on existing obesity standard, which resulted in slightly higher risks for male population having diseases.

Consequently, we recommend that BMI should control lower than 28.7听kg/m² and 23.3听kg/m², as well as WC should control lower than 95.9听cm and 83.1听cm for male and female respectively to prevent hypertension and diabetes. The recommendation is close to the result of research from Shi [15], that BMI in 24.0鈥墌鈥�31.9听kg/m² with lowest death risk for senior population. For population over 65 years old, it is unpractical and difficult to control BMI lower than 22.7听kg/m² and 18.1听kg/m², along with WC lower than 82.1听cm and 65.7听cm for male and female respectively. Simply because even with normal BMI and WC, there were risks for having dyslipidemia or abnormal lipid-related indices.

There are limitations of the study. First, there were two thirds of senior population about 6听million in Zhejiang province were not included in the study, due to some of the elderly didn鈥檛 attend the yearly medical examination (physical problems, time occupied, far traveling, etc.) or invalid EHRs. From research aspect, we collected all valid EHRs as many as possible without any subjective selection. However, it might present a bias that healthier senior population tend to take the examination. Second, the research was a cross-sectional study, which explained causal relationship with relatively lower sufficiency. Third, considering poor memory of the senior population, exact number of cigarettes consumption, volume of alcohol consumption, and frequency of physical exercise are not required to report, so there were no more detailed categories in the electronic health records.

Conclusions

Senior population were more likely to have central obesity over general obesity, and nearly half of them were having hypertension and dyslipidemia. Obesity negatively and strongly associated with chronic diseases in senior population, yet general obesity exerted more impact on hypertension, whereas central obesity exerted more impact on diabetes and dyslipidemia. Female population with general or central obesity were in higher risk than male having hypertension, diabetes, and dyslipidemia. We recommended senior population control BMI lower than 28.7听kg/m² and 23.3听kg/m², as well as WC lower than 95.9听cm and 83.1听cm for male and female, respectively. Optimal BMI and WC in senior population may be around the overweight or mild obesity range, which would challenge the application of international and national standard and guidelines of obesity in senior population. There were risks for having dyslipidemia or abnormal lipid-related indices even in senior population with normal BMI and WC. TC and TG were main reasons of dyslipidemia, and could be as major indicators of discovering disease and preventing senior population from severe dyslipidemia and complications.