Article Text

Original research
Associations of hypertension, diabetes and heart disease risk with body mass index in older Chinese adults: a population-based cohort study
  1. Haoxiang Lin,
  2. Nan Xiao,
  3. Shujun Lin,
  4. Meng Liu,
  5. Gordon G Liu
  1. Peking University Institute for Global Health and Development, Beijing, China
  1. Correspondence to Dr Gordon G Liu; gordonliu{at}nsd.pku.edu.cn

Abstract

Objective Obesity is a well-established risk factor for disease. Controversy exists regarding the relative risk of morbidity and mortality in individuals who are overweight or underweight compared with individuals with a normal body mass index (BMI). In this study, we investigated the associations between BMI and three non-communicable diseases (hypertension, diabetes and heart disease) in older adults.

Design Cohort study.

Setting This study used data from the China Health and Retirement Longitudinal Study. The baseline survey was carried out in 2011, and follow-up surveys were conducted in 2013, 2015 and 2018.

Participants Participants who reported having no doctor-diagnosed chronic disease at baseline were included in this study.

Main outcome measures We analysed the association between baseline BMI and disease incidence using Cox proportional hazards models. Disease information included self-reported diagnosed conditions. BMI was categorised according to the standard Chinese criteria: underweight (<18.5 kg/m2), normal body weight (18.5–23.9 kg/m2), overweight (24.0–27.9 kg/m2) and obese (≥28.0 kg/m2).

Results A total of 5605 participants were included at baseline. Based on the Kaplan-Meier estimation, the participants who were obese had the highest incidence of all three diseases. Compared with normal weight participants, overweight participants had a greater disease incidence (log-rank tests are p<0.01). Cox regression models showed that with increasing BMI, the HRs of diseases increased accordingly (eg, for hypertension, compared with the BMI group <18.5 kg/m2, the HRs for the BMI groups 18.5–23.9, 24.0–27.9 and ≥28.0 were 1.43 (95% CI 1.00 to 2.05), 2.19 (95% CI 1.51 to 3.18) and 2.89 (95% CI 1.91 to 4.36), respectively).

Conclusion A higher BMI was associated with an increased risk of hypertension, diabetes and heart disease in the population aged 45 years and older. Even within normal BMI ranges, a higher BMI was associated with an increased risk of disease. Actions are urgently needed at the population level to address the growing public health challenge of excess weight in the context of an ageing population.

  • Obesity
  • Epidemiology
  • Risk Factors

Data availability statement

Data are available on reasonable request. The data of the studies are accessible via Peking University.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • This study used nationally representative longitudinal data to examine the association between obesity and disease incidence.

  • Disease information was based on self-reports without biochemical verification.

  • A considerable proportion of participants may be unaware of their conditions or have undiagnosed diseases.

Introduction

Evidence prior to this study

The associations between body mass index (BMI) and various diseases have important implications, especially with the increased burden of metabolic diseases worldwide.1 Intervention studies revealed that weight loss reduces some disease symptoms, suggesting that BMI is not merely a factor associated with disease but may also have a causal relationship.2 3 In recent decades, the prevalence of overweight and obesity has increased dramatically both in developed and developing countries, leading to what has been called a global epidemic.2 4 Therefore, to change the trend and mitigate the disease burden, it is necessary to examine the association between BMI and disease incidence.

Many studies have shown that a higher BMI is associated with a lower quality of life and functional impairment, including many non-communicable diseases (NCDs), such as hypertension, diabetes and heart disease.5 6 In addition, based on a large cohort study, obesity is associated with a reduction of 3–8 years in disease-free life compared with healthy weight.7

Why research is needed

Although obesity is a well-established risk factor for morbidity, including NCDs and other health conditions, there are at least two important questions with strong implications that have been understudied in previous studies. First, there is a controversy regarding the relative risk of morbidity and mortality in overweight or underweight individuals compared with individuals who have a normal BMI. For example, a systematic review published in JAMA showed that relative to normal weight, overweight was associated with significantly lower all-cause mortality.8 In addition, abundant evidence indicates that overweight individuals have an increased risk of specific diseases.6 9 The associations between underweight and various diseases also need more evidence. Researchers found that underweight seemed to be associated with a decreased risk of some cancers10 and a lower lifetime risk of cardiovascular disease morbidity and mortality for men.6 Other researchers found that underweight and obesity were associated with increased mortality relative to the normal weight category, and overweight was not associated with excess mortality.11

Second, the discussion of disease factors in the context of specific age groups is worthy of further exploration. For example, people in the 45–65 years age group are of particular interest because they are in a prestage of less healthy years of life, which is associated with the greatest onset of NCDs.6 The estimation of lifetime risk factors for diseases is crucially meaningful but may cause difficulty for preventive policy implementation because a comprehensive health promotion strategy is costly. If we can identify the factors that have the most significant influence on NCDs for a specific age group, the government can use that information to set priorities for disease prevention and provide individuals with a more achievable health lifestyle goal than simply avoiding all risk factors.

To address these unanswered questions, our aim was to investigate the associations between BMI and the three most common NCDs (hypertension, diabetes and heart disease) in older adults using cohort data from the China Health and Retirement Longitudinal Study (CHARLS). Our report advances the field by including only healthy participants at baseline; we can observe disease onset, and with the more detailed stratification of BMI categories, this study can not only examine the association between obesity and disease but also reveal the trend of disease incidence with increasing BMI.

Method

Data

In this study, we used data from the CHARLS, which is a nationally representative longitudinal survey among middle-aged and elderly people aged 45 years and above in China. The CHARLS database is an open-access database, and researchers can use CHARLS data for research purposes through a Peking University-managed website (https://charls.charlsdata.com/pages/data/111/en.html). The national baseline survey of the CHARLS was carried out in 2011, and follow-up surveys were conducted in 2013, 2015 and 2018. The CHARLS collected sociodemographic, health status, health behaviour, physical function and other information from the participants through face-to-face personal interviews with a standardised questionnaire and physical measurements. The sample includes approximately 10 000 households and more than 17 000 respondents from 27 provinces. Further details about the study design of the CHARLS have been previously reported.12 13

For our analysis, we used data from all three waves. A total of 17 708 participants completed the baseline survey in 2011. Only the participants who reported having no doctor-diagnosed chronic disease at baseline were included in this study. We excluded 12 103 individuals who had doctor-diagnosed chronic diseases or who reported unclear answers. Therefore, 5605 participants were included in the cohort at baseline. The detailed selection process is shown in figure 1.

Measurements of outcomes and BMI

There are 14 doctor-diagnosed chronic diseases in CHARLS, including hypertension, dyslipidaemia, diabetes, cancer, chronic lung diseases, liver disease, heart disease, stroke, kidney disease, digestive disease, psychiatric disease, memory-related disease, arthritis or rheumatism and asthma. We focused on hypertension, diabetes and heart disease. In the survey, participants were asked, ‘Have you been diagnosed with (a disease) by a doctor?’

Weight and height at baseline were measured as part of the physical examination by trained research staff. The equipment used to measure height was a stadiometer. The interviewer first asked the respondent to stand erect on the floorboard of the stadiometer with his or her back to the vertical backboard of the stadiometer. The weight of the participants was evenly distributed on both feet. The heels of the feet were placed together with both heels touching the base of the vertical board. The respondents were asked to place their feet pointed slightly outwards at a 60° angle. The respondent’s head was maintained in the Frankfort Horizontal Plane Position when measuring. The equipment measured weight on a scale. The respondents were asked to stand on the scale with their shoes off. When taking the body weight measurement, the respondent must not be assisted or leaning against any object such as a wall or chair. The respondent was asked to look straight ahead, to stand very still, and to remain on the scale until the body weight on the scale did not change.12

We calculated BMI using the following formula: weight (kg) divided by height (m) squared. BMI was categorised according to the standard Chinese criteria: underweight (<18.5 kg/m2), normal body weight (18.5–23.9 kg/m2), overweight (24.0–27.9 kg/m2) and obese (≥28.0 kg/m2).14

Other covariates

The covariates included demographic characteristics, smoking behaviour, alcohol consumption behaviour, sleep duration, physical activity and depression symptoms. The demographic characteristics included age, sex and hukou status (non-agricultural or agricultural). Participants were also classified by their smoking status (non-smoker, ex-smoker and current smoker) or drinking status (non-drinker, less than once per month and more than once per month).

Statistical analysis

Our data analysis was conducted in three steps. First, we employed descriptive statistics to present the overall characteristics of our three cohorts. Second, we estimated the cumulative incidence of diseases by Kaplan-Meier curves. Third, we analysed the association between BMI and disease incidence using Cox proportional hazards models. A p<0.05 was considered indicative of statistical significance. All the statistical analyses were conducted by using SPSS V.19.0. No imputation was performed for missing data.

Step 1. Comparisons between different cohorts were performed by descriptive statistics. Categorical variables are presented as counts and percentages. Numerical variables are presented as the mean and SD.

Step 2. The Kaplan-Meier method was used to show the univariate association between BMI and disease incidence. We changed the traditional survival curves to disease incidence. The differences between curves were assessed using the log-rank test. We calculated the cumulative incidence of disease by 7 years of follow-up.

Step 3: The first Cox proportional hazards model was used to investigate the associations between traditional BMI categories and disease incidence from the perspective of multivariate analysis. We calculated the cumulative risk for specific disease events in each cohort. We determined the proportion of incident first-time reported specific diseases that occurred during the follow-up period. For example, if a respondent reported hypertension in 2013, then we coded the hypertension event variable as 1 since 2013. If a respondent reported no hypertension in any of the follow-up years, we coded 0 all the time. Adjusted HRs for the risk of disease incidence and their 95% CIs were calculated by multivariate analyses. We included BMI, sex, age, hukou status, smoking behaviour, alcohol consumption behaviour and sleep duration.

Step 4: To show the relationship between BMI and disease incidence more clearly, we classified more categories of BMI in the second Cox proportional hazards model. We included only variables that were statistically significant in the multivariate analysis. The mean of each BMI category and HR was plotted along with straight lines of best fit. The slopes of these lines were described in terms of the change in the mean and HR.

Patients and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination of our research.

Results

Table 1 shows the baseline characteristics of the patients included in this study. The mean age was approximately 57.0 years. Most of the participants were agricultural residents. The majority of the participants had a normal or overweight BMI. At the 7-year follow-up, the incidences of hypertension, diabetes and heart disease were 17.9%, 5.7% and 6.9%, respectively.

Table 1

Baseline characteristics

Figure 2 shows the univariate-level association between BMI and disease incidence. Participants who were obese (BMI≥28) had the highest disease incidence among all cohorts. Compared with those with a normal weight, overweight participants (BMI 24.0–27.9) had a greater disease incidence, especially after the fourth year of follow-up (all log-rank tests are p<0.01).

Figure 2

Relationship between disease incidence and BMI. BMI, body mass index.

Table 2 reports the results of the multivariate analysis. BMI had a significant influence on disease incidence, and the trend was similar to that in the univariate analysis. With increasing BMI, the HR of diseases increased accordingly (eg, for hypertension, compared with the BMI group <18.5, the HRs for the BMI groups 18.5–23.9, 24.0–27.9 and ≥28.0 kg/m2 were 1.43 (95% CI 1.00 to 2.05), 2.19 (95% CI 1.51 to 3.18) and 2.89 (95% CI 1.91 to 4.36), respectively).

Table 2

Relationship between disease incidence and BMI in a multivariate analysis with 7 years of follow-up

To show the relationship between BMI and disease incidence more clearly, we classified more BMI categories and constructed straight lines of best fit. The results are presented in figure 3 and online supplemental table 1. Among all cohorts, with increasing BMI, the risk of disease incidence increased dramatically. Even though for heart disease, the increased risk of incidence associated with increases in BMI was not statistically significant for the lower BMI group, the overall trend was unchanged for all three diseases.

Supplemental material

Figure 3

Associations between the HRs of disease incidence and BMI (with the best fit line). BMI, body mass index. *means P<0.05.

Discussion

This study provides new information for the ongoing debate regarding the association between BMI and the risk of disease and provides a comprehensive overview using national data for the population aged 45 years and above in China. The association between BMI and disease incidence was robust; although a difference existed in the lower BMI group, the trend of increased disease risk with increasing BMI was similar across the three disease subgroups and BMI subgroups.

Notably, our study revealed that a higher BMI was roughly linearly related to an increased risk of hypertension, diabetes and heart disease. Existing studies have almost reached a consistent conclusion that when comparing the obese population with the normal-weight population, obesity was associated with a significantly increased risk of developing NCDs. However, in recent years, there has been controversy about the health implications of overweight status, and some studies have shown similar or lower all-cause mortality rates and less excess mortality in overweight individuals than in normal weight individuals for cardiovascular disease.8 11 A possible explanation is that those studies usually did not account for the age at disease onset or the duration of life lived with a specific disease. Therefore, we strongly suggest including both disease-related mortality and morbidity in different BMI groups when reporting longitudinal observational research.

Underweighting is another topic that needs more study. Our results revealed that the risk of diabetes and heart disease in the underweight group was not different from that in the normal-BMI group, and the coefficients and HRs were lower in the underweight group. More studies are needed to explore the possible reasons and other possible confounders involved. If the results are robust and supported by population-based research from the perspective of lifetime risk, they will have important implications for public health; for example, in this case, what are the appropriate criteria for a normal BMI considering the current social and economic environment? Should different age groups have different criteria for normal BMI?

This study has several limitations. First, disease information was based on self-reports without biochemical verification, and the respondents’ understanding of some of the questions might vary. Second, we did not measure some important confounders that may have contributed to disease incidence, such as patient disease information and dietary information. Third, we only used baseline BMI and other characteristics without accounting for changes in those factors across follow-up. Fourth, we included only individuals who had BMI data, and the decision to measure weight might be related to the individual’s apparent weight or their health status; therefore, selection bias may exist. Fifth, a considerable proportion of participants may be unaware of their conditions or have undiagnosed diseases so a focus on diagnosed diseases may provide only a partial picture of the association between BMI and disease incidence.

Despite these limitations, the study provides potentially important information. BMI plays a crucial role in the incidence of hypertension, diabetes and heart disease in older adults, and similar to obesity, overweight status is associated with a greater risk of disease incidence, which is a finding of considerable public health importance. Therefore, at the population level, reducing the percentage of obese and overweight people could have a tremendous societal positive impact on public health. Health practitioners should conduct evidence-based interventions, such as behaviour change theory-based health education and dietary and physical activity interventions, to help people maintain a healthy BMI.15–17

Conclusion

Using the CHARLS data, this study showed that a higher BMI was associated with an increased risk of hypertension, diabetes and heart disease for the older adult population, emphasising the public health implications in view of the overall increase in population BMI distributions in many countries. Further research could be beneficial from two perspectives: (1) determining the most appropriate BMI for different age groups and (2) using more robust methods to prove the causal relationship between BMI and disease. Assuming that a causal relationship exists, actions are urgently needed at the population level to address the growing public health challenge of excess weight in the context of an ageing population.

Supplemental material

Data availability statement

Data are available on reasonable request. The data of the studies are accessible via Peking University.

Ethics statements

Patient consent for publication

Ethics approval

This study based on the CHARLS dataset was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015), and all participants were fully informed of the purpose of the research and signed informed consent forms before joining the study.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors HL finished the first draft. HL and GGL managed the study. NX conducted the statistical analysis. SL and ML helped draft and revise the paper. HL, NX and SL contributed equally to this study. All authors have approved the final paper for submission. HL accepts full responsibility for the work and the conduct of the study, had access to the data, and controlled the decision to publish.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.