Study design and participants
A secondary data analysis of a prospective longitudinal study was conducted using the Elderly Cohort Data of the National Health Insurance Service in South Korea between 2002 and 2013. This representative survey in Korea was originally designed to analyze the risk factors and prognosis of senile diseases and included socioeconomic information, hospital-use history, and diverse health examination results.
The participants in this study were 10,588 older adults (over 60 years of age) with hypertension. This Elderly Cohort Data surveyed the number of days of physical activity per week in 2002-2008, and the number of days participants engaged in intense, moderate, and walking activity per week in 2009-2015. Therefore, we identified physical activity as “yes” or “no”, regardless of its time and intensity. Participants were categorized into 4 groups by their changes in physical activity presence or not over time. Group I included the people who reported that they did not engage in physical activity in their daily life throughout the assessment period of 2002-2013 (the totally inactive group). Group II included those who did not engage in physical activity in 2002 but began to participate in physical activity by 2013 (the inactive to active group). Group III included those who reported that they were physically active in 2002 but did not sustain their participation in physical activity when assessed in 2013 (the active to inactive group). Finally, Group IV included those who reported remaining physically active throughout 2002-2013 (the totally active group).
As for CVD indicators, we extracted variables of BP, blood glucose, body mass index (BMI), and total cholesterol levels for 12 years from the database.
BP was measured as systolic/diastolic pressure (mmHg) on the participants. According to the European Society of Hypertension, BP over 160/100mmHg is considered Stage 2 (more severe) hypertension ; therefore, BP measurements that were 160/100mmHg or higher were used as the criteria of uncontrolled BP in this study.
In the cohort survey, pre-meal plasma glucose was checked in 2002-2010, and fasting plasma glucose was measured in 2011-2013. Blood glucose was considered to be controlled if it was less than 100mg/dL , so blood glucose greater than 100mg/dL was categorized as uncontrolled glucose in this study.
Body mass index
Height (cm) and weigh (kg) were measured on the participants. BMI was calculated from the data and was categorized as being normal or obese using the cut-off value of 25kg/m2 .
Total cholesterol level was measured for the participants. Participants were categorized as normal or having hypercholesterolemia using the cut-off value of 200mg/dL .
Age, gender, and economic status were identified as demographic characteristics. Economic status was divided into 4 groups: Low, Lower-Middle, Upper-Middle, and High. Smoking and alcohol consumption were included as characteristics of participants’ health behaviors. The number of comorbidities and family history of related diseases (i.e., hypertension, diabetes, heart disease, and cerebral disease) were included as disease-related characteristics.
Data cleaning and analysis was performed using SPSS version 23.0 (IBM SPSS Statistics, Armonk, NY). Demographic characteristics, health behaviors, and disease-related characteristics in 2002 were reported using descriptive statistics, such as frequency (percentage) and mean (standard deviation). Chi-square tests and one-way ANOVA were used to analyze differences among the physical activity groups. In addition, the Kaplan-Meier method was used to analyze the time to change from controlled to uncontrolled for BP and glucose level and from normal to high for BMI and total cholesterol level. The significance of the analyzed curves was tested using the log-rank test. A Cox proportional hazard model was also used to confirm the risk of cardiovascular indicators over time in each group.
This study was approved by the institutional review board of XX University (IRB No. AJIRB-SBR-EXP-19-127). After receiving approval, a request was made to access the Elderly Cohort Data to the National Health Insurance Data Sharing Services. All data were de-identified before conducting the analysis, so personal identification was protected.