The study population was drawn from the Korea National Health and Nutrition Examination Survey (KNHANES) VI, VII, and VIII. The KNHANES is conducted by the Korea Disease Control and Prevention Agency and contains structured health interviews and examinations conducted at mobile centres. To collect data representing all non-institutionalised Korean citizens, the KNHANES sampling plan follows a multistage cluster probability design. Once the survey area and households were randomly selected, the eligible individuals within those households were invited to participate by household visits, mails, and calls. A specialized survey team visited the area using a mobile examination vehicle to conduct medical examinations and health surveys. These mobile centres are equipped with medical and examination facilities, allowing participants to undergo anthropometric measurements, blood and urine samplings, eye and dental examinations, and other health-related assessments Anonymised raw data and instructions for analysis are accessible to the public on the KNHANES website (https://knhanes.kdca.go.kr/knhanes). All the KNHANES participants provided informed consent.

The inclusion criteria for the current study consisted of older adults aged 65 years or older who completed the questionnaires on physical activity and handgrip measurements. From 2014 to 2019, a total of 47,309 Koreans participated in the KNHANES. After excluding 37,484 participants aged below 65 years, 9,825 older adults aged 65 years or older (men, n = 4240, 43.16%; women, n = 5585, 56.84%) were included. Although the exact reasons for missing values are not reported on an individualized basis, participants could refuse to answer the physical activity questionnaires or the handgrip strength measurement. There were also cases of incomplete responses to the physical activity questionnaires. The handgrip strength of individuals with amputations of, paralysis of, or braces on the upper extremities was not measured. Those who had undergone wrist or hand surgeries within 3 months and with wrist or hand pain within 7 days prior to the KNHANES survey were also excluded from the handgrip strength measurement. After further excluding participants with missing data for physical activity (n = 1,213) or handgrip strength (n = 1,207), a total of 7,558 (men, n = 3,419, 45.24%; women, n = 4,139, 54.76%) older adults were included in the analysis.

Individuals without a history of the four major noncommunicable diseases (CVD, DM, CLD, and cancers) documented by the World Health Organization (WHO) were defined as participants without a major noncommunicable diseases [15]. Since the current study focused on CVD, DM, and CLD [11, 15], participants with these conditions were categorized into subgroups. Individuals with more than one noncommunicable disease (CVD, DM, or CLD) were included in more than one subgroup.

The Institutional Review Board of Uijeongbu St. Mary’s Hospital, Republic of Korea waived the need for ethical approval for this study because publicly available data were analysed.

In the KNHANES, the handgrip strength of both hands was measured in participants using a digital dynamometer (TKK 5401; Takei Scientific Instruments Co., Ltd., Niigata, Japan). The handgrip strength of each hand, starting with the dominant hand, was measured three times, with a 60-s rest between measurements. Measurements were performed in the standing position with the arms naturally lowered and the elbows and wrists extended.

The highest of the six handgrip measurement values was considered as the final handgrip strength of the participant [16]. In accordance with the Asian Working Group for Sarcopenia 2019 Consensus, dynapenia was defined using diagnostic cutoffs of handgrip strength, which is < 28.0 kg for men and < 18.0 kg for women [1].

Participation in aerobic exercise and SB were assessed using a validated Korean version of the WHO Global PA Questionnaire (GPAQ) [17, 18]. The GPAQ assesses the frequency (days/week) and duration (min/day) of participating in moderate and vigorous aerobic physical activities within a typical week that increases ones breathing or heart rate. Weekly time spent in moderate and vigorous aerobic exercises (min/week) was calculated from these frequency and duration variables, according to the GPAQ analysis guide [18]. The question 16 of the GPAG evaluates the time usually spent sitting or reclining on a typical day (min/day). Because the data on resistance exercise is not obtainable by the GPAQ, the frequency (days/week) of participation in resistance exercises such as push-ups, sit-ups, dumbbells, weights, or barbells in the last week was separately surveyed.

The obtained aerobic exercise and resistance exercise levels were categorised into binary variables (sufficient and insufficient) based on the WHO global recommendations on physical activity for health [19]. Sufficient aerobic exercise was defined as ≥ 75 min/week of vigorous-intensity aerobic physical activity, ≥ 150 min/week of moderate-intensity aerobic physical activity, or an equivalent combination of moderate- and vigorous-intensity activity according to the WHO global recommendations on physical activity for health [19]. Participants who did not meet these criteria for sufficient aerobic exercise were categorised as insufficient aerobic exercise. According to the WHO global recommendations on physical activity for health, resistance exercise ≥ 2 days/week was categorised as sufficient resistance exercise [19]. Individuals who performed resistance exercise < 2 days/week were classified as insufficient resistance exercise. Participants were categorized as ‘high sedentary time’ if they had ≥ 420 min of daily sedentary time and as ‘low sedentary time’ if they had < 420 min/day [20].

Participants diagnosed with stroke, angina, or myocardial infarction by a doctor were categorised as having a CVD [21]. Older adults with DM were defined as those diagnosed with DM by a doctor, taking hypoglycaemic agents or insulin injections, with an HbA1c ≥ 6.5%, or with a fasting blood glucose ≥ 126 mg/dL. Individuals diagnosed with chronic obstructive pulmonary disease or asthma by a doctor, with a forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) of < 0.7, or an FVC < 80% of predicted value on spirometry were categorised as having CLD. History of depression and cancer diagnosed by a doctor was also recorded.

Age, sex, and body mass index (BMI) were considered essential covariates due to their well-established associations with physical activity and handgrip strength [22,23,24]. After participants were asked to remove their shoes and stand upright against a stadiometer, the survey staff measured their height to the nearest mm. The staff recorded weight to the nearest 0.1 kg after the participants removed their heavy clothing and stood on a calibrated scale. Then BMI was obtained using the following formula: weight (kg) / height (m2). Age was calculated using the birth date of the participants.

Other covariates (alcohol, smoking, occupation, residence, and education) were selected based on their proposed associations with physical activity and handgrip strength in previous studies [25,26,27,28] and their clinical significance. Alcohol consumption data was collected through self-reported information regarding the frequency and quantity of alcohol consumption. Using this information, the average daily consumption of alcohol for each individual was calculated, as previously reported [29]. Participants were categorised as excessive alcohol drinkers if they consumed over 10 g (women) or 20 g (men) of alcohol per day. Information on smoking behaviour (never, past, or current), occupational status (yes or no), area of residence (urban or rural), and level of education (middle school or lower or high school or higher) was self-reported by participants during the interview.

The characteristics of study participants were evaluated using complex sample descriptive statistics. The weighted prevalence of dynapenia according to PA behavioural status was compared using a complex sample chi-square test. The association of participation in aerobic and resistance exercises and SB with dynapenia was analysed using complex sample multivariable-adjusted logistic regression models. To assess the potential confounding associations, four models were sequentially constructed. Furthermore, to evaluate the disease-specific association between PA and dynapenia, sequential logistic regression models were built separately for participants with CVD, DM, and CLD and participants without a major noncommunicable disease. To reduce the bias from missing data, unequal sampling, and non-response, we applied sampling weights considering the multistage cluster probability sampling design. We used complex-sample procedures of SPSS (version 24; IBM/SPSS Inc., Armonk, NY, USA).