All research data were retrieved from the National Health Insurance Service (NHIS) database. The NHIS is the only public health insurance in South Korea and contains and manages data on disease diagnoses and prescriptions for procedures and drugs. Physicians (from all outpatient clinics and hospitals) must register all prescription information on procedures, medications, and disease diagnoses in the NHIS database to obtain information from the government regarding treatment costs. Disease diagnoses were registered using the International Classification of Diseases, External 10th Revision (ICD-10) codes. In addition, the NHIS database contains demographic and socioeconomic status-related information for all patients in South Korea.

All adult patients (aged ≥ 20 years) who were admitted to the ICU during hospitalization between January 1, 2010, and December 31, 2018, were initially screened. Among the screened patients, we included those who had undergone standardized health examinations in the year prior to admission to the ICU. In South Korea, NHIS subscribers aged ≥ 40 years are recommended to undergo standardized health examinations every 2 years.10 Only the first ICU admission case on the earliest date was included in this study if a patient was admitted to the ICU two or more times during the study period. This exclusion criterion increased the homogeneity of our study population because the physical condition of the patients in the ICU might have been worse on the subsequent ICU admissions than on the first ICU admission. In addition, patients with missing data regarding lifestyle factors in the standardized health examinations were excluded from the final analysis.

Three lifestyle factors were evaluated prior to ICU admission: smoking status, alcohol consumption, and physical activity. Information regarding lifestyle factors was obtained through voluntary responses to questionnaires obtained during standardized health examinations. Smoking status was classified into two groups: current and non-current smokers. Based on previous reports,1112 alcohol consumption was divided into three groups: non-drinkers, mild drinkers, and heavy drinkers. The mild drinkers group was defined as alcohol consumption of ≤ 210 g and ≤ 140 g per week and the heavy drinkers group was defined as alcohol consumption of > 210 g and > 140 g per week in men and women, respectively. Physical activity was divided into three types according to intensity: intensive, moderate, and mild physical activity. Intensive physical activity was defined as strenuous exercise for more than 20 minutes, and moderate physical activity was defined as moderate exercise without breathlessness for more than 30 minutes. Mild physical activity was defined as walking for > 30 minutes. Each exercise intensity was divided into four groups: no exercise, 1–3 times per week, 4–5 times per week, and 6–7 times per week. These physical activities include all type of physical activities from work or leisure in all patients, and the questionnaire for physical activity is presented in Supplementary Data 1. As we used data from standardized health examinations in the year prior to admission to the ICU, physical activity was evaluated during the year prior to admission to the ICU.

Our study had two primary endpoints: 30-day mortality and 1-year all-cause mortality after ICU admission. Mortality was defined as death within 30 days or 1 year after ICU admission.

Information from the standardized health examination was collected and used as covariates. Body mass index data were collected and categorized into six groups: < 18.5, 18.5–24.9, 25.0–29.9, 30.0–34.9, 35.0–39.9, and > 40 kg/m². Waist circumference was measured and classified into two groups: ≥ 90 cm and < 90 cm. All laboratory results were obtained from venous blood samples during standardized medical examinations. Included individuals were diagnosed with anemia, according to the World Health Organization criteria, as hemoglobin levels of < 12 g/dL and < 13 g/dL for women and men, respectively. Cut-off values of 125, 200, 40, and 130 mg/dL were used for fasting serum blood glucose, serum total cholesterol, high-density lipoprotein, and low-density lipoprotein levels, respectively. Cut-off values of 40 IU/L were used for aspartate transaminase, alanine transaminase, and gamma-glutamyl transferase.

All covariates in this study were collected based on previous research on the epidemiology of ICU admissions in South Korea.13

The clinicopathological characteristics were presented as mean values with standard deviation (SD) for continuous variables and numbers with percentages for categorical variables. We constructed a multivariable logistic regression model to assess the prior lifestyle factors associated with 30-day mortality. All covariates were included in the multivariable adjustment in the model. The Hosmer–Lemeshow test was used to confirm that the goodness of fit in the model was appropriate. The results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). We also constructed a multivariable Cox regression model to determine prior lifestyle factors associated with 1-year all-cause mortality. All covariates were included for multivariable adjustment in the model, and log-log plots were used to confirm that the central assumption of the Cox proportional hazard models was satisfied. The results are presented as hazard ratios (HRs) with 95% CIs. There was no multicollinearity issue with the criterion of variance inflation factors < 2.0 between variables. All statistical analyses were performed using IBM SPSS Statistics for Windows (version 25.0; IBM Corp., Armonk, NY, USA), and statistical significance was set at P < 0.05.

This retrospective, population-based cohort study was conducted after obtaining approval from the Institutional Review Board (IRB) of Seoul National University Bundang Hospital (IRB approval number: X-2102-666-904). The Big Data Center of the NHIS also approved data sharing for this study (NHIS-2021-1-620). The requirement for informed consent was waived because the data analyses were performed retrospectively using anonymized data derived from the South Korean NHIS database.

Fig. 1 shows a flowchart of the patient selection process. From January 1, 2010 to December 31, 2018, there were 3,132,769 ICU admissions, of which 2,212,923 were adult patients. Among them, 695,897 ICU admissions of 665,732 patients who had undergone standardized health examinations in the year prior to admission to the ICU were initially screened. A total of 30,165 patients with ≥ 2 ICU admissions during the study period were excluded to focus only on the first ICU admission on the earliest date. Next, 80,349 patients with missing standardized health examination data in the year prior to admission to the ICU were excluded. Ultimately, 585,383 patients were included in the analysis. Among them, 59,075 (10.1%) and 113,476 (19.4%) patients died within 30 days and 1 year after ICU admission, respectively. Table 1 shows the clinicopathological characteristics of the patients included in this study. The patients were 62.9% male (368,343/585,383) with a mean age of 64.9 years (SD: 13.1 years). The mean length of hospital and ICU stays was 13.8 days (SD: 12.0 days) and 3.0 days (SD: 4.2 days), respectively. The surgery-associated hospital admissions were 72.9% (426,834/585,383). Table 2 shows the data regarding standardized health examinations in the year prior to ICU admission.

Table 3 shows the results of the multivariable logistic regression model for 30-day mortality after ICU admission. Current smoking (vs. non-current smokers: OR, 0.99; 95% CI, 0.96–1.02; P = 0.432), mild alcohol drinking (vs. no alcohol drinking: OR, 1.00; 95% CI, 0.96–1.03; P = 0.825), and heavy alcohol drinking (vs. no alcohol drinking: OR, 0.97; 95% CI, 0.92–1.01; P = 0.166) were not associated with 30-day mortality after ICU admission. One to 3 days per week of intensive physical activity was associated with lower odds of 30-day mortality (OR, 0.96; 95% CI, 0.93–0.99; P = 0.016) than without intensive physical activity. Four to 5 days per week (OR, 0.92; 95% CI, 0.88–0.98; P = 0.005) and 6–7 days per week (OR, 0.89; 95% CI, 0.84–0.94; P < 0.001) of moderate physical activity were associated with lower odds of 30-day mortality than without moderate physical activity. One to three days per week (OR, 0.95; 95% CI, 0.92–0.97; P < 0.001), 4–5 days per week (OR, 0.93; 95% CI, 0.89–0.96; P < 0.001), and 6–7 days per week (OR, 0.93; 95% CI, 0.90–0.96; P < 0.001) of mild physical activity were associated with lower odds of 30-day mortality than no physical activity. Other ORs with 95% CI for all other covariates are presented in Supplementary Table 1.

Table 4 shows the results of the multivariable Cox regression model for 1-year all-cause mortality after ICU admission. Current smoking was associated with a higher risk of 1-year all-cause mortality (HR, 1.02; 95% CI, 1.01–1.04; P = 0.020) than not currently smoking. Mild alcohol drinking (vs. no alcohol drinking: HR, 1.02; 95% CI, 0.99–1.03; P = 0.545) and heavy alcohol drinking (vs. no alcohol drinking: HR, 0.98; 95% CI, 0.93–1.02; P = 0.156) were not associated with 1-year all-cause mortality after ICU admission. One to 3 days per week (HR, 0.95; 95% CI, 0.93–0.97; P < 0.001) and 4–5 days per week (HR, 0.95; 95% CI, 0.91–0.98; P = 0.003) of intensive physical activity were associated with lower risk of 1-year all-cause mortality than no intensive physical activity. One to 3 days per week (HR, 0.97; 95% CI, 0.95–0.99; P = 0.004), 4–5 days per week (HR, 0.94; 95% CI, 0.91–0.97; P < 0.001), and 6–7 days per week (HR, 0.92; 95% CI, 0.91–0.97; P < 0.001) of moderate physical activity were associated with lower risk of 1-year all-cause mortality than no moderate physical activity. One to 3 days per week (HR, 0.94; 95% CI, 0.93–0.96; P < 0.001), 4–5 days per week (HR, 0.94; 95% CI, 0.91–0.97; P < 0.001), and 6–7 days per week (HR, 0.92; 95% CI, 0.89–0.95; P < 0.001) of mild physical activity were associated with lower odds of 1-year all-cause mortality than no mild physical activity. Other HRs with 95% CI for all other covariates are presented in Supplementary Table 2.