Data source and study subjects

Detailed information regarding data source are described in the Supplementary Methods. The present study was conducted in accordance with the Declaration of Helsinki and approved by the institutional review board of Yonsei University Health System Trial Center (approval no. 4-2018-0697). The informed consent requirement was waived owing to the retrospective nature of the analysis. Data usage was also approved by the national health information data request review committee of NHIS. According to the Act on the Protection of Personal Information Maintained by Public Institutions, the NHIS provides data on health examination results and detailed medical treatments after de-identification of individual-level data. Data were retrieved from the National Health Insurance Service-National Sample Cohort Database (NHIS-NSC DB), which is a retrospective population-based sample cohort constructed on a 2.2% representative sample of the Korean population. The detailed cohort profiles with respect to the development of the NHIS-NSC DB have been previously published [43]. Data acquired during the life transition health screening examination at age 66 years were considered as baseline information. The health screening examination results obtained thereafter were used as follow-up data (Figure 3).

Design of the study. Note: Data from the National Health Insurance Service-National Sample Cohort Database were used. Abbreviations: ICD-10, 10^th revision of the International Statistical Classification of Diseases and Related Health Problems.

Elderly adults who underwent the life transition health screening examination between 2007 and 2008 were initially screened for enrollment (N=42,132). Subjects who met the following criteria were excluded: (1) eGFR <60 mL/min/1.73 m², presence of proteinuria in dipstick urine examination at baseline visit, or history of any renal replacement therapy including renal transplantation and dialysis (2) missing data on TUG test, OLS test, follow-up health screening examinations, or lifestyle questionnaires that include information on smoking and alcohol use. A total of 30,871 subjects were included in the final analysis (Supplementary Figure 7).