This study used baseline records of 3,010 ambulatory participants who were aged 70–84 years, had geriatric assessments and TUG data that were assessed from 2016 to 2017 in the Korean Frailty Aging Cohort Study (KFACS), a nationwide multicenter longitudinal study conducted in 10 urban, rural, and suburban communities across Korea. Detailed descriptions of the KFACS design and measures are published elsewhere.¹⁴⁾ Briefly, the participants were recruited based on age- and sex-specific strata, and residents with no plans to move out during the following 2 years and with no difficulties in conversing were eligible to participate in this study. People with uncontrolled hypertension (>180/100 mmHg), cerebrovascular accident or myocardial infarction within the past 6 months, or active malignancy currently under treatment were excluded. The study protocol was reviewed and approved by the Institutional Review Board of Kyung Hee University Hospital (No. 2020-09-049) and complied with the ethical rules for human experimentation described in the Declaration of Helsinki. Informed consent was obtained from all participants or their proxy.

For TUG, time was measured as the time required for the participants to rise from a straight-backed chair at once, walk at a comfortable pace for 3 m, turn around and walk back to the chair, and sit down.⁶⁾ The participants were instructed to start the TUG maneuver immediately after hearing the “Start” command. The participants were allowed to use walking aids (e.g., cane or walker) during TUG. The usual gait speed over a distance of 4 m was measured using an automatic gait speed meter (Gaitspeedometer, Dyphi, Daejeon, Korea), with acceleration and deceleration phases of 1.5 m each.¹⁵⁾ The participants were asked to perform TUG by walking at their usual pace. The participants performed the test two times, and the results were averaged. The five-time-sit-to-stand test measured the time required to stand five times from a sitting position from a straight-backed chair as quickly as possible without using the arms. We also used SPPB, which included three components—standing balance, walking speed, and chair rise test—to assess physical performance according to recommendations from previous studies.¹⁵⁾ Each item of SPPB was scored based on a 0 to 4 point scale, with the total score ranging from 0 (worst) to 12 (best) points. According to literature, low physical performance was defined as an SPPB total score ≤9.¹⁶⁾ SPPB parameters were available for 1,429 men and 1,581 women.

To assess physical frailty, we used the CHS scale, which comprises five components, namely unintended weight loss, poor grip strength, exhaustion, reduced walking speed, and low physical activity level.⁴⁾ We defined unintended weight loss as ≥4.5 kg in the previous year. Poor grip strength was defined as the lower 20th percentile of grip strength (maximal grip strength in kg after measuring twice for each hand using a hand grip dynamometer [T.K.K.5401; Takei Scientific Instruments Co., Tokyo, Japan]), stratified by sex and body mass index (BMI) quartiles based on the KFACS baseline survey.¹⁷⁾ To quantify exhaustion, we used the following statements from the Center for Epidemiological Studies-Depression (CES-D) scale for 3 or more days in a week: “I felt that everything I did was an effort” or “I could not get going”.¹⁸⁾ For slowness, we used the lowest 20% of a 4-m gait speed stratified by sex and height based on KFACS data.¹⁷⁾ Low physical activity level was defined as an energy expenditure level 494.6 kcal per week for men and below 283.5 kcal per week for women according to the International Physical Activity Questionnaire. These values corresponded to the lowest 20% of the sex-specific total energy consumed according to a general population-based survey of older adults.¹⁷⁾ Participants positive for three or more of these items were classified as frail; those positive for 1–2 items were classified as prefrail, and those with no positive items for any of the criteria were classified as robust. The CHS frailty parameters were available for 1,382 men and 1,523 women.

Information on participant age, marital status, education level, drinking status, smoking status, comorbidities, and functional capacities was collected during face-to-face interviews. Alcohol consumption was defined as three or more alcoholic drinks per week, while smoking was defined as a lifetime consumption of 100 or more cigarettes. To assess functional capacities, Korean Mini-Mental State Examination,¹⁹⁾ Geriatric Depression Scale Short Form,²⁰⁾ and minimal nutritional assessment questionnaires²¹⁾ were used. BMI was calculated as weight divided by the square of the height (kg/m²).

Continuous variables are shown as mean±standard deviation, while categorical variables are described as number and percentage. T-tests were used for continuous variables and χ² tests for categorical variables. Skewness was calculated, and histograms were used to display the distribution of TUG time in men and women. The commonly used percentile values for epidemiological studies were also calculated. We used linear regression analysis to assess the correlations between TUG time, total number of positive items according to the CHS criteria, and SPPB total score. Linear regression analyses were also used to assess the correlations between TUG time, item specific and total CHS frailty scale and SPPB scores. Receiver operating characteristic (ROC) analyses were performed to evaluate the classification performance of TUG for frailty assessment using the CHS criteria and low physical performance; the sensitivities and specificities for various TUG times (≥8 to ≥16 seconds) to classify frailty and low physical performance were also calculated. Statistical analyses were performed using Stata 15.0 (Stata Corp, College Station, TX, USA), with statistical significance set at a two-sided p-value <0.05.

The mean age of the 3,010 participants was 76.5±3.9 years, and 1,429 (47.5%) participants were men. The baseline demographic, anthropometric, and functional characteristics are shown in Table 1. The mean TUG times were 10.3±2.7 seconds in men and 10.8±3.0 seconds in women. The TUG time was significantly higher in women than in men (p<0.001). The histogram in Fig. 1 displays the distribution of TUG times in men and women, with commonly used percentile cut-offs of TUG time in the study population. Specifically, the TUG cut-off times for the worst quintile was 11.8 seconds in men and 12.5 seconds in women. Assessment showed that TUG was right skewed in both men and women (p<0.001).

TUG time was positively associated with age (standardized beta [B]=0.38, p<0.001) and CHS frailty scale score (B=0.24, p<0.001) and negatively associated with SPPB total score (B=-0.36, p<0.001) in the linear regression analysis (Supplementary Fig. S1). The correlation between TUG time and CHS frailty scale score remained significant (B=0.22, p<0.001) after adjusting for age and sex. Similarly, the correlation between TUG time and SPPB total score remained significant (B=-0.33, p<0.001) in the multivariate linear regression analysis that included age and sex. Fig. 2 shows the corresponding TUG time distribution according to the CHS frailty scale score and SPPB total score. To assess content validity, correlations between TUG time and component-specific and total scores of the CHS frailty scale and SPPB were calculated using linear regression analysis, as shown in Table 2.

In the ROC analysis, the performance of TUG in identifying physical frailty, calculated as the area under the curve (AUC), was 0.87, while the AUC of TUG in identifying low physical performance according to SPPB was 0.86 (Fig. 3).

The calculated sensitivities and specificities of TUG time to classify physical frailty and low physical performance in men and women are displayed in Table 3. Based on the worst quintile TUG time cut-off of ≥11.8 seconds for men and ≥12.5 seconds for women, the sensitivity and specificity were 74.74% and 83.92%, respectively, in men and 65.04% and 85.07%, respectively, in women for frailty and 68.57% and 86.04%, respectively, in men and 55.08% and 90.56%, respectively, in women for low physical performance.