The study was accepted by the University of Health Sciences, Gülhane Scientific Research Ethics Committee (2023 − 151). All participants were informed about the study, and they signed a voluntary participation form that adhered to the protocols of the Helsinki Declaration before enrollment in the study. The inclusion criteria were; 1) aged ≥ 65 years and b) sufficient comprehension and speaking skills in Turkish. They were excluded if they were not community-dwellers and had neurological diseases. In this cross-sectional study, 593 community-dwelling older adults were reached using the method of snowball sampling from the province of Ankara, Turkey’s capital. Initially, 25 primary participants representing the target population of our study were identified from the researchers’ network. The primary participants were then asked to invite other individuals who might be eligible for the study. To obtain data, face-to-face interviews were conducted between May 2023 and Agust 2023 at the faculty of health sciences and a questionnaire including demographic characteristics, malnutrition risk, sarcopenic risk, frailty status, and physical activity (PA) level was administered by the researchers.

The Mini Nutritional Assessment-Short Form (MNA-SF), which is recommended by the European Society for Clinical Nutrition and Metabolism to be used in older adults [19], was used for malnutrition screening. It is a validated 6-item instrument that provides a continuous score (0–14 points) and categorizes participants as having healthy nutritional status (12–14 points), at risk of malnutrition (8–11 points), or malnourished (0–7 points) [20]. The validity study of the MNA-SF in Turkish older adults was conducted by Sarıkaya et al. [21].

The SARC-F questionnaire, which addressed strength, assistance required for walking, standing up from a chair, climbing stairs, and falls, was used for sarcopenia screening. It is a validated 5-item tool that provides a continuous score (0–10 points), with a score of ≥ 4 indicating a risk of sarcopenia [22].

Frailty screening was executed using the 5-item FRAIL Scale, which measures fatigue, endurance, ambulation, illnesses, and loss of weight. Participants who scored 0 were regarded as robust, whereas those who scored 1–2 and ≥ 3 were regarded as pre-frail and frail, respectively [23]. Turkish validation of the FRAIL Scale was carried out by Hymabaccus et al. [24].

Physical activity (PA) level was assessed using the International Physical Activity Questionnaire Short Form (IPAQ-SF). It records the activity of four intensity levels, namely, vigorous-intensity, moderate-intensity, walking, and sitting [25]. When determining the total score, the duration and frequency of various activities were noted. Activities that lasted at least 10 min at a time were considered. The duration, number of days of the activity per week, and metabolic equivalent (MET) values were multiplied to obtain the ‘MET-minute/week’ value. The duration in minutes of walking, moderate activity, and vigorous activity were multiplied by 3.3, 4, and 8 METs, respectively. According to the total MET value, individuals’ PA levels were classified as low if they were < 600 MET-minutes/week, moderate if they were between 600 and 3000 MET-minutes/week, and high if they were > 3000 MET-minutes/week [26].

Data analyses were performed with IBM SPSS Statistics for Windows v26.0 (SPSS Inc., Chicago, IL, USA) and Hayes PROCESS macro v4.2. Before the statistical analysis was performed, the distribution of the data was examined using analytical (KS-SW tests) and visual methods (histogram and probability graphs). As the examinations revealed all quantitative variables were normally distributed, the Pearson correlation test was performed to evaluate bivariate associations [27].

To explore the complex associations among malnutrition, sarcopenia, and frailty, a moderation model and a moderated mediation model were generated with 5000 random sample bootstrapping confidence intervals using the Hayes PROCESS macro [28]. First, the mediator effect of sarcopenia on the association between malnutrition and frailty was tested using Model 4. Then, the moderator effect of PA level on this mediator effect was tested using Model 7. Malnutrition, frailty, sarcopenia, and PA level were regarded as the independent, dependent, mediator, and moderator variables, respectively. Both models were controlled for gender and age. A p value less than 0.05 was determined as an indicator of statistical significance.