Community-dwelling adults  aged ≥65 years who attended routine primary care appointments between November 2022 and March 2023 and were able to complete the SARC-F questionnaire and the 3-trial dominant-hand grip-strength test were eligible for inclusion. Patients were excluded if acute illness, recent upper-limb injury, severe arthritis, neurologic disease, or marked cognitive impairment precluded safe grip testing or questionnaire completion. These criteria reflect pragmatic screening practices and maximize both patient safety and data validity.

The SARC-F was selected as the screening tool of interest in this study. The SARC-F is a five question self-report survey developed by Malmstrom et al [7] to detect clinical symptoms of sarcopenia. The SARC-F questions include asking the patients to report difficulties with strength, assistance walking, rising from a chair, climbing stairs, and falls. The first four items are scored as 0 (no difficulty), 1 (some difficulty), or 2 (a lot of difficulty). Number of falls in the past year is rated as 0 (no falls), 1 (between 1‐3 falls), or 2 (4 or more falls). The sensitivity is low to moderate, and the specificity is high to predict low muscle strength when a cutoff value of ≥4 is used.

Muscle strength is the criterion used to detect probable sarcopenia in clinical settings [6]. Grip strength was selected as the measure of skeletal muscle strength because it is a quick and easy tool to administer during physician visits. Diagnosis of probable sarcopenia was assessed using the gender-specific recommended cutoff values for grip strength by the EWGSOP2 [6,14]. These values are <27 kg for men and <16 kg for women [14]. All grip tests were performed in private exam rooms by the first author. Participants sat with elbows flexed at 90°, wrists in a neutral position, and feet flat. Using a calibrated digital dynamometer (Sutekus Digital), each participant performed three maximal efforts (3‐5 s) with 30‐60 seconds of rest. The highest value for the dominant hand was used for analysis.

This study was approved by the Barton College Institutional Review Board (IRB #2022000034; approval date  January 25,  2023). As SARC-F screening and grip-strength testing are standard components of routine visits for adults 65 years or older at the study clinic, informed consent was not required as the data were obtained from deidentified medical records in accordance with the Health Insurance Portability and Accountability Act. All patient information was anonymized prior to analysis to ensure confidentiality. The collected data were anonymized, and no compensation was provided to participants.

Deidentified encounter records supplied data on age, sex, BMI, SARC-F score, and dominant-hand grip strength. Normality was assessed using Kolmogorov-Smirnov tests and histograms. Between-group differences were analyzed with independent 2-tailed t tests (parametric) or Mann-Whitney U tests (nonparametric). Receiver operating characteristic (ROC) analysis evaluated the ability of SARC-F to detect probable sarcopenia (EWGSOP2 grip-strength thresholds) and generated area under the curve (AUC) estimates with 95% CIs. Sensitivity, specificity, predictive values, and accuracy were calculated at cut points 2 and 4. Effect sizes (Cohen d or r) quantified the magnitude of differences. Post hoc power for the ROC (n=204; AUC=0.75) was 98.6%.

A ROC curve was used to determine a threshold (SARC-F score) that optimized the balance between sensitivity and specificity for diagnosing probable sarcopenia. The AUC was calculated to present the ability of the SARC-F score to discriminate between probable sarcopenic and nonsarcopenic individuals. An AUC of 1.0 indicates perfect discrimination capability, 0.5 indicates discrimination capability equal to that of chance, and 0.0 indicates that all subjects are incorrectly classified.

Sensitivity, specificity, positive predictive value, negative predictive value, and false positive rate were calculated for SARC-F threshold scores. Sensitivity was calculated as the number of participants diagnosed with probable sarcopenia that were correctly identified by the SARC-F screening. Specificity was calculated as the number of participants not diagnosed with probable sarcopenia that correctly screened negative with the SARC-F. Positive predictive value was calculated as the number of participants diagnosed with probable sarcopenia that screened positive with the SARC-F. Negative predictive value was calculated as the number of participants without probable sarcopenia that screened negative with the SARC-F. The false positive rate was calculated as the ratio of the number of participants screened positive by the SARC-F without probable sarcopenia to the number of participants who were not diagnosed with probable sarcopenia. Accuracy was also calculated at each SARC-F threshold as the proportion of correctly classified patients (both true positives and true negatives).

Comparisons of muscle strength between groups determined by the SARC-F threshold of 2 and previously recommended SARC-F threshold of 4 were performed, following the between-group comparison procedures listed above. When variances were not equal, Welch t test was used. The α was set at .05. All statistical analyses were performed using RStudio software (version 2023.06.1; Posit PBC).

Probable sarcopenia was present in 12% (n=24) of participants. Participant characteristics for age, BMI, grip strength, and SARC-F score are presented in Table 1. There was a significant difference in grip strength between men and women (t_99.51=−14.25; P<.001; d=1.95) and SARC-F score (U=6307; P<.001; r=0.24). The sex-specific distribution of SARC-F scores is illustrated in Figure 1. There was no significant difference between BMI of men and women (t_145.3=1.39; P=.17) or age (t₂₀₁=−0.134; P=.89).

Figure 2 presents the combined ROC curve for thresholds ≥2 and ≥4. The AUC for both thresholds was 0.752 (95% CI 0.66-0.84). We compared the diagnostic performance of SARC-F across the two commonly used cut points (≥2 vs ≥4). Using DeLong test for paired ROC curves, the AUCs were not significantly different (AUC  0.752 vs 0.752; P=.98), supporting the clinical preference for the more sensitive ≥2 threshold. A post hoc power analysis for the ROC curve revealed statistical power of 99.5%. Calculations for sensitivity, specificity, positive predictive value, negative predictive value, false positive rate, and accuracy for SARC-F cutoff scores are presented in Table 2.

Using a SARC-F cut point of  ≥2, participants classified as having probable sarcopenia (n=64) had higher SARC-F scores (t₅₉=16.7; P<.001), were older (t₈₀=3.3; P=.001), had higher BMI (t_76.9= 2.7; P=.009), and demonstrated lower grip strength (t₁₂₁=8.0; P<.001) than those with SARC-F <2. Using a cut point of  ≥4, SARC-F scores and grip strength differed significantly (t_66.9= 7.8; P<.001), whereas age and BMI were similar (P=.05 and P=.06, respectively). Mean values are summarized in Table 3.

The composite ROC curve for the two cut points is presented in Figure 2. The AUC for both thresholds was 0.752 (95% CI 0.66‐0.84). DeLong test showed no significant difference between AUCs (P=.98), supporting the clinical use of the more sensitive ≥2 threshold. Post hoc power for the ROC analysis was 99.5%. Diagnostic operating characteristics for each threshold are provided in Table 2.

A SARC-F cut point of ≥2 balanced sensitivity and specificity better than the traditional ≥4 threshold, identifying probable sarcopenia in 31% (n=63) of community-dwelling adults 65 years or older without adding clinic burden. Men demonstrated higher grip strength and lower SARC-F scores than women, reaffirming sex-specific muscle-strength disparities.

Earlier studies reported high specificity but modest sensitivity when applying a SARC-F  ≥4 [6,7,15]; our findings replicated this pattern (58% sensitivity, 88% specificity) while confirming that lowering the threshold to  ≥2 improves case finding (78% sensitivity) while maintaining acceptable specificity (75%). Our AUC of 0.75 aligns with the AUC of 0.71 as reported by Erbas Sacar et al [16], supporting the tool’s value as a screening and not a stand-alone diagnostic test. Recent authors have advocated thresholds as low as  ≥1 for maximal sensitivity [14,16,17]; our operating characteristic table (Table 2) illustrates the same trade-off: as the cut point increases, sensitivity decreases and specificity increases. DeLong test showed no difference between AUCs for the two thresholds (P=.98), strengthening the argument for the more sensitive ≥2 cut point in primary care.

First, real-world implementation during annual visits increases external validity. Second, standardized grip-strength testing minimized measurement error. Third, the sample (N=204) provided 98.6% post hoc power for ROC analyses. Limitations included (1) the single-site, cross-sectional design limits generalizability and causal inference; (2) SARC-F relies on self-report and may incur recall bias; (3) potential confounders (physical activity, cognitive status, comorbidities) were not captured; and (4) grip strength was measured once, and functional measures such as gait speed were unavailable. Each factor may attenuate or inflate the observed associations, underscoring the need for multimodal assessment in future research.

Prospective, multicenter studies should validate the ≥2 threshold across diverse settings, incorporate additional functional tests, and examine longitudinal outcomes (ie, falls, hospitalization, disability). Cost-effectiveness analyses could further justify routine SARC-F screening in primary care, and digital integration of the questionnaire into electronic health records may streamline population-level implementation.

A SARC-F cut point of  ≥2 offers a feasible, time-efficient approach to flag older primary care patients who require confirmatory strength testing, aligning with EWGSOP2 recommendations for early clinical intervention. The tool should be used to complement—rather than replace—comprehensive diagnostic workups.