The application value of frailty assessment tools in the perioperative period: A narrative review Authors: Haozhou Shen, Manling Ju, Xinyue Wu, Shanshan Shi, Jie Sun, Mi Tian Published: 2026 ABSTRACT Frailty is a clinical syndrome characterized by reduced physiological reserves and multisystem dysfunction, which markedly elevates the risk of postoperative complications and delayed functional recovery, and increases mortality. Amid the rapidly aging global population, the proportion of surgical patients with frailty is steadily increasing, presenting substantial challenges for perioperative management. Conventional preoperative assessment tools have significant limitations in identifying frailty and predicting surgical outcomes, while the clinical utility of emerging assessment methods requires further systematic evaluation. This review provides a comprehensive overview of preoperative frailty assessment tools. It critically examines recent advances from 2019 to 2025 and incorporates seminal foundational studies that established core concepts and validated key early instruments. The analysis evaluates their comparative efficacy, clinical utility in perioperative care, and existing limitations, aiming to offer scholars and clinicians a balanced and insightful perspective rather than an exhaustive summary of all available evidence. Introduction The acceleration of global aging has significantly increased the proportion of frail patients undergoing surgery.1 Frailty, characterized by diminished physiological reserve and multisystem dysfunction, is closely associated with postoperative complications, delayed functional recovery, and mortality risk.2, 3 Accurate identification of frailty is therefore essential for optimizing perioperative management and improving patient outcomes. However, traditional preoperative risk-assessment systems exhibit notable limitations in the effective identification of frailty and its specific impact on perioperative outcomes.3 The widespread adoption of Enhanced Recovery After Surgery (ERAS) pathways has encouraged innovation in perioperative evaluation strategies, and the importance of preoperative frailty assessment in optimizing surgical outcomes is widely recognized, as reflected in its inclusion in clinical practice guidelines internationally.3, 4 This shift from theory to practice has highlighted key clinical decisions regarding tool selection, evaluation timing, and the interpretation of predictions for specific outcomes. Yet challenges remain in applying these tools, including limited adaptation to perioperative contexts and a lack of operational standardization.5 This narrative review summarizes and evaluates current preoperative frailty assessment tools. It examines existing evidence and addresses these practical questions, with a focus on integrating frailty assessment into structured anesthetic care pathways. Methods The detailed methodology for literature search, study selection, and inclusion/exclusion criteria is provided in the Supplemental Materials. The study selection process is summarized in Fig. 1.Fig. 1The literature screening process. The literature screening process. The core theoretical framework of frailty Definition of frailty Frailty is a complex clinical syndrome defined by reduced physiological reserve and multisystem dysfunction.6 This condition increases an individual’s vulnerability to stressors and impairs their recovery capacity. While highly prevalent among older adults, frailty is not inevitable in the elderly nor absent in younger populations. It may occur independently of chronic diseases. Alternatively, it can coexist with conditions such as diabetes and cardiovascular disease, often accelerating their progression.7 Notably, frailty frequently involves psychosocial dimensions, encompassing cognitive decline (e.g., memory impairment), psychological issues (e.g., depression, anxiety), and social withdrawal.8, 9 Together, they form a vulnerability triad spanning physical, cognitive, and psychological domains. Frailty assessment frameworks: from theory to tool development Frailty research fundamentally requires reliable tools for identification and measurement. The development of these tools is guided by specific theoretical models, which directly shape the design and application of assessment instruments. Currently, two primary frameworks guide practice: the Cumulative Deficit Model and the Phenotypic Model.10 The Integral Conceptual Model is also gaining recognition for its holistic perspective. These models differ significantly in their theoretical foundations, assessment dimensions, measurement methodologies, and clinical applications. Phenotypic model: the direct embodiment of theory The Phenotypic Model defines frailty as a distinct biological syndrome. It is driven by a decline in multisystem physiological reserve and is characterized by a set of observable, well-defined physical manifestations.11 This theory underpins an assessment strategy based on identifying and counting these predefined signs. The model enables rapid clinical screening through five key physiological indicators: unintentional weight loss, reduced grip strength, slower gait speed, diminished physical activity, and self-reported exhaustion.12 The corresponding assessment tool operationalizes these manifestations. Each criterion—such as measuring weakness via grip strength or slowness via gait speed—represents a direct mapping of a theoretical construct onto a measurable item. A diagnosis of frailty is made when three or more of these manifestations are present. In the perioperative setting, this model evaluates the physical function and quantifies physiological reserve of patients. It uses five specific indicators to support objective preoperative risk stratification. This identifies patients with limited physical capacity and rehabilitation potential, guiding anesthesia and surgical teams in developing tailored management plans. Cumulative deficit model: quantified aggregation The Cumulative Deficit Model conceptualizes frailty as a nonspecific state of vulnerability arising from the accumulation of health deficits across multiple domains.13, 14 The core tenet is that the risk of adverse outcomes increases continuously with the ratio of existing deficits to potential deficits.15 Consequently, the assessment tool is designed to aggregate diverse health information comprehensively. Its output is a continuous score between 0 and 1, quantifying the graded vulnerability associated with deficit accumulation. In perioperative assessment, this model quantifies the overall accumulation of health deficits, thereby informing risk management. However, its design limits the identification of specific, modifiable clinical targets. Thus, its integration with other assessment tools supports the development of more targeted clinical interventions. Integral conceptual model: a holistic view The Phenotypic Model offers objectivity and operationalizability but inadequately addresses psychological and social dimensions.11 Similarly, while the Cumulative Deficit Model incorporates a broader range of factors, its quantitative approach falls short in capturing the intrinsic relationships and interactions among deficits across different domains.13, 14 To achieve a more comprehensive understanding, the Integral Conceptual Model serves as an essential complementary framework. This model defines frailty as a multidimensional health state. It results from the dynamic interplay of decline in physical, psychological, and social domains.16, 17 Assessment within this framework requires tools that measure deficits across all three domains simultaneously. This is typically achieved through questionnaires.16 The output is a composite score, which may be continuous or categorical. It is derived by aggregating the scores from each dimension. In the perioperative setting, these tools are used in preoperative clinics. They need no special equipment or training and can be done by interview or remotely, thereby enabling fast, reliable screening. However, the cognitive status of patients may affect the results, potentially causing underestimation of frailty or suboptimal intervention planning. Table 1 provides an overview of the correspondence between these three models and their representative assessment tools.Table 1Comparison of core features of three frailty assessment frameworks and tools.Models (Representative Tools)Tool Design LogicKey CharacteristicsPerioperative RelevancePhenotypic Model(FFP)Physical signs checklistStrengths: Objective, standardized, rapidLimitations: Lack of cognitive/psychosocial assessmentPhysical function assessmentPreoperative risk stratificationCumulative Deficit Model(FI, CFS, RAI)Deficit ratio calculationStrengths: Comprehensive, reflects overall risk burdenLimitations: Lack of target specificityGlobal risk quantificationNon-targeted assessmentIntegrative Conceptual Model(TFI, EFS)Domain-specific questionnaireStrengths: Holistic perspective, multidimensionalLimitations: Susceptibility to cognitive statusClinic-based rapid screeningCognitive status dependencyAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; TFI, Tilburg Frailty Indicator. Table 1Comparison of core features of three frailty assessment frameworks and tools.Models (Representative Tools)Tool Design LogicKey CharacteristicsPerioperative RelevancePhenotypic Model(FFP)Physical signs checklistStrengths: Objective, standardized, rapidLimitations: Lack of cognitive/psychosocial assessmentPhysical function assessmentPreoperative risk stratificationCumulative Deficit Model(FI, CFS, RAI)Deficit ratio calculationStrengths: Comprehensive, reflects overall risk burdenLimitations: Lack of target specificityGlobal risk quantificationNon-targeted assessmentIntegrative Conceptual Model(TFI, EFS)Domain-specific questionnaireStrengths: Holistic perspective, multidimensionalLimitations: Susceptibility to cognitive statusClinic-based rapid screeningCognitive status dependencyAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; TFI, Tilburg Frailty Indicator. Comparison of core features of three frailty assessment frameworks and tools. Abbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; TFI, Tilburg Frailty Indicator. Clinical classic assessment tools This section provides a detailed review of six classic clinical frailty assessment tools, including the Clinical Frailty Scale (CFS), Fried's Frailty Phenotype (FFP), Frailty Index (FI), FRAIL Scale, Edmonton Frail Scale (EFS) and the Risk Analysis Index (RAI). These instruments are grounded in the above-mentioned three frailty assessment frameworks, and have emerged as the most widely adopted tools in perioperative settings. For these reasons, they have been selected for in-depth discussion here. Clinical frailty scale (CFS) Origin and overview The CFS was originally conceived as a tool to summarize findings from the Comprehensive Geriatric Assessment (CGA) and facilitate the creation of personalized care plans.18 The assessment, which requires only 2–5 min to complete, quantifies frailty on a 9-point ordinal scale. The specific descriptors for scores 1 through 9 are: 1. Very Fit; 2. Fit; 3. Managing Well; 4. Living with Very Mild Frailty; 5. Living with Mild Frailty; 6. Living with Moderate Frailty; 7. Living with Severe Frailty; 8. Living with Very Severe Frailty; 9. Terminally Ill. Originally proposed by Rockwood et al. as a 7-point scale, the CFS was later expanded to 9 points based on clinical observations that patients with the most severe impairments (CFS 8–9) required distinct nursing interventions. This led to the official adoption of the 9-point scale in 2020.19, 20 The latest iteration (Version 2.0) further refined the descriptors: for instance, renaming the score of 2 from "Well" to "Fit," replacing "Vulnerable" (CFS 4) with "Living with Very Mild Frailty," and optimizing the semantic descriptions for scores of 5–8.21 These modifications reflect the clinical imperative for enhanced granularity in the assessment of the CFS. Association with short-term and long-term mortality risk Valued for its simplicity and reliability, the CFS has gained significant importance in emergency settings. Its score shows a strong and graded association with mortality risk in patients undergoing emergency surgery, trauma, and critical illness. Regarding short-term prognosis, higher CFS scores show a clear dose-dependent association with increased mortality. In emergency laparotomy, Parmar et al.22 found that patients with CFS 5 and CFS 6–7 had a 90-day mortality risk 3.18-fold (adjusted odds ratio [aOR] = 3.18; 95% CI, 1.24–8.14) and 6.10-fold (aOR = 6.10; 95% CI, 2.26–16.45) higher, respectively, than those with CFS 1. A similar trend was observed for 30-day mortality. In major trauma patients, Carter et al.23 observed that postoperative in-hospital mortality increased with higher CFS scores. Notably, even a preoperative CFS of 3 was associated with a significantly higher mortality risk compared to CFS 1–2 (adjusted hazard ratio [aHR] = 1.82; 95% CI, 1.39–2.40). In critically ill patients, an individual patient data meta-analysis of older adults (≥ 65 years) admitted to the ICU also showed that frailty (CFS ≥ 5) was independently associated with a significantly increased cumulative 30-day mortality risk (HR = 1.34; 95% CI, 1.25–1.44).24 This collective evidence underscores the crucial role of CFS in identifying high-risk patients early. For long-term outcomes, the association between the CFS and mortality remains significant, though its pattern is characterized by two key features. Firstly, a marked increase in risk typically occurs at a higher score threshold. Braude et al.25 found that preoperative frailty based on the CFS was associated with 1-year mortality after trauma, with risk increasing with frailty severity. However, there was no significant difference in 1-year mortality between trauma patients with CFS 1–3 and those with CFS 4 (aHR = 1.73; 95% CI, 0.89–3.36; p = 0.11). The significant increase in mortality risk was primarily observed in patients with CFS ≥ 5. Secondly, the strength of the predictive association tends to attenuate over time. A retrospective multicenter cohort study showed that in patients admitted to the ICU after elective cancer surgery, frailty (CFS ≥ 5) was associated with worse survival (HR = 1.72; 95% CI, 1.59–1.86).26 This association was strongest early in follow-up (10 months, HR = 1.39; 95% CI, 1.03–1.86) and gradually weakened, disappearing by 4 years (HR = 1.96; 95% CI, 0.73–5.28). A similar pattern was observed in patients admitted to the ICU with metastatic cancer. Alamgeer et al.27 reported that frailty (CFS 5–8) was associated with worse 4-year overall survival (29.5% vs. 10.9%; p < 0.0001) and shorter survival time (HR = 1.52; 95% CI, 1.43–1.60). This effect was more pronounced in patients under 65 (HR = 1.66; 95% CI, 1.51–1.83), suggesting that long-term survival is influenced by more complex multifactorial processes. This evidence indicates that the CFS threshold demonstrating a significant association with adverse outcomes varies depending on the timeframe of observation, which underscores the importance of early risk stratification. It also reveals notable heterogeneity across studies in the specific CFS cut-offs used to define high-risk frailty. Distinguish frailty from comorbidity burden Comorbidity burden, characterized by the presence of two or more chronic diseases and assessed by both disease count and severity, is prevalent and critical in frail patients.28, 29 Commonly quantified using the Charlson Comorbidity Index (CCI) or Elixhauser Comorbidity Index (ECI), high comorbidity burden is typically defined as a CCI score > 3.30, 31 Significant overlap in the pathogenesis, manifestations, and assessment of frailty and comorbidities leads to their frequent conflation by anesthesiologists.32 An interview study by Hladkowicz et al.33 identified time constraints and tool complexity as key barriers to routine frailty assessment—factors likely driving anesthesiologists’ reliance on comorbidity counting as a "shortcut". This may overclassify high-risk patients by including those with significant comorbidities but no frailty, prompting unnecessary interventions such as invasive monitoring, prolonged ICU stays, or surgical delays. Conversely, failing to identify truly frail patients denies them individualized pre-rehabilitation (e.g., nutritional support, functional training), elevating risks of postoperative complications such as delirium and functional decline.34, 35 The core logic of the CFS lies in its functional status synthesis rather than disease accumulation, offering unique value in distinguishing frailty from comorbidity burden.24, 36, 37 Consider a preoperative evaluation involving two 75-year-olds: one with diabetes and coronary heart disease who actively gardens, and another with only hypertension who requires breaks when climbing stairs. The CFS directly captures the functional decline in the latter—rating it 5 (Living with Very Mild Frailty)—while traditional comorbidity assessment might erroneously prioritize the former.18 Optimize clinical assessment and prognosis prediction by combining multiple indicators The CFS possesses significant integrative potential, enabling synergy with other metrics to enhance assessment accuracy. For instance, it can be integrated with the 57-item Frailty Index based on Comprehensive Geriatric Assessment (FI-CGA) for a comprehensive evaluation of geriatric domains.20, 38 The CFS is increasingly integrated into predictive models as a critical measure of physiological reserve. Hajibandeh et al. developed the HAS model for patients undergoing emergency laparotomy by combining the CFS with ASA classification and sarcopenia.39 In a retrospective study of 143 patients, this model demonstrated exceptional discrimination in predicting 30-day mortality (Area Under the ROC Curve [AUC] = 0.96, p < 0.0001) with gold-standard calibration. Furthermore, combining the CFS with frailty biomarkers refines prognostic accuracy. Aakre et al. conducted a prospective observational study to evaluate a composite model of CFS and hs-cTnT in patients ≥ 75 years having emergency abdominal surgery.40 They found that this composite model demonstrated superior discriminative ability for 30-day mortality compared to either marker alone. The predictive power of this composite model for 30-day mortality (AUC = 0.79; 95% CI, 0.68–0.88) was significantly better than troponin concentration alone (AUC = 0.69, 95% CI, 0.55–0.83) or frailty alone (AUC = 0.69; 95% CI 0.57–0.82). Limitations The CFS assessment relies on the assessor’s subjective judgment. Without specific training, this can lead to variability in scores between different raters.18 While the tool effectively reflects baseline frailty status, it is less sensitive in capturing dynamic changes during the course of an illness.41 In specific populations, such as patients with end-stage renal disease, the CFS alone may not fully identify all disease-related functional limitations.42 Furthermore, defining a universal frailty threshold with the CFS is challenging. Its value lies in assessing a continuous spectrum, and forcing rigid categories may diminish its clinical applicability. The optimal predictive threshold varies by clinical context, and evidence suggests that intervention should be considered early, even for patients with a CFS of 3.23, 43 Future efforts should focus on integrating CFS gradations with other indicators for precise, scenario-specific risk assessment and early intervention, rather than seeking a single diagnostic cutoff. Fried's frailty phenotype (FFP) Origin and overview The Fried's Frailty Phenotype (FFP) was first developed in 2001 by a research team led by Professor Linda P. Fried, a geriatric specialist at Johns Hopkins University.11 It has since become one of the most evidence-based and standardized tools in comprehensive geriatric assessment. Using a multimodal approach, the FFP integrates objective biological measures with a standardized self-rating questionnaire, covering five core dimensions: reduced gait speed, reduced grip strength, low physical activity, unintentional weight loss, and self-reported exhaustion. A dichotomous scoring system is applied, assigning 1 point if a criterion is met and 0 otherwise. The cumulative score determines the frailty stage: a FFP of 0 indicates robustness, 1–2 indicates prefrailty, and ≥ 3 indicates frailty. Integrate the intrinsic association between frailty and comorbidities More than 63% of elderly patients present with at least two comorbidities before surgery, and this population shows a significantly higher incidence of frailty.44, 45 Underlying conditions such as cardiovascular disease, diabetes, and chronic kidney disease can accelerate physiological decline through mechanisms like metabolic disorders and chronic inflammation.46, 47 This process often interacts with frailty, malnutrition, and sarcopenia, forming a vicious cycle that further increases perioperative risk.48, 49 The FFP assesses frailty using five physical indicators, which often result from diseases or their treatments. For example, low physical activity can be caused by heart failure or chronic obstructive pulmonary disease (COPD).50, 51 Unintentional weight loss often indicates malnutrition, malignancy, or depression.52, 53 Exhaustion is commonly seen in conditions such as cardiovascular disease, autoimmune disorders, and post-COVID-19 syndrome.54, 55, 56 A prospective study showed that in heart failure patients after mitral valve repair, the prevalence of frailty as assessed by the FFP dropped by 17% within six weeks.57 This improvement in frailty corresponded with the alleviation of heart failure symptoms. This cluster of symptoms reflects the depletion of physiological reserve. Also, it demonstrates that the FFP can indirectly evaluate the severity of comorbidities through functional performance. The FFP is not intended to replace traditional comorbidity assessment. Tools such as the CCI are typically used to evaluate comorbidity burden by systematically listing and quantifying diseases.31 The FFP, in contrast, evaluates the functional consequences and overall severity arising from these conditions. The two approaches are complementary: the CCI provides a "structured inventory" of disease burden, whereas the FFP reveals its "holistic functional impact." Together, they enable a more comprehensive clinical assessment. Association with perioperative neurocognitive disorders (PND) The core FFP indicators relate to surgery stress mechanisms like inflammation and metabolic dysregulation.58 These factors interact to promote PND.59, 60 Thus, the FFP may serve as an indicator for PND risk, based on the shared pathophysiological link between frailty and reduced brain functional reserve. Multiple studies support a significant association between the FFP and adverse neurocognitive outcomes. A meta-analysis by Aucion et al.,61 comparing 35 preoperative frailty tools, reported that among the instruments studied, the FFP showed a particularly strong association with postoperative delirium (POD) (OR = 3.79; 95% CI, 1.75–8.22). Gracie et al.62 reached a similar conclusion, reporting that frailty assessed by the FFP was significantly associated with POD risk in patients aged ≥ 65 years undergoing elective surgery (OR = 2.14; 95% CI, 1.43–3.19). These findings suggest that the FFP indicators, such as grip strength and gait speed, may reflect cerebral oxygen metabolic reserve capacity. The observational study by Bortone et al.63 offers further insight. Assessing 133 patients preoperatively with the FFP, it found a significantly higher incidence of POD in both prefrail (48% incidence; risk difference = 35%; 95% CI, 10%–51%) and frail patients (48% incidence; risk difference = 35%; 95% CI, 7%–53%) compared to robust patients (13% incidence). This study also identified a threshold effect of the FFP on postoperative cognitive function. Frail patients showed a significant decrease in composite cognitive Z-scores at 1 month postoperatively, however, this difference was not statistically significant at 1 year. Limitations Operationally, the FFP is complex and time-consuming (averaging >12 min),61 and its gait speed and grip strength tests require standardized equipment, hindering its use in resource-limited settings.64, 65 Indicators like fatigue rely on subjective reports prone to recall bias. Theoretically, it omits multi-omics data (e.g., inflammatory markers like IL-6) linked to frailty progression66, 67 and overlooks interactions with cognitive, psychological, and social factors.68 Regarding population applicability, its Western-centric weight loss criterion may underestimate risk in Asians, who often lose muscle mass without significant weight change.69 Frailty index(FI) Origin and overview The Frailty Index (FI) was developed by researchers at Canada's Dalhousie University through the systematic integration of clinical assessment data (typically 30–70 health deficit items).70 It covers multidimensional health domains, including chronic diseases (e.g., diabetes, cardiovascular diseases), cognitive impairment (MMSE scores), activities of daily living (ADL/IADL), and laboratory parameters (e.g., albumin, inflammatory markers).71 Each variable is termed a "deficit" and is scored from 0 (no deficit) to 1 (maximal deficit) based on severity, with the total score calculated as the sum of deficits divided by the total number of items, ranging from 0 to 1.70 While minor variations exist across studies, the commonly adopted classification includes robustness (FI ≤ 0.12), prefrailty (0.12 < FI < 0.25) and frailty (FI ≥ 0.25). Robust adaptability, flexibility, and diversity Given the complexity of clinical contexts, variability in research data, and specificity of target conditions, the FI demonstrates significant adaptability. Researchers can adjust the core assessment items of the FI to achieve customization. This flexibility has led to the development of numerous derivative versions. Among these, the modified Frailty Index 11 (mFI-11) and modified Frailty Index 5 (mFI-5) were the first to emerge. The mFI‑11 is derived from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database and includes 11 binary (yes/no) variables: hypertension requiring medication, insulin‑dependent diabetes, chronic airway obstruction (asthma/COPD), congestive heart failure, history of myocardial infarction, history of percutaneous coronary intervention, history of angina, prior stroke or transient ischemic attack (TIA), preoperative non‑independent functional status, history of pneumonia, and dialysis‑dependent end‑stage renal disease.72, 73 Each item is scored 1 (present) or 0 (absent). The mFI‑11 score is the proportion of deficits present, with ≥ 0.27 indicating frailty. Subramaniam et al. simplified the mFI-11 into the mFI-5, which includes five variables: non‑independent functional status, diabetes, congestive heart failure, COPD, and hypertension.72 The mFI‑5 strongly correlated with the mFI-11 (Spearman’s ρ = 0.9) across nine surgical specialties, with both versions showing robust predictive power for mortality and postoperative complications. The Electronic Frailty Index (eFI), an automated tool derived from electronic health records (EHRs), uses 36 variables to stratify patients as robustness (eFI ≤ 0.10), prefrailty (0.10 < eFI ≤ 0.21), or frailty (eFI > 0.21).74 In a cohort of 33,449 patients aged ≥ 55 undergoing major non-cardiac surgery, Khanna et al.75 found that eFI-assessed frailty was associated with increased postoperative adverse events. Prefrail patients had significantly higher risks compared to robust patients (OR = 1.24; 95% CI, 1.18–1.30; p < 0.001) and frail (OR = 1.71; 95% CI, 1.58–1.82; p < 0.001), supporting eFI as an effective tool for standardized perioperative risk assessment. Beyond these, other versions such as the Chart‑derived Frailty Index (CFI), Laboratory Frailty Index (FI‑Lab), and Liver Frailty Index (LFI) have been developed for diverse needs, though they are not detailed here. While adapting items is a core FI strength, the proliferation of variants with different methods and thresholds compromises cross-study comparability and standardization. Longitudinal association and its temporal dynamics The FI demonstrates a dynamic, longitudinal association with patient outcomes. This means that the relationship between the FI score and outcomes such as mortality risk evolves over time, mirroring the longitudinal progression of the FI itself. This temporal evolution means that longitudinal FI measurements can more accurately capture an individual's evolving frailty status and its relationship to risk than a single baseline assessment. A German population-based cohort study by Li et al.76 utilized longitudinal FI measurements to analyze the frailty-mortality association. The analysis using time-dependent FI values revealed a robust and stable association with all-cause and cardiovascular mortality over 14 years. In contrast, the association estimated from a single baseline FI measurement attenuated over time. This highlights the value of longitudinal assessment in capturing the stable link between evolving frailty and long-term risk. A U.S. cohort study using NHANES data (1999–2010) reported that despite an increasing population-level frailty burden over time, the strength of the association between the FI and 5-year mortality risk remained consistent across different historical cohorts.77 This supports the FI's reliability as a robust indicator of mortality risk over the medium-to-long term. The stable longitudinal association between FI and mortality underscores its value for medium-to-long-term risk stratification in surgical populations, particularly in older adults. Moreover, because the FI encompasses a breadth of health deficits, it can also help identify potential targets for preoperative optimization, such as addressing nutritional status, managing comorbidities, or improving physical function. Limitations Although longitudinal FI assessment more accurately reflects risk association compared to single-timepoint measurements, it relies on repeated assessments that may be impractical in time-sensitive surgical settings.78 This limits its use in emergencies or during rapid preoperative evaluations, where clinical conditions always change acutely. To accommodate clinical heterogeneity, numerous derivatives of the FI have emerged. However, the proliferation of the FI variants widely differing methodologies and threshold definitions (e.g., frailty defined as the score > 0.21 vs. ≥ 0.25) introduces substantial inconsistency in cross-study comparability and weakens standardization.79, 80 The predictive performance of the FI may be suboptimal in specific patient populations (e.g., cancer or critically ill patients), necessitating further validation of its prognostic value in these cohorts.81 Furthermore, the FI implementation is contingent upon high-quality data input. The current limited recognition of frailty by surgeons and insufficient interdisciplinary collaboration lead to inconsistent data quality, which undermines the authenticity of the FI assessments.78 FRAIL scale Origin and overview Developed in 2008 by the International Association of Nutrition and Aging (IANA) expert panel, the FRAIL Scale integrates core frameworks including the FFP, the FI, Short Physical Performance Battery (SPPB), and Mini Nutritional Assessment (MNA) into a clinically efficient screening tool.82 This hybrid instrument synergizes the cumulative deficit and phenotypic models through five domains: Fatigue (subjective exhaustion), Resistance impairment (inability to climb one flight of stairs), Ambulation limitation (difficulty walking ≥ 100 m), Illness burden (≥ 5 comorbidities), and Loss of weight (unintentional > 5% weight loss in prior year). Using a binary scoring system ("yes" = 1; "no" = 0), it stratifies patients into: robustness (0), prefrailty (1–2), and frailty (≥ 3) with completion in ≤ 3 min via self-report or clinician interview.83 Association with short-term postoperative outcomes in the elderly Analogous to the CFS, the FRAIL Scale serves as a pragmatic rapid screening tool in resource-limited or urgent settings.84 Accumulating evidence indicates a pronounced association between frailty as identified by the FRAIL Scale and adverse short-term postoperative outcomes in elderly surgical patients. A meta-analysis by Gong et al.85 of 18 studies (4479 patients) found that frailty defined by the FRAIL scale was strongly associated with significantly higher odds of 30-day mortality (OR = 6.62; 95% CI, 2.80–15.61) and 6-month mortality (OR = 2.97; 95% CI, 1.54–5.72), but no significant increase in 1-year mortality (OR = 1.54; 95% CI, 0.91–2.58). Consistently, a retrospective analysis86 of 558 surgical patients aged ≥ 65 years reported that frail patients had an 11% absolute increase in 180-day postoperative mortality compared to non-frail patients (12.2% vs. 1.4%; 95% CI, 8.4–13.6). Further analysis within the same meta-analysis revealed that frailty per the FRAIL Scale was associated with a more than two-fold increase in the risk of overall postoperative complications (OR = 3.11; 95% CI, 2.06–4.68) and a particularly elevated risk of POD (OR = 2.65; 95% CI, 1.85–3.80).85 Another prospective cohort study87 reported frail or prefrail patients had nearly triple the risk of POD (OR = 2.7; 97.5% CI, 1.0–7.3) compared to robust patients, though no significant difference was observed in 3-month postoperative cognitive decline (POCD). These findings suggest that the predictive utility of the scale may be more specific to acute brain dysfunction (e.g., delirium) than to long-term cognitive impairment. The FRAIL Scale, derived from the Frailty Phenotype (FFP) framework, serves as its rapid screening counterpart. Given that the FFP itself is associated with POD, this shared theoretical basis may partly explain the observed strong association between the FRAIL Scale and POD. Therefore, the FRAIL Scale is best viewed as an efficient screening instrument for identifying at-risk individuals, rather than a substitute for comprehensive, multidimensional frailty assessment. Reveal the correlation between frailty and postoperative pain Frail patients exhibit reduced physiological reserve and declined metabolic function during the perioperative period, which may impair tissue repair. It is also associated with reduced metabolic function and sustained release of inflammatory factors, such as IL‑6 and CRP.88 Therefore, frail patients generally experience more intense postoperative pain compared to non‑frail individuals. Among various frailty assessment tools, the FRAIL scale has been specifically investigated for its association with postoperative pain outcomes. In a prospective study, Esses et al.89 found higher scores on the FRAIL scale were linked to a nearly four-fold increase in the risk of moderate-to-severe breakthrough pain at 3 months postoperatively (OR = 4.73; 95% CI, 1.24–18.09). Similarly, a cross-sectional study by Cai et al.90 demonstrated that preoperative screening with the FRAIL scale could identify patients at risk for more severe acute postoperative pain (B = 0.307, p < 0.01). These findings suggest that the FRAIL scale may serve as a useful clinical indicator for identifying patients at higher risk of severe postoperative pain. Further research should explore its predictive value across different surgical settings. Limitations In pre-elderly populations (55–60 years), the FRAIL scale exhibits low sensitivity (41.3%) despite high specificity (93.2%), reducing its reliability for early frailty detection.91 It also omits multidimensional domains (e.g., cognitive, psychological) and objective biological parameters, resulting in an incomplete assessment and lower precision compared to emerging biomarker-inclusive tools.83, 92, 93 Moreover, it is best suited for rapid community or primary care screening rather than for complex perioperative decision-making.83 Edmonton frail scale (EFS) Origin and overview The EFS, developed in 2006 by Canadian scholar Rolfson and his colleagues, is a multidimensional assessment instrument designed for efficient and comprehensive evaluation of frailty in older adults.94 It encompasses 11 items covering nine core domains: cognitive function (assessed by the Clock Drawing Test), overall health status, functional independence, social support, medication use, nutritional status, mood, continence, and physical performance (evaluated by the Timed Up and Go Test). The total score ranges from 0 to 17, with the severity of frailty stratified as follows: scores of 0–4 indicate fit/robust; 5–6 suggest vulnerable; 7–8 represent mild frailty; 9–10 correspond to moderate frailty and ≥ 11 signify severe frailty. Association with postoperative complications Postoperative complications represent a major cause of prolonged hospitalization, increased readmission rates, and functional decline.95 Higher EFS scores have been positively associated with both an increased risk and greater severity of postoperative complications. A prospective study in older Asian patients undergoing major abdominal surgery96 identified the EFS score as an independent factor associated with postoperative morbidity (OR = 1.35, p < 0.001). Also,this study found that each 1‑point increase in the EFS score raised the CCI score by approximately 3 points (Coefficient b = 2.944, p < 0.001), suggesting a dose‑response relationship between frailty and complication severity. A meta‑analysis by Aucoin et al. provided further support for this association.61 In comparing multiple frailty instruments, this analysis concluded that, among the outcomes studied, the EFS showed the strongest association with postoperative complications (OR = 2.93; 95% CI, 1.52–5.65). The included studies used varying definitions of “postoperative complications,” ranging from general criteria such as the Clavien-Dindo classification to procedure-specific event (e.g., sternal wound infection or delayed graft function for kidney transplant). Nevertheless, some evidence presents a contrasting view regarding the predictive utility of the EFS. Nishijima et al.97 compared frailty assessment tools for predicting composite adverse events (AEs) (Clavien-Dindo grade ≥ 2 complications or nursing facility discharge) after cancer surgery. They found the EFS score showed no significant predictive association with these outcomes, indicating its predictive validity is context-specific. This suggests that the predictive validity of the EFS may vary depending on the clinical context and the outcomes of interest. Pinpoint key intervention targets for loss of independence (LOI) Loss of independence (LOI), defined as needing increased care support after discharge, is a key patient-centered outcome for older surgical patients.98 By identifying high-risk patients through its multidimensional evaluation, the EFS offers distinct advantages for targeted interventions. In a retrospective cohort study by Owodunni et al.,99 patients who experienced LOI had significantly higher baseline EFS scores than those who did not (7 vs. 3 points; P < 0.001). Multivariable analysis confirmed that frailty assessed by the EFS was a strong independent predictor of LOI (OR = 12.72; 95% CI, 12.04–13.44; p < 0.001). Beyond risk stratification, the EFS also enables identification of specific modifiable factors for intervention. In a large retrospective cohort study, Goeddel et al.100 reported that the total EFS score was independently associated with LOI (OR = 1.37; 95% CI, 1.30–1.45; P < 0.001). They identified four intervenable domains: physical function, functional dependence, social support, and urinary incontinence. A prediction model using only these domains showed discrimination comparable to the full EFS (AUC = 0.838 vs. 0.840), highlighting the dual utility of the EFS for risk identification and prehabilitation targeting. Limitations The EFS relies heavily on patient self-reporting and cognitive engagement, which may lead to inaccurate assessments in older adults with cognitive impairment.94 It is also less sensitive to rapid functional decline caused by acute illness or sudden health changes, limiting its ability to capture dynamic deterioration.95 Furthermore, certain components (e.g., the clock-drawing test and timed up-and-go test) require standardized administration by trained professionals,94 and inconsistent execution could affect reliability. Notably, the EFS may be less suitable for rapid evaluation in settings such as emergency surgery or for clinically unstable patients. Because these cognitive and functional assessments in the EFS require time and patient cooperation. Finally, its measurement reliability and consistency across different cultural settings remain variable and require further validation.101 Risk analysis index (RAI) Origin and overview The Risk Analysis Index (RAI) is a frailty screening and perioperative risk assessment tool specifically designed for surgical populations. Daniel E. Hall et al. refined the Minimum Data Set (MDS) Mortality Risk Index–Revised (MMRI-R) to create the Clinical RAI (RAI-C) questionnaire, which incorporates 14 items derived from 11 variables and 2 statistical interaction terms.102, 103, 104 The RAI-C can be administered by healthcare providers in approximately two minutes using patient-reported information. Its core variables include: age, sex, dependence in activities of daily living, cognitive impairment, history of cancer, dyspnea, heart failure, weight loss, sleep disturbances, dialysis dependence, and non-independent living situation. Additionally, the tool accounts for two interaction effects: between functional dependence and cognitive impairment, and between cancer and weight loss. Scoring is based on a weighted cumulative system, with total scores ranging from 0 to 81. A score ≥ 11 serves as an initial screening cutoff (sensitivity 72%), while a score ≥ 21 indicates frailty (sensitivity 50%, specificity 82%). To support retrospective studies, the team developed the “Administrative RAI” (RAI-A), using variables commonly available in databases such as the Veterans Affairs Surgical Quality Improvement Program (VASQIP) or ACS-NSQIP. The RAI-A shows significant correlation with the RAI-C (r = 0.48). At a threshold of ≥ 21, only 3.7% of patients were classified as frail. Although the RAI-A demonstrated low sensitivity (25%) for predicting 180-day mortality, it achieved high specificity (97%). Consequently, the EFS is more appropriate for retrospective analysis than real-time clinical screening. To improve generalizability, Daniel E. Hall et al. further recalibrated the RAI using the VASQIP database, resulting in the Revised RAI (RAI-rev).105 The RAI-rev classifies patients into four categories: fit/robust (≤ 20), normal (21–29), frail (30–39), and very frail (≥ 40). This same criteria has likewise been adopted for use with the RAI score in some research studies.106, 107 Robust association with and prediction of postoperative mortality The RAI, which evolved from its predecessor the MMRI-R, was specifically developed and validated as a tool to predict postoperative mortality. Its predictive validity is supported by studies evaluating its measurement properties and comparative performance. Furthermore, the clinical utility of this predictive capacity is demonstrated by intervention studies. For instance, Varley et al.109 reported that implementing a frailty screening initiative (FSI) based on the RAI was associated with a significant 18% reduction in 365-day postoperative mortality (OR = 0.82; 95% CI, 0.72–0.92; p < 0.001).108 A prospective study110 quantified this dose-response relationship, demonstrating that each one-point increase in the RAI score corresponds to an approximately 10% increase in mortality hazard (Hazard Ratio [HR] = 1.099; 95% CI, 1.091–1.106; p < 0.001). The predictive performance of the RAI is further highlighted in head-to-head comparisons with other commonly used assessment tools. In a retrospective study of 121,095 emergency general surgery patients, Grudzinski et al.109 compared frailty tools including the FI, the RAI, and the Hospital Frailty Risk Score (HFRS). The results showed the RAI outperformed other frailty indices in predicting both 30-day and 365-day mortality, as measured by discrimination and net reclassification improvement (NRI). More decisively, a study (using Bayesian analysis)10 specifically comparing RAI-A and mFI-5 in adults who underwent bowel surgery or intermediate to high risk mixed noncardiac surgery reported that, only RAI-A had a ≥ 95% probability of a non-zero association with mortality. In this study, the predictive accuracy of the RAI-A was significantly superior to that of mFI-5 (minimum Bayes factor 3.25 × 10) and even surpassed the widely used NSQIP surgical risk calculator itself (minimum Bayes factor 4.27 × 10). Additionally, this predictive utility is also corroborated in an ambulatory surgery population (n = 28,059), where patients identified as frail by the RAI had a 2- to 5-fold higher mortality risk than those with normal scores.110 Together, evidence from calibration (graded association), discrimination (comparative performance), and impact (intervention study) studies consistently establishes the RAI as a robust and clinically actionable tool for stratifying postoperative mortality risk. Association with specialty-specific surgical outcomes Beyond mortality, the RAI shows significant associations with clinically relevant outcomes across various surgical specialties, such as non-home discharge (NHD). NHD reflects postoperative functional independence and heightened rehabilitation needs, particularly in fields like spinal and neurosurgery. Evidence supporting its use for risk stratification is strong. In spine surgery, Covell et al. demonstrated that the RAI had good discriminative accuracy for adverse discharge outcomes, including NHD(C-statistic: 0.71, 95% CI: 0.69–0.74, p < 0.001).111 A clear dose-response relationship is observed: higher RAI scores are strongly associated with an increased risk of NHD. This pattern is consistent in neurosurgical populations. For instance, in patients undergoing epilepsy surgery or ventriculoperitoneal shunt placement, NHD rates rise significantly with increasing RAI scores.107, 114 These findings support the utility of the RAI in identifying patients at high risk for loss of independence. Moreover, the RAI scores correlate strongly with major complications across specialties. In patients undergoing lower-extremity amputation, frailty status was significantly associated with higher risks of deep vein thrombosis, sepsis, myocardial infarction, acute kidney injury, and other complications.112 In endovascular aortic aneurysm repair, patients with an RAI ≥ 30 had a complication rate of 43%, substantially higher than the 21% observed in robust individuals.113 Collectively, these studies demonstrate a consistent, dose-dependent association between higher RAI scores and a greater risk of adverse surgical outcomes. Notably, while the RAI effectively predicts specialty-specific surgical outcomes, clinicians must consider that reversible symptoms may temporarily elevate scores, requiring careful interpretation. Redefine surgical risk stratification Frailty reevaluates the inherent risks of surgery by shifting the focus from the procedure alone to the patient’s physiological resilience. This paradigm shift is captured in the RAI, which emphasizes that postoperative risk arises from the interaction between frailty and surgical stress. A retrospective study by Hall et al.114 found that even in low-stress procedures like cystoscopy, RAI-defined frail and very frail patients had 30-day mortality rates of 1.55% and 10.34%, respectively. For moderate-stress surgeries like cholecystectomy, rates rose to 5.13% and 18.74%. Mortality rates were observed to increase over time, reaching 43.00% at 180 days among the very frail. These data demonstrate that even surgical procedures traditionally considered "low-risk" are associated with substantial mortality in patients identified as frail by the RAI. Another large retrospective cohort study (n = 984,550)107 confirmed that the aforementioned risk trend applies to both low-mortality-risk (≤ 1%) and high-mortality-risk (> 1%) surgeries. This clearly shows that even low- and moderate-stress operations can indeed carry substantial risks for frail individuals. Iterative development The RAI has undergone iterative development to enhance performance and expand applicability, resulting in multiple adapted versions. To improve predictive accuracy, the RAI-rev was developed from the original version.105 It showed significantly improved discrimination for 30-day (C-statistic: from 0.84 to 0.86), 180-day (from 0.81 to 0.84), and 365-day mortality (from 0.78 to 0.82; p < 0.001). Calibration improved markedly, with the median absolute difference between predicted and observed 180-day mortality decreasing from 8.45% to 1.23%. External validation using NSQIP data confirmed its reliability (30-day mortality C-statistic: 0.87; calibration error: 0.6%). To overcome dependence on clinical interviews, an International Classification of Diseases -based version (RAI-ICD) was developed by mapping 11 RAI parameters to 323 ICD-10-Clinical Modification codes.115 It demonstrated excellent discrimination for in-hospital mortality (C-statistic: 0.810) across multiple cohorts (C-statistic: 0.778–0.860) and proposed frailty strata: robust (< 27), normal (27–35), frail (36–45), and very frail (> 45). A modified RAI (mRAI) was also developed for specific settings, classifying patients as robust (0–10), frail (11–20), or very frail (≥ 21).116 Very frail patients had significantly higher rates of adverse outcomes (25.0% vs. 21.5%), non-home discharge (91.0% vs. 72.3%), and in-hospital mortality (29.2% vs. 10.9%). All these developments highlight the adaptability and clinical utility of the RAI system. Limitations First, inconsistent data standards across clinical databases complicate its mapping and may compromise score accuracy and generalizability of the RAI.117 Second, incorporating biomarkers like albumin into the RAI has not significantly improved mortality prediction.118 This suggests that simply augmenting the tool with additional objective physiological parameters may offer limited incremental predictive value. Third, the RAI score has limited interpretability, which may hinder transparent communication and shared decision-making between doctors and patients. Additionally, high-risk scores could bias clinicians towards more conservative management. Finally, the predictive performance of the RAI and its derived tools varies across racial and ethnic groups.119 Therefore, calibration for specific populations is needed before widespread implementation. Table 2 compares their basic operational properties, while Table 3 contrasts their applicability and efficacy.Table 2Comparison of core operational characteristics of classic frailty assessment tools.ToolsData SourceAdministration TimeEquipment RequiredTraining RequiredCFSClinical judgmenta1–2 minsNoneLowFFPDirect measurements & Questionnaireb> 10 minsObjective measurement toolsgHighFICumulative health deficitsc> 15 minsHealth deficit dataHighFRAIL ScalePatient self-report questionnaired1–3 minsNo equipment requiredLowEFSPatient interview & Simple testse5–10 minsObjective measurement toolshModerateRAIPatient self-report informationfAbout 2 minsNoneLowAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI-rev, Revised Risk Analysis Index; RAI-C, Clinical Risk Analysis Index; RAI-A, Administrative Risk Analysis Index. aThe assessor rates the patient on a 9-point scale based on overall health, comorbidities, and functional status, typically drawing from medical history, physical examination, and communication with the patient/family. bIncludes 5 criteria. Data comes from direct physical performance tests (e.g., grip strength, gait speed) and structured questionnaires (e.g., exhaustion, weight loss). cCalculates the proportion of health deficits present out of the total number of items (often 30–40 +) considered. Data can be sourced from comprehensive geriatric assessment (CGA), medical record review, or specific health surveys. dTypically administered via patient self-report or proxy response to 5 questions regarding fatigue, resistance, illnesses, weight loss, etc. eCombines a structured patient interview (inquiring about cognition, health status) with simple performance tests (e.g., Timed Up-and-Go test) for a multidimensional assessment. fThis pertains to the RAI-C, a clinical questionnaire with 14 items, completed by healthcare providers via a patient interview. The RAI-A or RAI-rev are computational tools based on optimized algorithms and databases. They do not involve direct patient assessment and therefore do not require training in traditional bedside assessment skills. gRequires a dynamometer (for grip strength), a stopwatch and a fixed distance (e.g., 3–5 m for gait speed), and a scale (for weight loss). hRequires an armchair, a stopwatch (for Timed Up-and-Go test) and simple cognitive test tools (e.g., paper and pen for Clock Drawing Test).Table 3Comparison of the applicability and efficacy of classic frailty assessment tools.ToolsOptimal SettingStrengthsLimitationsCFSaPreop clinic,CGA, ICU,Emergency and trauma surgery(1) Strong association with mortality(2) Rapid administration(3) Distinguishes frailty from comorbidity burden(1) Subjectivity(2) Limited dynamic tracking(3) Context-dependent thresholdsdFFPcPreop clinic,Research,Community(1) Strong association with PND(2) Reflects functional impact of comorbidities(3) Incorporates objective performance measures(1) Relatively complex operation(2) Omits psychological factors(3) Population applicabilityeFIbCGA,Research(1) Highly flexible and Customizable(2) Stable longitudinal association with long-term outcomes(3) Comprehensive dimensions(1) High data requirement(2) Inconsistent standards(3) Limited in emergency settingsFRAIL ScaleaPreop clinic, Community,Telephone screening,Primary care initial screening(1) Ultra-rapid screening(2) Association with postoperative pain outcomes(3) Identifies risk of short-term adverse outcomes(1) Limited sensitivity(2) Relatively narrow dimensionsEFSbPreop clinic,CGA, Community(1) Multidimensional assessment(2) Identifies risk and severity of postoperative complications(3) Pinpoints modifiable targets for prehabilitation(1) Requires training(2) Insensitive to rapid decline(3) Limited in cognitive impairmentRAIcPreop clinic,Research,Surgical risk stratification(1) Validated predictor of postoperative mortality(2) Identifies risk of specialty-specific adverse outcomes(3) Redefines risk stratification(1) Challenging data mapping,(2) Requires population-specific calibration(3) Potential management biasfAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; CGA, Comprehensive Geriatric Assessment; ICU, Intensive Care Unit; PND, Postoperative Neurocognitive Disorders; LOI, loss of independence; Preop, preoperative. aRapid Screening Tools. Ultra-fast (< 5 min), ideal for triaging patients in busy settings like the emergency department, primary care, or community screening. bComprehensive Assessment Tools. Multidimensional, thorough but more time-consuming. Provides a detailed frailty profile and specific deficit areas for developing intervention plans. cSpecialty/Research-Oriented Tools.Targets specific outcomes (e.g., surgery) or relies on specific data systems. High predictive value in specific domains. dThe optimal cutoff for predicting different outcomes (short-/long-term mortality) may vary with clinical context. eCriteria based on Western populations (e.g., weight loss) may not fully apply to others (e.g., Asians). fTotal score offers limited transparency for doctor-patient communication; high scores may unintentionally bias towards conservative management. Table 2Comparison of core operational characteristics of classic frailty assessment tools.ToolsData SourceAdministration TimeEquipment RequiredTraining RequiredCFSClinical judgmenta1–2 minsNoneLowFFPDirect measurements & Questionnaireb> 10 minsObjective measurement toolsgHighFICumulative health deficitsc> 15 minsHealth deficit dataHighFRAIL ScalePatient self-report questionnaired1–3 minsNo equipment requiredLowEFSPatient interview & Simple testse5–10 minsObjective measurement toolshModerateRAIPatient self-report informationfAbout 2 minsNoneLowAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI-rev, Revised Risk Analysis Index; RAI-C, Clinical Risk Analysis Index; RAI-A, Administrative Risk Analysis Index. aThe assessor rates the patient on a 9-point scale based on overall health, comorbidities, and functional status, typically drawing from medical history, physical examination, and communication with the patient/family. bIncludes 5 criteria. Data comes from direct physical performance tests (e.g., grip strength, gait speed) and structured questionnaires (e.g., exhaustion, weight loss). cCalculates the proportion of health deficits present out of the total number of items (often 30–40 +) considered. Data can be sourced from comprehensive geriatric assessment (CGA), medical record review, or specific health surveys. dTypically administered via patient self-report or proxy response to 5 questions regarding fatigue, resistance, illnesses, weight loss, etc. eCombines a structured patient interview (inquiring about cognition, health status) with simple performance tests (e.g., Timed Up-and-Go test) for a multidimensional assessment. fThis pertains to the RAI-C, a clinical questionnaire with 14 items, completed by healthcare providers via a patient interview. The RAI-A or RAI-rev are computational tools based on optimized algorithms and databases. They do not involve direct patient assessment and therefore do not require training in traditional bedside assessment skills. gRequires a dynamometer (for grip strength), a stopwatch and a fixed distance (e.g., 3–5 m for gait speed), and a scale (for weight loss). hRequires an armchair, a stopwatch (for Timed Up-and-Go test) and simple cognitive test tools (e.g., paper and pen for Clock Drawing Test). Comparison of core operational characteristics of classic frailty assessment tools. Abbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI-rev, Revised Risk Analysis Index; RAI-C, Clinical Risk Analysis Index; RAI-A, Administrative Risk Analysis Index. aThe assessor rates the patient on a 9-point scale based on overall health, comorbidities, and functional status, typically drawing from medical history, physical examination, and communication with the patient/family. bIncludes 5 criteria. Data comes from direct physical performance tests (e.g., grip strength, gait speed) and structured questionnaires (e.g., exhaustion, weight loss). cCalculates the proportion of health deficits present out of the total number of items (often 30–40 +) considered. Data can be sourced from comprehensive geriatric assessment (CGA), medical record review, or specific health surveys. dTypically administered via patient self-report or proxy response to 5 questions regarding fatigue, resistance, illnesses, weight loss, etc. eCombines a structured patient interview (inquiring about cognition, health status) with simple performance tests (e.g., Timed Up-and-Go test) for a multidimensional assessment. fThis pertains to the RAI-C, a clinical questionnaire with 14 items, completed by healthcare providers via a patient interview. The RAI-A or RAI-rev are computational tools based on optimized algorithms and databases. They do not involve direct patient assessment and therefore do not require training in traditional bedside assessment skills. gRequires a dynamometer (for grip strength), a stopwatch and a fixed distance (e.g., 3–5 m for gait speed), and a scale (for weight loss). hRequires an armchair, a stopwatch (for Timed Up-and-Go test) and simple cognitive test tools (e.g., paper and pen for Clock Drawing Test). Table 3Comparison of the applicability and efficacy of classic frailty assessment tools.ToolsOptimal SettingStrengthsLimitationsCFSaPreop clinic,CGA, ICU,Emergency and trauma surgery(1) Strong association with mortality(2) Rapid administration(3) Distinguishes frailty from comorbidity burden(1) Subjectivity(2) Limited dynamic tracking(3) Context-dependent thresholdsdFFPcPreop clinic,Research,Community(1) Strong association with PND(2) Reflects functional impact of comorbidities(3) Incorporates objective performance measures(1) Relatively complex operation(2) Omits psychological factors(3) Population applicabilityeFIbCGA,Research(1) Highly flexible and Customizable(2) Stable longitudinal association with long-term outcomes(3) Comprehensive dimensions(1) High data requirement(2) Inconsistent standards(3) Limited in emergency settingsFRAIL ScaleaPreop clinic, Community,Telephone screening,Primary care initial screening(1) Ultra-rapid screening(2) Association with postoperative pain outcomes(3) Identifies risk of short-term adverse outcomes(1) Limited sensitivity(2) Relatively narrow dimensionsEFSbPreop clinic,CGA, Community(1) Multidimensional assessment(2) Identifies risk and severity of postoperative complications(3) Pinpoints modifiable targets for prehabilitation(1) Requires training(2) Insensitive to rapid decline(3) Limited in cognitive impairmentRAIcPreop clinic,Research,Surgical risk stratification(1) Validated predictor of postoperative mortality(2) Identifies risk of specialty-specific adverse outcomes(3) Redefines risk stratification(1) Challenging data mapping,(2) Requires population-specific calibration(3) Potential management biasfAbbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; CGA, Comprehensive Geriatric Assessment; ICU, Intensive Care Unit; PND, Postoperative Neurocognitive Disorders; LOI, loss of independence; Preop, preoperative. aRapid Screening Tools. Ultra-fast (< 5 min), ideal for triaging patients in busy settings like the emergency department, primary care, or community screening. bComprehensive Assessment Tools. Multidimensional, thorough but more time-consuming. Provides a detailed frailty profile and specific deficit areas for developing intervention plans. cSpecialty/Research-Oriented Tools.Targets specific outcomes (e.g., surgery) or relies on specific data systems. High predictive value in specific domains. dThe optimal cutoff for predicting different outcomes (short-/long-term mortality) may vary with clinical context. eCriteria based on Western populations (e.g., weight loss) may not fully apply to others (e.g., Asians). fTotal score offers limited transparency for doctor-patient communication; high scores may unintentionally bias towards conservative management. Comparison of the applicability and efficacy of classic frailty assessment tools. Abbreviations: CFS, Clinical Frailty Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; EFS, Edmonton Frail Scale; RAI, Risk Analysis Index; CGA, Comprehensive Geriatric Assessment; ICU, Intensive Care Unit; PND, Postoperative Neurocognitive Disorders; LOI, loss of independence; Preop, preoperative. aRapid Screening Tools. Ultra-fast (< 5 min), ideal for triaging patients in busy settings like the emergency department, primary care, or community screening. bComprehensive Assessment Tools. Multidimensional, thorough but more time-consuming. Provides a detailed frailty profile and specific deficit areas for developing intervention plans. cSpecialty/Research-Oriented Tools.Targets specific outcomes (e.g., surgery) or relies on specific data systems. High predictive value in specific domains. dThe optimal cutoff for predicting different outcomes (short-/long-term mortality) may vary with clinical context. eCriteria based on Western populations (e.g., weight loss) may not fully apply to others (e.g., Asians). fTotal score offers limited transparency for doctor-patient communication; high scores may unintentionally bias towards conservative management. Emerging assessment tools The proven utility of classic tools is tempered by their practical limitations, spurring the advancement of new technologies. These emerging tools provide more automated and objective assessments, capturing previously unmeasured facets of frailty such as specific biomarkers, muscle mass, and real-time physiological data. Imaging technologies Sarcopenia, characterized by decline in muscle mass, strength, and physical performance, is a critical component of physical frailty.120 It accelerates frailty progression by reducing functions like gait speed, while frailty exacerbates muscle loss. Imaging technologies enable objective assessment, overcoming the limitations of conventional methods in detecting early muscle loss and reducing subjectivity. Musculoskeletal B-mode ultrasonography Musculoskeletal B-mode ultrasonography objectively quantifies muscle mass and quality by measuring thickness, cross-sectional area, and echo intensity.121 This radiation-free, portable technique facilitates bedside use and dynamic monitoring of muscle changes, effectively overcoming the limitations of traditional tools in providing continuous objective biomarkers. Specifically, rectus femoris ultrasound can noninvasively analyze echo intensity and fiber structure to predict adverse outcomes like falls, functional decline, and mortality in older adults.122 In a retrospective study by Mirón-Mombiela et al., 43 textural features were extracted from the rectus femoris and vastus intermedius to develop predictive models.123 These models identified 70%–87% of frailty cases, correlated with increased postoperative comorbidities (0.01 ≤ p ≤ 0.18), and predicted mortality risk in prefrail and frail patients (0.001 ≤ p ≤ 0.016). Despite these strengths, current practice tends to integrate ultrasound parameters into broader frailty indices or use them to provide objective support for the physical frailty phenotype. Thus, rather than replacing classic frailty assessments, ultrasonography complements them—offering structural insight where traditional tools focus on functional evaluation. Computed tomography (CT) -based analysis Beyond ultrasonography, CT-based analysis enables precise muscle volume quantification to predict postoperative outcomes. This was demonstrated by Guo et al. in a prospective cohort study.124 They assessed sarcopenia through CT-derived skeletal muscle index, grip strength, and 6-meter walk speed. It found that both sarcopenia and frailty—whether evaluated individually or jointly—were independently associated with the risk of 90-day postoperative complications in elderly patients (sarcopenia with frailty: OR = 2.21, 95% CI 1.26–3.89; sarcopenia alone: OR = 1.84, 95% CI 1.01–3.36; frailty alone: OR = 1.77, 95% CI 1.03–3.03). The psoas major muscle serves as a biomarker of systemic protein reserves. It also shows strong correlation with surgical outcomes when standardized by height (Psoas Muscle Index, PMI).125 A retrospective cohort by Batista et al.126 confirmed that PMI-defined sarcopenia independently predicted 2-year mortality (aOR = 1.79, 95% CI 1.03–3.10; p = 0.037), major complications (aOR = 1.69, 95% CI 1.04–2.74; p = 0.034), and complication severity in colorectal surgery patients measured at the L3 vertebral level. Moreover, a diagnostic accuracy study by Canales et al.127 established that ultrasound-measured quadriceps thickness (AUC = 0.80, 95%CI, 0.64–0.97) and CT-measured psoas area (AUC = 0.88, 95% CI, 0.76–1.00) had equivalent discriminatory power for frailty. Both methods accurately predicted unplanned discharge to skilled nursing facilities (AUC = 0.81, 95% CI, 0.61–1.00) and POD (AUC = 0.89, 95% CI, 0.77–1.00). CT-based analysis addresses the inability of traditional tools to accurately quantify body composition. Its value lies in retrospective or opportunistic screening, providing objective data for frailty assessment without additional scans. However, high cost, radiation exposure, and static nature limit its use as a first-line screening tool. Other technologies Beyond conventional imaging, several emerging technologies offer new approaches to frailty assessment. Among them, gait analysis and oral function assessments are closer to clinical translation, while voice analysis is a promising direction for future research. A strength of these methods is their ability to generate objective, quantifiable, and multidimensional physiological and behavioral data, thereby enabling the development of more precise and individualized assessment models. Gait analysis uses wearable sensors to measure walking speed, cadence, and gait variability.68 It objectively reflects mobility and fall risk, making it suitable for preoperative screening and postoperative monitoring. Oral function assessment focuses on measures like chewing efficiency and swallowing ability,128 which are closely tied to nutritional status and physiological reserve. It has been included in some pre-surgery optimization protocols to identify risks related to nutrition and swallowing. Additionally, vocal biomarkers—particularly in pitch variation, speech rate, and intensity in older adults—may correlate with frailty.129 This method analyzes acoustic features in speech to assess frailty, offering a non-invasive, low-cost option for potential large-scale remote screening.130 The parallel development of multiple digital assessment tools like gait, oral, and voice analysis presents both a key challenge and an opportunity to effectively synthesize these heterogeneous data streams to form a more comprehensive and dynamic understanding of frailty. In this context, artificial intelligence (AI) plays a key role in data integration and pattern discovery. The purpose of AI is not to replace established assessment tools, but to extract meaningful information from multi-source data and explore its potential associations with frailty and surgical risk. AI models can not only provide an integrated assessment of overall risk but also generate quantifiable probability estimates for specific postoperative complications and adverse outcomes. This enables more targeted and proactive clinical interventions, thereby supporting more precise decision-making. In summary, the value of emerging tools like gait, oral, and voice analysis lies in their provision of rich data sources. The future focus is on integrating these sources through AI-driven analytics to build more robust risk assessment models. However, broader implementation still faces challenges such as data standardization, model interpretability, and integration into clinical workflows. These technologies are summarized in Table 4.Table 4Comparison of emerging technologies for perioperative frailty assessment.TechnologiesMaturity LevelKey AdvantagesCost-EffectivenessMajor Feasibility BarriersMusculoskeletal B-mode ultrasonographyTranslational(1) Bedside, dynamic,(2) radiation-free muscle quality assessmentModerate(1) Lack of standardized protocols(2) operator dependencyCT-based analysisEstablished(only Research)(1) Highly precise body composition quantificationHigh(1) Radiation exposure(2) not for primary screening(3) requires specialized softwareGait analysisInvestigational(1) Objective(2) functional measure of physiological reserveUncertain(1) Lack of standardizedprotocols unvalidated for key outcomesOral function assessmentEarly research(1) Potential to capture physiological declineUncertain(1) Incremental value over existing tools not yet provenVocal analysisEarly research(1) Non-invasive,(2) highly scalable potential for remote monitoringUncertain(1) Specialized analysis resources required(2) early validation stageAbbreviations: CT, Computed tomography. Table 4Comparison of emerging technologies for perioperative frailty assessment.TechnologiesMaturity LevelKey AdvantagesCost-EffectivenessMajor Feasibility BarriersMusculoskeletal B-mode ultrasonographyTranslational(1) Bedside, dynamic,(2) radiation-free muscle quality assessmentModerate(1) Lack of standardized protocols(2) operator dependencyCT-based analysisEstablished(only Research)(1) Highly precise body composition quantificationHigh(1) Radiation exposure(2) not for primary screening(3) requires specialized softwareGait analysisInvestigational(1) Objective(2) functional measure of physiological reserveUncertain(1) Lack of standardizedprotocols unvalidated for key outcomesOral function assessmentEarly research(1) Potential to capture physiological declineUncertain(1) Incremental value over existing tools not yet provenVocal analysisEarly research(1) Non-invasive,(2) highly scalable potential for remote monitoringUncertain(1) Specialized analysis resources required(2) early validation stageAbbreviations: CT, Computed tomography. Comparison of emerging technologies for perioperative frailty assessment. Abbreviations: CT, Computed tomography. Conclusions This review summarizes the perioperative value of frailty assessment tools. Anesthesiologists ought to integrate frailty assessment into management pathways for more precise care (Fig. 2).Fig. 2Integration of frailty assessment into the perioperative anesthesia management pathway. This flowchart delineates the clinical pathway for systematically integrating frailty assessment across the preoperative, intraoperative, and postoperative phases. Based on the preoperative assessment results, a personalized in-depth evaluation and interventions are triggered for frail patients. Integration of frailty assessment into the perioperative anesthesia management pathway. This flowchart delineates the clinical pathway for systematically integrating frailty assessment across the preoperative, intraoperative, and postoperative phases. Based on the preoperative assessment results, a personalized in-depth evaluation and interventions are triggered for frail patients. As no universal assessment tool exists that is suitable for all clinical scenarios, we recommend selecting the most appropriate instrument based on specific contextual factors such as surgical type, urgency, and available resources. A clinical decision pathway (Fig. 3) has been developed to guide this selection process. Future efforts should focus on understanding tool features, addressing inconsistent results and implementation gaps. Building on this, we must advance the targeted integration of frailty assessment into perioperative decision-making to improve patient outcomes.Fig. 3Clinical decision pathway for frailty assessment in surgical patients. This algorithm guides the selection of frailty assessment tools based on surgical urgency (emergency vs. elective) and available clinical resources. The goal is to identify the most appropriate tool for efficient risk stratification and to guide subsequent management decisions. Abbreviations: CFS, Clinical Frailty Scale; EFS, Edmonton Frail Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; MDT, Multidisciplinary Team; RAI, Risk Analysis Index. Clinical decision pathway for frailty assessment in surgical patients. This algorithm guides the selection of frailty assessment tools based on surgical urgency (emergency vs. elective) and available clinical resources. The goal is to identify the most appropriate tool for efficient risk stratification and to guide subsequent management decisions. Abbreviations: CFS, Clinical Frailty Scale; EFS, Edmonton Frail Scale; FFP, Fried's Frailty Phenotype; FI, Frailty Index; FRAIL, Fatigue, Resistance, Ambulation, Illnesses, Weight Loss; MDT, Multidisciplinary Team; RAI, Risk Analysis Index. CRediT authorship contribution statement Shanshan Shi: Writing – review & editing, Investigation. Xinyue Wu: Writing – review & editing, Investigation. Manling Ju: Writing – review & editing, Investigation, Conceptualization. Haozhou Shen: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Mi Tian: Writing – review & editing, Supervision, Funding acquisition. Jie Sun: Supervision, Resources. All authors have read and agreed to the published version of the manuscript. Consent for publication Not applicable. Ethical statement Not applicable. Funding This research was funded by 10.13039/501100001809National Natural Science Foundation of China (No. 82202392) and Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction Funds (No. zdlyg14) from Mi Tian. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.