Feasibility, acceptability and prognostic value of muscle mass and strength measurement in patients with hip fracture: a systematic review Authors: James Prowse, Sharlene Jaiswal, Jack Gentle, Antony K. Sorial, Miles D. Witham Published: 2024 ABSTRACT Key summary pointsAimWe systematically reviewed whether muscle mass measurement was independently associated with adverse outcomes in patients with hip fracture. Our secondary aim was to systematically review whether muscle mass assessment was feasible and acceptable in this cohort.FindingsThere are no studies specifically assessing feasibility and acceptability of muscle mass assessment in patients with hip fracture. Low muscle mass is associated with increased mortality, worse mobility and physical performance in unadjusted analyses.MessageIncluded studies suggest muscle mass offers no additional prognostic benefit to muscle strength when assessing sarcopenia in acute hip fracture.Supplementary InformationThe online version contains supplementary material available at 10.1007/s41999-024-01102-x. Introduction Hip fractures are common, costly and life-threatening injuries both in the United Kingdom (UK) and worldwide [1]. Over the last 2 decades, national standardization of hip fracture management has improved quality of care and patient outcomes in the UK [2, 3]. Predicting post-operative outcomes is important when discussing prognosis with patients and family, to enable evidence-based discussions about care [4]. Current validated risk prediction tools in hip fracture include the Nottingham Hip Fracture Score (NHFS) and Almelo Hip Fracture score [5, 6]. Existing scores have moderate discriminative ability for in-hospital, 30- and 120-day mortality and functional outcomes but have only weak predictive ability for outcomes, such as length of stay and postoperative complications [7]. Current hip fracture risk prediction scores do not consider muscle strength and mass, yet sarcopenia (the age-related loss of muscle strength and mass) is known to be a key driver of falls, dependency, and earlier death, and an important component of the physical frailty syndrome [8–10]. Current definitions of sarcopenia require confirmation of low muscle strength, with low muscle mass also being a component of most definitions [11–13]. Measuring muscle strength in clinical practice (for example handgrip strength) is straightforward, inexpensive, and can easily be incorporated into the preoperative assessment of hip fracture [14]. Measuring muscle mass is more complex; simple techniques such as bioelectrical impendence assessment (BIA) are inaccurate and require equations calibrated to specific populations and devices; techniques such as dual X-ray absorptiometry (DXA), cross-sectional imaging via computed tomography (CT), or magnetic resonance imaging (MRI) are more expensive and less accessible. Before muscle mass measurement is considered for incorporation into routine clinical risk assessment for patients with hip fracture, it is important to determine whether these measures are feasible and acceptable for delivery at scale in routine clinical practice, and if they add valuable prognostic value to existing risk scores with or without muscle strength measures. The aim of this study was therefore to systematically review the available literature to determine whether: (a) muscle mass assessment measures are associated with adverse postoperative outcomes in patients with hip fracture, (b) measuring muscle mass provides additional prognostic value to existing scores that incorporate muscle strength measures in patients with hip fracture and (c) current tools used to assess muscle mass are feasible and acceptable for patients with hip fracture. Methods Search strategy and selection criteria A prespecified systematic review protocol was developed prior to commencing the work, and this was published on the PROSPERO online database in July 2021 (Registration number: CRD42021274981). A literature search was performed using the CENTRAL, Ovid MEDLINE, EMBASE, CINAHL and clinicaltrials.gov databases. Citation searching of studies was also undertaken to identify additional studies not detected in database searching. Supplementary Table 1 includes an example search strategy used for Ovid MEDLINE. Searches were conducted from database inception through to the end of November 2023. Studies were deemed eligible for inclusion if (a) a full-text paper was available; (b) the study population was adult hip fracture patients with a mean age of 60 years or above who were hospitalized for surgery; (c) studies included an assessment of muscle mass or strength in hip fracture patients before surgery or during acute hospital admission. Papers were excluded if inclusion criteria were not met. Data extraction 3 authors (J.P., S.J. or J.G.) independently reviewed titles and abstracts to determine eligibility for inclusion. Disagreement was discussed between reviewers to determine eligibility. 1 reviewer (J.P.) reviewed all full-text studies with a second reviewer (S.J. or J.G.) independently assessing a random sample of 10% of full-text studies. Data were extracted into a standardized template that included: mean age, gender, location, healthcare setting, reasons for patient exclusion, tool and units used to assess muscle mass/strength, timing of assessment, postoperative outcomes assessed, statistical analysis and adjustments made. For assessment of feasibility and postoperative outcomes, we aimed to compare either the same muscle mass assessment tool as used for the intervention or the currently accepted ‘gold standard’ modalities of CT or MRI. If CT or MRI was unavailable, DXA was employed as an alternative comparison. Prognostic value of muscle mass was assessed by statistical analyses examining the independent contribution of muscle mass, strength and prognostic scores. The full list of tools, muscle mass estimates, and characteristics extracted is summarized in Supplementary Tables 2 and 3. Data analysis Reviewers independently assessed the identified studies for risk of bias, with the second reviewer (S.J. or J.G.) assessing ~ 10% to mitigate errors. Non-randomized studies were assessed using the Cochrane Risk of Bias in Non-randomized studies of Interventions (ROBINS-I) tool [15]. As the review assessed diagnostic testing, domain 2 (bias in selection of participants) was amended accordingly to determine whether patient selection reflected the wider hip fracture population. Studies identified at critical risk of bias were excluded. Cross-sectional studies were assessed using the appraisal tool for Cross-Sectional Studies (AXIS) [16]. As only baseline data was taken from randomized controlled trials, it was decided that the AXIS tool would be appropriate. Assessment of feasibility and acceptability was determined using the proportion of patients able to complete muscle mass assessment, excluding patients who declined assessment or without data collection. This measure proposed to capture the number of patients who could have had sarcopenia assessed through alternative methods, but due to organizational delay or clinical cause, cognitive impairment and presence of a pacemaker precluding BIA, were not included. Reported effect sizes for the association between muscle mass and a broad range of postoperative outcomes was recorded for both univariate and multivariable analysis. Meta-analysis was not attempted due to heterogeneity of the populations studied, outcome measures included, and tools used to assess muscle mass or strength. We quantified evidence of an effect using vote counting based on direction of effect in line with synthesis without meta-analysis (SWiM) guidance [17]. We described studies narratively according to their association with adverse outcomes. We described an association as ‘positive’ if lower muscle mass was associated with adverse outcome and ‘negative’ if lower muscle mass was associated with favorable outcome. Positive outcomes in initial univariate analysis were noted, but if subsequent analyses adjusting for covariates found no significance, we deemed the overall outcome as non-significant. Results Study selection The search identified 3,317 titles, with 182 studies eligible for full-text review. A further 8 studies were included from citation searching during full-text review. A total of 190 full-text studies were reviewed, with 37 studies included in the final analysis. The full selection process and reasons for exclusion are summarized in Fig. 1.Fig. 1PRISMA flow diagram for selection of studies. Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram for selection of studies assessing muscle mass, strength feasibility, acceptability, and postoperative outcomes in hip fracture patients PRISMA flow diagram for selection of studies. Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram for selection of studies assessing muscle mass, strength feasibility, acceptability, and postoperative outcomes in hip fracture patients Study characteristics The 37 studies reviewed had a variety of designs. 12 were cross-sectional, 24 longitudinal in design (5 case–control and 19 cohort studies) and there was 1 randomized controlled trial with baseline data. Publication dates ranged from 1984 to 2023. Patients were recruited from acute trauma wards (29 studies) and rehabilitation hospitals (8 studies). Study sizes ranged from 40 to 1003 participants, with a total of 7860 participants with a mean age range of75 to 86. Full study details are provided in Supplementary Table 4. Risk of bias Of the 25 studies assessed using ROBINS-I, 7 were deemed to be at moderate risk of bias, 16 at serious risk and 2 at critical risk and therefore excluded. Studies at critical risk had inadequate reporting of participant information and missing data [18, 19]. Using AXIS, sample size was not justified in 8/12 there were concerns about non-response bias in 8/12 studies. Full results are displayed in Supplementary Figs. 1 and 2. Associations between muscle mass and hip fracture postoperative outcomes 16 studies were included, identifying 31 associations between muscle mass and hip fracture postoperative outcome. There were 7 ‘positive’ associations between lower muscle mass and worse outcomes, and 24 non-significant associations. Findings are summarized in Table 1 and Supplementary Table 5 gives details of each association by assessment tool.Table 1Summary of direction of association between muscle mass assessment and hip fracture postoperative outcomesDirection of effect with decreased muscle massPostoperative OutcomeTotalGreater MortalityLower Barthel IndexWorse MobilityIncreased length of stayMore perioperative complicationsWorse physical performance measuresPositive association3120017No statistical significance93135324Negative association0000000A positive association was determined if lower muscle mass was associated with adverse outcome Table 1Summary of direction of association between muscle mass assessment and hip fracture postoperative outcomesDirection of effect with decreased muscle massPostoperative OutcomeTotalGreater MortalityLower Barthel IndexWorse MobilityIncreased length of stayMore perioperative complicationsWorse physical performance measuresPositive association3120017No statistical significance93135324Negative association0000000A positive association was determined if lower muscle mass was associated with adverse outcome Summary of direction of association between muscle mass assessment and hip fracture postoperative outcomes A positive association was determined if lower muscle mass was associated with adverse outcome Mortality 7 cohort studies (n = 2663 total patients) contained 17 analyses that assessed the association between muscle mass and mortality [20–26]. 3 analyses showed a significant association between lower muscle mass and higher mortality; 5 unadjusted analyses that showed significant associations became non-significant when analyses were adjusted for other prognostic factors, and 9 analyses did not show significant associations. Full data are reported in Table 2.Table 2Summary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patientsMuscle mass assessment toolStudyMuscle mass assessmentMortality outcomeAnalysis performedStrength of association with lower muscle massDirection of associationDecreased muscle mass by DXAIida, 2021 [[20]]Low SMI(kg/m2)1-year mortalityMultivariable Coxproportional hazardsHR = 3.2 (1.1–9.3)p = 0.033PositiveIn-hospital mortalityStudent’s t-testp = 0.17Not significantKim, 2022 [21]Low ASM (kg/m2)1-, 2-, 5-year mortalityChi-squared tests“Associated”p = not providedPositiveMortality(Unclear timepoint)Cox proportional hazardsWomen HR = 1.3 (0.6–3.1)Men HR = 4.3 (0.5–36.1)Not significantDecreased muscle mass by BIAMenéndez-Colino, 2018 [24]SMI(kg/m2)1-year mortalityBivariate Coxproportional hazards, subsequent multivariable analysisCox: “Associated” HR = Not provided, p = 0.005MV analysis: “Dropped out”Not significantLow SMI(kg/m2)1-year mortalityBivariate CoxProportional hazardsHR = Not provided p = 0.23Not significantMalafarina,2019 [23]Low SMI(kg/m2)7-year mortalityCox proportional hazardsHR = 1.2 (1.0–1.5) p = 0.08Not significantSanchez-Torralvo, 2023 [51]Fat free muscle index(kg/m2)3-, 6-, 12-month mortalityChi-squared test3 m p = 0.536 m p = 0.0612 m p = 0.06Not significantDecreased muscle mass by TSFMeyer, 2000[25]Triceps skinfold thickness3.5-year mortalityRelative risk comparison to controlLowest quarter: RR = 1.7 (1.1–2.7),Second quarter: RR = 1.9 (1.1–3.0)Third quarter: RR = 1.3 (0.7–2.5)Highest quarter: RR = 1.2 (0.5–3.1)PositiveDecreased muscle mass by CTKim, 2018 [22]SMI (L3)1-year mortalityKaplan–Meier analysis and log-rank testMortality rate (MR) Sarcopenic vs non-sarcopenic 22.2 vs 19.6, p = 0.79Not significantSMI (L3)5-year mortalityKaplan–Meier analysis and log-rank testMortality rate (MR) Sarcopenic vs non-sarcopenic82.7 vs 52.7, p = 0.03PositiveSMI (L3)5-year mortalityMultivariable Cox analysis using backward selection modelHR: 2.1 (1.1–4.0)p = 0.02PositiveSummary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patients. A positive association was determined if lower muscle mass was associated with adverse outcome at most complex statistical analysisDXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASM Appendicular skeletal muscle mass, HGS Hand grip strength, BMI Body mass index, SONSQ The short portable mental status questionnaire, BMD Bone mineral density, ASA American society of anesthesiologists, L3:Third lumbar vertebra HR Hazard ratio, MR Mortality rate Table 2Summary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patientsMuscle mass assessment toolStudyMuscle mass assessmentMortality outcomeAnalysis performedStrength of association with lower muscle massDirection of associationDecreased muscle mass by DXAIida, 2021 [[20]]Low SMI(kg/m2)1-year mortalityMultivariable Coxproportional hazardsHR = 3.2 (1.1–9.3)p = 0.033PositiveIn-hospital mortalityStudent’s t-testp = 0.17Not significantKim, 2022 [21]Low ASM (kg/m2)1-, 2-, 5-year mortalityChi-squared tests“Associated”p = not providedPositiveMortality(Unclear timepoint)Cox proportional hazardsWomen HR = 1.3 (0.6–3.1)Men HR = 4.3 (0.5–36.1)Not significantDecreased muscle mass by BIAMenéndez-Colino, 2018 [24]SMI(kg/m2)1-year mortalityBivariate Coxproportional hazards, subsequent multivariable analysisCox: “Associated” HR = Not provided, p = 0.005MV analysis: “Dropped out”Not significantLow SMI(kg/m2)1-year mortalityBivariate CoxProportional hazardsHR = Not provided p = 0.23Not significantMalafarina,2019 [23]Low SMI(kg/m2)7-year mortalityCox proportional hazardsHR = 1.2 (1.0–1.5) p = 0.08Not significantSanchez-Torralvo, 2023 [51]Fat free muscle index(kg/m2)3-, 6-, 12-month mortalityChi-squared test3 m p = 0.536 m p = 0.0612 m p = 0.06Not significantDecreased muscle mass by TSFMeyer, 2000[25]Triceps skinfold thickness3.5-year mortalityRelative risk comparison to controlLowest quarter: RR = 1.7 (1.1–2.7),Second quarter: RR = 1.9 (1.1–3.0)Third quarter: RR = 1.3 (0.7–2.5)Highest quarter: RR = 1.2 (0.5–3.1)PositiveDecreased muscle mass by CTKim, 2018 [22]SMI (L3)1-year mortalityKaplan–Meier analysis and log-rank testMortality rate (MR) Sarcopenic vs non-sarcopenic 22.2 vs 19.6, p = 0.79Not significantSMI (L3)5-year mortalityKaplan–Meier analysis and log-rank testMortality rate (MR) Sarcopenic vs non-sarcopenic82.7 vs 52.7, p = 0.03PositiveSMI (L3)5-year mortalityMultivariable Cox analysis using backward selection modelHR: 2.1 (1.1–4.0)p = 0.02PositiveSummary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patients. A positive association was determined if lower muscle mass was associated with adverse outcome at most complex statistical analysisDXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASM Appendicular skeletal muscle mass, HGS Hand grip strength, BMI Body mass index, SONSQ The short portable mental status questionnaire, BMD Bone mineral density, ASA American society of anesthesiologists, L3:Third lumbar vertebra HR Hazard ratio, MR Mortality rate Summary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patients Low SMI (kg/m2) Multivariable Cox proportional hazards HR = 3.2 (1.1–9.3) p = 0.033 “Associated” p = not provided Mortality (Unclear timepoint) Women HR = 1.3 (0.6–3.1) Men HR = 4.3 (0.5–36.1) SMI (kg/m2) Bivariate Cox proportional hazards, subsequent multivariable analysis Cox: “Associated” HR = Not provided, p = 0.005 MV analysis: “Dropped out” Low SMI (kg/m2) Bivariate Cox Proportional hazards Low SMI (kg/m2) Fat free muscle index (kg/m2) 3 m p = 0.53 6 m p = 0.06 12 m p = 0.06 Lowest quarter: RR = 1.7 (1.1–2.7), Second quarter: RR = 1.9 (1.1–3.0) Third quarter: RR = 1.3 (0.7–2.5) Highest quarter: RR = 1.2 (0.5–3.1) Mortality rate (MR) Sarcopenic vs non-sarcopenic 82.7 vs 52.7, p = 0.03 HR: 2.1 (1.1–4.0) p = 0.02 Summary of associations between muscle mass and mortality outcomes alongside direction of association in hip fracture patients. A positive association was determined if lower muscle mass was associated with adverse outcome at most complex statistical analysis DXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASM Appendicular skeletal muscle mass, HGS Hand grip strength, BMI Body mass index, SONSQ The short portable mental status questionnaire, BMD Bone mineral density, ASA American society of anesthesiologists, L3:Third lumbar vertebra HR Hazard ratio, MR Mortality rate Activities of daily living 3 studies, (cohort, case–control and cross-sectional study, n = 785 total patients) assessed association using the Barthel Index [20, 27, 28]. 1 study (n = 337) found a significant association between low appendicular skeletal muscle index (ASMI) measured using DXA and lower (worse) discharge Barthel Index [20]. The 2 other studies (n = 448) performed 3 analyses which did not show an association between DXA-derived ASMI and Barthel Index on discharge after adjustment for other covariates [27, 28]. Mobility 3 cohort studies (n = 304 total patients) assessed association with mobility. 1 study (n = 71) reported a significant association between thigh cross-sectional area assessed by CT at 2 months and gait speed at 2 months [29]. This study also noted a significant association between limb cross-sectional area and post-operative improvement in gait speed. Another study (n = 143) reported no association between appendicular skeletal muscle index assessed by DXA and improvement in the timed up and go test in multiple linear regression [27]. The final study (n = 90) reported no association between muscle mass (kg) assessed by DXA and 12 month recovery in mobility function with scoring assessed by independence/assistance/inability to get in/out of bed, chair rise, walk 10 feet, walk 1 block and climb 5 stairs [30]. Length of stay in hospital 3 studies (1 case–control, 2 cohort, n = 484 total patients) assessed associations between muscle mess and length of hospital stay. They used appendicular skeletal muscle index measured by DXA, BIA or psoas muscle area at the third lumbar vertebra (L3) level on computed tomography [20, 31, 32]. No significant associations were found between these measures and length of stay. Postoperative complications 3 cohort studies (n = 448 total patients) assessed association between muscle mass and postoperative complications. 1 study (n = 57) assessed association between muscle mass using CT/DXA and Clavien–Dindo grading of surgical complications [33]. The ASMI assessed by DXA, total psoas volume at L3 and psoas area at L3 all showed no association with grade of complication [31]. Another study (n = 301) found no association between CT-reported gluteus maximum CSA and risk of a second hip fracture, with a final study finding no association between BIA-assessed ASMI and any postoperative complication. [32, 34]. Physical performance measures 2 studies (cross-sectional and cohort study, n = 487 total patients), assessed association between muscle mass and physical performance. 1 study compared DXA-derived muscle mass and a model of factors influencing physical activity assessed using the Yale Physical Activity Survey [35]. Muscle mass was associated with physical activity in females but not in males. The other study found no associated between BIA-derived SMI and change in motor functional independence measure score at 4 weeks [36]. Prognostic value of muscle mass when used in addition to muscle strength 7 studies (6 cohort and 1 cross-sectional) measured both muscle mass and muscle strength. Studies are summarized in Supplementary Table 4. 3 studies performed analyses of muscle strength and muscle mass separately rather than combining these measures in a single model [23, 25, 30]. 4 studies reported a multivariable analysis including both muscle mass and strength (n = 1139 total patients), with outcomes summarized in Table 3 [24, 27, 35, 36]. For mortality, 1 study (n = 509) found that handgrip strength was an independent predictor of 1-year mortality, while muscle mass was non-significant in multivariable analysis [24]. The remaining 2 studies performed separate analysis, with only handgrip strength reaching significance for mortality at 7 years and associated with a greater relative risk at 3.5 years [23, 25]. 2 studies assessed the effect of muscle strength and mass on physical activity measures. 1 study (n = 339) found a mixed association for DXA-derived total lean mass and handgrip strength with 2-month physical activity using the Yale Physical Activity Survey. Muscle mass but not strength was independently associated with physical activity for women; the converse was found for men [35]. The other study (n = 148) assessed the relationship between BIA-derived SMI, handgrip strength and change in motor functional independence measure (FIM) score at 4 weeks [36]. Higher handgrip was independently associated with improved FIM score, while SMI was not.Table 3Summary of studies containing assessment of both muscle mass and strength in multivariable analysisStudyCountryStudy typeAssessment ToolAnalysis performedVariables included in analysisOutcome under testWas assessment independently associatedMuscle massMuscle strengthDi Monaco, 2014 [27]ItalyProspective cohortDXA (ASMI),HGS (kg)Multiple linear regressionHGS, ASMI, BI on admission, age, # medications, pressure ulcers, # concomitant diseases, infection during stay, days between fracture and assessment included in analysisBI post-rehabilitationBI effectivenessTUG test performanceNoNoNoYesYesYesMenéndez-Colino, 2018 [24]SpainProspective cohortBIA (SMI)HGS (kg)Bivariate Cox regression analysis, subsequent multivariable analysisLow HGS, Sex, Age > 85, Baseline BI ≤ 60, SPMSQ > 3, BMI < 21 kg/m2, Heart disease, Vitamin D < 20 ng/ml and PTH ≥ 66 pg/ml, Hb < 12 g/L F < 13 g/L M1-year mortalityNoYesResnick, 2018 [35] USACross-sectionalDXA (TLM kg)HGS (kg)Structural equation modelling. Fit tested with Chi2Model included bone mineral density, TLM, HGS, total body fat, age, cognition, comorbidity, resilience scale, pain2-month Physical activity (Yale Physical Activity Survey)Men noWomen yesMen yesWomen noIrisawa, 2022 [36]JapanProspective cohortBIA(SMI kg/m2)HGS (kg)Multivariable logistic regressionGeriatric nutritional risk index, SMI, HGS, Phase angle4-week change in Functional Independence measure motor scoreNoYesDXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASMI Appendicular skeletal muscle mass index, TLM Total lean mass, TUG Timed up and go test, BI Barthel index, Hb Hemoglobin, SPSMQ Short portable mental status questionnaire Table 3Summary of studies containing assessment of both muscle mass and strength in multivariable analysisStudyCountryStudy typeAssessment ToolAnalysis performedVariables included in analysisOutcome under testWas assessment independently associatedMuscle massMuscle strengthDi Monaco, 2014 [27]ItalyProspective cohortDXA (ASMI),HGS (kg)Multiple linear regressionHGS, ASMI, BI on admission, age, # medications, pressure ulcers, # concomitant diseases, infection during stay, days between fracture and assessment included in analysisBI post-rehabilitationBI effectivenessTUG test performanceNoNoNoYesYesYesMenéndez-Colino, 2018 [24]SpainProspective cohortBIA (SMI)HGS (kg)Bivariate Cox regression analysis, subsequent multivariable analysisLow HGS, Sex, Age > 85, Baseline BI ≤ 60, SPMSQ > 3, BMI < 21 kg/m2, Heart disease, Vitamin D < 20 ng/ml and PTH ≥ 66 pg/ml, Hb < 12 g/L F < 13 g/L M1-year mortalityNoYesResnick, 2018 [35] USACross-sectionalDXA (TLM kg)HGS (kg)Structural equation modelling. Fit tested with Chi2Model included bone mineral density, TLM, HGS, total body fat, age, cognition, comorbidity, resilience scale, pain2-month Physical activity (Yale Physical Activity Survey)Men noWomen yesMen yesWomen noIrisawa, 2022 [36]JapanProspective cohortBIA(SMI kg/m2)HGS (kg)Multivariable logistic regressionGeriatric nutritional risk index, SMI, HGS, Phase angle4-week change in Functional Independence measure motor scoreNoYesDXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASMI Appendicular skeletal muscle mass index, TLM Total lean mass, TUG Timed up and go test, BI Barthel index, Hb Hemoglobin, SPSMQ Short portable mental status questionnaire Summary of studies containing assessment of both muscle mass and strength in multivariable analysis DXA (ASMI), HGS (kg) BI post-rehabilitation BI effectiveness TUG test performance No No No Yes Yes Yes BIA (SMI) HGS (kg) DXA (TLM kg) HGS (kg) Men no Women yes Men yes Women no BIA (SMI kg/m2) HGS (kg) DXA Dual-energy X-ray absorptiometry, TSF Triceps skinfold thickness, BIA Bioelectrical impedance analysis, SMI Skeletal muscle index, ASMI Appendicular skeletal muscle mass index, TLM Total lean mass, TUG Timed up and go test, BI Barthel index, Hb Hemoglobin, SPSMQ Short portable mental status questionnaire 1 study (n = 143) found that ASMI assessed by DXA was determined to be non-significant while handgrip strength was associated with improved change in timed up and go (TUG) test performance [27]. 1 study assessed recovery in mobility by interview using separate analyses[30]. Mobility recovery was not associated with change in muscle mass by DXA, while loss of handgrip strength was associated with worse mobility recovery. 1 study (n = 143) assessed Barthel Index after rehabilitation and Barthel ‘effectiveness’ as proportion of potential improvement including both mass and strength [27]. For both analyses, ASMI determined by DXA was non-significant, while handgrip strength was significantly associated with improved outcomes. No studies assessed the association between both muscle mass and strength with increased length of stay or postoperative complications. Feasibility and acceptability Few studies formally assessed feasibility at the time of patient hospitalization; 6 studies did not include any information detailing the reasons for patient exclusion from assessment. Table 4 summarizes the 31 studies assessing muscle mass, the tool used, and the proportion of patients assessed successfully. No papers included data on economic, patient or staff acceptability. Supplementary Table 6 summarizes the proportional completion and reason for patient exclusion for each study.Table 4Summary of studies assessing muscle mass in older hip fracture patients with sufficient data to assess feasibilityToolSufficient data to assess feasibilityPercentage of patients assessed successfullyDual-energy X-ray absorptiometry1468–92%Triceps skinfold thickness361–94%Computed tomography550–93%Bioelectrical impedance analysis950–95%Patients were assessed as successful by number of patients able to have muscle mass assessed within total number of patients who did not have assessment due to organizational factors or clinical reasons. We excluded patients who refused assessment or without data collection. Percentage range shows lowest-highest % of patients able to complete muscle mass assessment for each assessment tool for studies included in review Table 4Summary of studies assessing muscle mass in older hip fracture patients with sufficient data to assess feasibilityToolSufficient data to assess feasibilityPercentage of patients assessed successfullyDual-energy X-ray absorptiometry1468–92%Triceps skinfold thickness361–94%Computed tomography550–93%Bioelectrical impedance analysis950–95%Patients were assessed as successful by number of patients able to have muscle mass assessed within total number of patients who did not have assessment due to organizational factors or clinical reasons. We excluded patients who refused assessment or without data collection. Percentage range shows lowest-highest % of patients able to complete muscle mass assessment for each assessment tool for studies included in review Summary of studies assessing muscle mass in older hip fracture patients with sufficient data to assess feasibility Patients were assessed as successful by number of patients able to have muscle mass assessed within total number of patients who did not have assessment due to organizational factors or clinical reasons. We excluded patients who refused assessment or without data collection. Percentage range shows lowest-highest % of patients able to complete muscle mass assessment for each assessment tool for studies included in review Discussion To our knowledge, this is the first systematic review comparing muscle mass and strength assessment in hip fracture. Our results suggest that muscle mass measurement appears feasible (using BIA, DXA and CT), low muscle mass is associated with adverse outcomes in some, but not all studies, and that muscle mass assessment may add little additional prognostic value to assessment of muscle strength. We found that the overall quality of study design and reporting was low, with few studies specifically designed for diagnostic evaluation. Our review suggests that lower muscle mass is associated with worse postoperative patient outcomes, with more evidence for mortality than for other outcomes. A previous systematic review by Liu et al. reported sarcopenia as a predictor of all-cause mortality in community-dwelling older people, including either muscle mass or strength assessment [9]. A review by Chiang et al. noted an association between sarcopenia and poorer postoperative functional recovery in hip fracture and a further study by Willey et al. noted a significant association between serial BIA analysis of skeletal muscle mass and Patient-Reported Outcomes Measurement Information System (PROMIS) physical function domains[32, 37]. We also found evidence that lower muscle mass was also associated with worse activity of daily living metrics, mobility assessment and physical performance measures however, there were conflicting results from studies included. This finding aligns with previous literature on sarcopenia and activities of daily living [38]. We found no associations between low muscle mass and increased length of stay or postoperative complication scores. This is surprising, as in healthy older adults sarcopenia is associated with poor mobility and adverse postoperative outcomes in surgical patients [39, 40]. Other drivers (e.g. susceptibility to infection, organizational factors) may be more directly linked and thus more strongly associated with length of stay than muscle mass. In many of the included studies, there was a lack of adjustment for well-recognized confounders such as cognitive status. The observational studies were mostly deemed to have moderate or serious risk of bias in the domains of confounding, primarily due to the exclusion of medically unwell patients or patients with diagnoses such as dementia. Future research should ideally have a standardized design and produce outputs adjusted for known covariates. Perhaps most importantly for clinical practice, muscle mass did not appear to add additional prognostic value when muscle strength was also measured. All 4 studies directly comparing muscle strength and mass found muscle strength to be an independent predictor of adverse outcomes, with muscle mass failing to reach statistical significance in 3 studies when included in models already containing muscle strength. These finding support the EWGSOP2 focus on assessment of muscle strength prior to muscle mass in the diagnostic algorithm of sarcopenia. We hypothesize this is because current measures of muscle mass measure lean mass, which is a mixture of muscle and other tissues. Cross-sectional imaging techniques that purport to measure muscle mass may also contain other intramuscular tissues such as fat. It is also possible that muscle strength is influenced by a broader range of factors than muscle mass, potentially including factors that reflect whole-body health (e.g. neurological function, cardiorespiratory function) and that muscle strength therefore provides a more informative measure of integrative whole-body function than muscle mass for prognostic purposes. Furthermore, the EWGSOP2 guidelines highlight numerous studies recognizing muscle strength as being superior to mass in predicting adverse outcomes [41–43]. Given the acceptability, speed, and cost-effective nature of handgrip strength assessment in routine clinical practice, the lack of prognostic value added by measuring muscle mass suggests that the time and effort required to assess muscle mass in patients with hip fracture may not add clinical value [14]. Routine sarcopenia screening in clinical practice may enhance prediction of patient outcomes following hip fracture. This information can be used to estimate and discuss prognosis with patients and to enable evidence-based discussions about inpatient care, as well as to facilitate discharge planning and quantify recovery potential. Most studies included in our review did not report feasibility in a transparent or consistent way, and no studies assessed acceptability. However, some comparable studies have been undertaken with similar patient cohorts. In 2019, Ibrahim et al. examined the feasibility of assessing sarcopenia using muscle strength and mass in acutely unwell hospitalized older people. They found assessment of handgrip strength to be feasible, but that assessment of muscle mass presented challenges [42]. A 2013 review by Mijnarends et al. determined that BIA was feasible and acceptable in community-dwelling older people, but that further evaluation was required, which aligns with our findings [44]. By comparison, implementation of handgrip strength assessment in hip fracture patients has been shown to be acceptable to staff and cost-effective to implement [14]. We postulate that due to the baseline characteristics of the hip fracture cohort, it can be more challenging to assess feasibility/acceptability. This is likely due to exclusion of patients with cognitive impairment, comorbidities and acute illness from such assessments. This may be important to consider when trying to implement muscle mass and strength assessment in routine clinical practice. Almost all the studies included in this review assessed muscle mass in the days after admission to hospital for fracture. While assessment of muscle mass in the acute setting may be challenging, hip fractures have been shown to have a substantial negative effect on muscle mass and strength measured after surgery [45]. The lack of studies assessing muscle mass using MRI or ultrasound indicates further research using these tools is necessary to establish their value. It may therefore be more feasible to assess muscle mass by DXA, BIA or CT either during admission or as part of routine perioperative assessment. Sarcopenia assessment in routine practice is part of a suite of outcome prediction biomarkers alongside clinical review of the patient. Alternative methods of sarcopenia assessment such as muscle strength assessment using HGS may be cheaper and easier to deploy in the hip fracture cohort. Strengths of our review included a best-practice approach when planning, conducting and interpreting a systematic review. We used a prespecified protocol, with a clearly defined clinical question. There was a rigorous approach to assessment of bias using either the ROBINS-I or AXIS tool. A second reviewer was used to mitigate errors at the data-extraction stage. A number of limitations of our review methods also require highlighting. We identified 37 papers and report some key findings, however, our review included a limited evidence base, potentially narrow inclusion criteria and heterogeneity of analyses which precluded meta-analysis of postoperative outcome measures. The broad scope of our literature search may have included older studies, where data collected are no longer relevant due to evolution of clinical practice. It is also possible that the search may have missed some studies, although measures were taken to avoid this. The timing of the sarcopenia assessment may have affected the assessment of muscle mass in the included studies. Furthermore, there was a difference in diagnostic cut-off thresholds between studies. The tools we reviewed are susceptible to gender, ethnicity, and height variations in measurement, and are therefore not a perfect measure of whole-body muscle mass [12]. The studies included in the review used broad exclusion criteria, and therefore excluded patients who could not complete assessment, were acutely unwell, or cognitively impaired. Similarly, no measures were taken to adequately address and categorize non-responders, beyond a reason cited for exclusion from studies. It is recognized that patients with cognitive impairment, multimorbidity and acutely unwell patients are underserved in clinical research[46]. Consequently, it is likely that the included studies are at risk of selection bias and exclusion of the most unwell patients. Due to the lack of analysis about true/false positives or negatives of the included tests, it was not possible for us to produce a summary receiver operating characteristic curve for the feasibility of muscle mass assessment tools. Our review also highlighted a paucity of studies using measures of muscle mass to assess sarcopenia. EWGSOP2 recommends use of DXA, BIA, CT and MRI in research studies of sarcopenia[12]. Few studies used the modalities of CT, with none for MRI, which may be influenced by the cost, inconvenience, and high clinical demand for these modalities in the acute setting. While publication bias may contribute, the fact that 24 out of 31 of the associations found in this review had no statistical significance indicates this is less likely. The widespread use of CT/MRI imaging for undiagnosed hip fracture patients provides a window for opportunistic assessment of muscle mass for some patients [47]. Further research in a clinical setting is recommended. The small number of studies assessing both muscle mass and strength and their risk of bias (4 serious, 3 moderate) also needs to be considered. Overall, studies performed poorly in the domains of confounding and selection of participants. Analyses performed in individual studies did not control for possible covariates and selected healthier participants. The results may therefore not be representative of the overall hip fracture population. Emerging methods of estimating muscle mass An alternative method of estimating muscle mass, the D3-creatine (D3-Cr) dilution, has been shown to be a simple, clinically feasible risk predictor of hip fracture in healthy older men [48]. There are currently limited numbers of studies in a clinical setting, with a narrative review by Pagano et al. in 2024 identifying 15 studies correlating D3-Cr to clinical outcomes [49]. This study noted a positive correlation between D3-Cr and MRI/DXA assessment of muscle mass, with less consistent associations for measures of muscle strength including HGS. Barriers to widespread clinical use include isotope cost, requirements for laboratory analysis and logistics of isotope transport to patients. Furthermore, D3-Cr values may be affected by muscle breakdown in hip fracture patients [50]. Nevertheless, ongoing research and future assessment of sarcopenia may include D3Cr assessment. Conclusion Although limited by the quality and scope of available literature, our findings suggest that muscle mass measurement offers no additional prognostic benefit to muscle strength when assessing sarcopenia in patients with acute hip fracture. Additional high-quality studies of different muscle mass measurement techniques, specifically designed to evaluate the acceptability, feasibility and prognostic value of muscle mass measures added to muscle strength measures would be desirable. Given that muscle strength assessment is cheaper, easier, and more strongly associated with adverse outcomes, our findings suggest that at present, routine assessment of muscle mass in hip fracture patients has limited value in clinical practice. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 443 KB) Supplementary file1 (DOCX 443 KB)