Muscle parameters in fragility fracture risk prediction in older adults: A scoping review

Abstract Half of osteoporotic fractures occur in patients with normal/osteopenic bone density or at intermediate or low estimated risk. Muscle measures have been shown to contribute to fracture risk independently of bone mineral density. The objectives were to review the measurements of muscle health (muscle mass/quantity/quality, strength and function) and their association with incident fragility fractures and to summarize their use in clinical practice. This scoping review follows the PRISMA‐ScR guidelines for reporting. Our search strategy covered the three overreaching concepts of ‘fragility fractures’, ‘muscle health assessment’ and ‘risk’. We retrieved 14 745 references from Medline Ovid SP, EMBASE, Web of Science Core Collection and Google Scholar. We included original and prospective studies on community‐dwelling adults aged over 50 years that analysed an association between at least one muscle parameter and incident fragility fractures. We systematically extracted 17 items from each study, including methodology, general characteristics and results. Data were summarized in tables and graphically presented in adjusted forest plots. Sixty‐seven articles fulfilled the inclusion criteria. In total, we studied 60 muscle parameters or indexes and 322 fracture risk ratios over 2.8 million person‐years (MPY). The median (interquartile range) sample size was 1642 (921–5756), age 69.2 (63.5–73.6) years, follow‐up 10.0 (4.4–12.0) years and number of incident fragility fractures 166 (88–277). A lower muscle mass was positively/not/negatively associated with incident fragility fracture in 28 (2.0), 64 (2.5) and 10 (0.2 MPY) analyses. A lower muscle strength was positively/not/negatively associated with fractures in 53 (1.3), 57 (1.7 MPY) and 0 analyses. A lower muscle function was positively/not/negatively associated in 63 (1.9), 45 (1.0 MPY) and 0 analyses. An in‐depth analysis shows how each single muscle parameter was associated with each fragility fractures subtype. This review summarizes markers of muscle health and their association with fragility fractures. Measures of muscle strength and function appeared to perform better for fracture risk prediction. Of these, hand grip strength and gait speed are likely to be the most practical measures for inclusion in clinical practice, as in the evaluation of sarcopenia or in further fracture risk assessment scores. Measures of muscle mass did not appear to predict fragility fractures and might benefit from further research, on D3‐creatine dilution test, lean mass indexes and artificial intelligence methods.


Introduction
Osteoporosis is characterized by a generalized loss of bone mass and altered microarchitecture, leading to an increased risk of fracture. 1Over the age of 50, a fifth of men and half women will have a fragility (or osteoporotic) fracture, developed spontaneously or after a minor trauma, such as a fall from a standing height. 1 Major osteoporotic fractures (MOFs) include hip, vertebral, humeral and forearm fractures.Fragility fractures are a major age-related adverse event due to their consequences and high incidence. 2Osteoporotic fractures account for more days of hospitalization than acute myocardial infarction, chronic obstructive pulmonary disease or breast cancer. 3In Europe, the direct costs were estimated at 37.4 billion euros in 2010 and 56.9 billion euros in 2019 2 and will continue to increase as the population aged over 65 and over 80 is expected to double and triple respectively between 2020 and 2050. 4Bone fragility can be prevented and treated.However, the gap in its management consists in the limited capacities to detect and predict fragility fractures. 5he gold standard for assessing bone mineral density (BMD) is dual-energy X-ray absorptiometry (DXA).The World Health Organization (WHO) defines osteoporosis as a BMD of 2.5 standard deviations below the mean peak BMD of young female adults. 6However, half of fractures occurs in individuals with a normal BMD. 7Risk scores have thus been developed and have improved fracture prediction, by taking into consideration other clinical risk factors for fractures 8 ; the most widely used fracture risk score is FRAX® (Fracture Risk Assessment Tool). 8Although FRAX with BMD performs better than BMD alone in predicting incident fractures, there is still room for improvement in risk prediction, potentially through inclusion of additional measures, such as falls, that are independent of BMD. 9 Muscles lose 40% of their volume between the ages of 20 and 80. 10 Since the first mention of the muscles mass loss as sarcopenia by Rosenberg in 1989, 11 many parameters of muscle health have been studied using a variety of measures such as radiological imaging, strength measurements, functional assessments and blood tests.In parallel, the definition of sarcopenia has evolved to a composite loss of muscle mass, strength and function, and its association with adverse outcomes, including fragility fractures. 12arcopenia and osteoporosis are both associated with ageing and similar risk factors in a close interaction. 13][22] A scoping review is a structured approach to summarize and map the evidence and gaps on a topic.This type of knowledge synthesis is particularly useful for planning future research on heterogeneous and broad topics.So far, only one scoping review studied muscle health and its association with adverse outcomes. 23The authors focused on three defini-tions of sarcopenia and their ability to predict various adverse outcomes.Of the 11 included studies in this previous review, only one analysed fragility fractures. 24The currently available studies on muscle health parameters and their association with incident fragility fractures have not been fully reviewed.
The objectives of this scoping review were (1) to review muscle health assessment techniques (muscle mass/quantity/quality, strength and function) and their association with incident fragility fractures and (2) to summarize the clinical use of the parameters associated with fragility fractures risk.

Methodology
This scoping review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Review (PRISMA-ScR) guidelines for reporting and the JBI methodology for writing. 25,26The PRISMA-ScR checklist is provided in the supporting information.The study protocol is available online in the OSF (Open Science Framework) registry at https://archive.org/details/osf-registrations-2fmtg-v1(registration DOI: 10.17605/OSF.IO/2FMTG).

Inclusion criteria
The studies included in this review fulfilled the following criteria: (1) original study; (2) participants over 50 years of age recruited from the general population (communitydwelling) without gender, racial, geographic or cultural restriction.Studies where the participants were recruited on the basis of a medical condition (e.g., frailty, osteoporosis and cancer) were excluded to minimize selection bias; (3) assessment of at least one muscle health parameter; (4) prospective studies; (5) fragility fracture as outcome: a low-trauma fracture at any specific osteoporotic site or a combination of sites; and (6) the association of each muscle health parameter with the fragility fracture incidence was examined.No language restrictions were performed.Meta-analyses, systematic reviews and, text/opinion papers relevant to the current review's question were considered for the qualitative and critical evaluation and interpretation.

Source of evidence and search strategy
A systematic search strategy was developed with a research librarian to cover the three overarching concepts of the research: 'fragility fractures', 'muscle health assessment' and 'risk'.The search syntax contains free and index/mesh terms, a filter to exclude animal studies and a general filter for the study types.Relevant articles were also compared to better define the keywords and index terms of the equations.The search strategy was translated for the following databases: Medline Ovid SP, EMBASE and Web of Science Core Collection.A complementary search equation was developed for Google Scholar.Systematic search syntaxes are available in the supporting information.Unpublished studies and grey literature were not screened.Backward and forward citation chasing of eligible studies was also done.We also undertook hand searching of references within records and on specific authors to identify further eligible studies.The search included article published from inception of the databases to 27 April 2023.

Study selection
The identified citations from the systematic search were de-duplicated (J.E.) in EndNote ™ (Clarivate Analytics, Philadelphia, PA, USA) and transferred (C.V.) to Rayyan (free web application for systematic reviews 27 ).One author (C.V.) screened the titles and abstracts for eligibility and retrieved the full texts of the selected articles.The reasons for exclusion were recorded at full text reading.The study's selection process is fully reported using the PRISMA 2020 flow diagram (cf. Figure 1).

Data extraction and qualitative assessment
The data were extracted from the included articles by one author (C.V.) using an Excel table.For each study, qualitative and quantitative data were extracted 25 : first author, year of publication, country, design, duration of follow-up, population, sex, mean age at baseline, sample size, muscle health parameter, fracture type, number of fractures, statistical approach, model adjustments and fracture risk estimates for the muscle parameters studied.When one association had multiple models, we kept the model considering the strongest predictor of fragility fractures including age and/or BMD.Multiple adapted forest plots were used to visually demonstrate the overall trends of associations between each muscle parameter and the fracture risk.The results were grouped by mass/quantity/quality (Figures 3-5), strength and function and by fracture type (A-F).The results were ordered by parameter, measure subtype, sex and publication date.The muscle mass mostly refers to lean mass (LM) (or its estimation) while quantity also includes volumes and areas.Muscle quality is a broad terminology and includes muscle density, muscle texture, myosteatosis, muscle fat infiltration and some ultrasound measures. 28In order to homogenize the reporting and to facilitate the interpretation of the results, we always reported the fracture risk ratios for a lower/slower/deteriorated muscle parameter (e.g., 'the risk ratio for 1 SD decrease in lean mass').Most of the original articles had reported the fracture risk ratio per unit of deterioration in the muscle parameter studied, and these values were reported identically; if the original article had reported the fracture risk ratios per increase in the muscle parameter studied, we calculated and reported the 1/risk ratio.The rationale is that a worsened/unhealthy muscle parameter is associated with a higher risk of fracture.Finally, the most frequently cited muscle health assessment parameters in the included articles are briefly discussed in terms of their generalizability and availability in clinical practice. 29Additionally, the best predictors of fragility fractures are reported, including the total person-year.

Characteristics of the included studies
Of the 13 745 studies extracted from the databases and the approximately 1000 studies screened using additional methods (Figure 1: PRISMA flow chart), 67 studies were included in this review, comprising 2.8 million person-years: median sample size (1st-3rd quartile) of 1642 (921-5756) participants, follow-up of 10.0 (4.4-12.0)years, age of 69.2 (63.5-73.6)years and number of incident fragility fractures of 166 (88-277).  The neral characteristics of the included studies are summarized in Table 1 and detailed for each article in Table 2.The most cited cohorts were MrOS (USA, China and Sweden; 13 articles), DOES (Australia; 6 articles), SOF (USA; 5 articles), Health ABC (USA; 4 articles) and EPIDOS (France; 4 articles).Within the studies, 37 analysed women, 30 men and 13 both together.All results and references are presented visually and summarized in multiple stacked plots (Figures 3-5).The 67 included studies investigated 60 different muscle parameters and were grouped into 6 types of fragility fracture: hip (Figure 2B

Muscle mass, quantity and quality
Evaluation of muscle mass and quantity has been performed by very different methods, from radiological images (i.e., DXA and computed tomography [CT]), biological measures (creatine dilution test) or even anthropometric prediction equations.Globally, a lower muscle mass or quantity was associated with risk of incident fragility fracture in 28 (2034 thousand person-years [TPY]) analyses, no risk in 66 (2633 TPY) analyses and lower risk in 10 (230 TPY) analyses (Figures 2, 3A-E and S3f).Body composition analysis by DXA was the most used method.Several DXA-derived muscle mass parameters were analysed: appendicular lean mass (ALM), change in ALM, ALM/height, ALM/height 2 , change in ALM/height 2 , ALM/weight, ALM/body mass index (BMI), total LM, change     39 The Heymsfield equation is based on the triceps skinfold thickness and midarm circumference. 96A lower muscle mass derived from these two equations was positively and not associated with fractures in four (1381 TPY) and one (395 TPY) analyses, respectively.Using the creatine and creatinine-derived parameters (D3-creatine dilution test and estimated glomerular filtration rate [eGFR]), a lower parameter was associated with a higher, no and a lower fracture risk in 4 (170 TPY), 12 (507 TPY) and 2 (88 TPY) analyses, respectively.

Muscle strength
Muscle strength was mostly assessed using the maximum isometric contraction of a specific muscle group.No analysis showed a negative association between muscle strength and fractures.A lower muscle strength was positively associated with incident fragility fractures in 53 (1.3 TPY) analyses and not associated in 57 (1.7 TPY) analyses.Hand grip strength (HGS) was associated with a higher and no fracture risk in 37 (1181 TPY) and 39 (1312 TPY) analyses, respectively.A lower triceps strength was associated with a higher and no fracture risk in two (29 TPY) and three (46 TPY) analyses, respectively.A lower quadriceps strength (QS) was associated with a higher and no fracture risk in 13 (131 TPY) and 15 (389 TPY) analyses, respectively.
One study also analysed a lower arm and leg strength together and found a positive association (2 TPY) with fractures.

Muscle function
Muscle function refers to tests that assess specific tasks, mobility and balance.As for muscle strength, none showed a negative association between muscle function's assessment    and fractures.A lower muscle function was positively associated with incident fragility fracture in 63 (1901 TPY) analyses, not associated in 45 (972 TPY) analyses and negatively associated in 0 analyses.Gait speed (GS) refers to the usual walking speed over a distance of 4-6 m.A slower GS or loss of GS over time was associated with a higher and no fracture risk in 32 (1121 TPY) and 17 (391 TPY) analyses, respectively; it was positively associated with MOF in all the eight concerned studies (333 TPY). 31,32,41,44,54,55,58The different walking and chair rising tests were associated with a higher and no fracture risk in 19 (572 TPY) and 12 (299 TPY) analyses, respectively.They included five assessments: timed get up and go test (TGUG), change in TGUG, five-time sit-to-stand test (5×STS), Δ 5×STS and squat/jump.Balance tests were associated with a higher and no fracture risk in 11 (184 TPY) and 10 (196 TPY) analyses, respectively.These included three different assessments: one-leg standing test (OLST), narrow/tandem walk and single-foot coordination.Multi-item tests were associated with a higher and no fracture risk in one (24 TPY) and six (86 TPY) analyses, including three assessments: Short Physical Performance Battery (SPPB) test, sarcopenia screening questionnaire (SARC-F) and a speed/reaction test.

Discussion
In this scoping review, we investigated the association between 60 different muscle parameters with incident fractures risk in 322 separate analyses within 67 studies.Overall, low muscle mass was poorly/not associated with fracture risk, while low muscle strength and low muscle function were associated with higher risk of fracture.The results showed heterogeneity between the studies, in terms of studies' populations, measurement methods and statistical analysis.Our conclusion is a summary of the observed trends in this review and is not comparable to a meta-analysis.

Muscle mass, quantity and quality
Muscle mass, quantity and quality are objective and reproducible assessments of muscle health. 98The accuracy and the reliability of these assessments mostly depend on the technique used, for which the time available, the radiation dose, the costs and the patient involvement must also be considered.0][101] In this review, we did not find any studies using MRI.DXA and BIA were more studied as part of the diagnostic criteria of most sarcopenia definitions.The muscle quantity can be estimated from its volume using the muscle length and cross-sectional area.As these two properties are also important components of muscle strength, 102,103 the hypothesis is that a low muscle quantity leads to weaker muscle (dynapaenia), which then lead to disbalance and falls. 104At the same time, we know that a tailored exercise programme reduces the risk of fall-related fragility fractures. 105However, the relationship between low muscle mass and fractures has been repeatedly questioned. 12,23,45The results of our scoping review also suggest that a higher muscle mass, as assessed by different parameters, has little protective effect on the occurrence of fragility fractures.Indeed, seven analyses (within three studies) showed even opposite results with an increased risk of fragility fractures with higher muscle mass 44,52,66 : six (110 TPY) analyses for hip fractures and one (15 TPY) analysis for MOF.[57]66,67 The use of LM indexes in fracture prediction models is complex because anthropometric measures are correlated with LM and are associated with fractures.The literature describes weight as a protective factor, height as a risk factor and BMI as having a U-shaped association with fragility fractures. 106he stratification of LM analyses for body size or shape would enable a better estimation of its association with fragility fracture.Note that these considerations differ between the fragility fracture types and the sex (Figure 3A-E).We also know that measures of LM include water, joints and ligaments 107 and may not be specific enough of muscle mass.
Muscle density is a more recent concept.It was first used in CT scans by measuring the X-ray absorption in the different muscle voxels (3D pixels) but is now also available in DXA. 66It is used as a proxy for intramuscular fat infiltration (as fat absorbs less X-rays than bone or muscle) and has been associated with fragility fractures in this review. 37,66,73The bottleneck to more widespread use of CT scanning, including in larger studies, is the increased radiation dose and costs.
Muscle mass/quantity has also been investigated using biological tests, with promising results in fracture prediction.Blood creatine, a breakdown product of muscle, is associated with functional and clinical outcomes. 108Cystatin or its ratio showed a positive association in women with low eGFR and humerus fractures, but it showed conflicting results in men. 36Using the D3-creatine dilution test, Cawthon et al. found a positive association between low eGFR and hip fractures and MOF. 33A review summarizes the necessary assumptions of the creatine dilution test, including individual variation (diet, age, activity level and disease state) that lead to underestimation or overestimation of the measurement. 108As a result, the clinical implementation of blood tests should be further investigated.
Newer methods are being developed such as ultrasound (e.g., with muscle thickness, cross-sectional area, pennation angle and echogenicity) 109 or image analysis (classification, segmentation, texture/pattern analysis and radiomics) using artificial intelligence (AI). 110,111AI models could help us to extract the full information from the DXA scans (or other imaging modalities) and potentially measure new markers of muscle health.Pickhardt et al. analysed low-dose CT scans using deep learning to predict lumbar muscle myosteatosis and cross-sectional area. 112The prediction of hip fracture at 5 years was similar between their model (area under the curve [AUC] 0.709, 95% confidence interval [CI] 0.639-0.778)and the FRAX® (AUC 0.708, 95% CI 0.629-0.787). 112I seems to be a suitable tool to analyse DXA body composition images and to search for unanticipated complex interactions between the available parameters.
The role of muscle mass in fragility fracture remains unclear.The assessment of muscle mass/quantity through the D3-creatine dilution tests and muscle density assessment by DXA and CT imaging seem promising and could be object of further research.Furthermore, AI will undoubtedly influence musculoskeletal imaging and provide novel muscle mass assessments.

Muscle strength
Muscle strength is highly correlated with muscle quantity (length and cross-sectional area), but with greater variability, 102 and is influenced by the conservation of peripheral and central neurological structures. 103Fifty per cent of the total body muscle mass lies in the lower body, while the upper body represents only 25%. 113Even if the quadriceps and psoas muscles make standing and walking possible, HGS has been shown to correlate with leg strength and is similarly predictive of low GS. 114From a clinical perspective, HGS is the most widely used test to assess muscle strength due to its low cost, accessibility, widespread use and reliability, whereas quadriceps testing is more complex and requires more equipment. 45This is probably the reason why fewer studies analysed QS.In this review, both lower HGS and lower QS were significantly associated with higher fracture risk in 37 and 13 (131 TPY) studies, respectively; 41 analyses showed no association between HGS and fracture risk and 15 (389 TPY) analyses between lower QS and fracture risk.
Muscle strength may be useful in predicting fracture risk using grip strength as a practical and reliable proxy of muscle strength.

Muscle function
Muscle function is the most multifactorial determinant of muscle health.It correlates with both muscle mass and strength and is defined as the ability of the muscle to perform a certain task or movement.The assessment of muscle function, as for muscle strength, also depends on peripheral and central neurological structures.In addition, muscle function is closely linked to the brain (mostly through the cerebellum, motor, pre-motor and supplementary motor cortex) when testing balance, coordination or complex tasks.The reasons for variation in measures of muscle function are similar to those for strength testing and are mainly analytical and/or methodological variations.Based on the observations of this review, GS shows a robust association with fracture risk, as all studies showed a significant association between slow GS and higher risk of MOF.The 5×STS was the second most commonly used muscle function test, with comparable results to QS.The 5×STS is a proxy of the thigh strength in addition to coordination ability.These observations emphasize the importance of assessing muscle function during a clinical consultation.Indeed, physicians are trained to assess the risk of falling (and therefore, to some extent, muscle function) by observing the patient walking around the examination room, sitting in the chair, changing clothes and so forth.For example, the chair stand tests (including 5×STS), the timed up and go test (TUGT), the SPPB and the tandem walk test have been validated to assess the mobility status and fall risk in older adults. 115arious muscle functional tests are available and provide an objective assessment of the patient muscle status, and they give an additional information on the patient's risk of fragility fracture.They include more variability than muscle strength or mass assessment but stay reliable overall.These tests were not designed to predict the fracture risk, but as they are associated with multiple medical conditions including neurological and musculoskeletal diseases, their association with fracture is also multifactorial.

Clinical implications
In the field of sarcopenia, the association between muscle parameters and fragility fractures remains subject to debate.In the SDOC sarcopenia definition (2020), the authors argue against the use of muscle mass in further definitions because of insufficient evidence of its association with sarcopenia outcomes (including fractures) and the cost of DXA. 45 Our scoping review similarly suggests that low muscle mass, as currently defined, is not robustly associated with fragility fractures and that an adjustment or stratification for body size is necessary.As we analysed each muscle health component separately and did not assess the other sarcopenia endpoints, our study does not allow us to directly challenge the composite definitions of sarcopenia.On the other hand, the observed association of GS and HGS with fragility fractures supports their use in the diagnostic workflow of current sarcopenia definitions.These muscle parameters provide objective measures of the muscle health and insights on its association with fragility fractures.Ideally, a test or score would be developed to specifically identify the fracture risk associated with sarcopenia, at best independently from the risk of fall.
In the field of osteoporosis, the relationship between bone and muscle has been studied from various angles.Falls are important risk factors for fracture occurrence.They often, but not always, precede the fracture. 9In the causal hypothesis linking muscle mass to fragility fractures, falls are more likely to be a mediator in the equation, involving both dependent and independent pathways, rather than just an intermediate factor.In this scoping review, only few studies demonstrated that the relation between muscle mass, 33,37,55,57 strength 69 and function 31,32,34,55,64,69 with incident fracture was positive and independent from falls.At the cellular level, a cross-talk between muscle and bone has been discussed in studies about osteo-sarcopenia. 13At the organ level, the bone mechanostat hypothesis explains that the properties of load-bearing bones are primarily influenced by their functions, rather than the influence of load and gravitational forces. 116Our study could support this hypothesis considering that muscle function and strength have an additive discriminative value in fragility fractures prediction models, assuming that bone properties are related in the same way.However, muscle mass and quantity, as it currently stands, do not appear to have an independent effect on fracture susceptibility.Heymsfield et al. insisted on the importance of muscle 'form' (size and shape) and not only muscle function in the pathophysiology of adverse events (cf.OFF hypothesis: Outcome follow function, follow form), based on the axiom that without the physical form of the muscle, there would be no function. 117The overall lack of association between muscle mass/quantity and fractures that we highlight in this review does not discredit its importance in the pathophysiology of osteoporosis and sarcopenia.Further research is needed on muscle mass, quantity and quality in the prediction of fracture risk, including a judicious use of anthropometric measures.The D3-creatine dilution test and the CT-scan measures showed promising results, while LM, its indexes and the new statistical approaches using AI need to be further investigated.
Muscle health parameters are important in the prevention and diagnostic of sarcopenia and in the assessment of osteoporotic patients.This scoping review highlights the benefits and the gaps of muscle health tests in clinical setting and in community-dwelling older adults.

Strengths and limitations
This study has some limitations.First, a common limitation to scoping reviews is the publication bias.Positive studies are more likely to be published, whereas negative studies may be discontinued.However, most of the results analysed are inconclusive (no association) and some are even negative and contra-intuitive (e.g., the positive association between muscle mass and fragility fracture risk), suggesting that the data observed and discussed here are undistorted.
Second, the overall quality and risk of bias of the included studies were not systematically assessed.However, this is not a requirement for conducting a scoping review.As shown in Tables 1 and 2, the majority of the included studies have large sample sizes and long follow-up periods and come from recognized and well-conducted national or international cohorts.Finally, although not related to the scoping review itself, the included studies have some limitations that weaken their interpretation, such as the consideration of non-MOF fractures as fragility fractures (Figures S3f-S5f ); the lack of a clear fragility fracture definition 30,34,41,69,71,78,83,88,90 ; and the lack of systematic radiographic assessment for fracture detection, as some incident fractures were only collected based on questionnaires and general practitioners.
To the best of our knowledge, this is the first review, based on a systematic search, that thoroughly reviews studies that investigated the association of incident fracture risk with muscle mass/quantity/quality, strength and/or functional parameters.The rigorous systematic search, under the supervision of medical library experts, adds value to the current study.The inclusion of only prospective studies is a major strength, as prospective studies have a temporal framework to assess causality (outcome occurring after exposure), which positions them as strong scientific evidence.In addition, most of the analyses were performed with the muscle parameter as a continuous variable, assuming that the risk is proportional to the parameter in question.Some studies had previously categorized the variables using percentiles or a specific value (cf., which lost statistical information but made it easier to use in clinical practice.Furthermore, following the PRISMA checklist for reporting (cf.supporting information) and the JBI methodology for writing improves the transparency, reproducibility and, ultimately, the overall quality of this review.Moreover, we visualize the trend of associations between muscle parameters and fracture risk using adapted forest plots.Finally, our review highlights muscle parameters that could be further analysed in a meta-analysis.

Conclusions
This scoping review gives a broad overview of the gaps and evidences in the relationship between muscle parameters and fragility fractures.Poorer muscle function followed by lower muscle strength were the parameters mostly related to a higher risk of incident fragility fractures.For daily clinical practice, this review suggests that measures of HGS and GS are the most useful methods to assess muscle-dependent fracture risk.This supports their use in the evaluation of sarcopenia.This review also confirms that muscle mass, as currently defined, is a poor independent predictor of fragility fracture.For future research and development of fragility fracture prediction models, it will be necessary to determine whether muscle-associated fracture risk is fully independent from other risk factors.In addition, further investigation of DXA images, including body composition, using AI methods may reveal new complex interactions between muscle tissue and fragility fractures.

Figure 1
Figure 1 PRISMA 2020 flow diagram of the study.

Figure 2 (
Figure 2 (A-F) Summary of the 322 analyses for each muscle assessment and each fracture types.

Figure 3 (
Figure 3 (A-E) Muscle mass/quantity/quality parameters and risk of incident fragility fractures.

Figure 4 (
Figure 4 (A-E) Muscle strength parameters and risk of incident fragility fractures.

Figure 5 (
Figure 5 (A-E) Muscle function parameters and risk of incident fragility fractures.

Table 1
Summary of the 67 included studies and main characteristics

Table 2
Characteristics of included studies

Table 2 (
continued) 2. The ultrasonography of the quadriceps (US) was not associated with fragility fractures in one (2 TPY) analysis using quadriceps quantity/quality.The parameters derived from the CT scan (lower thigh muscle cross-sectional area representing muscle mass and lower thigh muscle attenuation representing muscle quality) were positively and not associated with fractures in three (63 TPY) and five (105 TPY) analyses, respectively.Muscle mass can also be estimated using anthropometric prediction equations.The Lee equation includes height, weight, waist circumference, serum creatinine level and health behaviour factors.