Dysmobility Syndrome Independently Increases Fracture Risk in the Osteoporotic Fractures in Men (MrOS) Prospective Cohort Study

Abstract

We proposed the term “dysmobility syndrome” (DS) to identify individuals with impaired musculoskeletal health, a risk factor for falls and fractures. Whether DS is associated with increased risk of incident fracture is unknown.

The “Osteoporotic Fractures in Men” study enrolled 5994 men ages ≥65 years, between March 2000 and April 2002. We used baseline data to determine whether DS increased fracture risk, independent of FRAX. Men met DS criteria at baseline if they had ≥3 of the following: appendicular lean mass/height² <7.26 kg/m², total body fat >30%, spine or hip T-score ≤−2.5, grip strength <30 kg, gait speed <1.0 m/s, and ≥1 fall within 12 months. We examined whether baseline DS increased the risk of hip and major osteoporotic fractures (MOF) over a median of 14 (IQR 9, 15) years.

Among 5834 men mean age 74±6 years, 471 (8%) had DS and 635 (11%) experienced a MOF, including 274 (5%) hip fractures. Age (per SD increase) conferred a HR of 1.72 (95% CI, 1.59, 1.86), DS conferred a HR of 3.45 (95% CI 2.78, 4.29,) and FRAX calculated with BMD (per %) conferred a HR of 1.10 (95% CI 1.08, 1.11) for MOF. Prediction of MOF using the FRAX score provided a concordance value of 0.67 (SE 0.012). Concordance increased to 0.69 (SE 0.012) by adding DS and to 0.70 (SE 0.012 by adding DS and age to the multivariate model. Kaplan Meier curves indicated that men with both DS and a FRAX risk above the National Osteoporosis Foundation treatment thresholds had higher MOF (HR 6.23, 95% CI 3.10, 12.54) and hip (HR 7.73, 95% CI 5.95, 10.04) fracture risk than men with neither condition.

We suggest further studies to determine the optimal criteria for DS, and to test DS as a predictor of falls and fractures, especially in women.

INTRODUCTION

Fragility fracture risk increases significantly with increasing age and is a major cause of morbidity and mortality in older adults.^(1,2) Worldwide, approximately 8.9 million fragility fractures occur annually.⁽³⁾ More than half of these fractures occur in adults with a bone mineral density (BMD) T-score better than −2.5.⁽⁴⁾ The introduction of fracture risk calculators, such as the Fracture Risk Assessment Tool (FRAX), has markedly improved clinicians’ ability to predict osteoporotic fractures over BMD values alone.⁽⁵⁾

However, FRAX does not directly account for impaired musculoskeletal health characterized by falls or sarcopenia, both of which increase fracture risk.^(6–12) Indeed, impaired physical function and sarcopenia predict fractures, and sarcopenia combined with FRAX improves fracture prediction.^(10,13–15) We and others have suggested terms such as “sarcoporosis,” “sarco-osteopenia,” “sarco-osteoporosis,” “osteo-sarcopenia,” “sarcopenic obesity” and “osteosarcopenic obesity” to denote impaired muscle function and reduced BMD. These terms all intend to describe the increased risk of adverse musculoskeletal health outcomes (e.g. impaired mobility, falls, and fractures) when low BMD and impaired musculoskeletal health occur together.^(16–20)

We proposed the concept of “dysmobility syndrome” (DS) to help clinicians identify patients with impaired musculoskeletal health⁽²¹⁾ and recently suggested that this concept could potentially ameliorate the current crisis in osteoporosis care.⁽²²⁾ In this score-based approach, individuals meet criteria for DS if they have ≥3 of the following 6 conditions: low appendicular lean mass/height², high percent body fat, osteoporosis based on a spine or hip T-score ≤−2.5, low grip strength, low gait speed or ≥1 fall in the prior year.

Since the concept of DS was proposed, some studies have evaluated its utility in predicting adverse health outcomes. For example, in 121 postmenopausal women, DS was more common in those with a prior fragility fracture compared to women without fracture.⁽¹⁸⁾ Additionally, among 2975 individuals in the National Health and Nutrition Examination Survey (NHANES), DS was associated with increased mortality.⁽²³⁾ However, whether baseline DS increases the risk of fragility fractures remains unknown. Thus, the aim of this study was to examine whether DS identifies individuals at greater risk of incidence fragility fracture. We used data from the Osteoporotic Fractures in Men (MrOS) prospective cohort study to evaluate whether DS at baseline increased incident fragility fracture risk. We also evaluated whether the risk conferred by DS was independent of the risk predicted by age and FRAX.

METHODS

A total of 5994 men, age ≥65 years without hip arthroplasty and able to walk without assistance, were initially enrolled in the MrOS multicenter prospective cohort study. The design of MrOS, including methods used to measure BMD and physical function, is described elsewhere.^(24,25) Men were recruited between March 2000 and April 2002. The main outcome measures for MrOS were falls and fractures, which were recorded at the baseline visit and then captured by questionnaires, mailed every four months to participants. All fractures were adjudicated by physician review of radiology reports. For the current study, men were followed for a median of 14 (interquartile range; 9, 15) years.

We defined baseline DS as having three or more of the following six factors: appendicular lean mass/height² <7.26 kg/m², percent body fat >30%, lumbar spine or hip (femur neck or total proximal femur) T-score ≤−2.5, gait speed <1.0 meter/second, grip strength <30 kg and reported history of ≥1 fall in the last 12 months. The cutoff values for DS criteria were chosen using consensus documents when possible, as previously described.⁽²¹⁾

We excluded 160 men from the current analysis because they had incomplete data for at least one of the six DS measures. These individuals were significantly older and had higher FRAX scores than the 5834 men with complete data who are included in this analysis (Supplemental Table 1).

STATISTICAL ANALYSIS

We used quantile against theoretical quantile plots (QQ plots) to determine whether data exhibited a normal distribution, then summarized data using the mean ± standard deviation or median (interquartile range). We compared men with and without DS using chi-square and t-tests. We used survival analyses to explore how time to MOF or hip fracture were related to various risk factors. Men were censored at the time of first event (fracture) or last observation. We fit single-variable Cox proportional hazards models using age, DS and FRAX scores. We analyzed FRAX scores in four ways: with and without femoral neck BMD, and as a continuous or dichotomous variable (FRAX ≥20% for MOF, and ≥3% for hip fracture). We next fit multiple variable Cox proportional hazard models including combinations of age, DS and FRAX values. We used concordance to evaluate the predictive accuracy and discriminatory power of the models. A concordance value of 1.0 would indicate a model with perfect discrimination, a value of 0.5 works as well as flipping a coin, and values ≥0.6 are good models. Finally, we compared time to first major osteoporotic and hip fracture in men with DS, a FRAX score above the treatment threshold, both conditions and neither condition using Cox proportional hazard models, Kaplan-Meier survival curves and log-rank tests.

Four components of DS are continuous variables (body fat, muscle mass, grip strength, gait speed) with proposed thresholds for abnormal. We generated receiver operator curves to determine the ability of these criteria to predict MOF as measured by area under the curve. To determine an optimal cutoff for prediction, we then applied the Youden index, which identifies the value at which the sum of the true positive and true negative values is maximal (e.g. the thresholds at which sensitivity and specificity were both maximized as a means of predicting MOF). In a sensitivity analysis, we controlled for age in all models.

We used Analyse-it for Microsoft Excel (version 2.20, Analyse-it Software, Ltd. http://www.analyse-it.com/; 2009) to graph data for normality and R (www.r-project.org) for all other analyses.

RESULTS

Among 5834 men observed for a median of 14 (interquartile range; 9, 15) years, 635 (11%) experienced a major osteoporotic fracture, including 274 (5%) men with hip fractures. Table 1 summarizes the demographic and functional status of the cohort. Men with DS at baseline (n=471, 8%) were older (78±7 vs. 73±6 years, p<0.0001), had lower BMI (26.7±4.4 vs. 27.4±3.7 kg/m², p=0.0001), and higher FRAX scores (calculated with femoral neck BMD) than men without DS. More specifically, the MOF FRAX score expressed as median (IQR) was 8.5% (6.2, 12.4%) vs. 6.3% (4.8, 8.9%, p<0.0001) and the hip FRAX score was 2.9% (1.6, 5.2%) vs. 1.3% (0.6, 2.6%, p<0.0001) in men with DS compared to men without DS. By definition, men with DS had higher prevalence of osteoporosis, lower muscle mass, grip strength and gait speed, higher percent body fat and more falls than men without DS (all p < 0.0001, Table 1).

Table 1:

Characteristics of Participants with and without Dysmobility Syndrome

Characteristic	All Men, n=5834	Men with DS, n=471	Men without DS, n=5363	p-value
Age, years	74 ± 6	78 ± 7	73 ± 6	<0.0001
Caucasian Race	5219 (89%)	411 (87%)	4808 (90%)	0.05
Body Mass Index, kg/m2	27.4 ± 3.8	26.7 ± 4.4	27.4 ± 3.7	0.0001
FRAX score, major osteoporotic fracture	6.5 (4.8, 9.2)	8.5 (6.2, 12.4)	6.3 (4.8, 8.9)	<0.0001f
FRAX score, hip fracture	1.4 (0.65, 2.80)	2.9 (1.6–5.2)	1.3 (0.6, 2.6)	<0.0001f
ALM/ht2	8.0 ± 0.9	7.2 ± 1.0	8.1 ± 0.9	<0.0001
Low ALM/ht2a	1274 (22%)	330 (70%)	944 (18%)	<0.0001
Body Fat, %	26 ± 5	29 ± 6	26 ± 5	<0.0001
High Body Fatb	1406 (24%)	274 (58%)	1132 (21%)	<0.0001
Gait Speed, meters/second	1.25 ± 0.24	1.00 ± 0.25	1.27 ± 0.22	<0.0001
Low Gait Speedc	774 (13%)	272 (58%)	502 (9%)	<0.0001
Grip Strength, kg	42 ± 9	32 ± 8	42 ± 8	<0.0001
Low Grip Strengthd	374 (6%)	189 (40%)	185 (3%)	<0.0001
Falls in last 12 months	1220 (21%)	301 (64%)	919 (17%)	<0.0001
Lowest Spine or Hip T-Score	−1.3 ± 1.0	−2.0 ± 1.1	−1.2 ± 1.0	<0.0001
Osteoporosise	567 (10%)	195 (41%)	372 (7%)	<0.0001
Incident Major Fracture	635 (11%)	98 (21%)	537 (10%)	<0.0001
Incident Hip Fracture	274 (5%)	45 (10%)	229 (4%)	<0.0001

FRAX scores are reported as the median (interquartile range).

^aappendicular lean mass/height² <7.26 kg/m²

^bpercent body fat >30%

^cgait speed <1.0 meter/second

^dgrip strength <30 kg

^elumbar spine, femoral neck or total hip T-score ≤−2.5

^fKruskal-Wallis test

In Supplemental Table 2, we present correlations between the individual DS criteria and age. Age was positively associated with falls and negatively associated with lowest T-score, muscle mass, grip strength and gait speed. As expected, many criteria for DS were co-linear. For example, muscle mass was positively associated with grip strength, gait speed, T-score and body fat. Further associations are found within the table.

Both MOF and hip fractures became more common as the DS score increased (p<0.0001, Figure 1). When analyzed separately, each of the six DS criteria showed a positive relationship with hazard of major osteoporotic fracture (Table 2), but multivariate analyses including each of the components revealed that high body fat was no longer significant. When analyzed separately in univariate models, all DS criteria except for high body fat showed a positive relationship with hazards of hip fracture. However, high body fat and low grip strength were no longer significant in multivariate models predicting hip fracture. Further adjustment for age did not alter these findings.

Figure 1: Percent of Men with Incident Major Osteoporotic Fracture and Hip Fracture by Baseline Dysmobility Syndrome Score.

The figure depicts the proportion of men with incident major osteoporotic fractures, including hip fractures, as a function of the number of baseline criteria for DS (DS score). Only one man had a DS score of six; this data point is omitted. Increasing DS score was associated with increased odds of major osteoporotic fractures (p<0.0001), including hip fractures (p<0.0001).

Table 2.

Hazards of Fracture as a Function of the Dysmobility Syndrome Criteria in 5834 men

		Hazard Ratio for Major Osteoporotic Fracture	Hazard Ratio for Hip Fracture
Univariate Analysis	Falls within 12 months	1.51 (1.26, 1.80)	1.69 (1.30, 2.19)
	Osteoporosis	3.33 (2.77, 4.01)	3.28 (2.48, 4.33)
	High Body Fat	1.27 (1.07, 1.51)	1.01 (0.77, 1.34)
	Low Muscle Mass	1.73 (1.46, 2.06)	1.86 (1.43, 2.41)
	Low Grip Strength	2.02 (1.51, 2.70)	1.92 (1.22, 3.04)
	Slow Gait Speed	1.94 (1.56, 2.40)	2.79 (2.08, 3.74)
Age, in Tertiles	65–70 years old 71–75 years old 76–100 years old	Referent 1.77 (1.42, 2.20) 3.26 (2.67, 3.97)	Referent 2.39 (1.65, 3.44) 5.59 (4.01, 7.79)
Multivariate Analysis	Falls within 12 months	1.41 (1.18, 1.68)	1.54 (1.19, 2.01)
	Osteoporosis	3.04 (2.51, 3.68)	2.92 (2.19, 3.90)
	High Body Fat	1.19 (1.00, 1.42)	0.91 (0.68, 1.21)
	Low Muscle Mass	1.39 (1.16, 1.66)	1.48 (1.13, 1.94)
	Low Grip Strength	1.41 (1.04, 1.90)	1.22 (0.76, 1.96)
	Slow Gait Speed	1.73 (1.39, 2.16)	2.60 (1.92, 3.52)
Multivariate Analysis, adjusted for Age	Age 71–75 years	1.69 (1.36, 2.11)	2.28 (1.57, 3.29)
	Age 76–100 years	2.74 (2.23, 3.37)	4.55 (3.23, 6.41)
	Falls within 12 months	1.35 (1.13, 1.62)	1.47 (1.13, 1.92)
	Osteoporosis	2.95 (2.44, 3.57)	2.78 (2.09, 3.71)
	High Body Fat	1.21 (1.01, 1.44)	0.93 (0.70, 1.23)
	Low Muscle Mass	1.21 (1.01, 1.45)	1.24 (0.95, 1.62)
	Low Grip Strength	1.20 (0.89, 1.62)	0.99 (0.62, 1.58)
	Slow Gait Speed	1.45 (1.16, 1.80)	2.02 (1.49, 2.73)

The hazard ratio is presented with its 95% confidence intervals. Hazard ratios represent the hazards of major osteoporotic or hip fracture in men with the condition, compared to men without the condition, at any given time point.

In Cox proportional hazard models (Table 3), age, DS and FRAX (continuous variable, determined with or without BMD) each showed a positive relationship with the hazards of major osteoporotic fractures. Age (per SD) conferred a HR of 1.72 (95% CI, 1.59, 1.86, concordance 0.65), DS conferred a HR of 3.45 (95% CI 2.78, 4.29, concordance 0.56) and FRAX determined with BMD (per %) conferred a HR of 1.10 (95% CI 1.08, 1.11, concordance 0.67) for MOF. By itself, prediction of MOF using the FRAX score determined with BMD provided a concordance value of 0.67 (SE 0.012). Adding DS to the model increased concordance to 0.69 (SE 0.012). Concordance increased to 0.70 (SE 0.012) when including FRAX, DS and age in a multivariate model. Findings were very similar when using the FRAX score determined without femoral neck BMD (Table 3).

Table 3:

Cox Proportional Hazards Models for Incident Fracture, using FRAX as a continuous variable with or without femur neck BMD

		Major Osteoporotic Fracture 635 events	Hip Fracture 274 events
	Model	Hazard Ratio (95% Confidence Interval)	Hazard Ratio (95% Confidence Interval)
	Dysmobility Syndrome	3.45 (2.78, 4.29) Concordance 0.56 (SE 0.005)	3.71 (2.69, 5.12) Concordance 0.57 (SE 0.007)
	Age, per SDa	1.72 (1.59, 1.86) Concordance 0.65 (SE 0.012)	2.15 (1.90, 2.42) Concordance 0.71 (se 0.018)
FRAX determined without BMD	FRAX Score, per 1%b	1.09 (1.07, 1.10) Concordance 0.64 (SE 0.012)	1.11 (1.09, 1.13) Concordance 0.70 (SE 0.018)
	Dysmobility Syndrome +FRAX Scorec	3.26 (2.05, 5.19)d 1.08 (1.07, 1.10) Concordance 0.67 (SE 0.012)	3.13 (2.26, 4.33)e 1.10 (1.08, 1.13) Concordance 0.72 (SE 0.018)
	Dysmobility Syndrome, FRAX Scorec and Age (per SD)	2.42 (1.93, 3.03)f 1.05 (1.04, 1.07) 1.44 (1.32, 1.57) Concordance 0.68 (SE 0.012)	2.28 (1.63, 3.18)g 1.04 (1.01, 1.07) 1.85 (1.61, 2.13) Concordance 0.72 (SE 0.018)
FRAX determined with BMD	FRAX Score, per %	1.10 (1.08, 1.11) Concordance 0.67 (SE 0.012)	1.11 (1.09, 1.13) Concordance 0.77 (SE 0.018)
	Dysmobility Syndrome + FRAX Scorec	3.34 (2.26, 4.95)h 1.09 (1.08, 1.11) Concordance 0.69 (SE 0.012)	2.98 (2.15, 4.13) 1.11 (1.09, 1.12) Concordance 0.76 (SE 0.018)
	Dysmobility Syndrome, FRAX Scorec and Age (per SD)	2.16 (1.73, 2.71)i 1.08 (1.06, 1.09) 1.52 (1.40, 1.65) Concordance 0.70 (SE 0.012)	2.14 (1.53, 2.99) 1.08 (1.06, 1.10) 1.85 (1.63, 2.10) Concordance 0.74 (SE 0.018)

^aThe hazard of fracture by age is described by standard deviation (SD) of age (5.9 years)

^bIn crude models, we analyzed the hazards of fracture using dysmobility syndrome alone, and the hazard of fracture using the FRAX score for major and hip fractures.

^cIn adjusted models, we analyzed the hazards of fracture as a function of both dysmobility syndrome and the FRAX score as a continuous variable.

^dP-value = 0.285 for the interaction between FRAX and dysmobility syndrome

^eP-value = 0.050 for the interaction between FRAX, dysmobility syndrome and age

^fP-value = 0.285 for the interaction between FRAX and dysmobility syndrome

^gP-value = 0.051 for the interaction between FRAX, dysmobility syndrome and age

^hP-value = 0.230 for the interaction between FRAX and dysmobility syndrome

ⁱP-value = 0.151 for the interaction between FRAX, dysmobility syndrome and age

Likewise, in Cox proportional hazard models of hip fracture (Table 3), age, DS and FRAX (continuous variable, determined with or without BMD) each showed a positive relationship with risk of hip fracture. Age (per SD) conferred a HR of 2.15 (95% CI, 1.90, 2.42, concordance 0.71), DS conferred a HR of 3.71 (95% CI 2.69, 5.12, concordance 0.57) and FRAX determined with BMD (per %) conferred a HR of 1.11 (95% CI 1.09, 1.13, concordance 0.77) for MOF. A multivariate model of FRAX and DS provided a nearly identical concordance value of 0.76 (SE 0.018) to that obtained with the FRAX score alone. The final multivariate model of FRAX, DS and age reduced concordance to 0.74 (0.018).

We next used FRAX as a dichotomous variable, where the probability of MOF and hip fracture were ≥20% and ≥3% respectively. As expected when changing continuous data to a dichotomous variable,⁽²⁶⁾ the FRAX score provided lower concordance in models predicting MOF and hip fractures (Supplemental Table 3).

We next divided men into one of four groups: those with both a MOF FRAX score ≥20% and DS (n=27), a high FRAX score but no DS (n=97), DS but not a high FRAX score (n=444) and men with neither condition (referent, n=5266). Kaplan Meier survival curves indicate that men with both DS and a FRAX score at or above the NOF treatment threshold had higher fracture risk compared to men with neither or one condition (Figure 2, Table 4). More specifically, men with both DS and FRAX risk above treatment threshold had a higher risk for MOF (HR 6.23, 95% CI 3.10, 12.54) than those with neither condition (referent). Similarly, men with DS alone (HR 3.45, 95% CI 2.75, 4.32) or FRAX risk above treatment threshold alone (HR 3.29, 95% CI 2,22, 4.89) had a higher MOF risk than men with neither condition (p<0.001).

Figure 2: Cumulative Major Osteoporotic Fractures (2a) and Hip (2b) Fractures:

2a.) For this analysis, men were divided into one of four groups: those with both a high overall MOF FRAX score and DS (n=27), a high FRAX score but no DS (n=97), DS but not a high FRAX score (n=444) and men with neither condition (referent, n=5266). Men with both DS and a FRAX MOF score ≥20% experienced more fractures than the referent (Table 4, p<0.001). Likewise, men with DS had more fractures than men with neither condition (p<0.001). Similarly, men with a FRAX score ≥20% had more fractures than men with neither condition (p<0.001).

2b.) Men were divided into four groups: those with both a high hip FRAX score and DS (n=229), a high hip FRAX score but no DS (n=1088), those with DS but not a high hip FRAX score (n=242) and men with neither condition (referent, n=4275). As shown in the figure and Table 4, men with both DS and a hip FRAX score ≥3% experienced a greater risk of fracture than men with a hip FRAX score ≥3% (p<0.0001), DS (p<0.0001) or neither condition (p<0.0001). Men with DS had no difference in hip fractures, compared to men with neither condition (p=0.705). Men with a high FRAX score sustained a more fractures than the referent (p<0.0001).

Table 4:

Hazards for Fracture in Men with Dysmobility Syndrome, High Fracture Risk, Both or Neither Conditions

	Hazard Ratio for Major Osteoporotic Fracture	Hazard Ratio for Hip Fracture
FRAX determined with femoral neck bone mineral density, Figure 2
Neither condition	referent	referent
High Fracture Risk	3.29 (2.22, 4.89)	2.89 (2.43, 3.45)
Dysmobility Syndrome	3.45 (2.75, 4.32)	2.25 (1.53, 3.31)
Both Conditions	6.23 (3.10, 12.54)	7.73 (5.95, 10.04)
	Concordance 0.57	Concordance 0.64
FRAX determined without femoral neck bone mineral density, Supplemental Figure 1
Neither condition	referent	referent
High Fracture Risk	2.56 (1.74, 3.77)	2.32 (1.96, 2.75)
Dysmobility Syndrome	3.38 (2.69, 4.23)	3.25 (2.19, 4.83)
Both Conditions	8.47 (4.37, 16.42)	6.20 (4.75, 8.09)
	Concordance 0.57	Concordance 0.64

In similar manner, we divided men were divided into four groups based on their hip fracture risk: those with both a hip FRAX score ≥3% and DS (n=229), a high hip FRAX score but no DS (n=1088), those with DS but not a high hip FRAX score (n=242) and men with neither condition (referent, n=4275). Men with both DS and a hip FRAX score ≥3% had a higher hip fracture risk (HR 7.73, 95% CI 5.95, 10.04) than men with neither condition (referent, Table 4). Similarly, men with DS alone (HR 2.25, 95% CI 1.53, 3.31) and men with a hip fracture FRAX score ≥3% alone had higher risk of hip fracture (HR 2.89, 95% CI 2.43, 3.45) than the referent.

Receiver operator curves of the four continuous variables comprising DS revealed that none had high area under the curve (AUC) value for predicting MOF (Supplemental Figures 2–5). The AUC values for body fat, muscle mass, grip strength and gait speed were 0.524, 0.556, 0.568 and 0.533 respectively. Youden index analyses suggested that differing cut-points could be used, to maximize both sensitivity and specificity of these four continuous variables as a means of predicting MOF. Proposed cut points are provided in the supplemental figures.

DISCUSSION

Among men enrolled in the MrOS prospective cohort study, age, FRAX scores and DS were each positively associated with the hazard ratio of future major osteoporotic fracture and hip fracture. The addition of DS to a model including FRAX improved concordance, but improvements were very small in magnitude. However, Kaplan Meier curves indicated that men with both DS and an elevated FRAX score had a significantly higher fracture risk, compared to men with just one or neither condition.

Whether DS is a feasible approach to clinically improving fracture prevention is unknown. However, current approaches to risk reduction are suboptimal, as many high-risk patients do not receive appropriate treatment to reduce future fractures,^(27,28) a situation deemed scandalous by experts in the field.^(29,30) Our study suggests that individuals with DS are at greater risk of fractures. Such individuals might benefit from additional interventions such as formal falls risk assessment, nutritional evaluation, physical therapy, muscle strengthening programs and/or bone anabolic therapy, rather than simple prescription of a bone-active medication.

Previous studies have suggested that sarcopenia increases the risk of both falls and fractures.^(7,8,31) Individuals with both sarcopenia and osteoporosis have a greater risk of fracture than osteoporotic non-sarcopenic individuals.^(20,32) Sarcopenia was an independent predictor of fracture in the Geneva Retirees Cohort, with sarcopenia improving fracture risk prediction beyond FRAX.⁽¹⁰⁾ Similarly, Yu et al. reported that sarcopenia predicted fractures in the Chinese cohort of the MrOS study and additionally showed improved fracture prediction when combined with FRAX.^(13,14) Similar observations were reported for sarcopenic obesity and osteosarcopenic obesity.^{(16,19,20,33)} Although previous data from the US MrOS study found that impaired physical function predicted hip fractures,⁽¹⁵⁾ current sarcopenia definitions did not consistently predict fractures.⁽³⁴⁾ These studies suggest that incorporating functional status into fracture risk assessment would improve ability to predict fractures. The dysmobilty syndrome concept might be a step in this direction.

Experts coined the term “dismobility” to identify individuals at high risk for adverse health outcomes based on their very low gait speed (below 0.6 m/s).^(35,36) Dysmobility syndrome includes slow gait speed (although using a higher cut-off of 1.0 m/s) but adds additional factors that identify impaired musculoskeletal health. These factors and their cut-offs were chosen based on the available literature and previous studies.^(21,22) Although the current definition of DS (the factors included, thresholds used to diagnose each criterion, and weight of each factor) is far from perfect, our study results support its potential ability to identify individuals at greater risk of osteoporotic fracture.

Other groups have linked DS with adverse health outcomes. In a retrospective case-control study, Iolascon et al. found DS was more common in postmenopausal women who sustained a fragility fracture, compared to those without fracture.⁽¹⁸⁾ In the NHANES cohort, Looker found that DS was associated with higher mortality in US men and women ≥50 years old.⁽²³⁾ Lee and colleagues reported similar findings in Asians.⁽³⁷⁾ In the Hertfordshire Cohort Study, DS was associated with falls, but not fractures.⁽⁷⁾ Hong and colleagues recently published findings that older Koreans with DS had a higher prevalence of several geriatric conditions such as cognitive impairment, malnutrition or depression.⁽³⁸⁾

Our study has several strengths. We used the large prospective MrOS cohort study, in which nearly all participants had baseline measures permitting categorization of DS, to test our hypothesis. The median 14-year follow-up time and large number of osteoporotic fractures are additional strengths. Our study is novel, using a score-based approach to identify individuals with impaired musculoskeletal health. Such an approach is not included in widely used fracture risk calculators such as FRAX. However, we recognize limitations of the current DS approach and this study. First, the six components of DS were selected from review of the literature. Second, we weighted each of the six DS factors equally. The inclusion of all six criteria, optimal diagnostic threshold and weighting of each factor clearly deserves future study. Third, we acknowledge some overlap between the factors included in DS and FRAX, including obesity/body fat and bone mineral density. Fourth, the feasibility of diagnosing DS in clinical practice might be limited. Fifth, our study focused on men. The utility of DS must be tested in other populations, including other racial groups and postmenopausal women. Additional studies are clearly needed to determine if other factors (such as symptomatic osteoarthritis and diabetes mellitus) should be considered in a syndrome-based approach to optimally identify those at increased fracture risk.

In conclusion, our study highlights the importance of impaired musculoskeletal health when determining an individual’s fracture risk. In our study, men with DS and a FRAX score above the US treatment threshold had a higher fracture risk, compared to men with neither condition. Further studies are necessary to optimize DS, test it in other populations, and evaluate whether DS predicts injurious falls. Finally, studies should test the logical hypothesis that interventions in those with DS reduce falls and fractures.

Question:

Does a composite measure of impaired musculoskeletal health termed “dysmobility syndrome” independently increase the risk of incident osteoporotic fractures?

Findings:

In this prospective cohort study of almost 6,000 older men, dysmobility syndrome at baseline was an independent risk factor for fracture after accounting for age and FRAX score. Men with dysmobility syndrome and a high FRAX score had a higher risk of fracture than men with neither condition.

Meaning:

Dysmobility syndrome identifies men at higher risk of fragility fracture.

All authors:

• Made substantial contributions to study design, data acquisition or statistical analysis and interpretation of data

• Participated in writing the manuscript or revising it

• Approved the final manuscript

• Agreed to be accountable for the work and to answer questions related to the accuracy or integrity of the work