Asymmetries Identified in Sit-to-Stand Task Explain Physical Function After Hip Fracture

Abstract

Background: 

Several known demographic and functional characteristics combine to predict physical function after hip fracture. Long-term weight-bearing asymmetries, evident during functional movements after hip fracture, contribute to limited mobility and large asymmetries in muscle function are linked to a high rate of injurious falls. Although postfracture mobility is commonly measured as whole body movement, a force-plate imbedded chair can identify individual limb contributions to an important task like moving from a sitting to standing position. The modified Physical Performance Test (mPPT) and stair climb test (SCT) are reliable, valid measures of function that predict independence after hip fracture. The purpose of this study was to determine to what extent asymmetry during a sit-to-stand task (STST) predicts function (mPPT, 12-step SCT), above and beyond other known predictors.

Methods: 

Thirty-one independent community-dwelling older adults, recently discharged from usual care physical therapy (mean [standard deviation], 77.7 [10.5] years, 10 male), within 2 to 8 months postfracture, volunteered for this study. Participants performed an STST on a force-plate-imbedded chair designed to identify individual limb contributions during an STST. Asymmetry magnitude during the STST was determined for each individual. In addition, mPPT and SCT were assessed and regression analyses were performed to determine the contribution of asymmetry to the variance in these physical function scores beyond other factors predicting function.

Results: 

Demographic factors (sex, time since fracture, repair type, and body mass index) were not significantly related to function in this sample. Age, gait speed, knee extension strength, balance confidence, and functional self-report were each significantly related to both mPPT (r = 0.43–0.86) and SCT (r = 0.40–0.83), and were retained in the regression model. Included variables accounted for 83.4% of the variance in mPPT score, and asymmetry during the STST did not significantly contribute to explaining variability in mPPT (P = .23). Variables in the regression model accounted for 78.0% of the variance in SCT score, and STST asymmetry explained 7.1% (P < .005) of the variance in SCT score.

Discussion: 

In this small sample, asymmetry contributed significantly to explaining the variability in SCT performance, but not mPPT score. The SCT requires greater unilateral strength and control than the battery of items that comprise the mPPT. This contributes to the disproportionate number of falls occurring during stair ambulation (>10% of all fall-related deaths), relative to the minimal time typically involved in stair negotiation. Our results indicate potential benefit to identifying injured limb asymmetries as they predict function in challenging, high-risk functional tasks after hip fracture.

Conclusion: 

Although gait speed is the best explanator of physical function in older adults after hip fracture, lower extremity asymmetry during an STST provides a unique contribution to explaining high-level ambulatory performance after hip fracture. Efforts to reduce weight-bearing asymmetry during rehabilitation following hip fracture may improve function and recovery.

INTRODUCTION

Enduring asymmetry is evident in both muscle force output and vertical ground reaction (vGRF) forces during a sit-to-stand task (STST) following hip fracture.^1–3 Because the surgical limb typically experiences long-lasting deficits, lower extremity asymmetries often endure,² and have been implicated in gait limitations^4,5 and elevated fall risk^6,7 among frail older adults, particularly after fracture. These weight-bearing asymmetries persist after lower limb surgical repairs despite gradual strength gains and pain reduction following fracture.^8–10

Asymmetries in muscle function (strength and power) and vGRF during the STST should be addressed, as mobility impairments and an increased fall risk linked to asymmetries may contribute to poor balance confidence, increased sedentary behavior,¹¹ and a resulting cascade of health problems.^12–18 Because half of those who experience a hip fracture will fall within 6 months after hospital discharge,¹⁹ and because hip fracture survivors are up to 5 times more likely to experience an additional fall-related fracture within 1 year after hip fracture,²⁰ identifying and integrating strategies to mitigate falls and improve mobility in this vulnerable population is important. Asymmetries in muscle function and vGRFs may be modifiable risk factors following hip fracture and thus could inform new rehabilitation strategies.

Performance of the STST is one of the more challenging tasks older adults perform daily. Although several variables contribute to the successful completion of an STST,^21,22 strength is a key contributor.^23–26 Knee extension strength predicts the lowest seated surface from which one can rise,²⁷ and inability to consistently rise from a chair predicts pending disability.²⁸ As older adults experience an immediate strength decline of up to 50% after hip fracture,²⁹ it is expected that many will experience a subsequent decline in physical function and difficulty performing this challenging task.

One less frequently addressed factor that might impact physical function after hip fracture is weight-bearing asymmetry during physical task performance. Typically, whole body measurements in physical movements are collected to quantify physical function after hip fracture (ie, time required to walk 10 m or to climb stairs), with little effort to identify individual lower limb contributions to the specific task.² Asymmetry contributes to a measurable decline in performance of functional tasks such as STST, ambulation, and balance recovery following a misstep.^{1–4,30–33} Asymmetry in muscle function is apparent after hip fracture, with larger asymmetries being associated with injury risk, fall frequency, and mobility impairments.^6,7,34 The magnitude of asymmetry varies across different weight-bearing tasks, with evidence that more complex tasks may be more demonstrative of lower limb deficits and residual asymmetries.⁹ Although one study reports absolute lower limb power, rather than power or strength asymmetry as most critical in identifying those likely to fall,³⁵ the consensus is that asymmetry apparent during mobility tasks negatively impacts mobility and increases injurious fall risk.^1,4,36

The purpose of this study was to determine the unique contribution of weight-bearing asymmetry during an STST on physical function after hip fracture. We examined correlations between vGRF asymmetry variables during an STST, the modified Physical Performance Test (mPPT) score,³⁷ and a timed stair climb test (SCT).³⁸ We hypothesized that asymmetry in vGRF variables during an STST would provide a unique contribution to physical function beyond that identified by known contributors to physical function after controlling for covariates.

METHODS

Participants

A convenience sample of 31 community-dwelling older adults, who had recently incurred a hip fracture, participated in this study. Participants (age range: 53-90 years, mean [standard deviation], 77.7 [10.5] years) were recruited from University of Utah (UU) and Intermountain Healthcare (IH) hospitals in Salt Lake City, Utah, between January 2014 and February 2015. Eligible participants were required to independently transfer and ambulate at least 50 ft without physical assistance, have incurred a hip fracture in the last 2 to 8 months, be 50 years or older, and have minimal cognitive impairments (>23/30 Montreal Cognitive Assessment). In addition, participants had been discharged from “usual care,” typically consisting of 8 to 10 weeks of physical therapy that included balance, mobility, and strength training in skilled, home health, and outpatient facilities following hip fracture. Baseline characteristics of the sample are summarized in Table 1. Exclusion criteria included known serious medical or neurological diagnosis (eg, cancer, chronic obstructive pulmonary disease, and multiple sclerosis), visual impairments, vestibular disorders, bilateral hip fracture, significant range of motion limitations, and/or painful osteoarthritis in the hip or knee. Exclusion criteria were selected in an effort to minimize factors other than hip fracture that might contribute to asymmetries in task performance. Institutional review boards of the UU and IH both approved the study and all participants provided informed consent before enrollment (UU: IRB_00062639, IH: IRB #1040261).

Table 1. - Descriptive Characteristics of 31 Eligible Participants

Variable	Summary Statistics
	Means (SD)	Range
Age, y	77.7 (10.5)	53-90
Sex, female/male	21/10
Height, cm	166.1 (10.7)	152.4-195.6
Weight, kg	70.3 (18.0)	40.9-112.5
Body mass index, kg/m2	25.3 (5.6)	15.5-40.6
Time since fracture, m	4.1 (1.4)	2.0-6.5
Normalized peak strength, N/kg	3.1 (1.5)	0.9-6.5
Gait speed, m/s	0.9 (0.3)	0.5-1.5
ABC, points	68.7 (17.0)	3.1-6.5
LEM, points	75.1 (11.4)	47-100
AREA, N·s/kg	1.3 (0.8)	0.1-6.6
mPPT, points	25.7 (5.5)	15-35
Stair climb, s	12.3 (6.2)	3.1-26.3

Abbreviations: ABC, Activities of Balance Confidence scale; AREA, asymmetry during rising phase; LEM, lower extremity measure (self-reported function); mPPT, modified Physical Performance Test; SD, standard deviation.

Procedures

All participants completed a series of questionnaires and physical performance tests. In order to determine the vGRF during the STST, participants were tested using an instrumented chair (Figure 1). Muscle function was assessed by examining unilateral isometric strength of the knee extensors and lower extremity extension power. In order to document physical function, both performance and self-report measures were used. The mPPT and SCT were chosen to represent physical function, and the lower extremity measure (LEM) and Activities-specific Balance Confidence Scale (ABC) were used as self-report measures.

Figure 1.: Instrumented chair designed for sit-to-stand trials. This portable, instrumented chair incorporated 4 imbedded force plates to measure vertical ground reaction forces of each extremity during sit-to-stand transition.

vGRFs During STST Analysis

Participants were seated on the front half of the instrumented chair (Figure 1), with the mid-length of the thigh aligned with the edge of the chair and ankles placed in approximately 15° of dorsiflexion. The chair seat height was adjusted so that a 90° hip and 90° knee flexion angle was achieved when the participant was seated. Allowing for the use of arms to assist with the task, participants were instructed to stand up as “quickly as you safely can.” One practice trial was performed before recording data from 3 separate STST trials, allowing 30-second rest between trials.

The chair was instrumented to detect vGRFs measured under each foot, each arm, and the seat (Figure 2). Force sensors (NMB Technologies Corporation [Menibia], Chatsworth, California) mounted in 2 Wii platforms were amplified with SGA/A signal conditioners (Mantracourt Electronics Ltd, Devon, UK) and fed into a computer using a 16-bit analog to digital converter (Model: USB 1608G). Additional force sensors (Menibia, Chatsworth, California) were also mounted in each arm and on the seat to record seat-off and arm push. The arm force signals were also amplified and converted to a 16-bit signal output. During each trial, the vGRF through each force plate was recorded at a sampling rate of 1000 Hz and exported to excel using TracerDAQ 2.2.0 software (Measurement Computing, Norton, Massachusetts). During calibration, correlations of captured data regarding vGRF magnitude to known weights of each arm and footplate were high (r = 0.99).

Figure 2.: Graphical display of sit-to-stand task trial output. This is an example of a single participant trial of the sit-to-stand task with graphical depiction of the resulting vGRF output. Two phases of the sit-to-stand movement are identified: preparation phase and rising phase. The moment of seat-off, measured by the seat force plate (vGRF_seat), determines transition from preparation to rising phase. The RFD during the preparation phase was calculated as the slope from 25% to 50% of the force value at seat-off for each lower extremity. The unilateral measures of vGRF_{INvolved/UNINvolved} were determined from the left and right force plates. Asymmetry during the rising phase was calculated as AREA between the vGRF_{INvolved/UNINvolved} throughout rising phase. Note: Each trial was recorded over a 10-second duration, though depicted output is abbreviated to show task completion during initial 3 seconds. AREA, asymmetry measure during rising phase of sit-to-stand task; DRW, drop in resting weight (initiation of sit-to-stand trial); RBW, reestablishment of body weight (completion of sit-to-stand trial); RFD, rate of force development; vGRF, vertical ground reaction force.

Two phases of the STST were identified from the sum of vGRF_INVolved and vGRF_UNINVolved (vGRF_Bilateral).^2,39 The preparation phase was initiated by a 5N decrease in vGRF_Bilateral. This brief unweighting of the lower limbs is a countermovement, typically occurring just prior to the ensuing rapid lower limb loading. The end of the preparation phase occurred at seat-off, marked as the instant when vGRF_Seat was below 5N. The rising phase began at seat-off and ended when vGRF_Bilateral equaled body weight, following the first peak of vGRF_Bilateral. The STST time was measured from the beginning of the preparatory phase to the end of the rising phase (Figure 3).

Figure 3.: Sit-to-stand task performance on instrumented chair. Depicted is an individual performing sit-to-stand trial with corresponding movement during recorded trial. Typical posthip fracture vGRF output of bilateral and unilateral lower extremity contributions during the trial is presented. Red line corresponds with (left) involved lower extremity. Blue line corresponds with (right) uninvolved lower extremity. Green line corresponds with summation of vGRF output from both lower extremities. RFD is recorded by 25% to 50% of vGRF at seat-off divided by time from countermovement DRW that initiates the sit-to-stand until moment of seat-off. AREA is established as the difference in vGRFs demonstrated by the lower extremities during the trial from moment of seat-off, to RBW. Note: Arm impulse measurement data removed on this image for clarity, though individuals did use arms to assist in pushoff during sit-to-stand trial. AREA, asymmetry measure during rising phase of sit-to-stand task; DRW, drop in resting weight; RBW, reestablishment of body weight; RFD, rate of force development; vGRF, vertical ground reaction force.

To capture the vGRF developed by each limb during the preparation phase, the rate of force development (RFD) was calculated. The RFD was calculated as the slope of the vGRF data (vGRF_INVolved and vGRF_UNINVolved). The slope of the force between 25% and 50% of force at time of seat-off (end of preparation phase) was calculated for each limb separately (RFD_INVolved and RFD_UNINVolved) and summed (RFD_Bilateral). Higher slopes indicate more rapid development of force, which correlates to faster rising time.

To capture the vGRF developed by each limb during the rising phase, AREA variables were calculated, as has been done previously for STST measurements using a similar force-plate-imbedded chair.² The magnitude of the vGRF_INVolved and the vGRF_UNINVolved was calculated by obtaining the area under the curve from the beginning to the end of the rising phase (AREA_INVolved, AREA_UNINVolved, and summed AREA_Bilateral). A higher area value occurs from either a longer rising period or higher force amplitude over the rising phase. Lower area values are the result of shorter rising periods or lower force amplitudes over the rising phase. An AREA score was calculated as the difference between AREA_INVolved and AREA_UNINVolved to indicate the difference in contribution of each limb to rising. Higher AREA during rising phase suggests lower symmetry or greater reliance on one limb (typically the nonsurgical limb). Good reliability has been previously established (0.84–0.91) for vGRF variables identified during STST performance among older adults who have recently incurred a hip fracture.⁴⁰

The average of 3 STST trials normalized to body mass were recorded for RFD and AREA to represent STST performance during preparatory and rising phases, respectively. AREA was the asymmetry measure captured for further analysis, as this occurs during the longest phase (rising), and is the portion of STST that places the individual most at risk for balance loss, due to rapid lower limb loading during transition from sit-to-stand, while the seat no longer contacts the chair.

Muscle Function

An isokinetic dynamometer (KinCom, Chattanooga Inc, TN) was used to determine unilateral knee extension strength. Participants were positioned with their hips at 90° and knee at 60° of flexion. A maximal voluntary isometric contraction of the knee extensors was recorded in newton (N). The average of 3 trials (with 30-second rest between trials) normalized to body mass were used for analysis. This method has excellent reliability (0.81–0.98).⁴¹ Extension power of each lower limb was unilaterally measured on a Nottingham power rig (Medical Engineering Unit, Nottingham, UK), and recorded in watts (W). Participants were seated in an upright position with arms folded across their chests. The seat was adjusted until comfortable extension of the knee with full depression of the foot pedal was reached. Participants were instructed to depress the foot pedal as hard and fast as possible. After 3 warm-up trials at 50%, 75%, and 100% effort, 6 trials were performed and the average of the 3 highest trials, normalized to body mass, were used for analysis. The lower limb extension power rig is a valid, reliable, and feasible means of assessing muscle power across the lifespan in both sexes.⁴²

Physical Function

Usual gait speed was measured over a 15.24-m distance. Participants were instructed to “walk at your normal daily pace.” The average score from 2 trials was used for analysis. Gait speed is a quick, inexpensive, reliable measure of mobility with established predictive value for major health-related outcomes among older adults.^43,44 The ABC scale, a 16-item, validated, reliable self-report scale, was used to determine balance confidence. Possible scores range from 0 to 100. Scores below 67 indicate high risk of falling,⁴⁵ with fall prevalence twice that of individuals with balance confidence scores reaching 82.⁴⁶ The LEM, a reliable, valid, and responsive 29-item self-report scale,⁴⁷ was used to determine perceived postfracture mobility and performance. Possible scores range from 0 to 100. Scores of 75 to 85 indicate moderate limitations in mobility, whereas scores above 85 indicate normal mobility.⁴⁷

The mPPT is a composite 9-item standardized assessment of physical function, mimics activities of daily living, and includes various mobility measures such as standing balance, putting on and removing a jacket, picking up an object from the floor, walking, and stair climbing.^48,49 This composite test of physical function yields possible scores of 0 to 36 and categorizes frailty, with a score of 17 to 24 indicating moderate frailty, 25 to 31 considered mildly frail, and 32 to 36 indicating no frailty.³⁷ Per protocol, 2 trials were performed and the average score for each measure was used to determine the overall mPPT score. The compilation of scored tasks combined to calculate total mPPT score. SCT performance, a component of mPPT, was thus scored categorically (0-3). The SCT was also measured to represent physical function. Nonstandardized methods of applying this test (eg, varying number and height of steps and inconsistent arm rail usage) have resulted in a lack of normative data in an older population; yet the SCT has good construct validity, and is highly reliable.^38,50 The SCT is a clinically relevant measure of lower limb power impairments⁵¹ that is meaningfully associated with mobility performance,^50,52 strength,⁵³ independence,⁵³ and self-report of physical function,⁵⁰ and thus suitable for clinical settings in which impairment-mobility relationships are of interest. One practice trial, followed by 3 recordings of timed SCT performance were collected and averaged to determine SCT time. Participants were allowed to use stair handrail if desired and were instructed to climb the twelve 7-inch stairs as quickly as they safely could upon the command, “Ready... Go!”After returning to base of stairs, 30-second rest was given between each trial.

Data Analysis

Data management and statistical analyses were performed with SPSS statistical software (SPSS Version 22.0, Armonk, New York). Descriptive data were calculated for demographic variables and dependent measures and are presented as means (standard deviation). The relationships between selected demographic, clinical, and vGRF variables and the physical function (mPPT, SCT) variables were examined with univariate and multivariate analyses. Pearson (or Spearman's ρ) correlation coefficients were calculated to determine the relationship for continuous variables, whereas a point-biserial coefficient was used for the dichotomous variable of sex. All variables with a significant univariate correlation 0.35 or more were retained for further analysis. Variables with correlations less than r = 0.35 were not further analyzed as the low correlation suggested the limited utility of these variables as significant contributors to predict physical function. The relative contribution of each vGRF asymmetry variable to explaining variability in the physical function outcomes was examined using hierarchical linear regression, after controlling for covariates that have been reported to be related to physical function and exceeded our entry criterion. Each of the vGRF variables derived from STST trials as well as the muscle function variables (strength and power) was normalized to body mass (kg). Criterion for entry to the model was a significance level of P < .05. For each variable entered into the final model, the part correlation was examined to determine the unique amount of variance in the physical function outcome that was accounted for by the variable.

RESULTS

Participant characteristics are presented in Table 1. The sample is typical of older adults after sustaining a hip fracture, demonstrating persistent functional deficits relative to healthy community-dwelling older adults, despite having been discharged from usual care.

The bivariate correlations of demographic variables with physical performance variables (mPPT, SCT) revealed age to be the single demographic variable with a significant moderate correlation (r =−0.43, 0.40, respectively, P < .05) (Table 2). The bivariate correlations of other clinical measures expected to explain physical function showed significant moderate to strong correlations with both mPPT score (r range =−0.47 to 0.86) and SCT score (r range =−0.47 to 0.83) (Table 2).⁵⁴ The direction of the correlations indicates that older age is associated with lower physical function (mPPT score) whereas faster gait speed, higher balance confidence, muscle strength, and self-reported function are associated with higher physical function (mPPT score). Older age is also associated with slower SCT time, whereas faster gait speed, higher balance confidence, muscle strength, and self-reported function are associated with faster SCT performance. RFD, the initial pushoff from the chair during the STST, did not show significant correlation with physical function variables and was not included in further analysis.

Table 2. - Bivariate Correlations (95% Confidence Intervals) Between Selected Variables Expected to Influence Function and Measured Physical Function^a

Variable	mPPT	SCT
Age, y	−0.43b (−0.68 to −0.09)	0.40 (0.05 to 0.66)
Sex, male/female	0.33 (−0.03 to 0.61)	−0.28 (−0.58 to 0.08)
Body mass index, kg/m2	0.05 (−0.31 to 0.40)	−0.09 (−0.43 to 0.27)
Gait speed, s	0.86 (0.73 to 0.93)	−0.83 (−0.92 to −0.68)
ABC, points	0.77 (0.57 to 0.88)	−0.65 (−0.82 to −0.39)
Peak strengthINV, N/kg	0.55 (0.25 to 0.76)	−0.53 (−0.74 to −0.22)
LEM, points	0.55 (0.25 to 0.76)	−0.47 (−0.71 to −0.14)
AREA, N·s/kg	−0.42 (−0.67 to −0.08)	0.57 (0.27 to 0.77)

Abbreviations: ABC, Activities-specific Balance Confidence Scale; AREA, asymmetry during rising phase of sit-to-stand task; LEM, lower extremity measure; mPPT, modified Physical Performance Test; SCT, stair climb test.

^aAll variables listed were considered for inclusion in the regression model(s). Significant correlations were retained in further analysis.

^bItalic indicates significant correlations (95% confidence intervals). Values are represented as Pearson correlation coefficients, except sex presented as point-biserial correlation, AREA as Spearman's ρ.

The regression analysis with mPPT score as the dependent variable revealed that the explanators as a group accounted for 83.4% of the variance in mPPT score, with ABC (P < .001), Gait Speed (GS) (P < .001), and LEM (P = .05) each significantly contributing to the final model. The part correlation of ABC was 0.32, of GS was 0.41, of LEM was 0.20, indicating that ABC, GS, and LEM explained 10.4%, 16.6%, and 4.0% of the variance in the mPPT score, with all other model variables held constant (Table 3).

Table 3. - Results of Hierarchical Regression Examining Association Between Asymmetry in Sit-to-Stand Task Performance and Physical Function

Variable	Regression Coefficient (95% CI)	Standardized Coefficient (β)	P Value	Part Correlation	R 2
mPPT, points					83.4
Age, y	0.02 (−0.08 to 0.12)	0.04	.69	0.03
ABC, points	0.20 (0.10 to 0.29)	0.61a	<.001	0.32
Gait speed, s	11.47 (7.15 to 15.18)	0.64	<.001	0.41
Strength, N/kg	0.06 (−0.64 to 0.76)	−0.02	0.86	0.01
LEM, points	−0.17 (−0.30 to −0.04)	0.06	<.05	−0.20
AREA, N·s/kg	−0.46 (−1.23 to 0.32)	−0.11	0.23	−0.09
Stair climb test, s					78.0
Age, y	−0.03 (−0.15 to 0.10)	−0.05	0.65	−0.04
ABC, points	−0.14 (−0.26 to −0.01)	−0.37	<.05	−0.20
Gait speed, s	−14.1(−19.64 to −8.40)	−0.69	<.001	−0.44
Strength, N/kg	0.04 (−0.88 to 0.95)	0.01	0.94	0.01
LEM, points	0.17 (0.01 to 0.34)	0.32	<.05	0.18
AREA, N·s/kg	1.52 (0.51 to 2.53)	0.31	<.005	0.27

Abbreviations: ABC, Activities-specific Balance Confidence Scale; AREA, asymmetry measure during rising phase of sit-to-stand task; CI, confidence interval; LEM, lower extremity measure; mPPT, modified Physical Performance Test.

^aItalic indicates variable is a significant predictor (P < .05) in the regression model.

The regression analysis with SCT score as the dependent variable revealed that the predictors as a group accounted for 78.0% of the variance in SCT score, with ABC (P = .03), GS (P < .001), LEM (P = .04), and AREA (P = .006) each contributing to the final model (P < .001). The part correlation of ABC was −0.20, of GS was −0.44, of LEM was 0.18, of AREA was 0.27, indicating that ABC, GS, LEM, and AREA explained 3.8%, 19.4%, 3.4%, and 7.1% of the variance in the SCT score, respectively (Table 3).

DISCUSSION

After accounting for the expected contributors to physical function following hip fracture, asymmetry during the performance of an STST emerged as a significant explanator for SCT performance. Specifically, age, balance confidence, gait speed, normalized muscle strength, and self-reported function were each tested to determine contributions of these variables to physical function following hip fracture. Interestingly, asymmetry was not a significant explanator for a composite physical function score (mPPT). Although others have suggested a relationship between asymmetry and physical performance, this is the first study to identify the unique and shared contribution of identified asymmetry during an STST on stair-climbing performance after hip fracture.

Relationships between asymmetry and physical function have been described with asymmetry during an STST showing moderate to high correlations with gait speed, balance, and self-reported function in the year following hip fracture.^1,2 Asymmetry in muscle function is common among independent community-dwelling women older than 65 years, and large asymmetries in leg extension power have been linked to falls.⁶ This is clinically relevant to the physical therapist as over 50% experience a fall within 6 months of hospital discharge after hip fracture.^19,55 Mobility limitations are more prevalent among individuals with higher asymmetry,⁴ and Portegijs et al⁴ identified that large asymmetries in power correlate with slower stair climb at week 1 and week 13 after hip fracture.

Climbing stairs is among the most challenging tasks of daily living for older individuals. Stair ambulation requires as much as 3-fold greater peak knee extensor strength than level walking³⁸ and also necessitates coordinated unilateral limb contributions. Demanding nearly 90% of maximum capacity for many, compared to 40% for younger, nonimpaired adults,⁵⁶ there may be little strength reserve to cope with unexpected challenges, contributing to a high fall risk during stair ambulation among older adults who are frail. Falls on stairs are the leading cause of accidental death, contributing to over 10% of all fatal falls among individuals older than 65 years,³⁸ a large number considering adults spend only a small fraction of their day performing stair ambulation.³⁸

Asymmetry did not surface as a significant explanator for physical function as defined by the mPPT score. This was an unexpected finding, as the mPPT score provides an objective, valid, reliable, and responsive measure of physical function.^37,48,57 In this composite physical performance test, climbing stairs has been identified as the most difficult single item.⁴⁸ Many of the simpler tasks of the mPPT (eg, donning/doffing jacket, reaching to a shelf, and static balance measurements) do not unilaterally challenge individuals to the extent that SCT does providing a potential explanation for the inability of asymmetry during STST performance to predict mPPT score. Clinically, this provides support for the physical therapist to more closely examine individual limb muscle contributions rather than typical whole body assessments prior to discharge.

Gait speed, ABC score, and LEM score all emerged as significant explanators of physical function. This was expected, as each has been associated with mobility and physical performance among older adults following hip fracture.^{32,47,57–61} Although normalized strength revealed a significant bivariate relationship with both measures of physical function, it did not remain significant in the regression analysis. This was surprising considering the strength requirements of the SCT. However, strength and power are curvilinear.⁶² It may be that a larger sample with more heterogeneity of function may be necessary to fully understand the relationship between strength and function. Frail older adults, for example, require a higher relative percentage of their maximum capacity collectively, and from each limb in order to accomplish a specified task.

Despite the ability of variables included in our model to explain 70% of the variance in stair climbing prior to the inclusion of an asymmetry measure, AREA uniquely explained over 7% of the SCT performance. As anticipated, GS explained the largest portion of variability in timed stair climb performance. The identified asymmetry provided a higher unique contribution than either self-reported function (3.4%) or balance confidence (4%). This suggests that asymmetry may be important to challenging tasks that require unilateral limb contributions that may predispose one to falls. Examples of this include stair ascent and descent, stepping off a curb, walking on uneven surfaces, and recovering balance after a perturbation or misstep.

Although gait speed explains a large percentage of the variability in stair climb capacity, lower extremity asymmetry, balance confidence, and self-report of function also contribute. Rehabilitation including a multicomponent approach to recovery may be most effective in this population. Improving lower extremity symmetry during functional movement may contribute to improving high-level mobility after fracture, thus rehabilitation strategies that address this aspect of recovery should be explored.

The results of the study should be considered in light of some limitations. Our convenience sample of older adults recovering from hip fracture, by virtue of their interest and ability to volunteer for this study, may have had better physical function than other postfracture subpopulations. Our sample showed limb symmetry in the STST to be approximately 0.77, whereas larger asymmetries (>0.70) have been reported.^1–3 Each participant in our sample was able to climb stairs (with use of a handrail) without assistance. Reports suggest that not all can perform this task at 3 to 6 months after fracture.^49,63 Performance capacity should be considered when generalizing our results. However, there is the possibility that a sample with lower physical ability may demonstrate an even more dramatic influence of asymmetry on physical function. A strong relationship (r > 0.70) between mPPT score and asymmetry that we noted in a subgroup of 12 participants scoring less than 24 on the mPPT lends support to this notion. Finally, although we identified the unique contribution of asymmetry of the STST on SCT performance, this study was underpowered to thoroughly explain the variance in mPPT score and SCT score from predictor variables and to investigate potential interactions.

CONCLUSION

Asymmetry during an STST is a significant and unique contributor to explaining SCT performance after hip fracture. With potential for improving high-level mobility, and thereby reducing fall risk, interventions that address surgical limb deficits after hip fracture may be beneficial in this vulnerable population and should be explored.