Aberrant Lumbopelvic Movements Predict Prospective Functional Decline in Older Adults with Chronic Low Back Pain

Abstract

Objective:

To investigate if clinically observable aberrant lumbopelvic movements are associated with physical function at 12-month follow-up in older adults with chronic low back pain (CLBP), both directly and indirectly through baseline physical function.

Design:

Secondary analysis of a yearlong prospective cohort study.

Setting:

Clinical Research Laboratory.

Participants:

Community-dwelling older adults with CLBP.

Interventions:

Not applicable.

Main Outcome Measures:

Data from 239 participants were analyzed. Participants were screened at baseline for aberrant lumbopelvic movements during active trunk flexion; total observable aberrant movements were recorded and summed (range 0–4). Latent constructs of physical function were developed from an array of perception-based and performance-based outcome measures at baseline and 12-months, respectively. Structural Equation Modeling was used to assess the direct effect of baseline aberrant movement score on the latent construct of 12-month physical function, and its indirect effect through baseline physical function.

Results:

Aberrant movements were present in the majority of participants (64.9%) and had a significant negative total effect on 12-month physical function (γ= −0.278, p<.001). Aberrant movement score’s direct effect and indirect effect, through baseline functioning, were significantly negatively associated with physical function at 12-months, after adjusting for covariates (γ=−0.068, p=.038; γ= −0.210, p<.001, respectively).

Conclusion:

Aberrant lumbopelvic movements are associated with decreased physical function at 12-month follow-up in older adults with CLBP, independent of baseline physical function and covariates. Future studies should evaluate if screening for aberrant movements may inform prognostic and interventional efforts in this patient population.

Low back pain (LBP) is a leading cause of disability among adults¹; given that LBP prevalence continues to rise,² this trend may continue. Older adults with chronic LBP (CLBP) are particularly susceptible to functional decline and resultant disability.^3–5 Despite susceptibility to poor health outcomes, older adults are often excluded from LBP intervention studies.⁶ To develop effective interventions, it is essential to identify potentially modifiable precursors of functional decline in older adults with CLBP.

Research suggests that adults with LBP may exhibit altered lower extremity and lumbar spine kinematics (i.e., coordination deficits) during functional tasks.^7–13 Rehabilitation professionals can screen for aberrant lumbopelvic movements during forward flexion from a standing position, which are considered to be indicative of clinical lumbar spine instability.¹⁴ Prior evidence has established that clinical observation of aberrant movements has adequate interrater reliability^{15, 16} and moderate agreement with kinematic analysis.¹⁷ Further, Biely et al.¹⁵ established the utility of an aggregate score of aberrant movements in adults with nonspecific LBP. We propose that an aberrant movement score, as a clinical index of lumbopelvic incoordination, may importantly contribute to physical function in older adults with CLBP.

Physical function is an important consideration in the management of older adults with CLBP, given that functional decline predisposes older adults to adverse health outcomes (e.g., mortality)¹⁸ while maintenance of function promotes independence in basic and instrumental activities of daily living.¹⁹ Evidence suggests that performance and perception-based measures both contribute to function^{20, 21} and downstream health outcomes such as mortality.¹⁸ Thus, use of one type of functional outcome measure may incompletely represent this broad construct. A latent variable comprised of performance and perception-based measures may better characterize the underlying construct of physical functioning.

The primary purpose of this study is to examine the longitudinal relationship between aberrant movement score and physical function at 12-months in older adults with CLBP. Secondarily, we sought to examine if the association between aberrant movement score and 12-month physical function was mediated by baseline physical function. Our hypothesis was that aberrant movement score would be negatively associated with physical function at 12-months, both independently and by its association with baseline physical function. Better understanding of these relationships may guide rehabilitation professionals in plan of care development by informing prognosis and intervention selection, and may lead to improved outcomes for this understudied patient population.

METHODS

Study Design and Participants

We performed a secondary analysis of a prospective cohort study of older adults with CLBP that was conducted from March 2013 to December 2016.²² Participants were recruited from northern Delaware via posted fliers, newspaper advertisements, community outreach events, and visits to senior living facilities. Participants between the ages of 60–85 years were screened for non-specific CLBP (i.e., pain intensity≥3/10, pain frequency≥4 days/week, pain duration≥3 months)²³ that interfered with daily life. People were excluded if they had non-mechanical LBP symptoms (i.e., unremitting night pain), non-radicular lower extremity pain greater in severity than LBP, current fractures in the back or hips, significantly impaired mobility (i.e., using an assistive device more supportive than a cane), a diagnosis of a progressive neurological disorder, or a diagnosis of a terminal illness. With these criteria, 250 participants were enrolled out of 432 individuals that underwent screening (FIGURE 1).

FIGURE 1.

Study flow diagram

After providing written informed consent, enrolled participants underwent assessments by a licensed physical therapist (PT) at baseline and 12 months post-baseline. Assessments included clinical hip and low back examinations as well as a battery of performance-based physical function tests. Study assessments were conducted in the clinical research laboratory at the University of Delaware. The study protocol was approved by the University of Delaware’s Institutional Review Board for Human Subjects Research, and adhered to guidelines from the Helsinki Declaration.

Personal Characteristics

Demographic (i.e., age, sex, race) and anthropometric (i.e., height and weight) data were collected; body mass index (BMI) was calculated as kg/m². Participants rated their CLBP intensity with pain thermometers on a 0–10 scale (0=no pain, 10=worst pain imaginable), which has been shown to be reliable and valid in this patient population.²⁴ Composite LBP intensity was calculated by averaging each participant’s current pain rating with their best and worst pain intensity in the preceding 24 hours. Hip extension strength was measured via hand-held dynamometry using a reliable approach²⁵; the weaker of the two hips represented hip extension strength. Increased age,²⁶ female sex,²⁶ race,²⁷ elevated BMI,²⁸ LBP,²⁹ and worse lower extremity function³⁰ have been associated with functional decline. We included hip extension strength as a proxy measure of lower extremity function.

Aberrant Lumbopelvic Movement Score

Prior to physical performance testing, participants were screened for four aberrant movements during three trials of forward trunk flexion from a standing position: instability catch, painful arc, reversal of lumbopelvic rhythm, and Gower’s sign. Operational definitions for each aberrant movement are presented in TABLE 1.³¹ Presence or absence (coded as 1 or 0, respectively) of the movements was observed and recorded by a PT. Participants’ were not educated on their performance, and were not given opportunities to correct movement patterns. Aberrant movement score was calculated by summing the observations together (range of 0–4), with higher values indicating greater lumbopelvic incoordination. Literature has established that observation of aberrant movement patterns in patients with LBP has fair to excellent interrater reliability^{15, 16} and criterion validity¹⁷; further, the construct validity of an aggregated aberrant movement score has been previously established.¹⁵

TABLE 1,

Operational Definitions for Aberrant Movements

Aberrant Movement	Definition
Instability Catch	Deviation from the sagittal plane (i.e., rotation or lateral bending) or abrupt changes in the speed of movement (i.e., acceleration or deceleration) during trunk flexion and/or the return from a flexed position
Painful Arc	Onset of pain at particular ranges of motion (i.e., “arcs”) during either trunk flexion or the return from a flexed position; pain must be absent at all ranges other than the arc(s)
Reversal of Lumbopelvic Rhythm	Initiation of the return to standing from a flexed position via knee flexion and anterior shift of the pelvis
Gower’s Sign (Thigh Climbing)	Use of upper extremity assistance on the thighs or an external surface to facilitate the return to standing from a flexed position

Measures of Physical Function

Physical function was operationally defined by both perception-based and performance-based outcomes: Quebec Back Pain Disability Scale, Repeated Chair Rise Test, self-selected gait speed, Timed Up-and-Go Test, Six-Minute Walk Test, self-selected Stair Climbing Test, and the Figure-of-Eight Walk Test.

Perception-Based Function

Quebec Back Pain Disability Scale (QBPDS)

The 20-item QBPDS was utilized to measure LBP-related disability.³² Each item is scored on a scale from 0–5 (0=‘not difficult at all’; 5=‘unable to do’), with a possible score between 0–100; higher scores on the QBPDS reflect greater LBP-related disability. The QBPDS is reliable and valid among older adults with LBP.³³

Performance-Based Function

Repeated Chair Rise Test (RCR)

The RCR was conducted as in the Health ABC Performance Battery.³⁴ Participants were instructed to complete five sit to stand transitions from a standard height armchair with their arms crossed over their chest. The RCR is reliable and valid in both LBP³⁵ and community-dwelling older adult³⁶ populations, and is predictive of falls³⁷ and adverse health events such as lower extremity decline and mortality.³⁸

Self-Selected Gait Speed (SSGS)

Self-selected GS (SSGS) was captured by the electronic GaitMat II® system (EQ Inc, Chalfont, PA), which has excellent test-retest reliability.³⁹ Participants were instructed to ambulate on the 4-meter long walkway at their usual, comfortable speed; use of a cane was permissible. After three practice trials, three trials were recorded and averaged for subsequent analyses. SSGS is predictive of incident disability^{30, 40} and mortality.⁴¹

Timed Up-and-Go Test (TUG)

The TUG was conducted with a standard height armchair and a 3-meter walkway.⁴² Participants were instructed to stand up from the chair, walk 3-meters, turn around the demarcation, walk back to the chair, and sit down. Three trials were averaged for subsequent analyses. The TUG has excellent test-retest reliability and discriminant validity,⁴² and is predictive of falls in community-dwelling older adults.⁴³

Six-Minute Walk Test (6MWT)

Participants were instructed to cover as much ground as possible in six minutes on a continuous rectangular pathway.⁴⁴ One trial was completed, with standard encouragement provided to participants every 30 seconds. The 6MWT has high test-retest reliability and construct validity of physical function in older adults.⁴⁵

Stair Climbing Test (SC)

Participants were instructed to ascend and descend a standard flight of stairs (12 steps) at their usual pace;²⁶ upper extremity assist and assistive devices were permissible if necessary. One trial was conducted, and the ascent and descent task durations were summed together for subsequent analyses. Stair ascent and descent tasks are predictive of multiple falls in older adults.³⁷

Figure-of-Eight Walk Test (F8WT)

For the F8WT, cones were set up 1.52 meters apart with a line of tape designating the halfway point between the cones.⁴⁶ Participants were asked to stand with their toes on this line, and were instructed to walk quickly yet safely in a figure-of-eight pattern from this midway point. The F8WT has excellent interrater and test-retest reliability as well as validity,⁴⁶ and has been associated with falling in older adults.⁴⁷

Statistical Analysis

Of the 250 participants from the Delaware Spine Studies longitudinal cohort, 239 participants were included in this analysis (FIGURE 1). Baseline characteristics of participants with and without aberrant movements were compared with chi-square, independent t-tests and Mann-Whitney U tests. Structural Equation Modeling (SEM) was used to test our hypotheses that aberrant movement score at baseline was related to 12-month physical functioning, and that this path would be mediated by baseline functioning, while adjusting for potential confounders including demographics and personal characteristics. SEM provides the ability to analyze associations among observed variables and unobservable latent constructs, which are defined by a set observed indicators, while removing the impact of measurement error.^{48, 49} SEM has two components: a measurement component relating all indicators to their latent constructs, and a structural component modeling the relationships between the observed and latent variables. Prior literature supports the use of SEM in clinical research^{50, 51} and advocates for the particular approach used in this analysis.⁵² The analysis was conducted with MPlus software.⁵³ Statistical significance was set at α<0.05.

Physical Function

We treated physical function as a latent construct using seven validated indicators of functioning. The measurement component of our model consisted of latent variables for physical function at two time points with the indicators of QBPDS, RCR, 6MWT, SSGS, TUG, F8WT, and SC. Physical function was invariant across time, ensuring that the underlying construct of function would be equally measured on both occasions (for brevity these results are not reported). We accounted for the lack of independence of the indicators by correlating the disturbances (i.e., residuals) of the baseline indicators with their 12-month counterpart (i.e., baseline QBPDS and 12-month QBPDS).

The structural component of our model included aberrant movement score at baseline predicting physical function at baseline and at 12 months, as well as baseline physical function predicting 12-month physical function (see path diagram in FIGURE 2). Both latent physical function constructs and aberrant movement scores were adjusted for relevant covariates (age, sex, race, BMI at baseline, composite LBP at baseline, and hip extension strength at baseline). This model allows for testing of the indirect effect of aberrant movement score on 12-month physical function through its effect on baseline physical function. All SEM path coefficients (γ’s) are reported as standardized values, allowing for inferences to be made about the relative contribution of aberrant movement score to physical function at both time points.

FIGURE 2. Path diagram showing the relationship between aberrant movements and physical function.

Values in the path model are presented as standardized path coefficient (standard error), and * indicates significance with p<0.05. The solid arrow represents the significant negative association between aberrant movements and the latent construct of baseline physical function, while the dashed-dot arrow represents the significant positive association between baseline and 12-month physical function. The indirect effect of aberrant movements on the latent construct of 12-month physical function is represented by the solid and dashed-dot arrows. The double-line arrow represents the direct effect of aberrant movements on the latent construct of 12-month physical function. The total effect of aberrant movements on the latent construct of 12-month physical function is the sum of the indirect and direct effects. Abbreviations QBPDS = Quebec Back Pain Disability Scale; RCR = Repeated Chair Rise Test; 6MWT = Six-Minute Walk Test; SSGS = self-selected gait speed; TUG = Timed Up-and-Go Test; F8WT = Figure-of-Eight Walk Test; SC = Stair Climbing Test; 0 = Baseline; 12 = 12-months.

RESULTS

Participant characteristics (n=239) are detailed in TABLE 2. Aberrant movements were observed in 64.9% of participants, with an average aberrant movement score of 0.97 (±0.90). Participants with aberrant movements had significantly worse baseline measures of performance and perception-based function (all p<0.05). There was no multicollinearity among covariates and aberrant movements, as the magnitude of the correlations were all less than 0.396.

TABLE 2,

Baseline Participant Characteristics

Variable	Mean±SD or %
	No Aberrant Movements (n=85)	Aberrant Movements (n=154)
Age (y)	68.0±5.9	70.6±7.1*‡
Sex (female)	53.6%	48.7%
Race (Caucasian)	85.7%	86.9%
BMI (kg/m2)	29.5±5.8	29.2±5.4
Composite LBP (0–10)	2.9±1.3	3.0±1.4
Hip Extension Strength (kg)	6.6±3.1	6.2±3.7
QBPDS (0–100%)	24.6±16.5	28.9±15.8*‡
RCR (s)	13.5±10.4	14.5±7.2*†
SSGS (m/s)	1.06±0.18	1.00±0.24*‡
TUG (s)	9.5±1.9	10.7±2.9*†
6MWT (m)	519.7±103.9	485.6±130.7*
SC (s)	14.9±4.1	17.7±7.8*†
F8WT (s)	6.6±1.5	7.3±2.2*‡
Aberrant Movement Count
1		59.7%
2		33.1%
3		5.2%
4		1.9%
Aberrant Movement Type
Instability Catch		63.6%
Painful Arc		55.2%
Reversal of Lumbopelvic Rhythm		19.5%
Gower’s Sign		11.0%

Abbreviations: BMI = body mass index; LBP = low back pain; QBPDS = Quebec Back Pain Disability Scale; RCR = Repeated Chair Rise Test; SSGS = self-selected gait speed; TUG = Timed Up-and-Go Test; 6MWT = Six-Minute Walk Test; SC = Stair Climbing Test; F8WT = Figure-of-Eight Walk Test; SD = standard deviation

^{* =}p<0.05;

^{‡ =}equal variances not assumed;

^{† =}Mann-Whitney U Test

The structural equation model consisted of 58 parameters; fit indices overall indicated good fit (CFI = .915, TLI =.899, SRMR =.068, and RMSEA = .086; see TABLE 3 for reference values⁴⁹) but with a significant X²(15) =442.55 (p<.001). All indicator variables were significant at both time points (p’s < .001), and accounted for a significant amount of variance in baseline (R² = 0.479, p<.001) and 12-month (R² = 0.862, p<.001) physical functioning. All indicators were significantly correlated with their respective repeated measure (all p’s ≤ .002; see Supplementary Table 1). Supplementary Table 2 contains results for the covariates.

TABLE 3:

Structural equation modelling output, model fit indices, and R² values for latent function variables

Model Fit Indices
	Values	Reference Values
CFI	0.915	>0.90=acceptable; >0.95=good
TLI	0.899	>0.90=acceptable; >0.95=good
SRMR	0.068	<0.08=good
RMSEA	0.086	<0.10=acceptable; <0.08=good

Abbreviations: CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; SRMR = Standardized Root Mean Square Residual; RMSEA = Root Mean Squared Error of Approximation

The relationship of aberrant movement score with latent physical function constructs while controlling for covariates and baseline physical function is displayed in FIGURE 2. The total effect of baseline aberrant movement score on 12-month physical function was significant (γ= −0.278, p<.001), such that more pronounced lumbopelvic incoordination at baseline is associated with worse physical function at 1-year. Baseline aberrant movement score was significantly negatively associated with baseline physical function (γ=−0.240, p<.001) and 12-month physical function (γ=−0.068, p=.038). Baseline physical function was significantly positively associated with 12-month physical function (γ=.0.874, p<.001). The indirect effect of aberrant movement score on 12-month physical function was also statistically significant (γ= −0.210, p<.001).

DISCUSSION

After controlling for covariates, aberrant movement score had a significant negative association with the construct of latent physical function at 12-months. Further, aberrant movement score had an indirect negative association with 12-month physical function through the construct of baseline physical function. Our hypotheses were supported at both time points, establishing clinically observable aberrant movements as correlates and potential drivers of physical function in older adults with CLBP.

The negative total effect between aberrant movement score and 12-month physical function suggests that lumbopelvic incoordination may contribute to functional decline among older adults with CLBP. Of the variables included in the structural component of the model, only aberrant movement score was significantly associated with physical function constructs at both baseline and 12-month follow-up. Compared to other longitudinal coefficients, the negative direct effect between aberrant movement score and 12-month physical function ranks as the third strongest independent association. This direct effect is smaller in magnitude than the negative indirect effect of aberrant movement score on 12-month physical function, through baseline physical function. From a clinical sense, these results suggest that one way to mitigate the process of functional decline in this patient population may be to intervene upon aberrant movements; future studies need to address this research question.

The majority of our cohort demonstrated aberrant movements (64.9%); this prevalence is similar to middle-aged adults with LBP (59.3%),⁵⁴ and indicates that this impairment may be common among older adults with CLBP. Greater aberrant movement scores may be the consequence of underlying impairments that are prevalent among adults with CLBP, such as fat infiltration of lumbar multifidi,^{29, 55, 56} reduced corticomotor excitability of lumbar multifidi,⁵⁷ and decreased multifidi and erector spinae muscle function.⁵⁸ Multifidi muscles function as segmental spinal stabilizers and are capable of producing the most force in a flexed spinal posture⁵⁹; thus, multifidi impairments may manifest as lumbopelvic incoordination during a forward bending task. Furthermore, morphological and functional trunk impairments have been found to negatively affect balance in older adults with and without LBP;^{29, 60, 61} this evidence provides a plausible context from which lumbopelvic incoordination could decrease movement efficiency and result in poorer performance during lower extremity functional tasks. Though our study did not address whether lumbopelvic incoordination is a cause or effect of LBP, the potential link between underlying trunk impairments, aberrant movement scores, and decreased physical performance provides a rational framework in which aberrant movements are one determinant of functional decline in older adults with CLBP.

Aberrant movement scores may additionally carry prognostic implications, given that our indicators of physical function have prospective associations with fall risk,^{37, 43, 47} lower extremity decline and incident disability,^{30, 38, 40} and mortality^{38, 41}. No studies to date have investigated whether aberrant movement scores are associated with adverse health events; however, our results complement prior literature that has established connections between trunk impairments and falling in older adults.⁶² Future investigations are needed to elucidate whether aberrant movement scores are associated with adverse health events, such as falling, in this patient population.

Biomechanical,⁶³ electromyograhpical^{64, 65} and ultrasonographical⁶⁶ evidence altogether indicate that lumbopelvic incoordination and multifidi dysfunction may persist after resolution of LBP symptoms, suggesting a potential need for intervention. Aberrant movements are one component of a prediction rule for successful response to an exercise program that incorporates submaximal contractions of muscles that stabilize the spine: transversus abdominus, erector spinae, multifidi, quadratus lumborum, and oblique abdominals.⁵⁴ This exercise approach aligns with biomechanical evidence, which suggests that functional spinal stability is achieved by overlapping contributions from multiple muscle groups.⁶⁷ Thus, the exercise regimen from Hicks et al.⁵⁴ may modify aberrant movements by encouraging muscle activation patterns that enhance dynamic spinal stability. Future studies should evaluate if aberrant movements are amenable to stabilization interventions, and address whether this modifiability is temporally related to improvements in physical function in older adults with CLBP.

Our latent constructs of physical function were stable over time, and each indicator significantly contributed to baseline and 12-month physical functioning. These results substantiate our analytic strategy, and support the notion that physical function is comprised of complementary yet distinct components. If possible, future studies should endeavor to examine physical function as a latent construct with many measures; such a practice may unveil which clinical correlates are most strongly associated with overarching physical function.

Our study has appreciable strengths, including the longitudinal design, sample size, and robust and well-fitting analytical approach.⁴⁹ In addition to model significance and agreeable fit indices, the consistent directionality of variable coefficients reinforces the merit of our analyses and results. Our findings are clinically applicable and generalizable, given that our latent physical function construct encapsulates both performance-based and perception-based measures of function.

Study Limitations

Our results should be viewed in light of several limitations. The study is observational, therefore the reported associations between aberrant movements and decreased physical function do not amount to causality. Nonetheless, the intent of our study was to assess whether or not aberrant movements were longitudinally associated with physical function in this population; future research could investigate whether resolution of aberrant movements attenuates or prevents declines in physical function. Another limitation is that there are many other known prognostic factors for LBP that could have been included as covariates in our analysis. In addition, our findings are specific to tertiary prevention and may not be applicable to primary or secondary prevention efforts. Finally, our cohort was recruited for study participation and may not be directly comparable to patients seeking medical care for LBP.

CONCLUSIONS

Baseline aberrant lumbopelvic movement score is significantly associated with decreased physical function at 12 months in older adults with CLBP; aberrant movements appear to be one determinant of functional decline in this patient population. Future investigations are required to determine whether screening for aberrant movements aid in the prognosis of and treatment for older adults with CLBP.

HIGHLIGHTS.

• Aberrant movements had an independent negative effect on 12-month physical function

• Aberrant movements had an indirect effect on 12-month function through baseline function

• Aberrant movements appear to be predictive of functional decline in this population

Abbreviations:

BMI

body mass index

CLBP

chronic low back pain

CPR

clinical prediction rule

F8WT

Figure-of-Eight Walk Test

LBP

low back pain

NMES

neuromuscular electrical stimulation

PT

physical therapist

QBPDS

Quebec Back Pain Disability Scale

RCR

Repeated Chair Rise Test

SSGS

self-selected gait speed

6MWT

Six Minute Walk Test

SC

Stair Climbing Test

SEM

Structural Equation Modeling

TUG

Timed Up-and-Go Test