Patients with low back pain use stiffening strategy to compensate for movement control during active prone hip rotation: A cross-sectional study

Abstract

BACKGROUND:

New motor adaptation to pain theory suggests that patients with low back pain (LBP) use the lumbopelvic stiffening strategy by redistribution of within and between muscle activities to protect painful structure. This could result in an altered postural control of the lumbopelvic region during active prone hip rotation (PHR).

OBJECTIVE:

To investigate coordination and timing of lumbopelvic and hip movements, and smoothness of the lumbopelvic control during PHR between participants with and without LBP.

METHODS:

Eight participants with LBP and eight participants without LBP were recruited. The electromagnetic tracking system was used to record kinematic data during PHR. Cross-correlation between hip rotation and lumbopelvic movement in the transverse plane was calculated. Correlation at zero time-lag, time-lag, correlation at time-lag, and maximal lumbopelvic motion were derived. Frequency of movement disruption was identified. An independent $t$ -test was used in conjunction with the effect size and 95% minimal detectable difference (MDD ${}_{95}$ ) to determine the difference in kinematic parameters.

RESULTS:

Participants with LBP demonstrated a significant delay (exceeding MDD ${}_{95}$ ) in lumbopelvic motion while nonsignificant frequency of disrupted motion on the painful side PHR demonstrated a trend with a large effect size that exceeded MDD ${}_{95}$ . There were trends with moderate to large effect sizes and differences exceeding MDD ${}_{95}$ in delay of lumbopelvic motion with greater movement disruption on the nonpainful side in participants with LBP.

CONCLUSION:

Participants with LBP used a lumbopelvic stiffening strategy for postural control to protect painful structures; however, the stiffening might complicate efforts to smoothly control lumbopelvic movement.

Keywords

Prone hip rotation low back pain motion analysis motor adaptation

1. Background

There is a generally accepted relationship between pain and altered lumbopelvic movement [1]. This could be driven by motor adaptation to pain. However, conflicting evidence exists regarding stereotypical increases or decreases in trunk muscle activity impacting the control of functional movements [2]. Recently, Hodges and Tucker have proposed the “motor adaptation to pain” theory [2]. Unlike conventional theories in which patients with pain would be expected to respond in one way, this theory suggests that motor adaptations to pain are primarily a redistribution of activity within and between muscles designed to enhance protection to painful structures [2, 3, 4]. This redistribution of muscle activity could present as decreased movement and variability, increased movements, or altered segment coordination.

Typical lumbopelvic motion is characterized by a gradual increase and decrease in range of motion [5, 6]. A sudden disruption of motion during a task has been proposed as one indicator of poor neuromuscular control [5, 6]. Several studies demonstrated that patients with low back pain (LBP) use a stiffening strategy for lumbopelvic control during functional movements (e.g., standing, walking, bending) [7, 8, 9]. This strategy could result in an observed inability to separate motion of adjacent body segments or in a delay of lumbopelvic motion during multi-segmental tasks. Thus, the coordination and timing between limb and lumbopelvic movement could be used to identify a lumbopelvic stiffening strategy [10], while frequency of movement disruption (sudden change in lumbopelvic movement direction) could represent impaired control of movement [5, 6].

Although some studies found that patients with LBP use a stiffening strategy, other studies demonstrated an early and excessive lumbopelvic motion during lower limb movements [11, 12, 13]. Based on movement system impairment, excessive lumbopelvic motion during lower limb movements can cause accumulation of soft tissue stress leading to microtrauma and eventually LBP symptoms [12, 13, 14]. Studies found early lumbopelvic motion during prone hip rotation in active persons with chronic LBP [12, 13]. The researchers suggested that this early lumbopelvic motion consistently repeated during daily activities in the same direction would eventually cause excessive lumbopelvic motion [14]. The underlying factors associated with different lumbopelvic movements are still unknown. Therefore, these conflicting theories need to be further investigated.

Prone hip rotation (PHR) is one movement test clinicians use to assess lumbopelvic control during active lower extremity motion [11, 13, 14, 15]. It is believed that excessive lumbopelvic movement during lower extremity motion can increase stress to the soft tissue surrounding the lumbar spine leading to LBP [14]. This underlying mechanism is a well-accepted theory that focuses on the origin of the pain caused by repetitive movements associated with a preferred movement strategy that, when utilized throughout the day, leads to neuromuscular adaptation (e.g., increased peak lumbopelvic rotation and earlier onset of lumbopelvic motion) [12, 13, 15]. A focus on greater and an earlier onset of lumbopelvic motion does not consider an altered presentation of delayed onset of lumbopelvic motion and reduced smoothness of the motion, which could signal the use of a stiffening strategy to protect painful structures.

During PHR, clinicians observe the coordination between hip rotation and lumbopelvic movement for signs of maintaining dynamic stability, as well as for movement disruption of the lumbopelvic segment control while rotating the hip [14]. Based on the “motor adaptation to pain” theory [2] and the reported stiffening strategy adopted by some individuals with LBP, clinicians could also see delayed lumbopelvic movement during PHR, and disruption during movement caused by inadequate dynamic compensation. However, research evidence to support this presentation and underlying mechanism is limited. This may be in part because prior research focused on active indviduals with chronic LBP that reported low levels of current pain and disability [12]. It may be that prior studied cohorts of LBP did not adopt a stiffening strategy given the longetivity of symptoms, low levels of pain and their reported activity level [12, 13].

The objective of this study was to determine differences in quantity and quality of lumbopelvic rotation in the transverse plane between participants with and without subacute LBP during PHR. We hypothesized that participants with subacute LBP and higher levels of pain intensity would demonstrate delayed and reduced lumbopelvic motion (quantity) and increased frequency of movement disruption (quality) during active PHR.

Figure 1.

Electromagnetic sensor location and prone hip rotation task starting from prone with 90 degrees knee flexion. To complete one repetition, the participant performed external hip rotation followed by internal hip rotation, and returned to the starting position.

2. Methods

2.1 Participants

Eight participants with subacute ( ${<}$ 3 months) unilateral LBP, and eight age-, sex-, and BMI-matched participants without LBP were recruited for this study. Participants with LBP were included if they were between 20 and 40 years of age, were in a current episode of LBP that caused them to seek healthcare, and their back pain was mechanical in nature (associated with movement) and not associated with a specific pathology (fracture, infection) based on behavior of symptoms during the history taking [16]. Inclusion criteria for participants without LBP were age between 20 and 40 years, no episode of LBP prior to 3 months of participating in the study, and no regular exercise routine focusing on trunk muscles. We selected this age range to minimize the effect of degenerative diseases, such as degenerative disc or spondylosis on spine mechanics. Exclusion criteria for both groups included clinical signs of a systematic disease, neurological signs, previous spinal surgery, and any lower extremity condition that would potentially change lumbopelvic and lower extremity movements. Participants provided a written informed consent prior to data collection. This cross-sectional study was part of an intervention study that had a pre-specified sample size of eight participants with LBP and eight participants without LBP; thus, we did not perform a sample size calculation. However, effect size and the required sample size were calculated for future replications of this study.

2.2 Instruments and measures

Kinematic data were collected by using an electromagnetic tracking system (EMT; 3D Guidance trakSTAR, Ascension Technology Corp., Vermont, USA). Electromagnetic sensors were attached to 1) lumbar spinous process of L1 (lumbar sensor), 2) sacral spinous process of S2 (pelvic sensor), 3) right lateral malleolus (right leg sensor), and 4) left lateral malleolus (left leg sensor) (Fig. 1) [17]. Kinematic data (linear and angular displacement) from each electromagnetic sensor were simultaneously collected at 100 Hz through a custom LabVIEW program (LabVIEW program version 2012, National Instruments Corp., Texas, USA).

Based on our electromagnetic tracking system specification, maximal distance from the sensors to the source was 1 meter (1000 millimeters) and the resolution stated by the manufacturer was 0.003 inches (0.08 millimeters). Therefore, the system was able to detect 0.005 degrees (arctan 0.08/1000) at the 1-meter distance. We used two decimal points as our angle precision. Therefore, the system should be able to detect lumbopelvic rotation in the transverse plane.

2.3 Procedure

The study protocol was approved by the university’s institutional review board (COA No. 2015/050.3004). Data were collected from August 2016 to January 2017. Demographic data including age, sex, and BMI were recorded for all participants. The numeric pain rating scale (0–10) was used to assess back pain intensity for participants with LBP. Pain ratings were taken for current pain (at the time of the test) and the range of the participant’s pain as the highest and lowest intensity experienced within the last 24-hours. The Oswestry Disability Index was also collected for participants with LBP. Then, EMT sensors were mounted on thermoplastic material and attached to the participant’s body landmarks [6, 17].

The participant was placed in prone position with 90 degrees of knee flexion on a treatment table with a face hole. To establish the range of movement that the participant was to perform, the researcher passively rotated the participant’s leg into hip external and internal rotation in transverse plane until the pelvis started to rotate with the hip. The researcher performed passive hip external and internal rotation three times within this range to familiarize the participant with the range of interest. This range was selected because lumbopelvic movement within this range relies primarily on neuromuscular control to adequately stabilize the lumbopelvic segment while the hip rotates [6, 12, 13, 18]. In addition, several studies have demonstrated that altered trunk neuromuscular control can be clinically observed during the early phase of the movements prior to the end of hip motion, which engages passive hip structures (e.g., joint capsule, ligaments) attached to the pelvis thus rotating the lumbopelvic segment [6, 12, 13, 18].

The researcher placed both hands at the passive hip rotation motion boundary and, to familiarize the participants with the task, asked them to perform one repetition of active prone hip rotation that touched each hand. Visual inspection was performed to ensure that the participants moved through the same range as passive hip rotation. For consistency, participants were positioned prone with 90 degrees of knee flexion and hip in neutral position. This position was arbitrarily set as zero reference. Then, the researcher asked the participant to perform two sets of three consecutive repetitions of active PHR in this range without any specific instruction relative to the lumbopelvic region (Fig. 1). The participant was asked to pause at the neutral position for one second between repetitions, and keep the knee in 90 degrees throughout the task to maintain the lever arm length. Data were averaged across the six repetitions to stabilize the measurement. Additionally, the researcher closely monitored the performance to ensure that the participant correctly performed the task.

2.4 Data analysis

Data analysis was performed using a custom LabVIEW program. Kinematic data were converted to lumbopelvic transverse plane motion (pelvic sensor with respect to lumbar sensor as a local reference frame) and hip rotation during PHR (right and left lateral malleolus sensors with respect to the source as a global reference frame). Data were filtered using a dual-pass Butterworth filter (second order low pass frequency at 5 Hz) [6]. The zero reference was used to separate the time-series into three separate repetitions. Pelvic and malleolus sensors that rotated in the same direction as hip external rotation were considered positive values. Peak hip external and internal rotations (HER and HIR, respectively) and peak lumbopelvic motion during hip external and internal rotations (LPER and LPIR, respectively) were identified. Hip rotation velocity was also calculated. Time series of hip rotation and lumbopelvic motion in the transverse plane were plotted (Fig. 2). Cross-validation was used to determine correlation at zero time-lag (r ${}_{0}$ ), time-lag at maximum curve correlation (t), and correlation value at time-lag (r ${}_{\text{t}}$ ) [10]. The r ${}_{0}$ and r ${}_{\text{t}}$ represent movement coordination between hip and lumbopelvic rotation motions. A value close to zero can be interpreted as lumbopelvic and hip segments are moving in a dissociated manner, whereas a value close to one represents both segments are moving in a coordinated manner. The $t$ value represents movement latency between hip and lumbopelvic motions. A positive $t$ value can be interpreted as lumbopelvic motion occurring after hip motion, while a negative $t$ value means lumbopelvic motion occurs prior to the hip motion [10]. Disruption or sudden change in movement direction of lumbopelvic motion over full range of motion (Fig. 2) was identified by the number of local minimum occurrences (Lmin) [5, 6]. These disruptions were used to represent alteration in lumbopelvic control. A greater number of disruptions in lumbopelvic angular displacement reduces the smoothness of control and thus represents impaired lumbopelvic control [5, 6]. Kinematic data from the first and second sets (average scores within each set) were used to establish within-session test-retest reliability of measurement and 95% confidence minimal detectable difference (MDD ${}_{95}$ ) of parameters of interest (Table 1).

Table 1
Test-retest reliability of measurement and 95% confidence minimal detectable difference of kinematic parameters

Parameter	Right PHR		Left PHR
	ICC ${}_{3,\text{k}}$	MDD ${}_{95}$	ICC ${}_{3,\text{k}}$	MDD ${}_{95}$
LPER (degrees)	0.84	1.00	0.72	1.48
LPIR (degrees)	0.67	1.73	0.69	2.04
r ${}_{0}$	0.72	0.35	0.80	0.28
t (seconds)	0.89	14.38	0.89	12.10
r ${}_{\text{t}}$	0.87	0.09	0.79	0.15
Lmin (occurrences)	0.96	1.67	0.98	0.93

PHR $=$ prone hip rotation, ICC ${}_{3,\text{k}}$ $=$ intraclass correlation coefficient, MDD ${}_{95}$ $=$ 95% confidence minimal detectable difference, LPER $=$ lumbopelvic motion in the same direction of hip external rotation, LPIR $=$ lumbopelvic motion in the same direction of hip internal rotation, r ${}_{0}$ $=$ correlation at zero time-lag, t $=$ time-lag that maximizes correlation, r ${}_{\text{t}}$ $=$ correlation at time-lag, Lmin $=$ local minimum occurrences.

Figure 2.

Example of a participant with low back pain on the painful side (A) and an age-, sex-, and BMI-matched participant without low back pain (B). Kinematic parameters derived from hip rotation (above) and lumbopelvic motions (below) including hip external and internal rotations (HER and HIR, respectively), hip motion velocity, lumbopelvic motion in the same direction of hip external rotation (LPER), lumbopelvic motion in the same direction of hip internal rotation (LPIR), correlation at zero time-lag (r ${}_{0}$ ), time-lag (t), correlation at time-lag (r ${}_{\text{t}}$ ), and local minimum occurrences (Lmin).

Our preliminary analysis demonstrated a significant difference in lumbopelvic rotation during painful-side prone hip external rotation in participants with LBP. In addition, our previous study demonstrated that pain changed lumbopelvic motion during movement tests, and van Dillen et al. demonstrated asymmetrical lumbopelvic movement in patients with LBP [13, 18]. Accordingly, we renamed kinematic variables in the LBP group from right and left to painful (PS) and non-painful side (NS), PS corresponds to where the participants with LBP performed lower leg movement on the same side of LBP (e.g., a participant with right unilateral LBP performed right lower leg movement), and NS corresponds to lower leg movement on the side opposite their LBP. For participants without LBP, we used the same side prone hip rotation as their age-, sex-, and BMI-matched participants with LBP. For example, if the participant with right unilateral LBP performed right lower leg movement (both external and internal hip rotation), kinematic variables from right leg movement in a matched participant without LBP were used for comparison. These side-matched kinematic variables were used for statistical analyses.

2.5 Statistical analysis

Statistical analyses were performed using SPSS version 21 (IBM Corp., New York, USA). The Shapiro-Wilk test was performed to determine normality assumption. Transformation including square root and natural log was performed as needed. An independent $t$ -test was used to determine the differences in kinematic parameters between participants with and without LBP if the parametric assumptions were met. Otherwise, the non-parametric Mann-Whitney U test was used. The Chi-square test was used to compare sex between groups. Significance level ( $\alpha$ ) was held at 0.05 for all analyses. Effect sizes were calculated using Cohen’s d. Differences in the group means were used in conjunction with the MDD ${}_{95}$ and ES to interpret meaningful differences.

Table 2
Comparisons of demographic data, hip range of motion and velocity between participants with low back pain (painful and non-painful side) and side-matched participants without low back pain

Parameter	Non-LBP Mean $\pm$ SD (Min-Max)	LBP Mean $\pm$ SD (Min-Max)	$p$ value
Age (years)	27.7 $\pm$ 5.0	29.4 $\pm$ 5.2	0.54
Sex (%female)	42.9	42.9	1.00
BMI (kg/m ${}^{2}$ )	22.1 $\pm$ 2.3	24.5 $\pm$ 2.2	0.08
NPRS_Current	N/A	5.7 $\pm$ 1.9 (3-8)	N/A
NPRS_High	N/A	5.8 $\pm$ 2.4 (4-10)	N/A
NPRS_Low	N/A	2.5 $\pm$ 1.8 (0-4)	N/A
ODI (%)	N/A	19.7 $\pm$ 7.5 (6-26)	N/A
PS_HER (degrees)	40.30 $\pm$ 10.54	50.17 $\pm$ 21.90	0.27
PS_HIR (degrees)	39.57 $\pm$ 10.46	46.03 $\pm$ 11.70	0.26
PS_Velocity (degrees/second)	35.39 $\pm$ 10.70	31.80 $\pm$ 12.84	0.55
NS_HER (degrees)	47.30 $\pm$ 22.60	49.76 $\pm$ 23.99	0.81 ${}^{\text{a}}$
NS_HIR (degrees)	35.31 $\pm$ 11.41	45.58 $\pm$ 7.78	0.07
NS_Velocity (degrees/second)	35.80 $\pm$ 10.68	28.58 $\pm$ 13.40	0.25

Non-LBP $=$ no low back pain group, LBP $=$ low back pain group, SD $=$ standard deviation, BMI $=$ body mass index, NPRS_Current $=$ numeric pain rating scale at “time of testing”, NPRS_High $=$ numeric pain rating scale “highest level of pain in the last 24-hours”, NPRS_Low $=$ numeric pain rating scale “lowest level of pain in the last 24-hours”, ODI $=$ Oswestry Disability Index, PS $=$ painful side, NS $=$ non-painful side, HER $=$ hip external rotation, HIR $=$ hip internal rotation, ${}^{\text{a}}$ $=$ independent $t$ -test using square root transformed data, N/A $=$ non-applicable.

3. Results

The NS_HER data was not normally distributed. However, square root transformed data demonstrated normal distribution. Therefore, an independent $t$ -test was used to compare differences between groups. Demographic data demonstrated no significant difference ( $p>$ 0.05) in age, sex, and BMI between groups. BMI did trend ( $p=$ 0.08) toward a difference between groups, but both groups were in the normal range ( ${<}$ 25 kg/m ${}^{2}$ ); however, 2 out of 8 (25%) participants in the LBP group had BMI greater than 25 kg/m ${}^{2}$ (26.6 and 27.8 kg/m ${}^{2}$ ). Data also showed that both groups similarly performed PHR ( $p{>}$ 0.05) in terms of hip range of motion (HER and HIR) and average movement velocity (Table 2).

Our data, except for PS_r ${}_{0}$ and NS_r ${}_{\text{t}}$ , were normally distributed; therefore, an independent $t$ -test was used to determine the differences in kinematic parameters between participants with and without LBP. Square root and natural log transformations for PS_r ${}_{0}$ and NS_r ${}_{\text{t}}$ were not successful; therefore, the non-parametric Mann-Whitney U test was used to analyze these two kinematic parameters. In addition, the LBP group demonstrated significantly longer time-lag or movement latency of the lumbopelvic region relative to initiation of hip motion on the painful side (PS_t, $p=$ 0.05) compared to the No-LBP group with a large effect size (Cohen’s d $=$ 1.07) and mean difference exceeding MDD ${}_{95}$ . The LBP group demonstrated a trend ( $p=$ 0.053) toward less lumbopelvic rotation on the painful side in the same direction as hip external rotation (PS_LPER); however, the difference between groups did not exceed MDD ${}_{95}$ . Other time series kinematic parameters also demonstrated trends showing the LBP group had greater time lag on the non-painful side (NS_t, $p=$ 0.12) and greater number of movement disruptions on both the painful (PS_Lmin, $p=$ 0.19) and non-painful side (NS_Lmin, $p=$ 0.17). In all cases, the comparison mean difference between groups exceeded MDD ${}_{95}$ with large effect sizes (Table 3).

Table 3
Comparisons of kinematic parameters (lumbopelvic motion, correlation, time-lag, and local minimum occurrences) between participants with low back pain (painful and non-painful sides) and side-matched participants without low back pain

Parameter	Non-LBP Mean $\pm$ SD	LBP Mean $\pm$ SD	$p$ value		Mean difference		MDD ${}_{95}$	ES (Cohen’s d)
PS_LPER (degrees)	2.58 $\pm$ 1.16	1.58 $\pm$ 0.59	0.	053	1.	00	1.48	1.09
PS_LPIR (degrees)	1.76 $\pm$ 0.97	2.07 $\pm$ 1.20	0.	57	$-$ 0.	31	2.04	0.28
PS_r ${}_{0}$	0.65 $\pm$ 0.20	0.68 $\pm$ 0.25	0.	80 ${}^{\text{a}}$	$-$ 0.	03	0.35	0.13
PS_t (seconds)	0.29 $\pm$ 0.16	0.49 $\pm$ 0.21	0.	05 ${}^{*}$	$-$ 0.	20 $\dagger$	0.14	1.07
PS_r ${}_{\text{t}}$	0.83 $\pm$ 0.10	0.84 $\pm$ 0.15	0.	90	$-$ 0.	01	0.15	0.08
PS_Lmin (occurrences)	9.6 $\pm$ 4.4	15.3 $\pm$ 10.7	0.	19	$-$ 5.	7 $\dagger$	1.67	0.70
NS_LPER (degrees)	2.08 $\pm$ 1.04	2.52 $\pm$ 0.96	0.	39	$-$ 0.	44	1.48	0.44
NS_LPIR (degrees)	1.73 $\pm$ 1.30	1.53 $\pm$ 0.74	0.	70	0.	20	2.04	0.19
NS_r ${}_{0}$	0.61 $\pm$ 0.26	0.64 $\pm$ 0.28	0.	81	$-$ 0.	03	0.35	0.11
NS_t (seconds)	0.28 $\pm$ 0.16	0.47 $\pm$ 0.28	0.	12	$-$ 0.	19 $\dagger$	0.14	0.83
NS_r ${}_{\text{t}}$	0.81 $\pm$ 0.16	0.86 $\pm$ 0.13	0.	30 ${}^{\text{a}}$	$-$ 0.	05	0.15	0.34
NS_Lmin (occurrences)	9.4 $\pm$ 4.0	14.5 $\pm$ 8.9	0.	17	$-$ 5.	10 $\dagger$	1.67	0.74

Non-LBP $=$ no low back pain group, LBP $=$ low back pain group, SD $=$ standard deviation, MDD ${}_{95}$ $=$ 95% confidence minimal detectable difference, ES $=$ effect size, PS $=$ painful side, NS $=$ non-painful side, LPER $=$ lumbopelvic motion in the same direction of hip external rotation, LPIR $=$ lumbopelvic motion in the same direction of hip internal rotation, r ${}_{0}$ $=$ correlation at zero time-lag, t $=$ time-lag that maximizes correlation, r ${}_{\text{t}}$ $=$ correlation at time-lag, Lmin $=$ local minimum occurrences, ${}^{\text{a}}$ $=$ non-parametric Mann-Whitney U test, ${}^{*}$ $=$ significant at $p<$ 0.05, $\dagger$ $=$ mean difference exceeds MDD ${}_{95}$ .

4. Discussion

The results of our study partially support our hypothesis. Our participants with LBP demonstrated delayed onset of lumbopelvic motion as indicated by a greater time-lag between initiation of hip and lumbopelvic motion associated with task performance most evident during performance of hip rotation on their painful side. The LBP group also demonstrated a greater time-lag on the non-painful side, however the difference was not statisitically significant. These differences exceeded the MDD ${}_{95}$ and represents a large and moderate ES, respectively. In addition, there was tentative support for reduced quantity and quality of lumbopelvic motion as our finding demonstrated trends toward reduced lumbopelvic rotation motion on the painful side of the back only and decreased smoothness of movement during performance of active hip motion on both the painful and non-pain sides in the LBP group. Although these differences did not reach statistical significance the group differences in movement control (Lmin), represent a moderate effect size that exceeded MDC ${}_{95}$ . This suggests that these differences are not likely entirely measurement error. In this case, our findings suggest that patients with subacute LBP tend toward diminished quality of lumbopelvic control [5, 6].

The findings of time delay for lumbopelvic rotation during active hip rotation could be associated with attempted increase in spinal stiffness as a mechanism to protect against or in anticipation of increased pain caused by the active hip rotation [2, 3, 4]. This movement timing delay is consistent with the study investigating postural steadiness that found patients with LBP used a lumbar stiffening strategy to compensate for control deficits during single leg standing [8]. Another study investigated the instantaneous center of rotation movement in patients with LBP during active forward bend and reported a decrease in total center of rotation displacement and radius in patients with LBP which the authors interpreted as use of a stiffening strategy during trunk movement [9]. Lamoth et al. also reported a more rigid transverse plane movement pattern of the lumbar spine in patients with LBP during walking [7]. These changes appear consistent with the altered strategy proposed by the “motor adaptation to pain” theory of Hodges and Tucker [2]. They described that the goal of adaptation to pain is to reduce pain and protect the painful structure by modifying muscle activity leading to decrease movement and variability [2].

However, our findings of delayed time of onset of lumbopelvic transverse plane rotation relative to initiation of active hip rotation are inconsistent with other studies of PHR comparing participants with and without chronic LBP [12]. Prior studies have reported greater peak lumbopelvic rotation and earlier onset of lumbopelvic motion in participants with chronic LBP, while we found no significant difference in lumbopelvic rotation range, and delayed onset of lumbopelvic transverse plane rotation motion. We do not believe that our participants with subactue LBP demonstrated differences in hip joint motion that could influence the findings, as our LBP group’s hip motion was similar to that reported for patients with LBP in other studies directly investigating kinematics during the PHR test. However, differences in findings could be driven by different characteristics of LBP participants (e.g., activites, pain intensity and body structure) and methodology.

Prior studies have particularly investigated the difference between extremes of activity characteristics (individuals with LBP who played rotation-related sports and individuals without LBP who did not play rotation-related sports). Repetitive movements over time can cause cumulative micro-trauma to the spinal tissue; thus, individuals engaged in repetitive rotational sports may be prone to developing excessive lumbar motion during PHR [14]. Thus, participants who played rotation-related sports may demonstrate increased and earlier lumbopelvic movement during hip rotation [12, 13]. Our participants with LBP did not currently participate in recreational sports that involved repeated lumbopelvic rotation in transverse plane. In this case, their lack of familiarity with the movement might have resulted in anticipated increased pain during PHR that led to stiffening their lumbopelvic region to protect the painful area or as a compensation for decreased movement control [2, 3, 4].

Pain intensity is another characteristic that could influence the amount or control pattern of movement. In our study, participants with subacute LBP who were currently seeking care for their symptoms had a moderate pain level at rest (5.7 $\pm$ 1.9), while previous studies reported lower levels of current pain (2.9 $\pm$ 1.7) in their sample of indidviduals with chronic LBP [12]. Greater pain intensity could cause greater protection of painful structures through co-contraction and stiffening of the lumbopelvic region leading to differences in findings between studies.

A potenttial confounding factor in our study was our subject’s characteristic of BMI. The BMI trended toward a difference ( $p=$ 0.08) between groups. This was driven by two out of eight (25%) participants in the LBP group that had a BMI greater than 25 kg/m ${}^{2}$ , which put them in the overweight category. Our sample was consistent with the BMI reported by Scholtes et al. [12]. While several studies demonstrate the association between high BMI and LBP [19, 20], our literature review did not produce any studies that directly investigated the association between BMI and lumbopelvic control. We performed a statistical analysis on our BMI and kinematic data. There was no association between BMI and the kinematic variables. Therefore, we were more confident that BMI did not directly influence lumbopelvic movement in our study.

Another plausible explanation for differences in findings comes from variations in methodology. In particular, the approach used to calculate lumbopelvic motion and the number of repetitions used in the calculation were different. Our approach was to calculate transverse plane pelvic rotation relative to the lumbar spine by using a local reference frame. This was done to capture movement between the lumbar region and the pelvis. The approach of previously published studies was to calculate lumbopelvic motion by using pelvic motion relative to a global reference frame and the pelvic starting position. Using the global reference frame and marker locations on the pelvis segment for quantifying pelvic transverse plane rotation does not capture motion between the pelvis and lumbar spine during PHR. In addition, prior studies used a single repetition of PHR, while our study averaged performance over six repetitions potentially stabilizing the measurement. Our time delay and coordination variables were derived using different equations.

The time-delayed lumbopelvic movement could be interpreted in the context of altered patterns of muscle activity [2, 21, 22]. Based on the “motor adaptation to pain” theory, the finding of our study could be associated with increased net amplitude of trunk muscles [3, 23, 24, 25]. Several studies show that variability in trunk muscle activation patterns and net muscle activity is generally reported as increased with varying redistributions of activity between muscles (deep and superficial) [2, 3, 4]. Presumably these changes are driven by a need to increase spinal stability to protect the painful tissue or movement that causes increase in pain [2, 3, 26]. Stiffening strategies may provide a short-term benefit for patients with LBP, but stiffening the lumbopelvic region can increase joint load which could further induce tissue damage. A stiffening strategy would also reduce joint movement that could compromise the capacity of the lumbopelvic region to handle external perturbation and regain postural equilibrium [27, 28]. Increased lumbopelvic stiffness can decrease movement variability which would limit load sharing among lumbopelvic structures [9, 29, 30]. While we did not collect EMG data, the parameters of interest in our study are the direct output of muscle activation and insufficient redistribution of activation could contribute to movement control changes. However, to confirm this interpretation, simultaneous collection of movement patterns and muscle activity during the PHR test is necessary.

Several studies presume that increased muscle activation in patients with LBP is a coping strategy to decrease pain [24, 25, 31]. One study found that even though participants with and without LBP have the same amount of lumbopelvic movement, the lumbopelvic control in participants with LBP was not optimized demonstrating greater angular velocity disruption during forward bend and return to upright movement [6]. This result highlights the importance of quality of the movement. Therefore, clinicians may derive important information about quality of control by observing smoothness of the movement in addition to timing of the lumbopelvic segment motion during active PHR.

Cross-correlations between lower leg and lumbopelvic rotation motions suggest similar lumbopelvic coordination between participants with and without LBP. No significant difference in coordination found in this study could be due to the fact that cross-correlation takes the entire time-series into consideration and both hip and lumbopelvic motions are sinusoidal waveforms. This would result in high correlation. Our findings suggest that cross-correlation might not be the best approach to quantify hip and lumbopelvic coordination. Other approaches, such as a dynamic systems approach, may be more sensitive to small dissociation between two segments [5, 6].

Key limitations in our study include the small sample size that limited generalizability and study power. Evidence for this study being underpowered exists based on our group differences exceeding MDD ${}_{95}$ and representing moderate to large ES related to delay lumbopelvic movement timing and poor quality of movement in our LBP sample. Our sample was part of an intervention study that pre-specified a sample size of 16 (eight participants in each group). This sample size was not large enough to detect significant differences in non-painful side time-lag at maximum curve correlation, and disruption in angular displacement for both painful and non-painful side between participants with and without LBP. A minimum sample size of 68 (34 participants per group) would be required to detect significant differences in these parameters using the calculated ES, 80% power and confidence level of 0.05. In addition, a replication of this study with concurrent EMG data is needed to confirm whether increases in net activation alters trunk muscle activity that significantly contributes to the delayed onset of lumbopelvic movement during active PHR. Based on inclusion criteria, we might have included individuals in our non-LBP group who had a history of LBP. The individuals with a history of LBP may perform prone hip rotation similar to those with current LBP. However, we still found a significant delay in lumbopelvic motion. One study found no significant difference in performance between individuals with and without a history of LBP [9]. Further studies should take this issue into consideration when recruiting participants without LBP. Finally, while we asked the participants whether they had increased pain after each set of prone hip rotation and provided resting period if needed, we did not record the pain intensity for each set. Therefore, we do not have direct evidence to support whether the participants were successful in using a stiffening strategy to reduce their current symptoms during the task.

5. Conclusion

This study investigated the amount, timing, and quality of lumbopelvic movement during active PHR. Although both groups demonstrated a similar amount of lumbopelvic movement during active PHR, participants with LBP demonstrated greater time-delays in lumbopelvic movement and more frequent movement disruptions during PHR. These findings suggest that participants with LBP may alter muscle activation of their lumbopelvic region, resulting in delayed movement and loss of smooth control. This suggests that clinicians should also observe the quality (timing and smoothness) of the movement in addition to the amount of lumbopelvic movement to fully interpert clinical assessment of the PHR test.

Author contributions

PW has significantly contributed to the conception, research design, data collection, data analysis, and drafting and revising the manuscript. SPS has substantially contributed to the conception, data analysis, and editing and revising the manuscript. HKW has significantly contributed to the data analysis and editing and revising the manuscript.

Availability of data and material

The datasets used and/or analyzed during this study are available from the corresponding author upon reasonable request.

Funding

This study was funded in part by the Thailand Research Fund (TRG5880133) and the Faculty of Physical Therapy Research Assistant Fund (Wattananon).

Footnotes

Acknowledgments

We would like to thank the Motor Control and Neural Plasticity Laboratory at Mahidol University for providing data collection space and equipment. We would also like to thank Ms. Tanatta Chichakan and Mr. Pisit Suwanimit for their help with data collection. We would lastly like to thank all participants for participating in this study.

Conflict of interest

The authors declare that they have no competing interests.

References

Laird

Gilbert

Kent

Keating

. Comparing lumbo-pelvic kinematics in people with and without back pain: a systematic review and meta-analysis. BMC Musculoskelet Disord. 2014; 15: 229.

Hodges

Tucker

. Moving differently in pain: a new theory to explain the adaptation to pain. Pain. 2011; 152(3 Suppl): S90-8.

Hodges

Coppieters

MacDonald

Cholewicki

. New insight into motor adaptation to pain revealed by a combination of modelling and empirical approaches. Eur J Pain. 2013; 17(8): 1138-46.

Hodges

Tsao

Sims

. Gain of postural responses increases in response to real and anticipated pain. Exp Brain Res. 2015; 233(9): 2745-52.

Spinelli

Wattananon

Silfies

Talaty

Ebaugh

. Using kinematics and a dynamical systems approach to enhance understanding of clinically observed aberrant movement patterns. Man Ther. 2015; 20(1): 221-6.

Wattananon

Ebaugh

Biely

Smith

Hicks

Silfies

. Kinematic characterization of clinically observed aberrant movement patterns in patients with non-specific low back pain: a cross-sectional study. BMC Musculoskelet Disord. 2017; 18(1): 455.

Lamoth

Meijer

Daffertshofer

Wuisman

Beek

. Effects of chronic low back pain on trunk coordination and back muscle activity during walking: changes in motor control. Eur Spine J. 2006; 15(1): 23-40.

Sung

Danial

. Analysis of relative kinematic index with normalized standing time between subjects with and without recurrent low back pain. Eur Spine J. 2017; 26(2): 518-27.

Wattananon

Intawachirarat

Cannella

Sung

Silfies

. Reduced instantaneous center of rotation movement in patients with low back pain. Eur Spine J. 2018; 27(1): 154-62.

10.

Nelson-Wong

Howarth

Winter

Callaghan

. Application of autocorrelation and cross-correlation analyses in human movement and rehabilitation research. J Orthop Sports Phys Ther. 2009; 39(4): 287-95.

11.

Luomajoki

Kool

de Bruin

Airaksinen

. Movement control tests of the low back; evaluation of the difference between patients with low back pain and healthy controls. BMC Musculoskelet Disord. 2008; 9: 170.

12.

Scholtes

Gombatto

Van Dillen

. Differences in lumbopelvic motion between people with and people without low back pain during two lower limb movement tests. Clin Biomech (Bristol, Avon). 2009; 24(1): 7-12.

13.

van Dillen

Gombatto

Collins

Engsberg

Sahrmann

. Symmetry of timing of hip and lumbopelvic rotation motion in 2 different subgroups of people with low back pain. Arch Phys Med Rehabil. 2007; 88(3): 351-60.

14.

Sahrmann

Azevedo

Dillen

. Diagnosis and treatment of movement system impairment syndromes. Braz J Phys Ther. 2017; 21(6): 391-9.

15.

Tong

Mousavi

Kiers

Ferreira

Refshauge

van Dieën

. Is there a relationship between lumbar proprioception and low back pain? A systematic review with meta-analysis. Arch Phys Med Rehabil. 2017; 98(1): 120-36.e2.

16.

Balagué

Mannion

Pellisé

Cedraschi

. Non-specific low back pain. Lancet. 2012; 379(9814): 482-91.

17.

Siegler

Allard

Kirtley

Leardini

Rosenbaum

, et al. ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion – part I: ankle, hip, and spine. J Biomech. 2002; 35(4): 543-8.

18.

Wattananon

Silfies

Tretriluxana

Jalayondeja

. Lumbar multifidus and erector spinae muscle synergies in patients with nonspecific low back pain during prone hip extension: a across-sectional study. Pm r. 2019; 11(7): 694-702.

19.

Heuch

Hagen

Heuch

Nygaard

Zwart

. The impact of body mass index on the prevalence of low back pain: the HUNT study. Spine (Phila Pa 1976). 2010; 35(7): 764-8.

20.

Shiri

Karppinen

Leino-Arjas

Solovieva

Viikari-Juntura

. The association between obesity and low back pain: a meta-analysis. Am J Epidemiol. 2010; 171(2): 135-54.

21.

Lund

Donga

Widmer

Stohler

. The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity. Can J Physiol Pharmacol. 1991; 69(5): 683-94.

22.

Streisfeld

Bartoszek

Creran

Inge

McShane

Johnston

. Relationship between body positioning, muscle activity, and spinal kinematics in cyclists with and without low back pain: a systematic review. Sports Health. 2017; 9(1): 75-9.

23.

Marras

Davis

Ferguson

Lucas

Gupta

. Spine loading characteristics of patients with low back pain compared with asymptomatic individuals. Spine (Phila Pa 1976). 2001; 26(23): 2566-74.

24.

Silfies

Squillante

Maurer

Westcott

Karduna

. Trunk muscle recruitment patterns in specific chronic low back pain populations. Clin Biomech (Bristol, Avon). 2005; 20(5): 465-73.

25.

Stanton

Kawchuk

. The effect of abdominal stabilization contractions on posteroanterior spinal stiffness. Spine (Phila Pa 1976). 2008; 33(6): 694-701.

26.

van Dieen

Cholewicki

Radebold

. Trunk muscle recruitment patterns in patients with low back pain enhance the stability of the lumbar spine. Spine (Phila Pa 1976). 2003; 28(8): 834-41.

27.

Mok

Brauer

Hodges

. Failure to use movement in postural strategies leads to increased spinal displacement in low back pain. Spine (Phila Pa 1976). 2007; 32(19): E537-43.

28.

Mok

Brauer

Hodges

. Changes in lumbar movement in people with low back pain are related to compromised balance. Spine (Phila Pa 1976). 2011; 36(1): E45-52.

29.

Colloca

Hinrichs

. The biomechanical and clinical significance of the lumbar erector spinae flexion-relaxation phenomenon: a review of literature. J Manipulative Physiol Ther. 2005; 28(8): 623-31.

30.

Olson

Solomonow

. Flexion-relaxation response to gravity. J Biomech. 2006; 39(14): 2545-54.

31.

Panjabi

. Clinical spinal instability and low back pain. J Electromyogr Kinesiol. 2003; 13(4): 371-9.

Patients with low back pain use stiffening strategy to compensate for movement control during active prone hip rotation: A cross-sectional study

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Background

2.1 Participants

2.2 Instruments and measures

2.3 Procedure

2.4 Data analysis

Table 1 Test-retest reliability of measurement and 95% confidence minimal detectable difference of kinematic parameters

Table 2 Comparisons of demographic data, hip range of motion and velocity between participants with low back pain (painful and non-painful side) and side-matched participants without low back pain

Table 3 Comparisons of kinematic parameters (lumbopelvic motion, correlation, time-lag, and local minimum occurrences) between participants with low back pain (painful and non-painful sides) and side-matched participants without low back pain

5. Conclusion

Author contributions

Availability of data and material

Funding

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Test-retest reliability of measurement and 95% confidence minimal detectable difference of kinematic parameters

Table 2
Comparisons of demographic data, hip range of motion and velocity between participants with low back pain (painful and non-painful side) and side-matched participants without low back pain

Table 3
Comparisons of kinematic parameters (lumbopelvic motion, correlation, time-lag, and local minimum occurrences) between participants with low back pain (painful and non-painful sides) and side-matched participants without low back pain