Journal of Sport Rehabilitation, 2014, 23, 307-318 http://dx.doi.org/10.1123/jsr.2013-0057 © 2014 Human Kinetics, Inc.
www.JSR-Journal.com ORIGINAL RESEARCH REPORT
Changes in Muscle Thickness After Exercise and Biofeedback in People With Low Back Pain Shandi L. Partner, Mark Alan Sutherlin, Shellie Acocello, Susan A. Saliba, Eric M. Magrum, and Joe M. Hart Context: Individuals with low back pain (LBP) have reduced function of the transversus abdominis (TrA) and lumbar multifidus (LM) muscles. Biofeedback during exercise may increase the ability to contract the TrA and LM muscles compared with exercise alone. Objective: To compare TrA preferential activation ratio (PAR) and the percent change in LM-muscle thickness in patients with LBP history before and after exercise with or without biofeedback. Design: Controlled laboratory study. Setting: University research laboratory. Patients: 20 LBP individuals, 10 exercise alone and 10 exercise with biofeedback. Interventions: Patients were allotted to tabletop exercises in isolation or tabletop exercises with visual, auditory, and tactile biofeedback. Main Outcome Measures: TrA PAR and percent change in LM-muscle thickness. Results: There were no differences between groups at baseline (all P > .05). Nonparametric statistics showed decreased resting muscle thickness for total lateral abdominal-wall muscles (P = .007) but not TrA (P = .410) or LM (P = .173). Percent TrA thickness increased from table to standing positions before (P = .006) and after exercise (P = .009). TrA PAR increased after exercise (pre 0.01 ± 0.02, post 0.03 ± 0.04, P = .033) for all patients and for exercise with biofeedback (pre 0.02 ± 0.01, post 0.03 ± 0.01, P = .037) but not for exercise alone (pre 0.01 ± 0.02, post 0.02 ± 0.05, P = .241). No group differences were observed for TrA PAR before (exercise 0.01 ± 0.02, exercise with biofeedback 0.02 ± 0.01, P = .290) or after exercise (exercise 0.02 ± 0.05, exercise with biofeedback 0.03 ± 0.01, P = .174). There were no group differences in LM percent change before exercise (P = .999) or after exercise (P = .597). In addition, no changes were observed in LM percent change as a result of exercise among all participants (P = .391) or for each group (exercise P = .508, exercise with biofeedback P = .575). Conclusion: TrA PAR increased after a single session of exercises, whereas no thickness changes occurred in LM. Keywords: transversus abdominis, preferential activation ratio, lumbar multifidus Low back pain (LBP) is one of the most common conditions reported,1 with annual costs exceeding $25 billion annually.2 Nearly one-third of all LBP individuals will go on to experience additional episodes of LBP.3 The individual and financial burdens of this condition emphasize the importance of treatment options and availability to decrease LBP episodes. Stability of the spine involves a coordinated system of activated spinal muscles, passive spinal-column anatomy, and neural components.4 The contribution of musculature to spinal stability is derived from both global and local muscles.4 Global musculature refers to prime movers of the spine, such as the erector spinae, whereas Partner is with the Fairfax County Public School Athletic Training Education Program, McLean High School, McLean, VA. Sutherlin, Acocello, Saliba, and Hart are with the Kinesiology Program, University of Virginia, Charlottesville, VA. Magrum is with Healthsouth Outpatient Sports Medicine, University of Virginia, Charlottesville, VA. Address author correspondence to Shandi Partner at [email protected].
local stabilizing muscles refer to the smaller muscles that are thought to provide segmental stability of the vertebral column. Coordination between prime movers and local stabilizers has been theorized as an essential neuromuscular component of maintaining a healthy, nonpainful low back.4 LBP individuals present with altered neuromuscular control of the local muscles, which could result in lumbar instability.5,6 Therefore, methods for promoting activation of local stabilizing musculature of the low back have become a prevalent component of spine rehabilitation in patients with LBP. The transversus abdominis (TrA) is a local abdominal-wall muscle that contributes primarily to lumbarspine stability.5 The TrA aids in movement stability by activating before limb movement in a normal population.5 Individuals with LBP exhibit a reduction in this anticipatory activation, resulting in increased vertebral motion, which can cause trauma to the structures of the spine.7 The lumbar multifidus (LM) is a small, intrinsic posterior spine muscle that also contributes to spinal stability.8 Atrophy of the LM muscle has been observed in computed tomography 9 and magnetic resonance 307
308 Partner et al
imaging techniques in approximately 80% of LBP individuals.10 Rehabilitative efforts for LBP have focused on improving local spinal-stabilizing-muscle activation. In addition, educational techniques focused on training patients to activate the TrA may reduce LBP symptoms and recurrence of LBP episodes.11 Programs designed to teach patients to preferentially activate the local stabilizers often involve the use of biofeedback in the LBP population.12 Biofeedback can be a verbal description and encouragement during activation exercises such as the abdominal drawing-in maneuver (ADIM),13–16 visual feedback via real-time ultrasound imaging,17,18 or tactile feedback through pressure-cuff systems.19,20 The preferential activation ratio (PAR) is a calculation comparing the TrA-muscle thickness with the entire lateral abdominal-muscle-wall thickness in both the contracted and rested states13 through the use of either the ADIM13 or abdominal-hollowing exercises.21,22 The TrA PAR indicates the ability to contract the TrA independently of the lateral abdominal wall, which may be used as a clinical measure to assess lumbar stability of the TrA by observing an increase in TrA without increase in thickness of the internal or external oblique muscles.13 While the LM muscle has both deep and superficial layers that may have different contributions to spinal control,23 changes in muscle thickness of the entire muscle have been previously reported to quantify stabilization of the LM.24 Clinical methods for improving activation of the TrA and LM using several forms of biofeedback increase activation of the TrA and LM during non-weight-bearing positions25; however, the current literature regarding activation of local stabilizing musculature during functional tasks is limited.21,26–36 Translation of activation patterns learned during tabletop exercises to weight-bearing positions is an important component of rehabilitation progression in patients with LBP. The purpose of this study was to determine if exercise in isolation or the use of visual, verbal, and tactile biofeedback during TrA and LM muscle-activation exercises would increase TrA PAR and LM-muscle thickness in patients with a history of LBP during functional weight-bearing positions. We hypothesized that both exercise alone and exercise with biofeedback would increase TrA PAR and LM-muscle thickness after a single session of exercise.
Methods Design A pretest–posttest repeated-measures design was used to measure the effects of exercise alone or exercise with biofeedback on TrA PAR and percent change in LMmuscle thickness. The dependent variables were TrA PAR and percent change in LM-muscle thickness. The independent variables were group (exercise alone or exercise with biofeedback) and time (pretest, posttest).
Participants The current study was part of another study examining lumbopelvic neuromuscular differences between LBP and healthy individuals. A total of 22 patients with a history of LBP were included and allocated to 1 of 2 groups, exercise alone or exercise with biofeedback. Group assignment was counterbalanced to ensure equal number of patients with a history of LBP per group (Figure 1). Patients were eligible to participate in this study if they had a self-reported history of LBP within the last 6 months, had had no lower-extremity injuries or surgery in the past 6 months, and were classified as a subgroup of patients with LBP who would benefit from stabilization exercise.37 Requirements for the inclusion of patients with LBP to be eligible for stabilization exercises were determined if individuals presented with a minimum of 3 of the following criteria: age under 40 years, flexibility measured by straight-leg-raise range of motion greater than 91°, deviant movements during trunk flexion and extension, and a positive prone instability test.37 Including a subgroup of LBP individuals based on clinical recommendation factors that support the application of stabilization exercises may provide greater changes in segmental muscle thickness and result in better clinical outcomes.37 All participants were recruited through flyers in a university setting and then prescreened for inclusion before testing. Participants were excluded if they self-reported a previous medical diagnosis or receiving previous medical treatment for spondylolysis, spondylolisthesis, neurologic pathology, radiculopathy, nerve-root impingement, or disc pathology. Patients were also excluded if they had received additional medically supervised isolated TrA- or LM-activation exercise. Approval of this study was obtained from the university institutional review board, informed consent was obtained for each subject before any screening procedures were performed, and all rights of the subjects were protected. The order of the procedures and study is outlined in Figure 1.
Instruments A GE LOGIQ Book XP (GE Medical Systems, Milwaukee, WI) with an 8-MHz linear-array transducer was used for patient education through visual biofeedback and to capture still images of the TrA and LM muscles. Images were exported, and muscle thickness for the abdominal muscles and LM was then measured using Image J image-processing software (Version 1.41o, Wayne Rasband National Institutes of Health, USA). All PAR and LM percent-change calculations were measured using Microsoft Excel software (Microsoft Corp, Redman, WA). Pressure biofeedback was performed using a LabTron sphygmomanometer (GF Healthy Products Inc, Atlanta, GA) during the tactile intervention of the educational protocol.
Changes in Muscle Thickness After Exercise 309
Figure 1 — CONSORT flowchart.
Subjective Measures Low Back Impairment. Low back impairment was measured using a 6-point scale to rate current levels of disability or impairment through a series of 10 questions related to activities of daily living, pain intensity, traveling, social activities, and the changing degree of pain. Reportable answers ranged from zero, indicating low to minimal disability or impairment, to 5, corresponding to the greatest impairment or disability for each item. Percentage of impairment was calculated by taking the total score reported and dividing the total score possible, similar to other subject questionnaires that measure disability in patients with LBP.38,39 Fear-Avoidance Beliefs. The Fear Avoidance Beliefs Questionnaire (FABQ) consists of 16 questions that are divided into 2 subgroups to measure pain responses to daily living activities.40 The FABQ was originally developed to identify behavioral changes to LBP focusing on physical activity and work and has good test reproducibility.40,41 Scoring for each subscale was determined by combining the total response scores reported for each subsection.
Visual Analogue Scale. Pain was measured using an 11-point Likert scale with zero indicating no pain and 10 maximal pain. Patients completed the visual analogue scale based on current pain levels before testing.
Procedures Participants reported to the university research laboratory for a single session. After screening, locations of the TrA and LM muscles were identified using ultrasound in the tabletop position (Figure 2), followed by recording muscle activity at rest standing and during a functional task (Figure 3). Patients were then allocated to receive a single session of exercises with or without biofeedback followed by repeated measures of TrA and LM activation in standing (Figure 1). Tabletop Measures. Both TrA and LM measurements
were assessed in the tabletop position before testing to identify proper muscle identification and to ensure proper technique of the ADIM during the intervention period. The TrA was measured supine on a table with patients’ legs in a crook-lying position. The ultrasound transducer was originally placed transversely superior to the iliac
310 Partner et al
Figure 2 — Ultrasound-transducer placement and musclethickness measurements in tabletop positions: (A) transversus abdominis ultrasound placement, (B) transversus abdominis and lateral abdominal-wall-thickness measurements, (C) lumbar multifidus ultrasound placement, and (D) lumbar multifidus thickness measurement.
Figure 3 — Testing measurement for the transversus abdominis and lumbar multifidus muscles: (A) resting measure for transversus abdominis, (B) contracted measure for transversus abdominis, (C) resting measure for lumbar multifidus, and (D) contracted measure for lumbar multifidus.
crest approximately 10 cm lateral from the umbilicus42 along the midaxillary line.13,19,43,44 Standardization for final ultrasound-transducer placement occurred when the fascial border of the TrA and thoracolumbar fascia was visible on the lateral side of the ultrasound screen and the best images of the TrA, external oblique, and internal oblique were obtained (Figure 2).13,15,43,44 Resting images were obtained through a static image immediately after exhalation.43,44 Thickness measures for the TrA were reported in centimeters and calculated at the thickest position on the ultrasound screen between the superior and deep fascial borders,13,19,43 while the total lateral abdominal-wall thickness was measured from the deep border of the TrA to the superficial border of the external oblique.13,19,43 Two markings were placed opposite the transducer head against the skin to reference the starting transducer position (Figure 2). Transducer placement for the LM occurred longitudinally at the L4–L5 level with the patient in the prone position.24,45 The patient’s arms were positioned at the sides for tabletop measures, and the ultrasound transducer was rotated medially and toward the spine until the L4– L5 facet joint was visible on the ultrasound screen.24,45 Measurements were recorded at the L4–L5 facet, as differences have been seen at the L5 level between healthy and LBP patients.46 Two lines were placed on the opposite side of the transducer head to identify starting reference location for the standing measurements. Muscle thickness was obtained at the L4–L5 facet and measured as the distance in centimeters between the zygapophyseal joint and the superficial facial border of the LM with the subcutaneous tissue (Figure 2).24,45,47–49 Standing Measures. Resting standing measurements of the TrA and LM were taken bilaterally with participants positioned in quiet standing with their arms at their sides (Figure 3).27 Ultrasound-transducer placements for the TrA and LM were applied using the markings from the tabletop positions. Measurements for the TrA occurred when the fascial border of the TrA was on the lateral portion of the ultrasound screen, as previously measured in the tabletop position. LM-muscle images were obtained with the L4–L5 facet joint on the ultrasound screen, similar to the prone position. Contracted measurements were recorded without the addition of a voluntary contraction during 1 of 2 functional tasks (Figure 3). Functional measurements of the contracted TrA were recorded on the stance limb during single-leg stance while individuals maintained approximately 90° of hip and knee flexion on the non-weight-bearing leg. Functional measurements of the contracted LM were obtained on the trailing limb during a forward-stepping task while holding the trailing foot plantar flexed in the toe-off position. Images were recorded in the same positions bilaterally. Subjects were given instructions of the functional positions, but no additional instructions or feedback was administered during the functional tasks. Intervention. Participants performed exercises in isolation or exercises with biofeedback that may be commonly used during clinical rehabilitation. Exercises
Changes in Muscle Thickness After Exercise 311
between the 2 groups were paired, so both groups completed similar workloads. Participants in the exerciseonly group completed tabletop TrA exercises in both prone and supine positions and were verbally instructed in the ADIM consistent with previous techniques27 by pulling the umbilicus toward the table during exhalation. No additional verbal feedback was given beyond the instruction technique before exercise. Participants completed a total of 15 ADIM repetitions with a 5-second hold in both positions. LM exercises consisted of raising the coccyx toward the ceiling while in a prone position and performing a prone active leg raise for a total of 15 repetitions with a 5-second hold time during each repetition for each exercise. The exercise-with-biofeedback group received biofeedback from a combination of 3 different learning tactics: verbal, tactile, and visual feedback during exercise. During TrA exercises, patients were given visual biofeedback through ultrasound imaging and verbal biofeedback consisting of additional auditory cues during exercise on breathing and focusing contractions on the lower abdominal area simultaneously while contracting the TrA. Tactile biofeedback was provided while in a prone position with the use of a pressure cuff during TrA contraction. The pressure cuff was inflated to 80 mm Hg, and each participant was instructed to perform an ADIM, decreasing the cuff between 6 and 8 mm Hg and then maintaining the pressure during the contraction. Patients completed both exercises for 15 repetitions each with 5 seconds duration. LM verbal, visual, and tactile biofeedback were administered in the prone position. Tactile feedback was administered during the prone active leg raise against resistance, where the tester put one hand over the LM so the participant could feel the contraction, and as the tester provided a small distraction on the leg instructing the participant to pull the hip back into the socket. Both of these were performed 7 times on each leg with durations of 5 seconds. Verbal and visual biofeedback using ultrasound were provided as the patients tried to lift their coccyx to the ceiling for 15 repetitions each lasting 5 seconds. The exercises for both groups are outlined in Table 1.
Data Processing Preintervention and postintervention images of the TrA and LM in the standing positions were visualized and measured using Image J image-processing software. Before thickness measurements, images were scaled using a 1-cm reference line to determine the number of pixels measured. This length was converted to a measurement scale that could be used to measure muscle thickness. Muscle thicknesses from 3 images during quiet standing and respective functional-movement conditions were calculated bilaterally and averaged by a blinded data processor.48 Images from both sides were averaged for statistical analysis. Muscle thickness of the TrA, LM, and total lateral were recorded in both resting and contracted states. Percentages TrA-thickness ratios at rest while supine and standing were calculated by observing the percent of the TrA thickness to the total lateral abdominal wall,50 and preferential activation ratios13 were calculated using the following formulas: Percent TrA thickness = TrARest/(TrA + IO + EO)Rest TrA preferential activation = TrAContracted/(TrA + IO + EO)Contracted – TrARest/(TrA + IO + EO)Rest Total thickness of the LM muscle was calculated in the resting state and during a forward-step maneuver. Percent LM-muscle-thickness changes were calculated for the total thickness with the following formula17:
LM thickness change = (LMContractedThickness – LMRestingThickness)/LMRestingThickness × 100
Statistical Analysis All statistical analyses were performed using SPSS version 19.0 (SPSS, Inc, Chicago, IL). Descriptive statistics were compared using independent t tests for continuous outcomes and Mann-Whitney U tests for categorical variables. Preliminary data diagnostics revealed that TrA
Table 1 Exercises for Intervention Exercise Alone Muscle
Exercise With Biofeedback
Maneuver
Routine
Maneuver
Routine
ADIM supine ADIM prone
15 reps 5 s 15 reps 5 s
ADIM supinea,b Pressure-cuff ADIMc
15 reps 5 s 15 reps 5 s
Sacrum to ceiling PALR
15 reps 5 s 15 reps 5 s each leg
Sacrum to ceilinga,b PALRc Hip to socketc
15 reps 5 s 7 reps 5 s each leg 7 reps 5 s each leg
Transversus abdominis
Lumbar multifidus
Abbreviations: ADIM, abdominal drawing-in maneuver; PALR, prone active leg raise. a Verbal biofeedback. b Visual biofeedback. c Tactile biofeedback.
312 Partner et al
PAR and LM percent-change data were nonnormally distributed. Nonparametric statistics were used for both within and between-groups analyses. Resting muscle thickness was compared using a Friedman test with pairwise comparisons for significant findings. Betweengroups differences were analyzed using 4 separate MannWhitney U tests (standing TrA PAR preexercise, standing TrA PAR postexercise, standing LM percent change in muscle thickness preexercise, and standing LM percent change in muscle thickness postexercise). Within-group preexercise to postexercise differences of the standing TrA PAR and LM percent change in muscle thickness were analyzed using Wilcoxon-sign ranked tests (TrA PAR, LM percent change in muscle thickness for all patients, TrA PAR and LM percent change for exercise only and exercise with biofeedback groups) along with percent TrA thickness ratios while supine and standing. The level of significance was set a priori at P ≤ .05. Effect sizes were calculated using Cohen d with 95% confidence intervals (CI) and interpreted as weak (.8)
Results A total of 20 patients (11 men, 9 women; age 23 ±4 y, height 175.26 ±12.90 cm, mass 74.64 ±15.69 kg) completed the single exercise session (Table 2). Two patients were not included in the final analysis due to 1 not completing the exercise training and a second due to ultrasound imaging error. No differences were observed between groups for age, height, mass, body-mass index, visual analogue scale, low back impairment, or FABQ subscale scores; however, there was a difference in the number of men and women per group (Table 2). Painrecurrence frequency and duration were reported for 19 of the 20 subjects and are listed in Table 3. Overall, patients with a history of LBP reported a median pain level of 2
ranging from 0 to 3 on a visual analogue scale, along with reported impairment of 22.7% (Table 2).
Resting Muscle Thickness Resting muscle thickness for both table and standing measures can be found in Table 4. There were no differences between table and standing measures before or after exercise for the TrA (χ2 = 2.880, P = .410) or LM (χ2 = 4.980, P = .173) muscles for all patients. A significant decrease was observed in resting muscle thickness for the total lateral wall (χ2 = 12.180, P = .007). Pairwise comparisons showed lower resting muscle thickness during both standing before exercise (P = .013) and standing after exercise (P = .042) than tabletop measures before exercise. The TrA accounted for a larger portion of the total lateral abdominal-wall thickness in the standing position both before exercise (Z = –2.725, P = .006) and after exercise (Z = –2.613, P = .009) compared with the table measures (Table 4).
TrA PAR Table 5 shows the TrA PAR values before and after exercise. There were no between-groups differences in TrA PAR at baseline (Z = –1.134, P = .257) or after exercise
Table 3 Reported Pain-Symptom Duration and Frequency, N = 19 Pain frequency
n
Pain duration
n
Daily Weekly Monthly Annually
2 11 4 2
All day 12+ h 1–12 h
www.JSR-Journal.com ORIGINAL RESEARCH REPORT
Changes in Muscle Thickness After Exercise and Biofeedback in People With Low Back Pain Shandi L. Partner, Mark Alan Sutherlin, Shellie Acocello, Susan A. Saliba, Eric M. Magrum, and Joe M. Hart Context: Individuals with low back pain (LBP) have reduced function of the transversus abdominis (TrA) and lumbar multifidus (LM) muscles. Biofeedback during exercise may increase the ability to contract the TrA and LM muscles compared with exercise alone. Objective: To compare TrA preferential activation ratio (PAR) and the percent change in LM-muscle thickness in patients with LBP history before and after exercise with or without biofeedback. Design: Controlled laboratory study. Setting: University research laboratory. Patients: 20 LBP individuals, 10 exercise alone and 10 exercise with biofeedback. Interventions: Patients were allotted to tabletop exercises in isolation or tabletop exercises with visual, auditory, and tactile biofeedback. Main Outcome Measures: TrA PAR and percent change in LM-muscle thickness. Results: There were no differences between groups at baseline (all P > .05). Nonparametric statistics showed decreased resting muscle thickness for total lateral abdominal-wall muscles (P = .007) but not TrA (P = .410) or LM (P = .173). Percent TrA thickness increased from table to standing positions before (P = .006) and after exercise (P = .009). TrA PAR increased after exercise (pre 0.01 ± 0.02, post 0.03 ± 0.04, P = .033) for all patients and for exercise with biofeedback (pre 0.02 ± 0.01, post 0.03 ± 0.01, P = .037) but not for exercise alone (pre 0.01 ± 0.02, post 0.02 ± 0.05, P = .241). No group differences were observed for TrA PAR before (exercise 0.01 ± 0.02, exercise with biofeedback 0.02 ± 0.01, P = .290) or after exercise (exercise 0.02 ± 0.05, exercise with biofeedback 0.03 ± 0.01, P = .174). There were no group differences in LM percent change before exercise (P = .999) or after exercise (P = .597). In addition, no changes were observed in LM percent change as a result of exercise among all participants (P = .391) or for each group (exercise P = .508, exercise with biofeedback P = .575). Conclusion: TrA PAR increased after a single session of exercises, whereas no thickness changes occurred in LM. Keywords: transversus abdominis, preferential activation ratio, lumbar multifidus Low back pain (LBP) is one of the most common conditions reported,1 with annual costs exceeding $25 billion annually.2 Nearly one-third of all LBP individuals will go on to experience additional episodes of LBP.3 The individual and financial burdens of this condition emphasize the importance of treatment options and availability to decrease LBP episodes. Stability of the spine involves a coordinated system of activated spinal muscles, passive spinal-column anatomy, and neural components.4 The contribution of musculature to spinal stability is derived from both global and local muscles.4 Global musculature refers to prime movers of the spine, such as the erector spinae, whereas Partner is with the Fairfax County Public School Athletic Training Education Program, McLean High School, McLean, VA. Sutherlin, Acocello, Saliba, and Hart are with the Kinesiology Program, University of Virginia, Charlottesville, VA. Magrum is with Healthsouth Outpatient Sports Medicine, University of Virginia, Charlottesville, VA. Address author correspondence to Shandi Partner at [email protected].
local stabilizing muscles refer to the smaller muscles that are thought to provide segmental stability of the vertebral column. Coordination between prime movers and local stabilizers has been theorized as an essential neuromuscular component of maintaining a healthy, nonpainful low back.4 LBP individuals present with altered neuromuscular control of the local muscles, which could result in lumbar instability.5,6 Therefore, methods for promoting activation of local stabilizing musculature of the low back have become a prevalent component of spine rehabilitation in patients with LBP. The transversus abdominis (TrA) is a local abdominal-wall muscle that contributes primarily to lumbarspine stability.5 The TrA aids in movement stability by activating before limb movement in a normal population.5 Individuals with LBP exhibit a reduction in this anticipatory activation, resulting in increased vertebral motion, which can cause trauma to the structures of the spine.7 The lumbar multifidus (LM) is a small, intrinsic posterior spine muscle that also contributes to spinal stability.8 Atrophy of the LM muscle has been observed in computed tomography 9 and magnetic resonance 307
308 Partner et al
imaging techniques in approximately 80% of LBP individuals.10 Rehabilitative efforts for LBP have focused on improving local spinal-stabilizing-muscle activation. In addition, educational techniques focused on training patients to activate the TrA may reduce LBP symptoms and recurrence of LBP episodes.11 Programs designed to teach patients to preferentially activate the local stabilizers often involve the use of biofeedback in the LBP population.12 Biofeedback can be a verbal description and encouragement during activation exercises such as the abdominal drawing-in maneuver (ADIM),13–16 visual feedback via real-time ultrasound imaging,17,18 or tactile feedback through pressure-cuff systems.19,20 The preferential activation ratio (PAR) is a calculation comparing the TrA-muscle thickness with the entire lateral abdominal-muscle-wall thickness in both the contracted and rested states13 through the use of either the ADIM13 or abdominal-hollowing exercises.21,22 The TrA PAR indicates the ability to contract the TrA independently of the lateral abdominal wall, which may be used as a clinical measure to assess lumbar stability of the TrA by observing an increase in TrA without increase in thickness of the internal or external oblique muscles.13 While the LM muscle has both deep and superficial layers that may have different contributions to spinal control,23 changes in muscle thickness of the entire muscle have been previously reported to quantify stabilization of the LM.24 Clinical methods for improving activation of the TrA and LM using several forms of biofeedback increase activation of the TrA and LM during non-weight-bearing positions25; however, the current literature regarding activation of local stabilizing musculature during functional tasks is limited.21,26–36 Translation of activation patterns learned during tabletop exercises to weight-bearing positions is an important component of rehabilitation progression in patients with LBP. The purpose of this study was to determine if exercise in isolation or the use of visual, verbal, and tactile biofeedback during TrA and LM muscle-activation exercises would increase TrA PAR and LM-muscle thickness in patients with a history of LBP during functional weight-bearing positions. We hypothesized that both exercise alone and exercise with biofeedback would increase TrA PAR and LM-muscle thickness after a single session of exercise.
Methods Design A pretest–posttest repeated-measures design was used to measure the effects of exercise alone or exercise with biofeedback on TrA PAR and percent change in LMmuscle thickness. The dependent variables were TrA PAR and percent change in LM-muscle thickness. The independent variables were group (exercise alone or exercise with biofeedback) and time (pretest, posttest).
Participants The current study was part of another study examining lumbopelvic neuromuscular differences between LBP and healthy individuals. A total of 22 patients with a history of LBP were included and allocated to 1 of 2 groups, exercise alone or exercise with biofeedback. Group assignment was counterbalanced to ensure equal number of patients with a history of LBP per group (Figure 1). Patients were eligible to participate in this study if they had a self-reported history of LBP within the last 6 months, had had no lower-extremity injuries or surgery in the past 6 months, and were classified as a subgroup of patients with LBP who would benefit from stabilization exercise.37 Requirements for the inclusion of patients with LBP to be eligible for stabilization exercises were determined if individuals presented with a minimum of 3 of the following criteria: age under 40 years, flexibility measured by straight-leg-raise range of motion greater than 91°, deviant movements during trunk flexion and extension, and a positive prone instability test.37 Including a subgroup of LBP individuals based on clinical recommendation factors that support the application of stabilization exercises may provide greater changes in segmental muscle thickness and result in better clinical outcomes.37 All participants were recruited through flyers in a university setting and then prescreened for inclusion before testing. Participants were excluded if they self-reported a previous medical diagnosis or receiving previous medical treatment for spondylolysis, spondylolisthesis, neurologic pathology, radiculopathy, nerve-root impingement, or disc pathology. Patients were also excluded if they had received additional medically supervised isolated TrA- or LM-activation exercise. Approval of this study was obtained from the university institutional review board, informed consent was obtained for each subject before any screening procedures were performed, and all rights of the subjects were protected. The order of the procedures and study is outlined in Figure 1.
Instruments A GE LOGIQ Book XP (GE Medical Systems, Milwaukee, WI) with an 8-MHz linear-array transducer was used for patient education through visual biofeedback and to capture still images of the TrA and LM muscles. Images were exported, and muscle thickness for the abdominal muscles and LM was then measured using Image J image-processing software (Version 1.41o, Wayne Rasband National Institutes of Health, USA). All PAR and LM percent-change calculations were measured using Microsoft Excel software (Microsoft Corp, Redman, WA). Pressure biofeedback was performed using a LabTron sphygmomanometer (GF Healthy Products Inc, Atlanta, GA) during the tactile intervention of the educational protocol.
Changes in Muscle Thickness After Exercise 309
Figure 1 — CONSORT flowchart.
Subjective Measures Low Back Impairment. Low back impairment was measured using a 6-point scale to rate current levels of disability or impairment through a series of 10 questions related to activities of daily living, pain intensity, traveling, social activities, and the changing degree of pain. Reportable answers ranged from zero, indicating low to minimal disability or impairment, to 5, corresponding to the greatest impairment or disability for each item. Percentage of impairment was calculated by taking the total score reported and dividing the total score possible, similar to other subject questionnaires that measure disability in patients with LBP.38,39 Fear-Avoidance Beliefs. The Fear Avoidance Beliefs Questionnaire (FABQ) consists of 16 questions that are divided into 2 subgroups to measure pain responses to daily living activities.40 The FABQ was originally developed to identify behavioral changes to LBP focusing on physical activity and work and has good test reproducibility.40,41 Scoring for each subscale was determined by combining the total response scores reported for each subsection.
Visual Analogue Scale. Pain was measured using an 11-point Likert scale with zero indicating no pain and 10 maximal pain. Patients completed the visual analogue scale based on current pain levels before testing.
Procedures Participants reported to the university research laboratory for a single session. After screening, locations of the TrA and LM muscles were identified using ultrasound in the tabletop position (Figure 2), followed by recording muscle activity at rest standing and during a functional task (Figure 3). Patients were then allocated to receive a single session of exercises with or without biofeedback followed by repeated measures of TrA and LM activation in standing (Figure 1). Tabletop Measures. Both TrA and LM measurements
were assessed in the tabletop position before testing to identify proper muscle identification and to ensure proper technique of the ADIM during the intervention period. The TrA was measured supine on a table with patients’ legs in a crook-lying position. The ultrasound transducer was originally placed transversely superior to the iliac
310 Partner et al
Figure 2 — Ultrasound-transducer placement and musclethickness measurements in tabletop positions: (A) transversus abdominis ultrasound placement, (B) transversus abdominis and lateral abdominal-wall-thickness measurements, (C) lumbar multifidus ultrasound placement, and (D) lumbar multifidus thickness measurement.
Figure 3 — Testing measurement for the transversus abdominis and lumbar multifidus muscles: (A) resting measure for transversus abdominis, (B) contracted measure for transversus abdominis, (C) resting measure for lumbar multifidus, and (D) contracted measure for lumbar multifidus.
crest approximately 10 cm lateral from the umbilicus42 along the midaxillary line.13,19,43,44 Standardization for final ultrasound-transducer placement occurred when the fascial border of the TrA and thoracolumbar fascia was visible on the lateral side of the ultrasound screen and the best images of the TrA, external oblique, and internal oblique were obtained (Figure 2).13,15,43,44 Resting images were obtained through a static image immediately after exhalation.43,44 Thickness measures for the TrA were reported in centimeters and calculated at the thickest position on the ultrasound screen between the superior and deep fascial borders,13,19,43 while the total lateral abdominal-wall thickness was measured from the deep border of the TrA to the superficial border of the external oblique.13,19,43 Two markings were placed opposite the transducer head against the skin to reference the starting transducer position (Figure 2). Transducer placement for the LM occurred longitudinally at the L4–L5 level with the patient in the prone position.24,45 The patient’s arms were positioned at the sides for tabletop measures, and the ultrasound transducer was rotated medially and toward the spine until the L4– L5 facet joint was visible on the ultrasound screen.24,45 Measurements were recorded at the L4–L5 facet, as differences have been seen at the L5 level between healthy and LBP patients.46 Two lines were placed on the opposite side of the transducer head to identify starting reference location for the standing measurements. Muscle thickness was obtained at the L4–L5 facet and measured as the distance in centimeters between the zygapophyseal joint and the superficial facial border of the LM with the subcutaneous tissue (Figure 2).24,45,47–49 Standing Measures. Resting standing measurements of the TrA and LM were taken bilaterally with participants positioned in quiet standing with their arms at their sides (Figure 3).27 Ultrasound-transducer placements for the TrA and LM were applied using the markings from the tabletop positions. Measurements for the TrA occurred when the fascial border of the TrA was on the lateral portion of the ultrasound screen, as previously measured in the tabletop position. LM-muscle images were obtained with the L4–L5 facet joint on the ultrasound screen, similar to the prone position. Contracted measurements were recorded without the addition of a voluntary contraction during 1 of 2 functional tasks (Figure 3). Functional measurements of the contracted TrA were recorded on the stance limb during single-leg stance while individuals maintained approximately 90° of hip and knee flexion on the non-weight-bearing leg. Functional measurements of the contracted LM were obtained on the trailing limb during a forward-stepping task while holding the trailing foot plantar flexed in the toe-off position. Images were recorded in the same positions bilaterally. Subjects were given instructions of the functional positions, but no additional instructions or feedback was administered during the functional tasks. Intervention. Participants performed exercises in isolation or exercises with biofeedback that may be commonly used during clinical rehabilitation. Exercises
Changes in Muscle Thickness After Exercise 311
between the 2 groups were paired, so both groups completed similar workloads. Participants in the exerciseonly group completed tabletop TrA exercises in both prone and supine positions and were verbally instructed in the ADIM consistent with previous techniques27 by pulling the umbilicus toward the table during exhalation. No additional verbal feedback was given beyond the instruction technique before exercise. Participants completed a total of 15 ADIM repetitions with a 5-second hold in both positions. LM exercises consisted of raising the coccyx toward the ceiling while in a prone position and performing a prone active leg raise for a total of 15 repetitions with a 5-second hold time during each repetition for each exercise. The exercise-with-biofeedback group received biofeedback from a combination of 3 different learning tactics: verbal, tactile, and visual feedback during exercise. During TrA exercises, patients were given visual biofeedback through ultrasound imaging and verbal biofeedback consisting of additional auditory cues during exercise on breathing and focusing contractions on the lower abdominal area simultaneously while contracting the TrA. Tactile biofeedback was provided while in a prone position with the use of a pressure cuff during TrA contraction. The pressure cuff was inflated to 80 mm Hg, and each participant was instructed to perform an ADIM, decreasing the cuff between 6 and 8 mm Hg and then maintaining the pressure during the contraction. Patients completed both exercises for 15 repetitions each with 5 seconds duration. LM verbal, visual, and tactile biofeedback were administered in the prone position. Tactile feedback was administered during the prone active leg raise against resistance, where the tester put one hand over the LM so the participant could feel the contraction, and as the tester provided a small distraction on the leg instructing the participant to pull the hip back into the socket. Both of these were performed 7 times on each leg with durations of 5 seconds. Verbal and visual biofeedback using ultrasound were provided as the patients tried to lift their coccyx to the ceiling for 15 repetitions each lasting 5 seconds. The exercises for both groups are outlined in Table 1.
Data Processing Preintervention and postintervention images of the TrA and LM in the standing positions were visualized and measured using Image J image-processing software. Before thickness measurements, images were scaled using a 1-cm reference line to determine the number of pixels measured. This length was converted to a measurement scale that could be used to measure muscle thickness. Muscle thicknesses from 3 images during quiet standing and respective functional-movement conditions were calculated bilaterally and averaged by a blinded data processor.48 Images from both sides were averaged for statistical analysis. Muscle thickness of the TrA, LM, and total lateral were recorded in both resting and contracted states. Percentages TrA-thickness ratios at rest while supine and standing were calculated by observing the percent of the TrA thickness to the total lateral abdominal wall,50 and preferential activation ratios13 were calculated using the following formulas: Percent TrA thickness = TrARest/(TrA + IO + EO)Rest TrA preferential activation = TrAContracted/(TrA + IO + EO)Contracted – TrARest/(TrA + IO + EO)Rest Total thickness of the LM muscle was calculated in the resting state and during a forward-step maneuver. Percent LM-muscle-thickness changes were calculated for the total thickness with the following formula17:
LM thickness change = (LMContractedThickness – LMRestingThickness)/LMRestingThickness × 100
Statistical Analysis All statistical analyses were performed using SPSS version 19.0 (SPSS, Inc, Chicago, IL). Descriptive statistics were compared using independent t tests for continuous outcomes and Mann-Whitney U tests for categorical variables. Preliminary data diagnostics revealed that TrA
Table 1 Exercises for Intervention Exercise Alone Muscle
Exercise With Biofeedback
Maneuver
Routine
Maneuver
Routine
ADIM supine ADIM prone
15 reps 5 s 15 reps 5 s
ADIM supinea,b Pressure-cuff ADIMc
15 reps 5 s 15 reps 5 s
Sacrum to ceiling PALR
15 reps 5 s 15 reps 5 s each leg
Sacrum to ceilinga,b PALRc Hip to socketc
15 reps 5 s 7 reps 5 s each leg 7 reps 5 s each leg
Transversus abdominis
Lumbar multifidus
Abbreviations: ADIM, abdominal drawing-in maneuver; PALR, prone active leg raise. a Verbal biofeedback. b Visual biofeedback. c Tactile biofeedback.
312 Partner et al
PAR and LM percent-change data were nonnormally distributed. Nonparametric statistics were used for both within and between-groups analyses. Resting muscle thickness was compared using a Friedman test with pairwise comparisons for significant findings. Betweengroups differences were analyzed using 4 separate MannWhitney U tests (standing TrA PAR preexercise, standing TrA PAR postexercise, standing LM percent change in muscle thickness preexercise, and standing LM percent change in muscle thickness postexercise). Within-group preexercise to postexercise differences of the standing TrA PAR and LM percent change in muscle thickness were analyzed using Wilcoxon-sign ranked tests (TrA PAR, LM percent change in muscle thickness for all patients, TrA PAR and LM percent change for exercise only and exercise with biofeedback groups) along with percent TrA thickness ratios while supine and standing. The level of significance was set a priori at P ≤ .05. Effect sizes were calculated using Cohen d with 95% confidence intervals (CI) and interpreted as weak (.8)
Results A total of 20 patients (11 men, 9 women; age 23 ±4 y, height 175.26 ±12.90 cm, mass 74.64 ±15.69 kg) completed the single exercise session (Table 2). Two patients were not included in the final analysis due to 1 not completing the exercise training and a second due to ultrasound imaging error. No differences were observed between groups for age, height, mass, body-mass index, visual analogue scale, low back impairment, or FABQ subscale scores; however, there was a difference in the number of men and women per group (Table 2). Painrecurrence frequency and duration were reported for 19 of the 20 subjects and are listed in Table 3. Overall, patients with a history of LBP reported a median pain level of 2
ranging from 0 to 3 on a visual analogue scale, along with reported impairment of 22.7% (Table 2).
Resting Muscle Thickness Resting muscle thickness for both table and standing measures can be found in Table 4. There were no differences between table and standing measures before or after exercise for the TrA (χ2 = 2.880, P = .410) or LM (χ2 = 4.980, P = .173) muscles for all patients. A significant decrease was observed in resting muscle thickness for the total lateral wall (χ2 = 12.180, P = .007). Pairwise comparisons showed lower resting muscle thickness during both standing before exercise (P = .013) and standing after exercise (P = .042) than tabletop measures before exercise. The TrA accounted for a larger portion of the total lateral abdominal-wall thickness in the standing position both before exercise (Z = –2.725, P = .006) and after exercise (Z = –2.613, P = .009) compared with the table measures (Table 4).
TrA PAR Table 5 shows the TrA PAR values before and after exercise. There were no between-groups differences in TrA PAR at baseline (Z = –1.134, P = .257) or after exercise
Table 3 Reported Pain-Symptom Duration and Frequency, N = 19 Pain frequency
n
Pain duration
n
Daily Weekly Monthly Annually
2 11 4 2
All day 12+ h 1–12 h
