239
Journal of Back and Musculoskeletal Rehabilitation 28 (2015) 239–246 DOI 10.3233/BMR-140509 IOS Press
Discriminant analysis of neuromuscular variables in chronic low back pain Denise Martineli Rossia,∗, Mary Hellen Morcellia , Adalgiso Coscrato Cardozob, Benedito Sérgio Denadaib , Mauro Gonçalvesb and Marcelo Tavella Navegaa,b a
b
Department of Physical Therapy and Occupational Therapy, São Paulo State University, Marília, Brazil Department of Physical Education, São Paulo State University, Rio Claro, Brazil
Abstract. BACKGROUND: Investigation and discrimination of neuromuscular variables related to the complex aetiology of low back pain could contribute to clarifying the factors associated with symptoms. OBJECTIVE: Analysing the discriminative power of neuromuscular variables in low back pain. METHODS: This study compared muscle endurance, proprioception and isometric trunk assessments between women with low back pain (LBP, n = 14) and a control group (CG, n = 14). Multivariate analysis of variance and discriminant analysis of the data were performed. RESULTS: The muscle endurance time (s) was shorter in the LBP group than in the CG (p = 0.004) with values of 85.81 (37.79) and 134.25 (43.88), respectively. The peak torque (Nm/kg) for trunk extension was 2.48 (0.69) in the LBP group and 3.56 (0.88) in the GG (p = 0.001); for trunk flexion, the mean torque was 1.49 (0.40) in the LBP group and 1.85 (0.39) in the CG (p = 0.023). The repositioning error (degrees) before the endurance test was 2.66 (1.36) in the LBP group and 2.41 (1.46) in the CG (p = 0.664), and after the endurance test, it was 2.95 (1.94) in the LBP group and 2.00 (1.16) in the CG (p = 0.06). Furthermore, the variables showed discrimination between the groups (p = 0.007), with 78.6% of the individuals with low back pain correctly classified in the LBP group. In turn, variables related to muscle activation showed no difference in discrimination between the groups (p = 0.369). CONCLUSION: Based on these findings, the clinical management of low back pain should consist of both resistance and strength training, particularly in the extensor muscles. Keywords: Torque, physical endurance, electromyography, abdominal muscles
1. Introduction Low back pain is the leading cause of absenteeism and musculoskeletal dysfunction in adults younger than 45 years of age, resulting in high economic costs to health and social security systems [1]. Low back pain is nonspecific in 75% to 85% approximately of cases [1]. This symptom occurs quite frequently in younger populations, in which some individuals de∗ Corresponding author: Denise Martineli Rossi, São Paulo State University, Higyno Muzzi Filho, No: 737, University Campus – 17.525-900 – Marília, Brazil. Tel.: +55 1434021300; Fax: +55 1434021302; E-mail: [email protected].
velop a chronic condition with recurrent complaints over time [2]. Muscle fatigue stands out among the possible causes of musculoskeletal disorders of the spine. There are direct relationships of muscle fatigue with the isometric endurance of the erector spinae muscles and endurance time on the Biering-Sorensen test [3,4]. Shorter isometric endurance times in individuals with low back pain have been associated with changes in EMG parameters, such as a decrease in median frequency [5]. It has been suggested that fatigue of the erector spinae muscles impairs the ability to feel changes in the lumbar position, and proprioceptive dysfunction could be related to chronic low back pain [6]. However, the methodological diversity of studies investigating pro-
c 2015 – IOS Press and the authors. All rights reserved ISSN 1053-8127/15/$35.00
240
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
prioception has generated conflicting results, leaving the literature inconclusive regarding this issue [7]. Isometric muscular endurance and muscle strength are useful tools for measuring the contractile state of the trunk muscles. The literature has shown a deficit in trunk muscle strength in patients with chronic low back pain [3,8]. Yahia et al. [8] suggested that subjects with low back pain showed lower peak torque of the trunk flexor and extensor. The peak torque of the extensor muscles was selectively deficient and approximately 30% to 50% lower than in asymptomatic subjects [8]. Neuromuscular control is essential for strength generation during the performance of motor tasks and posture maintenance [9]. Individuals with low back pain have strategic compensation of the abdominal muscles in situations of postural imbalance [9,10] and of the posterior trunk muscles during isometric extension [10], to maintain spinal stability and produce torque [7]. However, despite knowledge about the changes in physical abilities in low back pain, little attention has been paid to the influence of variables such as isometric endurance, muscle strength, proprioception and muscle activation in this musculoskeletal dysfunction. In this regard, investigating the discriminative power of each neuromuscular variable associated with this symptom could contribute to clarifying cause-andeffect relationships and could assist in developing intervention programs. The aims of this study were to evaluate and verify the discriminative power among subjects with and without symptoms such as isometric muscle strength and the effects of muscle fatigue on trunk proprioception, peak torque of trunk flexor and extensor, muscle activation behavior at the time of peak torque, such as antagonist activation and the balance of muscle activation amplitude between body sides. Thus, the hypothesis was that subjects with low back pain would perform poorly on endurance and muscle strength evaluations. Moreover, it was expected that reduction in physical abilities would be associated with proprioceptive deficits and changes in trunk muscle activation, with increased antagonist activation and muscle activation imbalance at the moment of peak torque in subjects with chronic low back pain. 2. Method 2.1. Subjects This case-control study included 28 female subjects, recruited from a university population, between 18 and
30 years old and classified as physically active (active/very active) through the application of the International Physical Activity Questionnaire (IPAQ). The sample size was estimated with an effect size of 1.04, probability error of α = 0.05 and power of 0.85. Thus, the subjects were divided into two groups: low back pain (LBP) (n = 14) and a control group (CG) (n = 14). The inclusion criteria for the LBP group were reports of low back pain for at least six months prior to the study. The exclusion criteria were neurological symptoms, discrepancy between limbs greater than 2 cm, rheumatoid arthritis, herniated discs, tumours or vertebral fractures [3]. The exclusion criteria for the CG were the presence of low back pain or absence from work/school due to pain. All subjects in both groups, the LBP group and CG, showing body mass index (BMI) greater than 29.9 or who had prior pelvis or spine surgery were excluded from the sample. This study was approved by the Local Ethics Research Committee, and all participants signed informed consent form. 2.2. Procedures Data collection was performed on two days. On the first day, body mass, height, leg length, physical activity level (IPAQ), disability level (Roland Morris Disability Questionnaire) and pain level (Visual Analogue Scale – VAS) were measured. Later, the volunteers were submitted to trunk proprioceptive evaluations. Then, they were evaluated for resistance of the trunk extensor muscles, and immediately thereafter, trunk proprioceptive evaluation was performed again. On the same day, the subjects performed familiarisation with an isometric protocol for trunk flexion and extension. On the second day, only the isometric evaluation protocol was repeated, and at the same time, the electromyographic signals of the rectus abdominal (RA), internal oblique (IO), multifidus (MU) and longissimus thoracis (LT) muscles were obtained. Synchronisation of the dynamometric and electromyographic data was performed through plate receiver synchronisation G2 TM 2400 Analog Input Board (Noraxon , Scottsdale, Arizona, USA). 2.2.1. Pain assessment Pain intensity was verified using the VAS, with the subject indicating the pain intensity prior to the performance of tests. The duration of pain in years was also evaluated.
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
241
vice when they perceived the previously familiarised position [12]. The equipment provided a measurement, in degrees, of the position at which each volunteer triggered the device and the distance from the target position. Each volunteer performed test twice, and the mean of the two measurements was used [12].
Fig. 1. Isokinetic dynamometer evaluation position.
2.2.2. Isokinetic dynamometer Assessments in the Isokinetic Dynamometer System 4 PRO were performed by coupling a special chair (Dual position Back Ex/Flex Attachment) in the Seated-Compressed mode (Fig. 1). Thus, the subjects remained seated and were stabilised by means of a device for maintaining the knee at 90◦ , and special belts were fixed on the chest, in the region of hip joint and in the middle third of the thigh. The mechanical axis of the dynamometer was aligned with the anterior superior iliac spine, as recommended by the equipment manufacturer. 2.2.3. Proprioceptive assessment Trunk proprioception was evaluated using an isokinetic dynamometer for a test of passive motion reproduction, in which the individual remained blindfolded. From the sitting position with 90◦ of hip flexion, the dynamometer moved at an angular speed of 1◦ /s in the trunk flexion direction over a total range of 20◦ . The volunteers were asked to memorise this position, which was maintained for ten seconds. Then, the chair returned to its initial position. From this time point, the equipment moved again at the same speed and in the same direction, and the volunteers were instructed to stop the amplitude by moving a manual de-
2.2.4. Muscle endurance assessment Muscle endurance assessment was performed using the Biering-Sorensen test. In this test, the subject was kept in the prone posture and was firmly stabilised with belts at the levels of the hips, knees and feet. The upper trunk remained suspended at the level of the ASIS, with the upper limbs crossed over the trunk (Fig. 2). To assess the resistance of the trunk extensor muscles, the test measured the time that the subject was able to maintain the upper body in the horizontal position [3]. The volunteers were instructed to keep the trunk in the horizontal position, which was controlled by an electrogoniometer (Inline1D/2D Electrical Goniometer Noraxon , Scottsdale, Arizona, USA) positioned on the mid-axillary line of the trunk at the ASIS level, and a monitor showed the trunk angle as a form of feedback (Fig. 2). The subjects were instructed to maintain the position for as long as possible. The criterion adopted for test interruption was inability to maintain the horizontal position of the trunk. A threshold amplitude of 10◦ trunk movements was determined, and if the subject exceeded this limit or if she voluntarily lowered her trunk due to an inability to maintain the required position, the test was stopped [3,4]. 2.2.5. Isometric muscle torque assessment Trunk isometric muscle torque was assessed on two different days, with the values obtained from the second evaluation used for analysis. To perform isometric contraction, the trunk was kept at 60◦ of flexion, and the volunteers performed six isometric contractions held for five seconds each: three for trunk flexion and three for trunk extension. A 30-second rest interval was provided between each contraction. Verbal encouragement during muscle contraction was standardised. Trunk isometric muscle torque was synchronously assessed using electromyography by means of a TM 2400 G2 receiver Analog Input Board synchronisation plate (Noraxon , Scottsdale, Arizona, USA). 2.2.6. Electromyography assessment Electromyographic signals were obtained using a 16-channel biological signal acquisition module on
242
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
Fig. 2. Biering-Sorensen test position.
a Telemyo 900 Myoresearch telemeter (Myoresearch Electromyograph Noraxon , Scottsdale, Arizona, USA) (common mode rejection 100 dB, input impedance greater than 10 MOhm and base noise less than 1 UV RMS), Myoresearch software (Noraxon ), connected to a 16-bit analogue-digital converter calibrated with a sampling frequency of 2000 Hz and total gain of 2000 times (20 times on the sensor and 100 times on the machine). Ag/AgCl surface electrodes in bipolar configuration, 1 cm in diameter and at a distance of 2 cm from each other, were used. The skin was shaved and cleaned with alcohol prior to the placement of the electrodes to avoid potential interference with the electromyographic signal [13]. The electrodes were placed bilaterally (right and left sides) on the global or multisegmental trunk muscles: the rectus abdominis (RA), 3 cm above the umbilicus and 2 cm lateral to the midline, vertically oriented; and the longissimus thoracic (LT), vertically oriented at 2 cm lateral to the midline from spinous process of L1. On local or unisegmental trunk muscles, the electrodes were placed as followed: internal oblique (IO), 2 cm medially and inferiorly to the anterior superior iliac spine [14] and the multifidus (MU), 2 cm lateral to the midline toward spinous process L5, oriented towards the line between the interspinous space of L1L2 and the posterior inferior iliac spine, according to SENIAM [15] (Hermens et al., 2000). The reference electrode was positioned at the radial styloid process. 2.3. Data analysis 2.3.1. Data processing Isometric assessment analysed the torque and electromyographic signal data through specific routines developed in Matlab software (Mathworks 7.0 Natick, Massachusetts, USA). The torque signal was processed with a 4th order Butterworth digital filter with a cutoff frequency of 15 Hz. Thus, peak torque in trunk flexion and extension was observed. The torque measurements obtained were normalised based on the body mass of
each subject to consider the differences between individuals and to ensure more precise functional muscle performance. The analysis of muscle activation at peak torque (PT) used a window of 100 ms (50 ms before and 50 ms after PT) for trunk flexion and extension. The EMG signal was analysed in time domain by means of root mean square (RMS) values. A Butterworth 4th order digital bandpass filter of 20–500 Hz was used. In trunk flexion, RMS of the right and left muscles’ IO and RA were calculated, and for trunk extension movement, RMS of the right and left MU and LT muscles were calculated. The RMS values were normalised using the RMS peak of maximal isometric contractions. Furthermore, the ratio of activation amplitude, measured by RMS at peak torque between bilateral trunk muscles, was also determined. In other words, during trunk flexion, the RMS ratio was calculated for RA right and RA left and for IO right and IO left, and this calculation was also performed for the trunk extensor muscles in extension movement. Obtaining the ratios of bilateral muscles allowed for an analysis of the equilibrium amplitude of trunk muscle activation, in which a value close to one indicated good muscular balance. For antagonist activation during isometric evaluation, the amplitude of EMG signal was calculated by means of RMS values at the peak torque of antagonistic muscles. In other words, for trunk flexion, RMS of the MU and LT muscles were bilaterally calculated, and for trunk extension, RMS of the RA and IO muscles were bilaterally calculated at peak torque. 2.3.2. Statistical analysis Data analysis was performed using SPSS Inc. statistical software package. Data normality was verified using the Shapiro-Wilk test, and multivariate analysis of variance and discriminant analysis of data were subsequently performed to investigate the discriminating power of variables between groups with and without low back pain. For intra-group analysis of antagonist activation variables, the paired t-test for normality of assumed
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
243
Table 1 Mean and standard deviation of age, anthropometric data, Biering-Sorensen test, proprioception through passive repositioning error, extensor and flexor peak torque normalized according to body mass of the low back pain (LBP) and control group (CG) Age (years) Mass (kg) Height (m) BMI (kg/m2 ) Biering Sorensen (s) RE before (◦ ) RE after (◦ ) PT Extensor (Nm/kg) PT Flexor (Nm/kg) ∗p
LBP (n = 14) 24.14 (3.13) 61.68 (7.19) 1.66 (0.05) 22.31 (2.12) 85.81 (37.79) 2.66 (1.36) 2.95 (1.94) 2.48 (0.69) 1.49 (0.40)
CG (n = 14) 22.21 (3.40) 58.2 (8.73) 1.61 (0.06) 22.23 (1.98) 134.25 (43.88) 2.41 (1.46) 2.00 (1.16) 3.56 (0.88) 1.85 (0.39)
p 0.131 0.256 0.044∗ 0.912 0.004∗ 0.664 0.062 0.001∗ 0.023∗
F − − − − 9.791 0.193 3.799 13.178 5.846
< 0.05, LBP different of the CG. Legend: Repositioning Error (RE), Peak Torque (PT), Body Mass Index (BMI).
difference and Wilcoxon’s test for normality of nonassumed difference were used. The significance level adopted was 0.05.
3. Results The LBP group showed a pain onset time of 4.85 (3.73) years and pain intensity of 12.84 (17.05) mm on the VAS. Disability level was measured by the Roland Morris Disability Questionnaire, on which zero points indicates no disability, and 24 points indicates maximum disability. Regarding disability, the LBP group had a mean of 3.46 (2.50) points. Table 1 shows comparisons between the groups regarding age, anthropometric data and some independent variables, such as proprioception through passive repositioning error, muscle fatigue assessed by the Biering-Sorensen test and extensor and flexor peak torque normalised according to body mass. The LBP group had greater mean height (m) values, compared to the control group (p < 0.05). Multivariate analysis of variance identified significant differences (p = 0.008, F = 3.761) between the CG and LBP group for independent variables, as shown in Table 1. The results of the Biering-Sorensen test showed differences between groups, with the LBP group showing a 36.08% shorter endurance time (p < 0.05) than the CG. In contrast, regarding proprioception, represented by the error values of passive trunk repositioning, there were no differences between individuals with and without low back pain, either before or after the Biering-Sorensen test (Table 1). Peak isometric extension and flexion in the CG were greater than in the LBP group: 30.33% and 19.45%, respectively (Table 1). RMS values of the IO, RA, MU and LT muscles at peak torque showed no significant differences (p =
Fig. 3. Antagonist activation of the trunk between control group (GC, n = 14) and low back pain group (LBP, n = 14).
0.546, F = 0.991) between the groups for either trunk flexion or extension. Only for the left LT muscle during extension were there differences between the groups (p < 0.05) (Table 2). Regarding the activation ratio between the muscles of the right and left hemispheres during peak torque, there were no differences between the groups (p > 0.05), as shown in Table 2. However, antagonist activation of the trunk showed inter-group and intra-group differences for comparisons between local and global muscles. Figure 3 shows differences between the groups in antagonist activation only for local extensor muscles: bilaterally, the right MU (p = 0.027, F = 5.509) and left MU (p = 0.048, F = 4.309) during trunk flexion. The
244
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
Table 2 RMS values of the IO, RA, MU and LT muscles at peak torque and activation ratio between the muscles of the right and left hemispheres during peak torque. Mean and standard deviation values
RMS (µV) Extension
Flexion
Ratio Extension Flexion
CG (n = 14) Mean (SD)
LBP (n = 14) Mean (SD)
p
F
Right LT Left LT Right MU Left MU
64.1 (20.04) 71.29 (20.56) 59.93 (21.47) 72.44 (19.2)
54.1 (22.85) 54.85 (15.84) 57.31 (23.87) 61.44 (27.08)
0.229 0.025# 0.763 0.226
1.517 5.617 0.093 1.536
Right RA Left RA Right IO Left IO
67.82 (13.41) 63.64 (22.04) 60.06 (21.2) 64.15 (13.56)
63.98 (18.97) 62.37 (20.33) 60.4 (18.94) 62.98 (18.91)
0.541 0.876 0.965 0.852
0.383 0.025 0.002 0.036
LT MU
0.91 (0.27) 0.85 (0.34)
0.99 (0.4) 0.98 (0.32)
0.569 0.300
0.332 1.116
RA IO
1.22 (0.65) 0.99 (0.41)
1.07 (0.3) 0.99 (0.4)
0.422 0.967
0.422 0.967
Longissimus thoracic (LT), multifidus (MU), rectus abdominais (RA) and internal oblique (IO). Control group (CG) and low back pain (LBP) group. #p < 0.05 for CG higher than LBP.
CG showed, on average, 40% greater antagonist activation of extensor muscles (MU, bilaterally) during trunk flexion, compared to the LBP group. In intra-group analysis (local X global), the CG showed increased activation of antagonist muscles, the right MU (p = 0.003) and left MU (p = 0.009), in trunk flexion, compared to the right and left LT muscles, respectively. The LBP group showed increased antagonist activation of right IO and left IO muscles during extension, which were 52% and 44% greater than that of the right RA and left RA, respectively. In trunk flexion, the LBP group also showed greater antagonist activation, with RMS of the right MU muscle 35% higher than that of the right LT (p = 0.005) and the left MU 27% higher than that of the left LT (p = 0.001). Discriminant analysis showed significant discrimination between the groups (WL = 0.347, χ2 = 22.745, df = 9, p = 0.007) for endurance, proprioception, extensor and flexor peak torque variables, in which 78.6% of individuals with low back pain were correctly classified in the LBP group, and 71.4% of asymptomatic individuals were correctly classified in the CG. Furthermore, the variables with greater discriminative power were, in descending order, extensor peak torque, endurance time, and flexor peak torque. Regarding variables related to muscle activation, such as the RMS of each muscle at peak torque, the ratio between the RMS of the bilateral trunk muscles and antagonist activation, discriminant analysis showed no significant differences between the groups (WL = 0.261, χ2 = 21.49, df = 20, p = 0.369). Thus,
muscle activation variables, such as amplitude, ratio and antagonist activation, were able to classify only 57.1% of young individuals with pain in the LBP group and 42.9% of young people without back pain in the CG.
4. Discussion Discrimination of neuromuscular variables in a nonspecific musculoskeletal dysfunction with a multifactorial aetiology could contribute, among other factors, to clarification of cause-and-effect relationships. It was found that isometric endurance time and the peak torque of the trunk extensors showed greater discriminatory power between individuals with and without this symptom. However, muscle activation variables, such as the activation amplitude at peak torque, activation ratio and antagonist activation, did not contribute to discrimination or correct classification of each subject within the groups. As noted in the literature [3,4], in this study, the Biering-Sorensen test showed accuracy in diagnosing individuals with and without low back pain. The endurance time of asymptomatic individuals was, on average, 134 seconds, in agreement with Johnson et al. [16], who found a mean value of 138 seconds for healthy adolescents. In association with physiological changes in the muscle fatigue process, such as changes in the speed conditions of muscle fibres due to lactate accumulation [5], psychological influences are also related to
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
lower resistance of the extensor muscles in populations with back pain. Factors such as motivation, discomfort tolerance related to muscle fatigue and, particularly in clinical situations, pain or fear of pain could influence this outcome [17]. Anthropometric factors, such as BMI, can interfere with performance of the Biering Sorensen test [16]. In this study, BMI was not different between the groups, but height in the LBP group was greater than in the CG. Johnson et al. [16] suggested a lack of correlation between isometric endurance time and height, so the difference in height between the LBP group and CG did not influence the results. The variable with the greatest discriminating power demonstrated in this study was the isometric peak torque of the trunk extensor. Data from this study corroborated the findings of Yahia et al. [8], who showed impaired muscle torque of the trunk extensor and flexor in individuals with chronic low back pain, with greater differences between groups for the peak torque of the trunk extensor. The lower peak torque in trunk flexion and extension movements in the LBP group could be related to fear of feeling pain or fear of re-injury associated with vertebral movements. Thus, the anticipation of pain or fear of inducing pain can result in inability or unwillingness to exert the maximum effort, resulting in lower torque values [18]. Similarly, neural adaptations and changes in motor control, such as delay in activation onset during balance disturbances in individuals with low back pain [9,10], which persist even after the elimination of painful stimuli, have also been indicated as contributors to weakness of the trunk muscles [19]. The current study did not identify differences in trunk proprioception between the groups and similarly found no effects of extensor muscle fatigue induced by the Biering-Sorensen test on the ability to reproduce a pre-determined angle of the trunk. Lee et al. [20] suggested that active and passive repositioning tests were less sensitive for identifying proprioceptive deficits because there was an influence of memory on the task, which in turn varied according to amount of time between the presentation of the target position and movement execution. It has also been suggested that the differences found in the literature are due not only to different assessment protocols used but also to the ages of the participants. Thus, proprioception has been shown to be more impaired in older subjects (mean age of 40 years old) compared to younger subjects (mean age of 19 years old) as a result of the effects of the natural aging process on neural control and proprioception [7, 20].
245
In contrast with evidence found in the literature regarding changes in the responses of trunk muscle activation in individuals with low back pain in cases of postural imbalance or in cases of rapid trunk flexion [21], in this study, changes were not observed during the task of producing maximum force in trunk flexion and extension. The LBP group did not exhibit different patterns of muscle activation variables. Thus, muscle activation variables showed poor discriminative power, with only 42.9% correct classification. However, Andersen et al. [22] suggested that, under conditions of chronic musculoskeletal pain, the ability to activate muscles quickly and to generate strength might be more severely impaired than the maximum amplitude of muscle activation and the capacity to produce maximum force. Due to anatomical and functional differentiation of the trunk muscles, there is clinical importance in analysing the segmental and multi-segmental erector and abdominal muscles of individuals with low back pain [23]. Thus, intra-group comparisons were also performed for antagonist activation amplitude. The LBP group showed higher local antagonist activation than globally in the trunk flexion and extension movements. Jones et al. [24] suggested that individuals with low back pain in situations of postural imbalance increased trunk coactivation regardless of disturbance direction to compensate for a lack of accurate proprioceptive information, aiming to restrict movement to avoid pain. As a result, trunk coactivation leads to a lower capacity to produce torque in these individuals [24].
5. Conclusion Among the neuromuscular variables investigated, extensor peak torque, muscle endurance, pain intensity and flexor peak torque showed significant discrimination and correct classification of individuals within each group. However, trunk proprioception and the muscle activation amplitude variables showed poor discriminative power. Based on these findings, clinical treatment of young individuals with low back pain should include strength training and endurance, particularly in the extensor muscles.
Acknowledgement D.M. Rossi was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
246
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
References [1]
Posadzki P, Lizis P, Hagner-Derengowska M. Pilates for low back pain: A systematic review. Complementary Therapies in Clinical Practice. 2011; 17: 85-89. doi: http://dx.doi. org/10.1016/j.ctcp.2010.09.005. [2] Smith DR, Leggat PA, 2007. Back pain in the young: A review of studies conducted among school children and university students. Current Pediatric Reviews. 2007; 3(1): 69-77. doi: 10.2174/157339607779941624. [3] Gruther W, Wick F, Paul B, Leitner C, Posch M, Matzner et al. Diagnnostic accuracy and reliability of muscle strength and endurance measuremenfts in patients with chronic low back pain. Journal of Rehabilitation Medicine. 2009; 41: 613-619. doi: 10.2340/16501977-0391. [4] Latimer J, Maher CG, Refshauge K, Colaco L. The reliability and validity of the Biering-Sorensen test in asymptomatic subjects and subjects reporting current or previous nonspecific low back pain. Spine. 1999; 24: 2085-2090. doi: 10.1097/ 00007632-199910150-00004. [5] Biedermann HJ, Shanks, GL, Forrest WJ, Inglis J. Power spectrum analyses of electromyographic activity: Discriminators in the differential assessment of patients with chronic low back pain. Spine. 1991; 16(10): 1179-1184. [6] Ribeiro F, Mota J, Oliveira J. Effect of exercise-induced fatigue on sense of the knee in the elderly. Journal of Applied Physiology. 2007; 99: 379-385. doi: 21-006-0357-8. [7] Silfies SP, CholewickI J, Reeves NP, Greene HS. Lumbar position sense and the risk of low back injuries in college athletes: A prospective cohort study. BMC Musculoskelet Disorders. 2007; 8(129): 1-7. doi: 10.1186/1471-2474-8-129. [8] Yahia A, Jribi S, Ghroubi S, Elleuch M, Baklout S, Elleuch MH. Evaluation of the posture and muscular strength of the trunk and inferior members of patients with chronic lumbar pain. Joint Bone Spine. 2011; 78: 291-297. doi: http://dx.doi. org/10.1016/j.jbspin.2010.09.008. [9] Newcomer KL, Jacobson TD, Gabriel DA, Lardson, DR, Brey RH, An K. Muscle activation patterns in subjects with and without low back pain. Archives of Physical Medicine and Rehabilitation. 2002; 83(6): 816-821. doi: http://dx.doi. org/10.1053/apmr.2002.32826. [10] Oddson LIE, De Luca CJ. Activation imbalances in lumbar spine muscles in the presence of chronic low back pain. Journal of Applied Physiology. 2003; 94: 1410-1420. doi: 10.1152/ japplphysiol.01183.2001. [11] Larivière C, Arsenault AB, Gravel D, Gagnon D, Loisel P. Surface electromyography assessment of back muscle intrinsic properties. Journal of Electromyography and Kinesiology. 2003; 13: 305-318. doi: http://dx.doi.org/10.1016/S10506411(03)00039-7. [12] Yilmaz B, Yaar E, Takaynatan MA, Goktepe AS, Tucu I, Yazcolu K. et al. Relationship between lumbar muscle strength and proprioception after fatigue in men with chronic low back pain. Turkey Journal Rheumatology. 2010; 25: 68-71. doi: 10.5152/tjr.2010.04. [13] Gonçalves M, Marques NR, Hallal CZ, Van dieen JH.
Electromyographic activity of trunk muscles during exercises with flexible and non-flexible poles. Journal of Back and Musculoskeletal Rehabilitation. 2012; 24: 209-214. doi: 10.3233/BMR-2011-0297. [14] Marshall P, Murphy B. The validity and reliability of surface EMG to assess the neuromuscular response of the abdominal muscles to rapid limb movement. Journal of Electromyography and Kinesiology. 2003; 13: 477-489. doi: http://dx.doi. org/10.1016/S1050-6411(03)00027-0. [15] Hermens JH, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. Journal of Electromyography and Kinesiology. 2000; 10: 361-374. doi: http://dx.doi.org/10.1016/ S1050-6411(00)00027-4. [16] Johnson OE, Mbada CE, Akosil CO, Agbeja AO. Isometric endurance of the back extensors in school-aged adolescents with and without low back pain. Journal of Back and Musculoskeletal Rehabilitation. 2009; 22: 205-211. doi: 10.3233/ BMR-2009-0235. [17] Mannion AF, Riordan D, Dvorak J, Masharawi Y. The relationship between psychological factors and performance on the Biering-Sorensen back muscle endurance test. The Spine Journal. 2011; 11: 849-857. doi: 10.1016/j.spinee.2011. 08.004. Epub 2011 Sep 8. [18] Davariana S, Maroufia N, Ebrahimia I, Farahmand F, Parnianpour M. Trunk muscles strength and endurance in chronic low back pain patients with and without clinical instability. Journal of Back and Musculoskeletal Rehabilitation. 2012; 25: 123-129. doi: 10.3233/BMR-2012-0320. [19] Tsao H, Galea MP, Hodges PW. Reorganization of the motor cortex is associated with postural control deficits in recurrent low back pain. Brain. 2008; 131: 2161-2171. doi: 10.1093/ brain/awn154. Epub 2008 Jul 18. [20] Lee AS, Cholewicki J, Reeves P, Zazulak BT, MysliwieC LW. Comparison of trunk proprioception between patients with low back pain and healthy controls. Archives of Physical Medicine and Rehabilitation. 2010; 91: 1327-1331. doi: 10.1016/j.apmr.2010.06.004. [21] Marshall P, Murphy B. Delayed abdominal muscle onsets and self-report measures of pain and disability in chronic low back pain. Journal of Electromyography and Kinesiology. 2009, doi:10.1016/j.jelekin.2009.09.005. [22] Andersen LL, Holtermann A, Jorgensen MB, Sjogaard G. Rapid muscle activation and force capacity in conditions of chronic musculoskeletal pain. Clinical Biomechanics. 2008; 23: 1237-1242. doi:10.1016/j.clinbiomech.2008.08.002. Epub 2008 Oct 2. [23] Sung PS, Lammers AR, Danial P. Different parts of erector spinae muscle fatigability in subjects with and without low back pain. The Spinal Journal. 2009; 9: 115-120. doi: 10.1016/j.spinee.2007.11.011. Epub 2008 Feb 14. [24] Jones SL, Henry SM, Raasch CC, Hitt JR, Bunn JY. Individuals with non-specific low back pain use a trunk stiffening strategy to maintain upright posture. Journal of Electromyography and Kinesiology. 2012; 22: 13-20. doi: 10.1016/j. jelekin.2011.10.006. Epub 2011 Nov 18.
Copyright of Journal of Back & Musculoskeletal Rehabilitation is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Journal of Back and Musculoskeletal Rehabilitation 28 (2015) 239–246 DOI 10.3233/BMR-140509 IOS Press
Discriminant analysis of neuromuscular variables in chronic low back pain Denise Martineli Rossia,∗, Mary Hellen Morcellia , Adalgiso Coscrato Cardozob, Benedito Sérgio Denadaib , Mauro Gonçalvesb and Marcelo Tavella Navegaa,b a
b
Department of Physical Therapy and Occupational Therapy, São Paulo State University, Marília, Brazil Department of Physical Education, São Paulo State University, Rio Claro, Brazil
Abstract. BACKGROUND: Investigation and discrimination of neuromuscular variables related to the complex aetiology of low back pain could contribute to clarifying the factors associated with symptoms. OBJECTIVE: Analysing the discriminative power of neuromuscular variables in low back pain. METHODS: This study compared muscle endurance, proprioception and isometric trunk assessments between women with low back pain (LBP, n = 14) and a control group (CG, n = 14). Multivariate analysis of variance and discriminant analysis of the data were performed. RESULTS: The muscle endurance time (s) was shorter in the LBP group than in the CG (p = 0.004) with values of 85.81 (37.79) and 134.25 (43.88), respectively. The peak torque (Nm/kg) for trunk extension was 2.48 (0.69) in the LBP group and 3.56 (0.88) in the GG (p = 0.001); for trunk flexion, the mean torque was 1.49 (0.40) in the LBP group and 1.85 (0.39) in the CG (p = 0.023). The repositioning error (degrees) before the endurance test was 2.66 (1.36) in the LBP group and 2.41 (1.46) in the CG (p = 0.664), and after the endurance test, it was 2.95 (1.94) in the LBP group and 2.00 (1.16) in the CG (p = 0.06). Furthermore, the variables showed discrimination between the groups (p = 0.007), with 78.6% of the individuals with low back pain correctly classified in the LBP group. In turn, variables related to muscle activation showed no difference in discrimination between the groups (p = 0.369). CONCLUSION: Based on these findings, the clinical management of low back pain should consist of both resistance and strength training, particularly in the extensor muscles. Keywords: Torque, physical endurance, electromyography, abdominal muscles
1. Introduction Low back pain is the leading cause of absenteeism and musculoskeletal dysfunction in adults younger than 45 years of age, resulting in high economic costs to health and social security systems [1]. Low back pain is nonspecific in 75% to 85% approximately of cases [1]. This symptom occurs quite frequently in younger populations, in which some individuals de∗ Corresponding author: Denise Martineli Rossi, São Paulo State University, Higyno Muzzi Filho, No: 737, University Campus – 17.525-900 – Marília, Brazil. Tel.: +55 1434021300; Fax: +55 1434021302; E-mail: [email protected].
velop a chronic condition with recurrent complaints over time [2]. Muscle fatigue stands out among the possible causes of musculoskeletal disorders of the spine. There are direct relationships of muscle fatigue with the isometric endurance of the erector spinae muscles and endurance time on the Biering-Sorensen test [3,4]. Shorter isometric endurance times in individuals with low back pain have been associated with changes in EMG parameters, such as a decrease in median frequency [5]. It has been suggested that fatigue of the erector spinae muscles impairs the ability to feel changes in the lumbar position, and proprioceptive dysfunction could be related to chronic low back pain [6]. However, the methodological diversity of studies investigating pro-
c 2015 – IOS Press and the authors. All rights reserved ISSN 1053-8127/15/$35.00
240
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
prioception has generated conflicting results, leaving the literature inconclusive regarding this issue [7]. Isometric muscular endurance and muscle strength are useful tools for measuring the contractile state of the trunk muscles. The literature has shown a deficit in trunk muscle strength in patients with chronic low back pain [3,8]. Yahia et al. [8] suggested that subjects with low back pain showed lower peak torque of the trunk flexor and extensor. The peak torque of the extensor muscles was selectively deficient and approximately 30% to 50% lower than in asymptomatic subjects [8]. Neuromuscular control is essential for strength generation during the performance of motor tasks and posture maintenance [9]. Individuals with low back pain have strategic compensation of the abdominal muscles in situations of postural imbalance [9,10] and of the posterior trunk muscles during isometric extension [10], to maintain spinal stability and produce torque [7]. However, despite knowledge about the changes in physical abilities in low back pain, little attention has been paid to the influence of variables such as isometric endurance, muscle strength, proprioception and muscle activation in this musculoskeletal dysfunction. In this regard, investigating the discriminative power of each neuromuscular variable associated with this symptom could contribute to clarifying cause-andeffect relationships and could assist in developing intervention programs. The aims of this study were to evaluate and verify the discriminative power among subjects with and without symptoms such as isometric muscle strength and the effects of muscle fatigue on trunk proprioception, peak torque of trunk flexor and extensor, muscle activation behavior at the time of peak torque, such as antagonist activation and the balance of muscle activation amplitude between body sides. Thus, the hypothesis was that subjects with low back pain would perform poorly on endurance and muscle strength evaluations. Moreover, it was expected that reduction in physical abilities would be associated with proprioceptive deficits and changes in trunk muscle activation, with increased antagonist activation and muscle activation imbalance at the moment of peak torque in subjects with chronic low back pain. 2. Method 2.1. Subjects This case-control study included 28 female subjects, recruited from a university population, between 18 and
30 years old and classified as physically active (active/very active) through the application of the International Physical Activity Questionnaire (IPAQ). The sample size was estimated with an effect size of 1.04, probability error of α = 0.05 and power of 0.85. Thus, the subjects were divided into two groups: low back pain (LBP) (n = 14) and a control group (CG) (n = 14). The inclusion criteria for the LBP group were reports of low back pain for at least six months prior to the study. The exclusion criteria were neurological symptoms, discrepancy between limbs greater than 2 cm, rheumatoid arthritis, herniated discs, tumours or vertebral fractures [3]. The exclusion criteria for the CG were the presence of low back pain or absence from work/school due to pain. All subjects in both groups, the LBP group and CG, showing body mass index (BMI) greater than 29.9 or who had prior pelvis or spine surgery were excluded from the sample. This study was approved by the Local Ethics Research Committee, and all participants signed informed consent form. 2.2. Procedures Data collection was performed on two days. On the first day, body mass, height, leg length, physical activity level (IPAQ), disability level (Roland Morris Disability Questionnaire) and pain level (Visual Analogue Scale – VAS) were measured. Later, the volunteers were submitted to trunk proprioceptive evaluations. Then, they were evaluated for resistance of the trunk extensor muscles, and immediately thereafter, trunk proprioceptive evaluation was performed again. On the same day, the subjects performed familiarisation with an isometric protocol for trunk flexion and extension. On the second day, only the isometric evaluation protocol was repeated, and at the same time, the electromyographic signals of the rectus abdominal (RA), internal oblique (IO), multifidus (MU) and longissimus thoracis (LT) muscles were obtained. Synchronisation of the dynamometric and electromyographic data was performed through plate receiver synchronisation G2 TM 2400 Analog Input Board (Noraxon , Scottsdale, Arizona, USA). 2.2.1. Pain assessment Pain intensity was verified using the VAS, with the subject indicating the pain intensity prior to the performance of tests. The duration of pain in years was also evaluated.
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
241
vice when they perceived the previously familiarised position [12]. The equipment provided a measurement, in degrees, of the position at which each volunteer triggered the device and the distance from the target position. Each volunteer performed test twice, and the mean of the two measurements was used [12].
Fig. 1. Isokinetic dynamometer evaluation position.
2.2.2. Isokinetic dynamometer Assessments in the Isokinetic Dynamometer System 4 PRO were performed by coupling a special chair (Dual position Back Ex/Flex Attachment) in the Seated-Compressed mode (Fig. 1). Thus, the subjects remained seated and were stabilised by means of a device for maintaining the knee at 90◦ , and special belts were fixed on the chest, in the region of hip joint and in the middle third of the thigh. The mechanical axis of the dynamometer was aligned with the anterior superior iliac spine, as recommended by the equipment manufacturer. 2.2.3. Proprioceptive assessment Trunk proprioception was evaluated using an isokinetic dynamometer for a test of passive motion reproduction, in which the individual remained blindfolded. From the sitting position with 90◦ of hip flexion, the dynamometer moved at an angular speed of 1◦ /s in the trunk flexion direction over a total range of 20◦ . The volunteers were asked to memorise this position, which was maintained for ten seconds. Then, the chair returned to its initial position. From this time point, the equipment moved again at the same speed and in the same direction, and the volunteers were instructed to stop the amplitude by moving a manual de-
2.2.4. Muscle endurance assessment Muscle endurance assessment was performed using the Biering-Sorensen test. In this test, the subject was kept in the prone posture and was firmly stabilised with belts at the levels of the hips, knees and feet. The upper trunk remained suspended at the level of the ASIS, with the upper limbs crossed over the trunk (Fig. 2). To assess the resistance of the trunk extensor muscles, the test measured the time that the subject was able to maintain the upper body in the horizontal position [3]. The volunteers were instructed to keep the trunk in the horizontal position, which was controlled by an electrogoniometer (Inline1D/2D Electrical Goniometer Noraxon , Scottsdale, Arizona, USA) positioned on the mid-axillary line of the trunk at the ASIS level, and a monitor showed the trunk angle as a form of feedback (Fig. 2). The subjects were instructed to maintain the position for as long as possible. The criterion adopted for test interruption was inability to maintain the horizontal position of the trunk. A threshold amplitude of 10◦ trunk movements was determined, and if the subject exceeded this limit or if she voluntarily lowered her trunk due to an inability to maintain the required position, the test was stopped [3,4]. 2.2.5. Isometric muscle torque assessment Trunk isometric muscle torque was assessed on two different days, with the values obtained from the second evaluation used for analysis. To perform isometric contraction, the trunk was kept at 60◦ of flexion, and the volunteers performed six isometric contractions held for five seconds each: three for trunk flexion and three for trunk extension. A 30-second rest interval was provided between each contraction. Verbal encouragement during muscle contraction was standardised. Trunk isometric muscle torque was synchronously assessed using electromyography by means of a TM 2400 G2 receiver Analog Input Board synchronisation plate (Noraxon , Scottsdale, Arizona, USA). 2.2.6. Electromyography assessment Electromyographic signals were obtained using a 16-channel biological signal acquisition module on
242
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
Fig. 2. Biering-Sorensen test position.
a Telemyo 900 Myoresearch telemeter (Myoresearch Electromyograph Noraxon , Scottsdale, Arizona, USA) (common mode rejection 100 dB, input impedance greater than 10 MOhm and base noise less than 1 UV RMS), Myoresearch software (Noraxon ), connected to a 16-bit analogue-digital converter calibrated with a sampling frequency of 2000 Hz and total gain of 2000 times (20 times on the sensor and 100 times on the machine). Ag/AgCl surface electrodes in bipolar configuration, 1 cm in diameter and at a distance of 2 cm from each other, were used. The skin was shaved and cleaned with alcohol prior to the placement of the electrodes to avoid potential interference with the electromyographic signal [13]. The electrodes were placed bilaterally (right and left sides) on the global or multisegmental trunk muscles: the rectus abdominis (RA), 3 cm above the umbilicus and 2 cm lateral to the midline, vertically oriented; and the longissimus thoracic (LT), vertically oriented at 2 cm lateral to the midline from spinous process of L1. On local or unisegmental trunk muscles, the electrodes were placed as followed: internal oblique (IO), 2 cm medially and inferiorly to the anterior superior iliac spine [14] and the multifidus (MU), 2 cm lateral to the midline toward spinous process L5, oriented towards the line between the interspinous space of L1L2 and the posterior inferior iliac spine, according to SENIAM [15] (Hermens et al., 2000). The reference electrode was positioned at the radial styloid process. 2.3. Data analysis 2.3.1. Data processing Isometric assessment analysed the torque and electromyographic signal data through specific routines developed in Matlab software (Mathworks 7.0 Natick, Massachusetts, USA). The torque signal was processed with a 4th order Butterworth digital filter with a cutoff frequency of 15 Hz. Thus, peak torque in trunk flexion and extension was observed. The torque measurements obtained were normalised based on the body mass of
each subject to consider the differences between individuals and to ensure more precise functional muscle performance. The analysis of muscle activation at peak torque (PT) used a window of 100 ms (50 ms before and 50 ms after PT) for trunk flexion and extension. The EMG signal was analysed in time domain by means of root mean square (RMS) values. A Butterworth 4th order digital bandpass filter of 20–500 Hz was used. In trunk flexion, RMS of the right and left muscles’ IO and RA were calculated, and for trunk extension movement, RMS of the right and left MU and LT muscles were calculated. The RMS values were normalised using the RMS peak of maximal isometric contractions. Furthermore, the ratio of activation amplitude, measured by RMS at peak torque between bilateral trunk muscles, was also determined. In other words, during trunk flexion, the RMS ratio was calculated for RA right and RA left and for IO right and IO left, and this calculation was also performed for the trunk extensor muscles in extension movement. Obtaining the ratios of bilateral muscles allowed for an analysis of the equilibrium amplitude of trunk muscle activation, in which a value close to one indicated good muscular balance. For antagonist activation during isometric evaluation, the amplitude of EMG signal was calculated by means of RMS values at the peak torque of antagonistic muscles. In other words, for trunk flexion, RMS of the MU and LT muscles were bilaterally calculated, and for trunk extension, RMS of the RA and IO muscles were bilaterally calculated at peak torque. 2.3.2. Statistical analysis Data analysis was performed using SPSS Inc. statistical software package. Data normality was verified using the Shapiro-Wilk test, and multivariate analysis of variance and discriminant analysis of data were subsequently performed to investigate the discriminating power of variables between groups with and without low back pain. For intra-group analysis of antagonist activation variables, the paired t-test for normality of assumed
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
243
Table 1 Mean and standard deviation of age, anthropometric data, Biering-Sorensen test, proprioception through passive repositioning error, extensor and flexor peak torque normalized according to body mass of the low back pain (LBP) and control group (CG) Age (years) Mass (kg) Height (m) BMI (kg/m2 ) Biering Sorensen (s) RE before (◦ ) RE after (◦ ) PT Extensor (Nm/kg) PT Flexor (Nm/kg) ∗p
LBP (n = 14) 24.14 (3.13) 61.68 (7.19) 1.66 (0.05) 22.31 (2.12) 85.81 (37.79) 2.66 (1.36) 2.95 (1.94) 2.48 (0.69) 1.49 (0.40)
CG (n = 14) 22.21 (3.40) 58.2 (8.73) 1.61 (0.06) 22.23 (1.98) 134.25 (43.88) 2.41 (1.46) 2.00 (1.16) 3.56 (0.88) 1.85 (0.39)
p 0.131 0.256 0.044∗ 0.912 0.004∗ 0.664 0.062 0.001∗ 0.023∗
F − − − − 9.791 0.193 3.799 13.178 5.846
< 0.05, LBP different of the CG. Legend: Repositioning Error (RE), Peak Torque (PT), Body Mass Index (BMI).
difference and Wilcoxon’s test for normality of nonassumed difference were used. The significance level adopted was 0.05.
3. Results The LBP group showed a pain onset time of 4.85 (3.73) years and pain intensity of 12.84 (17.05) mm on the VAS. Disability level was measured by the Roland Morris Disability Questionnaire, on which zero points indicates no disability, and 24 points indicates maximum disability. Regarding disability, the LBP group had a mean of 3.46 (2.50) points. Table 1 shows comparisons between the groups regarding age, anthropometric data and some independent variables, such as proprioception through passive repositioning error, muscle fatigue assessed by the Biering-Sorensen test and extensor and flexor peak torque normalised according to body mass. The LBP group had greater mean height (m) values, compared to the control group (p < 0.05). Multivariate analysis of variance identified significant differences (p = 0.008, F = 3.761) between the CG and LBP group for independent variables, as shown in Table 1. The results of the Biering-Sorensen test showed differences between groups, with the LBP group showing a 36.08% shorter endurance time (p < 0.05) than the CG. In contrast, regarding proprioception, represented by the error values of passive trunk repositioning, there were no differences between individuals with and without low back pain, either before or after the Biering-Sorensen test (Table 1). Peak isometric extension and flexion in the CG were greater than in the LBP group: 30.33% and 19.45%, respectively (Table 1). RMS values of the IO, RA, MU and LT muscles at peak torque showed no significant differences (p =
Fig. 3. Antagonist activation of the trunk between control group (GC, n = 14) and low back pain group (LBP, n = 14).
0.546, F = 0.991) between the groups for either trunk flexion or extension. Only for the left LT muscle during extension were there differences between the groups (p < 0.05) (Table 2). Regarding the activation ratio between the muscles of the right and left hemispheres during peak torque, there were no differences between the groups (p > 0.05), as shown in Table 2. However, antagonist activation of the trunk showed inter-group and intra-group differences for comparisons between local and global muscles. Figure 3 shows differences between the groups in antagonist activation only for local extensor muscles: bilaterally, the right MU (p = 0.027, F = 5.509) and left MU (p = 0.048, F = 4.309) during trunk flexion. The
244
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
Table 2 RMS values of the IO, RA, MU and LT muscles at peak torque and activation ratio between the muscles of the right and left hemispheres during peak torque. Mean and standard deviation values
RMS (µV) Extension
Flexion
Ratio Extension Flexion
CG (n = 14) Mean (SD)
LBP (n = 14) Mean (SD)
p
F
Right LT Left LT Right MU Left MU
64.1 (20.04) 71.29 (20.56) 59.93 (21.47) 72.44 (19.2)
54.1 (22.85) 54.85 (15.84) 57.31 (23.87) 61.44 (27.08)
0.229 0.025# 0.763 0.226
1.517 5.617 0.093 1.536
Right RA Left RA Right IO Left IO
67.82 (13.41) 63.64 (22.04) 60.06 (21.2) 64.15 (13.56)
63.98 (18.97) 62.37 (20.33) 60.4 (18.94) 62.98 (18.91)
0.541 0.876 0.965 0.852
0.383 0.025 0.002 0.036
LT MU
0.91 (0.27) 0.85 (0.34)
0.99 (0.4) 0.98 (0.32)
0.569 0.300
0.332 1.116
RA IO
1.22 (0.65) 0.99 (0.41)
1.07 (0.3) 0.99 (0.4)
0.422 0.967
0.422 0.967
Longissimus thoracic (LT), multifidus (MU), rectus abdominais (RA) and internal oblique (IO). Control group (CG) and low back pain (LBP) group. #p < 0.05 for CG higher than LBP.
CG showed, on average, 40% greater antagonist activation of extensor muscles (MU, bilaterally) during trunk flexion, compared to the LBP group. In intra-group analysis (local X global), the CG showed increased activation of antagonist muscles, the right MU (p = 0.003) and left MU (p = 0.009), in trunk flexion, compared to the right and left LT muscles, respectively. The LBP group showed increased antagonist activation of right IO and left IO muscles during extension, which were 52% and 44% greater than that of the right RA and left RA, respectively. In trunk flexion, the LBP group also showed greater antagonist activation, with RMS of the right MU muscle 35% higher than that of the right LT (p = 0.005) and the left MU 27% higher than that of the left LT (p = 0.001). Discriminant analysis showed significant discrimination between the groups (WL = 0.347, χ2 = 22.745, df = 9, p = 0.007) for endurance, proprioception, extensor and flexor peak torque variables, in which 78.6% of individuals with low back pain were correctly classified in the LBP group, and 71.4% of asymptomatic individuals were correctly classified in the CG. Furthermore, the variables with greater discriminative power were, in descending order, extensor peak torque, endurance time, and flexor peak torque. Regarding variables related to muscle activation, such as the RMS of each muscle at peak torque, the ratio between the RMS of the bilateral trunk muscles and antagonist activation, discriminant analysis showed no significant differences between the groups (WL = 0.261, χ2 = 21.49, df = 20, p = 0.369). Thus,
muscle activation variables, such as amplitude, ratio and antagonist activation, were able to classify only 57.1% of young individuals with pain in the LBP group and 42.9% of young people without back pain in the CG.
4. Discussion Discrimination of neuromuscular variables in a nonspecific musculoskeletal dysfunction with a multifactorial aetiology could contribute, among other factors, to clarification of cause-and-effect relationships. It was found that isometric endurance time and the peak torque of the trunk extensors showed greater discriminatory power between individuals with and without this symptom. However, muscle activation variables, such as the activation amplitude at peak torque, activation ratio and antagonist activation, did not contribute to discrimination or correct classification of each subject within the groups. As noted in the literature [3,4], in this study, the Biering-Sorensen test showed accuracy in diagnosing individuals with and without low back pain. The endurance time of asymptomatic individuals was, on average, 134 seconds, in agreement with Johnson et al. [16], who found a mean value of 138 seconds for healthy adolescents. In association with physiological changes in the muscle fatigue process, such as changes in the speed conditions of muscle fibres due to lactate accumulation [5], psychological influences are also related to
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
lower resistance of the extensor muscles in populations with back pain. Factors such as motivation, discomfort tolerance related to muscle fatigue and, particularly in clinical situations, pain or fear of pain could influence this outcome [17]. Anthropometric factors, such as BMI, can interfere with performance of the Biering Sorensen test [16]. In this study, BMI was not different between the groups, but height in the LBP group was greater than in the CG. Johnson et al. [16] suggested a lack of correlation between isometric endurance time and height, so the difference in height between the LBP group and CG did not influence the results. The variable with the greatest discriminating power demonstrated in this study was the isometric peak torque of the trunk extensor. Data from this study corroborated the findings of Yahia et al. [8], who showed impaired muscle torque of the trunk extensor and flexor in individuals with chronic low back pain, with greater differences between groups for the peak torque of the trunk extensor. The lower peak torque in trunk flexion and extension movements in the LBP group could be related to fear of feeling pain or fear of re-injury associated with vertebral movements. Thus, the anticipation of pain or fear of inducing pain can result in inability or unwillingness to exert the maximum effort, resulting in lower torque values [18]. Similarly, neural adaptations and changes in motor control, such as delay in activation onset during balance disturbances in individuals with low back pain [9,10], which persist even after the elimination of painful stimuli, have also been indicated as contributors to weakness of the trunk muscles [19]. The current study did not identify differences in trunk proprioception between the groups and similarly found no effects of extensor muscle fatigue induced by the Biering-Sorensen test on the ability to reproduce a pre-determined angle of the trunk. Lee et al. [20] suggested that active and passive repositioning tests were less sensitive for identifying proprioceptive deficits because there was an influence of memory on the task, which in turn varied according to amount of time between the presentation of the target position and movement execution. It has also been suggested that the differences found in the literature are due not only to different assessment protocols used but also to the ages of the participants. Thus, proprioception has been shown to be more impaired in older subjects (mean age of 40 years old) compared to younger subjects (mean age of 19 years old) as a result of the effects of the natural aging process on neural control and proprioception [7, 20].
245
In contrast with evidence found in the literature regarding changes in the responses of trunk muscle activation in individuals with low back pain in cases of postural imbalance or in cases of rapid trunk flexion [21], in this study, changes were not observed during the task of producing maximum force in trunk flexion and extension. The LBP group did not exhibit different patterns of muscle activation variables. Thus, muscle activation variables showed poor discriminative power, with only 42.9% correct classification. However, Andersen et al. [22] suggested that, under conditions of chronic musculoskeletal pain, the ability to activate muscles quickly and to generate strength might be more severely impaired than the maximum amplitude of muscle activation and the capacity to produce maximum force. Due to anatomical and functional differentiation of the trunk muscles, there is clinical importance in analysing the segmental and multi-segmental erector and abdominal muscles of individuals with low back pain [23]. Thus, intra-group comparisons were also performed for antagonist activation amplitude. The LBP group showed higher local antagonist activation than globally in the trunk flexion and extension movements. Jones et al. [24] suggested that individuals with low back pain in situations of postural imbalance increased trunk coactivation regardless of disturbance direction to compensate for a lack of accurate proprioceptive information, aiming to restrict movement to avoid pain. As a result, trunk coactivation leads to a lower capacity to produce torque in these individuals [24].
5. Conclusion Among the neuromuscular variables investigated, extensor peak torque, muscle endurance, pain intensity and flexor peak torque showed significant discrimination and correct classification of individuals within each group. However, trunk proprioception and the muscle activation amplitude variables showed poor discriminative power. Based on these findings, clinical treatment of young individuals with low back pain should include strength training and endurance, particularly in the extensor muscles.
Acknowledgement D.M. Rossi was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
246
D.M. Rossi et al. / Discriminant analysis of neuromuscular variables in chronic low back pain
References [1]
Posadzki P, Lizis P, Hagner-Derengowska M. Pilates for low back pain: A systematic review. Complementary Therapies in Clinical Practice. 2011; 17: 85-89. doi: http://dx.doi. org/10.1016/j.ctcp.2010.09.005. [2] Smith DR, Leggat PA, 2007. Back pain in the young: A review of studies conducted among school children and university students. Current Pediatric Reviews. 2007; 3(1): 69-77. doi: 10.2174/157339607779941624. [3] Gruther W, Wick F, Paul B, Leitner C, Posch M, Matzner et al. Diagnnostic accuracy and reliability of muscle strength and endurance measuremenfts in patients with chronic low back pain. Journal of Rehabilitation Medicine. 2009; 41: 613-619. doi: 10.2340/16501977-0391. [4] Latimer J, Maher CG, Refshauge K, Colaco L. The reliability and validity of the Biering-Sorensen test in asymptomatic subjects and subjects reporting current or previous nonspecific low back pain. Spine. 1999; 24: 2085-2090. doi: 10.1097/ 00007632-199910150-00004. [5] Biedermann HJ, Shanks, GL, Forrest WJ, Inglis J. Power spectrum analyses of electromyographic activity: Discriminators in the differential assessment of patients with chronic low back pain. Spine. 1991; 16(10): 1179-1184. [6] Ribeiro F, Mota J, Oliveira J. Effect of exercise-induced fatigue on sense of the knee in the elderly. Journal of Applied Physiology. 2007; 99: 379-385. doi: 21-006-0357-8. [7] Silfies SP, CholewickI J, Reeves NP, Greene HS. Lumbar position sense and the risk of low back injuries in college athletes: A prospective cohort study. BMC Musculoskelet Disorders. 2007; 8(129): 1-7. doi: 10.1186/1471-2474-8-129. [8] Yahia A, Jribi S, Ghroubi S, Elleuch M, Baklout S, Elleuch MH. Evaluation of the posture and muscular strength of the trunk and inferior members of patients with chronic lumbar pain. Joint Bone Spine. 2011; 78: 291-297. doi: http://dx.doi. org/10.1016/j.jbspin.2010.09.008. [9] Newcomer KL, Jacobson TD, Gabriel DA, Lardson, DR, Brey RH, An K. Muscle activation patterns in subjects with and without low back pain. Archives of Physical Medicine and Rehabilitation. 2002; 83(6): 816-821. doi: http://dx.doi. org/10.1053/apmr.2002.32826. [10] Oddson LIE, De Luca CJ. Activation imbalances in lumbar spine muscles in the presence of chronic low back pain. Journal of Applied Physiology. 2003; 94: 1410-1420. doi: 10.1152/ japplphysiol.01183.2001. [11] Larivière C, Arsenault AB, Gravel D, Gagnon D, Loisel P. Surface electromyography assessment of back muscle intrinsic properties. Journal of Electromyography and Kinesiology. 2003; 13: 305-318. doi: http://dx.doi.org/10.1016/S10506411(03)00039-7. [12] Yilmaz B, Yaar E, Takaynatan MA, Goktepe AS, Tucu I, Yazcolu K. et al. Relationship between lumbar muscle strength and proprioception after fatigue in men with chronic low back pain. Turkey Journal Rheumatology. 2010; 25: 68-71. doi: 10.5152/tjr.2010.04. [13] Gonçalves M, Marques NR, Hallal CZ, Van dieen JH.
Electromyographic activity of trunk muscles during exercises with flexible and non-flexible poles. Journal of Back and Musculoskeletal Rehabilitation. 2012; 24: 209-214. doi: 10.3233/BMR-2011-0297. [14] Marshall P, Murphy B. The validity and reliability of surface EMG to assess the neuromuscular response of the abdominal muscles to rapid limb movement. Journal of Electromyography and Kinesiology. 2003; 13: 477-489. doi: http://dx.doi. org/10.1016/S1050-6411(03)00027-0. [15] Hermens JH, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. Journal of Electromyography and Kinesiology. 2000; 10: 361-374. doi: http://dx.doi.org/10.1016/ S1050-6411(00)00027-4. [16] Johnson OE, Mbada CE, Akosil CO, Agbeja AO. Isometric endurance of the back extensors in school-aged adolescents with and without low back pain. Journal of Back and Musculoskeletal Rehabilitation. 2009; 22: 205-211. doi: 10.3233/ BMR-2009-0235. [17] Mannion AF, Riordan D, Dvorak J, Masharawi Y. The relationship between psychological factors and performance on the Biering-Sorensen back muscle endurance test. The Spine Journal. 2011; 11: 849-857. doi: 10.1016/j.spinee.2011. 08.004. Epub 2011 Sep 8. [18] Davariana S, Maroufia N, Ebrahimia I, Farahmand F, Parnianpour M. Trunk muscles strength and endurance in chronic low back pain patients with and without clinical instability. Journal of Back and Musculoskeletal Rehabilitation. 2012; 25: 123-129. doi: 10.3233/BMR-2012-0320. [19] Tsao H, Galea MP, Hodges PW. Reorganization of the motor cortex is associated with postural control deficits in recurrent low back pain. Brain. 2008; 131: 2161-2171. doi: 10.1093/ brain/awn154. Epub 2008 Jul 18. [20] Lee AS, Cholewicki J, Reeves P, Zazulak BT, MysliwieC LW. Comparison of trunk proprioception between patients with low back pain and healthy controls. Archives of Physical Medicine and Rehabilitation. 2010; 91: 1327-1331. doi: 10.1016/j.apmr.2010.06.004. [21] Marshall P, Murphy B. Delayed abdominal muscle onsets and self-report measures of pain and disability in chronic low back pain. Journal of Electromyography and Kinesiology. 2009, doi:10.1016/j.jelekin.2009.09.005. [22] Andersen LL, Holtermann A, Jorgensen MB, Sjogaard G. Rapid muscle activation and force capacity in conditions of chronic musculoskeletal pain. Clinical Biomechanics. 2008; 23: 1237-1242. doi:10.1016/j.clinbiomech.2008.08.002. Epub 2008 Oct 2. [23] Sung PS, Lammers AR, Danial P. Different parts of erector spinae muscle fatigability in subjects with and without low back pain. The Spinal Journal. 2009; 9: 115-120. doi: 10.1016/j.spinee.2007.11.011. Epub 2008 Feb 14. [24] Jones SL, Henry SM, Raasch CC, Hitt JR, Bunn JY. Individuals with non-specific low back pain use a trunk stiffening strategy to maintain upright posture. Journal of Electromyography and Kinesiology. 2012; 22: 13-20. doi: 10.1016/j. jelekin.2011.10.006. Epub 2011 Nov 18.
Copyright of Journal of Back & Musculoskeletal Rehabilitation is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
