Available online at www.sciencedirect.com

Journal of Electromyography and Kinesiology 18 (2008) 997–1005 www.elsevier.com/locate/jelekin

Assessment of the validity of the Biering-Sørensen test for measuring back muscle fatigue based on EMG median frequency characteristics of back and hip muscles Pascal Coorevits a,b,*, Lieven Danneels b, Dirk Cambier b, Herman Ramon c, Guy Vanderstraeten b a

RAMIT vzw, c/o Department of Medical Informatics and Statistics, Ghent University Hospital, Building K3, 5th Floor, De Pintelaan 185, B-9000 Ghent, Belgium b Department of Rehabilitation Sciences and Physiotherapy, Ghent University, Belgium c Department of Agro-Engineering and Economics, Catholic University Leuven, Belgium Received 19 December 2006; received in revised form 30 October 2007; accepted 31 October 2007

Abstract The aims of the present study were (1) to investigate the differences in median frequency characteristics between back and hip muscles of healthy subjects during a Biering-Sørensen test, (2) to determine if the Biering-Sørensen test is a valid test for measuring back muscle fatigue, and (3) to standardise the Biering-Sørensen test by using objective movement analysis when defining endurance time, and compare this to the original method based on tactile feedback. Twenty healthy subjects participated in this experiment. The electromyographic activity of eight back and hip muscles was bilaterally measured. In addition three-dimensional data of the lumbar region were collected with an ultrasound movement analysis system. Median frequencies were computed from the EMG power spectra. Two methods of determining the endurance time of the Biering-Sørensen test yielded highly correlated but significantly different normalized median frequency slope values (NMFslope). Significant differences in NMFslope values between several back and hip muscles could be demonstrated. Low to moderate correlation coefficients were shown between NMFslope values and endurance time. Multiple stepwise linear regression analyses revealed that only NMFslope of the thoracic part of the iliocostalis lumborum muscle could significantly predict the test endurance time. The findings of the present study support the validity of the Biering-Sørensen test for measuring back muscle fatigue. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Electromyography; Validity; Muscle fatigue; Median frequency; Isometric contraction

1. Introduction The electromyographic (EMG) spectrum analysis has been widely used to monitor the development of localized muscle fatigue (Biedermann et al., 1990; Dolan et al., 1995; Koumantakis et al., 2001; Mannion et al., 1997, *

Corresponding author. Address: RAMIT vzw, c/o Department of Medical Informatics and Statistics, Ghent University Hospital, Building K3, 5th Floor, De Pintelaan 185, B-9000 Ghent, Belgium. Tel.: +32 9 3328926; fax: +32 9 3323439. E-mail address: [email protected] (P. Coorevits). 1050-6411/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2007.10.012

1998; Mannion and Dolan, 1994; Mayer et al., 1989; Ng and Richardson, 1996; Ng et al., 1997; Tsuboi et al., 1994; Van Diee¨n et al., 1993). Fatigue causes a decrease of the frequency content of the EMG signal, usually described as a decline of the mean power or median frequency parameters of the EMG power spectrum. The Biering-Sørensen test, a trunk holding test in an antigravity prone position, is commonly used to measure the endurance capacity of the back muscles (Biering-Sørensen, 1984; Chok et al., 1999; Cooper and Stokes, 1994; Mannion and Dolan, 1994; Ng and Richardson, 1996; Ng et al., 1997; Tsuboi et al., 1994; Umezu et al., 1998).

998

P. Coorevits et al. / Journal of Electromyography and Kinesiology 18 (2008) 997–1005

The change in the median frequency of the EMG spectrum obtained during this test has been shown to be a better predictor of first-time low back pain (LBP) acquisition than the simple measure of endurance time (Mannion et al., 1997). On the other hand, Adams et al. reported that the median frequency parameters were not significant predictors of first-time LBP, although endurance time during the Biering-Sørensen was (borderline) significant at some of their follow-ups (Adams et al., 1999). Several studies however have shown that chronic LBP patients often suffer from excessively fatigable back extensor muscles (BieringSørensen, 1984; Chok et al., 1999; Jorgensen and Nicolaisen, 1987; Mayer et al., 1989; Roy et al., 1989; Tsuboi et al., 1994). Despite recommendations of several authors (Mannion et al., 1997; Roy et al., 1989; Sparto et al., 1997), most studies analyzing back muscle fatigue have relied on analyses of EMG signals detected from only a few muscle locations (Biedermann et al., 1990; Koumantakis et al., 2001; Mannion and Dolan, 1994; Ng and Richardson, 1994; Ng et al., 1997). The back muscles cannot be accurately assessed as one homogeneous muscle mass because they are composed of several fascicles that act synergistically to produce trunk extension tasks (Larivie`re et al., 2002). Moreover, Mannion and Dolan (1994) have found that endurance time during an isometric back extension task is highly correlated with the muscle that shows the greatest decline in median frequency than any one muscle alone. Therefore it is important to simultaneously measure several muscle locations in order to avoid a reduced and oversimplified view of the EMG trunk muscle performance (Mathieu and Fortin, 2000; Roy et al., 1989; Tsuboi et al., 1994). Relevant within this context is that Macintosh and Bogduk showed that the erector spinae muscle consists of two muscles – the iliocostalis lumborum and longissimus thoracis – each with distinct thoracic and lumbar parts (Macintosh and Bogduk, 1987). However, to the authors’ current knowledge no previous studies have reported simultaneously collected electromyographic data from these four different parts of the erector spinae muscles during a Biering-Sørensen test. During this test it was also demonstrated that the hip extensor muscles (gluteus maximus and biceps femoris) showed muscle fatigue, which was closely related to the test endurance time (Kankaanpa¨a¨ et al., 1998). Consequently, the validity of the Sørensen back endurance test to measure specifically the paraspinal localized muscle fatigue has been questioned (Kankaanpa¨a¨ et al., 1998). Furthermore, it has been shown that – among other factors (e.g. changes in muscle force, displacement of electrodes, motor unit synchronization, etc.) – the frequency content of the EMG signal can be altered by variations in muscle length throughout the test (Mannion and Dolan, 1996; Sparto et al., 1999). Because the length of the erector spinae muscles is largely determined by the curvature of the lumbar spine in the sagittal plane (Dolan et al., 1995; Mannion and Dolan, 1996), measurements of the lumbar lordo-

sis of the subject should be taken into account when studying EMG trunk muscle results. Few authors, however, have incorporated recordings of the posture of the subject during the Biering-Sørensen test (Chok et al., 1999; Mannion et al., 1997; Mannion and Dolan, 1994; Ng and Richardson, 1996; Sparto et al., 1997). The aims of the present study were (1) to investigate the differences in median frequency characteristics between back and hip muscles of healthy subjects during a Biering-Sørensen test, (2) to determine if the Biering-Sørensen test is a valid test for measuring back muscle fatigue, and (3) to standardise the Biering-Sørensen test by using objective movement analysis when defining endurance time, and compare this to the original method based on tactile feedback. 2. Materials and methods 2.1. Subjects Seven males (mean age of 32.9 ± 14.3 years, mean height of 178.7 ± 7.5 cm and mean weight of 76.7 ± 7.8 kg) and 13 females (mean age of 23.5 ± 1.1 years, mean height of 169.1 ± 7.1 cm and mean weight of 61.5 ± 8.0 kg) voluntarily participated in this study. All participants were free of back complaints, had no known pathology, and were habitually active. The local University Hospital’s ethics committee approved the study protocol. Subjects gave their written informed consent prior to participation. 2.2. Design The subjects had to perform a modified Biering-Sørensen test until fatigue (modified means that the original Biering-Sørensen test was used with minor variations e.g. fewer than three belts to strap the subject’s body or variations in hand position, etc.). Therefore, the subjects were placed in a prone position on an examination couch in such a way that their superior border of the anterior iliac crest was positioned at the edge of the couch and that their upper body was extending beyond the edge of the examination couch. Two belts were used to strap the lower bodies of the subjects (one at the level of the hips, and one just below the knees). During the test the examiner instructed the subjects to maintain their unsupported body in this position as long as possible. The subjects had to hold their hands touching their forehead, with their elbows out to the side and leveled with the trunk. The subjects were also instructed to hold their head in a neutral position, and to look downward at a visual fixation point. Throughout the test verbal instructions and encouragement were given by the examiner. The subjects also received tactile feedback by means of a rope hung up between two vertical stands and placed over their seventh thoracic vertebra (Fig. 1). The EMG activity of eight back and hip muscles was bilaterally recorded during the test and a three-dimensional movement analysis system was used to measure the lumbar curvature of the subjects. 2.3. Equipment The EMG signals were recorded with a 16 channel surface EMG system (MyoSystem 1400, Noraxon USA Inc., Scottsdale,

P. Coorevits et al. / Journal of Electromyography and Kinesiology 18 (2008) 997–1005

Fig. 1. Testing position of a female subject performing the BieringSørensen test.

AZ). All raw EMG signals were analogue bandpass-filtered between 10 and 500 Hz, amplified (common mode rejection ratio >100 dB, overall gain 1000, noise .05). For both the longissimus thoracis and iliocostalis lumborum muscles, the NMFslope was not significantly different between their thoracic and lumbar parts (all p > .05). The latissimus dorsi showed the lowest NMFslope, but not significantly different from ILT, GM and BF (p > .05). Of the hip muscles, the NMFslope was not significantly different between GM and BF (p > .05). Furthermore, both GM and BF were not significantly different from LD, ILT and ILL (all p > .05).

LTT

LTL

ILT

ILL

MF

GM

BF