Journal of Electromyography and Kinesiology 23 (2013) 362–368

Contents lists available at SciVerse ScienceDirect

Journal of Electromyography and Kinesiology

journal homepage: www.elsevier .com/locate / je lek in

Lumbar posture and muscular activity while sitting during office work

Falk Mörl a,⇑, Ingo Bradl b,c

a Forschungsgesellschaft für angewandte Systemsicherheit und Arbeitsmedizin mbH, Dubliner Straße 12, 99091 Erfurt, Germanyb German Social Accident Insurance Institution for the Foodstuffs and Catering Industry, Department of Prevention, Biomechanics, Dubliner Straße 12, 99091 Erfurt, Germanyc University Hospital Jena, Clinic for Trauma-, Hand- and Reconstructive Surgery, Division for Motor Research, Pathophysiology and Biomechanics, 07740 Jena, Germany

a r t i c l e i n f o

Article history:Received 20 June 2012Received in revised form 11 September 2012Accepted 2 October 2012

Keywords:Lumbar spineLong term EMGSittingOffice workLordosisLumbar posture

1050-6411/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.jelekin.2012.10.002

⇑ Corresponding author.E-mail address: falk.moerl@apz-erfurt.de (F. Mörl)

a b s t r a c t

Purpose: Field study, cross-sectional study to measure the posture and sEMG of the lumbar spine duringoffice work for a better understanding of the lumbar spine within such conditions.Scope: There is high incidence of low back pain in office workers. Currently there is little informationabout lumbar posture and the activity of lumbar muscles during extended office work.Methods: Thirteen volunteers were examined for around 2 h of their normal office work. Typical taskswere documented and synchronised to a portable long term measuring device for sEMG and postureexamination. The correlation of lumbar spine posture and sEMG was tested statistically.Results: The majority of time spent in office work was sedentary (82%). Only 5% of the measured time wasundertaken in erect body position (standing or walking). The sEMG of the lumbar muscles under inves-tigation was task dependent. A strong relation to lumbar spine posture was found within each task. Themore the lumbar spine was flexed, the less there was activation of lumbar muscles (P < .01). Periods ofvery low or no activation of lumbar muscles accounted for about 30% of relaxed sitting postures.Conclusion: Because of very low activation of lumbar muscles while sitting, the load is transmitted bypassive structures like ligaments and intervertebral discs. Due to the viscoelasticity of passive structuresand low activation of lumbar muscles, the lumbar spine may incline into de-conditioning. This may be areason for low back pain.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Office work and sitting at a desk for longer periods is commonfor people in western civilisation. Orthopaedists and physical ther-apists assume de-conditioning of the trunk and lumbar spinestructures due to long-term sitting without longer active periodsof standing, walking or running. This de-conditioning may be a rea-son for low back pain and accelerated degeneration of lumbarspine structures. Looking at the incidence of low back pain andthe inability to work because of low back pain in office workersconfirms this assumption (Burdorf et al., 1993; Hemingway et al.,1997; Janwantanakul et al., 2008; Juul-Kristensen and Jensen,2005; Juul-Kristensen et al., 2004; Riihimäki et al., 1989, 1994;Spyropoulos et al., 2007; Törner et al., 1991; Videman and Battie,1999). In summary, there is high prevalence of low back pain in of-fice workers with the risk of getting low back pain comparable tomore demanding work. However, there is currently little informa-tion available about the behaviour of the lumbar spine over longperiods because of a lack of adequate measurement devices. In lab-oratory settings, no coherence of lumbar flexion angle and lumbar

ll rights reserved.

.

muscle activity was found (O’Sullivan et al., 2006; Callaghan andDunk, 2002). One study documents different movement patternsof the lumbar spine during sitting for low back pain developersand asymptomatic subjects (Dunk and Callaghan, 2010). Onlyone field study measured global angles of trunk and thighs and cor-related these posture measurements to the activation of lumbarmuscles (Mork and Westgaard, 2009). They did not show clear cor-relations (�.44 < r < .80) because the measurements used for trunkposture are not precise enough.

The load on lumbar discs while sitting is not to be underesti-mated and is greater than in erect positions like standing or re-clined (Nachemson, 1966). Newer studies support this data onthe whole (Wilke et al., 2001). Further, the flexion–relaxation phe-nomenon is present in flexed postures of the trunk, so there is noactive muscular support or stabilisation while resting in such posi-tions (Schultz et al., 1985; Sihvonen et al., 1988). This means acti-vation of the lumbar muscles is not required in such body positionsand it would seem that this is comparable to the situation in sed-entary work. Most studies on the flexion–relaxation phenomenononly show global angles for trunk inclination. For the lumbar spine,however, it is assumed that the curvature (lordosis, flat or kypho-sis) is the main impacting factor on activating the lumbar muscles.There have been no studies providing this information as yet.

F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368 363

Low back pain patients show atrophy of lumbar muscles (Hadaret al., 1983; Cooper et al., 1992; Hides et al., 1994; Danneels et al.,2000; Barker et al., 2004; de-las Peñas et al., 2008). Since physicalinactivity and a lack of activation of muscles are reasons for atro-phy (Hayashi et al., 1992; Salminen et al., 1993; Bloomfield,1997; Hodges et al., 2006; Hides et al., 2007; Hyun et al., 2007;Belavy et al., 2008) de-conditioning or atrophy of lumbar musclescan be assumed due to long periods of sedentary work withoutactive leisure.

Physical changes are not only present in the muscles. Passivestructures like ligaments or intervertebral discs are characterisedby viscoelasticity. This means that (also low) cyclic but long-lastingloading leads to the creep of discs or ligaments (Solomonow et al.,1998; Adams and Dolan, 1996). The main function of passive struc-tures, which is to guide the kinematics of a joint, decreases becauseof the decrease in stiffness (Solomonow et al., 1998; Adams andDolan, 1996). Moreover, ligaments have neural connections tomuscles and are mechanical receptors for critical situations(Solomonow et al., 1987; Johansson et al., 1991). Due to the creepof the passive structures, the mechanism that triggers reflexes de-creases and finally disappears (Solomonow et al., 1999, 2003). Itshould be noted that recovery of ligaments takes more than 8 hof total rest (Gedalia et al., 1999).

Both changes (de-conditioning of lumbar muscles and creep ofpassive structures) may be characterised as detuning of the lumbarspine. For example, an important reflex is triggered too late andwithin a wrong joint-angle and the adynamic muscle is not ableto protect the joint.

With this in mind, the purpose of this paper was to measure thenormal behaviour of the lumbar spine during two hours of officework. Healthy subjects were investigated. The lumbar postureand sEMG of lumbar muscles were recorded. As a more precisemeasure for lumbar posture (than pelvic angle and trunk angle),the curvature of lumbar spine was deduced. The aim of this paperis to investigate the coherence between lumbar curvature and lum-bar muscle activity. The results may provide the first impetus forfurther discussion on whether sedentary work is disadvantageousor unhealthy.

2. Materials and methods

2.1. Subjects and procedure

Thirteen subjects (8 $; 5 #) were recruited from an insurancecompany (n = 4), a software development company (n = 2) and ahealth care company (n = 7). All the subjects were investigatedwhile undertaking their normal sedentary work at a desk (Table 1).Inclusion criteria was no period of acute low back pain during the12 months before measurement. Exclusion criteria were acute lowback pain, acute pain or injury of lower extremities, deformation ofthe spine and known protrusion or prolapse of an intervertebraldisc.

Before work, the subjects were equipped with a small and por-table measuring device. At the start of data collection, a calibrationprocedure lasting approximately 1–2 min was carried out (seeSection 2.2).

Table 1Age, body height, and body weight as mean (standard deviation) for all investigatedsubjects.

Gender (n) Age (years) Body height (cm) Body weight (kg)

Female (8) 36.0 (7.0) 168 (3.2) 61.8 (6.1)Male (5) 41.2 (11.7) 182 (4.2) 81.2 (8.4)

All the subjects were investigated for a minimum of 2 h. Duringthe measurement, different tasks were manually marked by the re-searcher and written online to the data file for later identification.The observed tasks were standing, walking, unsupported sitting,supported sitting on backrest of the chair, long-lasting periods ofkeyboard use and telephoning. Each subject performed each ofthese tasks at least once, as triggered by the work to be done.The backrest of the chair did not affect the positions of the EMG-electrodes and motion sensors. None of the subjects reportedrestrictions from the measuring device. All other office-typicaltasks and periods in which the subjects could not be viewed bythe researcher were summarised using the ‘‘miscellaneous’’ marker(Table 2).

All the subjects were volunteers and gave written informedconsent under the terms of the Declaration of Helsinki. The exper-imental protocol was reviewed and approved by the local ethicscommission.

2.2. Measurements

Using the PS11-UD measuring device (Thumedi, Jahnsbach,Germany) the posture of the lumbar spine, the sEMG of selectedlumbar muscles and the cardiogram (for identification withinEMG signals) were measured synchronously. It is possible to mea-sure and collect bipolar sEMG and posture data for up to 8 h withthis device. Lumbar posture was monitored by three gravity-basedsensors. The selected lumbar muscles were the longissimus muscleat lumbar level 1 and the multifidus muscle at lumbar level 4bilaterally.

The motion sensors (size: 24 � 24 � 10 mm) were applied on thelumbar spine at level L5, L3 and L1 using hypoallergenic double-sided adhesive tape. Each motion sensor measures spatialorientation with an accuracy of 0.1� within the field of gravity (e.g.inclination from vertical axis) and is comparable to devices de-scribed in the literature (Aminian and Najafi, 2004). The angular dif-ference between the sensors in the sagittal plane was calculated as ameasure of the lumbar spine curvature (Mörl and Blickhan, 2006).

Abrasive lotion was used for skin preparation for bipolar sEMGand ECG measurements. Where there was pronounced growth ofhair at the application position, the subjects were shaved prior toskin preparation. After this the skin was fumigated and dried.The electrodes used were Ag/AgCl-electrodes (H93SG, Tyco Health-care, Germany) with a circular uptake area of 10 mm and an inter-electrode distance of 25 mm. The electrode positions of the fourinvestigated muscles were in line with the recommendations ofSENIAM (Hermens et al., 1999).

Before data collection a calibration procedure was carried outfor identification and elimination of the ECG signal from EMG sig-nals, for spinal posture offset-adjustment and for normalisation ofthe EMG. The first step of calibration was to measure the raw ECGsignal on each EMG channel while the subject sat relaxed and sup-ported by the back of the chair. While occurrence of the ECG signal(detected by ECG channel) during data collection, the device elim-inates the main part of crosstalk by subtracting the ECG signal fromthe sEMG signal for each single EMG channel (Mörl et al., 2010).

The second step of calibration was to eliminate the angular off-set for lateral bending and axial rotation of the spine (not discussedin this paper) due to inaccurate sensor application. The normalshape of the lordosis in standing position was not offset-adjustedand was given in normal angular positions.

The golden standard for EMG normalisation are records duringmaximum voluntary contractions (MVC). These measurements arelaborious and depend on the subject’s motivation. MVC measure-ments are nearly impossible at the workplace, especially for lum-bar spine muscles. A special normalisation posture was thereforeundertaken by the subjects during the third step of the calibration

Table 2Summary of distribution [%, relating to individual measuring time] of different tasks during office work for all subjects.

Subject Standing Walking Uns. sitting Sup. sitting Keyboard Phoning Misc

S1 1.3 1.4 1.4 0.1 54.2 20.7 18.9S2 4.1 2.4 23.7 41.3 0.1 4.1 23.5S3 5.0 2.7 3.6 0.1 49.0 6.5 32.5S4 3.6 2.9 60.2 3.4 22.2 3.0 3.7S5 0.1 1.5 65.5 23.7 0.1 0.1 6.8S6 7.0 1.1 75.2 2.8 0.1 8.3 4.6S7 2.1 1.6 59.6 5.3 14.2 20.6 0.8S8 11.2 5.2 60.0 0.4 6.7 5.8 10.0S9 6.7 2.6 38.9 28.5 9.6 - 12.6S10 2.7 1.7 21.4 22.3 37.4 8.5 4.7S11 – 1.3 55.2 25.8 – 14.7 0.1S12 2.6 0.1 19.0 54.9 15.0 – 6.9S13 0.1 2.7 44.0 16.1 12.9 6.8 15.9

unsupported sitting supported sitting0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

rel R

MS

lumbar muscular activity during sitting

Mu L4 riMu L4 leLo L1 riLo L1 le

Fig. 1. Median of normalised lumbar muscle activation (LoL1 – longissimus muscleat lumbar level L1, MuL4 – multifidus muscle at lumbar level L4, le – left and ri –right side, respectively) during sitting. In comparison to normalisation posture theactivation of lumbar muscles is low (P < .01, for all muscles).

364 F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368

procedure. For this, the body position of the subjects was as fol-lows: standing position, slightly bent knees and hips about 45�,the trunk tilted about 45�. The researcher advised the subjects tomaintain lordosis of the lumbar spine. Within this posture the tor-que of the tilted trunk has to be compensated for mostly by thelumbar muscles under investigation. Activation was moderatelycomparable to an exercise in gymnastics. Subjects had to hold thisposition (isometric contraction) for about 30–60 s. Repeated mea-surements carried out previously in the laboratory showed ade-quate reliability and validity. All given sEMG data wasnormalised to this activation (i.e. fractions of this activation).

2.3. Data processing and analysis

The PS11-UD measures and stores the raw sEMG data (4–650 Hz) at a sampling frequency of 4096 Hz. Amplification of thesignal was selected to get a resolution of 688 nV per bit. The rawsignals were high pass filtered (16 Hz), low pass filtered (1 kHz)and band pass filtered (moving average at multiplies of 50 Hz) torepress crosstalk from power supply lines.

During measurement, the device processes and stores the recti-fied and averaged activity (RMS, root mean square) at 8 Hz by

RMS ¼ k

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1

fu � fl

Z fu

fl

A2ðf Þdf

sð1Þ

with the gain k, frequency limits fl/u, amplitude A and frequency f foramplitude information. It also calculates and stores the median fre-quency (not discussed in this paper). The angular data of the motionsensors on the spine were also stored synchronously at 8 Hz. Thelumbar flexion angle was calculated by

aflex ¼ ðaS1 � aS2Þ þ ðaS2 � aS3Þ ð2Þ

with inclination angle from vertical axis aSn of each single motionsensor (S1 at lumbar level L1 . . .S3 at level L5) in sagittal plane.All the data presented here is based on the calculated RMS andangular position data.

The measurements taken were summarised for each task ob-served and presented as percentages for all subjects. For the tasks‘‘unsupported sitting’’ and ‘‘supported sitting’’ the lumbar curva-ture angle in the sagittal plane was plotted by a defined histogram.Bins of a range of 5� within the total range from �40� (maximumlordosis) to 20� (maximum kyphotic posture of lumbar spine) wereused. Sporadic outliers below �40� or over 20� lumbar flexion an-gle were allocated to the first or last bin respectively. The coinci-dental RMS measurements of the investigated muscles werepresented as medians and quartiles over the occurrence of lumbarposture. The tasks ‘‘standing’’, ’’walking’’, ‘‘keyboard use’’ and ‘‘tele-phoning’’ were excluded from this analysis because of high varia-tion in muscle activation due to movements of the arms.

To detect statistical differences between the different officetasks, the Mann–Whitney U-test was used. For variations withindifferent postures during one task the Kruskal–Wallis-test (non-parametric version of classical one-way ANOVA) was used. The sig-nificance level for all tests was P = .05. Both nonparametric testswere used because there was no normal distribution of sEMG data.

3. Results

3.1. Occurrence of office tasks

The majority (over 82%) of office work was sedentary. Onlystanding and walking occurred in an erect body position and theseaccounted for just 5% of the total office time. In the sitting posture,unsupported sitting (41%) made up the main part, with supportedsitting only accounting for 17%. Other tasks like keyboard use (16%)and phoning (8%) were not that frequent. In summary, office workwas very passive, with 5% undertaken in an erect body position andonly 2% walking (Table 2).

3.2. Different tasks and lumbar muscle activation

This passivity was accentuated by the differences in lumbarmuscle activation within different seating postures. Neither typeof sitting activated the lumbar muscles a quarter as much as nor-malisation posture (Fig. 1). In unsupported sitting, the longissimusmuscle showed up to 20% of the normalisation posture whereas

−40 −30 −20 −10 0 10 200

0.5

1

1.5

rel R

MS

30

15

22.5

7.5

0

occu

renc

e [%

]

Lordosis Kyphosislumbar flexion angle [°]

occurenceLoL1 leLoL1 riMuL4 leMuL4 ri

lumbar posture and muscular activity in unsupported sitting

Fig. 2. Median and quartiles of lumbar muscle normalised activation (LoL1 –longissimus muscle at lumbar level L1, MuL4 – multifidus muscle at lumbar levelL4, le – left and ri – right side, respectively) and distribution of lumbar posture (greybars, right y-axis) during unsupported sitting. The activation of all lumbar muscleswithin lordotic postures is moderate but very low within kyphotic curvature.Lumbar muscle activation depends on posture (P < .001, for all muscles).

F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368 365

the multifidus muscle had lower activity at the level of 10% of nor-malisation. Supported sitting was effectively passive. The longissi-mus muscle showed up to 12% of normalised activation. Theactivity of the multifidus muscle was near the limit of the resolu-tion of the measurement device at only 5%. This low activation ofmultifidus muscle was below 5 lV (non-normalised value) formost cases. The differences between the described tasks are highlysignificant (P < .01).

3.3. Lumbar muscle activity and dependency on posture

The main outcome is the dependency of lumbar muscle activityon lumbar posture. In walking, as a reference, the lumbar spine had

−40 −30 −20 −10 0 10 200

0.5

1

1.5

rel R

MS

30

15

22.5

7.5

0

occu

renc

e [%

]

Lordosis Kyphosislumbar flexion angle [°]

occurenceLoL1 leLoL1 riMuL4 leMuL4 ri

lumbar posture and muscular activity in supported sitting

Fig. 3. Median and quartiles of lumbar muscle normalised activation (LoL1 –longissimus muscle at lumbar level L1, MuL4 – multifidus muscle at lumbar levelL4, le – left and ri – right side, respectively) and distribution of lumbar posture (greybars, right y-axis) during supported sitting on the backrest of the chair. Incomparison to Fig. 2 (unsupported sitting) the distribution of postures is shiftedright. There is very low activation of all lumbar muscles within kyphotic curvature.Lumbar muscle activation depends on posture (P < .001, for all muscles).

lordotic curvature (negative flexion angle). A permanent changefrom low (�10�) to pronounced lordosis (up to �40�) was foundwith regard to the gait cycle. Thus, there is little tilting of the trunkforwards (�10�) or backwards (�40�) during the gait cycle.

In contrast, the lumbar posture was flat or in kyphotic curvaturein sitting. The lumbar spine showed 47% lordotic curvature inunsupported sitting and only 30% in supported sitting. Both kindsof sitting showed only low lordotic angles (at minimum �25�) incomparison to walking. Further, the highest lumbar muscle activitywas to be found within the lordotic posture of the lumbar spine(Figs. 2 and 3). The statement ‘‘the more the lumbar spine is flexed,the lower the lumbar spine’s muscular activity’’ is true for bothkinds of sitting. The dependency of lumbar muscle activation onlumbar posture is highly significant for all investigated muscles(P < .001). In supported sitting, the main curvature of the lumbarspine was pronounced kyphosis (>15�, 30% of time). For this, stepby step, the pelvis slid onto the front edge of the seat and was tiltedbackwards, whereas the thighs were not supported. Within thisposture, the activation of lumbar muscles was very low and belowthe resolution limit of the measurement device (<5 lV) most of thetime.

3.4. Gaps in muscular activity during sitting

In phases of sitting with kyphotic lumbar posture, long periodsof very low (near the resolution limit) or no activity of lumbarmuscles were found. In unsupported sitting, the longissimus mus-cle at L1 showed 28% and the multifidus muscle 36% of no activity,respectively. In supported sitting this increased for the multifidusto 45%. The online filter for removing the ECG signal from the sEMGsignal of the muscles does not work perfectly. With this in mind itis appreciable that there is more time without lumbar muscle acti-vation: for the longissimus muscle, the periods increase to 32% inboth tasks; for the multifidus muscle this increases to 41% inunsupported sitting and to 51% in supported sitting. In summary,during sedentary office work, the lumbar muscles were deacti-vated about 40% of working time.

4. Discussion

Inactivity of spinal paravertebral muscles during trunk flexionhas been described for decades (Floyd and Silver, 1951, 1955;Ahern et al., 1988; Dolan and Adams, 1993; McGill and Kippers,1994; Newman and Gracovetsky, 1995). What is new in this studyis the accurate measurement of postures during long-term sitting.There was found to be a strong dependency of lumbar muscularactivity on lumbar spine posture in the sedentary body position.Only one study currently documents comparable data for women(Mork and Westgaard, 2009). In contrast to the cited study, herethe angle of curvature of the lumbar spine in the sagittal planewas deduced as a more precise measure of lumbar spine posture.Furthermore the lumbar curvature has an impact on the sEMG oflumbar muscles. Only global angles for trunk and thighs were mea-sured in the earlier study (Mork and Westgaard, 2009). It is possi-ble to incline or bend the trunk with the lumbar spine in slightlordotic curvature. This may be the reason why the activity of lum-bar muscles is more affected by the pelvis inclination angle andless by the trunk inclination angle. A flexed lumbar spine (kyphoticposture) leads to very low activity or phases of no activity of thelumbar muscles. This means that the flexion–relaxation phenome-non is present in sitting. In laboratory settings, examining the flex-ion–relaxation phenomenon during sitting showed no clear effecton the lumbar muscles (O’Sullivan et al., 2006; Callaghan andDunk, 2002). The measurement times in the cited studies werevery short in comparison to the data documented here. In the field

366 F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368

setting, while doing their own work at their own workplace, thesubjects focused on the task in hand and their lumbar musclesrelaxed.

Because the load on the lumbar spine in the sitting posture isnot small, and there is no adequate muscular support, passivestructures like ligaments and intervertebral discs or passive muscleproperties have to carry the load (Nachemson, 1966; Wilke et al.,2001; Schultz et al., 1985; Sihvonen et al., 1988). When sitting sup-ported by the backrest of the chair, stress on the disc is reduced,whereas intradiscal stress is about three times greater in unsup-ported flexed sitting (Wilke et al., 1999). There is no knowledgeabout stress on the ligaments during sedentary work in vivo cur-rently. Biomechanical models may give hints, but depending onthe optimisation parameter used and components included (e.g.ligaments, muscles) different results were predicted (McGill,1986; Cholewicki et al., 1995; McGill et al., 1994; Arjmand andShirazi-Adl, 2005, 2006; McGill et al., 2006; Brown and McGill,2008).

Due to their viscoelasticity and the long-lasting loadingcaused by sitting, creep of passive structures can be predicted(Adams and Dolan, 1996; Solomonow et al., 1998, 1999). Stepby step, this may lead to changes in mechanical properties anddysfunction of these structures and also to muscular dysfunction(Solomonow et al., 2003). The load on the human lumbar spineduring sitting is probably not directly comparable to the citedexperiments, but consideration of the daily reiterative long-lasting periods of sitting during office work (around 8 h) andthe long times for recovery of passive structures may supportthis assumption (Gedalia et al., 1999). First studies also giveevidence of this hypothetical behaviour in humans (Olsonet al., 2009; Shin et al., 2009). In reality, low back pain doesnot develop after some hours of office work, but the fact that lifein western societies is more and more physically passive (sittingat desks, sitting to drive cars, resting in an elevator, sitting onthe couch and watching TV, etc.) may contribute to the long-term hypothesis described.

Distinct passivity due to bedrest and low back pain are asso-ciated with atrophy or fatty replacement of paravertebral mus-cles (Hadar et al., 1983; McConnell and Daneman, 1984;Mattila et al., 1986; Cooper et al., 1992; Hides et al., 1994,2007; Bloomfield, 1997; Ng et al., 1998; Danneels et al., 2000;Kader et al., 2000; Yoshihara et al., 2001, 2003; Barker et al.,2004; Hyun et al., 2007; Belavy et al., 2008; Lee et al., 2008).In an animal study, the main reason for atrophy of lumbar mus-cles was nonexistent neural activation (Hodges et al., 2006). Onlyone study documents the coherence of physical inactivity, lowback pain and atrophy of paravertebral muscles in children(Salminen et al., 1993). It is currently speculative to assume thatyears of office work without active leisure lead to atrophy ofparavertebral muscles. However, the increase in the cross sec-tional area of back muscles due to training in patients sufferingfrom low back pain would support this assumption (Rissanenet al., 1995). If there is atrophy because of years of office work,the muscle function to stabilise the spine is reduced and may bea reason for the high prevalence of low back pain. Longitudinalstudies on the cross sectional area, function or force of lumbarmuscles are necessary to investigate this further.

Paravertebral muscles and passive structures of the spine arenot independent from each other. Changes in the mechanical prop-erties of both structures and the disappearance of reflexes due tolong-lasting (also low) loading, can cause the whole spine to beout of tune (Solomonow et al., 2003; Panjabi, 1992a,b). In this con-dition of dysfunction, increased degeneration and low back paincan be assumed.

The sample of 13 subjects is small and limits the explanatorypower of this study. More subjects may lead to more variation in

lumbar posture and lumbar muscle activation within the investi-gated tasks. In general, significant deviations from the results pre-sented here should not be assumed: Firstly, low or no activation ofthe lumbar muscles in flexed postures of the trunk also seem to bemechanically linked in lifting (Schultz et al., 1985; Toussaint et al.,1995). Secondly, the medians of lumbar muscle activation pre-sented here are (near) zero within kyphotic lumbar posture duringsedentary work. Thus, 50% of time within such postures there islow or no activation of the lumbar muscles.

5. Conclusions

In conclusion, to reduce the high prevalence of low backpain in sedentary work, reasonable prevention is necessary.Considering the low activation of lumbar muscles in the sittingposture, the use of instable seats or special office chairs to pro-tect the spine or to train the paravertebral muscles will fail.This is because lumbar muscle activation depends more onthe task than on the office chair used (van Dieën et al.,2001). Further, lumbar muscle activation does not differ whenseated on an exercise ball, different dynamic office chairs oron a reference chair (McGill et al., 2006; Ellegast et al., 2012).Again the spine has to be seen as a complex of muscles, passivestructures and neural control: In the sitting posture, the passivestructures are stretched/loaded, the muscles only have low acti-vation within the stretched condition, and the lumbar spine isin flat or kyphotic curvature. A natural way to activate the par-avertebral muscles within the normal length conditions withinthe lordotic curvature of the lumbar spine may be breaks forwalking, working in standing position and active leisure in erectbody position.

Conflict of interest

The authors have no conflict of interest, no funding.

Acknowledgments

Thanks to Carol Keelan for language assistance. Thanks toSabine Franke for data collection.

References

Adams M, Dolan P. Time-dependent changes in the lumbar spine’s resistance tobending. Clin Biomech (Bristol, Avon) 1996;11(4):194–200.

Ahern DK, Follick MJ, Council JR, Laser-Wolston N, Litchman H. Comparison oflumbar paravertebral EMG patterns in chronic low back pain patients and non-patient controls. Pain 1988;34(2):153–60.

Aminian K, Najafi B. Capturing human motion using body-fixed sensors: outdoormeasurement and clinical applications. Comput Anim Virtual Worlds2004;15(2):79–94.

Arjmand N, Shirazi-Adl A. Biomechanics of changes in lumbar posture in staticlifting. Spine 2005;30(23):2637–48.

Arjmand N, Shirazi-Adl A. Role of intra-abdominal pressure in the unloading andstabilization of the human spine during static lifting tasks. Eur Spine J2006;15(8):1265–75.

Barker KL, Shamley DR, Jackson D. Changes in the cross-sectional area of multifidusand psoas in patients with unilateral back pain: the relationship to pain anddisability. Spine 2004;29(22):E515–9.

Belavy DL, Hides JA, Wilson SJ, Stanton W, Dimeo FC, et al. Resistive simulatedweightbearing exercise with whole body vibration reduces lumbar spinedeconditioning in bed-rest. Spine 2008;33(5):E121–31.

Bloomfield SA. Changes in musculoskeletal structure and function with prolongedbed rest. Med Sci Sports Exerc 1997;29(2):197–206.

Brown SHM, McGill SM. How the inherent stiffness of the in vivo human trunkvaries with changing magnitudes of muscular activation. Clin Biomech (Bristol,Avon) 2008;23(1):15–22.

Burdorf A, Naaktgeboren B, de Groot HC. Occupational risk factors for low back painamong sedentary workers. J Occup Med 1993;35(12):1213–20.

Callaghan JP, Dunk NM. Examination of the flexion relaxation phenomenon inerector spinae muscles during short duration slumped sitting. Clin Biomech(Bristol, Avon) 2002;17(5):353–60.

F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368 367

Cholewicki J, McGill SM, Norman RW. Comparison of muscle forces and joint loadfrom an optimization and EMG assisted lumbar spine model: towardsdevelopment of a hybrid approach. J Biomech 1995;28(3):321–31.

Cooper RG, Forbes WSC, Jayson MI. Radiographic demonstration of paraspinalmuscle wasting in patients with chronic low back pain. Br J Rheumatol1992;31(6):389–94.

Danneels LA, Vanderstraeten GG, Cambier DC, Witvrouw EE, Cuyper HJD. Ct imagingof trunk muscles in chronic low back pain patients and healthy control subjects.Eur Spine J 2000;9(4):266–72.

van Dieën JH, de Looze MP, Hermans V. Effects of dynamic office chairs on trunkkinematics, trunk extensor EMG and spinal shrinkage. Ergonomics2001;44(7):739–50.

Dolan P, Adams MA. The relationship between EMG activity and extensor momentgeneration in the erector spinae muscles during bending and lifting activities. JBiomech 1993;26(4-5):513–22.

Dunk NM, Callaghan JP. Lumbar spine movement patterns during prolonged sittingdifferentiate low back pain developers from matched asymptomatic controls.Work 2010;35(1):3–14.

Ellegast RP, Kraft K, Groenesteijn L, Krause F, Berger H, Vink P. Comparison of fourspecific dynamic office chairs with a conventional office chair: impact uponmuscle activation, physical activity and posture. Appl Ergon2012;43(2):296–307.

Floyd WF, Silver PH. Function of erectores spinae in flexion of the trunk. Lancet1951;1(6647):133–4.

Floyd WF, Silver PH. The function of the erectores spinae muscles in certainmovements and postures in man. J Physiol 1955;129(1):184–203.

Gedalia U, Solomonow M, Zhou BH, Baratta RV, Lu Y, Harris M. Biomechanics ofincreased exposure to lumbar injury caused by cyclic loading. Part 2. Recoveryof reflexive muscular stability with rest. Spine (Phila Pa 1976)1999;24(23):2461–7.

Hadar H, Gadoth N, Heifetz M. Fatty replacement of lower paraspinal muscles:normal and neuromuscular disorders. AJR Am J Roentgenol 1983;141(5):895–8.

Hayashi N, Tamaki T, Yamada H. Experimental study of denervated muscle atrophyfollowing severance of posterior rami of the lumbar spinal nerves. Spine1992;17(11):1361–7.

Hemingway H, Shipley MJ, Stansfeld S, Marmot M. Sickness absence from back pain,psychosocial work characteristics and employment grade among officeworkers. Scand J Work Environ Health 1997;23(2):121–9.

Hermens H, Freriks B, Merletti R, Stegeman D, Blok J, Rau G, et al. EuropeanRecommendations for Surface ElectroMyoGraphy, results of the SENIAMproject. Roessingh Research and Development B.V., 1999.

Hides JA, Belavy DL, Stanton W, Wilson SJ, Rittweger J, et al. Magnetic resonanceimaging assessment of trunk muscles during prolonged bed rest. Spine2007;32(15):1687–92.

Hides JA, Stokes MJ, Saide M, Jull GA, Cooper DH. Evidence of lumbar multifidusmuscle wasting ipsilateral to symptoms in patients with acute/subacute lowback pain. Spine 1994;19(2):165–72.

Hodges P, Holm AK, Hansson T, Holm S. Rapid atrophy of the lumbarmultifidus follows experimental disc or nerve root injury. Spine2006;31(25):2926–33.

Hyun JK, Lee JY, Lee SJ, Jeon JY. Asymmetric atrophy of multifidus muscle inpatients with unilateral lumbosacral radiculopathy. Spine 2007;32(21):E598–602.

Janwantanakul P, Pensri P, Jiamjarasrangsri V, Sinsongsook T. Prevalence of self-reported musculoskeletal symptoms among office workers. Occup Med (Lond)2008;58(6):436–8.

Johansson H, Sjölander P, Sojka P. A sensory role for the cruciate ligaments. ClinOrthop Relat Res 1991(268):161–78.

Juul-Kristensen B, Jensen C. Self-reported workplace related ergonomic conditionsas prognostic factors for musculoskeletal symptoms: the bit follow up study onoffice workers. Occup Environ Med 2005;62(3):188–94.

Juul-Kristensen B, Søgaard K, Strøyer J, Jensen C. Computer users’ risk factors fordeveloping shoulder, elbow and back symptoms. Scand J Work Environ Health2004;30(5):390–8.

Kader DF, Wardlaw D, Smith FW. Correlation between the MRI changes in thelumbar multifidus muscles and leg pain. Clin Radiol 2000;55(2):145–9.

Lee JC, Cha JG, Kim Y, Kim YI, Shin BJ. Quantitative analysis of back muscledegeneration in the patients with the degenerative lumbar flat back using adigital image analysis: comparison with the normal controls. Spine2008;33(3):318–25.

Mattila M, Hurme M, Alaranta H, Paljärvi L, Kalimo H, Falck B, et al. The multifidusmuscle in patients with lumbar disc herniation. A histochemical andmorphometric analysis of intraoperative biopsies. Spine 1986;11(7):732–8.

McConnell JR, Daneman A. Fatty replacement of muscles adjacent to spinal osteoidosteoma. J Comput Assist Tomogr 1984;8(1):147–8.

McGill S. Partitioning of the l4–l5 dynamic moment into discs, ligamentous andmuscular components during lifting. Spine 1986;7:666–78.

McGill S, Seguin J, Bennett G. Passive stiffness of the lumbar torso in flexion,extension, lateral bending, and axial rotation. Effect of belt wearing and breathholding. Spine (Phila Pa 1976) 1994;19(6):696–704.

McGill SM, Kavcic NS, Harvey E. Sitting on a chair or an exercise ball: variousperspectives to guide decision making. Clin Biomech (Bristol, Avon)2006;21(4):353–60.

McGill SM, Kippers V. Transfer of loads between lumbar tissues during the flexion–relaxation phenomenon. Spine (Phila Pa 1976) 1994;19(19):2190–6.

Mork PJ, Westgaard RH. Back posture and low back muscle activity in femalecomputer workers: a field study. Clin Biomech (Bristol, Avon)2009;24(2):169–75.

Mörl F, Anders C, Graßme R. An easy and robust method for ecg artifact eliminationof SEMG signals. In: ISEK. 2010.

Mörl F, Blickhan R. Three-dimensional relation of skin markers to lumbar vertebraeof healthy subjects in different postures measured by open MRI. Eur Spine J2006;15(6):742–51.

Nachemson A. The load on lumbar discs in different positions of the body. ClinOrthop 1966;45:107.

Newman NM, Gracovetsky SA. Flexion–relaxation phenomenon (transfer of loadsbetween lumbar tissues during the flexion–relaxation phenomenon). Spine(Phila Pa 1976) 1995;20(15):1739–40.

Ng JK, Richardson CA, Kippers V, Parnianpour M. Relationship between musclefiber composition and functional capacity of back muscles in healthysubjects and patients with back pain. J Orthop Sports Phys Ther1998;27(6):389–402.

Olson MW, Li L, Solomonow M. Interaction of viscoelastic tissue compliance withlumbar muscles during passive cyclic flexion–extension. J ElectromyogrKinesiol 2009;19(1):30–8.

O’Sullivan P, Dankaerts W, Burnett A, Chen D, Booth R, Carlsen C, et al. Evaluation ofthe flexion relaxation phenomenon of the trunk muscles in sitting. Spine (PhilaPa 1976) 2006;31(17):2009–16.

Panjabi MM. The stabilizing system of the spine. Part I. Function, dysfunction,adaptation, and enhancement. J Spinal Disord 1992a;5(4):383–9 [discussion397].

Panjabi MM. The stabilizing system of the spine. Part II. Neutral zone and instabilityhypothesis. J Spinal Disord 1992b;5(4):390–6 [discussion 397].

de-las Peñas CF, Albert-Sanchís JC, Buil M, Benitez JC, Alburquerque-Sendín F.Cross-sectional area of cervical multifidus muscle in females with chronicbilateral neck pain compared to controls. J Orthop Sports Phys Ther2008;38(4):175–80.

Riihimäki H, Tola S, Videman T, Hänninen K. Low-back pain and occupation: a cross-sectional questionnaire study of men in machine operating, dynamic physicalwork, and sedentary work. Spine 1989;14(2):204–9.

Riihimäki H, Viikari-Juntura E, Moneta G, Kuha J, Videman T, Tola S.Incidence of sciatic pain among men in machine operating, dynamicphysical work, and sedentary work: a three-year follow-up. Spine1994;19(2):138–42.

Rissanen A, Kalimo H, Alaranta H. Effect of intensive training on the isokineticstrength and structure of lumbar muscles in patients with chronic low backpain. Spine 1995;20(3):333–40.

Salminen JJ, Erkintalo-Tertti MO, Paajanen HE. Magnetic resonance imaging findingsof lumbar spine in the young: correlation with leisure time physical activity,spinal mobility, and trunk muscle strength in 15-year-old pupils with orwithout low-back pain. J Spinal Disord 1993;6(5):386–91.

Schultz AB, Haderspeck-Grib K, Sinkora G, Warwick DN. Quantitative studies ofthe flexion–relaxation phenomenon in back muscles. J Orthop Res1985:189.

Shin G, D’Souza C, Liu YH. Creep and fatigue development in the low back in staticflexion. Spine (Phila Pa 1976) 2009;34(17):1873–8.

Sihvonen T, Partanen J, Hänninen O. Averaged (rms) surface EMG in testing backfunction. Electromyogr Clin Neurophysiol 1988;28(6):335–9.

Solomonow M, Baratta R, Zhou BH, Shoji H, Bose W, Beck C, et al. The synergisticaction of the anterior cruciate ligament and thigh muscles in maintaining jointstability. Am J Sports Med 1987;15(3):207–13.

Solomonow M, Baratta RV, Zhou BH, Burger E, Zieske A, Gedalia A. Musculardysfunction elicited by creep of lumbar viscoelastic tissue. J ElectromyogrKinesiol 2003;13(4):381–96.

Solomonow M, Zhou BH, Baratta RV, Lu Y, Harris M. Biomechanics of increasedexposure to lumbar injury caused by cyclic loading: Part 1. loss of reflexivemuscular stabilization. Spine 1999;24(23):2426–34.

Solomonow M, Zhou BH, Harris M, Lu Y, Baratta RV. The ligamento-muscularstabilizing system of the spine. Spine 1998;23(23):2552–62.

Spyropoulos P, Papathanasiou G, Georgoudis G, Chronopoulos E, Koutis H,Koumoutsou F. Prevalence of low back pain in greek public office workers.Pain Phys 2007;10(5):651–9.

Toussaint HM, de Winter AF, de Haas Y, de Looze MP, van Dieen JH, Kingma I.Flexion relaxation during lifting: implications for torque production by muscleactivity and tissue strain at the lumbo-sacral joint. J Biomech1995;28(2):199–210.

Törner M, Zetterberg C, Andén U, Hansson T, Lindell V. Workload andmusculoskeletal problems: a comparison between welders and office clerks(with reference also to fishermen). Ergonomics 1991;34(9):1179–96.

Videman T, Battie M. The influence of occupation on lumbar degeneration. Spine1999;24(11):1164–8.

Wilke H, Neef P, Hinz B, Seidel H, Claes L. Intradiscal pressure together withanthropometric data – a data set for the validation of models. Clin Biomech(Bristol, Avon) 2001;16(1):111–26.

Wilke HJ, Neef P, Caimi M, Hoogland T, Claes LE. New in vivo measurements ofpressures in the intervertebral disc in daily life. Spine (Phila Pa 1976)1999;24(8):755–62.

Yoshihara K, Nakayama Y, Fujii N, Aoki T, Ito H. Atrophy of the multifidus muscle inpatients with lumbar disk herniation: histochemical and electromyographicstudy. Orthopedics 2003;26(5):493–5.

368 F. Mörl, I. Bradl / Journal of Electromyography and Kinesiology 23 (2013) 362–368

Yoshihara K, Shirai Y, Nakayama Y, Uesaka S. Histochemical changes in themultifidus muscle in patients with lumbar intervertebral disc herniation. Spine2001;26(6):622–6.

Falk Mörl received his Master in Sport Science fromthe University of Jena in 1999 and his PhD in 2004.From 2000 to 2001 he has been postgraduate at theChair of Motion Science at the University of Jena.Since 2002 he has been a scientist at FSA mbH. Hisresearch focuses on passive properties of the lumbarspine and biomechanical modelling.

Ingo Bradl studied Physics and received his PhD(theoretical physics) in 1987 from the TechnicalUniversity ‘‘Otto von Guericke’’, Magdeburg. Since1995 he works at the Biomechanics group within thePrevention Department of the German Social Acci-dent Insurance Institution for the foodstuffs andcatering industry. Since 2003 he is the head of theBiomechanics group. He is guest scientist at theUniversity Hospital Jena, Clinic for Trauma-, Hand-and Reconstructive Surgery, Division for MotorResearch, Pathophysiology and Biomechanics. Hisresearch focuses on the prevention of occupationalinduced musculoskeletal disease and on the devel-

opment and application of related mobile measurement methods.