at SciVerse ScienceDirect

Manual Therapy 17 (2012) 312e317

Contents lists available

Manual Therapy

journal homepage: www.elsevier .com/math

Original article

Comparing lower lumbar kinematics in cyclists with low back pain (flexionpattern) versus asymptomatic controls e field study using a wirelessposture monitoring system

Wannes Van Hoof a,*, Koen Volkaerts a, Kieran O’Sullivan b, Sabine Verschueren a, Wim Dankaerts a

aMusculoskeletal Research Unit, Department of Rehabilitation Sciences, Faculty of Kinesiology and Rehabilitation sciences, Katholieke Universiteit Leuven,Tervuursevest 101, B-3001 Leuven, BelgiumbDepartment of Physiotherapy, Faculty of Education and Health Sciences, University of Limerick, Limerick, Ireland

a r t i c l e i n f o

Article history:Received 13 July 2011Received in revised form16 February 2012Accepted 22 February 2012

Keywords:Low back pain (LBP)Flexion patternCyclingLower lumbar kinematics

* Corresponding author. Tel.: þ32 16 32 91 24; fax:E-mail address: wannes.vanhoof@faber.kuleuven.b

1356-689X/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.math.2012.02.012

a b s t r a c t

The aim of this study was to examine lower lumbar kinematics in cyclists with and without non-specificchronic low back pain (NS-CLBP) during a cross-sectional cycling field study. Although LBP is a commonproblem among cyclists, studies investigating the causes of LBP during cycling are scarce and are mainlyfocussed on geometric bike-related variables. Until now no cycling field studies have investigated therelationship between maladaptive lumbar kinematics and LBP during cycling. Eight cyclists with NS-CLBPclassified as having a ‘Flexion Pattern’ (FP) disorder and nine age- and gender-matched asymptomaticcyclists were tested. Subjects performed a 2 h outdoor cycling task on their personal race bike. Lowerlumbar kinematics was measured with the BodyGuard� monitoring system. Pain intensity during andafter cycling was measured using a numerical pain rating scale. The NS-CLBP (FP) subjects weresignificantly more flexed at the lower lumbar spine during cycling compared to healthy controls(p ¼ 0.018), and reported a significant increase in pain over the 2 h of cycling (p < 0.001). One-wayrepeated measures ANOVA revealed a significant main effect for group (p ¼ 0.035, F ¼ 5.546) whichremained just significant when adding saddle angle as a covariate (p ¼ 0.05, F ¼ 4.747). The difference inposture between groups did not change over time. These findings suggest that a subgroup of cyclists withNS-CLBP (FP) demonstrate an underlying maladaptive motor control pattern resulting in greater lowerlumbar flexion during cycling which is related to a significant increase in pain.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Cycling is one of the most popular sports throughout the world,with cyclists experiencing a variety of injuries (Mellion, 1991;Callaghan and Jarvis, 1996; Rivara et al., 1997; Thompson andRivara, 2001; Stone and Broughton, 2003; Tin Tin et al., 2010).Overuse injuries like low back pain (LBP) are mostly mild but canbecome chronic and can lead to functional impairment requiringmedical attention (Clarsen et al., 2010). The prevalence of LBPamong cyclists ranges between 31 and 60% (Mellion, 1991; Wilberet al., 1995; Callaghan and Jarvis, 1996; Salai et al., 1999; Clarsenet al., 2010).

In the majority of chronic LBP disorders there is no detectablepatho-anatomic/radiologic abnormality (Dankaerts et al., 2006)

þ32 16 32 91 97.e (W. Van Hoof).

All rights reserved.

leading to a “non-specific” chronic LBP (NS-CLBP) classification(Dillingham, 1995; Waddell, 1998). The NS-CLBP population hasbeen proposed to consist of a heterogeneous group requiring sub-classification (Dankaerts et al., 2006). A bio-psycho-social classifi-cation framework for NS-CLBP patients, based on the underlyingpain mechanisms, has been developed (O’Sullivan, 2005a). One ofthe proposed subgroups is the Flexion Pattern (FP) which is themost common clinical LBP pattern (O’Sullivan, 2000, 2005b) withinthe ‘maladaptive motor control impairment’ subgroup (O’Sullivan,2000, 2005a, 2005b). The repetitive and sustained near end-range flexion strain associated with this disorder (O’Sullivanet al., 2003) can adversely affect the spinal tissues (i.e. ligaments,intervertebral discs, zygaphophyseal joints and capsular structures)and aggravate LBP (Shirazi-Adl and Drouin, 1987; McKenzie, 1989;Pope, 1989; Adams et al., 1994; Callaghan and McGill, 2001; Dunket al., 2009). Cyclists with LBP are thought to commonly presentwith this FP disorder (Burnett et al., 2004). One of the few LBPstudies among cyclists intermittently monitored and compared thelumbar kinematics among cyclists with NS-CLBP (FP) and non-LBP

Table 1Inclusion and exclusion criteria for NS-CLBP patients.a

Inclusion criteriaDiagnosis of non-specific Chronic Low Back Pain with a ‘Flexion Pattern LBP

disorder’ considered directly attributable to the activity of cyclingThe Flexion Pattern was confirmed by two physiotherapists using a ‘Flexion

Pattern sub-classification schema’ that is based on the literature (O’Sullivan,2005b)

Mechanical provocation of pain: cycling and sitting are dominant aggravatingactivities

No or minimal pain (NPRS: <2/10) at start, developing pain during ‘prolongedcycling’ (within 1 h)

Continuous or recurrent symptoms for 3 months or moreModerate ongoing LBP, average cycling pain (NPRS) > 3/10

Exclusion criteriaHistory of structural pathology at the lower back region, with evidence

of specific diagnosis (e.g. spondylolisthesis, stenosis, inflammatory disease)History of significant surgical intervention at the lower back regionSigns of neurologic involvement (radicular pain)Non-mechanical painMore generalized pain (widespread non-specific pain disorder)Decision of a misclassified person after careful consideration between the two

physiotherapists when there was no agreement upon the Flexion Patternsub-classification of an LBP patient between the two physiotherapists

Presence of red flag pathologyCardiac disorders

NPRS: Numerical Pain Rating Scale; LBP: Low Back Pain.a All features within the inclusion criteria had to be present.

W. Van Hoof et al. / Manual Therapy 17 (2012) 312e317 313

subjects. They found that NS-CLBP (FP) cyclists demonstrated anincreased flexion/rotation strain across the lower lumbar spineclinically linked with the development of LBP (Burnett et al., 2004).

However to date, the causes of LBP in cyclists have mainly beenrestricted to non-cycling laboratory biomechanical studies. Fromthis research, several patho-mechanical mechanisms have beenproposed to explain the development of LBP in cyclists. Threeplausible mechanisms, related to sustained and repeated lowerlumbar flexion can be extrapolated to the FP LBP subgroup. Firstly,the mechanical loads generated by the lower limbs during cyclingare transferred through a flexed and/or flexed/rotated position ofthe thoracolumbar spine (Nachemson, 1999; Drake et al., 2005;Burnett et al., 2008; Drake and Callaghan, 2008). Secondly, theflexion-relaxation phenomenon (FRP) might occur, which refers tomyo-electrical silence in the back extensor muscles at the mid-toend-range of trunk flexion (Callaghan and Dunk, 2002; Olsonet al., 2004; Colloca and Hinrichs, 2005; O’Sullivan et al., 2006a).Thirdly, sustained flexion may result in mechanical creep of thespine, a deformation of visco-elastic structures under constantforce (McGill and Brown, 1992; Solomonow et al., 1999; Little andKhalsa, 2005).

As a consequence of adopting and sustaining more end-rangelower lumbar flexion during cycling, resulting from a maladaptivecontrol impairment (e.g. FP), all thesemechanisms can (individuallyor together) play an important role in the provocation of LBP duringcycling through overloading the spinal structures (Kuslich et al.,1991; Dolan et al., 1994; McGill and Kippers, 1994; Gupta, 2001).

Table 2Baseline characteristics of both the NS-CLBP and non-LBP group.

Age (y) Weight (kg) Height (cm) BMI (kg/m

LBP (n ¼ 8) 28.3 (8.7) 76.2 (8.5) 184.9 (4.1) 22.3 (2.7)non-LBP (n ¼ 9) 28.4 (9) 75.1 (7.7) 181.2 (2.7) 22.8 (1.9)

Values aremean (SD); LBP: Low Back Pain; non-LBP: no Low Back Pain; BMI: BodyMass Indweight, height or BMI were all p > 0.05.

a Negative value indicates the degrees above 90� .

No cycling studies have yet performed continuous monitoring oflumbar spine kinematics in the field despite its potential impor-tance (Marsden and Schwellnus, 2010). The benefits of field-testinginclude; allowing subjects to use their own usual bicycle, incor-porating real-life variables such as wind resistance and reducingthe risk of changing the cyclist’s performance due to the feeling ofbeing observed in a laboratory testing environment.

Therefore the main aim of this study was to examine lowerlumbar kinematics in cyclists with and without NS-CLBP (FP)during a 2 h cycling field study.

2. Methods

2.1. Subjects

Twenty-four male cyclists were recruited from seven differentlocal cycling clubs by email. Because of the specific logistics of thetesting (strain gauges loosening and protocol independent personalreasons) four NS-CLBP (FP) subjects and three non-LBP subjectswere unable to be further evaluated. The cyclists with NS-CLBPwere (independently) assessed by two physiotherapists usingstrict inclusion and exclusion criteria (Table 1). Only cyclists pre-senting with a clear FP LBP disorder, considered directly attribut-able to the activity of cycling, were selected.

LBP and non-LBP cyclists were matched for age (�2 y) andgender. Independent t-tests showed no between group significantdifferences in age, weight, height or Body Mass Index (all p > 0.05).The non-LBP group contained nine cyclists without a history ofsignificant LBP (requiring intervention) and without signs andsymptoms of LBP in the preceding three months. The LBP groupcontained eight cyclists with a clear history of NS-CLBP. Subjectcharacteristics are outlined in Table 2.

The studywas approved by the local university ethics committeeand according to the declaration of Helsinki. Since subjects regularlycycled for prolonged periods and reported LBP doing this, theywereinformed that they were likely to experience LBP during the 2-hcycling task. Subjects were informed that they could stop thetesting at any moment if they wanted. All subjects gave writteninformed consent and were instructed not to take part in any sport,physical activity or training 24 h prior to the testing day.

2.2. Experimental protocol and instrumentation

Subjects performed a 2-h outdoor cycling task on a standard flatparcours on their personal race bike. They were instructed to cycleas usual and were guided by a heart rate monitor (Polar, Finland) tomaintain a heart rate between 60 and 70% of their age-predictedmaximum throughout the cycling task (Burnett et al., 2004;ACSM, 2006). Each cycling task started at 1.30 pm to avoid thepotential negative influence of diurnal variation (Russell et al.,1992; McGill, 2004).

Lower lumbar kinematics were measured using a remoteposture monitoring system (BodyGuard�) (Sels Instruments nv,

2) Average pain(NPRS; 0e10) 4wprior (cycling)

Average pain(NPRS; 0e10) 4wprior (ADL)

Years ofcycling

Saddleangle (�)

5.6 (1.2) 3.3 (1.8) 7.3 (2.5) �0.1 (2.9)a

0 0 8.4 (5.1) 2.2 (2.6)

ex; NPRS: Numerical Pain Rating Scale; w: weeks; differences between group in age,

W. Van Hoof et al. / Manual Therapy 17 (2012) 312e317314

Belgium) (Fig. 1). The BodyGuard� consists of a very thin straingauge connected with a small and lightweight signal processingunit (56� 71�15mm). Each strain gauge has two rectangular end-pieces (10 � 28 mm) (Fig. 1), with the distance between these twoends increasing during spinal flexion. Elongation of the strain gaugealters its internal resistance and therefore the voltage of the gauge.This alteration in voltage occurs in a linear manner in response toelongation. Therefore, the voltage output is directly related to thelength (flexion vs. extension) of the strain gauge. By initially cali-brating this output relative to end-range flexion and extension,subsequent lower lumbar kinematics can be expressed asa percentage of total lumbo-pelvic flexion (% Fl ROM). The resis-tance of the gauge to elongation is minimal, with a tensile strengthof 1300 psi, and 50 g being sufficient to elongate a 100 mm straingauge by 100%. The device can measure and store data for a rela-tively long period of time (24 h), which makes it useful for longi-tudinal field studies with minimal to no task interference. Thesampling frequency was 20 Hz. Intra- and inter-rater reliabilityhave been shown to be excellent (ICC values: 0.837e0.874 and0.914e0.940 respectively) (O’Sullivan et al., 2011a). Recently, theconcurrent validity of the BodyGuard� has been supported foranalysing lower lumbar kinematics during functional tasks(rs ¼ 0.88, r2 ¼ 0.78, mean difference <10% Fl ROM) (O’Sullivanet al., 2012). Additionally, the correlation of BodyGuard� datawith a laboratory-based motion analysis system (CODA) (Charn-wood Dynamics Ltd, Leicestershire UK) during ergometer cyclingwas evaluated (n ¼ 12) in advance and was strong (r ¼ 0.8), witha mean difference of three degrees.

2.2.1. Procedures of BodyGuard placementSensors were attached with tape (Strappal (2.5 cm) and leuko-

tape) at the level of S2 and L3 spinous processes (Fig. 1). These twopoints were chosen for several reasons. Firstly, they are used todefine the lower lumbar angle and the region between S2 and L3 isregarded as the lower lumbar region (Dankaerts et al., 2006);Secondly, it has been shown that the upper (T12eL3) and lowerlumbar (L3eS1) region have a degree of functional independenceand that both regions should be considered separately (Mitchellet al., 2008). Finally, the lower lumbar spine is the most symp-tomatic region in LBP (Burnett et al., 2004; Dankaerts et al., 2006).For each individual the optimal length of strain gauge was chosen(ranging from 50 to 120 mm) to allow maximal flexion. To ensurethe equipment did not slip off and based on pilot testing, skin wascleaned with alcohol and sensor position was checked after thecycling task (based on reference lines on S2 and L3, see Fig. 1).

2.2.2. Calibration of BodyGuardTo calibrate the BodyGuard� subjects were instructed

and manually guided by a physiotherapist to perform maximal

Fig. 1. Cyclist with BodyGuard� and strain gauge attached with tape at the level of S2and L3 spinous processes.

extension and flexion of the lower lumbar spine while sitting ona stool with knees and hips 90� flexed. End-range lumbar extensionwas achieved by maximally anterior rotating the pelvis withoutmoving the thoracic spine. End-range flexion was achieved byrotating the pelvis maximally posterior and bending the trunkforward in between the legs. This movement was repeated threetimes and the fourth time subjects had to maintain the end-rangepositions for 5 s. End-range extension was set as the 0%-flexion-position and end-range flexion was set as the 100%-flexion-posi-tion. To ensure data quality, re-calibration was done in exactly thesame manner after the cycling task.

A long arm goniometer (Gymna, Belgium) was used to measurethe saddle angle of each subject’s personal race bike (Salai et al.,1999). Saddle design was documented descriptively.

The Numerical Pain Rating scale (NPRS) was used to measurethe level of LBP. This is an 11-point scale ranging from 0 (no pain) to10 (worst imaginable pain) that has been demonstrated to be valid,reliable and appropriate for use in clinical practice (Williamson andHoggart, 2005). A two-point change on the NPRS has been identi-fied as the minimal clinically important difference (Childs et al.,2005; Ostelo et al., 2008). The level of pain of each LBP subjectwas measured at start, every 15 min during cycling (by an inves-tigator cycling with them) and at 30 min, one, two and 24 h aftercycling.

2.3. Data analysis

BodyGuard� data were downloaded to a personal computer,uploaded to Microsoft Excel and compressed from 20 Hz to anaverage value for each minute and an average value per 10 min ofcycling. Data were normally distributed (KolmogoroveSmirnov,p > 0.05). An independent t-test was used to determine if differ-ences in saddle angle, mean lumbo-pelvic posture or posturalvariation (using Standard Deviation, SD) over the entire 2 h cyclingperiod existed between groups. To determine if the level of LBPreported by the cyclists with LBP changed during testing, a repeatedmeasures ANOVA (with post-hoc Bonferroni) was used. Sincea repeated measures ANOVA demonstrated no significant change(p > 0.05) in lumbo-pelvic posture in either group over time,a separate one-way repeated measures ANOVA (with post-hocBonferroni) was used to determine if lumbo-pelvic posturechanged in the two groups over time across the 12 intervals of10 min. This one-way repeated measures ANOVA was thenrepeated with saddle angle as a covariate. All statistical analyseswere performed using SPSS Version 16.0. The significance level wasset at p � 0.05, and adjusted appropriately using the Bonferronicorrection where indicated.

3. Results

Fig. 2 shows the averge pain scores of both groups during cyclingand the follow-up periods. The level of pain reported by the LBPgroup increased significantly over time during cycling (p < 0.001,F ¼ 51.4), whereby the pain peaked towards the end of the cyclingperiod and increased significantly from baseline after 90 (p ¼ 0.05),105 (p ¼ 0.04) and 120 (p ¼ 0.01) minutes of cycling.

After cycling the level of pain decreased but remained slightlyelevated compared to baseline (although not significantly) up to24 h past cycling.

There was no significant differences in the amount of kinematicvariation (measured as SD) between the LBP and non-LBP groups(p ¼ 0.388, t ¼ �0.890).

The LBP subjects had a slightly more posteriorly tilted saddleangle (�0.1��2.9�) compared to those without LBP (2.2��2.6�),although this difference was not statistically significant between

Fig. 4. Percentage of total lower lumbar flexion (�SD) over entire period per 12intervals of 10 min per group. LBP: low back pain; Non-LBP: no low back pain.

Fig. 2. The average pain scores (NPRS; 0e10) (�SD) during and after cycling per group.The vertical dotted black line indicates the end of the cycling task and the start of the24 h follow-up period. NPRS: Numerical Pain Rating Scale; LBP: low back pain; Non-LBP: no low back pain; h: hour.

W. Van Hoof et al. / Manual Therapy 17 (2012) 312e317 315

both groups (p ¼ 0.112, t ¼ 1.689). Three of the non-LBP and one ofthe LBP group used a medial-cutout saddle.

The mean percentage of total lower lumbar flexion over theentire 2 h of cycling was significantly increased in the LBP groupcompared to the non-LBP group, namely 74.1(�7.9)% Fl ROMcompared to 63.6(�9.8)% Fl ROM respectively (p ¼ 0.018,t ¼ �2.668).

Fig. 3 shows the amount of time spent near end-range flexion inboth groups. This analysis demonstrated that NS-CLBP (FP) subjectsspend on average more than 38.5% of their total cycling time in anend-range posture exceeding 80% of total lumbo-pelvic flexion, incontrast to only 4% for the non-LBP group.

The lower lumbar kinematics of both groups during cycling aredisplayed in Fig. 4. One-way repeated measures ANOVA revealeda significant main effect for group (p ¼ 0.035, F ¼ 5.546) with theNS-CLBP (FP) group exhibiting significantly greater lumbo-pelvicflexion which remained just significant when adding saddle angleas a covariate (p ¼ 0.05, F ¼ 4.747). The degree of flexion did notchange between groups over time (p ¼ 0.076, F ¼ 2.574). The

Fig. 3. Time (min) expressed as a % of the total 2 h cycling period spent in the availablelower lumbar flexion ROM (expressed as a % of the total lower lumbar flexion ROM).LBP: low back pain; Non-LBP: no low back pain.

interaction effect between group and posture was also not signifi-cant (p ¼ 0.592, F ¼ 0.950).

In summary, the LBP group reported increased pain over the 2 hof cycling. They exhibited a significantly increased lower lumbarflexed posture compared to the non-LBP group, and the differencein kinematics between groups did not change over time.

4. Discussion

4.1. Spinal kinematics

The results of this study clearly demonstrate that cyclists withNS-CLBP (FP) adopt significantly greater lower lumbar flexioncompared to asymptomatic cyclists during a 2-h cycling field test.Maintaining this increased flexionwas associated with a significantincrease in LBP during cycling. Although the pain only increased toa statistically significant level from baseline after >90 min ofcycling, the increase was already clinically significant (>2/10 onNPRS) after 30 min of cycling, and remained elevated to this clin-ically significant level for up to 2 h after cycling. Similar to thesefindings, Burnett et al. (2004) demonstrated an increased lowerlumbar flexion (and rotation) in NS-CLBP (FP) cyclists duringa laboratory cycling study. Studies investigating activities otherthan cycling, such as rowing (McGregor et al., 2002; Caldwell et al.,2003; Holt et al., 2003; Perich et al., 2006; Ng et al., 2008), manualhandling (O’Sullivan et al., 2006b), and work in a special school(Wong et al., 2009) have also described LBP subjects adoptinggreater lumbar flexion. For example, Ng et al. (2008) also reportedLBP subjects spending more time near end-range lumbar spineflexion (>90% Fl ROM) during prolonged ergometer rowing. Simi-larly, O’Sullivan et al. (2006b) reported that the usual sittingposture of manual workers with LBPwas significantly closer to end-range lumbar flexion.

Measuring lumbar kinematics relative to an individual’smaximum flexion range may be useful in LBP (O’Sullivan et al.,2006b), since bending moments increase near end-range whichplaces more stress on spinal structures (Dolan and Adams, 1993;Adams et al., 1994). The propensity for the cyclists with NS-CLBP(FP) to adopt and sustain a near end-range lower lumbar flexion,as clearly shown in this study, suggests a potential for increasedloading of lower lumbar soft tissues and additional pain provoca-tion. This is consistent with the concept of sustained flexion-relatedpain provocation (McKenzie, 1989; Callaghan and Dunk, 2002;O’Sullivan et al., 2006b; Dunk et al., 2009).

In addition, increased lumbar flexion could increase spinal loads(McGill, 2004; Burnett et al., 2008) and reduce trunk muscle acti-vation (Burnett et al., 2004), neither of which was measured in this

W. Van Hoof et al. / Manual Therapy 17 (2012) 312e317316

study. Additionally, while increased lower lumbar flexion may beinversely correlated with back muscle endurance (O’Sullivan et al.,2006b), in this study lower lumbar kinematics did not deteriorateover time during cycling but instead stayed relatively consistent.This suggests that an altered motor control pattern could be theprimary driver for LBP during cycling, rather than simply reducedendurance. Further research is required to clarify the contributionof fatigue in cyclists with LBP.

It has been reported that spinal creep is associated withincreased stretch, microdamage and acute inflammation in thevisco-elastic tissues (Solomonow et al., 2003a, 2003b, 2003c)potentially leading to LBP (Solomonow et al., 1999). However,regarding the relatively constant lower lumbar kinematics in thisstudy we cannot clearly support the concept of spinal creep asa causative factor for developing LBP during cycling (Burnett et al.,2004). However, it is still reasonable to hypothesise that creep mayhave occurred on a microscopic-level, not measurable with theexternal device used in this study or former studies. The authorsrecommend further investigations under the form of in-vivo labo-ratory studies investigating tissue creep.

The FRP has been reported to occur for spinal stabilizingmusclesfrom beyond mid-to end-range of spinal flexion in sitting(O’Sullivan et al., 2006a; Mork andWestgaard, 2008). Therefore themore end-range lower lumbar flexed kinematics observed amongthe NS-CLBP (FP) cyclists could possibly induce the FRP. This couldfurther increase loading of passive lumbar structures and intensifyor provoke existing back pain (O’Sullivan et al., 2006a). Furtherinvestigation by means of EMG studies during cycling is necessaryto clarify this hypothesis.

The findings of this field study suggest that a well selectedsubgroup of cyclists with NS-CLBP present with increased flexion inthe lower lumbar spine during cycling. Interestingly, the differencein kinematics between the two groups did not only occur after theonset of increased pain. Instead, differences in kinematics werepresent from baseline, and remained relatively consistentthroughout testing despite a gradual increase in the severity of theLBP experienced. This may suggest an inherent maladaptive motorcontrol impairment rather than a reflex response to pain. Further,with the onset and accumulation of LBP, the LBP subjects generatedno positional adaptation at the painful lower lumbar spine to reducethe flexion strain, strengthening the hypothesis of an inherentunderlying motor control dysfunction in the NS-CLBP (FP) subjects.The irony in this finding is that the subjects seem to assume a mal-adaptive posture thatmaximally stresses their pain sensitive tissueswithout being aware of doing this (Burnett et al., 2004; O’Sullivan,2005a; Dankaerts et al., 2006). In the literature it has been docu-mented that a chronic pain state can be maintained by furthersensitisation of habitually end-range loaded pain sensitive tissue(O’Sullivan et al., 2006b). In this context, end-range lower lumbarflexion during cycling, as observed in this study, could act as anongoing peripheral nociceptive generator and predispose them tothe development, or further provocation, of LBP.

This study also demonstrated a non-significant trend for the NS-CLBP group to have their saddle slightly more posteriorly tilted.This saddle position has a tendency to induce a posterior pelvic tilt,thereby contributing to lower lumbar flexion and perhapsexplaining difficulties controlling their lumbar spine in a morefavourable anterior pelvic tilt. Indeed, adding saddle angle asa covariate reduced the significance of the difference in lowerlumbar kinematics between both groups. This might be importantto consider during rehabilitation/prevention. Further research isnecessary to discern the relative contribution of changing lowerlumbar kinematics and/or geometric bike-related variables inreducing LBP in cyclists. It is not unlikely that a combination tar-geting both factors may help.

4.2. Limitations and recommendations for further research

Although the BodyGuard� is able to allow longitudinal real-time measurements with minimal to no task interference, somelimitations of the current study have to be considered. Due to skindistraction, the use of any measurement system utilizing surfacelandmarks to quantify spinal posture can result in an over-estimation of the true lumbo-pelvic posture (Burnett et al., 2004,2008). However, since the same measurement system was used inboth groups this aspect does not explain the significantdifferences observed between the groups. Further, the Body-Guard� measurement system was only capable of measuringsagittal plane movements. Additional movements of the lumbarspine in other planes (transversal/frontal), especially whencombined with a sustained end-range flexed position, may bea greater risk factor for the development of LBP (Adams and Dolan,1995; Nachemson, 1999; Burnett et al., 2008). Therefore the abilityof the BodyGuard� to identify biomechanical risk factors duringcycling was limited to sagittal kinematics. Although the findings ofthe present study are significant the small sample size, and thespecific homogeneous subgroup targeted, limits the ability togeneralize the conclusions for all cycling-related LBP. A furtherconsequence of the sub-classification is the exclusion of subjectswith an active extension pattern. In this subgroup it has beensuggested that over-activation of the paravertebral muscles couldbe a contributing factor to LBP (Mellion, 1994; Usabiaga et al., 1997;Indahl, 1999).

The cross-sectional nature of the study design limits the abilityto make definite conclusions regarding a causal relationshipbetween the flexion pattern and LBP. A prospective study witha large sample size is necessary to further clarify this potentialcausal relationship.

Since the findings of this study suggest that LBP during cycling isrelated to maladaptive lower lumbar kinematics, trying to regaincontrol over the lower lumbar region during cycling could berelevant in the rehabilitation/prevention of LBP in this subgroup. Acognitive functional training intervention including biofeedback tomonitor the lower lumbar kinematics, to facilitate a less end-rangeflexed cycling posture has been recently explored by our researchgroup (Van Hoof et al., 2011). The results revealed that an inter-vention targeting this maladaptive control at the symptomaticlower lumbar region resulted in a significant decrease of the nearend-range lower lumbar flexion and a substantial reduction of LBPduring cycling. Additional studies are necessary to further test thisinterventional approach.

5. Conclusion

The findings of this first field study evaluating lower lumbarkinematics during cycling suggest that a well selected subgroup ofcyclists with NS-CLBP (FP) adopted and sustained increased lowerlumbar flexion during cycling. This appears to reflect an inherentmaladaptive motor control pattern at the lower lumbar spineduring cycling. The difference in posture between groups did notchange over time. This posture is maintained and associated witha significant increase of LBP.

Acknowledgements

We would like to thank Sels Instruments nv, Belgium (http://www.sels-instruments.be/) for the provision of the BodyGuard�measuring systems. One of the authors (O’Sullivan K.) is currentlysupported by a Health Research Board of Ireland researchfellowship.

W. Van Hoof et al. / Manual Therapy 17 (2012) 312e317 317

References

ACSM. ACSM guidelines for exercise testing and prescription. 7th ed.; 2006.Adams M, Dolan P. Recent advances in lumbar spinal mechanics and their clinical

significance. Clinical Biomechanics 1995;10(1):3e19.Adams M, McNally D, Chinn H, Dolan P. Posture and the compressive strength of the

lumbar spine. Clinical Biomechanics 1994;9(1):5e14.Burnett A, Cornelius M, Dankaerts W, O’Sullivan PB. Spinal kinematics and trunk

muscle activity in cyclists: a comparison between healthy controls and non-specific chronic low back pain subjects-a pilot investigation. Manual Therapy2004;9(4):211e9.

Burnett A, O’Sullivan P, Ankarberg L, Gooding M, Nelis R, Offermann F, et al. Lowerlumbar spine axial rotation is reduced in end-range sagittal postures whencompared to a neutral spine posture. Manual Therapy 2008;13(4):300e6.

Caldwell JS, McNair PJ, Williams M. The effects of repetitive motion on lumbarflexion and erector spinae muscle activity in rowers. Clinical Biomechanics2003;18(8):704e11.

Callaghan J, Dunk N. Examination of the flexion relaxation phenomenon in erectorspinae muscles during short duration slumped sitting. Clinical Biomechanics2002;17(5):353e60.

Callaghan M, Jarvis C. Evaluation of elite British cyclists: the role of the squadmedical. British Journal of Sports Medicine 1996;30(4):349e53.

Callaghan J, McGill S. Low back joint loading and kinematics during standing andunsupported sitting. Ergonomics 2001;44:280e94.

Childs J, Piva S, Fritz J. Responsiveness of the numeric pain rating scale in patientswith low back pain. Spine 2005;30(11):1331e4.

Clarsen B, Krosshaug T, Bahr R. Overuse injuries in professional road cyclists.American Journal of Sports Medicine 2010;38(12):2494e501.

Colloca CJ, Hinrichs RN. The biomechanical and clinical significance of the lumbarerector spinae flexion-relaxation phenomenon: a review of literature. Journal ofManipulative and Physiological Therapeutics 2005;28(8):623e31.

Dankaerts W, O’Sullivan P, Burnett A, Straker L. Differences in sitting postures areassociated with nonspecific chronic low back pain disorders when patients aresubclassified. Spine 2006;31(6):698e704.

Dillingham T. Evaluation and management of low back pain: an overview. State ofthe Art Reviews 1995;9:559e74.

Dolan P, Adams M. Influence of lumbar and hip mobility on the bending stressesacting on the lumbar spine. Clinical Biomechanics 1993;8(4):185e92.

Dolan P, Mannion AF, Adams MA. Passive tissues help the back muscles to generateextensor moments during lifting. Journal of Biomechanics 1994;27(8):1077e85.

Drake J, Callaghan J. Do flexion/extension postures affect the in vivo passive lumbarspine response to applied axial twist moments? Clinical Biomechanics 2008;23(5):510e9.

Drake J, Aultman C, McGill S, Callaghan J. The influence of static axial torque incombined loading on intervertebral joint failure mechanics using a porcinemodel. Clinical Biomechanics 2005;20(10):1038e45.

Dunk N, Kedgley A, Jenkyn T, Callaghan J. Evidence of a pelvis-driven flexionpattern: are the joints of the lower lumbar spine fully flexed in seated postures?Clinical Biomechanics 2009;24:164e8.

Gupta A. Analyses of myo-electrical silence of erectors spinae. Journal of Biome-chanics 2001;34(4):491e6.

Holt P, Bull A, Cashman P, McGregor A. Kinematics of spinal motion during pro-longed rowing. International Journal of Sports Medicine 2003;24:597e602.

Indahl A. Low back pain: a functional disturbance. PhD. Thesis, Centre for Ortho-paedics, Norway: University of Oslo; 1999.

Kuslich S,UlstromC,Michael C. The tissueorigin of lowbackpainand sciatica: a reportof pain response to tissue stimulation during operations on the lumbar spineusing local anesthesia. Orthopedic Clinics of North America 1991;22:181e7.

Little JS, Khalsa PS. Human lumbar spine creep during cyclic and static flexion:creep rate, biomechanics, and facet joint capsule strain. Annals of BiomedicalEngineering 2005;33(3):391e401.

Marsden M, Schwellnus M. Lower back pain in cyclists: a review of epidemiology,pathomechanics and risk factors. International Sportmed Journal 2010;11(1):216e25.

McGill S. Linking latest knowledge of injury mechanisms and spine function to theprevention of low back disorders. Journal of Electromyography and Kinesiology2004;14(1):43e7.

McGill S, Brown S. Creep response of the lumbar spine to prolonged full flexion.Clinical Biomechanics 1992;7:43e6.

McGill S, Kippers V. Transfer of loads between lumbar tissues during the flexion-relaxation phenomenon. Spine 1994;19(19):2190e6.

McGregor A, Anderton L, Gedroyc W. The assessment of intersegmental motion andpelvic tilt in elite oarsmen. Medicine and Science in Sports and Exercise 2002;34:1143e9.

McKenzie R. The lumbar spine: mechanical diagnosis and therapy. Waikanae: NewZealand Spinal Publications; 1989.

Mellion MB. Common cycling injuries. Management and prevention. Sports Medi-cine 1991;11(1):52e70.

Mellion M. Neck and back pain in bicycling. Clinics in Sports Medicine 1994;13(1):137e64.

Mitchell T, O’Sullivan PB, Burnett AF, Straker L, Smith A. Regional differences inlumbar spinal posture and the influence of low back pain. BMC MusculoskeletDisord 2008;9:152.

Mork P, Westgaard R. Back posture and low back muscle activity in femalecomputer workers: a field study. Clinical Biomechanics 2009;24(2):169e75.

Nachemson A. Back pain: delimiting the problem in the next millennium. Inter-national Journal of Law and Psychiatry 1999;22(5e6):473e90.

Ng L, Burnett A, O’Sullivan P. Spino-pelvic kinematics and trunk muscle activation inprolonged ergometer rowing: mechanical etiology of non-specific low backpain in adolescent rowers [abstract]. In: Kwon Y, Shim J, Shim J, Shin I, editors.Seoul National University; 2008. p. 270e3.

Olson MW, Li L, SolomonowM. Flexion-relaxation response to cyclic lumbar flexion.Clinical Biomechanics 2004;19(8):769e76.

Ostelo RW, Deyo RA, Stratford P, Waddell G, Croft P, Von Korff M, et al. Interpretingchange scores for pain and functional status in low back pain: towards interna-tional consensus regarding minimal important change. Spine 2008;33(1):90e4.

O’Sullivan K, Galeotti L, Dankaerts W, O’Sullivan L, O’Sullivan P. The between-dayand inter-rater reliability of a novel wireless system to analyse lumbar spineposture. Ergonomics 2011a;54(1):82e90.

O’Sullivan K, O’Sullivan L, Campbell A, O’Sullivan P, Dankaerts W. Towards moni-toring lumbo-pelvic posture in real-life situations: concurrent validity ofa novel posture monitor and a traditional laboratory-based motion analysissystem. Manual Therapy 2012;17(1):77e83.

O’Sullivan P. Lumbar segmental ‘instability’: clinical presentation and specificstabilizing exercise management. Manual Therapy 2000;5(1):2e12.

O’Sullivan P. Diagnosis and classification of chronic low back pain disorders: mal-adaptive movement and motor control impairments as underlying mechanism.Manual Therapy 2005a;10(4):242e55.

O’Sullivan P. ‘Clinical instability’ of the lumbar spine: its pathological basis,diagnosis and conservative management. In: Jull G, Boyling J, editors.Grieve’s modern manual therapy. 3rd ed. Amsterdam: Elsevier; 2005b. p.311e22.

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. Spine2006a;31(17):2009e16.

O’Sullivan P, Mitchell T, Bulich P, Waller R, Holte J. The relationship between postureand back muscle endurance in industrial workers with flexion-related low backpain. Manual Therapy 2006b;11(4):264e71.

O’Sullivan PB, Burnett A, Floyd AN, Gadsdon K, Logiudice J, Miller D, et al. Lumbarrepositioning deficit in a specific low back pain population. Spine (Phila Pa1976) 2003;28(10):1074e9.

Perich D, Burnett A, O’Sullivan P. Low back pain and factors associated with it:examination of adolescent female rowers. In: XXIVth symposium of the inter-national society of biomechanics in sports. Salzburg: University of Salzburg;2006. p. 355e8.

Pope MH. Risk indicators in low back pain. Annals of Medicine 1989;21(5):387e92.Rivara FP, Thompson DC, Thompson RS. Epidemiology of bicycle injuries and risk

factors for serious injury. Injury Prevention 1997;3(2):110e4.Russell P, Weld A, Pearcy MJ, Hogg R, Unsworth A. Variation in lumbar spine

mobility measured over a 24-hour period. British Journal of Rheumatology1992;31(5):329e32.

Salai M, Brosh T, Blankstein A, Oran A, Chechik A. Effect of changing the saddleangle on the incidence of low back pain in recreational bicyclists. British Journalof Sports Medicine 1999;33(6):398e400.

Shirazi-Adl A, Drouin G. Load-bearing role of facets in a lumbar segment undersagittal plane loadings. Journal of Biomechanics 1987;20:601e13.

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):2426e34.

Solomonow M, Baratta RV, Banks A, Freudenberger C, Zhou BH. Flexion-relaxationresponse to static lumbar flexion in males and females. Clinical Biomechanics2003a;18(4):273e9.

Solomonow M, Baratta RV, Zhou BH, Burger E, Zieske A, Gedalia A. Musculardysfunction elicited by creep of lumbar viscoelastic tissue. Journal of Electro-myography and Kinesiology 2003b;13(4):381e96.

Solomonow M, Hatipkarasulu S, Zhou BH, Baratta RV, Aghazadeh F. Biomechanicsand electromyography of a common idiopathic low back disorder. Spine (PhilaPa 1976) 2003c;28(12):1235e48.

Stone M, Broughton J. Getting off your bike: cycling accidents in Great Britain in1990e1999. Accident Analysis and Prevention 2003;35(4):549e56.

Thompson MJ, Rivara FP. Bicycle-related injuries. American Family Physician 2001;63(10):2007e14.

Tin Tin S, Woodward A, Ameratunga S. Injuries to pedal cyclists on New Zealandroads, 1988e2007. BMC Public Health 2010;10:655.

Usabiaga J, Crespo R, Iza I, Aramendi J, Terrados N, Poza J. Adaptation of the lumbarspine to different positions in bicycle racing. Spine 1997;22(17):1965e9.

Van Hoof W, Volkaerts K, O’Sullivan K, Verschueren S, Dankaerts W. Cognitivefunctional therapy intervention including biofeedback for LBP during cycling e

a single case study. Sport & Geneeskunde 2011;44(4):20e6.Waddell G. The back pain revolution. New York: Churchill Livingstone; 1998.Wilber CA, Holland GJ, Madison RE, Loy SF. An epidemiological analysis of overuse

injuries among recreational cyclists. International Journal of Sports Medicine1995;16(3):201e6.

Williamson A, Hoggart B. Pain: a review of three commonly used pain rating scales.Journal of Clinical Nursing 2005;14(7):798e804.

Wong K, Lee R, Yeung S. The association between back pain and trunk posture ofworkers in a special school. BMC Musculoskeletal Disorders 2009;10(43).