Whiplash associated disorders (WAD) have been studied mainly in comparison to asymptomatic subjects and it remains controversial whether WAD represent a diagnostic entity different from chronic neck pain. WAD patients are separated from chronic neck pain merely on the history of trauma [1]. The trauma designation is however not reflected in structural injuries of the cervical spine or the cerebrum [2, 3]. Consequently, somatosensory and motor deficits have gained research attention. There are consistent findings of hypersensitivity and widespread pain in WAD compared to healthy subjects [4–7], and when comparing WAD patients to chronic neck pain patients [5, 6, 8]. While this points to centrally mediated somatosensory alterations in WAD, it is not clear whether motor areas are also affected. If such changes exist, altered motor control as well as kinaesthetic change should be present in WAD, which would also provide important information for clinicians.

There is conflicting evidence for presence of kinaesthetic change in WAD patients, as measured by head repositioning or joint position error (JPE) in comparison to healthy controls [9–14], and the reported repositioning errors are generally small, i.e. 2–5°. Non-traumatic chronic neck pain patients do not seem to have reduced kinaesthetic sense compared to healthy subjects [12, 15, 16]. The smooth pursuit eye movement test, believed to reflect cervical afferent dysfunction, has also shown inconsistent results [17–19].

There is consistent documentation for reduced standing balance and increased sway in WAD patients compared to healthy subjects [20–25]. Small differences, and only for difficult balance tasks, have been found between WAD and non-traumatic neck pain patients [23]. Less attention has been given to local motor control in the neck. Proprioceptive information from the neck greatly influence head and trunk position sense [26] and motor control [27] and mechanisms controlling head motion may be different from those that control the trunk [28]. WAD patients present with reduced cervical range of motion (ROM) relative to asymptomatic individuals [29, 30], but they also show increased variability in ROM [31]. These tests may however be affected by motivational factors. For trajectory head movement patterns WAD patients showed greater variation between days than asymptomatic controls [32]. More jerky movement patterns were found in WAD and chronic neck pain patients compared to healthy subjects, but there were no differences between the two pain groups [31]. Although the latter two studies address relevant motor control issues, they were both small and neither could provide evidence for a difference between WAD and chronic non-traumatic neck pain. In fact, of the above studies on postural control only two included non-traumatic chronic neck pain patients in addition to WAD and healthy subjects [23, 31]. In other words, the other studies were not designed to reveal whether postural control deficits are related to the trauma or just long-standing pain. If WAD patients are distinguished by altered motor control strategies, their movement patterns should be different from that seen in both non-traumatic neck pain patients and in healthy controls.

Pain has been shown to induce decreased variability in postural strategies in low back pain [33, 34], and stiffening of spinal movements [33]. Cervical movements in the associated planes relative to the primary movement plane have been named conjunct motion, and might reflect protective postural control strategies. Conjunct motion was explored in this study to get an impression of motor function or "freedom of movement" in the neck. Conjunct motion at the end of primary range was investigated in a previous study which found small differences between WAD and healthy controls, but the study did not include non-traumatic neck pain patients [30]. In addition, movement deviations would more likely be found during execution of the primary motion which from a motor control perspective is more complicated.

In order to explore motor function characteristics of WAD patients in particular, control groups of non-traumatic neck patients and asymptomatic subjects were included in this study. The main purpose was to investigate motor control deficits in WAD compared to the two control groups in relation to conjunct motion, JPE, ROM and ROM variability.

A case-control study with a total of 173 participants was conducted in the period Jan 2004 – Oct 2006. The study group consisted of persons with persistent WAD for more than 6 months (WAD group). Two control groups were included; a group of chronic non-traumatic neck pain patients (chronic neck pain group) and a group of asymptomatic volunteers (asymptomatic control group). All subjects gave written informed consent and the study was conducted in accordance with the Helsinki Declaration and approved by the Regional Ethics Committee.

Data management and statistical analysis

Maximal cervical ROM was calculated for the entire range of each primary movement plane by the mean and standard deviation (SD) of three trials. The SD of three repetitions of maximal ROM was used to express ROM-variability for each cardinal plane, and the mean value (SD_mean) of the three planes was used in the analyses. Mean ROM in each primary plane was summed to express total cervical ROM. Conjunct motion was calculated as the maximum deviation range (°) in each associated conjunct motion plane during execution of the primary movement, i.e., the maximal ROM trials (Fig 1). The sum of conjunct motion in the two associated planes was taken to express total conjunct motion for each primary motion. JPE was registered in degrees as the absolute maximum repositioning error regardless of plane, i.e. the largest error value of the three planes was selected as the JPE. The mean JPE value from four test repetitions (two left, two right) was used for analysis.

Differences between groups in conjunct motion and maximal cervical ROM were analyzed by general linear modelling (GLM) and post-hoc by the Tukey-Kramer multiple comparison test. Conjunct motion was analyzed both with and without introduction of covariates and adjusted for age, gender, neck pain intensity on day of testing and range of motion in the primary plane. For some variables residual plots showed an increasing variance with increasing predicted values. Transformation of these variables using natural logarithms (Ln) gave acceptable residual plots and similar statistical results. Thus, the non-transformed values are reported. The Kruskall-Wallis test was used for evaluating group differences in JPE and ROM-variability. Differences in pain levels and ROM between subjects with and without ongoing compensation claims within the WAD group were evaluated by the two sample t-test. Group differences were considered significant at the p < 0.05 level. Analyses were performed using SPSS 14.0 and NCSS 2007 (Utah, USA).

The main finding in this study was that patients with chronic neck pain showed altered motor control in the cervical spine. We found no indications for differential effects between traumatic and non-traumatic chronic neck patients in terms of conjunct motion. No group differences were found in ROM-variability or joint position error. Maximal cervical range of motion in the primary planes was however significantly reduced among the WAD patients compared to both control groups.

Conjunct motion was included in this study to get an impression of motor function or motion stiffness during execution of primary plane movements, believed to reflect central nervous motor control strategies [40, 41]. Conjunct motion in the cervical spine has been studied in asymptomatic subjects [42–45], but only two studies have included WAD and healthy subjects [30, 46]. Marginal reductions in conjunct motion in the WAD group were observed in both studies. A limitation in the study by Dall'Alba et al was that conjunct motion was recorded only at end of primary motion. In our study, conjunct motion was recorded during primary movements and is to our knowledge the first to compare WAD patients to both non-traumatic neck pain and healthy subjects. The study groups were separated on most conjunct motion measures in the unadjusted analyses (Tab. 3). It is possible that conjunct motion mainly happens towards end of physiological range and is therefore reduced due to reduced primary ROM. However, differences between the pain groups and the healthy controls remained after controlling for ROM in the analysis. Neither could the differences between the pain groups and healthy subjects be explained by current neck pain intensity. This indicates that stiffer or more guarded movement patterns are a response to long-standing pain. This response was irrespective of a trauma history. For all groups the largest conjunct motion ranges were found during primary cervical rotation (horizontal plane) and, interestingly, stiffer movement patterns in the pain groups were also most evident during cervical rotation. This complies with a general understanding that cervical rotation demands more complex physiological coupled motions due to the anatomic structures, such as the orientation of cervical zygapophysial and uncovertebral joints [47]. Changes in motor control due to chronic pain may thus be observable only in more complex movement patterns.

Whether reduced conjunct motion reflects altered motor control strategies needs further investigation. If confirmed, it remains to reveal if such changes are peripherally driven or merely confined to central processing mechanisms. Perceived position of the head is greatly influenced by proprioceptive signals from the neck [26]. Hypothetically, cervical microtrauma involving mechanoreceptors may cause corrupted afferent information and sensorimotor mismatch [48]. The head relocation test, equivalent to the JPE test, was introduced to detect alterations in cervical proprioception [49]. We thus expected that disturbances in proprioception would be reflected in the JPE test, but we found no difference between the groups. Consequently, proprioceptive dysfunction gives a poor explanation for the observed alterations in conjunct motion in the pain groups. Since current pain intensity could neither account for reduced conjunct motion, it may be argued that the observed changes in motor control reflect central nervous motor processing adaptations due to long-standing pain rather than disturbances in concurrent afferent inputs. Musculoskeletal pain has been reported to induce cortical reorganization in both somatosensory and motor areas [40]. Maladaptive cortical sensimotor integration with chronic neck pain has been suggested [50] and neuromagnetic imaging has shown altered responses in the motor cortex due to chronic pain [51]. Although local trauma to mechanoreceptors in the neck cannot be excluded, our findings favour centrally driven alterations in motor control due to long-standing pain.

Kinaesthetic sense deficits in patients with cervical pain have in previous studies been demonstrated as reduced accuracy, or increased cervical joint position error, when relocating the head to the initial position after a maximal or sub-maximal cervical movement [9, 11, 14, 49, 52, 53]. The evidence is inconclusive, however, as other studies have found only small or no differences in JPE among neck pain patients [13, 15, 22, 54] or greater JPE only among those with the highest levels of pain and disability [10]. Various procedures have been used to record JPE of the cervical spine and the variation in set-up may account for the conflicting findings. In our study, group differences in JPE were very small (<1°) and statistically non-significant.

As expected, and in agreement with other studies [29, 30, 55, 56], primary ROM differentiated between WAD patients and healthy controls also in this study. The WAD patients also had significantly less cervical ROM compared to the chronic neck pain group, even after adjusting for current pain intensity, gender and age. Although ROM differentiated between the two groups of neck pain patients, the diagnostic value of ROM-testing can be questioned. Patients could intentionally simulate less motion for beneficial reasons in cases of unclarified insurance compensation. However, the subgroup of WAD-patients with ongoing compensation claims in this study performed similarly to the rest of the WAD group, which leaves no suspicion of simulation of abnormal motion.

ROM-variability from repeated recordings of maximal cervical range has been introduced as a measure that could distinguish between maximal and sub-maximal effort in the ROM testing situation [57, 58]. Hypothetically, both simulation of ROM and pain-related fear of movement may explain a sub-maximal effort. Sub-maximal effort may presumably increase variability between repeated trials of maximal ROM. This is due to an assumption that physiological end-of-range in the cervical spine is not reached with a sub-maximal effort, resulting in greater variability in endpoints of the repeated movements. Two studies have reported larger ROM-variability among WAD patients compared to healthy controls [31, 56]. In our study, no increase in ROM-variability among WAD patients or chronic neck pain patients was found. However, the use of SD_mean as measure of variability implies that the "size" of ROM did not influence the variability measure. An alternative that would normalize the variation to the mean ROM is the coefficient of variation (CV). However, the small variability of the nominator (SD_mean) relative to the large variability in the denominator (mean ROM) would make the CV heavily depend on the latter and thus add little information beyond that of mean ROM. A correlation analysis was performed between mean ROM and SD_mean to investigate whether SD_mean was dependent on the "size" of mean ROM and thus needed to be normalized. No correlation was found in any of the study groups, and SD_mean was considered the best variability measure. When direction-specific maximal ROM is repeatedly recorded from neutral position, variability in the neutral position starting point could hypothetically add to the variability in ROM. This factor was eliminated by recording total plane ROM in our study. Since increased ROM-variability was not found between any of the groups, sub-maximal effort due to simulation, pain or fear of pain was not confirmed in this study.

Compared to asymptomatic controls both whiplash and chronic neck pain patients showed reduced conjunct motion, particularly during primary cervical rotation. We found no indications for a difference between traumatic and non-traumatic neck pain patients. Reduced conjunct motion was not explained by variation in current neck pain or range of motion in the primary plane. Less flexible movement patterns in the two pain groups compared to the asymptomatic group may reflect motor control adaptations, possibly due to long lasting pain. JPE and maximal ROM-variability did not separate the study groups. The whiplash patients showed reduced maximal cervical ROM compared to both control groups, and the chronic neck pain group had less ROM compared to the asymptomatic controls. No significant differences in ROM and pain intensity were found between WAD-patients with ongoing insurance compensation claims and the rest of the WAD-group.