Sensory hypersensitivity occurs soon after whiplash injury and is associated with poor recovery

Pain 104 (2003) 509 517 www.elsevier.com/locate/pain Sensory hypersensitivity occurs soon after whiplash injury and is associated with poor recovery Michele Sterling a, *, Gwendolen Jull a, Bill Vicenzino a, Justin Kenardy b a The Whiplash Research Unit, Department of Physiotherapy, The University of Queensland, Brisbane 4072, Australia b Department of Psychology, The University of Queensland, Brisbane 4072, Australia Received 25 November 2002; received in revised form 11 February 2003; accepted 19 February 2003 Abstract Hypersensitivity to a variety of sensory stimuli is a feature of persistent whiplash associated disorders (WAD). However, little is known about sensory disturbances from the time of injury until transition to either recovery or symptom persistence. Quantitative sensory testing (pressure and thermal pain thresholds, the brachial plexus provocation test), the sympathetic vasoconstrictor reflex and psychological distress (GHQ-28) were prospectively measured in 76 whiplash subjects within 1 month of injury and then 2, 3 and 6 months post-injury. Subjects were classified at 6 months post-injury using scores on the Neck Disability Index: recovered (, 8), mild pain and disability (10 28) or moderate/severe pain and disability (. 30). Sensory and sympathetic nervous system tests were also measured in 20 control subjects. All whiplash groups demonstrated local mechanical hyperalgesia in the cervical spine at 1 month post-injury. This hyperalgesia persisted in those with moderate/severe symptoms at 6 months but resolved by 2 months in those who had recovered or reported persistent mild symptoms. Only those with persistent moderate/severe symptoms at 6 months demonstrated generalised hypersensitivity to all sensory tests. These changes occurred within 1 month of injury and remained unchanged throughout the study period. Whilst no significant group differences were evident for the sympathetic vasoconstrictor response, the moderate/severe group showed a tendency for diminished sympathetic reactivity. GHQ-28 scores of the moderate/severe group were higher than those of the other two groups. The differences in GHQ-28 did not impact on any of the sensory measures. These findings suggest that those with persistent moderate/severe symptoms at 6 months display, soon after injury, generalised hypersensitivity suggestive of changes in central pain processing mechanisms. This phenomenon did not occur in those who recover or those with persistent mild symptoms. q 2003 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. Keywords: Whiplash; Quantitative sensory testing; Hypersensitivity 1. Introduction Chronic whiplash associated disorders (WAD) have been shown to be associated with sensory and motor hypersensitivity that likely reflects underlying changes in neurobiological processing of pain mechanisms. Sensory changes include hypersensitivity to a variety of stimuli including mechanical and electrocutaneous stimulation and induced muscle pain (Koelbaek-Johansen et al., 1999; Curatolo et al., 2001; Moog et al., 2002; Sterling et al., 2002a). In addition, there is evidence of increased sensitivity of the flexor withdrawal response in patients with chronic WAD (Curatolo et al., 2002). This is supported to some extent by findings of * Corresponding author. Tel.: þ61-7-3365-4568; fax: þ61-7-3365-2775. E-mail address: m.sterling@shrs.uq.edu.au (M. Sterling). hypersensitive motor responses to a clinical test of flexor withdrawal (Sterling et al., 2002b). In contrast to the expanding knowledge of mechanisms involved in chronic WAD, the acute phase of injury and the processes involved in the transition to either recovery or symptom persistence have been largely unexplored. A greater knowledge of these phases of the condition may help with early identification and management of those at risk of developing persistent pain estimated at up to 40% of those experiencing whiplash injury (Barnsley et al., 1994). Kasch et al. (2001a,b) demonstrated mechanical hyperalgesia locally within the cervical spine of acutely injured WAD subjects but failed to identify any such effects at remote uninjured sites (as occurs in subjects with chronic WAD). However, these authors did not explore the possibility that such changes may occur only in some patients, a likely 0304-3959/03/$20.00 q 2003 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. doi:10.1016/s0304-3959(03)00078-2

510 M. Sterling et al. / Pain 104 (2003) 509 517 important factor supported by Yerner et al. (2001) who showed that altered sensory responses in cutaneous areas supplied by the trigeminal nerve occurred only in a subgroup of WAD patients. Furthermore, higher levels of pain and disability in acute WAD have been accepted as a sign of poor outcome (Cote et al., 2001; Radanov et al., 1995), suggesting the importance of differentiation between those with higher pain and disability levels from those with lesser symptoms. Involvement of sympathetic nervous system (SNS) activity as contributing to symptoms of WAD is yet to be pursued, although suggestions to its presence have been made (Adeboye et al., 2000; Munglani, 2000). Using the sympathetic vasoconstrictor reflex (SVR), changes in SNS activity have been shown to occur in chronic musculoskeletal pain syndromes such as frozen shoulder and lateral epicondylalgia (Mani et al., 1989; Smith et al., 1994) and complex regional pain syndrome type 1 (Schurmann et al., 1999). As some symptoms of WAD such as vasomotor disturbances, burning pain and cold intolerance mimic those of complex regional pain syndrome, investigation of SNS activity may provide further information of the underlying processes of this condition. Those with chronic WAD show evidence of psychological distress (Peebles et al., 2001; Radanov et al., 1995), which is probably not surprising in view of persisting symptoms. Psychological factors such as anxiety and fear have been shown to affect measures of both pain threshold and pain tolerance (Rhudy and Meagher, 2000). For this reason, it is important that psychological factors are taken into account when measuring pain responses. Our study addressed the lack of information of changes in sensory and SNS function soon after injury and the time course of such changes to 6 months post-injury (a time frame after which symptoms change little; Mayou and Radanov, 1996). The aims of the study were threefold: to investigate the differences in sensory and SNS function between those who recover and those who report persistent symptoms based on their status at 6 months; to investigate the prospective longitudinal development of such changes following whiplash injury and to determine the effect of psychological distress on sensory measures. 2. Methods 2.1. Study design A prospective longitudinal design was used to study persons who sustained a whiplash injury from within 1 month of injury to 6 months post-injury. They were assessed on four occasions within 1 month of injury, 2, 3 and 6 months post-injury. An asymptomatic control group was assessed three times, each 1 month apart. 2.2. Subjects Eighty volunteers (24 males, 56 females, mean age 36.27 ^ 12.69 years) reporting neck pain as a result of a motor vehicle crash and 20 asymptomatic volunteers (8 males, 12 females, mean age 40.1 ^ 13.6 years) participated in the study. The whiplash subjects were recruited via hospital accident and emergency departments, primary care practices (medical and physiotherapy) and from advertisement within radio and print media. They were eligible if they met the Quebec Task Force Classification of WAD II or III (Spitzer et al., 1995). Subjects were excluded if they were WAD IV, experienced concussion, loss of consciousness or head injury as a result of the accident and if they reported a previous history of whiplash, neck pain or headaches that required treatment. The asymptomatic control group was recruited from the general community from print media advertisement and were included provided they had never experienced any prior pain or trauma to the cervical spine, head or upper quadrant. Ethical clearance for this study was granted by the Medical Research Ethics Committee of The University of Queensland. 2.3. Pressure pain thresholds (PPTs) PPTs were measured using a pressure algometer with a probe size of 1 cm 2 and application rate of 40 kpa/s (Somedic AB, Farsta, Sweden). PPTs were measured at two bilateral cervical spine sites (over the articular pillars of C2/3 and C5/6), at three bilateral upper limb sites (over the three main peripheral nerve trunks) and at a bilateral remote site (tibialis anterior). These sites have been previously used in investigation of chronic WAD (Sterling et al., 2002b). The subjects were requested to push a button when the sensation changed from one of pressure alone to one of pressure and pain (Brennum et al., 1989). Triplicate recordings were taken at each site and the mean values used for analysis. 2.4. Thermal (hot, cold) pain thresholds Thermal pain thresholds were measured bilaterally over the cervical spine using the Thermotest system (Somedic AB, Farsta, Sweden). The thermode was placed over the skin of the mid-cervical region and preset to 308C with the rate of temperature change being 18C/s. To identify cold pain thresholds (CPTs) and heat pain thresholds (HPTs), subjects were asked to push a patient-controlled switch when the cold or warm sensation first became painful (Hurtig et al., 2001). Triplicate recordings were taken at each site and the mean values used for analysis. 2.5. Brachial plexus provocation test (BPPT) The BPPT was performed as described previously and in

M. Sterling et al. / Pain 104 (2003) 509 517 511 the following sequence: gentle shoulder girdle depression, glenohumeral abduction and external rotation in the coronal plane, wrist and finger extension and elbow extension (Elvey, 1979; Selvaratnam et al., 1994). The range of elbow extension was measured at the subjects pain threshold using a standard goniometer aligned along the mid-humeral shaft, medial epicondyle and ulnar styloid (Balster and Jull, 1997; Clarkson and Gilewich, 1989). If the subject did not experience pain, the test was continued until the end of available range. At the completion of this test, the subjects were asked to record their pain on a 10 cm visual analogue scale (VAS). 2.6. Sympathetic vasoconstrictor reflex The SVR was used as an indication of SNS activity (Mani et al., 1989; Schurmann et al., 1999). Using laser Doppler flowmetry (flolab Monitor, Moor Instruments, Devon, UK), the skin blood flow in the fingertips of both hands was measured. Data were sampled at 20 Hz. A provocation manoeuvre (inspiratory gasp), which is known to cause a short sympathetic reaction and cutaneous vasoconstriction, was performed (Schurmann et al., 1999). A program using Labview software was written which calculated two quotients that represented the change in blood flow following the inspiratory gasp. These were taken after Schurmann et al. (1999) and included the SRF parameter (sympathetic reflex) that represents the relative drop in the curve after provocation and the quotient of integrals (QI) that also takes into account the duration of perfusion decrease. A high QI and low SRF are indicative of an impaired vasoconstrictor response. 2.7. Questionnaires Self-reported pain and disability was measured in all whiplash subjects using the Neck Disability Index (NDI) (Vernon and Mior, 1991). They also completed the GHQ-28 (Goldberg, 1978) as an indicator of the general psychological well being. Subjects pain intensity was measured using a 10 cm VAS scale. 2.8. Procedure The following measures were undertaken at each of the four time points. The whiplash subjects first completed the NDI and GHQ-28 questionnaires. VAS measures of resting pain were recorded. Testing of both whiplash and asymptomatic subjects was performed in the following sequence: SVR, BPPT, PPTs, HPTs and CPTs. The same examiner performed all tests. This examiner remained blind to the subjects responses on the NDI and GHQ-28 questionnaires. For all tests, no verbal cues/feedback were given to the subjects about their performance. After completion of the questionnaires, the subjects lay supine and the laser Doppler blood flow sensors were attached to the tips of the middle fingers using double-sided adhesive discs. Subjects rested their hands on their abdomen and an electric heating pad was placed over the hands in order to obtain a uniform increase in blood flow of the fingertips. Subjects rested in this position for 10 15 min. After this time and as soon as a stable blood flow baseline was obtained for at least 30 s, the provocation manoeuvre of inspiratory gasp was performed. Subjects were requested to inspire as deeply as possible and then to expire with a deep sigh. The moment of full inspiration was marked with an electronic footswitch. Recording of blood flow continued for another 30 s. The SVR testing was performed in a temperature-controlled laboratory. The temperature was set at 208C, lights were dimmed and ambient noise was kept low. Further testing was completed in a standard laboratory. The BPPT was then performed, followed by PPTs measures in the following order: tibialis anterior, median, radial and ulnar nerves, C5/6 and C2/3. After this, thermal pain thresholds were measured over the cervical spine; CPTs followed by HPTs. For BPPT, PPTs and thermal pain thresholds, the left side was tested first followed by the right side. 2.9. Statistical analysis The whiplash subjects were classified into one of three groups based on results of the NDI at 6 months post-injury. The groups were recovered (, 8 NDI), mild pain and disability (10 28 NDI) and moderate/severe pain and disability (. 30 NDI) (Vernon, 1996). This grouping was validated by a cluster analysis (K-means algorithm), which showed no significant difference between the analytical clustering and the NDI groups as proposed by Vernon (1996). A repeated measures mixed model analysis of variance (ANOVA) with a between subjects factor of Group (four levels: asymptomatic, recovered, mild, moderate/severe) and a within subjects factor of Time (four levels:,1 month, 2, 3 and 6 months post-injury) was performed. Age and gender were used as covariates in this analysis. Differences between the groups were analysed with a priori contrasts. A repeated measures mixed model ANOVA with a timechanging covariate of GHQ-28 total scores was used to assess the effect of psychological distress on the sensory measures of the whiplash groups. Significance was set at P, 0:01. 3. Results 3.1. Subject classification at 6 months post-injury Of the 80 subjects who entered the study, four withdrew during the study period, all after the initial assessment point. The reasons given for withdrawal included relocation to

512 M. Sterling et al. / Pain 104 (2003) 509 517 Table 1 The age, gender and classification of subject groups at 6 months according to the NDI scores (Vernon, 1996) Group Number Age (years) (mean ^ SD) Gender (% female) NDI classification NDI (mean ^ SD) Recovered group 29 29.3 ^ 11.72 50,8 2.9 ^ 2.9 Mild pain and disability group 30 34.3 ^ 12.5 77 10 28 16.5 ^ 5.6 Moderate/severe pain and disability group 17 43.7 ^ 14.5 94.30 42.8 ^ 12.2 Control group 20 40.1 ^ 13.6 60 another city (two subjects), a head injury several weeks following the whiplash injury (one subject) and no reason given (one subject). The remaining 76 subjects formed the 6 month classification. The NDI of the recovered group was 3 ^ 2.9 (mean ^ SD), the mild group 16.5 ^ 5.6 and the moderate/severe group 42.8 ^ 12.2. Thirty-eight percent of the whiplash subjects reported recovery by 6 months postinjury, 39.6% reported persistent mild pain and disability and 22.4% persistent moderate/severe pain and disability based on NDI scores at 6 months. Age and gender distribution of the four groups is illustrated in Table 1. There was an uneven distribution of males and females and differences in ages between the groups approached significance ðp ¼ 0:03Þ. As a consequence, age and gender were included as covariates in the initial analysis. NDI scores and VAS scores of pain intensity at each time point are shown in Table 2. After the accident, the onset of pain was immediate in 34% of subjects, occurred within 24 h in 46% and after 24 h in 20%. Forty-six percent of collisions were rear impact, 22% were front on impact, 21% were combined rear and front on collisions and 11% were side impact. Sixty-one percent of subjects filed a compensation claim. This study did not aim to investigate the effect of treatment. Subjects were free to pursue any form of treatment. The types and numbers of treatments received (including medication) were similar between the three whiplash groups (Table 3). 3.2. Pressure pain thresholds The marginal means (^SEM) of the four groups for measures of PPT at cervical spine, median nerve and tibialis anterior sites, are presented in Fig. 1. There was a significant main effect for Group ðp, 0:01Þ for all test sites. There was a significant interaction effect between Group and Time for both cervical spine sites (C2/3, C5/6) ðp, 0:01Þ. The group with moderate/severe symptoms at 6 months showed lower PPTs at all sites when compared with controls and the other two whiplash groups ðp, 0:01Þ. PPTs of this group did not significantly change over the study period and remained less than all other groups at 6 months post-injury ðp, 0:01Þ. The recovered and mild pain groups showed lower PPTs at both cervical spine sites (C2/3, C5/6) than control subjects ðp, 0:01Þ at entry into the study (, 1 month). However, both these groups improved over time ðp, 0:01Þ and by 2 months post-injury were no longer different from control subjects. There was no effect of age on PPTs ðp. 0:2Þ but there was an effect of gender at all sites ðp, 0:01Þ with females having lower PPTs than males. 3.3. Thermal pain thresholds There was a significant group difference for both heat and CPTs ðp, 0:01Þ but there was no interaction effect between Group and Time for either measure of thermal pain threshold ðp. 0:09Þ indicating that thermal pain thresholds remained stable in all groups over the study period (Fig. 2). HPT of the moderate/severe group was 39.5 ^ 0.48, which was significantly lower than that of the control group (43.2 ^ 0.48), the recovered whiplash group (42.6 ^ 0.38) and those with mild symptoms (43.1 ^ 0.38) (all P, 0:01). CPT of the moderate/severe group was 19.19 ^ 1.58, which was significantly higher than that of the control group (9.66 ^ 1.48), the recovered whiplash group (11.57 ^ 1.18) Table 2 Mean (SD) NDI and VAS scores for each whiplash group (recovered, mild pain and disability and moderate pain and disability) at each time point,1 month 2 months 3 months 6 months NDI Recovered 19.14 (12.7) 8 (8.2) 5.4 (6.8) 2.9 (2.9) Mild pain and disability 36.1 (19.4) 25.6 (10.8) 21.45 (12.6) 16.3 (5.6) Moderate/severe pain and disability 55.6 (13.4) 49.1 (15.1) 47.4 (15.4) 42.8 (12.2) VAS Recovered 2.3 (0.9) 1.5 (0.8) 0.6 (0.1) 0.3 (0.1) Mild pain and disability 3.2 (1.2) 2.6 (0.9) 0.9 (0.2) 2.0 (0.7) Moderate/severe pain and disability 3.2 (1.3) 3.8 (1.3) 1.3 (0.3) 3.4 (1.0)

M. Sterling et al. / Pain 104 (2003) 509 517 513 Table 3 The numbers and types of treatment and medication received by the three whiplash groups Group N (%) who received treatment No. of treatments (average/study period) Treatment type N (%) N (%) on medication Medication type Recovered (n ¼ 29) 14 (48.3%) 10.6 Physiotherapy 29 (100%) 7 (24%) Simple analgesics (3), NSAIDS (4), codeine (1), anti-depressants (0), steroids (1), opioids (0) Mild symptoms (n ¼ 30) 19 (63%) 14.4 Physiotherapy 14 (46.7%), chiropractic 4 (13.3%), acupuncture 1 (3.3%) Moderate/severe symptoms (n ¼ 17) 9 (52.9%) 18.4 Physiotherapy 8 (47.1%), chiropractic 1 (5.8%) 13 (43.3%) Simple analgesics (2), NSAIDS (10), codeine (2), Anti-depressants (1), steroids (0), opioids (1) 12 (70.5%) Simple analgesics (2), NSAIDS (7), codeine (2), anti-depressants (2), steroids (0), opioids (1) and the group with mild symptoms (11.39 ^ 1.18). There was no effect of age on either measure ðp. 0:13Þ but a significant effect of gender on both measures ðp, 0:001Þ with females having lower CPTs and lower HPTs than males. 3.4. Brachial plexus test There was a significant Group effect for both elbow extension and pain scores with the BPPT ðp, 0:01Þ and a significant interaction effect between Group and Time for both measures ðp, 0:01Þ. Both the groups with moderate/severe and mild symptoms at 6 months postinjury showed less range of elbow extension (234.27 ^ 3.48, 233.97 ^ 2.6, respectively), and higher VAS scores (4.1 ^ 0.5, 3.2 ^ 0.5) at entry into the study (,1 month) than both the control group (220.67 ^ 3.128, 1.8 ^ 0.4) and the whiplash group who recovered at Fig. 1. PPTs at cervical spine, median nerve and tibialis anterior sites (means and SEM) for all groups (control, recovered, mild pain and moderate/severe pain) over time (1, 2, 3 and 6 months post-injury).

514 M. Sterling et al. / Pain 104 (2003) 509 517 Fig. 2. HPTs and CPTs (means and SEM) for all groups (control, recovered, mild pain and moderate/severe pain) over time (1, 2, 3 and 6 months post-injury). 6 months (223.95 ^ 2.48, 1.8 ^ 0.4) ðp, 0:01Þ (Fig. 3). The group with mild symptoms improved over time ðp ¼ 0:004Þ and were no different from controls by the 2 month assessment point. However, the group with moderate/severe symptoms showed no change over time ðp. 0:09Þ and continued to demonstrate less elbow extension and higher VAS scores than controls at 6 months post-injury ðp ¼ 0:002Þ. There was no effect of age or gender on either measure of the BPPT ðp. 0:4Þ. 3.5. Sympathetic vasoconstrictor reflex There was no significant effect for Group on both quotients of the SVR (QI and SRF) ðp. 0:07Þ nor was there any interaction effect between Time and Group for either measure ðp. 0:98Þ (Table 4). However, the moderate/ severe group tended to show higher QI and lower SRF values than the other two whiplash groups. 3.6. Psychological distress (GHQ-28) There was a significant main effect for Group for GHQ-28 total score ðp, 0:01Þ and a significant interaction between Group and Time ðp, 0:001Þ. As can be seen in Fig. 4, the groups with moderate/severe (41 ^ 3) or mild symptoms (33 ^ 3) both had a total GHQ-28 score above the threshold of 23/24 at entry into the study (,1 month). Both groups significantly improved over the 6 month study period but the moderate/severe group (33.5 ^ 3) continued to show a GHQ-28 total score above the threshold at 6 months post-injury, whereas the mild group returned to below threshold levels (21.3 ^ 2). When GHQ-28 total scores were included in the analysis of the three whiplash groups, group differences remained Table 4 Estimated marginal means (SEM) of the sympathetic vasoconstrictor reflex parameters (QI and SRF) after inspiratory gasp in all groups (all P. 0:07) Group QI SRF Recovered 54 ^ 17.2 0.79 ^ 0.17 Mild pain and disability 53.1 ^ 16.9 0.79 ^ 0.18 Moderate/severe pain and disability 64.8 ^ 18.2 0.69 ^ 0.15 Controls 52.4 ^ 18.4 0.71 ^ 0.18 significant for all measures ðp, 0:01Þ. There was no interaction between Group and GHQ-28 total score for any measure, suggesting the effect of psychological distress (as measured by the GHQ-28) is similar irrespective of group allocation. The effect size for GHQ-28 (total) on the measures of sensory function and SNS activity was small (h 2 ranged from 0.027 to 0.147). 4. Discussion Little is known about the continuum of WAD from the time of injury through transition to either recovery or chronicity. The results of this study provide the first evidence that the presence of generalised sensory hypersensitivity can differentiate those with persistent moderate/severe symptoms at 6 months following whiplash injury from those who have largely recovered. These sensory disturbances occurred independently of psychological distress, within a month of injury and persisted unchanged to 6 months post-injury. It is likely that such changes in sensory function reflect altered nociception within the central nervous system. Supporting previous longitudinal studies (Gargan et al., 1997; Radanov et al., 1995), 61% of our whiplash subjects reported ongoing pain and disability of varying degrees at 6 months post-injury. Interestingly, of this group with persistent symptoms, there exists a subgroup of subjects (22% of the total cohort) in which there was evidence of altered nociception and it is this subgroup who reported more disabling pain levels. This may reflect different underlying mechanisms between those with higher pain levels and those with lesser symptoms. All whiplash injured subjects, irrespective of the level of reported symptoms, demonstrated early local mechanical hyperalgesia within the cervical spine possibly reflecting sensitisation of peripheral nociceptors resulting from injured neck structures as proposed in previous studies of acute whiplash (Kasch et al., 2001a; Yerner et al., 2001). Local mechanical hyperalgesia had resolved by 2 months post-injury in recovered whiplash subjects and those with lesser symptoms whereas it persisted in those with ongoing moderate/severe symptoms. This may reflect healing of the underlying soft tissue injury in recovered subjects. Generalised hypersensitivity including widespread

M. Sterling et al. / Pain 104 (2003) 509 517 515 Fig. 3. Range of elbow extension at pain threshold and VAS scores of pain (means and SEM) with the BPPT for all groups (control, recovered, mild pain and moderate/severe pain) over time (1, 2, 3 and 6 months post-injury). mechanical and thermal hyperalgesia and heightened responses to the BPPT was the differentiating feature of those with persistent moderate/severe symptoms. Widespread hypersensitivity to blunt pressure has been proposed to occur as a result of sensitisation of central nervous system nociceptive pathways or changes in endogenous descending pain modulation mechanisms (Koelbaek-Johansen et al., 1999; Ren et al., 2000; Treede et al., 2002). This phenomenon has been demonstrated in subjects with chronic WAD (Koelbaek-Johansen et al., 1999; Sterling et al., 2002a) but its existence in the earlier stages of the condition has been disputed (Kasch et al., 2001b). However, past investigation of such changes have not differentiated whiplash subjects on the basis of levels of pain and disability and as such may have overlooked identification of important sub-groups within this condition. The results of our study emphasise that whiplash injury is not a homogenous condition. Heightened responses to the BPPT and thermal (heat and cold) hyperalgesia were seen soon after injury in the group with persistent moderate/severe symptoms. Bilateral loss of elbow extension and higher pain levels with the BPPT in chronic WAD have been interpreted as reflecting both hyperalgesic motor and sensory responses as a consequence of central sensitisation (Quintner, 1989; Sterling et al., 2002b). Reduced HPT was only present in the more severe group and may be a feature of nociceptor sensitisation (Kilo et al., 1994). Enhanced sensitivity to innocuous heat has also Fig. 4. GHQ-28 total scores (mean and SEM) for all groups (control, recovered, mild pain and moderate/severe pain) over time (1, 2, 3 and 6 months post-injury). been proposed to occur due to convergence of fibres activated by noxious stimuli and heat upon sensitised dorsal horn neurons (Kosek and Ordeberg, 2000). As such the reduced HPT seen in the moderate/severe group may be another reflection of this group s general hypersensitive state. Cold hyperalgesia may be due to changes in the central mediation of pain (Berglund et al., 2002) but has also been shown to be a feature of pain due to peripheral nerve injury (de Medinaceli et al., 1997) and disturbances of SNS activity (Frost et al., 1988). The possibility of nerve injury as a contributor to symptoms of those with persistent moderate/severe symptoms cannot be discounted. Cold hyperalgesia, together with heightened responses to the BPPT and mechanical hyperalgesia over peripheral nerve trunk sites may be indicative of such injury. Injury to nerve tissue such as nerve roots and dorsal root ganglia has been demonstrated in cadaveric studies following severe whiplash injury (Taylor and Taylor, 1996) and in clinical studies where evidence of nerve tissue irritation and ensuing mechanosensitivity has been shown to be present in chronic WAD (Ide et al., 2001; Sterling et al., 2002b). Disturbances in SNS function are not a feature of WAD as shown on SVR testing. However, there was a tendency for higher QI and lower SRF quotients in the whiplash group with moderate/severe symptoms at 6 months post-injury. A closer inspection of the SVR data in this group revealed that seven of the 17 subjects showed values for both the QI and SRF quotients outside approximated normal physiological ranges (mean values ^ 2 SD) (Schurmann et al., 1996) and similar to those of patients with complex regional pain syndrome type 1 (Schurmann et al., 1999). This may suggest that SNS dysfunction could exist in some whiplash patients and further investigation involving larger subject numbers is indicated. All whiplash subjects in our study were psychologically distressed to some degree. Both groups with persistent moderate/severe or mild pain and disability had GHQ-28 total scores above the threshold at entry into the study (, 1 month post-injury) with the recovered group also approaching this threshold. All three groups improved over time until at 6 months post-injury, only the moderate/severe

516 M. Sterling et al. / Pain 104 (2003) 509 517 group continued to score above the threshold. This is consistent with previous studies showing elevated psychological distress in chronic WAD, likely as a result of their ongoing pain and disability (Peebles et al., 2001; Radanov et al., 1996). The presence of hypersensitivity in WAD has been suggested as being due to the patient s psychological distress (Ferrari, 2001). However, when GHQ-28 total scores were included in the analysis of our data, group differences on the quantitative sensory tests remained significant and effect sizes of the GHQ-28 scores on sensory variables was small. Furthermore, whilst the GHQ-28 scores of the moderate/severe group significantly decreased over the study period, all hypersensitive responses remained unchanged. If the hypersensitivity of the moderate/severe group seen in this study was merely as a consequence of psychological distress, a similar pattern of change would be expected for both psychological and sensory variables. An alternative explanation for the hypersensitive responses in the moderate/severe group is disturbances in central pain processing mechanisms. The results of this study have implications for the early management of WAD. Acute WAD is not a homogenous condition and identification of those with early sensory disturbances may be important. It has been argued that appropriate expeditious treatment may help to prevent transition from acute pain into persistent pain (Cousins, 2002). In the case of the at risk patients identified in this study, this may involve appropriate early pharmaceutical pain management. Acknowledgements This work was supported by Suncorp Metway Insurance, Queensland and Centre of National Research on Disability and Rehabilitation Medicine (CONROD). References Adeboye K, Emerton D, Hughes T. Cervical sympathetic chain dysfunction after whiplash injury. J R Soc Med 2000;93:378. Balster S, Jull G. Upper trapezius activity during the brachial plexus tension test in asymptomatic subjects. Man Ther 1997;2:144 9. Barnsley L, Lord S, Bogduk N. Clinical review. Whiplash injury. Pain 1994;58:283 307. Berglund B, Harju E-L, Kosek E, Lindblom U. Quantitative and qualitative perceptual analysis of cold dysesthesia and hyperalgesia in fibromyalgia. Pain 2002;96:177 87. Brennum J, Kjeldsen M, Jensen K, Jensen K. Measurements of human pressure pain thresholds of fingers and toes. Pain 1989;38:211 7. Clarkson H, Gilewich G. Musculoskeletal assessment: joint range of motion and manual muscle strength. Baltimore, MD: Williams and Wilkins; 1989. Cote P, Cassidy D, Carroll L, Frank J, Bombardier C. A systematic review of the prognosis of acute whiplash and a new conceptual framework to synthesize the literature. Spine 2001;26:E445 58. Cousins M. Evidence for persisting pain as a disease entity: clinical implications. Australian Pain Society, 23rd Annual Scientific Meeting, Sydney, Australia, 2002 Curatolo M, Petersen-Felix S, Arendt-Nielsen L, Giani C, Zbinden A, Radanov B. Central hypersensitivity in chronic pain after whiplash injury. Clin J Pain 2001;17:306 15. Curatolo M, Banic B, Petersen-Felix S, Andersen O, Radanov B, Villeger P, Arendt-Nielsen L. Preliminary electrophysiological evidence for central hypersensitivity in whiplash pain and fibromyalgia. 10th World Congress on Pain, San Diego, CA: IASP; 2002. Elvey RL. Brachial plexus tension test and the pathoanatomical origin of arm pain. In: Glasgow E, Twomey L, editors. Aspects of manipulative therapy. Melbourne: Lincoln Institute of Health Sciences; 1979. p. 105 10. Ferrari R. Whiplash and symptom amplification. Pain 2001;89:203 302. Frost S, Raja S, Cambell J, Meyer R, Khan A. Does hyperalgesia to cooling stimuli characterise patients with sympathetically maintained pain? In: Dubner R, Gebhart G, Bond M, editors. 8th World Congress on Pain. Amsterdam: Elsevier; 1988. p. 151 6. Gargan M, Bannister G, Main C, Hollis S. The behavioural response to whiplash injury. J Bone Joint Surg 1997;79-B:523 6. Goldberg D. Manual of the general health questionnaire. Windsor: NFER- Nelson; 1978. Hurtig I, Raak R, Kendall S, Gerdle B, Wahren L. Quantitative sensory testing in fibromyalgia patients and in healthy subjects: identification of subgroups. Clin J Pain 2001;17:316 22. Ide M, Ide J, Yamaga M, Takagi K. Symptoms and signs of irritation of the brachial plexus in whiplash injuries. J Bone Joint Surg Br 2001;83: 226 9. Kasch H, Stengaard-Pedersen K, Arendt-Nielsen L, Jensen T. Headache, neck pain and neck mobility after acute whiplash injury. Spine 2001a; 26:1246 51. Kasch H, Stengaard-Pedersen K, Arendt-Nielsen L, Staehelin Jensen T. Pain thresholds and tenderness in neck and head following acute whiplash injury: a prospective study. Cephalalgia 2001b;21:189 97. Kilo S, Schmelz M, Koltzenburg M, Handwerker H. Different patterns of hyperalgesia induced by experimental inflammation in human skin. Brain 1994;117:385 96. Koelbaek-Johansen M, Graven-Nielsen T, Schou-Olesen A, Arendt- Nielsen L. Muscular hyperalgesia and referred pain in chronic whiplash syndrome. Pain 1999;83:229 34. Kosek E, Ordeberg G. Abnormalities of somatosensory perception in patients with painful osteoarthritis normalize following successful treatment. Eur J Pain 2000;4:228 38. Mani R, Cooper C, Kidd B, Cole J, Cawley M. Use of laser doppler flowmetry and transcutaneous oxygen tension electrodes to assess local autonomic dysfunction in patients with frozen shoulder. J R Soc Med 1989;82:536 8. Mayou R, Radanov B. Whiplash neck injury. J Psychosom Res 1996;40: 461 74. de Medinaceli L, Hurpeau J-C, Merle M, Begorre H. Cold and posttraumatic pain: modeling of the peripheral nerve message. BioSystems 1997;43:145 67. Moog M, Quintner J, Hall T, Zusman M. The late whiplash syndrome: a psychophysical study. Eur J Pain 2002;6:283 94. Munglani R. Neurobiological mechanisms underlying chronic whiplash associated pain. J Musculoskelet Pain 2000;8:169 78. Peebles J, McWilliams L, MacLennan R. A comparison of symptom checklist 90-revised profiles from patients with chronic pain from whiplash and patients with other musculoskeletal injuries. Spine 2001; 26:766 70. Quintner J. A study of upper limb pain and paraesthesiae following neck injury in motor vehicle accidents: assessment of the brachial plexus tension test of Elvey. Br J Rheumatol 1989;28:528 33. Radanov B, Begre S, Sturzenegger M, Augustiny K. Course of psychological variables in whiplash injury a 2-year follow-up with age, gender and education pair-matched patients. Pain 1996;64: 429 34.

M. Sterling et al. / Pain 104 (2003) 509 517 517 Radanov B, Sturzenegger M, Di Stefano G. Long-term outcome after whiplash injury. A 2-year follow-up considering features of injury mechanism and somatic, radiologic, and psychological findings. Medicine 1995;74:281 97. Ren K, Zhuo M, Willis W. Multiplicity and plasticity of descending modulation of nociception: implications for persistent pain. In: Devor M, Rowbotham M, Wiesnfeld-Hallin Z, editors. IXth World Congress on Pain, Vol. 16. Vienna: IASP; 2000. p. 387 400. Rhudy J, Meagher M. Fear and anxiety: divergent effects on human pain thresholds. Pain 2000;84:65 75. Schurmann M, Gradl G, Furst H. A standardized bedside test for assessment of peripheral sympathetic nervous function using laser doppler flowmetry. Microvasc Res 1996;52:157 70. Schurmann M, Gradl G, Andress H, Furst H, Schildberg F. Assessment of peripheral sympathetic nervous system function for diagnosing early post-traumatic complex regional pain syndrome type I. Pain 1999;80: 149 59. Selvaratnam P, Matyas T, Glasgow E. Noninvasive discrimination of brachial plexus involvement in upper limb pain. Spine 1994;19:26 33. Smith R, Papadopolous E, Mani R, Cawley I. Abnormal microvascular responses in lateral epicondylitis. Br J Rheumatol 1994;33:1166 8. Spitzer W, Skovron M, Salmi L, Cassidy J, Duranceau J, Suissa S, Zeiss E. Scientific monograph of Quebec Task Force on whiplash associated disorders: redefining Whiplash and its management. Spine 1995;20: 1 73. Sterling M, Treleaven J, Edwards S, Jull G. Pressure pain thresholds in chronic whiplash associated disorder: further evidence of altered central pain processing. J Musculoskelet Pain 2002a;10:69 81. Sterling M, Treleaven J, Jull G. Responses to a clinical test of mechanical provocation of nerve tissue in whiplash associated disorders. Man Ther 2002b;7:89 94. Taylor J, Taylor M. Cervical spinal injuries: an autopsy study of 109 blunt injuries. J Musculoskelet Pain 1996;4:61 79. Treede R-D, Rolke R, Andrews K, Magerl W. Pain elicited by blunt pressure: neurobiological basis and clinical relevance. Pain 2002;98: 235 40. Vernon H. The neck disability index: patient assessment and outcome monitoring in whiplash. J Musculoskelet Pain 1996;4:95 104. Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. J Manipulative Physiol Ther 1991;14:409 15. Yerner S, Toolanen G, Knibestol M, Gerdle B, Hildingsson C. Prospective study of trigeminal sensibility after whiplash trauma. J Spinal Disord 2001;14:479 86.