Introduction 2 25 5 6 12 7 8 7 8 We therefore conducted a double blind placebo-controlled randomized trial to confirm if botulinum toxin in treatment of neck pain is indeed effective in chronic whiplash syndrome. Methods Patients 24 Study design The trial had a double-blind, placebo-controlled parallel design. After patient’s consent was obtained, we recorded demographics (age, gender). During a baseline-period of 4 weeks, patients used a diary to record the presence of neck pain, number of neck pain hours per day, number of days on which medication was taken, and number of tablets taken per day. At the end of this baseline-period (week 0) patients recorded their mean neck pain intensity scored on 100 mm Visual Analogue Scale of the previous 4 weeks (week –4 to 0, VAS-pre-treatment). At week 0, patients were randomly assigned to receive 100 units botulinum toxin (Botox®) or placebo (saline) in 2 cc syringes. At 12 weeks follow-up they recorded again their mean neck pain intensity of the previous 12 weeks (week 0 to 12, VAS-post-treatment). The local trial pharmacist prepared the drug, coded the syringes, and kept treatment codes. The pharmacist had no further participation in the trial. An experienced clinical neurophysiologist (D.T.) performed all injections. Injection sites were selected individually (i.e. ‘follow the pain’ approach) in muscles with clinically increased muscle tone or muscle tenderness, and included m. occipitofrontalis, m.temporalis, m. masseter, m. sternocleidomastoideus, m. splenius capitis, m. trapezius and m. semispinalis. Outcomes After patients were injected at week 0 in one single treatment-session, clinical characteristics were recorded using diaries during the 12 weeks post-treatment. The primary outcome measure was the intensity of the neck pain of the total study period, week 0–12, scored on the VAS at week 12 (one single measurement) compared with baseline (VAS week –4 to 0). Patients were asked to rate (in a single value) severity of neck pain which at times may be mild and perhaps at others severe, by “integrating severity over time”. Secondary outcome measures were average number of neck pain days, average number of neck pain hours per day, average number of days on which symptomatic treatment was taken, mean number of symptomatic tablets per day at 12 weeks compared with baseline. Additionally, a blinded investigator asked patients whether there was any improvement in their condition after 4, 8 and 12 weeks in the two treatment groups (five-point scale: great worsening, any worsening, no improvement, any improvement and great improvement). Range of cervical motion was measured with a cervical range of motion (CROM) device to measure the six conventional movements of the spine at week 12 compared with baseline. For measurements of CROM, the patients were seated in a chair with their shoulders supported on a backrest. Patients were instructed to assume a comfortable position and then to perform each required movement and return to the start position. The movements were assessed in this order: right rotation, left rotation, right flexion, left flexion, extension, flexion and then summated to determine the mean total range of motion. We also divided all patients in responders and non-responders at week 12. We considered at least 45% reduction in the VAS compared with baseline to be clinically relevant (e.g. responder). All adverse events during the 3 months follow-up were recorded by the investigators. Sample size determination At baseline we expected mean neck pain intensity on the VAS of 40/100 mm. The estimated placebo effect was a 8 mm reduction to a VAS intensity of 32/100 at 12 weeks. We hypothesized a 18 mm reduction in the treatment group to a VAS intensity of 22/100, although there are no earlier data on treatment effect of BTX in these patients. To detect this clinically relevant effect (8 versus 18 mm improvement) at a 5% level of statistical significance (two-tailed), with a power of 80%, a total of 40 patients was needed. Statistical analysis We calculated mean differences between baseline VAS data and VAS data after 12 weeks, using the t-test; statistical uncertainty was expressed in 95% confidence intervals. The responders at week 12 in each treatment group and the differences in improvement scores were compared by use of the Chi-square test. All these analysis were done with SPSS (version 10.0). Results Between November 1999 and March 2003 40 patients were included at Leyenburg hospital The Hague. Of the 59 patients assessed for eligibility, 19 patients were excluded. Four patients did not meet the inclusion criteria, eight patients refused to participate and seven patients withdrew in the baseline period. No patient was lost to follow-up. 1 Table 1 Baseline characteristics Characteristic Botulinum toxin (n = 20) Placebo (n = 20) Mean age, years (range) 39 (20–58) 34 (20–49) Female sex (%) 14 (70%) 13 (65%) Ongoing claims 15 (75%) 18 (90%) Mean duration of pain, months (SD) 40 (71) 37 (20) Neck pain severity, mean VAS (SD) 64.5 (14.8) 62.1 (20.3) Mean number of neck pain days, % (SD) 1 96% (12) Mean neck pain duration per day, hours (SD) 3 11.7 (4.5) Mean number of medication days, % (SD) 1 36% (40) Mean number of analgesics per day, tablets (SD) 1.4 (1.9) 1.3 (2.1) Mean total cervical range of motion, degrees (SD) 1 1 1  3  At 12 weeks the VAS scores were 52.0 mm (SD = 29.2) in the botulinum group and 56.7 mm (SD = 29.6) in the placebo group. Mean differences between VAS pre- and post-treatment was 12.5 in the botulinum group and 5.5 in the placebo group. The therapeutic gain ( = difference in improvement between the two treatment groups measured in mm) was –7.0 (p = 0.31; 95% CI (–20.7 to + 6.7)). Six patients in the botulinum group were responders as defined, versus four in the placebo group (NS, p = 0.5). 2 Table 2 Improvement scale Improvement At 4 weeks At 8 weeks At 12 weeks Botulinum toxin (n = 20) Placebo (n = 20) Botulinum toxin (n = 20) Placebo (n = 20) Botulinum toxin (n = 20) Placebo (n = 20) Great worsening 4 0 0 2 1 2 Any worsening 2 4 1 4 2 3 No improvement 6 8 8 7 6 8 Any improvement 5 5 5 4 5 4 Great improvement 2 3 6 3 6 3 Total 1 20 20 20 20 20 1  3 Table 3 Secondary outcome measures Outcome Botulinum toxin (n = 20) Placebo (n = 20) Therapeutic gain (95% CI) Neck pain days (% ± SD)   Baseline (4 weeks) 1 96% (12)    Post-treatment (12 weeks) 2 2    Mean difference 8% (18) 7% (22) −1% (−15% to + 13%) Neck pain duration (hours/day ± SD)   Baseline (4 weeks) 3 11.7 (4.5)    Post-treatment (12 weeks) 7 2    Mean difference 0.86 (2.2) 0.72 (4.7) −0.14 (−3.0 to + 2.7) Days on which analgesics were taken (% ± SD)   Baseline (4 weeks) 49% (42%) 35% (40%)    Post-treatment (12 weeks) 3 2    Mean difference 6.4% (16%) 5.7% (24%) 0.7% (−15% to + 13%) Number of analgesics taken per day (tablets/day ± SD)   Baseline (4 weeks) 1 1.3 (2.1)    Post-treatment (12 weeks) 2 1    Mean difference 0.32 (0.8) 0.18 (0.6) −0.14 (−0.6 to + 0.4) Total cervical range of motion (degrees ± SD)   Baseline 1 1    Post-treatment (week 12) 2 3    Mean difference 22 (43) 11 (36) −11 (−40 to + 17) 1 2 3 7 9 Twenty-seven patients reported minor side-effects, 15 in the botulinum toxin group and 12 in the placebo group. The main complaint was “short-lasting pain at the injection sites”. Discussion Our study could not demonstrate botulinum toxin to be more effective than placebo in treatment of neck pain in patients with chronic whiplash syndrome. 12 7 8 This study, however, has some methodological shortcomings. Primary outcome was improvement of headache intensity scored on the VAS after 4 weeks compared with baseline. The post treatment difference between the two treatment groups was reported to be significant. However, baseline-characteristics were significantly different, with a significant higher headache intensity in the botulinum group at baseline. (VAS 6.5 (range 2–9) versus 3.0 (0–8) in the placebo group (p < 0.01)). The reported improvement may mainly have been caused by the unequal distribution of the patients at baseline. In their second publication ( on the same trial) the mean total pain score was reported, ranging from 0–30, expressing the total pain experience for neck pain, headache and shoulder pain. After 4 weeks the mean total pain was 10.0 (SE 1.3) in the botulinum group and 14.1(SE 21) in the placebo group, which was also reported to be significant. These data are difficult to interpret without individual scores. Calculation of the standard deviation could suggest that some patients even had a negative mean total pain score. In any case, the distribution was truncated and skew to the right. 10 16 20 16 14 17 23 26 Patients with chronic whiplash syndrome also form a very heterogeneous group, with a wide spectrum of symptoms, varying from neck pain to headache, with pain of different degrees. Many different factors seem to play a role in the development of acute and chronic pain after a whiplash injury. 4 21 1 13 4 15 18 22 11 19 3 Based on present evidence BTX cannot be recommended as treatment for neck pain in chronic whiplash patients. Future studies directed on possible central mechanisms of this complicated chronic pain syndrome are warranted. Contributors The trial was designed by S.F.T.M. de Bruijn, D.L.J. Tavy and M. Padberg. Statistical analysis was performed by M. Padberg. Conflict of interest statement Non declared.