Anticholinergic and GABAergic drugs (benzodiazepines and baclofen) are both effective in reducing the symptoms of cervical dystonia including the pain. In the former category, trihexyphenidyl (6–30 mg/day) and benztropine (1–3 mg day) are the two most commonly used drugs. Baclofen (30–60 mg day) and diazepam (10–30 mg/day) also improve symptoms of CD. Clonazepam 2.5–1 mg three times daily is also helpful especially when CD is associated with cervical myoclonus. All aforementioned drugs need to be started at low dose and gradually built up over several weeks. Elderly patients have poor tolerance for anticholinergic medications, and many such patients also find baclofen hard to tolerate. Patients who fail pharmacological treatment may benefit from selective cervical rhizotomy (cutting the cervical nerve roots) under careful electrophysiological guidance. The procedure should be performed in specialized centers. In recent years, deep brain stimulation has been employed for the treatment of recalcitrant cases of cervical dystonia. Although the number of treated patients is still small, pallidal stimulation seems to be effective as a secondary therapeutic option for severe CD (Walsh et al. 2013).

Introduction of BoNTs for the treatment of cervical dystonia has revolutionized the management of this disorder. All types of BoNTs (ona, abo, inco, and rima, types A and B) have been proven effective with a rate of efficacy of over 80 % and safety profile unmatched by any other therapeutic modality. Treatment improves all major symptoms of CD and prevents development of contractures and radiculopathy (Jankovic 2004). More favorable results have been reported in one study in which BoNT therapy was combined with physical therapy (Tassorelli et al. 2006). In a recent systematic review and meta-analysis (Marsh et al. 2014) which assessed 18 studies and over 1,900 patients, the mean duration of onaA effect was 93.2 and 95.2 days for fixed and random effect models, respectively. Doses of over 180 units of onaA for the treatment of cervical dystonia produced longer-lasting effects (107–109 days vs. 86–88 days for doses of <180 units).

One of the most important early observations in treatment of CD with botulinum neurotoxins (in this case onaA) was the recognition that neck pain relief in CD often occurred before improvement of posture and limitation of head movement (Jankovic and Schwartz 1990). This important early observation suggested an analgesic effect for botulinum toxins in human, independent from its other effects which proved to be correct in the following years.

Among the double-blind studies which have reported on the efficacy of BoNTs in cervical dystonia, 14 included assessment of pain.

In the pioneering study of Tsui et al. (1986), 14 of 16 patients with CD reported significant reduction of neck pain after administration of onabotulinumtoxinA into the neck and shoulder muscles (0.002). In a larger study of 55 patients, Greene et al. (1990) also reported significant pain relief of their subjects 6 weeks after administration of onaA for CD. In another study of 23 patients, 19 of whom complained of pain, Lorentz et al. (1991) reported pain relief in 12 of 19 patients who were injected by onaA, but only in 1 of 19 subjects injected by saline (P = 0.002). Lew et al. (1997) conducted an efficacy and safety study on 122 patients with CD evaluating the effects of 2,500, 5,000, and 1,000 units of rimaB against placebo (saline). Pain was assessed through the pain subset of TWSTRS and the visual analog scale (VAS). At 4 weeks, all three doses had produced significant pain relief compared to the placebo (P < 0.05). This relief was more pronounced for the largest dose used in the study (<0.004). Poewe et al. (1998), in a study of 31 patients with neck pain and CD, also found a clear difference in pain improvement in favor of aboA (compared to placebo) at 4 weeks. The difference between the three dose groups of toxin (250, 500, and 1,000 units), however, was not significant.

In 1999, two studies assessed the efficacy of rimabotulinumtoxinB (rimaB) in cervical dystonia and associated pain. In one study (Brashear et al.), the investigators compared the efficacy of 5,000 and 10,000 units of rimaB with placebo in 109 patients using visual analog scales. Compared to placebo, at 4 weeks, significant reduction of pain was noted in both toxin groups compared to placebo; 5,000 unit group (P = 0.001) and 10,000 unit group (P = 0.0002). Overall, the TWSTRS scores also improved more in the 10,000 unit group. In the same year, Brin et al. (1999) published the results of another investigation on efficacy of rimaB versus placebo in 77 patients with CD. Administration of 10,000 units of rimaB improved neck pain significantly at 4 weeks (P < 0.001). Wissel et al. (2001) studied 68 patients with CD (with minimum Tsui score of 9) comparing the effect of aboA (500 units) with placebo (saline). Patients were assessed at weeks 4, 8, and 12 with Tsui scale rating the severity of CD and pain. At week 4, 49 patients in aboA group were pain-free versus 33 patients in the saline group (P = 0.02). In the following open phase of the study, the aboA group demonstrated significant pain relief (P = 0.011). Truong et al. (2005) investigated the efficacy and safety of abobotulinumtoxinA (aboA) (500 units) in 80 patients with CD. Participants were followed up for 4–20 weeks, until they needed further treatment. The efficacy was assessed with TWSTRS at baseline and weeks 2, 4, 8, 12, 16, and 20 after treatment. Pain was evaluated via pain subset of TWSTRS on VAS scale of 0–100 mm. At 4 weeks, the reduction in VAS score was 13.4 for aboA group and 1.9 for the placebo group (P = 0.02). This significant degree of pain reduction was also noted at week 8.

In the past 3 years, four multicenter studies in a sizeable number of patients with CD-related pain and botulinum toxin therapy have been published (Truong et al. 2010; Comella et al. 2011; Charles et al. 2012; Fernandez et al. 2013). Truong et al. (2010) reported on the results of a multicenter study of 116 patients (61 placebos) with CD after administration of 500 units of abobotulinumtoxinA (aboA) into the neck and shoulder muscles. Four weeks after administration of aboA, the VAS score was reduced 3.7 for the onaA group and 1.4 for the placebo group, respectively. Comella et al. (2011) reported on the efficacy of two doses of incobotulinumtoxinA (incoA), 120 and 240 units on 233 patients with cervical dystonia. Both doses were equally effective in improving all subsets of the TWSTRS including pain. The pain subset of TWSTRS (0–20) was markedly improved (P < 0.0001) at 4, 8, and 12 weeks. In another study (Charles et al. 2012), the efficacy of onabotulinumtoxinA (mean dose 241 unit) versus placebo was assessed in 170 patient with cervical dystonia (88 onaA, 82 placebo) using dystonia severity scale and physicians global assessment scale at baseline and 6 weeks after injection. Evaluation of pain subset showed significant improvement (P < 0.05) at 2, 4, and 6 weeks posttreatment in the toxin group versus placebo. In a multicenter double-blind, placebo-controlled study, Fernandez et al. (2013) studied the effect of two doses (120 and 240 units) of incobotulinumtoxinA in 233 patients with CD. Pain was assessed through TWSTRS pain subset. At 4 weeks postinjection, patients in both 120 and 240 groups demonstrated significant reduction of neck pain (P < 0.0001). There was no significant difference in pain response between the two prescribed doses of the neurotoxin. In contrast, Kaji et al. (2013), using the TWSTRS pain subset, found doses of 2,500 and 5,000 units of rimabotulinumtoxinA (rimaA) ineffective in alleviating the pain of cervical dystonia at 4 weeks following toxin administration. The 10,000 unit dose, however, improved the pain significantly (P < 0.05).