Therapeutic Use of Botulinum Toxin Type A in Treating Neck and Upper-Back Pain of Myofascial Origin: A Pilot Study

Participants 

Twenty-nine adult subjects with diagnoses of cervical or upper-back pain of myofascial origin participated in the study. Through flyers and advertisements, subjects were recruited from the general population and the patient population of the physical medicine and rehabilitation clinic in a university-affiliated tertiary hospital. The study was approved by the institutional review board of the hospital’s affiliated university. Subjects were included if they (1) had been diagnosed as having neck or upper-back pain of myofascial origin within the past 2 to 6 months or had a previous diagnosis of neck or upper-back pain of myofascial origin but had experienced an exacerbation of pain symptoms within the same period, (2) were aged 18 to 70 years, and (3) had a VAS pain score of 5 or greater for the 4-week period before injection. The exclusion criteria were the following: (1) allergy to BTX-A; (2) any medical condition that put a subject at risk with exposure to BTX-A; (3) acute pathology such as infection, inflammation, cervical radiculopathy, or any operative pathology, as shown on physical examination or magnetic resonance imaging; (4) use of aminoglycoside antibiotics, curare-like agents, or other agents that may interfere with the neuromuscular junction; (5) history of gastroesophageal reflux disease (GERD); (6) abnormal swallowing test results on baseline; and (7) pregnancy, breastfeeding, or planned pregnancy. Use of concomitant pain medication and physical therapy was allowed, and no instructions were given to subjects to alter their current regimen.

Subjects were referred by other physicians and their history, physical examinations, and pertinent laboratory and other examination results were reviewed during the screening visit. Initial screening included history taking to include medication and surgical history and duration of pain, as well as completion of preinjection assessments for the VAS for average and maximum pain, the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) to assess functional status, and the Neck Disability Index (NDI). In addition to the screening review, a physical examination was performed to determine the most tender cervical or upper-back muscles and to rule out operative, radiculopathic, and other contraindicated conditions.

Study Intervention 

Qualified, consenting subjects were randomized to the control group (injection with normal saline solution) or the treatment group (injection with BTX-A [Botox]). A computer-generated randomization scheme was used. Subjects were assigned a randomization number, which corresponded to a treatment allocation schedule and was stored by the hospital’s research pharmacist. Hence, for each injection, both the subject and the investigator were blinded. Fourteen subjects were included in the treatment group, and 15 subjects participated in the control group (fig 1). Each vial of BTX-A (100U) or saline was reconstituted by 2mL of normal saline. The BTX-A or saline vials were prepared and reconstituted before receipt by the investigator, who received them in identically appearing syringes, which ensured that he remained blinded as to whether the syringe contained medication or placebo.

View Large Image Download to PowerPoint

Fig 1. Flow of participants in the study.

The dose per injection site was 50U. The total dose did not exceed 200U per treatment and 100U per side. No more than 2 muscles were selected on each side, and only painful muscles were injected. Trigger points were determined before the injection procedure by deep finger pressure. Injections were performed without electromyographic guidance at the site of the trigger points. The muscles that were injected include the trapezius, levator scapulae, splenius capitis, and other posterior neck muscles. For the control group, the same determined volume was injected, but with normal saline. Injection sites were sterilized, and BTX-A or normal saline was administered by a disposable 27-gauge needle with disposable syringe. Each injection was performed slowly, over 15 seconds per location.

The following outcome measures were used: (1) VAS for pain, (2) SF-36, and (3) NDI. Evaluations were performed at baseline, week 2, and months 1, 2, 3, 4, and 6. Subjects were monitored for possible adverse events after BTX-A injection by telephone follow-up at 3 and 10 days postinjection. Monitoring of adverse events continued throughout the entire study period.

Outcome Assessment 

Subjects were evaluated at baseline, then at 2 weeks, and at 1, 2, 3, 4, and 6 months postinjection. For the baseline and second-week assessments, subjects were evaluated in person. For the succeeding evaluation periods, subjects were requested to complete the questionnaires for the outcome measures and to return them by mail to the study coordinator. The study used the following outcome measures to determine subjective pain relief as reported by each patient.

Data Analysis 

The results of the study were analyzed by an independent data management group. SAS softwarea was used for statistical analysis. Patients were included in the analysis on the basis of intention to treat. The hypothesis, that the efficacy in pain relief would be better with BTX-A, was tested through between- and within-group comparisons in the average pain VAS, SF-36, and NDI scores. The normality of variables was tested using Shapiro-Wilk tests. According to the normality of variable, the 1-sided t test (Wilcoxon rank-sum test) and 2-sided paired t test (Wilcoxon signed-rank test) were used for between- and within-group comparisons, respectively. Demographic variables were compared using the Fisher exact test and t test. Frequencies in adverse events were compared with the chi-square and Fisher exact tests.

Demographics 

Thirty subjects consented and were randomized to the study. One subject was excluded in the analysis because his age exceeded 70 years, which was a protocol deviation. There were 20 men and 9 women among the 29 subjects included in the analysis. An analysis of the group characteristics regarding age, sex, and race yielded no significant group differences between the treatment and control subjects (table 1). A comparison of the BTX-A and control groups showed no baseline differences in the outcome measures used in this study: NDI, VAS (average and maximum pain), and the summary measures for SF-36 (average total score, mental health, and physical health).

Table 1. Demographics and Baseline Characteristics of the BTX-A and Control Groups

	Characteristic	BTX-A (n=14)	Control (n=15)	P
	Sex
	Male	8 (57.1)	12 (80.0)	.245⁎
	Female	6 (42.9)	3 (20.0)
	Age (y)	48.7±9.4	48.5±13.4	.955†
		32.0–62.0	23.0–70.0
	VAS pain score (range, (0–10)
	Average pain	6.06±2.00	4.77±1.52	.065†
	Maximum pain	7.89±1.67	7.05±1.90	.218†
	SF-36 summary measures
	Average total score	43.7±15.1	54.9±20.2	.105†
	Physical health	40.7±18.0	52.3±23.5	.149†
	Mental health	46.8±13.8	57.5±19.4	.101†
	NDI (range, 0–50)	20.0–9.4	16.3–11.3	.149†

NOTE. Values are n (%), mean ± standard deviation, and minimum and maximum.

⁎ Two-sided Fisher exact test. † Two-sided t test.

Table 2 shows the efficacy results for between-group change (BTX-A vs control) from baseline comparisons for the outcome measures (NDI, VAS, and SF-36 bodily pain and mental health scales). A negative difference in NDI and VAS and a positive difference in SF-36 summary measures indicate a more favorable result in BTX-A change from baseline. Although we found no significant between-group differences for most of the main outcomes (see table 2), we used the 95% confidence interval (CI) and Cohen d to describe treatment effects. An absolute value of the Cohen d effect size less than 0.2 was considered small, whereas 0.2 to 0.5 and greater than 0.5 were considered as moderate and large, respectively. A P value less than .025 was considered significant.

Table 2. Efficacy Outcomes of the BTX-A and Control Groups

	Follow-Up	NDI	VAS		SF-36
			Average Pain	Maximum Pain	Bodily Pain	Mental Health
	Week 2
	Mean	−0.50	−0.71	0.27	7.83	3.26
	95% CI	−4.67 to 3.67	−2.62 to 1.20	−1.66 to 2.21	−9.43 to 25.09	−10.05 to 16.58
	Cohen d	−0.11	−0.35	0.13	0.44	0.23
	P	.824⁎	.498⁎	.787⁎	.376⁎	.640⁎
	Month 1
	Mean	−0.53	−0.92	−1.07	10.01	16.40
	95% CI	−4.14 to 3.07	−2.75 to 0.92	−3.21 to 1.08	−3.57 to 23.58	4.29 to 28.51
	Cohen d	−0.12	−0.43	−0.40	0.59	1.10
	P	.765⁎	.320⁎	.337⁎	.144⁎	.010⁎
	Month 2
	Mean	−1.85	−1.19	−0.55	19.38	0.77
	95% CI	−7.16 to 3.46	−3.23 to 0.86	−3.22 to 2.12	2.86 to 35.91	−18.08 to 19.62
	Cohen d	−0.30	−0.51	−0.17	1.09	0.03
	P	.502⁎	.267⁎	.703⁎	.017⁎	.582†
	Month 3
	Mean	−0.31	−0.58	0.30	8.51	−2.53
	95% CI	−6.34 to 5.72	−2.66 to 1.50	−2.27 to 2.86	−7.53 to 24.55	−21.75 to 16.70
	Cohen d	−0.05	−0.25	0.10	0.49	−0.10
	P	.918⁎	.590⁎	.828⁎	.261⁎	.801⁎
	Month 4
	Mean	1.31	−1.11	0.09	16.22	9.79
	95% CI	−4.09 to 6.71	−2.81 to 0.60	−2.04 to 2.22	1.61 to 30.84	−12.68 to 32.26
	Cohen d	0.22	−0.57	0.04	0.98	0.36
	P	.619⁎	.219⁎	.936⁎	.033⁎	.408⁎
	Month 6
	Mean	2.75	−1.49	−1.00	6.90	16.23
	95% CI	−2.71 to 8.22	−3.57 to 0.59	−3.33 to 1.33	−5.92 to 19.72	−9.97 to 42.43
	Cohen d	0.45	−0.59	−0.36	0.45	0.50
	P	.276⁎	.179⁎	.405⁎	.273⁎	.265⁎

Abbreviation: CI, confidence interval.

⁎ Two-sided t test. † Two-sided Wilcoxon rank-sum test.

VAS for Pain 

Between-group (BTX-A vs control) comparisons did not reach statistical significance. However, as shown in table 2, BTX-A treatment effect as determined by the Cohen d showed moderate to large effect sizes across all the evaluation periods from week 2 through month 6 in the average VAS scores. As shown in figure 2, the largest differences in mean changes between the BTX-A and control groups were seen at month 4 (BTX-A change from baseline, −2.09±1.93 vs control group change from baseline, −0.98±1.94; d=−.57) and at month 6 (BTX-A group change from baseline, −2.21±.2.42 vs control change from baseline, −0.72±2.65, d=−.59).

View Large Image Download to PowerPoint

Fig 2. Mean changes in pain scores on the VAS from baseline (BTX-A vs control). Negative change indicates improved outcome.

The SF-36 

We found statistically significant differences between the BTX-A and control groups only in the bodily pain and mental health scales of the SF-36 but not in the other scales nor in the summary measures. We found better outcomes in the BTX-A treatment group, compared with controls (fig 3), at months 2 and 4 for the bodily pain scale (2-sided, P=.009, P=.016, respectively). It should be noted that the BTX-A group showed improvements in the bodily pain scale compared with baseline across all the evaluation periods with the greatest improvement seen at the aforementioned months 2 and 4, in which the improvements reached statistical significance compared with the control group.

View Large Image Download to PowerPoint

Fig 3. Improvement in mean scores on the SF-36 bodily pain scale from baseline (BTX-A vs control). *Statistically significantly better outcome for the BTX-A group (P<.025).

In the mental health scale (fig 4), there was a statistically significant difference between the BTX-A and control groups only at the 1-month follow-up evaluation (2-sided, P=.005).

View Large Image Download to PowerPoint

Fig 4. Improvement in mean scores on the SF-36 mental health scale from baseline (BTX-A vs control). *Statistically significantly better outcome for the BTX-A group (P<.025).

Neck Disability Index 

The highest effect of BTX-A on neck disability was detected at month 2 (BTX-A group change from baseline, −4.0±6.3 vs control group change from baseline, −2.2±6.2; d=−.30). The difference between the groups’ mean changes from baseline is within the 95% CI (−7.16 to 3.46). The opposite effect on neck disability was detected at month 6, when the NDI score returned to the baseline value in the BTX-A group but we saw a strong reduction in neck disability in the control group (BTX-A group change from baseline, 0.1±7.7 vs control group change from baseline, −2.6±4.2; d=.45). The insufficient number of values provided at both time points (powers of the tests were 10% and 18%, respectively) did not permit us to detect statistical significance. Moreover, the 2-point difference on the 50-point NDI scale could hardly be regarded as a clinically significant result.

There was no serious adverse event noted during the study period. Likewise, there was no difference in the number of side effects between the BTX-A and control groups.

Study Limitations 

We realize that our pilot study had limited power and a limited population base. Further, a limitation of this study may be the injection paradigm used. Studies using a trigger point injections paradigm for administering BTX-A have reported variable success.20 It has been suggested that targeting the “belly” of the muscle implicated in the pain may result in improved efficacy.20, 21 In fact, when looking at studies reporting positive outcomes, the majority have used injection paradigms other than targeting trigger points.22 In addition, our choice to restrict the total dose to 200U, in 4 or fewer muscles, may have resulted in suboptimal efficacy in those patients in whom more than 4 muscles were implicated in the pain.

Another limitation to the generalizability of our data is that we could not control for the concurrent treatments that the subjects received during the study period. We agree with Abram’s recommendations23 on the need to develop better criteria that could define the conditions in which BTX-A injections would yield the most benefit for patients suffering from chronic pain. To achieve this goal, there is a need to conduct larger-scale, well-powered multicenter studies and to do so with better-defined inclusion and exclusion criteria, injection technique, therapeutic regimen, and dosing schedule and perhaps a crossover design. The results of this double-blind, placebo-controlled pilot study provide effect sizes that can be used to properly design and power future studies by our group and by others.