Abstract
Background
Occipital nerve (ON) injections with corticosteroids and/or local anesthetics have been employed for the acute and preventive treatment of migraine for decades. However, to date there is no randomized, placebo-controlled evidence to support the use of occipital nerve block (ONB) for the prevention of migraine.
Objective
The objective of this article is to determine the efficacy of ONB with local anesthetic and corticosteroid for the preventive treatment of migraine.
Participants and methods
Patients between 18 and 75 years old with ICHD-II-defined episodic (> 1 attack per week) or chronic migraine (modified ICHD-II as patients with > 10 days with consumption of acute medications were permitted into the study) were randomized to receive either 2.5 ml 0.5% bupivacaine plus 0.5 ml (20 mg) methylprednisolone over the ipsilateral (unilateral headache) or bilateral (bilateral headache) ON or 2.75 ml normal saline plus 0.25 ml 1% lidocaine without epinephrine (placebo). Patients completed a one-month headache diary prior to and after the double-blind injection. The primary outcome measure was defined as a 50% or greater reduction in the frequency of days with moderate or severe migraine headache in the four-week post-injection compared to the four-week pre-injection baseline period.
Results
Thirty-four patients received active and 35 patients received placebo treatment. Because of missing data, the full analysis of 33 patients in the active and 30 patients in the placebo group was analyzed for efficacy. In the active and placebo groups respectively, the mean frequency of at least moderate (mean 9.8 versus 9.5) and severe (3.6 versus 4.3) migraine days and acute medication days (7.9 versus 10.0) were not substantially different at baseline. The percentage of patients with at least a 50% reduction in the frequency of moderate or severe headache days was 30% for both groups (10/30 vs nine of 30, Δ 0.00, 95% CI –0.22 to 0.23).
Conclusions
Greater ONB does not reduce the frequency of moderate to severe migraine days in patients with episodic or chronic migraine compared to placebo.
The study was registered with ClinicalTrial.gov (NCT00915473).
Keywords
Introduction
Migraine is one of the most common neurological disorders with a lifetime incidence in women and men of 43% and 18%, respectively (1). Chronic migraine (2) has a prevalence of 1.4% to 2.2% in adults and is highly disabling (3,4). Although some patients benefit from a daily preventive medication, others continue to suffer from severe, frequent, debilitating headache and associated symptoms. Limited efficacy, poor compliance, intolerable side effects and drug-drug interactions may explain why preventive medications are used in only 28.3% and 44.8% of episodic and chronic migraine patients’, respectively (5).
The anatomical and functional convergence of sensory afferents from C2 and the trigeminal nerve within the trigeminal cervical complex has been well established (6,7). The greater occipital nerve (GON) is composed of sensory fibers that originate predominantly at the C2 spinal level with a cutaneous distribution covering the posterior head up to the vertex. GON block with corticosteroids and/or local anesthetics has been employed for the acute (8,9) and prophylactic treatment of migraine (10), cervicogenic headache (11) and cluster headache (12,13) for more than 50 years. However, there have been no randomized, placebo-controlled trials examining the effect of GON block in patients with migraine. The absence of proper placebo-controlled studies is a significant shortcoming and limits the interpretation of open-label studies and expert opinion.
A long-acting anesthetic and corticosteroid are often combined in GON (14–17) although anesthetic agents (10,18) or corticosteroids (19) have been used alone. One randomized study demonstrated no additional benefit of triamcinolone to local anesthetic when patients with transformed migraine received both GON block and trigger point injections (14). Although there was a reduction in the headache intensity pre- and 20 minutes post-injection in both groups, this benefit was not sustained at four weeks and no difference was noted between the local anesthetic alone and the combination of anesthetic and corticosteroid (14). This randomized trial was underpowered and it did not directly measure efficacy beyond four weeks.
A prospective study examined the benefit of GON block using 2% lidocaine and 80 mg methylprednisolone in various primary headache types (15). Complete response (no pain) occurred in nine out of 54 migraineurs while partial response (> 30% decrease in severity or frequency) occurred in 17 out of 54 migraineurs (15). The mean duration of complete and partial response was nine days and 61 days, respectively (15). In another prospective study, GON blocks were performed in 150 consecutive chronic migraine patients (16). The injection consisted of 10 ml of bupivacaine with 20 mg of triamcinolone into the suboccipital region. At the one-month follow-up time point, 52% of patients reported at least a 50% reduction in the headache days compared to the month prior to injection (16). In a retrospective review of 97 migraine patients who had either a single or repeated GON consisting of lidocaine and methylprednisolone, headache improvement was noted in 54% of patients for up to six months (17).
Factors that predict efficacy of treatment to GON block have been investigated in previous studies and are potentially helpful in clinical practice to determine which patients may be better responders. For example, tenderness over the GON was strongly associated with a better response (20,21). In another study, tenderness over the suboccipital region was not associated with relief, nor was the duration of daily headache, presence of constant head pain, acute medication overuse, or previous prophylactic medication use (16). Other expected predictive factors include description and location of the head pain. It is possible that the directional description of pain (e.g. imploding versus exploding) may predict the response to a subcutaneous injectable therapy and reflect an extracranial origin of pain through the activation of meningeal sensory afferents that innervate the periosteum through the calvarial sutures (22).
We performed the first randomized, placebo-controlled trial of occipital nerve blockade in patients with episodic and chronic migraine to determine the short-term preventive efficacy, and evaluate several potentially predictive factors.
Participants and methods
The study was approved by the Mayo Clinic Arizona Research Committee and Mayo Clinic Institutional Review Board. All patients provided an informed and signed consent prior to randomization. The study was registered with ClinicalTrial.gov (NCT00915473). The study was sponsored by a Center for Translational Science Activities (CTSA) grant from Mayo Clinic.
Patients were recruited sequentially. Allocation of patients to active and placebo treatment was randomized in a one to one ratio. The randomized treatment allocation schedule was performed by a computer random number generator. The randomized treatment allocation schedule had seven blocks. Ten patients were in each block totaling 70 patients. Allocation was concealed by storing the treatment allocation schedule in the Mayo Clinic pharmacy.
Participants
The inclusion criteria were: (i) age 18–75 years, (ii) International Classification of Headache Disorders, second edition (ICHD-II)-defined episodic or chronic migraine, (iii) at least one migraine attack per week.
The exclusion criteria were: (i) individuals with continuous headache (no headache-free periods), (ii) individuals using maintenance opioid medication, (iii) individuals who had started a medication with placebo-controlled evidence for migraine-preventive efficacy within the past two months, (iv) known hypersensitivity or allergic reaction to any of the study ingredients (lidocaine, bupivacaine, any local anesthetics, or corticosteroids), (vi) use of any investigational medication within 90 days of the initial screening visit and/or concurrent enrollment in an investigational study, (vi) injection site infection or systemic infection at the injection visit or fever at time of injection, (vii) presence of cranial bone defect, (viii) individuals with chronic cluster headache, new daily persistent headache, hemicrania continua, or chronic tension-type headache, (ix) individuals with a history of an unstable medical condition (e.g. cardiovascular, hepatic, renal, endocrine) that may impair reliable participation in the study or necessitate the use of medications not permitted in this study, (x) individuals with a history (within the past six months) of a major psychiatric disorder that in the opinion of the investigator may preclude the individual from completing the requirements of the study, (xi) females who are pregnant or nursing, (xii) individuals with a history of drug or alcohol abuse within the past two years, (xiii) individuals with a history of poor compliance with past drug therapies, as judged by the investigator.
Procedure and intervention
Patients were equally randomized to receive either 2.5 ml 0.5% bupivacaine plus 0.5 ml 20 mg methylprednisolone over the ipsilateral (unilateral headache) or bilateral (bilateral headache) occipital nerve or 2.75 ml normal saline plus 0.25 ml 1% lidocaine without epinephrine (placebo). The procedure and intervention with regards to the type of medication, dosages, and site of injection were designed based on our experience. The GON injection site was at the medial third of the distance between the occipital protuberance and the mastoid process. Patients either received single or bilateral injection of the GON depending on whether their headache was unilateral or bilateral. The greater occipital nerve was the only injection site and all injectors had extensive training for the injection technique by the same investigator (DWD) and therefore delivered the injections consistently using the same technique. A 25-gauge, 7/8-inch needle was used. There were three patient visits in this study: screening, injection (post-baseline) and 28-day post-injection. In an effort to ensure adequate blinding, 0.25 ml of short-acting 1% lidocaine without epinephrine was used as the placebo arm and was designed to create numbness over the dermatomal region of the GON for blinding but at a minimal dose and duration to presumably avoid any impact on headache. In order to ensure adequate blinding of the investigator providing the injection, each syringe and needle hub was covered with opaque tape by the pharmacy. A total of three investigators provided injections (two physicians and one physician assistant). The blinded investigator who evaluated the study participant four weeks after injection was not necessarily the investigator who provided the injection.
Baseline characteristics collected included age, sex, body mass index (BMI), age at onset of migraine, location of headache (frontal, ocular, parietal, temporal, neck), tenderness of the GON to palpation (0 - to 10-point severity scale), directionality of headache (exploding, imploding, ocular), presence of migraine with aura, use of prophylactic analgesic (no, yes), use and frequency of consumption of acute analgesic medications (acetaminophen, butalbital, nonsteroidal anti-inflammatory drug (NSAID), opioid, triptan, other), and bilateral injection (no, yes). The directionality of pain was determined clinically through a detailed history, standardized descriptors of each type, and illustrations to help visualization. Headache measures were the number of calendar days with mild (one to three/10), moderate (four to six/10), and severe (seven to 10/10) headache (of any type), the number of calendar days with migraine headache, the number of hours with moderate or severe migraine headache, and the number of calendar days in which an acute medication was used. All participants received headache diary training and used the Mayo Clinic Headache Diary (paper version). Patients were asked open-ended questions regarding any adverse events. Adverse events were coded using the World Health Organization Adverse Reactions Terminology (WHOART) dictionary of preferred terminology and recorded with onset or worsening during the follow-up period.
Information from the headache diary was transcribed and calculated by the study coordinator and verified by a clinical studies unit volunteer with training at the data processing unit at Mayo Clinic. Prophylactic medication use was stable for two months prior to the study and could not be changed during the study.
Outcome measures
The primary outcome measure was defined as at least a 50% reduction in the frequency of moderate or severe migraine headache days in the active versus placebo group during the 28-day post-injection period compared to the 28-day baseline period. The frequency of days with moderate or severe migraine headache days was chosen as the primary measure based on International Headache Society Clinical Trials Subcommittee guidelines (23).
Secondary outcome measures were the reduction in mean frequency of all migraine headache days, mean number of hours with moderate or severe migraine headache, the mean number of days with acute medication consumption, percentage of patients with moderate to severe migraine headache ≥ 15 days, and percentage of patients with acute medication use ≥ 15 days.
Statistical analysis
Sample size calculations were based on 80% power (alpha = 0.05) if the percentage of responders in the target population were 40% with active treatment and 10% with placebo group. The assumptions regarding the estimated between group difference was based on the effect size from a placebo-controlled study of ONB study in patients with cluster headache (10).
The analysis included all randomized patients who provided follow-up data. Headache days per four weeks was calculated using: 28 × headache days/(diary end date – diary start date + 1). The primary outcome was also assessed for all treated patients.
The adequacy of the sample size is addressed by reporting confidence intervals (CIs) for the difference between treatments, and also by reporting Bayes posterior probabilities of differences. Analysis assuming that individuals who did not return were not responders was also performed. The upper confidence limit for the difference between groups excludes differences larger than 25 percentage points regardless of which sample is used.
The percentage of responders in the active group was compared to the placebo group using the Pearson chi-square test. The mean migraine headache frequency, migraine duration (hours) or acute medication frequency in the active group was compared to the placebo group using the two-sample t-test. Percentage of patients with moderate to severe migraine headache days ≥ 15 days and percentage of patients with acute medication use ≥ 15 days were compared using the Pearson chi-square test. The Fisher exact test was used instead of the Pearson chi-square test if the minimum expected cell count was less than five. The incidence of adverse events was assessed by using the Fisher exact test. Adjusted comparisons were made by using multiple logistic regression. Calculations were performed using SAS software version 9.3. Data were recorded on paper case report forms and typed into SAS data files by using double-entry and validation.
Results
Demographics
A total of 97 eligible patients consented to participate (Figure 1, Consolidated Standards of Reporting Trials (CONSORT) flow diagram). Patients were recruited by referral from multiple neurologists in the division of Headache at Mayo Clinic and the number of patients assessed for eligibility, excluded from participation, or declined consent were not recorded. Patients were recruited from June 2009 to October 2012. Thirty-five patients were allocated to each treatment group. One patient in the active group withdrew consent and did not receive the assigned treatment. Thirty-four patients received active treatment and 35 patients received placebo treatment. One patient had missing follow-up efficacy data in the active group, and five patients had missing follow-up efficacy data in the placebo group. Therefore, 33 patients in the active group, and 30 patients in the placebo group, were analyzed for efficacy.
Participant flow.
Baseline.
BMI: body mass index; NSAID: nonsteroidal anti-inflammatory drug.
(a) Baseline, all randomized.
(b) Baseline, all treated.
The baseline frequency of days with moderate or severe migraine headache ranged from one to 24 days per 28-day period, with a mean of 9.6 days (SD 4.8). The mean baseline frequency of moderate or severe migraine headache differed between the active and placebo groups by less than 0.3 days per four weeks. The mean baseline frequency of migraine headache days of any severity was 12.7 (SD 5.3) for the active group and 13.0 (SD 6.7) for the placebo group. The mean baseline frequency of days with moderate or severe migraine headache in the missing data group that was randomized to placebo was 6.75 days (moderate migraine) and six days (severe migraine). All headaches days were migraine days.
Moderate to severe migraine headaches occurring ≥ 15 days per four weeks at baseline was 18% in the active group and 10% in the placebo group. Acute medication use ≥ 15 days per four weeks was 12% in the active group and 23% in the placebo group.
The side of the injection was similar between both groups with bilateral injections occurring in 79% in the active group and 77% in the placebo group.
Mean age differed by less than two years. Mean age was less than 5% younger in the placebo group, and the standard deviations were 11 and 16. Age differences were even smaller in the analysis of all treated patients. Adjusting for triptan use or opioid use did not increase the difference in response. No other variable differed by at least 10 percentage points.
Outcome data
Outcomes.
At least 50% reduction in frequency of days with at least moderate migraine. Primary outcome measure. p(δ > 0.10) = 20%, p(δ > 0.20) = 4%. CI: confidence interval.
Triptans were used by only 65% (20/31) of patients in the active group versus 80% (24/30) in the placebo group, and opioids were used by 23% (seven of 31) in the active group versus only 13% (four of 30) in the placebo group. Adjustment for triptan use or opioid use did not increase the difference in response rates. Counting participants who did not provide follow-up data as non-responders did not alter the conclusion.
There were no significant differences in any of the secondary outcome measures. The CIs and p values for the mean frequency of days with a migraine headache of any intensity, the mean frequency of at least moderate migraine headache, the mean frequency of days with severe migraine headache, the mean number of hours with moderate or severe migraine headache, the mean number of days with acute medication use, percentage of patients with moderate to severe migraine headache ≥ 15 days and the percentage of patients with acute medication use ≥ 15 days are summarized in Table 2. The reduction in migraine headache days between the active and placebo group did not differ based on intensity of the migraine headache.
Adverse events.
(a) Total adverse events.
Discussion
This double-blind, randomized, placebo-controlled study demonstrated that GON block with corticosteroid and local anesthetic does not reduce the short-term (28-day) frequency of moderate/severe migraine headaches days, migraine days of any pain intensity, migraine hours, or acute medication use in patients with episodic and chronic migraine when compared to placebo. There was no correlation between response and the diagnosis of episodic or chronic migraine, though the study was not powered to detect a difference in these subgroups.
These results stand in contrast to patient outcomes in open-label studies or case series in addition to the perception of benefit among clinicians. There may be multiple reasons for these results. The first is that there is truly no short-term preventive effect of ONB with corticosteroid and local anesthetic, reinforcing the well-recognized need for randomized, placebo-controlled studies to evaluate treatments that are used based on anecdote and clinical experience. It is possible, however, that the small amount of lidocaine used in the placebo group to protect the blind could have had a therapeutic effect by blocking afferent input along C2 sensory afferents, although only one patient in the placebo group reported hypoesthesia (prolonged period of numbness) over the GON. However, the response rate in the placebo group at one month is not substantially different from placebo response rates demonstrated in other preventive migraine trials (24), and it is well recognized that placebo response rates increase in studies that employ a more invasive treatment modality (injection versus tablet) (25,26). Moreover, a double-blind crossover trial comparing bupivacaine alone to normal saline did not demonstrate a difference in the number and duration of headache attacks (18). It should also be noted that the addition of a significant dose of steroid in the active group, which has been shown to be effective for the short-term prevention of cluster headache, failed to separate the active from placebo group.
It is difficult to perform placebo-controlled studies using anesthetic or needling owing to the impact of the effect of needling independent of the pharmacological effects of the medication, and the unblinding that can occur with hypoesthesia produced by local anesthetic. One option to consider in the future is to inject normal saline for the placebo group after the administration of topical lidocaine in both groups so as to produce hypoesthesia in both groups.
The innervation of periosteum from meningeal sensory afferents that traverse calvarial sutures has led to speculation that the origin of pain during some migraine attacks and/or some patients may have an extracranial pathophysiology. It has been suggested that such patients may describe their pain as either imploding or ocular, and may preferentially respond to injectable treatments that have the potential to access and modulate the activity of meningeal sensory afferents though subcutaneous injections (22). This prompted us to evaluate for a possible relationship between response to ONB and the directionality of pain. However, there was no difference in the response rate in this study between the active and placebo group for the different directional patterns of headache (exploding, imploding or ocular).
The study did not address the minimum volume of injection required for a therapeutic effect, nor whether a higher volume of either or both the local anesthetic and corticosteroid will yield higher efficacy. In one prospective study, a higher volume (10 ml) of bupivacaine was used with a therapeutic benefit (16). The percentage of patients with > 50% reduction in “functionally incapacitating headache days” was 54% (16); however, this consecutive study of 150 patients included only chronic migraine sufferers. In a separate prospective study by Afridi et al., the GON block contained 2% lidocaine and 80 mg methylprednisolone and showed >30% reduction in severity and frequency of the migraine (15). Dose-response effect was not performed in this study although standard protocols for the GON block include 20 mg of methylprednisolone or alternative steroid of equivalence (27). In our study, 79% and 77% of the active and placebo groups received bilateral GON blocks with 20 mg methylprednisolone on each side. There was no significant difference in the response between the two groups regardless of whether they received single or bilateral GON blocks; however, the study was not powered to explore this difference.
The study did not evaluate the onset or duration of benefit of the GON block. While previous studies have demonstrated the benefit of GON block may occur within five to 20 minutes of the injection (8,9), this study did not evaluate acute response to injection, and not all patients were experiencing headache pain at the time of injection. Consequently, a placebo-controlled trial evaluating GON block for acute relief of migraine pain is warranted. In addition, future randomized, controlled studies evaluating the effect of corticosteroid alone to placebo, which have been demonstrated to be effective for the short-term prevention of cluster headache, are warranted and would be easier to blind than a study that employs the use of a local anesthetic. Future studies could explore monthly serial nerve blocks over three to four cycles as this study involved only GON block at a single time frame and not repetitive injections.
In conclusion, this double-blind, randomized, placebo-controlled study suggests that a single GON block using corticosteroid and local anesthetic does not provide a short-term preventive effect, as measured by a significant reduction in moderate/severe migraine days in the one month following injection.
Clinical implications
Occipital nerve blocks are commonly used therapies in treating headaches including migraine and chronic migraine. Greater occipital nerve block using anesthetic and corticosteroid does not reduce the frequency of moderate to severe migraine days in chronic migraine.
Footnotes
Funding
This work was supported by CTSA grant number UL1 TR000135 from the National Center for Advancing Translational Sciences (NCATS), a component of the National Institutes of Health (NIH). Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NIH.
Conflict of interest
None declared.