Exploding vs. imploding headache in migraine prophylaxis withBotulinum Toxin A

Moshe Jakubowski, Ph.D.1, Peter J. McAllister, M.D.3, Zahid H. Bajwa, M.D.1, Thomas N.Ward, M.D.4, Patty Smith, R.N.1, and Rami Burstein, Ph.D.1,2

1 Departments of Anesthesia and Critical Care, Beth Israel Deaconess Medical Center, Harvard MedicalSchool, Boston, MA 02115

2 Program in Neuroscience, Harvard Medical School, Boston, MA 02115

3 Associate Neurologists of Southern Connecticut, Fairfield, CT

4 Dartmouth-Hitchcock Medical Center, Hanover, NH

AbstractMigraine headache is routinely managed using medications that abort attacks as they occur. Analternative approach to migraine management is based on prophylactic medications that reduce attackfrequency. One approach has been based on local intramuscular injections of Botulinum Toxin TypeA (BTX-A). Here, we explored for neurological markers that might distinguish migraine patientswho benefit from BTX-A treatment (100 units divided into 21 injections sites across pericranial andneck muscles). Responders and non-responders to BTX-A treatment were compared prospectively(n=27) and retrospectively (n=36) for a host of neurological symptoms associated with their migraine.Data pooled from all 63 patients are summarized below. The number of migraine days per monthdropped from 16.0±1.7 before BTX-A to 0.8±0.3 after BTX-A (down 95.3±1.0%) in 39 responders,and remained unchanged (11.3±1.9 vs. 11.7±1.8) in 24 non-responders. The prevalence of aura,photophobia, phonophobia, osmophobia, nausea, and throbbing were similar between respondersand non-responders. However, the two groups offered different accounts of their pain. Among non-responders, 92% described a buildup of pressure inside their head (exploding headache). Amongresponders, 74% perceived their head to be crushed, clamped or stubbed by external forces (implodingheadache), and 13% attested to an eye-popping pain (ocular headache). The finding that explodingheadache was impervious to extracranial BTX-A injections is consistent with the prevailing viewthat migraine pain is mediated by intracranial innervation. The amenability of imploding and ocularheadaches to BTX-A treatment suggests that these types of migraine pain involve extracranialinnervation as well.

IntroductionMigraine is a throbbing headache routinely managed using medications that abort attacks asthey occur. An alternative approach to migraine management is based on prophylacticmedications that reduce attack frequency. One approach has been based on local intramuscularinjections of Botulinum Toxin Type A (BTX-A), a neurotoxin that causes transient muscleparalysis (Black and Dolly 1986a; Black and Dolly 1986b; Dolly 2003; Foran et al. 2003;

Correspondence and reprint requests: Rami Burstein, Ph.D., Department of Anesthesia and Critical Care, Harvard Institutes of Medicine,Room 830, 77 Avenue Louis Pasteur, Boston, MA 02115. Tel: 617 667 0806, Fax: 617 975 5329, email: rburstei@caregroup.harvard.edu.Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resultingproof before it is published in its final citable form. Please note that during the production process errors may be discovered which couldaffect the content, and all legal disclaimers that apply to the journal pertain.

NIH Public AccessAuthor ManuscriptPain. Author manuscript; available in PMC 2007 December 5.

Published in final edited form as:Pain. 2006 December 5; 125(3): 286–295.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Lawrence and Dolly 2002a; Lawrence and Dolly 2002b; Simpson 1981). Clinical studies onthe efficacy of BTX-A treatment for migraine prophylaxis have generally failed to demonstrateits superiority over placebo using analysis of pooled data from all participants (Binder et al.2000; Dodick et al. 2005; Evers et al. 2004; Ondo et al. 2004; Silberstein et al. 2000; Tepperet al. 2004). However, many clinicians are well aware that in some patients BTX-A treatmentdramatically reduces attack frequency, whereas in others it produces no benefit. The goal ofthe present study was to prospectively search for a neurological marker that would differentiatebetween migraine patients who respond to BTX-A treatment and those individuals who do notbenefit from such treatment.

BTX-A therapy has been used in a variety of disorders associated with painful muscles spasm(Berweck and Heinen 2004; Chen and Swope 2003; Comella et al. 2000; Defazio and Livrea2004; Jankovic 2004; Marjama-Lyons and Koller 2000; Tilton 2004) as well as migraine(Binder et al. 2000; Blumenfeld 2003; Dodick et al. 2005; Evers et al. 2004; Tepper et al.2004; Troost 2004). Because migraine attacks are frequently associated with muscle tenderness(Jensen et al. 1988; Tfelt-Hansen et al. 1981), it is generally believed that intramuscular BTX-A may prevent abnormal sensory signals in the affected muscle from arriving to the centralnervous system. If abnormal muscle physiology can trigger migraine, one would predict thatBTX-A treatment should work prophylactically only in patients whose migraine attacksdevelop at the heels of episodic or chronic muscle tenderness. According to another view, BTX-A may block sensitization of nociceptive neurons in the dorsal horn (Aoki 2005; Cui et al.2004; Oshinsky et al. 2004). Considering that central sensitization during migraine ismanifested as cutaneous allodynia (Burstein et al. 2000a; Burstein et al. 1998; Burstein et al.2000b), one would expect BTX-A to block the development of allodynia during migraine inBTX-A responders who were allodynic before treatment.

The results of the prospective study indicated that BTX-A responders did not differ from non-responders in terms of muscle tenderness, cutaneous allodynia or any other known feature ofmigraine that we examined. Much to our surprise, however, responders and non-responderswere sharply divided in the way they described their migraine headache. To test the validityof this finding, we conducted a triple-blind retrospective study in which patients from two otherclinics were interviewed after they had already undergone routine prophylactic BTX-Atreatment.

MethodsPatient selection

The studies herein were carried out in compliance with 1983 revision of the 1975 HelsinkiDeclaration, and according to the ethical standards of the Beth Israel Deaconess Medical Center(BIDMC) Committee on Clinical Investigation on Human Experimentation, including aninformed consent form. Included in the study were patients who met the criteria of theInternational Headache Classification Committee for Migraine. (Headache-classification-committee-of-the-International-Headache-Society 1988). Excluded were patients with chronicdaily headache, tension-type headache, peripheral nervous system injuries, and patients usingopioids for reasons other than migraine. With the exception of 2 patients in the retrospectivestudy, no patient was taking any prophylactic drugs for a period of 3–6 months prior and 3–4month after BTX-A injections.

Prospective studySixty migraine patients (25–57 year-old) were recruited for the prospective study at the BIDMCpain clinic, Boston, MA. Of those, 18 patients dropped out of the study without receiving BTX-

Jakubowski et al. Page 2

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

A treatment. The remaining 42 patients were designated to visit the clinic 5 times as describedbelow.

Visit 1—This visit took place when the patients were migraine-free for at least 2 days. Theywere interviewed for migraine history and accompanied symptoms, and tested for their baselineskin sensitivity in the absence of migraine.

The interview included the following items: family history; migraine history (age of onset andnumber of years with migraine); past and present medication usage; attack frequency/durationand headache intensity; associated neurological symptoms (aura, nausea, vomiting,photophobia, phonophobia, osmophobia). Patients were asked if they experience neck muscletenderness during migraine, and if so, whether it starts before and after the onset of theheadache. In addition, patients were asked to give a detailed account of their migraine headache,including the affected area (e.g., ocular, frontal, parietal, occipital, unilateral, bilateral) and thecharacteristics of the pain (e.g., pressuring, throbbing). A significant portion of the interviewwas devoted to a novel set of questions aimed at delineating the directionality of the pain. Thesequestions were constructed on the basis of subjective testimonies that we have collected frommigraine patients over the years. Collectively, these subjective testimonies suggested that somepatients perceived a painful buildup of pressure inside the head (e.g., “my head is exploding”)whereas others perceived a painful assault on the outside of the head (e.g., “I have a crushingheadache”). Initially, patients were prompted to provide a free-form description of how theyfelt their headache using words or metaphors of their choosing. Next, they were specificallyasked whether their pain felt like a pressure inside the head pointing outward (i.e., inside tooutside direction) or a crushing/piercing pain outside the skull pointing inward (i.e., outside toinside direction). Finally, they were asked to chose from the following options: (a) it feels asif my head is exploding, (b) it feels as if my head is being crushed or pierced on the outside.

After the interview, baseline skin sensitivity was measured using quantitative sensory testing(QST) according to the method of limits (Fruhstorfer et al. 1976; Yarnitsky 1997) as describedbefore (Burstein et al. 2000b). Pain thresholds to thermal and mechanical stimuli of the skinwere measured at the temple on the usual side of the headache. Patients were supine in a dimly-lit room and were unable to observe the source of the stimulation at use.

At the end of the visit patients were asked to keep records of attack frequency, attack durationand pain intensity for the duration of the study.

Visit 2—This visit took place within 3 month of visit 1. Patients arrived at the clinic during amigraine attack, 4 h after the onset of the pain. The time of onset was defined as the time atwhich the headache begun to throb spontaneously, or intensify in response to physicalactivities. Patients were asked to rate their headache intensity on a visual analog scale rangingfrom 0 (no pain) to 10 (worst pain). Their skin sensitivity was measured again as per visit 1,only this time it was measured during an attack. At the end of the visit, patients were offeredthe choice of receiving rescue medication to terminate the attack.

Visit 3—This visit took place within 7 days of visit 2 (typically in the absence of migraine)in order to administer BTX-A. A total of 100 units of BTX-A (Allergan) dissolved in 4 mlsaline were divide into 21 intramuscular injections using a 30-gauge, 0.5-inch-long needle.The following muscles were injected: frontal/glabellar (7 sites, 5 units each); occipital (1injection/side, 10 units each); temporalis (2 injections/side, 3.25 units each); trapezius (2injections/side, 5 units each); semispinalis (1 injection/side, 2.5 units each); splenius capitis (1injection/side, 2.5 units each). No placebo group was employed as we were only interested infinding a marker that would distinguish responders from non-responders, rather than evaluatingthe efficacy of BTX-A.

Jakubowski et al. Page 3

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Visit 4—This visit took place between week 4 and week 12 after visit 3. As per visit 2, patientsarrived at the clinic 4 h after the onset of migraine headache, and their skin sensitivity andheadache intensity were recorded again. Patients who became migraine-free between weeks 4and 12 after treatment were asked to wait until they underwent another attack, once the effectof BTX-A wore off.

Visit 5—This visit took place 12 weeks after visit 3. Patients provided informationsummarizing their experience during weeks 4–12 post-treatment regarding attack frequency,attack duration and pain intensity during an attack.

Allodynia—Whether or not a patient was allodynic during a migraine attack was determinedbased on the pain thresholds at visits 2 and 4 (migraine), with reference to the correspondingvalues from visit 1 (baseline). In healthy subjects, pain thresholds for heat, cold and mechanicalskin stimuli were shown to range between 42–47 °C (Lindblom 1994), 5–18 °C (Harrison andDavis 1999; Lindblom 1994), and 75–281 g (Strigo et al. 2000), respectively. Using somewhatmore stringent criteria, we considered a patient to be allodynic if the pain threshold was below40 °C for heat, above 20 °C for cold, and below 30 g for skin indentation. Meeting the criterionfor any one modality was sufficient to determine that the patient was allodynic at the time oftesting.

Retrospective studyThirty six migraine patients (25–70 year-old) were interviewed retrospectively at their doctor’soffice in two headache clinics unaffiliated with BIDMC: 26 patients were from the AssociateNeurologists of Southern Connecticut, Fairfield, CT; 10 patients were from the Dartmouth-Hitchcock Medical Center, Hanover, NH. These patients received BTX-A injections for thepurpose of migraine prophylaxis before (and independent of) the prospective study at BIDMC.

The interview was conducted by the same investigator and in the same way described in theprospective study. To minimize any bias, the interview was carried out in a triple-blind fashion:the interviewer remained uninformed about the outcome of the treatment until the very end ofthe process, while the tending physician and the patient were both oblivious to the findings ofthe prospective study as well as which item might be associated with the treatment outcome.Our goal was to recruit as many non-responders as responders, but we found that fewer non-responders were motivated to participate in the study.

Statistical analysisData analysis included only patients who were clear-cut responders or non-responders. Inresponders, the number of migraine days per month (attack frequency × attack duration) hadto drop 80% or more below the pre-BTX-A level. In non-responders, the number of migrainedays per month either remained unchanged or dropped by no more than 33%. A third subsetof 15 patients who were unsure whether they benefited from BTX-A treatment were excludedfrom data analysis, albeit the type of headache (see Results).

Data analysis was performed using non-parametric statistics (Siegel 1956). Ordinalmeasurements (e.g., age, attack frequency, attack duration, pain intensity) were comparedbetween responders and non-responders using Mann-Whitney U test. Repeated ordinalmeasurements (i.e., before and after treatment) were compared within each group usingWilcoxon matched-pairs signed-rank test. Nominal data were analyzed using either simple χtest or Fisher exact probability test, depending on sample size and expected frequencies (Siegel1956). The level of significance was set at 0.05

Jakubowski et al. Page 4

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

ResultsProspective study

During the period between weeks 4 and 12 after BTX-A treatment, 27 of the 42 patients whoparticipated in the study were clear responders or non-responders (Fig. 1): non-responders(n=13) experienced no improvement, and responders (n=14) reported significant drops inattack frequency (down 88.4±3.1%, p<0.002), attack duration (down 82.0±7.6%, p<0.002),and headache intensity (down 53.4±11.9%, p<0.005). These 27 patients (25 women, 2 man)were 41.9±1.7 years of age (mean±SEM) with a history of 24.3±2.1 years of migraine (onsetat 17.7±1.7 years of age). According to their migraine history documented prior to BTX-Atreatment (visits 1–3), they experienced 9.0±1.3 attacks/month, each lasting 2.0±0.3 days, withheadache intensity rated at 8.3±0.3 on a 0–10 VAS.

At the end of the study, information collected in visits 1, 2 and 4 was analyzed in search forany difference between responders and non-responders. The number of years with migraine(mean±SEM) was similar between the two groups (23.6±3.2 vs. 24.9±2.9, p>0.9). Analysis ofthe following migraine symptoms indicated they were as prevalent in responders as in non-responders (p>0.6): aura (50 vs. 38%, respectively); photophobia (100 vs. 92%); phonophobia(93 vs. 92%); osmophobia (86 vs. 69%); nausea (86 vs. 77%); vomiting (57 vs. 46%); throbbing(93 vs. 92%); or unilateral headache (57 vs. 69%). Analysis of the incidence of neck-musclestenderness showed no significant difference between responders and non-responders (79 vs.54%; p>0.4). Analysis of medication overuse, as defined by Silberstein et al (Silberstein et al.2005), yielded no difference between the two groups of patients (36 vs. 38%). Analysis of theincidence of cutaneous allodynia during migraine (Table 1) showed no significant differencebetween responders and non-responders before BTX-A treatment (86 vs. 62%, p>0.3).Moreover, between weeks 4 and 12 post-treatment, all allodynic patients but one (a non-responder) continued to exhibit allodynia during migraine.

Our last set of data analysis dealt with the way patients described their headache and itsdirectionality. Patients were specifically asked to make a distinction between the universalperception of throbbing, which was typically pointed and always emanating outward, and theirperception of pain in the larger area affected by the headache as described in Methods. Basedon the testimonies, we came to recognize 2 classes of migraine pain, which we denoted‘exploding headache’ and ‘imploding headache’. Patients with exploding headache perceiveda buildup of pressure inside their head as if the skull was about to split open (Fig. 2A). Patientswith imploding headache perceived their skull to be assaulted by external forces typicallyreferred as crushing, clamping or stubbing (Fig. 2B, 2C). We also encountered a third subgroupof patients whose testimonies did not fit the description of either exploding or implodingheadache. These patients attested to an eye-popping pain which we denoted ‘ocularheadache’ (Fig. 2D). Typical quotes for each of the 3 categories are summarized in Fig. 2.

Much to our astonishment, responders and non-responders offered very different accounts oftheir pain: Of 14 responders, 79% had imploding headache and 14% ocular headache;conversely, all 13 non-responders had exploding headache.

Of the 42 patients who participated in the study, 15 were unsure as to whether they benefitedfrom BTX-A treatment. Of those, 4 had exploding headache, 4 had imploding headache, and7 testified to having both types of headache.

Retrospective studyPatients included in this study (35 women, 1 man) were 48.3±1.7 years of age (mean±SEM)with a history of 27.3±2.3 years of migraine (onset at 20.9±2.3 years of age). They testifiedthat their migraine history prior to BTX-A treatment consisted of 13.3±1.7 attacks/month, each

Jakubowski et al. Page 5

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

lasting 1.3±0.1 days, with headache intensity rated at 8.9±0.2 on a 0–10 VAS. After BTX-Atreatment, the patients fell into two categories (Fig. 3): non-responders (n=11), whoexperienced no improvement, and responders (n=25), who reported significant drops in attackfrequency (down 86.2±3.2%, p<0.001), attack duration (down 65.6±8.3%, p<0.001), andheadache intensity (down 59.4±7.7%, p<0.001).

In agreement with the prospective study, there was no significant difference in the number ofyears with migraine (mean±SEM) between responders and non-responders (29.3±3.0 vs. 22.8±3.0, p>0.1). Likewise, every migraine symptom analyzed was as prevalent in responders asin non-responders (p>0.4): aura (36 vs. 55%, respectively); photophobia (96 vs. 100%);phonophobia (96 vs. 73%); osmophobia (76 vs. 82%); nausea (92 vs. 100%); vomiting (68vs. 64%); throbbing (88 vs. 91%), and unilateral headache (56 vs. 73%). Nor was there asignificant difference between the two groups in the incidence of neck-muscles tenderness (76vs. 91%; p>0.9), or medication overuse (36 vs. 18%; p>0.4).

As with the prospective study, the only factor that distinguished responders from non-responders was the way they described the headache and its directionality. Of the 25 responders,72% had imploding headache and 12% had ocular headache, while only 16% had explodingheadache. Conversely, of the 11 non-responders, 82% had exploding headache, and only 18%had imploding headache.

Summary of data combined from both studiesSince the findings of the prospective and retrospective studies were not significantly differentfrom each other, we pooled the data for the following analyses.

Response to BTX-A—Dissection of the data on the basis of the type of headache (Table 2)indicated that non-responders constituted 81% of the patients with exploding headache (n=27),whereas responders constituted 94% of the patients with imploding headache (n=31) and 100%of those with ocular headache (n=5). Fig. 4 shows that the mean±SEM number of days ofmigraine per month (attack frequency × attack duration) dropped a whopping 90.0±0.05% inimploding headache patients (p<0.0001), 96.5±0.03% in ocular headache patients (p<0.05) butremained unchanged in exploding headache patients (p>0.9).

Episodic vs. chronic migraine—Of the 63 patients, 36 had episodic migraine (1–14 daysof migraine per month) and 27 had chronic migraine (15–30 days of migraine per month). Ofthe episodic migraine patients, 58% were responders and 42% were non-responders. Of thechronic migraine patients, 67% were responders and 33% were non-responders. Thedistribution of responders and non-responders between episodic and chronic migraine patientsproved non-significant (χ2=0.84; p>0.5).

Neck-muscle tenderness—Contrary to our working hypothesis, the data in Table 3demonstrate that the distribution of responders and non-responders was similar among patientswho developed muscle tenderness before (n=21) or after (n=22) the onset of a migraine attack,as well as among patients who had no muscle tenderness at all (n=15).

Addendum: retrospective analysis of migraine patients who received placebo injectionsAfter finding a relationship between headache type and outcome of BTX-A treatment (initiallypresented at the International Headache Conference in Kyoto, Japan, October 2005), we madeit a general practice in the 3 participating clinics to inquire patients about their headache type,regardless of whether they had received any BTX-A treatment. We then asked Allergan toprovide us with code names of patients that had received placebo injections in our 3 clinics inan earlier phase 2, double-blind, randomized, placebo-controlled study that evaluated BTX-A

Jakubowski et al. Page 6

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

injections as a preventative treatment for migraine headache. A summary of the placebotreatment in the earlier Allergan study was provided to the attending physicians and analyzedagainst the type of migraine headache in individual patients. Twelve placebo subjects werethus identified, and none of them was exposed to any BTX-A treatment. As shown in Fig. 5,the mean±SEM number of days of migraine per month remained unchanged in patients withimploding/ocular headache (n=6) or exploding headache (n=6). The maximal drop in numberof migraine days per month after placebo injections was 21, 24, 29 and 45% in 4 patients (1with imploding, 2 with exploding, 1 with ocular headache, respectively). Among the remaining8 patients, the drop in migraine days/month was either minimal (<10%) or nil (4 withimploding, 4 with exploding headache).

DiscussionThe success rate of preventing migraine attacks prophylactically using BTX-A was associatedwith the type of headache individual patients had. Patients who testified to ‘imploding’ or‘ocular’ headache were likely to experience a drastic decline in the number of migraine daysper month between weeks 4 and 12 after a single treatment session. In contrast, patients whotestified to ‘exploding’ headache were unlikely to respond to BTX-A treatment. Thus, migraineperceived as exploding headache appears to be fundamentally different in pathophysiologyfrom migraine perceived as imploding or ocular headache, suggesting that the subjectiveperception of pain be incorporated into migraine diagnosis.

Considering that migraine prophylaxis may be associated with a placebo effect (Binder et al.2000; Dodick et al. 2005; Evers et al. 2004; Ondo et al. 2004), we examined whether theincidence of migraine could be selectively reduced by placebo injections in patients withimploding or ocular headache, but not in those with exploding headache. Retrospective analysisindicated that the mean number of migraine days per month remained unchanged after placeboinjections in 6 patients with imploding/ocular headache, as well as in 6 patients havingexploding headache. Patient-by-patient analysis indicated that 4 out of 12 patients (33%)experienced a 21–45% drop in migraine days/month. It should be noted, however, that thedefinition of response to BTX-A in the current study was much higher (>80% drop in migrainedays/month), suggesting that the selective response to BTX-A treatment in our patients withimploding/ocular headache was unlikely to be due to a placebo effect. Using a lower thresholdfor response to BTX-A (e.g., >50% rather than >80% drop in migraine days/month), as in otherstudies, would slikely increase the size of the placebo effect.

Migraine patients typically describe their headache using familiar terms such as throbbing,pounding, pressing or tightening (Olesen et al. 2005). In recent years, we have become awarethat some patients perceive their pain as pressure buildup inside their head (explodingheadache) while others perceive their pain to be inflicted from the outside (implodingheadache). Subjective accounts of exploding headaches often consisted of phrases such as:“My head is exploding”, or “my brain feels so swollen that there isn’t enough room inside”,or “I wish someone would drill a hole in my head to let the pressure out”. Subjective accountsof imploding headaches included phrases such as: “someone is tightening a vise around myhead and it is going to crush”, or “an elephant is sitting on my head”, or “someone is stubbinga knife in my head”. We must admit that this novel delineation of the type of headache accordingto its directionality may not be a simple task for either the interviewing physician or the patient.So far, we have employed a three-tier approach in the interview: first, allow the patients tovolunteer their own answer to the best of their recollection; second, ask the patients to choosebetween inward and outward headache; third, present a sample of common testimonies (Fig.2) and ask the patients which fits best their perception of migraine pain. To minimize inter-observer variability, it will first be necessary to optimize a standardized line of questioning

Jakubowski et al. Page 7

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

presented to the patient and, second, apply the standardized interview in different clinics inorder to validate the consistency of the diagnosis.

Although we do not understand the pathophysiological difference between exploding andimploding headaches, we have made it a routine to ask patients to describe their headache andits directionality. We found that the proportion of migraineurs with imploding headache was42 and 44% in the prospective and retrospective study, respectively. Exploding headache wasfound in 54% of the patients in the prospective study and 33% of the patients in the retrospectivestudy. While the relative prevalence of exploding and imploding headaches remains to bedetermined in the general population of migraineurs, our preliminary survey of 50 patients inBIDMC headache clinic suggests that 70% have exploding and 30% imploding or ocularmigraine headache.

A number of factors proved unrelated to the prophylactic action of BTX-A. Symptoms suchas aura, photophobia, phonophobia, osmophobia, nausea, vomiting or throbbing, were equallyprevalent among responders and non-responders. Also unrelated to the prophylactic action ofBTX-A is the frequency of migraine attacks: among responders 54% of the patients were‘episodic’ and 46% were ‘chronic”. Given that medication overuse often renders patientsrefractory to common migraine therapy, the finding that medication overuse was equallyprevalent between responders and non-responders suggests that medication overuse does notinterfere with the prophylactic action of BTX-A.

The finding that migraine prophylaxis was unrelated to concurrent muscle tenderness suggeststhat the effect was independent of the inhibitory action of BTX-A at the level of theneuromuscular junction. The notion that migraine prophylaxis may be achieved through directinhibition at the level of the central nervous system did not gain support either. Theresponsiveness to BTX-A was not associated with any centrally-mediated migraine symptomssuch as aura and cutaneous allodynia. Considering that cutaneous allodynia is driven bysensitized central trigeminovascular neurons (Burstein et al. 2000b), it is worth noting thatBTX-A proved virtually ineffective in blocking allodynia during migraine in both respondersand non-responders.

Rather than acting centrally, BTX-A may exert its prophylactic action on migraine throughinhibition of peripheral sensory neurons (Aoki 2005). That BTX-A can inhibit sensory neuronsis consistent with evidence that the toxin inhibits release of substance P from embryonic dorsalroot ganglion neurons in vitro (Welch et al. 2000), or CGRP from trigeminal ganglion neuronsin vitro (Durham and Cady 2004), or glutamate from peripheral nociceptors terminating in thedorsal horn in vivo (Cui et al. 2004). Using quantitative sensory testing, however, Blersch etal (2002) showed that subcutaneous injection of BTX-A did not change pain thresholds to localcold, heat and electrical stimulation, and concluded that BTX-A may block pain throughchemodenervation or anti-inflammatory action, rather than by direct peripheral anti nociceptiveeffect.

Regardless of the exact peripheral mechanism of BTX-A action, at issue here is whether theperipheral action is exerted intracranially or extracranially. The finding that explodingheadache was impervious to extracranial BTX-A injections is consistent with the general viewthat migraine pain is mediated by intracranial nociceptors innervating the meninges andsinuses of the dura. Such a concept may also be tested in the future using patients whoseheadache stems from a clear intracranial pathophysiology. On the other hand, the amenabilityof imploding and ocular headaches to BTX-A treatment suggests that these types of migrainepain also involve extracranial sensory fibers adjacent to the injection sites. Involvement ofextracranial tissue in the pathophysiology of certain migraine patients has been proposed somefive decades ago (Selby and Lance 1960; Wolff et al. 1953). If so, it is tempting to speculate

Jakubowski et al. Page 8

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

that imploding headache involves activation of extracranial nociceptors that innervate scalptissues, such as bone and periostium.

The dramatic difference between responders and non-responders in the incidence of migraineafter BTX-A treatment suggests that previous studies that pooled data from all treated patientswithout such distinction (Binder et al. 2000; Dodick et al. 2005; Evers et al. 2004) may havemissed the full magnitude of the prophylactic effect of the toxin at the level of the individualpatient. Of the 42 subjects that participated in our prospective study, only 14 (33%) were clear-cut responders, while 13 (31%) were clear-cut non-responders. It remains to be determinedwhat types of headache are present in patients whose attack frequency after BTX-A droppedbetween 33 and 80% (‘partial responders’).

Finally, we would like to emphasize that the present study does not intend and cannot be usedto recommend the use of BTX-A as prophylactic migraine therapy. The only goal of this studywas to explore for neurological markers that might single out migraine patients who respondto BTX-A treatment from those who do not.

Acknowledgements

Supported by an unrestricted grant from Allergan, and by NIH grant NS051484 to Dr. Burstein. We thank Virginiaartist Lorraine Fink for the illustrations in FIG. 2. Dr. Jakubowski had full access to all of the data in the study andtakes responsibility for the integrity of the data and the accuracy of the data analysis.

ReferencesAoki KR. Review of a proposed mechanism for the antinociceptive action of botulinum toxin type A.

Neurotoxicology 2005;26:785–93. [PubMed: 16002144]Berweck S, Heinen F. Use of botulinum toxin in pediatric spasticity (cerebral palsy). Mov Disord

2004;19:S162–7. [PubMed: 15027070]Suppl 8Binder WJ, Brin MF, Blitzer A, Schoenrock LD, Pogoda JM. Botulinum toxin type A (BOTOX) for

treatment of migraine headaches: an open-label study. Otolaryngol Head Neck Surg 2000;123:669–76. [PubMed: 11112955]

Black JD, Dolly JO. Interaction of 125I-labeled botulinum neurotoxins with nerve terminals. I.Ultrastructural autoradiographic localization and quantitation of distinct membrane acceptors for typesA and B on motor nerves. J Cell Biol 1986a;103:521–34. [PubMed: 3733877]

Black JD, Dolly JO. Interaction of 125I-labeled botulinum neurotoxins with nerve terminals. II.Autoradiographic evidence for its uptake into motor nerves by acceptor-mediated endocytosis. J CellBiol 1986b;103:535–44. [PubMed: 3015983]

Blersch W, Schulte-Mattler WJ, Przywara S, May A, Bigalke H, Wohlfarth K. Botulinum toxin A andthe cutaneous nociception in humans: a prospective, double-blind, placebo-controlled, randomizedstudy. J Neurol Sci 2002;205:59–63. [PubMed: 12409185]

Blumenfeld A. Botulinum toxin type A as an effective prophylactic treatment in primary headachedisorders. Headache 2003;43:853–60. [PubMed: 12940806]

Burstein R, Cutrer FM, Yarnitsky D. The development of cutaneous allodynia during a migraine attack:clinical evidence for the sequential recruitment of spinal and supraspinal nociceptive neurons inmigraine. Brain 2000a;123:1703–1709. [PubMed: 10908199]

Burstein R, Yamamura H, Malick A, Strassman AM. Chemical stimulation of the intracranial durainduces enhanced responses to facial stimulation in brain stem trigeminal neurons. J Neurophysiol1998;79:964–982. [PubMed: 9463456]

Burstein R, Yarnitsky D, Goor-Aryeh I, Ransil BJ, Bajwa ZH. An association between migraine andcutaneous allodynia. Annals Neurol 2000b;47:614–624.

Chen JJ, Swope DM. Essential tremor: diagnosis and treatment. Pharmacotherapy 2003;23:1105–22.[PubMed: 14524643]

Comella CL, Jankovic J, Brin MF. Use of botulinum toxin type A in the treatment of cervical dystonia.Neurology 2000;55:S15–21. [PubMed: 11188979]

Jakubowski et al. Page 9

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Cui M, Khanijou S, Rubino J, Aoki KR. Subcutaneous administration of botulinum toxin A reducesformalin-induced pain. Pain 2004;107:125–33. [PubMed: 14715398]

Defazio G, Livrea P. Primary blepharospasm: diagnosis and management. Drugs 2004;64:237–44.[PubMed: 14871168]

Dodick DW, Mauskop A, Elkind AH, DeGryse R, Brin MF, Silberstein SD. Botulinum toxin type a forthe prophylaxis of chronic daily headache: subgroup analysis of patients not receiving otherprophylactic medications: a randomized double-blind, placebo-controlled study. Headache2005;45:315–24. [PubMed: 15836567]

Dolly JO. Synaptic transmission: Inhibition of neurotransmitter release by botulinum toxins. Headache2003;43:S16–24. [PubMed: 12887390]

Durham PL, Cady R. Regulation of calcitonin gene-related peptide secretion from trigeminal nerve cellsby botulinum toxin type A: implications for migraine therapy. Headache 2004;44:35–42. [PubMed:14979881]discussion 42–3

Evers S, Vollmer-Haase J, Schwaag S, Rahmann A, Husstedt IW, Frese A. Botulinum toxin A in theprophylactic treatment of migraine--a randomized, double-blind, placebo-controlled study.Cephalalgia 2004;24:838–43. [PubMed: 15377314]

Foran PG, Mohammed N, Lisk GO, Nagwaney S, Lawrence GW, Johnson E, Smith L, Aoki KR, DollyJO. Evaluation of the therapeutic usefulness of botulinum neurotoxin B, C1, E, and F compared withthe long lasting type A. Basis for distinct durations of inhibition of exocytosis in central neurons. JBiol Chem 2003;278:1363–71. [PubMed: 12381720]

Fruhstorfer H, Lindblom U, Schmidt WC. Method for quantitative estimation of thermal thresholds inpatients. J Neurol Neurosurg Psychiatry 1976;39:1071–5. [PubMed: 188989]

Harrison JL, Davis KD. Cold-evoked pain varies with skin type and cooling rate: a psychophysical studyin humans. Pain 1999;83:123–35. [PubMed: 10534583]

Headache-classification-committee-of-the-International-Headache-Society, Classification and diagnosiscriteria for headache disorders, cranial neuralgia and facial pain. Cephalalgia 1988;8:1–92.

Jankovic J. Treatment of cervical dystonia with botulinum toxin. Mov Disord 2004;19:S109–15.[PubMed: 15027062]Suppl 8

Jensen K, Tuxen C, Olesen J. Pericranial muscle tenderness and pressure-pain threshold in the temporalregion during common migraine. Pain 1988;35:65–70. [PubMed: 3200599]

Lawrence GW, Dolly JO. Ca2+-induced changes in SNAREs and synaptotagmin I correlate withtriggered exocytosis from chromaffin cells: insights gleaned into the signal transduction using trypsinand botulinum toxins. J Cell Sci 2002a;115:2791–800. [PubMed: 12077369]

Lawrence GW, Dolly JO. Multiple forms of SNARE complexes in exocytosis from chromaffin cells:effects of Ca(2+), MgATP and botulinum toxin type A. J Cell Sci 2002b;115:667–73. [PubMed:11861772]

Lindblom, U. Analysis of abnormal touch, pain, and temperature sensation in patients. In: Boivie, J.;Hansson, P.; Lindblom, U., editors. Touch, temperature and pain in health and disease: mechanismand assesments. 3. IASP press; Seattle: 1994. p. 63-84.

Marjama-Lyons J, Koller W. Tremor-predominant Parkinson's disease. Approaches to treatment. DrugsAging 2000;16:273–8. [PubMed: 10874522]

Olesen, J.; Goadsby, P.; Ramadan, NM.; Tfelt-Hansen, P.; Welch, KMA. The Headaches. LippincottWilliams & Wilkins; Philadelphia: 2005.

Ondo WG, Vuong KD, Derman HS. Botulinum toxin A for chronic daily headache: a randomized,placebo-controlled, parallel design study. Cephalalgia 2004;24:60–5. [PubMed: 14687015]

Oshinsky M, Poso-Rosich P, Luo J, Hyman S, Silberstein SD. Botulinum toxin A blocks sensitization ofneurons in the trigeminal nucleus caudalis. Cephalalgia 2004;24:781.

Selby G, Lance JW. Observations on 500 cases of migraine and allied vascular headache. J NeurolNeurosurg Psychiat 1960:23–32. [PubMed: 14444681]

Siegel, S. Nonparametric Statistics for the Behavioral Sciences. McGrow-Hill Kogakusha; Tokyo: 1956.p. 312

Silberstein S, Mathew N, Saper J, Jenkins S. Botulinum toxin type A as a migraine preventive treatment.For the BOTOX Migraine Clinical Research Group. Headache 2000;40:445–50. [PubMed:10849039]

Jakubowski et al. Page 10

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Silberstein SD, Olesen J, Bousser MG, Diener HC, Dodick D, First M, Goadsby PJ, Gobel H, Lainez MJ,Lance JW, Lipton RB, Nappi G, Sakai F, Schoenen J, Steiner TJ. The International Classification ofHeadache Disorders, 2nd Edition (ICHD-II)--revision of criteria for 8.2 Medication-overuseheadache. Cephalalgia 2005;25:460–5. [PubMed: 15910572]

Simpson LL. The origin, structure, and pharmacological activity of botulinum toxin. Pharmacol Rev1981;33:155–88. [PubMed: 6119708]

Strigo IA, Carli F, Bushnell MC. Effect of ambient temperature on human pain and temperatureperception. Anesthesiology 2000;92:699–707. [PubMed: 10719949]

Tepper SJ, Bigal ME, Sheftell FD, Rapoport AM. Botulinum neurotoxin type A in the preventivetreatment of refractory headache: a review of 100 consecutive cases. Headache 2004;44:794–800.[PubMed: 15330826]

Tfelt-Hansen P, Lous I, Olesen J. Prevalence and significance of muscle tenderness during commonmigraine attacks. Headache 1981;21:49–54. [PubMed: 7239900]

Tilton AH. Management of spasticity in children with cerebral palsy. Semin Pediatr Neurol 2004;11:58–65. [PubMed: 15132254]

Troost BT. Botulinum toxin type A (Botox) in the treatment of migraine and other headaches. ExpertRev Neurother 2004;4:27–31. [PubMed: 15853612]

Welch MJ, Purkiss JR, Foster KA. Sensitivity of embryonic rat dorsal root ganglia neurons to Clostridiumbotulinum neurotoxins. Toxicon 2000;38:245–58. [PubMed: 10665805]

Wolff HG, Tunis MM, Goodell H. Studies on migraine. Arch Internal Medicine 1953;92:478–484.Yarnitsky D. Quantitative sensory testing. Muscle Nerve 1997;20:198–204. [PubMed: 9040659]

Jakubowski et al. Page 11

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fig. 1.Prospective study: effects of BTX-A on mean±SEM attack frequency, attack duration and painintensity in responders and non-responders. Gray and black bars depict values before and aftertreatment, respectively.

Jakubowski et al. Page 12

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fig. 2.Illustrations of migraine pain and its directionality based on patients’ testimonies. A. explodingheadache, B. Imploding (crushing and stabbing) headache, C. Ocular headache.

Jakubowski et al. Page 13

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fig. 3.Retrospective study: effects of BTX-A on mean±SEM attack frequency, attack duration andpain intensity in responders and non-responders. Gray and black bars depict values before andafter treatment, respectively.

Jakubowski et al. Page 14

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fig. 4.Effects of BTX-A on mean±SEM number of days of migraine per month (attack frequency ×attack duration) according to the type of headache in imploding (n=31), ocular (n=5) andexploding (n=27) headache patients pooled from the prospective and retrospective studies.Gray and black bars depict values before and after treatment, respectively.

Jakubowski et al. Page 15

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fig. 5.Effects of placebo injections on mean±SEM number of days of migraine per month (attackfrequency × attack duration) according to the type of headache in imploding/ocular (n=6) andexploding (n=6) headache patients. Gray and black bars depict values before and aftertreatment, respectively.

Jakubowski et al. Page 16

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Jakubowski et al. Page 17

TABLE 1Prospective study: classification by QST of skin sensitivity to heat, cold and mechanical stimulation in individualmigraine patients before and after BTX-A treatment. ○ normal sensitivity; ● hypersensitivity. A, allodynic; NA,non-allodynic.

Before BTX-A After BTX-APatient Cold Heat Mechanical Allodynia Allodynia Cold Heat Mechanical

Responders LD ○ ○ ● A A ○ ○ ● KF ○ ○ ○ NA NA ○ ○ ○ RN ● ● ● A A ● ● ● JS ● ○ ● A A ● ● ● CS ● ● ● A A ○ ● ● SW ● ○ ● A A ● ● ● ES ○ ● ● A A ○ ● ● RS ● ○ ● A A ● ○ ● TS ○ ● ● A A ○ ○ ● DC ● ○ ● A A ● ○ ● RD ● ○ ● A A ● ○ ● AS ● ● ● A A ● ● ● KV ● ○ ○ A A ● ○ ○ MS ● ● ● A A ● ● ●Non-responder MC ● ● ● A A ○ ● ● NG ● ● ○ A A ● ● ○ RL ○ ○ ○ NA NA ○ ○ ○ MM ○ ○ ○ NA NA ○ ○ ○ SM ○ ○ ○ NA NA ○ ○ ○ JN ● ● ● A A ● ● ● DN ○ ○ ● A A ○ ○ ● ND ● ● ○ A A ● ● ○ CP ○ ○ ● A NA ○ ○ ○ RG ○ ○ ● A A ○ ○ ● CM ○ ○ ● A A ○ ○ ● EG ● ● ● A A ● ● ● MC ● ● ● A A ● ● ●

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Jakubowski et al. Page 18

TABLE 2Combined tudies: distribution of responders and non-responders according to the classification of the headache

Type of headache N Responders Non-responders

Exploding headache 27 5 (19%) 21 (81%)Imploding headache 31 29 (94%) 2 (6%)Ocular headache 5 5 (100%) 0

Pain. Author manuscript; available in PMC 2007 December 5.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Jakubowski et al. Page 19

TABLE 3Combined studies: classification of muscle tenderness among responders and non-responders

Muscle tenderness Responders (n=39) Non-responders (n=24)

None 9 (23%) 7 (29%)After onset of migraine 14 (36%) 8 (33.5%)Before onset of migraine 13 (33%) 8 (33.5%)Constant 3 (8%) 1 ( 4%)

Pain. Author manuscript; available in PMC 2007 December 5.