SEARCH

SEARCH BY CITATION

Keywords:

  • migraine;
  • headache directionality;
  • ocular headache;
  • pain;
  • onabotulinum toxin;
  • headache classification

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References

Background

The study aims to compare methods of determining headache directionality (imploding, exploding, and/or ocular headaches) in women with migraine, investigate the concordance between physician assignment and patient self-assignment of pain directionality, and evaluate whether patients assigned their headaches to the same direction when queried using different methods. Directionality of migraine headache pain (imploding, exploding, or ocular) may reflect differences in the underlying pathogenesis of individual migraine attacks among and within individuals. Emerging evidence suggests that directionality of pain in migraine sufferers may predict response to onabotulinumtoxin A. The best method of determining headache directionality in migraine sufferers has not been systematically explored.

Methods

We conducted a prospective cross-sectional survey study of 198 female patients with migraine presenting to a Women's Health Clinic. Patients determined the directionality (imploding, exploding, and/or ocular) of their own migraine pain by choosing among 3 pictures graphically representing directionality and also by responding to a written question regarding directionality. Clinicians then classified directionality of migraine pain using structured interviews. Concordance between clinician assignment of directionality and patient self-assignment was determined with Kappa coefficients.

Results

Subjects were females between the ages of 18 and 77 years (mean 48 years). According to patient selection of representative pictures, 62 (31.6%) had imploding headaches with or without ocular pain, 36 (18.4%) had exploding headaches with or without ocular pain, 78 (39.8%) had ocular pain only, and 20 (10.2%) had imploding and exploding headaches with or without ocular pain. Two subjects did not respond. According to patient responses to a written question, 80 (41.0%) had imploding headaches with or without ocular pain, 53 (27.2%) had exploding headaches with or without ocular pain, 46 (23.6%) had ocular pain only, and 16 (8.2%) had imploding and exploding headaches with or without ocular pain. Three subjects did not respond. For physician assignment, 69 (34.9%) subjects had imploding headaches with or without ocular pain, 89 (45%) had exploding headaches with or without ocular pain, 14 (7.1%) had ocular pain only, and 26 (13.1%) had imploding and exploding headaches with or without ocular pain. The concordance (Kappa coefficient) between physician assignment of headache directionality with patient response to the written question was 0.33 (weak agreement), between physician assignment and patient assignment via selection of representative pictures was 0.35 (weak agreement), and between patient assignment via written question and via selection of representative pictures was 0.35 (weak agreement).

Conclusions

The assignment of headache directionality varied substantially depending upon the method of determination. The concordance between clinician assignment, patient-self assignment via answering a written question, and patient self-assignment via choosing a representative picture was weak. Improved methods of determining pain directionality are needed.

Abbreviation
ICHD-II

International Classification of Headache Disorders 2nd edition

Migraine affects approximately 18% of women and 6% of men in the USA.[1] It is estimated that in the general population about 1/4 of people with migraine should be offered migraine prophylactic therapy.[1] However, responses to prophylactic therapy are highly variable, with only 40-50% of patients responding to any one of the first-line prophylactic medications.[2] To date, there are not clinical factors that reliably predict response to an individual treatment, rendering selection of acute and preventive therapy for the individual patient a process of trial and error.

The direction that a patient feels their headache, such as from the outside of the head inward (ie, imploding) or from deep inside the head outward (ie, exploding), may reflect differences in the underlying pathogenesis of individual migraine attacks among and within individuals.[3, 4] Headache pain directionality has been explored as a headache characteristic that may be useful in predicting treatment response.3,5-8 Initial reports from 2006 showed a marked difference in response to onabotulinumtoxin A between patients describing imploding vs exploding migraine headache.[3] Responders were more likely to describe their headache as “imploding or ocular,” while nonresponders were more likely to describe “exploding” headache. Other studies looking at patient perception of migraine directionality as a predictor of responsiveness to onabotulinumtoxin treatment have shown similar results.[5, 7, 8] The pathophysiology underlying the difference in these 2 groups is not clear, but it has been suggested that imploding or ocular headache may have an extracranial origin that is mediated by activation of meningeal nerves that infiltrate the periosteum through the calvarial sutures.[4]

The best methods to differentiate imploding, ocular, and exploding headache types in migraine sufferers have not been systematically explored. Clinical observations regarding the difficulty with correct assignment of headache directionality have been discussed in the literature.[6, 10] Clinicians have observed that headache patients often have difficulty consistently describing and assigning directionality to their headache pain.[6, 10] Currently, no specific criteria exist for defining headache directionality, nor are there agreed upon descriptors to aid the clinician and patient in assigning directionality.

The purpose of our study was to investigate different methods of determining imploding, exploding, and or ocular headaches in women with migraine, to investigate the concordance between physician assignment and patient self-assignment of pain directionality, to assess interattack and intra-attack variability in headache directionality, and to evaluate the consistency with which patients assigned a direction to their usual headache when queried using different methods.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References

We conducted an institutional review board approved prospective, cross-sectional survey study of 198 consecutive patients seen in an outpatient women's health practice at our institution from January 2008 to October 2012. Female patients between the ages of 18 and 77 who fulfilled the International Classification of Headache Disorders 2nd edition (ICHD-II) diagnostic criteria for migraine with or without aura[11] were asked to participate in the study. A chief complaint of headache was not required for participation in the study. Patients with migraine headache were identified through direct questioning or chart review at time of clinic appointment and when patients requested a prescription refill of a migraine specific therapy. If identified through prescription refill request, patients were asked to participate in the study at the time of their next appointment. If no appointment was scheduled, patients were asked to come in to complete a survey. Patients were excluded from the study for headache not fulfilling ICHD-II criteria for migraine, an inability to read English, visual or communication impairment that led to an inability to complete the survey, long-term maintenance opioid therapy for headache or another chronic pain condition, and patient refusal.

Following informed consent, patients completed a self-administered survey regarding pain directionality and number of headache days within the last 90 days. This was followed by a provider-administered survey regarding pain directionality.

Survey questions were designed to allow the patient to express pain directionality utilizing simple written and pictorial representations of pain (Fig. 1). The design of the questions and diagrams were similar to those utilized by Jakubowski et al.[3] The survey classified migraine pain directionality as “exploding alone, imploding alone, ocular alone, or mixed (ie, combination of any 2 types or all 3 types).” Patients were given a paper survey with the pictorial representations listed first followed by the written question, “Is your headache pain pushing in or pushing out of your head or is it located within your eye socket (ocular).” After the patient completed the written survey, the clinician conducted a scripted interview from which headache directionality, interattack variability in headache directionality and intra-attack variability in headache directionality were determined (Fig. 2). The clinician was blinded to the patient self-assignments of headache directionality.

figure

Figure 1. Pictorial representation of headache directionality. Imploding, ocular, exploding.

Download figure to PowerPoint

figure

Figure 2. Physician administered headache survey to determine headache directionality.

Download figure to PowerPoint

Statistical Methods

Summary statistics were used to describe the study sample. Mean ± standard deviation was reported for continuous variables, and percentage/frequency count was computed for nominal variables. For comparison between groups, Kruskal–Wallis test and Pearson's chi-square test (or Fisher's exact test, if applicable) were applied. If the overall test was significant, pairwise comparison was performed using Bonferroni adjustment. Using a 2-sided test, a P value < .05 for overall test was considered statistically significant. Kappa coefficients were calculated to determine concordance between the different methods of assigning headaches to 1 of 4 pain directionalities: imploding ± ocular, exploding ± ocular, ocular, or imploding and exploding. Kappa coefficients were considered weak if less than 0.41, moderate if between 0.41 and 0.60, and strong if 0.61 or greater.[9] Analyses were conducted using SAS 9.2 (SAS Institute, Inc., Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References

Patient Demographics (Table 1)

Table 1. Patient Demographics
 Imploding ± Ocular (N = 69)Exploding ± Ocular (N = 89)Ocular Only (N = 14)(Imploding + Exploding) ± Ocular (N = 26)Total (N = 198)P Value
  1. †Kruskal Wallis; ‡Chi-square.

  2. There were no differences in age, race or highest level of education attained between subjects with different pain directionality.

Age at examination     .0864
N61861425186 
Mean (SD)44.3 (13.3)49.6 (11.7)48.7 (12.3)50.4 (11.3)47.9 (12.4) 
Median44.150.048.455.549.7 
Q1, Q333.4, 55.743.1, 57.240.0, 57.843.7, 57.140.0, 57.2 
Range(20.0-70.6)(20.5-77.2)(25.8-65.4)(19.9-63.6)(19.9-77.2) 
Race     .6186
Missing1 (%)1 (%)0 (%)0 (%)2 
White63 (92.6%)76 (86.4%)13 (92.9%)21 (80.8%)173 (88.3%) 
African American1 (1.5%)2 (2.3%)0 (0.0%)2 (7.7%)5 (2.6%) 
Asian/Pacific Islander1 (1.5%)5 (5.7%)0 (0.0%)2 (7.7%)8 (4.1%) 
Others3 (4.4%)5 (5.7%)1 (7.1%)1 (3.8%)10 (5.1%) 
Education     .7338
Missing1 (%)0 (%)0 (%)0 (%)1 
Grade 11 or less1 (1.5%)0 (0.0%)0 (0.0%)0 (0.0%)1 (0.5%) 
Graduated high school0 (0.0%)4 (4.5%)1 (7.1%)2 (7.7%)7 (3.6%) 
Some college or technical school23 (33.8%)25 (28.1%)4 (28.6%)8 (30.8%)60 (30.5%) 
Graduated college21 (30.9%)28 (31.5%)6 (42.9%)10 (38.5%)65 (33.0%) 
Some graduate work5 (7.4%)8 (9.0%)2 (14.3%)1 (3.8%)16 (8.1%) 
A graduate degree18 (26.5%)24 (27.0%)1 (7.1%)5 (19.2%)48 (24.4%) 

One hundred ninety-eight female patients between the ages of 18 years and 77 years were included in the study. Mean age was 48 ± 12.4 years, and median age was 50 years. Race was self-reported as white by 88%, Asian/Pacific Islander by 4%, African American by 3%, and Other by 5%. Highest level of education was graduated college or higher by 65.5%, some college or technical school by 31%, graduated high school by 4%, and grade 11 or less by 1%. There were no differences in age, race, or highest level of education when comparing subjects with different headache directionality. Ninety-five percent (n = 188) of patients in the study reported episodic migraine (<45 headache days in 90), while 5% (n = 10) reported chronic migraine (>45 headache days in 90 days).

Pain Directionality (Table 2; Fig. 3)

figure

Figure 3. Pain directionality according to method of assignment. The number of subjects with headaches described as imploding with or without ocular pain, exploding with or without ocular pain, ocular only, or imploding and exploding with or without ocular pain varied widely according to the method by which pain directionality was assigned. There was weak agreement between physician assignment of pain directionality according to answers given in a scripted questionnaire, patient self-assignment via answering a written question, and patient self-assignment via choosing among images representing pain directionality.

Download figure to PowerPoint

Table 2. Headache Type by Method of Assignment
Headache DirectionalityPatient Assigned Pictorial RepresentationPatient Assigned Written DescriptionPhysician Assigned
  1. Patients assigned to headache directionality utilizing 1 of 3 methods: physician assigned, patient assigned utilizing pictorial representation of headache directionality, and patient assigned utilizing a written description of headache directionality.

Imploding only29 (14.8%)40 (20.51%)37 (18.69%)
Exploding only16 (8.16%)21 (10.77%)45 (22.73%)
Ocular only78 (39.80%)46 (23.59%)14 (7.07%)
Imploding + ocular33 (16.84%)40 (20.51%)23 (16.16%)
Imploding + exploding9 (4.59%)7 (3.59%)18 (9.09%)
Exploding + ocular20 (10.20%)32 (16.41%)44 (22.22%)
Imploding + exploding + ocular11 (5.61%)9 (4.62%)8 (4.04%)

According to patient selection of pictures that best represented their pain directionality, 14.8% of patients had imploding headaches, 8.2% exploding, 39.8% ocular, 16.8% imploding + ocular, 4.6% imploding + exploding, 10.2% exploding + ocular, and 5.6% had all 3 types (imploding + exploding + ocular). Two patients did not choose a pictorial representation of headache. According to patients' responses to the question “Is your headache pain pushing in or pushing out of your head or is it located within your eye socket (ocular)? (check all that apply),” 20.5% of patients had imploding headaches, 10.8% exploding, 23.6% ocular, 20.5% imploding + ocular, 3.6% imploding + exploding, 16.4% exploding + ocular, and 4.6% had all 3 types (imploding + exploding + ocular). Three patients did not choose a written descriptor of headache. Reasons for nonresponse were not elucidated. According to physicians' diagnoses according to scripted questionnaire, 18.7% of patients had imploding headaches, 22.7% exploding, and 7.1% ocular headaches, while 16.2% had imploding + ocular, 9.1% imploding + exploding, 22.2% exploding + ocular, and 4% had all 3 types (imploding + exploding + ocular). Ten subjects (5%) had pain directionality that varied within an individual migraine attack (ie, intra-attack variability), and 11 (6%) subjects had different pain directionalities from 1 migraine attack to another (ie, inter-attack variability).

A total of 77 patients used prophylactic medications, and among them, 14 (18.2%) had imploding, 18 (23.4%) had exploding, 6 (7.7%) had ocular only, 7 (9.1%) had imploding + ocular, 9 (11.7%) had imploding + exploding, 19 (24.7%) had exploding + ocular, and 4 (5.2%) had all 3 types. One hundred twenty-one patients did not use prophylactic meds: 23 (19.0%) had imploding, 27 (22.3%) had exploding, 8 (6.6%) had ocular only, 25 (20.7%) had imploding + ocular, 9 (7.4%) had imploding + exploding, 25 (20.7%) had exploding + ocular, and 4 (3.3%) had all 3 types. There was no difference in the distribution of headache directionality between subjects using prophylactic medication and subjects not using such medications (P = .4549).

Concordance Between Different Methods of Assigning Pain Directionality

There was weak agreement, Kappa coefficient 0.33 (P < .0001) between physician diagnosis of pain directionality and patient self-assignment via answering the written question about pain directionality. There was weak agreement, Kappa coefficient 0.35 (P < .0001), between physician diagnosis of pain directionality and patient self-assignment via selection of representative pictures. There was weak agreement, Kappa coefficient 0.35 (P = .0005), between subject self-assignment of pain directionality via answering the written question about pain directionality and choosing from representative pictures.

Concordance between methods of assigning pain directionality was also determined for each pain direction separately. For imploding headaches, there was moderate agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via selection of representative pictures (Kappa coefficient 0.50, P < .0001), weak agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via answering the written question (Kappa coefficient 0.39, P < .0001), and moderate agreement between patient self-assignment via selection of representative pictures and patient self-assignment via answering the written question (Kappa coefficient 0.43, P < .0001). For exploding headaches, there was weak agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via selection of representative pictures (Kappa coefficient 0.33, P < .0001), weak agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via answering the written question (Kappa coefficient 0.35, P < .0001), and weak agreement between patient self-assignment via selection of representative pictures and patient self-assignment via answering the written question (Kappa coefficient 0.39, P < .0001). For ocular headaches, there was moderate agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via selection of representative pictures (Kappa coefficient 0.42, P < .0001), weak agreement between physician diagnosis according to scripted questionnaire and patient self-assignment via answering the written question (Kappa coefficient 0.37, P < .0001), and moderate agreement between patient self-assignment via selection of representative pictures and patient self-assignment via answering the written question (Kappa coefficient 0.57, P < .0001).

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References

Responses to migraine therapies vary substantially among patients. For example, when measuring response to prophylactic therapy as at least a 50% reduction in headache frequency, less than one-half of patients treating with a first-line therapy are responders.[4] Identification of clinical factors that predict a patient's likelihood of responding to a specific migraine therapy would transition the treatment of migraine from a process of trial-and-error to a process of individualized medicine, maximize patient outcomes, and minimize patient exposure to the potential adverse events from medications to which they are unlikely to respond.

Published reports have suggested that migraine pain directionality is predictive of a response to onabotulinumtoxin A therapy. Studies have found an association between headache pain directionality and response to onabotulinumtoxin A[3, 5, 8] and more recently to botulinum toxin B.[7] Headache pain directionality has been described as imploding (a vice-like pain and pressure squeezing in), exploding (pain and pressure pushing outward), or ocular (pain focused on the eye).[3, 7] However, methods for determination of headache pain directionality have not been standardized. A number of different methods for determining headache pain directionality have been described. Written descriptors, drawings, or a combination of both were utilized in different studies to help patients assign directionality to their headache pain.[3, 5, 7, 8] The accuracy and concordance of these different methods to assign pain directionality are unknown.

In this study, we compared different methods of assessing headache directionality in a cohort of patients presenting to a women's health center for evaluation of issues not necessarily specific to headache. Participants in the study were representative of patients seen in an internal medicine practice rather than in a specialty headache practice, accounting for the fact that the majority of patients had episodic migraine. This setting is particularly relevant because migraine is most often managed in the primary care setting.[12] Our results demonstrate that the assignment of directionality varied substantially depending on the method of determination, and the concordances between different methods of assignment were generally weak. The concordance between clinician assignment and patient-self assignment was weak regardless of the method used by the patient to describe the headache pain directionality (choice of a representative picture or answering a written question). When considering specific pain directions individually, concordances between methods of assigning pain direction were weak to moderate. Intra-attack and interattack variability in headache directionality make assignment of headache directionality more complex and would likely contribute to weak concordance between different assignment methods. However, few subjects in this study had variability in headache directionality, suggesting that headache variability could have only had a minor impact on study results. It is possible that the use of migraine prophylactic medications could alter pain directionality. However, in this study, there was no difference in headache directionality in patients who reported the use of prophylactic medications compared with those who did not. Concerns have previously been raised regarding an individual patient's ability to adequately and consistently describe headache pain and its directionality.[6, 10] Our study supports these concerns and suggests the need for further development and study of methods to assign migraine pain directionality.

In this study, we sought to evaluate a novel method to assess and compare different methods of assigning headache pain directionality and concordance between the different methods in a group of women seeking care in a primary care setting. The study did not intentionally exclude potential subjects based on gender, race, or cultural background, but the patient population studied was relatively homogenous reflecting the demographics of the women's health clinic. This population may be more reflective of the general migraine population than that seen in a specialty headache practice. However, findings reflect a population drawn from a single institution and do not include men. We consider the use of a relatively homogenous patient population a strength of this study. If the methods of assignment are not concordant within a homogeneous population, they are likely to perform even worse in a heterogeneous population. The results of this study provide a starting point for testing methods of assigning pain directionality. Failure to achieve internal concordance between different methods of assigning pain directionality may have related to the design of the drawings and the specific verbal descriptors that were utilized. Drawings utilized to describe headache directionality were unique and drawn specifically for this study. Thus, we cannot necessarily assume that our study findings (ie, weak concordance between different methods of assigning pain directionality) are generalizable to methods of assigning pain directionality used in other studies. We recommend that investigators validate their methods of assigning pain directionality prior to using them in future studies aimed at predicting treatment response based on pain directionality. A limitation of our study was that we did not conduct a pilot study to test the methods of assigning pain directionality prior to enrolling subjects. Additionally, consideration could be given to providing text descriptions and drawings in the same assessment tool as a means of improving reliability of assigning directionality. Furthermore, 3 patients did not answer the question about headache directionality presented as a picture, and 2 failed to answer the written question in this study. The missing data for these subjects could not be included in calculations of concordance between methods of assigning pain directionality.

A possible influence on a patient's ability to assign headache directionality is their lack of familiarity with this concept. Migraine patients have been asked about aura, nausea, vomiting, throbbing, laterality, etc, for many years by many doctors and thus know how to answer questions about these headache characteristics. In contrast, questions about headache directionality are uncommonly asked, and unfamiliarity with the concept may influence the ability to assign directionality. It is possible that informing patients about this concept and asking them to prospectively record headache directionality in a prospective headache diary might improve the ability to assign directionality. Furthermore, while the clinical interviews in this study utilized a structured questionnaire, 7 physician interviewers participated in interviews, each of whom may have influenced results through their individualized use of gestures, inflections, and emphasis of certain words when administering the scripted interview.

In conclusion, headache pain directionality may have implications for predicting treatment responses to specific migraine prophylactic therapies. However, valid methods of determining pain directionality are not yet established. In this study, the concordance between assigning pain directionality via clinician assignment, patient-self assignment via answering a written question, and patient self-assignment via choosing a representative picture was weak. Improved methods of determining pain directionality are needed.

Statement of Authorship

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References

Category 1

  • (a)
    Conception and Design
    Julia A. Files; Todd J. Schwedt; Paru S. David; Bert B. Vargas; Beverly S. Tozer; David W. Dodick
  • (b)
    Acquisition of Data
    Julia A. Files; Paru S. David; Megan Hunt; Salma Patel; Marcia G. Ko
  • (c)
    Analysis and Interpretation of Data
    Julia A. Files; Todd J. Schwedt; Anita P. Mayer; Bert B. Vargas; Yu-Hui Chang; Megan Hunt; Salma Patel; Rami Burstein; David W. Dodick

Category 2

  • (a)
    Drafting the Manuscript
    Julia A. Files; Todd J. Schwedt; Paru S. David; Yu-Hui Chang; Megan Hunt; Salma Patel; Beverly S. Tozer; Rami Burstein; David W. Dodick
  • (b)
    Revising it for Intellectual Content
    Julia A. Files; Todd J. Schwedt; Anita P. Mayer; Bert B. Vargas; Marcia G. Ko; David W. Dodick

Category 3

  • (a)
    Final Approval of the Completed Manuscript
    Julia A. Files; Todd J. Schwedt; Anita P. Mayer; Paru S. David; Bert B. Vargas; Yu-Hui Chang; Megan Hunt; Salma Patel; Marcia G. Ko; Beverly S. Tozer; Rami Burstein; David W. Dodick

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Statement of Authorship
  7. References