Abstract
Objective
To develop and validate a questionnaire for assessing cutaneous allodynia (CA), and to estimate the prevalence and severity of CA in the migraine population.
Methods
Migraineurs (n = 11,388) completed the Allodynia Symptom Checklist, assessing the frequency of allodynia symptoms during headache. Response options were never (0), rarely (0), less than 50% of the time (1), ≥50% of the time (2), and none (0). We used item response theory to explore how well each item discriminated CA. The relations of CA to headache features were examined.
Results
All 12 questions had excellent item properties. The greatest discrimination occurred with CA during “taking a shower” (discrimination = 2.54), wearing a necklace (2.39) or ring (2.31), and exposure to heat (2.1) or cold (2.0). The factor analysis demonstrated three factors: thermal, mechanical static, and mechanical dynamic. Based on the psychometrics, we developed a scale distinguishing no CA (scores 0–2), mild (3–5), moderate (6–8), and severe (≥9). The prevalence of allodynia among migraineurs was 63.2%. Severe CA occurred in 20.4% of migraineurs. CA was associated with migraine defining features (eg, unilateral pain: odds ratio, 2.3; 95% confidence interval, 2.0 –2.4; throbbing pain: odds ratio, 2.3; 95% confidence interval, 2.1–2.6; nausea: odds ratio, 2.3; 95% confidence interval, 2.1–2.6), as well as illness duration, attack frequency, and disability.
Interpretation
The Allodynia Symptom Checklist measures overall allodynia and subtypes. CA affects 63% of migraineurs in the population and is associated with frequency, severity, disability, and associated symptoms of migraine. CA maps onto migraine biology.
Cutaneous allodynia (CA) is characterized by pain provoked by stimulation of the skin that would ordinarily not produce pain.1 The underlying mechanism of facial CA is sensitization of the nociceptive neurons in the trigeminal nucleus caudalis, which receives convergent afferent input from the dura mater and periorbital skin.2,3 Clinic-based studies suggest that about two thirds of migraine sufferers experience development of CA.4–6 As a marker of central sensitization, allodynia has been proposed as a risk factor for progression to chronic migraine.7–9 Therefore, CA has significant implications for our understanding of the pathophysiology of migraine attacks, for the implementation of treatment, and for assessing prognosis.
CA is usually assessed by quantitative sensory testing (QST). QST requires specialized equipment, training, and testing; it is too cumbersome and costly for wide-spread use in clinical practice or epidemiological research and is subject to temporal sampling error. As a consequence, most studies on headache and CA come from a small number of headache centers and the highly selected patients treated there.5,10,11 There is an urgent need to develop and validate simple methods for assessing CA to better characterize CA in representative samples and to facilitate clinical practice.
Several groups have developed questionnaires to assess CA,10,12,13 but only one instrument has been validated against QST.13 Most questionnaires use dichotomous response options and score CA as present or absent. However, CA is not an all-or-none phenomenon; it varies over time, emerging and remitting during the course of a migraine attack.13 In addition, the available questionnaires assess CA as a unitary dimension, 12,13 though at least three types of CA are well described (ie, thermal, dynamic mechanical, and static mechanical allodynia).1,14 Thermal allodynia is tested with QST by measuring nociceptive thresholds to hot and cold. It is mediated by C nociceptive and Aδ fibers. 1,14 Dynamic mechanical allodynia (brush allodynia), assessed by brushing the skin, is likely mediated by Aβ mechanoreceptive and capsaicin-insensitive Aβ fibers.14 Von Frey hair filaments have been used to assess static mechanical (or pressure) allodynia, which is mediated by Aδ nociceptive fibers.15
We developed the Allodynia Symptom Checklist (ASC) by modifying Jakubowski and colleagues’14 instrument to provide graded response options. Our goals were to quantify CA overall and to determine whether there were natural subtypes of allodynia. We administered the ASC to a population sample of severe headache sufferers identified through the American Migraine Prevalence and Prevention (AMPP) project. Herein we describe the questionnaire and its psychometric attributes in the general migraine population We also describe the prevalence distribution and characteristics of allodynia in this population.
Subjects and Methods
This study was conducted as a part of the AMPP project, the details of which are reported elsewhere.15 This research protocol was approved by the Institutional Review Board at the Albert Einstein College of Medicine. In brief, we first screened a sample of 120,000 US households selected to be representative of the US population using a validated headache questionnaire (AMPP phase 1). Of 162,576 individual respondents, 30,721 reported having at least one severe headache in the prior year that was not caused by a head injury or by a disease, such as the flu. We selected a random sample of 24,000 of these severe headache sufferers and sent them a second mailed questionnaire in phase 2. This questionnaire was used to determine headache diagnoses for up to three types of headache based on the Second Edition of the Inter-national Classification of Headache Disorders criteria.16 It was also used to assess details of headache frequency, burden, treatment, and comorbidities. This phase 2 instrument also included the ASC, as detailed in the following section.
Allodynia Symptom Checklist
The ASC included 12 questions (Fig) about the frequency of various allodynia symptoms in association with headache attacks. For individuals with more than one type of headache, questions were directed to the “most severe type of headache,” based on the prior evidence indicating that the most severe type was likely to be migraine.17–19 Instead of using a dichotomous option (yes or no), the response categories were “never,” “rarely,” “less than half the time,” and “half the time or more.” Based on prior studies, the option “rarely” was considered a negative response to reduce false-positive symptom reporting.17,18,20 In addition, subjects could also indicate that an item “does not apply to me.” That option was used by someone who never shaved their face or someone who never wore a ponytail.
Fig. 1.
The 12-item Allodynia Symptom Checklist (ASC-12).
Data Analysis
ASC items were scored as 0 (ie, never, rarely, or does not apply to me), 1 (less than half the time), and 2 (half the time or more), yielding scores that ranged from 0 to 24. Alternative scoring strategies were evaluated but did not materially alter our results.21,22
Exploratory factor analysis was performed to determine whether the 12 items could be reduced to a set of independent factors representing the various dimensions of CA. Factor analysis was completed using ordinary least-squares extraction and oblique equimax rotation of the solutions.23 Ordinary least squares was used for factor extraction instead of maximum likelihood because the items are categorical. Items were retained in the factor on which they had highest loading. Three factors emerged from the exploratory factor analyses (see Results).
Item response theory was used to determine the relation between individual items and allodynia severity. We first applied a traditional item response model, treating allodynia as a single underlying construct. We then applied a nonlinear mixed-effects model to each of the factors in the three-factor solution.21 Interfactor correlations were estimated among the three allodynia domains. We modeled our data using Gaussian adaptive quadrature, with four quadrature points per dimension. 24
To assess the severity of allodynia, we estimated two parameters: discrimination and threshold. Discrimination corresponds to the slope of a logistic function. High item discrimination values (values greater than 1) indicate that the item distinguishes subjects with high values on the CA latent variable from those with lower values. Threshold indicates the score of allodynia associated with a 50% chance of endorsing a given response category for an item. Because there are three levels of responses to each item on the ASC, there are two thresholds. The first threshold is between “never” or “rarely” and “less than the half the time,” and the second is between “less than half the time” and “half the time or more.” The first threshold demarcates the location on the latent variable scale beyond which subjects are more likely to indicate that the symptoms occur “less than half the time” rather than “never” or “rarely.” The second threshold demarcates the location on the allodynia scale beyond which the probability of endorsing “half the time or more” exceeds “less than half the time,” “never,” or “rarely.” Because we assume that latent variables follow standard normal distributions, the thresholds can be interpreted as the number of standard deviations above or below the mean of the latent variable required for a respondent to have a 50% chance of endorsing the response category. Because 98% of the standard normal distribution is bounded between three standard deviations, items with thresholds beyond two standard deviations discriminate at very high or low levels of the latent variable.
Missing Data
Missing data were handled through multiple imputation (missing data were assumed missing at random) to minimize impact on subsequent parameter estimates. Our approach performs multiple categorical imputations under a saturated multinomial model using an established algorithm.23–25
Migraine and Allodynia
We modeled CA as the outcome variable in a series of Poisson regression models. In Model 1, we adjusted for demographic variables (eg, age, sex, race, income). In Model 2, we added to adjustments for headache frequency, severity, and duration of illness. In Model 3, we added to Model 2 adjustments that included comorbidities, use of preventive medication, use of triptans, and use of opioids.
Results
Study Population
Of 24,000 mailed questionnaires, 16,577 headache sufferers returned usable surveys (69.1% response rate). Among these respondents, 11,388 met Second Edition of the International Classification of Headache Disorders criteria for migraine. Demographics for the returned sample closely matched those of the outgoing sample of past year headache sufferers (Table 1).
Table 1.
Demographic Features for the Target Sample and Respondent Sample
| Demographic Features | Target Sample (N = 24,000), n (%) |
Returned Sample (N = 16,577), n (%) |
Response Rate |
|---|---|---|---|
| Sex | |||
| Male | 7,077 (29.5) | 4,053 (24.4) | 57% |
| Female | 16,923 (70.5) | 12,524 (75.6) | 74% |
| Race | |||
| White | 20,528 (85.5) | 14,364 (86.6) | 70% |
| Black | 2,021 (8.4) | 1,311(7.9) | 65% |
| Asian, Pacific Islander | 243 (1.0) | 142 (0.8) | 60% |
| American, Indian | 216 (0.9) | 124 (0.7) | 60% |
| Other | 362 (1.5) | 225 (1.3) | 62% |
| Unknown/no answer | 630 (2.6) | 411 (2.5) | 65% |
| Age, yr | |||
| 18–24 | 1,768 (7.4) | 741 (4.5) | 42% |
| 25–34 | 4,179 (17.4) | 2,478 (14.9) | 59% |
| 35–44 | 5,414 (22.6) | 3,693 (22.3) | 68% |
| 45–54 | 6,191 (25.8) | 4,616 (27.9) | 75% |
| 55–64 | 3,706 (15.4) | 2,977 (17.9) | 80% |
| 65–74 | 1,676 (6.9) | 1,321 (8.0) | 79% |
| ≥75 | 1,066 (4.4) | 751 (4.5) | 70% |
| Region | |||
| New England | 1,104 (4.6) | 758 (4.6) | 69% |
| Middle Atlantic | 3,313 (13.8) | 2,259 (13.6) | 68% |
| East North Central | 3,827 (15.9) | 2,682 (16.3) | 70% |
| West North Central | 1,696 (7.1) | 1,200 (7.2) | 71% |
| South Atlantic | 4,672 (19.5) | 3,213 (19.4) | 69% |
| East South Central | 1,824 (7.6) | 1,306 (7.9) | 72% |
| West South Central | 2,791 (11.6) | 1,904 (11.5) | 68% |
| Mountain | 1,550 (6.5) | 1,087 (6.6) | 70 |
| Pacific | 3,223 (13.4) | 2,168 (13.0) | 67% |
| Urbanization | |||
| <100,000 | 3,883 (16.2) | 2,770 (16.7) | 71% |
| 100,000–499,999 | 4,174 (17.4) | 2,916 (17.6) | 70% |
| 500,000–1,999,999 | 5,772 (24.1) | 3,987 (24.0) | 69% |
| ≥2,000,000 | 10,171 (42.4) | 6,904 (41.7) | 68% |
| Household size | |||
| 1 member | 4,527 (18.9) | 3,129 (18.9) | 69% |
| 2 members | 7,950 (33.1) | 5,680 (34.3) | 71% |
| 3 members | 4,421 (18.4) | 3,045 (18.4) | 69% |
| 4 members | 4,018 (16.7) | 2,706 (16.3) | 67% |
| ≥5 members | 3,084 (12.9) | 2,017 (12.2) | 65% |
| Family annual income | |||
| <$22,500 | 6,378 (26.6) | 4,267 (25.7) | 67% |
| $22,500–$39,999 | 4,893 (20.4) | 3,312 (19.9) | 68% |
| $40,000–$59,999 | 4,390 (18.3) | 3,094 (18.7) | 70% |
| $60,000–$89,999 | 4,234 (17.6) | 2,993 (18.1) | 70% |
| ≥$90,000 | 4,105 (17.1) | 2,911 (17.7) | 71% |
Identifying Questionnaire Factors
Examination of the screen plot and eigenvalues suggested three CA factors (Table 2). We interpreted the three factors to represent thermal, mechanical static, and mechanical dynamic allodynia, and labeled them accordingly. The thermal factor, which reflects pain sensitivity to heat and cold, includes five items (ie, shaving your face, taking a shower, resting your face or head on a pillow, exposure to heat, and exposure to cold). The mechanical static factor is composed of five items (ie, wearing eyeglasses, wearing contact lenses, wearing earrings, wearing a necklace, and wearing tight clothing) and reflects pressure in a fixed locus. The mechanical dynamic factor comprises two items (ie, combing your hair and pulling your hair back) and reflects a more dynamic pressure across an area of skin. The three factors were intercorrelated. Items on the thermal allodynia factor were correlated to the mechanical static factor (0.44) and the mechanical dynamic factor (0.41). Items on the mechanical static and mechanical dynamic factors were also correlated (0.42).
Table 2.
Exploratory Factor Analysis for 12-Item Allodynia Symptom Checklist in Second Edition of the International Classification of Headache Disorders Migraine Sufferers from the General Population (n = 11,194)
| Items | Communality Estimates |
Loading SE | ||
|---|---|---|---|---|
| Thermal Allodynia |
Static Mechanical Allodynia |
Dynamic Mechanical Allodynia |
||
| Combing your hair | 0.654 | 0.212 (0.018) | 0.041 (0.015) | 0.677 (0.032) a |
| Pulling your hair back | 0.998 | −0.176 (0.018) | −0.034 (0.020) | 1.071 (0.052)a |
| Shaving your face | 0.257 | 0.381 (0.013)a | 0.258 (0.011) | −0.093 (0.012) |
| Wearing eyeglasses | 0.352 | 0.182 (0.011) | 0.346 (0.011)a | 0.219 (0.011) |
| Wearing contact lenses | 0.195 | −0.011 (0.009) | 0.383 (0.011)a | 0.119 (0.010) |
| Wearing earrings | 0.776 | 0.040 (0.011) | 0.784 (0.015)a | 0.155 (0.011) |
| Wearing a necklace | 0.876 | 0.037 (0.012) | 0.874 (0.016)a | 0.095 (0.011) |
| Wearing tight clothing | 0.559 | 0.157 (0.011) | 0.575 (0.012)a | 0.161 (0.011) |
| Taking a shower | 0.686 | 0.622 (0.012)a | 0.165 (0.011) | 0.213 (0.011) |
| Resting your face or head on a pillow |
0.497 | 0.613 (0.012)a | 0.017 (0.010) | 0.166 (0.011) |
| Exposure to heat | 0.661 | 0.704 (0.013)a | 0.020 (0.011) | 0.194 (0.011) |
| Exposure to cold | 0.649 | 0.719 (0.013)a | 0.004 (0.010) | 0.171 (0.011) |
The measures that load on each factor.
SE = standard error. Values in parentheses are standard errors of factor loading.
Item Analysis
The first item analysis treated allodynia as a single underlying construct. The model demonstrated excellent item properties (data not shown). The discrimination score for an item measures how well it distinguishes those with high scores on the allodynia scale from those with low scores. Scores greater than 1 indicate items that are highly discriminating. We then applied a nonlinear mixed-effects model treating the three allodynia factors separately (mechanical static, mechanical dynamic, and thermal), as shown in Table 3. The items with the highest discrimination included “combing your hair” (discrimination =4.89), “pulling your hair back” (4.15), “wearing a necklace” (4.04), as well as “wearing earrings” (3.63) or “taking a shower” (3.02). Even the items with the lowest levels of discrimination, “resting your face or head on a pillow” (1.75) and “wearing contact lenses” (1.89), were highly discriminating.
Table 3.
Three-Factor Confirmatory Graded Item Response Model Analysis of 12-Item Allodynia Symptom Checklist in Individuals with Second Edition of the International Classification of Headache Disorders Migraine (n = 11,194)
| Item Stem | Mechanical Static | Thermal | Mechanical Dynamic | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Discrimination | Threshold 1 |
Threshold 2 |
Discrimination | Threshold 1 |
Threshold 2 |
Discrimination | Threshold 1 |
Threshold 2 |
|
| Wearing eyeglasses |
1.9514 | 0.1196 | 0.8856 | ||||||
| Wearing contact lenses |
1.8997 | 0.6416 | 1.1307 | ||||||
| Wearing earrings |
3.6352 | 1.223 | 1.6683 | ||||||
| Wearing a necklace |
4.0447 | 1.2601 | 1.6846 | ||||||
| Wearing tight clothing |
2.5968 | 0.7158 | 1.2445 | ||||||
| Shaving your face |
2.0753 | 1.2749 | 1.896 | ||||||
| Taking a shower | 3.0162 | 0.8284 | 1.4668 | ||||||
| Resting your face or head on a pillow |
1.756 | 0.3824 | 1.2418 | ||||||
| Exposure to heat |
2.6084 | 0.3216 | 1.0032 | ||||||
| Exposure to cold |
2.4227 | 0.5997 | 1.3494 | ||||||
| Combing your hair |
4.8975 | 0.1732 | 0.6593 | ||||||
| Pulling your hair back |
4.1534 | −0.02689 | 0.375 | ||||||
The discrimination score reflects how well the item distinguishes subjects with high levels on the factor from those with low levels on the factor. Larger discrimination parameters indicate stronger measurement of the latent variable. Threshold 1 contrasts “never” and “rarely” versus “less than half the time” as response option. Threshold 2 contrasts “never,” “rarely,” and “less than half the time” versus “half the time or more.”
Each of the items has three levels (never/rarely, less than half the time, and half the time or more). The first threshold (see Table 3) indicates how well an item discriminates severity of allodynia in respondents who say never/rarely versus less than half the time. For example, in the “exposure to heat” item, Threshold 1 is 0.32. This indicates that subjects whose level on the thermal allodynia factor was 0.32 or more standard deviations greater than the mean were more likely to endorse “less than half the time” than “never” or “rarely.” The second threshold of 1.0 indicated that subjects with a thermal allodynia factor score more than 1 standard deviation greater than the mean were more likely to say that exposure to heat increased their pain “half the time or more” than to say that it happened “less than half the time.”
The item responses span the thresholds from 0.28 standard deviation less than the mean (Threshold 1 for pulling your hair) to 1.89 standard deviations greater than the mean (Threshold 2 for shaving the face). The item response model factors were highly correlated, with interfactor correlations exceeding those observed in the exploratory factor analysis. Thermal allodynia correlated 0.76 with static mechanical allodynia, and 0.71 with dynamic mechanical allodynia. Static mechanical and dynamic mechanical allodynia were highly correlated as well, with an interfactor correlation of 0.69.
Allodynia Severity and Migraine Symptoms
Based on the distribution of scores, we developed a CA scale defining no CA (scores 0–2), mild CA (3–5), moderate CA (6–8), and severe CA (≥9). Using these scores, we found that 63.2% of migraine sufferers had CA. Allodynia was absent in 36.8% of migraine sufferers, mild in 25.1%, moderate in 17.1%, and severe in 20.4%.
We examined the relation of these categories of CA with the headache symptoms in our migraine sample (Table 4). Severe CA was associated with headache frequency. Although just 12.9% of those with less than 6 headaches in the past year had severe CA, 20.3% of those with headaches 13 to 24 days per year had it (odds ratio [OR], 1.8; 95% confidence interval [CI], 1.5–2.2). The proportion increased to 25.9% in those with headaches 104 to 179 days per year (OR, 2.5; 95% CI, 2.0=3.0). Severe CA was also more frequent in those with throbbing versus nonthrobbing migraines (25.9 vs 18.1%; OR, 2.3; 95% CI, 2.1=2.6), unilateral versus bilateral pain (24.9 vs 18.8%; OR, 2.2; 95% CI, 2.0=2.4), moderate or severe headaches versus mild headaches (25.8 vs 18%; OR, 2.9; 95% CI, 2.3–3.1), and being aggravated by physical activities versus not aggravated (29.3 vs 17.9%; OR, 2.8; 95% CI, 2.5–3.1).
Table 4.
Severity of Cutaneous Allodynia as Measured by the Allodynia Symptom Checklist-Questionnaire and Migraine Headache Features and Disability in Second Edition of the International Classification of Headache Disorders Migraine Sufferers from the General Population
| Headache Features |
None | Mild | Moderate | Severe | OR (95% CI) | ||||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | n | % | ||
| Total | 3,724 | 36.8 | 2,545 | 25.1 | 1,794 | 17.7 | 2,061 | 20.4 | |
| Headache frequency (attacks/yr)a |
|||||||||
| <6 | 431 | 54.4 | 167 | 21.1 | 92 | 11.6 | 102 | 12.9 | 1 |
| 6−12 | 840 | 42.5 | 452 | 22.9 | 348 | 17.6 | 336 | 17.0 | 1.6 (1.3–1.9) |
| 13−24 | 516 | 38.3 | 396 | 29.4 | 251 | 18.6 | 273 | 20.3 | 1.8 (1.5–2.2) |
| 24−51 | 1,136 | 35.4 | 819 | 25.5 | 589 | 18.4 | 664 | 20.7 | 2 (1.6–2.4) |
| 52−103 | 455 | 28.4 | 435 | 27.1 | 298 | 18.6 | 416 | 25.9 | 2.5 (2.0–3.0) |
| 104−179 | 205 | 28.0 | 175 | 23.9 | 162 | 22.2 | 189 | 25.9 | 2.5 (2.0–3.0) |
| Throbbinga | |||||||||
| No | 2,846 | 39.6 | 1,811 | 25.2 | 1228 | 17.0 | 1,297 | 18.1 | 1 |
| Yes | 878 | 29.8 | 734 | 24.9 | 566 | 19.2 | 764 | 25.9 | 2.3 (2.1–2.6) |
| Unilaterala | |||||||||
| No | 2,964 | 39.1 | 1,896 | 24.9 | 1,301 | 17.1 | 1,429 | 18.8 | 1 |
| Yes | 760 | 29.9 | 649 | 25.6 | 493 | 19.4 | 632 | 24.9 | 2.2 (2.0–2.4) |
| Aggravated by physical activitiesa |
|||||||||
| No | 3,163 | 39.5 | 2,031 | 25.4 | 1,366 | 17.1 | 1,439 | 17.9 | 1 |
| Yes | 561 | 26.4 | 514 | 24.2 | 428 | 20.1 | 622 | 29.3 | 2.8 (2.5–3.1) |
| Moderate or severea | |||||||||
| Yes | 2,822 | 29.8 | 1,809 | 24.3 | 1,187 | 20.0 | 1,280 | 25.8 | 1 |
| No | 902 | 39.7 | 736 | 25.4 | 607 | 16.7 | 781 | 18.0 | 2.4 (2.2–2.7) |
| Disability | |||||||||
| MIDAS I | 2,174 | 46.7 | 1,153 | 24.8 | 698 | 15.0 | 630 | 13.5 | 1 |
| MIDAS II | 620 | 32.3 | 499 | 26.0 | 401 | 20.9 | 397 | 20.7 | 1.7 (1.5–1.9) |
| MIDAS III | 459 | 27.3 | 455 | 27.1 | 333 | 19.8 | 435 | 25.9 | 2.2 (1.9–2.5) |
| MIDAS IV | 280 | 19.9 | 333 | 23.7 | 299 | 21.3 | 493 | 35.1 | 3.4 (3.0–3.9) |
Complete data for these analyses were obtained from 10,124 migraineurs.
For clarity and brevity, the “do not know–do not remember” categories were omitted.
OR = odds ratio; CI = confidence interval; MIDAS = Migraine Disability Assessment questionnaire.
Severe CA was associated with the level of disability experienced by migraineurs, as measured by the Migraine Disability Assessment (MIDAS) questionnaire (see Table 4). Whereas 13.5% of those with no disability had severe CA, the proportion increased to 20.7% in those with mild disability (OR, 1.7; 95% CI, 1.5– 1.9), to 25.9% in those with moderate disability (OR, 2.2; 95% CI, 1.9 –2.5), and to 35.1% in the severely disabled (OR, 3.4; 95% CI, 3.0 –3.9).
The presence of severe CA was also related with the presence of associated symptoms in individuals with migraine (Table 5). It was more common in those who typically experience nausea (27.1 vs 19.2%; OR, 2.3; 95% CI, 2.1–2.6), photophobia (26.9 vs 18.9%; OR, 2.4; 95% CI, 2.1–2.7), and phonophobia (26.9 vs 17.7%; OR, 2.4; 95% CI, 2.1–2.7).
Table 5.
Severity of Cutaneous Allodynia, as Measured by the Allodynia Symptom Checklist-Questionnaire and Associated Symptoms in Second Edition of the International Classification of Headache Disorders Migraine Sufferers from the General Population
| Associated Symptoms |
None | Mild | Moderate | Severe | OR (95% CI) for Severe Allodynia |
||||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | n | % | ||
| Nausea | |||||||||
| No | 3,298 | 38.0 | 2,192 | 25.2 | 1,518 | 17.4 | 1,669 | 19.2 | 1 |
| Yes | 426 | 29.4 | 353 | 24.4 | 276 | 19.1 | 392 | 27.1 | 2.3 (2.1−2.6) |
| Photophobia | |||||||||
| No | 3,209 | 38.6 | 2,080 | 25.1 | 1,438 | 17.3 | 1,567 | 18.9 | 1 |
| Yes | 515 | 28.1 | 465 | 25.4 | 356 | 19.4 | 494 | 26.9 | 2.4 (2.1−2.7) |
| Phonophobia | |||||||||
| No | 3,127 | 38.9 | 3,030 | 25.1 | 1,391 | 17.3 | 1,500 | 18.7 | 1 |
| Yes | 597 | 28.6 | 525 | 25.1 | 403 | 19.3 | 561 | 26.9 | 2.4 (2.1−2.7) |
| Aura | |||||||||
| No | 2,831 | 43.5 | 1,690 | 25.9 | 1,060 | 16.2 | 927 | 14.2 | 1 |
| Yes | 893 | 24.7 | 855 | 23.6 | 734 | 20.2 | 1,134 | 31.3.2 | 3.5 (3.2−3.8) |
OR = odds ratio; CI = confidence interval.
Finally, a greater proportion of individuals with migraine with aura (31.3%) experienced severe CA compared with migraineurs without aura (14.2%) (OR,3.5; 95% CI, 3.2–3.8).
Multivariate Analyses of Allodynia in Migraineurs
Table 6 summarizes predictors of CA in a series of Poisson regression models. Among migraineurs, CA was more common in Asians and Native Americans. Women had a greater prevalence of CA than men in all three models (Model 3: prevalence ratio [PR], 1.43; 95% CI, 1.28 –1.59).
Table 6.
Prevalence of Allodynia and Adjusted Prevalence Ratios in Individuals with Migraine
| Characteristics | With Allodynia |
Total Sample, N |
PR (95% CI) | ||||
|---|---|---|---|---|---|---|---|
| n | % | Model 1 | Model 2 | Model 3 | |||
| Race | |||||||
| No answer | 98 | 38.58 | 275 | Not calculated | Not calculated | Not calculated | |
| White | 3,307 | 37.43 | 9,696 | 1 | 1 | 1 | |
| Black | 308 | 42.66 | 784 | 1.14 (1.04–1.25) | 1.07 (0.98–1.17) | 1.08 (0.97–1.20) | |
| Asian | 32 | 40 | 86 | 1.07 (0.82–1.40) | 1.18 (0.91–1.54) | 1.28 (1.02–1.61) | |
| Native American | 44 | 51.16 | 96 | 1.37 (1.11–1.68) | 1.33 (1.08–1.63) | 1.32 (1.07–1.64) | |
| Other | 66 | 44.9 | 157 | 1.20 (1.00–1.44) | 1.15 (0.96–1.37) | 1.10 (0.88–1.39) | |
| Sex | |||||||
| Male | 476 | 24.56 | 2,192 | 1 | 1 | 1 | |
| Female | 3,379 | 41.28 | 8,902 | 1.68 (1.55–1.82) | 1.68 (1.54–1.82) | 1.43 (1.28–1.59) | |
| Highest education level | |||||||
| Junior high or less | 55 | 48.67 | 117 | 1 | 1 | 1 | |
| Some high school | 202 | 44.59 | 501 | 0.92 (0.74–1.14) | 0.82 (0.67–1.01) | 0.88 (0.67–1.16) | |
| HSD or GED | 914 | 38.19 | 2,628 | 0.78 (0.65–0.95) | 0.68 (0.57–0.82) | 0.80 (0.62–1.02) | |
| Some college | 1,614 | 41.34 | 4,256 | 0.85 (0.70–1.03) | 0.73 (0.61–0.88) | 0.85 (0.66–1.08) | |
| Bachelor’s degree | 665 | 32.52 | 2,259 | 0.67 (0.55–0.82) | 0.58 (0.48–0.69) | 0.69 (0.53–0.89) | |
| Graduate school | 371 | 32.92 | 1,232 | 0.68 (0.55–0.83) | 0.61 (0.50–0.75) | 0.71 (0.55–0.92) | |
| Age, yr | |||||||
| 18–24 | 185 | 37.91 | 541 | 1 | 1 | 1 | |
| 25–34 | 682 | 40.77 | 1,798 | 1.08 (0.95–1.22) | 0.99 (0.88–1.13) | 1.02 (0.87–1.20) | |
| 35–44 | 959 | 39.32 | 2,659 | 1.04 (0.92–1.17) | 0.97 (0.86–1.09) | 0.96 (0.82–1.14) | |
| 45–54 | 1,113 | 38.55 | 3,168 | 1.02 (0.90–1.15) | 0.95 (0.84–1.07) | 0.90 (0.77–1.07) | |
| 55–64 | 636 | 36.87 | 1,917 | 0.97 (0.85–1.11) | 0.90 (0.79–1.02) | 0.86 (0.72–1.02) | |
| 65–74 | 199 | 31.54 | 700 | 0.83 (0.71–0.98) | 0.78 (0.67–0.92) | 0.78 (0.62–0.97) | |
| >74 | 81 | 28.83 | 311 | 0.76 (0.61–0.94) | 0.66 (0.53–0.84) | 0.64 (0.43–0.95) | |
| Illness duration, yr | |||||||
| <10 | 648 | 43.37 | 1,611 | 1 | 1 | 1 | |
| 10–19 | 594 | 49.13 | 1,304 | 1.13 (1.04–1.23) | 1.16 (1.07–1.27) | 1.15 (1.06–1.25) | |
| 20–29 | 363 | 46.9 | 831 | 1.08 (0.98–1.19) | 1.22 (1.10–1.35) | 1.19 (1.08–1.32) | |
| 30–39 | 145 | 43.81 | 356 | 1.01 (0.88–1.16) | 1.21 (1.05–1.41) | 1.22 (1.05–1.43) | |
| 40–49 | 63 | 47.01 | 149 | 1.08 (0.90–1.31) | 1.31 (1.04–1.66) | 1.28 (1.01–1.64) | |
| 50–59 | 663 | 41.44 | 1,730 | 0.96 (0.88–1.04) | 0.95 (0.88–1.04) | 0.95 (0.87–1.03) | |
| >60 | 20 | 42.55 | 53 | 0.98 (0.70–1.37) | 1.46(0.95–2.26) | 1.12 (0.63–1.98) | |
| Attack frequency, attacks/yr |
|||||||
| <6 | 194 | 24.49 | 891 | 1 | 1 | 1 | |
| 6–12 | 684 | 34.62 | 2,181 | 1.41 (1.23–1.62) | 1.37 (1.19–1.58) | 1.21 (1.00–1.46) | |
| 13–24 | 524 | 36.49 | 1,577 | 1.49 (1.30–1.71) | 1.41 (1.22–1.62) | 1.19 (0.98–1.44) | |
| 24–51 | 1,253 | 39.06 | 3,507 | 1.59 (1.40–1.82) | 1.51 (1.32–1.72) | 1.15 (0.96–1.39) | |
| 52–103 | 714 | 44.51 | 1,744 | 1.82 (1.59–2.08) | 1.67 (1.46–1.92) | 1.19 (0.98–1.44) | |
| ≥104 | 351 | 48.02 | 781 | 1.96 (1.70–2.26) | 1.40 (1.31–1.51) | 1.22 (1.00–1.50) | |
| Disability | |||||||
| No | 1,328 | 28.53 | 5,185 | 1 | 1 | 1 | |
| Mild | 798 | 41.63 | 2,113 | 1.46 (1.36–1.56) | 1.51 (1.41–1.62) | 1.32 (1.21–1.45) | |
| Moderate | 768 | 45.66 | 1,796 | 1.72 (1.54–1.91) | 1.60 (1.49–1.72) | 1.88 (1.76–2.00) | |
| Severe | 792 | 56.37 | 1,483 | 1.98 (1.85–2.11) | 1.26 (1.17–1.37) | 1.61 (1.46–1.78) | |
| Body mass index | |||||||
| Normal | 1,141 | 35.34 | 3,547 | 1 | 1 | 1 | |
| Overweight | 1,039 | 37.01 | 3,085 | 1.05 (0.98–1.12) | 1.15 (1.07–1.24) | 1.11 (1.03–1.20) | |
| Obese | 728 | 39.54 | 2,014 | 1.12 (1.04–1.20) | 1.22 (1.14–1.31) | 1.16 (1.06–1.26) | |
| Morbidly obese | 816 | 42.86 | 2,063 | 1.21 (1.13–1.30) | 1.40 (1.32–1.48) | 1.11 (1.02–1.21) | |
In Model 1, adjustment was conducted for demographic variables (eg, age, sex, race, income). In Model 2, adjustment was made for demographics and also headache frequency, severity, and duration of illness. In Model 3, we used the same adjustments for Model 2 and also included comorbidities, use of preventive medication, and use of opioids.
PR = prevalence ratio; CI = confidence interval; HSD = high school diploma; GED = general equivalency diploma.
CA did not vary as a function of education level, and its relative frequency decreased with age over 65 in all 3 models; whereas 28.8% of those migraineurs 75 years or older had CA, 37.9% of those aged 18 to 24 had CA (Model 3: PR, 0.64; 95% CI, 0.43=0.95).
The prevalence of CA increased in those with long duration of illness relative to those with migraine for less than 10 years, peaking after 40 to 49 years of illness (PR, 1.28; 95% CI, 1.01–1.64). CA was also associated with high frequency of attacks (from 24.9% in those with less than 6 migraine attacks/year to 48% in OR, those with 2–3 attacks/week) and higher disability (Migraine Disability Assessment IV vs I: PR, 1.61; 95% CI, 1.46 –1.75). Finally, CA was more common in those with increased body mass index in all models (see Table 6).
Discussion
We modified Jakubowski and colleagues’14 allodynia questionnaire and administered it to a population sample of severe headache sufferers. In this article, we focus on results obtained in more than 11,000 migraine sufferers from the general population. Our goals were to refine the measurement of CA using a questionnaire, to estimate the prevalence and severity of CA in the migraine population, to develop measures for allodynia subtypes (thermal, static mechanical, and dynamic mechanical), and to contrast migraine sufferers with and without CA on the ASC. In this discussion, we first consider the scoring of the ASC instrument and then the relation of ASC scores to the features of migraine. Next, we discuss the three subscales that emerge from factor analysis corresponding with different types of allodynia. After considering the limitations of this study, we discuss its implications and directions for future research.
In the questionnaire that Jakubowski and colleagues’ 14 developed, subjects who gave a positive response to any of the 12 CA questions were considered to have allodynia. The sensitivity of the questionnaire was 84.8%, whereas the specificity was 52.2%, using QST as a gold standard. If we accept the gold standard, the low specificity may reflect false-positive symptom reporting, meaning that nonallodynic patients report allodynia symptoms. Based on prior work, to reduce false-positives, we provided graded response options and considered “rarely” as a negative response. 17–19 These graded responses also created an opportunity to measure the severity of allodynia.
The psychometric analysis confirms that CA can be considered as a quantitative trait. This quantitative perspective is consistent with clinical impressions of allodynia.Some patients do not develop allodynia at all. Others develop it for some attacks, particularly if they are left untreated for a period of time. In chronic migraine, allodynia may be present between attacks.26 Thus, duration and severity of allodynia may depend both on within-person biological factors and environmental factors, including treatment strategy.2,3,26 Our goal was to develop a measure that summarized a patient’s allodynia experience over multiple attacks. Quantitative response options lend themselves to the development of a robust allodynia scale.
Using the ASC, we found that in the population, CA was present in about 63.2% of migraine sufferers, a number remarkably consistent with estimates based on QST in clinical samples.12,13 The convergence of these estimates lends credibility to our findings. This result also demonstrates that allodynia is a common feature of migraine, not just in specialty care but in the population. We found a distribution of scores: 36.8% had no allodynia, and 25.1% had mild, 17.7% had moderate, and 20.4% had severe CA using our somewhat arbitrary categorical scale.
The presence and severity of allodynia was associated with many aspects of migraine. The odds of CA more than doubled in migraineurs with nausea, photophobia, and phonophobia. This association has not been reported previously in representative sample of migraineurs. Particularly striking was the relation between allodynia and headache frequency. As headache frequency increased, the proportion of migraine sufferers with severe allodynia also increased. In a cross-sectional study, it is not possible to disentangle causal sequence. However, these findings are consistent with the idea that repeated attacks of migraine lead to the development of allodynia.27,28 On the other hand, allodynia may be a marker of risk for frequent and severe attacks. We plan to explore these possibilities in the longitudinal data from AMPP. Finally, the odds of severe CA increased 3.5-fold in migraine with aura. Because the association is stronger for aura than for other associated symptoms, aura may have a stronger link with CA. Studies demonstrate that cortical spreading depression, the physiological basis of aura, can activate brainstem regions involved in the processing of nociceptive information via trigeminovascular mechanisms.29,30 Perhaps this process lowers the threshold of nociceptive neurons at the trigeminal nucleus caudalis and thalamus, predisposing to CA.
Though allodynia is sometimes considered a unitary factor, physiological research suggests that there are several subdomains of allodynia.14 Our exploratory factor analysis demonstrated factors that correspond to the subdomains of thermal, mechanical dynamic, and mechanical static allodynia. Thermal allodynia (mediated by C nociceptive and Aδ nociceptive fibers) is measured by questions that tap sensitivity to heat and cold (eg, taking a shower, washing your face).14 It is interesting to note that shaving loaded with the thermal factor, perhaps because hot water is often used when shaving. Mechanical dynamic or brush allodynia is likely mediated by Aβ mechanoreceptors, which are capsaicin insensitive; it was measured by questions about combing hair and pulling the hair back.14 Mechanical static or pressure allodynia is mediated by Aδ nociceptive fibers; it loaded on questions about pain while wearing earrings, necklace, or tight clothing. These subtypes of allodynia can be assessed in relation to direct measures of psychophysical thresholds using QST. It is not clear whether these different types of allodynia as measured by ASC are differentially associated with clinically important outcomes such as poor response to acute treatment late in the migraine attack or to headache progression.14 In future work, we will map these subtypes of allodynia onto domain-specific QST, as well as clinical features and outcomes.
This study should be interpreted with caution for several reasons. First, we have not yet compared allodynia classification based on ASC with classification based on QST. Though QST is sometimes regarded as the gold standard for determining whether a patient has allodynia at a particular point in time, it is subject to temporal sampling error.13,14 For example, if testing is conducted 5 minutes into an attack, and CA emerges just 2 hours into the attack, CA will not be detected by QST. Therefore, it may also be of greater clinical interest to validate the ASC against either physiological measures or patient outcomes. Second, this report does not examine the prevalence of CA as defined by the ASC in individuals with other types of headache; in a separate report, we show that the prevalence of allodynia based on the ASC is greatest in transformed migraine, followed by migraine and probable migraine.28 Third, measuring the severity of symptoms and signs for chronic disorders with episodic manifestations such as migraine or epilepsy is challenging. Chronic disorders with episodic manifestations are characterized by episodic attacks superimposed on an enduring predisposition to attacks.31 In measuring migraine severity, pain intensity, number and severity of associated symptoms, and disability all index the severity of a particular attack. Because attacks are transient and recurrent, time enters into the assessment of severity. ASC measures severity in the conventional sense by tapping into a range of partially correlated symptoms and summarizing them. To assess the severity of allodynia at a particular point in time, we ascertained the presence or absence of each ASC item at that point in time. But to ascertain the overall severity of allodynia, we included a measure of frequency for each of the 12 items. Thus, our ASC score is based on the assumption that both number and frequency of symptoms can be scaled together to measure severity. The psychometric validation using item response theory provides strong support for this assumption. Severity measures for other chronic disorders with episodic manifestations, such as asthma, combine symptom number and frequency.32 In addition, the definitions of mild, moderate, and severe allodynia are somewhat arbitrary. In longitudinal analysis, we will assess the relation between ASC score and headache prognosis to develop an empirical foundation for cut scores. Finally, though allodynia is associated with attack frequency at cross section, its relation with headache progression has not yet been assessed.
Despite these limitations, a simple, quantitative tool for assessing allodynia should find applications in research and, we hope, in clinical practice. If CA predicts response to triptan therapy,6 knowing whether a patient has allodynia has clinical implications. In addition, migraine is now viewed as a sometimes progressive disorder that may lead to chronic migraine.31,33 Because migraine progresses to chronic migraine in some but not most individuals, identifying the risk factors for progression has emerged as an important public health priority. We have divided risk factors for progression into those that are modifiable (eg, obesity, attack frequency) and those that are nonmodifiable (eg, female sex).27 CA may be a modifiable risk factor for migraine progression if pharmacological or nonpharmacological strategies reduce the frequency of CA by reducing the frequency of attacks.
Acknowledgments
This study was sponsored by the National Headache Foundation, which is supported by Ortho-McNeil Neurologics (grant number 991836659 [R.L. M.B., D.S., M.R.]
We thank S. Simons, Dr K. M. Fanning, and K. Ward for help with data management and statistical analyses.
Appendix
The members of the AMPP advisory group are Richard B. Lipton, MD (principal investigator), Marcelo E. Bigal, MD, PhD, Dawn Buse, MD, Michael L. Reed, PhD, Walter Stewart, PhD, Merle Diamond, MD, Frederick Freitag, DO, Elisabeth Hazard, PhD, Jono-than Tierce, CPhil, Elizabeth Loder, MD, Paul Winner, MD, Stephen Silberstein, MD, Suzanne Simons, and Seymour Diamond, MD.
Footnotes
Members of the American Migraine Prevalence and Prevention Advisory Group are listed in the Appendix on page xx.
References
- 1.Woolf CJ, Wall PD. Relative effectiveness of C primary afferent fibers of different origins in evoking a prolonged facilitation of the flexor reflex in the rat. J Neurosci. 1986;6:1433–1442. doi: 10.1523/JNEUROSCI.06-05-01433.1986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Burstein R, Cutrer MF, Yarnitsky D. The development of cutaneous allodynia during a migraine attack: clinical evidence for the sequential recruitment of spinal and supraspinal nociceptive neurons in migraine. Brain. 2000;123:1703–1709. doi: 10.1093/brain/123.8.1703. [DOI] [PubMed] [Google Scholar]
- 3.Sessle BJ, Hu JW, Amano N, Zhong G. Convergence of cutaneous, tooth pulp, visceral, neck and muscle afferents onto nociceptive and non-nociceptive neurones in trigeminal subnucleus caudalis (medullary dorsal horn) and its implications for referred pain. Pain. 1986;27:219–235. doi: 10.1016/0304-3959(86)90213-7. [DOI] [PubMed] [Google Scholar]
- 4.Selby G, Lance JW. Observations on 500 cases of migraine and allied vascular headache. J Neurol Neurosurg Psychiatry. 1960;23:23–32. doi: 10.1136/jnnp.23.1.23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Burstein R, Yarnitsky D, Goor-Aryeh I, et al. An association between migraine and cutaneous allodynia. Ann Neurol. 2000;47:614–624. [PubMed] [Google Scholar]
- 6.Burstein R, Collins B, Jakubowski M. Defeating migraine pain with triptans: a race against the development of cutaneous allodynia. Ann Neurol. 2004;55:19–26. doi: 10.1002/ana.10786. [DOI] [PubMed] [Google Scholar]
- 7.Bigal ME, Lipton RB. Modifiable risk factors for migraine progression. Headache. 2006;46:1334–1343. doi: 10.1111/j.1526-4610.2006.00577.x. [DOI] [PubMed] [Google Scholar]
- 8.Burstein R, Yamamura H, Malick A, Strassman AM. Chemical stimulation of the intracranial dura induces enhanced responses to facial stimulation in brain stem trigeminal neurons. J Neurophysiol. 1998;79:964–982. doi: 10.1152/jn.1998.79.2.964. [DOI] [PubMed] [Google Scholar]
- 9.Bigal ME, Lipton RB, Reed M, et al. Prevalence and characteristics of allodynia in headache sufferers: results from the American Prevalence and Prevention Study. Neurology. doi: 10.1212/01.wnl.0000310645.31020.b1. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bahra A, Matharu MS, Buchel C, et al. Brainstem activation specific to migraine headache. Lancet. 2001;357:1016–1017. doi: 10.1016/s0140-6736(00)04250-1. [DOI] [PubMed] [Google Scholar]
- 11.Young WB, Richardson ES, Shukla P. Brush allodynia in ospitalized headache patients. Headache. 2005;45:999–1003. doi: 10.1111/j.1526-4610.2005.05180.x. [DOI] [PubMed] [Google Scholar]
- 12.Ashina S, Bendtsen L, Ashina M, et al. Generalized hyperalgesia in patients with chronic tension-type headache. Cephalalgia. 2006;26:940–948. doi: 10.1111/j.1468-2982.2006.01150.x. [DOI] [PubMed] [Google Scholar]
- 13.Mathew NT, Kailasam J, Seifert T. Clinical recognition of allodynia in migraine. Neurology. 2004;63:848–852. doi: 10.1212/01.wnl.0000137107.27585.f7. [DOI] [PubMed] [Google Scholar]
- 14.Jakubowski M, Silberstein S, Ashkenazi A, Burstein R. Can allodynic migraine patients be identified interictally using a questionnaire? Neurology. 2005;65:1419–1422. doi: 10.1212/01.wnl.0000183358.53939.38. [DOI] [PubMed] [Google Scholar]
- 15.Pitcher GM, Henry JL. Cellular mechanisms of hyperalgesia and spontaneous pain in a spinalized rat model of peripheral neuropathy: changes in myelinated afferent inputs implicated. Eur J Neurosci. 2000;12:2006–2020. doi: 10.1046/j.1460-9568.2000.00087.x. [DOI] [PubMed] [Google Scholar]
- 16.Lipton RB, Bigal ME, Diamond M, et al. Migraine prevalence, disease burden and the need for preventive therapy. Neurology. 2007;68:343–349. doi: 10.1212/01.wnl.0000252808.97649.21. [DOI] [PubMed] [Google Scholar]
- 17.Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders: 2nd edition. Cephalalgia. 2004;24 suppl 1:9–160. doi: 10.1111/j.1468-2982.2003.00824.x. [DOI] [PubMed] [Google Scholar]
- 18.Lipton RB, Stewart WF, Diamond S, et al. Prevalence and burden of migraine in the United States: data from the American Migraine Study II. Headache. 2001;41:646–657. doi: 10.1046/j.1526-4610.2001.041007646.x. [DOI] [PubMed] [Google Scholar]
- 19.Lipton RB, Stewart WF, Simon D. Medical consultation for migraine: results from the American Migraine Study. Headache. 1998;38:87–96. doi: 10.1046/j.1526-4610.1998.3802087.x. [DOI] [PubMed] [Google Scholar]
- 20.Bigal ME, Sheftell FD, Rapoport AM, et al. Chronic migraine is the early stage of transformed migraine in adults. Neurology. 2005;65:1556–1561. doi: 10.1212/01.wnl.0000184477.11569.17. [DOI] [PubMed] [Google Scholar]
- 21.Little RJA, Rubin DB. Hoboken, NJ: John Wiley & Sons; 2002. Statistical analysis with missing data. [Google Scholar]
- 22.Samejima F. Psychometrika Monograph Supplement, No. 17. Richmond, VA: The William Byrd Press; 1969. Estimating latent ability using a pattern of graded scores; p. 14. [Google Scholar]
- 23.Ake CF. Rounding after multiple imputation with non-binary categorical covariates; Paper presented at: 30th annual SUGI meeting; Philadelphia, PA. 2005. [Google Scholar]
- 24.Browne MW, Cudeck R, Tateneni K, Mels G. [Last accessed June 28, 2007];CEFA: Comprehensive Exploratory Factor Analysis, version 2.00 [computer software and manual] Available at http://quantrm2.psy.ohio-state.edu/browne.
- 25.Schilling S, Bock RD. High-dimensional maximum marginal likelihood item factor analysis by adaptive quadrature. Psychometrika. 2005;70:533–555. [Google Scholar]
- 26.Thissen D, Chen W-H, Bock RD. Multilog (version 7) [computer software] Lincolnwood, IL: Scientific Software International; 2003. [Google Scholar]
- 27.Young WB, Richardson ES, Shukla P. Brush allodynia in hospitalized headache patients. Headache. 2005;45:999–1003. doi: 10.1111/j.1526-4610.2005.05180.x. [DOI] [PubMed] [Google Scholar]
- 28.Haut SR, Bigal ME, Lipton RB. Chronic disorders with episodic manifestations: focus on epilepsy and migraine. Lancet Neurol. 2006;5:148–157. doi: 10.1016/S1474-4422(06)70348-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Moskowitz MA, Nozaki K, Kraig RP. Neocortical spreading depression provokes the expression of c-fos protein-like immuno-reactivity within trigeminal nucleus caudalis via trigeminovascular mechanisms. J Neurosci. 1993;13:1167–1177. doi: 10.1523/JNEUROSCI.13-03-01167.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Bolay H, Reuter U, Dunn AK, et al. Intrinsic brain activity triggers trigeminal meningeal afferents in a migraine model. Nat Med. 2002;8:136–142. doi: 10.1038/nm0202-136. [DOI] [PubMed] [Google Scholar]
- 31.Plaza V, Serra-Batles J, Comella A, Badiola C. Differences in asthma clinical outcomes according to initial severity. J Asthma. 2005;42:207–211. [PubMed] [Google Scholar]
- 32.Bigal ME, Lipton RB. When migraine progresses: transformed or chronic migraine. Expert Rev Neurother. 2006;6:297–306. doi: 10.1586/14737175.6.3.297. [DOI] [PubMed] [Google Scholar]
- 33.Lipton RB, Pan J. Is migraine a progressive brain disease? JAMA. 2004;291:493–494. doi: 10.1001/jama.291.4.493. [DOI] [PubMed] [Google Scholar]