Validation of an 8-item self-administered questionnaire for assessing migraine-related sensory hypersensitivities (MHQ-8)

Abstract

Background

In addition to headache, migraine patients often experience sensory hypersensitivity to external stimuli. While photophobia and phonophobia are part of the diagnostic criteria of migraine, many patients also exhibit cutaneous allodynia and osmophobia. However, the presence and intensity of these four hypersensitivities are rarely assessed systematically and simultaneously due to the lack of a simple and rapid self-report questionnaire.

Methods

We have identified existing questionnaires for allodynia, photophobia and phonophobia and selected one of each, that were translated in French and validated (according to COSMIN’s recommendations). We also proposed a 2-item questionnaire (presence and intensity) for each of these 3 hypersensitivities plus osmophobia, resulting in the 8-item Migraine Hypersensitivity Questionnaire (MHQ-8) exploring these four hypersensitivities. In addition, the headache impact test (HIT-6), the migraine disability assessment (MIDAS) and the hospital anxiety and depression scale (HADS) were also answered. The survey was conducted in Pain and Neurology departments during specialised consultations for headaches. Content validity, structural validity, internal consistency, transcultural validity, reliability, criterion validity, construct validity and responsiveness were tested. Confirmatory factor analysis (CFA) was used to test the dimensionality of the questionnaires.

Results

The study sample consisted in N = 329 patients with a mean age of 43.7 ± 13.2 and a mean number of 10.2 ± 7.0 migraine days per month; 84% of them were women and 27% had chronic migraine. Overall, 312 to 327 questionnaires were usable for the hypersensitivity questionnaires. The reliability of the MHQ-8 was good to excellent with a Cronbach’s alpha of α = 0.88 (photophobia), α = 0.89 (phonophobia), α = 0.91 (allodynia), α = 0.95 (osmophobia), and α = 0.85 for the whole questionnaire. The intraclass correlation coefficient assessing test-retest reliability was 0.83, 0.77, 0.87, and 0.90, respectively; it was 0.88 for the whole questionnaire. The factor analysis on the MHQ-8 items showed excellent exploratory results, and the indicators of the CFA showed good performances with CFI and TLI at 0.999, RMSEA at 0.054 and SRMR at 0.021.

Conclusions

The MHQ-8 developed in this study is valid and reliable. It serves as a new diagnostic tool for the four sensory hypersensitivities that can occur during migraine attacks and may be useful in both clinical research and daily practice.

Supplementary Information

The online version contains supplementary material available at 10.1186/s10194-025-02067-3.

Introduction

The prevalence of migraine is currently around 15% worldwide, with a major impact on years lived with disabilities [1]. During a migraine attack, patients suffer not only from headache, but also from other debilitating symptoms such as photophobia (light aversion or discomfort to visual stimulations, or an exacerbation of headache by those same stimulations) and phonophobia (sound aversion or discomfort, or an exacerbation of headache by auditory stimulations). These symptoms are regarded as sensory hypersensitivity markers to environmental stimuli and are taken into account as diagnostic criteria [2]. Other hypersensitivity symptoms associated with migraine include cutaneous allodynia (pain in response to a normally non-painful stimulus) and osmophobia (aversion to or discomfort from odors, or exacerbation of headache by olfactory stimuli). These last two signs are rarely evaluated in clinical practice or clinical trials, even though osmophobia could represent a specific clinical marker of migraine [3]. Schwedt [4] argued that because migraine symptoms such as headache and hypersensitivity often occur together and may influence each other, it is insufficient to assess each symptom separately. Examining the symptom complex as a whole, and their interactions, has the potential to improve clinical and scientific understanding of migraine.

Previous studies using various questionnaires have confirmed that more than 90% of migraine patients report photophobia during attacks [5], more than 90% report phonophobia [5] and more than 40% report olfactory hypersensitivity [3, 6, 7]. In a study involving more than 6000 migraine patients in the USA, photophobia was the most bothersome symptom [5]. In the migraine population, the prevalence of cutaneous allodynia has been estimated to be between 40% and 65%, with a higher prevalence in patients with chronic migraine compared with episodic migraine [8–10]. Interestingly, these hypersensitivities can occur individually or in combination in migraine patients, leading to diverse migraine phenotypes. However, to date, there is no concise, reliable and valid questionnaire to rapidly assess both the presence and intensity of the four types of hypersensitivity simultaneously. While a validated 12-item questionnaire for cutaneous allodynia exists in English, it has not yet been validated in French [8]. A previous French study assessed self-reported sensitivity to visual, auditory and olfactory stimuli, but did not include questions on cutaneous allodynia [11]. Furthermore, there is no widely accepted and well-validated questionnaire to easily detect and measure the three other hypersensitivities, photophobia, phonophobia and osmophobia in migraine patients.

The lack of a comprehensive questionnaire for assessing hypersensitivities in migraine poses several challenges. In clinical practice, patients may under-report or be unaware of their symptoms, complicating monitoring and treatment evaluation. In research, inconsistent data hinder the identification of symptom correlations and disease severity. This absence of stratification obscures potential subgroups with distinct mechanisms, limiting personalized treatment approaches and response predictions.

The aim of this study was to develop a concise questionnaire to assess the four types of hypersensitivity that may occur during a migraine attack, including cutaneous allodynia, photophobia, phonophobia and osmophobia. To achieve this, we first translated into French and validated existing questionnaires for cutaneous allodynia [8], photophobia [12] and phonophobia [13] in migraine patients. We then used those questionnaires as standard to further validate the 8-item Migraine Hypersensitivity Questionnaire (MHQ-8).

Materials and methods

General considerations

The validation study of the 8-item Migraine Hypersensitivity Questionnaire (MHQ-8) has been conducted through a multicentre prospective observational study conducted across twelve hospitals in France within Pain and Neurology departments during specialised consultations for headaches. It had been approved by a referent ethics committee (Comité de Protection des Personnes Sud-Ouest et Outremer III, ID-RCB 2018-A03108-47), and registered at ClinicalTrials.gov under the reference NCT04580641. Eligible participants were adults aged 18 years or older who met the ICHD-3 diagnostic criteria for migraine [2]. Written informed consent was obtained from all participants. The processes of translation and validation followed recommendations of Beaton and Bullinge [14] and of the COSMIN (COnsensus-based Standards for the selection of health Measurement INstruments) group [15]. The evaluation of the measurement properties of the questionnaire was conducted in accordance with the COSMIN guidelines. This included the assessment of content validity, structural validity, internal consistency, transcultural validity, reliability, criterion validity, construct validity and responsiveness.

Available sources and translation processes

An 8-item questionnaire (Photo-8) has been developed in 2009 by a Korean team to assess photophobia in migraine [12]. Each item is quoted as absent/present (0/1) and the total score is the sum of all, ranging from 0 to 8. A score greater than or equal to 1 is considered sufficient to define photophobia. Firstly, two French native persons independently translated the English version provided by the Korean authors into French. Afterwards, a third French native bilingual translator compared the two translations and checked for inconsistency and adapted the vocabulary. Any discrepancies and differences were resolved by discussion, and finally a consensus version was reached. Finally, a native Korean speaker fluent in both English and French checked the concordance between the original version and the French one. The validated French version can be found in Appendix 1.

A 14-item questionnaire (Phono-14) has been developed in 2002 by a French team to diagnose hyperacusis, although it was not for migraine patients [13]. Each item is quoted from 0 to 3 [0,1,2,3], and the total score is the sum of all, ranging from 0 to 42. Severe hyperacusis is defined as a score of at least 28.

The 12-item Allodynia Symptom Checklist (ASC-12) has been developed in 2007 by a US team to assess cutaneous allodynia in migraine [8]. Each item is quoted from 0 to 2 [0,0,0,1,2] and the total score is the sum of all, ranging from 0 to 24. Allodynia is then categorised as mild (3 to 5), moderate (6 to 8) or severe (9 or more). The building of a French version was the same as for the photophobia questionnaire, and a backward translation was conducted by two people, including a professional native English speaker, unaware of the concepts explored. The authors checked the concordance between the original version and the French one. The validated French version can be found in Appendix 2.

Creation of the 8-item migraine hypersensitivity questionnaire (MHQ-8)

This questionnaire has been created by two physicians currently involved in care and clinical research in migraine, and aware of the abovementioned questionnaires. In the same way for each domain of hypersensitivity, a 2-item sub-questionnaire was built. The four domains consecutively addressed were cutaneous allodynia (defined by “disturbance by skin stimulation such as when brushing hair, wearing eyeglasses, or wearing tight clothes”), photophobia (defined by “disturbance by bright light of luminous stimulation”), phonophobia (defined by “disturbance by noise, e.g. noisy environment, busy place like a bar, shouting children…”), and osmophobia (defined by disturbance by smells usually acceptable such as fragrance, cooking, cleaning products…”). The first item focused on the frequency of the presence of the hypersensitivity symptom during migraine attacks (“During your worst headaches, how often are you disturbed by [the symptom]?”) and the possible answers were “never”, “rarely”, “sometimes”, “very often”, and “all the time”, to be quoted 0/1/2/3/4 respectively. The second item focused on the intensity of the hypersensitivity symptom during migraine attacks (“During your worst headaches, how intense is the disturbance due to [the symptom]?”) and the possible answers were “none”, “slight”, “moderate”, “severe”, and “very severe”, to be quoted 0/1/2/3/4 respectively. A score was calculated for each domain by multiplying the frequency and the intensity sub-scores, for a result ranging from 0 to 16. A score of one or more indicated the presence of hypersensitivity in each domain. We defined “Relevant hypersensitivity” as a calculated score of 2 or more, and a “severe hypersensitivity” as a calculated score of 6 or more. These threshold values were defined on the basis of clinical relevance: the clinical equivalent of “Relevant” (score ≥ 2) is when patient starts to mention the existence of a hypersensitivity, whereas the clinical equivalent of “Severe” (score ≥ 6) is when the hypersensitivity starts to hinder the patient.

See Fig. 1 for an overview of the questionnaires that are processed in the present paper.

Fig. 1.

Description of the questionnaires processed in the present paper, with number of items per questionnaire, object (validation/translation) and sources

The MHQ-8 validated in French and a suggested translation in English (not validated) are available at https://neurodol.fr/francois-gabrielli/.

Conduction of the survey

Each patient was offered to participate to the study by the physician in charge during a current follow-up visit for migraine care. The physician reported demographic and morphometric data, as well as information about the history of migraine. Then, in a quiet room, the patient answered consecutively on a paper sheet: the French version of the ASC-12, the French version of the hypersensitivity to light questionnaire by Choi et al. (Photo-8), the hyperacusis questionnaire by Khalfa et al. (Phono-14), the MHQ-8 and French validated versions of the Headache Impact Test-6 (HIT-6) [16, 17], the Migraine Disability Assessment (MIDAS) [17, 18], and the Hospital Anxiety and Depression scale (HADS) [19].

For the test-retest reliability, the same set of hypersensitivity questionnaires, plus the HADS, was delivered to a group of participants of the coordinating centre at Clermont-Ferrand, two weeks later during another visit.

Sample size

The number of subjects to include in the present study was evaluated from the number of items of the longest questionnaire, i.e. the phonophobia questionnaire, containing 14 items. We considered that 15 subjects per item, for a total of 210 subjects, would bring sufficient accuracy in the validation process. We estimated that 10% of questionnaires would not be correctly filled, raising the needed minimal subject number to 230. Considering the number of items in the survey, 230 subjects allow a good validation of the factorial analysis [20], as well as bring the significant level down to 5% and the accuracy up to 95%.

In addition, a minimum of 50 subjects were needed for the test-retest of the survey. We considered a reply rate of 80%, and therefore we planned to recruit a minimum of 60 participants from Clermont-Ferrand University Hospital to fill the survey twice.

Data analysis

All data was entered into REDCap and double checked. Numerical data were expressed as mean ± standard deviation (SD); nominal data were expressed as headcount and percentage.

The validation of the French versions of the ASC-12 [8] and the Photo-8 [12], the validation of the Phono-14 [13], as well as the validation of the MHQ-8, were conducted according to current recommendations [15, 21].

Content validity

Prior to implementation in this study, the acceptability of the entire questionnaire was tested and confirmed on N = 27 patients (23 women, 46 ± 13 years old, HIT-6 at 63 ± 7). All questions were found to be easy to understand and the average time to complete was less than 15 min. No missing data were noted.

Structural validity

The structural validity of the questionnaires was tested by means of an exploratory factorial analysis (EFA) and a confirmatory factor analysis (CFA), which was first applied to each questionnaire to confirm its transcultural validity and then to the 36 items of the survey (12 for allodynia, 8 for photophobia, 14 for phonophobia and 2 for osmophobia). Finally, both an EFA and a CFA were performed on the 8 items of the MHQ-8 to test whether 4 factors could be confirmed. The outputs of the EFA were the loadings, and the outputs of the CFA were the Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI), the Root Mean Squared Error of Approximation (RMSEA) and the Standardised Root Mean Square Residual (SRMR). As the data were considered categorical, we used the diagonally weighted least squares (DWLS) estimator [22].

Internal consistency

Cronbach’s α was calculated to assess internal consistency.

Transcultural validity

Confirmatory factor analysis was also carried out to confirm the transcultural validity of the long questionnaires, namely ASC-12, Photo-8 and Phono-14.

Reliability

Test-retest reliability was analysed by calculating the intraclass correlation coefficient (ICC) and by calculating Cohen’s κ coefficient for nominal outcomes (obtained after transforming the scores into classes).

Criterion validity

Patient reported outcomes, which often focus on subjective perceptions and opinions, almost always lack a gold standard. An exception is when we want to develop a shorter questionnaire for a construct where a long version already exists. In this case, the long version could be considered as the gold standard [23]. Thus, the criterion validity of the MHQ-8 against questionnaires exploring similar domains as gold standard (namely, all except osmophobia) were analysed by calculating Pearson’s correlation coefficient (or Spearman’s in the case of a non-Gaussian distribution).

Construct validity

Construct validity against external outcomes (namely, the other study questionnaires) were also analysed by calculating Pearson’s correlation coefficient (or Spearman’s in the case of a non-Gaussian distribution).

Responsiveness

Spearman Correlation between MHQ-8 scores (except osmophobia) and their respective gold standards were done from the test-retest data, and correlation coefficients were compared using a two-sided Wilcoxon signed rank test.

Software tools

All data were processed with MATLAB (2015. Natick, Massachusetts: The MathWorks Inc.) using the Statistics and Machine Learning Toolbox and Confirmatory Factorial Analysis was done using R^® and Lavaan package.

Missing data

Whenever an item is missing for a participant, the participant was removed from the corresponding questionnaire analysis, but was kept for the other questionnaires. As a consequence, the number of available questionnaires varies within analysis.

Data availability

Raw anonymized data will be available on publication on a public depository held by University Clermont Auvergne: https://entrepot.recherche.data.gouv.fr/dataverse/univ-clermont-auvergne.

Results

Descriptive analyses

Study sample

A total of 329 participants were included into the study. The characteristics of the study sample are presented in Table 1. As expected, the sample was predominantly made of middle-aged women. Overall, 83% of the patients presented at least 4 migraine days per month and 26% had chronic migraine. The vast majority had a severe impact of their migraine (82%), 38% of participants had definite anxiety, 11.5% had definite depression, 81.8% had severe impact of migraine on their daily life according to HIT-6 and 44.7% had severe disability according to MIDAS.

Table 1.

Demographic characteristics and migraine characteristics of the study sample (n  = 329)

Age, years	43.7 ± 13.2 [19.0–80.0]
Female sex	276 (83.9%)
BMI, kg/m²	24.1 ± 5.1 [14.5–46.3]
Physical activity > 2 h / week	117 (35.6%)
Tobacco consumption	47 (14.3%)
Alcohol consumption > 1 glass/week	97 (29.5%)
Family history of migraine	218 (66.3%)
Age of migraine onset	18.7 ± 10.3 [2.0–55.0]
Pain intensity during attacks (out of 10)	5.7 ± 1.1 [2.0–8.0]
Maximal pain intensity during attacks (out of 10)	8.4 ± 1.3 [2.0–10.0]
Presence of aura during attacks	102 (31.0%)
No. of days with migraine per month, in the last 3 months	10.2 ± 7.0 [0.0–30.0]
Chronic migraine	86 (26.1%)
Use of preventive treatment	241 (73.3%)
Type of acute treatment used
NSAID	200 (60.8%)
Triptans	246 (74.8%)
Opioid	64 (19.4%)
Acetaminophen	121 (36.8%)
Combination	61 (18.5%)
Acute treatment overuse	120 (36.5%)
Anxiety score (HADS) (out of 21)a
none (score < 8)	114 (35.4%)
possible (score 8-to-10)	87 (27.0%)
relevant (score > 10)	123 (38.2%)
Depression score (HADS) (out of 21)a
none (score < 8)	232 (72.0%)
possible (score 8-to-10)	54 (16.8%)
relevant (score > 10)	37 (11.5%)
HIT-6 score (out of 78)
Low impact (score < 50)	15 (4.6%)
Moderate impact (score 50-to-55)	14 (4.3%)
Important impact (score 56-to-59)	31 (9.4%)
Severe impact (score ≥ 60)	269 (81.8%)
MIDAS scoreb
Little or No Disability (score < 6)	66 (22.4%)
Mild Disability (score 6-to-10)	25 (8.5%)
Moderate Disability (score 11-to-20)	72 (24.4%)
Severe Disability (score > 20)	132 (44.7%)

Numerical data are expressed as mean ± sd [range]; nominal data are expressed as headcount and percentage. Abbreviations: BMI: body mass index; HADS, hospital anxiety and depression scale; HIT-6, 6-item headache impact test; MIDAS, migraine disability assessment. Notes: A, 322 usable questionnaires; B, 295 usable questionnaires

Responses to the questionnaires

Table 2 presents the distribution of the total scores calculated from the survey questionnaires. The proportion of useable questionnaires was very good for all hypersensitivity questionnaires, but always better for the MHQ-8 items (≥ 99%) than for the longer questionnaires (≥ 95%), confirming the excellent acceptability of the new questionnaire.

Table 2.

Description of the responses to the different questionnaires, expressed as scores

	No. of useable questionnaires (%)	Total score	Any hypersensitivity	Relevant hypersensitivity	Severe hypersensitivity
Cutaneous allodynia
ASC-12 (out of 24)	312 (94.8)	3.8 ± 4.0 [0–18]	159 (51.0)	99 (31.7)	42 (13.5)
MHQ-8-Allo (out of 16)	326 (99.1)	2.6 ± 3.2 [0–16]	213 (65.3)	151 (46.3)	59 (18.1)
Photophobia
Photo-8 (out of 8)	316 (96.0)	5.8 ± 1.8 [0–8]	310 (98.1)	NA	NA
MHQ-8-Photo (out of 16)	327 (99.4)	9.4 ± 4.6 [0–16]	321 (98.2)	305 (93.3)	261 (79.8)
Phonophobia
Phono-14 (out of 42)	323 (98.2)	18.7 ± 10.1 [0–42]	NA	NA	66 (20.4)
MHQ-8-Phono (out of 16)	327 (99.4)	8.7 ± 4.9 [0–16]	321 (98.2)	298 (91.1)	237 (72.5)
Osmophobia
MHQ-8-Osmo (out of 16)	326 (99.1)	5.1 ± 5.1 [0–16]	250 (76.7)	213 (65.3)	132 (40.5)

Numerical data are expressed as mean ± SD [range]; nominal data are expressed as headcount and percentage. Abbreviations: ASC, allodynia symptom checklist; MHQ, migraine hypersensitivity questionnaire

Figure S1 presents the distribution of the scores calculated from the survey, expressed as a rate of response per item. As it is calculated by multiplication, the score cannot take all values between 0 and 16, and values of 3 and 8 are rare. The thresholds (relevant or severe hypersensitivity) have not been established on the basis of the distribution of responses, but solely on the basis of clinical relevance.

Hypersensitivities

Photophobia and phonophobia were the most common hypersensitivities, both present in 98% of patients in our sample, regardless of which questionnaire (long version or 2-item version) was used. Severe photophobia was present in 80% of patients and severe phonophobia in 73%. The 14-item score showed that 20% of participant showed a severe hyperacusis. Osmophobia was present in 77% and severe in 41%. Cutaneous allodynia was present in 51% of patients according to the ASC-12 and in 65% according to the MHQ-8 subscore. It was severe in 13 or 18% of subjects, depending on which questionnaire was used. All results are shown in Table 2.

Structural validity: factor analysis

In order to match the structure of the survey, we chose 4 factors: cutaneous allodynia, photophobia, phonophobia and osmophobia. We kept only the most significant coefficients, with a threshold set to 0.3.

Results of the exploratory factor analysis on the 36 items showed that cutaneous allodynia, phonophobia and osmophobia were successfully identified as factors, but not photophobia: instead, the 4th factor revealed to be a subset of phonophobia score. We replicated the analysis with 5 factors. Those new results showed that most of the items (33 over 36) fitted in 4 factors that corresponded to the 4 scores of hypersensitivities. Consequently, the 5th factor was identified as a subset of phonophobia score: the additional factor was identified with items that scored for the effect of sound environment during migraine attack (Table S1 and S2).

From the CFA on the 36 items, the model with 5 factors was considered acceptable as the CFI was higher than 0.95, the TLI was higher than 0.9 and the RMSEA was less than 0.05 (Table S1). Only the SRMR (0.183) was higher than 0.08.

The factor analysis on the MHQ-8 items showed excellent exploratory results (Table S2.C), and the indicators of the CFA showed good performances with CFI and TLI at 0.999, RMSEA at 0.054 and SRMR at 0.021. Overall, the new scores exhibited excellent structural validity.

Internal consistency

The internal consistency within each questionnaire has been tested by measurement of the Cronbach’s α coefficient, along with the floor and ceiling effects and Cohen’s κ coefficient (Table 3). Of note, the Cronbach’s α for the whole MHQ-8 was 0.846.

Table 3.

Analyses of internal consistency and test-retest reliability for the different questionnaires

	α	Floor (%)	Ceiling (%)	ICC (95%CI) N = 55 patients
Cutaneous allodynia
ASC-12 (out of 24)	0.744	28.0	0.3	0.865 (0.783–0.918)
MHQ (2 items) (out of 16)	0.913	34.3	0.9	0.874 (0.800-0.923)
Photophobia
Hypersensitivity to light (out of 8)	0.708	1.8	18.5	0.866 (0.785–0.917)
MHQ (2 items) (out of 16)	0.880	1.8	19.5	0.826 (0.726–0.892)
Phonophobia
Hyperacusis inventory (out of 42)	0.904	0.9	0.6	0.848 (0.757–0.906)
MHQ (2 items) (out of 16)	0.890	1.8	19.1	0.772 (0.646–0.857)
Osmophobia
MHQ (2 items) (out of 16)	0.948	23.1	9.4	0.898 (0.836–0.938)

The Cronbach’s α coefficient is the principal parameter of internal consistency; values ≥ 0.7 indicate acceptable reliability, ≥ 0.8 indicate good reliability and ≥ 0.9 indicate excellent reliability. Floor and ceiling effects are defined when the rates of responders who endorsed respectively the lowest and the highest value are ≥ 5%. The intraclass correlation coefficient (ICC) is the principal parameter of test-retest reliability; values between 0.75 and 0.90 indicate good reliability and values over 0.90 indicate excellent reliability. 95%CI = 95% confidence interval

Transcultural validity

Confirmatory factor analysis was also carried out to confirm the transcultural validity of the questionnaire. For the ASC-12 [8], the items were successfully classified into the 3 factors (“mechanical static”, “mechanical dynamic”, “thermal”), except for “shaving your face” and “wearing contact lenses” (see Table S3), which remained below the 0.3 salient threshold.

Regarding Phono-14 [13], the items were mainly classified into 3 factors, but with some differences: “earplugs” was not salient enough, “set aside noise” was misclassified in the emotional dimension, “noise tolerance” was also classified in the emotional dimension, “intelligibility in noise” was completely misclassified, “street noise” was also classified in the emotional dimension.

Where applicable, the CFA successfully confirmed the factors in the literature.

Test-retest reliability

Overall, N = 70 patients were asked to complete the questionnaires twice and N = 55 were usable for the test-retest analysis. The ICCs for each questionnaire are shown in Table 3, suggesting a good reliability of the scores. Of note, the ICC for the whole MHQ-8 was 0.880 (95%CI = 0.848–0.906).

Criterion validity for the MHQ-8

How the newly developed MHQ-8 was able to represent the different aspects of hypersensitivity has been tested by analysing the relationship of each dimension of the MHQ-8 (but osmophobia) with its corresponding validated tool. The relationship was expressed as the ρ correlation coefficient and the significance was defined as a difference between ρ and the null value (Figure S2). The Spearman’s correlation was significant for cutaneous allodynia (MHQ-8-Allo vs. ASC-12; ρ = 0.709; p < 0.0001), photophobia (MHQ-8-Photo vs. Photo-8; ρ = 0.652; p < 0.0001), and phonophobia (MHQ-8-Phono vs. Phono-14; ρ = 0.566; p < 0.0001). Of note, values of ρ between 0.5 and 0.7 indicate moderate correlation, and values ≥ 0.7 indicate a high correlation.

After the scores had been categorised into binary classes, the concordance between categories has also been tested on contingency tables with a Fisher’s exact test. The results indicate a strong agreement between the classification methods, this for mild cutaneous allodynia (p < 0.0001), severe cutaneous allodynia (p < 0.0001), photophobia (p = 0.0066), and phonophobia (p < 0.0001).

Construct validity

Construct validity has been tested by analysing the relationship of each score (for the already existing questionnaires) and sub-score (for the MHQ-8) with the four criteria recorded concomitantly, i.e. anxiety, depression, impact of headache, and migraine disability. The relationship was expressed as the ρ correlation coefficient and the statistical significance was defined as a difference between ρ and the null value. The results are given in Table 4; statistical significance was reached for most relationship, although the level of correlation was from poor (ρ < 0.3) to fair (ρ between 0.3 and 0.5). The correlations of the MHQ subscales with the HIT-6 and MIDAS can be considered a measure of convergent validity, and the correlations of the MHQ subscales with the HADS (especially the depression subscale) can be considered a measure of divergent validity.

Table 4.

Construct validity as analysed by the relationship of each score (for the already existing questionnaires) and sub-score (for the MHQ-8) with the four parameters recorded concomitantly, i.e. anxiety, depression, impact of headache, and migraine disability

	Anxiety (HADS)	Depression (HADS)	Impact of headache (HIT-6)	Migraine disability (MIDAS)
Cutaneous allodynia
ASC-12	0.11 (0.051)	0.04 (0.476)	0.24 (< 0.0001)	0.27 (< 0.0001)
MHQ-8-Allo	0.17 (0.002)	0.04 (0.524)	0.21 (0.0001)	0.25 (< 0.0001)
Photophobia
Photo-8	0.21 (0.0002)	0.12 (0.031)	0.14 (0.014)	0.20 (0.0005)
MHQ-8-Photo	0.08 (0.144)	0.00 (0.956)	0.32 (< 0.0001)	0.23 (0.0001)
Phonophobia
Phono-14	0.32 (< 0.0001)	0.34 (< 0.0001)	0.19 (0.0005)	0.29 (< 0.0001)
MHQ-8-Phono	0.12 (0.033)	0.10 (0.065)	0.37 (< 0.0001)	0.26 (< 0.0001)
Osmophobia
MHQ-8-Osmo	0.15 (0.006)	0.04 (0.452)	0.27 (< 0.0001)	0.22 (0.0001)

The relationship is expressed as the Ρ correlation coefficient and the significance is defined as a difference between Ρ and the null value (p value). Abbreviations: HADS, hospital anxiety and depression scale; HIT-6, headache impact Test-6; MIDAS, migraine disability assessment scale

Responsiveness

Spearman correlation coefficients between MHQ-8 scores and corresponding gold standards were [0.715, 0.652, 0.555] and [0.724, 0.664, 0.606] for respectively test and retest values, and no significant difference was found between those data sets (p = 0.25).

Between-domains associations for the MHQ-8

After the scores for each domain of the MHQ-8 had been categorised into binary classes, the associations between each positive case of relevant (score ≥ 2/16) cutaneous allodynia, photophobia, osmophobia, and phonophobia are described in a Venn’s diagram (Fig. 2). It shows that 31.6% of the responders were positive for each domain of hypersensitivity, and that (31.6 + 27.7) = 59.3% of the responders were positive for photophobia, phonophobia and osmophobia.

Fig. 2.

Venn diagram showing the association between the four types of hypersensitivity as defined by the MHQ-8 (relevant [score ≥ 2/16] cutaneous allodynia, photophobia, osmophobia, and phonophobia)

Discussion

This study successfully developed and validated the MHQ-8, an innovative 8-item questionnaire for evaluating four key migraine-associated hypersensitivities: cutaneous allodynia, photophobia, phonophobia, and osmophobia. Each hypersensitivity dimension was identified through two straightforward items (presence and intensity), demonstrating robust psychometric properties, including good internal consistency and test-retest reliability. To achieve this objective, we have also validated the French versions of the Allodynia Symptom Checklist (ASC-12) [8], a photophobia questionnaire (Photo-8) [12] and an assessment tool for hyperacusis in a migraine population (Phono-14) [13]. The introduction of the MHQ-8 represents a notable advance, as it is the first tool to simultaneously assess both the presence and intensity of these hypersensitivities in a concise format. This innovation facilitates the exploration of potential interactions between them, while improving usability. This concise format enhances its usability for both clinical practice and large-scale research.

Hypersensitivity prevalence and intensity

The literature indicates wide variability in the reported prevalence of ictal sensory hypersensitivities associated with migraine, especially for osmophobia (from 35 to 86%) [24, 25] and allodynia (from 15 to 75%) [26, 27]. This inconsistency highlights the need for standardized, reliable assessment tools, as many studies have relied on interviews or non-validated, semi-structured questionnaires [3, 24]. This also suggests significant variability depending on the population studied, with the prevalence of allodynia varying from 15% in a Korean study [26] to 40–63% in the two largest studies conducted in the USA [8, 10] using the same tool. This supports the need to investigate all hypersensitivities using a validated questionnaire across diverse populations to determine their potential for guiding personalized treatments.

Photophobia and phonophobia, although key diagnostic criteria for migraine, are not universally present [5]. Photophobia and phonophobia were reported by 98% of participants in this study, with significant symptoms in over 90%, in agreement with previous studies [5]. Notably, photophobia, phonophobia, and nausea consistently rank among the top three most bothersome symptoms, although their order may vary across regions [5, 28, 29]. Of course, there is a bias because these three symptoms are part of the diagnostic criteria.

The tool we have just validated will make it possible to assess the prevalence of the symptom as well as its intensity in a rapid and standardised way, providing potentially interesting additional information. A link has been shown between allodynia and poorer response to treatment [27, 30]. It is possible that resistance to treatment is more closely related to the intensity of the symptom than simply the frequency with which it occurs. Therefore, we need an appropriate tool to test this hypothesis. Another useful element could be to assess the presence and intensity of the four different hypersensitivities between attacks, which could be envisaged by adapting the wording of the present questionnaire.

Clinical implications for migraine management

The MHQ-8 offers a practical tool for comprehensive assessment of migraine-related hypersensitivities, complementing traditional measures such as attack frequency, intensity, and duration. By also addressing hypersensitivities, migraine treatments can achieve better outcomes and enhance patients’ quality of life.

An interesting finding from our study was the lower-than-expected prevalence of severe hyperacusis (20%), with scores comparable to non-migraine populations [13]. This suggests that existing hyperacusis criteria may not fully capture the phonophobia experienced by migraine patients. Our newly developed questionnaire appears to align more closely with patient-reported complaints. Factor analysis identified a distinct dimension related to environmental noise stress during migraine attacks, warranting further investigation.

For significant osmophobia, the prevalence was 65%, consistent with recent studies reporting rates between 53% and 67% [3, 6, 31]. Despite its clinical relevance, no simple and validated questionnaire currently exists for standardized assessment. Administering four different questionnaires, each of which takes a few minutes to complete, is too time consuming to be done systematically in daily clinical practice and even in most research protocols. The MHQ-8 could fill this gap and facilitate more comprehensive evaluations.

Regarding photophobia, we opted to use the questionnaire developed by Choi and colleagues [12]. However, other validated tools, such as the Photosensitivity Assessment Questionnaire (PAQ) [32] and the Leiden Visual Sensitivity Scale [33], are available. The PAQ includes 16 items—8 focused on photophobia and 8 on photophilia—and has demonstrated associations between photophobia and poor sleep quality [34]. The Leiden Visual Sensitivity Scale, which was published after the initiation of this project, contains 9 well-validated items and shares the same objective as our newly developed 2-item questionnaire. Future studies should compare these questionnaires, to determine whether the shorter 2-item format offers comparable diagnostic utility.

In the present study, allodynia was reported by 51% of participants using the ASC-12, consistent with previous findings [8, 10, 35, 36]. As shown in the initial validation of the ASC-12 [8] and in the German and Turkish validations, two items (“shaving your face” and “wearing contact lenses”) were not discriminant [36, 37]. Our new questionnaire identified symptoms in 65% of participants, with 46% meeting criteria for relevant hypersensitivity. Mechanical cutaneous allodynia has been suggested as a risk factor for migraine chronification [35]. As such, it should be evaluated more systematically for clinical and research purposes. Of note, our new questionnaire targets only the most common form of allodynia (mechanical), while the ASC-12 also includes thermal allodynia.

From a practical point of view, identifying these hypersensitivities is of interest for treatment. Indeed, there is some evidence that it is better to be exposed to different stimuli than to avoid them. Avoidance of triggers such as noise, light or odors is common among migraineurs. However, learning to cope with triggers, with gradual and regular exposure, as a kind of desensitization, has been shown to be more effective than avoiding them, resulting in fewer migraine attacks and less medication [38].

Practical applications of the MHQ-8

The 2-item surveys introduced in this study exhibited strong correlations with corresponding scores from larger scales, highlighting their potential as efficient screening tools. This short format is well-suited for both clinical and research settings, enabling rapid assessment without compromising diagnostic accuracy.

The observed correlations between hypersensitivity scores and HIT-6 were particularly strong for phonophobia (ρ = 0.37) and photophobia (ρ = 0.32), supporting the hypothesis that hypersensitivities contribute to migraine-related disability. These are consistent with previous studies linking photophobia to migraine-induced disability using the PAQ [39].

The MHQ-8 is a fully self-administered questionnaire, making it practical for use in both in-person and remote evaluations, including digital health platforms. A score of 0 or 1 rules out hypersensitivity. If the score is between 2 and 4 (a score of 5 is impossible because of the multiplication used to obtain the score), hypersensitivity is present but generally does not have too great an impact on the patient’s functioning. With a score of 6 or more, the effect can be disabling. The MHQ-8 can be used to monitor treatment response or to stratify patients in therapeutic trials based on their sensory phenotype. In addition, it is important for the patient to recognise the first signs of hypersensitivity in order to limit aggravating factors during the attack. On the other hand, it has been shown that over-protection against a potential trigger can be harmful [38]. In interictal situations, promoting active coping strategies — such as sensory desensitization, gradual exposure, or environmental adaptations — may help limit their long-term impact on quality of life.

Strength and limits

Our psychometric analyses confirmed the reliability and validity of both the newly developed MHQ-8 and the translated questionnaires (ASC-12 and Photo-8). Internal consistency was supported by satisfactory Cronbach’s alpha values, indicating strong structural coherence across all tools. In addition, excellent test-retest reliability, reflected by high correlation coefficients, demonstrated the robustness and stability of these instruments over time. In addition, the correlations between the MHQ-8 and the HADS (especially the depression subscale) tend to be lower than some other hypersensitivity measures (e.g. MHQ-8-Photo vs. Photo-8), which can be interpreted as an indication of good divergent validity for the MHQ-8. Furthermore, the correlations between the MHQ-8 and the headache impact questionnaires (HIT-6 & MIDAS) tend to be better or similar to some other hypersensitivity measures. In other words, the pattern of correlations suggests good and, in some cases, even better construct validity in favour of the MHQ-8. However, the classification into divergent and convergent measures was made post hoc and is therefore exploratory.

Despite these strengths, several limitations should be acknowledged. First, our study population was primarily recruited from tertiary care centres, introducing a potential selection bias. This cohort included over 25% chronic migraine patients and more than 80% reporting severe migraine impact, as measured by the HIT-6 score, along with a high prevalence of anxiety. Such characteristics likely contributed to an overestimation of hypersensitivity severity compared to what might be observed in a broader population of individuals with episodic migraines and lower disability levels. Extending validation efforts to more diverse populations will be essential to enhance the generalizability of these findings. Rasch models were not used in this study because the sample did not include the full spectrum of migraine patients with varying severity levels. As the model requires a representative distribution of the trait being measured, our current population was not sufficiently heterogeneous. However, Rasch analyses are planned in a larger and more diverse cohort to further validate the questionnaire and ensure its applicability across different migraine subgroups. Additionally, the thresholds we established for defining relevant (≥ 2 points) and severe (≥ 6 points) hypersensitivity were determined based on expert clinical input. Although these thresholds offer a practical starting point for research and clinical use, further studies are needed to refine and validate these cut-off values to ensure their accuracy and relevance in different patient populations. Another limitation of this study is that the questionnaire was not co-designed with patients, as recommended by the PROMIS guidelines. Other dimensions such as nausea, vomiting, vestibular symptoms or kinetosis may have emerged as important symptoms that bother patients during migraine attacks and could have been included in the questionnaire. Such an addition could be suggested in future versions. A final limitation of this study is the absence of data on whether patients completed the questionnaire during or outside a migraine attack. This factor could potentially influence responses, despite the wording of each question explicitly referring to “During your most severe headaches,” similar to the approach used in the ASC-12.

Conclusion

In conclusion, this study introduced the MHQ-8, a practical 8-item self-scoring questionnaire for assessing cutaneous allodynia, photophobia, phonophobia, and osmophobia in migraine patients. The MHQ-8 provides a reliable and efficient tool for both clinical practice and research, enabling comprehensive assessments and guiding more personalized treatment strategies to mitigate hypersensitivity intensity and frequency.

Future studies should further explore the interactions between hypersensitivities, their correlation with neurophysiological measures of sensory discomfort and pain thresholds, and refine assessment criteria to optimize patient care and research outcomes.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Substantial contributions to the conception or design of the work: FG, MZ, CM, CD, RD, XM. Substantial contributions to the acquisition, analysis, or interpretation of data for the work: SSA, BP, JD, SDG, GD, ADo, ADu, PG, EGM, CL, JM, MN, CR, EKVO. Drafting the work: FG and XM. Reviewing the work critically for important intellectual content: MZ, CM, SSA, BP, JD, SDG, GD, ADo, ADu, PG, EGM, CL, JM, MN, CR, EKVO, CD, RD. All authors read and approved the final manuscript.

Funding

This study received financial support from Novartis to cover the costs associated with the study (translation, contracting with all centres, printing and mailing of questionnaires, double data entry). Novartis had no role in designing the questionnaire, analysing the data, or writing the article.

Data availability

Raw data will be available on a public depository held by University Clermont Auvergne: https://entrepot.recherche.data.gouv.fr/dataverse/univ-clermont-auvergne.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

F Gabrielli has nothing to disclose. M Zuel has nothing to disclose in relation to this study. C Magaud currently works for PPD, Part of Thermo Fisher Scientific. Sophia Sickout Arondo received travel, accommodation, and meeting expenses from SOS oxygen and Vertex. B Pereira has nothing to disclose. J Dassa has nothing to disclose in relation to this study. S De Gaalon has nothing to disclose in relation to this study. G Demarquay has received financial support from Allergan-Abbvie, Lilly, Lundbeck, Novartis, Pfizer and Teva not related to the submitted work. A Donnet reports fees from Allergan/Abbvie, Lilly, Lundbeck, Novartis, Orion pharma, Perfood, Pfizer, Teva, not related to the submitted work. A Ducros has nothing to disclose in relation to this study. P Giraud has received financial support from Allergan-Abbvie, Lundbeck, Merck-Serono, Pfizer, Roche and Sanofi. E Guégan-Massardier has received financial support from Allergan-Abbvie, Boehringer Ingelheim, Lilly, Lundbeck, Teva, Medtronic, Novartis, Pfizer not related to the submitted work. C Lucas has received financial support from Abbvie, Lilly, Lundbeck, Orion pharma, Pfizer and Teva, not related to the submitted work. J Mawet reports fees from Abbvie, Lilly, Lundbeck, Orion pharma, Pfizer, Teva and non-financial support from Dr Reddy’s, Ipsen pharmaceutical and SOS Oxygène, not related to the submitted work. M Najjar has nothing to disclose in relation to this study. C Roos has nothing to disclose in relation to this study. EK Van Obberghen has nothing to disclose in relation to this study. C Dualé has nothing to disclose. R Dallel has nothing to disclose. X Moisset has received financial support from Allergan-Abbvie, Amgen, Aptis Pharma, Biogen, Lilly, Lundbeck, Teva, Merck-Serono, Novartis, Orion pharma, Pfizer, Roche, and Sanofi-Aventis and non-financial support from Dr Reddy’s, Elivie, SOS Oxygène and Vertex, not related to the submitted work.