Determinants of the severity of comorbid migraine in multiple sclerosis
- First Online:
- Cite this article as:
- Villani, V., De Giglio, L., Sette, G. et al. Neurol Sci (2012) 33: 1345. doi:10.1007/s10072-012-1119-5
- 216 Views
A high co-morbidity between multiple sclerosis (MS) and migraine has been reported, especially in young female patients affected by a relapsing-remitting (RR) course of MS. In this study, we aimed to elucidate the determinants of the severity of comorbid migraine in MS. Demographic, clinical and psychometric variables were collected from a cohort of 205 RR-MS patients regularly attending to an Italian outpatient MS Centre. Of them, 102 (49.8 %) were diagnosed as affected by comorbid migraine. About one-third of MS patients with comorbid migraine have asked the attending neurologist a specific anti-migraine treatment. Despite this, only few MS patients (10.8 %) reported a prior use of prophylactic drugs, and even fewer (2.9 %) took triptans as pain killers; these proportions were significantly lower when compared with those of a control group of 63 migraineurs subjects without MS (p < 0.0001 for both comparison). Factors associated with a moderate or severe disability (MIDAS grades III or IV) due to comorbid migraine in MS patients were the depressive state (OR = 4.294; p = 0.001), the anxiety trait (OR = 5.786; p = 0.004) and an ongoing IFNB treatment (OR = 2.337; p = 0.028). Likewise, depression (OR = 3.453; p = 0.048) and anxiety (OR = 4.582; p = 0.014) were both independent predictors for having a MIDAS grades of III or IV also in migraineurs subjects without MS. Investigating the determinants of migraine severity may allow a better management of MS patients with comorbid migraine. In these patients, a tailored therapeutic approach is warranted to improve their quality of life and reduce the burden of these two chronic and disabling conditions.
Several studies suggest a high comorbidity between primary headache and multiple sclerosis (MS). The prevalence of headaches in MS people has been reported from 35.5 to 61.8 % [1–7]. A recent systematic review reported the prevalence of primary headaches in MS of about 51 % . Studies providing separate data for migraine and tension-type headache (TTH) showed that migraine represents the most common type of primary headache, especially in young females patients affected by a relapsing-remitting (RR) course of MS [2–4, 6, 7]. Moreover, whereas the frequency of migraine was three times higher (34 vs. 10 %) in MS than in general population , that of TTH is similar (21 vs. 20–34 %) .
Despite its prevalence and impact, migraine is often misdiagnosed in patients with comorbid diseases ; as a consequence, it is under-diagnosed and under-treated also in MS people . Comorbidities are also associated with more rapid progression of disability  and decrements in health-related quality of life of MS patients [13–15]. Moreover, patients suffering from MS-associated pain conditions tended to have more than one concurrent painful condition, especially osteoarthritis and migraine [7, 16]. Nevertheless, no studies have still investigated which factors are associated with the severity of comorbid migraine in MS population, even though some studies focused on the impact of MS-specific treatment, especially interferon beta (IFNB), in triggering “de novo” primary headaches or worsening pre-existing migraine [3, 17–20].
In the present study, we aimed to identify demographic, clinical and psychometric determinants related with the severity of comorbid migraine in a large cohort of patients affected by MS and in a control group of migraineurs subjects without MS.
We consecutively recruited patients affected by RR-MS according to the McDonald criteria , regularly attending the MS Centre of S. Andrea Hospital in Rome. As exclusion criteria, we considered an age <18 years, the presence of cognitive impairment or severe psychiatric disorders, and the current use of prophylactic treatment for migraine.
As control group, migraineurs subjects without MS regularly attending the Neurological Headache Centre of S. Andrea Hospital were consecutively enrolled. We excluded subjects having probable migraine, structural lesions demonstrable by brain imaging, or currently using prophylactic treatment for migraine.
All participants underwent a semi-structured interview according to the ICHD-II criteria  to make a specific diagnosis of headache. Patients with MS diagnosed as affected by migraine, as well as control subjects, were asked to collect a diary for a 3-month period to measure the mean days with migraine per month. In MS patients, the differential diagnosis of primary or secondary headache (i.e. attributable to a MS-specific treatments) was also made according to the definition from the ICHD-II criteria: ‘‘headache as an acute adverse event attributed to medication used for other indications” (8.1.10) . We diagnosed a secondary headache considering (1) the temporal relationship between the headache onset and MS-specific treatment assumption, and (2) the association with one or more symptoms of flu-like reaction, such as chills, fever, fatigue, tiredness and myalgia.
Controls and MS patients diagnosed as affected by migraine also underwent the Migraine Disability Assessment Scale (MIDAS)  for the assessment of migraine disability. Finally, we asked MS patients with comorbid migraine to recall whether they had explicitly requested to the attending neurologist a treatment for their migraine.
We collected demographic and clinical variables for each subject: gender, educational level, marital status, number of children. The Beck Depression Inventory (BDI) , State and Trait Anxiety Inventory (STAI-1 and -2) , and the Toronto Alexithymia Scale (TAS-20)  were also administered for evaluating depression, anxiety and alexithymia, respectively.
In the MS group only, we also collected data on age at MS onset, previous and ongoing MS-specific therapies, and level of disability assessed by the Expanded Disability Status Scale (EDSS)  and the Fatigue Severity Scale (FSS) .
Local Ethics Committee Board provided an exemption of approval for post-marketing studies. Informed consent was obtained by each participant before any study-related procedure.
The BDI is a 21-question multiple-choice self-report inventory, one of the most widely used instruments for measuring the severity of depression. Each question has a set of at least four possible answer choices, ranging in intensity (from 0 to 3); total scores may range from 0 to 63, with higher scores indicating more depressive states: scores from 0 to 9 correspond to no or minimal depression (grade I); scores from 10 to 18 correspond to mild to moderate depression (grade II); scores from 19 to 29 corresponds to moderate to severe depression (grade III); scores of 30 or more correspond to severe depression (grade IV) .
The State-Trait Anxiety Inventory (STAI-1 and 2) is a psychological inventory based on a 4-point scale (from 1 = completely disagree to 4 = fully agree); the STAI measures the anxiety about an event (form 1), and trait anxiety (form 2); scores may range from 20 to 80, with higher scores correlating with greater anxiety .
The TAS-20 is a 20-item scale consists of three subscales (difficulty of identifying feelings, difficulty of describing feelings, externally oriented thinking) with 5 levels of agreement (from 1 = strongly disagree to 5 = strongly agree); the total score may range from 20 to 100 points, with higher scores indicating greater levels of alexithymia . Only patients with a score of 61 or more were considered as having the alexithymic trait .
The MIDAS is a 7-item assessment tool which measures the influence of migraine on three domains of activity over the preceding 3 months: paid and school work, household work, leisure activities with family or in social situations. The last two items investigate the total number of days with migraine attacks and the mean pain intensity. MIDAS score is calculated on the basis of answers to the first five questions and indicates the number of days in which migraine interfered with these activities: minimal disability (grade I) when total score is from 0 to 5, mild disability (grade II) when total score is from 6 to 10; moderate disability (grade III) when total score is from 11 to 20; severe disability (grade IV) when total score of 21 or more .
The EDSS score provides a non-linear measurement of the disability status of MS people; scores from 1.0 to 4.5 indicate a high degree of ambulatory ability, the subsequent levels (from 5.0 to 9.5) refer to the loss of ambulatory ability, and 10 death due to MS .
The FSS is a self-report 9-item scale with seven levels of agreement (from 1 = strongly disagree to 7 = strongly agree); the final score is derived from the mean of nine items and may range from 1 to 7, with higher scores indicating more severe fatigue . Only patients with a score of 4 or more were considered as affected by MS-related fatigue .
All values are reported as proportion or mean (±standard deviation-SD) or median (range), as appropriate.
Differences between groups were tested by the χ2 test (for dichotomous variables), the linear-by-linear test for trend (for ordinal variables), and non-parametric Mann–Whitney U test (for continuous variables). Multigroup comparisons were assessed using an ANOVA analysis.
Demographic, clinical and psychometric variables considered in the statistical analysis
Type (values, if applicable)
MS patients with migraine
Migraine patients without MS
Dichotomous (female or male)
MS duration (years)
Migraine duration (years)
Dichotomous (≤13 years or >13 years)a
Categorical (single, married or divorced)
No. of children
Ordinal (0, 1 or ≥2)
Familial history of primary headache
Dichotomous (yes or not)
Ongoing MS-specific therapy
Categorical (none, interferon beta, glatiramer acetate, natalizumab, immunosuppressants)
Fatigue: FSS score
Dichotomous (<4 or ≥4)
Depressive state: BDI grade
Ordinal (I, II, III or IV)
Anxiety state: STAI-1 score
Dichotomous (≤40 or >40)a
Anxiety trait: STAI-2 score
Dichotomous (≤40 or >40)a
Alexithymic trait: TAS-20 score
Dichotomous (<60, ≥61)
All p < 0.05 in either directions were considered as significant. Analyses were carried out using a PC version of Statistical Package for Social Sciences 16.0 (SPSS, Chicago, IL, USA).
Patients affected by MS
A total of 205 consecutive RR-MS patients (145 females, 60 males) with a mean age of 36.4 ± 9.4 (range 18–65) years, mean MS duration of 8.6 ± 6.7 (range 1–33) years, and median EDSS score of 2.5 (0–6.5) were included in the present study from January 2010 to November 2010. According to ICHD-II criteria, 102 (49.8 %) patients suffered from migraine (all without aura), 22 (10.7 %) from tension-type headache (TTH), while 81 (39.5 %) were free of primary headache. In patients treated with MS-specific treatments (n = 175), such as IFNB (n = 66), glatiramer acetate (n = 6), natalizumab (n = 89) or immunosuppressants (n = 14; 10 mitoxantrone and 4 cyclophosphamide), we performed an accurate classification aimed to differentiate headaches secondary to the therapy from primary headaches, according to ICHD-II criteria  and suggestions coming from a previous study .
Characteristics of the whole MS sample (n = 205) according to the ICHD-II diagnosis
Migraine (n = 102)
Tension-type headache (n = 22)
No primary headache (n = 81)
Mean (SD) age, years
Education level (years)
No. of children
Mean (SD) MS duration, years
Median (range) EDSS score
Ongoing MS-specific therapy
On the other hand, patients with MS and comorbid TTH were older (p < 0.0001), had a longer MS duration (p < 0.0001), and were more likely to be married (p = 0.016) and to have a greater number of children (p = 0.002) than those without headache or migraineurs.
There were no other differences between the groups regarding the demographic and clinical characteristics, and psychometric variables, such as fatigue, depression, and alexithymia.
Comparisons between MS patients with comorbid migraine and migraineurs subjects without MS
Demographic and clinical characteristics of MS patients with comorbid migraine and migraineurs subjects without MS
Migraineurs subjects without MS (n = 63)
MS patients with comorbid migraine (n = 102)
Mean (SD) age, years
No. of children
Mean (SD) migraine duration, years
Familial history of headache, n (%)
No. of days with migraine per month, median (range)
MIDAS score, median (range)
MIDAS grade, n (%) I
Triptan-naïve, n (%)
Prior assumption of prophylactic drugs
Despite the two groups were quite similar in most clinical characteristics, we found significant differences regarding the therapeutic approach to migraine (see also Table 3). All the 63 migraineurs subjects without MS have used NSAIDs at least once, and 22 took triptans as pain killers; moreover, 30 (47.6 %) of them had a prior exposure to one or more anti-migraine prophylactic drugs.
Among the 102 MS patients affected by comorbid migraine, we found that the majority (n = 98, 96.1 %) used analgesics as pain killers, especially non-steroidal anti-inflammatory drugs (NSAIDs) (n = 95), but only 3 patients took triptans (p < 0.0001 vs. migraineurs patients without MS). No prior use of prophylactic drugs was recorded in the majority of MS patients (n = 91, 89.2 %) (p < 0.0001 vs. migraineurs patients without MS), while 11 of them reported a previous assumption of antiepileptic agents (n = 7), or amytriptiline (n = 4); no differences were found between MS patients who had been treated with prophylactic drugs and those who did not (data not shown).
Determinants of migraine severity
Stepwise logistic regression model showing the variables associated with the risk of having a MIDAS grade of III or more; the patients who reached the predefined outcome were 32 (31.4) and 27 (42.8) in the MS and control groups, respectively
95% Confidence intervals
MS group (patients with MS and comorbid migraine)
BDI (each grade)
STAI-2 score ≥ 40
Ongoing IFNB therapy
Control group (migraineurs subjects without MS)
BDI (each grade)
STAI-2 score ≥ 40
This logistic regression model explains a considerable amount of the variance (Nagelkerke pseudo R2 = 0.48) in the risk of having a MIDAS grade of III or IV. The other demographic and clinical characteristics considered in the present study did not contribute to better fitting the model.
Finally, we observed similar findings also found in the control group regarding the variables associated with the risk of having a moderate or severe MIDAS grade (III or IV) (see also Table 4), but this model explains only a modest amount of the variance (Nagelkerke pseudo R2 = 0.25) in the risk of reaching the aforementioned outcome.
The main findings of the present study are that the severity of comorbid migraine in a cohort of 102 RR-MS patients was associated with concomitant anxiety, depression and IFNB assumption.
The risk of having a MIDAS grade of III or IV was 4.3-fold increased for each BDI grade and 5.8-fold increased for a STAI-2 score ≥ 40. These results carried out from the multivariate logistic regression are not only similar to those found in the control group (i.e. migraineurs subjects without MS), but also consistent with the previous literature data. In fact, depression and stressful life events were reported as independent risk factors for migraine clinical progression and associated with a more frequent and disabling migraine in general population . Moreover, anxiety and depression may play a role in the processing and perception of pain, and depressed subjects are more vulnerable to painful physical symptoms [33, 34]. At this regard, a recent study reported that migraine in MS patients was significantly associated with cranial and facial neuralgia, Lhermitte’s sign, temporo-mandibular joint pain and a past history of depression .
In our study, IFNB-treated patients had a 2.3-fold increased risk of having a MIDAS grade of III or more. Other therapies different from IFNB, including natalizumab, did not seem trigger a migraine exacerbation in our sample, according to the results from the stepwise logistic regression.
The role of IFNB in exacerbating pre-existing migraine or triggering “de novo” headache has been extensively recognized [3, 17–20]. Headache which started after IFNB initiation is reported to be only about 15 % [3, 18], suggesting that the majority of patients had an increased frequency of a pre-existing headache [17–19], as also observed in the present study. A dose-dependent effect has been also suggested, with patients receiving high-frequency IFNB treatment reporting a greater number of migraine attacks than those receiving once weekly IFNB administration . Nevertheless, headache or exacerbation of pre-existing migraine was investigated as a side effect only in few trials on IFNB .
In our sample, we recorded a diffuse utilization of NSAIDs as pain killers; despite this, it has to be considered that (1) more than one half of patients (about 58 %) had at least a mild level of disability (as assessed by the MIDAS); (2) less than one-third have asked neurologist a specific treatment for their migraine; (3) only three patients have used triptans as pain killers, (4) just over 10 % of MS patients with comorbid migraine have taken prophylactic drugs in the past.
By contrast, among the group of migraineurs subjects without MS, we observed a lower proportion of triptan-naïve individuals (34.9 %), and a more diffuse use prophylactic drugs in the past (47.6 %). These findings seem to be consistent with a recent review suggesting that migraine is under-treated in MS patients, despite their high comorbidity . A possible explanation may be that both patients and neurologists mainly focused their attention only on MS, to the detriment of migraine. Another hypothesis encompasses the possibility that some patients may consider the migraine as a MS symptom.
Possible implications for the management of comorbid migraine in MS patients
The first clinical implication is that the possibility of comorbid migraine should be considered in all patients with MS and those presenting with frequent episodic and disabling migraine should be also screened for anxiety and depression .
Secondly, the use of triptans in MS setting should be encouraged, even in patients who have previously taken NSAIDs as pain killers. This suggestion is supported by studies on general population demonstrating that (1) triptans are at least as effective as NSAIDs or other analgesic , (2) triptans improve effectiveness and tolerability ratings when compared with the previous acute therapies ; a combination of a triptan and other analgesics may offer improved clinical benefits over monotherapy in some selected cases .
Thirdly, migraine prophylaxis agents should be more extensively used, and symptomatic therapy tailored according to the individual characteristics of MS patients, especially in those starting IFNB therapy to improve their adherence and compliance .
As in general population, drugs that may worse depression have to be avoided (flunarizine, beta-blockers), while tricyclic antidepressant (amitriptyline) and some AEDs—such as topiramate, divalproax sodium, pregabalin, gabapentin, and lamotrigine—may be useful for both mood disorders and migraine. However, it has also to consider that beta-blockers and flunarizine may help if anxiety is present . In some cases, the use of GABAergic molecules (pregabalin, gabapentin) may allow a better control not only of migraine, but also of multiple MS-related symptoms (neuropathic pain, spasticity, etc.) .
The side effects of some drugs should be carefully recognized, especially for therapies affecting the cognitive performance (e.g. amitriptyline, topiramate, zonisamide) or the balance skill (e.g. antidepressant, benzodiazepines), given the high prevalence of cognitive disturbances and accidental falls even in the early stage of MS [41, 42].
Selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake inhibitors (SNRIs) represent first-line therapy for treating depression and anxiety disorders [43, 44], but they have still poor evidence of efficacy in preventive treatment of migraine, and in most cases a second drug must be added . However, the use of both SSRIs and SNRIs in MS is recommended not only for treating mood disorders, but even for enhancing adherence to MS-specific therapy . Interestingly, fluoxetine has also been reported as able to reduce the development of brain inflammation, as seen on magnetic resonance imaging of RR-MS patients .
Finally, non-pharmacological techniques—including relaxation and biofeedback trainings, behavioural approaches, acupuncture, chiropractic spinal manipulative therapy—should be also taken into consideration [48–50].
The limitations of the present study mainly concern the cross-sectional design and the reliance on only one MS centre; thus it is important to replicate results in other populations.
Moreover, the small sample size of some subgroups of patients, especially those treated with GA or immunosuppressants, cannot allow us to better elucidate whether other MS-specific therapies may have the same activity as IFNB in triggering migraine. However, a previous study reported that GA did not affect the severity of comorbid migraine , and immunosuppressants are usually administered for short periods of time, owing to their potential serious adverse events [51, 52].
Despite these limitations, we are confident that this study might highlight some important aspects of comorbid migraine in MS people. At this regards, we think that statistical comparisons between MS patients with comorbid migraine and migraineurs subjects without MS allowed us to provide more consistent results.
Both migraine and MS are disabling and costly diseases, with a considerable burden on patients, their families and society. Therefore, the possibility of comorbid migraine should be considered in all patients with MS. Investigating the determinants of migraine frequency and severity may also allow a better management of these patients.
This research was carried out using information collected during normal patient care, and extra time spent in data analysis and interpretation was part of educational programmes within the University (V.V., L.P., L.D.G.); no external source of funding was required.
Conflict of interest
Prof. Pozzilli has received honoraria for consultancy or speaking from Sanofi-Aventis, Biogen Idec, Bayer Schering, Merck Serono, and Novartis and has received research grants from Sanofi-Aventis, Merck Serono, and Bayer Schering. Prof. Salvetti has received research grants from Sanofi-Aventis, Merck Serono, and Bayer Schering, and lecture fees from Biogen Idec. Dr. Prosperini has received honoraria for consultancy from Merck Serono and lecture fees from Biogen Idec. Drs. Villani, De Giglio, Sette do not have any disclosures to report.