Migraine is common in patients with sarcoidosis

Abstract

Objective

To describe the frequency of migraine and predictors of having migraine in sarcoidosis patients.

Methods

The ID migraine questionnaire was administered to a well-phenotyped observational cohort of sarcoidosis patients (most of whom were seeking specialty care) and healthy controls. Predictors of migraine status were examined using univariate and multivariable logistic regression.

Results

Migraine was seen in 29% of 96 patients with sarcoidosis and 13% of 39 healthy controls, (p = 0.049). Among those with sarcoidosis, in univariate regression analysis only female sex was predictive of having migraine, and in a multivariable regression female sex remained significant (OR 4.6, 95% CI 1.2–18.2). There was no association between migraine and age, depression, dyspnea, immunosuppression use, or ESR.

Conclusions

Migraine is a common comorbidity in sarcoidosis patients. As such, better recognition and targeted treatment of migraine has the potential to improve quality of life as part of a comprehensive care plan for sarcoidosis patients.

Introduction

Sarcoidosis is a multisystem inflammatory disorder characterized by granulomatous inflammation (1). Depending on which organs are involved, symptoms can be diverse and broad-ranging, often leading to diagnostic uncertainty as to whether a complaint is secondary to sarcoidosis or due to another etiology. Headache can be a particularly troubling symptom to parse, with clinicians often wondering whether the patient has neurosarcoidosis as the cause of headaches. Migraine is common in the population, with a one-year prevalence of 18% in women and 6% in men in the USA (2). We aim to describe, in a well-phenotyped cohort of individuals with sarcoidosis, the frequency with which migraine is seen as a comorbidity.

Methods

This cohort study of individuals with sarcoidosis has been previously described (3). In brief, these adult subjects with sarcoidosis and healthy controls were recruited into the University of California, San Francisco Sarcoidosis Research Program cohort between January 2010 and May 2015. This led to a total cohort size of 126 subjects with sarcoidosis and 64 healthy controls. At study enrollment, subjects completed a general neurological symptom screening questionnaire. Given a high affirmative response rate to “headache”, to probe this question further we modified the study protocol to incorporate the ID Migraine questionnaire going forward. At the time of this analysis, 96 sarcoidosis subjects and 39 healthy controls had completed ID Migraine screening, and here we report that analysis. All subjects gave written informed consent to participate, and the study was approved by the UCSF Institutional Review Board (IRB number 10-02323).

Sarcoidosis diagnosis was confirmed by study clinicians based on clinical history, exclusion of alternate causes, and biopsy showing non-necrotizing granulomas according to accepted criteria (3,4). Those with comorbid cancer, chronic infection, other autoimmune disease, and other pulmonary diseases were excluded.

ID migraine, a validated three-question instrument with high positive predictive value for a migraine diagnosis (5), was utilized to screen for comorbid migraine. Baseline questionnaires were analyzed.

Participants were also screened for depression using the nine-item Patient Health Questionnaire (PHQ-9) (6,7). Patient-reported degree of dyspnea was assessed using the baseline dyspnea index (BDI2) (8).

Statistical analysis was conducted using Stata v.13.0 (College Station, Texas). Summary statistics on demographic variables were performed. Student’s t-tests or Chi-Squared/Fisher’s Exact were used as appropriate. Univariate and multivariable logistic regression models were examined. As depression and migraine are associated in the general population, depression symptoms were controlled for in the logistic regression model using PHQ-9 with a cut-point of ≥ 10, indicating moderate or greater depression, as this cut-point has been previously used in migraine epidemiology research (9). To control further for somatic symptom sensitivity, patient-reported degree of dyspnea was also controlled for using the baseline dyspnea index (BDI2). As a measure of systemic inflammation, ESR values from collected blood samples were analyzed. Analyses were stratified based on presence or absence of known neurosarcoidosis, with neurosarcoidosis based on clear evidence of granulomatous infiltration in the nervous system (10).

Results

There were 96 individuals (55% female) with sarcoidosis and 39 healthy controls (56% female). Of those with sarcoidosis, 18 (19%) had neurosarcoidosis. Table 1 shows the demographic and clinical variables of the study sample.

Table 1.

Clinical and demographic characteristics of the 96 individuals with sarcoidosis and 39 healthy controls.

	All sarcoidosis	Sarcoidosis without neurosarcoidosis	Sarcoidosis with neurosarcoidosis	Unaffected controls
N	96	78	18	39
Age in years; mean (SD); range	51.1 (10.7); 30–76	51.2 (11.0); 30–76	50.4 (9.0); 38–68	46.1 (15.3); 22–73
Female n (%)	53 (55%)	42 (54%)	11 (61%)	22 (56%)
Currently on immunosuppression for treatment of sarcoidosis	76 (79%)	59 (76%)	17 (94%)	– (0%)
BDI2 mean (SD)*	8.7 (2.3)	8.4 (2.8)	9.8 (1.6)	10.7 (1.1)
PHQ9 ≥ 10, n (%)	21 (22%)	19 (24%)	2 (11%)	1 (3%)
ESR (mm/hr)**	14.5 (15.5)	13.3 (15.0)	19.2 (16.9)	9.5 (7.5)
ID migraine screen positive	28 (29%)	22 (28%)	6 (33%)	5 (13%)

SD: standard deviation; PHQ9: Patient Health Questionnaire 9, higher scores on this instrument indicate greater depression symptomatology; BDI2: Baseline Dyspnea Index 2, lower scores on this instrument indicate greater dyspnea symptomatology; ESR: Erythrocyte sedimentation rate, mm/hr.

*

BDI2 scores were available on 38 controls, 65 of those without neurosarcoidosis, 15 of those with neurosarcoidosis, and 80 of the total sarcoidosis group.

**

ESR values were available on 32 controls, 69 of those without neurosarcoidosis, 18 of those with neurosarcoidosis, and 87 of the total sarcoidosis group.

Thirteen percent (n = 5) of healthy controls screened positive on ID migraine.

Of those with sarcoidosis, 29% (n = 28) screened positive on ID migraine (p = 0.049, Fisher’s Exact), which included 28% (n = 22) of patients without neurosarcoidosis and 33% (n = 6) with neurosarcoidosis (p = 0.67 for the difference).

In univariate analyses of subjects with sarcoidosis, female sex was the only significant predictor of having migraine (see Table 2). In a multivariable logistic regression model of the same group, again, only female sex was associated significantly with migraine (OR 4.6, 95% CI 1.2–18.2, p = 0.03). Being on immunosuppression, ESR, depression (PHQ-9 ≥ 10), and dyspnea (BDI2) were not associated with having migraine in the multivariable model. Having neurosarcoidosis also did not predict having migraine (OR 0.7, 95% CI 0.14–3.5).

Table 2.

Univariate regression model results for predictors of migraine among the 96 subjects with sarcoidosis.

	Univariate regression model (OR, 95% CI)
Female sex	2.7 (1.02–6.84)
Age	1.0 (0.96–1.05)
Neurosarcoidosis	1.3 (0.4–3.8)
ESR	1.0 (0.98–1.03)
Immunosuppression	2.8 (0.8–10.4)
PHQ-9 ≥ 10	2.2 (0.8–6.1)
BDI2	0.9 (0.7–1.05)

OR: odds ratio; CI: confidence interval; ESR: erythrocyte sedimentation rate, mm/hr; PHQ9: Patient Health Questionnaire 9; BDI2: Baseline Dyspnea Index 2.

In a multivariable logistic regression model including all subjects (cases and healthy controls), female sex was again the only predictor associated with migraine (OR 5.4, 95% CI 1.6–18.7, p < 0.01).

Discussion

Migraine is common in the general population; our study demonstrates that migraine is also common among those with sarcoidosis. Migraine should be considered in the differential diagnosis of recurrent headaches in patients with sarcoidosis, and better recognition and treatment of migraine has the potential to improve quality of life in this population.

Female sex, a predictor of migraine in the general population (2), was also predictive of migraine in people with sarcoidosis. Of note, those with neurosarcoidosis were not more likely to have migraine than those without neurosarcoidosis, although the cohorts were small. The two conditions can certainly coexist.

Strengths of this study include the use of a validated migraine instrument, ID migraine, in a well-phenotyped, biopsy-confirmed sarcoidosis research cohort. ID migraine has a 93% positive predictive value for a migraine diagnosis (5). The frequency of migraine observed in healthy controls in our study was 13%, which is in line with the North American population-based prevalence rates (2), albeit ID migraine has three-month recall and population-based estimates are one-year period prevalences. Furthermore, the proportion of migraine in women in our dataset was approximately threefold higher than men, again in line with migraine population-based data (2).

Limitations of this study include not being able to confirm migraine status by direct interview, and the relatively small sample size. This cohort was also not designed to be population-based and many subjects were recruited based on interest, seeking specialty care for sarcoidosis. While we are unable to generalize our results to individuals with sarcoidosis who were either asymptomatic, not seeking specialty care, or undiagnosed, we believe that the study population is representative of sarcoidosis patients seeking specialty care at a tertiary center. Furthermore, not all subjects had neuroimaging and/or cerebrospinal fluid examinations, tests that can be helpful for diagnosing CNS neurosarcoidosis. However, clinical presentation, care, and follow-up make the likelihood of undiagnosed CNS sarcoidosis in the non-neurosarcoidosis group relatively low.

While migraine was more frequent in those with sarcoidosis than in controls in our dataset (29% vs. 13%, p = 0.049), given our relatively small sample size and attendant risk for Type 2 error we do not think this clearly demonstrates that migraine is more common in those with sarcoidosis. Rather, we consider the main finding of our study to be that migraine is common in those with sarcoidosis and as common in sarcoidosis as in healthy controls, and thus migraine should be considered in the differential when patients with sarcoidosis present with recurrent headaches and diagnosed and treated accordingly. Larger studies would be needed to determine whether migraine is actually more prevalent in sarcoidosis than in the general population. Clinicians caring for people with sarcoidosis should screen patients for migraine, either through direct interview or with a well-validated screening tool like ID migraine. Neurological consultation may be helpful in some cases in evaluating for neurosarcoidosis and other causes of headache. Under-diagnosis of migraine in the general population is common, with only 48% of those with migraine ever having received a physician diagnosis of migraine (11). Better recognition and targeted treatment of migraine has the potential to improve quality of life in people with sarcoidosis.

Key findings

Migraine is common in patients with sarcoidosis.

Female sex is a risk factor for migraine in those with sarcoidosis, as it is in the general population.

Age, immunosuppression, inflammatory markers, or presence of neurosarcoidosis do not predict migraine status in those with sarcoidosis.