Incidence of Status Migrainosus in Olmsted County, Minnesota, United States

Abstract

Background and Objectives

SM is recognized as a complication of migraine in which pain and/or associated symptoms are unremitting and debilitating for more than 72 hours. The epidemiology of SM in the general population is not known. The aim of this study is to determine the incidence, recurrence rate, and clinical associations of status migrainosus (SM) in care-seeking residents of Olmsted County, Minnesota.

Methods

The Rochester Epidemiology Project was used to identify the incident cases of SM according to the International Classification of Headache Disorders, Third Edition criteria and based on the first physician-encountered case in the record. The clinical characteristics of the incident cases were abstracted from the medical record. One-year recurrence-free survival was evaluated and compared between clinically relevant groups, including baseline demographics, migraine characteristics, and treatment exposures.

Results

Between January 1, 2012, and December 31, 2017, 237 incident cases of SM were identified. The median age was 35 (IQR 26–47) years, and 210 (88.6%) were female. A history of chronic migraine was recorded in 82/226 (36.3%) and a history of aura in 76/213 (35.7%). At the time of the incident case, medication reconciliation included a triptan or ergotamine in 127/233 (53.6%) and/or an opioid-containing analgesic in 43/233 (18.5%). The overall age- and sex-adjusted incidence rate was 26.60 per 100,000 [95% CI, 23.21–29.97], with a peak incidence between ages 40 and 49 years. The median (95% CI) attack duration was 5 (4.48–5.42) days. The most frequent triggers were stress (40/237, 16.9%) and too much or too little sleep (27/237, 11.4%). Recurrence occurred in 35/237 (14.8%) at a median of 58 (IQR 23–130) days following the initial attack. In our age- and sex-adjusted multivariable model, too much or too little sleep as a trigger was associated with 12-month risk of recurrence (adjusted OR 3.59 [95% CI 1.58–8.14], p = 0.0022).

Discussion

Our study provides a population-based estimate of SM incidence. We identified aberrant sleep patterns as a potentially modifiable risk factor for 1-year SM recurrence.

Migraine is common with a 1-year prevalence of 12% in the United States, more often seen in females than males.¹ The prevalence has been observed to increase from age 12 to 40 years and then steadily decline.

Status migrainosus (SM) is recognized by the International Classification of Headache Disorders, Third Edition (ICHD-3) as a complication of migraine in which pain and/or associated symptoms are unremitting and debilitating for more than 72 hours.²

The epidemiology of SM has not been well characterized. The condition is reported to have been seen in practice by 97% of surveyed headache clinicians, implying that it is not rare.³ In a retrospective study interrogating the 2013 Nationwide Readmissions Database, SM accounted for 14.4% of 12,448 hospital admissions for migraine.⁴ In children, a history of SM was recorded in 10.5% in a retrospective cohort of 5,316 persons with migraine.⁵ However, this is in contrast with retrospective data from a French tertiary center, where SM was only observed in 3% (24/8,821) of individuals over an 11-year observation period.⁶ It might be possible that SM is overrepresented among certain subsets of individuals with migraine, such as those with menstrually related migraine or migraine with aura.^5,7-10

To characterize the epidemiology of SM, we performed a population-based incidence study using the Rochester Epidemiology Project (REP). The goals of the study included determining the incidence of SM in care-seeking residents of Olmsted County, Minnesota (MN), and describing clinical characteristics and 1-year recurrence for SM following the incident event.

Methods

REP Database

This study used the population-based resource of the REP. Olmsted County is isolated from other urban/suburban centers with only a few medical providers. These include Mayo Clinic and Olmsted Medical Center and their affiliated clinics. Because of the unique isolated nature of the region and few providers, billing data on all medical services are collated through the REP.^11,12 This enables identification of the incident diagnosis of medical conditions, such as SM, and permits review of treatments, evaluations, autopsy reports, and death certificates.

Data Collection and Case Ascertainment

Using the medical indexing of the REP, the International Classification of Diseases, Ninth Edition (ICD-9) (before October 1, 2015)/the International Classification of Diseases, Tenth Edition [ICD-10] (on or after October 1, 2015) codes for intractable migraine or status migrainosus were searched, and all cases were manually reviewed. Between January 1, 2012, and December 31, 2017, male and female residents of Olmsted County of any age who met the criteria for SM according to the ICHD-3 were included.² Given that assignment of an ICHD-3 diagnosis of SM requires establishing a prior diagnosis of migraine without aura and/or migraine with aura, we followed the convention of prior REP investigators to not require specific documentation of 5 prior attacks if the diagnosis was otherwise typical.^13,14

As a study standard, incident cases were the first physician-encountered case in the record. Individuals were excluded if (1) reported SM events were less than 72 hours, (2) diagnosis of SM was based on the physician documentation of retrospective self-reported history alone, (3) red flags were present in the clinical history without sufficient diagnostic evaluation to safely conclude that a secondary etiology was not present, or (4) chart review revealed a subsequent discovery of a secondary headache disorder as cause for refractory headache rather than the index event of SM. Red flags were identified according to convention (e.g., SNOOP4).¹⁵ A recurrent case of SM required the same criteria as the incident attack, except occurring within 1 year from the date of the index encounter. All cases were independently reviewed by 2 study investigators, with arbitration by a third investigator used in instances of disagreement.

A final diagnosis was recorded at last follow-up as determined by the clinician in the medical record. In cases in which new daily persistent headache (NDPH) and episodic or chronic migraine were each recorded as diagnostic possibilities, the ICHD-3 guidance on making NDPH the default diagnosis was used.²

Data Management and Statistical Plan

Study data were collected and managed using Research Data Capture. Age- and sex-adjusted incidence rates were calculated based on the population data from REP.

Demographics and clinical presentation characteristics, encounter setting, medication list and potential triggers at incident event, and SM treatments were presented. Triggers were recorded if they were attributed by the patient as contributing to their incident attack of SM. Because of the retrospective study design, it is likely that not all data points of interest (e.g., triggers) were systematically assessed. Both baseline prescriptions (prior to treatment of the incident attack of SM), in addition to medications used for treatment of the attack, were recorded. Medications used for treatment of the attack could have either been present at baseline or prescribed at the time of the encounter. Attack duration was defined as the duration in days between the onset and termination of the attack, according to that reported in the medical record.

The association between recurrence and change in migraine subtype between episodic and chronic migraine from initial encounter to final diagnosis was assessed using the χ² test. Exploratory univariate and age- and sex-adjusted multivariable Cox proportional hazard models of 12-month recurrence were conducted. The median attack duration and 95% CI were estimated by bias-corrected and accelerated bootstrap with 2000 samples.¹⁶ All statistical tests were 2 sided, with p < 0.05 considered statistically significant. Analyses were performed in SAS v9.4 (SAS Institute; Cary, NC).

Standard Protocol Approvals, Registrations, and Patient Consents

This study was approved by the institutional review boards of Mayo Clinic and Olmsted Medical Center. We reviewed the records of Olmsted County residents seen between January 1, 2012, and December 31, 2017, who participated in the REP and granted permission for review of their medical records.

Data Availability

The data cannot be made public because all residents were in a specific Minnesota County. Any geographic subdivision smaller than a state cannot be deidentified according to the Health Insurance Portability and Accountability Act of 1996 definition of protected health information.¹⁷

Results

Using the medical indexing of the REP, we identified 1,442 possible cases of SM between January 1, 2012, and December 31, 2017, based on ICD-9 (before October 1, 2015)/ICD-10 (on or after October 1, 2015) codes for intractable migraine or status migrainosus (Figure 1). After elimination of duplicates and review for inclusion and exclusion criteria, 237 cases of incident SM were identified. The majority of excluded cases did not qualify due to not meeting the threshold for attack duration of more than 72 hours. Nine cases were additionally excluded whose migraine attack fell outside of the dates of the prespecified study window.

Figure 1. Identification and Classification of Individuals With Status Migrainosus.

Demographic and Clinical Characteristics

Demographic and clinical characteristics of individuals with SM are shown in Table 1. The 237 individuals had a median age of 35 (IQR 26–47) years and were predominantly White (200/237, 84.4%) and female (210/237, 88.6%). At the time of the incident SM episode, the majority had episodic migraine (144/226, 63.7%) and migraine without aura (137/213, 64.3%). The majority had a family history of migraine (126/176, 71.6%).

Table 1.

Demographic and Clinical Characteristics of Individuals With Status Migrainosus in Olmsted County, 2012–2017

Table 2 outlines the encounter setting, medication list at presentation, and potential triggers for the incident SM event. Presentation of SM to the emergency department/urgent care (137/237, 57.8%) was slightly more common than in the outpatient setting (100/237, 42.2%). Most individuals had an nonsteroidal anti-inflammatory drug (158/233, 67.8%) and/or triptan (126/233, 54.1%) listed among their medications at the time of presentation. About 1 in 5 individuals had an opioid-containing analgesic, and about 1 in 10 individuals had a butalbital-containing analgesic listed at presentation. Most individuals did not have a reported trigger for their episode of SM (159/237, 67.1%). Stress (40/237, 16.9%) and too much/too little sleep (27/237, 11.4%) were the most commonly named triggers. The median (95% CI) attack duration was 5 (4.48–5.42) days. The mean (95% CI) attack duration for the full sample was 8.6 (7.29–9.92) days.

Table 2.

Encounter Setting, Medication List, Potential Triggers, and Treatment at Incident Status Migrainosus Event

As outlined in Table 3, the most common treatments for SM in the outpatient setting were NSAIDs (54/87, 62.1%), triptans (31/87, 35.6%), and/or antiemetics (29/87, 33.3%). In the emergency department/inpatient setting, the most common treatments for SM were antiemetics (118/146, 80.8%), NSAIDs (116/146, 79.5%), and/or opioids (28/146, 19.2%).

Table 3.

Treatment of Incident Status Migrainosus Event

Incidence of SM

The overall age- and sex-adjusted incidence rate among care-seeking Olmsted County residents was 26.60 per 100,000 (95% CI), 23.21–29.99), with a peak incidence occurring between ages 40 and 49 years (Figure 2). Age-adjusted incidence rates per 100,000 were higher in females (46.97 [95% CI, 40.61–53.32]) compared with males (6.23 [95% CI, 3.88–8.58]).

Figure 2. Age- and Sex-Adjusted Incidence of Status Migrainosus in Olmsted County, Minnesota.

SM Outcomes

Recurrence of SM occurred in 35/237 (14.8%) at a median of 58 (IQR 23–130) days following the initial attack. Exploratory univariate analysis identified too much or too little sleep as a trigger as a potential predictor of 12-month risk of recurrence (unadjusted OR 2.94 [95% CI 1.33–6.47], p = 0.0075). In an age- and sex-adjusted multivariable model, too much or too little sleep as a trigger was associated with over 3 times increased risk of 12-month recurrence (adjusted OR 3.59 [95% CI 1.58–8.14], p = 0.0022) (Table 4).

Table 4.

Univariate and Age- and Sex-Adjusted Multivariable Analysis for 12-Month Recurrence Risk

At last follow-up within 1 year, 43.9% (104/237) had a diagnosis of episodic migraine, 47.7% had a diagnosis of chronic migraine (113/237), and 3.8% (9/237) had a diagnosis of NDPH. Eleven patients had no clear follow-up diagnosis. A statistically significantly higher proportion of those that changed from episodic migraine diagnosis at incident SM to chronic migraine diagnosis at last follow-up experienced SM recurrence within 1 year than those that remained episodic (23.1% vs 7.4%, p = 0.011) (Table 5).

Table 5.

Recurrence Rate Among Individuals With Episodic and Chronic Migraine at Incident Event Compared With Follow-up

Discussion

Using the REP, the overall age- and sex-adjusted incidence rate of SM was 26.60 per 100,000 (95% CI, 23.21–29.97) among the care-seeking residents of Olmsted County, MN. There was a peak incidence occurring between ages 40 and 49 years. Age-adjusted incidence rates per 100,000 were higher in females (46.97 [95% CI, 40.61–53.32]) compared with males (6.23 [95% CI, 3.88–8.58]).

The median age of incident SM cases was 35 (IQR 26–47) years, which is in keeping with the age for peak prevalence of migraine.¹⁸ Approximately 10% of cases of incident SM (25/237; 22 female and 3 male) were observed in individuals aged 17 years and younger compared with approximately 4% of cases among individuals aged 60 years and older (10/237; 8 female and 2 male). At all age ranges, there was a female predominance. The majority of individuals were White (84.4%), which mirrors the demographics of Olmsted County, MN.¹⁹ Approximately 35% (76/237) of individuals had migraine with aura. This is comparable to previously reported prevalence rates of aura in individuals with migraine in the United States.²⁰ The presence of aura on univariate analysis did not show an association with 12-month recurrence risk of SM.

Although most incident SM cases were in individuals with episodic migraine, chronic migraine (36.3%, 82/237) was enriched in this population compared with what would be excepted from previously reported chronic migraine rates in the general population (approximately 2%).²¹ At latest follow-up within 1 year from the incident SM event, there was an increase in the number of individuals with diagnosis of chronic migraine (47.7%, 113/237). Individuals who transitioned from episodic to chronic migraine had a higher rate of SM recurrence than those who remained in the episodic category (23.1% vs 7.4%, p = 0.011). There were an insufficient number of individuals converting from chronic to episodic migraine to permit statistical analysis. SM may be considered a candidate risk factor for chronic migraine progression, although this study was not designed to address that hypothesis. A trajectory of increasing headache frequency among individuals with a history of SM has been suggested based on prior work.^5,8

Over half of those with SM (57.8%, 137/237) presented to the emergency department/urgent care. This is not surprising given that SM is defined by its unremitting and debilitating nature and therefore may require an escalation in therapy.² At presentation with SM, most individuals had an NSAID (67.8%, 158/233) and/or triptan (54.1%, 126/233) on their medication list; however, the available data do not allow for assessment of whether they had taken any of their listed medications at the time of presentation for the incident SM event. In addition, although approximately 1 in 5 individuals had a listed opioid-containing analgesic and 1 in 10 had a listed butalbital-containing analgesic at the time of presentation, it is not known whether the opioid- or butalbital-containing analgesics were previously prescribed to the individual for indication of migraine or another comorbid pain condition. Previous studies have reported that 12% of Olmsted County residents receive opioid prescriptions.²² The population of interest in our study, people with migraine, has a slightly higher rate of opioid prescriptions. This increased incidence is concerning as it goes against best practices, with current migraine guidelines recommending against the use of opioids for migraine treatment.²³ A possible explanation for the observed rates of opioid- and butalbital-containing analgesics may be that the identified individuals reflect a more refractory or severely affected cohort that perhaps had not found 1st-line treatments to be helpful or had comorbidities that contraindicated use. The data available do not allow for exploration of this. Of note, listed medication at presentation (triptan vs NSAID vs opioid vs butalbital) did not show an association with 12-month recurrence risk of SM.

Little is known about the duration of SM episodes beyond its defining feature of lasting a minimum of 72 hours, with no guidance on an upper limit of attack duration.² Our data show, with 95% confidence, that the mean attack duration in this general population sample has an upper limit of 9.92 days. The clinical relevance of an upper limit attack duration would be to alert the clinician toward earlier recognition of a more persistent phenotype, such as chronic migraine or NDPH. In such cases, an attack duration surpassing a defined upper limit would facilitate an earlier shift in management to include other modalities beyond acute interventions alone. Of note, NDPH appears to be overrepresented in this cohort enriched for SM compared with the general population.²⁴ Clinicians should be aware that NDPH represents an uncommon diagnostic trajectory among individuals presenting with SM.²⁵

Recurrence of SM occurred in 35 individuals (14.8%) at a median of 58 (IQR 23–130) days following the initial attack. The window of recurrence (approximately 2 months) is similar to the minimum amount of time that patients are counseled it may take to assess benefit from an oral preventive treatment. Available data do not allow examination of whether any of the individuals were started on preventive treatments after their incident SM episode. The initiation of a preventive treatment would be an interesting factor to examine in future studies as to its effect on SM recurrence risk.

Most individuals did not have an identified trigger for their incident SM episode. Stress (16.9%, 40/237) and too much or too little sleep (11.4%, 27/237) were the most identified triggers. Menstrually related migraine has been previously identified as a contributor to SM.²⁶ However, hormonal changes were identified as a trigger in only 4.6% of individuals (11/237) despite a female predominance. The current study examined the role of sex and hormonal changes in SM, future research should examine the role of gender, as its influence likely extends beyond that of sex hormones and impact on SM needs further clarification.^27,28

Among the more common listed possible triggers, too much or too little sleep, in the age- and sex-adjusted multivariable model, was the only trigger potentially associated with 12-month risk of recurrence (adjusted OR 3.59 [95% CI 1.58–8.14], p = 0.0022). Confirmatory investigation of triggers associated with SM recurrence is warranted, but sleep has been previously identified as a modifiable risk factor for migraine chronification and now has been shown as a potentially modifiable risk factor for SM recurrence risk.²⁹

Just under 2/3 of individuals were treated in the emergency department/inpatient setting for their SM episode. These individuals received an antiemetic (80.8%, 118/146) and/or NSAID (79.5%, 116/146) most commonly; both treatments have high evidence and are guideline recommended.^{23,26,27,30,31} Triptans (5.5%, 8/146) and dihydroergotamine (6.2%, 9/146) were minimally used in the emergency department/inpatient setting, despite having high evidence.^30,31 The low rate of triptan and dihydroergotamine use may relate to individuals having already tried triptans or dihydroergotamine prior to presentation because 55.4% of the overall SM population had a triptan/dihydroergotamine listed among their medications. Alternatively, clinicians might be uneasy with using vasoactive medications in patients with whom they do not have an established care relationship and are concerned about precipitating vascular complications. Opioids were administered to about 1 in 5 individuals, despite minimal evidence and guidelines advising against their use.^23,30,31 The rate of opioid prescribing was slightly lower than previously reported rates based on national database prescribing patterns.³² In the outpatient setting, an astounding 1 in 10 individuals did not receive a documented treatment for their SM episode. It is possible that these patients were sent to the emergency department/inpatient setting for treatment. Nevertheless, this is a particularly concerning observation given that SM is defined by debilitating symptoms and individuals with SM compared with those without have been identified to be at higher risk for suicide.³³ Unfortunately, inadequate care for migraine is a common trend with previous reports documenting only 26.3% of individuals with episodic migraine and 4.5% of individuals with chronic migraine having seen a clinician for migraine, receiving an appropriate diagnosis, and obtaining guideline-recommended treatments.^34,35 From an outpatient treatment standpoint, 62.1% (54/87) received an NSAID, 35.6% (35/87) a triptan, and 33.3% (29/87) an antiemetic. Nerve blocks were minimally used in both the outpatient and emergency department/inpatient setting (3.4% and 0%, respectively), which may speak to limited availability of trained health care practitioners for this procedure. Steroids were also minimally used in both the outpatient and emergency department/inpatient setting (13.8% and 15.8%, respectively), despite literature supporting their effect on reducing headache recurrence.³⁰ The observed choices for SM treatment suggest areas for improvement. Education of clinicians regarding the best evidence acute treatments for migraine is important, recognizing that data specific to SM treatment are limited.³⁶ The study design did not allow for assessment of the efficacy of the administered treatments and did not explore how the timing of treatment in an episode of SM could affect the outcome. Of note, type of SM treatment did not show an association with 12-month recurrence risk of SM.

The major strengths of this study were the identification of a population-based cohort across a contemporary time frame (2012–2017) to characterize a poorly understood migraine complication. Clinical record review offers advantages over population surveys, which are more costly and can be subject to selection and participation bias. This strength, however, may also have been a limitation given that SM case ascertainment required contact with a health care practitioner, and it has been previously reported that many individuals with migraine may not seek care for migraine and clinicians may not recognize migraine.^34,35 However, given the severity of SM episodes, it would be assumed that this condition would prompt seeking medical care and appropriate recognition by health care practitioners. To mitigate potential risk of recall bias of SM attack characteristics by retrospective self-report alone, the incident cases were based on the first physician-encountered case. This contrasts with a patient who reports in the record having had SM a decade prior to the office visit, for example. In addition, the ability to review medical records and use of physician-identified cases ensured that SM cases were collected rather than severe or refractory headaches due to a secondary headache disorder. Furthermore, review of medical records and physician-based diagnosis compared with self-report also should have more accurately allowed for determining initial and follow-up headache diagnosis (such as NDPH, episodic migraine, and chronic migraine). The authors used a previously described approach to not require 5 previous episodes of migraine to establish diagnosis, as has been previously shown to be both valid and necessary to mitigate the impracticality of this criterion in a chart review–based protocol.^13,14 However, this could contribute to lack of diagnostic accuracy. Cases of SM were ascertained from a population of patients who had received either an ICD-9 or ICD-10 diagnostic code for intractable migraine or status migrainosus so it is also possible that some cases of SM were missed if they had not been coded as such. Given this, the reported SM incidence can be viewed as a minimum estimation. Among individuals diagnosed with NDPH, some had more than one follow-up diagnosis (e.g., NDPH and chronic migraine). This may speak to the clinical practice of diagnosing patients with NDPH and providing a diagnosis for the phenotype of the headache (i.e., migrainous) as well. In such cases, the precedent outlined by the ICHD-3, to use NDPH as the primary diagnosis, was followed.² Chart review is also limited to what data are available in the chart. For example, trigger data, such as too much or too little sleep, were based on what was documented in the chart rather than more objective measures such as questionnaires or actigraphy. Furthermore, in exploratory analysis of predictors of recurrence, several clinically relevant variables were tested for association without adjustment for multiple testing, which may have led to identification of a false-positive result.

Our study provides a population-based perspective on SM incidence among care-seeking residents of Olmsted County, MN. SM may be inherently enriched among persons with chronic migraine. We identified aberrant sleep patterns as a potentially modifiable risk factor for 1-year SM recurrence. Of interest, our data suggest that recurrence of SM is associated with conversion from episodic to chronic migraine, highlighting SM as a candidate risk factor for chronic migraine progression. The care of individuals with SM represents an important medical need among persons with migraine, with possible prognostic significance.

Glossary

ICD-9

International Classification of Diseases, Ninth Revision

ICD-10

International Statistical Classification of Diseases, Tenth Revision

ICHD-3

International Classification of Headache Disorders, Third Edition

NFPH

new daily persistent headache

REP

Rochester Epidemiology Project

SM

status migrainosus

Appendix. Authors

Study Funding

This study used the resources of the Rochester Epidemiology Project (REP) medical records-linkage system, which is supported by the National Institute on Aging (NIA; AG 058738), by the Mayo Clinic Research Committee, and by fees paid annually by REP users. The content of this article is solely the responsibility of the authors and does not represent the official views of the National Institutes of Health (NIH) or the Mayo Clinic.

Disclosure

J.H. VanderPluym reports research grant from Amgen. K. Mangipudi, A.A. Mbonde, and D. Gritsch report no disclosures. E. Caronna has received honoraria from Novartis and Chiesi. R.B. Halker Singh reports research grant from Amgen. R.J. Butterfield reports no disclosures. J.H. Smith is currently a full-time employee of AbbVie and holds stock in AbbVie. The content does not reflect the views of AbbVie. Go to Neurology.org/N for full disclosures.