Economic burden and risk factors of migraine disease progression in the US: a retrospective analysis of a commercial payer database

Abstract

Aims

To evaluate the prevalence and risk factors of migraine progression and to assess the incremental burden of migraine progression on healthcare systems.

Materials and methods

Adult patients were required to have a migraine diagnosis in IQVIA’s US adjudicated claims database between 1 January 2012 and 30 June 2016, continuous enrollment ≥12 months before and after the index date (i.e. the first observed migraine diagnosis), and ≥1 additional migraine diagnosis claim during the 12-month post-index period. A previously-developed algorithm identified patients with prevention-eligible episodic migraine (EM). All-cause healthcare resource utilization (HCRU) and costs were evaluated at baseline, over the follow-up period and pre/post progression from prevention-eligible EM to chronic migraine. Cox proportional hazards models were used to evaluate risk factors associated with progression.

Results, limitations, and conclusions

Of the 125,436 patients with prevention-eligible EM that were initially identified, 5,790 (4.6%) were further identified as progressed. Patients who progressed had higher healthcare costs and higher medication use at baseline compared to patients that did not progress. Mean (SD) all-cause total costs per patient per month were $1,790 ($3,788), significantly higher in the post-progression period compared to $1,414 ($2,456) in the pre-progression period in patients who progressed (p < .0001). Younger age, female sex, initial diagnosis by a neurologist, chronic pain, and use of triptans and/or non-specific acute medications were all significant progression risk factors. Results are limited by the use of a heterogeneous population (incident, prevalent, treated, and untreated patients), coding biases, and lack of information on non-prescription drug utilization and plan limits. Limitations aside, there are substantial HCRU and cost burden associated with migraine progression. Younger age, female sex, and the use of specific drug classes are likely to increase migraine disease progression risk.

Keywords:

• migraine
• migraine progression
• resource use
• cost
• risk factors

JEL CLASSIFICATION CODE:

• I12
• I19

Introduction

Migraine is one of the most prevalent health disorders and is one of the most debilitating diseases worldwide^1–5. Migraine is associated with personal and societal burdens of pain, disability, and financial cost. Globally, migraine is underdiagnosed and undertreated, with only 40% of people with migraine and tension type headache report being diagnosed by physicians^1,4,5. The global burden of diseases (GBD) report indicates that migraine is one of the five leading causes of years lived with disability (YLD)³. The American Migraine Prevalence and Prevention Study estimates the prevalence of migraine at 11.79% in the US population aged 12 and above⁶. The prevalence of migraine is higher in females, with an estimated 17.27% of females and 5.72% of males affected by the disease⁶. Episodic migraine (EM) is characterized by patients having 14 or fewer headache days per month. The International Classification of Headache Disorders (third revision) diagnostic criteria defines chronic migraine (CM) as having 15 or more headache days per month, of which at least 8 days per month have migraine features for at least 3 months^7,8. Patients with CM represent around 8.8% of all migraine cases, and the percentage increases with age⁶. CM is associated with higher disability leading to loss of household productivity, work productivity, and missed time for family, leisure, and social activities when compared to EM^9,10. CM is also associated with higher healthcare resource use and costs when compared to EM^10,11.

Previous studies based on survey data have shown that patients progress from EM to CM at a rate of 2.5–3% per year^9–12. There is limited understanding of the pathophysiology of progression from EM to CM. However, previous epidemiological studies have identified risk factors for progression to CM. These established risk factors include overuse of medications such as triptans, high headache frequency, allodynia, obesity, caffeine consumption, snoring, depression, neck/head trauma, and other comorbidities^12–20. Socioeconomic factors such as low income and low education were found to be significantly associated with progression of migraine^13,14. Considering the higher burden of CM, identification of risk factors for migraine progression is essential for development of appropriate treatment guidelines and prevention strategies^12–17. Though existing studies have established the higher burden of patients with CM when compared to patients with EM, there is limited real-world data on comparison of healthcare resource use and costs in patients with migraine who progressed to CM when compared to patients who did not. In other words, there is limited information that directly overlaps with the timing of progression given that most studies have focused on the economic burden of EM versus CM, irrespective of the progression. The current study aimed to evaluate the prevalence and assess the incremental burden of migraine progression on healthcare systems using a real-world claims database. Further, the study aimed to complement the existing literature regarding risk factors for migraine progression to help develop better treatment strategies for prevention of migraine progression.

Methods

Study design and data source

This retrospective cohort analysis was conducted using IQVIA’s US adjudicated claims database [PharMetrics Plus (P+)], using data from January 2011 through June 2017. IQVIA’s US adjudicated claims database is one of the largest US health plan claims databases. The aggregated database is comprised of adjudicated claims from more than 150 million unique enrollees across the United States and has diverse representation of geography, employers, payers, providers, and therapy areas. Records in the database are considered representative of the national, commercially insured population in terms of age and gender. Standard fields include inpatient and outpatient diagnoses and procedures, and retail and mail order prescription records and payments. All data are HIPAA compliant to protect patient privacy.

Patient selection

Eligible patients were required to have a migraine diagnosis in the database during the sample selection window (1 January 2012 − 30 June 2016), evidenced by ≥2 non-ancillary outpatient claims > 30 days apart or ≥1 inpatient claim with a migraine diagnosis code (ICD9: 346.xx, ICD10:G43.xxx). The date of the first observed diagnosis was the index date. Continuous health plan enrollment ≥12 months pre- and ≥12 months post-index was required. Patients were required to be aged ≥18 years at index and to have ≥1 additional non-ancillary claim with a migraine diagnosis code within the 12-month post-index period. Patients with missing demographic data (age, gender, and/or geographic region) were excluded.

A previously-developed algorithm was used to identify prevention-eligible migraine patients. The algorithm was developed by using a sample of patients from an electronic medical records database with headache frequency data to generate criteria that could be used to identify patients in claims data that do not have headache frequency data available (positive predictive value: 0.973 with specificity: 0.655; sensitivity: 0.817). The following criteria during a 12-month post-index period (including the index date) was identified²¹: (1) ≥1 CM diagnosis codes (ICD-9/10 codes of 346.7x or G43.7xx); (2) ≥2 claims for ≥2 unique classes of migraine preventive medications (e.g. anticonvulsants, antidepressants, antihypertensives); (3) ≥15 claims of acute migraine medications (i.e. triptans, ergotamine/dihydroergotamine, isometheptene, NSAIDs, barbiturates, neuroleptics/antiemetics and/or opioids); (4) ≥2 claims (< 90 days apart) for the same class of preventive migraine medications²²; (5) ≥24 triptan equivalents (1 triptan equivalent = 2 tablets per day of a triptan, regardless of dose); ≥24 opioid equivalents (1 opioid equivalent = 4 tablets per day regardless of dose); and/or ≥15 butalbital equivalents (1 butalbital equivalent = 4 tablets per day regardless of dose)^21,22.

The purpose of this analysis was to look at patients who progress from episodic to chronic migraine, thus, patients who met criteria 1 and 3 were not included as they were assumed to already have chronic migraine. Patients who met criteria 2, 4, and 5 were included and were required to have continuous health plan enrollment ≥24 months post-index.

Measures

Baseline demographic characteristics and comorbid conditions

Baseline demographics were evaluated including age, gender, health plan type, payer type, and geographic region at index.

Patient comorbid conditions were identified during the 12-month pre-index period using the International Classification of Diseases, Ninth and Tenth Revisions (ICD-9/10) codes. A list of chronic comorbid conditions across several organ systems included alcohol/drug use, asthma, anxiety, cardiac arrhythmia, cardiac valvular disease, cerebrovascular disease, chronic kidney disease, chronic pain/fibromyalgia, congestive heart failure, chronic obstructive pulmonary disease, dementia/Alzheimer’s disease, depression, diabetes, dyslipidemia, epilepsy/seizure disorders, hepatitis, HIV/AIDs, hypertension, diseases of the liver/gallbladder/pancreas, myocardial infarction/coronary artery disease, osteoarthritis, paralysis/hemiplegia/paraplegia, peptic ulcer disease, peripheral vascular disease, renal failure/dialysis, rheumatologic disease, schizophrenia, sinusitis, sleep disorders, thyroid disease, and current or previous smoking. Charlson Comorbidity Index (CCI) scores were calculated using the Dartmouth-Manitoba adaptation^23,24.

Utilization and costs

All-cause healthcare resource utilization (HCRU) was reported at baseline (12-month pre-index period) and over the 24-month post-index period (including the index date). HCRU was presented as the number of different medical services included in the following mutually exclusive categories: outpatient [including physician office visits, and other (i.e. lab, radiology, and surgery)]; emergency department (ED), inpatient services; and pharmacy. Total pharmacy utilization was reported along with acute (migraine specific and non-specific medications were reported separately) and preventive categories (reported separately). Migraine specific acute medications included triptans, ergotamine/dihydroergotamine, and isometheptene. Non-specific medications included NSAIDs, barbiturates, neuroleptics/antiemetics, and opioids. Preventive migraine medications were based on evidence-based treatment guidelines and included antidepressants, antihypertensives, anticonvulsants, and onabotulinumtoxinA²⁵. Health plan-allowed amounts were reported in total, as well as for the following subcategories: outpatient [including physician office visits, and other (i.e. lab, radiology, and surgery)]; emergency department (ED), inpatient services; and pharmacy. Per patient per month (PPPM) costs were determined by dividing each patient’s total costs accrued during the follow-up period by the number of months (1 month = 30 days) that the patient had continuous enrollment and calculating the group average of all the individual patient averages. Because of the skewness of the data, medians (including median PPPM metrics) were used to report costs. All costs were adjusted to 2017 US dollars using the Medical Care Consumer Price Index for All Urban Consumers²⁶.

Migraine progression definition

The progression date was defined as the date of the specific CM diagnosis or the date when acute medication(s) increased in frequency, whichever occurred first. An increase in medication frequency was defined as the first 3-month window at any time post-index in which there were ≥4 prescriptions where the date of the fourth prescription served as the progression date. Baseline demographic and clinical characteristics, and HCRU/cost outcomes were reported for patients with prevention-eligible EM who progressed to CM during the 24 months post-index and for those who did not progress during the 24-month post-index period. Acute and preventive medication use (number of fills/prescriptions), HCRU, and costs were reported per-patient per-month (PPPM) during the 24-month post-index (follow-up) period. Length of follow-up time (months) before and after progression (from prevention-eligible EM to CM) was also reported. PPPM measures (acute and preventive medication use, HCRU and costs) were reported and compared before and after the progression date.

Statistical analyses

Descriptive analyses included reporting the frequency (number of patients [n]) and percentage (%) for each cohort for categorical measures. For continuous variables, both the mean (standard deviation [SD]) and median were reported. Two cohort comparisons were conducted on baseline characteristics, post-index HCRU and costs in patients with and without progression using 2-sample t-tests (means) and Wilcoxon rank-sum tests (medians) for continuous variables and Chi-square/Fisher exact tests for categorical variables. Post-index HCRU and costs before and after progression within the progressed sub-cohort were tested with paired t-tests (mean) and Wilcoxon signed-rank tests (median) for continuous variables and McNemar’s tests for categorical variables. Multivariable Cox proportional hazards models were used to evaluate significant risk factors associated with migraine progression from prevention-eligible EM to CM (time to progression to CM). Risk factors included age group, gender, region, payer type, physician specialty associated with the index migraine diagnosis, total all-cause costs during the 12 months pre-index, baseline acute and preventive medication use, and the number of unique classes of preventive medication used at baseline, along with comorbidities of interest. Comorbidities adjusted for in the models were anxiety, chronic pain/fibromyalgia, sleep disorders, osteoarthritis, thyroid disease, depression, hypertension, dyslipidemia, and epilepsy/seizure disorder. Patients were followed until the progression date or censored at 2 years for patients who were not observed to progress during the 2-year post-index period. All analyses were conducted using SASⁱ version 9.3.

Results

Patient characteristics

After the application of study inclusion and exclusion criteria, 125,436 patients with prevention-eligible EM were initially identified. Of these patients, 5,790 (4.6%) were further identified as progressed. Patient attrition is described in detail in Figure 1.

Figure 1. Study sample attrition.

Demographic and clinical characteristics of the sample are detailed in Table 1. Patients who progressed were slightly younger compared to patients that did not progress [mean (SD) age = 42 (12) vs. 43 (12), p < 0.0001]. Around 40% of patients were in the Southern region, 25% in Midwest and Northeast, and the remaining 10% were located in the West. Most patients (80%) were commercially-insured in a PPO plan. Gender distribution was significantly different between progressed and non-progressed cohorts, with a higher proportion of females in the progressed cohort (87.9% vs 84.3%, p < 0.0001). The most frequently observed comorbid conditions were osteoarthritis, dyslipidemia, and hypertension. Patients who progressed were more likely to have significantly higher baseline rates of depression, chronic pain, sleep disorders, and thyroid disease, and significantly lower rates of hypertension, dyslipidemia, smoking, and diabetes compared with the non-progressed patients. Mean (SD) CCI score was lower in patients who progressed compared to patients that did not progress [0.4 (1.0) vs 0.5 (1.0), p = 0.0038]. A significantly higher proportion of patients who progressed were initially diagnosed by a neurologist, compared to the cohort of patients who did not progress (27.6% vs 20.5%; p < 0.0001).

Table 1. Patient characteristics.

Characteristic	Prevention-eligible EM progressed to CM (n = 5,790)	Non-progressed prevention-eligible EM (n = 119,646)
Age, mean years (SD)	42.3 (11.7)	43.4 (12.5)*
Sex, % female	87.9%	84.3%*
Geographic region, %*
Northeast	28.6%	24.9%
Midwest	23.1%	25.5%
South	37.2%	39.6%
West	11.1%	10.0%
Payer type, %
Commercial	58.4%	57.2%
Self-insured	3.2%	3.6%
Medicare Advantage	0.3%	0.3%
Medicaid	37.4%	38.3%
Unknown	0.8%	0.6%
Charlson Comorbidity Index (CCI), mean (SD)	0.4 (1.0)	0.5 (1.0)*
Baseline comorbid conditions, %
Asthma	10.4%	9.9%
Anxiety	19.1%	19.2%
Cardiac arrhythmia	4.7%	5.7%*
Chronic pain/fibromyalgia	17.5%	13.5*
Depression	24.1%	22.4%*
Diabetes	5.9%	6.8%*
Dyslipidemia  (including hyperlipidemia)	22.6%	26.0%*
Hypertension	19.9%	25.5%*
Osteoarthritis	44.2%	40.3%
Sleep disorders	16.2%	14.5%
Smoking or history of smoking	6.5%	7.5%*
Thyroid disease	13.4%	12.5%*
Baseline migraine medication use, %
Acute medications  (specific & non-specific)**	81.5%	72.4%*
Preventive medications***	66.2%	61.9%*

* Statistically significant (i.e. p < 0.05). Univariate analyses were conducted with 2-sample t-tests to compare mean values. Chi-square tests/Fisher exact tests were used to compare categorical variables.

** Acute migraine medications included triptans; ergotamine/dihydroergotamine; isometheptene mucate; NSAIDs; butalbital combination analgesics; neuroleptics/antiemetics; opioids.

*** Preventive migraine medications included antidepressants (e.g. amitriptyline, SSRIs, other (e.g. mirtazapine, venlafaxine, phenelzine); antihypertensives (e.g. propranolol, lisinopril, candesartan, calcium channel blockers); anticonvulsants (i.e. gabapentin, topiramate, divalproex sodium, sodium valproate); onabotulinumtoxin A.

Resource utilization and costs

Baseline (12-months pre-index) utilization and costs

At baseline, significantly more patients with prevention-eligible EM who progressed to CM utilized acute migraine medications, anticonvulsants, tricyclic antidepressants (TCAs), or other antidepressants compared to patients that did not progress. Within the triptan category, patients who progressed had more baseline triptan prescriptions PPPM compared to patients that did not progress (0.5 vs 0.3, p < 0.0001). Within the beta-blockers category, patients who progressed had fewer baseline prescriptions PPPM than patients that did not progress (0.4 vs 0.5, p = 0.0007). All-cause total costs, outpatient, pharmacy, and inpatient costs were significantly higher at baseline in patients that progressed vs those that did not progress (p < 0.05).

Post-index (24-months post-index) utilization and costs

Patients who progressed had higher HCRU compared to the non-progressed cohort in the 24-month post-index period (all p < 0.0001). Significantly more patients who progressed (p < 0.0001) used acute or preventive medications (except for SSRIs, p = 0.40), compared to the cohort of patients that did not progress in the 24-month post-index period. Patients who progressed had higher mean numbers of prescription fills for triptans, opioids, NSAIDs, and TCAs PPPM, and lower mean numbers of prescription fills for antihypertensives or other antidepressants PPPM compared to the cohort of patients that did not progress in the 24-month post-index period (p < 0.01). Patients who progressed were associated with higher total all-cause costs and costs by type of service (all p < 0.0001 except for inpatient costs), compared to patients who did not progress in the 24-month post-index period. Patients who progressed were associated with significantly higher migraine-specific total costs and costs by type of service (i.e. outpatient, inpatient, ED, and pharmacy, all p < 0.0001), compared to patients who did not progress in the 24-month post-index period.

Pre–post-healthcare resource utilization and costs for progressed cohort

Detailed resource utilization and cost data are shown in Table 2. The mean (SD) post-index time before progression was 16.5 (4.5) months. Mean (SD) post-index time after progression was 7.5 (4.5) months. HCRU was significantly higher in the post-progression period, compared to the pre-progression period (all p < 0.0001). Utilization (number of prescriptions) of acute and preventive medications was higher after disease progression, compared to the pre-progression period (all p < 0.0001). All-cause total costs, outpatient and pharmacy costs significantly increased after disease progression (all p < 0.0001, Figure 2). Migraine-specific total costs, outpatient, pharmacy costs (all p < 0.0001) and inpatient costs (p < 0.05) also increased significantly after disease progression.

Figure 2. All-cause healthcare costs by medical service type before vs after progression. *p < 0.0001.

Table 2. All-cause healthcare resource utilization and costs over the 24-month follow-up period (per patient per month) before vs after progression.

	Prevention-eligible EM before progression (n = 5,790)	Prevention-eligible EM after progression (n = 5,790)
Total number of healthcare visits and prescriptions among all patients
Mean (SD)	6.0 (3.8)	7.2 (5.6)*
Median	5.2	5.8*
Total healthcare costs among all patients
Mean (SD)	$1,414 ($2,456)	$1,790 ($3,788)*
Median	$779	$874*
Outpatient visits
Incidence of ≥1 visit, %	99.9%	99.9%
Number of visits among all patients
Mean (SD)	2.6 (2.3)	3.1 (3.3)*
Median	2.0	2.2*
Number of visits among those with ≥1 visit
Mean (SD)	2.6 (2.3)	3.1 (3.3)*
Median	2.0	2.2*
Outpatient visit costs
Total costs among all patients
Mean (SD)	$753 ($1,428)	$999 ($2,277)
Median	$424	$493
Total outpatient costs among those with ≥1 visit
Mean (SD)	$754 ($1,428)	$1,000 ($2,278)*
Median	$424	$493*
Physician office visits
Incidence of ≥1 visit, %	99.9%	99.1%*
Number of visits among all patients
Mean (SD)	1.6 (1.5)	1.9 (2.4)*
Median	1.1	1.3*
Number of visits among those with ≥1 visit
Mean (SD)	1.6 (1.5)	2.0 (2.4)*
Median	1.1	1.3*
Physician office visit costs
Total costs among all patients
Mean (SD)	$192 ($202)	$258 ($504)*
Median	$135	$153*
Total costs among those with ≥1 visit
Mean (SD)	$192 ($202)	$261 ($505)*
Median	$135	$154*
Emergency department visits
Incidence of ≥1 visit, %	45.0%	23.8%*
Number of visits among all patients
Mean (SD)	0.1 (0.2)	0.1 (0.3)
Median	0.0	0.0*
Number of visits among those with ≥1 visit
Mean (SD)	0.2 (0.3)	0.3 (0.6)*
Median	0.1	0.2*
Emergency department visit costs
Total costs among all patients
Mean (SD)	$92 ($320)	$94 ($466)
Median	0.0	0.0*
Total costs among those with ≥1 visit
Mean (SD)	$204 ($453)	$393 ($892)*
Median	$92	$156*
Inpatient visits
Incidence of ≥1 visit, %	14.3%	6.7%*
Number of visits among all patients
Mean (SD)	0.0 (0.1)	0.0 (0.1)*
Median	0.0	0.0*
Number of visits among those with ≥1 visit
Mean (SD)	0.1 (0.1)	0.3 (0.5)*
Median	0.1	0.1*
Inpatient visit costs
Total costs among all patients
Mean (SD)	$251 ($1,284)	$300 ($2,272)
Median	$0	$0*
Total costs among those with ≥1 visit
Mean (SD)	$1,748 ($2,981)	$4,474 ($7,646)*
Median	$905	$2,009*
Pharmacy utilization (outpatient)	100.0%	97.5%
Incidence of ≥1 prescription, %
Number of prescriptions among all patients	3.4 (2.2)	4.0 (3.2)*
Mean (SD)	2.9	3.3*
Median
Number of prescriptions among those with ≥1 prescription
Mean (SD)	3.4 (2.2)	4.1 (3.2)*
Median	2.9	3.4*
Pharmacy costs (outpatient)
Total costs among all patients
Mean (SD)	$318 ($553)	$398 ($804)
Median	$168	$175
Total costs among those with ≥1 prescription
Mean (SD)	$318 ($553)	$408 ($812)*
Median	$168	$183*

* Statistically significant (i.e. p < 0.05). Paired t-tests were used to compare mean values before and after progression. Wilcoxon signed-rank tests were used to compare median values. McNemar’s tests were used to compare categorical variables. For outcomes reported among users (e.g. patients with ≥1 prescription), independent sample tests were used.

Risk factors for migraine progression

Results from the multivariate Cox proportional hazard model are presented in Figure 3. Younger age groups (i.e. 18–34, 35–44, and 45–54 compared to the ≥65 years reference) were associated with an increased risk of migraine progression. The hazard of progression was significantly higher in females vs males (HR = 1.17; p < 0.0001) and in patients whose index diagnosis was made by a neurologist vs any other type of prescriber type (HR = 1.36; p < 0.0001). Chronic pain and the use of triptans, butalbital, anticonvulsants or other non-specific acute medications of interest (i.e. NSAIDs and neuroleptics/antiemetics) were all significantly associated with an increased risk of progression. Anxiety, hypertension, and dyslipidemia were all significantly associated with a lower risk of progression.

Figure 3. Risk factors associated with migraine progression from prevention-eligible EM to CM. The HRs are adjusted for potential confounders in the model such as demographics, baseline (12-month pre-index) comorbidities, medication use, and costs; also adjusted for region, payer type, and number of unique classes of preventive medication use at baseline (not shown in the figure).

Discussion

The current study provides detailed real-world resource use and cost data in patients with prevention-eligible EM using a large US health insurance database. Such data suggest that patients with prevention-eligible EM who progressed to CM were more likely to have higher baseline healthcare costs, and higher baseline acute or preventive medication use, compared to the cohort of patients that did not progress during the 24-month follow-up period. This analysis indicated that 4.6% of patients with prevention-eligible EM progressed within 24 months of the initial observed migraine diagnosis. Bigal et al.¹⁰ reported an annual rate of progression from EM to CM of 2.5% based on self-reported survey data from a sample of 14,450 patients with EM that participated in the American Migraine Prevalence and Prevention (AMPP) 2005 study. Similarly, Castillo et al.²⁷ reported the prevalence of progression to be 2.4% (95% CI = 1.8–3.3) based on a sample of 1,883 survey respondents in Cantabria, Spain. As the current study utilized a sample of patients with more severe disease (i.e. patients with prevention-eligible migraine) compared to Bigal et al. and Castillo et al., the higher rate of progression observed in the current study seems clinically plausible. In addition, our progression rate was estimated within a time window of 24 months following the initial observed migraine diagnosis (not an annual rate) and was not based on self-reported data. As our study utilized more recent data, it is also possible that chronic migraine has become better diagnosed and coded in the claims data over the years.

All types of HCRU and costs (outpatient, inpatient, pharmacy, etc.) increased significantly after patients with prevention-eligible EM progressed to CM. One possible reason for this finding is that patients who progressed had worsening cases of EM at the point of progression which resulted in higher costs. Though there is a paucity of literature reporting HCRU and cost data before and after progression in the same sample of patients, this finding is consistent with prior studies that have reported higher HCRU and costs in patients with CM compared to patients with EM¹¹.

Multivariate analysis indicated that younger age, female sex, initial diagnosis by a neurologist, comorbid chronic pain, and use of triptans and/or non-specific acute migraine medications (butalbital, NSAIDs, and neuroleptics/antiemetics) were all significant risk factors for disease progression. While younger age has not been reported previously as a risk factor, female sex, and medication use have been reported by others as risk factors for progression to CM^13,16. The association between initial diagnosis by a neurologist and disease progression may be due to progressing EM patients having worsening disease, which may lead to a referral to a neurologist. Additionally, a new CM diagnosis code may be a result of a decision to consider initiation of onabotulinumtoxin A treatment, which is more likely to occur by a neurologist. In this study, comorbidities such as hypertension, dyslipidemia, anxiety, and depression showed either slightly lower risk of progression or no significant association, which appears to be inconsistent with other research that show an increased risk^28–32. However it should be interpreted in the context that the regression model also included migraine medications that could also be used to treat these comorbidities, and these medication variables might modify the effect of the comorbidity on migraine progression. Hypertension was associated with a significantly lower risk of progression which may be an effect of antihypertensive therapies also treating migraine. Dyslipidemia was also associated with a significantly lower risk of progression which, along with the finding that hypertension was associated with a lower risk of progression, may be due to an age effect as both comorbidities increase with age while the risk of progression decreases with age. It is unclear why anxiety was associated with a slightly lower risk of progression and adjusted results showed that depression or antidepressants were not significantly associated with the risk of migraine progression. Similar to the hypertension findings, it may be an effect of antidepressant/anti-anxiety medications also treating migraine.

Limitations

Results from retrospective studies should be interpreted with an understanding of their inherent limitations and in the context of results from other similar studies. The current study was subject to several limitations. The current study sample included a mix of incident and prevalent, treated and untreated patients with migraine, which may have been subjected to confounding by indication/bias. Coding biases (e.g. a payer mandated CM diagnosis for onabotulinumtoxin A claim approval) and errors may have influenced the findings from the current study. Additionally, the generalizability of the study findings may be limited by the fact that patients ≥65 years are under-represented in the database and information from patients that do not participate in commercial plans (e.g. uninsured patients and those covered only by non-commercial Medicare or non-commercial Medicaid) is not included. Generalizability may have been limited further by the lack of information on utilization of non-prescription drugs and by requiring ≥15 claims of acute migraine medications in the operational definition of CM (potentially excluding patients with CM who did not have the requisite number of claims due to plan limits around the number of prescription claims permitted). The finding of higher baseline utilization of acute migraine medications in patients who progressed to CM may have been limited by the fact that acute medication use (i.e. ≥15 claims) was used to define CM. Similarly, the impact of progression on the use of acute medications is difficult to interpret with acute medication use as part of the CM definition. Finally, results pertaining to the effects of preventive medication on migraine progression may be limited as many of the preventive medications are also used to treat conditions other than migraine (e.g. hypertension, anxiety, and depression) which may or may not be used to treat a comorbid condition rather than migraine. In addition, some preventive medications that were included (e.g. calcium channel blockers) are reported to have conflicting or inadequate evidence to support preventive use in AHS/AAN guidelines.

Conclusions

The results of this study suggest there are substantial healthcare resource use and cost burdens associated with migraine progression. Approximately 5% of the study sample of patients with preventive eligible EM progressed to CM. Younger age, female gender, and use of specific drug classes were all associated with increased progression. Additional studies are needed to substantiate these findings.

Transparency

Declaration of funding

This study was sponsored by Eli Lilly and Company. Eli Lilly and Company employees were involved in study design, interpretation of data, writing of the report, and the decision to submit the report for publication.

Declaration of financial/other interests

CCC, CBM, and RLW are employees of IQVIA. YD is a former employee of IQVIA. SAF, OM, PM, and WY are employees and stockholders of Eli Lilly and Company. TRS has received compensation for contracted work on clinical trials, consulting and for his work as a member of Eli Lilly and Company’s speakers’ bureau. SJ has received compensation as a member of Eli Lilly and Company’s speakers’ bureau and advisory board.

A peer reviewer on this manuscript has disclosed that they have received honoraria for participation in clinical trials, contribution to advisory boards or oral presentations from: Alder, Allergan, Amgen, Electrocore, Ipsen, Lilly, Medtronic, Novartis, Pfizer, Teva and Weber & Weber. They have received financial support for research projects from Electrocore. Their headache research is supported by the German Research Council (DFG), the German Ministry of Education and Research (BMBF), and the European Union. A peer reviewer on this manuscript has disclosed that they serve on the Speakers’ Bureau for Lilly for Emgality and Reyvow, which are both migraine specific medications. They have also served on Advisory Boards for Lilly Pharmaceutical. The peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Acknowledgements

Sasikiran Nunna (IQVIA) provided assistance drafting the manuscript. Yi-Chien Lee (IQVIA) provided analytic support.

Notes

i SAS version 9.3 is a registered trademark of SAS Institute Inc., Cary, NC, USA.