Sage Journals: Discover world-class research

Abstract

Objectives

To describe long-term treatment patterns in migraine patients initiating prophylactic therapy and to evaluate acute medication use and adverse events associated with opioids.

Methods

This study used the 2005–2014 IBM MarketScan® databases to evaluate migraine patients initiating prophylactic medication. Outcome measures included persistence with prophylactic migraine medications over 2–5 years. Acute medication use and gastrointestinal-related adverse events and opioid abuse following opioid use were evaluated. Cox proportional hazards models were used to evaluate predictors of non-persistence and predictors of gastrointestinal-related AEs and opioid abuse associated with long-term opioid use.

Results

In total, 147,832 patients were analyzed. Non-persistence was observed in 90% of patients; 39% switched, 30% restarted, and 31% discontinued treatment. Over the follow-up, 59.9% of patients received triptans, 66.6% non-steroidal anti-inflammatory drugs, 77.4% opioids, and 2.6% ergotamines. Among opioid users, 16.6% experienced nausea/vomiting, 12.2% had constipation, and 10.4% had diarrhea. Opioid abuse was reported in <1% of opioid users. Gastrointestinal-related adverse events increased with increasing number of days’ supply of opioids.

Conclusions

Non-persistence to prophylactic treatment was frequent among migraine patients. Opioid use was common in migraine patients and the risk of gastrointestinal-related adverse events and opioid abuse increased with long-term use of opioids. These results suggest a need for more effective prophylactic migraine treatments.

Keywords

Migraine prophylactic and acute medications treatment patterns opioid-related adverse events

Introduction

Migraine is a chronic neurological disorder that affects 38 million people in the US (1). Evidence suggests migraine substantially reduces productivity and impairs functionality, which in turn translates into poor health-related quality of life (2). More than half of all individuals with migraine are never diagnosed, and of those who are diagnosed, the majority do not seek medical treatment (1).

A number of acute and daily prophylactic therapies are available to treat or prevent migraine, and treatment approach depends on the severity and frequency of headaches (3 –5). Acute medications, like over the counter and prescription analgesics, triptans, and ergotamines, are used on a short-term basis (2–3 days per week), with or without background prophylactic migraine medication, to reverse or abort an ongoing attack and stop progression to severe pain (4). Opioid analgesics are widely prescribed as acute medication for treating migraine (6). Statistics show that as many as 65% of individuals with migraine receive short-term opioid therapy (<3 weeks) and approximately 3% to 4% of the US adult population receive long-term opioid therapy for chronic pain (7,8). Although acute therapies provide necessary pain relief for migraineurs, there is mounting evidence that their overuse can lead to worse and more frequent headaches, decreased responsiveness to other migraine medications, and poorer long-term outcomes (9,10). Furthermore, both short- and long-term opioid use lead to multiple adverse events (AEs) (11). As many as 80% of patients were found to experience at least one opioid-related AE, with gastrointestinal AEs being the most common (12 –14).

Prophylactic migraine therapies are usually indicated in patients with recurrent and frequent migraine attacks (≥2 disabling attacks/month, or ≥2 attacks/week, or increasing attacks over time), in patients who ineffectively respond to or experience AEs from acute treatments, or in patients at risk of acute medication overuse (4,15). Treatment guidelines recommend use of prophylactic migraine medications like topiramate, other anticonvulsants, tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors (SNRIs), calcium channel blockers, anti-hypertensive and other migraine medications for a continued period of at least 2–3 months; however, evidence recommending long-term use is lacking (16). Additionally, guidelines published by the American Academy of Neurology and American Headache Society cite a lack of clinical evidence regarding long-term safety, efficacy, and compliance for current prophylactic migraine medications (17).

Despite availability of multiple prophylactic treatment options, only 12–20% of patients with migraine currently use prophylaxis (18 –21), whereas it is estimated that approximately 38% of migraineurs may meet clinical criteria for receiving prophylactic therapy (20). Of those who receive prophylactic therapy, data from several real-world studies report poor adherence and persistence to these medications (19,22 –27), and AEs were the most common reason for discontinuation (23). While previous studies have examined treatment patterns of prophylactic therapy among migraine patients with <12 months’ follow-up period (19,23,24,26 –28), there is a paucity of data on the long-term treatment patterns and persistence among migraine patients initiating prophylaxis in the US. Therefore, the objective of this study was to describe long-term treatment patterns in migraine patients initiating prophylactic therapy and to evaluate the use of acute migraine medication and the prevalence of opioid-related AEs in real-world clinical practice.

Methods

Study design and data source

This retrospective cohort analysis used administrative claims data between 2005–2014 from the IBM MarketScan® Commercial Claims and Encounters (Commercial) and the Medicare Supplemental and Coordination of Benefits (Medicare) databases. These databases contain inpatient and outpatient medical, and outpatient prescription drug experience of individuals with employer-sponsored primary or those with Medicare supplemental health insurance. The study databases satisfy the conditions set forth in Sections 164.514 (a)-(b)1ii of the Health Insurance Portability and Accountability Act of 1996 privacy rule regarding the determination and documentation of statistically de-identified data. As this study used statistically de-identified patient data, Institutional Review Board review was not required in accordance with the policy of our institution.

Patient selection

A previously published administrative-claims based algorithm published by Woolley et al. (29) was used to identify patients with migraine. Selected patients had evidence of migraine based on a) ≥1 inpatient claim, ≥1 emergency room claim or ≥2 outpatient claims (7–180 days apart) with a diagnosis of migraine (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnosis code 346.xx), b) ≥1 outpatient claim with a diagnosis of migraine and ≥1 outpatient pharmacy claim for a prophylactic migraine medication (either a triptan or ergotamine) (7–180 days apart), or c) ≥2 outpatient pharmacy claims for a triptan or ergotamine and initiation of treatment with a prophylactic migraine medication (other anticonvulsants, topiramate, tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors [SNRIs], calcium channel blockers, anti-hypertensive and other migraine medications) between 1 January 2006 and 31 December 2012. The initial prophylactic migraine medication prescription date was the index date. Patients initiating treatment were required to have a washout period with no prescription claims for a prophylactic migraine medication in the 12 months prior to the index date. All patients were required to be 18 years of age or older and to have continuous health plan coverage with medical and pharmacy benefits for at least 12 months before and 24 months following initiation of prophylactic medication. Patients who had a diagnosis of epilepsy during the 12-month baseline period were excluded. The follow-up period was of variable length and patients were followed for at least 24 months and up to 60 months from their index date until disenrollment, death, or end of study (31 December 2014), whichever occurred first.

Study outcomes

Treatment patterns were evaluated during the variable-length follow-up period and included proportion of patients initiating prophylactic migraine medication, treatment persistence (continuous use of any prophylactic migraine medication without a gap of >60 days after the previous prophylactic migraine prescription supply ran out, and separately by continuous use of index prophylactic medication class), and non-persistence (presence of a gap of >60 days in supply after the previous prescription ran out). Treatment modifications such as switching (adding a new prophylactic medication while discontinuing the initial medication), restarting (restarting a prophylactic migraine medication following a gap of >60 days in supply after the previous prescription ran out), discontinuation (ending treatment with an index medication without evidence of another refill within 60 days of exhausting the drug supply for the prior prescription), and augmentation (initiation of a prophylactic migraine medication prior to a 60-day gap in the supply of the previous prophylactic migraine prescription claim) were evaluated.

Utilization of acute migraine medications (triptans, ergotamines, NSAIDs, and opioids) during the baseline period and up to five years following the initiation of migraine prophylactic medications was evaluated. Incidence rates of a gastrointestinal (GI)-related AE and opioid abuse following opioid use were reported among opioid users in the follow-up period and among opioid users stratified by the number of days’ supply of opioids received during the follow-up period. GI-related AEs were identified by a medical claim with a diagnosis of nausea/vomiting, diarrhea, constipation (ICD-9-CM diagnosis codes 564.xx, 787.0x, 787.91) or by a prescription claim for the constipation medications lubiprostone or linaclotide. Opioid abuse was identified by a medical claim with a diagnosis of opioid abuse (ICD-9-CM diagnosis codes 304.0x, 304.7x, 305.5x, 965.00, 965.02, 965.09).

Covariates

Patient demographic characteristics, including age, gender, geographic region (US census division), urbanicity (based on urban or rural metropolitan statistical area (MSA)), and payer type, were measured on the index date. Other clinical characteristics were measured in the 12-month baseline period. Clinical characteristics included headache, chronic pain, hypertension, low back pain, depression, anxiety, osteoarthritis, inflammatory bowel disease, nausea/vomiting, asthma, fibromyalgia, menopause or menopause-related symptoms, constipation, cardiac arrhythmia, chronic migraine, diabetes, diarrhea, endometriosis, irritable bowel syndrome, menorrhagia, myocardial infarction, osteoporosis, and stroke.

Statistical analyses

Descriptive statistics were used to summarize all study variables, including demographic and clinical characteristics and treatment patterns. Categorical variables were summarized as counts and percentages while those for continuous variables were presented as means and standard deviations. Cox proportional hazards models were used to evaluate predictors of first-line treatment discontinuation (>60-day gap), switching and restarting. In addition, Cox proportional hazard models were used to assess the hazard of clinical events associated with the increasing number of opioid days supplied. Clinical events included opioid abuse and GI-related events, including constipation, nausea/vomiting, or diarrhea. All models were controlled for patients’ demographic (e.g. age, sex, payer type, geographic region), clinical characteristics (e.g. comorbid conditions), and first-line prophylactic medication (other anticonvulsants, tricyclic antidepressants, antihypertensives, beta blockers, calcium channel blockers, other migraine medications, and SNRIs). Hazard ratios (HRs) and p-values were calculated; a p-value <0.05 was set as the threshold for statistical significance.

Results

Patient disposition

A total of 147,832 patients met the study eligibility criteria and were included in the study (Figure 1). Of these, 95,793 (64.8%) patients had at least 3 years of follow-up post-prophylactic medication initiation and 39,999 (27.1%) patients had at least 5 years of follow-up.

Figure 1.

Patient selection.

Patient characteristics

Table 1 summarizes the baseline demographics and clinical characteristics of the study population. Mean age of the study cohort was 43.6 years (SD = 12.4), with women comprising 83% of the cohort. The patients were demographically similar to the overall population of patients in the MarketScan Research Databases. Patients were mostly located in the South region of the United States (42.1%), resided in an urban MSA (83.7%), and were insured through a commercial insurance (96.0%). The average duration of follow-up was approximately 4 years. The most common baseline comorbidities were headache, chronic pain, hypertension, low back pain, and depression.

Table 1.

Demographic and clinical characteristics of patients initiating first prophylactic migraine medication.

Characteristics	All prophylactic migraine treatments (n = 147,832)
Mean (SD) age, in years	43.6 (12.4)
Sex, n (%)
Female	122,726 (83.0)
Geographic region, n (%)
Northeast	19,445 (13.2)
North Central	34,608 (23.4)
South	62,178 (42.1)
West	30,364 (20.5)
Unknown	1237 (0.8)
Payer, n (%)
Commercial	141,863 (96.0)
Medicare	5969 (4.0)
Urbanicity, n (%)
Urban	123,768 (83.7)
Rural	22,931 (15.5)
Unknown	1133 (0.8)
Mean (SD) duration of follow-up, in days	1495.9 (625.1)
Deyo-Charlson Comorbidity Index, mean (SD)	0.3 (0.9)
Comorbid conditions* n (%)
Headache	47,418 (32.1)
Chronic pain	33,752 (22.8)
Hypertension	25,948 (17.6)
Low back pain	23,631 (16.0)
Depression	19,445 (13.2)
Anxiety	14,623 (9.9)
Osteoarthritis	13,952 (9.4)
Inflammatory bowel disease	13,519 (9.1)
Nausea/vomiting	12,667 (8.6)
Asthma	10,512 (7.1)
Fibromyalgia	8604 (5.8)
Menopause or menopause- related symptoms	8268 (5.6)
Constipation	7389 (5.0)

SD: standard deviation.

Comorbidities ≥5% are presented. Other less prevalent comorbid conditions (<5%) were cardiac arrhythmia, chronic migraine, diabetes, diarrhea, endometriosis, irritable bowel syndrome, menorrhagia, myocardial infarction, osteoporosis, and stroke.

Treatment patterns of first-line prophylactic migraine medications

Overall, 9.7% of first-line patients were persistent on therapy (average, 3.6 years). However, the vast majority of migraine patients (90.3%) were found to be non-persistent to the first-line prophylactic medications. Of the non-persistent patients, 38.9% switched, 30.1% restarted, and 31.0% discontinued treatment. The average time to switch, restart, or discontinue therapy was 186 days (Table 2).

Table 2.

Prophylactic migraine treatment patterns with first-line medication.

	All prophylactic migraine treatments (n = 147,832)
Overall duration of treatment, in days, mean (SD)¹	297.2 (475.6)
Treatment persistence², n (%)	14,314 (9.7)
Days of persistence, mean (SD)	1331.1 (552.8)
Treatment non-persistence³, n (%)	133,518 (90.3)
Days to non-persistence, mean (SD)	186.4 (301.4)
Reason for treatment non-persistence
Switch (within 60 days), n (%)	51,981 (38.9)
Restart (gap of >60 days with restart), n (%)	40,136 (30.1)
Discontinuation (gap of >60 days with no restart), n (%)	41,401 (31.0)
Augmentation⁴, n (%)	23,179 (15.7)

SD: standard deviation.

The overall duration of treatment is calculated among persistent and non-persistent patients prior to 60-day gap or end of follow-up.

Persistence is defined as continuous use of any prophylactic migraine medication without a gap >60 days following the days of supply of the previous prophylactic migraine prescription claim running out.

Non-persistence with a prophylactic migraine medication is defined as a >60-day gap in therapy following the days of supply of the previous prescription running out.

Augmentation is defined as initiation of a prophylactic migraine medication (other anticonvulsants, topiramate, beta-blockers, tricyclic antidepressants, SNRIs, calcium channel blockers, antihypertensives, and other) prior to a 60-day gap in days of supply of a previous prophylactic migraine prescription claim.

After adjusting for covariates in the proportional hazards model, the risk of discontinuation was significantly lower for patients aged between 35–64 years, male sex, rural MSA, having Medicare coverage, having an index year between 2006–2012, and having lower pre-index Deyo Charlson Comorbidity Index (CCI) score (all p < 0.05). A significantly lower risk of discontinuation was also observed among patients with select pre-index comorbidities (anxiety, asthma, chronic migraine, chronic pain, depression, diabetes, fibromyalgia, headache, hypertension, IBD, myocardial infarction, nausea, and osteoarthritis), and first-line prophylactic medication use with other anticonvulsants (i.e. non-topiramate), tricyclic antidepressants, antihypertensives, beta blockers, other migraine medications, and SNRIs (all p < 0.05) (Table 3). Results for treatment switching and restarting showed similar trends for most predictors, though magnitude of effect estimate and statistical significance of individual predictor varied.

Table 3.

Hazard ratio of switching, restart, and discontinuation among migraine patients initiating prophylactic medication.

Parameter	Discontinuation		Switch		Restart
HR	p-value	HR	p-value	HR	p-value
Age groups (reference 18–34)
35–44	0.74	<0.001	0.95	<0.001	0.93	<0.001
45–54	0.70	<0.001	0.86	<0.001	0.84	<0.001
55–64	0.70	<0.001	0.76	<0.001	0.78	<0.001
65–74	0.91	0.40	0.66	<0.001	0.77	0.004
75+	0.85	0.20	0.65	<0.001	0.72	0.001
Gender (reference female)
Male	0.90	<0.001	1.00	0.88	0.99	0.38
Plan type (reference other)
Fee for service	0.94	0.07	1.04	0.14	1.08	0.02
HMO	0.98	0.30	1.03	0.09	1.08	0.001
Point of sale	0.97	0.14	1.13	<0.001	1.10	<0.001
Preferred provider	0.98	0.27	1.02	0.18	1.02	0.46
Region (reference South)
Northeast	1.10	<0.001	0.88	<0.001	0.87	<0.001
North Central	0.97	0.04	0.94	<0.001	0.90	<0.001
West	1.07	<0.001	1.02	0.08	0.93	<0.001
Unknown	1.70	0.003	1.28	0.14	2.02	<0.001
Payer (reference commercial)
Medicare	0.74	0.005	1.17	0.03	0.99	0.93
Urbanicity (reference urban)
Rural	0.91	<0.001	1.02	0.098	0.97	0.03
Unknown	0.58	0.004	0.64	0.01	0.45	<0.001
Index year
2006	0.64	<0.001	1.18	<0.001	0.98	0.45
2007	0.69	<0.001	1.17	<0.001	1.00	0.84
2008	0.78	<0.001	1.11	<0.001	1.02	0.26
2009	0.84	<0.001	1.08	<0.001	1.02	0.38
2011	1.10	<0.001	0.94	<0.001	0.96	0.03
2012	1.23	<0.001	0.85	<0.001	0.91	<0.001
2013	0.93	0.86	0.39	0.11	0.94	0.87
Pre-period conditions
CCI	0.98	0.03	1.01	0.40	0.99	0.07
Anxiety	0.93	<0.001	1.07	<0.001	1.00	0.96
Asthma	0.96	0.05	1.07	<0.001	1.03	0.17
Arrhythmia	1.03	0.19	0.93	0.005	1.02	0.41
Chronic migraine	0.62	<0.001	0.84	<0.001	0.71	<0.001
Chronic pain	0.90	<0.001	1.09	<0.001	0.99	0.48
Constipation	1.02	0.66	0.99	0.68	0.98	0.53
Depression	0.77	<0.001	1.07	<0.0001	0.97	0.07
Diabetes	0.80	<0.001	1.07	0.003	1.13	<0.001
Endometriosis	0.94	0.22	1.10	0.011	0.97	0.57
Fibromyalgia	0.82	<0.001	1.10	<0.001	0.97	0.15
Headache	0.97	0.02	1.15	<0.001	0.97	0.002
Hypertension	0.6	<0.001	0.99	0.22	1.02	0.19
IBD	0.94	0.02	1.09	<0.001	1.04	0.09
IBS	0.95	0.26	0.99	0.78	0.94	0.20
Lower back pain	1.02	0.41	1.07	<0.001	1.06	0.02
Menopausal symptoms	1.02	0.32	1.00	0.87	1.00	0.81
Menorrhagia	1.03	0.29	1.08	0.002	1.08	0.005
Myocardial infarction	0.70	<0.001	0.70	<0.001	0.76	0.001
Nausea	0.96	0.031	1.121	<0.0001	1.01	0.46
Osteoporosis	1.01	0.799	1.003	0.903	0.99	0.56
Osteoarthritis	0.85	<0.0001	1.100	<0.0001	0.99	0.47
Stroke	0.97	0.364	1.010	0.746	0.89	0.002
First-line drug (reference topiramate)
Other anticonvulsants	1.45	<0.001	1.31	<0.001	1.39	<0.001
Tricyclic antidepressants	1.32	<0.001	1.38	<0.001	1.08	<0.001
Antihypertensives	0.80	<0.001	0.63	<0.001	0.95	0.02
Beta blockers	0.88	<0.001	0.85	<0.001	0.99	0.43
Calcium channel blockers	1.03	0.36	1.24	<0.001	0.77	<0.001
Other migraine medications	2.41	<0.001	2.09	<0.001	2.31	<0.001
SNRIs	0.95	0.003	0.76	<0.001	0.73	<0.001

CCI: Charlson comorbidity index; HR: hazard ratio; HMO: health maintenance organization; IBD: inflammatory bowel disease; IBS: irritable bowel syndrome; SNRIs: serotonin-norepinephrine reuptake inhibitors.

Treatment with acute migraine medications

During the entire duration of follow-up, approximately 59.9% of patients received acute migraine treatment with triptans, 66.6% received NSAIDS, and 77.4% of patients received strong or weak opioids. Ergotamine use was rare and was observed in less than 3% of patients. Among patients who initiated prophylactic treatment, the mean total number of prescription claims per patient during the entire follow-up period was 13.5 (SD = 17.3) for triptans, 13.6 (SD = 24.2) for opioids, and 6.4 (SD = 9.0) for prescription NSAIDs (Table 4).

Table 4.

Acute migraine treatment patterns in the follow-up period.

Treatment patterns	Entire follow-up (n = 147,832)
Any acute migraine medication
Acute migraine medication use, n (%)	139,673 (94.5)
Total number of prescription claims, mean (SD)	24 (31.4)
Triptans
Triptan use, n (%)	88,502 (59.9)
Total number of prescription claims, mean (SD)	13.5 (17.3)
Ergotamines
Ergotamine use, n (%)	3830 (2.6)
Total number of prescription claims, mean (SD)	3.6 (7.9)
Prescription NSAIDs
Prescription NSAIDs use, n (%)	98,431 (66.6)
Total number of prescription claims, mean (SD)	6.4 (9.0)
Opioids (strong and weak)¹
Opioid use, n (%)	114,417 (77.4)
Total number of prescription claims, mean (SD)	13.6 (24.2)

NSAIDs: nonsteroidal anti-inflammatory drugs; SD: standard deviation.

NSAID and opioid combination drugs are counted in the opioid category and not double counted.

GI-related AEs among opioid users

An analysis evaluating the incidence of AEs was conducted for the 49,676 prophylactic migraine patients receiving acute migraine treatment with opioids in the follow-up period. In total, 27.7% experienced any GI-related AE. Specifically, 16.6% were reported to have had nausea/vomiting, 12.2% constipation, and 10.5% diarrhea during the entire length of follow-up period. Less than 1% of patients were reported to have had an opioid abuse diagnosis over five years of follow-up. Incidence rates per 100 person-years was 7.9 for any GI-related AEs, 6.1 for nausea/vomiting, 3.7 for diarrhea, and 4.4 for constipation. The majority of opioid users (79.8%) had 1–45 days’ supply of opioids on hand, which decreased to 13.3% for 46–180 days’ supply, 3.0% for 180–365 days’ supply, and 3.9% for 365+ days’ supply. Incidence of GI-related AEs and opioid abuse increased with increasing number of days’ supply of opioids (Table 5).

Table 5.

Potential incident opioid-related clinical events among prophylactic migraine patients.

Potential opioid-related clinical events	Opioid users, from initiation of opioids to end of follow-up¹ (n = 49,676)	Opioid users by days’ supply
1–45 days (n = 39,651)	46–180 days (n = 6588)	181–365 days (n = 1483)	>365 days (n = 1954)
Any gastrointestinal event
Number of patients at risk for clinical event, n (%)	42,731 (86.0)	34,313 (86.5)	5516 (83.7)	1251 (84.4)	1651 (84.5)
Number of patients with an incident clinical event, n (%)	11,835 (27.7)	8282 (24.1)	2166 (39.3)	562 (44.9)	825 (50.0)
Incidence rate of clinical event, in 100 person-years (95% CI)	7.9 (7.9–7.9)	6.9 (6.9–6.9)	11.2 (11.2–11.3)	13.6 (13.5–13.6)	14.5 (14.5–14.6)
Constipation diagnosis or prescription constipation medication (lubiprostone or linaclotide)
Number of patients at risk for clinical event, n (%)	46,965 (94.5)	37,544 (94.7)	6195 (94.0)	1386 (93.5)	1840 (94.2)
Number of patients with an incident clinical event, n (%)	5738 (12.2)	3793 (10.1)	1162 (18.7)	301 (21.7)	482 (26.2)
Incidence rate of clinical event, in 100 person-years (95% CI)	4.4 (4.4–4.4)	3.9 (3.9–3.9)	5.5 (5.4–5.5)	6.3 (6.2–6.3)	7.0 (6.9–7.0)
Nausea/vomiting diagnosis
Number of patients at risk for clinical event, n (%)	46,000 (92.6)	36,850 (92.9)	6006 (91.2)	1359 (91.6)	1785 (91.4)
Number of patients with an incident clinical event, n (%)	7626 (16.6)	5128 (13.9)	1498 (24.9)	391 (28.8)	609 (34.1)
Incidence rate of clinical event, in 100 person-years (95% CI)	6.1 (6.1–6.1)	5.5 (5.5–5.5)	7.6 (7.6–7.6)	8.6 (8.6–8.7)	9.4 (9.3–9.4)
Diarrhea diagnosis
Number of patients at risk for clinical event, n (%)	47,719 (96.1)	38,148 (96.2)	6268 (95.1)	1424 (96.0)	1879 (96.2)
Number of patients with an incident clinical event, n (%)	4987 (10.4)	3352 (8.8)	1005 (16.0)	262 (18.4)	368 (19.6)
Incidence rate of clinical event, in 100 person-years (95% CI)	3.7 (3.7–3.7)	3.4 (3.3–3.4)	4.6 (4.6–4.6)	5.2 (5.1–5.2)	4.9 (4.9–5.0)
Opioid abuse
Number of patients at risk for clinical event, n (%)	49,594 (99.8)	39,621 (99.9)	6572 (99.8)	1474 (99.4)	1927 (98.6)
Number of patients with an incident clinical event, n (%)	313 (0.6)	85 (0.2)	53 (0.8)	42 (2.8)	133 (6.9)
Incidence rate of clinical event, in 100 person-years (95% CI)	0.2 (0.2–0.2)	0.08 (0.08–0.08)	0.2 (0.2–0.2)	0.7 (0.7–0.7)	1.6 (1.6–1.6)

CI: confidence interval.

Clinical events for opioid users were identified following the initial opioid prescription claim in the follow-up period.

After adjusting for baseline demographic and clinical characteristics with Cox proportional hazards models, the adjusted hazard of any GI event was 66% higher in patients with 46–180 days of opioid supply (HR: 1.66; p < 0.001) and increased to 95% higher (HR: 1.95; p < 0.001) in those with >365 days’ supply, compared to patients with 1–45 days of medication supply. Relative to patients with 1–45 days of medication supply, the risk of constipation, nausea/vomiting, and diarrhea increased by 34–47%, 47–68% and 34–45% (all p < 0.001) compared to patients with 46 to >365 days of medication supplied, respectively. Similarly, the risk of opioid abuse increased 3–8 times in patients with 46 to >365 days of opioid supply, while it increased 23 times among those with >365 days’ supply (both p < 0.001) (Table 6).

Table 6.

Hazard ratio of risk of any GI event, constipation, nausea, diarrhea, and opioid abuse among migraine patients initiating prophylactic medication.

Parameter	Any GI event		Constipation		Nausea		Diarrhea		Opioid abuse
HR	p-value	HR	p-value	HR	p-value	HR	p-value	HR	p-value
Days’ supplied (reference 1–45 days)
46–180	1.66	<0.001	1.34	<0.001	1.47	<0.001	1.34	<0.001	3.00	<0.001
181–365	1.84	<0.001	1.43	<0.001	1.56	<0.001	1.42	<0.001	7.61	<0.001
>365	1.95	<0.001	1.47	<0.001	1.68	<0.001	1.45	<0.001	22.54	<0.001
Age groups (reference 18–34)
35–44	0.85	<0.001	1.00	0.90	0.75	<0.001	0.86	<0.001	0.59	<0.001
45–54	0.76	<0.001	0.97	0.22	0.59	<0.001	0.77	<0.001	0.42	<0.001
55–64	0.74	<0.001	1.09	0.004	0.54	<0.001	0.82	<0.001	0.32	<0.001
65–74	0.85	0.09	1.27	0.04	0.63	<0.001	0.76	0.03	0.19	<0.001
75+	1.00	0.99	1.65	<0.001	0.73	0.01	0.92	0.53	0.15	<0.001
Gender (reference female)
Male	1.52	<0.001	1.65	<0.001	1.60	<0.001	1.31	<0.001	0.62	<0.001
Plan type (reference other)
Fee for service	1.00	0.90	0.95	0.32	0.98	0.66	0.92	0.11	1.25	0.08
HMO	1.03	0.12	0.97	0.41	0.97	0.39	0.94	0.10	1.05	0.60
Point of sale	0.99	0.60	0.91	0.02	0.93	0.02	0.90	0.007	1.05	0.64
Preferred provider	1.03	0.25	1.00	0.97	0.97	0.32	0.95	0.11	1.03	0.71
Region (reference South)
Northeast	0.82	<0.001	0.99	0.66	0.72	<0.001	0.97	0.24	1.28	<0.001
North Central	0.86	<0.001	0.89	<0.001	0.82	<0.001	1.01	0.73	0.88	0.01
West	0.90	<0.001	0.93	0.002	0.85	<0.001	1.00	0.94	1.38	<0.001
Unknown	0.68	0.09	1.03	0.92	0.51	0.04	0.53	0.15	0.00	0.92
Payer (reference commercial)
Medicare	1.04	0.64	1.15	0.20	0.94	0.54	1.04	0.71	1.03	0.90
Urbanicity (reference urban)
Rural	1.05	0.003	1.00	0.87	1.11	<0.001	1.01	0.70	0.97	0.58
Unknown	1.36	0.20	0.89	0.72	1.92	0.06	2.12	0.10	1627.85	0.92
Index year
2006	0.93	0.001	0.87	<0.001	0.71	<0.001	0.89	<0.001	0.61	<0.001
2007	0.93	<0.001	0.86	<0.001	0.78	<0.001	0.90	<0.001	0.72	<0.001
2008	0.96	0.06	0.91	0.001	0.85	<0.001	0.95	0.07	0.76	<0.001
2009	1.01	0.70	0.95	0.08	0.95	0.03	1.03	0.32	0.92	0.20
2011	0.99	0.63	1.05	0.07	1.01	0.77	1.03	0.33	1.25	0.002
2012	1.03	0.17	1.09	0.006	1.18	<0.001	1.15	<0.001	1.83	<0.001
2013	1.11	0.86	0.00	0.82	2.68	<0.05	0.93	0.9	0.00	0.96
Pre-period conditions
CCI	1.07	<0.001	1.05	<0.001	1.07	<0.001	1.07	<0.001	1.00	0.85
Anxiety	1.19	<0.001	1.18	<0.001	1.20	<0.001	1.22	<0.001	1.44	<0.001
Asthma	1.16	<0.001	1.15	<0.001	1.13	<0.001	1.22	<0.001	1.09	0.24
Arrhythmia	1.14	<0.001	1.11	0.009	1.14	<0.001	1.19	<0.001	1.08	0.42
Chronic migraine	1.05	0.42	0.97	0.74	1.02	0.75	1.02	0.85	0.90	0.59
Chronic pain	1.14	<0.001	1.18	<0.001	1.14	<0.001	1.13	<0.001	1.23	0.003
Constipation	1.09	0.13	1.78	<0.001	0.96	0.40	0.74	<0.001	1.14	0.25
Depression	1.16	<0.001	1.14	<0.001	1.17	<0.001	1.18	<0.001	1.62	<0.001
Diabetes	1.15	<0.001	1.03	0.45	1.29	<0.001	1.18	<0.001	1.06	0.49
Endometriosis	1.17	0.002	1.31	<0.001	1.12	0.03	1.15	0.04	1.11	0.41
Fibromyalgia	1.13	<0.001	1.28	<0.001	1.10	<0.001	1.15	<0.001	1.04	0.55
Headache	1.18	<0.001	1.13	<0.001	1.23	<0.001	1.11	<0.001	1.14	0.001
Hypertension	1.07	<0.001	1.02	0.45	1.13	<0.001	1.06	0.008	1.00	0.97
IBD	1.44	<0.001	1.27	<0.001	1.40	<0.001	1.77	<0.001	1.07	0.37
IBS	1.01	0.86	1.22	0.02	1.00	0.95	1.41	<0.001	0.81	0.12
Lower back pain	1.01	0.70	0.96	0.15	1.01	0.84	0.99	0.84	1.11	0.12
Menopausal symptoms	1.16	<0.001	1.20	<0.001	1.13	<0.001	1.13	0.001	0.94	0.50
Menorrhagia	1.10	0.001	1.05	0.24	1.07	0.04	1.04	0.36	0.96	0.70
Myocardial infarction	0.95	0.64	0.93	0.59	0.99	0.95	0.90	0.43	1.26	0.45
Nausea	1.21	<0.001	1.31	<0.001	1.66	<0.001	1.36	<0.001	1.33	<0.001
Osteoporosis	1.13	<0.001	1.15	<0.001	1.11	0.001	1.08	0.05	0.94	0.50
Osteoarthritis	1.06	0.001	1.12	<0.001	0.98	0.44	0.99	0.77	1.17	0.003
Stroke	1.06	0.10	1.09	0.08	0.97	0.53	1.03	0.53	0.75	0.07
First-line drug (reference topiramate)
Other anticonvulsants	1.04	0.05	0.99	0.66	1.02	0.40	0.98	0.42	1.31	<0.001
Tricyclic antidepressants	1.01	0.70	0.98	0.56	1.05	0.03	0.99	0.73	1.13	0.10
Antihypertensives	0.92	0.001	0.81	<0.001	0.98	0.56	0.96	0.23	0.95	0.61
Beta blockers	0.96	0.04	0.89	<0.001	0.98	0.39	0.98	0.49	1.16	<0.05
Calcium channel blockers	0.93	<0.05	0.84	0.001	0.96	0.37	0.95	0.36	0.85	0.29
Other migraine medications	1.00	0.85	0.90	0.002	0.97	0.31	0.96	0.24	1.54	<0.001
SNRIs	1.00	0.87	0.95	0.09	0.99	0.63	0.95	0.12	1.42	<0.001

CCI: Charlson comorbidity index; GI: gastrointestinal; HR: hazard ratio; HMO: health maintenance organization; IBD: inflammatory bowel disease; IBS: irritable bowel syndrome; MPR: medication possession ratio; SNRIs: serotonin-norepinephrine reuptake inhibitors.

Discussion

In this US administrative claims-based analysis of adults with migraine who initiated prophylactic therapy, long-term persistence with first-line prophylactic medications was poor. Over 90% of the migraine patients either discontinued, restarted or switched from their initial prophylactic treatment a little after the first 6 months of initiation. Additional treatment with acute migraine medications (opioids, NSAIDs, triptans, and ergotamine) to reverse or abort an ongoing migraine was common among migraine patients that were receiving prophylactic therapy. Furthermore, GI-related AEs and opioid abuse related to use acute migraine medications increased with longer duration of treatment with opioids.

Reports on short-term persistence with prophylactic therapy have been published previously (19,23,24,26 –28,30). Data from our study showed that less than 10% of migraine patients persisted with index prophylactic therapy over a 4 year average duration of follow-up, which is consistent with the rates reported in previous studies (23,24,26,27,31). Furthermore, our study suggested younger age, lower pre-index CCI score, history of select comorbidities such as anxiety, asthma, chronic migraine, chronic pain, depression, diabetes, fibromyalgia, headache, hypertension, IBD, myocardial infarction, nausea, and osteoarthritis, and first-line medication use with other anticonvulsants, tricyclic antidepressants, antihypertensives, beta blockers, other migraine medications, and SNRIs in the pre-index period were significantly associated with lower risk of discontinuation from the initial prophylactic treatment. Taken together, these findings suggest that migraine patients receiving treatment with existing prophylactic medications have poor persistence in the real world. Although this study did not investigate the reasons for non-persistence, the relatively low persistence rate can possibly be due to a lack of perceived efficacy and/or undesirable side effects associated with the use of prophylactic medication in migraine, which appears to be even more distinct among patients who are older and have a higher disease burden (23,24,27,32,33).

Frequent and long-term use of acute therapies are not recommended; however, the majority of patients initiating migraine prophylaxis use both specific and non-specific acute medications to supplement prophylactic migraine medications to achieve rapid pain relief during a migraine (6). In an analysis of data from the 2005 American Migraine Prevalence and Prevention (AMPP) study, 91.7% of respondents meeting criteria for episodic migraine used acute treatment for their headaches (34). Similarly, in a recent US claims-based analysis that assessed the use of specific and non-specific acute migraine medication before and after prophylaxis, the majority of patients (84%) were found to use at least one acute medication within 1 year of initiating prophylaxis, which increased to 97% over 5 years of follow-up (35). Consistent with these findings, the use of acute migraine medication was widespread in this cohort of patients who initiated prophylactic treatment. Analysis of the data from the 2009 AMPP study demonstrated that opioids are widely used as acute treatment for migraine (36). Furthermore, 40–90% of opioid-treated patients experienced any GI-related AE (13,14,37 –39), and the rate of AEs was found to increase with the duration of opioid prescriptions (40,41). Data from the current study extends support to these findings, as opioids were the most commonly used acute medications in this study cohort, and the risk of any GI-related AEs (constipation, nausea/vomiting, and diarrhea) and opioid abuse increased with increasing days of opioid supply.

This analysis is subject to several limitations. First, this analysis evaluated individuals with migraine initiating prophylactic treatments; therefore, the results may not be generalizable to migraine patients not treated with prophylactic treatments or those previously treated with prophylactic treatments. Furthermore, this analysis evaluated a patient sample that reflected the geographic distribution of the overall population in the MarketScan Research Databases population and was not balanced to control for differences in urban versus rural MSA or by geographic region. Moreover, incidence rates of AEs were evaluated within the subset of patients in the study population treated with opioids. This study did not describe whether opioid users differed from the overall patient sample or whether opioid users differed by increasing days’ supply of opioids.

Several limitations that are inherent to claims database analyses apply to this study as well. As with all administrative healthcare database, the MarketScan Research databases rely on administrative claims data for clinical detail; therefore, the data may be subject to coding limitations, potentially introducing measurement error. Furthermore, patients were selected based on the presence of diagnoses based on ICD-9-CM diagnosis codes and the receipt of prescription medication based on NDC codes and HCPCS codes. Under-diagnosis, as well as coding errors, may affect study selection to an extent that cannot be quantified. To limit the impact of migraine undercoding in administrative claims data, a previously published algorithm using the combination of a migraine diagnosis and prescription claims for prophylactic migraine medications was used to identify patients with chronic migraine. However, this database does not enable the characterization of migraine when medical care was not provided. Furthermore, patients who did not meet the continuous enrollment criteria, without any health coverage, or who died during the follow-up, were excluded. Therefore, it may be possible that the patients included in the analysis may have been healthier than the total population of patients in this claims database. Medication exposure was based on fill prescriptions. While we assumed these patients took medications as prescribed, confirmation of this assumption was not possible. Lastly, this study primarily included migraine patients enrolled with either Commercial or Medicare Supplemental health coverage, thus results of this analysis may not be generalizable to all migraine patients with other insurance or populations without health insurance coverage.

This study does offer some potential advantages – in contrast to most previous retrospective studies, the current study provides contemporary long-term data on the persistence of prophylactic treatment among patients with migraine from two large administrative claims databases. Furthermore, the wide range of information on patients’ demographic factors, medical and medication history allows multivariable adjusted analysis to include many important potential confounders.

In summary, the results of this study suggest that long-term persistence with prophylactic medications in migraine patients was poor, reflecting an unmet need for improved prophylactic therapies. Use of acute migraine medications, especially opioids, coupled with increased rates of opioid-related AEs, were common in migraine patients treated with prophylactic therapy. New treatment options with improved efficacy and tolerability are warranted to address these concerns.

Clinical implications

Long-term persistence with prophylactic medications in migraine patients is poor.

Acute migraine medications, including opioids, are used frequently to manage migraine pain.

Increasing use of opioids is associated with higher rates of opioid-related adverse events.

There is an unmet need for improved prophylactic therapies to manage migraine.

Footnotes

Acknowledgements

Editorial/writing assistance for this manuscript was provided by Santosh Tiwari, who is an employee of IBM Watson Health.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: This study was conducted by IBM Watson Health. All listed authors meet the criteria for authorship set forth by the International Committee for Medical Journal Editors. Machaon Bonafede and Kathleen Wilson are employees of IBM Watson Health, which received compensation from Amgen Inc. to conduct the study. Fei Xue is an employee and stockholder of Amgen Inc.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Amgen Inc., USA.

References

Migraine Research Foundation. Migraine fact sheet, http://wwwmigraineresearchfoundationorg/fact-sheet.html (2017, accessed 1 March 2017).

Shaik

Hassan

Tan

et al.

Quality of life and migraine disability among female migraine patients in a tertiary hospital in Malaysia. BioMed Res Int 2015; 2015: 523717–523717.

Gilmore

Michael

. Treatment of acute migraine headache. Am Fam Physician 2012; 15(Suppl 1): S33–S39.

Miller

. The acute and preventative treatment of episodic migraine. Ann Indian Acad Neurol 2012; 15: 33–33.

Silberstein S, Saper J and Freitag F. Migraine: Diagnosis and treatment. In: Silberstein SD, Lipton RB, D’Alessio DJ (eds) Wolff’s Headache and Other Head Pain. New York: Oxford University Press, 2001, pp.121–237.

Brown M, Warren K, Thornhill L, et al. Management of migraines. Alabama Pharm Association, https://wwwaparxorg/resource/resmgr/CEs/CE_Management_of_Migrainespdf (2014, accessed March 2017).

Boudreau

Von Korff

Rutter

et al.

Trends in long‐term opioid therapy for chronic non‐cancer pain. Pharmacoepidem Drug Safety 2009; 18: 1166–1175.

Volkow

McLellan

Cotto

et al.

Characteristics of opioid prescriptions in 2009. JAMA 2011; 305: 1299–1301.

Bigal

Lipton

. Excessive acute migraine medication use and migraine progression. Neurology 2008; 71: 1821–1828.

10.

Bigal

Serrano

Buse

et al.

Acute migraine medications and evolution from episodic to chronic migraine: A longitudinal population‐based study. Headache 2008; 48: 1157–1168.

11.

White

Arnold

Norvell

et al.

Pharmacologic management of chronic low back pain: Synthesis of the evidence. Spine 2011; 36: S131–S143.

12.

Baldini

Von Korff

Lin

. A review of potential adverse effects of long-term opioid therapy: A practitioner’s guide. Prim Care Companion CNS Disord 2012; 14: PCC.11m01326–PCC.11m01326.

13.

Cook

Lanza

Zhou

et al.

Gastrointestinal side effects in chronic opioid users: Results from a population‐based survey. Aliment Pharmacol Therap 2008; 27: 1224–1232.

14.

Kalso

Edwards

Moore

et al.

Opioids in chronic non-cancer pain: Systematic review of efficacy and safety. Pain 2004; 112: 372–380.

15.

D’Amico

Tepper

. Prophylaxis of migraine: General principles and patient acceptance. Neuropsych Dis Treat 2008; 4(6): 1155–1167.

16.

Estemalik

Tepper

. Preventive treatment in migraine and the new US guidelines. Neuropsych Dis Treat 2013; 9: 709–720.

17.

Silberstein

Holland

Freitag

et al.

Evidence-based guideline update: Pharmacologic treatment for episodic migraine prevention in adults, report of the Quality Standards Subcommittee of the American Academy of Neurology and the American Headache Society. Neurology 2012; 78: 1337–1345.

18.

Diamond

Bigal

Silberstein

et al.

Patterns of diagnosis and acute and preventive treatment for migraine in the United States: Results from the American Migraine Prevalence and Prevention study. Headache 2007; 47: 355–363.

19.

Lafata

Tunceli

Cerghet

et al.

The use of migraine preventive medications among patients with and without migraine headaches. Cephalalgia 2010; 30: 97–104.

20.

Lipton

Bigal

Diamond

et al.

Migraine prevalence, disease burden, and the need for preventive therapy. Neurology 2007; 68: 343–349.

21.

Muzina

Chen

Bowlin

. A large pharmacy claims-based descriptive analysis of patients with migraine and associated pharmacologic treatment patterns. Neuropsych Dis Treat 2011; 7: 663–672.

22.

Berger

Bloudek

Varon

et al.

Adherence with migraine prophylaxis in clinical practice. Headache 2013; 53: 1009–1010.

23.

Hepp

Bloudek

Varon

. Systematic review of migraine prophylaxis adherence and persistence. J Managed Care Pharm 2014; 20: 22–33.

24.

Hepp

Dodick

Varon

et al.

Persistence and switching patterns of oral migraine prophylactic medications among patients with chronic migraine: A retrospective claims analysis. Cephalalgia 2017; 37: 470–485.

25.

Hepp

Dodick

Varon

et al.

Adherence to oral migraine-preventive medications among patients with chronic migraine. Cephalalgia 2015; 35: 478–488.

26.

Lenz

Bonafede

Maiese

et al.

Prophylaxis and acute medication treatment patterns in migraine patients initiating migraine prophylactic therapy. Neurology 2016; 86(16 Suppl): P2.206–P2.206.

27.

Yaldo

Wertz

Rupnow

et al.

Persistence with migraine prophylactic treatment and acute migraine medication utilization in the managed care setting. Clin Ther 2008; 30: 2452–2460.

28.

Rapoport

Mauskop

Diener

et al.

Long‐term migraine prevention with topiramate: Open‐label extension of pivotal trials. Headache 2006; 46: 1151–1160.

29.

Woolley

Bonafede

Maiese

et al.

Prophylaxis and acute medication treatment patterns in migraine patients receiving prophylaxis. Headache 2015; 55: 175–175.

30.

Elston Lafata

Moon

Leotta

et al.

The medical care utilization and costs associated with migraine headache. J Gen Int Med 2004; 19: 1005–1012.

31.

Wong

Stock

Godley

et al.

Adherence and persistence to oral migraine prophylaxis medications in patients with chronic migraine. Value Health 2016; 19: A64–A64.

32.

Blumenfeld

Bloudek

Becker

et al.

Patterns of use and reasons for discontinuation of prophylactic medications for episodic migraine and chronic migraine: Results from the second International Burden of Migraine Study (IBMS‐II). Headache 2013; 53: 644–655.

33.

Stark

Valenti

Miller

. Management of migraine in Australian general practice. Med J Australia 2007; 187: 142–146.

34.

Lipton

Buse

Serrano

et al.

Acute medication use patterns in episodic migraine: Results of the American Migraine Prevalence and Prevention Study (AMPP). Cephalalgia 2009; 29(Suppl 1): 17–17.

35.

Bonafede M, Wilsonn K, Spencer D, et al. Acute medication utilization among migraine patients initiating prophylaxis. Headache 2016; 56: 22–23.

36.

Buse

Pearlman

Reed

et al.

Opioid use and dependence among persons with migraine: Results of the AMPP study. Headache 2012; 52: 18–36.

37.

Bell

Panchal

Miaskowski

et al.

The prevalence, severity, and impact of opioid‐induced bowel dysfunction: Results of a US and European patient survey (PROBE 1). Pain Med 2009; 10: 35–42.

38.

Holzer P. Methylnaltrexone for the management of unwanted peripheral opioid effects. Therapy 2008; 5: 531–543.

39.

Tuteja

Biskupiak

Stoddard

et al.

Opioid‐induced bowel disorders and narcotic bowel syndrome in patients with chronic non‐cancer pain. Neurogastroent Motil 2010; 22: e424–e496.

40.

Camilleri

. Opioid-induced constipation: Challenges and therapeutic opportunities. Am J Gastroent 2011; 106: 835–842.

41.

Herzig

Rothberg

Cheung

et al.

Opioid utilization and opioid‐related adverse events in nonsurgical patients in US hospitals. J Hosp Med 2014; 9: 73–81.

Long-term treatment patterns of prophylactic and acute migraine medications and incidence of opioid-related adverse events in patients with migraine

Abstract

Objectives

Methods

Results

Conclusions

Keywords

Introduction

Methods

Study design and data source

Patient selection

Study outcomes

Covariates

Statistical analyses

Results

Patient disposition

Patient characteristics

Treatment patterns of first-line prophylactic migraine medications

Treatment with acute migraine medications

GI-related AEs among opioid users

Discussion

Clinical implications

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

References