This was a retrospective cohort study using IQVIA open-source US pharmacy (LRx) and medical (Dx) claims data. The LRx database contains more than 250 million unique patients, across multiple payer and coverage structures including cash and manufacturers’ discount and coupon programs [7]. Data are collected via direct feeds from pharmacy suppliers capturing adjudicated and dispensed prescriptions sourced from retail, mail, long-term care and specialty pharmacies [7] including patient demographics, payer type, product information, 3-digit zip as well as prescription relevant information including prescriber, date of service, refill (medication claims after the initial prescription claim) number, quantity dispensed and days supply. The Dx medical claims database provides patient-level diagnoses, procedures, and administered therapeutics for over 1.5 billion claims per year for visits to US office-based physician, ambulatory, and general healthcare sites; and captures claims from commercial, Medicare, Medicaid, and cash payers [7]. Dx data was used in this study to obtain patients’ comorbidity history. All data in both databases are compliant with the Health Insurance Portability and Accountability Act to protect patients’ privacy.

Patients with ≥ 1 pharmacy claim for an nAMSM (ubrogepant, rimegepant, or lasmiditan) in the LRx database between January 1, 2020, through September 30, 2020 (the index period), were identified and assigned an index date on the date of their first nAMSM claim. This index period was selected to allow for a 6-month pre-index (6-month period prior to the index date) and a 6-month post-index period (6-month period after and including the index date) for assessment of patient characteristics. Additional patient identification criteria required for study inclusion were ≥ 18 years of age on the index date, linkage to Dx, ≥ 1 pharmacy claim of any type prior to the 6-month pre-index period and ≥ 1 pharmacy claim of any type after the end of the 6-month post-index period (ensuring pharmacy claims visibility) in LRx, use of a pharmacy that consistently contributed data during the 6-month pre-index and the 6-month post-index periods in LRx, ≥ 2 medical claims ≥ 30 days apart in the 6-month pre-index period (ensuring medical claims visibility for comorbidities) in Dx, and no data quality issues (defined as having ≥ 2 distinct index medications on index date or missing age or sex information).

Demographic characteristics (age and sex), prescriber of index nAMSM, and index payer type were assessed on the index date. Comorbidities were identified using International Classification of Diseases (ICD)-10 codes in the pre-index period with the most frequent comorbidities (> 5% of patients) being presented. In addition, the Charlson Comorbidity Index score was computed for each patient based on comorbidities identified during the pre-index period [18, 19].

For nAMSMs (ubrogepant, rimegepant, or lasmiditan), the index claim, dose, and quantity dispensed were identified. In the 6 months following the index nAMSM, the number of refills of the index nAMSM and time between refills were assessed, as well as any switches from the index nAMSM to a different nAMSM.

Use of anti-CGRP pathway mAbs, which included subcutaneously administered erenumab, fremanezumab, and galcanezumab, was evaluated in the pre-index period and in the post-index period. Intravenous administered eptinezumab was excluded from the analysis due to limited sample size.

Use of traditional acute and preventive anti-migraine medications (defined as having ≥ 1 claim) was evaluated in the pre-/post-index periods. Traditional acute anti-migraine medications included triptans, opioids, ergots (dihydroergotamine and ergotamine-containing products), and non-steroidal anti-inflammatory drugs (NSAIDs). Traditional preventive anti-migraine medications included select anticonvulsants (carbamazepine, gabapentin, levetiracetam, pregabalin, topiramate, valproate sodium, valproic acid, divalproex sodium, zonisamide), select antihypertensives (atenolol, bisoprolol, metoprolol, nadolol, nebivolol, pindolol, propranolol, timolol, verapamil, candesartan, clonidine including transdermal patches, lisinopril, olmesartan), select antidepressants (duloxetine, desvenlafaxine, venlafaxine, amitriptyline, desipramine, doxepin, imipramine, nortriptyline, protriptyline, clomipramine, escitalopram, citalopram, sertraline, mirtazapine), select botulinum toxins (abobotulinumtoxin A injection, incobotulinumtoxin A injection, onabotulinumtoxin A injection, rimabotulinumtoxin B injection), and other medications (carisoprodol, cyproheptadine, guanfacine, memantine, methysergide, milnacipran, tizanidine).

Discontinuation of traditional acute and preventive anti-migraine medications was reported at the class level. The discontinuation definitions differed for traditional acute and preventive anti-migraine medications, as traditional acute anti-migraine medications were taken as needed, while traditional preventive anti-migraine medications were taken on a scheduled basis. Discontinuation of traditional acute anti-migraine medications was evaluated among the subset of patients with ≥ 1 claim for the class in the pre-index period and ≥ 1 claim for the class in the post-index period and was defined as no claims for the same class in the 60 days prior to the end of the post-index period. Discontinuation of traditional preventive anti-migraine medications was evaluated among the subset of patients with concomitant preventive medication use on the index date (≥ 1 claim for the preventive medication class with days supply overlapping with the index date) and was defined as ≥ 60-day gap (≥ 90-day gap for botulinum toxins) in days supply of the preventive medication class in the post-index period.

Changes in equivalent dose of traditional acute anti-migraine medications (triptans, opioids, and ergots) from the pre-index period to the post-index period were measured among the subset of patients with pre-index acute anti-migraine medication use who continued use of the traditional acute anti-migraine medication during the post-index period. Triptans were converted to oral sumatriptan milligram (mg) equivalents, opioids were converted to oral morphine mg equivalents, and ergots were converted to oral ergotamine mg equivalents During each period (pre-/post-index period), the patient’s equivalent dose per month was calculated using the following formula:

$$Equivalent\;dose\;per\;month=\frac6$$

The change in equivalent dose was calculated as the post-index equivalent dose minus the pre-index equivalent dose, resulting in negative values that represent a reduction in dose and positive values representing an increase in dose in the post-index period.

All analyses were descriptive and were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC). Frequencies and percentages are presented for categorical variables, while means, standard deviations, and medians are presented for continuous variables. Analyses were conducted among the entire study population (all patients initiating nAMSMs) and then stratified among patients with prior mAb use (≥ 1 pre-index claim for erenumab, fremanezumab, or galcanezumab) and patients without prior mAb use.