This study demonstrated the continued real-world clinical benefits of erenumab over 12 months of follow-up among a large US patient population. Results from this study showed a decrease from the pre-index to the post-index period in acute and traditional migraine-preventive medication use as well as in ER visits and brain/head-imaging studies, with more-pronounced reductions among patients who were adherent versus those who were non-adherent to erenumab. Notably, these reductions in utilization resulted in a total cost decrease of nearly two thousand dollars per patient for adherent patients over the year of follow-up, reflecting reductions in both acute and traditional preventive-migraine medication costs (− $864) and associated medical utilization (− $1083).

Results from this study are generally consistent with other real-world analyses of patients initiating erenumab. The erenumab patient profile was similar to those in published studies, including a mean age between 44 and 51 years and a majority who were female (79–92%) [4]. The change in migraine type prevalence observed in this study is consistent with results observed in the Chandler et al. study, which also used MarketScan data but with a 6-month follow-up [10].

The reduction in acute migraine medication utilization after erenumab initiation observed in this study is well documented in the literature. Using Optum’s combined electronic health record and claims database, Tepper and colleagues observed significant reductions in migraine-specific acute medication use and HRU in the 6 months after erenumab initiation: a significant decrease in migraine-specific office visits from 86.2% to 77.6% and a decrease in claims for health care utilization of 10–19% [12, 13]. Hines and colleagues used IQVIA’s open-source claims database to report the discontinuation or reduced consumption of acute migraine-specific medication in the 6 months following erenumab initiation; among those patients with pre- and post-medication use, 36.8% discontinued their acute migraine-specific medication, based on there being no refills for ≥ 60 days [14]. Less information exists about the change in traditional migraine-preventive medication usage after erenumab initiation. Chandler et al. demonstrated a reduction in preventive medication use during the 6-month pre- to post-index period: the use of preventive medication (which included non-erenumab anti-CGRP pathway mAbs) changed from 87.3% in the pre-index period to 80.2% in the post-index period [10].

In this study, the proportion of patients using each specific acute and traditional migraine-preventive medication class decreased from the pre-index to the post-index period (e.g., triptan use decreased from 72.7% to 65.4% and anti-convulsants from 56.6% to 44.3%). Of note, this study included data through September 2020, allowing for the capture of newly approved medications (gepants/lasmiditan) in the post-index period that were not readily available during the pre-index period due to the relative timing of FDA approval.

While a reduction in migraine-preventive medication was observed from the pre-index to the post-index period, the use of traditional migraine-preventive medication was not replaced completely with the initiation of erenumab (e.g., 44.3% used anti-convulsants, 50.9% used anti-depressants, and 33.2% used anti-hypertensives in the post-index period). The literature supports the use of combination therapy in migraine prevention, potentially leading to better outcomes and lower dosages [15]. Approximately 40% of patients do not significantly benefit from monotherapy for migraine prevention, so using multiple agents from at least two different preventive medication classes may be beneficial [16]. In a retrospective analysis of medical charts from five US headache centers, chronic migraine patients treated with erenumab continued to be managed via a polypharmacy approach despite the demonstrated effectiveness of erenumab [17]. Finally, many of the traditional preventative medications are indicated not just for migraine but for other conditions (e.g., depression and hypertension), which can also contribute to their continued use. Particularly in this analysis using claims data, the indication (e.g., migraine versus depression) cannot be confirmed. But utilizing medications that result in positive benefits for coexisting conditions may be beneficial.

Despite the introduction of gepants and lasmiditan in the post-index period, the total medication cost (for acute and traditional migraine-preventive medications) decreased from the pre-index to the post-index period by $764. The medication cost reduction was nearly $900 for erenumab-adherent patients. The additional 6 months of follow-up in this study confirm the trends observed in Chandler et al.; total associated costs for healthcare resources among adherent patients were $770 lower during the 6-month follow-up period compared to the corresponding pre-index timeframe in Chandler et al. and $1083 lower during the 12-month follow-up period compared to the corresponding pre-index timeframe in this study [10]. A recent publication by Tepper et al. using US claims data from the Komodo Health database included erenumab costs and demonstrated an initial entrance cost associated with erenumab via an increase in migraine-related prescription costs during the 6-month pre- to post-index period; however, this was offset by reduced migraine-related medical costs and lower odds of migraine-related ER, office, and neurologist visits [18].

It is noted that differences in associated non-medication costs appear to be driven by small differences in the mean number of services. For example, the difference in ER visits per 100 patients was a reduction of 9.6 overall (− 18.5 in adherent and − 4.1 in non-adherent patients), and the difference in costs was a reduction of $167 overall (− $260 in adherent and − $109 in non-adherent patients). But for OP office visits, while the number of visits decreased among non-adherent patients (− 49.1 per 100 patients), an increase in OP visit costs of $816 ($9320 pre-index vs $10,136 post-index) was observed, suggesting that non-adherent patient visits were more intensive and therefore more costly.

This study measured persistence with and adherence to erenumab, with the results stratified by adherence and non-adherence. At 12 months, the reported persistence in this study was 40.1%, based on a 45-day gap. In Chandler et al., the reported persistence at 3 and 6 months following erenumab initiation was 72.6% and 47.3%, respectively, based on the same 45-day gap [10]. Mean PDC was 0.68 in the 6-month follow-up period in Chandler et al., compared to the value of 0.60 observed during the 12-month follow-up period in this study [10]. However, the 38.5% adherence over the 12-month post-index period in this study is higher than the 30.8% adherence measured over the 6-month post-index period using IQVIA’s open-source claims database [19].

The adherence to and persistence with erenumab reported in this study as well as others are higher than those for medications traditionally used for migraine prevention [20, 21]. A study using MarketScan data to compare anti-CGRP pathway mAbs (including erenumab, fremanezumab, and galcanezumab) to traditional migraine-preventive medications showed that 32.7% were adherent to anti-CGRP pathway mAbs, compared to 18.7% who were adherent to medications considered the standard of care in a 12-month follow-up [11].

This study was not intended as an economic analysis, so associated costs of anti-CGRP pathway mAbs were not included. Rather, the costs of acute or traditional migraine-preventive medications and the associated HRU were included as an indicator of the impact of erenumab that accounts for both changes in the proportion of patients utilizing and the extent of that utilization. Economic analyses of cost-effectiveness, clinical effectiveness, and value-based price for erenumab in the US have been previously published [22,23,24].

The results from this study highlight the impact of adherence to erenumab on associated medication and healthcare resource cost savings in the post-index period. Patients who were adherent to erenumab had an overall cost reduction (− $1947), whereas non-adherent patients had a cost increase ($101). Future studies may consider capturing both indirect costs from negative impacts on work productivity as well as direct costs, as this can provide a more complete story of the significant burden of migraine [24, 25]. For example, a Finnish study reported a 73.9% decrease in headache-related sick leave days in the 12 months following erenumab treatment among a cohort of employed patients with migraine [26].

The limitations of this study include those inherent in any retrospective claims-based analysis. This study was limited to medical claims from US-centric databases. Results of this analysis may not be generalizable to migraine patients overall or to other countries. The potential for misclassification of migraine status, covariates, or study outcomes was present, as patients were identified through administrative claims data as opposed to medical records. Claims data lack information on the frequency or severity of migraine attacks and do not distinguish among migraine subtypes (while there is an ICD-10-CM code for chronic migraine [G43.7*], there is no similar code for episodic migraine). A potential imbalance in healthcare resource utilization by migraine severity may result in less-precise cost estimates. As with any claims databases, the MarketScan Research Databases rely on administrative claims data for clinical detail. These data are subject to data-coding limitations and data entry error. Medication use was based on filled prescriptions. Patients were assumed to take the medications as prescribed, but whether patients actually took the medications cannot be confirmed. The intended indication for the traditional migraine-preventive medications prescribed cannot be established, whether they were intended for migraine, another condition, or both, and whether they were used acutely or on a regular schedule, such as daily use. Patients with discontinuation due to a change in insurance drug coverage of erenumab may confound the assessment of adherence. Outpatient pharmacy claims do not include the diagnosis or prescribing physician information; therefore, the prescribing physician was approximated based on proximal ICD-10-CM or HCPCS medical claims. Because this study aimed to quantify the changes in utilization and direct cost associated with acute and traditional migraine-preventive medications when initiating erenumab, the associated cost of anti-CGRP pathway mAbs was not included. Therefore, reductions in cost reported here are specifically within the context of the medication categories reported and do not represent cost decreases related to overall migraine medication use. Finally, there may be systematic differences between the adherent and non-adherent cohorts that account for differences found in healthcare costs and utilization.