This multicenter, retrospective, longitudinal chart review study used data extracted from EMRs of eligible patients treated at five clinical sites in the United States: Los Angeles Headache Center—Los Angeles (Los Angeles, CA, USA), Los Angeles Headache Center—San Diego (San Diego, CA, USA), DENT Neurologic Institute (Buffalo, NY, USA), Diamond Headache Clinic (Chicago, IL, USA), and Neuroscience Group (Neenah, WI, USA), and had initiated the combination between November 1, 2021 and November 30, 2022. The index date was defined as the date of the onabotulinumtoxinA treatment visit at which atogepant was prescribed. A baseline period of 1–3 months prior to index was used to assess the effectiveness of onabotulinumtoxinA monotherapy treatment. Charts were reviewed up to 8 months prior to the index date to confirm inclusion criteria. Patients were followed for up to ~ 6 months post-index (corresponding to up to two onabotulinumtoxinA treatments; Fig. 1).

Eligible patients were identified based on a physician’s diagnosis of CM in the patient chart and met the following criteria: ≥ 18 years of age, received ≥ 2 consecutive treatments with onabotulinumtoxinA prior to starting atogepant treatment, and received combination treatment for ≥ 1 month with onabotulinumtoxinA and atogepant (one cycle), initiated between November 1, 2021, and November 30, 2022. Patients were excluded if the time interval between onabotulinumtoxinA treatment and atogepant prescription was more than 4 weeks or if they received a CGRP treatment (monoclonal antibody or gepant) indicated for preventive use in migraine throughout the period between the last onabotulinumtoxinA treatment prior to the index date and the index date (~ 3 months), and if there were no migraine-specific treatments post index date. No a priori power or sample size calculations were performed as this was a descriptive study.

The site Principal Investigator (or their designee) identified potentially eligible patient charts at each site. Data were entered by staff under their supervision into an electronic case report form (eCRF) according to the protocol and study materials. The eCRF was made available in the electronic data capture (EDC) system Dacima (Montréal, QC, Canada). Dacima Clinical Suite is a secure, validated EDC software, and is fully compliant with FDA 21 CFR Part 11 requirements. The eCRF was developed to help optimize the collection of clean and homogeneous data by allowing for built-in data checks. An ‘Unknown’ option was made available in the EDC for variables where data may have been missing from the patient chart. This allowed for more complete data to be collected in the EDC by avoiding the option for variables to be left blank. Free-text entries were limited as much as possible in the eCRF to enhance data quality. Each site was provided with access to the secure online database in order to enter the data.

Baseline demographics, migraine-relevant clinical history, concomitant migraine medication(s) use, and duration of onabotulinumtoxinA treatment prior to starting atogepant treatment were collected. Treatment patterns were measured by onabotulinumtoxinA treatment dose and duration between treatment doses, the dose of atogepant prescribed, and any change of type, dose, or regimen.

Outcome measures to evaluate the potential benefits of combination treatment with onabotulinumtoxinA and atogepant were assessed at ~ 3 and 6 months post-index. Outcomes assessing treatment effectiveness included the mean change from baseline in monthly headache days (MHDs) at ~ 3 months and ~ 6 months and the proportion of participants with any, ≥ 25%, ≥ 50%, ≥ 75%, or 100% reduction from baseline in MHDs at ~ 3 and 6 months post-index.

Safety was evaluated by the frequency, types, and severity of adverse events (AEs) assessed from the patient charts. Treatment discontinuation was also evaluated by patient chart and categorized as discontinuation of onabotulinumtoxinA treatment, discontinuation of atogepant treatment, or concurrent discontinuation of both treatments and the reported reason for discontinuation when available.

Statistical analyses were descriptive in nature and were conducted using SAS v9.04 and R version 4.1.3. For all analyses, continuous variables were summarized by mean, standard deviation (SD), median, range (i.e., minimum–maximum), and number and percentage of patients with unknown values. Categorical variables were summarized by number and percentage of patients in each possible category and those with unknown values. Additionally, the total number of patients at each visit and/or the total number of patients during follow-up were presented. Effectiveness outcomes and adverse events were described at the first (~ 3-month) and second (~ 6-month) post-index onabotulinumtoxinA treatment visits. Discontinuations were described cumulatively for the follow-up period for which the patient was receiving combination treatment.

Headache frequency recorded on a 90-day basis was converted to a 30-day basis at each follow-up interval. Distributions of the headache frequency converted from a 90-day to a 30-day basis were compared to those that were captured on a 30-day basis to ensure there were no statistically significant differences in distribution. If data were collected for a post-index visit dating more than 7.5 months post-index, this was considered an additional visit and not included in the analysis. Descriptive statistics related to headache frequency and mean change from baseline in headache day frequency are presented as days per month and were calculated for each follow-up visit.

This real-world retrospective study of the addition of atogepant to onabotulinumtoxinA treatment for CM was conducted on secondary data collected for the purpose of health care, not research. Missing values may be the result of several circumstances including data not being applicable for a specific patient, or data not being available in the medical chart. Not all data fields of study interest were available for each patient at a particular time point. Additionally, not all study patients contributed data to both post-index onabotulinumtoxinA administrations of study interest due to loss to follow-up or discontinuation. Missing data were captured explicitly by allowing for an unknown option in the eCRF wherever appropriate. The number and percentage of unknown values for a variable were reported. No imputation was performed, and all analyses were conducted on available data only, with the exception of date variables where day and/or month were missing, for which the first day of the month and/or the first month of the year were imputed, respectively. In case the sample size for exploratory variables was not sufficient, then the analyses were not reported.

The New England Independent Review Board reviewed the study protocol prior to study initiation and determined the study as exempt from review. Written informed consent from participants was not required for this study in accordance with the national legislation and the institutional requirements as only deidentified data were collected. The study was conducted in accordance with the International Conference for Harmonisation guidelines, applicable regulations, and the Declaration of Helsinki.