Response to “Early- and later-stage persistence with antiobesity medications: a retrospective cohort study”

TO THE EDITOR: We have read and appreciate the hypothesis and the data collected and analyzed to report on patient persistence with antiobesity medications (AOMs) using electronic health records [(1)]. Although we agree that this study explores an interesting topic, we are concerned that the study methodology may have inadvertently introduced bias, thereby impacting the conclusions.

We have concerns with the classification of patients by the last AOM prescribed and the categorization of patients switched from brand-name AOMs to non-Food and Drug Administration (FDA)-approved generic combinations. We recognize that classification of patients to the last AOM prescribed among patients who switched AOMs during the first year of the study period may be appropriate when looking strictly at overall persistence, irrespective of AOMs (primary aim); however, this artificially inflates the observed persistence with novel AOMs (e.g., semaglutide). Furthermore, the classification of patients to FDA-approved, branded, fixed-dose combination AOMs (e.g., phentermine-topiramate, naltrexone-bupropion) who were switched to generic combinations during the study period is concerning because the safety and efficacy of the combined generic components have not been demonstrated and cannot be compared to or combined with that of the FDA-approved treatment.

Another concern is that the entirety of the patient journey (patients often undergo ≥1 obesity treatment) was not considered. The unbalanced exclusion of patients (e.g., patients undergoing bariatric surgery, patients with diabetes on injectable forms of semaglutide and liraglutide) is of concern because clinical evidence has reliably demonstrated that AOM effectiveness for weight loss among patients with diabetes is generally less than that for patients without diabetes [(2, 3)]. This study reports that patients with diabetes were less persistent with AOM therapy; however, excluding patients with diabetes using semaglutide and liraglutide but not excluding patients using the other AOMs may have falsely inflated the persistence and weight loss achieved from semaglutide and liraglutide. Additionally, many issues of persistence are not reflective of treatment response or patient participation but, instead, are affected by coverage and availability [(4)].

Finally, the study failed to account for the vastly different drug titration schedules of the AOMs (e.g., 2 weeks for phentermine-topiramate, 4 weeks for naltrexone-bupropion, 16 weeks for semaglutide) [(5-7)]. These differences in titration schedule are important to consider because the variations observed in persistence may be due, in part, to the time to reach the respective maintenance dose of each AOM.

We acknowledge that all studies have limitations that cannot be overcome; however, careful consideration must be taken when contextualizing those implications. Manuscripts such as the one in question by Gasoyan and colleagues inform payer and provider decisions and therefore may lead to further restriction of patient options for the treatment of obesity. Objective-based, scientifically driven methodology is therefore critical in the design of future studies to accurately inform patients, caregivers, and payers on the treatment options in obesity.

This work was sponsored by Currax Pharmaceuticals, LLC.

Angela Fitch reports work on advisory boards for Novo Nordisk A/S, Gelesis, Vivus, Jenny Craig, FoundHealth, Ms. Medicine, and Eli Lilly and Company. Amber Huett-Garcia declared no conflict of interest.