Treatment transitions in neuromyelitis optica spectrum disorder increase risk for disease advancement

•Treatment transitions are common in people with NMOSD.•Each non-medical/tolerability switch after the first, increased risk of harm.•The same risk was not observed for each additional switch due to a medical reason.•Transition time due to non-medical/tolerability switches was markedly longer.•Prior treatment experience impacts the duration of the treatment transition period. People with neuromyelitis optica spectrum disorder (PwNMOSD) commonly switch between disease modifying therapies, yet the consequence of transitions remains unknown. We aimed to understand if treatment transitions due to medical, non-medical, and tolerability reasons were related to disease progression. A retrospective study of medical records for PwNMOSD was performed between 2008 and 2022. A comprehensive clinical timeline was created for each person including details related to treatment history and associated clinical and radiological outcomes (i.e., hospital admission, relapses, and MRI advancement). If a transition occurred, the reason for the switch was categorized as being due to medical, non-medical, or tolerability issues. A proportional hazards model was created, and the assumptions were tested based on weighted residuals. The cohort included 164 aquaporin-4 IgG positive NMOSD subjects with 89 (79 female; median disease duration (range) = 10.1 years (y) (1.7–32.8)) people switching therapies at least once (once: 42; twice: 26; three times: 12; four times: 6; 5 or more times: 3). A similar amount of higher efficacy therapies was used by PwNMOSD that switched due to a non-medical/tolerability or a medical-related reason. The results of the recurrent event survival analysis revealed that after an initial transition due to non-medical/tolerability reasons, the risk of a hospital admission, relapse, and MRI advancement decreases by 40.3 % (p = 0.005), 53.1 % (p = 0.002), and 65.9 % (p = 0.005), respectively. However, with each additional discontinuation due to non-medical/tolerability reasons, the risk of hospitalization increased by 25.2 % (p = 0.0003) and risk for MRI advancement increased by 41.9 % (p = 0.03). For transitions due to medical reasons, a significant increased risk of MRI advancement by 32.2 % (p = 0.005) for the first switch was identified with no associated observed risk with each additional discontinuation (p = 0.33). Within the first six months after stopping a medication due to non-medical/tolerability reasons, the rate of starting a new medication was less