OBJECTIVES This study aims to report treatment patterns of relapsing-remitting multiple sclerosis (RRMS) in the Brazilian Public Health System (SUS).

METHODS Retrospective analysis of the frequency of RRMS treatment sequence, number of RRMS treatment switching per patient and the time between switches was developed according to Ambulatory Information System (SIA/SUS) database, from October/2011 to July/2018. A drug survival analysis was performed to compare the time to first treatment switch or interruption. The Cox proportional-hazards model was used and the Kaplan-Meier survival curves and their hazard ratios were estimated. Only patients who started with first line therapy, according to Brazilian guideline, were considered.

RESULTS In the period, 633 different RRMS treatment sequences were found including interferon-β (IFN-β) 1a or 1b, glatiramer acetate (GLA), fingolimod and natalizumab. Teriflunomide and dimethyl fumarate were not available in the database at the data cutoff. IFN-β 1a 30 mcg-GLA was the most common (6.34%) treatment sequence, followed by switch between IFN-β 1a 22 mcg and IFN-β 1a 44 mcg (5.83%), GLA-natalizumab (5.56%) and GLA-fingolimod (5.45%). The number of switches per patients showed that 26.1% switched only one time, 9.3% had two switches, 3.3% three switches, 2.3% four switches or more and the others 59.1% did not switch at any time . On average, patients switched 1.6 times (excluding patients without switches). The mean time to the second treatment was 24.7 months (SD: 19.7). The Kaplan Meier survival curve of platform therapies showed that all drugs had a similar risk of switch or interruption, except for IFN-β 1a 22, which showed an statistical significant higher risk in all comparisons.

CONCLUSIONS Data from Brazilian Public Health System showed the switch profile of RRMS treatment considering IFN-β 1a or 1b, GLA, fingolimod and natalizumab. Initial therapy survival was similar for all drugs, except for IFN-β 1a 22.