Mr. Lucisano and colleagues sought to examine the long-term effect of DMTs on CDP and irreversible disability milestones (i.e., EDSS scores of 4 and 6) in relapsing-remitting MS. The researchers used marginal structural proportional models, a novel technique that enables them to correct modeling for confounders that vary with time in longitudinal observational studies. Such confounders include treatment switches, on-treatment relapses, and treatment gaps.
The investigators selected patients with 10 or more years’ follow-up and one or more EDSS score evaluations per year from the BMSD pooled cohort. Using marginal structural proportional models, the investigators evaluated cumulative hazards of 3-, 12- and 24-month CDP (i.e., CDP3, CDP12, CDP24) events in 6-month periods. They created stabilized inverse probability of treatment weights (IPTWs) at each 6-month period using survival models according to treatment status (i.e., treated versus untreated). Treatment status was assigned for each patient according to the percentage of time that he or she spent receiving DMT in each 6-month period. A patient who received treatment for 70% or more of the period studied was considered treated; patients who did not meet this threshold were considered untreated. The weights were calculated on the basis of sex, age, occurrence of relapse, EDSS score, and registry source. Finally, the researchers used Cox regression models estimating the effect of DMTs on the risk of reaching CDP3, CDP12, and CDP24, adjusted by the IPTWs, to compare cohorts that remained treated or untreated throughout follow-up.
The investigators identified a cohort of 15,602 patients with relapsing-remitting MS, and this group had 312,040 EDSS score evaluations. Approximately 28% of patients were male. Median age at disease onset was 28.3 years, and median disease duration was 18.7 years. Median follow-up duration was 13.8 years.
During follow-up, 43.3% of patients had CDP3, 27.7% had CDP12, and 14.4% had CDP24 events. In addition, 23.6% of patients reached an EDSS score of 4, and 11.2% reached an EDSS score of 6.
Cox models adjusted by IPTW demonstrated increasing positive evidence of the effect of cumulative treatment exposure, compared with cumulative untreated epochs, according to the length of confirmation time used for defining the CDP. The investigators did not observe an effect of treatment on the probability of reaching CDP3 (hazard ratio [HR], 1.02), but treatment had a protective effect on CDP12 (HR, 0.90) and CDP24 (HR, 0.65) endpoints. During treated epochs, the HR of EDSS 4 was 0.89, and the HR of EDSS 6 was 0.86. Sensitivity analyses largely confirmed the results of the main analysis.
Two of the researchers are employees of Biogen International, which supported the research. Several investigators received compensation or funding from various pharmaceutical companies.