Predictive validity of NEDA in the 16-and 21-year follow-up from the pivotal trial of interferon beta-1b

Background: Long-term follow-up from the randomized trial of interferon beta-1b (IFNB-1b) permitted the assessment of different definitions of no evidence of disease activity (NEDA) for predicting long-term outcome in multiple sclerosis (MS). Objective: To examine the predictive validity of different NEDA definitions. Methods: Predictive validity for negative disability outcomes (NDOs) at 16 years and survival at 21 years post-randomization were assessed. NEDA in the first 2 years was defined as follows: clinical NEDA: no relapses or Expanded Disability Status Scale (EDSS) progression from baseline to Year 2; NEDA-3a: no relapses, no confirmed > 1-point EDSS progression, and no new T2-active lesions; NEDA-3b: no relapses, no EDSS progression, and no increase in T2 burden of disease (T2-BOD); and NEDA-4: no relapses, no EDSS progression, and no increase in T2-BOD or atrophy. NDOs were defined as death, need for wheelchair, EDSS > 6, or progressive MS. Results: A total of 245 and 371 patients were evaluated at 16 and 21 years, respectively. Clinical NEDA predicted NDOs (p = 0.0029), as did baseline EDSS (p < 0.0001), baseline T2-BOD (p < 0.0001), and change in T2-BOD (p = 0.0033). IFNB-1b treatment (p = 0.0251), relapse rate in the 2 years before study start (p = 0.0260), T2-BOD at baseline (p = 0.0014), and change in T2-BOD (p = 0.0129) predicted survival at 21 years. Conclusion: Clinical NEDA predicted long-term disability outcome. By contrast, definitions of NEDA that included on-therapy changes in magnetic resonance imaging variables did not increase the predictive validity.