Objective: This article describes the design of and difficulties inherent in the execution of a long-term, observational trial that sought to assess the validity of short-term measures of multiple sclerosis (MS) (eg, relapse rate, inflammatory lesions) for long-term disease outcomes.
Methods: In the original double-blind, placebo-controlled interferon (IFN)-p1b study, 372 patients with relapsing-remitting MS (Expanded Disability Status Scale score 0.0-5.5) were randomly assigned to IFN-beta1b 50 ug (n = 125), IFN-beta1b 250 microg (n = 124), or placebo (n = 123) for 2 years. These patients were recruited 16 years later for participation in this long-term follow-up (LTF) study, which had no exclusion criteria or drug interventions.
Results: The 11 centers identified 88.2% (328/372) of the original study patients at LTF; however, 10.8% (n = 40) refused to participate and 9.4% (n = 35) were deceased. Detailed evaluations were available for 260 patients, which included 7 deceased patients. No differences in demographic or baseline disease characteristics were found between individuals who did and did not participate in the LTF. More patients randomly assigned to placebo in the original trial were deceased (20/123 [16.3%]) than those assigned to IFN-beta1b 50 microg (9/125 [7.2%]; uncorrected P = 0.044) or IFN-beta1b 250 microg (6/124 [4.8%]; uncorrected P = 0.003).
Conclusions: Although most patients (88.2%) were identified at LTF, ascertainment was incomplete. This was attributable to patients' refusal to participate, loss to follow-up, or death. Delays in the registration of death data and recent privacy legislation provided further barriers. Mortality was lower for patients originally randomized to receive IFN-beta1b rather than placebo. We recommend that all short-term trials on chronic diseases include provisions for LTF.