ADAGIO: rasagiline

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

C. Warren Olanow, Olivier Rascol and colleagues in the ADAGIO study group; sponsored by Teva, which sells rasagiline as Azilect. Published in the New England Journal of Medicine.

What happened

1,176 people with untreated, very early Parkinson's in 14 countries were split into "early-start" groups (rasagiline 1 mg or 2 mg a day for 72 weeks) and "delayed-start" groups (placebo for 36 weeks, then the drug for 36 weeks). The logic: a drug that only eases symptoms would let the late starters catch up; a drug that slows the disease would leave the early starters permanently ahead. To count as positive, a dose had to pass three tests on the UPDRS rating scale (NCT00256204). Rasagiline 1 mg passed all three; for example, at week 72 the early group had worsened by 2.82 points against 4.52 for delayed-start (P = 0.02). The 2 mg dose failed: 3.47 points early-start against 3.11 delayed-start (P = 0.60), so the early group was not ahead. The authors urged caution because the two doses disagreed.

A later imaging trial (University of Florida, NCT02789020). A Phase 2 placebo-controlled study gave rasagiline 1 mg a day or placebo for 12 months to people with early Parkinson's. 96 were enrolled; the registry shows it completed on 11 May 2023 with results posted (a later date than the paper's 2021 online publication, unexplained). The main measure was a brain-scan marker, free water in the substantia nigra. The paper (Arpin et al., Movement Disorders 2022;37:325-333) found no difference between groups (p=0.346 and p=0.228 on the two scan settings, with 90 and 75 scans analysed) and no difference in any secondary clinical measure. The authors found no evidence that 1 mg a day is disease-modifying over 1 year. The scans did show free water rising over the year in the whole group on one setting (p=0.025).

Why it moved us toward a cure

The later Florida imaging trial used free-water MRI as its measure (landmark 224). It followed the first delayed-start analysis of rasagiline in TEMPO (landmark 162). It put the delayed-start design, a way to tell protection from symptom relief that DATATOP lacked, into a large trial. Its muddled result also showed the design's weakness: one dose positive and one not leaves the answer open.

Outcome

Dead end for a slowing claim. On 17 October 2011 the FDA advisory committee voted 17-0 that ADAGIO did not convincingly show the 1 mg dose met its prespecified goals, and 17-0 that Teva had not shown substantial evidence that Azilect slows progression. Azilect, approved in the US in May 2006, remained approved only for symptoms. No drug has since been approved to slow Parkinson's (README status).

Sources

• Olanow et al., N Engl J Med 2009;361:1268–1278, DOI 10.1056/NEJMoa0809335 (abstract via Lehigh Valley repository; PubMed 19776408)
• PharmaTimes, 28 Sep 2009: 14 countries, Teva's filing plans · Teva, FDA approval of Azilect, 17 May 2006
• Michael J. Fox Foundation on the FDA panel, 18 Oct 2011 · Medthority
• Arpin et al., "Diffusion magnetic resonance imaging detects progression in Parkinson's disease: a placebo-controlled trial of rasagiline," Mov Disord 2022;37:325-333, DOI 10.1002/mds.28838, PMID 34724257 · ClinicalTrials.gov NCT02789020 · ADAGIO, NCT00256204
• Not yet verified: how the 96 enrolled map onto the 90 and 75 scans analysed; why the Florida registry completion date (2023) is later than the paper; whether the FDA formally issued a final refusal after the panel; the 14-country count (PharmaTimes only; the registry lists no sites).