EMA two-step rule for 'disease-modifying'

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

The EMA's Efficacy and later CNS Working Party and its drug committee (CHMP), with comments from industry groups such as Lundbeck, EFPIA and Biogen/Vernalis.

What happened

The guideline on clinical investigation of medicines in Parkinson's was adopted by CHMP on 24 July 2008 (effective 1 February 2009; first drafted for consultation on 19 July 2007) and updated on 21 June 2012 (effective 2 July 2012; reference EMA/CHMP/330418/2012 rev. 2). EMA is a public regulator, so there is no funder or deal to record. For disease modification it asks for "a two-step procedure": first show a delay in clinical progression, for example change in UPDRS or time to needing levodopa; then show an effect on the underlying disease process. It recommends randomised, placebo-controlled, long trials ("probably up to 24 months" in early untreated disease) and prefers no single design. Imaging markers of dopamine uptake "cannot be considered sufficient surrogate biomarkers."

In the 2008 response to comments, EMA declined to add specific guidance on delayed-start designs (the approach used in ADAGIO, landmark 62), saying the value of one specific design over another "remains to be established", and stated there was then "no accepted biomarker showing an effect on the underlying pathophysiological process."

A concept paper (EMA/322471/2025), agreed by EMA's Central Nervous System Working Party on 25 September 2025 and adopted by CHMP for consultation on 4 November 2025, proposes replacing the guideline. Reasons: biomarker-based enrolment and enrichment (citing the 2024 biological definitions), guidance for drugs aimed at the underlying disease on top of symptomatic treatment, digital endpoints and decentralised trials. It would also widen the scope to multiple system atrophy, corticobasal degeneration, progressive supranuclear palsy and essential tremor. Consultation ran 30 November 2025 to 31 March 2026. The paper plans a draft for consultation "not later than Q4 2026"; EMA's guideline page on 19 Oct 2026 listed no draft.

Why it moved us toward a cure

It defines what evidence will count, and so shapes every trial design, from DAT-scan enrichment to the seed assay (landmark 9). The planned revision is where biological definitions (landmark 28) could first gain regulatory force.

Outcome

Open. The 2012 text still stands; the revision is at concept-paper stage, with no draft published as of 19 Oct 2026. The two-step requirement is demanding, which is part of why slowing has never been proven (see landmark 61).

Sources

• EMA guideline (EMA/CHMP/330418/2012 rev. 2)
• EMA overview of comments, 30 Jul 2008
• EMA concept paper on revision, EMA/322471/2025, 4 Nov 2025
• EMA guideline page, with version history (read 19 Oct 2026)
• Not yet verified: any US FDA guidance of this kind (none found; the FDA's 2018 draft was for early Alzheimer's); whether the draft revision appears by the end of 2026.