Trial
PREDIMED
Research Observatory
Evidence Update
Current evidence
PREDIMED compared Mediterranean dietary patterns supplemented with extra-virgin olive oil or nuts against control dietary advice in 7,447 adults at high cardiovascular risk. The 2018 corrected replacement, not the retracted 2013 report, governs current numerical findings. PREDIMED's withdrawn 2013 report must not serve as the current efficacy source; the 2018 replacement is a reanalysis of the same experiment after randomization deviations. PREDIMED's replacement analysis found lower major cardiovascular events with supplemented Mediterranean dietary patterns, with randomization deviations remaining material.
What changed
PREDIMED's replacement is the same experiment after allocation departures; its adjustment and sensitivity analysis do not erase those departures.
What did not change
The 2013 and 2018 articles cannot be counted as independent replication.
Why it matters
Preserves a material benefit alongside its corrected provenance and design limitations.
Main limitation
Replacement does not erase deviations or create independent replication. Original retained only as historical provenance. Adjusted baseline/propensity methods and sensitivity exclusions support bounded interpretation, not an unqualified pristine-RCT claim or a causal single-food effect.
Next proof
Independent well-randomized replication of the dietary-pattern comparison, with transparent assignment and clinical cardiovascular outcomes.
Summary
PREDIMED compared Mediterranean dietary patterns supplemented with extra-virgin olive oil or nuts against control dietary advice in 7,447 adults at high cardiovascular risk. The 2018 corrected replacement, not the retracted 2013 report, governs current numerical findings. PREDIMED's withdrawn 2013 report must not serve as the current efficacy source; the 2018 replacement is a reanalysis of the same experiment after randomization deviations. PREDIMED's replacement analysis found lower major cardiovascular events with supplemented Mediterranean dietary patterns, with randomization deviations remaining material.
Intelligence Take
Preserves a material benefit alongside its corrected provenance and design limitations.
Limitations
Replacement does not erase deviations or create independent replication. Original retained only as historical provenance. Adjusted baseline/propensity methods and sensitivity exclusions support bounded interpretation, not an unqualified pristine-RCT claim or a causal single-food effect.