A dose map needs
more than a dose.
Trials at similar labeled doses can reach different answers because they test different products, people, comparators and outcomes.
The 4 g example
REDUCE-IT tested 4 g/day of prescription icosapent ethyl and found fewer primary cardiovascular events in selected high-risk, statin-treated patients. STRENGTH tested 4 g/day of an EPA+DHA carboxylic-acid product and did not find a significant primary-event reduction. Both reported an atrial-fibrillation signal.
These studies were not a direct head-to-head comparison. The result cannot prove that DHA prevents EPA from working, or that all EPA products have the same effects.
A point estimate is only part of the picture.
An estimate below a risk ratio of 1 can still have a confidence interval that includes no effect. Different trials also use different composite endpoints. Comparing isolated dots without intervals and study context can be misleading.
Explore seven actual study doses ↗