The review did not assign one certainty label to every cannabinoid finding. Remission, inflammatory markers and adverse-event outcomes had separate assessments based on the evidence supporting each one.
GRADE followed the endpoint
In the capsule comparison, the 2018 review rated remission and clinical-response evidence low certainty. It rated the reported C-reactive-protein and quality-of-life comparisons moderate certainty, while serious-event and adverse-event-withdrawal evidence was low certainty. Sparse data were a recurring reason for downgrading.
These judgments concerned confidence in an effect estimate, not whether the study drug was desirable or whether a particular result reached statistical significance.
Certainty and benefit are different dimensions
Moderate-certainty evidence can concern an uncertain or small difference. Low-certainty evidence can have a favorable point estimate without establishing a reliable treatment benefit. A publication’s randomized design alone does not make every outcome high certainty, especially when events or participants are few.
Use the label with its supporting comparison
A claim about remission should use the remission assessment rather than borrowing a stronger label from an inflammatory-marker endpoint. The review was a historical synthesis of two small trials, not an updated census of present-day research. Its graded results support precise endpoint-specific language rather than a blanket statement that cannabis is either proven or disproven for all colitis outcomes.
Related reading: why endpoint definitions matter.
Research cited
Kafil and colleagues (2018). Cannabis for the treatment of ulcerative colitis. DOI: 10.1002/14651858.CD012954.pub2.
Leave a Reply