If you can think of a plausible third variable that explains the association, be cautious. If the study controlled for known confounders, that strengthens the claim.