Being a small biotech is a high risk situation. The resources aren’t really there to do full scale drug development from scratch for a major disease, so small biotechs are a mish-mash of risky bets on existing molecules, creative tweaks such as repurposing or bets on the relatively safe route of a minor tweak to an existing drug class.
Trevana was trying to get established with this last strategy. The company made what seemed like a safe bet on an already approved mechanism of action for opioid withdrawal. Their drug, oliceridine, was designed to be a more selective version of opiod withdrawal drugs, meaning it was designed to be active more narrowly against the desired drug targets. Their assumption was that eliminated off-targets were responsible for adverse events, notably respiratory depression. They hoped to gain an approval via the greater safety route by eliminating this off target activity.
Trevana might have done better by projecting their more selective profile to the tissues in which those remaining targets were expressed. They likely did not because it's not that easy to do. Well, until now.
Trevana may have been hoping to massage the data, an indication that their MOA did not blow away the prior standard for reducing respiratory depression. The FDA did not agree with their methods of defining respiratory depression. Better to have a real pharmacologic advance, on which to base your approval.