Collecting this in one place because it comes up every few weeks and the answer is always assembled from scratch. It is about nausea, and it is deliberately narrow — everything I am not confident about is marked as such.
What is actually established
The line between titrate-through and stop is not severity, it is trajectory and what else is present. Nausea that peaks and improves within a week is the expected pattern. Nausea that is escalating, or that comes with severe upper-abdominal pain radiating to the back, or that prevents fluids for more than a day, is a different conversation and belongs with a clinician the same day.
The condition it depends on
The meal advice is right and incomplete without the hydration point. People stop drinking because drinking makes them feel full, then attribute dehydration symptoms to the drug.
The practical version
Trial-level incidence runs roughly 20 to 25% for nausea at the higher dose tiers and 12 to 17% for diarrhoea, with most events mild to moderate and concentrated in the weeks after each escalation.
What I am not sure about
The question I want answered is what distinguishes the nausea you can titrate through from the nausea that means stop. I have searched first, so if this is covered somewhere point me at it and I will read it.
— Dr.SurgeonPGH · corrections welcome and will be edited into this post with credit