Overall, I agree with your train of thought - that targeting non-HDL-C is a better path than TGs - but there's a small problem IRL application - a normal lipid panel does not directly measure LDL-C, only TC, HDL-C and TGs, so there is no way for a PCP to distinguish non-HDL-C from LDL-C, both are wrapped up within the TC number.
However, I see no reason why AMRN could not alter their post R-IT sNDA to change the lipid target from TGs to non-HDL-C, as data from the trial will provide the numbers they need to do so.