Pretty much because previous people decades ago decided that LCM oscillators don't really count. For example, Bill Gosper alluded to the existence of an LCM-based p24 oscillator in September 1994, a month before he discovered 186P24, referring to the latter as the "first nonboring p24". Dean Hickerson referred to it as the "first nontrivial p24", evidently using a different definition of "trivial" and thus implicitly excluding the boring p24. I don't really agree with this definition (particularly due to the existence of borderline cases like 87P26) but I can see where he's coming from, considering that these oscillators are easy to manually construct and ultimately rather useless when it comes to interacting with other objects.
Regardless, this definition has stuck when it comes to the first-discovered oscillators of a given period on the wiki, though weirdly not for the smallest known oscillators. I've tried using footnotes on the respective wiki pages to help resolve the confusion between these two definitions of triviality.