It is most probably either a regular comparative or one **built by analogy to other regular comparatives**.

Not sure if it is possible to prove or disprove the existence of and adj. _*anterus_ to which *anterior* would be the comparative, in the same way that *posterior* is of *[posterus][1]*. *Disclaimer: this is out of my comfort zone, and has to be taken as a not-very-informed hypothesis.*

*Posterus* is the adj., *postea* the adverb. _*Anterus_ would be the adj. corresponding to adv. *antea*.

About the existence of _*anterus_ and the hypothesis that *anterior* is its regular comparative, I found two pieces of evidence: [a XIX century lexicon by a Latin Professor at Oxford][2] and [a XV century book][3] actually written in Latin by a *[prominent Flemish humanist][4]* (according to wp).

Now, [L&S][5] says *anterior* is late Latin and cites [Sulpicius Severus][6] using the word as early as the V century. Whether _*anterus_ existed or not (at least in spoken) at that time, I haven't been able to found any evidence. The fact that it doesn't show up in good dictionaries or corpora is a good argument against its popularity. But even if it did not exist, the -r- could have been derived (as an innovation) by analogy with postea > posterus > posterior

As noted by Colin and Hugh, the same analogy works with other adverb/adjective pairs:

* infra/inferus/inferior
* intus,inter//interior
* extra/exterus/exterior