I've been told that the first syllable of abiciō is long by position, because it's actually an underlying *abjiciō, which causes it to be syllabified as *ab-ji-ci-ō before the *ji simplifies to i. So the first syllable has a coda consonant, despite the next syllable being headless.
That's different from the account that I've heard. According to Alex B.'s answer here, which cites Allen 1978 and Leumann 1977, the spelling "abicio" was used along with heavy scansion of the first syllable in Classical Latin because of orthographic, not phonological, simplification of double letter
I. The pronunciation in this case would start with /ab.ji/, a VC syllable followed by a CV syllable. A page away, Allen discusses a possibly related orthographic phenomenon, the avoidance of double letter
I as a representation of phonologically geminate intervocalic /jj/.
When the /j/ was lost (which seems to have been possible in some time periods), the consonant before the /i/ actually would have been resyllabified into the onset of the following syllable. This normally would create a pronunciation with a light initial syllable (/a.bi.ki.oː/) but some speakers may have used a pronunciation with a lengthened vowel /aː.bi.ki.oː/ for the sake of scansion when reading poetry, since Aulus Gellius complains about hearing speakers using long vowels [oː] and [uː] in the first syllables of words like obiciebat and subicit (Attic Nights 4.17.1).
I don't know of any examples apart from the compounds of jaciō.