According to Mallory we have the following PIE words (in this notation, g = palatal/plain ġ=plain/uvular):
a̯enghu̯is snake (> уж)
a̯enĝhus narrow (> узко)
a̯enĝhnos fear, constriction (> ужас)
This hints at that the words could be related if the exact value of the g was determined wrongly.
I also wonder whether the following words are related:
e̯oğhis snake, serpent (in Fortson, e̯eĝhu̯is). Mallory claims this is unrelated to the a̯enghu̯is. Fortson claims it is related to the following entry and reconstructs the same consonant in the root. Mallory claims that the "traditional school" derived hedgehog as "snake-eater". The Brill etymological dictionary of Latin reconstructs the initial a̯- here and links this root to the a̯nğhis, claiming that the -n- infix appeared in it due to analogy.
e̯eĝhis hedgehog (> russ. ёж, eng. hedgehog)
e̯eĝherom lake (> russ. озеро)
If e̯oghu̯is, it could be connected to e̯eghu̯ti drinks