Ignoring intonation and punctuation for a moment, there's two structures in French that can produce a surface form like "Ça énerve d'attendre":
The first is an impersonal construction, where the logical subject of a sentence appears in object position and a dummy pronoun (il or ça) fills the object position:
La neige tombe - Il tombe de la neige
Les touristes arrivaient alors par dizaine - Il arrivait alors des touristes par dizaines
Attendre est énervant - Il est / c'est énervant d'attendre
Two things to note is that the subject pronoun is always neutral and triggers masculine singular agreement and that marking of the erstwhile subject with de is obligatory.
The second is right dislocation, where an argument is moved outside of the sentence (in this case, at the end) and reflected in its core by a pronoun:
La neige tombe - Elle tombe, la neige
Les touristes arrivaient alors par dizaine - Ils arrivaient alors par dizaines, les touristes
Attendre est énervant - C'est énervant, d'attendre
Attendre énerve - Ça énerve, d'attendre
If we contrast right dislocation and the impersonal construction, two major differences appear:
Dislocated subject noun phrases are unmarked like normal subjects while the impersonal construction has to be marked with de. Yet right dislocated infinitive phrases have to be marked with de.
The resumptive pronouns of a dislocated element conserve their agreement features: gender in the first two sentences and number in the second. In the case of infinitives, they have to be echoed by ça.
Both the impersonal construction and left dislocation serve the same purpose: put the focus on the verb phrase by moving it to the start of the sentence, as Luke Sawczak noted in a comment.
Which construction is used is largely a question of register, with the impersonal one being privileged in higher and more formal registers and dislocation dominating everyday colloquial speech.
Yet it's interesting to note that both constructions can produce the exact same form when the displaced subject is an infinitive:
Te voir lui a fait plaisir - Ça lui a fait plaisir de te voir - Ça lui a fait plaisir, de te voir
The main differences between the impersonal construct and left dislocation of an infinitive are the pronoun used (either il or ça in the first case but only ça in the second) and the specific prosodic pattern of the dislocated element, marked in writing by a comma (often said to be a pause, but this is not supported by the literature, that describes it as a flat intonational contour of the right dislocated element instead). However, this prosodic signature can be absent and the dislocated element fully integrated to the rest of the sentence (see references at the bottom of the answer).
When digging further however, the constructions do not necessarily coincide. It's the case of the sentence that prompted the question: A more formal counterpart of "Ça énerve d'attendre" would be "Il m'énerve d'attendre" but this is not a well formed sentence.
So if "Ça énerve d'attendre" has to be the dislocated counterpart of "Attendre m'énerve", why does the comma intonation seems not only unnecessary to the question writer and to many commenters, but outright wrong?
A possibility is that in spoken language both constructions have been conflated, giving rise to a new construction with the syntax of dislocation (ça as subject pronoun) and the prosody of the impersonal construction. But this will have to remain idle speculation (at least until I find something more conclusive in the literature).
BUTHKE Carolin et al. "Les sujets nominaux du français et leur reprise pronominale: leur reprise pronominale: dislocation ou doublement ?" (conference paper) http://prosodia.upf.edu/membres/sichel_bazin/presentations/Buthke_et_al_PFC_NO_10.pdf
AVANZI Matthieu, "L'interface prosodie/syntaxe en français : Dislocations, incises et asyndètes" http://www.theses.fr/2011PA100065
