Someone determined enough to serve you ads through a DNS sinkhole could do something like this: the page injects the ads using js and waits for a pingback that is triggered after the ad is loaded. On a timer, if the ping was not received, the page can inject another ad with hard-coded ip address, instead of DNS.
The reason why this does not happen is that companies understand that someone who hates ads so much they are willing to deploy an ad blocker, a DNS sinkhole, etc will probably not click on an ad either way.
Also, note that pfblocker-ng contains lists of millions of (automatically updated) IPs known to serve Ads (or malicious content), redirecting those, too. So unless you are willing to switch precious IPs for your hardcoded IP-Ad-Serving Servers, injection through Javascript won't work either.
The only actual direct consequence to clients I have noticed is that if you directly want to resolve Blocked-Ad-Links, it won't work - e.g., Google usually displays some "Ad"-Links first as results - clicking these won't get you anywhere. For my private LAN users, I think this is a good lecture to have - realizing that you're bullied into clicking valid-looking links that are instead Ads, from people that pay others to distract you from what you were actually looking for, or distort the natural ranking by paying for being ranked first.
I have a printout at our house for visitors who complain about this: They can manually set their DNS to 8.8.8.8 and my router won't interfere (it could, but I found spoofing direct requests to Google's DNS a bit too much).
