It really depends on the environment. If the scanner isn’t able to determine the targets hostname (e.g. via a Reverse-DNS lookup or other means) and the target is serving a different web-page when using the IP vs. when accessing it via the hostname you could miss various vulnerabilities in this case.
This should be already explained in the previous linked documentation:
In environments with virtual hosts, the scan reports will have less results because duplicates are avoided.
Some more technical details around this topic are available in the following discussion thread: