There are impacts to testing, but, not to the test in one sense. Load is load, one agent or many will provide the same input (geography aside). Setting up multiple agents may not provide benefit, or, it may be required. It might allow for additional options, or remove them. Then of course, there is the setup and maintenance of the agents, and the cost/benefit of it.
For large volume tests, you'll need multiples.
As mentioned, you might get geographic benefit.
It makes it ultra easy to work with load balancers, with sticky IP policies.
Setup, maintenance and install.
Potential firewall issues. Connectivity.
At conclusion, extra time to collate results.