I'm not sure how they can no that without restarting extensive trials, but as a sibling commenter points out, viruses mutate constantly so I guess dealing with variants is pretty common when it comes to vaccination.
EDIT: Actually, sounds like a couple of the mutations are in the spike protein and there is some evidence of reduced antibody effectiveness against the mutated version.
The deletion seems to reduce antibody neutralization, but:
- In the preprint where this was shown, only 4 convalescent sera were tested;
- The same 4 sera had large variation in neutralization activity per se;
- There is no investigation on potential impaired T cell reactivity (cellular immunity): FTR, the "mink mutation", although it exhibited slightly lower antibody neutralization, did not change the reaction of T cells to it.