A thing to consider is that ESD (test) waveforms are in the 5-9ns
range, which is pretty low frequency by modern RF standards.
A balun primary meant for 5GHz might look pretty much like
a short (so eliminating a voltage breakdown concern, provided
that the ESD return path is Gnd as is the balun primary return)
at 100-200MHz. Presumably there would be some kind of bandpass
filter before the LNA (to keep out of band crud from saturating it,
if nothing else) and that might be the main protection. Or the first
victim, if an amp in 10ns will blow up your filter's caps. I'd bet there's
art about how to best arrange the low and high, and their components,
to get rugged at the antenna.