Very few, but we've found that the w3wp.exe instance for DefaultAppPool was only spiking CPU because it accounts for the queued connections. Turns out this was a by-product of the real issue, which seems to be code on one specific site which is deadlocking SQL connections and maxing out the TCP stack. If that one site is taken down, the queued connections and CPU usage return to normal.
Thanks for the help though! Dev team is looking into rolling back the updates now until we can figure out the problem with the new code.