Do semiclassical zero temperature black holes exist?

The semiclassical Einstein equations are solved to first order in epsilon = (h) over bar/M-2 for the case of a Reissner-Nordstrom black hole perturbed by the vacuum stress energy of quantized free fields. Massless and massive fields of spin 0, 1/2, and 1 are considered. We show that in all physically realistic cases, macroscopic zero temperature black hole solutions do not exist. Any static zero temperature semiclassical black hole solutions must then be microscopic and isolated in the space of solutions; they do not join smoothly onto the classical extreme Reissner-Nordstrom solution as epsilon --> 0.