We argue that low-temperature effects in QED can, if anywhere, only be quantitatively interesting for bound electrons. Unluckily the dominant thermal contribution turns out to be level independent, so that it does not affect the frequency of the transition radiation.