There is nothing wrong with N=81 if the effect is large enough. They did use VO2Max to measure CRF.