Why do confidence limits for fixed effects differ between lme and lmer in a simple linear mixed effects model

Question

I\'m trying to model a gaussian output with two fixed effects (treatment and sequence) and a single random effect patient. Patient is nested within sequence (different patie