This study aimed to investigate long-term stability and variability of diurnal cortisol and alpha-amylase patterns. Diurnal cortisol and alpha-amylase secretion patterns were assessed on a single workday with three waves of measurement across a total time period of 24months in 189 participants. Separate hierarchical linear models were analyzed, with and without a number of potential predictor variables (age, BMI, smoking, chronic stress, stress reactivity). While low long-term stability was found in diurnal cortisol, the stability of diurnal alpha-amylase was moderate across the time period of 24months. Several predictor variables had a positive impact on diurnal cortisol and alpha-amylase secretion patterns averaged across waves. Our findings underpin the notion that long-term stability is not necessarily warranted in longitudinal studies. It is important to choose an appropriate study design when attempting to disentangle clinically and biologically relevant changes from naturally occurring variations in diurnal cortisol and alpha-amylase.