HLM Longitudinal Analysis with 2 Time Points

Does longitudinal HLM with only two time points differ radically from linear growth curve modeling? I know Raudenbush says you need 3 or more timepoints to test for a linear or other trend but one can still test for change over two time points.