PROC GLIMMIX

[GLIMMIX]

Hi, I am new to this procedure and analysis of continuous outcomes. Yet i have a data set where y is a continuous, badly skewed outcome even when log-transformed. I have a categorical predictor x with 3 levels. and another continuous predictor z in the model. I have read about Glimmix but cannot come up with a good syntax that would give me odds ratios for variable x in the simple model in relation to y. Please help!
I am trying this: proc glimmix; class x (but it does not let me do the referent category like in logistic), model y=x z /s oddsratio; run.
I am also not sure how to do order statement (did not work out) or contrast statement to obtain the p-values and OR with CI for the 3-level variable x.
Please HELP with correct syntax, thanks!