Administration route could affect Q?

Dear Nonmem users, I am modelling animal data of a new drug in dogs. I have one iv dose group and 3 oral dose groups (of the same range). I started modelling the iv data to obtain good initial estimates of the systemic parameters and then, including the oral data I modelled the presystemic parameters. Once I obtained a reasonably good model I started with the covariate model. At this step, I observedthat the administration route was statistically significant as categorical variable in the inter-compartimental clearance (Q), something that I can’t explain well. I have heard several times “don´t try covariables that you don’t think could affect your parameters” but in this case does any of you have experience with similar situations? This effect has not been observed in rats. Couldthe blood disposition of the drug (faster and higher after Iv dose than after oral) modify the Q? Please give me references. Thank you in advance for your kind help. BR. Mark.