How to incorporate cointegrating equations into the main regression model of multivariate time series?

In multivariate time-series analysis with non-stationary data, I have used the commands varsoc, vecrank, and vec, and accordingly obtained the cointegrating equation(s). How do we interpret these results with respect to our main (original) model? Specifically, how do we arrive at the corrected set of coefficients for the explanatory variables in our main model?