Without some code to look at, it is difficult to say definitively. However, most analytic procedures allow you to define a set of variables that will each be analysed in the specified context. Sometimes the syntax is similar to "Model ( A, B, C) * C;"
Reference to the online help for the procedure should assist.
There is a temptation to send the variable names to a macro where each regression is run in sequence. I would suggest you think very carefully before undertaking such a solution.
If you define a P value of perhaps 0.05, then you run a single analysis with a given risk of getting a false positive for that test. If you run 20 such tests, without adjusting your parameters, then you run a more severe risk of getting a false positive. This is usually called experiment wide risk.
The analysis that said the risk of an O ring failing was acceptable failed to take into account the actual cumulative risk of having a series of these O rings exposed to similar stresses. You may recall that particular analysis was partially responsible for the destruction of the space shuttle Challenger. Which serves to remind that repeated single tests may not be suitable for accurately calculating the probability of a given outcome.
Further and more definitive answers on probability tests can be found in the SUGI papers by David Cassell which are available on the SAS support website.