Test of trend using published odds ratios and confidence intervals

[PT-109]

I have done a meta-analysis and have ORs and 95% CIs for three levels of an exposure (all published data).  I know the SAS codes for doing a trend test with RAW data, but because these data are not raw data, but already adjusted and shown in previous publications, I don't know if or how to do a test of trend on those data.  Perhaps some sort of meta-regress or something, but hopefully there is a more direct way.  Thanks very much!  paul