Besides trimming and winsorizing, there are other methods to deal with outliers in regression, such as transforming the data, using robust estimators, or adding dummy variables.