The finding from secondary analyses that a higher baseline total cholesterol was predictive of LDL response to b-glucan consumption in responders is consistent with previous studies, but regression to the mean may have been a confounder in this analysis.