Preliminary reports have suggested that microcolumn technology might be too sensitive for direct antiglobulin testing (DAT). We studied 228 samples from patients with autoimmune diseases and 30 samples from healthy controls to determine the sensitivity of column techniques. Both Sephadex® gel and protein A/G columns were compared with manual methods using rabbit or murine polyspecific reagents. Of the 187 samples that were negative by both manual methods, an additional 29 (15%) and 42 (22%) samples gave weakly positive reactions with the Sephadex® and protein A/G bead columns, respectively. Subsequently, there was poor correlation between manual and column techniques (