We are having trouble measuring the amount of antigen specific human IgM in an ELISA. We know that part of the problem is antigen specific IgG but have not had success removing IgG with protein A (both the IgG and the IgM reactions were reduced by about half). Any suggestions?