Summary. In five patients (including two sisters) an unusual anaemia was characterized by erythroblastic multinuclearity, ineffective erythropoiesis and a positive acidified‐serum test. Unlike PNH, the sugar‐water test was always negative in these patients. The lysis of their cells by acidified normal sera indicated their abnormal sensitivity to an agglutinating and complement‐binding antibody present in some normal subjects. The patients cells gave high agglutination scores with anti‐i, and were unusually susceptible to lysis by anti‐i and anti‐I. The disorder is apparently inherited as an autosomal recessive character.

Review of direct antiglobulin testing (DAT) in 88 patients with multiple myeloma (MM) and five with Waldenstrom's macroglobulinemia revealed 26 cases with a positive DAT. Twenty-two of these had immunoglobulin G-M protein, three had light chain MM, and one had immunoglobulin A-MM protein. None of the immunoglobulin GD-MM (n = 2), nonsecretory MM (n = 5), or Waldenstrom's macroglobulinemia patients (n = 5) were positive. None of the patients had hemolysis attributable to the adsorption of the M protein. The serum concentration of M protein was higher in DAT-positive patients (57.6 +/- 3.8 g/L, mean +/- SEM) than in the negative ones (35.7 +/- 6.4 g/L; probability value of the difference was less than 0.01). The erythrocyte eluates from DAT-positive patients contained a single immunoglobulin, of the same class as the M protein, and did not react with a panel of ABO-compatible erythrocytes. Addition of melphalan during incubation did not affect the results. The M protein of DAT-positive patients was of immunoglobulin G-3 subclass in 7 of 10 patients. A positive direct antiglobulin test frequently is seen in patients with multiple myeloma, the reaction is due to passive adsorption of the M protein onto the erythrocytes, is most frequently observed with immunoglobulin G3-MM, and usually does not produce hemolysis.

An analysis of the results of a compulsory proficiency testing programme in immunohaematology is presented. Error rates have been calculated for the determination of ABO and Rh(D) groups, the direct antiglobulin test and antibody detection according to defined criteria. The introduction of proficiency testing has been associated with alterations in error rates for some determinations. An educational programme introduced for laboratories with poor performance has proved effective in improving their results in the proficiency testing programme.

Investigation of the serum of three patients with positive antibody detection tests demonstrated the cause in each to be an antibody against chloramphenicol, a bacteriostatic agent used in commercial red blood cell reagents. Washing of these red cells prior to use prevented agglutination. All three examples of anti-chloramphenicol antibody were IgM and were in low titer when tested at room temperature and 37 C in saline. Two of the antibodies bound complement. The possibility of an antibody to an ingredient of the commercial preservative solution should be considered if problems are encountered in tests with unwashed commercial red blood cell reagents.