Two patients with acute immune haemolytic anaemia caused by diclofenac are described. Both patients had developed IgG drug-independent autoantibodies and drug-dependent antibodies. The drug-dependent antibodies in one patient reacted with red blood cells (RBC) only in the presence of urine from patients receiving diclofenac (ex vivo antigen) but not in the presence of the drug itself or its known metabolites. This antibody appeared to recognize a trace metabolite which has not yet been identified. The drug-dependent antibodies in the second patient reacted with RBC in the presence of urine as ex vivo antigen as well as in the presence of the drug itself and its main metabolites. Incorrect diagnosis of such cases often is due to the occurrence of autoantibodies and/or unusual drug metabolism and excretion.

Directed blood donations from family members are frequently desired for pediatric patients. A reduced risk for transfusion-transmitted deseases could not be prooved. There are indications that recommendation of directed blood donation could oblige members of risk groups for HIV or hepatitis infection to donate blood since their voluntariness ist not guaranteed and the possibility of self-exclusion is omited. In some instances it may be necessary to perform directed blood donations from relatives because of immunological or medical principles, e.g. transfusion of maternal platelets in neonatal alloimmune thrombocytopenia. In contrast, directed blood danations should not be realized for most hematologic-oncologic patients because they bear the risk of immunisation against alloantigens of the donor and can thus prevent a later bone marrow transplantation from relatives. Inspite or veservations against directed blood donations for individual cases the pediatrician and the transfusion medicine service will have to decide on blood donations from relatives. In any case the blood products should be irradiated with 30 Gy to avoid a transfusion-related graft-versus-host-reaction.

Directed blood donations from family members are frequently desired for pediatric patients. A reduced risk for transfusion-transmitted diseases could not be proved. There are indications that recommendation of directed blood donation could oblige members of risk groups for HIV or hepatitis infection to donate blood since their voluntariness is not guaranteed and the possibility of self-exclusion is omitted. In some instances it may be necessary to perform directed blood donations from relatives because of immunological or medical principles, e.g. transfusion of maternal platelets in neonatal alloimmune thrombocytopenia. In contrast, directed blood donations should not be realized for most hematologic-oncologic patients because they bear the risk of immunisation against alloantigens of the donor and can thus prevent a later bone marrow transplantation from relatives. Inspite of reservations against directed blood donations for individual cases the pediatrician and the transfusion medicine service will have to decide on blood donations from relatives. In any case the blood products should be irradiated with 30 Gy to avoid a transfusion-related graft-versus-host-reaction.

A third patient with autoimmune hemolytic anemia due to autoantibodies against Gerbich antigens is described. The patient's serum contained strong hemagglutinating antibodies of the IgA plus IgG classes which reacted with all red blood cells (RBC) tested, but not with Gerbich-negative cells. Although the patient was typed as Gerbich positive, his serum failed to react with his own RBC, and the sensitization of his erythrocytes with autoantibodies was only demonstrable if eluates of his RBC were used. The failure of the autoantibodies to react with autologous RBC at the peak of hemolysis most likely reflects a weakening of Gerbich antigens during the course of autoimmune hemolytic anemia.

During the last 4 years, we have studied a total of 531 adults and 68 children with clinically and serologically well-defined forms of immune hemolytic anemias. Among these, Donath-Landsteiner (DL) hemolysis was the underlying disease in 22 of the 68 children (32.4%), but was not observed in adults. All children with DL hemolysis suffered from acute infections presumably of viral origin. In none of the cases was the DL hemolysis suspected clinically. Boys were more often affected than girls. The hemolytic episodes were severe, but resolved within few weeks. Serologically, all patients had a strongly positive direct antiglobulin test (DAT) using anti-C3d reagents, but a weak (n = 6) or negative (n = 16) IgG-DAT. DL hemolysins were always weak and transient, detectable with enzyme-treated red blood cells (RBC) in all, with untreated RBC in 12 of 22 sera. To explore the reason why these weak antibodies can cause extensive hemolysis in vivo, we compared the action of DL antibodies and of cold agglutinins (anti-I) on RBC by several reincubations at 4 and at 37 degrees C. The data obtained from this experiment demonstrate a stronger aggravation of hemolysis by DL than by anti-I antibodies, presumably due to low-affinity interaction between the cells and DL antibodies.

A cold autoantibody detected in the serum of a patient with chronic idiopathic cold agglutinin disease and hemolytic anemia is described. The antibody reacted with adult as well as with cord red cells, and its reactivity was strongly diminished by treatment of the cells with neuraminidase and to a lesser degree by treatment with protease. Thus, the specificity of the antibody is distinct from those of all cold antibodies yet described.

A third patient with autoimmune hemolytic anemia due to autoantibodies against Gerbich antigens is described. The patient's serum contained strong hemagglutinating antibodies of the IgA plus IgG classes which reacted with all red blood cells (RBC) tested, but not with Gerbich-negative cells. Although the patient was typed as Gerbich-positive, his serum failed to react with his own RBC, and the sensitization of his erythrocytes with autoantibodies was only demonstrable if eluates of his RBC were used. The failure of the autoantibodies to react with autologous RBC at the peak of hemolysis most likely reflects a weakening of Gerbich antigens during the course of autoimmune hemolytic anemia.

Summary. S protein is a plasma glycoprotein (Mr= 78 000) which binds to nascent C5b‐7 complexes upon complement activation in the fluid phase in whole serum. It thereby protects innocent bystander cells from complement mediated lysis. It is unknown whether S protein also functions as complement inhibitor on cell surfaces. We here report that S protein is recognized on red blood cells (RBC) from patients with paroxysmal nocturnal haemoglobinuria (PNH), but not on normal RBC. RBC from eight PNH patients showed 12–48% haemolysis subsequent to complement activation in the fluid phase, while normal RBC did not respond. Preincubation of the PNH cells with affinity‐purified antibodies against human S protein resulted in a three‐ to five‐fold increase of haemolysis, while preincubation of these cells with S protein decreased haemolysis by 40%. In contrast, haemolysis remained unaffected by other unrelated antibodies, i.e. IgG anti‐Rh(D) and anti‐A. If PNH RBC, normal RBC pretreated with 2‐amino‐ethylisouronium bromide (AET), or untreated normal RBC, respectively, were incubated with purified S protein in vitro, the uptake of antibodies against S protein was significantly enhanced with PNH and with AET‐treated, but not with untreated normal RBC. Additionally, while normal RBC did not respond to reactive lysis initiated by purified C5b‐6 and C7, PNH as well as AET‐RBC showed significant haemolysis that could be inhibited by S protein in a dose‐dependent fashion. These findings strengthen the assumption that the increased sensitivity of PNH cells towards reactive complement lysis is either due to the lack of an inhibitor of the terminal complement sequence and/or enhanced insertion of the membrane attack complex. These defects of PNH RBC may partly be overcome by the fluid phase complement inhibitor S protein which binds to PNH RBC and may thereby suppress homologous cytolysis.

Hemolysis mediated by human antibodies is generally ascribed to the attack of red blood cells (RBC) by complement. We here extend earlier in vitro observations which indicate that potent cold agglutinins can directly cause lysis of RBC without the participation of complement. We have noted that EDTA plasma taken from patients with cold agglutinin disease is frequently reddish if the plasma is not immediately separated from the cells at 37 degrees C. Moreover, eluates prepared in such cases from plasma or heat-inactivated serum (30 min at 56 degrees C) by absorption (at 4-20 degrees C) and elution (at 37 degrees C) are usually contaminated with hemoglobin, and a large number of RBC used for absorption is lost during the procedure. To characterize this phenomenon further, we examined the effect of different hemagglutinating antibodies in vitro on normal RBC in the absence of complement. Hemolysis (5-17%) of RBC only occurred after treating the cells with potent antibodies at low temperatures (0-20 degrees C). This hemolysis increased 2- to 3-fold when the RBC were treated with an enzyme and decreased with rising temperature. Unlike cells hemolyzed by complement activation, no C5b-9 complexes could be detected on RBC damaged by this mechanism.

Immune hemolytic anemia (IHA) related to cephalosporins is rare and generally considered to be the result of a drug-adsorption mechanism. In previously reported cases, the hemolysis was usually extravascular and the causative antibodies were IgG, incapable of activating complement, and demonstrable by the direct or indirect antiglobulin test using red cells (RBCs) pretreated in vitro with cephalosporin. The authors report a patient in whom acute intravascular hemolysis developed while she was receiving cefotaxime (a cephalosporin as yet not reported to cause IHA). The patient's RBCs were coated only with complement fragments (C3d), even at the peak of the hemolytic episode. Her serum and eluates repeatedly yielded negative results when tested against normal or cephalosporin-coated RBCs. However, strong hemagglutination and C5b-9-mediated hemolysis were observed if the patient's serum was tested against RBCs in the presence of the drug, its ex vivo antigen and, to a lesser degree, cephalothin and ceftriaxon, but not in the presence of penicillin and other related cephalosporins. The positive reactions were not changed by preincubating the serum with different amounts of the drugs. All of these findings reflect the typical picture of drug-induced IHA by the so-called "immune complex" mechanism and not by the drug-adsorption mechanism. The authors conclude that cephalosporin can cause immune hemolysis in two ways: the drug-adsorption mechanism and, as described here, the "immune complex" mechanism.

900 pregnancy sera were screened for monocyte antibodies. 23 sera (2.6%) were found to be reactive with monocytes in an allotypic pattern distinct from blood group ABO, HLA-A,B,C, DR and DQ specificities. Because of strong reactions and high reproducibility, 4 sera with pure endothelial monocyte (EM) reactivity and 2 sera also positive with B lymphocytes of certain donors were selected for population and family studies. The frequency of positive reactions in a panel of 26 random donors obtained with the pure EM sera was 13, 8, 8 and 13% with clearly distinct patterns. Analysis of the tentatively designated EM antigens 1.1, 1.2, 2.1, and 2.2 in 7 families revealed definite segregation with HLA. By means of two crossing-over events within the HLA region the two hypothetical EM loci were localized centromeric of HLA-A and one of both loci centromeric of HLA-B.