Preventing incompatible transfusions

Each year in Britain 2.2 million units of red cells are transfused and perhaps four deaths due to preventable causes subsequently occur. Providers of health care and the public express most concern about the risk of infections transmitted by transfusion, particularly HIV infection, but procedural errors resulting in ABO incompatibility cause most of the deaths immediately after blood transfusion.^1,2 ABO antibodies are “naturally occurring” and are present in all subjects except those of blood group AB after the first 3 to 6 months of life. Although these antibodies can cause intravascular haemolysis, with disseminated intravascular coagulation and renal failure, most incompatible transfusions are uneventful or result in minor signs and symptoms.^3,4 Only about 1 in 10 ABO incompatible transfusions is fatal.2

Studies in the United States, where all deaths associated with transfusions must be reported to the Food and Drug Administration, show that patients with blood group O are most at risk of dying from ABO incompatibility^1,2 because they have the highest concentrations of potent antibodies directed against A and B antigens. Other causes of death from transfusions, excluding viral infections, are generally not preventable. They include, in order of frequency of reports, acute lung injury related to transfusion, bacterial contamination of red cells or platelet concentrates, delayed haemolytic transfusion reaction due to non-ABO antibodies, transfusion of inadvertently damaged cellular components, and, very rarely, graft versus host disease.

Most incompatible transfusions result not from laboratory errors but from mistakes in identification such as the faulty labelling of samples taken for testing before transfusion or misidentification of the patient receiving the transfusion. Most transfusion errors result from blood being given to the wrong patient.^2,5,6 In a study in Glasgow of 20 000 recipients of more than 60 000 units of blood, an estimated 1 in 3 300 patients received ABO incompatible blood - all because of the failure to identify patients or blood samples correctly.5 In fact, the true frequency with which the wrong blood was transfused must have been at least three times higher as the distribution of ABO blood groups in these patients is such that when the wrong blood is given by chance it is ABO incompatible in only one in three instances. This means that two out of three patients were fortunate enough to receive blood of their own or of a compatible ABO group even though they were not the intended recipient. Moreover, only those patients who had a reaction were investigated, which accounts for only a small proportion of those receiving incompatible transfusions. Recent studies show that the incidence of transfusion errors has not decreased.6

Data from the Food and Drug Administration suggest that during 1976-8, 39 deaths were primarily attributable to transfusions. Incompatible transfusions were responsible for the deaths of 24 patients, and all but two of these deaths were due to ABO incompatibility. The commonest cause of ABO incompatible transfusion was failure to identify the recipient correctly, and this mistake occurred most commonly in the operating theatre.7 In a recent study in two London teaching hospitals of 400 units of red cells transfused into 200 patients, 56 of the patients were given at least one unit of blood without adequate documentation. Two thirds of these transfusions were given in the operating theatre (C Costello, personal communication).

Many systems have been described for minimising risks of misidentification.8 Guidelines on documentation and procedures, prepared by the Blood Transfusion Task Force, were published in 19919 but have been largely ignored. Most hospitals still do not fully document transfusions.

McClelland and Phillips in this week's journal calculate a minimum estimate of the current rate of errors in the administration of blood in Britain (p 1205).10 The authors identified 111 incidents over two years, which resulted in the death of six patients and morbidity in 12. These outcomes represent at least one death per 550 000 units transfused and one incident in which the wrong blood was given to the patient per 29 000 units of red cells transfused. Worryingly, only 40% of the laboratories that took part in the study indicated that they had a system for documenting these incidents, and over half the respondents provided information from memory.

Although in all hospitals haematologists are in charge of blood transfusion departments, most transfusion errors occur at the patient's bedside and are beyond the haematologists' direct responsibility. Most transfusions are administered and documented by junior doctors, anaesthetists, and nurses - mainly a rotating group of staff - with hardly any proper training or awareness of local and national guidelines on operating procedures. It makes sense, therefore, to place overall responsibility for transfusions in hospital under single management. Most serious transfusion errors arise from breaches of current established codes of practice resulting in blood being given to the wrong patient.*RF 1-3* Human errors are inevitable, but systems should be in place to minimise their occurrence. Improvement is possible. During 1986-93, 268 000 units of blood were transfused at the Mayo Clinic, and on only one occasion was ABO incompatible blood transfused owing to misidentification.11 At the clinic only dedicated staff, including nurse transfusionists, are responsible for giving and monitoring blood transfusions.

All hospitals need to set up standard procedures for accurately identifying patients and recognising the signs of transfusion reactions. These procedures must be simple enough to be followed by staff in emergencies. Multidisciplinary audits of the quality of blood administration are invaluable to ensure compliance8 and should be carried out regularly.

Finally, as McClelland and Phillips recommend, a central reporting system for collating data on all serious incidents related to transfusion, similar to that which exists for reporting adverse reactions to drugs, is needed. Until now the main concern has been to improve the quality and microbiological safety of the blood supply to hospitals by the blood transfusion service. It is time for the same resources and attention to be given at hospital level to ensure that blood is prescribed appropriately and administered safely to the intended recipient by well trained staff.