key: cord-281395-qxu41hta authors: Rogers, M. A. M.; Rohde, J. M.; Blumberg, N. title: Haemovigilance of reactions associated with red blood cell transfusion: comparison across 17 Countries date: 2015-12-21 journal: Vox Sang DOI: 10.1111/vox.12367 sha: doc_id: 281395 cord_uid: qxu41hta BACKGROUND AND OBJECTIVES: The recent establishment of the National Healthcare Safety Network Hemovigilance Module in the United States affords an opportunity to compare results with those of other developed nations. MATERIALS AND METHODS: Using data from national haemovigilance systems, reactions associated with red blood cell (RBC) transfusion and residual risks of transfusion‐transmitted infectious diseases were assembled from 17 nations. Country‐specific rates of adverse events were pooled using random‐effects Poisson regression. RESULTS: Febrile non‐haemolytic and delayed serologic transfusion reactions were the most frequent adverse events reported after RBC transfusion, occurring in 26 patients per 100 000 RBC units and 25 patients per 100 000 RBC units administered, respectively. Rates of allergic, febrile non‐haemolytic and delayed haemolytic transfusion reactions in the United States were significantly greater than the pooled rates from other countries. Frequencies of adverse events generated from the national haemovigilance programme in the United States were considerably lower than when obtained through active surveillance. CONCLUSION: Haemovigilance reports of adverse events in the United States are comparable to, or greater than, reports from other developed countries. Rates generated from haemovigilance programmes are lower than those obtained through active surveillance. The lack of universal leucoreduction of RBC units may be a contributing factor to the higher rate of some adverse events in the United States. First established in France and Japan in 1993 as a response to the vulnerability of the blood supply after the emergence of HIV [1, 2] , haemovigilance systems aim to improve the safety of the blood supply through systematic surveillance of transfusion-related adverse events. Such systems generate recommendations so as to encourage safety throughout the entire process, from blood donation through monitoring of recipients. Several directives from the European Union relate to the regulation, monitoring and safety of blood products, with the haemovigilance component helping to direct member nations' reporting of serious transfusion-related adverse reactions and events [3] . Therefore, haemovigilance systems are widespread in European nations, with prime examples being the United Kingdom with its voluntary Serious Hazards of Transfusion reporting system (99Á5% participation in 2013) [4] , and the Netherlands with its Transfusion and Transplantation Reactions in Patients haemovigilance programme (98% participation in 2013) [5] . The USA opened its first nationwide haemovigilance system in 2010 as a voluntary module of the National Healthcare Safety Network [6] . Results from this new system are now available and can be compared with existing haemovigilance programmes. As more nations begin to implement haemovigilance programmes, an assessment of the abilities of national systems to report reactions and a comparative review of event rates can be informative. Haemovigilance data on transfusion-related adverse reactions and transfusion-transmitted infections were available from the national haemovigilance systems in each of the following countries: Australia [7] , Canada [8] , Denmark [9] , Finland [10] , France [11] , Germany [12] , Ireland [13] , Japan [14] [15] [16] , the Netherlands [5] , New Zealand [17] , Norway [18] , Portugal [19, 20] , Spain [21] , Sweden [22] , Switzerland [23] , United Kingdom [4] and the USA [6] . Data regarding adverse events after RBC transfusion were collected: allergic reactions, anaphylactic reactions, febrile non-haemolytic transfusion reactions (FNHTR), acute and delayed haemolytic transfusion reactions, hypotensive reactions, transfusion-related acute lung injury (TRALI), transfusionassociated dyspnoea, transfusion-associated circulatory overload (TACO), delayed serologic transfusion reactions, post-transfusion purpura and transfusion-transmitted bacterial sepsis. Some countries reported additional categories and these were included as well (acute transfusion-related pain, haemosiderosis). For each country, data from the most recently available report were utilized. Risks of transfusion-transmitted viruses were available for agents with mandated testing. Reporting methods differ by country, although in the USA, data regarding residual risk of viral transmission were derived from donor testing since long-term follow-up of recipients is not feasible [24] . Because it is theoretically possible that a donor could have newly acquired a virus but is seronegative at the time of the blood donation, seroconversion rates during this infectious window period are used to predict the residual risk of viral infection [24] . Rates of non-infectious adverse events were calculated for each country using the number of events (numerator) and the number of RBC units administered (denominator). Rates were reported for each country as events per 100 000 RBC units for non-infectious adverse events and events per million RBC units for infectious adverse events. Canada and Spain were exceptions; adverse events were recorded for components combined (RBCs, platelets, plasma). In the USA, TACO was the only event in which reporting was performed for components combined (RBCs, platelets, plasma). In secondary analyses, adverse event rates from active surveillance in the USA were compared with the rates generated from the passive haemovigilance programme; in these instances, the rates were given using several denominators (per patient, per unit and per transfusion-related hospital stay). Following current recommendations for summarizing rates [25] , random-effects Poisson regression models were generated to pool rates of transfusion-related adverse events with 95% confidence intervals. That is, the numbers of events were fit to a Poisson distribution with the number of RBC units included as an offset. Countries that reported zero events were included and, in such instances, a one-sided exact Poisson 97Á5% confidence interval was used for the country-specific rate. Heterogeneity was assessed using a gamma density function (shape parameter = 2, scale = 0Á5) for the variance of the random intercept (between-country variance) in the Poisson model. This Rate Index of Heterogeneity (RIH) is unit-independent, encompasses a range from 0% to 100% heterogeneity and is directly derived from the underlying random-effects Poisson model for assessment of rates. Analyses were conducted using Stata/MP 13Á1 (College Station, Texas, USA). Allergic reactions after RBC transfusion occurred in 11 patients per 100 000 RBC units administered (95% CI: 6Á55/100 000 to 18Á08/100 000) with a RIH of 58Á5% (Fig. 1) . The rate of allergic reactions in the USA was significantly greater (53Á61/100 000; 95% CI: 49Á59/ 100 000-57Á87/100 000) than the pooled rate of the other developed countries combined (9Á7/100 000; 95% CI: 5Á93/100 000-15Á85/100 000). Some countries reported anaphylactic reactions separately and these occurred at a lower rate of 0Á9 per 100 000 RBC units (RIH = 17Á1%). In the United Kingdom, acute transfusion reactions were defined as instances of anaphylaxis or severe allergic reactions, severe febrile reactions, severe hypotensive reactions and severe mixed reactions; these occurred at a rate of nine patients per 100 000 RBC units. There was variability in the rates of FNHTR across countries (RIH = 94Á0%). Overall, the rate was 26 patients per 100 000 RBC units, although the Netherlands, New Zealand and the USA recorded rates in excess of 100 patients per 100 000 RBC units, or approximately one in one-thousand units (Fig. 2) . The rate of FNHTR in the USA was significantly greater (106Á32/100 000; 95% CI: 100Á63/100 000-112Á25/100 000) than the pooled rate of the other developed countries combined (22Á85/100 000; 95% CI: 9Á74/100 000-53Á63/100 000). Hypotensive reactions after RBC transfusion occurred less frequently, in 2 patients per 100 000 RBC units (RIH = 34Á3%). In Spain, hypotensive reactions were combined with FNHTR, yielding a rate of 48 patients in 100 000 RBC units. Acute transfusion-related pain was reported separately in the Swedish haemovigilance system, occurring in one patient per 100 000 RBC units (Fig. 2) . Rates of acute and delayed haemolytic transfusion reactions are shown in Fig. 3 . Acute haemolytic transfusion reactions occurred in 1 patient per 100 000 RBC units administered (RIH = 8Á3%), while a delayed haemolytic transfusion reaction occurred in approximately two patients per 100 000 RBC units (RIH = 37Á6%). In Germany, Spain and the United Kingdom, both acute and delayed haemolytic transfusion reactions were combined, yielding a pooled rate of 1Á4/100 000 units (RIH = 57Á0%). Both Ireland and the USA tended to have delayed haemolytic transfusion reactions at a greater frequency than other countries, although the confidence intervals for Ireland were wide indicating fewer RBC units due to a smaller population. The rate of delayed haemolytic transfusion reactions in the USA was significantly greater (6Á94/100 000; 95% CI: 5Á54/100 000-8Á58/ 100 000) than the pooled rate of the other developed countries combined (1Á6/100 000; 95% CI: 0Á9/100 000-2Á8/100 000). Respiratory reactions to RBC transfusion are shown in Fig. 4 . Transfusion-associated dyspnoea was reported in two patients per 100 000 RBC units (RIH = 77Á4%). New Zealand was an outlier in this category. TRALI associated with RBC units was quite low (Fig. 4) , at a pooled rate of 0Á35 per 100 000 RBC units (RIH = 73Á2%). TACO occurred in approximately three patients for each 100 000 RBC units transfused, although there was some variation in rates across countries ( Fig. 5 ; RIH = 76Á9%). Haemosiderosis was reported separately in France and Spain and occurred at a rate of 0Á3 for every 100 000 units transfused (RIH = 99Á6%). The results for delayed serologic transfusion reactions are given in Fig. 6 , showing heterogeneity across countries (RIH = 91Á5%; rate = 24Á6/100 000 units), with elevated rates in the Netherlands and Switzerland. Posttransfusion purpura occurred rarely, at a rate of 0Á08 per 100 000 RBC units administered ( Fig. 6 ; RIH = 0%). Other transfusion-related reactions, as shown in Fig. 6 , varied by country and were not described in great detail in the country-specific annual reports. Documented cases of transfusion-transmitted bacterial sepsis occurred rarely, approximately once in every million RBC units administered ( Fig. 7 ; RIH = 57Á8%). anaphylaxis, hypotension and purpura, the rates were quite consistent across countries. Rates of some reactions, however, were notably elevated in particular countries. Rates of allergic reactions were higher in the USA and New Zealand; rates of FNHTR were elevated in the USA, the Netherlands and New Zealand; and TACO was more commonly reported in Canada, France, Ireland, the Netherlands and the USA. New Zealand reported a greater rate of transfusion-associated dyspnoea in 2013. However, in a recent report, more than half of the dyspnoea cases were reclassified when additional information was retrieved (many being reclassified as TACO) [45] . Overall, TACO occurred at a rate of three cases per 100 000 units when pooled from haemovigilance reports, with the incidence of TACO in the USA alone being one in 10942 units. However, based on active surveillance studies from the USA, 0Á6% to 8% of patients receiving RBC transfusions develop TACO. This disparity between active and passive surveillance rates may stem from insufficient reporting of reactions to haemovigilance systems. For example, researchers from the Mayo Clinic noted that there were 176 cases of TACO in their investigation but only three were in the transfusion database that housed adverse reactionsand even then, these three cases were not labelled as TACO [31] . Given the more comprehensive monitoring of fever after RBC transfusion, the variability of rates of FNHTR across countries may be a reflection of dissimilarities in patient populations, differential use of pretransfusion antipyretic medications or the preparation of RBC units. Of note, the USA has not adopted universal leucoreduction (70Á5% RBC units leucoreduced in 2011)[46] and FNHTR has been reported more frequently in patients receiving non-leucoreduced products [33, 39, 41, 42, 47] . FNHTR rates in the USA are higher than the pooled FNHTR rates in other developed countries and the use of non-leucoreduced products could partially account for these findings. The variability in rates of allergic reactions across nations may stem from differences in case definitions. In Australia, for example, a severe allergic reaction would be recorded if rash, allergic dyspnoea, angioedema, generalized pruritis or uticaria occurred within 24 h of the transfusion [7] . In the USA, a severe allergic reaction would be recorded if two or more symptoms (conjunctival oedema; oedema of lips, tongue and uvula; erythema and oedema of the periorbital area; generalized flushing; hypotension; localized angioedema; maculopapular rash; pruritus; respiratory distress or bronchospasm; urticaria) occurred during or within 4 h of the cessation of transfusion [48] . Moreover, in some countries such as New Zealand, both non-severe and severe allergic reactions are reported [17] (perhaps reflecting their higher rates as shown in Fig. 1 ) while in the USA, reporting of nonsevere allergic reactions is not required [48] . Other reasons for differences in rates may be variations in hospital-specific reporting practices and the mandatory/ voluntary nature of the reporting. For events such as delayed serologic transfusion reaction and delayed haemolytic transfusion reaction, the definition hinges on laboratory testing and not symptoms. Rate differences may be a reflection of the resources afforded for testing of alloantibodies and the feasibility of longitudinal observation. Information regarding follow-up procedures or standardization of procedures would enhance the interpretability of rates. Alternatively, haemovigilance programmes may reconsider the usefulness of expending resources to capture such data when the likelihood of consistent, meaningful information is low. Perhaps periodic systematic testing in high users of transfusion would be more appropriate such as in patients with sickle-cell disease or cancer, or in participants of cardiovascular disease registries. Results from this investigation indicate that the residual risks from currently tested infectious agents are very low and, in many instances, could be considered negligible. However, it is the potential risks from other infectious agents (not yet tested) which are more pertinent at this time. These include vCJD, hepatitis E virus, dengue viruses, chikungunya virus, Babesia spp., West Nile virus and Middle East respiratory syndrome coronavirusalthough the list of potential threats includes approximately 70 such agents [49] . For example, in the USA, transfusion-transmitted babesiosis has been documented in more than 200 instances and is associated with the greatest case fatality rate of transfusion-related infections [50] . Haemovigilance for transfusion-transmitted infection is meant to capture pathogen transmission from donor to recipient. The requirements include the demonstration of the pathogen in the transfused component, or in the donor at the time of donation, or in an additional component from the same donation, or in an additional recipient of a component from the same donation [48] . Also required is documentation that there are no other potential exposures to the pathogen that could be identified in the recipient and either evidence that the recipient was not infected with the pathogen prior to transfusion or evidence that the pathogen strains are related by molecular or extended phenotypic comparison testing [48] . This is a difficult target to reach when such in-depth testing is not usually conducted with each transfusion. Therefore, the pooled rate of transfusion-transmitted bacterial sepsis was very low (0Á83/1 million RBC units). There are few published reports of prospective surveillance for comparison purposes. Barrett and colleagues [51] found a bacterial contamination rate of 3Á2/100 000 RBC units, while Dzieczkowski and colleagues [37] reported this rate to be 14Á1/100 000 RBC units. In a more recent report from Denmark, 35% of RBC units contained viable bacteria which were obtained from donors 50 years of age or older [52] . There are avenues for improvement in existing haemovigilance systems. The addition of active surveillance components may enhance reporting. Expansion of electronic medical records to incorporate more information regarding blood administration may augment comprehensiveness. Adjunct programming routines could be enacted to capture pre-and post-transfusion temperatures so that detection of FNHTR could be improved. Likewise, electronic capture of pre-and post-transfusion blood pressures could assist the detection of hypotensive reactions. Some have already merged records from transfusion and apheresis medicine with the main electronic medical record [53] [54] [55] . The SCANDAT2 is a successful example of linked donor and recipient health information in Sweden and Denmark [55] . In addition, international networks (e.g. International Haemovigilance Network [56], Global Vigilance and Surveillance Database for Medical Products of Human Origin [57] , World Health Organization Global Database on Blood Safety [58] ) provide opportunities for surveillance and the improvement of blood safety. National haemovigilance programmes afford an important aspect of patient safety. Such systems generally capture serious transfusion reactions but, with improvement, could generate more comprehensive information. Crosscountry comparisons can provide the basis for discovery and continuous process improvement. decreases the incidence of febrile non- 2·24 (1·03, 4·87) Herv e P: L'organisation de l'h emovigilance en France The new Japanese blood law: its impact on blood safety and usage European Blood Alliance: The revision of the European blood directives: a major challenge for transfusion medicine TRIP annual report: Hemovigilance extended version Transfusion-related adverse reactions reported to the National Healthcare Safety Network Hemovigilance Module, United States National Blood Authority: Australian Haemovigilance Report: data for 2009-10 and 2010-11 Public Health Agency of Canada: Transfusion transmitted injuries surveillance system (TTISS): summary results for Danish Registration of Transfusion Risk (DART) hemovigilance report for Adverse transfusion events in Finland French Agency for the Safety of Health Products: Annual Haemovigilance report Haeomovigilance report of the Paul-Ehrlich Institute: Assessment of the reports of serious adverse transfusion reactions Japanese Red Cross Society: Nonhemolytic adverse transfusion reactions reported to JRC Blood Centers Japanese Red Cross Society: Transfusion-related acute lung injury Japanese Red Cross Society: Infectious cases of probably related to transfused blood components National Haemovigilance Group. National haemovigilance programme Norwegian Knowledge Centre for the Health Services: Monitoring of Blood: haemovigilance report Report of the Portuguese Haemovigilance System Activity report of the services of blood and transfusion medicine Quality and Innovation: Hemovigilance Report from the BIS -Blood Monitoring in Sweden Swissmedic Haemovigilance Annual report The risk of transfusion-transmitted viral infections. The Retrovirus Epidemiology Donor Study Simulation-Based Comparison of Methods for Meta-Analysis of Proportions and Rates. AHRQ Publication No. 13(14)-EHC084-EF Transfusion-associated circulatory overload in orthopedic surgery patients: a multi-institutional study An analysis of blood management in patients having a total hip or knee arthroplasty Transfusion-related acute lung injury and pulmonary edema in critically ill patients: a retrospective study Incidence and transfusion risk factors for transfusion-associated circulatory overload among medical intensive care unit patients Hemotherapy bedside biovigilance involving vital sign values and characteristics of patients with suspected transfusion reactions associated with fluid challenges: can some cases of transfusion-associated circulatory overload have proinflammatory aspects? Characterizing the epidemiology of perioperative transfusion-associated circulatory overload Transfusion-associated circulatory overload (TACO) and potential risk factors among the inpatient US elderly as recorded in Medicare administrative databases during Reduction of febrile but not allergic reactions to RBCs and platelets after conversion to universal prestorage leukoreduction Premedication with acetaminophen or diphenhydramine for transfusion with leucoreduced blood products in children A prospective, randomized, doubleblind controlled trial of acetaminophen and diphenhydramine pretransfusion medication versus placebo for the prevention of transfusion reactions Febrile transfusion reaction: what blood component should be given next? Characterization of reactions after exclusive transfusion of white cell-reduced cellular blood components Characterization of reactions after transfusion of cellular blood components that are white cell reduced before storage Prestorage universal WBC reduction of RBC units does not affect the incidence of transfusion reactions Allergic transfusion reactions: an evaluation of 273 consecutive reactions Universal leukoreduction Stramer SL: Current perspectives in transfusion-transmitted infectious diseases: emerging and re-emerging infections Blood Products Advisory Committee, Food and Drug Administration: Strategies for implementation of antibody and nucleic acid-based testing for Babesia microti in blood donors Strategies for the avoidance of bacterial contamination of blood components Viable bacteria associated with red blood cells and plasma in freshly drawn blood donations Development of electronic medical record charting for hospital-based transfusion and apheresis medicine services: Early adoption perspectives Electronic health record surveillance algorithms facilitate the detection of transfusion-related pulmonary complications The new Scandinavian Donations and Transfusions database (SCANDAT2): a blood safety resource with added versatility The Global Vigilance and Surveillance Database for All authors contributed to the design, analysis or data interpretation; made critical revisions of the manuscript; and approved the submitted final version.