Delayed haemolytic transfusion reaction in adults with sickle cell disease: a 5-year experience

Transcription

1 research paper Delayed haemolytic transfusion reaction in adults with sickle cell disease: a 5-year experience Jennifer B. Vidler, 1,2 * Kate Gardner, 1,2 * Kenneth Amenyah, 2 Aleksandar Mijovic 2 and Swee L. Thein 1,2 1 Molecular Haematology, Faculty of Life Sciences & Medicine, King s College London, and 2 Department of Haematological Medicine, King s College Hospital NHS Foundation Trust, London, UK Received 19 November 2014; accepted for publication 19 January 2015 Correspondence: Swee L. Thein, Molecular Haematology, Faculty of Life Sciences & Medicine, King s College London, The James Black Centre, 125 Coldharbour Lane, London SE5 9NU, UK. sl.thein@kcl.ac.uk *Equal contribution. Summary Delayed haemolytic transfusion reactions (DHTR) are potentially lifethreatening complications in patients with sickle cell disease (SCD). Between 1 August 2008 and 31 December 2013, 220 of 637 adult patients in our centre had at least one red blood cell (RBC) transfusion in 2158 separate transfusion episodes. Twenty-three DHTR events occurred in 17 patients (13 female) including 15 HbSS, one HbSC and one HbSb 0 thalassaemia, equating to a DHTR rate of 77% of patients transfused. Mean interval from RBC transfusion to DHTR event was d, and typical presenting features were fever, pain and haemoglobinuria. Twenty of the 23 (870%) DHTR episodes occurred following transfusion in the acute setting. Notably, 11/23 (478%) of DHTRs were not diagnosed at the time of the event, most were misdiagnosed as a vaso-occlusive crisis. 16/23 DHTRs had relative reticulocytopenia, which was more common in older patients. Seven of 23 episodes resulted in alloantibody formation, and three caused autoantibody formation. DHTRs are a severe but uncommon complication of RBC transfusion in SCD and remain poorly recognized, possibly because they mimic an acute painful crisis. Most of the DHTRs are triggered by RBC transfusion in the acute setting when patients are in an inflammatory state. Keywords: delayed haemolytic transfusion reaction, hyperhaemolysis, acute pain episode, allo-antibodies, sickle cell disease. Blood transfusion is increasingly used in the prevention and treatment of many of the complications of sickle cell disease (SCD) (Drasar et al, 2011). The benefits of transfusion must be weighed against its risks, including iron overload, infections and haemolytic transfusion reactions. A delayed haemolytic transfusion reaction (DHTR), occurring more than 24 h after a red blood cell (RBC) transfusion, is a potentially fatal complication, but can easily be missed. The main cause of DHTR is alloimmunization, and in SCD, this is partly due to the donor-recipient red cell antigen mismatches from ethnic differences, between largely Caucasian donors and largely African and African-Caribbean recipients (Yazdanbakhsh et al, 2012). Pathogenic mechanisms of DHTR in SCD are complex; not only does alloimmunization not necessarily lead to DHTR, but evidence of alloimmunization is not always found in DHTR episodes. The diagnosis of DHTRs is further challenged in SCD where the clinical presentation mimics a vaso-occlusive crisis (VOC). In severe cases of DHTR, haemoglobin drops to below pre-transfusion levels; some authors have described this as a hyper-haemolytic transfusion reaction or hyperhaemolysis syndrome (Davey et al, 1980; Petz et al, 1997). While alloimmunization rates of 5 36% have been quoted in SCD populations (King et al, 1997; Talano et al, 2003), there are few data on the incidence of DHTR. Aygun et al (2002) described a rate of clinically manifested (rather than purely serological) DHTR events of 1 in 62 (16%) adult patients with SCD transfused over a 4-year period. We have reviewed our large cohort of adult patients with SCD to assess the prevalence, natural history and outcome of clinically manifested DHTR in SCD. Patients and methods A retrospective review was undertaken of all red cell transfusions in the 637 adult patients (over 16 years of age) with SCD treated at King s College Hospital, London (KCH), between 1 August 2008 and 31 December First published online 5 March 2015 doi: /bjh ª 2015 John Wiley & Sons Ltd

2 DHTR in Sickle Cell Disease Data were collected by interrogation of the Electronic Patient Record (EPR) system in KCH and National Health Service Blood and Transplant (NHSBT). The EPR contains complete clinical, laboratory and transfusion information for all patients treated at our institution, but not information from other hospitals. However, most of our patients live locally and attend KCH for both acute admissions as well as in steady state for follow-up. We therefore have good clinical, transfusion and laboratory histories for the vast majority of our patients. Our institution s indications for transfusing patients (either simple or exchange) in the acute setting include acute chest syndrome (ACS), stroke, severe anaemia (e.g. due to transient red cell aplasia or organ sequestration, or >20% fall in haemoglobin compared to steady-state baseline values), multi-organ failure, severe sepsis (often with associated anaemia) and occasional cases of acute single organ impairment. Our indications for a chronic transfusion programme include primary and secondary stroke prevention, frequent acute pain when hydroxycarbamide therapy has failed, intra-hepatic cholestasis, leg ulceration refractory to compression bandaging, complicated pregnancies, priapism refractory to medical treatment and pre-/post-solid organ transplantation. Throughout this period, our transfusion protocol in SCD includes extended red cell phenotyping pre-transfusion, as per British Committee for Standards in Haematology guidelines (i.e. including C, c, E, e, K, k, Jka, Jkb, Fya, Fyb, S, s antigens), plus red cell antigen matching (Rh and Kell) pretransfusion (British Committee for Standards in Haematology et al, 2013). All transfused red cells were leucodepleted. A DHTR was defined as a significant drop of >25% in haemoglobin (Hb), between 24 h and 21 d after a red cell transfusion; where the Hb drop was associated with at least one of: new red cell antibodies, haemoglobinuria, accelerated HbS% increase post-transfusion with concomitant fall in HbA (if Hb fraction measured), relative reticulocytopenia or prominent lactate dehydrogenase (LDH) rise (over double baseline LDH). Cases were excluded if an alternative cause was more likely (e.g. perioperative blood loss, transient red cell aplasia due to infection, e.g. Parvovirus B19). To detect potential DHTR events, we identified all RBC transfusion episodes and examined the post-transfusion Hb trend, looking for a sharp drop in the 21 d following transfusion. If, on review of the medical notes, at least one of the above associated features of DHTR was present, this was identified as an index case and further laboratory and clinical data were collected. Laboratory data recorded included: Hb, reticulocyte count, LDH, direct antiglobulin test (DAT) and red cell antibody screen. Baseline laboratory values for LDH, Hb and reticulocytes were also reviewed and a steady state calculated as a mean average of all steady state laboratory values taken in haematology clinics during the previous 10 years. Clinical details, including treatment and outcome data, were documented. Our clinical practice is to send positive antibody screen samples to the NHSBT reference laboratory for further immunohaematological analysis. Results During the 54-year study period, 220 of 637 adult SCD patients had at least one RBC transfusion, with a total of units in 2158 separate transfusion episodes. Of the 2158 transfusions, 591 (274%) were given in the acute setting (i.e., when a patient was admitted to a medical ward), and the remainder, in the elective setting. The genotype distribution in the 637-patient cohort was 401 (629%) HbSS, 202 (317%) HbSC, 29 (45%) HbSb + thalassaemia, 4 (06%) HbSb 0 thalassaemia and 1 HbS with hereditary persistence of fetal haemoglobin, compared to 185 (841%) HbSS, 27 (123%) HbSC, 5 (23%) HbSb + thalassaemia and 2 (09%) HbSb 0 thalassaemia in the transfused cohort. From these transfusions, we identified 23 DHTR episodes in 17 patients (13 female), including 15 with HbSS, one with HbSC and one with HbSb 0 -thalassaemia. Thus, 77% (17/220) transfused patients with SCD had one or more DHTRs; alternatively, DHTR complicated 11% of the total 2158 transfusion episodes. Nine of 17 (529%) patients had previously identified red-cell antibodies (allo- and auto-) prior to the trigger transfusion for the described DHTR event. Seven of 17 (412%) patients had a previous DHTR, but these cases overlapped with, rather than being a subset of, those patients with prior alloimmunization. The clinical details of the 23 DHTR episodes are summarized in Table I. The mean age at DHTR episode onset was years. Eleven of the 23 transfusion episodes (478%) occurred in the setting of a known red-cell alloantibody. In 9 of 23 (391%) episodes there was a history of previous DHTR. Trigger transfusion The indications for the trigger transfusion for the 23 episodes were: 18 emergency transfusions for acute VOC with or without ACS, one for acute stroke, one for acute blood loss (post-partum haemorrhage), one elective transfusion pre-operatively and two elective transfusions in patients on chronic transfusion programmes (Table I). Of the 23 episodes of DHTRs, 20 (870%) occurred following transfusion in the acute setting. Given that, in total, 591 of the 2158 transfusion episodes in the study period occurred during acute medical admission, these data provide different DHTR rates for acute and elective transfusions: 20/591 (34%) and 3/1567 (02%), respectively. Six of the 23 DHTR episodes followed exchange blood transfusions, the remainder followed simple transfusions. The mean number of units transfused during the trigger transfusion was units. DHTR episode The DHTR episode presented in two clinical settings: (i) progression from the initial event where the patient remained an inpatient, with clinical deterioration and evidence of ª 2015 John Wiley & Sons Ltd 747

3 J. B. Vidler et al Table I. Clinical characteristics of the 23 DHTR episodes in 17 patients. Demographics Steady state laboratory values Trigger transfusion DHTR event Age Pt Genotype (years)/sex Previous abs Hb (g/l) LDH (iu/l) Retics (10 9 /l) Acute/ elective Pre-Tx Hb (g/l) Units transfused (n) Post-Tx Hb (g/l) Reason for Tx Time to DHTR (days) Clinical presentation of DHTR Min Hb (g/l) Peak LDH (iu/l) Peak Retics (10 9 /l) New antibodies 1 SC 56/F Acute* Stroke 8 Haemoglobinuria, pain, fever Anti-S, autoanti-e 2 SS 17/M Anti-Fy a Acute VOC 10 Pain, fever SS 18/M Anti-Fy a Acute* ACS 2 Haemoglobinuria, pain, fever 3 SS 35/M Acute* ACS 11 Haemoglobinuria, fever, MOF 4 SS 18/M Acute ACS 15 Haemoglobinuria, pain, fever 5 SB0 53/F Elective Pre-op 16 Haemoglobinuria, pain, fever Panreactive autoantibody 55 Haemolysed 285 Anti-Jk b,-s Anti-Fy a, -Jk b, -K, -M, autoantibody SS 51/F Anti-Le a Acute ACS 9 Pain, fever Anti-E 7 SS 49/F Anti-S Acute VOC 3 Pain, fever auto-anti-i 8 SS 37/F Anti-Kp a, -Lu a Elective TP 15 Pain SS 39/F Anti-Kp a, -Lu a Acute Sepsis 18 Pain and AKI Anti-S SS 42/F Anti-Kp a, -Lu a, -S Acute Sepsis 20 Pain SS 16/M Acute ACS 17 Pain Panreactive autoantibody SS 19/M Autoantibody Acute* ACS 9 Pain SS 20/F Anti-C, auto-anti-e Elective* EP 3 Cholecystitis, fever, pain 82 ND 160 Panagglutination in eluate 11 SS 49/F Acute ACS 8 Pain, fever SS 54/F Anti-C, -E, -Fy3, -Jk b, -McCa, -HI, cold-auto- Anti-I Acute ACS 9 Pain, fever, ACS Anti-S, panreactive autoantibody 13 SS 59/F Acute* ACS 1 Haemoglobinuria, pain 14 SS 23/F Acute ACS 11 Pain, fever SS 25/F Acute VOC 7 Pain, fever SS 25/F Acute VOC 16 Pain 38 ND 543 Anti-Jk b 15 SS 41/F Anti-S, -McCa Acute ACS 6 Haemoglobinuria, ACS ª 2015 John Wiley & Sons Ltd

4 DHTR in Sickle Cell Disease Table I. (Continued) Steady state laboratory values Trigger transfusion DHTR event Demographics Peak Retics (10 9 /l) New antibodies Peak LDH (iu/l) Min Hb (g/l) Clinical presentation of DHTR Time to DHTR (days) Reason for Tx Post-Tx Hb (g/l) Units transfused (n) Pre-Tx Hb (g/l) Acute/ elective Retics (10 9 /l) LDH (iu/l) Hb (g/l) Age Pt Genotype (years)/sex Previous abs 16 SS 36/F Acute PPH 5 Haemoglobinuria, pain 17 SS 39/F Anti-K, -Kp a, Acute Sepsis 13 Pain, viral illness SI(a), -S, Autoantibody Pt, patient; DHTR, delayed haemolytic transfusion reaction; Tx, transfusion; ACS, acute chest syndrome; VOC, vaso-occlusive crisis; pre-op, pre-operative; TP, top-up blood transfusion programme; EP, exchange blood transfusion programme; PPH, post-partum haemorrhage; AKI, acute kidney injury; MOF, multi-organ failure; Hb, haemoglobin; LDH, lactate dehydrogenase; retics, absolute reticulocyte count; ND, not done. Reference ranges: Hb g/l; LDH < 240 iu/l; reticulocyte count /l. *Exchange transfusion. haemolysis (six episodes), (ii) readmission with acute pain after discharge (17 cases). The most common presenting symptoms of the DHTR were fever, pain and haemoglobinuria. The average time from transfusion to nadir Hb was d. This may be skewed (increased) by those representing and is a result of how swiftly a patient re-presents to hospital with symptoms, particularly in the context of a recent admission. Notably, 11/23 (478%) of DHTR episodes were not diagnosed at the time, and have only been identified subsequently in clinic or during this auditing process. These unrecognized events were most frequently misdiagnosed as a VOC. Laboratory results Measures of haemolysis. Mean Hb pre-trigger transfusion was g/l. Mean Hb immediately post-trigger transfusion was g/l. The mean nadir Hb during DHTR was g/l. Mean peak LDH was iu/l (reference range LDH < 240 iu/l). All 23 DHTR episodes had nadir Hb significantly below the immediate post-transfusion Hb, 21 of 23 (913%) of the episodes had a nadir Hb lower than the pre-transfusion Hb. The reticulocyte count at peak haemolysis ranged from 106 to /l with a mean of /l. Fifteen of 23 episodes (652%) had a reticulocyte count at peak haemolysis below the patient s steady-state mean reticulocyte count, a relative reticulocytopenia. Reticulocytopenia was, on average, more common in older patients; mean reticulocyte count for those over 40 years of age was significantly lower ( /l) compared to those under 40 years of age ( /l, P = 0037). Immunohaematology. Direct antiglobulin test was requested at least once in 19 (826%) episodes, and was positive in 15 of the 19 (789%) cases. In 7 of these 15 positive cases, DAT results revealed only IgG 1+ positivity and is of uncertain significance. The time to DAT positivity was very variable, with a mean of d from the trigger transfusion date. Notably, a minority (3 of 15) of DAT results were initially negative at peak haemolysis but subsequently became positive when the patient was clinically improving. Furthermore, the strength of DAT positivity did not correlate with either severity of haemolysis as measured by LDH, or clinical severity. Antibodies. Eleven of 23 (478%) episodes of DHTR occurred in the context of previous known red cell alloantibodies. Ten of 23 episodes (435%) resulted in the identification of at least one new red cell antibody (and in some cases, several new antibodies, see Table II), after an average interval of d from trigger transfusion date. Seven of these 10 episodes included alloantibodies, while three of the ten cases involved autoantibodies only. All new red cell alloantibodies were clinically significant. Three episodes required red cell eluate analysis to accurately identify the antibodies. ª 2015 John Wiley & Sons Ltd 749

5 J. B. Vidler et al Table II. New alloantibodies formed following a delayed haemolytic transfusion reaction episode and number of times represented in our cohort. Antigen group New alloantibodies Number Rh Anti-E 1 Kell Anti-K 1 Duffy Anti-Fy a 1 Kidd Anti-Jk b 3 MNS Anti-M 1 Anti-S 4 Treatment The treatment and outcomes of the 23 episodes are outlined in Fig 1. Five of the 23 DHTRs at the severe end of the spectrum required treatment with immunosuppression including steroids and intravenous immunoglobulin (IVIg). Our prescribing practice has been to give methylprednisolone 500 mg IV over two consecutive days, and, in very severe cases, adding IVIg 1 g/kg/d delivered on two consecutive days. In one case refractory to steroids and IVIg, rituximab was given weekly for four weeks (375 mg/m 2 on week 1 and 100 mg on weeks 2, 3 and 4). Four of five patients treated with immunosuppression subsequently received further red cell transfusions in the same acute setting. An erythropoiesisstimulating agent (ESA) was given at high doses in the same episode refractory to first-line immunosuppression (a lifethreatening DHTR with profound anaemia and respiratory compromise), in an effort to optimize erythropoiesis and Hb. This patient also received IV iron (known not to have iron overload). At the mild end of the spectrum, four episodes were managed conservatively; they were not transfused subsequent to the trigger transfusion. These patients showed a slow Hb response, preceded by reticulocytosis. In 14 episodes with moderate severity, further RBC transfusion was given without immunosuppression; eight of these episodes were not recognized as DHTR events at the time, and most of the remaining six were only recognized after the RBC transfusion had been delivered. Nine of the 14 had escalation of haemolysis with this second phase transfusion, with subsequent Hb falling below pre-second transfusion Hb levels. Four did not display significant further haemolysis with the second transfusion. The final case unfortunately died 1 d after admission, and we were unable to assess the Hb trend. Outcomes The clinical severity of five DHTR episodes warranted critical care unit admission. These admissions were to allow safe transfusion of further units of blood in two cases and for organ support in three cases. The mean length of hospital stay (including critical care) was d, compared with d amongst our general SCD population. Further complications experienced by patients during DHTR included acute liver injury, intracranial haemorrhage, ACS, acute kidney injury and cholecystitis. Two of the 17 patients died from events that were precipitated during peak haemolysis; one of acute liver injury/multi-organ failure and the other, of intracranial haemorrhage. Recurrent DHTR events and subsequent RBC transfusions Twenty-three DHTR events occurred in 17 patients over the 54-year study period, four patients had recurrent events; two patients had two DHTR episodes and two patients had three DHTRs. In both patients with two DHTR episodes, an active decision was made to try to avoid future transfusion, due to the clinical severity of their last DHTR. One patient with three DHTRs has continued on a simple transfusion programme throughout the study period. Of the 13 patients with a single DHTR in the study period, seven patients have not received any further transfusions four have had active clinical decisions to avoid transfusions (patients 4, 6, 12 and 16), although none of these have been Fig 1. Treatment and outcomes of patients with DHTR. DHTR, delayed haemolytic transfusion reaction; RBC, red blood cell; Hb, haemoglobin; ICH, intracranial haemorrhage; ALI, acute liver injury. 750 ª 2015 John Wiley & Sons Ltd

6 DHTR in Sickle Cell Disease challenged with a clinically severe acute event); two (patients 5 and 13) have had no further precipitating events requiring transfusions and two patients (3 and 15) have died. The remaining five patients have since received transfusions: two (patients 1 and 10) are on exchange transfusion programmes for stroke prevention delivered uneventfully and three (patients 7, 11 and 17) have had further sporadic transfusions in acute settings, delivered without complication. In some of the patients in whom we want to avoid further transfusion, steady state Hb has been optimized medically. Three patients have had hydroxycarbamide initiated or doseincremented and two patients have had an ESA initiated in the steady state setting (both with normal baseline renal function). Discussion Our data suggest that DHTRs are an uncommon but potentially fatal complication of transfusion. While uncommon, DHTR events are not rare; over a 5-year period, in our cohort of 220 transfused patients with SCD, 17 (77%) had DHTRs. Historical reports suggested higher DHTR rates [11% by Vichinsky et al (1990), 16% by Diamond et al (1980)] where red cell matching for Rh and Kell antigens was not routine. More recently, Aygun et al (2002) reported a rate of 16% of clinically manifested DHTR, where transfused units were cross-matched compatible for Rh. Our prevalence falls between these DHTR rates, despite that Rh- and Kell-compatible blood is used in all patients. Our most significant finding is that the DHTR events were poorly recognized, with nearly half (478%) of our cases not identified at the time of the event. Unrecognized cases were mostly misdiagnosed as VOCs, probably because the clinical presentation of DHTR mimics that of an acute painful crisis, as others have reported (Diamond et al, 1980; Talano et al, 2003; de Montalembert et al, 2011). The unrecognized cases afford us an insight into the natural history of untreated DHTR episodes. In several milder cases, some patients even received further blood transfusions (often deemed inappropriate) uneventfully. At the other end of the spectrum, two cases resulted in death, with DHTR a contributing factor. As well as DHTR, the fatal cases experienced other complications (stroke, and acute liver injury/multi organ failure, respectively). Mortality in DHTR and other complications have also previously been reported (El-Husseini & Sabry, 2010). As in the literature, timing of DHTR in SCD appears to be very variable. Previous studies have described DHTRs occurring up to 15 d post-transfusion in SCD. We used a 21-d cut-off when seeking a DHTR to ensure that we captured as many episodes as possible. Five of 23 (217%) of DHTR episodes occurred more than 15 d post-transfusion. Clinical presentation of DHTR was uniform, with key symptoms of pain, fevers and haemoglobinuria, described by patients as a typical VOC episode, although the pain was frequently unusually severe. In terms of investigations, the first alerts to a DHTR event are a rapid decline in total Hb and sharp rise in LDH. In cases where Hb fraction was performed, the Hb decline was mirrored with a concomitant fall in HbA% (as donor cells are lysed) and a rise in HbS%. Synchronous, significantly elevated LDH (and notably higher than the patient s own typical VOC LDH level) was seen. Reticulocyte count was variable, with brisk and appropriate reticulocytosis in some patients, and reticulocytopenia in others. We noted that older patients have a higher risk of lack of appropriate reticulocyte response, and this absolute/ relative reticulocytopenia may prolong the DHTR episode. Delayed reticulocytosis in older patients may be related to the lower physiological reserve of the bone marrow from repeated bone marrow infarctions and reduced erythropoiesis. In SCD, immunohaematological findings of a DHTR do not always appear to be typical, with potentially negative antibody screens and DAT results even in the presence of significant haemolysis. In our series, there was a time delay between haemoglobin nadir and first DAT positivity/antibody screen positivity of 4 d. Some of this is due to delayed testing on the part of clinicians. However, three patients were DAT-negative during peak haemolysis, and subsequently became DAT-positive during recovery. This suggests that sequential immunohaematology testing may be useful to confirm the diagnosis and identify antibodies; this is crucial to document for future transfusion requirements. Positive DATs did not seem to predict clinical nor haemolytic severity, and so clinicians should be cautious in interpreting DAT results. However, for those patients with more than a weakly positive DAT, alloantibodies could be identified subsequently. Clearly, DAT positivity should prompt alloantibody screening, but we suggest that DAT positivity is not a prerequisite for DHTR diagnosis. Many alloimmunized patients do not develop DHTR, but our experience also confirms the converse, that DHTR events do not always have evidence of alloimmunization. In our own patients with SCD, we are now careful to label patients with a history of alloimmunization and DHTR separately; that is, to identify as separate diagnoses those who have red cell alloantibody formation and those with a clinically significant DHTR episode. The apparent risk factors for DHTR in our case series concur with those reported in the literature (Reisner et al, 1987; Schonewille et al, 2006): higher rates of DHTR in women compared to men (13 female vs. 4 male); previous DHTR (9 of the 23 episodes occurred in patients with known previous DHTR events); pre-existing red cell alloantibodies (11 of the 23 episodes had previous alloantibodies) and HbSS genotype. It is unclear if higher rates are seen in HbSS solely because HbSS patients are transfused more than patients with other genotypes. Murao and Viana (2005) reported that HbSC patients have a 28-fold higher risk of alloimmunization than HbSS patients, the risk being highest in female HbSC patients, but DHTRs were not included in the study. ª 2015 John Wiley & Sons Ltd 751

7 J. B. Vidler et al Given the lack of contemporaneous identification of many of these DHTR episodes, our series provide an insight into the natural history of an untreated DHTR event, which seems very variable. The mild cases demonstrate that it is possible to manage some DHTRs conservatively, albeit with close monitoring. Furthermore, second phase transfusions exacerbated the haemolysis in 10 of 16 cases where a second RBC transfusion is given without immunosuppression (most of these second phase transfusions were given in the context of an undiagnosed DHTR). This underlines the importance of identifying a DHTR and avoiding transfusion during a DHTR event if possible. For those with profound haemolysis and severe anaemia, our cases demonstrate success with immunosuppression. We adopted a policy of first line immunosuppression with steroids and IVIg, followed by repeated courses of immunosuppression (either more steroids if previously successful, or rituximab if refractory to first line therapy). The clinical context of the trigger transfusion appears to be a striking risk factor for the development of a DHTR. Although the majority of blood transfusions in our patients with SCD are given in the elective setting, most of the DHTRs occurred after a trigger transfusion given in the acute setting. This gave marked differences in DHTR rates for SCD patients for transfusions given in the acute setting (34%) versus elective setting (02%). Hence, while we should pay attention to DHTR events in patients on long-term transfusion programmes, which lead to significant cumulative red cell unit burden, our data suggests that we should be particularly vigilant in patients who are sporadically transfused in the acute setting. Recent studies have also suggested that the risk of alloimmunization in SCD is increased if the transfusion is delivered in an acute setting when the patient is in an inflammatory state (Yazer et al, 2009; Fasano et al, 2015). We acknowledge some limitations of this study. Diagnosis of a DHTR event was made retrospectively, based on clinical/laboratory data lookback and therefore we suspect that we may have underestimated true numbers, and data are not complete for some cases. In the same way that physicians are misled by DHTR symptoms, patients may also attribute a DHTR as a typical VOC rather than a transfusion reaction and may be less inclined to present than patients without SCD. While a prospective study into DHTRs in SCD (including optimal immunosuppressive treatment) is urgently warranted, current emphasis must be placed on both healthcare provider and patient awareness of DHTR events to capture true incidence. In conclusion, our retrospective analysis provides an insight into the frequency, recognition and natural history of DHTRs. A wide range of DHTR severity has been observed, from mildly symptomatic anaemia to severe complications. While a DHTR can be life-threatening in itself, it has also triggered life-threatening sickle-related complications including ACS, cholecystitis, multi-organ failure and stroke. DHTR should be always excluded in any recently transfused patients with SCD who present with acute pain; Hb and LDH trend, together with haemoglobinuria and immunohaematology results, help distinguish a DHTR from other acute SCDrelated events, notably VOCs. Acknowledgements We thank Claire Steward and Robin Swabey for help in preparation of the manuscript. Authorship JBV, KG and KA extracted data; JBV and KG analysed data; AM and SLT advised on analysis; JBV, KG, AM and SLT wrote the manuscript. All authors gave feedback on the manuscript. Conflict of interest The authors declare no competing financial interest. References Aygun, B., Padmanabhan, S., Paley, C. & Chandrasekaran, V. (2002) Clinical significance of RBC alloantibodies and autoantibodies in sickle cell patients who received transfusions. Transfusion, 42, British Committee for Standards in Haematology, Milkins, C., Berryman, J., Cantwell, C., Elliott, C., Haggas, R., Jones, J., Rowley, M., Williams, M. & Win, N. (2013) Guidelines for pre-transfusion compatibility procedures in blood transfusion laboratories. British Committee for Standards in Haematology. Transfusion Medicine (Oxford, England), 23, Davey, R.J., Gustafson, M. & Holland, P.V. (1980) Accelerated immune red cell destruction in the absence of serologically detectable alloantibodies. Transfusion, 20, Diamond, W.J., Brown, Jr, F.L., Bitterman, P., Klein, H.G., Davey, R.J. & Winslow, R.M. (1980) Delayed hemolytic transfusion reaction presenting as sickle-cell crisis. Annals of Internal Medicine, 93, Drasar, E., Igbineweka, N., Vasavda, N., Free, M., Awogbade, M., Allman, M., Mijovic, A. & Thein, S.L. (2011) Blood transfusion usage among adults with sickle cell disease a single institution experience over ten years. British Journal of Haematology, 152, El-Husseini, A. & Sabry, A. (2010) Fatal hyperhemolytic delayed transfusion reaction in sickle cell disease: a case report and literature review. American Journal of Emergency Medicine, 28, e1065 e1068. Fasano, R.M., Booth, G.S., Miles, M., Du, L., Koyama, T., Meier, E.R. & Luban, N.L. (2015) Red blood cell alloimmunization is influenced by recipient inflammatory state at time of transfusion in patients with sickle cell disease. British Journal of Haematology, 168, King, K.E., Shirey, R.S., Lankiewicz, M.W., Young- Ramsaran, J. & Ness, P.M. (1997) Delayed hemolytic transfusion reactions in sickle cell disease: simultaneous destruction of recipients red cells. Transfusion, 37, de Montalembert, M., Dumont, M.D., Heilbronner, C., Brousse, V., Charrara, O., Pellegrino, B., Piguet, C., Soussan, V. & Noizat-Pirenne, F. (2011) Delayed hemolytic transfusion reaction in children with sickle cell disease. Haematologica, 96, Murao, M. & Viana, M.B. (2005) Risk factors for alloimmunization by patients with sickle cell disease. Brazilian Journal of Medical and Biological Research, 38, ª 2015 John Wiley & Sons Ltd

8 DHTR in Sickle Cell Disease Petz, L.D., Calhoun, L., Shulman, I.A., Johnson, C. & Herron, R.M. (1997) The sickle cell hemolytic transfusion reaction syndrome. Transfusion, 37, Reisner, E.G., Kostyu, D.D., Phillips, G., Walker, C. & Dawson, D.V. (1987) Alloantibody responses in multiply transfused sickle cell patients. Tissue Antigens, 30, Schonewille, H., van de Watering, L.M., Loomans, D.S. & Brand, A. (2006) Red blood cell alloantibodies after transfusion: factors influencing incidence and specificity. Transfusion, 46, Talano, J.A., Hillery, C.A., Gottschall, J.L., Baylerian, D.M. & Scott, J.P. (2003) Delayed hemolytic transfusion reaction/hyperhemolysis syndrome in children with sickle cell disease. Pediatrics, 111, e661 e665. Vichinsky, E.P., Earles, A., Johnson, R.A., Hoag, M.S., Williams, A. & Lubin, B. (1990) Alloimmunization in sickle cell anemia and transfusion of racially unmatched blood. New England Journal of Medicine, 322, Yazdanbakhsh, K., Ware, R.E. & Noizat-Pirenne, F. (2012) Red blood cell alloimmunization in sickle cell disease: pathophysiology, risk factors, and transfusion management. Blood, 120, Yazer, M.H., Triulzi, D.J., Shaz, B., Kraus, T. & Zimring, J.C. (2009) Does a febrile reaction to platelets predispose recipients to red blood cell alloimmunization? Transfusion, 49, ª 2015 John Wiley & Sons Ltd 753