Hyperhaemolysis syndrome (HHS) is a potentially fatal complication of transfusion that is likely underrecognized in sickle cell disease (SCD).1, 2 The American Society of Hematology defines HHS as a delayed haemolytic transfusion reaction associated with a rapid decline in haemoglobin to below the pretransfusion level.3 First-line management typically involves immunomodulatory therapy with corticosteroids and intravenous immunoglobulin (IVIg). Terminal complement inhibitors, such as eculizumab, block the cleavage of complement C5 into C5a and C5b, preventing formation of the membrane attack complex (sC5b-9). The alternative complement pathway is thought to serve as a critical driver for refractory HHS.4 Thus, patients with refractory HHS may benefit from treatment with eculizumab; however, a major hurdle is a lack of data to inform a dosing strategy. We hypothesize that manifestations of HHS may differ depending on the complement-fixing efficiency of the triggering Red blood cell (RBC) antibody. An antibody may activate complement based on features such as immunoglobulin subtype, and also density and spatial proximity of antigen binding sites on the RBC surface. Shorter distances between two surface-bound Immunoglobulin G (IgG) antibodies are thought to increase the likelihood of forming a doublet capable of binding complement C1.5 An RBC-bound antibody fixing C1 could activate the classical complement pathway, thereby feeding into the amplification loop of the alternative pathway.6 Furthermore, several RBC antigens are expressed on non-erythroid tissues, raising the possibility that complement deposition on tissue surfaces may result in end organ damage. Here, we describe two patients with refractory HHS following RBC transfusion who were treated with eculizumab. We highlight the challenges in antibody identification and propose that refractory HHS may result from antibodies that efficiently activate complement. We monitored changes in sC5b-9 and CH50 to assess adequacy of complement inhibition (Figure 1; Table S2). In the first case, a 46-year-old woman with HbSS disease presented to our hospital with acute pain. Five days prior to this admission (denoted ‘Day −5’), she had received two units of Packed red blood cells at another hospital. Her blood group was B+ and antibody screen at our hospital showed anti-E and a cold-reactive anti-M (Table S1a). Her initial hemoglobin (Hb) was 10.9 g/dL; however, over the next 4 days, her Hb rapidly decreased (Figure 1), reaching a nadir of 3.7 g/dL on day 5. The absolute reticulocyte count was 81 × 109 cells/L, and lactate dehydrogenase (LDH) was greater than 3000 units/L by that time. The direct antiglobulin test (DAT) was negative multiple times; however, Hb electrophoresis showed disproportionately low HbA% despite receiving transfusion 9 days prior. Repeat antibody screen at the time of the Hb nadir showed a single cell positive, characterized as a new antibody of unknown specificity (AUS), in addition to the anti-E and anti-M. The anti-M now demonstrated a component reactive at 37°C. RBC antigen genotyping demonstrated no uncommon variants. Meanwhile, the patient became unresponsive and hypotensive, and she was urgently transfused two units of red blood cells that were E-negative, M-negative and crossmatch-compatible. The patient was started on high-dose steroids, IVIg and epoetin-alpha. The patient's Hb again rapidly declined. Repeat serological testing at this time was notable for a new anti-S on the antibody screen, with DAT positive for IgG2 only (negative for complement) and anti-S in the eluate. Segments from the two transfused units were found to be S antigen positive. On hospital day 6, 1 day after starting steroids and IVIg, the patient's sC5b-9 level was 371 ng/mL (normal 60 U/mL. Due to concern for further decompensation, the patient was started on eculizumab 900 mg intravenously. After this first dose, the patient's reticulocyte count improved, LDH decreased and Hb stabilized. CH50 on day 8–2 days after the first dose of eculizumab—was undetectable. Steroids were tapered. A second dose of eculizumab 900 mg was given on day 12. The patient was discharged on day 14 with an Hb of 6.7 g/dL, absolute reticulocyte count of 308 × 109 cells/L and LDH 769 units/L. The patient returned to the hospital with recurrent pain, 19 days after her initial admission and 24 days after the trigger transfusion. Hb nadir on this second admission was 5.2 g/dL. She was managed without transfusions: the steroid taper was continued; hydroxyurea (hydroxycarbamide) was restarted with opposing erythropoietin- stimulating agent. At discharge from the second hospitalization, total Hb was 9.2 g/dL with undetectable HbA. In a second case, a 62-year-old woman with HbSC disease and baseline haemoglobin of 11 g/dL was transfused with two units of RBCs due to acute anaemia of 8.1 g/dL from gastroenteritis triggering a pain crisis. She was blood type O+ with a history of antibodies against Jkb, Bga, C, Jsa, K, s and V as well as an antibody of undetermined specificity and a warm autoantibody noted over 10 years prior. All antibody screens in the last 10 years had been negative. The two units transfused were crossmatch compatible and negative for C, K and s. The Jkb status of the units was not assessed, as RBC antigen genotyping had shown that the patient expressed Jkb. She was discharged from the hospital in improved condition with total Hb 10.1 g/dL. Ten days after transfusion, she presented with worsened pain and confusion. She was found to have an Hb of 8.7 g/dL, with a corresponding absolute reticulocyte count of 231 × 109 cells/L. Lactate dehydrogenase was 1132 units/L (Figure 1). She had acute kidney injury (Cr 2.26 mg/dL, baseline 0.4 mg/dL). Haemoglobin electrophoresis showed an absence of HbA. Initial blood bank work-up was notable for what seemed to be an anti-Jkb on the antibody screen; however, on repeat testing, this reactivity disappeared. Multiple DATs as well as subsequent antibody screens were negative (Table S1b). sC5b-9 was elevated at 469 ng/mL (reference range <244 ng/mL). The patient's haemoglobin decreased to 5.6 g/dL on day 14. Due to concerns that steroids would precipitate a more severe pain crisis and that IVIg would further worsen her creatinine, the patient was treated with eculizumab 900 mg on day 15. The patient experienced improvement in her haemoglobin to 7.1 g/dL on day 21–6 days after the first dose of eculizumab—with improvement in LDH to 400 units/L. Creatinine returned to baseline. However, by day 24, sC5b-9 had increased to 338 ng/mL with a corresponding drop in Hb to 6.6 g/dL. Therefore, a second dose of eculizumab 900 mg was given on day 25 with improvement in Hb to 8.2 g/dL and LDH to 367 units/L by day 29. Deep sequencing of the patient's Kidd blood group locus showed that she expressed conventional Jkb. The transient anti-Jkb reactivity was ultimately thought to be from an autoantibody. Her course was further complicated by choledocholithiasis, deep vein thrombosis and osteonecrosis. These cases illustrate the limitations of our current management algorithms for HHS. They serve as a call for rigorous antibody identification by the hospital transfusion service as well as the refinement of diagnostic and prognostic biomarkers for dosing complement inhibitors. The optimal eculizumab dosing regimen in HHS has not been established. Reported cases have generally used induction regimens approved for paroxysmal nocturnal haemoglobinuria (PNH) or atypical haemolytic uremic syndrome (aHUS), despite significant differences in dosing depending on the level of complement inhibition required.7, 8 In Case 1, HHS developed in conjunction with elevated sC5b-9 and an anti-S antibody formation. sC5b-9 levels improved following eculizumab. Anti-S antibodies are known to sensitize RBCs to complement in about one-third of cases.5 A major challenge in HHS is that up to one-third of HHS cases lack an identifiable alloantibody.9, 10 In Case 2, no antibody trigger was found conclusively. However, a history of anti-Jkb in the patient's medical record, plus questionable anti-Jkb reactivity noted at the time of HHS, is concerning. Possible mechanisms to explain the observed anti-Jkb reactivity include immune stimulation from the transfusion that triggered formation of a transient autoanti-Jkb. IgG Kidd antibodies are generally felt to be capable of fixing complement, leading to both intravascular and extravascular haemolysis.11 Interestingly, Kidd antigens are expressed on both RBCs and in the kidney, and renal injury was a prominent feature of Case 2. Other differences between the two cases bear mention. Hyperhaemolysis is thought to involve immunologic pathways outside of complement activation, including macrophage activation and phagocytosis.12 IVIg and steroids, as administered in Case 1 with eculizumab, are thought to suppress macrophage activation and thus interfere with extravascular haemolysis. However, in SCD, steroids are associated with pain crises.13, 14 Certain preparations of IVIg may worsen renal injury.15 The risks and benefits of treatment with steroids and IVIg may differ depending on whether the offending antibody predominantly triggers extravascular haemolysis through macrophage activation, or whether the antibody predominantly triggers complement. The difference in SCD genotype between the two cases also likely influenced LDH trends and complement biomarker values, as much as the offending antibody. See Supporting Information S3 for further discussion of the potential role of SCD genotype. Our case series demonstrates that a ‘one size fits all’ approach may not work for the management of HHS. While the mechanisms of complement activation by alloanti-S in Case 1 and possible autoanti-Jkb are hypothetical, we argue for a better mechanistic understanding of how specific anti-RBC antibodies activate complement in vivo. To do this, haematologists must work in conjunction with the blood bank to identify allo- and autoantibodies as well as develop robust, accessible biomarkers for HHS treatment response. Mamie M. Thant conceptualized and wrote the manuscript. Jahnavi Gollamudi conceptualized the manuscript, created the figures and edited the manuscript. Stefanie W. Benoit conceptualized and edited the manuscript. Satheesh Chonat reviewed and edited the manuscript. Jose A. Cancelas reviewed and edited the manuscript. We would like to acknowledge Sara Staker MLS ASCP and Rachel Moore MLS ASCP for their assistance with blood bank record review. There was no funding for this work. The authors report no relevant conflicts of interest. Informed consent could not be obtained from the patient in Case 1, as they have left the country and cannot be contacted despite reasonable efforts. This case report has been sufficiently anonymized to protect the patient's identity, and the authors believe that publication is justified on grounds of public interest and educational value. Informed consent for the publication of this case report, including clinical details and lab values, was provided by the patient in Case 2. The data that support the findings of this study are available from the corresponding author upon reasonable request. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Thant et al. (Thu,) studied this question.