Page 207 - 2020_11-Haematologica-web

of +3.8 g/L for each pRBC unit transfused, and ongoing hemolysis at a rate of -4.4 g/L for each passing day (the fixed effects of the model). However, after anti- C5 antibody infusion, Hb levels gradually increased, with a basal Hb concentration of 46.2 g/L: the effect of each transfusion was an increase of 1.58 g/L for each pRBC unit transfused and an increase in Hb levels of 0.94 g/L.
Case Reports
Figure 1. Best mixed-effects model for total hemoglobin and lactate dehydrogenase during delayed hemolytic transfusion reaction, before and after anti-C5 antibody infusion. Hemoglobin (Hb) levels are predicted for a “theoretical” patient receiving anti-C5 antibody infusions on days 0 and 7 and no packed red blood cell (pRBC) transfusions. Before anti-C5 antibody infusion, basal total Hb concentration in this model was 51.0 g/L (the intercept of the model), with an increase
organ failure and, in some cases, death.
These pathophysiological findings suggest that
inhibitors of complement activation may be useful for treating DHTR with hyperhemolysis. Eculizumab is a monoclonal anti-C5 antibody that inhibits the cleavage of C5 into C5a by the C5 convertase, thereby preventing the late stages of the complement cascade. Anti-C5 ther- apy has been offered to several SCD patients for DHTR treatment at our SCD referral center in France since 2013.1,7 Other teams have also treated DHTR in patients with and without detectable allo- or auto-antibody for- mation with Eculizumab, with promising results.8-11 The American Society of Hematology (ASH) guidelines include a conditional recommendation for the use of anti- C5 antibodies in patients with SCD presenting DHTR and ongoing hyperhemolysis, based on currently very low levels of certainty.12
This retrospective study focuses on the biological and clinical findings and the effects of anti-C5 therapy on DHTR, for patients treated between 2013 and 2019 who experienced particularly severe DHTR.
DHTR was diagnosed1,2 on the basis of VOC signs occurring 5-20 days after pRBC transfusion, with no other identifiable cause of intravascular hemolysis, in association with at least one of the following signs:
- rapid decrease in, or unexpectedly low, hemoglobin A (HbA) concentration (the diagnostic nomogram for DHTR diagnosis was used),2
- hemoglobinuria, as revealed by dark urine,
- positive direct antiglobulin test (DAT) results or new antibody formation.
The criteria for the use of anti-C5 therapy was based either on the existence, at the time anti-C5 infusion was decided, of one or more organs with dysfunction and/or very low total Hb concentration (< 50 g/L), and/or a rap- idly worsening clinical state.7
Data were collected retrospectively from patient
records. The clinical and biological findings available at transfusion, at the time of DHTR diagnosis and during follow-up were collected. We also noted patient sex, age, history of DHTR, pRBC transfusion, and antibody screens. We recorded the number of pRBC units and the indication of the transfusion(s) occurring within a time- frame compatible with DHTR (some patients had received pRBC on several occasions during the 5-20 days preceding DHTR). Clinical (hemoglobinuria, pain and VOC signs, organ failure) and biological (hemoglobin concentration, reticulocyte count, LDH, total bilirubin) findings at DHTR diagnosis were collected. The first clin- ical signs compatible with DHTR were noted, particular- ly pain indicative of VOC recurrence, and hemoglobin- uria indicative of intravascular hemolysis. We collected follow-up data for biological tests, intensive care unit admission and discharge, organ failure and treatments. This study was performed in accordance with the Declaration of Helsinki.
We used R3.6.1 and lme413 for a linear mixed-effects model analysis of the relationship between Hb and lac- tate dehydrogenase (LDH) levels and anti-C5 treatment. Hb and LDH levels were modeled before and after treat- ment. An initial blind statistical analysis was performed, and several models were then proposed, with days and pRBC transfusions as fixed effects, and different combi- nations of random effects for subjects, days and pRBC transfusions. We obtained P-values for likelihood ratio tests of the full model with random effects against the model without additional terms, which we used to select the best model.
Eighteen SCD patients received anti-C5 treatment for DHTR with hyperhemolysis. All patients had signs of VOC 5-20 days after pRBC transfusion, with low HbA concentrations (<10 g/L) in five patients, a rapid decrease in HbA concentration in 10 patients (estimated by the nomogram2 as a high (n=4) or intermediate (n=6) risk of
haematologica | 2020; 105(11)
2695