Direct Antiglobulin Test, Ferritin and GPI-Negative Cell Populations in Patients with Paroxysmal Nocturnal Hemoglobinuria (PNH) during Eculizumab Therapy: Extravascular Hemolysis Unmasked by Efficient Eculizumab-Mediated Inhibition of Intravascular Hemolysis.

Introduction: Eculizumab is a humanized monoclonal antibody inhibiting the cleavage of complement factor C5 and preventing terminal complement activation. Therapy with eculizumab leads to a highly significant reduction of complement-mediated intravascular hemolysis and related morbidities including thromboembolic events in Paroxysmal Nocturnal Hemoglobinuria (PNH). It is a new targeted, disease-modifying treatment of PNH. In a placebo-controlled, double-blind randomized study of eculizumab more than 50% of patients (pts) achieved transfusion independence with the remaining pts showing a 44% reduction in the number of red blood cell (RBC) units transfused. In this study we analyzed the proportion of cells with deficiency of GPI-anchored proteins, i.e. PNH cell populations, during eculizumab therapy. Furthermore, we examined a potential mechanism to explain residual hemolysis in some pts despite complete blockade of the terminal complement cascade.

Methods: We performed flow cytometric analysis of GPI-anchored proteins, measured levels of hemolysis (LDH), direct bilirubin, hemoglobin and ferritin, and carried out a direct antiglobulin test in samples from 15 (7 female, 8 male) PNH pts during eculizumab therapy. To identify PNH cells, flow cytometric analyses of CD58/CD59 on reticulocytes and erythrocytes, CD66b/CD24/CD16 on granulocytes as well as CD14/CD48 on monocytes were performed. For comparison, the direct antiglobulin test was also performed in 22 (11 female, 11 male) PNH pts without eculizumab therapy.

Results: Flow cytometric results of the 15 PNH pts before eculizumab (baseline) and at the time of latest follow-up during eculizumab were compared: GPI-deficient erythrocyte population doubled (29.3% +/−19.3% vs. 56.9% +/−24.0%). The proportion of GPI-deficient reticulocytes (79.1% +/−6.8% vs. 84.6% +/−8.7%), granulocytes (87.5% +/−9.7% vs. 85.5% +/−12.4%) and monocytes (79.0% +/−20.6% versus 73.6%+/−16.2%) did not change significantly. We observed a mean decrease in LDH from 2802 U/l +/− 1015 U/l to 240 U/l +/− 62 U/l; while total bilirubin was stable (47.2 +/− 33.3 μmol/l versus 45.0 +/− 27.8 μmol/l), direct bilirubin showed a threefold increase (3.2 +/− 3.2 μmol/l versus 9.6 +/− 2.4 μmol/l). An increase of ferritin levels was also observed (270 +/− 609 ng/ml versus 651 +/− 652 ng/ml). We performed a direct antiglobulin test in the control group of 22 pts not receiving eculizumab therapy and in 12 pts that received eculizumab. 10 of the 12 pts were tested before and during eculizumab therapy. Only three of all 34 pt who had not received eculizumab at time of antiglobulin test showed a positive antiglobulin test with polyspecific antiglobulin serum, two of them with monospecific anti-IgG and one of them was positive with monospecific anti-C3d. In contrast, all but one of the 12 pts during eculizumab therapy reacted positive with polyspecific antiglobulin serum and monospecific anti-IgG and anti-C3d; in addition, one pt was positive with anti-C3c. The only patient with a negative antiglobulin test during eculizumab therapy was treated for PNH-related thromboembolic events without having transfusion dependent hemolytic anemia. Two of the 12 pts had anti-erythrocytic allo-antibodies (anti-D, -C, -E, -Lua, and -Kpa) before initiation of eculizumab therapy and one of them had additional warm auto-antibodies.

Conclusions: 1) A positive direct antiglobulin test (due to C3d loading of PNH RBC) seems to be a consistent finding during eculizumab therapy, which has therefore to be taken into account as a new differential diagnosis of a positive direct antiglobulin test. 2) Eculizumab does not block the early stages of complement activation and therefore C3 cleavage products may occur and bind to CD55/CD59-deficient RBC. The inhibition of terminal membrane attack complex formation by eculizumab allows the survival of these cells which otherwise would undergo a rapid intravascular hemolysis due to C5b-9 formation. 3) The endogenous population of GPI–anchored protein deficient RBC increases during eculizumab therapy. 4) Our data are preliminary and need to be confirmed in larger patient groups: For this purpose, we recommend the monitoring of signs of extravascular hemolysis and ferritin during eculizumab therapy.