ANALYSIS OF BASELINE CHARACTERISTICS, DISEASE BURDEN AND LONG-TERM FOLLOW-UP OF 167 PATIENTS WITH BRAZILIAN PAROXYSMAL NOCTURNAL HEMOGLOBINURIA - ANOTHER NATURAL HISTORY

Abstract

Paroxysmal nocturnal hemoglobinuria (PNH) may occur as classical hemolytic disease or a small PNH clone found in a patient with bone marrow failure. To describe clinical features and long-term follow-up of 167 patients with PNH and demonstrate differences between disease categories. Multiparametric flow cytometry performed on 1025 patients referred from Jan/2000 to Dec/2019 found 167 (16.3%) confirmed PNH clone, 87M/70F. Clinical characteristics at first visit, and laboratory results, incidence of thrombosis, treatment and outcome during follow-up visits were considered for statistical analysis. Most patients (89.2%) had hypocellular bone marrow at diagnosis; 55(32.9%) developed hemoglobinuria, and 22(13.2%) developed thrombosis during monitoring. Mean age at aplasia diagnosis 28.4 years (range 5.6 - 71.2), and mean age at PNH clone detection 31.0 years (7.2 - 71.3 years). Clonal evolution occurred in 77 (46.1%) patients at a median of 4 years (range 6-281 months) after the onset of cytopenia. The cohort was divided into 15 classic PNH, 55 hemolytic PNH with bone marrow hypoplasia (PNH/AA), and 97 definitive bone marrow disease with minor PNH clones - subclinical PNH (sc-PNH). Subclinical patients had lower erythrocyte (2.0% vs 24.0% vs 57.8%) and granulocyte PNH clones (11.7% vs 58.8% vs 81.2%) than PNH/AA and classic PNH, respectively. LDH, reticulocytes, absolute neutrophils and bone marrow cellularity were all lower in sc-PNH than in the hemolytic group. Pearson's analysis showed a statistically significant correlation (p < 0.001) between PNH granulocyte clone and RBC clone (0.697), LDH (0.602), absolute neutrophils (0.310), platelet count (0.430), absolute reticulocytes (0.546), total bilirubin (0.330), and bone marrow biopsy (0.435). PNH granulocyte clone greater than 50% was found in 16/22 thrombotic patients. Mean granulocyte clone was 89.1% (3.4%-99.2%) versus 12.3%(0.1%-34.0%), p < 0.001, and erythrocytes 28.4% (3.4-99.8) versus 0.7% (0.1-14.0) in thrombotic versus the remaining 145 patients. Treatment was highly heterogeneous during this 20-year follow-up: immunosuppressive therapy with CSA and/or thymoglobulin was given to 72 patients in the subclinical group (73.1%) and 45 (64.3%) in the hemolytic group; 52 patients received HSCT (31.1%) of which 40 in sc-PNH and 11 in PNH/AA; a total of 19 patients received Eculizumab after 2008 and 3 patients never received treatment. Incidentally, there were 17 thrombotic events in hemolytic patients (38.6%) diagnosed by 2007, compared to only five events (19.2%) in patients diagnosed after 2008. The average follow-up period was 74 months (1 - 330 months), with 31 deaths in total (80.4% OS). Sepsis was the leading cause of sc-PNH mortality (16/18, 88.8%), and thrombosis in both hemolytic groups (11/13, 84.6%). Log rank test shows similar survival when comparing patients with diagnosis until and after 2007 (p = 0.04). this study suggests clonal evolution in the long-term follow-up of aplastic patients and confirms the observation that high levels of PNH and LDH clones are associated with life-threatening hemolysis and thrombosis, while small PNH clones and young age are associated with the subclinical form of the disease. Although without statistical significance, the cumulative incidence of thrombosis in the group diagnosed after 2008 appears to be lower than the previous period, perhaps due to the routine use of Eculizumab.