Insights Into The Natural History Of Paroxysmal Nocturnal Hemoglobinuria (PNH): Analysis Of The Presenting Clinical, Haematological and Flow Cytometric Features Of 705 Patients Leads To Improved Classification and Prediction Of Clinical Course

In past 22 years, we have identified using flow cytometry 705 patients with detectable PNH (GPI deficient) populations of granulocytes, monocytes and red cells in the peripheral blood in samples sent for diagnosis. We undertook an analysis of presenting clinical features, blood count data and PNH clone sizes in order to better understand the natural history and provide a more objective classification of disease. Based on serial flow cytometry measurements of PNH clone sizes, we also studied disease stability, frequency of recovery and progression with an aim to guiding future management of individual patients.

Clinical classification of patients at presentation was as follows; aplastic anemia (58%), hemolytic anemia (36.1%); myelodysplasia (2.5%); thrombosis (2.4%); hemolysis & thrombosis (0.6%), myeloproliferative neoplasm (0.3%); Fanconi anemia (0.1%). Median age at presentation was 45 years (observed range 0.5 – 90 years) and the Male:Female ratio was 1.05. Descriptive statistical analysis of presenting blood count data revealed novel gender related features not previously described in PNH. At presentation, pancytopenia was found in 61% of male and 47% of female patients; a normal blood count was present in only 0.3% of males and 4% of females. A combined low red blood cell count (RBC) and white cell count (WBC) was the most frequent bicytopenia affecting 19% males and 22% females. Leucopenia as a sole abnormality did not occur in males and was present in<1% of females; Leucocytosis was present in<1% of cases. Platelet counts were low in 79% of patients and normal in 20.3%. Low absolute numbers of RBC (mean 3.002 x 10¹²/L; IQR 2.52-3.47) were present in 97.6% of males and 87.6% of females (mean 3.009 x 10¹²/L; IQR 2.60 - 3.42). Similar findings were obtained for hemoglobin (Hb) values: males (mean 100.5 g/L; IQR 85-115): females (mean: 100.7 g/L; IQR 88-116). Unexpectedly, there were no statistically significant differences between male and female values for Hb or RBC counts. In a more detailed comparison, significant differences between aplastic PNH patients (n = 394) and hemolytic PNH patients (n = 245) were identified. Aplastic patients had significantly lower WBC, neutrophil, monocyte and platelet counts, and MCV (all P<0.01), though lymphocyte count was higher than in the hemolytic cohort. No statistical differences could be demonstrated for Hb or RBC counts between disease types or between males and females, confirming earlier results. For PNH clone sizes at presentation, aplastic patients showed a median red cell PNH clone size of 0.34% (IQR 0.05 - 2.14) and a median granulocyte PNH clone size of 2.58% (IQR 0.42 - 12.48). In contrast, hemolytic patients had a median red cell PNH clone size of 32.35% (IQR 22.04 - 51.67) and a median granulocyte PNH clone size of 90.61% (IQR 77.07 - 97.41). Follow up flow cytometry studies of 154 aplastic and hemolytic patients over a minimum period of 18 months (mean follow up 81 months, range 20-216 months), provided important insights into biology of the disease. Firstly, aplastic patients with granulocyte PNH clones of<1% at presentation, did not evolve to hemolytic PNH. Any increase in PNH clone sizes and progression to hemolytic disease within the aplastic cohort was associated with granulocyte clones of >5% at presentation, though this occurred in only 10/154 (6.5%) cases. For patients presenting with hemolytic disease, PNH granulocyte clones continued to increase in size in 44% of cases most likely reflecting a combination of on going selection in favour of the PNH clone and prompt diagnosis. In the 38 patients that presented with >95% granulocyte PNH clones, 92% remained stable over time (mean follow up 80 months (many on Eculizumab therapy)) with only 8% showing a gradual fall in clone size.

The study shows that pancytopenia is a consistent feature of hemolytic and aplastic PNH patients. The degree of anemia is the same in both major groups of patients, but appears to be less severe for females. Not only are PNH clone sizes larger in hemolytic patients, but they also show higher platelet and leucocyte counts compared to aplastic patients, most likely reflecting a more active bone marrow. The data support the model that bone marrow failure is the primary underlying pathology in >95% of PNH patients and that sub classification on the basis of degree of aplasia, hemolysis (with or without thrombosis) and PNH clone size at presentation can be a powerful predictor of clinical course.