Paroxysmal nocturnal hemoglobinuria is not a cause of anemia in patients with myelofibrosis

Abstract

Paroxysmal nocturnal hemoglobinuria (PNH) is a disease characterized by the expansion of a non-malignant hematopoietic stem cell clone harboring a somatic mutation in the PIGA gene, which is required for biosynthesis of the glycophosphatidylinositol (GPI) moiety that anchors proteins to cell membranes [1,2]. Erythrocytes and granulocytes derived from this mutant clone are defi cient in several GPI-anchored membrane proteins (APs), two of which are necessary for activation of the complement system (CD55 [decay-accelerating factor (DAF)] and CD59 [membrane inhibitor of reactive lysis]) [3]. Symptomatic patients generally present with intravascular hemolysis, hemoglobinuria and thrombophilia [4]. PNH has also been shown to be closely associated with other bone-marrow disorders such as aplastic anemia (AA) and myelodysplastic syndromes (MDS) [5 – 8]. Approximately 50 – 60% of patients with AA and 15 – 20% of patients with low-risk MDS that clinicopathologically mimics AA have a detectable population of GPI-AP defi cient erythrocytes and granulocytes [4 – 7]. However, in 90% of these cases, the PNH clone represents less than 25% of the total neutrophil population [8], with no clinical evidence of hemolysis, and therefore does not require treatment. More importantly, the subclinical presence of the PNH clone among patients with low-risk MDS and AA has been associated with a benign disease course and predicts a favorable response to immunosuppressive therapy [6]. A few case reports of PNH associated with myelofi brosis (MF) have been reported [9,10]. However, the actual incidence of PNH among patients with MF is unknown, despite the fact that patients with MF commonly have high lactate dehydrogenase, low haptoglobin and sometimes elevated indirect bilirubin, all possibly indicative of intravascular hemolysis. We retrospectively searched clinical records of patients with MF seen at Th e University of Texas M. D. Anderson Cancer Center over the last 5 years, in whom treating physicians ordered testing for PNH to determine whether PNH was a reason for signifi cant anemia in the patients. Th e study was based on a retrospective chart review protocol approved by the Institutional Review Board at M. D. Anderson. Testing for PNH, therefore, was solely based on a clinical suspicion of a treating physician; i.e. testing was done as a part of standard practice based on the involved physician ’ s medical judgement. Although more than 1000 patients with MF were seen during that period of time, and the great majority had clinically relevant anemia, only 62 patients with MF and signifi cant anemia (hemoglobin 10 g/dL) were identifi ed in whom testing for PNH was done: 41 with primary MF, 13 with MF secondary to polycythemia vera, and eight with MF secondary to essential thrombocytosis. Both peripheral blood red cells and granulocytes were the primary cell populations used to identify cells with the PNH phenotype, defi cient in GPI-APs (test sensitivity of 5%) by four-color fl ow cytometry immunophenotypic analysis [5]. Th e fl ow assay we used was a screening test to demonstrate clones at a level associated with clinical PNH syndrome, and was not sensitive enough to detect small subclinical clones ( 1%) now possible with fl uorescent aerolysin (FLAER)-based assay. Interestingly, none of the 62 patients evaluated were found to have a concomitant PNH clone. In trying to understand what might have prompted treating physicians to order testing for PNH in these patients, we identifi ed that all of them had signifi cantly elevated lactate dehydrogenase, all but fi ve had signifi cantly lower than normal haptoglobin, but only four had a major thrombotic event (e.g. portal vein or renal artery thrombosis). In another study of 136 patients with hematologic malignancies, including fi ve with primary MF, none had peripheral blood granulocytes harboring the PNH phenotype [5]. Although the number of patients with MF in that study was very small, the result correlates with our fi ndings. In conclusion, our study suggests that the presence of granulocytes with the PNH phenotype in general is not present in patients with MF and signifi cant anemia. PNH is likely an extremely rare complication in patients with MF, and it is appropriate to look for PNH only in those with unexplained