Spontaneous remissions in paroxysmal nocturnal hemoglobinuria: True remission or clonal evolution to malignancy?

Abstract Background Paroxysmal nocturnal hemoglobinuria (PNH) is a rare clonal hematopoietic disorder driven by a deficiency of glycosylphosphatidylinositol-anchored proteins, most notably the complement regulatory molecules CD55 and CD59. It is characterized by complement-mediated intravascular hemolysis (IVH), bone marrow (BM) failure and thrombosis. Eculizumab and ravulizumab are complement protein C5 inhibitors (C5i) that prevent IVH and its consequences, improving hemoglobin levels and reducing transfusion dependence. In a small subset of patients, the PNH clone may decline allowing discontinuation of treatment. However, in some cases this may herald clonal evolution including development of acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS). We evaluated spontaneous remissions in C5i treated patients with disease recurrence and clonal evolution following discontinuation of anti-complement therapy. Methods We retrospectively reviewed records of all 490 patients with PNH treated with anti-complement therapy (including 292 currently) at the Leeds PNH center (June 2002 - July 2025). Inclusion criteria comprised patients achieving spontaneous remission defined as their neutrophil and monocyte clone falling to below 10% who subsequently discontinued anti-complement treatment. All the patients who discontinued treatment were on C5i. Data collected included age at C5i initiation/discontinuation, corresponding treatment dates, treatment indications, thrombotic history and BM assessment at C5i cessation. Laboratory parameters comprised lactate dehydrogenase (LDH) and PNH clone size at initiation and discontinuation. Continuous variables were reported as mean ± standard deviation (±SD). Results Of 490 patients, 28 (5.7%) met the inclusion criteria. All initially received eculizumab; 9 patients (31.1%) later transitioned to ravulizumab. At eculizumab initiation, mean age was 49 years (range 17-87), mean granulocyte clone 70.2% (± 22) and mean LDH 1794iu/L (± 990.3). A history of thrombosis was present in 13 patients (46.4%). Mean C5i duration was 85.8 months (± 58.8). Indications for C5i included IVH in 18 patients (64.3%), IVH and thrombosis in 8 patients (28.6%), 1 pregnancy (3.6%) and 1 case of significant fatigue (3.6%). At C5i discontinuation, mean neutrophil clone size was 5.9% (± 3.4). On stopping treatment, 4 patients (4.3%) had undetectable clones. Five patients (17.9%) had disease recurrence and resumed C5i, with 1 patient restarting twice. Mean duration of follow up from C5i discontinuation was 52.4 months (± 49.4). Seven patients (25%) died during the period of follow up after spontaneous remission 3 from clonal evolution (MDS n=1, AML n=2), 1 from aplastic anemia, 1 from neutropenic sepsis, 1 from intracerebral hemorrhage and 1 from old age. Clonal evolution occurred in 5 patients (17.9%); with cytogenetic changes (n=1), MDS (n=3) and AML (n=1). BM examination at cessation was performed in 17 patients (60.7%); 4 patients (14.3%) had new cytogenetic abnormalities (monosomy 7; absent Chr X der(16) 1q12qter 46,X,add (Y)(q12)). Next-generation sequencing (NGS) showed mutations in 28.6% (8/28): isolated ASXL1 (n=2), BCOR (n=2), SETBP1VUS (n=1), combined ASXL1 and SRSF2 (n=2), and a complex panel (TET2, ASXL1, RUNX1, URAF1; n=1). Six patients had incomplete BM assessments due to missing cytogenetics (n=3) or NGS data (n=3). 2 patients (7.1%) did not undergo BM assessment due to their age and physical condition; 9 patients (32.1%) lacked a documented reason for not undergoing a BM examination. Discussion Spontaneous remission post-C5i therapy in PNH is rare and may reflect true hematologic recovery or underlying clonal evolution. In this cohort, 5 patients (17.9%) exhibited clonal evolution including MDS/AML transformation, and 5 (17.9%) had disease recurrence and reinitiated C5i. This data highlights 3 considerations: (1) monitoring the size of the PNH clone in patients to identify those in whom levels fall 10% during treatment with anti-complement therapy may allow safe discontinuation of costly anti-complement therapies; (2) spontaneous remission may not be equated with disease eradication, and continued surveillance post treatment cessation is essential to detect disease recurrence and guide treatment reinitiation; and (3) BM evaluation on discontinuing anti-complement therapy, with cytogenetic analysis and NGS, is recommended to help distinguish remission from clonal evolution.