Paroxysmal nocturnal hemoglobinuria (PNH) is a rare, potentially life-threatening, acquired clonal non-neoplastic hematopoietic stem cell disorder. PNH is characterized by complement-mediated chronic hemolytic anemia, acquired thrombophilia and bone marrow deficiency (1, 2).

The primary genetic defect in PNH comprises somatic mutations in the X-linked PIGA gene (2). PIGA encodes a protein that is involved in the biosynthesis of the glycosyl-phosphatidyl-inositol (GPI), a glycolipid that acts as an anchor that mediates attachment of over 150 proteins to cell membrane (GPI-anchored proteins) (2). Mutations in PIGA result in a significant reduction or lack of synthesis of the GPI anchor, with consequent partial or complete deficiency of GPI-anchored cell surface proteins such as the complement-inhibiting proteins, CD55 “(DAF - decay accelerating factor)” and CD59 (MIRL – membrane inhibitor of reactive lysis) (2, 3). CD55 by destabilizing the complement C3 convertase complex inhibits complement activation beyond C3 while CD59 exerts a direct inhibitory effect on the membrane attack complex (MAC), the final product of complement activation (2). Therefore, due to CD55 and CD59-deficient red cells, complement activation results in intravascular hemolysis and anemia (2).

In addition, PNH has been characterized as “the most vicious acquired thrombophilic state known in medicine” underscoring the high risk of thrombosis and the associated morbidity and mortality in untreated patients (4). The mechanisms of thrombosis in PNH are multifactorial and seem to involve complex interactions between the coagulation and complement cascades (5). It has been postulated that complement-mediated intravascular hemolysis and PNH platelet activation are also key pro-thrombotic factors (5). Other factors favoring a prothrombotic state include lack of nitric oxide bioavailability because of intravascular hemolysis, pro-inflammatory mediators, a dysfunctional fibrinolytic system, and damage to the endothelium (5). Patients diagnosed with PNH are at an increased risk for developing thrombosis, and clinical thrombotic episodes can occur in up to 45% of patients across all vascular sites (5, 6). Most thrombotic events are venous (85%) while arterial thrombosis is seen in 15% of cases and often affects the cerebral and coronary arteries (5, 6). Thrombosis is the leading cause of death in PNH, accounting for 40-67% of PNH deaths with known causes (5).

Thrombotic events (TE) necessitate initiation of anticoagulation treatment. However, anticoagulants alone are insufficient to prevent new TEs and therefore disease-specific therapy with complement inhibitors is indicated (4, 5, 7). Eculizumab, the first-in-class anti-C5 humanized antibody, has shown to decrease hemolysis and the frequency of serious and life-threatening morbidities such as thrombosis in PNH patients (8, 9).

In this study, we provide real-world data on TEs occurred in a cohort of 41 patients with PNH who were identified and treated over the past two decades.

We evaluated 41 patients diagnosed with PNH at 14 different hematology centers in Greece between 2002 and 2022. Only patients with adequate data were included. Their medical records were reviewed, and demographic and clinical data were collected. Symptoms and medical history, lab results at the time of diagnosis, and the size and type of PNH clone were all considered. The sites of thrombosis, the time of onset in relation to the course of the disease, any potential contributory factors, and the type and duration of anticoagulant therapy (primary or secondary thromboprophylaxis) were all evaluated. Blood transfusions requirements and other information on therapeutic interventions, whether disease-specific or not, were also assessed. Statistical analysis of the data was performed using the IBM SPSS 26 statistical package program. The chi-square test was used to determinate the relationship between groups and categorical variables. For the comparisons of numeric variables, independent samples t-tests were conducted. Differences were considered statistically significant when p-value was <0.05.

All patients with PNH are at risk for thrombosis, and up to 45% of patients can experience clinical thrombotic episodes at both typical and atypical sites (5, 6). In the present study, 39% (16/41) of PNH patients experienced thrombosis during the course of their disease, a frequency comparable to that reported in the medical literature (5, 7, 9, 11).

In 17.1% (7/41) of the patients, the thrombotic episode was the primary cause of the PNH investigation. It is recognized that thrombosis in patients with PNH does not “respect” any blood vessel and can occur at any site (5). Intra-abdominal and cerebral veins are frequently affected by thrombosis for unknown reasons (5). In accordance with earlier findings, 29.3% (12/41) of our patients had intra-abdominal thrombosis.

In addition, it is anticipated that one-fifth of PNH patients develop thrombosis at numerous sites (5). Seven people (17.1%) in our study presented with thrombosis at several sites.

Unusual venous thrombotic events, intra-abdominal or Budd-Chiari syndrome, and cerebral vein thrombosis account for two-thirds of the thrombotic episodes (5, 6). Seven patients in our cohort were diagnosed with Budd-Chiari syndrome, which corresponds to 17.1% of all patients examined and 43.8% of patients who experienced TEs. Patients with Budd-Chiari syndrome primarily complained of abdominal discomfort, which was considerably worse in patients with a TE than in patients without a TE (p<0.001). Hepatic dysfunction and splenomegaly were observed to be substantially more prevalent in patients with a TE (p=0.001 and p=0.03 respectively). Ocular involvement is a rare manifestation of thrombosis in patients with PNH (5). Two patients, or 12.5% of all patients who experienced a TE, developed ocular thrombosis in our study. Both patients suffered serious vision loss as a result of these thrombi, which had catastrophic effects on their quality of life.

Several other hematological disorders, including aplastic anemia and, to a lesser extent, myelodysplastic syndrome, have been linked to the presence of a PNH clone (2, 12–15). In contrast, the combination of PNH and myeloproliferative diseases (MPN) is rare and is mostly associated with primary myelofibrosis (13–17). The existence of a PNH clone in PV is extremely uncommon and especially challenging to detect when an intra-abdominal thrombosis occurs. In fact, less than 40 cases have been described in the literature to date where PNH may be associated with a MPN (12, 17–19). In our dataset, four (26.7%) individuals had concomitant MDS, two (13.3%) had primary myelofibrosis and one (6.7%) had PV. The PV patient is a female who was diagnosed with the coexistence of PNH clone and a JAK2V617F-positive PV, with atypical thromboses and not overt hemolysis (18).

It is generally accepted that the size of the PNH clone is one of the most significant risk factors for thrombosis (5, 8, 9, 20–22). In this study, the median size of the granulocyte and monocyte PNH clone in patients with a thrombotic episode was 63% (range, 2.6-96.6%) and 60% (range, 4.6-91.2%) respectively. Patients with PNH and >50% granulocyte clones have a ten-year risk of venous thrombosis of 44%, according to previous findings (9, 23). In patients with a smaller clone size, this proportion was only 5.8%, which is still higher than what is observed in healthy subjects (9, 23). In our study, among 16 patients with a thrombotic event, 10 (62.5%) patients had granulocyte clone sizes greater than 50%.

Elevated LDH at diagnosis is associated with an increased risk of thrombosis (24, 25). In our study group, the median laboratory values of baseline LDH was increased in both patients with and without a TE (858 U/L, range 242-4,024 U/L and 1,020, range 156-3,716 U/L respectively) with no significant difference between the two groups (p=0.83).

It is reasonable to initiate prophylactic anticoagulant therapy at the time of PNH diagnosis; but, long-term anticoagulants are an additional burden for PNH patients and pose a risk of bleeding for half of individuals who will never develop thrombosis (4, 9). No patient in our study received preventive anticoagulant therapy.

Anticoagulants such as heparin or LMWH are still the preferred treatment for acute thrombotic events in PNH patients (26). DOACs are probably equivalent, but their use in PNH has not been exhaustively investigated (26). However, it is widely recognized that anticoagulants alone could not control thrombosis for an extended period of time (6, 26). This requires treating TEs with a complement blocker, which targets the thrombus’ underlying etiology. To prevent the activation of the complement system, eculizumab, ravulizumab or another approved complement inhibitor should be administered with the anticoagulants as soon as possible, if they are available (6, 26). Eculizumab, the first-in-class anti-C5 humanized antibody, improves patient outcome by reducing hemolysis and the incidence of life-threatening morbidities, such as thrombosis, in PNH patients (8, 9). Long-acting C5 inhibitor Ravulizumab is non inferior to eculizumab in terms of efficacy, safety, and tolerance (27–29). Pegcetacoplan is the first C3 inhibitor approved for adult PNH treatment. Both extravascular and intravascular hemolysis are prevented. Anti-C3 compounds, anti-factor D agents, and anti-factor B agents are the strategic targets of many other newly developed therapeutics (30, 31).

The majority of patients examined in our study received LWMH as anticoagulant therapy after the thrombotic episode, with six patients receiving acenocoumarol shortly after LWMH treatment, and three started on DOACs. According to international guidelines, eight patients were started on eculizumab immediately after the TE. Although a considerable reduction in TEs has been demonstrated, the risk exists for patients receiving eculizumab (6, 22, 32). In our cohort, 3 of 13 patients with PNH who did not receive eculizumab treatment developed a TE during the course of the disease, whereas 3 of 26 patients treated with eculizumab were reported to have TEs. Although we have no information regarding the possible cause of thrombosis in patients treated with eculizumab, administration delays may be an issue, but additional information is required. Notably, each of those three patients had a concomitant underlying disease, one PMF, one PV and one homozygous thalassemia.

Due to a lack of evidence on the safe duration of anticoagulant therapy following a thrombotic event in complement inhibitor-treated patients, the decision to discontinue should be made on a case-by-case basis and addressed with the patient (26, 33, 34). No data exist regarding the discontinuation of anticoagulants in our patient group.

In addition to anti-complement mAbs, hematopoietic stem cell transplantation (HSCT) has a curative role but is related to serious or fatal complications in PNH patients; hence, it can be considered for only certain cases, such as patients with severe cytopenias or patients with severe hemolysis or thrombosis refractory to complement inhibition (35). Even in these cases, eculizumab can be utilized to minimize morbidity prior to and during HSCT, as demonstrated by several case reports (36).

PNH is an acquired thrombophilic disorder, and the major cause of death is thrombosis, particularly in the pre-eculizumab era (5–7, 37). Further research indicates that the prevalence of PNH thrombosis has been greatly underestimated. An MRI finding of subclinical pulmonary embolism or early cardiac scarring in 60% and 20% of a small cohort of young PNH patients, respectively, supports this (5, 38). In our study, however, none of the 16 patients who had PNH-induced thrombosis died as a result of the TE. Of course, there were two deaths in our cohort for unspecified reasons, which may imply a possible pathogenetic correlation.

Our study has the limitations of a retrospective study with small number of patients. However, due to the rarity of the disease, prospective studies are difficult to conduct. Moreover, as this is a multicenter study, the cases reported represent a significant number of cases diagnosed and treated in Greece.

In conclusion, regardless of the small sample size, our study demonstrated that thrombosis is a serious consequence of PNH that affects more than one-third of patients. The majority of thrombotic episodes occurred at unusual sites. In addition, the study reveals that the incidence of thrombosis and its link with the size of the PNH clone are similar with prior findings. Ongoing treatment with complement inhibitors has considerably reduced the risk of thrombosis, although it has not eliminate it totally.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

SC collected, analyzed the data and wrote the manuscript. MA, JVA, NEL, TC, PZ, AB, ZB, EK, CF, PC, DS, VD, ASi, EM, MPs, HAP, TM, NAV, SK, FK, ASk, GV, IK, VP, CL, IB, SD, and MPag provided data. EH, MPap designed the study, provided, analyzed the data and wrote the manuscript. All authors contributed to the article and approved the submitted version.