Thrombosis in PNH: largely gone, but not forgotten

The most dramatic and lifesaving consequence of the introduction of effective therapy for paroxysmal nocturnal hemoglobinuria (PNH) was a marked reduction in thrombosis risk. In this issue of Blood, Gurnari et al¹ address several largely unanswered, but frequently asked, questions in relation to the modern prophylaxis and management of PNH-related thrombosis including the need for prophylactic anticoagulants, the place of direct oral anticoagulants, and the recognition of any new predictors for thromboembolic phenomena.² The need to improve understanding of thrombosis in this setting is underscored by the natural history of untreated patients with PNH who historically had a high risk of thrombosis, many of which were atypical and resulted in a high rate of premature mortality.³ The reported study was of 267 patients in 4 academic centers in the United States. Patients either had classical hemolytic PNH or aplastic anemia with a clinically significant PNH clonal population of cells. The analysis was of aggregated retrospective data, and patients were investigated and treated according to local clinical standards.

It is important to understand the extent of the clinical problem of thrombosis in patients with PNH. Although the causation of thrombosis in this disease is complex, it has been at least partially elucidated,³ and it is now understandable why prophylactic anticoagulation with heparin products or vitamin K antagonist were minimally efficacious given the indirect effect of direct complement activation and the cellular and biochemical consequences of intravascular hemolysis itself. In initial studies of the C5 inhibitor eculizumab, prophylactic anticoagulation was required to be continued if already in place at trial entry but was not able to be initiated after study entry.⁴ Although there was a dramatic reduction in thrombosis events after initiation of anti-C5 therapy, anticoagulation had no apparent influence on subsequent thromboembolic events.

In this report by Gurnari et al, a slight majority of the patients were not on treatment with anticomplement agents due to the absence of a significant hemolytic phenotype. With reference to thromboembolism, 21% of the patients experienced this, 43% of the events were the initial manifestation of the disease, and 7% of the events occurred when anticomplement therapy was discontinued. The majority of the patients who experienced thromboses while on anticomplement therapy were suboptimal responders to that therapy. As has been shown in other studies, the rate of thromboembolism increased with an increase in both granulocyte and erythrocyte clone size.⁵ The multivariate analysis done in this study showed a particular correlation with a predominant type II (cells only partially deficient in glycosylphosphatidylinositol-anchored proteins) erythrocyte population.⁶ Other interesting molecular correlations included that of the variant allele frequency of PIGA and clonal mosaicism, where there were multiple PIGA mutations, being associated with thromboembolism occurrence.

In the management of thrombosis, standard anticoagulation or antiplatelet therapy was immediately commenced except in 1 patient who developed Budd-Chiari syndrome and was unable to receive anticoagulation therapy because of severe thrombocytopenia; the patient subsequently died of progression of this thrombotic phenomenon. Direct-acting oral anticoagulants were used in some patients without apparent inferiority of outcome, confirming preliminary observations from elsewhere.⁷ Notably, anticoagulants were able to be withdrawn without recurrence in those patients in whom anticomplement therapy was commenced.

The importance of complete and relevant data sets in rare diseases is exemplified by this study. A well-constructed, professionally run academic registry should build on this knowledge and understand the impact of new diagnostics and therapies on this phenomenon and others. Fortunately, the International PNH Interest Group⁸ has established such a registry (The International PNH Interest Group PNH Registry), which, in the coming years, should be a valuable resource for research into this fascinating disease.

These analyses should encourage confidence and improve accuracy among treating physicians in the prediction and management of thromboembolic phenomena in patients with PNH. The question of whether complement blockade should be used in patients with a predominant type II red blood cell phenotype who do not otherwise meet standard criteria for therapy is intriguing, given the clear association with an increased risk of thromboembolism in this study.

Conflict-of-interest disclosure: J.S. serves on the advisory committee and speaker bureau for, is a consultant to, and receives travel assistance for Alexion Pharmaceuticals; serves on the advisory committee and speaker bureau for, is a consultant to, and receives travel assistance from Sobi Pharmaceuticals; serves on the advisory committee and is a consultant to Novartis; and is a consultant to ADARx, Kira Pharmaceuticals, and Samsun Bioepis.