ABSTRACT
The clinical significance of PNH Type II white blood cells (WBCs) remains unclear. We assessed the relative percentage (rel%) of Type II neutrophils in 355 patients with a PNH clone ≥ 1% on neutrophils enrolled by 33 flow cytometry laboratories in the 5‐year French nation‐wide multicenter prospective observational study. We first analyzed 127 of 133 patients with a major PNH clone (≥ 50%) and evaluable Type II neutrophil rel%. Using hemolysis data available for 107 patients, a threshold of 3% Type II neutrophils rel% distinguished two groups: 22 “high Type II” and 105 “low Type II” patients. Hemolysis was less frequent in the “high Type II” group. Thrombosis was more frequent at diagnosis and during cumulative long‐term follow‐up in this group. High neutrophil rel% ≥ 3% was independently associated with thrombotic events. Among 222 patients with smaller PNH clones (1%–50%), 207 displayed evaluable Type II neutrophil rel%, of which 87 were classified as “high Type II” and 120 as “low Type II”. Although no thrombosis was present at diagnosis, two thrombotic events occurred during follow‐up in the “high Type II” group, whereas none were observed in the “low Type II” group. Long‐term follow‐up showed no major changes in Type II neutrophil rel%, irrespective of variations in total clone size. These findings suggest that, regardless of the total PNH clone size, PNH Type II neutrophil rel% is a stable parameter and may represent a reliable predictor of thrombosis, potentially supporting earlier consideration of anticomplement therapy initiation.
1. Introduction
Thrombosis remains the most feared complication of paroxysmal nocturnal hemoglobinuria (PNH), accounting for 40%–67% of PNH‐related deaths prior to the advent of complement inhibition [1, 2, 3, 4, 5]. Thrombotic events in PNH may occur at both venous and arterial territories, with overrepresentation of atypical sites, most notably intra‐abdominal veins, where Budd‐Chiari syndrome is the most frequent manifestation, as well as cerebral veins [6, 7, 8]. Although the pathogenesis of PNH is now well characterized [9], the mechanisms underlying thrombosis remain incompletely understood and are likely multifactorial with multiple proposed contributors [6, 10, 11, 12]. Notably, the complement cascade plays a central role through its close bidirectional interplay with the coagulation system [6], providing a mechanistic rationale for the profound clinical impact of anti‐complement therapy. Indeed, complement inhibition has revolutionized the natural history of PNH by effectively controlling hemolysis and markedly reducing thrombotic risk [13, 14, 15, 16]. Consistently, recent real‐world studies demonstrated that complement inhibition was associated with an approximately 50% reduction in thrombotic incidence [17] and a marked decrease in thromboembolism rate to 1.05 events/100 patient‐years [18], notably for splanchnic and cerebral sites.
PNH is caused by a somatic mutation in the phosphatidylinositol glycan class A (PIGA), which results in a block of glycosyl phosphatidylinositol (GPI) anchor biosynthesis, leading to a deficiency in GPI–anchored proteins [19, 20]. The underlying genetic abnormality may consist of a complete deletion, a frameshift, a splicing, or a point mutation, all of which either completely inactivate or markedly reduce PIGA expression [9]. Mature blood cells derived from the hematopoietic clone may exhibit either a complete deficiency (Type III) or a partial deficiency (Type II) of GPI‐linked proteins or a mixture of both, and almost invariably coexist with residual cells showing normal expression of these proteins (Type I) [21, 22, 23]. The clinical features of PNH stem from the loss of GPI‐anchored complement regulatory proteins, such as CD55 and CD59, rendering red blood cells (RBCs) susceptible to complement‐mediated attack [24]. Flow cytometry has become firmly established as the method of choice for identifying GPI anchor–deficient cell populations for both the diagnosis and the monitoring of PNH [25, 26, 27]. Reporting the distinct proportions of Type II and Type III RBCs may have clinical relevance, as Type III cells exhibited greater susceptibility to lysis than Type II cells [28]. The introduction of fluorescently labeled inactive variant of the protein aerolysin (FLAER) reagent [29] in 2010 [26] markedly improved the diagnosis of PNH by enabling highly sensitive detection of GPI‐deficient white blood cell (WBC) clones [26, 30]. In particular, FLAER facilitates the identification of Type II WBC populations, although their clinical significance has not yet been established. Therefore, reporting separate Type II and Type III PNH populations in neutrophils and monocytes is not recommended, unlike the total clone size [26, 30].
Although hemolysis is a major contributor to thromboembolism in PNH and the risk of thrombosis increases with larger granulocyte PNH size [4, 31], it has been reported that overt hemolysis is not always required for thrombosis [32] and that PNH patients with a higher proportion of Type II RBCs appear to be at increased thrombosis risk [17, 32, 33]. Because the proportion of Type II RBCs can be influenced by hemolysis of Type III RBCs and transfusion history, we sought to assess the clinical relevance of Type II WBCs. To this end, we analyzed herein data from the national French PNH Observatory, which enrolled both newly and previously diagnosed PNH patients over a 5‐year period.
2. Methods
2.1. PNH Patients
The French National PNH Observatory was a prospective observational multicenter study conducted by the French PNH working group within the French‐speaking flow cytometry society, CytHem (www.cythem.fr). Patients with a diagnosed PNH clone or GPI‐deficient cells ≥ 0.01% on neutrophils detected by flow cytometry, prior to the opening or during a 5‐year period, in France, regardless of age, were enrolled by 33 French laboratories between February 2016 to February 2021. Patients' baseline assessment was considered as the first quantification of the PNH clone by flow cytometry. Peripheral blood collected from each patient was analyzed locally in each participating center, where the referent cytometrist provided clinical and biological information on a dedicated eCRF. All centers participated in an ongoing PNH flow cytometry quality control program [34]. PNH patients were followed up during the 5‐year opening period or for more time if the PNH clone detection was performed before the opening.
The study was approved by the Consultative Committee for the Data Processing in Health Research (CCTIRS), authorized by the French Data Protection Authority (CNIL) and sponsored by Alexion Pharmaceuticals France. As this study was observational, human participants did not have to give written informed consent, according to French regulatory rules.
2.2. PNH Flow Cytometry Analysis
All French laboratories assessed PNH clone or GPI‐deficient cells using flow cytometry panels recommended by the first international guidelines published in 2010 [26]. Most laboratories used antibodies against GPI‐anchored proteins, anti‐CD24 for neutrophil analysis and anti‐CD14 for monocyte analysis; two laboratories used anti‐CD157 instead of these markers. The FLAER reagent was included in the panel from 2010 onward but was not used for analyses performed before that date. RBCs were analyzed with a panel including anti‐CD59, as recommended by the guidelines [26]. Samples were acquired on FACSCanto II or FACSLyric (Beckton Dickinson), or on Navios or DxFLEX (Beckman Coulter) cytometers.
Cut‐off values for PNH clone detection were defined as ≥ 0.01% for neutrophils, ≥ 0.1% for monocytes and ≥ 0.01% for RBCs. For WBCs, these thresholds were based on results obtained from the multicenter workshop organized by the PNH French working group [35].
Only patients with a PNH clone ≥ 1% were included in the present analysis. The distribution of Type II and Type III cells among neutrophils, monocytes and RBCs was independently validated by both principal investigators (O.W.‐B. and M.L.G.‐T.). Type II or Type III WBCs were evaluated only when the flow cytometry panel included FLAER.
2.3. Statistical Analysis
Median and interquartile ranges (IQRs) were used to tabulate continuous variables, whereas numbers and percentages were used for categorical variables. Accordingly, Fischer and χ 2 tests were used to compare frequencies when appropriate. The area under the curve (AUROC, Receiver operator characteristic) was calculated to determine the optimal threshold for Type II neutrophil relative percentages based on presence or absence of hemolysis (as defined by biological parameters) using the MedCalc Statistical software version 12.7.5 (Ostend, Belgium). Univariable logistic regression models were used to assess associations between relevant biological variables and thrombotic events. A multivariable logistic regression model, adjusted for relevant covariates, was used to evaluate these associations.
3. Results
3.1. 355 Patients With a PNH Clone Above 1% Enrolled in the French National Observatory
Thirty‐three participating flow cytometry laboratories across France enrolled 637 patients in the 5‐year French nation‐wide multicenter observational study (Figure 1 upper left panel). Prior to inclusion, all the clinical and biological information as well as flow cytometry raw data files provided for each patient underwent systematic double verification by both investigators (O.W.‐B. and M.L.G.‐T.), notably for the validation of the PNH Type II and Type III cell quantification in the three lineages (See flow chart Figure 1). This led to the exclusion of 23 incorrectly registered patients as having rare cells with PNH phenotype between 0.01% and 0.1% (corresponding to false positive cases, i.e., patients displaying less than 20 events of Type III neutrophils according to the international recommendations [26, 30]). As this was an Observational study with no planned intervention, 23 patients who were registered early (corresponding to patients with a PNH clone detected prior to the opening of Observatory with no subsequent follow‐up points during the 5‐year enrollment period) couldn't be included. In addition, 46 patients were belatedly registered (corresponding to patients with a PNH clone detected after the closure of the Observatory, as a prolongation period was initially announced but later withdrawn due to changes in French regulatory rules).
FIGURE 1.
Flow chart of the 5‐year French Observational study of PNH clones. Upper left panel: 33 French flow cytometry laboratories including overseas; Upper right panel: Inclusion rhythm during the 5‐year Observatory opening for patients with a PNH clone above 1% (shown in blue) and below 1% (shown in orange); Middle panel: Repartition of the total clone size in neutrophils among the 545 patients included; Lower panels: Venn diagrams showing the clinical indications for the first PNH clone quantification for both the patients with a major clone ≥ 50% and those with a smaller clone between 1% and 50%. [Color figure can be viewed at wileyonlinelibrary.com]
Eventually, 545 patients with PNH clone or GPI‐deficient cells ≥ 0.01% newly‐ or previously diagnosed were included in this study and divided into 190 patients with PNH clone < 1% and 355 patients with a PNH clone ≥ 1%, who were enrolled at a steady rate between 2016 and 2021 (Figure 1 upper right panel). For 169 patients, the first quantification of the PNH clone had been performed by flow cytometry before the Observatory opened, whereas 186 patients were newly diagnosed. The total PNH clone size in neutrophils exhibited the characteristic bimodal distribution typically observed [30] (Figure 1 middle panel). The indications for initial PNH clone quantification aligned with expectations. Among the 133 patients with a major PNH clone (i.e., ≥ 50%), most presented with hemolytic anemia, with bone marrow failure and thrombosis occurring less frequently. In contrast, among the 222 patients with a smaller PNH clone (1%–50%), bone marrow failure was the predominant indication (for details, see Venn diagrams in Figure 1).
3.2. The 3% Threshold of Type II Neutrophil Relative Percentage
To enable more straightforward quantification of PNH Type II neutrophils, we first focused on the 133 patients with a major PNH clone. Of these, 127 displayed evaluable Type II neutrophil percentages (See algorithm Figure 2A), in whom we calculated the proportion of Type II neutrophils among the total PNH clone, that is, the relative percentage (rel%) of Type II cells, obtained by the following operation (Type II/(Type II + Type III) × 100). Using the 107 patients with available data on hemolysis status, we performed an AUROC test based on the presence of hemolysis (displayed by 95 patients while no hemolysis was found by 12 patients). We determined a threshold of 3% of Type II neutrophil rel%, which enabled the distinction of two groups of patients, those with ≥ 3% of Type II rel% (hereafter referred to “high Type II”) and those with < 3% (“low Type II”).
FIGURE 2.
Establishment of two patient groups based on the relative proportion of Type II PNH neutrophils (“high Type II” vs “low Type II”). (A) Algorithm used to determine the 3% threshold for the Type II neutrophil relative percentage (rel%). Example of rel% Type II neutrophil calculation for the patient shown: Rel% Type II = (Type II = 1.19%)/(II = 1.19% + III = 65.22%) × 100 = 1.79%. (B) Classification of patients into “high Type II” or “low Type II” groups for patients with a major clone ≥ 50% or with a clone size between 1% and 50%. [Color figure can be viewed at wileyonlinelibrary.com]
3.3. Patients With High Type II Neutrophil Proportion Displayed Less Hemolysis
By applying the previously determined 3% threshold on the 127 patients with evaluable PNH Type II neutrophils, we identified 22 patients with high rel% (5.98%; IQR, 4.75%–12.26%) and 105 patients with low rel% of Type II neutrophils (0.00%; IQR, 0.00%–0.69%; p < 0.0001) (Figure 2B; Table 1; Figure S1). In addition, Type II monocyte rel% were higher in the “high Type II” group (7.76%; IQR, 5.13%–11.07%) compared to the “low Type II” group (0.00%; IQR, 0.00%–0.93%; p < 0.0001). The limited number of RBC data points prevented from reaching statistical significance, although a numerically higher Type II RBC rel% was observed in the “high Type II” (22.19%; IQR, 0.00%–65.04% vs. 0.32%, IQR, 0.00%–24.15%; p = 0.1786) (Figure 3A,B left panels, Table 1).
TABLE 1.
Demographic, clinical and biological characteristics at first PNH clone quantification by flow cytometry.
| Variable | PNH clone ≥ 50% | PNH clone 1%–50% | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Low Type II | n | High Type II | n | p | All cases | n | Low Type II | n | High Type II | n | p | All cases | n | |
| Demographics | ||||||||||||||
| Age at diagnosis, y, median (IQR) | 42 (26–63) | 105 | 47 (34–55) | 22 | 0.3809 | 44 (28–62) | 127 | 42 (29–61) | 120 | 46 (29–62) | 87 | 0.3177 | 43 (29–61) | 207 |
| Sex ratio, Male/Female | 0.84 | 105 | 1.20 | 22 | 0.4506 | 0.90 | 127 | 0.88 | 120 | 0.45 | 87 | 0.0235 | 0.67 | 207 |
| Male, % | 46% | 48 | 55% | 12 | 47% | 60 | 47% | 56 | 31% | 27 | 40% | 83 | ||
| Female, % | 54% | 57 | 45% | 10 | 53% | 67 | 53% | 64 | 69% | 60 | 60% | 124 | ||
| Available clinical information | 95.2% | 100 | 95.5% | 21 | 95.3% | 121 | 93.3% | 112 | 96.6% | 84 | 94.7% | 196 | ||
| Hemolytic anemia | 54.0% | 54 | 23.8% | 5 | 0.0119 | 48.8% | 59 | 14.3% | 16 | 6.0% | 5 | 0.0668 | 10.7% | 21 |
| Bone marrow failure (with hemolytic anemia, %) | 40.0% (32%) | 40 | 47.6% (20%) | 10 | 0.5192 | 41.3% (30%) | 50 | 85.7% (6%) | 96 | 94.0% (3%) | 79 | 0.0668 | 89.3% (5%) | 175 |
| Aplastic anemia (with hemolytic anemia, %) | 70.0% (29%) | 28 | 70.0% (29%) | 7 | 0.6241 | 70.0% (29%) | 35 | 81.2% (8%) | 78 | 82.3% (3%) | 65 | 0.2274 | 81.7% (6%) | 143 |
| Hypoplastic MDS (with hemolytic anemia, %) | 22.5% (56%) | 9 | 30.0% (0%) | 3 | 0.4353 | 24.0% (42%) | 12 | 4.2% (0%) | 4 | 6.3% (0%) | 5 | 0.5014 | 5.15% (0%) | 9 |
| Unexplained cytopenia (with hemolytic anemia, %) | 7.5% (0%) | 3 | 0.0% (0%) | 0 | 1.000 | 6.0% (0%) | 3 | 14.6% (0%) | 14 | 11.4% (0%) | 9 | 0.7007 | 13.15% (0%) | 23 |
| Thrombosis (with hemolytic anemia, %) | 6.0% (33%) | 6 | 28.6% (50%) | 6 | 0.0063 | 9.9% (42%) | 12 | 0% (0%) | 0 | 0% (0%) | 0 | 1.000 | 0% (0%) | 0 |
| CBC and reticulocytes | ||||||||||||||
| Hb, g/dL, median (IQR) | 9.3 (8.2–10.7) | 96 | 9.5 (8.8–10.6) | 21 | 0.2556 | 9.3 (8.3–10.6) | 117 | 10.0 (8.0–11.5) | 109 | 9.2 (7.9–11.2) | 79 | 0.4052 | 9.8 (8.0–11.3) | 188 |
| MCV, fL, median (IQR) | 98.8 (93.7–105.7) | 93 | 94.8 (90.4–99.1) | 21 | 0.0291 | 98.0 (93.0–105.0) | 114 | 100.0 (94.0–105.0) | 109 | 99.9 (91.2–107.4) | 77 | 0.6689 | 99.9 (92.8–106.1) | 186 |
| WBC, ×109/L, median (IQR) | 4.4 (3.4–5.5) | 96 | 4.2 (3.1–4.8) | 21 | 0.3102 | 4.4 (3.4–5.5) | 116 | 2.9 (2.2–3.9) | 108 | 2.8 (2.0–3.6) | 78 | 0.4865 | 2.8 (2.0–3.9) | 186 |
| Neutrophil count, ×109/L, median (IQR) | 2.6 (1.7–3.7) | 96 | 2.5 (1.4–3.0) | 21 | 0.3938 | 2.5 (1.7–3.5) | 117 | 1.1 (0.5–1.9) | 106 | 1.0 (0.6–1.7) | 75 | 0.5623 | 1.0 (0.6–1.8) | 181 |
| Lymphocyte count, ×109/L, median (IQR) | 1.2 (0.9–1.5) | 92 | 1.1 (0.8–1.4) | 21 | 0.1780 | 1.2 (0.9–1.5) | 113 | 1.4 (1.0–1.9) | 105 | 1.4 (1.0–2.0) | 71 | 0.5093 | 1.4 (1.0–1.9) | 176 |
| Monocyte count, ×109/L, median (IQR) | 0.3 (0.2–0.5) | 93 | 0.3 (0.2–0.4) | 21 | 0.1853 | 0.3 (0.2–0.5) | 114 | 0.2 (0.1–0.3) | 105 | 0.2 (0.1–0.3) | 71 | 0.9975 | 0.2 (0.1–0.3) | 176 |
| Platelets, ×109/L, median (IQR) | 144 (96–202) | 93 | 135 (59–156) | 21 | 0.0501 | 141 (90–193) | 114 | 44 (22–83) | 108 | 30 (12–60) | 79 | 0.0432 | 36 (16–75) | 187 |
| Absolute reticulocyte count, 109/L, median (IQR) | 123 (100–176) | 78 | 92 (51–128) | 18 | 0.0019 | 117 (100–171) | 96 | 41 (25–69) | 87 | 47 (23–65) | 61 | 0.9224 | 45 (24–67) | 148 |
| Biochemical markers | ||||||||||||||
| LDH, U/L, median (IQR) | 1002 (489–1731) | 73 | 808 (494–1232) | 16 | 0.3868 | 918 (489–1709) | 89 | 256 (205–351) | 74 | 251 (200–301) | 57 | 0.5586 | 251 (205–338) | 131 |
| LDH > 1.5 N, % | 67% | 72 | 75% | 16 | 0.5174 | 68% | 88 | 24% | 74 | 21% | 57 | 0.6586 | 23% | 131 |
| Total bilirubin, μmol/L, median (IQR) | 21 (13–29) | 74 | 14 (7–20) | 19 | 0.0079 | 19 (12–27) | 93 | 10 (8–17) | 85 | 12 (8–16) | 61 | 0.7746 | 11 (8–17) | 146 |
| Haptoglobin, g/L, median (IQR) | 0.10 (0.08–0.20) | 65 | 0.15 (0.08–0.27) | 14 | 0.2229 | 0.10 (0.08–0.20) | 79 | 0.10 (0.08–0.70) | 57 | 0.20 (0.10–1.10) | 37 | 0.3695 | 0.13 (0.08–0.90) | 94 |
| Creatinine, μmol/L, median (IQR) | 65 (56–80) | 75 | 67 (54–75) | 17 | 0.7474 | 65 (56–79) | 92 | 71 (62–85) | 88 | 66 (57–81) | 64 | 0.3047 | 70 (59–82) | 152 |
| PNH total clone size | ||||||||||||||
| Neutrophil clone, %, median (IQR) | 92.15 (81.76–98.20) | 105 | 90.69 (75.44–96.90) | 22 | 0.6215 | 92.15 (80.28–97.84) | 127 | 6.57 (2.62–17.66) | 120 | 4.74 (1.81–11.53) | 87 | 0.0289 | 5.57 (2.32–13.99) | 207 |
| Monocyte clone, %, median (IQR) | 91.48 (78.52–96.74) | 104 | 87.53 (76.10–93.01) | 22 | 0.1976 | 90.65 (77.71–96.49) | 126 | 9.20 (3.06–21.85) | 114 | 5.01 (2.13–13.54) | 87 | 0.0091 | 7.09 (2.32–16.48) | 201 |
| RBC clone, %, median (IQR) | 30.48 (16.15–45.73) | 56 | 13.21 (0.60–32.79) | 14 | 0.0563 | 28.90 (13.05–45.15) | 70 | 1.06 (0.35–5.52) | 73 | 0.80 (0.24–2.67) | 50 | 0.4475 | 0.90 (0.31–4.20) | 123 |
| PNH Type II clone size | ||||||||||||||
| Neutrophil Type II, %, median (IQR) | 0.00 (0.00–0.53) | 105 | 5.80 (3.91–10.92) | 22 | < 0.0001 | 0.15 (0.00–1.50) | 127 | 0.00 (0.00–0.02) | 120 | 0.57 (0.23–2.00) | 87 | < 0.0001 | 0.06 (0.00–0.51) | 207 |
| Monocyte Type II, %, median (IQR) | 0.00 (0.00–0.81) | 104 | 6.62 (3.27–8.48) | 22 | < 0.0001 | 0.04 (0.00–2.50) | 126 | 0.00 (0.00–0.02) | 114 | 0.46 (0.05–1.39) | 87 | < 0.0001 | 0.00 (0.00–0.51) | 201 |
| RBC Type II, %, median (IQR) | 0.03 (0.00–4.22) | 56 | 0.27 (0.00–9.55) | 14 | 0.4941 | 0.03 (0.00–7.20) | 70 | 0.00 (0.00–0.03) | 73 | 0.00 (0.00–0.48) | 50 | 0.0001 | 0.00 (0.00–0.21) | 123 |
| Relative % of PNH Type II cells | ||||||||||||||
| Neutrophil Type II, rel%, median (IQR) | 0.00 (0.00–0.69) | 105 | 5.98 (4.75–12.26) | 22 | < 0.0001 | 0.23 (0.00–1.75) | 127 | 0.00 (0.00–0.26) | 120 | 11.96 (5.77–28.57) | 87 | < 0.0001 | 1.38 (0.00–8.91) | 207 |
| Monocyte Type II, rel%, median (IQR) | 0.00 (0.00–0.93) | 104 | 7.76 (5.13–11.07) | 22 | < 0.0001 | 0.07 (0.00–2.79) | 125 | 0.00 (0.00–0.44) | 113 | 7.05 (0.36–21.20) | 87 | < 0.0001 | 0.00 (0.00–8.12) | 200 |
| RBC Type II, rel%, median (IQR) | 0.32 (0.00–24.15) | 56 | 22.19 (0.00–65.04) | 14 | 0.1786 | 0.47 (0.00–32.80) | 70 | 0.00 (0.00–4.81) | 73 | 10.10 (0.00–37.50) | 50 | 0.0005 | 0.00 (0.00–23.60) | 123 |
Note: Bold values indicate statistical significance.
Abbreviations: Hb, hemoglobin; IQR, interquartile range; LDH, lactate deshydrogenase; MCV, mean corpuscular volume; RBC, red blood cells; WBC, white blood cells; y, years.
FIGURE 3.
Relative percentage of PNH Type II neutrophils (A) and total PNH clone size (B) in neutrophils, monocytes and RBC, shown for patients with a major clone ≥ 50% (left panels) and for those with a clone of 1%–50% (right panels). [Color figure can be viewed at wileyonlinelibrary.com]
However, the total clone size was similar for both groups in WBCs (neutrophils: 90.69%, IQR, 75.44%–96.90% vs. 92.15%, IQR, 81.76%–98.20%; p = 0.6215; monocytes: 87.53%, IQR, 76.10%–93.01% vs. 91.48%, IQR, 78.52%–96.74%; p = 0.1976) as well as in RBCs (13.21%, IQR, 0.60%–32.79% vs. 30.48%, IQR, 16.15%–45.73%; p = 0.0563), in the “high Type II” and the in the “low Type II” group, respectively (Table 1).
The median age of these patients at first PNH clone quantification (male‐to‐female ratio, 1.20 vs. 0.84; p = 0.4506) was similar in both groups (47 years, IQR, 34–55 years vs. 42 years, IQR, 26–63 years; p = 0.3809).
Hemolysis, which was the criterion used to build the ROC, was less frequent in “high Type II” group: 61.1% compared to “low Type II” group: 94.4% (p < 0.001). Accordingly, patients with “high Type II” displayed lower absolute reticulocyte count (92 × 109/L, IQR, 51–128 × 109/L vs. 123 × 109/L, IQR, 100–176 × 109/L; p = 0.0019), lower mean corpuscular volume (MCV) (94.8 fL, IQR, 90.4–99.1 fL vs. 98.8 fL, IQR, 93.7–105.7 fL; p = 0.0291) and lower total bilirubin (14 μmol/L, IQR, 7–20 μmol/L vs. 21 μmol/L, IQR, 13–29 μmol/L; p = 0.0079). All the other complete blood count (CBC) parameters were similar between the two patient groups (Table 1).
3.4. “High Type II” Patients Had a Higher Thrombosis Rate Regardless of the Total Clone Size
Thrombotic events appeared to be more frequent at diagnosis in “high Type II” group: 28.6% vs. 6.0% in the “low Type II” group (p = 0.0063) and when analyzing the cumulative events during long‐term follow‐up (31.7% vs. 7.6%; p = 0.0048) (Table 1). Most of these thromboses involved the venous system in both the “high Type II” group (100%) and the “low Type II group” (83.3%). The majority occurred at unusual sites, with Budd‐Chiari syndrome being the most frequent (42.1%), followed by deep vein thrombosis (DVT, 26.3%), pulmonary embolism (5.3%), cerebrovascular (5.3%), splanchnic (5.3%), and superficial veins (5.3%). Only 4.5% of thrombotic events originated in the arterial system, affecting both cerebrovascular and coronary arteries (Figure 4, Table 2). Notably, at the time of the three thrombotic events observed during the follow‐up in both groups, the total clone size in neutrophils had increased in all cases. Univariate logistic regression analysis further demonstrated a significantly increased risk of thrombotic events in patients with a high proportion of Type II neutrophils (OR, 5.66; 95% CI, 1.79–17.89; p = 0.0032). No other biological parameters were associated with thrombotic risk in this cohort (Table 3). In multivariable logistic regression, a high Type II rel% neutrophils remained independently associated with thrombotic events (OR, 4.32; 95% CI, 1.30–14.31; p = 0.0167) (Table 4).
FIGURE 4.
Thrombotic events in patients with PNH high Type II and low Type II neutrophils, adapted from [17] (A) Pie charts illustrating the type of thrombotic events in patients with high Type II (upper pies) and low Type II (lower pies) with a major clone above 50% (left pies) and a clone between 1% and 50% (right pies). (B) Anatomic site of origin of all thrombotic events. (C) Longitudinal follow‐up and anticomplement therapy in patients with PNH low Type II and high Type II neutrophils experiencing thrombotic events (patients are shown by decreasing total PNH clone sizes); y, years. [Color figure can be viewed at wileyonlinelibrary.com]
TABLE 2.
PNH cases with thrombotic events: demographics, CBC and reticulocytes, biochemical markers, PNH total clone size and relative percentage of Type II cells at the time of the first PNH clone evaluation.
| Case no. | Age, y | Sex | Clinical presentation | Thrombosis localization | CBC and reticulocytes | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Hb, g/dL | MCV, fL | WBC, ×109/L | Neut count, ×109/L | Mono count, ×109/L | Plts, ×109/L | Retic count, ×109/L | |||||
| Low Type II | |||||||||||
| PNH1 | 22 | F | Venous thrombosis | Budd‐Chiari | 5.7 | 98 | 11.8 | 8.7 | 1.9 | 280 | NE |
| PNH2 | 42 | M | Venous thrombosis | Budd‐Chiari | 10.0 | 86 | 9.1 | 6 | 1.3 | 240 | 105 |
| PNH3 | 58 | F | Venous thrombosis | Deep vein | 9.2 | 107 | 3.4 | 2.3 | 0.4 | 193 | 109 |
| PNH4 | 35 | M | Venous thrombosis ‐ HA | Budd‐Chiari | 9.6 | 81 | 8.1 | 4.5 | 0.4 | 44 | 277 |
| PNH5 | 49 | M | Venous thrombosis | Deep vein | 8.9 | 99 | 3.7 | 2.7 | NE | 226 | 67 |
| PNH6 | 21 | F | Venous thrombosis | Budd‐Chiari | 8.8 | 96 | 14.3 | 11.1 | 1.4 | 86 | NE |
| PNH7 | 55 | F | HA | FU: Deep vein | 9.1 | 95 | 4.4 | 2.4 | 0.4 | 77 | 100 |
| PNH8 | 68 | F | HA | FU: Deep vein | 10.7 | 98 | 2.8 | 1.7 | 0.2 | 86 | 87 |
| High Type II | |||||||||||
| PNH9 | 30 | M | HA | FU: Budd‐Chiari | 15.0 | 90 | 4.5 | 3,0 | 0.4 | 115 | 66 |
| PNH10 | 44 | M | Venous thrombosis | Budd‐Chiari | 8.8 | 83 | 4.8 | 2.9 | 0.4 | 147 | 128 |
| PNH11 | 71 | M | Venous thrombosis | Splanchnic vein | 11.1 | 94 | 6 | 3.1 | 0.7 | 148 | 61 |
| PNH12 | 24 | F | Venous thrombosis | Deep vein | 10.3 | 96 | 5.3 | 4.6 | 0.3 | 90 | NE |
| PNH13 | 54 | M | Arterial thrombosis | Cerebrovascular & coronary arteries | 9.1 | 105 | 3.4 | 2.5 | 0.3 | 106 | 135 |
| PNH14 | 32 | M | Venous thrombosis | Cerebrovascular vein | 11.5 | 100 | 1.7 | 0.8 | 0.1 | 103 | 113 |
| PNH15 | 50 | Venous thrombosis | Budd‐Chiari and pulmonary embolism | 11.0 | 92 | 4.2 | 2.5 | 0.3 | 135 | 135 | |
| PNH16 | 75 | F | HA | FU: Budd‐Chiari | 9.3 | 100 | 3.9 | 2.5 | 0.4 | 23 | 90 |
| PNH17 | 40 | F | Aplastic anemia | FU: superficial vein | 12.2 | 96 | 3.7 | 0.8 | 0.5 | 113 | 34 |
| Case no. | Biochemical markers | PNH total clone size | Relative % of Type II cells | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| LDH, U/L | Total bilirubin, μmol/L | Hapto, μmol/L | Creat, μmol/L | Neut clone, % | Mono clone, % | RBC clone, % | Neut Type II, rel% | Mono Type II, rel% | RBC Type II, rel% | |
| Low Type II | ||||||||||
| PNH1 | NE | NE | NE | NE | 99.83 | 99.98 | NE | 0.12 | 0.07 | NE |
| PNH2 | 689 | 25 | 0.2 | 40 | 95.28 | 93.95 | NE | 0.00 | 0.00 | NE |
| PNH3 | 411 | 10 | 0.62 | 65 | 94.28 | 69.08 | NE | 1.05 | 4.39 | NE |
| PNH4 | 3485 | 25 | 0.08 | 59 | 93.83 | 88.12 | NE | 0.29 | 0.50 | NE |
| PNH5 | 1462 | 17 | 0.08 | 90 | 87.29 | 67.47 | NE | 0.00 | 0.00 | NE |
| PNH6 | 727 | 9 | 0.00 | 49 | 87.01 | 82.48 | 45.15 | 0.16 | 0.52 | 0 |
| PNH7 | 1002 | 12 | 0.30 | 60 | 79.89 | 77.59 | 17.01 | 0.00 | 0.00 | 100.00 |
| PNH8 | NE | 9 | 0.10 | 48 | 62.90 | 72.48 | 48.6 | 0.00 | 0.00 | 0 |
| High Type II | ||||||||||
| PNH9 | 292 | 39 | NE | 75 | 99.94 | 99,00 | 2.99 | 18.91 | 35.83 | 17.39 |
| PNH10 | 914 | 21 | 0.08 | 70 | 89.13 | 83.57 | 31.45 | 8.47 | 7.23 | 27.00 |
| PNH11 | 433 | 11 | 0.27 | 109 | 85.90 | 91.50 | 3.14 | 4.73 | 0.00 | 0.00 |
| PNH12 | NE | 7 | NE | 63 | 80.67 | 80.27 | 28.89 | 45.85 | 31.98 | 65.04 |
| PNH13 | 808 | 43 | 0.08 | NE | 80.47 | 79.23 | 32.79 | 9.89 | 8.23 | 32.78 |
| PNH14 | 870 | 13 | 0.08 | 57 | 70.79 | 63.76 | NE | 5.06 | 5.13 | NE |
| PNH15 | 502 | 20 | 0.20 | 83 | 65.82 | 61.99 | NE | 5.94 | 8.55 | NE |
| PNH16 | 377 | 12 | 0.08 | 78 | 21.77 | 14.80 | 4.20 | 5.65 | 9.05 | 11.43 |
| PNH17 | NE | 7 | NE | 97 | 4.74 | 5.83 | 2.99 | 60.34 | 38.42 | 17.39 |
Note: Thrombotic events that occurred during the follow‐up are shaded.
Abbreviations: F, female; FU, follow‐up; HA, hemolytic anemia; Hapto, haptoglobin; Hb, hemoglobin; LDH, lactate deshydrogenase; M, male; MCV, mean corpuscular volume; Mono, monocytes; NE, not evaluated; Neut, neutrophils; Plts, platelets; RBC, red blood cells; Retic, reticulocytes; WBC, white blood cells; Y, years.
TABLE 3.
Univariate logistic regression analysis of predictors of thromboembolic events in PNH patients (PNN ≥ 50%).
| Variable | OR | Low 95% CI | High 95% CI | p |
|---|---|---|---|---|
| Age > 40 years | 1.450 | 0.485 | 4.331 | 0.5058 |
| Hb < 10 g/dL | 0.714 | 0.240 | 2.121 | 0.5446 |
| MCV < 100 fL | 2.780 | 0.738 | 10.472 | 0.1309 |
| ANC < 1.5 ×109/L | 1.125 | 0.309 | 4.099 | 0.8583 |
| Platelets < 100 ×109/L | 1.371 | 0.429 | 4.374 | 0.5945 |
| Reticulocytes > 100 ×109/L | 0.446 | 0.104 | 1.905 | 0.2757 |
| LDH > 400 U/L | 2.167 | 0.257 | 18.263 | 0.4772 |
| Total bilirubin < 12 μmol/L | 2.444 | 0.716 | 8.351 | 0.1539 |
| Creatinine > 1 mg/dL | 0.937 | 0.186 | 4.733 | 0.9373 |
| Neutrophil PNH total clone size > 70% | 1.247 | 0.260 | 5.981 | 0.7822 |
| Relative % of PNH Type II neutrophils (≥ 3%) | 5.659 | 1.790 | 17.890 | 0.0032 |
Note: The threshold values of the variables were those used by Gurnari and colleagues [17]. Bold values indicate statistical significance.
TABLE 4.
Multivariate analysis of predictors of thromboembolic events in PNH patients (PNN ≥ 50%).
| Variable | OR | Low 95% CI | High 95% CI | p |
|---|---|---|---|---|
| Age > 40 years | 1.157 | 0.356 | 3.756 | 0.8084 |
| Hb < 10 g/dL | 0.695 | 0.212 | 2.285 | 0.5492 |
| MCV < 100 fL | 2.367 | 0.592 | 9.457 | 0.2229 |
| Platelets < 100 ×109/L | 1.182 | 0.335 | 4.172 | 0.7945 |
| Neutrophil PNH total clone size > 70% | 1.447 | 0.272 | 7.703 | 0.6649 |
| Relative % of PNH Type II neutrophils (≥ 3%) | 4.319 | 1.304 | 14.311 | 0.0167 |
Note: Variables with a high proportion of missing data (> 25%) were not included in the multivariate analysis in order to preserve sample size and model robustness. In addition, these variables showed no clinically meaningful association with the occurrence of thrombosis in univariable analysis (ANC, reticulocytes, LDH, total bilirubin and creatinine). This approach allowed the inclusion of 112 out of 127 patients (88%) in the multivariate analysis. The threshold values of the variables were those used by Gurnari and colleagues [17]. Bold values indicate statistical significance.
We then focused on 207 out of the 222 patients with a lower PNH clone size between 1% and 50% and evaluable Type II neutrophil percentages (Figure 2B; Figure S1). The rel% threshold of 3% allowed the distinction of 87 with high Type II rel% neutrophils (11.96%, IQR, 5.77%–28.57%) and 120 patients with low Type II rel% neutrophils (0.00%, IQR, 0.00%–0.26%; p < 0.0001). Similarly, Type II monocyte rel% were higher in the “high Type II” group (7.05%, IQR, 0.36%–21.20%) compared to the “low Type II” group (0.00%, IQR, 0.00%–0.44%; p < 0.0001), as well as Type II RBC rel% (10.10%, IQR, 0.00%–37.50% vs. 0.00%, IQR, 0.00%–4.81%; p = 0.0005) (Table 1). Conversely, the clone size was lower on WBCs in the “high Type II” group compared to the “low Type II” group (neutrophils: 4.74%, IQR, 1.81%–11.53% vs. 6.57%, IQR, 2.62%–17.66%; p = 0.0289, monocytes: 5.01%, IQR, 2.13%–13.54% vs. 9.20%, IQR, 3.06%–21.85%; p = 0.0091), while similar on RBCs (0.80%, IQR, 0.24%–2.67% vs. 1.06%, IQR, 0.35%–5.52%; p = 0.4475) (Figure 3A,B right panels, Table 1).
The median age of these patients at first PNH clone detection was similar in both groups (46 y, IQR, 29–62 y vs. 42 y, IQR, 29–61 y; p = 0.3177) and we observed a higher proportion of women in the high Type II group (male‐to‐female ratio, 0.45 vs. 0.88; p = 0.0235).
As for the patients with major PNH clone, hemolysis was less frequent in “high Type II” group compared to the “less Type II” group (19% vs. 40%; p < 0.01). All the other CBC parameters, but platelet count, were similar between the two patient groups. A more marked thrombopenia was indeed observed in the patients from the “low Type II” group (30 × 109/L, IQR, 12–60 × 109/L vs. 44 × 109/L, IQR, 22–83 × 109/L, IQR; p = 0.0432) (Table 1). However, bone marrow failure was not more frequent in the “high Type II” group than in the “low Type II” group (94.0% vs. 85.7%; p = 0.0668).
While no thrombosis was present at diagnosis, two thromboses occurred during the follow‐up of the “high Type II” group and none in the “low Type II” group (Figure 4). At the time of the two venous thrombotic events, both cases showed a marked expansion of the neutrophil clone, with total PNH clone size rising above 50% (21.77%–88.06% and 4.74%–59.20%).
3.5. Type II Neutrophil Relative Percentages Remain Stable Despite Clone Size Changes, Providing a Reliable Predictor of Thrombosis
Long‐term follow‐up up to 14.2 years of 138 out of the 355 patients with a PNH clone ≥ 1%, of whom 69 (50.0%) were receiving anti‐C5, showed the three classical patterns of clone size evolution. Sixty‐one patients (44.2%), of whom 32 (52.5%) were receiving anti‐C5, displayed an increase of the total PNH clone size in neutrophils (defined by a relative increase above 5% in the total clone size), (Figure 5A left panel), while 35 (25.4%) of whom 29 (82.9%) were receiving anti‐C5, remained stable (i.e., with a relative variation between −5% and +5% of the total clone size) (Figure 5B left panel) and 42 (30.4%), of whom 8 (19.0%) were receiving anti‐C5 and 12 (28.6%) underwent bone marrow transplantation, exhibited a decrease (defined by a relative decrease above 5% in the total clone size) (Figure 5C left panel). No major modification in Type II neutrophil rel% was observed, irrespective of changes in the total clone size (Figure 5 right panels).
FIGURE 5.
Evolution of total PNH clone size on neutrophils (left panels) and relative percentage of PNH Type II neutrophils (right panels) over time in 138 patients. Among the 69 patients (49.6%) who received anti‐C5, 32 (46.4%) showed an increase in the total PNH clone size in neutrophils (A), 29 (42.0%) were stable (B), and 8 (11.6%) showed a decrease (C); y, years. [Color figure can be viewed at wileyonlinelibrary.com]
4. Discussion
We reported here a large, longitudinal, prospective multicenter study of patients with a clone above 1%, supported by robust flow cytometric data. The French laboratories involved in this Observatory were fully trained through an interlaboratory comparison program led by our dedicated CytHem‐HPN working group [35]. This program was conducted for 3 years prior to the opening of the study and continued throughout its 5‐year enrollment period [34]. Moreover, all the flow cytometry data provided by each center, both at inclusion and at follow‐up points for the patients they enrolled, were thoroughly reviewed by both principal investigators (O.W‐B and M.LG‐T). We successfully mobilized the French flow cytometry community to actively participate in this study, as illustrated by the steady pace of inclusions. The absence of follow‐up during the 5‐year enrollment period for the 23 patients registered early illustrates suboptimal adherence to international recommendations, particularly in patients with aplastic anemia (the main clinical setting in our study), who should undergo regular testing regardless of whether a clone is absent, present, or newly emerging [27].
Regarding our multicenter cohort of 355 French patients with a PNH clone ≥ 1%, demographic and clinical features resembled those of a large retrospective monocentric cohort of 1081 British patients collected over 25 years [33]. The Leeds study reported a median age of 49 years (range, 1–91) versus 42 years (range, 8–88) in our study. However, the male‐to‐female ratio (1.04:1) was higher than that observed in our study (0.76:1). We described two major categories of features underlying PNH clone detection, cytopenia(s) and hemolytic anemia, consistent with findings from other retrospective studies conducted either in Europe [33] or in China [36] or in South America [37].
In line with previous reports [3, 17, 38], all the patients in our study exhibited a total clone size on granulocytes exceeding 50% at the time of thrombosis. Here, we described for the first time the relevance of accurately quantifying PNH Type II cells among WBCs, particularly within the neutrophils. We demonstrated that a high proportion of PNH Type II neutrophils, defined as a relative percentage ≥ 3%, is independently associated with thrombotic events, although the absolute number of thrombotic events was low. We also found that “high Type II” patients displayed a higher proportion of PNH Type II among monocytes and RBCs when sufficient data were available. Even though this observational study was not designed to specifically assess thrombotic risk and potential biases may exist, we believe our findings are accurate, as they are based on robust multicentric flow cytometric data. Although the collected clinical and biological data related to the thrombotic events were limited, we focused on the 17 patients who experienced a thrombotic event in our study and asked the participating center that had enrolled them to provide additional information from both prior and after the 5‐year opening period in order to generate Panel C of Figure 4.
Moreover, our results are consistent with previous studies reporting that patients with the highest proportions of Type II RBCs were at increased risk of thrombosis [17, 32, 33]. Notably, Type II cells among WBCs were not analyzed in those studies. The quantification of Type II neutrophils remains accurate regardless of the clinical and/or therapeutic context, whereas the proportion of Type II cells within RBC may be influenced by hemolysis, with preferential destruction of Type III cells, and by transfusions. Furthermore, we demonstrated that the proportion of Type II cells on neutrophils remained stable over time, irrespective of changes in total clone size. Thus, even if the relative percentage of Type II neutrophils was not determined at diagnosis, this parameter can be assessed during follow‐up and may therefore serve as a reliable predictor of thrombotic event.
As thrombotic events in PNH are multifactorial, the mechanism relying on the thrombotic risk linked to a relative proportion of PNH Type II cells remains unexplained. Several mechanisms may be involved. Neutrophil extracellular trap (NET) formation, that is a distinct antimicrobial defense in which neutrophils release chromatin and granular proteins to trap pathogens, provides, within the vasculature, a scaffold for platelet adhesion and fibrin deposition that promotes thrombus formation, a process termed immunothrombosis [39]. Moreover, the presence of a high proportion of Type II cells across multiple lineages implies that the defect in GPI‐anchored proteins extends to platelets, which may in turn contribute to an increased thrombotic risk [40]. Patients who have both PNH II and PNH III cells may have at least two PNH clones [41]. Additionally, recent evidence indicates that deregulation of specific genes may contribute to an increased risk of thrombosis in PNH patients [42]. A more detailed analysis of gene deregulation specifically in Type II WBCs would provide valuable insight.
As anticomplement therapy has been proven effective to prevent thromboembolic complications in PNH [14], accurate quantification of Type II neutrophils may influence clinical management, since it could support consideration of treatment initiation, particularly in patients with a PNH clone below 50% for whom the current therapeutic guidelines are less clear. In such cases, a relative proportion of PNH Type II neutrophils exceeding 3% could provide an additional argument to support the clinician's decision to initiate therapy. Current indications for antiC5 therapy (eculizumab or ravulizumab) include transfusion‐dependent hemolysis, arterial or venous thrombosis, a large PNH clone (> 50%), or pregnancy [43, 44]. At present, there is no indication for prophylactic antithrombotic treatment in PNH patients with a clone size < 50%, as no reliable predictive risk factors for thrombosis have yet been identified. Of note, a greater proportion of our cohort consisted of patients with a clone size below 50%, in whom the Type II neutrophils were more frequent, as they exhibited less hemolytic anemia than patients with larger PNH clones ≥ 50%.
There are no current guidelines for the precise gating of PNH Type II cells, the latest international recommendations to report only the total PNH clone size in neutrophils and monocytes [30]. The delineation of such cells, even with the use of the FLAER reagent, could be tricky and was part of the collegial decision of the two principal investigators when they double‐checked the raw flow cytometry data files. There is a need to educate the clinical laboratories for harmonization before reporting such parameters, as the French interlaboratory comparison program showed a higher coefficient of variation for the Type II populations [35].
A dedicated prospective study is needed to validate our results, yet could be difficult to lead as PNH is a rare disease and most of the patients are currently being treated by anti‐complement therapy [17], and decide whether the reporting of PNH Type II neutrophil relative percentages may be included in the report to the clinician in future guidelines. In the meantime, in agreement with the French reference center (Prof Regis Peffault de Latour), we have decided that the strong network of cytometrists we built for many years [35] might be of help. Where there is a suspicion of high Type II neutrophils in patients with a clone between 1% and 50%, the French clinical laboratory might ask the three centers with the French label of reference laboratory within the French PNH working group to reanalyze the raw flow cytometry data and even retest the sample if needed, in order to help the clinician to initiate anticomplement therapy.
Author Contributions
Orianne Wagner‐Ballon and Magali Le Garff‐Tavernier designed the study, contributed to the design of the eCRF, collected the data, both verified all the data including flow cytometry data, analyzed the data and wrote the manuscript; Anne‐Catherine Lhoumeau, Anna Raimbault, Camille Debord, Jean‐Philippe Vial, Valérie Soenen, Mélanie Pannetier, Edouard Bonneville, Elsa Bera, Caroline Bret, Nicolas Chapuis, Véronique Latger‐Cannard, Caroline Mayeur‐Rousse, Véronique Harrivel, François Vergez, Marie Loosveld, Juliette Vo Xuan, and Stéphanie Mathis enrolled the patients and completed the eCRF; Gérard Socié, Flore Sicre de Fontbrune, and Régis Peffault de Latour contributed to the design of the clinical section of the eCRF and critically reviewed the manuscript; Bernard Drenou and Agathe Debliquis contributed to the design of the study and the eCRF.
Funding
This study was supported by Alexion Pharmaceuticals France.
Ethics Statement
The study was approved by the Consultative Committee for the Data Processing in Health Research (CCTIRS), authorized by the French Data Protection Authority (CNIL). As this study was observational, human participants did not have to give written informed consent, according to French regulatory rules.
Conflicts of Interest
O.W.‐B. received honoraria for conferences and symposia from Alexion Pharma France, Alexion Pharma Spain, and Novartis, and grants for research from Alexion Pharma France, Novartis, and Sobi. F.V. received honoraria for conferences and symposia from Alexion Pharma France. M.L. received honoraria for conferences and symposia from Sanofi Genzyme and Alexion Pharma France. J.V.X. received grants for research from Alexion Pharma France, Novartis, and Sobi. F.S.d.F. received grants from Pfizer, Novartis, Alexion, Sobi, and honoraria for lectures, presentations, boards, manuscript writing or educational events from Alexion, Sobi, Novartis, Samsung. R.P.d.L received honoraria for conferences and symposia from Alexion Pharma, Novartis, Sobi, and Pfizer as well as research grants. B.D. received honoraria for boards from Alexion Pharma France. A.D. received honoraria for conferences and boards from Alexion, Beckman Coulter, Novartis, and Sobi. M.L.G.‐T. received honoraria for conferences and symposia from Alexion Pharma France, Novartis, Janssen, and Abbvie, and grants for research from Alexion Pharma France. The other authors declare no conflicts of interest.
Supporting information
Figure S1: Distribution of the Type II neutrophil relative percentages across the entire cohort of 334 patients, including 127 patients with a major PNH clone ≥ 50% (in blue) and 207 patients with a PNH clone size between 1% and 50% (in orange). The inset illustrates the distribution between 2% and 4%.
Acknowledgments
This study was supported by Alexion Pharmaceuticals France. The authors are indebted to the webmaster Aurélien Richaud for the eCRF development and maintenance, and to Florian Rigaudier from CEN CRO for the statistical analysis. The authors thank the following members of the French working group dedicated to PNH, the “CytHem‐HPN”, for having enrolled a few patients: Prof. Thomas Boyer, Prof. Mikaël Roussel, Dr. Marie Guillon, Dr. Sabrina Bouyer, Dr. Isabelle Kittler, Dr. Franck Geneviève, Dr. Julien Guy, Dr. Sophie Brun, Dr. Claude Capron, Dr. Sébastien Lachot, Prof. Francine Garnache‐Ottou, Dr. Richard Veyrat‐Masson, Dr. Corine Brunet, Dr. Caren Brumpt, and Dr. Ludovic Lhermitte. Finally, the authors are grateful to Dr. Louis Terriou and Dr. Nicolas Blin for enlightening discussions. Open access publication funding provided by COUPERIN CY26.
The authors used ChatGPT (OpenAI, San Francisco, CA, USA) to assist with language editing and improving clarity of the manuscript. All content was reviewed and approved by the authors, who take full responsibility for the final version.
Contributor Information
Orianne Wagner‐Ballon, Email: orianne.wagnerballon@aphp.fr.
the French PNH working group CytHem‐HPN:
Thomas Boyer, Mikaël Roussel, Marie Guillon, Sabrina Bouyer, Isabelle Kittler, Franck Geneviève, Julien Guy, Sophie Brun, Claude Capron, Sébastien Lachot, Francine Garnache‐Ottou, Richard Veyrat‐Masson, Corine Brunet, Caren Brumpt, Ludovic Lhermitte. Finally, Louis Terriou, and Nicolas Blin
Data Availability Statement
For original data, please contact orianne.wagnerballon@aphp.fr or magali.legarff@aphp.fr.
References
- 1. Hillmen P., Lewis S. M., Bessler M., Luzzatto L., and Dacie J. V., “Natural History of Paroxysmal Nocturnal Hemoglobinuria,” New England Journal of Medicine 333, no. 19 (1995): 1253–1258, 10.1056/NEJM199511093331904. [DOI] [PubMed] [Google Scholar]
- 2. Parker C., Omine M., Richards S., et al., “Diagnosis and Management of Paroxysmal Nocturnal Hemoglobinuria,” Blood 106, no. 12 (2005): 3699–3709, 10.1182/blood-2005-04-1717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. de Latour R. P., Mary J. Y., Salanoubat C., et al., “Paroxysmal Nocturnal Hemoglobinuria: Natural History of Disease Subcategories,” Blood 112, no. 8 (2008): 3099–3106, 10.1182/blood-2008-01-133918. [DOI] [PubMed] [Google Scholar]
- 4. Moyo V. M., Mukhina G. L., Garrett E. S., and Brodsky R. A., “Natural History of Paroxysmal Nocturnal Haemoglobinuria Using Modern Diagnostic Assays,” British Journal of Haematology 126, no. 1 (2004): 133–138, 10.1111/j.1365-2141.2004.04992.x. [DOI] [PubMed] [Google Scholar]
- 5. Luzzatto L., Gianfaldoni G., and Notaro R., “Management of Paroxysmal Nocturnal Haemoglobinuria: A Personal View,” British Journal of Haematology 153, no. 6 (2011): 709–720, 10.1111/j.1365-2141.2011.08690.x. [DOI] [PubMed] [Google Scholar]
- 6. Hill A., Kelly R. J., and Hillmen P., “Thrombosis in Paroxysmal Nocturnal Hemoglobinuria,” Blood 121, no. 25 (2013): 4985–4996, 10.1182/blood-2012-09-311381. [DOI] [PubMed] [Google Scholar]
- 7. Ziakas P. D., Poulou L. S., Rokas G. I., Bartzoudis D., and Voulgarelis M., “Thrombosis in Paroxysmal Nocturnal Hemoglobinuria: Sites, Risks, Outcome. An Overview,” Journal of Thrombosis and Haemostasis 5, no. 3 (2007): 642–645, 10.1111/j.1538-7836.2007.02379.x. [DOI] [PubMed] [Google Scholar]
- 8. Gurnari C., Colak C., Zawit M., and Maciejewski J. P., “Chronic Budd–Chiari Syndrome in Paroxysmal Nocturnal Haemoglobinuria,” Lancet 398, no. 10304 (2021): e14, 10.1016/S0140-6736(21)01542-7. [DOI] [PubMed] [Google Scholar]
- 9. Luzzatto L. and Nakao S., “Pathogenesis of Paroxysmal Nocturnal Hemoglobinuria,” Blood 145, no. 26 (2025): 3077–3088, 10.1182/blood.2024025975. [DOI] [PubMed] [Google Scholar]
- 10. Sims P. J., Rollins S. A., and Wiedmer T., “Regulatory Control of Complement on Blood Platelets. Modulation of Platelet Procoagulant Responses by a Membrane Inhibitor of the C5b‐9 Complex,” Journal of Biological Chemistry 264, no. 32 (1989): 19228–19235. [PubMed] [Google Scholar]
- 11. Wiedmer T., Hall S. E., Ortel T. L., Kane W. H., Rosse W. F., and Sims P. J., “Complement‐Induced Vesiculation and Exposure of Membrane Prothrombinase Sites in Platelets of Paroxysmal Nocturnal Hemoglobinuria,” Blood 82, no. 4 (1993): 1192–1196. [PubMed] [Google Scholar]
- 12. Van Bijnen S. T. A., Van Heerde W. L., and Muus P., “Mechanisms and Clinical Implications of Thrombosis in Paroxysmal Nocturnal Hemoglobinuria,” Journal of Thrombosis and Haemostasis 10, no. 1 (2012): 1–10, 10.1111/j.1538-7836.2011.04562.x. [DOI] [PubMed] [Google Scholar]
- 13. Hillmen P., Young N. S., Schubert J., et al., “The Complement Inhibitor Eculizumab in Paroxysmal Nocturnal Hemoglobinuria,” New England Journal of Medicine 355, no. 12 (2006): 1233–1243, 10.1056/NEJMoa061648. [DOI] [PubMed] [Google Scholar]
- 14. Hillmen P., Muus P., Dührsen U., et al., “Effect of the Complement Inhibitor Eculizumab on Thromboembolism in Patients With Paroxysmal Nocturnal Hemoglobinuria,” Blood 110, no. 12 (2007): 4123–4128, 10.1182/blood-2007-06-095646. [DOI] [PubMed] [Google Scholar]
- 15. Loschi M., Porcher R., Barraco F., et al., “Impact of Eculizumab Treatment on Paroxysmal Nocturnal Hemoglobinuria: A Treatment Versus No‐Treatment Study,” American Journal of Hematology 91, no. 4 (2016): 366–370, 10.1002/ajh.24278. [DOI] [PubMed] [Google Scholar]
- 16. Weitz I. C., Razavi P., Rochanda L., et al., “Eculizumab Therapy Results in Rapid and Sustained Decreases in Markers of Thrombin Generation and Inflammation in Patients With PNH Independent of Its Effects on Hemolysis and Microparticle Formation,” Thrombosis Research 130, no. 3 (2012): 361–368, 10.1016/j.thromres.2012.04.001. [DOI] [PubMed] [Google Scholar]
- 17. Gurnari C., Awada H., Pagliuca S., et al., “Paroxysmal Nocturnal Hemoglobinuria‐Related Thrombosis in the Era of Novel Therapies: A 2043‐Patient‐Year Analysis,” Blood 144, no. 2 (2024): 145–155, 10.1182/blood.2024023988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Kelly R. J., Holt M., Vidler J., et al., “Treatment Outcomes of Complement Protein C5 Inhibition in 509 UK Patients With Paroxysmal Nocturnal Hemoglobinuria,” Blood 143, no. 12 (2024): 1157–1166, 10.1182/blood.2023021762. [DOI] [PubMed] [Google Scholar]
- 19. Takeda J., Miyata T., Kawagoe K., et al., “Deficiency of the GPI Anchor Caused by a Somatic Mutation of the PIG‐A Gene in Paroxysmal Nocturnal Hemoglobinuria,” Cell 73, no. 4 (1993): 703–711. [DOI] [PubMed] [Google Scholar]
- 20. Kinoshita T., “Biosynthesis and Biology of Mammalian GPI‐Anchored Proteins,” Open Biology 10, no. 3 (2020): 190290, 10.1098/rsob.190290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Bessler M., Mason P. J., Hillmen P., et al., “Paroxysmal Nocturnal Haemoglobinuria (PNH) is Caused by Somatic Mutations in the PIG‐A Gene,” EMBO Journal 13, no. 1 (1994): 110–117, 10.1002/j.1460-2075.1994.tb06240.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Hall S. E. and Rosse W. F., “The Use of Monoclonal Antibodies and Flow Cytometry in the Diagnosis of Paroxysmal Nocturnal Hemoglobinuria,” Blood 87, no. 12 (1996): 5332–5340. [PubMed] [Google Scholar]
- 23. Hillmen P., Bessler M., Mason P. J., Watkins W. M., and Luzzatto L., “Specific Defect in N‐Acetylglucosamine Incorporation in the Biosynthesis of the Glycosylphosphatidylinositol Anchor in Cloned Cell Lines From Patients With Paroxysmal Nocturnal Hemoglobinuria,” Proceedings of the National Academy of Sciences of the United States of America 90, no. 11 (1993): 5272–5276, 10.1073/pnas.90.11.5272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Holguin M. H., Fredrick L. R., Bernshaw N. J., Wilcox L. A., and Parker C. J., “Isolation and Characterization of a Membrane Protein From Normal Human Erythrocytes That Inhibits Reactive Lysis of the Erythrocytes of Paroxysmal Nocturnal Hemoglobinuria,” Journal of Clinical Investigation 84, no. 1 (1989): 7–17, 10.1172/JCI114172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Richards S. J., Rawstron A. C., and Hillmen P., “Application of Flow Cytometry to the Diagnosis of Paroxysmal Nocturnal Hemoglobinuria,” Cytometry 42, no. 4 (2000): 223–233. [PubMed] [Google Scholar]
- 26. Borowitz M. J., Craig F. E., Digiuseppe J. A., et al., “Guidelines for the Diagnosis and Monitoring of Paroxysmal Nocturnal Hemoglobinuria and Related Disorders by Flow Cytometry,” Cytometry Part B, Clinical Cytometry 78, no. 4 (2010): 211–230, 10.1002/cyto.b.20525. [DOI] [PubMed] [Google Scholar]
- 27. Dezern A. E. and Borowitz M. J., “ICCS/ESCCA Consensus Guidelines to Detect GPI‐Deficient Cells in Paroxysmal Nocturnal Hemoglobinuria (PNH) and Related Disorders Part 1 ‐ Clinical Utility,” Cytometry. Part B, Clinical Cytometry 94, no. 1 (2018): 16–22, 10.1002/cyto.b.21608. [DOI] [PubMed] [Google Scholar]
- 28. Packman C. H., Rosenfeld S. I., Jenkins D. E., Thiem P. A., and Leddy J. P., “Complement Lysis of Human Erythrocytes. Differeing Susceptibility of Two Types of Paroxysmal Nocturnal Hemoglobinuria Cells to C5b‐9,” Journal of Clinical Investigation 64, no. 2 (1979): 428–433, 10.1172/JCI109479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Brodsky R. A., Mukhina G. L., Li S., et al., “Improved Detection and Characterization of Paroxysmal Nocturnal Hemoglobinuria Using Fluorescent Aerolysin,” American Journal of Clinical Pathology 114, no. 3 (2000): 459–466, 10.1093/ajcp/114.3.459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Illingworth A., Marinov I., Robert Sutherland D., Wagner‐Ballon O., and DelVecchio L., “ICCS/ESCCA Consensus Guidelines to Detect GPI‐Deficient Cells in Paroxysmal Nocturnal Hemoglobinuria (PNH) and Related Disorders Part 3 ‐ Data Analysis, Reporting and Case Studies,” Cytometry Part B, Clinical Cytometry 94, no. 1 (2018): 49–66, 10.1002/cyto.b.21609. [DOI] [PubMed] [Google Scholar]
- 31. Hall C., Richards S., and Hillmen P., “Primary Prophylaxis With Warfarin Prevents Thrombosis in Paroxysmal Nocturnal Hemoglobinuria (PNH),” Blood 102, no. 10 (2003): 3587–3591, 10.1182/blood-2003-01-0009. [DOI] [PubMed] [Google Scholar]
- 32. Griffin M., Hillmen P., Munir T., et al., “Significant Hemolysis Is Not Required for Thrombosis in Paroxysmal Nocturnal Hemoglobinuria,” Haematologica 104, no. 3 (2019): e94–e96, 10.3324/haematol.2018.198846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Richards S. J., Dickinson A. J., Cullen M. J., et al., “Presentation Clinical, Haematological and Immunophenotypic Features of 1081 Patients With GPI‐Deficient (Paroxysmal Nocturnal Haemoglobinuria) Cells Detected by Flow Cytometry,” British Journal of Haematology 189, no. 5 (2020): 954–966, 10.1111/bjh.16427. [DOI] [PubMed] [Google Scholar]
- 34. Debliquis A., Wagner‐Ballon O., Le Garff‐Tavernier M., et al., “Evaluation of Paroxysmal Nocturnal Hemoglobinuria Screening by Flow Cytometry Through Multicentric Interlaboratory Comparison in Four Countries,” American Journal of Clinical Pathology 144, no. 6 (2015): 858–868, 10.1309/AJCPWMWJ7XDNQ3SN. [DOI] [PubMed] [Google Scholar]
- 35. Wagner‐Ballon O., Le Garff‐Tavernier M., Lestestu R., Drenou B., Debliquis A., and PNH French Working Group «CytHem‐HPN» , “The PNH French Working Group Experience: Building a Strong Network of Cytometrists,” International Journal of Laboratory Hematology 48, no. 2 (2026): 242–249, 10.1111/ijlh.14449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Du Y., Yang Y., Yang C., Chen M., and Han B., “Clinical Characteristics of 512 Eculizumab‐Naive Paroxysmal Nocturnal Hemoglobinuria Patients in China: A Single‐Center Observational Study,” Hematology 27, no. 1 (2022): 113–121, 10.1080/16078454.2021.2022849. [DOI] [PubMed] [Google Scholar]
- 37. de Azambuja A. P., Oliveira M. M., Bitencourt M. A., Bonfim C., Malvezzi M., and Pasquini R., “Analysis of Baseline Characteristics, Disease Burden and Long‐Term Follow‐Up of 167 Patients With Paroxysmal Nocturnal Hemoglobinuria at a Single Center in Brazil,” Blood Cells, Molecules & Diseases 92 (2021): 102605, 10.1016/j.bcmd.2021.102605. [DOI] [PubMed] [Google Scholar]
- 38. Höchsmann B., Peffault de Latour R., Hill A., et al., “Risk Factors for Thromboembolic Events in Patients With Paroxysmal Nocturnal Hemoglobinuria (PNH): A Nested Case‐Control Study in the International PNH Registry,” Annals of Hematology 102, no. 11 (2023): 2979–2988, 10.1007/s00277-023-05402-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Engelmann B. and Massberg S., “Thrombosis as an Intravascular Effector of Innate Immunity,” Nature Reviews. Immunology 13, no. 1 (2013): 34–45, 10.1038/nri3345. [DOI] [PubMed] [Google Scholar]
- 40. Gurnari C., Graham A. C., Efanov A., et al., “Frequency and Perturbations of Various Peripheral Blood Cell Populations Before and After Eculizumab Treatment in Paroxysmal Nocturnal Hemoglobinuria,” Blood Cells, Molecules & Diseases 87 (2021): 102528, 10.1016/j.bcmd.2020.102528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Bessler M., Mason P. J., Hillmen P., and Luzzatto L., “Mutations in the PIG‐A Gene Causing Partial Deficiency of GPI‐Linked Surface Proteins (PNH II) in Patients With Paroxysmal Nocturnal Haemoglobinuria,” British Journal of Haematology 87, no. 4 (1994): 863–866. [DOI] [PubMed] [Google Scholar]
- 42. Du Y., Wang D., Hu Q., et al., “Identifying Genes Associated With Proliferation, Immunity and Thrombosis in Paroxysmal Nocturnal Haemoglobinuria,” Journal of Cellular and Molecular Medicine 28, no. 23 (2024): e70295, 10.1111/jcmm.70295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Brodsky R. A., “How I Treat Paroxysmal Nocturnal Hemoglobinuria,” Blood 137, no. 10 (2021): 1304–1309, 10.1182/blood.2019003812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Kelly R. J., Höchsmann B., Szer J., et al., “Eculizumab in Pregnant Patients With Paroxysmal Nocturnal Hemoglobinuria,” New England Journal of Medicine 373, no. 11 (2015): 1032–1039, 10.1056/NEJMoa1502950. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Distribution of the Type II neutrophil relative percentages across the entire cohort of 334 patients, including 127 patients with a major PNH clone ≥ 50% (in blue) and 207 patients with a PNH clone size between 1% and 50% (in orange). The inset illustrates the distribution between 2% and 4%.
Data Availability Statement
For original data, please contact orianne.wagnerballon@aphp.fr or magali.legarff@aphp.fr.