Reversible skin microvascular hyporeactivity in patients with immune-mediated thrombocytopenic thrombotic purpura

Acquired or immune-mediated thrombotic thrombocytopenic purpura (TTP) is a rare thrombotic microangiopathy characterized by thrombocytopenia and hemolytic anemia [1‐3]. Immune TTP is due to the presence of neutralizing anti-ADAMTS-13 autoantibodies responsible for impaired cleavage of the von Willebrand factor (VWF) mega multimers [4, 5]. Ultimately, VWF–platelet aggregates provoke microvascular thrombosis leading to inadequate microvascular blood flow, tissue ischemia and multiorgan failure. Therefore, the iTTP pattern combines hemolytic anemia, thrombocytopenia, often with neurologic, cardiac, or renal abnormalities, still associated with a 10–20% death rate [6, 7]. Current treatment consists of plasma exchange (PE) [8‐10] combined with immunosuppressive therapy (e.g., glucocorticoids and rituximab) and caplacizumab [11], an anti–VWF factor monoclonal humanized antibody inhibiting interaction between VWF multimers and platelets [12, 13].

Arteriolar and capillary microthrombosis due to the accumulation of VWF–platelet aggregates lead to life-threatening organ hypoperfusion, affecting the heart and the brain. However, the consequences of VWF–platelet aggregates on the endothelium, a key regulator of blood flow, remain unknown. This observational study aimed to explore endothelial-dependent microvascular reactivity in iTTP patients in the intensive care unit (ICU) at admission and after treatment.

We conducted a prospective observational study in our tertiary university hospital. We included iTTP patients (> 18-year old) experiencing their first acute event, admitted to our ICU between January 2016 and September 2022. Clinical and biological parameters were recorded. Skin microcirculatory reactivity in the right forearm area was recorded at ICU admission (baseline) and after PE. As a control group for microvascular reactivity, we used data from a previously published cohort of young diabetic patients recorded after the correction of metabolic acidosis [14]. The local ethical committee approved the protocol (Comité de Protection des Personnes, Hôpital Saint-Louis, France, No 2015/64NI), and the database was registered according to the French legislation (No 2,228,742), and all patients consented to anonymous data use for academic research and publication. It was a noninvasive observational study without any specific intervention. All patients were managed following international guidelines for TTP [15] and in collaboration with the physician of the thrombotic microangiopathy national reference center. All patients received urgent therapeutic PE (1.5 plasmatic mass, 100% fresh frozen plasma (FFP)), corticosteroids, and 17/18 patients received caplacizumab on the first day of ICU admission.

Eighteen consecutive iTTP patients were included, 72% were female, aged 43 ± 16-year-old. ADAMTS13 activity at baseline was below 10% in all included patients [16] (below 5% in 15 patients (83.3%)) and all included patients had circulating anti-ADAMTS13 autoantibodies, which confirmed the final diagnosis of iTTP. At ICU admission, 55% had neurological abnormalities, 50% cardiac (troponin elevation, EKG, or echocardiography) issues and 27.8% had stage ⩾1 acute kidney injury according to KDIGO classification [17]. At baseline, all included patients had severe thrombocytopenia with median platelet count at 19 G/mL [10–37] G/L, mild regenerative anemia (hemoglobin: 9.6 g/dl [7.6–10], reticulocytes 178 [128–285]G/L) associated with hemolysis markers (median bilirubin: 36 µmol/L [23–59], haptoglobin: 0 g/dL [0–0.035]). Only one patient was under mechanical ventilation and received vasopressors. None received renal replacement therapy. All patients were treated by PE (median number of PE: 4 [2–5]), 100% corticosteroids (methylprednisolone 1 mg/Kg/Day i.v), and 88.9% caplacizumab (one injection i.v. before the first PE, then 10 mg/Day s.c). Patients' baseline characteristics are reported in Table 1 and initial treatment and ICU stay characteristics are summarized in Additional file 1: Table S1. Overall, the in-ICU length of stay was 6.4 [4.8–8] days, and one patient died in ICU (5.5%).

First, we compared the microvascular reactivity of iTTP patients with a cohort of diabetic patients admitted to our ICU after correction of keto-acidosis. Such a control cohort was relevant because patients were young with rare co-morbidities (Table 1) and no severe organ failure. At admission, when compared to the control group (Fig. 1 and Additional file 1: Table S2), we observed that iTTP patients had twofold lower skin microvascular blood flow (5.97 ± 4.5 vs. 10.1 ± 6.3 PU, P = 0.03) (Fig. 1A). In addition, we found marked impaired endothelial-mediated microvascular reactivity in iTTP patients characterized by a lower peak after Ach iontophoresis (31.9 ± 19.1 vs. 67.7 ± 39.9, P = 0.001) (Fig. 1B) and ultimately a lower AUC (9627 ± 8122 vs. 16,475 ± 11,738, P = 0.03) (Fig. 1C) (Fig. 2).

Next, on iTTP patients, we analyzed the impact of combined treatment on endothelial-dependent microvascular hyporeactivity during ICU stay. Acetylcholine iontophoresis was repeated after the first and the second PE session. After the first PE session, platelet count significantly increased (26 ± 28 vs 42 ± 38 G/mL, P = 0.0003) while hemolysis parameters improved (LDH 1870 ± 1440 vs. 650 ± 203 UI/mL, P < 0.0001, Haptoglobin 0.09 ± 0.2 vs. 0.56 ± 0.16 g/L, P < 0.0001) and biological recovery was more pronounced after the second session (Additional file 1: Table S1, Fig. 3 and Additional file 1: Fig. S2A). We observed that global microvascular blood flow significantly increased after the first PE session (Baseline perfusion index: 5.97 ± 4.5 PU at admission vs 11.38 ± 8.6 post-PE1, P = 0.027,) and even more after the second session (Baseline perfusion index 12.89 ± 6.9 PU, P = 0.008 vs admission). Global microvascular reactivity improved after the first PE session (AUC: 9627 ± 8122 vs 16,558 ± 10,699, P = 0.007, respectively, baseline and post-PE1) and much more after the second session (26,431 ± 23,181, P = 0.04 post-PE1 vs post-PE2) (Fig. 2A–C and Additional file 1: Table S2). Changes in microvascular reactivity after PE were heterogeneous, some patients improved after the first PE while others improved their microvascular reactivity after the second. Finally, few iTTP patients had no variation of skin microvascular response to Ach (Additional file 1: Fig. S3). Figure 4 shows an archetypical example of endothelium-mediated microvascular hyporeactivity in a single patient which improved after the plasma exchange session and much more after the second session. Interestingly, microvascular blood flow across time positively correlated with platelet but not the vasoreactivity (Additional file 1: Fig. S2A). Hemolysis biomarkers (LDH and bilirubin) negatively correlated with microvascular flow and reactivity (Fig. 3). Conversely, we observed no correlation between microvascular flow/ reactivity and hemoglobin, haptoglobin, schizocytes or reticulocytes (Additional file 1: Fig. S2B).

In this prospective study, we showed a markedly impaired skin microvascular endothelial-mediated reactivity in iTTP patients which recovered quickly after plasma exchange therapy.

This profoundly impaired microvascular vasoreactivity is similar to what our group previously observed in other critical conditions characterized by patent endothelial dysfunction such as critically ill COVID-19 [18] septic shock [19] or severe keto-acidosis [14]. Endothelial cell (EC) involvement in the pathophysiology of thrombotic microangiopathy-associated organ failure has been suggested in animal models but remains poorly demonstrated in humans. Indeed, experimental models indicated that Adamts-13 deficiency, by itself, is not sufficient to trigger thrombotic microangiopathy. Endothelial activation is another necessary step to induce microvascular disease, probably by releasing of a large amount of UL-VWF [20‐22]. In the same line, in primates, the injection of human anti-ADAMTS-13 neutralizing autoantibody provokes a transient biological thrombotic microangiopathy but not a severe disease responsible for organ failure [23]. In iTTP patients, circulating EC number and plasma biomarkers reflecting endothelial activation are increased and correlated with the outcome supporting a role of the endothelium in the pathophysiology of iTTP [24]. Recently, Tellier et al., reported that several plasmatic components, including anti-ADAMTS13 IgG, free heme and possibly others converge to induce EC activation ex vivo [25]. Interestingly the intensity of the ex vivo EC activation induced by iTTP patients' plasmas correlated with the disease severity [25]. Among soluble factors released in iTTP, hemolysis-derived products (including cell-free heme, free-hemoglobin and bilirubin at high concentration [26, 27]) are well known to be highly toxic for the endothelium [28, 29]. Moreover, in a murine model, Frei et al. reported that hemolysis causes direct vascular injury and functionally impaired vasodilation via increased scavenging of nitric oxide by plasma free hemoglobin [30]. Interestingly, in our study, we observed a negative correlation between endothelium-mediated skin microvascular vasodilation and hemolysis parameters.

We showed that the microvascular reactivity of iTTP patients is impaired and, therefore, could participate in organ injury besides the thrombotic process. Moreover, the endothelial vasoreactivity is restored after plasma exchange therapy at the same time than platelet count recovered. However, the improvement of endothelial reactivity could not be directly linked to exchange plasma therapy because at the same time, almost all the patient received additional treatment including steroids and Caplacizumab. Currently, there is no experimental data about the effect of Caplacizumab on endothelial microvascular reactivity [31].

We acknowledge some limitations to this observational and translational study. First, this is a single-center study with a limited number of patients. Second, given the multiple treatments received simultaneously and the synchronous correction of thrombocytopenia and hemolysis, one cannot speculate on which biological mechanism is responsible for the microcirculatory improvement. Next, as a control group, we used a previously published cohort of young diabetic patients after correction of ketoacidosis where we showed that microvascular hyporeactivity recovered after acidosis correction [14] Given that patients with cardiovascular risk factors are susceptible to a lower microvascular reactivity; difference between healthy subjects and iTTP patients could be even more important than differences between diabetic and iTTP patients [32‐35]. The device used in this study only allows exploration of the skin microvasculature and we did not investigate the endothelium of key organs affected by iTTP such as the brain, heart and kidney microcirculation [36, 37]. Finally, we did not explore the endothelium-independent vasodilation, which requires either heating or nitroprusside challenge. Thus, we cannot rule out that iTTP patients have impaired endothelium-independent vasodilation reserve [38] or decreased NO bioavailability [39], on top of the observed impaired endothelium-mediated vasoreactivity.

This prospective observational study highlights a marked endothelium-mediated microvascular hyporeactivity in acute iTTP patients that could participate in organ injury pathophysiology. Moreover, endothelium-mediated vasoreactivity dysfunction quickly recovered after PE therapy.