Etiology and outcomes: Thrombotic microangiopathies in pregnancy

Abstract

A State of the Art lecture titled “Etiology and Outcomes of Thrombotic Microangiopathies in Pregnancy” was presented at the International Society on Thrombosis and Haemostasis Congress in 2022. First, it is important to understand changes in laboratory parameters in normal pregnancy, including complement levels, specifically the increase in C3, C4, C3a, and C4a throughout pregnancy. Complement is critical in normal pregnancy for implantation and for placental development. Complement–mediated hemolytic uremic syndrome (CM-HUS) and thrombotic thrombocytopenic purpura (TTP) can present anytime from the first trimester to the postpartum period. In comparison, Thrombotic microangiopathies specific to pregnancy, such as preeclampsia (PET) or hemolysis, elevated liver enzymes, and low platelets (HELLP), present from the second trimester. C5b-9 deposition (following terminal complement pathway activation) is demonstrated in CM-HUS cases, and in HELLP and few PET cases. PET can also be confirmed and related to severity using soluble fms-like tyrosine kinase-1/placental growth factor ratios. Diagnosis of CM-HUS and TTP in pregnancy can be further complicated by clinical overlap at presentation with PET or occasionally HELLP. Management is aided by ADAMTS-13 analysis to confirm or exclude TTP. Treatment of CM-HUS, in conjunction with supportive care, is complement inhibitor therapy (eculizumab or ravulizumab). Acute TTP requires standard therapy, but caplacizumab should be avoided. Confirmation of congenital or immune subtypes informs care in subsequent pregnancies. Finally, we summarize relevant new data on this topic presented during the 2022 International Society on Thrombosis and Haemostasis Congress.

1. Introduction

Thrombotic microangiopathy (TMA) is a syndrome defined by thrombocytopenia, microangiopathic hemolytic anemia and end-organ damage. Definitions of the main conditions associated with TMA during pregnancy and the postpartum period have been presented by an international working group [1] (Figure 1). A distinction can be made between TMAs that can be triggered by pregnancy but also occur outside pregnancy, such as thrombotic thrombocytopenic purpura (TTP) and complement–mediated HUS (CM-HUS), (“pregnancy–associated” TMA), and secondly, “pregnancy–specific” TMAs, including preeclampsia (PET) and hemolysis, elevated liver enzymes, and low platelets (HELLP). HELLP is now generally regarded to represent a severe form of PET rather than a separate clinical entity.

Figure 1.

Adapted from Fakhouri et al. [1]

The pregnancy–specific TMAs are much more common than pregnancy–associated TTP and CM-HUS; about 5% of pregnancies are complicated by PET and 0.1% by HELLP, compared with incidences of TTP and CM-HUS in pregnancy of 1 in 200,000 and 1 in 25,000, respectively. However, TTP and CM-HUS are important to consider and diagnose as they are medical emergencies and have specific treatments that are different from pregnancy–specific TMAs.

2. Pathophysiology

The pathophysiological bases of TTP and CM-HUS are well-established and distinct. TTP is associated with a severe deficiency of ADAMTS-13 (a disintegrin and metalloproteinase with a Thrombospondin type 1 motif, member 13), the metalloprotease required for the cleavage of ultra-large von Willebrand factor (VWF) multimers from endothelial cells into smaller, less hemostatically active multimers. ADAMTS-13 deficiency is the result of either autoantibodies against ADAMTS-13 in the case of immune–mediated TTP or of pathogenic mutations in the ADAMTS-13 genes in the case of congenital TTP, which is much rarer in the general population. ADAMTS-13 has been shown to be present in trophoblasts and villous core fetal vessel endothelium during pregnancy, highest in the 1st trimester and reducing to term. However, ADAMTS-13 expression is significantly reduced, for example in PET [2]. CM-HUS meanwhile is associated with dysregulation of the alternative complement system, with an identifiable complement genetic mutation in 40% to 60% of cases. Review of an international cohort of pregnancy–associated CM-HUS cases [3] demonstrated similar demographic and clinical features as nonpregnancy–associated cases, supporting a shared complement–mediated basis.

The pathogenesis of PET and HELLP syndrome remains to be fully elucidated but deficient placentation is commonly cited as a root cause, with resulting hypoperfusion of the fetoplacental unit. This triggers release of mediators into the maternal circulation and causes a so-called angiogenic imbalance, leading to the maternal syndrome. Two now well-established angiogenic mediators are soluble fms-like tyrosine kinase-1 (sFlt-1) and placental growth factor (PIGF) ratio. PIGF has important proangiogenic effects on fetoplacental circulation and supports trophoblast growth. sFlt-1 is a soluble form of vascular endothelial growth factor receptor type 1 produced by the placenta, and at higher levels in preeclamptic pregnancies than in normotensive pregnancies [4]. sFlt-1 antagonizes the effects of both VEGF and PIGF. The ratio of sFLT-1/PIGF is therefore high in PET and has been shown to be useful diagnostically and prognostically. A ratio of sFlt-1/PIGF <38 excludes PET, between 38 and 85 is associated with mild PET, and >85 is associated with severe PET and indeed significant adverse pregnancy outcomes [5,6]. Furthermore, the ratio can be used in subsequent pregnancies to monitor for PET and predict the likeliness of its recurrence.

3. Clinical Presentation

There is significant clinical overlap however between pregnancy–associated TTP and CM-HUS and PET/HELLP syndrome, and diagnostic challenges can result from this.

Laboratory features must be interpreted in the context of pregnancy. Thrombocytopenia is not an uncommon finding in pregnancy, affecting 5% to 10% of pregnancies [7]. The most common cause is gestational thrombocytopenia, which accounts for around 75% of cases and generally presents in the second trimester, but it is only very rarely associated with a platelet count <75 × 10⁹/L. A total of 3% to 4% of cases are due to immune thrombocytopenia and are more likely to be associated with more severe and earlier-onset drops in platelet count. Approximately 50% of cases of PET are associated with thrombocytopenia, and this is the most common cause of thrombocytopenia with evidence of TMA. However, if the platelet nadir is <50 × 10⁹/L, TTP and CM-HUS should be excluded. At the same time, it should be noted in a large multi-European collaboration of 87 pregnancy–associated HUS cases, thrombocytopenia was generally mild and in 15% cases, the platelet count was normal [8].

Lactate dehydrogenase levels are not normally raised in pregnancy and follow nonpregnant ranges. However, renal function does change in pregnancy due to direct or indirect effects of hormonal changes. An increased glomerular filtration rate and a resulting fall in serum creatinine (to about 80% of nonpregnant values [9]) is seen as a consequence of reduced average oncotic pressure and increased ultrafiltration capacity [10]. Creatinine levels, which would be normal outside pregnancy, can represent a significant acute kidney injury (AKI) in pregnancy. A recent systematic review [9] suggests that a serum creatinine >77 μmol/L (0.87 mg/dL) should raise the possibility of AKI, which is defined by a 1.5-time increase from baseline or 26 μmol/L increase in 48 hours [11]. Although AKI can be a feature of TTP and PET/HELLP, this is normally mild, in contrast to CM-HUS, where significant renal impairment is more common. In the European case series, two-thirds of cases of pregnancy–associated HUS required dialysis [8]. Marked transaminitis is not commonly seen in TTP or CM-HUS and would be more in keeping with PET/HELLP (though acute fatty liver of pregnancy needs also to be excluded).

The timing of presentation of the TMA in pregnancy may aid in diagnosis. Both TTP and CM-HUS can present from the first trimester up until any time in the postpartum period. In contrast, PET and HELLP can present any time after 20 weeks of gestation, with a peak incidence in the third trimester, but a small minority of cases may present postpartum.

Although TTP may present in any trimester of pregnancy, peak incidence is during the third trimester or the postpartum period. Of note is the higher frequency of congenital TTP (cTTP) compared to immune TTP (iTTP) presenting in pregnancy [12]. A review of symptoms at presentation does not aid in differentiating between TTP subtypes; however, neurological symptoms (such as migraine, headache, and transient ischemic attack/stroke) and renal impairment were more common in cTTP. In both iTTP and cTTP, presentation may be with fetal death and/or fetal growth restriction, typically in the second trimester. Histology demonstrates infarcts of various ages in the placenta, often with central hemorrhage. This is not specific to TTP and indeed may be a feature of any TMA or severe thrombophilia, but explains fetal impact. In TTP, microvascular thrombi related to ADAMTS-13 deficiency are the cause for the placental findings.

CM-HUS presents most commonly in the postpartum period. In a European collaboration of cases [8], 76% cases were presented in the postpartum period, with the remainder occurring throughout pregnancy, including during the first trimester. A total of 58% of cases presented in the first pregnancy and 42% in subsequent pregnancies. An underlying complement mutation was detected in 56% of women, a similar proportion to in nonpregnant cases. In those with a defined complement mutation, presentation was equally likely to occur in first versus subsequent pregnancies. However, in the 44% of women without a detected complement mutation, presentation was more likely in the first pregnancy. The gestation at presentation was comparable between those with and without a detectable genetic abnormality.

4. Complement and Pregnancy

The complement system plays a critical role in normal pregnancy, and changes in levels of complement proteins and activation products compared to the nonpregnant state are seen.

Complement can be activated by the classical, alternative), and lectin pathways, with the end result of all 3 pathways being assembly of the C3 and C5 convertases on the targeted surface. These cleave plasma proteins C3 and C5 to form bioactive fragments (C3a and C3b and C5a and C5b, respectively), which together opsonize and undergo direct phagocytosis (C3b and its fragments), act as anaphylatoxins (C3a and C5a), modulate the immune response, and lead to cell lysis via formation of the membrane attack complex (C5b-9, MAC) in the terminal pathway. Complement activation is kept in check by complement regulators, including plasma proteins factors H and I, and the membrane proteins CD35, CD46 (membrane cofactor protein), CD55 (decay accelerating factor), and CD59.

In normal pregnancy, there are increases in C3 and C4, as well as in the activation products C3a and C5a. There is also a rise in the regulatory factor H however, and there is no significant change in levels of the soluble version of the MAC (sC5b-9), suggesting a rebalanced state of complement homeostasis [13].

Aside from maintaining host defenses, the complement system has been shown to play an important role in all stages of pregnancy, from implantation to placentation, fetal development, and even in parturition [14] (Figure 2). For example, to ensure adequate trophoblast invasion and healthy implantation, fetal cells need to avoid recognition and activation of the complement system. To this end, placental expression of the membrane–bound complement regulators (CD55, CD59, and CD46 [MAC]) prevents MAC assembly on membrane surface. The classical complement pathway also has an important role, via C1q, for trophoblast migration and spiral artery remodeling for normal placental development. Activation of the lectin pathway holds an important role in removing apoptotic cells and promoting further placentation [15].

Figure 2.

Role of complement in normal and abnormal pregnancy adapted from Girardi et al. [14]

However, there is also evidence for complement dysregulation, for example, excessive activation or impaired regulation in the pathogenesis of a range of pregnancy disorders, including PET, as well as miscarriage, fetal growth restriction, preterm birth [14], and pregnancy–associated CM-HUS.

The evidence for a role for complement in the pathogenesis of PET/HELLP is the subject of a recent review [16], and includes demonstration of increased C5b-9 and C4d deposition on trophoblasts, increased complement activation products (eg, C3a, C5a, sC5b-9, Bb, and C4d) in maternal plasma, and ex vivo experiments in which PET serum induces complement deposition on cells ex vivo.

In one such ex vivo experiment [17], cultured HMEC-1 endothelial cells were exposed to sera or activated-patient plasma, and C5b-9 deposition was analyzed by immunofluorescence. Activated plasma from cases of CM-HUS lead to significantly higher levels of C5b-9 deposition in the acute phase compared to controls, and deposition correlated with disease activity. Normal pregnant activated plasma produced similar deposition levels as controls, but C5b-9 deposition was significantly raised in 3 out of 3 HELLP cases, but to a lesser degree in 6 out of 7 cases of PET [17].

5. Management Of Pregnancy–Specific and Pregnancy–Associated Tmas

PET and HELLP syndrome require supportive care such as blood pressure control, close fetal monitoring, and delivery if there is progressive TMA. Delivery remains the only definitive management, and normalization of laboratory parameters occurs within 48 hours to 72 hours following delivery. However, with a worsening of renal function postdelivery, CM-HUS should be considered, and if cardiac or neurological symptoms occur, TTP is more likely.

If a diagnosis of TTP is suspected, plasma exchange (PEX) should be considered while ADAMTS-13 activity levels are analyzed. In TTP, ADAMTS-13 activity is <10%, whereas it is either normal or mildly reduced in CM-HUS, usually >20% [18]. The diagnosis of CM-HUS remains clinical.

6. Management of Pregnancy–Associated Cm-Hus

In postpartum cases of CM-HUS, aside from supportive care, eculizumab should be initiated early in the disease course for improved renal outcomes (to a similar degree as in non-pregnancy–associated cases [3]). The newer long-acting C5 inhibitor, ravulizumab also has safety and efficacy data in postpartum CM-HUS, with a documented complete response in 87.5% in a median of 31 days, and discontinuation of dialysis in all cases by 21 days [19]. However, if complement inhibitor therapy is not available, there is still a role for PEX [1] (Figure 3).

Figure 3.

Summary of treatment of pregnancy–related and pregnancy–associated TMAs adapted from Scully [18]

When CM-HUS is diagnosed in pregnancy rather than post-partum, and urgent delivery is not feasible, the decision to start complement inhibitor therapy is more complex. However, there is evidence, mainly from use in paroxysmal nocturnal hemoglobinemia (PNH) [20], that eculizumab use in pregnancy is safe.

The optimal management of subsequent pregnancies following a diagnosis of pregnancy–associated CM-HUS remains to be determined. There is a risk of recurrence of CH-HUS, as well as of preeclampsia and HELLP syndrome, as illustrated in a Dutch and Belgian cohort of 17 women and 39 pregnancies, of whom 10 women had complement variants (total of 14 pregnancies). Seven out of 39 pregnancies were complicated by HUS, and 5 out of 39 had HELLP or preeclampsia (unrelated to HUS in pregnancy). A total of 35 out of 39 pregnancies resulted in a live birth but in 8 women there was a progression to end stage renal failure and 3 cases received eculizumab treatment. Of the 14 pregnancies in 10 women with rare complement variants, 10 were uncomplicated pregnancies [21]. Use of complement inhibitor therapy prophylaxis in subsequent pregnancies appears safe, and may be reasonable in high-risk cases, but may not prevent the development of preeclampsia, especially in women with baseline chronic kidney disease [22]

7. Management of Pregnancy–Associated Ttp

Diagnosis of acute TTP in pregnancy requires standard care with plasma exchange, and immunosuppression for immune–mediated cases. Little is known regarding the safety of caplacizumab in pregnancy, though its use in pregnancy has recently been reported in a single case [23]. In this case, the patient had severe ADAMTS-13 deficiency (<10%) with antibodies for ADAMTS-13 at the onset of pregnancy; the pregnancy resulted in an intrauterine death, but it not associated with fetal bleeding. Transplacental transfer of caplacizumab was demonstrated, but at much lower levels than in the patient. Given its small molecular size, crossing of the placenta and inhibition of fetal VWF is possible. Differentiation of iTTP from cTTP may not be immediately obvious from ADAMTS-13 assays and regular follow-ups after achieving clinical remission are suggested.

Subsequent pregnancy following a diagnosis of iTTP is possible, and central to a positive outcome is a normal ADAMTS-13 activity level at the beginning of pregnancy. ADAMTS-13 activity levels should be checked at least in each trimester (in conjunction with routine laboratory parameters) and also in the postpartum period; even with adequate levels during pregnancy, there may be a precipitous drop following delivery. If ADAMTS-13 is reduced or decreases during pregnancy, options for therapy include azathioprine and steroids. If ADAMTS-13 activity becomes severely reduced, PEX and/or rituximab may need to be considered. For the fetus, there is a risk of growth restriction, prematurity, and even intrauterine death. These risks may be exacerbated if ADAMTS-13 activity levels are not optimized [18].

It has been established that women with cTTP require ADAMTS-13 replacement in pregnancy, and the mainstay of treatment currently is plasma infusion (PI). In a Japanese cohort of 38 pregnancies, outcomes of pregnancies occurring before and after a diagnosis of cTTP were compared [24]. Thirteen of a total of 14 cases of intrauterine fetal demise (IUFD) occurred in pregnancies preceding a diagnosis of cTTP, in which fresh frozen plasma was not administered. Of 10 pregnancies following a diagnosis of cTTP, in which fresh frozen plasma prophylaxis was administered, there was 1 case of IUFD. Fetal survival without PI was 40% compared to approximately 90% with PI. Placental histology in an untreated cTTP case associated with IUFD confirmed necrotic villi due to focal placental infarction [24]. Comparable findings have been published from the UK TTP registry [12]. A treatment pathway involves low-dose aspirin and thromboprophylaxis from confirmation of pregnancy and PI at least every 2 weeks, (10 mL/kg). This may need to be increased to weekly infusions from the second or early trimester, guided by routine laboratory parameters and fetal monitoring, including uterine artery dopplers. PI should continue into the postpartum period. Induction of labor should be discussed from 37 weeks of gestation; thereafter cTTP symptoms/TMA laboratory counts are more likely [12].

Finally, a poor obstetric and neonatal history may not be because of maternal cTTP but because of affected offspring, in the very rare event that both parents are carriers of ADAMTS-13 mutations.

8. International Society On Thrombosis and Haemostasis Congress Report

A variety of markers have been investigated in PET. Markers of endothelial dysfunction, including endothelin-1 and endothelial microvesicles, were explored as potential predictors of recurrent PET [25]. VWF:ADAMTS-13 ratio and sC5b-9 were shown to be raised in PET compared to normal pregnancies and pregnancies complicated by normotensive thrombocytopenia and fetal growth restriction, and sC5b-9 levels were correlated with sFlt-1 levels [26]. Changes in platelet aggregation in PET compared to normal pregnancy were described [27]. Plasma levels of tissue factor were found to be increased, with a decrease in tissue factor pathway inhibitor levels, in pregnant women who developed severe early onset PET, but placental tissue factor expression was reduced [28]. Therapeutically, the hemostatic effects of tranexamic acid in PET were evaluated ex vivo, with no effect on D-dimer or plasmin-antiplasmin levels in PET compared to normal pregnancies, suggesting no increased thrombotic risk in PET [29].

In a case of cTTP with fetal demise, proteomics analysis identified a 6- to 7-fold overexpression of S100A8 and S100A9, which was not evident in plasma from cases of PET or pregnancy–induced hypertension, whereas detailed fluorescence imaging revealed increased platelet-neutrophil activation and interaction. A hypercoagulable proinflammatory state was demonstrated in cTTP compared to both normal pregnancy and other pregnancy–related TMAs [30].

Results of targeted screening approach for TTP in complicated pregnancies illustrated that fetal growth restriction may in few cases be the initial presenting feature of pregnancy–associated TTP, and there is potential for misdiagnosis as PET/HELLP [31].

9. Future Directions

Now that complement inhibitor therapy (specifically eculizumab and longer-acting ravulizimab) in CM-HUS (and PNH) is well-established, the role of complement in other pathological scenarios is an active area of investigation. The evidence for increased complement activation in pregnancy–related TMAs, such as PET and HELLP, may provide novel therapeutic approaches, especially given the growing repertoire of complement inhibitors, including C3, factor B, and factor D inhibitors. Currently, such treatments are undergoing trials in PNH. In pregnancy–associated CM-HUS, both eculizumab and ravulizumab appear safe and are effective. The important future development in TTP will be the availability of recombinant ADAMTS-13. While plasma infusion has improved maternal and fetal outcomes in congenital TTP, the increment in ADAMTS-13 levels is limited. Given that recombinant ADAMTS-13 normalizes ADAMTS-13 activity levels, it remains to be seen whether fetomaternal outcomes can be further optimized.