Postpartum Hemorrhage: Wherefore Art Thou, Hyperfibrinolysis?

See Article, p 1373 In 2017, tranexamic acid (TXA)—a well-known antifibrinolytic agent—was hailed as a life-saving treatment for women experiencing postpartum hemorrhage, according to a major study published in the Lancet.1 In this pragmatic, randomized placebo-controlled multicenter trial (The World Maternal Antifibrinolytic Trial [WOMAN]), 20,060 women with postpartum hemorrhage from 21 low- and middle-income countries received TXA (1–2 g) or placebo. Trial data showed a 19% decrease in the relative risk of death from exsanguination in women given TXA versus placebo (mortality rate: 1.9% vs 1.5%, TXA versus placebo groups, respectively; relative risk = 0.81, 95% confidence intervals [CI], 0.65–1.0; P = .045). Further, the treatment benefit appeared strongest if TXA was administered within 3 hours of birth. Following publication, experts have debated whether the WOMAN trials' findings are statistically significant and generalizable to women who deliver in well-resourced hospitals in high-income countries.2,3 There has been less debate surrounding how TXA induces this survival benefit. One may assume that TXA mitigates severe postpartum bleeding by attenuating fibrinolysis or excess clot lysis (hyperfibrinolysis) after delivery. To answer this assumption, 2 questions need to be asked. First, what changes occur in the fibrinolytic system at the time of delivery? Second, what is the frequency of hyperfibrinolysis among women with postpartum hemorrhage? Fibrinolysis describes the dissolution of fibrin clots into soluble fibrin degradation products by plasmin and other fibrinolytic proteases. This process prevents low-flow microcirculatory occlusion. During the peridelivery period, observational data suggest that the maternal fibrinolytic system is activated, evidenced by increased plasma urokinase-plasminogen activator (u-PA) or tissue plasminogen activator (t-PA) levels at or after delivery and decreased plasminogen activator inhibitor 1 (PAI-1) levels after placental delivery.4–8 However, the magnitude of these changes is not consistent across these studies, likely because of differences in assay techniques and small sample sizes. Further, activation of the fibrinolytic system does not occur in isolation. Tissue factor–dependent activation and increased platelet activation occur within 2 hours of placental separation.9,10 Therefore, the increase in clotting system activity may lessen the impact of the concomitant increase in fibrinolytic activity. Data that are more limited exist describing the extent of fibrinolytic changes during episodes of postpartum hemorrhage. Two studies have reported increased postpartum D-dimer levels among women with postpartum hemorrhage compared to women without hemorrhage.11,12 These data require cautious interpretation for several reasons. The half-life of D-dimers in the circulation is long (9–10 hours), no cutoff exists for diagnosing disseminated intravascular coagulation in obstetric hemorrhage, and comparing D-dimer levels across institutions is challenging as D-dimer assay techniques are not standardized.13 The development and standardization of other tests of pro- and antifibrinolytic proteins (eg, t-PA, u-PA, plasminogen, PAI-1) also lag behind other coagulation tests. Plus, tests for these markers of fibrinolysis are more difficult and time-consuming to perform and are typically only used in a research setting. Given that high-throughput fibrinolysis screening tests remain a work in progress,13 identifying other means of detecting hyperfibrinolysis during postpartum hemorrhage is of scientific and clinical interest. The use of whole blood viscoelastic hemostasis assays (eg, thromboelastography [TEG] and rotational thromboelastometry [ROTEM]) for detecting hyperfibrinolysis have been discussed in the surgical and trauma literature but less so in obstetrics. In this issue of Anesthesia& Analgesia, Arnolds and Scavone14 used TEG data to identify hyperfibrinolysis in a cohort of 118 women who experienced postpartum hemorrhage. Hyperfibrinolysis was classified by clot lysis at 30 minutes (LY30) ≥3% using kaolin-activated TEG. The median LY30 among all women was 0.2% (interquartile range: 0%–0.8%), which was substantially lower than the hyperfibrinolysis threshold. Only 15 women (12.7%; 95% CI, 7.9–19.9) were identified with hyperfibrinolysis. Because LY30 values may also be high in the presence of platelet-mediated clot retraction, the authors performed post hoc analysis of the functional fibrinogen TEG LY30 profiles in 13 women with hyperfibrinolysis. Examining data from kaolin TEG and functional fibrinogen TEG allowed the investigators to examine the independent contribution of platelets on individual TEG parameters, including LY30. The results were surprising in that the functional fibrinogen TEG LY30 values were 0% for these 13 women, suggesting that the elevated LY30 values were due to platelet-mediated clot retraction and not hyperfibrinolysis. Based on these results, they estimated the upper limit for "true" hyperfibrinolysis (detected by kaolin TEG LY30 ≥3% and a nonzero functional fibrinogen TEG LY30 ≥3%) to be approximately 3%. The authors correctly acknowledge that TEG and ROTEM may lack sensitivity in detecting nonsevere hyperfibrinolysis and that different fibrinolytic subphenotypes (eg, hyperfibrinolysis, physiological fibrinolysis, and fibrinolytic shutdown) were not examined. Other important limitations should be highlighted. Their small sample size impedes an accurate estimation of the overall incidence of true hyperfibrinolysis and platelet-mediated clot retraction in women with elevated LY30 values. It is unclear whether the incidence of hyperfibrinolysis varies according to hemorrhage severity and etiology. In their study, the mean blood loss was not excessively high (1236 mL) and subgroup sizes were too small for analyses of lysis data by hemorrhage etiology. Finally, it is unclear whether the incidence of hyperfibrinolysis is lower in women with a mechanical etiology—such as atony—compared with women with etiologies linked to consumptive processes, such as placental abruption. Given the paucity of data in the obstetric literature, our understanding of the pathophysiology of postpartum-hemorrhage coagulopathy is limited. In stark contrast, the trauma literature is replete with detailed studies characterizing the multiple phenotypes, mechanisms, and mediators of trauma-induced coagulopathy.15 It is therefore unsurprising that some anesthesiologists and other care providers may assume that study findings and guidelines for bleeding management from the trauma literature apply to the management of coagulopathy in obstetrics. However, these populations' baseline physiology and their respective bleeding etiologies are very different. Therefore, there is a need for large observational studies to study the pathophysiology of coagulation activation and fibrinolysis across all hemorrhage etiologies as well as the complex biological and clinical interactions associated with the pharmacological and nonpharmacological treatment of coagulopathy. In the meantime, we should applaud Arnolds and Scavone14 for shedding light on this poorly understood obstetric morbidity. DISCLOSURES Name: Alexander J. Butwick, MBBS, FRCA, MS. Contribution: This author wrote the manuscript and approved the final version. This manuscript was handled by: Jill M. Mhyre, MD.