Obstetric hemorrhage is a major cause of maternal morbidity and mortality associated with acute fatty liver of pregnancy.1–6 When the syndrome was first described, the accompanying profuse obstetric hemorrhage was attributed to the severe coagulopathy caused by liver failure characteristic of acute fatty liver of pregnancy.7,8 Later, however, disseminated intravascular coagulation (DIC) was implicated as a primary cause of the hemostatic derangement,9 a concept further propagated by Castro et al1,10 in their description of 28 women with convincing evidence of severe DIC. Meanwhile, our previous report of clinical outcomes associated with acute fatty liver of pregnancy suggested that both mechanisms—diminished production of procoagulants and increased use by intravascular coagulation—were likely related to the coagulopathy.6
There is also controversy related to whether hemolysis is a clinically significant component in women with acute fatty liver of pregnancy. Whereas Sheehan11 concluded that there was no discernible hemolysis, Burroughs and colleagues7 reported that accelerated red cell destruction was a major facet of acute fatty liver of pregnancy. In our previous clinical observations of 51 women with acute fatty liver of pregnancy, we reported that brisk ongoing hemolysis was a common finding that contributed to hyperbilirubinemia as well as the need for red cell transfusions.6
Because of these disparate observations, we designed the present investigation with three principal aims. First, we sought to expand our previous observations concerning the frequency and severity of DIC complicating acute fatty liver of pregnancy. Second, we wanted to elucidate the respective contributions to hemostatic dysfunction from DIC and defective procoagulant synthesis. Our final aim was to characterize the cause and extent of associated hemolysis.
PATIENTS AND METHODS
From 1975 through 2012, one or more of the investigators was involved in the care of women admitted to Parkland Hospital and diagnosed with acute fatty liver of pregnancy. With approval from the institutional review boards at the University of Texas Southwestern Medical Center at Dallas and Parkland Hospital, clinical and laboratory data were extracted from medical records of these women. These cases were entered into a registry specific for acute fatty liver of pregnancy as previously described.6 Patient information was deidentified and stored in a computerized database. Maternal demographic information, clinical findings, laboratory results, management, delivery and pregnancy outcomes, and recovery after delivery were reviewed. Laboratory findings were those determined by methods contemporaneously in use for various biochemical, hematologic, and coagulation tests. Additional hematologic and coagulation studies were performed as available by certified medical technologists in the Obstetrical Hematology Research Laboratory.12 For clinical management, hepatic tissue to confirm microvesicular fatty infiltration was obtained by liver biopsy in nine women because of uncertain diagnosis and at autopsy in the two who died. Clinical management was directed using real-time results available contemporaneously from the Parkland Hospital Clinical Laboratory.
For this observational study, women with acute fatty liver of pregnancy were classified and analyzed according to the International Society of Thrombosis and Haemostasis DIC score (Box 1).13,14 Briefly, this scoring system uses a combination of laboratory tests to provide a five-step diagnostic algorithm to calculate the DIC score. These tests include quantification of platelets, fibrinogen, fibrin-related markers (fibrin monomers and fibrin degradation products), and prothrombin time in combination with the underlying disorder known to be associated with DIC.13 Composite International Society of Thrombosis and Haemostasis DIC scores 2 or greater are suggestive of DIC and scores 5 or greater are considered to define unequivocal DIC.13 This scoring system provides a sensitivity and specificity of DIC with 93% and 98%, respectively.14 For the current study, a composite International Society of Thrombosis and Haemostasis DIC score was provided for each woman at 24 hours epochs after delivery.
Studies of hepatic and hemostatic function included fibrinogen, fibrin–fibrinogen split products, coagulation studies, and cholesterol. Fibrinogen in plasma and fibrinogen–fibrin split products in serum were measured using methods previously described by Pritchard et al.12 Briefly, plasma was collected in citrate-Trasylol tubes and fibrinogen was quantified using the thrombin-clottable protein method of Ratnoff and Menzie.15 Serum obtained after thrombin clotting was assayed for fibrin–fibrinogen split products using the Thrombo-Wellcotest immunologic particle agglutination method.
To portray normal fibrinogen synthesis after acute defibrination in women with normal hepatic function, serial plasma fibrinogen and serum fibrin-degradation product concentrations were determined in a cohort of 25 women with placental abruption and hypofibrinogenemia. These women were identified from a repository of women with fetal demise and placental abruption confirmed at delivery.16 Comparisons among fibrinogen and fibrin-degradation product measurements for those women with acute fatty liver of pregnancy compared with women with severe placental abruption were made using mixed-effect modeling with interaction and within-subjects modeling, log likelihood χ2 test for goodness of fit, and Kaplan-Meier analysis for length of time for recovery to specified fibrinogen values. Statistical analysis was accomplished using SAS 9.2. P values <.05 were considered statistically significant.
Concomitant with another ongoing study,17 in some of these women, erythrocyte morphology was quantified using scanning electron microscopy. Briefly, these methods included preparation of erythrocytes from peripheral blood collected into ethylenediamine tetraacetic acid-containing tubes was fixed in 1.25% glutaraldehyde, and 1,000 erythrocytes in contiguous fields were scanned using a JEOL JSM-35 electron microscope. Red cells were classified as normal discocytes or as abnormal schizocytes, echinocytes, or spherocytes according to Bessis.18
Between 1975 and 2012, there were 51 women with acute fatty liver of pregnancy cared for at our institution. The clinical outcomes of this cohort have been previously described,6 and briefly, the frequency of acute fatty liver of pregnancy was 1 per 10,000 births and distribution over this time period was relatively homogeneous, mean maternal age was 27.4±7.3 years (range 15–42 years), and 41% were nulliparous. Their overall demographic characteristics were similar to those of women from our general obstetric population. Blood or component transfusions were required in 28 of 51 women, and seven (14%) required platelet transfusions. Importantly, 12 (42%) of the women were given additional red cell transfusions beyond 48 hours after delivery with indications being persistent, severe anemia. Aside from obstetric hemorrhage associated with coagulopathy and acute fatty liver of pregnancy, we found only 5 of 51 women might have had an associated event that contributed to the hematologic aberrations; specific comorbid conditions included three abruptions, one uterine rupture, and one subcapsular hematoma.
As shown in Figure 1, for the entire cohort, the mean International Society of Thrombosis and Haemostasis DIC score was 5.9±1.8 at the time of delivery, confirming that 80% of these women had unequivocal DIC defined as composite score of 5 or greater. As discussed, inherent in this scoring system is the presence of an underlying disorder known to be associated with overt DIC, scored as yes or no. As evidenced by the clinical hemostatic dysfunction, a score of 2 was applied to all members of the cohort on the day of delivery. Global coagulation tests were then serially scored on days after delivery, and the most influential marker scored was elevated fibrin-split products that were present in all cases studied during the first 48 hours after delivery. Importantly, composite International Society of Thrombosis and Haemostasis DIC scores remained elevated for up to 4–5 days suggesting persistent hemostatic dysfunction after delivery (Fig. 1).
To demonstrate the effect of liver dysfunction on procoagulant synthesis, serial plasma fibrinogen and fibrin-degradation products were plotted and compared for the two cohorts of women. For the 25 women with placental abruption, hypofibrinogenemia, and normal hepatic function, the mean age was 22.6±4.6 years (range 16–31 years), 20% were nulliparous, and mean gestational age and birth weight of their stillborn neonates was 35±1.5 weeks and 2,290±740 g, respectively. In Figure 2, their serial values are compared with women with acute fatty liver of pregnancy. The initial mean fibrinogen level (±standard error of mean) for women with acute fatty liver of pregnancy were 188±24 mg/dL compared with that of 153±10 mg/dL for women with an abruption. As can be seen in Figure 2, fibrinogen recovery for women with an abruption was rapid, linear, and returned to near normal levels within 36 hours of delivery. For women with acute fatty liver, however, several days were required to reach similar levels. Random effects modeling for both linear and quadratic regression curves were significantly different for days after delivery (P<.001 for interaction). Kaplan-Meier survival analysis of fibrinogen recovery to a set point of 280 mg/dL after delivery was also statistically different between the two groups with median times to recovery for acute fatty liver of pregnancy and abruption being 4.2 and 1.7 days, respectively (P=.046 using log-rank test). The groups of women with acute fatty liver of pregnancy and abruption were also compared for fibrin-degradation product clearance, and as shown in Figure 3, these differences were also significant using random effects modeling for both linear and quadratic regression curves for days after delivery (P<.001 for interaction). Although we have previously described continuing hepatic dysfunction for days after delivery, it was not possible to accurately quantify the contribution of decreased synthesis versus continuing consumption of fibrinogen—only that both were present.
Reticulocytosis, nucleated red blood cells, and serum bilirubin levels together are indicative of ongoing hemolysis in the cases of acute fatty liver of pregnancy. As shown in Figure 4, levels of serum bilirubin remained increased after delivery in women with acute fatty liver of pregnancy in conjunction with elevated reticulocyte counts. Nucleated red blood cells were identified in 40% of women at delivery with acute fatty liver of pregnancy (Fig. 5). The median (quartile1, quartile3) nucleated red blood cells at delivery was 4 (2, 11)/100 white blood cells, and these cells were present for up to 4–5 days postpartum. In three women, erythrocyte morphology was determined using scanning electron microscopy. Echinocytes were the predominant abnormal form (Fig. 6), and these women initially had 22–39% echinocytes at the time of delivery. This proportion compares with normally pregnant women whose mean value ranges from 0.5 to 1.9%. In these women, normal discocyte forms were present within 2–3 days. During this same time, schizocyte proportions were less than 2% and spherocytes were negligible.
Our observations in these 51 women with acute fatty liver of pregnancy confirm that profound hemostatic dysfunction commonly exacerbates obstetric hemorrhage and contributes to the need for blood and component transfusions. At the time of delivery, half of these women now reported had a plasma fibrinogen level less than 150 mg/dL, and 12 (48%) of these required either fresh-frozen plasma or cryoprecipitate to achieve hemostasis because of dangerously depressed fibrinogen concentrations in the setting of active bleeding. We also found that the cause of hemostatic dysfunction was twofold and the result of a combination of procoagulant deficiency from liver failure as well as procoagulant consumption from continuing DIC. Another observation was that brisk hemolysis in these women substantively adds to the requirement for continued transfusions well beyond the immediate time of delivery.
We have previously presented evidence for continuing hemostatic dysfunction in these women manifest by prolongation of the thrombin and partial thromboplastin times for 4–6 days after delivery.6 When we plotted the time course of plasma, fibrinogen concentrations and fibrin degradation products found that continuing fibrinogen deficiency is caused by diminished production as well as increased use (Figs. 2 and 3). Evidence for decreased production was seen with abnormally low plasma fibrinogen concentrations persisting for the first several days after delivery along with only mild to moderately elevated fibrin-degradation products. By way of comparison, in the same figures, fibrinogen values are plotted for 25 women who sustained a placental abruption severe enough to kill the fetus. These women with hypofibrinogenemia who had normal hepatic function exhibited the anticipated response with a rapid return within 24 hours to a normal plasma fibrinogen concentration along with simultaneous clearance of fibrin degradation products. This was in contrast to the women with hepatic dysfunction from acute fatty liver of pregnancy whose plasma fibrinogen levels remained static over a period of several days after delivery.
At the same time, there is also evidence for continuing increased procoagulant consumption caused by ongoing DIC. As shown in Figure 1, most of these women had evidence for persistent consumptive coagulopathy as assessed by the DIC score of the International Society of Thrombosis and Haemostasis.13,14 As shown in Figure 3, evidence for ongoing consumptive coagulopathy is provided by the modestly elevated levels of fibrin degradation products in the face of depressed plasma fibrinogen concentrations. Although this provides evidence of ongoing DIC, we cannot exclude the possibility that elevated fibrin degradation product levels are at least partially related to their diminished clearance because of hepatic dysfunction.
We also found that these women with fatty liver of pregnancy had brisk hemolysis that continued for several days after delivery. This was of clinical significance because of the frequent need for ongoing red cell transfusions after delivery and at a time after which surgical and obstetric hemostasis was secured. Specifically, one-fourth of these women required additional erythrocyte transfusions for persistent severe anemia without evidence of ongoing hemorrhage. There are at least three other findings that document hemolysis. One marker is the increasing serum bilirubin levels that continued to rise at a time when there is improving, albeit impaired, hepatic bilirubin clearance. As shown in Figure 4, bilirubin concentrations peaked at 5–7 days after delivery at a time when reticulocytosis was also maximal. Another potent marker for accelerated hemolysis with hemopoiesis is the appearance of remarkably elevated levels of nucleated red blood cells (Fig. 5), which are seldom encountered in adults except with massive hemolysis or hypoxia.19–21 The third marker indicative of hemolysis is the remarkably high proportion of echinocytes in the three women studied. This cell morphotype may be induced by abnormal plasma and red cell membrane levels of cholesterol and other blood lipids. Echinocytes either undergo premature hemolysis or they can revert back to normal discocytes. At the same time, the paucity of schizocytes supports the observations by Burroughs et al7 that there is negligible microangiopathic hemolysis.
Our study has several limitations. First, because of the observational design, we could have missed some women with acute fatty liver of pregnancy. Second, some of these women may have been incorrectly diagnosed to have hepatocellular steatosis. However, as we have previously described,6 the criteria for diagnosis of acute fatty liver of pregnancy included evidence of acute liver failure with characteristic clinical findings accompanied by laboratory evidence that confirmed hepatic dysfunction along with collateral multiorgan system aberrations. We also applied both the “Swansea criteria” proposed by Ch'ng et al22 and the “acute fatty liver of pregnancy triad” of Vigil-de Gracia et al5 to confirm the diagnoses. Another drawback is that observations from this study provide only circumstantial evidence of the hemostatic dysfunction associated with acute fatty liver of pregnancy as related to substantive ongoing DIC in concert with reduced procoagulant synthesis.
In summary, our findings support the hypothesis that a combination of substantive DIC and procoagulant deficiency from hepatic dysfunction both contribute significantly to hemostatic dysfunction that characterizes acute fatty liver of pregnancy. To compound the issue, this syndrome includes an element of clinically significant hemolysis. Because prolonged persistence of coagulopathy occurs in some women, it seems prudent to monitor coagulation studies until clinical recovery and derangements normalize.
1. Castro MA, Fasset MJ, Reynolds TB, Shaw KJ, Goodwin TM. Reversible peripartum liver failure: a new perspective on the diagnosis, treatment, and cause of acute fatty liver of pregnancy based on 28 cases. Am J Obstet Gynecol 1999;181:389–95.
2. Knight M, Nelson-Piercy C, Kurinczuk JJ, Spark P, Brocklehurst P; UK Obstetric Surveillance System. A prospective national study of acute fatty liver of pregnancy in the UK. Gut 2008;57:951–6.
3. Fesenmeier MF, Coppage KH, Lambers DS, Barton JR, Sibai BM. Acute fatty liver of pregnancy in 3 tertiary centers. Am J Obstet Gynecol 2005;192:1416–9.
4. Lau HH, Chen YY, Huang JP, Chen CY, Su TH, Chen CP. Acute fatty liver of pregnancy in a Taiwanese tertiary care center: a retrospective review. Taiwan J Obstet Gynecol 2010;49:156–9.
5. Vigil-de Gracia P, Montufar-Rueda C. Acute fatty liver of pregnancy: diagnosis, treatment, and outcome based on 35 consecutive cases. J Matern Fetal Neonatal Med 2011;24:1143–6.
6. Nelson DB, Yost NP, Cunningham FG. Acute fatty liver of pregnancy: clinical outcomes and expected durations of recovery. Am J Obstet Gynecol 2013;209:456.e1–7.
7. Burroughs AK, Seong NG, Dojcinov DM, Scheuer PJ, Sherlock SV. Idiopathic acute fatty liver of pregnancy in 12 patients. Q J Med 1982;51:481–97.
8. Sherlock S. Acute fatty liver of pregnancy and the microvesicular fat diseases. Gut 1983;24:265–9.
9. Hozbach R. Acute fatty liver of pregnancy with disseminated intravascular coagulation. Obstet Gynecol 1974;43:740–4.
10. Castro MA, Goodwin TM, Shaw KJ, Ouzounian JG, McGehee WG. Disseminated intravascular coagulation and antithrombin III depression in acute fatty liver of pregnancy. Am J Obstet Gynecol 1996;174:211–6.
11. Sheehan HL. The pathology of acute yellow atrophy and delayed chloroform poisoning. J Obstet Gynecol 1940;47:49–62.
12. Pritchard JA, Cunningham FG, Mason RA. Coagulation changes in eclampsia: their frequency and pathogenesis. Am J Obstet Gynecol 1976;124:855–64.
13. Taylor FBJ, Toh CH, Hoots WK, Wada H, Levi M; Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH). Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86:1327–30.
14. Bakhtiari K, Meijers JC, de Jonge E, Levi M. Prospective validation of the international society of thrombosis and haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med 2004;32:2416–21.
15. Ratnoff OD, Menzie C. A new method for the determination of fibrinogen in small samples of plasma. J Lab Clin Med 1951;37:316–20.
16. Pritchard JA, Brekken AL. Clinical and laboratory studies on severe abruption placentae. Am J Obstet Gynecol 1967;97:681–700.
17. Cunningham FG, Lowe T, Guss S, Mason R. Erythrocyte morphology in women with severe preeclampsia and eclampsia: preliminary observations with scanning electron microscopy. Am J Obstet Gynecol 1985;153:358–63.
18. Bessis M. Living blood cells and their ultrastructure. New York (NY): Springer; 1973.
19. Stachon A, Holland-Letz T, Krieg M. High in-hospital mortality of intensive care patients with nucleated red blood cells in blood. Clin Chem Lab Med 2004;42:933–8.
20. Stachon A, Segbers E, Holland-Letz T, Kempf R, Hering S, Krieg M. Nucleated red blood cells in the blood of medical intensive care patients indicate increased mortality risk: a prospective cohort study. Crit Care 2007;11:R62.
21. Stachon A, Becker A, Kempf R, Holland-Letz T, Friese J, Krieg M. Re-evaluation of established risk scores by measurement of nucleated red blood cells in blood of surgical intensive care patients. J Trauma 2008;65:666–73.
© 2014 by The American College of Obstetricians and Gynecologists.
22. Ch'ng CL, Morgan M, Hainsworth I, Kingham JG. Prospective study of liver dysfunction in pregnancy in Southwest Wales. Gut 2002;51:876–80.