Best Practice & Research Clinical Obstetrics and Gynaecology Vol. 15, No. 4, pp. 623±644, 2001 doi:10.1053/beog.2001.0204, available online at http://www.idealibrary.com on 9 Disseminated intravascular coagulation Elizabeth A. Letsky MB BS FRCPath FRCOG FRCPCH Consultant Perinatal Haematologist/Hon Senior Lecturer Imperial College School of Medicine, Queen Charlotte's Hospital, Hammersmith Hospitals Trust, Hammersmith House, 2nd Floor, Du Cane Road, London, W12 0HS, UK Healthy pregnancy is accompanied by changes in the haemostatic system which convert it into a hypercoagulable state vulnerable to a spectrum of disorders ranging from venous thromboembolism to disseminated intravascular coagulation (DIC). This latter is always a secondary phenomenon triggered by speci®c disorders such as abruptio placentae and amniotic ¯uid embolism due to release of thromboplastin intravascularly or endothelial damage resulting from pre-eclampsia and sepsis. In modern obstetric practice the most common cause is haemorrhagic shock with delay in resuscitation leading to endothelial damage. The initial management of massive obstetric haemorrhage is the same whether associated with coagulopathy initially or not. Low-grade DIC, associated with pre-eclampsia, is monitored haematologically by serial platelet counts and serum ®brin degradation products (FDPs). Supportive measures and removal of the triggering mechanism are the key to successful management. Outcome depends primarily on our ability to deal with the trigger and not on direct attempts to correct the coagulation de®cit. Key words: haemostatis; coagulopathy; ®brinolysis; massive haemorrhage; transfusion; abruptio placentae; amniotic ¯uid embolism; low-grade DIC; pre-eclampsia; sepsis. HAEMOSTASIS AND PREGNANCY Healthy haemostasis depends on normal vasculature, platelets, procoagulant factors with their inhibitors and ®brinolysis. These act together to con®ne the circulating blood to the vascular bed and arrest bleeding after trauma. Normal pregnancy is accompanied by dramatic changes in the coagulation and ®brinolytic systems1,2 where there is a marked increase in some of the procoagulant factors, particularly ®brinogen and a suppression of ®brinolysis. These changes, together with the increased blood volume, help to combat the hazard of haemorrhage at placental separation, but play only a secondary role to the unique process of myometrial contraction which reduces the blood ¯ow to the placental site. They also produce a vulnerable state for intravascular clotting, which is expressed as a whole spectrum of disorders in pregnancy ranging from thromboembolism to bleeding resulting from disseminated intravascular coagulation (DIC).1,3 Summary of changes in haemostasis in pregnancy and delivery The changes in the haemostatic system in normal healthy pregnancy are consistent with a continuing low-grade process of coagulant activity. Suppression of ®brinolysis 1521±6934/01/04062322 $35.00/00 c 2001 Harcourt Publishers Ltd. * 624 E. A. Letsky leads to ®brin deposition in the intervillous space of the placenta and in the walls of the spiral arteries supplying the placenta.4 As pregnancy advances, the elastic lamina and smooth muscle of these spiral arteries are replaced by a matrix containing ®brin. This allows expansion of the lumen to accommodate an increasing blood ¯ow and reduces the vascular resistance of the placenta. At placental separation during normal childbirth, a blood ¯ow of 500±800 ml/minute has to be staunched within seconds, or serious haemorrhage will occur. Myometrial contraction plays a vital role in securing haemostasis by reducing the blood ¯ow to the placental site. Rapid closure of the terminal part of the spiral artery will be further facilitated by removal of the elastic lamina. The placental site is rapidly covered by a ®brin mesh following delivery. The increased levels of ®brinogen and other coagulation factors will be advantageous to meet the sudden demand for haemostatic components. The changes also produce a vulnerable state for intravascular clotting and a whole spectrum of disorders involving coagulation which occur in pregnancy. DISSEMINATED INTRAVASCULAR COAGULATION (DIC) The ®rst problem with DIC is in its de®nition. It is never primary, but always secondary to some general stimulation of coagulation activity by release of procoagulant substances into the blood (Figure 1). Triggers of this process in pregnancy include endothelial damage, the release of placental tissue, amniotic ¯uid, incompatible red cells or bacterial products into the maternal circulation. There is a great spectrum of manifestations of the process of DIC (Table 1). ranging from a compensated state with no clinical manifestation but evidence of increased production and breakdown of coagulation factors, to the condition of massive uncontrollable haemorrhage with very low concentrations of plasma ®brinogen, pathological raised levels of ®brin degradaption products (FDPs) and variable degrees of thrombocytopenia. Further cause for confusion is that there appears to be a transient state of intravascular coagulation during the whole of normal labour, maximal at the time of birth.2,5,6 Fibrinolysis is stimulated by DIC, and the FDPs resulting from the process interfere with the formation of ®rm ®brin clots causing a vicious circle which results in further disastrous bleeding (Figure 1). FDPs also interfere with myometrial function and possibly cardiac function and therefore in themselves aggravate both haemorrhage and shock (Figure 2). Obstetric conditions associated with DIC include abruptio placentae, amniotic ¯uid embolism, septic abortion and intrauterine infection, retained dead fetus, hydatidiform mole, placenta accreta, pre-eclampsia and prolonged shock from any cause (see Figure 3). Despite the advances in obstetric care and highly developed blood trasfusion services, haemorrhage still constitutes a major factor in maternal mortality and morbidity.7,8 There have been many reports concerning small series of patients or individual patients with coagulation failure during pregnancy. However, no signi®cant controlled trials of the value of the many possible therapeutic measures have been carried out. This is mainly because no one person or unit is likely to see enough cases to randomize patients into groups in which the numbers would achieve statistical signi®cance. Also, the complex and variable nature of the conditions associated with DIC, which are often self-correcting and treated with a variety of measures, make it dicult to draw helpful conclusions from the published reports. Disseminated intravascular coagulation 625 Trigger Local activation of coagulation Recovery Intravascular coagulation Fibrin deposition Consumption of clotting factors and platelets Fibrinolysis FDPs Bleeding Impairment of myometrial function Thrombosis of small vessels Organ dysfunction Figure 1. Disseminated intravascular coagulation: stimulation of coagulation activity and its possible consequences. Table 1. Spectrum of severity of DIC: its relationship to speci®c complications in obstetrics. Severity of DICa In vitro ®ndings Stage 1 Low-grade compensated FDPs " Platelets # Stage 2 Uncompensated but no haemostatic failure As above, plus ®brinogen # Platelets # # Factors V and VIII # Small abruptio Severe pre-eclampsia Stage 3 Rampant with haemostatic failure Platelets # # Gross depletion of coagulation factors, particularly ®brinogen FDPs " " Abruptio placentae Amniotic ¯uid embolism Eclampsia Obstetric conditions commonly associated Pre-eclampsia and related syndromes a Rapid progression from stage 1 to stage 3 is possible unless appropriate action is taken. Haematological management of the bleeding obstetric patient The management of the bleeding obstetric patient is an acute and frightening problem. Because of the urgency of the situation there should be a routine planned practice agreed by haematologist, physician, anaesthetist, obstetrician and nursing sta in all maternity units, to deal with this situation whenever it arises. Good, reliable, continuing communication between the various clinicians, nursing paramedical and laboratory sta is essential. There should be frequent `rehearsals'7 or `®re-drills' to ensure that this relatively rare emergency is dealt with eciently. It is imperative that the source of bleeding, often an unsuspected uterine or genital laceration, be located and dealt with. Prolonged hypovolaemic shock, or indeed shock from any cause, may also trigger DIC and this may lead to haemostatic failure and further prolonged haemorrhage. 626 E. A. Letsky Interfere with platelet function POTENT ANTICOAGULANT ACTIVITY Anti-thrombin Interfere with polymerization of fibrin monomer Myometrium TOXICITY Myocardium Figure 2. Properties of FDPs. Trigger mechanisms of DIC during pregnancy Collagen Pre-eclampsia Hypovolaemia ENDOTHELIAL INJURY XII XIIa XI XIa IX IXa VII VIII X Xa Septicaemia Abruptio placentae Amniotic fluid embolism Retained dead fetus Abortion induced with hypertonic fluids Intrauterine sepsis THROMBOPLASTIN Hydatidiform mole Placenta accreta Intravascular haemolysis Incompatible blood transfusion V PHOSPHOLIPID Large feto-maternal bleed II IIa Septicaemia Fibrinogen Fibrin Figure 3. Interactions occur in many of these obstetric complications. The management of haemorrhage is virtually the same whether the bleeding is initiated or augmented by coagulation failure. The clinical condition usually demands urgent treatment and there is no time to wait for results of coagulation factor assays or sophisticated tests of the ®brinolytic system activity for precise de®nition of the extent of haemostatic failure. (Blood may be taken for this purpose and analysed later once the emergency is over.) Disseminated intravascular coagulation 627 Emergency laboratory assessment of haemostatic status In order to avoid testing artefacts it is essential that the blood be obtained by a quick, ecient, non-traumatic technique remote from indwelling lines. Thromboplastin release from damaged tissues may contaminate the specimen and alter the results. This is likely to occur if diculty is encountered in ®nding the vein, if the vein is only partly canalized and the ¯ow is slow, or if there is excessive squeezing of tissues and repeated attempts to obtain a specimen with the same needle. In such circumstances the specimen may clot in the tube in spite of the presence of anticoagulant, or the coagulation times of the various tests will be altered and will not re¯ect the true situation in vivo. The platelets may aggregate in clumps and give a falsely low count. Heparin characteristically prolongs the partial thromboplastin time and thrombin time out of proportion to the prothrombin time. As little as 0.05 units of heparin per millilitre will prolong the coagulation test times. It is customary, though not desirable, to take blood for coagulation tests from lines which have been washed through with ¯uids containing heparin to keep them patent. It is almost impossible to overcome the eect of such ¯uids on the blood passing through such a line however much blood is taken and discarded before obtaining a sample for investigation. This necessitates taking blood from another site remote from heparin infusions. Useful rapid screening tests for haemostatic failure include the platelet count, partial thromboplastin time, or accelerated whole-blood clotting time (which tests intrinsic coagulation), prothrombin time (which tests extrinsic coagulation), the thrombin time and estimation of ®brinogen (Figure 4). The measurement of FDPs provides an indirect test for ®brinolysis. In obstetric practice the measurement of FDPs is usually part of the investigation of suspected acute or chronic DIC. In the acute situation raised FDPs only con®rm the presence of DIC, but are not diagnostic, and once the specimen is taken the laboratory measurement should be delayed until after the emergency is over. Although frequently measured, the level of soluble ®brin has not been shown in a large controlled trial to be of diagnostic or prognostic value in obstetric patients with haemorrhage.9 Skilled laboratory workers can be performing a much more valuable service in providing results of coagulation screening tests and in providing blood and blood products suitable for transfusion. Of the tests of coagulation, probably the thrombin time, an estimation of the thrombin clottable ®brinogen in a citrated sample of plasma, is the most valuable overall rapid screen of haemostatic competence of coagulation factors in a previously haemostatically normal bleeding patient. The thrombin time of normal plasma is adjusted in the laboratory to 10±15 seconds, and the ®brin clot formed is ®rm and stable. In the most severe forms of DIC there is no clottable ®brinogen in the sample, and no ®brin clot appears even after 2±3 minutes. Indication of severe DIC is obtained usually by a prolonged thrombin time which is observed not only with depleted ®brinogen but in conditions where FDPs are increased. There is no point whatsoever in the obstetrician, anaesthetist or nursing sta wasting time trying to perform bedside whole-blood clotting tests which furnish little information of practical value and only create more panic. The valuable hands at the bedside are of more use doing the things they are trained to do in this emergency situation rather than wasting time performing a test which is time-consuming, of little value or signi®cance unless performed under strictly controlled conditions, and will not contribute much, if anything, to management. The alerted laboratory worker will be able to provide helpful results on which the obstetrician can act within 30 minutes 628 E. A. Letsky INTRINSIC SYSTEM XII EXTRINSIC SYSTEM Contact factor Tissue thromboplastin XI IX VIII Activated Partial Thromboplastin time (APTT) VII X Xa Prothrombin time (PT) V II I Thrombin time (TT) Fibrin clot Figure 4. In vitro screening tests of coagulation and their relationship to the systems involved. of receiving the specimen in the laboratory. The tests referred to above are straightforward and should be available from any routine haematology laboratory. Management of severe haemorrhage Management of severe haemorrhage must include prompt and adequate ¯uid replacement in order to avoid renal shutdown.10 If eective circulation is restored without too much delay FDPs will be cleared from the blood mainly by the liver, which will further aid restoration of normal haemostasis. Plasma substitutes. There is much controversy around which plasma substitute to give to any bleeding patient. The remarks which follow relate to the supportive management of acute haemorrhage from the placental site and/or birth canal and should not be taken to apply to those situations in which hypovolaemia may be associated with severe hypoproteinaemia such as occurs in septic peritonitis, burns and bowel infarction. The choice lies between simple crystalloids, such as Hartmann's solution or Ringer lactate, and arti®cial colloids, such as dextrans, hydroxyethyl starch and gelatin solution or the very expensive preparations of human albumin (albuminoids). If crystalloids are used, two to three times the volume of estimated blood loss should be administered because Disseminated intravascular coagulation 629 the crystalloid remains in the vascular compartment for a shorter time than colloids when renal function is maintained. The infusion of plasma substitutes, i.e. plasma protein, dextran, gelatin and starch solutions may result in adverse reactions. Although the incidence of severe reactions is rare, they are diverse in nature, varying from allergic urticarial manifestations and mild fever to life-threatening anaphylactic reactions due to spasm of smooth muscle, with cardiac and respiratory arrest.11 Dextrans adversely aect platelet function, may cause pseudo-agglutination and interfere with interpretation of subsequent blood grouping and cross-matching tests. They are, therefore, contraindicated in the woman who is bleeding due to a complication associated with pregnancy where there is a high chance of there being a serious haemostatic defect already. The anaphylactoid reactions accompanying infusion of dextrans are probably related to IgG and IgM antidextran antibodies which are found in high concentrations in all patients with severe reactions. Of greater signi®cance in modern obstetric practice, acute fetal distress has been reported in mothers given Dextran 70 who suered anaphylactoid reactions.12 These Dextran-induced anaphylactoid reactions have resulted in uterine hypertonia with subsequent severe fetal bradycardia even though immunoprophylaxis with Dextran Hapten has been administered.13 There are many suitable superior alternatives for plasma expansion, and the Royal College of Obstetricians recommends that Dextran should be avoided in obstetric practice.14 Many studies suggested that the best way to deal with hypovolaemic shock initially is by transfusing simple balanced salt solutions (crystalloid) followed by red cells and fresh frozen plasma (FFP).15±17 Some advocate the use of a derivative of bovine gelatin ± polygeline (Haemaccel) ± as a ®rst-line ¯uid in resuscitation. It has a shelf-life of 8 years and can be stored at room temperature. It is iso-oncotic and does not interfere with platelet function or subsequent blood grouping or crossmatching. Renal function is improved when it is administered in hypovolaemic shock. Haemaccel is generally considered to be non-immunogenic and therefore does not trigger the production of antibodies in humans, even on repeated challenge. The reactions which occur related to Haemaccel infusion are thought to be due to histamine release18, the incidence and severity of reactions being proportional to the extent of histamine release. There have been a few reports of severe reactions with bronchospasm and circulatory collapse, although rarely, but there has been only one report of a fatality.19 Nevertheless, whatever substitute is used, it is only a stop gap until suitable blood component therapy can be administered. The use of blood and component therapy Whole blood may be the treatment of choice in coagulation failure associated with obstetric disorders, but whole fresh blood is not available from regional centres in the UK. To release blood components earlier than the usual 18±24 hours would increase the risk of serologically incompatible transfusions, viral hepatitis and human immunode®ciency virus (HIV). Syphilis, cytomegalovirus and Epstein±Barr virus are examples of other infections which may be transmitted in fresh blood. Their viability diminishes rapidly on storage at 48C. These infections, particularly in immunosuppressed or pregnant patients, can be particularly hazardous. All red cell concentrates from units in the UK are now ®ltered, removing virtually all white cells ± which further reduces the hazard of transfusion-transmitted infection.20 630 E. A. Letsky Fresh frozen plasma (FFP). FFP contains all the coagulation factors present in plasma obtained from the whole blood within 6 hours of donation. Frozen rapidly and stored at ÿ308C, the factors are well preserved for at least 1 year. Cryoprecipitate. Although cryoprecipitate is richer in ®brinogen than FFP it lacks antithrombin (AT) which is rapidly consumed in obstetric bleeding associated with DIC.21 The use of cryoprecipitate also exposes the recipient to more donors and the potential associated hazards. Platelets. Platelets, an essential haemostatic component, are not present in FFP and their functional activity rapidly deteriorates in stored blood. The platelet count re¯ects both the degree of intravascular coagulation and the amount of bank blood transfused. A patient with persistent bleeding and a very low platelet count (520 109 l) may be given concentrated platelets, although they are seldom required in addition to FFP to achieve haemostasis in obstetric haemorrhage. A spontaneous recovery from the coagulation defect is to be expected once the uterus is empty and well contracted, provided that blood volume is maintained by adequate replacement monitored by central venous pressure and urinary output. Quite frequently the platelet count continues to fall in the ®rst 12±24 hours after massive haemorrhage even though haemostasis has been achieved. This is not an indication for platelet transfusion in the non-bleeding patient with normal coagulation screening tests. Red cell transfusion. Cross-matched blood should be available within 40 minutes of the maternal specimen reaching the laboratory. If the woman has had regular antenatal care her blood group and antibody screen will be documented. There is a good case for giving uncross-matched blood of her own group should the situation warrant it, provided that blood has been properly processed at the transfusion centre. If the blood group is unknown, uncross-matched group O Rh (D) negative blood may be given if necessary. By this time laboratory screening tests of haemostatic function should be available. If these prove to be normal, but vaginal bleeding continues, the cause is nearly always concealed trauma or bleeding from the placental site due to failure of the myometrium to contract. It is imperative that the source of bleeding, often an unsuspected uterine or genital laceration, should be located and dealt with. Prolonged hypovolaemic shock, or indeed shock from any cause, may trigger DIC and this may lead to haemostatic failure and further prolonged haemorrhage. If the blood loss is replaced only by stored bank blood which is de®cient in the labile clotting factors V and VIII and platelets, then the circulation will rapidly become depleted in these essential components of haemostasis even if there is no DIC initially as the cause of haemorrhage. It is advisable to transfuse 2 units of FFP for every 4±6 units of bank red cells administered. It seems sensible in any event, whatever the cause of bleeding, to change the initial plasma substitute and transfuse FFP once it has thawed, while waiting for compatible blood to be available. A spontaneous recovery from the coagulation defect is to be expected once the uterus is empty and well contracted, provided that blood volume is maintained by adequate replacement monitored by central venous pressure and urinary output. The single most important component of haemostasis at delivery is contraction of the myometrium, stemming the ¯ow from the placental site. Massive transfusion of all clotting factors and platelets will not stop haemorrhage if the uterus remains ¯abby. Disseminated intravascular coagulation 631 Vaginal delivery will make a less severe demand on the haemostatic mechanism than delivery by Caesarean section, which requires the same haemostatic competence as any other major surgical procedure. Should DIC be established with the fetus in utero, rather than embarking on heroic surgical delivery it is better to correct the DIC and wait for spontaneous delivery if possible, or stimulate vaginal delivery, avoiding softtissue damage. Arterial embolization for haemorrhage in the obstetric patient (see Chapter 5) Transcatheter arterial embolization has been used for the control of pelvic haemorrhage resulting from trauma, malignancy and radiation since the late 1960s. During the last two decades there has been an increasing use of transcatheter embolization of the internal iliac or uterine artery for control of post-partum haemorrhage.22,23 The general concensus is that selective emergency arterial embolization is an eective means of controlling severe intractable post-partum haemorrhage. High-risk surgery is avoided, and reproductive ability is maintained. DIC in clinical conditions Abruptio placentae Premature separation of the placenta, or abruptio placentae, is the most frequent obstetric cause of coagulation failure. Many of the problems which confront the attendant in this situation are common to other conditions associated with DIC in pregnancy so that abruption will be used as the central focus to discuss management controversies. Abruptio placentae can occur in apparently healthy women with no clinical warning or in the context of established pre-eclampsia. It is possible that clinically silent placental infarcts may pre-dispose to placental separation by causing low-grade abnormalities of the haemostatic system such as increased Factor VIII consumption and raised FDPs.24 There is a great spectrum in the severity of the haemostatic failure in this condition25 which appears to be related to the degree of placental separation. Only 10% of patients with abruptio placentae show signi®cant coagulation abnormalities.26 In some small abruptions there is a minor degree of failure of haemostatic processes and the fetus does not succumb. When the uterus is tense and tender and no fetal heart can be heard, the separation and retroplacental bleeding are extensive. No guide to the severity of the haemorrhage or coagulation failure will be given by the amount of vaginal bleeding. There may be no external vaginal blood loss, even when the placenta is completely separated, the fetus is dead, the circulating blood is incoagulable and there is up to 5 litres of concealed blood loss resulting in hypovolaemic shock. Haemostatic failure may be suspected if there is persistent oozing at the site of venepuncture or bleeding from the mucous membranes of the mouth or nose. Simple rapid screening tests, as described above and referred to below, will con®rm the presence of DIC. There will be a low platelet count, greatly prolonged thrombin time, low ®brinogen, together with raised FDPs, due to secondary ®brinolysis stimulated by the intravascular deposition of ®brin.27 A recent report28, which linked elevated thrombomodulin with abruptio placentae, showed that the combination of thrombomodulin and ultrasound as a `double-marker' detected all cases of abruption in this small series. The authors suggest that thrombomodulin may prove useful in the diagnosis of abruption associated with trauma or unexplained vaginal bleeding and may become a 632 E. A. Letsky useful routine screening test. The mainstay of treatment is to restore and maintain the circulating blood volume. This not only prevents renal shutdown and further haemostatic failure caused by hypovolaemic shock, but helps clearance of FDPs which in themselves act as potent anticoagulants. It has also been suggested that FDPs inhibit myometrial activity, and serious post-partum haemorrhage in women with abruptio placentae was found to be associated with high levels of FDPs.29,30 High levels of FDPs may also have a cardiotoxic eect, resulting in low cardiac output and blood pressure despite a normal circulating blood volume (Figure2). If the fetus is dead, the aim should be prompt vaginal delivery avoiding soft-tissue damage, once correction of hypovolaemia is underway. Following emptying of the uterus, myometrial contraction will greatly reduce bleeding from the placental site, and spontaneous correction of the haemostatic defect usually occurs shortly after delivery, if the measures recommended above have been taken. However, post-partum haemorrhage is a not an infrequent complication and is the commonest cause of death in abruptio placenta.7 In cases where the abruption is small and the fetus is still alive, prompt Caesarean section may save the baby, if vaginal delivery is not imminent. FFP, bank red cells and platelet concentrates should be available to correct the potentially severe maternal coagulation defect. In rare situations where vaginal delivery cannot be stimulated and haemorrhage continues, Caesarean section is indicated even in the presence of a dead fetus. In these circumstances normal haemostasis should be restored as far as possible by the administration of FFP and platelet concentrates if necessary, as well as transfusing red cells before surgery is undertaken. Despite extravasion of blood throughout the uterine muscle, the function of the uterine muscle is not impaired, and good contraction will follow removal of fetus, placenta and retroperitoneal clot. Regional anaesthesia or analgesia is contraindicated. Expansion of the lower limb vascular bed resulting from regional block can add to the problem of uncorrected hypovolaemia. In the presence of haemostatic failure there is the additional hazard of bleeding into the epidural space. In recent years, heparin has been used to treat many cases of DIC, whatever their cause. There is, however, no objective evidence to demonstrate that its use in abruptio placentae decreased morbidity and mortality although anecdotal reports in the older literature continued to suggest this.31 Very good results have been achieved without the use of heparin.32 Its use, with an intact circulation, would be sensible and logical to break the vicious circle of DIC, but in the presence of already defective haemostasis with a large bleeding placental site it may prolong massive local and generalized haemorrhage.33 Treatment with anti®brinolytic agents such as EACA or Trasylol (Aprotinin) can result in blockage of small vessels of vital organs, such as the kidney or brain, with ®brin. It has been suggested30,34 that Trasylol may be helpful in the management of abruptio placentae, particularly in those cases with uterine inertia associated with high levels of FDPs. There is a high incidence (1.5%) of abruptio placentae in the obstetric admissions (18 000/annum) at the Groote Schuur Obstetric Unit in Cape Town where the ®rst study was carried out. The selection of Trasylol depended on its alleged anticoagulant activity in addition to its well-known anti®brinolytic properties.35 There has been a resurgence of use of Trasylol recently, particularly in cardiac surgery.36 where signi®cant reduction in blood loss has been shown following cardiac bypass operations. This is thought to be due predominantly to platelet sparing. It is doubtful whether Trasylol would have any advantage in management of obstetric DIC. Disseminated intravascular coagulation 633 Obstetricians appear unconvinced of the bene®ts of Trasylol in the treatment of DIC and abruptio placentae. Prompt supportive measures alone, maintaining central venous pressure and replacing blood loss together with essential coagulation factors, will of course result in reduction in FDPs. This will improve myometrial function and contribute to the return of healthy haemostasis. A report on reducing the frequency of severe abruption from Dallas, Texas37 noted that the reduction in fetal death associated with abruption by 50% over a period covering more than 30 years (1958±90) could be accounted for by the decrease in women of very high parity and to an increase in the proportion of Latin-American as opposed to black women in the population served. Abruption in the latter part of the study recurred in 12% of subsequent pregnancies and proved fatal to the fetus in 7%, which was unchanged from earlier experience. With modern supportive measures, maternal death due directly to abruption is now extremely rare. Placenta accreta The morbid adherence of the placenta to the uterus results from the invasion of placental villi into the myometrium without underlying decidua and is described as accreta, increta or pancreta, depending on the extent of the penetration. It usually presents in the early post-partum period with massive haemorrhage during dicult removal of the placenta. It is associated with placenta praevia and with uterine scar from a previous Caesarean section38 but can occur in a normally sited placenta without any previous scar. The preferred management of this condition in the presence of severe haemorrhage is hysterectomy, but arterial embolization is being used increasingly with preservation of the uterus22,23 (see Chapter 5). A recent report of controlling haemorrhage by subendometrial vasopressin infusion during Caesarean section in six women from one centre, thus avoiding hysterectomy, deserves further investigation.39 As the incidence of Caesarean section rises world wide this complication of pregnancy is likely to increase, and in many centres placenta accreta is becoming the main cause of massive intractable haemorrhage at delivery, leading to hysterectomy.40 In one centre, a review of emergency hysterectomy in obstetric practice showed a reduction over the decade 1985 to 1994 but a changing pattern in the primary cause. In the ®rst 5 years it was uterine atony (42%) followed by placenta accreta (25.5%). However in the second period, 1990±94, the predominant indication was placenta accreta (41.7%) followed by uterine atony (29.2%).41 In spite of every precaution and prompt management, placenta accreta remains a potent cause of the rare maternal mortality in the developed world.41,42 It is also, in my experience, a relatively common cause of DIC resulting from shock due to massive haemorrhage before hysterectomy is undertaken. Amniotic ¯uid embolism7,43,44 This obstetric disaster usually occurs during or shortly after a vigorous labour with an intact amniotic sac, but can occur during a Caesarean section. It is thought that amniotic ¯uid enters the maternal circulation via lacerations of membranes and placenta. Platelet ®brin thrombi are formed and trapped within the pulmonary blood vessels: profound shock follows, accompanied by respiratory distress and cyanosis. There is a high mortality at this stage from a combination of respiratory and cardiac failure.45 If the mother survives long enough, massive intravascular coagulation with almost total 634 E. A. Letsky consumption of coagulation factors invariably follows. There is bleeding from venepuncture sites and severe haemorrhage from the placental site after delivery. Con®rmation of diagnosis is usually made post-mortem by ®nding histological evidence of amniotic ¯uid and fetal tissue within the substance of the maternal lungs; occasionally, similar material may be aspirated from a central venous pressure (CVP) catheter line ± see below. It is, therefore, dicult to assess the value of therapeutic measures taken, in the few reports which have appeared, for the successful management of a clinical picture which can usually only be suggestive of amniotic ¯uid embolism.46±50 A signi®cantly reduced maternal mortality rate was reported in a population-based study in California where the clinical diagnosis of amniotic ¯uid embolism in survivors was included.49 The major dierential diagnoses of amniotic ¯uid embolism in the collapsed patient are primary cardiovascular catastrophes such as pulmonary embolus or aspiration in the anaesthetized patient. Apart from the bleeding and evidence of DIC associated with amniotic ¯uid embolism, pulmonary embolus has speci®c features. Aspiration is usually associated with bronchospasm which is very rare in amniotic ¯uid embolus.43 At any time, if there is doubt about the diagnosis, the rapid ¯uid infusion necessary for the treatment of amniotic ¯uid embolus should be controlled by careful assessment of the CVP. Such rapid ¯uid infusion would cause a marked rise in CVP in patients with pulmonary embolus, and could well lead to fatal ¯uid overload. In addition, the presence of the CVP line will allow aspiration of fetal material from the great veins, supportive of the diagnosis of amniotic ¯uid embolus.50 Such material can also be found in maternal sputum.51 Since fetal material has been found in pulmonary arterial blood in some women who did not have amniotic ¯uid embolus.52 such a ®nding should not be considered pathognomonic for amniotic ¯uid embolus. The object of the treatment is to sustain the circulation while the intravascular thrombin in the lungs is cleared by the ®brinolytic response of the endothelium of the pulmonary vessels. If bleeding from the placental site can be controlled by stimulation of uterine contraction, then the logical treatment is carefully monitored transfusion of FFP and packed red cells with heparin administration and, if necessary, ventilation. Retention of the dead fetus The question of intrauterine fetal death and haemostatic failure has been reviewed in the past.53 There is a gradual depletion of maternal coagulation factors following intrauterine fetal death, and the changes are not usually detectable in vitro until after 3±4 weeks. Thromboplastic substances released from the dead tissues in the uterus into the maternal circulation are thought to be the trigger of DIC in this situation, which occurs in about one-third of patients who retain the dead fetus for more than 4±5 weeks. Problems arising from defective haemostasis are not observed in modern obstetric practice because labour is induced promptly following diagnosis of fetal death before clinically signi®cant coagulation changes have developed. Rupture of the membranes is recommended once induced labour is established in such patients, as there is a risk of precipitate labour and amniotic ¯uid embolism has been known to occur. Retained dead fetus and living twin The occurrence of single fetal death in a pre-term multiple pregnancy poses unique therapeutic dilemmas. The incidence of this problem is unknown but is likely to be Disseminated intravascular coagulation 635 observed more frequently with the advent of widespread use of ultrasound in obstetrics. In addition, on occasion, selective termination of the life of the aected twin is being oered in situations where only one fetus has been shown to be aected with a genetic disorder or involved in twin/twin transfusion syndrome. Haemostatic failure appears to be a hazard for the remaining fetus rather than for the mother. Induced abortion Changes in haemostatic components consistent with DIC have been demonstrated in patients undergoing abortion induced with hypertonic solutions of saline and urea.54±57 This combination appears to be particularly hazardous58 in comparison to the use of urea and prostaglandin or oxytocin.59 The stimulus appears to be the release of tissue factor into the maternal circulation from the placenta, which is damaged by the hypertonic solutions. In later pregnancy DIC has been described with both dilatation and evacuation60 and also with prostaglandin and oxytocin termination.61 The haemorrhage resulting may be massive and has resulted in maternal deaths. Prompt restoration of the blood volume and transfusions with red cells and FFP, as described above, should resolve the situation which, once the uterus is empty, is self-limiting. A unique case of DIC associated with chronic ectopic pregnancy has been reported.62 Intrauterine infection63 Endotoxic shock associated with septic abortion and ante-partum or post-partum intrauterine infection can trigger DIC.64,65 Infection is usually with Gram-negative organisms. Fibrin is deposited in the microvasculature owing to endothelial damage by the endotoxin, and secondary red cell intravascular haemolysis with characteristic fragmentation, so-called micro-angiopathic haemolysis, is characteristic of the condition. The patient is usually alert and ¯ushed with a rapid pulse and low blood pressure. Transfusion has little or no eect on the hypotension in comparison to its bene®t in the obstetric emergencies complicated by DIC triggered by hypovolaemic shock. Elimination of the uterine infection remains the most important aspect of management; this is probably best performed by a short intensive period of antibiotic therapy followed by evacuation of the uterine contents. If the uterus is empty and contracted, there is no undue risk of severe bleeding from the placental site. If there is evidence of a consumptive coagulopathy, heparin may be useful as part of the management of this hazardous emergency63,66,67 but this remains controversial. In recent years this complication of pregnancy is rarely seen. If not recognized and treated promptly it is associated with a very high mortality rate. Purpura fulminans This rare complication of infection sometimes occurs in the puerperium, precipitated by Gram-negative septicaemia. Extensive haemorrhage occurs into the skin in association with DIC. The underlying mechanism is unknown but there appears to be an acute activation of the clotting system resulting in the deposition of ®brin thrombi within blood vessels of the skin and other organs.68 The extremities and face are usually involved ®rst, the purpuric patches having a jagged and erythematous border, which can be shown histologically to be the site of a leukocytoclastic vasculitis. Rapid enlargement of the lesions which become necrotic and gangrenous is associated with 636 E. A. Letsky shock, tachycardia and fever. Without treatment the mortality rate is high, and among those who survive, digit or limb amputation may be necessary. The laboratory ®ndings are those of DIC with leukocytosis. In this situation treatment with heparin should be started as soon as the diagnosis is apparent. It will prevent further consumption of platelets and coagulation factors. It should always be remembered, however, that bleeding from any site in the presence of defective coagulation factors will be aggravated by the use of heparin. Survival in purpura fulminans is currently much improved because of better supportive treatment for the shocked patient and eective control of the triggering infection, together with heparin therapy. Low-grade DIC, pre-eclampsia and related syndromes Rampant uncompensated DIC results in severe haemorrhage with the characteristic laboratory ®ndings described above. However, low-grade DIC does not usually give rise to any clinical manifestations, although the condition is a potentially hazardous one for both mother and fetus (Table 1). In vitro detection of low-grade DIC Many in vitro tests have been claimed to detect low-grade compensated DIC, and space does not allow an account of all of these. FDPs. Estimation of FDPs will give some indication of low-grade DIC if these are signi®cantly raised when ®brinogen, platelets and screening tests of haemostatic function appear to be within the normal range. Soluble ®brin complexes. The action of thrombin on ®brinogen is crucial in DIC. Thrombin splits two molecules of ®brinopeptide A and two molecules of ®brinopeptide B from ®brinogen. The remaining molecule is called a ®brin monomer and polymerizes rapidly to ®brin (Figure 5). Free ®brinopeptides in the blood are a speci®c measure of thrombin activity and high levels of ®brinopeptide A have been shown to be associated with compensated DIC in pregnancy.6 Soluble ®brin complexes made up of ®brin±®brinogen dimers are increased in conditions of low-grade DIC.69 These complexes are generated during the process of thrombin generation and the conversion of soluble ®brinogen to insoluble ®brin (Figure 5). Levels of soluble ®brin complexes are increased in patients with severe preeclampsia and with a retained dead fetus.9,70 Factor VIII. During normal pregnancy the levels of both von Willebrand factor (vWF) and Factor VIII coagulation activity (VIIIC) rise in parallel.71,72 An increase in the ratio of vWF to Factor VIIIC has been observed in conditions accompanied by low-grade DIC, whether associated with pregnancy or not. The stages in the spectrum of severity of DIC (see Table 1) are not strictly delineated, and there may be rapid progression from low-grade compensated DIC, as diagnosed by paracoagulation tests described above, to the rampant form with haemostatic failure. Many investigators now believe that gestational hypertension with or without proteinuria, intrauterine growth retardation (IUGR), haemolysis, elevated liver enzymes and low platelets (HELLP) syndrome, acute fatty liver of pregnancy (AFLP) and eclampsia are part of the same disease process73 which presents with the related Disseminated intravascular coagulation 637 Prothrombin (II) (IIa) Thrombin Fibrinogen Fibrin monomers + fibrinopeptides A+B XIII XIIIa Fibrin polymers Ca++ Fibrin Figure 5. The ®nal common pathway of the coagulation cascade: conversion of ®brinogen to ®brin induced by thrombin. signs and symptoms depending on the organ targeted by small-vessel coagulopathy and varying degrees of DIC. Haemostatic changes in pre-eclampsia (Table 2) There have been many reports and reviews showing that the circulating platelet count is reduced in pre-eclampsia.74,75 The platelet count can be used to monitor severity of the disease process as well as an initial screening if there is concern about signi®cant coagulation abnormalities.76 A fall in the platelet count may precede any detectable rise in serum FDPs in women subsequently developing pre-eclampsia. The combination of a reduced platelet life span and a fall in the platelet count without platelet-associated antibodies (see below) indicates a low-grade coagulopathy. Platelets may either be consumed in thrombus formation or may suer membrane damage from contact with abnormal surfaces and be prematurely removed from the circulation. Rarely, in very severe pre-eclampsia, the patient develops micro-angiopathic haemolytic anaemia. These patients have profound thrombocytopenia and this leads to confusion in the dierential diagnosis between pre-eclampsia, haemolysis, elevated liver enzymes, low platelet count syndrome (HELLP)77 and thrombotic thrombocytopenic purpura (TTP) (see below). Activation of the haemostatic mechanisms in normal pregnancy has led to the view that the haematologic manifestations of pre-eclampsia merely represent augmentation of the hypercoagulable state which accompanies normal pregnancy. In this respect many studies have been carried out on levels of individual coagulation factors. No clear Table 2. Haemostatic changes in pre-eclampsia. . . . . . . a Increased ratio von Willebrand factor (VWF)/factor VIIIC Decreased prostacyclin generation Decreased plasma antithrombin (AT) concentration Increased soluble ®brinogen products ± particularly ®brinopeptide A Decreasing platelet counta Increased serum ®brin degradation productsa Most useful markers of severity and outcome. 638 E. A. Letsky Non-Pregnant State At Delivery Rapid blood loss leads to vasoconstriction to compensate for drop in blood volume. Plasma volume increase over next few days. Hypervolaemia modifies response. Remain vasodilated even if loss excessive. Fall in plasma volume due to diuresis in puerperium. Fall in haematocrit proportionate to blood loss. Haematocrit gradually increases as blood volume returns to normal. Figure 6. The unique maternal response to acute blood loss at parturition. pattern emerges but there appear to be some signi®cant correlations of severity of the disease process with both the Factor VIII complex.78 and antithrombin (AT).79 A readily available and sensitive indicator of activation of the coagulation system is assay of ®brinopeptide A concentration in the plasma. Although in mild pre-eclampsia patients may have a normal or only slight increase in ®brinopeptide A levels, marked increases occur in patients with severe pre-eclampsia6 (Figure 5). Most studies in pre-eclampsia have shown increased levels of FDPs in serum and urine. Plasma levels of soluble ®brinogen±®brin complexes are also raised in preeclampsia compared with normal pregnancies. Once the disease process is established the most relevant coagulation abnormalities appear to be the platelet count, Factor VIII and FDPs. Those women with the most marked abnormalities in these parameters suer the greatest perinatal loss.80 In reality, most units use serial platelet counts to monitor the activity of disease. It has been shown that coagulation abnormalities do not occur unless the platelet count falls below 100 109/l.76 Following the reported association between thrombophilia factors, severe preeclampsia and intrauterine growth retardation (IUGR)81 it has been suggested that investigation of pre-eclampsia should also include a search for defects in the naturally occurring anticoagulants so that future pregnancies can be managed appropriately. Acute fatty liver of pregnancy (AFLP) AFLP is a rare complication of pregnancy included in this section because it is often, if not always, associated with variable degrees of DIC ± which contributes signi®cantly to its morbidity and mortality.33,82 Only the haematological aspects will be discussed here. Early diagnosis and subsequent delivery are essential for improving survival of both mother and child. Most patients have prodromal symptoms for at least 1 week before jaundice develops. The Royal Free series83 draws attention to a characteristic blood picture of neutrophilia, thrombocytopenia and normoblasts. Some of the blood ®lms available for review also showed basophilic stippling and giant platelets and the authors suggest that these appearances might help towards an early diagnosis of AFLP. However, these features are not speci®c to AFLP and may be seen in any condition of additional stress on a bone marrow already working to capacity in the last trimester of pregnancy. DIC complicating severe liver failure is an extremely complex topic. In AFLP the haemostatic defect is frequently resistant, probably owing to prolonged activation of coagulation combined with very low to undetectable AT levels.33,84,85 The replacement Disseminated intravascular coagulation 639 of AT with plasma or AT concentrate to shorten the period of DIC, and thereby decrease morbidity and mortality of AFLP, has been suggested.33 AT concentrate has also been used in the successful management of a patient with AFLP.85 Heparin therapy can be very dangerous.86 Thrombotic thrombocytopenic purpura (TTP) and haemolytic-uraemic syndrome (HUS) These conditions, extremely rare in pregnancy, share so many features that they have been considered as one disease in the past, with pathological eects con®ned largely to the kidney in HUS and being more generalized in TTP. Both are due to the presence of platelet thrombi in the microcirculation that cause ischaemic dysfunction and microangiopathic haemolysis. They are included here because they can be confused with severe pre-eclampsia. Unlike severe pre-eclampsia, HELLP and AFLP, there is no evidence that prompt delivery aects the course of HUS or TTP favourably. However, most clinicians would recommend delivery if these conditions are present in late pregnancy so that the mother can be treated vigorously without fear of harming the fetus. Therapeutic strategies hinge on intensive plasma exchange or replacement. It is essential that the correct diagnosis is made to obtain optimal outcome. The pentad of fever, normal coagulation tests with low platelets, haemolytic anaemia, neurological disorders and renal dysfunction are virtually pathogonomic of TTP. The thrombocytopenia may range from 30 to 100 109/l. The clinical picture is severe, with a high maternal mortality. A crucial problem when dealing with TTP is to establish a correct diagnosis, because this condition can be confused with severe preeclampsia and HELLP syndrome, especially if DIC (very uncommon in TTP) is triggered. All of these conditions are associated with thrombocytopenia and a characteristic macroangiopathic blood ®lm with red cell fragments and burr cells. TTP is associated with abnormal patterns of multimers of von Willebrand (vWF) factor in the plasma. A proteolytic enzyme present in normal plasma cleaves peptide bonds in monomeric subunits of von Willebrand factor, thus degrading the large multimers. This proteolytic enzyme prevents the unusually large multimers from causing vWF-mediated adhesion of platelets to sub-endothelium after vascular damage.87 Exciting new research by two groups of investigators has elucidated the role of vWF-cleaving protease activity in TTP. In all the patients studied with acute single episode TTP there was little or no vWF-cleaving protease activity associated with an IgG auto-antibody during the illness but the activity returned to normal following recovery.88 In patients with chronic relapsing TTP unusually large multimers of vWF are found in the plasma, even between acute episodes, together with absence or low levels of vWF-cleaving protease activity.89 What are the implications of these new ®ndings? . The diagnosis of TTP which has been confused with severe pre-eclampsia and HELLP syndrome may be facilitated by a single laboratory test measuring the vWF-cleaving protease activity in the plasma and its inhibitor. . This test will also help clinicians to distinguish between TTP and HUS, thought in the past to be a spectrum of clinical expressions of the same disease. In HUS the levels of vWF-cleaving protease are normal or approaching normal, and this also explains why plasma exchange has disappointing results in HUS. . The response of patients in the past to fresh frozen plasma component infusions which contain the vWF cleaving protease activity is explained. 640 E. A. Letsky . The fact that plasmapheresis is often required in the acute phase to achieve a response is also clari®ed as this will remove auto-antibodies as well as the abnormal large vWF multimers. In summary, objective diagnosis of TTP should now be facilitated by measurement of vWF-cleaving protease and its inhibitor. The basis of treatment depends on supplying the defective missing enzyme by infusion of plasma or cryosupernatant, preferably in severe cases, by exchange transfusion. Virally inactivated plasma products retain vWFcleaving protease activity, ensuring the safety of this treatment. SUMMARY ± DISSEMINATED INTRAVASCULAR COAGULATION As emphasized, DIC is always a secondary phenomenon, and the mainstay of management is therefore to remove the initiating stimulus if possible.90 Low-grade DIC may be con®rmed by ®nding increased FDPs and ®brin monomer complexes with a falling platelet count in the presence of normal coagulation screening tests. With rampant DIC and haemorrhage, increased thrombin time and a low ®brinogen level will con®rm the coagulopathy and help to monitor the success of treatment measures. Recovery will usually follow delivery of the patient provided the blood volume is maintained and shock due to hypovolaemia is prevented. An eciently acting Practice points Aims . to recognize the mother who has excessive blood loss peripartum in the early stages so that appropriate replacement therapy will be given promptly in order to prevent the development of DIC which is not present initially in the majority of cases of abnormal obstetric haemorrhage . to recognize low-grade DIC, particularly in pre-eclampsia and related syndromes, by appropriate monitoring, and to take measures to deal with the trigger Management . the management of DIC depends on dealing with the trigger and has to be modi®ed depending on the speci®c problem . the therapy of fulminant DIC with haemostatic failure should be automatic with a frequently rehearsed protocol. Replacement therapy with crystalloids, colloids and blood components is the same for massive obstetric haemorrhage, whether or not there is coagulation failure initially Investigations . for chronic low-grade DIC, rising FDPs and falling platelets are probably the best rapidly available investigations. Serum ®brin monocomplexes and ®brinopeptides may give additional supportive information. Examination of a blood ®lm may reveal large platelets, indicating increased turnover and, less frequently, microangiopathic red cell changes . for fulminant DIC, an FBC with platelets, a coagulation screen with particular emphasis on the thrombin time and the ®brinogen concentration will identify coagulation failure. Blood can be taken for FDPs but should be measured after the emergency is over Disseminated intravascular coagulation 641 Research agenda . the evaluation of the measurement of thrombomodulin for the diagnosis of placenta abruptio28 . the value of sub-endometrial vasopressin infusion during Caesarean section to control intractable placental haemorrhage39 . regular audit of management of massive obstetric haemorrhage in every centre with exchange of information because this problem does not lend itself to randomized controlled trials myometrium post-delivery will stem haemorrhage from the placental site. Measures taken to achieve a ®rm contracted uterus will obviously contribute one of the most important factors in preventing continuing massive blood loss. Supportive measures must include urgent restoration of blood volume with a plasma substitute followed rapidly by blood component therapy to replace blood loss and correct the coagulopathy. Each obstetric unit must have frequent rehearsals and an established protocol for this emergency. It is of interest that the maternal mortality of DIC associated with placental abruption is less than 1%, whereas that associated with infection and shock is 50±80%. The mortality rate reported in series of patients with DIC due to various aetiologies is 50±85% and the wide variation probably re¯ects the mortality rate of the underlying disorder, not of DIC per se.33 There is no doubt that the major determinant of survival is our ability to identify the underlying trigger and manage it successfully. REFERENCES * 1. Forbes CD & Greer IA. Physiology of haemostasis and the eect of pregnancy. In Greer IA, Turpie AGG & Forbes CD (eds) Haemostasis and Thrombosis in Obstetrics and Gynaecology, pp 1±25. London: Chapman & Hall, 1992. 2. Stirling Y, Woolf L, North WR et al. Haemostasis in normal pregnancy. Thrombosis and Haemostasis 1984; 52: 176±182. 3. Levi M & Ten Cate H. Disseminated Intravascular Coagulation. New England Journal of Medicine 1999; 341: 586±592. 4. Sheppard BL & Bonnar J. The ultrastructure of the arterial supply of the human placenta in early and late pregnancy. Journal of Obstetrics and Gynaecology of the British Commonwealth 1974; 81: 497±511. 5. Gilabert J, Aznar J, Parilla J et al. Alteration in the coagulation and ®brinolysis system in pregnancy, labour and puerperium, with special reference to a possible transitory state of intravascular coagulation during labour. Thrombosis and Haemostasis 1978; 40: 387±396. 6. Wallmo L, Karlsson K & Teger Nilsson AC. Fibrinopeptide A and intravascular coagulation in normotensive and hypertensive pregnancy and parturition. Acta Obstetricia et Gynecologica Scandinavica 1984; 63: 637±640. 7. Department of Health. Why Mothers Die. Report on Con®dential Enquiries into Maternal Deaths in the United Kingdom 1994±1996. London: HMSO, 1998. * 8. Bonnar J. Massive obstetric haemorrhage. BaillieÁreÂs Clinical Obstetrics and Gynaecology 2000; 14: 1±18. * 9. Ostund E, Bremme K & Wiman B. Soluble ®brin in plasma as a sign of activated coagulation in patients with pregnancy complications. Acta Obstetricia et Gynecologica Scandinavica 1998; 77: 165±169. 10. Hewitt PE & Machin SJ. Massive blood transfusion. In Contreras M (ed.) ABC of Tranfusion, 3rd edn, 49±52. London: BMJ Books, 1998. 11. Doenicke A, Grote B & Lorenz W. Blood and blood substitutes in management of the injured patient. British Journal of Anaesthesia 1977; 49: 681±688. 12. Barbier P, Jonville AP, Autret E & Coureau C. Fetal risks with dextrans during delivery. Drug Safety 1992; 7: 71±73. 642 E. A. Letsky 13. Berg EM, Fasting S & Sellevold OF. Serious complications with dextran-70 despite hapten prophylaxis. Is it best avoided prior to delivery? Anaesthesia 1991; 46: 1033±1035. 14. RCOG Working Party.Report of the RCOG Working Party on Prophylaxis against Thromboembolism in Gynaecology and Obstetrics. London: The Royal College of Obstetricians and Gynaecologists, 1995. 15. Carey LC, Cloutier CT & Lowery BD. The use of balanced electrolyte solution for resuscitation. In Fox CL & Nahas GG (eds) Body Fluid Replacement in the Surgical Patient. New York: Grune & Stratton, 1970. 16. Moss G. An argument in favour of electrolyte solutions for early resuscitation. Surgical Clinics in North America 1972; 52: 3±17. 17. Virgilio RWK, Rice CL & Smith DE et al. Crystalloid versus colloid resuscitation: is one better? A randomized clinical study. Surgery 1979; 85: 129±139. 18. Lorenz W, Doenicke A, Messmer K et al. Histamine release in human subjects by modi®ed gelatin (Haemaccel) and dextran: an explanation for anaphylactoid reactions observed under clinical conditions. British Journal of Anaesthesia 1976; 48: 151±165. 19. Freeman M. Fatal reaction to haemaccel. Anaesthesia 1979; 34: 341±343. *20. Williamson L, Heptonstall J & Soldan K. A SHOT in the arm for safer blood transfusion. British Medical Journal 1996; 313: 1221±1222. 21. Schinzel H & Weilemann LS. Antithrombin substitution therapy. Blood coagulation and ®brinolysis 1998; 9: S17±S21. *22. Hansch E, Chitkara U, McAlpine J et al. Pelvic arterial embolization for control of obstetric hemorrhage: a ®ve-year experience. American Journal of Obstetrics and Gynecology 1999; 180: 1454±1460. 23. Pelage J-P, Le Dref O, Jacob D et al. Selective arterial embolization of the uterine arteries in the management of intractable post-partum hemorrhage. Acta Obstetricia et Gynecologica Scandinavica 1999; 78: 698±703. 24. Redman CWG. Coagulation problems in human pregnancy. Postgraduate Medical Journal 1979; 55: 367±371. 25. Gilabert J, Estelles A, Aznar J & Galbis M. Abruptio placentae and disseminated intravascular coagulation. Acta Obstetricia et Gynecologica Scandinavica 1985; 64: 35±39. 26. Naumann RO & Weinstein L. Disseminated intravascular coagulation ± the clinician's dilemma. Obstetrical and Gynecological Survey 1985; 40: 487±492. 27. Estelles A, Gilabert J, Espana F et al. Fibrinolysis in preeclampsia. Fibrinolysis 1987; 1: 209±214. 28. Magriples U, Chan DW, Bruzek D et al. Thrombomodulin: a new marker for placental abruption. Thrombosis and Haemostasis 1999; 81: 32±34. 29. Basu HK. Fibrinolysis and abruptio placentae. Journal of Obstetrics and Gynaecology of the British Commonwealth 1969; 76: 481±496. 30. Sher G. Pathogenesis and management of uterine inertia complicating abruptio placentae with consumption coagulopathy. American Journal of Obstetrics and Gynecology 1977; 129: 164±170. 31. Thragarajah S, Wheby Ms, Jarn R et al. Disseminated intravascular coagulation in pregnancy. The role of heparin therapy. Journal of Reproductive Medicine 1981; 26: 17±24. 32. Pritchard JA. Haematological problems associated with delivery, placental abruption, retained dead fetus and amniotic ¯uid embolism. Clinics in Haematology 1973; 2: 563±580. 33. Feinstein DI. Diagnosis and management of disseminated intravascular coagulation: the role of heparin therapy. Blood 1982; 60: 284±287. 34. Sher GI & Statland BE. Abruptio placentae with coagulopathy: a rational basis for management. Clinical Obstetrics and Gynecology 1985; 28: 15±23. 35. Amris CJ & Hilden M. Anticoagulant eects of Trasylol: in vitro and in vivo studies. Annals of the New York Academy of Sciences 1968; 146: 612±694. 36. Bidstrup BP, Royston D, Sapsford RN & Taylor KM. Reduction in blood loss and blood use after cardiopulmonary bypass with high dose aprotinin (Trasylol). Journal of Thoracic and Cardiovascular Surgery 1989; 97: 364±372. *37. Pritchard JA, Cunningham FG, Pritchard SA & Mason RA. On reducing the frequency of severe abruptio placentae. American Journal of Obstetrics and Gynecology 1991; 165: 1345±1351. 38. Manyonda IT & Varma TR. Massive obstetric hemorrhage due to placenta previa/accreta with prior cesarean section. International Journal of Gynaecology and Obstetrics 1991; 34: 183±186. 39. Lurie S, Appleman Z & Katz Z. Subendometrial vasopressin to control intractable placental bleeding. Lancet 1997; 349: 698. 40. Zaki ZMS, Bahar AM, Ali ME et al. Risk factors and morbidity in patients with placenta previa accreta compared to placenta previa non-accreta. Acta Obstetricia et Gynecologica Scandinavica 1998; 77: 391±394. 41. GuÈrkan Zorlu C, Turan C, Isik AZ et al. Emergency hysterectomy in modern obstetric practice ± changing clinical perspective in time. Acta Obstetricia et Gynecologica Scandinavica 1998; 77: 186±190. 42. Greene MF, Roberts DJ & Mark EJ. Cardiovascular collapse after vaginal delivery in a patient with a history of Cesarean section. New England Journal of Medicine 1998; 338: 821±826. Disseminated intravascular coagulation 643 43. Morgan M. Amniotic ¯uid embolism. Anaesthesia 1979; 34: 20±32. 44. Clark SL, Hankins GDV, Dudley DA et al. Amniotic ¯uid embolism: analysis of the national registry. American Journal of Obstetrics and Gynecology 1995; 172: 1158±1169. 45. Herbert WNP. Complications of the immediate puerperium. Clinical Obstetrics and Gynecology 1982; 25: 219±232. 46. Chung AF & Merkatz IR. Survival following amniotic ¯uid embolism with early heparinisation. Obstetrics and Gynecology 1973; 42: 809±814. 47. Skjodt P. Amniotic ¯uid embolism ± a case investigated by coagulation and ®brinolysis studies. Acta Obstetricia et Gynecologica Scandinavica 1965; 44: 437±457. *48. Fletcher SJ & Parr MJ. Amniotic ¯uid embolism: a case report and review. Resuscitation 2000; 43: 141±146. 49. Gilbert WM & Danielsen B. Amniotic ¯uid embolism: decreased mortality in a population based study. Obstetrics and Gynecology 1999; 93: 973±977. 50. Resnik R, Swartz WH, Plumer MH et al. Amniotic ¯uid embolism with survival. Obstetrics and Gynecology 1976; 47: 395±398. 51. Tuck CS. Amniotic ¯uid embolus. Proceedings of the Royal Society of Medicine 1972; 65: 94±95. 52. Clark SL, Pavlova Z, Greenspoon J et al. Squamous cells in the maternal pulmonary circulation. American Journal of Obstetrics and Gynecology 1986; 154: 104±106. 53. Romero R, Copel JA & Hobbins JC. Intrauterine fetal demise and hemostatic failure: the fetal death syndrome. Clinical Obstetrics and Gynecology 1985; 28: 24±31. 54. Grundy MFB & Craven ER. Consumption coagulopathy after intra-amniotic urea. British Medical Journal 1976; 2: 677±678. 55. Mackenzie IZ, Sayers L, Bonnar J et al. Coagulation changes during second trimester abortion induced by intra-amniotic prostaglandin E2 and hypertonic solutions. Lancet 1975; ii: 1066±1069. 56. Spivak JL, Spangler DB & Bell WR. De®brinogenation after intraamniotic injection of hypertonic saline. New England Journal of Medicine 1972; 287: 321±323. 57. Stander RW, Flessa HC & Glueck HI Changes in maternal coagulation factors after intraamniotic injection of hypertonic saline. Obstetrics and Gynecology 1971; 37: 660±666. 58. Clarkson AR, Sage RE & Lawrence JR. Consumption coagulopathy and acute renal failure due to Gram negative septicaemia after abortion. Complete recovery with heparin therapy. Annals of Internal Medicine 1969; 70: 1191±1199. 59. Burkman RT, Bell WR, Atizenza MF et al. Coagulopathy with midtrimester induced abortion. Association with hyperosmolar urea administration. American Journal of Obstetrics and Gynecology 1977; 127: 533±536. 60. Davis G. Midtrimester abortion. Late dilation and evacuation and DIC. Lancet 1972; ii: 1026. 61. Savage W. Abortion: methods and sequelae. British Journal of Hospital Medicine 1982; 27: 364±384. 62. Collier CB & Birrell WRS. Chronic ectopic pregnancy complicated by shock and disseminated intravascular coagulation. Anaesthesia in Intensive Care 1983; 11: 246±248. 63. Beller FK. Sepsis and coagulation. Clinical Obstetrics and Gynecology 1985; 28: 46±52. 64. Grae H, Ernst F & Bocaz JA. Evaluation of hypercoagulability in septic abortion. Haemostasis 1976; 5: 285±294. 65. Steichele DF & Herschlein HJ. Intravascular coagulation in bacterial shock. Consumption coagulopathy and ®brinolysis after febrile abortion. Medizinische Welt 1968; 1: 24±30. 66. Beller FK & Uszynski M. Disseminated intravascular coagulation in pregnancy. Clinical Obstetrics and Gynecology 1974; 17: 264±278. 67. Bonnar J. Haemorrhagic disorders during pregnancy. In Hathaway WE & Bonnar J (eds) Perinatal Coagulation Monographs in Neonatology. New York: Grune & Stratton, 1978. 68. McGibbon DH. Dermatological purpura. In Ingram GIC, Brozovic M & Slater NCP (eds) Bleeding Disorders ± Investigation and Management, pp 220±242. Oxford: Blackwell Scienti®c Publications, 1982. 69. Aznar J, Gilabert J, Estelles A et al. Evaluation of the soluble ®brin monomer complexes and other coagulation parameters in obstetric patients. Thrombosis Research 1982; 27: 691±701. 70. Hafter R & Grae H. Molecular aspects of de®brination in a reptilase treated case of `dead fetus syndrome'. Thrombosis Research 1975; 7: 391±399. 71. Fournie A, Monrozies M, Pontonnier G et al. Factor VIII complex in normal pregnancy, pre-eclampsia and fetal growth retardation. British Journal of Obstetrics and Gynaecology 1981; 88: 250±254. 72. Whigham KAE, Howie PW, Shaf MM & Prentice CRM. Factor VIII related antigen and coagulant activity in intrauterine growth retardation. Thrombosis Research 1979; 16: 629±638. 73. de Swiet M. Some rare medical complications of pregnancy. British Medical Journal 1985; 290: 2±4. 74. Baker P & Cunningham F. Platelet and coagulation abnormalities. In Lindheimer M, Roberts J & Cunningham F (eds) Chesley's Hypertensive Disorders in Pregnancy, pp 349±373. Stamford, CT: Appleton & Lange, 1999. 644 E. A. Letsky 75. Romero R, Mazor M, Lockwood CJ et al. Clinical signi®cance, prevalence, and natural history of thrombocytopenia in pregnancy-induced hypertension. American Journal of Perinatology 1989; 6: 32±38. *76. Leduc L, Wheeler JM, Kirshon B et al. Coagulation pro®le in severe preeclampsia. Obstetrics and Gynecology 1992; 79: 14±18. 77. Isler CM, Rinehart BK, Terrone DA, et al. Maternal mortality associated with HELLP (hemolysis, elevated liver enzymes, and low platelets) syndrome. American Journal of Obstetrics and Gynecology 1999; 181: 924±928. 78. Redman CW, Denson KW, Beilin LJ et al. Factor-VIII consumption in pre-eclampsia. Lancet 1977; ii: 1249±1252. 79. Weiner CP, Kwaan HC, Xu C et al. Antithrombin III activity in women with hypertension during pregnancy. Obstetrics and Gynecology 1985; 65: 301±306. 80. Thornton JG, Molloy BJ, Vinall PS et al. A prospective study of haemostatic tests at 28 weeks gestation as predictors of pre-eclampsia and growth retardation. Thrombosis and Hemostasis 1989; 61: 243±245. 81. Kupferminc MJ, Eldor A, Steinman N et al. Increased frequency of genetic thrombophilia in women with complications of pregnancy. New England Journal of Medicine 1999; 340: 9±13. 82. Laursen B, Frost L, Mortensen JZ et al. Acute fatty liver of pregnancy with complicating disseminated intravascular coagulation. Acta Obstetricia et Gynecologica Scandinavica 1983; 62: 403±407. 83. Burroughs AK, Seong NG, Dojcinoov DM et al. Idiopathic acute fatty liver of pregnancy in 12 patients. Quarterly Journal of Medicine 1982; 51: 481±497. 84. Hellgren M, Hagnevik K, Robbe H et al. Severe acquired antithrombin III de®ciency in relation to hepatic and renal insuciency and intra-uterine death in late pregnancy. Gynecologic and Obstetric Investigation 1983; 16: 107±108. 85. Laursen B, Mortensen J, Frost L & Hansen KB. Disseminated intravascular coagulation in hepatic failure treated with antithrombin III. Thrombosis Research 1981; 22: 701±704. 86. Goodin RC. Acute fatty liver of pregnancy. Acta Obstetricia et Gynecologica Scandinavica 1984; 63: 379±380. *87. Moake JL. Moschcowitz, Multimers, and Metalloprotease [Editorial]. New England Journal of Medicine 1998; 339: 1629±1631. 88. Tsai H & Lian E-Y. Antibodies to von Willebrand factor-cleaving protease in acute thrombotic thrombocytopenic purpura. New England Journal of Medicine 1998; 339: 1585±1594. 89. Furlan M, Robles R, Galbusera M et al. von Willebrand factor-cleaving protease in thrombotic thrombocytopenic purpura and the hemolytic-uremic syndrome. New England Journal of Medicine 1998; 339: 1578±1584. *90. Lurie S, Feinstein M & Mamet Y. Disseminated intravascular coagulopathy in pregnancy: thorough comprehension of etiology and management reduces obstetricians' stress. Archives of Gynecology and Obstetrics 2000; 263: 126±130.