Chapter 8: Hematology During Pregnancy

Maternal blood volume increases by an average of 40 to 50 percent above the nonpregnant level.¹ Plasma volume begins to rise early in pregnancy, with most of the escalation taking place in the second trimester and prior to week 32 of gestation.² Red cell mass increases significantly beginning in the second trimester and continues to expand throughout pregnancy, but to a lesser extent than plasma volume.² Erythropoietin levels increase throughout pregnancy, reaching approximately 150 percent of their prepregnancy levels at term.^3,4 The overall effect of these changes in most women is a slight drop in hemoglobin concentration, which is most pronounced at the end of the second trimester and slowly improves approaching term.

PLATELET AND WHITE CELL COUNTS

The effect of pregnancy on maternal platelet count is somewhat more controversial; some studies demonstrate a mild decline in platelet count over the course of gestation,⁵ whereas others do not.⁶ In general, white cell counts rise during pregnancy with the occasional appearance of myelocytes or metamyelocytes in the blood.⁷ During labor and the early puerperium, there is a rise in the leukocyte count. Leukocytosis appears to be linearly related to the duration of labor.⁸

PLASMA PROTEINS

The levels of some plasma proteins also increase during pregnancy. In particular, C-reactive protein concentration is higher in pregnant women and rises even further during labor.⁹ Erythrocyte sedimentation rate (ESR) rises during pregnancy, and is affected by both hemoglobin concentration and gestational age.¹⁰ The rise in ESR during pregnancy, in large part a result of an increase in levels of plasma globulins and fibrinogen, makes its use as a marker of inflammation difficult. The levels of many of the procoagulant factors increase during pregnancy whereas activity of the fibrinolytic system diminishes in preparation for the hemostatic challenge of delivery. Plasma levels of von Willebrand factor (VWF), fibrinogen, and factors VII, VIII, and X all increase markedly, whereas factors II, V, IX, and XII are essentially unchanged and factor XIII declines.¹¹ Levels of protein C and antithrombin remain stable throughout pregnancy whereas total and free protein S fall with increasing gestational age.¹² Fibrinolysis is also impaired by increases in plasminogen activator inhibitors I and II, the latter a product of the placenta.¹³

IRON DEFICIENCY

Worldwide, the contribution of anemia to maternal and fetal morbidity and mortality is well recognized; in some parts of Africa, more than 75 percent of pregnant women are anemic, and there is a significant correlation between maternal mortality and anemia.¹⁴ It has been suggested that iron deficiency may protect against placental malaria, but epidemiologic studies have not been conducted to verify this supposition.¹⁵ In pregnant women, anemia is defined as a hemoglobin concentration of less than 11 g/dL in the first and third trimesters, and less than 10.5 g/dL in the second trimester.¹⁵ In both the industrialized and the developing world, iron-deficiency anemia (Chap. 43) is the commonest cause of anemia.¹⁶ On average approximately 1 g of iron is required during a normal pregnancy; 300 mg of iron are required by the fetus and the placenta, whereas expansion of the maternal red cell mass requires 500 mg, and 200 mg are lost via excretion.¹⁷ These requirements exceed the iron storage of most young women and in general cannot be met by the diet. Even in cases of maternal iron deficiency, the fetal requirements for iron are always met; thus there is no correlation between the hemoglobin of the fetus and that of the mother.¹⁸

Iron-deficiency anemia during the first two trimesters of pregnancy is associated with a twofold increased risk for preterm delivery and a threefold increased risk for delivery of a low-birth-weight infant.¹⁹ However, a large randomized trial comparing routine iron prophylaxis in pregnancy versus iron supplementation given only as needed demonstrated no significant differences in adverse maternal or fetal outcomes.²⁰ As in nonpregnant individuals, iron-deficiency anemia can generally be diagnosed using laboratory values such as serum ferritin, and transferrin saturation levels (Chap. 43). Pica, the ingestion of nonnutritive substances, is said to be more common among iron-deficient pregnant women than among other populations with iron deficiency. Ice, clay or dirt, and starch are the most frequent substances ingested (Chap. 43); to some extent, however, the choice appears to be cultural and much more widespread than most practitioners realize.²¹

FOLATE AND VITAMIN B₁₂ DEFICIENCY

Apart from iron deficiency, folate deficiency is the next most frequent nutritional deficiency leading to anemia in pregnant women. In the United States, where foodstuffs are supplemented with folate and the level of awareness of the association between folate deficiency and neural tube defects in the embryo is high, folate deficiency is relatively unusual. Folate requirements in pregnancy are roughly twice those in the nonpregnant state (800 mcg/day vs. 400 mcg/day), and if diet is insufficient may exceed the body’s stores of folate (5–10 mg) relatively quickly.²² Anemia related to folate deficiency most often presents in the third trimester and responds to folate supplementation with reticulocytosis within 24 to 72 hours.¹⁶ Reports of severe pancytopenia and even states resembling the HELLP (hemolysis, elevated liver enzymes, and low platelet count) syndrome as a result of folate deficiency in pregnancy have appeared in the literature.^23,24 Despite these case reports, a review of 21 trials measuring the effect of folate supplementation on biochemical and hematologic parameters and pregnancy outcome (excluding neural tube defects) revealed improvement in low hemoglobin level in late pregnancy, but had no measurable effect on any substantive measures of pregnancy outcome (Chap. 41).²⁵

Vitamin B₁₂ (cobalamin) deficiency during pregnancy is rare, in part because deficiency of this vitamin leads to infertility. Serum cobalamin levels are known to fall during pregnancy.²⁶ A shift from the serum to tissue stores is proposed to account for the drop in serum B₁₂ levels. However, values less than 180 pmol/L usually are not observed in healthy women, and these low-normal levels are not accompanied by increased levels of methylmalonic acid, an indicator of cellular deficiency of cobalamin (Chap. 41).²⁷

RED CELL APLASIA

A rare cause of anemia in pregnancy is pure red cell aplasia (Chap. 36). In pure red cell aplasia, anemia tends to occur early in pregnancy and often resolves within weeks of delivery. The pathogenic mechanism leading to red cell aplasia does not appear to be transferred to the fetus, but does tend to recur in subsequent pregnancies.^28,29 Conservative treatment, if feasible, is probably best until delivery; successful prenatal treatments with glucocorticoids and with intravenous immunoglobulin have been reported.^30,31

Bleeding disorders in pregnancy require consideration of maternal bleeding and hemorrhagic complications in the newborn. Data on the fetus are often lacking, and the practitioner must base decisions on past experience and the mother’s previous reproductive history.

DISSEMINATED INTRAVASCULAR COAGULATION

Life-threatening bleeding is seen with some pregnancy-unique complications, resulting in disseminated intravascular coagulation (DIC). Because of the changes in coagulation factor levels, D-dimer, and platelet count during pregnancy, the normal range for tests routinely used to diagnose DIC in a nonpregnant state cannot be extrapolated directly to DIC in pregnancy. Serial measurement of the prothrombin time (PT), partial thromboplastin time (PTT), D-dimer, and fibrinogen are likely to be more helpful than measuring a single value.³² The DIC score developed by the International Society on Thrombosis and Hemostasis has been modified for pregnancy and this score may be more useful in identifying DIC.³³

Complications of pregnancy that lead to DIC include placental abruption, a retained dead fetus, and amniotic fluid embolism (Chap. 129). Although amniotic fluid embolism is a significant cause of maternal death in developed countries, the mortality decreased from 86 percent in 1979 to less than 30 percent in 1994 and 1995, perhaps from a better supportive therapy.³⁴ Amniotic fluid embolism is most likely to occur in older multiparous women whose pregnancies have gone beyond the 40th week and during tumultuous labor. Amniotic fluid enters the maternal circulation through tears in the chorioamniotic membranes, injury to the uterine veins, or uterine rupture. Its onset is heralded by maternal vascular collapse with dyspnea, hypotension, and cardiac arrhythmias followed by DIC that is manifested by oozing from intravenous lines, hematuria, hemoptysis, and excessive uterine bleeding. Atypical presentations have also been reported in which there is rapid deterioration of the fetus, followed by maternal respiratory and cardiovascular deterioration with development of DIC.³⁵

In amniotic fluid embolism, DIC appears to involve an abnormal host response to exposure to various foreign antigens with the subsequent release of endogenous mediators which drive the clinical manifestations.³⁶ Treatment is not significantly different than in other cases of DIC with bleeding (Chap. 129); however, there are some reports of successful management with uterine artery embolization.³⁷

Placental abruption has also led to development of DIC, and the spectrum of hemostatic failure is broad and appears to be related to the degree of placental separation.³⁸ Volume resuscitation, delivery of the fetus, and infusion of blood products to correct the maternal coagulation defect are indicated. Regional anesthesia is contraindicated because of the risk of bleeding in the epidural space and of the pooling of blood in the lower limb vascular bed, which could worsen hypovolemia.³⁸ Fetal trophoblast cells have distinct properties which may activate coagulation including expression of tissue factor, suppression of fibrinolysis, and exposure of anionic phospholipids.³⁹ Finally, intrauterine fetal death can also lead to DIC. Thromboplastic substances and specifically tissue factor released from dead fetal tissues into the maternal circulation are thought to trigger DIC; however, this is not usually detectable by laboratory tests until 3 or 4 weeks after fetal demise. Overt DIC is present in approximately 50 percent of women who retain a dead fetus for 5 weeks or longer.⁴⁰

VON WILLEBRAND DISEASE

Although von Willebrand disease (VWD) is transmitted in an autosomal dominant fashion, women appear to be disproportionately affected with bleeding symptoms, primarily menorrhagia and postpartum hemorrhage (Chap. 126). In normal women and in types 1 and 2 (but not type 3) VWD patients, levels of factor VIII and VWF rise during pregnancy, with the most pronounced increase in the third trimester.³⁹ As a result, prophylactic administration of VWF-containing factor concentrates at delivery is often unnecessary in type 1 and type 2 VWD patients; however, the risk of postpartum hemorrhage is significant (13–29 percent) as levels fall rapidly after birth.⁴¹ Thus in type 1 patients, factor VIII levels should be tested not only late in the third trimester, but also for 1 to 2 weeks postpartum. These patients should be monitored for increases in menstrual blood flow for at least 1 month. Risk of bleeding appears to be minimal when factor VIII levels are greater than 50 U/dL. There are several reports of severe thrombocytopenia developing late in pregnancy in patients with type 2B VWD,^42,43 and at least one of these patients developed a pulmonary embolus while receiving cryoprecipitate for postpartum hemorrhage. Despite the possible risk of thrombosis, these patients may require treatment with plasma-derived VWF-containing concentrates at delivery or postpartum if there is abnormal bleeding, and with platelets if thrombocytopenic bleeding is not controlled with infusion of VWF concentrate. Type 3 VWD patients require infusion of a plasma-derived VWF-containing concentrate at delivery, typically 40 to 80 IU/kg, followed by doses of 20 to 40 IU/kg daily for a week then tapered over the next few weeks.⁴⁴ Use of desmopressin acetate (DDAVP) antepartum is controversial because of the theoretical risk of vasoconstriction and placental insufficiency and the risk of maternal hyponatremia. Guidelines for management of VWD at delivery and during the puerperium have been published and are also reviewed in Chap. 126.^45,46

COAGULATION FACTOR DEFICIENCIES

Carriers of hemophilia A and B generally have factor levels approximately 50 percent of normal; however, a wide range of values have been reported as a result of random inactivation of the X chromosome (Chaps. 10 and 123).^47,48 Ideally, carriers are identified before pregnancy when prenatal counseling can be offered. Baseline factor levels should be tested at the first visit during pregnancy and again in the third trimester, but it should be noted that factor IX levels generally do not rise during the course of the pregnancy.⁴⁷ The sex of the fetus should be determined to guide the obstetrician at delivery. With the recognition that maternal serum contains cell free fetal DNA, genomic strategies have been developed to determine fetal gender as early as 7 weeks of gestation. Similarly, strategies to determine whether a male fetus is affected by hemophilia based on testing of maternal blood have now been developed and will doubtless enter the clinical arena in the near future.⁴⁹ Cranial hemorrhage is the commonest site of bleeding in newborns with severe hemophilia, and has the highest potential for long-term serious sequelae. Risk factors for cranial hemorrhage include prolonged labor and use of instruments during delivery.⁴⁸ To protect a potentially affected or known hemophiliac fetus, vacuum extraction should be avoided at delivery and forceps should be used only with caution. All intramuscular injections should be withheld from the newborn until hemophilia testing is completed. If an infant’s hemophilia status is not known, testing should be done on cord blood to avoid potential bleeding or bruising after a blood draw.⁴⁸ The mother’s factor level should be followed for a few days after delivery and menstrual bleeding should be monitored to ensure adequate hemostasis.

There is also an association between pregnancy and acquired hemophilia caused by factor VIII autoantibodies (Chap. 127). This condition usually appears 1 to 4 months postpartum, but emerges during pregnancy in up to 14 percent of patients.⁵⁰ In general, the Bethesda titer of the inhibitor is low and in most cases the inhibitor disappears spontaneously. Inhibitors can recur in subsequent pregnancies.⁵¹

Rarely, pregnant women with factor deficiencies other than factors VIII and IX may be identified. The most important of these to recognize is deficiency of factor XIII, which is associated with habitual hemorrhagic abortions and postpartum hemorrhage. In rare pregnancies reaching term, bleeding complications, including intracranial hemorrhage in the infant, have been observed.^52,53 Treatment of this deficiency with fresh-frozen plasma, cryoprecipitate, or plasma-derived factor XIII concentrates (now available in the United States) prevents abortion in women, although there are no controlled studies.⁵⁴ Most authorities recommend more frequent prophylactic therapy during pregnancy (every 3 weeks vs. every 5–6 weeks) with booster doses during labor or before cesarean section to ensure a level of 5 percent or greater.⁵⁵ Although rare, congenital afibrinogenemia, hypofibrinogenemia, and dysfibrinogenemia (Chap. 125) can cause hemorrhagic and thrombotic pregnancy complications. Most experts recommend fibrinogen replacement (using cryoprecipitate or fibrinogen concentrate) to maintain a level of 60 to 100 mg/dL during pregnancy and for 6 weeks postpartum.⁵⁶

CAUSES OF THROMBOCYTOPENIA

Thrombocytopenia in pregnancy is relatively common, with up to 5 percent of all pregnant women exhibiting asymptomatic thrombocytopenia.⁵⁷ Many causes of thrombocytopenia in pregnancy are identical to those seen in the nonpregnant state, with some predisposing to bleeding whereas others predispose to clotting. However, there are several conditions leading to thrombocytopenia that are unique to pregnancy, including gestational thrombocytopenia, preeclampsia/HELLP syndrome/eclampsia, and acute fatty liver of pregnancy.

Gestational and Immune Thrombocytopenia

Gestational thrombocytopenia and idiopathic thrombocytopenic purpura (ITP) are best discussed together as they can be difficult to differentiate and, in fact, may be two extremes of a spectrum of disease. In general, gestational thrombocytopenia is asymptomatic and is said to occur later in pregnancy and be less severe than ITP. Most sources suggest that gestational thrombocytopenia occurs in the second and third trimesters, with platelet counts rarely falling below 70,000/μL.⁵⁸ Gestational thrombocytopenia can sometimes be diagnosed with certainty only after delivery; usually there is no past history of low platelets, except perhaps with previous pregnancies, the platelet count returns to normal after delivery, and there is no association with fetal thrombocytopenia. It is not clear whether or not gestational thrombocytopenia is a variant of immune-mediated platelet destruction (Chap. 117).⁵⁸

In contrast to gestational thrombocytopenia, ITP can occur at any point in pregnancy and the fall in platelet count can be severe. Diagnosis is essentially the same as it would be in any patient in that alternative causes of thrombocytopenia must be ruled out. As in other cases, treatment of ITP in pregnancy must take into account the severity of the thrombocytopenia and the presence or absence of symptoms. In general, platelet counts less than 10,000/μL require treatment regardless of the trimester; platelet counts of 30,000 to 50,000/μL without bleeding require no treatment, and platelet counts of 10,000 to 30,000/μL in later trimesters or in the presence of bleeding require treatment. Although glucocorticoid and intravenous immunoglobulin are safe in pregnancy, they may have no effect on fetal counts and should only be used to treat the mother.⁵⁹ Splenectomy for ITP in pregnancy is best done in the second trimester if platelet counts are extremely low and unresponsive to treatment.⁵⁸ One small study evaluated the safety of anti-D antibodies during pregnancy; all 10 of the women studied achieved a platelet count greater than 30,000/μL, but larger studies are needed before this intervention can be recommended.⁶⁰ Similarly, there are case reports of rituximab administration for treatment of refractory ITP in pregnancy; at least one report demonstrated transient inhibition of neonatal B-lymphocyte development.⁶¹ There are no adequate and well-controlled studies of either eltrombopag or romiplostim in pregnant women and both are considered pregnancy category C drugs. In animal studies, both drugs crossed the placental and fetal effects included thrombocytosis, postimplantation loss, increase in fetal mortality, but no major structural malformations were reported.⁶² Case reports describing the use of romiplostim in pregnancy have appeared and in one report the newborn had severe thrombocytopenia at birth complicated by intracranial hemorrhage.⁶³ Maternal platelet counts of greater than 50,000/μL usually are safe for both vaginal and cesarean delivery. In most cases, spinal anesthesia should not be used if the platelet count is less than 75,000/μL.⁶⁴ Less than 5 percent of babies born to mothers with ITP have platelet counts less than 20,000/μL, although there does seem to be some correlation between very severely depressed maternal platelet count and severe thrombocytopenia in the newborn.⁶⁵ No clear recommendations can be given for measuring fetal platelet count prior to or at delivery as measurements are fraught with error; however, if the fetal platelet count is known to be less than 20,000/μL, cesarean section is probably reasonable. Newborns of mothers with ITP should be monitored for 5 to 7 days after delivery to ensure that the platelet count does not drop (Chap. 117).

Eclampsia and HELLP Syndrome

The spectrum of hypertensive disorders of pregnancy ranging from preeclampsia to severe preeclampsia and HELLP syndrome to eclampsia (see Chap. 51) may also result in thrombocytopenia, although thrombosis is more of an issue than is bleeding. There is some debate in the literature as to whether thrombocytopenia can be diagnosed in preeclampsia without HELLP syndrome; however, data from one large study⁵⁷ indicated that approximately 15 percent of cases of preeclampsia are complicated by thrombocytopenia. In general, the symptoms of preeclampsia, including hematologic manifestations, resolve with delivery; however, in a small proportion of cases they persist, worsen, or even develop immediately postpartum. When symptoms persist postpartum, the differentiation from thrombotic thrombocytopenic purpura (TTP)/hemolytic uremic syndrome becomes more difficult. Some data suggest that maternal recovery from the HELLP syndrome is accelerated by administration of intravenous dexamethasone⁶⁶; however, a meta-analysis demonstrated no clear advantage to the use of glucocorticoids to decrease maternal or perinatal morbidity or mortality.⁶⁷ A collaborative randomized controlled trial of glucocorticoids in HELLP syndrome (COHELLP) is underway to determine the effectiveness of dexamethasone to accelerate the postpartum recovery of patients with class I HELLP syndrome.⁶⁸ Observation or treatment of HELLP with glucocorticoids alone postpartum should probably not persist beyond the third postpartum day. If the patient is not clearly improving, plasma exchange should be initiated as one would do for thrombotic thrombocytopenic purpura (TTP).^69,70 Although not associated with hypertension, acute fatty liver of pregnancy is another rare disorder that can present in the third trimester with severe liver dysfunction, but thrombocytopenia, if present, is generally mild and does not require treatment (Chaps. 51, 117, and 129).

FETAL LOSS AND COMPLICATIONS

Pregnancy is a prothrombotic state. Inherited prothrombotic conditions contribute to 50 percent of the cases of venous thromboembolism and pulmonary embolism, as well as to stroke in pregnancy and the puerperium. Hereditary thrombophilias (Chap. 130) may also predispose to fetal loss through placental vascular disorders. The best evidence for an association between a thrombophilia, albeit acquired, and recurrent fetal loss exists for antiphospholipid antibody syndrome in which the association between the antibodies and pregnancy loss has been recognized for more than 20 years.⁷¹ As many as 20 percent of women with recurrent fetal loss have antiphospholipid antibodies,⁷² and studies show that without treatment up to 90 percent will experience fetal loss.⁷³ One study suggests that poor pregnancy outcomes occur more frequently in primary antiphospholipid syndrome when patients have more than one positive laboratory test (lupus anticoagulant, immunoglobulin [Ig] G/IgM anticardiolipin, IgG/IgM antihuman β₂-glycoprotein I antibodies).⁷⁴ In a randomized controlled trial including 90 women with a history of recurrent miscarriage associated with phospholipid antibodies (or antiphospholipid antibodies), lupus anticoagulant, and cardiolipin antibodies (or anticardiolipin antibodies), the rate of live births with low-dose aspirin (75 mg/day) and unfractionated heparin (5000 U subcutaneously twice per day) was 71 percent (32 of 45 pregnancies) and 42 percent (19 of 45 pregnancies) with low-dose aspirin alone (odds ratio: 3.37 [95% confidence interval: 1.40 to 8.10]).⁷⁵ A rarer acquired cause of fetal loss and thrombosis in pregnancy is paroxysmal nocturnal hemoglobinuria (PNH; Chap. 40). Though no clinical trials data exist, recommendations are to provide prophylactic or intermediate-dose low-molecular-weight heparin antepartum and for 6 weeks postpartum. Eculizumab should be consider both prior to and following delivery.⁷⁶

Although an association between inherited thrombophilias and pregnancy loss has been elusive and there are inconsistencies between studies, there is an association between factor V Leiden and recurrent fetal loss.^77,78 Less-convincing data exist for prothrombin 20210A, but there is no clear association with homozygous methylene tetrahydrofolate reductase C677T polymorphism (hyperhomocysteinemia).⁷⁹ Clinical trials examining the use of low-molecular-weight heparins or aspirin to prevent adverse pregnancy outcomes in women with inherited thrombophilias have been criticized for lack of an untreated control group.^80,81,82 Pregnancy outcomes from a large ongoing trial conducted with a no-treatment control have yet to be published.⁸³ Studies evaluating a role for inherited thrombophilias in preeclampsia and intrauterine growth retardation indicate that these factors may not be causative, but may contribute to disease severity.^80,81

THROMBOEMBOLIC EVENTS

Risk Factors

Estimates place the relative risk of arterial and venous thromboembolism (VTE) in pregnant women (Chaps. 133 and 134) at two to six times that of nonpregnant women.^11,82 Factors specific to pregnancy that increase the risk of VTE include obstruction of venous return by the gravid uterus, acquired prothrombotic changes in hemostatic proteins, and venous atonia caused by hormonal factors.⁸⁴ Additional risk factors include cesarean section (especially emergency), obesity, and increasing age. Approximately 80 percent of deep vein thromboses in pregnancy occur in the iliofemoral veins on the left, probably as a consequence of compression of the left iliac vein by the right iliac and ovarian arteries.^85,86 Rates of VTE immediately postpartum are difficult to assess as many occur after the patient is discharged; one large study estimates rates as much as 5 times higher in postpartum women than pregnant women.⁸⁷ Inherited thrombophilias (Chap. 130) are associated with more than half of VTE in pregnancy. Factor V Leiden and the prothrombin gene mutation are the commonest abnormalities associated with VTE in pregnancy, accounting for 44 and 17 percent, respectively. In general 1 in 500 factor V Leiden heterozygotes and 1 in 200 heterozygous carriers of the prothrombin 20210 gene mutation will have a VTE in pregnancy. Homozygosity for either of these mutations increases risk about fourfold. Based on recent studies, the risk for VTE during pregnancy is 1 in 113 for protein C deficiency, approximately 1 in 3 for type I antithrombin deficiency, and 1 in 42 for type II antithrombin deficiency. Risk for carriers of protein S deficiency is similar to that for protein C deficiency.^88,89,90

Diagnostic Methods

Diagnosis of VTE in pregnancy is complicated both because the presenting complaints—leg edema, back pain, and chest pain—are common in pregnancy, and because radiologic studies used to make the diagnosis in nonpregnant individuals are relatively contraindicated in pregnant women. Compression ultrasonography is the initial test of choice in pregnant women. If this test is nondiagnostic, several other tests may be considered. If pulmonary embolus is suspected, lung ventilation perfusion scanning, which gives relatively low-dose radiation, may be used. Magnetic resonance imaging or magnetic resonance venography are also informative if available. Measurement of D-dimers is a useful adjunct in nonpregnant patients to rule out VTE (D-dimers are sensitive, but not specific, for VTE). However, D-dimer levels rise over the course of normal pregnancy^91,92 and with several complications of pregnancy, including preterm labor, hypertension, and placental abruption,⁹³ and thus may not be useful in excluding VTE.

Prophylaxis

Prophylaxis for VTE is a controversial issue as only a few prospective studies have been done to assess the risk of use.^94,95 There is general agreement, however, that because of its teratogenic potential, warfarin should not be used during pregnancy and that low-molecular-weight heparins are the anticoagulant of choice because they do not cross the placenta and have a lower risk of osteoporosis and heparin-induced thrombocytopenia than unfractionated heparin.⁹⁶ Most experts now agree that risk assessment should be based on both personal and family history of VTE as well as the presence of a known hypercoagulable state. Based on the most recent Chest Guidelines,⁹⁷ all women with prior history of VTE should be offered postpartum prophylaxis with prophylactic or intermediate dose low molecular weight heparin for 6 weeks. For pregnant women with a low risk of recurrence (e.g., a single VTE with a transient risk factor unrelated to pregnancy or estrogen use) surveillance is recommended antepartum, whereas those with higher risk should receive prophylactic or intermediate dose low molecular weight heparin prior to delivery. Women with no prior VTE are divided into four categories. Those with higher risk of VTE in pregnancy including Factor V Leiden or Prothrombin 20210 homozygotes or compound heterozygotes with a family history of VTE should receive prophylactic or intermediate dose low molecular weight heparin antepartum and postpartum prophylaxis with prophylactic or intermediate dose low molecular weight heparin for 6 weeks. Factor V Leiden or Prothrombin 20210 homozygotes or compound heterozygotes without a family history of VTE should undergo surveillance antepartum and should receive postpartum prophylaxis with prophylactic or intermediate dose low molecular weight heparin for 6 weeks. Pregnant women with no personal history of VTE with any other thrombophilia and a family history of VTE should undergo surveillance antepartum and should be offered postpartum prophylaxis with prophylactic or intermediate-dose low-molecular-weight heparin for 6 weeks. Finally, women with lower risk thrombophilias and no personal or family history of VTE should be offered surveillance post antepartum and postpartum. Patients with two or more episodes of VTE should probably be treated throughout pregnancy and the puerperium.^{98,99,100,101,102} As noted above, women who meet the criteria for antiphospholipid antibody syndrome should receive antepartum prophylactic or intermediate dose unfractionated or prophylactic low molecular weight heparin and low dose aspirin throughout pregnancy. Treatment of VTE in pregnancy should be with full-dose low-molecular-weight heparin. Ideally, women on treatment doses of heparin have elective induction of labor. Heparin is usually discontinued 24 hours prior to induction; however, women deemed to be at very high risk of recurrent VTE can then receive intravenous heparin up to 4 to 6 hours prior to delivery.^100,101 Great care should be taken with epidural anesthesia, and it should be avoided if there is any question of a significant anticoagulant effect. Heparins and warfarin are safe postpartum, even when breastfeeding.¹⁰³

Although not common, leukemias and lymphomas do occur in pregnancy and present problems with proper diagnosis, staging, and treatment (Chaps. 88–90). The literature suggests that the incidence of Hodgkin lymphoma is 1:1000 to 1:6000 pregnancies, whereas the incidence of non-Hodgkin lymphoma is manyfold lower.¹⁰⁴ Leukemia in pregnancy is uncommon.

HODGKIN LYMPHOMA

Neither the histology nor the outcome of patients who present during pregnancy is worse than that of other patients.¹⁰⁴ Diagnosis, usually by biopsy of a lymph node, is usually not problematic, but staging can be difficult. Posterior–anterior chest films with abdominal shielding and marrow biopsy (in the presence of B symptoms, leukopenia, or thrombocytopenia) should be done and present little risk to the fetus. Laboratory studies, including blood counts, liver functions tests, and ESR, should be done, but care should be taken in interpreting the alkaline phosphatase and ESR measurements, which both rise during the course of a normal pregnancy. Evaluation for the presence of abdominopelvic disease is difficult because computed tomography imaging is contraindicated in pregnant women. Abdominal ultrasonograms are safe, but provide limited information. If necessary, magnetic resonance imaging scans can probably be done safely in pregnancy; however, this is rarely necessary and most experts recommend waiting until after the first trimester. The toxicities of treatment and the risks of delaying treatment until later in pregnancy or postpartum need to be considered carefully in each case. Fetal risks of chemotherapy are greatest in the first trimester during the period of organogenesis, with folate antagonists and antimetabolites carrying the largest risk.¹⁰⁵ Despite the changes in physiology that occur during pregnancy, there is no evidence that dosing should be changed. If chemotherapy is indicated, it should be delayed until the second trimester; however, single-agent vinblastine has been given in the first trimester with a low incidence of fetal abnormalities.¹⁰⁶ Treatment should be timed so that there is the maximum amount of time possible between the last dose of chemotherapy and delivery to avoid cytopenias in either the mother or the fetus. In some cases, radiotherapy may be a feasible alternative in the second and third trimesters of pregnancy. Of 16 patients who received radiotherapy for supradiaphragmatic Hodgkin lymphoma (clinical stages IA and IIA) during pregnancy, 11 received full mantle irradiation, and all patients had lead shielding of the uterus.¹⁰⁷ All 16 pregnancies were carried to completion with full-term deliveries of normal infants. However, a review of the records of 382 women treated with radiotherapy for Hodgkin lymphoma suggests that the risk of breast cancer after radiation therapy is nearly sevenfold greater with irradiation around the time of pregnancy.¹⁰⁸ Additional studies are needed to confirm these findings, but this potential risk should be borne in mind by the clinician when making therapeutic decisions.¹⁰⁹ Relapse of Hodgkin lymphoma usually occurs within the first 2 years following treatment, and patients are counseled to avoid pregnancy during this period. Vigilance for second cancers in these patients is also advised as is monitoring for hypothyroidism in those who receive radiation therapy, especially during subsequent pregnancies when hypothyroidism could have profound maternal and fetal effects.

NON-HODGKIN LYMPHOMA

As compared with Hodgkin lymphoma, other lymphomas (Chap. 95) are less frequent in pregnancy, tend to present with a higher stage disease, and have a poorer prognosis.¹¹⁰ Burkitt or Burkitt-like lymphoma can involve the breasts of young pregnant or lactating women and typically behaves aggressively.^111,112 A recent review of more than 100 cases of Non-Hodgkin lymphoma in pregnancy, 75 percent of the patients had stage IV disease at diagnosis and nearly half had involvement of reproductive organs, primarily the breast. Very few cases of placental or fetal involvement were observed.¹¹³ In patients with high-grade lymphomas, chemotherapy often cannot be delayed and difficult decisions must be made. However, in one report of 16 pregnant patients who received aggressive chemotherapy for non-Hodgkin lymphoma during their pregnancies, all survived to delivery.¹¹⁴ Half of the 16 patients received chemotherapy in their first trimester and all 16 delivered healthy infants despite episodes of myelosuppression during the pregnancies. In a subsequent report, the health of 84 children born to mothers who received chemotherapy for hematologic malignancies during pregnancy revealed no abnormalities in physical or cognitive development and no increase in cancers at a median followup of 18.7 years.¹¹⁵ Rituximab in pregnancy, both as a single agent and in combination with chemotherapy, has not been associated with abnormalities of the newborn when given in the first, second, or third trimester¹¹⁶; however, there is one report of prolonged lymphopenia in a neonate whose mother received rituximab in pregnancy.^61,116,117 A retrospective study of 231 pregnancies associated with maternal rituximab exposure identified 153 pregnancies with known outcomes. Among these pregnancies there were 90 live births, 22 of which were premature. There were four neonatal infections and two congenital malformations.¹¹⁸

ACUTE LEUKEMIA

Leukemia is distinctly uncommon in pregnancy; estimates derived from studies beginning in the 1950s place the incidence at approximately 1:75,000 pregnancies (Chaps. 88 and 91).^119,120 Acute leukemias make up nearly 90 percent of the total, followed by chronic myeloid leukemia, which comprises an additional 10 percent; chronic lymphocytic leukemia is extremely rare.¹²¹ The acute leukemias require urgent treatment, and while pregnancy itself does not alter the course of the leukemia, the outcome is much worse if treatment is delayed.¹²² A summary of data on 96 pregnant women reported in the literature from 1983 to 1995 who were treated with cytotoxic chemotherapy for leukemias (most of which were acute) revealed that most patients received regimens that included multiple drugs and were not different from those given to nonpregnant patients.¹²³ Nearly one-third of patients were treated in the first trimester of pregnancy. Among the 96 pregnancies, there were 2 maternal deaths, 2 children were stillborn, 2 therapeutic abortions were performed, 1 child had chromosomal abnormalities, and 8 had congenital defects. Seven of the eight children born with congenital defects were born to mothers who had been treated in the first trimester. It was not possible to identify a drug (or drugs) that was most likely responsible for adverse outcomes. Treatment in the first trimester carries a high risk of fetal anomaly or miscarriage. For patients with acute myelogenous leukemia (AML) receiving a standard induction regimen of cytarabine and an anthracycline, it is probably best to avoid idarubicin, which is more lipophilic with elevated placental transfer. Doxorubicin has been extensively studied in pregnancy women with breast cancer and is probably the anthracycline of choice in pregnant patients with AML.¹²⁴ Although few cases of fetal cardiotoxicity related to anthracyclines have been reported, fetal cardiac function should be monitored in pregnancy. Very few data exist for consolidation therapy in pregnancy. Case reports of treatment of acute promyelocytic leukemia in pregnancy with all-trans-retinoic acid^125,126,127 suggest that it may be safe after the first trimester. Arsenic trioxide is not recommended for use at any stage of pregnancy because of its high potential for embryotoxicity.¹²⁸ For patients who require chemotherapy postpartum, breastfeeding is not recommended so as to avoid exposure of the newborn to cytotoxic drugs in the breast milk.¹²⁹ Patients with chronic myeloid leukemia have been successfully treated in pregnancy with interferon-α, hydroxyurea, leukapheresis, and even busulfan.^130,131,132 A review of 125 women treated with imatinib mesylate during pregnancy revealed that most pregnancies had successful outcomes; however, 12 infants had abnormalities, 3 of which involved complex malformations.¹³³

Supportive treatment with antiemetics, including ondansetron, aprepitant and metoclopramide, is safe, and these agents do not cause congenital malformations.¹³⁴ Little data exist for the effects of growth factors including granulocyte colony-stimulating factors, but there are no reports of teratogenic effects.¹³⁵ Care should be taken with prescribing antibiotics and quinolones in specific should be avoided in pregnancy as should sulfonamides. When antifungal therapy is required, amphotericin may be the drug of choice as there have been no reports of teratogenicity with this agent. Fluconazole appears to be safe at doses less than 150 mg per day, but ketoconazole and voriconazole can cause fetal malformations and should avoided altogether.¹³⁶

The management of pregnant patients with essential thrombocythemia (ET) is a challenge because thrombosis is the main complication of ET (Chap. 85) and is accentuated by the prothrombotic state of pregnancy. In addition, of all the myeloproliferative neoplasms, ET has the highest proportion of affected females of child-bearing age. One study reviewed 155 pregnancies in 86 women with ET, and only 59 percent of these pregnancies resulted in a live neonate.¹³⁷ First-trimester abortion was seen in 31 percent of pregnancies, the main cause being placental infarction. Maternal thrombotic or hemorrhagic complications were infrequent, but were more common than in normal pregnancy. Pregnancy did not appear to adversely affect the course and prognosis of ET.

A meta-analysis revealed a benefit for aspirin treatment, whereas the benefit of heparin prophylaxis has not been established, but may have a role in selected cases.¹³⁸ If cytoreductive therapy becomes necessary, interferon-α is the drug of choice. A similar incidence of pregnancy complications in patients with ET was reported in a series from the Mayo clinic.¹³⁹ Another large single institution study of 68 young ET patients demonstrated that for both polycythemia vera (PV) and ET, most thromboses in young patients occurred at the time of diagnosis and also suggested, but did not prove, the benefit of aspirin.¹⁴⁰ The most detailed analysis was published by the Italian Society of Hematology in its guidelines.¹³⁸ The Society’s report analyzed pooled outcome data from 461 pregnancies in women with ET. The mean age of the pregnant patients was 29 years, and the mean platelet count at the beginning of pregnancy was 1000 × 10⁹/L, which declined to 400 × 10⁹/L in the second trimester. This decrease in the platelet count during pregnancy documented for the first time the anecdotal observation that some women with ET spontaneously normalize their platelet count during their pregnancy. (The authors of this chapter have rarely observed this phenomenon; however, in one of their ET patients a spontaneous, but transient, ET remission occurred in the first pregnancy, but not in the following pregnancy.) The Italian study found that 44 percent of pregnancies were unsuccessful in women with ET, a figure that is threefold higher than in the general population. Among the 461 pregnancies there were 13 pre- or postpartum significant bleeding events. The median duration of gestation was 38 weeks because of abortions and preterm deliveries. Cesarean section was necessary in 15 percent of the patients. The platelet count at the beginning of pregnancy did not predict pregnancy outcome. Placental infarctions were reported in 18 pregnancies and these were associated with intrauterine fetal growth retardation (11 pregnancies). Placental abruption was reported in 3.6 percent of ET pregnancies compared to 1 percent in the non-ET population. Preeclampsia was seen at a rate equal to that seen in non-ET pregnancies. Postpartum thrombotic episodes were reported in 5.2 percent of the pregnancies and included venous thrombosis, pulmonary embolism, sagittal sinus thrombosis, transient ischemic attacks, and Budd-Chiari syndrome (rates for all problems were significantly higher than in non-ET pregnancies). The impact of therapy was difficult to evaluate because management of ET pregnancies was heterogeneous; no specific therapy for ET was given in 48 percent of the pregnancies. Aspirin therapy at doses ranging from 75 to 500 mg per day was used in 106 pregnancies, low-molecular-weight heparin (pre-/postpartum) was used in 26 pregnancies, interferon-α was used in 19 pregnancies, and a handful of patients had various chemotherapies and radioactive phosphorus. When the outcome of the ET pregnancies was reviewed, 74 percent of patients treated with aspirin during pregnancy had successful pregnancies, whereas 55 percent of the patients not receiving aspirin had successful pregnancies. Based on the detailed analyses of all variables, this panel of experts felt that there was no direct evidence of the efficacy of aspirin in pregnant ET women, but that “it seems possible that aspirin increases the rate of successful pregnancies.” The panel also recommended that ET patients with a thrombotic episode (peripheral or placental) during pregnancy should receive low-molecular-weight heparin at therapeutic doses and oral anticoagulant therapy (PT international normalized ratio 2 to 3) for at least 6 weeks postpartum. Longer periods of anticoagulation were recommended for patients with familial thrombophilia. Pregnant women deemed candidates for platelet-lowering therapy (a history of major thrombosis, or of major bleeding, platelet count greater than 1000 × 10⁹/L, familial thrombophilia or cardiovascular risk factors) were recommended to receive interferon. The Italian panel also recommended avoidance of anagrelide in pregnancy because of uncertainty about its teratogenic potential; however, several normal infants have been born to women who inadvertently took this drug during pregnancy (FDA documents submitted by the manufacturer). Although the risk of congenital anomalies among infants of women treated with hydroxyurea during pregnancy was thought to be substantial, of 15 infants born to women treated with hydroxyurea at conception and/or during pregnancy, no malformations were observed, and only one stillbirth was reported in a woman who also had eclampsia. At least one publication has identified the presence of the JAK2 V617F mutation as a risk factor for pregnancy complications; however, to date there is no consensus on whether to manage these patients differently.¹⁴¹ A review of 158 young women with ET experiencing 237 pregnancies demonstrated that pregnancy complications are associated with higher risk of subsequent thrombosis.¹⁴²

Although there is significant overlap in the clinical features of PV and ET, there are some noteworthy differences (Chap. 84). In PV, the number of reported pregnancies is low because most PV patients are past childbearing age, and comorbid conditions are more frequent. One authoritative review suggests maintaining the hematocrit below 45 percent¹⁴¹ in pregnancy, and another recommends using interferon-α when myelosuppression is indicated.¹⁴³ Another noted authority in PV recommends that the hematocrit be kept lower than 35 percent in pregnancy.¹⁴⁴ However, because of a dearth of data and controlled studies, optimal management of PV pregnancies is poorly defined and agreed upon protocols are not available. None of the available information allows definite therapeutic recommendations; however, some authorities recommend that, at a minimum, all pregnant patients with PV be treated with low-dose aspirin.¹⁴⁴

SICKLE SYNDROMES

Although pregnancy in patients with sickle cell trait is typically uneventful, these patients probably have an increased risk for urinary tract infection.¹⁴⁵ Earlier studies suggested an increased risk for preeclampsia in patients with sickle cell trait, but a large study demonstrated that sickle cell trait is not an independent risk factor for preeclampsia (Chap. 49).¹⁴⁶

Patients with sickle cell anemia should receive at least 1 mg of folate per day; however, they should not receive iron supplementation until a ferritin level is checked and iron deficiency is documented.¹⁴⁷ Because of the risk of fetal malformation, hydroxyurea should be discontinued at least 3 months before pregnancy. However, successful outcomes have been reported in sickle cell disease patients who were exposed to the drug while pregnant.¹⁴⁸ Women with sickle cell anemia and their fetuses have an increased risk of complications during pregnancy. A large study of more than 17,000 deliveries to women with sickle cell anemia compared with controls demonstrated that infectious complications, including pneumonia, systemic inflammatory response syndrome, and sepsis, were more common in the women with sickle cell anemia. Furthermore, thrombotic complications such as cerebral vein and deep vein thrombosis and pregnancy complications including preeclampsia, eclampsia, abruption and antepartum bleeding were significantly more common in patients with sickle cell anemia. As in previous reports, rates of intrauterine growth retardation and preterm labor, as well as rates of maternal mortality, were higher in mothers with sickle cell anemia.¹⁴⁹ The issue of prophylactic versus need-based transfusion in sickle cell patients is controversial. A Cochrane Review identified only two small randomized studies conducted in the United States in the 1980s addressing this issue. These studies demonstrated no difference in perinatal outcome between the offspring of mothers with sickle cell disease who were assigned to treatment with prophylactic transfusions and those who were not.^1,150

Although the incidence of cesarean section in sickle cell patients is reported to be as high as 36 percent,¹⁵¹ delivery can generally be accomplished vaginally. Most experts recommend avoiding induction of labor as this can lead to sickle crisis.¹⁵² Epidural anesthesia is reported to be safe and to decrease the risk of peripartum painful crises.¹⁵³

THALASSEMIA SYNDROMES

β-Thalassemia Syndromes

Preconception evaluation of patients with β-thalassemia syndromes is recommended and should include assessment of transfusion needs, chelation therapy, body iron status and organ function, and the presence of antibodies to red cell antigens.¹⁵⁴ Patients with β-thalassemia minor generally tolerate pregnancy well; however, doses of at least 4 mg of folate per day are recommended in the preconception period and the first trimester as there is some data to suggest an increased risk of neural tube defects in their offspring.¹⁵⁵ Transfusion and iron chelation therapy has improved both life expectancy and fertility in patients with β-thalassemia intermedia and major, and successful pregnancies have been reported in both disorders.¹⁵⁶ A high rate of maternal mortality is reported in women with thalassemia and cardiac iron overload emphasizing the need for aggressive management of iron status prior to undertaking a pregnancy.¹⁵⁷ During pregnancy, regular transfusions are recommended to keep the hemoglobin level at 10 mg/dL and transfusion requirements often increase as compared to prepregnancy values.¹⁵⁸ Iron-chelation therapy with deferoxamine in pregnancy is controversial and most authorities recommend a hiatus during pregnancy; however, no fetal abnormalities have been reported in pregnancies in which it was continued (Chap. 48).¹⁵⁹

α-Thalassemia Syndromes

Patients with the silent carrier state or α-thalassemia trait have no increase in pregnancy complications; however, identification of patients with heterozygous α-thalassemia trait is important in assessing the risk of having a fetus that has hemoglobin H or hemoglobin Bart. Although women with hemoglobin H are generally able to have successful pregnancies, the chronic anemia often worsens, requiring blood transfusion. Patients with hemoglobin H are sensitive to oxidizing compounds and medications, which should be borne in mind, particularly during pregnancy (Chap. 48).