How to Treat Thrombocytopenia in Pregnancy (BLOOD)

How I Treat Series HEMATOLOGIC COMPLICATIONS IN PREGNANCY How I treat thrombocytopenia in pregnancy Annemarie E. Fogerty and David J. Kuter Hematology Division, Massachusetts General Hospital, Boston, MA Thrombocytopenia is a common hematologic abnormality in pregnancy, encountered in ~10% of pr...

How I Treat Series HEMATOLOGIC COMPLICATIONS IN PREGNANCY How I treat thrombocytopenia in pregnancy Annemarie E. Fogerty and David J. Kuter Hematology Division, Massachusetts General Hospital, Boston, MA Thrombocytopenia is a common hematologic abnormality in pregnancy, encountered in ~10% of pregnancies. There are many possible causes, ranging from benign conditions that do not require intervention to life-threatening dis- orders necessitating urgent recognition and treatment. Although thrombocytopenia may be an inherited condition or predate pregnancy, most commonly it is a new diagnosis. Identifying the responsible mechanism and predicting its course is made challenging by the tremendous overlap of clinical features and laboratory data between normal pregnancy and the many potential causes of thrombocytopenia. Multidisciplinary collaboration between hematology, obstetrics, and anesthesia and shared decision-making with the involved patient is encouraged to enhance diagnostic clarity and develop an optimized treatment regimen, with careful consideration of management of labor and delivery and the potential fetal impact of maternal thrombocytopenia and any proposed therapeutic intervention. In this review, we outline a diagnostic approach to pregnant patients with thrombocytopenia, highlighting the subtle dif- ferences in presentation, physical examination, clinical course, and laboratory abnormalities that can be applied to focus the differential. Four clinical scenarios are presented to highlight the pathophysiology and treatment of the most common causes of thrombocytopenia in pregnancy: gestational thrombocytopenia, preeclampsia, and immune thrombocytopenia. Introduction General pathophysiologic aspects of Thrombocytopenia, defined as a platelet count 9 million pregnancies worldwide each year.7,8 both mother and fetus and require timely intervention. Accu- Despite its frequency, the precise mechanism of thrombocyto- rately identifying the cause is the critical element of manage- penia is not known. As such, GT remains a diagnosis of exclu- ment but can be challenging given tremendous overlap in the sion that can only be confirmed after a postpartum complete symptoms and laboratory data between normal pregnancy blood count (CBC) reveals recovery to a normal platelet count. physiology and pathologic thrombocytopenias, particularly at The salient features of GT are listed in Table 2. onset. Confidence in the diagnosis often comes only with time or depending on response to initial therapies, because few Numerous hypotheses for the mechanism of GT have been processes are definitively identified with a singular confirmatory proposed, including hemodilution, increased von Willebrand test. factor (VWF), insufficient thrombopoietin (TPO) response, or reduced ADAMTS13 activity levels, but none provide complete Figure 1 displays the relative proportion of causes of throm- explanation. Although there is a physiologic increase in VWF bocytopenia in pregnancy, with Figure 2 depicting the most production in all pregnancies,12-14 this quantitative change likely cause based on trimester of onset and severity, which are alone is not likely responsible for GT, which occurs in only 10% the most important features to direct an appropriate differen- of pregnancies. ADAMTS13 levels decrease progressively in tial. Table 1 highlights the salient features of presentation and pregnancy, starting at about week 12 to 16 through the early management of commonly encountered or sinister thrombo- postnatal period,15 but without correlation to platelet count or cytopenic processes in pregnancy. Among the many potential identified association with GT. Women with GT had a higher causes, this review focuses on the 3 most common: gestational TPO level (187 pg/mL) than that of pregnant women with thrombocytopenia (GT), preeclampsia, and immune thrombo- normal platelet counts (65 pg/mL), nonpregnant women with cytopenia (ITP). ITP (88 pg/mL), and nonpregnant controls (37 pg/mL), but this 29 FEBRUARY 2024 | VOLUME 143, NUMBER 9 747  driven. Prior publications have demonstrated that some genetic 3% 1% VWF polymorphisms can result in enhanced platelet adhesion only under conditions of high shear flow.14,20,21 20% Although observation alone is appropriate for women with GT, GT its management is a frequently encountered clinical predica- Preeclampsia ment, particularly when the GT diagnosis is uncertain. There are ITP no randomized data defining a safe platelet threshold for 76% Other epidural anesthesia, which can lead to variability between institutions in accepted platelet targets and may drive unnec- essary “treatment” or denial of epidural anesthesia when an arbitrary target is strictly enforced. Several retrospective studies have shown no complications associated with regional anes- thesia to parturients with platelets 70 × 109/L in women with from ITP but do not explain the mechanism of GT. GT, ITP, or hypertensive disorders of pregnancy in the absence of additional risk factors.27 We encourage regular dialogue An investigation into potential mechanisms of GT suggests a between multidisciplinary providers to develop guidelines that unique maternal physiology. In a case control study of >3500 incorporate these data and revisit any protocols that may have pregnancies17 in which 12% of cases were diagnosed with GT, been established by custom or uncertainty. platelet counts in women with GT declined 31.8%, which was significantly greater than that of controls with an 18.3% decrease. There was also no difference in maternal weight, ITP weight gain, or hemoglobin (Hb) decline between women with ITP is defined as a platelet count 100 × 109/L purpura/hemolytic uremic syndrome. The figure was modified from Cines and Levine.6 All platelet counts GT PEC GT ITP Platelets 50 × 109/L, removal is indicated. recognizing the increased hemostatic demand of labor and delivery.40,41 The great success of recombinant human TPO (rhTPO) and the TPO receptor agonists (TPO-RA) in ITP51-53 may carry over to Treatment of the pregnant patients with ITP differs from that of the pregnant patient, but data remain scant. Only 1, rhTPO, the nonpregnant patient only in the exclusion of medications available in China, does not cross the placenta. In the only regarded as toxic to the fetus or insufficiently studied for prospective study of TPO agents in pregnant patients,54 31 safety. The general treatment approach is either to decrease women with platelets 30 × 109/ pregnancy is corticosteroids and IV immunoglobulin (IVIG). In a L, and 10 developed platelet counts >100 × 109/L. There were retrospective study of 137 pregnancies, there was no signifi- no bleeding events, and adverse events were minimal. There cant difference in maternal platelet count response with were no observed effects on the neonates. treatment of IVIG or corticosteroids (39% vs 38%).42 Although there was suggestion of reduced efficacy of corticosteroids and All TPO-RAs (romiplostim, eltrombopag, and avatrombopag) IVIG in pregnant patients with ITP compared with nonpregnant cross the placenta. Despite evidence of fetal effects in preclinical patients with ITP,43 this has not been confirmed by others44,45 animal toxicology studies only at levels much higher than those and is not our experience. The choice of corticosteroids is clinically used,55 these agents are not approved for pregnant limited to prednisone, which appears only in small amounts in patients, and we do not advocate their routine use. A retrospec- cord blood compared with dexamethasone, which reaches tive analysis of the use of TPO-RAs in the third trimester during 17 higher fetal concentrations due to less efficient placental pregnancies (8 women managed with eltrombopag and 7 women metabolism.46 The onset of prednisone effect is slightly slower with romiplostim) showed that baseline platelet counts of 10 × than with dexamethasone but is manifest in most patients 109/L rose to a median of 94 × 109/L with no bleeding at delivery. within 2 to 3 days.47 No serious adverse events or clinically symptomatic thromboem- bolic events were reported, but 2 patients reported mild head- Prednisone can usually be deferred until about week 35 or 36 in ache.56 Median TPO-RA exposure was limited to 4.4 weeks. Six of preparation for delivery to minimize exposure and associated the neonates (43%) with available platelets counts had thrombo- maternal side effects, which can include aggravated gestational cytopenia, and 1 had a platelet count of 558 × 109/L. There were diabetes, hypertension, mood lability, or insomnia.48 Once no fetal losses or birth defects attributable to TPO-RA; 29% were started, prednisone is continued until spontaneous delivery, delivered preterm (34-38 weeks). thereafter tapered if the platelet count is >20 × 109/L. Some consider a brief trial of prednisone 40 mg daily for 5 days at Our approach has been to restrict the use of TPO-RAs only for about week 30 to 32 to assess response, which may guide patients who have an inadequate response to corticosteroids management for labor and delivery, because steroid response and IVIG, for whom there is no time for other options (rituximab, is often consistent in an individual patient. This practice is cyclosporine, and azathioprine) to act, and only in the last usually reserved for more significant thrombocytopenia, trimester. If used, we prefer romiplostim, given its potentially because failed response to prednisone started at week 35 or 36 fewer off target effects, for example, no iron chelation as is seen still provides time to add IVIG in most patients. with eltrombopag. 750 29 FEBRUARY 2024 | VOLUME 143, NUMBER 9 FOGERTY and KUTER Preeclampsia which always requires active management. Onset and severity Preeclampsia is diagnosed in about 3% to 4% of all pregnancies of thrombocytopenia (Figure 2) further focus the differential, and is defined as maternal hypertension (blood pressure, guiding the appropriate supporting laboratory data and phys- 140/90 mm Hg) with proteinuria and/or end-organ dysfunction ical examinations. after 20 weeks.57 It may also occur in the early postpartum window. Preeclampsia is a thrombotic microangiopathy (TMA) A thorough history is important, particularly relevant family and thus characterized by schistocytes present on the periph- history, remote CBCs for baseline platelet count, and available eral blood film. The pathogenesis includes both abnormal data from prior pregnancies. The presence of other prior or placentation and angiogenic imbalance,58 with excess secretion concurrent autoimmune conditions, fevers, infectious symp- of the placentally derived antiangiogenic factor, soluble fms- toms, new drug, or toxin exposures should be reviewed. like tyrosine kinase-1 (sFlt-1). sFlt-1 sequesters proangiogenic molecules vascular endothelial growth factor and placental growth factor. Increased sFlt-1/placental growth factor ratios Cases are seen in preeclampsia, both at diagnosis and weeks before Patient 1 the onset of clinical symptoms,58-60 suggesting causality. A 27-year-old female (G3P2002) at 34 weeks is referred for a Aspirin has consistently been shown to prevent or delay onset platelet count of 108 × 109/L. She feels well. She is unsure of of preeclampsia.61,62 A Cochrane analysis of >36 000 women platelet nadirs in her prior pregnancies but recalls mention of showed an 18% reduction. thrombocytopenia. She received neuraxial anesthesia and delivered vaginally without complications. On physical exami- A platelet count