Thrombocytopenia in pregnancy

Thrombocytopenia occurs in up to 10% of pregnancies, with etiologies ranging from the benign to life-threatening.¹  As obstetricians are experts in managing the hypertensive disorders of pregnancy (eg, preeclampsia [PEC]), a hematology consult often signals the patient has atypical features for these disorders. Hematologists are also consulted for treatment recommendations in patients with preexisting chronic thrombocytopenia. This review presents 2 cases highlighting an approach to managing these challenging consults.

A 32-year-old primigravida woman at 34 weeks gestation presents with abdominal pain. Her blood pressure is elevated at 150/90  mm Hg. Her platelet count has dropped to 80 × 10⁹/L from 150 × 10⁹/L 2 months prior. White blood cell count is 4.0 × 10⁹/L and hemoglobin is 9  g/dL.

An initial laboratory workup includes review of a peripheral blood smear (evaluating for pseudothrombocytopenia and any morphologic abnormalities) and an assessment of renal and hepatic function. Hematologists should consider both “pregnancy-specific” and “general” causes of thrombocytopenia (Table 1).²  Similar to nonpregnant patients, current medications and comorbidities should be reviewed as potential culprits. It is essential to inquire about the presence of any neurologic, infectious, or B-symptoms. A history of thrombocytopenia during prior pregnancies or family history of thrombocytopenia can be revealing.³  On physical examination, special attention must be paid to any elevation in blood pressure, bruising, hepatosplenomegaly, or lymphadenopathy.³  Additional complete blood count abnormalities significantly change the differential, with possible exceptions for thrombocytopenia with concomitant mild anemia due to increased plasma volume in the later trimesters (physiologic anemia) or a microcytic anemia (iron deficiency). These generally mild anemias are common in pregnancy and may be unrelated to the underlying etiology of the thrombocytopenia.⁴  Pancytopenia warrants a bone marrow biopsy if an offending medication and/or vitamin deficiency cannot be identified.⁵ 

Knowledge of the expected platelet count trend during pregnancy is important to distinguish benign from life-threatening etiologies. A study by Reese et al¹  of ~7000 pregnant vs nonpregnant individuals showed platelet counts were lower during pregnancy, beginning in the first trimester, and lowest at time of delivery (Figure 1). This benign phenomenon of steadily decreasing platelet counts throughout pregnancy with spontaneous recovery in the nonpregnant state is known as gestational thrombocytopenia (GT). It is believed to be partly due to hemodilution secondary to increased plasma volume and splenic sequestration.⁶  However, the degree of thrombocytopenia is important to consider before invoking GT as the cause. Platelet counts at delivery <100 × 10⁹/L occur in 1% and <80 × 10⁹/L in only 0.1% of women with uncomplicated pregnancies.¹  Given the rarity of this degree of thrombocytopenia in uncomplicated pregnancies, platelets <100 × 10⁹/L warrant careful consideration of etiologies other than GT.¹ 

The trimester of onset of the thrombocytopenia is also helpful in determining the etiology (Figure 2).³  In the first trimester, only 1.8% of patients with uncomplicated pregnancies have a platelet count <150 × 10⁹  L. This increases to 4.8% and 8.5% in the second and third trimesters, respectively.¹  Thus, GT most commonly becomes apparent in the later trimesters.^1,3  On the differential for isolated thrombocytopenia in the first trimester is hereditary thrombocytopenia, immune thrombocytopenia (ITP), and, more rarely, type IIB von Willebrand disease (a diagnosis to consider in a patient with a personal and/or family history of bleeding as the associated coagulopathy may complicate delivery).⁷  The hypertensive disorders of pregnancy (eg, PEC and hemolysis, elevated liver enzymes, and low platelets [HELLP]) present after 20 weeks (mid to late second trimester), with most occurring after 34 weeks (third trimester).⁸ 

A peripheral blood smear reveals no platelet clumping and 2 to 3 schistocytes per high-power field. Creatinine is elevated at 1.7  mg/dL (baseline creatinine, 1.0  mg/dL). Liver function is within normal limits, except for an elevated total bilirubin at 1.9  mg/dL. Lactate dehydrogenase is elevated at 600 I U/L.

The presence of schistocytes and elevated lactate dehydrogenase supports the diagnosis of a microangiopathic hemolytic anemia. Microangiopathic hemolytic anemia with thrombocytopenia is concerning for a thrombotic microangiopathy (TMA). Table 2 outlines features of TMAs that may present during pregnancy. It is of utmost importance for hematologists to recognize differences between PEC with severe features/HELLP and thrombotic thrombocytopenia purpura (TTP)/complement-mediated TMA (cm-TMA) as their management differs greatly (with urgent delivery indicated only for the former). A brief schema for evaluating a TMA during second and third trimesters of pregnancy is shown in Figure 3. For an in-depth review, we refer readers to recent an international working group report by Fakhouri et al.⁹ 

PEC with severe features/HELLP is by far the most common cause of a TMA in the late second/third trimesters.¹⁰  The pathogenesis of PEC is incompletely understood, but it is hypothesized that abnormal placental angiogenesis contributes to the disorder.¹¹  PEC is defined by new-onset hypertension (systolic blood pressure ≥140  mm Hg or diastolic blood pressure ≥90  mm Hg) and proteinuria (although not required if other features of organ impairment are present).⁸  A platelet count <100 × 10⁹/L is a criterion for PEC with severe features (as is other organ impairment such as renal dysfunction or elevated liver function tests).⁸  While there is no definitive test to predict PEC, one potential biomarker is the soluble fms-like tyrosine kinase 1/placental growth factor ratio. A prospective analysis of over 1000 women showed that a soluble fms-like tyrosine kinase 1/placental growth factor ratio of ≤38 had a 99.3% negative predictive value for short-term development of PEC among women in their 37th week or less of pregnancy.¹²  Some centers have adopted this as a tool in predicting early PEC, but further prospective studies are needed to determine their widespread clinical utility.⁸  Other pregnancy-related TMAs to consider are acute fatty liver of pregnancy (a very rare cause of a TMA with profound liver dysfunction).⁶  Disseminated intravascular coagulation is not specific to pregnancy but is mentioned here as it is often triggered by obstetric complications (eg, acute peripartum hemorrhage, amniotic fluid embolism, or retained products of conception).^13,14 

Pregnancy is a clear trigger for both cm-TMA and TTP. Thus, while these are rare disorders, it is not uncommon for their first presentation to occur during pregnancy or postpartum. TTP results from an acquired deficiency in ADAMTS13 in the case of immune TTP (iTTP) or an inherited deficiency (congenital TTP [cTTP]), leading to uncleaved von Willebrand multimers binding platelets and causing arterial and venous thrombosis.⁹  Unlike in the nonpregnant adult population, in whom iTTP is far more common, the frequency of cTTP presenting as a first episode in pregnancy is nearly equal to iTTP.¹⁵  Features that should prompt consideration of TTP as the etiology of a TMA include presentation prior to 20 weeks, severe neurologic symptoms, or cardiac injury.¹⁶  TTP should be easy to discern from other TMAs due to its supportive laboratory diagnostics (ADAMTS13 < 10%). In practice, however, the turnaround time for the ADAMTS13 assay (a send-out test at most institutions) is several days, adding complexity to the diagnosis. When TTP is suspected, treatment with plasma exchange should not be delayed while awaiting confirmatory ADAMTS13 testing given high mortality without prompt treatment.¹⁶  Once the diagnosis is confirmed by ADAMTS13 < 10%, ADAMTS13 antibody testing should be performed. Patients with cTTP will lack a detectable antibody. However, not all patients with iTTP will have detectable inhibitors if they have low-titer or nonneutralizing antibodies. The lack of ADAMTS13 recovery in remission and undetectable antibody should prompt ADAMTS13 sequencing to further confirm the diagnosis of cTTP.¹⁷ 

cm-TMA (formerly known as atypical hemolytic uremic syndrome) is associated with inherited genetic mutations in complement or autoantibodies that result in overactivation of the complement pathway.¹⁸  Severe acute kidney injury (stage 3 by RIFLE [Risk of renal dysfunction, Injury to the kidney, Failure of kidney function, Loss of kidney function, and End-stage kidney disease] or AKIN [Acute Kidney Injury Network] criteria) should raise high suspicion for this disorder.¹⁶  Kidney injury requiring dialysis is not a typical feature of PEC or TTP and is in fact a rare occurrence in pregnancy (a recent study reported only 188 cases in nearly 2 million pregnancies).^16,19  Hematologists should also consider this diagnosis when patients present with new-onset or worsening TMA postpartum.¹⁶ 

A diagnosis of TTP is considered, and ADAMTS13 testing is sent. A multidisciplinary discussion with hematology and obstetric teams believes the presentation is most consistent with PEC with severe features based on onset in the third trimester, hypertension, and degree of acute kidney injury. Urgent induction of labor is recommended. Three days following delivery, the platelet count remains 80 × 10⁹/L, and the patient becomes anuric with a creatinine of 3.0  mg/dL. Hepatic function and prothrombin time are normal. ADAMTS13 returns at 40%. Given the lack of clinical resolution after delivery and severe renal injury, cm-TMA is suspected. The team initiates eculizumab with rapid improvement in her platelet count followed by renal recovery over the next few weeks.

Once the diagnosis is established, the management of pregnancy-related TMA is driven by the obstetrics team. With PEC with severe features/HELLP, delivery is recommended at or beyond 34 weeks, and expectant management before 34 weeks is only appropriate when strict selection criteria are applied and with close monitoring of maternal and neonatal status.⁸ 

As case 1 highlights, the diagnosis of TTP or cm-TMA can be difficult to distinguish from the more common PEC with severe features/HELLP in the late second/third trimester. However, efforts should be made to attempt to identify cases of TTP and cm-TMA as delivery is not typically required for these disorders, nor is it definitive management. For a detailed review of the management of TTP and cm-TMA, we refer readers to a recent American Society of Hematology education review by Scully.¹⁰  Briefly, TTP and cm-TMA are managed similarly in pregnancy as they would be in the general population.²⁰  Key caveats for the management of iTTP during pregnancy are as follows: (1) caplacizumab does not have sufficient safety data to be routinely recommended for use, and (2) a more nuanced risk/benefit analysis of rituximab is needed due to neonatal risks.^20,21  cTTP can be managed with fresh-frozen plasma infusions alone instead of plasma exchange if the diagnosis is known. For cm-TMA, the C5 complement inhibitor, eculizumab, has safety data to support its use during pregnancy and breastfeeding.^10,22  Both cm-TMA and TTP can be associated with good maternal and fetal outcomes when promptly diagnosed and managed, even in early pregnancy, and thus do not necessitate urgent delivery when responding to therapy.

A 30-year-old woman with chronic ITP treated with eltrombopag desires to conceive. Before starting eltrombopag, she had multiple ITP relapses with platelets <10 × 10⁹/L without major bleeding, which responded to steroids. Her platelets have remained in the 100 to 200 × 10⁹/L range since starting eltrombopag 2 years ago. After discussion with her hematologist, she decides to discontinue eltrombopag while trying to conceive. Two months later, she becomes pregnant. Her platelet count remains above 30 × 10⁹/L during early pregnancy, and thus she remains off medications. She notes petechiae when she is 28 weeks pregnant and is found to have a platelet count of 5 × 10⁹/L.

ITP is caused by autoantibodies to platelets that accelerate their clearance.²³  Pregnancy is not contraindicated in a patient with chronic ITP but would ideally be delayed in the setting of recent major bleeding, refractory severe thrombocytopenia, and/or thrombocytopenia requiring teratogenic medications.³  Like the general population, patients with ITP can also develop gestational thrombocytopenia (or other pregnancy-specific causes of thrombocytopenia), resulting in lower platelet counts than their typical baseline. Reese et al¹  reported that in pregnant patients with a preexisting diagnosis of ITP, platelet counts of less than 80 × 10⁹/L occurred in over 50% (13 of 24). A retrospective analysis of data from a tertiary care center found approximately one-third of patients with known ITP require treatment during pregnancy.²⁴  Additionally, they reported 42% of patients with preexisting ITP before pregnancy responded to first-line therapies when they required treatment during pregnancy.²⁴  Prenatal counseling in a patient with ITP should include discussion of what therapies may be required during pregnancy. For some patients, the desire for a future pregnancy may play a role in choosing their long-term management strategy for chronic ITP, as many of the medications for chronic ITP have limited data in pregnancy (outlined below).²⁵  Splenectomy as a treatment modality for chronic ITP offers the best chance of a durable response (occurs in ~50%-70% of patients) and may limit the need for medications during pregnancy, although this cannot be guaranteed.^25,26 

Most patients who can maintain a platelet count ≥30 × 10⁹/L during pregnancy do not require treatment until near delivery. Table 3 outlines therapies for ITP and considerations for pregnancy. First-line treatment is generally corticosteroids, but recommended starting dose varies, with some clinicians trialing lower doses (10-20  mg prednisone daily) if thrombocytopenia is not severe and delivery is not imminent.²  Intravenous immunoglobulin (IVIG) also has good safety data in pregnancy and can be used when a more rapid platelet count rise is needed, such as prior to delivery. A retrospective review of 149 patients with ITP during pregnancy found similar responses with corticosteroids vs IVIG (38% vs 39%, P = .71).²⁷  Notably, response to both of these first-line therapies was lower than those typically reported in nonpregnant women. Still, ITP refractory to steroids/IVIG requiring second-line therapies is a rare entity in pregnancy, although incidence is not well defined.

In a patient with ITP refractory without any response to first-line therapies, an alternative etiology for thrombocytopenia should be considered. This is particularly important to consider in patients who lack even a short-lived response to IVIG and without a history of ITP outside of pregnancy. In a patient with other cytopenias, a bone marrow biopsy may be considered (noting this is typically not necessary for the diagnosis of ITP).^28,29  For second-line treatment of ITP in pregnancy, splenectomy has been described (although not commonly performed and may carry increased risk when performed in pregnant patients) until late in the second trimester.^30,31  Rituximab has data in pregnancy but carries potential risks of neonatal B-cell depletion.³²  While use of thrombopoietin receptor agonists (TPO-RAs) in ITP management is increasingly common, to date, use of TPO-RAs in pregnancy has been limited and caution is required given they are likely to cross the placenta.³³ Table 4 reviews outcomes from 1 clinical trial recombinant human TPO (rhTPO), available only in China,³⁴  and the largest retrospective review of TPO-RA in pregnancy to date.³⁵  Michel et al³⁵  examined outcomes from 15 women (17 pregnancies) with preexisting ITP treated with romiplostim or eltrombopag in preparation for delivery (58%), refractory symptomatic ITP (4 pregnancies), or continuation of treatment prior to pregnancy (3). Median exposure during pregnancy to these agents was 4.4 weeks. Response to TPO-RA was seen in 77% of cases. No maternal or fetal thrombotic events were noted. Prospective studies on the use of TPO-RA in refractory ITP during pregnancy are needed, but unfortunately, at the time of this writing, clinicaltrials.gov lists no active trials other than with rhTPO (NCT05333744). Given the paucity of data, the use of the TPO-RA in pregnancy should be limited to patients with refractory disease and requires a thorough risk/benefit discussion with the patient. Another novel agent for ITP in nonpregnant individuals, fostamatinib, is associated with reported adverse effects on organogenesis in animal studies and thus should not be used during pregnancy or lactation.³⁶ 

The patient is started on 1  mg/kg prednisone with minimal initial improvement in her platelet count IVIG (1 g/kg) daily for 2 days is added with a transient rise to 50 × 10⁹/L platelets but a fall to 9 × 10⁹/L 2 weeks later. She is concerned with infectious risk of the immunosuppressants. After discussing with her hematologist again about the limited data for TPO-RAs in pregnancy, she opts for restarting eltrombopag. Two weeks after initiation of eltrombopag, her platelet count improves to 80 × 10⁹/L. Prednisone is weaned off over the course of 1 month. At 39 weeks' gestation, she goes into spontaneous labor, and her platelet count is 90 × 10⁹/L. She is able to receive epidural anesthesia given stable platelet counts and has an uncomplicated vaginal delivery.

Thrombocytopenia is particularly challenging near time of delivery due to the hemostatic challenges of childbirth and risks with epidural anesthesia. The platelet threshold necessary for epidural placement is not well studied and can vary by institution but is generally avoided at platelet counts less than 80 × 10⁹/L.³⁷  Platelet counts of 30 to 50 × 10⁹/L are otherwise felt to be sufficient for vaginal delivery or cesarean section in the absence of another coagulopathy. Mode of delivery should be guided by obstetric indication.²⁴  Tranexamic acid administered shortly after delivery can be a useful adjunct in patients with thrombocytopenia.³⁸  As for neonatal outcomes with maternal ITP, studies report 10% to 15% of neonates are born with a platelet count less than 150 × 10⁹/L³⁹  Severe bleeding, including intracranial hemorrhage, appears rare in the neonate, and its presence should raise concern for an alternate etiology, such as neonatal alloimmune thrombocytopenia.^6,40 

Determining the etiology of thrombocytopenia in pregnancy requires a multidisciplinary approach between hematology and the obstetrics team, particularly to distinguish cm-TMA/TTP from the more common hypertensive disorders of pregnancy. In patients with refractory ITP, further prospective studies evaluating TPO-RA use in pregnancy are needed.

Allyson M. Pishko is supported by a 2019 Hemostasis and Thrombosis Mentored Research Award supported by Sanofi Genzyme. Figures and visual abstract made with biorender.com.

Allyson M. Pishko has received research funding on behalf of her institution from an educational grant from Sanofi Genzyme.

Ariela L. Marshall: no competing financial interests to declare.

Thrombopoietin-receptor agonists in pregnancy is discussed.