Diagnosis and management of maternal thrombocytopenia in pregnancy

Authors

  • Bethan Myers

    Corresponding author
    1. Obstetric Haematologist Nottingham University Hospitals, Nottingham, , UK
    • Department of Haematology, Lincoln County Hospital, Lincoln, UK
    Search for more papers by this author

Correspondence: Dr B. Myers, Department of Haematology, Lincoln County Hospital, Greetwell Rd, Lincoln LN2 5QY, UK. E-mail: bethan.myers@ulh.nhs.uk

Summary

Thrombocytopenia is a common finding in pregnancy, occurring in approximately 7–10% of pregnancies. It may be a diagnostic and management problem, and has many causes, some of which are specific to pregnancy. Although most cases of thrombocytopenia in pregnancy are mild, and have no adverse outcome for either mother or baby, occasionally a low platelet count may be part of a more complex disorder with significant morbidity and may be life-threatening. Overall, about 75% of cases are due to gestational thrombocytopenia, 15–20% secondary to hypertensive disorders; 3–4% due to an immune process, and the remaining 1–2% made up of rare constitutional thrombocytopenias, infections and malignancies. In this review, a diagnostic approach to investigating thrombocytopenia in pregnancy is presented, together with antenatal, anaesthetic and peri-natal management issues for mother and baby, followed by a detailed discussion on the specific causes of thrombocytopenia and the management options in each case.

The platelet count in pregnancy is slightly lower than in non-pregnant women (Abbassi-Ghanavati et al, 2006). Most studies report a reduction in platelet count during pregnancy, resulting in levels about 10% lower than pre-pregnancy level at term (Boehlen et al, 2000; Jensen et al, 2011). The majority of women still have levels within the normal range; however, if pre-pregnancy levels are border-line, or there is a more severe reduction, the level may fall below the normal range. The mechanisms for this are thought to be dilutional effects and accelerated destruction of platelets passing over the often scarred and damaged trophoblast surface of the placenta (Fay et al, 1983). Platelet counts may also be lower in women with twin compared with singleton pregnancies, possibly related to greater increase of thrombin generation (Sunoda et al 2002).

Thrombocytopenia is a common finding in pregnancy, occurring in 7–10% pregnancies (Verdy et al, 1997). It may be a diagnostic and management problem, and has many causes, some of which are specific to pregnancy.

Women with low platelet counts in pregnancy are generally less symptomatic due to the procoagulant state induced by increased levels of fibrinogen, factor VIII and von Willebrand Factor (VWF), suppressed fibrinolysis and reduced protein S activity (Calderwood, 2006).

Although most cases of thrombocytopenia in pregnancy are mild, with no adverse outcome for mother or baby, occasionally a low platelet count may be part of a complex disorder with significant morbidity and may (rarely) be life-threatening.

Overall, about 75% of cases are due to gestational thrombocytopenia; 15–20% secondary to hypertensive disorders; 3–4% due to an immune process, and the remaining 1–2% made up of rare constitutional thrombocytopenias, infections and malignancies (Burrows & Kelton, 1990). Table 1 summarizes the causes. Non-pregnancy specific causes will only be discussed with respect to specific pregnancy-related differences in management.

Table 1. Causes of maternal thrombocytopenia in pregnancy
  1. HELLP, Haemolysis, Elevated Liver enzymes, and Low Platelets; CMV, cytomegalovirus; EBV, Epstein–Barr virus; HIV, human immunodeficiency virus.

Pregnancy-specific conditions
Gestational thrombocytopenia~75% cases, commonest cause of thrombocytopenia in pregnancy (Burrows & Kelton, 1990)

Hypertensive disorders including pre-eclampsia

HELLP syndrome

Thrombocytopenia occurs in 50% cases of pre-eclampsia (McCrae et al, 1992)

0·5–0·9% pregnancies (Kirkpatrick, 2010)

Acute fatty liver of pregnancy1:7000–1:20 000 pregnancies (Knight et al, 2008)
Pregnancy-associated conditions: (not specific but increased association)
Thrombotic thrombocytopenic purpuraOne in 25 000 pregnancies (Dashe et al, 1998)
Disseminated intravascular coagulation

In HELLP syndrome: 20% all cases 84% severe cases (Sibai & Ramadan, 1993; Sibai et al 1993)

In Pre-eclampsia: rarely, in severe cases

Haemolytic uraemic syndromeOne in 25 000 pregnancy-associated cases (Fakhouri et al, 2010)
Non-pregnancy associated
Spurious0·1% general population (Garcia et al, 1991)
Immune thrombocytopenia

0·1–1/1000 pregnancies (Segal & Powe, 2006)

Commonest cause thrombocytopenia in 1st and 2nd trimesters

HereditaryRare
Marrow diseaseVery rare
Viral infectionCommon temporary cause; consider screening for CMV, EBV, hepatitis C. HIV and hepatitis B are routinely screened in pregnancy

Folate/B12 deficiency

Drugs/hypersplenism

All rareAll rare

Diagnostic approach to investigating thrombocytopenia in pregnancy

As for the non-pregnant patient, careful history, examination and laboratory workup is helpful. An approach to diagnostic assessment is summarized in Table 2, and a suggested approach to management in the haematology obstetric clinic setting is presented in Fig 1. The extent of testing for each woman should be individualized depending on the platelet count and clinical features.

Figure 1.

Suggested approach to management in the haematology obstetric clinic.

Table 2. Diagnostic assessment of thrombocytopenia in pregnancy
  1. HELLP, Haemolysis, Elevated Liver enzymes, and Low Platelets; APTT, activated partial thromboplastin time; DIC, disseminated intravascular coagulation; (a)HUS, (atypical) haemolytic uraemic syndrome.

1. History: record – any previous obstetric experience, neonatal blood counts; past response to treatment, such as corticosteroids for immune thrombocytopenia, family history of bleeding, auto-immune disorders
2. Physical examination: usually normal aside from bleeding manifestations, these are unusual unless the count is very low. Check blood pressure (pre-eclampsia), and for abdominal tenderness (HELLP syndrome)
3. Relevant laboratory tests depend on the stage of pregnancy and other factors:
 (a) At all stages:
  Repeat blood count, blood film to confirm thrombocytopenia, and urgently exclude presence red cell fragments (microangiopathies)
  Coagulation screen – care in interpretation as APTT shortens through pregnancy (DIC) Consider checking auto-immune profile in selected cases
 (b) 2nd trimester onwards: Liver function, haemolysis screen – (HELLP)
 (c) Post-partum: Renal function (HUS/aHUS)
 (d) Family history bleeding/thrombocytopenia – von Willebrand screen for type 2b; blood film for hereditary macrothrombopathies
  (e) Changes in white cell count +/− lymphadenopathy: consider bone marrow N.B. left shift in white cell series is normal in pregnancy

Antenatal management – general aspects

The key initial assessment is a blood film to confirm that the thrombocytopenia is genuine and to urgently exclude the presence of microangiopathy. In those with no adverse features, antenatal management depends on the extent of the thrombocytopenia. In general, in women with platelet counts above 100 × 109/l, monthly checks by the midwife or general practitioner are sufficient, with instructions to refer in to the specialized Haematology Obstetric clinic if platelet counts fall below 80–100 × 109/l, or if there is unexplained bleeding or extensive bruising.

In the first and second trimesters, low platelet counts are most likely due to an immune process, or rarely, a platelet production defect. In these cases, a multidisciplinary approach, involving obstetricians and haematologists antenatally, and later anaesthetists and neonatologists, is required for optimal care.

The aim of antenatal management in these cases is to achieve and maintain a ‘safe’ rather than normal platelet count. An experienced team and an individualized approach are important. Although there are no robust trials, expert opinion and retrospective studies advise that asymptomatic patients with immune thrombocytopenia (ITP) and platelet counts >20–30 × 109/l do not need any treatment until the third trimester (Provan et al, 2010). In the rare cases of bone marrow production defects, such as myelodysplasia, levels may need to be higher to avoid bleeding as platelet function may also be impaired, and platelet transfusion may be necessary (Provan et al, 2010). A conservative approach minimizes the risks to mother and fetus from exposure to therapeutic agents, such as prednisolone (see below).

Treatment will be necessary during the first two trimesters if the patient is symptomatic, the platelet count falls below 20–30 × 109/l or a procedure is required. A platelet count of 50 × 109/l is usually taken as adequate for procedures (Provan et al, 2010).

The full blood count (FBC) can be checked monthly until the 3rd trimester and then 2-weekly, increasing to weekly as term approaches (Provan et al, 2010). This pattern should be adjusted according to the absolute platelet count and the rate of decline.

From 35 to 36 weeks, or prior to any intervention, it is usual to give treatment to raise the platelet count to at least 50 × 109/l, and higher levels are advised (usually ≥80 × 109/l) to enable epidural anaesthesia or analgesia (see below), although these levels are not evidence-based.

General measures, including avoidance of aspirin, non-steroidal anti-inflammatory drugs, and intramuscular injections, should also be considered depending on the platelet level and stability. As low dose aspirin is now frequently prescribed in pregnancy for many indications, this should not be withheld unless risk of bleeding is considered high.

Perinatal management

Considerations for mother

The main maternal concern is haemorrhage, during delivery or post-partum. Serious haemorrhage following vaginal delivery remains uncommon, even with severe thrombocytopenia, in ITP (Webert et al 2003), but may be of concern post-natally given the fall in pro-coagulant factors at this time. Haemorrhage is unusual with platelet counts >50 × 109/l. The aim, usually, is to attain a platelet count of ≥80 × 109/l, to allow for regional anaesthesia/analgesia, as many anaesthetists would not consider an epidural below this level. An anaesthetic consultation in the third trimester to discuss options for delivery is helpful.

The mode of delivery should be based on obstetric considerations given there is no evidence that Caesarean section is safer for the fetus with thrombocytopenia than an uncomplicated vaginal delivery, which is usually safer than caesarean section for the mother.

Where the maternal platelet count remains low (<50 × 109/l) around the time of delivery, platelets should be available on standby, but are likely to be destroyed quickly after transfusion if due to immune process, so should be timed judiciously, given in well-established rather than early labour, if there are increased bleeding complications.

Anaesthetic management

The decision regarding regional anaesthesia should ideally made before delivery with an obstetric anaesthetist. Epidural analgesia is of concern, as even a small increase in venous haemorrhage has the potential for spinal cord compression. There is little data to support a minimum ‘safe’ platelet count and each case must be risk-assessed individually. Patients with a platelet count >80 × 109/l in the absence of pre-eclampsia or additional risk factors are unlikely to have abnormal platelet function and most experienced anaesthetists would use this threshold (van Veen et al, 2010). There are too few reports of epidural haematoma following regional blockade in obstetric patients to give an accurate incidence of this complication (Douglas, 2005). The current international guidelines on the management of ITP recommend a threshold of 75 × 109/l, based on expert opinion (Provan et al, 2010).

Some experienced obstetric anaesthetists consider a platelet count of 50 × 109/l to be adequate (Letsky & Greaves, 1996). Spinal anaesthesia may allow a Caesarean section to be performed without the need for a general anaesthetic, as the risk of vascular damage is likely to decrease with needle size.

Considerations for baby

Historically, the major concern regarding delivery in mothers with thrombocytopenia of undetermined cause has been the risk of neonatal thrombocytopenia and intracranial haemorrhage (ICH). The two main differential diagnoses that may be difficult to separate until after delivery are gestational thrombocytopenia and ITP, as both are diagnoses of exclusion. The former is considered a completely benign condition for mother and baby whereas ITP may result in the fetus having thrombocytopenia from the passage of antibodies across the placenta. Antibodies produced in ITP are immunoglobulin G (IgG) in nature and are therefore able to cross the placenta, with the potential to cause thrombocytopenia in the fetus. There is variability in percentages reported for neonatal thrombocytopenia, with 15–50% infants with platelet counts <100 × 109/l (Neylon et al, 2003; Ozkan et al, 2010; Gasim, 2011), 8–30% <50 × 109/l, and 1–5% with severe thrombocytopenia (<20 × 109/l). A recent retrospective study of 11 797 maternal-neonatal pairs found no correlation between individual maternal and neonatal platelet counts. However, if the lowest maternal count was <50 × 109/l the relative risk of thrombocytopenia in the neonate was 4·6 [95% confidence interval (CI) 1·8–33·3], and of severe neonatal thrombocytopenia was 7·8 (CI 1·8–33·3) (Jensen et al, 2011).

A history of splenectomy, previously thought to be an indicator of increased risk of neonatal thrombocytopenia, has recently been questioned (Cines & Bussel, 2005). After splenectomy, the maternal platelet count may be normal, but antibodies are still circulating and the infant may be affected. Maternal ITP refractory to splenectomy correlates with a higher risk of ICH in the fetus or neonate (Koyama et al, 2012).

Neonatal thrombocytopenia is more likely if there is a sibling with thrombocytopenia (Christiaens et al, 1997). Percutaneous umbilical cord sampling can be performed to measure the fetal platelet count prior to delivery, but is not recommended, as it carries a fetal mortality risk of 1–2%, at least as high as that from ICH (Scioscia et al, 1988).

In contrast to neonatal alloimmune thrombocytopenia, in maternal ITP, neonatal platelet counts are rarely <10 × 109/l. The most feared form of severe bleeding, ICH, has a very low risk (0–1·5%), and this risk is not increased by vaginal delivery (Samuels et al, 1990; Payne et al, 1997). In addition, most haemorrhagic events in neonates occur 24–48 h postpartum at the nadir of the platelet count. Current guidelines therefore recommend that the mode of delivery is determined by obstetric indications, with avoidance of procedures associated with increased haemorrhagic risk to the fetus: fetal scalp electrode/samples, ventouse and rotational forceps (Cines & Bussel, 2005).

A cord sample should be taken to check neonatal platelet count, and intramuscular injection of vitamin K deferred until the platelet count is known. Consider giving orally if the platelet count is <50 × 109/l. Infants with subnormal counts should be monitored, as platelet counts tend to fall to a nadir on days 2–5 after birth (Burrows & Kelton, 1990) In those infants with a platelet count <50 × 109/l at delivery, transcranial ultrasonography is recommended even if the neonate is asymptomatic. Treatment of the neonate is rarely required, but in those with clinical haemorrhage or platelet counts <20–30 × 109/l, treatment with intravenous immunoglobulin (IVIG) 1 g/kg produces a rapid response, usually by 24 h (Ballin et al, 1988). Life-threatening haemorrhage is treated with platelet transfusion combined with IVIG.

As severe thrombocytopenia in neonates due to maternal ITP is rare, when it occurs, testing for parental platelet antigen incompatibility [check parental human platelet antigen (HPA) status] is recommended to exclude feto-neonatal alloimmune thrombocytopenia.

Neonatal thrombocytopenia secondary to maternal ITP may occasionally last for months and, in these cases, monitoring and, occasionally, further doses of IVIG are needed.

Specific conditions causing thrombocytopenia in pregnancy and their management: isolated thrombocytopenia

Gestational thrombocytopenia

Gestational thrombocytopenia is the commonest cause of thrombocytopenia in pregnancy (75% of cases), and is not associated with any adverse events for either mother or baby (McCrae et al, 1992; Verdy et al, 1997). It is a diagnosis of exclusion, generally causes only mild thrombocytopenia, and occurs in the latter half of pregnancy, from mid-second or third trimester. Table 3 summarizes the main features of gestational thrombocytopenia. Most experts consider this diagnosis less likely if the platelet count dips below 70 × 109/l; the main differential diagnosis at this level or lower is ITP. However, there are reports of more severe thrombocytopenia that showed no response to steroids, and which resolved postnatally, consistent with gestational thrombocytopenia (Win et al, 2005). It is not possible to differentiate between the more severe form of gestational thrombocytopenia, and ITP, as both are diagnoses of exclusion. The degree of thrombocytopenia is not generally low enough to increase risk of bleeding at delivery, but may compromise the ability to receive epidural anaesthesia (see below). Some suggest that a short trial of prednisolone (generally 20 mg/d) may be helpful diagnostically and therapeutically when the platelet count is around 50–70 × 109/l (McCrae, 2010).

Table 3. Features of gestational thrombocytopenia
1. No specific diagnostic test; diagnosis of exclusion
2. Causes mild thrombocytopenia, usually >70 × 109/l
3. Not associated with maternal bleeding
4. No past history of thrombocytopenia outside pregnancy
5. Occurs in mid second/third trimester
6. No associated fetal thrombocytopenia
7. Spontaneous resolution after delivery
8. May recur in subsequent pregnancies

The maternal platelet count normalizes shortly after delivery, usually very quickly, but within 1–2 months in all cases.

Immune thrombocytopenia

Primary immune thrombocytopenia (ITP) occurs in 1/1000–1/10 000 pregnancies, accounting for around 3% of women thrombocytopenic at delivery (Gill & Kelton, 2000; Segal & Powe, 2006). Although an uncommon cause of thrombocytopenia in pregnancy, it is the commonest cause of a low platelet count in the first and second trimesters (Gill & Kelton, 2000). A recent consensus group meeting recommended using a platelet count of <100 × 109/l to define ITP (Provan et al, 2010). Most studies report two-thirds of women with ITP have pre-existing disease, and one-third are diagnosed for the first time in pregnancy (Webert et al 2003). Despite improved understanding of the pathophysiology, there is no specific diagnostic test and, like gestational thrombocytopenia, it remains predominantly a diagnosis of exclusion. Differentiating ITP from alternative causes of thrombocytopenia encountered in pregnancy can therefore sometimes present a diagnostic challenge. The presence of other autoimmune phenomena or a low platelet count pre-pregnancy may help diagnostically.

Although some studies report exacerbation or relapse of ITP during pregnancy (Fujimura et al, 2002), others found no increase in relapse rate in those with a history of severe ITP (Baili et al, 2009). There are also conflicting reports from studies on maternal and perinatal outcomes. Belkin et al (2009) reported a significant association with adverse outcomes for mother and baby: hypertensive disorders and diabetes mellitus, pre-term delivery (<34 weeks) and perinatal mortality when compared with patients without ITP.

Most other studies have reported a favourable outcome for neonates and mothers (Ozkan et al, 2010; Gasim, 2011; Fujita et al 2010). This may reflect differing approaches to treatments. Current recommendations for management are based on experience and expert-consensus (Provan et al, 2010).

Specific treatments for ITP

Primary treatment options for maternal ITP are corticosteroids and intravenous IVIG (Letsky & Greaves, 1996; Provan et al, 2010). The decision regarding therapy depends on the urgency of the platelet increment, the duration the increment is required for and potential side effects, and should be made on an individual patient basis.

ITP patients with moderate/severe thrombocytopenia (<20–30 × 109/l)

The usual first-line treatment is prednisolone, but at much lower doses than used out with pregnancy, 10–20 mg daily for a week, adjusting to the minimum dose that achieves a haemostatically effective platelet count (Provan et al, 2010). Response time is as for non-pregnant women, 3–7 d, as is response rate. Maternal risks include inducing or exacerbating gestational diabetes, maternal hypertension, osteoporosis, weight gain, and psychosis. Platelet count may fall more rapidly near term, so defer dose adjustment until after delivery, and then taper slowly to avoid a rapid fall in platelet count and to ensure the mother's psychological state is not affected.

In the short term, low-dose steroids are considered safe and effective for the mother. However, low dose maternal steroids do not have any beneficial effect on the fetal platelet count (Christiaens et al, 1990). Prednisolone is metabolized by the placenta but high doses have the potential to cause effects on the fetus including premature rupture of membranes, adrenal suppression, and a small increase in cleft lip after use in the first trimester (Ostensen et al, 2006; Gisbert, 2010).

ITP patients with very severe thrombocytopenia (≤10 × 109/l) or significant bleeding

Intravenous immunoglobulins are used in similar dosage to the non-pregnant population with comparable response rates (Provan et al, 2010). Some pregnant women tolerate high dose IVIG poorly. IVIG has the same potential side effects as in the non-pregnant population. Duration of response is short, usually lasting 2–3 weeks, and repeated infusions may be needed to prevent haemorrhagic complications and attain an adequate platelet count for delivery.

In life-threatening bleeding platelet transfusion plus combination treatment with IVIG and a pulse of methyl prednisolone (0·5–1 g/d for 3 d) is necessary.

Intravenous anti-D at a dose of 50–75 μg/kg has been used in non-splenectomized Rh (D)-positive patients. Experience is limited: some describe positive results with response rates comparable to IVIG with the advantage of bolus administration (Michel et al, 2003); however, complications, though uncommon, may be life-threatening, and monitoring is required for neonatal jaundice, anaemia and direct anti-globulin test positivity after delivery. Therefore this is best avoided until further safety data is available. This product is currently unavailable in the UK.

Management of refractory ITP cases: - azathioprine, ciclosporin, splenectomy, rituximab, thrombopoietin receptor agonists

Occasionally, a patient may be refractory to standard treatments. In these circumstances, a bone marrow examination is indicated to exclude rare alternative diagnoses (see Table 1). For pregnant patients with refractory disease, the options are limited by the toxic and teratogenic effects of cytotoxic agents. Vinca alkaloids, cyclophosphamide and androgen analogues should not be used. For other treatments, listed below, the balance of risks of bleeding versus potential toxic effects of more aggressive therapies needs to be assessed on an individual basis.

Azathioprine

Older data available for pregnancy use of azathioprine in systemic lupus erythematosus and renal transplantation suggested that it is safe during pregnancy (Erkman & Blythe, 1972; Price et al, 1976), and this is supported by a consensus paper (Provan et al, 2010). Recent studies suggest a possible association between azathioprine and intrauterine growth restriction and preterm delivery (Cleary & Kallen, 2009; Gisbert, 2010), although it remains unclear whether these outcomes are manifestations of the underlying disease. The US Federal Drug Administration (FDA) states: ‘Positive evidence of fetal risk is available, but the benefits may outweigh the risk if life- threatening or serious disease.’ (Category D) (http://depts.washington.edu/druginfo/Formulary/Pregnancy.pdf). Despite this, it has been used successfully in pregnant patients for many years. The very delayed onset of action limits its use as a steroid-sparing agent in pregnancy.

Ciclosporin A

Ciclosporin A has not been associated with significant toxicity to mother or fetus during pregnancy. Data mainly comes from its use in transplant and inflammatory bowel disease, and show no association with congenital abnormalities or premature delivery (Reindl et al, 2007; Reddy et al, 2008). As hypertension is a common side-effect, and seizures have been reported, it is not often used, but could be considered for refractory cases (Provan et al, 2010).

Splenectomy

This is very rarely necessary during pregnancy. When considered essential, the standard recommendation is that it is best performed in the second trimester (Provan et al, 2010). It may be performed laparoscopically but the technique may be difficult beyond 20 weeks' gestation. Pre-splenectomy immunizations can be safely given in pregnancy, as all of the pre-splenectomy vaccines required are inactivated. However, the British National Formulary (BMJ Publishing Group, 2011; Joint Formulary Committee, 2011) recommendation is that inactivated vaccines should be administered only if protection is required without delay.

Rituximab

There is limited information regarding the anti-CD20 monoclonal antibody, rituximab, in pregnancy. The manufacturer recommends that women of reproductive age avoid pregnancy during treatments and for 12 months following treatment. It has been assigned to pregnancy category C by the FDA, which states: ‘animal studies have not been conducted (or adverse results reported), and there are no controlled data in human pregnancy’ (http://depts.washington.edu/druginfo/Formulary/Pregnancy.pdf). One study identified 153 pregnancies associated with maternal rituximab exposure for varied indications, with known outcomes, from which there were 90 live births; of these, there were 22 premature births, one neonatal death (6 weeks), 11 haematological abnormalities (no infections), and two congenital malformations (clubfoot in one twin, and cardiac malformation in a singleton) (Chakravarty et al, 2011). One maternal death from pre-existing autoimmune thrombocytopenia occurred. It is difficult to draw firm conclusions from this review as the maternal conditions were serious and likely to have an adverse effect on pregnancy outcome. There are several reports of good outcome when rituximab was used for ITP in the 2nd/3rd trimesters of pregnancy (Gall et al, 2010; Schmid et al, 2011).

Rituximab crosses the placenta and may cause a delay in neonatal B-lymphocyte maturation. This seems to normalize at between 4 and 6 months postpartum, and the reports to date have not described any significant clinical consequences (Klink et al, 2008; Gall et al, 2010).

Thrombopoietin receptor agonists

There is no clinical information in pregnancy about the new thrombopoietin receptor agonists, romiplostim and eltrombopag. Animal studies have shown evidence of thrombocytosis, post-implantation losses, and an increase in mortality with romiplostim (http://www.medicines.org.uk/emc/medicine/23117/SPC#PREGNANCY). Both drugs are classified as category C in pregnancy (see above).

ITP and venous thromboembolism (VTE) prophylaxis

There are a number of studies that suggest that ITP confers a prothrombotic state, (Severinsen et al, 2011). Thromboprophylaxis risk assessment should be carried out as normal, and thromboprophylaxis considered unless the platelet count is <50 × 109/l or there is bleeding (Provan et al, 2010). Below this level, discussion with a haematologist with experience in this area should be sought regarding thromboprophylaxis.

Pre-pregnancy counselling

Women with an established diagnosis of ITP who wish to become pregnant should be offered pre-pregnancy counselling. Table 4 summarizes the relevant issues that should be addressed. In general, pregnancy should not be discouraged, but the risks of possible relapse or worsening of ITP, possible need for treatment, risks of haemorrhage and neonatal thrombocytopenia should be discussed.

Table 4. Key points in pre-pregnancy counselling for women with immune thrombocytopenia (ITP)
1. Explain that ITP may worsen or relapse during pregnancy
2. Up to one-third of women require treatment in pregnancy, especially peri-natally
3. Detail treatments for ITP, their side-effects and risks to mother and baby
4. Explain increased risk of bleeding, but this risk is small, even with a very low platelet count
5. Epidural analgesia/anaesthesia may not be possible, but anaesthetic review will be arranged to discuss alternatives
6. Serious adverse maternal outcomes are very rare
7. Maternal platelet count is not predictive of the baby's platelet count
8. The risk of severe haemorrhage, such as intracranial haemorrhage, for the fetus/neonate is very low and is not reduced by caesarean section
9. Explain measures that are put in place to minimize trauma to baby's head at delivery

Hypertensive and microangiopathic disorders of pregnancy: thrombocytopenia + microangiopathy

These include the hypertensive disorders pre-eclampsia, HELLP (Haemolysis, Elevated Liver enzymes, and Low Platelets) syndrome (see below), haemolytic uraemic syndrome (HUS) and thrombotic thrombocytopenic purpura (TTP). There is overlap between these conditions with the common features of thrombocytopenia and microangiopathy, which may cause considerable diagnostic difficulties. Typical characteristics of each are shown in Table 5, which may help distinguish between them. Diagnosis can be especially problematic post-natally, because all the conditions may present in the immediate postnatal period. However, delay in diagnosis, especially in TTP, is rapidly life-threatening, and so where there is diagnostic uncertainty, immediate referral to a centre with expertise is advised, for consideration of plasma exchange (PEX).

Table 5. Hypertensive and microangiopathic disorders of pregnancy: usual characteristics
CharacteristicTTPHUSHELLPPre-eclampsiaAFLPAPSSLE
  1. AFLP, acute fatty liver of pregnancy; APS, antiphospholipid syndrome; CNS, central nervous system; DIC, disseminated intravascular coagulation; HELLP, haemolysis, elevated liver enzymes and low platelet count syndrome; HUS, haemolytic uraemic syndrome; MAHA, microangiopathic haemolytic anaemia; SLE, systemic lupus erythematosus; TTP, thrombotic thrombocytopenic purpura.

  2. Reproduced with permission from Myers (2009).

OnsetAny timePostpartumThird trimesterThird trimesterThird trimesterAny timeAny time
HypertensionNoNo+/−++++/−±±
MAHA++++++++/−+−/±±/+++
Thrombocytopenia++++++++/−+/±++
DICNo+/−+++/−+++±±
Liver disease+/−+/−++++/−+++±
Renal disease+/−+++++±±+/++
CNS disease++++/−+/−+/−±++
ManagementPlasma exchangeSupportiveDeliver fetus, blood productsDeliver fetus

Correct metabolic defect

Deliver fetus

LMWH

Aspirin

Aspirin

(LMWH)

(Hydroxy-chloroquine)

Pre-eclampsia/toxaemia (PET)

PET affects 6% of all first pregnancies. Criteria for the diagnosis include: new-onset hypertension with ≥140 mmHg systolic or ≥90 mmHg diastolic after 20 weeks' gestation together with proteinuria (ACOG Committee on Practice Bulletins–Obstetrics, 2002). It is characterized by multisystem involvement, with systemic endothelial dysfunction affecting many organ systems, especially the kidney (Mirza & Cleary, 2009), together with activation of the coagulation system. Thrombocytopenia is the commonest abnormality, occurring in up to 50% of women with pre-eclampsia, and the severity parallels the PET. The hypercoagulability of normal pregnancy is accentuated, even when platelet counts appear normal. PET also affects the liver, with elevated aspartate aminotransferase and lactic dehydrogenase levels, although increments are small, except when the HELLP syndrome supervenes. The pathogenesis is thought to be due to inadequate placentation. Elevated levels of various growth factors are present in ‘pre-eclamptic’ placentae from late first trimester, characterized as vascular endothelial growth factor receptor type1 (Maynard et al, 2005) and endoglin, a tumour growth factor-β receptor derived from the endothelium (Venkatesha et al, 2006). There is decreased nitric oxide production and endothelial dysfunction as a result of interaction of these with placental growth factor (Myatt & Webster, 2009).

Due to the other well recognized features of this condition, thrombocytopenia in PET is unlikely to be confused with the previously described conditions. However, there may be overlap with the HELLP syndrome.

HELLP syndrome

The HELLP syndrome is a serious complication specific to pregnancy characterized by: Haemolysis, Elevated Liver enzymes, and Low Platelets. It occurs in about 0·5–0·9% of pregnancies and in 10–20% of cases with severe pre-eclampsia (Kirkpatrick, 2010). A 70% of cases develop antenatally, usually in the third trimester, and the remainder within 48 h after delivery. The syndrome can occur without proteinuria (25%) or hypertension (40%), and the diagnosis may then be missed. Presenting features are often vague, with nausea, malaise and upper abdominal pain. There should be a low threshold for checking blood count and liver function. The pathophysiology is similar to that of pre-eclampsia, with endothelial damage and release of tissue factor and coagulation activation. A recent study identified mutations in genes that regulate the alternative complement system in four of 11 patients with HELLP, suggesting that excessive complement activation may be involved in the pathogenesis (Fakhouri et al, 2008), similar to atypical HUS. Haemolytic anaemia, thrombocytopenia and red cell fragments are characteristic, secondary to the endothelial damage. Liver function shows raised liver enzymes, and disseminated intravascular coagulation (DIC) complicates >80% severe cases. This is a progressive condition and serious complications are frequent. Abdominal pain reflects obstructed blood flow in hepatic sinusoids, with the development of hepatic necrosis. Medical stabilization, with blood products, followed by delivery, is indicated if HELLP occurs after the 34th week or the fetal and/or maternal conditions deteriorate. Vaginal delivery is preferable. Conservative management for very early HELLP is controversial, with need for risk-benefit assessment between gaining extra time for the baby versus maternal risks. Close surveillance of the mother should be continued for at least 48 h after delivery, as the process may worsen before improvement occurs.

The severity and frequency of complications for mother and infant was shown in a study by Celik et al (2003), who reported 36% incidence of acute renal failure, 17% placental abruption, 6% DIC, one case of adult respiratory distress syndrome, and two (of 36) patients died. Intrauterine death occurred in 19% and 30% of birth-weights were below 1·5 kg. Five of 11 babies died in the neonatal period. Several studies have reported the rarer complications of hepatic rupture and liver transplantation (Zarrinpar et al 2007; Smyth, 2011). These emphasize the need for prompt recognition of HELLP syndrome, close observation and transfer to obstetric centres with expertise in this field, and a multidisciplinary approach to improve materno-fetal prognosis.

Thrombotic thrombocytopenic purpura (TTP)

The reader is referred to the newly updated British Committee for standardization in Haematology (BCSH) TTP guideline (Scully et al, 2012) for detailed diagnostic and management information.

TTP is a life-threatening condition characterized by microangiopathic haemolytic anaemia, thrombocytopenia, fever, neurological abnormalities and renal dysfunction, due to a deficiency of VWF cleaving protein, ADAMTS13. The coagulation screen is normal in TTP, and may help to differentiate it from other microangiopathies. TTP is more common in women (3:2), and nearly 50% cases occur in women of child bearing age. TTP occurs in approximately one in 25 000 pregnancies. Unlike HELLP, it is not specific to pregnancy, but occurs with increased frequency in association with pregnancy, in 5–25% of cases (Vesely et al, 2004; Scully et al, 2008). In one series, 55% of cases occurred in the second trimester (Sadler, 2008). There is a risk of relapse during subsequent pregnancies, although women with normal levels of ADAMTS13 pre-pregnancy have a lower risk of relapse (Ducloy-Bouthors et al, 2003; Scully et al, 2006).

There is a normal drift down of ADAMTS13 levels in the 2nd and 3rd trimesters of pregnancy, resolving post-partum (Sanchez-Luceros et al, 2004).

Management of TTP in pregnancy

Any delay in treatment for TTP is life-threatening for mother and baby; treatment with PEX should be commenced without delay. Regular PEX may enable pregnancy to continue successfully (Scully et al, 2006). The optimal frequency of PEX during pregnancy is unknown. Where possible, delivery is the treatment of choice for pregnancy-associated microangiopathies, however delivery does not guarantee remission (Hovinga & Meyer, 2008).

The recommendations for pregnancy management from the 2012 guideline are:

  1. If a thrombotic microangiopathy (TMA) cannot be explained by a non-TTP pregnancy-related TMA, then the diagnosis of TTP must be considered and PEX started.
  2. Mothers with congenital TTP should attend a specialist centre and receive ADAMTS13 supplementation regularly throughout pregnancy and post-partum.
  3. Close liaison with an obstetrician with a special interest in feto-maternal medicine is required in mothers with TTP. Serial fetal monitoring with uterine artery dopplers should be used to assess fetal growth and placental blood flow.
  4. In mothers with acquired TTP, ADAMTS13 activity should be monitored throughout pregnancy to help predict the need for adjuvant therapy and outcome.
  5. Pre-conceptual counselling is advised for subsequent pregnancies.

Counselling women – risk of recurrence

The rarity of the condition and possible reporting bias makes precise advice difficult. In one study, following a pregnancy-associated event, the risk of recurrence was 50%, with a live birth rate of 67% (Vesely et al, 2004). Prophylactic treatment with PEX in subsequent pregnancies may be appropriate, on a fortnightly basis. The baseline ADAMTS13 level in early pregnancy may be useful in predicting likelihood of relapse, and hence help to guide use of PEX. Reduction in ADAMTS13 activity (<10%) in early pregnancy may require elective therapy.

Supportive therapy: This is a prothrombotic condition and thromboprophylaxis with low molecular weight heparin (LMWH) is recommended once the platelet count is >50 × 109/l (Yarranton et al, 2003). Red cell transfusions and folate supplementation are required during active haemoylsis. The clinical efficacy of antiplatelet agents in TTP is unproven in terms of response rate in recent trials (Bobbio-Pallavicini et al, 1997) but it is relatively safe, and may help to reduce placental thrombosis.

Haemolytic uraemic syndrome (HUS)

HUS is a microangiopathy similar to TTP, but with predominant renal involvement. An atypical form is usually seen in association with pregnancy, with no evidence of infection. Although most patients with HUS have renal insufficiency as the prominent component, there is still considerable overlap with TTP. The levels of ADAMTS13 are generally not severely reduced, but management decisions must usually be made before this result is available. A useful clinical feature distinguishing atypical HUS from TTP is the timing of onset: most cases of HUS occur several weeks postpartum. There are recently published guidelines on the diagnosis and management of atypical HUS and the reader is referred to this text for more detailed discussion (Taylor et al, 2010).

In one retrospective study of 100 women with atypical HUS, 21% had experienced pregnancy-associated HUS symptoms, of which 79% occurred postpartum. Complement abnormalities were detected in 90% of the cases with pregnancy-related disease. A 76% developed end-stage renal disease (Fakhouri et al, 2010). The poor outcome in this study underlines the gravity of this syndrome. There is poor response to PEX, although a trial of PEX should be undertaken, especially as the diagnosis may be uncertain between atypical HUS (aHUS) and TTP. Anticoagulants and antiplatelet agents are not beneficial.

Eculizumab, a terminal complement inhibitor, appears a promising agent for aHUS (Nurnberger et al, 2009). The most recent study reports sustained normalization of 13/15 aHUS patients with thrombocytopenia and four of five patients became dialysis-free (Greenbaum et al, 2011). Eculizumab is approved for use in aHUS in the USA and reports of eculizumab in pregnancy (in PNH) support its safety in this setting (Kelly et al, 2010).

Acute fatty liver of pregnancy (AFLP)

AFLP is a rare, serious disorder that presents in the third trimester or postpartum with nausea, vomiting, right upper quadrant pain and cholestatic liver function. It occurs in 1:7000–1:20 000 deliveries, and still has a maternal mortality rate of around 15%. Laboratory findings include low platelet count, prothrombin time, low fibrinogen, and low antithrombin levels in addition to raised bilirubin levels. This results in a clinical picture similar to DIC; however, in AFLP, the values are abnormal not due to consumption of the clotting factors but rather to decreased production by the damaged liver. Treatment requires intensive supportive care with blood product support for the coagulopathy (McCrae, 2006).

Neonatal outcome in microangiopathies

The prognosis for the baby in all the microangiopathies described is poor, because of extensive placental ischaemia.

Miscellaneous causes of thrombocytopenia not specific to pregnancy

Disseminated intravascular coagulation (DIC)

Pregnancy-related DIC often presents a dramatic clinical picture. The obstetric events associated with DIC include placental abruption, amniotic fluid embolism, and uterine rupture, where there is profound activation of the clotting system and severe consumption of clotting factors. DIC may, however, develop more gradually in the case of retained fetal products, and thrombocytopenia may be the presenting feature (McCrae & Cines, 1997).

Antiphospholipid syndrome (APS)

A detailed account is presented elsewhere (Myers & Pavord, 2011). Immune thrombocytopenia may be associated with APS, usually mild, or moderate in degree. In those with primary or secondary APS, standard management in pregnancy is to give LMWH and aspirin, which improves neonatal outcome. Occasionally, the degree of thrombocytopenia may make this treatment more difficult to achieve. In this circumstance, steroids may be required to boost the platelet count, as the mechanism of the thrombocytopenia in APS is the same as in ITP. See also Table 5.

Familial thrombocytopenias

Inherited thrombocytopenias, such as the May–Hegglin anomaly, may first come to light in pregnancy. These conditions are usually due to defective platelet production, and, depending on the level may require platelet transfusion (or on standby) for delivery. Patients with a bleeding history should receive tranexamic acid (see below) postpartum. As there is a risk of neonatal thrombocytopenia, traumatic delivery should be avoided, and a cord sample taken at birth.

Type 2B von Willebrand disease (type 2B VWD)

This is a rare subtype of VWD, with increased affinity for platelet receptor glycoprotein 1b, which binds to platelets inducing spontaneous platelet aggregation, accelerating platelet clearance and hence causing thrombocytopenia.

During pregnancy, the abnormal VWF increases and thrombocytopenia may become evident or more pronounced, with the platelet count occasionally falling as low as 20–30 × 109/l. Women with type 2B VWD (and other types) should be managed by a multidisciplinary team at a haemophilia centre. There is variability of response of VWD to pregnancy, and regular monitoring of the VWF, factor VIII (FVIII) level, and platelet count is essential. These should be taken at booking, before invasive procedures, and at 28 and 34 weeks.

Levels of VWF and FVIII should be increased to >50 iu/dl, to cover any surgical procedures, spontaneous miscarriage, and delivery. For delivery, infusion should start at the onset of established labour, with the aim of maintaining levels >50 iu/dl for ≥3 d after vaginal delivery and ≥5 d after caesarean section. Platelet transfusions may be required if the platelet count falls to <20 × 109/l antenatally, prior to invasive procedures, if bleeding, or if <50 × 109/l at delivery, together with the factor replacement. Epidural anaesthesia is not recommended, as fully normalized coagulation cannot be guaranteed, even with treatment. Whilst normal vaginal delivery is the aim, prolonged second stage should be avoided (as judged by obstetricians), with early recourse to caesarean section if necessary, to reduce risk of trauma to mother and baby. Fetal scalp monitoring/sampling, ventouse and rotational forceps should be avoided. Postpartum, ensure surgical haemostasis and uterine contraction. Prophylactic tranexamic acid should be given (Pavord, 2010).

Malignant haematological disorders

These are very rare, and include infiltrative marrow disorders, such as metastatic disease, and bone marrow syndromes, such as myelodysplasia. In some, platelet counts may also be dysfunctional, making bleeding risk more likely for any given platelet count. This is one of the few indications for a bone marrow during pregnancy to clarify the diagnosis and plan appropriate management. The management plan will depend on the exact diagnosis and stage of pregnancy, and the risk to the mother of delay in treatment if deferred until after delivery. Neonates may also be at risk of low platelet counts.

Nutritional deficiencies

Severe deficiency of either folic acid or vitamin B12 may cause low platelet counts, usually accompanied by a low red and white cell count. However, both are rare in pregnancy. Folic acid supplementation preconceptually and in early pregnancy is recommended to reduce risk of neural tube deformatities, and has reduced the deficiency as a cause of FBC changes. Vitamin B12 deficiency is rare in pregnancy because those with established B12 deficiency are subfertile. B12 levels may be difficult to interpret in pregnancy; in standard laboratory testing, the levels fall during pregnancy due to changes in protein binding.

Medication

Although there is less use of drugs in pregnancy, medication should always be considered as a possible cause of thrombocytopenia.

Heparin-induced thrombocytopenia (HIT) has been described, very rarely, in pregnancy. Greer and Nelson-Piercy (2005) reviewed prophylactic dose of LMWH use during 2777 pregnancies and identified no cases of HIT, and only one possible case was reported in another three studies (Warkentin et al, 2008); therefore routine monitoring is not recommended for prophylactic doses in pregnancy. The BCSH guideline recommends monitoring when therapeutic doses are used, or when unfractionated heparin is used. In this case the occurrence of HIT is still uncommon (0·1–1%: Lee & Warkentin, 2007). HIT is an intensely thrombotic process, despite the low platelet count and, after stopping heparin, alternative anticoagulation is needed to avoid thrombotic risk. The alternative drugs used for HIT include lepirudin, danaparoid and more recently, fondaparinux. Danaparoid sodium is a heparinoid, distinct from heparin and thus has little cross-reactivity. It does not cross the placenta and is not secreted in breast milk. There are anecdotal reports of its use in pregnancy (Myers et al, 2003; Gerhardt et al, 2009). Fondaparinux is not licensed for use in HIT, but due to the ease of use and favourable outcomes of its use in HIT outside pregnancy, there are occasional reports of use in pregnancy, for women with allergy to LMWHs and, in one case of HIT, following use of unfractionated heparin in pregnancy (Ciurzynski et al, 2011). Lepirudin is a synthetic direct thrombin inhibitor. It crosses the placenta, and should be avoided in pregnancy, although there are a few case reports of its use in pregnancy, in the second and third trimesters (Harenberg et al, 2005).

Tranexamic acid

Tranexamic acid is an antifibrinolytic that is regularly used outside pregnancy to stabilize clotting in bleeding disorders, including thrombocytopenia. It crosses the placenta, but has been used to treat bleeding during pregnancy in a limited number of cases of type 2B VWD and other bleeding disorders without reported adverse fetal effects (Lindoff et al, 1993). There are no adequate well-controlled trials of its use in pregnancy and its systemic use should be considered only in refractory patients with ongoing symptoms after the 1st trimester. It is a category B drug (Pregnancy Category B – Animal reproduction studies have failed to demonstrate a risk to the fetus but there are no adequate and well-controlled studies in pregnant women).

A recent meta-analysis concluded it is safe (with respect to thrombotic risk) and effective in pregnancy-associated use (Peitsidis & Kadir, 2011)..

Conclusion

In conclusion, there are a large number of causes of thrombocytopenia in pregnancy. By far the most common causes are mild, but awareness of the complex disorders in which thrombocytopenia occurs is essential to ensure prompt diagnosis and referral into a centre with expertise in these rare conditions, to optimize outcome for mother and baby.

Ancillary