Recommendation for investigation of suspected ITP in pregnancy:
As in non-pregnant patients, the diagnosis of ITP is one of exclusion. Benign gestational thrombocytopenia is 100 times more common but, as there are no definitive diagnostic tests, it may be impossible to distinguish these conditions until a non-pregnant platelet count is available. All women with platelet counts < 100×109/ l should be screened for clinical or laboratory evidence of pre-eclampsia, coagulopathy or autoimmune disease. Bone marrow examination is unnecessary unless there is suspicion of leukaemia or lymphoma. The routine measurement of PAIg or platelet antibodies is not recommended (Evidence Levels III / IV, Grade C recommendations).
OPTIONS FOR TREATMENT
Optimum management of ITP in pregnancy requires close collaboration between the obstetrician, haematologist and paediatrician. There are no high quality prospective studies or randomized clinical trials to inform management of the mother, delivery or the neonate. Recommendations are based on clinical experience and expert consensus (Grade C recommendation).
Management of the pregnant woman with ITP
The decision to treat the pregnant woman with ITP is based on assessment of the risk of significant haemorrhage. The count usually falls as pregnancy progresses, the greatest rate of decline and nadir occurring in the third trimester (Burrows & Kelton, 1992). Therefore, careful planning is required to ensure a 'safe' platelet count at the time of delivery. Frequency of monitoring depends on the individual case, taking into account the absolute platelet count, rate of change and proximity of delivery.
Targets for treatment
Asymptomatic patients with platelet counts > 20 ×109/ l do not require treatment until delivery is imminent but should be carefully monitored, both clinically and haematologically (Letsky & Greaves, 1996). Platelet counts of > 50×109/ l are regarded as safe for normal vaginal delivery (Letsky & Greaves, 1996) and some experts extend this to levels of 30–50×109/ l (Lichtin, 1996). A platelet count of > 50×109/ l is also safe for Caesarian section but would preclude the use of epidural anaesthesia for which the platelet count should be > 80×109/ l. Because of the theoretical risk of haematoma formation and neurological damage, spinal or epidural anaesthesia is not recommended if the platelet count is < 80> 50×109/ l (Letsky & Greaves, 1996). There is no evidence that the template bleeding time predicts the risk of haemorrhage in this situation.
Treatment options
The major treatment options for maternal ITP are corticosteroids or IVIg. Vinca alkaloids, androgens and most immunosuppressive drugs should not be used in pregnancy although azathioprine has been used safely in transplant patients.
If the duration of treatment is likely to be short, i.e. starting in the third trimester, corticosteroids are a costeffective option. An initial dose of 1 mg / kg (based on prepregnancy weight) is recommended (Burrows & Kelton, 1992; Letsky & Greaves, 1996), subsequently tapered to the minimum haemostatically effective dose. Patients must be carefully monitored for signi?cant side effects such as hypertension, hyperglycaemia, osteoporosis, excessive weight gain and psychosis. As 90% of the administered dose of prednisolone is metabolised in the placenta (Smith & Torday, 1982), serious fetal side effects such as adrenal suppression are unlikely. The only systematic study showed no bene?cial effect of low dose maternal steroids on the fetal platelet count (Christiaens et al, 1990).
If steroid therapy is likely to be prolonged, signi?cant side effects occur or an unacceptably high maintenance dose is required (perhaps > 7.5 mg prednisolone daily) IVIg therapy should be considered. There are no published comparative trials of steroids and IVIg in pregnancy. The conventional dose of IVIg is 0.4 g / kg / d for 5 d although 1 g / kg for 2 d has been used successfully and may be more convenient (Burrows & Kelton, 1992). The response rate (80%) and duration of response (2–3 weeks) is similar to non-pregnant patients. After an initial response, repeat single infusions can be used to prevent haemorrhagic symptoms and ensure an adequate platelet count for delivery. The ability of maternal IVIg therapy to improve fetal platelet counts remains controversial (Nicolini et al, 1990). IVIg carries the same potential risks and side effects as in the non-pregnant patient and the ?nancial cost is much higher than corticosteroids.
Therapeutic options for those women with severely symptomatic ITP refractory to oral steroids or IVIg include high dose intravenous methyl prednisolone (1000 mg), perhaps combined with IVIg, or azathioprine (Letsky & Greaves, 1996). From available data, the latter appears to cause no signi?cant problems to either mother or fetus (Erkman & Blythe, 1972; Price et al, 1976; Alstead et al, 1990). Splenectomy in pregnancy is now rarely performed. If essential, it is best carried out in the second trimester and may be successfully performed by the laparoscopic route, although this may be technically dif?cult beyond 20 weeks gestation.
Recommendation: (All Grade C)
♦ Asymptomatic women with platelet counts > 20×109/ l do not need treatment until delivery is imminent
♦ Platelet counts > 50×109/ l are safe for normal vaginal delivery in patients with otherwise normal coagulation
♦ Platelet counts > 80×109/ l are safe for caesarean section, spinal or epidural anaesthesia in patients with otherwise normal coagulation
♦ In women who need treatment, both oral corticosteroids and IVIg appear to have a similar response rate to the use of these agents in the non-pregnant patient Although many clinicians now favour the use of IVIg in pregnancy there are no good comparative studies and the decision must take into account maternal clinical factors and preference in addition to the expense, availability and (remote) risks of microbial transmission by IVIg.
♦ There are no convincing data on the effect (bene?cial or otherwise) of corticosteroids or IVIg on the fetal ? neonatal platelet count.
♦ Severe refractory ITP may respond to high dose IV methyl prednisolone ± IVIg or azathioprine. If essential, splenectomy may be performed (ideally in the second trimester) and the laparoscopic route may have clinical advantages similar to those seen in nonpregnant patients.
Management of delivery and the newborn infant
Historically, management of the delivery in mothers with ITP was dominated by concerns over the risk of severe neonatal thrombocytopenia and haemorrhage. In 1976, caesarean section was recommended for all patients based on a reported perinatal mortality of 12–21%, largely due to birth trauma and ICH (Murray & Harris, 1976). These data were clearly selective and excessively pessimistic; more recent reviews suggesting a neonatal mortality of around 0.6% (Burrows & Kelton, 1992). Large prospective studies published in the 1990s show an incidence of severe neonatal thrombocytopenia (< 50×109/ l) of 8.9–14.7% with ICH occurring in 0–1.5% of infants (Bussel et al, 1991b; Burrows & Kelton, 1993). There is no evidence that caesarean section is safer for the thrombocytopenic neonate than uncomplicated vaginal delivery (which is unequivocally safer for the mother). In any event, most haemorrhagic events in neonates occurred 24–48 h after delivery at the nadir of the platelet count.
Attempts have been made to predict which neonates will have severe thrombocytopenia to inform management decisions. It is reported (evidence level III) that the fetal or neonatal platelet count cannot be reliably predicted from the maternal platelet count, maternal platelet antibody levels (by a variety of techniques) or a history of maternal splenectomy for ITP (Samuels et al, 1990; Burrows & Kelton, 1992; Letsky & Greaves, 1996; Sainio et al, 1998; Boehlen et al, 1999). Fetal blood sampling by cordocentesis has been explored (Scioscia et al, 1988) but this procedure carries a mortality of 1–2%, which is at least as high as the risk of ICH. Scalp blood sampling in early labour to measure the fetal platelet count has also been advocated (Scott et al, 1980). This procedure is technically dif?cult, commonly produces artefactually low results because of clotting due to exposure to vernix or amniotic ?uid and can cause significant haemorrhage. For these reasons, both cordocentesis and fetal scalp blood sampling have now been largely abandoned in the management of ITP in pregnancy.
In view of the difficulty in predicting those few neonates with severe thrombocytopenia and the very low risk of serious haemorrhage, it is now generally agreed (Level III evidence, Grade B recommendation) that the mode of delivery in ITP should be determined by purely obstetric indications (Bussel et al, 1991c; Burrows & Kelton, 1992; Letsky & Greaves, 1996).
Following delivery, a cord blood platelet count should be determined in all cases (Level III evidence, Grade C recommendation). Those infants with subnormal counts should be closely observed clinically and haematologically as the platelet count tends to fall further to a nadir between d 2 and 5 after birth (Burrows & Kelton, 1992). Treatment of the neonate is rarely required. In those with clinical haemorrhage or platelet counts < 20×109/ l treatment with IVIg 1 g / kg produces a rapid response. Life-threatening haemorrhage should be treated by platelet transfusion combined with IVIg (Burrows & Kelton, 1992).
Severe thrombocytopenia and clinical haemorrhage in neonates are suf?ciently unusual in association with maternal ITP that where they occur, neonatal alloimmune thrombocytopenia should be excluded by laboratory testing. This is important not only for the appropriate management of the neonate, but also in relation to the antenatal management of subsequent pregnancies.
