Management of refractory immune thrombocytopaenia in pregnancy

Abstract

A 25-year-old woman with a history of immune thrombocytopaenia (ITP) in childhood was referred to haematology clinic for review with a platelet count of 50 μ/L at 9 weeks gestation, gravida 2, para 0. She developed progressive severe thrombocytopaenia as the pregnancy progressed, with associated bleeding complications. The thrombocytopaenia was refractory to standard therapies. This led to a need for a planned delivery, which was performed via caesarean section under general anaesthetic with platelet transfusion support, Intravenous Immune Globulin (IVIG), high-dose corticosteroid and the thrombopoietin (TPO) mimetic romiplostim. Both the mother and the neonate survived; however, the neonate required treatment for severe prolonged neonatal thrombocytopaenia. The patient subsequently re-presented 15 months later with recurrent ITP complicating another pregnancy, refractory to rituximab but responsive to romiplostim. She had a successful elective caesarean section under epidural anaesthesia, but the neonate once again suffered severe thrombocytopaenia, which was responsive to IVIG.

Background

Thrombocytopaenia in pregnancy is a relatively common presentation with an important list of differential diagnoses to consider.¹ Immune thrombocytopaenia (ITP) in pregnancy may occur de novo in pregnancy, but often occurs in women who have a history of the condition. Standard approaches to management include close observation, corticosteroids or Intravenous Immune Globulin (IVIG) in steroid-refractory cases.¹

This case provides an example of the difficulties faced in managing a pregnant woman with severe treatment-refractory ITP and highlights the risk of associated severe neonatal thrombocytopaenia.

Case presentation

A 25-year-old woman from a regional town attended the haematology outpatient clinic with her partner on referral by her obstetrician for a platelet count of 50×10⁹/L at 9 weeks gestation. The patient’s history included childhood ITP diagnosed at age 9, prior to emigrating to Australia. She underwent splenectomy at that time as the condition had persisted despite multiple lines of therapy, but had not received any formal follow-up for the condition since. She gave a history of occasional episodes of petechiae and epistaxis in adulthood, but no episodes of major bleeding nor menorrhagia. Her medical history was otherwise unremarkable, with no medications nor alternative therapies and no significant family history. She was asymptomatic, with no evidence of bruising or petechiae on examination.

Initial management involved a ‘watch and wait’ approach. She was counselled about the likelihood of the platelet count falling as the pregnancy progressed and the likely need for a planned delivery, along with interventions to improve the platelet count if it remained <75/μL closer to delivery. The risks associated with spinal anaesthesia and various modes of delivery (standard vaginal, instrumental, caesarean) in the setting of thrombocytopaenia were introduced to the couple and they were referred on to the high-risk antenatal clinic.

The patient subsequently presented to her local regional hospital at 20 weeks gestation following a mechanical fall with a platelet count of 13×10⁹/L. She did not sustain any significant injuries and had no significant bruising, with a normal fetal ultrasound. She commenced 1 mg/kg daily of oral prednisolone before being discharged home. Her platelet count rose to a peak of 89×10⁹/L within 2 weeks but had fallen rapidly to 16 ×10⁹/L several weeks later at 23 weeks gestation, when an attempt was made to taper the dose to 50 mg daily. An episode of self-limiting epistaxis, new lower limb bruising and petechiae followed. Given the relatively early stage of her pregnancy and intolerance to steroid side effects at full dose, the patient did not wish to increase the dose. She was given 2 g/kg of IVIG, split over 2 days. Ongoing weekly blood tests were monitored. The plan was to continue maintenance IVIG at 1 g/kg fortnightly.

Within a week she re-presented following a low-speed motor vehicle accident with a seatbelt strain to the lower abdomen. Her platelet count had risen to 41×10⁹/L and her abdominal and fetal ultrasound was normal. She was administered a dose of anti-D immunoglobulin (prior to realisation of maternal Rh(D) positivity) and was admitted for observation and 48 hours of cardiotocography (CTG) monitoring. Kleihauer test was negative for fetal cells.

Despite fortnightly IVIG dosing, the patient became refractory. At 28 weeks she presented with haematuria in the setting of a urinary tract infection, with a platelet count of 3×10⁹/L within 10 days of an IVIG dose. Eltrombopag was commenced with a short burst of high-dose prednisolone and antibiotics.

At 31+ weeks, she presented with widespread petechiae, oral mucosal wet purpura, palate petechiae and scattered bruising. She was admitted again for CTG monitoring, which confirmed fetal well-being, and was administered a platelet transfusion, further high-dose corticosteroids and regular tranexamic acid, in addition to ongoing eltrombopag. She was discharged home once the oral lesions had resolved despite a platelet count of 6×10⁹/L.

Rituximab was discussed as the next line of therapy once it became clear that the ITP was refractory to eltrombopag, which the patient declined. Instead, at 35 weeks, with a platelet count of 6×10⁹/L, the patient elected to commence romiplostim as outlined in the Treatment section.

Investigations and differential diagnosis

Baseline routine blood tests to exclude other diagnoses or secondary causes of thrombocytopaenia associated with ITP were performed and were unremarkable.

During the second successful pregnancy, further testing for platelet antibodies, including work-up for neonatal alloimmune thrombocytopaenia (NAIT), was arranged using blood samples from the first child, the father and the patient. However, the results were inconclusive due to the mother’s recent treatment with IVIG interfering with the testing and will need to be repeated down the line to exclude NAIT as a contributing cause of the severe neonatal thrombocytopaenia.

Genetic testing for inherited thrombocytopaenias was negative.

Treatment

The need for a coordinated multidisciplinary approach with a planned delivery and expert neonatal, obstetric, anaesthetic and haematological support was emphasised in this case. Open disclosure and involvement of the patient and her partner in all decision making were imperative.

Standard front-line therapy with corticosteroids and then IVIG was administered in this case during both pregnancies. The patient showed initial responses to both agents, but as the pregnancies progressed these responses became short-lived, less significant and ultimately refractory.

During the first pregnancy, rituximab was offered to the patient as a third-line therapy in the second trimester. At that time, the couple felt uncomfortable with the use of the agent during pregnancy and instead opted to trial TPO mimetics. Eltrombopag was trialled for 8 weeks with the maximum dose escalation to 75 mg daily with no effect. The corticosteroids were concurrently tapered.

At 35 weeks gestation the eltrombopag was ceased and weekly romiplostim commenced at a dose of 3 μg/kg with rapid dose escalation.

A multidisciplinary meeting was held to discuss the patient’s management and delivery plan at 37+2 weeks gestation, at which time her platelet count was 2×10⁹/L. The consensus was to proceed with an elective caesarean section under general anaesthetic at 38 weeks. All attempts to avoid instrumental delivery, use of fetal scalp electrodes and need for prolonged topical pressure after intramuscular vitamin K administration were documented, as well as the plan for urgent umbilical cord full blood count testing post delivery and daily for 3–5 days post partum.

The day prior to delivery, the platelet count remained 2×10⁹/L. Regular tranexamic acid 1 g intravenously administered four times a day was commenced. After high-dose IVIG at 1 mg/kg, 5 μg/kg of romiplostim on day −1 (the fourth dose of the drug), and 1 mg/kg intravenous methylprednisolone daily on day −1 and on the day of the procedure, her platelet count was 10×10⁹/L. Platelets were transfused during anaesthetic induction and again on delivery of the neonate. Platelet count was 69×10⁹/L intraoperatively and 50×10⁹/L immediately postoperatively. Estimated maternal blood loss was 400 mL and the patient recovered well without further bleeding complications. Her platelet count rose over subsequent days, reaching 191×10⁹/L on day +4.

The neonate was covered in petechiae on delivery, with a platelet count of 2×10⁹/L. He was treated with IVIG, prophylactic platelet transfusions and a short course of corticosteroids. Serial cranial ultrasound monitoring was performed to exclude spontaneous intracranial haemorrhage. There were no significant bleeding complications and the infant’s platelet count showed spontaneous rise at approximately 6 weeks post partum.

After initial good responses to ongoing romiplostim for the first few weeks post partum, the patient found it difficult to continue to attend for therapy while caring for a newborn. At 6 weeks post partum, she declined further blood tests and romiplostim dosing. She expressed concerns about the unknown effect of potential passage of the drug through her breast milk and felt confident managing her own symptoms without regular blood test monitoring.

The patient was then lost to follow-up until once again referred with the most recent pregnancy. She had unfortunately suffered two early miscarriages during that interval. Bloods once again revealed thrombocytopaenia along with iron deficiency anaemia. Initial management was akin to the first pregnancy, with corticosteroids and IVIG administered once the platelet count fell <10×10⁹/L, resulting initially in good responses but soon becoming refractory. Prednisolone was weaned to and continued at a dose of 7.5 mg once daily. The couple decided to proceed with rituximab (anti-CD20 monoclonal antibody) therapy in the second trimester, and 4× weekly doses of 375 mg/m² were administered intravenously, with an initial rise in platelet count to 87×10⁹/L at 6 weeks after commencement and soon fell back to <10×10⁹/L. Fortnightly IVIG was continued (for the ITP but also has NAIT had not been excluded). Romiplostim was recommenced at 29+ weeks at the previously successful dose of 5 μg/kg and with the dose titrated according to response. The platelet count rose to and was maintained above 100×10⁹/L from 33 weeks at a dose of 8 μg/kg romiplostim weekly. This allowed elective caesarean section (obstetric and maternal choice) under epidural anaesthetic to be performed at 37+ weeks gestation with no maternal bleeding complications.

The neonate once again experienced severe thrombocytopaenia with a platelet count of 19×10⁹/L, and remained <20×10⁹/L for the following 4 days. Several doses of IVIG were administered during a 1-week period of postnatal observation with good effect, and the mother and the baby were discharged home with the infant’s platelet count at 51×10⁹/L post IVIG on day +5, with ongoing follow-up underway.

Outcome and follow-up

Both the patient and the newborn remain well several months following delivery. Any future pregnancies will require careful medical planning and early institution of maternal romiplostim if required, while acknowledging high-risk or recurrent neonatal thrombocytopaenia.

No adverse effects to either infant have been observed following intrauterine exposure to the agents used.

Discussion

Standard management of ITP during pregnancy aims to minimise toxicity to the fetus while optimising maternal platelet count to allow safe delivery. Traditional treatment guidelines suggest observation, corticosteroids, IVIG and even splenectomy. Immunosuppressive and chemotherapeutic agents have been used in refractory cases but tend to have a slow onset of action and low success rates with additional toxicities.¹

Outside the pregnancy setting, the novel TPO mimetics, eltrombopag and romiplostim, have provided a useful salvage option for cases of refractory ITP. Documented response rates range from 50% to 90%, and these agents have been shown to significantly reduce bleeding rates and the need for rescue treatments.² Limited data are available regarding their use and safety in pregnancy and are based predominantly on case reports of off-label use.

Michel et al³ recently published a multicentre observational retrospective case series of 15 patients (17 pregnancies) who received the agents for ITP in pregnancy. No major safety concerns were raised; however, there was one case of neonatal thrombocytosis. They observed a response rate of 77%.

A novel recombinant TPO mimetic (rhTPO) has shown promising results in this setting but is currently not readily accessible outside of China. Kong et al⁴ studied the use of rhTPO in 31 pregnant patients with refractory ITP and observed a response rate of 74.2%. No major safety concerns were noted in mothers or infants, with a median follow-up time of 53 weeks post partum.⁴

Learning points.

• Combined multidisciplinary team management and multiagent medical therapies, particularly the thrombopoietin (TPO) mimetic romiplostim, was successful in elevating maternal platelet count in this patient with immune thrombocytopaenia (ITP) to allow safe deliveries during her two pregnancies.

• However, it is prudent to note that this did not protect the infants against severe neonatal thrombocytopaenia, which, although rare (~1.5%), is more commonly seen in the setting of refractory maternal ITP.

• It is imperative that early testing and close monitoring of the neonate’s platelet count be performed and that obstetric and paediatric management plans be prepared for this possibility.

• Early case reports suggest TPO mimetics may be useful as a salvage treatment for cases of ITP in pregnancy refractory to standard therapies.

Ethics statements

Patient consent for publication

Obtained.