Clinical management of paroxysmal nocturnal haemoglobinuria in pregnancy: three case reports

Introduction

First described in 1928, paroxysmal nocturnal haemoglobinuria (PNH) is a rare disease resulting from an acquired defect of one or more lineages of haematopoietic stem cells. The prevalence is estimated to be approximately 1–10 cases per million people, with the disease mainly affecting young adults, including women of childbearing age¹.

PNH is characterised by paroxysmal intravascular haemolysis with anaemia, neutropenia, thrombocytopenia, bone marrow disorders, a thrombotic tendency and increased susceptibility to infections^2,3. Until 2007, the treatment of this condition was mainly supportive, including transfusions of red blood cells and platelets, folic acid and iron replacement, glucocorticoids and anticoagulation therapy. Bone marrow transplantation appears to be the only potentially curative option, but it carries a 10–20% risk of early mortality³. Eculizumab, a humanised monoclonal antibody, has recently revolutionised the treatment of PNH. The use of this antibody for the treatment of PNH was approved by the United States Food and Drug Administration in March 2007 and subsequently by the European Medicines Agency in June 2007^3,4.

We describe the clinical evolution during pregnancy and post-partum period of three women diagnosed with PNH and referred to the obstetric outpatient clinic of our tertiary care level hospital (Table I). The diagnosis of PNH was confirmed by laboratory studies in all three cases. Venous blood samples were collected in EDTA K3 and a Redquant^® kit from Biocytex was used. The samples were treated according to standard protocols and submitted to flow cytometry in a FACS Calibur^® (Becton-Dickinson, Enzifarma Portugal) using the "Cell Quest" program. The data were analysed with "Paint-a Gate" PRO software. Three cell lineages were studied after appropriate labelling: red blood cells (CD55-FITC and CD59-FITC), monocytes (CD48-FITC/CD87-PE/CD14-PerCP/CD4-APC), and granulocytes (CD24-FITC/CD87-PE/CD16-PerCP). Diagnostic criteria were the existence of at least two positive markers in a cell line, with at least two cell lines affected.

Table I.

Description of cases.

	Case 1	Case 2	Case 3
Maternal age (years)	29	25	27
Gravida/Para	G1 P0	G1 P0	G1 P0
PNH diagnosis	5 years prior to pregnancy	14th week of gestation	8 years prior to pregnancy
Flow cytometry	Erythrocytes: 22% CD55−; 58% CD59−. Monocytes: 68% CD48−, CD87− and CD14−. Granulocytes: 81% CD24−, CD87− and CD16−	Erythrocytes: 30% CD55−; 82% CD59−. Monocytes: 83% CD48−, CD87− and CD14−. Granulocytes: 82% CD24−, CD87− and CD16+	Erythrocytes: 63% CD55−; 84% CD59−. Monocytes: 99% CD48−, CD87− and CD14−. Granulocytes: 99% CD24−, CD87− and CD16+
First appointment	9th week	14th week	18th week
Reason for referral	Anaemia	Anaemia Thrombocytopenia	Anaemia Previous thrombosis
Treatment	LMWH from week 30	LMWH from week 16	Warfarin up to week 18 LMWH after week 18
Transfusion therapy	1 RBC unit in weeks 14, 21, 29, and 35	5 RBC units in week 14 14 units of platelets during labour	1 unit of RBC every 2 weeks
Maternal complications	Sinus longitudinal thrombosis	Severe thrombocytopenia Preterm labour	Haemolytic crisis
Delivery	Elective caesarean section at 37th week	Induced vaginal delivery at 34th week	Induced vaginal delivery at 34th week
Anaesthesia	General anaesthesia	None	Epidural anaesthesia
Newborn	2,300 g, Apgar 9/10	2,240 g, Apgar 9/10	2,340 g, Apgar 8/9
Anticoagulation after delivery	Warfarin	Warfarin	Warfarin
Contraception	Levonorgestrel-IUD	Levonorgestrel-IUD	Male condom

Case reports

Case 1

A 29-year-old primipara of African descent was referred to our outpatient clinic in the 9^th week of gestation for monitoring of her pregnancy. She had been diagnosed with PNH 5 years previously, following investigations for persistent fatigue and haemolytic anaemia (haemoglobin [Hb]=8.2 g/dL, lactate dehydrogenase=2,384 U/L). After oral iron replacement, she had remained asymptomatic without needing regular transfusion therapy. She had no prior history of thrombosis but she is heterozygous for the MTHFR mutation with a normal homocysteine level. At the time of referral she was taking aspirin (100 mg/day), advised by her general practitioner, and we decided not to discontinue this until the 35^th week of pregnancy. At the first appointment, she had moderate haemolytic anaemia (Hb=9.4 g/dL) but, during pregnancy, she needed to be transfused with a total of 7 units of red blood cells (in weeks 14, 21, 29 and 35), in order to maintain her Hb level around 10 g/dL.

Pregnancy was uncomplicated until the 33^rd week, when she began to complain of a strong frontal headache and was given prophylactic subcutaneous Low Molecular Weight Heparin (LMWH), (enoxaparin 4,000 UI anti-Xa, daily). Despite this, 2 weeks later, the patient was admitted to hospital because of persistent headache; she was treated with intravenous analgesia and was discharged home after clinical improvement and cerebral magnetic resonance imaging, which showed no abnormalities.

At the beginning of the 37^th week she was readmitted to our hospital because of worsening symptoms and computed tomography of the brain showed a longitudinal sinus thrombosis. Anticoagulation therapy was started with unfractionated intravenous heparin and, in order to avoid the stress of vaginal delivery, we performed an elective caesarean section. The woman's baby, a female, weighed 2,300 g, had an Apgar score of 9 at the 1^st minute and 10 at the 5^th minute and did not require neonatal intensive care. The puerperium was uneventful and the patient began anticoagulation with warfarin, before being discharged home on the 16^th day after delivery. Cerebral magnetic resonance imaging performed 1 month later showed complete resolution of the thrombosis. Contraception with a levonorgestrel-releasing intrauterine device (IUD) was instituted 3 months after delivery.

Case 2

A 25-year old previously healthy woman in her first pregnancy was referred to our outpatient clinic in the 14^th week of gestation because of severe thrombocytopenia (platelets=43x10⁹/L) and anaemia (Hb=10.6 g/dL) found during first trimester laboratory screening tests. She went through our protocol study for the investigation of thrombocytopenia and, after laboratory evidence of macrocytic anaemia and intravascular haemolysis, the diagnosis of PNH was confirmed by flow cytometry. Anticoagulation with subcutaneous LMWH (dalteparin, 5,000 UI anti-Xa, twice daily) was started in the 16^th week of gestation (Hb=9.3 g/dL; platelets=53x10⁹/L) and maintained throughout pregnancy. Haematological and foetal growth parameters were closely monitored on an outpatient basis until the third trimester of pregnancy and five units of red blood cells were administered during that period because of worsening anaemia (lowest Hb=6 g/dL).

By the 30^th week of pregnancy, the woman was experiencing symptoms of preterm labour and recurrent epistaxis. At that time, she had severe thrombocytopenia (platelets=18x10⁹/L) and anaemia (Hb=9.2 g/dL), and was admitted to hospital for transfusion therapy and intensive clinical monitoring. She had no need for tocolysis but, after induction of foetal lung maturity with betamethasone in the 34^th week, we decided to induce labour with an oxytocin infusion, with concurrent transfusion of platelets. Epidural analgesia was not administered because of the severe thrombocytopenia. The woman had a vaginal delivery, giving birth to a live male infant weighing 2,240 g with a normal Apgar score (9/10). The baby was transferred to the neonatal intensive care unit because of prematurity and had a favourable evolution. The mother had no haemorrhagic, thrombotic or infectious complications during the post-partum period, and began oral anticoagulation with warfarin before being discharged home. All haematological values had returned to baseline within 8 weeks of delivery and, by that time, she had had a levonorgestrel-releasing IUD introduced for contraception. A bone marrow biopsy showed no marrow failure and she undergoes regular haematological surveillance with sporadic episodes of haemoglobinuria and fatigue, but without needing regular transfusion therapy.

Case 3

A 27-year old primigravida, who was a regular smoker and an occasional user of cocaine, was referred for evaluation in the 18^th week of gestation because of severe anaemia (Hb=6 g/dL) and increased thromboembolic risk. She had been diagnosed as having PNH in another country when she was 19 years old, following hepatic vein thrombosis, and had been on warfarin since then. She had had frequent episodes of haemolysis with symptomatic anaemia (lowest Hb=6.8 g/dL) and needed to be transfused with one unit of RBC every 2 months. In the 18^th week of pregnancy, after a complete clinical and laboratory assessment, she was switched to full anticoagulation with LMWH (nadroparin, 5,700UI anti-factor Xa, subcutaneously, twice daily), and was transfused with two units of RBC. Despite this, she had to be admitted to hospital 1 week later because of severe anaemia following a haemolytic crisis (Hb=5 g/dL; platelets=93x10⁹/L), and was transfused with three more units of RBC.

She was followed regularly throughout the pregnancy on an outpatient basis: foetal growth and status were monitored and RBC transfusions were administered regularly when required (approximately every 2 weeks). The second trimester ultrasound scan showed foetal unilateral ventriculomegaly, so amniocentesis was performed, revealing a normal female karyotype.

Because of worsening haematological values, labour was induced in the 34^th week of gestation. The woman was given epidural analgesia and she delivered vaginally a female neonate weighing 2,340 g with an Apgar score of 8/9. The neonate did not require admission to the intensive care unit and the mother's post-partum course was uneventful. The patient was discharged home 11 days after delivery on oral anticoagulation. She continues to require regular red blood cell transfusions because of persistent anaemia.

Discussion

Most eukaryotic cell membrane proteins are embedded in the phospholipid double layer surrounded by a polypeptide chain of hydrophobic amino acids. Others, however, are attached to the outer leaflet of the cell membrane by a glycophosphatidy-linositol (GPI) anchor, synthesised in the endoplasmic reticulum. PNH is caused by a somatic mutation of the phosphatidylinositol glycan complementation class A (PIG-A) gene on chromosome X (Xp22.1), which encodes the glycosyltransferase enzyme that catalyses the first step in the production of the GPI-anchor. This defect results in a partial or complete inability to construct the GPI-anchor, leading to the deficiency or absence of a variety of cell membrane proteins⁵. Two of these proteins, CD55 (DAF, decay-accelerating factor) and CD59 (MIRL, membrane inhibitor of reactive lysis), are complement regulatory proteins and their deficiency renders PNH erythrocytes susceptible to haemolysis. Intravascular haemolysis releases free haemoglobin into the plasma, causing depletion of nitric oxide in tissues which accounts for numerous manifestations of the disease, including oesophageal spasm, male erectile dysfunction, renal insufficiency and thrombosis⁶. PNH can be divided into two subcategories: (i) classic PNH, with a predominance of haemolytic manifestations, and (ii) aplastic anaemia/PNH syndrome in which there is also bone marrow failure⁶.

The diagnosis of PNH has evolved over time⁷. The original assays used to detect PNH erythrocytes (Ham's test, the sucrose haemolysis test and the complement lysis assay) were abandoned because they are both less sensitive and less quantitative than flow cytometry. Since the late 1980s, flow cytometric analysis using antibodies directed against GPI-anchored complement regulatory proteins (GPI-AP) has been the most sensitive and informative assay available for the diagnosis of PNH^7,8. Ideally, at least two different monoclonal antibodies, directed against two different GPI-AP, on at least two different cell lineages are recommended to exclude the possibility that the clinical process is a consequence of an inherited, isolated deficiency of a single GPI-AP. As in patient 2, morphological analysis of the bone marrow aspirate and biopsy and cytogenetics are needed for proper classification, because PNH often occurs in association with bone marrow failure syndromes and clonal myelopathies.

As for most rare disorders, the primary challenge to a successful management of this condition during pregnancy is to consider the diagnosis. PNH should always be suspected when faced with a pregnant young woman with unexplained severe anaemia and/or thrombocytopenia and intravascular haemolysis, as in patient 2. In fact, PNH can be underdiagnosed and mistaken for numerous other conditions in pregnancy⁹.

Although it does not always modify the prognosis, pregnancy complicates the management of PNH and significantly increases the risk for both mother and foetus^9,10. The maternal mortality rate may be as high as 20% and the perinatal mortality rate up to 10%^2,7. The most common causes of maternal mortality are venous thrombosis or embolism (commonly in cerebral or intra-abdominal veins) and infection, and the risk of major complications is higher in the post-partum period. We present one case of cerebral venous thrombosis in the last trimester of pregnancy in a woman who was not receiving anticoagulation therapy; no thrombotic complications occurred in the other two patients who received anticoagulation promptly at referral in the first half of pregnancy. No infectious complications were reported and the post-partum period was uneventful in all three cases.

The major cause of perinatal mortality of babies born of mothers with PNH is severe iatrogenic prematurity, because preterm labour occurs in about half of this pregnancies^2,4. Only in case 1, the pregnancy was allowed to continue until term, but a caesarean section had to be performed at 37 weeks because of a maternal thrombotic event; the other two pregnancies were terminated at 34 weeks because of worsening haematological values, but no neonatal complications related to prematurity occurred.

Supportive transfusion therapy and anticoagulation were the most important treatment modalities used in the management of these pregnancies. We suggest that transfusion therapy should be used to maintain the maternal Hb concentration around 10 g/dL and prophylactic or therapeutic anticoagulation should be started in the first trimester or soon after diagnosis and continued until 4–6 weeks post-partum, to avoid thrombotic complications such as those described in the first case. In patients 2 and 3, anticoagulation was started early in the second trimester and no thrombotic events occurred.

In future cases, we should consider treatment with eculizumab, a humanised monoclonal antibody directed against the terminal complement protein C5, which results in inhibition of complement-mediated cell lysis¹¹. It reduces intravascular haemolysis, stabilises haemoglobin levels, decreases transfusion requirements and improves quality of life in patients with PNH. Pregnancy was an exclusion criterion in the earliest clinical studies of this therapeutic agent^12,13, but in recent studies^14,15 eculizumab was successfully administered to pregnant women, without adverse events and with good maternal and foetal outcomes. The effects of eculizumab on the foetus are unknown, but it does not appear to cross the placenta or to be excreted in breast milk. Indeed, its use appears to be safe and it would be expected to reduce the complications of PNH during pregnancy. However, the data available so far have been collected from a small number of cases and eculizumab is still a pregnancy category C medicine. Patients must be vaccinated to prevent Neisseria infection. More evidence is needed in order to expand the use of eculizumab during pregnancy. Unfortunately, this monoclonal antibody is not currently available for use in our country.

We consider that planned delivery in a hospital capable of providing both obstetric care for high-risk cases and haematological support should be arranged after the 34^th week of gestation, if the woman has normal haematological values and a favourable clinical status. Knowing that the risk of thrombosis is lower with vaginal delivery than with caesarean section, the former should be the preferred choice whenever possible. We always try to provide epidural analgesia for our patients during labour, unless it is contraindicated because of severe thrombocytopenia, as in our second case, or because of the timing of LMWH therapy.

Pregnancy is a prothrombotic state and is usually discouraged in women with PNH. Only a small number of cases have been reported in the literature^2,10. The three cases we present here demonstrate the high morbidity associated with these two conditions when they occur concomitantly. Nevertheless, we believe that prompt referral of the pregnant woman with diagnosed or suspected PNH to a specialised obstetric care centre and management by a multidisciplinary team, familiar with rare haematological disturbances, is the key to successful clinical and obstetric outcomes.