Pregnancy-induced thrombocytopenia and TTP, and the risk of fetal death, in Upshaw–Schulman syndrome: a series of 15 pregnancies in 9 genotyped patients

Authors


Prof Yoshihiro Fujimura. Department of Blood Transfusion Medicine, Nara Medical University, 840 Shijo-cho, Kashihara City, Nara, 634-8522, Japan. E-mail: malon@naramed-u.ac.jp

Summary

Upshaw–Schulman syndrome (USS) is a congenital thrombotic thrombocytopenic purpura (TTP) due to mutations in the gene that encodes for ADAMTS13 (ADAMTS13), but its clinical signs may be mild or absent during childhood. We have identified 37 patients with USS (24 females, 13 males) belonging to 32 families. The nine women from six families who were diagnosed during their first pregnancy are the focus of this report. Six of the nine women had episodes of thrombocytopenia during childhood misdiagnosed as idiopathic thrombocytopenic purpura. Thrombocytopenia occurred during the second–third trimesters in each of their 15 pregnancies, with 16 babies (one twin pregnancy), often followed by TTP. Of 15 pregnancies, eight babies were stillborn or died soon after birth, and the remaining seven were all premature except one, who was born naturally following plasma infusions to the mother that had started at 8 weeks’ gestation. All nine USS women had severely deficient ADAMTS13 activity. ADAMTS13 analyses demonstrated that eight women were compound heterozygotes of Y304C/G525D (2 siblings), R125VfsX6/Q1302X (2 siblings), R193W/R349C (2 siblings), I178T/Q929X, and R193W/A606P; one woman was homozygous for R193W. Only the R193W mutation has been previously reported. These observations emphasize the importance of measuring ADAMTS13 activity in the evaluation of thrombocytopenia during childhood and pregnancy.

Upshaw–Schulman syndrome [USS, also described as congenital thrombotic thrombocytopenic purpura (TTP)] is a congenital deficiency of ADAMTS13 (A Disintegrin-like And Metalloprotease with ThromboSpondin type 1 motifs-13) activity due to ADAMTS13 gene mutations (Kinoshita et al, 2001; Levy et al, 2001; Kokame et al, 2002). When ADAMTS13 activity is deficient, unusually large von Willebrand factor (VWF) multimers accumulate in the circulation that can cause platelet thrombi under high shear stress of the microcirculation (Yagi et al, 2001; Moake, 2002). In 2005, the presence of roughly 100 patients with USS in 80 families worldwide was estimated (Kremer-Hovinga et al, 2005). But still USS patients may often be overlooked in clinical practice. Therefore, the frequency of USS is presumably underestimated.

Acquired TTP, caused by neutralizing or non-neutralizing autoantibodies to ADAMTS13 that reduce plasma ADAMTS13 activity, is a disorder with an incidence of 1·7 per million population; 73% of the patients were women and most women were of child-bearing age (Terrell et al, 2005). Although acquired TTP and USS have some similarities, USS has important differences in clinical features from acquired TTP (Furlan & Lämmle, 2001; Fujimura et al, 2002).

During the past 10 years, we have diagnosed 37 patients with USS by assaying ADAMTS13 activity and its inhibitor titres in the laboratory of Nara Medical University. Further, analysis of the natural history and ADAMTS13 mutations in these patients showed that severe neonatal jaundice requiring exchange blood transfusions, a classic hallmark of USS, was seen in only 16 (43%) of 37 patients. Twenty-nine (79%) of the 37 patients had a history of thrombocytopenia during childhood that was misdiagnosed as idiopathic thrombocytopenic purpura (ITP). Nine women from six families were first diagnosed during pregnancy. Further, we documented that thrombocytopenia inevitably developed during the second or third trimester of all 15 pregnancies in these nine women. Often, the initial isolated thrombocytopenia was followed by overt signs of microangiopathic haemolytic anaemia and TTP. Notably, of 15 pregnancies (one twin pregnancy), eight babies were stillborn or died soon after birth, and seven babies were all premature but survived. Only one mature baby was born at full term following plasma infusions to the mother that had started at 8 weeks gestation. Six of these nine USS-women had episodes of severe to mild thrombocytopenia during childhood that had been incorrectly diagnosed as ITP.

Because of the important association between USS and pregnancy we report the detailed natural history and clinical characterization of these nine women as well as their ADAMTS13 mutations. Our experience emphasizes the importance of measuring ADAMTS13 activity as part of the evaluation of thrombocytopenia during childhood or pregnancy.

Patients, materials and methods

Patients with USS

Between July 1998 and April 2008, 37 patients with USS (24 females and 13 males) belonging to 32 different families were identified at Nara Medical University. Severe neonatal jaundice that required exchange blood transfusion, a hallmark of USS, was found in 16 (43%) of these 37 patients. Notably, nine USS patients belonging to 6 families (designated as K, L, M, O, R and Z) had been pregnant and they all developed thrombocytopenia, often followed by TTP, at the second or third trimester. Thus, these 9 patients were extensively studied here. None of these nine women had received contraceptive agents before their pregnancies. It is also important to note that the remaining 15 USS women had no pregnancies. These data will be published elsewhere.

Assays of ADAMTS13 activity and ADAMTS13 inhibitor

Plasma levels of ADAMTS13 activity and ADAMTS13 inhibitor were measured by both a classic VWF multimer assay (Furlan et al, 1998) and a chromogenic ADAMTS13-activity-enzyme-linked immunosorbent assay (ADAMTS13 act-ELISA) (Kainos Inc., Tokyo) (Kato et al, 2006). The ADAMTS13 activity of pooled normal plasma was defined as 100%, and the detection limits of classic VWF multimer assay and the act-ELISA were 3% and 0·5% of the normal controls respectively. The ADAMTS13 inhibitor titres were expressed in Bethesda units (BU), where one inhibitor unit was defined as the amount necessary to reduce ADAMTS13 activity to 50% of the control level. A titre of ≥0·5 BU/ml in both the assays was considered to be significant.

Assays of ADAMTS13 antigen

Plasma levels of ADAMTS13 antigen were determined by a quantitative sandwich antigen ELISA (ag-ELISA) by using two anti-ADAMTS13 monoclonal antibodies (mAbs); a neutralizing mAb A10 as the coating antibody and a non-neutralizing mAb C7 labeled with horseradish peroxidase (HRP) as the detection antibody (Yagi et al, 2007). The epitopes recognized by mAbs A10 and C7 were determined to reside in the disintegrin-like domain and the 7th and 8th thrombospondin type-1 domains respectively (Uemura et al, 2005). The ADAMTS13 antigen of pooled normal plasma was defined as 100%, and the detection limit was 0·1% of the normal controls.

Further, ADAMTS13 antigen was analysed by Western blot under reducing conditions, quantitatively and qualitatively (Ishizashi et al, 2007). Briefly, this was performed as the follows. An aliquot (2 μl) of diluted or undiluted plasma sample per lane were separated by sodium dodecyl sulfate-5% polyacrylamide gel electrophoresis (SDS-5%PAGE) under reducing conditions. After electrophoresis, the proteins were electrophoretically blotted onto polyvinylidene difluoride (PVDF) microporous membranes. We probed the blots for ADAMTS13 antigen with WH2-11-1 as the primary mAb, followed by secondary staining with HRP-conjugated goat anti-mouse IgG (Kirkegaard & Perry Lab, Gaithersburg, MO, USA). The epitope of mAb WH2-11-1 resides in the fourth thrombospondin type-1 domain (Soejima et al, 2006). After incubation with Western Lighting Chemiluminescence Reagent (PerkinElmer Life Sciences, Shelton, CT), the blots were exposed to X-ray film. Densitomeric analysis of ADAMTS13 antigen was performed for the 190-kD band using the National Institutes of Health (NIH) image J (developed by NIH, http://rsb.Info.gov/noh-image/). The detection limit of plasma ADAMTS13 antigen by this method was 3% of the normal controls.

ADAMTS13 gene analysis

All DNA analyses were performed with the permission of Ethics Committees of both the sample-collecting hospitals and the institute that performed the gene analysis. Written informed consent was obtained from all patients. The nucleotide sequences of the entire 29 exons of ADAMTS13, including the intron–exon boundaries, were determined by direct sequencing of polymerase chain reaction (PCR) products, as previously described in detail (Kokame et al, 2002; Matsumoto et al, 2004). All the disease-causing ADAMTS13 mutations reported here were excluded as common polymorphisms by screening 96 individuals in the Japanese general population.

Results

The natural history, clinical characterization, and ADAMTS13 activity of the 9 women with USS and their family members are summarized in Table I. Detailed descriptions of these patients are provided in the text.

Table I.   Clinical characterization of 9 female patients with Upshaw–Schulman Syndrome who developed TTP-bouts during pregnancy.
NumberUSS-patientsYear of birthNewborn periodChildhoodAdulthood (pregnancy)Year of diagnosis for USS by ADAMTS13 assay
BeforeAfterTherapy and outcome
Exchange blood transfusion for severe jaundiceClinical diagnosis (age)Clinical diagnosisAge (years)Gestation (weeks)TherapyGestation (weeks)DiagnosisBabiesMothers
  1. USS, Upshaw–Schulman Syndrome; ITP, idiopathic thrombocytopenic purpura; TTP, thrombotic thrombocytopenic purpura; HUS, hemolytic uremic syndrome; APS, anti-phospholipid syndrome; ST, steroid; FFP, fresh frozen plasma; IVIg, high dose immunoglobulin G; HELLP, haemolysis, elevated liver enzyme and low platelet count; CS, Caesarean section; IUFD, interuterine fetal death.

1K-31976NoITP (6 years)ITP2724ST + IVIg25TTPIUFD at 25 weeks gestationPE (2 times), CS, necrosis of placenta and uterus, hysterectomy, remission2003
2K-41978YesITP (4 years)ITP2520FFP infusions29ITPLive-birth at 29 weeks gestation, premature babyCS under FFP infusions, remission2003
3L-21967NoNoneNone2727Conservative28HELLP syndrome or APSIUFD at 27 weeks gestationCS, remission
APS283rd trim.Aspirin37APSLive-birth at 37 weeks gestation under aspirin intake, Premature babyCS, remission
APS303rd trim.Aspirin32APSLive-birth at 37 weeks gestation under aspirin intake, Premature babyCS, remission
APS333rd trim.Aspirin32APSLive-birth at 32 weeks gestation under aspirin intake, Premature twinsCS, remission2003
4L-31972NoITP (3 years)ITP2524ST25Evans syndromeStillbirth at 25 weeks gestationST, remission after vaginal delivery
ITP2723PE, Haemodialysis24HUSLive-birth of premature baby at 24 weeks gestation, but it died soon after delivery.PE (2 times) and haemodialysis, Remission after vaginal delivery2003
5M-31969NoNoneNone3317PE20TTPMiscarriage at 20 weeks gestationPE (2 times), remission2003
6M-41971NoNoneNone3028CS HELLP syndromeStillbirth at 28 weeks gestationCS, remission
 3133Conservative36ITPLive-birth at 36 weeks gestation CS, remission2003
7O-41958NoITP (5 years)ITP2623Conservative25ITPLive-birth of premature baby at 25 weeks gestation, but it died soon after birthTTP-bout after vaginal delivery, PE (2 times), remission
Chronic relapsing TTP318FFP infusions39Chronic relapsing TTPLive-birth weighed 2·984 kg at 39 weeks gestationNatural delivery2004
8R-51982NoITP (8 months)ITP2323Conservative30TTP (after delivery)IUFD at 30 weeks gestationTTP-bout after CS, PE (8 times) and ST, remission2005
9Z-31971NoITP (7 years)ITP2512Conservative32TTP (after CS)Live-birth at 32 weeks gestation, Premature babyTTP-bout after CS, PE (times unclear), remission1998

Family K, clinical data

Proposita K-3, born in 1976, was the older of 2 siblings born to non-consanguineous parents (Fig 1, top). She had no history of severe neonatal jaundice. At the age of 6 years, she developed nasal bleeding with thrombocytopenia and was diagnosed with ITP. She then had repeated episodes of thrombocytopenia during childhood. Her first pregnancy was at the age of 27 years. The initial platelet count was normal (180 × 109/l), but at 24 weeks gestation the platelet count decreased to 17 × 109/l. She received steroid therapy and then high-dose immunoglobulin G (IVIg) infusions in a local hospital without clinical improvement. She subsequently developed microangiopathic haemolytic anaemia with neurological and renal abnormalities, was diagnosed with TTP, and treated with plasma exchange. Her fetus died at 25 weeks gestation and was removed by Caesarean section; necrotic lesions and intravascular thrombi were identified in a large area of the placenta and also the uterus, necessitating hysterectomy. Following hysterectomy and plasma exchange (twice during 2 d), a normal platelet count (≥150 × 109/l) was achieved, but a few weeks after cessation of plasma exchange the platelet count dropped again requiring an additional plasma exchange treatment. Following recovery, her plasma ADAMTS13 activity and ADAMTS13 inhibitor were assayed by the VWF multimer method, demonstrating a severe deficiency of ADAMTS13 activity but without an inhibitor, suggesting a diagnosis of USS.

Figure 1.

 Pedigree and ADAMTS13 analyses of USS-Family K and L. Squares and circles indicate males and females, respectively, and arrows with P indicate the proposita. Filled symbols represent USS-patients. The half-filled symbols represent asymptomatic carriers. The cross indicates deceased. The ADAMTS13 activity by both act-ELISA and VWF multimers are shown as a percentage of the normal control. The ADAMTS13 antigen by ag-ELISA is shown as a percentage of the normal control. Mutations found in ADAMTS13 are shown as one-letter amino acid abbreviations.

At the same time, her younger sister (K-4), born in 1978 (age 25 years), was a primigravida at 22 weeks gestation, and she also had mild thrombocytopenia (59 × 109/l) and anaemia (Hb 91 g/l). She had a history of severe neonatal jaundice that required exchange blood transfusion. At the age of 4 years, she had an episode of severe thrombocytopenia with bleeding (site unknown) that required blood transfusion, and was diagnosed as ITP. Because of her elder sister’s diagnosis of USS, this patient was tested for ADAMTS13 activity and ADAMTS13 inhibitor, demonstrating the same results as her elder sister. Thus, prophylactic FFP infusions (320 ml every 2 weeks) were instituted to maintain pregnancy. Under these circumstances, at 29 weeks gestation the patient had a premature boy by Caesarean section.

After the diagnosis of USS, both the siblings had receiving 320 ml of FFP infusions every 3 weeks. In 2005, the plasma levels of ADAMTS13 activity in both K-3 and K-4 were examined by the act-ELISA, and indeed they were shown to be severely deficient (<0·5% of the normal) but without its inhibitors (<0·5 BU/ml).

ADAMTS13 analysis (heterozygous Y304C/G525D mutation)

The ADAMTS13 mutations of both the affected siblings (K-3 and K-4) were shown to be the compound heterozygotes of Y304C (c. 911A > G, exon 8) and G525D (c. 1574G > A, exon 13), and their parents were heterozygous carriers of either of the mutations. Five common single nucleotide polymorphisms (SNPs) found in the two affected siblings are listed in Table II.

Table II.   Disease-causing mutations and common SNPs of ADAMTS13 gene in 9 patients with Upshaw–Schulman syndrome.
Patients with USSExon/ IntronNucleotidesAmino acidRemarks
K-3
K-4
exon 8911 A > GY304Cdisease causing mutation
exon 91016 C > GT339Rcommon SNP
exon 121342 C > GQ448Ecommon SNP
exon 131574 G > AG525Ddisease causing mutation
exon 151716 G > Asilentcommon SNP
exon 161852 C > GP618Acommon SNP
exon 192280 C > Tsilentcommon SNP
L-2
L-3
exon 43712 ins GTR125VfsX6disease causing mutation
exon 91016 C > GT339Rcommon SNP
exon 121342 C > GQ448Ecommon SNP
exon 151716 G > Asilentcommon SNP
exon 161852 C > GP618Acommon SNP
exon 192280 C > Tsilentcommon SNP
exon 283904 C > TQ1302Xdisease causing mutation
M-3
M-4
exon 6577 C > TR193Wdisease causing mutation
exon 91045 C > TR349Cdisease causing mutation
exon 192280 C > Tsilentcommon SNP
O-4exon 5533T > CI178Tdisease causing mutation
exon 192280 C > Tsilentcommon SNP
exon 222785 C > TQ929Xdisease causing mutation
R-5exon 6577 C > TR193Wdisease causing mutation
exon 91016 C > GT339Rcommon SNP
exon 121342 C > GQ448Ecommon SNP
exon 151716 G > Asilentcommon SNP
exon 161816 G > CA606Pdisease causing mutation
exon 161852 C > GP618Acommon SNP
exon 192280 C > Tsilentcommon SNP
Z-3exon 6577 C > TR193Wdisease causing mutation
exon 192280 C > Tsilentcommon SNP

Plasma levels of ADAMTS13 antigen by ag-ELISA were 0·4% of normal in both the affected siblings, and 31% and 36% in their father and mother respectively. Further, plasma levels of the respective ADAMTS13 antigen by Western blot were shown to be 5·1%, 4·8%, 39% and 34%. Thus, the Y304C/G525D mutation was apparently secreted into the plasma, but with a markedly reduced level (Fig 1, top right and Fig 2).

Figure 2.

 Western blot analyses of ADAMTS13 antigen in plasma milieu of members in USS-Family K∼Z. Plasma ADAMTS13 antigen was qualitatively and quantitatively analyzed by Western blot, as described in Materials and methods. Briefly, each 2 μl of diluted or undiluted plasma sample per lane were separated by SDS-5%PAGE under reducing conditions. After electrophoresis, the proteins were blotted onto polyvinylidene difluoride (PVDF) membranes. The ADAMTS13 antigen was incubated with WH2-11-1 as the primary mAb, followed by secondary staining with HRP-conjugated goat anti-mouse IgG, and finally detected by luminography. Densitomeric analyses of ADAMTS13 antigen were performed for the 190-kD band using NIH image J.

Family L, clinical data

Proposita L-2, born in 1967, was the second of five siblings to non-consanguineous parents. Of her five siblings, the first child died at 20 weeks gestation with a diagnosis of intrauterine fetal death of unknown aetiology, and two brothers (the third and fourth siblings) were apparently healthy. She had no history of severe neonatal jaundice or childhood thrombocytopenia. She was a primigravida at the age of 27 years, and her platelet count was 180 × 109/l at 6 weeks gestation, which dropped to 38 × 109/l at 27 weeks, complicated with acute renal failure, followed by intrauterine fetal death. She was suspected to have HELLP (Hemolysis, Elevated Liver enzyme, and Low Platelet count) syndrome or anti-phospholipid antibody syndrome. Interestingly, soon after Caesarean section her platelet count returned to normal. Because of this episode, her physician recommended that she took a low-dose aspirin (81 mg/d) whenever she had thrombocytopenia in association with pregnancy. At the age of 28 years, she had her second pregnancy, and again developed thrombocytopenia (21 × 109/l) and proteinuria during the third trimester, and low-dose aspirin was administered. She delivered by Caesarean section at 37 weeks gestation. After delivery, her platelet count soon returned to normal (150 × 109/l). At the age of 30 years, she had her third pregnancy and was treated similarly as above, and was delivered by Caesarean section at 37 weeks gestation. Finally, at the age of 33 years, she had her fourth pregnancy with twins. Even with the administration of low-dose aspirin, she developed haemolytic anaemia with thrombocytopenia (53 × 109/l) at 30 weeks gestation, which was suspected to be Coombs’-negative Evans’ syndrome and therefore she was treated with oral steroids and IVIg therapy. Her platelet count did not increase, she developed generalized oedema and pleural effusions at 32 weeks gestation, and her premature twins were delivered by Caesarean section. One week after delivery, her platelet count spontaneously increased to 200 × 109/l. Since then, she has never taken steroid or aspirin.

The younger sister (L-3), born in 1972, had no history of severe neonatal jaundice. At the age of 3 years, she had mild thrombocytopenia (70 × 109/l) at the time of a respiratory infection, and was diagnosed with ITP. At the age of 25 years, she was a primigravida, and at 24 weeks gestation developed proteinuria and thrombocytopenia (20 × 109/l) that was suspected to be Evans’ syndrome and was treated with steroids. Then she had a stillbirth at 25 weeks gestation. After delivery, her platelet count soon recovered to normal. She had her second pregnancy when aged 27 years, and at 23 weeks gestation she developed thrombocytopenia (14 × 109/l) and gross hematuria accompanied by renal failure. She was suspected to have haemolytic-uremic syndrome (HUS) and was treated with plasma exchange for 2 d (once per day) and haemodialysis. Despite this intensive care, she showed no clinical improvement, and at 24 weeks gestation she delivered a premature baby transvaginally but it died soon after birth. Her platelet count recovered within 1 week after delivery. These two sisters were referred to Nara Medical University in 2003, and plasma levels of ADAMTS13 activity and ADAMTS13 inhibitor were assayed by the VWF multimer method, demonstrating a severe deficiency of ADAMTS13 activity but without an inhibitor, consistent with a diagnosis of USS. In 2005, the act-ELISA revealed that the plasma levels of ADAMTS13 activity in both L-2 and L-3 were severely deficient (<0·5% of the normal control), but without an inhibitors (<0·5 BU/ml).

ADAMTS13 analysis (heterozygous R125VfsX6/Q1302X mutation)

The ADAMTS13 mutations were the compound heterozygotes of R125VfsX6 (c.372insGT, exon 4) and Q1302X (c. 3904C > T, exon 28) in both the affected siblings (L-2 and L-3). Their parents were heterozygous carriers of either of the mutations. Seven common SNPs found in the two affected siblings are listed in Table II.

Plasma levels of ADAMTS13 antigen by the ag-ELISA were severely deficient in both patients, <0·1% of the normal, and 58% in their mother. The ADAMTS13 antigen analysis by Western blot confirmed these results: <3% of normal in both the patients, 79% in their mother, and no other bands with a molecular weight less than 190 kDa. Thus, neither mutant protein, R125VfsX6 or Q1302X, was present in plasma (Fig 1, bottom right and Fig 2).

Family M, clinical data

Proposita M-3, born in 1969, was the second of four siblings to non-consanguineous parents (Fig 3, top). She had no history of severe neonatal jaundice or thrombocytopenia during childhood. The 1st child in this family died of utero-placental thrombosis at 32 weeks gestation, and the 4th child (M-5) has never been pregnant and is apparently healthy. At the age of 33 years, M-3 was a primigravida and at 17 weeks gestation she developed petechiae and oedema in both the lower extremities, and was diagnosed as TTP at 19 weeks gestation. She was treated with plasma exchange twice for 2 d but eventually miscarried at 20 weeks gestation. Thereafter, her platelet count quickly recovered, but 1 month later her platelet count again dropped to 24 × 109/l. She received plasma infusion, which increased her platelet count to normal.

Figure 3.

 Pedigree and ADAMTS13 analyses of USS-Family M and O. Symbols are the same shown in Fig 1. Mutations found in ADAMTS13 are shown as one-letter amino acid abbreviations.

Her younger sister (M-4), born in 1971, also had no history of severe neonatal jaundice or thrombocytopenia during childhood. At the age of 30 years, she had her first pregnancy and at 28 weeks gestation she developed thrombocytopenia, hypertension, and proteinuria, suggesting a diagnosis of HELLP syndrome. A Caesarean section was performed, but the child was stillborn. The next year, she had her second pregnancy, and at 33 weeks gestation she developed thrombocytopenia, and Caesarean section was performed at 36 weeks gestation, delivering a live female infant.

Both the affected siblings (M-3 and M-4) in this family were diagnosed with USS in 2003 after determination of ADAMTS13 activity and ADAMTS13 inhibitor by VWF multimer assay. In 2005, the act-ELISA revealed that the plasma levels of ADAMTS13 activity in both M-3 and M-4 were indeed severely deficient (<0·5% of the normal), without inhibitors (<0·5 BU/ml).

ADAMTS13 analysis (heterozygous R193W/R349C mutation)

The ADAMTS13 mutations were compound heterozygotes of R193W (c. 577C > T, exon 6) and R349C (c. 1045C > T, exon 9) in both affected siblings (M-3 and M-4); their parents were heterozygous carriers of either of the mutations. One common SNP was found in these two affected siblings (Table II).

Plasma levels of ADAMTS13 antigen by the ag-ELISA were 8·6% and 7·7% of the normal in the patients, 57%, and 63% in their parents. Further, plasma levels of ADAMTS13 antigen by Western blot were 24% and 23% in two patients, and 45% and 43% in their parents. Thus, the R193W/R349C mutation is apparently secreted into plasma, but at a significantly reduced level (Figs 2 and 3, top right).

Family O, clinical data

The proposita 0–4, born in 1958, was the second child of two siblings to non-consanguineous parents (Fig 3, bottom). She had no history of severe neonatal jaundice or thrombocytopenia during childhood. At the age of 26 years, she had her first pregnancy, and at 23 weeks gestation she developed thrombocytopenia (20 × 109/l) and a diagnosis of ITP was made. At 25 weeks gestation she delivered a premature infant, but it died soon after birth. After delivery, she developed overt TTP, and received plasma exchange for 2 d (one each day). She was diagnosed with chronic relapsing TTP and received prophylactic plasma infusions (400 ml) every 2–4 weeks. At the age of 31 years, she had her second pregnancy and was managed with prophylactic plasma infusions every 1–2 weeks, but at 8 weeks gestation she developed proteinuria and thrombocytopenia (93 × 109/l), and therefore she received more frequent plasma infusions (fresh frozen plasma, about 10 ml/kg; 400–480 ml per week) beginning at this time. Under these circumstances, at 39 weeks gestation, she naturally delivered a mature baby who weighed 2·984 kg. The platelet count of this patient decreased to 73 × 109/l at 7 d after delivery and plasma was administered every 3–4 weeks. In this patient, plasma levels of ADAMTS13 activity and ADAMTS13 inhibitor were assayed in 2004 by the VWF multimer method, showing a severe deficiency of enzyme activity but without an inhibitor, suggesting a diagnosis of USS. In 2005, the act-ELISA revealed that the plasma level of ADAMTS13 activity in this patient was severely deficient (<0·5% of the normal), without an inhibitor (<0·5 BU/ml).

ADAMTS13 analysis (heterozygous I178T/Q929X mutation)

The ADAMTS13 mutation was the compound heterozygote of I178T (c. 533T > C, exon 5) and Q929X (c.2785C > T, exon 22) in the proposita (0–4), and her parents, elder brother, and 2nd child (girl) were heterozygote carriers of either of the mutations. One common SNP was found in the patient (Table II).

Plasma level of ADAMTS13 antigen by the ag-ELISA was <0·1% of the normal in the proposita, and ADAMTS13 antigen levels of her 4 family members were 44∼65%. These results were also confirmed by Western blot studies (Fig 2). Further, no other band with a molecular weight less than 190 kDa was found. Thus, neither the mutant protein of I178T nor Q929X was present in plasma (Fig 2).

Family R, clinical data

The proposita R-5, born in 1982, was the last child of 3 siblings to non-consanguineous parents (Fig 4, top). She had no history of severe neonatal jaundice. At the age of 8 months she developed generalized petechiae with thrombocytopenia (20 × 109/l) and was diagnosed with ITP. Until aged 5 years she had repeated episodes of purpura accompanied by fever that improved spontaneously within a few days. Her two elder siblings were asymptomatic with no history suggesting TTP. At the age of 23 years, she had her first pregnancy. At 23 weeks gestation, she developed mild thrombocytopenia (82 × 109/l), which further decreased to 47 × 109/l at 29 weeks gestation, and was diagnosed with ITP exacerbated by pregnancy. At 30 weeks gestation she had headache and nausea, and a week later she had sudden intrauterine fetal death. After Caesarean section, she developed overt TTP, treated with steroids and plasma exchange, which was performed eight times during the following 23 d. On this occasion in 2005, she was diagnosed with USS after determinations of ADAMTS13 activity (<0·5% of the normal) and ADAMTS13 inhibitor (<0·5 BU/ml) by the act-ELISA. After the diagnosis was confirmed, she has received prophylactic plasma infusions (320 ml every 2 weeks).

Figure 4.

 Pedigree and ADAMTS13 analyses of USS-Family R and Z. Symbols are the same shown in Fig 1. Mutations found in ADAMTS13 are shown as one-letter amino acid abbreviations.

ADAMTS13 analysis (heterozygous R193W/A606P mutation)

The ADAMTS13 mutation was the compound heterozygote of R193W (exon 6) and A606P (c.1816G>C, exon 16) in the proposita (R-5), and her parents and elder sister were both heterozygote carriers of either of the mutations. Five common SNPs were found in this patient (Table II).

Plasma level of ADAMTS13 antigen by the ag-ELISA was 1·6% of normal in the proposita; the levels in her three family members were 25∼65%. These results were also confirmed by Western blot studies (Fig 2). Thus, the R193W/A606P mutant protein is secreted into plasma, but at a markedly reduced level (Fig 2).

Family Z, clinical data

The proposita Z-3, born in 1971, was the last of three siblings of consanguineous parents; the patient’s grandparents were cousins (Fig 4, bottom). She had had neonatal jaundice of moderate severity. She was diagnosed of ITP at the age of 7 years, and had been treated with steroids until 15 years of age. She had her first pregnancy at age 25 years, and at 12 weeks gestation she developed thrombocytopenia (15 × 109/l) and was diagnosed with pregnancy-associated ITP. At 32 weeks gestation, she had a live birth by Caesarean section, after which she developed overt TTP, which was treated effectively with daily plasma exchange. This patient was referred to our laboratory in 1998, and plasma levels of ADAMTS13 activity and ADAMTS13 inhibitor were assayed by the VWF multimer method, demonstrating a severe deficiency of enzyme activity but without an inhibitor, suggesting a diagnosis of USS. Prophylactic plasma infusions were not initiated in this patient, and she had more than 5 episodes of TTP during 1998–2005. Each episode was treated with a single plasma infusion (320 ml). In 2006, when she had TTP triggered by a respiratory infection, infusion of plasma induced a severe anaphylactic reaction requiring steroid treatment. The act-ELISA confirmed her plasma level of ADAMTS13 activity was severely deficient (<0·5% of the normal) and ADAMTS13 inhibitor was negative (<0·5 BU/ml).

ADAMTS13 analysis (homozygous R193W mutation)

The ADAMTS13 mutation was homozygous R193W (exon 6) in the proposita (Z-3), and her parents and child were heterozygote carriers of this mutation. One common SNP was found in this patient (Table II).

The plasma level of ADAMTS13 antigen by the ag-ELISA was 11% of normal in the proposita; the levels in her three family members were 17∼83%. These results were also confirmed by Western blot studies (Fig 2). Thus, the R193W mutation is apparently secreted into plasma, but at a reduced level.

Discussion

Coombs’-negative haemolysis with neonatal jaundice requiring exchange blood transfusions and thrombocytopenia have been considered to be a clinical hallmark of USS (Fujimura et al, 2002). However, only 16 (43%) of our 37 USS patients had episodes of exchange blood transfusion for severe jaundice in this period. Since USS is an autosomal recessive order, the ratio of female-to-male patients should be 1-to-1, but there was female predominance (24-to-13) among our 37 patients. The reason for this apparent gender disparity is probably because many USS patients have a mild disorder and thrombocytopenia is misdiagnosed as ITP. Physicians may overlook this important sign, as in fact 29 (78%) of our 37 USS patients had a history of thrombocytopenia; the remaining 8 were unclear as to whether their platelet count had been checked during their childhood.

Episodes of pregnancy and infections are well known precipitating factors for overt TTP in USS patients, but it is not well understood how often and why TTP episodes are induced by these factors (Weiner, 1987; Vesely et al, 2004; Terrell et al, 2005; Scully et al, 2006). In contrast, a chemical compound, DDAVP (1-deamino-8-D-arginine vasopressin), is well documented to induce TTP in USS patients (Hara et al, 1986; Veyradier et al, 2006), presumably through a rapid increase of plasma VWF by releasing ultra large(UL)-VWF multimers from vascular endothelial cells, leading to platelet thrombi formation under high shear stress (Yagi et al, 2001).

This study has clearly shown that pregnancy consistently induces thrombocytopenia during the second–third trimester, often followed by overt TTP. In normal pregnancies, plasma VWF increases steadily during gestation and reaches a maximum level (200∼500% of the normal control) at term with appearance of UL-VWF multimers. The elevated plasma levels of VWF return to normal within one week after delivery (Stirling et al, 1984). Thus, the plasma milieu during the third trimester and at delivery appears to be similar to that seen after DDAVP administration. Thus, it is conceivable that the rapidly increased plasma level of UL-VWF multimers plays a critical role in precipitating overt TTP in pregnant women with USS.

More interestingly, however, the onset of clinically overt TTP observed in our USS patients during pregnancy is in contrast to mice with ADAMTS13 knock-out (Motto et al, 2005; Banno et al, 2006). As previously shown, the ADAMTS13 knock-out mice have UL-VWF multimers in their plasma and were potentially thrombogenic in in vitro studies using a parallel plate flow-chamber system under high shear stress. However, no thrombotic complications were found in the mice after pregnancy. We assume that it is important to note that haemodynamics formed by anatomically different vasculature networks in these two species significantly differs.

It has been proposed that there may be clinically distinct phenotypes of USS, with early-onset and late-onset types (Furlan & Lämmle, 2001; Camilleri et al, 2008). Our experience supports this hypothesis. In fact, three out of the nine women who were diagnosed with USS during pregnancy appeared to lack any clinical signs of TTP during their childhood. The distinction of USS patients into two phenotypes by their plasma levels of ADAMTS13 activity is less likely, because all nine of our USS patients had undetectable ADAMTS13 activity (<0·5% of the normal control) by the sensitive act-ELISA. In the nine women, eight of the ADAMTS13 mutations were compound heterozygotes of Y304C/G525D (2 siblings), R125VfsX6/Q1302X (2 siblings), R193W/R349C (2 siblings), I178T/Q929X, and R193W/A6060P, and one was homozygous for R193W/R193W. Except for R193W, these mutations were all novel, and distributed from exons 4–28. The location of the ADAMTS13 mutations in our nine USS patients, seven in the metalloprotease domain and the remaining two on the disintegrin/cysteine-rich domains, may be relevant to the clinical presentation. Camilleri et al (2008) recently reported that prevalence of R1060W missense mutation at the seventh thrombospondin-1 domain was associated with late-onset adult TTP, but we have not found the R1060W mutation in our 37 Japanese patients with USS (unpublished). Thus, presently there is no clear evidence dividing USS into two phenotypes on a basis of ADAMTS13 activity levels or ADAMTS13 mutations. Our observations below also support this concept; two siblings of family M (M-3 and M-4) were the late-onset type, whereas other two siblings of family L (L-2 andL-3) shared two different onset types.

Veyradier et al (2003) described that prophylactic plasma infusions (fresh frozen plasma at a single dose of 10 ml/kg) every 3–4 weeks were efficiently performed to prevent periods of TTP in their USS-patients, but more plasma infusions (10 ml/kg during 2–5 consecutive days) were required during each relapse as early as possible. However, a standard therapeutic protocol for plasma infusions for pregnant women with USS has not been established. Our experience here clearly demonstrates that women with USS during pregnancy expose their infants to high risk without prophylactic plasma infusions. The risks for fetal loss are presumably caused by the disturbance of utero-placental circulation with platelet thrombi formed, as shown in Patient K-3 with extensive necrotic lesions in both the placenta and uterus. Among 15 pregnancies (one twin pregnancy), eight infants were stillborn or died soon after birth, and the remaining eight survived but they were premature except for one mature baby (O-5 in Fig 3). Further, seven of eight surviving infants were born by Caesarean section; one with prophylactic plasma infusions, two without plasma infusions, and four with aspirin. Notably, one mature baby (O-5) was born at 39 weeks gestation by natural delivery following plasma infusions to the mother that had started in the very early stage of pregnancy (8 weeks gestation). This experience clearly indicated that prophylactic plasma infusions (fresh frozen plasma, about 10 ml/kg; 400–480 ml per week) at this stage of pregnancy allowed an uneventful delivery. But the best protocol for plasma infusions to women with USS during pregnancy remains to be determined in future studies.

It is clear that deficient ADAMTS13 activity in maternal plasma is harmful to both the mother and fetus. Retrospectively, we suspect that the first two children from asymptomatic carriers of L-1 (Fig 1, bottom) and M-1/-2 (Fig 3, top) might have had USS, because the former died of intrauterine fetal death at 20 weeks gestation and the latter was a stillbirth at 32 weeks gestation. This speculation is based on the observation of UL-VWF multimers in normal newborn cord blood, despite of subnormal levels of ADAMTS13 activity (Johnson et al, 1981; Tsai et al, 2002).

Because of these clinical observations, here we strongly recommend that ADAMTS13 activity, and ADAMTS13 inhibitor should be measured in patients with thrombocytopenia during childhood and in association with pregnancy.

Conflict-of-interest disclosure

YF is a member of clinical advisory boards for Baxter BioScience.

Acknowledgements

This work was supported in part by research grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan and from the Ministry of Health, Labor, and Welfare of Japan.

Ancillary