Neonatal alloimmune thrombocytopenia in Norway: poor detection rate with nonscreening versus a general screening programme
Dr H Tiller, Department of Immunology and Transfusion Medicine, University Hospital of North Norway, N-9038 Tromsø, Norway. Email email@example.com
The implementation of an antenatal screening programme for neonatal alloimmune thrombocytopenia (NAIT) is currently under debate. We evaluated the detection rate for NAIT in a nonscreened population of 661 200 births where NAIT was diagnosed on clinical indication. We did a cross-sectional comparison with a population of 100 448 human platelet antigen 1a (HPA1a)-screened pregnancies from three of the five health regions in Norway. In a nonscreening situation, 7.5 cases of NAIT were detected per year compared with 53 cases when screening was applied. The detection rate of NAIT in Norway was therefore 14% of the expected rate.
Severe thrombocytopenia (platelet count <50 × 109/l) is seen in 0.12–0.14% of unselected newborns.1 The most frequent cause of isolated severe thrombocytopenia in term newborns is neonatal alloimmune thrombocytopenia (NAIT).1
NAIT is the result of maternal alloimmunisation against fetal platelet antigens. Maternal immunoglobulin G antibodies can cross the placenta, and fetal antibody-coated platelets will be removed from the circulation in the reticuloendothelial system.2
NAIT has various clinical courses. In most cases, the thrombocytopenia is transient and resolves within days without clinical symptoms.2 However, neonates with severe NAIT (platelets <50 × 109/l) may display unexpected bleeding in utero or within a few hours after birth. Intracranial haemorrhage (ICH) is the major cause of mortality and long-term morbidity and is reported in up to 26% of untreated NAIT cases.2
The immunodominant antigen for NAIT in the Caucasian population is the human platelet antigen 1a (HPA1a). Two percent of Caucasian women are HPA1a negative (HPA1bb) and are therefore at risk of being immunised. Interestingly, only 10% of HPA1bb pregnant women develop anti-HPA1a antibodies.3 Anti-HPA1a antibodies cause 85% of NAIT cases and almost all severe NAIT cases.1 The frequency of anti-HPA1a antibody-induced NAIT is reported to be 1:1100 live births.3
At present, there is no general agreement on how to handle pregnant women with anti-HPA1a antibodies. Treatment regimens have varied from in utero fetal blood sampling with subsequent platelet transfusion to more conservative approaches such as weekly intravenous gammaglobulin to the mother (with or without additional steroid therapy) and careful delivery by caesarean section.4 Due to high rate of mortality and morbidity, intrauterine approaches are now abandoned in most institutions.4
A recent Norwegian screening study including 100 448 pregnancies concluded that neonatal mortality and morbidity can be significantly reduced by antenatal screening for HPA1a-negative pregnancies followed by intervention.3 Furthermore, the study showed that it is possible to establish an antenatal screening programme for NAIT that is cost-effective.5 However, the implementation of a national screening programme for NAIT is still under debate.4
The aim of this study was to estimate the detection rate of anti-HPA1a antibody-induced NAIT in Norway in a nonscreened population compared with a screened population.
Materials and methods
The nonscreened population
All cases of suspected NAIT referred to our laboratory at the University Hospital of North Norway during the period 1994–2007 were studied. Since the laboratory is the National reference laboratory for clinical platelet immunology, the database contains all suspected cases of NAIT in Norway in the time period studied. Testing for NAIT was performed on clinical indication, and this population is therefore defined as the nonscreened population.
Maternal HPA status, antibody specificity and antibody quantitation were enrolled in a Microsoft Access database. If paternal and/or neonatal HPA genotyping was performed, this information was registered. Platelet count in the newborn was listed when known. Extended clinical information was written as a commentary when available.
The mean birth rate in Norway during the 13 years study period was 58 000 each year (Medical Birth Registry, Norway, annual reports 1994–2007). However, since a screening programme for NAIT was introduced in parts of Norway during the study period, the screened population was excluded from the nonscreened population. The total birth population of the screened population was 94 421 (see below) and equivalent to 1.6 years of births in Norway. When the screened population was excluded, births in the nonscreened population corresponded to an effective study period of 11.4 years or 661 200 births.
The screened population
The expected NAIT rate per live birth in Norway was calculated from a large and recent screening study involving HPA1 typing of 100 448 pregnancies in Norway.3 The screening study was undertaken in three out of the five health regions in Norway. Pregnancies from the northernmost region were included from 1995 to 2004, and pregnancies from the two southern health regions were screened from 2001 to 2004. HPA1 typing was carried out in all children except in two cases. In 153 of 173 immunised pregnancies, the fathers were also HPA1 genotyped. All HPA1a-negative women were tested for anti-HPA1a antibodies. Immunised women were delivered by careful caesarean section 2–4 weeks prior to term. By ‘careful’, we mean delivery of the baby’s head without rupturing of the membranes when possible and also making a liberate incision to gain good access when delivering the head. The rationale was that forceful manipulation when delivering the head might contribute to ICH if the neonate was severely thrombocytopenic. In addition, only experienced obstetricians performed these caesarean sections, and we always used two operating doctors. Furthermore, platelets from HPA1a-negative donors were prepared for immediate transfusion if the platelet count in the newborn was <35 × 109/l. None of the women were given intravenous gammaglobulin or steroids. Further details regarding the accomplishment of the screening study and results regarding morbidity and mortality from the study have already been published and are therefore only mentioned briefly in this study.3
To obtain the expected NAIT rate per live birth from the screening study, we estimated the equivalent number of births from 100 448 pregnancies screened. Termination of pregnancies was not considered because this patient group was not included in the screening study. Of a total of 2111 HPA1a-negative women detected in the screening study, 127 women (6%) dropped out of the study before delivery. Pregnant women who were typed HPA1a positive in the study were not followed up. We therefore chose a 6% drop-out rate for the total screening population, and the expected NAIT rate was estimated from a population of 94 421 births.
Definition of NAIT
When the mother had the platelet type HPA1bb, had detectable anti-HPA1a antibodies and her child was thrombocytopenic at birth (platelet count <150 × 109/l), the case was defined as NAIT. A case was defined as NAIT even if we did not have HPA status of child or father. Thrombocytopenic newborns from HPA1bb mothers without detectable antiplatelet antibodies were not diagnosed as NAIT. Severe NAIT was defined as neonatal platelet count <50 × 109/l. Other possible causes of thrombocytopenia, such as pre-eclampsia or prematurity, were excluded when making the diagnosis of NAIT.
HPA1 phenotyping was performed by flow cytometry or enzyme-linked immunosorbent assay (Dia-Med AG, Cressier, Switzerland) and genotyping by sequence specific primer-polymerase chain reaction (SSP-PCR) or real-time polymerase chain reaction with melting curve analysis.6 For identification of anti-HPA1a antibodies, platelets with known HPA phenotype were isolated and a modified monoclonal antibody immobilisation of platelet antigen (MAIPA) assay was performed.6 When MAIPA was positive for anti-HPA1a antibodies, quantitative MAIPA was performed.6
Data from the nonscreened database were exported to Microsoft Excel files for further statistical analysis using SPSS for Windows program. Ninety-five percent confidence intervals for detection rates were calculated from the formula: detection rate ± 1.96 SD, where (p = percent and n = number of cases).
The nonscreened population
Samples from 545 referrals for suspected NAIT were investigated in the period from 1994 to 2007. The majority (66%) of the investigations were related to neonatal thrombocytopenia in either the current or a previous child. The remaining referrals were due to other reasons, such as recurrent miscarriage, intrauterine fetal death or maternal thrombocytopenia (data not shown). The referral rate for each of the five health regions varied from 0.5 to 2.3 referred cases per 1000 live births (data not shown).
From 545 referrals, 94 cases of NAIT were diagnosed in total when all antibody specificities were included. The overall predictive value of diagnosing NAIT based on nonscreening was therefore 17%. When investigation was based on having a previous child with neonatal thrombocytopenia, the predictive value was higher (31%). The majority (67%) of NAIT cases were diagnosed postpartum based on clinical signs or symptoms.
Of 94 NAIT cases, 86 (91.5%) cases were due to anti-HPA1a antibodies. Among the 86 NAIT cases due to anti-HPA1a antibodies, 64 (74%) cases were severe.
The screened population
In the screening study, 85 cases of NAIT were diagnosed in a population of 94 421 births. This corresponds to a NAIT incidence of 1:1100 births, equivalent to 53 NAIT cases every year in Norway. Among all NAIT cases detected, 55 cases were severe (Table 1).
Table 1. Anti-HPA1a NAIT detection rate
NAIT detection rate
NAIT due to anti-HPA1a antibodies detected in the nonscreened population corresponded to 7.5 cases per year. When comparing 7.5 observed cases per year with 53 expected NAIT cases per year, the detection rate for NAIT in Norway without an implemented screening programme was 14% of the expected rate (95% CI 4.7–23.3%, Table 1).
Severe NAIT cases due to anti-HPA1a antibodies detected in the nonscreened population corresponded to 5.8 cases per year. The screening study equally detected 34 severe cases per year when adjusting for the total birth population. When calculating the NAIT detection rate using only severe NAIT cases, the detection rate was 16% (95% CI 3.7–28.3%, Table 1).
Sixty-four HPA1bb mothers in the nonscreened study group were investigated because they delivered a thrombocytopenic neonate (data not shown). In nine (14%) of these mothers, anti-HPA1a antibodies could not be detected, and in four (6%)cases, we did not detect any platelet reactive antibodies at all. Interestingly, all four HPA1bb women without detectable antibodies delivered neonates with severe thrombocytopenia (range platelet count 15–49 × 109/l, data not shown).
We have demonstrated a poor detection rate of NAIT in a nonscreened population in Norway. This phenomenon has been observed in other populations. A retrospective study was performed in the Irish population from 1992 to 2000, reporting a NAIT detection rate of 1:16 500 live births.1 However, our study is unique in having prospective data from a large screening study, instead of comparing with historical controls. The study is further strengthened by the fact that the nonscreened and screened population were studied during the same time period and within the same area.
Since the NAIT cases in the nonscreened population are mostly diagnosed because of clinical symptoms, thrombocytopenia will on average be more severe compared with NAIT cases found with a general screening in pregnancy. ICH is only a risk when thrombocytopenia is severe. In fact, the mean platelet count for all NAIT cases in the nonscreened population was 35 × 109/l (range 1–140, data not shown). It can therefore be argued that the detection rate would be higher and more clinically relevant if we adjust for severe NAIT only. However, the detection rate in severe NAIT cases was not significantly higher than in the whole group (Table 1).
We used a drop-out rate due to miscarriages of only 6%, which is a low estimate. In the screening and intervention study, twice as high drop-out rate (12.2%) was used. However, 12.2% is an estimated rate of miscarriages from week 6 in pregnancy and onwards.3 In the screening study, participants were enrolled at time of their first prenatal appointment, which usually takes place during week 8–12 in Norway (Antenatal care: routine for the healthy pregnant women, The Directorate for Health and Social Affairs, Norway 2005). To avoid bias towards poor NAIT detection rate, we used the lower drop-out rate.
Almost all (91%) the NAIT cases detected were due to anti-HPA1a antibodies. This is in accordance with previous studies of severe NAIT.1 Compared with the 2% reported frequency of HPA1bb in the Caucasian population,3 we found an excess (14%) of HPA1bb-positive mothers without detectable anti-HPA1a antibodies who delivered newborns with thrombocytopenia. In four cases of severe neonatal thrombocytopenia with noncompatible HPA1 antigens between mother and child, maternal antibodies could not be detected (range platelet count 15–49 × 109/l). There are at least two possible explanations: thrombocytopenia is not mediated by anti-HPA1a antibodies or HPA antibodies are present but cannot be detected by our methods. If the latter is true, we must consider our methods critically. It is a great challenge to explain why HPA1bb-positive mothers without detectable antibodies give birth to thrombocytopenic, otherwise healthy HPA1a-positive neonates. We suspect that the answer might give us new insight in the field of fetal–maternal immunology.
Three slightly different screening strategies have previously been evaluated to estimate the costs and cost-effectiveness of a screening and intervention programme.5 Depending on the screening strategy chosen, a screening and intervention programme was estimated to cost €620,000–910,000 per 100 000 pregnancies screened. Compared with no screening, the screening programmes would generate between 210 and 230 additional quality-adjusted life years among 100 000 children. Acknowledging the limitations of using nonrandomised controlled trial data as control group, all three screening and intervention strategies were found to be cost-effective.5
In the screening study, neonatal morbidity and mortality was reduced to approximately one-fourth compared with prospective historical controls.3 For the screened population, computed tomography caput was performed in all cases of severe neonatal thrombocytopenia. Similarly, cerebral ultrasound was part of the study protocol for the control group.3 The data from the nonscreened population used to evaluate detection rate were collected in retrospect, and complete data on haemorrhage in the neonates were not obtainable. However, the aim of this study was to evaluate the detection rate of NAIT. Our material adds valid data on detection rate of NAIT in a nonscreening situation to the existing data on clinical effect and cost-effectiveness of a screening and intervention programme.
In Norway, we would recommend a screening programme that identifies high-risk pregnancies. Briefly, this can be achieved by HPA1a typing during pregnancy in the samples already obtained for RhD typing. All HPA1a-negative pregnancies are followed up by analysing blood samples in gestational weeks 22 and 34 for detection of anti-HPA1a antibodies. An intervention programme consisting of delivery by careful caesarean section around 3 weeks prior to term, combined with compatible platelets available for immediate transfusion, should be offered only to women with anti-HPA1a antibody levels >3 IU/ml.6
Recent studies suggest that the immunisation pattern for NAIT is more similar to haemolytic disease of the newborn than previously thought. The possibility of preventing immunisation against HPA1a in a similar way as for the RhD immunisation has therefore been proposed.6 To prevent immunisation, we need to detect pregnancies at risk. This can only be performed by general screening for HPA1bb pregnancies.
The discrepancy between observed and expected cases of NAIT is huge. Without a screening programme, the detection rate of NAIT in Norway is only 14% of expected. Screening or not, our level of recognising NAIT in Norway today cannot be accepted. The ambition must be to detect all cases of NAIT.
Disclosure of interest
No conflict of interest.
Contribution to authorship
H.T. has participated in planning of the study. She was the main responsible for analysis and interpretation of the data, drafting and writing of the manuscript. M.K.K., B.S., P.Ø. and A.H. were involved in planning of the study. B.S. and A.H. were accountable for conception of the study. All authors revised the article critically for important intellectual content and participated in the final approval of the version to be published.
Details of ethics approval
The study was approved by the Regional Committee for Medical Research Ethics, North Norway, 9 June 2005 with reference number 42/2005.
The study was funded by a grant from the North Norwegian Health Authorities.
The authors would like to thank Prof Ganesh Acharya at the University Hospital of North Norway for contribution in planning of the study.