Authors: Siw Leiknes Ernstsen (Norwegian National Unit for Platelet Immunology, University Hospital of North Norway, Department of Laboratory Medicine, Division of Diagnostics, University Hospital of North Norway, Tromsø, Norway), Maria Therese Ahlen (Norwegian National Unit for Platelet Immunology, University Hospital of North Norway, Department of Laboratory Medicine, Division of Diagnostics, University Hospital of North Norway, Tromsø, Norway; Immunology Research Group, Department of Medical Biology, UiT The Arctic University of Norway, Tromsø, Norway), Eirin Listau Bertelsen (Immunology Research Group, Department of Medical Biology, UiT The Arctic University of Norway, Tromsø, Norway), Jens Kjeldsen‐Kragh (Norwegian National Unit for Platelet Immunology, University Hospital of North Norway, Department of Laboratory Medicine, Division of Diagnostics, University Hospital of North Norway, Tromsø, Norway), Anne Husebekk (Immunology Research Group, Department of Medical Biology, UiT The Arctic University of Norway, Tromsø, Norway), Heidi Tiller (Women's Health and Perinatology Research Group, Department of Clinical Medicine, UiT The Arctic University of Norway, Tromsø, Norway; Department of Obstetrics and Gynecology, University Hospital of North Norway, Tromsø, Norway)
Categories: Fetal Medicine, alloimmunization, antenatal management, immunoglobulins, intracranial hemorrhage, newborn, thrombocytopenia
Source: Acta Obstetricia et Gynecologica Scandinavica
Doi: 10.1111/aogs.70031
Improving prediction of fetal/neonatal intracranial hemorrhage in HPA‐1a alloimmunized pregnancies to guide antenatal An observational cohort study
Authors: Siw Leiknes Ernstsen, Maria Therese Ahlen, Eirin Listau Bertelsen, Jens Kjeldsen‐Kragh, Anne Husebekk, Heidi Tiller
Abstract
Introduction
Maternal alloimmunization against human platelet antigen‐1a (HPA‐1a) may lead to severe intracranial hemorrhage (ICH) in the fetus or newborn as a life‐threatening complication of fetal neonatal alloimmune thrombocytopenia (FNAIT). Most women who are HPA‐1a‐alloimmunized do not have a fetus/neonate with ICH. In the absence of predictive tools to identify pregnancies with high risk of ICH outcome, most countries offer weekly antenatal IVIg to all recognized HPA‐1a‐alloimmunized pregnancies. Norwegian FNAIT guidelines are restrictive regarding antenatal IVIg administration and have a long‐standing tradition of longitudinal anti‐HPA‐1a antibody measurements when at‐risk pregnancies are identified, facilitating exploration of the natural history of alloimmunized pregnancies. We aimed to explore associations between maternal anti‐HPA‐1a antibody levels and risk of fetal/neonatal ICH in non‐IVIg treated HPA‐1a alloimmunized pregnancies and assess if an antibody level threshold can be useful for identifying pregnancies with increased ICH risk.
Material and Methods
We compared anti‐HPA‐1a levels both from clinically referred and prospectively identified, non‐IVIg treated, HPA‐1a‐immunized pregnancies stratified by previous neonatal FNAIT outcome (ICH or FNAIT without ICH) in Norway 1997–2023.
Results
Anti‐HPA‐1a levels in pregnancies with ICH outcome were higher (median 29.6 IU/mL, range 0.1–222.1, n = 15) compared to no ICH FNAIT pregnancies (median 10.4 IU/mL, range 0.0–83.1, n = 55; p = 0.046, Mann–Whitney U test). A suggestive anti‐HPA‐1a threshold of 70 IU/mL was chosen based on receiver operating characteristic (ROC) analysis, with high specificity values (96.4%).
Conclusion
Antenatal anti‐HPA‐1a levels may be useful when assessing the risk of ICH outcome and may enable a more targeted antenatal IVIg treatment both in a nonscreening and screening situation.
Key messagePrediction of severe FNAIT is currently limited. Pregnancies complicated by FNAIT‐related ICH were found to have significantly higher anti‐HPA‐1a levels compared to those without ICH outcome. This opens the door for targeted antenatal treatment to avoid both under‐ and overtreatment.
INTRODUCTION
Fetal neonatal alloimmune thrombocytopenia (FNAIT) is the leading cause of severe intracranial hemorrhage (ICH) in term‐born neonates ^1^ and is associated with severe morbidity and mortality. ^2^ , ^3^ FNAIT is caused by maternal alloantibodies against human platelet antigens (HPAs) targeting paternally inherited HPA antigens on fetal platelets. In whites, 80% of FNAIT cases are caused by anti‐HPA‐1a antibodies. Although considered a rare disease, FNAIT occurs in 1000 pregnancies, ^4^ with ICH observed in 10 000 pregnancies. ^4^ , ^5^
No country has yet implemented antenatal screening to detect pregnancies at risk for FNAIT and no prophylaxis is currently available. Unlike the red cell counterpart, hemolytic disease of the fetus and newborn (HDFN), severe FNAIT outcome may also occur already during the first alloimmunized pregnancy. ^2^ , ^3^ Consequently, FNAIT is not only underdiagnosed, ^6^ but prevention is limited to subsequent pregnancies—provided the condition was recognized in a prior sibling.
The only known and acknowledged factor for increased risk of ICH due to FNAIT is a previous sibling with an ICH complication. The recurrence risk of ICH is high, ranging from 29% to 79%. ^7^ , ^8^ In contrast, if the previous child had FNAIT, but without severe bleeding complications, the risk of ICH is substantially lower. ^7^ , ^9^ Consequently, HPA‐1a alloimmunized pregnancies are typically risk‐stratified depending on prior obstetric FNAIT history— with a prior ICH‐affected fetus/newborn defined as high risk and a prior FNAIT history but no ICH defined as standard risk. ^10^
Nearly 40 years ago, it was reported that giving antenatal intravenous immunoglobulin G (IVIg) to HPA‐1a alloimmunized women increased platelet counts in neonates affected by FNAIT compared to their older siblings affected by the same condition. ^11^ Also, antenatal IVIg treatment to mothers who previously had a fetus/newborn with anti‐HPA‐1a induced ICH has been reported to reduce ICH recurrence risk, ^2^ and there is wide international support for IVIg treatment in such high‐risk pregnancies. ^10^ Current management strategies in most Western countries routinely recommend off‐label, weekly high‐dose IVIg to all HPA‐1a alloimmunized women, regardless of prior FNAIT history, ^10^ including HPA‐1a alloimmunized pregnant women without obstetric FNAIT history. In contrast, the Norwegian guidelines only recommend antenatal IVIg treatment to women in the high‐risk group. We recently evaluated our practice and found it to be safe, ^7^ and moreover, it seems that treatment of all HPA‐1a alloimmunized pregnancies with IVIg may not be necessary. ^7^ , ^12^ In a future FNAIT screening setting, the majority of identified alloimmunized women will have no prior obstetric history to guide ICH risk prediction. To reduce the risk of fetal/neonatal ICH, it would be helpful to have tools to identify the few pregnancies among standard risk and screened pregnancies which could benefit from IVIg treatment, ^7^ , ^13^ but no such tools are currently available.
Maternal anti‐HPA‐1a antibody levels have shown good positive and negative predictive values for severe neonatal thrombocytopenia in HPA‐1a alloimmunized pregnancies. ^14^ , ^15^ Serial measurements of maternal anti‐HPA‐1a antibody level during pregnancy have been part of the National clinical guidelines in Norway the last 15 years to guide ante‐ and perinatal management. Also, in France, sequential maternal anti‐HPA‐1a quantifications are used for risk prediction and antenatal therapy guidance in pregnancies at risk of FNAIT. ^16^ Recently, the Netherlands revised their clinical guidelines to include anti‐HPA‐1a antibody levels to assist in antenatal management considerations for HPA‐1a alloimmunized pregnancies at standard risk of FNAIT (Verweij EJT, Leiden University Medical Center, personal communication) and they have also explored if inclusion of antibody levels as a tool to guide management in a future an HPA‐1a screening program would be cost‐effective. ^17^
Although high levels of maternal anti‐HPA‐1a antibodies have been reported in ICH cases, ^2^ , ^13^ , ^18^ , ^19^ , ^20^ antibody levels have not yet been considered useful in ICH risk prediction. ^18^ , ^20^ The primary aim of this study was to investigate the association between maternal anti‐HPA‐1a antibody levels and ICH in the fetus/newborn in Norwegian FNAIT cohorts. Second, we aimed to identify an antibody level that may be useful in predelivery risk stratification and clinical management decisions.
MATERIAL AND METHODS
Study design and participants
The study was designed as a retrospective comparison of anti‐HPA‐1a levels in non‐IVIg treated pregnancies with different FNAIT outcomes.
HPA‐1a alloimmunized pregnancies were identified from two pregnancies referred for clinical investigation at the Norwegian National Unit for Platelet Immunology (NNUPI) at the University Hospital of North Norway, and pregnancies identified through a previous Norwegian HPA‐1a screening study (1995–2004). ^21^ NNUPI serves to investigate and diagnose all cases of suspected/confirmed FNAIT in Norway. All clinically referred HPA‐1a alloimmunized pregnancies to NNUPI from 1997 to 2023 with FNAIT were included and stratified on ICH outcome or not. From the former Norwegian screening study, HPA‐1a‐alloimmunized, HPA‐1 incompatible pregnancies with data on neonatal FNAIT outcome were included. Subsequent HPA‐1a incompatible pregnancies from women identified in the screening study were also included.
The pregnancies were categorized into four Pregnancies with ICH outcome (FNAIT with ICH, group I), clinically referred FNAIT cases without ICH (FNAIT no ICH, group II), HPA‐1a alloimmunized pregnancies identified prospectively with FNAIT outcome without ICH (group III) and prospectively identified pregnancies without FNAIT (group IV). Data from groups III and IV originate from the screening study, and these data were included to assess how a chosen antibody cutoff value to decide on antenatal treatment would perform if applied in a screening situation. Exclusion criteria were no written informed consent (from mother or children >16 years), antenatal IVIg treatment, HPA‐1 compatible fetuses, lack of available samples, or missing data. One pregnancy was excluded due to trisomy 13 in the neonate.
Definitions
FNAIT was defined as a neonatal platelet count <150 × 10^9^/L, with detectable maternal anti‐HPA‐1a antibodies and HPA‐1 incompatibility between mother and child. If data on neonatal HPA‐1 platelet type was missing, the pregnancy was considered HPA‐1 incompatible if the paternal genotype was HPA‐1aa. An index pregnancy was defined as the first pregnancy where FNAIT was diagnosed. We also defined a pregnancy as complicated by FNAIT if the fetus/neonate had ICH even if the neonatal platelet count was missing.
Antibody quantitation and time of sampling
There were limited data on anti‐HPA‐1a quantitation during pregnancy for the ICH cohort (2/15) and from some of the referred FNAIT cases with no ICH (24/55). Thus, antibody levels measured at the time of delivery were used for comparison. A sample was defined as taken at delivery if drawn after gestational week 34 or within 4 weeks before delivery, including up to 5 days after delivery. We included samples taken during pregnancy whenever available. All pregnancies from screening (n = 145) had longitudinal measurements. If there was more than one sample available during pregnancy, the highest antibody level was registered. If the only available sample was taken at delivery, this was included as the highest antibody level. Samples taken more than 5 days after delivery were excluded.
The level of maternal anti‐HPA‐1a antibodies was quantified at NNUPI at the time of FNAIT investigation or follow‐up using monoclonal antibody immobilization of platelet antigen (MAIPA) as previously described. ^14^ There were no significant changes to the quantitative MAIPA protocol during the study period. Samples from 1994 to 2007 were calibrated against a local anti‐HPA‐1a control plasma in AU/ml (arbitrary units) at the time of analysis and recalculated to IU/ml after the implementation of the NIBSC 03/152 anti‐HPA‐1a standard (100 IU) in 2007. For two of the pregnancies with ICH outcomes (IDs 1 and 2, Table 1) maternal anti‐HPA‐1a levels were quantified after years in cryo storage (−20 to −70°C).
HPA‐genotyping and platelet counts
HPA‐1 incompatibility data between mother and newborn were available from diagnostic FNAIT lab workup (HPA‐1 genotyping with TaqMan allele discrimination). We retrieved platelet counts from clinical records and from the previous screening study. ^21^
Obstetric history and ICH bleeding assessment
Obstetric history, course, and outcome of each pregnancy was retrieved from the medical records of mothers and neonates. If no medical record for the child was established during the perinatal period from the hospital where the delivery took place, we concluded that there had been no clinical suspicion of ICH. Data on pregnancies with ICH outcomes is presented in Table 1.
Statistical analysis
All data was analyzed using GraphPad Prism 10.3.1. Unpaired Mann–Whitney U test was used to compare anti‐HPA‐1a levels and neonatal platelet counts in the different study groups. A p‐value <0.05 was considered significant. MedCalc version 19.6 was used for evaluating diagnostic test performance by calculating receiver operating characteristic (ROC).
RESULTS
From referrals for investigation at NNUPI, we identified 145 HPA‐1a alloimmunized, non‐IVIg treated, HPA‐1a incompatible pregnancies where the newborn had FNAIT. Following exclusion criteria, we included 69 pregnancies classified as FNAIT, of whom 14 fetuses/newborns had ICH (group I) and 55 had FNAIT without ICH (group II) (Figure 1). From the previous Norwegian screening study, ^21^ 242 HPA‐1a alloimmunized pregnancies were identified, with 121 pregnancies from 115 women meeting our inclusion criteria. From these women, additionally 25 subsequent HPA‐1a alloimmunized pregnancies from 2005 to 2014 were included, in total 146 pregnancies. Of these, 99 pregnancies resulted in FNAIT without ICH (group III), 46 had no FNAIT outcome (group IV), while one woman from screening had a child affected by ICH (group I), Figure 1.
ICH cohort (group I)
Among 15 non‐IVIg treated HPA‐1a alloimmunized pregnancies complicated by fetal/neonatal ICH, 14 were index pregnancies. Key characteristics of all 15 ICH‐complicated pregnancies are reported in Table 1. All but one (ID 3) were clinical referrals, and most (n = 12) were first‐borns. One woman had two children with ICH (ID 4 and 5), with both pregnancies occurring before antenatal IVIg treatment to women with prior FNAIT history with ICH outcome was included in the Norwegian clinical guidelines in 2014. ^22^ Two of the ICH cases were from twin pregnancies (ID 1 and 12). Data from five of the pregnancies with ICH outcome (ID 1–5) were previously included in an international study on ICH caused by FNAIT. ^2^ Two of the ICH cases were from 1997 to 2000; six between 2001 and 2010 and seven between 2011 and 2023.
Antibody levels
The anti‐HPA‐1a levels at delivery among ICH‐complicated pregnancies (group I) were significantly higher (median 29.6 IU/mL, range 0.1–222.1 IU/mL) compared to pregnancies with FNAIT, but no ICH outcome referred to NNUPI (group II) (median 10.4 IU/mL, range 0.0–83.1 IU/mL, n = 55; p = 0.046, Mann–Whitney U test; Figure 2A). In five pregnancies, the antibody levels exceeded 100 IU/mL, whereas in four, antibody levels were below 1 IU/mL. Antibody levels at delivery in group I were also higher compared to those of combined groups II + III (FNAIT without ICH outcome from clinical referrals and pregnancies identified by prospective screening), median 12.2 IU/mL, range 0.0–133.6 IU/mL, n = 131; p = 0.041, Mann–Whitney U test (Figure 2B). There was no significant difference in neonatal platelet counts between referred FNAIT cases without ICH (median platelet count 16 × 10^9^/L, range 2–122 × 10^9^/L) and ICH cases (median platelet count 11 × 10^9^/L, range 1–46 × 10^9^/L, p = 0.286, Figure 2C). Of note, 82% of neonates without ICH had severe FNAIT with platelet counts <30 × 10^9^/L. The association between maternal anti‐HPA‐1a levels and the corresponding neonatal platelet count is shown in Figure 2D.
To assess the feasibility of using antibody levels for identification of pregnancies at higher risk of ICH, we analyzed test performance ROC analysis (Figure S1). Since data on anti‐HPA‐1a quantitation during pregnancy were only available in two pregnancies from the ICH cohort, we compared antibody levels from samples taken at delivery in FNAIT pregnancies with no ICH referred to NNUPI (n = 55) and FNAIT with ICH outcome (n = 15, Figure 1). A cutoff value of 70 IU/mL was chosen by evaluating sensitivity versus specificity values, resulting in a positive predictive value (PPV) of 75.0% and negative predictive value (NPV) of 85.5%. The sensitivity was only 40.0%, however, the specificity was estimated to 96.4% (Table 2).
As outlined above, women without an obstetric history of ICH^22^ are in Norway not generally offered antenatal IVIg, but instead receive more closely clinical follow‐up if anti‐HPA‐1a levels are ≥3 IU/mL. We therefore did a subanalysis of pregnancies with antibody levels ≥3 IU/mL (at delivery) to evaluate whether a further risk stratification within this group could be useful. The anti‐HPA‐1a antibody level was significantly higher in the ICH group (median 71.5 IU/mL, n = 11) compared to the no ICH FNAIT group (n = 39, median 21.5 IU/mL, p = 0.0025, Mann–Whitney U test). ROC analysis of the subanalysis group with the selected threshold of 70 IU/mL gave similar test characteristics (Table 2). The maximum Youden's index (sensitivity + specificity − 1) from ROC data for the subcohort (≥3 IU/mL at delivery) was 70.7 IU/mL and in line with the selected cutoff at 70 IU/mL.
Two pregnancies with ICH outcomes also had additional samples taken during first or second trimesters. Notably, the antibody levels in these samples were higher (154 IU/mL and 97 IU/mL) than the corresponding anti‐HPA‐1a concentrations at delivery (41 IU/mL and 72 IU/mL, respectively). Longitudinal antibody levels during pregnancy for no ICH FNAIT pregnancies (group II) and ICH pregnancies (group 1) are shown in Figure 3.
To evaluate the potential impact on antenatal treatment decisions if a cutoff value of 70 IU/mL had been applied in these pregnancies, we assessed the frequency of pregnancies with antibody levels above this cutoff across the study pregnancies with ICH outcome (group I), clinically referred FNAIT cases without ICH (group II) and HPA‐1a alloimmunized pregnancies identified prospectively (group III + IV). The highest antibody level during pregnancy was used when available; otherwise, measurements from the time of delivery were used (Figure 2E). From clinically referred FNAIT pregnancies without ICH, only two out of 55 pregnancies (3.6%) had antibody levels above the suggested 70 IU/mL threshold. Similarly, 6.9% (10 out of 145) HPA‐1a alloimmunized pregnancies identified through screening had levels >70 IU/mL (Figure 2E), representing the frequency of pregnancies that may have been offered antenatal treatment if this cut‐off were implemented.
DISCUSSION
In the present study, we show an association of maternal anti‐HPA‐1a antibody levels and fetal/neonatal ICH outcome among non‐IVIg treated pregnancies. Our study comprises the largest cohort of non‐IVIG treated HPA‐1a immunized pregnancies for comparison of antibody levels in pregnancies with or without FNAIT‐related ICH. Our findings open the possibility for the use of antibody levels to assist in predicting risk for fetal/neonatal ICH in pregnancies today classified as standard FNAIT risk as well as pregnancies where a prior FNAIT history is lacking. Thus, a strategy including risk stratification based on antibody measurements during pregnancy may aid in preventing severe brain bleeds by targeted antenatal treatment to the few women with very high antibody levels while at the same time avoiding overtreatment.
A limitation of this study is the lack of maternal longitudinal samples for antibody quantitation among pregnancies complicated by fetal/neonatal ICH. Antibody measurements from samples taken around the time of delivery are not optimal to identify antibody thresholds to be used to guide clinical management during pregnancy, since antibody levels may fluctuate during pregnancy. However, anti‐HPA‐1a levels have been reported to decline toward delivery in many HPA‐1a alloimmunized pregnancies. ^14^ It is therefore not unlikely that antibody levels may have been even higher during pregnancy. We only had repetitive antibody level data from two ICH‐complicated pregnancies, notably both with descending antibody levels.
In contrast to many other areas of perinatal medicine, postnatal treatment of FNAIT has not changed during the nearly 30 years inclusion period of this study; if FNAIT is suspected, the treatment has consistently been immediate transfusion of compatible platelets. Although antenatal treatment practice has changed slightly during this period, possibly affecting the incidence of ICH over time, this study assesses associations between antibody levels and ICH, not ICH rates. The distribution of ICH cases included was evenly distributed, and there is no reason to think that these have changed over the past 30 years. We therefore consider our study population to be representative of a current population.
Norway's historically restrictive antenatal treatment policy regarding FNAIT, combined with performing anti‐HPA‐1a antibody level measurements, is unique and allows for an assessment of antibody levels and the risk of FNAIT‐related ICH. Previous association studies of anti‐HPA‐1a levels have primarily predicted the severity of the neonatal thrombocytopenia. Data showing the usefulness of antibody levels in ICH risk prediction has been missing, either due to no significant differences, lack of appropriate control groups, limited sample sizes, and a broad range in antibody levels. ^18^ , ^19^ The long‐standing tradition of treating all HPA‐1a alloimmunized pregnant women with IVIg, especially in a non‐screening setting, constrains the opportunity to identify better predictive tools. There are currently several ongoing clinical trials related to preventative treatment for FNAIT (IPA2202; ClinicalTrials.gov ID NCT06435845 and FREESIA‐1; ClinicalTrials.gov ID NCT06449651) and the results of these clinical trials can potentially support the implementation of antenatal HPA‐1 screening of pregnant women. If antenatal HPA‐1a screening is introduced, the relevance of identifying the subgroup of HPA‐1a alloimmunized pregnancies, where treatment is necessary and effective, will increase significantly.
We found significantly higher maternal anti‐HPA‐1a antibody levels in ICH cases compared to pregnancies with FNAIT without ICH outcome. This is in accordance with data from Ghevaert et al. ^19^ who also reported higher levels of maternal anti‐HPA‐1a levels in ICH cases with a broad range, although not significantly different compared to pregnancies with FNAIT without ICH outcome. Delbos et al. ^20^ reported even higher median maternal anti‐HPA‐1a levels among ICH cases (median 99 IU/mL), which is in accordance with the one ICH case identified through the recent Dutch HPA‐1 screening study. ^13^ Together, the studies support both the association across populations, but also that our chosen cutoff level may be transferable to other cohorts.
Our antibody quantity data suggest that there are subcohorts among ICH cases, including one group with extremely high concentrations (>100 IU/mL) and one group where antibodies are barely detectable. This widespread range in antibody levels is noteworthy, but also in line with previous reports. ^18^ , ^19^ , ^20^ One possible interpretation of this phenomenon is that factors other than the total amount of anti‐HPA‐1a antibodies may contribute to bleeding risk, especially among those individuals with very low antibody levels. Several aspects of the maternal anti‐HPA‐1a antibody characteristics, such as glycosylation and effector function, have been indicated as possible predictors of severe FNAIT. ^23^ , ^24^ It has further been suggested that the glycoprotein dimer‐complex subspecificity, rather than the quantity, is a critical factor in the risk of ICH. ^24^ , ^25^ However, this association has not yet been confirmed in other international FNAIT cohorts, nor are any standardized tests available for such antibody subspecificity detection. An alternative explanation is that the antibody level has declined significantly towards delivery and was much higher during pregnancy. ^14^ Only future larger‐scale studies, including longitudinal measurements in HPA‐1a alloimmunized pregnancies, can clarify this.
As highlighted in the 19th report of the Platelet Immunology workshop by the ISBT, ^26^ the protocols for MAIPA are not standardized, neither are the quantitative in‐house MAIPAs. ^15^ This is exemplified by different antibody reagents for capture, glycoprotein targeting and detection, incubation times, number of platelets used per well, and buffers. However, all quantitative MAIPA protocols are calibrated against the NIBSC international quantitative standard (03/152). Some comparative studies were performed upon the development of this standard, as well as laboratory‐initiated comparative efforts, however, using a limited number of samples. This means that our chosen cutoff value may not be directly applicable to other labs or cohorts. To implement an international antibody threshold for antenatal treatment, the method for antibody quantitation should be standardized. Although there are other types of assays to detect anti‐HPA‐1a antibodies, ^27^ most do not allow quantitation.
The ROC analysis indicated that an antibody threshold of 70 IU/mL gives a sensitivity of 40% and a specificity of 96.4% for predicting the risk of ICH, the latter being particularly important to avoid overtreatment. ^7^ Albeit the low sensitivity, due to only barely detectable antibodies in some pregnancies with ICH outcome, the positive and negative predictive values warrant a potential use of the test.
Acknowledging that our antibody levels used for analyzing the association with ICH were collected around the time of delivery, a cutoff value of 70 IU/mL applied to the current data set, and if used as a rule‐in test, would have resulted in unnecessary antenatal IVIg treatment of 3.6% of the clinical referrals where the neonate had FNAIT, but not ICH. Similarly, based on the prospectively identified cohort, ^21^ this would correspond to 6.9% of alloimmunized pregnancies or about 1% of HPA‐1bb women being treated. We consider this a reasonable rate of potential overtreatment. On the other hand, by applying our selected cutoff to the ICH cohort as a rule‐out test, less than half would have been offered antenatal treatment (7/15, 46.7%). However, as Figure 2E illustrates, the alternative of treating all standard‐risk pregnancies, as well as all HPA‐1a alloimmunized pregnancies identified through screening, would imply a substantial overtreatment, given that the goal of treatment is bleeding prevention. The risk of fetal/neonatal ICH among standard‐risk pregnancies is, moreover, found to be very low even without antenatal IVIg. ^7^ , ^9^
Avoiding overtreatment is important for several reasons. The efficacy of weekly antenatal IVIg treatment of alloimmunized women has never been documented in a placebo‐controlled clinical trial. High‐dose IVIg treatment also has a negative impact on the quality of life of a significant number of the treated pregnant women. ^28^ , ^29^ Further, such off‐label use of IVIg is not only extremely costly, but it also involves tremendous donor effort to produce. ^7^
CONCLUSION
Our data indicate that including maternal anti‐HPA‐1a antibody levels could be helpful when assessing ICH risk and thus the need for antenatal treatment in pregnancies at risk of FNAIT. This could be relevant for pregnancies where the risk of ICH is either considered low due to prior history of FNAIT without ICH or where there is no obstetric history to guide such risk. To avoid undertreatment in Norway, we propose revising our National clinical guidelines and offering antenatal IVIg even to standard risk HPA‐1a alloimmunized pregnancies if maternal anti‐HPA‐1a antibody levels measured at any time during pregnancy exceed 70 IU/mL. If other countries were to adopt a similar approach, it would result in a clear reduction in the number of pregnancies being treated with antenatal IVIg, thus reducing overtreatment. Use of anti‐HPA‐1a antibody level thresholds could also achieve a more targeted treatment if antenatal screening to detect all HPA‐1a alloimmunized pregnant women were introduced. More data on maternal anti‐HPA‐1a antibody levels during the first and early second trimester in pregnancies complicated by ICH are desirable to improve predictive values but difficult to obtain due to the rarity of this complication.
AUTHOR CONTRIBUTIONS
SLE, MTA, ELB, and HT conceptualized the study, collected, and analyzed the data. SLE, MTA, and HT drafted the manuscript. SLE, MTA, ELB, AH, JKK, and HT made substantial contributions to the design of the study, interpretation of data, and critically revising the manuscript for important intellectual content. All authors approved the final version to be published.
FUNDING INFORMATION
This study received no external funding. No entity other than the authors listed played any role in the design of the study; the collection, analysis, or interpretation of data; writing of the report; or in the decision to submit the paper for publication.
CONFLICT OF INTEREST STATEMENT
HT is a member of a steering committee for Janssen Pharmaceuticals as part of an ongoing clinical trial related to FNAIT (FREESIA‐1). HT reports previous payment from Prophylix AS related to a patent on a monoclonal anti‐HPA‐1a antibody. HT is the local study site PI, and MTA and SLE are part of the local study team in two multicenter studies on FNAIT (IPA2202 sponsored by Rallybio and FREESIA‐1 sponsored by Janssen). AH and JKK belong to a group of founders and owners of Prophylix AS, a Norwegian biotech company that has produced a hyperimmune anti‐HPA‐1a IgG (NAITgam) for the prevention of HPA‐1a alloimmunization and FNAIT. JKK is also a member of the Scientific Advisory Board for the FREESIA‐1 study sponsored by Janssen Pharmaceuticals. In addition, JKK is a consultant for Rallybio IPA, LLC, a US biotech company developing a new drug in class (RLYB212) for the prophylaxis of HPA‐1a immunization and FNAIT. ELB declares no conflict of interest.
ETHICS STATEMENT
The study was approved by the Regional Ethics committee, North Norway (REKNORD 2009/1585) on January 20, 2009. All women gave informed written consent on the use of biological material and clinical data.