Introduction

Alloimmunisation against the Rhesus D (RhD) red cell antigen is still the most common cause of severe hemolytic disease of the fetus and newborn (HDFN). [1], [2] Affected pregnancies require close surveillance in order to detect fetal anemia in a timely manner, and the neonates often require treatment for hyperbilirubinemia and anemia. RhD immunisation is the commonest indication for intrauterine blood transfusion and neonatal exchange transfusion. [3], [4] Once a woman is RhD immunised, all her subsequent pregnancies with an RhD positive fetus will be affected.

Since the beginning of the 1970s, the risk of RhD immunisation has been reduced from around 13 percent to 1–2 percent by administration of anti-D immunoglobulin (IgG) after delivery. [5] It has been shown that the most important residual cause of immunisation is silent fetomaternal haemorrhage (FMH) during pregnancy, mostly during the third trimester. [6], [7] Routine antenatal anti-D prophylaxis (RAADP) at the beginning of the third trimester has been introduced in several countries, reducing the incidence of RhD immunisation to 0.2–0.3 percent. There is strong evidence that RAADP is effective in the prevention of RhD immunisation. [8]–[11] Anti-D prophylaxis consists of pooled polyclonal anti-D IgG from human plasma donors. Generally, RAADP has been administered to all RhD negative women, even though 40 percent will carry a compatible RhD negative fetus, and so are not at risk for RhD immunisation. RAADP in such women unnecessarily exposes them to a human plasma product and is wasteful of the limited supply of anti-D immunoglobulin, which should be given only when strictly indicated. The discovery of cell-free fetal DNA in maternal plasma has led to the development of non-invasive methods to determine the fetal RHD genotype.[12]–[14] Several studies have confirmed the safety and high diagnostic accuracy of this approach.[15]–[19] This enables administration of RAADP selectively to RhD negative women with an RHD positive fetus and avoids giving it to women who are not at risk. This strategy has recently been introduced to Danish and Dutch screening programs, but no study has yet been published of the effectiveness of this approach in reducing the incidence of new RhD immunisation. [20].

Routine antenatal anti-D prophylaxis has not previously been offered in Sweden. The aim of this study was to estimate the incidence of RhD immunisation after implementation of first trimester non-invasive fetal RHD screening using cell-free fetal DNA in maternal plasma to selectively provide routine antenatal anti-D prophylaxis (RAADP) to RhD negative women carrying an RHD positive fetus.

Materials and Methods

Results

Between 1^st of January 2010 and 31^st of March 2012, 9380 RhD negative women gave birth in Stockholm. Non-invasive fetal RHD genotyping was performed in 8374 pregnancies (89%) (1 Sept 2009 to 1 Sept 2011). 3270 fetuses were determined to be RHD negative (39%) and 5104 (61%) RHD positive. 4521 women participating with 4590 pregnancies (90%) received targeted RAADP due to an RHD positive fetus. Sixty-nine women participated with two pregnancies. Twenty-five women with anti-D antibodies in the first trimester of pregnancy and RHD positive fetal genotype were excluded from the study. Data on all included and excluded women and outcomes are described in Figure 1. The median gestational week for receiving RAADP was 29 weeks (25^th to 75^th percentile: 29 to 30 weeks).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Study participants and outcomes.

https://doi.org/10.1371/journal.pone.0070984.g001

In the reference cohort, the incidence of RhD immunisation was 0.46 percent (95% CI 0.37 to 0.56) (86/18546). Twenty-four new cases of RhD immunisation have been confirmed during the study period, resulting in an incidence of 0.26 percent (24/9380) (95% CI 0.15–0.36%). Of these new cases, two women had not received RAADP and fetal RHD genotyping had not been performed. This happened at the beginning of the study period when compliance with the new program was incomplete. Frequencies of antibody screening tests in the study cohort were 88 percent in the second trimester (around 25 weeks of GA), 74 percent at delivery, 71 percent post partum (median 7.5 months after delivery). In 91 percent of participants at least one sample was taken at delivery or at post partum follow-up. The frequency of antibody tests in the second and third trimester in the reference cohort was unknown. The risk ratio (RR) for sensitisation with the new program was 0.55 (95% CI 0.35 to 0.87). The absolute risk difference was 0.20 percent, which corresponds to a number needed to treat (NNT) of 500. The risk reduction was statistically significant (P = 0.009).

Data on all new RhD immunisations during the study period are described in Table 1. Fifty-eight percent (14/24) of the women were already sensitised before being scheduled for anti-D prophylaxis. Four women were sensitised in spite of receiving RAADP and had seroconverted at delivery. In two of these, the frozen sample was not available and the possibility of sensitisation before prophylaxis cannot be ruled out. Two women had a negative antibody test at delivery but anti-D was detected at 10 months postpartum or in the first trimester in a subsequent pregnancy. In two cases it was unclear if the women had become immunised during the third trimester in spite of receiving RAADP or after delivery. The antibody test at delivery was missing in these two cases. In the two cases where the women had not received RAADP, one woman already had detectable anti-D antibodies at 28 weeks of gestation. Seven of the immunised women were nulliparous. One of those had anti-D antibodies with high titres detected at routine screening at 25 weeks of gestation. She had rapidly increasing antibody concentrations and multiple intrauterine blood transfusions were required due to fetal anemia. A boy was delivered at 34 weeks and congenital cytomegalovirus (CMV) infection was diagnosed soon after birth in addition to severe hemolytic disease of the newborn. Investigation of the mother, neonate and the blood donors indicated reactivated CMV infection in the mother during pregnancy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Details on women RhD immunised during study period.

https://doi.org/10.1371/journal.pone.0070984.t001

Discussion

First trimester non-invasive antenatal screening for fetal RHD combined with routine antenatal anti-D prophylaxis given selectively to women with an RHD positive fetus significantly reduces the incidence of RhD immunisations. The risk reduction is comparable to that reported in studies evaluating the outcome of giving RAADP to all RhD negative women.[9]–[11], [23] To our knowledge, this is the first large study to evaluate the reduction of RhD immunisation in a community after introduction of such a program. The number needed to treat (NNT) to prevent one sensitisation was 500, which means that non-invasive RHD screening would have to be performed in 500 women, but only about 60 percent (NNT 300) of these would need RAADP.

Of the two cases in which the women had not received antenatal prophylaxis, one might have been avoided with prophylaxis. In the third woman, anti-D was already detectable at 28 weeks of gestation and sensitisation would therefore not have been prevented with the current program. Large fetomaternal haemorrhages that overwhelm the dose of anti-D IgG administered may lead to sensitisation in spite of prophylaxis. Analysis for fetal red blood cells in maternal blood after delivery to assess the magnitude of FMH is not routinely performed in Sweden, but in a few women the standard dose of 250–300 µg will not be sufficiently protective. Risk factors for RhD immunization in spite of prophylaxis include assisted delivery and caesarean section, postmaturity, pregnancy-related red blood cell transfusion and younger age. [24] Three women immunized in spite of RAADP in our study cohort delivered after 41 weeks of gestation. The effect of RAADP depends not only on the size of silent FMH during the third trimester, but also on the time between administration and delivery. Most women have detectable levels of anti-D IgG 10 weeks after receiving 250–300 µg RAADP but after that levels are variable. [25] [26] This means that some pregnant women will be unprotected in gestational weeks 38 to 42. Routine administration of anti-D prophylaxis in gestational weeks 38–39 instead of the postnatal injection could protect throughout this period. A high enough dose at this gestational age will normally protect against sensitisation after delivery.

Screening in the first trimester has the clinical advantage that only individuals at risk are exposed to a human plasma product. Women will also benefit from knowing the RHD type of the fetus in case of miscarriages, abortions and invasive prenatal testing such as amniocentesis and chorionic villus sampling. A disadvantage of first trimester screening is the lower sensitivity of the test compared to later in pregnancy. [21] However, for a screening purpose, the sensitivity of 98.9 percent is still very high and first trimester testing is safe to use. For a diagnostic purpose, as in already RhD immunised pregnant women, we recommend repeated testing in gestational week 20 for confirmation when fetal RHD genotype is determined as negative. Knowing the fetal genotype early in pregnancy also enables new preventive strategies to reduce sensitisation in the second trimester.

A relatively high number (14/9380, incidence approximately 0.15%)) of cases of new RhD immunisations in the study cohort occurred during the second trimester of pregnancy. Our findings are in line with the evaluation of the Dutch RhD prevention program, in which the incidence of immunisations occurring between 12 to 30 weeks of pregnancy was 0.24 percent before introduction of RAADP and 0.25 percent after. [10] In Bowman’s classical studies, a comparable frequency of women were immunised before the 28^th week of gestation. The most likely reason for early sensitisation in spite of a negative first trimester antibody test is FMH occurring early in pregnancy. But it is also possible that some women may have been sensitised in a previous pregnancy but did not develop detectable IgG antibodies until the second trimester of the subsequent pregnancy. Routine anti-D prophylaxis could be added in 16 weeks of pregnancy to prevent second trimester sensitisation in women with RHD positive fetuses to protect until the dose at week 28.

We performed a community-based study aiming at evaluating the outcome of a new prevention program in a routine setting. The aim was to include all RhD negative women in our population, but during the first six months of the study period recruitment to the program was slow, which explains the discrepancy between the number of RhD negative women who gave birth during the study period and the number of women who underwent screening for fetal genotype. The compliance of both women and health care professionals with the protocol has increased with time. The use of an intention-to-treat analysis should mean our study provides a good indication of the expected effectiveness of this new program in our health care system. An evaluation of the first six months of the Danish national antenatal RHD screening program noted that adoption of the program was slow with only about 80 percent of the expected number of maternal samples being sent to the laboratories. Compliance with antenatal administration of anti-D was unknown and can be anticipated to be lower. [20] The strengths of our study include a large study sample, prospective collection of data, central management of all cases of red cell immunisation in our region and reliable registers of these pregnancies. We used two time points, both at delivery and six to ten months postpartum, to find women who seroconverted in the study cohort. The frequency of received samples at follow-up post partum was 71 percent and at least at delivery or postpartum 91 percent. In the reference cohort, antibody screening was performed only at 37 weeks of pregnancy and not at delivery or post partum. Not all women in the reference cohort had a subsequent pregnancy when antibodies from sensitisation late in the third trimester or at delivery in the previous pregnancy would be found. We believe the detection rate of new RhD immunisations in the reference cohort to be somewhat underestimated due to this. Since the optimal time point to assess the incidence of alloimmunisation is in the first trimester of a subsequent pregnancy, we expect a few more cases to appear in the study cohort on long-term follow-up since some women may have been sensitised at delivery, but anti-D antibodies will not be detected until the first trimester of a subsequent pregnancy. Some new RhD immunisations may also be detected in future in women in the reference cohort who have another child. The incidence of RhD immunisation was as low as 0.46 percent in the reference cohort. The prevalence of RhD immunisation during the five-year period 2004–2008 was 0.7 percent. Since the study sample was large enough this did not affect the power of the study. The low frequencies compared to previous reports might be due to improved obstetric management of RhD negative women in situations with risk for FMH and also improved compliance to anti-D prophylaxis recommendations over the years. Most reports on the prevalence of RhD immunisation with a postnatal prophylaxis program were published in the 1970s and 1980s and reports already in the 1990s show a lower prevalence compared to the earlier studies. [27], [28] The incidence of RhD immunisation in the Dutch population before introduction of routine antenatal anti-D prophylaxis in 1998 was reported to be 0.5 percent, which is similar to our findings. [10].

We chose an observational cohort study to assess the effectiveness of the new intervention and not a randomised controlled study. In our opinion, there is already strong evidence for the effectiveness of RAADP in preventing sensitisation to the RhD antigen.[8]–[11], [23] A randomised study in which RAADP was administered to all RhD negative women in one arm, and targeted RAADP offered exclusively to women with an RHD incompatible fetus in the other would have been complicated to carry out in our setting and it is questionable whether such a study would have been successful or have provided more reliable outcome data. The use of a historical control group has its limitations. However, obstetric and transfusion practice and laboratory methods have remained essentially unchanged during the reference and study periods. We believe no important selection bias was introduced in our study.

Acknowledgments

We thank the many pregnant women who took part in this study and all the maternity care centres and delivery hospitals in Stockholm for their co-operation. We are grateful to Elisabeth Berg for assistance with the statistical analysis, to Professor Dick Oepkes for valuable feedback in writing the manuscript and to Birgit Örn for registering follow-up data in the study database.