High-throughput non-invasive prenatal testing for fetal RHD genotype

The committee considered the current standard of care offered to pregnant women who are rhesus‑D (D) negative. It heard from clinical experts on the committee that current care for women who are not sensitised to the D antigen involves routine antenatal anti‑D prophylaxis, additional doses of anti‑D immunoglobulin if a woman has a potentially sensitising event, and postpartum testing of cord blood and anti‑D prophylaxis if cord blood typing shows the baby to be D positive. The committee noted that introducing these methods for preventing sensitisation of women to the D antigen has dramatically reduced the number of sensitisations and the rates of haemolytic disease of the fetus and newborn over the last 40 years. The committee also heard from a clinical expert that there are effective treatments for D‑negative women who are sensitised to D antigen, which means that deaths from severe haemolytic disease of the fetus and newborn are very rare.

The committee considered whether there were any problems with the current care offered to pregnant women who are D negative and not sensitised to the D antigen. It heard from a clinical expert that errors do sometimes occur, for example, a small number of women at risk of sensitisation do not have anti‑D immunoglobulin, or do not receive the correct dose of anti‑D immunoglobulin at the correct time. The committee also heard from a clinical expert that many sensitisations result from clinically silent fetomaternal haemorrhage events – potentially sensitising events without a known cause or clinical symptoms. The committee further heard that cord blood typing to determine the Rh blood group of the baby after birth may be affected by errors, such as sampling the blood of the mother instead of the baby, or incorrect sample labelling. The committee concluded that although anti‑D prophylaxis is very effective for reducing sensitisations and therefore haemolytic disease of the fetus and newborn, it is not perfect because sensitisations do still happen.

The committee considered the possible disadvantages of using anti‑D immunoglobulin. It heard from experts that there have previously been shortages of supply because it is a blood product and therefore a finite resource. The committee also considered the potential future risks from unknown prions or pathogens associated with using a blood product such as anti‑D immunoglobulin. The committee concluded that it would be beneficial to conserve supplies by only using anti‑D immunoglobulin for those in whom it is necessary.

The committee heard from a patient expert that for some women, having anti‑D immunoglobulin may not be acceptable for personal, cultural or religious reasons. It noted that using high-throughput non‑invasive prenatal testing (NIPT) for fetal RHD genotype would allow women whose fetus was identified as D negative to avoid having unnecessary anti‑D immunoglobulin, while women identified as having a D‑positive fetus would be able to make an informed decision about whether to have anti‑D immunoglobulin.

The committee considered the diagnostic performance of high-throughput NIPT for fetal RHD genotype. It noted that good-quality evidence was available and that the test is accurate after 11 weeks of gestation. The committee then considered how the diagnostic accuracy of the test affected clinical effectiveness. It noted that there is a small increase in the false-negative rate for high-throughput NIPT to determine fetal RHD genotype (0.21%; 95% CI [confidence interval] 0.09 to 0.48), compared with the current practice of postpartum cord blood typing. This means that some women with a D‑positive fetus would be incorrectly identified as having a D‑negative fetus and would not be offered routine antenatal anti‑D prophylaxis (RAADP) or anti‑D immunoglobulin after potentially sensitising events. Because of this, more women could become sensitised to the D antigen and may have complications in later pregnancies, although the committee heard from a clinical expert that the severity of these complications is hard to predict. The committee noted that the rate of sensitisations with current practice was estimated to be 281 per 100,000 D‑negative pregnancies. If offering RAADP and anti‑D immunoglobulin after potentially sensitising events was based on the results from NIPT for fetal RHD genotype, the rate of sensitisations would increase by 3, to 284 sensitisations per 100,000 D‑negative pregnancies. The committee considered that this relatively small increase in the number of sensitisations could be accepted in the context of other potential benefits of NIPT associated with avoiding unnecessary treatment with blood products.

The committee considered the results of the economic model. It noted that the quality-adjusted life year (QALY) losses in the model were 0.46 per 100,000 pregnancies if the postpartum testing strategy stayed the same as current practice (cord blood typing for all D‑negative women regardless of the NIPT result). The committee noted that although this is a reduction in clinical effectiveness compared with current practice, the reduction is extremely small (0.0000046 QALYs per pregnancy). The committee therefore concluded that the clinical effectiveness of using high-throughput NIPT for fetal RHD genotype to guide antenatal anti‑D prophylaxis is comparable with offering antenatal anti‑D prophylaxis to all D‑negative women, provided that there are no changes to postpartum practice.

The committee considered the cost savings in the economic model. It noted that cost savings in the models were £583,538 per 100,000 pregnancies in D‑negative women if the postpartum testing strategy stayed the same as current practice (cord blood typing for all D‑negative women regardless of the NIPT result, referred to as postpartum scenario 1 [PP1] in the economic analysis). The committee noted that the cost savings are relatively small, at £5.84 per pregnancy, and on their own might not justify the risks that could be associated with making substantial changes to current practice. The committee then considered the base-case incremental cost-effectiveness ratio (ICER) for PP1. It noted that although the ICER appears to be large, at £1,269,100 saved for each QALY lost, it is very sensitive to changes in the numerator (change in cost) or denominator (change in QALYs), and is therefore subject to substantial uncertainty. The committee concluded that the total costs for using high-throughput NIPT for fetal RHD genotype to guide antenatal anti‑D prophylaxis are not substantially different from the total costs for the current practice of offering antenatal anti‑D prophylaxis to all D‑negative women, provided that there are no changes to postpartum practice.

The committee considered the different postpartum testing strategies presented in the diagnostics assessment report (see section 4.24). It noted that in the base-case analysis, NIPT using PP1 was the most cost-effective strategy compared with current practice. The committee heard from clinical experts that the postpartum care in PP1 is the same as used in current clinical practice. It also noted that with different assumptions on postpartum testing, other postpartum scenarios could be associated with greater cost savings, but increased QALY losses compared with PP1. The committee concluded that it is preferable to minimise the QALY losses.

The committee considered the results of a scenario analysis that made different assumptions on postpartum testing (PP5). It noted that this postpartum scenario was associated with greater cost savings and equivalent QALY losses compared with PP1. The committee heard from clinical experts that postpartum testing involves taking a cord blood sample soon after the birth, and that although midwives are used to doing this they also have multiple other tasks to complete at this time. The committee was concerned that if midwives had to get the NIPT result and then make a decision on whether to take a cord blood sample in the period immediately after the delivery, errors could be made, for example, not taking a cord blood sample from a fetus predicted to be D negative. The committee concluded that although alternative postpartum strategies may potentially have greater cost savings, they would be complicated to implement in clinical practice and may result in errors, which could affect costs and clinical effectiveness. The committee also discussed an alternative approach in which cord blood samples would be taken from all babies born to D‑negative women and the laboratory would then decide whether to test the cord blood sample. The committee concluded that the postpartum testing of cord blood should not be changed from current practice. This is because without cord blood typing, false-negative NIPT results would not be identified and women with false-negative NIPT results would not have postpartum anti‑D prophylaxis. The committee noted the difficulties of taking a blood sample from the cord and that the consequences of a sampling error may include having to take repeat blood samples from a neonate. It decided that further research on the practicalities of implementing alternative postpartum testing strategies would be valuable (see section 6.2).

The committee discussed the input used in the model for the cost of care of a pregnant woman who has been sensitised to the D antigen in an earlier pregnancy. It heard from a clinical expert that some women who are sensitised to the D antigen will be identified as having a D‑negative fetus, and others will be identified as having a D‑positive fetus, but will not experience problems during their pregnancy. These 2 groups of women would not need many extra appointments with a specialist obstetrician. A third group of sensitised women will be identified as having a D‑positive fetus and will experience problems during their pregnancy. These women will need more frequent surveillance and treatment for the baby before and after the birth. The committee concluded that if a weighted average is taken of the cost of care for these 3 groups of pregnant women, then an input of £3,167 per sensitised pregnancy is reasonable.

The committee considered the cost of the test that was used in the economic model and noted that the cost did not include sample transport. It heard from the current provider of the test, the International Blood Group Reference Laboratory (IBGRL), that blood samples are transported around the country to their laboratory on a daily basis using the NHS Blood and Transplant (NHSBT) transport network. It heard further, that because of this established transport network, there should be no cost for sample transport. The committee was concerned that although there may be no cost for sample transport between the NHSBT units and the IBGRL, there may be a cost for transporting the sample from the maternity clinic to the NHSBT unit. It was also concerned about the length of time it may take to transport samples from rural areas to the IBGRL, and that longer sample transport times may result in increased rates of failed tests. The committee also heard from the IBGRL that the unit cost of the test depends on the expected annual sample throughput and on a royalty fee, which is currently under negotiation. The committee concluded that the test cost is uncertain.

The committee considered whether there were any costs associated with implementing high-throughput NIPT for fetal RHD genotype that had not been included in the economic model. It noted that extra time to explain the test, take the blood sample, give the test results, and provide counselling, that could result in extra midwife appointments, were not included in the model. The committee heard from a clinical expert that in the south-west of England where high-throughput NIPT for fetal RHD genotype has been implemented, the blood sample for the test is normally taken at the routine 16‑week antenatal appointment; therefore, no additional appointments are needed. It also heard from the clinical expert that the main issue when implementing the test was educating midwives and other healthcare professionals so they understood the test and could explain it to women and their families. It noted that a patient information leaflet explaining the test and its results is available from NHSBT. The committee heard from the external assessment group that none of the studies in the review of implementation, included costs associated with implementation. The committee concluded that the costs associated with implementing high-throughput NIPT for fetal RHD genotype were uncertain.

The committee considered a threshold analysis done by the external assessment group on the unit cost of the test. It noted that results show that the cost effectiveness of high-throughput NIPT for fetal RHD genotype is sensitive to small increases in costs associated with doing the test, for example, sample transport, the need for repeat tests, midwife time, or the cost of the test itself. The committee also noted that increasing the test cost to £24.64 or more per test would result in high-throughput NIPT for fetal RHD genotype no longer being cost effective compared with current practice, using a maximum acceptable ICER of £20,000 per QALY gained. The committee concluded that high-throughput NIPT for fetal RHD genotype has the potential to be cost effective, but that the cost savings are volatile with respect to the cost of the test (see section 5.11) and the costs associated with implementation (see section 5.12). The committee also concluded that the overall cost of testing below which high-throughput NIPT for fetal RHD genotype can be considered cost effective should not be stated to 2 decimal places in the recommendation. This is because there is substantial uncertainty about the results of the model. The committee decided that £24.64 should be rounded down to £24 rather than up to £25 to increase the chance of high-throughput NIPT for fetal RHD genotype being cost effective.

The committee noted its conclusions on the comparable clinical effectiveness of high-throughput NIPT for fetal RHD genotype and current practice (see section 5.6), and the uncertainty about cost savings (see section 5.13). The committee also noted its conclusion that it would be beneficial to avoid inappropriate use of anti‑D immunoglobulin (see section 5.3). The committee decided that although the cost savings are potentially small, recommending high-throughput NIPT for fetal RHD genotype would be an effective way of reducing unnecessary use of anti‑D immunoglobulin, and that this reduction could affect a large number of women.

The committee considered the effect that ethnicity has on NIPT results. They heard from the provider of the test, IBGRL, that D‑negative women of black African family origin are more likely to have an RHD pseudogene, and so are more likely to have an inconclusive or false-positive NIPT result compared with women from other ethnic family origins. The committee noted that women with an inconclusive or false-positive NIPT result would be offered antenatal anti‑D prophylaxis (that is, they would have the same care as they would have in current practice), and so would not be at a greater risk of sensitisation to the D antigen than women from other ethnic family origins. It noted further that although unnecessary anti‑D immunoglobulin use would be reduced in women of black African family origin, these women would be more likely to have unnecessary anti‑D immunoglobulin than women of white European family origin. The committee concluded that this is a proportionate means of achieving a reduction in anti‑D immunoglobulin use in the population as a whole.

The committee considered the current level of interest in high-throughput NIPT for fetal RHD genotype. It heard from a clinical expert that there have been many enquiries about the test from healthcare professionals and women who would like to have the test but do not live in an area where it has been implemented. It also heard from another clinical expert that the level of knowledge and understanding of NIPT is growing because of the publicity around NIPT for Down's syndrome and other aneuploidies. The committee concluded that based on the current level of interest, the timing was right for making a recommendation of high-throughput NIPT for fetal RHD genotype, but noted that additional data collection from areas beginning to implement the test would help confirm the cost of implementing the test given the uncertainty about this.