Is ultrasound screening for vasa praevia clinically justified and a financially viable screening test? A literature review

Introduction

Vasa praevia occurs when fetal vessels run through the chorioamniotic membranes, unsupported by the umbilical cord or Wharton’s jelly, close to or over the internal cervical os. ¹ If undiagnosed prenatally, vasa praevia can result in rupture of the fetal blood vessels and exsanguination of the fetal blood volume when the membranes rupture, ultimately resulting in fetal death. ² As fetal blood volume towards term is so small, even the loss of a small amount can be disastrous. In fetuses that survive delivery, fetal anaemia requires immediate rectification on emergency delivery. ³ If undetected the associated fetal mortality rate is 44% whereas if antenatally detected, the fetal survival rate is noted as 97%. ⁴

The condition is rare with reported incidence rates in the literature varying between one in 2500 ⁵ to one in 6000. ⁴ However, the incidence rates may be under-reported due to the difficulty in prenatal diagnosis and the inability to accurately diagnose once delivered. ¹ Vasa praevia can be diagnosed prenatally with ultrasound; however, current guidelines within the UK do not support screening for this condition. ⁶

There are two types of vasa praevia, type 1 (velamentous cord insertion) and type 2 (bilobed placenta) which are reported within the literature as being associated with certain risk factors, namely low lying placenta/placenta praevia, bilobed placenta, twin gestations, assisted reproduction techniques, antepartum haemorrhage (APH) and velamentous cord insertion. ⁷

Within the UK there is a current focus on improving pregnancy outcomes, reducing the stillbirth and neonatal death rate through the ‘saving babies lives’ ⁸ care bundle and the MBRRACE report. ⁹ As vasa praevia undiagnosed is known to be associated with adverse outcomes, this review was conducted to assess the current evidence on vasa praevia detection to assess whether screening is clinically justified or financially viable.

National and international guidelines

Literature search

A literature search was conducted using the University of Derby EBSCOhost multiple databases, ScienceDirect and the Cochrane library. The search was restricted to English language and to articles published after 2000. These restrictions were placed due to time restrictions of the review and also to ensure the most recent relevant literature was included. The search terminology used included variations on vasa praevia, ultrasound and Doppler (including pulsed wave and colour flow Doppler).

The literature search identified a total of 44 articles across the databases. After duplicates were removed, articles were reviewed for relevance by initially reading the abstract to assess the content of the literature. If the article was not assessing the ultrasound detection of vasa praevia then it was rejected. After further review of the literature content for quality of the research conducted, a total of 13 research articles remained. Figure 1 demonstrates the process of identifying articles and exclusions to reach the final group of literature for review. OPEN IN VIEWER

Of the remaining 13 articles there was one prospective review, four were retrospective cohort reviews and five case reviews which although they do not provide primary research data, do provide some insight into specific cases. Two further articles were also included which specifically examined the identification of a velamentous cord insertion. Only one article was identified which focussed on cost-effectiveness and so this was included but reviewed separately due to the difference in information provided.

Review

Four retrospective and one prospective study of prenatal diagnosis of vasa praevia are included in this review.^{13,15– 18} Table 1 provides a summary of the studies included and how cases were identified.

OPEN IN VIEWER

Table 1.

Summary of the reviewed literature

Author	Design	Inclusion	Population	Cases of vasa praevia	Screening test	Reference
Lee et al. 15	Retrospective cohort study	All deliveries between January 1991 and December 1998	93,874	18	Standardised greyscale imaging of cervix. Transvaginal scan and colour Doppler if clinically indicated	Placental pathology on examination
Rebarber et al. 16	Retrospective cohort study	All referred patients between June 2005 and June 2012	27,573	32	Standardised greyscale imaging. Transvaginal imaging in suspected cases and high-risk group. Standard colour flow Doppler and pulsed wave Doppler	Confirmation at delivery and on placental pathologic reports
Bronsteen et al. 17	Retrospective cohort study	All deliveries with vasa praevia diagnosis between January 1990 and June 2010	Not stated	60	Transabdominal imaging of the lower segment. Placental cord insertion site imaged in twin gestations. Transvaginal and colour flow Doppler imaging utilised if clinically needed	Delivery confirmation and placental pathology reports
Oyelese et al. 13	Retrospective chart review	Cases registered with the Vasa Praevia Foundation and chart review of patients identified as discharged with vasa praevia diagnosis	Not stated	155	No standardisation of screening due to variety of case sources	Delivery record and placental pathology reports
Catanzarite et al. 18	Prospective cohort study	All deliveries between January 1991 and December 1998	33,208	10	Greyscale imaging over lower segment. Colour Doppler when vessels over cervix. From 1993 onwards identification of placental cord insertion site. Transvaginal sonography when clinically needed	Obstetrician’s operative findings at caesarean section and placental pathological reports

A total of 265 cases of vasa praevia were identified across the four retrospective studies ranging between 18 cases ¹⁵ and 155. ¹³ With all four retrospective reviews the purpose of the studies was to assess the prenatal detection of vasa praevia and how this affected the patient outcome and management. In all four studies, the diagnosis was confirmed on visual and/or pathologic examination of the placenta and membranes post-delivery.

Three of the retrospective studies described the ultrasound technique used within the centres to identify vasa praevia present. ^{15 – 17} Two studies routinely used transabdominal scanning of the lower uterine segment with the use of colour flow Doppler to assess for fetal blood vessels in the lower segment however only used transvaginal ultrasound if there was a clinical need to assist in the view of the cervix.^15,17 The study by Rebarber et al. ¹⁶ however, used a standard screening protocol which incorporated the routine use of transvaginal ultrasound if the patient had a defined risk factor for vasa praevia.

In two studies the natural progression of vasa praevia was also examined through serial scans to assess the resolution of the condition through gestations.^15,16 The resolution of vasa praevia was seen in three of 18 ^{1 5} cases and five of 32 ^{1 6} cases showing the need for serial scans after diagnosis. Further to this, Rebarber et al. ¹⁶ noted that the resolution of vasa praevia was seen in 23.8% of cases diagnosed during the second trimester; however, if the vasa praevia was diagnosed in the third trimester then no resolution was identified.

Catanzarite et al. ¹⁸ conducted a prospective cohort study with selection of vasa praevia cases to assess follow-up, outcome and management of prenatal diagnosis. The study was conducted on a cohort of women who were already identified as high risk due to being referred to the authors’ centre with a placental abnormality. ¹⁸ Sonographic findings were used to identify 11 prenatally diagnosed cases of vasa praevia from a cohort of 33,208. Throughout the course of this study, the technique used with ultrasound to detect the condition evolved as technology advanced from using colour flow Doppler over the cervix in the early days of the study to identification of the placental cord insertion site and sweep of the lower segment with colour flow Doppler in the latter stage. ¹⁸ Transvaginal ultrasound was only used when needed to confirm the vasa praevia as noted by the authors but never in the presence of vaginal bleeding. ¹⁸ As was described by the retrospective studies, Catanzarite et al. ¹⁸ noted the surveillance of vasa praevia and scanning in the third trimester to confirm the diagnosis. There is no mention of any cases of resolution though during the course of the study.

Of the 11 cases prenatally diagnosed, 10 were confirmed in the prospective study by examination of maternity records and pathological examination of the placenta and membranes at delivery. ¹⁸ One case which was prenatally diagnosed, at delivery was found to be a marginal placenta praevia rather than vasa praevia. ¹⁸ Catanzarite et al. ¹⁸ also commented on two potential missed cases of vasa praevia which had normal prenatal ultrasounds; however, in both cases there was fetal demise at delivery after membrane rupture followed by blood. In both cases, a velamentous cord insertion was identified on examination of the placenta and membranes.

Five articles were found discussing case reports of 10 cases of vasa praevia of which eight were prenatally diagnosed, two were discovered post-delivery and had resulted in neonatal death of the fetus. ^{19 – 23} The case reports discuss differing techniques used to prenatally diagnose the vasa praevia; however, as these case reports describe individual cases, no standard protocol is discussed. The diagnostic technique used varied between transabdominal scanning with the use of colour flow Doppler in the uterine segment ¹⁹ to transabdominal scanning with colour flow Doppler and three-dimensional imaging ²¹ and then transabdominal, transvaginal and colour flow Doppler.^19,20 Two of the authors discussing case reports do not discuss what ultrasound scanning protocol was used on each of the patients so no conclusion can be made.^22,23

Results from the retrospective reviews found that of the 265 cases reported, 59 intrapartum or neonatal deaths occurred due to the vasa praevia presence.^13,15,17 Oyelese et al. ¹³ also reported a pregnancy which resulted in a still birth; however, this was not thought to be linked to the diagnosis of vasa praevia. In the review by Rebarber et al., ¹⁶ all pregnancies resulted in live births of healthy fetuses; however, two cases had unknown outcomes due to loss of medical records. In the prospective review, all 10 cases of prenatally diagnosed vasa praevia resulted in the live birth of healthy fetuses. ¹⁸ These findings are in keeping with the findings of the retrospective reviews by Rebarber et al. ¹⁶ and Bronsteen et al. ¹⁷ who found that when prenatal diagnosis was made, patient management subsequently led to earlier delivery of a healthy fetus. Likewise these echo the results by Oyelese et al. ¹³ who found that when prenatal diagnosis was made, the survival rate was much greater at 97% whereas when not prenatally diagnosed the survival rate was greatly reduced at 44%.

Within the series of case reports, similar survival rates were identified with all cases which were prenatally diagnosed resulting in a live birth of a healthy fetus. ^{19 – 22} In the two case reports where vasa praevia was not diagnosed, both authors report the rupture of membranes followed by vaginal bleeding and clots and fetal bradycardia. Emergency caesarean section was performed in both cases; however, due to loss of blood and pale floppy, anaemic fetus this resulted in the neonatal death on day 1 in both cases.^22,23

The selected studies were limited in the data available for review; therefore, no statistical analysis could be completed. It should also be acknowledged that of the studies reviewed, the authors note that there is likely to be some degree of bias present due to the motivation to conduct the research and case selection within a ‘high risk population’.^13,16 This combined with changing technology and ultrasound methods may therefore have affected the validity of the studies reviewed.

Patient management

Management of a pregnancy with vasa praevia is discussed across the retrospective and prospective reviews with varying strategies identified as having occurred. Elective admission at a mean gestational age of 33 weeks was found in the study by Rebarber et al. ¹⁶ so that antenatal steroids could be administered to enable elective early delivery by caesarean section. This study also noted that twin gestations were admitted at an earlier stage of 28–30 weeks; however, this was not quantified with an explanation of the reasoning. Emergency admission and delivery was also noted as significant with six of the 32 cases resulting in this due to APH and preterm labour. ¹⁶ None of the other retrospective reviews discuss elective admission or timing of admission prior to delivery. Elective caesarean section is noted as being significant for improving outcomes when prenatally diagnosed in the further two studies with the mean gestational age of elective delivery by caesarean being 35.4 weeks and 34.9 weeks.^13,15 The prenatal diagnosis does not, however, stop the need for emergency admission and delivery with Bronsteen et al. ¹⁷ noting seven emergency deliveries, 11 urgent deliveries and 11 cases of spontaneous labour prompting the need for delivery and in the series by Oyelese et al. ¹³ noting that 80 emergency caesareans were performed in the case series described. In the prospective review by Catanzarite et al., ¹⁸ similar findings on elective delivery are noted with two deliveries being emergency prior to confirmation of lung maturity and overall the deliveries being between gestational age of 31.5 and 37.8 weeks. No detail is, however, provided on the optimum management strategy for admission and elective delivery in prenatally diagnosed cases. This is also in keeping with the current UK guidelines which acknowledge that there is no agreed management strategy where vasa praevia is detected and this could lead to more patient anxiety and unnecessary early caesarean deliveries. ¹¹

Lee et al. ¹⁵ discussed the possibility of missed cases within the review, noting that 81 sets of patient notes which were reviewed identified the pregnancy as having either vasa praevia or a velamentous cord insertion; however, only one case is identified by the authors as being possibly a missed vasa praevia. Two missed cases are discussed by Catanzarite et al. ¹⁸ which resulted in fetal demise and in both cases the authors reviewed the ultrasound imaging and found that the vasa praevia was identifiable on the ultrasound videotape. False positive diagnosis is discussed by two authors reporting five cases and one case of false positive; however, no further information is provided on the nature of the false positive so this cannot be expanded further.^16,17

Risk factors

Throughout the literature, various risk factors are identified as being highly indicative for vasa praevia with placenta praevia and APH as being the most supported risk factors.^{13,15– 18} Table 2 provides a summary of the incidence of risk factors reported in the literature.

OPEN IN VIEWER

Table 2.

Summary of the incidence of risk factors

Author	Low lying/ placenta praevia	Bilobed/ succenturiate lobe	Twin gestation	Velamentous/ marginal cord insertion	IVF	Antepartum haemorrhage
Lee et al. 15	✓ (44%)	✓	✓ (17%)	✓		✓ (33%)
Rebarber et al. 16	✓ (78%)	✓ (17%)	✓ (24%)	✓ (72%)	✓ (41%)	✓
Bronsteen et al. 17	✓ (62%)		✓	✓		✓
Oyelese et al. 13	✓ (61%)	✓ (32%)	✓ (5%)		✓ (10%)	✓ (36.1%)
Catanzarite et al. 18	✓	✓		✓		✓ (80%)

Cost effect

Only one piece of literature examined the cost effect of screening strategies based on population figures for a one-year cohort of women. ²⁵ The results of this analysis concluded that in singleton pregnancies, a screening strategy of performing a transvaginal scan on all women with a high-risk indicator would result in 8726 extra women having at least one transvaginal ultrasound scan during the course of the pregnancy. ²⁵ In the context of ultrasound departments, this would result in increased examination times, additional examinations and increase in the demand for ultrasound. The authors also note that this would result in 33 more caesarean sections but with a reduction in the number of emergency caesarean sections and would ultimately reduce the number of late fetal or neonatal deaths by 19, whereas universal screening would result in 23 fewer deaths. ²⁵ The resultant cost effect of a screening strategy of high-risk singleton gestations is estimated as $3.4 million, of which a majority ($2.2 million) would be due to the cost of screening and training within the first year whereas universal screening for singletons would dramatically increase costs by $10.6 million for each cohort. ²⁵ This would result in substantial financial implications for hospitals due to the increased demand for ultrasound scans and through the management of the pregnancy, early delivery and possible involvement of the neonatal units. ²⁵ Cipriano et al. ²⁵ rightly noted that although this strategy would reduce the number of late fetal and neonatal deaths, the rate of maternal deaths and complications would potentially increase per year due to more interventions being performed and the increased rate of caesarean sections being completed. This is one consideration which is not noted by any of the authors previously discussed.

In agreement with the UK National Screening committee review in August 2017 and 2013, Cipriano et al. ²⁵ concluded that when all options are considered, the most cost-effective strategy would be to not screen at all; however, in their view this would be negligent.^10,11 Targeted screening of twin gestations was found to be cost effective as it would reduce the number of late fetal and neonatal deaths per year. ²⁵ Providing targeted screening for women with high-risk factors and all twin gestations would provide the optimum screening strategy as the majority of these women would already be accessing ultrasound services due to the risk factors and thus this would have the least impact financially. ²⁵

Discussion

All studies discuss the routine use of greyscale imaging with the use of transvaginal imaging and colour flow Doppler appearing to be intermittent. Only one study discusses a standard protocol which utilises transabdominal, transvaginal and colour flow Doppler techniques. ¹⁶ The use of colour flow Doppler does not appear to be a standard technique but this is likely due to changes in technology during the course of these studies which may have contributed to the lack of use of colour flow Doppler in earlier studies.^14,15

The two studies by Bronsteen et al. and Catanzarite et al. both comment on the identification of the placental cord insertion site but note that this is not standard practice.^17,18 When discussing the techniques for screening for vasa praevia though, identification of the placental cord insertion site should be considered as this could be used to eliminate the possibility of a velamentous cord insertion. This is not specifically discussed by these studies though.

Sepulveda et al. ²⁶ specifically discussed the detection of velamentous cord insertion as a screening method for vasa praevia. A prospective cross sectional study reported successful identification of the placental cord insertion site in 99% of cases in less than 1 minute. ²⁶ The authors used a standardised protocol using greyscale and colour flow Doppler techniques and suggested that this would be an effective technique which would not impact on the ultrasound examination time or method. ²⁶ The study concluded that a velamentous cord insertion was present in 1% of cases reviewed; however, no cases of vasa praevia were identified even though placenta praevia was present in 2% of the cases examined. ²⁶

The authors acknowledge that the rate of visualisation of the placental cord insertion site may have been higher as the single operator was a fetal medicine specialist who had training in umbilical cord sampling. The authors believe though that the visualisation of the placental cord insertion site does not require specialist training as long as the screening is performed in a systematic way. ²⁶ It could be argued that within an obstetric ultrasound department within the UK, these detection rates may not be achievable as it is not an area that is routinely viewed during the ultrasound examination. This would need subsequent training of sonographers, obstetricians, fetal medicine specialists and would also have a potential impact on a need to change theoretical and practical experience provided by university institutions to ultrasound students. Sepulveda et al. ²⁶ acknowledged that a previous case of velamentous cord insertion with an adverse outcome had instigated this study being performed and so this may have influenced the study as the operator would most likely have been motivated to try and make every effort to identify the placental cord insertion due to previous experiences and the potential outcome if missed.

Ebbing et al. ²⁷ conducted a retrospective population-based review of births registered as having a velamentous cord insertion present and found that a velamentous cord insertion was present in 1.5% of pregnancies which is in line with the previous study. ²⁷ This figure increased to 5.9% in multiple gestations however was not found to increase in the case of pregnancies conceived by assisted reproduction techniques. ²⁷ In patients who had experienced APH, this doubled the risk of there being an anomalous cord insertion present supporting the notion that APH should be seen as a warning sign for vasa praevia.^1,27

Several authors discuss various techniques of screening for vasa praevia and due to the variation with no standardisation between them, this makes it impossible to provide any comparative data or even sensitivity or specificity for the techniques used. ^{15 –18,26} The majority of studies discussed do advocate some form of targeted screening should be adopted to try and improve outcomes. The recommendations of each study can be seen in Table 3. Universal screening is also recommended by several authors; however, it should be considered the financial impact that this would have, not just for the patient management but also the large impact on staff and student training and implementation of some form of quality assurance programme. The most cost-effective strategy as identified by Cipriano et al. ²⁵ would be to provide targeted screening as this would cost the least per cohort whilst still preventing more late fetal or neonatal deaths than performing no screening at all.

OPEN IN VIEWER

Table 3.

Summary of recommendations

Author	Year	Recommendation
Lee et al. 15	2000	No recommendations made
Rebarber et al. 16	2014	Targeted screening for high-risk patients with transvaginal scanning, greyscale and colour flow Doppler
Bronsteen et al. 17	2013	Targeted transvaginal ultrasound of high-risk women
Oyelese et al. 13	2004	Placental cord insertion ultrasound should be included in second trimester scan. Targeted transvaginal ultrasound for high-risk patients
Catanzarite et al. 18	2001	Combined transvaginal ultrasound in second trimester with targeted screening for resolving placenta praevia or bilobed placenta
Sepulveda et al. 26	2003	No recommendations made
Ebbing et al. 27	2013	More investigation needed on ascertaining if identification of placental cord insertion site can reduce the risk of adverse outcome. No recommendations regarding ultrasound screening
Cipriano et al. 25	2010	Targeted transvaginal screening in high-risk group (multiple gestation, low lying placenta, IVF, accessory lobes, velamentous or marginal cord insertion)
Al-Khaduri et al. 22	2007	Targeted transvaginal ultrasound with colour flow Doppler for IVF gestations
Ghandi et al. 21	2008	Targeted screening for twins, low lying placenta or conceived with assisted reproduction techniques

What should also be noted at this point is the results of a systematic review conducted by Ruiter. ²⁸ The aim of the review was to assess the accuracy of ultrasound diagnosis of vasa praevia and covered some of the same research as this review. The authors found that there were similar concerns over the bias within the literature and validity of results as has been found in this review. What was concluded though was that ultrasound can be accurate in detection of vasa praevia if a transvaginal scan is performed with the addition of colour Doppler. ²⁸ No consideration is given to the process of screening and whether this should be universally provided to all pregnant women or targeted to the patients who are at high risk. It is acknowledged that further research is needed to provide evidence on the risk factors as well as the potential of whether an early caesarean section is the correct management pathway. ²⁸

Conclusion

Within the UK, screening for vasa praevia is not supported as there is uncertainty regarding the epidemiology of the condition.^{6,10,11,20,29} This together with discrepancies between authors over what is classed as a high risk makes it difficult to find a consensus on screening strategy and patient pathways.^25,30 It is evident that placenta praevia and APH are particularly important risk factors for the condition.^15,17,18

Consideration has to be given to the point that screening for vasa praevia is an emotive subject and people are likely to be motivated to implement screening if they have been personally affected by a case with an adverse outcome. It is also difficult to conduct thorough research within this field as randomised controlled trials would not be ethical. ³¹ Consequently, it is difficult to conclude whether it is clinically justified to perform screening for vasa praevia as a technique and management strategy is not clearly defined. What is apparent though is that when a prenatal diagnosis is made, follow-up is important to assess for resolution of the condition.

When acknowledging that screening can help save lives, even with the most standard protocol cases would still be missed. Therefore, further research is needed to assess any impact that implementation of screening for vasa praevia would have and clarification is also needed of the proportion of ‘high-risk pregnancies’ that are present in within a general UK population. ¹⁸

It is clear from the evidence reviewed that prenatal diagnosis is crucial in order to improve outcomes and fetal deaths are mostly preventable with timely management and delivery.^29,32 Given that there is a current focus in the UK to improve the stillbirth and neonatal death rate with care strategies being implemented such as the ‘saving babies lives’ ⁸ care bundle and the MBRRACE report, ⁹ it could be argued that research should be conducted to understand the cause and identification of this condition in pregnancy.

With awareness of vasa praevia increasing over recent years, there will inevitably be incidental findings where no screening or management strategy is in place. This therefore warrants discussion. If prenatal diagnosis can make a significant impact on pregnancy outcomes, then professionals, sonographers and ultrasound departments have a duty of care to be doing what they can to detect this condition and potentially change the outcome for the fetus and mother.