Screening for a potentially lethal disease that is not clinically apparent seems intuitively worthwhile. However, the controversies regarding screening for presymptomatic breast or prostate carcinoma, which have raged both in the popular and medical media, illustrate that this can be a vexed subject. The journey to date for prenatal screening for major congenital heart disease (mCHD) has likewise not been straightforward.

When prenatal screening for mCHD was first suggested more than 30 years ago, it was anticipated that high-risk pregnancies, such as those with family history of heart defects, could be targeted.1 However, it soon became evident that most mCHD detected prenatally was to be diagnosed because the front-line fetal ultrasonographer suspected a cardiac abnormality at the mid trimester anomaly scan in pregnancies not previously designated as high risk.2 Population-based screening was advocated, initially on the basis of the four-chamber view of the fetal heart, obtained from a transverse imaging section of the fetal thorax, with the subsequent addition of outflow tract views in more recent guidelines.3

Prenatal detection of mCHD confers two potentially beneficial opportunities:

• provision of time for preparation by parents and professionals, including the consideration of termination of pregnancy, and

• perinatal management of the neonate with mCHD in an environment where the required expertise is available.

The value of counselling to inform considered decision-making, which includes options of invasive fetal chromosome analysis, termination of pregnancy and plans for postnatal management, should not be underestimated. It is notable that in the prospective study by Gardiner et al,4 28/41 (68%) of mothers chose termination of pregnancy after receiving a fetal diagnosis of fetal hypoplastic left heart syndrome (HLHS). In a subset of cases, this decision was made in the context of associated aneuploidy or extracardiac malformation.

Plans for postnatal management of prenatally diagnosed mCHD requires careful thought in order to strike the right balance between patient safety and unnecessary intervention. A recent French study demonstrated a greatly increased incidence of caesarean section after prenatal diagnosis of complete transposition of the great arteries (TGA), without improved neonatal outcome.5 With a prospective management plan in place, it may be safe and effective for neonates with mCHD to deliver normally in a selected obstetric unit located closer to the family home, with subsequent transfer of the newborn to the tertiary centre.6 An exception is TGA, where urgent treatment for hypoxaemia with a Rashkind balloon atrial septostomy may be required immediately after birth,7 and therefore delivery in close proximity to facilities for this is advisable. A recent study from Texas incorporating data from all hospitals statewide, which represents a relatively large and varied geographical area, demonstrated that birth at closer proximity to the tertiary centre benefitted neonates with HLHS whether or not they were diagnosed prenatally.8 One interpretation of these data is that they support a close collaboration between tertiary centres and hospitals within their referral networks, such that pathways of care are put in place, including the selection of where births of neonates with prenatally diagnosed mCHD will occur as well as their postdelivery management plans, all with a common goal of ensuring optimal patient outcomes.

Evidence indicating that prenatal diagnosis benefits outcome is predominantly indirect, drawn from single-centre reports demonstrating that prenatally diagnosed neonates are in a more stable condition prior to undergoing essential cardiac surgery,7 ,9–11 and that in turn avoidance of cardiovascular instability from mCHD through planned postnatal management leads to a reduction in early perioperative neurological insults12 and to better long-term neurodevelopmental outcome.13

Early survival rates for mCHD are now generally excellent,14 and although audit of 30-day mortality rates remains an important component of quality assurance for congenital heart programmes, there is strong interest in shifting the focus of audit towards a wider range of metrics, including process measures, among them, the rate of prenatal diagnosis, as well as further clinical measures reflecting functional outcome.15 ,16 The paper by Gardiner et al provides important contemporary information about the accuracy of prenatal diagnosis of mCHD in the UK with different models of screening. Their manual linkage of maternal to postnatal infant data is commendable. The suggested mandatory use of maternal National Health Service (NHS) number to link maternal data to infant outcomes is potentially very valuable. The unique NHS number is an underused resource, not just for audit but also for clinical management when patients move between different health providers.

Gardiner et al report rates of prenatal detection for two sentinel abnormalities, TGA and coarctation of the aorta (coarctation). These defects require use of outflow tract fetal cardiac views rather than fetal cardiac screening confined to the four-chamber view and are lesions where neonatal intervention is required, with susceptibility to circulatory collapse if undiagnosed in the presymptomatic phase. In the university-based fetal medicine unit with fetal cardiologists on site (hospital ‘A’), TGA was diagnosed in 7/9 (78%) and coarctation in 9/11 (82%) of cases. By comparison, the rate of diagnosis of TGA was 3/11 (27%) and coarctation 2/13 (15%) in two maternity hospitals linked to the referral centre (hospitals ‘B’ and ‘C’). These findings suggest that both operator expertise and continuous on-site training are key to promoting performance approaching the optimum achievable.

Population screening for mCHD is challenging, as evidenced from the variable rates for successful prenatal detection of mCHD across different healthcare systems. In the UK, the rate of prenatal diagnosis for mCHD types that led to intervention in infancy is currently 35%.14 Statewide audit found a similar rate of prenatal diagnosis in Utah, USA, at 39%.17 A statewide review of a range of selected types of mCHD from Victoria, Australia, reached a higher rate of prenatal diagnosis at 53%.18 Remarkably, a national programme of prenatal diagnosis reported a rate of 80% in the Czech Republic,19 which is a similar rate of prenatal diagnosis to that reported by hospital ‘A’ in Gardiner et al.

How could this outstandingly high level of prenatal diagnosis be achieved across all NHS obstetric units?

Fetal ultrasonographers would require the necessary training and support required to detect all, or nearly all, of the anticipated 3.5 cases of mCHD occurring per 1000 unselected mid trimester fetal anomaly scans. This is an ambitious target, demanding strong links between the tertiary paediatric cardiac units and their networks to support it. Shared care, which includes training of ultrasonographers in diagnosis of fetal mCHD, perinatal joint management after prenatal diagnosis of mCHD and long-term follow-up in local outreach clinics in the ‘hub and spoke’ network of congenital cardiac care, seems the appropriate model to pursue. Rates of prenatal diagnosis require an ongoing transparent process of audit, which is more likely in the future given the presence of this process measure on the congenital heart diseases ‘dashboard’ provided by NHS England. In order to complete the audit cycle, healthcare providers (both obstetric units and tertiary paediatric cardiac centres) should receive feedback on their performance in terms of their rates of prenatal diagnosis, alongside annual targets for improvement.