Objective
Compare the performance of first trimester ultrasound biparietal diameter (BPD) screening for open spina bifida (OSB) when BPD is adjusted for crown‐rump length (CRL) or abdominal circumference (AC).
Methods
For 63 OSB and 24 265 unaffected pregnancies, BPD was expressed as multiple of the normal median (MoM) based on CRL and on AC, and as the ratio BPD/AC. Screening performance was assessed by the Mahalanobis distance, the observed detection rate with normal fifth and 10th percentile cut‐offs and the area under the receiver‐operator characteristic curve (AUC).
Results
Mahalanobis distance for BPD MoM was considerably higher when based on AC than on CRL: 1.69 versus 1.14. Screening performance was also higher: using a fifth percentile cut‐off, the detection rate was 59% compared with 41%; using a 10th percentile cut‐off, the rates were 63% and 51%. Whilst the false‐positives rates were slightly higher too—5.3% versus 5.1% and 10.8% versus 9.9%—the AUC was statistically significantly higher: 0.872 (95% CI, 0.816‐0.928) compared with 0.735 (95% CI, 0.664‐0.806). BPD/AC had intermediate performance.
Conclusion
The best results are obtained when AC, rather than CRL, is used to express BPD values in MoMs. First trimester OSB screening can detect half to two‐thirds of cases.