Postnatal gestational age determination by ultrasonographic measurement of transverse cerebellar diameter: A simulation study

Background: Accurate determination of gestational age (GA) has an important role in clinical practice, especially prevention/anticipation of prematurity-related complications. In low- and middle-income countries, the availability and accuracy of the gold standard for GA determination is very limited. In fact, prenatal ultrasonography in the first trimester is strongly affected by high costs, lack of skills and poor maternal access to health service. Furthermore, the determination by Last Menstrual Period has a very low accuracy, as well as postnatal clinical tools such as the Ballard Score. Recent studies in high-income countries tested TCD measurement by cranial ultrasound for GA determination with reported acceptable accuracy. The development of technology over the years is leading to a progressive spread of cranial ultrasound also in LMICs. Objective: To assess the performance of a published equation in estimating GA using fetal TCD measurements from low-income settings. Methods: Data were simulated using available information from relevant studies conducted in LMICs. Estimated GA was calculated as GA in weeks = 0.470 × TCD in mm + 13.162. The performance was assessed using modified calibration plots and Bland-Altman plots. Pooled bias and 95% limits of agreement (LoA) were also calculated. Results: Four datasets (n = 400, 257, 450, 500) were generated based on data from relevant studies. For GA 23–32 weeks, the equation was prone to overestimate and pooled bias was 6.0 days (95% LoA −6.7 to 18.7 days). For GA 33–36 weeks, the equation was prone to underestimate in one dataset and pooled bias was −1.5 days (95% LoA −16.8 to 13.8 days). For GA 37–40 weeks, the equation showed bad calibration and pooled bias was −5.0 days (95% LoA −25.8 to 15.8 days). When extending the original interval to GA 23−40 weeks, pooled bias was 2.8 days (95% LoA −13.0 to 18.6 days). Conclusions: Within the limitations of a simulation study, our findings suggest that the TCD-based equation may provide some benefits in postnatal estimation of GA between 23–32 weeks in low-resource settings. Lower precision, yet acceptable for clinical use, may be expected beyond 32 weeks’ gestation. Future studies may validate the equation in real newborns in low-resource settings where precise GA dating is available.