To what extent does ultrasound estimation of gestational age affect the rate of preterm births in a rural population in sub‐Saharan Africa (sSA)?

Objectives: To assess the impact of ultrasound (US) in resource poor settings, we tested the hypothesis that US estimates of gestational age (USGA) would result in a different preterm birth (PTB: <37 weeks) rate to that derived from clinical estimates based upon menstrual dates and/ or fundal height measurements (CGA) in a rural Kenyan population at high risk of adverse birth outcomes (estimated perinatal mortality = 80-120/1000). Methods: PTB rates according are calculated using USGA and CGA estimates measured for all women attending Kilifi District Hospital for antenatal care at <24 weeks (CGA). Clinical information including the date of the last menstrual period is recorded by government staff as part of routine care. Locally trained sonographers measure fetal crown–rump length (<14 weeks) or head circumference in the trans-thalamic plane (>14 but <24 weeks): USGA is calculated from the mean of 3-blinded measurements using published formulae. Outcomes are determined using the largest demographic surveillance system in sSA. Results:In the first year, 2459 women attended KDH for antenatal care. Of the 856 (35%) eligible women <24 weeks (CGA), 781 (92%) were scanned. From the 592 (76%) who have delivered so far, outcome data are available for 560 (95%). The median and interquartile ranges for USGA and CGA estimates at delivery were 275 (IQR:13.5) days and 276 (IQR:16) days, respectively. CGA classified significantly more PTBs (108/560=19%) than USGA (80/560=14%, p=0.025). Conclusions: Bias associated with CGA estimates significantly impacts upon PTB rates. As one of the largest US dated birth cohorts in sSA, this on-going study will provide a more precise understanding of the epidemiology of PTB.