Selective versus universal third trimester ultrasound: Time for a rethink? An audit of current practices at a metropolitan Sydney hospital

Abstract

Introduction Studies supporting risk assessment-based over routine third trimester ultrasound (T3US) may not be applicable to current Australian practice. We assessed incidence, indications, and utility of singleton T3US at St George Hospital (SGH), Sydney. Methods Singleton births October-December 2012 (n = 623) underwent retrospective file review. Singleton pregnancies booked February-April 2013 with continued care at SGH (n = 440) were assessed prospectively via clinician survey and file review. Data collected included maternal demographics, pregnancy risk factors, first and second trimester ultrasound results, whether T3US was performed, indications for and results of T3US, and impact on clinical practice. Obstetric and neonatal outcomes were also collected. Results After excluding 30 cases with inadequate data, 54.2% (560/1033) had at least one T3US. 82.7% had protocol supported/ medical indications. The remainder had no documented indication (7.0%), an indication outside SGH protocol (e.g. 'borderline' low-lying placenta, 4.1%), or was at maternal request (1.4%). Main indications for initial T3US were low-lying placenta (18.8%), reduced fundal height (10.2%), and morphology scan abnormalities (8.6%). 28.9% (162/560) of first T3US were abnormal, with 64.2% resulting in a clinical action (repeated/more specialised ultrasound, change in mode of care/delivery, paediatric follow-up). 129 (12.6%) had two or more T3US, with the majority for the same indication as the first (43.9%) or required to follow-up abnormalities from the first T3US (36.2%). Repeated T3US were more frequently performed at specialised centres (18.1% for 1st T3US, 25.6% for 2n4 T3US and 34.2% for 3r4 T3US, P = 0.015). While no (0/32) small for gestational age at birth (SGA) cases were identified in the no T3US group, T3US identified 22% (8/37) fetuses with SGA <10t8 centile (P = 0.04 versus no T3US) and 44% (14/32) <3rd centile (P = 0.004 versus no T3US). Of high-risk women requiring a protocol-driven T3US, 98% (43/44) with gestational insulin-requiring diabetes, 92% (51/ 56) with body mass index ≥35 kg/m2 and 75% (27/36) with maternal age ≥40 received one. There were no significant differences between retrospective and prospective cohorts apart from midwives ordering a larger percentage of initial T3US in the prospective cohort (69.2 versus 51.1% P < 0.001). Conclusions The majority of singleton pregnancies at SGH receive at least one T3US, most for accepted medical indications. Many abnormal ultrasound results changed clinical management, but also incidentally detected abnormalities requiring follow-up. SGA identification was superior in women receiving T3US but still <50%. Future multi centre research would help to determine whether routine T3US improves outcomes and is cost-effective in current clinical practice.