Discussion
An increase in clinical intervention by early delivery, including by labour induction, associated with the SGA GTG guideline and the introduction of the GROW package was evidenced by the rapid rise in average monthly induction rate from 21% in January 2015 pre introduction, to 31.5% in January 2016 and a peak of 36.4% in July 2018. An overall increase of 73% more inductions contemporaneous with the introduction of GROW. There may have been other components to this increase but there were no other changes in clinical policies in the unit.
The purpose of SBL was to reduce stillbirths. The Plymouth singleton stillbirth rate had fallen between 2010-13, believed to be due to the introduction of a RCOG Certificate of Completed Training holder (Consultant or Fellow), resident on delivery suite, including every night, for 164 hours a week in 2012. The Stillbirth rate did not show a significant change following the introduction of GROW and the SGA GTG. Introduction of GROW was also contemporaneously associated with a massive rise in the number of (post anomaly) growth ultrasound scans. This put enormous pressure on the obstetric ultrasound service and Fetal Medicine unit who had to provide 40% scans.
The SGA GTG changed the primary aim of fetal wellbeing screening from detection of fetal growth restriction (FGR) to detection of SGA fetuses, despite the guideline identifying that the majority of SGA fetuses will be normal (constitutionally) small. Concentration on detecting SGA fetuses would not identify FGR fetuses who were destined to be in the upper 50 percentiles but were growth restricted to a lower centile, unless they went below the tenth centile. The majority of FGR fetuses will not be detected by simply detecting SGA fetuses.
Since moving to IG21, the induction rate rise stopped, and the numbers of infants allocated as SGA decreased. The question is whether neither GROW or IG21 are “correct”. The UK90 reference ranges were adopted across the UK in the late 1990s 8,10 and were reanalysed in 20093 for inclusion in the UK-WHO growth charts, still the standards in the Personal Child Health Record (red book). There are more recent gender specific UK birthweight centiles11 but these have not been adopted nationally. There is no software package analogous to GROW or IG21, useable on a daily basis to allocate SGA at birth for either UK90 or others. Of the three birthweight references tested, the UK90 gave the best true positive and least false negative rates. It would appear to be the most appropriate for our routine use.
Screening for fetal wellbeing is a complex process. The primary screening tools are patient history and SFH measurement. SFH has been in use for 50 years 11 and there are over 20 uncustomised SFH references in use 12 but only one customised.13 If ultrasound is indicated there are multiple charts for each part of the fetus. In the version of ultrasound database (Viewpoint 5.6.25.281, GE Healthcare, UK) we are currently using there are 17 formulae available for calculating estimated fetal weight (EFW). UPHT policy dictates the Hadlock four parameter (biparietal diameter, head circumference, abdominal circumference and femur length) formula 14 which has been published as the most accurate, although accuracy of EFW is at best ± 10-20%, with a systematic over estimation.15
The GROW process compiles the errors at each step in the process. If a fetus is suspected of being small by SFH, with it’s inter and intraobserver error, an ultrasound is performed. The abdominal circumference and EFW are calculated, each with their own inter and intraobserver error dependant on the formulae used. The EFW is then manually transferred to the GROW SFH/EFW paper chart stored in the patient’s notes, with manual transcription errors combined with plotting errors due to the size of the chart. The GROW EFW has its own error range. At delivery the newborn is weighed and the weight is entered onto GROW which has another reference range with its errors.
When the change was made to the IG21 process, we introduced IG21 SFH charts, changed our ultrasound database to use IG21 fetal biometry references and entered the birthweight into the IG21 software, thus using references all derived from one dataset 16 in the hope of reducing accumulated errors. The finding that the IG21 process introduced a significant a skew to the allocation, albeit in the opposite direction to GROW, was disappointing.
The focus of all the additional work introduced by SBL and SGA GTG was to reduce avoidable stillbirth. An aim all clinicians support fully. The review of 200 consecutive stillbirths, identified that there were similar numbers of stillbirths who were suspected of being large for gestational age as there were who were SGA. In addition several stillbirths were ‘expected’, either as a result of parental choice to interrupt (e.g. delayed selective feticide giving the survivor twin a better chance at later gestation; feticide at <24 weeks but delivery occurring after 24 weeks) or where parental choice was to allow a fetus to die rather than actively end the pregnancy. With 10-12 stillbirths per year a small number of expected cases has a significant effect on the total rate, so the process set in place over the past decade by external guidance, concentrating solely on detection of SGA rather than fetal health will always have limitations in reducing avoidable stillbirths.
Screening for SGA is not a Public Health England Screening programme, unlike screening for trisomies or the anomaly scan. 17
A future larger multicentre study should link neonatal outcome with antenatal policies to determine which predict poor outcomes, rather than simply being sizeist.
This study suggests that maternity units should assess the impact on local intervention rates, clinical outcomes and changes in workload associated with introduction of programmes like GROW and Intergrowth 21, but also hints at the need for a centrally organised, properly funded, PHE specified programme of assessment of fetal growth screening. A central part of such a programme would be validated birthweight centiles. As in the trisomy programme 17, sonographers should be centrally monitored and recertified for biometry measurements. Standardised ultrasound measurements and EFW formulae using appropriate longitudinal statistical analysis18,19 should be used in all units. This approach is already standard for both Crown Rump Length and Nuchal Translucency measurements and there is a national quality assurance programme monitoring the process, with local and national reporting. Surely, if the aim is truly to reduce stillbirth and perhaps further, to improve all fetal heath monitoring, such a national Growth screening programme is now required.