Background
Large-for-gestational age (LGA) offspring increase the risk of maternal and neonatal birth complications, yet non-standardised definitions of LGA complicate clinical assessment and interpretation of randomised controlled trials (RCTs). The aim of this study was to evaluate how different growth standards defining LGA affected the interpretation of an RCT.
Methods
Data from the Swedish Changing Diagnostic Criteria for Gestational Diabetes (CDC4G) stepped wedge RCT (ISRCTN41918550) were used. Women with risk factors for gestational diabetes (GDM) or random glucose ≥9.0 mmol/l were referred for a 2-hour 75g oral glucose tolerance test (OGTT). The primary aim of CDC4G was to investigate whether the switch from old Swedish (SWE-GDM: fasting ≥7.0 mmol/l and/or 2-hr glucose 8.9-11.1 mmol/l depending on region) to the World Health Organization (WHO-2013) GDM diagnostic criteria reduced LGA risk. Different reference growth charts were used to define LGA (birthweight >90th percentile) and severe LGA (>97th percentile) including a) new Swedish reference ranges for fetal weight (SWE-REF), b) Gestation Related Optimal Weight (GROW) and c) International Fetal and Newborn Growth Consortium for the 21st Century (Intergrowth-21st) reference charts, respectively. Adjusted relative risks (aRRs) with 95% confidence intervals were calculated adjusting for potential confounders using multilevel mixed Poisson regression. Analyses were performed according to the intention to treat (ITT) principle across the whole population and on a subgroup discordant for the definition of GDM (OGTT values above the WHO-2013 diagnostic criteria, but below the former Swedish diagnostic criteria for GDM (SWE-GDM)).
Results
In the total ITT population, the LGA incidence was only reduced after the switch from SWE-GDM (n=26169) to the WHO-2013 GDM criteria (n=28509) when using the SWE-REF definition (aRR 0.92, 95% CI 0.86-0.98). There was no change in severe LGA demonstrated using any of the included reference charts. In the subgroup (n=966 vs 1244), the incidence of LGA was demonstrated to be decreased using both the SWE-REF and the GROW reference charts (aRRs, 95% CIs: 0.68, 0.50-0.91 and 0.78, 0.63-0.95, respectively), but not when using the Intergrowth-21st reference chart. Severe LGA was demonstrated to be reduced only when using the Intergrowth-21st reference chart (aRR 0.73, 95% CI 0.62-0.85).
Conclusion
The choice of LGA definition significantly affected the interpretation of the CDC4G results, with evidence of reduced severe LGA risk using SWE-REF and GROW, but not Intergrowth. Clinical guidelines for GDM should review the emphasis on RCT primary outcome and primary analysis, as definitions may be based upon relatively arbitrary choices.