Customized vs INTERGROWTH-21^st standards for the assessment of birthweight and stillbirth risk at term

Background: Fetal growth abnormalities are linked to stillbirth and other adverse pregnancy outcomes, and use of the correct birthweight standard is essential for accurate assessment of growth status and perinatal risk.

Objective: Two competing, conceptually opposite birthweight standards are currently being implemented internationally: customized gestation-related optimal weight (GROW) and INTERGROWTH-21^st. We wanted to compare their performance when applied to a multiethnic international cohort, and evaluate their usefulness in the assessment of stillbirth risk at term.

Study design: We analyzed routinely collected maternity data from 10 countries with a total of 1.25 million term pregnancies in their respective main ethnic groups. The 2 standards were applied to determine small for gestational age (SGA) and large for gestational age (LGA) rates, with associated relative risk and population-attributable risk of stillbirth. The customized standard (GROW) was based on the term optimal weight adjusted for maternal height, weight, parity, and ethnic origin, while INTERGROWTH-21^st was a fixed standard derived from a multiethnic cohort of low-risk pregnancies.

Results: The customized standard showed an average SGA rate of 10.5% (range 10.1-12.7) and LGA rate of 9.5% (range 7.3-9.9) for the set of cohorts. In contrast, there was a wide variation in SGA and LGA rates with INTERGROWTH-21^st, with an average SGA rate of 4.4% (range 3.1-16.8) and LGA rate of 20.6% (range 5.1-27.5). This variation in INTERGROWTH-21^st SGA and LGA rates was correlated closely (R = ±0.98) to the birthweights predicted for the 10 country cohorts by the customized method to derive term optimal weight, suggesting that they were mostly due to physiological variation in birthweight. Of the 10.5% of cases defined as SGA according to the customized standard, 4.3% were also SGA by INTERGROWTH-21^st and had a relative risk of 3.5 (95% confidence interval, 3.1-4.1) for stillbirth. A further 6.3% (60% of the whole customized SGA) were not SGA by INTERGROWTH-21^st, and had a relative risk of 1.9 (95% confidence interval, 3.1-4.1) for stillbirth. An additional 0.2% of cases were SGA by INTERGROWTH-21^st only, and had no increased risk of stillbirth. At the other end, customized assessment classified 9.5% of births as large for gestational age, most of which (9.0%) were also LGA by the INTERGROWTH-21^st standard. INTERGROWTH-21^st identified a further 11.6% as LGA, which, however, had a reduced risk of stillbirth (relative risk, 0.6; 95% confidence interval, 0.5-0.7).

Conclusion: Customized assessment resulted in increased identification of small for gestational age and stillbirth risk, while the wide variation in SGA rates using the INTERGROWTH-21^st standard appeared to mostly reflect differences in physiological pregnancy characteristics in the 10 maternity populations.