Association between risk of infant death and birth-weight z scores according to gestational age: A nationwide study using the Finnish Medical Birth Register.

Objective: To investigate the association between infant mortality and birth weight using estimated fetal weight (EFW) versus birth-weight charts, by gestational age (GA).

Methods: This nationwide population-based study used data from the Finnish Medical Birth Register from 2006 to 2016 on non-malformed singleton live births at 24-41 weeks of gestation (N = 563 630). The outcome was death in the first year of life. Mortality risks by birth-weight z score, defined as a continuous variable using Maršál's EFW and Sankilampi's birth-weight charts, were assessed using generalized additive models by GA (24-27, 28-31, 32-36, 37-38, 39-41 weeks). We calculated z score thresholds associated with a two- and three-fold increased risk of infant death compared with newborns with a birth weight between 0 and 0.675 standard deviations.

Results: The z score thresholds (with corresponding centiles in parentheses) associated with a two-fold increase in infant mortality were: -3.43 (<0.1) at 24-27 weeks, -3.46 (<0.1) at 28-31 weeks, -1.29 (9.9) at 32-36 weeks, -1.18 (11.9) at 37-38 weeks, and - 1.34 (9.0) at 39-41 weeks according to the EFW chart. These values were - 2.43 (0.8), -2.62 (0.4), -1.34 (9.0), -1.37 (8.5), and - 1.43 (7.6) according to the birth-weight chart.

Conclusion: The association between birth weight and infant mortality varies by GA whichever chart is used, suggesting that different thresholds for the screening of growth anomalies could be used across GA to identify high-risk newborns.