Trends Over Time in Cognitive Outcomes of Children Born Very Preterm: A Systematic Review and Meta-Analysis

Progress in perinatal care has improved survival for children born very preterm (VPT), but these children remain at higher risk of cognitive impairment compared with children born at term.

To synthesize cohort studies on childhood cognitive ability following VPT birth to investigate trends over time.

All studies from 5 previous meta-analyses of VPT birth and cognition published before 2019 were included, and PubMed, Web of Science, and PsycInfo were searched for new studies published up to June 2024.

Studies reporting IQ scores of children (aged <18 years) born VPT (<32 weeks' gestational age [GA] or birth weight <1500 g) with a term-born comparison group were included.

Two reviewers independently selected studies, extracted data, and evaluated study quality using a modified version of the Newcastle-Ottawa Scale. Unique cohorts were identified to avoid duplicate measures from studies on the same children.

The standardized mean difference (SMD) of IQ scores between VPT-born and term-born children was calculated, and mixed-effects metaregression was used to investigate linear and nonlinear associations between median birth year and the SMD. The main analysis focused on cohorts with IQ measured between 4 and 7 years of age to allow comparison at similar assessment ages. Secondary analyses were conducted in all cohorts using IQ obtained at the latest assessment age.

A total of 257 studies reported data from 131 cohorts of 25 746 individuals born from 1977 to 2016 (15 548 born VPT and 10 198 at term). In the 61 cohorts assessed at age 4 to 7 years (13 842 children born between 1977 and 2014 [8847 born VPT and 4995 at term]; mean [SD] GA, 28.2 [1.7] weeks for the VPT cohorts), IQ was lower for VPT-born children compared with term-born children (SMD = -0.88; 95% CI, -0.97 to -0.79). The linear model showed no association with birth year (β = -0.002; 95% CI,-0.012 to 0.008). Three types of nonlinear models were fit, with no nonlinear associations observed. Adjustment for GA and study characteristics did not change the results (β = -0.001; 95% CI, -0.013 to 0.011). Secondary analysis of 131 cohorts found a similar difference between VPT and term groups (SMD = -0.84; 95% CI, -0.90 to -0.79), with no time trend (β = 0.001; 95% CI, -0.005 to 0.007).

On average, children born VPT had significantly lower IQ scores than term-born children, and this deficit did not decrease in studies conducted over 4 decades.