Adoption or Placement in Foster Care and Catch-up in Linear Growth and Development: A Meta-Analysis of Individual Participant Data.

The ability of children to recover from linear growth retardation, often referred to as catch-up growth, has intrigued researchers for many decades. Whether adoption from a low-income to a high-income setting, which provides a comprehensive improvement in the conditions that cause children to not grow well, leads to catch-up growth is unknown. We estimated the association of adoption (or placement in foster care) with catch-up in linear growth and child development before 5 y of age. We conducted a 2-stage meta-analysis using individual participant data for linear growth. We obtained study-specific and subgroup estimates and pooled the estimates using random-effects models. Sensitivity analyses were used to assess the robustness of our findings. A review of child-development outcomes was conducted. We included 485 children under 5 y of age from 9 adoption studies. At baseline, children had a mean age of 15.8 mo and a length deficit of 3.9 cm. Adoption reduced this gap by 77% or 3.0 cm (95% confidence interval [CI]: 1.9, 4.1 cm; mean age: 32.3 mo). Catch-up growth was found in both girls (3.6 cm; 95% CI: 2.9, 4.2 cm) and boys (2.5 cm; 95% CI: 1.9, 3.1 cm) and in children adopted after the age of 24 mo (2.2 cm; 95% CI: 0.6, 3.7 cm). The sensitivity analyses did not change any of the substantive findings. The magnitude of catch-up in child development (mean reduction in deficit of 46%) was smaller than that in linear growth. Catch-up in linear growth in children under 5 is biologically possible when the environment is improved profoundly and comprehensively. Partial reversal of the accumulated height deficit is more likely than recovery in developmental outcomes, which highlights the need to ensure all children grow and develop in environments that prevent deficits from occurring rather than trying to correct them. This review was registered at PROSPERO as CRD42022298715 (https://www.crd.york.ac.uk/PROSPEROFILES/298715_PROTOCOL_20220429.pdf).