Proactive neonatal treatment at 22 weeks of gestation: a systematic review and meta-analysis

OBJECTIVE: The objective of this study was to provide a systematic review and meta-analysis to quantify prognosis and identify factors associated with variations in reported mortality estimates among infants who were born at 22 weeks of gestation and provided proactive treatment (resuscitation and intensive care). DATA SOURCES: PubMed, Scopus, and Web of Science databases, with no language restrictions, were searched for articles published from January 2000 to February 2020. STUDY ELIGIBILITY CRITERIA: Reports on live-born infants who were delivered at 22 weeks of gestation and provided proactive care were included. The primary outcome was survival to hospital discharge; secondary outcomes included survival without major morbidity and survival without neurodevelopmental impairment. Because we expected differences across studies in the definitions for various morbidities, multiple definitions for composite outcomes of major morbidities were prespecified. Neurodevelopmental impairment was based on Bayley Scales of Infant Development II or III. Data extractions were performed independently, and outcomes agreed on a priori. STUDY APPRAISAL AND SYNTHESIS METHODS: Methodological quality was assessed using the Quality in Prognostic Studies tool. An adapted version of the Grading of Recommendations Assessment, Development and Evaluation approach for prognostic studies was used to evaluate confidence in overall estimates. Outcomes were assessed as prevalence and 95% confidence intervals. Variabilities across studies attributable to heterogeneity were estimated with the I-2 statistic; publication bias was assessed with the Luis Furuya-Kanamori index. Data were pooled using the inverse variance heterogeneity model. RESULTS: Literature searches returned 21,952 articles, with 2034 considered in full; 31 studies of 2226 infants who were delivered at 22 weeks of gestation and provided proactive neonatal treatment were included. No articles were excluded for study design or risk of bias. The pooled prevalence of survival was 29.0% (95% confidence interval, 17.2-41.6; 31 studies, 2226 infants; I-2=79.4%; Luis Furuya-Kanamori index=0.04). Survival among infants born to mothers receiving antenatal corticosteroids was twice the survival of infants born to mothers not receiving antenatal corticosteroids (39.0% vs 19.5%; P<.01). The overall prevalence of survival without major morbidity, using a definition that includes any bronchopulmonary dysplasia, was 11.0% (95% confidence interval, 8.0-14.3; 10 studies, 374 infants; I-2=0%; Luis Furuya-Kanamori index=3.02). The overall rate of survival without moderate or severe impairment was 37.0% (95% confidence interval, 14.6-61.5; 5 studies, 39 infants; I-2=45%; Luis Furuya-Kanamori index=-0.15). Based on the year of publication, survival rates increased between 2000 and 2020 (slope of the regression line=0.09; standard error=0.03; P<.01). Studies were highly diverse with regard to interventions and outcomes reported. CONCLUSION: The reported survival rates varied greatly among studies and were likely influenced by combining observational data from disparate sources, lack of individual patient-level data, and bias in the component studies from which the data were drawn. Therefore, pooled results should be interpreted with caution. To answer fundamental questions beyond the breadth of available data, multicenter, multidisciplinary collaborations, including alignment of important outcomes by stakeholders, are needed.