Customized vs population‐based growth charts to identify neonates at risk of adverse outcome: systematic review and Bayesian meta‐analysis of observational studies

• Published in: Ultrasound in obstetrics & gynecology 2017-08, Vol.50 (2), p.156-166
• Main Authors: Chiossi, G., Pedroza, C., Costantine, M. M., Truong, V. T. T., Gargano, G., Saade, G. R.
• Format: Article
• Language: English
• Subjects: adverse perinatal outcome; Bayes Theorem; Bayesian analysis; Birth Weight; Charts; Clinical trials; Customization; customized growth charts; Death; Evidence-based medicine; Female; Fetal Macrosomia; Fetuses; Growth Charts; Humans; Hypoglycemia; Infant; Infant Mortality; Infant, Newborn; Infant, Small for Gestational Age - growth & development; intrauterine growth disturbance; large‐for‐gestational age; Libraries; Medical research; Meta-analysis; Mortality; Neonates; Newborn babies; Observational studies; Polyethylene glycol; Population; population growth charts; Population-based studies; Pregnancy; Pregnancy Outcome; Risk factors; Shoulder; small‐for‐gestational age; Systematic review; Translations
• ISSN: 0960-7692; 1469-0705
• DOI: 10.1002/uog.17381

ABSTRACT Objective To compare the effectiveness of customized vs population‐based growth charts for the prediction of adverse pregnancy outcomes. Methods MEDLINE, ClinicalTrials.gov and The Cochrane Library were searched up to 31 May 2016 to identify interventional and observational studies comparing adverse outcomes among large‐ (LGA) and small‐ (SGA) for‐gestational‐age neonates, when classified according to customized vs population‐based growth charts. Perinatal mortality and admission to the neonatal intensive care unit (NICU) of both SGA and LGA neonates, intrauterine fetal demise (IUFD) and neonatal mortality of SGA neonates, and neonatal shoulder dystocia and hypoglycemia as well as maternal third‐ and fourth‐degree perineal lacerations in LGA pregnancies were evaluated. Results The electronic search identified 237 records that were examined based on title and , of which 27 full‐text articles were examined for eligibility. After excluding seven articles, 20 observational studies were included in a Bayesian meta‐analysis. Neonates classified as SGA according to customized growth charts had higher risks of IUFD (odds ratio (OR), 7.8 (95% CI, 4.2–12.3)), neonatal death (OR, 3.5 (95% CI, 1.1–8.0)), perinatal death (OR, 5.8 (95% CI, 3.8–7.8)) and NICU admission (OR, 3.6 (95% CI, 2.0–5.5)) than did non‐SGA cases. Neonates classified as SGA according to population‐based growth charts also had increased risk for adverse outcomes, albeit the point estimates of the pooled ORs were smaller: IUFD (OR, 3.3 (95% CI, 1.9–5.0)), neonatal death (OR, 2.9 (95% CI, 1.2–4.5)), perinatal death (OR, 4.0 (95% CI, 2.8–5.1)) and NICU admission (OR, 2.4 (95% CI, 1.7–3.2)). For LGA vs non‐LGA, there were no differences in pooled ORs for perinatal death, NICU admission, hypoglycemia and maternal third‐ and fourth‐degree perineal lacerations when classified according to either the customized or the population‐based approach. In contrast, both approaches indicated that LGA neonates are at increased risk for shoulder dystocia than are non‐LGA ones (OR, 7.4 (95% CI, 4.9–9.8) using customized charts; OR, 8.0 (95% CI, 5.3–10.1) using population‐based charts). Conclusions Both customized and population‐based growth charts can identify SGA neonates at risk for adverse outcomes. Although the point estimates of the pooled ORs may differ for some outcomes, the overlapping CIs and lack of direct comparisons prevent conclusions from being drawn on the superiority of one method. Future clini