Ultrasound prediction of fetal macrosomia in pregnancies complicated by diabetes mellitus: a systematic review and meta-analysis

Abstract

Abstract Objectives To report the diagnostic accuracy of ultrasound in identifying fetuses with macrosomia in pregnancies complicated by gestational or pregestational diabetes. Methods Medline, Embase and Cochrane databases were searched. Inclusion criteria were singleton pregnancies complicated by diabetes undergoing third-trimester ultrasound evaluation. The index test was represented by ultrasound estimation of fetal macrosomia (estimated fetal weight EFW or abdominal circumference AC >90th or 95th percentile). Subgroup analyses were also performed. Sensitivity, specificity, positive and negative likelihood ratios, and diagnostic odds ratio were computed using the hierarchical summary receiver–operating characteristics model. Results Twenty studies were included in the systematic review including 8,530 pregnancies complicated by diabetes. Ultrasound showed an overall moderate accuracy in identifying fetuses with macrosomia with a sensitivity of 71.2 % (95 % CI 63.1–78.2), a specificity of 88.6 % (95 % CI 83.9–92.0). The interval between ultrasound and birth of two weeks showed the highest sensitivity and specificity (71.6 %, 95 % CI 47.9–87.3 and 91.7, 95 % CI 86.2–95.5). EFW sensitivity and specificity were 76.6 % (95 % CI 70.1–82.3) and 82.9 % (95 % CI 80.9–84.8), while AC 84.8 % (95 % CI 78.2–90.0) and 73.7 % (95 % CI 71.0–76.4). Conclusions Ultrasound demonstrates an overall good diagnostic accuracy in detecting fetal macrosomia in pregnancies with diabetes.