OBJECTIVE To use recommendations made by the Evidence-Based Medicine Working Group to determine the accuracy of sonographically estimated fetal weight to diagnose fetal growth restriction (FGR) (estimated or actual birth weight < 10% for gestational age [GA]) or discordancy in dichorionic and monochiorionic twin gestations. METHODS The inclusion criteria for this retrospective analysis were nonanomalous twins, reliable GA and sonographic estimate of fetal weight within 21 days of live births of both fetuses beyond 26 weeks. For likelihood ratio (LR) and odds ratio (OR), 95% confidence intervals (CIs) were calculated. RESULTS Over 52 months, 126 twin gestations met the inclusion criteria; 29% were monochorionic and 71% dichorionic. The incidence of FGR in 1 of monochorionic twin fetuses (46%) was twice as high as in dichorionic (26%; OR 2.01; 95% CI 1.00, 4.03), but the rates of discordance > or = 20% were similar (21% vs. 18%, respectively; OR 1.26; 95% CI 0.64, 2.47). The LR (95% CI) of detecting discordant growth > or = 20% was 6.1 (2.5, 15.1) for dichorionic twins and 6.0 (1.8, 20.0) for monochorionic. The LR of identifying FGR in twin A and B in dichorionic twins was 6.3 (2.0, 19.7) and 8.7 (2.9, 26.9), respectively; the corresponding LR for monochorionic twins was 14.5 (1.9, 112.3) and 18.7 (2.6, 135.1). To determine the number of twins required for a tight confidence interval around an LR of 10, post hoc sample size calculations indicated: (1) regardless of chorionicity, attaining LR > 10 is not feasible; (2) in dichorionic twins, detection of FGR with twin A or B is not possible; and (3) about 500 monochorionic pregnancies are necessary to ensure that estimated fetal weight can accurately identify FGR. CONCLUSIONS Using guidelines proposed by the Evidence-Based Medicine Working Group, at present it may not befeasible to identify aberrant growth in twins.