To investigate the value of intertwin discordance in fetal crown-rump length (CRL) at the 11-13-week scan in the prediction of adverse outcome in dichorionic (DC), monochorionic diamniotic (MCDA) and monochorionic monoamniotic (MCMA) twin pregnancies. This was a retrospective analysis of prospectively collected data on twin pregnancies undergoing routine ultrasound examination at 11-13 weeks' gestation between 2002 and 2019. In pregnancies with no major abnormalities, we examined the value of intertwin discordance in fetal CRL in DC, MCDA and MCMA twins in the prediction of fetal loss at < 20 and < 24 weeks' gestation, perinatal death at ≥ 24 weeks, preterm delivery at < 32 and < 37 weeks, birth of at least one small-for-gestational-age (SGA) neonate with birth weight < 5th percentile and intertwin birth-weight discordance of ≥ 20% and ≥ 25%. First, the study population of 6225 twin pregnancies included 4896 (78.7%) DC, 1274 (20.4%) MCDA and 55 (0.9%) MCMA twin pregnancies. Second, median CRL discordance in DC twin pregnancies (3.2%; interquartile range (IQR), 1.4-5.8%) was lower than in MCDA twins (3.6%; IQR, 1.6-6.2%; P = 0.0008), but was not significantly different from that in MCMA twins (2.9%; IQR, 1.2-5.1%; P = 0.269). Third, compared to CRL discordance in DC twin pregnancies with two non-SGA live births at ≥ 37 weeks' gestation, there was significantly larger CRL discordance in both DC and MCDA twin pregnancies complicated by fetal death at < 20 and < 24 weeks' gestation, perinatal death at ≥ 24 weeks, preterm birth at < 32 and < 37 weeks, birth of at least one SGA neonate and birth-weight discordance ≥ 20% and ≥ 25%, and in MCDA twin pregnancies undergoing endoscopic laser surgery. Fourth, the predictive performance of CRL discordance for each adverse pregnancy outcome was poor, with areas under the receiver-operating-characteristics curves ranging from 0.533 to 0.624. However, in both DC and MCDA twin pregnancies with large CRL discordance, there was a high risk of fetal loss. Fifth, in DC twin pregnancies, the overall rate of fetal loss at < 20 weeks' gestation was 1.3% but, in the small subgroup with CRL discordance of ≥ 15%, which constituted 1.9% of the total, the rate increased to 5.3%. Sixth, in MCDA twin pregnancies, the rate of fetal loss or endoscopic laser surgery at < 20 weeks was about 11%, but, in the small subgroups with CRL discordance of ≥ 10%, ≥ 15% and ≥ 20%, which constituted 9%, < 3% and < 1% of the total, the risk was increased to about 32%, 49% and 70%, respectively. Seventh, in MCMA twin pregnancies, there were no significant differences in CRL discordance for any of the adverse outcome measures, but this may be the consequence of the small number of cases in the study population. In both DC and MCDA twin pregnancies, increased CRL discordance is associated with an increased risk of fetal death at < 20 and < 24 weeks' gestation, perinatal death at ≥ 24 weeks, preterm birth at < 37 and < 32 weeks, birth of at least one SGA neonate and birth-weight discordance ≥ 20% and ≥ 25%, but CRL discordance is a poor screening test for adverse pregnancy outcome. However, in DC twins, CRL discordance of ≥ 15% is associated with an increased risk of fetal loss at < 20 and < 24 weeks' gestation and, in MCDA twins, CRL discordance of ≥ 10%, and more so discordance of ≥ 15% and ≥ 20%, is associated with a very high risk of fetal loss or endoscopic laser surgery at < 20 and < 24 weeks and this information is useful in counseling women and defining the timing for subsequent assessment and possible intervention. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.