Objective To systematically review the literature to assess first trimester ultrasound tests for prediction of TTTS in uncomplicated MCDA twins.

Methods Bibliographic databases were searched and an EndnoteX5 database constructed. The title/abstracts were examined and full manuscripts obtained. Study selection criteria included: MC twins, first trimester test, reference standard-development of TTTS, in trials/cohort studies. Quality assessment was based on QUADAS 2. Data were extracted, and test accuracies, with meta-analysis where appropriate, were calculated (using Meta-DiSc, STATA 11, Statsdirect, p < 0.05).

Results The search yielded 25,140 citations, of which 33 were thought relevant, and 10 were included. Extractable data was available for 4 tests (CRL discordance/NT discordance/NT >95^th percentile/ductus venosus abnormality) in 1451 pregnancies. A “positive test” had a high reliability, but a negative test result did not exclude TTTS (sensitivity of all tests poor- 0.21–0.53). CRL discordance and NT > 95^th percentile had specificities of 92% and gave post-test probabilities for a positive test of 31% and 28% respectively (doubling the pre-test probability). CRL discordance greater than 10% (four studies, 646 pregnancies) appeared to perform the best.

Conclusion This review suggests that routine first trimester ultrasound (ie CRL and NT) could be utilised in screening for TTTS. NICE guidance recommends scanning MC twins fortnight from 16 weeks, so if the screening test was negative these pregnancies would still receive regular follow up. However, as treatment for TTTS is only offered at regional and sometimes supra-regional centres, early referral could be reserved for women whose pregnancies screen positive on first trimester test.