Objective: To assess if amniotic fluid alpha-fetoprotein (AFAFP) could be useful to determine if both sacs are sampled during an amniocentesis for twin pregnancies. Method: We reviewed all amniocenteses performed on twin pregnancies over a 5-year period. Inclusion criteria were restricted to pregnancies where both karyotypes and AFAFP were available on each fetus. Pregnancies complicated by fetal anomalies were excluded. The following information was obtained: maternal age, gestational age at the procedure, karyotypes, AFAFP values, pregnancy and neonatal outcome. Placental pathology reports were used to confirm chorionicity. Analysis was performed to evaluate the impact of the fetal gender and chorionicity on the AFAFP values. Results: 260 pregnancies were reviewed. Mean maternal age was 36.9 years (33.6, 40.1). Gestational age at the time of the procedure was 16.2 weeks (14.5, 17.9). Complications included 1.8% of misdiagnosis (discrepancy between karyotype and gender). The difference of AFAFP values between the two fetuses was statistically larger in dichorionic pregnancies than in monochorionic gestations. Fetal gender had no influence on the AFAFP. Conclusion: Amniocentesis in twin pregnancies is associated with a 1.8% risk of misdiagnosis. AFAFP can help to assess the chorionicity of a twin pregnancy. When the difference between the two values is <0.2 MoM and the chorionicity was thought to be dichorionic and the two karyotypes are similar, then failure to sample both sacs should be suspected.