For the purpose of improving the clinical efficacy of alpha-fetoprotein (AFP)-L3% in prenatal screening for trisomy 21, we calculated the multiple of the median (MoM) of AFP-L3% (L3 MoM) and the ratio of L3 MoM to AFP MoM (L3 MoM/AFP MoM) in maternal serum. Maternal serum samples from 1822 women (maternal age 37.3 +/- 3.8 years, and weeks of gestation 16.0 +/- 1.0; mean +/- SD) with unaffected pregnancies and 28 women (37.6 +/- 4.6 years, 16.6 +/- 3.1) pregnant with of trisomy 21 fetuses were obtained. The AFP concentration and AFP-L3% in maternal serum were measured using a liquid-phase binding assay. The areas under the receiver operating characteristic curves (AUCs) of AFP MoM, AFP-L3%, L3 MoM, and L3 MoM/AFP MoM were 0.750, 0.868, 0.949 and 0.946, respectively. The AUCs of L3 MoM and L3 MoM/AFP MoM were significantly higher than AFP-L3% (P < 0.05) and AFP MoM (P < 0.0005). However, no statistical difference was observed between the AUCs of L3 MoM and L3 MoM/AFP MoM. In conclusion, the L3 MoM should be an effective replacement for AFP-L3% in prenatal trisomy 21 screening.