To evaluate the distribution of first- and second-trimester maternal serum markers in twin pregnancy with and without pre-eclampsia.Methods
One-hundred and forty-four twin and 109 unaffected singleton pregnancies were recruited from the same institution. First- and second-trimester maternal blood samples were stored and measured retrospectively for serum placental growth factor (PlGF), pregnancy-associated plasma protein-A (PAPP-A), free β-human chorionic gonadotropin (β-hCG) and α-fetoprotein (AFP). All had measurement of first-trimester serum markers, and 167 (66%) had second-trimester tests. Values were expressed in multiples of the gestation-specific median (MoMs) in singletons, adjusted for maternal weight, as appropriate.Results
Pre-eclampsia was diagnosed in 12 (9.0%) twin pregnancies of 133 continuing beyond 22 weeks. In unaffected twin pregnancies, all serum markers were statistically significantly increased (P < 0.0001), consistent with a doubling of concentration. Among twin pregnancies, those with pre-eclampsia had a significantly reduced median PlGF compared with surviving unaffected twin pregnancies (0.96 MoM vs 1.46 MoM; P < 0.0002, two-tailed), whilst median PAPP-A, which is known to be reduced in affected singleton pregnancies, was increased (3.91 MoM vs 2.43 MoM; P < 0.0005, two-tailed). The levels of free β-hCG (P < 0.02) and AFP (P < 0.05) were also significantly raised, but to a lesser extent than was the level of PAPP-A. Using a logistic regression algorithm based on first- and second-trimester PlGF and PAPP-A, together with previously published uterine artery Doppler and mean arterial pressure measurements in the same series, the predicted pre-eclampsia detection rate was 65% for a 10% false-positive rate.Conclusions
In twin pregnancy, the predicted detection rate of pre-eclampsia using first- and second-trimester maternal serum and biophysical markers is good. In contrast to singleton pregnancy, PAPP-A levels are raised in the first trimester of twin pregnancies destined to develop pre-eclampsia and therefore a different prediction algorithm is needed.