• Kaplan-Meier algorithm;
  • preterm delivery;
  • total activin A



To retrospectively evaluate whether increased serum levels of total activin A (t-activin A) are found in women who subsequently experience preterm delivery (PTD).


Data on maternal serum t-activin A concentrations were available from a total of 84 singleton pregnant women and included 14 PTD pregnancies, each matched for gestational age and length of freezer storage, with 5 control pregnancies having term delivery (TD). Analyte values were expressed as multiple(s) of the control median.


The median t-activin A for controls and cases was 1.00 ± 0.45 and 1.27 ± 0.53 MoM, respectively. Univariate analysis of the MoM values was performed using the Kaplan-Meier algorithm. Differences in the rate of delivery using a t-activin A MoM cut-off of ≥1 SD (equivalent to 1.26 MoM) were analysed using the log rank test. The cumulative rate of PTD (<37 weeks) was significantly higher for women with t-activin A concentrations ≥1.26 MoM than those with t-activin A concentrations below this cut-off (40% vs.. 10%, p-value = 0.0218 log rank test).


T-activin A concentration is higher in women who will develop PTD in a low-risk population. T-activin A values are inversely proportional to the time elapsed from blood test to delivery. Prospective studies would determine the precise discriminability of this marker for PTD and the best week for performing the blood test, allowing for a proper calculation of the detection rate and a positive predictive value. Copyright © 2006 John Wiley & Sons, Ltd.