PREDICTION OF FETAL GROWTH BASED ON MATERNAL SERUM CONCENTRATIONS OF HUMAN CHORIONIC-GONADOTROPIN, HUMAN PLACENTAL-LACTOGEN AND ESTRIOL

Citation
T. Markestad et al., PREDICTION OF FETAL GROWTH BASED ON MATERNAL SERUM CONCENTRATIONS OF HUMAN CHORIONIC-GONADOTROPIN, HUMAN PLACENTAL-LACTOGEN AND ESTRIOL, Acta obstetricia et gynecologica Scandinavica, 76, 1997, pp. 50-55
Citations number
24
Categorie Soggetti
Obsetric & Gynecology
ISSN journal
00016349
Volume
76
Year of publication
1997
Supplement
165
Pages
50 - 55
Database
ISI
SICI code
0001-6349(1997)76:<50:POFGBO>2.0.ZU;2-8
Abstract
Background. The purpose was to determine the usefulness of maternal se rum concentrations of human chorionic gonadotropin (hCG), human placen tal lactogen (hPL) and estriol as predictors of fetal growth. Method. From a large cohort serum obtained serially at 17, 25, 33 and 37 weeks of gestation were analyzed for randomly selected pregnancies resultin g in small for gestational age (SGA, n = 102) and non-SGA (n = 112) in fants. Results. There were no significant correlations between birthwe ight ratio (ratio of birthweight to mean weight for gestational age) a nd hCG, but between birthweight ratio on one hand and estriol for all stages of pregnancy (r = 0.19-0.38, p < 0.01 - p < 0.001) and hCL exce pt at 33 weeks (r = 0.11 - 0.40, p ns - p < 0.001) on the other. There were statistically significant, but small median differences and subs tantial overlaps between the SGA and non-SGA infants for hCG at 17 and 37 weeks, for hPL at 17, 33 and 37 weeks, and for estriol at all the stages of pregnancy. The sensitivity and positive predictive value of low hormone concentrations (below the 10th percentile) in predicting t he birth of an SGA infant were in the range of 6 - 26% and 17 - 39%, r espectively. The corresponding specificity and prediction of a non-SGA infant from normal levels were 91 - 93% and 85 - 88%. Conclusions. HP L and estriol, but not hCG concentrations, are positively related to t he size of the fetus, but the relationships are too weak to be of pred ictive value in an unselected population.