Prenatal diagnosis of neural tube defects: predictive value of AF-AFP in a low-risk population.

Abstract

Amniotic fluid alpha-fetoprotein (AF-AFP) determinations were performed on 1,215 women who were at low risk for fetal neural tube defects and who were undergoing mid-trimester amniocentesis for cytogenetic indications, primarily age-related aneuploidy. Maternal sera obtained before amniocentesis and amniotic fluids were assayed in duplicate for alpha-fetoprotein by radioimmunoassay. Of the 1,215 low-risk women, eight (0.7%) had significant elevations of AF-AFP (greater than or equal to +5 SD). In none of the cases was the elevation associated with a fetal neural tube defect. Two cases with elevated AF-AFP were associated with chromosome aberrations; one with impending fetal demise; one with fetal blood contamination; and one case was due to a laboratory error. In one case, no source for the elevated AFP was found, and a normal infant was delivered at term. In the final two cases, the cause of the elevated AF-AFP was a fetal abdominal wall defect (one gastroschisis and one omphalocele). The predictive value of an elevated AFP varies with the population screened, and is reduced by routine ultrasonography before amniocentesis, which at least identifies anencephaly. In a low-risk population, an elevated AF-AFP is most often not associated with a fetal neural tube defect. Because of the low predictive value and the nonspecificity of AF-AFP, genetic counselors should reconsider the recommendation of routine AF-AFP in low-risk maternal populations.