Importance of including all pregnancy outcomes to reduce bias in epidemiologic studies of neural tube defects--Texas, 1999 to 2005.

BACKGROUND

Neural tube defects (NTDs) often result in fetal death or elective termination; therefore, not all cases are captured in typical hospital-based surveillance. We examined sociodemographic differences among pregnancy outcomes to assess sources of bias in NTD surveillance and research.

METHODS

We used 1999 to 2005 Texas Birth Defects Registry data, a population-based active surveillance system, and calculated crude and adjusted prevalence ratios (aPRs). We then assessed the association of anencephaly and spina bifida with the selected characteristics, stratified by pregnancy outcomes (fetal death, elective termination, or live birth).

RESULTS

Data were available for 1852 NTD cases (anencephaly, 677; spina bifida, 954; and encephalocele, 221), resulting in 1211 live births, 236 fetal deaths, and 405 elective terminations. For both anencephaly and spina bifida, a significant excess of Hispanic mothers was observed among live-birth cases (aPRs=1.2-2.4), but not among mothers experiencing other pregnancy outcomes. Mothers of anencephaly cases resulting in a non-live birth were more likely to be adolescents (aPRs=2.4-2.7 for ages<20 years old vs. ages 25-29 years old), but this pattern was not observed for live-birth cases. A trend of increasing anencephaly risk with increasing parity was demonstrated only among fetal-death cases. For spina bifida, mothers of fetal-death (but not live-birth) cases were less likely to live along the Texas-Mexico border (aPR=0.30).

CONCLUSIONS

Demographic differences across NTD pregnancy outcomes exist and are a potential source of bias. Inclusion of all pregnancy outcomes in NTD surveillance is vital in NTD monitoring and research.