You are here:
The sensitivity of length of pregnancy using clinical estimate of gestation versus last menstrual period: an example with particulate matter and preterm birth
Citation:
Rappazzo, K., D. Lobdell, L. Messer, C. Poole, AND J. Daniels. The sensitivity of length of pregnancy using clinical estimate of gestation versus last menstrual period: an example with particulate matter and preterm birth. Presented at Society for Pediatric and Perinatal Epidemiologic Society, Boston, MA, June 17 - 18, 2013.
Impact/Purpose:
Estimating gestational age is usually based on last menstrual period date (LMP) or clinical estimation (CGA); both approaches introduce error and potential bias. Differences in the two methods of gestational age assignment may lead to misclassification and differences in risk estimates, particularly if exposure assignment is also dependant on gestation. This methods paper assesses the differences between the two methods to obtain gestational age and the effects it may have on results when looking at an environmental exposure - air pollution.
Description:
Estimating gestational age is usually based on last menstrual period date (LMP) or clinical estimation (CGA); both approaches introduce error and potential bias. Differences in the two methods of gestational age assignment may lead to misclassification and differences in risk estimates, particularly if exposure assignment is also dependant on gestation. To examine how these gestational age estimation approaches produced different cohort compositions, we constructed two 20-week gestational age cohorts of pregnancies between 2000 and 2005 from live birth certificates: one defined preterm birth (PTB) status using CGA and the other using LMP. Within these two populations, we estimated risk of birth in 4 preterm categories(PTBs per 106 pregnancies) and risk differences(RD (95% confidence intervals)) for exposure to particulate matter under 2.5 micrometers in aerodynamic diameter (PM2.5). Though the overall cohort was 10% smaller (1,730,969 versus 1,940,212), substantially more births were classified as preterm using LMP compared with CGA estimation (278,031(16%) versus 156,325(8%)). Divergences in RDs between the two cohorts increased as exposure weeks approached delivery. For example, among births between 28-31 weeks, RDs for PM2.5 exposure at week 7 were 44(13, 76) and 46(23, 69) for clinical and LMP populations, respectively; while RDs with exposure at week 26 were 11(-10, 33) and -19(-49, 11). Results are sensitive to the choice of gestational age estimation, though the degree of sensitivity varies across exposure weeks. Estimation choice is critical, particularly when both outcome and exposure are dependent on the result. This abstract does not reflect EPA policy.