You are here:
A TIME SERIES ANALYSIS OF AIR POLLUTION AND PRETERM BIRTH IN PENNSYLVANIA, 1997-2001
Citation:
Sagiv, S., P Mendola, D. Loomis, A. H. Herring, L Neas, D. A. Savitz, AND C. Poole. A TIME SERIES ANALYSIS OF AIR POLLUTION AND PRETERM BIRTH IN PENNSYLVANIA, 1997-2001. Presented at International Society for Environmental Epidemiology, New York, NY, August 1-4, 2004.
Description:
Introduction: Small increases in risk for preterm delivery in relation to air pollution have been reported, but prior investigations may have inadequately controlled for individual factors, such as maternal smoking. To eliminate confounding by known and unknown individual risk factors that do not vary over short periods of time, we conducted a times-series analysis to investigate the effect of ambient outdoor particulate matter with diameter of 10 micrometers or less (PM10) and sulfur dioxide (SO2) on risk for preterm delivery. Methods: Daily counts of preterm births were obtained from birth records in four Pennsylvania counties (Allegheny, Beaver, Lackawanna and Philadelphia) between January 1, 1997 and December 31, 2001. PM10 and SO2 levels were averaged over a six-week window directly preceding birth. Full, adjusted models controlled for long-term seasonal trends, co-pollutants (nitrogen dioxide, carbon monoxide and SO2 in the PM10 analysis/PM10 in the SO2 analysis), and were offset by the number of live births in the population. Long-term seasonal trends were adjusted using county-specific parametric splines and county-level information was incorporated using a mixed model with a random intercept. Results: Of 187,997 singleton births, 21,450 (11.4%) were born preterm. Mean six-week air pollution for the four counties ranged from 8.67 to 68.85 g/m3 for PM10 and from 0.79 to 17.02 parts per billion (ppb) for SO2. In full, adjusted models, we observed approximately 1 excess preterm birth for every 100 births exposed to a 50 ug/m3 increase in PM10 (Risk Ratio=1.09; 95% CI: 1.01, 1.19) and 2 excess preterm birth for every 100 births exposed to a 15 ppb SO2 increase (RR =1.20%, 95% CI: 1.04, 1.39). Discussion: These findings provide evidence of small absolute increases in preterm birth risk with exposure to both PM10 and SO2 in the six weeks preceding birth. Results from this time-series analysis of air pollution and preterm delivery are consistent with prior investigations using conventional analyses. This suggests that confounding at the individual level is probably not explaining the observed association between air pollution and preterm delivery. Further analyses are planned to disentangle the effects of PM10 and SO2 and to investigate a possible acute effect of these pollutants in the seven days preceding birth.
This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.