Fine particle sources and adverse events in infants using home cardiorespiratory monitors

Abstract

Background: Recent research has provided a wealth of knowledge of the contributions of air pollution to adverse cardiovascular and respiratory events in sensitive populations. Source apportionment methods can be used to apportion and identify ambient sources of air pollution, which can then be used to estimate health effects of air pollution sources. Infants are thought to be particularly susceptible to air pollutant sources; however, little research has been conducted. The objective of this study was to examine the associations of ambient source apportioned fine particulate matter with bradycardia (low heart rate) and apnea (cessation of respiration) events in a cohort of infants prescribed home cardiorespiratory monitors. Methods: We utilized data from 3,629 infants within the Atlanta metropolitan statistical area who used home cardiorespiratory monitors between November 19, 1998 and December 31, 2002. Home monitors were used to record respiratory effort and heart rate to detect bradycardia and apnea events. Chemical mass balance (CMB) and positive matrix factorization (PMF) source apportionment methods for fine particulate matter (PM2.5) were used to produce 14 source profiles. Repeated-measures unconditional logistic regression using generalized estimating equations (GEEs) was used to associate cardiorespiratory events with air pollution sources. A stationary 45-dependent correlation structure was used to account for the correlation of multiple event-days for a patient. The model included age of the infant, the squared age of the infant, average daily temperature, the square of average daily temperature and indicator variables for weekend and federal holiday. Our analysis used a day, day-squared and day-cubed set of variables for the full term/normal birth weight apnea analysis to adjust for time. Cubic splines with seasonal knots for time were used to adjust for long term temporal trends in the remainder of the presented final results for apnea analyses and all of the bradycardia analyses. We performed separate analyses for zero and one-day lags of pollution. We used odds ratios (ORs) and 95% confidence intervals (CI's) as our measure of effect size to describe the odds of an event occurring. ORs from the analysis were calculated for an inter-quartile range (IQR) increase in each of the single pollutant source models. Apnea and bradycardia were evaluated separately. Results: We observed a pattern of suggestive positive odds ratios, especially in the primary analysis, such as in the woodsmoke source, which were consistent across source apportionment method and lag structure. We observed positive associations in the positive matrix factorization models for the woodsmoke source in the apnea zero-day lag for the primary analysis with an odds ratio of 1.031 (95% CI: 1.001-1.061; IQR: 0.93 µg/m3 increase). We also observed positive associations in the positive matrix factorization models for the woodsmoke source in the apnea one-day lag analysis for primary and premature/normal birth weight with an odds ratio of 1.048 (95% CI: 1.017-1.080; IQR: 0.93 µg/m3 increase) and 1.041 (95% CI: 1.006-1.077; IQR: 0.93 µg/m3 increase), respectively. The results for the full term/normal birth weight strata had stronger odds ratios than for both the primary and premature/low birth weight strata. Conclusions: Although we did observe wide confidence intervals and some protective odds ratios, we also observed stronger odds ratios in the one-day lag models compared to the zero-day lag models for the apnea events across both source apportionment methods. We observed some suggestive associations between cardiorespiratory events and source apportioned fine particulate matter that contributes to the body of air pollution literature. The access to such a large cohort of infants with the apnea and bradycardia data made this study a contribution to the understanding of the associations between cardiorespiratory events and source apportioned fine particulate matter in infants at high risk for cardiorespiratory events.