Exposure metrics for smoke waves from wildland fires and their associations with preterm birth in California

Wildland fire smoke has been linked to preterm birth, but the relative contributions of exposure frequency, duration, and intensity remain unclear. We defined a smoke wave as two or more consecutive days with wildfire-specific particulate matter of aerodynamic diameter ≤ 2.5 microns concentrations exceeding thresholds (5 and 20 μg/m³) and assessed the associations of preterm birth with several different metrics of exposure based on smoke waves experienced during pregnancy. These exposure metrics emphasized different aspects of exposure: intensity was always included via the threshold concentrations but was particularly emphasized in metrics quantifying the average or maximum concentration during smoke waves; frequency was considered in the count of smoke waves experienced; and duration was emphasized in metrics quantifying the total number of smoke wave days or the maximum length of a single smoke wave. Finally, all three aspects contribute to the cumulative exposure during smoke waves. We used Cox proportional hazards models to compare pregnancies at the same gestational age, adjusting for season of conception, health insurance type, and birthing parents' age, education level, race/ethnicity, and smoking status. Increases in each exposure metric were associated with higher hazards of preterm birth. For example, an increase of 3 smoke waves based on a 5 μg/m³ threshold was associated with a hazard ratio of 1.02 (95% confidence interval: 1.01, 1.02), and an increase of 10 smoke wave days based on a 20 μg/m³ threshold was associated with a hazard ratio of 1.03 (1.02, 1.05). Metrics incorporating duration or frequency of smoke waves showed stronger associations with preterm birth than those focusing mainly on intensity.