Table 4.

Mixture effects on spontaneous premature rupture of membranes for one quartile increase in the air pollution mixture and PM_2.5 mixture throughout the pregnancy based on quantile-based g computation.

Mode	Contribution to positive/negative effect (%)	Positive/Negative coefficient (β)a	Overall mixture coefficient (β [95% CI])b	Overall mixture effect (OR [95% CI])
Model 1: Krigged air pollutants
Positive mixturec			0.14 (0.12, 0.17)	1.15 (1.12, 1.18)
NO2	25.0	0.16	(p <.001)	(p <.001)
O3	49.6
PM2.5	25.4
Negative mixtured
PM10	100.0e	−0.02
Model 2: PM2.5 components
Positive mixturec			0.07 (0.05, 0.09)	1.07 (1.05, 1.09)
PM2.5 sulfate	34.1	0.10	(p <.001)	(p <.001)
PM2.5 nitrate	43.3
PM2.5 organic matter	22.6
Negative mixtured
PM2.5 ammonium	100.0e	−0.03

Note: The exposure to four krigged air pollutants was measured based on empirical Bayesian kriging. The exposure to PM_2.5 components was measured based on a fine-resolution geoscience-derived model. Models were adjusted for maternal age, race/ethnicity, education level, income level, pre-pregnancy body mass index, smoking status, parity, year of infant birth, and season of conception. Abbreviation: CI, confidence interval; OR, odds ratio.

^aThe coefficient for the positive or negative mixture.

^bThe overall mixture coefficient is the sum of coefficients of the positive mixture and negative mixture.

^cThe positive mixture includes pollutants positively associated with the outcome in the model.

^dThe negative mixture includes pollutants negatively associated with the outcome in the model.

^eThe contribution of the pollutant is 100.0% as it is the only pollutant negatively associated with the outcome in the model.