Table 5.

Change in metabolite intensity associated with urinary paraben concentration (ng/mL, specific gravity–adjusted) after adjusting for confounding using multivariable linear regression in 42 participants in the Early Pregnancy Study.

Metabolite number (model)a	Sample time point	Paraben	Intensity change when paraben concentration doubles (95% CI)	Adjusted $R^2$
1 (a)	Conception cycle	Propylparaben	28% ($-4\%$, 71%)	$-0.02$
2 (a)	Conception cycle	Methylparaben	75% (45%, 111%)	0.69
2 (b)	Conception cycle	Propylparaben	32% (0%, 73%)	0.19
2 (c)b	Conception cycle	Methylparaben	103% (49%, 175%)	0.70
		Propylparaben	$-15\%$ ($-35\%$, 12%)
2 (d)	Early pregnancy	Methylparaben	62% (38%, 90%)	0.60
3 (a)	Conception cycle	Methylparaben	$-14\%$ ($-26\%$, $-1\%$)	0.27
4 (a)	Conception cycle	Propylparaben	15% ($-7\%$, 41%)	0.16
5 (a)	Conception cycle	Methylparaben	59% (33%, 89%)	0.59
5 (b)	Conception cycle	Propylparaben	32% (5%, 66%)	0.21
5 (c)b	Conception cycle	Methylparaben	72% (27%, 133%)	0.58
		Propylparaben	$-8\%$ ($-30\%$, 20%)
5 (d)	Early pregnancy	Methylparaben	69% (49%, 93%)	0.70
6 (a)	Conception cycle	Propylparaben	15% ($-5\%$, 39%)	$-0.09$
7 (a)	Early pregnancy	Propylparaben	2% ($-9\%$, 13%)	0.01

Note: The adjusted $R^2$ is for the full model, and full regression results are included in supplementary materials (Table S4). The full model includes paraben concentration; acetaminophen, aspirin, antibiotic, and vitamin use; age; BMI; marijuana use; alcohol consumption; caffeine intake; smoking status; recent oral contraceptive use; education; income; health care occupation; year of specimen collection; season of specimen collection; and whether a pooled sample includes weekend days. BMI, body mass index; CI, confidence interval.

^a Metabolite numbers (1–7) correspond to the metabolites described in Table 4, and model letters (a–d) differentiate models run for the same metabolite.

^b Bipollutant model, including both methylparaben and propylparaben.