Table 3:

Exploratory multiple linear regression models examining birth weight and length z score and percentile in relation to the percent of underhydrated urine samples in the second and third trimesters among the full sample

Independent variables	Model 2.1a Birth weight z score (SE)	Model 2.2a Birth weight percentile (SE)	Model 2.3a,b Birth length z score (SE)	Model 2.4a,b Birth length percentile (SE)
Percent of underhydrated urine samples in the 2nd trimester (per 10% point difference) c	−0.32** (0.13)	−9.31** (3.33)	−0.21 (0.16)	−3.33 (3.39)
Percent of underhydrated urine samples in the 3rd trimester (per 10% point difference) c	0.27* (0.14)	7.45** (3.33)	0.27* (0.14)	2.71 (2.69)
Intervention (control = reference)	−0.41 (0.52)	−10.26 (12.61)	0.04 (0.60)	4.57 (12.12)
Age (years)	0.03 (0.08)	0.03 (2.20)	0.02 (0.14)	−1.29 (3.03)
GWG (lbs)	0.07** (0.03)	1.65** (0.62)	0.03 (0.03)	1.38* (0.75)
Pre-pregnancy BMI	0.001 (0.05)	0.05 (104)	−0.04 (0.05)	0.01 (0.85)
Percent active time (%weeks with >7500 steps)	0.01 (0.01)	0.18 (0.35)	0.00 (0.02)	0.20 (0.30)
Observations	26	26	25	25
R-squared	0.34	0.37	0.22	0.28

^***p<0.01

^**p<0.05

^*p<0.1

Note: Results from multiple linear regression analyses with robust standard errors that included all listed predictors and covariates simultaneously. Coefficients for the percent of underhydrated urine samples represent the estimated difference in birth outcomes in relation to a 10-percentage-point difference in the percent of urine samples that had a Uosm≥500 mOsm/kg while controlling for other covariates.

^a1 woman missing gestational weight gain (GWG);

^b1 woman missing birth length record

^cUnderhydrated defined as urine osmolality≥500 mOsm/kg