Skip to main content

Table 5 Linear regression analyses: Relationship between maternal exposure to smoking, preeclampsia and fetal growth

From: Smoking in preeclamptic women is associated with higher birthweight for gestational age and lower soluble fms-like tyrosine kinase-1 levels: a nested case control study

Model Unadjusted β
(95% CI)
p-value Adjusted β#
(95% CI)
p-value
Model 1a
z-score BWGA = smoker + preeclampsia + interaction term
    
Smoker -0.29
(-0.51,-0.08)
0.008 -0.38
(-0.61, -0.15)
0.0013
Preeclampsia -0.67
(-0.91,-0.43)
< 0.0001 -0.67
(-0.93, -0.41)
< 0.0001
Smoker* preeclampsia 0.86
(0.36, 1.36)
0.0008 0.86
(0.33, 1.39)
0.0015
Model 1b
z-score BWGA = smoker + preeclampsia + interaction term + z-score PWGA
    
Smoker -0.30
(-0.50, -0.09)
0.004 -0.32
(-0.54, -0.09)
0.0067
Preeclampsia -0.61
(-0.86, -0.35)
< 0.0001 -0.60
(-0.87, -0.33)
< 0.0001
Smoker* preeclampsia 0.74
(0.25, 1.23)
0.003 0.69
(0.16, 1.22)
0.0112
z-score PWGA 0.56
(0.48, 0.65)
< 0.0001 0.52
(0.43, 0.61)
< 0.0001
Model 2a
z-score BWGA = log hair nicotine + preeclampsia + interaction term
    
Log nicotine -0.12
(-0.16, -0.05)
0.0004 -0.086
(-0.16, -0.01)
0.0234
Preeclampsia -0.31
(-0.55, -0.07)
0.011 -0.36
(-0.63, -0.09)
0.0087
Log hair nicotine * preeclampsia 0.28
(0.11, 0.46)
0.002 0.26
(0.07, 0.46)
0.0091
Model 2b
z-score BWGA = log hair nicotine + preeclampsia + interaction term + z-score PWGA
    
Log nicotine -0.09
(-0.16, -0.03)
0.005 -0.06
(-0.13, 0.01)
0.0912
Preeclampsia -0.31
(-0.54, -0.08)
0.009 -0.33
(-0.60, -0.07)
0.0131
Log hair nicotine * preeclampsia 0.25
(0.07, 0.42)
0.007 0.25
(0.06, 0.44)
0.0095
z-score PWGA 0.56
(0.48, 0.64)
< 0.0001 0.52
(0.42, 0.62)
< 0.0001
  1. Dependent variable for all models: z-score birthweight for gestational age (BWGA), as defined in Kramer et al [14]
  2. PWGA = placental weight for gestational age; z-score PWGA calculated as defined by McNamara [15, 16]
  3. # analyses adjusted for age, pre-pregnancy body mass index, maternal language other than French or English, diabetes, chronic hypertension and parity (nullipara vs. other). For adjusted models 1a and 1b, diabetes was statistically significant (p = 0.03 and p = 0.03, respectively). No other variables were statistically significant in any of the models.
  4. R-square for models: Model 1a unadjusted, 0.057, Model 1a adjusted, 0.096; Model 1b unadjusted, 0.388, Model 1b adjusted, 0.391; Model 2a unadjusted, 0.062, Model 2a adjusted, 0.100; Model 2b unadjusted, 0.395, Model 2b adjusted, 0.384.