Table 2

Linear regression coefficients from a multivariate regression analysis showing the association between weight loss at day 3 as dependent variable, and maternal, obstetric and neonate's characteristics with a stratification by feeding mode

Tests for interactionsModel 1: all neonates, n=1495Model 2: breastfed neonates (exclusively breastfed or mixed- feeding (breast and formula)), n=1084Model 3: formula-fed neonates (exclusively formula), n=411
ParameterInteraction p value*Linear regression coefficientSEp ValueLinear regression coefficientSEp ValueLinear regression coefficientSEp Value
Maternal characteristics
 Recruitment centrePoitiers vs nancy<0.00011.070.17<0.0001−0.430.250.09
 Age at deliveryFor 10 years0.190.460.160.004
 ParityMultipara vs primipara0.0008−0.810.18<0.00010.350.270.20
 Pre-pregnancy BMIFor 10 kg/m20.00020.680.200.0007−0.650.280.02
 Gestational diabetesYes vs no0.830.640.290.03
 Gestational weight gainFor 10 kg0.120.450.180.01−0.050.250.84
Obstetric characteristics
Method of deliveryPlanned caesarean section vs vaginal0.291.530.29<0.0001
 Emergency caesarean section vs vaginal0.720.250.004
Infant characteristics
 SexMale vs female0.13−0.100.170.54−0.570.240.02
 Gestational ageFor 1 week of gestation0.25−0.400.06<0.0001
 Birth weightFor 1 kg0.561.070.18<0.0001
  • * Interaction p value between each explanatory variable and feeding mode, tested in all the neonates.

  • Due to the significant interactions, it is inappropriate to provide these estimates.

  • The interpretation is, for example in model 1: D3WL decreases 0.4 % for each additional week of gestation (β=−0.4): while D3WL increases of 1.07 % for a 1 kg difference (β=1.07).