The purpose of this work was to correlate the maternal feed intake, ponderal and metabolic parameters with the neonatal birth weight and the offspring number to obtain linear models that could predict dependent variables in pregnant Wistar rats and their neonates fed either 4 or 20 % dietary protein concentration. The independent variables considered were (a) maternal feed, energy and protein intakes; (b) maternal body weight before and after Caesarean section (rats were humanely killed), and maternal body weight gain before and after Caesarean section; and (c) maternal metabolic factors including plasma insulin and glucose, insulin resistance (insulin/glucose ratio), growth hormone (GH) and insulin/GH ratio.

Best-fit stepwise multiple regression analysis was used to select factors (independent variables) that could be entered in the models. Independent variables were grouped in two blocks: (a) feed and ponderal variables, and (b) metabolic variables. The statistical criterion for a variable to enter in a given model was a probability of F < 0.05. The variance found in the birth weight was explained in a 32 % by the maternal protein intake (R ² = 0.32) and in a 47 % by the maternal insulin resistance (R ² = 0.47). Maternal body weight gain after a Caesarean section was the independent variable, among the feed and ponderal parameters, that had the strongest correlation (R ² = 0.26) with the number of offspring. Also, the linear model obtained for the offspring number with metabolic variables included two factors: glucose and insulin/GH. In the first step, glucose entered with F < 0.001 and R ² = 0.33. In the second step, insulin/GH increased R ² to 0.44 (F = 0.027).

In conclusion, these results support that various maternal ponderal and metabolic factors may modify neonatal growth and offspring number in pregnant rats fed 4 and 20 % dietary protein concentration. The higher predictive factors were insulin resistance for the birthweight, and both maternal plasma glucose and insulin/GH ratio for the offspring number.