The incidence and some determinants of LBW were studied in two population-based cohorts of singleton live birth from families residing in Ribeirão Preto, state of São Paulo, southeast Brazil, the first including infants born from June 1978 to May 1979 (6750 birth-population survey) and the second including infants born from May to August 1994 (2990 birth-sample survey). The incidence of LBW was 7.2% in 1978/79 and 10.6% in 1994. After adjustment for confounding factors by multiple logistic regression, the following determinants remained significant: female gender, maternal age of 35 years and more, pre-term delivery, less than 4 prenatal visits, maternal smoking, lower maternal schooling and manual jobs/unemployment in 1978/79 and pre-term delivery, maternal smoking and cesarean section (CS) in 1994. The adjusted odds ratios comparing these two series decreased for female gender, maternal age less than 20 or 35 years and more, pre-term birth, medicare insurance, lower maternal schooling and manual workers and was increased for non-cohabiting women, primiparity, multiparity, less than four prenatal visits, CS and maternal smoking. The population attributable risk percent fell for the majority of risk factors and increased for CS delivery, pre-term birth, multiparity (≥5), primiparity and non-cohabiting women. The increase in LBW rate, comparing 1978/79 with 1994, was higher for the families with more qualified occupations, occurred only for infants delivered from 36 to 40 weeks of gestational age and weighing 1500-2499 g., who are those at risk for elective CS. The CS rate rose from 30.3% in 1978/79 to 51.1% in 1994. CS delivery had the highest increment in population attributable risk percent between 1978/79 and 1994. In 1978/79, CS increased with birth weight and gestational age (33,2% for pre-term, 32,9% for term and 33,9% for post-term birth); in 1994, the highest CS rate occurred among term birth (54,6%). The mean and median birth weight decreased for the entire newborn population and this phenomena cannot be attributable to registration artifacts. Therefore, the increase in LBW rate was probably due to iatrogenic practices associated with elective CS.