Objective: To estimate the prevalence of preterm birth by categories of birth weight, and to obtain an equation to correct the estimates.
Methods: Systematic review of the Brazilian literature published from 1990 to 2012, to identify studies with primary collection of data on birth weight and gestational age. Twelve studies were selected and contributed for tabulations of preterm prevalence according to 100 g birth weight categories. These results were combined using sex-specific fractional polynomial equations and the resulting curves were compared with results from the Live Birth Information System for the years 2000, 2005, 2010 and 2011.
Results: For all birth weight categories, preterm prevalence estimates based on primary studies had a higher prevalence than those of the the Live Birth Information System. The prevalence reported by the Live Birth Information System was of 7.2% in 2010, about 38.0% lower than the estimated prevalence of 11.7% obtained with the correctional equation.
Conclusions: Information reported by the Live Birth Information System on preterm prevalence does not reflect the true magnitude of the problem in Brazil, and should not be used without the correction factors proposed in the present analyses.