OBJECTIVE To evaluate if clinical and laboratory parameters could assist in the differential diagnosis of intra and extra-hepatic neonatal cholestasis (NC). METHODS Retrospective study of NC patients admitted at the Pediatric Hepatology Outpatient Clinic of the teaching hospital of Universidade Estadual de Campinas (UNICAMP), Campinas, Brazil, between December 1980 and March 2005. The approach to the diagnosis of NC was standardized. According to diagnosis, patients were classified into two groups: I (intra-hepatic neonatal cholestasis) and II (extra-hepatic neonatal cholestasis). In order to verify if there was association with the categorical variable, the chi-square and Mann-Whitney tests were used, with corrections for age for the covariance analysis (ANCOVA). The determination of accuracy of the clinical and laboratory variables for differentiation of the groups was made using the analysis of the ROC curve. RESULTS One hundred and sixty-eight patients were evaluated (group I = 54.8% and group II = 45.2%). In the patients with less than 60 days of life there was predominance of intra-hepatic causes, whereas, in those older than 60 days, there was predominance of extra-hepatic etiology (p < 0.001). Median birth weight was lower in group I (p = 0.003), as well as length at birth (p = 0.007). Median values of direct bilirubin were higher in group II (p = 0.006). Values of gamma-glutamyltransferase (GGT) (10 times higher than the limit of normality) presented sensitivity of 56.3%, specificity of 91.5%, and accuracy of 75.7% for the diagnosis of extra-hepatic cholestasis. CONCLUSIONS In the present study, extra-hepatic NC presented greater weight and length at birth, fecal hypocholia/acholia, choluria, hepatomegaly, increase in GGT (10.8 times higher than the limit of normality), and a delay for investigation in the tertiary center.