Breast cancer is a heterogeneous disease, previously associated with genomic instability. Our aim was to analyze microsatellite markers in order to determine patterns and levels of instability, as well as possible correlations with histopathological parameters. Polymerase chain reaction was used to characterize microsatellite instability (MSI) and loss of heterozygosity (LOH) in 107 breast carcinomas at twelve microsatellite loci. Some of the markers were selected because of their relation to steroid hormone metabolism, which seems to be related to sporadic breast cancer risk. D5S346 and D17S250 markers showed a statistically significant frequency of MSI. LOH in D3S1611, D17S250, AR and ER-β were associated with some parameters of worse prognosis. Marker group analysis showed that CYP19, AR and ER-β were related to histological grade III, ER-negative and PR-negative cases. Our results suggest that marker group analysis may be preferred to the single marker strategy, being predictive of worst prognosis when single markers are unable to provide such information. A further evaluation of steroid metabolism genes and their association with low penetrance genes in breast cancer may be useful.