Bioelectrochemical systems (BES) application was proposed for a variety of specific uses, due to these systems’ characteristics: electrodes can act as virtually inexhaustible electron acceptors/donors, offering a growth-support surface for microorganisms, and stimulating naturally-occurring microbial degradation activities. In situ, groundwater denitrification therefore seems to be a potential candidate for their use. In this study, buried biocathodes were operated in laboratory settings for the simulation of in situ groundwater denitrification. Two alternative configurations were tested: biocathode buried in sand, and biocathode buried in gravel. A control test with a biocathode in absence of sand/gravel was also performed. In all the cases, biocathodes were driven by power supply or potentiostat to guarantee a steady electron flux to the cathode. The presence of sand and gravel strongly influenced the denitrification process: in both configurations, accumulation of intermediate N-forms was detected, suggesting that the denitrification process was only partially achieved. In addition, a significant decrease (in the 20–36% range) in nitrate removal rates was measured in sand and gravel setups compared to the control reactor; this issue could be attributed to lack of recirculation that limited contact between substrate and electrode-adherent biofilm. Biocathodes buried in gravel obtained better results than those buried in sand due to the lower packing of the medium. The results of this study suggest that, in order to achieve successful in situ treatment, special design of submerged-biocathodic BESs is necessary.