Constructed wetlands (CWs) polish wastewater prior to discharge into aquatic receptacles. Size variabilities of substrates used in CWS may have effects on the treatment efficiencies though there is scanty information regarding this aspect in VSSF CWs. To address this knowledge gap, a laboratory scale mesocosm experiment was set up to investigate the potential of a VSSF CW to reduce organic matter and FIB using various gravel substrate aggregate sizes. This consisted of three gravel size treatment units; <12.5, 12.5-18 and 18-24 mm in triplicates. 70 liters of pre-treated wastewater from the final wastewater stabilization pond (WSP) of Egerton University’s WSPs system was added to the units, allowed to settle for 6 weeks for development of biofilms, followed by periodic feeding of equal wastewater quantity on weekly basis and influent and effluent samples collected for 8 weeks for analysis. Results disclosed reduction efficiency of 95.2, 94.3 and 88.4 % for E coli in the fine, medium and coarse gravel aggregates respectively. Less than 20 % reduction efficiency was recorded for BOD5 in all gravel aggregate sizes. There was no significant variation on performance of the three gravel aggregate sizes in reduction of both FIB and BOD5 (p˃0.05). Poor performance in BOD5 reduction was related to absence of wetland macrophytes in the study. The relatively high reduction efficiency for FIB was attributed to other factors and processes such as predation, mechanical interactions, starvation, microbial interactions and natural die-offs. The study recommends assessing the combined effect of increasing the retention time, use of wetland macrophytes and incorporating various gravel aggregate sizes in order to increase the efficiency of VSSF CW in reduction of FIB and organic matter.