Whole-body counters (WBC) are used to detect potential intake of radioisotopes in the human body. To achieve a precise measurement of an incorporated activity, a WBC needs to be properly characterized and calibrated. However, due to technical limitations, not every geometry can be calibrated using adequate physical phantoms. To extend the use of a standard calibration to other geometries, we performed GEANT4 Monte Carlo (MC) simulations characterizing the WBC at the Spiez Laboratory, Switzerland. The aim of our project was threefold: (1) to validate our MC model against an experimental model, (2) to study the impact of a person's size on the detection efficiency of the WBC and (3) to study the impact of intake scenarios on the estimation of whole-body activity. First, we simulated the calibration of the WBC with the IGOR phantoms. Then, we computed the detection efficiency of the WBC for homogenous whole-body contamination using ICRP adult voxel phantoms scaled to sizes. We computed the bias of the activity that would be measured by the WBC relative to the activity contained in the body in case of inhalation, modeled by the ICRP biokinetic models. Our computed detection efficiency obtained with the IGOR phantoms agreed within 5% with the calibration measurements. We found that WBC efficiency depends on the weight over height ratio (w/h). We proposed a criterion for the calibration curve selection adequate for any person, and we validated this criterion using ICRP pediatric phantoms. We estimated that the precision of a whole-body activity measurement in the case of a homogeneous activity distribution is at the level of 20%. Finally, for a non-homogenous contamination of the body by inhalation, we showed that the WBC measurements can overestimate intake up to 40-80% depending on the absorption type of the substance into the blood, and the size of the scanned person. In conclusion, our study confirms the need of using phantoms with different sizes to calibrate a WBC and shows that in case of a highly non-homogenous distribution of the radioisotope in the body, the WBC may significantly overestimate the whole-body activity. To limit this bias, we recommend scanning the entire body with the detectors instead of setting them at fixed positions or increasing the distance between the detectors and the phantom.