Background: Thermal effects of radiofrequency electromagnetic fields (REMF) exposure of humans may be assessed by calculations of the parameter recognized as SAR (specific energy absorption rate) in virtual human body models, which actually do not represent anthropometric properties of the entire population. Therefore, it is important to determine the relations between SAR values and anthropometric parameters that enable individualization of SAR estimation independently of body properties of a given person. Material and Methods: The analysis concerned 48 exposure scenarios of 4 virtual body models (male and female) to vertically or horizontally polarized REMF of 27 MHz or 100 MHz frequency of various directions of propagation. Results: In the subgroup of results 100 MHz / vertical polarization statistically significant (strong; p < 0.05) correlations were identified between SAR averaged in the whole body and height, mass, BMI, circumference of chest, waist, neck and frontal cross-section area, and between local SAR in head and neck and the height, mass, circumference of chest waist or neck and frontal cross-section area. Identified relations and SAR in the Gustav model were used to estimate the variety of SAR in Polish population of adults (5-95. percentile of female and male): ±30% for SAR averaged in the whole body, ±50% for localized SAR. Conclusions: It was demonstrated that in the preliminary classified type of assessed REMF exposure (e.g., in terms of field polarization and frequency) it is possible to identify statistical relations between various SAR parameters and anthropometric properties of the exposed body. Related quantities can be used for individualized assessment of worker's electromagnetic hazards. Med Pr 2014;65(3):351–360