The aim of the study was to determine the prevalence of orthorexia nervosa among school-age youth from Pomeranian and Warmian-Masurian voivodeships, as well as search for factors which enhance the risk of orthorexia nervosa. An attempt was made to find differences in occurrence of orthorexia nervosa among youth from big cities and small towns. The study group consisted of 864 subjects (599 females and 265 males). The mean age of female participants was 20.21+/-3.27 years, and of male participants - 18.93+/-3.67 years. In the study, we used a proprietary questionnaire to collect patient data, as well as following diagnostic questionnaires: ORTO-15 by Donini et al. (Polish version validated by Stochel, Janas-Kozik et al.), EAT-26 by Garner and Garfinkel, MOCI (Maudsley Obsessive-Compulsive Inventory) by Hodgson and Rachman, and BDI-II (Beck Depression Inventory II) by Beck. The results were statistically analyzed. 27% of subjects were found to be at risk of orthorexia nervosa (score of 35 was consideredas cut-off point). The highest score of risk was observed in the group of students of junior secondary school, the lowest in the group of students of senior secondary school. There were no statistically significant differences in the risk of orthorexia in groups from big city and small town. Studied social factors did not show impact on the risk of orthorexia. Individuals with suspected orthorexia have notably higher BMI. There were no statistically significant differences in occurrence and severity of depression in the study group. Depression occurred in 25% of the subjects;the link between depression and orthorexia was not confirmed. On the basis of the study, it was indicated that the group of the highest risk are students of junior secondary school, probably because of the great interest in physical attractiveness in this period of life, as well as individuals with higher BMI. We believe that for more effective diagnosis it would be advisable to adopt a cut-off point for orthorexia in the ORTO-15 at the level of 35 points, as postulated by Stochel, Janas-Kozik et al. The 40-point threshold is associated with considerable overdiagnosis of the phenomenon.