Purpose:Body composition of adolescent athletes is often evaluated scientifically and in sports by using reference equations developed from nonathlete adolescent populations. The aim of this study was to analyze the validity of predictive equations based on skinfold measurements, as compared with dual-energy X-ray absorptiometry (DXA), to estimate body fat in adolescent modern pentathlon athletes.Methods:51 athletes, 27 male (mean age = 15.1 years; standard deviation, SD = 1.5 years) and 24 female (mean age = 14.2 years; SD = 2.5 years), were assessed using DXA, anthropometric parameters, sports practice anamnesis, and pubertal stages. Agreement between methods was tested with boxplots of mean comparisons using Student’s t test (p < .05), and Bland-Altman plots.Results:The body density equations of Durnin & Rahaman (1967) and Durnin & Womersley (1974) showed better agreement with DXA than the other predictive equations, for both females (difference between means=-2.03; 2SD = 8.44) and males (difference between means = 0.98; 2SD = 7.30). There were no mean differences between these equations and the reference method (DXA; p > .05), but they did display high variability (2SD).Conclusion:The high variability among results indicated imprecision. Predictive skinfold equations developed for nonathlete adolescents do not offer good validity for modern adolescent pentathlon athletes, and should be avoided.