Cardiopulmonary exercise testing (CPET) is used for the diagnosis and prognosis of cardiovascular and pulmonary conditions in children and adolescents. Several authors have published reference values for pediatric CPET, but evaluation of their validity is lacking. The aim of this study was to review pediatric CPET references values published between 1980 and 2014. We specifically assessed the adequacy of the normalization methods used to adjust for body size. Articles that proposed references values were reviewed. We abstracted information on exercise protocols, CPET measurements and normalization methods. We then evaluated the studies’ methodological quality and assessed them for potential biases. Thirty-four studies were included. We found important heterogeneity in the choice of exercise protocols and in the approach to adjustment for body size or other relevant confounding factors. Adjustment for body size was principally done using linear regression for age or weight. Assessment of potential biases (residual association, heteroscedasticity and departure from the normal distribution) was mentioned in only a minority of studies. Our study shows that contemporary pediatric reference values for CPET have been developed based on heterogeneous exercise protocols and variable normalization strategies. Furthermore, assessment of potential bias has been inconsistent and insufficiently described. High-quality reference values with adequate adjustment for confounding variables are needed in order to optimize CPET’s specificity and sensitivity to detect abnormal cardiopulmonary response to exercise.