Objective To investigate the effect of different premolar extraction models on postextraction Bolton overall ratio among different classes of malocclusion. Methods 180 patients with equivalent skeletal and dental classifications were chosen（Group Ⅰ, Ⅱ and Ⅲ, with 60 per group）. According to dental casts, the Bolton overall ratios of each group were measured and compared before and after hypothetical premolar extraction（the four premolar extraction models were all first premolars, all second premolars, upper first and lower second premolars, and upper second and lower first premolars）. Chi-square Test analysis was taken within each group in order to compare the abnormal rates of postextraction Bolton overall ratios. Results The differences of pretreatment Bolton overall ratios among three types of malocclusions were found statistically significant（P=0.000）. For Class Ⅰ and Class Ⅱ group, the four extraction models made no statistically significant differences in the abnormal rates of postextraction Bolton overall ratio（P>0.05）. For Class Ⅲ group, the four extraction models made statistically different results that was the upper second and lower first premolar extraction model creating lower abnormal rates than the upper first and lower second premolar extraction model（P=0.002）. Conclusion As far as the postextraction Bolton overall ratios are concerned, Class Ⅰ and Ⅱ malocclusions can choose any of four premolar extraction models, while to Class Ⅲ malocclusions, the upper second and lower first premolar extraction model fits better than the upper first and lower second premolar extraction model.
