Purpose: The aim of the present study was to adjust a multivariate model to explain each of the response variables for the occurrence of surgical difficulty during the removal of impacted lower third molars.

Patients and methods: A prospective cohort study was carried out involving patients submitted to at least one surgical removal of an impacted lower third molar. A total of 285 patients fulfilled the eligibility criteria and 473 surgeries were performed. Preoperative variables indicative of surgical difficulty were recorded. All surgical procedures were performed under the same conditions by two surgeons who were unaware of the data collected in the pre-selection phase. Either Pearson’s chi-square test or Fisher’s exact test was used for the data analysis ( P < 5.0%).

Results: Root number ( P ⁽¹⁾ < 0.004 ^* ) and morphology ( P ⁽¹⁾ < 0.031 ^* ), tooth position ( P ⁽¹⁾ = 0.001 ^* ), periodontal space ( P ⁽²⁾ < 0.004 ^* ) and second molar relation ( P ⁽¹⁾ = 0.001 ^* ) were significant predictors of surgical difficulty, whereas patient age ( P ⁽¹⁾ = 0.097), gender ( P ⁽¹⁾ = 0.470), body mass index ( P ⁽¹⁾ = 0.719), associated pathologies ( P ⁽¹⁾ = 0.237), relation with mandibular canal ( P ⁽¹⁾ = 0.384) and width of 3rd molar crown ( P ⁽¹⁾ = 0.154) were not significant predictors.

Conclusion: Many factors contribute to surgical difficulty, but considering these factors individually, some are only determinants of either difficulty or complications. Thus, not all significant predictors of surgical difficulty should be considered indicators of complications.

Conflict of interest: None declared.