RESUMO
AIM: to evaluate differences in craniofacial morphologic features and several anthropometric measures between schizophrenia patients and non-psychiatric controls, and to find the best-fit model to differentiate between two groups. METHODS: 40 morphologic features of the head and face, and 5 craniofacial anthropometric measures were evaluated using the Lane Dysmorphology Scale in 58 patients and 46 controls. Total MPA score and subscores for different craniofacial regions were calculated. Individual items were examined using logistic regression analyses to define a model that can discriminate between patient vs. control status. RESULTS: total MPA score, and several subscores (general asymmetry, nasal, lip, ear and tongue) were significantly higher in the patient group. Patients were distinguished by significantly higher measures of maxillary and mandibular facial arcs, general and subtle facial asymmetries presented as deviation of facial landmarks from the vertical facial midline and horizontals, more variable vermilion of the upper lip, tongue surface, frenulum and anterior hair margin, and more adherent and underdeveloped earlobes. A final regression model including maxillary are, facial asymmetry, and adherent earlobes as independent predictors proved useful to efficiently recognize schizophrenia patients (specificity and positive prediction value of 100% when all the three items were present in an individual) or to exclude risk for schizophrenia (sensitivity and negative prediction value of 96.6% and 84.6%, respectively, in cases no one of the three items was present). CONCLUSIONS: schizophrenia patients evidenced significantly more craniofacial dysmorphology than controls. The model revealed in the study needs to be verified in larger samples and other populations.