RESUMO
OBJECTIVE: To find the preoperative predictors of obesity recurrence. MATERIAL AND METHODS: We analyzed 81 patients with obesity who underwent laparoscopic sleeve gastrectomy. Patients were followed-up for 5 years and divided into 2 groups (64 (79.0%) patients were in the persistent weight loss group and 17 (21.0%) ones were in the obesity recurrence group). RESULTS: We tried to make a mathematical model for predicting obesity recurrence after bariatric surgery based on various parameters (age, height, weight, hemoglobin A1C, lipoprotein A, T3, T4, triglycerides, TSH, cholesterol, arterial hypertension, diabetes, metabolic syndrome, percentage of fat mass). We trained the following classification models: logistic regression, naive Bayes classifier, decision tree. None classifier provided sufficient prediction quality. This may indicate that these data do not allow predicting of obesity recurrence after sleeve gastrectomy. CONCLUSION: We failed to construct a mathematical model based on preoperative data of patients with obesity predicting recurrence in long-term period after sleeve gastrectomy.
