RESUMO
PURPOSE: A Hartmann's procedure is performed in perforated diverticulitis, but in some patients the colostomy is never closed. Identification of patients at risk for Hartmann's nonreversal would be helpful to determine the extent of resection. The aim of this study was to quantify the risk of nonclosure by deriving a predictive score. METHODS: Patients undergoing a Hartmann's procedure for diverticulitis were identified from database. They were separated into those who underwent Hartmann's reversal within one year of the initial operation, and those who did not. The data were analyzed in univariable and multivariable logistic regression. A predictive scoring system of Hartmann's reversal was created. The predictive power of the multivariable models, the predictive scoring system, and colorectal POSSUM physiology scores were compared. RESULTS: Eighty of 117 patients (68.4%) had their colostomy reversed. Multivariate analysis identified age, American Society of Anesthesiologists' score, pulmonary comorbidity, preoperative blood transfusion, perforation, and anticoagulants as the factors of failure for stoma reversal. All 36 patients with a predictive score of less than 14 had their stoma reversed. Twenty-two of 25 patients (88%) with scores greater than 18 did not have a reversal. Predictive power was similar when using only colorectal POSSUM physiology scores alone, or with preoperative data. CONCLUSIONS: More than 30% of patients undergoing a Hartmann's procedure for diverticulitis will not have their stoma reversed within a year. If this scoring system can be validated in an independent group of patients, it will be useful in allowing surgeons to strategize accurately and to counsel patients realistically.
