Large bowel obstruction: predictive factors for postoperative mortality.

PURPOSE

The aims of this study were to assess the prognostic value for mortality of several factors in patients with colonic obstruction and to study the differences between proximal and distal obstruction.

METHODS

Two-hundred and thirty-four consecutive patients who underwent emergency surgery for colonic obstruction were studied. Patients with an obstructive lesion distal to the splenic flexure were assessed as having a distal colonic obstruction. Resection and primary anastomosis was the operation of choice in selected patients. Alternative procedures were Hartmann's procedure in high-risk patients, subtotal colectomy in cases of associated proximal colonic damage, and colostomy or intestinal bypass in the presence of irresectable lesions. Obstruction was considered proximal when the tumor was situated at the splenic flexure or proximally and a right or extended right colectomy was performed. A range of factors were investigated to estimate the probability of death: gender, age, American Society of Anesthesiologists score, nature of obstruction (benign vs. malign), location of the lesion (proximal vs. distal), associated proximal colonic damage and/or peritonitis, preoperative transfusion, preoperative renal failure, and laboratory data (hematocrit < or = 30 percent, hemoglobin < or = 10 g/dl, and leukocyte count >15,000/mm3). Univariate and multivariate forward steptwise logistic regression analysis was used to study the prognostic value of each significant variable in terms of mortality.

RESULTS

One or more complications were detected in 109 patients (46.5 percent). Death occurred in 44 patients (18.8 percent). No differences were observed between proximal and distal obstruction. Age (>70 years), American Society of Anesthesiologists III-IV score, preoperative renal failure, and the presence of proximal colon damage with or without peritonitis were significantly associated with postoperative mortality in the univariate analysis. Only American Society of Anesthesiologists score, presence of proximal colon damage, and preoperative renal failure were significant predictors of outcome in multivariate logistic regression.

CONCLUSION

Large bowel obstruction still has a high of mortality rate. An accurate preoperative evaluation of severity factors might allow stratification of patients in terms of their mortality risk and help in the decision-making process for treatment. Such an evaluation would also enable better comparison between studies performed by different authors. Principles and stratification similar to those of distal lesions should be considered in patients with proximal colonic obstruction.