Back to Annual Meeting Posters

Background: Colorectal surgery is performed for a wide array of gastrointestinal disease, in patients with varied preoperative fitness. Preoperative diagnosis, in addition to patient and operative factors, may influence the rate of postoperative complications.
Aims: To determine predictors for bleeding, reoperation, and mortality after colorectal surgery.
Methods: This is a retrospective cohort study using the 2010 Nationwide Inpatient Sample (NIS), a nationally representative database of inpatient admissions. Adult patients were included if they had ICD-9 CM code for gastrointestinal malignancy, ulcerative colitis (UC), Crohn's disease (CD), ischemic colitis (IC), diverticulitis, diverticulosis, or diverticular bleeding. Inclusion was limited to patients with ICD-9 CM procedure code for colectomy, proctectomy, or colostomy. Charlson Comorbidity Index score was calculated for each patient. Outcomes included postoperative bleeding, reoperation, and mortality. Postoperative bleeding was defined as ICD-9 CM code for intraoperative bleeding, intraoperative hematoma, or lower gastrointestinal bleeding during the postoperative period in the setting of postoperative packed red blood cell transfusion. Univariate logistic regression models were performed to determine significant predictors, which were entered into multivariate logistic regression models controlling for patient demographics and hospital characteristics to obtain adjusted odds ratios (AOR). Statistical significance was established if p<0.05.
Results: 214,933 patients met inclusion criteria in 2010. Mean age was 60.3 ±0.2 years and mean Charlson score was 1.17 ±0.04. 1528 patients (0.7%) had postoperative bleeding requiring blood transfusion. 5439 patients (2.53%) required reoperation during the inpatient admission. 5715 patients (2.66%) experienced mortality during the inpatient admission.
Adjusted odds ratios for predictors of postoperative outcomes are shown in Table 1. Urgent or emergent admission, age >65, gastrointestinal malignancy, UC, and diverticulosis were significant predictors of postoperative bleeding. Urgent or emergent admission and IC were significant predictors of reoperation. Urgent or emergent admission, age >65, and IC were significant predictors of mortality.
Conclusion: Preoperative diagnosis, patient factors, and admission factors have significant association with postoperative outcomes after colorectal surgery. Urgent or emergent admission, age >65, and ischemic colitis are significant and independent predictors of mortality. Identification of modifiable factors that increase operative and postoperative risk in these patient groups would be of benefit in improving patient outcomes.
Table 1: Multivariable logistic regression

	Postoperative bleeding		Reoperation		Mortality
	AOR (95%CI)	p value	AOR (95%CI)	p value	AOR (95%CI)	p value
Urgent/emergent	2.8 (2.53-3.12) *	<0.01	2.0 (1.75-2.24) *	<0.01	3.6 (3.15-4.02) *	<0.01
Age >65	1.1 (1.06-1.19) *	<0.01	1.1 (0.98-1.20)	0.13	2.5 (2.24-2.79) *	<0.01
Malignancy	2.4 (2.17-2.57) *	<0.01	0.54 (0.46-0.63)	<0.01	0.21 (0.18-0.24)	<0.01
UC	2.0 (1.60-2.46) *	<0.01	0.90 (0.58-1.38)	0.61	0.49 (0.31-0.78)	<0.01
CD	1.1 (0.87-1.29)	0.55	0.45 (0.29-0.69)	<0.01	0.19 (0.10-0.34)	<0.01
IC	1.1 (0.93-1.21)	0.41	2.3 (1.88-2.72) *	<0.01	2.1 (1.83-2.39) *	<0.01
Diverticulitis	1.1 (0.96-1.19)	0.21	0.44 (0.37-0.52)	<0.01	0.24 (0.20-0.29)	<0.01
Diverticulosis	1.7 (1.42-1.96) *	<0.01	0.47 (0.29-0.75)	<0.01	0.40 (0.26-0.61)	<0.01

* denotes significant increase