Colon ischemia following abdominal aortic aneurysm repair in the era of endovascular abdominal aortic repair

Methods 

Since 1985, all vascular procedures performed in the Department of Vascular Surgery at Henri Mondor Hospital, University Paris XII, are stored prospectively in a specifically designed data base (Logit). The data base sheets include preoperative risk factors, anatomical features, intraoperative data, and early and late postoperative outcomes. Pre- and intraoperative data are filled by senior surgeons. In-hospital outcomes are filled by residents when patients leave the hospital. A weekly review of all cases allows a double check of the data and coded items, and finally, the medical staff reviews every 3 months the records of patients who have died or who had life threatening complications. In the data base, the definitions of preoperative risks factors fulfilled the criteria of the ad-hoc committee for reporting standards.19 The variables examined in this report were age, gender, preoperative risk factors (diabetes, coronary disease, ventricular dysfunction, renal insufficiency defined by serum creatinin >200 micromoles, pulmonary disease, previous colectomy), indication for surgery (ruptured vs elective), type of treatment (EVAR vs open), and operative details (duration of operation, blood loss, inferior mesenteric artery reconstruction, site of distal anastomoses, hypogastric artery embolization or coverage).

For the purpose of this study, we reviewed the 1174 consecutive infrarenal AAA patients who were treated between January 1995 and December 2005. A total of 682 (58%) were treated by open surgery and 492 (42%) by EVAR. When performing EVAR, we used commercially available stent grafts: Stentor (Mintec, Bahamas) n = 8 (1.6%), Talent (Medtronic, Santa Rosa, Calif) n = 8 (1.6%), Ancure (EVT Guidant, San Francisco, Calif) n = 14 (2.8%), AneurX (Medtronic) n = 24 (4.9%), Excluder (Gore, Flagstaff, Ariz) n = 25 (7%), Vanguard (Boston Scientific, Natick, Mass) n = 69, (13%), and Zenith (Cook) n = 334 (68%). All were EEC marked. Since 2001, the date of creation of this agency, all were approved by the AFSSAPS (Agence Française de Sécurité Sanitaire des Produits de Sante), the French version of the FDA in the United States.

In the open repair group, we used standard Dacron grafts (Datascope La Ciotat France or Edwards Maurepas 78000 France). Treatment allocation was influenced by anatomical factors and followed the French regulation, which so far, prohibits EVAR in relatively good risk patients unless they are enrolled in clinical trials. From a technical point of view, in open repair patients, we reattached fully patent IMA when the back flow was poor or when intraoperative Doppler ultrasound study of the colonic arcade failed to demonstrate pulsatile flow. In EVAR patients whose aneurysm extended below the iliac bifurcation, the internal iliac artery was covered and coil embolized prior to stent graft placement. However, in a few cases with occluded contralateral internal iliac, we bypassed the internal iliac artery from the ipsilateral common femoral artery. When coil embolization was performed, we strove, whenever feasible, for blocking the origin of the internal iliac artery.

Postoperatively, we did not perform systematic colonoscopy; however, our indications for colonoscopy were relatively liberal: patients with postoperative severe abdominal pain, early diarrhea, rectorrhagia, unexplained hemodynamic disturbances, unexplained organ failures, and/or elevated lactate enzymes were assessed for colonic ischemia either by colonoscopy and/or by redo surgery.

Colectomy was performed (1) at the time of AAA repair, in case of obvious transmural necrosis of the colon and (2) postoperatively, when colonic ischemia upon colonoscopy was either extensive or transmural or when CI was associated with clinical and/or biological symptoms. Observation with repeat colonoscopy was done in pauci symptomatic cases with ischemia seemingly limited to the mucosa. The records of patients identified with colonic ischemia in the data base were reviewed to assess clinical presentations and specific outcomes. Statistics were performed by the Department of Biomedical statistics of Henri Mondor Hospital (Dr Roudot-Thoraval). Statistical comparisons were made by means of χ2 test or Fisher exact test if appropriate for qualitative data, and Student t test or Mann Whitney nonparametric test if appropriate for continuous data. P values < .05 were considered significant. Variables associated with colonic ischemia in univariate analysis were tested in a stepwise logistic regression analysis. Adjusted odds ratios (ORs) with 95% confidence intervals were estimated from the model. Statistical analysis was performed with SPSS 13.0 and BMDP statistical software.

Results 

Among the 1174 consecutive infrarenal AAA patients, they were 1097 (93, 4%) males and 77 females. Mean age was 70.8 ± 9.0 years. Six hundred and eighty-two were treated by open surgery and 492 by EVAR. Postoperative colonic ischemia occurred in 34 patients (2.9%). Colonic ischemia was diagnosed at the time of aneurysm repair in five cases (15%): three were ruptured aneurysms and two were elective difficult cases. Colectomy was performed at the end of the intervention due to obvious transmural necrosis. Postoperative colonic ischemia was diagnosed in 29 patients (85%): by colonoscopy in 21/29 cases (72%) and by early reoperation in eight cases. For those patients with colonic ischemia diagnosed postoperatively, diarrhea was present in 15/29 (53%), rectorrhagia in 9 (31%), septic shock in 16/29 (57%), persistent low blood pressure with multiple organ failure in 24/29 (84%), and fever (temperature >38° Celsius) in 20/29 (69%). Review of biological parameters showed a high level of lactates in 23/29 cases (81%), elevated creatinin (ie, increase by at least 30% of the preoperative value) in 18/29 (65%), and elevated leucocytes (>10,000 c/mm3) in 20/29 (72%).

Table I shows the statistically significant variables identified in the univariate analysis. In patients without CI and in patients with CI, the median (quartiles range) blood loss was 1000 ml (500-1800) and 2000 ml (650-3350), respectively (P =.008). The median (quartiles range) operating time was 150 mn (120-180) and 180 mn (120-255), respectively (P = .0005). Analysis of technical factors with regard to CI occurrence is shown in Table II.

Table I. Predictive risk factor of colonic ischemia following AAA repair (univariate analysis)

	(N° at risk)	% Colon ischemia	P value
	Gender:
	Male (1097)	2.8	.59
	Female (77)	3.9
	Diabetes:
	No (1046)	2.9	.76
	Yes (116)	1.7
	Coronary disease:
	No or asymptomatic (874)	2.3	.27
	Angina pectoris stable or unstable (254)	3.5
	Left ventricular function:
	Normal or mild (1002)	2.5	.66
	Severely impaired (59)	3.4
	Kidney:
	Creatinin < 200 μmol (1133)	2.7	.01
	Creatinin > 200 μmol (30)	13.3
	Respiratory function:
	Normal or moderately impaired (1005)	2.5	.07
	Severely impaired (157)	5.1
	Previous colectomy: (44)	4	.89
	Presentation:
	Nonruptured (1086)	2.1	<.0001
	Ruptured (88)	12.5
	Operative duration:
	≤2 h (518)	1.7
	2-4 h (558)	2.9	.001
	>4 h (66)	13.6
	Procedure:
	Open surgery (682)	4	.01
	EVAR (492)	1.4

AAA, Abdominal aortic aneurysm; EVAR, endovascular abdominal aortic repair.

Table II. Technical details of patients treated by open repair and by EVAR regarding the occurrence of colon ischemia

	Open repair (N° at risk)	% Colon ischemia	P value
	Tube graft (239)	4.02
	Aorto uni or bi iliac graft (395)	2.7
	Aorto bifemoral graft (21)	22	.001
	IMA reimplantation
	Yes (39)	4
	No (602)	3.9	ns
	EVAR
	Hypogastric artery loss
	No (379)	.7
	Yes (106)	3.6	.046

EVAR, Endovascular abdominal aortic repair; IMA, inferior mesenteric artery.

Table III shows the results of the multivariate analysis. Aneurysm rupture, length of operation (>4 hours), and preoperative renal insufficiency were independently associated with postoperative colonic ischemia. The prognosis of colonic infarction was poor. Eighteen patients (52%) died within the first postoperative month. The mortality due to colonic ischemia was not statistically different between open surgery 14/27 (52%) and EVAR 4/7 (50%).

Table III. Multivariate analysis of factors associated with colonic ischemia following AAA repair

	Risk factors	Odd ratio	IC 95% Odds ratio	P value
	Presentation:
	Nonruptured	1⁎	—
	Ruptured	6.03	2.68–13.5	.0001
	Operative duration:
	≤2 h (518)	1⁎
	2-4 h (558)	1.41	.60–3.28	.43
	>4 h (66)	5.73	2.06–15.9	.001
	Kidney:
	Creatinin ≤ 200 μmol	1⁎	—
	Creatinin > 200 μmol	4.67	1.39–15.7	.028

⁎ Reference category.

Twenty patients (59%) were treated by colectomy. Twelve of them died (60%). Colon resection was performed between 0 and 13 days with the majority at day 3. In six patients, colectomy was not performed because of extensive necrosis and multiple organ failure. All of them died. In the eight remaining cases, conservative therapy was attempted because ischemia was deemed less extensive. None of these patients died. Colonic continuity was re-established in five out of six surviving patients after colonic resection. Major postoperative complications were seen in the vast majority of patients with colonic ischemia. Cardiac complications occurred in 18/34 (53%), pulmonary complications in 20/34 (61%), and renal insufficiency in 20/34 (61%) of whom 14 (41%) required hemodialysis.

Discussion 

Colonic necrosis remains an infrequent but devastating complication following AAA repair. In the current series, which included open and EVAR patients, the overall rate of colonic ischemia was 2.9%. This rate is similar to previous reports investigating open surgery before EVAR was introduced. It was 1% out of 1420 patients in Brewster’s series,1 and 2.6% out of 1800 patients in the SWEDAC registry.2 In EVAR series, the rate of colonic ischemia was 2.9% out of 278 patients for Dadian14 and 1.2% out of 311 patients for Maldonado.13 In the univariate analysis, we found more colonic ischemia following open surgery (4%) than after EVAR (1.4%). However, in the multivariate analysis the type of AAA repair was not an independent variable, underlining the weight of stronger parameters such as rupture, length of operation, and renal insufficiency.

Conclusion 

Colonic ischemia remains a serious complication following AAA repair. Although in univariate analysis, EVAR was associated with a lower rate of colonic ischemia, multivariate analysis showed that the sole independent factors of CI were rupture, long duration of operation, and prior renal disease. Of note, within the two treatment modalities, the mortality rate was similar.

Author contributions