From the Eastern Vascular Society
neurs ofof thof coth aned wia. P
nstrmortry. Ced inopy.
bowel resection because of grade III ischemia; two procedures were performed because of worsening ischemia discoveredat repeat colonoscopy. In patients with colonoscopic findings of bowel ischemia the mortality rate was 50% (13 of 26patients). In those with grade III necrosis who underwent resection the mortality rate was 55%. Elevated lactate levels,
79immature white blood cells, and increased fluid sequestration were all variables associated with the occurrence of colonischemia.Conclusions: Bowel ischemia is a frequent postoperative complication (42%) of repaired rAAA. Performing mandatorysurveillance colonoscopy in these patients may be associated with a decrease in overall mortality and improved survival inpatients with transmural bowel necrosis with no comorbid condition. (J Vasc Surg 2004;39:792-6.)
Surgical repair of a ruptured abdominal aortic aneu-m (rAAA) continues to be associated with high morbid-and mortality.1 Despite improvements in pre-hospitale, changes in operative strategies, and advancements instoperative care, the mortality of rAAA repair has notnged over several decades, with a rate of 50% to 70%.1,2
tients who survive rAAA are susceptible to a variety ofrbid conditions, including colonic ischemia. Clinically andoscopically proved ischemia of the colon complicates elec-
e reconstruction in 1% to 7% of cases, and 60% of patientsh rAAA.3 Prospective studies with selective colonoscopyve documented ischemic mucosal changes in 3% to 20%patients undergoing elective aortic procedures, and in asny as 60% of patients with rAAA. When transmuralcrosis is present the mortality rate can be as high as 90%.4
Although postoperative clinical assessment with physi-cal examination and laboratory tests is unreliable in predict-ing ischemic colitis, several intraoperative methods havebeen promising. These include inferior mesenteric artery(IMA) stump pressure measurements, transserosal tissueoxygen tension measurements (tPO2), laparoscopy, andtonometry.2,4,5 Of these, selective ligation of the IMA onthe basis of intraoperative bowel inspection, colonic mes-enteric Doppler signals, and IMA stump pressure has beenthe most encouraging.6 However, use of these techniquesto predict colon ischemia are single intraoperative measure-ments and may not reflect subsequent ischemic events.
Colonoscopy remains the diagnostic procedure ofchoice for assessing ischemic colitis. Although clinical pre-dictors of ischemic colitis after rAAA are lacking and mor-tality is high, mandatory routine colonoscopy after rAAAhas not been universally adopted.
Most reports of ischemic colitis after rAAA are based onpatients with profound clinical symptoms. We instituted anaggressive approach with routine colonoscopy after rAAAin an attempt to identify the true incidence of postoperativebowel ischemia and to reduce overall mortality throughearly detection and treatment.
m the Institute for Vascular Health and Disease, Albany Medical College.mpetition of interest: none.sented at the Seventeenth Annual Meeting of the Eastern Vascularociety, New York, NY, May 1-4, 2003.rint requests: R. Clement Darling III, MD, Vascular Institute (MC157),lbany Medical College, 47 New Scotland Ave, Albany, NY 12208
pyright 2004 by The Society for Vascular Surgery.:10.1016/j.jvs.2003.12.002
2Outcome of aggressive suruptured abdominal aorticBradley J. Champagne, MD, R. Clement Darling III,MD, Edward C. Lee, MD, Manish Mehta, MD, MPHPhilip S. K. Paty, MD, Kathleen J. Ozsvath, MD, and
Purpose: Emergent repair of ruptured abdominal aortic amorbidity and mortality. One of the significant complicationas high as 80% when this condition is realized. The objectiveeffect of mandatory postoperative colonoscopy on outcomeMethods: From July 1995 to September 2002 all patients wiwere included in this review. All colonoscopies were performtreatment was initiated per protocol based on grade of ischemwith medical management and repeat colonoscopy. All patPreoperative, intraoperative, and postoperative variables werpredictors of bowel ischemia.Results: Eighty-eight patients underwent emergent aortic recoage was 73 years, and 64 patients (72%) were men. Operative24 hours, and 24% died between 1 and 30 days after surgesurvived more than 24 hours. Bowel ischemia was documentgrade I or grade II ischemia at both initial and repeat endosceillance colonoscopy inneurysm, Mani Daneshmand, BS, Paul B. Kreienberg,an P. Roddy, MD, Benjamin B. Chang, MD,iraj M. Shah, MD, Albany, NY
ysms (rAAAs) is associated with high perioperativethis surgery is bowel ischemia. Reports detail mortalityis project was to determine both the incidence and thelon ischemia after rAAA.rAAA who underwent emergent aortic reconstructionithin 48 hours, ischemia was graded consistently, andatients with grades I and II ischemia were followed upwith grade III ischemia underwent bowel resection.
ected to assess possible independent risk factors for and
uction because of rAAA in the study period. Their meanality was 42%. Eighteen percent of patients died withinolonoscopy was performed in 62 of 72 patients who26 of the 72 patients (36%). Of these, 16 patients hadNine patients underwent exploratory laparotomy with
JOURNAL OF VASCULAR SURGERYVolume 39, Number 4 Champagne et al 793ETHODS
Using data from our vascular surgery registry, we ret-pectively reviewed the charts of all patients with rAAAo underwent repair between July 1995 and September02. Data were evaluated for demographics and postop-tive factors, including symptoms, laboratory values, he-dynamics, aortic crossclamp time, perioperative hypo-sion, blood transfusions, and fluid sequestration.
rioperative hypotension was defined as mean arterialod pressure less than 65 mm Hg for at least 10 minutes.ta were analyzed with an unpaired Student t test, and.05 indicated statistical significance.Patients with rAAA underwent emergent repair via a
t retroperitoneal approach, as described by Shah et al.7
sessment of ischemic colitis included visual inspectiond Doppler scan evaluation of mesenteric blood flow. Allients with rAAA were offered colonoscopy within 48urs of completion of the repair, regardless of clinical andoratory findings. In all cases in which colonoscopy wast performed the reason was patient or family refusal.The bowel was prepared with a sodium phosphate
ema (Clisma Fleet, 133 mL) to avoid fecal soiling. Tapter enemas were used in patients with evidence of renalsfunction (creatinine concentration 2). Colonoscopys performed to at least 40 cm in all patients, except whenep ulcers or suspected necrosis was identified, to avoidrforation. The rectosigmoid junction was always exam-d.Furthermore, all instances of colon ischemia were
ded consistently on the basis of published criteria. Gradechemia was defined as mucosal ischemia; grade II isch-ia was characterized by involvement of the mucosa andmuscularis layers; and grade III ischemia was described
transmural ischemia, gangrene, and perforations.Severe ischemia (grade III) was treated with immediateloratory laparotomy and resection of the ischemic bowelue. Mild and moderate ischemia (grades I and II) wasnaged nonoperatively, and repeat colonoscopy was per-med every 48 hours until improvement, to ascertain theponse to management. Nonoperative management con-ed of a conservative regimen of bowel rest, intravenous
dration, parenteral nutrition, and broad-spectrum anti-tic therapy. Low-grade ischemia that progressed at re-
at colonoscopy was treated with exploratory laparotomyd resection of the ischemic bowel tissue.
From July 1995 to September 2002, 88 patients (64n, 24 women) underwent rAAA repair. Their averagewas 73.4 years. Postoperative mortality was 18% (16 ofpatients) within 24 hours and 24% (21 of 88 patients)hin 30 days. Seventy-two patients survived longer thanhours and were offered colonoscopy. Sixty-two patients%) underwent colonoscopy, 9 patients (13%) declinedprocedure, and 1 patient required laparotomy and artmann procedure while awaiting colonoscopy. riskSixty-one percent (38 of 62 patients) had no evidencecolonic ischemia. Twenty-nine percent (18 of 62 pa-nts) had evidence of grade I or grade II ischemia, and% (6 of 62 patients) had evidence of grade III ischemia attial colonoscopy. Of 38 patients with normal findings atlonoscopy, cholecystitis developed in 1, and in this pa-nt transmural colonic necrosis was discovered at laparot-y. Of 18 patients with grade I or II ischemia, 2 patients%) had progression to grade III ischemia at repeat
lonoscopy performed within 48 hours. Overall, transmu-ischemia (grade III) developed in 10 patients (14%).
ne required laparotomy and a Hartmann procedure, ande family declined further intervention and withdrewport (Table I).Of 37 patients without evidence of ischemic colitis, 10
ients (27%) died secondary to adult respiratory distressdrome (n 3), multisystem organ failure with with-wal of support (n 3), cardiac arrest (n 2), cerebralcular accident (n 1), or massive gastrointestinal hem-hage (n 1). Of 16 patients with grade I or II ischemia,atients (31%) died, of cardiac arrest (n 2), multipletem organ failure (n 1), or respiratory embarrassment 1); 1 patient was declared brain dead, and the familyhdrew support on postoperative day 3. Of nine patientso underwent bowel resection to treat grade III ischemia,
e patients (55%) died secondary to multiple system organlure (n 4) or cardiac arrest (n 1). Four of these fiveients had combined colon and rectal necrosis at laparot-y. Only one patient with ischemia limited to the colond after laparotomy and Hartmann procedure (Table II).Logistic regression analysis for all grades of ischemia,
mparing patients with documented colonic ischemiah those with no evidence of ischemia, is summarized inble III. Elevations in neutrophilic bandemia, serum lac-e, and fluid sequestration all correlated with an increase
ble I. Colonoscopy
clined colonoscopy 9 12.5nical findings warrantedmergent celiotomy
thout Colonoscopyevidence of colonic ischemia 38 52.8
lon ischemia found ateliotomy to treatholecystitis
ade I or grade II ischemia 18 25.0ade III ischemia at repeatolonoscopy
ade III ischemia 6 8.3tal number of patientsndergoing colonoscopy
tal number of patientsffered colonoscopy
wel ischemia found atolonoscopy
wel ischemia foundndependent of colonoscopy
2for development of bowel ischemia, to 95% confidence.
JOURNAL OF VASCULAR SURGERYApril 2004794 Champagne et alriables that increase the risk for development of transmu-necrosis (grade III) are addressed in Table IV. The
tistically significant variables, with 95% confidence, werentical to those in Table IV, with the addition of preop-tive hypotension. The odds of development of grade IIIcrosis increase by a factor of 1.35 for every 10 minutes ofoperative hypotension.
The diagnosis and early recognition of ischemic colitiser rAAA is necessary to avoid the profound morbidityd mortality associated with transmural bowel necrosis.tients described in four recent large series who survived
initial operation for repair had an incidence of colonichemia ranging from 15% to 35%.8 This disastrous conse-ence has resulted in 90% to 100% mortality in previousdies.4 Despite increased awareness of this fatal postop-tive complication and the adoption of various intraop-tive methods to predict bowel ischemia, most centersntinue to report overall mortality rates greater than 50%.ose patients with full-thickness necrosis require prompt
ble II. Mortality: 1 to 30 days
ischemia 10 21.7ade I or grade II ischemia 5 31.2erall grade III ischemia 6 60.0ated grade III ischemia 5 55.5
ble III. Binary logistic regression: Odds of grade I, II,III ischemia
Odds ratio Lower Upper
operative hypotension10 min)
1.26 0.99 1.61
raoperative hypotension10 min)
1.27 0.98 1.65
ssclamp time (min) 0.95 0.84 1.09od loss (L) 1.06 0.87 1.29
stoperative day 1max 1.13 0.80 1.59BC 0.99 0.90 1.09ands 1.08 1.02 1.14actate (mmol/L) 2.15 1.27 3.65luid sequestration (L) 1.17 1.03 1.33
stoperative day 2max 1.01 0.70 1.44BC 0.94 0.83 1.07ands 1.07 1.02 1.14actate (mmol/L) 2.97 1.37 6.44luid sequestration (L) 1.41 1.09 1.82
dface denotes statistical significance to 95% confidence.ax , Maximum temperature (F) over 24 hours; WBC, peripheral whiteod cell count, reported as thousands of cells per high-power field; bands,centage of peripheral band neutrophils.loration and resection, whereas those with ischemia Thited to the mucosa will recover without operative resec-n.The pathogenesis of intestinal ischemia after rAAAears to be multifactorial. Long periods of hypotension
d preoperative shock, extrinsic mesenteric compressionm a large retroperitoneal hematoma, improper ligationthe IMA during reconstruction, embolization duringture or repair, and operative trauma all have been impli-ed as potential risk factors.Previous attempts to detect preoperative and postoper-
ve risk factors have been made. Piotrowski et al4 foundt colon ischemia was more prevalent in patients withoperative shock and greater intraoperative blood loss.issner and Johansen9 found no demographic, clinical, or
erative factors that correlated with the development ofhemia, other than low perioperative cardiac output.eze et al10 found that D-lactate levels may be predictorsbowel ischemia after rAAA at admission to the intensivee unit. More recently, Levison et al11 demonstrated thatlon ischemia after rAAA may be predicted with thesence of two or more specific perioperative factors.In the present study, increased immature white blood
ls, elevated lactate levels, and fluid sequestration allreased the odds for development of bowel ischemiaable III). These variables were addressed independentlyh logistic regression analysis. Although an associations observed, this does not imply that elevations in theseues are specific for bowel ischemia. Furthermore, 2 of 10ients in whom grade III ischemia developed did not
monstrate elevations in band count or lactate levels.
ble IV. Binary logistic regression: Odds of grade IIIhemia
Odds ratio Lower Upper
operative hypotension10 min)
1.35 1.02 1.78
raoperative hypotension10 min)
1.33 0.94 1.71
ssclamp time (min) 1.01 0.87 1.18od loss (L) 1.37 0.95 1.80
stoperative day 1max 0.97 0.62 1.52BC 0.96 0.83 1.12ands 1.07 1.01 1.14actate (mmol/L) 2.01 1.21 3.34luid sequestration (L) 1.21 1.05 1.39
stoperative day 2max 1.21 0.69 2.12BC 0.88 0.71 1.08ands 1.16 1.06 1.28actate (mmol/L) 2.5 1.21 5.18luid sequestration (L) 1.62 1.15 2.30
dface denotes statistical significance to 95% confidence.ax , Maximum temperature (F) over 24 hours; WBC, peripheral whiteod cell count, reported as thousands of cells per high-power field; bands,centage of peripheral band neutrophils.erefore, although the presence of these elevated values
JOURNAL OF VASCULAR SURGERYVolume 39, Number 4 Champagne et al 795rioperatively should heighten the suspicion for intestinalhemia, they do not seem to be specific enough to deter-ne the need for repeat exploration. No statistical signif-nce was found between groups with and without anywel ischemia for the categories of aortic clamp time,rioperative blood pressure, and postoperative tempera-e measurements.Variables that increased the odds for development of
nsmural necrosis (grade III) are specifically addressed inble IV. The same risk factors discussed above, with thedition of preoperative hypotension, were found to in-ase the odds of requiring laparotomy in these 10 pa-nts. This interesting finding may suggest that the mostmpelling risk factor for transmural necrosis occurs beforeerative repair is undertaken.
More recently, alternative techniques for diagnosis ofestinal ischemia after rAAA have been proposed. Some ofse methods include laser duplex flometry, pressure mea-ements in the stump of the IMA, and intraluminalO2-pH ratios. Despite early optimism for the utility of
er duplex flometry, the variability in colonic measure-nts obtained limits its usefulness for detecting levels of
lonic perfusion.12 Intraoperative IMA stump pressuresy enable determination of when safe ligation of thissel is feasible.2 Mean pressure less than 40 mm Hg hasen associated with the possible development of ischemiclitis in previous studies. However, dependence on thismber assumes that all subsequent mean arterial pressureordings will be at least that of the pressure recorded attime of the measurement. Fiddian-Green et al13 studieduse of intracolonic silicone tonometry to measure in-
luminal PCO2 as an indirect measure of intramural pH inpatients at high risk undergoing aortic surgery. Theynd that a low normal pH was predictive of ischemic
litis and that the duration of pH evidence for ischemia onday of operation was the best predictor for the devel-
ment of signs and symptoms of ischemic colitis and forath after operation.9 This technique was complex andmbersome, and has not been subsequently studied. Indition, it was not predictive of which patients woulduire colonic resection. Use of these intraoperative meth-s to predict colon ischemia is limited in that they aree-time calculations and will not reflect subsequent isch-ic events.Hagihara et al14 first reported the use of flexible
lonoscopy to establish the diagnosis of ischemic colitis inctive and rAAA. As described, colonoscopy to 40 cm isficient to detect greater than 95% of colon ischemia afterrtic reconstruction.2 In the same year, Forde et al15
nd that the endoscopic appearance at colonoscopy cor-ated well with subsequent microscopic pathologic find-s. More recently, Brandt et al16 used flexible sigmoidos-
py on a selective basis when there was clinical suspicion ofwel ischemia, and abnormalities were found in 14 of 18ients. Since that time several authors have recom-nded routine postoperative lower endoscopy in all pa-
nts who initially survive repair of rAAA, although the fore incidence of bowel ischemia with this method has noten reported. Furthermore, no reports to date have doc-ented the effect of mandatory colonoscopy after rAAAsurvival in patients with transmural necrosis.As described, the lack of a precise and sensitive intraop-
tive or postoperative predictive measure of bowel isch-ia in rAAA has failed to improve the high mortalityociated with this condition. In our study we took anressive approach to identifying all patients with ischemic
litis by performing mandatory colonoscopy in all patientsh rAAA who survived 24 hours, over follow-up of 7rs.Results of this study suggest that the overall incidence
bowel ischemia in those patients who survive greatern 24 hours after rAAA is 36%. It should be mentionedt nine patients who survived longer than 24 hours didt receive colonoscopy secondary to family or patientusal or other random factors. These patients all survived,it can safely be assumed that grade III necrosis did notvelop, yet occurrence of grade I or grade II disease mayve gone undetected. The average distance of 40 cm forlonoscopy should be adequate for detection of ischemiclitis after abdominal aortic reconstruction, inasmuch as
lesions center on the rectosigmoid junction. The sensi-ity of colonoscopy in detecting bowel ischemia was 96%.e patient had negative findings at colonoscopy, and
wel necrosis was discovered at cholecystectomy. In thisient it is impossible to determine the true cause ofptoms and the declining clinical picture, because of
th disease processes. Furthermore, there were no com-cations from colonoscopy.
Mortality from transmural necrosis after rAAA has beennsistently reported at 80% to 100%.17 In unpublishedta from our vascular registry (1985-1995), the mortalitye from this disastrous consequence was 75%. In thesent study the mortality rate from grade III necrosis wasculated to be 55% (five of nine patients) for those whoderwent resection. All patients but one with bowel ne-sis limited to the sigmoid region survived. Four of the
e patients who died after laparotomy had some involve-nt of the rectum. In addition, the overall mortality rate42% after rAAA in this study is less than that in mostviously reported studies.In conclusion, the high incidence of and mortality from
hemic colitis after rAAA is well-recognized. In our expe-nce, the use of routine surveillance colonoscopy can bed safely, and enables early treatment of colon ischemia.e treatment of grades I and II bowel ischemia withtibiotic therapy, bowel rest, and repeat colonoscopy, andmediate laparotomy in patients with grade III ischemia,y decrease the mortality associated with this condition.
better demonstrate the effectiveness of mandatorylonoscopy, a prospective trial including a control groupdergoing laparotomy based on clinical variables only androup undergoing mandatory colonoscopy could be per-
1. Hsiang YN, Turnbull RG, Nicholls SC, McCullough K, Chen JC,Lokananthan R, et al. Predicting death from ruptured abdominal aortic
abdominal aortic aneurysms. J Vasc Surg 1996;24:614-23.
9. Meissner MH, Johansen KH. Colon infarction after ruptured abdomi-nal aortic aneurysm. Arch Surg 1992;126:979-85.
10. Poeze M, Froon AH, Greve JW, Ramsay G. D-Lactate as an early
Available online Feb 9, 2004.
JOURNAL OF VASCULAR SURGERYApril 2004796 Champagne et alMENTORED PATIENT-ORIENTED RESEA
The Lifeline Foundation and the National Heannounce their joint sponsorship of the NHLBI Mement Award (K23), which will provide supplementastages of their research careers.
The purpose of the Mentored Patient-Orientedsupport the career development of investigators whendeavors on patient-oriented research. The objecCareer Development Award (K23) are (1) to encourdent research skills and gain experience in advancedconduct patient-oriented research and (2) to increapatient-oriented studies, capitalizing on the discoveclinical settings.
The application deadline is June 1, 2004. For email@example.com or visit VascularWeb (httH CAREER DEVELOPMENT AWARD
Lung and Blood Institute (NHLBI) wish toed Patient-Oriented Research Career Develop-ding for vascular surgeon scientists in the early
earch Career Development Award (K23) is tove made a commitment to focus their researchs of the Mentored Patient-Oriented Researchresearch-oriented clinicians to develop indepen-thods and experimental approaches needed toe pool of clinical researchers who can conductof biomedical research and translating them to
information, contact the Lifeline Foundation atwww.vascularweb.org/).aneurysms. Am J Surg 2001;181:30-5.Ernst CB, Hagihara PF, Daughtery ME, Sachatelo CR, Griffen WO Jr.Ischemic colitis incidence following abdominal aortic reconstruction: aprospective study. Surgery 1976;80:417-21.Zelenock GB, Strodel WE, Knol JA, Messina LM, Wakefield TW,Lindenauer SM, et al. A prospective study of clinically and endoscop-ically documented colonic ischemia in 100 patients undergoingaortic reconstructive surgery with aggressive colonic and direct pelvicrevascularization, compared with historic controls. Surgery 1989;106:771-9.Piotrowski JJ, Ripepi AJ, Yuhas JP, Alexander JJ, Brandt CP. Colonicischemia: the Achilles heel of ruptured aortic aneurysm repair. Am Surg1996;62:557-61.Sheridan WG, Lowndes RH, Young HL. Intraoperative tissue oximetryin the human gastrointestinal tract. Am J Surg 1990;159:314-9.Iberti TJ, Salky BA, Onofrey D. Use of bedside laparoscopy to identifyintestinal ischemia in postoperative cases of aortic reconstruction. Sur-gery 1989;105:686-9.Shah DM, Chang BB, Paty PS, Kaufman JL, Kaslow AR, Leather RP.Treatment of abdominal aortic aneurysm by exclusion and bypass: ananalysis of outcome. J Vasc Surg 1991;13:15-22.Chen JC, Hildebrand HD, Salvian AJ, Taylor DC, Strandberg S,Myckatyn TM, et al. Predictors of death in nonruptured and ruptured
marker of intestinal ischemia after ruptured aortic aneurysm repair. Br JSurg 1998;85:1221-4.
11. Levison JA, Halpern VJ, Kline RG, Faust GR, Cohen JR. Perioperativepredictors of colonic ischemia after ruptured abdominal aortic aneu-rysm. J Vasc Surg 1999;29:40-5.
12. Redaelli CA, Schilling MK, Carrel TP. Intraoperative assessment ofintestinal viability by laser Doppler flowmetry for surgery of rupturedabdominal aortic aneurysms. World J Surg 1998;22:283-9.
13. Fiddian-Green RG, Amelin PM, Herrmann JB, Arous E, Cutler BS,Schiedler M, et al. Prediction in the development of sigmoid ischemiaon the day of aortic operations: indirect measurements of intramural pHin the colon. Arch Surg 1986;121:654-60.
14. Hagihara PF, Ernst CB, Griffen WO Jr. Incidence of ischemic colitisfollowing abdominal aortic reconstruction. Surg Gynecol Obstet 1979;149:571-3.
15. Forde KA, Lebwohl O, Wolff M, Voorhees AB. The endoscopy corner:reversible ischemic colitis; correlation of colonoscopic and pathologicchanges. Am J Gastroenterol 1979;72:182-5.
16. Brandt CP, Piotrowski JJ, Alexander JJ. Flexible sigmoidoscopy: areliable determinant of colonic ischemia following ruptured abdominalaortic aneurysm. Surg Endosc 1997;11:113-5.
17. Tollefson DFJ, Ernst CB. Colon ischemia following aortic reconstruc-tion. Ann Vasc Surg 1991;5:485-9.
Submitted May 14, 2003; accepted Dec 4, 2003.
Outcome of aggressive surveillance colonoscopy in ruptured abdominal aortic aneurysmMETHODSRESULTSDISCUSSIONREFERENCES