Resection margin involvement with Crohn ' s disease and postoperative anastomotic integrity ]

A prospective double-blind trial was conducted in Crohn's disease 
subjects in whom a resection was being performed, to assess the hypothesis 
that marginal disease would adversely influence healing of the anastomosis. Of 
106 eligible patients. 51 completed a protocol of pathological assessment of the 
surgical specimen and a water soluble contrast enema 10 to 15 days after the 
surgical procedure. Six were found to have radiological leaks and three additional 
subjects had clinical leaks from the anastomosis. The proportion of leaks, 
both clinical and radiological, was nine of 54. There was no trend to increasing rate 
of anastomotic breakdown with increasing marginal disease.


°C LECTIVE SURGICAL INTERVENTION IN
LCrohn's disease frequently involves resection of a segment of grossly involved intestine with restoration of continuity by an end-to-end anastomosis.There is considerable controversy over the intraoperative surgical decision on where the intestine should be cut.Some workers maintain that it is advisable to use all possible means, including frozen section.to ensure disease-free margins as this will decrease the incidence of immediate postoperative complications ( 1,2) and reduce the rare ofrecurrence (3-12).while others have not found any change in the rate of postoperative complications or of recurrence to correlate with the histological state of the resection margins (13)(14)(15)(16)(17)(18).
The authors have examined the short term complications of an intestinal anastomosis in Crohn's disease subjects relative to the presence of disease at the margins in a prospective double-blind trial, using a radiological technique to examine anastomotic incegrtty.The autho rs' findings support the view that disease-free margins have little advantage in improving anastomotic healing.

PATIENTS AND METHODS
In a three year period, all subjects undergoing elective resection of Crohn 's disease by 11 general surgeons were notified to a central office.As for as the authors are aware, rhese surgeons performed no resections for Crohn 's disease during this period which were not reported.If the proposed procedure involved an anastomosis to the colon (commonly ileocolic or colocolic) the patient was asked to cooperate in the trial according to a protocol approved by the University of Calgary Ethics Committee.One hundred and six subjects were approached and 96 agreed to take part in rhe trial.Of the remaining subjects who agreed to c"ter, three cases were excluded because the radiological diagnosis of Crohn 's disease was not confirmed at laparotomy and a further 17 were excluded because the actual procedure performed differed from that planned and a resection with anastomosis was not performed (commonly an ileostomy was carried out) Anastomoses were performed according to the preference o( the individual surgeon and were both stapled and sewn.Each resection specimen was examined by a pathologist in the fresh.unfixed ~rate.Having been opened, the segment was fixed after screrching on a cork board.Sections were taken as d ictated by the gross appearance and in a routine fas hion from each resection margin.A written report was made on each specimen without reference to eith er the operative record or the postoperative progress.
Marginal disease was defined as the microscopic or macroscopic presence of mucosa!ulceration or regeneration within 1 cm of the cut edge.together with either su bm ucosal or transm ucosal acute inflammation.Perivascular and serosal 96 acute inflammation was not taken into account.and generalized nonspecific increase in chronic inflammatory cells of the lamina propria was also ignored.
Ten to 15 days after the resection, each eligible subject had a water soluble contrast enema using an iodine containing proprietary contrast agent (iothalamate meglumine 17%).The radiologist was informed of the site of the a nastomosis.but had no access to the pathological diagnosis or the records of the subject's postoperative course.Prior to the x-ray, oral intake was restricted to fluids for 12 to 18 h, but no attempt was made at other bowel preparation.
Of the 76 subjects eligible following a suitable resection with anastomosis, three developed clinical signs and symptoms o( a major anastomotic leak before the 15th day and, as previously agreed, were not subjected to contrast radiology.In two subjects, the radiologist was unable to reach the site of the anastomosis and the examination was abandoned.A further l l subjects withdrew from the trial at this stage and nine did not have a contrast enema performed within the 10 to 15 day postoperative period because of miscellaneous problems, including missed appointments and administrative errors.This left 51 subjects in whom both pathological and radiologic assessment was completed and three who had pathological evaluation but developed a clinical leak. A radiological leak was defined as extravasation of contrast from the presumed anastomotic site for a distance of at least l cm.Trends in both clinical and radiological leaks were related to marginal involvement and are listed in Tab le l ( 19).

RESULTS
There were no deaths in the postoperative period in th is study.Of the 54 subjects who completed pathological evaluation and had contrast radiology.or had a clinical leak, margins were clear of d isease in 34 (63%).one margin was involved in 17 (31%) and both margins were involved in three (6%).Evidence of disru ption of anastomotic integrity were termed 'clinical' in three subjects (clear evidence ofleak of bowel contents in fi rst 15 days following surgical procedure) and 'radiological' in another five subjects (extravasation of contrast as defined above LO to 15 days after the surgical procedure).As noted elsewhere ( 20-24), radiological lea ks were not as• sociated with any discernible effect on the subjects' postoperative course.
The results are tabulated in Table l which shows a leak rate (both cl inical and radiological) of l7% when both margins were clear, 13% when one margin was involved and one leak in two cases where both margins were diseased.Assessment of subgroups according to anatomical segment resected, method of anastomosis (stapled versus sewn).in-Jivid ual surgeon's results, degree of disease activity in the resected intestine.sex or age provided no further useful information.
The three subjects with clinical leaks all were covered with parente ra l hydrocortisone in the perioperacive period: four of the five subjects with radiological leaks also had steroid cover.Of the 54 subjects who completed the trial, 48 received perioperative steroids.This small series, therefore, does not permit any conclusion about the role of perioperative steroid therapy on the presence or absence ofleaks at intestinal anastomoses in Crohn 's disease.
The proportion of e ligible subjects who chose not to participate in the trial, either when approached initially or who withdrew later, was high (21 of 106) and nine ochers failed to complete the protocol.Retrospective review of these patients showed no obvious differences from chose included in the analysis.

DISCUSSION
Evidence to suggest that histological Crohn 's disease at the intestinal margins The present findings support the view chat intestinal anastomoses heal without difficulty even in the presence of active disease ( l3-18) This concept ts upheld by reports (25.26) that plastic procedures carried out in areas of stricture caused by Crohn's disease (stricturoplasty) heal without excess of clinical leaks.Perhaps the finding is also a reflecuon of the view thatCrohn'sd1sease frequently involves most of the gastrointestinal tract, albeit at a low level of activity ( 2 7).a nd hence truly disease-free margins cannot be obtained.
The After colonic anastomoses, 'radiological' leaks, as defined m this study, appear to have no discernible effect on the subject's recovery from the procedure ( 20-24 ).Since these small leaks may be truly benign, it would be possible to ar• gue that any dern;ion on dis<.'ase-freemargins should nor be made wtth reference to radiological leaks.In fact, in this study the three clinical (as opposed to radiological) leaks occurred 111 subjects in whom the margins W<'re free from disease (Table I).
Du ring the rhree year period of this stu<ly, 106 patients were identified who were electively proposed for a resection of a segment of C ro hn 's disease with an anastomosis.Of this group only 5 I completed the protocol for a postoperative wate r soluble enema T wenry-one paucnts refused to participate and the others were excluded for miscellaneous reasons After review of the results, tt appeared unlikely that continuauon of the study to rncrease the numbers of parttnpants would add significantly to the conclusions and the trial was terminatl'd 18 Spt'ranza V S11rn Submit your manuscripts at http://www.hindawi.com CAN J GASTROENTEROL VOL 3 No 3 JUNE 1989 mcreascd the likelihood of anascomotic leaks has been primarily anec<lotal ( 1,2).