INDICATIONS FOR SURGERY IN CROHN'S DISEASE. Analysis of 500 cases

Transcription

1 GASTROENTEROLOGV 71:24f1--250, by The Williams & Wilkins Co. Vol. 71, No. 2 Print.din U.S.A. INDICATIONS FOR SURGERY IN CROHN'S DISEASE Analysis of 500 cases RICHARD G. FARMER, M.D., WILLIAM A. HAWK, M.D., AND RUPERT B. TURNBULL, JR., M.D. Department of Gastroenterology, Department of Pathology, and Department of Colon and Rectal Surgery, The Clevelond Clinic Foundation and The Cleveland Clinic Educational Foundation, Cleveland, Ohio Data for 500 patients with Crohn's disease who underwent operations were analyzed (316 patients, 1966 to 1969; 184 patients, 1972 to 1973) by comparison of various anatomic disease locations (clinical pattern): (1) ileocolic, 225 patients; (2) small intestinal, 130 patients; (3) colonic, 127 patients; (4) anorectal, 18 patients. Indications for surgery were tabulated and compared using statistical analysis for the three large patterns. For patients with ileocolic Crohn's disease, the primary surgical indications were internal fistula and abscess, 44%, intestinal obstruction, 35%, and perianal disease, 12%. For patients with Crohn's disease of the small intestine, the primary surgical indications were intestinal obstruction, 55%, and intestinal fistula and abscess, 32%. Patients with colonic Crohn's disease had a significantly more diverse surgical indication, with poor response to medical therapy, 26%, internal fistula and abscess, 23%, toxic megacolon, 20%, and perianal disease, 19%. These values were highly statistically significant (P < 0.00(1) in all instances but one. This study demonstrates that statistically significant differences occur in the surgical indication depending on the location of Crohn's disease. Patients with ileocolic, small intestinal, and colonic involvement have striking differences in clinical course. It is concluded that Crohn's disease is not a homogeneous entity, but should be recognized as having a varying course depending on clinical pattern. Treatment of Crohn's disease continues to be a difficult problem. Results of medical therapy are often unsatisfactory, and many patients require operation. Attempts have been made to define the reasons for operation, to determine which operation should be done at what time in the course of disease, and to assess results of the operation. Despite many studies, there is still confusion, and considerable subjectivity in decisionmaking continues mainly because of the capriciousness of the disease. Difficulty in determining the indication for surgery in Crohn's disease has been exemplified by two recent publications. In a general discussion of surgery for Crohn's disease, Janowitz i stated that patients were "treated surgically for the major catastrophies of this disease (hemorrhage, obstruction, or perforation), and even more frequently for their failure to respond to the nonspecific medical agents." Glotzer and Silen,2 in regard to Crohn's disease of the colon, reported that "intractability is the most frequent indication for operation, as it is in ulcerative colitis." Two long term studies of patients with Crohn's disease, from England and Received September 29, Accepted January 13, Address requests for reprints to: Dr. Richard G. Farmer, Depart Ilient of Gastroenterology, The Cleveland Clinic Foundation, 9500 EUclid Avenue, Cleveland, Ohio The authors thank J. N. Berrettoni, Ph.D., Professor of Statistics, Case Western Reserve University, for his assistance in the statistical 8tudy. 245 Sweden, respectively, showed that 90% required operation, but, again, mainly because of presumed poor response to medical therapy." In a recent study of 615 consecutive cases of Crohn's disease seen at the Cleveland Clinic, ~ we proposed the thesis that initial anatomic involvement in Crohn's disease was a major determinant in prognosis. Analysis of these cases indicated what we referred to as a "clinical pattern" in that identifiable patterns of the disease process could be observed which were dependent on initial location of disease. This led us to analyze the reasons why patients underwent operation in that series and to expand this work to include additional years. Thus, we have attempted to define the surgical indication for patients for whom the decision for operation was made and for whom the definitive (at the time) operation was also performed at the Cleveland Clinic. The clinical, surgical, and pathological data were present for review by the same group of investigators. Materials and Methods Description of Series As previously reported,' 615 consecutive cases of Crohn's disease seen at the Cleveland Clinic during the years 1966 to 1969 were analyzed. Patients were registered into the series when sufficient data were available for the diagnosis of Crohn's disease, and at the time we made the original diagnosis or verified the diagnosis. Of the 615 patients, 380 underwent operation. Of the 380 patients, 316 had the definitive operation

2 246 FARMER ET AL. Vol. 71, No.2 performed at the Cleveland Clinic. Thus, data were available concerning the decision-making process, and an indication for surgery could be defined from analysis of the records of these patients. Further, the decision regarding need for operation was made by one or two of a small group of physicians and surgeons with a special interest in inflammatory bowel disease. The prevailing philosophy regarding need for operation for patients with inflammatory bowel disease has been the development of a complication of the disease process rather than simply presence of disease itself. Because these patients were studied as part of a larger and long term study on natural history and prognosis of Crohn's disease, it seemed desirable to have a more recent group of patients for comparative purposes and to continue analysis with regard to indication for surgery. The years 1972 and 1973 were chosen, because of availability of data and in order to utilize more recent experience with operation for inflammatory bowel disease. All patients who underwent operation for Crohn's disease during these 2 years were evaluated. There were 184 patients for whom the decision for operation was made at the Cleveland Clinic and the operation was then performed. Therefore, in the 6-year period of study (1966 to 1969 and 1972 to 1973) there were 500 patients for whom primary surgical indications could be determined. In no case was simple incision and drainage of an abscess considered as a "primary operation." All surgical procedures were intended to be definitive. Definition of Clinical Pattern Continuing the concept of clinical patterns as previously described,' the cases were analyzed according to initial anatomic involvement. Thus, classification was based on location of disease at the time of diagnostic documentation, and not necessarily at the time of operation. There were 225 patients with ileocolic involvement (terminal ileum and right colon), 130 patients with small intestinal involvement only, 127 patients with colonic involvement only, and 18 patients for whom the initial involvement was anorectal without proximal involvement. All patients with anorectal involvement subsequently had colonic involvement; interestingly, none had small intestinal disease. Thus, in this group of 500 patients for whom operation was performed for Crohn's disease, 45% had ileocolic involvement, 26% had small intestinal involvement, 25% had colonic involvement, and 4% had anorectal involvement. Statistical Analysis The three large groups were then treated as separate entities statistically, and data analysis was made comparing the three groups. Tests of statistical significance were made by the use of x and t-tests and Kolmogorov-Smirnov two-sample statistics. Definition of Criteria Used in Assessment In determining the indication for surgery, the following symptom complex descriptive terms were used. Perianal disease. This included perianal and perineal fistulae and abscesses, rectovaginal and rectoscrotal fistulae. Significant anal canal disease was also included, but specifically excluded were minor anal fissures and -any rectal mucosal disease itself. Thus, perianal disease was recorded as a complication which could be specifically identified by physical examination. Internal fistula and abscess. These included fistulae which could be documented roentgenographically (ileocecal, ileovesical, other types of enteric fistulae), as well as enterocutaneous fistulae other than perianal. Further, documentation of an abscess was by history and physical findings (fever and definable abdominal mass) or by roentgenography (ileocecal, intraabdominal, pelvic). All abscesses were also conclusively established at the time of operation. Therefore, all internal fistulae and abscesses were defined by objective means. Intestinal obstruction. This was defined by history and physical findings (persistent and progressive abdominal cramping pain occurring in a consistent manner and over a period of time-not simply sporadic abdominal pain associated with abdominal distention-plus abdominal tenderness); roentgenographic documentation, including plain or kidney, ureter, and bladder film showing distention of proximal loops of intestine; and documentation by barium study of marked stenosis of a segment of the small or large intestine (string sign) in association with the symptoms described. In a few cases, colonoscopy documented stenotic areas, but this was not the primary modality for diagnosis. Again, in all cases these findings were confirmed at operation. Specific separation was made between intestinal obstruction and development of abscess, even though at times there was some overlap in initial clinical symptomatology. Toxic megacolon. This was defined as a 7-cm or greater dilation of the transverse colon on plain or kidney. uterer, and bladder film of the abdomen. In addition, the symptom complex used to identify this condition included a rapid worsening of the patient's clinical course (rapid development of fever, abdominal pain, and abdominal distention, and the physical findings of abdominal distention). No case was included as toxic megacolon without specific roentgenographic documentation ofthis and documentation at time of operation. Poor response to medical therapy. This category, obviously the most difficult to define, included patients whose clinical progress was not satisfactory, despite what was regarded as optimal medical therapy. Although a degree of subjectivity was present in this category, the patients were those for whom none of the specific complications necessarily constituted a surgical indication, but for whom significant impairment of activities had occurred because of the disease. Many of these patients were malnourished (loss of 10% from ideal body weight), many had intermittent abdominal cramping and pain, and almost all had significant diarrhea. Also included in this category were patients with significant systemic manifestations (arthritis, pyoderma, growth retardation), and those for whom the indication was failure to respond to medical therapy or "intractability. " Results Presentation of Clinical Data and Definition of Surgical Indication Ileocolic Pattern. Of the 225 patients, 120 were males and 125 females. The mean age at time of operation was 29 years, and 59 patients were under age 20. The mean duration of symptoms was 4.7 years before surgery. Surgical indications are listed in table 1. In general, the indications varied relatively little whether the operation was performed during 1966 to 1969 or during 1972 to As shown in table I, 91 % of the surgical indications for patients with ileocolic Crohn's disease were for internal fistula and abscess (44%), intestinal obstruction (35%), and perianal disease (12%). Thus, the vast majority of patients with ileocolic Crohn's disease required operation for a specific rather than a nonspecific reason. Among those with poor response to medical therapy was one with growth retardation, and one with a

3 August 1976 SURGERY IN CROHN'S DISEASE 247 TABLE 1. Indication.. for surgery among patients with ileocolic pattern No. No. Indication ofpat.ents of patients Total % Perianal disease Internal fistula and ab scess Intestinal obstruction Toxic megacolon Poor response to medical therapy Total TABLE 2. Ileocolic pattern subgroups Predominant involvement Indication for surgery Ileum Colon Equal No. % No. % No. % Perianal disease Intestinal obstruction Internal fistula and abscess Megacolon Poor response to medical ther apy Total massive hemorrhage requiring operation. Among those having had some type of operation previously were 106 patients (75 in the 1966 to 1969 group and 31 in the 1972 to 1973 group)_ Ten of these patients had a previous appendectomy with subsequent development of enterocutaneous fistula. The data pertaining to patients in the ileocolic pattern were then analyzed as to the predominant initial involvement, and comparison was made between tho~e with predominantly ileal disease and secondary colomc involvement, and vice versa. For a large number ~f patients, the involvements appeared to be approximately equal. These were compared with the other subgroups. As can be seen in table 2, all 5 patien~s who subsequently developed toxic megacolon were In the subgroup with predominantly colonic involvement, as well as most of the patients for whom the indication for surgery was poor response to medical therapy. Internal fistula and abscess were somewhat less frequent among those with predominantly colonic involvement, and those with predominantly ileal involvement had less frequent perianal diseas~.. ' Small intestinal pattern. Of the 130 patients With small intestinal pattern, 70 were males and 59 females_ The mean age at the time of operation, 3L6 years, was similar for those operated on during the years 1966 to 1969 and those of 1972 to The mean duration of symptoms was 3.5 years. In this group, 99 patients ha? ileal involvement only, and 31 patients had more proximal areas of involvement. Six patients had duodenal involvement and the remainder had jejunal involvement. There were only 9 patients with no evidence of ileal involvement. Thirty-five of these patients had undergone previous operations. 1 ne indications for surgery at the Cleveland Clinic are listed in table 3. Two indications made up 87% of the surgical indications for patients with small intestinal Crohn's disease: intestinal obstruction, 55%, and internal fistula and abscess, 32%_ Differences from ileocolic pattern can be noted, but again, in the vast majority of cases, patients underwent operation for specific reasons. Of those patients with poor response to medical therapy, growth retardation was a primary indication in 1. Despite the duration of symptoms from diagnosis to operation, it is interesting to note that in virtually all cases the operation was performed for disease in the small intestine; only 7 patients had initial small intestinal involvement and subsequent colonic involvement requiring operation. Colonic pattern. Of the 127 patients with colonic pattern, 69 were males and 58 females. The mean age at the time of operation was 32 years, again, similar in the 1966 to 1969 and 1972 to 1973 groups. The mean duration of symptoms was 4.6 years before operation. Among these 127 patients who had operation, 93 had total colonic involvement and 34 had segmental colonic involvement. Thirty-one patients has previously undergone some type of operation at the Cleveland Clinic_ Table 4 lists the indication f6r surgery among patients with the colonic pattern_ As can be seen, a far greater number of patients in the colonic pattern had indication primarily for severe disease and poor response to medical therapy than did those in the ileocolic and small intestinal patterns. Among those in the poor response to medical therapy group was 1 patient whose primary indication was arthritis, 1 with pyoderma gangrenosum, and 1 with hemorrhage. Among the 15 patients with intestinal obstruction as an indication for surgery were 9 for whom the obstructed area was in the small intestine; all of these patients had had a previous operation. These were the only patients in this pattern with any small intestinal involvement. Six patients had stenotic areas in the colon, with symptoms of intestinal obstruction requiring operation. Anorectal pattern. The primary initial manifestation of Crohn's disease in 18 patients was anorectal, and operation was subsequently performed_ The disease was perianal in 12 and involved the rectal mucosa initially in 6; in the latter instances, the rectal mucosa was characteristic for Crahn's disease on sigmoidoscopic exam ina- TABLE 3. Indicatiom for surgery among patients with S1n4l/ intestinal pattern No. N o. lndication ofpatlenu of patients Total Perianal disease Internal fistula and ab Cess Intestinal obstruction Toxic megacolon Poor response to medical therapy Total %

4 248 FARMER ET AL. Vol.71,No.2 TABLE 4. Indications for surgery among patients with colonic pattern No. No. Indication ofpatiento of patients Total % Perianal disease Internal f.. tula and ab BCess Intestinal obstruction Toxic megacolon Poor response to medical tberapy Total tion. As previously indicated, colonic Crohn's disease subsequently developed in all of these patients; small intestine involvement developed in none. Indication for surgery in 12 of these 18 patients (12 males and 6 females with a mean age of 36 years) was severe perianal disease. Four additional patients underwent surgery because of indolent colonic Crohn's disease. An internal (sigmoid) fistula developed in 1 patient, and in 1 other, toxic megacolon developed. Comparison of the Three Major Patterns Table 5 lists the surgical indication by percentage among the three major patterns and the statistical significance of each surgical indication. Thus, it can be seen that there are significant statistical differences in indication for surgery based on anatomic location of Crohn's disease. Perianal disease was statistically signif icant as a surgical indication for patients in both ileocolic and colonic patterns. Intestinal obstruction was more frequent for those with small intestinal pattern than those with ileocolic pattern (P < 0.(03), but both were quite significant when compared to colonic pattern patients. For patients with ileocolic pattern, internal fistulae were more significant as surgical indications than they were with small intestinal pattern (P < 0.025), but both were significant when compared with colonic pattern patients. Both toxic megacolon and poor response to medical therapy were significant for colonic pattern patients. The relevance of the concept of the "clinical pattern" is emphasized by these data, which indicate that the operation was performed usually in the area of initial anatomic involvement. Thus, disease of the colon was unusual after initial small intestine involvement. For patients with initial colon disease, small intestine involvement occurred only after operation and usually as a preanastomotic or prestomal recurrence. Types of Surgery Performed As shown in table 6, the type of operation performed correlated directly with the anatomic location of disease. Resection with anastomosis was the predominant operation performed for patients with small intestinal involvement, and ileostomy with resection (usually subtotal colectomy) was the predominant operation for those with colonic disease. For patients with ileocolic disease, when the indication was intestinal obstruction, resection with anastomosis was often possible. When abscess was present, frequently an ileostomy was performed in addition to the resection. Ileostomy was usually required for patients with perianal disease. Bypass procedures were performed for some patients with abscess formation, but mainly for those patients in the 1966 to 1969 group. Bypass procedures were seldom employed in the 1972 to 1973 group of patients, and this was the primary difference in surgical technique between these two periods. Ileorectal anastomosis, a particular interest of ours,' was performed for certain patients with extensive colonic disease but with normal rectums and without perianal disease. Patients with toxic megacolon were treated by decompression colostomy and loop ileostomy, as advocated by Turnbull et a1. Carcinoma of the colon was found in 1 patient. This patient, with the ileocolic pattern and a 20-year history of Crohn's disease, had operation because of progressive intestinal obstruction. At operation, a carcinoma of the rectosigmoid was found in an area involved with Crohn's disease. This was the only patient of these 500 operated for Crohn's disease in whom a carcinoma was found in an involved area of intestine. Of these 500 patients, 37 died, 18 of whom were in the ileocolic group, 8 of whom were in the small intestinal group, and 11 of whom were in the colonic group. Thirty-two of the patients who died were in the group of the 316 patients operated on in 1966 to 1969; only 5 patients in the 1972 to 1973 group died. TABLE 5. Comparison o/surgical indications among the three major clinical patterns Indication Ileocolic Pattern Small int.. tinal Colonic Statistical significance no.(%) P Perianal disease 28 (12)" 6 (5) 24 (19)" <0.002 Intestinal obstruction 79 (35)" 72 (55)" 15 (12) < Internal fistula and ab 98 (44) 41 (32)" 30(23) < Beess Toxic megacolon 5(2) 0 25 (20)" < Poor response to medi 15(7) 11 (8) 33 (26)" < cal therapy Total Statistically significant. Operation TABLE 6. Types of operation perfomred Ileocolic Pattern Small intestinal Colonic no. Anor«tal neostomyonly neostomy and resection Resection only Bypass Ileorectal anastomosis Total

5 August 1976 SURGERY IN CROHN'S DISEASE 249 Discussion In recent years there has been considerabl~ attentio,n directed toward postsurgical recurrences m Crohn s disease.' ' Study has also been made regarding ~evel~pment of Crohn's disease after primary resectlon with anastomosis," 10 and a great deal of emphasis has bee.n placed on recurrences or lack of recurrences after colome resection However, there has been much less attention directed to the original surgical indication, except to note that most patients with Crohn's disease at some time or another during their clinical course do undergo surgery. In three recent studies of natu~al history and long and short term prognosis,... t~is pomt was again made. However, now, in the decade. smce t.he controversy regarding diagnosis and clinical differentiation of Crohn's disease and ulcerative colitis has been largely resolved, there still have not been large stud~es on why patients with Crohn's disease undergo surgery In the first place. The series reported from Leeds".. consi.s~s of 332 patients followed for a long period, but the ongmal surgical indication is not well defined. In th~ Scandanavian study of 186 patients,' and in two studies from United States,14... there were 186, 105, and 92 patients, respectively, and it was observed that between 66% and 90% of the patients underwent operation. Howev~r, no~e of these studies addressed the possible relatlonshlp between anatomic location of disease and surgical indication. Krause and co-workers' attempted to define the indications for surgery, but not with numerical or anatomical associations. They stated "the most common indications were (1) decline in general condition, with iron deficiency anemia, raised sedimentation r~te, low serum protein, and loss of weight; (2) severe dlarrhe~, often combined with nutritional disturbancjl; (3) s~blleus or ileus attacks; (4) abdominal or intraabdommal fistulae; (5) anal nstulae or abscesses which had not healed after simple surgical treatment; and (6) re~arde? physical or mental development in younger patients. They went on to state that "in most of the patients. there was a combination of these indications." Not all mvestigators have agreed that the site of disease mak~s any difference in the prognosis, and Fromm and his coworkers" stated that "there appeared to be no correlation between site and extent of regional enteritis at. the time of diagnosis and the incidence of major operations and mortality.". As stated previously, the thesis of this study IS that initial anatomic involvement in Crohn's disease stron~ly influences the clinical course, prognosis, and indicatlo~ for surgery. It is important to recognize that Crohn s disease is not a homogeneous entity and that the indication for surgery, as well as the type of surgery performed, will probably depend on t~e. nature an~ location of the disease process and the climcal ~attern: Although more patients in this st~dy ~ad IleocolIc involvement (45%) than either small mtestmal. (2?%) or colonic involvement (25%) alone, there were Significant statistical differences in the indications!or. surger:y depending on clinical pattern. For patients With Ileoc~hc Crohn's disease, more than 90% had an operation performed for three specific reasons: internal fistula and abscess, intestinal obstruction, and perianal disease. For those with Crohn's disease of the small intestine, 87% underwent operation for two specific indications: internal fistula and abscess and intestinal obstruction. Those with colonic Crohn's disease had a greater diversity of indications for surgery and included the only group in which "intractability" was a major indication (26%), followed by internal fistula and abscess (23%), toxic megacolon (20%), and perianal disease (19%). The development of toxic megacolon in colonic Crohn's disease is of particular significance, in view of the acuteness of the process and potential danger to the patient. Although the clinical patterns characteristic of small intestinal Crohn's disease and colonic disease are generally distinct from each other, there is some merging of clinical features in the ileocolic pattern. Nevertheless, most patients have a reasonably well defined clinical pattern, with surprisingly little overlap of clinical features after the initial anatomic location of disease. The indications for surgery which were recognized seldom were the result of a change in clinical pattern, but were related to the initial anatomic location of Crohn's disease. Although most patients with Crohn's disease eve.ntually require surgery, the indication for surgery can be anticipated by recognition of the concept of clinical patterns, and the type of surgery required can also be predicted in many instances. The highly significant problem of postsurgical recurrences remains, but it is believed that by a clear definition of the reason for the original surgery, study of this problem may be facilitated in the future. REFERENCES 1. Janowitz HD: Problems in Crohn's disease: evaluation of the results of surgical treatments. J Chronic Dis 28: G10tzer OJ. Silen W: Indications for surgical treatment in chronic ulcerative colitis and Crohn's disease of the colon. In Inflammatory Bowel Disease. Edited by JB Kirsner, R Shorter. Philadelphia. Lea and Febiger, p Goligher JC. dedombal IT. Burton I: Crohn's disease. with special reference to surgical management. Prog Surg 10: Krause U, Bergman L. Norlen BJ: Crohn's disease. A clinical study based on 186 patients. Scand J GastroenteroI6:97-1OS, Farmer RG. Hawk WA, Turnbull RB Jr: Clinical patterns in Crohn's disease: a statistical study of 615 cases. Gastroenterology 68:627~5, Lefton HB. Farmer RG, Fazio V: neorectal anastomosis for Crohn's disease of the colon. Gastroenterology 69: , Turnbull RB Jr, Hawk W A. Weakley FL: Surgical treatment of to~ic megacolon. Ileostomy and colostomy to prepare patients for colectomy. Am J Surg 122: , Nugent FW, Veidenlteimer Me. Meissner WA, et al: PrognOllis after colonic resection for Crohn's disease of the colon. Gastroenterology 65: , Lennard-Jones JE, Stalder GA: Prognosis after resection of chronic regional ileitis. Gut 8: dedombal IT, Burton I, Goligher JC: Recurrence of Crohn'. disease after primary excisionalsurgery. Gut 12: , Korelitz BI, Present DH, Alpert U, et al: Recurrent regional ileitis after ileostomy and colectomy for granulomatoll8 colitis. N Engl J

6 250 FARMER ET AL. Vol. 71, No.2 Med 287: , Korelitz BI, Janowitz HD: Controversy on recurrent ileitis after ileostomy: background and speculation. Gastroenterology 65: , Steinberg DM, Allan RN, Brooke BN, et al: Sequelae of colectomy and ileostomy: comparison between Crohn's colitis and ulcerative colitis. Gastroenterology 68:33-39, Willwerth B, DeCosse JJ, Dworken HJ, et al: Natural history of regional enterocolitis. Arch Surg 103: , Fromm H, Wilson FA, Rodgers JB Jr, et al: Granulomatous bowel (Crohn's) disease: a retrospective study of the course and treat ment. Arch Intern Med 128: , dedomhal IT, Burton II., Clamp SE, et al: Short-term course and prognosis of Crohn's disease. Gut 15: , 1974