Search

Background Lower gastrointestinal (GI) bleeding frequently recurs after negative technetium 99m–labeled red blood cell (RBC) scintigraphy. Methods Between July 1, 1999 and July 31, 2005, 84 negative 99m Tc-labeled RBC scintigrams were obtained for acute lower GI bleeding. Medical records were abstracted for age, gender, prior history of lower GI hemorrhage, length of hospitalization, initial hematocrit (Hct) and Hct nadir, transfusion requirements, cause of bleeding, use of anticoagulants and/or antiplatelet medications, and rebleeding episodes. Results The overall rate of rebleeding was 27% (n = 23). There were no significant associations between any of the patient variables investigated and rebleeding. Conclusions Despite negative 99m Tc-labeled RBC scintigraphy, more than 25% of patients experience recurrent lower GI bleeding. Patient age, bleeding source, use of anticoagulant/antiplatelet medications, length of stay, admission Hct, Hct nadir, transfusion requirements, and gender are not predictive of the patients who will rebleed.

PURPOSE: The role of colonoscopy in the prevention of colorectal cancer has been accepted, not only by the medical community but by the federal government as well. This study sought to document the current role of colonoscopy in the practices of colorectal surgeons. METHODS: A survey was mailed to members of The American Society of Colon and Rectal Surgeons detailing the scope of colonoscopy in their practices. RESULTS: Surveys were mailed to 1,800 members of The American Society of Colon and Rectal Surgeons; responses were received from 778 (43.2 percent). The mean age was 48 ± 10 (range, 27-79) years; the mean number of years in practice was 14 ± 10 (range, 0.2-48). The majority of respondents (91 percent) were male. Responses were received from 47 U.S. states and 30 foreign countries. Seventy-four respondents (9.5 percent) reported not performing colonoscopy; the most common reason cited was "referring physicians' preference" (45 percent). Seven-hundred four respondents (90.5 percent) reported performing colonoscopy as part of their clinical practice and reported an average of 41 ± 41 colonoscopies in the last month (range, 0-635) and 457 ± 486 in the last year (range, 2-7,000). Colonoscopy accounted for 23 ± 16 percent of responding physicians' clinical time (range, 1-100 percent) and 27 ± 19 percent of total charges (range, 0-100 percent). Nearly all respondents (97 percent) anticipated maintaining or increasing their volume of colonoscopy in the coming year. Eighty-four percent of respondents reported receiving some or all of their training in colonoscopy during a colon and rectal surgery fellowship. More than one-half of respondents (55 percent) believed that there should be more of an emphasis on colonoscopy on the American Board of Colon and Rectal Surgery board examination, and 81 percent believed that the annual meeting of The American Society of Colon and Rectal Surgeons should include lectures and/or courses covering colonoscopy. CONCLUSIONS: Colonoscopy plays a major role in the practices of colorectal surgeons across the world, accounting for approximately one-quarter of clinical time and total charges. Based on the expectation that this trend will continue, The American Society of Colon and Rectal Surgeons needs to aggressively support its members not only in the technical aspects of colonoscopy but also in the practice management issues.

PURPOSE: The aim of this study was to determine the appropriate surveillance for patients with a history of adenomatous polyps whose last colonoscopic examination was normal. METHODS: This was a retrospective review of a database of 7,677 colonoscopies (1990 to 1996). In patients under colonoscopic surveillance, we reviewed cases of patients who had received three colonoscopies (an index (initial) colonoscopy positive for adenomas and 2 follow-up colonoscopies (interim and final)). The risk of adenomas and cancers at final follow-up colonoscopy was compared between patients having a normal interim colonoscopy and those with a positive interim colonoscopy. The risk at final colonoscopy was also stratified by time interval and the size and number of adenomas at the initial index colonoscopy. RESULTS: Two hundred four patients undergoing surveillance for adenomas met inclusion criteria. At index colonoscopy the median polyp size was 1 cm and median frequency was three polyps. At all follow-up colonoscopies, we detected 493 adenomas and one cancer (median follow-up, 55 months). At 36 months patients with a normal interim colonoscopy (n=91) had significantly fewer polyps than patients with a positive interim colonoscopy (n=113; 15vs. 40 percent;P=0.0001). By 40 months, adenomas were detected in more than 40 percent of patients in both groups. The risk after a normal interim colonoscopy was not affected by time interval or number or size of polyps. Adenomas found subsequent to a normal interim colonoscopy were dispersed throughout the colon in 28 patients and isolated to the rectosigmoid in 6 patients. CONCLUSIONS: In patients with a history of adenomas, a normal follow-up colonoscopy is associated with a statistically but not clinically significant reduction in the risk of subsequent colonic neoplasms. These patients require follow-up surveillance colonoscopy at a four-year to five-year interval.

PURPOSE: Colonoscopic surveillance is recommended for patients with adenomatous polyps. Significant cost savings would result from identification of subgroups of patients in whom less costly surveillance would suffice. This study was performed to determine the natural history of patients undergoing removal of isolated rectosigmoid adenomas and to establish whether flexible sigmoidoscopy might be adequate for follow-up. METHODS: A retrospective review of a database of 7,677 colonoscopies, from 1990 to 1996, identified patients who had a minimal follow-up of two years after removal of adenomatous polyps isolated to the rectosigmoid. Polyps detected on surveillance colonoscopy were categorized as distal (≤60 cm from anal verge), proximal (>60 cm from anal verge), and diffuse (proximal plus distal). The risk of polyp formation was determined by actuarial analysis using the Kaplan-Meier method. RESULTS: Sixty-two patients undergoing surveillance for adenomas met inclusion criteria. At the index colonoscopy, 124 isolated rectosigmoid polyps were identified. The median polyp size was 1 cm and median frequency was one polyp. The median follow-up time for the entire cohort (N = 62) was 53 months. At follow-up surveillance colonoscopy, 105 additional adenomas were discovered and removed in 40 patients. No malignant polyps were detected. The pattern of polyps detected were proximal (n=19), rectosigmoid (n=16), and diffuse (n=5). CONCLUSIONS: The majority (65 percent) of patients with isolated rectosigmoid polyps have additional polyps on long-term surveillance, and 60 percent of patients will have these polyps located proximal to the reach of a sigmoidoscope. Therefore, flexible sigmoidoscopy is not a safe alternative for surveillance of patients with isolated rectosigmoid polyps.

Continuous mucosal involvement from the rectum proximally is one of the hallmarks of ulcerative colitis. However, recent pathologic series report appendiceal ulcerative colitis in the presence of a histologically normal cecum, representing a "skip" lesion. The clinical significance of this finding has not been established. Eighty patients, 54 males and 26 females, average age 37.9 years (range 14 to 82 years) who underwent proctocolectomy for ulcerative colitis from January 1990 to September 1995 were examined to determine the rate of discontinuous appendiceal involvement. Excluded were 12 patients with prior appendectomy and 11 with fibrotic obliteration of the appendiceal lumen. Of the remaining 57 patients, seven (12.3%) had clear appendiceal involvement in the presence of a histologically normal cecum. These seven patients clinically were indistinguishable from the 50 patients without skip involvement of the appendix in terms of age at surgery, pretreatment medications, type of surgery, interval from diagnosis to definitive procedure, complications, functional results, and clinical course. Discontinuous appendiceal involvement was found in 12.3% of patients undergoing proctocolectomy for ulcerative colitis, and clinically these patients behave as those without this feature.

PURPOSE: This study was designed to evaluate the long-term outcome and survival of patients treated for malignant colonic polyps. METHODS: A retrospective review of 15,975 cases of colonoscopies with 8,685 endoscopic polypectomies performed between 1972 and 1990 was undertaken. In 65 patients, the polypectomy specimens contained invasive carcinoma. Six patients were excluded (follow-up

PURPOSE: This study was undertaken to evaluate the incidence, diagnostic methods, and treatment of hemorrhage occurring after colonoscopic polypectomy. METHODS: A retrospective chart review was conducted of 12,058 patients who underwent colonoscopy at an academic referral center between January 1989 and July 1993. Of these, 6,365 patients required polypectomies or biopsies. RESULTS: After these procedures, 13 patients (0.2 percent) developed lower gastrointestinal hemorrhage requiring hospitalization. All bleeding episodes occurred within 12 days of polypectomy or biopsy (mean=8 days). Twelve patients (92 percent) underwent technetium-tagged red blood cell scintigraphy, which localized bleeding in four patients (31 percent). In the eight patients with normal scintigrams, hemorrhage did not recur, and no further evaluation was performed. Five patients (38 percent) underwent arteriography. Arteriogram was positive in two of four patients with positive scintigrams, and bleeding was controlled with selective vasopressin infusion. The fifth patient had arteriography without prior diagnostic studies because of massive hemorrhage; the bleeding site was identified and controlled with selective vasopressin infusion. Three patients had lower gastrointestinal endoscopy, with endoscopic identification of bleeding site in two patients, and endoscopic electrocautery controlled the bleeding in one patient. In the 13 patients with hemorrhage, cessation of bleeding occurred with intestinal rest and hydration in nine patients (69 percent), selective vasopressin infusion in three patients (23 percent), and endoscopic electrocautery in one patient (8 percent). Eight patients (62 percent) required blood transfusion with a mean of 4.8 units (excluding one patient on warfarin sodium who required 14 units of blood). No patient required surgical intervention. CONCLUSIONS: Incidence of hemorrhage after colonoscopic polypectomy or biopsy is low, and in our series, hemorrhage resolved without the need for surgical intervention. Management includes initial stabilization followed by diagnostic evaluation. Technetium-tagged red blood cell nuclear scintigraphy identifies ongoing bleeding and identifies patients in whom additional invasive procedures (arteriography, lower gastrointestinal tract endoscopy) are warranted.

Previously we demonstrated consistency in perioperative steroid dosing among colon and rectal surgeons. To determine whether patterns have changed and if dosing schedules differ across surgical specialties, we evaluated multiple specialties. Questionnaires were mailed to members of the American Society of Colon and Rectal Surgeons (CRS) (n = 1523), American Society of Transplant Surgeons (TS) (n = 988), American Society of General Surgeons (GS) (n = 2750), and American Association of Endocrine Surgeons (ES) (n = 278). Surveys addressed demographic factors and factors in dosing, whether steroids are managed by surgeon alone or in collaboration with colleagues, and the most common taper regimens used. Four hundred fifty surveys were returned. Sixty-four respondents had retired or answered less than 50 per cent; 386 (211CRS, 116GS, 45TS, and 14ES) were available for analysis. The majority managed both perioperative (85.5%) and tapers (77%) themselves; TS and ES were significantly less likely to use other physicians (P0.001). The preoperative dose used most frequently was 100 mg hydrocortisone intravenously (76% CRS, 64% GS, 22% TS, and 93% ES). Most CRS (44.5%) and GS (24.1%) taper intravenous steroids over 3 days, whereas TS (33.3%) and ES (50%) return patients to prednisone within 1 to 2 days. Discharge steroid use was inconsistent with CRS (46.4%) tapering prednisone over greater than 21 days, GS (19%) over less than 21 days, and TS (20%) and ES (21.4%) taper over 21 days to preoperative prednisone doses (P0.001). In the absence of standard guidelines for perioperative corticosteroid administration, significant variations exist in the regimens used by surgeons in multiple specialties.

Seventy-three patients underwent total colectomy, rectal mucosectomy, creation of J or S ileal reservoir, and ileal pouch-anal anastomosis from 1982 to 1989. Mean follow-up was 38 months, with a minimum of 3 months in 15 patients being followed long-term at another institution. Forty-eight (66 %) patients had histologically proven ulcerative colitis and 25 (34%) patients had familial polyposis. Thirty-eight J reservoirs and 35 S reservoirs were constructed. There were no perioperative deaths. The failure rate (loss of pouch) was 3%. Thirty-six complications in 34 (47%) patients were reported, 14 (19%) patients required surgery. Bowel obstruction was the most common postoperative complication (16%), followed by pouchitis (15%), and cuff infection (5%). Seventy-eight percent of the complications were associated with the J pouch. Average stool frequency at 1 year was 4 per 24-hour period. Other complications included postoperative pneumonia (1), peroneal nerve palsy (1), and temporary sexual dysfunction (1). Seven of 15 complications requiring surgical intervention occurred in the first 2 years of the study period, illustrating the learning curve associated with the procedure. Blood loss, transfusion requirements, and length of operation were not associated with higher complication rates. Use of the J pouch and experience of the individual surgeon affected morbidity.

The records of 201 asymptomatic patients who underwent colonoscopy based solely on a family history of colon cancer were reviewed. Eighty-five patients (42 percent) had a total of 166 lesions. Fifty-four (27 percent) patients of the screened population had neoplastic lesions, while 31 (15 percent) patients had nonneoplastic polyps. Four carcinomas were found. Twenty-five of the patients with polyps (29 percent) had no polyps distal to the splenic flexure; these proximal polyps (and two carcinomas) would have been missed on screening with fiberoptic sigmoidoscopy. Nineteen of these 25 patients had polyps smaller than 0.5 cm, which likely would have been missed with contrast enemas. Almost one half (47 percent) of all polyps discovered at screening colonoscopy were proximal to the descending colon. Only one patient younger than 40 years old had adenomas. The yield of polyps and cancer in patients with familial risk indicates screening colonoscopy should be considered after age 40.

This retrospective study defines a population with neoplastic colonic polyps who have had colonoscopic polypectomy and, in follow-up within one year, a repeat colonoscopic evaluation. The population was broken down into two groups, one group that had polyps at the second examination and one group that did not. This study determined which factor(s) were significant among this population in distinguishing whether new polyps would be found at one year follow-up. The authors found that among the many variables studied, only polyp multiplicity was significant in predicting polyp recurrence. More than one polyp found at index colonoscopy led to a significant chance of having a new polyp after only one year. Also, it was demonstrated that these "new" polyps were unlikely to have been "missed" polyps from the initial colonoscopy. Because of the shifting location, smaller size, and fewer instances of histologic atypia in these polyps compared with those at index examination, the authors believe that polyps found after one year may be assumed to have arisen de novo. Finally, the authors show that a significant number of polyps occur beyond the reach of the flexible sigmoidoscope (approximately 60 cm). The authors recommend that patients who have polyps undergo a colonoscopic examination. When patients are re-evaluated after having colonoscopic neoplastic polypectomy, they should undergo repeat colonoscopy.

PURPOSE: To determine the physiologic alteration resulting in fecal seepage and soiling, results of anorectal manometric testing were evaluated in patients with varying degrees of fecal incontinence. METHODS: Anal manometric studies performed on 170 patients with fecal incontinence were reviewed. Results of their studies, including mean resting pressure, maximum resting pressure, maximum squeezing pressure, minimum rectal sensory volume, and minimum volume at which reflex relaxation first occurs, were compared with those of 35 control group subjects with normal fecal continence. Manometric studies were performed using a four-channel, water-perfused catheter. Incontinent patients were divided into three groups based on presenting complaints: complete incontinence (incontinence of gas and liquid and solid stool), partial incontinence (incontinence of gas and liquid), and seepage and soiling (incontinence of small amounts of liquid and solid stool without immediate awareness). RESULTS: Resting pressures were significantly lower in complete incontinence, partial incontinence, and seepage and soiling groups than in the controls (P

PURPOSE: The aim of this study was to assess perioperative warfarin management and complications in patients requiring colonoscopy. METHODS: We retrospectively reviewed 109 cases of colonoscopies performed on 94 patients requiring anticoagulation with warfarin. Patients stopped their warfarin three days before colonoscopy. Coagulation profiles obtained just before the colonoscopy showed a median prothrombin time of 13.4 seconds with a range of 11.1 to 29.1 (normal range, 10.9–13) and a median international normalized ratio of 1.2 with a range of 0.9 to 2.6. Patients restarted warfarin the day after the examination. RESULTS: During the 109 colonoscopies, 47 percent of the patients underwent either hot biopsy or snare polypectomy. One examination that included several biopsies was associated with a hemorrhagic complication (0.92 percent) requiring hospitalization and transfusion. Subset analysis of the therapeutic (biopsy and snare polypectomy) group indicated a slightly higher complication rate (1.96 percent) with a median international normalized ratio of 1.3 (range, 1–2.3) and a median prothrombin time of 13.7 (range, 11.6–25.9). CONCLUSION: Patients taking warfarin for anticoagulation may safely undergo colonoscopy. The risk of hemorrhagic complications increases slightly with hot biopsy or snare procedures. Further studies are needed to refine guidelines for colonoscopy in the patient requiring anticoagulation.

This study was designed to assess the medical and surgical treatment of colonoscopic perforations.A retrospective review of colonoscopic perforations from 1970 to 1999 was performed.In 30 years, 34,620 colonoscopies resulted in 31 (0.09 percent) perforations. Eighteen (58 percent) resulted from therapeutic colonoscopies, whereas 13 (42 percent) occurred after diagnostic colonoscopies. Sixteen perforations (52 percent) were identified during the procedure, 13 (42 percent) within 24 hours, and two (6 percent) within 48 hours. Twenty patients (65 percent) underwent surgical therapy, and 11 (35 percent) were treated medically with intestinal rest and intravenous antibiotics. In the medically treated group, one patient required rehospitalization for percutaneous drainage of an intra-abdominal abscess, and one patient died after requesting no further treatment because of an underlying terminal medical condition. Three patients failed medical treatment and required surgical intervention. One underwent repair with proximal diversion, whereas the remaining two received a colorrhaphy without resection or diversion. In the surgical treatment group, nine patients received colorrhaphy without diversion, seven underwent resection with primary anastomosis, and four had resection with diversion.Selected patients with colonoscopic perforation may be safely treated nonoperatively. Surgical treatment is reserved for patients with a large perforation or diffuse peritonitis.

PURPOSE: This study was performed to evaluate whether the time interval from injection of technetium Tc 99m (99mTc)-labeled red blood cells to the time of a radionuclide “blush” (positive scan) can be used to improve the efficacy in predicting a positive angiogram. METHOD: A retrospective review revealed 160 patients who received99mTc-labeled red blood cell scintigraphy for evaluation of massive lower gastrointestinal hemorrhage between 1989 and 1994. Patients were included who demonstrated signs of shock on admission, had an initial decrease in hematocrit of ≥6 percent, or required a minimum transfusion of two units of packed red blood cells. Scanning duration was 90 minutes, with imaging every 2 minutes. Time interval from injection to a positive scan was analyzed to determine predictability of a positive angiography. RESULTS: Of 160 patients, 86 demonstrated positive scans, of whom 47 underwent angiography. These 47 patients were divided into two groups according to scan results. Group 1 (n = 33) had immediate appearance of blush; Group 2 (n = 14) had blush after two minutes. In Group 1, 20 of 33 patients had a positive angiogram, yielding a positive predictive value of 60 percent (P = 0.033). Of the 14 patients with negative angiograms (13 from Group 1, and 1 with a negative scan), 6 had radiographic occlusion of the inferior mesenteric artery and 1 had spasm of the right colic artery, with scans that blushed in the respective distributions. Excluding these seven patients yielded a positive predictive value of 75 percent (P = 0.0072) for angiography. In patients with a delayed blush (Group 2), 13 of 14 had negative angiograms, yielding a negative predictive value of 93 percent (92 percent excluding those with nonvisualization of the inferior mesenteric artery). Twenty of 21 (95 percent) positive angiograms occurred in Group 1 patients. Of the 27 patients with negative angiograms, 13 were Group 2 patients. CONCLUSION: Patients with immediate blush on99mTc-labeled red blood cell scintigraphy required urgent angiography. Patients with delayed blush have low angiographic yields. These data suggest that patients with delayed blush or negative scans may be observed and evaluated with colonoscopy.

PURPOSE: This study was performed to determine costeffective colonoscopy guidelines for patients with prior colorectal adenocarcinoma. METHOD: A retrospective review was performed of patients who had been treated for colorectal adenocarcinoma and later underwent follow-up colonoscopy from 1984 to 1994. RESULTS: During this study period, 389 patients previously treated for colorectal adenocarcinoma underwent follow-up colonoscopy. All patients had perioperative colon evaluation for other neoplasms. Ages ranged from 26 to 89 (mean, 65.8) years, and 46.8 percent were female. Recurrent or metachronous cancer or a neoplastic polyp constituted a positive examination. Results of 389 first follow-up colonoscopies were compared with 259 second (66.6 percent), 165 third (42.4 percent), and 83 fourth (21.3 percent) follow-up examinations. Median interval between all colonoscopies was 13 months. Positive examination rates for the first two yearly examinations were 18.3 and 18.5 percent, respectively. Slightly lower, third-year and fourth-year positive examination rates were 16.4 and 14.5 percent, respectively. Fouryear examinations yielded the following: first year-1 carcinoid, 1 new adenocarcinoma, and 100 polyps; second year-1 anastomotic recurrence and 68 polyps; third year-55 polyps; and fourth year-1 recurrent cancer and 17 polyps. CONCLUSIONS: These data suggest that 1) annual follow-up colonoscopy for two years after colorectal cancer surgery is beneficial for detecting recurrent and metachronous neoplasms and 2) the interval between subsequent examinations may be increased depending on the result of the most recent examination.

Increasing experience with colonoscopy has altered recommendations for the frequency of follow-up surveillance examinations for adenomatous polyps and colorectal cancer. Current recommendations include a follow-up colonoscopy at 1 year for patients with more than two adenomatous or highly suggestive polyps and after curative surgery for colorectal cancer. Other patients can safely receive a follow-up colonoscopy at longer intervals of 3 years. Published data and a review of the Ochsner Clinic experience are presented to support these recommendations.

The significance of mucinous carcinoma has been controversial since first described by Parham in 1923. Previous reports have suggested that mucinous tumors affect young patients, involve the more proximal colon, are more advanced at diagnosis, and have a poorer prognosis than nonmucinous colon carcinoma. More recent reports have refuted these results. In an effort to clarify the significance of mucinous histology, a retrospective review of cases of invasive colon cancer treated at the Ochsner Clinic between 1982 and 1985 was undertaken. Mucinous adenocarcinoma, as defined by > or = 50 percent mucin, was found in 52 patients. During the same period, 343 nonmucinous adenocarcinomas were resected. The mean age, distribution within the colon, stage at diagnosis, and survival of mucinous carcinoma patients were compared with those with nonmucinous tumors. Mucinous tumors presented at a statistically significant more advanced stage (38 percent vs. 22 percent Dukes C lesions; P < 0.01). No significant differences were seen in age at presentation, distribution within the colon, or stage-for-stage survival when the entire group was analyzed. Mucinous carcinomas of the rectum occurred at an advanced stage more frequently (P < 0.05) than nonmucinous rectal carcinomas and had a markedly worse five-year survival (11 percent vs. 57 percent; P < 0.002).

Forty-eight cases of Ogilvie's syndrome, colonic pseudo-obstruction, presenting between 1983 and 1989 were retrospectively reviewed to assess the results of colonoscopic decompression and to identify potential etiologic factors. Three patients had spontaneous resolution with medical treatment. Forty-five patients required 60 colonoscopic decompressions: 38 (84 percent) were successfully treated using colonoscopy; five (11 percent) required an operation; and two died within 48 hours of colonoscopy from medical causes. No complications or deaths were the result of colonoscopy. Twenty-nine patients (64 percent) were successfully treated with a single colonoscopy. One-third of patients required serial decompressions. Average cecal diameter in patients with successful colonoscopic decompression was 12.4 cm but was larger for patients requiring more than one colonoscopy (13.3 cm) and for those who failed colonoscopic therapy (13.4 cm). The spine or retroperitoneum had been traumatized or manipulated in 52 percent of patients. Patients with Ogilvie's syndrome were being treated with narcotics (56 percent), H-2 blockers (52 percent), phenothiazines (42 percent), calcium-channel blockers (27 percent), steroids (23 percent), tricyclic antidepressants (15 percent), and epidural analgesics (6 percent) at diagnosis. Electrolyte abnormalities included hypocalcemia (63 percent), hyponatremia (38 percent), hypokalemia (29 percent), hypomagnesemia (21 percent), and hypophosphatemia (19 percent). Colonoscopic decompression in Ogilvie's syndrome is safe and effective management. Multiple pharmacologic and metabolic factors, as well as spinal and retroperitoneal trauma, appear to alter autonomic regulation of colonic function, resulting in colonic pseudo-obstruction.

A retrospective review of patients with Crohn's disease treated at our institution from 1973 to 1986 revealed 35 patients operated upon for anorectal fistulas. Twenty-nine had low intermuscular fistulas (multiple in seven), and six had high intermuscular (supralevator) fistulas. Fistulotomy alone was performed in 19 patients, and eight underwent partial fistulotomy and seton insertion. Five additional patients had proximal fecal diversion before fistulotomy. Three patients with severe colonic and anorectal disease underwent proctocolectomy as the initial procedure. Of the 32 patients who had fistulotomy performed, complete healing occurred in 30. Seven patients who healed required more than one operation for fistula. One patient was left with an asymptomatic fistula, and one required proctectomy for persistent symptomatic fistula and proctitis. Success of operation correlated with absence of rectal disease and quiescent disease elsewhere in the gastrointestinal tract. Aggressive medical treatment is required to control bowel disease preoperatively. In the majority of patients, subsequent surgery is justified and healing can be anticipated.

A retrospective review of patients with Crohn's disease treated at our institution from 1973 to 1986 revealed 12 patients operated on for rectovaginal fistula. Disease involved the large intestine in 10 patients. Primary fistula repair was performed in four patients and four others had staged repair with preliminary fecal diversion. Four patients with severe colonic and anorectal disease had proctocolectomy performed as the first procedure. Of eight patients who underwent fistula repair, complete healing occurred in six. One patient has a persistent fistula, which is minimally symptomatic, and the other required proctocolectomy after three unsuccessful repairs. Success of operation correlated with quiescent intestinal disease and absence of rectal involvement. In selected patients with symptomatic fistulas, surgical repair is indicated and healing can be anticipated.

Perforation of the rectum or sigmoid colon complicated 5 of 2200 barium-enema examinations performed during a 4-year period. Three patients with rectal perforations manifested by air extravasation were successfully treated with intravenous antibiotics and complete bowel rest. Two patients with barium extravasation were treated with immediate operation and colostomy. All five patients recovered. Perforation was found to be associated with a rectal stricture due to ulcerative colitis, a rectal cancer, an incarcerated inguinal hernia, fulminant ulcerative colitis, and a normal colon in an elderly patient. To determine the pressure in the rectum that could potentially be generated during a barium-enema examination, the pressures created by a standard barium delivery set were measured, using 1-meter columns of water, 25 percent diatrizoate sodium (Hypaque), 20 percent barium, and 80 percent barium. The columns generated pressures of 70, 85, 95, and 120 mm Hg respectively. Squeezing the delivery bag increased the pressure 21 to 79 percent or a maximum of 55 mm Hg. Colorectal perforation during barium-enema examination that was not accompanied by barium extravasation could be successfully treated nonoperatively. The associated pathology and our studies of pressures generated during a barium-enema examination allow us to suggest that the incidence of colorectal perforation during barium-enema radiography can be reduced by 1) performing proctoscopy prior to barium enema, 2) avoiding the use of the rectal balloon in patients with known rectal lesions, 3) avoiding barium studies in patients with active colitis, 4) avoiding generation of pressure greater than that created by a column of barium suspension of one meter, and 5) using a lower concentration of barium when possible.

A case of nonHodgkin's lymphoma of the perianal region in a patient with AIDS is reported. The unusual features of AIDS-related lymphoma and the possible role of immunodeficiency increasing susceptibility to oncogenic viruses are discussed.