The value of specialization--is there an outcome difference in the management of fistulas complicating diverticulitis.

PURPOSE: The value of specialization has frequently been challenged by many health care institutions and providers. This review was conducted to determine whether there were any outcome differences in the management of fistulas complicating diverticulitis. METHODS: We conducted an historical cohort study using hospital charts of all cases of fistulas complicating diverticulitis that were operated on in four university-affiliated hospitals between 1975 and 1995. There were 122 patients, with 37 under the care of fully trained colorectal surgeons and 85 under the care of general surgeons. RESULTS: There were no significant differences in patient demographics, preoperative comorbidities, or the number of preoperative diagnostic investigations between the two groups. The colorectal surgeons performed more intraoperative ureteral stenting (Colorectal Surgery 55.5 percent vs. General Surgery 24.4 percent, P = 0.001). The general surgeons performed more initial diverting Hartmann's and colostomy procedures (Colorectal Surgery 5.4 percent vs. General Surgery 27 percent, P = 0.013). The patients in the General Surgery group had longer preoperative lengths of stay (median Colorectal Surgery 3 (range, 1-28) days vs. General Surgery 8 (range, 0-29) days; P < 0.001), longer postoperative lengths of stay (median Colorectal Surgery 11 (range, 5-40) days vs. General Surgery 14 (range, 2-80) days; P = 0.001), and longer total lengths of stay (median Colorectal Surgery 14 (range, 6-62) days vs. General Surgery 24 (range, 6-100) days; P < 0.001). The patients in the General Surgery group experienced a higher rate of wound infections (Colorectal Surgery 5.4 percent vs. General Surgery 12.9 percent), and a larger proportion of them experienced complications (Colorectal Surgery 27 percent vs. General Surgery 41.2 percent). CONCLUSIONS: We conclude that specialization in colon and rectal surgery contributed to an improved outcome, with a lower rate of diverting procedures, a shorter hospital stay, and a lower rate of complications.

The effect of intraoperative thoracic epidural anesthesia and postoperative analgesia on bowel function after colorectal surgery: a prospective, randomized trial.

PURPOSE: Colorectal surgery is associated with postoperative ileus, which contributes to delayed discharge. This study was designed to investigate the effect of thoracic epidural anesthesia and analgesia on gastrointestinal function after colorectal surgery under standardized controlled postoperative care. METHODS: Forty-two patients diagnosed with either colonic cancer, diverticulitis, polyps, or adenoma, and scheduled for elective colorectal surgery, were randomly assigned to either postoperative patient-controlled analgesia (PCA) with intravenous morphine (n = 21) or epidural analgesia with a mixture of bupivacaine and fentanyl (n = 21). Postoperative early oral feeding and assistance to mobilization were offered to all patients. Pain visual analog scale (1-100 mm), passage of flatus and bowel movements, length of hospital stay, and readiness for discharge were recorded. RESULTS: Pain visual analog scale (visual analog scale, 1-100 mm) at rest, on coughing, and daily on mobilization was significantly lower in the epidural group compared with the patient-controlled analgesia group. Median values for the visual analog scale group were 7 (95 percent confidence interval, 2-18) mm, 19 (95 percent confidence interval, 4-38) mm, and 10 (95 percent confidence interval, 5-33) mm, respectively, and, for the patient-controlled analgesia group, were 24 (95 percent confidence interval, 18-51) mm, 59 (95 percent confidence interval, 33-74) mm, and 40 (95 percent confidence interval, 29-79) mm, respectively (P < 0.01). Intake of protein and calories and time out of bed were similar in both groups. Mean time intervals +/- standard deviation from surgery to first flatus and first bowel movement occurred earlier in the epidural group, 1.9 +/- 0.6 days and 3.1 +/- 1.7 days, respectively, compared with patient-controlled analgesia, 3.6 +/- 1.5 days and 4.6 +/- 1.6 days, respectively (P < 0.01). Postoperative complications occurred in 33 percent of the patient-controlled analgesia group and 28 percent of the epidural group. There was no significant difference in length of hospital stay between the two groups with a mean of 7.3 +/- 3.7 days in the patient-controlled analgesia group and 8.5 +/- 4.2 days in the epidural group. Readiness for discharge was similar in both groups. CONCLUSION: Thoracic epidural analgesia has distinct advantages over patient-controlled analgesia morphine in providing superior quality of analgesia and shortening the duration of postoperative ileus. However, discharge home was not faster, indicating that other perioperative factors influence the length of hospital stay.

Postoperative hypocalcemic tetany caused by fleet phospho-soda preparation in a patient taking alendronate sodium: report of a case.

This case report describes a patient who was previously prescribed alendronate (Fosamax) and presented with postoperative hypophosphatemia and hypocalcemic tetany after bowel preparation with Fleet Phospho-Soda. This report suggests that patients taking bone metabolism regulators may not be able to respond appropriately to hypocalcemic stressors.

Do surgical residents rated as better teachers perform better on in-training examinations?

BACKGROUND: There have been no attempts to objectively compare resident teaching ability with resident knowledge level. METHODS: Resident teaching ability, as rated by medical students and junior surgical residents, was compared with resident knowledge level, estimated by in-training examination results, for 18 PGY5 and PGY4 surgical residents at McGill University (September 1996 to July 1997). RESULTS: There was a trend to suggest that greater teaching ability is associated with higher in-training examination scores; this did not achieve statistical significance. PGY4 residents were rated as better teachers than PGY5 residents. Resident self-evaluation revealed a high degree of interest in teaching; inadequate time was the principal deterrent to resident teaching; enjoyment and learning during teaching were found to be the most common incentives. CONCLUSIONS: Our results suggest an association between resident level of knowledge and teaching ability. The principal deterrent to teaching--inadequate time--must be addressed to effectively assist surgical resident teaching.

Fistulas complicating diverticulitis.

This study was undertaken to assess the appropriate management of patients with diverticulitis complicated by fistula formation. A retrospective chart review was conducted on patients with symptoms of a fistula who presented between 1975 to 1995. There were 42 patients (32 women, 76%; 10 men, 24%) who ranged in age from 46 to 89 years (mean 69.8 +/- 9.8). Six patients had multiple fistulas. The types of fistulas included colovesical (48%), colovaginal (44%), colocutaneous (4%), colotubal (2%), and coloenteric (2%). Operative procedures consisted of resection and primary anastomosis in 38 patients and a Hartmann's operation in one. Three patients were managed conservatively with antibiotics (two due to poor performance status, the third due to resolution of symptoms). There were no operative deaths. The postoperative course was uncomplicated in 69%, while 12 patients (31%) experienced 19 complications (40%). These consisted of urinary tract infection (9.5%), atelectasis (7.1%), prolonged ileus (4.8%), arrhythmias (4.8%) and renal failure, myocardial infarction, pseudomembranous colitis, peroneal nerve palsy, unexplained fever, pulmonary edema (2.4% each). There were no anastomotic leaks and no deaths. Hospital stay ranged from 6 to 31 days (mean 12.3 +/- 7.6). Fistulas due to diverticulitis were safely managed by resection and primary anastomosis without mortality and with acceptable morbidity in this series. Patients deemed to be poor operative risks can be managed with a course of nonoperative treatment.

Toxic megacolon complicating pseudomembranous enterocolitis.

PURPOSE: Toxic megacolon is a rare complication of pseudomembranous enterocolitis. We reviewed our recent experience with this complication. METHODS: The first five patients of the series were studied retrospectively, and six others were followed prospectively. RESULTS: Between June 1992 and May 1994, 11 patients (8 male, 3 female) developed toxic megacolon secondary to pseudomembranous enterocolitis. Mean age was 60.7 +/- 11.8 (range, 40-79) years. Presenting symptoms and signs included diarrhea, 100 percent; malaise, 91 percent; abdominal pain, 82 percent; abdominal distention, 82 percent; abdominal tenderness, 72 percent; anemia less than 12 gm, 72 percent; albumin less than 3 gm, 64 percent; tachycardia greater than 100, 55 percent; fever greater than 38.5 degrees celsius, 45 percent; shock or hypotension, 45 percent. Predisposing factors included antibiotics, 64 percent; immunosuppressants or chemotherapy, 36 percent; antidiarrheals, 27 percent; and barium enema in one patient. Five patients (45 percent) had more than one predisposing factor. X-rays showed transverse colon dilation and loss of haustrations in eight patients (72 percent), with a mean diameter of 9.9 +/- 3.4 cm. Flexible proctosigmoidoscopy showed pseudomembranes in all scoped patients, and toxin assay for Clostridium difficile was positive in all patients. One patient had emergency surgery. Ten patients were initially treated medically with nasogastric suction and intravenous resuscitation (90 percent) and antibiotics (100 percent), usually in the intensive care unit (80 percent). Four patients did not respond and underwent surgery; two others improved, then deteriorated, and also underwent surgery. Altogether, 7 of 11 patients (64 percent) underwent surgery. Three patients (27 percent) responded well to medical treatment. One patient was deemed too ill to undergo surgery and died. Mean delay to surgery was 3.0 +/- 1.3 days. No sealed or overt perforation was found at laparotomy. All patients who underwent surgery had a subtotal colectomy, with either a Hartmann's stump (71 percent) or a mucous fistula (29 percent). Eventually, five of seven patients who were operated on and two of four medically treated patients died (overall mortality, 64 percent). Only one patient underwent closure of ileostomy and anastomosis. CONCLUSION: Toxic megacolon complicating pseudomembranous enterocolitis is a serious problem that carries a high morbidity and mortality rate, regardless of treatment.

Progression toward malignancy of hamartomas in a patient with Peutz-Jeghers syndrome: case report and literature review.

Peutz-Jeghers syndrome, inherited in an autosomal dominant fashion, is characterized by hamartomatous polyps of the gastrointestinal tract and by mucocutaneous pigmentation. The frequency of gastrointestinal malignant disease in this syndrome is estimated to be 2% to 3%. The authors review reports associating Peutz-Jeghers syndrome with malignant disease and present a patient who had advanced jejunal adenocarcinoma in association with Peutz-Jeghers syndrome. It has not been determined with certainty whether the malignant lesions arise from hamartomas, from associated adenomatous polyps or from the normal mucosa. Histologic examination of the excised specimen from the patient reported in this paper showed areas typical of a hamartoma as well as areas of hyperplasia, adenoma with mild to severe dysplasia and carcinoma in situ all in the same polyp. These findings suggest that the hamartomatous polyps found in Peutz-Jeghers syndrome have the potential to undergo malignant transformation.

Antibiotics bioavailability in acute experimental pancreatitis.

Pancreatic sepsis in acute pancreatitis is the most lethal complication of the disease. This study was done to create a rational basis for the choice of antibiotics used in the treatment of severe acute pancreatitis. We postulated that, unless the antibiotics were present in therapeutic concentrations in the pancreatic tissue during pancreatitis, their use was of no value. Six mongrel dogs were used to test each antibiotic, each dog acting as its own control. The doses were based on the weight of the dogs: 15.0 milligrams per kilogram of clindamycin; 50.0 milligrams per kilogram of chloramphenicol; 10.0 milligrams per kilogram of metronidazole; 5.0 milligrams per kilogram of gentamicin; 12.5 milligrams per kilogram of cefazolin, and 50.0 milligrams per kilogram of ampicillin. Baseline serum and pancreatic tissue levels were obtained after intravenous injection of the antibiotics. Bile-trypsin hemorrhagic pancreatitis was induced one week later, and the serum and pancreatic tissue level antibiotics were measured again. The results showed significant differences in bioactive levels of antibiotics between blood and the pancreas. Ampicillin, gentamicin and cefazolin reached therapeutic blood levels, but did not achieve a parallel therapeutic level in the normal pancreatic tissue or during pancreatitis. Only three of the antibiotics tested, clindamycin, metronidazole and chloramphenicol, achieved therapeutic tissue penetrance in the normal and inflamed pancreas. After 1982, based on these results, clindamycin became our prophylactic antibiotic of choice in instances of acute severe pancreatitis. This resulted in the eradication of Bacteroides as a cause of pancreatic sepsis between 1980 and 1985. In 1993, our recommendation is to use a broad-spectrum gram-negative and gram-positive antibiotic with good penetration of the pancreatic tissue, such as cefotaxime or imipenem.

[Ferguson's method of closed hemorrhoidectomy]. / L'hémorroïdectomie fermée de Ferguson.

The technique of closed hemorrhoidectomy (Ferguson hemorrhoidectomy) is described in details. The surgical indications, advantages, pre-operative preparation, choice of anesthesia and technical details are addressed in sequence. The original references by Dr. Ferguson are included.

Persistence of the circadian variation and altered response to hepatectomy of hepatic ornithine decarboxylase activity with malignant tumor burden.

We measured the effect of MC-26 mouse colon cancers (of different sizes) on the circadian rhythm of hepatic ornithine decarboxylase (ODC) activity and hepatic ODC activity during the 24 hr after 60% hepatectomy. Tumor-free control mice showed a normal circadian rhythm of ODC activity with the highest levels at 1100 hr and the lowest levels at 2300 hr. The amplitude of the rhythm was diminished significantly in mice with a large tumor burden (3% of their body weight), and hepatic ODC activity was significantly less than in the tumor-free mice at every point during the 24 hr of the study. In mice with "early" tumors (0.3% of body weight), basal activity of ODC was normal and there was no reactive increase in activity following hepatectomy. In contrast, mice with "late" (3% of body weight) tumors had significantly lower basal ODC activities and the increase in ODC activity following hepatectomy was prolonged and exaggerated. We concluded that tumor burden is associated with abnormal ODC activity and that these differences are exaggerated after hepatectomy. Furthermore, although average ODC concentrations in tumor-bearing mice fell precipitously, the circadian rhythm in hepatic ODC persisted. This finding indicates early recognition by the host of tumor presence, which has a profound negative regulatory effect on hepatic ODC. Apparently, this effect does not impinge on circadian control mechanisms, indicating that these signals act independently.

Substance P increases in vitro lymphokine-activated-killer (LAK) cell cytotoxicity against fresh colorectal cancer cells.

Gut-associated lymphoid cells are modulated by several gut hormones. We postulated that lymphokine-associated-killer (LAK) cell cytotoxicity of lymphocytes isolated from the gut mucosa may be increased by substance P (SP). Intestinal lamina propria mononuclear cells (LPMC) and colonic cancer cells were isolated from operative specimens by successive mechanical and enzymatic dissociation methods. Effector LAK cells were induced by culturing LPMC with recombinant interleukin-2 at a concentration of 250 U/ml. Substance P (10(-5) M) was added to the culture medium. Targets consisted of fresh colon cancer cells, HT-29 (cultured human colon cancer cell line), and control cell lines. After 4 days of incubation, cytotoxicity was measured using a 4-h 51Cr release assay. LAK cells alone showed moderate cytotoxicity against HT-29 and none against fresh colon cancer cells. LAK cells generated in the presence of substance P showed moderate cytotoxicity against HT-29 and strong cytotoxicity against fresh colorectal cancer cells. The percentage of cytotoxicity +/- SEM at various effector to target ratios was [(*) denotes P < 0.05 compared with above]: [table: see text] We conclude that substance P significantly increases LAK cell cytotoxicity against fresh colon cancer cells, but not against cultured cells.

Establishment of a human gastrinoma in nude mice.

We report the first establishment and characterization of functioning gastrinoma from a human being transplanted into nude mice. Tissue was obtained at operation from a gastrinoma liver metastasis from a patient with the Zollinger-Ellison syndrome. The tumor was implanted subcutaneously in five athymic nude mice. Serum gastrin was measured by means of radioimmunoassay in specimens of mouse blood taken before and 5 minutes after intraperitoneal injection of secretin (100 micrograms/kg). In a second experiment serum gastrin was measured 30 minutes after injection of somatostatin analogue, SMS 201-995 (300 micrograms/kg). Studies were also done in 10 control mice. At passage, the fundus of each tumor-bearing mouse was weighed and examined microscopically. The gastrinoma (tumor line, PT) has been maintained for 34 months through four passages with a tumor doubling time of 37 to 45 days. The histology is similar to the original tumor. Immunocytochemistry showed that PT contained gastrin. In two mice metastasis developed 9 months after implantation. Gastrin levels in mice bearing PT have ranged from 216 to 12,000 pg/ml. Gastrin levels of control mice ranged from 0 to 63 pg/ml. Secretin increased gastrin levels in three of five mice tested and decreased gastrin levels in two mice. Repeat secretin tests showed identical results. SMS 201-995 decreased gastrin levels from basal values. Fundic weight of mice bearing PT (397 +/- 93 mg) was significantly greater than control fundic weight (180 +/- 26 mg). Gastrinomas growing in nude mice produce physiologically active gastrin as shown by elevated serum gastrin levels and by hyperplasia of the stomach. Two distinct subpopulations of gastrinoma cells respond differently to secretin. This model should provide important information on mechanisms of growth control and on gastrin release by gastrinomas in human beings.

Lymphokine-activated killer (LAK) cells from human intestinal mucosa: cytotoxic activity against tumor cell lines and modified self but not autologous and allogeneic colon cancer cells.

Patients with colorectal cancer respond poorly to in vivo immunotherapy with lymphokine-activated killer (LAK) cells generated from peripheral blood mononuclear cells (PBMC). We postulated that gut-derived immune cells could be a more relevant source of LAK cells directed against colorectal cancer. Intestinal lamina propria mononuclear cells (LPMC) and colonic adenocarcinoma cells were isolated from operative specimens by combination of mechanical and enzymatic dissociation methods. LAK cells were generated by culturing PBMC and LPMC with recombinant interleukin 2 (IL2), with and without OKT3 monoclonal antibody, in short- (4 days) and long-term (21 days) cultures. Other cultured tumor cells, normal intestinal fibroblasts, and hapten-modified autologous LPMC were used as control targets. Cytotoxicity was measured by a 4-hr 51Cr release assay. Short-term cultured LAK cells exhibited a strong to moderate degree of killing against normal intestinal fibroblasts, hapten-modified self cells, and four different tumor cell lines. Instead, fresh colon cancer cells were resistant to cytotoxicity, regardless of their degree of histologic differentiation and the autologous or allogeneic nature of the LAK cells. Long-term culture with IL2 remarkably increased LAK cell activity against all tumor targets, but not against colonic adenocarcinoma cells. The results of this study, showing that freshly isolated colon cancer cells are intrinsically resistant in vitro to highly activated cytotoxic effector cells, may explain the poor clinical results observed in human trials with in vivo administration of IL2 or LAK cells.

Colonoscopic diagnosis and treatment of arteriovenous malformations in chronic lower gastrointestinal bleeding. Clinical accuracy and efficacy.

The authors reviewed their experience with diagnosis and treatment of lower gastrointestinal bleeding secondary to colonic arteriovenous malformations (AVM). A diagnosis was established exclusively by endoscopy in 80 percent of the patients. Twenty-eight patients were treated by endoscopic coagulation; bleeding stopped after one or more treatments in 67.9 percent of these patients. There were no complications or mortality as a consequence of endoscopic treatment. Surgery controlled the recurrent bleeding in six of seven cases of failed endoscopic coagulation, and in 13 of 17 cases where surgery only was undertaken, for an overall success rate 79.2 percent. It is concluded that colonoscopy can accurately establish the diagnosis of colonic AVMs in chronic lower gastrointestinal bleeding. Endoscopic coagulation is a useful adjunct in the treatment of this condition, and is safe, effective, and leaves other options open.

Surgery for ulcerative colitis in the pediatric population. Indications, treatment, and follow-up.

To evaluate the outcome of surgery for ulcerative colitis in pediatric and adolescent patients, the experience at the Cleveland Clinic Foundation was reviewed retrospectively. Fifty-nine percent of the patients presented with acute toxic colitis and sepsis; 94 percent underwent staged procedures with creation of a temporary or permanent ileostomy. Sepsis was the most frequent postoperative complication (38 percent), and accounted for all three deaths (5 percent). Long-term disability was minimal, and 90 percent of the patients were fully active at the time of follow-up.

Response of hypertrophic heart myocardial glycogen to GIK and hypovolemic shock.

Glucose-insulin-potassium (GIK) given during myocardial ischemia or anoxemia results in improved myocardial function and augments energy reserves of myocardial glycogen (MG). Because many patients with heart disease also have myocardial hypertrophy, our purpose was to examine whether similar elevations in MG can occur in hypertrophic hearts with GIK administration and to study the effect of hypovolemic shock on those MG levels. Mongrel dogs (n = 5) with myocardial hypertrophy underwent serial myocardial biopsies of the left (LV) and right (RV) ventricles, and blood samples were followed by GIK infusion (14.5 ml/kg/hr) for 2 hr. after which the dogs were subjected to 2 hr of hypovolemic shock (mean arterial pressure = 40 mmHg). It was found that after GIK infusion MG was consistently elevated in both RV (.43 +/- .02 to .60 +/- .04 g%) and LV (.63 +/- .07 to .71 +/- .01 g%) and FFA declined (.20 +/- .05 to .05-.01 mEq/liter). The MG responded to hypovolemia by further significant elevations (RV 1.16 +/- .33; LV .82 +/- .17), as did FFA (.38 +/- .21). These results indicate that hypertrophic hearts can indeed respond to GIK infusion by increasing MG in both the RV and LV, as do normal hearts. These hearts then submitted to hypovolemic shock showed a further elevation of MG. The elevated insulin levels post-GIK resulted in suppression of FFA. Thus GIK administration may have a sparing effect on energy stores of the heart during hypovolemic shock, which could have clinical implications in the treatment of patients with hypertrophic myocardia.

Prophylactic use of antibiotics in pancreatic sepsis: a 25-year reappraisal.

The efficacy of antibiotics given prophylactically in cases of acute pancreatitis is controversial. The authors carried out a retrospective review of 528 cases of pancreatitis seen between 1955 and 1980 and noted a marked increase in the number of septic complications and in the mortality related to sepsis. In most cases, infection was due to organisms sensitive to the antibiotics used, suggesting that they are not effective in preventing sepsis. The authors suggest a new approach to prophylactic administration of antibiotics, based on bioactive tissue levels in pancreatitis.

The fat/fiber antagonism in experimental colon carcinogenesis.

Current epidemiologic and experimental data emphasize the importance of nutrition in colon carcinogenesis. The interactions between dietary fat and dietary fiber were studied in the dimethylhydrazine model for colon cancer. A strict pair-feeding system was developed, and rats were fed one of five semisynthetic diets containing various amounts and types of fat and fiber. The incidence and number of tumors per animal were increased in all fat-fed groups. Adding fiber to the diets provided either partial (polyunsaturated fat) or complete (saturated fat) protection against the deleterious effect of fat. The antagonism between fat and fiber and the clinical relevance of these data are discussed.