Some Additional Data That Might Be Useful for Diastasis Recti Assessment.

Background: Diastasis recti (DR) is characterized by separation of both rectus muscles and protrusion of the median bulging, but besides median bulging DR can also entail global abdominal bulging. On other note, DR classification is based on the width of divarication, but measurement values are different at rest and at effort due to muscle contraction. Aim of the study is to provide additional features concerning the type of bulging and the width of divarication. Methods: Findings were retrospectively drawn from the data prospectively collected in the records of a continuous cohort of 105 patients (89 females, 16 males) referred for diastasis and concomitant ventral hernia repair. Results: There was a median bulging alone in 45 (42.9%) cases, a global bulging alone in 18 (17.1%) cases, both types combined in 37 (35.2%) cases and no bulging in 5 (4.8%). On 55 patients with a global bulging, 51 were females. Tape measurements values of DR width were closer to the values measured on the CT scan at leg raise than at rest. The differences were significant at rest as well as at leg raise. Though the difference at rest was highly significant (p = 0.000), the difference at effort was not far from being not significant (p = 0.049). Conclusion: Besides median bulging, presence or absence of the global bulging should be included in DR assessment. The difference between width of divarication at rest and on exertion raises the question of which value should be used for DR classification. The question is worth being debated.

[Who are the recipients of labile blood products? A multicenter nation-wide study--a "donation day." Blood banks, health facilities]. / Qui sont les receveurs de produits sanguins labiles (PSL)? Une étude nationale multicentrique - un jour donné. Etablissement de transfusion sanguine (ETS) - établissements de santé (ES).

During the years 1994-2001, a progressive decrease of the number of blood units transfused has been reported in France. In contrast, since 2002, there is an increasing number of blood units issuing (+7.6% between 2001 and 2006) and this must be investigated. On behalf of the French Society of Blood Transfusion, the "Recipients" working group promoted a nation wide survey with the support of the regional blood transfusion centres. This survey was aimed at describing the profiles of the transfused patients: socio-demographical patterns, and reasons of the blood transfusion (main and associated diagnoses). A cross-sectional survey was designed. All the patients who received a blood unit during a specific day were considered as the population of the study. They were identified by the regional transfusion centres by means of the "individual issuing form". Survey forms were fully filled for 90% of the patients. It has been considered as a good answer rate. Seven thousand four hundred and twenty-two blood units, delivered to 3450 patients were analyzed. Three groups of pathologies were found as a reason of transfusion: haematology-oncology (52.70% of the prescriptions) with 892 patients (27.8%) for haematological malignancies; surgical procedures (23.99%); intensive care and medicine procedures (21.92%). More than 50% of the recipients are 70 years old and more. This result is explained by the age distribution of inpatients. In a context of lack of donors and consequently difficulties to provide patients with optimal number of blood units, this study is helpful. Variability of blood unit issuings must be detected, analyzed and monitored in real time by the actors of the transfusion process, using computerized dashboards: the blood units provider (in order to adjust the strategy of blood units provision) and the health care establishment as well as care blood components prescribers (reasons of blood transfusion and evaluation of practices).

Transinguinal preperitoneal repair with the Polysoft patch: prospective evaluation of recurrence and chronic pain.

OBJECTIVE: The Polysoft patch was conceived to facilitate performance of the transinguinal preperitoneal patch method and combine the advantages of patch placement in the preperitoneal space and the open inguinal approach. The aim of this prospective study was to evaluate the rate of recurrence and chronic pain at midterm follow-up. METHODS: In a cohort of 200 hernia repairs involved in a prospective evaluation, midterm results of 171 cases operated on from 1 July 2004 to 31 December 2005 were assessed. The anesthesia was spinal in 136 (79.5%) cases, local in 26 (15.2%), and general in nine (5.3%). A questionnaire was sent to patients on 30 November 2006 asking about recurrence, chronic groin pain, and satisfaction. RESULTS: With a median follow-up of 21.9 months (11.6-29.4), 167 (97.7%) patients were evaluated, two were dead, and two were lost to follow-up. There were two (1.2%) recurrences that were reoperated on; both consisted of an indirect sac that protruded between the branches of the split patch. Eleven (6.6%) patients alleged the feeling of a foreign body, and 12 (7.2%) reported pain that occurred occasionally or upon effort but did not prevent activity. In one case, the pain present before operation was unchanged, and in three cases, the pain could clearly be attributed to an origin other than the hernia repair. No case of pain that impaired activity was observed. With regard to results, 98.2% of patients were satisfied and 97.6% declared that they would adopt the same method in case they had to be operated on for another hernia. CONCLUSION: These results suggest that the technique provides a low rate of recurrence and a low percentage of chronic pain that did not impair activity.

[Supplying surgery blocs in red cell concentrates: a collective approach to improve practices]. / Approvisionnement des blocs opératoires en concentrés érythrocytaires: une démarche collective d'amélioration des pratiques.

The need to adapt red blood cells concentrates management in surgery blocs and resuscitation to the changes of the legal framework has lead to a collective approach to improve practices. Gathered by the regional hemovigilance coordinators of the Drass Ile-de-France (regional office of health and social actions), representatives of doctors' ordering transfusions and hemovigilance correspondents of the Assistance publique-Hôpitaux de Paris and representatives of the EFS (French blood establishment) Ile-de-France, together with representatives of the Afssaps (French health products safety agency), have coordinated an assessment of local transfusion practices in surgery blocs and resuscitation that have to be compliant. Each hospital then offered local improvement actions, approved by regional and national instances. We present this original and collective approach of assessing practices leading to offers that both respond to a legal framework and improve blood products flows without damaging transfusion security.

The Polysoft patch: prospective evaluation of feasibility, postoperative pain and recovery.

INTRODUCTION: The Polysoft patch was conceived to associate the advantages of placement of the patch in the preperitoneal space (PPS) and easiness of the inguinal incision. The aim of this study was to evaluate prospectively the feasibility and postoperative outcome of this method. METHODS: Two hundred Polysoft repairs were performed in 191 consecutive patients with Nyhus types III and IV hernias. The anesthesia was spinal in 146 patients (76.4%), local in 36 (18.8%) and general in 9 (4.7%). The patch was placed in the PPS through the hernial orifice in the fascia in direct hernias and through the internal orifice in indirect ones. The technical points and postoperative course data were prospectively recorded and postoperative pain was assessed daily by visual analogue scale (VAS) in 25 patients. RESULTS: The method was achieved in all the cases. The type of the hernias was as follows: 59 direct, 129 indirect (including 28 pantaloon and 16 sliding), 3 inguino-femoral and 9 recurrent. The size of the patch used was medium in 152 cases (76%) and large in 48 (24%). The large patch was used more in pantaloon, sliding and inguino-femoral hernias. In direct hernias the patch was not split; in indirect hernias the patch was split (so the wings recreate a new internal orifice around the spermatic cord) in 79 cases (61.2%) and not split (with the cord parietalized) in 50 cases (38.8%). The length of operation, postoperative hospital stay, return to daily activity, to work and analgesics consumption were [median +/- SD (extremes)]: 35 min +/- 9.1 (20-60), 1 day +/- 0.4 (0-5), 3 days +/- 1.8 (0-8), 15 days +/- 9.7 (1-30) and 3 days +/- 2.3 (0-10), respectively. The median number of analgesics units was 8 +/- 5.9 (0-32). The values of pain assessed daily by VAS (on 100) varied from 20.4 +/- 19.7 (0-60) at day 0, 25.0 +/- 24.5 (0-80) at day 1 to 7.5 +/- 13.7 (0-50) at day 7; the mean value for the week was 16.7 +/- 16.7 (0-57). There were 14 (7%) benign postoperative complications. CONCLUSION: This study permitted the definition of some technical points and showed that the Polysoft patch can be used for all types of hernias with a weak posterior wall, including complex cases (big scrotal, pantaloon, sliding and recurrent), with a low risk of postoperative complications, a low level of postoperative pain and a short recovery time.

[Inguinal hernia. What techniques are available for the surgeon? Theoretical and practical advantages and disadvantages]. / Hernie inguinale. De quelles techniques dispose le chirurgien? Avantages et inconvénients théoriques et pratiques.

There are many hernia repair techniques. Among the most popular, the Shouldice Hospital method was the most widely used in the 1980s. Since then, methods employing prosthetic mesh have taken over, mainly because they are tension-free and therefore seem to lead to less recurrence and cause less postoperative pain. There are several ways of placing the prosthesis, which differ according to the approach used and the superficial or deep situation of the prosthesis. The Lichtenstein technique consists of placing and fixing the prosthesis on the posterior wall of the inguinal canal through an inguinal incision: it is currently the most widely used procedure because of its simplicity. The concept of placing the prosthesis in the subperitoneal space was developed by the French school (Rives, Stoppa). Initially the operation was performed through an inguinal (Rives) or midline (Stoppa) incision. These techniques are no longer adapted to the minimally invasive concept of surgery that prevails today and therefore are less used. The laparoscopic approach places the prosthesis through the trocars instead of through a large incision. The transabdominal approach opens the preperitoneal space through the abdominal cavity, breaching the peritoneum. The extraperitoneal approach consists of direct penetration of the preperitoneal space without violating the peritoneum, as in the Stoppa technique. This is undoubtedly the most elegant laparoscopic technique, but more difficult to perform. Other, simpler methods can also reinforce the posterior wall with a preperitoneal prosthesis. The Polysoft prosthesis, placed through an inguinal incision and inguinal ring, is a modern substitute for the Rives technique, but can be performed under local or locoregional anesthesia. Other types of prostheses, consisting of two, both superficial and deep, components are available: the Prolene Hernia System, for instance, is composed of two circular plates, linked together by a central stalk; there are also several types of plugs. As for anesthesia, there are several evidence-based arguments that indicate that local anesthesia is best.

Inguinal hernia: preperitoneal placement of a memory-ring patch by anterior approach. Preliminary experience.

The aim of this prospective study was to set up and evaluate a technique allowing, by the mean of a memory ring, easy placement of the patch in the preperitoneal space (PPS), directly via the hernia orifice, so as to associate the advantages of the preperitoneal patch, anterior approach and minimally invasive surgery. The memory-ring patch was made by basting a PDS cord around a 14 x 7.5 cm oval shaped polypropylene mesh. The hernia sac was dissected, blunt dissection of the PPS was carried out through the hernia orifice and the patch was introduced in the PPS via the orifice. Spreading of the patch in the PPS was facilitated by the memory-ring. One hundred and twenty nine hernias, classified as Nyhus Type IIIa, IIIb and IV, were operated on 126 patients; 11 were big pantaloon or sliding hernias. The anesthesia was spinal in 116 cases and local in 10 cases. There were three benign postoperative complications (2.3%) related to the hernia repair. Ninety six percent of the patients were evaluated with a mean follow up of 24.5 months (12-42). Two recurrences (1.6%) occurred, 7 patients (5.6%) felt some degree of light pain, but not any case of disabling pain was observed. This technique offers many advantages. It is tension-free and almost sutureless. The patch is placed in the PPS through the hernia orifice without any remote opening in the abdominal wall. The patch applied directly to the deep surface of the fascia reinforces the weak inguinal area by restoring the normal anatomic disposition. The good preliminary results are encouraging and justify further randomized evaluation.

[Subperitoneal inguinal hernioplasty by anterior approach, using a memory-ring patch. Preliminary results]. / Hernioplastie inguinale sous-péritonéale par voie antérieure, à l'aide d'une prothèse à mémoire de forme. Résultats préliminaires.

OBJECTIVE: Incidence of chronic pain is lower following laparoscopic hernioplasty than open surgery, probably due to the location of the patch in the preperitoneal space. But since laparoscopy is more demanding, the rates of complications and recurrences are higher. The aim of this study was to evaluate the results of a procedure consisting of placement of a patch, endowed with some memory of shape, provided by a thin peripheral memory ring, in the preperitoneal space, by inguinal incision, through the hernia orifice, under spinal or local anesthesia. METHODS: The study was carried out in two stages. The first study consisted of evaluating the results of a prosthesis made of a polypropylene mesh endowed with a memory-ring made of a PDS cord (Ethicon SAS, 92787 Issy-les-Moulineaux). The second study evaluated the results of the Polysoft patch manufactured by Bard C degrees (Bard France, 78960 Voisins-le-Bretonneux), according to this concept. RESULTS: The first study involved 129 hernias operated on 126 patients of mean age 60 years (27-84). There were 3 (2.3%) benign complications. With a median follow-up of 24.5 months (12-42), 124 hernias (96%) were evaluated. There were 2 recurrences (1.6%) and 7 cases (5.6%) of chronic pain. The second series involved 150 hernias operated on 139 patients of mean age 60 years (21-94). Four (2.7%) benign complications occurred. The median length of surgery was 36 min (20-60), the median postoperative hospital stay was 1 day (0-5), the median time to return to normal activity was 3 days (0-8) and the median time off work was 18 days (1(30). The patients took paracetamol for 3 days (0-10) and the total units number was 8 (0-28). CONCLUSION: These results suggest that the method, easily reproducible, provides a low rate of complications and recurrences, as well as a low level of postoperative and chronic pain. They are worth being confirmed by a randomised comparison to the laparoscopic and Lichtenstein techniques.

[Pre-peritoneal continuous infusion of local anesthetics for pain relief after laparotomy. A preliminary report]. / Infiltration continue prépéritonéale d'anesthésiques locaux pour l'analgésie après laparotomie. Données préliminaires.

INTRODUCTION: Continuous administration of local anesthetic through a catheter placed in the scar of a laparotomy is a postoperative analgesic technique, which seems effective but remains little developed and poorly codified. METHODS: In this prospective evaluation, we present a series of 25 observations of adult patients scheduled for abdominal laparotomy, to which a multiperforate catheter was placed at the end of the intervention by the surgeon in pre-peritoneal position, allowing the continuous perfusion of ropivacaïne over the first 48 postoperative hours. Patients received intravenous paracetamol associated with ketoprophene or nefopam. Opiates were given as rescue analgesics, in case of failure in pain relief, defined on objective criteria measured on visual analogic scale (VAS). RESULTS: The feasibility of the technique was excellent, except in one case of catheter obstruction. Pain was adequately relieved, with a majority of patients having VAS scores lower than 3/10 cm with the VAS, as well as rest as during mobilization. Only 9 patients needed morphine rescue analgesics. There was no sign of clinical overdose nor parietal complication related to the technique. Blood dosages of ropivacaine, carried out among 5 patients having received 600 mg daily, showed serum concentrations below the thresholds of toxicity. CONCLUSIONS: These results reveal a good effectiveness of the method, with moderate pain intensity and a low analgesic consumption. The local and general tolerance was excellent.

[Reducing the hospital stay after colorectal resection]. / Réduire la durée d'hospitalisation après résection colorectale.

INTRODUCTION: The postoperative hospital stay after colorectal resection is about 15 days in France, when some authors have published a postoperative stay of 2 to 5 days. The aim of this work was to obtain a postoperative hospital stay less than 7 days. PATIENTS AND METHODS: Sixty-one patients who underwent a colorectal resection performed by laparotomy were included in the study: 16 right hemicolectomies, 9 left hemicolectomies, 15 sigmoidectomies and 21 anterior resections were performed. The operation was performed through a midline incision extended over the umbilicus in 13 cases, limited below the umbilicus in 22 cases and elective in 26 cases (right transverse in 16 and left iliac fossa in 10 cases). The protocol comprised epidural analgesia or wound infusion with ropivacaine, restricted intravenous fluids, early oral feeding and active mobilisation. RESULTS: The median and mean times of discharge were 6 and 7.3 days respectively; 36 patients (59%) were discharged on postoperative days 3 to 6, 8 patients (13%) on days 7 and 17 (28%) after day 7. A nasogastric tube was necessary in 2 cases (3.3%). Ten (16%) postoperative complications and 3 (5%) readmissions occurred. There were no deaths. CONCLUSION: Although the postoperative stay cannot be reduced in all the cases, a median hospital stay inferior to which is currently observed can easily be obtained by applying some simple and inexpensive means. This is advantageous for the patient, whose recovery is faster, and contributes to reduce the cost, which is of crucial importance today.

Preoperative radiochemotherapy in rectal cancer: long-term results of a phase II trial.

PURPOSE: To assess toxicity and long-term results of preoperative chemoradiotherapy in rectal cancer. METHODS AND MATERIALS: Between 1989 and 1997, as a phase II study, 66 patients with T3 M0, rectal cancer received preoperatively a 45 Gy dose pelvic radiotherapy (XRT) combined with two 5-day chemotherapy courses (CT) of 5-Fluorouracil (5-FU) and Leucovorin (LV) delivered the first and fifth week of XRT. For each CT course, LV:20 mg/m2/d1-d5,. While the 5-FU dose was variable from 450 to 350 mg/m2/d first course and 370 to 350 mg/m2/d second course. Surgery was planned 3 weeks later. RESULTS: XRT-CT was stopped in 1 patient due to progressive disease. CT was stopped in 1 patient due to toxicity. Grades 2 and 3 diarrhea were observed in 8 and 3 patients, respectively. One patient died from acute diarrhea due to deviation from recommendations; 60 patients went to surgery. Among the 58 patients operated on for cure, 5 had an R1-resection. After a 4.5-year median follow-up, the 5-year pelvic disease-free survival was 92% for the whole group and 96% in the R0-resection group. CONCLUSION: Preoperative combined XRT-5-FU-LV is feasible if optimal XRT and patients are carefully managed. The recommended 5-FU daily dose is 350 mg/m2 for both CT courses. This approach is currently tested in a large EORTC phase III trial.

Conventional external irradiation alone as adjuvant treatment in resectable pancreatic cancer: results of a prospective study.

Between 1/85 and 1/90, 14 consecutive patients were entered into a prospective study of conventional adjuvant post-operative external beam radiotherapy after complete resection for a pancreatic adenocarcinoma. The surgical procedure was a Whipple resection in nine patients, a distal pancreatectomy in four patients and a total pancreatectomy in one patient. There were three T1b, eight T2 and three T3 tumours (UICC 1987); nodal involvement was present in five cases. The radiotherapy was delivered using a four-field box technique with a 23 x MV photon beam. All patients received a total dose of 54 Gy to the tumour bed. The mean treated volume was 900 cm3. Acute toxicities consisted mainly of weight loss (mean: 2 kg). Two patients had a grade 2 diarrhoea and two patients a grade 2 gastritis. Late effects were minimal and only observed in two patients. The overall locoregional recurrence (LR) rate was 50%. The median disease-free survival was 12 months, and the median survival was 23 months. This post-operative conventional radiotherapy treatment gives results that are comparable to the results of the GITSG-adjuvant study using a combination of split-course radiotherapy and 5-fluorouracil (5-FU).

Is prophylactic pelvic drainage useful after elective rectal or anal anastomosis? A multicenter controlled randomized trial. French Association for Surgical Research.

OBJECTIVE: We investigated the role of drainage in the prevention of complications after elective rectal or anal anastomosis in the pelvis. Anastomotic leakage after colorectal resection is more prevalent when the anastomosis is in the distal or infraperitoneal pelvis than in the abdomen. The benefit of pelvic drains versus their potential harm has been questioned. Drain-related complications include (1) those possibly benefiting from drainage (leakage, intra-abdominal infection, bleeding) and (2) those possibly caused by drainage (wound infection or hernia, intestinal obstruction, fistula). METHODS: Between September 1990 and June 1995, 494 patients (249 men and 245 women), mean age 66 +/- 15 (range 15 to 101) years, with either carcinoma, benign tumor, colonic Crohn's disease, diverticular disease of the sigmoid colon, or another disorder located anywhere from the right colon to the midrectum undergoing resection followed by rectal or anal anastomosis were randomized to undergo either drainage (n = 248) with 2 multiperforated 14F suction drains or no drainage (n = 246). The primary end point was the number of patients with one or more postoperative drain-related complications. Secondary end points included severity of these complications as assessed by the rate of related repeat operations and associated deaths as well as extra-abdominally related morbidity and mortality. RESULTS: After withdrawal of 2 patients (1 in each group) both groups were comparable with regard to preoperative characteristics and intraoperative findings. The overall leakage rate was 6.3% with no significant difference between those with or without drainage. There were 18 deaths (3.6%), 8 (3.2%) in those with drainage and 10 (4%) in those without drainage. Five patients with anastomotic leakage died (1%), 3 of whom had drainage. There were 32 repeat operations (6.5%) for anastomotic leakage 11 in the group with drainage and 4 in the group with no drainage. The rate of these and the other intra-abdominal and extra-abdominal complications did not differ significantly between the 2 groups. CONCLUSION: Prophylactic drainage of the pelvic space does not improve outcome or influence the severity of complications.

Stapled coloanal anastomosis with reservoir procedure.

A closed-lumen procedure of coloanal anastomosis with a reservoir, performed exclusively with staplers to reduce spillage, is described. This stapling procedure provides three advantages: simplicity, reduction of spillage, and the preservation of sphincteric structures. In 33 procedures performed, 2 (6%) anastomotic leakages and 5 (15%) dilatable stenoses occurred.

Inguinal hernia: the size of the mesh.

The size of the mesh used for surgical repair of groin hernias differs significantly from one technique to another. Such differences are not unimportant, since implantation of a large amount of prosthetic material can induce some drawbacks and may perhaps be avoided in many cases. The weak inguinal area, where inguinal hernias are exteriorized, is smaller than the myopectineal orifice, and a mesh 8-9 cm long and 5-6 cm wide is sufficient to cover this area. The results of methods using a small mesh are good, and large patches provide at best the same results with some real drawbacks. A large mesh spread in the preperitoneal space on the bladder and iliac vessels can induce some difficulties in case of further operation on these organs, and when we operate on a man 30 to 50 years old, we ignore whether he may need that kind of operation many years later. Consequently, in most cases a small patch covering only the weak inguinal area is preferable and a large preperitoneal patch should be used only in selected cases at high risk of recurrence.

Inguinal hernia: a patch covering only the myopectineal orifice is effective.

A large mesh in the preperitoneal space, spread on iliac vessels and bladder, can cause future difficulties in a case of urogenital or vascular disease. Since all recurrences are located in the area of the myopectineal orifice, a mesh covering only this area, as in the Rives procedure, is effective and avoids these drawbacks. However, the Rives procedure is demanding, which is why we tried to simplify the technique. A mesh 8-10 cm long and 6-7 cm wide was spread in the preperitoneal space using the inguinal approach, without any fixation or with only a few stitches; the fascia was then simply approximated over the mesh without tension. A total of 161 hernias (99% Nyhus type III and IV) were repaired using this technique. Only six benign complications (3.7%) occurred. With a median follow-up of 63 months (range, 18-94 months), only one recurrence occurred (0.7%), due to misplacement of the mesh, eight patients complained of mild pain, and no late complications were observed. In conclusion, placement of a small mesh covering only the myopectineal orifice using the inguinal approach is effective and is associated with a low risk of drawbacks and complications.

Cholecystectomy by minilaparotomy without muscle section: a short-stay procedure.

The aim of this prospective study was to assess the results of cholecystectomy performed by an incision characterized by its small size and absence of muscle section, to minimize postoperative discomfort, and shorten hospital stay. Surgical access was via a transverse division of the linea alba extending 1.5-2 cm on the rectus sheath, on both sides, without section of muscle. One hundred and ninety-one patients were operated on. Only 1 (0.5%) peroperative complication (injury of cystic artery) and 4 (2%) benign postoperative complications occurred. The mean postoperative hospital stay was 4.4 +/- 1.7 days (range 2 to 12); for the 90 patients operated on in 1989-1990 it was 3.3 +/- 1.7 days (range 2 to 12). Long-term results were assessed in 113 patients. Only 1 (0.6%) complication occurred, namely migration of a retained common duct stone. Neither incisional hernias nor a need for reoperation occurred. The cosmetic results were assessed by the patients themselves as very good in 95 (84.1%) cases and good in 16 (14.2%) cases.

[Value of blood assay of tumor markers for detecting recurrence of colorectal cancer]. / Intérêt du dosage sanguin des marqueurs tumoraux dans le dépistage des récidives du cancer colorectal.

Clinicians have at least 12 serial tumoral markers for the follow-up of colorectal cancer patients. The literature analysis indicates that serial carcinoembryonic antigen determination (CEA), is the most effective. The determination of other markers together with CEA does not add any benefit. CEA determination seems to be associated with a high rate of recurrence resection. However, the influence on overall survival of a planned follow-up including CEA determination is missing.

[Recognition of predictive prognostic factors in recurrent colonic cancer. Apropos of a retrospective series of 178 cases]. / Reconnaissance des facteurs pronostiques prédictifs de récidive du cancer colique. A propos d'une série rétrospective de 178 cas.

Prognostic factors influencing the outcome of colon cancer were studied retrospectively in 178 patients who underwent curative resection between 1975 and 1985. Monovariate and multivariate analysis have been performed. Median follow-up was 8.7 years. Three factors increased the risk of recurrences: age when equal or greater than 70, poorly differentiated tumor and advanced Duke's stage. Two factors increased the risk of intra-abdominal recurrence: age when equal or greater than 70 and advanced Duke's stage. Two factors increased the risk of metastasis (including the liver): poorly differentiated tumor and advanced Duke's stage. Multivariate analysis showed that the risk of recurrences was correlated to Duke's stage, age and histological differentiation. Patterns of failure and prognostic factors reported in the literature are reviewed and implications of these prognostic factors on adjuvant treatment indications are discussed.