Incorporation of skin preparation guidelines in local surgical facility protocols: what kind of barriers does it face? A multi-centre study in France.

BACKGROUND: Surgical site infection (SSI) is the most frequent and severe adverse event after surgery. Among preventive measures, the preoperative skin preparation (PSP) is known to be heterogeneously implemented in routine practice. A prerequisite would be the actual incorporation of guidelines in French surgical local protocols. AIM: To assess whether PSP recommendations have been incorporated in local protocols and to identify the reasons for the non-incorporation. METHODS: An online survey was proposed to all infection control teams (ICTs) in facilities participating in the French national surveillance and prevention of SSI network Spicmi. The reference recommendations were based on the French Society for Hospital Hygiene guidelines. FINDINGS: In all, 485 healthcare facilities completed the questionnaire. The incorporation of recommendations in the facility protocol varied between 30% and 98% according to the recommendation. The measures most frequently incorporated were antisepsis with an alcoholic product and cessation of systematic hair removal. The least frequently incorporated were the use of plain soap for preoperative shower and the non-compulsory skin cleaning in the operating room. Barriers reported were either specific to PSP (e.g. 'Concern about an increase of SSI', 'Scepticism about recommendations', 'Force of habit') or non-specific (e.g. 'The protocol not yet due to be updated'). CONCLUSION: We suggest that although some major prevention measures have been incorporated in the local protocol of most facilities, local protocols still frequently include some non-evidence based former recommendations. Communication about evolution of SSI rates, diffusion of guidelines by learned societies, and exchange with judiciary experts could make clear the conditions for applying recommendations.

Pandemic A/H1N1/2009 influenza in a paediatric haematology and oncology unit: successful management of a sudden outbreak.

Viral respiratory infections are potentially life-threatening among children treated for cancer. We report a nosocomial outbreak of six cases of pandemic influenza A/H1N1/2009 on a paediatric haematology and oncology ward. Three patients developed pneumonia and two of them sustained haemodynamic collapse. The source was probably a relative of the first infected patient. The outbreak was probably spread by cross-infection between patients during communal activities. A few days' delay in identifying the outbreak promoted spread of the influenza. Infection control measures included the use of oral oseltamivir treatment for all hospitalised patients, isolation of the infected patients, strict personal protective controls and a restriction on visitors. No new cases occurred after implementation of these containment measures. At the time when the outbreak was identified, all the patients were already isolated for other reasons. We conclude that A/H1N1/2009 influenza may spread rapidly and cause severe infection in paediatric cancer patients but can be efficiently contained. Identification of isolated or clustered cases should lead to the rapid implementation of appropriate infection control measures.

The changing pattern of Crohn's disease incidence in northern France: a continuing increase in the 10- to 19-year-old age bracket (1988-2007).

BACKGROUND: Crohn's disease incidence rates have stabilised in industrialised countries since the 1980s. Conversely, a continuing increase in childhood-onset Crohn's disease incidence has been reported. AIM: To confirm trends in inflammatory bowel disease (IBD) incidence in northern France over an extended time period (1988-2007) with a focus on childhood-onset Crohn's disease. METHODS: The IBD patients recorded in the EPIMAD registry between 1988 and 2007 were included. Standardised incidence rates were calculated for Crohn's disease and ulcerative colitis in the entire population, and separately according to age. Evolution of phenotypes at diagnosis was also studied. RESULTS: A total of 12 084 incident IBD cases (7428 Crohn's disease and 4656 ulcerative colitis) were recorded. Crohn's disease incidence rates increased from 5.2 cases/100 000 persons in 1988-1990 to 6.7 in 2006-2007 (+29%), stabilising after a peak at 7.1 in 1997-1999. Crohn's disease incidence rates in the 10-19-year age category increased by 71%, from 6.5 (1988-1990) to 11.1 (2006-2007). The frequency of initial ileo-colonic localisation increased from 52.9% in 1988-1990 to 68.6% in 2006-2007 (P<0.0001). Ulcerative colitis incidence rates decreased during the same period. CONCLUSIONS: From 1988 to 2007, Crohn's disease incidence increased by 29% in northern France and by 71% in the 10-19-year-old age group. Consequently, studies on Crohn's disease risk factors should focus on the population under 20 years of age.

[Efficacy of a multidisciplinary team for preventing hospital-acquired invasive aspergillosis: five years' experience]. / Efficacité des mesures environnementales dans la prévention de l'aspergillose invasive nosocomiale liée aux travaux: bilan de cinq années d'expérience.

Invasive hospital-acquired aspergillosis (IA) is responsible for lethal outbreaks. In 2002, an interdisciplinary team was created in the teaching hospital of Rouen in order to organize the surveillance of construction sites by the implementation of environmental measures of prevention. The aim of our study was to estimate the efficiency of these measures using an indirect indicator, reflecting the incidence of the cases of invasive nosocomial aspergillosis (AI): the consumption of antifungals. From the nominative prescriptions established, we studied the medical files about 210 patients to track down the number of IA cases in intensive care unit (ICUI) and in pediatric hematology-oncology units between 2002 and 2006. The incidence of the cases was put in parallel with the various periods of level 5-risk works during these five years. The relative risk of appearance of the disease was calculated. In pediatric haematology-oncology unit, 35 cases were diagnosed on 99 medical files which have been studied and in ICU 19 cases were classified on 93 studied files. The follow-up of the incidence in both units stake in parallel with the periods of level 5-risk works does not show increase of the number of cases. The calculated relative risk indicates the same result: the level 5-risk works are not a factor facilitating the appearance of invasive aspergillosis cases. This study shows the importance of the environmental measures of prevention during the periods of works within services for risk. The coordination of the actors within an interdisciplinary cell seems thus essential for the prevention of AIN.

Knowledge and opinions of surgical patients regarding nosocomial infections.

UNLABELLED: Sixty-five inpatients in various surgery departments were questioned about their knowledge and opinions regarding nosocomial infection, the information they were given on nosocomial infection, and their supposed attitude should they contract a nosocomial infection. RESULTS: Seventeen (26%, [16-39%]) were able to describe nosocomial infections as infections acquired in hospital. Identification of nosocomial infections as hospital-acquired infections was significantly associated with a high educational level and with having a member of their own family working in a health-related field. Fifty-two patients (80.0%, [68.2-88.9%]) stated that during their hospitalization they had received no information concerning nosocomial infections and 50 patients (76.9% [64.8-86.5]) mentioned that patients would welcome information about nosocomial infections. Thirty-three patients [50.8, 95% CI(38.6-62.9%)] declared that they would seek legal action against the hospital should they contract a nosocomial infection. There was a trend toward a higher probability of legal action in patients who rated their own risk of nosocomial infection as low or absent versus those who rated their own risk of nosocomial infection as medium or high (58.0% vs. 28.6%, p=0.051). The intention of seeking legal action against the hospital in case of nosocomial infection was not significantly influenced by patients' opinion regarding nosocomial infection preventability.

Environmental risk factors in paediatric inflammatory bowel diseases: a population based case control study.

BACKGROUND: Environmental exposures in early life have been implicated in the aetiology of inflammatory bowel disease. OBJECTIVE: To examine environmental risk factors prior to the development of inflammatory bowel disease in a paediatric population based case control study. METHODS: A total of 222 incident cases of Crohn's disease and 60 incident cases of ulcerative colitis occurring before 17 years of age between January 1988 and December 1997 were matched with one control subject by sex, age, and geographical location. We recorded 140 study variables in a questionnaire that covered familial history of inflammatory bowel disease, events during the perinatal period, infant and child diet, vaccinations and childhood diseases, household amenities, and the family's socioeconomic status. RESULTS: In a multivariate model, familial history of inflammatory bowel disease (odds ratio (OR) 4.3 (95% confidence interval 2.3-8)), breast feeding (OR 2.1 (1.3-3.4)), bacille Calmette-Guerin vaccination (OR 3.6 (1.1-11.9)), and history of eczema (OR 2.1 (1-4.5)) were significant risk factors for Crohn's disease whereas regular drinking of tap water was a protective factor (OR 0.56 (0.3-1)). Familial history of inflammatory bowel disease (OR 12.5 (2.2-71.4)), disease during pregnancy (OR 8.9 (1.5-52)), and bedroom sharing (OR 7.1 (1.9-27.4)) were risk factors for ulcerative colitis whereas appendicectomy was a protective factor (OR 0.06 (0.01-0.36)). CONCLUSIONS: While family history and appendicectomy are known risk factors, changes in risk based on domestic promiscuity, certain vaccinations, and dietary factors may provide new aetiological clues.

[Efficiency of hepatitis C virus screening before and after blood transfusion]. / Evaluation médicoéconomique du suivi sérologique péritransfusionnel du virus de l'hépatite C.

UNLABELLED: Efficiency of a viral hepatitis C screening strategy before and after blood transfusion has to be evaluated. METHODS: Four screening strategies were virtually applied to the population of transfused patients at Rouen University Hospital during 1996 and then compared : the first without any systematic HCV screening test; the second with systematic testing both before and 3 months after transfusion; the third with systematic testing both before and 6 months after transfusion ; the last defined as systematic testing before transfusion only. The efficacy (i.e. number of positive tests), the efficiency (i.e. average cost per positive test) and the marginal costs of moving from a strategy to another one were assessed using decision analysis. RESULTS: The efficacy of systematic screening test before transfusion only (361 per positive test), systematic testing both before and three months after (523 per positive test) or six months after (488 per positive test) transfusion was similar, but the efficacy of the strategy without any systematic screening test (385 per positive test) was lower. The systematization of screening test both before, and three months, or 6 months after transfusion lead to a marginal cost of 619 , and 559 per positive test respectively. The systematization of testing before transfusion only lead to a marginal cost of 343 per positive test. Adding systematic testing after transfusion lead to a marginal cost of 5824 per positive test. CONCLUSION: Systematic screening tests before transfusion only can be considered as the most efficient strategy.

Opposite evolution in incidence of Crohn's disease and ulcerative colitis in Northern France (1988-1999).

BACKGROUND: Northern France was characterised by a high incidence of Crohn's disease (CD) and a low incidence of ulcerative colitis (UC) according to the first inquiry undertaken in the late 1980s. AIMS: To assess the trends in the incidence of inflammatory bowel disease (IBD) over a 12 year period (1988-1999) in the same area of Northern France. PATIENTS: Patients living in Northern France (Nord, Pas-de-Calais, Somme, and Seine Maritime--total of 5,790,526 inhabitants) between 1988 and 1999 were included in the study. Case ascertainment was established according to methodology previously described. METHODS: Trends in incidence were studied using a Poisson regression model in four three year periods (1988-90, 1991-93, 1994-96, and 1997-99) adjusted for age at diagnosis and sex. Incidence rates were standardised for age with the European standard population. RESULTS: During 1988-99, 7066 cases of IBD were recorded (56.8% CD, 37.7% UC, and 5.5% indeterminate colitis). Mean annual incidence rate of CD increased from 5.2/100,000 inhabitants in 1988-90 to 6.4 in 1997-99 (adjusted p for trend <0.001). In contrast, the incidence of UC decreased from 4.2 to 3.5 (adjusted p for trend <0.001). The ileocolonic subtype of CD increased by 25% even though median age at diagnosis and frequency of digestive investigations were not different. CONCLUSIONS: Contrary to what has been reported in other countries in Northern Europe, the incidence of CD increased by 23% in 12 years in Northern France while that of UC decreased by 17% during the same period. This indicates that some factors which influence IBD frequency (in both directions) are still at work in this area of Europe, and that further studies aimed at identifying these should be performed. The rising incidence of CD could enhance the burden of this disease on the public health system in France.

[Factors affecting quality of care for elderly subjects undergoing surgery for hip fracture: review of the literature]. / Facteurs de qualité de la prise en charge des sujets âgés opérés d'une fracture de l'extrémité supérieure du fémur: revue de la littérature.

Hip fractures are one of the leading causes for admission of elderly subjects to healthcare facilities. Because of population aging, the incidence of hip fractures has increased considerably over the last years and will continue to increase in industrialized countries. Hip fracture in an elderly subject may be life threatening and has a significant functional and social impact not only because of the fracture itself, but also because of the risk of complications related to the patient's health status and the long hospital stay. The purpose of this work was to identify in the published literature professional practices, excepting the surgical procedure, associated with better early and long-term outcome in elderly patients with hip fracture. Questions raised concerning the patient's hospital stay include factors related to the preoperative phase (time to surgery, usefulness of traction), the operation itself (antibiotic prophylaxis, anesthesia technique), and the postoperative phase (prevention of venous thrombosis, malnutrition, episodes of confusion, duration of indwelling bladder catheter, correction of anemia, geriatric care during the stay in the orthopedic ward, early and intense rehabilitation, prevention of recurrence). Among these factors, several appear to be associated with better outcome, including long-term outcome--surgery as early as possible in light of the patient's general status, antibiotic prophylaxis in accordance with standard recommendations (SFAR), prevention of venous thrombosis with low-molecular-weight heparin initiated at admission and associated with elastic contention. Oral nutritional support is probably beneficial and should be proposed for all patients. Particular attention must be given to prevention of confusion in order to reduce the rate of institutionalization. The rythm of rehabilitation exercises should be at least five sessions per week. Finally, there are several methods, which are effective in preventing recurrence, taking into account osteoporosis, risk of falls. Preventive measures should be instituted for all patients undergoing surgery for hip fracture.

Mediastinitis after coronary artery bypass graft surgery: influence of the mammary grafting for diabetic patients.

Mediastinitis is a severe complication of coronary artery bypass graft surgery (CABG) particularly when harvesting internal mammary arteries (IMA). CABG in diabetic patients often uses two IMA because the saphenous graft is damaged. To our knowledge this risk of mediastinitis has not previously been reported in diabetic patients. All consecutive diabetic patients undergoing CABG over a three-year period from 1998 to 2000 were included in the study. Data recorded were: age, sex, duration of stay, whether one or two IMA were used, diagnosis of mediastinitis. Calculation of relative risk and analysis of trends by chi2 trend tests was also performed. In total 256 diabetic patients were included in the cohort. The incidence of mediastinitis was 4.3% (11/256). The risk of mediastinitis was higher in patients with two IMA than in patients with one IMA (relative risk 5.97, 95 CI 1.63-21.93, P=0.004). Age and sex were not confounding factors. No patients with mediastinitis died. Bilateral IMA grafting is associated with higher risk of mediastinitis in diabetic patients. The authors suggest that the risk of mediastinitis in diabetic patients should be taken into consideration when cardiac surgeons choose unilateral or bilateral IMA harvesting for surgery.

Assessment of prolonged hospital stay attributable to surgical site infections using appropriateness evaluation protocol.

BACKGROUND: The accepted standard in estimating the stay prolongation attributable to surgical site infections is the matched-cohort study method (MCS), which is associated with selection bias. The Appropriateness Evaluation Protocol (AEP) has been used to estimate stay prolongation attributable to nosocomial infections but has not been validated specifically for surgical site infections. AIM OF THE STUDY: To compare estimates of stay prolongation attributable to surgical site infections after digestive surgery, obtained by AEP and by MCS. METHODS: Sixty-five surgical site infections after digestive tract surgery were analyzed by AEP and MCS. AEP stay prolongation was the number of days judged specifically appropriate for the care of surgical site infections. MCS stay prolongation was the difference of stay duration in surgical site infection cases and two controls matched by age, sex, and diagnosis-related groups. Sensitivity and specificity of AEP, and agreement between both methods, were calculated. RESULTS: The mean AEP stay prolongation was 3.5 days vs 7.2 days for MCS. The sensitivity of AEP was 58% and the specificity was 75%. The agreement between the two methods was poor. CONCLUSION: Surgical site infections after digestive tract surgery increased the hospital stay. Accurate estimations of a prolongation of stay will vary according to the method selected.

Postoperative management of stage II/III colon cancer: a decision analysis.

BACKGROUND & AIMS: Two separate decisions must be made for the management of patients with resected stage II/III colon cancer: whether to begin adjuvant chemotherapy and whether patients should be included in a follow-up protocol consisting of regular monitoring of carcinoembryonic antigen level and of colonoscopy and imaging. The standard management for these patients is adjuvant chemotherapy for stage III patients and follow-up for stage II/III patients with resected colon cancer. METHODS: Decision analysis was used to compare the effectiveness (5-year survival rate) and cost-effectiveness ratio of 7 strategies of treatment and follow-up. RESULTS: The most cost-effective strategies were adjuvant chemotherapy for all patients with stage II/III resected colon cancer, with either no follow-up or follow-up only for patients aged less than 75 years with a seric preoperative carcinoembryonic antigen level of >5 ng/mL (5-year survival, 62.3% or 62.7%; cost per surviving patient, $8254 or $8657, respectively). The order of efficacy of the strategies was insensitive to changes in the values of the studied variables. The method of follow-up does little to improve 5-year survival but adds substantial cost. CONCLUSIONS: The current standard strategy may not be the most cost-effective strategy for the management of patients with resected colon cancer.

[Risk factors of contamination by hepatitis C virus. Case-control study in the general population]. / Facteurs de risque de contamination par le virus de l'hépatite C. Etude cas-témoins en population générale.

UNLABELLED: In 30% of patients with hepatitis C virus, the source of infection is unknown. OBJECTIVE: To identify the risk factors of infection by hepatitis C virus in a case-control study. METHODS: Cases had positive hepatitis C virus serology, and were living in Fecamp (Normandy, France). Controls (2 for each case) were age and sex-matched subjects with negative hepatitis C virus serology, living in Fecamp. Demographic, medical, professional, and environmental data were collected. Statistical analysis included chi 2 or Fisher's exact test and multiple logistic regression. RESULTS: The risk factors of hepatitis C virus by univariate analysis were: history of transfusion, high number of sexual partners, hepatitis C virus infection in a relative, history of digestive or genitourinary surgery, an invasive medical procedure, digestive endoscopy, biopsy, lumbar or pleural puncture, medical care after an accident, injections, multiple deliveries or abortion. Risk factors of hepatitis C virus infection by multivariate analysis: hepatitis C virus infection in a relative (Odds ratio: 4.58), history of transfusion (Odds ratio: 2.32), of a surgical procedure (Odds ratio: 2.50), of medical care after an accident (Odds ratio: 1.51), of injections (Odds ratio: 2.24). CONCLUSION: Our results suggest the possible nosocomial transmission of hepatitis C virus. Intrafamilial transmission is also possible.

[Cost of early management of chronic inflammatory intestinal disease]. / Coût de la première prise en charge des maladies inflammatoires chroniques de l'intestin.

OBJECTIVES: The aim of this study was to assess the cost of the first management of inflammatory bowel disease (IBD) from the onset of first symptoms until 6 weeks after the diagnosis. This cost was calculated in French francs (FF) for all IBD and namely for Crohn's disease (CD), ulcerative colitis (UC), and ulcerative proctitis (UP). MATERIAL AND METHODS: Data concerning 258 patients were collected by the mean of a standardized questionnaire from 3 different sources: the patient, his general practitioner, and his gastroenterologist. RESULTS: Two hundred and fifty eight patients were included: 144 CD (55.8%), 76 UC (29.5%), 30 UP (11.6%), and 8 chronic unclassifiable colitis (CUC) (3.1%). The mean direct costs of the diagnosis (m +/- SD) were 23,116 +/- 40,820 FF for CD, 10,628 +/- 17,316 FF for UC and 3,451 +/- 2,743 FF for UP. Although unplanned hospitalizations occurred in only 38% of the patients (98/258), they represented the 3/4 of the mean costs: 78.2% for CD and 64% for UC. Indirect costs generated by days off work were 4,719 +/- 6,610 FF for CD, 2,996 +/- 6,897 FF for UC and 1,230 +/- 3,622 FF for UP. CONCLUSION: The first management of a patient with CD was twice more expensive than the one with UC and 6.5 times than the one with UP.

[Diagnostic and therapeutic management of patients with chronic inflammatory bowel disease]. / Prise en charge diagnostique et thérapeutique des patients atteints de maladie inflammatoire chronique de l'intestin.

BACKGROUND: The period of time required for the diagnosis of a chronic illness depends on initial clinical symptoms and their perception by the patient and the physicians. The aim of this study was to describe the procedures of diagnosis of incident cases of Inflammatory Bowel Disease (IBD). METHODS: Patients reported by the Registry of inflammatory bowel disease of northern France (EPIMAD) in 1994 were included. Standardized questionnaires describing clinical history, patient behavior, medical consultations and examinations were collected by an interviewer practitioner from three sources: patients, general practitioners (GP) and gastroenterologists (GE). Patients were divided in 2 groups according to the time between symptom onset and diagnosis: more than 9 months or less than 9 months (D > 9 and D < or = 9). RESULTS: 258 patients were included: 144 Crohn's disease (CD) (56%), 106 ulcerative colitis (UC) (41%) and 8 chronic unclassifiable colitis (CUC). Median time between symptom onset and diagnosis was 3 months, 196 (76%) patients belonged to the group D < or = 9 and 62 (24%) to the group D > 9. There was no difference between the 2 groups for initial clinical symptoms. The delay between symptom onset and the consultation to the GP and the GE was longer in the group D > 9: respectively 1 month vs 0 and 7.6 vs 2. Thirty-five percent of patients in the group D > 9 had consulted more than one GP vs 14% (p < 0.05). Diagnosis management by the GE was the same in both groups. Patients of group D < or = 9 had more often perceived their symptoms as serious (p < 0.05). CONCLUSIONS: Delay to diagnosis in a quarter of patients with IBD was more than 9 months. This later diagnosis was not due to patient management by the GE but rather to a longer delay to consulting the GP and between GP and GE referral. Patient interpretation of the symptoms could also explain the variability of this delay.

Incidence of rectosigmoid adenomatous polyps in subjects without prior colorectal adenoma or cancer: a prospective cohort study.

BACKGROUND: Subjects without known colorectal adenomas or cancer constitute a large majority of the population where 85% of all cases of colorectal cancer are thought to occur. Consequently these people should be considered for screening to decrease mortality from colorectal cancer in the general population. AIMS: To estimate the incidence rate of rectosigmoid adenomas in these subjects. METHODS: Subjects without adenomas or cancer at a previous examination which had visualised the rectosigmoid underwent a fibre endoscopy every three years. Endoscopic data and population characteristics were collected prospectively. RESULTS: A total of 450 subjects fulfilled the selection criteria; 287 (64%) underwent at least two examinations, and 163 had three or more. At the second examination, with a mean delay of 39 months, the incidence rate of rectosigmoïd adenomas was 1.50% per patient year. The rate was 1.75% per patient year (95% CI 0.80-3.33) at the third endoscopy with an additional mean delay of 38 months. The cumulative incidence rate at six years was 7.3% (95% CI 4.3-10.3), representing a mean of 1.2% per patient year. This rate increased with age and was higher for men than for women after age adjustment (p< 0.03). CONCLUSIONS: The incidence rates are very low compared with those of patients with prior adenomas. These results should be considered in establishing rectosigmoid adenoma screening strategies.

[Upper digestive hemorrhage in patients over 80 years of age: incidence and prognostic factors]. / Hémorragies digestives hautes du sujet âgé de plus de 80 ans: fréquence et facteurs pronostiques.

OBJECTIVES: The aim of this study was to assess the incidence and the prognosis of upper gastrointestinal hemorrhage in patients aged over 80 years. METHODS: Between January and December 1993, among 360 patients admitted for upper gastrointestinal hemorrhage, 63 were older than 80 years (18%). Data were prospectively collected in all patients. RESULTS: Before admission, 8 experienced a lipothymia and 3 a shock. The median initial hematocrit was 29%. Endoscopy was performed in 59 patients and a cause was determined in 49 (83%). The main cause of bleeding was gastric and duodenal ulcer (n = 26, 53%) and ten of them were graded Forrest < or = IIb. Endoscopy did not contribute in 10 patients and was impossible in 4 because of an hemodynamic failure. Gastrotoxic drugs intake was found in 28 patients: non steroidal anti-inflammatory drugs (n = 14) and aspirin (n = 14). Endoscopic injection therapy was performed in the 10 patients with gastric or duodenal ulcer < or = Forrest IIb and permanent hemostasis was achieved in 8 out of 10. No further bleeding was seen in 53 patients (84%), while bleeding persisted in 4 and rebleeding occurred in 6 (9.5%) (3 duodenal ulcers, 2 gastric carcinomas and 1 esophageal varices bleeding). Only one patient required emergency surgery (rebleeding duodenal ulcer). The overall mortality was 12/63 (19%):50% of the deaths were related to hemorrhagic complications and 50% to concomitant disease. Only one patient among those taking gastrotoxic drug died from bleeding. The risk factors of death from bleeding were: initial shock (P = 0.02), lipothymia before admission (P = 0.02), rebleeding (P < 0.01), persistence of bleeding (P < 10(-4)). Gastrotoxic drugs intake was associated with a favorable prognosis (P < 0.05). The prognosis was not significantly affected by an initial hematocrit < 30% or blood units transfused > or = 4. CONCLUSIONS: People older than 80 years account for a large proportion of upper gastrointestinal bleeding (18%). The more common cause is gastric or duodenal ulcer. Mortality in these patients is high (19%). An initial shock or lipothymia, rebleeding or persistence of bleeding worsens prognosis. Gastrotoxic drugs intake is frequent (44%), but is associated with a good prognosis.