The upper and lower gastrointestinal bleeding, not related to portal hypertension, is a common disorder in the elderly. Indeed, in 1996, in a French study, the median age of patients hospitalized for upper gastrointestinal bleeding was 68. During the same period in the studies reported in English the median age was 71. If epidemiological data concerning lower gastrointestinal bleeding are rare, the average age of hospitalized patients varies from 63 to 77 depending on the study. Due to improvement in endoscopic haemostatic procedures and current resuscitation methods, gastrointestinal bleeding prognosis has greatly improved, whereas anaemia related to a bleeding episode remains a frequent complication of gastrointestinal bleeding in elderly patients.

Among elderly patients over 65, the prevalence of anaemia varies from 8 to 44% depending on the criteria used and populations studied. The occurrence of a bleeding episode can either induce anaemia or exacerbate pre-existing anaemia. Physicians in charge of gastrointestinal bleeding are often unaware of anaemic consequences in the elderly patients which can often be serious. Various studies have shown that anaemia increases morbidity and mortality rates in the elderly. Life expectancy is independently significantly lower for anaemic patients over 65, than for non-anaemic subjects. Anaemia is also a risk factor for the occurrence of cardiovascular and neurological complications, impairment in cognitive function and increased risk of falling.

Iron deficiency and anaemia induced by bleeding episodes in patients over 65 hospitalized for upper or lower gastrointestinal bleeding should be corrected rapidly and effectively. Currently, the cost and risks of infection or cardiovascular-related complications of transfusions lead to limiting red blood cell transfusion with a goal average of 9 g/dL haemoglobin. It is also necessary to develop alternatives to massive transfusions. The correction of iron deficiency promotes erythropoiesis and can quickly correct anaemia.

In clinical practice, the effectiveness of iron intake by the oral route is limited by the frequent occurrence of significant gastrointestinal side effects that limit patient compliance and limited absorption necessitating prolonged treatment to correct iron deficiency.

The black colour of stools caused by taking oral iron supplements also makes it difficult to detect a possible recurrence of bleeding after hospitalization.

The prescription of intravenous iron seems more suitable for a rapid and complete correction of iron deficiency after gastrointestinal bleeding. The main objective of our study is to evaluate efficacy of intravenous iron for the correction of anaemia, measured by haemoglobin at week 6 (W6) in patients aged over 65, after gastrointestinal bleeding. Secondary objectives were to assess the speed of anaemia correction, the tolerance of intravenous iron supplementation, the rate of re-hospitalization within 6 months after discharge and patients quality of life. This is a prospective multicenter randomized study versus placebo. After obtaining informed consent, all patients aged over 65 admitted with upper or lower gastrointestinal bleeding, with successful outcome, not related to portal hypertension, responsible for persistent anaemia (definition: Hb < 11 g / dL) after hospitalization will be included in the study. Patients will be treated for their bleeding event in the usual manner of each centre with target for transfusion of 9 g / dL haemoglobin. The absence of external bleeding and haematocrit and/or constant haemoglobin levels will be considered as the end of bleeding.

Day 1 was arbitrarily defined as the day the patient left hospital. The protocol at Day - 1 included: obtaining informed consent of the patient, determination of iron and ferritin blood levels and complete blood count. and randomization intravenous iron injection , (Ferinject) versus Placebo. Intravenous iron injection will be performed at Day 0. A complete blood count will be performed at week 6 and month 6. Patients will be reviewed in consultation at week 6 and at month 6 to obtain related intercurrent events and assess their quality of life.

The results of this study could lead to changes in the care of older patients hospitalized for gastrointestinal bleeding.

Inclusion Criteria:

1. Patients aged over 65 hospitalized for upper or lower GI bleeding with positve outcome during hospitalization without surgery, and with persistent anaemia (Hb <11g/dL),
2. Signed informed consent,
3. Patients with National Health Insurance,

Exclusion Criteria:

1. Uncontrolled haemorrhage defined by any new externalizing and / or a decrease of haemoglobin and haematocrit levels,
2. GI bleeding related to portal hypertension or malignancy,
3. The absence of anaemia,
4. Cancer evolution,
5. Patient under guardianship, curatorship or unable to supply consent,
6. Iron overload,
7. History of asthma
8. History of eczema
9. Hypersensitivity to any component
10. Decompensated liver cirrhosis
11. Infection during treatment or uncontrolled infection 12 Rheumatoid arthritis

13. Acute renal failure