The efficacy of octreotide therapy in chronic bleeding due to vascular abnormalities of the gastrointestinal tract


Nardone Dr Cattedra di Gastroenterologia, Università degli Studi di Napoli ‘Federico II’ via Pansini, 5, 80131 Napoli, Italy. E-mail:



: The treatment of angiodysplasia and watermelon stomach, vascular abnormalities implicated in gastrointestinal bleeding of obscure origin, is a major clinical problem.


: To determine the efficacy of octreotide in patients with long-standing gastrointestinal bleeding due to acquired angiodysplasia and watermelon stomach, resistant to previous treatments and not suitable for surgery because of old age and/or concomitant disorders.

Patients and methods

: We treated 17 patients (seven had isolated angiodysplasia, seven had multiple upper and lower gastrointestinal angiodysplasia, and three had watermelon stomach) with octreotide (0.1 mg subcutaneous t.d.s. for 6 months). Six of the patients had liver cirrhosis, one had Glanzmann-type platelet derangement, two had cardiovascular diseases and one had chronic uraemia.


: Octreotide treatment stopped bleeding in 10 patients. A transient improvement was observed in four, who needed subsequent cyclical retreatment to correct low haemoglobin levels. No effect was observed in three, probably due to the severity of the concomitant disorders.


: Octreotide is a safe drug that may be useful to control the recurrent gastrointestinal bleeding due to acquired angiodysplasia and watermelon stomach, especially in patients who are not candidates for surgery due to old age and/or concomitant disorders.


Hereditary and acquired vascular abnormalities of the gastrointestinal tract have been variously classified and are described as angiodysplasia, teleangectasias, vascular ectasias, diffuse antral vascular ectasias ‘watermelon stomach’, angiomas, and arterovenous malformation, often used interchangeably.1[2]–3 Here we focus on angiodysplasia and watermelon stomach, which are well-defined acquired vascular abnormalities, not associated with skin lesions, systemic vascular diseases or hereditary disorders.

Gastrointestinal vascular abnormalities are responsible for 2–8% of all cases of bleeding, and are the most common cause of gastrointestinal bleeding of obscure origin in the elderly.1[2][3][4][5][6][7][8]–9 The symptoms range from acute, recurrent, self-limited digestive bleeding to chronic iron deficiency anaemia and the diagnosis, generally difficult, can be delayed by up to 40 months after initial symptoms.1[2][3][4][5][6][7][8][9][10][11]–12 In 90% of cases angiodysplasia lesions appear endoscopically as flat cherry-red spots, 2–10 mm in size1[2]–3 and their histological features are difficult to identify without complex techniques.13, 14

The treatment of vascular abnormalities is a major clinical problem. The therapeutic endoscopic or surgical approach is often ineffective and intraoperative mortality ranges from 10% to 20%, while symptoms recur in 30–40% of cases.3, 10, 15 The treatment of angiodysplasia often fails because the lesions are spread throughout the gastrointestinal tract often in sites inaccessible to endoscopy.1[2]–3, 5, 11, 15[16]–17 Angiodysplasia lesions are restricted to the right colon in 54–100% of cases1[2]–3, 8, 9, 15[16][17][18][19]–20 but lesions spread throughout the upper, lower and middle gastrointestinal tract are present in one-third of cases.1[2][3]–4, 17, 19, 20

In water melon stomach, which appears as tortuous dilated vessels restricted to the gastric antrum, laser endoscopy is an alternative to surgical antrectomy.10, 21, 22, 23 However, multiple endoscopic sessions are required, extensive lesions are difficult to ablate completely with this procedure, and instances of perforation of the stomach and rebleeding are not insignificant.3, 10 Finally, patients with vascular abnormalities are usually in the seventh or eighth decade of life when invasive procedures are often contraindicated by concomitant disorders.1[2]–3, 8, 15, 16, 24

Various drugs have been proposed for the treatment of vascular abnormalities, but thus far, none has been widely accepted.4, 10, 25[26][27][28]–29 We describe 17 patients with acquired vascular abnormalities (14 with angiodysplasia and three with watermelon stomach) resistant to previous endoscopic or arteriographic treatment and not candidates for surgery, in whom the octreotide treatment stopped or reduced bleeding during a long-term follow-up.


Seventeen patients were referred to our unit between 1991 and 1998 because of a long history of chronic and/or acute gastrointestinal bleeding requiring several blood transfusions and repeated cycles of iron therapy. Neither treatment was sufficient to ensure stable haemoglobin levels and clinical tolerance to anaemia. Most were females (12 out of 17) the mean age was 63 years, ranging from 23 to 86 years; four subjects were less than 60 years old (23, 30, 45 and 52 years). No patient had skin involvement or a personal or family history of haemorrhagic disorders.

Anaemia due to occult or overt digestive bleeding, lasted from 1 to 12 years before the correct diagnosis (Table 1) and blood transfusions ranged from 1 to 80 units per year before our observation. All patients underwent upper and lower endoscopy; seven (recently enrolled patients) underwent enteroscopy; small bowel X-ray was carried out in 10 because enteroscopy was not available at the time of diagnosis.

Table 1.  . Clinical characteristics of the patients with angiodysplasia of the gastrointestinal tract Thumbnail image of

After written informed consent was obtained from each patient, octreotide was given to all subjects at a dosage of 0.1 mg t.d.s. subcutaneous (sc) for 6 months.


The Wilcoxon’s matched pair test was used to evaluate haemoglobin levels and blood unit transfusions in patients before and after 1 year of octreotide treatment. P-values of less than 0.05 were considered to be statistically significant.


The clinical data of 17 case records are listed in Table 1. Seven patients had isolated angiodysplasia, four in the right colon (patients nos 1, 5, 8, and 17), two in the small bowel (patients nos 3 and 14) and one in the stomach (patient no. 4), while seven patients had multiple angiodysplasia involving the upper and lower gastrointestinal tract (patients nos 2, 6, 7, 9, 13, 15, 16). Three patients, all female, had watermelon stomach (patients nos 10, 11 and 12). Concomitant disorders were viral cirrhosis in six subjects (3 out of 6 with watermelon stomach), cardiac valve alteration in two (one aortic stenosis and one mitralic failure), Glanzmann’s thrombopathy and uraemia undergoing dialysis, in two subjects, respectively.

Endoscopic treatment (electrocoagulation and laser) and arteriographic embolization did not reduce the rate of bleeding in any of the treated patients (11 out of 17). Surgical resection was contraindicated because of age and/or associated disorders in all subjects with the exception of one, who refused.

The treatment was stopped after 6 months; the haemoglobin and iron levels and faecal occult blood test were monitored during therapy and thereafter in all subjects. The efficacy of therapy was evaluated by the monitoring of haemoglobin levels and the need for blood transfusions or iron therapy (Table 2 and Figures 1 and 2). The average haemoglobin level increase was from 5.7 to 11.1 g/dL (< 0.0005), while blood units transfused decreased from an average of 8.8 to 1.5 units per year (< 0.0005). Octreotide was successful in 10 subjects, who therefore no longer required iron supplementation. A transient improvement was obtained in four patients (nos 11, 12, 13 and 14). These patients required, 1 to 2 months after cessation of therapy, retreatment with octreotide which was continued cyclically depending on haemoglobin level (on an average of 10 days every 2–3 months). During this time, only patient no. 11 received iron supplementation; the other three patients developed iron intolerance. No effect was observed in three (patients nos 15, 16 and 17) who required recurrent cycles of iron therapy and blood transfusions (Table 2, and Figures 1 and 2). No significant side-effects occurred in any subject. Gastrointestinal endoscopy was repeated in each patient six or 12 months after octreotide treatment. Small lesions (isolated dots) disappeared (patients nos 1, 2, 7 and 9), while larger lesions become smaller and lighter in colour (patients nos 4 and 10). No change was observed in patients resistant to treatment (patients nos 15, 16 and 17).

Table 2.  . Effect of ocreotide treatment (0.1 mg t.d.s. for 6 months) in the 17 patients of the study Thumbnail image of
Figure 1.

. Transfusion requirement (red cell per unit per year) in patients before and during octreotide treatment (0.1 mg t.d.s.). * P Wilcoxon test for paired data; § Responders; ^ Transient improvement; # non responders.

Figure 2.

. Haemoglobin levels in patients before and during octreotide treatment (0.1 mg t.d.s.). * P Wilcoxon test for paired data; § Responders; ^ Transient improvement; # non responders.


Case no. 15

Patient number 15, a woman, aged 45 years, was admitted to our department with a history of chronic anaemia lasting 2 years (average haemoglobin level 5.5 g/dL) due to chronic faecal occult blood and sporadic rectal bleeding, treated with recurrent cycles of iron therapy and about 12 blood transfusions a year. She reported previous surgery for appendicitis, gallstones and uterine leiomyoma. Lower endoscopy revealed multiple dot angiodysplasia in the rectum and left colon. Enteroscopy showed multiple (at least 10) angiodysplasia lesions in the proximal jejunum. No lesions were detected in the upper gastrointestinal tract. She was treated with octreotide (0.1 mg t.d.s. sc), but during the 4-year follow-up she required iron supplementation and blood transfusions (6 units per year).

Case no. 16

This 75-year-old woman affected by hereditary platelet disturbance (Glanzmann type) was admitted to our department in 1997 with a 12-year history of chronic iron deficiency anaemia and recurrent episodes of melaena mixed with blood (rectal haemorrhage). She had been given recurrent cycles of iron therapy and multiple blood transfusions (an average of five units per year) but because of clinical worsening she was given an average of one blood unit per week. Lower endoscopy showed multiple angiodysplasia lesions in the right colon and in the small bowel. The mesenteric arterial embolization did not modify the rate of gastrointestinal bleeding. She was treated with octreotide (0.1 mg t.d.s. sc) and haemoglobin levels remained stable for eight months, after which they decreased to 7 g/dL notwithstanding treatment. An increase in the octreotide dosage (0.1 mg × 6 per diem) did not result in an improvement; she currently needs an average of two blood units per month.

Case no. 17

This 67-year-old man with renal failure, severe chronic anaemia (haemoglobin blood level 6 g/dL) lasting one year and occult faecal blood, had received a total of 10 blood units and several cycles of intravenous iron therapy. Upper endoscopy revealed gastro-duodenal erosions with positive Helicobacter pylori infection; lower endoscopy showed angiodysplasia lesions in the right colon. He was given octreotide (0.1 mg t.d.s. sc) and H. pylori-eradication therapy which cured the ulcer (15 days of clarithromycin 500 mg b.d. plus amoxocillin 1 g b.d. plus omeprazole 20 mg b.d.), but a further episode of melaena caused deterioration in renal function, and the patient underwent dialysis. During the 2-year follow-up, the haemoglobin level decreased slowly and the patient was transfused six times a year.


It is postulated that vascular abnormalities could be the result of age-related degenerative processes.1[2]–3 In fact, most patients are in the seventh to eighth decades of life.1[2]–3, 8, 9, 15, 24 Furthermore, angiodysplasia lesions are found prevalently in the right colon where the wall pressure is high1[2]–3, 15, 17[18][19]–20 and watermelon stomach is restricted to the antrum where peristaltic activity is vigorous and may causes the prolapse of the mucosa.1[2]–3, 30

The association of angiodysplasia with other liver, renal, cardiac and lung chronic diseases seems to be casual, and could depend on the higher prevalence of these disorders in advanced age. Differently, the close association of watermelon stomach with cirrhosis seems to suggest that portal hypertension plays a key role in this condition.10, 30 Indeed, all our patients with watermelon stomach had cirrhosis. In any event, these associated disorders, by enhancing the haemorrhagic tendency through portal hypertension, coagulopathy, central venous hypertension, uraemia, etc., can contribute to the development of bleeding. In these cases the treatment, when possible, of the associated disorders does not change the basic mechanism or the clinical outcome of the gastrointestinal bleeding lesion.15, 17, 31

A conservative therapeutic approach is recommended in asymptomatic subjects whose lesions are discovered incidentally during endoscopic examination, but in symptomatic subjects the high rate of rebleeding (about 50%) justifies immediate treatment.1[2]–3, 15, 17, 32[33]–34

Most of these patients are treated by endoscopic procedures: monopolar35 or bipolar36 electrocoagulation, argon37 or Nd–Yag23, 38, 39 laser, or injection sclerotherapy.40 These procedures represent the gold standard. However, watermelon stomach affects the whole gastric antrum,3, 10 and the angiodysplasia lesions are often spread throughout the gastrointestinal tract or are not accessible to the endoscope.1[2]–3,5, 11, 15, 20 Moreover, the rebleeding rate ranges from 14% to 53% for a follow-up of 1–3 years.15, 17, 19, 32

An alternative is surgery; however, surgery has a high morbidity, significant mortality (from 10% to 20%) and a recurrence rate ranging from 30% to 40%.1[2]–3, 10, 15, 17

Drug treatment should be used in cases of failure of previous treatment, diffuse or inaccessible lesions or other contraindications to invasive treatment. The outcome of medical therapy with oestrogen is not consistent.15, 25[26]–27, 41, 42 In 1991 Torsoli reported a clinical case of intractable gastrointestinal bleeding due to diffuse angiodysplasia of the small intestine that was stopped with octreotide 0.05 mg b.d.43 Subsequently, octreotide therapy was successful in four cases of chronic bleeding due to angiodysplasia of the small intestine.44, 45 It is routinely used in acute gastrointestinal bleeding because it reduces splanchnic arterial blood flow.46

Octreotide is a long-acting somatostatin analogue which exerts many biological effects in the gastrointestinal tract, i.e. inhibition of gastric acid, of intestinal and endocrine secretion (e.g. gastrin, cholecystokinin, vaso-intestinal peptide and motilin), and is of therapeutic value in patients with carcinoid tumour.46, 47 In this contest, it is noteworthy that proliferation of neuroendocrine cells has been observed in the resected antrum of patients with watermelon stomach.48 Furthermore, octreotide inhibits the growth factors EGF,49 b-FGF50 and IGF-1,51 and is responsible for angiogenesis suppression.52 In fact, in experimental models of vascularization, the number of blood vessels is reduced in animals receiving octreotide.53, 54

Interestingly, the endoscopic evolution of vascular abnormalities during the 12-month follow-up in our successfully treated patients supports an angiogenetic inhibition effect: the tiny angiodysplasia lesions disappeared while larger lesions became smaller and lighter in colour.

Finally, octreotide induces a decrease in smooth muscle contractility,55 and according to Boley, the development of vascular abnormalities may also depend on repeated low grade obstruction of submucosal veins at the point where they cross the muscolaris layers.24

In our study, octreotide modified the clinical course in 14 out of 17 patients: in 10 it corrected haemoglobin levels, which became stable during the follow-up, while in four patients with overt rebleeding or anaemia after cessation of treatment, cyclic octreotide was required. Unlike a recent study,56 we found that octreotide stopped the bleeding caused by watermelon stomach in one patient and brought about a transient improvement in the remaining two over a follow-up period ranging from 36 to 48 months.

In our three non-responding patients the associated disorders and/or their treatment may have negatively affected the therapy. In patient no. 15, a history of multiple abdominal surgical interventions suggests the presence of an artero-venous intestinal fistula, but the patient refused arterography; patient no. 16 had thrombocytopathy (Glanzmann’s disease); patient no. 17, affected by uraemia, received heparin in dialysis; systemic heparinization together with platelet olysfunction of chronic renal failure render these patients more susceptible to bleeding from multiple sites.57

In conclusion, this is the first study with a prolonged follow-up (between 48 and 84 months in seven subjects) showing that octreotide is an effective and safe drug that may be useful to control the recurrent gastrointestinal bleeding due to acquired vascular abnormalities, especially in patients who are not candidates for surgery due to old age and/or concomitant disorders.


This study was supported by a grant from MURST (40–60%) Rome, Italy.