Gastric antral vascular ectasia in children, rare presentation
- Matthew Pizzuto ,
- Sarah Ellul and
- Mohamed Shoukry
- Mater Dei Hospital, Msida, Malta
- Correspondence to Dr Matthew Pizzuto; matthew.a.pizzuto@gov.mt
Abstract
A 14-year-old boy, a known case of perinatal hypoxic cerebral palsy, presented to paediatric emergency with acute melaena and blood staining around feeding gastrostomy site. Physical examination revealed pallor, but no signs of distress with an unremarkable abdominal examination. Routine blood tests revealed normochromic. Abdominal ultrasound scan and Meckel’s scan were unremarkable. The patient underwent examination under anaesthesia of the perianal area and joint upper and lower gastrointestinal endoscopy. Streak-like gastritis with no signs of active bleeding lesions were noted and patchy areas of colitis involving the descending and sigmoid colon and the rectum. All clinical findings and evidence-based diagnosis matched gastric antral vascular ectasia. He was successfully managed conservatively with elemental hydrolysed feeding formula.
Background
Gastric antral vascular ectasia (GAVE) is a capillary-type vascular malformation, commonly affecting the gastric antrum. GAVE is a rare cause of non-variceal upper gastrointestinal (GI) bleeding.1 GAVE typically presents in middle-to-older age patients, mostly women (71%), with an average age of 73 years (range between 53 and 89 years).2 Prevalence in adults is 4%, but prevalence in children is unknown.2 GAVE commonly presents with iron-deficiency anaemia, but patients can present with acute massive fresh bleeding requiring ample amounts of red cell transfusion or with chronic GI bleeding.
Although pathogenesis has not been clearly identified in paediatrics, many theories have been postulated, including hormonal, autoimmune factors, mechanical stresses and haemodynamic alterations.3
Case presentation
A 14-year-old boy, known to suffer from cerebral palsy (due to perinatal hypoxic encephalopathy and Christianson syndrome), presented to paediatric emergency with a 2-day history of melaena unassociated with abdominal pain, coffee ground stained vomiting or bleeding from gastrostomy site. Christianson syndrome is an X-linked syndrome, associated with microcephaly, seizures and intellectual disability due to mutation in SLC9A6 gene. Patient was a wheelchair user and fed via gastrostomy tube. He had similar episodes during the previous 2 years and was treated conservatively with regular proton pump inhibitors and H2-receptor antagonist.
Physical examination revealed a malnourished boy weighing 34 kg, with tachycardia and pallor but no tachypnea. On palpation, he had a soft, non-tender abdomen and no masses were felt. Blood stains were noted around the gastrostomy site. Haemoglobin was 57 g/L, haematocrit 23.2% and mean cell volume 96.3 fL. Urine dipstick, renal profile, liver enzymes and coagulation screen were all within normal range.
After resuscitation, the patient was transfused 350 mL of packed red blood cells and treated conservatively with intravenous medications consisting of proton pump inhibitor (omeprazole), H2-receptor blocker (ranitidine) and sucralfate. Abdominal ultrasound scan (USS) and Meckel’s scan (MS) were performed. Both results were normal. Post-transfusion haemoglobin level was 8.4 g/dL.
The paediatric gastroenterology team was consulted and advised feeding change to elemental hydrolysed formula (amino acid-based formula). After clinical stabilisation and adequate preparation, the patient underwent oesophagogastroduodenoscopy (OGD) and full colonoscopy. Endoscopy findings revealed the presence of lower one-third oesophagitis, with no variceal lesions. No active bleeding was visible from the stomach wall mucosa. Lining of the gastric mucosa revealed hyperemic steak-like gastritis, raised on the surface and alternating with patchy normal-looking mucosa along the antrum and pylorus. No definite ulcers or polyps were otherwise visible; significant duodenitis over the first and second part of the duodenum were present. Macroscopically, there were areas of non-velvet appearance of the duodenal mucosa, distributed irregularly (figure 1). Endoscopic mucosal biopsies were obtained from different duodenal, gastric and oesophageal sites for histopathology review. Prior to colonoscopy, examination under anaesthesia (EUA) was performed, revealing healing posterior anal fissure with no ulcers. The colonoscopy showed colitis throughout the whole descending and sigmoid colon and the rectum (figure 2).
Oesophagogastroduodenoscopy showing the stomach pylorus (left) and stomach antrum (right).
Colitis present throughout the descending and sigmoid colon along with the rectum.
Biopsies obtained from the oesophagus (figure 3) and stomach (figure 4) showed mild chronic reflux changes, with the stomach showing foveolar hyperplasia. Biopsies taken from the colonoscopy noted a preserved crypt architecture with normal lamina propria cellularity. No evidence of active inflammation or microscopic colitis were visible.
Oesophageal biopsy showing stratified squamous epithelium featuring a basal infiltrate of lymphocytes, in keeping with chronic oesophagitis, usually seen in the presence of gastro-oesophageal reflux disease.
Stomach biopsy showing gastric antrum with a moderately dense lymphoplasmacytic infiltrate in the lamina propria associated with reactive foveolar hyperplasia and stromal fibromuscularisation, in keeping with chronic reactive gastritis. No intestinal metaplasia is present. No Helicobacter organisms are noted.
Clinical and endoscopic findings were discussed at the paediatric multidisciplinary team meeting. The gastroenterology team suggested that patient’s findings mimic those found in GAVE syndrome. It was suggested to convert intravenous medications to enteral form via gastrostomy. The final plan was to repeat the endoscopic procedure, while proceeding with an endoscopic spray coagulation if bleeding persisted or recurred.
The patient achieved full recovery and received follow-up appointments with both teams. Enteral medications were continued at home, including proton pump inhibitor, H2-receptor antagonist and sucralfate. Three weeks later, a small bowel MRI was carried out and showed no pathological abnormality. Clinical improvement and satisfactory results were confirmed by surveillance OGD and consecutive follow-up.
Investigations
Haemoglobin on admission was 5.7 g/dL (11.0–14.5 g/dL), which improved to 8.4 g/dL, with haematocrit of 23.2% (33.9%–43.5%) and mean cell volume of 96.3 fL (76.7–89.2 fL). Renal profile, liver enzymes and coagulation screen were normal. Radiological investigations carried out, including USS, MS and MRI of the small bowel showing normal findings. OGD and colonoscopy with perianal EUA were performed.
Differential diagnosis
Maelena and/or haematemesis are suggestive of an upper GI bleed. In younger children, the likely aetiologies include stress ulcers and mucosa erosions while in older children, causes could be trauma (Mallory-Weiss tear), duodenal ulcers, oesophagitis or gastritis,4 rarely coagulation disorders and Dieulafoy lesions (DL).
Oesophagitis is inflammation of the lining of the squamous oesophageal epithelium, due to gastro-oesophageal reflux disease (regurgitated gastric and duodenal fluids include pepsin, trypsin and gastric acid), infections, eosinophilic oesophagitis, food allergies or iatrogenic causes, with the former three being the most common in the paediatric age group.
Peptic ulcers (PUs) are present as discontinuities in gastric or duodenal mucosal tissues, with penetration up to the muscularis mucosa.5 PU in children is uncommon with an annual incidence of 5.4/100 000.6 Primary ulcers, mostly found in the antrum or the lesser curve of the stomach, arise due to an increased production of hydrochloric acid and pepsin. Secondary ulcers are caused by pathological extragastric events such as drugs (non-steroidal anti-inflammatory drugs and immunosuppressive medications), stress or infections (Helicobacter pylori).7
DL is uncommon and represents 6% of fatal non-variceal bleeding of upper GI tract. This is a vascular abnormality, which consists of large calibre, tortuous and persistent submucosal artery. This artery protrudes through a small mucosal defect yet shows no other arterial wall abnormality. Ulcerations are limited to the overlying mucosa, with the surrounding mucosa appearing histologically normal.8
Treatment
The patient was initially transfused packed red blood cells. He was treated conservatively with intravenous medications consisting of proton pump inhibitor, H2-receptor blocker and oral sucralfate. Paediatric gastroenterology team advised establishing elemental hydrolysed feeding formula.
Outcome and follow-up
After full recovery, he was monitored thoroughly by both surgical and medical teams and underwent follow-up surveillance OGD, which noted remarkable clinical improvement. There was no indication for surgical antrectomy as the patient recovered well. He was kept on his regular medications along with proton pump inhibitors and H2-receptor antagonist.
Discussion
GAVE was initially described in 19539 as erosive atrophic gastritis with marked venocapillary ectasia. This rare pathology is an important cause of GI bleeding and anaemia in adults. GAVE is exceedingly rare in children, with few cases ever reported. GAVE presents with a variety of symptoms, ranging from occult bleeding to severe acute upper GI bleeding, requiring frequent blood transfusions. Early recognition and management are essential in all age groups and any delay in recognition can result in fatal life-threatening bleeding emergency. Most patients are transfusion-dependent despite being on oral iron supplementation.10
GAVE can be associated with other chronic illnesses, mainly liver conditions such as liver cirrhosis, present in 30% of cases,11 and connective tissue disorders.12
In non-cirrhotic patients, autoimmune diseases are present in up to 62% of patients, Raynaud’s phenomenon is common in up to 31% and sclerodactyly in up to 20%.3
At endoscopy, a pathognomonic pattern is characteristic: visible columns of red tortuous ectatic vessels all along antral longitudinal folds.13 This is visualised as red spots, which are either arranged in a diffuse pattern, described as ‘honeycomb stomach’ or organised in stripes, which are radially departing from the pylorus, defined as ‘watermelon stomach’.1 GAVE is usually found in the gastric antrum but can also be found in the cardia,14 duodenum, jejunum15 or rectum.16 In this case, both gastric antrum and pylorus had visible red tortuous vessels alternating with normal-looking mucosa. The duodenum was also noted to have patchy erythematous areas in keeping with duodenitis.
The pathophysiology of GAVE is poorly understood and various theories have been postulated, which include hypergastrinemia, low pepsinogen levels and achlorhydria. It is primarily thought to arise from a partial prolapse of gastric mucosa in the antrum, leading to intermittent obstruction of submucosal vessels leading to vascular ectasia. This theory is strengthened by histological features suggestive of hyperplasia of lamina propria and dilatation of mucosal capillaries, the latter due to occlusion by fibrin thrombi and occasionally associated with inflammation.17 The histological changes seen under microscopy include lamina propria fibromuscular proliferation and vascular dilatation (both of which are commonly associated with thrombosis), which are still poorly understood.18 In this case, the histological biopsies from the gastric mucosa revealed foveolar hyperplasia along with mild chronic inflammatory infiltrate. In 1989, a scoring system (0–4) was devised to score GAVE severity in adults.17 It incorporated the presence or absence of fibrin thrombi and/or vascular ectasia and spindle cell proliferation. This was updated in 20131 to also include fibrinhyalinosis (scoring system is now 0–5). Using the 2013 scoring system (since no system currently exists for paediatrics) (figure 4) would score 3 (2 for fibrin and vascular ectasia and 1 for increased spindle cell proliferation).
Increased levels of gastrin and prostaglandin E2-hormones (both are potent vasodilators) were observed in some patients and it was hypothesised that liver failure may lead to a build-up of these hormones and contribute to the final pathogenesis.19
The current available treatment options are highly controversial and indeed a plethora of medical, endoscopic and surgical techniques have been implicated. Patients are treated based on their symptoms; stable and asymptomatic patients treated with iron supplementation, whereas symptomatic patients are given blood products transfusions.
Multiple endoscopic therapies using neodymium-yttrium-aluminium garnet (Nd:YAG) laser, endoscopic band ligation or argon plasma electrocoagulation (APE) have been implicated (Nd: YAG laser and APE being the commonly used).20 Both APE and Nd:YAG laser are highly effective and equally efficient and show low levels of recurrence rates after minimal sessions. Nd:YAG laser coagulation has been successful in controlling GAVE-related bleeds and this has led to a reduction in the blood transfusion requirements. APE is a widely used technique due to its high rate of availability21 and reversible complications, including distension (due to gas) and GI pneumatosis. Other endoscopic options include use of nitrous oxide cryotherapy, laser photo ablation and coagulation with haemospray. The latter endoscopic treatment options carry potentially more serious side effects (perforation and gastric ulcers), resulting in higher rates of morbidity and mortality.21
In this case, since no active bleeding or ulceration was evident during gastroscopy and the patient’s haemoglobin level remained stable after red cell concentrate transfusion, the decision taken was to withhold argon plasma use and monitor response.
Surgical procedures such as total gastrectomy, subtotal gastrectomy with Biliroth II anastomosis with deployment of a portocaval shunt have limited success.
In this case, dual endoscopy was suggested by the paediatric team in view of the presenting symptoms. Colonoscopy revealed no active bleeding areas but noted colitis of the distal part of his colon. Gastroscopy revealed streak-like gastritis and duodenitis with no active bleeding. In view of the latter, his medical treatment was continued as there was no indication for argon laser treatment.
The case was discussed during a paediatric multidisciplinary team meeting and a possible diagnosis of GAVE was suggested. A novel management of changing his feeds was implemented, leading to clinical and symptomatic improvement. No further episodes of maelena were recorded, with no further need for blood transfusions.22 23
GAVE is a condition in the spectrum leading to upper GI vascular ectasia in children. It is a challenging cause of upper GI bleeding, which needs careful individual management. Careful exploration of all current available methods is ideal for better optimisation of treatment.
Learning points
-
In this case, the patient was noted to be stable yet found to be anaemic. Therefore, he was initially stabilised and successfully treated conservatively.
-
Since no active bleeding was visible at endoscopy, medical treatment was instituted with no need for coagulation endoscopic therapy as addressed in the literature.
-
A different management approach was taken whereby a change in the patient’s dietary requirement was made: a change to an elemental diet, which resulted in a significant clinical and symptomatic improvement.
-
Even though the pathogenesis remains quite unclear, it is a recognised cause of occult upper gastrointestinal bleeding and therefore high index of suspicion and careful monitoring should take place.
Footnotes
-
Contributors MP was involved in the case summary, background information, case presentation, differential diagnosis, treatment, follow-up and a small part of the discussion. SE was involved in the write-up of background information, case presentation, investigations, differential diagnosis, treatment, follow-up and a small part of the discussion. MS was involved in all aspects, constantly adjusting all sections of the paper as well as contributing heavily towards the discussion aspect. MP and SE were both involved in all sections and liaised regularly with MS.
-
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
-
Competing interests None declared.
-
Patient consent for publication Parental/guardian consent obtained.
-
Provenance and peer review Not commissioned; externally peer-reviewed.
- © BMJ Publishing Group Limited 2020. No commercial re-use. See rights and permissions. Published by BMJ.
References
Use of this content is subject to our disclaimer