Portal de Periódicos da CAPES

Vomiting-Induced Hematemesis in Children: Mallory–Weiss Tear or Prolapse Gastropathy?

2000; Lippincott Williams & Wilkins; Volume: 30; Issue: 4 Linguagem: Inglês

10.1097/00005176-200004000-00017

1536-4801

Phyllis R. Bishop, Michael J. Nowicki, Paul H. Parker,

Potassium and Related Disorders

Hematemesis in children provokes anxiety both in parents and healthcare providers. In children its cause is somewhat age related. Hematemesis after forceful vomiting in children is often attributed to a Mallory–Weiss tear (MWT); however, this is an uncommon finding in children, with fewer than 25 cases reported (1–8). A second postemetic condition leading to hematemesis is prolapse gastropathy syndrome (PGS), characterized by gastric mucosal hemorrhage, typically limited to a small area in the proximal stomach. Since the first description in 1972 (9), fewer than 50 cases have been reported (10–15). Although MWT is a well-known cause of hematemesis, PGS remains a poorly recognized entity. It is not mentioned in often-cited textbooks of adult (16,17) and pediatric (18,19) gastroenterology. Reports of large series of adults (20) and children (21–25) with hematemesis fail to mention PGS as an endoscopic finding. Although articles mention PGS as a cause of hematemesis in children (26,27), to date, there has been no report describing the clinical features and endoscopic findings of PGS in children. We report seven children with diagnosed PGS. Further, we review the role of MWT and PGS in vomiting-induced hematemesis in children. METHODS AND PATIENTS We retrospectively reviewed findings in all diagnostic upper endoscopies performed in children at the University of Mississippi Medical Center from August 1998 through July 1999. During this period, 224 diagnostic upper endoscopies were performed, 27 for hematemesis. The endoscopic findings in patients with hematemesis are listed in Table 1. No patient had an MWT, whereas in seven, PGS was identified. Overall, PGS was found in 3.1% of all patients who underwent diagnostic endoscopy, which compares favorably with the frequency of PGS in previously reported adult series (2–5.5%) (9,14,28). More impressive is that PGS accounted for 25.9% of patients with hematemesis in our series. The clinical features of patients with PGS are discussed in this report.TABLE 1: Endoscopic findings in children with a history of hematemesisCASE REPORTS Patient 1 A 13 year-old boy with a lifelong history of abdominal pain was referred for further evaluation. He reported intermittent pain, which typically lasted for hours, localized to the epigastric region. Vomiting occurred infrequently. There were no identifiable precipitating factors. There was no associated nocturnal awakening, diarrhea, constipation, weight loss, or joint symptoms. There was no history of hematemesis, coffee ground emesis, melena, or hematochezia. The pain persisted despite treatment with an H2 blocker for 2 years. There was no history of ethanol, aspirin, or nonsteroidal antiinflammatory drug (NSAID) use. There were no underlying medical conditions. Family history was significant for a mother with Helicobacter pylori– positive peptic ulcer disease. Abdominal examination revealed normal bowel sounds, no hepatosplenomegaly, and no tenderness. Rectal examination was positive for occult blood. Hemoglobin (13.5 g/dl), mean corpuscular volume (MCV; 81.7 fl), and platelet count (202 × 103/μl) were normal. Upper endoscopy revealed grossly normal-appearing mucosa in the second and third portions of the duodenum. The duodenal bulb had a discrete area of nodular mass-like lesion just distal to the antrum and extending approximately 180° around the bulb. Analysis of biopsy specimens revealed gastric heterotopia. The antrum had grossly normal-appearing mucosa. Specimen analysis by both rapid urease test and histology was negative for H. pylori. A discrete area of erythema and friability was seen on the lesser curvature of the stomach, just distal to the lower esophageal sphincter (Fig. 1). The esophagus had grossly normal-appearing mucosa of the upper and midesophagus, and the lower esophagus showed thickening, ulceration, and friability. Specimen analysis results were consistent with reflux esophagitis. During the endoscopic examination, gastric prolapse was seen repeatedly during retching (Fig. 2).FIG. 1.: Endoscopic picture (retroflexed view) obtained in patient 1 revealing a well-demarcated area of edematous, erythematous mucosa on the lesser curvature of the stomach. Arrowheads demarcate the borders of the prolapse gastropathy; an asterisk demarcates the endoscope.FIG. 2.: Endoscopic picture obtained in patient 1 revealing the prolapse of the stomach into the esophagus during retching.Patient 2 A 17-year-old boy with a previous history of duodenal ulcer (H. pylori–positive) was admitted for evaluation of epigastric pain and hematemesis. The patient had been asymptomatic while taking ranitidine for 6 months after treatment for duodenal ulcer (metronidazole, amoxicillin, bismuth sulfate, and ranitidine). He had acute onset of epigastric pain associated with vomiting. After multiple episodes of vomiting, the emesis contained streaks of red blood without coffee ground material. There was no history of ethanol, aspirin, or NSAID use. There were no underlying medical conditions. Abdominal examination showed normal bowel sounds; there was no tenderness, hepatosplenomegaly, or mass. Hemoglobin (13.9 g/dl), MCV (86.8 fl), and platelet count (151 × 103/μl) were normal. The time between the last episode of hematemesis and upper endoscopy was less than 12 hours. Upper endoscopy showed a normal esophagus, antrum, and duodenum. The only finding was a discrete area of erythema and friability on the upper lesser curvature. There was no frank bleeding at the time of the endoscopy. Specimens from the antrum, analyzed by rapid urease test and histology, were negative for H. pylori. Biopsy of the lesion on the lesser curvature showed acute, focal hemorrhage, without an inflammatory component. Patient 3 A 10-year-old boy with a 5-year history of recurrent abdominal pain was referred for further evaluation. He reported periumbilical pain that occurred 3 to 4 days per week and lasted from minutes to hours. No precipitating factors were identified. The pain was associated with vomiting that occurred one to two times each week. Nocturnal awakening occurred infrequently. There was no response to H2 blocker. Symptoms worsened 2 weeks before evaluation with the development of hematemesis, described as "bright red blood." There was no history of aspirin or NSAID use. There were no underlying medical conditions. Abdominal examination showed no tenderness, organomegaly, or masses. Rectal examination was not performed. Hemoglobin (12.7 g/dl), MCV (84.6 fl), and platelet count (269 × 103/μl) were normal. The time between last episode of hematemesis and upper endoscopy was 7 days. Upper endoscopy showed normal esophagus, antrum, and duodenum, with the only abnormality being a discrete area of erythema and friability on the lesser curvature, just distal to the lower esophageal sphincter. There was no frank bleeding at the time of the endoscopy. Patient 4 A 10-year-old boy with a history of cyclic vomiting reported vomiting for 36 hours. Hematemesis developed after 24 hours of vomiting, consisting of streaks of red blood and brown flecks. He had never vomited blood with previous episodes of cyclic vomiting. There was no history of hematochezia, melena, or epistaxis. He was taking no medications, including NSAIDs and aspirin. Abdominal examination showed diffuse discomfort to deep palpation, but there was no hepatosplenomegaly and no masses were appreciated. Rectal examination showed soft, brown stool that was negative for occult blood. Hemoglobin (14.2 g/dl), MCV (83.5 fl), and platelet count (289 × 103/μl) were normal. The time between the last episode of hematemesis and upper endoscopy was less than 24 hours. Upper endoscopy to the third portion of the duodenum revealed normal-appearing mucosa throughout, except for a discrete area of erythema and friability on the lesser curvature of the stomach, just distal to the esophagogastric junction. Vomiting ceased after rehydration and treatment with intravenous ondansetron (Zofran; GlaxoWellcome, Research Triangle Park, NC, U.S.A.). Hematemesis did not recur after the endoscopy. Patient 5 A 10-year-old boy with a history of recurrent abdominal pain and gastroesophageal reflux had vomited for 2.5 days. Hematemesis developed after 24 hours, consisting of "red blood and clots." He had never vomited blood in the past. There was no history of hematochezia, melena, or epistaxis. His only medication was ranitidine; he had taken no NSAIDs or aspirin. Abdominal examination showed no discomfort to deep palpation. There was no hepatosplenomegaly and no masses were appreciated. Rectal examination revealed soft, brown stool that was negative for occult blood. The hemoglobin (12.1 g/dl), MCV (85.0 fl), and platelet count (220 × 103/μl) were normal. The time between the last episode of hematemesis and upper endoscopy was 24 hours. At upper endoscopy, the esophageal mucosa was seen to be thickened with decreased vascular markings. Analysis of esophageal tissue specimens confirmed esophagitis. During endoscopy, coffee ground material was seen in the dependent portion of the stomach. The stomach mucosa was normal throughout except for a discrete area of erythema and friability on the lesser curvature of the stomach, which was oozing blood. Biopsy specimens were not obtained from the lesion in the cardia. Results of rapid urease test and histologic examination of antral specimens were negative for H. pylori. An erosion was seen in the duodenal bulb, and tissue specimens showed chronic duodenitis with gastric metaplasia. Treatment was initiated with intravenous ranitidine, which was converted to oral omeprazole at discharge from the hospital. Hematemesis did not recur after endoscopy. Patient 6 An 8-year-old boy with a history of abdominal pain for 12 months that was unresponsive to H2-blocker was referred to pediatric gastroenterology after development of hematemesis. Episodes of vomiting bright read blood occurred intermittently for 3 months. There was no history of hematochezia, melena, or epistaxis. The only medication was ranitidine; no use of NSAIDs or aspirin was admitted. Family history was significant for a mother with H. pylori–positive peptic ulcer disease. Abdominal examination showed no discomfort to deep palpation; no hepatosplenomegaly or masses were appreciated. Rectal examination revealed soft stool that was positive for occult blood. Hemoglobin (13.8 g/dl), MCV (82.1 fl), and platelet count (311 × 103/μl) were normal. The patient vomited blood just before endoscopy. At endoscopy the esophagus appeared normal. Coffee ground material was noted in the stomach. The stomach showed antral erythema, without ulceration, erosion, or nodularity. On retroflexion, a discrete area of erythema and friability was seen on the lesser curvature of the stomach; blood was oozing from this area. Biopsy specimens were not obtained from the lesion in the cardia; specimens of the antrum were consistent with chronic, active gastritis. Results of investigation for H. pylori by a rapid urease test and histological examination were negative. Patient 7 A 16-year-old boy had had nausea, epigastric pain, and vomiting for 10 days. Hematemesis, consisting of coffee ground material, developed after 3 days of vomiting. The patient reported a decreased appetite and a 10-lb weight loss. There was no history of nocturnal awakening, hematochezia, melena, epistaxis, or use of NSAIDs and aspirin. He reported alcohol use on weekends but denied that there had been any recent binges. A similar episode 9 months earlier was treated by his primary care physician with lansoprazole. Family history was negative for peptic ulcer disease. The abdominal examination revealed epigastric tenderness; there was no hepatosplenomegaly, and no masses were appreciated. Rectal examination revealed soft, brown stool that was negative for occult blood. Hemoglobin (13.4 g/dl), MCV (82.7 fl), and platelet count (294 × 103/μl) were normal. The time between the last episode of hematemesis and upper endoscopy was 2 days. At endoscopy, the esophageal mucosa was observed to be thickened, with loss of vascular markings; analysis of biopsy specimens confirmed esophagitis. Coffee ground material was noted during endoscopy. The gastric mucosa was normal throughout except for a discrete area of erythema and friability on the lesser curvature of the stomach, which was oozing blood (Fig. 3). Biopsy specimens were not obtained from the lesion in the cardia. Rapid urease test results and histologic examination of specimens of the antrum were negative for H. pylori. The duodenum was normal in appearance.FIG. 3.: Endoscopic picture (retroflexed view) obtained in of patient 7 revealing a well-demarcated area of edematous, erythematous mucosa on the lesser curvature of the stomach. Arrowheads demarcate the borders of the prolapse gastropathy; an asterisk demarcates the endoscope.DISCUSSION In children, hematemesis after vomiting and/or retching is often attributed to an MWT; however, in our experience, PGS is a far more common finding. We retrospectively reviewed all findings in diagnostic upper endoscopies performed in children at the University of Mississippi Medical Center from August 1998 through July 1999. During this period, 224 diagnostic endoscopies were performed, 27 for hematemesis. No patient had an MWT, but seven were identified with PGS. Overall, PGS was found in 3.1% of all children undergoing a diagnostic endoscopy and in 25.9% of children with hematemesis. Mallory–Weiss tear is often cited as the cause of hematemesis in children, although it is a rarely reported condition in this age group. Since the original description by Mallory and Weiss in 1929 (29), less than 25 cases have been reported in children (1–8). This condition may be underreported for one or more reasons. Because MWT heals rapidly, delaying endoscopy may result in missing the lesion. Possibly, MWT has been underreported because of the relatively recent advent of endoscopy in children. Alternatively, MWT may be commonly seen, but unreported. It is also possible that MWT is truly a rare condition in children. Combining our series with five previously reported series evaluating the cause of hematemesis–upper gastrointestinal bleeding, not a single case of MWT was reported in nearly 200 children (21–25). This contrasts with the experience in adults. In nearly 4000 upper endoscopies performed for hematemesis in adults, MWTs were the cause in 4.2% to 5.5%(20,30). A more common cause of hematemesis in our series was PGS. It was first described in 1972 as a second postemetic condition causing hematemesis (9). Since the original description, less than 50 cases have been reported, all but 1 in adults (10–15). To date, there has been no characterization of PGS in children. Ament and Christie (22) were the first to mention this condition in a child, a patient with "submucosal hemorrhage caused by retching," in a series of 38 children evaluated by endoscopy for upper gastrointestinal bleeding. A second child was included in a series of patients with PGS reported by Byfield et al. (15) Potential reasons for the paucity of reports of PGS in children are underreporting, failure to recognize the condition, or rarity of the condition. In five previously reported series in which the cause of hematemesis–upper gastrointestinal bleeding was evaluated in 171 children, only a single case of PGS was reported (22). The experience in our series suggests that PGS is a common finding in children with hematemesis–upper gastrointestinal bleeding. We specifically evaluated the cardia and esophagogastric junction by retroflexion during all upper endoscopies, particularly in those with hematemesis–upper gastrointestinal bleeding. Differentiation between MWT and PGS by clinical features may be difficult. Comparing our series of children with PGS (Table 2) with previously reported children with MWT, we found the following. Use of NSAIDs, ethanol, or corticosteroids is typically not present in either condition. Both conditions most commonly manifest with hematemesis after vomiting, although blood may be present in the initial vomiting episode. Less commonly, hematemesis is absent; evidence of upper gastrointestinal bleeding manifested in melena or guaiac–positive stools. The character of hematemesis is similar in PGS and MWT, including bright red blood, coffee ground material, and, rarely, clots. Bleeding tends to be more significant in children with MWT and is more often associated with low hemoglobin and more necessity for transfusion (≥50%) than in children with PGS. These findings are similar to those reported in adults, in whom bleeding in PGS tends to be minimal and transfusion is uncommon (14%) (10–15), whereas in adults with MWT, transfusion is required in 60% to 75%(31–33) and large (≥4 units) blood transfusions are required in more than 25% of patients (30). Surgical control of bleeding from MWT in adults varies from 8% to 40%(30–32).TABLE 2: Clinical characteristics of 7 male children with prolapse gastropathy syndromeClinical features of PGS in children are similar to those previously reported in adults with PGS. The most common indication for upper endoscopy in adults with PGS is a history of vomiting or retching, with the majority (65%) of patients experiencing hematemesis (9–15). However, PGS may be present without hematemesis. Epigastric pain is the second most common indication for endoscopy in adults with identified PGS (14). In our series, hematemesis was the indication for endoscopy in six (86%) of seven children. One patient had a history of abdominal pain and guaiac–positive stools. Potential contributing factors to hematemesis—NSAID use and ethanol ingestion—were seen in 46% of reported adult cases. None of our patients had a history of NSAID use and only one admitted alcohol use. None of the children reported had placement of a nasogastric tube for diagnostic lavage, eliminating tube trauma as contributing to hematemesis. Any condition causing retching or vomiting can lead to PGS. In our series, three of seven patients had a history of gastrointestinal disease before endoscopy, and four of seven had endoscopic and histologic evidence of inflammation, in addition to PGS (Table 2). The typical endoscopic appearance of PGS is a focal patch of well-demarcated, erythematous, congested mucosa—sometimes with a reticulated-mosaic pattern reminiscent of hypertensive gastropathy (15). Superficial ulcerations have been reported with PGS, including the original description by Delumeau et al. (9) The presence of ulceration is not a prerequisite for bleeding. Oozing may be seen from traumatized, albeit intact, mucosa. Prolapse gastropathy is located in the upper stomach, just distal to the esophagogastric junction. In previously reported adults, when stated, the location of the gastropathy in decreasing order of frequency was the greater curvature (46%), lesser curvature (43%), anterior wall (7%), and circumferential (4%) (10,12,14). Endoscopic features of the patients we report are identical with those reported in adults; however, all lesions were noted on the lesser curvature. Histologic features of PGS include acute inflammation, chronic inflammation, submucosal hemorrhage, and superficial ulceration, either alone or in combination (14,15). The single lesion sampled in our series showed acute, focal hemorrhage without inflammatory changes. Why children rarely have MWT and appear to have PGS fairly commonly is uncertain. It has been suggested that tonicity of the lower esophageal sphincter plays a role in determining the development of PGS or MWT. A lax lower esophageal sphincter favors the development of PGS by allowing the stomach to prolapse into the lower esophagus during vomiting and/or retching, leading to mechanical trauma (13). Trauma leads to focal hemorrhage and the potential for mucosal inflammation, hematemesis, and/or ulceration. An MWT develops in the setting of high intragastric pressure and a tight lower esophageal sphincter (13,34,35). It has been suggested that children are less averse to vomiting (2) and therefore do not generate the same high intragastric pressure seen with retching in adults who are resisting vomiting. Also, children are less apt to have contributing factors, such as alcoholism, tasks involving heavy lifting, and childbirth (36). Others have attributed the increased risk of MWT in adults to the presence of hiatal hernia, foreshortening of the esophagus, diaphragmatic abnormalities, or abnormal attachments of the phrenoesophageal ligament (37,38). Protection may be afforded the pediatric patient by the greater tensile strength of the gastrointestinal tract in children compared with that in adults (39,40). In summary, these data suggest that PGS is a relatively common finding in children with hematemesis, whereas MWT remains an uncommon finding. Clinical features are similar in PGS and MWT; however, bleeding appears to be less severe in the former condition. The endoscopic features of PGS in children are similar to those described in adult patients. In our experience, PGS appears to be a benign condition with minimal bleeding. Any condition leading to vomiting and/or retching can lead to PGS. Endoscopy often reveals disease that is amenable to medical therapy. The reasons for the commonality of PGS and rarity of MWT in children remain speculative. Other large series are needed to determine whether our results can be generalized to children in other parts of the world.