Phlegmonous gastritis (PG) is a rare bacterial infectious disease characterized by neutrophil-based purulent inflammation of the gastric wall. The most representative causative bacterium is Streptococcus pyogenes, followed by Staphylococcus, Pneumococcus and Enterococcus. Hepatic portal venous gas (HPVG) is considered a potentially fatal condition and is rarely associated with PG.

The white blood cell count of a 70-year-old woman with acute lymphocytic leukemia in complete remission dropped to 100/μL after consolidation chemotherapy. Her vital signs were consistent with septic shock. Venous blood culture revealed the presence of Bacillus cereus. Abdominal computed tomography (CT) and esophagogastroduodenoscopy (EGD) showed marked thickening of the gastric wall. As with the other findings, CT was suggestive of HPVG, and EGD showed pseudomembrane-like tissue covering the superficial mucosa. Histopathological examination of gastric biopsy specimens showed mostly necrotic tissue with lymphocytes rather than neutrophils. Culture of gastric specimens revealed the presence of Bacillus cereus. We finally diagnosed this case as PG with Bacillus cereus-induced sepsis and HPVG. This patient recovered successfully with conservative treatment, chiefly by using carbapenem antibiotics.

The histopathological finding of this gastric biopsy specimen should be called "neutropenic necrotizing gastritis".

Phlegmonous gastritis (PG) is a rare bacterial infectious disease characterized by neutrophil-based purulent inflammation of the gastric wall starting from the submucosa and muscularis layers[1]. In approximately 70% of causes, the causative bacterium is Streptococcus pyogenes, and other causative bacteria include Staphylococcus, Pneumococcus and Enterococcus[2]. Alcoholism, mucosal injury, gastric hemorrhage and immunosuppression have been reported to be risk factors for PG[2,3]. PG is regarded as a disease with a high mortality rate of approximately 40%, even if treated in a timely manner[3,4]. There are only a few reports on PG with hematological malignancies[5,6]. These patients developed PG after having gastric lymphoma or extramedullary leukemia in the stomach, and all died early. On the other hand, hepatic portal venous gas (HPVG) is also still considered a potentially fatal condition that can occur during intestinal ischemia or necrosis and is rarely associated with PG[7,8].

Previously, we reported a case of PG with Bacillus cereus-induced sepsis in which the neutrophil count was markedly reduced during chemotherapy for acute lymphocytic leukemia (ALL)[9]. In this study, we report a case of a similar pathological condition accompanied by HPVG followed by multiple liver abscesses.

Histopathological analysis of gastric biopsy specimens showed mostly necrotic tissue with fibrin precipitation and partial infiltration of inflammatory cells by lymphocytes rather than neutrophils (Figure 5A). It was unique findings that should be called "neutropenic necrotizing gastritis". Bacilli showing positive Gram staining were also observed (Figure 5B). Culture of gastric specimens later proved that this bacterial group was Bacillus cereus. We finally diagnosed this case as PG with Bacillus cereus-induced sepsis and HPVG followed by multiple liver abscesses. The causative bacteria of the liver abscesses were not investigated.

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Histopathology of the gastric biopsy. A: Mostly necrotic tissue and partly infiltration by lymphocytes were observed (hematoxylin & eosin stain); B: Bacilli were gram-positive (gram stain).

The clinical course of this patient is shown in Figure 1. After day 11, while use of G-CSF was continued for severe infection, administration of carbapenem antibiotics (initially biapenem 0.3 g, 2 times/d and then meropenem 0.5 g, 2 times/d) was used in combination with intravenous immunoglobulin (5 g/d) for 3 d. In addition to antibiotics, septic shock was treated by administration of hydrocortisone and fluid. Red blood cells and platelets were transfused once each. No invasive treatment was administered for liver abscess.

This patient recovered successfully with conservative treatment, chiefly administration of antibiotics, without any sequelae. She then received intrathecal injection twice and repeated consolidation therapy for ALL with blinatumomab (a CD19/CD3 bispecific T-cell engager) for 3 cycles. She has no recurrence of PG more than one year and is maintaining CR for ALL. Currently, she is being followed-up with no treatment as an outpatient.

Our patient with ALL developed PG due to Bacillus cereus infection during the neutrophil depletion period after consolidation chemotherapy and was temporarily in a fatal state of septic shock. We reported a similar case in 2012 in which severe inflammation, such as an increase in CRP levels to 52.97 mg/dL, was observed, and we were able to save the lives of patient at that time[9] (since then, the patient has been able to lead a completely normal daily life for more than 11 years). Regarding the current patient, on the day that WBC dropped to 100/μL, the patient complained of gastrointestinal symptoms as well as worsening vital signs such as fever, decreased blood pressure, tachycardia and tachypnea, and we strongly suspected PG because CT showed marked thickening of the gastric wall. Based on our previous experience, we started administering carbapenem antibiotics immediately, before the causative organism was identified; the initial response was effective, and we were able to save this patient’s life.

In general, hematological malignancies tend to be accompanied by a decrease in immunocompetence, and myelosuppression by the administration of anticancer drugs further promotes immunodeficiency. There have been 6 reports of patients with hematological malignancies who developed PG during the neutropenic phase after the administration of anticancer drugs, including 1 of our previous patients[9-14]. Table 2 summarizes the clinical features of a total of 7 patients, including the patient experienced this time. Notably, Enterococcus, which is one of the representative causative bacterial species of PG, was detected in only 2 cases, whereas Bacillus, which has rarely been reported as a causative agent of PG, was detected in 4 cases, including our 2 cases. Neutrophil-based purulent inflammation of the gastric wall is considered to be the main cause of PG, whereas in 5 of the 7 cases, the neutrophil count was reduced to less than 100/μL. It was presumed that there is a difference in the onset process as a bacterial infectious disease. Of 7 cases, only our two cases were histopathologically examined with gastric biopsy, and unlike typical PG findings, most were covered with necrotic tissue, in which gram-positive bacilli were found. Infiltration of inflammatory cells by lymphocytes but not neutrophils was partly observed. This finding should be called "neutropenic necrotizing gastritis", which has never been proposed, rather than the typical PG. Bacillus, one of the residents of the bacterial flora in the small intestine, infected the gastric wall under immunosuppression and caused PG in this patient. Bacillus cereus produces various toxins[15], including necrotizing enterotoxin which may have been present in our two patients.

Gastric acid secretion inhibitors were used in 5 cases (two cases were not described). Recently, it was reported that the risk of intestinal infectious diseases such as Clostridium difficile is increased by disruption of the function of the barrier that prevents the invasion of pathogens into the gastrointestinal tract by a gastric acid secretion inhibitor[16]. It was reported that vonoprazan, a potassium-competitive acid blocker used in this patient, showed a stronger gastric acid secretion inhibitory effect than proton pump inhibitor and induced a stronger change in the composition of the intestinal flora[17]. Shi et al[14] expressed concern that the use of gastric acid secretion inhibitors increases the bacterial abundance in the gastric juice and thus may result in risk for the development of PG. Currently, there are few patients with PG with or without hematological malignancies, and scientific evidence is limited. In the future, a large number of patients will need to be aggregated for further study.

This patient also developed HPVG and liver abscess. HPVG was previously a sign of poor prognosis, but in recent years, the mortality rate of patients with HPVG detected by CT has been lower (29%) than that previously reported, and HPVG alone cannot predict prognosis[18]. In our case, a small amount of HPVG was found in the initial CT that suggested PG, and we were able to conservatively respond before it spread. In addition, formation of liver abscess could be followed by CT from the beginning. We have not been able to identify the bacterial species responsible for the liver abscess; however, we believe that the causative agent was Bacillus cereus, which spread through the portal vein. Furthermore the liver abscesses could also be conservatively cured without drainage.

We reported a case of PG due to Bacillus cereus infection during the neutropenic phase after the administration of anticancer drugs for ALL. We analyzed 7 cases with similar pathological conditions reported previously, including the 2 cases we experienced. The histopathological findings in our two patients should be called "neutropenic necrotizing gastritis", which has never been proposed.