Nonocclusive Mesenteric Ischemia after Chemotherapy in an Adolescent Patient with a History of Three Allogeneic Hematopoietic Stem Cell Transplantations for Acute Lymphoblastic Leukemia

Nonocclusive mesenteric ischemia (NOMI) is induced by intestinal vasospasm without thromboembolic occlusion. The typical patient with NOMI is critically ill with severe cardiac disease or sepsis while undergoing inotropic support.^,  Thus, NOMI is associated with high morbidity and mortality rates. The estimated overall incidence of autopsy-verified fatal NOMI is 2.0 cases/100,000 person-years, although no pediatric or adolescent cases have yet been reported. In this paper, we report a case of NOMI occurring in an adolescent girl after being administered chemotherapy for a third hematological relapse of acute lymphoblastic leukemia (ALL).

An 18-year-old woman had been diagnosed with B-cell precursor ALL in March 2005 at the age of 10 years. Because of marked leukocytosis (191 × 10⁹/L) and a poor response to prednisolone monotherapy for 7 days, we selected the L04-16 protocol of the Tokyo Children's Cancer Study Group to estimate the risk associated with allogeneic hematopoietic stem cell transplantation (allo-HSCT). In August 2005, she achieved complete remission after induction therapy and then received an allogeneic bone marrow transplant from her human leukocyte antigen (HLA)-matched mother; she had been administered a preparative regimen comprising 8-Gy of total body irradiation, cyclophosphamide at 120 mg/kg, and fludarabine at 150 mg/m². In July 2006, she experienced her first relapse. Chemotherapy and donor lymphocyte infusion were administered, but the patient failed to achieve complete remission. In July 2007, she received an allogeneic peripheral blood stem cell transplant from her mother, after conditioning with cytarabine at 10 g/m², idarubicin at 36 mg/m², cyclophosphamide at 120 mg/kg, and fludarabine at 150 mg/m². However, she experienced a second relapse in October 2008. In December 2008, she received a third allo-HSCT a transplant from a sibling with one HLA allele mismatch during the active disease, after she had been conditioned with busulfan at 12.8 mg/kg, cyclophosphamide at 120 mg/kg, and fludarabine at 150 mg/m². In September 2009, she developed type 2 diabetes mellitus. After her third relapse in August 2012, she received a combination therapy of prednisolone, L-asparaginase, vincristine, and bortezomib, beginning on September 19, 2012 (Day 1, Figure 1). Physical examination and laboratory findings—which included blood, urine, echocardiography, thoracoabdominal contrast-enhanced computed tomography (CT), and brain magnetic resonance imaging—were normal before chemotherapy with the exception of glucose intolerance and abnormal bone marrow examination results. She received a continuous infusion of insulin after starting chemotherapy. During chemotherapy, only mild neutropenia (500‒1000/μL) was present and her general status was relatively good. After the initiation of chemotherapy, the blast count in peripheral blood decreased and disappeared by Day 16. Because of the continuous drip infusion before beginning chemotherapy, she did not develop severe dehydration. However, she often required furosemide to maintain sufficient urine flow during chemotherapy. She complained of severe abdominal pain on Day 16. Contrast-enhanced CT showed thickening of the bowel wall, decreased bowel wall enhancement by contrast media, an elevated CT value around the ascending colon, and accumulation of ascitic fluid (Figure 2A). No obvious thromboembolic occlusion of the mesenteric arteries was evident. An exploratory laparotomy revealed extreme edema of the bowel wall from the ileum to the ascending colon. Pulsation of the ileocolic, right colic, and middle colic arteries was clearly palpable. No bacterial pathogens were detected by cultures of feces and ascitic fluid. Because no macroscopically abnormal intestinal lesions were detected, we decided only to insert an indwelling drainage catheter. However, because her abdominal pain persisted, we again performed abdominal contrast-enhanced CT on Day 27, which revealed thinning of the large bowel without enhancement of the mucosa (Figure 2B). These findings strongly suggested necrosis of the ascending colon; therefore, we performed a right-sided hemicolectomy. Surgical findings confirmed necrosis from the ileum to the ascending colon (Figure 3A). Blood flow within the ileocolic artery that supplied the necrotic colon was confirmed. The colonic mucosa adjacent to the necrotic lesions was characterized by decreased crypts with reduced goblet cells. Furthermore, bleeding to the lamina propria, congestion, edema, and ectasia of the submucosa were present. These findings indicated chronic ischemia of the colon (Figure 3B). There were no findings of vascular lesions, leukemia, arteriosclerosis, or graft-versus-host disease (GVHD). We diagnosed NOMI because there were no findings of vascular lesions, invasion of leukemia cells, or GVHD by contrast-enhanced CT, pathological examination, or surgical findings. No further abdominal complications occurred after restarting chemotherapy. In April 2013, she received unrelated cord blood transplantation. In June 2013, she died of invasive aspergillosis.

In this paper, we reported a case of an adolescent patient who developed NOMI after chemotherapy for a third hematological relapse of ALL. It was possible that ischemia, infection, hematoma, the use of chemotherapeutic agents, and/or GVHD may have caused colonic complications in this patient. Infection, hematoma, and GVHD were excluded by blood tests, imaging, bacterial culture, and pathological findings. Because there have been no reports of necrosis of the large sections of the colon after the administration of chemotherapeutic agents, we excluded chemotherapeutic agents as the cause of colonic necrosis in our patient. Because no arterial thromboembolic occlusion was observed by contrast-enhanced CT imaging, surgical findings, or pathological examination, we diagnosed NOMI in the right colon. The right side of the colon appears particularly sensitive to NOMI because this site frequently lacks a well-developed and consistent marginal collateral vascular network. In addition, right colonic involvement is associated with a particularly poor prognosis.^, 

A large proportion of patients with NOMI reportedly develop severe complications such as congestive heart failure, sepsis, myocardial infarction, and renal/hepatic disease. The use of digitalis, emulgents, and beta-blockers is also a risk factor for NOMI. Because our patient received furosemide to maintain sufficient urine flow during chemotherapy, we could not exclude the possibility that this emulgent possibly triggered NOMI. Tichelli et al report that long-term survivors after allo-HSCT were likely to have an increased risk of premature cardiovascular accidents. Therefore, dysfunction of the mesenteric arteries caused by multiple allo-HSCT may have a role in the development of NOMI.

Mitsuyoshi et al and Kamimura et al report that early treatment with prostaglandin E1 can effectively increase survival in patients with NOMI. Therefore, if our patient had been diagnosed with NOMI at the onset of abdominal pain and received prostaglandin E1 therapy, the development of colonic necrosis could have been avoided. However, NOMI is difficult to diagnose in the early phase because it lacks clear clinical symptoms. Selective digital subtraction angiography of the mesenteric arteries is the gold standard for the diagnosis of NOMI. Multidetector CT is also capable of depicting the condition of the arteries in sufficient detail to enable an early diagnosis of NOMI.^,  In retrospect, we should have performed digital subtraction angiography and/or multidetector CT to diagnose NOMI because thromboembolic occlusion of the mesenteric arteries was not detected at the onset of abdominal pain.

In conclusion, NOMI should be considered during differential diagnosis of intestinal complications, even in pediatric and adolescent patients when there is no apparent thromboembolic occlusion of the mesenteric arteries.

The authors have no conflicts of interest to declare.