A 78-year-old woman who had undergone a total arch replacement six years earlier arrived at a Japanese hospital vomiting blood from a tear between her aorta and esophagus, a complication so lethal that reported mortality ranges from 33 percent to nearly 100 percent. Surgeons stabilized her with a stent graft, removed the infected esophagus two days later, and rebuilt her digestive tract 69 days after that. She went home 94 days after the first operation and, seven months after the stent procedure, remains well with no evidence of recurrent infection or bleeding. The case, published in General Thoracic and Cardiovascular Surgery Cases, describes a staged, graft-preserving strategy that the authors say may offer a reasonable path to survival for carefully selected high-risk patients.
The condition, known as a secondary aortoesophageal fistula, arises when the aorta and the esophagus, which lie close together in the chest, erode into one another after thoracic aortic surgery. The authors report that the incidence of secondary fistulas after arch replacement or thoracic endovascular aortic repair is approximately 1 to 2 percent, with mortality ranging from 33 percent to nearly 100 percent. Conservative treatment without surgery is uniformly fatal. The mechanism of fistula formation remains poorly understood, which makes prevention difficult and places the entire burden of survival on rapid recognition and coordinated treatment.
When the woman arrived, she was hemodynamically stable but severely malnourished, with a body mass index of 15.4. Laboratory tests showed a hemoglobin level of 8.9 grams per deciliter, a white blood cell count of 13,860 per microliter, a C-reactive protein level of 17.99 milligrams per deciliter, and a procalcitonin level of 4.49 nanograms per milliliter, together with a blood urea nitrogen level of 29.4 milligrams per deciliter and a serum creatinine level of 1.18 milligrams per deciliter. Blood cultures were negative. An upper gastrointestinal endoscopy performed at the referring hospital had already revealed a deep ulcerative lesion in the mid-esophagus with necrotic debris, a protruding blood clot, and exposed underlying tissue that the endoscopist judged highly suggestive of a communication with the aorta and impending active hemorrhage.
Contrast-enhanced computed tomography at the receiving hospital showed gas and fluid collection surrounding the descending aorta, with loss of the fat plane that normally separates the aorta from the esophagus, findings the team considered strongly suggestive of a secondary fistula. The distal anastomosis of her six-year-old prosthetic graft lay adjacent to the suspected fistulous region. She had originally undergone total arch replacement with a four-branched J-graft prosthesis for a distal aortic arch dissecting aneurysm, and her postoperative course at that time had been uneventful without evidence of graft infection. She had stopped attending outpatient follow-up two years before this presentation, so the complication emerged outside any structured surveillance.
On the day of admission, the team began empiric intravenous meropenem after obtaining blood cultures and performed emergency thoracic endovascular aortic repair, or TEVAR, using a 34 by 34 by 200 millimeter thoracic stent graft that covered the suspected fistulous segment from Zone 3. The graft was oversized by approximately 19 percent relative to the reference aortic diameter. Completion angiography demonstrated no type I or type III endoleak and adequate exclusion of the suspected fistulous segment. The authors emphasize that emergency TEVAR is widely recognized as the most effective initial intervention for achieving rapid hemostasis in patients presenting with massive hemorrhage, but that it cannot eradicate the underlying infection. When TEVAR is used as sole therapy, long-term mortality rates of approximately 55 percent have been reported, because persistent infection, recurrent bleeding, and death remain common without further treatment.
Two days after the stent procedure, the patient was transferred to a collaborating university hospital, where surgeons performed a thoracoscopic esophagectomy, a cervical esophagostomy that brought the remaining esophagus out through the neck, a feeding jejunostomy, and omental flap coverage of the infected mediastinal area. Intraoperative cultures yielded Parvimonas micra, an anaerobic bacterium, and antimicrobial therapy was de-escalated from meropenem to sulbactam/ampicillin based on susceptibility testing, then continued for a total of eight weeks. During that period, enteral nutrition through the jejunostomy and structured rehabilitation preserved her nutritional and functional status. Her inflammatory markers gradually improved without evidence of persistent or disseminated infection, and follow-up CT showed reduction of the perigraft gas and fluid collection without endoleak or new perigraft complications.
The authors argue that this early source control was the pivotal step. Oesophagectomy was performed on postoperative day 2 only after confirming hemodynamic stability and the absence of recurrent bleeding following TEVAR, and the short interval enabled early definitive infection control while minimizing perioperative risk. The timing of esophagectomy in fistula care remains controversial, because current guidelines advocate early intervention but do not define an optimal timeframe. The omental flap provided well-vascularized biological coverage around the retained stent graft, which the authors believe likely contributed to local infection control and isolation of the contaminated field. Previous reports have suggested that failures of graft-preserving strategies are often associated with inadequate source control or insufficient biological coverage, and the combination of early esophagectomy and omental coverage may have prevented progression of graft infection in this patient.
Graft preservation itself remains a contested choice. The traditional principle of definitive treatment for an aortoesophageal fistula is complete removal of the infected aorta and prosthetic material, followed by either in situ reconstruction or extra-anatomical bypass. In this case, explantation of the infected descending aorta and stent graft was considered but deemed prohibitive because of the patient’s advanced age, severe frailty, and high surgical invasiveness. Graft explantation is generally recommended in cases of uncontrolled sepsis, persistent infection despite adequate source control, recurrent bleeding, endoleak-related failure, progressive graft infection, or structural complications such as pseudoaneurysm formation. Graft-preserving strategies may instead be considered in selected patients with controlled infection and prohibitive surgical risk. The authors acknowledge that preservation cannot be regarded as definitive eradication of infection and should be reserved for highly selected patients in whom radical explantation is prohibitively invasive.
Once infection control had been firmly established, surgeons performed a presternal ileocolonic reconstruction on postoperative day 69 to restore gastrointestinal continuity. This extra-anatomical route deliberately routed the new digestive conduit in front of the sternum, avoiding the previously infected and scarred posterior mediastinum, a technical choice the authors describe as strategic and likely important to her favorable recovery. The 69-day interval allowed recovery from sepsis and nutritional optimization through intensive enteral feeding and rehabilitation, addressing the marked malnutrition with which she had arrived. The postoperative course after reconstruction was uneventful, with adequate nutritional tolerance and stable respiratory and circulatory status. Because prosthetic material was retained in a potentially contaminated environment, oral suppressive therapy with amoxicillin/clavulanate was started after the reconstruction and is planned to continue for the rest of her life, a decision made with the infectious disease team that accepts risks of antimicrobial resistance and microbial substitution in exchange for a practical strategy when complete eradication is not feasible.
She was discharged home on postoperative day 94 in good general condition. At the latest follow-up, seven months after TEVAR and five months after the reconstructive surgery, her inflammatory markers had gradually normalized, no recurrent fever or positive blood cultures had been observed, and she remained free from clinical or radiological evidence of stent-graft infection or esophageal complications. The authors stress that the seven-month outcome, while favorable, rests on a relatively short follow-up period, and that long-term durability must still be evaluated. Their follow-up protocol includes periodic contrast-enhanced CT and monitoring of inflammatory markers, and they describe lifelong surveillance as essential for any patient managed with graft preservation. The team is explicit that the case does not propose a novel surgical concept; rather, it demonstrates that a staged, graft-preserving approach may be considered in carefully selected high-risk patients when radical surgery is not feasible, with success depending on early hemostasis, prompt infection source control, effective biological coverage, nutritional optimization, carefully planned delayed reconstruction, and strict postoperative surveillance within a multidisciplinary framework.
Subject of Research: Staged multidisciplinary surgical management of a secondary aortoesophageal fistula after total arch replacement
Article Title: Life-saving staged management of a secondary aortoesophageal fistula after arch replacement: a case report
Article References: Kamikawa, Y., Machida, Y., Fujii, H., Hata, H., & Dohi, S. (2026). Life-saving staged management of a secondary aortoesophageal fistula after arch replacement: a case report. General Thoracic and Cardiovascular Surgery Cases, 5(1), Article 31. https://doi.org/10.1186/s44215-026-00267-0
Image Credits: AI Generated
DOI: 10.1186/s44215-026-00267-0
Keywords: aortoesophageal fistula, TEVAR, total arch replacement, esophagectomy, stent-graft infection, omental flap, delayed reconstruction, Parvimonas micra, case report, aortic surgery, suppressive antibiotics, mediastinal sepsis
News Source: Ophelia Keating. (November 5, 2026). Staged Surgery Saves Woman From Deadly Aortic Graft Fistula. Scienmag.
