Surgical Treatment of Persistent Pseudomonas aeruginosa Bacteraemia After n-Butyl-2-cyanoacrylate for Gastric Bleeding Varices

A 34-year-old man with a history of Wilson’s disease diagnosed 10 years prior and with secondary liver cirrhosis (Child–Pugh class B7) was admitted to the hospital as a result of upper gastrointestinal bleeding secondary to fundic varices that were treated by NCBA in April 2022. Eight weeks after the bleeding episode he was readmitted with a 14-day history of fever and chills. Physical examination revealed absence of abdominal pain, bleeding signs, diarrhoea, respiratory or urinary symptoms. Laboratory tests showed a slight increase in inflammation parameters (leukocyte count was 4.6 × 10⁹/L with 92% neutrophils, and C-reactive protein was 4.42 mg/dL) and liver function tests showed aspartate aminotransferase (AST) of 43 U/L, alanine transaminase (ALT) of 50 U/L, alkaline phosphatase (ALP) of 115 U/L, gamma-glutamyl transferase (GGT) of 39 U/L, total bilirubin of 0.9 mg/dL, and INR of 1.32, similar to previous levels. Chest X-ray and urinalysis were normal and respiratory virus nucleic acid testing (influenza, respiratory syncytial virus and SARS-CoV-2) was negative. Blood and urine cultures were obtained. Ceftriaxone was started empirically and d-penicillamine was discontinued as a possible cause of drug-related fever. Once the patient was admitted, Pseudomonas aeruginosa with no pattern of resistance was isolated from blood cultures and therefore he was switched to ceftazidime. Despite the correct antibiotic treatment, he persisted with fever and chills and had repeated blood cultures with P. aeruginosa 2 weeks after starting ceftazidime 2 g every 8 h. The treatment was switched to ceftolozane/tazobactam 2 g every 8 h plus ciprofloxacin 750 mg every 12 h (MIC < 0.125 µg/ml) and amikacin 1 g every 24 h (MIC < 4 µg/ml). A thoracoabdominal computerized tomography and an echocardiography did not detect any source of infection. Two days later, PET-CT revealed inflammatory changes around the NCBA plug with no other findings; hence, in the absence of alternative infectious foci and given the persistence of bacteraemia despite effective antibiotic therapy against the bacterial strain, a presumptive diagnosis of NCBA plug infection was considered (Fig. 1). Moreover, the patient’s fever persisted and liver function was getting worse. The case was discussed in a multidisciplinary committee (involving staff from gastroenterology, hepatology, transplant surgery, infectious diseases, abdominal radiology, anaesthesia and resuscitation departments) and liver transplantation was approved as a definitive treatment of the underlying liver disease (decompensated cirrhosis).

Partial gastrectomy was extensively deliberated upon but ultimately dismissed owing to anatomical peculiarities, specifically the presence of multiple feeding collaterals beneath the NCBA plug. Similarly, placement of a transjugular intrahepatic portosystemic shunt (TIPS) as a bridge therapy to decompress varices prior to surgery was also discarded because of the potential risk of infection (endotipsitis) as positive blood cultures persisted. The Catalonian Liver Transplant Board was consulted, and accepted prioritised inclusion of the patient on the transplant waiting list. The patient received a deceased donor liver transplant 29 days later. Two days after that, ceftolozane-tazobactam was stopped; however, 10 days after discontinuing antibiotics the patient presented with another episode of P. aeruginosa bacteraemia, and antibiotics were therefore restarted. It was confirmed that the source of infection was the plug, even after the decompression of gastric varices and the resolution of portal hypertension. It was then decided to surgically remove the NCBA plug with a vertical gastrectomy 1 month after the transplant (Fig. 2). P. aeruginosa with identical resistance pattern as in the blood cultures was isolated in the surgical sample of the NCBA plug. After the surgery, fever resolved after 72 h, and the patient was discharged with a plan to complete a 6-week course of ceftazidime plus ciprofloxacin. Blood cultures performed 48 h after stopping antibiotic therapy were negative and the patient has remained asymptomatic ever since. Figure 3 shows clinical and analytical characteristics as well as the susceptibility pattern of the isolates.