Introduction
Mediastinitis is a rare but serious complication of congenital cardiac surgery with a high mortality rate. Its incidence in paediatric patients has been reported to be less than 1%. Most cases are caused by bacterial pathogens, whereas fungal infection has been identified as a significant independent risk factor for mortality, with reported mortality rates as high as 80%. Reference Bernheim, Chikhi and Foufa1 Trichosporon infections in children are extremely rare, but have been reported in neonates as well as in premature infants and patients with malignancies or haematologic disorders. Reference Kourti and Roilides2 Breakthrough infections have been described in patients receiving echinocandin therapy, and successful treatment with voriconazole has also been reported. On the other hand, paediatric Trichosporon mediastinitis is extremely rare, and its clinical characteristics and optimal management approaches have not been established. We report a case of Trichosporon austroamericanum mediastinitis in an infant that developed after congenital cardiac surgery. Reference Francisco, Desnos-Ollivier and Dieleman3
Case report
A male infant was diagnosed with right isomerism, common atrioventricular valve, double-outlet right ventricle, and total anomalous pulmonary venous return. At the age of 1 month (weight 5.3 kg), he underwent systemic-to-pulmonary artery shunt surgery and was discharged home. At the age of 4 months (weight 6.5 kg), he presented with marked hypoxaemia at a routine outpatient visit. Emergency examination revealed pulmonary venous obstruction, prompting total anomalous pulmonary venous return repair and common atrioventricular valvuloplasty. Although he was successfully weaned from cardiopulmonary bypass, progressive desaturation required venovenous extracorporeal membrane oxygenation. He was then admitted to the paediatric ICU with an open chest. On postoperative day 5, he was successfully weaned from extracorporeal membrane oxygenation. Due to acute kidney injury, a haemodialysis catheter was subsequently placed directly into the right atrium.
On the next day, his haemodynamics deteriorated. Fungi were cultured from tissue samples taken from the mediastinum, and after debridement and irrigation, negative-pressure wound therapy was initiated. Antifungal therapy was empirically initiated with micafungin; however, based on the microscopic findings and the lack of response to micafungin, Trichosporon infection was suspected, and the treatment was subsequently switched to voriconazole. Antifungal susceptibility testing revealed minimum inhibitory concentrations of >16 μg/mL for micafungin, 1 μg/mL for fluconazole, 0.06 μg/mL for voriconazole, and 2 μg/mL for amphotericin B. Voriconazole was initiated with a loading dose of 9 mg/kg twice daily, and subsequent dosing was adjusted according to therapeutic drug monitoring to achieve a target trough concentration of 1–4 μg/mL. Because the organism could not be identified by mass spectrometry at our institution, it was referred to the Medical Mycology Research Center, Chiba University, for further analysis. Sequencing of ribosomal DNA loci, including the intergenic spacer region 1 and the internal transcribed spacer region, identified the pathogen as T. austroamericanum, a novel yeast species first reported in 2024 (Figure 1a,b).Reference Francisco, Desnos-Ollivier and Dieleman3 The isolate has been preserved as IFM 69578 through the National BioResource Project, Japan (JPMBRP-202228).
( a ) Culture on Sabouraud dextrose agar shows cream-colored, dull, wrinkled cerebriform colonies. ( b ) A microscopic image demonstrates the formation of budding yeast cells and hyphae. ( c ) Intraoperative photograph during wound debridement showing a substantial amount of necrotic tissue with a moist wound environment. ( d ) Photograph demonstrating daily negative-pressure wound therapy.
Wound irrigation and negative-pressure wound therapy foam exchanges were initially performed on alternate days; nevertheless, each exchange revealed abundant mycelium-covered nonviable tissue, and cultures remained persistently positive (Figure 1 c , d ). When irrigation was intensified to a daily regimen, wound cultures eventually showed negative results. Although this process was prolonged, a minimally acceptable wound condition was ultimately achieved, and chest closure was scheduled on postoperative day 52. Considering the extremely high risk of recurrence, chest closure was performed using omental flap interposition to reinforce infection control, and pectoralis major myocutaneous flap coverage to reduce tension at the wound site and ensure secure closure (Figure 2a–f). The patient recovered without recurrence. Voriconazole was administered intravenously for 2 months, followed by oral therapy, which was continued until after bilateral Glenn surgery at the age of 1 year and discontinued 6 months postoperatively.
( a ) The midline incision in the chest was extended to the abdomen, and the omentum was harvested. ( b ) Although the volume of the omentum was not entirely satisfactory, it did reach the artificial blood vessel. ( c ) The patient’s hemodynamic status was unstable, making it difficult to fully close the sternum. ( d ) Wound closure was made possible by harvesting a large pectoralis major flap. ( e ) Intraoperative photograph at the completion of surgery. ( f ) Postoperative photograph after suture removal.
Discussion
Trichosporon spp. are uncommon pathogens, typically affecting immunocompromised adult patients with haematologic malignancies or who are undergoing chemotherapy. These infections are even rarer in children, but have also been reported in patients with haematologic malignancies and in preterm infants. Reference Kourti and Roilides2 In our case, T. austroamericanum was identified as the causative pathogen. Reference Francisco, Desnos-Ollivier and Dieleman3 Although postoperative open-chest management with extracorporeal membrane oxygenation support was required, the patient had no apparent immunosuppressive conditions; nonetheless, the presence of asplenia syndrome may have conferred an underlying susceptibility to infection. Trichosporon is resistant or less susceptible to echinocandins, which probably accounted for the poor response to micafungin. In addition, cases of poor response to and resistance to polyene antifungal agents have also been reported. Reference Kourti and Roilides2,Reference Matsue, Uryu and Koseki4 Therefore, early transition to voriconazole—an agent with superior in vitro and clinical activity against Trichosporon—was crucial.
Our standard approach for treating bacterial mediastinitis includes extensive debridement followed by negative-pressure wound therapy, with foam exchanges on alternate days. In general, within one or two exchanges, the wound bed improves remarkably, necrotic tissue disappears, cultures become negative, and primary closure without filler tissue becomes feasible. Using this approach, we have previously encountered no prolonged treatment courses or recurrence. Nevertheless, the patient’s case followed an unusual course. Despite repeated debridement and alternate-day negative-pressure wound therapy, the wound bed did not improve, adhesions failed to progress, necrotic tissue persisted, and cultures remained positive. Only after implementing daily debridement and copious irrigation, were negative cultures finally achieved. Nonetheless, this intensive regimen imposed significant physiological stress. Each procedure was accompanied by high fever, vasodilation, and circulatory collapse. Although the wound environment remained suboptimal, progressive improvement was observed, whereas daily invasive procedures became increasingly unsustainable. Moreover, vascular access became progressively more difficult, and renal function continued to deteriorate. Consequently, these factors collectively led us to proceed with chest closure. Unlike our standard practice of direct closure without additional tissue, we opted to use omental filling in this case. This approach was intended to reduce dead space and fill the infected area with well-vascularized tissue, thereby improving the delivery of immune cells and antifungal agents. Also, because the skin had atrophied due to prolonged exposure of the chest opening, and direct closure would have placed excessive tension on the skin and significantly increased the risk of dehiscence, we extensively dissected the pectoralis major muscle and used it as a pectoralis major musculocutaneous flap to close the wound. Our goal was to achieve definitive closure along with reducing the risk of recurrent infection.
To our knowledge, this is one of the first reported paediatric cases of T. austroamericanum mediastinitis. Awareness of these unique features is essential because conventional treatment protocols for bacterial mediastinitis may be inadequate for this rare but life-threatening infection.
Acknowledgements
We would like to thank Takashi Yaguchi, Koji Yamashita, and Sayaka Ban, from the Medical Mycology Research Center, Chiba University, for their assistance in fungal identification.
Financial support
The authors declare none.
Competing interests
The authors declare none.
