A 47-year-old female who had suffered from obstructive jaundice and ascites for two months was admitted to our hospital for further examinations. A peripheral blood smear revealed lymphoblastic cells with blue cytoplasm and cytoplasmic vacuoles, and a CT scan showed swelling of multiple lymph nodes in the peritoneal cavity. She was diagnosed as having acute lymphoblastic leukemia of Burkitt's type confirmed by karyotype analysis of a bone marrow aspiration sample showing 47, XX, t(8;14)(q24;q32), +mar in 9 out of 20 cells and FISH analysis of bone marrow showing IgH/c-MYC fusion signals. Combination chemotherapy with rituximab+hyper CVAD followed by high-dose methotrexate/cytosine arabinoside was started using a central venous catheter (CVC), but she could not obtain complete remission (CR) because of early regrowth of lymphoblasts in bone marrow after the induction chemotherapy (Fig. 1). Thereafter, salvage chemotherapy with rituximab+IVAC and rituximab+ CODOX-M was started, but she could not obtain CR because of central nervous system involvement of leukemic cells. Fluconazole (FLCZ) at 200 mg/day had been administered as a prophylactic anti-fungal drug since the start of initial chemotherapy. Three months after the start of chemotherapy, serum βD-gulucan became positive, and the anti-fungal drug was therefore changed from FLCZ to VRCZ because of covering aspergillosis even without characteristic features of aspegillosis shown by a CT scan of the chest and sinus. After that, galactomannan and βD-gulucan remained negative, but she was administered prophylactic VRCZ at 300 mg BID because of transient increase of galactomannan and continuous myelosuppression after repeated chemotherapy (Fig. 1). Nine months after the diagnosis, she suffered from high fever in a severe neutropenic period (white blood cell count of 0.1-0.6×10⁶/L) after salvage chemotherapy. Venous blood culture taken through the CVC showed Corynebacterium species and Trichosporon mucoides (T. mucoides) by a conventional morphological identification method. At that time, both Cryptococcus antigen and βD-gulucan were negative and culture of the removed CVC was also negative. Although the optimal therapy for invasive trichosporonosis is still not known and the identified Trichosporon was susceptible to VRCZ in vitro (MIC was 0.060 µg/mL), clinical anti-fungal effect of the drug was not sufficient. Therefore, VRCZ was switched to L-amphotericin B (L-AMB). Thereafter, Trichosporon species was not identified by repeated venous blood samplings. She died of refractory leukemia 2 weeks after the diagnosis of fungemia by Trichosporon. The stocked isolate was genetically analyzed later and identified as belonging to T. dermatis by DNA sequencing of the intergenic spacer 1 region [4].

Several species of Trichosporon have been reported to cause opportunistic fungemia [5]. T. dermatis was newly detected from an infected human skin lesion and was also recognized as a cause of summer-type hypersensitivity pneumonitis (SHP) [3, 6]. T. dermatis is closely related to T. mucoides morphologically and biochemically and is easily misdiagnosed as T. mucoides [7]. Therefore, previous reports concerning trichosporonemia should be reevaluated. Although it had been unknown whether this species was pathogenic to humans, fungemia caused by T. dermatis has been reported since 2006 [7, 8]. Retrospective gene analysis of 22 Trichosporon sp. blood stream isolates sequentially obtained from different patients in Brazil showed only one case of T. dermatis [9]. Similarly, Ruan et al. reported that no T. dermatis infection was identified from 14 patients with fungemia [1]. Although Rodriguez-Tudela et al. documented eight cases of T. dermatis causing infections, only one case showed bloodstream infection [10]. Therefore, fungemia caused by T. dermatis is very rare. Although Fekkar et al. reported that T. dermatis antigens were cross-reactive with both Aspergillus galactomannan and Cryptococcus capsular antigen, serum of our patient did not show cross-reactivity [8].

Fungemia caused by T. dermatis in our patient occurred during administration of prophylactic VRCZ and levofloxacin at standard doses. Although amphotericin B and echinocandins were not active against Trichosporon isolates in vitro, azoles, especially VRCZ, have been reported to show good potency [1]. In vitro susceptibility testing of T. dermatis to antifungal agents has been reported to be similar to that of other pathogenic Trichosporon species [3, 9], and it was in fact also susceptible to VRCZ in vitro in this case. Profound immunosuppression after repeated chemotherapy, severe neutropenia and oral administration might be the main reasons for the ineffectiveness of VRCZ at a standard dose. Monitoring plasma VRCZ levels might be necessary to avoid subtherapeutic levels [11]. Risk factors for invasive trichosporonosis are prior antibiotic therapy, use of a central catheter, malignancy, prior chemotherapy and neutropenia, all of which were present in our patient and all of which are also risk factors for candidemia [1, 12]. Since fungemia by T. dermatis was resolved after removing the central venous catheter in this patient, the influence of central venous catheter might be a major reason of fungemia. Although the clinical effect of L-AMB in our patient was not clear because of early death after the switch of agents, Trichosporon species was no longer identified by repeated venous blood samplings. Careful attention should be paid to breakthrough fungal infection in patients with severe innmunosupression and neutropenia even if they are administered wide-spectrum anti-fungal drugs.

In conclusion, this is the first report of T. dermatis infection during administration of VRCZ. Careful attention should be paid to invasive fungal infection in patients with immunosuppression even with administration of prophylactic anti-fungal drugs because of the high mortality rate of invasive trichosporonosis, especially in patients with malignancies [1]. Molecular techniques are necessary for accurate and definitive diagnosis of trichosporonosis.