Trichosporon asahii infection after total knee arthroplasty: A case report and review of the literature.

Reports of fungal infection after total knee arthroplasty are extremely rare. In most reports, the infecting organism is a Candida species. The present report describes a case involving a 73-year-old immunocompetent woman who underwent total knee arthroplasty and presented one month later with signs of prosthetic infection. She underwent joint debridement and the fluid was sent for culture and sensitivity testing. The culture showed growth of Trichosporon asahii. The patient was administered intravenous and intra-articular injections of amphotericin B, followed by antifungal treatment with voriconazole for one year. At 26 months of follow-up, there was no evidence of infection and the patient was ambulating with a walker. To the authors' knowledge, the present case is the first report of T asahii infection following knee replacement. Early detection, prompt institution of the appropriate antibiotics and regular follow-up are recommended.

Incidences of periprosthetic fungal infections following arthroplasty are extremely rare. In the few reported cases, Candida species were found to be the primary infecting organisms (1). Trichosporon asahii is widely present in the natural environment (2), and T asahii infection is more common in patients with immunodeficiency. T asahii can infect the lung, skin, hair, lymph nodes or can present as a systemic disseminated infection. The present report describes a case involving primary total knee arthroplasty in a patient without immunodeficiency who presented with a postoperative T asahii infection. A review of the literature on the treatment strategy for patients with postoperative fungal infections is also presented.

CASE PRESENTATION

A 73-year-old woman with bilateral osteoarthritis underwent bilateral single-stage total knee replacement. Routine preoperative and preanesthetic investigations were performed. She had a history of hypertension and type 1 diabetes mellitus, and was receiving appropriate medications. Her preoperative fasting blood sugar level was in the range of 6 mmol/L to 10 mmol/L. She did not have a history of intra-articular injections of steroids or any other medications.

The bilateral knee replacement surgery was uneventful. Ceftezole (a first-generation cephalosporin) was administered before the skin incision as a preoperative prophylactic antibiotic. A tourniquet was used bilaterally. A midline skin incision was made and a medial parapatellar approach was used. The surgery lasted for 130 min.

Her postoperative blood sugar levels were regularly monitored, and appropriate medications were administered. The volume of drainage from both knees was larger than usual for the first six days after the procedure. At day 6, the drainage volume was <50 mL and, thus, the drainage tubes were removed. A wound inspection in the right knee revealed that although there was some evidence of superficial poor skin healing, the deep fascia had completely healed. In addition, the patient had a continuous low-grade postoperative fever (37.3°C to 38°C). Laboratory examinations revealed a white blood cell count of 13.6×109/L and a neutrophil percentage of 85.6%. For the first six days, she received a first-generation cephalosporin.

In view of the fever and findings of the wound check, the antibiotic was changed to imipenem/cilastatin sodium, which was given for an additional six days. A wound swab and right knee arthrocentesis culture results were normal at this time. The specimens were inoculated on Columbia Blood agar and China Blue Lactose agar at 35°C in ambient air and cultivated with a Bact/Alert3D Automated Blood Culture System (bioMérieux, France) for three days. Three weeks after surgery, the wound was debrided and resutured due to right knee wound dehiscence. Injectable vancomycin was administered for six days and, subsequently, the wound healed well. One month after the primary surgery, the patient presented with a persistent low-grade fever and recurrent pain on the right side of the left knee joint, which was accompanied by swelling. The temperature of the surrounding skin was also elevated.

Laboratory examinations revealed a normal blood cell count, C-reactive protein level of 27.2 mg/L (normal range 0 mg/L to 5 mg/L) and erythrocyte sedimentation rate of 32 mm/h (normal range 0 mm/h to 15 mm/h). Arthrocentesis revealed that the aspirate from the left knee joint effusion was light yellowish in colour, while aspirate from the right knee effusion was pale, bloody fluid. Both were viscous, not muddy and cultures showed no bacterial growth.

Bilateral joint debridement and lavage was performed. During debridement, pus and necrotic tissue were not present, the synovial fluid was clear and transparent, the surface of the prosthesis was clean and smooth, all of the prosthetic components were stable without any loosening and the bone cement was fixed well. Synovium was completely resected and a pulse lavage system was used to ensure a thorough wash of the joint. In addition, the poly liners were changed. Five days after debridement, intraoperative sampling culture showed no bacterial growth in the right knee; however, the left knee joint fluid culture grew T asahii (Figures 1 and 2). The cultures were grown on Columbia Blood agar and China Blue Lactose agar for 24 h to 48 h. The API 20C identification system (BioMérieux, France) was used to identify T asahii, which was sensitive to amphotericin B and voriconazole. Intravenous amphotericin B (1 mg) was administered and increased to 5 mg every subsequent day. However, due to the patient’s worsening condition, on the third day, the amphotericin B dose was increased to 25 mg. The patient reported marked improvement in pain three days later. Due to the patient’s poor appetite and persistent hypokalemia, the dosage of amphotericin B was not increased to the recommended maximum. Voriconazole alone was started after five weeks, and then after 10 days of intravenous therapy, the dose was changed to 200 mg orally, twice per day. Intra-articular injections of amphotericin B (25 mg once daily) were administered 17 days after surgery. Clinically, the patient continued to improve, her temperature stabilized (with occasional spikes of 37.5°C) and the range of motion of the joint was 0° to 90°.

Figure 1)

Colonies of Trichosporon asahii after 48 h on Columbia Blood agar at 35°C in ambient air. Morphological features of T asahii: T asahii is Gram-positive and produces blastoconida of various shapes, well-developed hyphae, pseudohyphae and arthroconida. (Gram stain, magnification ×1000)

Figure 2)

Colonies of Trichosporon asahii after 48 h on Columbia Blood Agar at 35°C in ambient air. Colonies were cream-coloured and smooth. They become dry, moist, shiny, folded, cerebriform and elevated with marginal zones with age

The patient received oral voriconazole for one year. At the most recent follow-up, which was at 26 months following the primary procedure and 11 months after stopping all antibiotics, including oral voriconazole, the patient’s general condition was good, body temperature was normal, function of the right knee had returned to normal, the left knee joint was mildly swollen (Figure 3), there was no joint effusion and the range of motion of the left knee was similar to the preoperative measure. However, there was pain on weight bearing and the patient required a cane for ambulation. Laboratory re-examinations revealed normal C-reactive protein level, erythrocyte sedimentation rate and routine blood tests, as well as liver and kidney function tests. X-rays showed that the prosthesis of the left knee was not loose, there was no evidence of osteolysis, and the prosthesis was well positioned (Figure 4). Laboratory tests and clinical manifestations revealed no drug toxicity and side effects. Left knee arthrocentesis repeatedly showed no growth of T asahii.

Figure 3)

Gross morphology of the left knee 26 months after surgery. The left knee joint was mildly swollen

Figure 4)

X-ray of the left knee 26 months after surgery. The prosthesis was not loose and there was no subsidence or migration

DISCUSSION

Periprosthetic fungal infection cases reported from 1979 to 2012 after artificial joint replacement totaled 57 reports, including 91 cases (3–59). Of these, there were 41 cases of hip replacements, 46 cases of total knee replacements, three cases of shoulder replacement and one case of metacarpophalangeal joint replacement. Candida albicans was the most commonly reported pathogen and was cultured in 38 cases. Candida parapsilosis was reported in 18 cases, Candida glabrata in 10 cases, Candida tropicalis in six cases, Aspergillus fumigatus in three cases, Pseudallescheria boydii in two cases, and one case of each of the following: Candida guillermondii, Aspergillus niger, Cryptococcus neoformans, Candida lipolytica, Phialemonium curvatum, Histoplasma capsulatum, zygomycosis, Syncephalastrum racemosum, Absidia corymbifera, Sporotrichum schenckii and Rhodotorula minuta. In addition, there was one case of coinfection by C parapsilosis and C albicans, and one case of coinfection by C glabrata and C albicans. The prostheses in 17 of the 91 patients were retained after antifungal therapy, of whom one had persistently painful knee reports during the follow-up, one needed continuous application of fluconazole (63 months), one had sinus tract persistence and three died (the cause of death was not related to joint diseases). Thirty of the 91 patients underwent two-staged reimplantation; swelling and pain during follow-up occurred in four, postoperative bacterial infection occurred in one (antibacterial treatment was successful), and there was one case of secondary infection after surgery (the patient died of heart failure). In addition, 28 of the 91 patients required prosthesis removal, of whom five needed continual treatment with fluconazole and four died. Finally, two of the 91 patients underwent successful single-staged replacement, while arthrodesis was performed in nine of the 91 patients and five of the 91 patients resulted in amputation.

T asahii is widely present in the environment and is present in the normal flora of the human skin, gastrointestinal tract and respiratory tract (60). Formation of biofilms on medical biomaterials in the hospital is very common. Studies have shown the presence of T asahii in the wash basins in operating rooms (61). However, the overall incidence of T asahii infection is low. As a conditional pathogen, it often causes a systemic disseminated infection in patients with transplantation, blood diseases, malignant tumours, immunodeficiency and leukopenia. The incidence of infection in nonimmunodeficient patients is rare, with only one case affecting the central nervous system and lung being reported in an immunocompetent individual (62). There are slightly more than 100 cases of disseminated infection caused by T asahii reported worldwide (63). In previous reports, there were no reported cases of prosthesis infection with T asahii after arthroplasty. T asahii can easily lead to disseminated trichosporosis in particular populations; however, in the present case, blood cultures did not grow T asahii, which may be due to the fact that the individual was immunocompetent.

Periprosthetic fungal infections are related to the use of glucocorticosteroids or immunosuppressants, or are observed in patients with malignant tumours. In addition, an increased risk for infection is related to excessive use of antibiotics for a prolonged period, indwelling catheters and patients in the intensive care unit. The patient in the present case underwent multiple surgeries due to postoperative fever and suspicion of bacterial infection. This particular patient may have been at risk for fungal infection due to prolonged use of antibiotics, the presence of indwelling catheters for long periods and multiple instances of recatheterization secondary to multiple surgeries.

Joint infections following arthroplasty are most often bacterial in origin. Because the overall incidence of fungal infection following arthroplasty is very low, microbiology experts and orthopedic surgeons may suspect contamination and incorrectly label the specimen as a false positive. It is difficult to detect fungal infections on pathogenic examination, and arthrocentesis often fails to reveal a positive fungal culture. In the present case, pathogens were not detected even after several synovial fluid cultures before debridement.

In the present case, detection of T asahii was successful following debridement and immediate inoculation in blood culture media. Studies have shown that this sampling method can significantly improve the detection rate of specimens (35,51). This method led to a prompt diagnosis, especially given the unexplained signs of infection. The present case report suggests that in cases in which the diagnosis is in doubt due to a confounding clinical picture, such as fever, pain and swelling of unknown origin, a fungal infection must be considered. Such cases warrant an extended culture time, specific culture conditions and a high degree of suspicion for the presence of rare microorganisms.

Common Candida infections are sensitive to amphotericin B and triazole drugs (fluconazole, itraconazole, voriconazole and posaconazole). T asahii strains are sensitive to amphotericin B and itraconazole (64). In addition, some reports indicate that T asahii is highly sensitive to fluconazole and voriconazole in vitro (65); however, fluconazole resistance can occur. In the present case, susceptibility testing of the synovial fluid culture suggested resistance to fluconazole, itraconazole and 5-fluorocytosine; however, the patient was sensitive to amphotericin B and voriconazole. Therefore, it is important to select appropriate drugs and consider the resistance pattern.

Studies have suggested that biofilm formation of T asahii is the main factor leading to persistent infection. T asahii has a complex three-dimensional structure under electron microscopy. Interestingly, relative to free cells, the resistance of T asahii biofilm to voriconazole is 16,000 times greater (66), making it difficult to eradicate with voriconazole alone.

The efficacy of amphotericin B treatment of Candida and other fungal infections is well established. However, its renal toxicity can lead to sustained hypokalemia, nausea, vomiting and other side effects. Therefore, the dose is significantly limited in such patients. In the present case, because the application of amphotericin B led to sustained hypokalemia, we administered local intra-articular injections of amphotericin B to increase the local drug concentration. The joint swelling increased after intra-articular injection, but there was no local erythema or rash 24 h after injection. In addition, the patient’s skin temperature decreased, the patient reported pain relief and her joint swelling subsided after one week.

As recommended by the Infectious Diseases Society of America, the appropriate time of antifungal treatment for Candida arthritis is between six and 12 months (67). In the present case, the patient was treated with antifungal voriconazole for one year. Symptoms did not reappear after withdrawal of the medication, even after 11 months. Laboratory tests and clinical manifestations indicated no adverse reactions. In comparison, Fabry et al (45) reported a case involving C albicans infection of the prosthesis, in which symptoms reappeared two weeks after treatment withdrawal. Thus, there may be a relationship between the duration of the medication and patient recovery after surgery.

Because the patient in the present case refused further surgical treatment, we did not perform a two-stage reimplantation, which was most likely the ideal treatment. In the literature, there are more cases of two-stage reimplantation (30 of 91 cases) and removal of the prosthesis (28 of 91 cases) compared with other surgeries. After the prosthesis is removed, instability and dysfunction may occur and, thus, a two-stage reimplantation is recommended. In two reports by Anagnostakos et al (27) and García-Oltra et al (28), the treatment outcomes were drastically different following two-stage reimplantation. In seven two-stage reimplantation cases, Anagnostakos et al (27) reported only one case that did not accept the reoperation, and treatment of the remaining six cases was successful. In contrast, García-Oltra et al (28) reported that only one case was successful following two-stage replantation after debridement, while the remaining six cases failed. In two-stage replacement surgery, the first step involves the removal of the joint prosthesis and insertion of an antibiotic spacer, as well as long-term application of antifungal treatment. Only two single-staged replacement cases have been successful (6,39). In addition, amputation and arthrodesis are becoming rarer because patients do not accept the outcome. Thus, if revision surgery is not possible or if there are contraindications, long-term use of antifungal therapy with retention of the original prosthesis may be an alternative long-term suppression strategy.

CONCLUSION

Cases of periprosthetic fungal infections after total knee arthoplasty are rare, and the ideal treatment is two-stage revision implantation. However, if the patient’s general condition is poor and they are not willing to undergo the revision surgery, long-term treatment with conservative antifungal drugs can be considered on the basis of joint debridement. For immunocompetent patients infected with T asahii, the keys to effective treatment are early detection and long-term application of antifungal treatment.