Background. Chronic disseminated candidiasis (CDC) is typically observed during neutrophil recovery in patients with acute leukemia and requires protracted antifungal therapy.
Objective. Our objective was to document the efficacy and tolerance of corticosteroid therapy (CST) in patients with symptomatic CDC, including those who experienced fever and abdominal pain despite ongoing antifungal therapy.
Methods. We performed a retrospective, multicenter study involving 10 pediatric and adult patients who experienced ongoing symptomatic CDC despite receipt of appropriate antifungal therapy for whom adjuvant oral CST was initiated.
Results. All cases of CDC were proven or probable, as determined on the basis of the European Organization for Research and Treatment of Cancer-Mycosis Study Group definition criteria, and occurred in patients with leukemia. CDC-attributable clinical symptoms resolved with CST, which was started a mean of 33.8 days after antifungal therapy had been initiated. Fever and abdominal pain disappeared a median of 4–5 days, and serum fibrinogen and C-reactive protein levels returned to normal values within 14–30 days. The median duration of hospitalization after CST initiation was 8.8 days. Hepatosplenic microabscesses decreased or disappeared within a mean period of 107 days (range, 30–210 days). No relapses of CDC were observed during a median duration of follow-up of 6.5 years (range, 4–9 years).
Conclusions. In children and adults who experience persistently symptomatic CDC despite ongoing receipt of antifungal therapy, CST involving a prednisone equivalent at a dosage of ⩾0.5 mg/kg per day for at least 3 weeks is associated with a prompt resolution of symptoms and of inflammatory response. These findings support the pathophysiological hypothesis that CDC belongs to the spectrum of fungus-related immune reconstitution inflammatory syndrome.
Chronic disseminated candidiasis (CDC), also called “hepatosplenic candidiasis,” is a severe invasive fungal infection principally observed during neutrophil recovery in almost 5% of patients with acute leukemia treated with intensive chemotherapy [1]. The pathophysiology of CDC is not fully understood, but its clinical, radiological, and biological features have been well characterized [2]. It is assumed that initial colonization of the gastrointestinal tract with Candida species, fol-lowed by liver invasion through portal venous circulation, is the most probable scenario for the onset of this disease [3, 4].
The efficacy of treatment with various systemic antifungals has been reported for CDC. Nevertheless, several months of antifungal therapy are usually required, thus delaying completion of antileukemia treatment. Thus, CDC negatively impacts the survival of these patients [1, 5–10].
CDC often becomes clinically apparent in the course of neutrophil recovery, typically with concomitant negative results of blood and tissue fungal cultures and focal granulomatous reaction noted by examination of liver biopsy tissue specimens [4]. Thus, CDC appears to share several features with immune reconstitution inflammatory syndrome (IRIS) associated with invasive fungal infections (IFIs), as has been described in patients who have HIV infection or who have undergone solid organ transplantation, in whom corticosteroid therapy has been shown to be beneficial [11, 12].
We hypothesized that corticosteroid therapy could also accelerate the recovery from CDC-associated clinical symptoms (abdominal pain and fever) and biological parameters (inflammatory response and elevated liver enzyme levels) that persist despite ongoing antifungal therapy. We report the results of a multicenter, retrospective study conducted in university hospitals in Paris, France, that included all children and adult patients for whom we prescribed CST for severe CDC.
Patients. A multicenter, retrospective study involving 5 University hospitals of the Paris area (Robert Debré, Kremlin-Bicêtre, Necker-Enfants Malades, Hôtel Dieu, and Percy hospitals) was coordinated by the Centre d'Infectiologie Necker-Pasteur to review medical files for 6 adults and 4 children with leukemia who had CDC that was treated with adjuvant CST after administration of appropriate antifungal therapy during 1991–2004. For all cases, CDC occurred after intensive antineoplastic chemotherapy.
Diagnostic criteria of CDC. Patients were classified according to the 2002 European Organization for Research and Treatment of Cancer-Mycosis Study Group criteria for IFI [13]. Cases were considered to be “proven” if histopathologic or cytopathologic examination of needle aspiration or liver biopsy samples revealed yeast cells or if liver biopsy specimens had positive culture results and the liver and/or spleen exhibited clinical/biological or radiologic symptoms consistent with infection. Cases were classified as “probable” in all patients with hematologic malignancies for whom small, peripheral, “target-like” images in the liver and/or spleen were observed by CT or ultrasound and who had elevated serum alkaline phosphatase levels (and/or elevated γ-glutamyl transferase levels in children, because serum alkaline phosphatase level correlates with bone growth and is not a reliable marker for cholestasis), without any evidence of an alternative diagnosis.
Antineoplastic and antimicrobial therapies. Antineo-plastic chemotherapy was administered as described in table 1. All neutropenic patients received conventional broad-spectrum antibacterial therapy in accordance with the guidelines of the Infectious Diseases Society of America [16], as well as empirical amphotericin B therapy (except for patient 1), 1 mg/kg per day, at 5–15 days after the initiation of antibacterial therapy in cases involving prolonged persistent fever. None of the patients received systemic antifungal therapy as a prophylactic regimen before the occurrence of neutropenia. Treatment with granulocyte colony-stimulating factor, which was started between days 5 and 10 of neutropenia, was prescribed for 4 patients.
Corticosteroid therapy. Daily oral CST was introduced at a dose of 0.5–0.8 mg/kg for adult patients and 1 mg/kg for children (except for 2 patients who received 0.5 mg/kg and 2 mg/kg) after confirmation of CDC diagnosis and always in association with systemic antifungal therapy (table 2). CST was initiated in association with daily fluconazole at 6–9 mg/kg in 3 children (weight, <35 kg), 200 mg in 1 child (weight, 45 kg) and 2 adults, 400 mg in 1 adult and 1 child (weight, 45 kg), with liposomal amphotericin B in 1 patient and amphotericin B lipid complex in 1 patient (3 days per week). Three of our patients experienced relapse shortly after CDC diagnosis, received a long duration of steroid therapy, and finally died of leukemia.
A chemistry panel, including liver function tests and determination of the alkaline phosphatase level, γ-glutamyl transferase level, blood urea nitrogen level, creatinine level, and electrolyte level, was obtained at least 3 times per week from all patients during their hospital stay. Complete blood cell counts were obtained daily during hospitalization.
The following criteria were used as end points for evaluating the efficacy of antifungal and corticosteroid therapies: complete resolution of clinical symptoms attributable to CDC, disappearance of CDC-associated visceral images on abdominal CT and/or ultrasound, and normalization of serum levels of alkaline phosphatase and of inflammatory markers.
Centralized review of histopathologic analysis of liver biopsy specimens. A single pathologist reviewed all liver biopsy specimens. Formalin-fixed embedded biopsy specimens were sectioned and stained with hematoxylin and eosin, Giemsa, and Periodic acid schiff stains. Immunohistochemistry was performed using a streptavidin peroxidase method and amino-ethyl carbazole as a chromogen. Monoclonal antibodies to CD3 and UCHL-1 were used to label T lymphocytes (Dako) and a polyclonal anti-Candida albicans antibody (Biogenesis; Argene).
The main characteristics of the patients are summarized in table 1. All patients had an implanted vascular device and experienced mucositis and acute diarrhea during chemotherapy.
Clinical and radiological findings. The clinical and radiological signs of CDC are summarized in table 1. In the 3 patients with acute myeloblastic leukemia, there was an interval of 4–18 days without symptoms of CDC after neutrophil recovery. Two of these patients were discharged from the hospital during this period and were readmitted for persistent fever. In patients with acute lymphoblastic leukemia, the clinical symptoms of CDC appeared before or immediately after neutrophil recovery. Two patients received steroids before the CDC diagnosis (for respiratory distress syndrome in one patient and for consolidation chemotherapy by COPADM [i.e., vincristine, methotrexate, cyclophosphanamide, doxorubicin, and prednisone] in the other); in these 2 patients, clinical signs reappeared after discontinuation or tapering of CST, and clinical symptoms of CDC appeared 5 days thereafter, with a concomitant rapid increase in the neutrophil count.
Biological findings. Biological abnormalities included elevated C-reactive protein and/or fibrinogen serum levels in all patients, elevated alkaline phosphatase serum levels in 3 patients, elevated γ-glutamyl transferase serum level in 5 patients, and cholestasis in a total of 8 patients. The median duration of neutropenia was 17.8 days (range, 5–30 days), and the median neutrophil count at the time of CDC diagnosis was 19,372 cells/mL (range, 6400–36,000 cells/mL). The mean time between neutrophil recovery and diagnosis of CDC was 5.9 days (range, 1–18 days) (table 1).
Microbiological findings. Results of microbiological investigations included blood cultures positive for C. albicans in 1 patient. Eight patients had positive stool fungal cultures (C. albicans in 7 patients [70%], with Geotrichum candidum coinfection in 1 patient). One patient had a urine culture positive for C. albicans, and 1 patient had a urine culture positive for Sacccharomyces species (table 1). The results of liver biopsy cultures were negative for all patients who underwent biopsy (9 of 10 patients), except for 1 who did not receive antifungal treatment before undergoing liver biopsy (table 1).
Liver biopsy specimen analysis. Laparoscopy or coelioscopy was performed in 3 patients and revealed multiple whitish nodules in the liver and/or spleen, with peritoneal inflammatory reaction. Ten transcutaneous liver biopsies were performed in 6 patients; 3 patients underwent 1 biopsy, 2 underwent 2 biopsies (one before and the other after the appearance of liver micronodules), and 1 patient underwent 3 biopsies (the first was on day 60, when liver micronodules appeared; the second was 1 month later, after receipt of fluconazole treatment; and the last was 3 months after receipt of CST and fluconazole treatment). At diagnosis, yeasts were observed on direct examination of a cytological smear of the liver biopsy specimen for 3 patients.
Histologic examination revealed yeasts with pseudomycelia in 2 patients; this was associated with necrotic areas that contained numerous polymorphonuclear cells surrounded by lymphocytes, a few plasma cells, and fibroblasts embedded within areas of fibrosis in 4 patients (data not shown). Granuloma were observed in 2 patients, with clusters of epithelioid macrophages, which were sometimes vacuolated and were always CD68 positive (data not shown). Images suggestive of yeasts and pseudohyphae debris labelled by the anti-Candida antibody could be detected in these necrotic areas (figure 1A). T lymphocytes were observed, predominantly at the periphery of granulomatous lesions (figure 1B). Overall, the diagnosis of CDC was proven for 6 patients and probable for the 4 other patients.
A, Immunolabeled yeasts and pseudohyphae debris of Candida albicans surrounded by polymorphonuclear cells in the necrotic areas (hematoxylin counter-staining; original magnification, ×200). B, CD3-positive lymphocytes surrounding granuloma areas (hematoxylin counter-staining; original magnification, ×100).
Antifungal therapy for CDC prior to CST initiation included amphotericin deoxycholate, liposomal amphotericin B, and fluconazole (table 2). After a median duration of 33.8 days (range, 6–60 days) of treatment with amphotericin B alone (except for patient 10, who was treated with multiple antifungals during 120 days), fever persisted in 9 patients in association with other clinical symptoms suggesting CDC evolution. Hepatosplenic and/or kidney and lung micronodules persisted, despite the receipt of antifungal therapy, with biological signs of persisting inflammation (table 1).
CST administration. The median interval between CDC diagnosis and CST initiation was 40 days (range, 1–120 days). Fever disappeared after a mean period of 1.2 days (range, 1–3 days) after onset of CST, except in 1 patient for whom disappearance of fever occurred within 30 days after onset. All patients left the hospital within a mean of 8.8 days (range, 1–30 days) after initiation of CST. Serum fibrinogen and/or C-reactive protein levels decreased 1 week after commencement of CST and returned to normal values within a median period of 14.4 days (range, 2–30 days). The findings of radiologic investigations returned to normal or to with residual microcalcification after a mean period of 107 days (range, 30–210 days) after initiation of CST (table 2).
Antifungal treatment and patient outcome after commencement of corticosteroid treatment (CST).
Full-dose CST was maintained for a mean duration of 21 days (range, 10–60 days), after which the CST dose was tapered progressively on the basis of clinical and biological parameters. The mean duration of CST, always given in association with antifungal therapy, was 124.9 days (range, 49–240 days). In 3 patients, CST was maintained for a long period (180–240 days) because of a relapse of leukemia, without any adverse event.
Antifungal therapies administered after commencement of CST include liposomal amphotericin B, which was given for 1 month to 2 patients. All patients received fluconazole, with a median duration of treatment of 109 days (range, 49–240 days) (table 2).
Outcome. Of the 10 patients, 5 (3 adults and 2 children) underwent bone marrow transplantation 90 days to 4 years after CDC diagnosis. Neither reactivation of CDC nor emerging systemic fungal infection was observed in any of the patients. During the conditioning regimen, 4 children switched their treatment from fluconazole to deoxycholate amphotericin B (1 mg/kg per infusion) for 30 days after bone marrow transplantation 3 times per week. Six patients died of relapse of leukemia without evidence of evolving infection, and 4 patients were alive 4–9 years after the onset of CDC. In our study, the median survival rate was 40.8% at 110 months (data not shown), whereas, in a historical study, the rate was 26% at 96 months [1]. In these 2 studies, there was no mortality directly attributable to CDC.
We show—to our knowledge, for the first time—that adjuvant CST, in addition to antifungal therapy, is beneficial for CDC-related clinical symptoms and inflammatory response in adults and children. Indeed, CDC-related clinical symptoms disappeared as early as 1 day after initiation of CST in almost all patients, and inflammatory markers decreased within 1 week, reaching normal values in a median of 2 weeks. In addition, our patients were able to leave the hospital after 1 week. In contrast, in historical studies, the time to resolution of fever has exceeded 4 weeks (range, 9–196 days) [4, 6, 17]; nausea and vomiting has taken 6–8 weeks to resolve [18], and increased levels of liver enzymes persisted for a mean of 4.4 months (range, 1 week to 12 months) [19]. Of note, in our study, CDC-related radiological abnormalities disappeared within a mean of 106 days (range, 30–210 days)—a result not significantly different from that of historical series (mean, 100 days; range, 28–390 days) [18], thereby suggesting that adjuvant CST does not contribute to the disappearance of radiological lesions.
Several months of antifungal treatment are typically recommended for eradication of CDC [5–8, 20]. The use of amphotericin B (or its lipid formulations), fluconazole, and flucytosine may be considered, with amphotericin B typically considered to be the therapeutic cornerstone [5, 20]. More recently, voriconazole and caspofungin have also been used, with encouraging results and better tolerance than polyenes [7, 21, 22]. However, in patients with CDC who are treated with antifungals alone, clinical symptoms often persist during a very prolonged period, and a median cumulative dose of up to 2 g has been advocated for amphotericin B [5, 6].
It should also be emphasized that the clinical improvement observed in all of our patients allowed us to minimize the interval between diagnosis of CDC and initiation of chemotherapy or allogeneic stem cell transplantation. The safety of resuming antineoplastic chemotherapy in association with antifungal therapy has already been advocated by Walsh et al. [9] and Masood et al. [20]. One might argue that corticosteroid therapy prescribed early in the course of CDC (i.e., close to the time of neutrophil recovery) might be associated with an increased risk of subsequent infection—particularly infections of fungal origin. However, no reactivation of candidiasis was observed in our patients, and no other new opportunistic infections were diagnosed. The efficacy and safety of steroid therapy observed here emphasizes that they may be considered as adjuvant therapy in association with antifungal treatment in cases of severe inflammatory-related symptoms during CDC.
It is well known that CDC becomes clinically overt on recovery from neutropenia in the context of negative liver fungal culture results, in accordance with the major role of neutrophils in the host defense against pseudohyphae, the invasive form of Candida species [23–25]. Of note, the classic histologic image of CDC includes large infiltrates of neutrophils surrounding a few Candida pseudohyphae and blastoconidia, in association with lymphocytes and macrophages in the liver, as noted here and elsewhere [2].
Studies of mice have shown that the development of a protective Th1 response against invasive candidiasis requires the coordination of a series of cytokines (namely, IFN-γ, TGF-β, IL-6, TNF-α, and IL-12) and the relative absence of cytokines (e.g., IL-4 and IL-10) that inhibit Th1 response [23, 26–29]. IL-10 has been shown to suppress phagocytosis of neutrophils against C. albicans blastoconidia and to impair neutrophil-induced hyphal damage [29]. Increased serum concentrations of IL-10, but not of IL-4, have been reported in patients with CDC, compared with control subjects; this may, in fact, be the consequence of an exacerbated Th1 response rather than a causal factor in CDC [26].
Of note, in addition to classic risk factors for CDC (e.g., younger age, prolonged neutropenia, and antibiotic prophylaxis [30]), a genetic basis has been advocated for the occurrence of CDC. Indeed, in one study, a common haplotype of the IL-4 promoter was overrepresented in patients with CDC [31]. However, a study by the same group challenged this result by showing that common polymorphisms in 6 critical genes of innate immunity did not contribute to an increased risk of CDC [32]. Knowing that steroids inhibit Th1 responses [33, 34] and IL-10 synthesis [26, 35], we propose that CST may modulate an inappropriate immune response associated with the onset of CDC by modulating T cell activation and mononuclear phagocyte function.
Overall, the occurrence of CDC during neutrophil recovery, the liver histopathologic findings in patients with CDC, and the beneficial effect of steroids towards its inflammatory symptoms strongly suggest that CDC belongs to the IFI-induced IRIS. IFI-induced IRIS was initially described in the context of profoundly immunosuppressed patients with the AIDS, often with initial severe or disseminated fungal infections, such as Pneumocystis jiroveci associated pneumonia, disseminated histoplasmosis, or Cryptococcus neoformans infection [36–40]. IRIS typically occurs after the commencement of HAART, when the HIV load markedly decreases, with or without dramatic recovery of the CD4 cell count. IRIS has also been recently reported during pulmonary aspergillosis in patients with cancer who are recovering from neutropenia [41]. Of note, reascension of the CD4+ T lymphocyte count is also known to occur after intensive antineoplastic chemotherapy [42] and might also be involved in the pathogenesis of CDC, as suggested by the presence of T lymphocytes in inflammatory areas, as noted in our patients' liver biopsy samples. A similar complication was recently reported in solid organ transplant recipients whose immunosuppressive therapy, which contained Th1-inhibiting drugs, was markedly reduced after the occurrence of severe C. neoformans infection and who, in some cases, experienced additional complications associated with graft loss [43]. Interestingly, in the context of IRIS occurring in patients who have AIDS or who have undergone solid organ transplantation, the addition of steroids to appropriate antimicrobial therapy may also be effective and safe to control this paradoxical reaction [38, 44], as evidenced in our study of CDC.
We thank Dr. Emmanuel Roilides for his helpful advice.
Potential conflicts of interest. All authors: no conflicts.
IDSA Members: For your free access to this journal, log in via the IDSA members area.
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy