A 62-year-old male presented in April 2005 with fever and malaise. Blood counts revealed pancytopenia with 330 neutrophils/μl and 65% blasts. A bone marrow aspirate established a diagnosis of acute myelogenous leukemia (M2) with a normal cytogenetic study. He received induction chemotherapy with idarrubicine (12 mg/m2 × 3 days) and cytarabine (200 mg/m2 × 7 days). The first day of chemotherapy was designated day 0. Because of the persistence of blasts in the bone marrow, further treatment with mitoxantrone (10 mg/m2 × 3 days) and cytarabine (400 mg/m2 × 5 days) was given. Prophylactic fluconazole was administered from day +3 after induction chemotherapy until day +39. A brief episode of fever of unknown origin developed between days +3 and +7 and was treated empirically with imipenem. On day +42 after induction chemotherapy, maculopapular skin lesions appeared, disseminating in 48 h and involving the face, trunk, and limbs (up to 42 papules with central ulceration were counted). Some of the lesions resembled ecthyma gangrenosum. A skin biopsy sample showed dermoepidermic necrolysis and fibrin thrombi. No cultures were obtained at the time. The clinical picture was considered to be compatible with a septic process, although a toxic epidermal necrolysis was also possible. When the skin lesions appeared, the patient was neutropenic and was receiving empirical antimicrobial therapy with imipenem and teicoplanin. From day +39 after induction chemotherapy, the patient was started on caspofungin because of a relapsing neutropenic fever. Four days later, he became afebrile and his neutrophil count increased to more than 500/μl. Caspofungin and antibiotics were withheld, and the patient was discharged on day +50 after induction chemotherapy without any antifungal drug. The skin lesions gradually improved.
Two more consolidation chemotherapy cycles were administered in June and August 2005 (23 mg idarrubicine × 2 days and 390 mg cytarabine × 5 days in both cycles). No fever or skin lesions appeared, and no antifungal therapy was administered during the first consolidation chemotherapy. Two new skin lesions identical to the previous ones appeared on day +10 after the second consolidation cycle, before he became neutropenic on day +12. Because of the appearance of these lesions, fluconazole was started on day +10 and maintained until day +28 after consolidation chemotherapy, when the patient was discharged. The lesions gradually disappeared.
In September and October 2005, the patient's leukemia was in complete remission and his peripheral blood counts were normal. During this period of time, while the patient was waiting for autologous stem cell transplantation (SCT), a new flare-up of disseminated skin lesions reappeared that was clinically similar to the previous ones. A second skin biopsy was performed (October 2005), and the initial interpretation was of an acute inflammatory dermal lesion with a mixed neutrophilic and histiocytic infiltrate. During this time, numerous lesions continued reappearing and resolving spontaneously. No antifungal therapy was administered during this time. In November 2005, he was admitted to undergo autologous peripheral blood SCT. On admission, he had four skin lesions in resolution (arm, thigh, and trunk) (Fig. ). Deeper sections of the second skin biopsy were performed at this time and revealed a nidus of fungi in the dermis with broad hyphae surrounded by a neutrophilic and histiocytic infiltrate without leukemic infiltration. The lesion was morphologically interpreted as possible mucormycosis (Fig. ). Because of these findings, on day +8 after SCT, liposomal amphotericin was started but had to be suspended after one dose because of a severe adverse reaction, and no further antifungal therapy was administered. During transplantation, no new lesions appeared. A third biopsy of one of the four resolving lesions was obtained on day +9 after SCT, and it showed characteristics similar to those of the previous sample. Biopsy cultures were negative. A galactomannan test was negative, and a high-resolution computed tomography chest scan was normal. New sections of the first skin biopsy were reviewed, and abundant hyphae were seen with special stains within the necrotic epidermis. The patient was discharge on day +12 after SCT. Briefly, all three of the biopsy samples that were examined histologically (the first, second, and third), showed dense fungi in the dermis, and in one of them (the second) they also invaded the subcutaneous tissue.
On day +26 after SCT, two new lesions appeared on the thigh. A new biopsy was performed, but no histopathology study was done. Cultures were negative. Voriconazole was started, and the lesions disappeared. Voriconazole was stopped on day +78 after SCT, and there were no further skin lesions at the time of the last follow-up (June 2006). No evident port of entry was found for the fungal infection.
Tissue samples from the four skin biopsy samples were sent to the Mycology Reference Laboratory of the Spanish National Center for Microbiology. Portions of tissue homogenates were plated on Sabouraud dextrose agar, on brain heart infusion with 5% sheep blood, and on malt extract agar (2%; Oxoid Unipath, Madrid, Spain). Cultures were invariably negative.
Subsequently, the tissue samples were analyzed genetically by using a real-time PCR-based assay to identify the fungus observed (4
). The LightCycler technology (Roche Applied Science, Madrid, Spain) was used for the assay. Quantification was performed by online monitoring of fluorescence and identification of the exact time point at which the logarithmic linear phase could be distinguished from the background (crossing point). The level of fluorescence was proportional to the amount of DNA generated during the PCR process. The LightCycler Probe Design Software 1.0 (Roche) was used to optimize the design of primers and probes. The primers used were ITS 1 (forward, 5′ TCCGTAGGTGAACCTGCGG 3′) and ITS 4 (reverse, 5′ TCCTCCGCTTATTGATATGC 3′). These assays were designed to amplify internal transcribed spacer (ITS) regions 1 and 2 from the fungal rRNA gene complex. The primers used were based on the nucleotide sequence of the ITS regions of 1,800 isolates of 150 distinct fungal species. These sequences belong to the database of the Department of Mycology of the Spanish National Center for Microbiology, and the design was done with the help of Fingerprinting II informatix software, version 3.0 (Bio-Rad, Madrid, Spain).
Standardized commercial kits were used for the extraction of DNA from clinical samples (Wizard SV Genomic DNA Purification System; Promega, Madrid, Spain). The manufacturer's instructions were strictly followed. The PCR mixture (20 μl) consisted of a 0.5 mM concentration of each of the primers and 5 mM MgCl2. Each reaction mixture contained a 2-ml aliquot of extracted specimen together with 18 ml of the master mix. The cycling conditions included a first step for preincubation (activation of the enzyme) and denaturation of the template DNA at 95°C for 10 min. The next step included an amplification program of 45 cycles as follows: denaturation, 95°C, 10 s; annealing, 56°C, 5 s; and extension, 72°C, 30 s. In addition, quantification standards were run in conjunction with each set of samples. To monitor for contamination, aliquots of distilled water were prepared concurrently. The experiments were repeated two times on different days.
Fungal DNA was detected in biopsy samples taken on day +44 after induction chemotherapy and in October 2005. Melting-curve analysis was performed in the range of 47 to 95°C. PCR products were subjected to electrophoresis in 2% agarose gels (Pronadisa, Madrid, Spain) to verify the PCR results. Amplified fragments were sequenced (ABI Prism 377 DNA sequencer; Applied Biosystems, Madrid, Spain). Sequences were assembled and edited by the SeqMan II and EditsEq software packages (DNAstar, Inc. Lasergene, Madison, WI). Sequence analysis was performed by comparison with the nucleotide sequence database available in the laboratory and with the GenBank database (http://www.ncbi.nih.gov/GenBank/
After analyzing DNA segments comprising the ITS1 and ITS2 regions and comparing them with those of the reference strains, the sequence of the clinical isolate matched those of Metarrhizium anisopliae. The percentage of similarity was 100%. A phylogenetic analysis was also conducted with Fingerprinting II informatix software. The methodology used was maximum-parsimony clustering. Phylogram stability was assessed via parsimony bootstrapping with 1,000 simulations. Analysis was done with sequences of 1,800 organisms including seven M. anisopliae strains that were used as reference strains.
Patients with hematological malignancies treated with intensive chemotherapy are prone to several opportunistic infections, including invasive fungal infections. Although Candida
are the fungi more often isolated from these patients, emerging fungal species are gradually increasing because of several factors (12
). We report here the first case of proven disseminated skin infection due to M. anisopliae
in an immunocompromised adult patient. In this case, the infection, limited to the skin without involvement of other distant organs, followed a striking mild and intermittent evolution, considering that the patient was severely immunosuppressed and that he did not receive antifungal treatment during most of the time because of a delayed diagnosis.
is an entomopathogenic fungus that belongs to the class Hyphomycetes
). It is a common insect pathogen (1
) and rarely infects animals or humans (5
). It has a wide range of host species, including arachnids and five other orders of insects (10
), and a worldwide distribution in soil, although it has been more often isolated from soil-dwelling insects living in tropical and Mediterranean areas because of its high thermal requirements (7
). Its first description was under the name Entomophthora anisopliae
as a pathogen of the wheat cockchafer in 1879 by Metschnikoff. Sorokin renamed it M. anisopliae
in 1883 (11
). It is now being used for biological control of insect populations belonging to the orders Coleoptera
(locusts and beetles). The disease produced in insects is called green muscardine because of the green color of its conidial masses (13
). It is also being investigated as a biological control agent for adult malaria and filariasis vector mosquitoes (9
To date, there have been only five reported cases of disease in humans: two cases of fungal keratitis, one in Colombia (3
) and one in the United States (5
); two cases of sinusitis in immunocompetent hosts in the United States (8
); and one disseminated invasive infection with skin, lung, and brain involvement in an Australian child with acute lymphoblastic leukemia (2
). In the last case, the child had disseminated skin lesions consistent with ecthyma gangrenosum that never resolved.
Our patient also presented with skin involvement with disseminated skin lesions as in the case reported by Burgner et al. (2
), some of them resembling ecthyma gangrenosum, but did not show any other tissue involvement. The clinical course of our case was quite atypical. It was surprising that our patient had at least two flare-ups of disseminated skin lesions and two flare-ups of localized lesions without a clear relationship with neutropenia evolution or antifungal therapy. It is also striking that although the patient underwent several periods of severe neutropenia, most of them without any antifungal treatment, the infection course was mild with just skin involvement. The initial port of entry was not identified, and as far as we know the patient did not have previous contact with commercial products containing M. anisopliae
, not even with bees or other types of insects. There were no more cases of M. anisopliae
identified in the hospital before or after this case. The organism was seen histologically and identified by PCR techniques in three and two different skin biopsy samples, respectively, taken from different skin areas and in different flare-ups but was not isolated from cultures. Histological examination of three out of four biopsy samples was done, PCR assay of two out of four biopsy samples was done, and culture of two out of four biopsy samples was done. This case shows the clinical utility of molecular methods in the diagnosis of invasive mycosis due to emerging pathogens. Without the use of PCR techniques, this case would have been classified as a proven invasive mold infection, probably aspergillosis.
There is no standard treatment for this infection. The case of keratitis described in Colombia was successfully treated with topical natamycin (3
). The two cases of sinusitis reported in immunocompetent hosts were treated with surgery without antifungal agents, and the patients recovered without complications (8
). Antifungal susceptibility testing of five isolates was done by Clinical and Laboratory Standards Institute (formerly National Committee for Clinical and Laboratory Standards) methods, and all were resistant to amphotericin B, 5-flucytosine, and fluconazole (8
). The MICs of itraconazole and voriconazole were lower, suggesting that they could be more effective. The Australian child with the disseminated infection was treated with amphotericin B and 5-flucytosine, but the treatment was ineffective and the patient died (2
). In vitro susceptibility testing (tablet diffusion method) was performed, and the results suggested that the organism was sensitive to amphotericin B but resistant to 5-flucytosine, fluconazole, and itraconazole (2
). In the veterinary literature, we found a case of a cat with invasive rhinitis successfully treated with itraconazole (6
Our patient was receiving caspofungin when the skin lesions appeared for the first time, suggesting that this drug was ineffective. After the transplantation and with normal white cell counts, two new lesions appeared. As we were at that time aware of the diagnosis of an invasive fungal infection, we decided to treat him with voriconazole with prompt resolution of the lesions. Nevertheless, a biopsy was performed on one of these lesions and both PCR and cultures were negative. Likewise, considering the previous evolution, it is difficult to know if the lesions would have also resolved without treatment.
In summary, in this case we can highlight several interesting points. First, M. anisopliae should be added to the growing list of emergent fungal pathogens able to cause invasive fungal infections in immunosuppressed adult patients. Second, the clinical picture caused by this pathogen can be very different from what we are used to with other molds. Third, a diagnosis of invasive fungal infection based only on histological findings can underdiagnose uncommon pathogens. Finally, a collaboration of clinicians, microbiologists, and pathologists is necessary for the correct diagnosis of difficult infections.