Central nervous system mucormycosis (CNS-M) is a severe disease with difficult and often delayed diagnosis, leading to high mortality. The aim of this new study was to assess clinical and radiological presentation according to underlying conditions and dissemination routes to optimize diagnostic strategies. We conducted a retrospective national study including 54 CNS-M cases diagnosed between 2005 and 2020, with brain imaging reviewed by two neuroradiologists. CNS-M resulted from presumed hematogenous dissemination in 29 patients (54%) and from direct extension in 25 (46%), known as rhino-orbito-cerebral mucormycosis (ROCM). No neurological symptoms were found in 10/54 (19%), regardless of dissemination route. Hematogenous CNS-M mainly affected highly immunocompromised (HM or SOT) patients (90%), including 43% neutropenic. Radiology showed abscesses (87%) and small-vessel disease (39%). In ROCM, two patterns emerged depending on osteolysis (19/25, 76%) or its absence (6/25, 24%). ROCM without bone lysis, mostly in severely immunosuppressed patients, caused meningitis without abscess, whereas osteolytic ROCM led to abscess formation (11/18, 60%). Without osteolysis, perineural spread along the optic nerve occurred in 2/3 cases. Serum Mucorales PCR was positive in 91% of hematogenous and 64% of ROCM cases. Fungal co-infections occurred in 26%. This study underscores distinct invasion patterns and the need for extensive workup in CNS-M, highlighting the diagnostic value of MRI with gadolinium and serum Mucorales qPCR based on dissemination route and underlying condition. MRI is particularly useful in ROCM for detecting meningitis (80%), large-vessel disease (30%), and perineural involvement (8%).

ABSTRACT Invasive fungal diseases are difficult to diagnose because of the limited sensitivity of culture. Panfungal PCR amplicon sequencing assays (targeting ribosomal RNA, such as 18S, 28S, ITS) are recommended for fungal identification in histopathology samples showing fungal elements. However, data describing its overall performance and consistency are lacking. This systematic literature review and meta-analysis assessed the performance of panfungal PCR on formalin-fixed paraffin-embedded (FFPE) and non-fixed (fresh or frozen) tissue samples. A systematic literature search was performed to include studies reporting the use of panfungal PCR for fungal identification in FFPE or non-fixed tissue samples. PCR sensitivity and specificity were assessed using the reference standard of histopathology showing fungal elements. Quality assessment was performed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. Pooled estimates were obtained using random-effects meta-analysis. Twenty-eight studies were included. In FFPE samples (18 studies, 852 samples), sensitivity and specificity were 75.4% (95% confidence interval [CI], 59.2–86.6) and 93.5% (70.2–98.9), respectively. Sensitivity in non-fixed samples (13 studies, 207 samples) was 86.5% (74.7–93.3), while specificity could not be assessed (insufficient data). Comparative analyses showed a significantly higher sensitivity of panfungal PCR over culture (88.2%; 76–94.7 vs 52.2%; 39–65, P = 0.001). Sub-analyses could not demonstrate the superiority of one PCR target over another due to limited data. Panfungal PCR exhibited adequate sensitivity and good specificity in FFPE samples. Sensitivity was even higher in non-fixed samples and largely superior to culture. Nevertheless, large interstudy variability was observed, warranting interlaboratory studies to define the optimal PCR target and standardized protocols. IMPORTANCE Invasive fungal diseases are difficult to diagnose because of the low sensitivity of culture. Panfungal PCRs are widely used for fungal identification in tissue specimens but suffer from heterogeneous procedures and performance. This meta-analysis shows an acceptable sensitivity (75.4% and 86.5% in fixed and non-fixed samples, respectively) and good specificity (93.5%) of panfungal PCR, supporting its use, not only on histopathology-positive fixed samples but also in non-fixed samples concomitantly with other diagnostic tools (cultures and fungal-specific PCRs if available). These results provide a strong basis for further standardization of panfungal PCR techniques via interlaboratory assays to assess reproducibility and optimize analytical protocols. CLINICAL TRIALS This study is registered with PROSPERO as CRD42023461148 .

Wickerhamomyces anomalus is a yeast rarely involved in human invasive fungal diseases (IFD). We retrospectively analyzed 44 episodes of W. anomalus IFD in France during 2012–2024. Injecting drug use (IDU) was the main risk factor among 26/35 (74.3%) incident cases. Most infections were community acquired; overall 3-month mortality rate was 1/30 (3.3%). Short tandem repeat (STR) genotyping and whole-genome sequencing analyses revealed substantial genetic diversity among isolates. However, 1 STR genotype was shared by 2 IDU patients, suggesting common exposure. In addition, 1 isolate obtained from a cotton filter used for drug preparation was identical by STR genotyping to the bloodstream isolate from the same patient, indicating direct inoculation via contaminated material or poor injection practices. Our findings highlight the increased risk for W. anomalus IFD among IDU patients and emphasize the importance of targeted preventive measures within that population.

The most frequent invasive mycoses in France are candidemia, Pneumocystis jirovecii pneumonia, and invasive aspergillosis. They occur primarily in immunocompromised patients. The risk of Pneumocystis pneumonia and cryptococcosis has notably decreased for people living with HIV through the past two decades. Patients with diabetes or an autoimmune disease represent an increasing proportion of these cases. Mortality of invasive mycoses is still high, particularly in cases of candidemia and invasive aspergillosis. Azole-resistant Aspergillus spp. isolates remain rare in France and the proportion of resistant strains in cases of invasive candidiasis is stable. The emergence of Candida auris, which is associated with outbreaks in healthcare facilities, is closely monitored. Except for candidemia, investigation of predisposing genetic conditions should mainly be conducted when invasive mycosis occurs in patients without any clear risk factor.

Fungi in the family Trichosporonaceae are rarely involved in invasive disease but are frequently associated with colonization or respiratory allergic infection. Trichosporonaceae exhibit intrinsic resistance to echinocandin antimicrobial drugs, posing challenges for treatment and contributing to high mortality rates. We complied a nationwide analysis of 112 cases of invasive disease caused by Trichosporon spp. and related fungi, diagnosed in France over 20 years, that combined clinical data, susceptibility profiles, and molecular identification. We identified 12 species; T. asahii was the most common species recovered, and the new species T. austroamericanum was next. Comparison of clinical data highlighted species and genotypic differences, such as a much higher proportion of children infected by T. asahii and major differences in antimicrobial drug susceptibility. Correct identification is not only of epidemiologic interest but also necessary for patient management because of the varying clinical and microbiological characteristics found in different species.

Background:Antimicrobial resistance (AMR) in Candida species is an emerging global health threat. This study aims to document the implementation of standardized surveillance tools from the Global Antimicrobial Resistance and Use Surveillance System (GLASS) and to describe the distribution and antifungal resistance profiles of Candida species in bloodstream infections generated through the study outputs. Methods:The GLASS-Fungi pilot was a multicentre observational demonstration study where twenty-four laboratories were invited to collect and share Candida bloodstream infections (BSIs) data with the World Health Organization (WHO) using a standardized surveillance protocol and WHONET software. Data were collected for patients with laboratory-confirmed Candida bloodstream infections between January 2017 and July 2021. Participating sites were trained to collect, deduplicate, and report clinical and microbiological data. Patient demographic and microbiological data were summarized using descriptive statistics. The percent distribution of Candida species and resistance profiles were analysed with a 95% confidence interval (CI) calculated using robust standard errors clustered at the site level. Participating sites shared experiences with using the standardised GLASS surveillance tools and procedures through structured feedback forms and consultative meetings. This study was not designed to support causal inference or population-level generalization; analyses are descriptive and intended to illustrate outputs generated through pilot surveillance implementation. Findings:Fourteen laboratories from 13 countries contributed data from 3447 patients with candidemia. Overall, Candida albicans was the predominant species (37.6%; 95% CI: 33.8-41.5). C. albicans was most common in the Americas, Europe, and Africa, whereas Candida tropicalis was more prevalent in Southeast Asia. Candida auris was detected by two laboratories in Southeast Asia and Africa. Among isolates with interpretable antifungal susceptibility testing (AFST) results and corresponding established breakpoints, 13.6% (95% CI: 10.1-17.2%) were resistant to at least one antifungal, with fluconazole resistance highest among C. parapsilosis isolates (29.7%; 95% CI: 18.2-41.2). Challenges to scaling up fungal AMR surveillance globally included limited fungal laboratory testing capacity, restricted access to antifungal susceptibility testing, and lack of sustained funding, particularly in low and middle-income countries. Interpretation:The study was a global, multi-centre initiative to systematically collect and report surveillance data on Candida BSIs using standardised data collection tools and reporting procedures. The study identified major training capacity and infrastructure gaps. Addressing those is an essential step for anticipating and responding to emerging invasive fungal infections. Funding:This work was supported by the United States Centers for Disease Control and Prevention.

Wickerhamomyces anomalus is a yeast rarely involved in human invasive fungal diseases (IFD). We retrospectively analyzed 44 episodes of W. anomalus IFD in France during 2012-2024. Injecting drug use (IDU) was the main risk factor among 26/35 (74.3%) incident cases. Most infections were community acquired; overall 3-month mortality rate was 1/30 (3.3%). Short tandem repeat (STR) genotyping and whole-genome sequencing analyses revealed substantial genetic diversity among isolates. However, 1 STR genotype was shared by 2 IDU patients, suggesting common exposure. In addition, 1 isolate obtained from a cotton filter used for drug preparation was identical by STR genotyping to the bloodstream isolate from the same patient, indicating direct inoculation via contaminated material or poor injection practices. Our findings highlight the increased risk for W. anomalus IFD among IDU patients and emphasize the importance of targeted preventive measures within that population.

Abstract In the OPTIFIL study that included 39 patients with localized invasive pulmonary aspergillosis complicating hematological malignancies, Aspergillus-specific IFN-γ-producing T cells were detectable in 41% of the patients and predicted better 6-week outcomes, while their absence correlated with uncontrolled disease. Lymphopenia was the main limitation to Aspergillus ELISpot assay interpretation.

Invasive fungal diseases are life-threatening complications, particularly in immunocompromised patients, and require rapid and accurate diagnosis to improve clinical outcomes. Although major advances in fungal diagnostics that includes antigen detection, molecular assays, and Matrix-Assisted Laser Desorption/Ionization - Time Of Flight (MALDI-TOF) mass spectrometry, have transformed diagnostic strategies, access to these tools remains heterogeneous. In France, national data on diagnostic capacities for invasive fungal diseases have been lacking. Using the framework of the national prospective surveillance program for invasive fungal diseases (SINFONI network), we conducted a survey to assess laboratory diagnostic practices in France. A secured 116-item questionnaire was distributed to 58 participating laboratories, of which 48 responded (83%). Automated blood culture systems and MALDI-TOF mass spectrometry for yeast identification were universally available, with 40 (83%) of the 48 participating laboratories also using MALDI-TOF for mould identification. Antifungal susceptibility testing was performed on-site in 47 (98%) centres for yeasts and in 37 (77%) for moulds. Antigen-based biomarkers were widely available on-site, particularly cryptococcal antigen (n = 43, 90%) and Aspergillus galactomannan (n = 39, 81%), whereas β-D-glucan testing was available in only 26 (54%) of the centres. PCR-based diagnostics were implemented on-site in 43 (88%) centres, most commonly for Pneumocystis jirovecii, Aspergillus spp., and Mucorales. Systematic screening for Candidozyma auris colonization in at-risk patients was performed in 30 (63%) centres, predominantly using culture-based methods (n = 24). Overall, this survey provides the first national overview of diagnostic capacities for invasive fungal diseases in France, highlighting a strong laboratory mycology infrastructure while identifying remaining gaps in access to specific biomarkers, molecular assays, and mould antifungal susceptibility testing.

BACKGROUND:Pneumocystis pneumonia (PCP) is a well-known infectious complication of organ transplantation requiring prophylaxis at least within the first 6 month to 1 year post transplantation. RESEARCH QUESTION:Few data exist comparing the characteristics of Pneumocystis pneumonia associated with kidney, heart, liver or lung transplant recipients. STUDY DESIGN & METHOD:We here conducted a cross-sectional study nested within the surveillance of our national reference center for invasive mycoses to analyze the prospectively declared cases of PCP occurring in solid organ transplant recipients over a period of 11 years. RESULTS:We found that the median occurrence of PCP post-transplantation varies from the organ recipients with PCP occurring earlier in liver recipients and later in other organs reaching a median of 3.9 years in kidney recipients. We also found a clear increase the proportion of positive mycological criteria in PCP cases occurring within 2 years post-transplantation. Age and ICU hospitalization were major variables associated with 3 month-mortality with liver and lung recipient having a better outcome than renal transplant patients upon adjustment including age. A trend toward the role of additional risk factors (such as HIV, Cancer of hematological malignancy) in the outcome of PCP was also observed. INTERPRETATION:Altogether, this study described on a large patient cohort, some key mycological and clinical information associated with PCP in solid organ transplant patients. The characteristics of PCP in kidney and heart recipients seems similar.

Invasive fungal diseases (IFDs) are common and often fatal in severe burn patients due to skin barrier loss and immune dysfunction. However, current definitions of invasive mold infections are poorly adapted to this population. This study evaluated the characteristics of various diagnostic criteria and their combinations in relation to clinical outcomes in burn patients. We conducted a retrospective cohort study of all patients admitted to the Burn ICU from 2014 to 2023 with ≥15% total burn surface area and at least one sample sent to the mycology lab. Criteria included direct microscopy, culture (respiratory, skin, or tissue), species-specific quantitative PCR (qPCR) (Aspergillus, Mucorales, and Fusarium) on plasma/tissue/bronchoalveolar lavage fluid, and serum galactomannan. Among 276 patients, 489/6,184 samples were positive, including 281 skin biopsies (direct examination and conventional culture) and 132 plasma specimens (qPCR). Positive diagnostic criteria ≥1 was found in 93 patients (33.7%): Aspergillus (25.7%), Mucorales (10.9%), and Fusarium (9.8%). Twenty-seven patients (9.8%) had ≥2 criteria involving ≥2 mold types. Mortality rose with the number of positive criteria: 12.7% (0), 10.7% (1, 2), 27.3% (3, 4), and 46.7% (≥5) (P < 0.001). Plasma qPCR was positive in 81.3% of Mucorales, 40% of Aspergillus, and 15.4% of Fusarium cases with skin involvement. Skin biopsies (direct examination and conventional culture) combined with species-specific plasma qPCR enhance timely and reliable IFD diagnosis in burn patients. Mortality correlated with the number of positive criteria and coexistence of multiple mold species, underscoring the need for broad antifungal coverage and the value of multi-criteria diagnostics to guide treatment.IMPORTANCEInvasive mold infections are frequent and often fatal complications in patients with severe burns, occurring in up to 20% of cases with a total burn surface area exceeding 15%. Despite their severity, no standardized case definition currently exists to guide research or clinical management in this population. The performance of existing mycological diagnostic criteria remains unknown in burn patients. In this 10-year retrospective study, we evaluated the diagnostic performance of individual and combined mold-related criteria in relation to patient outcomes, analyzing more than 6,000 clinical samples. These findings provide a first comprehensive assessment of mold diagnostic markers in the burn population.

Penicillium, Talaromyces, and Purpureocillium species (so called Penicillium-like) are increasingly recognised as opportunistic fungal pathogens capable of causing a wide range of infections, from superficial to invasive. Although ocular infections caused by these environmental fungi have been reported in the literature, they remain relatively underrecognised. We conducted a multicentric retrospective study of Penicillium-like ocular cases reported in the RESeau de Surveillance des Infections Fongique (RESSIF) database of the French National Reference Center for Invasive Mycoses and Antifungals between 2012 and 2021. Among 55 cases of Penicillium-like invasive infections reported to the RESSIF network in France, a total of 24 cases (44%) were identified, which predominantly presented as keratitis (n = 22, 92%). Predisposing factors were as follows: contact lens wear (n = 11, 46%), topical corticosteroid use (n = 9, 37.5%), ocular trauma (n = 5, 21%), pre-existing corneal disease (n = 4, 17%), and previous ocular surgery (n = 2, 8%). Purpureocillium lilacinum was the predominant pathogen (n = 19, 79%), followed by four Penicillium species and one Talaromyces species. Direct examination of ocular samples was positive in 13 cases (54%). Antifungal susceptibility testing (European Committee on Antimicrobial Susceptibility Testing) revealed that all species exhibited high minimal inhibitory concentrations to amphotericin B (> 1 mg/l). Most patients were treated with a combination of topical voriconazole and amphotericin B (n = 15, 62.5%), which was administered alongside oral voriconazole in nine cases (37.5%). Surgical intervention was required in six cases (25%) and included keratoplasty (n = 4) and enucleation (n = 2). This study provides a valuable overview of Penicillium-like fungal ocular infections and highlights the importance of systematically surveilling filamentous fungal keratitis.

OBJECTIVES:Invasive aspergillosis is a difficult-to-diagnose fungal disease, the diagnosis of which relies on many criteria including mycological criteria. Recently, the European Organisation for Research and Treatment of Cancer/Mycoses Study Group Education and Research Consortium promoted Aspergillus PCR as a specific marker given the significant progress made in optimizing and standardizing Aspergillus real-time quantitative PCR (qPCR) assays. However, there is no unique and optimal assay in the literature, particularly for optimally detecting aspergillosis due to the main Aspergillus species responsible for invasive aspergillosis. METHODS:We therefore designed, optimized, and validated a new type of real-time PCR assay using a diagnostic case-control study approach. This assay is based on reverse transcriptase qPCR methodology and multi-Aspergillus section-specific primers (multi Aspergillus species reverse transcriptase qPCR [MAS-RTqPCR]). This assay detects whole nucleic acids from species belonging to six Aspergillus sections (Fumigati, Flavi, Nigri, Terrei, Nidulantes, and Usti). RESULTS:The limit of detection of this assay was five copies per reaction, and the analytical specificity was 100% based on the screening of >90 fungal species. We used three groups of patients to validate the assay, testing 120 patients and 223 plasma samples. The clinical sensitivity was 0.91 (95% CI, 0.71-0.98) and clinical specificity was 1.00 (95% CI, 0.93-1.00). Follow-up of positive patients revealed superiority of the MAS-RTqPCR assay compared with the reference assay. In addition, our MAS-RTqPCR assay detected three patients with non-Aspergillus fumigatus infections. CONCLUSIONS:Overall, this study demonstrates that our MAS-RTqPCR assay is a promising diagnostic for invasive aspergillosis and monitoring PCR-positive patients.

BACKGROUND:Despite tuberculosis being a well-known concern in patients with advanced human immunodeficiency virus (HIV), the STATIS trial, which focused on its management, highlighted significant mortality rates. Histoplasmosis, a fungal disease endemic in sub-Saharan Africa, presents with similar clinical manifestations as tuberculosis. Therefore, it may be prevalent and potentially responsible for deaths in patients with advanced HIV in this region. We conducted an ancillary study of the ANRS STATIS trial to provide the first prevalence estimates of histoplasmosis among individuals with advanced HIV in Côte d'Ivoire. METHODS:We analyzed urine samples from patients previously enrolled in the STATIS trial in Côte d'Ivoire. These ambulatory patients with newly diagnosed HIV infection, CD4+ T-cell counts <100/µL, and eligible for antiretroviral therapy (ART) were randomized to receive either systematic or test-guided tuberculosis treatment. We performed Histoplasma antigen enzyme immunoassay on their urine samples. RESULTS:The prevalence of Histoplasma antigenuria was 68/280 (24.3%; 95% CI: 19.5%-29.8%), including 52/280 (18.6%; 95% CI: 14.3%-23.7%) symptomatic patients. Of 22 tuberculosis cases documented at inclusion, 8 (36.4%) also had Histoplasma antigenuria. In patients who died within the 48-week follow-up, the prevalence of Histoplasma antigenuria was 15/42 (35.7%% 95% CI: 22.0%-52.0%) compared with 22.3% (95% CI: 17.3%-28.2%) in those surviving. These survivors had a higher body mass index, CD4+ T-cell count, and hemoglobin and platelet count than those who died. CONCLUSIONS:The prevalence of Histoplasma antigenuria was comparable to that of tuberculosis, and histoplasmosis was potentially responsible for preventable deaths. Prospective studies are needed to confirm these findings and promote screening strategies in sub-Saharan Africa.

Cryptococcus neoformans/gattii and Histoplasma capsulatum var. capsulatum may present atypical histopathological features inducing diagnostic errors. We aimed to estimate the frequency of these atypical features in formalin-fixed tissue (FT) samples and to assess the relevance of an integrated histomolecular diagnosis using specific H. capsulatum PCR and panfungal PCR followed by Sanger sequencing and/or targeted massive parallel sequencing (MPS). A total of 27 FT from 23 patients with a histopathological diagnosis of cryptococcosis (n = 16 FT from 13 patients) or histoplasmosis (n = 11 FT from 10 patients) were retrospectively included. All FT were consultation cases. Mycological identifications on equivalent fresh tissue were available for 11/23 (47.8%) patients. The expert pathologist review modified the diagnosis suggested by the initial pathologist in 7/27 (25.9%) FT. Fungal morphology and tissue inflammation were compared between both mycoses. The most discriminant atypical criterion was the presence of dented-looking yeasts, observed in 68.75% (11/16) of C. neoformans/gattii and none (0/11) of H. capsulatum var. capsulatum (P = .002). For the 12/23 (52.2%) patients without mycological identification on fresh tissue, an integrated histomolecular diagnosis on FT using specific PCR or panfungal PCR followed by Sanger sequencing and/or MPS led to fungal identification in 9/12 (75%) cases; for cryptococcosis, the targeted MPS sensitivity was higher than that of Sanger sequencing (P = .041). Thus, because atypical histopathological features may be tricky, integrated histomolecular diagnosis is essential for optimal patient care. Pathological features of Cryptococcus neoformans showed more atypical dented-looking yeasts than Histoplasma capsulatum. An integrated histomolecular diagnosis improved fungal identification in 75% of the cases, with a higher sensitivity of targeted massive parallel compared to Sanger sequencing for cryptococcosis.

BACKGROUND:The burden of invasive fungal diseases (IFDs) in patients with complicated alcoholic hepatitis (CAH)-defined by ≥ 2 hepatic (ascites, jaundice, liver failure, encephalopathy) or extrahepatic (coagulopathy, shock, kidney or respiratory failure) dysfunctions within 30 days-remains poorly characterised. AIMS:To assess the burden of IFDs in CAH and compare it with bacterial pneumonia (BP). METHODS:We conducted a retrospective nationwide cohort study of adult CAH patients in France (2012-2021). The primary exposure were IFDs. The primary outcome was 3-month mortality or liver transplantation. Associations were assessed with adjusted odds ratios (aORs) in complete-case and propensity score-matched cohorts. A 6-week landmark analysis and time-dependent Cox models were used to evaluate time-varying effects. RESULTS:Among 11,434 CAH patients (median age 55 years; 72% male), 2.2% and 15% developed IFDs and BP, respectively. Three-month survival was 17.5% (95% CI: 13.0-23.0) in IFDs, 46.8% (44.3-49.3) in BP and 60.0% (59.4-61.4) in those without either (p < 0.001). IFDs occurred in 44.3% of patients with BP, and BP increased IFD risk (aOR 2.93, 95% CI: 2.23-3.84). In matched analyses, IFDs were associated with a fourfold increase in mortality (aOR 4.58, 95% CI: 3.02-7.20), while BP showed a lower association (aOR 1.23, 95% CI: 1.06-1.43). IFDs were strong time-dependent predictors of death. CONCLUSIONS:IFDs affected 1 in 50 CAH patients and carried a disproportionate mortality risk, compared with BP. These findings support the implementation of targeted screening and early antifungal strategies in CAH management, as for BP.

Mucormycosis is a fungal infection typically affecting immunocompromised patients. Here, we report a severe case of invasive cutaneous and peritoneal mucormycosis caused by Rhizopus microsporus, successfully treated with a combination of antifungal therapy, PD-1 inhibitor, and interferon-gamma. We highlight the importance of personalized immunotherapy in refractory cases of invasive mucormycosis.

Invasive aspergillosis (IA) caused by Aspergillus flavus remains poorly described. We retrospectively analyzed 54 cases of IA caused by A. flavus reported in France during 2012-2018. Among cases, underlying IA risk factors were malignancy, solid organ transplantation, and diabetes. Most (87%, 47/54) infections were localized, of which 33 were pleuropulmonary and 13 were ear-nose-throat (ENT) infection sites. Malignancy (70% [23/33]) and solid organ transplantation (21% [7/33]) were the main risk factors in localized pulmonary infections, and diabetes mellitus was associated with localized ENT involvement (61.5%, [8/13]). Fungal co-infections were frequent in pulmonary (36%, 12/33) but not ENT IA (0 cases). Antifungal monotherapy was prescribed in 45/50 (90%) cases, mainly voriconazole (67%, 30/45). All-cause 30-day case-fatality rates were 39.2% and 90-day rates were 47.1%, and rates varied according to risk factor, IA site, and fungal co-infections. Clinicians should remain vigilant for A. flavus and consider it in the differential diagnosis for IA.