Bloodstream infections (BSIs) are life-threatening systemic invasions predominantly caused by bacterial or fungal pathogens (Dolatabadi et al., 2024; Najafzadeh et al., 2024). Notably, fungemia in critically ill patients carries serious prognostic implications, with 30-day attributable mortality rates remaining between 35 and 50% despite advancements in diagnostic modalities and targeted therapies (Oren and Paul, 2014; Ruhnke et al., 2018; Arias et al., 2017). Yeasts of the Candida genus are the predominant etiological agents of invasive fungal diseases (IFDs) in hospitalized populations, accounting for 42.7% of all IFDs (Ruiz-Ruigómez et al., 2018). As the first national surveillance system for IFDs, the China Hospital Invasive Fungal Surveillance Network (CHIF-NET) offers valuable epidemiological evidence. However, our analysis of fungemia data from 2019 to 2023 showed that there are regional differences in the distribution of fungal pathogens and antifungal resistance (Enoch et al., 2017). Understanding the epidemiological patterns and in vitro antifungal susceptibility characteristics of the causes of fungal BSIs in this region is crucial for clinical diagnosis and treatment. These evidence-based findings directly inform clinical decision-making algorithms for selecting empirical antifungal therapies and help optimize provincial-level antimicrobial stewardship programs.

The number of patients with invasive mycoses is rising due to increasing risk factors such as immunocompromised states, neoplastic diseases, leukemia, and invasive operations (Pappas et al., 2018; Puig-Asensio et al., 2014). The annual number of fungal BSIs reported from 31 hospitals in the Sichuan provincial center of the CHIF-NET increased steadily from 2019 to 2023, with the exception of 2022 when a decline occurred likely due to COVID-19-related disruptions. This is the largest recent study of fungemia in Sichuan. Few other studies in China have conducted province-wide surveillance over a 5-year period, particularly in comparing different types of hospitals. In our study, men accounted for 58.6% (1,191/2,034) of the patients, with 80.0% (1,627/2,034) being middle-aged or elderly patients (>46 years old). Previous studies have shown that fungal BSIs are highly prevalent in middle-aged and elderly populations. In addition, 36.8% (748/2,034) of the cases were from ICUs, followed by the Department of Internal Medicine and Surgical Departments. IFDs in ICUs represent a critical clinical challenge, with a global incidence increasing by 6.8% annually. This rise is largely due to the severity of illness in ICU patients and risk factors such as invasive procedures, prolonged bed rest, and extended antifungal use, all of which contribute to an increased likelihood of infection (Vincent et al., 2009).

Candida species dominated the fungal distribution (88.7%, 1,805/2,034). The most common species was C. albicans, followed by the C. parapsilosis complex, the C. glabrata complex, and C. tropicalis. This distribution differs from reports in other countries and shows regional variation compared to national data from the CHIF-NET (Xiao et al., 2018a). The pathogenicity of fungal BSIs shifts from C. albicans to non-albicans Candida species, and there is also variation in the virulence of different Candida species (Arendrup, 2013). In this study, non-albicans Candida species outnumbered C. albicans as causes of BSIs. This shift in pathogen distribution may reflect the widespread use of antifungal agents, which can alter the epidemiology of fungal BSIs (Pfaller et al., 2019). Two strains of C. auris were detected in 2023, representing only 0.1% (2/2,034) of the cases. However, C. auris is an emerging fungus that can cause serious illnesses and is easily spread among patients in healthcare facilities. The emergence and rapid increase of C. auris should be considered a threat. C. auris was first discovered in China in 2018 (Wang et al., 2018; Chen et al., 2018). As of December 2023, C. auris has been reported in 10 provinces across China. A study reported 312 cases of C. auris infections across 18 hospitals. The researchers found significant differences in prevalence between years, with the lowest number of cases occurring in 2020–2021. Notably, the number of infections increased dramatically to 182 in 2023, marking a 450% rise compared to 2022 (33 cases) (Bing et al., 2024). C. auris causes serious infections including bacteremia, wound infections, and catheter-associated infections (Tsay et al., 2017), and it can also colonize the urinary tract, respiratory system, digestive tract, and central nervous system (Eyre et al., 2018). The fungus persists in hospital environments and colonizes patients’ skin, posing significant infection control risks. Primarily affecting immunocompromised individuals through nosocomial transmission, C. auris carries a high mortality rate (up to 60% in BSI cases) (Boutin and Luong, 2024). Its propensity for causing hospital outbreaks jeopardizes the safety of patients and healthcare workers. On the other hand, C. auris is naturally prone to carry antifungal resistance-related genes, and more than 90% of clinical isolates of C. auris are resistant to azole antifungal agents (Chowdhary et al., 2023). Resistance to other antifungal agents (e.g., amphotericin B or echinocandins) is also frequently observed, complicating treatment strategies for C. auris infections and significantly increasing healthcare costs, compared to other fungal infections.

The isolation rates of different fungi varied across hospitals. For example, the C. parapsilosis complex was the most frequently isolated fungal pathogen at the Second Hospital of West China, Sichuan University, accounting for 53.6% (22/41) of cases. This may be because the hospital is a specialized children’s hospital (Yılmaz-Ciftdoğan et al., 2021). The C. parapsilosis complex comprises C. parapsilosis, C. metapsilosis, and C. orthopsilosis. Although these species were not individually listed in the Results section of this study, it should be noted that they exhibit differences in antimicrobial susceptibility breakpoints and phenotypic profiles. Similar considerations apply to the C. glabrata complex. Unlike in adults, the C. parapsilosis complex may be the predominant pathogen causing fungal BSIs in children. Among Candida species, the C. parapsilosis complex ranks second only to C. albicans in biofilm-forming capability among clinically relevant species (Mancera et al., 2021). Biofilms formed by this complex are structurally less complex and thinner than those of C. albicans, facilitating their attachment to medical devices (Larkin et al., 2018). In children, the gut is not fully developed, allowing the C. parapsilosis complex to enter the bloodstream through damaged mucosal barriers. This leads to infections in children. It can also be transmitted to immunocompromised children by healthcare workers through contact.

An important limitation of our study is the variability in antifungal susceptibility testing methods among the participating hospitals, which may affect the comparability of resistance rates for certain agents. Nevertheless, the data presented in this study can still truly reflect the prevalence of fungal BSIs in Sichuan Province. In this study, the best in vitro susceptibility of Candida to the antifungal agents analyzed was observed with echinocandins, followed by amphotericin B. Among the azoles, fluconazole (except C. albicans) and voriconazole showed sensitivities of no more than 90%. This may be due to the extensive clinical use of fluconazole, which has driven natural selection in Candida itself, leading to mutations in certain genes that make it less sensitive to fluconazole (Andes et al., 2006). The C. parapsilosis complex showed resistance rates of 14.3 and 5.3% to fluconazole and voriconazole, respectively, both of which were higher than those reported in 20 consecutive years of surveillance data from the SENTRY Antimicrobial Surveillance Program (Pfaller et al., 2019). C. tropicalis exhibited resistance rates of 36.2 and 34.8% to fluconazole and voriconazole, respectively, with rates exceeding 50% in 2021. These values are notably higher than those reported by the CHIF-NET (fluconazole 21.0% vs. voriconazole 21.4%) (Xiao et al., 2018a). As with the threat posed by the emergence and rapid spread of C. auris, the increasing azole resistance rate of C. tropicalis also warrants serious attention. The rate of drug resistance in C. tropicalis has been increasing annually, as shown in some studies (Fan et al., 2017). The increase in isolation rates of azole-resistant strains was not associated with the amount of azole antifungal used. However, several mechanisms of resistance to azoles in C. tropicalis have been identified, including mutations and overexpression of the ERG11 gene (Morio et al., 2017), mutations and overexpression of the UPC2 gene (Choi et al., 2016), and overexpression of the MDR1 and CDR1 genes (Fan et al., 2019). The observed increase in fluconazole resistance poses a direct threat to the effectiveness of first-line therapy for C. tropicalis infections. In regions where resistance rates exceed 10%, Infectious Diseases Society of America (IDSA) guidelines recommend avoiding fluconazole for the empirical treatment of BSIs. Our data suggest that local stewardship programs should prioritize the use of echinocandins or voriconazole in high-risk patients (Pappas et al., 2016). In this study, resistance rates to fluconazole and voriconazole were higher than those observed for itraconazole. Further investigation is needed to determine the underlying reasons for this difference. The IDSA guidelines recommend echinocandins as the best treatment for IFDs because resistance to azole antifungal agents has increased. The European Committee on Antimicrobial Susceptibility Testing (EUCAST) has not established a breakpoint for caspofungin because susceptibility results for caspofungin vary significantly across laboratories. The EUCAST also does not recommend using caspofungin MIC values for clinical evaluation. Some laboratories found that certain C. glabrata complex isolates were susceptible to anidulafungin but not to caspofungin when tested using methods such as the E-test and Sensititre YeastOne for sensitization. This discrepancy may be related to the variability of the in vitro assay for caspofungin (Arendrup and Pfaller, 2012; Eschenauer et al., 2014). Therefore, the EUCAST recommends using anidulafungin and micafungin as markers for caspofungin susceptibility. Until 2018, when the first case of echinocandin-resistant C. glabrata infection was reported in China. The main mechanism of its resistance to echinocandins is a mutation in FKS1/FKS2, and the FKS2 E655K mutation (Fks2 HS1) is a recently identified echinocandin resistance site (Xiao et al., 2018b). Therefore, when testing for caspofungin, the result can be directly reported as “Sensitive.” When an “Intermediate” or “Resistant” result is obtained, it is necessary to confirm the result using the following methods: 1. Additional testing for micafungin or anidulafungin; 2. DNA sequencing analysis, including screening for FKS1 point mutations in all Candida species and FKS2 point mutations specifically in C. glabrata; and 3. sending the isolate to a reference laboratory for further confirmation. In addition, Candida spp. should report resistance to all echinocandins (including caspofungin) if they are resistant to anidulafungin or micafungin or carry an FKS point mutation.

The preferred treatment for Cryptococcus infections is amphotericin B (Iyer et al., 2021). In our study, 8.3% (14/169) of Cryptococcus strains were classified as non-wild-type for amphotericin B, which is inconsistent with the results of a global drug sensitivity study of 3,590 novel Cryptococcus strains conducted by Espinel-Ingroff et al. (2012). A total of five Cryptococcus strains exhibited non-wild-type susceptibility to amphotericin B (four strains with an MIC = 1 μg/mL; one strain with an MIC = 2 μg/mL). This reduced susceptibility may reflect prolonged amphotericin B exposure in patients with disseminated cryptococcosis, where antifungal pressure selects for resistant subpopulations (Córdoba et al., 2016). In addition, since most laboratories use Sensititre YeastOne, which is not the gold standard (broth microdilution method), this may result in the observation of higher levels of not-wild-type strains.

Our findings reveal a critical geospatial divergence in both fungal pathogen distribution and resistance profiles. Given the high prevalence of azole-resistant C. tropicalis and the emergence of C. auris, empirical treatment with echinocandins should be considered for high-risk patients, particularly in settings where resistance rates exceed established thresholds. Accurate knowledge of the regional epidemiology of pathogenic fungi and suspected fungal BSIs is essential for the early selection of appropriate empirical antifungal treatment or intervention, even before definitive pathogenic evidence is obtained.