Aspergillus species are saprophytic molds widely distributed throughout the environment and are easily transported in the air and inhaled into the airway due to the small size of spores (Chabi et al., 2015). Aspergillus spp. can cause many human diseases, ranging from non-invasive allergic bronchopulmonary aspergillosis (ABPA) and chronic pulmonary aspergillosis to invasive aspergillosis (IA), and their clinical manifestations and prognosis vary widely (Cadena et al., 2016). IA is a life-threatening opportunistic infection associated with high morbidity and mortality rates, occurring mainly in immunocompromised patients, such as those with organ transplants or hematological malignancy and those receiving certain types of chemotherapy or immunosuppression therapy (Verweij et al., 2016). Aspergillus fumigatus is the most common Aspergillus spp. causing infection in humans, accounting for 70–80% of cases. The incidences of infections with other species, such as Aspergillus flavus, Aspergillus niger, and Aspergillus terreus, have been increasing, especially in immunocompromised hosts (Richardson and Lass-Flörl, 2008).

Triazoles have a broad spectrum of in vitro antifungal activity against molds and are currently the first-line antifungals for the treatment of IA, including itraconazole (ITC), voriconazole (VRC), and posaconazole (POS). Caspofungin (CAS) and amphotericin B (AMB) are important therapeutic agents for the systemic treatment of refractory IA as well as empirical or prophylactic therapy (Walsh et al., 2008). Despite advances in IA treatment, the mortality rates remain high, especially in immunosuppressed hosts. Antifungal resistance development is one of the major threats (Verweij et al., 2016). Resistance has emerged during the past decade, and the prevalence is increasing in some areas of the world. Therefore, identifying strains of Aspergillus spp. with different susceptibilities has important implications for understanding the susceptibility trend and selecting the correct antifungal agents.

Antifungal susceptibility testing of Aspergillus spp. has been standardized by both the Clinical and Laboratory Standards Institute (CLSI) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) (CLSI, 2017). Breakpoints are used to determine whether the microorganisms are susceptible or resistant to the tested antifungals, which is an important basis for clinicians to select antifungal agents for the treatment of pathogenic infections. Breakpoints include clinical breakpoints (CBPs) and epidemiological cutoff values (ECVs). Currently, CBPs based on the minimum inhibitory concentration (MIC) distributions, pharmacokinetic and pharmacodynamics (PK/PD) parameters, animal data, and clinical outcomes for molds have not been established by the CLSI except for the CBP of VRC for A. fumigatus (CLSI, 2020). In the absence of sufficient data such as (PK/PD) parameters and clinical outcomes, CBPs are unable to be established. ECVs are used to evaluate the susceptibility of strains, to distinguish wild-type (WT) strains from strains with decreased susceptibility or acquired resistance mechanisms, and to monitor resistance development. ECVs for AMB, ITC, VRC, POS, CAS, and Aspergillus spp. have been defined by the CLSI and EUCAST using the broth microdilution (BMD) method (Espinel-Ingroff et al., 2010, 2011a,b; CLSI, 2018).

In the present study, we analyzed the antifungal activities of AMB, triazoles, and CAS against an extensive, geographically diverse collection of 706 Aspergillus spp. isolates from a clinical laboratory in China collected from 1999 to 2019. Furthermore, we applied CLSI ECVs to detect the emergence of resistant isolates. More importantly, WT distributions and tentative ECVs for AMB, ITC, VRC, POS, and CAS against A. fumigatus and A. flavus were proposed for the first time based on the gathered MIC/minimum effective concentration (MEC) data obtained from 20 years of antifungal susceptibility surveillance in China.

In the present study, A. fumigatus was the most frequently isolated species among 706 Aspergillus strains, accounting for 63% of the isolates collected in this study, similar to the rates observed in several epidemiological studies in other countries (Krishnan et al., 2009; Alastruey-Izquierdo et al., 2013a). A. flavus was the second most commonly isolated species, which was in accordance with previous studies in the United States (Balajee et al., 2009), Europe (Alastruey-Izquierdo et al., 2013b), and Brazil (Negri et al., 2014). A. terreus and A. niger were the third and fourth most common isolates, respectively. However, A. terreus and A. niger were reported as the second most common species in Austria (Lackner et al., 2016) and Korea (Heo et al., 2015), respectively, indicating that there are geographical variations in the prevalence of different species. Additionally, sputum was the most common source of the clinically isolated Aspergillus spp. over the 20 years. A. fumigatus and A. flavus were mainly isolated from respiratory tract specimens. However, A. terreus and A. niger were mainly isolated from ear specimens.

The results of in vitro susceptibility testing listed in Table 2 showed that AMB was the most active against 100% of A. fumigatus isolates and that triazoles were effective against 95.7% of A. fumigatus isolates. In agreement with our finding, the susceptibility profiles of 159 clinical A. fumigatus isolates collected from different areas in China also showed that all isolates were susceptible to AMB and that 95.6% were susceptible to triazoles (Deng et al., 2017). A retrospective surveillance study in Portugal also indicated that AMB MICs of all A. fumigatus isolates were ≤2 μg/ml, while ITC, VRC, and POS were effective against 95.8, 97.4, and 84.7% of A. fumigatus isolates, respectively (Pinto et al., 2018). Gomez-Lopez et al. (2003) also demonstrated that AMB inhibited all A. fumigatus isolates at a concentration of 2 μg/ml. Likewise, the susceptibility result of a study from multiple hospitals in Shanghai, China, also showed that AMB was more effective than triazoles against A. fumigatus (Xu et al., 2020). Therefore, AMB appears to be the most active antifungal agent against A. fumigatus. This may be due to the fact that AMB is a polyene fungicidal agent with excellent property of high activity, while azole compounds inhibit fungal growth (Wiederhold, 2017). The fungicidal ability of AMB is higher than that of triazoles (Perlin et al., 2017). On the other hand, since the clinical use of AMB is limited due to nephrotoxicity, triazoles are recommended as the first-line therapy for IA, leading to the emergence of triazole resistance (Perlin et al., 2017; Wiederhold, 2017).

As shown in Table 4, the mutations in the cyp51A gene are the most common reasons conferring resistance to triazoles in A. fumigatus. Among 19 non-WT A. fumigatus strains, six types of the cyp51A mutations (M220I, G54W, G54R, TR₃₄/L98H, TR₃₄/L98H/S297T/F495I, and TR₄₆/Y121F/T289A) were observed. The emergence of A. fumigatus harboring TR₃₄/L98H/S297T/F495I mutation in China was first reported from a global surveillance study conducted in 2008 and 2009 (Lockhart et al., 2011). Chen et al. (2015) reported the first isolation of TR₄₆/Y121F/T289A A. fumigatus strain from a Chinese patient without triazole exposure in 2015, which was related to two strains from clinical and environmental samples obtained in the Netherlands by genetic analysis based on microsatellite genotyping. Clinical and environmental A. fumigatus strains harboring TR₃₄/L98H and TR₃₄/L98H/S297T/F495I mutations were isolated in China; genetic typing and phylogenetic analysis showed that TR₃₄/L98H isolates had a clonal expansion worldwide, while the TR₃₄/L98H/S297T/F495I isolates harbored a distinct genetic background with resistant isolates from other countries (Chen et al., 2016). These mutations were also reported in many countries (Sharma et al., 2015; Chowdhary et al., 2017; Leonardelli et al., 2017). Additionally, one strain without mutations in the cyp51A gene was isolated in 2019 in this study. Recent studies have demonstrated the occurrence of triazole-resistant A. fumigatus without cyp51A gene mutations (Hagiwara et al., 2018; Sharma et al., 2019). Therefore, the strain may have non-cyp51A-mediated resistance mechanisms, and further research is worth conducting.

As shown in Table 2, triazoles were likely to be more effective than AMB against A. flavus, A. terreus, and A. niger. Since the total number of non-fumigatus strains isolated was low during the 10-year period and the susceptibility result of a single-center study was insufficient, it is difficult to consider triazoles as the potent agents against non-fumigatus strains, although the susceptibility result was in accordance with the previous study showing triazoles were the most active compounds against non-fumigatus species (Gomez-Lopez et al., 2003). However, the emergence of AMB resistance in A. flavus and A. niger strains should raise concerns. A. terreus is considered intrinsically resistant to AMB (Lass-Flörl et al., 2005; Blatzer et al., 2015).

As shown in Figure 2, the 5-year proportion of resistant A. fumigatus isolates was 5.06% for 1999–2004, 6.25% for 2005–2009, 0.58% for 2010–2014, and 8.43% for 2015–2019. The increase in resistance for 2015–2019 compared with 2010–2014 was statistically significant (χ² = 8.859, p < 0.05). In various time periods, the prevalence of resistance was inconsistent and dynamic. Among 19 non-WT A. fumigatus, nine strains of BMU02731, BMU02810, BMU02816, BMU02998, BMU03908, BMU03941, BMU03942, BMU04053, and BMU04758, harboring the single point mutations in the cyp51A gene, were isolated from one patient with pulmonary aspergilloma and chronic cavitary tuberculosis before 2009 (Chen et al., 2005). The patient had a history of long-term ITC therapy (Chen et al., 2005). The presence of a cavity allows for asexual sporulation to occur; and with chronic azole exposure, numerous spontaneous mutations may occur in the conidia (Rivero-Menendez et al., 2016; Buil et al., 2019). Hence, due to ITC exposure, nine triazole-resistant strains were continuously isolated from the patient, resulting in high resistance rates in 1999–2004 (5.06%) and 2005–2009 (6.25%). Additionally, strains harboring TR₃₄/L98H and TR₄₆/Y121F/T289A were gradually isolated after 2009. Since TR₃₄ or TR₄₆ mutations are believed to be primarily driven by the use of azole fungicides in the environment (Chen et al., 2016; Rivero-Menendez et al., 2016), patients are infected by inhaling resistant conidia that already harbor azole resistance mechanisms under environmental exposure, presumably the consequence of exposure to azole fungicides used in agriculture (Chen et al., 2016; Rivero-Menendez et al., 2016).

Although the prevalence of resistance showed fluctuations, the overall trends toward decreasing susceptibility to triazoles were observed in A. fumigatus over the past 20 years. Previous studies were consistent with our finding. Nine-year susceptibility trends of 1,789 clinical Aspergillus spp. isolates obtained from more than 60 medical centers worldwide showed that the proportion of non-WT A. fumigatus isolates ranged from 0.7 to 4.0% for ITC, from 1.1 to 5.7% for POS, and from 0.0 to 1.6% for VRC (Pfaller et al., 2011). A 23-year continuous surveillance study for A. fumigatus in the Netherlands also observed an increasing trend of triazole resistance in clinical A. fumigatus isolates (Buil et al., 2019). Hence, the increasing prevalence of triazole resistance is a worldwide concern (Choukri et al., 2015; Chowdhary et al., 2017; Buil et al., 2019). In 2018, Chen et al. (2020) reported that 10.2% of A. fumigatus isolates collected from 63 soil samples from agricultural farms or greenhouses were azole resistant, and 18 resistant strains were cultured from soil samples acquired from strawberry fields. In a study in the United Kingdom, azole-resistant A. fumigatus isolates were identified in several products that originated from China, including tea and pepper (Chen et al., 2020). The total amount of azole fungicides used in agriculture accounted for more than one-third of fungicides used in China during 2013–2016 (Zhang et al., 2011). All these findings suggested that more resistant A. fumigatus, harboring the TR₃₄ or TR₄₆ mutations, were isolated from the environment in China. Similarly, azole-resistant A. fumigatus isolates have been recovered from the environment in an increasing number of countries (Choukri et al., 2015; Rivero-Menendez et al., 2016; Verweij et al., 2016; Perlin et al., 2017), with the frequency of azole resistance varying widely. Additionally, it has been reported that more resistant strains from China harbored the identical genetic background with resistant isolates from other countries by genetic analysis based on microsatellite genotyping (Chen et al., 2016; Deng et al., 2017), suggesting that the spread of resistant strains in the environment is increasing worldwide. Considering that the TR₃₄/L98H was the most common type of mutation in this study, we speculated that the decreasing trend in susceptibility of A. fumigatus to triazoles shown in Figure 2 may be correlated with the increased exposure and transmission of resistant strains in the environment.

The possible reasons for the fluctuations in the prevalence of resistance in this study are as follows. Firstly, the number of resistant strains in the environment is increasing. Secondly, the triazole exposure in patients is increasing. Thirdly, the number of isolates collected was low, and the susceptibility result of a single-center study was limited. Finally, all isolates of Aspergillus spp. were recovered from clinical samples only.

Additionally, since only four isolates of A. flavus were non-WT to AMB, the trend in susceptibility of A. flavus to AMB seemed to be stable over the 10-year period, which was inconsistent with other studies showing that more A. flavus isolates were resistant to AMB (Gonçalves et al., 2013; Reichert-Lima et al., 2018; Rudramurthy et al., 2019). Given that A. flavus was the second leading species in our collection, continuous susceptibility surveillance will be still needed.

Cryptic species were found to account for over 10% of Aspergillus spp. clinical isolates and to be resistant to at least one of the antifungals available in Brazil, Spain, and Korea (Alastruey-Izquierdo et al., 2013a; Negri et al., 2014; Won et al., 2018). While only four isolates of cryptic species were identified, including A. sydowii, A. penicillioides, A. tamarii, and A. undagawae, and all were susceptible to the antifungals tested. Albeit the number of cryptic species was low in this study, attention should be paid to the emergence of resistant cryptic species.

Considering that the CBPs of Aspergillus spp. for common antifungals have not been determined except for the CBP of VRC for A. fumigatus (CLSI, 2020), the ECVs proposed for A. fumigatus and A. flavus were calculated. The proposed ECVs were in agreement with those defined by the CLSI except the ECV of ITC for A. flavus (Espinel-Ingroff et al., 2010, 2011a,b; CLSI, 2018). The susceptibility of A. flavus isolates to ITC was re-evaluated using the proposed ECV, showing that the prevalence of non-WT A. flavus isolates reduced significantly from 6 to 0.6%. The ECVs, established by CLSI using statistical methods combined with the “eyeball” method, must include MIC distributions (≥100 MIC results per species and antifungal agent) from multiple (≥3) independent laboratories (Espinel-Ingroff et al., 2010, 2011a,b), while the ECVs defined in this study were based on the gathered MIC data of strains collected from multiple geographic locations and preserved in a single center in China over 20 years. The proposed ITC ECV of A. flavus was higher than that defined by the CLSI, probably because the susceptibility of A. flavus isolates to ITC in China may be lower than that of other countries. On the other hand, the susceptibility result of a single-center study could not fully represent the susceptibility of A. flavus strains nationwide. Therefore, multicenter surveillance data in China need to be collected to establish accurate ECVs in the future.

In conclusion, this is the first 20-year retrospective surveillance study for clinically isolated Aspergillus spp. in China. Several species of Aspergillus spp. have developed drug resistance and even multidrug resistance, and the resistance mechanisms of non-WT A. fumigatus strains have been also described. Moreover, the proposed ECVs were established for A. fumigatus and A. flavus, which will be an essential step in developing CBPs and be useful for resistance surveillance in China. Hence, the significance of this study is to establish continuous susceptibility surveillance and determine antifungal susceptibility trends, providing a basis for clinical medication.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

XY, WC, TL, JT, WxL, YS, QW, HX, LL, YZ, QqW, and YgS contributed to the species identification and antifungal susceptibility testing. WL designed all the experiments of the study. XY, QW, and YZ contributed to molecular biology experiments. XY and WL contributed to writing and reviewing the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.