A seven-year surveillance study of the epidemiology, antifungal susceptibility, risk factors and mortality of candidaemia among paediatric and adult inpatients in a tertiary teaching hospital in China

Background There are no current national estimates of the candidaemia burden in China, and epidemiological candidaemia data from the underdeveloped region of China are lacking. Methods A 7-year retrospective study was carried out to analyse the prevalence, species distribution, antifungal susceptibility, risk factors and inpatient mortality of candidaemia among paediatric and adult patients in a regional tertiary teaching hospital in China. Results During the seven-year study period, a total of 201 inpatients with candidaemia were identified. The median age of the patients was 65 years (range, 1 day to 92 years), and 114 of the patients (56.7%) were male. The mean annual incidence of candidaemia was 0.26 cases per 1000 admissions (0.42 cases per 1000 paediatric admissions vs 0.24 cases per 1000 adult admissions, P < 0.05). Candida albicans was the most common fungal species (81/201, 40.3%) in all patients, Candida glabrata was the most common fungal species (18/35, 51.4%) in paediatric patients. Most isolates were susceptible to flucytosine (99.0%) and amphotericin B (99.0%), and the activity of antifungal agents against Candida species was no significant difference in satisfaction between paediatric and adult patients (P > 0.05). The all-cause mortality rate was 20.4% (paediatric patients: 11.4% vs adult patients:22.3%, P > 0.05). Fewer univariate predictors of poor outcomes were identified for paediatric patients than for adult patients (4 vs 11 predictors). Respiratory dysfunction and septic shock were independent predictors of 30-day mortality for all patients. Conclusions The epidemiological data of candidaemia in paediatric and adult patients are only different in the distributions of Candida species and the mean annual incidence of candidaemia. Flucytosine and amphotericin B can be used as first-choice agents when no antifungal susceptibility test results are available.


Background
Candidaemia is the most common fungal disease among hospitalised patients worldwide and is the fourth to tenth most common bloodstream infection (BSI) in most population-based studies [1,2]. Candidaemia is associated with significant morbidity and mortality [3]. The main risk factors for candidaemia include critical illness, a long intensive care unit (ICU) length of stay, haematologic malignant disease, solid-organ transplantation, solid-organ tumours, low birth weight in neonates and preterm infants, broad-spectrum antimicrobial agent use, central venous catheterization (CVC), total parenteral nutrition, haemodialysis, abdominal surgery, and aggressive chemotherapy [1]. With the increase in related research, reports have shown that the incidence of candidaemia is age-specific, with maximum rates observed in those with older age (over 65 years) [1,4,5].
More than 40 Candida species can cause candidaemia in humans [6]. Five species of Candida (Candida albicans, Candida glabrata, Candida parapsilosis, Candida tropicalis and Candida krusei) are the most common species and account for more than 90% of all the isolates [2]. The variability in the relative proportions of Candida isolates has been associated with clinical condition or risk factors such as age, underlying comorbidities, the extensive use of antifungal agents and geography. Candida albicans is the primary cause of candidaemia and one of the most common species in many countries, Candida glabrata is the second or third most common species in the USA and Europe, and Candida parapsilosis is predominant in neonates in South America, southern Europe and Asia [2]. The global incidence of candidaemia varies from 0.3 to 5 per 1000 admissions according to geographical region, local epidemiology, age and other factors [7]; the 30-day mortality among all patients with candidaemia has been reported to be between 22 and 70% [8], and the cost of candidaemia treatment has been reported to be US $40,000 per patient [1,9,10].
In China, the epidemiology of candidaemia varies widely among different areas [11]. Epidemiological surveillance of candidaemia has focused on ICUs and single centres in China, and national surveillance systems are usually absent. Most of the existing epidemiological surveillance of candidaemia has focused on adults or children, and little information about general populations (including neonates, children and adults) is known. Therefore, in the present study, we performed a seven-year retrospective study to evaluate the epidemiology, antifungal susceptibility, risk factors and mortality of candidaemia among all inpatients in a tertiary teaching hospital in China.

Patient data collection
We conducted a retrospective observational study of electronic laboratory records. The fungal specimen data were collected from inpatients with candidaemia in the Affiliated Hospital of Southwest Medical University (Luzhou, China), which is a 3200-bed tertiary care teaching hospital with 43 wards and approximately 120,000 annual admissions, from January 2013 to December 2019. The diagnostic criteria of candidaemia were based on the guidelines for the diagnosis and treatment of Candidiasis: the expert consensus issued by the Chinese Medical Association [12]; these criteria were also in accordance with the European Society of Clinical Microbiology and Infectious Diseases (ESCMID)* guidelines for the diagnosis and management of Candida diseases 2012 [13,14] and the Infectious Diseases Society of America (IDSA) Guidelines for the Management of Candidiasis: 2016 Update [15]. For each patient, only the first episode was included in our analysis. Patient cultures with two or more Candida species were excluded from the analysis, and all data were collected from electronic medical records. The following data were retrospectively collected from all patients: demographic characteristics, underlying comorbidities, Candida species, susceptibility to antifungal agents and mortality. Data on the following risk factors associated with candidaemia were also collected: gestational age and weight of neonates, indwelling central vascular catheter, mechanical ventilation, systemic corticosteroid treatment (a dose equivalent to prednisone 10 mg/d for at least 14 days), total parenteral nutrition, chemotherapy, abdominal surgery, ICU admission, neutropenia (absolute neutrophil count < 500 cells/μl), concomitant bacterial infections, septic shock, haemodialysis, broad-spectrum antibiotic use and treatment with antifungal agents. The study protocol was approved by the ethics committee of the hospital (Project No. KY2020043). The need for informed consent was waived by the Clinical Research Ethics Committee.

Microorganism identification and antifungal susceptibility
According to the manufacturer's instructions, blood was inoculated into both aerobic and anaerobic BacT/AlerT 3D vials (Bruker Diagnostics Inc., USA). All positive cultures were manually sampled and inoculated onto CHROMagar Candida medium (CHROMagar Company, France) to ensure viability and purity. The identification of all species was confirmed by a MicroScan WalkAway 96 Plus System (Siemens, Germany) and Microflex LT (Bruker Diagnostics Inc., USA) matrix-assisted laser-desorption/ionization timeof-flight (MALDI-TOF) mass spectrometry (MS) system.
Antifungal susceptibility tests for fluconazole (FCA), itraconazole (ITR), voriconazole (VRC), flucytosine (5-FC) and amphotericin B (AMB) were performed for all Candida strain isolates by using an ATB FUNGUS 3 kit (bioMérieux, France). The minimal inhibitory concentrations (MICs) of the antifungal agents were judged by visualization in our laboratory according to the manufacturer's instructions. The quality control strains were C. parapsilosis ATCC 22019 and C. krusei ATCC 6258. The results were interpreted using the Clinical and Laboratory Standards Institute M27-A3 microbroth dilution method.

Statistical analyses
The data were analysed using Microsoft Excel (version 2016, Redmond, USA) and IBM SPSS software version 24 for Windows (IBM, Armonk, USA). Categorical data were compared using chi-square or Fisher's exact tests. Continuous data were analysed using Student's t-test or the Mann-Whitney U test. Multivariable logistic regression analysis was performed to identify independent predictors of candidemia and 30-day hospital mortality. Biologically plausible variables with a value of P < 0.1 according to the univariate analyses were included in the multiple logistic regression model. Statistical significance was determined using two-tailed tests, and P < 0.05 was considered statistically significant.
The results of in vitro susceptibility testing of Candida strain isolates are summarized in Table 3. All isolates were highly susceptible to AMB (99.0%) and 5-FC (99.0%). The resistance rates of ITR, VRC and FCA were 24.9 19.4 and 18.5%, respectively. C.tropicalis had the highest antifungal agent resistance rate among the Candida species and was resistant to FCA (39.3%), ITR (39.3%) and VRC (42.9%). The activity of antifungal agents against Candida species was not significantly different in terms of satisfactory outcomes between paediatric and adult patients (P > 0.05). The detailed data are shown in Table 3.
The univariate predictors of poor outcomes due to candidaemia are shown in Table 4. For paediatric patients with candidaemia, the variables associated with 30-day mortality were as follows: length of hospital stay, respiratory dysfunction, chronic/acute renal failure and septic shock. For adult patients with candidaemia, the variables associated with 30-day mortality were as follows: age, length of hospital stay, respiratory dysfunction, pulmonary infection, cardiovascular disease, chronic/ acute renal failure, other invasive catheters, mechanical ventilation, septic shock, C. albicans infection, concomitant bacterial infection and haematologic (nonmalignant) disease. The results of the multivariate analysis are listed in Table 5. Because the total numbers of paediatric patients (35 patients) and deaths (3 patients) were very small, multivariable logistic regression analysis was not performed for paediatric patients. Respiratory dysfunction and septic shock were independent predictors of 30-day mortality in all patients and adult patients. The length of hospital stay was a protective factor for 30-day mortality in all patients and adult patients, and other invasive catheters were only the protective factor for 30day mortality in all patients. The prognostic factors for 30-day mortality in all patients and adult patients were almost the same, and the independent predictors were the same, with no significant differences (Table 5). In addition, we compared the independent protective factors and independent risk factors in this study with those in other studies and found that the protective factors (length of hospital stay and other invasive catheters) in our study were different from those in other studies, while the independent risk factor of septic shock in our study was also identified in some others studies, but the independent risk factor of respiratory dysfunction was not reported in other studies (Table 6).

Discussion
This was a 7-year retrospective study of candidaemia in a regional tertiary teaching hospital in Southwest China. We not only analysed the epidemiological characteristics, including the basic information of patients, underlying comorbidities, risk factors, the distributions of Candida species, antifungal agent use, antifungal agent susceptibility results and patient outcomes, but also performed epidemiological comparisons between paediatric patients and adult patients. To our knowledge, this is   Our data showed that there was no significant difference in the sex ratio, length of hospital stay or mortality between adult and paediatric patients (P > 0.05). However, the proportions of underlying comorbidities in paediatric patients, including pulmonary infection, neurological diseases, congenital malformations/ syndromes and haematologic (nonmalignant) disease, were higher than those in adult patients (P < 0.05), and the other proportions in adult patients were similar or higher than those in paediatric patients (Table 2). Among the risk factors, only CVC, other invasive catheters and abdominal surgery in adult patients had higher risks than those in paediatric patients (P < 0.05), and other risk factors in children had higher or similar risks as those in adult patients (Table 2). Fewer univariate predictors of poor outcomes were identified for paediatric patients than for adults patients (4 vs 11 predictors),   Table 4. This situation has not been clearly shown in other studies, and more epidemiological investigations are needed for confirmation. The incidence of candidaemia among paediatric patients was significantly higher than that among adults (P < 0.05) ( Table 2); however, no significant difference in mortality was found between paediatric patients and adult patients (P > 0.05) ( Table 2) in contrast to other studies [16,17].
Our data showed that the median age of patients with candidaemia and the proportion of males were similar to those in other studies [8,[18][19][20][21][22][23]. Moreover, our study showed that the patients with candidaemia were hospitalised mostly in internal medicine wards, which is different from other studies that reporting hospitalisation mainly in ICU wards [8,22,[24][25][26][27], but similar to other studies [28][29][30][31]. This phenomenon may be related to the demographic characteristics of the inpatients in our hospital, most of whom had more than two underlying diseases and were hospitalised in internal medicine wards. However, the incidence of candidaemia was still the highest in the ICU, similar to other studies [8,[30][31][32][33][34].
In accordance with other studies [17-19, 24, 25, 30, 32, 35, 36], C. albicans was the most common cause of candidaemia in the whole hospital, but the proportion of non-C. albicans infections was higher than that of C. albicans infections. Moreover, the proportions of C. glabrata in surgical, internal medicine and paediatric wards were the highest, which is different from other studies in China [18,19,[35][36][37] but similar to other studies in other countries [4,22,27,29,32]. This may be due to the large number of elderly patients and the increasing use of azole antifungal agents.  Our data showed that the incidence of candidaemia increased from 0.20 episodes/1000 admissions in 2013 to 0.37 episodes in 2016 and then dropped to 0.26 between 2017 and 2019. The change in the annual incidence rate was mainly due to the change in the incidence rate in paediatric patients. The reasons may be due to the gradual easing of restrictions of China's two-child policy since 2013. The number of geriatric pregnant women has increased annually, resulting in an increase in the incidence of neonatal diseases. The change trend was similar to that reported by Oeser et al. [38]. The overall morbidity and 30-day mortality in ICUs and hospitals in   this study were similar to those in another hospital in this region of China [18], but lower than those in hospitals in other regions of China [35,37] and other countries [5,8,16,20,21,23,25,30]. The overall mortality rate of candidaemia has been reported to be 20-49% globally [39], and the mortality rate was 20.4% in our hospital, which is low compared to the global rate. This may be because the demographic characteristics and underlying diseases of patients in this region are different from those in other regions or countries, and few severe patients were admitted to our hospital. With regard to resistance, resistance to FCA, ITR and VRC were common in C. albicans and non-C. albicans species (Table 3). In our study, AMB and 5-FC were highly active against all Candida species. In paediatric patients, the resistance rate of ITR was higher than that in adult patients, but the resistance rates of FCA and VRC were lower than those in adult patients; however, and the resistance rate of Candida species was no significant difference in satisfaction between paediatric and adult patients(P > 0.05). Moreover, FCA was highly active against all Candida species in paediatric patients and could be used in paediatric patients with candidaemia as a first-line agent. In the whole hospital, the resistance rate to azole was higher than those reported in other regions [18,19,36] and countries [17,19,25,29,30,34]. This may be related to the long-term use of empirical prophylactic drugs by clinicians. Therefore, it was necessary to conduct an epidemiological analysis of antifungal agent susceptibility and guide clinicians to choose the rational antifungal agents to avoid the continuous increase in resistance rates.
In this study, septic shock was an independent predictor of 30-day mortality; which has been reported in many other studies [18,35]. However, the other factors reported here have rarely been reported in other studies [35,[40][41][42][43], possibly because the demographic characteristics, underlying diseases and risk factors of the patients in our study were different from those in other studies; which may be the reason that the independent predictors and protective factors in this study were different from those in other studies [5,35,[40][41][42][43]. The independent predictors and protective factors in different regions and countries are shown in Table 6.
The limitations of this study must be acknowledged. First, this was a single-centre retrospective study, and the total number of patients(166 adult and 35 paediatric patients) were small. Our data might be influenced by the number of patients, the level of medical intervention, and the distribution of patient types. Second, due to technical limitations of the clinical microbiology laboratory and the impact of hospital policies, no data on echinocandins were available in our hospital. Therefore, the results may not be generalizable to all patients with candidaemia in China.

Conclusion
C. albicans was the main Candida species, but C. glabrata has become the second most common species in this region. FCA was the main antifungal agent for paediatric patients. AMB and 5-FC were highly active against all Candida species. The morbidity and mortality rates in elderly patients were the highest. Respiratory dysfunction and septic shock were independent predictors of 30-day mortality. Further multi-centre studies on candidaemia in different geographical regions in all patients should be conducted to help infection specialists assess the distribution and trends in patients with suspected fungal infections.