Comparison of the incidence, clinical features and outcomes of invasive candidiasis in children and neonates

Background Invasive candidiasis differs greatly between children and neonates. We aimed to investigate the different therapeutic approaches and their effects on treatment outcomes of these two groups. Methods Episodes of neonatal invasive candidiasis were compared with non-neonatal pediatric episodes during a 12-year cohort study. Clinical isolates were documented by matrix-assisted laser desorption/ionization-time of flight mass spectrometry and DNA sequencing, and antifungal susceptibility testing was performed. Results A total of 342 episodes of invasive candidiasis (113 neonatal and 229 non-neonatal pediatric episodes) in 281 pediatric patients (96 neonates and 185 children) were identified. Candida albicans was the most common pathogen causing invasive candidiasis in neonates and children (47.8% vs. 44.1%). The antifungal susceptibility profiles were not significantly different between neonates and children. More neonates received amphotericin B as therapy, whereas more children received fluconazole or caspofungin. Compared with children, neonates had a significantly longer duration of fungemia, higher rates of septic shock (34.5% vs. 21.8%; P = 0.013), sepsis-attributable mortality (28.3% vs. 17.5%; P = 0.024) and in-hospital mortality (42.7% vs. 25.4%; P = 0.004) than children. Independent risk factors for treatment failure of invasive candidiasis were septic shock (odds ration [OR] 16.01; 95% confidence interval [CI] 7.64–33.56; P <  0.001), delayed removal of intravenous catheter (OR 6.78; 95% CI 2.80–17.41; P <  0.001), renal failure (OR 5.38; 95% CI 1.99–14.57; P = 0.001), and breakthrough invasive candidiasis (OR 2.99; 95% CI 1.04–8.67; P = 0.043). Conclusions Neonatal invasive candidiasis has worse outcomes than non-neonatal pediatric candidiasis. Neonatologists and pediatricians must consider age-specific differences when developing treatment and prevention guidelines, or when interpreting studies of other age groups.


Background
Candida species are the fourth most common cause of nosocomial infection and are the leading cause of invasive fungal infection among hospitalized patients [1,2]. Invasive candidiasis deserves greater attention because it is associated with a high mortality rate, especially in severely ill patients [3][4][5]. Recent population-based surveillance studies have shown an increased incidence of invasive candidiasis in intensive care units (ICUs) during the past decade [6,7]. An increase of susceptible hosts who receive intensive care or immunosuppressive therapies and the widespread use of broad-spectrum antibiotics may account for the increase of invasive Candida infections [8][9][10]. Furthermore, uses of antifungal drugs such as azoles for prophylaxis and echinocandins for treatment are reported to be associated with a continuous shift from C. albicans to various non-albicans Candida species [11,12].
The microbiological and clinical characteristics of invasive Candida infections vary widely among different geographic areas, patient characteristics and ages, and institutions [13,14]. Although some studies concluded that the mortality of candidemia was higher in adults than in children [15], a recent study found a poor prognosis among infants (<1 year of age) and elderly patients (>60 years) [16]. Furthermore, we recently documented fungemia as an independent risk factor for treatment failure in the neonatal ICU (NICU) [17]. The reported incidence of candidemia in pediatric patients generally ranges between 0.21 and 10.5 cases per 1000 admissions [15,[18][19][20]; however, patients in the NICU, pediatric ICU (PICU), and pediatric wards were not studied as separate and distinct groups [15,21]. In order to clarify and assess unique characteristics of invasive candidiasis in neonates and children, we compared the epidemiology and clinical features of Candida spp. identified by matrix-assisated laser desorption/ionization time-offlight mass spectrometry (MALDI-TOF) causing invasive candidiasis in these two populations.

Methods
We included all hospitalized patients in the Department of Pediatrics, Chang Gung Memorial Hospital (CGMH) from January 2004 through December 2015, for whom ≥1 blood culture and/or sterile site cultures were positive for Candida spp. and who had symptoms, signs, or laboratory findings consistent with fungal infection. We retrospectively reviewed electronic medical records for demographic, clinical and laboratory data for the onset of invasive candidiasis (defined as the day of blood or sterile site collection for culture), and we reviewed risk factors within the preceding 30 days, major comorbidities, complications of invasive candidiasis, treatments and outcomes. The study was approved by the Institutional Review Board and Human Research Ethics Committee of CGMH, and a waiver of informed consent for anonymous data collection was also approved.
Isolation and identification of all Candida spp. isolates in blood and sterile site cultures were performed using a standard API 32C AUX yeast identification kit (bioMérieux SA, Marcy l'Étoile, France) and chromogenic culture media (CHROMagar; Becton Dickinson and Company, USA). Since December 2013, we have used MALDI-TOF (Bruker Biotype, software version 3. 0, Ewing, NJ, USA) and large-subunit (18S) ribosomal RNA gene D1/D2 domain sequencing to re-confirm all these species. Antifungal susceptibility was tested using the Clinical and Laboratory Standards Institute broth microdilution reference method [22]. For uncommon Candida spp., clinical breakpoints are undefined; therefore, isolates that showed minimum inhibitory concentrations (MICs) higher than the epidemiologic cutoff value were considered potentially resistant [23]. We excluded unidentified Candida species and selected only the first isolate recovered from the blood or sterile sites if a patient had several cultures that were positive for the same Candida spp.

Definitions
Invasive candidiasis included candidemia and deepseated candidiasis, which were defined as the recovery of a Candida species from blood or a sterile site, respectively [24,25]. An episode of candidemia was considered to be catheter-related only if the catheter tip culture was positive for the same Candida spp. [26]. Episodes were considered to be separate if they occurred ≥1 month apart. Breakthrough invasive fungal disease was defined as candidemia or positive Candida spp. isolated from a sterile site in a patient who had undergone therapy or prophylaxis with any systemic antifungal drug for ≥3 consecutive days before the index blood culture [8,27]. Invasive candidiasis-attributable mortality was defined when the patient died within 7 days after onset of invasive candidiasis or in the presence of persistent clinical sepsis or persistent candidemia, or if the patient died of candidemia associated complications [27,28]. Combined with the antifungal susceptibility results, treatment failure was defined as an infection that led to attributable mortality or cases of persistent candidemia ≥7 days after initiation of effective antifungal therapy. Patient responses to antifungal therapy following invasive candidiasis were defined according to the consensus criteria of the Mycoses Study Group and the European Organization for Research and Treatment of Cancer [29].

Statistical analysis
Clinical data were analyzed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Variables associated with invasive candidiasis in the NICU were compared with non-neonatal episodes. Univariate analyses were performed using Student's t-test or non-parametric tests as appropriate (for continuous variables) or the chi-square or Fisher's exact tests (for categorical variables). All tests were 2-tailed, and a P value of < 0.05 was considered significant. We performed multiple logistic regression analyses to identify clinical risk factors that were associated with treatment failure of invasive candidiasis. All risk factors that were significant at 0.10 in the univariate analysis were included in the corresponding multivariate analysis.

Incidence and demographic data
Among a total of 20,545 neonatal admissions and 153,372 pediatric non-neonatal admissions (of which 14,018 were PICU admissions), there were a total of 342 invasive candidiasis episodes in 281 children that occurred during the study period; of these, neonatal episodes accounted for 113 episodes (33.0%, in 96 neonates). The incidence rates of invasive candidiasis in neonates (NICU) and non-neonatal pediatric patients were 26.9 episodes per 100,000 inpatient days and 32.6 episodes per 100,000 inpatient days, respectively. However, the incidence rate of invasive candidiasis in the PICU was 147.2 episodes per 100,000 inpatients days, which was significantly higher than the rates in the NICU and in the general pediatric ward (both P < 0. 001). Overall, the annual incidence rates of invasive candidiasis did not change significantly throughout the study period and varied between 27.3 and 36.7 episodes per 100,000 inpatients days (data not shown).
Invasive candidiasis occurred in 152 boys (54.1% of all patients) and 129 girls. The mean age for non-neonatal pediatric patients was 6.2 ± 5.7 years (range, 3 months to 18 years). Overall, 214 (62.6%) episodes of invasive candidiasis occurred in children ≤3 years old (Fig. 1). Most of the invasive candidiasis episodes were primary bloodstream infections (228 episodes, 66.7%), followed by catheterrelated bloodstream infections (69 episodes, 20.2%), and intra-abdominal infections (31 episodes, 9.1%). The sites of isolation and Candida species distributions were not significantly different between the neonatal and nonneonatal groups ( Table 1), except that only two episodes in the NICU were caused by C. tropicalis. C. albicans was the most common Candida species that caused invasive candidiasis in children (45.3%, 155 episodes), followed by C. parapsilosis (27.8%, 95 episodes), C. tropicalis (6.4%, 22 episodes) and C. glabrata (6.1%, 21 episodes). Polyfungal isolates (i.e., two different Candida species yielded on cultures of blood samples that were obtained simultaneously) were recovered from three episodes, and the majority of ascites cultures (in 23 episodes, 74.2%) were polymicrobial isolates that also yielded gram-positive cocci, or aerobic and anaerobic gram-negative bacilli.

Risk factors
The majority of patients had multiple underlying illnesses and other risk factors that have been associated with invasive candidiasis ( Table 2). The majority of neonatal invasive candidiasis cases occurred in very low birth weight infants (birth weight < 1500 g) (98 episodes, 86.7%), and the mean (SD) gestational age in this group was 27.8 ± 3.9 weeks. The most common predisposing factors were use of central intravenous catheter (CVC) (94.2%), use of broad-spectrum antibiotics (91.8%), stay in an ICU (69.3%), receipt of parenteral nutrition (64. 6%), and underlying neurological sequelae (36.0%). For 282 episodes (82.4%), ≥ 4 risk factors and/or underlying illness were identified. However, the underlying illness and major predisposing factors were significantly different between neonatal invasive candidiasis and nonneonatal pediatric episodes. While neonates with invasive candidiasis were significantly more likely to have chronic lung disease and receive total parenteral nutrition, non-neonatal pediatric candidemia were more likely to occur in children with underlying neurological sequelae, cancer treated immunosuppressive agents, neutropenia and artificial devices other than CVC.

Clinical presentations
No patient developed endophthalmitis, endocarditis, or osteomyelitis that was clinically evident, but five had CNS infection and four had an obstructing renal fungus ball during the follow-up period. In addition, 31 had intraabdominal abscesses or peritonitis and four patients had positive Candida isolates from pleural fluids. 36.8% of invasive candidiasis presented with severe sepsis, and 26.0% had septic shock at the onset of sepsis. After effective antifungal treatment, 17.8% had progressive and deteriorated candidiasis, and 14 (4.1%) had disseminated candidiasis. Neonates with invasive candidiasis had significantly higher severity of illness than children based on the surrogate marker of severe sepsis and septic shock (Table 3).
More than half of the episodes (180 episodes, 52.6%) were characterized by fungemia or persistent invasive candidiasis of > 1 day's duration, whereas 165 (48.2%) and 78 (22.8%) were characterized by fungemia or persistent invasive candidiasis of ≥3 days and ≥ 7 days, respectively. The mean duration of fungemia was 4.4 days (± 6.2 days). The longest duration of fungemia was 34 days, which occurred in a patient who had end-stage renal disease and who received long term hemodialysis.

Treatment and outcomes
Of the 342 episodes, 332 (97.1%) were treated with an antifungal agent, and there were 41 episodes of breakthrough invasive candidiasis. Ten episodes (2. 9%) were not treated because of the patient's death before or at the time of the diagnosis was established. Antifungal therapy was initiated after a mean of 1. 81 days (range, 0-6) following the acquisition of the first diagnostic blood and/or sterile site culture and was significantly later in neonates than in children (2. 1 ± 1.3 vs. 1.7 ± 1.4 days, P = 0.009). The mean duration of all antifungal therapy per episode was 18. 5 days (range, 1-68). Of those 332 episodes for which an antifungal agent was used, 151 episodes (45.5%) had modification of the antifungal regimens during the treatment course, mainly due to the patient's poor response to initial antifungal therapy (101 episodes, 66.9%), suspicious antifungal resistance after confirmation of Candida spp.(36 episodes, 23.8%), or no reason was documented (14 episodes, 9.3%). Among the antifungal regimens for treatment, fluconazole was the most commonly prescribed initially (62.3%), followed by amphotericin B (24.7%) and caspofungin (4.5%). However, the final treatment regimens were fluconazole/Voriconazole (39.5%), amphotericin B (29.2%) and echinocandin (28.9%), with significant differences between neonates and children (Table 3).
Catheter removal was done within 3 days after illness onset in only one-third of patients with invasive candidiasis (107 episodes, 32.2%), and in 22 episodes, the candidemia resolved only after removal of the intravenous catheter. Neonates with invasive candidiasis had a longer period of fungemia than children, and a higher rate of treatment failure was also noted (31.0% vs. 19.7%, P = 0. 015) ( Table 3). Invasive candidiasis in neonates was associated with a significantly higher rate of sepsisattributable mortality than that in children (28.3% vs. 17.5%, P = 0.024). After invasive candidiasis, neonates had a higher rate of in-hospital mortality than children (42.7% vs. 25.4%, P = 0.004, and P = 0.005 by log rank test [ Fig. 2]).

Susceptibility studies
In vitro susceptibility to various commonly prescribed antifungal agents in our hospital was determined for 295 isolates ( Table 4). The rate of fluconazole-R or S-DD Candida was 14.6% (43 of 295 isolates) overall. The antifungal susceptibility profiles of Candida spp. in neonates were not significantly different between those in non-neonatal pediatric episodes. No trend toward higher minimum inhibitory concentrations was noted when earlier isolates (i.e., isolates recovered during 2004-2009) were compared with those obtained later (i.e., those recovered during 2010-2015).

Discussion
The epidemiology and choice of therapy for candidemia or invasive candidiasis are rapidly changing, and vary  [15,28,30]. The crude mortality rates are generally lower in younger (pediatric) than older (adult) patients with candidemia regardless of the Candida species [15,30]. A recent prospective multicenter surveillance study of candidemia has showed a higher 30-day mortality rate in neonates with candidemia than children (40% vs. 28%, P = 0.02) [31]. Our study further demonstrated that neonatal invasive candidiasis has worse responsiveness to antifungal therapy, more prolonged fungemia, more likely to have severe sepsis and septic shock, and higher rates of sepsis attributable mortality and in-hospital mortality. Several factors can affect the treatment outcomes of candidemia, including underlying chronic comorbidities, microbiological factors, treatment policies and timely administration of antifungal agents, illness severity, and treatment with an infected catheter in situ or removal [3,14,[32][33][34][35]. The underlying illness and predisposing factors for neonatal candidiasis and non-neonatal  [29] pediatric candidiasis are basically different. We found extreme prematurity and related comorbidities, including use of total parenteral nutrition and underlying chronic lung disease to be the major predisposing factors for neonatal candidemia [31]. In non-neonatal pediatric candidiasis, underlying chronic conditions, especially neurological sequelae, hemodialysis, hematological/cancer patients on immunosuppressive treatment and resulting neutropenia accounted for the majority of the susceptible hosts.
Administration of antifungal agents is less frequently delayed in the non-neonatal pediatric setting because the underlying chronic comorbidities would remind clinicians of the possibility of opportunistic infections. Furthermore, we identified incidental cases of congenital candidemia without any of these underlying illnesses or risk factors, which has rarely been reported [36]. Because most of our Candida isolates were sensitive to all commonly prescribed antifungal agents, treatment  S-DD or R 2 (0.7) 0 (0) 2 (1.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 2 (3.6) All data were expressed as number (percentage %), unless indicated otherwise S-DD susceptible-dose dependent, R resistant failures most likely were caused by the failure of infectious sources control and underlying illness. Our results are similar to those of Grim et al. [37], who concluded that a high mortality rate among patients with candidemia (34% mortality within 30 days), with underlying cirrhosis and HIV infection, and increased illness severity were the independent risk factors, despite timely receipt of appropriate antifungal therapy. We documented that removal of an infected catheter was an important factor for successful treatment of invasive candidiasis [38,39]. Although breakthrough candidemia was not associated with nonsusceptible fluconazole isolates [8,40], it was independently associated with treatment failure and also more prolonged fungemia in our cohort. This study included all episodes of invasive candidiasis in children [41]. Although some episodes were blood culture-negative intra-abdominal candidiasis and some were Candida spp. positive only in the pleural fluid, all cases of invasive candidiasis were evaluated by the infection specialist and were found to require treatment, unless mortality preceded the diagnosis of invasive fungal infection. Currently no clinical study has assessed the need to treat Candida peritonitis [42,43], but we excluded cases of probable Candida colonization [44]. Furthermore, the high proportion of mixed-flora peritonitis is one limitation in this study because the pathogenic role of Candida in this polymicrobial form of infection is a matter of debate. However, many experts still consider positive Candida cultures from intraabdominal fluid in patients with peritonitis to be clinically significant even in the presence of concomitant bacterial growth [45]. Antifungal therapy is recommended by the European Society of Clinical Microbiology and Infectious Diseases and the Infectious Disease Society of America guidelines on management of complicated intraabdominal abscess or peritonitis that is positive for Candida spp. [46,47]. Recent studies described an overall mortality of 17.2-46.2% among children with invasive candidiasis [15, 19-21, 30, 48-50], and mortality attributable to candidemia has been reported to be 12-22% [15,30,[48][49][50]. Some Candida species were associated with worse outcomes. For example, C. tropicalis, C. krusei and C. glabrata related candidemia are associated with higher mortality rates than C. parapsilosis related candidemia in adults [28,48]. However, our cohort did not include enough patients to support any firm conclusions in the pediatric settings. We found the choice of antifungal therapy did not appear to have a significant impact on treatment result [51][52][53]. In our cohort, patients who received fluconazole alone had a more favorable outcome than did patients who received other regimens, reflecting a bias toward the administration of fluconazole therapy to patients who were less ill.
The strengths of our study include the systematic identification of specific species of Candida isolates, the systemic collection of clinical data, and the fact that this is the first report that focuses specifically on the difference between neonatal and non-neonatal pediatric settings. However, there were some limitations in this study. First, this study was a retrospective study and conducted in a single center. Therefore, extrapolation of the findings to other institutions must be done cautiously. Second, severity of illness was not documented because the scoring systems were different in the neonatal and non-neonatal period. Third, Candida surveillance cultures were not obtained, and there were no data regarding prior colonization [19]. We failed to document all the risk factors for invasive candidiasis in this study. Furthermore, the policies regarding antifungal therapy may be changed over such an extended study period in our non-interventional study, which should be considered as a description of clinical practice only.

Conclusions
In conclusion, this study demonstrates some significant differences of invasive candidiasis between neonates and children. Neonatologists and pediatricians must consider age-specific differences when developing treatment and prevention guidelines, or when studies of other age groups are interpreted. Furthermore, because the clinical signs of invasive candidiasis are not specific and early microbiological documentation remains a major challenge, intensive research dedicated to the development of alternative tools for early diagnosis of invasive candidiasis is urgently warranted.

Funding
Chang Gung Medical Research Program Foundation (grants CMRPG3E1491). The role of the funding body is to support the isolation and identification of all Candida spp. isolates in blood and sterile site cultures, and to perform the antifungal susceptibility testing.

Availability of data and materials
The datasets used/or analyzed during the current study available from the corresponding author on reasonable request. Ethics approval and consent to participate This study was approved by the institutional review board of Chang Gung Memorial Hospital, with a waiver of informed consent because all patient records and information were anonymized and de-identified prior to analysis.

Competing interests
The authors declare that they have no competing interests.

Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.