Epidemiology of Invasive Fungal Disease in Children

Abstract

Considerable progress has been made in the prevention, diagnosis, and management of pediatric patients with invasive fungal disease (IFD). The reported decreasing trend in the incidence of invasive candidiasis (IC) over the past 15 years in both neonates and children has been encouraging. Nevertheless, due to the growing number of immunocompromised children at risk for IFD, this disease continues to be associated with significant morbidity and death and with increased financial burden to the health care system. Therefore, it is important to understand the contemporary epidemiology of IFD. Incidence rates of IFD in children are affected by geographical, population, and time variability. There is an ongoing effort to constantly document and update the incidence of IFD and species distribution among different pediatric populations as a means to direct preventative, diagnostic, and therapeutic resources to the most appropriate subset of patients. Children with a hematologic malignancy or a primary or secondary immunodeficiency, those undergoing solid organ or hematopoietic stem cell transplantation, and premature neonates are the major subsets of pediatric patients at risk of developing IFD. In this review, we focus on fungal disease epidemiology with a specific emphasis on the 2 most common pediatric IFDs, IC and invasive aspergillosis (IA).

Invasive fungal disease (IFD) is a major cause of morbidity and death among immunocompromised and hospitalized pediatric patients [1, 2]. There has been a significant increase in the number of pediatric patients at risk of IFD, primarily because of the increasing use of immunosuppressive medications across many medical specialties. Simultaneous advances have been made in the management of IFD via novel fungal diagnostic tests and evidenced-based use of antifungal agents for prophylaxis and treatment. These factors collectively have altered the epidemiology and outcomes of IFD over the past 15 years [3, 4].

The spectrum of pediatric patients vulnerable to IFD is wide and includes children who receive chemotherapy for a malignancy, pediatric hematopoietic stem cell transplantation (HCT) or solid organ transplant (SOT) recipients, children with a primary immunodeficiency (PID), children who receive immunomodulating therapy for an autoimmune condition, and those with an acquired immunodeficiency. Beyond these patient groups, neonates and children hospitalized in an intensive care unit (ICU), among other groups, are also at risk for IFD [5–10]. The wide range of pediatric populations at risk for IFD makes it challenging to maintain contemporary estimates of epidemiology to guide clinical decision-making.

Despite these challenges, a recent increased focus on IFD in the pediatric literature has provided clinicians with reasonable estimates of IFD incidence and species distribution in populations at risk. Candida spp remain the leading cause of IFD among pediatric patients and are the fourth most common pathogen detected in hospital-acquired pediatric bloodstream infections in the United States and Europe [2, 11–13]. Aspergillus spp and organisms from the Mucorales family remain the leading causes of invasive mold disease (IMD) [14, 15].

In this review, we summarize the contemporary literature on the epidemiology of IFD in pediatric patients with malignancies, pediatric transplant recipients, children with a PID, and those managed in a pediatric ICU (PICU) or neonatal ICU (NICU). As previously noted, other subpopulations of children are at risk for IFD, but because data on IFD epidemiology in these populations are limited, they are not included in this discussion. Future investigations are necessary to better define the risk of IFD in these other populations.

OVERALL IFD EPIDEMIOLOGY

Both candidemia and invasive aspergillosis (IA) are associated with significant increases in hospital lengths of stay and overall in-hospital mortality rates of 15.8% for pediatric patients with candidemia and 18% for children with IA [12, 13]. In an attempt to estimate the financial burden on the health care system attributable to IFD, a cost-analysis study revealed that total hospital charges for treating invasive candidiasis (IC) for nonneonatal pediatric patients have a mean increase of $92 266 (95% confidence interval, $65 058–$119 474) per episode [12]. Similarly, for children with IA, the median healthcare cost can reach $49309 per episode [13].

A number of pediatric multicenter studies have been performed in different countries to document the incidence of IC and IA (Table 1) [16–21] and the distribution of fungal pathogens in children (Tables 2 and 3). The largest international collaborative studies to have assessed the incidence of IC and IA in children were conducted by the International Pediatric Fungal Network (IPFN [see www.ipfn.org]) [14, 22]. A predominance of non-albicans Candida spp in pediatric (56%) and neonatal (52%) patients was found by the IPFN [22], with similar distributions of Candida albicans and Candida parapsilosis, as reported in other pediatric studies performed in Latin America, the United States, and Europe [23–25]. In particular, a South American surveillance study found that C albicans (pediatric, 43.8%; neonatal, 35.7%) and C parapsilosis (pediatric, 27.0%; neonatal, 26.3%) prevailed [23]. A higher incidence of C parapsilosis infection in neonates (42%) and pediatric patients (38%) was noticed in an Australian prospective candidemia study, which revealed a possible difference in geographic Candida niches [26]. A European multicenter study performed to define Candida species distribution among pediatric patients (the EURO-CANDY study) is currently ongoing and is led by the European Pediatric Mycology Network (EPMyN) [27].

Table 1.

Incidence of IC and IA in Contemporary Multicenter Pediatric Studies

Author [ref]	IFD incidence	Time period	Cases (N)	Mortality	Type of study	Comments
Invasive Candidiasis (IC)
Zaoutis et al. [12]	4.3 /10. 000 pediatric admissions	2000	1118	15.8%	Multicenter US study (KID 2000) & (NIS 2000)	Analysis showed an absolute 10.0% increase in mortality attributable IC
Oeser et al. [16]	15.2/10.000 person-years	2000–2009	1473	NR	Multicenter EU study (England & Wales)	Decrease in IC incidence after 2007: 2.09/100. 000 (2007) versus 1.53/100. 000 (2009) Difference in IC incidence among age groups: Highest in <1 year old patients (11.0/100.000) and lowest in 10–14 year old patients: (0.47/100. 000)
Blyth et al. [26]	4.6/10.000 admissions 4.39/100. 000 population (neonates) 0.92/100 000 population (children)	2001–2004	1005	10% children 22% neonates	Multicenter study in Australia
Fisher et al. [17]	2.46/10. 000 inpatient days (2003) 0.77/10. 000 inpatient days (2011)	2003–2011	4456	14%	Multicenter US study	Decrease in IC incidence:72% for pediatric and 91% for neonatal cases Mortality varied: 17.3% (2003) versus 11.6% (2011)
Santolaya et al. [23]	8.1/10.000 pediatric admissions	2008–2010	302	28%	Multicenter study in Latin America
Cleveland et al. [19]	ATL: 13.3/ 100. 000 person-years BTM: 26.2 / 100. 000 person-years	2008–2011	1863*	29% 28%	Multicenter US study population-based surveillance	Significant decrease of IC for both pediatric and <1 year of age groups.
Cleveland et al. [18]	19/ 10.000 person-years (children) 33.8/ 100. 000 person-years (neonates)	2008–2013	3848*	NR	Multicenter US study population-based surveillance	Baltimore: Decrease in IC incidence in neonatal but not in pediatric patients: Reported increase 17% (2.0/100.000 in 2008 to 2.4/100,000 in 2013); Atlanta: the decline was greatest for persons aged <1 year: reported decrease 60% (41.7/100,000 in 2008 to 16.6/ 100.000 in 2013)
Invasive Aspergillosis (IA)
Zaoutis et al. [13]	437/100. 000 (0.4%) hospitalized immunocompromised children	2000	666	18%	Multicenter US study	Children with IA had a significantly higher mortality and longer median length of hospital stay (16 days) than immunocompromised children without IA (3 days)

^a Mixed population of children and adults (children, n = 121; <1-y-olds, n = 113).

Table 2.

Distribution of Candida spp. Causing IFD Among Pediatric Patients From Multicenter Studies Between 2000–2017

Author [ref]	Invasive Candidiasis (IC) Fungal species distribution		Time period	Cases (N)	Mortality	Type of study
	Candida albicans Ν (%)	Non-albicans Candida spp. N (%) (three most frequently reported)
Oeser et al. [16]	815 (55.3)	C parapsilosis 320 (21.7) C glabrata 60 (4.1) other 153 (10.4)	2000–2009	1473	NR	Multicenter EU study (England & Wales)
Blyth et al. [26]	47 (43.9)	C parapsilosis 41 (38.3) C glabrata 3 (2.8) C krusei 2 (1.9) C tropicalis 2 (1.9) C orthopsilosis 2 (1.9)	2001–2004	80 pediatric	NR	Multicenter study in Australia
Blyth et al. [26]	13 (39.4)	C parapsilosis 14 (42.4) C glabrata 3 (9.1) C tropicalis 1 (3.0)	2001–2004	24 neonatal	NR	Multicenter study in Australia
Steinbach et al. [22]	87 (44)	C. parapsilosis 45 (22) C. glabrata 21 (11) C. lusitaniae 8 (4)	2007–2011	196 pediatric	19%	Multicenter US & EU study (IPFN)
Steinbach et al. [22]	12 (48)	C. parapsilosis 7 (28) C. glabrata 1 (4) Other 6 (24)	2007–2011	25 neonatal	8%	Multicenter US & EU study (IPFN)
Santolaya et al. [23]	115 (38.1)	C. parapsilosis 80 (26.5) C. tropicalis 44 (14.6) C. guilliermondii 31 (10.3)	2008–2010	302 pediatric	28%	Multicenter study in Latin America

Abbreviations: IPFN, International Pediatric Fungal Network; NR, not reported.

Table 3.

Distribution of Aspergillus spp. Causing IFD Among Pediatric Patients From Multicenter Studies Between 2000–2017

Author [ref]	Invasive mold Diseases (IMDs) Fungal species distribution			Time period	Cases (N)	Mortality	Type of study
	Aspergillus spp. N (%)	Non-Aspergillus Molds N (%)
Burgos et al. [15]	A. fumigatus 67 (52.8) A. flavus 20 (15.7) A. terreus 6 (4.7) A. niger 6 (4.7)	NR		2000–2005	139 IA	A. fumigatus 38 (52) A. flavus 11 (15)	Multicenter EU study
Pana et al. [29]	NR	Rhizopus spp. (39.7) Lichtheimia spp. (17.5) Mucor spp. (12.7)		2005–2014	63 MC	33%	Multicenter study (Fungiscope & zygomyco.net)
Wattier et al. [14]	A. fumigatus 26 (20) A flavus 7 (5) A. niger 6 (5)	Mucormycoses  Rhizopus spp  Mucor spp Other mold  Curvularia spp  Exserohilum spp  Fusarium spp	17 (13) 9 (7) 3 (2) 22 (17) 4 (3) 4 (3) 4 (3)	2007–2011	131 IMIS: 98 IA 17 MC	IMIs 39 (30) IA 30 (31) MC 6 (35)	Multicenter EU and US study (IPFN)

Abbreviations: IA, Invasive Aspergillosis; IMIs, Invasive Mold infections; MC, Mucorales; NR, Not reported.

Epidemiologic studies of IMDs in children revealed that Aspergillus fumigatus and Aspergillus flavus are the predominant molds isolated [14, 15]. Beyond Aspergillus spp, pathogens from the Mucorales family were the causative agent in 13%, with Rhizopus and Mucor spp prevailing [14]. A single-center study found comparable fungal epidemiologies among pediatric patients; Aspergillus spp accounted for 40% of the IMDs, followed by Mucorales spp (20%) and Fusarium spp (11%) [28]. Two large international registries (Zygomyco.net and FungiScope) characterized pediatric-specific data surrounding the underlying fungal epidemiology in mucormycosis; Rhizopus spp predominated (39.7 %), followed by Lichtheimia spp (17.5 %) and Mucor spp (12.7 %) [29].

IFD IN PEDIATRIC PATIENTS WITH MALIGNANCIES AND HCT RECIPIENTS

It is challenging to define the incidence of IFD in children with cancer, because the incidence varies according to chemotherapy regimen and supportive care practices [30, 31]. Furthermore, the criteria for defining and diagnosing IFD have varied over time, and application of these definitions varies according to study [30–33]. Inconsistencies in the diagnostic criteria for IFD might affect the true estimate of IFD rates among these patients and therefore make the comparison among different chemotherapy protocol groups difficult [31]. Despite these challenges, early diagnosis and prompt initiation of effective antifungal therapy remains one of the important actions necessary for improving IFD outcomes in these patients [4, 30, 32, 33].

Epidemiology

The incidence of IFD in children receiving chemotherapy for cancer and those undergoing HCT remains high and is associated with increased morbidity and death [30]. Two studies nicely illustrated the changing landscape of IFD within similar cohorts of children with acute myeloid leukemia (AML) [34, 35]. The rate of IFD reached almost 5% for those in whom the AML Berlin-Frankfurt-Münster (BFM) 93 chemotherapy protocol was used, and in the same population, the IFD incidence decreased to 3% with use of the more-intensified BFM-AML 2004 protocol [34, 35]. The lower incidence of IFD in the BFM-AML 2004 study might be attributed partially to the broader administration of antifungal prophylaxis (in >70% of the chemotherapy cycles), with a preference for drugs with antimold activity [35]. As a comparator, in a French study that included 387 children with AML who were receiving the ELAM 02 chemotherapy protocol between 2005 and 2011, the incidence rate of IFD was 6.7% [36]. In the United States, a higher incidence of IFD in children with AML enrolled on the Children’s Cancer Group (CCG) 2961 protocol was reported [37]. These differences in the incidence of IFD have been explained partially by international variations in supportive infection care practices among the BFM and CCG groups for pediatric patients with AML [31]. In particular, BFM centers provided antifungal prophylaxis with antimold activity more frequently than the Children’s Oncology Group centers (63.8% vs 14.4%, respectively) [31].

The inconsistent use of antifungal prophylaxis and the choice of prophylactic agent also have an effect on reported IFD rates among groups of pediatric patients with cancer. Without antifungal prophylaxis, a study of a mixed population of such patients reported rates of IFD to between 2.9%–7.8% [38, 39]. In specific populations of pediatric patients with cancer, an IFD rate of 6.1% was observed in patients with promyelocytic AML without antifungal prophylaxis and 8.4% was found for patients with nonlymphoblastic leukemia patients [40, 41]. In a mixed population of patients with leukemia or lymphoma who received oral amphotericin B or intravenous fluconazole as prophylaxis, Watanabe et al reported an IFD rate of 3.8% (6/158 cases) [42]. Koyabashi et al in a mixed population of pediatric patients with hematologic malignancies or children undergoing HCT receiving antifungal prophylaxis reported an IFD rate of 6.9% (23/334 cases) [43]. In comparison, Kaya et al reported an IFD rate of 13.6% (proven rate, 7.2%) among children with leukemia who were receiving fluconazole prophylaxis [44].

Candida and Aspergillus spp are the predominant pathogens that cause IFD in children with cancer, but an increasing shift toward other non-Aspergillus molds (Fusarium, Scedosporium, and Mucorales) was observed recently [3, 4, 30, 45]. During a 10-year period, the average annual incidence of candidemia among pediatric patients with cancer and pediatric HCT recipients was 1.25 cases per 1000 hospital discharges [45]. Although C albicans is the most frequent species isolated, an increasing trend for non-albicans Candida spp in children with cancer, most frequently C parapsilosis and Candida tropicalis, has also been found [22, 30, 45]. A fumigatus is the most common cause of IA in children with hematologic malignancy, followed by A flavus and Aspergillus terreus [14, 15, 28]. In 1 study, among non-Aspergillus molds, pathogens from the Mucorales family caused 13% of the IFD cases, and all other non-Aspergillus and non-Mucorales molds caused 17% [14].

Outcomes

The overall case-fatality rate attributable to IFD ranges between 10% and 70%; higher rates have been observed in specific subpopulations, such as patients with disseminated IFD, central nervous system (CNS) involvement, or persistent neutropenia [30, 41, 43]. Kobayashi et al [43] reported IFD case-fatality rates of 48.2% and up to 71.4% in patients with lung involvement. For CNS aspergillosis, case-fatality rates before 1990 reached 80%, whereas after 1990, mortality rates decreased significantly to 39.5% [46]. For IC, overall fatality rates range between 10% and 25% but can reach close to 50% in patients with ICU admission [30]. In a French study of children with AML, the overall survival rate at 24 months for children diagnosed with IFD was 72% [36]. The case-fatality rates for IMD in most studies are between 20% and 50% and increase to approximately 80% in patients who have undergone allogeneic HCT [14, 15, 30, 47]. Pana et al [29] found a case-fatality rate of almost 40% for mucormycosis in children suffering from a hematologic malignancy and 80% for HCT patients respectively [29].

IFD IN PEDIATRIC SOT RECIPIENTS

Epidemiology

The true burden of IFD and species distribution after an SOT have been evaluated in few studies. A US multicenter prospective study (TRANSNET) that used IFD surveillance among mainly adult SOT recipients reported a marginal increase of IFD from 2000 to 2006; the highest rates were observed among those who underwent small bowel, lung, and liver transplantation, respectively [48]. In an attempt to analyze only the pediatric SOT recipients’ cases from the TRANSNET database, Knapp et al [49] reviewed 49 IFD episodes among 41 pediatric SOT recipients (3% of all SOT recipients in the TRANSNET cohort). The most common organisms detected were Candida spp (78%) and Aspergillus spp (8%).

Organ-specific data on IFD in pediatric SOT recipients have been limited [50–55]. A study with 98 pediatric liver transplant recipients revealed that 31% presented with a Candida infection [50]. In a more recently published study, the incidence rate of IC in children undergoing liver transplantation is estimated to be 2.5% (10 of 397) [51]. Among the 10 cases of IC reported, C albicans prevailed (50%), followed by C parapsilosis (20%), Candida lusitaniae (20%), and Candida guillermondii (10%) [51]. One study dedicated to pediatric heart transplant recipients found that Candida infections accounted for 66% of all IFDs, followed by IMDs at 16% (82% of them attributable to Aspergillus spp) [52]. Among 83 IFDs attributed to yeast infection, C albicans was responsible for the majority of them (55%), followed by C parapsilosis (13%), Candida krusei (4%), Candida glabrata (2%), and C tropicalis (2%) [52]. Among 22 IFDs attributed to mold infection, 18 were caused by Aspergillus spp (82%), followed by zygomycetes (13.6%) and Exserohilum spp (4.5%) [52]. Results from a multicenter US and European study that analyzed children undergoing lung transplantation showed that the proven and probable IFD rates reached 10.5% with almost equal distribution of Candida and Aspergillus spp [53]. In a single-center study with 55 pediatric lung transplant recipients (2002–2007), 11 patients accounted for 14 proven or probable IFD events (20%) [54]. Although pediatric data are lacking, few studies in adult lung transplant recipients have indicated that pretransplant Aspergillus spp lung colonization could be implicated with the presence of post-transplant bronchiolitis obliterans associated with Aspergillus spp pulmonary infection [56].

Contemporary data among 548 pediatric SOT recipients between 2000 and 2013 from a single center in the United States revealed a low overall IFD incidence of 2.2% (13 of 584), or 14.3 IFD events per 100000 patient-days, and a decreasing trend over time [55]. Differences in IFD rates were reported among organ transplant types and over 2 time periods. In particular, higher IFD rates were observed for heart/lung (12.5%), lung only (11.4%), and liver (4.7%) recipients than kidney and heart recipients (0%). In addition, over the 2 time periods selected (2000–2006 and 2007–2013), a decrease in the IFD rate was noted, from 4% to 1% (25.5 to 3.9 events per 100000 patient-days, respectively), along with a stable number of patients with IFD in each period. The number of patients who received antifungal prophylaxis increased over time from 6% for the first time period to 9% for the second period, which might explain some, but likely not all, of the decrease in the incidence of IFD between the 2 time periods [55].

Outcomes

The mortality rate associated with IFD varies according to type of SOT, type of IFD, and time period. In 1999, Gladdy et al [50] reported a 33.3% case-fatality rate in pediatric liver transplant recipients with invasive Candida infection. On the contrary, in a recently published study, only 1 of the 10 patients with IC died (mixed infection with C parapsilosis and IA) [51]. In a study of pediatric heart transplant recipients with IFD, the case-fatality rate reached almost 50% [52]. More specifically, 13 (59%) of 22 patients with IMD and 43 (47%) of 92 patients with yeast infection died [52]. The case-fatality rate from the aforementioned cohort of 548 pediatric SOT recipients was 21.4% (3 of 14 patients), including 2 lung recipients and 1 heart/lung recipient [55].

IFD IN PEDIATRIC PATIENTS WITH A PID

Epidemiology

Among all PIDs, the epidemiology of IFD has been defined most clearly for chronic granulomatous disease (CGD), an inborn error of the phagocyte nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex. Although children with CGD are at risk for a wide range of yeast and mold pathogens, Aspergillus spp and Candida spp are the most common causes [5, 57–64]. Among 155 patients with CGD in a French study from 1976 to 2008, 42.6% (66 of 155) developed at least 1 IFD [58]. In particular, IMDs represented 61.3% (49 of 80) of all IFD events. Aspergillus spp accounted for 65.3% of these IMDs (32 of 49); A fumigatus (28.5%) and Aspergillus nidulans (22.4%) were the most common Aspergillus spp [58]. It was notable that itraconazole prophylaxis had a significant effect on the incidence of IFD [58].

Another congenital immunodeficiency associated with an increased susceptibility to IFD is hyperimmunoglobulin E syndrome (HIES) (ie, Job syndrome) [5, 65]. Invasive pulmonary aspergillosis occurs in almost 20% of patients with HIES, almost exclusively secondary to the presence of pneumatocysts and bronchiectasis caused by recurrent bacterial infections and impaired local STAT3-dependent lung epithelial immunity [65, 66]. Although rare, dissemination of fungal infection to the CNS in these patients has been reported occasionally [65, 67]. A recent literature review reported 16 cases of patients with HIES with rare endemic/dimorphic fungi such as Coccidioides, Cryptococcus, and Histoplasma, which underscores the vulnerability of this patient group to a wide range of fungal pathogens [66].

The caspase recruitment domain-containing protein 9 (CARD9) represents an essential molecule of the innate immunity that controls Dectin-1-mediated myeloid cell activation and cytokine production. CARD9 deficiency is a PID with impaired Candida species killing [67–69]. Although the available cohorts are not large enough to define the true incidence and case-fatality rates of IFD in these patients, a case series report suggested that the gastrointestinal tract and, in particular, the CNS are common anatomical locations for Candida infection [69]. Other important anatomic locations are bones and eyes [70]. A subsequent case series found that CARD9-deficient patients also can suffer from isolated IA in the CNS and gastrointestinal tract [71].

Outcomes

The overall IFD case-fatality rate for patients with CGD reached 17% in 1 study, although a reduction in the mortality rate over time, from 43% (1985–1990) to 6% (1991–2009), was reported [62]. The decrease in deaths for patients with CGD over the past 15 years has been attributed to high clinical awareness and the implementation of itraconazole prophylaxis [58, 62, 64]. Nevertheless, IA remains a major cause of death for patients with CGD [64]. In patients with HIES, a 17% IFD case-fatality rate has been reported [65].

IFD IN PICU PATIENTS

Children in PICUs represent a heterogeneous pediatric population with a well-documented increased risk for developing IFD as a result of a unique combination of critical and complex clinical conditions, including prolonged need for hospitalization, frequent invasive interventions, and the presence of foreign devices, such as catheters and endotracheal tubes [72]. The predominant cause of IFD in the PICU is IC, while IA is observed mainly in children admitted to the PICU with an underlying hematologic malignancy.

Epidemiology

The incidence of IC and Candida species distribution vary among different PICUs and among different time periods. These differences can reflect the peculiarities of specific institutions associated with differences in critical care practices, differences in geographical niches of Candida spp, and the expansion of antifungal prophylactic regimens. For example, from 2005 to 2009, the incidence of IC among 7 PICUs in Greece ranged from 0 to 14.1 cases per 1000 admissions (median, 6.4 cases per 1000 admissions) [73]. Comparable incidences have been reported from Spain (6.9 cases per 1000 admissions during a 2-year period [1996–1998]) [74]. In a study in the United States, a slightly lower incidence of 3.5 cases per 1000 admissions between 1997 and 2004 was found [75], similar to the results from Egypt (3 cases per 1000 inpatient-days) [76]. Over a 10-year-period, the incidence in PICU patients in a single-center study in Germany was 0.59 per 1000 hospital discharges (95% confidence interval, 0.02–1.09 per 1000 hospital discharges) [45]. A more recent update from Spain for the period 2008–2009 reported an incidence of 4.22 cases per 100 PICU admissions [77].

Richards et al [78] reported that almost 10% of bloodstream infections in US PICUs were attributed to Candida spp, whereas a study in Israel reported that 14.4% of bloodstream infections in PICUs were candidemia [79]. C albicans remains the leading cause of IC in the PICU, but an increasing trend of non-albicans Candida spp has been found worldwide. In Europe, C albicans prevails, with a proportion that ranges between 37.6% and 55.5%, which is comparable to those found in US studies, which have reported 46% of IC cases to be caused by C albicans [73, 75–77, 80, 81]. C parapsilosis is the second leading etiology of IC at approximately 20%. The high percentage of C parapsilosis infections isolated in the PICU emphasizes the need to implement additional infection control bundle measures, because its origin is mainly exogenous, either through horizontal transmission or adherence to foreign devices (such as catheters and other devices) [75]. Other Candida spp, most prominently C tropicalis, C glabrata, C krusei, and C lusitaniae, account for 10% to 15% of isolates. Differences in the distribution of these species among different PICUs have been associated with local practices; therefore, it is necessary to learn a center’s local epidemiology.

Outcomes

The case-fatality rate for IC in the PICU is difficult to estimate because of the high clinical complexity and severity of the patients’ underlying conditions. Zaoutis et al [75] compared the case-fatality rates in children with IC and controls in a PICU and found a statistically significant higher rate of death in children with IC (44% vs 14%). The same group observed prolonged median PICU and hospital lengths of stay for children with IC (35 and 46 days, respectively) [75]. Hegazi et al [76] found a similar case-fatality rate (42.4%) in PICU patients, and the candidemia case-fatality rate was estimated to be 16.7%, similar to the 18.2% reported by Vogiatzi et al [73].

The effect of specific Candida species on mortality rates among children in PICUs has been evaluated; however, results have been conflicting. In another study, children with candidemia caused by non-albicans Candida spp were twice as likely to die than children with candidemia caused by C albicans [82]. In contrast, other studies found no significant difference between different Candida spp and case-fatality rates [76, 79]. In 1 study, the main species associated with higher mortality rates were C glabrata, C krusei, and C tropicalis, and the authors felt this result was most likely a result of decreased susceptibility and/or resistance to fluconazole in C glabrata and C krusei [76].

IFD IN NICU PATIENTS

A significant increase in the incidence of IC was reported initially during the 1990s to be temporally associated with increased survival rates of premature very-low-birth-weight neonates, but in the past 15 years, there has been an overall decrease in neonatal IC within European countries and the United States [16, 17, 20, 21, 83, 84]. The cause of this decrease is likely multifactorial and has been correlated with prophylactic use of fluconazole and infection control bundle measures to eliminate catheter-related bloodstream infections [17, 84].

Epidemiology

In a large cohort study that included 6956 very-low-birth-weight neonates, C albicans was the third most common pathogen causing late-onset sepsis (6%) [85]. Results from a multicenter study (19 centers in the United States) of extremely-low-birth-weight (ELBW) neonates revealed significant variability in the incidence of IC among different centers (2%–28%) [86]. Aliaga et al [84] were among the first to report a significant decrease of neonatal IC, which dropped in the United States from 3.6 per 1000 infants in 1997 to 1.4 per 1000 infants in 2010. Similar decreases have been reported since in a number of smaller European studies [16, 19]. In a UK study, lower median ages of diagnosis were reported for infants <90 days of age with C albicans (11 days) or C glabrata (9 days) infection compared to those infected with other species, such as C parapsilosis (18 days), C tropicalis (20 days), and C lusitaniae (23 days) [16]. Irrespective of the age of diagnosis, C albicans remains the most frequent Candida species associated with neonatal IC, followed by C parapsilosis and C tropicalis; C glabrata and C krusei are encountered less frequently [87–89]. The incidence of C parapsilosis infection in NICU patients was rather stable in a comparison of the time periods before 2000 (33.5%) and after 2000 (27%) [90].

Outcomes

Despite the decreasing incidence, neonatal IC is associated with a high case-fatality rate; the overall estimated rate is approximately 20%, which increases to 50% for ELBW infants. Increased numbers of deaths and long-term neurodevelopmental abnormalities have been associated with neonatal IC, particularly with the occurrence of hematogenous Candida meningoencephalitis [86, 91, 92]. Almost 50% of infants who survive neonatal IC will have severe long-term neurodevelopmental deficits, including cerebral palsy, ocular, hearing, or cognitive impairment, or periventricular leukomalacia [91–95]. Additional poor prognostic factors for neonatal IC outcome include the early onset of IC, delayed catheter removal, and delayed initiation of antifungal therapy [96–98]. A recently published meta-analysis found that fluconazole prophylaxis in ELBW infants contributed to not only a significant reduction of IC but also a reduction in case-fatality rates [99].

CONCLUSION

A global effort to constantly update our knowledge on the incidence and distribution of pathogens that cause pediatric IFD is ongoing. Continuing and actually expanding this effort is necessary to better understand the changing incidence and outcomes of IFD and to identify emerging populations at risk. Local epidemiologic monitoring is also necessary to understand the burden of IFD at the institutional level. These data are of utmost importance for tailoring preventive measures, focusing resources on the most susceptible hosts, and implementing institution-based infection control strategies. Although the spectrum of children who are vulnerable to IFD is wide, the majority of cases are inclusive of the patient populations reviewed in detail here. The results of recent studies on the epidemiology of Candida infections suggest that a decrease in infection rates has occurred in the past decade. The reason for this decline is not exactly known but is often attributed to the use of antifungal prophylaxis and improved infection control practices. A gradual shift from C albicans to non-albicans Candida infection has been recorded also, whereas a stable incidence of IA has been observed. Mortality rates remain high, depending on the fungal pathogen isolated and underlying condition of the pediatric patient. Future work is needed to improve diagnostic capabilities to better understand the epidemiology of these infections and to enable earlier initiation of appropriate therapeutic interventions that will result in improved survival in those with one of these devastating infections.

Notes

Financial support. A.W. is supported by the Wellcome Trust Strategic Award (grant 097377) and the MRC Centre for Medical Mycology at University of Aberdeen (grant MR/N006364/1).

Supplement sponsorship. This article appears as part of the supplement “State of the Art Diagnosis of Pediatric Invasive Fungal Disease: Recommendations From the Joint European Organization for the Treatment of Cancer/Mycoses Study Group (EORTC/MSG) Pediatric Committee,” sponsored by Astellas.

Potential conflicts of interest. Z.D.P. has no conflict of interests. E.R. is an independent contractor (research grants) of significant value: Astellas, Gilead, Pfizer, Scientific Advisor (Review Panel of Advisory Committee): Astellas, Gilead, Pfizer, Merck, and Speaker’s Bureau: Astellas, Gilead, Merck, Pfizer. T.Z. is a Consultant: T2 Biosystems, Nabriva, and received research support: Merck. A.H.G. has received grants from Gilead, Merck, Sharp & Dohme, and Pfizer; is a consultant to Astellas, Basilea, Gilead, and Merck, Sharp & Dohme, and served at the speakers’ bureau of Astellas, Basilea, Gilead, Merck, Sharp & Dohme, Pfizer, Schering-Plough, and Zeneus/Cephalon.