Secondary Logo

Journal Logo

Original Studies

Candidemia in Children

A 16-year Longitudinal Epidemiologic Study

Silvester, Eloise J. MD*,†; Watanabe, Melissa M. Y. MBBS; Pittet, Laure F. MD, PhD; Boast, Alison MD; Bryant, Penelope A. BM, BCh, PhD*,†,‡,§; Haeusler, Gabrielle M. MBBS†,‡,¶,‖; Daley, Andrew J. MBBS, MMed(ClinEpi)†,**; Curtis, Nigel MBBS, PhD*,†,‡; Gwee, Amanda MBBS, PhD*,†,‡

Author Information
The Pediatric Infectious Disease Journal: June 2021 - Volume 40 - Issue 6 - p 537-543
doi: 10.1097/INF.0000000000003082

Abstract

Systemic fungal infections cause significant morbidity and mortality. Candida species are the third leading cause of nosocomial bloodstream infections in children in the United States and Europe and are associated with significant economic burden.1–4 Rates of candidemia are increasing in adults in Australia5,6 and in adults and children in other countries.5,7 Potential reasons for this include the prolonged survival of immunocompromised patients and the increased use of invasive devices.5,7 Of additional concern is the emergence of antifungal resistance with increasing reports of multi-drug resistant Candida auris.8,9

The spectrum of disease for disseminated candidiasis in patients with candidemia has varied over time, with rates of ocular fungal disease decreasing.10 Well-established risk factors for candidemia in children include presence of a central venous catheter (CVC), prior broad-spectrum antibiotic exposure, immunocompromised states and gastrointestinal disease.11,12Candida spp. are commonly found in the normal human microbiota, and antibiotic exposure allows overgrowth of gastrointestinal Candida spp. and subsequent translocation to the blood.13,14 Characterizing the local epidemiology of candidemia is essential for identifying risk factors, monitoring emerging strains and directing prophylaxis and treatment to improve patient outcomes. We aimed to describe the epidemiology and spectrum of disease of candidemia in children admitted to a tertiary pediatric hospital.

METHODS

Study Design

We retrospectively reviewed the case notes of all children who had a positive blood culture with Candida spp. over a 16.5-year period (January 2002 to June 2018) at the Royal Children’s Hospital (RCH) in Melbourne, a tertiary referral center with a 34-bed oncology and hematopoietic stem cell transplantation unit. Positive blood cultures were identified through the laboratory information management system. Additionally, positive stool cultures and corresponding ward location were identified. Ethics approval was obtained from RCH human research ethics committee (HREC 37259A).

Episodes of candidemia were included if Candida spp. was isolated from blood culture and it was deemed significant by the clinical team and treated. Clinical data were collected on patient demographics, medical comorbidities, other sites of infection and colonization, Candida spp. isolated, antifungal susceptibility, antifungal prophylaxis and treatment, and clinical outcome. Comorbidities and predisposing factors were identified in the following age groups: <30 days (neonates); 30 days to <1 year (infants); and 1 to ≤18 years of age (children).15,16

Definitions

A new episode was defined as the isolation of a new Candida spp. or the same species after a >30-day interval. Duration of candidemia was calculated from the day of the first positive blood culture until the first negative blood culture, with persistent candidemia defined as positive blood cultures >3 days post-CVC removal.17 Colonization was defined as a positive isolate with Candida spp. from a non-sterile site before the episode of candidemia. If a patient died in the setting of persistent candidemia or Candida spp. infection, and it was documented as a cause of death on the patient’s death certificate, mortality was attributed to candidemia. Additional investigations for source or complications of infection (such as renal and abdominal ultrasound, positron emission tomography, chest and abdominal computed tomography, echocardiogram and ophthalmologic assessment) were at the discretion of the treating physician. Results for echocardiograms were available from 2011 to 2018.

Microbiologic Methods

Blood samples were cultured using the BacT/ALERT automated blood culture system (bioMerieux Inc., Durham, NC) and positive samples were sub-cultured onto Sabouraud dextrose agar and CHROMID Candida agar (bioMerieux Inc.) to obtain single colonies. Formal yeast identification was done using the VITEK 2 YST Card (bioMerieux Inc.) and susceptibility testing was done using Clinical and Laboratory Standards (Wayne, PA, www.clsi.org) disk diffusion18 or the Sensititre YeastOne Y09 AST broth microdilution system (Thermo Fisher, Waltham, MA). At RCH, stool samples are routinely cultured and speciated via CHROMID Candida agar (bioMerieux Inc.) for oncology, immunocompromised and immunosuppressed patients.

Statistical Analysis

The incidence of candidemia was determined using hospital admission data available from July 2003 to June 2018. Patient factors associated with the outcomes of all-cause mortality at 30 days were identified using univariate logistic regression. Risk factors with a P < 0.2 were included in a stepwise backward multivariate logistic regression analysis. We performed a sensitivity analysis using one candidemia episode per patient. If there was >1 episode per patient, the last episode was used. Categorical variables such as neutropenia, fungal colonization, fungal renal disease, prior episodes of candidemia, antifungal prophylaxis and CVC removal were analyzed using the χ2 or Fisher-exact test.

Data were analyzed using Stata v.15 (StataCorp, College Station, TX). All tests were 2-tailed and a P < 0.05 was considered statistically significant. Nonparametric tests were used for nonparametric variables.

RESULTS

Patient Demographics and Incidence of Candidemia

During the study period, there were 151 potential episodes of candidemia. Of these, 12 were excluded; 9 patients had a single positive blood culture and remained well despite not receiving antifungal therapy with no adverse outcomes; and for 3, the medical records were unavailable. Overall, 139 episodes of candidemia in 124 patients were included (13/139, 9% in neonates; 17/139, 12% in infants; 109/139, 78% in children) (Table 1). The median number of episodes per patient was 1 (range 1–4) episode. Overall, 76/139 (55%) episodes occurred in males and the median age was 6.1 years (range 3 days to 18 years). The average annual incidence of candidemia between July 2003 and June 2018 was 2.2 (range 0.9–3.8) per 10,000 hospitalizations and this remained stable over the study period.

TABLE 1. - Patient Characteristics
Neonate < 30 d (n = 13) Infant 30 d to <1 y (n = 17) Child 1 to ≤18 y (n = 109) All (n = 139)
Demographic characteristics
 Total episodes (n) 13 17 109 139
 Age, median (range) 14 (3–28) d 4 (1–11) m 8 (1–18) y 6 (3d–18y) y
 Sex, n of male (%) 6 (46) 12 (71) 58 (53) 76 (55)
Comorbidities
 Gastrointestinal disease, n (%)* 6 (46) 6 (35) 53 (49) 65 (47)
 Respiratory disease, n (%) 8 (62) 8 (47) 36 (33) 52 (37)
 Malignancy, n (%) 0 3 (18) 48 (44) 51 (37)
  Hematologic, n (%) 0 2 (12) 32 (29) 34 (24)
  Solid, n (%) 0 1 (6) 16 (15) 17 (12)
 Cardiac disease, n (%) 5 (38) 4 (24) 19 (17) 28 (20)
 Renal disease, n (%) 2 (15) 4 (24) 15 (14) 21 (15)
 Neurologic disease, n (%) 2 (15) 0 13 (12) 15 (11)
 Liver disease, n (%) 2 (15) 4 (24) 5 (5) 11 (8)
 Congenital disease, n (%) 1 (8) 0 10 (9) 11 (8)
 Hematopoietic stem cell transplant, n (%) 0 1 (6) 6 (6) 7 (5)
 Diabetes, n (%) 0 0 4 (4) 4 (3)
 Premature birth, n (%) 8 (62) 5 (29)
  Gestational age weeks, median (range) 31 (24–42) 38 (24–41)
 LBW (1500–2500 g), n (%) 2 (15) 1 (6)
 VLBW (1000–1500 g), n (%) 1 (8) 0
 ELBW (<1000 g), n (%) 4 (31) 3 (18)
Risk factors
 Central venous catheter, n (%) 10 (77) 16 (94) 104 (95) 130 (94)
 Total parenteral nutrition, n (%) 8 (62) 9 (53) 43 (39) 60 (43)
 Bacteremia (30 d prior), n (%) 2 (15) 6 (35) 35 (32) 43 (31)
 Severe neutropenia (<0.5 × 109/L), n (%) 0 3 (18) 31 (28) 34 (24)
 Mucositis, n (%) 0 1 (6) 21 (19) 22 (16)
 Burns/trauma, n (%) 0 0 3 (3) 3 (2)
 Antibiotics (30 d prior), n (%) 12 (92) 16 (94) 92 (84) 120 (86)
  Aminoglycosides, n (%) 11 (95) 12 (71) 68 (62) 91 (65)
  Penicillin combinations, n (%) 2 (15) 7 (41) 60 (55) 69 (50)
  Vancomycin, n (%) 5 (38) 8 (47) 40 (37) 53 (38)
 Immunosuppressive treatment (30 d prior), n (%) 0 6 (35) 55 (51) 61 (44)
  Chemotherapy, n (%) 0 4 (24) 46 (42) 50 (36)
  Corticosteroids, n (%) 0 5 (29) 35 (32) 40 (29)
Site of fungal infection
 Imaging for renal fungal disease (US/CT/PET), n (%) 11 (85) 16 (94) 87 (80) 114 (82)
  Renal fungal disease, n (% of those with imaging)§ 2 (18) 3 (19) 4 (5) 9 (8)
 Ophthalmologic review, n (%) 10 (77) 13 (76) 79 (72) 102 (73)
  Fungal endophthalmitis, n (% of those with reviews) 0 1 (8) 7 (9) 8 (8)
 Imaging for hepatosplenic nodules, n (%) 5 (38) 8 (47) 79 (72) 92 (66)
  Hepatosplenic nodules, n (% of those with imaging) 0 1 (12.5) 4 (5) 5 (5)
 Imaging for pulmonary nodules, n (%) 0 2 (12) 36 (33) 38 (27)
  Pulmonary nodules, n (% of those with imaging) 0 0 9 (27) 9 (24)
 Echocardiogram performed, n (%) 5 (38) 3 (18) 31 (28) 39 (28)
  Endocarditis, n (% of those with imaging) 0 0 1 (3) 1 (3)
 Candiduria, n (%) 2 (15) 2 (12) 6 (6) 10 (7)
Outcomes
 CVC removed ≤2 d, n (% of episodes with CVC prior) 3 (30) 5 (31) 29 (28) 37 (28)
 CVC removed ≤7 d, n (% of episodes with CVC prior) 6 (60) 8 (50) 63 (61) 77 (59)
 Prolonged candidemia (>3 d post CVC removal), n (% of episodes with CVC prior) 4 (40) 2 (13) 14 (13) 20 (15)
 30-day all-cause mortality, n (%) 3 (23) 3 (18) 10 (9) 16 (12)
*Gastrointestinal disease included: inflammatory/infectious conditions, for example, pancreatitis, gastroenteritis; requirement of PEG feeding, Hirschsprung disease.
Data not available for: gestational age for 1 neonate and 7 infants, birth weight for 2 neonates and 10 infants; bacteremia 1 child; severe neutropenia 2 neonates and 3 children.
Bacteremia did not include bacteria isolated from same culture as Candida spp.
§Renal fungal disease included fungal balls, hypoechoic lesions, echogenic foci/debris.
LBW indicates low birth weight; VLBW, very low birth weight; ELBW, extremely low birth weight.

The most common predisposing factors for candidemia were the presence of a CVC (130/139, 94%), prior broad-spectrum antibiotic exposure (120/139, 86%), total parenteral nutrition (60/139, 43%) and bacteremia in the prior 30 days (43/139, 31%) (Table 1). Immunosuppressive treatment was also a frequent predisposing factor (61/139, 44%; chemotherapy 50/139, 36%; corticosteroid treatment 40/139, 29%) and severe neutropenia (<0.5 × 109/L) was present in 34/139 (24%) of patients, including 34/51 (67%) with a cancer diagnosis.

Comorbidities

The most common comorbidities across all age groups were gastrointestinal disease (65/139, 47%) followed by respiratory disease (52/139, 37%), and malignancy (51/139, 37%) (Table 1). Of those with malignancy, 34/51 (67%) were hematologic. In the 100 days before developing candidemia, 7 patients (7/139, 5%) had received a hematopoietic stem cell transplant and 6 of these for hematologic malignancy. The proportion of patients with comorbid hematologic malignancy increased over the study period from 11/64 (17%) in the first 8 years (2002–2009) to 23/75 (31%) in the remaining 8.5 years (2010–2018).

In neonates, prematurity (8/13, 62%) was the most common comorbidity (median gestational age 31.0, range 24.0–41.7 weeks). Birth weight data were available in 11 neonates and 2 were low birth weight (1500–2500 g), 1 was very-low birth weight (1000–1500 g) and 4 were extremely-low birth weight (<1000 g).

Site of Infection

To assess for other sites of infection, patients were investigated with imaging for renal fungal disease (fungal balls, hypoechoic lesions and echogenic foci/debris) in 114/139 (82%) episodes, for hepatosplenic nodules in 92/139 (66%) episodes and for pulmonary dissemination in 38/139 (27%) episodes. Echocardiograms were performed in 38/62 (61%) episodes (from 2011 to 2018) and specialist ophthalmologic review in 102/139 (73%) episodes. Infection at these other sites occurred in a minority of children: pulmonary dissemination occurred in 9/38 (24%) episodes, evidence of renal fungal disease was present in 9/114 (8%) episodes, fungal endophthalmitis in 8/102 (8%) episodes, and hepatosplenic nodules in 5/92 (5%) episodes. There was only one episode of endocarditis (1/38, 3%) in a 5-year-old child with acute lymphoblastic leukemia and a single episode of meningitis in a 5-day-old premature neonate. Children with persistent candidemia had a higher frequency of disseminated disease than those whose CVC was removed within 3 days (6/19, 37% vs. 12/76, 16%, P = 0.04). In 10/139 (7%) episodes, the patient had candiduria in the 30 days before candidemia and in 8 of these episodes, the same species was cultured from urine and blood.

Microbiology and Antifungal Susceptibility

Of the 139 episodes of candidemia, 2 episodes had 2 different Candida spp. isolated, therefore there were 141 Candida isolates (Table 2). Overall, the most common Candida spp. was C. albicans (65/141, 46%), followed by C. parapsilosis (36/141, 26%), C. krusei (14/141, 10%) and C. glabrata complex (9/141, 6%). Non-albicans candidemia (NAC) accounted for approximately half of the isolates (76/141, 54%) with the rates of NAC being higher in children [67/111, 60%, 95% confidence interval (CI): 50.6–69.5] compared with neonates (3/13, 23%, 95% CI: 5.0–53.8) and infants (6/17, 35%, 95% CI: 14.2–61.7; P < 0.01). During the study period, there was an increase in the proportion of NAC [42% (2002–2009) to 68% (2010–2018)] (Figure, Supplemental Digital Content 1, https://links.lww.com/INF/E310). This was primarily due to an increase in C. krusei from 2% to 17% (2002–2009 and 2010–2018, respectively) (Figure, Supplemental Digital Content 1, https://links.lww.com/INF/E310). C. albicans and C. parapsilosis were the most common cause of candidemia across all patient comorbidities except those patients with hematologic malignancy in whom C. krusei (12/34, 35%) was the predominant isolate.

TABLE 2. - Candida Spp. by Age Group
Neonate, n (%) (n = 13) Infant, n (%) (n = 17) Child, n (%) (n = 111) All, n (%) (n = 141)
C. albicans 10 (77) 11 (65) 44 (40) 65 (46)
Non-albicans candidemia 3 (23) 6 (35) 67 (60) 76 (54)
C. parapsilosis 2 (15) 5 (29) 29 (26) 36 (26)
C. krusei 0 0 14 (13) 14 (10)
C. glabrata complex 1 (8) 0 8 (7) 9 (6)
C. guilliermondii 0 0 5 (5) 5 (4)
C. lusitaniae 0 0 4 (4) 4 (3)
C. tropicalis 0 1 (6) 4 (4) 5 (4)
C. famata 0 0 1 (1) 1 (1)
Candida spp. (not speciated)* 0 0 2 (2) 2 (1)
*Candida spp. not speciated were non-albicans spp.

There were 59 (42%) episodes in which the patient had received antifungal prophylaxis within the 30 days before developing candidemia. The prophylaxis received included nystatin 42/59 (71%), an azole antifungal 17/59 (29%; 7 itraconazole, 10 fluconazole), amphotericin B 4/59 (7%) and an echinocandin 3/59 (5%) (4 patients received >1 agent). Of the 17 candidemia episodes with azole prophylaxis, almost half were due to C. krusei (8/17, 47%), followed by C. albicans (4/17, 24%), C. guilliermondii (3/17, 18%), C. parapsilosis (1/17, 6%) and C. glabrata (1/17, 6%). Azole-resistant C. krusei was isolated in 3/8 (38%) episodes and azole-intermediate in 3/8 (38%) episodes with azole prophylaxis.

Antifungal susceptibility data were available for fluconazole in 127/141 (90%), amphotericin B in 121/141 (86%) and echinocandins in 13/141 (9%) isolates (Table 3). Excluding C. krusei due to its inherent resistance to fluconazole, 111/115 (97%) isolates were susceptible to fluconazole. Fluconazole resistance occurred in 8 isolates (8/127, 6%), including 7 C. krusei (7/12, 58%) and one C. glabrata (1/9, 11%). All 7 patients with fluconazole-resistant C. krusei had hematologic malignancy and 3 had received azole prophylaxis. A further 7 isolates (7/127, 6%) had intermediate minimum inhibitory concentrations to fluconazole (4 C. krusei and one each of C. albicans, C. parapsilosis and C. lusitaniae). The C. albicans and C. lusitaniae isolates were from the same patient with CHARGE syndrome who was receiving prophylactic nystatin and was admitted to the intensive care unit post-cardiac surgery. There was no amphotericin B or echinocandin resistance in the 121 and 13 isolates tested, respectively.

TABLE 3. - Antifungal Sensitivities for Episodes With Susceptibility Testing
Fluconazole, n (% of Isolates) Amphotericin, n (% of Isolates) Echinocandin*, n (% of Isolates)
n S I R n S n S
Overall (n = 141 isolates) 127 (90) 112 (79) 7 (5) 8 (6) 121 (86) 121 (86) 13 (9) 13 (9)
C. albicans (n = 65) 56 (86) 55 (85) 1 (2) 0 54 (83) 54 (83) 3 (5) 3 (5)
C. parapsilosis (n = 36) 35 (97) 34 (94) 1 (3) 0 35 (97) 35 (97) 3 (8) 3 (8)
C. krusei (n = 14) 12 (86) 1 (7) 4 (29) 7 (50) 8 (57) 8 (57) 4 (29) 4 (29)
C. glabrata complex (n = 9) 9 (100) 8 (89) 0 1 (11) 9 (100) 9 (100) 1 (11) 1 (11)
C. lusitaniae (n = 5) 5 (100) 4 (80) 1 (20) 0 5 (100) 5 (100) 0 0
C. guilliermondii (n = 5) 4 (80) 4 (80) 0 0 4 (80) 4 (80) 1 (20) 1 (20)
C. tropicalis (n = 4) 4 (100) 4 (100) 0 0 4 (100) 4 (100) 1 (20) 1 (20)
Candida spp. (not speciated) (n = 2) 1 (50) 1 (50) 0 0 1 (50) 1 (50) 0 0
C. famata (n = 1) 1 (100) 1 (100) 0 0 1 (100) 1 (100) 0 0
*Echinocandin minimum inhibitory concentration not well established.
No fluconazole minimum inhibitory concentration for C. krusei.
One isolate of C. krusei had minimum inhibitory concentration of 0.25 mg/L to caspofungin.
N indicates number of isolates tested for sensitivity; S, sensitive; I, intermediate; R, resistant.

Candida Colonization

In 40% (55/139) of episodes, there was evidence of colonization with Candida spp. in the 30 days before candidemia (Table 4). The most common sites of colonization were stool (23/55, 42%), skin (14/55, 26%), respiratory tract (13/55, 24%) and the oropharynx (8/55, 15%). Of those colonized, 33/55 (60%) had the same Candida spp. isolated from blood. Patients who had more than one episode of candidemia had similar rates of colonization to those with a single episode (7/15, 47% vs. 48/124, 39%, respectively, P = 0.55). Stool colonization rates were higher in neutropenic patients (15/34, 44% neutropenic vs. 8/100, 8%, non-neutropenic, P < 0.01). Also, a higher proportion of colonized patients received antifungal prophylaxis compared with those who were not colonized (30/59, 51% vs. 25/80, 31%, P = 0.02). Overall, of the 463 patients admitted to the oncology ward with positive stool samples for Candida spp. between January 2002 and July 2018, 5% (22/463 95% CI: 3.0–7.1) developed candidemia.

TABLE 4. - Colonization With Candida Spp. in 30 Days Before Candidemia
All (n = 139)
Colonization, n (%) Same Candida spp. Colonization and Blood, n (%)
Colonization 30 days before candidemia* 55 (40) 33 (24)
 Stool 23 (17) 10 (7)
 Skin 14 (10) 10 (7)
 Respiratory 13 (10) 7 (5)
 Oropharynx 8 (6) 5 (4)
 Drain 5 (4) 5 (4)
 Central venous catheter tip 3 (2) 3 (2)
 Gastrointestinal 2 (1) 0
*Some episodes had multiple sites of colonization.
Gastrointestinal included nasogastric aspirate and peritoneal fluid.

Treatment

Amphotericin B was the most common empiric therapy for candidemia (87/139, 63%), followed by an azole (37/139, 27%; 36 fluconazole, 1 voriconazole) and echinocandins (14/139, 10%) (one child was palliated before therapy). The median duration of treatment was 20 (range 0–210) days (note: one child died on the day of candidemia diagnosis), with a median duration of treatment after clearance of candidemia of 15 (range 0–204) days.

Clinical Outcomes and Mortality

The median duration of candidemia was 3 (range 1–40) days with no difference in the duration in neonates (3 days, range 1–11) compared with infants (2 days, range 1–9) and children (3 days, range 1–40) (P = 0.37). Two patients did not have a negative blood culture to confirm clearance of candidemia, and for these, the date of discharge was used (duration 4 and 5 days).

When CVCs were present before developing candidemia, they were removed within 2 days in 37/130 (28%) and within 7 days of candidemia in 77/130 (59%) episodes. The duration of candidemia was longer in patients who retained a CVC for 3 or more days compared with those who had them removed within 2 days (median 4 days, interquartile range 2–6 vs. median 2 days, interquartile range 1–3, P < 0.01). There was no difference in the duration of candidemia for different empiric antifungal agents (amphotericin B median 3 days, range 1–16; azole median 4 days, range 1–40; echinocandin median 3 days, range 1–7; P = 0.16).

All-cause 30-day mortality rate was 12% (16/139, 95% CI: 6.7–18.0). There was no difference in 30-day mortality rate in neonates (3/13, 23%, 95% CI: 5.0–53.8), infants (3/17, 18%, 95% CI: 3.8–43.4) or children (10/109, 9%, 95% CI: 4.5–16.2; P = 0.18) (Table 5). The 30-day mortality attributable to candidemia was 5% (7/139). Significant risk factors for all-cause 30-day mortality in the multivariate model were male sex (P = 0.01), liver disease (P = 0.03) and mucositis (P = 0.03). There was no difference in the mortality rate for episodes due to C. albicans versus NAC (8/74 vs. 8/66, respectively, P = 0.73; episodes with 2 Candida spp. excluded) or for different empirical antifungal regimens (amphotericin 9/87, 10%; azole 4/37, 11%; echinocandin 2/12, 17%; P = 0.84).

TABLE 5. - Mortality Univariate Analysis
Died (n = 16) Survived (n = 123) P
Age, median (range) 5.2 y (25 d to 18 y) 6.2 y (3 d to 18 y) 0.56
Age group
 Neonate (n = 13), n (% neonates) 3 (23) 10 (77)
 Infant (n = 17), n (% infants) 3 (18) 14 (82)
 Children, n = 110, n (% children) 10 (9) 99 (90)
Male sex, n (%) 13 (81) 63 (51) 0.03*
Comorbidity
 Gastrointestinal disease, n (%) 9 (56) 56 (45) 0.42
 Respiratory disease, n (%) 8 (50) 44 (36) 0.27
 Malignancy, n (%) 5 (31) 46 (37) 0.63
  Hematologic, n (%) 3 (19) 31 (25) 0.57
  So lid, n (%) 2 (13) 15 (12) 0.97
 Cardiac disease, n (%) 2 (13) 26 (21) 0.42
 Renal disease, n (%) 5 (31) 16 (13) 0.07*
 Neurologic disease, n (%) 2 (13) 13 (11) 0.81
 Congenital disease, n (%) 2 (13) 9 (7) 0.48
 Liver disease, n (%) 3 (19) 8 (7) 0.10*
 Hematopoietic stem cell transplant, n (%) 2 (13) 5 (4) 0.17*
 Diabetes, n (%) 0 4 (3)
 Premature, n (% of neonates and infants) 2 (8) 11 (46) 0.49
Risk factors
 Central venous catheter, n (%) 14 (88) 116 (94) 0.31
 Total parenteral nutrition, n (%) 9 (56) 51 (42) 0.27
 Bacteremia (30 d prior), n (%) 7 (44) 36 (30) 0.25
 Neutropenia (<0.5 × 109/L), n (%) 3 (20) 31 (27) 0.61
 Mucositis, n (%) 5 (31) 17 (14) 0.08*
 Burns/trauma, n (%) 0 3 (2)
 Antibiotic (30 d prior), n (%) 16 (100) 104 (85)
 Immunosuppression (30 d prior), n (%) 7 (44) 54 (44) 0.99
  Chemotherapy, n (%) 5 (31) 45 (37) 0.68
  Corticosteroid, n (%) 6 (38) 34 (28) 0.38
Site of infection
 Renal fungal disease, n (%) 1 (11) 8 (8) 0.74
 Fungal endophthalmitis, n (%) 0 8 (9)
 Hepatosplenic nodules, n (%) 0 5 (4)
 Pulmonary nodules, n (%) 0 9 (7)
Other
 CVC removal <7 d, n (%) 5 (31) 70 (57) 0.33
 Fever during candidemia, n (%) 8 (67) 102 (85) 0.12
 Colonization (30 d prior), n (%) 7 (44) 48 (39) 0.72
  Same Candida spp. colonizing and causing candidemia, n (%) 4 (25) 29 (24) 0.90
*Significant factors added to multi-logistic regression.
No info available for: gestational age and birth weight 1 neonate and 7 infants, neutropenia 5 episodes, bacteremia 1 episode, fever 6 episodes.

DISCUSSION

To our knowledge, this study describes the epidemiology of candidemia in children over the longest period to date. Our study showed low rates of antifungal resistance and mortality. However, it demonstrated higher rates of disseminated infection involving the kidneys, liver, spleen and lungs in episodes of persistent candidemia (>3 days post-CVC removal). Also, consistent with previous literature, we found that the duration of candidemia was longer in episodes where the CVC was retained for 3 or more days.

Our study found an increase in NAC over the 16.5-year period, mainly due to a rise in C. krusei, and low rates of antifungal resistance. These results are consistent with other studies in children and adults that have reported an increase in NAC, although these studies reported a predominant increase in C. glabrata complex.19–22C. krusei has previously been found to be the most common NAC causing candidemia in patients with hematologic malignancy, consistent with our study.23–25 However, while there was an increasing proportion of children with hematologic malignancies in our study, this does not entirely explain the rise in C. krusei at our center. Another potential driver may be the use of fluconazole prophylaxis given the inherent resistance of C. krusei to this agent.23,26,27 This has implications for antifungal prophylaxis in patients with high-risk hematologic malignancies. Although multi-drug resistant C. auris is a major public health concern, no C. auris was isolated in this study nor have there been any reported episodes of C. auris in children in Australia.9C. auris can be difficult to identify using the VITEK 2 YST card, however, the 2 species (C. haemulonii and C. duobushaemulonii) that C. auris may be misidentified as were not found in our study.

The finding of low rates of ocular and renal disease is consistent with recent studies showing that only 0%–7% of children with candidemia had disseminated disease.28–31 Screening for fungal endophthalmitis is crucial given the limited ocular penetration of amphotericin B and caspofungin and the need for treatment with intraocular antifungal therapy.32

A previous study of infants showed that prolonged candidemia for >7 days increased the odds of comorbid fungal renal disease [P = 0.03; odds ratio 7.57 (3.59–16.0)] and fungal ball uropathy [P = 0.01; odds ratio 6.00 (1.29–28.00)].29 Similarly, we found that disseminated infection was more likely to occur in children with persistent candidemia for more than 3 days after CVC removal. Despite low rates, screening for end-organ disease is important in determining the appropriate duration of antifungal therapy or requirement for surgery.32

The need for early removal (within 2 days) of the CVC in candidemic children was highlighted in our study. This is supported by a small cohort study of 50 infants with candidemia that reported a shorter duration of candidemia when the CVC was removed within 3 days (3 vs. 6 days, P = 0.0002)33 as well as a study in adults (2.6 vs. 5.6 days, P < 0.001).34 Although Candida spp. can persist on CVCs through biofilm formation these studies may be confounded by the severity of illness of the patient, with more critically unwell patients more likely to retain their CVC and have impaired responses to infection.33,34 Interestingly, CVC removal within 7 days of candidemia was not associated with a reduction in mortality in our study, despite many studies in children showing a survival advantage with removal.23,35 This may be due to the lower mortality rate in our study (12%) compared with previous studies (17.2%–20%).23,35

Stool colonization in children with candidemia has been reported in 2 previous retrospective studies which found that 39.0% and 45.6% of patients were colonized.35,36 Of note, one of these studies also found that 39% of children with candidemia had prior colonization compared with 4% of non-candidemic controls, although the time period for prior colonization was not defined.36 This is supported by a small case-control study of 24 children with acute lymphoblastic leukemia that showed that a high stool burden (>4-log) of Candida spp. was associated with a 25-fold increased risk of developing candidemia.12 In this study, vancomycin exposure increased the amount of Candida spp. in stool, however, this high stool burden of Candida was prevented when patients were treated with both vancomycin and amphotericin B.12 However, 42% of the patients in our study developed candidemia despite receiving antifungal prophylaxis and large prospective studies are needed to determine the role of antifungal treatment for colonized patients.

The limitations of our study include the retrospective nature therefore some clinical data, including the presence of skin lesions, were missing. Our dataset included 9 episodes where Candida spp. was isolated from blood culture and yet considered contaminants. Although 8 of the 9 children had a subsequent negative blood culture despite not receiving antifungal therapy, it is unclear from the case notes why these were considered contaminants at the time. Further, cultures of other sites to determine Candida colonization were not routinely obtained. In addition, this manuscript describes the epidemiology of candidemia at a single site.

CONCLUSIONS

This 16.5-year longitudinal study found an increase in NAC in children with candidemia primarily due to an increase in C. krusei. Disseminated candidiasis including pulmonary, hepatosplenic, renal and ophthalmologic involvement was rare. Antifungal resistance remained low; however, with resistance emerging, ongoing surveillance for resistant strains remains important to guide antifungal therapy. Finally, our study suggests that prior stool colonization may be an important risk factor for candidemia, and further large prospective studies are required to determine the role of preemptive antifungal therapy in selected colonized patients.

REFERENCES

1. Wisplinghoff H, Seifert H, Tallent SM, et al. Nosocomial bloodstream infections in pediatric patients in United States hospitals: epidemiology, clinical features and susceptibilities. Pediatr Infect Dis J. 2003;22:686–691.
2. Zaoutis TE, Argon J, Chu J, et al. The epidemiology and attributable outcomes of candidemia in adults and children hospitalized in the United States: a propensity analysis. Clin Infect Dis. 2005;41:1232–1239.
3. Raymond J, Aujard Y. Nosocomial infections in pediatric patients: a European, multicenter prospective study. European Study Group. Infect Control Hosp Epidemiol. 2000;21:260–263.
4. Harrington R, Kindermann SL, Hou Q, et al. Candidemia and invasive candidiasis among hospitalized neonates and pediatric patients. Curr Med Res Opin. 2017;33:1803–1812.
5. Chapman B, Slavin M, Marriott D, et al.; Australian and New Zealand Mycoses Interest Group. Changing epidemiology of candidaemia in Australia. J Antimicrob Chemother. 2017;72:1103–1108.
6. Chen S, Slavin M, Nguyen Q, et al.; Australian Candidemia Study. Active surveillance for candidemia, Australia. Emerg Infect Dis. 2006;12:1508–1516.
7. Celebi S, Hacimustafaoglu M, Ozdemir O, et al. Nosocomial candidaemia in children: results of a 9-year study. Mycoses. 2008;51:248–257.
8. Lee WG, Shin JH, Uh Y, et al. First three reported cases of nosocomial fungemia caused by Candida auris. J Clin Microbiol. 2011;49:3139–3142.
9. Warris A. Candida auris, what do paediatricians need to know? Arch Dis Child. 2018;103:891–894.
10. Benjamin DK Jr, Poole C, Steinbach WJ, et al. Neonatal candidemia and end-organ damage: a critical appraisal of the literature using meta-analytic techniques. Pediatrics. 2003;1123 pt 1634–640.
11. Viudes A, Pemán J, Cantón E, et al. Candidemia at a tertiary-care hospital: epidemiology, treatment, clinical outcome and risk factors for death. Eur J Clin Microbiol Infect Dis. 2002;21:767–774.
12. Richet HM, Andremont A, Tancrede C, et al. Risk factors for candidemia in patients with acute lymphocytic leukemia. Rev Infect Dis. 1991;13:211–215.
13. Samonis G, Gikas A, Anaissie EJ, et al. Prospective evaluation of effects of broad-spectrum antibiotics on gastrointestinal yeast colonization of humans. Antimicrob Agents Chemother. 1993;37:51–53.
14. Nucci M, Anaissie E. Revisiting the source of candidemia: skin or gut? Clin Infect Dis. 2001;33:1959–1967.
15. Blyth CC, Chen SC, Slavin MA, et al.; Australian Candidemia Study. Not just little adults: candidemia epidemiology, molecular characterization, and antifungal susceptibility in neonatal and pediatric patients. Pediatrics. 2009;123:1360–1368.
16. Nucci M, Queiroz-Telles F, Alvarado-Matute T, et al.; Latin American Invasive Mycosis Network. Epidemiology of candidemia in Latin America: a laboratory-based survey. PLoS One. 2013;8:e59373.
17. Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;49:1–45.
18. Clinical and Laboratory Standards Institute. Method for Antifungal Disk Diffusion Susceptibility Testing of Yeasts. 2018. 3rd ed. Clinical and Laboratory Standards Institute; (CLSI guideline M44).
19. Dzierzanowska-Fangrat K, Romanowska E, Gryniewicz-Kwiatkowska O, et al. Candidaemia in a Polish tertiary paediatric hospital, 2000 to 2010. Mycoses. 2014;57:105–109.
20. Neu N, Malik M, Lunding A, et al. Epidemiology of candidemia at a Children’s hospital, 2002 to 2006. Pediatr Infect Dis J. 2009;28:806–809.
21. Trick WE, Fridkin SK, Edwards JR, et al.; National Nosocomial Infections Surveillance System Hospitals. Secular trend of hospital-acquired candidemia among intensive care unit patients in the United States during 1989-1999. Clin Infect Dis. 2002;35:627–630.
22. Pfaller MA, Diekema DJ, Turnidge JD, et al. Twenty years of the SENTRY antifungal surveillance program: results for Candida species from 1997-2016. Open Forum Infect Dis. 2019;6(suppl 1):S79–S94.
23. Gokcebay DG, Yarali N, Isik P, et al. Candida associated bloodstream infections in pediatric hematology patients: a single center experience. Mediterr J Hematol Infect Dis. 2016;8:e2016018.
24. Muñoz P, Sánchez-Somolinos M, Alcalá L, et al. Candida krusei fungaemia: antifungal susceptibility and clinical presentation of an uncommon entity during 15 years in a single general hospital. J Antimicrob Chemother. 2005;55:188–193.
25. Abbas J, Bodey GP, Hanna HA, et al. Candida krusei fungemia. An escalating serious infection in immunocompromised patients. Arch Intern Med. 2000;160:2659–2664.
26. Berrouane YF, Herwaldt LA, Pfaller MA. Trends in antifungal use and epidemiology of nosocomial yeast infections in a university hospital. J Clin Microbiol. 1999;37:531–537.
27. Orozco AS, Higginbotham LM, Hitchcock CA, et al. Mechanism of fluconazole resistance in Candida krusei. Antimicrob Agents Chemother. 1998;42:2645–2649.
28. Noyola DE, Fernandez M, Moylett EH, et al. Ophthalmologic, visceral, and cardiac involvement in neonates with candidemia. Clin Infect Dis. 2001;32:1018–1023.
29. Chapman RL, Faix RG. Persistently positive cultures and outcome in invasive neonatal candidiasis. Pediatr Infect Dis J. 2000;19:822–827.
30. Donahue SP, Hein E, Sinatra RB. Ocular involvement in children with candidemia. Am J Ophthalmol. 2003;135:886–887.
31. Fierro JL, Prasad PA, Fisher BT, et al. Ocular manifestations of candidemia in children. Pediatr Infect Dis J. 2013;32:84–86.
32. Pappas PG, Walsh TJ, Zaoutis TE, et al. Clinical practice guideline for the management of candidiasis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis. 2015;62:e1–e50.
33. Karlowicz MG, Hashimoto LN, Kelly RE Jr, et al. Should central venous catheters be removed as soon as candidemia is detected in neonates? Pediatrics. 2000;106:E63.
34. John HR, John EB, Alan MS, et al. Intravascular catheter exchange and duration of candidemia. Clin Infect Dis. 1995;21:994.
35. Sutcu M, Salman N, Akturk H, et al. Epidemiologic and microbiologic evaluation of nosocomial infections associated with Candida spp in children: a multicenter study from Istanbul, Turkey. Am J Infect Control. 2016;44:1139–1143.
36. García San Miguel L, Pla J, Cobo J, et al. Morphotypic and genotypic characterization of sequential Candida parapsilosis isolates from an outbreak in a pediatric intensive care unit. Diagn Microbiol Infect Dis. 2004;49:189–196.
Keywords:

candidemia; Candida; pediatric; colonization; resistance

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.