Temporal trends and clinical outcomes of pediatric invasive fungal diseases: a ten-year retrospective study from a tertiary-care center in Thailand Authors: Chanittha Niyomthammarat, Kornvalee Meesilpavikkai, Wariya Chintanapakdee, Jiratchaya Sophonphan, Suvaporn Anugulruengkitt, Thanyawee Puthanakit, Watsamon Jantarabenjakul Published: 2026 ABSTRACT BackgroundInvasive fungal diseases (IFDs) remain an important cause of morbidity and mortality among immunocompromised children. Changes in diagnostic criteria and increasing use of antifungal prophylaxis may influence the epidemiology and clinical presentation of pediatric IFDs. Longitudinal data from Southeast Asia remain limited. This study aimed to characterize temporal trends, clinical features, and outcomes of pediatric IFDs using the 2020 EORTC/MSGERC criteria.MethodsWe retrospectively reviewed children aged 0–15 years diagnosed with IFDs at a tertiary-care hospital in Bangkok, Thailand, from 2014 to 2023. IFD episodes were classified as proven, probable, or possible according to the 2020 EORTC/MSGERC criteria and categorized by causative organism. Temporal trends, clinical characteristics, antifungal prophylaxis exposure, and outcomes were analyzed.ResultsAmong 156 patients, 171 IFD episodes were identified: invasive aspergillosis (60 episodes), invasive candidiasis (61), Pneumocystis pneumonia (33), cryptococcosis (5), other invasive mold infections (8), and endemic mycoses (4). Invasive aspergillosis and Pneumocystis pneumonia demonstrated greater year-to-year variability with higher case numbers in later years, whereas invasive candidiasis occurred consistently throughout the study. A proportion of IFD episodes occurred despite antifungal prophylaxis, including 23.3% of invasive aspergillosis, 8.2% of invasive candidiasis, and 12.1% of Pneumocystis pneumonia. Hematologic malignancies were the most common underlying conditions among children with invasive aspergillosis (68.3%) and Pneumocystis pneumonia (27.3%). Invasive candidiasis primarily affected infants and young children with chronic conditions, including chronic liver or gastrointestinal disease (25.4%) and congenital heart disease (25.4%), and was strongly associated with central venous catheter use (88.9%). Candida albicans remained the most common pathogen, with overall low antifungal resistance. The clinical manifestations were non-specific and overlapped across pathogens, most commonly presenting with fever and organ-specific symptoms. Ninety-day all-cause mortality was 20.3% for aspergillosis, 24.8% for candidiasis, and 12.9% for Pneumocystis pneumonia.ConclusionsPediatric IFDs remained associated with substantial mortality despite evolving diagnostic criteria and increasing use of antifungal prophylaxis, underscoring the need for improved diagnostic and preventive strategies.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12879-026-13849-8. Background Invasive fungal diseases (IFDs) are important causes of morbidity and mortality among immunocompromised and critically ill children. In pediatric populations, invasive candidiasis and invasive aspergillosis remain the most frequently encountered IFDs, with reported mortality rates ranging from 10% to 40%, depending on host factors, pathogen, and access to timely diagnosis and treatment [1–6]. The epidemiology of pediatric IFDs varies substantially across geographic regions, patient populations, and time periods, underscoring the importance of local longitudinal data to guide prevention strategies, diagnostic approaches, and empirical antifungal therapy [3]. Over the past decade, substantial changes have occurred in the classification and diagnosis of IFDs. The revised 2020 EORTC/MSGERC definitions broadened host and clinical criteria, incorporated newer mycological tools, and refined disease entities relevant to pediatric and intensive care unit populations. While these updates have improved diagnostic consistency, they complicate comparisons with earlier pediatric studies, in which invasive procedures and microbiologic confirmation are often limited [7, 8]. Concurrently, antifungal prophylaxis strategies in children with hematologic malignancies and other high-risk conditions have evolved, shifting from predominantly fluconazole-based regimens to risk-adapted use of mold-active agents [9, 10]. While these strategies have reduced the incidence of some fungal infections, concerns remain regarding suboptimal drug exposure, variable absorption, and the occurrence of IFDs despite prophylaxis, which may influence pathogens’ distribution and clinical outcomes in contemporary pediatric practice. In Thailand and other parts of Southeast Asia, long-term data describing pediatric IFDs remain scarce. Previous single-center studies from our institution reported substantial mortality associated with invasive aspergillosis and invasive candidiasis using earlier diagnostic criteria; however, data incorporating updated definitions and recent prophylaxis practices are limited [11, 12]. Temporal trends across multiple fungal pathogens in children have not been well characterized. To address these gaps, we conducted a retrospective study describing a 10-year experience with pediatric IFDs at a large tertiary-care hospital in Thailand. The primary objective was to characterize temporal trends, clinical features, and outcomes of pediatric IFDs using the 2020 EORTC/MSGERC criteria. Secondary objectives included describing the distribution of organisms, antifungal susceptibility patterns among Candida species, and the proportion of IFD episodes diagnosed during antifungal prophylaxis. Methods Study design and setting We conducted a retrospective observational study at King Chulalongkorn Memorial Hospital (KCMH), a 1,500-bed tertiary-care academic medical center in Bangkok, Thailand, with 300 pediatric inpatient beds, 20 beds neonatal intensive care unit (NICU), 8 beds Pediatric intensive care unit and a dedicated pediatric hematology/oncology and other immunocompromised unit. Furthermore, the institution serves as a major referral center housing advanced cellular therapy, including hematopoietic stem cell transplantation and solid organ transplant programs, managing highly complex and critically ill patient population. We enrolled hospitalized children aged 0–15 years diagnosed with invasive fungal diseases (IFDs) between January 1, 2014, and December 31, 2023. Case identification and definitions Potential IFD cases were identified through review of International Classification of Diseases, Tenth Revision (ICD-10) codes, microbiology laboratory records (including fungal cultures, histopathology, and cytopathology), and fungal biomarker results, including serum and bronchoalveolar lavage (BAL) galactomannan, serum β-D-glucan, cryptococcal antigen testing, and Pneumocystis jirovecii polymerase chain reaction (PCR). Each IFD episode was classified as proven, probable, or possible according to the 2020 EORTC/MSGERC definitions [7, 8]. For Pneumocystis infection, cases were classified as proven (microscopic visualization), probable (positive PCR or β-D-glucan with typical radiology), or possible (highly suggestive clinical/radiological features with negative PCR, managed as clinical PCP). Episodes that did not meet the required host factors or clinical criteria were excluded after chart review. IFDs were categorized by causative organism into six groups: invasive aspergillosis, invasive candidiasis, Pneumocystis infection, cryptococcosis, other invasive mold diseases, and endemic or rare fungal infection (including rare yeasts and endemic mycoses). Patients could contribute more than one episode if infections were caused by different organisms or occurred at least 30 days apart, provided there was documented clinical improvement or microbiological clearance between episodes. Clinical data collection Demographic and clinical data were extracted from electronic medical records, including age, sex, hospital unit, underlying diseases, and clinical features within seven days of IFD diagnosis. Predisposing factors were collected according to organism group. For invasive aspergillosis, variables included neutropenia (absolute neutrophil count < 500 cells/mm³ for ≥ 10 consecutive days within one month before diagnosis), exposure to chemotherapy or immunosuppressive agents, intensive care unit (ICU) admission, mechanical ventilation, central venous catheter use, and prior broad-spectrum antibiotic use. For invasive candidiasis, additional factors included urinary catheterization, recent surgery, total parenteral nutrition, and Candida colonization. For Pneumocystis infection, variables included lymphopenia (absolute lymphocyte count < 1,500 cells/mm³), prior broad-spectrum antibiotic use, and trimethoprim–sulfamethoxazole prophylaxis. Antifungal prophylaxis exposure was defined as receipt of any systemic antifungal agent at the time of IFD diagnosis. Data on duration, adherence, and therapeutic drug monitoring were unavailable and not analyzed. Microbiological methods Blood cultures For fungal hemoculture, blood specimens were collected into Bactec Peds Plus/F bottles or Bactec Myco/F Lytic bottles (BD). Upon a positive signal, aliquots were examined by direct microscopy (Gram stain with acid-fast bacilli stain when organisms were not visualized) and subcultured onto blood agar and Sabouraud-based media, with additional CHROMagar subculture when budding yeasts were observed. Subculture tubes were incubated at 25 ± 2 °C and 35 ± 2 °C, and Myco/F bottles were retained for 30 days. Candida albicans was identified phenotypically during 2003–2007. From 2008 onward, yeasts were identified using API 20 C AUX (bioMérieux, France), and since 2017, species identification has been routinely performed using the Vitek 2 XL with the YST card [13]. Antifungal drug susceptibility Before December 2019, susceptibility testing was performed using the Etest method; therefore, micafungin MICs data were unavailable. From December 2019 onward, broth microdilution was performed using Sensititre™ YeastOne™ (Thermo Scientific). Isolates were standardized to 0.5 McFarland, inoculated into YeastOne broth, and incubated at 35 °C for 24–25 h. MICs were read using the Sensititre Vizion™ and interpreted with SWIN™ software. The panel included anidulafungin, micafungin, 5-fluorocytosine, amphotericin B, fluconazole, itraconazole, caspofungin, voriconazole, and posaconazole [13]. Antifungal susceptibility and resistance were primarily interpreted using the European Committee on Antimicrobial Susceptibility Testing (EUCAST) v11.0 (2024) criteria [14]. However, because EUCAST does not provide clinical breakpoints for caspofungin, the Clinical and Laboratory Standards Institute (CLSI) criteria were applied specifically to interpret caspofungin MIC values for Candida isolates, classifying them as susceptible, intermediate, or resistant based on species-specific CLSI thresholds [15, 16]. Fungal culture Clinical specimens were examined by direct microscopy using a KOH wet mount when indicated and inoculated onto blood agar and selective Sabouraud-based media together with Sabouraud dextrose agar tubes. The cultures were incubated at 25 ± 2 °C and 35 ± 2 °C for up to 30 days and examined three times weekly. Growth was reported as yeast/mold followed by organism identification. Molds identification was based on colony morphology and microscopic features on lactophenol cotton blue stain, with PCR-based identification from pure culture when confirmatory testing was required [17]. Galactomannan and cryptococcal antigen detection Aspergillus galactomannan (GM) antigen testing has been available since January 2005 and was performed on serum and bronchoalveolar lavage (BAL) specimens using the PLATELIA™ Aspergillus EIA kit (Bio-Rad). A GM index ≥ 0.5 was considered positive. Cryptococcal antigen (CrAg) testing was performed on serum and cerebrospinal fluid (CSF) using a lateral flow immunochromatographic assay. Titers were determined using a latex agglutination assay when clinically indicated. PCR for fungus Molecular identification was performed by PCR amplification of the internal transcribed spacer (ITS) and D1/D2 regions, followed by Sanger sequencing. Genomic DNA was extracted from single-colony cultures, purified, and quantified prior to sequencing [17]. Outcomes The primary outcome was all-cause mortality at 30 days, 90 days, and 12 months after IFD diagnosis. Deaths were classified as all-cause mortality regardless of attribution. Statistical analysis Continuous variables were summarized as medians with interquartile ranges (IQRs), and categorical variables as frequencies and percentages. To analyze temporal trends from 2014 to 2023, the annual incidence rates of specific IFDs were calculated and expressed as the number of cases per 10,000 pediatric admissions to account for variations in hospital volume over time. To determine the disease-specific burden, the prevalence of IA within high-risk hematologic malignancy subgroups -specifically newly diagnosed ALL, AML, and relapse ALL/AML was calculated. The denominator for each subgroup was defined as the total number of pediatric patients with that specific oncology cohort who were admitted or received treatment at our center during the 10-year study period. Similarly, the subgroup-specific prevalence of PCP was determined for high-risk pediatric cohorts, specifically those with SLE and ALL. Mortality was estimated using the Kaplan–Meier method and compared using the log-rank test. Poisson regression analysis was used to identify factors associated with mortality, with results reported as incidence rate ratios (IRRs) with 95% confidence intervals (CIs). Variables with a P value < 0.20 in univariable analyses were included in the multivariable model. All tests were two-sided, and statistical significance was defined as P < 0.05. Analyses were performed using Stata version 19.5 (StataCorp, College Station, TX). Results Study population and case classification Between January 2014 and December 2023, 168 children (183 episodes) with invasive fungal diseases (IFDs) were identified. After exclusion of 12 episodes that did not fulfill the 2020 EORTC/MSGERC host or clinical criteria and one episode with an alternative diagnosis, 156 patients with 171 evaluable IFD episodes were included (Fig. 1). Some experienced more than one episode, caused by the same or different organisms. The 171 IFD episodes were classified into six organism groups: invasive aspergillosis (60episodes), invasive candidiasis (61 episodes in 55 patients), Pneumocystis infection (33 episodes), cryptococcosis (5 episodes), other invasive mold diseases (8 episodes), and endemic or rare fungal infection (4 episodes). Fig. 1Enrollment flow diagram of pediatric invasive fungal episodes. Enrollment flow diagram and classification of pediatric invasive fungal episodes, which were classified according to the 2020 EORTC/MSGERC criteria Enrollment flow diagram of pediatric invasive fungal episodes. Enrollment flow diagram and classification of pediatric invasive fungal episodes, which were classified according to the 2020 EORTC/MSGERC criteria Temporal trends of invasive fungal diseases The temporal trends of the three major IFDs, expressed as incidence rates per 10,000 admissions, are shown in Fig. 2. The incidence of invasive aspergillosis remained relatively stable between 2014 and 2018 (ranging from 6.6 to 8.2 per 10,000 admissions), followed by a sharp decline in 2019 and a marked, progressive increase in recent years, peaking at 13.3 per 10,000 admissions in 2023. Invasive candidiasis demonstrates a fluctuating pattern, with a notable peak in 2016 (10.1 per 10,000 admissions) before stabilizing, then rising to 12.0 per 10,000 admissions in 2023. Additionally, the incidence rate of PCP showed a distinct surge in 2020, reaching 10.2 per 10,000 admissions. By 2023, a simultaneous increase across all three major IFDs was observed, representing the highest overall burden of the decade under analysis. Fig. 2Temporal trends in the incidence rates of major invasive fungal diseases (IFDs) per 10,000 pediatric admissions from 2014 to 2023 Temporal trends in the incidence rates of major invasive fungal diseases (IFDs) per 10,000 pediatric admissions from 2014 to 2023 Baseline characteristics Baseline demographic and clinical characteristics are summarized in Table 1. Children with invasive aspergillosis and Pneumocystis infection were older (median ages 9.7 and 9.3 years, respectively), whereas invasive candidiasis predominantly affected younger children and infants (median age 1.1 years). Hematologic malignancy was the most common underlying condition among children with invasive aspergillosis (68.3%) and Pneumocystis infection (27.3%), while invasive candidiasis was most frequently associated with chronic liver or gastrointestinal disease and congenital heart disease. Pulmonary involvement predominated in invasive aspergillosis (83.3%) and Pneumocystis infection (100%), whereas bloodstream infection was the most common manifestation of invasive candidiasis (73.7%). A proportion of IFD episodes were diagnosed while patients were receiving antifungal prophylaxis, including 23.3% of invasive aspergillosis episodes, 8.2% of invasive candidiasis episodes, and 12.1% of Pneumocystis infection episodes. Table 1Baseline demographic and clinical characteristics of children with invasive aspergillosis, invasive candidiasis, and pneumocystis infectionBaseline characteristicAspergillosis(N = 60)Candidiasis(N = 55)Pneumocystis(N = 33)Age (years), median (IQR)9.7 (4.5–13.0)1.1 (0.2–3.5)9.3 (4.1–13.2)Sex, N (%) Female28 (46.7)24 (43.6)16 (48.5)Underlying, N (%) Hematologic malignancy41 (68.3)10 (18.2)9 (27.3) Solid organ malignancy1 (1.7)1 (1.8)6 (18.2) Transplantation3 (5.0)1 (1.8)4 (12.1) Autoimmune disease7 (11.7)2 (3.6)6 (18.2) Immunodeficiency disease5 (8.3)1 (1.8)3 (9.1) Chronic liver disease or GI disease2 (3.3)14 (25.5)#1 (3.0) Congenital heart disease014 (25.5)0 Preterm06 (10.9)0 Other4 (6.6)7 (12.7)6 (18.2)Site of infection, N (%) Pulmonary50 (83.3)0*33 (100) Blood045 (73.7)*0 Other organs2(3.4)7 (11.5)*0 Disseminated disease8 (13.3)9 (14.8)*0Clinical features, N (%) Fever50 (83.3)52 (85.2)*23 (69.7) Respiratory symptoms26 (43.3)0*32 (97.0) shock011(18.0)*2 (6.0)Breakthrough infection, N (%)(previous antifungal prophylaxis)14 (23.3)5 (8.2)*4 (12.1)Preexisting factor, N (%) Previous broad-spectrum antibiotics56 (93.3)55 (100)23 (69.7) Chemotherapy38 (63.3)11 (20.4)13 (39.4) Immunosuppressive drug14 (23.3)3 (5.5)15 (45.5) History of lymphopeniaNANA20 (62.5) History of neutropenia35 (63.6)9 (17.7)3 (9.1)Data are presented as number (%) unless otherwise indicated#Chronic liver disease or GI disease, including biliary atresia with cirrhosis, short bowel syndrome, TE fistula and omphalocele*Percentages for invasive candidiasis are calculated per episode (n = 61)Abbreviations: IQR, interquartile range; GI, gastrointestinal; ICU, intensive care unit; NA, not applicable Table 1Baseline demographic and clinical characteristics of children with invasive aspergillosis, invasive candidiasis, and pneumocystis infectionBaseline characteristicAspergillosis(N = 60)Candidiasis(N = 55)Pneumocystis(N = 33)Age (years), median (IQR)9.7 (4.5–13.0)1.1 (0.2–3.5)9.3 (4.1–13.2)Sex, N (%) Female28 (46.7)24 (43.6)16 (48.5)Underlying, N (%) Hematologic malignancy41 (68.3)10 (18.2)9 (27.3) Solid organ malignancy1 (1.7)1 (1.8)6 (18.2) Transplantation3 (5.0)1 (1.8)4 (12.1) Autoimmune disease7 (11.7)2 (3.6)6 (18.2) Immunodeficiency disease5 (8.3)1 (1.8)3 (9.1) Chronic liver disease or GI disease2 (3.3)14 (25.5)#1 (3.0) Congenital heart disease014 (25.5)0 Preterm06 (10.9)0 Other4 (6.6)7 (12.7)6 (18.2)Site of infection, N (%) Pulmonary50 (83.3)0*33 (100) Blood045 (73.7)*0 Other organs2(3.4)7 (11.5)*0 Disseminated disease8 (13.3)9 (14.8)*0Clinical features, N (%) Fever50 (83.3)52 (85.2)*23 (69.7) Respiratory symptoms26 (43.3)0*32 (97.0) shock011(18.0)*2 (6.0)Breakthrough infection, N (%)(previous antifungal prophylaxis)14 (23.3)5 (8.2)*4 (12.1)Preexisting factor, N (%) Previous broad-spectrum antibiotics56 (93.3)55 (100)23 (69.7) Chemotherapy38 (63.3)11 (20.4)13 (39.4) Immunosuppressive drug14 (23.3)3 (5.5)15 (45.5) History of lymphopeniaNANA20 (62.5) History of neutropenia35 (63.6)9 (17.7)3 (9.1)Data are presented as number (%) unless otherwise indicated#Chronic liver disease or GI disease, including biliary atresia with cirrhosis, short bowel syndrome, TE fistula and omphalocele*Percentages for invasive candidiasis are calculated per episode (n = 61)Abbreviations: IQR, interquartile range; GI, gastrointestinal; ICU, intensive care unit; NA, not applicable Baseline demographic and clinical characteristics of children with invasive aspergillosis, invasive candidiasis, and pneumocystis infection Breakthrough infection, N (%) (previous antifungal prophylaxis) Data are presented as number (%) unless otherwise indicated #Chronic liver disease or GI disease, including biliary atresia with cirrhosis, short bowel syndrome, TE fistula and omphalocele *Percentages for invasive candidiasis are calculated per episode (n = 61) Abbreviations: IQR, interquartile range; GI, gastrointestinal; ICU, intensive care unit; NA, not applicable Invasive aspergillosis Sixty episodes of invasive aspergillosis (IA) were identified, including 14 proven, 34 probable, and 12 possible cases. Most patients had underlying hematologic malignancies (68.3%), particularly acute myeloid leukemia (AML) (21.7%), acute lymphoblastic leukemia (ALL) (16.7%), or relapsed disease (13.3%). Among hematologic malignancies, IA prevalence was 4.8% (95% CI, 2.3–8.6%) in ALL (10/209) and 20.0% (95% CI 11.0–31.8%) in AML 13/65), peaking at 21.6% (95%CI 9.8–38.2%) in relapsed ALL/AML (8/37). Pulmonary infection was the predominant presentation (83.3%), while extrapulmonary involvement (3.4%) and disseminated diseases (13.3%) occurred mainly among proven cases. Among patients with invasive pulmonary aspergillosis, chest computed tomography commonly demonstrated wedge-shaped consolidation (57%), well-circumscribed dense lesions (34.6%), or cavitary lesions (23%); air-crescent signs were uncommon (6%). Overall, an elevated serum galactomannan (≥ 1) was detected in 24.5% (13/53) of tested cases; however, when excluding possible cases (which, by definition, lack mycological criteria), the positivity rate among proven and probable cases was 31.7% (13/41). Similarly, BAL galactomannan (≥ 1) was detected in 80.0% (28/35) of test cases, increasing to 93.3% (28/30) when evaluated strictly within the proven and probable subgroups. Concomitant elevation of both serum galactomannan (≥ 0.7) and BAL galactomannan (≥ 0.8) was observed in 15.1% (5/33) of test cases. Fourteen episodes occurred during antifungal prophylaxis. Most shortly after itraconazole initiation or in the setting of subtherapeutic drug levels. Voriconazole was used as definitive therapy in most patients (93.3%), whereas amphotericin B formulations were frequently used as initial empirical treatment. All-cause mortality at 30 days, 90 days, and one year was 11.7% (95%CI 5.7–22.9), 20.3% (95%CI 12.1–33.0), and 33.6% (95%CI 22.5–48.3), respectively (Fig. 3). Mortality was highest among proven cases. Elevated serum galactomannan was associated with increased mortality in univariable analysis (IRR 1.57 (95%CI 1.06–2.33)) (Supplement Table 1). Fig. 3Kaplan-Meier estimates of all-cause mortality among children with invasive aspergillosis, invasive candidiasis, and pneumocystis infection. Mortality was assessed at 30 days, 90 days, and 12 months following diagnosis of invasive fungal disease Kaplan-Meier estimates of all-cause mortality among children with invasive aspergillosis, invasive candidiasis, and pneumocystis infection. Mortality was assessed at 30 days, 90 days, and 12 months following diagnosis of invasive fungal disease Regarding deaths suspected to be directly attributable to IA, one case was confirmed by autopsy in a previously healthy child with severe influenza A infection complicated with acute liver failure and acute respiratory distress syndrome (ARDS). Two additional deaths were suspected to be IA-related, both in patients with proven disseminated disease. One who had systemic lupus erythematosus (SLE) and died from an intracerebral hemorrhage. The second patient, who was ALL, died from pulmonary hemorrhage. One fatal case was initially not diagnosed as invasive aspergillosis due to the absence of clinical features; however, disseminated invasive aspergillosis was subsequently confirmed by autopsy. Invasive candidiasis Fifty-five patients experienced 61 episodes, all of which were classified as proven disease. Most episodes (90.1%) occurred in infants and young children (Table 1). The most common underlying conditions were chronic liver or gastrointestinal disease (25.4%) and congenital heart disease (25.4%). Bloodstream infection accounted for 73.7% of episodes, and central venous catheters were present in 88.9%. Five episodes (8.2%) in four patients occurred during antifungal prophylaxis, including fluconazole prophylaxis in two patients and itraconazole prophylaxis in two patients. Serum β-D-glucan testing was not performed because of the prolonged turnaround time. Candida albicans remained the most common pathogen, followed by Candida tropicalis (Fig. 4). Overall antifungal resistance rates were low; however, C. tropicalis demonstrated higher rates of azole resistance (3 of 14 reported resistance to fluconazole and 7 of 8 reported resistance to itraconazole) compared with C. albicans (1 of 19 reported resistance to fluconazole and no reports of resistance to itraconazole). C. parapsilosis was all susceptible to micafungin. Infections diagnosed while patients were receiving antifungal prophylaxis occurred in 8.2% of episodes, including cases involving azole-resistant or non-albicans Candida species. Fig. 4Distribution of Candida species causing invasive candidiasis overtime among hospitalized children. Data adapted from previously published institutional data [11] and the current study cohort Distribution of Candida species causing invasive candidiasis overtime among hospitalized children. Data adapted from previously published institutional data [11] and the current study cohort The 30-day mortality rate (95%CI) was 16.6% (9.0-29.5), increasing to 24.8% (15.2–38.9) at 90 days and 31.5% (20.5–46.3) at one year (Fig. 3). No clinical or demographic factors were significantly associated with mortality (Supplement Table 2). Regarding cases with documentation of candidiasis-related mortality, one death was directly attributable to invasive candidiasis, with disseminated C. tropicalis infection confirmed at autopsy in a patient with hemophagocytic lymphohistiocytosis. Three additional deaths were suspected to be related to candidemia. Pneumocystis infection Thirty-three episodes of Pneumocystis infection (PCP) were identified, comprising 5 proven, 22 probable, and 6 possible cases. All presenting with pulmonary disease. Underlying conditions included hematologic malignancies, solid organ malignancies, autoimmune diseases, and transplantation (Table 1). The 6 possible cases demonstrated typical clinical and radiological features of PCP but had negative PCR results, often after initiating empiric anti-PCP therapy. During the 10-year study period, PCP occurred in 5 of 143 patients with SLE (3.5%; 95%CI, 1.1-8.0%) and in 8 of 219 patients with ALL (3.7%; 95% CI, 1.6–7.1%). Infections diagnosed while patients were receiving trimethoprim–sulfamethoxazole prophylaxis occurred in 12.1% of cases, including patients with hematologic malignancy, HIV infection, and post–liver transplantation, but none of the SLE patients had received trimethoprim–sulfamethoxazole. All-cause mortality was 0% at 30 days, 12.9% (95%CI 5.1–30.8) at 90 days, and 20.2% (95%CI 9.6–39.6) at one year (Fig. 3). Autoimmune disease was associated with an increased risk of mortality (IRR 10.93(1.00-59.69)) (Supplement Table 3). One patient with medulloblastoma was confirmed to be attributable to PCP, based on lung necropsy. Other invasive fungal diseases A total of 17 cases of other IFDs were identified, comprising cryptococcosis (n = 5), other invasive mold (n = 8), and endemic or rare fungal infections (n = 4). Among the cryptococcosis cases, three occurred in children with HIV infection, while the others had osteosarcoma or ALL; these patients presented with central nervous system (n = 2), pulmonary (n = 2), or bloodstream infection (n = 1), with 100% survival reported among those with available follow-up. The other invasive mold disease group included mucormycosis (n = 3), Exserohilum (n = 2), Curvularia (n = 1), and Basidiobolus (n = 1), and zygomycota (n = 1), affecting both immunocompromised hosts – such as those with ALL or post-hematopoietic stem cell transplant and two previously healthy children. These infections primarily involved the sinuses and subcutaneous tissues, with one death occurring at the one-year follow-up in a post-transplant patient. In contrast, the endemic and rare fungal group showed high mortality; all three patients with Penicillium marneffei and Trichosporon spp. infections died within 30 days, while one patient with disseminated pythiosis survived. Discussion In this 10-year retrospective study, we describe the epidemiology, clinical characteristics, and outcomes of invasive fungal diseases (IFDs) among hospitalized children at a tertiary care center in Thailand, using the 2020 EORTC/MSGERC definitions. Despite advances in diagnostics and evolving antifungal prophylaxis practices, pediatric IFDs remained associated with considerable morbidity and mortality. Our findings highlight ongoing challenges in the prevention, diagnosis, and management of immunocompromised children, particularly those with hematologic malignancies and complex chronic conditions. Regarding temporal trends, our study highlighted a notable surge in the incidence rate of PCP in 2020, which might be relative to previous years, was primarily driven by a concomitant drop in total annual hospital volume (7,828 admissions) during the COVID-19 pandemic, rather than a massive influx of absolute cases. Furthermore, the remarkable surge in the incidence of invasive fungal disease in 2023, particularly invasive aspergillosis, could be largely attributed to a critical clinical challenge of a shortage of itraconazole oral solution, forcing a temporary switch to the tablet formulation for anti-fungal prophylaxis, which is well-known to have highly unpredictable bioavailability and often fails to achieve therapeutic serum concentrations required for effective prophylaxis. The European Conference on Infections in Leukemia (ECIL) has recently downgraded its recommendation for itraconazole prophylaxis in patients at high risk of hematological malignancies from Grade B to Grade C [18]. To mitigate this issue and optimize patient outcomes, our center is currently implementing voriconazole for antifungal prophylaxis in this high-risk population. Invasive aspergillosis (IA) was the most frequent mold infection and occurred predominantly among children with hematologic malignancies, especially AML and relapsed ALL. This distribution is consistent with previous studies [12, 19–22]. IA also occurred in children with non-malignant immunocompromising conditions and, occasionally, in previously healthy children with severe viral pneumonia, underscoring the heterogeneity of risk profiles and limitation of current risk stratification strategies in pediatric populations. A proportion of IA episodes were diagnosed while patients were receiving antifungal prophylaxis. Although these findings should not be interpreted as a failure of prophylaxis, many episodes occurred shortly after prophylaxis initiation or in the setting of subtherapeutic levels, highlighting challenges associated with itraconazole use in children [9, 10, 23]. Due to the lack of a standardized surveillance protocol, therapeutic drug monitoring (TDM) data were limited. However, among patients with breakthrough IA who underwent TDM, 57.1% (4/7) had subtherapeutic trough levels (itraconazole level < 0.5 mcg/mL). In two additional cases, clinicians preemptively switched prophylaxis to therapeutic voriconazole before TDM could be performed. Variable absorption, drug–drug interactions, and limited access to therapeutic drug monitoring may compromise effective mold prevention. These findings support consideration of antifungal agents with more predictable pharmacokinetics and underscore the importance of therapeutic drug monitoring when azoles are used for prophylaxis in high-risk pediatric patients [24]. Diagnostic challenges persisted. Serum galactomannan positivity was low, consistent with previous pediatric studies and likely reflecting prior antifungal exposure, and differences in immune response compared with adults [24–26]. Furthermore, our center’s practice utilized serum GM primarily as a broad, serial screening tool during prolonged fever and neutropenia, which naturally diluted its overall yield. In contrast, BAL galactomannan demonstrated high positivity, reinforcing its diagnostic utility when bronchoscopy is clinically feasible. At our center, BAL GM was highly selective and reserved strictly for patients with abnormal CT findings who were stable enough to undergo the procedure. It was frequently delayed or prohibited in the early phase of infection due to common clinical contraindications in high-risk children, such as hemodynamic instability and severe thrombocytopenia. Radiologic findings were broadly consistent with pediatric literature, although classical adult signs such as the air-crescent sign were uncommon, emphasizing the need for pediatric-specific interpretation [12, 24]. Mortality associated with IA remained high, particularly among patients with proven or disseminated disease. Elevated serum galactomannan was associated with increased mortality, likely reflecting greater disease severity. The occurrence of fatal IA in a previously healthy child highlights emerging at risk-populations and the need for heightened clinical awareness [27]. Invasive candidiasis primarily affected infants and young children with chronic conditions frequently associated with central venous catheter use, mirroring patterns reported in other pediatric and ICU populations [1, 28, 29]. Candida albicans remained the predominant pathogen; we observed a notable emergence of Candida tropicalis, which aligns with recent epidemiological trends across Asia [30]. Given the higher rates of azole resistance characteristic of C. tropicalis, this shift underscores critical challenges for empirical antifungal selection and highlights the urgent need for robust antifungal stewardship, particularly in settings with widespread prior azole exposure. Despite advances in antifungal therapy, mortality associated with invasive candidiasis remained substantial, with nearly one-third of patients dying within one year. Deaths directly attributable to candidiasis were uncommon, suggesting that underlying disease severity and comorbid substantially influenced outcomes. The absence of identifiable predictors of mortality likely reflects patient heterogeneity and limited statistical power. Pneumocystis infection accounted for a notable proportion of IFDs, highlighting critical gaps in current preventive strategies. Infections diagnosed during trimethoprim–sulfamethoxazole prophylaxis suggests issued with adherence, underdosing, or periods of intensified immunosuppression [31]. Crucially, none of the SLE patients who developed PCP have received prophylaxis, representing a major missed preventive opportunity. Given that mortality was highest among children with autoimmune disease, our funding underscores an urgent need to standardize and mandate TMP-SMX prophylaxis in pediatric SLE patients undergoing intensive immunosuppressive therapy to prevent this avoidable burden. Other invasive mold infections, cryptococcosis, and endemic mycoses were less common but contributed to significant clinical complexity, underscoring the broad spectrum of fungal pathogens encountered in pediatric tertiary-care settings. This study has several limitations. First, its retrospective, single-center design at a tertiary care center may limit the generalizability of the finding. Second, a small sample size for some pathogens and limited use of autopsies might have led to an underestimation of disease burden. Finally, fungal antigen testing (serum and BAL galactomannan) was not performed routinely across all suspected cases, and the inclusion of possible cases in the overall case may have further diluted the reported diagnostic yield. Clinical Implications Based on our findings, specific pediatric sub-populations demand targeted clinical strategies due to their disproportionately high risk for IFDs. (Table 2) For invasive aspergillosis, patients with AML, or relapsed ALL/AML are at highest risk, and clinicians should pursue earlier CT scan and bronchoscopy with BAL galactomannan testing to optimize early diagnosis. For invasive candidiasis, which primarily affects infants with chronic liver or gastrointestinal diseases and central venous catheters (CVCs), more aggressive line management and strict catheter stewardship are warranted. Lastly, for Pneumocystis pneumonia (PCP), children with autoimmune diseases particularly SLE – represent a critical at-risk group that suffered the highest mortality. Consequently, clinicians must urgently reassess current protocols to standardize and mandate trimethoprim-sulfamethoxazole prophylaxis in high-risk pediatric autoimmune patients undergoing intensive immunosuppression. Table 2Summary of risk profiles, clinical presentations, and 90-day mortality across major pediatric invasive fungal diseasesIFD typeHighest-risk groupsCommon presentationKey findings90-day mortalityClinical implicationInvasive aspergillosisAML, relapsed ALL, prolonged neutropeniaFever, pulmonary nodules/infiltratesBAL GM positivity higher than serum GM20.3%Consider early chest CT and bronchoscopy/BAL GM in high-risk hematology patientsInvasive candidiasisInfants, chronic GI/hepatobiliary disease, CVC useCandidemia/ sepsisCVC present in 88.9%; C. albicans most common24.8%Prompt catheter evaluation/removal and early antifungal therapy may improve outcomesPneumocystis pneumoniaAutoimmune disease, immunosuppressive therapyHypoxemic pneumoniaMany cases occurred without prophylaxis12.9%Reassess TMP-SMX prophylaxis strategies in high-risk patientsOther mold infectionsSevere immunocompromisePulmonary/disseminated infectionRare molds are identified sporadicallyvaryMaintain awareness of breakthrough mold infections despite prophylaxis Table 2Summary of risk profiles, clinical presentations, and 90-day mortality across major pediatric invasive fungal diseasesIFD typeHighest-risk groupsCommon presentationKey findings90-day mortalityClinical implicationInvasive aspergillosisAML, relapsed ALL, prolonged neutropeniaFever, pulmonary nodules/infiltratesBAL GM positivity higher than serum GM20.3%Consider early chest CT and bronchoscopy/BAL GM in high-risk hematology patientsInvasive candidiasisInfants, chronic GI/hepatobiliary disease, CVC useCandidemia/ sepsisCVC present in 88.9%; C. albicans most common24.8%Prompt catheter evaluation/removal and early antifungal therapy may improve outcomesPneumocystis pneumoniaAutoimmune disease, immunosuppressive therapyHypoxemic pneumoniaMany cases occurred without prophylaxis12.9%Reassess TMP-SMX prophylaxis strategies in high-risk patientsOther mold infectionsSevere immunocompromisePulmonary/disseminated infectionRare molds are identified sporadicallyvaryMaintain awareness of breakthrough mold infections despite prophylaxis Summary of risk profiles, clinical presentations, and 90-day mortality across major pediatric invasive fungal diseases Conclusion Pediatric IFDs remain a significant and heterogeneous clinical problem. The occurrence of infections during antifungal prophylaxis, shifting pathogen distribution, and persistently high mortality highlight the need for improved risk stratification, optimized prophylaxis strategies, early diagnostic pathways, and antifungal stewardship tailored to local epidemiology. Prospective, multicenter studies are needed to refine prevention and management strategies for high-risk pediatric populations. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1