Is the ERICAP Candida Score Useful for the Diagnosis of Candidemia in a Pediatric Critical Care Unit?

To the Editors: Early diagnosis of invasive Candida infections (ICI), candidemia particularly, is a challenge for pediatricians, intensivists and microbiologists.1–3 In 2014, a scoring system for per-species risk factors and predictors of ICI in patients admitted to a pediatric intensive care unit (PICU) was developed by the Incidence of invasive fungal disease and scale of risk of candidiasis in pediatric population hospitalized in intensive care units in spain (ERICAP) study, to help pediatricians to identify ICI in a species-specific manner species and facilitate adequate empiric treatment.4 Also, a new molecular method in the diagnosis of candidemia (T2MR) has been recently developed to detect the 5 most common Candida species (T2Candida Panel) directly from whole blood.3 Our aim was to compare the Candida score developed by the ERICAP study for candidemia in pediatric patients (0–17 years) admitted to PICU with candidemia suspicion at our institution and with a proven diagnosis by T2Candida assay from May 2017 to January 2022. Blood culture/T2Candida test results, demographic data and clinical features are shown in Table 1. Evaluation of the Candida score was performed by Receiver operating characteristic analysis to calculate sensitivity, specificity and threshold. Table 1. - Demographic and Clinical Characteristics of the Patients with T2Candida Panel and Simultaneous Blood Cultures Performed Characteristics All patients (N = 80) T2Candida positive patients (N = 26) T2Candida negative patients (N = 54) P value OR (CI 95%) Demographic Sex (% male) 29 (36.25) 14 (53.85) 15 (27.78) 0.043 2.98 (1.12–8.14) Median age and IQR (years) 4.5 (10.5) 5 (8.73) 3.5 (11) 0.138 NA Underlying disease Oncological disease 7 (8.75) 3 (11.53) 4 (74) 0.676 1.63 (0.28–8.45) Hematological disease 16 (20) 4 (15.38) 12 (22.2) 0.676 0.65 (0.16–2.16) Abdominal/digestive disease 20 (25) 5 (19.23) 15 (27.78) 0.581 0.63 (0.18–1.92) Cardiovascular disease 13 (16.25) 5 (19.23) 8 (14.81) 0.748 1.37 (0.36–4.74) Immunologic disorders 6 (7.5) 1 (3.84) 5 (9.26) 0.658 0.43 (0.02–3.05) Other 18 (22.5) 8 (30.77) 10 (18.51) 0.346 1.94 (0.63–5.82) Risk factors for candidemia Previous broad spectrum antibiotherapy 77 (96.25) 26 (100) 51 (94.4) 0.547 NA Previous antifungal therapy 68 (85) 24 (92.30) 44 (81.48) 0.319 2.56 (0.60–19.40) Central venous catheter 78 (97.5) 26 (100) 52 (96.29) 1 NA Endotracheal tubes 57 (71.25) 21 (80.77) 36 (66.67) 0.298 2.04 (0.69–7.10) Total parenteral nutrition 53 (66.25) 19 (73.08) 34 (62.96) 0.520 1.57 (0.57–4.70) Previous bacterial infection 45 (56.25) 20 (76.9) 25 (46.29) 0.019 3.74 (1.34–11.82) Hemodialysis 14 (17.5) 5 (19.23) 9 (16.67) 0.763 1.20 (0.32–4.01) Fever at the moment of admission 37 (46.25) 14 (53.85) 23 (42.59) 0.480 1.56 (0.60–4.09) Neutropenia 24 (30) 9 (34.62) 15 (27.78) 0.715 1.37 (0.49–3.78) Thrombocytopenia 54 (67.5) 18 (69.23) 36 (66.67) 1 1.11 (0.41–3.21) Previous abdominal surgery 37 (46.25) 14 (53.85) 23 (42.59) 0.48 1.11 (0.41–3.21) Chronic metabolic disease 6 (7.5) 5 (19.23) 1 (1.85) 0.012 11.05 (1.57–304.97) Urine sample with Candida species 12 (15) 3 (11.54) 9 (16.67) 0.742 0.67 (0.13–2.57) At least one clinical specimen with Candida isolate 29 (36.25) 12 (46.15) 17 (31.48) 0.303 1.85 (0.70–4.93) Median days of previous PICU stay (IQR) 7.5 (17.5) 18 (54.5) 6 (13.75) 0.037 NA Median number of Candida scores met (IQR) 2 (2) 3 (1) 2 (1) 0.004 NA CI indicates confidence interval; IQR, Interquartile range; OR, odds ratio; PICU, pediatric intensive care unit. Eighty cases with T2Candida test/blood cultures collected due to suspicion of candidemia from 73 pediatric patients were included in the study. A total of 26 patients had a positive T2Candida assay (17 Candida parapsilosis, 7 Candida albicans/tropicalis and 2 Candida krusei/glabrata). Three of them had a positive blood culture (2 Candida parapsilosis and 1 Candida tropicalis) taken at the same time, and in another one, the second blood culture taken was positive (C. parapsilosis). These patients had the 4 ERICAP Candida scores met. All the negative T2Candida results had negative blood cultures. All the patients positive by the T2Candida assay had risk factors for candidemia with different scores according to the ERICAP Candida score.4 Analysis of associations showed statistical significance with previous bacterial infection (P = 0.019), chronic metabolic disease (P = 0.012), median days of previous PICU stay (18 days, P = 0.037) and median number of Candida scores met (at least 3, P = 0.004) (Table 1). The comparison of the evaluation of the Candida score4 with the positive T2Candida assay results showed a threshold of 5.5, a sensitivity of 71% and a specificity of 60% for C. parapsilosis [AUC: 0.653, 95% confidence interval (CI): 0.492–0.814] and a threshold of 4.5, a sensitivity of 62% and a specificity of 67% for Candida sp. (AUC: 0.657, 95% CI: 0.528–0.787) as shown in Receiver operating characteristic curves (see Figures, Supplemental Digital Content 1, https://links.lww.com/INF/F12 and Supplemental Digital Content 2, https://links.lww.com/INF/F13). The use of the T2Candida panel in these patients could be optimized for the diagnosis after 2 weeks of PICU admission due to a significantly higher risk of candidemia after this time. Also, we observed that at least 3 Candida scores met might be used as clinical predictors of candidemia allowing early and adequate empiric or prophylactic antifungal treatment. However, we obtained different thresholds to those reported by Jordan et al.4 (4.5 vs. 5 for Candida sp. and 5.5 vs. 3 for C. parapsilosis). Our sensitivity and specificity values in both scores were rather low. This issue could have been due to the ERICAP Candida score was created and validated for invasive candidiasis in PICU patients,4 and we evaluated it only for candidemia. The ERICAP Candida score seems to be not optimal for diagnosing candidemia episodes. In conclusion, our data showed that the validated thresholds of the ERICAP Candida score are not useful in the specific context of candidemia in our media. However, if a patient admitted at PICU met at least 3 Candida scores, would be useful as a predictor of candidemia.