Original Studies
Risk Adjustment for Congenital Heart Surgery Score as a Risk Factor for Candidemia in Children Undergoing Congenital Heart Defect Surgery Fabio de Araujo Motta, MD, PhD, MSc,* Libera Maria Dalla-Costa, PharmD, PhD,*† Marisol Dominguez Muro, PharmD, PhD,*† Andrea Lenzi, MD, MSc,‡ Gledson Luiz Picharski,* and Marion Burger, MD, PhD§ Background: Candida species are the primary cause of invasive fungal infection in hospitalized children. There are few data on risk factors for postoperative candidemia in pediatric patients with congenital heart defects. This study aimed to identify risk factors for candidemia in patients with congenital heart defects who underwent cardiac surgery. Methods: This was a case-control study conducted in patients admitted to a pediatric cardiology intensive care unit from January 2006 to December 2013. Candidemia cases were matched with control patients without candidemia. Multivariate analyses were conducted to determine predictive probabilities for the incidence of candidemia at a risk higher than 10%. Results: Thirty patients diagnosed with candidemia (incidence: 0.7 cases/1000 patient days) were matched with 75 controls. Risk factors independently associated with candidemia included Risk Adjustment for Congenital Heart Surgery (RACHS-1) category ≥3 [odds ratio (OR) = 3.165, 95% confidence interval: 1.377–8.467], use of acid suppression therapy (OR = 1.9, 95% confidence interval: 0.949–3.979) and thrombocytopenia (OR = 2.2, 95% confidence interval: 1.2–4.2). Predictive probabilities ranged from 11% (only in RACHS-1 category ≥3) to 58% (combined RACHS-1 ≥3, thrombocytopenia and acid suppression therapy use). The case fatality rate within 30 days after candidemia was 50%. Conclusion: This is the first report using the RACHS-1 category as a risk factor for invasive candidiasis in patients with congenital heart defects in the pediatric intensive care unit. Further studies must be conducted to validate the risk factors for candidemia in this pediatric population.
lence of 44% in those 18-year-olds among all identified cases (672 episodes). The most prevalent species were C. albicans, C. parapsilosis and C. tropicalis. A low incidence of resistance to the major antifungal drugs used was also reported.2 Pediatric intensive care units (PICUs) are considered high-risk locations for death from candidemia, and the incidence may reach 3.5 cases/1000 patient days.3 This same risk occurs in patients with congenital heart disease (CHD) who undergo cardiac surgery; mortality from the underlying disease can reach 39%, and mortality from candidemia 10%.4 However, there are surprisingly few data about the risk factors for candidemia in postoperative pediatric patients with CHD. Knowledge of these risk factors will enable the development of more specific prevention strategies for candidemia and/or invasive candidiasis in this population. Antifungal prophylaxis has been an effective strategy to prevent candidemia in specific pediatric populations, including neonates and oncology patients.5,6 However, the potential benefit of prophylactic strategies in a large group of patients with a low incidence of events (candidemia) must be considered in relation to the potential risks of emergence of resistance or toxicity to the drug used.3 This study aimed to investigate risk factors and propose a predictive probability model for the incidence of candidemia in patients with CHD who underwent cardiac surgery in this hospital service.
Key Words: candidemia, congenital heart surgery, risk factor, RACHS
MATERIALS AND METHODS
(Pediatr Infect Dis J 2016;35:1194–1198)
C
andida species are the primary cause of invasive fungal infection in hospitalized children, and may be the fourth most common single agent to cause bloodstream infection associated with health care services in some Brazilian hospitals.1 An epidemiologic study collecting cases from Latin America recently highlighted the magnitude of candidemia in children in this region, with a preva-
Accepted for publication May 17, 2016. From the *Faculdades e Instituto de Pesquisa Pelé Pequeno Príncipe, Curitiba, Paraná, Brazil; †Bacteriolgy Laboratory, Hospital de Clínicas, Universidade Federal do Paraná Curitiba, Paraná, Brazil; ‡Cardiac Intensive Care Unit, Hospital Pequeno Príncipe, Curitiba, Paraná, Brazil; and §Secretaria Municipal da Saúde de Curitiba, Paraná, Brazil. F.A.M. has received a grant from Pfizer Inc. and Merck Sharp & Dohme Corp. related to clinical trial with antifungal drug and is on the speaker’s bureau and for Astellas Pharma S.A. M.B. is on the speaker’s bureau and has received support for travel to meetings from GlaxoSmithKline PLC and Sanofi S.A. The other authors have no funding or conflicts of interest to disclose. Address for correspondence: Fabio de Araujo Motta, MD, PhD, MSc, Rua Desemb. Motta, 1070 6.andar, 80.250-060 Bairro Água Verde, Curitiba, Paraná, Brazil. E-mail: [email protected]. Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/16/3511-1194 DOI: 10.1097/INF.0000000000001277
1194 | www.pidj.com
The Hospital Pequeno Príncipe exclusively provides pediatric service and the philanthropic private type; with about 390 beds, the hospital is located in Southern Brazil, in Curitiba, Paraná. It has a PICU exclusively for postoperative use in CHD and has approximately 550 admissions per year. A matched, case-control study was conducted in patients admitted to the cardiology PICU of Hospital Pequeno Príncipe from January 2006 to December 2013. The study was approved by the Hospital Pequeno Príncipe Research Ethics Committee in 2008, under the registration number 0591-08.
Case Definition All patients ≤ 18 years old were diagnosed with candidemia (presence of Candida in the blood) 72 or more hours after admission.7
Control Definition Matched patients admitted to the same PICU as case patients, in the same period (initial range of up to ±30 days), with the same age in days (range ±10%) and absence of Candida growth from a sterile site. For cases in which at least 2 controls were not available, a further time criterion was searched, allowing a further expansion of ±30 days. The matching goal was at least 2 and up to 4 controls for each case.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Exclusion Criteria Case patient: Candida growth in blood samples collected before completion of 72 hours of admission to the PICU. Control patient: empirically administered systemic antifungals, such as amphotericin B (in any formulation) or echinocandins, as therapy for suspected candidemia.
Data Collection An online data collection tool developed by the authors to retrieve data from the medical records of case and control patients was used. The data obtained included age, sex, length of hospital stay before infection and infection date. Data on the admission diagnosis and heart disease signs and symptoms, whether cyanotic or acyanotic, and the details of the surgery (with or without extracorporeal circulation) were also collected. All patients were classified according to the Risk Adjustment for Congenital Heart Surgery (RACHS-1) score. This is an easily applicable risk scoring method based on categorization into 6 levels (1–6) of several palliative or corrective surgical procedures, grouped by similar hospital mortality rates. Thus, heart diseases were distributed among 6 categories according to their respective expected mortality rates.8 In this study, RACHS-1 categories were rated according to the risk for candidemia. All data on antimicrobials used until the date of onset of candidemia were recorded for case patients. The duration of use (days) of antimicrobial drugs during hospitalization was considered for control patients. Comorbidity conditions at the time of admission in the study were considered as potential confounding factors. These included cancer, acute renal failure (including the need for peritoneal dialysis or hemodialysis and days of dialysis), liver disease, neurologic disease, total white blood cell count and thrombocytopenia (10% risk) were considered for the predictive model, according to Figure 1. Based on these results, the predictive probability of developing candidemia ranged from 11% to 58%. Among the 3 independent factors found, the patients with only RACHS-1 score ≥3 had an 11% risk of developing candidemia (95% CI: −2 to 24). Patients with combined RACHS-1 score ≥3, thrombocytopenia and AST use had a 58% risk of developing candidemia (95% CI: 38–78).
Demographic Characteristics and Clinical Features The demographic characteristics and clinical features are shown in Table 1. The average age of candidemia cases was 264 days (9 months) and the median was 126 days (4 months). The following variables were significant for candidemia in the univariate analysis: RACHS-1 score ≥3 (P = 0.01), use of AST (P = 0.01), thrombocytopenia (P = 0.01) and use of more than 5 antimicrobial drugs (P = 0.05), as shown in Table 1.
Multivariate Analysis The following factors remained independently associated with candidemia: RACHS-1 score ≥3 [odds ratio (OR) = 3.16, 95% CI: 1.38–8.47 and P = 0.01), thrombocytopenia (OR = 2.19, 95% CI: 1.16–4.22 and P = 0.02) and use of AST (OR = 1.89, 95% CI: 0.95–3.98 and P = 0.08).
Outcomes The mortality rate for these cases was 50%, compared with controls, at 51% (OR = 0.97). The average length of PICU stay for cases was 55 days, while that of controls was 52 days. The average number of days until the onset of candidemia was 33, with a median of 24. The average number of days postoperatively before onset of candidemia was 24 with a median of 17.
DISCUSSION To our knowledge, this is the third study published to date describing risk factors for candidemia in patients with CHD undergoing cardiac surgery. The first study refers to an outbreak of 6 cases of candidemia after CHD surgical correction.10 The second study, published in 2006, evaluated 75 patients between 1995 and 2000, using the Paediatric Risk of Mortality I score and the Therapeutic Intervention Scoring System. The study demonstrated that the Therapeutic Intervention Scoring System identified candidemia as an independent risk factor when applied 7 days after the admission of a patient to the PICU.11 www.pidj.com | 1195
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
de Araujo Motta et al
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
FIGURE 1. Predictive probability and 95% CI for candidemia in PICU with a combination of risk factors. AST indicates acid suppression therapy; Low plt., low platelets; Norm. Plt., normal platelets. However, a wide range of risk factors in CHD was identified, hampering consistent assessments, and these PICU scores are no longer used for patient profiles. In 2002, Jenkins et al8 proposed the use of a new, easily applicable risk score, termed RACHS-1. This score was based on the categorization of several palliative or corrective surgical approaches with similar hospital mortality rates. Accordingly, CHD has been divided into 6 categories, according to the expected mortality rate.8 RACHS-1 categories are applied worldwide in most PICUs for postoperative children with CHD, thus highlighting the importance of their use as risk factors for candidemia. The present study, in addition to reporting the largest series to date, with 105 patients enrolled among cases and controls (30/75), is the first to assess RACHS-1 category as a risk factor for candidemia. A RACHS-1 score ≥3 was also identified as an independent risk factor for candidemia (OR = 3.16, 95% CI: 1.38–8.47). Despite evaluation in several studies, the use of AST is not usually a factor commonly found in reports on risk factors for candidemia, both in adults and in children, as stated in some reviews.12,13 However, their use is recognized as a frequently associated factor.14 This study identified the use of AST as a risk factor for candidemia (OR = 1.89, 95% CI: 0.95–3.98) and is the first to describe this risk in PICU CHD patients. In 2000 and 2001, Saiman et al15,16 reported H2 blockers as a risk factor both for candidemia and Candida colonization, respectively, in neonates. Similar to AST, thrombocytopenia is an important associated comorbid condition in postoperative cardiac surgery patients.
1196 | www.pidj.com
This is due to the frequent use of some drugs that potentially cause thrombocytopenia, not because it is a direct cause of candidemia, but is instead a confounding factor. Jordan et al17 recently described thrombocytopenia as a risk factor for candidemia, specifically because of C. parapsilosis and C. tropicalis. Similarly, our study identified this condition as an independent risk factor for candidemia (OR = 2.189, 95% CI: 1.161–4.225), regardless of the species. Another important study conducted in a cohort of neonates by Benjamin et al18 also reported thrombocytopenia as a risk factor for candidemia Several studies investigated the relationship between the use of antibiotics and the occurrence of candidemia.19−21The present study separately evaluated all antimicrobials used in cases and controls, as well as the duration of use of each antimicrobial, and found no difference in incidence of candidemia between those who did or did not receive these drugs. Both case-control studies published before this investigation evaluated postoperative cardiac surgery patients and found that the significant use of antibiotics enhances the incidence of candidemia, which is also increased by the long-term use of these agents.10,11 However, it is noteworthy that different results for this analysis can be found, depending on the selection criteria adopted for the controls used in each study. Moreover, we evaluated the number of antimicrobial agents used, grouped as 1 to 2, 3 to 4 or 5 to 6. Patients who used 5 to 6 antimicrobial drugs showed a significant P value for incidence of candidemia only in univariate analysis (OR = 8.89). Multiple empiric antimicrobial therapies were observed due to the occurrence of © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Risk Factor for Candidemia in Children
TABLE 1. Univariate Analysis Comparing Demographic and Clinical Variables for Case Patients With Candidemia and Control Patients
Age (d) Female Cyanotic heart disease RACHS-1 category* 3–6 1 and 2 Liver disease Kidney disease Dialysis Neurologic disease CPB use MV TPN Use of corticosteroids Use of AST Thrombocytopenia Death Number of antimicrobial agents used 1–2 3–4 5–6 None Length of use of fourth-generation cephalosporins (d) 1–3 >3 0 Length of use of carbapenems (d) 1–3 >3 0 Length of use of piperacillin/tazobactam (d) 1–3 >3 0 Length of use of vancomycin (d) 1–3 >3 0 Length of use of linezolid (d) 1–3 >3 0
Case (%)
Control (%)
OR
P Value
217.60 12 (40.0) 18 (60.0)
299.80 35 (46.7) 40 (54.1)
— 0.76 1.27
0.58 0.66 0.66
25 (92.6) 2 (7.4) 0 (0) 5 (16.7) 4 (13.3) 1 (3.3) 21 (77.8) 30 (100.0) 13 (43.3) 13 (43.3) 26 (86.7) 25 (34.7) 15 (50.0)
41 (64.1) 23 (35.9) 5 (6.7) 7 (9.3) 8 (10.70) 2 (2.70) 51 (79.7) 69 (92.0) 39 (52.0) 23 (30.7) 43 (57.3) 17 (65.4) 38 (50.7)
7.14 0.14 — 1.96 1.28 1.27 0.89 — 0.7 1.72 4.8 3.55 —
0.01 Ref 0.32 0.32 0.74 1.00 1.00 0.18 0.52 0.26 0.01 0.01 0.97
33 (44.0) 23 (30.7) 9 (12.0) 10 (13.3)
9 (30.0) 12 (40.0) 8 (26.7) 1 (3.3)
2.73 5.22 8.89 Reference
0.67 0.14 0.05 Reference
3 (4) 20 (26.7) 52 (69.3)
3 (10) 12 (40) 15 (50)
3.47 2.08
0.155 0.148
2 (2.7) 32 (42.7) 41 (54.7)
1 (3.3) 12 (40) 17 (56.7)
1.21 0.9
1 1
4 (5.3) 8 (10.7) 63 (84)
1 (3.3) 7 (23.3) 22 (73.3)
0.72 2.51
1 0.126
8 (10.7) 26 (34.7) 41 (54.7)
5 (16.7) 14 (46.7) 11 (36.7)
2.33 2.01
0.279 0.161
2 (2.7) 8 (10.7) 65 (86.7)
2 (6.7) 3 (10) 25 (83.3)
2.6 0.98
0.576 1
*Three patients in the case group and 11 patients in the control group did not undergo surgery. CPB indicates cardiopulmonary bypass; TPN, total parenteral nutrition.
previous infection/sepsis in these patients. All results related to the use of antimicrobials indicate a high rate of use of these drugs in this unit, but significant differences between the 2 groups (case and control) were not found. In previous studies on risk factors in the PICU, the use of a CVC and MV were also important.3,22,23 In our investigation, neither was identified as a risk factor for candidemia. However, this study has some limitations related to the lack of accurate data in the medical records on the duration of MV and CVC use. This study has several other limitations, such as the lack of evaluation for previous colonization by Candida, described by many authors as an important association, because of the likelihood of colonization preceding infection.24−26 Another limitation is the small number of studied patients with candidemia, because it was not possible to conduct a multicenter PICU study. A third limitation is the fact that the diagnosis of candidemia was only made from a positive blood culture and did not identify how many patients had a diagnosis of infection related to a CVC through matching of central and peripheral cultures. Finally, another investigation that was not conducted in this study was the previous incidence of bloodstream © 2016 Wolters Kluwer Health, Inc. All rights reserved.
infection, which was also identified as a risk factor for hospitalacquired candidemia in neonatal patients.5,27,28 Briefly, the present findings enhance the little knowledge we have about risk factors for candidemia in pediatric populations with CHD who undergo cardiac surgery. According to our results, the use of AST and thrombocytopenia are independent risk factors for candidemia, as is RACHS-1 category ≥3. This study identified several combinations of predictors with a >10% risk of developing candidemia. Given these combinations, some patients could potentially benefit from the prophylactic use of antifungal agents. To date, only 1 study proposing predictive criteria for candidemia in PICU patients has been published.17 However, there is no predictive score for postoperative cardiac surgery PICU patients. A validated predictive model for this type of patient would be useful for intensive care physicians. In addition, preventive strategies could be proposed using such a validated predictive model. Further research focusing on the validation of risk factors identified in this study should be conducted to develop interventions to prevent candidemia in critically ill children. www.pidj.com | 1197
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
de Araujo Motta et al
REFERENCES 1. Colombo AL, Nucci M, Park BJ, et al; Brazilian Network Candidemia Study. Epidemiology of candidemia in Brazil: a nationwide sentinel surveillance of candidemia in eleven medical centers. J Clin Microbiol. 2006;44:2816–2823. 2. 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. 3. Zaoutis TE, Prasad PA, Localio AR, et al. Risk factors and predictors for candidemia in pediatric intensive care unit patients: implications for prevention. Clin Infect Dis. 2010;51:e38–e45. 4. San Miguel LG, Cobo J, Otheo E, et al. Candidemia in pediatric patients with congenital heart disease. Diagn Microbiol Infect Dis. 2006;55:203–207. 5. Manzoni P, Farina D, Leonessa M, et al. Risk factors for progression to invasive fungal infection in preterm neonates with fungal colonization. Pediatrics. 2006;118:2359–2364. 6. Robenshtok E, Gafter-Gvili A, Goldberg E, et al. Antifungal prophylaxis in cancer patients after chemotherapy or hematopoietic stem-cell transplantation: systematic review and meta-analysis. J Clin Oncol. 2007;25:5471–5489. 7. Hope WW, Castagnola E, Groll AH, et al; ESCMID Fungal Infection Study Group. ESCMID* guideline for the diagnosis and management of Candida diseases 2012: prevention and management of invasive infections in neonates and children caused by Candida spp. Clin Microbiol Infect. 2012;18(suppl 7):38–52. 8. Jenkins KJ, Gauvreau K, Newburger JW, et al. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg. 2002;123:110–118. 9. McCullagh P, Nelder JA. Generalized Linear Models. London: Chapman and Hall; 1989. 10. Chakrabarti C, Sood SK, Parnell V, et al. Prolonged candidemia in infants following surgery for congenital heart disease. Infect Control Hosp Epidemiol. 2003;24:753–757. 11. Garcia-San Miguel L, Cobo J, Martos I, et al. Risk factors for candidemia in pediatric patients with congenital heart disease. Infect Control Hosp Epidemiol. 2006;27:576–580. 12. Brissaud O, Guichoux J, Harambat J, et al. Invasive fungal disease in PICU: epidemiology and risk factors. Ann Intensive Care. 2012;2:6. 13. Yapar N. Epidemiology and risk factors for invasive candidiasis. Ther Clin Risk Manag. 2014;10:95–105. 14. França JC, Ribeiro CE, Queiroz-Telles Fd. [Candidemia in a Brazilian tertiary care hospital: incidence, frequency of different species, risk factors and antifungal susceptibility]. Rev Soc Bras Med Trop. 2008;41:23–28.
1198 | www.pidj.com
15. Saiman L, Ludington E, Pfaller M, et al. Risk factors for candidemia in neonatal intensive care unit patients. The National Epidemiology of Mycosis Survey study group. Pediatr Infect Dis J. 2000;19:319–324. 16. Saiman L, Ludington E, Dawson JD, et al; National Epidemiology of Mycoses Study Group. Risk factors for Candida species colonization of neonatal intensive care unit patients. Pediatr Infect Dis J. 2001;20:1119–1124. 17. Jordan I, Balaguer M, López-Castilla JD, et al; ERICAP Study Group. Perspecies risk factors and predictors of invasive Candida infections in patients admitted to pediatric intensive care units: development of ERICAP scoring systems. Pediatr Infect Dis J. 2014;33:e187–e193. 18. Benjamin DK Jr, DeLong ER, Steinbach WJ, et al. Empirical therapy for neonatal candidemia in very low birth weight infants. Pediatrics. 2003;112(3 pt 1):543–547. 19. Bross J, Talbot GH, Maislin G, et al. Risk factors for nosocomial candidemia: a case-control study in adults without leukemia. Am J Med. 1989;87:614–620. 20. Karabinis A, Hill C, Leclercq B, et al. Risk factors for candidemia in cancer patients: a case-control study. J Clin Microbiol. 1988;26:429–432. 21. Pittet D, Monod M, Suter PM, et al. Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg. 1994;220:751–758. 22. Ağın H, Devrim I, Işgüder R, et al. Risk factors for candidemia in pediatric intensive care unit patients. Indian J Pediatr. 2014;81:1158–1162. 23. MacDonald L, Baker C, Chenoweth C. Risk factors for candidemia in a children’s hospital. Clin Infect Dis. 1998;26:642–645. 24. Bonten MJ, Gaillard CA, Johanson WG Jr, et al. Colonization in patients receiving and not receiving topical antimicrobial prophylaxis. Am J Respir Crit Care Med. 1994;150(5 pt 1):1332–1340. 25. Martino P, Girmenia C, Micozzi A, et al. Prospective study of Candida colonization, use of empiric amphotericin B and development of invasive mycosis in neutropenic patients. Eur J Clin Microbiol Infect Dis. 1994;13:797–804. 26. Voss A, Hollis RJ, Pfaller MA, et al. Investigation of the sequence of colonization and candidemia in nonneutropenic patients. J Clin Microbiol. 1994;32:975–980. 27. Feja KN, Wu F, Roberts K, et al. Risk factors for candidemia in critically ill infants: a matched case-control study. J Pediatr. 2005;147:156–161. 28. Linder N, Levit O, Klinger G, et al. Risk factors associated with candidaemia in the neonatal intensive care unit: a case-control study. J Hosp Infect. 2004;57:321–324.
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Risk Adjustment for Congenital Heart Surgery Score as a Risk Factor for Candidemia in Children Undergoing Congenital Heart Defect Surgery Fabio de Araujo Motta, MD, PhD, MSc,* Libera Maria Dalla-Costa, PharmD, PhD,*† Marisol Dominguez Muro, PharmD, PhD,*† Andrea Lenzi, MD, MSc,‡ Gledson Luiz Picharski,* and Marion Burger, MD, PhD§ Background: Candida species are the primary cause of invasive fungal infection in hospitalized children. There are few data on risk factors for postoperative candidemia in pediatric patients with congenital heart defects. This study aimed to identify risk factors for candidemia in patients with congenital heart defects who underwent cardiac surgery. Methods: This was a case-control study conducted in patients admitted to a pediatric cardiology intensive care unit from January 2006 to December 2013. Candidemia cases were matched with control patients without candidemia. Multivariate analyses were conducted to determine predictive probabilities for the incidence of candidemia at a risk higher than 10%. Results: Thirty patients diagnosed with candidemia (incidence: 0.7 cases/1000 patient days) were matched with 75 controls. Risk factors independently associated with candidemia included Risk Adjustment for Congenital Heart Surgery (RACHS-1) category ≥3 [odds ratio (OR) = 3.165, 95% confidence interval: 1.377–8.467], use of acid suppression therapy (OR = 1.9, 95% confidence interval: 0.949–3.979) and thrombocytopenia (OR = 2.2, 95% confidence interval: 1.2–4.2). Predictive probabilities ranged from 11% (only in RACHS-1 category ≥3) to 58% (combined RACHS-1 ≥3, thrombocytopenia and acid suppression therapy use). The case fatality rate within 30 days after candidemia was 50%. Conclusion: This is the first report using the RACHS-1 category as a risk factor for invasive candidiasis in patients with congenital heart defects in the pediatric intensive care unit. Further studies must be conducted to validate the risk factors for candidemia in this pediatric population.
lence of 44% in those 18-year-olds among all identified cases (672 episodes). The most prevalent species were C. albicans, C. parapsilosis and C. tropicalis. A low incidence of resistance to the major antifungal drugs used was also reported.2 Pediatric intensive care units (PICUs) are considered high-risk locations for death from candidemia, and the incidence may reach 3.5 cases/1000 patient days.3 This same risk occurs in patients with congenital heart disease (CHD) who undergo cardiac surgery; mortality from the underlying disease can reach 39%, and mortality from candidemia 10%.4 However, there are surprisingly few data about the risk factors for candidemia in postoperative pediatric patients with CHD. Knowledge of these risk factors will enable the development of more specific prevention strategies for candidemia and/or invasive candidiasis in this population. Antifungal prophylaxis has been an effective strategy to prevent candidemia in specific pediatric populations, including neonates and oncology patients.5,6 However, the potential benefit of prophylactic strategies in a large group of patients with a low incidence of events (candidemia) must be considered in relation to the potential risks of emergence of resistance or toxicity to the drug used.3 This study aimed to investigate risk factors and propose a predictive probability model for the incidence of candidemia in patients with CHD who underwent cardiac surgery in this hospital service.
Key Words: candidemia, congenital heart surgery, risk factor, RACHS
MATERIALS AND METHODS
(Pediatr Infect Dis J 2016;35:1194–1198)
C
andida species are the primary cause of invasive fungal infection in hospitalized children, and may be the fourth most common single agent to cause bloodstream infection associated with health care services in some Brazilian hospitals.1 An epidemiologic study collecting cases from Latin America recently highlighted the magnitude of candidemia in children in this region, with a preva-
Accepted for publication May 17, 2016. From the *Faculdades e Instituto de Pesquisa Pelé Pequeno Príncipe, Curitiba, Paraná, Brazil; †Bacteriolgy Laboratory, Hospital de Clínicas, Universidade Federal do Paraná Curitiba, Paraná, Brazil; ‡Cardiac Intensive Care Unit, Hospital Pequeno Príncipe, Curitiba, Paraná, Brazil; and §Secretaria Municipal da Saúde de Curitiba, Paraná, Brazil. F.A.M. has received a grant from Pfizer Inc. and Merck Sharp & Dohme Corp. related to clinical trial with antifungal drug and is on the speaker’s bureau and for Astellas Pharma S.A. M.B. is on the speaker’s bureau and has received support for travel to meetings from GlaxoSmithKline PLC and Sanofi S.A. The other authors have no funding or conflicts of interest to disclose. Address for correspondence: Fabio de Araujo Motta, MD, PhD, MSc, Rua Desemb. Motta, 1070 6.andar, 80.250-060 Bairro Água Verde, Curitiba, Paraná, Brazil. E-mail: [email protected]. Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/16/3511-1194 DOI: 10.1097/INF.0000000000001277
1194 | www.pidj.com
The Hospital Pequeno Príncipe exclusively provides pediatric service and the philanthropic private type; with about 390 beds, the hospital is located in Southern Brazil, in Curitiba, Paraná. It has a PICU exclusively for postoperative use in CHD and has approximately 550 admissions per year. A matched, case-control study was conducted in patients admitted to the cardiology PICU of Hospital Pequeno Príncipe from January 2006 to December 2013. The study was approved by the Hospital Pequeno Príncipe Research Ethics Committee in 2008, under the registration number 0591-08.
Case Definition All patients ≤ 18 years old were diagnosed with candidemia (presence of Candida in the blood) 72 or more hours after admission.7
Control Definition Matched patients admitted to the same PICU as case patients, in the same period (initial range of up to ±30 days), with the same age in days (range ±10%) and absence of Candida growth from a sterile site. For cases in which at least 2 controls were not available, a further time criterion was searched, allowing a further expansion of ±30 days. The matching goal was at least 2 and up to 4 controls for each case.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Exclusion Criteria Case patient: Candida growth in blood samples collected before completion of 72 hours of admission to the PICU. Control patient: empirically administered systemic antifungals, such as amphotericin B (in any formulation) or echinocandins, as therapy for suspected candidemia.
Data Collection An online data collection tool developed by the authors to retrieve data from the medical records of case and control patients was used. The data obtained included age, sex, length of hospital stay before infection and infection date. Data on the admission diagnosis and heart disease signs and symptoms, whether cyanotic or acyanotic, and the details of the surgery (with or without extracorporeal circulation) were also collected. All patients were classified according to the Risk Adjustment for Congenital Heart Surgery (RACHS-1) score. This is an easily applicable risk scoring method based on categorization into 6 levels (1–6) of several palliative or corrective surgical procedures, grouped by similar hospital mortality rates. Thus, heart diseases were distributed among 6 categories according to their respective expected mortality rates.8 In this study, RACHS-1 categories were rated according to the risk for candidemia. All data on antimicrobials used until the date of onset of candidemia were recorded for case patients. The duration of use (days) of antimicrobial drugs during hospitalization was considered for control patients. Comorbidity conditions at the time of admission in the study were considered as potential confounding factors. These included cancer, acute renal failure (including the need for peritoneal dialysis or hemodialysis and days of dialysis), liver disease, neurologic disease, total white blood cell count and thrombocytopenia (10% risk) were considered for the predictive model, according to Figure 1. Based on these results, the predictive probability of developing candidemia ranged from 11% to 58%. Among the 3 independent factors found, the patients with only RACHS-1 score ≥3 had an 11% risk of developing candidemia (95% CI: −2 to 24). Patients with combined RACHS-1 score ≥3, thrombocytopenia and AST use had a 58% risk of developing candidemia (95% CI: 38–78).
Demographic Characteristics and Clinical Features The demographic characteristics and clinical features are shown in Table 1. The average age of candidemia cases was 264 days (9 months) and the median was 126 days (4 months). The following variables were significant for candidemia in the univariate analysis: RACHS-1 score ≥3 (P = 0.01), use of AST (P = 0.01), thrombocytopenia (P = 0.01) and use of more than 5 antimicrobial drugs (P = 0.05), as shown in Table 1.
Multivariate Analysis The following factors remained independently associated with candidemia: RACHS-1 score ≥3 [odds ratio (OR) = 3.16, 95% CI: 1.38–8.47 and P = 0.01), thrombocytopenia (OR = 2.19, 95% CI: 1.16–4.22 and P = 0.02) and use of AST (OR = 1.89, 95% CI: 0.95–3.98 and P = 0.08).
Outcomes The mortality rate for these cases was 50%, compared with controls, at 51% (OR = 0.97). The average length of PICU stay for cases was 55 days, while that of controls was 52 days. The average number of days until the onset of candidemia was 33, with a median of 24. The average number of days postoperatively before onset of candidemia was 24 with a median of 17.
DISCUSSION To our knowledge, this is the third study published to date describing risk factors for candidemia in patients with CHD undergoing cardiac surgery. The first study refers to an outbreak of 6 cases of candidemia after CHD surgical correction.10 The second study, published in 2006, evaluated 75 patients between 1995 and 2000, using the Paediatric Risk of Mortality I score and the Therapeutic Intervention Scoring System. The study demonstrated that the Therapeutic Intervention Scoring System identified candidemia as an independent risk factor when applied 7 days after the admission of a patient to the PICU.11 www.pidj.com | 1195
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
de Araujo Motta et al
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
FIGURE 1. Predictive probability and 95% CI for candidemia in PICU with a combination of risk factors. AST indicates acid suppression therapy; Low plt., low platelets; Norm. Plt., normal platelets. However, a wide range of risk factors in CHD was identified, hampering consistent assessments, and these PICU scores are no longer used for patient profiles. In 2002, Jenkins et al8 proposed the use of a new, easily applicable risk score, termed RACHS-1. This score was based on the categorization of several palliative or corrective surgical approaches with similar hospital mortality rates. Accordingly, CHD has been divided into 6 categories, according to the expected mortality rate.8 RACHS-1 categories are applied worldwide in most PICUs for postoperative children with CHD, thus highlighting the importance of their use as risk factors for candidemia. The present study, in addition to reporting the largest series to date, with 105 patients enrolled among cases and controls (30/75), is the first to assess RACHS-1 category as a risk factor for candidemia. A RACHS-1 score ≥3 was also identified as an independent risk factor for candidemia (OR = 3.16, 95% CI: 1.38–8.47). Despite evaluation in several studies, the use of AST is not usually a factor commonly found in reports on risk factors for candidemia, both in adults and in children, as stated in some reviews.12,13 However, their use is recognized as a frequently associated factor.14 This study identified the use of AST as a risk factor for candidemia (OR = 1.89, 95% CI: 0.95–3.98) and is the first to describe this risk in PICU CHD patients. In 2000 and 2001, Saiman et al15,16 reported H2 blockers as a risk factor both for candidemia and Candida colonization, respectively, in neonates. Similar to AST, thrombocytopenia is an important associated comorbid condition in postoperative cardiac surgery patients.
1196 | www.pidj.com
This is due to the frequent use of some drugs that potentially cause thrombocytopenia, not because it is a direct cause of candidemia, but is instead a confounding factor. Jordan et al17 recently described thrombocytopenia as a risk factor for candidemia, specifically because of C. parapsilosis and C. tropicalis. Similarly, our study identified this condition as an independent risk factor for candidemia (OR = 2.189, 95% CI: 1.161–4.225), regardless of the species. Another important study conducted in a cohort of neonates by Benjamin et al18 also reported thrombocytopenia as a risk factor for candidemia Several studies investigated the relationship between the use of antibiotics and the occurrence of candidemia.19−21The present study separately evaluated all antimicrobials used in cases and controls, as well as the duration of use of each antimicrobial, and found no difference in incidence of candidemia between those who did or did not receive these drugs. Both case-control studies published before this investigation evaluated postoperative cardiac surgery patients and found that the significant use of antibiotics enhances the incidence of candidemia, which is also increased by the long-term use of these agents.10,11 However, it is noteworthy that different results for this analysis can be found, depending on the selection criteria adopted for the controls used in each study. Moreover, we evaluated the number of antimicrobial agents used, grouped as 1 to 2, 3 to 4 or 5 to 6. Patients who used 5 to 6 antimicrobial drugs showed a significant P value for incidence of candidemia only in univariate analysis (OR = 8.89). Multiple empiric antimicrobial therapies were observed due to the occurrence of © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
Risk Factor for Candidemia in Children
TABLE 1. Univariate Analysis Comparing Demographic and Clinical Variables for Case Patients With Candidemia and Control Patients
Age (d) Female Cyanotic heart disease RACHS-1 category* 3–6 1 and 2 Liver disease Kidney disease Dialysis Neurologic disease CPB use MV TPN Use of corticosteroids Use of AST Thrombocytopenia Death Number of antimicrobial agents used 1–2 3–4 5–6 None Length of use of fourth-generation cephalosporins (d) 1–3 >3 0 Length of use of carbapenems (d) 1–3 >3 0 Length of use of piperacillin/tazobactam (d) 1–3 >3 0 Length of use of vancomycin (d) 1–3 >3 0 Length of use of linezolid (d) 1–3 >3 0
Case (%)
Control (%)
OR
P Value
217.60 12 (40.0) 18 (60.0)
299.80 35 (46.7) 40 (54.1)
— 0.76 1.27
0.58 0.66 0.66
25 (92.6) 2 (7.4) 0 (0) 5 (16.7) 4 (13.3) 1 (3.3) 21 (77.8) 30 (100.0) 13 (43.3) 13 (43.3) 26 (86.7) 25 (34.7) 15 (50.0)
41 (64.1) 23 (35.9) 5 (6.7) 7 (9.3) 8 (10.70) 2 (2.70) 51 (79.7) 69 (92.0) 39 (52.0) 23 (30.7) 43 (57.3) 17 (65.4) 38 (50.7)
7.14 0.14 — 1.96 1.28 1.27 0.89 — 0.7 1.72 4.8 3.55 —
0.01 Ref 0.32 0.32 0.74 1.00 1.00 0.18 0.52 0.26 0.01 0.01 0.97
33 (44.0) 23 (30.7) 9 (12.0) 10 (13.3)
9 (30.0) 12 (40.0) 8 (26.7) 1 (3.3)
2.73 5.22 8.89 Reference
0.67 0.14 0.05 Reference
3 (4) 20 (26.7) 52 (69.3)
3 (10) 12 (40) 15 (50)
3.47 2.08
0.155 0.148
2 (2.7) 32 (42.7) 41 (54.7)
1 (3.3) 12 (40) 17 (56.7)
1.21 0.9
1 1
4 (5.3) 8 (10.7) 63 (84)
1 (3.3) 7 (23.3) 22 (73.3)
0.72 2.51
1 0.126
8 (10.7) 26 (34.7) 41 (54.7)
5 (16.7) 14 (46.7) 11 (36.7)
2.33 2.01
0.279 0.161
2 (2.7) 8 (10.7) 65 (86.7)
2 (6.7) 3 (10) 25 (83.3)
2.6 0.98
0.576 1
*Three patients in the case group and 11 patients in the control group did not undergo surgery. CPB indicates cardiopulmonary bypass; TPN, total parenteral nutrition.
previous infection/sepsis in these patients. All results related to the use of antimicrobials indicate a high rate of use of these drugs in this unit, but significant differences between the 2 groups (case and control) were not found. In previous studies on risk factors in the PICU, the use of a CVC and MV were also important.3,22,23 In our investigation, neither was identified as a risk factor for candidemia. However, this study has some limitations related to the lack of accurate data in the medical records on the duration of MV and CVC use. This study has several other limitations, such as the lack of evaluation for previous colonization by Candida, described by many authors as an important association, because of the likelihood of colonization preceding infection.24−26 Another limitation is the small number of studied patients with candidemia, because it was not possible to conduct a multicenter PICU study. A third limitation is the fact that the diagnosis of candidemia was only made from a positive blood culture and did not identify how many patients had a diagnosis of infection related to a CVC through matching of central and peripheral cultures. Finally, another investigation that was not conducted in this study was the previous incidence of bloodstream © 2016 Wolters Kluwer Health, Inc. All rights reserved.
infection, which was also identified as a risk factor for hospitalacquired candidemia in neonatal patients.5,27,28 Briefly, the present findings enhance the little knowledge we have about risk factors for candidemia in pediatric populations with CHD who undergo cardiac surgery. According to our results, the use of AST and thrombocytopenia are independent risk factors for candidemia, as is RACHS-1 category ≥3. This study identified several combinations of predictors with a >10% risk of developing candidemia. Given these combinations, some patients could potentially benefit from the prophylactic use of antifungal agents. To date, only 1 study proposing predictive criteria for candidemia in PICU patients has been published.17 However, there is no predictive score for postoperative cardiac surgery PICU patients. A validated predictive model for this type of patient would be useful for intensive care physicians. In addition, preventive strategies could be proposed using such a validated predictive model. Further research focusing on the validation of risk factors identified in this study should be conducted to develop interventions to prevent candidemia in critically ill children. www.pidj.com | 1197
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 35, Number 11, November 2016
de Araujo Motta et al
REFERENCES 1. Colombo AL, Nucci M, Park BJ, et al; Brazilian Network Candidemia Study. Epidemiology of candidemia in Brazil: a nationwide sentinel surveillance of candidemia in eleven medical centers. J Clin Microbiol. 2006;44:2816–2823. 2. 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. 3. Zaoutis TE, Prasad PA, Localio AR, et al. Risk factors and predictors for candidemia in pediatric intensive care unit patients: implications for prevention. Clin Infect Dis. 2010;51:e38–e45. 4. San Miguel LG, Cobo J, Otheo E, et al. Candidemia in pediatric patients with congenital heart disease. Diagn Microbiol Infect Dis. 2006;55:203–207. 5. Manzoni P, Farina D, Leonessa M, et al. Risk factors for progression to invasive fungal infection in preterm neonates with fungal colonization. Pediatrics. 2006;118:2359–2364. 6. Robenshtok E, Gafter-Gvili A, Goldberg E, et al. Antifungal prophylaxis in cancer patients after chemotherapy or hematopoietic stem-cell transplantation: systematic review and meta-analysis. J Clin Oncol. 2007;25:5471–5489. 7. Hope WW, Castagnola E, Groll AH, et al; ESCMID Fungal Infection Study Group. ESCMID* guideline for the diagnosis and management of Candida diseases 2012: prevention and management of invasive infections in neonates and children caused by Candida spp. Clin Microbiol Infect. 2012;18(suppl 7):38–52. 8. Jenkins KJ, Gauvreau K, Newburger JW, et al. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg. 2002;123:110–118. 9. McCullagh P, Nelder JA. Generalized Linear Models. London: Chapman and Hall; 1989. 10. Chakrabarti C, Sood SK, Parnell V, et al. Prolonged candidemia in infants following surgery for congenital heart disease. Infect Control Hosp Epidemiol. 2003;24:753–757. 11. Garcia-San Miguel L, Cobo J, Martos I, et al. Risk factors for candidemia in pediatric patients with congenital heart disease. Infect Control Hosp Epidemiol. 2006;27:576–580. 12. Brissaud O, Guichoux J, Harambat J, et al. Invasive fungal disease in PICU: epidemiology and risk factors. Ann Intensive Care. 2012;2:6. 13. Yapar N. Epidemiology and risk factors for invasive candidiasis. Ther Clin Risk Manag. 2014;10:95–105. 14. França JC, Ribeiro CE, Queiroz-Telles Fd. [Candidemia in a Brazilian tertiary care hospital: incidence, frequency of different species, risk factors and antifungal susceptibility]. Rev Soc Bras Med Trop. 2008;41:23–28.
1198 | www.pidj.com
15. Saiman L, Ludington E, Pfaller M, et al. Risk factors for candidemia in neonatal intensive care unit patients. The National Epidemiology of Mycosis Survey study group. Pediatr Infect Dis J. 2000;19:319–324. 16. Saiman L, Ludington E, Dawson JD, et al; National Epidemiology of Mycoses Study Group. Risk factors for Candida species colonization of neonatal intensive care unit patients. Pediatr Infect Dis J. 2001;20:1119–1124. 17. Jordan I, Balaguer M, López-Castilla JD, et al; ERICAP Study Group. Perspecies risk factors and predictors of invasive Candida infections in patients admitted to pediatric intensive care units: development of ERICAP scoring systems. Pediatr Infect Dis J. 2014;33:e187–e193. 18. Benjamin DK Jr, DeLong ER, Steinbach WJ, et al. Empirical therapy for neonatal candidemia in very low birth weight infants. Pediatrics. 2003;112(3 pt 1):543–547. 19. Bross J, Talbot GH, Maislin G, et al. Risk factors for nosocomial candidemia: a case-control study in adults without leukemia. Am J Med. 1989;87:614–620. 20. Karabinis A, Hill C, Leclercq B, et al. Risk factors for candidemia in cancer patients: a case-control study. J Clin Microbiol. 1988;26:429–432. 21. Pittet D, Monod M, Suter PM, et al. Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg. 1994;220:751–758. 22. Ağın H, Devrim I, Işgüder R, et al. Risk factors for candidemia in pediatric intensive care unit patients. Indian J Pediatr. 2014;81:1158–1162. 23. MacDonald L, Baker C, Chenoweth C. Risk factors for candidemia in a children’s hospital. Clin Infect Dis. 1998;26:642–645. 24. Bonten MJ, Gaillard CA, Johanson WG Jr, et al. Colonization in patients receiving and not receiving topical antimicrobial prophylaxis. Am J Respir Crit Care Med. 1994;150(5 pt 1):1332–1340. 25. Martino P, Girmenia C, Micozzi A, et al. Prospective study of Candida colonization, use of empiric amphotericin B and development of invasive mycosis in neutropenic patients. Eur J Clin Microbiol Infect Dis. 1994;13:797–804. 26. Voss A, Hollis RJ, Pfaller MA, et al. Investigation of the sequence of colonization and candidemia in nonneutropenic patients. J Clin Microbiol. 1994;32:975–980. 27. Feja KN, Wu F, Roberts K, et al. Risk factors for candidemia in critically ill infants: a matched case-control study. J Pediatr. 2005;147:156–161. 28. Linder N, Levit O, Klinger G, et al. Risk factors associated with candidaemia in the neonatal intensive care unit: a case-control study. J Hosp Infect. 2004;57:321–324.
© 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
