Secondary Logo

Journal Logo

Brief Reports

RISK FACTORS FOR MORTALITY IN CHILDREN WITH CANDIDEMIA

Zaoutis, Theoklis E. MD*‡§; Coffin, Susan E. MD, MPH*‡; Chu, Jaclyn H. MHS*; Heydon, Kateri MS*; Zhao, Huaqing MA; Greves, H Mollie MD; Walsh, Thomas J. MD

Author Information
The Pediatric Infectious Disease Journal: August 2005 - Volume 24 - Issue 8 - p 736-739
doi: 10.1097/01.inf.0000172938.76561.8e
  • Free

Abstract

Candidemia is a frequent and potentially life-threatening complication experienced by many chronically or critically ill children.1 Infections caused by Candida are the most common invasive fungal infections in hospitalized patients, and these organisms have been identified as the fourth most common cause of nosocomial bloodstream infection in the United States.2 Candidemia has also been associated with significant morbidity and mortality. Mortality rates in children with candidemia are from 19–26% to as high as 43–54% among infants.3–8 Candidemia is often associated with signs and symptoms of sepsis and fungal infections possess the second highest case fatality rate (13%) among all causes of sepsis in children.9

To date, most studies of risk factors for mortality in patients with candidemia have focused on adult populations. However, risk factors for mortality identified in adults may not be relevant in pediatric patients. For example, hospitalized children with candidemia often have unique underlying diagnoses and comorbidities that might result in a different set of predictors of mortality than in adults.6 This study was undertaken to determine the risk factors for mortality in a large cohort of hospitalized children with candidemia.

METHODS

This retrospective cohort study was conducted at a large tertiary care children's hospital. All study patients were identified through a systematic record review of the clinical microbiology laboratory. All patients who had an episode of candidemia (defined as a positive blood culture for any Candida species obtained from a peripheral vessel or vascular catheter) from January 1, 1998 through December 31, 2001 were eligible for inclusion in the study. If a patient experienced multiple episodes of infection during the study period, only the first episode was included in the analysis of risk factors for mortality, because the risk of death could change with subsequent episodes of candidemia.

Data on demographic and clinical characteristics, presence of comorbid conditions, antifungal treatment and infecting Candida species were collected on all patients. To assess a possible association between total placement duration and disposition of an intravascular catheter during infection and the risk of death, detailed information on exposure to central venous catheters (CVCs) during infection was collected. A catheter was considered replaced if the patient was still candidemic (positive blood cultures for Candida) on the day of new catheter placement. The presence of arterial catheters was also assessed. The primary outcome was defined as in-hospital death (from any cause) within 1 month (30 days) of the initial positive culture for Candida.

Univariate analyses were conducted to determine the association between potential risk factors and in-hospital death among children with candidemia. Categoric variables were compared by Fisher's exact test, and continuous variables were compared by the Wilcoxon rank sum test. All variables with a P value of <0.10 on univariate Cox analysis were considered for inclusion in the multivariable model, as well as other variables determined a priori to be key factors for predicting mortality among patients with candidemia (eg, cancer, neutropenia, prematurity, total parenteral nutrition, CVC use, antifungal treatment, transplantation). We assessed central venous catheter exposure during infection and antifungal use as time-dependent variables, because these exposures can change over the course of observation, respectively. A 2-tailed P value of <0.05 was considered statistically significant. Statistical calculations were performed with the computing packages Stata version 8.0 (Stata Corp., College Station, TX) and SAS version 9.1 (SAS Corp., Cary, NC).

RESULTS

We identified a total of 172 eligible patients during the study period. Eleven patients (6%) had 2 or more episodes of infection. Of the 172 patients, complete records were available for review for 168 (98%) of the patients.

The microbiology data for this cohort have been previously reported.10 In summary, Candida albicans accounted for 43% of the cases of candidemia followed by Candida parapsilosis (27%), Candida glabrata (10%), Candida tropicalis (10%), Candida krusei (3%) and Candida lusitaniae (2%).

Table 1 summarizes the patient demographics and the univariate analysis of risk factors for mortality. Overall 29 (17%) children died within 1 month of the first positive culture of Candida.

TABLE 1
TABLE 1:
Risk Factors for Mortality in Children With Candidemia

The types of antifungal therapy administered after the onset of candidemia are reported in Table 1. Twelve patients were not treated with any antifungal therapy after the first positive Candida culture was reported, for reasons including death before receipt of culture results, clinician determination of a catheter-related infection that resolved after a line was pulled or unexplained resolution of infection by subsequent negative culture without initiation of antifungal treatment.

The multivariable Cox proportional hazard model revealed that patients located in the pediatric intensive care unit (PICU) at the time of infection were at higher risk for death of when compared with patients located outside the PICU at the time of infection [hazard ratio (HR); 6.3, 95% confidence interval (CI), 1.6–24.3]. The overall 30-day mortality rate for patients in the PICU was 37% compared with 10% for patients not located in the PICU. The other independent predictor for death in children with candidemia was the presence of an arterial line (HR 2.4; 95% CI 1.1–5.8). Neutropenia approached but did not reach statistical significance (HR 2.4; 95% CI 0.09–6.2). In multivariate analysis, we found that time-dependent exposure to catheters and antifungal therapies during infection were not significantly associated with mortality.

DISCUSSION

In our cohort of 168 children with candidemia, we found that significant independent risk factors for death were location in the PICU and the presence of an arterial catheter at the time of infection. Both of these variables are likely strong markers for severity of illness.

Thirty-seven percent of children with candidemia in the PICU died compared with 10% in all other areas including the neonatal intensive care unit. No association was seen between the time with a central venous catheter while blood cultures remained positive for Candida or duration of antifungal use and death.

The mortality rate in children with candidemia has been reported to range between 17–24%.3–8 Few studies have reported PICU-specific rates of mortality rates. One study of PICU patients from Spain reported a crude mortality rate of 22%.7 The variability in rates may reflect different definitions of mortality (ie, 30-day, all-cause, in-hospital). Multiple studies have reported risk factors for mortality in adult patients with candidemia. However, the outcomes of children with candidemia have not been extensively studied. Pappas et al,6 in a prospective cohort study that included 144 children, identified neutropenia and endotracheal intubation as independent risk factors for mortality. In our study, patients with neutropenia had a higher HR for death, but this did not reach statistical significance. Because endotracheal intubation occurs only in an intensive care unit, our finding of an association between presence in the PICU and an increased risk of death is probably comparable with the findings of Pappas et al.6 Furthermore we found no association between duration of CVC use while patients were candidemic and increased mortality. However, our sample size might have not been sufficient to detect a small difference if one existed.

Amphotericin B was the drug of choice for the treatment of candidemia in children. This can reflect a more conservative approach to the management of children with candidemia as well as the lack of clinical trials of azoles and other newer antifungal agents in children. Although fluconazole was available during the study period, newer triazole antifungals such as voriconazole and echinocandins (caspofungin) were not available. These and other newer antifungal agents can change the treatment strategies used to treat candidemia in children in the future. Previous studies have provided evidence supporting the recommendation that all candidemic patients should receive antifungal therapy.6,11 Our analysis revealed that almost all children with candidemia who did not receive antifungal therapy survived. However, this finding is limited by the small number of patients in this group and should not be considered as evidence for not treating children with candidemia.

Several potential limitations exist in our study. We defined 30-day mortality as the outcome of interest. Although this defined end point is unambiguous, it does not differentiate according to the cause of death so that some deaths may not have been directly attributable to candidemia. Autopsies were not performed on most of diseased patients. In addition, efforts to designate outcomes as “attributable” to a specific infection are often very subjective and inconsistently applied.12 Therefore our identified associations do not necessarily imply biologic causality. However, our results may indicate a secondary effect of candidemia in worsening the health of critically ill pediatric patients that subsequently may increase the risk of death. Additionally we did not assess birth weight or antifungal resistance as predictors of mortality, which may have resulted in residual confounding. Finally our study was conducted in a large academic tertiary care medical children's hospital, and the results might thus not be generalizable to other institutions.

REFERENCES

1. Hostetter MK. New insights into candidal infections. Adv Pediatr. 1996;43:209–230.
2. Edmond MB, Wallace SE, McClish DK, Pfaller MA, Jones RN, Wenzel RP. Nosocomial bloodstream infections in United States hospitals: a three-year analysis. Clin Infect Dis. 1999;29:239–244.
3. 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.
4. Stamos JK, Rowley AH. Candidemia in a pediatric population. Clin Infect Dis. 1995;20:571–575.
5. MacDonald L, Baker C, Chenoweth C. Risk factors for candidemia in a children's hospital. Clin Infect Dis. 1998;26:642–645.
6. Pappas PG, Rex JH, Lee J, et al. A prospective observational study of candidemia: epidemiology, therapy, and influences on mortality in hospitalized adult and pediatric patients. Clin Infect Dis. 2003;37:634–643.
7. Rodriguez-Nunez A. Incidence and mortality of proven invasive Candida infections in pediatric intensive care patients. Infect Control Hosp Epidemiol. 2001;22:477–478.
8. Turner RB, Donowitz LG, Hendley JO. Consequences of candidemia for pediatric patients. Am J Dis Child. 1985;139:178–180.
9. Watson RS, Carcillo JA, Linde-Zwirble WT, Clermont G, Lidicker J, Angus DC. The epidemiology of severe sepsis in children in the United States. Am J Respir Crit Care Med. 2003;167:695–701.
10. Zaoutis TE, Greves HM, Lautenbach E, Bilker WB, Coffin SE. Risk factors for disseminated candidiasis in children with candidemia. Pediatr Infect Dis J. 2004;23:635–641.
11. Rex JH, Walsh TJ, Sobel JD, et al. Practice guidelines for the treatment of candidiasis: Infectious Diseases Society of America. Clin Infect Dis. 2000;30:662–678.
12. Cosgrove SE, Carmeli Y. The impact of antimicrobial resistance on health and economic outcomes. Clin Infect Dis. 2003;36:1433–1437.
Keywords:

pediatric; candidemia; fungal infections; invasive candidiasis; mortality

© 2005 Lippincott Williams & Wilkins, Inc.