The incidence of invasive candidiasis (IC), the third most common cause of infection in intensive care units (ICUs) worldwide, accounts for 17% of infections and has been increasing throughout the world in recent years. Due to the high associated mortality and hospital costs,[2,3] IC is an increasing concern, especially for critically ill patients. Candida albicans have been the most commonly isolated strain in hospital patients for the past 20 years; however, the epidemiological trend of IC has changed towards increasing rates of infection with non-albicans strains in recent years. Particularly, Candida glabrata is now responsible for 15% to 20% of Candida infections in most countries, and the susceptibility of Candida to the azole antifungals is reduced.[5,6]
Knowledge of IC epidemiology, including geographical differences, is an important guide to prescribing practices and health policy and has far reaching clinical implications. In addition, the epidemiology of IC varies with the time of onset,[7,8] and candidaemia occurring more than 48 hours after admission is usually described as being healthcare-associated candidaemia (HCA).[9–11] There are few studies on the clinical characteristics and outcomes of early-onset IC (EOIC) with variable time definitions, ranging from 2 to 14 days after hospital admission.[12–14] In a cohort of critically ill patients with candidaemia, onset within 14 days of hospital admission, higher mortality, and hospital costs were closely associated with inadequate initial antifungal therapy. To date, there has been no scientific agreement on the definitions of EOIC and LOIC. Various times for EOIC definition in a few studies on the clinical characteristics and outcomes of EOIC have ranged from 2 to 14 days after hospital admission, with 1 study comparing the clinical characteristics with prognosis of candidaemia within or 10 days after admission, whereas most studies identified EOIC as IC onset within 48 hours of hospital admission.[12,15,16]
Because no clinical data for EOIC or LOIC in patients in Chinese ICUs have been reported so far, and the development of IC caused by opportunistic pathogens may be slow, we conducted an observational, multicentre study in 67 ICUs in China between November 2009 and April 2011 and compared the epidemiology, clinical characteristics, treatment, and risk factors in patients with EOIC (≤10 days) and LOIC (>10 days) after hospital admission in Chinese ICUs.
2.1 Study design and patients
The patients were from the China-SCAN study conducted from November 2009 to April 2011. In brief, 67 closed ICUs in general hospitals distributed throughout China participated in the observational multicentre study of proven IC. Patient inclusion criteria were age ≥18 years with clinical signs of infection and at least 1 of the following diagnostic criteria: histopathological, cytopathological, or direct microscopic confirmation of yeast cells in a specimen obtained by needle aspiration or biopsy from a normally sterile site; at least 1 peripheral blood culture positive for Candida; or positive Candida culture from a sample obtained by sterile technique from a normally sterile site. Samples from a drain, including urine and respiratory tract secretions that were not confirmed to be sterile were excluded. Patients were grouped according to the time of onset of the disease after hospital admission: ≤10 days for EOIC or >10 days for LOIC. APACHE II and Sequential Organ Failure Assessment (SOFA) scores were used to evaluate disease severity at the time of patient ICU admission and IC diagnosis. The time of onset refers to the date of diagnosis of IC, which was defined as the date when the first positive specimen was obtained. Chronic hepatic insufficiency was defined as biopsy-proven cirrhosis and portal hypertension, episodes of past upper gastrointestinal bleeding attributed to portal hypertension, or prior episodes of hepatic failure/encephalopathy/coma. Gastrointestinal dysfunction was defined as hemorrhage, food intolerance, perforation, surgery, acalculouschole cystitis, or intra-abdominal hypertension. Severe sepsis referred to the presence of known or suspected infection in association with systemic inflammatory response syndrome and organ dysfunction. Septic shock was defined as sepsis-induced hypotension that persisted despite adequate fluid resuscitation. Isolated Candida species were cultured and identified in vitro thereafter tested against amphotericin B, fluconazole, voriconazole, itraconazole, and caspofungin according to the CLSI M27-A3 microbroth dilution method at the Research Center for Medical Mycology, Peking University First Hospital, Beijing, China. The sensitivity cut-offs values were ≤8 mg/L for fluconazole, ≤1 mg/L for voriconazole, ≤2 mg/L for caspofungin, ≤0.125 mg/L for itraconazole, and ≤2 mg/L for amphotericin B. The epidemiologic data, pathogen-related, clinical characteristics, and therapeutic data were collected from all eligible patients treated in the participating ICUs. The study was approved by the Ethics Committee of Zhongda Hospital of Southeast University, the lead investigation site.
2.2 Statistical analysis
Continuous variables were described as mean ± standard deviation (SD) or median and range and compared with Student t tests if normally distributed or by Wilcoxon tests if non-normally distributed. Categorical variables were described as frequencies and percentages and compared with chi-squared tests or Fisher exact tests. Multivariate analysis with Cox proportional hazard model was used to identify independent risk factors for mortality among all variables with a P value <.05 on chi-squared test. P values <.05 were considered as statistically significant. All statistical analyses were performed using SAS 9.1 (SAS Institute INC, Cary NC).
3.1 Patient characteristics
Three hundred sixty-six patients were diagnosed with IC among a total of 96,060 ICU patients from 67 centers throughout China were included in this study. Out of the 306 patients, 290 (94.8%) were diagnosed solely based on at least 1 positive blood culture; 12 (3.9%) cases were diagnosed based on positive fluid culture from a normally sterile site (3 cases of CNS candidiasis, 8 cases of intra-abdominal candidiasis, and 1 case of candida pleuritis); and 3 were diagnosed based on candidaemia and positive culture from a normally sterile site. One pulmonary candidiasis was confirmed by histopathology.
Of these 306 patients, 105 cases were classified as EOIC and 201 cases as LOIC. The median (Q1, Q3) times between ICU admission and confirmed diagnosis of IC were 4 days (Q1–Q3, 1–7 days) and 17 days (Q1–Q3, 10–33 days) in the EOIC and LOIC groups, respectively. The main characteristics of the patients are detailed in Table 1.
The mean (SD) SOFA scores at diagnosis were 11.7 (±3.8) and 10.2 (±3.3; P < .001) in the EOIC and LOIC groups, respectively. The most common underlying diseases in the 2 groups were type 2 diabetes, chronic cardiac dysfunction, chronic obstructive pulmonary disease (COPD), and solid tumors. There were more patients with solid tumors in the LOIC group (P = .001) than in the EOIC group. Compared with patients in the LOIC group, more patients were treated with surgery in the EOIC group (46.7% vs 33.8%, P = .035). Other risk factors present in 40% to 60% of patients during the 2 weeks prior to study entry were gastrointestinal dysfunction and total parenteral nutrition in the 2 groups. More than 50% of patients in the EOIC and LOIC groups (59.0% vs 89.1%, P < .001) had a recent history of antibiotic therapy, and more than 70% of patients (76.2% vs 77.6%, P = .659) had invasive mechanical ventilation and central venous catheterization (76.2% vs 86.1%, P = .03; Table 1).
3.2 Microbiology data
Although the IC diagnosis was confirmed for all 306 patients in local laboratories, not all hospitals sent isolates to the central laboratory for confirmation (reasons included individual hospital policy and suboptimal storage or handling of isolates). Thus, the dataset was incomplete in this regard. A total of 387 isolates from 244 patients, 86 patients in the EOIC group, and 158 patients in the LOIC group were analyzed. C albicans was the most prevalent species in both EOIC and LOIC cases (47.7% and 36.1%, P = .101 in the EOIC group and LOIC group, respectively), while the proportion of different non-C albicans infections tended to be different between patients with EOIC and LOIC. Especially for Candida tropicalis and Candida parapsilosis infections, C parapsilosis was more prevalent in LOIC cases than in EOIC cases (10/86, 11.6% vs 45/158, 28.5%, P = .002), whereas C tropicalis was more common in EOIC cases (20/86, 23.3% vs 21/158, 13.3%, P = .05). There were no statistical differences among the distributions of other species within the 2 groups (Table 2).
3.3 Treatment and outcomes
The overall mortality rate of all the patients was 36.6% (112/306), and the mortality in the EOIC group was significantly lower than that in the LOIC group (30/105, 28.6% vs 82/201, 40.8%, P = .045). A total of 268/306 (87.6%) patients (89 in the EOIC group and 179 in the LOIC group), received antifungal therapy. The most commonly used antifungal agent was fluconazole, followed by the echinocandins. There was no significant difference between the EOIC and LOIC groups in the first-line antifungal agents used or in the susceptibility to the first-line agent (Table 3).
3.4 Risk factors associated with EOIC mortality
The results of univariate analysis of the factors associated with mortality among 105 patients with EOIC (30 nonsurvivors and 75 survivors) are presented in Table 4: patients with a higher SOFA score at IC diagnosis, severe sepsis, invasive mechanical ventilation, or gastrointestinal dysfunction had a higher mortality. However, multivariate analysis by COX model showed that only SOFA at ICI diagnosis tended to be associated with decreased survival (hazard ratio = 1.09, P = .0543; Table 5).
3.5 Risk factors associated with LOIC mortality
The results of univariate analysis for 201 patients with LOIC (82 nonsurvivors and 119 survivors) are presented in Table 6. Older patients and patients with severe sepsis, chronic renal insufficiency, and higher SOFA score had a higher risk of mortality. Susceptibility to the first-line agent was associated with a lower risk of mortality. Multivariate analysis by the Cox model showed that interval from ICU admission to ICI occurrence was a risk factor for mortality (hazard ratio = 1.009, P = .001; Table 7).
Investigations of IC according to the time of onset after hospital admission are rare. To our best knowledge, this is the first study of the epidemiology, treatment, outcomes, and risk factors for mortality in EOIC and LOIC patients in Chinese ICUs in which IC was diagnosed by the same strict criteria used in another study. In this study, we found that SOFA scores were higher at either ICU admission or IC diagnosis, and more cases of septic shock at IC diagnosis occurred in patients with EOIC than in those with LOIC, suggesting that the clinical conditions of patients with EOIC are more severe than those with LOIC both at ICU admission and at IC diagnosis.
C albicans was found to be the most prevalent Candida species in patients with both EOIC and LOIC; however, non-albicans species constituted 63.9% of isolates in patients with LOIC and 52.3% of isolates in patients with EOIC. These distributions of C albicans and non-albicans species were in accordance with those in recent reports both in China and other countries.[20–22] The reasons for the gradually increasing incidence of non-alibicans infections may be ascribed to the increased use of relatively low-cost azoles, an ageing population, the clinical condition of patients, and central venous catheter placement. The proportions of different non-albican species were different between patients with EOIC and LOIC. The reason C tropicalis was the most prevalent non-albicans species in EOIC may be that there were severe complications associated with hospitalization, more surgeries, and higher SOFA scores in this group, whereas the most common non-albicans isolate in LOIC was C parapsilosis, which may be caused by the longer duration of admission and intravenous catheter indwelling.
The crude mortality rate in the EOIC group was significantly lower than that in the LOIC group. This is in line with previous findings from a study in which mortality was significantly lower in patients with EOIC (≤10 days after admission) compared with patients with candidaemia diagnosed 10 days after admission. The explanation for this lower mortality in the EOIC group is likely related to less severe infection. However, the EOIC group had higher SOFA scores, at ICI, which showed a contradictory result. Furthermore, Cox hazard ratio analysis showed only SOFA score at ICI diagnosis in the EOIC group was a risk factor for mortality. The reason could be that despite the severity in the EOIC group, these patients were younger and had used less antibiotics previously. Finally, the interval from ICU admission to ICI occurrence was identified as a risk factor for mortality in the LOIC group, which was logical given that a prolonged ICU stay itself is often related to poor outcomes.
We found that the most commonly used antifungal drug was fluconazole followed by the echinocandinsin both the EOIC and LOIC groups. Based on the previous China-SCAN study, the resistance to itraconazole is reasonably high, and the sensitivity to fluconazole is relatively low among non-albicans species; similarly, resistance to fluconazole is widespread, especially among C glabrata and C parapsilosis infections in another study. Therefore, it is important to monitor the susceptibility and earlier use of effective non-azole antifungals in patients with probable or proven azole resistance, and an early and effective treatment of IC is critical for improved prognosis.
At least 3 limitations exist in the present study. First, not all isolates were sent to the central laboratory for our study population, so that the distribution of Candida species between EOIC and LOIC might not be precisely represented by our evaluation; second, differences in clinical practices across centers might influence the diagnosis and management of IC, and thus, the mortalities in the 2 different groups might not be precise. The average number of patients included from each center was 4.6 patients/center, and a mixed model might need a larger sample size. Third, the pathogenesis of IC most likely reflects a multistep process in which comorbidities, host factors, and colonization contribute to the invasion of the Candida spp. Thus, a further study with a larger sample size is needed to confirm the findings of our present study.
Nevertheless, significant differences were observed between patients with LOIC and EOIC in terms of mortality rates and non-albican Candidia species infection. SOFA score at ICI diagnosis was associated with higher mortality in EOIC, whereas a longer interval from ICU admission to ICI onset played a more important role for LOIC. Also, susceptibility to the first-line agent was a predictor of mortality in LOIC. Our findings highlight the need for earlier use of effective antifungals to reduce mortality.
The authors would like to thank the patients and investigators who participated in this study. They also acknowledge the investigators at each study site.
. Vincent JL, Rello J, Marshall J, et al. International study of the prevalence and outcomes of infection in intensive care units. JAMA 2009;302:2323–9.
. Falagas ME, Roussos N, Vardakas KZ. Relative frequency of albicans and the various non-albicans Candida spp among candidemia isolates from inpatients in various parts of the world: a systematic review. Int J Infect Dis 2010;14:e954–66.
. Olaechea PM, Palomar M, Leon-Gil C, et al. Economic impact of Candida colonization and Candida infection in the critically ill patient. Eur J Clin Microbiol Infect Dis 2004;23:323–30.
. Mean M, Marchetti O, Calandra T. Bench-to-bedside review: Candida infections in the intensive care unit
. Crit Care 2008;12:204.
. Pfaller MA, Diekema DJ. Epidemiology
of invasive candidiasis: a persistent public health problem. Clin Microbiol Rev 2007;20:133–63.
. Bassetti M, Righi E, Costa A, et al. Epidemiological trends in nosocomial candidemia in intensive care. BMC Infect Dis 2006;6:21.
. Kriengkauykiat J, Ito JI, Dadwal SS. Epidemiology
and treatment approaches in management of invasive fungal infections. Clin Epidemiol 2011;3:175–91.
. Concia E, Azzini AM, Conti M. Epidemiology
, incidence and risk factors
for invasive candidiasis in high-risk patients. Drugs 2009;69(Suppl):5–14.
. Hajjeh RA, Sofair AN, Harrison LH, et al. Incidence of bloodstream infections due to Candida species and in vitro susceptibilities of isolates collected from 1998 to 2000 in a population-based active surveillance program. J Clin Microbiol 2004;42:1519–27.
. Gulia J, Aryal S, Saadlla H, et al. Health care-associated candidemia—a distinct entity? J Hosp Med 2010;5:298–301.
. Friedman ND, Kaye KS, Stout JE, et al. Health care—associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med 2002;137:791–7.
. Chen S, Slavin M, Nguyen Q, et al. Active surveillance for candidemia, Australia. Emerg Infect Dis 2006;12:1508–16.
. Zilberberg MD, Kollef MH, Arnold H, et al. Inappropriate empiric antifungal therapy for candidemia in the ICU and hospital resource utilization: a retrospective cohort study. BMC Infect Dis 2010;10:150.
. De Rosa FG, Trecarichi EM, Montrucchio C, et al. Mortality in patients with early- or late-onset candidaemia. J Antimicrob Chemother 2013;68:927–35.
. Shorr AF, Gupta V, Sun X, et al. Burden of early-onset candidemia: analysis of culture-positive bloodstream infections from a large U.S. database. Crit Care Med 2009;37:2519–26. quiz 2535.
. Kung HC, Wang JL, Chang SC, et al. Community-onset candidemia at a university hospital, 1995-2005. J Microbiol Immunol Infect 2007;40:355–63.
. Guo F, Yang Y, Kang Y, et al. Invasive candidiasis in intensive care units in China: a multicentre prospective observational study. J Antimicrob Chemother 2013;68:1660–8.
. Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818–29.
. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on sepsis-related problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22:707–10.
. Azie N, Neofytos D, Pfaller M, et al. The PATH (prospective antifungal therapy) alliance (R) registry and invasive fungal infections: update 2012. Diagn Microbiol Infect Dis 2012;73:293–300.
. Montagna MT, Caggiano G, Lovero G, et al. Epidemiology
of invasive fungal infections in the intensive care unit
: results of a multicenter Italian survey (AURORA Project). Infection 2013;41:645–53.
. Wang H, Xiao M, Chen SC, et al. In vitro susceptibilities of yeast species to fluconazole and voriconazole as determined by the 2010 National China Hospital Invasive Fungal Surveillance Net (CHIF-NET) study. J Clin Microbiol 2012;50:3952–9.
. Liu W, Tan J, Sun J, et al. Invasive candidiasis in intensive care units in China: in vitro antifungal susceptibility in the China-SCAN study. J Antimicrob Chemother 2014;69:162–7.
. Leon C, Ruiz-Santana S, Saavedra P, et al. Usefulness of the “Candida score” for discriminating between Candida colonization and invasive candidiasis in non-neutropenic critically ill patients: a prospective multicenter study. Crit Care Med 2009;37:1624–33.
Keywords:Copyright © 2017 The Authors. Published by Wolters Kluwer Health, Inc. All rights reserved.
early-onset invasive candidiasis; epidemiology; intensive care unit; late-onset invasive candidiasis; risk factors