Medical Mycology December 2003, 41, 521 /528
Epidemiology, presentation, management and outcome of candidemia in a tertiary care teaching hospital in the United Arab Emirates, 1995 2001
/
MICHAEL ELLIS*,%, ULLA HEDSTROM*, PAULINE JUMAA§ & ABDULBARI BENER$ Departments of *Medicine and $Community Medicine, Faculty of Medicine and Health Sciences, UAE University, %Tawam Hospital, and §Department of Microbiology, Tawam Hospital, United Arab Emirates
Sixty episodes of candidemia among hospitalized patients in the United Arab Emirates (0.77/1000 discharges) in 1995 2001 were identified through case retrieval. All patients had malignancy (65%) or serious non-malignant disease (35%). Candida albicans accounted for 45% of isolates. Non-C. albicans Candida species occurred more frequently than C. albicans in adults (67%), hematologicmalignancy patients (58%), and cases of breakthrough candidemia (83%) and were prevalent overall in 2000 2001 (67 73%). C. tropicalis was identified in 15% of cases, C. glabrata in 5%, C. parapsilosis in 5%, C. inconspicua in 2%, C. famata in 2% and C. lusitaniae in 1%. Delayed diagnosis or treatment was common, as was Karnofsky scale 5 40%, septic shock, and inadequate dosage or duration of antifungal drug therapy. Crude mortality was 50%, and mortality attributable to candidemia was 30%. Univariate analysis indicated patients were more likely to die (odds ratio for death [95% CI]) if they had been stationed in the intensive care unit (ICU) (4.76 [1.31 17.2]), had a Karnofsky scale 5 40% (38.76 [4.66 322.47]), or suffered septic shock (9.88 [2.9 33.65]). They were more likely to survive in cases with concomitant bacteremia (0.25 [0.07 0.91]), adequate antifungal dose (0.28 [0.08 0.94]), and removal of central lines (0.26 [0.07 0.95]). The high association of bacteremia with candidemia (70% of cases) is unusual. The apparent survival benefit experienced by patients who had bacteremia (odds ratio for survival on multivariate analysis 2.40 [0.28 20.17], P B 0.03) is novel. /
/
/
/
/
/
/
/
/
/
/
/
Keywords
/
/
candidemia, diagnosis, treatment, outcome
Introduction
Materials and methods
Candidemia is the fourth highest cause of bloodstream infection in hospitalized patients in Europe and North America [1]. Scant information has been reported from developing countries. This study describes patients with candidemia in a teaching, tertiary referral United Arab Emirates (UAE) hospital.
Patients hospitalized at Tawam Hospital, UAE, a tertiary referral center, between June 1995 and May 2001 (6 years) who developed candidemia were identified from the Microbiology laboratory records. Demographic, clinical, treatment and outcome data were obtained by retrospective chart review.
Definitions and terminology Received 24 September 2002; Accepted 14 April 2003 C o r re s p o n d e n c e : Fa x : 0 0 9 7 1 3 7 2 2 3 9 4 2 ;
[email protected]
– 2003 ISHAM
E-mail:
Candidemia 1 positive blood culture for Candida from a peripheral vein or central venous catheter. DOI: 10.1080/13693780310001645337
522
Ellis et al.
Results Neutropenia neutrophil count of B/500/mm3 (profound B/100; protracted /10 days).
Microbiology Specimen processing Blood samples were processed using the Vital (Biome´rieux, Marcy l’Etoile, France) blood culture analyzer. Bottles indicating positive were Gram stained and subcultured onto appropriate media. Where yeasts were observed the samples were subcultured onto Sabouraud agar. Identification The Candida species were identified by morphological and biochemical characteristics and examined for germ tube production. Biochemical tests were performed using the API 20C AUX yeast identification system (Biome´rieux). Performance status This was assessed at occurrence of candidemia using the Karnofsky scale (5/40 /moderate disability requiring total support).
Data collection, processing, statistical methods and analysis The data were collected, coded and processed using the Statistical Packages for the Social Sciences, Norusis (SPSS, Chicago, IL, USA) [2]. In assessing risk factors for death, patients who died were compared with patients who survived the episode of candidemia. Student’s t -test was used to ascertain the significance of differences between mean values of two continuous variables and the Mann/Whitney test for data with nonparametric distributions. Chi-squared analysis was performed to test for differences in proportions of categorical variables between two or more groups. In 2/2 tables the Fischer exact test (two-tailed) was used instead of Chi-squared when the sample size was small. Kaplan /Meier survival curves were constructed in patients with C. albicans and those with Candida species other than C. albicans. The differences between the two patient groups were tested by the log-rank test. Multivariate analysis by the Cox model of proportional hazard was used to incorporate all the explanatory variables [3]. Forward stepwise procedures and likelihood ratio tests were used to select independent variables with the greatest prognostic value. The level P 5/0.05 was used as the cut-off value for significance.
Sixty patient episodes among 58 patients (29 male; 29 female) were identified during 1995 /2001. This represented an overall incidence of 0.77 episodes per 1000 discharges, 4.51 among cancer patients and 0.33 in those with nonmalignant conditions. The annual incidence was constant during the study period. The median age for 31 adults was 42 (range 15 /87) years, and for 22 children was 3 years (range 8 months /13 years). Five patients were neonates, including four extremely premature infants born at 25, 25, 27 and 28 weeks gestation. Candidemia was nosocomial in 55 episodes. It occurred after the fourth day in 51 patients (median time of 17.5 days; range 1/540 days) and in four patients discharged from hospital 2 weeks previously. The principal admitting diagnoses were hematological malignancy and lymphoma in 27 episodes (45%), solid tumors in 12 (20%) and nonmalignant conditions in 21 (35%) (Table 1). The units in which the patients were hospitalized before candidemia were: adult intensive care unit (ICU) (16 admissions; 27%), adult hematology-oncology (14; 23%), adult medical (4; 7%), adult surgical (2; 3%), neonatal ICU (5, 8%), pediatric medical (6, 10%) and pediatric surgical (13, 22%).
Risk factors Fifty-seven patients had received antibiotics within 4 weeks of candidemia, an average of 3.3 courses per patient for a median duration of 16 (range 4 /31) days. 27 patients had been given ]/4 antibiotics. Twenty-three of 60 patient admissions (38%) developed neutropenic fever unresponsive to broad-spectrum antibiotics. The median duration of neutropenia prior to candidemia was 14 (range 2/42) days. In 17 patients, the neutropenia was profound. In eight it was profound and protracted. Bacteremia preceded candidemia within 4 weeks in 42 episodes (70%), and occurred within 3 days in 16. Enteric Gram-negative organisms accounted for 29 episodes. Polymicrobial bacteremia, excluding coagulase-negative staphylococci and Corynebacterium species, occurred in 13 episodes. Four of 16 patients (25%) with bacteremia in the preceding 3 days died, compared to 25 of 44 patients (57%) bacteremia-free in the same period (P /0.04).
Candida species C. albicans was documented in 27/60 (45%) isolates (Table 2). Non-C. albicans species identified were – 2003 ISHAM, Medical Mycology, 41, 521 /528
Candidemia in the UAE
Table 1
523
Risk factors in United Arab Emirates candidemia patients (number of admissions)
Primary diseases at admission
Adults
Children
Total
Hematological malignancy/lymphoma Acute myeloid leukemia Acute lymphocytic leukemia Chronic myelocytic leukemia Aplastic anemia Lymphoma Unspecified Total
8a,* 3d 1 1 2 2 17
3 4 / / 1 2 10
11 7 1 1 3 4 27
Solid tumors Breast carcinoma Bladder carcinoma Gastric carcinoma Bronchial carcinoma Osteosarcoma Wilm’s tumor Primitive neuroectodermal tumor Medulloblastoma Unspecified Total
1 1g 1b,i 1l 1 / / / 2b,j 7
/ / / / / 1 1f 1 2 5
1 1 1 1 1 1 1 1 4 12
4e,m 3c 1 1 1 1c / / / / / / / / / 11
4 3 1 1 1 1 1 1 1 1 1 1 2 1 1 21
Non-malignant diseases Severe prematurity Multiple congenital abnormalities Hirschsprung’s disease/short bowel Idiopathic malabsorption Coxsackie myocarditis Chronic lung disease Bowel necrosis/ischemia Multiple trauma after road traffic accident Thrombotic thrombocytopenic purpura Ulcerative colitis Hyperosmotic nonketotic diabetic coma Alcoholism Chronic renal failure Necrotizing fasciitis Unspecified Total Secondary risk factors Antibiotic use within 4 weeks Antibiotic-resistant neutropenic fever Profound neutropenia Corticosteroids Chemotherapy Central venous line Parenteral nutrition Indwelling urinary catheter Surgery within 4 weeks Immunosuppressant therapy Bacteremia within 4 weeks
/ / / / / / 1b 1 1k 1 1 1h 2* 1 1 10
571 232 173,4 33 27 545 20 10 22 6 396
a Cerebral thrombosis; bDiabetes mellitus; cEpilepsy; dCrohn’s disease; eThalassaemia trait; fDubowitz syndrome and dysmorphism; gSenile, nonambulatory; hHepatorenal syndrome; iRenal and cardiac failure; jNecrotizing fasciitis; kParaplegia; lBronchiectasis/renal failure/ tracheostomy; mMultiple congenital abnormalities. *Represents one patient with two admissions. 1, Median duration 16 (4 /31) days. Antibiotics were: Aminoglycosides (45 occasions), penicillins (41), cephalosporins (26) and carbapenems (22). Others were glycopeptides (19), quinolones (11), folate acid antagonists (7) and miscellaneous (26). 2, Median duration 14 (1 /42) days. 3, Median duration 6 (2 /21) days. 4, Eight patients had profound and prolonged neutropenia. 5, median duration 14 (1 /90) days. 6, Thirteen patients had bacteremia within 3 days (concomitant bacteremia).
– 2003 ISHAM, Medical Mycology, 41, 521 /528
524
Ellis et al.
Table 2
Candida species: percentages isolated in candidemia patients in the United Arab Emirates and in comparable studies worldwide
Candida albicans C. glabrata C. parapsilosis C. tropicalis C. krusei C. inconspicua C. famata C. lusitaniae Candida unspecified all Candida non-C. albicans, unspecified
USAa
Canadaa
Europeb
South Americaa
Tawam Hospital, UAE
54 22 15 6 1 0 0 0 2 /
70 12 7 5 2 0 0 0 4 /
53 12 21 6 1 1 2 0 4 /
45 9 19 20 2 0 0 0 6 /
45 5 5 15 0 5 1 1 / 23
Data from aRef. [9], bRef. [16].
mainly Candida tropicalis (9 isolates; 15%). C. krusei was not isolated. 16/28 isolates (57%) in children were C. albicans compared to 11/33 (33%) in adults (P / 0.06). Candida parapsilosis was exclusive to children aged 5/1 year. Among the five neonates, C. albicans was identified in three, C. parapsilosis in one and unspecified non-C. albicans in one. Eight of 14 patients (57%) with solid tumor had C. albicans, compared to 10/24 patients (42%) with hematologic malignancy and 9/23 patients (39%) with non-malignant diseases. The percentage of C. albicans isolations decreased from 60% prior to 1999 to 46% in 1999, and fell further to between 27 and 33% in 2000/2001. Twelve patients developed candidemia (C. albicans in 33%) whilst receiving antifungal prophylaxis; this is referred to as ‘breakthrough candidemia’. Three of nine patients (33%) receiving fluconazole had C. albicans and 6/9 (67%) had non-C. albicans fungemia (Candida glabrata 2/9, C. inconspicua 3/9 and unidentified nonC. albicans 1/9).
Clinical presentation Fever occurred in 57/60 episodes. Gastrointestinal symptoms were documented in 15, respiratory signs in 20 and neurological features in four. Focal embolic lesions were documented in the skin in five episodes, kidney in three and spleen in one. The Karnofsky performance scale was 5/40% in 41 (68%) of the patient episodes. Multiorgan failure occurred in 19. Severe sepsis was present in 18 and septic shock in 23 episodes [4]. The median time from onset of fever to first positive blood culture was 3 (range 1 /28) days, clinical diagnosis 4 (range 1 /29) days and antifungal drug treatment 6 (range 1/29) days.
Management An antifungal drug was used in 50/60 episodes. Fluconazole was used in 31 episodes, liposomal amphotericin B in 19, conventional amphotericin B in 15, itraconazole in five and 5-flucystosine in six. Monotherapy was given in 27 episodes. 23 episodes received two or more antifungal drugs. Approximately 50% of treatment episodes were judged to have featured an inadequate dose or duration of antifungal therapy [5] (Table 3). The central venous catheter was removed as an adjunctive therapeutic maneuver in 20/54 patients.
Outcome The median length of stay or time to death for the 60 episodes after onset of candidemia was 20 (range 1 / 206) days. The median time to defervescence was 3 (range 1/21) days in febrile patients who responded. Thirty of 60 episodes ended in death. Eighteen deaths were directly attributable to candidemia. Candidemia was contributory in nine and unrelated in three. The most frequent cause of death was multi-organ failure combined with septic shock (Table 4). Twenty-three of 27 deaths in which candidemia was judged causal or contributory occurred within 30 days of candidemia (median 6 [range 1 /143] days. Nine deaths occurred within 4 days and 15 within 7 days. At the time of death 17 patients were neutropenic; 12 were receiving colony-stimulating factors. Variables showing a significant association with outcome are detailed in Table 5. In univariate analysis the odds ratio for death (95% CI in parentheses) for patients developing candidemia on an ICU station was 4.76 (1.31/17.2), in patients with septic shock was 9.88 (2.9 /33.65) and in patients with Karnofsky scale 5/40 was 38.76 (4.66/322.47). Growth factor usage showed – 2003 ISHAM, Medical Mycology, 41, 521 /528
Candidemia in the UAE
Table 3
525
Dosages and durations of therapeutic antifungals administered to candidemia patients
Drug
No. patients
Dose (mg/kg/day)
Duration (days) [mean9/SD]
Adequate dosagea No. patients
Adequate durationb
Fluconazole Liposomal amphotericin B Conventional amphotericin B
28 17 12
4.89/2.3 2.69/1.5 0.99/0.3
12.89/8.3 9.39/6.4 12.39/9.4
8 (29%) 9 (53%) 9 (75%)
14 (50%) 8 (47%) 8 (67%)
a
Fluconazole /6mg/kg/day; liposomal amphotericin B /3mg/kg/day; conventional amphotericin B /0.8 mg/kg/day [5] with appropriate dosage adjustment for renal dysfunction; choice of antifungal appropriate for predicted susceptibility of isolate. (Average 46% episodes treated were with adequate dosages.) b ]/10 days. (Average 53% episodes treated were with adequate durations.) Reasons for no antifungal drug treatment in 10 episodes were: delayed/post-mortem diagnosis (five episodes), successful treatment by central venous line removal alone (two episodes), patient refusal (one episode), unknown (two episodes).
Table 4 Cause of death in candidemia patients
Multiorgan failure and septic shock Respiratory distress or pneumonia Uncertain Anaphylaxis to conventional amphotericin B Bacterial septic shock Single organ failure
Attributable to Candidemia
Associated with Candidemia
Other
Total
11 5 2 0 0 0
3 0 2 1 2a 1b
0 1c 0 0 1d 1e
14 6 4 1 3 2
TOTAL 30 Figures are number of patients. aUnspecified Gram negative (1); Streptococcus pneumoniae (1). bRespiratory failure. cDue to underlying bronchopulmonary dysplasia. dEnterobacter spp. eChronic renal failure.
Table 5
Univariate and multivariate analysis of factors for death in patients with candidemia
Factor
Present
P-valuea
Absent
Total Alive Dead Total Alive Dead Intensive care unit Karnofsky Performance score 5/40 Septic score 4 Concomitant bacteremia Adequate antifungal dose Growth factors Removal central venous line
16 41 23 16 23 12 17
4 13 4 12 15 3 13
12 28 19 4 8 9 4
44 18 37 44 24 42 37
27 17 26 19 8 25 16
17 1 11 25 16 17 21
Odds ratiob (95% Adjusted odds ratioc (95% confidence intervals) confidence intervals)
B/0.03 4.76 (1.31 /17.2) B/0.0001 38.76 (4.66 /322.47) B/0.0003 9.88 (2.9 /33.65) 0.05 0.25 (0.07 /0.91) 0.05 0.28 (0.08 /0.94) 0.07 4.2 (1.0 /17.5) B/0.05 0.26 (0.07 /0.95)
0.64 (0.08 /5.08) 22.4d (1.46 /342.4) 2.79 (0.37 /20.81) 0.41d,e (0.05 /3.49) 0.20f (0.02 /1.65) 2.26 (0.27 /18.76) 0.24g (0.03 /1.76)
a For difference between risk factor present or absent and death. bUnivariate analysis. cMultivariate analysis. dP B/0.03; eodds ratio for survival is 2.40 (0.28 /20.17). fodds ratio for survival is 4.8 (0.60 /38.34). godds ratio for survival is 4.04 (0.56 /28.82).
a trend towards a worse outcome. Odds ratios for death were low in patients who had an adequate antifungal dosage (0.28 [0.08/0.94]), preceding bacteremia (0.25 [0.07/0.91]) and removal of central venous catheter (0.26 [0.07 /0.95]). Notable factors that were not significantly associated with outcome included underlying disease, neutropenic status, non-C. albicans candidemia, breakthrough candidemia, class of antifungal drug and duration of antifungal therapy. On multivariate analysis poor performance scale was the – 2003 ISHAM, Medical Mycology, 41, 521 /528
only variable that retained a significant independent association with death, whilst the other variables lost their statistical significance for association but retained a trend. (Table 5). Multivariate analysis showed that the influence of the ICU station and of bacteremia on death was to favor survival. Overall episode mortality for patients with C. albicans was 10/26 (39%) and for non-C. albicans infections was 18/34 (53%). Patients with C. glabrata died in 1/3 episodes. With C. tropicalis 6/9 episodes
526
Ellis et al.
Fig. 1 Survival of candidemia patients infected with Candida albicans (squares) and non-C. albicans Candida species (solid line).
were fatal, with C. parapsilosis 3/3, with C. inconspicua 1/3, with C. famata 0/1, with C. lusitaniae 1/1 and with unidentified non-C. albicans species 5/14. Overall survival curves for patients with non-C. albicans species and C. albicans were not statistically significant (Fig. 1).
Discussion In this UAE tertiary referral center the candidemia incidence of 0.77 episodes per 1000 discharges is low compared to rates seen in the USA (3.8 /4.9) [6,7]. The constancy seen between 1995 and 2001 differs from the findings of other reports [6,7], but this may be due to a stabilizing effect mediated by fluconazole subsequent to its introduction into Tawam Hospital in 1995 [8]. Candidemia in Tawam Hospital does not occur in patients without serious underlying disease. The 14-fold prevalence increase among patients with hematologic malignancy and solid tumors confirms the association of candidemia with malignancy. Nevertheless one-third of all candidemias occurred in patients with serious nonmalignant diseases. The highest proportion of cases (27%) occurred in patients on the ICU, similar to the rate (21%) recorded in Europe [9], but lower than that seen in the Americas (50%) [10]. A much higher occurrence of candidemia was seen in patients in the
adult hematology service unit (23%) or who had malignancy (65%) than was seen in a European study (9%) [9]. This difference might be due to poor infection control in these patient groups in our facility, or to the large volume of patients with malignancy (14% of all discharges). The common feature linking our patients is immunocompromise, related either directly to the index disease (e.g. neutrophil dysfunction) or indirectly to treatment of the index disease (e.g. use of central venous devices). C. albicans occurred in 45% of cases overall, but varied with different risk groups and was inexplicably more prevalent in children of all ages (affecting 57%) than in adults (33% of whom were affected) (P /0.06). In 2001, C. albicans fell to 33% of cases, a rate that differs from the 47% seen in the Americas [10], other parts of the Middle East [11] and Europe [9], but that is broadly similar to the 37% rate seen in Brazilian tertiary care facilities [12]. The emergent proportional increase of non-C. albicans species may be a consequence of fluconazole use [13,14]. The appearance of C. tropicalis as the most frequent non-C. albicans species (15% of all isolates) and C. glabrata as a relatively uncommon but regularly occurring agent (5%) is a pattern that more closely resembles that seen in South America rather than that found in North America and Europe [9,15] (Table 2). This observation supports the concept that the frequency of various Candida species is partly determined by factors in the local institution. Nominally, more of the patients with C. albicans died than did those with non-C. albicans Candida , and there also appeared to be more early deaths among C. albicans patients; however, for both these factors, the difference was not statistically significant and may have been attributable to chance alone. Certain non-C. albicans species (e.g. C. tropicalis ) are more virulent than C. albicans, whilst others (such as C. parapsilosis) are less virulent. These trends, however, are not always consistent [16] and the relation of causal species and outcome is unclear. Breakthrough candidemia was predominantly caused by non-C. albicans species, several of which are expected to have low susceptibility to fluconazole. This suggests that fluconazole use selects these species. We did not see an increased mortality in the breakthrough candidemia group, in contrast to the situation reported in other studies [17]. The high rate of preceding bacteremia is of interest. Recent bacteremia is predictive for candidemia [18]. Possible explanations are antibiotic induced disturbance of normal flora, or that bacteremia is a surrogate – 2003 ISHAM, Medical Mycology, 41, 521 /528
Candidemia in the UAE
marker of severity of illness. The observed improved survival in patients with bacteremia is surprising. Antibiotics given at the time of candidemia might have been beneficial through a biological modifying effect [19] or perhaps the bacteremic patients received closer clinical monitoring. A striking presentation feature is the poor general condition of patients reflected by a Karnofsky score of 5/40 in 68% episodes and by septic shock in 33%. It is unlikely that the preceding bacteremia contributed to the high sepsis scores since all patients cleared the bacteremia with appropriate antibiotic treatment and the mortality of patients with bacteremia was lower than that of patients without bacteremia. Most clinical features were nonspecific and bacterial infection was suspected in all patients, leading to delay in accurate diagnosis and treatment. The 6-day therapeutic delay after onset of fever could have been a factor for death in several patients. Some received no antifungal therapy, as culture results were unknown until after death. Guidelines for antifungal dosage and duration of therapy in patients with candidemia are available [5], but in only 50% of treated cases in the present study was the therapy given in compliance with the guidelines. Short-duration antifungal therapy in some patients reflects advanced severity of illness: many died prior to before they could receive adequate treatment. One in three patients who died did so within 4 days of microbiologically confirmed diagnosis, suggesting overwhelming infection. In such circumstances an adequate initial antifungal dosages is of great importance, as is shown in the multivariate analysis by the reduced chance of death in patients given adequate dosages of fluconazole or amphotericin B. Removal of existing central venous catheters is strongly advocated for treating candidemia [5] to minimize dissemination to distal organs and to reduce mortality [20], as we have confirmed (Table 5). The 50% crude and 30% attributable mortality rates approximate the results of other studies [6], partially reflecting the poor prognosis associated with the underlying illnesses. However, univariate analysis did not detect an influence of the underlying diseases on attributable mortality. Rather, septic shock, poor performance status and ICU station significantly predicted for death. Retention of poor performance as the only significant independent predictor for death may reflect a lack of statistical power due to small numbers. However, the substantial reversal in multivariate analysis of the high odds-ratio for death seen in the univariate analysis in patients on ICU station is of interest. This suggests that although patients in the – 2003 ISHAM, Medical Mycology, 41, 521 /528
527
ICU are at risk for candidemia, the immediate, aggressive support that patients receive once they develop candidemia actually improves their outcome. This study suggests areas for possible improvement of diagnosis and management of patients with candidemia. Diagnostic delay could be shortened with heightened diagnostic acumen whilst outcome could be improved with: adequate dosage and earlier administration of antifungals; removal of central lines; more effective correction of risk factors, for example neutropenia; rationalization of antibiotic use; and use of intensive care.
References 1 Centers for Disease Control and Prevention. National nosocomial infection surveillance (NNIS) report, data summary from October 1986 /April 1996. Am J Infect Control 1996; 24: 380 /388. 2 Norusis MJ. SPSS/PC/ For Windows, base system and advanced statistics user’s guide. Windows version 11.0. Chicago: SPSS, 1998. 3 Cox DR. Regression models and life tables. J Roy Stat Assoc 1972; 29: 187 /120. 4 American College of Chest Physicians: Society of Critical Care Medicine Consensus Conference. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992; 20: 864 /875. 5 Rex JH, Walsh TJ, Sobel JD, et al . Practice guidelines for the treatment of candidiasis. Clin Infect Dis 2000; 30: 662 /678. 6 Beck-Sague CM, Jarvis WR. National Nosocomial Infections Surveillance System. Secular trends in the epidemiology of nosocomial fungal infections in the United States, 1980 /1990. J Infect Dis 1993; 167: 1247 /1251. 7 Jarvis RW, Martone WJ. Predominant pathogens in hospital infections. J Antimicrob Chemother 1992; 29: 19 /24. 8 Baran Jr J, Muckatira B, Khatib R. Candidemia before and during the fluconazole era: prevalence, type of species and approach to treatment in a tertiary care community hospital. Scand J Infect Dis 2001; 33: 137 /139. 9 Pfaller MA, Jones RN, Doern GV, et al . International surveillance of blood stream infections due to Candida species in the European SENTRY Program: species distribution and antifungal susceptibility including the investigational triazole and echinocandin agents. SENTRY Participant Group (Europe). Diagn Microbiol Infect Dis 1999; 35: 19 /25. 10 Pfaller MA, Jones RN, Doern GV, et al . International Surveillance of Bloodstream Infections due to Candida species: frequency of occurrence and antifungal susceptibilities of isolates collected in 1997 in the United States, Canada, and South America for the SENTRY Program. J Clin Microbiol 1998; 36: 1886 /1889. 11 Rennert G, Rennert HS, Pitlik S, et al . Epidemiology of candidemia / a nationwide survey in Israel. Infection 2000; 28: 26 /29. 12 Colombo AL. Epidemiology and treatment of hematogenous candidiasis: a Brazilian perspective. Braz J Infect Dis 2000; 4: 113 /118. 13 Abi-Said D, Anaissie E, Uzun O, et al . The epidemiology of hematogenous candidiasis caused by different Candida species. Clin Infect Dis 1997; 24: 1122 /1128.
528
Ellis et al.
14 Ellis ME, Qadri SMH, Spence D, et al . The effect of fluconazole as prophylaxis for neutropenic patients on the isolation of Candida species from surveillance cultures. J Antimicrob Chemother 1994; 33: 1223 /1228. 15 Pfaller MF, Jones RN, Doern GV, et al . Bloodstream infections due to Candida species: SENTRY antimicrobial surveillance Program in North America and Latin America, 1997 /1998. Antimicrob Agents Chemother 2000; 44: 747 /751. 16 Viscoli C, Girmenia C, Marinus A, et al . Candidemia in cancer patients: a prospective, multicenter surveillance study by the Invasive Fungal Infection Group (IFIG) of the European Organization for Research and Treatment of Cancer (EORTC). Clin Infect Dis 1999; 28: 1071 /1079.
17 Uzun O, Ascioglu S, Anaissie E, et al . Risk factors and predictors of outcome in patients with cancer and breakthrough candidemia. Clin Infect Dis 2001; 32: 1713 /1717. 18 Marr KA, Seidel K, White TC, et al . Candidemia in allogeneic and marrow transplant recipients: evolution of risk factors after the adoption of prophylactic fluconazole. J Infect Dis 2000; 181: 309 /316. 19 Labro MT. Cefodizime as a biological response modifier: a review of its in-vivo, ex-vivo and in-vitro immunomodulatory properties. J Antimicrob Chemother 1990; 26(Suppl. C): 37 /47. 20 Eppes SC, Troutman JL, Gutman LT. Outcome of treatment of candidemia in children whose central catheters were removed or retained. Pediatr Infect Dis 1989; 8: 99 /104.
– 2003 ISHAM, Medical Mycology, 41, 521 /528
