Original article

Candidaemia in a London teaching hospital: analysis of 128 cases

over a 7-year period

Candidamie in einer Londoner Universitatsklinik: Analyse von 128

Fallen uber einen Zeitraum von sieben Jahren
S. Schelenz and W. R. Gransden
Division of Microbiology, Department of Infection, Guys, Kings and St. Thomas School of Medicine, London, UK

Summary In a retrospective analysis of 128 cases of Candida bloodstream infections in a London

teaching hospital between 1995 and 2001, the incidence of candidaemia increased
from 0.2/1000 admissions in 1995 to 0.5 and 0.4/1000 admissions in 2000 and
2001, respectively. Risk factors for candidaemia included the presence of intravascular
(IV) lines (88%), admission to intensive care (51%), parenteral nutrition (35%),
multiple antibiotics (74%), corticosteroid therapy (12%), cancer chemotherapy (11%),
renal transplantation (5%) and neutropenia (3%). The sources of infection were IV
lines (77%), the urinary tract (7%) and the gastrointestinal tract (7%). Serious infective
complications (endocarditis, endophthalmitis or brain abscess) were noted in 6% of
cases. The most frequently isolated species were Candida albicans (64%), C. glabrata
(20%), C. tropicalis (9%) and C. parapsilosis (5%). The overall fluconazole-resistance rate
of Candida spp. was 7% (MIC 64 mg l)1). All the C. albicans isolates were sensitive to
fluconazole (MIC 8 mg l)1) whereas 20% of non-C. albicans isolates (27% of
C. glabrata and 14% C. tropicalis) were resistant. The mortality rate (35%) was lower
than in other reports and may be due to the early recognition of candidaemia and the
prompt removal of IV lines together with the initiation of appropriate antifungal
therapy. Regular surveillance of local Candida species, resistance profiles and risk
factors is important in order to identify patients at risk and to develop empirical
treatment protocols to reduce the incidence and mortality of candidaemia.

Zusammenfassung Eine retrospektive Analyse von 128 Blutstrominfektionen in einem Londoner

Lehrkrankenhaus in der Zeit von 1995 bis 2001 ergab einen Anstieg von 0.2/1000
Einweisungen 1995 auf 0.5 und 0.4/1000 Einweisungen 2000 und 2001.
Candidamie-Risikofaktoren waren intravasale Zugange (88%), Intensivpflege (51%),
parenterale Ernahrung (35%), multiple Antibiotika (74%), Corticoid-Therapie
(12%), Krebs-Chemotherapie (11%), Nierentransplantation (5%) und Neutropenie
(3%). Infektionsquellen waren intravasale Zugange (77%), Harnwege (7%)
und Gastrointestinaltrakt (7%). Schwere Infektionskomplikationen (Endokarditis,
Endophthalmitis und Gehirnabszesse) traten in 6% auf. Die am haufigsten isolierten
Species waren Candida albicans (64%), C. glabrata (20%), C. tropicalis (9%) und
C. parapsilosis (5%). Die Fluconazolresistenzrate fur Candida spp. insgesamt lag bei 7%
(MIC 64 mg ml)1). Samtliche C. albicans-Isolate waren empfindlich fur Fluconazol

Correspondence: Dr Silke Schelenz, Department of Infection, 5th floor,

North Wing, St. Thomas Hospital, Lambeth Palace Road, London SE1 7EH,
UK. Tel.: +20 7 928 9292 extn 1493. Fax: +20 7 928 0730.
E-mail: silke.schelenz@kcl.ac.uk

Accepted for publication 11 November 2002

390  2003 Blackwell Publishing Ltd Mycoses, 46, 390396

 Candidaemia in a London teaching hospital

(MIC 8 mg ml)1), wahrend 20% der Nicht-albicans-Isolate resistent waren (27% von
C. glabrata und 14% von C. tropicalis. Die Mortalitat lag mit 35% niedriger als in
anderen Berichten, was an der fruhen Erkennung der Candidamie und der
konsequenten Beseitigung intravasaler Zugangen gleichzeitig mit dem Beginn einer
angemessenen antimykotischen Therapie liegen mag. Regelmaige Uberwachung der
lokalen Candida-Species, der Resistenzprofile und Risikofaktoren sind unverzichtbar, um
die Risikopatienten zu erkennen und empirische Behandlungsprotokolle zu erstellen,
damit die Candidamie-Inzidenz und -Mortalitat gesekt werden konnen.

Key words: Candida, Candida glabrata, candidaemia, epidemiology, fluconazole resistance.

Schlusselworter: Candida, Candida glabrata, Candidamie, Epidemiologie, Fluconazol-Resistenz.

deliver acute as well as tertiary referral services. The

Introduction
Candida species have emerged as a major cause of
nosocomial fungal infections contributing to morbidity
and mortality in hospitalized patients.1,2 Candidaemia is
now a common complication of many surgical and
medical therapies, often related to treatment with
broad-spectrum antibiotics,3,4 parenteral nutrition,3
immunosuppressive therapy for organ transplants and
malignancies5,6 and the increased use of intravascular
(IV) devices.7 In the USA Candida is the fourth most
prevalent isolate in blood cultures.8 There is a concern
that the incidence of serious Candida infections is also
increasing in Europe.9 Of additional concern is that
resistance to triazoles, such as fluconazole has emerged
in these organisms and there is now an increasing
prominence of species other than C. albicans.10,11
Regular surveillance of Candida species, resistance pro-
files and risk factors is important in order to establish
empirical treatment protocols and in judging the
potential impact of new antifungal agents.
We have analysed the current trend of Candida
bloodstream infections in a London teaching hospital
between 1995 and 2001. The incidence, underlying
risk factors, complications, management and outcome
of candidaemia are presented. In this study, we report
on recent species distribution and antifungal sensitivity
of Candida causing bloodstream infection.
Methods
specialties include paediatric, neonatal and adult inten-
sive care, general adult and paediatric medicine, cardiac
and general surgery, urology, renal medicine including a
renal transplant unit, haematology/oncology, obstetrics/
gynaecology, dermatology, neurology and ENT.
The clinical data were obtained from the bactaeremia
database of the Department of Infection described previ-
ously.12 Patients with positive blood cultures were
regularly seen by a medical microbiologist and if cultures
were judged clinically significant the relevant clinical
details were recorded on a computerized database.
Definitions
Candidaemia was defined as the isolation of any
pathogenic species of Candida from at least one blood
culture specimen.13 A source of candidaemia was
defined as proven if the same organism was cultured
from other relevant sites such as IV line tips, urine or
intra-abdominal wound specimens and if no other
source of infection was found. A presumed source was
based on clinical evidence without microbiological
proof. Endocarditis was diagnosed on the basis of the
Duke criteria.14
The diagnosis of endophthalmitis was based on
positive microscopy and culture of Candida from vitreous
fluid in conjunction with a positive fundoscopy. Neu-
tropenia was defined as a neutrophil count of
1 109 cells l)1.

Patients Microbiology

We carried out a retrospective study of candiaemias in Clinical isolates were detected using an automated
patients admitted to Guys and St. Thomas Hospital Trust continuous monitoring blood culture system (VITAL,
between 1 January 1995 and 31 December 2001. Both BioMerieux, Marcy lEtoile, France) which was used
teaching hospitals combinedly have 1200 beds and throughout the study. Blood cultures were incubated for

 2003 Blackwell Publishing Ltd Mycoses, 46, 390396 391

S. Schelenz & W. R. Gransden

up to 7 days and subcultured thereafter if macroscop- Patients 50 years and infants <1 year of age were
ically cloudy. If yeasts were observed on Gram staining, the most commonly affected age groups (54 and 18%,
blood culture supernatants were cultured directly onto a respectively; Table 1). The number of male and female
chromogenic agar for speciation (CHROM-agar; MAST patients was similar (66 and 62, respectively).
Diagnostics, Merseyside, UK). Fluconazole disc sensitiv-
ity was measured using yeast nitrogen-based agar. All
Risk factors, source of candidaemia and complications
isolates were sent to the Mycology Reference Laboratory
(Bristol, UK) for MIC sensitivity testing (NCCLS) and More than half the patients were nursed on an intensive
species confirmation. care unit (Table 2). The majority of children (78%) were
premature infants nursed on the neonatal intensive care
unit (ITU) (Table 1). Intravascular devices were in situ
Statistical analysis
in 88% of cases (Table 2) and 35% of patients received
The significance of differences between groups was parenteral nutrition through a central line. The major-
determined by the chi-square test. Values of P < 0.05 ity of patients were treated with multiple antibiotics
were considered significant. before development of candidaemia (Table 2). Many
were immunocompromised because of treatment with
corticosteroids (11.7%), chemotherapy (10.9%), renal
Results
transplantation (4.6%) or neutropenia (3.1%) (Table 2).
Twenty-four per cent of patients suffered from a
Demographics
malignancy.
The overall incidence of hospital-acquired candidaemia Intravascular lines were the most frequent proven or
was 0.3 cases per 1000 hospital admissions. Over the presumed cause (77%) of infection followed by urinary
7-year study period between 1995 and 2001, 128 tract infections and gastrointestinal disorders (Table 3).
patients had at least one episode of positive Candida
blood culture. Candida represented 1.9% of all significant
Table 1 Age, gender and mortality of patients with candidaemia.
blood culture isolates (annual mean of total significant
bacteraemias/fungaemias are 970). The mean annual Age (years) Total no. Male Female No. of deaths (%)
number of episodes of candidaemia was 18.3 (SD 5.1).
<1 23 12 11 11 (47.8)
Since 1995 there was an increase in the incidence of 19 4 3 1 0 (0)
candidaemia from 0.23 per 1000 hospital admissions in 1029 6 1 5 0 (0)
1995 to 0.43 per 1000 admissions in 2001 (Fig. 1). 3049 26 12 14 7 (26.9)
The majority of cases (98%) were hospital acquired with 5069 39 20 19 10 (25.6)
>69 30 18 12 17 (56.7)
only three patients presenting with community-
Total 128 66 62 45 (35.2)
acquired infection because of intravenous drug use
(IVDU).

Table 2 Predisposing condition/risk factors.

0.6
Condition No. of patients* (%)
No. of candidaemia episodes per 1000

0.5 Intravascular catheter 113 (88.3)

Multiple antibiotics 95 (74.2)
hospital admissions

0.4 Intensive care 66 (51.6)

Parenteral hyperalimentation 45 (35.2)
0.3 Recent surgery 41 (32)
Malignancy 31 (24.2)
0.2 Renal failure 22 (17.2)
Premature neonate 17 (13.3)
0.1 Steroids 15 (11.7)
Chemotherapy 14 (10.9)
0
Diabetes 10 (7.8)
95 96 97 98 99 2000 2001 Transplant (renal) 6 (4.6)
Year Neutropaenia 4 (3.1)
Intravenous drug user 3 (2.3)
Figure 1 Incidence of candidaemia episodes between 1995 and
2001 at Guys and St. Thomas Hospital. *n 128.

392  2003 Blackwell Publishing Ltd Mycoses, 46, 390396

 Candidaemia in a London teaching hospital

Table 3 Source of 128 episodes of candidaemia.
Source Proven Presumed
Intravascular line 56 (44) 42 (33)
Urinary tract 7 (5.5) 2 (1.6)
Gastrointestinal 3 (2.3) 6 (4.7)
Unknown source
The values in parentheses are percentages.
Three patients had an infected nephrostomy stent in situ
and nine patients had underlying gastrointestinal dis-
orders such as perforated appendix or necrotizing
gastroenteritis. In 6.3% no microbiological or clinical
cause could be found.
Serious infectious complications were noted in 6.3%
of all patients. Endocarditis occurred in 3.1% of
candidaemia cases and included three patients who
developed endocarditis as a complication of colonized IV
lines and one IVDU presented with prosthetic heart
valve endocarditis on admission. Three patients (2.3%)
had endophthalmitis and one patient (0.8%) developed
a brain abscess.
Fifty-four per cent of the patients were colonized with
Candida before the positive blood culture and 46% had
Candida isolated from more than one site including
oropharynx, urine, wounds, rectum or line tips (data
not shown).
Candida species and sensitivity
Candida albicans was the most frequently isolated species
(Table 4). Non-C. albicans species contributed to 36%
with C. glabrata being the most common isolate followed
by C. tropicalis, C. parapsilosis, C. krusei and C. lusitaneae
(Table 4). One patient had a dual infection with
C. albicans and a fluconazole-resistant C. glabrata in
the same blood culture. Compared with 1995 there has
been an increase in non-C. albicans isolates causing
candidaemia mainly because of a rise in C. glabrata
isolates, although this difference did not reach statistical
significance (data not shown).
Total
Candida glabrata was mainly seen in the age group 10
98 (77) to >69 years but not in children <10 years of age
9 (7)
(Fig. 2). In neonates, C. albicans and C. parapsilosis were
9 (7)
8 (6.3)
isolated in 74 and 26%, respectively, and represented
the only two species in this group of patients (Fig. 2).
One case of C. krusei occurred in an oncology patient on
chemotherapy and the one C. lusitaneae case was seen in
a 5-year-old child.
The overall fluconazole resistance rate was 7%
(MIC 64). All C. albicans isolates were sensitive to
fluconazole (MIC 8 mg l)1) whereas 20% of non-C.
albicans isolates were resistant (MIC 64 mg l)1) and
11% had a reduced susceptibility (intermediate; MIC,
1632 mg l)1) (Table 4). The most common fluconaz-
ole-resistant species was C. glabrata (26.9%) followed by
C. tropicalis (14.3%). None of the C. parapsilosis strains
were resistant but 9% of isolates showed dose-dependent
fluconazole susceptibility.
Management and outcome
In 78% (88/113) of patients, IV lines were removed
and over 71% (91/128) of patients received antifun-
gal therapy as part of their management. The
majority of patients (82.4%) received monotherapy
with amphotericin formulations (20/91), fluconazole
(53/91) or flucytosine (2/91). Combination treatment
was given in 17.6% (amphotericin plus flucytosine
15/91 or in one case fluconazole and flucytosine).
Over the last 2 years fluconazole superseded ampho-
tericin as the first-line drug of choice (data not
shown).
The overall mortality of bloodstream infection with
Candida spp. was 35.2% (Table 1) and was greatest in
the elderly (>69 years) and the very young (<1 year)
(56.7 and 47.8%, respectively; Table 1). No deaths
occurred in the age range of 129 years.

Table 4 Candida species recovered

from blood cultures, resistance to flucon- Fluconazole (%)
azole and mortality (n 129).
Candida species No. of isolates (%) S R I Mortality (%)

C. albicans 83 (64.3) 83 (100) 0 (0) 0 (0) 32 (38.6)

C. glabrata 26 (20) 16 (61.5) 7 (26.9) 3 (11.5) 8 (30.8)
C. tropicallis 11 (8.5) 5 (71.4) 1 (14.3) 1 (14.3) 3 (27.3)
C. parapsilosis 7 (5.4) 10 (90.9) 0 (0) 1 (9.1) 1 (14.3)
C. krusei 1 (0.8) 0 1 (100) 0 (0) 1 (100)
C. lustaneae 1 (0.8) 1 (100) 0 (0) 0 (0) 0 (0)

S, sensitive (MIC 8 mg l)1); I, intermediate or dose-dependent susceptibility (MIC,

1632 mg l)1); R, resistant (MIC 64 mg l)1).

 2003 Blackwell Publishing Ltd Mycoses, 46, 390396 393

S. Schelenz & W. R. Gransden
60
50
Percent of candidaemias
40
30
20
10
0 extremes of age and in our study 78% of all children
<1 19 1029 3049
Age group (years)
Figure 2 Effect of age on the distribution of C. albicans,
C. glabrata and C. parapsilosis.
Candida albicans was associated with the highest
mortality (38.6%) followed by C. glabrata (30.8%),
C. tropicalis (27.3%) and C. parapsilosis (14.3%) (Table 4).
One elderly oncology patient with C. krusei died whereas
the paediatric case with C. lusitaneae infection survived.
The removal of IV lines alone without additional
antifungal therapy was associated with significantly
higher survival rate (13/18) compared with those who
had no intervention (2/9) (P < 0.025). The additional
treatment with antifungal agents in conjunction with
line removal was not associated with significantly
greater survival (line removal plus antifungals 44/65
vs. line removal and no antibiotics 13/18; P < 1).
However, if for some reason the IV line could not be
removed, the treatment with antifungal agents alone
(8/12) was associated with greater survival compared
with those patients not treated (2/9) (P < 0.5).
Discussion
Several recent studies have indicated a steady increase
in the incidence of Candida bloodstream infections over
the last two decades.1,4,8,15 This increase is multifacto-
rial in origin and reflects the advances in medical and
surgical technology. The improved chemo- and immu-
nosuppressive therapy, transplant medicine and inten-
sive care technology has decreased the mortality of
many life threatening diseases but has also led to an
increase in patients vulnerable to infections. In the
United States, Candida spp. are now the fourth most
common cause of nosocomial bloodstream infection and
in many countries Candida represents 57.6% of all
blood culture isolates.16,17 Our study confirms this
upward trend in candidaemias with an increase in
incidence from 0.2/1000 hospital admissions in 1995
to 0.5 and 0.4/1000 admissions in the last two study
394
years, 2000 and 2001, respectively. However, in our
C. albicans study Candida represented only 1.9% of all significant
C. glabrata blood culture isolates and remains lower than those
C. parapsilosis
reported previously. Almost all (98%) of the candidae-
mias in our study are hospital acquired. The three
patients with community-acquired infections were
intravenous drug users.
There was a trend to the previous reported2,16 male
predominance although this was not statistically signi-
ficant. There is a higher incidence of candidaemia in the
5069 >69
were premature neonates. This is in concordance with
recent CDC data (Centers for Disease Control, Atlanta,
CA, USA), where 14% of candidaemias occurred in
012-month-old children.2
Risk factors for the development of candidaemia
include administration of parenteral hyperalimentation
and broad-spectrum antibiotics.16,18 Antibiotics such as
vancomycin and meropenem are known to affect the
protective bacterial gut flora and have been shown to be
risk factors for the development of candidaemia.19 Other
risk factors are those associated with an impaired
immune response including diabetes, malignancies, ster-
oids, chemotherapy, immunosuppressive drugs for the
management of organ transplants and neutropaenia.2,16
Those predisposing factors were similar to our series.
More than half (52%) of the patients in our study were
cared for on an ITU where they developed candidaemia.
The SCOPE study in 1999 also showed that 57% of all
candidaemia blood stream infections occurred in the
ITU.17 The majority of these patients have intravascular
devices in situ as part of their advanced life support which
is a major risk factor for the development of many line-
related infections.2, 16, 18 The urinary and gastrointesti-
nal tracts were the second most common sources of
infection. Only in 6.3% of cases we were not able to
determine the source of infection.
Candida albicans remains the most common Candida
species accounting on average for 64% of all candidae-
mia blood stream isolates in our hospital. This observa-
tion is higher than the 5258% reported from recent
surveillance studies performed by CDC,2 National Epi-
demiology of Mycoses Survey (NEMIS),11 SCOPE,17
SENTRY20 and EIEIO15 which have found a relative
increase in non-C. albicans species. Overall, the distri-
bution of Candida species other than C. albicans has not
significantly changed in our hospital over the 7-year
study period. However, we have observed a greater
percentage of C. glabrata species in 2000 and 2001 and
this species is the second most common Candida blood
culture isolate in our hospital (20%). This is similar to
the recent SENTRY surveillance data from the United
 2003 Blackwell Publishing Ltd Mycoses, 46, 390396

States (21%) but in contrast to data from other
European countries (10%), Canada (12%) and South
America (6%).20 C. glabrata did not occur in any
children younger than 10 years of age but was mainly
seen amongst 30 to >60-year-old patients. The increas-
ing importance of C. glabrata in the adult population is
well known and similar observation has been recently
reported by the EIEIO and SENTRY surveillance group
and others.15,20,21
In contrast to C. glabrata, the majority of C. parapsi-
losis candidaemias were found in children <1 year of
age. Although C. parapsilosis has been increasingly
associated with neonatal sepsis2,20,22,23 its prevalence
among neonates is not well understood and nosocomial
transmission in ITU nurseries has been suggested.11,24
The fluconazole resistance of all Candida spp. was
more than twice as high (7%) compared with 3%
reported from other surveillance study groups.15 As
much as 30.4% of all non-C. albicans species expressed
resistance or dose-dependent susceptibility to fluconaz-
ole whereas all C. albicans remain sensitive. Recent
surveillance data suggest that 712% of C. glabrata
isolates are resistant to fluconazole.15,20,25 In our study
we have seen an unusually high percentage (27%) of
fluconazole resistance with MICs >64 mg l)1 and a
further 11% of isolates demonstrated a dose-dependent
susceptibility. As C. glabrata expresses increasing resist-
ance to fluconazole and is now the second most
common species causing candidaemia in many coun-
tries it is important to speciate and to test sensitivities of
blood stream isolates in order to guide optimal treat-
ment.
Resistance was also reported in C. tropicalis which is
normally reported to be fully sensitive to fluconazole.15,20
The increase in fluconazole resistance of Candida blood
culture isolates in the UK and many other countries is of
great concern because of the increasing use of flucon-
azole as the first-line treatment for Candida infections.26
The mortality in patients with candidaemia is high,
ranging from 3479% as shown by studies including
patients from teaching and community hospital as well
as oncology centres.1,18 The overall crude mortality in
our study was relatively low (35%). This may be
because of an early recognition of the condition and the
prompt removal of the IV line together with the
initiation of appropriate antifungal therapy. The rate
for secondary infectious complications, such as endo-
carditis, endophthalmitis and brain abscess was 6.3%
but this might be an underestimation, as post-mortem
examinations were not always performed.
In conclusion, our study demonstrates that Candida is
still an important fungal pathogen with a rising
 2003 Blackwell Publishing Ltd Mycoses, 46, 390396
Candidaemia in a London teaching hospital
incidence and a high mortality. Resistance to fluconaz-
ole is increasing in non-C. albicans isolates in partic-
ularly C. glabrata which is now the second most
common cause of candidaemia in our hospital. Regular
surveillance of local Candida species and antifungal
susceptibility is important to establish effective treat-
ment protocols for the management of candidaemia.
A stringent antibiotic policy, early recognition of the
infection and the prompt removal of IV lines may help to
decrease the incidence and mortality from candidaemia.
Acknowledgement
We would like to thank the staff of the Mycology
Reference Laboratory, Bristol, UK for testing all our
Candida isolates.
References
1 Wey SB, Mori M, Pfaller MA, Woolson RF, Wenzel RP.
Hospital acuqired candidaemia: attributable mortality and
excess lenth of stay. Arch Intern Med 1988; 148: 26425.
2 Kao AS, Brandt ME, Pruitt WR et al. The epidemiology of
candidemia in two United States cities: results of a popu-
lation-based active surveillance. Clin Infect Dis 1999; 29:
116470.
3 Wey SB, Motomi M, Pfaller MA, Woolson RF, Wenzel RP.
Risk factors for hospital-acquired candidemia: matched
case-control study. Arch Intern Med 1989; 149: 234953.
4 Fraser VJ, Jones M, Dunkel J, Storfer S, Medoff G, Dunagan
WC. Candidaemia in a tertiary care hospital: epidemiol-
ogy, risk factors and predictors of mortality. Clin Infect Dis
1992; 15: 41421.
5 Maksymiuk AN, Thongprasert S, Hopfer R, Luna M,
Fainstein V, Bodey GP. Systemic candidiasis in cancer
patients. Am J Med 1984; 77(suppl 40): 207.
6 Meunier-Carpentier F, Kiehn TE, Armstrong D. Fungemia
in the immunocompromised host: changing patterns,
antigenemia, high mortality. Am J Med 1981; 71:
36370.
7 Lowder JN, Lazarus HM, Herzig RH. Bacteremias and
fungemias in oncology patients with central venous
catheters: changing spectrum of infection. Arch Intern Med
1982; 142: 14569.
8 Jarvis WR. Epidemiology of nosocomial fungal infections,
with emphasis on Candida species. Clin Infect Dis 1995; 2:
152630.
9 Vincent JL, Bihari DJ, Suter PM et al. The prevalence of
nosocomial infection in intensive care units in Europe
(EPIC). JAMA 1995; 274: 63944.
10 Nguyen MH, Peacock JE Jr, Morris AJ et al. The changing
face of candidemia: emergence of non-Candida albicans
species and antifungal resistance. Am J Med 1996; 6:
61723.
395
S. Schelenz & W. R. Gransden
11 Pfaller MA, Messer SA, Houston A et al. National Epi-
demiology of mycoses survey: a multicenter study of strain
variation and antifungal susceptibility among isolates of
Candida species. Diagn Microbiol Infect Dis 1998; 31:
28996.
12 Gransden WR, Eykyn SJ, Phillips J. Staphylococcus aureus
bacteraemia: 400 episodes in St. Thomass Hospital. BMJ
1984; 288: 3003.
13 Bodey GP, Anaissie EJ, Edwards JE. Definitions of Candida
Infections. In: Bodie GP (ed.), Pathogenesis, Diagnosis and
Treatment. New York: Raven Press Ltd., 1993; pp. 4078.
14 Durack DT, Lukes AS, Bright DK, the Duke Endocarditis
Service. New criteria for the diagnosis of infective endo-
carditis: utilization of specific echocardiographic findings.
Am J Med 1994; 96: 2009.
15 Diekema DJ, Messer SA, Brueggemann AB et al. Epidemi-
ology of candidemia: 3-year results from the emerging
infections and the epidemiology of Iowa organism study.
J Clin Microbiol 2002; 40: 1298302.
16 Al Soub H, Estinoso W. Hospital-acquired candidaemia:
experience from a developing country. J Hosp Infect 1997;
35: 1417.
17 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: 23944.
18 Stratov I, Gottlieb T, Bradbury R, OKane GM. Candidae-
mia in an Australian teaching hospital: relationship to
central line and TPN use. J Infect 1998; 36; 2037.
19 Richet HM, Andremont A, Tancrede C, Pico JL, Jarvis
WR. Risk factors for candidemia in patients with acute
396
lymphocytic leukaemia. Rev Infect Dis 1991; 156;
2115.
20 Pfaller MA, Diekema DJ, Jones RN et al. International
surveillance of bloodstream infections due to Candida
species: frequency of occurrence and in vitro susceptibility
to fluconazole, ravuconazole, and voriconazole of isolates
collected from 1997 through 1999 in SENTRY antimi-
crobial surveillance program. J Clin Microbiol 2001; 39:
32549.
21 Kauffman CA. Fungal infections in older adults. Clin Infect
Dis 2001; 33: 5505.
22 Saiman L, Ludington E, Pfaller S et al. Risk factors for
candidemia in neonatal intensive care unit. Pediatr Infect
Dis J 2000; 19: 31924.
23 Kossoff EH, Buescher ES, Karlowicz MG. Candidaemia in a
neonatal intensive care uniy: trends during fifteen years
and clinical features of 111 cases. Pediatr Infect Dis J 1998;
17: 5048.
24 Brown J, Froese-Fretz A, Luckey D, Todd JK. High rate off
hand contamination and low rate of hand washing before
infant contact in a neonatal intensive care unit. Pediatr
Infect Dis J 1996; 15: 90810.
25 Trick WE, Fridkin SK, Edwards JR et al. Secular trend of
hospital-acquired candidaemia among intensive care unit
patients in the United States during 19891999. Clin
Infect Dis 2002; 35: 62730.
26 Berrouane YF, Herwaldt LA, Pfaller MP. Trends in anti-
fungal use and epidemiology of nosocomial yeast infec-
tions in a university hospital. J Clin Microbiol 1999; 37:
5317.
 2003 Blackwell Publishing Ltd Mycoses, 46, 390396