TempleArtesian Poster

1
Risk Factors for Bloodstream Infections due to
Fluconazole-Resistant Candida glabrata A
Case-Case-Control Study Ingi Lee MD, MSCE Neil
O. Fishman MD Theoklis E. Zaoutis MD, MSCE
Knashawn H. Morales ScD Mark G. Weiner MD Marie
Synnestvedt PhD Irving Nachamkin DrPH, MPH
Ebbing Lautenbach, MD, MPH, MSCE University of
Pennsylvania School of Medicine, Philadelphia, PA
Ingi Lee, MD, MSCE 3400 Spruce Street, 3
Silverstein, Suite E Philadelphia, PA 19104
USA ingi.lee_at_uphs.upenn.edu
POTENTIAL LIMITATIONS
RESULTS
ABSTRACT
METHODS

• Case-case-control study at three hospitals in the
  University of Pennsylvania Health System.
• Each case group 1) all subjects with
  fluconazole-resistant Cg-BSI (MIC 16 µg/mL) and
  2) all subjects with fluconazole-susceptible
  Cg-BSI (MIC 8 µg/mL) compared to same control
  group, a random sample of hospitalized patients
  without Cg-BSI frequency matched to cases by
  quartiles based on time at risk.
• Primary risk factor of interest was prior
  fluconazole use.
• Multivariable analyses performed for each
  case-control study using multiple logistic
  regression.
• Final models compared qualitatively to identify
  unique risk factors for fluconazole-resistant and
  fluconazole-susceptible Cg-BSI.

Background Bloodstream infections (BSI) due to
Candida glabrata (Cg) have significantly
increased. Cg is often associated with
fluconazole resistance. However, risk factors for
healthcare-associated fluconazole-resistant
Cg-BSI have not been studied. Objective To
identify risk factors for healthcare-associated
fluconazole-resistant Cg-BSI. Methods A
case-case-control study was conducted at three
hospitals from January 2003 to May 2007. The case
groups were 1) patients with fluconazole-resistant
Cg-BSI minimum inhibitory concentration (MIC)
16 µg/mL and 2) patients with fluconazole-suscept
ible Cg-BSI MIC 8 µg/mL. Hospitalized
patients without Cg-BSI were randomly selected
for the control group, and frequency matched to
cases by quartiles based on time at risk. Two
case-control studies were performed using this
shared control group. The primary risk factor of
interest, prior fluconazole use, was evaluated in
multivariable analyses, adjusting for
demographics, co-morbidities, and antimicrobial
exposures. Results We included 76 patients with
fluconazole-resistant Cg-BSI, 68 patients with
fluconazole-susceptible Cg-BSI, and 512 control
patients. Prior fluconazole use and linezolid use
were independent risk factors for
fluconazole-resistant Cg-BSI prior cefepime use
and metronidazole use were independent risk
factors for fluconazole-susceptible
Cg-BSI. Conclusions Prior fluconazole use and
linezolid use were significant risk factors for
healthcare-associated fluconazole-resistant
Cg-BSI. Future studies will be needed to evaluate
the impact of decreasing fluconazole use on the
rates of fluconazole-resistant Cg-BSI.

• Fluconazole-resistant Cg-BSI (n76),
  fluconazole-susceptible Cg-BSI (n68), and
  controls (n512) included.
• Prior fluconazole use and linezolid use were
  independent risk factors for fluconazole-resistant
  Cg-BSI.
• Prior cefepime use and metronidazole use were
  independent risk factors for fluconazole-susceptib
  le Cg-BSI.
• No effect modification by hospital, year of
  hospitalization, oncology status, or receipt of
  chemotherapy.

• Potential misclassification of risk factors.
• However, non-differential misclassification would
  bias results towards the null.
• Hospitals all located in Philadelphia which may
  limit generalizability.

CONCLUSIONS

• Prior fluconazole use is an independent risk
  factor for fluconazole-resistant Cg-BSI.
• Resistance could occur 1) de novo by one or more
  mechanisms, including upregulating efflux pumps
  or 2) by changing endogenous flora, allowing
  colonization and subsequent infection with
  fluconazole-resistant Cg.
• Consistent with studies in oncology patients
  where high rates of fluconazole use may lead to
  high rates of fluconazole-resistant Cg-BSI.
• Prior studies of the general inpatient
  population did not find this association perhaps
  due to case group selection.
• Future studies are needed to identify the impact
  of decreasing fluconazole use on the rates of
  fluconazole-resistant Cg-BSI.

Table 2. Unadjusted Risk Factors for
Fluconazole-Susceptible Cg-BSI (Bivariable
analysis) Variable Cases Controls OR (95CI) P
value1 (n68) (n512) TPN 4 (6) 6
(1) 5.3 (1.1, 22.8) 0.02 Immunosuppression21
(31) 98 (19) 1.9 (1.0, 3.4) 0.02 Fluconazole 1
6 (24) 67 (13) 2.0 (1.0, 3.9)
0.02 Cefepime 34 (50) 125 (24) 3.1 (1.8,
5.4) lt0.001 Metronidazole 39 (57) 164 (32) 2.9
(1.7, 5.0) lt0.001 Vancomycin 39 (57) 199
(39) 2.1 (1.2, 3.7) 0.004 Only variables with p
values lt 0.05 are shown. TPN total parenteral
nutrition. 1Chi-square or Fishers exact test
(categorical) Students t test or Wilcoxon Rank
Sum test (continuous).
Table 1. Unadjusted Risk Factors for
Fluconazole-Resistant Cg-BSI (Bivariable
analysis) Variable Cases Controls OR (95CI) P
value1 (n76) (n512) Hospital (HUP) 34
(45) 303 (59) 0.6 (0.3, 0.9) 0.02 Time at
risk2,3 23 (12-44) 17 (9-27) ---4 0.003 TPN 5
(7) 6 (1) 5.9 (1.4, 23.9) 0.008 Immunosuppressio
n24 (32) 98 (19) 1.9 (1.1, 3.4) 0.01 Fluconazol
e 31 (41) 67 (13) 4.6 (2.6, 8.0) lt0.001 Amphote
ricin 5 (7) 11 (2) 3.2 (0.8, 10.4) 0.04 SAM 15
(20) 47 (9) 2.4 (1.2, 4.7) 0.005 TZP 14
(18) 52 (10) 2.0 (1.0, 3.9) 0.03 Cefepime 38
(50) 125 (24) 3.1 (1.8, 5.2) lt0.001 Imipenem 3
(4) 3 (1) 7.0 (0.9, 52.7) 0.03 Gentamicin 20
(26) 86 (17) 1.8 (1.0, 3.2) 0.04 Levofloxacin 2
7 (36) 115 (23) 1.9 (1.1, 3.3) 0.01 Linezolid 2
2 (29) 23 (5) 8.7 (4.3, 17.4) lt0.001 Metronidazo
le 43 (57) 164 (32) 2.8 (1.6,
4.7) lt0.001 Vancomycin 49 (65) 199 (39) 2.9
(1.7, 4.9) lt0.001 Only variables with p values lt
0.05 are shown. HUP Hospital of the University
of Pennsylvania TPN total parenteral
nutrition SAM ampicillin/sulbactam TZP
piperacillin/tazobactam. 1Chi-square or Fishers
exact test (categorical) Students t test or
Wilcoxon Rank Sum test (continuous) 2Median
(interquartile range) 3Days from hospital
admission until positive culture for case groups
or until hospital discharge for control group
4OR unavailable for continuous variables.
INTRODUCTION

• Candida species are the fourth most common cause
  of healthcare-associated BSI.
• Rates of C. albicans BSI have decreased while
  rates of non-C. albicans BSI, particularly Cg,
  have increased.
• Cg is often associated with fluconazole
  resistance.
• Risk factors for fluconazole-resistant Cg-BSI
  have not been well described.
• This is the first study to evaluate independent
  risk factors for healthcare-associated
  fluconazole-resistant Cg-BSI.

ACKNOWLEDGEMENTS
Table 3. Adjusted Risk Factors for Cg-BSI
(Multivariable analyses) Variable
Fluconazole-Resistant1,2 Fluconazole-Susceptible1
,2 Cg Case Group Cg Case
Group Fluconazole 2.3 (1.3, 4.2),
p0.007 1.2 (0.6, 2.4), p0.53 Time at risk3
1.0 (1.0, 1.0), p0.16 ---4 Cefepime 1.6
(0.9, 2.9), p0.09 2.2 (1.2, 3.9),
p0.007 Linezolid 4.6 (2.2, 9.3),
plt0.001 ---4 Metronidazole 1.5 (0.8, 2.6),
p0.18 2.0 (1.1, 3.5), p0.02 Vancomycin 1.3
(0.7, 2.4), p0.35 1.3 (0.7, 2.3),
p0.34 1Adjusted OR (95 CI), p value 2Both
case groups were independently compared to the
control group 3Days from hospital admission
until positive culture for the case groups and
until hospital discharge for the control group
4Variable not included in multivariable model.

• Supported by the Institutional National Research
  Service Award T32 AI055435.
• Also supported by the Agency for Healthcare
  Research and Quality Centers for Education and
  Research on Therapeutics cooperative agreement
  (U18-HS016946 and U18-HS10399).