EORTC/MSG definitions in daily clinical practice: Care

Report
Directed therapy for
fungal infections - latest
advances
Rosemary Barnes
Focus on aspergillosis
Problem

Total UK antifungal
expenditure c £112
million
Rising by 9% pa
Antifungal expenditure is completely out
of proportion with the scale of the
problem



Incidence of IFD in ICU (candida) <0.6%
Aspergillus infection in haematological
malignancy (0.5-12%)
Aspergillus in SOT <5%
Harrison D et al Fungal Infection Risk Evaluation (FIRE) Study. Health Technol Assess 2013; 17(3).
Pagano L et al. Haematologica 2006; 91: 1068-1075
Pagano L et al. Clin Infect Dis 2007; 45: 1161-1170
Reasons




Infection associated with significant morbidity
and mortality
Signs and symptoms of systemic infection are
nonspecific
Conventional diagnostic techniques are
suboptimal
Delays in treatment associated with poorer
outcome
Choices choices choices
Febrile
despite
antibiotics
Empirical therapy
Diagnosis
Change antibiotics
Decision tree
yes
Fever
no
Decision tree
yes
yes
no
No diagnosis
Fever
no
yes
no
Decision tree
yes
no
yes
yes
no
yes
no
No diagnosis Anxiety
Fever
no
yes
yes
no
no
yes
no
Decision tree
yes
no
yes
no
yes
yes
no
yes
no
yes
no
yes
no
yes
no
No diagnosis Anxiety Out of hours
Fever
no
yes
yes
no
no
yes
no
yes
no
yes
no
yes
no
yes
no
Decision tree
yes
no
yes
no
yes
yes
no
yes
no
yes
no
yes
no
yes
no
No diagnosis Anxiety Out of hours
Fever
no
yes
yes
no
no
yes
no
Empirical
therapy
yes
no
yes
no
yes
no
yes
no
No therapy
Aim of a directed strategy
include
all patients likely to have invasive fungal
infection and treat them with the safest
and most effective drug
exclude
all patients unlikely to have invasive
fungal disease and adopt a WAIT-andSEE policy
Maertens et al. (2012) Haematologica 97(3): 325-327.
Consensus criteria


Aimed to provide definitions for proven, probable and
possible fungal infection that could facilitate clinical
research
Designed for use in clinical trials




Highly selective population
Not representative of real life clinical practice
Focus on specific radiological signs
Focus on defining DISEASE
 needs to shift towards INFECTION
 Needs a diagnostic approach

biomarkers
De Pauw et al CID 2008, 46
The biomarkers

Antigen tests




Galactomannan (aspergillus)
Beta D glucan (pan-fungal-ish)
Lateral flow device
Molecular



Aspergillus specific
Panfungal
Commercial (…….)
Galactomannan - in serum


useful test in surveillance: high NPV
Performance in






haematological malignancy better than in SOT
neutropenic > corticosteroid treated group
Adults >children
Influenced by pre-test probability (ie sensitivity
increases with prevalence)
EORTC/MSG criteria heavily dependent on test being
performed
Recommended by ECIL
Galactomannan – meta-analyses
30 studies > 7000 patients




cut-off 0.5: 100 patients:





2 patients with IA, will be missed,
17 patients will be treated unnecessarily
cut-off 1.5 OD:


Prevalence 7.7%
sensitivity78% (61% to 89%)
specificity 81% (72% to 88%).
3 IA patients will be missed
5 patients will be treated unnecessarily
results were very heterogeneous.
Insufficient data to look at clinical utility
http://www.thecochranelibrary.com
BAL


0.5 approved by FDA
On the basis of clinical validity


PUO
PPV of GM BAL is 100% at an OD index cutoff of ≥3
only 76% at ≥ 0.5 (but NPV is high)
GM pos
CT
BAL
Pre test probability
Maertens et al CID 2009
Beta D Glucan

4 different commercial tests




Sensitivity, specificity variable but NPV high





Heterogenous data : retrospective vs. prospective;
Different cut offs
“panfungal” – except cryptococcus and mucoracous moulds
High false-positives: up to 30% - bacteraemia, antibiotics,
pre-/analytical contaminations
complex analytical procedures
Analytical validity established
Utility data limited
Included in EORTC/MSG criteria
Beta D Glucan-meta-analysis

16 studies in 2979 patients







Included case-controlled studies
Included critical care, HM and solid
organ cancer patients
Cut off 10-1000 pg/ml
Sensitivity 76.8% (67.1%–84.3%)
specificity 85.3% (79.6%–89.7%)
“area under ROC curve 0.89”
“good diagnostic accuracy”
Karageorgopoulos et al Clin Infect Dis 2011;52(6):750
PCR

The UK Fungal PCR consensus group



2004 technically validated candida PCR
Made recommendations for aspergillus PCR
2006 European Aspergillus PCR Initiative set up


86 participants in 69 centres in 24 countries
defined a standard for PCR for Aspergillus



Whole blood
Serum
plasma

optimal methodology to evaluate the performance and impact

QCMD available
White et al J Molec Diagn 2006; 8: 376
White et al J Clin Micro 2010: 48 1231
www.eapcri.eu/
PCR
Study
ID
PCR=>2pos
Williamson 2000
Ferns 2003
Buchheidt 2004
Kawazu 2004
Florent 2006
Halliday 2006
Stenghele 2006
White 2006
Jordanides 2005
Subtotal (I-squared = 52.9%, p = 0.030)
.
PCR=>1pos
Hebart 2000 BJH
Hebart 2000 JID
Williamson 2000
Buchheidt 2001
Raad 2002
Ferns 2003
Buchheidt 2004
Kawazu 2004
Lass-Floerl 2004
Scotter 2005
Halliday 2006
Stenghele 2006
El-Mahallawy 2006
Subtotal (I-squared = 62.4%, p = 0.001)
.
Overall (I-squared = 57.1%, p = 0.001)
OR (95% CI)
%
Weight
325.00 (12.25, 8623.71)
0.75 (0.05, 11.31)
6.86 (1.65, 28.50)
11.92 (3.13, 45.34)
15.17 (6.00, 38.34)
81.00 (4.54, 1443.84)
12.50 (2.58, 60.45)
211.50 (24.37, 1835.39)
7.97 (1.52, 41.74)
15.97 (6.83, 37.34)
2.57
3.32
6.14
6.39
7.54
3.08
5.73
4.32
5.51
44.60
139.92 (7.21, 2716.71)
53.28 (2.87, 989.20)
93.00 (4.58, 1889.51)
47.88 (13.87, 165.35)
177.29 (8.35, 3765.07)
1.33 (0.09, 20.11)
3.04 (0.84, 11.06)
7.61 (0.94, 61.24)
1.43 (0.34, 6.08)
55.00 (2.44, 1238.39)
24.51 (1.39, 430.66)
8.66 (1.02, 73.47)
33.75 (8.90, 128.02)
16.41 (6.43, 41.88)
2.96
3.02
2.90
6.66
2.84
3.32
6.51
4.48
6.07
2.77
3.10
4.37
6.40
55.40
16.00 (8.60, 29.79)
100.00


Single negative PCR to
exclude disease
2 consecutive PCRs to
diagnose IA



Sensitivity 88%
Specificity 75%
DOR22
NOTE: Weights are from random effects analysis
.0001
1
10000
Mengoli et al Lancet Infectious Diseases. 2009; 9: 89-96
“Directed/pre-emptive” therapy
References
Country
Clinical
HRCT
GM
Micro-
biological
PCR
Non-comparative studies
Maertens 2005
Belgium
X
X
X
Girmenia 2010
Italy
X
(X)
X
Aguilar-Guisado 2010
Spain
X
Barnes 2009
UK
X
Dignan 2009
UK
X
X
X
X
X
X
Randomised, comparative studies
Cordonnier 2009
Hebart 2009
France
X
X
(X)
Germany
X
Observational studies
Pagano 2012
Italy
X
(X)
Galactomannan EIA


Open study
136 episodes of neutropenia




Patients receiving flucon
prophylaxis
daily EIA GM + early CT
scanning in neutropenic
febrile episodes
Antifungal given if 2
consecutive EIA GM results
+ve (index ≥ 0.5)
and confirmed by BAL or CT
Maertens et al. Clin Infect Dis 2005; 41: 1242
Maertens et al




35% of episodes met criteria for empirical
antifungal but only7.7% treated on basis of preemptive therapy
Duration of fever not affected
22 cases of IFD only one missed
3 breakthrough infections




2 candidaemias
1 mucorales
No excess mortality or fungal related death
No impact on overall antifungal usage despite
deceased empirical use
Cordonnieret al CID 2009 48:1043




293 patients randomised
empirical or pre-emptive therapy
empirical arm received antifungals if they had
persistent/recurrent fever after 4 days
pre-emptive patients given antifungal only if they
showed







clinical and radiological signs of pneumonia/sinusitis
positive GM index ≥ 1.5
Aspergillus colonization
Septic shock
CNS signs/periorbital inflammation
Diarrhoea/mucositis ≥ grade 3
fever > 14 days
Cordonnieret al



Survival was not significantly
“Non inferiority” demonstrated
pre-emptive patients had more IFI



9.1% vs 2.7%
pre-emptive patients received significantly
less antifungals
no significant cost savings were achieved

Used ampho B deoxycholate first -line
Empirical vs. pre-emptive
antifungal therapy
Empirical
Pre-emptive
IFI in Preemptive
IFI in Empirical
Cordonnier et al, Clin Infect Dis, 2009; 48: 1042-1051
Pagano et al Haematologica 2011; 96:1363

Observational: Empiric versus “pre-emptive”

Data collection 397 HM patients




190 empiric ; 207”pre-emptive”
More IFD in pre-emptive arm
Increased mortality and antifungal use in “preemptive arm”
Fever driven, no screening, diagnostic work up not
standardized


some GM usage, no PCR
Pre-emptive group largely diagnosed on basis of HRCT
PCR

Nested PCR to guide
antifungal therapy


42 patients with
cancer, neutropenia
AmB required in only
2 patients

randomised study of
a PCR directed
versus an empirical
antifungal


Lin et al. Clin Infect Dis.
2001;33:1621-1627
more than 400 SCT
patients
Safe


Improved survival at
30 days (not 100)
No reduction in
antifungal drug use.
Hebart et al. Blood 2004;104: 59A.
In Cardiff


549 high-risk haematology patients entering neutropenic
pathway 2005-2010
audited and followed up for a minumum of 12 months



Twice weekly antigen and PCR testing (or GvHD)
Itraconazole prophylaxis or AmBisome 7mg/kg/weekly
Empiric antifungals not used unless




Clinical/mycological evidence of disease
Itraconazole levels were subtherapeutic or unmeasured
First 125 patients analysed for safety and proof of concept
Data collected on compliance, incidence of IFD and
efficacy of prophylaxis
Barnes et al Journal of Clinical Pathology 2009
Incidence of IFD (2005-2011)

Invasive aspergillosis 9.6%

6 histologically proven (2 postmortem)





4 pulmonary (2 with dissemination)
2 invasive sinusitis
47 probable
(23 possible IA)
Invasive Candidal infection 2%

12 proven

4 C. albicans, 3C. glabrata, 2C. tropicalis, 1C. parapsilosis, 1C. guilliermondii, 1
mixedC. albicans + C. glabrata
1 probable
2 non-aspergillus moulds




1 Mucoraceous mould, 1 Scedosporium prolificans
Incidence of proven/probable IFD 12.3%
IA disease status of subjects
By EORTC/MSG diagnostic criteria

Proven – 6

Probable – 47

Possible – 23

NEF – 473
• 248 of NEF showed some signs suggestive of IA
•EIA
positive n=36
•PCR positive n=136
•EIA and PCR positive n=75
•Aspergillus isolated n=5
Diagnostic accuracy

Explore analytical validity. Clinical validity,
clinical utility



Sensitivity specificity
PPV, NPV, LR, DORs
Use ROC analysis to explore different
thresholds for defining “cases”





EORTC/MSG
EORTC – GM EIA
EORTC + PCR
Dual biomarker positivity
Multiple positives versus single
Positivity
Population
Assay
Provena, probableb vs No
IFDc
EIA
Provena, probableb,
possibled vs No IFDc
EIA
Provena,
probable
incorporating PCRe vs No
IFDc
EIA
Provena, probableb vs No
IFDc
PCR
Provena, probableb,
possibled vs No IFDc
PCR
Provena, probable
incorporating PCRe vs No
IFDc
PCR
Provena, probableb vs No
IFDc
Provena, probableb,
possibled vs No IFDc
Provena, probable
incorporating PCRe vs No
IFDc
Provena, probableb vs No
IFDc
PCR or
ELISA
PCR or
ELISA
PCR or
ELISA
Both
PCR
and
ELISA
Provena, probableb,
possibled vs No IFDc
Both
PCR
and
ELISA
Provena, probable
incorporating PCRe vs No
IFDc
Both
PCR
and
ELISA
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
LR+
LR-
DOR
Single
96.2 (87.3-99.0)
76.7 (72.7-80.3)
31.7 (25.0-39.2)
99.5 (98.0-99.9)
4.14
0.05
84.15
Multiple
62.3 (48.8-74.1)
91.8 (88.9-93.9)
45.8 (34.8-57.3)
95.6 (93.3-97.1)
7.55
0.41
18.36
Single
67.1 (55.9-76.6)
76.7 (72.7-80.3)
31.7 (25.0-39.2)
93.6 (93.7-97.6)
2.89
0.43
6.73
Multiple
43.4 (32.9-54.6)
91.8 (88.9-93.9)
45.8 (34.8-57.3)
91.0 (88.1-93.2)
5.27
0.62
8.54
Single
72.9 (61.5-81.9)
76.7 (72.7-80.3)
31.7 (25.0-39.2)
95.0 (92.4-96.8)
3.13
0.35
8.86
Multiple
47.1 (35.9-58.7)
91.8 (88.9-93.9)
45.8 (34.8-57.3)
92.1 (89.4-94.3)
5.72
0.58
9.93
Single
92.5 (82.1-97.0)
55.6 (51.1-60.0)
18.9 (14.6-24.1)
98.5 (96.2-99.4)
2.08
0.14
15.34
Multiple
73.6 (60.4-83.6)
79.9 (76.1-83.3)
29.1 (22.1-37.3)
96.4 (94.1-97.9)
3.66
0.33
11.08
Single
86.8 (77.5-92.7)
55.6 (51.1-60.0)
23.9 (19.3-29.3)
96.3 (93.4-98.0)
1.96
0.24
8.27
Multiple
65.8 (54.6-75.5)
79.9 (76.1-83.3)
34.5 (27.2-42.5)
93.6 (90.7-95.6)
3.28
0.43
7.65
Single
94.3 (86.2-97.8)
55.6 (51.1-60.0)
23.9 (19.3-29.3)
98.5 (96.2-99.4)
2.12
0.10
20.66
Multiple
71.4 (60.0-80.7)
79.9 (76.1-83.3)
34.5 (27.2-42.5)
95.0 (92.4-96.7)
3.56
0.36
9.95
Single
98.1 (93.4-100)
47.8 (42.0-53.5)
17.4 (11.9-22.9)
99.6 (98.5-100)
1.88
0.04
47.6
Multiple
79.2 (65.3-93.2)
76.7 (71.9-81.6)
27.6 (0.19-36.7)
97.1(94.9-99.2)
3.41
0.27
12.6
Single
90.8 (82.5-99.1)
47.8 (42.0-53.5)
21.8 (16.0-27.6)
97.0 (94.2-99.8)
1.74
0.19
9.0
Multiple
69.7 (56.6-82.9)
76.7 (71.9-81.6)
32.5 (23.341.7)
94.0 (91.0-97.1)
3.00
0.39
7.6
Single
98.6 (92.3-100)
47.8 (42.0-53.5)
21.8 (0.17-0.28)
99.6 (96.1-100)
1.89
0.03
63.13
Multiple
90.0 (80.8-95.1)
76.7 (71.9-81.6)
36.4 (27.6-46.2)
98.1 (95.5-99.3)
3.87
0.13
29.7
Single
90.6 (79.8-95.9)
84.4 (80.8-87.4)
39.3 (31.1-48.2)
98.8 (97.1-99.5)
5.79
0.11
51.76
Multiple (1 assay)
71.7 (56.2-87.2)
87.7 (84.0-91.5)
39.6 (27.1-52.1)
96.5 (94.3-98.7)
5.85
0.32
18.13
Multiple (both assays)
56.6 (43.3-69.1)
94.9 (92.6-96.6)
55.6 (42.4-68.0)
95.1 (92.8-96.7)
11.16
0.46
24.40
Single
63.2 (51.9-73.1)
84.4 (80.8-87.4)
39.3 (31.1-48.2)
93.4 (90.7-95.4)
4.04
0.44
9.24
Multiple (1 assay)
50.0 (35.7-91.5)
87.7 (84.0-91.5)
39.6 (27.1-52.1)
91.6 (88.4-94.90
4.08
0.57
7.16
Multiple (both assays)
39.5 (29.3-50.7)
94.9 (92.6-96.6)
55.6 (42.4-68.0)
90.7 (87.8-93.0)
7.78
0.64
12.20
Single
68.6 (57.0-78.2)
84.4 (80.8-87.4)
39.3 (31.1-48.2)
94.8 (92.2-96.5)
4.38
0.37
11.76
Multiple (1 assay)
54.3 (39.4-69.2)
87.7 (84.0-91.5)
39.6 (27.1-52.1)
92.8 (89.8-95.9)
4.43
0.52
8.5
Multiple (both assays)
42.9 (31.9-54.5)
94.9 (92.6-96.6)
55.6 (42.4-68.0)
91.8 (89.1-93.9)
8.45
0.60
14.03
Threshold
Statistical parameters
PCR or
GM EIA
PCR and
GM EIA
Specificity
Sensitivity
Proven,
probable
vs No IFD
PPV
NPV
LR+
LR-
DOR
Single
98.1
47.8
17.4
99.6
1.88
0.04 47.6
multiple
79.2
76.7
27.6
97.1
3.41
0.27 12.6
single
90.6
84.4
39.3
98.8
5.79
0.11 51.76
multiple
56.6
94.9
55.6
95.1
11.16
0.46 24.4
By EORTC/MSG criteria
EORTC
EORTC minus GM
EORTC plus PCR
EIA and PCR multiple (I assay)
EIA and PCR multiple (both assays)
EIA and PCR single
EIA or PCR multiple
EIA or PCR single
PCR multiple
PCR single
EIA multiple
EIA single
Proven/probable disease versus no IFD
Diagnostic odds ratio
Ascertainment bias
90
80
70
60
50
40
30
20
10
0
Positive likelihood ratio
16
14
12
10
8
6
4
2
0
15.8
10.5
11.2
EORTC
EORTC minus GM
EORTC plus PCR
Negative likelihood ratio
0.6
0.5
0.4
0.3 0.05
0.03
0.2
0.1
0
EORTC
EORTC minus GM
EORTC plus PCR
ROC plot
PCR + EIA
Proven/prob/poss
Proven/prob
curve (AUC): 0.910 (95% CI: 0.872-0.948)
Performance of PCR


Utility in proven/probable n=53
First marker positive





PCR in 23
EIA in 15
PCR and EIA simultaneously positive in 7
radiological features in 8
In 85% biomarkers preceded specific
radiological signs (range 1-118d)
Diagnostic accuracy


Screening by PCR AND GM EIA can enable a
diagnosis of IA to be excluded
Positive PCR +GM EIA or multiple positive PCRs
or EIAs can be used to accurately diagnosis IA

specificity 84.4%; sensitivity >90% DOR>50

Biomarkers are earliest markers in 85% of cases

Use antifungals more cost effectively
Antifungal expenditure
£400,000.00
£350,000.00
£300,000.00
Total Antifungal
expenditure
£250,000.00
Cost of testing
£200,000.00
Linear (Total
Antifungal
expenditure)
£150,000.00
£100,000.00
Linear (Cost of
testing)
£50,000.00
£0.00
2005

2006
2007
2008
2009
2010
Similar units typically spending £1-2 mill pa
Use of biomarkers





Regular screening throughout period or
risk
Screening during fever only
Diagnostic testing during refractory
fever only
Confirmation when specific radiological
signs are present
None - empiric therapy
Strategy Used

Influenced by

Risk of IFD

Prevalence affects utility of diagnostic tests


Prophylaxis used


ECIL recommend screening if IFD 5-10%
Mould active reduces utility of diagnostic tests
Availability of


Diagnostic tests
Protective environments/HEPA filtered air
Incidence of IFD after posaconazole
therapy
Pagano et al Haematological 2012; 97:963
Effect of antifungal therapy
Days post infection
1
Group
2
3
4
5
PCR GM
PCR GM
Test
qPCR GM PCR GM PCR GM
Infected controls 2/3
0/3 3/3 1/3 3/3 3/3 3/3 3/3
Amphotericin B
2/3 1/3 3/3 3/3 3/3 3/3
3/3 3/3
Caspofungin
1/3 1/3 1/3 3/3 2/3 3/3
3/3 2/3
Posaconazole
1/3 0/3 1/3 2/3 3/3 3/3
1/3 2/3
Uninfected
controls
McCulloch et al J Clin Path 2012; 65:83
0/3 0/3
Marr K A et al. Clin Infect Dis. 2005;40:1762-1769
Example
*For example: PCR and GM, or
Multiple GM
High risk patient
Prevalence8- ≥10%
No Mould active prophylaxis –
Screening regime
Mould active prophylaxis used
– Diagnostic regime
Twice weekly screening of blood
samples:
Galactomannan,
And Aspergillus PCR
HRCT and BAL when infection suspected
Diagnostic testing during refractory fever with Beta D glucan
(serum) and
Aspergillus PCR (BAL and blood or serum), galactomannan
(BAL and serum)
Single Positive biomarker
Continue screening process
>1 biomarker positive* triggers
diagnostic workup to include
relevant radiology and BAL if
indicated
No consistent clinical signs or
symptoms indicates need for possible
pre-emptive therapy
Targeted antifungal therapy for
clinically diagnosed infection only with
biomarker confirmation
Any consistent clinical signs or
symptoms indicates need for antifungal
therapy
at risk
Prophylaxis
exposure
infection
Pre-emptive
disease
Targetted
Any Questions?

similar documents