Table 1. Ventilator-associated pneumonia diagnostic algorithms utilized in this study.
Published Criteria (citation)
|
Systemic Criteria
|
Chest Criteria
|
Chest Radiography Criteria
|
Microbiologic Criteria
|
CDC/NHSN (33)
|
- Inflammatory response
- Temperature >38°C
- WBC >12,000/mm3 or <4,000/mm3
- OR new antimicrobial agent is started for ≥4 days
→ Infection-related ventilator-associated complication
|
After a period of stability or improvement on the ventilator (≥2 calendar days of stable or ↓ FiO2 or PEEP):
- Minimum daily FiO2 ↑ ≥0.20 lasting 2 days
- Or minimum daily PEEP values ↑ ≥3 cm H2O lasting 2 days
→ Ventilator-associated condition
|
----
|
Microbiologic quantitative (+), OR histologic (+), OR (+) for legionella, influenza, RSV, adenovirus, or parainfluenza virus
AND
Gram-stain evidence ≥25 neutrophils/lpf and ≤10 epithelial cells/lpf
→ Probable VAP
|
CPIS a
(34)
|
Fever:
- 38.5-38.9 (1 point)
- ≥39 or <36.5 (2 points)
WBC:
- <4,000/mm3 or >11,000/mm3 (2 points)
|
- Secretions but not purulent (1 point)
- Purulent secretions (2 points)
- PaO2/ FiO2 <240 without ARDS (2 points)
|
Diffuse infiltrate (1 point)
Localized infiltrate (2 points)
Progressive infiltrate (without cardiac disease or ARDS) (+2)
|
Positive (1 point)
|
HELICS b
(35)
|
At least 1 criterion:
- Temperature >38°C (with no other cause)
- WBC >12,000/mm3 or <4,000/mm3
- If age >70 years: AMS without other cause
|
At least 1 of following criteria (2 clinical pneumonia only = PN4 and PN5):
- New onset purulent sputum or change in sputum character (color, odor, quantity, consistency)
- Cough or dyspnea or tachypnea
- Suggestive auscultation (rales or bronchial breath sounds, rhonchi, wheezing)
- Worsening gas exchange (O2 desaturation, increasing FiO2 requirements or ventilation demands)
|
Image suggestive of pneumonia. (≥2 serial chest X-rays or CT scans with suggestive imaging for patients with underlying cardiac or pulmonary disease)
|
- PN1 – (+) quantitative Cx from minimally contaminated LRT specimen c
- PN2 – (+) quantitative Cx from possibly contaminated LRT specimen d
- PN3 – Alternative methods: e (+) blood or pleural Cx, pleural or pulmonary abscess, histology, or pathogen antigen or antibody testing.
- PN4 – (+) sputum Cx or non-quantitative LRT specimen Cx
- PN5 – No positive results
|
Johanson
(36)
|
Temperature >38.5°C
WBC >12,000/mm3
|
Purulent secretions
|
New or progressive consolidation
|
----
|
a Score >6 is suggestive of VAP.
b VAP diagnosis if criteria met and invasive respiratory device (even intermittently) in the 48 hours preceding the onset of infection.
c Either: (1) Broncho-alveolar lavage (BAL) with a threshold of ≥ 104 cfu/mL or ≥ 5 % of BAL obtained cells contains intracellular bacteria on direct microscopic exam; (2) Protected brush (PB Wimberley) with a threshold of ≥ 103 cfu/mL; (3) Distal protected aspirate (DPA) with a threshold of ≥ 103 cfu/mL.
d Quantitative culture of LRT specimen (e.g., endotracheal aspirate) with a threshold of 106 cfu/mL.
e Either: (1) positive blood culture not related to another source of infection; (2) positive growth in culture of pleural fluid; (3) pleural or pulmonary exam shows evidence of pneumonia; (4) positive exams for pneumonia with virus or particular germs (Legionella, Aspergillus, mycobacteria, Mycoplasma, Pneumocystis carinii) . The latter may include: (A) positive detection of viral antigen or antibody from respiratory secretions (e.g., EIA, FAMA, shell vial assay, PCR); (B) positive direct exam or positive culture from bronchial secretions or tissue; (C) seroconversion (ex: influenza viruses, Legionella, Chlamydia); or (D) detection of antigens in urine (Legionella).
AMS: altered mental status; ARDS: acute respiratory distress syndrome; CDC/NHSN: centers for disease control and prevention national health safety network; CPIS: clinical pulmonary infection score; Cx: culture; FiO2: fraction of inspired oxygen; HELICS: hospital in Europe link for infection control through surveilance; LRT: lower respiratory tract; PaO2: partial pressure of oxygen in arterial blood; RSV: respiratory syncytial virus; VAP: ventilator associated pneumonia; WBC: white blood cell
Table 2. Performance characteristics of ventilator-associated pneumonia diagnostic algorithms.
Criteria Studied
|
Year, (citation)
|
Population
|
Comparator
|
Sample Size
|
Sensitivity
|
Specificity
|
PPV
|
NPV
|
(+) LR
|
(-) LR
|
ROC AUC
|
Kappa (ĸ) index, agreement level a
|
CDC/NHSN
|
2015, (81)
|
Mixed b
|
CPIS
|
38
|
0.37
|
1.0
|
1.0
|
0.84
|
|
|
|
ĸ = 0.47, moderate
|
CPIS
|
1999, (42)
|
Mixed b
|
Pathology
|
23
|
0.77
|
0.42
|
|
|
|
|
|
|
CPIS c
|
2004, (85)
|
Mixed b
|
Quantitative Cultures
|
69
|
0.41
|
0.77
|
0.8
|
0.36
|
|
|
0.64
|
|
CPIS
|
2004, (86)
|
Mixed b
|
Quantitative Cultures
|
88
|
0.89
|
0.47
|
0.57
|
0.84
|
|
|
|
ĸ = 0.33, fair
|
CPIS
|
2007, (47)
|
Burn
|
Quantitative Cultures
|
28
|
0.30
|
0.80
|
0.70
|
0.50
|
|
|
|
|
CPIS
|
2010, (40)
|
Mixed b
|
Pathology
|
142
|
0.46
|
0.60
|
|
|
1.13
|
0.96
|
|
|
CPIS
|
2015, (48)
|
Surgical (mixed)
|
Quantitative Cultures
|
497
|
0.633
|
0.644
|
0.61
|
0.674
|
|
|
0.60
|
|
CPIS
|
2018, (87)
|
Surgical (acute care)
|
Quantitative Cultures
|
198
|
0.611
|
0.781
|
0.64
|
0.759
|
|
|
|
|
HELICS
|
2013, (59)
|
Mixed b
|
Not clearly specified
|
57 d
|
0.86
|
0.99
|
0.77
|
0.995
|
|
|
|
ĸ = 0.80, substantial
|
Johansen e
|
1999, (42)
|
Mixed b
|
Pathology
|
23
|
0.69
|
0.75
|
|
|
|
|
|
|
Johansen
|
2018, (87)
|
Surgical (acute care)
|
Quantitative Cultures
|
198
|
0.828
|
0.59
|
0.564
|
0.843
|
|
|
|
|
NTDB/NTR
|
2015, (88)
|
Trauma
|
CDC/NHSN
|
279
|
0.864
|
0.578
|
0.74
|
0.74
|
|
|
|
ĸ = 0.47, moderate
|
a Agreement based on score: ≤ 0 (no agreement); 0.01–0.20 (slight); 0.21–0.40 (fair); 0.41– 0.60 (moderate); 0.61–0.80 (substantial); and 0.81–1.00 (almost perfect agreement).
b A mixed population containing both medical and surgical patients. Studies that did not specify ICU type were by default classified as mixed.
c For CPIS threshold of >7, rather than current standard of >6. The AUC using threshold CPIS >6 was 0.54, other values not reported.
d Data from sub-population of a larger study assessing various types of ICU-acquired infections.
e The presence of all three criteria increased the specificity to 92% at the cost of a high beta error (sensitivity 23%).
CDC/NHSN: centers for disease control and prevention national health safety network; ICU: intensive care unit; MV: mechanical ventilation; NR: not reported; NTDB/NTR: national trauma data bank / national trauma registry; NPV: negative predictive value; PPV: positive predictive value; ROC AUC: receiver operating curve area under curve; (+) LR : positive likelihood ratio; (-) LR: negative likelihood ratio
Table 3. Patient demographic and clinical information.
Variable
|
All
|
VAP
n=45
|
No VAP
n=40
|
p-Value
|
Age, years, mean (SD)
|
46.9 (18.9)
|
44.2 (20.7)
|
49.9 (16.4)
|
0.159 a
|
Male, mean (SD)
|
62 (72.9)
|
33 (73.3)
|
29 (72.5)
|
0.931 c
|
Admission indication, N (%)
Trauma
Post-operative
|
54 (63.5)
31 (36.5)
|
30 (66.7)
15 (33.3)
|
24 (60)
16 (40)
|
0.652 c
|
Comorbidities, yes, N (%)
ARDS
Cancer
COPD
CHF
ESRD
Multiple trauma
|
7 (8.2)
13 (15.3)
7 (8.2)
24 (28.2)
14 (16.5)
20 (23.5)
|
3 (6.7)
6 (13.3)
4 (8.9)
13 (28.9)
9 (20)
10 (22.2)
|
4 (10)
7 (17.5)
3 (7.5)
11 (27.5)
5 (12.5)
10 (25)
|
0.932 b
|
Positive tracheal culture, N (%)
|
69 (81.2)
|
40 (88.9)
|
29 (72.5)
|
0.093 c
|
MDR organism, yes, N (%)
|
31 (36.5)
|
17 (37.8)
|
14 (35)
|
0.825 c
|
Procalcitonin, ng/mL, mean (SD)
|
4.03 (4.68)
|
3.53 (3.6)
|
4.6 (5.6)
|
0.308 a
|
APACHE II, mean (SD)
|
18.1 (2.84)
|
17.9 (3.43)
|
18.4 (1.98)
|
0.399 a
|
Duration of intubation, hours, mean (SD)
|
177.1 (39.61)
|
176.02 (38.7)
|
178.32 (41.09)
|
0.791 a
|
Reintubation, N (%)
|
32 (37.6)
|
14 (31.1)
|
18 (45)
|
0.262 c
|
MV duration prior to VAP, hours, median (IQR)
|
72 (54-87.5)
|
72 (52-87.5)
|
72 (64.5-88.5)
|
0.639 a
|
ICU duration prior to developing VAP, days, median (IQR)
|
7 (6-8)
|
7 (6-8.5)
|
7 (6-8)
|
0.118 a
|
VAP timing, mean (SD)
Early (< 5 days)
Late (≥ 5 days)
|
------
|
15 (33.3)
30 (66.7)
|
------
|
------
|
Length-of-stay, days, mean (SD)
ICU LOS
Non-ICU LOS
|
9.8 (3.0)
15.4 (3.1)
|
13.13 (3.27)
12.67 (3.34)
|
12.72 (2.75)
11.96 (2.99)
|
0.538 a
0.320 a
|
Mortality, N (%)
ICU
Hospital
|
17 (20)
22 (25.9)
|
8 (17.8)
12 (26.7)
|
9 (22.5)
10 (25)
|
0.787 c
0.861 c
|
VAP: ventilator-associated pneumonia; IQR: interquartile range; MDR: multiple drug resistant; APACHE: Acute Physiology and Chronic Health Evaluation; ICU: intensive care unit; VAP: ventilator-associated pneumonia; LOS: length-of-stay
a Independent sample t-test
b Fisher exact test
c Chi-square
Table 4. Sensitivity, specificity, and Youden index for assessed methods of ventilator-associated pneumonia diagnosis compared to the HELICS criteria as the reference standard.
Criteria
|
Ventilator-Associate Pneumonia
|
% Sensitivity
|
% Specificity
|
Youden index a
|
Positive
|
Negative
|
Total
|
CDC/NHSN
Positive
Negative
Total
|
45
0
45
|
38
2
40
|
83
2
85
|
54.22
|
100
|
0.542
|
CPIS
Positive
Negative
Total
|
44
1
45
|
20
20
40
|
64
21
85
|
68.75
|
95.23
|
0.640
|
Johanson
Positive
Negative
Total
|
44
1
45
|
21
19
40
|
65
20
85
|
67.69
|
95
|
0.627
|
a A measure of the maximum diagnostic accuracy, where 1 signifies a perfect test and 0 signifies no diagnostic value.
CDC/NHSN = centers for disease control and prevention national health safety network; CPIS = Clinical Pulmonary Infection Score, HELICS = Hospital in Europe Link for Infection Control through Surveillance.
Table 5. Correlation of serum procalcitonin and tracheal aspirate results with ventilator-associated pneumonia diagnostic algorithms.
Criteria
|
Serum Procalcitonin Level, ng/mL
|
Kappa (ĸ) index, agreement level a
(p-Value)
|
Tracheal Culture
|
Kappa (ĸ) index, agreement level a
(p-Value)
|
< 0.25
|
0.25-0.5
|
> 0.5
|
Total
|
Positive
|
Negative
|
Total
|
Johanson, n (%)
Positive
Negative
Total
|
10 (15.4)
18 (90)
28 (32.9)
|
9 (13.8)
0
9 (10.6)
|
46 (70.8)
2 (10)
48 (56.5)
|
65
20
85 (100)
|
0.47, moderate
(<0.001)
|
61 (93.8)
8 (40)
69 (81.2)
|
4 (6.2)
12 (60)
16 (18.8)
|
65
20
85 (100)
|
0.579, moderate
(<0.001)
|
CDC/NHSN, n (%)
Positive
Negative
Total
|
26 (31.3)
2 (100)
28 (32.9)
|
10 (12.0)
0
10 (11.8)
|
47 (56.6)
0
47 (55.3)
|
83
2
85 (100)
|
0.06, slight
(0.58)
|
67 (80.7)
2 (100)
69 (81.2)
|
16 (19.3)
0
16 (18.8)
|
83
2
85 (100)
|
0.04, slight
(0.49)
|
CPIS, n (%)
Positive
Negative
Total
|
11 (17.5)
17 (77.3)
28 (32.9)
|
8 (12.7)
1 (4.5)
9 (10.6)
|
44 (69.8)
4 (18.2)
48 (56.5)
|
63
22
85 (100)
|
0.42, moderate
(<0.001)
|
61 (96.8)
8 (36.4)
69 (81.2)
|
2 (3.2)
14 (63.6)
16 (18.8)
|
63
22
85 (100)
|
0.663, substantial
(<0.001)
|
a Agreement based on score: ≤ 0 (no agreement); 0.01–0.20 (slight); 0.21–0.40 (fair); 0.41– 0.60 (moderate); 0.61–0.80 (substantial); and 0.81–1.00 (almost perfect agreement).
CDC/NHSN: centers for disease control and prevention national health safety network; CPIS: Clinical Pulmonary Infection Score, HELICS: Hospital in Europe Link for Infection Control through Surveillance
Table 6. Kappa agreement coefficient among ventilator-associated pneumonia diagnostic methods.
Criteria
|
Kappa (ĸ) index, agreement level a
|
p-Value
|
CPIS and Johanson
|
0.874
|
<0.001
|
CDC/NHSN and Johanson
|
0.145
|
<0.001
|
CDC/NHSN and CPIS
|
0.129
|
0.015
|
a Agreement based on score: ≤ 0 (no agreement); 0.01–0.20 (slight); 0.21–0.40 (fair); 0.41– 0.60 (moderate); 0.61–0.80 (substantial); and 0.81–1.00 (almost perfect agreement).
CDC/NHSN: centers for disease control and prevention national health safety network; CPIS: Clinical Pulmonary Infection Score, HELICS: Hospital in Europe Link for Infection Control through Surveillance
Table 7. Correlation of individual variables with ventilator-associated pneumonia diagnostic methods.
Parameter
|
Kappa agreement coefficient
|
CDC/NHSN
|
CPIS
|
Johanson
|
PCT >0.5 ng/ml
|
0.061
|
0.423
|
0.470
|
Infiltrate on radiograph
|
-0.045
|
0.874
|
0.738
|
Temperature
|
-0.044
|
0.529
|
0.579
|
WBC
|
-0.044
|
0.739
|
0.729
|
PaO2
|
-0.038
|
0.094
|
-0.139
|
Tracheal culture
|
0.044
|
0.663
|
0.579
|
Blood culture
|
-0.011
|
0.238
|
0.165
|
CDC/NHSN: centers for disease control and prevention national health safety network; CPIS: Clinical Pulmonary Infection Score, HELICS: Hospital in Europe Link for Infection Control through Surveillance, PCT: Serum procalcitonin; WBC: White blood cell; PaO2: Partial pressure of O2 in arterial blood.