Abstract
Systemic kinetics of three inflammatory mediators (bactericidal/permeability-increasing protein [BPI], soluble intercellular adhesion molecule [sICAM], and soluble E-selectin [sE-selectin]) were studied during the development of ventilator-associated pneumonia (VAP) (n = 42), diagnosed on quantitative cultures of bronchoscopic samples. From a pool of collected samples, nested samples were used to measure mediators on Days − 4, − 2, 0, and + 2, relative to diagnosis. Correlations between systemic levels of mediators and clinical severity of infection (VAP with or without severe sepsis or septic shock) and patient outcome (mortality at Day 10 after diagnosis) were studied. Predictive values of inflammatory mediators were compared with daily Simplified Acute Physiology Score II (SAPS II) values and the logarithmic number of bacteria in bronchoscopic samples. During the development of VAP, increasing SAPS II scores and rising systemic mediator levels were only found in patients in whom VAP was accompanied with severe sepsis or septic shock. Values of SAPS II and plasma levels of BPI and sE-selectin, but not sICAM, increased from the day of diagnosis on in patients who died within 10 d of diagnosis. Systemic levels of inflammatory mediators did not better predict clinical severity or patient outcome than daily SAPS II scores. The logarithmic number of bacteria in bronchoscopic samples poorly correlated with circulating levels of inflammatory mediators, severity of infection, and patient outcome. Our findings show that a clinical scoring system (SAPS II score) is at least as good as a predictor for the clinical severity of infection and patient outcome, and provide new information on the kinetics of inflammatory mediators during the development of VAP.
Ventilator-associated pneumonia (VAP) is a frequently occurring infection among critically ill patients. Even when the diagnosis of VAP is based on quantitative culture results from bronchoscopic techniques, the clinical presentation of infection may range from a devastating illness with irreversible septic shock to mild and almost unnoticed (1).
The inflammatory response of the host during the development of VAP has not been studied extensively. The roles of systemic and localized inflammatory responses in the pathophysiology have not been elucidated, and it is unknown to what extent systemic levels of cytokines or inflammatory mediators predict the clinical severity of infection and patient outcome, and how they correlate with values of a clinical scoring system of illness.
In a previous case-control study of patients with and without VAP, only patients with VAP accompanied with a clinical presentation of severe sepsis or septic shock had elevated circulating levels of the cytokines interleukin-6 (IL-6) and interleukin-8 (IL-8) and an increased mortality rate at Day 10 after diagnosis (1). However, within individual patients the development of VAP was not associated with increasing concentrations of these cytokines and no correlation was found between the yield of quantitative culture results from bronchoscopic samples (bronchoscopic bacterial burden) and plasma levels IL-6 or IL-8 (1).
The primary aim of the present study was to determine whether systemic levels of three mediators, representing different parts of the inflammatory response to infection, correlated with clinical assessments of severity of illness and if they predicted the outcome of these patients. The severity of illness of patients with VAP was assessed on a daily Simplified Acute Physiology Score II (SAPS II) (2), and on the logarithmic number of bacteria isolated from bronchoscopic samples.
Additional aims of this study were to further determine the inflammatory response of critically ill patients during the development of VAP. Therefore, three parts of the inflammatory process were studied: bactericidal/permeability-increasing protein (BPI), soluble intercellular adhesion molecule (sICAM), and soluble E/selectin (sE-selectin). BPI is released by activated or killed neutrophils and circulating levels represent a measure of activation and/or killing of these cells (3, 4). ICAM is present on the surface of almost all cells and increased circulating levels of sICAM indicate the presence (and level) of inflammation (5-7). sE-selectin is only expressed on activated endothelial cells and is released by activation of these cells. Circulating levels, therefore, represent endothelial activation (5). Increased circulating levels of each of these inflammatory mediators have been associated with the presence of severe sepsis or septic shock in critically ill patients (6, 8-10).
The study was conducted at the intensive care unit (ICU) of the University Hospital, Maastricht, The Netherlands. The ICU is a 16-bed ward with a mixture of patients from the departments of surgery, internal medicine, traumatology, pulmonology, neurology, and neurosurgery. The 42 patients included in the present study have been studied previously in a matched-cohort analysis, and extensive clinical and microbiological characteristics of these patients have been reported (1).
VAP was considered ICU-acquired if a clinical condition fulfilling the criteria for pneumonia developed after the patient was in the ICU for at least 3 d. In case of a clinical suspicion of pneumonia, bronchoscopy with bronchoalveolar lavage (BAL) and protected specimen brush (PSB) was performed as described previously (11). The diagnosis VAP was established with either a positive quantitative culture of samples obtained by BAL (cutoff point 104 cfu/ml) or PSB (cutoff point 103 cfu/ml); a new or persistent infiltrate on chest radiograph; and when at least three of the following four criteria were met: (1) rectal temperature > 38.0° C or < 35.5° C; (2) blood leukocytosis (> 10.103/mm3) and/or left shift or blood leukopenia (< 3.103/mm3); (3) > 10 leukocytes per high-power field in Gram stain of tracheal aspirate; and (4) a positive culture from tracheal aspirate. The logarithmic number of bacteria present in quantitatively cultured samples obtained by BAL was defined as the “bronchoscopic bacterial burden.”
Plasma samples of all patients treated in the ICU from January 1, 1992 until January 1, 1994 were prospectively collected and frozen at −70° C until use. From this pool, a nested group of patients was selected and their samples were used for analyses. Circulating levels of BPI, sICAM and sE-selectin were measured in samples obtained at the day of diagnosis of VAP (D0), 4 d (D−4) and 2 d (D−2) before diagnosis, and 2 d after diagnosis (D+2). In addition, the changes in circulating levels of these cytokines between D−4 and D+2 and between D0 and D+2 were determined. When blood samples were not available on D−4 (n = 5) the levels of D−2 were used to calculate the difference with levels on D+2, and when samples were not available on D+2 (n = 3) the levels of D0 were used. Patients with missing data at D+2 (n = 3) were excluded for the analysis of differences between levels of D0 and D+2. The SAPS II was calculated on each study day as described by Le Gall and coworkers (2). Furthermore, all patients were grouped on the presence or absence of severe sepsis or septic shock on the day that VAP was diagnosed, defined according to the criteria from the American College of Chest Physicians and the Society of Critical Care Medicine (12), and on survival of the first 10 d after VAP was diagnosed. The period of 10 d was chosen on the assumption that the infection contributed directly to death in these patients or that contribution of VAP to death of any other cause could not be excluded.
Four types of analyses were performed: (1) values of SAPS II scores and levels of circulating cytokines, and differences of these values between D+2 and D−4, were correlated to the clinical severity of VAP (i.e., the presence of severe sepsis or septic shock) and (2) to patient outcome (i.e., mortality at Day 10); (3) correlations between the SAPS II score on D0 and the levels of circulating inflammatory mediators were determined; and (4) values of the bronchoscopic bacterial burden were correlated to SAPS II scores on D0 and to circulating levels of inflammatory mediators on D0.
Reagents and materials. Human recombinant (r) BPI was kindly provided by M. Marra (InCyte, Palo Alto, CA). sICAM and sE-selectin standards were obtained by purification of supernatant produced by NSO cells that produce sICAM, and CHO cells that produce sE-selectin, respectively (kindly provided by M. Robinson, Celltech, Slough, UK).
A BPI-specific and neutralizing monoclonal antibody (mAb) 4E3 (IgG1) was developed in our laboratory and described elsewhere (13). Polyclonal antibodies to human BPI were obtained by immunizing rabbits with human BPI.
For detection of ICAM-1, anti-ICAM-1 mAb HM2, and biotin- labeled HM1 were used (14). Anti-E-selectin mAb ENA-1 and biotin-labeled ENA-2 were used for measuring E-selectin (7). Peroxidase-conjugated streptavidin was purchased from Dakopatts (Glostrup, Denmark), and TMB (3′,5,5′-tetramethylbenzidine) substrate from KPL (Gaithersburg, MD). ELISA plates used were Nunc immuno maxisorp plates (Nunc, Roskilde, Denmark).
Immunoassays. Plasma BPI was measured using a sandwich ELISA as described elsewhere (15). In short, 96-well plates were coated with human BPI–specific mAb 4E3 and free sites were blocked with phosphate-buffered saline (PBS) 1% bovine serum albumin (BSA). Washing and dilution buffer used contained 80 mM MgCl. The use of Mg++ ions prevented disturbance by lipopolysaccharide (LPS) of BPI measurement. Human rBPI was used for standard titration curve. Plasma samples diluted in assay buffer (at least 1:2) were incubated for 2 h at room temperature. Next, a biotinylated polyclonal rabbit anti-human BPI IgG was used followed by peroxidase-labeled streptavidin. TMB was used as a substrate and photospectrometry (450 nm) was performed using a micro ELISA autoreader. Detection limit for the BPI was 200 pg/ml.
Plasma E-selectin and sICAM-1 concentrations were measured using mAb ENA-1 and mAb HM2 as capture antibodies, respectively, in a procedure largely parallel to the BPI ELISA mentioned previously. Plasma samples were diluted 1/20 and 1/10 for the sE-selectin and sICAM-1 ELISA, respectively. Samples and standard titration curves with sE-selectin and sICAM were added to the plates and subsequently incubated with biotinylated mAb ENA-2 or HM1 for detection of sE-selectin and sICAM, respectively. Substrate was added, and the optical density was read. All steps of the sE-selectin ELISA were done using a buffer containing calcium and magnesium. Detection limits for the sE-selectin and sICAM-1 assays were 0.1 ng/ml and 0.4 ng/ml, respectively.
All plasma samples were analyzed in the same run. When plasma levels exceeded the detection limit of the assay, samples were additionally diluted and analyzed in a separate run with an overlap to correct for interassay variation. The intra- and interassay coefficients of variance of the various assays performed were all < 10%.
Data are expressed as mean levels with standard deviations and ranges. Comparisons of median levels of BPI, sICAM, and sE-selectin were performed using the Mann-Whitney U test for nonparametric values. Differences in parametric values were tested with Student's t test. Categorical variables were compared by the chi-square test. Correlations were tested by the Spearman rank test. In general, a value of p < 0.05 was deemed significant. When analyzing differences in circulating levels of mediators at four different time points, the Bonferroni correction has been used and a value of p < 0.0125 was deemed statistically significant.
Forty-two patients who developed VAP were studied. The mean age was 59 ± 18 yr (range, 17 to 85) and the mean Acute Physiology and Chronic Health Evaluation (APACHE II) score, as obtained on admission, was 21 ± 7 (range, 7 to 36). The values of SAPS II scores and levels of circulating inflammatory mediators on each of the days of study are presented in Table 1. Although values of SAPS II scores were highest on D0, no evident increase of inflammatory mediators was observed during the development of VAP. VAP was diagnosed after 8 ± 4 d of mechanical ventilation. Fifteen patients had polymicrobial VAP, and Pseudomonas aeruginosa was isolated most frequently. VAP was monobacterial and caused by gram-positive microorganisms in only five patients (Staphylococcus aureus [n = 3] and Streptococcus pneumoniae [n = 2]). Therefore, no comparisons were made between infections with gram-positive and gram-negative bacteria. The bacterial burden of infection ranged from four in nine patients to 16 in one patient. Two patients fulfilled the criteria of severe sepsis at the day of diagnosis and eight (19%) had septic shock. Mortality at Day 10 was nine (21%) of 42 patients.
| Day Relative to Diagnosis of Ventilator-associated Pneumonia | ||||||||
|---|---|---|---|---|---|---|---|---|
| Day −4 | Day −2 | Day 0 | Day +2 | |||||
| SAPS II value, mean ± SD (range) | 39 ± 11 (19–66) | 37 ± 10 (17–64) | 43 ± 18 (20–102) | 42 ± 19 (14–121) | ||||
| Plasma samples available | 36 (85%) | 42 (100%) | 42 (100%) | 39 (93%) | ||||
| BPI, ng/ml (mean ± SD) | 4.12 ± 5.23 | 3.00 ± 3.12 | 4.18 ± 4.33 | 3.76 ± 4.12 | ||||
| sICAM, ng/ml (mean ± SD) | 224.2 ± 220.0 | 230.8 ± 244.7 | 252.0 ± 270.7 | 257.4 ± 222.4 | ||||
| sE-selectin, ng/ml (mean ± SD) | 79.1 ± 85.3 | 71.0 ± 62.3 | 73.3 ± 57.4 | 74.3 ± 53.4 | ||||
SAPS II scores. In 10 of 42 patients, VAP was associated with a clinical presentation of severe sepsis or septic shock. The mean SAPS II scores on D−4 were similar for patients with severe sepsis or septic shock (n = 10) and the remaining patients (n = 32), but from D−2 on, mean SAPS II scores increased for patients who developed severe sepsis or septic shock, whereas SAPS II scores remained stable in the other patients (Figure 1A).
Fig. 1. Mean SAPS II values and mean plasma levels of inflammatory mediators (with standard error) in patients with VAP which was associated (open circles) or was not associated (closed circles) with severe sepsis or septic shock. The data are according to the day of diagnosis of VAP (Day 0). (A) SAPS II values. (B) Plasma levels of BPI. (C ) Plasma levels of sE-selectin. (D) Plasma levels of sICAM. * p < 0.0125.
[More] [Minimize]Inflammatory mediators. On the 4 d of study, BPI was detectable in 89 to 97% of plasma samples and sICAM and sE-selectin were detectable in all plasma samples. Circulating levels of BPI, sICAM, and sE-selectin were higher for patients with a clinical presentation of severe sepsis or septic shock, although statistical significance was not reached (Figures 1B– 1D). Interestingly, differences in systemic levels of the inflammatory mediators between patients with and without severe sepsis or septic shock were not necessarily highest on D0, but levels tended to be higher in patients with severe sepsis or septic shock on all days of study.
SAPS II scores. Before diagnosis of VAP, SAPS II scores were similar for patients who survived and who did not survive the first 10 d after diagnosis (Figure 2A). From D0 on, patients who succumbed had higher SAPS II scores, although statistical significance was only approached on D0 (p = 0.06, t test). As a result, the mean differences in SAPS II scores between D−4 and D+2 were −1 ± 10 (range, −21 to 22) for patients who survived the first 10 d after VAP and 38 ± 31 (range, 0 to 89) for those who died within this time period (p = 0.0001, Mann-Whitney U test).
Fig. 2. Mean SAPS II values and mean plasma levels of inflammatory mediators (with standard error) in patients who died (open circles) or who survived (closed circles) the first 10 d after the diagnosis of VAP. The data are according to the day of diagnosis of VAP (Day 0). (A) SAPS II values. (B) Plasma levels of BPI. (C ) Plasma levels of sE-selectin. (D) Plasma levels of sICAM. * p < 0.0125.
[More] [Minimize]Inflammatory mediators. Circulating levels of BPI from D−4 to D0 were similar for patients who succumbed within 10 d of diagnosis and those who survived this period (Figure 2B). However, levels of BPI on D+2 were 5.78 ng/ml in patients who died within 10 d of diagnosis as compared with 1.65 ng/ml in survivors (p = 0.01, Mann-Whitney U test). The differences in BPI levels between D0 and D+2 were 3.22 ng/ml and −0.39 ng/ml for succumbing and surviving patients respectively (p = 0.0003, Mann-Whitney U test). An increase in circulating levels of BPI after the diagnosis of VAP was found in all (100%) patients who died within 10 d of diagnosis and from whom plasma samples were available (n = 6), but in only 11 (33%) of 33 patients who survived (p = 0.0007, chi-square test).
There were no significant differences between levels of sICAM on any of the four days of study for patients dying within or surviving the first 10 d after diagnosis of VAP, and for both groups, circulating levels of sE-selectin were similar on the first 3 d of study (Figures 2C and 2D). However, circulating levels of sE-selectin at D+2 were 114.3 ± 87.5 ng/ml and 67.0 ± 42.8 ng/ml (p = 0.04) for patients who died within 10 d of VAP and patients who survived, respectively. The differences in sE-selectin levels between D0 and D+2 were 67.2 ng/ml and −1.64 ng/ml for succumbing and surviving patients respectively (p = 0.06, Mann-Whitney U test).
Poor correlations were found between the values of the SAPS II scores on D0 and circulating levels of BPI (r = 0.08, p = 0.59), sICAM (r = 0.16, p = 0.30), and sE-selectin (r = 0.18, p = 0.25) on D0 (Figures 3A–3C).
Fig. 3. Correlations between the SAPS II values and the plasma levels of inflammatory mediators on the day of diagnosis of VAP. (A) Plasma levels of BPI. (B) Plasma levels of sE-selectin. (C ) Plasma levels of sICAM.
[More] [Minimize]Mean levels of the bronchoscopic bacterial burden (expressed as the logarithmic number of bacteria isolated from samples of BAL) were comparable for patients with and without severe sepsis or septic shock (6.7 ± 2.4 and 7.4 ± 3.5, respectively), and for patients surviving and not surviving the first 10 d after diagnosis of VAP (7.6 ± 3.4 and 5.2 ± 1.3, respectively). Correlations between the bronchoscopic bacterial burden of infection and any of the circulating inflammatory mediators on D0 as well as the SAPS II scores on D0 were poor: r = 0.17 (p = 0.32) for the SAPS II scores; r = 0.16 (p = 0.35) for BPI; r = −0.24 (p = 0.16) for sICAM; and r = 0.3 (p = 0.85) for sE- selectin (Figures 4A and 4B).
Fig. 4. Correlations between the logarithmic number of bacteria per milliliter in samples of BAL and SAPS II values and plasma levels of inflammatory mediators on the day of diagnosis of VAP. (A) SAPS II values. (B) Plasma levels of BPI. (C ) Plasma levels of sE-selectin. (D) Plasma levels of sICAM.
[More] [Minimize]The present study shows, in a mixed population of critically ill patients with bronchoscopically established diagnoses of VAP, three aspects about interactions between systemic inflammatory mediators, clinical presentation of infection, and prediction of outcome. The first and most important finding is that the SAPS II scoring system on the day of diagnosis correlated at least as well with the clinical severity of infection or 10-d mortality as with plasma levels of inflammatory mediators. Second, the development of VAP is accompanied by increased SAPS II values and a tendency of higher levels of BPI, sICAM, and sE-selectin only when there is a clinical presentation of severe sepsis or septic shock. When VAP is not complicated by this presentation, which applies for the majority of cases, concentrations of inflammatory mediators and SAPS II scores remain stable. Third, patients who succumb within 10 d after the diagnosis of VAP have rising values of SAPS II scores and rising systemic levels of BPI, and sE-selectin, but not sICAM, from the day of diagnosis on. Again, stable levels were found for patients who survived the first 10 d after diagnosis of VAP.
We found that the SAPS II score on the day of diagnosis of VAP was at least as good a predictor of the clinical severity of infection and the prognosis of the patient, as were circulating levels of BPI, sICAM, and sE-selectin. A simple bedside scoring system, therefore, seems to reflect the clinical severity of infection better than sophisticated measurements of the systemic inflammatory response. SAPS II scores can easily and rapidly be calculated on the patients' bedside, and this method, obviously, has many advantages over the time-consuming, labor-intensive, and expensive procedures to determine circulating levels of inflammatory mediators. In the present study the SAPS II score was used as a clinical severity scoring system, but other scoring systems such as the daily calculation of the APACHE II or APACHE III (16) score, or the Mortality Probablity Model (17), may prove to be useful as well. Whether circulating levels of other inflammatory mediators or cytokines correlate better with clinical severity of VAP and outcome of patients remains to be established. Friedland and coworkers found that circulating levels of TNF-α, IL-1β, IL-6, and IL-8 poorly correlated with the clinical severity of illness of ICU patients, and only the presence of TNF-α in plasma was an independent predictor of mortality (18). In contrast, relationships between elevated levels of cytokines and both a clinical condition of severe sepsis or septic shock and mortality have been demonstrated repeatedly (1, 19).
The kinetics of BPI, sICAM, and sE-selectin during the development of VAP resemble those of IL-6 and IL-8, supporting the idea that VAP, even when diagnosed with bronchoscopic techniques, is not associated with increasing levels of systemic inflammatory mediators in patients who do not have severe sepsis or septic shock. Furthermore, our data provide new information on the kinetics of systemic inflammatory mediators in a feared nosocomial infection in critically ill patients.
BPI is an endotoxin-neutralizing protein with potent bactericidal activity, providing protection against the toxic effects of LPS (3, 4). BPI is usually not detectable in plasma of healthy volunteers (15), but increased concentrations have been demonstrated in critically ill patients with bacteremia (8). In the present study, BPI was demonstrated in plasma samples of almost all patients developing VAP, both on the day of diagnosis as well as in the days preceding infection, and its detection was independent of a clinical status of severe sepsis or septic shock or the presence of bacteremia. Mortality within 10 d of diagnosing VAP was clearly accompanied with increasing levels of BPI. Our findings suggest that, in case of a serious infection, BPI is released from activated or killed neutrophils. Because intracellular killing of phagocytized bacteria is an important function of BPI, an increased systemic release of BPI may represent a general activation and presence of neutrophils in an effort to control the burden of infection. Moreover, high levels of BPI may reflect overwhelming inflammation, irrespective of an infectious etiology, and are, therefore, associated with mortality. However, more studies are needed to determine the role of systemic BPI release in these conditions. In a previous study we found that the BPI/neutrophil ratio, reflecting neutrophil activation, was associated with the presence of sepsis syndrome and death in bacteremic patients (8). In the present study the kinetics of BPI/neutrophil ratios and systemic BPI levels were comparable (data not shown).
The adhesion molecules ICAM and E-selectin have a critical role in the activation and sequestration of circulating blood neutrophils in microvessels of systemic organs. The interaction between ICAM and neutrophils initiates migration of neutrophils out of the vascular space. Increased levels of circulating sICAM have been demonstrated in adult patients with sepsis, and positive correlations between levels of circulating sICAM and intensity of sepsis and severity of shock and subsequent organ failure have been reported (6). Animal studies of acute lung injury indicated that sE-selectin is present in lung parenchyma and mediates neutrophilic inflammation (5). sE-selectin was detectable in plasma of healthy subjects (9). In patients, elevated levels of sE-selectin were found in bacteremic patients with hypotension (9) and in critically ill patients with microbiologically documented sepsis (10). Similar to BPI, our data demonstrate a tendency toward higher circulating levels of sICAM and sE-selectin when VAP was associated with the presence of severe sepsis or septic shock. Two days after diagnosis of VAP, levels of sE-selectin tended to be higher in patients who died within 10 d, but the association between plasma levels of sICAM and patient outcome is less obvious.
Hypothetically there are several explanations for our observation that VAP is not accompanied with a detectable systemic inflammatory response, unless the patient has a clinical condition of severe sepsis or septic shock.
First, one might question whether these patients really had pneumonia, or just extensive colonization of the lower respiratory tracts. This would imply that the specificity of quantitative cultures of bronchoscopically obtained samples from the distal airways is much lower than generally assumed.
Second, these patients may be immunocompromised because of their critical illness, and, therefore, unable to generate an efficient immunoresponse. Because an effective host defense against bacterial invasion is essential to survive an infection, our findings may provide evidence for the use of adjuvant treatment in these patients (20). Although interest in this field is rising, few clinical data (for example on the use of human granulocyte-colony stimulating factor) are yet available (20).
Third, VAP may be a compartmentalized infection, with an inflammatory response restricted to the lungs. The latter has been suggested to occur in unilateral community-acquired pneumonia (21, 22), and in experimental studies on pneumonia in rats (23). Recently, Fox-Dewhurst and coworkers described the relationships between intrapulmonary and systemic inflammatory responses in rabbits with gram-negative pneumonia (24). They found a dose relationship between the number of bacteria inoculated into the trachea and intrapulmonary and systemic inflammatory reactions. Animals treated with low inocula of bacteria had systemic changes which were comparable to those defined as systemic inflammatory response syndrome, and bacteria were cleared from the lungs. In contrast, those treated with high inocula failed to clear the bacteria from the lungs and developed severe inflammatory responses and septic shock (24). The findings of the present study partly confirm and partly disagree with these experimental data. When VAP is not associated with severe sepsis or septic shock or 10 d mortality, there is hardly any systemic inflammatory response, which may suggest that the threshold of intrapulmonary infection has not been exceeded, and bacteria were, with the help of antibiotics, cleared from the lungs. However, we failed to demonstrate an association between the bacterial inoculum in the lungs during infection (the bacterial burden) and the severity of the systemic inflammatory response, as was demonstrated in rabbits. However, our findings support the idea that VAP is a compartmentalized infection in most patients. Therefore, potential markers for infection would be concentrations of cytokines, inflammatory mediators, or endotoxin in samples of bronchoalveolar lavage or the logarithmic number of bacteria isolated from these samples (“the bronchoscopic bacterial burden”). Kollef and coworkers demonstrated that elevated concentrations of endotoxin in BAL fluid accurately predicted the presence of gram-negative bacterial pneumonia (25). Under the circumstances tested, the bronchoscopic bacterial burden poorly correlated with the clinical severity of infection (e.g., presence of severe sepsis or septic shock, and values of SAPS II scores), and levels of circulating inflammatory mediators. In another study, however, a correlation of 0.585 (p < 0.001) was found between the yield of BAL fluid quantitative cultures for gram-negative bacteria and BAL fluid endotoxin concentrations (25). Whether the bronchoscopic bacterial burden is associated with local cytokine production remains to be established.
| 1. | Bonten M. J. M., Froon A. H. M., Gaillard C. A., Greve J. W., de Leeuw P. W., Drent M., Stobberingh E. E., Buurman W. A.The systemic inflammatory response in the development of ventilator-associated pneumonia. Am. J. Respir. Crit. Care Med.156199711051113 |
| 2. | Le Gall J., Lemeshow S., Saulnier F.A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. J.A.M.A.270199329572963 |
| 3. | Elsbach P., Weiss J.Bactericidal/permeability increasing protein and host defense against Gram-negative bacteria and endotoxin. Curr. Opin. Immunol.51993103107 |
| 4. | Marra M. N., Wilde C. G., Collins M. S., Snable J. L., Thornton M. B., Scott R. W.The role of bactericidal/permeabability-increasing protein as a natural inhibitor of bacterial endotoxin. J. Immunol.1481992532537 |
| 5. | Pilewski J. M., Albelda S. M.Adhesion molecules in the lung: an overview. Am. Rev. Respir. Dis.148(Suppl.)1993S31S37 |
| 6. | Sessler C. N., Windsor A. C., Schwartz M., Watson L., Fisher B. L., Sugerman H. J., Fowler I.Circulating ICAM-1 is increased in septic shock. Am. J. Respir. Crit. Care Med.151199514201427 |
| 7. | Leeuwenberg J. F. M., Smeets E. F., Neefjes J. J., Shaffer M. A., Cinek T., Jeunhomme G. M. A. A., Ahern T. J., Buurman W. A.E-selectin and intercellular adhesion molecule-1 are released by activated human endothelial cells in vitro. Immunol.771992543549 |
| 8. | Froon A. H. M., Dentener M. A., Greve J. W. M., Ramsay G., Buurman W. A.Lipopolysaccharide toxicity-regulating proteins in bacteremia. J. Infect. Dis.171199512501257 |
| 9. | Newman W., Beall L. D., Carson C. W., Hunder G. G., Graben N., Randhawa Z. I., Gopal T. V., Wiener-Kronish J., Matthay M. A.Soluble E-selectin is found in supernatants of activated endothelial cells and is elevated in the serum of patients with septic shock. J. Immunol.1501993644654 |
| 10. | Cummings C. J., Sessler C. N., Beall L. D., Fisher B. J., Best A. M., Fowler A. A.Soluble E-selectin levels in sepsis and critical illness: correlation with infection and hemodynamic dysfunction. Am. J. Respir. Crit. Care Med.1561997431437 |
| 11. | Bonten M. J. M., Gaillard C. A., van Tiel F. H., Smeets H. G. W., van der Geest S., Stobberingh E. E.The stomach is not a source for colonization of the upper respiratory tract and pneumonia in ICU patients. Chest1051994878884 |
| 12. | American College of Chest Physicians and Society of Critical Care MedicineConsensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit. Care Med.201992864874 |
| 13. | Dentener M. A., Smit F. T., Francot G. J. M., Buurman W. A.Characterization of two monoclonal antibodies directed against bactericidal/permeability-increasing protein. J. Infect. Dis.170199414831489 |
| 14. | Bouma, M. G., M. P. Laan, M. A. Dentener, and W. A. Buurman. 1997. Analysis of soluble adhesion molecules. In A. P. Johnstone and M. W. Turner, editors. Immunochemistry 2: A Practical Approach. Oxford University Press, Oxford. 181–196. |
| 15. | Dentener M. A., Francot G. J. M., Smit F. T., Froon A. H. M., Pennings H.-J., Wouters E. F. M., Buurman W. A.Presence of bactericidal/permeability-increasing protein in disease: detection by ELISA. J. Infect. Dis.1711995739743 |
| 16. | Knaus W. A., Wagner D. P., Draper E. A., Zimmerman J. E., Bergner M., Bastos P. G., Sirio C. A., Murphy D. J., Lotring T., Damiano A., Harrell F. E.The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults. Chest100199116191639 |
| 17. | Lemeshow S., Teres D., Klar J., Avrunin J. S., Gehlbach S. H., Rapoport J.Mortality probability models (MPM II) based on an international cohort of intensive care unit patients. J.A.M.A.270199324782486 |
| 18. | Friedland J. S., Porter J. C., Daryanani S., Bland J. M., Screaton N. J., Vesely M. J. J., Griffin G. E., Bennett E. D., Remick D. G.Plasma proinflammatory cytokine concentrations, Acute Physiology and Chronic Health Evaluation (APACHE) III scores and survival in patients in an intensive care unit. Crit. Care Med.24199617751781 |
| 19. | Casey L. C., Balk R. A., Bone R. C.Plasma cytokine and endotoxin levels correlate with survival in patients with the sepsis syndrome. Ann. Intern. Med.1191993771778 |
| 20. | Standiford T. J.Cytokines and pulmonary host defenses. Curr. Opin. Pulm. Med.319978188 |
| 21. | Boutten A., Dehoux M. S., Seta N., Ostinelli J., Venembre P., Crestani B., Dombret M. C., Durand G., Aubier M.Compartmentalized IL-8 and elastase release within the human lung in unilateral pneumonia. Am. J. Respir. Crit. Care Med.1531996336342 |
| 22. | Dehoux M. S., Boutten A., Ostinelli J., Seta N., Dombret M. C., Crestani B., Deschenes M., Trouillet J. L., Aubier M.Compartmentalized cytokine production within the human lung in unilateral pneumonia. Am. J. Respir. Crit. Care Med.1501994710716 |
| 23. | Nelson S., Bagby G. J., Bainton B. G., Wilson L. A., Thompson J. J., Summer W. R.Compartmentalization of intraalveolar and systemic lipopolysaccharide-induced tumor necrosis factor and the pulmonary inflammatory response. J. Infect. Dis.1591989189194 |
| 24. | Fox-Dewhurst R., Alberts M. K., Kajikawa O., Caldwell E., Johnson M. C., Skerrett S. J., Goodman R. B., Ruzinski J. T., Wong V. A., Chi E. Y., Martin T. R.Pulmonary and systemic inflammatory responses in rabbits with Gram-negative pneumonia. Am. J. Respir. Crit. Care Med.155199720302040 |
| 25. | Kollef M., Eisenberg P. R., Ohlendorf M. F., Wick M. R.The accuracy of elevated concentrations of endotoxin in bronchoalveolar lavage fluid for the rapid diagnosis of gram-negative pneumonia. Am. J. Respir. Crit. Care Med.154199610201028 |
