Bacteremia and Ventilator-Associated Pneumonia: A Marker for Contemporaneous Extra-Pulmonic Infection

Abstract

Background:

Ventilator-associated pneumonia (VAP) is a well-known complication of mechanical ventilation in severely injured patients. A subset of patients with VAP develop an associated bacteremia (B-VAP), but the risk factors, microbiology, morbidity, and mortality in this group are not well described. The goal of this study was to examine the incidence, predictors, and outcome of B-VAP in adult trauma patients.

Methods:

We conducted a retrospective review of trauma patients who developed VAP or B-VAP from January 2007 to December 2009 at a single, university-affiliated medical center. Ventilator-associated pneumonia was defined as a clinician-documented instance of VAP together with confirmed positive respiratory cultures (bronchoalveolar lavage [BAL] fluid specimen with ≥10⁴ colony forming units (CFU)/mL or tracheal aspirate with moderate-to-many organisms and polymorphonuclear neutrophils [PMN]). Bacteremia associated with VAP (B-VAP) was defined as the blood culture of an organism that matched the pulmonary pathogen in a case of VAP. We reviewed the demographic data, injury severity, transfusion data, and microbiology of patients who developed VAP and B-VAP. Outcome data included the number of days of care in the intensive care unit (ICU) and hospital length of stay, number of days of mechanical ventilation, and survival. A Student t-test, χ² test, or logistic regression was used as appropriate for data analysis.

Results:

During the 36-mo period of the study, 4,018 adult patients were admitted to the hospital. Ventilator-associated pneumonia was diagnosed in 206 (5%) of these patients, and 26 of these latter patients (13%) had an associated bacteremia. The mean time from admission to the development of VAP was 5 d (95% CI 4.6–5.8). Patients who had B-VAP received significantly more units of red blood cell concentrates (PRBC) than those who did not have B-VAP (23 units vs. 9 units of PRBC, respectively, p<0.05). Patients with B-VAP also had higher rates of simultaneous non-pulmonary infections than those with VAP alone (69% vs. 38%, respectively), a greater number of days of mechanical ventilator support (24 d vs. 14 d, respectively, p<0.05), a greater number of days in the ICU (26 d vs. 17 d, respectively, p<0.05), and a greater hospital length of stay (50 d vs. 30 d, respectively, p<0.05). Patients with B-VAP showed a trend toward lower survival than those without B-VAP, but B-VAP was not an independent predictor of mortality.

Conclusions:

Trauma patients with B-VAP have a similar mortality but greater morbidity than those with VAP alone. The number of PRBC received is the most significant risk factor for developing B-VAP. More than two-thirds of patients with B-VAP have contemporaneous extra-pulmonic infections. Trauma patients with B-VAP may benefit from increased surveillance for additional concomitant infections and from more aggressive empiric antimicrobial coverage.

Ventilator-associated pneumonia (VAP) is the most common serious nosocomial infection in both medical and surgical intensive care units (ICUs), with a reported pooled incidence of 23% among mechanically-ventilated patients [1]. Such pneumonia is more prevalent in trauma patients than in other patient populations; the National Nosocomial Infections Surveillance (NNIS) System reports a mean rate of 15.2 cases of VAP per 1,000 ventilator-days for trauma ICUs, as compared with respective mean rates of 4.9 cases in medical ICUs, 9.3 cases in surgical ICUs, and 11.2 cases in neurosurgical ICUs per 1,000 ventilator-days [2]. Trauma patients are particularly susceptible to pneumonia because they often aspirate at the time of injury or intubation, are subject to prolonged mechanical ventilation, often have acute respiratory distress syndrome (ARDS), and have impaired host defenses secondary to their injuries [3]. Furthermore, VAP is a substantial problem in trauma ICUs, with each episode costing up to $57,000 [4]. Beyond this, VAP is associated with poor patient outcomes, with a reported mortality rate as high as 27% in medical ICUs [5].

Although bacteremia in VAP is uncommon, a subset of patients with VAP becomes bacteremic as a consequence of their pneumonia. In a medical ICU, patients with bacteremic VAP had twice the mortality of those with non-bacteremic VAP [6]. The risk factors, microbiology, morbidity, and mortality for bacteremic VAP in trauma patients have not been described. The goal of the study described here was to examine the incidence, predictors, and outcome of bacteremic VAP in adult patients admitted to a Level I urban trauma center.

Patients and Methods

Study design

A retrospective cohort analysis was performed of all adult trauma patients admitted between January 1, 2007 and December 31, 2009 to The New Jersey Trauma Center at University Hospital in Newark, New Jersey, with a diagnosis of VAP. These patients were subdivided into those with bacteremic VAP (B-VAP) and those with VAP without bacteremia.

Data source

A single investigator (AK) reviewed all of the patients in the study through records abstracted from the hospital trauma registry. Culture data and subsequent antibiotic data were obtained from the hospital's electronic database (Epic Systems, Verona, WI). Culture-specific data and antibiotic information were collected and reviewed by study personnel. Patient-specific demographic data were collected from the hospital trauma registry. Potential predictors of bacteremic VAP were reviewed, including injury severity, transfusion data, and mechanism of injury (blunt versus penetrating). Outcome data were also obtained, including number of ventilator days, length of stay in the ICU, hospital length of stay, and mortality.

Definitions

For the purposes of the study, VAP was defined as clinician-documented VAP together with confirmed positive respiratory cultures. Bronchoalveolar lavage (BAL) was the most common method of diagnosis (BAL fluid with ≥10⁴ CFU/mL). The basis for investigating the possibility of BAL at our institution is clinical suspicion of pneumonia in accordance with guidelines of the U.S. Centers for Disease Control and Prevention specifying: Fever or leukocytosis together with a change in the character of sputum and worsening gas exchange. A tracheal aspirate revealing moderate-to-many organisms together with polymorphonuclear leukocytes was also considered a positive culture. Bacteremia associated with VAP (B-VAP) was defined by an organism cultured from blood that matched the pathogen in the patient's lung. Patients were excluded from the study if they had no documented pneumonia during their initial hospital admission, culture data were lacking, or the patient was less than 18 y of age. Bacteremia was confirmed through isolation of the same organism in both of two blood cultures drawn from separate sites within 7 d of a diagnosis of VAP. Skin pathogens identified in only one of two blood cultures were deemed likely contaminants and were not considered to constitute bacteremia.

The date of diagnosis of VAP was based on clinician documentation together with supporting data from respiratory culture. The respiratory pathogen(s) in an initial instance of VAP were identified and sensitivity data for these pathogens were collected and recorded. Additional culture data collected within 7 d of the diagnosis of VAP were also reviewed. Recurrent VAP was defined as positive respiratory cultures identified ≥7 d after the initial diagnosis and initiation of treatment for VAP.

All charts of patients with VAP were reviewed to ascertain whether or not they had additional, extra-pulmonic sources of infection including urinary tract infection (UTI), wound or surgical site infection (SSI), intra-abdominal infection, infection of the central nervous system, and venous catheter-related infection. Urinary tract infections were diagnosed through urinalysis on the basis of leukocyte esterase or nitrite concentration accompanied by a finding of ≥10⁵/mL organisms on urine culture. Wound, surgical site infections and intra-abdominal infections were diagnosed through swab cultures without the use of quantitative data. Central nervous system infections were diagnosed through the finding of organisms cultured from cerebrospinal fluid together with fluid analyses suggesting possible infection. Venous catheter-related infections were diagnosed on the basis of a finding of >15 colony-forming units (CFU) plated from a removed central venous catheter. The site and causative organism were recorded in each of the foregoing infections.

For all patients with VAP, the antibiotic agent used for treatment, the timing of its dosing, and the duration of treatment were evaluated. The total duration of antibiotic treatment was recorded. We determined whether or not the patient received a full course of antimicrobial therapy of ≥7 d without interruption of therapy. Interruption was defined as a period of >24 h during which no antibiotic was given. We examined whether or not antibiotic treatment was begun on the day on which VAP was diagnosed (the day on which culture results were reported); antibiotic timing was considered inappropriate if the antibiotic was not initiated within 24 h of the diagnosis of VAP. We examined resistance patterns of the pathogens cultured and determined whether or not they were sensitive to the antimicrobial agent selected for empiric coverage. Additionally, we determined whether or not antibiotic dosages were reduced when sensitivity data became available, and whether or not the de-escalation of antimicrobial therapy on the basis of final culture results was associated with the development of B-VAP.

Data analysis

A Student t-test, χ² test, or logistic regression was used as appropriate. A value of p<0.05 was considered statistically significant.

Results

A total of 4,018 adult patients were admitted from January 2007 to December 2009. Of these, 1,014 were admitted to the trauma ICU, and 716 (71%) of these latter patients were ventilated mechanically. A diagnosis of VAP was made in 206 (29%) of the mechanically ventilated patients. These patients ranged in age from 18–87 y (mean age 45 y) and included 162 (79%) men and 44 (21%) women. One-hundred sixty-eight (82%) of the patients had sustained blunt injuries. The mean Injury Severity Score (ISS) for the study population was 27 points (95% CI 25–28) and the mean time to diagnosis of initial VAP was 5 d (95% CI 4.6–5.8). Of the 206 patients with a diagnosis of VAP, 26 (13%) had an associated bacteremia.

Forty different species of pathogens were involved in the initial episode of VAP (Table 1). Initial episodes of VAP caused by gram-negative pathogens were twice as common as those caused by gram-positive pathogens (109 vs. 49; 53% vs. 24%, respectively), and 47 (23%) cases of VAP were mixed gram-negative/gram-positive infections. The most common organism isolated from respiratory cultures in cases of initial pneumonia was Staphylococcus aureus (Table 1). Of the pathogens isolated in these infections, more than one-third consistsed of methicillin-resistant strains of S. aureus. The most common gram-negative organisms isolated were, in order of prevalence, Haemophilus influenzae, Enterobacter species, and Klebsiella species.

Table 1.

Organisms Isolated in Respiratory Cultures in Cases of Initial Pneumonia

Isolate			n (%)
Gram-positive
Enterococcus faecalis			5 (1.5)
Staphylococcus spp.			63 (19.0)
	MSSA	40
	MRSA	22
	Coagulase-negative Staphylococcus	1
Streptococcus spp.			47 (14.0)
	Streptococcus pneumoniae	18
	Viridans Streptococcus	17
	Beta-hemolytic Streptococcus	5
	Alpha-hemolytic Streptococcus	3
	Group C Streptococcus	1
	Streptococcus anginosus	1
	Streptococcus capitus	1
	Streptococcus salvarius	1
Peptostreptococcus spp.			2 (0.6)
Gram-negative
Acinetobacter baumannii			13 (4.0)
Alcaligenes xylosoxidans			2 (0.6)
Bacteroides spp.			1 (0.3)
Burkholderia cepacia			1 (0.3)
Capnocytophaga spp.			2 (0.6)
Citrobacter spp.			4 (1.2)
	Citrobacter freundii	2
	Citrobacter koseri	2
Corynebacterium spp.			2 (0.6)
Diptheroids			1 (0.3)
Eikenella spp.			2 (0.6)
Enterobacter spp.			38 (11.0)
	Enterobacter aerogenes	20
	Enterobacter cloacae	15
	Enterobacter gergoviae	3
Escherichia coli			21 (6.3)
Haemophilus spp.			42 (13.0)
	Haemophilus influenzae	41
	Haemophilus parahaemolyticus	1
Kingella kingae			2 (0.6)
Klebsiella spp.			35 (11.0)
	Klebsiella pneumoniae	34
	Klebsiella oxytoca	1
Kluyvera ascorbata			1 (0.3)
Moraxella catarrhalis			1 (0.3)
Neisseria spp.			7 (2)
Prevotella melaninogencia			1 (0.3)
Proteus mirabilus			7 (2)
Pseudomonas aeruginosa			23 (7)
Serratia marcescens			10 (3)
Stenotrophomonas maltophilia			1 (0.3)

The total number of isolates (n) exceeds 206 because many patients had mixed infections.

MRSA=methicillin-resistant Staphylococus aureus; MSSA=methicillin-sensitive Staphylococcus aureus.

Bacteremic VAP was more common in patients with VAP caused by gram-negative than by gram-positive organisms (16% vs. 10%, respectively). The two most common organisms isolated from blood cultures of bacteremic patients were S. aureus (23%) and Klebsiella pneumoniae (23%) (Table 2). Enterobacter species were grown in approximately 15% of positive blood cultures, and organisms of the family Enterobacteriaciae, inclusive of Klebsiella, were grown in 10 (38%) of the 26 positive blood cultures.

Table 2.

Organisms Isolated in Blood Cultures of Patients with Bacteremic Ventilator-Associated Pneumonia

Isolate	n (%)
Gram-positive
Methicillin-resistant Staphylococcus aureus	3 (12)
Methicillin-sensitive Staphylococcus aureus	3 (12)
Gram-negative
Klebsiella pneumoniae	6 (23)
Enterobacter spp.	4 (15)
Escherichia coli	3 (12)
Pseudomonas spp.	2 (8)
Proteus mirabilus	2 (8)
Acinetobacter baumannii	1 (4)
Burkholderia cepacia	1 (4)
Serratia marcescens	1 (4)

Several factors were analyzed to determine their association with the development of B-VAP (Table 3). Age, ISS, gender, and mechanism of injury were not associated with the development of B-VAP. Mean systolic blood pressure at the time of admission, and severe traumatic brain injury, were also not associated with the development of B-VAP. There was no statistically significant difference in the number of cases of recurrent VAP in the patient groups with VAP alone and B-VAP (data not shown), but patients who had B-VAP were given significantly more units of PRBC over the course of their hospital stay (Table 3). For patients who received large volumes of PRBC via transfusion, most of the PRBC were transfused in the first 48 h after hospital admission.

Table 3.

Patient Factors Associated with Ventilator-Associated Pneumonia and Bacteremic Ventilator-Associated Pneumonia

	Bacteremic VAP (n=26)	Nonbacteremic VAP (n=180)	p value
Mean age (95% CI)	46 (37–54)	45 (42–48)	NS
Number of males (%)	19 (73%)	143 (79%)	NS
Mean Injury Severity Score (95% CI)	27 (22–34)	27 (25–28)	NS
Penetrating injury (%)	7 (27%)	31 (17%)	NS
Admission mean systolic blood pressure (95% CI)	125 (108–142)	143 (137–148)	NS
Severe traumatic brain injury: Number with head AIS≥4 (%)	3 (12%)	53 (29%)	NS
Mean number of units of PRBC transfused (95% CI)	23 (13–34)	9 (7–11)	<0.05

AIS=Abbreviated Injury Scale; CI=confidence iknterval; PRBC=red blood cell concentrates; VAP=ventilator-associated pneumonia.

We also found that patients with B-VAP had higher rates of simultaneous non-pulmonary infections than did those with VAP alone (69% vs. 38%, respectively, p=0.005). Of the 26 patients with B-VAP, 18 had extra-pulmonic sources of infection, consisting of UTI in nine of these patients, intra-abdominal infections in six patients, and SSIs in three patients. Of these extra-pulmonic infections, 10 (56%) were caused by pathogens distinct from those isolated in the patients' respiratory and blood cultures.

Antibiotic factors were analyzed for their contribution to the development of B-VAP. The only significant factor associated with the development of bacteremia was initiation of an incorrect empiric antibiotic agent (Table 4). Thirty-two percent of patients in whom the pathogen responsible for B-VAP was resistant to the selected empiric agent developed bacteremia, as compared with only 11% of those treated with appropriate antimicrobial agents (p=0.04).

Table 4.

Evaluation of Contribution of Antibiotic Therapy to Development of Bacteremic Ventilator-Associated Pneumonia

	Rate of B-VAP	p value
Antibiotic starting time
Appropriate (<24 h after diagnosis)	13%	NS
Inappropriate (>24 h after diagnosis)	12%
Antibiotic agent (based on culture results)
Appropriate (pathogen sensitive)	11%	p<0.05
Inappropriate (pathogen resistant)	32%
Duration
Appropriate (≥7 d)	13%	NS
Inappropriate (<7 d)	12%
De-escalation (based on culture results)
Yes	7%	NS
No	17%

B-VAP=bacteremic ventilator-associated pneumonia; NS=not significant.

Outcomes of patients with B-VAP were significantly worse than those of patients with VAP alone (Table 5). Patients with B-VAP had more days of mechanical ventilator support (24 d vs. 14 d, respectively, p<0.05), a greater number of days in the ICU (26 d vs. 17 d, respectively, p<0.05), and a greater hospital length of stay (50 d vs. 30 d, respectively, p<0.05). Patients with B-VAP showed a trend toward having a lower rate of survival than those with VAP alone (23% vs. 8% mortality, respectively), but B-VAP was not an independent predictor of mortality.

Table 5.

Outcomes in Patients with Ventilator-Associated Pneumonia and Bacteremic Ventilator-Associated Pneumonia

	Bacteremic VAP (n=26)	Nonbacteremic VAP (n=180)	p value
Ventilator days (95% CI)	24 (17–30)	14 (13–15)	p<0.05
ICU days (95% CI)	26 (19–33)	17 (15–18)	p<0.05
Hospital days (95% CI)	50 (33–66)	30 (27–33)	p<0.05
Mortality (%)	6 (23%)	15 (8%)	p=0.06

CI=confidence interval; ICU=intensive care unit; VAP=ventilator-associated pneumonia.

Discussion

Few studies have evaluated the effect of bacteremia in trauma patients with VAP. Ventilator-associated pneumonia presents a profound problem in trauma ICUs owing to high associated rates of morbidity and mortality as well as to the significant cost associated with each episode of VAP [1 –5]. Given that B-VAP is associated with a higher mortality than is VAP alone [6], understanding the predictors of B-VAP may be beneficial.

We found the incidence of VAP in mechanically ventilated trauma patients to be 29% and the incidence of B-VAP in this population to be 13%, both of which percentages are similar to those reported in a systematic review of 89 studies of mechanically ventilated patients, which described a pooled incidence of VAP of 23% (95% CI 18.8–26.9%) [1]. These percentages of VAP are at the upper limits of what has been reported by other institutions, probably owing in part to our diagnostic threshold of 10⁴ CFU/mL in bronchoalveolar lavage specimens as well as to the higher than average percentage of trauma patients in our study-patient population. The rate of B-VAP in our study was comparable to the incidence reported by Agbaht et al. [6] and Magret et al. [7], who reported rates of B-VAP of 17.5% and 14.6%, respectively, in populations consisting of both medical and surgical patients.

Most of the cases of B-VAP diagnosed in this study were caused by S. aureus or K. pneumoniae, with each of these organisms being responsible for approximately one quarter of the bacteremic infections in our patient population. This differs from the bacteriology of the pneumonias reported by Magret et al. in a population of combined medical and surgical patients in 27 ICUs throughout Europe [7], in which Acinetobacter spp. and methicillin-resistant S. aureus (MRSA) were the predominant pathogens. In our study population B-VAP was caused largely by members of the family Enterobacteriaceae, inclusive of Klebsiella, constituting a distribution of infections that varies widely from what has been reported in the United States and Europe [7]. Increased frequencies of pneumonia caused by Klebsiella have been described in Asian populations, with a reported association with alcoholism [8]. The pathogenic differences in these and Western populations may be related to Western studies' reports of the development of B-VAP largely as a result of hospital-acquired pneumonias, and therefore as being caused by hospital-acquired pathogens. Conversely, critical illness and bacteremia in the patients described by Ko et al. [8] who presented with community-acquired pneumonia were the specific consequences of community acquired Klebsiella pneumoniae infections. Ko et al. [8] reported mortality rates ranging from 12%–60% in these bacteremic Klebsiella infections. Of the bacteremic Klebsiella pneumonias described in the present study, one-half originated as pneumonias diagnosed in the first 4 d after hospital admission, suggesting that the pathogen may have been community-acquired. Diabetes, human immunodeficiency virus (HIV) infection, and underlying liver disease have also been shown to predispose to bacteremic Klebsiella infection [8]. The world-wide, prospective, and multi-institutional study of Ko et al. suggested that K. pneumoniae is no longer a source of serious or important community-acquired infection in the United States, in documenting only four cases of community-acquired bacteremic Klebsiella pneumonia over a period of 2 y in nine large hospitals in the United States. Our data for trauma patients challenge this conclusion.

When examining risk factors for B-VAP, we found the transfusion of PRBC to be the most significant such factor. Blood stream infections are more common in patients with high transfusion requirements, and the most common sources of blood stream infections are pulmonary pathogens [10]. Other studies have also described an association between blood transfusion and bacteremia. Shorr et al. [11] studied ICU-acquired blood stream infections in 284 adult ICUs and found that transfusion more than doubled the risk of such infection (OR 2.23; 95% CI 1.43–3.52; p<0.001). Transfusion is also associated with a predisposition to other hospital-acquired infections, including pneumonia. Bochicchio et al. [12] demonstrated that blood product transfusion is an independent risk factor for the development of VAP in trauma patients.

Besides showing an association with transfusion, the development of VAP was associated with inappropriate empiric antibiotic therapy. We found that neither the timing nor the duration or the de-escalation of antibiotic therapy increased the risk of bacteremia. However, an inappropriate choice of initial antimicrobial coverage, did seem to play a pivotal role in this. Of the patients in our study-patient population, 32% became bacteremic, in whom the pathogen responsible for VAP was resistant to the chosen empiric agent as compared with only 11% of those treated with appropriate initial antimicrobial agents (p<0.05). Luna et al. [13] showed that among patients with VAP who received inappropriate early antibiotic coverage, mortality was more than twice that of those who received appropriate empiric coverage. In a later study, Luna et al. [14] demonstrated further that in patients with B-VAP, inadequate initial antimicrobial therapy was an independent predictor of mortality.

Interestingly, more than two-thirds of our patients with B-VAP had additional, extra-pulmonic infections. Although a few (5%) contemporaneous liver abscesses or CNS infections were reported by Ko et al. [8], our finding of a high rate of concomitant extra-pulmonic infection in trauma patients with B-VAP has not been described elsewhere. The reason for this is unclear. We speculate that trauma patients with B-VAP are immunosuppressed. Trauma patients are well known to be profoundly immunosuppressed; this has been attributed to trauma-induced bone marrow suppression, specifically demonstrated in trauma patients who have experienced major hemorrhage [4,15]. This contributes to increased susceptibility to infection. Furthermore, the immunosuppressive effects of blood transfusion have been well documented in the surgical literature since the mid-1960s [16]. This transfusion-induced immunosuppression results in significantly increased rates of infection in transfused trauma patients [17]. The high incidence of concurrent extra-pulmonic infection in B-VAP patients is therefore most likely a manifestation of the immunocompromised state of the massively transfused trauma patient. It is noteworthy that the causative organism of 56% of the extra-pulmonic infections in our study did not match the pathogen found in the patient's respiratory and blood cultures. The most common additional infection in our patients with VAP was UTI, followed closely by intra-abdominal abscesses and skin/soft tissue infections. The clinical implication of this is that even if one tailors antimicrobial coverage to treat a bacteremic pneumonia, another source of infection may be present and may not be treated adequately with the chosen antibiotic. Thus, in this setting of B-VAP, increased surveillance for additional sources of infection is indicated.

The mortality rates in our study were considerably lower for both VAP and B-VAP than those reported in previous studies. The mortality rate among our patients with diagnosed non-bacteremic VAP was 8%, and the rate for those who developed B-VAP was 23% (p-NS); however, lack of significance may represent a type II error. Agbaht et al. [6], in a retrospective reviewe of 199 cases of VAP, found that patients with positive blood cultures had a risk of death of 2.9 relative to that of non-bacteremic controls (95% CI 1.09,7.51). Magret et al. [7] also found greater mortality among patients with B-VAP (57%) than among those with VAP (33%).

Our population of patients differed from that in the the multi-center European study (EU-VAP study group) in consisting exclusively of trauma patients. The lower mortality among our patients may be attributed to a lower mean age. In a sub-analysis of trauma patients in the EU-VAP study group, Magret et al. [18] found that the subpopulation of patients with trauma was younger than the overall population (45.3±19.4 y vs. 61.1±16.7 y, p<0.0001), and that episodes of VAP in trauma patients were associated with a lower mortality than those in non-trauma patients (17.2% vs. 42.6% p<0.001). The mean age of our patients was 45 y, as compared with 57 y in the large EU-VAP study, and our mortality rates are more closely similar to those in the EU-VAP trauma subgroup. Although co-morbid factors were not well documented in our patients, one would suspect that younger patients would have fewer co-morbidities, perhaps explaining the lower overall mortality rates.

Not surprisingly, patients with B-VAP required more days of mechanical ventilation and had a longer ICU length of stay than those with VAP. The patient population with B-VAP may be a sicker patient population, with systemic infection and thus systemic inflammation, with the possible sequelae of septic shock and organ compromise. Our outcome data for ICU length of stay are comparable to those in the EU-VAP group (26 d vs. 29 d, respectively, for patients with B-VAP, and 17 d vs. 21 d, respectively, for patients with VAP alone).

Our study has several limitations. First, it is limited by its retrospective nature. Lack of prospective clinical data makes it difficult to determine exactly when each of the reported infections began, as the microbiologic data reflect only the dates on which confirmatory cultures were reported. Further, it is uncertain whether all additional infections were documented; this would potentially serve to make the incidence of concomitant infection even higher. The study is also limited by the relatively small number of bacteremic pneumonias that occurred during the study period, which may have specifically contributed to the failure to find a statistically significant difference in mortality in B-VAP vs. VAP. Lastly, lack of a standardized and uniform diagnostic and treatment algorithm for VAP in the trauma ICU contributed to a high variability in empiric antibiotic coverage and dosing.

Ventilator-associated pneumonia remains a frequent source of infection in our mechanically ventilated trauma patients, and B-VAP occurs in 13% of these patients. Transfusion is associated with and may contribute to the development of B-VAP. More than two-thirds of patients with bacteremic VAP have concurrent, other sources of infection, which are often attributed to pathogens distinct from those found in the blood, and this may reflect patients' immunocompromised states. Therefore, in trauma patients, an increased awareness of the likelihood of existence of multiple sources of infection, with appropriately increased surveillance and more aggressive empiric antimicrobial therapy, may be appropriate. Further prospective studies of B-VAP in trauma patients are warranted to ascertain whether more aggressive antimicrobial therapy may improve these patients' outcomes.

Author Disclosure Statement

No competing financial interests exits.

References

Safdar

, Dezfulian

, Collard

, et al. Clinical and economic consequences of ventilator-associated pneumonia: A systematic review. Crit Care Med, 2005; 33:2184–2193.

National Nosocomial Infections Surveillance System. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control, 2004; 32:470–485.

Barie

, Eachempati

, Pieracci

. Infection. In: Feliciano

, Mattox

, Moore

, eds. New York. McGraw-Hill, 2008:363–392.

Cocanour

, Ostrosky-Zeichner

, Peninger

, et al. Cost of a ventilator-associated pneumonia in a shock trauma intensive care unit. Surg Infect, 2005; 6:65–72.

Fagon

, Chastre

, Hance

, et al. Nosocomial pneumonia in ventilated patients: A cohort study evaluating attributable mortality and hospital stay. Am J Med, 1993; 94:281–288.

Agbaht

, Diaz

, Munoz

, et al. Bacteremia in patients with venitlator-associated pneumonia is associated with increased mortality: A study comparing bacteremic vs. nonbacteremic ventilator-associated pneumonia. Crit Care Med, 2007; 35:2064–2070.

Magret

, Lisboa

, Ignacio

, et al. Bacteremia is an independent risk factor for mortality in nosocomial pneumonia: A prospective and obervational multicenter study. Crit Care, 2011; 15:R62.

, Paterson

, Sagnimeni

, et al. Community-acquired Klebsiella pneumoniae bacteremia: Global differences in clinical patterns. Emerg Infect Dis, 2002; 8:160–166.

Lin

, Jeng

. Chen

, Fung

. Bacteremic community acquired Klebsiella pneumoniae: Clinical and microbiological characteristics in Taiwan 2001–2008. BMC Infect Dis, 2010; 10:307.

10.

Sifri

, Mohr

, Horng

, et al.

Do bloodstream infections affect outcomes In trauma patients?

Poster presentation at the Surgical Infection Society 30th Annual Meeting, Las Vegas, NV. April 17–20, 2010.

11.

Shorr

, Jackson

, Kelly

, et al. Transfusion practice and blood stream infections in critically ill patients. Chest, 2005; 127:1722–1728.

12.

Bochicchio

, Napolitano

, Joshi

, et al. Blood product transfusion and ventilator-associated pneumonia in trauma patients. Surg Infect, 2008; 9:415–422.

13.

Luna

, Vujacich

, Neiderman

, et al. Impact of BAL data on the therapy and outcome of ventilator-associated pneumonia. Chest, 1997; 111:676–685.

14.

Luna

, Videla

, Mattera

, et al. Blood cultures have limited value in predicting severity of illness and as a diagnostic tool in ventilator-associated pneumonia. Chest, 1999; 116:1075–1084.

15.

, Livingston

, Hauser

, et al. Trauma inhibits erythroid burst-forming unit and granulocyte-monocyte colony-forming unit growth through the production of TGF-β1 by bone marrow stroma. Ann Surg, 2001; 2:224–232.

16.

Brunson

, Alexander

. Mechanisms of transfusion-induced immunosuppression. Transfusion, 1990; 30:651–658.

17.

Agarwal

, Murphy

, Cayten

, et al. Blood transfusion increases the risk of infection after trauma. Arch Surg, 1993; 128:171–177.

18.

Magret

, Amaya-Villar

, Garnacho

, et al. Ventilator-associated pneumonia in trauma patients is associated with lower mortality: Results from EU-VAP study. J Trauma, 2010; 69:849–854.