Rapid MRSA PCR on respiratory specimens from ventilated patients with suspected pneumonia: a tool to facilitate antimicrobial stewardship.

Methicillin-resistant Staphylococcus aureus (MRSA) is an important cause of pneumonia in ventilated patients. Our objective was to evaluate the GeneXpert MRSA/SA SSTI Assay (Xpert MRSA/SA) (Cepheid, Sunnyvale, CA) for use in lower respiratory tract (LRT) specimens for rapid MRSA detection and to determine the potentially saved antibiotic-days if a culture-based identification method was replaced by this assay. Remnant LRT samples from ventilated patients submitted to the microbiology laboratory for routine culture were tested using conventional culture and Xpert MRSA/SA. One hundred of 310 LRT specimens met the inclusion criteria. Ten samples were positive for MRSA by Xpert MRSA/SA, while six were positive by routine culture methods. Xpert MRSA/SA correctly identified 5/6 positive and 89/94 negative MRSA specimens, for a sensitivity of 83.3%, specificity of 94.7%, positive predictive value of 45.6%, and negative predictive value of 98.9%. The assay also correctly detected 3/3 positive and 90/97 negative methicillin-susceptible S. aureus (MSSA) specimens, for a sensitivity of 100%, specificity of 92.8%, positive predictive value of 30%, and negative predictive value of 100%. A total of 748 vancomycin and 305 linezolid antibiotic-days were associated with the enrolled specimens. Vancomycin and linezolid utilization could decrease by 68.4% and 83%, respectively, if discontinued 1 day after negative polymerase chain reaction (PCR) results. The Xpert MRSA/SA SSTI rapid MRSA PCR assay performed well in respiratory samples from ventilated patients with suspected pneumonia and has the potential to facilitate stewardship efforts such as reducing empiric vancomycin and linezolid therapy.

The impact of the introduction of a rapid D-dimer assay on the diagnostic evaluation of suspected pulmonary embolism.

BACKGROUND: Rapid D-dimer assays are being used in the diagnostic evaluation of suspected pulmonary embolism (PE). Although this hypothesis is anticipated to decrease the use of ventilation-perfusion (VQ) scans and other diagnostic tests for PE, it has not been tested in a "real-world" environment. SUBJECTS AND METHODS: A randomized prospective trial was conducted on 470 of the 5390 enrolled patients aged 60 years and older who had previously undergone any diagnostic tests for PE at an urban teaching hospital. The use of D-dimer as part of the diagnostic evaluation for PE was promulgated in the 2 randomly chosen intervention firms. The remaining 2 firms served as controls. MAIN OUTCOME MEASURES: The number of VQ scans, spiral computed tomographic scans, and pulmonary angiograms performed. Secondary outcomes included mortality and thromboembolic or bleeding events during 3 months of follow-up. RESULTS: Of the 470 inpatients who underwent evaluation for PE on a per PE workup basis, fewer VQ scans were performed in the intervention firms (63.8% vs 81.3%; P<.01). However, the number of patients evaluated for PE nearly doubled in the intervention firms (304 vs 166; P<.01), so that more VQ scans were performed in the intervention than in the control firms (194 vs 135; P<.01). Ninety-four patients from the control firms and 160 patients from the intervention firms were diagnosed and treated for venous thromboembolic disease (P<.01). There were no differences in secondary outcomes during the 3-month follow-up. CONCLUSIONS: The introduction of a rapid D-dimer assay increased the number of VQ scans performed because the number of patients screened for PE increased. A larger number of patients in the intervention firms were diagnosed as having venous thromboembolic disease (PE and/or deep vein thrombosis). There were no perceived changes in mortality or venous thromboembolic events during the 3-month follow-up.

Noncardiogenic pulmonary edema following upper airway obstruction. 7 cases and a review of the literature.

Pulmonary edema is a relatively common problem facing most physicians. Its separation into cardiogenic and noncardiogenic or high-permeability variants is crucial to its proper early management. Our understanding of the disease processes producing noncardiogenic pulmonary edema has greatly expanded in the last 2 decades. Upper airway obstruction (UAO) is one of many recently recognized mechanisms which can produce noncardiogenic pulmonary edema. The UAO may be subtle in some patients, making its association with the subsequent pulmonary edema difficult especially for the physician unaware of this entity and the potential risk factors contributing to it. A high index of suspicion for this diagnosis is required in the right clinical settings. Our clinical results support a noncardiogenic basis for pulmonary edema occurring after UAO. Five of our 7 patients had at least 1 identifiable risk factor for the development of peri-intubation UAO and pulmonary edema. Additionally, the onset of pulmonary edema following UAO and the duration of the pulmonary edema varied considerably in our patients. Individuals with additional risk factors for the development of noncardiogenic pulmonary edema developed a more severe form of pulmonary edema associated with other organ-system disease. However, in most individuals, UAO-associated pulmonary edema appears to be a self-limited reversible process once it is recognized and properly treated.

Chronic pleural effusion following coronary artery revascularization with the internal mammary artery.

Two patients developed chronic symptomatic pleural effusions following coronary artery revascularization with the IMA. Both patients had evidence of left ventricular dysfunction and pleural effusions which only involved the side corresponding to the harvested IMA.

Fallibility of persistent blood return for confirmation of intravascular catheter placement in patients with hemorrhagic thoracic effusions.

Two patients are described with hemorrhagic thoracic effusions who required central venous catheterization. Presumed subclavian and internal jugular vein cannulation, ipsilateral to the hemorrhagic thoracic effusions, was confirmed by the operators in each case by the persistent aspiration of blood. Subsequent clinical and radiologic evaluation revealed that the vascular catheters were introduced into the pleural space. In both individuals, the persistent aspiration of extravascular hemorrhagic fluid mimicked intravascular catheter positioning. Physicians treating patients with hemorrhagic thoracic effusions need to be aware of this potential complication that can result in the delayed resuscitation of hemodynamically unstable patients.

Labetalol overdose successfully treated with amrinone and alpha-adrenergic receptor agonists.

Circulatory collapse and obtundation occurred in a 37-year-old woman following an iatrogenic overdose of labetalol. Conventional therapy with glucagon and alpha-adrenergic receptor-stimulating agents was ineffective in raising the patient's cardiac output or improving her mental status despite increasing the arterial pressure. The administration of amrinone was temporally associated with significant increases in the cardiac output accompanied by improved mental status. This case suggests that amrinone may be effective adjunctive therapy for beta-adrenergic receptor blocker overdoses by reversing their negative inotropic effects.

Loculated hemopneumothorax of a major fissure in patients with COPD and associated pleural disease.

Two patients with underlying COPD and bullous changes of the lungs were found to have a loculated hemopneumothorax within a major fissure by computed tomography. Both patients had associated pleural disease. In one of the patients, the hemopneumothorax was believed to be potentially compromising pulmonary function. Percutaneous needle drainage of the loculated hemopneumothoraces resulted in large symptomatic pneumothoraces requiring urgent insertion of closed thoracostomy tubes. Loculated hemopneumothorax is an unusual finding within a major fissure mimicking pulmonary abscess, infected bullae, pulmonary cyst, or a cavitary carcinoma of the lung. Physicians treating patients with COPD and associated pleural disease should be aware of this rare condition, since percutaneous needle drainage may be associated with serious life-threatening complications.

The role of selective digestive tract decontamination on mortality and respiratory tract infections. A meta-analysis.

PURPOSE: To review available clinical trials of selective digestive decontamination (SDD) in patients requiring intensive care. DATA SOURCES: All relevant English-language articles from 1982 through 1992 were identified through MEDLINE search and article bibliographies. STUDY SELECTION: Twenty-one articles were identified; 16 articles were selected for analysis based on inclusion and exclusion criteria. DATA EXTRACTION: Occurrence rates for mortality, acquired pneumonia, and acquired tracheobronchitis were extracted for patients treated with SDD and for control patients. Cumulative risk differences were calculated for each of these outcomes. RESULTS: There was no significant difference between cumulative mortality rates for control patients (0.262; n = 1,165) and patients receiving SDD (0.243; n = 1,105) (p = 0.291; beta error rate = 0.16). The acquired pneumonia greater than that in patients receiving SDD (0.074; n = 1,031) (p < 0.0001). The acquired tracheobronchitis rate in control patients (0.117; n = 549) was also significantly greater than that in patients receiving SDD (0.065; n = 494) (p = 0.004). The rate of acquired pneumonia due to Gram-positive bacteria was similar between the control patients (0.033; n = 660) and the SDD-treated patients (0.033; n = 646) (p = 0.933). Colonization with pathogenic Gram-positive bacteria and pneumonia due to antibiotic-resistant Gram-positive bacteria appeared to occur more frequently in SDD-treated patients. CONCLUSIONS: These results suggest that SDD decreases the overall incidence of acquired pneumonia and tracheobronchitis in patients requiring intensive care. SDD had no apparent effect on the hospital mortality rate. The routine use of SDD cannot be supported by this meta-analysis. SDD may be useful in specific circumstances where a particular ICU or ICU population is found to have an excessive incidence of acquired infections. Any use of SDD should include careful patient surveillance for the emergence of infection due to bacteria not covered by the prophylaxis regimen and due to antibiotic-resistant bacteria.

Determinants of outcome for patients in the medical intensive care unit requiring abdominal surgery: a prospective, single-center study.

OBJECTIVE: To identify objective factors, available at the time of surgical evaluation, associated with outcome for patients in the medical ICU undergoing abdominal surgery. DESIGN: Single-center, prospective observational study. SETTING: An academic tertiary care center. PATIENTS: The study included 1,617 consecutive patients in the medical ICU. INTERVENTION: Prospective patient surveillance and data collection. MEASUREMENTS: Patient demographics, severity of illness, organ system derangements, abdominal processes requiring surgery, and hospital mortality. RESULTS: Sixty-seven patients in the medical ICU (4.1 percent) developed an acute abdominal process potentially amenable to surgical intervention. Eleven of these patients (16.4 percent) elected not to undergo surgery (mortality = 100 percent). Forty-two of the 56 patients who underwent surgery survived (75.0 percent). Stepwise logistic regression analysis identified two independent objective predictors of mortality for this patient cohort (p < 0.05): an organ system failure index (OSFI) > 2 (adjusted odds ratio [AOR] = 19.5; 95 percent confidence interval [CI], 7.4 to 51.5; p < 0.001); and an APACHE II score > 18 (AOR = 9.4; CI = 3.1 to 28.3; p = 0.03). The observed mortality following surgery was stratified according to the presence or absence of these two factors: neither present, 5.1 percent; APACHE II > 18 present alone, 33 percent; OSFI > 2 present alone, 60 percent; and both present, 88.9 percent (p < 0.001). Surgical nonsurvivors and patients electing not to undergo surgery were similar without significant differences for demographics, severity of illness, or organ system derangements at the time of surgical evaluation. CONCLUSIONS: The number of organ system derangements and the severity of illness, as assessed by APACHE II, appear to be useful discriminators of outcome for patients in the medical ICU undergoing abdominal surgery. These data suggest potential outcome predictors for this selected group of patients in the ICU.

Reversible cardiogenic shock due to chest tube compression of the right ventricle.

A 62-year-old woman developed shock immediately after the insertion of a right-sided chest tube. A chest roentgenogram showed the chest tube to be overlying the heart and possibly compressing the right ventricle. An animal model was developed to replicate this clinical situation. Using a domestic goat model pulmonary artery, peripheral arterial catheters were inserted along with a right sided chest tube placed to suction. A second chest tube guided by a flexible fiberoptic bronchoscope placed within its lumen was positioned between the right ventricle and the sternum of the animals. Thirteen paired measurements in three goats (average of 4.3 measurements per animal) of cardiac output, heart rate, and mean arterial blood pressure were made at baseline and after chest tube placement over the right ventricle. The data were analyzed using a paired t test statistic. Compared with baseline measurements, there was a significant decrease in cardiac output (p less than 0.001) and mean arterial pressure (p less than 0.001) as well as an increase in heart rate (p = 0.0056) after placement of the chest tube across the right ventricle. We conclude that a misplaced chest tube compressing the right ventricle can impede cardiac output and lead to a low cardiac output state. Physicians inserting chest tubes in patients should be aware of this potential complication as it is easily treated by withdrawal of the chest tube.

Acute pulmonary embolism triggered by the act of defecation.

Pulmonary embolism associated with the act of defecation has not been previously well described. Recently, we reported our experience with four patients who presented to us over a 12-month period with syncope, near syncope, or sudden death following the act of defecation. In all four cases, acute pulmonary embolism was shown to be the etiology of the defecation-associated events. A retrospective chart review of all patients with the diagnosis of pulmonary embolism at our institution over a three-year period yielded five additional patients with the diagnosis of defecation-associated pulmonary embolism. These nine patients accounted for 6.8 percent of all patients with a discharge diagnosis of pulmonary embolism seen at our institution during the three-year study period. Six of the nine patients died from their defecation-associated pulmonary embolism. These six deaths accounted for 25 percent of all deaths from pulmonary embolism seen at our institution during the study period. Based on our experience, we suggest that the act of defecation may trigger the development of acute pulmonary embolism in some patients with deep vein thrombosis.

Intrinsic PEEP and unilateral lung hyperinflation. Pathophysiology and clinical significance.

Over a 14-month period, three patients developed unilateral lung hyperinflation during mechanical ventilation as a manifestation of intrinsic PEEP. These patients all had a history of CAO and an inflammatory or fibrotic disease process involving the contralateral lung. Physicians caring for patients with CAO need to be familiar with this presentation of intrinsic PEEP due to the associated barotrauma and cardiac dysfunction which can result from it.

A rapid qualitative assay to detect circulating endotoxin can predict the development of multiorgan dysfunction.

OBJECTIVE: To determine whether a rapid qualitative assay for the detection of circulating endotoxin (SimpliRED Endotoxin Test [SRE]; AGEN, Inc; Brisbane, Australia) can predict the occurrence of multiorgan dysfunction and hospital mortality. To compare the SRE to the limulus amebocyte lysate (LAL) assay as a predictor of clinical outcomes. DESIGN: Prospective, blinded, single-center study. SETTING: Medical ICU of Barnes-Jewish Hospital, St. Louis, a university-affiliated teaching hospital. PATIENTS: Included in the study were 265 adult patients requiring medical ICU admission. INTERVENTIONS: Daily collection of blood samples. MEASUREMENTS AND RESULTS: Daily detection for the presence of endotoxin in blood during intensive care and assessment for the development of multiorgan dysfunction (ie, an organ system failure index >2) or death. On ICU day 1, 55 (20.8%) patients had circulating endotoxin detected by the SRE. On ICU day 2, 29 of the 143 (20.3%) patients remaining in the ICU had a positive SRE. The development of multiorgan dysfunction was significantly greater among SRE-positive patients (44.8%) compared to SRE-negative patients (21.9%) on ICU day 2 (p=0.013). Multiple logistic regression analysis identified a positive SRE on ICU day 2 (adjusted odds ratio, 4.1; 95% confidence interval, 2.5 to 6.8; p=0.006) as being independently associated with the development of multiorgan dysfunction. A positive SRE test was not predictive of hospital mortality. Direct quantitative measurement of the concentration of circulating endotoxin using the LAL assay was not associated with either the development of multiorgan dysfunction or hospital mortality (p>0.1). CONCLUSIONS: Our preliminary data suggest that a bedside assay to qualitatively detect circulating endotoxin is predictive of the development of multiorgan dysfunction among patients admitted to a medical ICU. The rapid detection of circulating endotoxin could be useful for stratifying patients into various risk categories for the development of multiorgan dysfunction.

Acute respiratory failure due to extramedullary hematopoiesis.

Extramedullary hematopoiesis (EMH) associated with myelofibrosis uncommonly occurs within the thorax. We describe the first reported case of acute and rapidly fatal respiratory failure due to pulmonary interstitial EMH associated with myelofibrosis. Interstitial EMH should be considered in the differential diagnosis of patients with interstitial pulmonary infiltrates and respiratory failure accompanying a disease process known to predispose to the development of EMH.

The influence of mini-BAL cultures on patient outcomes: implications for the antibiotic management of ventilator-associated pneumonia.

STUDY OBJECTIVE: To determine the influence of mini-BAL culture results on subsequent changes in antibiotic therapy and patient outcomes. DESIGN: Prospective, single-center, cohort study. SETTING: Medical ICU of Barnes-Jewish Hospital, St. Louis, a university-affiliated teaching hospital. PATIENTS: One hundred thirty mechanically ventilated patients undergoing mini-BAL for suspected ventilator-associated pneumonia (VAP). INTERVENTIONS: Mini-BAL, prospective patient surveillance, and data collection. MEASUREMENTS AND RESULTS: Sixty (46.2%) patients had mini-BAL cultures that yielded at least one pathogen potentially accounting for the clinically suspected episode of VAP (64 bacterial, 3 viral, 2 fungal). Among the 60 patients with microbiologically positive mini-BAL cultures, 44 (73.3%) were classified as receiving inadequate antibiotic therapy (ie, identification of a microorganism resistant to the prescribed antibiotic regimen). Prior antibiotic administration or its absence remained unchanged in 51 (39.2%) patients based on the mini-BAL culture results, while in another 51 (39.2%) patients, antibiotic therapy was either begun (n=7) or the existing antibiotic regimen was changed (n=44), and in the remaining 28 (21.6%) patients, antibiotic therapy was discontinued altogether. The hospital mortality rates of these three groups were statistically different: 33.3%, 60.8%, and 14.3%, respectively (p<0.001). The most common pattern of antibiotic resistance resulting in an antibiotic change following mini-BAL was the identification of a Gram-negative bacteria resistant to a prescribed third-generation cephalosporin in 23 of 44 (52.3%) patients. Twenty-one of these 23 patients (91.3%) received prior therapy with a cephalosporin class antibiotic during the same hospitalization. Having an immunocompromised state (adjusted odds ratio [OR]=2.45; 95% confidence interval, 1.56 to 3.85; p=0.047) and the presence of a pathogen in the mini-BAL culture resistant to the empirically prescribed antibiotic regimen (adjusted OR=3.28; 95% confidence interval, 2.12 to 5.06; p=0.006) were identified as risk factors independently associated with hospital mortality by logistic regression analysis. CONCLUSIONS: These data suggest that antibiotic selection prior to obtaining the results of lower airway cultures is an important determinant of outcome for patients with suspected VAP. A delay in initiating adequate antibiotic therapy was associated with a greater mortality. Therefore, the initial selection of antibiotics for the empiric treatment of VAP should be broad enough to cover all likely pathogens, including antibiotic-resistant bacteria. This appears to be especially important in patients having received prior antibiotics.

Determinants of mortality and multiorgan dysfunction in cardiac surgery patients requiring prolonged mechanical ventilation.

OBJECTIVES: To identify characteristics associated with mortality and the development of multiorgan dysfunction in patients who had undergone cardiac surgery and required prolonged mechanical ventilation, ie, > 48 h. DESIGN: A prospective cohort study. SETTING: Barnes Hospital, St. Louis, an academic tertiary care center. PATIENTS OR OTHER PARTICIPANTS: A total of 107 consecutive patients undergoing cardiac surgery and requiring prolonged mechanical ventilation. INTERVENTIONS: Prospective patients surveillance and data collection. MAIN OUTCOME MEASURES: ICU mortality and multiorgan dysfunction. RESULTS: Among 472 consecutive patients admitted to the cardiac surgery ICU following surgery, 107 (22.7%) required prolonged mechanical ventilation. Twenty-one of these patients (19.6%) died during their hospitalization. In a logistic-regression analysis, the development of an organ system failure index (OSFI) of 3 or greater was the only characteristic independently associated with ICU mortality (p < 0.001). The occurrence of an antibiotic-resistant infection (adjusted odds ratio [AOR] = 6.1, 95% confidence interval [CI] = 2.5 to 14.6 p = 0.006), an aortic cross-clamp time equal to or greater than 1.25 h (AOR = 3.9, CI = 2.3 to 6.8, p = 0.016), the development of ventilator-associated pneumonia (AOR = 3.6, CI = 2.4 to 5.3, p < 0.001), and an APACHE III score equal to or greater than 30 (AOR = 3.1, CI = 1.8 to 5.3, p = 0.036) were independently associated with the development of an OSFI of 3 or greater. CONCLUSIONS: These data confirm that acquired multiorgan dysfunction is the best predictor of mortality in patients requiring prolonged mechanical ventilation following cardiac surgery. Additionally, they identify potential determinants of multiorgan dysfunction and suggest possible interventions for its reduction in this patient population.

Utility of transbronchial biopsy in patients with acute respiratory failure: a postmortem study.

STUDY OBJECTIVE: To determine the diagnostic yield of histologic specimens obtained by postmortem transbronchial biopsy (TBB) in patients with acute respiratory failure requiring mechanical ventilation. DESIGN: Standard postmortem histologic examination of lung tissue specimens. SETTING: An urban university-affiliated hospital. PATIENTS OR PARTICIPANTS: Thirty patients with diffuse pulmonary infiltrates and acute respiratory failure, who underwent postmortem examination. INTERVENTIONS: Following removal of the lungs from the thorax. TBBs were obtained from the lower lobe of each deflated lung and comparison was made to a 1-cm3 tissue block obtained from the ipsilateral lower lobe. MEASUREMENTS AND RESULTS: Standard postmortem histologic examination provided a specific diagnosis in 85% of the 60 lungs examined, and histologic evidence of acute pneumonia was present in 30% of the lungs. The overall yield of TBB was 48% for establishing a specific histologic diagnosis and 15% for the diagnosis of acute pneumonia. Using standard postmortem histologic examination as the gold standard, the sensitivity and specificity of TBB for making a specific diagnosis were 57% and 100% respectively, with corresponding positive and negative predictive values of 100% and 29%. For the histologic diagnosis of acute pneumonia, the sensitivity of TBB was 50%, the specificity was 100%, and the positive and negative predictive values were 100% and 82%, respectively. The kappa statistic for the agreement between the two diagnostic methods was 0.28 for establishing a specific diagnosis and 0.58 for the diagnosis of acute pneumonia. Obtaining 12 TBBs rather than six TBBs did not increase the diagnostic yield for TBB. CONCLUSIONS: These findings suggest poor overall agreement between standard postmortem histologic examination and TBB specimens. Although not performed in a clinical setting, this postmortem investigation suggests that TBB may be of limited value in mechanically ventilated patients with acute respiratory failure because of its low sensitivity.

Delayed pleuropulmonary complications following coronary artery revascularization with the internal mammary artery.

We have seen four cases of delayed postoperative pleuro-pulmonary complications associated with use of the internal mammary artery (IMA) conduit. In each case the left IMA was used as a bypass conduit to the left anterior descending (LAD) coronary artery. In two of the instances the complications were life-threatening to the patients. Each patient was left with symptomatic residual roentgenographic changes. The IMA is becoming the graft of choice for coronary artery revascularization. The potential for delayed pleuropulmonary complications associated with use of this graft is not well recognized.

The effect of late-onset ventilator-associated pneumonia in determining patient mortality.

STUDY OBJECTIVE: To determine whether the development of late-onset ventilator-associated pneumonia (VAP) is associated with an increased risk of hospital mortality. DESIGN: Prospective cohort study. SETTING: ICUs of two university-affiliated teaching hospitals. PATIENTS: Three hundred fourteen patients admitted to an ICU who required mechanical ventilation for greater than 5 days. INTERVENTIONS: Prospective patient surveillance and data collection. MEASUREMENTS: The primary outcome measures were the development of late-onset VAP (ie, occurring > 96 h after intubation) and hospital mortality. RESULTS: Late-onset VAP was observed in 87 patients (27.7%). Thirty-four (39.1%) patients with late-onset VAP died during hospitalization compared with 85 patients (37.4%) without late-onset VAP (relative risk, 1.04; 95% confidence interval [CI], 0.76 to 1.43). Twenty patients (6.4%) developed late-onset VAP due to a "high-risk" pathogen (ie, Pseudomonas aeruginosa, Acinetobacter sp, Xanthomonas maltophilia) with an associated mortality rate of 65%. Stepwise logistic regression analysis identified five variables as independent risk factors for hospital mortality (p < 0.05): an organ system failure index of 3 or greater (adjusted odds ratio [AOR], 3.4; 95% CI, 2.0 to 5.8; p < 0.001), having a nonsurgical diagnosis (AOR, 2.1; 95% CI, 1.3 to 3.6; p = 0.002), a premorbid lifestyle score of 2 or greater (AOR, 1.8; 95% CI, 1.1 to 2.9; p = 0.015), acquiring late-onset VAP due to a "high-risk" pathogen (AOR, 3.4; 95% CI, 1.2 to 10.0; p = 0.025), and having received antacids or histamine type-2 receptor antagonists (AOR, 1.7; 95% CI, 1.0 to 2.9; p = 0.034). Additionally, we found the occurrence of late-onset VAP due to high-risk pathogens to be the most important predictor of hospital mortality among patients developing VAP (AOR, 5.4; 95% CI, 2.8 to 10.3; p = 0.009). CONCLUSIONS: Nosocomial pneumonia due to certain high-risk microorganisms is an independent risk factor for hospital mortality among patients requiring prolonged mechanical ventilation. We suggest that future investigations of late-onset VAP stratify patient outcomes according to the distribution of high-risk pathogens when reporting their results.