Bacteremia During Ventilated-Associated Pneumonia is Associated with Increased Risk of Death

Abstract & Commentary

By Richard J. Wall, MD, MPH, Pulmonary, Critical Care, & Sleep Disorders Medicine Southlake Clinic, Valley Medical Center, Renton, WA

Dr. Wall reports no financial relationship to this field of study.

Synopsis: This study showed that developing bacteremia during an episode of ventilator-associated pneumonia (VAP) is associated with increased ICU mortality, and that bacteremia is most likely to occur when VAP is due to methicillin-resistant Staphylococcus aureus (MRSA).

Source: Agbaht K, et al. Crit Care Med. 2007;35(9):2064-2070.

This single-center, retrospective study examined ICU mortality and various risk factors among critically ill patients who developed bacteremia during their VAP episode. The investigators were interested in whether the presence of a positive blood culture at the time of VAP diagnosis was associated with excess mortality. Additional objectives were to compare the organisms seen in bacteremic and non-bacteremic VAP and identify the variables associated with bacteremia.

The study included 199 patients from a medical-surgical ICU who had microbiologically confirmed VAP during a 44-month observation period. Although only 35/199 (17.6%) developed bacteremia during their pneumonia episode, these individuals had a higher ICU mortality rate. In a multivariate model adjusting for severity of illness, bacteriology, and other relevant confounders, the only two variables associated with higher ICU mortality were vasopressor use (hazard ratio 2.43, 95% CI 1.23-4.82) and bacteremia (hazard ratio 2.55, 95% CI 1.25-5.23). In a secondary matched case-control analysis, the odds ratio of death for bacteremic cases compared with non-bacteremic controls was 2.86 (95% CI 1.09-7.51).

The most commonly isolated microorganisms in blood cultures were S. aureus (both MRSA and MSSA) and Pseudomonas aeruginosa. Although Haemophilus influenzae was the second most common organism in respiratory samples, it was never isolated from a single blood culture. In multiple logistic regression, the presence of MRSA as the culprit organism and recent prior hospitalization were both independent risk factors for development of bacteremia. Curiously, in a sub-analysis of individuals with VAP due to MRSA, there was no mortality difference between bacteremic and non-bacteremic patients. The authors propose this finding indicates that while MRSA may be a risk factor for bacteremia, ICU mortality is ultimately due to bacteremia not the MRSA itself.


VAP is one of the most commonly acquired infections in the ICU, accounting for >50% of all antibiotic prescriptions in medical ICUs. Understanding the risk factors associated with VAP is important for improving patient care. Reducing VAP is also financially important for hospitals because starting in 2009, the Centers for Medicare & Medicaid Services (CMS) will limit reimbursements for conditions that were not present at admission (eg, infections).1 Regardless of whether clinicians believe VAP is totally preventable, CMS will likely stop paying for VAP treatment in upcoming years.

Overall, this study shows that bacteremia is an independent risk factor for ICU mortality. The study supports the common sense notion that bacteremic VAP is more dangerous for patients than non-bacteremic VAP. However, it is important to remember that the majority of VAP patients never have bacteremia (in this study only 1/5 VAP patients had a positive blood culture). For this reason, clinicians cannot rely on negative blood cultures to exclude VAP. Quantitative cultures of the lower respiratory tract remain the gold standard for diagnosing VAP. Another key finding of this study is that S. aureus causes bacteremia more often than other organisms. The authors provide several theories to explain this interesting finding, but the reasons are still unclear.

This study has several strengths. First, diagnosis and management of VAP were protocolized in this ICU, thereby minimizing biases from variations in clinical care. Second, all cases were microbiologically confirmed using quantitative culture techniques. Unfortunately, the authors did not describe whether they used standardized protocols for ventilator weaning, blood product transfusion, glucose control, and sedation—all of these have been shown to affect VAP rates. Another limitation is that this was a single center study. In addition, "late" VAP episodes (≥10 days after intubation) were more frequent in the bacteremic group and this imbalance may have contributed to the observed mortality differences.

Of note, the Centers for Disease Control and Prevention revised their VAP definitions in May 2007.2 The most important change is that patients no longer need to be on the ventilator for 48 hours to diagnose VAP. The current study used the older definition (with the 48-hour criterion), but it seems unlikely that the change should affect the study's findings. Nonetheless, readers should be aware of this change, because it may affect interpretation of future VAP studies.


  1. Centers for Medicare & Medicaid Services [CMS-1533-FC]. "Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 2008 Rates," page 290ff.
  2. Centers for Disease Control and Prevention. The National healthcare safety network manual: Patient safety component protocol (updated May 24, 2007).