Vol. 13 •Issue 5 • Page 24
No Easy Answers to Diagnosing Ventilator-associated Pneumonia
Most intensive care units have more than one-third of their patients on mechanical ventilation every day.1 On average, these patients have a 1 percent chance per day of acquiring ventilator-associated pneumonia. Although the risk seems slight, the unfortunate few who develop VAP face high mortality rates of 40 percent to 67 percent.2,3
Whether VAP affects survival in the ICU has been the topic of controversy. Several studies that examined overall survival rate have failed to show an independent effect of pneumonia on mortality for patients with VAP. However, these studies didn’t account for the effect of treatment of the pathogen on the outcome for the VAP episode.
Using bronchoalveolar lavage (BAL) to assess possible VAP, there are three outcomes of these diagnostic procedures in patients with possible VAP: no organism recovered (no growth), recovery of an organism sensitive to current antibiotics (adequate therapy), and recovery of an organism resistant to current antibiotics (inadequate therapy).
Studies have demonstrated significantly higher mortality if the patient had received inadequate therapy for whatever organism was eventually recovered.4,5 The rate of mortality for those given inadequate therapy was higher than either acute respiratory distress syndrome or sepsis patients. These studies stress the importance of appropriate initial therapy.
Therefore, diagnosis of pneumonia is an essential part of management of VAP.6 However, there’s no perfect test for pneumonia. This lack of a gold standard has led to some debate about the relative merit of various techniques. Although these debates are useful to spur investigation into this area, clinicians still need to know how to diagnose pneumonia.
CLINICAL PULMONARY INFECTION SCORE
Several standard features are used as part of the daily evaluation to detect pneumonia in the ventilated patient. These include changes in the patient’s vital signs such as increased heart rate, temperature and respiratory rate. Heart rate and temperature may be compromised by the underlying condition or therapy of the patient. For the ventilated patient, changes in respiratory rate may be difficult to interpret.
On the other hand, serial observations of oxygenation usually are known for the ventilated patient. A worsening of oxygenation may be a clue for pneumonia. This is assessed by calculating the ratio of arterial oxygenation versus inspired oxygen concentration (PaO2/FIO2).
The most prominent physical finding of pneumonia is coarse crackles. These are best appreciated at the posterior aspect of the sitting patient. This is a difficult examination in the ICU.
Another important feature of pneumonia is an increase in quantity and purulence of secretions. Although mechanical ventilation can dry the secretions, most patients with pneumonia will start having increased secretions.
In the patient with increased secretions and worsening oxygenation, the next step in evaluation is the chest roentgenogram. Although chest X-ray is a standard part of the assessment of patients for possible VAP, there are multiple causes for pulmonary infiltrates. These include congestive heart failure, pulmonary emboli and even bland atelectasis. Perhaps the most difficult area is the patient with ARDS who develops a fever. In that situation, one has to determine whether there’s new infiltrate in addition to the diffuse infiltrate already present.
In a study of chest roentgenogram compared to subsequent autopsy confirmation of pneumonia, researchers looked at features associated with pneumonia.7 A new or worsening infiltrate was found in more than 80 percent of patients with pneumonia. However, of those patients with new or worsening infiltrate, only 25 percent had pneumonia.
The presence of an air bronchogram was a better predictor, but less sensitive. That study found more than 10 percent of patients with pneumonia at autopsy had no roentgenographic evidence of pneumonia within 24 hours of their death.
Given the lack of a single feature to diagnose pneumonia, researchers proposed a cumulative score (CPIS). In the original series, a CPIS of greater than 6 was associated with pneumonia.8 (See Table 1.)
Several others have used the CPIS with various levels of success. One difficulty with the original CPIS score was the need for culture results, which aren’t available for 48 to 72 hours. While this enhanced the diagnostic utility, it makes it less useful for a daily evaluation.
Researchers modified the CPIS and found the score would rise prior to development of pneumonia.9 The study also showed that patients who died within 28 days continued to have a high CPIS score, while those who survived had an improvement in their CPIS within three days of starting antibiotics.
CULTURE OF SAMPLES
The clinical information can be enhanced by the results of culturing of lower respiratory secretions. Several authors have found that the CPIS score increases specificity by adding the results of deep respiratory cultures.
The ventilated patient offers access to the lower airways, bypassing contamination of the mouth and pharynx. However, colonization of the tracheal secretions is common for the ventilated patient. The difficulty for the clinician is to separate the bacteria which is colonizing from that which is causing infection.
For some bacteria, one can assume that presence of the organism is the cause of pneumonia. For example, a patient who has M. tuberculosis in his sputum has pneumonia.
On the other hand, S. pneumoniae can be in cultures from healthy controls. However, a patient with clinical features of pneumonia who has S. pneumoniae identified would be assumed to have that as the cause of his pneumonia.
It would be unusual to recover P. aeruginosa in a healthy individual, although colonization with P. aeruginosa commonly occurs in a ventilated patient. Thus, semiquantitative cultures have proven useful to distinguish between colonization and infection of the ventilated patient.
For some organisms, such as C. albicans and diptheroids, the presence of even large numbers of organisms doesn’t necessarily mean pneumonia. These organisms can cause locally high concentrations without causing clinical pneumonia.
As clinicians have tried to clarify the significance of bacteria in lower respiratory samples, it’s become clear that semi-quantitative cultures can be useful. Current recommendations are that greater than 10,000 colony forming units (cfu)/mL of BAL fluid should be considered pneumonia. (See Table 1.) Based on the volume of sample tested, different cutoffs have been established for protected brush specimen (greater than 1,000 cfu/mL) and endotracheal aspirate (greater than 1,000,000 cfu/mL).
The use of bronchoscopic-directed protected specimen brush or BAL requires a physician trained in bronchoscopy. A simple endotracheal aspirate doesn’t require any additional skills than those already known by the caregivers in the ICU. In between the two tests is the nonbronchoscopic BAL procedure. The nonbronchoscopic BAL can be performed by respiratory therapists or other trained personnel. (See Table 2.)
SERIAL INFORMATION
The diagnosis of pneumonia shouldn’t be a static event. The initial information should be used to start treatment. However, clinicians always should reassess the patient after two to three days when further material has become available.
Among the additional material is the patient’s clinical status. The patient with an unchanged or rising CPIS isn’t responding to treatment. For this patient, clinicians must be sure that cultures from the original event don’t identify a resistant pathogen. In that situation, modification of antibiotics may be necessary.
For the patient with no resistant pathogen, clinicians should look for alternative causes for the patient’s problem. This would include infection elsewhere or another medical problem such as a pulmonary embolism.
Serial bacterial cultures of the lung can be useful. Look for superinfection. Also, the presence of persistent bacteria may be associated with increased mortality. In a study looking at serial BAL during therapy, there was a significant increase in mortality for those patients who failed to clear their pathogens in nonbronchoscopic BAL.10 This study needs verification by a larger trial. It’s supported by the observation that clearance of organisms in bacteremia and urinary tract infections is associated with a better prognosis.
For the patient who’s doing well, cutting back or even stopping antibiotic therapy is a reasonable approach. Improvement of the clinical status and reduction of the bacterial load in the lung seem to be markers of response to therapy.
A recent study demonstrated that eight days of antibiotic therapy were as effective as 15 days for treatment of VAP.11 Other studies have demonstrated that shorter courses of antibiotics are associated with a lower rate of super-infection with resistant bacteria. The policy of continuing on the same antibiotics for two to three weeks because the patient is doing better needs to be re-evaluated.
CONCLUSION
No single test will provide all the information needed to diagnose VAP.12 Clinicians must use multiple clinical criteria for the diagnosis. The use of culture techniques allows clinicians to enhance the sensitivity and specificity of the diagnosis of pneumonia.
In addition, the clinician can be kept up to date about potential new pathogens as they occur in the ICU. Knowing the local flora allows clinicians to make a better choice for empiric therapy for the VAP patient.
Dr. Baughman is a professor of medicine at the University of Cincinnati Medical Center. He has published more than 100 peer-reviewed articles in various aspects of pulmonary disease, including pneumonia. He’s interested in the diagnosis and management of pneumonia, including the application of techniques such as bronchoalveolar lavage.
For a list of references, please call John Crawford at (610) 278-1400, ext. 1499, or visit www.Respiratory-care-sleep-medicine.advanceweb.com.
Table 1: CPIS Scoring Key
Temperature
³ 36.5 ² 38.4 = 0
³ 38.5 ² 38.9 = 1
< 36 or ³ 39 = 2
WBC
³ 4,000 ² 11,000 = 0
11,000 = 1
Secretions
² small/day = 0
moderate/large = 1
purulent = 2
CXR
no infiltrate = 0
diffuse/patchy infiltrate = 1
localized infiltrate = 2
PF Ratio
> 240 without ARDS = 0
< 240 without ARDS = 2
Culture
< 10,000 bacteria or no growth = 0
> 10,000 bacteria = 1
positive Gram stain = 1
Table 2: Protocol for Nonbronchoscopic Bronchoalveolar Lavage15
Respiratory therapists at the University of Cincinnati use the following .protocol when performing nonbronchoscopic BAL. In an original series of 219 procedures, only two patients had significant hypoxia for more than 30 minutes after the procedure.
• Place patient on 100 percent oxygen.
• Be sure patient is well-oxygenated prior to procedure. If oxygen saturation isn’t greater than 92 percent, notify physician.
• Place adapter at end of tracheal tube to allow catheter to be placed in line.
• Pass catheter through tracheal adapter.
• Advance catheter approximately 30 cm. A black line on catheter should be lined up with lip (for endotracheal tube) or nares (for nasotracheal tube). This indicates catheter has been adequately advanced.
• Inner cannula advanced until it meets resistance (wedged).
• Two aliquots of 30 mL of saline are introduced and immediately aspirated.
• .Aspirated fluid is pooled and sent for Gram stain and semi-quantitative culture.
