NORTHBROOK, ILL--When presented with a suspected case of ventilator-associated pneumonia (VAP), clinicians can choose from a wide array of diagnostic modalities, including several invasive procedures whose benefits have not been clearly delineated. Until recently, however, no specific guidelines have been available to aid diagnosis. To address this problem, the American College of Chest Physicians (ACCP) convened a clinical practice guideline panel to develop an evidence-based assessment of diagnostic tests for VAP.[1]

In an interview with PULMONARY REVIEWS, panel member G. Douglas Campbell, MD, observed that there is a good deal of published information discussing diagnostic methods for VAP. "Unfortunately, these papers have tended to be scattered throughout the literature," said Dr. Campbell, who is a Professor of Pulmonology and Critical Care Medicine at Louisiana State University in Shreveport. "We felt it was important at this time to pull together these various reports and, based on a rigorous review of the literature, offer specific practice recommendations."

Donald E. Craven, MD, who was also a member of the ACCP panel, described another benefit of the new guideline. "Only recently have studies been based on outcomes from VAP; earlier investigations focused only on the sensitivity and specificity of clinical versus invasive diagnosis," said Dr. Craven, Director of the Clinical AIDS Program at Boston City Hospital. As a result, the best approach to diagnosis has remained controversial. "We think these new recommendations will help clinicians in utilizing outcome-based data," Dr. Craven added.

DIAGNOSTIC HALLMARKS FOR VAP

Temperature abnormalities, leukopenia or leukocytosis, and purulent tracheal secretions are the diagnostic hallmarks for VAP; decreased arterial oxygen tension also suggests the diagnosis. When two or more of these findings are accompanied by chest film abnormalities, VAP is likely. The specificity of this combination is quite low, however.

Therefore, the ACCP guideline suggests that a diagnosis of VAP is improbable in the absence of two or more of these clinical findings. If two or more are present, the next step is to order a chest film.

CHEST FILM ABNORMALITIES

The guideline acknowledges that "the accuracy of interpretation of chest radiographs has not been extensively evaluated." Also unclarified is the incidence of pneumonia in immunocompromised patients (such as those with Pneumocystis carinii pneumonia) who have clinical findings compatible with VAP but have normal chest film results. One problem with chest films is that findings considered typical of VAP, such as alveolar infiltrates, have only moderate sensitivity for pneumonia, and the specificity of these findings is unknown.

The authors of the guideline conclude, therefore, that when the chest film is normal, other causes of the clinical abnormalities should be investigated. The guideline authors also suggest that the presence of a single abnormal chest film finding does not substantially increase the likelihood of VAP, since patients receiving mechanical ventilation are prone to other disorders that can cause radiographic abnormalities.

However, when more than a single chest film abnormality is detected, or when such findings have worsened, follow-up testing is appropriate.

ENDOTRACHEAL ASPIRATION

One option for additional testing is qualitative cultures of endotracheal secretions. This approach is frequently used for diagnosis, particularly when an empiric antimicrobial regimen is started, because aspiration can be performed at the bedside by nonphysician health care personnel. If the results of the qualitative culture are negative, the guideline suggests, VAP is not likely to be present, except in patients previously given antibiotics.

Positive results must be interpreted with caution, because the samples may have been contaminated during collection. Furthermore, results can vary with the bacterial load and the duration of mechanical ventilation; they also can be altered if antimicrobials were previously used.

As a consequence, estimates of the sensitivity and specificity of qualitative cultures have ranged from quite low (14% to 38%) to quite high (100%). The authors of the guideline acknowledge that there is no clear evidence to confirm or refute the use of Gram staining and culture of endotracheal secretions obtained by aspiration.

The presence of either antibody coating or elastin fibers has not been found to be diagnostically sensitive or specific for VAP. The guideline states that testing for these variables is not recommended.

Another approach for additional testing is to use quantitative bronchoscopic or nonbronchoscopic techniques. Quantitative techniques include bronchoalveolar lavage (BAL), the protected specimen brush (PSB) method, and several other less invasive techniques.

Because these techniques appear to have similar diagnostic efficacy, the guideline recommends that the choice be based on "local expertise, experience, availability, and cost factors." However, the authors of the guideline admit that just as there is little evidence to support or refute the use of qualitative testing, there is "insufficient high-level evidence" to prove that quantitative testing produces a better clinical outcome than does the combination of qualitative testing and empiric treatment.

Although BAL has been used for more than a decade to diagnose VAP, bronchoscopic and bacteriologic techniques have not been standardized. The presence of 103 to 105 cfu/mL is usually considered a positive result for quantitative cultures; sensitivity corresponds inversely with the end point selected.

Estimates of the specificity of BAL range widely; however, specificity is improved if intracellular organisms are detected, and this finding has a high positive predictive value. Its sensitivity is unclear, though; estimates range between 37% and 100%.

The PSB method also has been used for a number of years, without standardization. The available evidence suggests that the accuracy of this technique increases if samples are obtained from the affected, rather than the unaffected, lobe. Accuracy also improves if multiple samples are obtained and tested. Overall, the PSB method appears to have greater specificity than sensitivity for the diagnosis of VAP.

The authors of the guideline stress that the risks associated with BAL and the PSB method have not been extensively evaluated. Bronchoscopy alone may lead to transient alterations in oxygenation in ventilated patients; it is not known whether use of the PSB increases this risk. In addition, bronchoscopy is often inconvenient and expensive, and it requires sophisticated technical expertise.

BLIND PROCEDURES

Because of the difficulties associated with bronchoscopy, other diagnostic tests have been developed, including three blind, nonbronchoscopic techniques: blind bronchial sampling (BBS), mini-BAL, and blind sampling with PSB (BPSB).

BBS involves blindly wedging a catheter into a distal bronchus; secretions are then aspirated without fluid instillation. In mini-BAL, fluid is instilled via a small catheter (usually, a plugged telescopic catheter is employed, but an unprotected catheter is sometimes used instead). BPSB employs a sterile brush that is inserted without bronchoscopic guidance.

"None of these tests have been standardized," states the guideline. Not surprisingly, therefore, the estimates of their sensitivity and specificity vary considerably. In addition, these newer procedures have not been validated in postmortem studies. However, the risks from blinded techniques appear to be minimal; at worst, they are no greater than those associated with fiberoptic bronchoscopy.

NONRESOLVING PNEUMONIA

When pneumonia fails to resolve despite antimicrobial therapy, clinicians are faced with a dilemma. At present, many obtain serial specimens (via a bronchoscope) in the hope of identifying the causative organism, but the utility of this approach has not been fully investigated. However, two recent studies have shown that frequent changes in the antimicrobial regimen based on serial culture results did not lower mortality. Thus, the guideline concludes that there is insufficient evidence to determine whether repeated bronchoscopy can improve outcome in patients with nonresolving pneumonia.

--Stanley Nelson

Reference
1. Grossman RF, Baughman R, Campbell GD, et al. Evidence-based assessment of diagnostic tests for ventilator-associated pneumonia. Chest. 2000;(4 suppl 2):1775-2185.

Return to table of contents