Approach to Nonresolving Pneumonia

Diagnostic Approach

In developing a diagnostic approach, it is important to first understand the capabilities as well as the limitations of the most commonly employed diagnostic tests. Careful consideration of the diagnostic yield, risks and benefits is critical in deciding whether additional invasive tests are warranted (Figure 11-1). The diagnostic tests that are most commonly employed in evaluating nonresolving pneumonia are:

• Chest radiographs
• Chest CT scans
• FOB
• Open-lung biopsy.

Enlarge 

Imaging Tools

As is clear from Natural History of Pneumonia, and Possible Etiologies of Nonresolving Pneumonia, radiographic findings alone are almost never specific for any one diagnosis. However, radiographic studies are useful in narrowing the differential diagnosis and suggesting groups of diagnostic possibilities for consideration. The primary imaging tools in assessing nonresolving pneumonia are chest radiograph and chest computed tomography (CT) scan. The evaluation of nonresolving pneumonia has benefited from the development of high-resolution computed tomography (HRCT) of the chest. HRCT involves using thin-section scanning at 1- to 2-mm collimation combined with a high spatial reconstruction algorithm. HRCT is superior to conventional techniques in several key areas that impact on the management of nonresolving pneumonia. Compared with conventional chest x-ray (CXR), HRCT allows superior detection of parenchymal abnormalities, including emphysema, air-space disease, interstitial disease and nodules. Detection of these structural abnormalities may narrow the differential diagnosis or suggest new possibilities.

Certain conditions, such as amiodarone toxicity and lymphangitic spread of malignancy, have specific HRCT characteristics that may suggest a diagnosis with reasonable specificity. In addition, the greater sensitivity of HRCT allows for better precision in assessing a patient’s response to therapy over time. This is especially useful when there is preexisting lung disease that makes it difficult to distinguish acute from chronic changes. HRCT also improves detection of localized collections, such as abscesses and empyema. Finally, the ability to better localize disease helps direct biopsy procedures and may improve diagnostic yield. In addition to HRCT, bedside ultrasound has added additional diagnostic capability when pleural effusion and associated atelectasis are the causes of persistent radiographic abnormality.

Bronchoscopy

The role of fiberoptic bronchoscopy (FOB) in the diagnosis of nonresolving pneumonia depends largely on the clinical scenario. The best accepted indication for FOB in the diagnosis of pneumonia is in an immunocompromised host with diffuse pulmonary infiltrates. In this setting, different organisms that require markedly different treatments may have similar or indistinguishable clinical presentations. While clinical and radiographic patterns may narrow the set of diagnostic possibilities, abnormal host factors, poor baseline cardiopulmonary reserve, and the wide spectrum of possible pathogens often make an empiric trial risky and therefore justify early FOB. Similarly, in cases of nonresolving pneumonia, the relative ease and low risk of FOB make this an appealing diagnostic procedure in a population of patients with a similarly wide spectrum of possible infectious and noninfectious etiologies.

Despite the frequency of its use for this indication, there are few studies that document the diagnostic yield of FOB for nonresolving pneumonia. Retrospective analysis of FOB for nonresolving pneumonia demonstrates that FOB is used successfully to diagnose 86% to 95% of patients who eventually have a specific diagnosis established. FOB is more likely to establish a specific diagnosis in younger, nonsmoking patients with multilobar involvement and prolonged duration of disease. Patients older than 55 years of age, smokers and those with immune defects are more likely to have a nondiagnostic bronchoscopy and slowly resolving pneumonia.

The utility of FOB in nonresolving pneumonia also depends upon the disease possibilities being considered and the population being studied. FOB is most useful in diagnosing unusual pathogens, some immunologic disorders (e.g., chronic and acute eosinophilic pneumonia (AEP)), and neoplastic diseases. Again, depending upon the diseases being considered, trans­bronchial biopsy may or may not be necessary. In other situations, FOB may have a relatively low diagnostic yield but may provide useful information in ruling out infectious processes. This is especially important if immunosuppressive therapy is being considered.

The role of bronchoscopy in ruling out bacterial infections in the setting of nonresolving pneumonia is unclear. Most recommendations are based upon extrapolating data from studies of community-acquired pneumonia (CAP) and hospital-acquired pneumonia (HAP). Since the causative organism in CAP is not isolated in >40% of cases, the initial role of FOB is limited. FOB for CAP, particularly if done prior to antibiotic therapy, increases the percentage of cases with a defined etiology. However, the additional pathogens that are isolated are almost always covered by routine empiric antibiotic therapy. Therefore, the role of FOB in identifying bacterial pathogens in nonresolving CAP is not easily defined. Unless unusual pathogens such as those for tuberculosis (TB) are present, the diagnostic sensitivity and specificity for pathogens in this population is probably limited.

Based upon studies of ventilator-associated pneumonia (VAP), however, several recommendations can be made. First, unprotected collection techniques, such as tracheal aspirates and unprotected bronchoalveolar lavage (BAL), are of little value for bacterial pathogens. While multiple protected bronchoscopic techniques have been utilized, each with their own particular advantages and disadvantages, it is unclear if any one technique is markedly superior. Techniques include protected specimen brush (PSB) and protected bronchoalveolar lavage (PBAL) with quantitative culture.

The important point with all of these techniques is to obtain specimens from the distal alveolar or respiratory bronchiole with minimal proximal airway contamination. Controversy persists whether diagnostic bronchoscopy should be performed on all patients with VAP since multiple studies have shown no survival benefit compared with that achieved with empiric therapy alone. Whether this applies to nonresolving pneumonia is unclear. Certainly FOB that detects noninfectious etiologies would be expected to alter therapy and presumably impact on survival. Given these limited data, it is still best to utilize a protected specimen technique if bacterial pathogens are suspected, realizing the limitations of the technique. Additional bronchoscopic capability has been added by the introduction of endoscopic ultrasound to enhance both lung parenchymal and mediastinal lymph node biopsy. Further improvements in diagnostic yield may accompany more widespread use of robotic bronchoscopic techniques. While specific utility in nonresolving pneumonia has not yet been demonstrated, these techniques are promising and may eventually be incorporated into diagnostic algorithms.

Open-Lung Biopsy

While definitive recommendations on the decision to proceed to open-lung biopsy cannot be made, several factors need to be considered when deciding on whether to proceed with open-lung biopsy. These factors include disease progression, the diagnostic possibilities being considered, and the effect that a positive open-lung biopsy will have on treatment. In general, if the disease is relatively stable, a period of careful observation may be warranted. If there is a high likelihood for a disease that would necessitate a dramatic change in therapy, open-lung biopsy is warranted. Diseases in this category generally include most vasculitis syndromes (WG) and inflammatory lung diseases that require immunosuppression. In these cases, the risk of immunosuppression in a patient who is currently infected requires a specific tissue diagnosis. The more potent the immunosuppression required, the more open-lung biopsy is warranted. Similarly, FOB in these cases can help to rule out concurrent infection, but open-lung biopsy to establish the diagnosis remains the gold standard. The safety of open-lung biopsy has been improved by the introduction of video-assisted and robotic thoracic surgery (VATS), which should be employed where possible.

Summary

The diagnostic evaluation of nonresolving pneumonia begins with a careful history, physical examination and review of the medical record. The goal is to determine whether the rate of resolution is within the range of expected norms, taking into consideration the patient’s underlying host factors, comorbidities, severity of illness and any known pathogens. If the patient is stable or slowly improving and has other comorbidities or host factors that are known to delay the rate of resolution of pneumonia, careful observation and continued therapy are warranted for 4 to 8 weeks. If there is no resolution or there is a progression of disease, a more aggressive diagnostic approach is warranted.

The physician must first determine whether the nonresolving pneumonia is due to an infectious or a noninfectious etiology. The initial evaluation should include a chest CT scan to look for unsuspected nodules or localized collections of fluid. Any significant pleural collections should be biopsied or drained. If this is unrevealing, bronchoscopy should be considered.

Several factors should be considered when deciding on whether to proceed with FOB. As mentioned earlier, in cases with stable but nonresolving pneumonia with impaired host defenses, it is reasonable to observe the patient, since the infection can be expected to take a longer time to clear. When infection fails to resolve in a patient without impaired host defenses or if there is clinical progression, FOB should be pursued. Similarly, if noninfectious etiologies or unusual pathogens are suspected, FOB is warranted. Positive results from FOB can serve to modify or optimize treatment regimens.

Similarly, a negative result has significant value. Patients with a negative FOB have a good chance of merely having a slowly resolving pneumonia and if they are stable can be observed. Similarly, a negative FOB will narrow the differential diagnosis in patients with progressive disease. Diseases that typically are not diagnosed with FOB that are progressive include pulmonary vasculitis syndromes, cryptogenic-organizing pneumonia (COP) and the various forms of diffuse alveolar damage. In these cases, a negative bronchoscopy with progressive symptoms should prompt consideration of an open-lung biopsy.