Early investigators of bronchoalveolar lavage (BAL) interpreted the presence of increased inflammatory cells in BAL to represent “alveolitis” based on the correlation between BAL findings and open lung biopsies from the same area (1). Investigators proposed cellular analyses of BAL as predictors of clinical outcome for various interstitial lung diseases (ILDs) (2). The cellular analyses of BAL seemed an ideal and relatively easy way to follow Sutton's law. When the American bank robber Willie Sutton was asked why he robbed banks, he is reported to have said “That's where the money is” (3).

Despite the attraction of using cellular analyses to predict clinical outcome in ILDs, it has not been clearly demonstrated that the cellular profile retrieved by BAL and change in cells after therapy correlate with clinical outcome of the disease. For example, although treatment with cyclophosphamide has been associated with reduction in the number of neutrophils in the BAL from some patients with interstitial pulmonary fibrosis, this was not correlated with changes in lung function. (4). Among patients with scleroderma-associated pulmonary fibrosis, a greater number of neutrophils in the BAL was associated with increased early, but not late, mortality (5).

Cyclophosphamide therapy appears to lead to improved clinical outcome in scleroderma-associated pulmonary fibrosis (6). White and colleagues found that increased BAL neutrophils or eosinophils predicted response to cyclophosphamide in treatment of scleroderma-associated pulmonary fibrosis (7). The Scleroderma Lung Study (SLS) included patients with BAL cellular analyses or high-resolution computed tomography (HRCT) abnormalities to recognize alveolitis. The study demonstrated a modest but significant improvement in measurements of FVC in patients treated with oral cyclophosphamide versus placebo after 1 year of therapy (8). Others reported a larger improvement with the use of intermittent, intravenous cyclophosphamide followed by oral azathioprine in scleroderma-associated pulmonary fibrosis (9).

In this issue of the Journal (pp. 91–98), Strange and colleagues report the lack of predictive value of BAL cellular analysis when obtained prior to therapy (10). The investigators report that patients with increased neutrophils or eosinophils in BAL had more extensive disease, but there was no difference between responders and nonresponders to therapy based on BAL findings. A retrospective analysis by Goh and coworkers also found that an abnormal BAL was not associated with response to therapy for patients with scleroderma (5).

Several reasons may explain the negative findings from the present study as well as the discrepancy between the findings observed in a previous report from a single-center trial (7). Variation in the technique for obtaining the BAL, including volume of saline instilled and lavage returns, as well as specimen handling, processing, analyses, and interpretation have led to guidelines for retrieving and handling BAL fluid (11). The SLS investigators used a standard lavage retrieval and analysis. However, they failed to address other potential reasons for the presence of increased neutrophils in the BAL sample as well as problems with single-area sampling.

Increased neutrophils in the BAL have been associated with many conditions. Thus, elevation of neutrophils in the BAL is nonspecific. In the study by Strange and colleagues (10), many patients with increased neutrophils in their BAL did not respond to cyclophosphamide. The SLS did not address other causes of increased neutrophils in the BAL fluid. Among these issues are the fact that neutrophils in the airways may contaminate the BAL sample. Some investigators discard the first aliquot of retrieved fluid for BAL cellular analyses (12). In the SLS, 40% of patients were past smokers and may have had airway inflammation. However, the investigators found no correlation between smoking exposure and BAL findings. Nevertheless, Schwartz and colleagues found that smoking history was a major determinant of BAL findings in idiopathic pulmonary fibrosis (13). Pulmonary infection can also cause increased BAL neutrophils (14). In a study of patients with scleroderma undergoing BAL, positive BAL cultures were identified in 10 of 42 patients (15). Eight of the 10 patients with positive cultures showed clinical deterioration over the following year. Occult aspiration secondary to esophageal dysmotility in scleroderma can also lead to increased neutrophils in the BAL fluid (16).

BAL sampling of the most appropriate site increases the amount of useful information available. In the SLS, the investigators chose to obtain BAL from right middle lobe regardless of parenchymal changes seen in HRCT. An attempt was made to correlate the BAL cellular findings with the HRCT images, but this was a post hoc analysis, and the time between the BAL procedure and the HRCT images obtained was unclear. Lavage in the right middle lobe has been a standard for diffuse disease (11). However, scleroderma may have patchy inflammation and fibrosis. Therefore, BAL obtained from the right middle lobe without evidence of ground-glass inflammation in this area may not accurately represent the inflammatory process of scleroderma-associated lung disease. In other studies of scleroderma-associated pulmonary fibrosis, BAL was usually performed in the lower lobes (5) or in the area most affected by HRCT (15). In addition, in scleroderma, there were different BAL results if the lavage was performed in areas of ground glass on HRCT versus HRCT showing honeycombing (17). When SLS investigators chose to lavage the same area in all patients, regardless of the parenchymal changes suggestive of inflammation and fibrosis, they ignored Mr. Sutton's advice.

The original studies suggesting that BAL results reflected the alveolitis of scleroderma were performed in areas of the lung that did not have extensive fibrosis on histologic examination (1). The SLS conclusion that BAL cellular analyses was not a predictor of response (10) only applies to lavage of the right middle lobe regardless of the HRCT images of the parenchymal abnormalities. The study by Goh and colleagues, which mostly included lower lobe lavage, tended to be in areas of honeycombing (5). Others have demonstrated that HRCT-directed bronchoscopy affects the BAL results in scleroderma (17). The lesson of sampling from areas of “active” disease (neither normal appearance nor end-stage fibrosis) has been learned in diagnosing idiopathic ILD by surgical lung biopsy. More information is therefore needed to determine the clinical role of BAL in evaluating inflammation and fibrosis in lungs from patients with scleroderma. Future studies are required to address confounding factors affecting BAL results and should include sampling from more than one area. In addition, the role of HRCT-directed BAL needs to be clarified. Sutton's law is: “Go where the money is.” Clearly, Mr. Sutton not only knew that there was money in the bank but he also knew exactly where to find it.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES