The diagnostic assessment of new pulmonary infiltrates in immunosuppressed patients remains a major problem for the clinician (Table 1). The number of patients with severe immunosuppression at risk for a pulmonary complication continues to increase as more patients receive cytotoxic chemotherapy, undergo organ transplantation, or become infected with the human immunodeficiency virus. The challenge to the clinician is to be aware of the changing patient populations and to recognize the relative risk of which complication is associated with a specific clinical manifestation. Valid diagnostic approaches developed in the 1960s and 1970s may have less applicability as newer technologies become available and provide for a less invasive and more rapid diagnosis (Table 2).
For example, in the 1970s (the period of my residency in internal medicine and my pulmonary fellowship at the Mayo Clinic), the accepted dictum was that an open-lung biopsy was the best diagnostic procedure for determining the specific cause of a pulmonary complication in immunosuppressed patients. In this context, “best” indicated that the procedure was safe, was cost-effective, and provided the highest diagnostic yield. In the 1990s, the same criteria support an ongoing role for open-lung biopsy in selected cases; however, in most cases, bronchoscopy in conjunction with bronchoalveolar lavage, with or without transbronchoscopic lung biopsy, is clearly the procedure of choice for evaluating new pulmonary infiltrates in immunosuppressed patients. The diagnostic yield of less invasive techniques has improved for two reasons: (1) an improvement in the ability to obtain adequate lung tissue and fluid samples and (2) an improvement in the sensitivity and specificity of microbiologic and histopathologic techniques.
Table 1Differential Diagnosis of New Pulmonary Infiltrates in an Immunosuppressed Patient
Table 2Diagnostic Options in the Evaluation of New Pulmonary Infiltrates in an Immunosuppressed Patient
In this issue of the Mayo Clinic Proceedings (pages 221 to 227), Pisani and Wright provide further support for the utility of bronchoalveolar lavage in the assessment of immunocompromised patients. In their study, the overall diagnostic yield of bronchoalveolar lavage (the number of positive bronchoalveolar lavage results divided by the total number of bronchoalveolar lavages) was 39%. These results were similar to a previously reported diagnostic yield of 33% in a similar population of immunosuppressed patients at the Mayo Clinic.
1The difference in yield between the two studies is, at least in part, related to the inclusion of more patients with Pneumocystis carinii pneumonia in the current study; the yield from bronchoalveolar lavage in these patients should exceed 90%. In any event, the diagnostic yields are comparable in these two studies but represent a lower percentage of specific diagnoses from bronchoalveolar lavage than that reported in most other studies in the literature (see Table 5 in the Pisani and Wright article, page 226). The most obvious explanation for the lower diagnostic yield is the considerably lower percentage of patients with acquired immunodeficiency syndrome (AIDS) in the two Mayo Clinic series than in the other published reports. Because patients with AIDS, in comparison with other patients, have a more profound level of immunosuppression and a resultant increase in microorganisms in the lung during infection,
2techniques such as bronchoalveolar lavage or even sputum analysis will have an associated higher diagnostic yield in patient populations that consist of a high percentage of patients with AIDS.
Pathogenicity Versus Colonization.
In the current study, an important effort is made to use specific criteria to determine whether microorganisms obtained by bronchoalveolar lavage are pathogens or “colonizers” in the lower respiratory tract. In Table 2 of their article (page 223), Pisani and Wright propose reasonable clinical criteria to determine whether isolated microorganisms are colonizers rather than pathogens. Unfortunately, the criteria necessitate subjective interpretation of observed findings, and “clinical judgment” cannot be quantified (nor perhaps should it be). Nonetheless, the inability to distinguish pathogens from colonizers remains an important obstacle to interpreting the results of bronchoalveolar lavage.
This issue of pathogenicity versus colonization relates most frequently to specific types of bacteria, fungi, and viruses. In the prior Mayo Clinic study, bacteria and fungi were considered pathogenic if the organisms could be either successfully cultured from a specimen (such as blood or pleural fluid) in addition to that obtained by bronchoalveolar lavage or confirmed by open-lung biopsy.
1In certain clinical conditions, however, the finding of a potential colonizing microorganism would indicate it should be classified as a pathogen. Bronchoalveolar lavage in a patient who has granulocytopenia after bone marrow transplantation may reveal a positive smear or culture for Aspergillus; with this finding, treatment with parenterally administered amphotericin B should be initiated, on the assumption that invasive aspergillosis is present. Compatible clinical findings such as nodular densities, pleuritic pain, hemoptysis, and fever would further support this diagnosis. Thus, the finding of a typical commensal organism in the lower respiratory tract may, in an appropriate clinical setting, be accepted as a true pathogen.
Perhaps one of the most difficult diagnoses to establish in immunosuppressed patients is cytomegalovirus (CMV) pneumonia. In the study by Pisani and Wright, CMV was detected in 29 of 150 patients, but it was classified as a pathogen in only 7 of the 29 patients (24%). Is CMV rarely a pathogen in the lung, or do these results suggest an underestimation of the true incidence of the disease? The answer is neither simple nor straightforward.
Clearly, bronchoalveolar lavage is a superb technique for detecting the presence of CMV in the lower respiratory tract.
3The difficulty is determining whether the CMV organism is responsible for the clinical findings or is present in the lower respiratory tract as an “innocent bystander.” The presence of CMV in the lower respiratory tract can be detected by three methods: (1) the shell vial method, (2) culture, and (3) the finding of a typical cytopathic effect in lung cells. CMV can be readily detected in the lower respiratory tract because both culture and shell vial methods with use of bronchoalveolar lavage are sensitive techniques.
3To demonstrate that CMV is the pathogen in the lung responsible for the observed clinical findings is much more difficult. Many investigators rely on detection of the virus (by immunofluorescence microscopy or culture) together with evidence of the typical inclusion bodies within the macrophages (bronchoalveolar lavage) or lung parenchymal cells (biopsy). This approach offers the promise of specificity (detection of the virus) and evidence of tissue invasion (adverse effect of the virus).
In special circumstances, even less diagnostic information can strongly suggest the presence of CMV pneumonia. For example, CMV infection is a common complication of bone marrow transplantation, especially after the first month postoperatively. In a study by Crawford and associates,
4an open-lung biopsy was performed in 111 patients who had diffuse pulmonary infiltrates after bone marrow transplantation. When infection was diagnosed (in 63% of all patients), CMV pneumonia was the cause in more than 90% of the cases. In contrast, a bacterial infection was diagnosed in only 2 of the 111 patients and P. carinii was detected in only 6. Springmeyer and colleagues
5(from the same institution) demonstrated that bronchoalveolar lavage detected CMV in more than 80% of the patients with CMV pneumonia.
Further evidence that CMV pneumonia is a major complication after bone marrow transplantation was provided by Schmidt and co-workers,
6who found CMV in the bronchoalveolar lavage fluid in 40 of 105 patients (38%). Interestingly, CMV pneumonia developed in 70% of patients in whom ganciclovir was not administered but in only 25% of those who received this drug. This result suggests that detection of CMV by bronchoalveolar lavage in patients who have undergone bone marrow transplantation should be strongly considered as evidence of the development or impending development of CMV pneumonia.
Differences in the timing of bronchoalveolar lavage after bone marrow transplantation may also yield varied results. For example, whereas CMV pneumonia is a common cause of diffuse pulmonary infiltrates after the first 4 to 6 weeks subsequent to bone marrow transplantation, CMV is relatively uncommon during the first month after transplantation. Diffuse pulmonary infiltrates during this period most likely will reflect pulmonary edema or alveolar hemorrhage. The latter diagnosis is far more common than previously realized. It is reasonable to include hemosiderin-stained cytopreparation smears as a routine part of the bronchoalveolar lavage analysis in patients at risk for alveolar hemorrhage—that is, those who have undergone bone marrow transplantation, have coagulopathies, or have platelet deficiencies. In patients with alveolar hemorrhage, the pulmonary infiltrates typically resolve rapidly unless rebleeding into the alveolar spaces occurs. Many patients with alveolar hemorrhage, however, will continue to have persistent bleeding in the lower respiratory tract with no clinical evidence of hemoptysis. Again, the addition of screening for alveolar hemorrhage would increase the diagnostic yield from bronchoalveolar lavage in this patient population.
Cytotoxic Drug Reaction.
A common pulmonary complication in the patient population described by Pisani and Wright is the presence of a cytotoxic drug reaction in the lung. The diagnosis of a cytotoxic drug reaction is typically based on compatible biopsy findings of cellular atypia and a patient history of having received cytotoxic therapy. Occasionally, patients may have clinical evidence of a “triggering event” for the toxicity, such as occurs with precipitation of bleomycin toxicity by exposure to high doses of oxygen. Frequently, patients will have fever, progressive respiratory symptoms, and diffuse pulmonary infiltrates suggestive of an infectious process. Whereas an open-lung biopsy provides the best confirmatory evidence of tissue inflammation and cellular atypia, bronchoalveolar lavage can yield essentially equivalent information.
7In either case, the clinical decision that a patient has a cytotoxic drug reaction necessitates compatible bronchoalveolar lavage or tissue findings in the appropriate clinical setting. Many investigators have estimated that cytotoxic drug reactions account for as much as a third of the cases of diffuse pulmonary infiltrates in immunosuppressed patients who do not have AIDS.
Bronchoalveolar lavage is clearly a useful diagnostic technique for assessing immunosuppressed patients. As the diagnostic yield from sputum analysis increases, however, the need for invasive techniques such as bronchoscopy and bronchoalveolar lavage may be decreased. Sputum analysis remains an excellent means of diagnosing P. carinii pneumonia in patients with AIDS, but its diagnostic utility for P. carinii in immunosuppressed patients who do not have AIDS is limited. Furthermore, the diagnostic yield of sputum analysis for other infections is substantially less than that of bronchoscopy. Although intuitively sputum analysis would seem more cost-effective than bronchoscopy, factors that must be considered are the cost of a delay in diagnosis, the repetition of expensive microbiologic and cytologic tests, and the potential need for bronchoscopy eventually. Clearly, sputum studies are of value in the preliminary assessment of a productive cough and other respiratory symptoms in patients with AIDS. Currently, however, its value in the assessment of immunosuppressed patients who do not have AIDS is unknown.
For nonpulmoriary clinicians, the diagnostic yield of 30 to 40% for bronchoalveolar lavage in immunosuppressed patients may seem low. Nevertheless, bronchoalveolar lavage likely detects most potentially treatable pulmonary disorders.
1Open-lung biopsy certainly has a greater diagnostic yield than does bronchoalveolar lavage (a diagnosis is obtained in 40 to 70% of patients); however, many of these “diagnoses” are nonspecific—such as interstitial pneumonitis or pulmonary fibrosis. In addition, they often are not amenable to specific therapy, and the “improved” diagnostic yield will not alter survival. Ultimately, the clinician must choose a diagnostic procedure on the basis of safety, useful diagnostic information, and cost. In most patients, bronchoscopy will be the procedure of choice.
Bronchoalveolar lavage is an excellent procedure for detecting opportunistic infection, alveolar hemorrhage, and malignant disease. In addition, it provides important confirmatory information to support the diagnosis of a cytotoxic drug reaction. Additional prospective studies will be needed to determine whether microorganisms isolated from specimens such as bronchoalveolar lavage fluid truly are pathogenic organisms or simply are colonizers of the lower respiratory tract. The current study by Pisani and Wright indicates that substantial progress has been made in attempts to establish criteria that will assist the clinician in making this difficult diagnostic decision. Future studies should continue to shed light on this diagnostic problem, and advances in technology should continue to improve the diagnostic yield of this procedure.
- Role of bronchoalveolar lavage in the assessment of opportunistic pulmonary infections: utility and complications.Mayo Clin Proc. 1987; 62: 549-557
- Pneumocystis carinii pneumonia: differences in lung parasite number and inflammation in patients with and without AIDS.Am Rev Respir Dis. 1989; 140: 1204-1209
- Rapid detection of cytomegalovirus in bronchoalveolar lavage specimens by a monoclonal antibody method.J Clin Microbiol. 1986; 23: 1006-1008
- Open lung biopsy diagnosis of diffuse pulmonary infiltrates after marrow transplantation.Chest. 1988; 94: 949-953
- Use of bronchoalveolar lavage to diagnose acute diffuse pneumonia in the immunocompromised host.J Infect Dis. 1986; 154: 604-610
- A randomized, controlled trial of prophylactic ganciclovir for cytomegalovirus pulmonary infection in recipients of allogeneic bone marrow transplants.N Engl J Med. 1991; 324: 1005-1011
- Utility of bronchoalveolar lavage in the diagnosis of drug-induced pulmonary toxicity.Acta Cytol. 1989; 33: 533-538
- Diffuse pulmonary infiltrates in the immunocompromised host.Clin Chest Med. March 1990; 11: 55-64
© 1992 Mayo Foundation for Medical Education and Research. Published by Elsevier Inc. All rights reserved.