Utility of bronchoalveolar lavage in assessing pneumonia in immunosuppressed renal transplant recipients.

PURPOSE

To determine if initial results obtained from diagnostic bronchoalveolar lavage (BAL) in immunosuppressed renal transplant patients with pulmonary infiltrates, fever, or hypoxemia can affect therapeutic decisions, morbidity, and mortality.

DESIGN

A retrospective study of all BAL specimens obtained from renal transplant patients from January 1985 through June 1991. Initial results of Gram stain, cytology, cell differential count, and semi-quantitative bacterial cultures, all available within 24 hours of bronchoscopy, were compared with clinical outcomes and final diagnoses.

SETTING

University hospital nephrology-transplant/pulmonary service.

PATIENTS

Seventy renal transplant patients with a suspected pneumonia were stratified into 3 groups. A total of 48 patients underwent 58 bronchoscopies. Group 1 was comprised of 32 BALs that yielded 1 or more infectious organisms and was considered diagnostic. Group 2 (n = 26) were those BALs in which no organism was isolated and were thus nondiagnostic. Twenty-two additional immunosuppressed renal transplant recipients with pneumonia were considered by the admitting transplant nephrologist to have an uncomplicated community-acquired lung infection and thus were empirically treated and did not undergo BAL (Group 3).

METHODS

BAL fluid analysis included cell differential count, cytopathologic examination, and culture for mycobacteria, legionella, fungi, viruses, and bacteria using a semi-quantitative technique. Etiologic diagnosis and the time of onset of the infectious processes were recorded. Therapeutic outcome and mortality were determined for each group.

RESULTS

Thirty-nine etiologic organisms were found in 32 patients, with 6 patients having more than 1 infection. Twenty-two patients had 26 negative BALs, and 8 of these patients were clinically believed to have a volume overload state. Eight of 13 (61%) patients with bacterial pneumonia had BAL neutrophil counts greater than 20%, whereas 11 of 13 (84%) patients without bacterial pneumonia had neutrophil counts less than 20% (p < 0.05). Those patients with an infectious etiology remained in the hospital longer than patients without a specific etiologic organism identified (p < 0.02). Therapeutic decisions leading to the institution of specific antibiotics were more frequently made in patients with a diagnostic BAL (p < 0.0001). An overall 3-month mortality (16%) was low compared with the historical rate (30%).

CONCLUSION

BAL is a useful procedure in the diagnosis of an infectious process in immunosuppressed renal transplant patients where initial results can alter therapy in more than 70% of cases.