Surgical Lung Biopsy in Immunosuppressed Pediatric Patients
Jennifer E Pogoriler, Aliya N Husain. University of Chicago Medical Center, Chicago, IL
Background: Surgical lung biopsies are typically performed in critically ill, immunosuppressed patients when less invasive techniques and broad spectrum anti-microbial treatments have failed. The differential can include infection, graft versus host disease, post transplant lymphoproliferative disease, recurrent or metastatic disease, drug reaction or radiation pneumonitis. Previous reviews of surgical lung biopsies in immunocompromised pediatric patients were performed in the 1980-90s and revealed rates of treatable etiologies in 54-82% of patients, with pneumocystis accounting for up to 70% of infections. Because of advances in treatment since that time we reviewed our series of lung biopsies to determine a more recent diagnostic yield for this procedure and to determine how well surgical pathology could identify organisms in comparison to culture.
Design: Of all pediatric surgical lung biopsy specimens from 1994-2012, we selected those from immunosuppressed patients with diffuse lung disease. We reviewed the chart for underlying cause of immunosuppression, radiology results, culture results and survival, and we reviewed all the lung biopsy slides and special stains.
Results: A total of 50 specimens were received from 46 patients; 19 (38%) had one or more types of infection identified. Pneumocystis remained the most common organism (8), followed by fungus (7), and CMV (4), and one case each of M Kansasii, RSV, parainfluenza, and Bacillus. 8 of the 19 infected patients (42%) did not survive the acute illness. Of the patients without identifiable infection with follow up, only 5 of 29 (17%) died of the acute illness (p=0.10). Surgical pathology failed to identify a specific agent detected by culture in three cases: one each of CMV, parainfluenza, and Aspergillus infection. Besides pneumocystis, microbiology cultures failed to detect one case of candida and one unclassifiable fungal organism. Dividing the cases into earlier (1994-2002, n=25) and later (2003-2012, n=25) time periods resulted in no significant differences in identification of organisms. In addition to infectious diseases, 3 cases of PTLD, 1 recurrent leukemia and 3 possible cases of drug reaction were identified. The remaining diagnoses were considered not specific and included diffuse alveolar damage, acute and organizing pneumonia, interstitial pneumonitis, and poorly formed granulomas.
Conclusions: Despite advances in treatment, infections such as pneumocystis and CMV are still among the most commonly identified agents in surgical lung biopsies. Identification of a treatable infection does not imply an improved outcome.
Wednesday, March 6, 2013 9:30 AM
Poster Session V # 251, Wednesday Morning