Open lung biopsy in pediatric bone marrow transplant patients

Open lung biopsy in pediatric bone marrow transplant patients

Open Lung Biopsy in Pediatric Bone Marrow Transplant Patients By Andrea Hayes-Jordan, Ely Benaim, Stacye Richardson, Javier Joglar, D. Kumar Srivastav...

47KB Sizes 0 Downloads 7 Views

Open Lung Biopsy in Pediatric Bone Marrow Transplant Patients By Andrea Hayes-Jordan, Ely Benaim, Stacye Richardson, Javier Joglar, D. Kumar Srivastava, Laura Bowman, and Stephen J. Shochat Memphis, Tennessee

Purpose: The aim of this study was to evaluate the clinical benefits of open lung biopsy in the diagnosis and treatment of pulmonary infiltrates in children who have undergone bone marrow transplantation. Methods: The authors retrospectively reviewed the medical records of all patients in whom pulmonary infiltrates developed within 6 months after bone marrow transplantation. Of 528 patients who received bone marrow transplants (313 allogeneic, 215 autologous) at St Jude Children’s Research Hospital between June 1991 and December 1998, 83 (16%) had radiographic evidence of pulmonary infiltrates after the procedure. Of these, 43 (52%) underwent bronchoalveolar lavage (BAL), 19 (23%) underwent open lung biopsy (OLB), 6 (7%) underwent needle biopsy, and 5 (6%) underwent transbronchial biopsy; 21 received medical therapy alone. The authors evaluated the outcome, culture results, histopathologic findings, radiographic findings, and clinical features of those who underwent OLB. Results: The 19 patients ranged in age from 0.9 to 19.8 years (median, 11.4 years). Histopathologic studies indicated an infectious process in 6 patients (30%), bronchiolitis obliterans organizing pneumonia (BOOP) in 5 (26%), interstitial pneumonitis (IP) in 4 (21%), gangliosidosis in 1, and lymphocytic infiltrate in 1. Although the clinical plan was changed on


HE USE of bone marrow transplantation (BMT) to treat children with leukemia or solid tumors is increasing. Pulmonary complications are one of the most common complications of bone marrow transplantation.1-5 Pulmonary infiltrates often are identified radiographically, and empiric therapy is begun frequently on the basis of the radiographic appearance of the lesions. However, standard empiric therapy does not always bring about clinical improvement; thus, establishing the

the basis of the histopathologic diagnosis for 17 of the 19 patients (90%), improvement in outcome was seen in only 8 (47.5%) of these patients. Postoperative morbidity (30 days) was 47% and included prolonged intubation (7 patients), pneumothorax (2 patients), and pleural effusion (1 patient). The 30-day survival rate was 63.2% ⫾ 10.6%. No patient with multisystem organ failure (MSOF), ventilator dependence, or a postoperative complication survived after OLB.

Conclusions: Histopathologic analysis of OLB specimens is very accurate in determining the cause of pulmonary infiltrates in pediatric patients who have undergone BMT, but OLB may not improve patient outcome. Because the postoperative morbidity and mortality rates associated with OLB are high, careful patient selection is necessary. The mortality rates of patients with MSOF or ventilator dependence are particularly high; therefore, less-invasive alternatives for diagnosis of pulmonary lesions should be considered before OLB is performed. J Pediatr Surg 37:446-452. Copyright © 2002 by W.B. Saunders Company. INDEX WORDS: Pulmonary infection, pulmonary infiltrates, lung biopsy, open lung biopsy, bone marrow transplant, complications.

From the Departments of Surgery and Hematology-Oncology, St Jude Children’s Research Hospital, Memphis, TN. Presented at the 32nd Annual Meeting of the American Pediatric Surgical Association, Naples, Florida, May 20-23, 2001. Supported in part by Cancer Center Support CORE Grant, P30 CA 21765 and by American Lebanese Syrian Associated Charities (ALSAC). Address reprint requests to Stephen J. Shochat, MD, Department of Surgery, St Jude Children’s Research Hospital, 332 North Lauderdale St, Memphis, TN 38105. Copyright © 2002 by W.B. Saunders Company 0022-3468/02/3703-0031$35.00/0 doi:10.1053/jpsu.2002.30854

definitive cause of these infiltrates is required. Minimally invasive procedures such as bronchoalveolar lavage may provide a sufficient amount of biopsy material to allow diagnosis of the cause of pulmonary infiltrates caused by infectious organisms. However, open lung biopsy (OLB) may be necessary when the cause is noninfectious, as with interstitial pneumonitis.5-8 OLB is the gold standard for obtaining the biopsy specimens used in diagnosing pulmonary disease and leads to an accurate diagnosis in 80% to 100% of adult patients.1,9-10 However, diagnosis based on OLB specimens has been reported to be less accurate when the patients are immunocompromised children.9,11-13 In addition, the complication rates in children after OLB vary widely, from 2% to 52%, depending on the patients’ immune status and the severity of the underlying disease.11 Previous reports of OLB in immunosuppressed children included patients who had undergone BMT, but because these patients were not evaluated as a separate group of immunocompromised children, the complications of OLB may have been underestimated. Other reports cited higher complication rates in association


Journal of Pediatric Surgery, Vol 37, No 3 (March), 2002: pp 446-452


with pancytopenia and other comorbid conditions. In an attempt to clarify these issues, we reviewed the role of OLB in the diagnosis and treatment of pulmonary infiltrates in an isolated group of pediatric patients who had undergone BMT. We also evaluated the short- and longterm outcome of these patients.


(30%). The most common infectious organism detected was aspergillus (4 patients). One patient was found to have a severe infection with the influenza virus and, another, a severe infection with Stenotrophomonas maltaphilia. The most common noninfectious diagnoses were BOOP, found in 5 patients (40%), and interstitial pneumonitis, found in 4 (31%).

MATERIALS AND METHODS At St Jude Children’s Research Hospital, 528 children received bone marrow transplants (313 allogeneic, 215 autologous) between June 1, 1991 and December 31, 1998. We conducted a retrospective review of the medical records of those patients who had experienced a pulmonary infiltrate within 6 months after BMT. Pulmonary infiltrates were identified in 92 children by computed tomography (CT) or chest radiography. Nine children whose infiltrate resolved after empiric therapy or whose pulmonary infiltrate occurred before BMT were excluded from this study. The 83 remaining patients had unilateral or bilateral pulmonary lesions that were visible on chest radiographs and fever that did not improve with empiric therapy. Of these, 21 continued to receive empiric therapy alone, and 62 underwent further evaluation for pulmonary infiltrates with one or more of the following methods: bronchoalveolar lavage, 43 patients; open lung biopsy, 19; needle biopsy, 6; and transbronchial biopsy, 5. We focused specifically on the 19 patients who had undergone OLB. We recorded demographic information, type of transplant (allogeneic, autologous, or donor antigen mismatch), radiographic distribution of pulmonary infiltrates, duration of antifungal and antibacterial therapy, therapeutic changes and clinical outcome after these changes, absolute neutrophil count (ANC) at the time of evaluation and treatment, use of steroids, comorbid factors, ventilator support at the time of procedure, complications, cause of death, and autopsy findings. For the purposes of this study, we defined empiric therapy as administration of a combination of broad-spectrum antibiotics to patients with pulmonary infiltrates and fever; the antibiotics used included ceftazidime, vancomycin, and meropenem. Amphotericin B or itraconozole was added when radiography cultures indicated fungal disease. High-dose steroids (prednisone or its equivalent at doses of at least 1 mg/d) were administered when bronchiolitis obliterans organizing pneumonia (BOOP) or interstitial pneumonitis was suspected. A change in therapy consisted of the addition or discontinuation of the use of antimicrobial agents, the addition of steroids for a diagnosis of BOOP, or the discontinuation of steroids when BOOP was not diagnosed. The addition or continuation of itraconozole or amphotericin B for patients with aspergillosis constituted a change in therapy, as did the administration of steroids once a diagnosis of interstitial pneumonitis was made and aspergillosis or another fungal infection was ruled out.


The 19 patients who underwent OLB ranged in age from 11 months to 19.8 years (median, 11.4 years). Four patients (21%) had acute lymphocytic leukemia (ALL), 6 (31%) had acute myelogenous leukemia (AML), 2 (10%) had solid tumors, and 2 (10%) had juvenile chronic myelogenous leukemia (JCML). Ten patients (53%) had relapsed or progressive disease. In all cases, OLB yielded sufficient biopsy material to allow a histopathologic determination of the cause of the infiltrate. The cause was determined to be noninfectious in 13 patients (58%) and infectious organisms in 6

Morbidity and Mortality No intraoperative complications were associated with OLB. Nine of the 19 patients (47%) experienced some form of postoperative morbidity, including prolonged intubation (7 patients), persistent pneumothorax (2 patients), and pleural effusion (1 patient, Table 1). Seven patients (37%) died within 30 days after OLB. Two of them had remained intubated for more than 2 days postoperatively, 2 had a persistent pneumothorax or pleural effusion, and all 3 patients who were intubated preoperatively could not be extubated and died postoperatively. Two patients required prolonged ventilatory support and died more than 2 months after OLB. Three of the 4 patients who received high-dose steroids preoperatively died after OLB. Three patients who had multisystem organ failure (MSOF) preoperatively died postoperatively. Two patients in whom MSOF developed postoperatively died, one 8 days after OLB and the other 77 days after OLB. Five of the 7 patients who died within 30 days after OLB were receiving high-dose steroids preoperatively, required preoperative mechanical ventilation, or had MSOF preoperatively. The other patients who died within 30 days after OLB had hemorrhagic pneumonitis or developed MSOF postoperatively. Change in Therapy The results of histopathologic diagnosis after OLB caused a change in therapy for 90% of the patients (17 of 19). However, this change in therapy improved the clinical condition of only 7 of these patients (42%). Four of these 7 patients were discharged within 30 days after OLB but died after being readmitted for complications unrelated to the pulmonary disease diagnosed by OLB (patients 1, 2, 4, and 7; Table 1). The remaining 3 patients whose condition improved clinically after a change in therapy (patients 9, 16, and 18) are alive after recovering from the pulmonary disease diagnosed by OLB. Thus, for only 3 of 17 patients (18%) did the change in clinical plan caused by the findings of histopathologic diagnosis after OLB result in clinical improvement and long-term survival. Bilateral or Unilateral Disease Fifteen patients had bilateral disease, and 4 had unilateral disease. A noninfectious cause was determined for



Table 1. Pediatric Bone Marrow Transplant Patients—Open Lung Biopsy Patient No.

BMT Type





2 3 4 5

Auto Allo Allo Allo


1,809 1,800 5,476 4,500

6 7 8 9 10 11

Allo Allo Allo Auto Auto Allo


3,920 0 1,955 0 0 15,925

12 13

Allo Allo

ALL Gang

4,661 6,460





15 16

Allo Auto


1,000 3,159

17 18 19

Allo Allo Allo


1,334 8,924 2,886



10,769 S S S, M M, V M, V





BOOP Interstitial pneumonitis Parainfluenza BOOP BOOP Interstitial pneumonitis Aspergillus Aspergillus Aspergillus Aspergillus Alveolitis Interstitial pneumonitis Gangliosidosis Interstitial pneumonitis


Diffuse alveolar damage BOOP Lymphoproliferative disorder BOOP Interstitial Fibrosis

Bilateral Bilateral Bilateral Bilateral


Intub Pn Intub

Cause of Death

D-21 mo

Respiratory failure†

D-8 mo D-22 d D-5 mo D-21 d

Brainstem herniation† Diffuse alveolar damage Liver hemorrhage† Respiratory failure Respiratory failure Liver failure, VOD† M, Pericarditis M Respiratory failure

Bilateral Bilateral Unilateral Bilateral Bilateral Bilateral


D-10 d D-31 d D-14 d A-1 yr D-8 d D-13 d

Bilateral Bilateral

Intub Intub

D-76 d D-77 d

M Progressive disease



D-15 d

Bilateral Unilateral


D-47 d A-7 yr

Hemorrhagic pneumonitis Pulmonary fibrosis and Senotrophmonas maltophilia Xanthomonas maltophilia?

Unilateral Bilateral Bilateral




A-7 yr A-6 yr D-34 d

Respiratory failure

Abbreviations: BMT, bone marrow transplant; Dx, diagnosis; ANC, absolute neutrophil count; Allo, allogeneic bone marrow transplant; ALL, acute lymphocytic leukemia; Auto, autologous bone marrow transplant; ST, solid tumor; S, on Steroids at time of biopsy; AML, Acute myelogenous leukemia; Intub, prolonged postoperative intubation; Pn, pneumothorax; CML, chronic myelogenous leukemia; JCML, juvenile chronic myelogenous leukemia; VOD, Veno-occlusive disease; M, multisystem organ failure; V, Ventilator dependent at the time of biopsy; A, Alive; D, Dead; SCID, severe combined immunodeficiency; Gang, gangliosidosis; APML, acute promyelocitic leukemia. *Complications after open lung biopsy. †Alive at discharge.

10 of the cases of bilateral disease and for 3 of the cases of unilateral disease. Five of the 10 patients with bilateral, noninfectious disease were alive 30 days after OLB, whereas all 3 patients with unilateral disease were alive. Eight (53%) of the 15 patients with bilateral disease suffered postoperative complications, and 5 of the 8 died. One (25%) of the patients with unilateral disease died of a postoperative complication. This patient had MSOF preoperatively. Prognostic Factors Patients who underwent autologous BMT were more likely to recover in the immediate postoperative period than were those who received allogeneic transplants. Three of 4 patients who underwent autologous BMT were alive 30 days after OLB, whereas only 9 (60%) of the patients who underwent allogeneic BMT survived for this length of time. No patient who underwent autologous BMT had postoperative complications, whereas 9

patients who underwent allogeneic BMT suffered postoperative complications. Sixteen patients had an ANC above 100 ⫻ 106/L at the time of OLB. Three patients had an ANC of 0 ⫻ 106/L at the time of OLB; 2 of them died, one 4 months and one 6 months after OLB. The third patient, who is alive, was given a transfusion of white blood cells and was treated with granulocyte colony-stimulating factor; the ANC recovered to more than 1,000 ⫻ 106/L within 2 days after OLB. The risk of postoperative complications is increased when the ANC is less than 500 ⫻ 106/L and is substantially increased when the ANC is 0 ⫻ 106/L. DISCUSSION

Previous reports of OLB in immunocompromised pediatric patients have included but not been limited to patients who underwent BMT.14-18 Because such patients are expected to experience different complications and are more severely immunocompromised than other pe-


diatric patients, the data from these patients should be analyzed separately when the outcome and complications of OLB are evaluated. Shorter et al,19 from the Children’s Hospital of Philadelphia, studied 126 children who had undergone BMT; 21 of these patients had also undergone OLB. In only 9 of the 21 cases was a specific diagnosis identified; the infectious organisms were fungal, viral, or bacterial and included Pneumocystis and cytomegalovirus (CMV). In the remaining 13 cases, the biopsy results were nonspecific, and the diagnosis given was “pneumonitis.” The overall mortality rate at 11 weeks was 65%; the mortality rate of patients whom an infectious organism was identified was 43%. Only 3 patients were alive 11 weeks after OLB; the investigators attributed the survival of 2 of these patients to a change in therapy brought about by the results of OLB. Since the report by Shorter et al,19 the treatment of pediatric patients who have undergone BMT has evolved and now includes the use of new prophylactic and therapeutic antimicrobial agents for pulmonary disease. Floreani et al9 recently reviewed 14 published reports of 625 immunocompromised pediatric and adults patients who underwent OLB. They found that the diagnostic yield of OLB in immunocompromised patients with pulmonary infections was 45% to 83%; changes in therapy were made on the basis of these results for 30% to 100% of patients. Snyder et al,10 from the University of Minnesota, reviewed 87 pediatric patients who had received bone marrow transplants from 1975 to 1986; the diagnostic yield of OLB was 60%. Coren et al12 evaluated 27 patients younger than 18 years of age who underwent OLB between 1991 and 1998 for the diagnosis of diffuse interstitial lung disease. A specific diagnosis was possible for 25 of 27 patients. In contrast, in our study, a specific histopathologic diagnosis was possible in all cases. Although the histopathologic diagnoses in our study were similar to those in the study by Snyder et al,10 the Minnesota study categorized the diagnosis as “nonspecific” when a specific pathologic entity was not detected (40% of cases). Histopathologic diagnoses are evolving continually, and this may account for the differences in diagnosis. Another difference between the 2 studies may be the indications for OLB. In our study, only 19 of 83 patients with pulmonary infiltrates underwent OLB; in the Minnesota study, a transplant center of a size equivalent to ours performed almost 4 times as many OLBs over a similar period. In addition, because the Minnesota study was done before the widespread use of CMV prophylaxis and empiric treatment of pulmonary infiltrates, CMV infection was a more prevalent fatal complication. Currently, broad-spectrum antibiotic therapy and agents such as acyclovir or ganciclovir are given before OLB is performed, thus, substantially reducing the incidence of CMV pneumonia. In fact, CMV


pneumonia did not occur in our study. Finally, at the time of the study by Snyder et al10 more than 15 years ago, diagnostic imaging methods such as spiral computed tomography (CT) scanners were unavailable; these methods can establish the exact location of specific infiltrates and may allow OLB to be more accurate, thereby reducing sampling error. The morbidity and mortality rates in this study were found to be proportional to the severity of the illnesses in our patient population. Pediatric patients who have undergone BMT are a unique group of patients for whom the onset of pulmonary disease may herald a poor outcome regardless of the diagnosis obtained by histopathologic analysis of samples obtained by OLB. Snyder et al10 reported a mortality rate of 74% and a complication rate of 21% after OLB over a 10-year period for pediatric patients who had undergone BMT. These findings are comparable to those of our study, which found an 8-year mortality rate of 85%. However, our complication rate of 47% is much higher. Coren et al12 reported that 3 of 27 children suffered complications after OLB. Our 30-day mortality rate after OLB was 37%. These differences in morbidity and mortality rates may be related to the severity of illness in the patients who underwent transplantation. In our study, many patients had very high-risk disease, and most underwent BMT after 1 or more disease relapses. Many underwent BMT after experiencing secondary solid tumors. Most of the patients in our study had either secondary AML (after ALL) or relapsed AML. In the studies by Davies et al11 and Snyder et al,10 not all patients underwent BMT, and specific diagnoses were not mentioned. In the study by Davies et al,11 most immunocompromised patients had acquired immunodeficiency syndromes; 55% of these patients experienced acute renal failure. Patients with immune deficiency and acute renal failure present different problems and should not be compared with recipients of bone marrow transplants. In addition, in our study, we included prolonged intubation as a postoperative complication; 8 of 9 patients with postoperative complications experienced prolonged intubation. In 2 patients, persistent bronchopleural fistula from the biopsy site contributed to postoperative respiratory failure. The studies by Davies et al11 and Snyder et al10 did not include prolonged intubation as a postoperative complication and may, therefore, have underestimated its incidence. Other studies of pediatric patients who have undergone BMT have not reported consistently the specific possible causes of increased morbidity and mortality rates after OLB other than pancytopenia. We have evaluated critically other factors such as steroid use, MSOF, infectious and noninfectious causes of pulmonary disease, and bilateral or unilateral pulmonary disease. The survival rates of patients with noninfectious dis-


ease, such as BOOP, appeared to be better than those of patients with infectious diseases. Of the 5 patients (26%) found to have BOOP, 4 (80%) were alive 30 days after OLB. Also, histopathologic analysis of the OLB specimen determined that 2 of the 4 long-term survivors had BOOP. BOOP is becoming a common histopathologic diagnosis at our institution. Few other studies of pediatric patients who have undergone BMT studies have mentioned their experience with BOOP. In our study, longterm survival was more likely for patients with BOOP than for patients with any other diagnosis facilitated by OLB. It is important to diagnose BOOP correctly in a patient who has undergone BMT and has pulmonary disease, because BOOP often can be treated successfully with high-dose steroids. Although some patients may have received high-dose steroids empirically before OLB to prevent clinical deterioration, it is necessary to confirm the diagnosis with histopathologic analysis of OLB specimens to determine whether there is a continued clinical need for steroids and to distinguish BOOP or other noninfectious causes of pulmonary infiltrates from infectious causes. High-dose corticosteroids most often are given to recipients of bone marrow transplants to treat graftversus-host-disease (GVHD). Such was the case for 4 of our patients. Infectious pulmonary organisms were detected in 3 of these 4 patients. With one exception, all patients receiving high-dose steroids became ventilator dependent postoperatively. None of these patients survived after OLB, and 3 died within 30 days postoperatively The single patient in this group who did not become ventilator dependent postoperatively had undergone autologous BMT and was able to recover from pulmonary disease; however, this patient died of brainstem herniation caused by a recurrent brain tumor. It appears that if high-dose steroids are necessary to treat GVHD or other nonpulmonary sequelae of BMT, survival is unlikely. Other investigators also have found that ventilator dependence is associated with a high mortality rate in patients who have undergone BMT. Adults who have undergone BMT and require mechanical ventilation as a result of pulmonary complications have a 70% to 90% chance of dying after becoming ventilator dependent.2,3 A mortality rate of almost 100% is seen in pediatric recipients of bone marrow transplants who have pulmonary disease and become ventilator dependent.20-22 A study by Keenan et al22 evaluated 121 pediatric patients who had undergone BMT and subsequently required mechanical ventilation. The mortality rate was 94% in the presence of pulmonary infection and 98% for patients with impairment of more than 1 organ system. These findings are consistent with ours. In our study, all 11 patients died who required mechanical ventilation pre-


operatively or more than 2 days postoperatively or who required intubation before OLB. All patients with MSOF died. Three patients with MSOF preoperatively died 10, 14, or 21 days postoperatively, whereas the 2 patients with MSOF postoperatively died, one 8 days and the other 47 days after OLB. In patients with a discrete solitary pulmonary nodule, OLB may be diagnostic and therapeutic. Most of our patients with discrete nodular disease were found to have aspergillosis. Three patients with 1 or 2 foci of aspergillus infection on radiographs of 1 or both lungs improved clinically after complete resection performed as part of the OLB and recovered within 30 days postoperatively. However, the condition of all 3 later deteriorated because of MSOF. One patient (patient 9) who has survived for one year with bilateral aspergillosis, had bilateral discrete nodules that were completely resected with multiple wedge resections and a right middle lobectomy. This child underwent autologous BMT. Although the ANC at the time of biopsy was initially 0 ⫻ 106/L, it quickly recovered postoperatively with the aid of granulocyte colony-stimulating factor and white blood cell transfusion. In patients with nodular disease, most of the disease burden can be removed by OLB; therefore, OLB provides a significant advantage in postoperative recovery from pulmonary disease. Patients with bilateral diffuse pulmonary infiltrates that do not respond to therapy may be unable to overcome such a diffuse process. Interstitial pneumonitis is one of the most common noninfectious causes of pulmonary disease in patients who have undergone BMT.12,23 This process usually is bilateral; in our study, all 4 patients with interstitial pneumonitis had bilateral disease. Three of these patients died of pulmonary complications or MSOF. This finding is comparable to those of other studies in which the mortality rate associated with interstitial pneumonitis is 40% to 70%.10,13 This entity is a bilateral diffuse process that represents a nonspecific pathologic diagnosis for which a directed treatment is unavailable, and steroids are the only effective treatment. In this study, the potential advantage of tailoring clinical therapy on the basis of the results of the histopathologic analysis of OLB specimens was not realized. Although therapy was changed for 17 of 19 patients on the basis of histopathologic results after OLB, this change in therapy resulted in prolonged clinical improvement for only 3 patients. Two of these patients were alive 7 years after OLB. One of the 2 patients for whom the clinical plan was not changed on the basis of OLB results survived for 6 years. This patient had BOOP and was receiving steroids empirically. Our study reflects the modern era, in which OLB rarely is required (only 3.5% of pediatric patients who undergo BMT need OLB for the diagnosis and treatment



of pulmonary infiltrates. Although OLB facilitates accurate diagnosis of the cause of pulmonary infiltrates in these patients, a change in treatment plan is unlikely to improve their short- and long-term clinical outcomes unless BOOP or a specific infectious agent can be treated. Because of the high complication rate associated with OLB, its use in recipients of bone marrow transplants should be limited to patients for whom the short-term prognosis is good. Removing discrete infectious lesions such as aspergillosis by OLB may cause clinical improvement in the immediate postoperative period, but the long-term prognosis

is poor. Because of the high morbidity and mortality rates associated with OLB in pediatric recipients of bone marrow transplants with pulmonary disease who are ventilator dependent, have MSOF, or who are receiving high-dose steroids, less-invasive alternatives to OLB should be sought to aid in diagnosis. Alternatively, earlier definitive diagnosis of pulmonary disease in pediatric patients who have undergone BMT may prevent complications such as ventilator dependence and MSOF, both of which are associated with high morbidity and mortality rates for patients who have undergone OLB.

REFERENCES 1. Cunningham I: Pulmonary infections after bone marrow transplant. Semin Respir Infect 7:132-138, 1992 2. Ewig S, Torres A, Riquelme R, et al: Pulmonary complications in patients with haemotological malignancies treated at a respiratory ICU. Eur Respir J 12:116-122, 1998 3. Hollmig KA, Soehngen D, Leschke M, et al: Long-term survival of recipients of allogeneic bone-marrow transplantation after mechanical ventilation. Eur J Med Res 2:62-66, 1997 4. Floreani AA, Sisson JH, Gurney J, et al: Thoracic complications related to bone marrow transplantation. Chest Surg Clin North Am 9:139-165, 1999 5. Leung AN, Gosselin MV, Napper CH, et al: Pulmonary infections after bone marrow transplantation: Clinical and radiographic findings. Radiology 210:699-710, 1999 6. Maschmeyer G, Link H, Hiddemann W, et al: Pulmonary infiltrations in febrile patients with neutropenia. Risk factors and outcome under empirical antimicrobial therapy in a randomized multicenter study. Cancer 73:2296-2304, 1994 7. Dichter JR, Levine SJ, Shelhamer JH: Approach to the immunocompromised host with pulmonary symptoms. Hematol Oncol Clin North Am, 7:887-912, 1993 8. Ellis ME, Spence D, Bouchama A, et al: Open lung biopsy provides a higher and more specific diagnostic yield compared to broncho-alveolar lavage in immunocompromised patients. Fungal Study Group. Scand J Infect Dis 27:157-162, 1995 9. Gentile G, Micozzi A, Girmenia C, et al: Pneumonia in allogenic and autologous bone marrow recipients. A retrospective study. Chest 104:371-375, 1993 10. Snyder CL, Ramsay NK, McGlave PB, et al: Diagnostic openlung biopsy after bone marrow transplantation. J Pediatr Surg 25:871876, 1990 11. Davies L, Dolgin S, Kattan M: Morbidity and mortality of open lung biopsy in children. Pediatrics 99:660-664, 1997 12. Coren ME, Nicholson AG, Goldstraw P, et al: Open lung biopsy

for diffuse interstitial lung disease in children. Eur Respir J 14:817-821, 1999 13. Kramer MR, Berkman N, Mintz B, et al: The role of open lung biopsy in the management and outcome of patients with diffuse lung disease. Ann Thorac Surg 65:198-202, 1998 14. Early GL, Williams TE, Kilman JW: Open lung biopsy. Its effects on therapy in the pediatric patient. Chest 87:467-469, 1985 15. Prober CG, Whyte H, Smith CR: Open lung biopsy in immunocompromised children with pulmonary infiltrates. Am J Dis Child 138:60-63, 1984 16. Ballantine TV, Grosfeld JL, Krapek RM, et al: Interstitial pneumonitis in the immunologically suppressed child: An urgent surgical condition. J Pediatr Surg 12:501-508, 1977 17. Rossiter SJ, Miller C, Churg AM, et al: Open lung biopsy in the immunosuppressed patient. Is it really beneficial? J Thorac Cardiovasc Surg 77:338-345, 1979 18. Imoke E, Dudgeon DL, Colombani P, et al: Open lung biopsy in the immunocompromised pediatric patient. J Pediatr Surg 18:816-821, 1983 19. Shorter NA, Ross AJ, August C, et al: The usefulness of open-lung biopsy in the pediatric bone marrow transplant population. J Pediatr Surg 23:533-537, 1988 20. Lupinetti FM, Behrendt DM, Giller RH, et al: Pulmonary resection for fungal infection in children undergoing bone marrow transplantation. J Thorac Cardiovasc Surg 104:684-687, 1992 21. Cerveri I, Zoia MC, Fulgoni P, et al: Late pulmonary sequelae after childhood bone marrow transplantation. Thorax 54:131-135, 1999 22. Keenan HT, Bratton SL, Martin LD, et al: Outcome of children who require mechanical ventilatory support after bone marrow transplantation. Crit Care Med 28:830-835, 2000 23. von Eiff M, Zuhlsdorf M, Roos N, et al: Pulmonary infiltrates in patients with haematologic malignancies: Clinical usefulness of noninvasive bronchoscopic procedures. Eur J Haematol 54:157-162, 1995

Discussion M. LaQuaglia (New York, NY): Was bronchoalveolar lavage done in the patients that had open lung biopsy, and, if so, what was the reason for the biopsy after the BAL? The second question is, exactly how did it change therapy in these situations? The third issue is, because a lot of your patients were

allogeneic, was graft-versus-host disease found in the biopsies and could that have been one of the reasons why they are on high-dose steroids beforehand? A. Hayes-Jordan (response): That was exactly the reason they were on high-dose steroids; that they had graft-versus-host disease. The answer to the second question, we included things


like changing antibiotics, eliminating antifungal therapy, making appropriate adjustments in steroids or starting steroids if steroids had not been started, or discontinuing or weaning steroids if steroids were inappropriate. The bronchoalveolar lavages were not done in the open lung biopsy group of patients. There was a separate group of patients who underwent bronchoalveolar lavage as part of their diagnosis. In this particular group of patients, the referring oncologist felt that it was not going to be an appropriate study for these patients and that they required an open lung biopsy. S. Rothenberg (Denver, CO): I enjoyed your presentation very much, and I think they are excellent data, but I would suggest that perhaps the conclusions you came to, except for your last slide, are wrong. I think the problem is that you are waiting to perform biopsy on these patients until too late. In fact, the oncologists and pulmonologist at our institution now often bypass bronchoalveolar lavage and go directly to thoracoscopic biopsy. They are trying to intervene earlier to get a specific diagnosis so we know whether to treat with steroids, we know whether to give or withhold antibiotics, and the morbidity of a thoracoscopic lung biopsy is nowhere near what you are showing. I think what you are showing us is the patient disease, not the morbidity of an open lung biopsy. Those patients that are dying probably were going that way anyway, and you did not affect their outcome by doing a biopsy, whether it is open or thoracoscopic, but the morbidity of a thoracoscopic biopsy certainly is much less. In our patients, and we presented this 2 years ago and are presenting some updated data in the oncology literature, there is no air leak, they do not have chest tubes, and they do not become ventilator


dependent. It is only when you wait until the patient is in the ICU, has been on the ventilator for a week, is crashing, and you are trying to get some last-minute idea to change your therapy that these factors become critical. If you would intervene earlier, when they start to show the infiltrates, start to show some deterioration, you can actually do the biopsy thoracoscopically. They are not intubated when they come out of the OR, you do not need to leave chest tubes in, and they do not have a high incidence of air leak. As opposed to your conclusions, I would argue that one slide that you gave that said it may be, if you intervene earlier, you will get the information you need, is very correct, and I encourage you to take a little bit different approach and come at it that way. A. Hayes-Jordan (response): I agree with you as far as the earlier intervention is concerned. It is very important to get involved early, and I think since we looked at these data there has been more of an effort to get us involved early, because that is definitely the key. I think, though, that the difference, like you said, in thoracoscopic biopsies may be significant, but there are some data to show that it is more than just the way that the biopsy is done, that it is the pulmonary status of these patients at the time that they undergo a general anesthetic that allows them to be at higher risk for being on the ventilator postoperatively. Of course, all those issues would not be relevant if you did the biopsy earlier, but in this particular group of patients I think that even though some patients were not on the ventilator before operation they had significant pulmonary disease, and I am not sure that doing their biopsy thoracoscopically would have provided them a significant advantage.