clinical investigations in critical care The Role of Open-Lung Biopsy in ARDS*

Transcription

1 clinical investigations in critical care The Role of Open-Lung Biopsy in ARDS* Sanjay R. Patel, MD; Dimitri Karmpaliotis, MD; Najib T. Ayas, MD; Eugene J. Mark, MD; John Wain, MD, FCCP; B. Taylor Thompson, MD; and Atul Malhotra, MD Study objectives: The role of open-lung biopsy in ARDS has been questioned due to potentially high morbidity and low diagnostic yield. The goals of this study were to better define the frequency of unexpected diagnoses made by open-lung biopsy, the frequency biopsy results lead to a change in clinical management, and the frequency of procedural complications. Design: Case series. Setting: A large tertiary referral center. Patients: All individuals with available records undergoing open-lung biopsy between 1989 and 2000 for evaluation of ARDS based on the American-European Consensus Conference definition. Interventions: None. Measurements and results: The mean age in this cohort of 57 patients was 53 years (SD, 18 years) with PaO 2 /fraction of inspired oxygen ratio of 145 mm Hg (SD, 61 mm Hg) at the time of biopsy. A pathologic diagnosis other than diffuse alveolar damage or fibroproliferation was found in 60% of patients. The most common alternative diagnoses were infection (n 8), alveolar hemorrhage (n 5), and bronchiolitis obliterans organizing pneumonia (n 5). Alternative diagnoses were as frequent in immunocompetent as immunosuppressed hosts (60% vs 59%, respectively). Biopsy results led to a change in management in the majority of patients, with addition of specific therapy in 60% and withdrawal of unneeded therapy in 37%. Although the overall complication rate was 39%, major complications occurred in only 7% of cases. No deaths were attributable to the procedure. Conclusions: In selected patients with clinical ARDS, open-lung biopsy can be performed safely, often reveals an unsuspected diagnosis, and frequently leads to alterations in therapy. (CHEST 2004; 125: ) Key words: ARDS; open-lung biopsy; pulmonary surgical procedures Abbreviations: AECC American-European Consensus Conference; BOOP bronchiolitis obliterans organizing pneumonia; DAD diffuse alveolar damage; Fio 2 fraction of inspired oxygen The American-European Consensus Conference (AECC) definition of ARDS includes the acute onset of bilateral pulmonary infiltrates, a Pao 2 /fraction of inspired oxygen (Fio 2 ) ratio 200 mm Hg, and no evidence of left atrial hypertension. 1 Although this clinical definition is an attempt to limit the pulmonary pathology to diffuse alveolar damage *From the Pulmonary and Critical Care Unit (Drs. Patel, Karmpaliotis, Thompson, and Malhotra), Department of Pathology (Dr. Mark), and Department of Surgery (Dr. Wain), Massachusetts General Hospital, Boston, MA; and Department of Medicine (Dr. Ayas), University of British Columbia, Vancouver, BC, Canada. This work was performed at the Pulmonary and Critical Care Unit, Massachusetts General Hospital. Support was provided by departmental funds. (DAD) in either acute or fibroproliferative phase, numerous underlying histologies can yield this syndrome. Despite these variable etiologies, the role of For editorial comment see page 5 open-lung biopsy in the patient with clinical ARDS remains unclear. Manuscript received July 31, 2002; revision accepted May 13, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( permissions@chestnet.org). Correspondence to: Sanjay R. Patel, MD, 221 Longwood Ave, RFB-486, Boston, MA 02115; spatel@partners.org CHEST / 125 / 1/ JANUARY,

2 Similar to the current state of knowledge regarding the relationship between ARDS and DAD, the clinical syndrome of idiopathic pulmonary fibrosis had at one point been equated with the pathologic diagnosis of usual interstitial pneumonia. It was only after careful pathologic investigation that other diagnoses, such as nonspecific interstitial pneumonia and desquamative interstitial pneumonia, with different natural histories and responses to therapy were found to also present clinically with pulmonary fibrosis. 2 With this knowledge, the importance of a pathologic diagnosis in individuals with the clinical syndrome of idiopathic pulmonary fibrosis has become apparent, making open-lung biopsy an important tool in the evaluation of patients presenting with clinical pulmonary fibrosis. In contrast, the role of open-lung biopsy in the management of ARDS remains unclear. Although guidelines have been suggested for the management of clinically diagnosed ARDS, 3 it is not clear that this syndrome is a homogeneous entity that can be treated uniformly. A number of specific diseases, including infectious pneumonia, bronchiolitis obliterans organizing pneumonia (BOOP), acute eosinophilic pneumonia, and pulmonary capillaritis, can produce the syndrome of ARDS and have specific therapies independent of ARDS. The frequency with which pathologies other than DAD occur in a population of patients with ARDS is unknown. In addition, it is not clear that knowledge of the underlying pathology affects clinical management or outcome. However, in light of data suggesting a therapeutic benefit of glucocorticoid therapy in the fibroproliferative phase of ARDS, 4 knowledge of the underlying lung pathology may have increased importance. Finally, accurate data regarding the risks of openlung biopsy in this critically ill population are needed in order to fully assess the risk/benefit balance of this procedure. The aim of this study was to evaluate the diagnostic utility of open-lung biopsy in a selected population of patients meeting the clinical diagnosis of ARDS, the frequency with which the results of biopsy alter management, and the rate of complications from lung biopsy in a critically ill population. Materials and Methods A retrospective review was performed using inpatient medical records at Massachusetts General Hospital, a tertiary care referral center, over a 12-year period from 1989 through Prior approval for this study was obtained from the appropriate institutional review board. Charts with a discharge diagnosis code of the International Classification of Diseases, Ninth Revision, Clinical Modification 5 suggesting ARDS not related to surgery or trauma, and a procedure code of 3328 (open-lung biopsy), 3402 (exploratory thoracotomy), or 3421 (transpleural thoracoscopy) were reviewed for potential inclusion in this study. The medical records of all patients retrieved from this search were reviewed to determine eligibility for this study. Inclusion criteria consisted of age 18 years at the time of biopsy and findings consistent with the AECC definition of ARDS at the time of the biopsy. 1 Specifically, patients were required to have bilateral pulmonary infiltrates, no clinical evidence of left atrial hypertension, and a Pao 2 /Fio 2 ratio 200 mm Hg within 48 h prior to biopsy. Patients who were not intubated and not receiving noninvasive ventilation at the time of biopsy were excluded since an accurate Fio 2 could not be ascertained. The medical records of all cases were reviewed and data regarding age; sex; comorbidities; dates of hospital admission, intubation, and biopsy; diagnostic tests performed prior to biopsy including BAL; and medications at the time of biopsy were recorded. In addition, laboratory, hemodynamic, and ventilatory parameters at the time of biopsy and over the following 48 h were abstracted from hospital records. Data regarding biopsy findings, changes in therapy, and complications of biopsy were also noted. Major complications were prospectively defined as death, myocardial infarction, stroke, institution of dialysis, or hemothorax within 48 h of surgery. Minor complications were defined as increase in serum creatinine level of 1 mg/dl at 48 h or ipsilateral air leak for 1 week. Statistical Analysis For normally distributed data, values are reported as mean SD. Medians and ranges are reported for all other data. Differences between subgroups were compared with a 2 test or the Fisher exact test if the expected number of events was less than five. Changes in measurements before and after surgery were compared with a paired t test. Statistical analyses were performed using SAS version 8 (SAS Institute; Cary, NC). Results A total of 68 patients underwent open-lung biopsy for evaluation of ARDS during the time period of this review. One chart could not be obtained, one patient had no arterial blood gas studies performed, and nine patients were breathing without ventilatory assistance prior to surgery. Therefore, a total of 57 patients are included in this review. Of these patients, 51 patients (89%) underwent thoracotomy, and the remainder had a thorascopic procedure. Baseline features of these patients are presented in Table 1. Approximately one third of patients were immunocompromised. The cause of immunosuppression was HIV infection in 2 patients, organ transplantation in 4 patients (two liver transplants, one renal transplant, and one bone marrow transplant), and active malignancy in 13 patients. Patients with active malignancy included five patients with neutropenia (absolute neutrophil count 500/ L) at the time of biopsy. The mean value for Pao 2 /Fio 2 ratio at the time of open-lung biopsy was mm Hg with a positive end-expiratory pressure of cm H 2 O. Three patients were receiving inhaled nitric 198 Clinical Investigations in Critical Care

3 Table 1 Population Characteristics* Characteristics Data Age, yr Pao 2 /Fio 2,mmHg Positive end-expiratory pressure, cm H 2 O Male gender 36 (63.2) Immunosuppressed 17 (29.8) BAL prior to biopsy 44 (77.2) Days from admission to biopsy 7 (1 48) Days from intubation to biopsy 3 (0 25) *Data are presented as mean SD, No. (%), or median (range). oxide for refractory hypoxemia at the time of biopsy. Two patients were not intubated prior to biopsy but were receiving continuous positive airway pressure, one at 5 cm H 2 O and the other at 15 cm H 2 O. The lung biopsy results are shown in Table 2. Biopsies revealed a diagnosis other than DAD in 34 patients (60%). Alternative diagnoses were as common in immunocompetent patients as in immunosuppressed ones (60% vs 59%, respectively; p 0.33). The most common diagnoses made after DAD were specific infections (n 8), diffuse alveolar hemorrhage (n 5), BOOP (n 5), and bronchiolitis (n 3). None of the baseline characteristics including age, gender, time to biopsy, and preoperative Pao 2 /Fio 2 ratio were predictive of a pathologic diagnosis of DAD. The specific infectious diagnoses consisted of three cases of cytomegalovirus, two cases of Pneumocystis carinii, two cases of Mycoplasma pneumoniae, and one case each of Staphylococcus aureus, Candida albicans, and influenza B. Preoperative BAL was performed in six of the nine patients with an infectious diagnosis. The BAL results predicted Table 2 Pathologic Diagnoses Diagnoses No. DAD 23 Acute phase 5 Fibroproliferative phase 18 Specific infection 8 Diffuse alveolar hemorrhage 5 BOOP 5 Bronchiolitis 3 Culture-negative purulent pneumonia 2 Drug reaction 2 Pulmonary lymphoma 2 Lymphangitic tumor 1 Organizing pneumonia 1 Desquamative interstitial pneumonia 1 Hypersensitivity pneumonitis 1 Chronic eosinophilic pneumonia 1 Allergic bronchopulmonary aspergillosis 1 Pulmonary edema 1 lung biopsy findings in only two cases. In one case (influenza B), the BAL culture result became positive only after the lung biopsy; in the other case (C albicans), the positive BAL culture result was not thought to represent a true pathogen. The etiologies of the diffuse alveolar hemorrhage cases were Wegener disease, cryoglobulinemia, paraneoplastic vasculitis, idiopathic vasculitis, and hemorrhage related to idiopathic thrombocytopenic purpura. BAL was performed in all five patients. A bloody lavage suggestive of the diagnosis was found in three patients. Overall, BAL was performed prior to the day of surgery in 44 patients (77%). Findings were suggestive of the pathologic diagnosis in a total of six patients. Besides the two infections and three cases of hemorrhage, BAL in a case of chronic eosinophilic pneumonia revealed a WBC count differential of 13% eosinophils. At the time of biopsy, most patients were receiving empiric antimicrobial therapy (52 patient were receiving antibacterials, 8 patients were receiving antifungals, and 7 patients were receiving antivirals). A total of 26 patients were receiving glucocorticoids at the time of biopsy: 7 patients for an underlying nonpulmonary condition, 3 patients for treatment of reactive airways disease, and 16 patients for empiric treatment of ARDS. In addition, four patients were receiving cyclophosphamide for an underlying disorder. Based on the results of the biopsy, 34 patients (60%) had a new therapy begun. Specific antibacterial therapy was begun in two patients, and antiviral therapy was begun in three patients. Steroids were begun in 26 patients, and the dose increased in another patient. Cyclophosphamide was begun in eight patients. In addition, based on the biopsy results, plasmapheresis was begun in a patient with cryoglobulinemia-related alveolar hemorrhage, and thoracic radiation was administered to a patient with pulmonary lymphoma. Results of the biopsy were also instrumental in discontinuing unneeded therapies in 21 patients (37%). Empiric antibacterial therapy was stopped in 17 patients, antifungal therapy was stopped in 2 patients, antiviral therapy was stopped in 2 patients, and steroid therapy was stopped in 3 patients. The operative procedure had little effect on hemodynamics or gas exchange. Mean arterial pressure 24 h postoperatively was not significantly different from preoperative values (mean change, 2.6 mm Hg; p 0.16). Twelve patients had an increase in vasopressor requirements, and 7 patients had a decrease. The Pao 2 /Fio 2 ratio 24 h postoperatively was also not significantly different from preoperative values (mean change, 1.8 mm Hg; p 0.74). CHEST / 125 / 1/ JANUARY,

4 Complications occurred in 22 patients (39%) and are listed in Table 3. Major complications occurred in 7%. The one death occurred in a severely ill patient who was hypotensive and dependent on vasopressors prior to surgery. The patient had an asystolic arrest while being placed on extracorporeal membrane oxygenation and then suffered a terminal arrest 6 h postoperatively. There was no difference in the rate of total or major complications between patients undergoing thoracotomy and thoracoscopy. As a whole, 47% of patients survived to hospital discharge. Survival was not different between patients with DAD compared to alternative diagnoses (52% vs 44%, respectively; p 0.74). Discussion In reviewing our experience with critically ill patients meeting the AECC definition of the ARDS, we have found open-lung biopsy to be useful and safe, often providing a diagnosis not previously suspected. This knowledge frequently led to the institution of specific therapies as well as the discontinuation of unnecessary (and potentially harmful) therapies. Given the retrospective nature of this study, it is impossible to determine whether these changes in therapy resulted in changes in survival. Complications were common after open-lung biopsy. However, most of these were minor. Severe complications were rare but occurred in 7% of patients. Death occurred in only one patient who was severely ill prior to the procedure. The findings that BP and Pao 2 /Fio 2 ratio did not significantly change after surgery suggest that biopsy does not greatly affect hemodynamic or gas exchange physiology. The rate of survival to hospital discharge in this study (47%) is within the range reported in the literature for ARDS. 6 8 This is the largest series of patients undergoing Table 3 Complications* Complications No. (%) Major 4 (7.0) Death 1 (1.8) Hemothorax 2 (3.5) New dialysis 1 (1.8) Minor 18 (31.6) Acute renal failure 6 (10.5) Persistent air leak 12 (21.1) Any 22 (38.6) *All complications were defined as occurring within 48 h of biopsy unless otherwise noted. Defined as a 1 mg/dl rise in serum creatinine level. Defined as an air leak present on the side of biopsy 1 week after surgery. open-lung biopsy for evaluation of clinical ARDS. Our results confirm and extend the findings of previous studies. Canver and Mentzer 9 reported on 27 patients with diffuse pulmonary infiltrates who underwent open-lung biopsy. The proportion of patients satisfying AECC criteria for ARDS is unknown as the study was performed prior to this definition. However, all patients were receiving mechanical ventilation at the time of biopsy. Similar to our findings, 67% of patients had a change in therapy based on results of the biopsy. They reported an overall complication rate of 40%, with the most common complication being a prolonged air leak. Warner et al 10 reported on 80 patients with diffuse pulmonary infiltrates and acute respiratory failure who underwent open-lung biopsy. This study was also performed prior to the AECC definition of ARDS; however, the proportion of patients meeting ARDS criteria was likely small, as only 20 patients were intubated prior to biopsy. In addition, only 14% of the patients underwent bronchoscopy prior to biopsy. The cohort was also unrepresentative of a general ARDS population, as 93% were immunosuppressed. Despite these caveats, it is notable that the findings were consistent with our study. A specific etiologic diagnosis was made in 66% of patients, and biopsy findings led to a change in therapy in 70% of patients. The complication rate was 19%; again, the most common complication was a prolonged air leak. Flabouris and Myburgh 11 reported on a case series of 24 patients with respiratory failure undergoing open-lung biopsy. The proportion of patients with ARDS is unclear: 10 patients (42%) were not receiving mechanical ventilation prior to biopsy, and another 3 patients had a preoperative Pao 2 /Fio 2 ratio 200 mm Hg. As in the study by Warner et al, 10 a large number of patients were immunosuppressed (71%). Again, despite these differences, the results were similar to ours. A specific diagnosis was evident in 46%, and an alteration in therapy based on these findings was made in 75%. As in the present study, there was no significant change in Pao 2 /Fio 2 ratio at 24 h postoperatively. Postoperative complications occurred in 17%, again due primarily to prolonged air leak. Papazian et al 12 reported on a cohort of 36 patients with ARDS as defined by the AECC criteria undergoing open-lung biopsy to exclude superimposed infection prior to the institution of corticosteroid therapy for the treatment of the fibroproliferative phase of ARDS. Biopsies were therefore done later in the course of illness, with a median of 14 days from intubation to biopsy in that study compared to 7 days in our study. Given the differing goals of the two studies, the results of biopsies are not directly comparable. As in our study, however, open-lung 200 Clinical Investigations in Critical Care

5 biopsy was found to be well tolerated with no worsening of Pao 2 /Fio 2 ratio, and a complication rate of 19%. In contrast to the prior studies (with the exception of that by Papazian et al 12 ), our report included only patients who met the AECC definition of ARDS. In addition, this is the largest study to address the safety and utility of open-lung biopsy in patients receiving mechanical ventilation. Finally, the majority of patients in this study had undergone BAL prior to surgery, thus ensuring that less invasive diagnostic methods had been exhausted. Without a randomized trial, it is impossible to definitively conclude whether open-lung biopsy has any effect on patient outcome. However, there are several reasons to believe lung biopsy can play a useful role in the management of patients with clinically defined ARDS. Our series demonstrates that lung biopsy can be done with relatively low morbidity in this critically ill population. The high percentage of therapeutic changes that were made as a result of the biopsy procedure suggests lung biopsy does contribute information that is useful to clinicians in decision making. Given the large number of diagnoses uncovered that have effective treatments, we believe that the changes in therapies we documented were meaningful ones that have the potential to improve outcome. There are however several limitations that should be noted. First, the results of this study cannot be generalized to all patients with clinical ARDS. During the time period of this study, a total of 1,707 patients had an International Classification of Diseases, Ninth Revision, Clinical Modification 5 diagnosis of (ARDS, not elsewhere classified). Assuming all of these patients met AECC criteria for ARDS, this would suggest a biopsy rate of only 4%, and even accounting for the possibility of misclassification, it is unlikely the proportion was 10%. In addition, because the decision to pursue lung biopsy was not made at random, the patients referred for biopsy were unlikely to be a representative sample of our ARDS population. Those with clinical features unusual for ARDS or who had failed to respond to standard treatment would be more likely to undergo biopsy. This selection bias would be expected to increase the probability of an alternative diagnosis. Another limitation is that due to its retrospective nature, our study cannot directly address the question of whether open-lung biopsy leads to a change in mortality or length of hospitalization. However, knowledge of a specific etiology would allow for initiation of specific treatment assuming such treatment is available. Many of the diagnoses discovered in this report (such as infectious pneumonia, chronic eosinophilic pneumonia, Wegener disease, and BOOP) have therapies that have been well established to have a positive effect on outcome. In addition, as our understanding of ARDS and therapeutic options improve, the ability of knowledge of the underlying etiology to positively affect outcome will only increase. For example, if corticosteroid therapy becomes standard of care for the management of fibroproliferative ARDS, then identifying those cases with unrecognized opportunistic infections may become more important. In addition, the withdrawal of unneeded therapy, as demonstrated in the ventilator-associated pneumonia literature, 13 may improve patient outcome. Our 12 years of experience suggests that clinically defined ARDS represents a heterogeneous collection of pathologic processes. Open-lung biopsy in a selected population of patients with clinical ARDS often provides a specific etiologic diagnosis. This information frequently leads to institution of specific therapy and also often allows for the withdrawal of potentially harmful treatments. We have also demonstrated that open-lung biopsy can be performed in a critically ill population with an acceptable rate of morbidity. Therefore clinicians should consider open-lung biopsy in patients with clinical ARDS if a cause of ARDS is not readily apparent. Further research should investigate the effect of lung biopsy on outcome in ARDS. References 1 Bernard GR, Artigas A, Brigham KL, et al. The American- European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149: Bjoraker JA, Ryu JH, Edwin MK, et al. Prognostic significance of histopathologic subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998; 157: Artigas A, Bernard GR, Carlet J, et al. The American- European Consensus Conference on ARDS, part 2: Ventilatory, pharmacologic, supportive therapy, study design strategies, and issues related to recovery and remodeling. Am J Respir Crit Care Med 1998; 157: Meduri GU, Headley AS, Golden E, et al. Effect of prolonged methylprednisolone therapy in unresolving acute respiratory distress syndrome: a randomized controlled trial. JAMA 1998; 280: Department of Health and Human Services. The International Classification of Diseases, 9th Revision, Clinical Modification: ICD-9-CM Vols. 1 and 3, 2nd ed. Washington, DC: Department of Health and Human Services, Milberg JA, Davis DR, Steinberg KP, et al. Improved survival of patients with acute respiratory syndrome (ARDS): JAMA 1995; 273: Amato MBP, Barbas CSV, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998; 338: The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with tradiwww.chestjournal.org CHEST / 125 / 1/ JANUARY,

6 tional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342: Canver CC, Mentzer RM Jr. The role of open lung biopsy in early and late survival of ventilator-dependent patients with diffuse idiopathic lung disease. J Cardiovasc Surg (Torino) 1994; 35: Warner DO, Warner MA, Divertie MB. Open lung biopsy in patients with diffuse pulmonary infiltrates and acute respiratory failure. Am Rev Respir Dis 1988; 137: Flabouris A, Myburgh J. The utility of open lung biopsy in patients requiring mechanical ventilation. Chest 1999; 115: Papazian L, Thomas P, Bregeon F, et al. Open-lung biopsy in patients with the acute respiratory distress syndrome. Anesthesiology 1998; 88: Fagon JY, Chastre J, Wolff M, et al. Invasive and noninvasive strategies for management of suspected ventilator-associated pneumonia: a randomized trial. Ann Intern Med 2000; 132: Clinical Investigations in Critical Care