Do We Really Know the Value of Surveillance Lung Biopsies?

reviewed in CHEST, 8 and a call for an analysis such as performed by Rennard et al was made. Now we have one, and the roles oflong-term pharmacologic therapy in CO PD needs further prospective study in important subsets of COPD, including early disease, as in the LHS patients with rapidly declining pulmonary function, and those at the end of a rapid decline such as the subjects in the study by Rennard et al. For the moment, in patients with advanced COPD, ipratropium appears to be the bronchodilator for regular use with a sustained benefit, little evidence of long-term adverse effects, and possibly some benefits beyond acute bronchodilation. Mark E. Deffebach, MD Portland, Oregon From the Portland VA Medical Center. REFERENCES

1 Anthonisen N, Connett J, Kiley J, eta!. Effects of smoking intervention and the use of an inhaled anticholinergic bronchodilator on the rate of decline of FEVJ. JAMA 1994; 272:1497-505 2 Gross N. The lung health study: disappointment and triumph. JAMA 1994; 272:1539-41 3 COMBNENT Inhalation Aerosol Study Group. In chronic obstructive pulmonary disease, a combination of ipratropium and albuterol is more effective than either agent alone: an 85-day multicenter trial. Chest 1994; 105:1411-19 4 Petty T. The combination of ipratropium and albuterol is more effective than either agent alone. Chest 1995; l07:183S-6S 5 Burrows B, Bloom J, Traver G, eta!. The course and prognosis of different forms of chronic airways obstruction in a sample from the general population. N Eng! J M ed 1987; 317:1309-14 6 van Schayck C, Dompeling E, van Herwaarden C, et a!. Bronchodilator treatment in moderate asthma or chronic bronchitis: continuous or on demand? a randomised controlled study. BJM 1991; 303:1426-31 7 Vollmer W, Johnson L, Buist S. Relationship of response to a bronchodilator and decline in forced expiratory volume in one second in population studies. Am Rev Respir Dis 1985; 132: 1186-93 8 Ziment I. The f)-agonist controversy: impact in COPD. Chest 1995; 107:198S-205S

Do We Really Know the Value of Surveillance Lung Biopsies?

Jread with interest the report by Baz and colleagues

in this issue of CHEST (see page 84) on the utility of bronchoscopies after lung transplantation at Duke University. Ibelieve that four important issues need to be emphasized from this study. First, we cannot be assured that there is an absolute need for surveillance biopsies. The lines are drawn, and opposing opinions today are not closer to resolution now than they were 3 to 4 years agoJ- 3 The issue gets cloudier when you include more recent reports from two of the largest programs in the world: The Papworth Program, leaning away from surveillance, and the 6

University of Pittsburgh, leaning toward it. 4•5 Next, at this time, we all need to concern ourselves with costs-and bronchoscopic biopsies are expensive. Surveillance biopsies may potentially cover costs by providing diagnosis of asymptomatic rejection and bronchiolitis obliterans syndrome (BOS). But it is not clear that early diagnosis today prevents BOS. To that end, the third and fourth issues become concerns-what is the evidence that surveillance biopsies prevent disease or improve survival? Currently, this evidence is lacking, even when one considers the unsuspected asymptomatic diagnoses made by surveillance biopsies .I·5 Part of the reason that the surveillance approach offers no clear advantage is that the old medical saw, "a high index of suspicion," leads to a similar number of biopsies per patient in the first 1 to 2 years in programs that biopsy for indication rather than surveillance (informal survey). What emerges as the greatest determinant of the frequency of biopsy is the number of clinical events in the first few months after transplantation. Essentially, those who are eventprone continue to have episodes that lead to more bronchoscopies and biopsies. Baz and colleagues address this very issue indirectly by showing, in a relatively small number of patients, that freedom from rejection or cytomegalovirus in the first 4 months portends an excellent outcome, for the duration of their observation. The controversy, however, continues because it is not clear that patients with early events would not have been diagnosed anyway by development of clinical signs within a short period. 3 The main value from the current paper may be the early identification of a group that can escape further biopsies by remaining event-free shortly after transplantation. Also of note, the results from Banda et al6 show that patients diagnosed early in the course of chronic rejection had a higher likelihood of achieving remission with aggressive therapy than did those with more advanced disease. This may lead us toward more surveillance. I, like the authors, urge that a definitive prospective study be conducted. It should be multicenter and designed to address the major issues: (1) the value of surveillance biopsy early after transplantation; (2) the cost benefit of surveillance biopsies; (3) the prevention of BOS; and (4) the improvement in survival. Edward R. Garrity, Jr., MD, FCCP Maywood, Illinois Associate Professor of Medicine and Medical Director, Lung Transplantation Program, Loyola University Medical Center. Reprint requests: Dr. Garrity, 2160 South First Atoenue, Maywood IL 60153 '

REFERENCES

1 Trulock EP, Ettinger NA, Brunt EM, et a!. The role of Editorials

2

3

4

5

6

transbronchial lung biopsy in the treatment of lw1g transplant recipients: an analysis of200 consecutive procedures. Chest 1992; 102:1049-54 Garrity ER, McCabe M, Husain A, et a!. The utility of transbronchial biopsies after isolated lung transplantation. Chest 1992; 102:73$ Solans EP, Garrity ER, McCabe M, eta!. Early diagnosis of cytomegalovirus pneumonitis in lung transplant patients. Arch Pathol Lab Med 1995; 119:33-5 Tamm M, Sharples LD, McNeil K, et a!. Effect of surveillance biopsies on the development of OB following heart-lung transplantation. Am J Respir Crit Care Med 1995; 151:A89 Guilinger RA, Paradis II, Dauber JH, et a!. The importance of bronchoscopy with transbronchial biopsy and bronchoalveolar lavage in the management of lung transplant recipients. Am J Respir Crit Care Med 1995; 152:2037-43 Bando K, Paradis IL, Similo S, et a!. Obliterative bronchiolitis after lung and heart-lung transplantation: an analysis of risk factors and management. J Thorac Cardiovasc Surg 1995; 110:4-14

Evaluating Pleural Fluid Most clinicians agree that classifYing pleural effusions as transudates vs exudates should be the first step in the laboratory evaluation of pleural effusions. The major objective is to identifY patients with exudates who require further diagnostic evaluation, especially for conditions with potentially preventable morbidity, such as tuberculosis, complicated parapneumonic effusions, pulmonary emboli, connective tissue diseases, or treatable metastatic malignancies. The converse of this objective is to identifY patients with transudates to avoid further costly evaluation. 1 Since 1972, Light's criteria have been used most commonly for this determination. 2 Despite reports of lower specificity than in the original data of Light and colleagues, the sensitivity for identifYing exudates ranged from 94 to 100%, with an average of 98% in 8 separate patient groups totalling 1,684 patients. 3-10 This high sensitivity is to be expected from the combination of multiple tests where any positive result is categorized as disease. Patients with pleural disease that is considered exudative but is classified as transudative by Light's criteria are very rare and frequently break the oslerian rule of "one patient, one disease." That is, they have both a transudative and exudative etiology for their effusion, such as congestive heart failure and malignancy.2.10 Further laboratory evaluation of pleural effusions classified as transudates by Light's criteria has an extremely low yield. 10 Unless the clinical history or examination findings are suspicious for an exudative process, it is appropriate and cost efficient to abandon any further laboratory testing on effusions classified as transudates. In my experience, this is almost never done. Instead, cell count, pH, glucose, amylase, Gram's stain, bacterial culture, acid-fast bacillus smear, tuber-

culosis culture, fungal smear, fungal culture, and even cytologic studies are ordered before the results of lactate dehydrogenase (LDH) and total protein are known. Stopping this practice would save more money than making the criteria more specific. In trying to improve specificity, a number of alternative criteria have been proposed, including the pleural fluid cholesterol, 6 the serum-effusion albumin gradient,8 and the combined pleural fluid cholesterol and LDH. 3 Although the latter criteria have the best reported combination of sensitivity and specificity (99 and 98%, respectively), they lack the track record of sensitivity that Light's criteria provide. 3 Of specific concern is that the fluid LDH greater than two thirds of the upper normal serum limit had the lowest sensitivity (66%) in a group of 297 patients reported by Romero et al.5 Contrary to the accompanying editorial11 to Costa et al, 3 I do not think we should abandon Light's criteria based on one publication.l 1 Once established, criteria should be prospectively confirmed on an independent, random population of patients. The report in this issue of CHEST (see page 97) by Garcia-Pachon and colleagues on the serum cholinesterase ratio for separating transudates and exudates should be evaluated with guarded optimism. As with Light's criteria, a simultaneous sample of both pleural fluid and serum improved the predictive value of the test and normalized the data from two different hospitals where two different assays were used. Using a cutoff level of 0.23 for the pleural fluid to serum cholinesterase ratio (all ratios greater than 0.23 are exudative), a sensitivity and specificity for exudates of 100 and 94%, respectively, were achieved. The sensitivity and specificity of Light's criteria were 97 and 74%, respectively. It would be interesting to know how the combination of pleural fluid cholesterol and LDH criteria performed, but those data are not available in the manuscript. Interesting data on tl1e results of the criteria in undiagnosed pleural effusions are available. Only one previous report by Peterman and Speicher10 provides similar data. In both reports, the majority of undiagnosed pleural effusions are classified as exudates using Light's criteria (30 of 32 in the present report and 56 of 62 in the Peterman and Speicher10 report). Because most of these patients would then get further testing and!or follow-up, the safety of the two-step laboratory approach (where the second step of further testing on effusions classified as transudates is abandoned) is enhanced. Clinical characteristics that have been associated with tuberculosis or malignancy in these patients include weight loss, fever, positive purified protein derivative, large effusions, and pleural fluid lymphocytosis.12 Aside from the relative obscurity of the serum cholinesterase in pleural effusion, the excellent perforCHEST/110/1/JULY, 1996

7