Early and Late Morbidity in Patients Undergoing Pulmonary Resection With Low Diffusion Capacity

Transcription

1 Early and Late Morbidity in Patients Undergoing Pulmonary Resection With Low Diffusion Capacity Michael Bousamra II, MD, Kenneth W. Presberg, MD, Joseph H. Chammas, MD, James S. Tweddell, MD, Barry L. Winton, MD, Mark R. Bielefeld, MD, and George B. Haasler, MD Department of Cardiothoracic Surge D' and Divisions of Pulmonary and Critical Care Medicine and Surgery, Medical College of Wisconsin, Milwaukee, Wisconsin Background. We sought to determine whether low diffiusion capacity of the lung to carbon monoxide (DLCO) is a predictor of high postoperative mortality and morbidity after major pulmonary resection and whether major pulmonary resection in patients with low DLCO results in substantial long-term morbidity. Methods. Sixty-two major pulmonary resections were performed in 61 patients with low DLCO (DLCO _<60% predicted for pneumonectomy or bilobectomy; _<50~ predicted for lobectomy). Contemporaneously, 262 other patients underwent 263 major pulmonary resections (group IIL Long-term morbidity was assessed in subsets of patients with low (n = 24) and high (n = 22; DLCO >60~ predicted) DLCO. Results. The hospital mortality rates were equivalent (4.8 7c low DLCO versus 4.9~4 group I), whereas respiratory complications were more frequent in patients with low DLCO (18% versus 9.5%; p = 0.05). In the subgroup analyses, patients with low DLCO had more hospitalizations for respiratory compromise and worse median dyspnea scores. Analysis of patients with substantial dyspnea revealed an association with extended pulmonary resection and postoperative radiation therapy in patients with low DLCO. Conclusions. Patients with low DLCO underwent major pulmonary resection with a low mortality rate and an acceptable, but increased, respiratory complication rate. Long-term respiratory morbidity was increased in patients with low DLCO; however, the extent of pulmonary resection and the use of postoperative radiation therapy may have contributed to the development of dyspnea in these patients. (Ann Thorac Surg 1996;62:968-75) O ver the past 40 },ears, progressive refinement in patient selection strategies has reduced the morbidity and mortality of major pulmonary resection [1, 2]. A minimum forced expiratory volume in 1 second (FEV~) of 1 L for lobectomy and a predicted postoperative FEV~ of 0.8 L after pneumonectomy are frequently quoted cutoff points for pulmonary resection [3, 4]. Pulmonary reserve also has been assessed by maximal voluntary ventilation [5, 6], maximal oxygen consumption [7, 8], and exercise desaturation studies 19], with improvement in the prediction of adverse outcomes after major pulmonary resection. Recent experience with lung volume reduction underscores the need to assess further the functional capacity of the involved lung, as removal of severely emphysematous tissue may be well tolerated in patients with advanced chronic obstructive pulmonary disease 110]. Less well studied, but also of importance, are the long-term morbidities related to pulmonary resection and the preoperative factors associated with their development. Chronic dyspnea is one such disabling syrup- Presented at the Thirty-second Annual Meeting of The Societx ot[tloracic Surgeons, Orlando, FL, Jan 29-31, Address reprint requests to Dr Bousamra, Department of Cardiothoracic Surgery, Medical College of Wisconsin, 9200 W Wisconsin Ave, Milwaukee, WI tom, which can result from pulmonary resection in marginal patients. In several pulmonary diseases, reduction in the diffusion capacity of the lung to carbon monoxide (DLCO) has been associated with exertionat dyspnea, pulmonary hypertension, and shorter life expectancy ]11-14]. The DLCO has also been reported to be an independent predictor of postoperative morbidity and mortality after major pulmonary resection [15]. In this retrospective study, we sought to determine whether patients with low DLCO were at increased risk of morbidity and mortality in both the perioperative period and during long-term follow-up after major pulmonary resection. Material and Methods We reviewed all adult patients at the Zablocki Veterans Administration Hospital and the Milwaukee County Medical Center who underwent standard lobectomy, bilobectomy, or pneumonectomy for nontraumatic diagnoses from January 1, 1991 to July 31, Case information spanning 26 perioperative variables was entered into a continually updated data base (Table 1). These data were supplemented by chart review as necessary. The single-breath DLCO was measured according to the method of Gaensler and Wright [16] and adjusted to correct for anemia, as described previously [17]. The 1996 bv The Society of Thoracic Surgeons /96/$15.00 Published by Elsevier Science lnc Pll S (96)

2 Ann Thorac Surg BOUSAMRA ET AL ;62: LOW DLCO IN PULMONARY RESECTION Table 1. Comparison of Perioperative Factors Between Patients With Low Diffusion Capaci~ of the Lung to Carbon Monoxide and Group II Patients" Low' DLCO Group II Characteristic (n 62) (n 263) p Value Demographic Age <0.01 Sex (male/female) 48/14 181/82... Smoking Current 40% (23) 33% (81)... Pack-years <0.001 Medications Inhaled BD 19% (11) 12% (31)... Oral BD 9% (5! 6% (16)... Cardiac 26% (15) 24% (58)... Steroids 1.6'!,, (1) 9"i, (15)... Symptoms Fatigue 26% (15) 23% (56)... Hemoptysis 19% (11) 17% (43)... Weight loss 22% (13) 16% (40)... Dyspnea 40% (23) 26% (65) <0.05 Cough 35% (20) 35% (86)... Chest pain 12% (7) 11% (27)... Past medical histo W Pulmonary operation 5". (3) 7% (17)... Cardiac disease 22% (13) 18% (44)... Pulmonary function DLCO (% predicted) ~ 16 <0.01 FVC (L} * 0.93 <0.01 FVC % 80 ~ ~. 17 <0.01 FEV~ (k) <0.01 FEV~ ". 74 ~ <0.01 PO~ (mm Hg) 76 ~ 9 77 * 9... PCO2 (ram Hg) ? 4... Activity level <0.05 (Karnofsky score) Lung cancer stage % (29) 68.5% (161) <0.001 li 10.5% (6) 6% (14)... III (A,B) 36.8% (21) 21.7% (51)... IV 1.6% (1) 3.8% (9)... Procedure Lobectomv 60% (37) 78", (206) <0.02 Bilobectomv 8% (5) 4.2% (11 )... Pneumonectomv 32% (20) 18% (46)... " Values are mean ~ standard deviation. Numbers in parentheses indicate absolute number of patients. BD bronchodilatur; DLCO diffusion capacity of lung to carbtm monoxide; FEV~ forced expiratory volume in 1 second; vital capacity; PCO z partial pressure of carbon dioxide; PO~ partial pressure of oxygen. FVC forced forced vital capacity and the FEV I were measured by a Collins DS Plus Spirometer System (W.E. Collins lnc, Braintree, MA). Patients were divided into two groups related to their DLCO: Patients with low DLCO were defined as having a DLCO of 60% or less predicted if undergoing pneumonectomy or bilobectomy and 50% or less predicted if undergoing lobectomy; the remaining patients had a higher or undetermined DLCO value (group II). Predicted postoperative values for the per- centage of predicted DLCO, FEV v percentage of predicted FEV 1, forced vital capacity, and percentage of predicted forced vital capacity were calculated by multiplying the index by the fraction of functional lung segments remaining after resection. The number of lung segments was assigned as described previously [5, 18]: right upper lobe 3; right middle lobe - 2; right lower lobe = 5; left upper lobe - 5; left lower lobe - 4. Performance status was assessed by the Karnofsky score [19].

3 970 BOUSAMRA ET AL Ann Thorac Surg LOW DLCO ln PULMONARY RESECTION 1996;62: Table 2. Mortality and Complications in Patients With Low Diffusion Capacity of the Lung to Carbon Monoxide and Group II Patients" Result Low DLCO Group II Hospital death 4.8% (3) 4.9% (13) Complication Respiratory" 18% (11) 9.5% (25) b Pleural space 10", (6) 5.7% (15) ~ Cardiac failure 0 (0) 2",, (5) Arrhythmia 5% (3) 3.3% (8) Pulmonary embolism 0 (0) 0.8% (2) Other 15% (9) 13% (33) a Numbers in parentheses indicate absolute number uf patients p not significant. DLCO - low diffusion capacity of lung to carbon monoxide. t~ p = Postoperative events are listed for patients with low DLCO and group II patients in Table 2. Respiratory complications included atelectasis, pneumonia, and adult respiratory distress syndrome. Atelectasis was considered substantial if it was lobar or if it prompted reintubation, bronchoscopy, transfer to the intensive care unit, or prolongation of the intensive care unit stay. Pneumonia was defined as the presence of a focal or diffuse infiltrate associated with pathologic organisms on sputum culture. The diagnosis of adult respiratory distress syndrome was reserved for cases with respiratory compromise associated with diffuse infiltrates, without identification of causative bacteria and in the absence of congestive heart failure. Prolonged air leak was defined as lasting for more than 10 days. Cardiac complications included cardiac failure requiring inotropic support, myocardial infarction confirmed bv enzymatic and electrocardiographic evidence, and arrhythmias resulting in hemodynamic compromise. Pulmonary embolism was documented by high-probability v/q scan, pulmonary arteriogram, or autopsy. Miscellaneous complications included renal insufficiencv requiring dialysis, stroke, prolonged ileus, intestinal ischemia, wound infection, and sepsis. Each of these complications individually occurred in less than 2% of patients. All available patients with low DLCO (n - 24) and a cohort of 22 group I patients with known DLCO greater than 60% of that predicted (high DLCO) were interviewed 6 months or more after major pulmonary resection to assess long-term respiratory morbidity. This cohort of group II patients was preselected to have age and spirometry data similar to those of their low-dlco counterparts. All patients were free of recurrent cancer. Dyspnea was graded by the Modified Medical Research Council of Great Britain dyspnea scale [20] (Appendix 1) and by the baseline dyspnea scale [21] (Appendix 2). The change in dyspnea related to major pulmonary, resection was graded by the transitional dyspnea index [20] (see Appendix 2). Negative changes in the transitional dyspnea index were assigned the modifiers "mild" (-1), "moderate" (-2), and "major" ( 3) to describe the deterioration in dyspnea. Patients were also questioned re- garding their level of dyspnea before operation to obtain a retrospective baseline dyspnea score, which was compared with the current score. Hospitalization for respiratory compromise was assessed by patient interview and confirmed by chart review. Continued supplemental oxygen use 1 month after discharge was noted. Survival comparison for lung cancer patients was made between group I (low DLCO) and group II patients using a Kaplan-Meier method of analysis. There were 41 group I patients and 172 group II patients with current data permitting analysis of survival and disease status. Comparisons between patients with low DLCO and group II patients were made by an unpaired Student's t test for continuous variables and by )(2 analysis for indicator variables. Dyspnea scores within groups were compared by the Wilcoxon signed rank test; scores between groups were compared by the Wilcoxon rank sum test. Missing data for individual variables were not included in the total number when calculating percentages. Results A total of 325 major pulmonary resections were performed in 323 patients: 243 lobectomies, 16 bilobectomies, and 66 pneumonectomies. Two hundred ninetytwo patients had lung cancer, and 33 had either benign disease or metastatic cancer to the lung. The criteria for Iow-DLCO classification (group I) were met in 62 cases, whereas the remainder (n 263) were classified as group II cases. Preoperative variables were compared between patients with low DLCO and group II patients (see Table 1). Patients with low DLCO were older, had a more extensive smoking history, were more likely to have had preoperative dyspnea, and had worse general health. Both measured and fractional spirometric values were lower in patients with low DLCO. Mean DLCO values were 44% _+ 9% in patients with low DLCO and 79% + 16% in group II patients. Patients with low DLCO underwent pneumonectomy or bilobectomy more frequently than their group II counterparts. Lung cancer was the predominant indication for resection in both groups. Hospital mortality rates were equivalent between the patient groups. Patients with low DLCO experienced a higher respiratory complication rate (18% versus 10%), and pleural space problems were twice as frequent among these patients (see Table 2). Individually and combined, other complication rates were similar between patients with low DLCO and group II patients. Among tung cancer patients, the stage of disease was more advanced in the low-dlco group (see Table 1). Long-term survival was similar between the groups, although there was a constant trend toward poorer survival in patients with low DLCO (Fig 1). The median survival was 41 months in patients with low DLCO and 45 months in group li patients. Twenty-six percent of the late deaths (19 of 71) were unrelated to cancer in group II, compared with 41% (7 of 17) in the Iow-DLCO group (p = not significant). Twenty-four patients with low DLCO and 22 group li

4 Ann Thorac Surg BOUSAMRA ET AL ;62: LOW DLCO IN PULMONARY RESECTION i 0,9 ] 0.8 -_. == o.7 ]~-- i '6 o.s~ ~ 0.4-.o a. 0.3 O.i ~ LOw DLCO~ High DLCO ~P!!!!!,,,,!~<g'... UL'4...,~q!,,,,... V,V..., Months post operauon Fig 1. Surt,ival among patients with low and high diffusion capacity of the lung to carbon monoxide (DLCO) with lung cancer. Median survival was 41 months for low DLCO; 45 months for high DLCO. patients with DLCO greater than 60% {high DLCO) were interviewed at least 6 months after resection to assess chronic respiratory compromise (Table 3). The subgroups were similar with respect to age and predicted postoperative spirometric values. Predicted postoperative DLCO was 58% in patients with high DLCO and 3~% in those with low DLCO (p < 0.01). Hospitalizations for respiratory compromise were significantly more frequent in patients with low DLCO. Three patients with low DLCO Table 3. Comparison qf Patients With Low and High Dif)h sion Capaci~ of the Lung to Carbon Monoxide Assessed for Chronic Respirato~ Morbidi~" Low DLCO High DLCO Measurement (n : 24) (n = 22) DLCO ('h,) ~'c (L) F%'C % FEV 1 (L) FEV~ 'I, Age (y) Percentage receiving postoperative radiation therapy Percentage undergoing pneumonectomy or bilobectomy Hospitalizations Home oxygen b ~ ~ :~ = z ~ b'~ 1 3 0,1 Values are mean * standard deviation. Diffusion capacity of the lung to carbon monoxide and spirometric data are given as predicted postoperative values, l~ p, ~ Eleven hospitalizations in 5 patients. Abbreviations are as in Table l. a I Low DLCO,4 I I I ] HighDLCO 4 a ~ ~2 2 ~, o o~ postoperative MMRC dyspnea score t pre- vs. postoperative (BDS) Fig 2. (a) Modified Medical Research Council (MMRC) median dyspnea scores for patients with low and high diffusion capacity of the lung to carbon monoxide (DLCO) after resection: Low DLCO is greater than high DLCO (p < by Wilcoxon rank sum test). (b) Preoperative and postoperative median baseline dyspnea scores (BDS).for patients with low and high DLCO. Post-low DLCO is less than pre-low DLCO (p < 0.01 by Wilcoxon signed rank test). required supplemental oxygen after resection, whereas all 22 patients with high DLCO were free of supplemental oxygen. Dyspnea assessed by the Modified Medical Research Council dyspnea scale was worse in patients with low DLCO than in their high-dlco counterparts (Fig 2a). In addition, pulmonary resection resulted in a significant decline in the median baseline dyspnea scores only among patients with low DLCO (Fig 2b). The transitional dyspnea index was used to identify patients with clinically significant worsening of dyspnea related to pulmonary resection (index scores of -2, -3) (Table 4). Seven patients with low DLCO experienced moderate (n = 4) or major (n = 3) worsening of dyspnea, compared with 2 patients with high DLCO who had moderate declines. These 9 patients with moderate or major worsening of dyspnea underwent more extensive pulmonary resections (bilobectomy, pneumonectomy; n = 4) and were more likely to have received postoperative radiation therapy (n - 4). Analysis by treatment modality showed that 3 of 6 patients with low DLCO and 1 of 7 patients with high DLCO who underwent bilobectomy or pneumonectomy had moderate or major deterioration of dyspnea. Similarly, 3 of 6 patients with low DLCO and I of 2 patients with high DLCO who had postoperative radiation therapy reported marked worsening of dyspnea. There was overlap within these subsets, as 3 patients underwent extended resection followed by radiation therapy; all experienced substantial worsening of dyspnea. Comment The data presented indicate that patients with low DLCO can undergo major pulmonary resection with a low mortality rate and an acceptable, though increased, incidence of respiratory complications. This result is notable in that patients with low DLCO also underwent pneumo- b

5 972 BOUSAMRA ET AL Ann Thorac Surg LOW DLCO IN PULMONARY RESECTION 1996;62: Table 4. Comparison qf Patients With and Without Substantial Deterioration in Dyspnea" Related to Major Pulmonary Resection ~' Deterioration in Radiation Bilobectomy, Dyspnea FVC (L) ~ FEV~ (L) ~ therapy Pneumonectomy Low/High DLCO d Yes (n 9) 1.84 ± % 44% 7/2 No (n 37) % 24'!', 17/20 '~ Transitional dyspnea index score of 2 or less. b Values are mean standard deviation. Predicted postoperative values, d Number of patients. Abbreviations are as in Table 1. nectomy more frequently and as a group were older, smoked more, and had more advanced cancer. We accepted the premise forwarded by Ferguson and colleagues [15] that these patients were at higher risk for lobectomy or pneumonectomy. In turn, patients with low DLCO were selected with relatively well-preserved spirometrv values (80% predicted forced vital capacity, 74% predicted FEV1). We also maintained a policy of longer observation in an intensive care setting, generally lasting 2 to 3 days. This program was reinforced by the observation in our institution that most fatal complications after major pulmonary resection were respiratory related and that the initial compromising insult usually occurred within the first 3 postoperative days. In addition to preoperative patient selection and prolonged postoperative intensive care, we advocated aggressive pain management using intrathecal or epidural agents. Together these measures may have reduced the incidence of respiratory complications and promoted earlier recognition of their presence, thus normalizing the mortality rate in patients with low DLCO. We hypothesized that because patients with low DLCO were subjected to greater respiratory morbidity, their survival would also be decreased because of related or comorbid factors. However, long-term survival was equivalent between the patient groups despite a higher stage distribution in patients with low DLCO. The incidence of death from causes other than recurrent cancer tended to be higher in patients with low DLCO, but the small populations prevented firm conclusions in this regard. Patients with low- DLCO experienced greater chronic respiratory morbidib, after major pulmonary resection. Specifically, major pulmonary resection led to worsening of the sense of dyspnea only in these patients. Patients with high DLCO, on average, remained stable. Radiation therapy and more extensive pulmonary resection may have had an additive effect on the worsening of dyspnea. Although the small sample size prohibited statistical significance, these patterns of respiratory decline in patients with lout DLCO subjected to extensive pulmonary resection with or without chest radiation therapy make intuitive sense and should be considered when counseling patients regarding therapeutic options. The data are also supported by experience in lung cancer patients undergoing primary radiation therapy. Abratt and Willcox [22] recently demonstrated worsening of clinical dyspnea scores in patients after chest radiation therapy for inoperable lung cancer. Associated with this finding was a 14% reduction in DLCO 6 months after radiation therapy. Choi and Kanarek [23], in a similar study of pulmonary function in patients undergoing radiation therapy for inoperable lung cancer, found that pulmonan' function scores were reduced by 22% in patients with higher baseline FEV, values (FEV~ ->50% predicted), whereas half of patients with lower baseline FEV 1 values actually had a modest improvement in pulmonary function 6 months after radiation therapy. With regard to the extent of pulmonary resection, Pelletier and associates [24] demonstrated that patients who underwent pneumonectomy were more likely to experience dyspnea and exercise desaturation after resection than were lobectomy patients. Patients with low DLCO have less functional reserve to tolerate eider more extensive pulmonary, resection or postoperative radiation therapy, and thus the prevalence of symptoms in this subgroup is not surprising. On the other hand, the majority of these patients did not experience marked worsening of dyspnea after pulmonary resection. In fact, postoperative dyspnea indices were equivalent between the patients with high and low DLCO if patients having radiation therapy and extended resection were excluded. Our findings confirm those of Ferguson and colleagues [15] that a reduction in DLCO is a predictor of respiratory complications after pulmonary resection. The DLCO may act as an independent variable with respect to other pulmonary function tests. The DLCO reflects the capillary surface area available for gas diffusion across the alveolus and thus indicates the lungs' ability to oxygenate blood. In scleroderma, a reduction in DLCO is associated with pulmonary hypertension, exercise desaturation, and reduced survival I13, 25]. Among patients with chronic obstructive pulmonary disease, Owens and coworkers [14] confirmed the association of exercise desaturation with reduction in DLCO. Because chronic obstructive pulmonary disease is the principal cause of low DLCO among lung cancer patients, major pulmonary resection may exacerbate exercise desaturation, resulting in exertional dyspnea. Likewise, this loss of reserve may lead to complications or death in the perioperative period. For individuals undergoing major pulmonary resection, particularly lung cancer patients, the struggle re-

6 Ann Thorac Surg BOUSAMRA ET AL ;62: LOW DLCO IN PULMONARY RESECTION mains to provide an "optimal" curative resection against the mounting risk of morbid and fatal postoperative events. The decision-making process is further complicated by the availability, of modern radiotherapeutic techniques and parenchyma-conserving pulmonary resections, which offer high-risk patients reasonable alternatives for lung cancer treatment. Thus, the quest to optimally assess perioperative risk continues. The DLCO measurement adds to the surgeon's armamentarium in evaluating patients for pulmonary resection and sheds light on the clinical course of these patients in the perioperative period and during long-term follow-up. Our data show that these patients can undergo major pulmonary resection with an equivalent mortality rate and an increased, but acceptable, respiratory complication rate. Chronic dyspnea and hospitalizations for respiratow decompensation are more common among patients with low DLCO; however, the development of substantial dyspnea in these patients is probably also dependent upon the extent of pulmonary resection and the use of postoperative chest radiation therapy. References 1. Gaensler EA, Cusell DW, Lindgren I, Verstraeten JM, Smith SS, Streider JW. The role of pulmonary insufficiency in mortality and invalidism following surgery for pulmonary tuberculosis. J Thorac Cardiovasc Surg 1955;29: Miller JI Jr. Physiologic evaluation of pulmonary function in the candidate for lung resection. J Thorac Cardiovasc Surg 1993;105: Jackson MCV. Preoperative pulmonary evaluation. Arch Intern Med 1988;148: Olsen GN, Block AJ, Swenson EW, Castle JE, Wynne JW. Pulmonary function evaluation of the lung resection candidate: a prospective study. Am Rev Respir Dis 1975;111: Mjorner G. 133Xe radiospirometry: a clinical method for studying regional lung function. Scand J Respir Dis 1968;64: Mittman C. Assessment of operative risk in thoracic surgery. Am Rev Respir Dis 1961;84: Eugene J, Brown SE, Light RW, Milne NE, Stemmer EA. Maximum oxygen consumption: a physiology guide to pulmonary resection. Surg Forum 1982;33: Smith 5FP, Kinasewitz GT, Tucker ~VY, Spillers WP, George RB. Exercise capacity as a predictor of post-thoracotomy morbidity. Am Rev Respir Dis 1984;129: Rao V, Todd TRJ, Kuus A, Buth KJ, Pearson FG. Exercise oximet D, versus spirometry in the assessment of risk prior to lung resection. Ann Thorac Surg 1995;60: Cooper JD, Trulock EP, Triantafillou AN, et al. Bilateral pneumonectomy (volume reduction) for chronic obstructive pulmonary disease. J Thorac Cardiovasc Surg 1995;109: Schwartz DA, Helmers RA, Galvin JR, et al. Determinants of survival in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1994;149: Peters-Golden M, Wise RA, Hochberg MC, Stevens MB, Wigley FM. Carbon monoxide diffusing capacity as predictor of outcome in systemic sclerosis. Am J Med 1984;77: Ungerer RG, Tashkin DP, Furst D, et al. Prevalence and clinical correlates of pulmonary, arterial hypertension in progressive systemic sclerosis. Am J Med 1983;75: Owens GR, Rogers RM, Pennock BE, Levin D. The diffusing capacity as a predictor of arterial oxygen desaturation during exercise in patients with chronic obstructive pulmonary disease. N Engl J Med 1984;310: Ferguson MK, Reeder LB, Mick R. Optimizing selection of patients for major lung resection. J Thorac Cardiovasc Surg 1995;109: Gaensler EA, Wright GW. Evaluation of respiratory impairment. Arch Environ Health 1996;12: American Thoracic Society Alta Conference. Single breath carbon monoxide diffusing capacity. (transfer factor): recommendations for a standard technique. Am Rev Respir Dis 1987;136: Wernly JA, DeMeester TR, Kirchner PT, Myerowitz PD, Oxford DE, Golomb HM. Clinical value of quantitative ventilation-perfusion lung scans in the surgical management of bronchogenic carcinoma. J Thorac Cardiovasc Surg 1980;80: Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: MacLeod, ed. Evaluation of chemotherapeutic agents. New York: Columbia University Press, 1949: Mahler DA, Weinberg DH, Wells CK, Feinstein AR. The measurement of dyspnea: contents, interobserver agreement, and physiologic correlates of two new clinical indexes. Chest 1987;85: Task Group on Screening for Respiratory Disease in Occupational Settings. Official statement of the American Thoracic Society. Am Rev Respir Dis 1982;126: Abratt RP, Willcox PA. The effect of irradiation on lung function and perfusion in patients with lung cancer. Int J Radiat Oncol Biol Phys 1995;31: Choi NC, Kanarek DJ. Toxicity of thoracic radiotherapy on pulmonary function in lung cancer. Lung Cancer 1994;10: $ Pelletier C, Lapointe L, LeBlanc P. Effects of lung resection on pulmonary function and exercise capacity. Thorax 1990; 45: Steen VD, Graham G, Conte C, et al. Isolated diffusing capacity reduction in systemic sclerosis. Arthritis Rheum 1992;35: Appendix 1. Modified Medical Research Council Dyspnea Scale Grade Symptoms Not troubled with breathlessness except with strenuous exercise Troubled by shortness of breath when hurrying on the level or walking up a slight hill Walks slower than people of the same age on the level because of breathlessness or has to stop for breath when walking at own pace on the level Stops for breath after walking about 100 yards or after a few minutes on the level Too breathless to leave the house, or breathless when dressing or undressing

7 974 BOUSAMRA ET AL Ann Thorac Surg LOW DLCO IN PULMONARY RESECTION 1996;62: Appendix 2. Baseline Dyspnea Scale and Transition Dyspnea Index Grade Symptoms Baseline dyspnea scale: functional impairment 4 No impairment. Able to do usual activities and occupation without shortness of breath. 3 Slight impairment. Distinct impairment in at least one activity, but no activities completely abandoned. Reduction in activity at work or in usual activities that seems slight or not clearly caused by shortness of breath. 2 Moderate impairment. Patient has changed jobs or has abandoned at least one usual activity because of shortness of breath. 1 Severe impairment. Patient is unable to work or has given up most or all usual activities because of shortness of breath. 0 Very severe impairment. Patient is unable to work and has given up most or all usual activities because of shortness of breath. Transition dyspnea index: change in functional impairment -3 Major deterioration. Has deteriorated two grades or more from baseline status. -2 Moderate deterioration. Formerly working but has had to stop working or has completely abandoned some of usual activities because of shortness of breath. -1 Minor deterioration. Has changed to a lighter job or has reduced activities in number or duration because of shortness of breath. Any deterioration is less than in preceding categories. 0 No change. No change in functional status caused by shortness of breath. -1 Minor improvement. Able to return to work at reduced pace or has resumed some customary activities with more vigor than previously because of improvement in shortness of breath. 2 Moderate improvement. Able to return to work at nearly usual pace or able to return to most activities with restriction only. +3 Major improvement. Able to return to work at former pace and able to return to full activities with only mild rest because of improvement in shortness of breath. DISCUSSION DR MARK K. FERGUSON (Chicago, 1L): I congratulate Drs Bousamra and Haasler on a very interesting paper. This represents one of the first prospective studies and is a necessary follow-up to the original reports that DLCO is an independent factor associated with the risk of morbidity and mortality in patients undergoing major lung resection. Their most important finding in this contemporary series of patients was that DLCO correlates with morbidity, particularly pulmonary_ morbidity. These data reinforce the utility of assessing DLCO in estimating the risk of major lung resection. They did not confirm that DLCO is correlated with mortality. Their mortality rate in both the low- and high-dlco groups was quite acceptable, and it is likelv that the equivalent rates were due in large part to improved postoperative care, including pain control and aggressive pulmonary toilet. However, I wonder whether the good outcomes in this study were due perhaps to spuriously low DLCO measurements in some of their patients. You mentioned that your DLCO values were corrected for anemia, and I presume that you mean corrected for hemoglobin, on all of the patients. I wonder whether they were also volume adjusted, which generally increases the DLCO value and might have put many of your patients into a low-risk category. 1 also note that the values of forced expiratory volume in 1 second were relatively high in both of your groups, and I would like you to reflect on whether you think this might have affected the outcome in your patients. I think the most important original finding in your report is that with long-term follow-up, there was relative pulmonary insufficiency in a large number of patients in the Iow-DLCO group. These types of analyses are growing more and more important in the managed care climate in which we find ourselves. Perhaps you could speculate whether these relatively poor long-term outcomes will affect the types of operations and the patients for whom you select major lung resection in the future. DR BOUSAMRA: Thank you, Dr Ferguson. I appreciate your remarks. First of all, i think this study does not contradict your previous study; we looked at these patients prospectively and believed that if we anticipated complications, we might be able to normalize the mortality rate. We may have been able to reduce the mortality rate by prolonging intensive care and by treating more aggressively postoperative complications related to pain. With regard to the measurement of diffusion capacity, this measurement is highly variable among laboratories. We did demonstrate reproducibility between the two laboratories in the study, one at the Veterans Hospital and one at the County Hospital. We did not normalize the diffusion capacity to volume. I believe that the DLCO corrected for lung volume is simply a separate value. You can choose either to measure the diffusion capacity or to normalize it to the total lung volume, and each value has merit. If the total lung volume is increased, as it probably is in most emphysema patients or in people with chronic obstructive pulmonary disease whom we are operating on, it will cause the diffusion capacity to be artificially increased, but that would not contradict our results of having a more normal operative mortality rate in patients with low DLCO. Finally, with regard to the question of chronic dyspnea, one quarter to one third of our patients experienced substantial chronic dyspnea. I think that fact should be considered in physicians' decision-making processes when deciding both whether to offer operation to a patient and whether they should undergo limited or extended pulmonary resection. We know that patients who had simple lobectomy were less likely to have chronic dyspnea than patients who underwent pneumonectomy. It might be that people who undergo a segmentectomy or a wedge resection would be less likely to have chronic dyspnea than patients who undergo a lobectomy. I do not think that the data are so convincing that we should steer people away from resection for fear of having certain degrees of dyspnea, because

8 Ann Thorac Surg BOUSAMRA ET AL ;62: LOW DLCO IN PULMONARY RESECTION the majority of these people were offered a curative resection at the price of a 25% risk of chronic dyspnea. DR THOMAS R. J. TODD (Toronto, Ont, Canada): 1 think as surgeons we have come to recognize what our medical colleagues have known for a long time, and that is that respiratory rehabilitation not only produces a subjective change in dyspnea, but also can improve objective measures in some of the spirometric studies that we undertake. Given that you were looking at chronic changes in the dyspnea scales, did any of these patients at any point engage in respirator}, rehabilitation programs, and did that influence the results? DR BOUSAMRA: I did not specifically question the patients in that regard, but in my chart reviews, I did not notice that they had undergone any specific pulmonary, rehabilitation process. Only a few patients returned to have postoperative spirometric functions remeasured. DR THOMAS R. CALHOUN (Houston, TX): What was the lower limit of DLCO that you accepted, and what is the lower limit that can be accepted for resection? DR BOUSAMRA: That is a recurring question for which I do not have a definite answer. We operated on patients whose diffusion capacity preoperatively was as low as 27% predicted for lobectomy and as low as the 40% range for pneumonectomy. We do not have a specific cutoff point. We do believe that the risk increases as the diffusion capacity decreases. But to put forward a cutoff point for DLCO would be premature because our mortality, rate in this series was only 5% in patients with low DLCO. DR THOMAS W. RICE (Cleveland, OH): Do you have any data concerning the lengths of intensive care unit and hospital stays and cost? Do you think that the equal early results are a result of two standards of care for your patients? DR BOUSAMRA: 1 do think that the normalization of mortality and the near normalization of morbidity are related to increased intensive care; that is, we recognized that patients with low DLCO would be high risk. The average lengths of stay were in the range of 12 to 14 days, 12 for the high-dlco and 14 for the low-dlco groups, but those data are somewhat skewed in that a large proportion of the patients were at the Veterans Hospital, where the impetus to send people home early is not the same as it is in the community. DR RICE: What about intensive care unit stay? That is very costly. DR BOUSAMRA: We did not measure intensive care unit stay. DR LEWIS WETSTEIN (Freehold, NJ): I also congratulate Dr Bousamra on some extremely insightful work. We have also been very interested in minimizing postoperative morbidity and mortality after lung resections, and therefore have evaluated another index: exercise oxygen consumption. We found that as an independent variable, ie, regardless of the static spirometry (forced expiratory, volume in 1 second), exercise oxygen consumption accurately predicts postoperative morbidity and mortality. Here, however, echoing what Dr Calhoun has asked you, I see no underlying limit of DLCO. As I quickly perused your data, the postoperative morbidity reflected your preoperative forced expiratory volume in 1 second regardless of whether the DLCO was high or low. If I am correct, then, there is no lower limit of DLCO; instead, it still depends on the other physiologic indices. My question is whether postoperative morbidity still really depends on the other physiologic indices. I commend you once again, because work such as this-- dissecting out every physiologic factor and attempting to evaluate it independently--continues to dramatically affect postoperative morbiditv and mortality. DR BOUSAMRA: The DLCO tended to correlate with the spirometric values. We performed a multivariate analysis on patients with complete data and found that age and low DLCO were predictors of increased postoperative morbidity with respect to respiratory complications, but, again, we remained without a specific cutoff point for low DLCO. However, there have been multiple other studies done analyzing exercise desaturation, oxygen consumption, and the like, and it is difficult to define a single value below which patients should not be offered a potentially curative resection. The basic problem is that we do not have any single test that one can measure such that operative resection is uniformly prohibited below a certain value. I think this paper plays a role in adding to the general fund of knowledge by giving us an overall assessment of who should be operated on and who should not.