Lymphangioleiomyomatosis (LAM), a disease that occurs primarily in women, is characterized by cystic lung lesions causing respiratory failure, which may require lung transplantation. Lung diffusion (DL^sub CO^) and/or FEV^sub 1^ are decreased, but frequently not in parallel with each other. Because cardiopulmonary exercise testing (CPET) provides information that is not obtainable from resting cardiopulmonary tests, we performed CPET in 217 LAM patients and correlated exercise data with clinical markers of severity, computed tomography scans, lung function, and histology. V(dot above)O^sub 2^max was decreased in 162 patients, of whom 28 did not reach anaerobic threshold; 29 had low oxygen uptake at anaerobic threshold, and 54 developed hypoxemia. Hypoxemia occurred even in patients with near normal DL^sub CO^ and FEV^sub 1^. V(dot above)O^sub 2^max decreased with an increasing score of histologic LAM severity and was correlated with computed tomography scans, the use of oxygen, and resting Pa^sub O^sub 2^^. DL^sub CO^ and FEV^sub 1^, however, were the only significant predictors of V(dot above)O^sub 2^max. We conclude that CPET uncovers the presence of exercise-induced hypoxemia and assists in grading the severity of disease and determining supplemental oxygen requirements in patients with LAM.
Keywords: interstitial lung disease; maximal oxygen uptake; diffusion capacity; computed tomography scan; lung histology
Lymphangioleiomyomatosis (LAM), a disease affecting primarily women, is characterized by progressive cystic lung lesions, recurrent pneumothoraces, chylous effusions, lymphatic tumors, and angiomyolipomas (1-3). The clinical course of LAM is highly variable. In some patients, the disease remains quiescent, and pulmonary function tests show only a slow decline in function. In others, a loss of function is rapid, and the time from the first symptoms to onset of respiratory failure and lung transplantation may be only a few years. Pulmonary function abnormalities in LAM consist primarily of impaired diffusion capacity (DL^sub CO^) and FEV^sub 1^ (1-4). However, abnormalities of DL^sub CO^ and FEV^sub 1^, that is, severity of impairment and rates of decline, do not parallel each other (4), which raises the question of how the two tests should be employed to grade the severity and progression of disease. Although the functional limitation in some patients appears to be related to ventilatory problems caused by airflow obstruction, other patients have almost exclusively an impairment in DL^sub CO^ with well-preserved flow rates. In patients with LAM, objective evidence of exercise limitation, along with exercise-induced hypoxemia, has led us to question whether standard pulmonary function tests are an adequate measure of disease severity.
Because cardiopulmonary exercise testing (CPET) evaluates all components of exercise responses and provides information that is not available from tests of pulmonary and cardiac function at rest (5), it may be a preferable method for grading the severity of disease in LAM. Based on this hypothesis, the principal aim of our study was to determine the causes of exercise intolerance in LAM and use the CPET data to grade the severity of disease. To accomplish these objectives, we compared cardiopulmonary exercise data with high-resolution lung computed tomography (CT) scans, pulmonary function tests, lung histology, and clinical markers of disease severity, in a large population of patients with LAM. Some of the results of these studies have been previously reported in the form of abstracts (6, 7).
The study population comprised 294 patients referred to National Institutes of Health for participation in an LAM protocol (NHLBI Protocol 95-H-0186) approved by the Institutional Review Board of the National Heart, Lung, and Blood Institute. All subjects gave informed consent before enrollment. The diagnosis of LAM was made by tissue biopsy in 170 patients and in the remainder by clinical and roentgenographic data (1-4). After exclusions because of death (n = 7), transplantation (n = 20), recent surgery (n = 3), joint disease (n = 4), cardiac disease (n = 3), and missed appointments (n = 40), data for analysis were available from 217 patients. One patient who participated in the study died subsequently from complications of abdominal surgery.
Pulmonary Function Tests
Lung volumes, flow rates, and DL^sub CO^ were measured using a computerized system (Master Screen PFT; Erich Jaeger, Wurzburg, Germany), according to American Thoracic Society recommendations (8, 9). Percentages of predicted normal values were derived from standard equations (10-12).
Cardiopulmonary Exercise Tests
Patients were exercised on a bicycle ergometer or treadmill and a computerized metabolic cart (Vmax 229 Cardiopulmonary Exercise System; Sensormedics, Yorba Linda, CA) using standard incremental protocols (5, 13). Sa^sub O^sub 2^^ was measured using a pulse oximeter (Nellcor Puritan Bennett, model 295). Tests were stopped when the patient reached an oxygen uptake plateau, when Sa^sub O^sub 2^^ fell below 88%, or when the patient became exhausted. The following variables were measured: work rate (watts), V(dot above)O^sub 2^max, heart rate, oxygen pulse (V(dot above)O^sub 2^/heart rate), blood pressure, V(dot above)E, respiratory rate, tidal volume, respiratory gas exchange ratio, and ventilatory equivalent for CO2 at anaerobic threshold (AT). Breathing reserve was calculated as MVV-V(dot above)E/MVV × 100 where MVV is the maximal voluntary ventilation (5, 13). MVV was estimated as FEV^sub 1^ × 40. AT was determined by the dual-methods approach (5, 13). V(dot above)O^sub 2^max was defined as the highest oxygen uptake observed during any 30-second measurement period. Oxygen-dependent patients (n = 31) were exercised while breathing from a 30-L bag filled with a gas mixture containing oxygen at concentrations set by a blender fed by compressed air and oxygen tanks. Tests were supervised by a physician. Predicted values for V(dot above)O^sub 2^max were calculated from standard equations (14) with a correction for body weight (13). For patients exercised on treadmill, predicted V(dot above)O^sub 2^max was obtained by multiplying ergocycle values by 1.11 (13). V(dot above)O^sub 2^max values below 85% of predicted and a decline in Sa^sub O^sub 2^^ of 4% or more were considered abnormal (5).
High-resolution CT scans of the lungs
The grade of severity of disease was determined from high-resolution CT scans, as previously reported (15). The extent of involvement of the lungs was graded as follows: grade 0, no involvement; grade 1, less than 30% involved; grade 2, from 30 to 60% involved; and grade 3, more than 60% involved.
An LAM histology score (LHS), based on the extent of replacement of lung tissue by cystic lesions and infiltration by LAM cells, was determined using open lung biopsy specimens and was scored as follows: LHS-1 = less than 25%, LHS-2 = 25 to 50%, and LHS-3 = more than 50% (16).
To determine the best predictors of V(dot above)O^sub 2^max, we first estimated the correlation coefficient between V(dot above)O^sub 2^max and explanatory variables, including duration of disease, oxygen therapy requirements (no oxygen, oxygen during physical activities, and continuous oxygen therapy), grade of CT scan abnormality, resting Pa^sub O^sub 2^^, and pulmonary function tests, to quantify any linear relationship between V(dot above)O^sub 2^max and any of these variables. Then, using a stepwise procedure, we conducted a multivariate regression analysis with V(dot above)O^sub 2^max as the dependent variable and all the independent variables found to be statistically significant at a 0.1 level in the univariate analysis.
To determine whether exercise capacity correlated with a measure of disease severity that was independent of pulmonary function tests, we defined disease severity scores based on CT scan grade: a severity score was 1, 2, or 3 if the CT scan grade was 1, 2, or 3. Because 31 of the CT scan grade 3 patients were receiving continuous supplemental oxygen, we defined an additional severity score of 4 for these patients. Then we used a one-way analysis of variance with a Bonferroni adjustment to compare the four severity groups. Analysis of variance was also used to compare other patient groups.
Unpaired Student's t test was employed to compare exercise and lung function data in patients exercised on room air with those exercised on supplemental oxygen and in patients with LHS of 1 with those with LHS of 2. All reported p values are two sided. Data are shown as mean ± SEM.
Population Characteristics and Exercise-limiting Symptoms
The mean age of the 217 patients at the time of testing was 45.0 ± 0.6 years (range of 19 to 77) and the time from diagnosis was 5.7 ± 0.3 years (range of 8 months to 22 years). Two patients were smokers (25.2 ± 2.7 pack-years), and 24 were ex-smokers (13.8 ± 1.9 pack-years).
Dyspnea was the major exercise-limiting symptom (40%), followed by leg fatigue (28%), severe hypoxemia (11%), and a combination of dyspnea and leg fatigue (7%). Three patients stopped because of dizziness and three more because of abdominal pain, whereas 14 reached a V(dot above)O^sub 2^max plateau. The remaining patients stopped because of general fatigue.
Exercise and Pulmonary Function Data
Table 1 shows CPET and pulmonary function data for all patients. V(dot above)O^sub 2^max (1,187 ± 29 ml/min), DL^sub CO^ (15.3 ± 0.4 ml/min/ mm Hg), FEV^sub 1^ (2.01 ± 0.05 L), and FEV^sub 1^/FVC ratio (63.6 ± 1.1%) were decreased.
The patients on continuous oxygen therapy who were exercised while breathing supplemental oxygen were significantly older and had significantly lower V(dot above)O^sub 2^max, oxygen pulse, breathing reserve, DL^sub CO^, FEV^sub 1^, and resting Pa^sub O^sub 2^^ than patients exercised on room air (Table 2). In addition, they had a significantly greater ventilatory equivalent for CO2 ratio at AT than patients not on continuous oxygen (Table 2).
Multiple abnormalities of ventilatory, gas exchange, and cardiovascular responses to exercise were observed in our patients. One hundred sixty-two of the 217 patients (75%) had low V(dot above)O^sub 2^max. Among these patients, 28 (17%) failed to reach AT and 29 (18%) reached AT at an oxygen uptake of less than 40% V(dot above)O^sub 2^max predicted. Of the 162 patients, 98 (60%) showed evidence of inefficient gas exchange, of whom 54 (33%) developed hypoxemia; 114 of the 162 patients (70%) had abnormal cardiovascular responses, of which 39 (24%) appeared to be limited by low heart rate reserve and 24 (15%) by low breathing reserve. Twenty-eight patients were limited by symptoms that could not be directly attributed to cardiorespiratory limitation.
Ten of 24 patients (41%) with DL^sub CO^ between 60 and 70% predicted and 7 of 29 patients (21%) with DL^sub CO^ between 70 and 80% predicted, exercised breathing room air, experienced exercise-induced hypoxemia. Of 69 patients with DL^sub CO^ and FEV^sub 1^ of 80% or more predicted exercised on room air, 7 (10%) had exercise-induced hypoxemia. DL^sub CO^ was the single best predictor of exercise-induced hypoxemia (r = 0.550, p
Relationship between V(dot above)O^sub 2^max and Clinical Markers of Disease Severity
One hundred twenty five of the 217 patients did not use supplemental oxygen. Sixty one used oxygen during physical activities, and 31 used oxygen continuously. There were significant differences among these groups of patients. As seen in Figure 2, V(dot above)O^sub 2^max in patients who did not use oxygen (1,360 ± 34 ml/min, 20.4 ml/kg/ min, 80.9 ± 1.9% predicted) was significantly greater than in patients who used oxygen during physical activities (1,041 ± 50 ml/ min, 14.9 ml/kg/min, 63.3 ± 2.6% predicted) and patients on chronic oxygen therapy (768 ± 46 ml/min, 10.9 ml/kg/mm, 49.8 ± 2.7% predicted). DL^sub CO^ and FEV^sub 1^ were also significantly higher in patients who did not use oxygen (18.4 ± 0.4 ml/min/mm Hg, 87.7 ± 1.9% predicted, and 2.32 ± 0.05 L, 85.7 ± 1.8% predicted, respectively) than in patients who used oxygen during physical activities (13.2 ± 0.5 ml/min/mm Hg, 63.8 ± 2.3% predicted and 1.88 ± 0.08 L, 71.9 ± 3.2% predicted, respectively) and patients on chronic oxygen therapy (7.4 ± 0.3 ml/min/mm Hg, 37.1 ± 1.6% predicted and 1.02 ± 0.06 L, 41.7 ± 2.6% predicted, respectively) (Figure 2).
Four of the 217 patients underwent lung transplantation. CPET and pulmonary function tests performed before transplantation snowed a V(dot above)O^sub 2^max of 712 ± 140 ml/min (10.5 ml/kg/ min, 41.7 ± 8.1% predicted), DL^sub CO^ of 7.3 ± 0.9 ml/min/mm Hg (34.0 ± 3.9% predicted) and FEV^sub 1^ of 0.93 ± 0.1 L (33.7 ± 4.5% predicted). These values are significantly lower than those observed in patients not on supplemental oxygen or patients who used oxygen only during physical activities.
Relationship between V(dot above)O^sub 2^max and Severity of Disease
Data for the four severity groups, divided according to CT scan grades and continuous use of supplemental oxygen, are shown in Table 3. It can be seen that as the severity of disease increases, V(dot above)O^sub 2^max decreases. Other exercise parameters such as oxygen pulse and work rate are also progressively reduced, whereas ventilatory equivalent for CO2 increases with disease severity. The proportion of patients who had reduced V(dot above)O^sub 2^max increased from 55% in group 1 to 88% in group 2, 93% in group 3, and 100% in group 4. Finally, the number of patients experiencing exercise-induced hypoxemia increased from 11 (11%) in groups 1 to 16 (40%) in group 2, and 32 (70%) in group 3. Despite use of supplemental oxygen during exercise testing, seven (23%) group 4 patients developed exercise-induced hypoxemia.
Relationship between V(dot above)O^sub 2^max and Histologic Severity of Disease
Of the 102 patients who underwent open lung biopsy, 18 and 12 patients, respectively, with LHSs of 1 and 2, had had lung biopsies within the year before CPET. We found that V(dot above)O^sub 2^max in patients with an LHS of 2 (991 ± 95 ml/min, 58.6 ± 5.2% predicted) was significantly lower (p = 0.005) than that in patients with an LHS score of 1 (1,359 ± 105 ml/min, 80.0 ± 4.6% predicted). In addition, DL^sub CO^ was significantly lower (p = 0.043) in patients with an LHS score of 2 (15.2 ± 1.7 ml/min/mmHg, 71.3 ± 7.0% predicted) than in patients with an LHS score of 1 (19.6 ± 1.2 ml/min/mmHg, 91.9 ± 5.1% predicted). There was no significant difference (p = 0.183) in FEV^sub 1^ between patients with an LHS score of 1 (2.25 ± 0.17 L, 80.8 ± 4.7% predicted) and patients with an LHS score of 2 (2.02 ± 0.22 L, 76.0 ± 8.5% predicted).
Predictors of V(dot above)O^sub 2^max
We found that of all the independent variables that were statistically significant at the 0.1 level in the univariate analysis (Table 4), only DL^sub CO^ and FEV^sub 1^ were statistically significant predictors of V(dot above)O^sub 2^max (p
In this study, which was conducted in a large population of patients with LAM, we found a high prevalence of abnormal exercise responses, suggesting the presence of gas exchange abnormalities, abnormal cardiovascular function, ventilatory abnormalities, and muscle fatigue. These heterogeneous abnormalities caused a decrease in exercise capacity in three fourths of the patients. A multivariate regression analysis showed that the only significant predictors of V(dot above)O^sub 2^max were DL^sub CO^ and FEV^sub 1^. Of note, exercise-induced hypoxemia occurred even in patients with seemingly mild compromise of lung function, with DL^sub CO^ being the best predictor of exercise-induced hypoxemia.
Pulmonary diseases can result in multiple abnormalities of both ventilatory, gas exchange, and cardiovascular responses to exercise (5, 17). Consistent with the clinical and functional heterogeneity of LAM, no single, unique pattern of response was observed in our patients. In some patients, there appeared to be a predominance of gas exchange abnormalities characterized by hypoxemia and/or an excessive ventilatory response to exercise. In others, low V(dot above)O^sub 2^ at AT, or failure to reach anaerobic threshold in the absence of ventilatory limitation, was observed. In a third group of patients, ventilatory abnormalities were a major component of exercise limitation (18). Finally, decreased exercise capacity without evidence of cardiorespiratory abnormalities was observed in some patients. The relative contributions of these factors into limiting exercise capacity varied from patient to patient in a manner that was not completely accounted for by their lung function, that is, DL^sub CO^, FEV^sub 1^. This probably explains the wide variance in V(dot above)O^sub 2^max. Our study, however, does not allow for specific identification of all the pathophysiologic processes involved in causing low exercise capacity in LAM.
The close correlation of V(dot above)O^sub 2^max with DL^sub CO^ and FEV^sub 1^ and between the decline in Sa^sub O^sub 2^^ and DL^sub CO^ may suggest that CPET is of no more value than standard pulmonary function tests in assessing the severity of disease in LAM. V(dot above)O^sub 2^max, however, cannot be fully explained by DL^sub CO^ and FEV^sub 1^, and exercise-induced hypoxemia occurred in the presence of near normal DL^sub CO^ and FEV^sub 1^, suggesting that in LAM, lung function tests do not consistently predict gas exchange abnormalities during exercise. Although correlating well with DL^sub CO^ and FEV^sub 1^, as previously reported (15, 19), CT scan grades of severity also appeared not to be good predictors of gas exchange abnormalities in LAM. Indeed, despite having the same CT scan grade as severity score group 3, severity group 4 patients were on continuous oxygen therapy and had significantly lower V(dot above)O^sub 2^max and resting Pa^sub O^sub 2^^. This finding is of importance because the prevalence and severity of exercise-induced hypoxemia in patients with LAM are probably even greater than those observed in our study. Indeed, the stipulated criterion of a decline in Sa^sub O^sub 2^^ of 4% or more is too stringent, and the most severely affected patients were tested on supplemental oxygen.
There was a close association between V(dot above)O^sub 2^max and LHSs. Patients with more severe scores had significantly lower V(dot above)O^sub 2^max. In addition, and as shown previously (4), DL^sub CO^ more closely followed LHSs than did FEV^sub 1^. Because LHSs are a predictor of death and time to transplantation (16), V(dot above)O^sub 2^max may also be a predictor of survival in patients with LAM. Long-term studies, however, will be required to determine whether V(dot above)O^sub 2^max is of value in predicting survival and time to transplantation.
In conclusion, the occurrence of exercise-induced hypoxemia in patients with mild degrees of impairment in lung function makes CPET an important measure of disease severity in LAM. This finding has both therapeutic, that is, treatment with oxygen, and prognostic implications. Measurement of DL^sub CO^ provides some general guidance regarding need for oxygen therapy, but exercise testing should be performed to determine the severity of gas exchange abnormality and supplemental oxygen requirements for the patients' level of physical activity.
CPET may also be of value in evaluating patients for referral to a lung transplantation center. V(dot above)O^sub 2^max was correlated with use of supplemental oxygen and was lowest in patients who subsequently underwent lung transplantation. Based on our pre-transplant data and the fact that the waiting time on a transplantation list can be several years, patients with V(dot above)O^sub 2^max below 50% predicted and DL^sub CO^ under 40% predicted should probably be considered for referral to a lung transplantation center.
Conflict of Interest Statement: A.M.T-D. has no declared conflict of interest; M.P.S. has no declared conflict of interest; C.J.H. has no declared conflict of interest; A.S.K. has no declared conflict of interest; N.A.A. has no declared conflict of interest; A.R. has no declared conflict of interest; W.DT. has no declared conflict of interest; J.M. has no declared conflict of interest.
Acknowledgment: The authors thank Drs. Martha Vaughan, Vincent C. Manganiello, and Stewart Levine for their helpful discussions and critical review of the article. They thank Xiaoling Chen for her assistance in compilation and analysis of the data (Ms. Xiaoling Chen was supported in part by a grant from the LAM Foundation). The authors also thank the LAM Foundation and the Tuberous Sclerosis Alliance for their assistance in recruiting patients, and they thank Mark Barton, CRTT, PFT, Pete McCraw, CRTT, PFT, and Clara Jolley, PFT, for performing the exercise tests. This study would not have been possible without the cooperation of patients with LAM, who in many cases traveled great distances to participate in our clinical research protocols.
1. Kitaichi M, Nishimura K, Harumi I, Izumi T. Pulmonary lymphangioleiomyomatosis: a report of 46 patients including a clinicopathologic study of prognostic factors. Am J Respir Crit Care Med 1995;151:527-533.
2. Chu SC, Horiba K, Usuki J, Usuki J, Avila NA, Chen CC, Travis WD, Ferrans VJ, Moss J. Comprehensive evaluation of 35 patients with lymphangioleiomyomatosis. Chest 1999;115:1041-1052.
3. Urban TJ, Lazor R, Lacronique J, Murris M, Labrune S, Valeyre D, Cordier JF. Pulmonary lymphangioleiomyomatosis: a study of 69 patients. Medicine (Baltimore) 1999;78:321-337.
4. Taveira-DaSilva AM, Hedin CJ, Stylianou MP, Travis WD, Matsui K, Ferrans VJ, Moss J. Reversible airflow obstruction, proliferation of abnormal smooth muscle cells and impairment of gas exchange as predictors of outcome in lymphangioleiomyomatosis. Am J Respir Crit Care Med 2001;164:1072-1076.
5. American Thoracic Society/American College of Chest Physicians. ATS/ACCP statement on cardiopulmonary exercise testing. Am J Respir Crit Care Med 2003;167:211-277.
6. Taveira-DaSilva AM, Hedin CJ, Stylianou MP, Kristof AS, Moss J. Maximal oxygen uptake (VO2 max) and lung diffusion capacity (DL^sub CO^) as predictors of severity of disease in lymphangioleiomyomatosis (LAM) [abstract]. Am J Respir Crit Care Med 2002;165:A701.
7. Taveira-DaSilva AM, Hedin CJ, Stylianou MP, Kristof AS, Moss J. Decline in maximal oxygen uptake (VO2 max) and lung function in 137 patients with lymphangioleiomyomatosis (LAM) [abstract]. Am J Respir Crit Care Med 2003;167:A952.
8. American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1994;152:1107-1136.
9. American Thoracic Society. Single-breath carbon monoxide diffusing capacity (transfer factor): recommendations for a standard technique-update. Am J Respir Crit Care Med 1995;152:2185-2198.
10. Morris JF, Koski A, Jonhson LC. Spirometric standards for healthy nonsmoking adults. Am Rev Respir Dis 1971;103:57-67.
11. Crapo RO, Morris AH, Clayton PD, Nixon CR. Lung volumes in healthy nonsmoking adults. Bull Eur Physiopathol Respir 1982;18:419-425.
12. Crapo RO, Morris AH. Standardized single breath normal values for carbon monoxide diffusing capacity. Am Rev Respir Dis 1981;123:185-189.
13. Wasserman K, Hansen JE, Sue DY, Casaburi R, Whipp BJ. Principles of exercise testing and interpretation. Baltimore: Lippincott, Williams & Wilkins; 1999.
14. Hansen JE, Sue DY, Wasserman K. Predicted values for clinical exercise testing. Am Rev Respir Dis 1984;129:S49-S55.
15. Avila NA, Chen CC, Chu SC, Wu M, Jones EC, Neumann RD, Moss J, et al. Pulmonary lymphangioleiomyomatosis: correlation of ventilation-perfusion scintigraphy, chest radiography, and CT with pulmonary function tests. Radiology 2000;214:441-446.
16. Matsui K, Takeda K, Yu Z, Travis WD, Barnes JP, Bechtle J, Falk R, Ferrans VJ, Moss J, Travis WD. Prognostic significance of pulmonary lymphangioleiomyomatosis histologic score. Am J Surg Pathol 2001;25:479-484.
17. Agusti AGN, Cotes J, Wagner PD. Responses to exercise in lung diseases. In: Roca J, Whipp BJ, editors. Clinical exercise testing. Sheffield, UK: European Respiratory Monographs; 1997, p. 32-50.
18. Crausman RS, Jennings CA, Mortensen RL, Ackerson LM, Irvin CG, King TE. Lymphangioleiomyomatosis: the pathophysiology of diminished exercise capacity. Am J Respir Crit Care Med 1996;153:1368-1376.
19. Moss J, Avila N, Barnes PM, Litzenberger RA, Bechtle J, Brooks PG, Hedin CJ, Hunsberger S, Kristof AS. Prevalence and clinical characteristics of lymphangioleiomyomatosis (LAM) in patients with tuberous sclerosis complex. Am J Respir Crit Care Med 2001;163:669-671.
Angelo M. Taveira-DaSilva, Mario P. Stylianou, Carolyn J. Hedin, Arnold S. Kristof, Nilo A. Avila, Antoinette Rabel, William D. Travis, and Joel Moss
Pulmonary-Critical Care Medicine Branch and Office of Biostatistics Research, National Heart, Lung, and Blood Institute; Department of Diagnostic Radiology, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, Maryland; and Department of Pulmonary and Mediastinal Pathology, Armed Forces Institute of Pathology, Washington, DC
(Received in original form June 21, 2002; accepted in final form September 2, 2003)
Supported by NHLBI Intramural Research.
Correspondence and requests for reprints should be addressed to Joel Moss, M.D., Ph.D., Pulmonary-Critical Care Medicine Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Building 10, Room 6D05, MSC 1590, Bethesda, MD 20892-1590. E-mail: firstname.lastname@example.org
Am I Respir Crit Care Med Vol 168. pp 1427-1431, 2003
Originally Published in Press as DOI: 10.1164/rccm.200206-593OC on September 4, 2003
Internet address: www.atsjournals.org
Copyright American Thoracic Society Dec 15, 2003
Provided by ProQuest Information and Learning Company. All rights Reserved