November 28, 2004

Predictors of Survival in Severe, Early Onset COPD*

Study objectives: Multiple risk factors for mortality in patients with COPD have been described, but most studies have involved older, primarily male subjects. The purpose of this study was to determine the mortality rate and predictors of survival in subjects with severe, early onset COPD.

Design, setting, and participants: The cohort of 139 probands in the Boston Early-Onset COPD Study was recruited from lung transplant and general pulmonary clinics between September 1994 and July 2002. Subjects were

Measurements and results: Subjects were young (mean age at enrollment, 47.9 years) and had severe airflow obstruction (mean baseline FEV^sub 1^, 19.4% predicted). A total of 72.7% of the subjects were women (p

Conclusion: In this cohort, recent smoking status predicted increased mortality independent of the effects of lifetime smoking intensity. Smoking cessation may confer a survival benefit even among patients with very severe COPD. (CHEST 2004; 126:1443-1451)

Key words: COPD; pulmonary emphysema; smoking; survival analysis

Abbreviations: ATS = American Thoracic Society; BMI = body mass index; CI = confidence interval; HR = hazard ratio; LVRS = lung volume reduction surgery

COPD is the fourth-leading cause of death in the United States, and the mortality rate from COPD has increased over the past 2 decades, while mortality rates from cardiovascular disease and cancer have decreased.1,2 Starting in the 1960s, numerous studies3- 11 have examined the prognosis of subjects with COPD. Multiple risk factors for mortality from COPD have been identified, including age, measures of pulmonary function (including bronchodilator and bronchoconstrictor responsiveness), arterial blood gas values, pulmonary hypertension, functional status, and comorbid illness. These factors were detailed in a review by Gerardi and ZuWallack.12 More recent studies13-15 have examined the effects on mortality of surgical treatments for COPD, including lung transplantation and lung volume reduction surgery (LVRS).

In the Boston Early-Onset COPD Study, an ongoing study to identify the genetic determinants of COPD, we have assembled a cohort of probands with severe, early onset COPD who were recruited primarily from lung transplant and LVRS programs in New England, and their families. Previous work from this study has shown that genetic factors may be involved in the susceptibility to develop COPD in response to cigarette smoking, as evidenced by an increased risk of reduced FEV^sub 1^, chronic bronchitis, and bronchodilator responsiveness among currently smoking and ex-smoking first-degree relatives of the study probands, compared to currently smoking and ex-smoking control subjects.16,17 Further analysis revealed that female first-degree relatives might be particularly susceptible to these effects.18 Linkage analyses for both quantitative and qualitative phenotypes in the early onset COPD study have been published.19-21

To provide insight into the natural history of severe, early onset COPD unrelated to severe α^sub 1^-antitrypsin deficiency, a follow-up questionnaire was administered to early onset COPD patients or their proxy respondents, at a mean of 3.54 years following study enrollment. This report outlines the baseline characteristics of the first 139 probands enrolled in the study, details the events recorded during the follow-up period, and presents a survival analysis of the cohort of subjects with severe, early onset COPD.


Study Participants

The recruitment of participants with severe, early onset COPD was performed in three phases. Details of the 44 probands in phase 1 (September 1994 to October 1996) and the 40 probands in phase 2 (January 1998 to June 1999) have been reported previously.16,18 Fifty-five additional probands were recruited in phase 3 (June 1999 to July 2002). In all three phases, participants with severe, early onset COPD were enrolled primarily from the lung transplant and LVRS programs at Brigham and Women's Hospital and Massachusetts General Hospital, as well as from pulmonary clinics at these hospitals and at the Brockton/West Roxbury Veterans Affairs Hospital. Eligibility criteria included an FEV^sub 1^ of

After giving written informed consent, participants completed a protocol that included a questionnaire, spirometry (before and after bronchodilator use), and a blood sample. The protocol was approved by the Human Research Committees of Partners Health Care (Brigham and Women's Hospital and Massachusetts General Hospital) and of the Brockton/West Roxbury Veterans Affairs Hospital.

Baseline Evaluation

Each participant completed a modified version of the 1978 American Thoracic Society (ATS)-Division of Lung Diseases epidemiology questionnaire.16,18,22 Smoking status was defined by the responses to this questionnaire. Number of pack-years of cigarette smoking as of study enrollment were computed as the product of the duration of smoking (in years) and the average number of cigarettes smoked per day, which was divided by 20 to convert the results to the number of packs. The diagnosis of chronic bronchitis was determined from responses to questions for chronic cough and chronic phlegm production for at least 3 months per year for at least 2 years. Body mass index (BMI) was calculated by dividing the self-reported weight (in kilograms) by the square of the measured height (in meters). Underweight was defined as a BMI of

Spirometry was performed (Survey Tach Spirometer; Warren E. Collins; Braintree, MA) in accordance with ATS specifications.24 Pulmonary function test results are expressed as a percentage of the predicted value, using prediction equations from Crapo and coworkers25 for adult white participants and from Hankinson and coworkers26 for African-American participants. Spirometry was repeated after the inhalation of 180 g (2 puffs) albuterol through a spacer device. Bronchodilator responsiveness was calculated as the absolute change in FEV^sub 1^ divided by the predicted value for FEV^sub 1^ and was expressed as a percentage.

For the subjects who had undergone LVRS prior to study enrollment, preoperative spirometry results were used in the analysis. For four additional subjects, outside spirometric values were used, because of geographic distance, previous major lung surgery, or illness, precluding performance of the test.

Follow-up Study

Surviving subjects and the next-of-kin of nonsurvivors were contacted for a telephone interview between May and November 2002. Respondents were questioned regarding treatment history, including lung transplantation and LVRS, hospitalizations, including ICU admissions and requirement for mechanical ventilation, comorbid diagnoses, cigarette smoking, and cause of death of the deceased subjects. Because data were collected from proxy respondents for a substantial number of participants, cigarette smoking since study enrollment was defined only as the presence or absence of any smoking during that period.

The Social Security Death Index was searched to ascertain the vital status of all subjects as of November 1, 2002, and to determine the dates of death of the decedents. When available, hospital records were accessed to determine the history of lung transplantation and LVRS of the participants who did not complete the telephone interview. The follow-up study was approved by the Human Research Committee of Partners Health Care.

Statistical Analysis

Differences in baseline characteristics and in treatments were compared using the Wilcoxon rank sum test, the Fisher exact test, or the χ^sup 2^ test, where appropriate. Univariate analysis of survival time from the date of study entry was performed using the Kaplan-Meier method, the log rank test, and Cox proportional hazards regression.27-29 Tests were performed using a statistical software package (SAS, version 8.2; SAS Institute; Cary, NC) on a personal computer. Lung transplantation and LVRS that occurred during the follow-up period were treated as time-varying explanatory variables, and survival time was dependent on the current value of the variable, as opposed to its value at study entry.30

To assess the effects of the baseline and \follow-up variables on mortality, a multivariable Cox proportional hazards model was constructed using the subjects with complete data on all pertinent variables. Because of the clinical relevance, adjustment for age and gender was included. Other risk factors with a significance level of p

The model then was repeated excluding the eight subjects who had already undergone LVRS at the time of study enrollment, as well as one other subject who had previously undergone bilateral apical bullectomy.


Baseline Characteristics of Early Onset COPD Subjects

A total of 139 probands were enrolled in the three phases of the study (Table 1). Details of the demographics and spirometry of the first 84 probands (phases 1 and 2) have been previously reported.16,18 The female predominance noted through the first two phases (71.4% women) persisted in the third phase (74.6% women), and both of these values are significantly different than the predicted equal sex distribution (p

There were no differences between the women and men in terms of age at enrollment, baseline FEV^sub 1^ as a percentage of predicted values, bronchodilator responsiveness, or diagnosis of chronic bronchitis. Men tended to have smoked more pack-years at baseline, but the difference was not significant (p = 0.11). Men were more likely to smoke during the follow-up period (men who reported smoking, 41.4%; women who reported smoking, 20.5%; p = 0.029). There was no difference in age, baseline FEV^sub 1^, or number of pack- years smoked between those who did and did not report smoking during the follow-up.

Table 1-Baseline Characteristics of Early Onset COPD Subjects*

The subjects had profound airflow obstruction (mean baseline FEV^sub 1^, 19.4% predicted), and 41.7% met the ATS criteria for the definition of chronic bronchitis.31 A history of cigarette smoking was nearly universal. Only three subjects were lifetime nonsmokers. The mean ( SD) lifetime cigarette consumption was 38.9 21.6 pack- years prior to study enrollment. The cohort was predominantly white, with only four African-American subjects. The mean response to an inhaled bronchodilator was an increase of 3.0% of the predicted FEV^sub 1^ value, and only 17 subjects had an absolute increase in FEV^sub 1^ of ≥ 200 mL.

Follow-up Study

Subjects were followed up for a mean period of 3.54 years (range, 2 months to 8.1 years). Over that period, there were 37 deaths among the 139 subjects (Table 2). The majority of the deaths were due to cardiorespiratory illness. Lung cancer had been diagnosed in five participants since study enrollment. Questionnaires were completed for 108 of the 139 subjects (77.7%). Most follow-up questionnaires were completed by the probands themselves (78.7%), and the remainder were obtained from interviews with proxy respondents.

Of those subjects for whom complete follow-up data were available, almost 90% had used home oxygen, and more than two thirds had completed pulmonary rehabilitation therapy. Twenty-six percent reported any cigarette smoking since study enrollment.

Twenty-one subjects had undergone single-lung or double-lung transplantation since the time of study enrollment. The mean baseline FEV^sub 1^ was significantly lower in the group of patients who had received lung transplants, compared to those who did not (15.3% vs 20.1% predicted, respectively; p = 0.0039). There were no differences in age, gender, or number of pack-years of smoking between the transplanted and nontransplanted groups.

A total of 33 subjects had undergone either unilateral or bilateral LVRS, with 14 having undergone LVRS prior to study enrollment, and 19 having undergone surgery during the follow-up period. In addition, two subjects had volume reduction surgery performed on the native lung subsequent to a single-lung transplant. For the purpose of this study, this was not considered to be LVRS.

Table 2-Follow-up Survey of Early Onset COPD Subjects*

A significantly greater proportion of the female participants underwent LVRS during the study period compared to male participants (19.6% vs 3.1%, respectively; p = 0.025). However, the 19 subjects who underwent LVRS during the study period were not different from those who did not in terms of baseline FEV^sub 1^, age, or number of pack-years of smoking.

Among the 40 subjects who received surgical therapy for COPD during the study period, the mean baseline FEV^sub 1^ was significantly lower in those who underwent lung transplantation compared to LVRS (15.3% vs 19.2% predicted, respectively; p = 0.019). Women tended to be treated with LVRS more commonly than with transplantation, although this did not reach statistical significance (p = 0.095). There were no differences in age or smoking history (in pack-years) comparing those subjects who had undergone transplantation with those who had undergone LVRS.

Survival Analysis of Early Onset COPD Subjects

Starting at the point of study entry, the 3-year survival rate was 85% and the 5-year survival rate was 72%. The median survival time since enrollment for the cohort was 7.0 years, which was estimated by the Kaplan-Meier method. The stratified Kaplan-Meier survival estimates are shown in Figure 1. Survival was not different between women and men (p = 0.16 [log rank test]). Baseline FEV^sub 1^ analyzed by quartile (p = 0.0058) and smoking since study enrollment (p = 0.0070) were both significant predictors of mortality. The intensity of cigarette smoking, defined by quartiles of pack-years, showed a trend toward significance (p = 0.098).

In univariate Cox proportional hazards models, the baseline FEV^sub 1^ percentage of predicted, the number of pack-years of cigarette smoking, and smoking status since study enrollment were significantly associated with survival time (Table 3). A higher FEV^sub 1^ percentage of predicted was linearly associated with an increased survival time. Over the range of values observed, a 1% increase in baseline FEV^sub 1^ percentage of predicted led to an 8% decrease in the risk of dying (hazard ratio [HR], 0.92; 95% confidence interval [CI], 0.86 to 0.97). Greater smoking intensity was associated with decreased survival, and the risk increased by 20% for each 10 pack-years of smoking (HR, 1.19; 95% CI, 1.05 to 1.35). Subjects who smoked during the study period had a risk of mortality that was almost three times that of abstainers (HR, 2.83; 95% CI, 1.29 to 6.21).

Age, gender, symptoms of chronic bronchitis, bronchodilator responsiveness, underweight, use of home oxygen, and completion of pulmonary rehabilitation were not significant predictors of survival. Treatment with LVRS did not predict survival in this cohort. In the univariate analysis, there was a trend toward a significant effect of lung transplantation (p = 0.065). The time interaction term was significant, indicating nonproportionality of the hazards; therefore, the risk of mortality varied with time following the surgery. Early after transplantation, the trend was toward a survival benefit, but this effect was attenuated over time.

In the multivariable Cox proportional hazards model, including the 107 subjects with complete data on all relevant predictors, the two variables measuring cigarette smoking (lifetime pack-years of smoking and smoking since study enrollment) were significantly associated with survival time, when adjusted for age, sex, baseline FEV^sub 1^ percentage of predicted, and transplant status (Table 4). For each 10 pack-years of smoking, the risk of dying increased by 20% (HR, 1.20; 95% CI, 1.02 to 1.40), as in the unadjusted model. The risk of dying in those who smoked following enrollment remained increased, and the HR of 2,50 (95% CI, 1.03 to 6.05) was similar to that seen in the unadjusted analysis. In the adjusted model, baseline FEV^sub 1^ was no longer a significant predictor of mortality in this cohort.

FIGURE 1. Kaplan-Meier survival curves for severe, early onset COPD subjects. Top left, A: stratified by sex (139 subjects). Bottom left, B: stratified by quartile of baseline FEV^sub 1^ percentage of predicted (139 subjects). The mean FEV^sub 1^ percentage of predicted values for each quartile were as follows: I, 29.5% predicted; II, 21.3% predicted; III, 15.6% predicted; IV, 10.9% predicted. Top right, C: stratified by smoking status since study enrollment (107 subjects). Bottom right, D: stratified by the quartile of the number of pack-years of cigarette smoking (n = 139). The mean numbers of pack-years for each quartile were as follows: I, 14.7; II, 30.9; III, 44.1; and IV, 67.1.

As in the unadjusted analysis, lung transplantation showed a trend toward significance in the multivariable model (p = 0.067). The trend continued to be toward early benefit, with a decrease in benefit over time, compared to those who did not undergo transplantation. In the multivariable model, the time interaction term was borderline significant, but it was retained in the model based on its significance in the unadjusted analysis.


Since the 1960s, multiple studies have investigated the risk factors that influence mortality in a variety of populations of COPD patients. In the current report, we examined the survival of a unique cohort of subjects with severe early onset COPD. The most striking finding was that two sepa\rate measures of cigarette smoking had the strongest effect on outcome. The number of pack- years smoked prior to study enrollment as well as smoking during the follow-up period were both independent predictors of mortality.

In the univariate analysis, a higher baseline FEV^sub 1^ predicted a lower mortality rate during the follow-up period, but this effect did not persist in the multivariable model, when adjusted for age, sex, lung transplantation, and the two measures of cigarette smoking. This is in contrast to the results of the majority of the prior studies. Previous studies have found inconsistent effects of smoking on survival in subjects with COPD; however, many studies, especially the earlier reports, did not measure or account for smoking-related variables. The unique features of this cohort (ie, relatively young age, severely reduced FEV^sub 1^, and female predominance) may explain the differences in the predictors of mortality in this cohort compared to those in other published studies.

Table 3-Univariate Cox Regression Models for Survival in Early Onset COPD Subjects*

Beginning with the earliest studios, multiple authors3-5,7 have shown the importance of measures of pulmonary function, usually expressed as FEV^sub 1^, as predictors of mortality in patients with COPD, including those with severe disease. More recent studies9,11,32-34 have confirmed the findings of these earlier studies. As in our cohort, the majority of the subjects in all of these studies had a history of cigarette smoking. Yet, many authors did not control for smoking, which is likely to confound the relationship between FEV^sub 1^ and sunival.

Throughout the literature, the reported effects of cigarette smoking on mortality in COPD patients have been inconsistent. This may he due to the use of different measures of cigarette smoking. Authors quantifying smoking intensity (in pack-years or in the number of cigarettes smoked per day) have found this measure to be a consistent predictor of mortality.6,34,35 However, studies6,7,11,35,36 using dichotomous measures of smoking (ie, either ever-smoking vs never-smoking or current smoking vs ex- smoking [and never-smoking]) have not always found an association with mortality.

Table 4-Multivariable Cox Regression Model for Survival in Early Onset COPD Subjects*

In the present population, the confounding effect of cigarette smoking, especially pack-year history of smoking, likely explains why FEV^sub 1^ was a significant predictor of mortality in the univariate analysis, but not in the multivariable model. In addition, the fact that all the subjects had severe airflow obstruction limits the ability to discern an effect of pulmonary function on survival across the narrow range of values of FEV^sub 1^ that were seen in this cohort. A similar result was found in a Dutch study8 of patients with severe COPD.

The majority of the previous studies4,5,8,11,32,34 has shown an effect of age on mortality in COPD patients, yet the present study did not confirm these results. However, there was an upper limit of age for study enrollment, leading to a younger cohort than in any of the previous studies. The mean age in the probands was 48 years, compared to means ranging between 54 and 69 years in the studies noted above. As with FEV^sub 1^, our cohort had a narrow range of ages, limiting the power to detect any effect of age on mortality.

The female predominance is a unique feature of the Boston Early- Onset COPD Study cohort, and previous work18 in this cohort has suggested that women may have a higher risk of the development of severe COPD. The persistence of a female predominance in the 55 previously unreported severe, early onset COPD probands provides further evidence for increased susceptibility to COPD in women. Despite the significant excess of female subjects, gender was not a significant predictor of outcome in the multivariable Cox regression model. Survival was not different between male and female subjects, so reduced survival among male patients with severe early onset COPD is unlikely to explain the observed female predominance in the Boston Early-Onset COPD Study.

Previous studies4 have reported that women with COPD may have decreased mortality compared to men. In a Japanese study of patients using long-term oxygen therapy, and in studies in Barcelona and Finland of emergency department and hospital patients, the survival rate of women with COPD was significantly better than that of men.10,37,38 However, none of these studies controlled for cigarette smoking. Higher rates of cigarette smoking in men, especially in Western Europe and Asia, may be sufficient to explain the gender- related survival differences seen in these studies.39 In one study34 that did control for smoking intensity, no effect of gender on survival in COPD patients was found.

Because the present cohort was younger and had more severe airflow obstruction than did subjects in the previous studies, the overall mortality rates between studies may not be directly comparable. Differences in standards of care, especially the use of home oxygen therapy, make comparisons to the earliest studies difficult. Two of the more recent studies33,34 reported survival rates that are similar to those seen in the present cohort. However, patients in both of these studies had less severe airflow obstruction than our patients, but were older.

Since we have hypothesized that early onset COPD is likely to have stronger genetic influences than late-onset disease, perhaps a more valid comparison group would be subjects with emphysema due to severe α^sub 1^-antitrypsin deficiency. Among 1,129 enrollees in the National Heart, Lung, and Blood Institute Registry of Patients with Severe Deficiency of α^sub 1^-Antitrypsin, the 5- year survival rate was 81%.40 The mean age among registry patients (46 years) was similar to that of our cohort, but the severity of airflow obstruction was not as great, with a mean FEV^sub 1^ of 47% predicted. Gender was not associated with survival in the α^sub 1^-Antitrypsin Deficiency Registry patients, but age and baseline FEV^sub 1^ were significant predictors. Initial smoking status was not related to survival. Approximately 10% of the registry patients underwent lung transplantation, and this was associated with an increase in the risk of mortality.

The data from the United Network for Organ Sharing and the International Society for Heart and Lung Transplantation Registry have suggested that lung transplantation does not confer a survival benefit in recipients with COPD.13 However, one study41 from a single transplant center revealed a significant survival benefit. In the present analysis, there was a suggestion of improved short-term survival following lung transplantation, with a decrease in benefit over time. However, these findings should be interpreted with caution. Our study was not designed to address this specific question, as the overall number of transplants was low and the results did not reach statistical significance.

In the early onset COPD cohort, LVRS was not found to have an effect on survival. This finding is even more difficult to interpret, for reasons similar to the limitations regarding transplantation, as well as the presence of 14 subjects who underwent LVRS prior to study enrollment. However, the multivariable survival model did not change appreciably when these subjects were excluded. In our population, we did not have the data that would be necessary to define the subgroups of patients with the greatest chance of benefit or the highest risk of harm from LVRS, based on the results of the National Emphysema Treatment Trial.14,15

There are additional limitations to the present analysis. The vital status could be accurately recorded for all 139 probands. However, the follow-up questionnaire was completed for only 108 subjects, so information on smoking during the study period was limited to these respondents. For deceased subjects, proxy respondents were interviewed, which may lead to recall bias. However, one would still expect proxies to know whether the subjects had been actively smoking following study enrollment, even though a detailed smoking history may not be available. Therefore, smoking since study enrollment could be analyzed only as a yes/no response. For participants without follow-up questionnaire data, history of lung transplantation was assessed by a review of medical records from the two adult lung transplant centers in New England. It is possible, but unlikely, that subjects could have undergone lung transplantation elsewhere.

The survival analysis was performed using all-cause mortality as the outcome. Due to incomplete data on the causes of death, the analysis could not be limited to respiratory deaths only. For the 23 subjects for whom cause of death was listed by proxy respondents, 19 (83%) listed a respiratory diagnosis as one of the causes of death. The two subjects who had cardiac causes of death listed also had respiratory causes reported. These events likely reflect the presence of sequelae of end-stage lung disease. Because of the predominance of cardiorespiratory causes of death listed for the deceased subjects, we contend that the use of all-cause mortality as an end point of the study was appropriate.

Current smokers are ineligible for lung transplantation or LVRS, so the effects of these surgical procedures may be seen as confounders of the effects of recent smoking status. However, this bias is unlikely to explain the observed association, since we found that LVRS was not associated with improved survival, and the marginal benefit of lung transplantation was controlled for by including this variable in the final model.

In a cohort of patients with severe early onset COPD, we have found that a greater cumulative history of cigarette smoking and recent smoking are both strong, independent predictors of increased m\ortality. This suggests a benefit of smoking cessation on survival even among patients with very severe airflow obstruction. Physicians should encourage smoking cessation in all patients with COPD, regardless of the severity of their disease.

ACKNOWLEDGMENT: We thank Ms. Kimberly Ladouceur for her tremendous assistance with recruitment and follow-up of the study subjects, and Drs. Harold Chapman, Frank Speizer, and Scott Weiss for their support and participation throughout all phases of the Boston Early-Onset COPD Study and for their helpful advice regarding this manuscript. We appreciate the assistance of many physicians in recruiting study participants. We are especially thankful for the enthusiastic support from the members of the early onset COPD families.

* From the Channing Laboratory (Drs. Hersh, DeMeo, Al-Ansari, Carey, and Silverman) and Division of Pulmonary and Critical Care Medicine (Dr. Reilly), Department of Medicine, Brigham and Women's Hospital, Boston, MA; Pulmonary and Critical Care Unit (Dr. Ginns), Department of Medicine, Massachusetts General Hospital, Boston, MA.

Dr. Silverman has received research grant support from GlaxoSmithKline. This research was supported by National Institutes of Health grants T32-HL07427 (training grant) and HL61575 (EKS).


1 National Institutes of Health. Morbidity & mortality: 2002 chart book on cardiovascular, lung, and blood diseases. Bethesda, MD: National Institutes of Health: National Heart, Lung, and Blood Institutem 2002

2 Ries L, Eisner M, Kosary C, et al, eds. SEER cancer statistics review, 1975-2000. Bethesda, MD: National Cancer Institute, 2003

3 Burrows B, Earle RH. Course and prognosis of chronic obstructive lung disease: a prospective study of 200 patients. N Engl J Med 1969; 280:397-404

4 Kok-Jensen A, Sorensen E, Damsgaard T. Prognosis in severe chronic obstructive pulmonary disease: relation to degree of obstruction, age and sex and electrocardiographic changes suggesting cor pulmonale. Scand J Respir Dis 1974; 55:120-128

5 Traver GA, Cline MG, Burrows B. Predictors of mortality in chronic obstructive pulmonary disease: a 15-year follow-up study. Am Rev Respir Dis 1979; 119:895-902

6 Kanner RE, Renzetti AD Jr, Stanish WM, et al. Predictors of survival in subjects with chronic airflow limitation. Am J Med 1983; 74:249-255

7 Anthonisen NR, Wright EC, Hodgkin JE. Prognosis in chronic obstructive pulmonary disease. Am Rev Respir Dis 1986; 133:14-20

8 Postma DS, Sluiter HJ. Prognosis of chronic obstructive pulmonary disease: the Dutch experience. Am Rev Respir Dis 1989; 140:S100-S105

9 Antonelli Incalzi R, Fuso L, De Rosa M, et al. Co-morbidity contributes to predict mortality of patients with chronic obstructive pulmonary disease. Eur Respir J 1997; 10:2794-2800

10 Vilkman S, Keistinen T, Tuuponen T, et al. Survival and cause of death among elderly chronic obstructive pulmonary disease patients after first admission to hospital. Respiration 1997; 64:281- 284

11 Piccioni P, Caria E, Bignamini R, et al. Predictors of survival in a group of patients with chronic airflow obstruction. j Clin Epidemiol 1998; 51:547-555

12 Gerardi D, ZuWallack R. Non-pulmonary factors affecting survival in patients completing pulmonary rehabilitation. Monaldi Arch Chest Dis 2001; 56:331-335

13 Hosenpud JD, Bennett LE, Keck BM, et al. Effect of diagnosis on survival benefit of lung transplantation for end-stage lung disease. Lancet 1998; 351:24-27

14 National Emphysema Treatment Trial Research Group. Patients at high risk of death after lung-volume-reduction surgery. N Engl J Med 2001; 345:1075-1083

15 Fishman A, Martinez F, Naunheim K, et al. A randomized trial comparing lung-volume-reduction surgery with medical therapy for severe emphysema. N Engl J Med 2003; 348:2059-2073

16 Silverman EK, Chapman HA, Drazen JM, et al. Genetic epidemiology of severem early-onset chronic obstructive pulmonary disease: risk to relatives for airflow obstruction and chronic bronchitis. Am J Respir Crit Care Med 1998; 157:1770-1778

17 Celedon JC, Speizer FE, Drazen JM, et al. Bronchodilator responsiveness and serum total IgE levels in families of probands with severe early-onset COPD. Eur Respir J 1999; 14:1009-1014

18 Silverman EK, Weiss ST, Drazen JM, et al. Gender-related differences in severe, early-onset chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2000; 162:2152-2158

19 Silvennan EK, Palmer LJ, Mosley JD, et al. Genomewide linkage analysis of quantitative spirometric phenotypes in severe early- onset chronic obstructive pulmonary disease. Am J Hum Genet 2002; 70:1229-1239

20 Silverman EK, Mosley JD, Palmer LJ, et al. Genome-wide linkage analysis of severe, early-onset chronic obstructive pulmonary disease: airflow obstruction and chronic bronchitis phenotypes. Hum Mol Genet 2002: 11:623-632

21 Palmer LJ, Celedon JC, Chapman HA, et al. Genome-wide linkage analysis of bronchodilator responsiveness and post-bronchodilator spirometric phenotypes in chronic obstructive pulmonary disease. Hum Mol Genet 2003; 12:1199-1210

22 Ferris BG. Epidemiology Standardization Project (American Thoracic Society). Am Rev Respir Dis 1978; 118:1-120

23 American Medical Association. Executive summary of the clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Arch Intern Med 1998; 158:1855- 1867

24 Expert Panel on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152:1107-1136

25 Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981; 123:659-664

26 Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med 1999; 159:179-187

27 Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53:457-481

28 Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50:163-170

29 Cox D. Regression models and life tables (with discussion). J R Stat Soc [B] 1972; 34:187-220

30 Hosmer DW, Lemeshow S. Applied survival analysis: regression modeling of time to event data. New York, NY: John Wiley & Sons, 1999

31 American Thoracic Society. Chronic bronchitis, asthma and pulmonary emphysema: a statement by the Committee on Diagnostic Standards for Nontuberculous Respiratory Diseases. Am Rev Respir Dis 1962; 85:762-768

32 Dubois P, Jamart J, Machiels J, et al. Prognosis of severely hypoxemic patients receiving long-term oxygen therapy. Chest 1994; 105:469-474

33 Bowen JB, Votto JJ, Thrall RS, et al. Functional status and survival following pulmonary rehabilitation. Chest 2000; 118:697- 703

34 Hansen EF, Phanareth K, Laursen LC, et al. Reversible and irreversible airflow obstruction as predictor of overall mortality in asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999; 159:1267-1271

35 Hospers JJ, Postma DS, Rijcken B, et al. Histamine airway hyper-responsiveness and mortality from chronic obstructive pulmonary disease: a cohort study. Lancet 2000; 356:1313-1317

36 Almagro P, Calbo E, Ochoa de Echaguen A, et al. Mortality after hospitalization for COPD. Chest 2002; 121:1441-1448

37 Sunyer J, Anto JM, McFarlane D, et al. Sex differences in mortality of people who visited emergency rooms for asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 158:851-856

38 Miyamoto K, Aida A, Nishimura M, et al. Gender effect on prognosis of patients receiving long-term home oxygen therapy: the Respiratory Failure Research Group in Japan. Am J Respir Crit Care Med 1995; 152:972-976

39 Mackay J, Eriksen M. The tobacco atlas. Brighton, UK: Myriad Editions Limited, 2002

40 The Alpha-1-Antitrypsin Deficiency Registry Study Group. Survival and FEV^sub 1^ decline in individuals with severe deficiency of α^sub 1^-antitrypsin. Am J Respir Crit Care Med 1998; 158:4959

41 Charman SC, Sharpies LD, McNeil KD, et al. Assessment of survival benefit after lung transplantation by patient diagnosis. J Heart Lung Transplant 2002; 21:226-232

Craig P. Hersh, MD, MPH; Dawn L. DeMeo, MD, MPH; Essam Al- Ansari, MD, MPH; Vincent J. Carey, PhD; John J. Reilly, MD, FCCP; Leo C. Ginns, MD, FCCP; and Edwin K. Silverrnan, MD, PhD

Manuscript received January 21, 2004; revision accepted June 18, 2004.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]).

Correspondence to: Edwin K. Silverman, MD, PhD, Channing Laboratory, Brigham and Women's Hospital, 181 Longwood Ave, Boston, MA 02115; e-mail: [email protected]

Copyright American College of Chest Physicians Nov 2004