A Pooled Analysis of FEV 1 Decline in COPD Patients Randomized to Inhaled Corticosteroids or Placebo*

CHEST A Pooled Analysis of FEV 1 Decline in COPD Patients Randomized to Inhaled Corticosteroids or Placebo* Joan B. Soriano, MD, PhD; Don D. Sin, MD, FCCP; Xuekui Zhang, MSc; Pat G. Camp, MSc; Julie A. Anderson, MD; Nick R. Anthonisen, MD; A. Sonia Buist, MD; P. Sherwood Burge, MD; Peter M. Calverley, MD; John E. Connett, PhD; Stefan Petersson, PhD; Dirkje S. Postma, MD; Wojciech Szafranski, MD; and Jørgen Vestbo, MD Original Research COPD Background: There is controversy about whether therapy with inhaled corticosteroids (ICSs) modifies the natural history of COPD, characterized by an accelerated decline in FEV 1. Methods: The Inhaled Steroids Effect Evaluation in COPD (ISEEC) study is a pooled study of patient-level data from seven long-term randomized controlled trials of ICS vs placebo lasting > 12 months in patients with moderate-to-severe COPD. We have previously reported a survival benefit for ICS therapy in COPD patients using ISEEC data. We aimed to determine whether the regular use of ICSs vs placebo improves FEV 1 decline in COPD patients, and whether this relationship is modified by gender and smoking. Results: There were 3,911 randomized participants (29.2% female) in this analysis. In the first 6 months after randomization, ICS use was associated with a significant mean ( SE) relative increase in FEV 1 of 2.42 0.19% compared with placebo (p < 0.01), which is quantifiable in absolute terms as 42 ml in men and 29 ml in women over 6 months. From 6 to 36 months, there was no significant difference between placebo and ICS therapy in terms of FEV 1 decline ( 0.01 0.09%; p 0.86). The initial treatment effect was dependent on smoking status and gender. Smokers who continued to smoke had a smaller increase in FEV 1 during the first 6 months than did ex-smokers. Female ex-smokers had a larger increase in FEV 1 with ICS therapy than did male ex-smokers. Conclusions: We conclude that in COPD in the first 6 months of treatment, ICS therapy is more effective in ex-smokers than in current smokers with COPD in improving lung function, and women may have a bigger response to ICSs than men. However, it seems that after 6 months, ICS therapy does not modify the decline in FEV 1 among those who completed these randomized clinical trials. (CHEST 2007; 131:682 689) Key words: COPD; corticosteroids; FEV 1 ; natural history; pooled analysis Abbreviations: ATS American Thoracic Society; BMI body mass index; CCLS Copenhagen City Lung Study; ERS European Respiratory Society; EUROSCOP European Respiratory Society Study on Chronic Obstructive Pulmonary Disease; ICS inhaled corticosteroid; IL interleukin; ISOLDE Inhaled Steroids in Obstructive Lung Disease in Europe; LHS Lung Health Study; RCT randomized controlled trial COPD represents an increasing burden throughout the world and is considered to be a major global epidemic. 1 It affects 5 to 15% of all adults in industrialized countries, its prevalence increases steeply with age, and it caused 2.7 million deaths worldwide in 2000. 2,3 Current international guidelines such as the Global Initiative for Chronic Obstructive Lung Disease and the American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines 4,5 define COPD as a modifiable and treatable disease. However, the natural history of COPD, which is characterized by an accelerated rate of lung function decline in terms of FEV 1, has been convincingly modified only by smoking cessation. 6 Both the Global Initiative for Chronic Obstructive Lung Disease 4 and the ATS/ERS guidelines 5 con- 682 Original Research

sider COPD as an inflammatory disease with pulmonary and systemic components. Inhaled corticosteroids (ICSs) are widely used in the treatment of COPD, at all stages of the disease, although guidelines recommend that they be restricted to individ- For editorial comment see page 648 uals with an FEV 1 of 50% predicted who have frequent exacerbations because of their efficacy in improving symptoms and quality of life and in reducing exacerbations of COPD. 4,5 However, there *From the Program of Epidemiology and Clinical Research (Dr. Soriano), Fundació Caubet-CIMERA Illes Balears, International Centre for Advanced Respiratory Medicine, Bunyola, Mallorca, Illes Balears, Spain; James Hogg icapture Center for Cardiovascular and Pulmonary Research (Dr. Sin, Mr. Zhang, and Ms. Camp), St. Paul s Hospital, Vancouver, BC, Canada; the Department of Statistics (Dr. Anderson), GlaxoSmihKline R&D, Greenford, Middlesex, UK; the Department of Medicine (Dr. Anthonisen), University of Manitoba, Winnipeg, MB, Canada; the Department of Medicine (Dr. Buist), Oregon Health and Science University, Portland, OR; the Department of Respiratory Medicine (Dr. Burge), Heartlands Hospital NHS Trust, Birmingham, UK; the Department of Medicine (Dr. Calverley), University Hospital Aintree, Liverpool, UK; the Division of Biostatistics (Dr. Connett), School of Public Health, University of Minnesota, Minneapolis, MN; the Department of Statistics (Dr. Petersson), AstraZeneca R&D, Lund, Sweden; the Department of Pulmonology (Dr. Postma), University of Groningen, the Netherlands; the Department of Lung Diseases (Dr. Szafranski), Voivodeship Specialist Hospital, Radom, Poland; and North West Lung Centre (Dr. Vestbo), South Manchester University Hospital NHS Trust, Wythenshawe Hospital, Manchester, UK. Interdisciplinary Capacity Enhancement: Bridging Excellence in Respiratory Disease and Gender Studies (ICEBERGS), which is supported by funding from Canadian Institutes of Health Research (IGH/ICRH), the Canadian Lung Association, and the Heart and Stroke Foundation of Canada (http://www.icebergs.ubc.ca). Dr. Soriano was an employee of GlaxoSmithKline, a manufacturer of respiratory drugs, up to 2005. Dr. Sin has received honoraria for speaking engagements from AstraZeneca and Glaxo- SmithKline, and has received consultancy fees and research funding from GlaxoSmithKline. Dr. Anderson is currently an employee of GlaxoSmithKline, a manufacturer of respiratory drugs. Dr. Anthonisen and Dr. Buist are members of a respiratory advisory board for GlaxoSmithKline. Dr. Calverley has received honoraria for speaking engagements and research funding from AstraZeneca and GlaxoSmithKline. Dr. Petersson is currently an employee of AstraZeneca, a manufacturer of respiratory drugs. Dr. Postma has received honoraria for speaking engagements and research funding from AstraZeneca and Glaxo- SmithKline. Dr. Vestbo has received honoraria for speaking engagements and research funding from AstraZeneca and Glaxo- SmithKline. Drs. Zhang, Camp, Burge, Connett, and Szafranski have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Manuscript received September 12, 2006; revision accepted October 2, 2006. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Joan B. Soriano, MD, PhD, Head, Program of Epidemiology and Clinical Research, Fundació Caubet-CIMERA Illes Balears, International Centre for Advanced Respiratory Medicine, Recinte Hospital Joan March, Carretera Soller Km 12; 07110 Bunyola, Mallorca, Spain; e-mail: jbsoriano@caubet-cimera.es DOI: 10.1378/chest.06-1696 is ongoing controversy over whether ICS therapy modifies the natural history of COPD. 7,8 Previous metaanalyses 9 11 based on the published results of randomized controlled trials (RCTs) of ICS therapy vs placebo in COPD were methodologically heterogeneous and produced inconclusive results. A relatively new tool, pooled analysis, which collects individual patient-level data and analyzes trials as a single, new trial, has advantages over classical metaanalyses as it avoids many of the methodological pitfalls related to classic metaanalytic techniques that rely exclusively on published data. 12 By pooling seven long-term RCTs, we recently reported 13 that ICS therapy was associated with a 27% lower all-cause mortality rate in COPD patients, and that this survival effect may be modulated by gender, smoking status, and postbronchodilator FEV 1. The current study is aimed at quantifying whether the regular use of ICSs vs placebo improves FEV 1 decline in COPD patients, and whether this relationship is modified by gender and smoking. Additionally, we aimed to determine whether short-term changes in lung function with ICS therapy vs placebo (ie, improvement in FEV 1 within the first 6 months of therapy) can predict survival in COPD patients. Materials and Methods The methods of the Inhaled Steroids Effect Evaluation in COPD study 13 have been reported. Briefly, we pooled data from all seven RCTs in which stable patients with COPD were randomly assigned to receive ICSs or placebo for at least 12 months. These trials included the Lung Health Study (LHS)-2, 14 the Copenhagen City Lung Study (CCLS), 15 Inhaled Steroids in Obstructive Lung Disease in Europe (ISOLDE), 16 ERS Study on COPD (EUROSCOP), 17 the Trial of Inhaled Steroids and Long Acting 2 -Agonists, 18 and the trials by Szafranski et al 19 and Calverley et al 20 (Table 1). All trials and centers were authorized by their corresponding institutional review boards, and all participants signed a written informed consent form, as per the international standards that were valid during the conduct of these trials. Anonymized data collected from the seven RCTs were sent from each trial site to the central Inhaled Steroids Effect Evaluation study coordinating center, where they were merged together for analytic purposes. As previously discussed, 13 each trial used slightly different methods, from recruitment and selection criteria to design, and excluded asthma patients somewhat differently. All trials pooled in this metaanalysis randomized COPD patients who were in the stable phase, and were considered by the investigator to be able to survive for at least the duration of the trial, producing, therefore, valid spirometry data. All metaanalyses included female participants in varying percentages (Szafranski et al 19, 14.5%; CCLS 15, 38.5%), and age inclusion criteria was 40 years. All trials included current smokers and ex-smokers, except for the LHS-2, 14 where current smoking was an inclusion criterion. COPD severity ranged from the mildest, in EUROSCOP, 17 to the most severe, in the 1-year trials of Szafranski et al 19 and Calverley et al. 20 All trials collected spirometry data every 6 months after randomization, around a period free of exacerbation, and performed spirometry and www.chestjournal.org CHEST / 131 / 3/ MARCH, 2007 683

Table 1 Characteristics of Individual Trials at the Time of Randomization* FEV 1 Study Patients, No. Age, yr Men Current Smoker L % Predicted Follow-up, mo Drug (Dose) LHS-2 14 1,057 55.8 (6.8) 665 (62.9) 951 (90.0) 2.25 (0.6) 67.3 (12.8) 24 Triamcinolone (1,200 g/d) CCLS 15 239 59.2 (8.6) 147 (61.5) 184 (77.0) 2.42 (0.8) 76.3 (18.5) 24 Budesonide (867 g/d) ISOLDE 16 520 64.0 (7.0) 394 (75.8) 248 (47.8) 1.46 (0.5) 48.4 (14.4) 30 Fluticasone (1,000 g/d) EUROSCOP 17 1,029 52.6 (7.6) 746 (72.5) 1,028 (99.9) 2.55 (0.7) 73.0 (12.7) 36 Budesonide (800 g/d) TRISTAN 18 515 63.2 (8.4) 373 (72.4) 277 (53.8) 1.41 (0.5) 47.0 (12.8) 12 Fluticasone (1,000 g/d) Szafranski et al 19 248 64.3 (8.5) 212 (85.5) 74 (30.0) 1.07 (0.4) 36.8 (11.5) 12 Budesonide (800 g/d) Calverley et al 20 303 63.1 (9.1) 233 (76.9) 94 (31.0) 1.25 (0.5) 43.2 (16.0) 12 Budesonide (800 g/d) Summary total 3,911 58.3 (0.14) 2,770 (70.8) 2,856 (73.0) 1.97 (0.1) 60.3 (0.1) 18 *TRISTAN Trial of Inhaled Steroids and Long-Acting 2 -Agonists. Values are given as the mean (SD). Variables are given as No. (%). Postbronchodilator values. Values may differ slightly from the original publications because we applied the prediction equations of Hankinson et al 22 to all of the raw FEV 1 values. recorded the findings with methods and machines that met the standards of the ATS or ERS guidelines that were available at the time of the conduct of each trial. 21 As the principal aim of this analysis was to determine the effects of ICS therapy on the rate of decline in FEV 1, we included only those participants who had three or more measurements of FEV 1 over time. All FEV 1 values (postbronchodilator) were converted into percent predicted values according to the reference equations of Hankinson et al, 22 and FEV 1 values for all time points were standardized as percentages of the baseline value. A mixed-effects model that takes into account the nested structure of the data (patients nested within individual trials) and the longitudinal structure of the FEV 1 measurements was used to analyze the FEV 1 data (PROC MIXED function, SAS, version 9.1; SAS Institute; Cary, NC). This model takes into account variances of FEV 1 measurements within individuals as well as between individuals. All estimates were adjusted for baseline values of age, body mass index (BMI), gender, current smoking status, and ICS therapy vs placebo. The effects of gender-ics and smoking-ics interactions on FEV 1 changes were considered in the analysis. There was insufficient statistical power to consider three-way interactions. Data from study dropouts and those persons lost to follow-up were censored (which is defined as a participant not contributing data beyond the last known visit) at the time of the last known visit. There was no extrapolation of any data in this report. As a stratified analysis, we determined the effect of ICS therapy on FEV 1 decline in trials that followed the patients for at least 3 years (ie, EUROSCOP, ISOLDE, CCLS, and LHS-2) to eliminate the potential confounding influence of short-term trials. Finally, to determine whether short-term changes in FEV 1 related to ICS therapy predicted the long-term mortality of COPD patients, we used a Cox proportional hazards model adjusted by sex, current smoking status, and treatment (PROC PHREG function, SAS, version 9.1; SAS Institute). Results Demographic and Clinical Characteristics There were 3,911 COPD participants randomized to receive ICSs or placebo who had three of more measurements of FEV 1. Of these, 29.2% were women (Table 1). Within all COPD participants analyzed in this pooled analysis, men were older, more frequently obese, and more likely to be current smokers than were women (p 0.05) [Table 2]. Regarding COPD severity measured by FEV 1,men had on average a higher FEV 1 in absolute terms, but a lower FEV 1 in terms of percent predicted than women (p 0.05). Table 2 Demographic and Clinical Characteristics of Participants at Baseline by Gender Characteristics Female Participants (n 1,141) Male Participants (n 2,770) Total (n 3,911) Age,* yr 57.1 (8.8) 58.8 (9.0) 58.3 (8.9) BMI,* kg/m 2 24.6 (5.1) 25.2 (4.9) 25.0 (5.0) ICS group 585 (51.3) 1,425 (51.4) 2,010 (51.4) Current smoker 224 (19.6) 830 (30.0) 1,054 (27.0) FEV 1 at baseline* L 1.66 (0.6) 2.10 (0.8) 1.97 (0.8) % predicted 60.1 (18.1) 57.7 (20.0) 58.4 (19.3) *Values are given as the mean (SD). Values are given as the No. (%). 684 Original Research

Treatment Effect (ICS vs Placebo in FEV 1 Decline Data) In the first 6 months after randomization, ICS use was associated with a significant increase in FEV 1 relative to placebo therapy (mean [ SE] additional increase in FEV 1 over 6 months, 2.42 0.19%; p 0.01 [ICS vs placebo use]) [Fig 1]. This relative increase in the percent predicted FEV 1 can be quantified in absolute terms as 42 ml in men and 29 ml in women over 6 months. The relative increase in FEV 1 related to ICS therapy in the first 6 months was apparent in both male and female participants, as well as in participants who were smokers and ex-smokers (Table 3). However, from 6 to 36 months, there was no significant difference between placebo and ICS therapy in FEV 1 decline ( 0.01 0.09%; p 0.86). Of interest, the treatment effect of ICSs vs placebo in FEV 1 decline data displayed a significant interaction with gender and smoking. The gendertreatment interaction and the smoking-treatment interaction were both significant in the first 6 months (Fig 2). Continued smoking attenuated the increase in FEV 1 related to ICS therapy in both men and women (Table 3). So, during the first 6 months of treatment, female ex-smokers who were receiving ICS therapy had a 6.5% increase in FEV 1 (or 65 ml) vs placebo, but corresponding figures for female current smokers were a 0.9% increase in FEV 1 (or 20 ml) vs placebo. Similarly, during the first 6 months of treatment, male ex-smokers who were receiving ICS therapy had a 3.4% increase in FEV 1 (or 45 ml) vs placebo, but corresponding figures for male current smokers were a 2.4% increase in FEV 1 (or 37 ml) vs placebo (all p 0.05). Additionally, within the first 6 months, female ex-smokers had a larger increase in FEV 1 with ICS therapy than did male ex-smokers who received ICS therapy (women receiving ICSs, 4.7% increase in FEV 1 ; men receiving ICSs, 1.5% increase in FEV 1 ) [Table 4], although differences in the absolute values of FEV 1 were minor. Stratified analysis for gender or smoking status showed no significant differences between ICS and placebo therapy on the rate of decline in FEV 1 after the first 6 months in women and men or in smokers and ex-smokers (Table 4). Actually, both percent predicted and absolute values were virtually identical in each one of the subgroups of gender and smoking in the ICS group from 6 to 36 months to those in the placebo group from 0 to 36 months. Stratified Analysis of Trials That Had at Least 3 Years of Follow-up As with the main analysis, in the 2,845 participants in EUROSCOP, ISOLDE, CCLS, and LHS-2, ICS therapy in the first 6 months was associated with a significant mean improvement in FEV 1 relative to placebo therapy in both current smokers (1.39 0.14%; p 0.001) and ex-smokers (3.97 0.41%; p 0.001). From 6 to 36 months of follow-up, however, ICS therapy did not have any significant impact on FEV 1 decline over time. In current smokers, there was a mean change of 0.02 0.09% (p 0.787) in FEV 1 over this time period in those receiving ICS therapy relative to placebo therapy; while in ex-smokers, the mean difference between ICS therapy and placebo was 0.16 0.23% (p 0.493). Survival Analysis Figure 1. Relative change in FEV 1 from baseline in COPD patients randomized to receive ICSs or placebo. The horizontal axis represents months (total number of participants at each visit). f participants who were randomized to receive ICSs; participants who were randomized to receive placebo; bars SE. The possibility that the initial increase in FEV 1 with ICS therapy over the first 6 months was a predictor of survival was explored (Table 5). Baseline FEV 1 (in liters) was inversely associated with mortality in both men and women. Among current smokers, men had a higher mortality rate than did women. Additionally, in ex-smokers there was a small trend toward improved survival in those participants who had a larger increase in FEV 1 from baseline values compared to those with reduced changes with ICS therapy, but it did not reach statistical significance (p 0.081). www.chestjournal.org CHEST / 131 / 3/ MARCH, 2007 685

Table 3 Effects of ICS on FEV 1 Changes in the First 6 Months From Baseline Values, by Gender and Smoking Status* Female Participants Male Participants Variables % Change ml % Change ml Current smokers ICS therapy 0.507 0.556 3 ( 13 to 18) 0.536 0.414 0 ( 14 to 14) Placebo 0.260 0.764 18 ( 35 to 2) 1.696 0.353 37 ( 50 to 24) Difference 0.847 0.477 20 (12 to 28) 2.360 0.221 37 (30 to 44) Ex-smokers ICS therapy 4.730 2.089 29 ( 9 to 67) 1.532 0.948 7 ( 17 to 30) Placebo 2.965 1.514 34 ( 69 to 1) 2.361 0.928 49 ( 72 to 26) Difference 6.490 1.060 65 (47 to 83) 3.423 0.532 45 (33 to 56) *Values are given as the mean SE or No. (95% confidence interval). p 0.05 A positive number denotes a larger increase in FEV 1 above baseline values in ICS therapy group than in the placebo group. A negative number denotes a larger increase in FEV 1 above baseline values in the placebo group than in the ICS group. Discussion The most important finding of the present study, which was conducted with primary individual-based data and was adjusted by major demographic and Figure 2. Relative change in FEV 1 from baseline in COPD patients who were randomized, by gender and treatment group. Top, A: ex-smokers. Bottom, B: current smokers. Πfemale ICS participants; F male ICS participants; E male placebo participants; female placebo participants; bars SE. clinical characteristics, was that ICS therapy in COPD patients did not affect the rate of decline in FEV 1 among those who completed these randomized clinical trials. ICS therapy only produced significant but small improvements in FEV 1 over the first 6 months of therapy, an effect that was most pronounced in female ex-smokers. Beyond this time frame, however, ICS therapy had no significant effect on the rate of decline in FEV 1. In ex-smokers, there was a modest (but not statistically significant) relationship between the initial improvement in FEV 1 and survival, which, while it is not definitive due to the limited sample size, raises the hypothesis that short-term improvements in FEV 1 related to ICS therapy may confer long-term benefits for COPD patients. A larger study is needed to confirm this initial observation. Previous metaanalyses on this topic, 9 11 because they relied largely on published data, could not differentiate the short-term effects from the longterm effects of ICS therapy. Moreover, they could not determine the potential modifying effects of gender and smoking status on FEV 1 changes related to ICS therapy. Additionally, because we had access to individualized data, we were able to adjust for salient confounding variables, namely, baseline age, BMI, gender, and current smoking status, and take into account intraindividual as well as interindividual variation in FEV 1 measurements over time. The short-term improvements in FEV 1 with ICS therapy are consistent with the antiinflammatory effects of ICSs in COPD patients. 23 So far, ICSs have been suggested to reduce the percentage of neutrophils in BAL fluid and the number of bronchial mast cells, 24 but without an effect on the number of bronchial CD8 lymphocytes and macrophages or the number of sputum neutrophils. 25 27 In individuals receiving ICS therapy, the bronchial 686 Original Research

Table 4 Effects of Inhaled Corticosteroids on FEV 1 Change From Baseline Values Throughout the Entire Follow-up Period by Gender and Smoking Status Variables ICS Group First 6 mo 6 36 mo % Change ml % Change ml Placebo Group (0 36 mo) % Change ml Current smokers Male 0.536 0.414 0 ( 14 to 14) 1.305 0.074 31 ( 37 to 25) 1.239 0.061 32 ( 35 to 29) Female 0.507 0.556 3 ( 13 to 18) 1.417 0.102 21 ( 28 to 15) 1.270 0.086 25 ( 28 to 22) Difference* 0.018 0.294 1 ( 8to9) 0.073 0.142 7 (2 to 12) 0.191 0.138 9 (4 to 14) Ex-smokers Male 1.532 0.948 7 ( 17 to 30) 1.245 0.104 25 ( 37 to 16) 1.434 0.108 24 ( 28 to 20) Female 4.730 2.089 29 ( 9to67) 1.109 0.222 24 ( 39 to 9) 1.532 0.188 17 ( 24 to 11) Difference* 3.020 0.860 23 (5 to 41) 0.178 0.366 2 ( 9to13) 0.120 0.213 5 ( 6to15) *Values are given as mean SE or No. (%). A positive number denotes a larger increase in FEV 1 above baseline values in female participants than in male participants. A negative number denotes a larger increase in FEV 1 above baseline values in male participants than in female participants. p 0.05. epithelium, which is a major source of proinflammatory cytokines in the lungs, produces less interleukin (IL)-6 and IL-8 than the bronchial epithelium of patients not receiving ICS therapy. 28 Interestingly, airway IL-6 levels have been associated with an increased frequency of exacerbations 29 and with an accelerated decline in lung function in COPD patients. 30 Why ex-smokers would experience larger benefits from ICS therapy than current smokers is uncertain. A similar finding has been reported also in asthma patients, both in short-term and long-term studies 31 33 assessing the effects of ICSs on lung function decline. Some investigators 34 have postulated that cigarette smoking induces a predominant neutrophilic response in the airways that is generally poorly responsive to corticosteroid therapy. Other investigators 35 have suggested that smoking down-regulates histone deacetylase activity in alveolar macrophages. Since this is a critical molecule in mediating the antiinflammatory effects of corticosteroids, the airway inflammation in active smokers may be expected to be less responsive to therapy with corticosteroids than the inflammation in ex-smokers. Additionally, cigarette smoking may induce drug-metabolizing enzymes, which may accelerate the breakdown of ICSs and their clearance from the lungs. 36 Whatever the mechanism, the findings from the present study suggest that ICSs work best in COPD patients who abstain from smoking. Our findings also suggest that female COPD patients, especially when they quit smoking, have larger benefits from ICS therapy than do male COPD patients. The reason underlying this observation is not entirely clear and requires further exploration in the future. Although there is ongoing controversy, adult women appear to have increased susceptibility to COPD when they smoke compared to men. 37,38 When they stop smoking, women gain more lung function than do men who stop smoking, 39 and in women methacholine reactivity is more strongly related to FEV 1 decline than in men. 38 ICSs Table 5 The Relationship of FEV 1 Changes Related to ICS Therapy in the First 6 Months and All-Cause Mortality During the Subsequent Follow-up Period* FEV 1 Smoking Status Baseline Change in First 6 mo With ICS Therapy Male vs Female Participants Age Ex-smoker Hazard rate 0.877 0.774 1.049 1.005 p Value 0.0373 0.0808 0.4620 0.2042 Current smoker Hazard rate 0.901 1.016 1.220 1.006 p Value 0.0019 0.8483 0.0003 0.0203 *All values adjusted for all variables in the table. Participants who died in the first 6 mo of follow-up were excluded from this analysis. For every 100-mL increase. For every 1-year increment. www.chestjournal.org CHEST / 131 / 3/ MARCH, 2007 687

attenuate airway reactivity to methacholine, 14 raising the possibility that ICSs may be able to amplify the beneficial effects of smoking cessation on airway responsiveness and inflammation. Future studies will be needed to validate this hypothesis. Consistent with previous research, 9 11 the magnitude of the changes in respiratory function with ICS therapy reported here is considered to be small, and some statistically significant differences in group averages might be of little clinical relevance at the individual patient level. Redelmeier et al 40 quantified that a change in FEV 1 of approximately 4%, which is equivalent to an absolute increase or decrease of 112 ml, was a minimum for a patient with severe COPD to detect a decrease or increase in perceived dyspnea. None of the comparisons here achieved this threshold. Therefore, any beneficial effect of ICS therapy in COPD patients might be due to mechanisms that are unrelated to respiratory function. There are certain limitations to the present study. We pooled data from all existing RCTs, and selection biases at recruitment within each of the seven RCTs cannot be ruled out (ie, more women came from the LHS-2, the RCT with the longest duration). 14 All statistical techniques used here are simplistic, in the sense that they create means of individual and group values and they assume a rather linear FEV 1 decline. It might appear attractive to differentiate slow vs rapid decliners in the response to ICS therapy, but, based on individual mean values, some (most) apparently fast decliners were actually driven by fewer data points with different RCT durations, making analysis and interpretation difficult. Further, lung function decline might prove to be nonlinear, and the effects of recurrent exacerbations, which may produce FEV 1 drops of 7 to 10 ml per exacerbation, need to be considered. 41 As mentioned earlier, we censored all participants who were lost to follow-up or had died at their last visit and excluded participants who did not have more than two serial FEV 1 measurements. This selection of participants with three or more time points will likely imply bias, but the alternative option was considered to be inappropriate. Data from ISOLDE 42 indicate that censored participants are generally sicker and experience a faster decline in lung function than those who complete clinical trials, and that they are more likely to be in the placebo group than in the activetreatment group. As such, it is likely that we underestimated the true effects of ICS therapy on FEV 1, especially after 6 months when the rate of number of dropouts from the study significantly increased. Since compliance rates decrease significantly over time, 43 this may have further diluted the effects of ICS therapy on FEV 1 decline and may have substantially reduced the power of the study to capture treatment benefits in terms of the decline in FEV 1. 44 Thus, the FEV 1 data beyond 1 year should be interpreted cautiously. Finally, we did not have measures of smoking history beyond baseline status as current smoker or ex-smoker; therefore, we could not determine whether differences in switchers, the number of pack-years of smoking or treatment groups could have influenced our findings. In summary, the present study indicates that ICS therapy slightly improves lung function within the first 6 months of therapy but has no material effect on the rate of decline in lung function thereafter among those patients who completed these randomized clinical trials. The initial improvement in lung function related to ICS therapy is most pronounced in female participants and in those participants who have quit smoking. ACKNOWLEDGMENT: We dedicate this article to the fond memory of our loving friend and colleague, Professor Romain Pauwels. References 1 Lopez AD. The evolution of the Global Burden of Disease framework for disease, injury and risk factor quantification: developing the evidence base for national, regional and global public health action. Global Health 2005; 1:1 8 2 Chapman KR, Mannino DM, Soriano JB, et al. Epidemiology and costs of chronic obstructive pulmonary disease. Eur Respir J 2006; 27:188 207 3 Pauwels RA, Rabe KF. Burden and clinical features of chronic obstructive pulmonary disease (COPD). Lancet 2004; 364:613 620 4 Pauwels RA, Buist AS, Calverley PM, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) Workshop summary. Am J Respir Crit Care Med 2001; 163:1256 1276 5 Celli BR, MacNee W, American Thoracic Society/European Respiratory Society Task Force. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 2004; 23:932 946 6 Anthonisen NR, Connett JE, Kiley JP, et al. Effects of smoking intervention and the use of an inhaled anticholinergic bronchodilator on the rate of decline of FEV 1 : the Lung Health Study. JAMA 1994; 272:1497 1505 7 Calverley PM. Inhaled corticosteroids are beneficial in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2000; 161:341 342 8 Barnes PJ. Inhaled corticosteroids are not beneficial in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2000; 161:342 344 9 Highland KB, Strange C, Heffner JE. Long-term effects of inhaled corticosteroids on FEV 1 in patients with chronic obstructive pulmonary disease: a meta-analysis. Ann Intern Med 2003; 138:969 973 10 Sutherland ER, Allmers H, Ayas NT, et al. Inhaled corticosteroids reduce the progression of airflow limitation in chronic obstructive pulmonary disease: a meta-analysis. Thorax 2003; 58:937 941 11 Sin DD, McAlister FA, Man SF, et al. Contemporary man- 688 Original Research

agement of chronic obstructive pulmonary disease: scientific review. JAMA 2003; 290:2301 2312 12 Blettner M, Sauerbrei W, Schlehofer B, et al. Traditional reviews, meta-analyses and pooled analyses in epidemiology. Int J Epidemiol 1999; 28:1 9 13 Sin DD, Wu L, Anderson JA, et al. Inhaled corticosteroids and mortality in chronic obstructive pulmonary disease. Thorax 2005; 60:992 997 14 Lung Health Study Research Group. Effect of inhaled triamcinolone on the decline in pulmonary function in chronic obstructive pulmonary disease. N Engl J Med 2000; 343: 1902 1909 15 Vestbo J, Sørensen T, Lange P, et al. Long-term effect of inhaled budesonide in mild and moderate chronic obstructive pulmonary disease: a randomised controlled trial. Lancet 1999; 353:1819 1823 16 Burge PS, Calverley PM, Jones PW, et al. Randomised, double blind, placebo controlled study of fluticasone propionate in patients with moderate to severe chronic obstructive pulmonary disease: the ISOLDE trial. BMJ 2000; 320:1297 1303 17 Pauwels RA, Lofdahl CG, Laitinen LA, et al. Long-term treatment with inhaled budesonide in persons with mild chronic obstructive pulmonary disease who continue smoking: European Respiratory Society Study on Chronic Obstructive Pulmonary Disease. N Engl J Med 1999; 340:1948 1953 18 Calverley PM, Pauwels R, Vestbo J, et al. Combined salmeterol and fluticasone in the treatment of chronic obstructive pulmonary disease: randomised controlled trial. Lancet 2003; 361:449 456 19 Szafranski W, Cukier A, Ramirez A, et al. Efficacy and safety of budesonide/formoterol in the management of chronic obstructive pulmonary disease. Eur Respir J 2003; 21:74 81 20 Calverley PM, Boonsawat W, Cseke Z, et al. Maintenance therapy with budesonide and formoterol in chronic obstructive pulmonary disease. Eur Respir J 2003; 22:912 918 21 Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J 2005; 26:319 338 22 Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general US population. Am J Respir Crit Care Med 1999; 159:179 187 23 Gan WQ, Man SF, Sin DD. Effects of inhaled corticosteroids on sputum cell counts in stable chronic obstructive pulmonary disease: a systematic review and a meta-analysis. BMC Pulm Med 2005; 5:3 24 Ozol D, Aysan T, Solak ZA, et al. The effect of inhaled corticosteroids on bronchoalveolar lavage cells and IL-8 levels in stable COPD patients. Respir Med 2005; 99:1494 1500 25 Keatings VM, Jatakanon A, Worsdell YM, et al. Effects of inhaled and oral glucocorticoids on inflammatory indices in asthma and COPD. Am J Respir Crit Care Med 1997; 155:542 548 26 Hattotuwa KL, Gizycki MJ, Ansari TW, et al. The effects of inhaled fluticasone on airway inflammation in chronic obstructive pulmonary disease: a double-blind, placebo-controlled biopsy study. Am J Respir Crit Care Med 2002; 165:1592 1596 27 Gizycki MJ, Hattotuwa KL, Barnes N, et al. Effects of fluticasone propionate on inflammatory cells in COPD: an ultrastructural examination of endobronchial biopsy tissue. Thorax 2002; 57:799 803 28 Patel IS, Roberts NJ, Lloyd-Owen SJ, et al. Airway epithelial inflammatory responses and clinical parameters in COPD. Eur Respir J 2003; 22:94 99 29 Bhowmik A, Seemungal TA, Sapsford RJ, et al. Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations. Thorax 2000; 55:114 120 30 Donaldson GC, Seemungal TA, Patel IS, et al. Airway and systemic inflammation and decline in lung function in patients with COPD. Chest 2005; 128:1995 2004 31 Chalmers GW, Macleod KJ, Little SA, et al. Influence of cigarette smoking on inhaled corticosteroid treatment in mild asthma. Thorax 2002; 57:226 230 32 Lange P, Scharling H, Ulrik CS, et al. Inhaled corticosteroids and decline of lung function in community residents with asthma. Thorax 2006; 61:100 104 33 Dijkstra A, Vonk JM, Jongepier H, et al. Lung function decline in asthma: association with inhaled corticosteroids, smoking and sex. Thorax 2006; 61:105 110 34 Baraldo S, Turato G, Badin C, et al. Neutrophilic infiltration within the airway smooth muscle in patients with COPD. Thorax 2004; 59:308 312 35 Ito K, Ito M, Elliott WM, et al. Decreased histone deacetylase activity in chronic obstructive pulmonary disease. N Engl J Med 2005; 12; 352:1967 1976 36 Zevin S, Benowitz N. Drug interactions with tobacco smoking: an update. Clin Pharmacokinet 1999; 36:425 438 37 Becklake MR, Kauffmann F. Gender differences in airway behaviour over the human life span. Thorax 1999; 54:1119 1138 38 Gan WQ, Man SFP, Postma DS, et al. Female smokers beyond the perimenopausal period are at increased risk of chronic obstructive pulmonary disease: a systematic review and meta-analysis. Respir Res 2006; 7:52 39 Connett JE, Murray RP, Buist AS, et al. Changes in smoking status affect women more than men: results of the Lung Health Study. Am J Epidemiol 2003; 157:973 979 40 Redelmeier DA, Goldstein RS, Min ST, et al. Spirometry and dyspnea in patients with COPD: when small differences mean little. Chest 1996; 109:1163 1168 41 Donaldson GC, Wedzicha JA. COPD exacerbations: epidemiology. Thorax 2006; 61:164 168 42 Waterhouse JC, Fishwick D, Burge PS, et al. What caused death in the ISOLDE study [abstract]? Eur Respir J 1999; 14(suppl):387S 43 Rand CS, Nides M, Cowles MK, et al. Long-term metereddose inhaler adherence in a clinical trial: the Lung Health Study Research Group. Am J Respir Crit Care Med 1995; 152:580 588 44 Calverley PM, Spencer S, Willits L, et al. Withdrawal from treatment as an outcome in the ISOLDE study of COPD. Chest 2003; 124:1350 1356 www.chestjournal.org CHEST / 131 / 3/ MARCH, 2007 689