Smoking cessation in patients with chronic obstructive pulmonary disease: a double-blind, placebo-controlled, randomised trial

Transcription

1 Smoking cessation in patients with chronic obstructive pulmonary disease: a double-blind, placebo-controlled, randomised trial D P Tashkin, R Kanner, W Bailey, S Buist, P Anderson, M A Nides, D Gonzales, G Dozier, M K Patel, B D Jamerson Summary Background Tobacco smoking is associated with chronic obstructive pulmonary disease (COPD) in more than 80% of cases. Our aim was to investigate the effect of sustainedrelease bupropion (amfebutamone) (SR) in promoting abstinence from smoking in patients with COPD. Methods In a double-blind, randomised, placebo-controlled trial 404 individuals with mild or moderate COPD who smoked 15 or more cigarettes per day, were assigned bupropion SR (150 mg twice daily) or placebo for 12 weeks. All patients received smoking cessation counselling. Study medication was taken for 1 week before patients attempted to stop smoking. The primary efficacy endpoint was the complete and continuous abstinence from smoking from the beginning of week 4 to the end of week 7. Participants were followed up at month 6. Analysis was by intention to treat. Findings All patients were chronic smokers with a smoking history of about 51 pack-years. Continuous smoking abstinence rates from week 4 to 7 were significantly higher in participants receiving bupropion SR than in those receiving placebo (28% [57/204] vs 16% [32/200], p=0 003). Continuous abstinence rates from weeks 4 to 12 (18% [36/204] vs 10% [20/200]) and weeks 4 to 26 (16% [32/204] vs 9% [18/200]) were also higher in participants receiving bupropion SR than in those taking placebo (p<0 05). Furthermore, symptoms of tobacco craving and withdrawal were attenuated in those receiving bupropion SR. Seven individuals discontinued study medication because of adverse events. Interpretation Bupropion SR is a well-tolerated and effective aid to smoking cessation in people with mild to moderate COPD. Lancet 2001; 357: See Commentary page??? Division of Pulmonary and Critical Care Medicine, Department of Medicine, UCLA School of Medicine, Los Angeles CA 90095, USA (D P Tashkin MD, M A Nides PhD); University of Utah, Salt Lake City, UT (R Kanner MD); UAB Pulmonary Medicine, AL (W Bailey MD); Oregon Health Sciences University, OR (S Buist MD, D Gonzales PhD); University of Arkansas for Medical Sciences, AR (P Anderson MD); and Glaxo Wellcome Research and Development, Reseach Triangle Park, NC (G Dozier MPH, M K Patel PharmD, B D Jamerson PharmD) Correspondence to: Dr D P Tashkin ( dtashkin@mednet.ucla.edu) Introduction Chronic obstructive pulmonary disease (COPD) is widespread and preventable. The disease is one of the leading causes of morbidity and mortality worldwide and is increasing in prevalence, especially in less-developed countries. Estimates suggest that in less than 20 years, COPD will be one of the five leading global medical burdens on society. 1 The prevention of COPD, therefore, has huge social and economic implications. The most important intervention in the management of COPD, as recommended by the Global Initiative for Chronic Obstructive Lung Diseases, is smoking cessation. 2 Tobacco smoking leads to COPD in more than 80% of cases, 3 and stopping smoking reduces the accelerated decline in pulmonary function and improves long-term prognosis. 4 Smoking-cessation programmes, as a primary method of disease prevention, are cost effective and have better economic value than many life-preserving medical interventions. 5,6 Well implemented programmes could, therefore, be judged the single most important therapeutic intervention in the treatment of COPD. Bupropion (amfebutamone) sustained release (SR) is the first non-nicotine-based pharmacotherapy for smoking cessation. The treatment has been available in USA since 1997 and has received marketing approval throughout most of the European Union. In smokers taking bupropion SR, cessation rates are about double those seen with placebo. 7,8 However, there are no data on smoking cessation rates for bupropion SR in smokers with COPD, a population that has additional barriers for success in smoking cessation such as older age, higher pack-year history, and higher nicotine-addiction scores than previously studied populations. We did a multicentre, double-blind, placebo-controlled randomised study to assess the efficacy of a 12-week course of bupropion SR in helping patients with stage I and II COPD to stop smoking. Continuous abstinence from smoking from week 4 to the end of week 7 was our primary endpoint. We also made secondary measurements of efficacy throughout the treatment and follow-up phases, and based subgroup analyses on participant characteristics and smoking histories. Methods Patients The study population consisted of current smokers with stage I or stage II COPD. All patients were aged 35 years or older, had smoked 15 cigarettes or more per day for the previous year, and had not stopped smoking for more than 3 months during that year. All participants were motivated to stop smoking. Recruitment procedures for this study were consistent across all study sites. We recruited potential study participants by means of print and radio advertisements and screened them specifically for the purposes of the study. They were assessed with spirometry and, if their lung function fell within defined limits, they were considered for inclusion in the trial, provided they fulfilled other entry criteria. We based diagnosis of COPD on spirometric findings of chronic airflow obstruction (ratio of forced THE LANCET Vol 357 May 19,

2 expired volume in 1 s to forced vital capacity [FEV 1 /FVC] 0 70) and presence of clinically defined COPD (emphysema, chronic bronchitis, or smoking-related smallairways disease according to previous definitions). 9 Smoking-related small-airways disease was characterised by chronic airflow limitation with or without symptoms of mucus hypersecretion or emphysema, and without other specific causes of chronic airflow obstruction. 10 Predicted spirometric values were based on Crapo standards. 11 We defined severity of COPD according to COPD-staging guidelines from the American Thoracic Society (stage I=FEV 1 50% predicted, stage II=FEV % predicted). 12 We excluded participants who had any serious or unstable medical disorders that might affect lung function or for which bupropion SR was contraindicated, or if they had a current diagnosis of major depression. Patients were allowed to continue to take medications for COPD except for oral corticosteroids and theophylline. Study design Our study was a parallel-group, randomised, double-blind, placebo-controlled trial done at 11 centres in USA, and it consisted of a 1-week screening phase, a 12-week treatment phase, and follow-up at month 6 (week 26). The study was approved by ethics committees or the institutional review boards of the participating centres, and was done in accordance with the Declaration of Helsinki. We obtained written informed consent from all participants. During the screening phase, eligible individuals were randomly assigned bupropion SR 150 mg once daily for days 1 3, followed by 150 mg twice daily for days 4 84, or matching placebo in a 1/1 ratio. Randomisation was done according to a randomisation code provided by Glaxo Wellcome, using block sizes of four stratified by centre. Within each block of four, two participants were assigned placebo and two bupropion SR. The randomisation codes were kept at the study sites during the trial and we instructed investigators to break the code only for a medical emergency. During the screening phase, individuals selected a date to stop smoking (target day) and were told not to attempt to stop before this day. Study medication was started 1 week before the target day and continued throughout the treatment phase. Placebo tablets were identical in appearance to bupropion SR. During the treatment phase, participants visited the clinic at weeks 1 7, 10, and 12. At each clinic visit, we measured exhaled carbon monoxide with Micro III Smokerlyzer monitors (Bedfont Scientific, Medford, NJ, USA). At all sites, a trained professional provided personalised counselling on stopping smoking by telephone 3 days after the target date, and brief face-to-face counselling at each clinic visit during the treatment phase to encourage smoking cessation and to prevent relapse. The trained counsellor was generally a nurse or other health professional and was formally certified by each clinic for proficiency in such counselling based on information provided by the National Cancer Institute. We asked participants to complete diaries each evening that included questions on the number of cigarettes smoked during the day. The frequency and nature of adverse events were recorded at each clinic visit during the treatment phase. The primary efficacy measure was continuous abstinence from smoking for 4 weeks from the start of week 4 to the end of week 7. We defined continuous abstinence as a participant report of zero cigarettes per day confirmed by exhaled carbon monoxide values of 10 parts per million (ppm) or less. 200 assigned placebo 411 patients enrolled 18 withdrew consent 16 lost to follow-up 10 adverse events 8 other 148 assessed at 12 weeks 5 lost to follow-up 14 other 129 assessed at 6 months Figure 1: Trial profile 404 took at least one dose of study treatment 7 lost to follow-up 204 assigned bupropion SR 12 withdrew consent 11 lost to follow-up 11 adverse events 3 other 167 assessed at 12 weeks 6 lost to follow-up 12 other 149 assessed at 6 months Secondary measures of efficacy were continuous abstinence during weeks 4 12 and 4 26, and point prevalence of abstinence at each clinic visit. Continuous abstinence for weeks 4 26 was defined by participants being continuously abstinent during weeks 4 12, having a diary cigarette count of zero during weeks 13 26, and having exhaled carbon monoxide values of 10 ppm or less at week 26. Point prevalence of abstinence (smoking abstinence during the previous 7 days) was assessed at each clinic visit during the treatment phase and at the 6 month follow-up. We measured bodyweight, and participants completed the University of Wisconsin Center for Tobacco Research and Intervention craving and withdrawal questionnaire at each clinic visit. This questionnaire was scored by rules of the Wisconsin smoking withdrawal scale. 13 Statistical analysis We did all analyses on the intention-to-treat population, which consisted of patients who took at least one dose of study medication. All participants who withdrew from the study were taken to be smokers thereafter. The study sample size was computed a priori 14 to test the null hypothesis of no difference in abstinence rates between bupropion SR and placebo groups, at the nominal 0 05 level of significance and 80% power. We chose a sample size of a minimum of 400 participants since true abstinence rates in COPD patients with bupropion SR were not known but were expected to be about 20 25% compared with 10% for placebo. 15 Data were analysed using SAS (version 6.12). To manage inflation in the rate of false positive findings, we separated the set of outcome variables into families according to clinical importance, and type I error rate was controlled within each family. Family 1 was the primary endpoint of continuous abstinence in weeks 4 7. This endpoint was chosen for its clinical importance and has been used in previous bupropion SR clinical trials. 7,8 Family 2 was point prevalence at weeks 7 and 12, continuous abstinence from smoking at week 12, continuous 1572 THE LANCET Vol 357 May 19, 2001

3 abstinence from smoking (permissive definition) at weeks 7 and 12, and continuous abstinence from smoking at the end of the treatment phase (weeks 9 12). Family 3 was abstinence endpoints relating to the follow-up phase and select withdrawal items on the Wisconsin smoking withdrawal scale. The endpoints in family 3 are deemed exploratory, hence control for type I error is not imposed. We studied the effect of COPD stage and age and sex of patients with multivariate logistic regression. 16 These variables were added to the basic model including treatment group assignment and centre. The analysis of the Wisconsin smoking withdrawal scale is based on change from baseline symptom and overall scores. We compared treatment groups by use of ANCOVA, with baseline values as a covariate. Results 411 patients were enrolled (350 stage-i COPD, 61 stage-ii COPD), of whom 404 took at least one dose of study medication and comprised the intention-to-treat population. The number of patients enrolled by each centre ranged from 15 to 56. Thus, no one centre contributed more than 14% of the study population. 200 individuals received placebo and 204 received bupropion SR (figure 1). The smoking histories of all randomised participants in each treatment group were closely similar for cigarettes smoked per day, number of previous attempts to stop smoking, and time since last attempt to stop (table 1). Additionally, the groups were similar for previous use of methods to stop smoking and pulmonary function test results. The proportion of patients with stage I or II COPD was evenly distributed between treatment groups, as were the participants with subdiagnoses of emphysema, bronchitis, or small-airways disease (table 1). The rates of continuous abstinence from week 4 to the end of week 7 differed significantly between patients receiving bupropion SR and those receiving placebo (p=0 003). 57 (28%) of 204 participants receiving bupropion SR remained abstinent compared with 32 (16%) of 200 receiving placebo. Bupropion SR-treated Bupropion SR Placebo (n=205) (n=206) Characteristics of patients Female/male 93 (45%)/113 (55%) 92 (45%)/113 (55%) White 194 (94%) 197 (96%) Mean (SD) age (years) 53 2 (9 0) 54 5 (9 5) History of depression 37 (18%) 47 (23%) Smoking characteristics Mean (SD) cigarettes smoked per day 28 7 (11 1) 27 6 (10 2) during previous year Mean (SD) pack-year history (years) 52 6 (24 8) 51 4 (23 8) Mean (SD) age when started smoking 16 5 (3 5) 17 3 (4 1) (years) Previous unaided quit attempts 157 (76%) 160 (78%) Previous use of nicotine patch 101 (49%) 100 (49%) Previous use of nicotine gum 66 (32%) 74 (36%) Other smokers in household 68 (33%) 57 (28%) Extremely sure of wanting to quit 138 (67%) 150 (73%) Quite sure of wanting to quit 66 (32%) 55 (27%) Mean Fagerström score (SD) 7 1 (1 7) 7 0 (1 7) Pulmonary function tests and subdiagnoses Stage-I COPD 175 (85%) 174 (85%) Stage-II COPD 31 (15%) 31 (15%) Mean (SD) FEV 1 % predicted 73 2 (19 4) 69 4 (17 3) Emphysema 78 (38%) 84 (41%) Bronchitis 136 (66%) 137 (67%) Small-airways disease 33 (16%) 35 (17%) Determined at screening visit using structured clinical interview for DSM-IV (SCID); score derived from Fagerström tolerance questionnaire, which measures nicotine dependence. Range of scores is 0 to 11, with scores of 6 or greater indicating high levels of nicotine dependence. Table 1: Baseline characteristics Proportion of patients abstinent (%) Bupropion SR Placebo Study week Figure 2: Rates of continuous abstinence for weeks 4 12 and 4 26 p<0 05. patients also had higher abstinence rates than placebotreated patients during the last 4 weeks of treatment (45 [22%] vs 24 [12%], p=0 011). Rates of continuous abstinence were significantly higher with bupropion SR than with placebo throughout the 12- week treatment phase and at the 26 week follow-up visit (figure 2). At the 6 month follow-up, significantly more participants receiving bupropion SR remained abstinent than those receiving placebo (p=0 04). Point-prevalence abstinence rates were greater in the bupropion SR group than in the placebo group (table 2). Bupropion SR was significantly better than placebo at weeks 2 12 during the treatment phase (p 0 015). These higher rates of pointprevalence abstinence in bupropion SR treated participants compared with placebo were maintained at the 6 month follow-up; but were not significantly different (23% [47] vs 16% [32], p=0 07). A main effects logistic-regression model was adequate for studying the probability of abstinence from week The model accounted for the main effects of COPD stage, smoking history, age, sex, centre, and treatment group assignment. The observed frequencies of abstinence are shown in table 3. The results of the multivariate logistic regression analysis showed significant effects for age (odds ratio per 10 years, 1 65 [95% CI ], p=0 003), smoking history (per 10 years 0 81 [ ], p=0 003), and treatment (2 5 [ ], p=0 001) on the probability of continuous abstinence for weeks 4 7. The association with COPD stage (0 48 [ ], p=0 11) and sex (1 7 [ ], 0 53) remains unconfirmed. Centre effects were significant (p=0 002), which suggests that some centres had higher rates of stopping smoking than others. The adequacy of the main-effects model implies absence of two and higher order interactions. In particular, the Week Continuous abstinence Point-prevalence abstinence Bupropion Placebo p Bupropion Placebp p SR (n=204) (n=200) SR (n=204) (n=200) 5 63 (31%) 35 (18%) (33%) 36 (18%) < (29%) 33 (17%) (30%) 39 (20%) (28%) 32 (16%) (32%) 39 (20%) (24%) 28 (14%) (28%) 32 (16%) (18%) 20 (10%) (29%) 33 (17%) (16%) 18 (9%) (23%) 32 (16%) No cigarettes from week 4 onwards. Participants abstinent from smoking during previous 7 days. Table 2: Abstinence rates during treatment and follow-up phases THE LANCET Vol 357 May 19,

4 Subgroup Bupropion SR Placebo (n/total) (n/total) COPD stage I 51/173 (29%) 30/170 (18%) II 6/31 (19%) 2/30 (7%) Sex Male 34/112 (30%) 22/110 (20%) Female 23/92 (25%) 10/90 (11%) Age, years <60 39/153 (25%) 19/138 (14%) 60 18/51 (35%) 13/62 (21%) Smoking history (pack-years) 30 10/35 (29%) 9/43 (21%) >30 47/169 (28%) 23/157 (15%) Table 3: Observed frequencies of abstinence from week 3 to end of week 7 treatment by centre interaction is negligible; there are no centre effects for the placebo-adjusted abstinence rates. Participants who were point-prevalent abstinent during week 12 were assessed for change in bodyweight. Abstinent men and women receiving bupropion SR showed less bodyweight increase than those receiving placebo during the treatment phase. For abstinent women, the increase in weight for the bupropion SR participants was 2 61 kg (n=25) compared with 4 55 kg (n=13) for the placebo participants. However, this was not a significant difference. For abstinent men, the increase in weight was 3 20 kg (n=34) for the bupropion SR participants and 3 64 kg (n=20) for the placebo group, also not significant. Additionally, weight changes 3 months after the end of treatment at week 26 did not differ significantly between treatment groups. Figure 3 shows the mean changes in overall withdrawal symptom scores from baseline for all participants (irrespective of whether or not they were smoking). Participants in the two groups showed an increase in change from baseline scores after week 1. However, the changes were smaller in the bupropion SR group than in the placebo group at weeks 4, 5, 7, 10, and 12. An analysis of overall and individual symptom scores at the end of treatment (week 12) showed lower change from baseline scores for craving, anger, anxiety, sadness, and overall scores for bupropion-sr treated patients than placebotreated patients (table 4, p<0 05). The two groups did not differ significantly for symptoms of concentration, hunger, and sleep at week 12. A post-hoc analysis of withdrawal symptoms only in abstinent patients showed similar observed differences between treatment groups (table 4). 60 (30%) patients receiving placebo and 90 (44%) receiving bupropion SR had an adverse event that was possibly related to the study treatment. Insomnia (placebo 23 [12%]; bupropion SR 49 [24%]), headache (11 [6%]; 12 [6%]), and dry mouth (10 [5%]; 12 [6%]) were the most frequently reported adverse events in the two Change from baseline score Bupropion SR Placebo Study week Figure 3: Mean change from baseline in overall Wisconsin smoking withdrawal scale scores Missing data imputed as last observation carried forward. Placebo n=199; bupropion SR n=198. p<0 05. p=0 06. treatment groups. 27 participants discontinued study medication because of an adverse event (placebo 13, bupropion SR 14). Six of these participants elected to remain in the study after discontinuing the study treatment (placebo three, bupropion SR three). The most common adverse events leading to discontinuation of study drug were anxiety (five) and insomnia (four) for the bupropion SR group, and headache (three) for the participants receiving placebo. Six patients experienced serious adverse events during the treatment phase (placebo five, bupropion SR one) that led to discontinuation of study medication. In the placebotreated group, transient ischaemic attack was the only serious adverse event thought to be related to study treatment. The serious adverse event in the bupropion SR treatment group was a lower-left extremity arterial occlusion that was judged unrelated to the use of study medication. Discussion Smoking cessation is the most important approach to the treatment of COPD. 17,18 Our results show that bupropion SR nearly doubles the rate of abstinence from smoking over 12 weeks in smokers with mild-to-moderate COPD compared with placebo. This advantage over placebo continued for 3 months after discontinuation of the drug. The smokers in our study were older, less healthy, and had smoked more cigarettes for longer periods than participants in previous studies of bupropion SR for smoking cessation, 7,8 and had more failed attempts at smoking cessation, many of which had relied on nicotine All participants Abstinent participants Bupropion SR Placebo p Bupropion SR Placebo p n Mean (SE) n Mean (SE) n Mean (SE) n Mean (SE) Anger (0 07) (0 07) (0 11) (0 15) Anxiety (0 06) (0 06) (0 10) (0 13) Concentration (0 06) (0 07) (0 11) (0 14) Craving (0 07) (0 07) (0 14) (0 18) Hunger (0 07) (0) (0 10) (0 14) Sadness (0 06) (0 06) (0 10) (0 13) Sleep (0 06) (0 06) (0 12) (0 16) Overall (0 04) (0 04) (0 07) (0 10) n=number of patients providing symptom score. Table 4: Change from baseline scores at week 12 for each symptom domain by Wisconsin smoking withdrawal scale 1574 THE LANCET Vol 357 May 19, 2001

5 replacement therapy. They might, however, have been more refractory to such treatments than other populations. Nevertheless, our results are comparable to those of previous studies of bupropion SR for smoking cessation. Nicotine craving and withdrawal are frequently cited as factors contributing to smoking relapse. 19 In our bupropion SR group, however, there was a reduction in overall mean symptom scores from week 4. Additionally, at the end of treatment, overall symptom scores and several individual symptom scores, including craving, significantly favoured the bupropion SR treatment group. When we assessed only abstinent patients, the observed differences were larger between treatment groups, yet were not significant. This finding is probably because of the small size of our population. Women are generally thought to experience more difficulty in stopping smoking than men 20 even if they have COPD. 21 This difference between sexes has been attributed partly to concerns over weight gain after cessation, 22 although other factors could have a role. In our study, women who were abstinent at the end of treatment in the bupropion SR group had gained a mean of about 2 kg less than those in the placebo group, although this difference was not significant. Additionally, whereas fewer women than men stopped smoking overall, the placebo-adjusted response rate to bupropion SR seems higher in women than in men (odds ratio 2 7 vs 1 7). However, these findings are only observational and larger studies are needed. Our study has several potential limitations. The study population was restricted to participants with mild to moderate COPD. We excluded patients with more severe COPD to avoid a disproportionate dropout rate, incidental medical complications, or unexpected drug interactions. The exclusion of such participants does, however, limit the relevance of the study results to patients with mild to moderate COPD. That said, we did note that patients with stage-ii disease had lower abstinence rates than those with stage-i COPD. However, participants with stage-ii COPD responded at least as favourably to treatment with bupropion SR compared with placebo as did those with milder disease, although larger numbers of participants would need to be studied to confirm these preliminary findings. Second, participants with mild to moderate COPD, who are commonly symptom-free and generally have less airflow obstruction than those with symptomatic disease, could be more susceptible to smoking-intervention therapy and thus not be representative of patients with more severe symptomatic COPD. Finally, we did not repeat pulmonary function tests during the study to find out the effects of smoking cessation on lung function. The relatively small size of the participant population and the short study duration were not judged adequate to support such measurements. Additionally, the Lung Health Study 18 has previously shown that predictable and measurable lung function benefits can be achieved with smoking cessation. 4 The adverse event profile in this population was much the same in frequency and quality to previous bupropion SR studies. 7,8 Insomnia, headache, and dry mouth were the most common adverse events observed overall, whereas anxiety and insomnia were the most frequent adverse events leading to discontinuation of bupropion SR. No patient had a seizure during the study, although the product information for bupropion SR describes a seizure rate of 0 1% (1/1000). Overall, bupropion SR was well tolerated. The favourable safety profile and beneficial effects on withdrawal and craving symptoms suggest that bupropion SR is a useful medication for smoking cessation in patients with COPD. Contributors D P Tashkin, R Kanner, W Bailey, S Buist, P Anderson, M A Nides, and D Gonzales were project leaders and contributed to the collection and interpretation of data. G Dozier did the statistical analysis, and M K Patel and B D Jamerson were involved in protocol design and coordinated the study. All investigators were involved in preparation of the report. Acknowledgments This study was funded by Glaxo Wellcome Inc, USA (AK1A4013). We thank Tim Ibbotson and Alok Krishen for writing and editing assistance, and Alan Metz. References 1 Hurd S. The impact of COPD on lung health worldwide: epidemiology and incidence. Chest 2000; 117: 1S 4S. 2 Pauwels RA. National and international guidelines for COPD: the need for evidence. Chest 2000; 117: 20 22S. 3 Roche N, Huchon GJ. Current issues in the management of chronic obstructive pulmonary diseases. Respirology 1997; 2: Scanlon PD, Connett JE, Waller LA, et al. Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease: the Lung Health Study. Am J Respir Crit Care Med 2000; 161: Parrott S, Godfrey C, Raw M, West R, McNeill A. Guidance for commissioners on the cost effectiveness of smoking cessation interventions. Thorax 1998; 53: S Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest 2000; 117: 5S 9S. 7 Hurt RD, Sachs DP, Glover ED, et al. A comparison of sustainedrelease bupropion and placebo for smoking cessation. N Engl J Med 1997; 337: Jorenby DE, Leischow SJ, Nides MA, et al. A controlled trial of sustained-release bupropion, a nicotine patch, or both for smoking cessation. N Engl J Med 1999; 340: Schapira RM, Reinke LF. The outpatient diagnosis and management of chronic obstructive pulmonary disease: pharmacotherapy, administration of supplemental oxygen, and smoking cessation techniques. J Gen Intern Med 1995; 10: Cosio M, Ghezzo H, Hogg JC, et al. The relations between structural changes in small airways and pulmonary-function tests. N Engl J Med 1978; 298: Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981; 123: American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995; 152: S Welsch SK, Smith SS, Wetter DW, Jorenby DE, Fiore MC, Baker TB. Development and validation of the Wisconsin Smoking Withdrawal Scale. Exp Clin Psychopharmacol 1999; 7: Fleiss JL. Statistical methods for rates and proportions, 2nd edn. New York: John Wiley, Crowley TJ, MacDonald MJ, Walter MI. Behavioural anti-smoking trial in chronic obstructive pulmonary disease patients. Psychopharmacology (Berl) 1995; 119: Hosmer DW Jr, Lemeshow S. Applied logistic regression. New York: John Wiley, Kanner RE, Connett JE, Williams DE, Buist AS. Effects of randomized assignment to a smoking cessation intervention and changes in smoking habits on respiratory symptoms in smokers with early chronic obstructive pulmonary disease: the Lung Health Study. Am J Med 1999; 106: 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 FEV1: the Lung Health Study. JAMA 1994; 272: Killen JD, Fortmann SP. Craving is associated with smoking relapse: findings from three prospective studies. Exp Clin Psychopharmacol 1997; 5: US Surgeon General. The health consequences of smoking for women: a report of the Surgeon General. Washington, DC: US Department of Health and Human Services, Bjornson W, Rand C, Connett JE, et al. Gender differences in smoking cessation after 3 years in the Lung Health Study. Am J Public Health 1995; 85; Pirie PL, McBride CM, Hellerstedt W, et al. Smoking cessation in women concerned about weight. Am J Public Health 1992; 82: THE LANCET Vol 357 May 19,