Bupropion for Smoking Cessation in Patients Hospitalized With Acute Myocardial Infarction

A Randomized, Placebo-Controlled Trial

Study design and patient population

The ZESCA (Zyban as an Effective Smoking Cessation Aid for Patients Following an Acute Coronary Syndrome) trial (NCT00689611) was a multicenter, double-blind, randomized, placebo-controlled trial that examined the efficacy and safety of bupropion for smoking cessation in patients with AMI (Online Table 1). A total of 392 patients who smoked at least 10 cigarettes per day and who were hospitalized with an enzyme-positive AMI were enrolled (Fig. 1). Eligible, consenting patients who met the inclusion and exclusion criteria (Online Table 2) were randomized at least 24 h before discharge in a 1:1 ratio to receive bupropion sustained-release or placebo. Randomization was done via an internet website using random blocks of 2 and 4 and was stratified by center to ensure that similar numbers of patients were randomized to the 2 arms of the study at each study center. Each center received ethics approval from their research ethics board before the start of the trial.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Randomization and Follow-Up of Study Patients

*Two patients who withdrew consent did not receive treatment. †One patient who withdrew consent did not receive treatment. ‡Includes all patients except those who died. For the ITT analysis, patients who were lost to follow-up or withdrew were assumed to have gone back to smoking at baseline rates. §Of the 392 patients enrolled, 3.8% died and 22.2% were lost to follow-up. Among those lost to follow-up, vital status was obtained via medical charts for 8.7% of patients. Vital status was obtained for 10 patients in the bupropion group and 9 patients in the placebo group among those who did not return for clinical follow-up. ITT = intention-to-treat; SR = sustained-release.

Baseline assessment, intervention, and follow-up

After randomization, a baseline questionnaire was used to document demographic, smoking, and clinical characteristics of patients. In addition, nicotine dependence and severity of depressive symptoms were measured using the Fagerström Test for Nicotine Dependence (15) and the Beck Depression Inventory II (16), respectively.

The behavioral counseling provided in the trial consisted of minimal clinical intervention that included brief advice to “stop smoking” and the importance of smoking cessation post-MI. This intervention was administered in-hospital by the attending physician before randomization and was approximately 2 min in length. Additional counseling was received at baseline (post-randomization) and at all follow-ups (both telephone and clinic visits). The sessions consisted of brief advice delivered in <20 min (average of 5 min) to “stop smoking” and was based on the “5 A's” model: 1) Ask about smoking; 2) Advise all smokers to quit; 3) Address willingness to quit; 4) Assist patients who want to quit; and 5) Arrange follow-up visits. The counseling consisted of a discussion of withdrawal symptoms, triggers for relapse, and strategies and motivation to help patients remain abstinent or strategies to achieve abstinence in patients who relapsed (Online Tables 3 and 4). Counseling was administered to all patients by the research nurses at baseline and follow-up visits. In addition to the counseling provided by the research nurses, patients were allowed to receive supplementary counseling if their site had a smoking cessation clinic.

Patients received their first dose of study medication within 24 h of randomization and while they were still in-hospital. Bupropion was administered as 150 mg daily for 3 days, followed by 150 mg twice daily for the remainder of the 9-week treatment period. Patients in the placebo group received a matching placebo administered with the same schedule. Patients stopped smoking at the time of hospital admission. At the time of randomization, they were instructed not to resume smoking. During the 9-week treatment period, patients experiencing intractable insomnia were allowed to reduce their study medication to 1 pill every morning. The study was funded by the Canadian Institutes of Health Research and by the Heart and Stroke Foundation of Quebec. No industry support was received, either financial or in the form of study medication. Bupropion and identical placebo tablets were purchased from Sandoz Canada Inc. (Boucherville, Quebec, Canada). An independent industrial pharmacist was responsible for coating, bottling, and packaging the medication for distribution to collaborating sites.

Follow-up consisted of weekly phone calls by research nurses at weeks 1 and 2, as well as clinic visits at weeks 4 and 9, and months 6 and 12 (Online Fig. 1). Phone calls were used to assess side effects and withdrawal symptoms. In addition, patients were asked if they had smoked in the past week and whether they were taking their study medication. During clinic visits, smoking status and clinical events were assessed. Smoking status assessment consisted of a self-report of smoking in the previous 7 days and biochemical validation by measurement of exhaled carbon monoxide (CO) levels. A threshold of exhaled CO ≤10 ppm on a CO monitor (Micro 4 Smokerlyzer, Bedfont Scientific Ltd., Rochester, United Kingdom) was used to confirm smoking abstinence (17).

Screening logs were not mandatory in this trial. Nevertheless, screening logs were forwarded to us from 15 of the 38 collaborating sites. From the logs of these 15 sites, 1,081 patients were screened to identify 778 current smokers (≥10 cigarettes/day). Of the 778 smokers, 584 were excluded for various reasons (e.g., did not meet myocardial infarction definition [39.6%], history or current diagnosis of a psychiatric disorder [10.6%], renal failure/impairment [5.0%], history of seizure or use of medication that lowers seizure threshold [4.1%], use of illegal drugs [4.5%], excessive alcohol use [3.1%], hepatic failure [2.7%], currently using a pharmacotherapy for smoking cessation [1.7%], not motivated to quit smoking [1.3%], participating in another study [1.4%], patient discharged [1.2%], too ill to participate [0.7%], recently underwent surgery [0.7%], physician refused [0.2%], and various other exclusion criteria [22.8%]), leaving a total of 194 eligible patients. Of the 194 eligible patients, 129 (18.1%) declined to participate. Thus, one third of the eligible smokers at these sites were enrolled in the trial.

Endpoint assessment

The primary end point was point prevalence smoking cessation at 12 months. Smoking cessation was defined as complete abstinence in the week before the 12-month clinic visit (documented by self-report) and a measurement of exhaled CO ≤10 ppm. Smoking cessation also was assessed at 4 weeks, 9 weeks, and 6 months. In addition to point prevalence smoking cessation, we also examined continuous abstinence (defined as negative self-report and CO ≤10 ppm at all follow-up periods) and the number of cigarettes smoked/day as secondary measures of smoking status. Secondary end points included the safety and tolerability of bupropion, based on the incidence of clinical events and nondangerous side effects. All clinical end points were adjudicated by members of the Endpoints Evaluation Committee who were blinded to treatment assignment (Online Table 5). An external Data and Safety Monitoring Board monitored the trial. An interim analysis was conducted after 24 months of patient enrollment and was repeated annually.

Data analysis

The initial descriptive analysis examined the balance of demographic, smoking, and clinical characteristics between the 2 treatment groups. Characteristics of the hospital course also were examined. Descriptive statistics were examined using medians, interquartile ranges, means, standard deviations (SDs), and percentages as appropriate, both overall and within treatment groups.

The primary data analysis examined smoking cessation rates at 4 and 9 weeks, and 6 and 12 months, with the primary end point being point prevalence smoking cessation at 12 months. The primary end point was analyzed as differences in cessation rates between bupropion and placebo groups with corresponding p values calculated using a Fisher exact test. The primary end point was analyzed on an intention-to-treat (ITT) basis. Because loss to follow-up rates >30% are not unusual in smoking cessation trials (8,18), our ITT analysis assumed that those who withdrew consent or were lost to follow-up had returned to smoking at their baseline rates. This assumption is common in smoking cessation trials, because >75% of patients return to smoking at 1 year in most trials even while receiving active pharmacotherapy. In addition to our ITT analyses, we also performed per protocol analyses that involved only patients who returned for follow-up visits.

Secondary data analyses examined serious adverse events, side effects, and medication compliance and were also calculated as differences between the 2 groups. Serious adverse events and nonharmful side effects were compared in the 2 treatment groups at weeks 1, 2, 4, and 9 and months 6 and 12 (cumulative events at 9 weeks being the primary measure for side effects). For safety analyses, patients who withdrew from the study, were lost to follow-up, or died were accounted for by censoring at the time of death or at the last follow-up contact. Statistical analyses were performed using SAS statistical software version 9.2 (SAS Institute Inc., Cary, North Carolina).

Baseline characteristics

Baseline demographic, smoking, and clinical characteristics were well balanced among patients in the 2 arms of the trial (Table 1). The only imbalance noted was a history of diabetes. Approximately two thirds of patients were enrolled at North American sites, and most patients were middle-aged men who had smoked for >3 decades. Patients smoked approximately 1 pack/day, two thirds reported at least 1 prior attempt to quit smoking, and one third reported living with other smokers. One third of all patients had previously used a nicotine replacement therapy or a non-nicotine replacement therapy in an attempt to quit smoking. The scores on the Fagerström Test for Nicotine Dependence ranged between 0 and 10, with 82.2% of patients having moderate or severe nicotine dependence (scores ≥4) (15). The scores on the Beck Depression Inventory II ranged between 0 and 63, with 19.9% having scores consistent with depressive symptomatology (scores ≥14) (16).

Hospital course

Characteristics of the hospital course also were well balanced between the 2 arms of the trial (Table 1). Approximately two thirds of patients presented with an ST-segment elevation myocardial infarction. Although all patients enrolled in the trial had positive cardiac enzymes (creatine kinase-myocardial band fraction, troponin I, or troponin T), more than three quarters had an enzyme elevation ≥3 times the upper limit of normal for their hospital. More than two thirds of patients had a coronary angiogram during the hospital admission, and more than half of patients underwent percutaneous coronary intervention or coronary artery bypass grafting. There were relatively few cases of congestive heart failure, recurrent ischemia, or ventricular arrhythmias. Median length of hospital stay was 4.0 days, and median time from admission to first dose of study medication was 2.0 days. Discharge medications (other than study medication) were similar in both groups.

Smoking cessation

Over the course of the trial, there was a progressive decline in smoking cessation rates so that by 12 months, two thirds of patients had returned to smoking (Fig. 2A). Although there were higher rates of smoking cessation among patients taking bupropion at 4 weeks, 9 weeks, 6 months, and 12 months, these differences did not achieve statistical significance. For the primary end point of smoking abstinence at 12 months, rates were 37.2% and 32.0%, respectively (p = 0.33). A secondary analysis, adjusting for between center differences showed an absolute difference between groups of 3.9%. Per protocol analyses showed similar results.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Smoking Cessation by Treatment Group (ITT Analysis)

Patients who withdrew consent or were lost to follow-up were considered to have returned to smoking. Patient numbers were as follows: at 4 weeks in the bupropion (N = 188) and placebo (N = 200) groups; at 9 weeks in the bupropion (N = 188) and placebo (N = 198) groups; at 6 months in the bupropion (N = 185) and placebo (N = 195) groups; and at 12 months in the bupropion (N = 183) and placebo (N = 194) groups. (A) Point prevalence abstinence. Patients were considered abstinent if they abstained from smoking in the 7 days before the visit through a self-report of 0 cigarettes smoked/day, confirmed by a CO monitor reading ≤10 ppm. (B) Continuous abstinence. Patients were considered abstinent if they abstained from smoking through a self-report of 0 cigarettes smoked/day, confirmed by a CO monitor reading ≤10 ppm at all follow-up periods. ITT = intention-to-treat.

In addition to point prevalence abstinence, we also performed an ITT analysis of continuous smoking cessation (Fig. 2B). These results parallel those of the point prevalence analysis, but further showed that <25% of patients were consistently abstinent during the 12-month follow-up with no statistically significant difference between groups.

Although two-thirds of patients with AMI returned to smoking by 12 months, many of these patients substantially reduced their daily cigarette consumption (Fig. 3). ITT analysis showed that among persistent smokers, there was a reduction in mean daily cigarette consumption from 23.2 ± 10.6 at baseline to 8.4 ± 11.3 at 12 months. Among patients who returned for follow-up at 12 months, cigarette consumption decreased from a mean of 22.8 ± 10.5 at baseline to 4.7 ± 7.3 at 12 months, with no difference between bupropion and placebo groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Mean Daily Cigarette Consumption by Treatment Group (ITT Analysis)

Nonsmokers were excluded from this analysis. Patients who were lost to follow-up or withdrew were considered to have returned to smoking at baseline rates. Patient numbers were as follows: at baseline in the bupropion (N = 190) and placebo (N = 198) groups; at 4 weeks in the bupropion (N = 186) and placebo (N = 198) groups; at 9 weeks in the bupropion (N = 186) and placebo (N = 196) groups; at 6 months in the bupropion (N = 183) and placebo (N = 193) groups; and at 12 months in the bupropion (N = 181) and placebo (N = 192) groups. ITT = intention-to-treat.

Counseling

The mean duration of counseling administered to patients over the course of the trial was 45.0 ± 24.3 min. The average time per counseling session was 6.4 ± 3.5 min. If patients relapsed, their treating physicians were allowed to prescribe any type of open-label pharmacotherapy for smoking cessation.

Study drug adherence, adverse events, and side effects

By the end of the 9-week treatment period, 72.3% of patients randomized to the bupropion group were taking at least 1 pill of study medication per day versus 82.0% among patients randomized to the placebo group (p = 0.05) (drug adherence data available for 230 patients). Serious adverse events were rare, and their occurrence was well balanced between the 2 arms of the trial (Table 2). Rates of neurologic or psychiatric events were low. A number of side effects were common during the 9-week treatment period, but no statistically significant differences were observed between treatment groups. Less than 5% of all patients received supplementary pharmacotherapy for smoking cessation beyond the treatment medication during the 12-month period.

Previous studies

To our knowledge, there have been no previous trials of pharmacotherapy for smoking cessation performed exclusively in a population with AMI. Many previous studies have demonstrated that both behavioral therapy and pharmacotherapy are effective for smoking cessation in otherwise healthy smokers without coronary artery disease (5,6). However, there have only been a limited number of trials examining the use of pharmacotherapies in patients with cardiovascular disease (19).

Three previous trials examined the use of bupropion in heterogeneous populations of patients with a variety of cardiovascular conditions. Unfortunately, none of these studies reported cessation rates in their AMI subgroups. Tonstad et al. (7) randomized 626 patients (including 307 with AMI >3 months previously) and found that smoking cessation rates at 12 months were 27% in the bupropion group and 12% in the placebo group (p < 0.05). Rigotti et al. (8) randomized 247 patients with acute cardiovascular disease (including 121 hospitalized with AMI) and found cessation rates of 25.0% and 21.3% (p = NS) at 12 months. The study by Rigotti et al. (8) used high-intensity behavioral counseling in addition to bupropion. Planer et al. (9) randomized 151 patients (including 60 hospitalized with AMI) and found cessation rates of 31% and 33% at 1 year (p = NS). Nicotine patch and nicotine gum were each examined in individual trials involving a variety of cardiovascular conditions, but these trials had disappointing results. Varenicline has been shown to have a benefit in outpatients with stable cardiovascular conditions (20). However, 2 studies have presented conflicting results regarding a possible increased incidence of cardiovascular events with the use of varenicline (12,13). Aside from our study, no trials have specifically focused on patients hospitalized with AMI, a patient population who are at high risk for recurrent cardiac events and in whom smoking cessation is of crucial importance.

Our study suggests that patients hospitalized with AMI may be more highly motivated to quit smoking than healthy smokers without coronary artery disease or patients with stable cardiovascular disease. Among the patients in the control arm of our trial, 47% were abstinent at 4 weeks and 32% were abstinent at 12 months. These results are better than those reported in the placebo arms of trials in healthy smokers without coronary artery disease and outpatients with stable cardiovascular disease and suggest that patients with AMI are more highly motivated to quit.

The “teachable” moment presented by a smoking-related hospital admission for AMI is thought to be an opportune time to encourage patients to quit smoking (10). In addition to our study, several observational studies have documented high initial quit rates after an AMI only to be followed by high relapse rates. Compared with smokers enrolled in previous smoking cessation trials, most patients with AMI are older and have been smoking for many decades. Despite their motivation to quit, their long-standing nicotine addiction may make them less responsive to smoking cessation therapies.

We observed that many persistent smokers reduced their daily cigarette consumption from approximately 1 pack/day at baseline to less than 10 cigarettes/day at 12 months. Although some studies suggest that reducing consumption does not affect disease risk (21), the INTERHEART study group found a linear relationship between the number of cigarettes smoked/day and the risk of a de novo AMI (22). Even if reduced consumption does not reduce disease risk, it indicates a subset of patients who could be retargeted for smoking cessation interventions.

Study limitations

Several potential limitations of our study should be noted. First, our study recruited only patients with AMI who were willing to attempt to quit smoking. A large number of patients were approached for inclusion in this study who were unwilling to even consider quitting. For this reason, the quit rates reported in our study are likely higher than those that would be observed in the AMI population at large. Second, similar to other smoking cessation trials, a not insubstantial number of patients withdrew from our study or were lost to follow-up. Previous smoking cessation trials have reported dropout rates >30%. Among the 377 surviving patients in our study, 22.2% withdrew consent or were lost to follow-up. These results are better than those observed in previous trials and are likely due to the higher level of motivation among patients with AMI. Third, our study included a low-intensity behavioral counseling intervention and similar to the study by Planer et al. (9) found no effect of bupropion. However, the study by Rigotti et al. (8) administered high-intensity behavioral counseling intervention in addition to bupropion therapy and found similar results. Finally, because of the relatively small numbers of serious adverse events that occurred, we had limited power to examine our secondary safety end points.