Randomized Trial of Ultrashort Psychotherapy vs Sustained-Release Bupropion for Smoking Cessation

Study design

This randomized clinical trial was designed to compare the efficacy and safety of a manual-based 1.5-day psychotherapeutic intervention, i.e., Psychodynamic Model Training®, with a medication, i.e., sustained-release bupropion (Zyban®), as smoking cessation aids. In order to render the data comparable to previously obtained bupropion data, the bupropion treatment regimen and study protocol closely followed that by Jorenby et al. 1999.(3) As an improvement over previous protocols(3-5), nonsmoking status was verified not only by determination of exhaled carbon monoxide (CO) levels but also by tandem mass spectrometric (LC/MS/MS) analysis of urine cotinine(6;7), a marker of higher sensitivity than CO.(8)

Subjects, screening, inclusion and exclusion criteria, and randomization

Female and male smokers were recruited between July 2005 and December 2005 by announcements in the local media (press, local radio stations) and in the newsletter (http://www.stgkk.at/mediaDB/104832.PDF) of the governmental Styrian Regional Health Care System (Steiermaerkische Gebietskrankenkasse, STGKK), the sponsor of the present study. Interested smokers were first screened by telephone interview and then scheduled for the study intake interview. The first subject was enrolled in July 2005, and follow-up was completed in January 2007. Of a total of 951 persons screened, 790 met the screening criteria and were enrolled and randomized to one of two treatments according to a randomization list in the care of the steering committee at the Medical University Innsbruck. Great care was taken to preserve treatment allocation concealment: Only after the subject had given his/her written consent, a sealed envelope containing the treatment allocation was broken by the interviewer and presented to the subject. The study protocol was registered with the European Medicines Agency (emea.europa.eu; EudraCT-Nr. 2005-006189-32) and was approved by the local governmental review board. The study was conducted in compliance with the ethical principles of the Declaration of Helsinki (www.wma.net/e/policy/b3.html ; accessibility verified May 16, 2007) and the standards on good clinical practice developed by the International Conference on Harmonization (www.ich.org/cache/compo/276-254-1.html ; accessibility verified May 16, 2007.)

Inclusion criteria were an age of at least 18 years, a cigarette consumption of at least 15 cigarettes per day in the last 3 months, and the willingness to contribute EUR 70 to the treatment costs, regardless of the treatment modality. Exclusion criteria, as assessed by the study-site physicians, were any psychiatric diagnosis except nicotine or caffeine dependence, any current psychopharmacological treatment, known hypersensitivity to bupropion, epilepsy, liver cirrhosis, brain tumor, pregnancy (verified by an immunological test at the intake interview and 3 weeks after the first day of the first menses occurring within treatment), lactation, any serious or unstable cardiac, renal, hypertensive, pulmonary, endocrine, or neurologic disorder, ulcers, dermatologic disorders, current use of other smoking cessation treatments, and regular use of any noncigarette tobacco product. After induction of the first 620 subjects, it became apparent that more subjects randomized to the bupropion group dropped out immediately after learning of their treatment allocation. To compensate for this higher rejecter rate in the bupropion group, the remaining 170 subjects were randomized to bupropion vs psychotherapy at a ratio of 2:1. Seven of the 420 subjects randomly allocated to the bupropion group and 4 of the 370 to the psychotherapy group were found to meet exclusion criteria by the study's physicians only after the randomization (bupropion: bulimia, 3 cases of major depression, 2 cases with antidepressant medication, anticoagulant medication; psychotherapy: 4 cases with a previously undisclosed psychiatric diagnosis) and had to be excluded before treatment started. Therefore, the final intention-to-treat (ITT) sample consisted of 413 subjects in the bupropion and 366 subjects in the psychotherapy group.

Treatments

The treatment period in the sustained-release bupropion (Zyban®) group was nine weeks(3), following the protocol given in Zyban®'s product information as provided by its manufacturer (GlaxoSmithKline, Vienna, Austria): After an initial medical consultation, a dose run-up period of 1 week to a final 150 mg sustained-release bupropion twice daily on day 7, a target quitting date was set for the second week, usually day 8. Subjects returned for a second visit, scheduled between days 26 and 33 of the treatment, to the study's physicians for a medical checkup and to receive the second 60-tablet package of Zyban®. Failure to show up for this second medical checkup was considered treatment failure.

Psychodynamic Model Training® is a manual(9)-based psychodynamically oriented ultrashort psychotherapeutic intervention, consisting of a very brief psychoeducation and an initial training in autosuggestion techniques that are taught during a single 1.5-day program (day 1, 0900-2000 h; day 2, 0900 - 1300 h). Subjects, approximately 30 per group, were introduced to five guided imagery scenarios aimed at strengthening the following psychotherapeutically defined aspects of functioning: (1) self-worth, self-determination, self-assurance; (2) self-acceptance; (3) health, vitality, relaxation, awareness of bodily functions; (4) creativity, self-worth, remuneration and emotional reward for one's work; and (5) ability to sustain intimate personal relationship.

Follow-up period

Subjects of either group were contacted by telephone to schedule follow-up assessments in the study physicians' offices at the Institute for Clinical Psychology, Psychotherapy, and Health Promontion, Graz, at the end of months 3, 6, and 12, and had to present for the appointment within 36 h of the telephone call in order to facilitate the detection of any smoking. During these assessments, subjects were asked about their non/smoking status, their breathalyzer CO levels were determined by associates of the study's sponsor, the STGKK, and they had to provide a urine sample for cotinine analysis. Family members of randomly selected subjects also received telephone calls from a staff member of the STGKK to confirm their non/smoking status. No relapse-prevention counseling was administered during the follow-up period.

Assessments

At baseline, demographic data, medical and smoking history were obtained, and vital signs and exhaled CO levels were determined. The Beck Depression Inventory(10) was used to assess depressive symptoms (scores of 0 to 9 are considered to be normal, scores of 10 to 18 indicated mild-to-moderate depression, scores of 19 to 29 indicated moderate-to-severe derpession, and scores of 30 to 63 indicate severe depression). The Fagerström Tolerance Questionnaire(11;12) was used to quantify the severity of nicotine dependence (scores of 0 to 11, with higher scores indicating more severe dependence). A separate questionnaire contained the portion of the Structured Clinical Interview for the Diagnostic and Statistical Manual version 4 (DSMIV)(13) concerning substance dependence disorders.

During follow-up interviews, CO levels were determined in the expired air by breathalyzer (piCOsmokerlyzer®, www.bedfont.com). Urine samples were obtained and quantified for cotinine, a major metabolite of nicotine with a urine elimination half-life around 20 h,(7) by HPLC coupled with tandem mass spectrometry (LC/MS/MS) as previously described.(14) The cotinine (mass transition 177 > 80 m/z) standard curve (internal standard, 50 ng/ml d3-methadone) was essentially linear up to the highest concentration tested, i.e., 2000 ng/ml (quadratic equation fit, weighting 1/x, r>0.99). The level of detection (LOD) was 13 ng/ml at a signal/noise ratio of 2/1). The level of quantification (LOQ) in the actual urine samples of the study was 30 ng/ml.

Measures of outcome

The primary outcome criterion was a biochemically/pharmacologically-confirmed 12-month continuous abstinence as evidenced by self-reports of nonsmoking status in all interviews (i.e., at 3, 6, and 12 months). Self-reports of nonsmoker status were confirmed by urine cotinine levels below the level of detection (13 ng/mL at a signal/noise-ratio of 2/1 in the LC/MS/MS analysis) at all 3 follow-up interviews. Subjects in the bupropion group had to report abstinence from the target quitting date onwards, subjects in the psychotherapy group had to report abstinence from the end of the 1.5-day program onwards. Subjects who self-reported nonsmoker status but for which urine samples could not be collected or at any of the 3 follow-up time points or which refused to provide a urine sample at any of the 3 follow-up time points were considered smokers. In order to better compare the data of the present study with previously published bupropion data,(3;4) nonsmoker status was confirmed, independently of the urine cotinine analysis, by exhaled carbon monoxide (CO) levels at or below 10 parts per million (ppm) at all 3 follow-up interviews, following the nonsmoker criterion of Jorenby and coworkers for low-frequency follow-ups.(3;4) The secondary outcome was point prevalence of self-reported abstinence confirmed by urine cotinine levels below the level of detection for each follow-up time point. Outcome criteria were assessed both in the ITT- and the completer-sample.

Statistical analysis

The sample size of approximately 380 subjects per group was determined such that, under standard conditions regarding power (80%) and two-sided alpha level (0.05), the following differences between the two treatment groups may be detected: continuous abstinence rates of 17% in one group vs 10% in the other or 23% vs 15 % or 29% vs 20%. Baseline comparisons of the two treatment groups were performed by the chi-square test for categorical data, Student's t-test for normally distributed and Mann-Whitney U-test for non-normally distributed numerical variables. Differences between the two groups regarding primary (12-month continuous abstinence rate) and secondary (point prevalence of abstinence) outcome criteria, TEAE rates, and rejecter rates were analysed by means of the chi-square test (two-sided). In order to quantify the differences between the two groups, the Odds Ratio (OR), Relative Benefit (RB; in analogy to the Relative Risk), and the Number Needed to Treat (NNT) were determined. Agreement between the two detection methods (cotinine level, CO level) was evaluated by the kappa coefficient, significant differences in continuous abstinence rates between the two detection methods were determined by means of the McNemar test. All analyses were performed by a certified statistician using the SPSS® software (www.spss.com). Measures of variance were either the standard deviation (S.D.) for the treatment groups' baseline characteristics or the 95 % confidence interval for OR, RB, and NNT.