No job name

RESEARCH REPORT
Varenicline in the routine treatment of tobacco
dependence: a pre–post comparison with nicotine
replacement therapy and an evaluation in those with
mental illness
John A. Stapleton1,2,3, Lucy Watson2, Lucy I. Spirling2, Robert Smith2, Andrea Milbrandt2,
Marina Ratcliffe2 & Gay Sutherland2,3
Health Behaviour Research Centre, Department of Epidemiology and Public Health, University College London, UK,1 South London and Maudsley NHSFoundation Trust, Tobacco Dependence Clinic, UK2 and Kings College London, Department of Psychological Medicine, Institute of Psychiatry, UK3 ABSTRACT
To compare the effectiveness of varenicline with nicotine replacement for smoking cessation and to evaluate the safety and effectiveness of varenicline in people with mental illness. Design
cases before and after the introduction of varenicline. Setting
National Health Service (NHS) tobacco dependence clinic in London, UK. Participants
A total of 412 cases receiving routine care. Intervention
port sessions over 6 weeks with either nicotine replacement therapy (NRT) (n = 204) or varenicline (n = 208).
Measurements
Verified abstinence 4 weeks after quit day, severity of withdrawal symptoms, incidence and severity of adverse drug symptoms, cost per patient treated and cost per successful short-term quitter. Findings
cessation rates were higher with varenicline than NRT (odds ratio = 1.70, 95% confidence interval = 1.09–2.67).
Varenicline was equally effective in those with and without mental illness. Craving to smoke, but not adverse mood, was less severe with varenicline than NRT. The cost per quitter was similar for varenicline and NRT. There was a higher incidence of adverse drug symptoms among those taking varenicline, but these were tolerated by most smokers. There was no evidence that varenicline exacerbated mental illness. Conclusions
varenicline appears to improve success rates over those achieved with NRT, and is equally effective and safe in those Keywords
Cost, craving, mental illness, nicotine replacement, smoking cessation, varenicline.
Correspondence to: John Stapleton, Health Behaviour Research Centre, Department of Epidemiology and Public Health, University College London, 2-16 Torrington Place, London WC1E 6BT, UK. E-mail: Submitted 18 September 2007; initial review completed 4 October 2007; final version accepted 24 October 2007 INTRODUCTION
UK) [4–8]. Smokers with mental illness were excluded in these trials, and some have expressed concern about the Varenicline (Champix/Chantix, Pfizer Ltd, Surrey, UK) is safety and efficacy of varenicline in this substantial sub- an alpha4beta2 nicotinic acetylcholine receptor partial agonist licensed for the treatment of tobacco dependence Nicotine replacement therapy (NRT) has become the by the USA FDA in May 2006 and by the European standard pharmacological treatment for tobacco depen- Agency for the Evaluation of Medicinal Products (EMEA) dence, due to its well-proven effectiveness, benign side- in September 2006 [1,2]. The National Institute for effect profile and easy availability through pharmacy and Health and Clinical Excellence (NICE) for England and general sales [12–14]. Without trials comparing vareni- Wales recently recommended varenicline to the National cline with NRT, a key question as to whether clinicians Health Service (NHS) [3], based on the results of should use varenicline in preference to NRT remains manufacturer-sponsored trials showing superior efficacy unanswered. Due to the large number of different NRT over placebo and bupropion (Zyban; GSK Ltd, Middlesex, products and doses from which smokers currently choose 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction it is unlikely that a blind trial that reflects fully the poten- from excessive NRT use [12]. The second product was tial of NRT will ever be conducted. To provide much- usually prescribed for about 6 weeks, or for 12 weeks at needed data on this issue we conducted an evaluation of half the dose. Those taking varenicline were also given a the first varenicline cases in an NHS tobacco dependence 12-week course on four prescriptions (2 weeks, 2 weeks, clinic, using NRT cases treated immediately prior to the 2 weeks, 6 weeks), with the last dispensed at session 7.
introduction of varenicline as a control cohort and Varenicline was introduced in the clinic at the beginning adjusting comparisons for a comprehensive set of ante- of January 2007, after which a minority of patients cedent patient characteristics potentially prognostic for chose to use NRT. Before this the majority had used NRT, outcome. We were also able to assess the efficacy and rather than bupropion. No patients were excluded from safety of varenicline in those with mental illness.
using NRT and only pregnant women (0), those breast feeding (0), those trying to conceive (1), those under 18 years old (2) and those with severe renal function impairment (3) were excluded routinely from using Patients and procedures
varenicline. NRT was dispensed under a Patient Group Direction by the clinic nurses and varenicline was pre- Patients attended the South London and Maudsley NHS scribed by the resident doctor using FP10 NHS prescrip- Foundation Trust Specialist Tobacco Dependence Clinic tions. NHS prescription charge rules applied to all as routine cases between May 2006 and April 2007. The treatment course comprised seven weekly group support sessions lasting 1–1.5 hours, plus NRT or varenicline.
Smokers were booked consecutively into groups of Materials and measures
between five and 25 depending on availability and pre- ferred time. Throughout the evaluation period sessions Patients completed standard clinic materials. One week were held at the same time of day (morning, afternoon or before the first session they were sent an appointment evening), with the same therapists and using the same letter, an information sheet detailing the treatments procedures. The course required smokers to stop from the and a self-completion questionnaire on demographics, third session (‘quit day’) onwards, following assessment smoking history, degree of tobacco dependence and and preparation sessions in previous weeks.
medical history. The responses were reviewed with a cli- At the first session (‘assessment’) a brief medical nician at the first session. At the start of sessions 2–7 history was taken and the suitability of medicines patients completed a weekly report detailing smoking assessed. In a group, smokers were then introduced to the throughout the week, tobacco withdrawal symptoms and clinic programme and the treatment options, as outlined potential adverse drug reactions. They were then seen previously in written material. At the end of the session individually, where these reports were reviewed, and an patients were asked to decide, subject to contraindica- expired-air carbon monoxide (CO) reading taken to tions, which medicine they preferred to take. Following the recommended schedule, those taking varenicline started treatment after the second group session (‘prepa- ration’) and NRT treatment started 1 week later, immedi- Tobacco withdrawal symptoms and adverse
ately after group session 3 (‘quit day’). Sessions 4–7 were drug reactions
designed to give support throughout the acute period of The self-completion tobacco withdrawal symptoms scale tobacco withdrawal symptoms, to dispense prescriptions consisted of seven known symptoms [15]. Each item was and follow patients until 4 weeks after the quit day.
rated from 1 (not at all) to 6 (extreme). An adverse mood (negative affect) score was created as the mean response Treatments
to the items: depression, irritability, restlessness and diffi- Supported by advice from a clinician, those using NRT culty concentrating. A craving score was created from could choose between all licensed preparations and doses.
the items: difficulty stopping smoking, urges to smoke During the period of this evaluation 60% used the nico- and strength of urges to smoke. Each week, patients we tine patch, 25% used the nasal spray, 11% used the gum asked to report suspected adverse drug reactions using or lozenge and 5% used the inhalator or microtab. NRT the question: ‘If you used Champix, Zyban or Nicotine was dispensed in three batches according to NICE guid- Replacement in the last week, please write below any ance (2 weeks, 2 weeks, 8 weeks), with the last batch unpleasant effects you think they may have caused’.
being dispensed at session 7 [13]. The clinic offered all When reported, the severity of each symptom was self- patients a second NRT product to be used in combination rated on a three-point scale: 1 = mild, 2 = moderate, with the first, although light smokers were discouraged 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction Short-term smoking cessation outcome
Data processing and statistics
To be classified as successful on the primary outcome Following usual practice, all data were entered on the measure (‘CO-verified abstinence’) participants had to clinic database for clinical, monitoring and evaluation report not smoking at all during the final 2 weeks of the report purposes. All patients gave signed consent for this course and record a CO level of less than 10 parts per at the first session. Smoking cessation rates and the inci- million (p.p.m.) at the last session (session 7). As a sec- dence and severity of adverse symptoms were compared ondary measure we also classified patients according to using the OR with 95% confidence intervals (CI), and the Department of Health (DH) criteria (‘DH Self-report logistic regression models were used to adjust for poten- abstinence’), as required of all NHS services [16]. Patients tially confounding antecedent characteristics. To allow successful on this measure were those who were either for possible group effects a beta-binomial model was used ‘CO-verified abstinent’ or, if they did not attend session 7 [19]. The severity of tobacco withdrawal symptoms were for CO verification, could instead self-report abstinence compared using the t-test, with adjustment using normal via telephone or letter. Those who did not attend the last session and who failed to respond to telephone calls and a letter were classified as smoking [17]. All those reporting abstinence in person also passed CO verification in these The varenicline and NRT cohorts were similar with respect to demographic, health history and smoking Evaluation cohorts
characteristics, apart from a higher proportion of patients of white European origin in the varenicline The cohort sizes were selected to be sufficiently large to cohort (Table 1). In particular, there was no difference in detect a difference between varenicline and NRT on the confidence in stopping smoking, as self-rated after the primary outcome measure. Because there are no pub- medicine had been chosen but before it had been lished data comparing varenicline and NRT on which to base this, an estimate of the short-term relative efficacy of varenicline compared with bupropion was used instead [odds ratio (OR) = 1.83] [4,5]. This was considered an Smoking cessation
appropriate substitute given the similar efficacy of bupro- pion and NRT in meta-analyses [12,13,18]. Applying this ratio to the expected NRT cessation rate known from pre- Cessation rates were significantly higher with varenicline vious DH monitoring returns (about 60%) gave an than NRT on both measures of short-term cessation, expected varenicline cessation rate of 73%. With alpha giving an estimated benefit of approximately 11% set at 0.05 and statistical power at 0.8, cohorts of at (Table 2). The effect of adjusting cessation rates for all the least 200 per treatment would be required to detect the characteristics shown in Table 1 was marginal, slightly 13% projected difference. Only entire treatment groups increasing the estimated relative effect of varenicline over were selected for evaluation, resulting in cohorts of 204 NRT. The only characteristics prognostic of cessation in and 208 smokers treated with NRT and varenicline, the multivariate model were: being older (more likely to stop), receiving state benefits (less likely to stop), smoking To achieve an evaluation cohort of at least 200 more (less likely to stop), having a smoking-related illness varenicline patients consecutive groups treated between (less likely to stop) and smoking cannabis (less likely to January and April 2007 were included. The NRT cohort stop). We also found no relation between the size of group for comparison was chosen as consecutive groups in which a patient was treated and their likelihood of treated immediately before the introduction of vareni- cline (May–November 2006). The small number using When group success rates were considered as the unit bupropion during this period were excluded. After of observation there was evidence that the group in varenicline was introduced 77% chose to use it. We which a smoker was treated had affected their likelihood excluded the 23% who did not wish to use varenicline of success as measured by DH self-report abstinence and used NRT instead. A decision to undertake this (c2 = 68.9, df = 44, P < 0.025) but not according to evaluation was taken after the completion of patient CO-verified abstinence (c2 = 46.5, df = 44, P = 0.30).
treatment but before the clinical data had been analy- Adjustment for the group effect using a beta-binomial sed, following a request from the clinical service com- model slightly increased the size of the confidence inter- missioners for an assessment of the cost implications of val for DH self-report abstinence (OR = 1.92, 95% CI = 1.04–3.56, c2 = 4.68, P < 0.05).
2007 The Authors. Journal compilation 2007 Society for the Study of Addiction Table 1 Demographic, health and smoking characteristics of those taking nicotine replacement therapy (NRT) and varenicline.
% Drinking more than 21 alcohol units weekly Number of daily cigarettes smoked mean (SD) % ‘very’ or ‘totally’ determined to stop Confidence in stopping† (1–10) mean (SD) % More than 3 attempts to stop in 5 years *Difference P < 0.05. †Recorded at session 2, after medication known.
Table 2 Short-term smoking cessation rates in those taking nicotine replacement therapy (NRT) or varenicline.
*Four weeks after ‘quit day’. †Adjusted for all characteristics shown in Table 1. CO: carbon monoxide; DH: Department of Health.
Table 3 Short-term smoking cessation rates in those taking single nicotine replacement therapy (NRT), combination NRT or
Adjusted odds ratio† Difference* NRT (n = 83) (n = 208) 57.9 (70/121) 66.3 (55/83) 72.1 (150/208) 1.88 (1.18–3.02) 1.96 (1.15–3.33) % DH self-report 66.1 (80/121) 74.7 (62/83) 80.3 (167/208) 2.09 (1.26–3.47) 2.27 (1.27–4.05) *Single NRT versus varenicline. †Adjusted for all characteristics shown in Table 1. CI: confidence interval; CO: carbon monoxide; DH: Department ofHealth.
Single-product NRT and combination NRT therapy and increased cessation rates by about 14%, equivalent to one additional success with varenicline for Among the 204 patients in the NRT cohort, 121 (59%) every seven smokers treated. However, there was no evi- used a single product and 83 (41%) used two products dence of a difference in success rates between varenicline simultaneously. None of the characteristics shown in and combination NRT (OR for CO-verified absti- Table 1 were associated with whether single or combina- nence = 1.32, 95% CI = 0.76–2.27 and OR for DH tion therapy was used. For both outcome measures the self-report abstinence = 1.38, 95% CI = 0.76–2.52).
observed cessation rates for combination NRT were higher Adjustment for background characteristics only margin- than for single NRT therapy (Table 3). Varenicline was ally altered this difference, increasing slightly the esti- significantly more effective than single-product NRT 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction Table 4 Short-term smoking cessation rates in those with mental illness.
*Adjusted for all characteristics shown in Table 1, with the exception of mental illness. NRT: nicotine replacement therapy. CI: confidence interval;CO: carbon monoxide; DH: Department of Health.
Table 5 Tobacco withdrawal symptom scores.*
*In those attending the clinic 1 week after ‘quit day’ and recording a carbon monoxide reading of less than 10 parts per million. †Scored: 1 = not at all,2 = slight, 3 = moderate, 4 = strong, 5 = very strong, 6 = extreme. ‡Adjusted for all characteristics shown in Table 1. NRT: nicotine replacement therapy;CI: confidence interval.
dence of nausea, disturbed sleep, vivid dreams, drowsi- ness, constipation, headache, dyspepsia, dry mouth, bad One hundred and eleven smokers (27%) reported that taste, low mood, diarrhoea and disorientation in those they were currently receiving treatment for mental illness taking varenicline (Table 6). Skin irritation (related to (primary diagnosis: depression 64, bipolar disorder 14, nicotine patch use) was the only reaction with a higher psychosis seven, psychosis and depression 24, eating dis- incidence in those using NRT. However, there was little order two). In these patients there was a similar, or evidence that when experienced, the severity of symp- slightly greater, advantage for varenicline over NRT than toms were different in the two cohorts, with the excep- seen in the whole group, although the confidence inter- tion of anxiety/panic which was reported as either vals were wide due to the reduced sample size (Table 4).
moderate or severe by all seven cases in the varenicline Adjustment for background characteristics increased the Among the 208 patients who started their quit attempt using varenicline, seven (two with mental Tobacco withdrawal symptoms
illness) switched to using NRT due to adverse symptoms.
Tobacco withdrawal symptoms self-rated 1 week after Six of these stopped smoking successfully. There was no the start of the quit attempt, when severity frequently evidence that adverse symptoms were experienced more peaks, were compared between the NRT and varenicline in those with mental illness, or that when experienced, cohorts [15]. To ensure that only those who could have the symptoms were more severe (Table 6). There were no potentially experienced withdrawal symptoms were reports of mental illnesses symptoms being exacerbated included the analysis was restricted to those who had by varenicline. Two reports were submitted under the either stopped smoking completely or who had reduced Medicines and Healthcare Regulatory Authority moni- their smoking sufficiently to record a CO level of less toring scheme. One recent eye surgery patient experi- than 10 p.p.m. There was no evidence of a difference in enced headaches and blurred vision and another patient experience of adverse mood, although craving severity had a severe psychological reaction likened to a ‘bad was lower among those taking varenicline (Table 5).
LSD trip’, including anxiety, paranoia, confusion and Adjustment for background characteristics did not impaired motor control. Neither was hospitalized.
The average drug costs per patient treated and per suc- Adverse drug reactions
cessful quitter at 4 weeks were calculated. Single-product Symptoms reported significantly more by patients in NRT treatment was costed at an average of £12 per week either cohort are shown in Table 6. Compared with (£144 for a full 12-week course) and combination NRT those using NRT there was a significantly higher inci- therapy at an average of £18 per week (£216 for a full 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction Table 7 Drug costs for nicotine replacement therapy (NRT) and varenicline users (£).
*% Carbon monoxide (CO) verified abstinence. †% Department of Health (DH) self-report abstinence.
12-week course). Approximately 70% of those using NRT ment was known to them. The integrity of these evalu- used products costing about £10 per week (mainly nico- ation results is also supported by our observation of tine patches) and 30% used products costing an average higher success rates with combination NRT therapy of £15 per week. A second product was used on average compared with single NRT therapy. The difference was for the equivalent of half a full 12-week course. Vareni- close to that seen in randomized controlled trials [12].
cline was costed at £13.65 per week or £163.8 per full Interestingly, we observed little difference between the 12-week course. The costs based on prescriptions dis- efficacy of varenicline and combination NRT therapy, pensed are shown in Table 7. The mean cost per patient although this evaluation was not designed with treated was slightly higher for varenicline than NRT adequate statistical power to test this.
overall, but the cost per short-term success was slightly Varenicline was associated with a number of adverse symptoms. Nausea and poor sleep/vivid dreams were the most common symptoms and were reported with a slightly higher incidence in these cohorts than in the DISCUSSION
manufacturer-sponsored clinical trials (nausea: 38% A key question for clinicians treating tobacco dependence versus 29%) (poor sleep/vivid dreams: 43% versus 31%) is whether to offer varenicline in preference to NRT. The [4,5]. The higher rate of sleep disturbance in this cohort published company-sponsored varenicline trials did not may have been because of the high abstinence rate and compare varenicline with NRT and hence did not provide because sleep disturbance is a recognized tobacco with- a direct answer. Indeed, it would be extremely difficult to drawal symptom for many. Additional to the symptoms design a blind trial that reflects adequately the potential identified in the clinical trials we also observed a higher of NRT to be tailored to individual patients. In preference incidence of low mood, disorientation and diarrhoea, to an open-label trial with inherent potential bias we each experienced by 10 patients. Due possibly to either opted to undertake a comparative evaluation of routine the moderate nature of these symptoms, the therapeutic varenicline cases using NRT cases treated immediately effect of varenicline, or the supportive nature of the before the introduction of varenicline as controls. In so group treatment programme, the majority of symptoms doing we were also able to assess varenicline in those with were tolerated by the majority of patients and only seven (3%) switched to NRT after starting treatment The results suggest that, with routine psychological with varenicline. We also found no evidence of more and behavioural group support, varenicline is more adverse symptoms being experienced by those with effective than NRT in aiding short-term smoking cessa- mental illness, although we cannot exclude the exist- tion. The magnitude of the benefit was similar to that ence of other adverse symptoms with low prevalence.
seen for varenicline over bupropion in clinical trials with Under the dispensing and prescribing schedule used individual treatment. The results also indicate that the cost per patient treated was about 7% higher and the varenicline is similarly effective in those with mental cost per short-term success was about 9% lower for vareni- illness, supporting the regulatory decision to allow cline compared with NRT. Other schedules, such as giving varenicline treatment in these patients. There was also the full 12-week course on a single prescription, would evidence that the therapeutic benefit of varenicline over lead to different costs. Given the extremely low cost of NRT may have been due, at least in part, to better these treatments relative to the life-years they gain, such control of urges to smoke, although there was no evi- small differences are unlikely to make cost an important dence of a difference on other withdrawal symptoms factor in prescribing decisions [13]. Although more effec- such as depression and poor concentration. We were tive compared with the single product NRT protocol able to adjust statistically all comparisons for a large usually advocated, varenicline was about 40% more number of antecedent characteristics, including the costly per patient treated and about 11% more costly per patient’s confidence in success recorded after the treat- successful short-term outcome. Again, this difference is 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction small enough not to be a major factor in commissioning idcplg?IdcService=SS_GET_PAGE&useSecondary=true&ssDocName=CON2031589 (accessed 14 November 2007).
3. National Institute for Health and Clinical Excellence (NICE).
Although all procedures, materials and staff were con- Final Appraisal Determination: Varenicline for Smoking Cessa- stant for NRT and varenicline cohorts and we were able to adjust for a variety of potentially confounding character- page.aspx?o=431452 (accessed 30 May 2007).
istics, there remains the possibility that unknown system- 4. Gonzales D., Rennard S. I., Nides M., Oncken C., Azoulay S., atic factors biased the results. To our knowledge, no Billing C. B. et al. for the Varenicline Phase 3 Study Group.
Varenicline, an alpha4beta2 nicotinic acetylcholine recep- major taxation or legal changes occurred during the tor partial agonist, vs sustained-release bupropion and evaluation period to affect the likelihood of success.
placebo for smoking cessation: a randomized controlled Perhaps a greater limitation is the fact that we had avail- trial. JAMA 2006; 296: 47–55.
able only short-term outcome at the end of treatment, 5. Jorenby D. E., Hays J. T., Rigotti N. A., Azoulay S., Watsky E.
whereas tobacco cessation studies are now recommended J., Williams K. E. et al. for the Varenicline Phase 3 StudyGroup. Efficacy of varenicline, an alpha4beta2 nicotinic to follow patients for at least 6 months [20]. However, acetylcholine receptor partial agonist, vs placebo or comparative smoking cessation rates, as measured here sustained-release bupropion for smoking cessation: a ran- by the OR, are usually highly stable over short- and long- domized controlled trial. JAMA 2006; 296: 56–63.
term follow-ups [21], and our results are consistent with 6. Nides M., Oncken C., Gonzales D., Rennard S., Watsky E. J., those from trials comparing varenicline with bupropion Anziano R. et al. Smoking cessation with varenicline, aselective alpha4beta2 nicotinic receptor partial agonist.
and trials comparing single NRT with combination NRT, all of which included 6- or 12-month follow-up.
bupropion-controlled trial with 1-year follow-up. Arch There remain several other gaps in the varenicline evi- Intern Med 2006; 166: 1561–8.
dence base in addition to those addressed here. Perhaps 7. Oncken C., Gonzales D., Nides M., Rennard S., Watsky E., foremost is the question of whether varenicline is effective Billing C. B. et al. Efficacy and safety of the novel selectivenicotinic acetylcholine receptor partial agonist, varenicline, when prescribed by general practitioners without several for smoking cessation. Arch Intern Med 2006; 166: 1571–7.
sessions of expert group support or counselling, as given 8. Cahill K., Stead L., Lancaster T. Nicotine receptor partial here, and in the company-sponsored trials and here, agonists for smoking cessation. Cochrane Database of Sys- respectively. Without such support sessions there is the tematic Reviews 2007, Issue 1. Art. No.: CD006103. DOI: possibility that side effects might cause undue fear and a loss of confidence, leading to failure. A varenicline trial 9. Action on Smoking and Health (ASH). Varenicline. Guidance with routine brief general practitioner support is needed for Health Professionals on a New Prescription-Only Stop Smok- ing Medication. London: ASH; 2006. Available at: DECLARATION OF INTEREST
10. Freedman R. Exacerbation of schizophrenia by varenicline.
J. S. and G. S. have acted as advisers to the manufacturers Am J Psychiatry 2007; 164: 1269.
of nicotine replacement therapy and varenicline, for 11. Kohen I., Kremen N. Varenicline-induced manic episode in a patient with bipolar disorder. Am J Psychiatry 2007; 164:
which they have received remunerations and hospitality, and have conducted smoking cessation trials with 12. Silagy C., Lancaster T., Stead L., Mant D., Fowler G. Nicotine support from the manufacturers of nicotine replacement replacement therapy for smoking cessation. Cochrane therapy. G. S. has given lectures on tobacco dependence Database of Systematic Reviews 2004, Issue 3. Art. No.: sponsored by the manufacturers of nicotine replacement CD000146. DOI: 10.1002/14651858.CD000146.pub2.
(accessed 12 November 2007).
therapy and varenicline, for which she has received 13. National Institute for Health and Clinical Excellence (NICE).
remunerations and hospitality. R. S. has given talks on Technology Appraisal Guidance. 11 April 2002.
smoking cessation sponsored by the manufacturers of varenicline, for which he has received remuneration.
L. W., R. S. and A. M. attended the 2007 UK Smoking statistics-on-nhs-stop-smoking-services-in-england-april- Cessation conference at the invitation of the manufactur- 2006-to-december-2006-q3--quarterly-report 15. Russell M. A. H., Stapleton J. A., Feyerabend C., Wiseman S.
ers of varenicline. This evaluation was conducted without M., Gustavsson G., Sawe U. et al. Targetting heavy smokers the involvement of any commercial organization.
in general practice: randomised controlled trial of transder-
mal nicotine patches. BMJ 1993; 306: 1308–12.
References
statistics/Publications/PublicationsPolicyAndGuidance/ Humans/EPAR/champix/champix.htm (accessed 26 Sep- 17. Foulds J., Stapleton J. A., Hayward M., Russell M. A. H., Feyerabend C., Fleming T. et al. Transdermal nicotine 2007 The Authors. Journal compilation 2007 Society for the Study of Addiction patches with low-intensity support to aid smoking cessation 19. Williams D. A. Extra-binomial variation in logistic linear in outpatients in a general hospital: a placebo-controlled models. Appl Stat 1982; 31: 144–8.
trial. Arch Fam Med 1993; 2: 417–23.
20. West R., Hajek P., Stead L., Stapleton J. A. Outcome criteria 18. Hughes J. R., Stead L. F., Lancaster T. Antidepressants for in smoking cessation trials: proposal for a common stan- smoking cessation. Cochrane Database of Systematic Reviews dard. Addiction 2005; 100: 299–303.
2007, Issue 1. Art. No.: CD000031. DOI: 10.1002/ 21. Stapleton J. A. Cigarette smoking prevalence, cessation and relapse. Stat Methods Med Res 1998; 7: 187–203.
2007 The Authors. Journal compilation 2007 Society for the Study of Addiction

Source: http://jieyan.dxy.cn/uploads/soft/1_090406164059.pdf