Smoking cessation

1 July 2010

The effectiveness of a medication to help quit smoking is examined in this edition’s clinically-appraised topic. 


Smoking cessation with varenicline increases the initial effectiveness of quitting compared to the application of nicotine patch (NRT) but the difference may not be maintained at one year.


You are a practice nurse in the primary health organisation. One day a 45-year-old man visits you and asks about quitting smoking. He is interested in nicotine patch but you have heard that varenicline is better.


Among smokers does varenicline increase smoking cessation compared to transdermal nicotine patches?


Pubmed – Clinical queries (therapy/narrow specific search): smoking cessation.


Aubin HJ, Bobak A, Britton JR, et al. (2008). Varenicline versus transdermal nicotine patch for smoking cessation: results from a randomised open-label trial. Thorax 2008; 63:717-24.


Two-arm, parallel, open-label randomised multicentre, phase III trial conducted in 24 centres in Europe and USA. Participants were recruited through smoking cessation clinics or via local advertising. Inclusion criteria were age 18–75 years, weight greater than 45.5 kg, body mass index (BMI) 15–38, smoking 15 or more cigarettes per day and less than three months abstinence from smoking in the previous 12 months. Exclusion criteria were: breastfeeding mothers, pregnant women or women at risk of becoming pregnant, history of cancer, any serious or unstable disease within previous six months, diagnosis or treatment for depression or other psychological disorder or drug or alcohol dependence within previous 12 months, allergic reactions to adhesive tapes or drugs, skin disorders, systolic blood pressure >150 mmHg or diastolic blood pressure >95 mmHg, significant renal or hepatic impairment, liver dysfunction or abnormal laboratory tests, taking medication that might interfere with study drugs, previously participated in varenicline study or used any form of NRT in previous six months. Overall, 957 smokers were assessed for eligibility and 757 randomised. Smokers received cessation booklet and brief counselling session at study start, a telephone call on target quit date (TQD) and weekly clinic visits during treatment and a mix of telephone and clinic visits during the non-treatment follow up phase.

Varenicline (N=376):

Full treatment period of varenicline was for 12 weeks, starting one week before TQD and titrated to 1 mg twice daily by TQD. Non-treatment follow-up was for 40 weeks.

NRT (N=370):

Application of transdermal nicotine patches each morning starting on TQD for 10 weeks. Doses were 21mg per day for first six weeks, 14mg per day for two weeks and 7mg per day for two weeks. Non-treatment follow-up was for 41 weeks.


Primary outcome measure was continuous self-reported abstinence (CAR) at four weeks confirmed by exhaled CO ≤10ppm. Secondary outcomes included CAR (CO confirmed) from last week of treatment to 24 and 52 weeks, seven-day point prevalence of abstinence (CO confirmed) at 24 and 52 weeks, measurement of craving, withdrawal, smoking satisfaction, and adverse and serious adverse events.


Individual smokers randomised 1:1 by computer-generated sequence without blocking. Allocation concealment was not mentioned. It was an open-label study and no participant blinding was possible. However, CO levels confirmed self-reported outcomes. Loss to follow-up was high (~7%), but only ~64% completed treatment and 52 week assessments (similar in both groups). Participants who withdrew were assumed to be smokers. The main analysis (primary analysis population – PAP) included participants who were randomised and took at least one dose. A sensitivity analysis included all randomised participants. Groups were similar at baseline. Missing carbon monoxide (CO) data or use of NRT during follow-up period was considered as non-smoker if other criteria were met. Overall, the methodological quality of the study was moderate.


There was significantly greater abstinence in the varenicline group for last four weeks of treatment (table). The seven-day point prevalence of abstinence score at week 12 was significant but not at 24 or 52 weeks. There were significantly lower rates for most craving and withdrawal symptoms and smoking satisfaction at the end of treatment with varenicline. There were more adverse events with varenicline but higher rates of serious adverse events (SAE) in NRT, although the difference was not significant.


Varenicline products are available in NZ but are not subsidised so must be paid for by the patient.

Generalisation may not be appropriate for the populations who met the exclusion criteria in this study.

Open-label study, which is similar to real-world clinical practice (smokers wouldn’t receive cessation treatment if they didn’t like it). ✚

Reviewers: Dong Eun Lee, 3rd year nursing student, University of Auckland; Dr Andrew Jull, associate professor, University of Auckland and nurse advisor – quality, Auckland District Health Board.