Tools for Practice

#27 Pharmacotherapy for Smoking: Which work and what to consider (Part II)?

In patients ready to make a smoking cessation attempt, how effective are registered first-line medications and what are the potential concerns?

Bupropion, nortriptyline and varenicline are all effective in smoking cessation, with varenicline more-so. Adverse events vary and may in part relate to quitting smoking, and require monitoring.

CFPCLearn Logo

Reading Tools for Practice Article can earn you MainPro+ Credits

Join Now

Already a CFPCLearn Member? Log in

See Part 1 for nicotine replacement therapy (NRT).
  • Antidepressants: Cochrane review of bupropion 65 and nortriptyline 10 randomized controlled trials (RCTs).1
    • Risk Ratio (RR) for cessation over placebo,
      • Bupropion (at 6-12 months): 1.62 (1.49-1.76)
      • Nortriptyline (at 6 months): 2.03 (1.48-2.78)
    • Serious adverse events:
      • Bupropion: Seizure (about 1/1000) and suicidal thoughts/behavior (association unclear) are rare.
    • SSRI (6 RCTs) and venlafaxine (1 RCTs): not effective.
  • Varenicline:
    • Cochrane review2 of 39 RCTs: RR for cessation at 6-12 months over placebo= 2.24 (2.06-2.43)
      • RR over bupropion= 1.39 (1.25-1.54)
      • RR over NRT = 1.25 (1.14-1.37)
    • Serious Adverse Events: RR 1.25 (1.05-1.49)
    • Neuropsychiatric safety: Early studies2,3 suggested possible increase in depression, irritability, and suicidal thoughts/attempts, however:
      • Systematic review4 of 39 RCTs: No increase versus placebo
      • RCT5 0f 8,144 (50% with psychiatric disorder): No difference versus placebo, NRT or bupropion
      • Cohort6 of ~120,000 patients: No difference between varenicline, bupropion or NRT.
    • Cardiovascular events: No increase (see updated Tools for Practice #71).
  • Assuming 10% placebo cessation rates (mean across studies), number needed to treat: Varenicline 8, Nortriptyline 10 and Buproprion 10.
  • Risk of bias in varenicline evidence:
    • Superiority of varenicline > buproprion is at risk of funding bias
    • Previously noted publication bias: In 2011, 75% of varenicline trials were unpublished.7
  • Health Canada recommends “thorough consideration” of NRT before varenicline or bupropion.8
  • Dosing:
    • Lower doses are effective:
      • Bupropion 150 mg is equivalent to 300 mg1,9
      • Varenicline 0.5 mg BID may be slightly less effective than 1 mg BID (with fewer adverse events)2,10
    • Nortriptyline: Can start at 25 mg qhs and increase by 25 mg every 3-4 days, if needed, to a maximum of 75-100 mg. Encouraged quit date 10 days in (or so) and continue for 10-12 weeks.

Latest Tools for Practice

#341 Forget about it? Statins and the risk of dementia (Free)

Do statins negatively affect cognition, memory, or dementia?
Read 0.25 credits available

#340 Crying babies: Can proton pump inhibitors help? (Free)

In infants (≤1year) with crying/irritability attributed to feeds, do proton pump inhibitors (PPIs) improve symptoms over placebo without additional harms?
Read 0.25 credits available

#339 Is acetaminophen under pressure?

Does regular use of acetaminophen increase blood pressure?
Read 0.25 credits available

This content is certified for MainPro+ Credits, log in to access


  • Charl Els MBChB FCPsych (Psychiatry)
  • G. Michael Allan MD CCFP

1. Hughes JR, Stead LF, Hartmann-Boyce J, Cahill K, Lancaster T. Cochrane Database System Rev. 2014;1:CD000031.

2. Cahill K, Lindson-Hawley N, Thomas KH, Fanshawe TR, Lancaster T. Cochrane Database System Rev. 2016;5:CD006103.

3. Gunnell D, Irvine D, Wise L, et al. BMJ. 2009; 339:b3805.

4. Thomas KH, Martin RM, Knipe DW, Higgins JPT, Gunnell D. BMJ. 2015;350:h1109.

5. Anthenelli RM, Benowitz NL, West R, et al. Lancet. 2016;387:2507-20.

6. Thomas KH, Martin RM, Davies NM, et al. BMJ. 2013; 347:f5704.

7. Brophy JM. Ann Intern Med. 2011 Oct 18; 155:JC4-5.

8. Health Canada. Accessed Nov. 7, 2013.

9. Hurt RD, Sachs DP, Glover ED, et al. New Engl J Med. 1997; 337:1195-202.

10. Fouz-Roson N, Montemayor-Rubio T, Almadana-Pacheco, et al. Addiction. 2017;112:1610-9.

Authors do not have any conflicts of interest to declare.