Smoking is the major avoidable cause of preventable morbidity and mortality in the UK and internationally.1,2 Smoking is also the principal cause of health inequalities and is responsible for most of the difference in healthy life expectancy between the richest and poorest in our society3 and those with and without mental health problems.4 Studies have estimated that smoking-related illnesses cost the NHS approximately £5B per year.5 In 1974, 45.6% of the population smoked; by 2016, this had fallen to 16.1% (Figure 1).
FIGURE 1 Proportion of the adult population in Great Britain who smoke, by age. Estimates from the Annual Population Survey and published by the Office for National Statistics.6 (Figure not reproduced here; see the original article.)
The main form of smoking cessation treatment is nicotine replacement therapy (NRT), such as nicotine patches and gums. Pharmaceutical treatments for nicotine addiction include bupropion and varenicline; however, bupropion is rarely used.7 Varenicline, a nicotinic receptor partial agonist, has been available in the UK from 2006. Varenicline has been shown to be the most effective smoking cessation medicine for short-term abstinence in randomised controlled trials (RCTs).7 However, there is relatively little evidence for its long-term effectiveness and impact on clinical outcomes that are relevant to health-care providers.
In addition, concerns have been raised that smokers taking varenicline have a higher risk of adverse events, including suicide, self-harm and cardiovascular events, than with other smoking cessation interventions.8 This led the Food and Drug Administration (FDA) to issue black box warnings about the possible adverse effects of varenicline between July 2009 and December 2016.9 There have been large randomised studies comparing the effectiveness and safety of varenicline with NRT, such as (EAGLES Evaluating Adverse Events in a Global Smoking Cessation Study).10 However, much of the evidence about the potential adverse effects of varenicline comes from observational studies that are prone to residual confounding. Furthermore, the experimental evidence about the effects of prescribing varenicline on health outcomes has limited follow-up.
The rationale for this study
We addressed this gap in the literature by investigating the adverse and beneficial effects of prescribing varenicline and NRTs using three statistical approaches to overcome confounding: multivariable adjusted, propensity score matched and instrumental variable regression. We investigated the effects of varenicline on all-cause mortality and smoking abstinence, and on all-cause primary and secondary care use because smoking increases morbidity and imposes major costs on health-care services.11 We examined the differences in smoking cessation medication effectiveness by socioeconomic position (SEP) because a recent systematic review reported that UK stop-smoking services might be helping to reduce inequalities in smoking prevalence by preferentially targeting smokers of lower SEP.12