The search identified 211 studies from a range of databases and other sources using comprehensive and sensitive search terms. After removing duplicates, 151 studies were assessed by reading their titles and abstracts, of which 132 studies were removed as they were not relevant to the review question. Finally, 19 full text articles were assessed, of which 10 studies were excluded. Reasons for study exclusion included combined intervention with behavioural therapy [33-35], measured cost-effectiveness of pharmacological smoking cessation therapies (36), pooled effect not provided [37-39], review not published in English [40], inclusion of non-randomized controlled trials in the review [41], and inclusion of study participants under the age of 15 [42]. The PRISMA flow diagram is depicted in Figure 1.
Characteristics of included reviews
Table 2 presents the detailed characteristics of the included systematic reviews. Of the nine included reviews, three assessed the effectiveness of nicotine replacement therapy [43-45], two assessed the effect of multiple pharmacotherapy [46, 47], one compared combination therapy (NRT+ varenicline vs varenicline alone) [48], one compared opioid antagonists to placebo or an alternative therapy [49], one evaluated the effectiveness of silver acetate products (gum, lozenge, spray) [50], and one compared the combined effect of nicotine replacement therapy of different formulations [51]. In total, 142 trials were included in 9 systematic reviews (mean per review: 14.2; range: 2-86). The included reviews consisted of a total of 63,568 study participants (mean per review: 7063.1). However, data from individual studies are likely to be represented more than once across the systematic reviews. Of the included reviews, two were Cochrane reviews. Six studies included in their reviews only studies that verified smoking cessation/abstinence using biochemical methods, while three included studies which used both self-reported and biochemical techniques (Table 3).
Methodological quality of included reviews
The reviews were assessed for methodological quality using the R-AMSTAR quality appraisal tool for systematic reviews. Table 4 presents the score of each item and the overall score of specific systematic reviews. The mean and standard deviation of each R-AMSTAR item have also been shown. A high mean value showed that most reviews attained a high quality level (maximum mean value possible is 4; minimum mean value possible is 1). The R-AMSTAR score for the nine included reviews ranged from 22 to 41 with a mean of 28.8. Based on R-AMSTAR scoring, one review scored ‘high’ for methodological quality [46], three scored ‘good’ for methodological quality [44, 48, 50], and five scored ‘fair’ for methodological quality [43, 45, 47, 49, 51]. All of the reviews conducted before the development of R-AMSTAR scored ‘fair’ for methodological quality. On the other hand, except for one review, all the reviews published after the inception of R-AMSTAR scored at least ‘good’ for methodological quality. Six reviews scored the maximum of 4 in the 6th item (“Were the characteristics of the included studies provided?”), and the 9th item (“Were the methods used to combine the findings of studies appropriate?”) and none of the reviews scored 4 in the 7th item (“Was the scientific quality of the included studies assessed and documented?”). The inter-rater agreement of individual items ranged from 0.57 to 1.00 and the overall Kappa score was 0.77 (95% CI: 0.65 - 0.89). Except for items 4 and 11, all items scored a kappa of >0.6 (Table 5).
The effectiveness of pharmacological interventions
Nicotine replacement therapy
In one review, which included 11 randomized controlled trials involving 1,808 study participants, researchers found that pharmacotherapy significantly increased the smoking cessation rate compared to the placebo group (RR = 1.88, 95% CI: 1.35, 2.57) at 6 weeks to 18-months follow-up. Likewise the pooled effect from a sub-analysis of three trials using only NRT indicated a significant positive effect on smoking cessation rate (RR = 7.74, 95% CI: 3.00, 19.94; 3 studies, 635 participants) [46]. In the quality appraisal, this paper was scored as having ‘excellent’ methodological quality. Another study assessing the pooled effect from 12 randomized controlled trials (‘fair’ methodological quality review) supported the favourable effect of NRT on sustaining smoking cessation beyond 12 months compared to a placebo (OR = 1.99, 95% CI: 1.50, 2.64) [43].
In a review of seven studies (‘fair’ methodological quality review), Moore et al. found that NRT increased smoking cessation for at least six months compared with the placebo (RR = 2.06, 95% CI: 1.34, 3.15; 5 studies) [45]. The pooled effect from a review that included 70 trials (n=28,343) found that the odds of smoking cessation at one year were higher among participants using NRT compared to the control group (OR = 1.71, 95%, CI: 1.55, 1.88) [47]. In this review, the finding was consistent across all NRT formulations (gum, patch). In addition, the pooled effect of 59 trials (n=25,294) demonstrated that NRT provided support for the efficacy of smoking cessation in the short-term follow-up (3 months) compared to the control group (OR = 1.98, 95% CI: 1.77, 2.21) [47]. In the quality appraisal, this review scored a ‘fair’ methodological quality. Conversely, a review by Lindson et al (‘good’ methodological quality review) that included eight studies and 2,813 participants found no significant effects of NRT over placebo for the treatment of smoking cessation in the short-term follow-up (4 to 12 weeks) (RR = 1.05, 95% CI: 0.92, 1.19) and long-term follow-up (6 to 12 months) (RR = 1.16, 95% CI: 0.97, 1.38) [44].
Non-nicotine pharmacotherapy
The pooled effect of a study including eight randomized control trials with 1,213 study participants identified that opioid antagonist therapy had no effect on smoking cessation rate based on the 6-month reported abstinence rate (RR = 0.97; 95% CI: 0.76, 1.24). Five studies that assessed the effect of naltrexone (long-acting form of opioid antagonist) compared to placebo also showed no significant effect on smoking abstinence rate (RR = 1.00; 95% CI: 0.66, 1.51) [49]. In the quality appraisal, this review was scored as having ‘fair’ methodological quality.
The pooled effect of 12 trials including 5,228 participants showed bupropion was more effective for smoking cessation compared to the control group at the one-year follow-up (OR = 1.56, 95% CI: 1.10, 2.21). Moreover, bupropion was more effective than placebo at the 3-month follow-up (OR = 2.13, 95% CI: 1.72, 2.64; 11 trials) [47]. The pooled effect of 4 studies (n=2,528) found that varenicline was effective for smoking cessation compared to placebo both at long-term follow-up (1 year) (OR = 2.96, 95% CI: 2.12, 4.12) and short-term follow-up evaluation (3 months) (OR = 3.75, 95% CI: 2.65, 5.30). Similarly, varenicline was more effective than bupropion at one year follow-up (OR = 1.58, 95% CI: 1.22, 2.05; 3 trials) and three-month follow-up (OR = 1.61, 95% CI: 1.16, 2.21; 3 trials) [47]. Silver nitrate was not effective for smoking cessation compared to placebo at a minimum of 6-month follow-up (RR = 1.04, 95% CI: 0.69, 1.57; 2 trials) [50]. This review was ranked as having a ‘good’ methodological quality.
Combination therapy
Chang et al (‘good’ methodological quality review) found that study participants on a combined regimen (NRT and non-NRT) were more likely to abstain from smoking compared with those in a non-NRT (varenicline) only treatment group, both during the short-term (measured at 4-12 months before treatment completion; OR = 1.50, 95% CI: 1.14, 1.97; 3 trials) and long-term (measured at the end of 2-24 months after treatment completion; OR = 1.62, 95% CI: 1.18, 2.23; 2 trials) [48]. Combining naltrexone and NRT did not favour smoking cessation compared to the placebo group based on the 6-month reported abstinence rate (RR = 0.95, 95% CI: 0.70, 1.30; 4 studies) [49].
In a ‘fair’ methodological quality review that included 2,204 study participants from 5 trials, a combination therapy of nicotine replacement patches with other nicotine formulation drugs (nicotine gum or nicotine inhaler or nicotine nasal spray) was found to be more effective than monotherapy at 3 months (RR = 1.42, 95% CI: 1.21, 1.67; 4 trials), at 6 months (RR = 1.54, 95% CI: 1.19, 2.00; 4 trials) and at the 12-month follow-up (RR = 1.58, 95% CI: 1.25, 1.99; 4 trials) [51].