Materials were tailored to the individual’s characteristics; several trials used computerized expert systems with tailoring according to baseline data. In all trials, materials were sent to participants without personal contact. Control conditions varied across studies and included assessment only, thank you letters only, and no intervention or information. No co-interventions were provided to either arm in all trials. Majority of studies reported continuous/sustained abstinence rates. No studies used biochemical validation.
Compared to those receiving no materials/no intervention, 20 more people per 1,000 receiving individually-tailored print-based self-help materials (no face-to-face contact) were abstinent at longest follow-up (6 + months) (95% CI: 11 to 31 more; n = 10, 14,359 participants; I2 = 0%) (rating down twice for risk of bias) (Appendix K Table 40; Fig. 2b (Appendix Q)) [74]. The certainty of the evidence was rated as low.
Intensive telephone counselling compared to minimal telephone counselling
Abstinence/cessation in general/mixed population of smokers
Three trials compared intensive telephone counselling, defined as three to five calls, to minimal telephone counselling consisting of a single call. Two trials provided a behavioural co-intervention to both arms, and one of these trials also provided pharmacotherapy to all study participants including those receiving control. Most trials reported point prevalence rates. Only one trial biochemically validated smoking abstinence.
At six months or more follow-up, 64 more people per 1,000 receiving intensive telephone counselling were abstinent compared to those receiving minimal telephone counselling (95% CI: 28 to 104 more; n = 3, 2,602 participants; I2 = 0%) (rating down twice for risk of bias and half for each indirectness and imprecision) (Appendix K Table 42; Fig. 2b (Appendix Q)) [75]. The certainty of the evidence was rated as very low.
Brief motivational telephone counselling compared to usual care telephone call
Abstinence/cessation in general/mixed population of smokers
One trial compared brief motivational telephone counselling (three 15-minute calls) to a usual care telephone call (one five-minute call). No co-interventions were provided to either arm.
At 12-months follow-up, 60 more people per 1,000 receiving brief motivational telephone counselling were point prevalence abstinent compared to those receiving usual care telephone call (95% CI: 4 to 190 more; n = 1, 374 participants) (rating down twice for risk of bias and twice for imprecision) (Appendix K Table 43; Fig. 2b (Appendix Q)) [75]. The certainty of the evidence was rated as very low.
Telephone counselling for smoking reduction compared to usual care telephone call
Abstinence/cessation in general/mixed population of smokers
Compared to a usual care telephone call (one five-minute call), 49 more people per 1,000 receiving telephone counselling (three 15-minute calls) for smoking reduction were point prevalence abstinent at 12-month follow-up (95% CI: 1 fewer to 167 more; n = 1, 375 participants) (rating down twice for risk of bias and twice for imprecision) (Appendix K Table 44; Fig. 2b (Appendix Q)) [75]. The certainty of the evidence was rated as very low.
Interactive voice response (IVR) systems compared to no intervention
Abstinence/cessation in general/mixed population of smokers
One trial compared IVR to no intervention. Prior to randomization, all participants received varenicline for 12 weeks and IVR. After randomization, the control group no longer received IVR. Review authors reported little to no difference between groups at 24-months follow-up (21.7% of intervention group versus 42.9% of control group; P = 0.13) (rating down once for risk of bias) (Appendix K Table 45; Fig. 2b (Appendix Q)) [76]. We were unable to rate the certainty of the evidence due to sample size analyzed not being reported for imprecision domain.
Physician advice (minimal or intensive interventions) compared to no advice (or usual care)
Abstinence/cessation in general/mixed population of smokers
Physician advice involved a 'stop smoking' message. Interventions were defined as minimal if delivered during a single session of less than 20-minute duration plus up to one follow-up visit with or without a leaflet. Intensive interventions involved a longer initial consultation, use of additional materials other than a leaflet, or more than one follow-up visit. Pharmacotherapy co-intervention was provided to intervention participants in some of the studies. Additional behavioural and/or pharmacotherapy co-interventions were provided to control participants in some studies. Trials varied with respect to how abstinence was measured and only a few biochemically confirmed smoking abstinence.
At longest follow-up, 36 more people per 1,000 receiving physician advice were abstinent in comparison to those receiving no advice (or usual care) (95% CI: 28 to 46 more; n = 26, 22,239 participants; I2 = 40%) (rating down twice for risk of bias) (Appendix K Table 46; Fig. 2b (Appendix Q)). The certainty of the evidence was rated as low. Subgroup differences by advice intensity (minimal or intensive) was not detected (test for subgroup differences: p = 0.31, I2 = 3%) (Appendix M). Review authors state that indirect comparisons across subgroups defined by number of advice sessions was suggestive of greater effects with multiple visits compared to a single visit. Authors indicate no important differences between subgroups according to use of aids [77].
Physician advice with follow-up compared to minimal intervention/advice with single visit
Abstinence/cessation in general/mixed population of smokers
Five studies compared physician advice with follow-up to a minimal intervention/advice with a single visit. Most studies reported continuous/sustained abstinence. Abstinence was biochemically confirmed in 60% of studies. One study provided a pharmacotherapy co-intervention to both arms.
Compared to control, 47 more people per 1,000 receiving physician advice with follow-up were abstinent at 6 + months (95% CI: 7 to 103 more; n = 5, 1,254 participants; I2 = 0%) (rating down twice for risk of bias and once for imprecision) (Appendix K Table 47; Fig. 2b (Appendix Q)) [77]. The certainty of the evidence was rated as very low.
Intensive advice compared to minimal advice
Abstinence/cessation in general/mixed population of smokers
Intensive interventions involved a longer initial consultation, use of additional materials other than a leaflet, or more than one follow-up visit. Trials varied with respect to how abstinence was measured; nearly half reported point prevalence rates and the reminder continuous/sustained rates. Biochemical validation was used in 60% of studies. One study provided a pharmacotherapy co-intervention to both arms.
Compared to minimal advice, 28 more people per 1,000 receiving intensive advice were abstinent at longest follow-up (6 + months) (95% CI: 15 to 42 more; n = 15, 9,775 participants; I2 = 32%) (rating down twice for risk of bias and once for indirectness) (Appendix K Table 48; Fig. 2b (Appendix Q)). The certainty of the evidence was rated as very low. The effect estimate was larger for the subgroup of participants at high risk of smoking-related diseases as compared to unselected smokers, but confidence intervals overlapped (test for subgroup differences: p = 0.02, I2 = 82%) (Appendix M) [77].
Group behaviour therapy compared to no intervention
Abstinence/cessation in general/mixed population of smokers
Group therapy was delivered over at least two sessions. Measures of abstinence were unclear or not reported for two-thirds of trials. Some studies provided both groups with co-interventions including pharmacotherapy; however, provision of co-interventions was unclear for most studies.
Compared to those receiving no intervention, 108 more people per 1,000 given group behaviour therapy were abstinent at six + months follow-up (95% CI: 54 to 186 more; n = 9, 1,098 participants; I2 = 55%) (rating down twice for risk of bias, once for inconsistency, once for imprecision) (Appendix K Table 51; Fig. 2b (Appendix Q)) [78]. The certainty of the evidence was rated as very low.
Interactive and tailored internet interventions compared to non-active controls
Abstinence/cessation in general/mixed population of smokers
Interactive interventions “involved a two-way flow of information between the internet and the participant” and programs were tailored to a participant's characteristics [79]. Across studies, control conditions consisted of printed self-help guides or usual care. One study provided NRT as an adjunct to the internet intervention. Another study offered free nicotine patches and bupropion to all participants, including those receiving control, and their partners who wanted to quit. Trials varied with respect to how abstinence was measured and only 25% used biochemical validation.
At longest follow-up (six to 12 months), 19 more people per 1,000 receiving interactive and tailored internet interventions were abstinent compared to non-active controls (95% CI: 1 to 39 more; n = 8, 6,786 participants; I2 = 58%) (rating down twice for risk of bias, once for indirectness, and inconsistency). (Appendix K Table 52; Fig. 2b (Appendix Q)) [79]. The certainty of the evidence was rated as very low.
Mobile phone short message service compared to control
Abstinence/cessation in smokers motivated/wishing to quit
One trial examined the effect of personalized text messaging with smoking cessation advice, support, and distraction sent between a health care provide or buddy (i.e., lay health worker or peer support) and the participant. The control group received biweekly text messages thanking them for their involvement and reminder of a free month of text messaging should they complete the study. The trial used various measures of abstinence.
Compared to control, 17 more people per 1,000 receiving the active text messaging intervention were point prevalent abstinent at 6-month follow-up. (95% CI: 21 fewer to 62 more; n = 1, 1705 participants) (rating down once for risk of bias and twice for imprecision). The certainty of the evidence was rated as very low. However, when point prevalence using last outcome carried forward, 76 more people per 1,000 receiving the active intervention were point prevalent abstinent in comparison to controls (95% CI: 30 to 131 more; n = 1, 1,705 participants) (rating down once for risk of bias and once for imprecision). The certainty of the evidence was rated as low.
When allowing for three or fewer lapses of two or fewer cigarettes per lapse, 29 more people per 1,000 receiving the active text messaging intervention were continuously abstinent compared to controls (95% CI: 5 to 65 more; n = 1, 1,705 participants) (rating down once for risk of bias and once for imprecision). The certainty of the evidence was rated as low. These effects were not sustained when allowing for no lapses (Appendix K Table 63; Fig. 2b (Appendix Q)) [64].
Adverse events in smokers motivated/wishing to quit
Compared to control, six fewer people per 1,000 receiving the active text messaging intervention had been in a car crash during the six-months follow-up period (rating down twice for imprecision). The certainty of the evidence was rated as low. However, it is possible that there is little to no difference between groups or that more participants receiving the intervention were in a car crash (Appendix K Table 63; Fig. 2b (Appendix Q)) [64].
Compared to control, five more people per 1,000 receiving the active text messaging intervention experienced pain in their thumb/finger joints during the six-months follow-up (rating down twice for imprecision). The certainty of the evidence was rated as low (95% CI: 15 fewer to 33 more; n = 1, 1705 participants) (Appendix K Table 63; Fig. 2b (Appendix Q)) [64].
Mobile phone-based interventions compared to usual care
Abstinence/cessation in smokers motivated/wishing to quit
Nearly all studies had text messaging (SMS) as the main component of the intervention; however, one trial differed in that participants received mobile phone-based counselling (cognitive behavioural therapy (CBT) and motivational). In-person visits or assessments were provided in addition to SMS in five studies. Control conditions varied across studies and included no intervention, text messages, written/internet untailored materials, untailored messages, and standard cessation advice and treatment. Trials varied with respect to how abstinence was measured with slightly more reporting point prevalence rates. Abstinence was biochemically confirmed in half of the studies.
At longest follow-up (six + months), 37 more people per 1,000 receiving mobile phone-based interventions were abstinent in comparison to those receiving usual care (95% CI: 26 to 50 more; n = 12, 11,885 participants; I2 = 59%) (rating down once for risk of bias and once for inconsistency) (Appendix K Table 69; Fig. 2b (Appendix Q)). The certainty of the evidence was rated as low. Sensitivity analyses according to abstinence definition, biochemical validation, and intervention characteristics showed no appreciable change in the relative effect estimate (Appendix M) [80].
3. Other therapy interventions
Hypnotherapy compared to placebo drug alone
Abstinence/cessation in smokers motivated/wishing to quit
Intervention details, including the number and duration of sessions, was not reported in the only included trial. At 12-months follow-up, 18 fewer people per 1,000 receiving hypnotherapy were point prevalent abstinent compared to those receiving the placebo drug (95% CI: 77 fewer to 166 more; n = 1, 114 participants) (rating down twice for risk of bias and twice for imprecision) (Appendix K Table 1; Fig. 2c (Appendix Q)) [52]. The certainty of the evidence was rated as very low.
Acupuncture compared to sham acupuncture
Abstinence/cessation in smokers motivated/wishing to quit
All trials selected acupuncture points (i.e., anatomic sites) for smoking cessation. Two studies used facial acupuncture, five used auricular acupuncture alone with/without continuous stimulation (i.e., needle or pressure device), and four used combined body and auricular acupuncture with/without continuous stimulation (i.e., indwelling needle or seed). Two trials, representing 4% of the evidence, used potentially active acupuncture points for the sham arm. Behavioural co-interventions were provided to both study arms in some trials and pharmacotherapy to all participants in one study. Trials varied with respect to how abstinence was measured. Few studies used biochemical validation.
At longest follow-up (6–12 months), 11 more people per 1,000 who received acupuncture were abstinent compared to those receiving sham, (95% CI: 15 fewer to 43 more; n = 11, 1892 participants, I2 = 23%) (rating down twice for risk of bias and twice for imprecision) (Appendix K Table 64; Fig. 2c (Appendix Q)). The certainty of the evidence was rated as low. Sensitivity analysis excluding studies which used potentially active acupuncture points for the sham arm yielded similar results [54].
Acupuncture compared to waiting list/no intervention
Abstinence/cessation in smokers motivated/wishing to quit
All trials selected acupuncture points (i.e., anatomic sites) for smoking cessation. One study used facial acupuncture and two used auricular acupuncture alone with or without continuous stimulation (i.e., needle or pressure device). One study provided a behavioural co-intervention to both study arms, with no co-interventions in the remaining studies. No studies biochemically verified abstinence.
At six to 12 months follow-up, 60 more people per 1,000 reported being smoking abstinent compared to those in the waitlist/no intervention group (95% CI: 2 fewer to 174 more; n = 2, 393 participants, I2 = 57%). We rated down twice for risk of bias, once for inconsistency, and twice for imprecision (Appendix K Table 65; Fig. 2c (Appendix Q)) [54]. The certainty of the evidence was rated as very low.
Continuous auricular stimulation compared to sham stimulation
Abstinence/cessation in smokers motivated/wishing to quit
Four studies used indwelling needles and the remainder used continuous acupressure. One trial, representing 5% of the evidence, used potentially active acupuncture points for the sham arm. Two studies provided a behavioural co-intervention to both groups. Half of the studies reported use of biochemical confirmation for smoking abstinence. The measure of smoking abstinence was unclear or not reported for half of the studies.
At six to 12-months follow-up, 26 more people per 1,000 given continuous auricular stimulation compared to sham stimulation stopped smoking (95% CI: 12 fewer to 98 more; n = 6, 570 participants, I2 = 22%) (rating down twice for risk of bias to also reflect concerns with publication bias and twice for imprecision) (Appendix K Table 66; Fig. 2c (Appendix Q)). The certainty of the evidence was rated as very low. Subgroup differences by type of stimulation (indwelling needles or continuous acupressure) was not detected (test for subgroup differences: p = 0.06, I2 = 72%) (Appendix M) [54].
Laser therapy compared to sham laser
Abstinence/cessation in smokers motivated/wishing to quit
Only one of the two studies reported the dose of laser used which was 50mW for 14 minutes. One study provided a behavioural co-intervention to both arms. Abstinence measures were unclear or not reported for both studies.
Studies were heterogeneous and could not be quantitatively synthesized (I2 = 97%). Review authors report that heterogeneity was possibly attributable to populations recruited and dose of laser administered. Results from one study were null-inclusive (RD 3 more per 1,000, 95% CI: 45 fewer to 94 more). The other study reported results favouring the intervention (RD 515 more per 1,000, 95% CI: 192 to 1,000 more) (rating down once for risk of bias, twice for inconsistency and twice for imprecision) (Appendix K Table 67; Fig. 2c (Appendix Q)) [54]. However, the certainty of the evidence was rated as very low.
Electrostimulation compared to sham electrostimulation
Abstinence/cessation in smokers motivated/wishing to quit
Electrostimulation was administered through surface electrodes over the mastoid bone in one study and to the ear in the other. No co-interventions were provided to either arm in both studies. Abstinence measures were unclear or not reported for both studies. Biochemical validation was used in one study.
Compared to sham electrostimulation, 34 fewer people per 1,000 receiving electrostimulation were abstinent at longest follow-up (6 to 12 months), (95% CI: 102 fewer to 60 more; n = 2, 405 participants, I2 = 46%) (rating down twice for risk of bias and twice for imprecision) (Appendix K Table 68; Fig. 2c (Appendix Q)) [54]. The certainty of the evidence was rated as very low.
Acupressure versus Sham
No studies were found [54].
Laser therapy versus Wait-list/No intervention
No studies were found [54].
Electrostimulation versus Wait-list/No intervention
No studies were found [54].
St John’s wort versus Placebo
Abstinence/cessation in smokers motivated/wishing to quit
One trial examined 900 mg/day of St John's wort for 14 weeks and the other examined both 900 and 1800 mg/day for 12 weeks; treatment arms in the latter study were combined in the analysis. Behavioural co-interventions were provided to all groups in both studies. The studies reported biochemically confirmed prolonged abstinence.
Compared to placebo, 10 fewer participants per 1,000 on St. John’s wort were abstinent at 6-month follow-up; (95% CI: 40 fewer to 83 more; n = 2, 261 participants, I2 = 29%) (rating down twice for imprecision) (Appendix K Table 27; Fig. 2c (Appendix Q)) [66]. The certainty of the evidence was rated as low.
S-Adenosyl-L-Methionine (SAMe) versus Placebo
Abstinence/cessation in smokers motivated/wishing to quit
A single study reported on 800 and 1600 mg/day SAMe; doses were combined in analysis by review authors. All study groups received a behavioural co-intervention.
Compared to placebo, 38 fewer participants per 1,000 receiving SAMe were point prevalence abstinent at six-months follow-up, (95% CI: 95 fewer to 134 more; n = 1, 120 participants) (rating down once for risk of bias and twice for imprecision) (Appendix K Table 28; Fig. 2c (Appendix Q)) [66]. The certainty of the evidence was rated as very low.
Exercise interventions versus No intervention
No studies were found [59].
4. Combination interventions
Behavioural support (advice) plus NRT plus phone calls versus No intervention
Abstinence/cessation in smokers not motivated/wishing to quit
In one trial, participants received an initial advice intervention aimed at encouraging reduction. Participants were also advised to quit; those who agreed received the cessation intervention. Intervention participants were offered a choice of NRT gum or patch. Control participants received assessment calls only. Co-interventions were not provided to either group. Smoking cessation was not biochemically verified.
Review authors report that point prevalence abstinence rates were significantly higher in the intervention group as compared to control at 6six-months follow-up (rating down twice for risk of bias and once or twice for imprecision) (Appendix K Table 34; Fig. 2d (Appendix Q)) [55]. The certainty of the evidence was rated as very low.
Reduction in number of cigarettes/day in smokers not motivated/wishing to quit
The same trial also reported on smoking reduction. Review authors indicate that the reduction rate was significantly higher in the intervention group compared to those in the control arm (rating down twice for risk of bias and once or twice for imprecision) (Appendix K Table 34; Fig. 2d (Appendix Q)) [55]. The certainty of the evidence was rated as very low.
Telephone counselling plus self-help materials versus Usual care
Abstinence/cessation in smokers not motivated/wishing to quit
Participants were instructed to reduce smoking by 50% or more; cessation was subsequently encouraged. Self-help materials were individually tailored newsletter and a targeted newsletter. Usual care consisted of usual care plus generic health mailings. Co-interventions were not provided to either group. Smoking abstinence was biochemically verified.
At 12-month follow-up, 22 more people per 1,000 receiving telephone counselling plus self-help were point prevalence abstinent as compared to those receiving usual care; (95% CI: 18 fewer to 124 more; n = 1, 320 participants) (rating down twice for imprecision) (Appendix K Table 33; Fig. 2d (Appendix Q)) [55]. The certainty of the evidence was rated as low.
Reduction in cigarettes/day of > 50% of baseline or cessation
In the same trial, 63 more people per 1,000 receiving telephone counselling plus self-help were abstinent or reduced the number of cigarettes smoked by more than 50% at 12 months compared to usual care; (95% CI: 22 fewer to 195 more; n = 1, 320 participants) (rating down once for risk of bias and twice for imprecision) (Appendix K Table 33; Fig. 2d (Appendix Q)) [55]. The certainty of the evidence was rated as very low.
Reduction in carbon monoxide > 50%
At 12 months, one fewer person per 1,000 receiving telephone counselling plus self-help reduced their carbon monoxide levels by more than 50% as compared to those receiving usual care; (95% CI: 59 fewer to 100 more; n = 1, 320 participants) (rating down twice for imprecision) (Appendix K Table 33; Fig. 2d (Appendix Q)) [55]. The certainty of the evidence was rated as low.
Reduction in number of cigarettes per day from baseline
Review authors state that the number of cigarettes per day decreased from baseline in both groups (mean change from baseline (SD): Intervention: 21.2 (9.4), Usual care: 20.1 (8.9). There was no difference between groups at 12-months follow-up (mean (SD): Intervention: 15.8 (10. 3), Usual care: 15.3 (9.2) (rating down once for risk of bias). (Appendix K Table 33; Fig. 2d (Appendix Q)) [55]. Unable to assess the certainty of the evidence and imprecision domain based on information reported.
Reduction in carbon monoxide from baseline
Carbon monoxide levels decreased from baseline in both groups. Review authors state that there was no significant between-group difference in the change from baseline but it is unclear whether this is based on confidence intervals or p-values; we interpret this to mean that the confidence interval around the best estimate of effect includes the possibility of little to no difference between groups and greater reduction in one group over another (Appendix K Table 33; Fig. 2d (Appendix Q)) [55]. Unable to assess the certainty of the evidence and imprecision domain based on information reported.
Hotline and self-help materials versus Minimal intervention
Abstinence/cessation in general/mixed population of smokers
Two trials examined the effects of telephone hotline and self-help materials. One study promoted a 24-hour hotline, daytime access to counsellors, and use of the American Lung Association self-help manual. The second trial consisted of a Quitline proactive contract, quit kits (national Quitline printed resources), individual counselling with a practice nurse (face-to-face) and three proactive telephone calls from counsellors. Control participants received a minimal intervention; this was usual care delivered by a primary care provider (i.e., advice, referral to Quitline, or both) in one study and a self-help manual in the second (American Lung Association). One study offered pharmacotherapy co-intervention (NRT patch) to both arms. Both trials reported continuous/sustained abstinence rates. Only one used biochemical validation.
At longest follow-up (12 to 18 months), 21 more people per 1,000 receiving hotline and self-help materials were abstinent compared to those receiving minimal intervention (95% CI: 5 more to 42 more; n = 2, 3327 participants; I2 = 0%) (rating down by 1.5 for risk of bias and 0.5 for imprecision) (Appendix K Table 41; Fig. 2d (Appendix Q)) [75]. The certainty of the evidence was rated as low.
Internet intervention plus behavioural support versus non-internet-based non-active control
Abstinence/cessation in general/mixed population of smokers
Participants in five trials received an internet intervention plus behavioural support which was provided by nurses, peer coaches, or tobacco treatment specialists. Control participants received a non-active control condition which varied across trials (usual care, printed self-help guides, standard smoking cessation advice). In two trials, all participants (including controls) were using or offered pharmacotherapy. Trials varied with respect to how smoking abstinence was measured, most reported prolonged rates. Biochemical validation was used in 40% of studies.
Compared to non-active controls, 54 more people per 1,000 receiving an internet intervention plus behavioural support were abstinent at longest follow-up of 6 to 12 months (95% CI: 23 more to 92 more; n = 5, 2334 participants; I2 = 60%) (rating down one to two times for risk of bias, once for inconsistency, and 0.5 for imprecision) (Appendix K Table 53; Fig. 2d (Appendix Q)) [79]. The certainty of the evidence was rated as very low.
Individual smoking cessation intervention based on cognitive behavioural therapy and motivational interviewing plus NRT patch compared to Routine care
Abstinence/cessation in smokers with schizophrenia or schizoaffective disorder
A single trial reported on the effects of a combination intervention which involved eight hours of individual contact for eight weeks and NRT patch for about 10 weeks (21 mg, 14 mg, 7 mg titrated down). Participants in both study groups received booklets on smoking cessation as a co-intervention. Review authors report no statistically significant difference between groups in biochemically validated point prevalence or continuous abstinence rates at six months, 12 months, or four years (threshold for statistical significance was p < 0.01) (rating down once for risk of bias) (Appendix K Table 57) [58]. We were unable to assess imprecision and provide an overall certainty rating due to missing information .
Reduction in cigarettes per day of > 50% of baseline in smokers with schizophrenia or schizoaffective disorder
Review authors report no statistically significant difference between groups in smoking reduction at six months, 12 months, or four years (rating down once for risk of bias) (Appendix K Table 57) [58]. We were unable to assess imprecision and provide an overall certainty rating due to missing information.
Standard treatment plus extended NRT and extended CBT compared to standard treatment in smokers with past depression
Abstinence/cessation in smokers with past depression
The only trial in this analysis recruited general smokers but provided pre-stated subgroup data for participants with past depression. Standard treatment consisted of sustained release bupropion (300mg/day) for 12 weeks, nicotine gum (2 mg and 4 mg) for 10 weeks, five group counselling sessions from a counsellor, and a self-help manual; no further treatment was provided after week 12. In addition to standard treatment, those in the active arm received extended NRT (i.e., until week 52), and extended CBT (i.e., 11 individual CBT sessions from week 10 to 52).
Compared to standard treatment, 259 more people per 1,000 receiving standard treatment plus extended NRT and extended CBT were point prevalence abstinent (follow-up timepoint unclear); (95% CI: six fewer to 786 more; n = 1, 57 participants) (rating down once for risk of bias and twice for imprecision) (Appendix K Table 62; Fig. 2d (Appendix Q)) [57]. The certainty of the evidence was rated as very low.
Combined pharmacotherapy and behavioural interventions compared to usual care or minimal intervention
Abstinence/cessation in general/mixed population of smokers
The typical intervention involved multiple contacts with a specialist cessation counsellor combined with pharmacotherapy. Review authors reported that most of the trials offered one or more types of NRT, or bupropion. Usual care or minimal intervention was typically brief advice and self-help materials. Additional co-intervention (behavioural with or without pharmacotherapy) was provided to intervention participants in a few trials. Some trials provided behavioural and/or pharmacotherapy co-interventions to control participants. Trials varied with respect to how abstinence was measured but most used point prevalence. Biochemical confirmation of abstinence was used in 65% of studies.
At longest follow-up (six + months), 71 more people per 1,000 receiving combination pharmacotherapy and behavioural support were abstinent in comparison to those receiving usual care or minimal intervention (95% CI: 58 more to 84 more; n = 52, 19,488 participants; I2 = 36%) (rating down twice for risk of bias to also reflect concerns for indirectness) (Appendix K Table 49; Fig. 2d (Appendix Q)). The certainty of the evidence was rated as low. Review authors reported a larger effect in the subgroup of trials recruiting participants from health care settings (Appendix M). Other variables tested in subgroup and meta-regression analyses did not modify the effect; this included participants’ motivation to quit and intensity of behavioural support [81].
Combined pharmacotherapy and intensive behavioural interventions compared to usual care or no intervention
Abstinence/cessation in general/mixed population of smokers
Review authors excluded one study with a more intensive intervention from the main analysis presented above [81]. The study intervention consisted of 12 group sessions over 10 weeks, with advice from physician on risk for COPD, and 2 mg of nicotine gum for six months. Intervention participants were also randomized to bronchodilator or placebo. The control group received no intervention or usual care. No co-interventions were provided to either arm. The study reported biochemically confirmed point prevalence rates.
Compared to usual care or no intervention, 260 more people per 1,000 receiving the combination intervention were abstinent at 12-month follow-up (95% CI: 212 more to 315 more; n = 1, 5887 participants) (rating down once for indirectness) (Appendix K Table 50) [81]. The certainty of the evidence was rated as moderate.
5. Electronic cigarette intervention
E-cigarette versus Placebo
Although the stage II of this evidence review aims at synthesizing evidence on the harms and benefits of e-cigs, we have included results from the Lindson-Hawley systematic review [55] in this overview of review for completeness. In the single trial identified by Lindson-Hawley [55], participants in the active treatment group received e-cigarettes containing either 7.2 or 5.4 mg of nicotine for 12 weeks to assist smoking reduction; active arms were combined in analyses. Control participants received e-cigarettes without nicotine. Neither group received co-interventions.
Abstinence/cessation in smokers not motivated/wishing to quit
The trial reported biochemically validated abstinence defined as “not even a puff” since the previous visit. At 12-months follow-up, 70 more people per 1,000 using e-cigarette with nicotine were abstinent in comparison to those receiving e-cigarettes without nicotine (95% CI: 1 fewer to 270 more; n = 1, 300 participants), but it is possible that there is little to no difference between groups or that more participants in the active treatment arm stopped smoking (rating down twice for imprecision) (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. The certainty of the evidence was rated as low.
Reduction in cigarettes/day of > 50% of baseline or cessation in smokers not motivated/wishing to quit
At 12-months follow-up, 45 more people per 1,000 using e-cigarette with nicotine reduced the number of cigarettes per day by more than 50% (including those who quit), (95% CI: 38 fewer to 187 more; n = 1, 300 participants) (rating down once for risk of bias and twice for imprecision) (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. The certainty of the evidence was rated as very low.
Reduction in number cigarettes/day
Review authors indicate that there was no statistically significant difference between groups (median cigarettes per day = 12–14 in all groups), but it is unclear whether this is based on confidence intervals or p-values; we interpret this to mean that the confidence interval around the best estimate of effect includes the possibility of little to no difference between groups and a greater reduction in one group over another (rating down once for risk of bias) (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. Unable to assess the certainty of the evidence because number of participants included in analysis not reported for imprecision domain.
Reduction in carbon monoxide
Authors indicate that there was no significant difference between groups (median CO = 15–17 ppm in all groups), but it is unclear whether this is based on confidence intervals or p-values; we interpret this to mean that the confidence interval around the best estimate of effect includes the possibility of little to no difference between groups and greater reduction in one group over another (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. Unable to assess the certainty of the evidence because number of participants included in analysis not reported for imprecision domain.
Adverse events in smokers not motivated/wishing to quit
Review authors indicate that the frequency of adverse events was similar across groups at baseline, 12 weeks, and 52 weeks. There was a reduction in reported symptoms from baseline to 12-month follow-up across groups (p < 0.001). Rates of shortness of breath were reduced from 20–4% from baseline to two weeks. (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. We were unable to assess imprecision and provide an overall certainty rating due to missing information (i.e., number of participants included in the analysis). Authors stated that no serious adverse events occurred [55].
Weight gain in smokers not motivated/wishing to quit
Review authors indicate that there was no significant difference in weight change within or between groups. Regarding the latter, it is unclear whether this is based on confidence intervals or p-values; we interpret this to mean that the confidence interval around the best estimate of effect includes the possibility of little to no difference between groups and more weight gain in one group over another (Appendix K Table 35; Fig. 2e (Appendix Q)) [55]. Unable to assess the certainty of the evidence because number of participants included in analysis not reported for imprecision domain. Based on reporting, it is uncertain but unlikely that the outcome is post-cessation weight gain (i.e., does not appear to be assessed in only abstinent smokers).