Pharmacotherapy for Smoking Cessation in Schizophrenia: A systematic review

Abstract

Introduction:

Rates of tobacco smoking are high in people with schizophrenia with greater difficulty of quitting smoking compared to the general population, which also relate to the increased cardiovascular and cancer risks in this co-occurring disorder. Therefore, effective smoking cessation pharmacotherapies addressing tobacco co-morbidity are imperative.

Areas covered:

In this review, the authors performed an extensive systematic electronic literature review examining the efficacy and safety of first-line pharmacotherapies for smoking cessation, including varenicline, sustained-release bupropion, and nicotine replacement therapies (NRT) using continuous abstinence rates over 10–12-week periods in smokers with schizophrenia. Twelve trials reporting smoking cessation outcomes using interventions in schizophrenia were included and risk ratio (RRs) were used.

Expert opinion:

Our findings support the efficacy and safety of first-line pharmacotherapies for the treatment of tobacco use disorder in smokers with schizophrenia. Further research on the long-term effectiveness and safety of these agents in community samples are warranted. Smoking cessation pharmacotherapies may warrant the consideration of the emerging use of electronic nicotine delivery systems while neuromodulation techniques also offer promise.

1. Introduction

Tobacco use disorder is a significant health concern in people with schizophrenia [1]. Approximately 1% of the general population has schizophrenia, which is characterized by positive (delusions, hallucinations, thought disorder) and negative (affect flattening, amotivation, anergia, social isolation) symptoms, and cognitive deficits. Schizophrenia is associated with higher rates of tobacco smoking, and lower abstinence rates [1, 2]. Additionally, studies have found that smokers with schizophrenia extract more nicotine from each cigarette and smoke more heavily compared to non-psychiatric control smokers [3]. Considering that heavy smoking has been associated with increased substance misuse, greater positive symptom severity, and more frequent psychiatric hospitalization [4], the importance of the cholinergic system relating to tobacco use disorder in the pathophysiology of schizophrenia is emphasized [5]. Thus optimizing current smoking cessation pharmacological therapies among individuals diagnosed with schizophrenia is imperative to reduce morbidity and mortality in this population.

First-line pharmacotherapies include nicotine replacement therapy (NRT), sustained-release bupropion, and varenicline, which are approved by US Food and Drug Administration (FDA) for smoking cessation [6, 7]. NRTs deliver nicotine with the purpose of reducing tobacco use by decreasing craving and withdrawal symptoms during abstinence [8]. Available forms include patch, inhaler, nasal spray, lozenge and gum [8]. Bupropion is a unicyclic atypical antidepressant approved by the FDA under the name Zyban® in the sustained-release (SR) formulation. Originally marketed as an antidepressant (Wellbutrin), in 1997 bupropion SR was the first non-nicotine medication approved for smoking cessation [9]. It acts as a dopamine and norepinephrine reuptake inhibitor, reducing the deficit of dopamine and brain reward function caused by nicotine withdrawal, and as well is a non-competitive antagonist of the nicotinic acetylcholine receptor (nAChR) α3β2 [10]. Since then, numerous studies have examined the efficacy of bupropion for smoking cessation compared to placebo in the general population [11]. Lastly, varenicline (Chantix®), is a selective α4β2 nAChR partial agonist and full α7 nAChR agonist that has recently become the most effective smoking cessation pharmacotherapy available [12]. Inspired by the natural product of (-)-cysteine and its partial agonist properties, varenicline was FDA approved in May of 2006 [13]. In the absence of nicotine from cigarette smoking, varenicline acts as an agonist resulting in release of dopamine and reduces withdrawal and craving symptoms [14]. While in the presence of nicotine with smoking, varenicline blocks the α4β2 receptor resulting in decreased pleasure of smoking and attenuating dopaminergic reward response to nicotine [15]. Varenicline has been reported to be more effective than bupropion and NRTs due to its partial agonist effect [16, 17].

Taken together, different treatments incorporate different mechanisms for smoking cessation. Numerous studies have examined and confirmed the promising efficacy of varenicline, bupropion and/or NRT for smoking cessation in the general population, with recent suggestions of optimal efficacy from combining therapies [18]. The EAGLES trial evaluated the safety of NRT, bupropion and varenicline in a psychiatric and non-psychiatric cohort. They found varenicline to be more effective than the other cessation therapies to help smokers achieve abstinence and no neuropsychiatric adverse events in any active treatment [16]. However, a limited number of studies have been completed in smokers with schizophrenia, a population more nicotine dependent and possibly differing in effectiveness of pharmacotherapies relating to some of the shared neurobiological pathways of the pharmacotherapies and systems implicated in schizophrenia and nicotine addiction risk. Individuals with schizophrenia tend to be heavier smokers (and more dependent), which could have implications for the effectiveness of NRT if it is not appropriately dosed. Therefore, in this systematic review our primary objective was to evaluate evidence for the efficacy and safety of using pharmacological strategies (i.e., varenicline, bupropion and NRTs) for smoking cessation in patients with schizophrenia.

2. Methods

Studies included randomized controlled trials examining the effectiveness and safety of the smoking cessation interventions (varenicline, bupropion and NRTs). Types of inclusion criteria for participants were: 1) smokers (10 or more cigarettes per day (CPD)), 2) between 18 and 70 years of age, 3) with schizophrenia or schizoaffective disorder and 4) no other Axis I Disorder. In some studies, non-abstinence treatment outcomes were encouraged such as smoking reduction [19]. Therefore, this review only includes studies with outcome measures of 7-day point prevalence or continuous abstinence (at end point of treatment or through-out) in the study trials with schizophrenia patients undergoing abstinence initiation (smoking cessation lasting 10–12 weeks) with a set target quit date (TQD), rather than relapse-prevention approaches.

This systematic review followed the guidelines of Preferred Reporting Items for Systematic Review and Meta-analysis (PRISM-A) [20]. The search strategy included the keywords: ‘schizophrenia’, ‘smoking cessation’, ‘varenicline’, ‘bupropion’, and ‘nicotine replacement therapy’. Keywords were also searched in combination (i.e., “schizophrenia” AND “smoking cessation” AND (“varenicline” OR “bupropion” OR “nicotine replacement therapy”)). We searched Google Scholar, PubMed, Ovid MEDLINE, and Ovid PsycInfo from database inception to October 30^th 2019 published in English. Data extraction by two authors (KK and TG) initially assessed titles and abstracts identified by the search, and then reviewed the full text of the remaining articles for inclusion, and entered into Excel and SPSS databases. Risk of bias across individual studies was not assessed and is a limitation of this review. We excluded reviews, case series/reports, comments, opinion, unpublished studies, conference posters, and abstracts only. Studies in the final stage of selection are in Table 1, the decision about whether or not to include each relevant study, and the reasons for excluding the articles not chosen.

Table 1.

Efficacy of pharmacotherapies for smoking cessation in smokers with schizophrenia

Study or Subgroup	Drug (n)	Placebo (n)	Quit	Not Quit	Continuous Abstinent Rates	Relative Risk (to quit)	95% CI	p =
1.1 Varenicline vs Placebo
Dutra et al. 2012 (open-label)	53		32	21	60.40%
Williams et al. 2012	84	43	16 vs 2	68 vs 41	19% vs 4.7%	4.1	0.99 to 17.00	0.052
Weiner et al. 2011	4	4	3 vs 0	1 vs 4	75% vs 0%	7	0.47 to 103.28	0.157
1.1 Combined Studies			19 vs 2	69 vs 45		5.07	1.24 to 20.85	0.024
1.2 Bupropion vs Placebo
Weiner et al. 2012	22	19	4 vs 2	18 vs 17	18% vs 11%	1.73	.36 to 8.41	0.499
Evins et al. 2005	25	28	4 vs 0	21 vs 28	16% vs 0%	10.04	0.57 to 177.66	0.116
George et al. 2002 (EOT)	16	16	8 vs 2	8 vs 14	50% vs 12.5%	4	1.00 to 15.99	0.05
Evins et al. 2001	9	9	1 vs 0	8 vs 9	11% vs 11%	3	0.14 to 65.16	0.484
1.2 Combined Studies			17 vs 4	55 vs 68		4.25	1.50 to 12.01	0.006
1.3 NRT
Horst et al. 2005 (single-blind)	8	9	8 vs 3	0 vs 6	100% vs 33.3%	3	1.19 to 7.56	0.02
George et al. 2000 (ATYP vs TYP) (open-label)			8 vs 5	10 vs 22	44.4% vs 18.5%	2.4	0.93 to 6.17	0.069
1.3 Combined Studies			16 vs 8	10 vs 28	61.5% vs 22.2%	2.77	1.40 to 5.48	0.003
1.4 Bupropion+NRT+CBT vs Bupropion+NRT vs Tau+CBT
Brody et al. 2017	34	N/A	5 vs 2 vs 1	6 vs 8 vs 12	27.6% at 10-week with CBT
George et al. 2008 (EOT) (BUP+NRT vs NRT)	29	29	10 vs 3	19 vs 26	34.5% vs 10.3%	3.33	1.02 to 10.88	0.046
Evins et al. 2007 (BUP+NRT vs NRT)	25	26	9 vs 5	16 vs 21	36% vs 19%	1.87	0.73 to 4.82	0.194
1.4 Combined Studies			19 vs 8	35 vs 47		2.42	1.16 to 5.05	0.019

Significant p<0.05

To examine the efficacy of each intervention, relative risk ratio (RRs) and pooled RRs were obtained using SPSS version 26 for Windows. Dichotomous outcomes (i.e., abstinent or not) were represented by arranging the observed counts into fourfold (2 (treatment vs placebo) by 2 (abstinent vs not)) tables along with 95% confidence interval (CI). Continuous abstinence was defined as being continuously abstinent for a period of two or more weeks after a quit date is set or the last 4 weeks of the trial. For some studies RRs could not be calculated because one of the four subgroups had a frequency of zero. This problem was resolved by obtaining pooled RRs by summating the frequencies of each subgroup for every study. Studies examining the effectiveness of NRTs alone were limited and open-label [21, 22], thus did not allow for comparative analysis for this intervention by calculating RRs. Two studies [23, 24] using combined treatments (bupropion and NRT), were included in the bupropion pooled RRs. If available, RRs of smoking cessation at longest follow-up was reported, as well as nicotine dependence measures (Table 2).

Table 2.

Efficacy of pharmacotherapies in follow-up phase for smoking cessation in smokers with schizophrenia

Study or Subgroup	Time-Frame	Counseling	Follow-Up	Time-Frame	Quit	Not Quit	Relative Risk (to quit)	95% CI	p =
1.1 Varenicline vs Placebo
Dutra et al. 2012	12 weeks	Yes	No	N/A	N/A	N/A	N/A	N/A	N/A
Williams et al. 2012	12 weeks	Yes	Yes	24 weeks	10 vs 1	74 vs 42	5.12	0.68 to 38.69	0.114
Weiner et al. 2011	12 weeks	Yes	No	N/A	N/A	N/A	N/A	N/A	N/A
1.1 Combined Studies
1.2 Bupropion vs Placebo
Weiner et al. 2012	12 weeks	Yes	No	N/A	N/A	N/A	N/A	N/A	N/A
Evins et al. 2005	12 weeks	Yes	Yes	24 weeks	1 vs 1	24 vs 27	1.12	0.07 to 16.98	0.935
George et al. 2002 (EOT)	10 weeks	Yes	Yes	26 weeks	3 vs 1	13 vs 15	3	0.35 to 25.87	0.318
Evins et al. 2001	12 weeks	Yes	Yes	24 weeks	3 vs 1	6 vs 8	3	0.38 to 23.68	0.297
1.2 Combined Studies					7 vs 3	43 vs 50	2.47	0.68 to 9.04	0.171
1.3 NRT
Horst et al. 2005	9 months	Yes	No	N/A	N/A	N/A	N/A	N/A	N/A
George et al. 2000	12 weeks	Yes	Yes	26 weeks	9 vs 5	16 vs 21	1.87	0.73 to 4.82	0.194
1.4 Bupropion+NRT+CBT vs Bupropion+NRT vs Tau+CBT
Brody et al. 2017	12 weeks	Yes	Yes	24 weeks	5 vs 2 vs 1	6 vs 8 vs 12	N/A	N/A	N/A
George et al. 2008 (EOT)	10 weeks	Yes	Yes	26 weeks	4 vs 0	25 vs 29	9	0.51 to 159.95	0.135
Evins et al. 2007	12 weeks	Yes	Yes	52 weeks	5 vs 2	15 vs 6	1	0.24 to 4.14	1
1.3 Combined Studies					9 vs 2	40 vs 35	3.4	0.78 to 14.80	0.103

Significant p<0.05

3. Results

A total of 413 studies were found in the overall search strategy, and 385 studies were excluded (193 were duplicates and the content of 192 studies was not pertinent) (Figure 1). 28 full text articles were assessed for eligibility, and 12 studies were included (reasons for exclusion in Supplementary Material).

Figure 1:

Study Selection PRISMA flow diagram: * Included animal studies, non- biological science studies, reviews, editorials, and commentaries.

This search resulted in a total of two double-blind, randomized, placebo-controlled trials using varenicline and one open-label study, and four placebo-controlled bupropion studies in patients with schizophrenia (Figure 2; Table 1). In addition, two NRT studies were included that were open label studies. Finally, three studies were included that used combined bupropion and NRTs. The remaining sixteen studies excluded are summarized in the methods section with reasons for their exclusions.

Figure 2.

Study Flow Chart

2.1. Varenicline

Since its appearance on the smoking cessation market, several studies have examined the effectiveness of varenicline in the general population with more than double the chances of successfully quitting [25]. However a limited number of studies, though with promising sample sizes, have examined its effect in the schizophrenia population.

In the current systematic review, we found two studies (a total sample size of 135 smokers with schizophrenia), examining the efficacy of 12-weeks treatment with varenicline (titrated up to 1mg BID) [26, 27]. A pooled RR of both showed that varenicline improved abstinence at endpoint compared to placebo group (RR=5.07, 95% CI=1 1.24–20.85; p=0.024). Williams et al. 2012 also provided data of a 12-week post-treatment follow-up phase, which showed that patients who took varenicline were approximately six times more likely to sustain abstinence compared to the placebo group; however, this difference was not significant (RR=5.12, 95% CI=0.68–38.690; p=0.114).

In an attempt to determine the safety of varenicline, we also examined the psychiatric rating scales (i.e. Brief Psychiatric Scale, Positive and Negative Symptom Scales) and the frequency of adverse events. Both studies showed no significant differences between varenicline and placebo groups in positive symptoms of schizophrenia throughout trials. Additionally, Williams et al. found no exacerbation of negative symptoms between groups and reported no differences between the treatment and follow-up phase [27], while Weiner et al. found no significant differences in depression symptoms [26]. Furthermore, both studies combined were similar in reporting the most common side effects of insomnia, constipation, headache and nausea, with the latter two found more significant in the varenicline group compared to placebo. Williams et al. 2012 reported 13 serious adverse events (SAEs), with two patients in the varenicline group having 3 SAEs considered to be study drug related, though both patients had a prior history of suicide attempts and hospitalization.

Another study examined 13 weeks of open treatment with varenicline (titrated up to 1mg BID for 11 weeks) [28]. Of the entire sample (n=53), 60.4% of the participants met 14-day point prevalence abstinence at week 12, with no SAEs reported. The authors also found that patients with lower baseline affective flattening scores were more likely to attain smoking cessation with varenicline (p=0.044).

2.2. Sustained-Release Bupropion

Our literature search revealed four studies examining the efficacy of bupropion (150 mg BID) for smoking cessation in smokers with schizophrenia in 10 to 12-week trials with a total sample size of 144 participants [29–32]. Pooled RR of the aforementioned studies was computed and found that participants in the bupropion group were almost five times as likely to continuously abstain from smoking than the placebo group (RR=4.25, 95% CI=1.50–12.01, p=0.006). In combined follow-up assessments for continued abstinence, bupropion was also found to be more effective than placebo, but this difference was not significant (RR=2.47, 95% CI=0.68–9.04, p=0.171).

All four studies found that compared with placebo, bupropion did not worsen positive, negative or depression symptoms throughout trials. Notably, the major side effects noted in bupropion versus placebo group were all non-significant differences with the exception of dry mouth. Weiner et al 2012 reported five participants dropping out from the bupropion group due to side effects including worsening of anxiety, psychosis and development of a rash, and any SAEs reported were deemed unrelated to study medication. Taken together, treatment with bupropion resulted in a significantly greater rate of abstinence in smokers with schizophrenia, when compared to placebo trials both at endpoint of treatment and follow-up.

2.3. Nicotine Replacement Therapies (NRTs)

To date there is still limited information about the outcomes of nicotine-patch therapy alone in smokers with schizophrenia using a double-blind study design. Two studies reviewed that are open-label NRT trials are worthy of noting as to compare the efficacy of NRT to the other pharmacotherapies discussed [21, 22]. One study explored the efficacy of NRT compared to placebo in smokers with schizophrenia over a 9-month period [22]. Participants who quit smoking completely in the first three months were then randomized into a single-blind phase of receiving nicotine patches or placebo. Nicotine patches were dosed with nicotine based on individual cotinine saliva levels to minimize cravings. The authors found that those with NRT were significantly more likely to continuously remain abstinent compared to the placebo group (RR=11.7, CI=t0.76–179.73, p=0.078), verified by bi-weekly CO levels.

Another group also investigated the effect of NRT in an open-label trial, however, with the novel approach of controlling and stratifying groups into atypical and typical antipsychotics [21]. Though at the end of this 12-week trial smoking cessation rates in smokers with schizophrenia were modest, it was found that patients taking atypical antipsychotics were marginally more likely to continuously abstain than the typical group (RR=2.40, CI=0.93–6.17, p=0.069). In the 6-month follow-up, the atypical group was not significantly more likely to continuously abstain when compared to the typical group (RR=1.87, CI=0.73–4.82, p=0.194). Pooled RR could not be calculated for these studies due to them being open-label studies. However, both studies reported no significant differences between NRT and placebo in any psychiatric symptoms of schizophrenia, as well as no serious adverse events.

2.4. Combined Bupropion and NRTs

The last group included studies that compared the effects of a combination treatment of bupropion and NRTs in smokers with schizophrenia. Brody et al. (2017) found that the best smoking cessation outcomes and abstinence rates resulted for participants in the combination extended treatment (COMB EXT) (bupropion, nicotine patch and lozenges) and therapy through home visits (HV) versus treatment as usual (regular medication) + therapy through home visits [33]. Similar results were obtained for the same group without HV, but the HV therapy increased the participants’ odds of remaining abstinent. Seven-day abstinence rates between the two groups were of 45% for COMB EXT + HV, 20% for COMB EXT without HV and 8% for treatment as usual (TAU, regular medication only) + HV, with rates significantly different between the COMB-EXT+HV vs. TAU group (χ²₍₁₎=4.8, P=0.03).

Similarly, George et al. (2008) and Evins et al. (2007) examined smoking abstinence outcomes in two groups: bupropion in combination with NRTs (transdermal nicotine patch or nicotine gum respectively) vs. placebo+NRTs. Both studies found greater rates of abstinence in the combination groups (Table 1). Pooled OR of these two latter studies was computed as these had comparable groups, and found that participants in the bupropion+NRT group were three times as likely to continuously abstain from smoking than the placebo+NRT group (RR=2.42, 95% CI=1.16–5.05, p=0.019). In combined follow-up assessments for continued abstinence, bupropion+NRT was also found to be more effective than placebo+NRT (RR=3.40, 95% CI=0.78–14.80, p=0.103).

George et al. (2008) did report three serious adverse events, (two in the placebo group and one in the bupropion group), however, they were all unrelated to the treatments. The other two studies reported no serious adverse events. Overall these studies demonstrate the tolerability of combination treatments.

2.5. Excluded Studies

A total of sixteen studies were excluded from this review, as they did not meet the inclusion criteria previously mentioned. Four studies [16, 34–36] were not specific to schizophrenia, as both mood and anxiety disorders were included in the participant pool.

Three studies were from a relapse prevention trial [37], two of which [38, 39] were based on a parent larger trial [40] that was a multi-site relapse-prevention trial which included patients with bipolar disorder, while this review only included trials with schizophrenia patients undergoing abstinence initiation (smoking cessation) with a set TQD. Six other studies were also excluded as no set “quit date” was set as part of the study design [41–46] or smoking cessation data [47]. In addition, Smith et al. (2016) only required a total of 6 cigarettes per day, while our inclusion criteria was of 10 cigarettes per day. Another study [48] was a secondary analysis of the George et al. (2002, 2008) studies mentioned in the Bupropion section.

3. Discussion

3.1. Efficacy and Safety

The findings of this systematic literature review suggest varenicline and bupropion SR are both effective pharmacotherapies for continuous smoking abstinence in smokers with schizophrenia as shown by significant RRs for smoking cessation when compared to placebo. Further, in contrast to the indications that prompted a 2009 black box warning, varenicline was not associated with increased psychiatric symptoms, suicidal ideation or depression [49]. Similarly, bupropion was also found to be well-tolerated and safe for use in patients with schizophrenia with no association of increased psychiatric or depressive symptoms. Nonetheless, individual participants experienced some adverse events (e.g. nausea, headache, insomnia) therefore clinical vigilance is warranted. A study conducted by George and colleagues found that atypical agents might be superior to typical agents in combination with NRTs for smoking cessation in schizophrenia [21]. In line with this, future studies may consider to investigate whether effects of varenicline and bupropion are similar and safe across these two agent groups, in both short-term and long-term smoking cessation trials, which could also account for why some trials were found to result in improved negative psychiatric outcomes. One should be cautious with these medications, because while smoking cessation does not alter the metabolism of varenicline or bupropion, cessation can lead to the de-induction of the liver enzyme, CYP1A2 [50], leading to a elevation of plasma levels of certain psychotropic medications (e.g. clozapine, olanzapine tricyclic antidepressants, caffeine) [51–53].

Follow-up assessments with respect to efficacy and safety of varenicline and bupropion appear promising. Thus beneficial effects of varenicline and bupropion are not only limited to smoking cessation but also have implications on long-term health outcomes. Although none of the above studies reviewed mentioned any long-term health outcomes, it is important to comment on downstream benefits of smoking cessation. These health benefits include decreased rates of other co-morbidities including cancers, cardiovascular, and respiratory diseases, which otherwise have ultimately led to higher rates of mortality associated with tobacco use [54]. Notably, one study found smoking cessation pharmacotherapies to increase life expectancies and decrease smokers for hospitalization for an acute myocardial infarction [55]. Thus clinical outcomes appear encouraging and pharmacotherapeutic aids such as varenicline and bupropion should be further investigated in smokers with schizophrenia in successfully preventing tobacco-related morbidity and mortality.

Reviews to date have been mixed, with similar findings to ours of higher incidence of smoking cessation following bupropion treatment [7, 56–58] and varenicline [7] as compared to the placebo group, and that all treatments are well tolerated (Table 3). Reviews also reported findings contrary to our results regarding the efficacy of varenicline, it should be noted that these reviews included smoking cessation trials of smokers with schizophrenia and bipolar disorder [59]. Our pooled RR analysis included studies with similar criteria to strengthen findings of the efficacy of smoking cessation aids. We found that compared to placebo, bupropion and varenicline both appear to be effective pharmacotherapies for smoking cessation in the schizophrenia population.

Table 3:

Prior review publications on pharmacotherapies for smoking cessation in smokers with schizophrenia

Study or Subgroup	Type of Review	Number of studies included	Total Sample (N)	Treatments Groups	Quit rate	Risk Ratio	95% CI	Effects on Psychaitric Symptoms (+, -, depressive)
Tsoi et al. 2010	Systematic	7	260	bupropion vs. placebo	bupropion>placebo	2.57 (p=0.004)		no differrences
Tsoi et al. 2013	Cochrane	7	340	bupropion vs. placebo	bupropion>placebo	3.03	1.69 to 5.42	no differrences
		2	137	varenicline vs. placebo	varenicline>placebo	4.74	1.34 to 16.71	no differences; 2 reports of suicidal ideation in varenicline group
Little evidence for the efficacy of transdermal nicotine patch (TNP)
Kishi and Iwata 2014	Meta-analysis	5	322	varenicline vs. placebo		0.79	0.58 to 1.08	no increase in suicide ideation and depression; more nausea than placebo group (RR=1.79, CI=1.20 to 2.67
Shawen and Drayton 2018	Non-systematic	18	NRTs, bupropion and varenicline all shown to be an effective treatment options					All three treatment options are well tolerated
Stubbs et al. 2015	Systematic	Conclusions: support for bupropion effectiveness; limited evidence in the effectiveness of NRT and varenicline						No increases in symptoms

This review adds to previously published reviews as we believe it provides the most rigorous assessment to date of smoking cessation aids in schizophrenia. with extensive eligibility criteria, and additional statistical approaches of RRs comparisons. Finally, to our knowledge this is the first review to support and compare the findings of the trials conducted between 2008–2019 when considering continuous abstinence study designs, necessary time periods, and eligibility requirements in smoking cessation treatment approaches for smokers with schizophrenia.

3.2. Limitations and Future Directions

There are several limitations to our review. First, standard smoking cessation therapies are effective in helping schizophrenia smokers quit in the short (10–12 weeks) but not in the long-term. Studies have shown that rapid relapse after six-months discontinuation of active pharmacotherapies, with approximately 35% of the general population relapsing compared to 60% of people with schizophrenia [31, 60]. Here, bupropion was found to have a higher rate of continuous abstinence in follow-up phases with smaller confidence intervals compared to the varenicline group, though the time frames varied and three studies did not assess follow-up abstinence rates [22, 26, 29]. However, efficacy of treatment effects across all three pharmacotherapies decreased in follow-up periods. The beneficial effects of nicotine from cigarette smoking have on clinical and cognitive deficits previously observed in schizophrenia smokers may explain this [31]. To add, other common determinants and high sensitivity to smoking cues may also account for the drop in abstinence rates in follow-up phases. Therefore, to determine optimal duration of maintenance pharmacotherapies, longer duration studies are needed with time to relapse assessments, as well as examining any exacerbations of positive symptoms of schizophrenia.

Second, most if not all studies reviewed were of short duration (e.g. 10–12 weeks) and therefore do not reflect the potential health benefits of long-term smoking cessation, such as decreased risk of cardiovascular and respiratory diseases. Acute and chronic smoking leads to different downstream health outcomes and should also be considered; varying from different cost effects, to differences in life expectancies [61, 62]. Therefore future research should examine time to relapse using relapse-prevention clinical trial designs (e.g. Evins et al., 2014), as well as any long-term heath benefits associated with smoking cessation, by incorporating greater clarity on the magnitude of benefits associated with varenicline and bupropion treatment. Because health benefits of smoking cessation increases with duration of abstinence, longer trials of combination and alone pharmacotherapies are suggested for prevention of relapse.

Third, many studies used in this review were limited to small sample sizes which may account for certain abstinence outcomes, as well as psychiatric and adverse events not reaching statistical significance when compared between different pharmacotherapy aids at end of treatment. A major limitation of directly evaluating all three pharmacotherapies by comparing RR relates to the limited number of NRT trials using a double-blind study design, and smoking cessation trials using varenicline, alone within the schizophrenia population (as apparent with wider confidence intervals when compared to bupropion). One solution might be the development of long-term treatment paradigms that compares the three pharmacotherapies in schizophrenia-smoking patients, with a long duration of follow-up (e.g. similar to studies by Hall et al., 2005 in major depressive disorder [63]). Adding home visits may also result in a doubling of treatment efficacy and abstinence rates due to potential higher awareness of behavioral smoking-related triggers in a familiar (home) environment [33].

4. Conclusions

This review supports the notion that maintenance pharmacotherapy with varenicline and sustained-release bupropion are both well-tolerated and do not worsen psychiatric outcomes over treatment periods as well as in follow-up phases. There are distinct mechanisms for the anti-smoking effects of NRT, bupropion and varenicline, which underlie their success as smoking cessation aids [6]. Each of these pharmacotherapies should be considered for long-term management of tobacco addiction in people with schizophrenia and other serious mental illnesses, which is consistent with previous studies in schizophrenia and bipolar disorder [27, 64]. Ultimately discussing what the next trials should be to clearly determine the best treatment for smoking in schizophrenia would be informative.

The importance of future research exploring the varying indirect health benefits of all three pharmacotherapies is warranted, with larger sample sizes resulting in adequate power to detect treatment effects and should also consider reporting time to relapse assessments. Continued exploration of reporting outcomes beyond end of treatment may lead to a better understanding of long-term benefits of smoking cessation in these patients.

Overall, evidence for effectiveness of approved pharmacotherapies in smokers with schizophrenia is promising. Conceptualizing smokers with schizophrenia as having a co-morbid disorder with multi-factorial determinants amenable to the combination of pharmacotherapy in combination with behavioral therapy may reduce smoking prevalence, as well as various health and cognitive problems present in these patients.

5. Expert Opinion

The high prevalence of smoking within schizophrenia is associated with lower abstinence rates and contributes to increased cardiovascular and cancer risks. Thus, smoking cessation pharmacotherapies such as varenicline, bupropion SR and NRTs may successfully prevent tobacco-related morbidity and mortality in people with schizophrenia. This systematic review found significantly higher continuous abstinence rates at endpoint of treatment with varenicline or bupropion, as well as showed support of the safety and tolerability of first-line pharmacotherapies for treatment of tobacco use disorder in smokers with schizophrenia. Combination treatments (bupropion and NRTs) were also found to result in high rates of continuous abstinence from smoking, as well as being well-tolerated and safe.

Each of these pharmacotherapies should be considered for longer term management of tobacco smoking cessation amongst smokers with schizophrenia. Determining which treatment is most effective depends on future research and replication studies with larger sample sizes to detect treatment effects. The conceptualization of smokers with schizophrenia as having a co-morbid disorder with multiple determinants amendable to the combination of pharmacotherapies may reach the goal of reduced tobacco smoking prevalence as well as improve other debilitating factors present within these patients, such as cognitive deficits.

Common challenges observed in randomized clinical trials of smoking cessation pharmacotherapies in smokers with schizophrenia are small sample sizes, as well as, effectiveness tapering in the long-term as indicated by follow-up periods. Thus to determine optimal duration of maintenance pharmacotherapies, longer and larger duration trials are warranted with time to relapse assessments included.

Moreover, smoking cessation pharmacotherapies may warrant considerations of the emerging use of electronic nicotine delivery systems (i.e., e-cigarettes and similar devices). Although suggestions of ENDS as harm reduction approaches or even as an aid for smoking cessation have been made [65], further research is greatly needed amongst vulnerable populations such as those with schizophrenia, before being recommended as a method for smoking cessation.

Finally, there is the potential of neuromodulation techniques having considerable promise as new a treatment modality for tobacco use in schizophrenia [66]. For instance, repetitive transcranial magnetic stimulation (rTMS) which is a non-invasive brain stimulation procedure, has shown tremendous promise and tolerability in short-term intervention trials for suppressing tobacco craving in schizophrenia [67] and lowering cigarette consumption [68]. However, study findings have been mixed [69], thus further studies using rTMS as well as other neuromodulation methods (i.e., transcranial direct current stimulation (tDCS)[70]), are needed to validate the efficacy of this treatment option.

Article Highlights Box.

• Rates of tobacco smoking amongst those with schizophrenia (~45–88%) are three times higher than the general population (<16%). Moreover, greater difficulty of quitting smoking is prevalent amongst this co-occurring disorder.

• One or more first-line pharmacotherapies for smoking cessation are recommended for patients, including varenicline, sustained-release bupropion (SR), and nicotine replacement therapies (NRT).

• This review examined the efficacy and safety of these pharmacotherapies using continuous abstinence rates (over 10–12 weeks) in smokers with schizophrenia.

• Varenicline and bupropion SR are both effective pharmacotherapies for smoking cessation in smokers with schizophrenia as shown by significant risk ratios when compared to placebo. These treatments are also well-tolerated and safe.

• Medications in combination (bupropion and NRTs) are also an effective treatment option that should be considered for patients motivated to quit smoking.