Review Article: Ulcerative Colitis, Smoking and Nicotine Therapy

P. C. Lunney; R. W. L. Leong


Aliment Pharmacol Ther. 2012;36(11):997-1008. 

In This Article


Ulcerative Colitis and Smoking

The risk of developing UC in current smokers compared with lifetime nonsmokers is reported at an odds ratio (OR) of 0.41 (95% CI: 0.34–0.48).[10] Conversely, the risk of development of UC in lifetime nonsmokers compared to current smokers has an OR of 2.9 (95% CI: 2.6–3.2).[10] There is overall consensus that there is demonstrable greater risk for the development of UC in lifetime nonsmokers compared to current smokers.[5] There also appears to be a rebound risk to the development of UC following the cessation of smoking.[1,4–12,16,21,22,29–32] This is generally reported as a pooled OR of 1.64 (95% CI: 1.36–1.98) when compared with lifetime nonsmokers' risk.[10] This risk is particularly high in the first few years following cessation of smoking.[2,5–8,20,21,29,31,32]

Ex-smokers have been observed to exhibit delayed onset of disease compared with lifetime nonsmokers.[8,21,32] In two studies, the mean age difference in men of disease onset between nonsmokers and those who previously smoked was found to be 15.2[32] and 16.1 years.[21] Surprisingly, this phenomenon is only true of male subjects, as there was no difference observed in disease onset between female nonsmokers and ex-smokers.[21,32] Interestingly, however, there does not appear to be any significant difference between ex-smokers and nonsmokers in disease extent, progression, or regression.[4,21]

On further examination of the relationship between UC and smoking, the trend of a dose-dependent relationship emerges.[1,4,5,10,16,24] That is to say, heavier smokers are less likely to develop UC than lighter smokers[4] and current smokers who have a diagnosis of UC are more likely to have milder disease than ex-smokers and nonsmokers with UC.[1,4,5,24] Indeed, relatively heavy smokers have been observed to have less macroscopic and histological evidence of disease on colonoscopic examination, as compared with lighter smokers, ex-smokers and nonsmokers with UC.[4] Intermittent smokers have often been reported to experience symptomatic exacerbation during periods of nonsmoking and alleviation of symptoms on recommencement of smoking.[2,5,6,15,20] The observation that heavier smokers who quit smoking are more susceptible to the rebound effect of developing UC, compared to lighter smokers, also corroborates the notion of a dose-dependent relationship.[6,9,10,16]

Current smokers with UC have been found to have less relapses,[33] require fewer hospitalisations[1,6,24,34,35] and demonstrate a reduced need for oral corticosteroid and immunosuppressant therapy for their disease, compared to ex-smokers and nonsmokers with UC.[1,5,24] However, not all studies have observed a significant difference in this regard between these patient groups.[6,35] Those patients who quit smoking after diagnosis of UC often experience an exacerbation in their disease course, and have been uniformly shown to have increased hospitalisation rates and need for oral corticosteroids and immunosuppressants.[1,2,6,24,34,36]

It remains divided as to whether smoking reduces the rate of surgery required in UC, particularly colectomy. Fraga et al.[11] found that nonsmokers and especially ex-smokers underwent surgery for their UC more frequently than current smokers. Odes et al.[35] also reported active smoking's protective effect against surgery in patients with UC. Aldhous et al.,[4] however, observed no difference in time from diagnosis of UC to colectomy in those patients requiring surgery, between current smokers and nonsmokers; however, found that ex-smokers demonstrated a shorter time period between diagnosis and colectomy than nonsmokers. Boyko et al.[34] found that ex-smokers more frequently underwent colectomy than current smokers or nonsmokers; however no such difference was observed between current smoker and nonsmoker groups. A meta-analysis of studies totalling 1489 patients with UC yielded an OR of 0.57 (95% CI: 0.38–0.85) for total colectomy in current smokers compared to nonsmokers.[37] Despite this, others have not found the overall proportion of patients from current smoker, ex-smoker and nonsmoker groups undergoing colectomy to be significantly different.[4–6,8,36] However, there appears to remain a predominance of opinion that current smoking in UC may reduce the need for colectomy.[1,5,11,24,37]

Smoking's influence over extraintestinal manifestations (EIMs) of disease remains diverse. One study found that current cigarette smoking in UC patients increased the risk of articular and dermatological EIMs, whereas no such increase was observed in ocular or hepatobiliary complications.[38] Current or previous smoking has also been shown to reduce instances of primary sclerosing cholangitis[1,5,6,24,39–41] and backwash ileitis.[24]

The literature remains divided, however, regarding smoking's effect on pouchitis following colectomy with ileal pouch-anal anastomosis. Although some cite a decreased rate of pouchitis among current smokers[5,42] others still observed no significant difference.[6,24,43] In fact, both van der Heide et al.[24] and Joelsson et al.[43] found a preponderance to pouchitis in current smokers compared to nonsmokers, but noted that the difference was not significant.

Conflicting results exist surrounding the effects of passive smoking on the development of UC.[6,44,45] Hypotheses have been formulated, postulating passive smoking may exert some protective effects similar to those observed in subjects who actively smoke. However, no association between passive smoking and future development of UC has been established.[5,24,44,45] Nonsmoking patients with a diagnosis of UC, who were regularly exposed to passive smoke, had no difference in the rate of hospitalisation, need for medication or need for surgery, when compared to other nonsmokers with UC who were not in regular contact with passive smoke.[24] Surprisingly, this research yielded results, which indicate that those patients regularly exposed to passive smoke had increased rates of pouchitis and backwash ileitis.[5,24]

Smoking may bear some influence over the determination of phenotype in IBD. The effect of smoking on Crohn's disease (CD) is positively correlated with its development and progression, juxtaposing it to the relationship between smoking and UC. The pooled OR of developing CD between current smokers and lifetime nonsmokers has been calculated as 2.0 (95% CI: 1.65–2.47) and 1.80 (95% CI: 1.33–2.51) between ex-smokers and lifetime nonsmokers.[10] The risk of smoking to the development of CD may be significantly greater to women than it is to men, with one study demonstrating a threefold relative risk increase.[45] Bridger et al.[46] conducted a study involving 87 sibling pairs with IBD who were discordant for smoking status at diagnosis. Of these 87, 23 sibling pairs were also discordant for disease type – one developed UC and the other CD. In 21 out of the 23 cases, CD developed in the active smoker and UC in the nonsmoker.[46] Similarly, other studies have confirmed the association of smoking with CD and nonsmoking with UC in the context of IBD phenotype expression within families.[47,48]

Current smoking has also been demonstrated to exert a deleterious effect on the clinical course of CD after diagnosis.[13,49–53] Interestingly, however, some studies have established a higher proportion of ileal involvement and a lower prevalence of colonic involvement in CD patients who actively smoke.[52–54] This suggests that smoking may provide a protective effect to the colon, common to both CD and UC.

Proposed Mechanisms in the Influence of Smoking on Inflammatory Bowel Diseases

No single mechanism has adequately explained the effects of smoking on IBDs and the exact pathogenic compound(s) remain unknown. Among the possible theories are the effects of smoking on cellular and humoral immunity, effects on cytokines, eicosanoid-mediated inflammation, antioxidant and oxygen free radicals, endogenous glucocorticoids, colonic mucus, mucosal blood flow, thrombosis, gut permeability and motility.[6] Recent research on microbiota changes with smoking may help explain the influence of smoking on IBD. The effects of smoking have been assessed in animal models and in clinical subjects. Animal studies, most frequently in mouse models of inflammation, may be criticised by their shorter exposure to smoke in comparison to chronic exposure in human smokers. Clinical observations of the temporal association between a change in smoking status with the development- or change in the clinical disease activity of IBD suggests that the pathogenic mechanism can be quite rapid in onset.

Animal and Cell Culture Studies. Nicotine is considered to be the active moiety of cigarettes and has been the main focus of animal and clinical studies.[55] The metabolites of nicotine may also be responsible for the anti-inflammatory effects of smoking. Animal models have demonstrated differential small bowel and colonic responses to cigarette smoking. Chronic nicotine exposure in rats decreased prostaglandin (PG) E2, increased nitric oxide synthetase (NOS) activity and there were no effects on the circulation of the jejunum. In the colon, however, there was increased microcirculation, but no effects on NOS and PGE2.[56] Cytokine effects also differ with chronic nicotine exposure in normal rats increasing jejunal interleukin (IL)-6 and IL-10 levels, whereas in the colon, there is decreased IL-2 without changes to IL-6 and IL-10.[57] These differential effects on the small bowel and colon may explain the divergent effects of smoking on CD and UC. No changes in the expression of the proinflammatory cytokines tumour necrosis factor (TNF)-α, IL-1β, TNF-γ, IL-6 were detected with exposure to cigarette smoke. IL-10 expression, however, significantly increased with smoke exposure.[58,59]

Studies have demonstrated that cell apoptosis may be either decreased[60] or increased[58,61] following nicotine exposure. In mice models, chronic smoke exposure induced the apoptotic index in the follicle-associated epithelium and promoted immune cell accumulation in the Peyer's patches. These changes were accompanied by upregulation of CCL9 and CCL20 mRNA, which was followed by increased dendritic cell recruitment. Cigarette smoke, however, had no effect on the villous epithelium of the ileum.[57]

Carbon monoxide (CO) also has anti-inflammatory properties. Inhalation of CO ameliorated TNBS-induced colonic injury and inflammation in mice. The proposed mechanisms of action include the heme oxygenase (HO-)1-dependent pathway and inhibition of TNF-α expression.[62–64] HO-1 augments the anti-inflammatory effects of IL-10 and also has cytoprotective antioxidant properties.[64] Inhibition of neutrophil aggregation also may underlie the anti-inflammatory effects of CO.[65] CO also increases regulatory T cells, which inhibit the T-helper 1 (Th1) cytokine response.[66] Due to its potential toxicity, CO has not been trialled in human subjects.

More recent attention has focused on the role of the effects of smoking on the gut microbiota.[67] Disruption of the host-microbiome interaction may result in loss of tolerance to commensal organisms residing in the intestinal tract and result in the development of various immune-mediated diseases, including IBD. Significant changes in gut microbiota occurred with side-stream cigarette smoke exposure in mice. Smoke exposure increased intestinal bacteria, especially Clostridium, but decreased Fermicutes (Lactococci and Ruminococcus), Enterobacteriaceae family and segmented filamentous bacteria compared to controls.[59] Smoking may also influence IBD through effects on intestinal mucosal barrier integrity and cell signalling pathways. Mouse colonic tight junction proteins, claudin 3 and zonula occludens (ZO)-2 are up-regulated and c-Jun N-terminal kinase and p38 mitogen-activated protein (MAP) kinase signalling are increased, whereas AMP-activated protein kinase are decreased.[59]

Human Studies. In humans, cytokine profiles varied following nicotine exposure. Lipopolysaccharide (LPS; 1 microg/mL) or phytohemagglutinin (PHA, 5 or 0.5 microg/mL)-stimulated production of IL-1β, IL-10, transforming growth factor (TGF)-β and TNF-α (P < 0.001), were reduced following exposure to nicotine (1, 10, 100 μg/mL), suggesting possible direct inhibitory effects on the innate immune response.[60] Mononuclear cells from CD smokers are functionally impaired compared to controls. Smokers with CD had inhibited LPS-induced IL-8 release compared with nonsmokers, but this was not found in healthy or UC smokers. Interestingly, nicotine was not responsible for this effect and nicotine replacement did not abrogate this effect.[68] Another study found that smoking protected against the development of granulomas in CD (OR: 0.16; 95% CI: 0.04–0.59).[69] These studies suggest that smoking may exert direct effects on mononuclear cells in CD. Aldhous et al. found that independent of disease activity and smoking chronicity, nicotine significantly decreased cellular proliferation and increased resting cells and these changes are mediated through IL-12/IL-23p40 and apoptosis. In CD, nicotine mildly reduced apoptosis in response to PHA or in unstimulated cells.[70] Nicotine may act on the MAP kinase family and affect TNF-α through inhibition of nuclear factor kappa B signalling.[2]

Loss of the intestinal mucosal barrier function may underlie the pathogenesis of IBD. Smoking has been proposed to help maintain normal intestinal permeability.[71] Indomethacin, a nonsteroidal inflammatory drug, increases intestinal permeability, but this effect is abrogated in smokers.[72] Another study on UC subjects, however, did not show significant improvement in intestinal permeability in smokers.[73]

Nicotine may influence inflammation in IBD through direct ligand-receptor binding of the nicotinic acetylcholine receptors (nAChR) of the small bowel. In UC, there is a depletion of the alpha-3 subunits of the nAChR, but the significance of this remains currently uncertain.[74] Nicotine also acts centrally through modification of the stress response on the intestinal tract.[2]

Disruption of the gastrointestinal microbiota has been identified in smokers with active CD.[75] Measuring the percentage of bacteria against the total microbiota load, smokers with or without CD had significantly increased Bacteroides-Prevotella in their faecal samples compared to nonsmokers (mean 38.8% in smokers vs. 28.3% in nonsmokers, P < 0.001). This study utilised 16sRNA fluorescent in situ hybridisation to detect the bacteria of interest. CD patients with ileal location of disease had significantly lower Faecalibacterium prausnitzii. The proportion of microbiota was not significantly different between age categories, sex or race, and was not affected by CD activity index, ileal resection, colonic involvement, C-reactive protein levels, current use of steroids or immunomodulators or disease duration.[75] As similar changes in intestinal microbiota were found in both non-IBD smokers as well as CD smokers, the dysbiosis appeared to be directly attributable to smoking, and therefore changes in the microbiota were unlikely to be simply secondary to the inflammation. Also, Bacteroides species has been previously shown to have proinflammatory effects.[76] There remains no accepted unifying pathogenic mechanism between smoking and the effects on IBD, but the microbial hypothesis may be the underlying key factor that results in the various intestinal cytokine and cellular changes. This hypothesis, however, does not explain the effects of smoke on in vitro studies, or the differences in response between CD and UC subjects and suggest that direct and indirect mechanisms may be responsible for the effects of smoking on IBD. The differential response of smoking on CD and UC also remains unexplained, but might be due to opposing effects of smoking on the small intestine and the large intestine.

Clinical Studies

Although nicotine and smoking are not synonymous, there has been substantial investigation into the potential of using nicotine as a therapy to UC. Summarised in Table 1 is the use of nicotine therapy in the induction of remission in UC and Table 2 the maintenance of remission in UC.

Nicotine as a Treatment in Ulcerative Colitis. Studies conducted to date have investigated the efficacy of nicotine therapy compared to placebo, as well as nicotine therapy compared to oral corticosteroids or mesalazine, in the induction and maintenance of remission in patients with UC.[3,20] Differing routes of administration have been used, including transdermal nicotine patches, nicotine chewing gum and a nicotine-based enema.[3]

Nicotine vs. Placebo Patch. Three randomised controlled trials (RCTs), Pullan et al.,[77] Sandborn et al.[78] and Thomas et al.[79] examined transdermal nicotine therapy compared to a placebo patch. Pullan et al.[77] conducted a randomised, double-blind study in which 72 patients with active UC were treated with either nicotine patch or a placebo patch for 6 weeks. Incremental doses of nicotine were given up to a tolerated maximum of 25 mg per day with most subjected tolerated 15–25 mg per day. All patients had been previously prescribed mesalazine and continued on their individual medication regimen throughout the study. Twelve patients were also on low-dose oral corticosteroids, which were also continued throughout the study. Clinical, sigmoidoscopic and histological assessment of each patient was conducted at the commencement and conclusion of the study. The complete remission rate in the treatment arm was 49% compared with 24% in the placebo group (P = 0.03). The nicotine group also had better improvement in global score of the severity of their colitis (P < 0.001) and the more objective histologic grade (P = 0.03). However, the nicotine group had increased side effects (P = 0.002) most commonly light-headedness, headache and sleep disturbance.[77]

Sandborn et al.[78] similarly compared nicotine with placebo patches in a randomised, double-blind trial incorporating 64 patients who were nonsmokers. Patients were stratified based on their smoking history and current medical regimen, then randomly assigned to the nicotine or placebo group. Nicotine therapy started at 11 mg per day and increased to a tolerated maximum of not more than 22 mg per day. Disease extent was measured at baseline and at the end of the 4-week study, by clinical and sigmoidoscopic assessment, as well as patient diary of symptoms. In the nicotine group, 39% experienced clinical improvement in their colitis, compared with only 9% of patients in the placebo group (P = 0.007). Nicotine-associated side effects were acute pancreatitis, headache and nausea.[78] The study concluded that nicotine was efficacious in the control of mild-to-moderate colitis manifestations.

McGrath et al.[20] included both these RCTs in a meta-analysis and concluded that the use of transdermal nicotine is superior to placebo for induction of remission in patients with UC. Nicotine patch was found to have an OR of 2.56 (95% CI: 1.02–6.45) for induction of remission and an OR of 2.72 (95% CI: 1.28–5.81) for clinical improvement and/or remission.[20] The absolute risk reduction in each case was calculated to be 0.13 (95% CI: 0.01–0.25) and 0.20 (95% CI: 0.06–0.35) respectively.[20] Therefore, the number needed to treat (NNT) with transdermal nicotine patches compared with an equivalent placebo, in addition to the standard therapy of oral mesalazine and corticosteroids, was calculated to be 8 for remission and 5 for clinical improvement.[20]

Thomas et al.[79] instead examined transdermal nicotine patches against placebo patches with regard to the maintenance of remission in UC. In a randomised, double-blind study, 80 patients who were in remission from UC were treated with either nicotine or placebo patch and followed for 6 months. Those randomised to the nicotine group were given incremental doses of nicotine for the first 3 weeks until a maintenance dose was reached, at which point this dose was continued for the duration of the study. Most patients tolerated 15 mg per day. All patients entering the study were initially on a mesalazine preparation, which was stopped once the maintenance dose was achieved. Clinical, sigmoidoscopic and histological assessments were made at the beginning and end of the study, or at relapse, and the patient kept a diary of symptoms throughout. Of 40 patients, 22 in the nicotine group were prematurely withdrawn from the study, 14 due to relapse and 8 for other reasons, including side effects, patient request or protocol violation. In the placebo group, 20 patients were withdrawn before 6 months, 17 because of relapse and 3 for other reasons. Side effects were common, with nausea, light-headedness and itching being the most prevalently cited. The authors concluded that transdermal nicotine alone was no more efficacious at maintenance of remission in UC than placebo. Interestingly, however, serum concentrations of nicotine and cotinine were lower than expected throughout the study in patients of the nicotine group, suggesting poor compliance to therapy.[79]

Nicotine Patch vs. Conventional Therapy. Three further RCTs compared transdermal nicotine to oral corticosteroids in the induction and maintenance of remission in UC.[80–82] Guslandi and Tittobello[80] compared 15 mg per day nicotine patches with a tapering dose of oral prednisone starting at 30 mg per day in mild-to-moderate left-sided UC. The randomised study followed the first 15 patients from each group to achieve clinical remission for a further 6 months, directly following the 5-week course of nicotine or prednisone therapy. Patient recruitment to the study continued until the goal of 15 patients in remission from each group was achieved. All patients received 1 g oral mesalazine twice a day throughout the study period. Clinical, sigmoidoscopic and histological examination was performed through to 6 months. Prednisone was found to be more efficacious in inducing remission, as the nicotine group required 21 patient recruitments prior to 15 patients achieving remission, compared with the prednisone group's 17. It was found that 60% of the patients receiving prednisone, compared with 20% of the patients receiving nicotine had a relapse of active colitis in the 6 months following therapy (P = 0.027), and that relapses occurred earlier in those receiving prednisone.[80] The impression that nicotine-induced remission of UC was longer lasting than that obtained by oral corticosteroids led to a 12-month follow-up cross-over study comparing mesalazine plus transdermal nicotine with mesalazine plus oral prednisolone. The relapse rate for prednisolone was 93% and nicotine was 47% (P = 0.007).[81]

Thomas et al.[82] conducted a randomised, double-blind study of 6 weeks duration, including 61 patients, comparing the maximum tolerated dose of transdermal nicotine with most patients tolerating between 15 and 25 mg per day, to a 15 mg daily dose of prednisolone. Mesalazine was ceased on day 10 of the study. In those who completed the 6-week study, both the nicotine and prednisolone groups improved symptomatically and on sigmoidoscopic scores, with results slightly favouring the prednisolone group. In this regard, however, statistical significance was not reached. In the transdermal nicotine group, the sigmoidoscopic remission rate at 6 weeks was 32% compared with 58% in those receiving prednisolone (P = 0.08). Side effects were more frequent in the nicotine group and mostly from nausea, light-headedness and tremor.[82]

Nikfar et al.[3] found during meta-analysis of the two induction studies[80,82] that the Relative Risk (RR) for clinical remission with nicotine therapy compared to oral corticosteroids was 0.74 (95% CI: 0.5–1.09). Therefore, the efficacy of this nicotine therapy compared to the conventional therapy of oral corticosteroids could not be supported.[3]

A fourth study[83] examined transdermal nicotine vs. oral mesalamine in treating distal UC refractory to mesalamine enema alone. Thirty patients were randomised to 15 mg nicotine patch or 800 mg mesalamine orally, three times daily for 4 weeks. Patients underwent clinical and sigmoidoscopic assessment at baseline and at the conclusion of the study. The remission rate in the nicotine group was 80% compared with 33% in the oral mesalamine group (P = 0.027). The study concluded that it appeared the addition of transdermal nicotine to rectal mesalamine was superior to the combination of oral and rectal mesalamine in treating distal UC refractory to mesalamine enema alone.[83] However, the small sample size and single-blind set up confers a lack of power for this data and makes it difficult for its findings to be recommended for clinical practice without further trials incorporating more patients.[20]

Nicotine Enema vs. Placebo Enema. A pilot study used nicotine tartrate liquid enemas in patients with left-sided UC unresponsive to conventional first-line therapy and found 71% to show clinical and endoscopic improvement after 4 weeks. The enema was prepared using 60 ml of sterile water mixed with 500 mg medium viscosity carboxymethylcellulose, 5 g of sorbitol, 5.23 g sodium phosphate and 0.05 g monobasic sodium phosphate yielding an iso-osmolar, buffered solution of pH 8.5. Nicotine tartrate was then added, 8.552 mg or 17.108 mg, to give a 3 mg or 6mg nicotine enema, respectively.[84]

A RCT[85] of active distal UC randomised 104 patients to either 6 mg nicotine enemas or placebo, whilst continuing on their regular medication regimen. Remission or improvement in disease was measured by clinical, sigmoidoscopic and histological means at 4 weeks. Only 27% of the treatment group achieved remission, compared with 33% of the placebo group (P = 0.55).[85] Ingram et al.[85] concluded that, although well tolerated, 6 mg nicotine enemas were not clinically efficacious in inducing remission of UC.

Oral Nicotine. Orally administered nicotine has yet to be studied in a controlled, randomised, double-blinded fashion, or with adequate numbers of participants for statistical analysis. Results from uncontrolled trials and one randomised, controlled cross-over study have yielded conflicting results.[3] Safety concerns over the links between nicotine and oropharyngeal and oesophageal cancers may have limited further investigations of this route of administration.