ORIGINAL ARTICLE

Does Smoking Influence Crohn’s Disease in the Biologic Era?The Tabacrohn StudyTiago Nunes, MD,1† Maria Josefina Etchevers, MD,1† Olga Merino, MD, PhD,2 Sonia Gallego, MD,3

Valle Garcı́a-S�anchez, MD, PhD,4 Ignacio Marı́n-Jim�enez, MD, PhD,5 Luis Mench�en, MD,5

Manuel Barreiro-de Acosta, MD, PhD,6 Guillermo Bastida, MD,7† Sara Garcı́a, MD,8 Elena Gento, MD,8

Daniel Ginard, MD, PhD,9 Fernando Gomoll�on, MD, PhD,10 Maite Arroyo, MD, PhD,10 David Monfort, MD,11

Esther Garcı́a-Planella, MD,12 Benito Gonzalez, MD,13 Carme Loras, MD, PhD,14 Carles Agustı́, MD,1†

Carolina Figueroa, MD,1† and Miquel Sans, MD, PhD1† for the TABACROHN Study Group of GETECCU, SpanishWorking Group in Crohn’s Disease and Ulcerative Colitis

Background: While most studies have found a negative effect of smoking on Crohn’s disease (CD) phenotype, more recent data have failed to

reproduce this association, which might be due to a current wider use of thiopurines and biologic therapy. The TABACROHN study aimed at

defining the impact of smoking on CD in the largest published series.

Methods: This multicenter cross-sectional study included 1170 CD patients. Patients were classified as nonsmokers, current smokers, or former

smokers according to their present smoking status. Clinical data regarding disease characteristics, treatment, and complications were collected.

Results: Smokers were more frequently under maintenance treatment when compared to nonsmokers. In addition, current smokers presented

higher use of biologic drugs compared to nonsmokers. Tobacco exposure and a higher tobacco load were independent predictors of need for

maintenance treatment and stenosing phenotype, respectively.

Conclusions: In the era of early and widespread use of immunosuppressants and biologics, tobacco exposure is an independent predictor of

need for maintenance treatment, specifically biologic therapy. The wider use of biologics and immunosuppressants could account for the exis-

tence of no major differences in disease behavior and complications between nonsmokers and current smokers.

(Inflamm Bowel Dis 2012;000:000–000)

Key Words: inflammatory bowel disease, Crohn’s disease, environmental factors, smoking, tobacco, disease phenotype

A ccording to our present understanding, Crohn’s disease

(CD) develops in genetically susceptible individuals

exposed to a variety of environmental factors. Although the

intricate interaction between genes and the environment

has not been completely unveiled, our knowledge of CD

pathogenesis has greatly evolved in the past two decades.1,2

One of the most studied environmental factors influencing

CD natural history is tobacco smoking.3,4 In 1984, Somer-

ville et al5 described for the first time that CD patients

were more likely to be smokers than control subjects, an

observation subsequently confirmed by others.6 Later on,

two large meta-analyses corroborated this association,

establishing tobacco exposure as the most robust evidence-

based risk factor to develop CD.7,8

Tobacco includes a countless number of potentially

detrimental substances, yet the exact mechanism by which

smoking contributes to the complex CD pathogenesis

remains unclear. Certain tobacco-derived products are

able to act on the intestinal barrier function, affecting the

superficial mucus layer, bowel epithelium, macrophage

function, and cytokine production.9–11 Since important

genes associated with CD are linked to autophagy and

immune response to bacteria,12 the most appealing pro-

posed mechanism for the role of tobacco in facilitating

CD development could be its impact on macrophage func-

tion.13 In that regard, in smokers macrophages present a

Additional Supporting Information may be found in the online version of this

article.

Received for publication February 24, 2012; Accepted February 28, 2012.

From the 1Hospital Clinic/IDIBAPS, Barcelona, 2Hospital de Cruces,

Barakaldo, 3Hospital Miguel Servet, Zaragoza, 4Hospital Reina Sofia, C�ordoba,5Hospital Gregorio Mara~n�on, Madrid, 6Hospital Clinico Universitario, Santiago

de Compostela, 7Hospital La Fe, Valencia, 8Hospital general Yague, Burgos,9Hospital Universitario Son Espases, Palma de Mallorca, 10Hospital Clinico,

Zaragoza, 11Consorci Sanitari de Terrassa, Terrassa, 12Hospital Sant Pau,

Barcelona, 13Hospital Juan Canalejo, A Coru~na, 14Hospital Mutua de Terrassa,

Terrassa, Spain.

The first two authors contributed equally to this work.†CIBEREHD: Centro de Investigaci�on Biom�edica en Red de Enfermedades

Hep�aticas y Digestivas.

Reprints: Miquel Sans, MD, PhD, Department of Gastroenterology, Hospital

Clı́nic i Provincial / IDIBAPS, 170 Villarroel, 08036 Barcelona, Spain (e-mail:

msans@clinic.ub.es).

Copyright VC 2012 Crohn’s & Colitis Foundation of America, Inc.

DOI 10.1002/ibd.22959

Published online in Wiley Online Library (wileyonlinelibrary.

com).

Inflamm Bowel Dis 1

selective functional deficiency in effectively killing intra-

cellular bacteria leading to an impaired innate immune

response.14

Several studies have investigated the impact of smok-

ing on CD clinical course and response to CD therapies.

Most of the available data support the notion that smoking

negatively impacts disease phenotype, resulting in a higher

prevalence of stenosing and penetrating behavior, perianal

disease, need for surgery, and postsurgical recurrence in

smokers.15–30 Also described is a higher incidence of ileal

disease and a lower prevalence of colonic involvement

among smokers.21,25,31–33 In relation to response to CD

therapies, an increased need for steroids and immunosup-

pressive treatment and a decreased clinical response to

infliximab have been linked to active smoking.17,34–36 In

addition, greater tobacco consumption has also been associ-

ated with a more complicated CD phenotype, suggesting a

dose-dependent impact.33 In that regard, it has been

observed that heavy smokers are more likely to present

structuring or penetrating phenotypes and less colonic

involvement compared to light smokers.17,33

Although the perception that smoking significantly

impacts the CD clinical course is widely accepted, a care-

ful review of the available literature does not fully support

this concept.37 In a large study aiming at evaluating CD

behavior throughout time, Cosnes et al38 found no effect

of smoking on disease progression to stenosing or pene-

trating phenotypes. Likewise, evidence showing no associ-

ation between smoking and CD location is also avail-

able.5,6,17,39 Additionally, a recent study specifically

aimed at defining the impact of smoking on CD phenotype

failed to find a direct association between tobacco and

severe phenotype or need for surgery.40 The discrepancy

between the most recent studies and previous ones might

suggest that the detrimental effect of smoking has been

overcome by the wider use of more potent drugs, such as

immunosuppressants and anti-tumor necrosis factor alpha

(TNFa) agents. To test this hypothesis we carried out a

multicenter study in which the influence of smoking on

CD clinical course was evaluated in the largest study

undertaken so far.

MATERIALS AND METHODS

Study DesignTo ascertain whether the smoking habit influences CD

clinical course we undertook an observational cross-sectional

study in a large population of CD patients. The Tabacrohn

Study was designed and supported by the Spanish Working

Group in Crohn’s Disease and Ulcerative Colitis (GETECCU)

with the participation of 14 Spanish referral IBD units. During

a 6-month inclusion period, 1217 consecutive patients with an

established diagnosis of CD confirmed by standard radiologi-

cal, endoscopic, and histological features were enrolled. Data

were collected from clinical records including demographics

and previous and past smoking habits. Disease location and

behavior were categorized according to the Montreal Classifi-

cation.41 Length of disease duration, current and previous

treatments, perianal disease, and past bowel resections were

also captured.

Smoking Habit DefinitionsThe definition of smoking was that of Cosnes et al,18

who defined smoking as the consumption of more than 7 ciga-

rettes per week for at least 6 months along with a report of

any cigarette smoking in the previous 6 months. Nonsmokers

were those who never smoked, had stopped smoking more

than 12 months before CD diagnosis, or smoked fewer than 7

cigarettes weekly. Former smokers were patients who were

active smokers but had quit at least 12 months before CD di-

agnosis or thereafter. The lifetime tobacco load was calculated

and measured in pack years where the number of packs/day

(assuming 20 cigarettes per pack) was multiplied by the num-

ber of years smoking. Current and former smokers were fur-

ther classified in relation to the median tobacco load (12 pack

years) as heavy smokers (�12 pack years) and light smokers

(<12 pack years).

Ethical IssuesAll data was anonymously analyzed to preserve patient

confidentiality. The study was approved by the Ethics Com-

mittee of the Hospital Clinic Provincial of Barcelona, Spain.

Statistical AnalysisQualitative variables were expressed using frequencies.

Continuous variables were expressed using median and inter-

quartile range. One-way analysis of variance (ANOVA) analy-

sis was performed for continuous variables and chi-square

tests were used to compare qualitative variables among

groups. Subsequent additional pairwise comparisons were per-

formed using Fisher’s or chi-square tests for categorical varia-

bles and Student’s t-tests for quantitative variables when nec-

essary. Given the impact of disease duration on CD behavior

and complications, all univariate results were adjusted by this

variable. All tests were two-tailed with a significance level set

at <0.05.

To determine the effect of risk factors (covariates) on

the prevalence of outcomes of interest regarding location,

behavior, and treatment, a logistic regression model was con-

structed. Multiplicative interaction between selected covariates

and exposure to tobacco were assessed. Candidate risk factors

were analyzed separately for each outcome by comparing

patients who presented the risk factor with those who did not.

For discrete variables, results were calculated and compared

by v2 or exact tests. Odds ratios, tests of significance, and

95% confidence intervals (CIs) were calculated. Student’s t-test (for normal data) or the Wilcoxon rank-sum test were

Inflamm Bowel DisNunes et al

2

used for continuous variables. Variables achieving a P value

of less than 0.1 were selected and subsequently included in

the multivariate analysis. The discrete variable ‘‘exposure to

tobacco’’ which included current and former smokers was

forced into all models even if it did not appear to be statisti-

cally significant in univariate analysis. Following adjustment

for all confounding variables, P < 0.05 was considered

significant.

RESULTS

Tabacrohn Study PopulationIn all, 1214 patients were included in the TABA-

CROHN study. Forty-four patients were subsequently

excluded due to lack of essential data regarding disease

phenotype or smoking status, leaving 1170 eligible patients

for further analysis. When smoking exposure was assessed,

there were 575 patients present or previously exposed to

tobacco (49%), including 420 current smokers and 155 for-

mer smokers (36% and 13% of the total study population,

respectively), and 595 nonsmokers (51%). Regarding the

lifetime tobacco load in the former and current smoker

population, 418 patients had complete data, allowing the

calculation of cigarette consumption with a median of 12

pack years. Additional information regarding gender, age,

disease duration, CD phenotype, and medical treatment is

summarized in Table 1.

Population Characteristics According to SmokingStatus

Age at Inclusion and at Disease Onset, Disease Du-ration, and Complications

No differences in age at study inclusion and age at

disease onset were observed between former, current, and

nonsmokers (Table 2). Former smokers presented longer

disease duration compared to the other groups. No differen-

ces were observed with respect to previous bowel resection

or prevalence of perianal disease (Table 2).

Disease Behavior and LocationNo differences in disease behavior between current

and nonsmokers were identified. Former smokers presented

less inflammatory disease than nonsmokers (Table 2). With

respect to disease location, ileocolonic disease was more

frequently observed in current smokers (Table 2).

TreatmentCurrent and former smokers were more frequently

under maintenance treatment than nonsmokers (Table 2).

In that regard, current smokers were more frequently under

biological treatment when compared to nonsmokers (Table

2).

Population Characteristics According to TobaccoLoad

Current and former smokers with �12 pack years of

tobacco load presented more stricturing and less inflamma-

tory CD phenotype than patients with <12 pack year (Ta-

ble 3). In addition, heavy smokers presented a trend toward

a higher prevalence of bowel resection when compared to

light smokers. Of note, patients with greater tobacco load

were more frequently male, older at disease onset and at

study inclusion, and presented longer disease duration

(Table 3).

Multivariate AnalysisMultiplicative interaction between selected covariates

assessed through logistic regression found that being

exposed to tobacco is an independent predictor of need for

TABLE 1. Population Characteristics RegardingDemographics, Disease Phenotype, Treatment, andComplications

Characteristics N¼1170 (%)

Age at inclusion

Median age 39

Interquartile range 31–49

Age at disease onset

Median age 30

Interquartile range 23–40

Disease duration

Median age 84

Interquartile range 41–140

Gender

Male 612 (52)

Location

Ileal 481 (41)

Colonic 228 (19)

Ileocolonic 420 (36)

Proximal 41 (4)

Behavior

Inflammatory 675 (58)

Stricturing 244 (21)

Penetrating 238 (21)

Treatment

No treatment 212 (18)

5-ASA 149 (13)

Thiopurine 494 (42)

Methotrexate 20 (2)

Biologics 144 (12)

Thiopurine þ Bio 139 (12)

Others 7 (1)

Perianal disease 322 (28)

Bowel resection 455 (39)

Inflamm Bowel Dis Smoking and CD Phenotype

3

maintenance treatment and, specifically, for biological ther-

apy. In addition, tobacco load, expressed as pack years,

was independently associated with stricturing disease

(Table 4).

DISCUSSIONThe influence of smoking on CD clinical course has

been the subject of intense research, but their results have

greatly varied with respect to nature and magnitude. While

in the last two decades a large body of evidence demon-

strated that smoking is associated with an increased risk of

developing complicated CD behavior, need for surgery,

and recurrence after surgical intervention,15–30 the most

recently published study, however, found no association

between smoking habit and the development of stricturing

or penetrating phenotypes, suggesting that tobacco does not

directly influence CD behavior or need for surgery.40

Two factors have probably contributed to the marked

heterogeneity of previously published studies. First, and

perhaps most important, the use of different tobacco expo-

sure definitions could partially account for this variability.

In that regard, in a recent literature review on the topic,

Mahid et al37 suggested the adoption of standardized smok-

ing exposure definitions in order to help increase homoge-

neity between studies to enable comparisons and future

meta-analyses. We have chosen the smoking status defini-

tion proposed by Cosnes et al18 because it takes into con-

sideration the temporal overlap between CD and the smok-

ing habit, classifying patients who quit smoking before

disease diagnosis as nonsmokers. Second, the lack of influ-

ence of tobacco in most recent studies may be due to the

current widespread use of immunosuppressant and biolog-

ics. In this respect, Cosnes et al17 have shown that the risk

for surgery is increased only in patients who smoked and

TABLE 2. Population Characteristics According to Smoking Status

Nonsmokers Current Smokers Former smokers Sig.

Age at inclusion

Median age 39 39 39 NS

Interquartile range 30–52 32–48 33–46

Age at disease onset

Median age 30 30 29 NS

Interquartile range 23–43 23–39 24–35

Disease duration

Median age 77 85 108 P < 0.05a

Interquartile range 36–132 37–135 65–168

Gender

Male 320 (54%) 210 (50%) 82 (53%) NS

Behavior

Inflammatory 361 (62%) 241 (58%) 73 (47%) P < 0.05b

Stricturing 110 (19%) 93 (22%) 41 (27%) NS

Penetrating 115 (19%) 83 (20%) 40 (26%) NS

Location

Ileal 255 (43%) 156 (37%) 70 (45%) NS

Colonic 125 (21%) 72 (17%) 31 (20%) NS

Ileocolonic 201 (34%) 173 (41%) 46 (30%) P < 0.05c

Proximal 14 (2%) 19 (5%) 8 (5%) NS

Treatment

No treatment 122 (21%) 61 (15%) 29 (19%) P < 0.05d

Thiopurine 314 (53%) 233 (56%) 86 (56%) NS

Biologic 126 (21%) 117 (28%) 40 (26%) P < 0.05e

Perianal disease 159 (27%) 113 (27%) 50 (32%) NS

Bowel resection 222 (37%) 165 (39%) 68 (44%) NS

aFormer smokers present longer disease duration compared to nonsmokers and current smokers.bFormer smokers present less inflammatory phenotype compared to nonsmokers.cCurrent smokers present more ileocolonic disease compared to nonsmokers and former smokers.dNonsmokers were more frequently found under no treatment compared to current smokers.eCurrent smokers presented more use of biological compared to nonsmokers.

Inflamm Bowel DisNunes et al

4

were not under immunosuppressive therapy. In addition,

Domenech et al42 more recently suggested that smoking

did not appear to influence the development of penetrating

or stricturing phenotypes in a population under thiopurine

therapy.

In keeping with these findings, our results suggest

that there are no major differences in disease behavior

and complications (intestinal resections and perianal dis-

ease) between nonsmokers and current smokers. Of note,

the lack of association between current smoking and

complicated disease was observed in a population in

which smokers in general and, specifically, current smok-

ers were more frequently under maintenance treatment,

including biologic therapy. In accordance with the uni-

variate analysis, the logistic regression indicated that the

odds for need for maintenance treatment, in particular

for biologic therapy, are significantly higher among

patients exposed to tobacco. Our results support the

notion that the advent of more effective therapies might

have undermined the negative impact of tobacco on CD,

leading to recent negative studies in which smoking does

not seem to play an important role in the development

of a more complicated clinical course. Conversely, there

is no straightforward explanation for the significantly

lower prevalence of inflammatory phenotype in former

smokers when compared to nonsmokers after adjusting

for disease duration.

TABLE 3. Population Characteristics According to theLifetime Tobacco Exposure

Lightsmokers

Heavysmokers Sig.

Age at inclusion

Median age 33 46 P < 0.05

Interquartile range 28–39 40–53

Age at disease onset

Median age 26 35 P < 0.05

Interquartile range 21–31 29–45

Disease duration

Median age 84 120 P < 0.05

Interquartile range 40–120 51–166

Gender

Male 92 (45%) 123 (58%) P < 0.05

Behavior

Inflammatory 128 (62%) 98 (46%) P < 0.05

Stricturing 35 (17%) 64 (30%) P < 0.05

Penetrating 42 (21%) 50 (24%) NS

Location

Ileal 70 (34%) 87 (41%) NS

Colonic 35 (17%) 39 (18%) NS

Ileocolonic 88 (43%) 76 (36%) NS

Proximal 12 (6%) 11 (5%) NS

Treatment

No treatment 26 (13%) 41(19%) NS

Thiopurine 111 (54%) 112 (53%) NS

Biologic 53 (26%) 61 (29%) NS

Perianal disease 57 (28%) 61 (29%) NS

Bowel resection 67 (33%) 102 (48%) P ¼ 0.06

The lifetime tobacco exposure was classified according to the mediantobacco load in pack years as light smokers (<12 pack year) and heavysmokers (�12 pack year). Pairwise comparisons were performed usingFisher or chi-square tests for categorical variables and Student’s t-tests forquantitative variables.

TABLE 4. Multivariate Analysis for Key Outcomes ofInterest Regarding Complications, Treatment, and DiseasePhenotype with Their Independently Associated Variables

OutcomeVariables Reaching

Significance*OddsRatio CI 95%

Bowel resection Disease duration 1.10 1.08–1.12

Stricturing disease 7.1 4.9–10.1

Penetrating disease 11 7.6–15.8

Isolated ileal disease 1.4 1.1–1.9

Thiopurine treatment 1.3 1.34–2.4

Tobacco exposure 0.9 0.7–1.3

Perianal disease Disease duration 1.04 1.02–1.06

Isolated colonic disease 4.3 2.9–6.5

Ileocolonic disease 3.3 2.3–4.6

Penetrating disease 2.2 1.5–3.1

Biologic treatment 3.0 2.2–4.2

Tobacco exposure 0.9 0.7–1.2

Any treatment Penetrating disease 1.6 1.1–2.6

Perianal disease 1.5 1.1–2.2

Tobacco exposure 1.4 1.1–1.9

Thiopurine Bowel resection 1.6 1.2–2.0

Tobacco exposure 1.0 0.9–1.3

Biologics Ileocolonic disease 1.9 1.4–2.4

Tobacco exposure 1.3 1.1–1.8

Inflammatory disease Isolated colonic disease 2.2 1.5–3.3

Bowel resection 0.1 0.8–0.2

Tobacco exposure 0.8 0.6–1.1

Stricturing disease Bowel resection 2.9 1.8–4.6

Tobacco load 1.8 1.1–3.0

Tobacco exposure 1.2 0.9–1.7

Penetrating Bowel resection 5.4 3.8–7.6

Perianal disease 2.0 1.5–2.8

Ileocolonic disease 1.4 1.1–1.9

Any treatment 1.6 1.1–2.6

Tobacco exposure 0.9 0.7–1.3

*The variable ‘‘tobacco exposure’’ was forced into all models regardless ofstatistical significance in the univariate analysis

Inflamm Bowel Dis Smoking and CD Phenotype

5

Interestingly, among smokers (current and former)

different accumulated tobacco loads were associated with

determined disease outcomes. In that regard, heavier smok-

ers presented less inflammatory disease with more strictur-

ing phenotype and a trend toward a higher prevalence of

bowel resection. Of note, the multivariate analysis also

indicated that a greater tobacco load is independently asso-

ciated with the development of stricturing disease. These

results have obvious implications in clinical practice, sug-

gesting that, although a complete smoking cessation must

be pursued in all CD patients, a reduction in the amount of

cigarettes smoked could also have clinical relevance.

Finally, the authors must acknowledge some limitations

of the present study. Due to our cross-sectional design, it is

impossible to precisely determine the time relationship

between smoking and the development of certain clinical out-

comes in order to perform a time-dependent analysis. An

exact account of smoking status at the moment of each com-

plication, prospectively, could further clarify the association

between tobacco and CD clinical course. In addition, some

drawbacks with respect to the self-report nature of the present

study also need to be mentioned. In this regard, it may be diffi-

cult for some patients to correctly recall or report their tobacco

consumption patterns, especially after long periods between

visits, when the smoking habit is irregular, or when they have

been told previously to quit smoking. Another limitation that

should be addressed in future studies is the lack of data regard-

ing the prevalence of other smoking-associated diseases and,

most important, second-hand smoking. The latter could func-

tion as a possible bias when evaluating the interaction between

smoking and the development of complicated disease.

In conclusion, in our study tobacco exposure was

found to be an independent predictor of the need for main-

tenance treatment and specifically for biologic therapy. At

the same time, nonsmokers and current smokers presented

no differences with respect to disease behavior, location,

and complications, which could imply that the detrimental

effects of smoking found in previous studies could have

been undermined by more potent therapies. Nevertheless,

among smokers, a higher tobacco load seems to negatively

impact disease behavior, mainly regarding the development

of stricturing disease. New, large, prospective studies are

necessary to determine the actual impact of smoking on

CD in the current era of early and widespread use of

immunosuppressants and biologic drugs.

ACKNOWLEDGMENTSThis work has been partially funded by fundaci�o

miarnau. Author contributions: Tiago Nunes, Maria Jose-

fina Etchevers, and Miquel Sans wrote the original article.

Tiago Nunes and Miquel Sans performed all statistical

analysis. Olga Merino, Sonia Gallego, Valle Garcia, Igna-

cio Marı́n, Luis Mench�en, Manuel Barreiro-de Acosta,

Guillermo Bastida, Sara Garcı́a, Elena Gento, Daniel

Ginard, Fernando Gomoll�on, Maite Arroyo, David Monfort,

Esther Garcı́a, Benito Gonzalez, Carme Loras, Carles

Agustı́, and Carolina Figueroa contributed to patient inclu-

sion, data collection, literature review, and original article

correction. All authors approved the final draft.

REFERENCES1. Fiocchi C. Inflammatory bowel disease: etiology and pathogenesis.

Gastroenterology. 1998;115:182–205.2. Scaldaferri F, Fiocchi C. Inflammatory bowel disease: progress and

current concepts of etiopathogenesis. J Dig Dis. 2007;8:171–178.3. Danese S, Sans M, Fiocchi C. Inflammatory bowel disease: the role of

environmental factors. Autoimmun Rev. 2004;3:394–400.4. Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease:

incidence, prevalence, and environmental influences. Gastroenterol-ogy. 2004;126:1504–1517.

5. Somerville KW, Logan RF, Edmond M, et al. Smoking and Crohn’sdisease. Br Med J (Clin Res Ed). 1984;289:954–956.

6. Tobin MV, Logan RF, Langman MJ, et al. Cigarette smoking andinflammatory bowel disease. Gastroenterology. 1987;93:316–321.

7. Calkins BM. A meta-analysis of the role of smoking in inflammatorybowel disease. Dig Dis Sci. 1989;34:1841–1854.

8. Mahid SS, Minor KS, Soto RE, et al. Smoking and inflammatorybowel disease: a meta-analysis. Mayo Clin Proc. 2006;81:1462–1471.

9. Johnson GJ, Cosnes J, Mansfield JC. Review article: smoking cessa-tion as primary therapy to modify the course of Crohn’s disease. Ali-ment Pharmacol Ther. 2005;21:921–931.

10. Cosnes J. Tobacco and IBD: relevance in the understanding of diseasemechanisms and clinical practice. Best Pract Res Clin Gastroenterol.2004;18:481–496.

11. McGilligan VE, Wallace JM, Heavey PM, et al. Hypothesis aboutmechanisms through which nicotine might exert its effect on the inter-dependence of inflammation and gut barrier function in ulcerative co-litis. Inflamm Bowel Dis. 2007;13:108–115.

12. Franke A, McGovern DP, Barrett JC, et al. Genome-wide meta-analy-sis increases to 71 the number of confirmed Crohn’s disease suscepti-bility loci. Nat Genet. 2010;42:1118–1125.

13. Carbonnel F, Jantchou P, Monnet E, et al. Environmental risk factorsin Crohn’s disease and ulcerative colitis: an update. GastroenterolClin Biol. 2009;33(Suppl 3):S145–S157.

14. King TE Jr, Savici D, Campbell PA. Phagocytosis and killing of Lis-teria monocytogenes by alveolar macrophages: smokers versus non-smokers. J Infect Dis. 1988;158:1309–1316.

15. Brant SR, Picco MF, Achkar JP, et al. Defining complex contributionsof NOD2/CARD15 gene mutations, age at onset, and tobacco use onCrohn’s disease phenotypes. Inflamm Bowel Dis. 2003;9:281–289.

16. Breuer-Katschinski BD, Hollander N, Goebell H. Effect of cigarettesmoking on the course of Crohn’s disease. Eur J Gastroenterol Hepa-tol. 1996;8:225–228.

17. Cosnes J, Carbonnel F, Beaugerie L, et al. Effects of cigarette smok-ing on the long-term course of Crohn’s disease. Gastroenterology.1996;110:424–431.

18. Cosnes J, Carbonnel F, Carrat F, et al. Effects of current and formercigarette smoking on the clinical course of Crohn’s disease. AlimentPharmacol Ther. 1999;13:1403–1411.

19. Cottone M, Rosselli M, Orlando A, et al. Smoking habits and recur-rence in Crohn’s disease. Gastroenterology. 1994;106:643–648.

20. Kane SV, Flicker M, Katz-Nelson F. Tobacco use is associated withaccelerated clinical recurrence of Crohn’s disease after surgicallyinduced remission. J Clin Gastroenterol. 2005;39:32–35.

21. Lindberg E, Jarnerot G, Huitfeldt B. Smoking in Crohn’s disease:effect on localisation and clinical course. Gut. 1992;33:779–782.

22. Louis E, Michel V, Hugot JP, et al. Early development of stricturingor penetrating pattern in Crohn’s disease is influenced by disease loca-tion, number of flares, and smoking but not by NOD2/CARD15 geno-type. Gut. 2003;52:552–557.

Inflamm Bowel DisNunes et al

6

23. Moskovitz D, McLeod RS, Greenberg GR, et al. Operative and envi-ronmental risk factors for recurrence of Crohn’s disease. Int J Colo-rectal Dis. 1999;14:224–226.

24. Reese GE, Nanidis T, Borysiewicz C, et al. The effect of smoking af-ter surgery for Crohn’s disease: a meta-analysis of observational stud-ies. Int J Colorectal Dis. 2008;23:1213–1221.

25. Russel MG, Volovics A, Schoon EJ, et al. Inflammatory bowel dis-ease: is there any relation between smoking status and disease presen-tation? European Collaborative IBD Study Group. Inflamm Bowel Dis.1998;4:182–186.

26. Sutherland LR, Ramcharan S, Bryant H, et al. Effect of cigarettesmoking on recurrence of Crohn’s disease. Gastroenterology. 1990;98(5 Pt 1):1123–1128.

27. Yamamoto T, Keighley MR. The association of cigarette smokingwith a high risk of recurrence after ileocolonic resection for ileocecalCrohn’s disease. Surg Today. 1999;29:579–580.

28. Yamamoto T, Keighley MR. Smoking and disease recurrence afteroperation for Crohn’s disease. Br J Surg. 2000;87:398–404.

29. Picco MF, Bayless TM. Tobacco consumption and disease durationare associated with fistulizing and stricturing behaviors in the first 8years of Crohn’s disease. Am J Gastroenterol. 2003;98:363–368.

30. Rocca G, Astegiano M, Sostegni R, et al. Effect of current smokingon the clinical presentation, medical therapy, and surgical therapy ofCrohn’s disease. Gastroenterology. 1997;112:A1074.

31. Bustamante M, Nos P, Hoyos M, et al. Relationship between smokingand colonic involvement in inflammatory bowel disease. Rev EspEnferm Dig. 1998;90:833–840.

32. Holdstock G, Savage D, Harman M, et al. Should patients withinflammatory bowel disease smoke? Br Med J (Clin Res Ed). 1984;288:362.

33. Aldhous MC, Drummond HE, Arnott DR, et al. Effect of smokinghabit and load on clinical presentation and progression in Crohn’s dis-ease and ulcerative colitis. BSG. 2005;Abstract A2.

34. Parsi MA. Does smoking decrease the response to infliximab inpatients with Crohn’s disease? Inflamm Bowel Dis. 2008;14(Suppl 2):S18–S19.

35. Parsi MA, Achkar JP, Richardson S, et al. Predictors of response toinfliximab in patients with Crohn’s disease. Gastroenterology. 2002;123:707–713.

36. Arnott ID, McNeill G, Satsangi J. An analysis of factors influencingshort-term and sustained response to infliximab treatment for Crohn’sdisease. Aliment Pharmacol Ther. 2003;17:1451–1457.

37. Mahid SS, Minor KS, Stevens PL, et al. The role of smoking inCrohn’s disease as defined by clinical variables. Dig Dis Sci. 2007;52:2897–2903.

38. Cosnes J, Cattan S, Blain A, et al. Long-term evolution of diseasebehavior of Crohn’s disease. Inflamm Bowel Dis. 2002;8:244–250.

39. Benoni C, Nilsson A. Smoking habits in patients with inflammatorybowel disease. Scand J Gastroenterol. 1984;19:824–830.

40. Aldhous MC, Drummond HE, Anderson N, et al. Does cigarettesmoking influence the phenotype of Crohn’s disease? Analysis usingthe Montreal classification. Am J Gastroenterol. 2007;102:577–588.

41. Silverberg MS, Satsangi J, Ahmad T, et al. Toward an integrated clini-cal, molecular and serological classification of inflammatory bowel dis-ease: report of a Working Party of the 2005 Montreal World Congressof Gastroenterology. Can J Gastroenterol. 2005;19(Suppl A):5–36.

42. Domenech E, Carrion S, Garcia-Planella E, et al. Smoking status andresponse to thiopurines in steroid-dependent inflammatory bowel dis-ease. Inflamm Bowel Dis. 2011;17:971–975.

Inflamm Bowel Dis Smoking and CD Phenotype

7