SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

Background

: Cigarette smoking is associated with a more severe course of Crohn’s disease, but individual factors determining this effect are poorly known and it is not clear whether smoking cessation is associated with an improvement in the disease activity.

Aim

: To assess the factors determining the harmful effect of smoking in individuals with Crohn’s disease.

Methods

: A total of 622 consecutive patients with Crohn’s disease and Crohn’s disease activity index <200 were enrolled in a prospective 12–18 month cohort study. Patients were classified as current smokers, former smokers, or non-smokers. Alcohol consumption, oral contraceptive use, body mass index, and blood lipid levels were also recorded. The main outcome measure was the rate of flare-up.

Results

: A total of 139 current smokers (46%) developed a flare-up, vs. 79 non-smokers (30%) and 13 former smokers (23%). The relative risk of flare-up adjusted for confounding factors was 1.35 (1.03–1.76) in current smokers. This risk was increased in patients with previously inactive disease and in those who had no colonic lesions. It became significant above a threshold of 15 cigarettes per day. Former smokers behaved like non-smokers. Obesity, dyslipidaemia, and alcohol consumption had no significant effect.

Conclusions

: Current smoking, particularly heavy smoking, markedly increases the risk of flare-up in Crohn’s disease. Former smokers have a risk similar to that of non-smokers.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

Cigarette smoking is clearly harmful in Crohn’s disease. More precisely, smoking is associated with a higher risk of developing the disease,1, 2 increased clinical activity,3, 4 lower quality of life,5 more frequent complications,6, 7 increased need for immunosuppressive drugs,8, 9 and faster recurrences after surgery.10[11]–12

Many questions remain unanswered, however. Factors determining the importance of the harmful effect of smoking in one individual are poorly defined. Retrospective studies have suggested that women and patients with small bowel involvement are particularly susceptible.6, 8, 10 The putative additive role of other atherogenic or thrombogenic risk factors, such as obesity or dyslipidaemia has not been studied. The effect of smoking seems to be dose-dependent, but the respective effects of the daily tobacco consumption and the total life exposure are not easy to separate.7, 8, 11 Importantly, although the risk of post-surgical recurrence may be decreased in former smokers, compared to current smokers, there is no clear evidence that former smokers have a day-by-day disease course similar to that of non-smokers.8, 11

To address these questions, we conducted a prospective longitudinal study in a large population of patients followed in a single centre.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

Patients

From February 1995 to January 1996, all consecutive patients with Crohn’s disease attending the out-patient clinic of our unit were evaluated for inclusion in the study. The inclusion criteria were an established diagnosis of Crohn’s disease, and little or no disease activity for 1 week prior to inclusion, as defined by the Crohn’s disease activity Index (CDAI).13, 14 Fifty-six patients with scores of 200 or more, and 32 who required urgent hospitalization, received appropriate therapy and were included after their CDAI had decreased to under 200.

Study design

The clinical course of the disease was compared over a period of 12–18 months in three groups of patients: current smokers, former smokers, and non-smokers, defined by their smoking status at baseline. At inclusion, patients had a physical examination and biological tests, and their CDAI was calculated. They were then followed-up regularly by their usual physician in our unit, with a visit every 3–6 months and instructions to telephone us in case of new symptoms. In the event of flare-up, patients were seen immediately, and CDAI was calculated retrospectively for the preceding week. The study period ended with the first visit after 12 months of follow-up, then its duration varied from 12 to 18 months. At this time, new clinical and biological tests were performed and the CDAI was again calculated. A small proportion of patients (12%) referred from another centre during the period of inclusion, were followed-up elsewhere. Data were collected in these patients by contacting their personal physician.

Evaluation criteria

The outcome was the flare-up rate and the time to flare-up. Flare-up was defined as a CDAI above 150 and an increase of 60 or more in CDAI compared with baseline. In patients with chronic perianal disease, flare-up was defined by development of an abscess or of a new fistula. Additional measures of evaluation included the need for stronger medical treatment (introduction of steroids or immunosuppressive drugs, change of immunosuppressive drug, course of enteral or parenteral nutrition), for surgery, or hospitalization, and the number of working days lost due to sickness.

Smoking habits

At inclusion, patients were asked about their smoking habits, using a fixed schedule questionnaire. They were classified as current smokers if they had smoked more than seven cigarettes per week for at least 6 months8, 10 and had reported any cigarette smoking in the previous 6 months. Former smokers were patients who were smoking at time of diagnosis of Crohn’s disease and had stopped smoking more than 6 months before inclusion. Non-smokers were those who never smoked, smoked a little occasionally, or had stopped smoking before diagnosis of Crohn’s disease.

Associated vascular risk factors

Patients were also asked about alcohol intake and the use of oral contraceptives. Patients who consumed 20 g of alcohol per day or more were considered as drinkers. Patients who were taking the pill were considered as oral contraceptive users. Detailed results concerning the effect of oral contraceptives in this series of patients were reported elsewhere.15 Obesity was defined as a body mass index over 25. In 477 patients, total cholesterol, HDL cholesterol, and triglycerides levels were determined once within the study period, during a period of clinical remission.

Characteristics of Crohn’s disease

The characteristics of Crohn’s disease were recorded according to the retrospective analysis of medical charts.8 The extent of previous intestinal resections in each individual was codified with the Post-surgical Handicap Index.16 The main events during the year preceding inclusion were noted, including flare-up episodes, chronic active disease, excisional surgery, and anoperineal complications. The disease was considered to be in medical remission if no such event had occurred during the year. It was also considered to be in surgical remission when a curative intestinal resection had been performed; in other cases it was considered as recently active. The location of the lesions was determined from the most recent small bowel X-ray images and the results of endoscopy.

Treatment policy

As maintenance treatment, we used aminosalicylates (sulphasalazine, olsalazine or mesalazine, 2–3 g daily) for asymptomatic or moderately active forms of the disease, and immunosuppressive drugs (azathioprine 2 mg/kg per day or intramuscular methotrexate 10–20 mg weekly) for severe forms. Flare-up episodes were treated with mesalazine (3–4 g daily) or prednisolone (1 mg/kg per day, progressively tapered after 4 weeks), according to their clinical severity. When steroid therapy failed, patients were given a 4-week course of enteral or parenteral nutrition. Metronidazole or ciprofloxacin were used in symptomatic anoperineal disease. Surgery was reserved for stenotic and extra-parietal complications, or intractable forms with limited anatomic damage.

Statistical analysis

Continuous data are expressed as means ± s.d., and differences between the two groups were tested for significance by one-way analysis of variance. Discrete data are given as percentages, and comparisons between the groups were made with the Pearson chi-square test; P-values of less than 0.05 were considered significant. Kaplan–Meier survival analysis was used to estimate the cumulative probability of developing a flare-up episode in the 12–18 months after inclusion. Survival curves were compared by means of a two-sided Log-rank test. A Cox proportional hazards regression model with a backward variable elimination procedure was used to assess the strength of the associations whilst controlling for possible confounding. All baseline variables suspected to be possible predictors of flare-up (age; gender; duration of disease; recent disease activity; delayed inclusion; previous intestinal surgery; CDAI value (< or > 150); small bowel lesions; colonic lesions; anoperineal lesions; 5-aminosalicylate; steroid or immunosuppressive therapy; current and former smoking status) were entered into the model. Interaction between smoking and other covariates was tested. The effect of alcohol use, obesity, and dyslipidaemia were tested in the same manner. Lastly, to evaluate the effect of the mean daily cigarette consumption, adjusted relative risks were calculated in different subgroups of current smokers defined by their daily cigarette consumption at baseline. All statistical tests were two-tailed. Calculations were performed using SAS statistical software.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

Baseline data

A total of 630 patients were considered for inclusion in the study. Eight were excluded for the following reasons: three patients were only seen during a flare-up episode and did not come back to our institution thereafter, two were living abroad and were unable to maintain further contact, and in the three remaining cases, the diagnosis of Crohn’s was not confirmed. Of the remaining 622 patients, 303 were classified as current smokers, 57 as former smokers, and 262 as non-smokers. The current smokers had been consuming cigarettes for 15 ± 8 years (11 ± 5 pack-years), and their daily tobacco consumption was 14 ± 9 cigarettes. The former smokers had consumed 15 ± 6 pack-years. The median interval between smoking cessation and inclusion was 46 months (range: 7–381). Among the non-smokers, 245 patients had never smoked and 17 had stopped smoking before the development of Crohn’s disease. The baseline characteristics of the three groups are shown in Table 1. Several significant differences were identified. Former smokers were older, with a disease of longer duration, which had been more often quiescent during the previous year. Conversely, when compared to the two other groups, more current smokers had had active disease during the previous year, had active intestinal lesions, and were taking steroids or immunosuppressive drugs.

Table 1.  . Baseline characteristics of Crohn’s disease patients according to smoking status Thumbnail image of

Outcome

The main events which occurred during the study period in the three groups are indicated in Table 2. Three patients died during the study period; one woman died suddenly at home, while she was on a tobacco weaning programme—no autopsy was performed. The other two deaths occurred in the non-smoker group. One of them was due to an acute attack of asthma, and the other was secondary to an endoscopic perforation of the colon. Current smokers were more likely than non-smokers to develop a flare-up during the study period (relative risk, 1.96 [95% CI, 1.37–2.82]). Accordingly, more current smokers than non-smokers needed stronger medication and were unable to work. However, no difference was observed regarding the need for excisional or anoperineal surgery, or for hospitalization. Former smokers behaved like non-smokers, with a disease significantly less severe than that of current smokers.

Table 2.  . Main events during the 1-year study period Thumbnail image of

Table 3 gives the data collected at the final visit. Compared to non-smokers, current smokers were more disabled, with a higher mean CDAI, even though more of them took steroids and immunosuppressive drugs. Former smokers exhibited the same features as non-smokers.

Table 3.  . Comparison of the two groups of patients at the end of the 1-year study period Thumbnail image of

Actuarial analysis

Figure 1 gives the actuarial rate of flare-up episodes in the three groups. The flare-up rate was significantly increased in current smokers compared to the two other groups, and was similar in former smokers and non-smokers. The regression model selected six factors predictive of flare-up: female gender; recently active disease; high CDAI; anoperineal lesions; steroid therapy; and current smoking status (Table 4). Current smoking therefore appeared as an independent factor of active disease. This effect was not significantly dependent on age, gender, duration of disease, CDAI, small bowel lesions, anoperineal lesions, steroid therapy, or immunosuppressive therapy. However, smoking interacted significantly with two covariates: recent disease activity (P=0.02) and presence of colonic lesions (< 0.01): the effect of smoking was more pronounced in patients in remission and in those who had no colonic lesions. Introduction of the interaction between smoking and recent disease activity as a covariate in the Cox model disclosed the actual effect of smoking (relative risk 2.40 [95% CI, 1.40–4.13]). Accordingly, the effect of smoking was not significant in patients whose disease was active during the previous year, whereas it was manifest in patients whose disease was in remission at baseline (Figure 2). This effect was particularly marked in patients in whom remission had been obtained surgically: the 1-year actuarial risk of flare-up was 6 ± 3% in non-smokers (n=48), 0% in former smokers (n=9), and 25 ± 7% in current smokers (n=36), respectively (P=0.02).

image

Figure 1. . Kaplan–Meier life-table analysis of the occurrence of a flare-up episode of Crohn’s disease in the study population, according to smoking status at baseline (former smokers vs. non-smokers, N.S.; current smokers vs. non-smokers, < 0.01; current smokers vs. former smokers, < 0.01).

Download figure to PowerPoint

Table 4.  . Predictors of flare-up according to multivariate Cox analysis Thumbnail image of
image

Figure 2. . Kaplan–Meier life-table analysis of the occurrence of a flare-up episode of Crohn’s disease in patients whose disease was in remission at baseline, according to smoking status (former smokers vs. non-smokers, N.S.; current smokers vs. non-smokers, < 0.01; current smokers vs. former smokers, < 0.01).

Download figure to PowerPoint

Effect of the daily cigarette consumption

The actuarial risk of flare-up was not significantly correlated to the daily cigarette consumption in the whole group of smokers. However, after taking into account the interaction between recent disease activity and smoking, a dose-dependent effect was observed, with a significantly increased risk of flare-up from a daily dose of 15 cigarettes (Figure 3).

image

Figure 3. . Relative risks of a flare-up episode according to the daily cigarette consumption at inclusion. The numbers indicate the number of patients at risk in each sub-group. Twenty-one smokers who had quitted less than 6 months before inclusion were classified as current smokers, 0 cigarettes per day; among them, 14 (67%) resumed smoking during the study period.

Download figure to PowerPoint

Effect of modification of smoking habits

During the study period, two patients started or resumed smoking and 27 smokers quitted. In addition, seven recent quitters at baseline (classified as smokers) remained abstinent, while 14 resumed smoking. Table 5 shows the time relationship between the occurrence of a flare-up and periods of smoking and non-smoking, respectively. The occurrence of flare-ups was more frequent while the patients were smoking than while they were not smoking (42% vs. 12%, < 0.002).

Table 5.   Modification of smoking habits during the study period and occurrence of a flare-up Thumbnail image of

Effects of other vascular risk factors

Other vascular risk factors in each group are indicated in Table 6. Alcohol intake was minimal in this population (median intake: 1 g per day), as most patients were occasional drinkers. Whereas smoking and alcohol consumption were significantly associated (P=0.006), disease activity was independent of alcohol use. Obesity was not associated with an increased risk of flare-up. Blood lipid levels were normal in most patients; high levels (total cholesteroverline> 230 mg/dL or triglycerides> 125 mg/dL) were not related to an increased risk of flare-up. None of these factors interacted with current smoking.

Table 6.  . Other vascular risk factors and mean lipid levels according to smoking status Thumbnail image of

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

This prospective study confirms the harmful effect of current smoking in an unselected large cohort of patients with Crohn’s disease. The effects of smoking were more marked in patients who had no colonic lesions and in those whose disease was in remission, particularly if they were smoking more than 15 cigarettes per day. Neither obesity nor dyslipidaemia increased the deleterious effect of smoking. Alcohol intake had no effect. More importantly, this study clearly showed that patients who had stopped smoking were doing as well as non-smokers.

The effect of smoking may only be apparent in some particular subgroups of patients with Crohn’s disease, therefore care was taken in this study to include all consecutive patients seen during a long enrolment period of 1 year. As some patients with chronic active disease never reach clinical remission, we chose to include patients with a CDAI of up to 199, comprising those with low 1-week disease activity. We believe that these criteria minimized selection bias and allowed us to obtain an accurate estimate of the real effects of smoking. Although our recruitment may have missed some asymptomatic patients who, as they had done well for several years, escaped medical surveillance, we found a relatively low proportion (37%) of patients who developed a flare-up episode during the 12–18-month study period. This figure fits well with the data of a cohort of unselected Crohn’s patients from Denmark demonstrating the existence of a remarkably constant annual proportion of 30% of the patients with a moderate to high activity state.17

Year-by-year analyses of changes from active disease to remission and from remission to active disease have shown that about 70% of patients maintain their state of activity during the following year.17 In our study, the main risk factor for flare-up was indeed the recent activity of the disease at baseline. Although this factor tended to outweigh the others, the effect of smoking was still apparent. In this regard, it should be noted that previous disease activity was in part correlated with smoking status. In our statistical analysis, we calculated the relative risk of flare-up for current smokers after adjustment for recent disease activity, but not for current smoking during the year preceding inclusion. Thus the actual effect of smoking, even adjusted, may be slightly underestimated. Compared to non-smokers, current smokers were in a worse condition at inclusion, during the year of follow-up, and at the end of the study. One could argue that patients with severe disease are more prone to smoke, and that smoking may be a consequence rather than a cause of severe disease. However, this interpretation does not agree with the fact that the noxious effect of smoking was pre-eminent in patients whose disease was in remission. In addition, our previous retrospective study showed that the effect of smoking was protracted and discernible during the long-term evolution, even though the activity of Crohn’s disease in one individual may vary over time, exhibiting cycles of 3–7 years’ duration.8, 17 The reality of the effect of smoking was demonstrated further in the present study by the trend for a dose–effect relationship. This result confirms data from retrospective long-term studies, which are compatible with a dose-dependent effect of cigarette consumption.8, 11 Finally, the present study showed that the harmful effect of smoking was not significantly modified by gender, age, duration of Crohn’s disease, presence of small bowel and anoperineal lesions, and current therapy. Only patients who had active colonic lesions seemed less sensitive to the effect of smoking, but further studies are needed to confirm this result. Thus at the present time, every patient with Crohn’s disease who smokes should be advised to stop smoking.

The effect of associated vascular factors could be questioned as Wakefield et al. demonstrated the important pathogenic role of multifocal infarction in Crohn’s disease.18 We did not find an increased prevalence of dyslipidaemia nor obesity in our cohort of patients, and, when present, these abnormalities were not associated with an increased risk of flare-up. Similarly, we have shown in patients from this series the absence of effect of oral contraceptive use.15 In addition, very few patients were drinking alcohol and no association could be demonstrated between alcohol consumption and either a benign or a severe disease course.

Finally, the most important finding of this study was the fact that former smokers were doing as well as non-smokers. Although these patients were older and had a disease of longer duration than patients from the two other groups, there is no evidence that these differences may explain such findings. In the present study, age and duration of the disease had no effect on the flare-up rate and in the literature there are no data supporting the view that the activity of Crohn’s disease may decrease significantly with time.17 Moreover, former smokers tended to have a particularly benign clinical course, and very few patients required intensification of medical therapy or excisional surgery. These results should be presented to current smokers in order to convince them to quit.

In conclusion, the effect of smoking is clearly harmful in Crohn’s disease, in both the short and the long-term, particularly in heavy smokers. On the other hand, this effect is not long-lasting and stops after smoking cessation. Thus, every effort should be made to encourage patients to stop smoking. The inactive phase of the disease, including the post-operative period after curative resection, is the time when the resulting benefit will be greatest.19

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References

The authors are indebted to the following physicians for their clinical care of study patients and for assistance in collecting this data: Hélène d’Almagne-Serrano, Philippe Baumer, Jérôme Bellanger, Jean-François Contou, Daniel Evard, Philippe Godeberge, Philippe Lamy, Yves Le Quintrec and Yann Ngô.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. References
  • 1
    Calkins BM. A meta-analysis of the role of smoking in inflammatory bowel disease. Dig Dis Sci 1989; 34: 1841 54.
  • 2
    Lindberg E, Tysk C, Andersson K, Jarnerot G. Smoking and inflammatory bowel disease, a case control study. Gut 1988; 29: 352 7.
  • 3
    Duffy LC, Zielezny MA, Marshall JR, et al. Cigarette smoking and risk of clinical relapse in patients with Crohn’s disease. Am J Prev Med 1990; 6: 161 6.
  • 4
    Timmer A, Sutherland LR, Martin F. The Canadian Mesalamine for Remission of Crohn’s Disease Study Group. Oral contraceptive use and smoking are risk factors for relapse in Crohn’s disease. Gastroenterology 1998; 114: 1143 50.
  • 5
    Russel MG, Nieman FH, Bergers JM, Stockbrugger RW. The South Limburg IBD Study Group. Cigarette smoking and quality of life in patients with inflammatory bowel disease. Eur J Gastroenterology 1996; 8: 1075 81.
  • 6
    Holdstock G, Savage D, Harman M, Wright R. Should patients with inflammatory bowel disease smoke? Br Med J 1984; 288: 362362.
  • 7
    Lindberg E, Jarnerot G, Huitfeldt B. Smoking in Crohn’s disease: effect on localisation and clinical course. Gut 1992; 33: 779 82.
  • 8
    Cosnes J, Carbonnel F, Beaugerie L, Le. Quintrec Y, Gendre JP. Effects of smoking on the long term course of Crohn’s disease. Gastroenterology 1996; 110: 424 31.
  • 9
    Russel MG, Volovics A, Van Schoon EJ, Wijlick EH, Logan RF, Shivananda S, Stockbrugger RW. Inflammatory bowel disease: is there any relation between smoking status and disease presentation? Inflamm Bowel Dis 1998; 4: 182 6.
  • 10
    Sutherland LR, Ramcharan S, Bryant H, Fick G. Effect of cigarette smoking on recurrence of Crohn’s disease. Gastroenterology 1990; 98: 1123 8.
  • 11
    Cottone M, Rosselli M, Orlando A, et al. Smoking habits and recurrence in Crohn’s disease. Gastroenterology 1994; 106: 643 8.
  • 12
    Breuer-Katschinski BD, Hollander N, Goebell H. Effect of cigarette smoking on the course of Crohn’s disease. Eur J Gastroenterol 1996; 8: 225 8.
  • 13
    Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol 1989; 24: 2 6.
  • 14
    Best WR, Becktel JM, Singleton JW. Rederived values of eight coefficients of the Crohn’s disease activity index (CDAI). Gastroenterology 1979; 77: 843 6.
  • 15
    Cosnes J, Carbonnel F, Carrat F, Beaugerie L, Gendre JP. Oral contraceptive use and the clinical course of Crohn’s disease. A prospective cohort study. Gut 1999; 45: 218 220.
  • 16
    Cosnes J, Parades V, Carbonnel F, et al. Classification of the sequelae of bowel resection for Crohn’s disease. Br J Surgery 1994; 81: 1627 31.
  • 17
    Munkolm P, Langholz E, Davidsen M, Binder V. Disease activity courses in a regional cohort of Crohn’s disease patients. Scand J Gastroenterol 1995; 30: 699 706.
  • 18
    Wakefield AJ, Sawyerr AN, Dhillon AP, et al. Pathogenesis of Crohn’s disease: multifocal gastrointestinal infarction. Lancet 1989; 2: 1057 62.
  • 19
    Simon JA, Solkowitz SN, Carmody TP, Browner WS. Smoking cessation after surgery. A randomized trial. Arch Intern Med 1997; 157: 1371 6.