Cancer risk in inflammatory bowel disease

AN ASSOCIATION BETWEEN cancer in different sites and inflammatory bowel disease (IBD) has been reported in many studies dating from the 1920s (1). Associations between every conceivable site and type have been proposed. There are, however, very few population-based studies with a size sufficiently large to give reliable risk estimates. In most instances the relationship between IBD and different malignancies has emanated from either small case series or follow-up studies of patient groups where the selection bias is a concern that most authors have not taken into consideration. It is also worth pointing out that between 20 and 30% of any population in western countries will be diagnosed with some sort of cancer during their life. Thus, to ensure reliable risk estimates there is a need for a control population, and such studies have to be performed outside referral centres where selection bias can be operating. In a population-based study of 3121 patients with ulcerative colitis and 1655 patients with Crohn’s disease the author’s group (2) confirmed an overall increased risk of cancer both among patients with ulcerative colitis – 202 observed cases of cancer compared with 142.1 expected (standardized incidence ratio [SIR] 1.6, 95% CI 1.4 to 1.8) – and with Crohn’s disease – 58 observed cases compared with 47.1 expected (SIR 1.2, 95% CI 0.9 to 1.6). Surveillance bias could be a concern in a cohort of patients with regular contacts with the health care system. HowSURGICAL ISSUES IN IBD

in different sites and inflammatory bowel disease (IBD) has been reported in many studies dating from the 1920s (1).Associations between every conceivable site and type have been proposed.There are, however, very few population-based studies with a size sufficiently large to give reliable risk estimates.In most instances the relationship between IBD and different malignancies has emanated from either small case series or follow-up studies of patient groups where the selection bias is a concern that most authors have not taken into consideration.It is also worth pointing out that between 20 and 30% of any population in western countries will be diagnosed with some sort of cancer during their life.Thus, to ensure reliable risk estimates there is a need for a control population, and such studies have to be performed outside referral centres where selection bias can be operating.
In a population-based study of 3121 patients with ulcerative colitis and 1655 patients with Crohn's disease the author's group (2) confirmed an overall increased risk of cancer both among patients with ulcerative colitis -202 observed cases of cancer compared with 142.1 expected (standardized incidence ratio [SIR] 1.6, 95% CI 1.4 to 1.8) -and with Crohn's disease -58 observed cases compared with 47.1 expected (SIR 1.2, 95% CI 0.9 to 1.6).
Surveillance bias could be a concern in a cohort of patients with regular contacts with the health care system.How-ever, the standardized mortality ratio (SMR) was also increased for malignant diseases.In the case of ulcerative colitis the SMR was 1.3 (95% CI 1.1 to 1.6) and 1.1 for Crohn's disease (95% CI 0.7 to 1.6).After excluding colorectal cancer the relative risk (RR) for cancer was close to 1.0 for both ulcerative colitis (SIR 1.0, 95% CI 0.9 to 1.2) and Crohn's disease (SIR 1.1, 95% CI 0.8 to 1.5).The risk estimate was almost the same: the SMR was 1.0 for ulcerative colitis (95% CI 0.7 to 1.5) and 0.9 for Crohn's disease (95% CI 0.8 to 1.3).Colorectal cancer therefore seems to be the major cause of the increased morbidity and mortality in malignancies in patients with IBD.There was a tendency towards an increased risk of cancer of the bile ducts among patients with extensive colitis (SIR 1.9, 95% CI 1.7 to 4.0), which is consistent with other studies (3,4), and an increased risk of cancer of the small intestine among patients with Crohn's disease (SIR 3.4, 95% CI 0.1 to 18.6), also similar to previous reports (5,6).The risk estimate in the study from the Uppsala Health Care Region, Sweden (2) is, however, lower than previously reported.One feasible explanation is that a combination of Crohn's disease and a bypassed intestine seems to carry the highest risk (5).This procedure is very rarely used in Sweden and Israel, where similar results have been reported (7).There was also an increased number of malignancies of the connective tissues (sarcomas) for both ulcerative colitis (SIR 4.0, 95% CI 1.0 to 10.2) and Crohn's disease (SIR 2.5, 95% CI 0.1 to 13.9), an association also previously reported (8).There was a decreased risk of cancer of the respiratory tract among patients with ulcerative colitis (SIR 0.5, 95% CI 0.2 to 1.0), giving further credence to nonsmoking status as a risk factor for ulcerative colitis (9).Surprisingly, the present author's group could not find any increased risk of cancer of the respiratory tracts among cases with Crohn's disease (SIR 0.8, 95% CI 0.2 to 0.4).This casts some doubt on smoking as a risk factor for Crohn's disease.Finally, the author and colleagues also found a statistically significant decreased risk of breast cancer among patients with extensive colitis (although the underlying reason for this remains unknown) (SIR 0.4, 95% CI 0.1 to 1.0), results similar to previous studies from New York (10) and Cleveland (3).
The risk of colorectal cancer among patients with Crohn's disease varies greatly in different studies, with risk estimates between 2.5 and 20.0 (6,(11)(12)(13).In the study from the Uppsala Health Care Region the author and co-workers ( 13) found an overall increased risk of 2.5 for patients with Crohn's disease (95% CI 1.3 to 4.3).There was no difference between females and males.However, in patients in whom the disease was confined to terminal ileum at diagnosis, the risk of colorectal cancer did not differ from that of the background population.On the other hand, patients with any colonic involvement at diagnosis had a substantially increased risk (SIR 5.6, 95% CI 2.1 to 12.2).
Age at diagnosis also seemed to affect the risk estimates substantially.In those with any colonic involvement and diagnosis before age 30 years, an SIR of 20.9 was observed (95% CI 6.8 to 48.7) compared with 2.2 (95% CI 0.6 to 5.7) among those diagnosed after age 30 years.This is consistent with results from a Mayo Clinic study (11) where an overall risk of 20.0 was found as 80% of their patients had colonic involvement at diagnosis.
pancolitis diagnosed before age 15 years will be diagnosed with a colorectal cancer before age 50 years.These findings suggest that a prophylactic proctocolectomy in this patient group might be an appropriate alternative compared with close surveillance in order to reduce the mortality in colorectal cancer, which is the main reason for the reduced long term relative survival in patients with ulcerative colitis (28).
It has been proposed that a patient with ulcerative colitis is at maximum risk of development of colorectal cancer at approximately age 50 years regardless of age at onset or duration of disease (22).In the Uppsala study, the author and colleagues ( 16) were unable to confirm that hypothesis.They found a consistently higher RR than for the normal population persisting among patients even after age 70 years, which seems to refute that hypothesis.
Extent of disease at diagnosis is also an independent risk factor of colorectal cancer (16).Patients with proctitis did not differ significantly in the risk of colorectal cancer from the background population (SIR 1.7, 95% CI 0.8 to 3.2) even after analyzing rectal cancer as a single entity.These findings imply that it is not the inflammation as such, but some additional exposure that is needed for a malignant transformation in this patient group.The increased risk of colorectal cancer among patients with left-sided colitis at diagnosis has been confirmed (21,27).In the analysis of the results from the Uppsala study (16), this increased risk was also found although the risk estimate was substantially lower (SIR 2.8, 95% CI 1.6 to 4.4) compared with patients with pancolitis at diagnosis (SIR 14.8, 95% CI 11.4 to 188.4).Moreover, the latency period seems to be 10 to 15 years longer compared with pancolitis before a patient with left-sided colitis is at an increased risk of colorectal cancer (Figure 1).Whether this increased risk is due to patients initially diagnosed with left-sided colitis in which the disease subsequently progressed to pancolitis, or the risk is present among patients with disease confined to the left colon remains unknown.
In one of the two studies in which no association was found between ulcerative colitis and colorectal cancer (14), the lack of association was probably due to the short follow-up -10 or fewer years after diagnosis.In the other study (15), however, the follow-up was substantially longer although the authors analyzed left-sided colitis and pancolitis as one entity which, at least partly, may explain the lack of association found.These authors suggest an alternative explanation: that modern treatment strategies using pharmacological compounds such as sulfasalazine could be one reason for the lack of any increased risk of colorectal cancer in their patient group.The same observation was also made in a study from the United Kingdom (29) of colonoscopic surveillance among patients with ulcerative colitis.These authors noted: "A review of the notes of our cancer patients suggests that virtually none of them were taking disease suppressive drugs, such as salazopyrin, regularly or at all".Moreover, a similar mechanism was proposed in the 1980s (30).To study whether additional risk factors or protective agents could be identified for colorectal cancer among patients with ulcerative colitis, Pinczowski et al (31) recently conducted a case-control study.Cases were found by a renewed follow-up of a previous study (16).
They identified 102 patients diagnosed with colorectal cancer after a previous diagnosis of ulcerative colitis.Two controls were identified for each case and were matched for sex, extent of disease at diagnosis and calendar year of ulcerative colitis diagnosis.Moreover, the controls had to be alive without colorectal cancer and with an intact or partly intact colon when the case's colorectal cancer was diagnosed.Six controls were excluded due to a proctocolectomy done before diagnosis of colorectal cancer in the case, leaving 196 controls.After retrieving the charts from cases and controls in a uniform manner, the extent to which pharmacotherapy and disease activity, among other things, had an impact on the risk of colorectal cancer was determined.
Disease activity imparted a protective effect when those with more than one exacerbation on average per year were compared with those with fewer than one.Those with a higher disease activity had a decreased risk (RR 0.70, 95% CI 0.44 to 1.14).Pharmacological therapy, especially sulfasalazine, imparted an even greater protective effect.A comparison of patients who had completed at least one three-month course of sulfasalazine medication with those who had not, showed a highly significant protective effect (RR 0.34, 95% CI 0.19 to 0.62).
The risk estimates for both disease activity and pharmacological therapy (sulfasalazine) prevailed and were only marginally changed in a multivariate analysis with mutual adjustment for the two factors.The results thus confirm the observations made in Denmark, the United Kingdom and Israel.One feasible explanation is, of course, the intervention of prostaglandin synthesis by salicylate compounds, which could inhibit various stages of the carcinogenic process (32), parallelling the reduced risk of sporadic colorectal cancer in acetylsalicylic acid users (33).

CONCLUSIONS
Colorectal cancer in patients with ulcerative colitis is the only malignancy that has any impact on the morbidity and mortality in cancer in patients with IBD.This increased risk can be reduced through pharmacological therapy.

Figure 1 )
Figure 1) Cumulative incidence of colorectal cancer for patients with pancolitis and left-sided colitis at diagnosis