Increased Proportion of Colorectal Cancer in Patients With Ulcerative Colitis Undergoing Surgery in the Netherlands

Abstract

INTRODUCTION: 

The aim of the current study was to assess whether there is an indication shift for surgery in patients with ulcerative colitis (UC) from refractory disease to malignant degeneration over the past 3 decades.

METHODS: 

All patients with histologically confirmed UC who underwent a colorectal resection between 1991 and 2020 were extracted from the nationwide Dutch Pathology Registry. The primary outcome was the proportion of colorectal cancer (CRC) in the colon specimens. Outcomes were compared between 3 periods (P1: 1991–2000, P2: 2001–2010, and P3: 2011–2020).

RESULTS: 

Overall, 6,094 patients with UC were included of which 4,854 underwent a (procto)colectomy and 1,240 a segmental resection. In 1,031 (16.9%) patients, CRC was demonstrated in the pathological resection specimen after a median disease duration of 11 years (IQR 3.0–19.0). The proportion of CRC increased from 11.3% in P1, to 16.1% in P2, and 22.8% in P3 (P < 0.001). Median disease duration at the time of resection increased from 4 years in P1, to 10 years in P2, and 17 years in P3 (P < 0.001). The proportion of patients diagnosed with advanced malignancy (pT3/T4) (P1: 61.2% vs P2: 65.2% vs P3: 62.4%, respectively, P = 0.633) and lymph node metastasis (N+) (P1: 33.0% vs P2: 41.9% vs P3: 38.2%, respectively, P = 0.113) did not change over time.

DISCUSSION: 

This nationwide pathology study demonstrated an increased proportion of surgery for CRC over the past 3 decades. We hypothesize that the expanding therapeutic armamentarium for UC leads to exhausting medical options and hence postponed colectomy. This, however, might be at the expense of an increased risk of CRC in the long term.

INTRODUCTION

Patients with ulcerative colitis (UC) have an increased risk of colorectal cancer (CRC), principally resulting from the proneoplastic effects of chronic intestinal inflammation (¹). The impact of this effect remains debated because data on the risk of CRC in patients with UC are conflicting. The first systematic review assessing the risk of CRC in UC from 2001 reported a cumulative risk of 2% at 10 years, 8% at 20 years, and 18% at 30 years (²). Lower CRC risk rates were found in following meta-analysis including population-based studies (^3–5). A recently published population-based cohort study including patients from Sweden and Denmark (n = 96,477) assessed the CRC incidence in patients with UC between 1969 and 2017 (6). They found a 1.7-fold increased risk of incident CRC in UC compared with the general population. This excess risk decreased over time (⁶).

A decrease in the risk of CRC in patients with UC might be related to the implementation of surveillance programs to detect precancerous lesions and the expanding medical options (e.g., biologicals and small molecules), leading to improved control of severe mucosal inflammation (⁷). Key question in this discussion is whether advanced medical therapies have a true modifying effect in the natural course of the disease.

If expanding treatment options indeed lead to postponed colectomy without modifying the natural course of disease, this might introduce an increased risk of CRC over time. Therefore, the aim of the current study was to assess whether the proportion of patients with UC undergoing resection for CRC has increased over the past 3 decades.

METHODS

Study design and data source

Patients with histologically confirmed UC who underwent a colorectal resection between January 1, 1991, and November 12, 2020, in the Netherlands were included in this study. Data were obtained from Dutch Pathology Registry (PALGA), the nationwide network and registry of histopathology and cytopathology in the Netherlands. PALGA was founded in 1971 and achieved nationwide coverage in 1991, comprising all 45 pathology laboratories serving the pathology needs of every hospital in the Netherlands (⁸).

All excerpts were retrieved matching the following search criteria: (“ulcerative colitis” OR “UC” OR “inflammatory bowel disease” OR “IBD”) AND “colorectal resection.” Of these patients, pathology reports of resection specimens and all previous colonic biopsies were included. The standardized excerpts contain an encrypted patient identification number (allowing for identification of multiple samples of 1 patient), age, date of arrival of the pathology specimen, indication for surgery or biopsy, and a full pathology report. Patients were excluded when they had a final diagnosis other than UC (i.e., indeterminate colitis, Crohn's disease, or diverticulitis), colectomy before 1991, no colon resection, colon resection before UC diagnosis, aged <18 years, and when the pathology report of the colon specimen was absent. As anonymized retrospective data were used, no ethical approval was required by the Institutional Review Boards.

Variable definitions

Data were collected on sex, age at diagnosis, age at colon resection, disease duration (time between first biopsy demonstrating UC in PALGA and colon resection), indication for colon resection (therapy-refractory disease, perforation, toxic megacolon, dysplasia/cancer, ischemia, bleeding, colonic stricture, diverticulitis, and other), type resection ([procto]colectomy and segmental resection), and colorectal neoplasia consisting of 4 separate groups: (i) high-grade dysplasia [HGD], (ii) CRC, (iii) neuroendocrine neoplasm, and (iv) other. As there continues to exist a well-known interobserver variability among pathologists in the assessment of dysplasia, this study focused on the presence of CRC (⁹). In patients with CRC, additional data on pathological staging and tumor location were extracted.

Outcomes

The primary outcome was the proportion of CRC in the colon specimens. Secondary outcomes were disease duration before colectomy and tumor stage at time of resection. In addition, the number of colonic resections per period was analyzed.

Statistical analyses

Baseline characteristics and outcomes were analyzed using descriptive statistics. Patients were grouped into 3 cohorts according to the date of their colorectal resection: P1 = 1991–2000, P2 = 2001–2010, and P3 = 2011–2020. Continuous variables were presented as median with interquartile range (IQR) and compared using the median test. Categorical variables were presented as frequencies and proportions and compared using the χ² test. A logistic regression analysis was performed to assess the effect of year of resection on the proportion of patients undergoing resection for CRC. Results were presented as odds ratios (OR) with a 95% confidence interval (CI). Two-sided P values of less than 0.05 were considered statistically significant. Statistical analysis was performed using SPSS Statistics for Windows version 25 (IBM, Armonk, NY).

RESULTS

A total of 14,249 patients with diagnosis UC or IBD and colorectal resection between January 1991 and November 2020 were extracted from the PALGA database. After screening the retrieved excerpts, 8,155 patients were excluded (Figure 1). The final study population consisted of 6,094 patients who underwent a (procto)colectomy (n = 4,854) or a segmental resection (n = 1,240). In 1,031 (16.9%) patients, pathological examination demonstrated CRC after an overall median disease duration of 11 years (IQR 3.0–19.0) (Tables 1 and 2).

Figure 1.: Selection of study population. IBD, inflammatory bowel disease; PALGA, Dutch Pathology Registry; UC, ulcerative colitis.

Table 1.: Baseline characteristics of 6,094 patients with UC undergoing colonic resection between 1991 and 2020 in the Netherlands extracted from the PALGA database

Table 2.: Clinical and pathological characteristics of patients with UC with CRC (n = 1,031)

Proportion of CRC in resected patients with UC

Baseline characteristics and pathological outcomes of patients with CRC are presented in Table 2. The proportion of CRC increased from 11.3% in P1, to 16.1% in P2, to 22.8% in P3 (P < 0.001, Table 1). Figure 2 demonstrates that the proportion of surgery for CRC is steadily increasing over a 30-year period. The OR for undergoing resection for CRC was 1.04 per year (CI = 1.035–1.053, P < 0.0001) and OR 1.24 per 5 years (CI = 1.192–1.297, P < 0.0001). The median age at colon resection for CRC (P1: 56 years [IQR 43.3–70.0] vs P2: 63 years [IQR 51.0–74.0] vs P3: 64 years [IQR 53.0–73.0], P < 0.001) and disease duration at the time of resection (P1: 4 years [IQR 0.0–9.0] vs P2: 10 years [IQR 3.0–16.0] vs P3: 17 years [IQR 7.0–23.0], P < 0.001) increased significantly over the past 3 decades. Excluding 69 patients with CRC (6.7%) without a diagnostic biopsy demonstrating UC before colonic resection did not change the results (see Supplementary Tables 1 and 2, https://links.lww.com/AJG/C797).

Figure 2.: Proportion of patients with ulcerative colitis undergoing resection for CRC per period (P < 0.001). CRC, colorectal cancer.

Pathological staging

Of the 1,031 patients with CRC, the majority (43.6%) had a tumor located in the rectosigmoid compared with 7.3% in the descending colon, 5.6% in the transverse colon, 32.8% in the right-sided colon, 6.5% at multiple sites, and in 4.2%, the location was not specified, P < 0.0001, respectively. The majority of patients were diagnosed with advanced disease (pT3/pT4; n = 651, 63.1%), and 38.4% had lymph node metastasis (Table 2). The percentage of patients diagnosed with advanced disease (pT3/T4) and lymph node metastasis (LN+) did not significantly change over time (tumor stage pT3/T4 P1: 61.2% vs P2: 65.2% vs P3: 62.4%, P = 0.633, and LN + P1: 33.0% vs P2: 41.9% vs P3: 38.2%, P = 0.113, respectively).

Number of colectomies

The absolute number of colonic resections remained comparable over the 3 periods (P1: n = 1,847 vs P2: n = 2,206 vs P3: n = 2,041, respectively, Table 1). Over time, the proportion of (procto)colectomies decreased (P1: 85.6%, P2: 80.4%, and P3: 73.4%), whereas the proportion of segmental resections increased (P1: 10.3%, P2: 16.0%, and P3: 22.9%, P < 0.001, Table 1). In total, 1,240 patients initially underwent a segmental resection of which 229 patients (18.5%) underwent an additional colonic resection. Reasons were recurrent cancer in the remaining colon (n = 40, 17.5%), cancer in the remaining colon without previous diagnosis of cancer (n = 4, 1.7%), and resection of the remaining colon for cancer prevention/active UC (n = 185, 80.8%). The median age of patients undergoing segmental resection was significantly higher compared with patients undergoing (procto)colectomy (66 years [IQR 55.0–74.0] vs 43 years [IQR 32.0–58.0], respectively, P < 0.001).

DISCUSSION

This Dutch nationwide study demonstrated that the proportion of CRC in the colonic specimen of patients with UC markedly increased over the past 30 years (from 11.1% in P1 to 22.8% in P3), and this was accompanied with an increased disease duration at the time of resection. The majority of patients with CRC (63.3%) were diagnosed with advanced disease (pT3/pT4) and 38.5% had lymph node metastasis, which did not decrease over time.

Findings of the current study are in line with previous reports from the United States, Korea, and Japan. Data from the Cleveland Clinic including 4,525 patients undergoing primary ileal pouch–anal anastomosis demonstrated an increasing proportion of dysplasia/cancer in the colectomy specimen (1983–1993 = 0.4%, 1994–2004 = 0.8%, and 2005–2015 = 9.0%, respectively) (¹⁰). The median disease duration before resection did not increase over time (1983–1993 = 7 years, 1994–2004 = 6 years, and 2005–2015 = 6 years, respectively). In another study from the United States, 16,556 patients undergoing colectomy for UC were extracted from the largest all-payer inpatient care database in the United States (National Inpatient Sample database) between 1993 and 2015 (¹¹). In this study, the proportion of colectomies for dysplasia increased from 3.5% to 5.6% (P trend <0.001) and for rectal cancer from 2.6% to 3.0% (P trend = 0.028). Proportions of colonic cancer remained stable in this cohort (4.5%–3.9%, P trend = 0.423). Interestingly, the absolute percentages of patients with UC undergoing colorectal resection for CRC in these studies are markedly lower than the percentages found in current one. This might be related to the start of inclusion (1983), but might also be reflective of differences in patient selection comparing patients from a specialized center undergoing pouch creation (younger patients) vs population-based data including patients undergoing segmental resection only. In addition, differences in disease duration before resection might play a role, and this was, however, not assessed in the study by Ni et al (¹¹).

A Japanese nationwide cohort study including 3,801 surgical cases with UC found an increase in surgery for cancer/dysplasia from 20.2% in 2007 to 34.8% in 2017, whereas the mean disease duration did not change (9.8 ± SD 8.1 years in 2007 to 10.1 ± SD 9.8 years in 2017) (P = 0.85) (¹²). A population-based study from Korea demonstrated a significant indication shift for surgery in patients with UC. The proportion of CRC increased from 27.0% in 2009 to 79.8% in 2015 (¹³). In this study, disease duration was not assessed. It is difficult to comment on whether this high proportion of CRC might be related to differences in treatment approach, individual risk factors, or phenotype of the disease between the Asian and Western population because details on clinical characteristics are not provided in both studies assessed.

The demonstrated surgical indication shift from refractory disease to CRC might be related to the introduction of many new treatment strategies including biologics and small molecules. As new treatment options will introduce the possibility to achieve increased rates of disease control, true mucosal healing is generally only seen in a minority of patients (⁷). Remaining smoldering disease activity in these patients with potentially continued cycles of inflammation and regeneration in the colon might lead to development of dysplastic lesions and consequently increased CRC rates over time. The finding of an increased proportion of CRC also suggests that despite the rapidly expanding medical armamentarium, we are still not able to truly change the natural course of this disease.

The proportion of patients diagnosed with advanced disease is suggested to be higher in patients with UC compared with data of the national Dutch registry demonstrating T3/T4 tumors in approximately 45% and lymph node metastasis in approximately 30% (Dutch Colorectal Audit) (¹⁴). Furthermore, patients with UC and CRC have worse survival compared with patients without UC with same tumor stage (^6,15). The increased risk of cancer development in patients with UC with ongoing disease activity is acknowledged by current guidelines, which all emphasize the need for continued screening in patients with UC. The European Crohn's and Colitis Organization Guideline and the ACG Clinical Guideline both state that a surveillance colonoscopy should be performed in all patients 8–10 years after disease diagnosis and immediately on diagnosis of Primary sclerosing cholangitis. A surveillance colonoscopy should be performed every 1–5 years based on risk factors for colorectal cancer (^16,17). Although these guidelines were introduced over more than 10 years ago, the amount of patients diagnosed with pT3 and pT4 disease did not significantly decrease over time.

The current data suggest that still most patients were diagnosed with advanced carcinomas. This should be considered as a worrisome message because it suggests that the current screening programs do not result in desired early diagnosis. Obviously, detection of a precancerous lesion is challenging especially when the mucosa is severely inflamed, which should be based in mind when offering therapy-refractory patients third- or fourth-line treatment options. Furthermore, potential differences in tumor biology and treatment with immunosuppressants might also lead to a faster progression of cancer in patients with UC. To improve patient-tailored surveillance, more details on the exact patient, disease, and treatment characteristics of these patients at risk of CRC should be obtained.

To avoid selection bias, all types of colonic resections were included in this study. Although a proctocolectomy with ileal pouch–anal anastomosis is the proposed operation in case of cancer (⁹), an increasing proportion of segmental resections performed in patients with CRC was observed. Patients undergoing segmental resection were significantly older compared with patients undergoing proctocolectomy (median age of 66 years [IQR 55.0–74.0] vs 43 years [IQR 32.0–58.0], P < 0.001). These results are in line with data from a population-based study from Canada, which demonstrated that a segmental resection was most commonly performed in patients with UC with CRC (273/599, 45.6%), followed by proctocolectomy (216/599, 36.1%) and total colectomy (110/599, 18.4%) (¹⁸). A retrospective analysis from the United States, including 59 patients with UC and CRC, demonstrated that 40.7% underwent a segmental resection and 59.3% proctocolectomy (¹⁹). Both studies concluded that segmental resection for CRC in the setting of UC is a safe option in selected patients (^19,20).

We must emphasize that this study was not designed to obtain data on the incidence of CRC. The required denominator to determine CRC incidence in the UC population to answer this question is almost impossible to obtain. Although we have a national pathology database covering all hospitals, most pathology reports describing biopsies do not discriminate UC from CD. Moreover, most reports only state "fitting chronic inflammation" without an official IBD diagnosis (and will not show in a search for IBD-diagnosis). Therefore, it will be impossible to extract a reliable denominator from this database. Although the absence of a dominator is suboptimal, this study was conducted with a different aim. Despite the fact that the overall surgical rate in patients with UC was not available, we cannot ignore the finding that the proportion of patients with UC undergoing resection for CRC has increased significantly. Even more worrisome is the finding that most patients are still diagnosed with advanced disease. A previous Dutch study demonstrated that the annual incidence of UC in South Limburg increased significantly from 11.67/100,000 to 21.47/100,000 between 1991 and 2010 (²¹). However, the authors suggest that a shift toward milder disease at diagnosis in combination with the improved diagnostics to detect low-grade inflammation might play an important role in the increasing incidence. This may explain why the numbers of resections in patients with UC remained stable over time.

Several limitations should be mentioned, specifically regarding the retrospective nature of this study. The absence of data on extent of bowel involvement (²), possible risk factors for CRC such as primary sclerosing cholangitis (²²), lifestyle risk factors, clinical characteristics, and family history, impaired us to assess associations between well-known risk factors for CRC (²³). In addition, the role of anti-inflammatory medication such as 5-aminosalicylic acid, thiopurines, biologicals, small molecules, and perhaps experimental drugs could not be assessed in their potential protective or harming effect on the risk of CRC. Finally, the disease duration might be underestimated in some patients who had their biopsy of diagnosis taken before nationwide coverage of the PALGA registry (between 1971 and 1991). However, excluding patients without a diagnosis date did not change the results, and the finding that age at resection shows a comparable increasing trend as disease duration suggests that this was not a major confounder.

The main strength of this study is its large population-based cohort with high external validity. All included patients had histologically confirmed UC and were extracted from the PALGA with reports of all resections performed. The long follow-up time permitted us to examine time trends in the proportion of CRC in resected patients with UC.

In conclusion, this nationwide pathology study demonstrated an indication shift for surgery from therapy-refractory disease to malignant degeneration over the past 30 years. The expanding therapeutic options for UC might lead to longer preservation of the colon, with possibly increased risk of CRC development. The high proportion of patients undergoing resection for advanced disease suggests need for improvement in early detection of (pre)cancerous lesions. Therefore, better understanding of patient and treatment characteristics related to CRC is needed to improve patient-tailored surveillance in UC.

CONFLICTS OF INTEREST

Guarantor of the article: Christianne J. Buskens, MD, PhD.

Specific author contributions: L.H., W.A.B., and C.J.B.: created the study concept and design. L.H. and H.H.: collected the data and conducted the analyses. L.H.: drafted the manuscript. All authors participated in the critical revision of the manuscript for intellectual content and approved the final version.

Financial support: This study was supported by a research grant from the European Crohn's and Colitis Organization.

Potential competing interests: L.H., H.H., H.J.S., and A.M. have no competing interests. G.R.D. has served as advisor for AbbVie, Ablynx, Active Biotech AB, AgomAb Therapeutics, Alimentiv, Allergan, Alphabiomics, Amakem, Amgen, AM Pharma, Applied Molecular Therapeutics; Arena Pharmaceuticals, AstraZeneca, Avaxia, Biogen, Bristol Myers Squibb/Celgene, Boehringer Ingelheim, Celltrion, Cosmo, DSM Pharma; Echo Pharmaceuticals, Eli Lilly, Engene, Exeliom Biosciences; Ferring, DrFALK Pharma, Galapagos, Genentech/Roche, Gilead, GlaxoSmithKline, Gossamerbio, Pfizer, Immunic, Johnson and Johnson, Kintai Therapeutics, Lument, Lycera, Medimetrics, Takeda, Medtronic, Mitsubishi Pharma, Merck Sharp & Dohme, Mundipharma, Nextbiotics, Novo Nordisk, Otsuka, PhotoPll, ProciseDx, Prodigest, Prometheus Laboratories/Nestle, Progenity, Protagonist, RedHill, Salix, Samsung Bioepis, Sandoz, Seres/Nestec/Nestle, Setpoint, Shire, Teva, Tigenix, Tillotts, Topivert, Versant, and Vifor. Received speaker fees from AbbVie, Biogen, Ferring, Galapagos/Gilead, Johnson and Johnson, Merck Sharp & Dohme, Mundipharma, Norgine, Pfizer, Samsung Bioepis, Shire, Millennium/Takeda, Tillotts, and Vifor. S.V. has received grants from AbbVie, J&J, Pfizer, Galapagos, and Takeda and has received consulting and/or speaking fees from AbbVie, AbolerIS Pharma, AgomAb, Alimentiv, Arena Pharmaceuticals, AstraZeneca, Avaxia, BMS, Boehringer Ingelheim, Celgene, CVasThera, Dr. Falk Pharma, Ferring, Galapagos, Genentech-Roche, Gilead, GSK, Hospira, Imidomics, Janssen, J&J, Lilly, Materia Prima, MiroBio, Morphic, MrMHealth, Mundipharma, MSD, Pfizer, Prodigest, Progenity, Prometheus, Robarts Clinical Trials, Second Genome, Shire, Surrozen, Takeda, Theravance, Tillots Pharma AG, and Zealand Pharma. A.D. has no competing interests. W.A.B. has received research funding from VIFOR; served as a consultant for Takeda and Braun; and received speakers fees from Takeda, Johnson, Braun, and Medtronic. C.J.B. has an unrestricted grant from Boehringer Ingelheim and Roche AND has received consultancy fees or speaker's honoraria from AbbVie, Merck Sharp & Dohme, Tillotts, Janssen, and Takeda.

Study Highlights

WHAT IS KNOWN

• ✓ Patients with ulcerative colitis have an increased risk of colorectal cancer (CRC).
• ✓ Improved disease control and surveillance programs are suggested to reduce the risk of CRC.

WHAT IS NEW HERE

• ✓ There is an indication shift for surgery from therapy-refractory disease to CRC.
• ✓ Disease duration at the time of colorectal surgery increased.
• ✓ The proportion of patients diagnosed with advanced malignancy did not change over time.