Abstract
Background:
Approximately 80% of patients with ulcerative colitis (UC) have intermittently active disease and up to 20% will require a colectomy, but little data available on predictors of poor disease course. The aim of this study was to identify clinical and genetic markers that can predict prognosis.
Methods:
Medical records of patients with UC with ≥5 years of follow-up and available DNA and serum were retrospectively assessed. Immunochip was used to genotype loci associated with immune mediated inflammatory disorders (IMIDs), inflammatory bowel diseases, and other single nucleotide polypmorphisms previously associated with disease severity. Serum levels of pANCA, ASCA, CBir1, and OmpC were also evaluated. Requirement for colectomy, medication, and hospitalization were used to group patients into 3 prognostic groups.
Results:
Six hundred one patients with UC were classified as mild (n = 78), moderate (n = 273), or severe disease (n = 250). Proximal disease location frequencies at diagnosis were 13%, 21%, and 30% for mild, moderate, and severe UC, respectively (P = 0.001). Disease severity was associated with greater proximal extension rates on follow-up (P < 0.0001) and with shorter time to extension (P = 0.03) and to prednisone initiation (P = 0.0004). When comparing severe UC with mild and moderate UC together, diagnosis age >40 and proximal disease location were associated with severe UC (odds ratios = 1.94 and 2.12, respectively). None of the single nucleotide polypmorphisms or serum markers tested was associated with severe UC, proximal disease extension or colectomy.
Conclusions:
Older age and proximal disease location at diagnosis, but not genetic and serum markers, were associated with a more severe course. Further work is required to identify biomarkers that will predict outcomes in UC.
Keywords: ulcerative colitis, proximal disease, prognosis, immunochip, serum markers, genetic markers
Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) primarily affecting the large intestine that has a highly variable course which can vary from localized to extensive colonic involvement and from minor symptoms with prolonged periods of remission1 to severe life-threatening presentations that can lead to hospitalization, colectomy, and death.2 Longitudinal cohort studies from Denmark, including over 1000 patients with UC followed for 25 years, have categorized patients with UC into those in prolonged remission, those with intermittent disease and those with continuous symptoms.3 In that study, there was a cumulative risk of over 80% with intermittently active disease, which contrasts with only an 11% chance of having relapse-free disease and overall a 20% risk of colectomy within 10 years. Similarly, over time, 70% of patients with UC progress from proctitis or left-sided colitis to extensive colitis4 that is associated with an increased risk of colectomy5 and colon cancer6 as well as a slight increase in mortality.7
To date, disease severity classification systems ranging from the early Truelove and Witts Score8 to the more recent Montreal classification9 assess activity at a single point in the disease course, which may be helpful to predict short-term outcomes but likely do not reflect the impact of UC longitudinally.10 Attempts have therefore been made to discover better approaches to identifying subjects with a worse prognosis in UC to potentially intervene with more effective management strategies. A population-based cohort study from Norway examined clinical predictors of prognosis in UC and showed that extensive endoscopic involvement, high ESR, Hb ≤10.5 g/dL, and younger age of onset were associated with colectomy. However, none of these factors nor smoking status or gender was associated with a pattern of disease activity over time.11 Another study from Hong Kong associated clinical factors with pattern of corticosteroids (CS) use and response in UC. In that retrospective study (n = 176), extensive colitis predicted CS dependency and the presence of anemia at diagnosis, and initial requirement of total parenteral nutrition predicted CS-refractory UC.12 Another retrospective study found that family history of IBD and extensive disease on endoscopy was associated with being CS refractory.13 Despite these and other studies, clinical predictors seem to be insufficiently sensitive and specific to be used for individualizing patient management.
There are increasing attempts to use various biomarkers as predictors of UC outcomes in the hope that such markers will improve the ability to characterize patients at risk of either mild or severe disease. Biomarkers of interest include serologic and genetic markers that have been found to be associated more generally with UC and IBD. Most studies evaluating serologic markers such as pANCA have not found associations with UC severity.14,15 Moreover, in an IBSEN cohort study, neither extent of colitis nor requirement of colectomy was associated with pANCA.16
In addition to the serological markers, there have also been attempts to determine whether genetic markers are associated with UC outcomes. Older studies have suggested that several loci are associated with risk of colectomy and extensive disease. Examples include the HLA region,13,17–19 IL1B,20 MDR1,21 and HSP70 loci.22 Conversely, single nucleotide polypmorphisms (SNPs) within IL23R were associated with beneficial response to infliximab,23 and a variant in ATG16L1 was protective from severe UC.24
More recently, there have been significant successes in the quest to identify inherited susceptibility factors that result in IBD. In 2012, Jostins et al reported the discovery of more than 160 genetic variants associated with IBD in a large international cohort. The majority of these variants were associated with both Crohn’s disease and UC with a minority being specifically associated with UC.25 As a result, there is now renewed interest in assessing the possibility that these recently discovered markers may be associated with UC severity and prognosis. In this study, we sought to incorporate the use of the Immunochip-based markers to perform a comprehensive assessment of genetic markers that may be associated with UC outcome in addition to evaluating clinical and serologic factors in a comprehensively phenotyped UC cohort.
METHODS
Patient Selection and Severity Group Allocation
This retrospective study was performed at the IBD tertiary care center of Mount Sinai Hospital (MSH) in Toronto, Canada. All the patients provided written informed consent, and the study was approved by the Research Ethics Board of MSH. Subjects were identified by a search of the IBD Genetics Database of MSH. A definitive diagnosis of UC was required by chart review, and detailed phenotyping was performed examining clinical, endoscopic, imaging, and pathology records verified by at least 1 expert, IBD-trained gastroenterologist. Only patients with confirmed UC, ≥5 years of follow-up and full out-patient and hospitalization records, and with genomic DNA available for genotyping were included. Pattern of corticosteroid use (mode of administration and cumulative dose), tumor necrosis factor alpha (TNF-α) antagonists use (indication), and colectomy requirements (timing and indications) were used to group the eligible patients into 3 severity groups: severe UC, moderate UC, and mild UC (Table 1).
TABLE 1.
Prognostic Groups—Inclusion Criteria
| Mild UC | Moderate UC | Severe UC | |
|---|---|---|---|
| IV steroid requirement | Never | Allowed, but less than twice during initial 2 years | Twice or more during initial 2 years |
| Colectomy | No, except for cancer/ dysplasia | Only for chronic active disease | For IV steroid failure, toxic megacolon or massive bleeding |
| Biological therapy | Never | Not as inpatient | Induction only as inpatient |
| Cumulative duration of prednisone use | Less than 3 months | More than 3 months | NA |
IV, intravenous; NA, not applicable.
The patients who met the inclusion criteria were treated and followed at MSH between 1980 and 2010. Patients with severe colitis requiring hospital admission were usually treated with intravenous corticosteroids. Colectomy was considered when intravenous corticosteroids failure was apparent or in chronically active UC when corticosteroids nonresponse or dependency was noted. TNF-α antagonists were increasingly in use starting in 2000 onwards. Intravenous and oral cyclosporine was not in use at MSH and in most of the IBD centers referring patients to MSH during the study period.
Clinical and Endocopic Data Collection
Clinical data collected included the following: date of birth, diagnosis date, age at diagnosis, gender, ethnicity, family history of IBD in first degree relatives, appendectomy, smoking at diagnosis (current vs. stopped/never), hospitalization for UC exacerbation, date of first oral steroid use, cumulative time on steroids, date of first use of azathioprine or TNF-α antagonist (infliximab or adalimumab) and response, oral 5-aminosalycilic acid (5-ASA) use and response, colectomy (date and indication), and the date of the last follow-up.
Endoscopic data collected were date of diagnostic (i.e., first) endoscopy, colitis extent at diagnostic endoscopy (proctitis, colitis up to the splenic flexure, and colitis proximal to the splenic flexure (E1, E2, and E3, respectively), as per the Montreal classification9), dates and disease extent on subsequent endoscopies, date of last endoscopy, and extent at last endoscopy. The retrospective nature of this study precluded analysis of disease severity on initial endoscopy as no standardized severity scale was applied. In patients who had Montreal E1 or E2 disease location on their diagnostic endoscopy, the date of proximal endoscopic extension proximally to the splenic flexure was noted and time to proximal endoscopic extension was calculated.
Response to a specific medication was defined as continuous therapy with the drug with a notation by the treating gastroenterologist that the patient was doing well on the last follow-up. Dynamic parameters calculated included time to first hospitalization, time to first steroid use, time to colectomy, time to proximal endoscopic extension, and follow-up duration from diagnosis to the last follow-up.
Serological Analysis
Peripheral blood samples were spun and serum aliquoted into sterile vials and stored at — 80°C for future testing. We excluded patients who had blood drawn for serological analysis after colectomy. Serologic assays for a panel of antimicrobial antibodies and autoantibodies including perinuclear antineutrophil cytoplasmic antibodies (pANCA), antibodies to Saccharomyces cerevisiae antigen (ASCA) IgA and IgG, antiflagellin (CBir1), and antibodies to Escherichia coli outer membrane porin antigen (anti-OmpC) were performed at Prometheus Laboratories (San Diego, CA) using enzyme-linked immunosorbent assay and immunofluorescence, as previously described.26 An antibody test was considered positive if the titers were greater than established cutoff limits: ASCA IgA >20 EU/mL; ASCA IgG >40 EU/mL; pANCA >12.1 EU/mL and pANCA IFA DNAse sensitive; anti-OmpC >16.4 EU/mL; CBir1 >21 EU/mL. Overall, serologic data were available for analysis in 139 patients. Median and range values for each serum marker were calculated in each severity group. In each severity group, the proportion of patients with 1, 2, 3, or 4 positive serum marker assays were calculated and compared (no patients had 5 positive serum markers).
Genotypic Data Collected and Analysis
Genomic DNA samples were obtained from peripheral blood collected from the study subjects and were genotyped using Immunochip. Immunochip is an Illumina Infinium genotyping chip, containing 196,524 polymorphisms (718 small insertion deletions, 195,806 SNPs) designed both to perform deep replication of major autoimmune and inflammatory diseases, and fine-mapping of established genome-wide association study–significant loci.27 Genotypic data were available for analysis in 465 patients. Determination of the minor allele frequency and genotype distribution among the 3 prognostic groups was performed. Logistic regression models were used to investigate association of genetic variations with prognostic grouping. Three different comparisons were made: severe UC versus mild and moderate UC; proximal endoscopic extension (transition from Montreal E1 or E2 to E3) versus persisting E1 or E2 disease at the last follow-up; UC requiring colectomy for active disease (i.e., medically refractory [MR]) versus UC not requiring colectomy. Genetic association analysis was performed using a logistic regression model assuming the additive genetic model and correcting for population stratification with 6 principal components as covariates (PLINK v.1.06). For genetic analysis, significance level was set at P ≤ 2.5 × 10−7 for whole Immunochip data and at P ≤ 3 × 10−4 for IBD- and severity-associated loci; the 2 strata represent SNPs for which we have different a priori hypothesis, and Bonferroni correction was performed separately for each strata.
Clinical Data Analysis
Comparisons of the clinical characteristics at diagnosis and endoscopic extent at diagnosis were made between 3 groups: severe UC, moderate UC, and mild UC. Chi-square or Fisher’s exact tests were applied for the comparison of categorical clinical variables. Kruskal-Wallis tests were applied on continuous factors. Odds ratios (OR) and 95% confidence interval (CI) were estimated for each marker when we compared severe with mild and moderate disease together. For the risk of endoscopic extension along time analysis, we excluded patients who had extensive colitis (Montreal E3) on their diagnostic endoscopy or when the endoscopic extent at diagnosis was unknown. The date of endoscopic disease extension was collected and time to extension was calculated. Clinical parameters at diagnosis and genotypic data were compared between patients whose disease extended and patients who remained with limited colitis at the last follow-up. Results were considered significant if P-values were ≤0.05.
RESULTS
Clinical Predictors at Diagnosis
Overall 601 patients with UC had complete records, adequate length of follow-up, and DNA available for analysis. Two hundred fifty patients met the criteria for severe UC, 273 had moderate UC, and 78 were included in the mild UC group. The clinical and endoscopic characteristics for the overall UC group and for each disease severity group are described in Table 2 and Table 3. None of the epidemiological and clinical characteristics at diagnosis (Table 2) were associated with increased severity. However, extensive disease (Montreal E3) on the initial endoscopy was more frequent in the moderate and severe groups (13% versus 21% versus 30% in mild, moderate, and severe UC, respectively, P = 0.001). Active smoking at diagnosis was more frequent in the mild group versus the moderate or severe group, although this trend did not reach statistical significance. When we compared severe UC with the combined group of mild and moderate UC (Table 4), patients with severe UC were more likely to be >40 years old and have extensive disease (OR of 1.66 and 2.12, respectively, P ≤ 0.02). On multivariate analysis, Montreal E3 was associated with severe UC (OR = 2.12; 95% CI, 1.39—3.24; P = 0.0005). Age >40 at diagnosis was also associated with severe UC (OR = 1.94; 95% CI, 1.15–3.25; P = 0.01). Neither active smoking nor ethnicity or a family history of IBD was associated with disease severity.
TABLE 2.
Clinical and Endoscopic Characteristics at Diagnosis of Patients with UC by Severity Group
| Overall UC (n = 601) | Mild UC (n = 78) | Moderate UC (n = 273) | Severe UC (n = 250) | P | |
|---|---|---|---|---|---|
| Gender (%) | 0.52 | ||||
| Male | 283 (47) | 41 (53) | 129 (47) | 113 (45) | |
| Female | 318 (53) | 37 (47) | 144 (53) | 137 (55) | |
| Age at diagnosis, median (range), yr | 27 (2–73) | 28 (9–66) | 26 (6–60) | 27 (2–73) | 0.17 |
| Age group (%) | 0.08 | ||||
| ≤16 | 80 (13) | 10 (13) | 33 (12) | 37 (15) | |
| 17–40 | 417 (70) | 54 (69) | 204 (75) | 159 (64) | |
| >40 | 104 (17) | 14 (18) | 36 (13) | 54 (22) | |
| Ethnicity (%) | 0.60 | ||||
| White | 507 (84) | 68 (87) | 231 (85) | 208 (83) | |
| Asian | 61 (10) | 6 (8) | 24 (9) | 31 (12) | |
| Black | 7 (1) | 2 (3) | 4 (1) | 1 (0) | |
| Hispanic | 2 (0) | 0 | 1 (0) | 1 (0) | |
| Others | 24 (5) | 2 (2) | 13 (5) | 9 (5) | |
| Family history—IBD (%) | 0.35 | ||||
| 1st degree | 120 (20) | 22 (28) | 49 (18) | 49 (20) | |
| 2nd degree or more | 92 (15) | 9 (12) | 45 (16) | 38 (15) | |
| None | 389 (65) | 47 (60) | 179 (66) | 163 (65) | |
| Extent at diagnosis (%) | 0.001 | ||||
| E1 | 89 (14) | 17 (22) | 44 (16) | 28 (11) | |
| E2 | 178 (30) | 33 (42) | 79 (29) | 66 (26) | |
| E3 | 142 (24) | 10 (13) | 56 (21) | 76 (31) | |
| Unknown | 191 (32) | 18 (23) | 93 (34) | 80 (32) | |
| Smoking at diagnosis (%) | 0.12 | ||||
| Yes | 83 (14) | 10 (13) | 47 (17) | 26 (11) | |
| Ex-smoker | 135 (22) | 13 (17) | 62 (23) | 60 (26) | |
| No | 357 (60) | 54 (69) | 154 (56) | 149 (63) | |
| Unknown | 26 (4) | 1 (1) | 10 (4) | 15 (6) | |
| Appendectomy before diagnosis (%) | 12 (2) | 2 (3) | 5 (2) | 5 (2) | 0.96 |
E1, proctitis; E2, colitis up to the splenic flexure; E3, colitis proximal to the splenic flexure.
TABLE 3.
Follow-up Clinical and Endoscopic Outcomes in Patients with UC by Severity Group
| Overall UC (n = 601) | Mild UC (n = 78) | Moderate UC (n = 273) | Severe UC (n = 250) | P | |
|---|---|---|---|---|---|
| Length of follow-up, mo | 162 (8–683) | 162 (42–494) | 180 (20–683) | 138 (8–594) | 0.0004 |
| Median number of endoscopic procedures | 3 (1–5) | 4 (1–5) | 3 (1–5) | 2 (1–5) | <0.0001 |
| Need for colectomy (%) | NA | ||||
| Yes | 385 (64) | 10 (13) | 167 (61) | 208 (83) | |
| No | 216 (36) | 68 (87) | 106 (39) | 42 (17) | |
| Colectomy indication (%) | NA | ||||
| IV steroid failure | 139 | 0 | 0 | 139 | |
| Megacolon | 21 | 0 | 0 | 21 | |
| Massive bleeding | 5 | 0 | 0 | 5 | |
| Chronic disease | 190 | 0 | 149 | 41 | |
| Cancer/dysplasia | 30 | 10 | 18 | 2 | |
| Time from diagnosis to colectomy (mo) | 60 (0–469) | 185 (60–303) | 97 (11–469) | 30 (0–436) | <0.0001 |
| Oral prednisone (%) | NA | ||||
| Yes | 498 (83) | 11 (14) | 264 (97) | 224 (90) | |
| No | 80 (13) | 64 (82) | 8 (3) | 8 (3) | |
| Unknown | 23 (4) | 3 (4) | 1 (0) | 18 (7) | |
| Prednisone duration (%) | NA | ||||
| Up to 3 mo | 8 (12) | 9 (12) | 8 (3) | 41 (19) | |
| 3–12 mo | 150 (32) | 0 | 68 (27) | 82 (39) | |
| More than 1 yr | 261 (55) | 0 | 173 (69) | 88 (42) | |
| Time from diagnosis to 1st prednisone (range), mo | 56 (0–424) | 48 (0–372) | 22 (0–618) | 5 (0–413) | 0.0004 |
| AZA/6-MP (%) | 31 (5) | 2 (3) | 20 (7) | 9 (4) | NA |
| Effective long term | 31 (5) | 2 (3) | 20 (7) | 9 (4) | |
| Stopped—ineffective | 108 (18) | 0 | 68 (25) | 40 (16) | |
| Stopped—AE | 68 (11) | 0 | 47 (17) | 21 (8) | |
| Never prescribed | 377 (63) | 76 (97) | 135 (50) | 166 (66) | |
| Unknown | 17 (3) | 0 | 3 (1) | 14 (6) | |
| TNF-α antagonists (%) | NA | ||||
| IFX effective | 35 (6) | 0 | 23 (8) | 12 (5) | |
| ADA effective | 7 (1) | 0 | 6 (2) | 1 (0) | |
| Ineffective | 41 (7) | 0 | 19 (7) | 22 (9) | |
| Never prescribed | 513 (85) | 77 (99) | 224 (82) | 212 (85) | |
| Unknown | 5 (1) | 1 (1) | 1 (0) | 3 (1) | |
| 5-ASA (%) | NA | ||||
| Effective long term | 139 (23) | 70 (90) | 49 (18) | 20 (8) | |
| Stopped—ineffective | 411 (69) | 6 (8) | 205 (75) | 201 (80) | |
| Stopped—AE | 30 (5) | 2 (2) | 16 (6) | 12 (5) | |
| Never prescribed | 11 (2) | 0 | 1 (0) | 10 (4) | |
| Unknown | 10 (2) | 0 | 2 (1) | 7 (3) | |
| PSC (%) | 0.19 | ||||
| Yes | 22 (4) | 4 (5) | 11 (4) | 7 (3) | |
| No | 565 (94) | 70 (90) | 259 (95) | 236 (94) | |
| Unknown | 14 (2) | 4 (5) | 3 (1) | 7 (3) | |
| EIM (%) | 0.13 | ||||
| Yes | 397 (66) | 45 (58) | 190 (70) | 162 (65) | |
| No | 204 (34) | 33 (42) | 83 (30) | 88 (35) | |
| Median time to colectomy, mo | 60 | 185 | 97 | 30 | <0.0001 |
| Median time to extension, mo | 56 | 75 | 74 | 37 | 0.03 |
| Median time to 1st oral CS, mo | 56 | 48 | 22 | 5 | 0.0004 |
ADA, adalimumab; AE, adverse effect; ASA, amino salicylic acid; AZA, azathioprine; EIM, extraintestinal manifestations; IFX, infliximab; IV, intravenous; 6-MP, 6-mercaptopurine; PSC, primary sclerosing cholangitis.
TABLE 4.
Clinical Parameters at Diagnosis Associated with Severe UC
| Moderate + Mild UC (n = 351) | Severe UC (n = 250) | OR (95% CI) | p | |
|---|---|---|---|---|
| Age (%), yr | 0.02 | |||
| >40 | 50 (14) | 54 (22) | 1.66 (1.08–2.54) | |
| ≤40 | 301 (86) | 196 (78) | ||
| Extent at diagnosis (%) | 0.0003 | |||
| E3 | 66 (19) | 76 (30) | 2.12 (1.40–3.21) | |
| E1/E2 | 173 (49) | 94 (38) | ||
| Unknown | 112 (32) | 80 (32) | ||
| Smoking (%) | 0.06 | |||
| Yes | 57 (16) | 26 (10) | 0.62 (0.38–1.02) | |
| Ex-smoker/no | 283 (81) | 209 (84) | ||
| Unknown | 11 (3) | 15 (6) | ||
| Ethnicity (%) | 0.12 | |||
| Asian | 30 (9) | 31 (12) | 1.52 (0.90–2.59) | |
| Others | 321 (91) | 219 (88) | ||
| Appendectomy before diagnosis (%) | 7 (2) | 5 (2) | 0.96 |
CI, confidence interval; E1, proctitis; E2, colitis up to the splenic flexure; E3, colitis proximal to the splenic flexure.
When we evaluated longitudinal disease activity patterns, a shorter time to oral steroid initiation, to colectomy, and to proximal extension on follow-up endoscopy were all associated with increased severity (Table 3). There were also significant differences in response to biological therapy between the moderate and severe groups. Effective long-term response rates to biologics was 64% (29 of 48) in moderate UC and only 29% (13 of 45) in severe UC (P = 0.002). However, there were no differences among the groups with respect to response to azathioprine or 6-mercaptopurine.
Serological Markers and Disease Severity
Serum markers were analyzed in 139 subjects in whom samples were available and never had a colectomy and those with a colectomy where serum was taken before surgery. The proportion of patients with positive tests for pANCA, ASCA IgA and IgG, anti-OmpC, and CBir1 were similar in mild UC (n = 24), moderate UC (n = 83), and severe UC (n = 32) (see Table, Supplemental Digital Content 1, http://links.lww.com/IBD/A933 ). When the median titers of each serum marker were compared between the mild, moderate, and severe groups, there were no significant differences in pANCA, ASCA, and anti-OmpC titers. However, compared with mild UC, patients with moder ate and severe UC had higher median titers of anti-CBir1 (P = 0.003). There were also no differences between the severity groups when assessing the distribution of patients with 0, 1, 2, 3, or 4 positive markers.
Genetic Markers and UC Severity
Genetic data were available for 465 patients (severe UC = 192, moderate and mild UC = 273). None of the genotyped loci on Immunochip were significantly associated with severe UC when compared with mild and moderate UC; SNPs with significance levels of P ≤ 1× 10−5 are reported in Table, Supplemental Digital Content 2, http://links.lww.com/IBD/A934. Of the 163 SNPs from Jostins et al., rs9264942 (upstream of HLA-C on chromosome 6) had the most significant uncorrected P-value of 0.004. Genetics data were available for 282 patients with UC requiring colectomy (designated previously as MRUC) and 184 non-MRUC patients. None of the loci on Immunochip were associated with MRUC. Of the 37 SNPs previously reported by Haritunians et al13 in their study of MRUC, the most significant SNP, rs2831462, has an uncorrected P-value of 0.004.
Markers at Diagnosis and Proximal Extension of UC
Complete data on endoscopic extent at diagnosis were available for 409 patients. One hundred forty-two patients had extensive disease at diagnosis and thus were excluded from this analysis. Of the 286 patients with limited colitis (Montreal E1 or E2) at their initial endoscopy, 174 progressed to have proximal disease (Montreal E3) on the last follow-up endoscopy and 112 continued to have only limited endoscopic disease extent. None of the clinical characteristics at diagnosis were associated with proximal disease extension (Table 5). When evaluating genetic markers and endoscopic disease extension, genetic data were available for 90 cases with limited colitis at the last follow-up and 140 who progressed to Montreal E3. None of the genotyped loci on Immunochip and none of the 163 IBD-associated regions were significantly associated with endoscopic extension. The most significantly associated SNP was rs6460071 (LIMK1) with a P-value of 2.7 × 10−6.
TABLE 5.
Clinical Markers at Diagnosis and Proximal Endoscopic Extension in UC
| Proximal Extension on Last Follow-up (n = 174) | Limited Colitis on Last Follow-up (n = 112) | P | |
|---|---|---|---|
| Age group distribution, yr | 0.09 | ||
| ≤16 | 20 | 5 | |
| 17–40 | 128 | 85 | |
| >40 | 26 | 22 | |
| Female gender, % | 58.6 | 56 | 0.59 |
| Smoking at diagnosis, % | 18.6 | 13.8 | 0.31 |
| Ethnicity (white), % | 71.8 | 61.6 | 0.07 |
Proximal extension, endoscopic extent proximal to splenic flexure.
DISCUSSION
This retrospective study evaluated various parameters at the time of diagnosis and their association with longitudinal outcomes of UC in a large cohort of patients in a tertiary care IBD center. In general, most of the clinical characteristics, serum antibody markers, and genetic markers tested were not associated with disease severity or risk for proximal endoscopic extension as defined by colectomy requirements, pattern of CS use, and TNF-α antagonist requirement. However, patients who were older (>40 yr) at diagnosis or had extensive disease (Montreal E3) at their initial endoscopy were at risk for severe UC with an OR of 1.94 and 2.12, respectively.
In this study, we proposed to use colectomy data (indication and timing from diagnosis), pattern of biological therapy use (indication), and pattern of systemic CS use to define severity phenotypes of UC. Previous studies defined disease severity phenotype in UC based on colectomy requirements, patterns of CS use,3,13 and hospitalization requirements.3 However, differentiation between colectomy for acute intravenous steroid-refractory disease and colectomy for chronically active disease has not been made, despite the fact that these may represent quite different patient populations. Moreover, need for biological therapy in defining longitudinal severity phenotypes, and not just in the context of acute severe UC, has not previously been used. We believe that our criteria may enhance and update definitions of mild and severe phenotypes in UC. Although our proposed definitions of UC severity are not validated, we believe that they reflect clinically useful parameters that would serve to separate the outcome groups in UC beyond just the outcome of colectomy. Some internal validation may be found, however, by the significant differences between the groups in independent variables such as response to biological therapy, rates of proximal endoscopic extension, and time to steroid initiation.
Most previous studies that have examined clinical and endoscopic parameters that predict prognosis in UC have focused on acute exacerbations and the need for colectomy. Relatively few studies reported the association of measurable parameters at diagnosis and disease severity over time.
In congruence with our findings, extensive endoscopic involvement has been shown to be associated with increased risk for steroid-refractory disease leading to colectomy.11,13 In a Dutch study evaluating long-term pattern of activity in UC, age <40 at diagnosis was associated with a slight risk for disease recurrence albeit only for the initial 10 years from diagnosis.28 However, in our cohort, age >40 at diagnosis was associated with the more severe phenotype. It should be noted that the endpoint in our study is defined differently than in the Dutch study. In the Dutch study, age was associated with recurrence which was defined as increased disease activity leading to either change in medication dosing or requiring surgery, but not assessing severity as was defined in our study. Another large retrospective cohort of UC patients (n = 861) did not find age of onset to be associated with MRUC leading to colectomy.13 Again, this may be due to the difference in definitions of outcomes. In our cohort, smoking showed a trend to association with mild colitis, although it did not reach statistical significance. This may well be a result of the relatively small sample size of patients with mild disease— a reflection of the referral bias typical for studies conducted at tertiary care centers. Indeed, other studies did show an association between active smoking and protection from hospitalization29 and colectomy.30 It is quite plausible had we had a larger sample of mild UC cases this inverse association between smoking and disease severity would have reached statistical significance. The small sample size of patients with mild UC significantly limited our power to compare mild UC with a combined group of patients with moderate and severe UC. Therefore, such analysis was not done. Further studies looking into characteristics of patients with mild UC are warranted indeed but would best be performed outside the setting of tertiary care centers.
We also found that a panel of 5 serum antibodies including ASCA, pANCA, and the antimicrobial antibodies anti-OmpC and antiflagellin (CBir1) was not associated with severe phenotypes of UC, although there was a slight increase in median anti-CBir1 titers in the more severe groups. The fact that serum samples have not been drawn in our patients close to the time of diagnosis makes these data difficult to interpret, especially when prediction of disease severity is sought at the time of diagnosis. However, a larger study did not find associations between severe UC and the risk for colectomy and pANCA positivity,16 in support of our negative finding.
The association of genetic factors and severe UC was investigated in several studies. Intuitively, a positive family history of UC would suggest a risk for more aggressive phenotype, and indeed, a previous study did show an association between steroid-refractory UC leading to colectomy and family history of IBD and UC in particular.13 However, we did not find that family history was associated with a severe phenotype, and this is supported by other studies that have not found a relation between family history of IBD and risk for colectomy.11,31 Several previous studies found associations between SNPs in several loci and a severe UC phenotype. It should be noted that major histocompatibility complex (MHC)-related loci have emerged as risk markers for severe UC in several studies,13,17–19 but a recent Dutch study on 561 patients with UC, which tested 20 UC-relevant SNPs, shows no association between severe UC and HLA-DRA.32 Another study using a genome-wide association approach on patients with MRUC identified a panel of 46 SNPs that were nominally associated. Ten of the SNPs were within the MHC region.13 As the Immunochip has more than 7000 SNPs within the extended MHC region 25 652 429–33 368 333 (build hg19), we were able to attempt replication of these findings. Specifically, we had data on 37 of the 45 MHC SNPs previously associated with MRUC. None of these 37 SNPs were associated with MRUC in our cohort.
Other loci lying outside the MHC region were also associated with UC severity. A study in Korean patients with UC found that the A allele of the Heat Shock Protein gene, HSP70-2 was associated with more severe UC.22 An SNP within the CASP9 gene on 1p36 was associated with need for colectomy in a cohort of British white patients.33Several SNPs within MDR1 were also associated with CS-refractory disease in patients with UC.21 Conversely, in Australian patients with UC, an inverse correlation with risk of colectomy and proximal disease extent was noted in carriers of the ATG16L1 T300A variant.24 Another study showed that carriers of certain IBD-risk SNPs within IL23R have a beneficial response to infliximab.23 Of those loci outside the MHC region, CASP9 and MDR1 are not well covered by the Immunochip, but Immunochip does cover the associated SNP in ATG16L1. We were not able to replicate any of these findings using rigorous statistical methods although power may be an issue. Although larger samples sizes may allow detection of markers that reach significance, their relevance may then become less meaningful due to diminishing ORs/relative risks.
Considering the fact that the targeted genotypic approach in our study and in previous studies have generally not resulted in reproducible associations with disease severity in UC and the fact that genome-wide association studies related to disease severity have not yet been replicated in severe UC, no firm evidence supporting the role of genetics in the determination of disease severity have yet emerged.
In summary, in this retrospective, large, well phenotyped cohort, severe UC is associated with extensive endoscopic involvement, older age at diagnosis, and possibly with nonsmoking at diagnosis. Serological and genetic factors were not associated with more severe phenotypes. Proximal extension of inflammatory endoscopic findings was not associated with clinical or genetic parameters at diagnosis. Our findings support the importance of early determination of disease extent, likely at the initial endoscopy in the prediction of severe disease behavior. Clinical criteria including indication for and timing of colectomy and biological and systemic steroid exposure may still be helpful in defining more severe phenotypes in UC. The clinical criteria suggested in this study to define disease severity need to be validated in external cohorts. Other epidemiological and clinical variables at diagnosis as well as the genetic and serological markers that we investigated do not seem to play an important role in prediction of disease severity on the long term. It is possible that other biomarkers in the serum, tissue or stool, such as the intestinal microbiome, may play a more important role in determining disease course in UC.
Supplementary Material
ACKNOWLEDGMENTS
The authors would like to acknowledge the International IBD Genetic Consortium for performing genotype calling on this dataset.
M. Waterman was supported by a Canadian Institutes for Health Research (CIHR)/ Canadian Association of Gastroenterology (CAG) Fellowship Grant. M. S. Silverberg is supported by the Gale and Graham Wright Chair in Digestive Diseases at the Zane Cohen Centre for Digestive Diseases at Mount Sinai Hospital, Toronto and by grants from Crohn’s and Colitis Canada (CCC) and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (DK-062423).
Footnotes
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.ibdjournal.org).
The authors have no conflicts of interest to disclose.
REFERENCES
- 1.Meucci G, et al. The natural history of ulcerative proctitis: a multicenter, retrospective study. Gruppo di Studio per le Malattie Infiammatorie Intestinali (GSMII). Am J Gastroenterol. 2000;95:469–473. [DOI] [PubMed] [Google Scholar]
- 2.Robert JH, et al. Management of severe hemorrhage in ulcerative colitis. Am J Surg. 1990;159:550–555. [DOI] [PubMed] [Google Scholar]
- 3.Langholz E, et al. Course of ulcerative colitis: analysis of changes in disease activity over years. Gastroenterology. 1994;107:3–11. [DOI] [PubMed] [Google Scholar]
- 4.Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology. 2004; 126:1504–1517. [DOI] [PubMed] [Google Scholar]
- 5.Leijonmarck CE, Persson PG, Hellers G. Factors affecting colectomy rate in ulcerative colitis: an epidemiologic study. Gut. 1990;31:329–333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Langholz E, et al. Colorectal cancer risk and mortality in patients with ulcerative colitis. Gastroenterology. 1992;103:1444–1451. [DOI] [PubMed] [Google Scholar]
- 7.Louis E, Belaiche J, Reenaers C. Do clinical factors help to predict disease course in inflammatory bowel disease? World J Gastroenterol. 2010;16: 2600–2603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Truelove SC, Witts LJ. Cortisone in ulcerative colitis; final report on a therapeutic trial. Br Med J. 1955;2:1041–1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Silverberg MS, et al. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can J Gastroenterol. 2005;19(suppl A):5–36. [DOI] [PubMed] [Google Scholar]
- 10.Satsangi J, et al. The Montreal classification of inflammatory bowel disease: controversies, consensus, and implications. Gut. 2006;55:749–753. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Solberg IC, et al. Clinical course during the first 10 years of ulcerative colitis: results from a population-based inception cohort (IBSEN Study). Scand J Gastroenterol. 2009;44:431–440. [DOI] [PubMed] [Google Scholar]
- 12.Chow DK, et al. Predictors ofcorticosteroid-dependent and corticosteroid-refractory inflammatory bowel disease: analysis ofa Chinese cohort study. Aliment Pharmacol Ther. 2009;29:843–854. [DOI] [PubMed] [Google Scholar]
- 13.Haritunians T, et al. Genetic predictors of medically refractory ulcerative colitis. Inflamm Bowel Dis. 2010;16:1830–1840. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zholudev A, et al. Serologic testing with ANCA, ASCA, and anti-OmpC in children and young adults with Crohn’s disease and ulcerative colitis: diagnostic value and correlation with disease phenotype. Am J Gastroenterol. 2004;99:2235–2241. [DOI] [PubMed] [Google Scholar]
- 15.Preda CM, et al. Prevalence and significance of perinuclear anti-neutrophil antibodies (pANCA) in Romanian patients with Crohn’s disease and ulcerative colitis. Rom J Gastroenterol 2005;14:357–360. [PubMed] [Google Scholar]
- 16.Solberg IC, et al. Predictive value of serologic markers in a population-based Norwegian cohort with inflammatory bowel disease. Inflamm Bowel Dis. 2009;15:406–414. [DOI] [PubMed] [Google Scholar]
- 17.de la Concha EG, et al. Susceptibility to severe ulcerative colitis is associated with polymorphism in the central MHC gene IKBL. Gastroenterology. 2000;119:1491–1495. [DOI] [PubMed] [Google Scholar]
- 18.Heresbach D, et al. Are HLA-DR or TAP genes genetic markers of severity in ulcerative colitis? J Autoimmun. 1996;9:777–784. [DOI] [PubMed] [Google Scholar]
- 19.Bouma G, et al. Genetic markers in clinically well defined patients with ulcerative colitis (UC). Clin Exp Immunol. 1999;115:294–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nemetz A, et al. IL1B gene polymorphisms influence the course and severity of inflammatory bowel disease. Immunogenetics. 1999;49:527–531. [DOI] [PubMed] [Google Scholar]
- 21.Potocnik U, et al. Polymorphisms in multidrug resistance 1 (MDR1) gene are associated with refractory Crohn disease and ulcerative colitis. Genes Immun. 2004;5:530–539. [DOI] [PubMed] [Google Scholar]
- 22.Nam SY, et al. Heat shock protein gene 70–2 polymorphism is differentially associated with the clinical phenotypes of ulcerative colitis and Crohn’s disease. J Gastroenterol Hepatol. 2007;22:1032–1038. [DOI] [PubMed] [Google Scholar]
- 23.Jurgens M, et al. Disease activity, ANCA, and IL23R genotype status determine early response to infliximab in patients with ulcerative colitis. Am J Gastroenterol. 2010;105:1811–1819. [DOI] [PubMed] [Google Scholar]
- 24.Fowler EV, et al. ATG16L1 T300A shows strong associations with disease subgroups in a large Australian IBD population: further support for significant disease heterogeneity. Am J Gastroenterol. 2008;103:2519–2526. [DOI] [PubMed] [Google Scholar]
- 25.Jostins L, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012;491: 119–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Landers CJ, et al. Selected loss of tolerance evidenced by Crohn’s disease-associated immune responses to auto and microbial antigens. Gastroenterology. 2002;123:689–699. [DOI] [PubMed] [Google Scholar]
- 27.Cortes A, Brown MA. Promise and pitfalls of the immunochip. Arthritis Res Ther. 2011;13:101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Romberg-Camps MJ, et al. Influence of phenotype at diagnosis and of other potential prognostic factors on the course of inflammatory bowel disease. Am J Gastroenterol. 2009;104:371–383. [DOI] [PubMed] [Google Scholar]
- 29.Boyko EJ, et al. Effects of cigarette smoking on the clinical course of ulcerative colitis. Scand J Gastroenterol. 1988;23:1147–1152. [DOI] [PubMed] [Google Scholar]
- 30.Szamosi T, et al. Early azathioprine/biological therapy is associated with decreased risk for first surgery and delays time to surgery but not reoperation in both smokers and nonsmokers with Crohn’s disease, while smoking decreases the risk of colectomy in ulcerative colitis. Eur J Gastroenterol Hepatol. 2010;22:872–879. [DOI] [PubMed] [Google Scholar]
- 31.Falcone RA Jr, Lewis LG, Warner BW. Predicting the need for colectomy in pediatric patients with ulcerative colitis. J Gastrointest Surg. 2000;4:201–206. [DOI] [PubMed] [Google Scholar]
- 32.Festen EA, et al. Genetic analysis in a Dutch study sample identifies more ulcerative colitis susceptibility loci and shows their additive role in disease risk. Am J Gastroenterol. 2010;105:395–02. [DOI] [PubMed] [Google Scholar]
- 33.Guo C, et al. Association of caspase-9 and RUNX3 with inflammatory bowel disease. Tissue Antigens. 2011;77:23–29. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.