Postoperative complications and prognosis based on type of surgery in ulcerative colitis patients with colorectal cancer: A multicenter observational study of data from the Japanese Society for Cancer of the Colon and Rectum

Patients with ulcerative colitis are reported to be at increased risk of colorectal cancer and are also at high risk of postoperative complications. However, the incidence of postoperative complications in these patients and how the type of surgery performed affects prognosis are not well understood.


| INTRODUC TI ON
Ulcerative colitis is a chronic inflammatory disease of the colon, and the mainstay of treatment is pharmacotherapy, including 5-aminosalicylic acid, steroids, and immunosuppressive agents. [1][2][3][4] Surgery is required in patients with ulcerative colitis if they develop colorectal cancer, when drug treatment is unsuccessful, or if the condition becomes an emergency, such as toxic megacolon or intestinal perforation. [5][6][7] The standard surgical procedure for ulcerative colitis is reconstruction of the intestinal tract by ileoanal anastomosis (IAA) or ileoanal canal anastomosis (IACA) after total colorectal resection. 8,9 When intestinal anastomosis is not possible for anatomical reasons such as inability to bring the two remaining ends of the intestinal tract close enough together, or when anastomosis of the small intestine and anus or anal canal in the pelvis is not desirable for oncological reasons such as advanced cancer, a permanent ileostomy is created.
The reported overall complication rates after surgery for colorectal cancer in ulcerative colitis patients range from 9% to 65%, with infectious complication rates of 10%-45% and mortality of 0%-5.2%. 10,11 Patients with ulcerative colitis are often treated with a variety of medications and the status of their intestinal tract is worse than that in patients without ulcerative colitis. There are postoperative complications specific to patients with ulcerative colitis, including pouch-related complications. Patients with ulcerative colitis are also at high risk of colorectal cancer and for developing postoperative complications. 12,13 Several types of surgery are performed for patients with ulcerative colitis who develop colorectal cancer, including IAA, IACA, and permanent placement of a stoma.
However, there is limited information on the risk of complications for each surgical procedure in these patients.
Postoperative complications have been reported to worsen both short-term and long-term prognosis in various types of cancer. [14][15][16][17][18][19] Although there have been reports suggesting an association between postoperative complications and worse prognosis after surgery for colorectal cancer or liver metastases of colorectal cancer, 20,21 little is known about this association in ulcerative colitis patients with colorectal cancer. In this study, we examined the association between postoperative complications and prognosis in these patients according to type of surgical procedure.

| Study design and setting
Information on patients who had undergone upfront radical surgery for up to stage III colorectal cancer was extracted from the Japanese Society for Cancer of the Colon and Rectum (JSCCR) data on ulcerative colitis patients with gastrointestinal cancer or dysplasia between January 1983 and December 2020 and were treated at any of 43 participating hospitals. 22 Patients for whom there was no information on postoperative complications and those who underwent other types of surgery were excluded. The final study population included patients who had undergone any one of the three typical surgical procedures for ulcerative colitis comorbid with colorectal cancer, namely, total colorectal resection with ileoanal anastomosis (the IAA group), total colorectal resection with ileoanal canal anastomosis (the IACA group), or total colorectal resection with permanent placement of a stoma (the stoma group).

| Statistical analysis
Categorical variables are shown as the number and percentage and were compared using Fisher's exact test. Five-year OS and RFS were estimated using the Kaplan-Meier method and compared using the log-rank test. Risk factors for complications and their effect on 5- year RFS were identified using univariable and multivariable Cox proportional hazards models. Clinical factors with the potential to have a confounding effect on 5-year RFS were adjusted for in the multivariable model. All statistical analyses were performed using JMP statistical software version 15 (SAS Institute Inc.). All p-values were two-sided and those less than 0.05 were considered statistically significant.  Table 1.

| Patient characteristics
Compared with patients in the IAA and IACA groups, those in the stoma group were older, had a shorter history of ulcerative colitis, had a higher carcinoembryonic antigen level, were more likely to have rectal cancer, more poorly differentiated cancer, a deeper tumor depth, and more lymph node metastasis, and were more likely to receive adjuvant chemotherapy. There were more Stage III patients in the stoma group and patients with more advanced cancer had a greater likelihood undergoing permanent stoma creation.
The relation between presence or absence of complications and the administration of adjuvant chemotherapy in Stage III patients are shown in Table S1. Patients with postoperative complications were less likely to receive adjuvant chemotherapy than those without postoperative complications in IAA and IACA groups.

| Postoperative complications
The incidence of overall complications was not significantly different among the IAA, IACA, and stoma groups (32.7%, 32.3%, and 37.7%, respectively; p = 0.510). The incidence of infectious complications was significantly higher in the stoma group (21.2%) than in the IAA (12.9%) group and the IACA group (14.6%) (p = 0.048). Meanwhile, the incidence of noninfectious complications was higher in the IAA (21.6%) group and the IACA (16.2%) group than in the stoma group (13.7%), but the difference was not statistically significant (p = 0.088).
The incidence of anastomotic leakage tended to be higher in the IACA group than in the IAA group (6.2% vs. 3.3%; p = 0.135).
The surgical site infection (SSI) rate was significantly higher in the stoma group (8.9%) than in the IAA and IACA groups (2.9% and 1.5%, respectively; p = 0.004). The incidence of high output syndrome/ diarrhea/dehydration tended to be higher in the IAA group than in the IACA and stoma groups (2.7% vs. 0.8% and 0.0%, respectively; p = 0.051). There were no significant between-group differences in the rates of other complications. Details of complications are shown in Table 2.

| Recurrence patterns based on surgical technique
The recurrence patterns are shown in Table S2. Lymph node recurrence and lung metastasis were significantly more frequent in stoma group (4.8%, 5.5%, respectively) than in IAA group (0.6%, 1.2%, respectively) and IACA group (0.8%, 0.8%, respectively). The other recurrences were not significantly different among the three groups.

| Five-year OS and RFS
The 5-year OS was 93.3% in the IAA group, 94.5% in the IACA group, and 82.5% in the stoma group; the difference among the three groups was statistically significant (p = 0.001), as was the 5-year RFS (90.2%, 88.1%, and 75.5%, respectively; p < 0.001).       In the stoma group, the 5-year RFS was significantly higher in patients without noninfectious complications (79.4% vs. 50.9%, p = 0.008).

| Prognostic factors affecting 5-year RFS
Prognostic factors affecting 5-year RFS were investigated in each group (Table S3). Five-year RFS was significantly associated with pT and pN disease and adjuvant chemotherapy in all surgery groups.
Pathology findings were significantly associated in the IAA and stoma groups, and sex was significantly associated in the stoma group. Age, sex, pathology findings, pT and pN disease, and adjuvant chemotherapy were entered in the multivariable analysis to determine the prognostic effect of complications on 5-year RFS.

| Effect of overall complications on 5-year RFS
In the multivariable analysis, there was no significant difference in

| Effect of infectious complications on 5year RFS
In the multivariable analysis, infectious complications had no sig-

| Subgroup analysis of 5-year OS and 5-year RFS according to stage
Five-year OS and RFS according to stage were shown in Table 4. Overall, patients in the IAA group did not have a worse prognosis due to complications ( Figure 2). However, patients with complications in stage II or stage III tended to have worse 5-year OS and 5-year RFS than those without complications in the IAA group.

F I G U R E 2
Five-year overall survival and recurrence-free survival for each surgical procedure according to presence or absence of overall complications. IAA, ileoanal anastomosis; IACA, ileoanal canal anastomosis; OS, overall survival; RFS, relapse-free survival.

| DISCUSS ION
In this study, we investigated three common surgical techniques used in ulcerative colitis patients with colorectal cancer, namely, IAA, IACA, and stoma creation, and found differences in patient characteristics and the incidence of postoperative complications among these techniques. We also showed that postoperative complications may worsen the prognosis in these patients. Postoperative complications can worsen the prognosis, not only directly but also indirectly, by causing delays in the start of drug treatment. 23,24 The mechanism by which postoperative complications worsen prognosis is not fully understood; one possibility is that they are associated with inflammation, which in turn induces an immunosuppressive state. 15,21,25 Furthermore, postoperative complications can lead to general debility and failure to receive adjuvant chemotherapy or delay its start, which has been shown to have an adverse prognostic impact in patients with colorectal cancer. 26,27 In this study, patients with postoperative complications were less likely to receive adjuvant chemotherapy than    in order to collect a larger number of cases, we were unable to control for the prognostic impact of improved surgical methods or advances in drug treatment over time.

| CON CLUS ION
The  I  II  III  I  II  III  I  II  III 5-year overall survival