What is the rate of definitive stoma after subtotal colectomy for inflammatory bowel disease? A nationwide study of 1860 patients

Some patients with inflammatory bowel disease (IBD) require subtotal colectomy (STC) with ileostomy. The recent literature reports a significant number of patients who do not undergo subsequent surgery and are resigned to living with a definitive stoma. The aim of this work was to analyse the rate of definitive stoma and the cumulative incidence of secondary reconstructive surgery after STC for IBD in a large national cohort study.


INTRODUC TI ON
Management of inflammatory bowel disease (IBD) is multidisciplinary with an increasing role for immunosuppressive treatments.
Nonetheless a substantial proportion of patients still require surgery.Indeed, about 70% of Crohn's disease (CD) patients [1] and 15% of patients with ulcerative colitis (UC) will need surgery [2].
Subtotal colectomy (STC) with ileostomy is recommended as the safest option in cases of acute colitis refractory to medical treatment or complicated by perforation, toxic megacolon or severe haemorrhage [3].In the elective setting, in the event of refractory disease or malignancy, STC may be preferred as the first step in surgical treatment for high-risk patients presenting with severe disease, high doses of immunosuppressive treatment or poor nutritional status [4], or in cases of indeterminate colitis, as the pathological examination can change the initial clinical diagnosis of IBD [5].STC is a quick and easy procedure with a low morbidity and mortality rate [6], and it is technically feasible to perform it laparoscopically.The absence of an anastomosis avoids the threat of leakage in a typically high-risk population and preserves the option of a future return to intestinal continuity.
After STC, completion proctectomy with ileal pouch-anal anastomosis (IPAA) is the restorative procedure of choice for UC and for selected patients with pancolonic CD [3,7,8].It entails the removal of all colonic and diseased rectal mucosa and restores bowel continuity, avoiding a definitive stoma.An improved quality of life and a high level of satisfaction are reported with IPAA surgery [9,10].It can be realized as a two-stage (completion proctectomy and IPAA with no defunctioning ileostomy) or a three-stage procedure (with a loop ileostomy) [11].An ongoing trial comparing modified two-stage and three-stage procedures is currently recruiting participants [12].
Ileorectal anastomosis (IRA) can be offered to highly selected patients as an alternative to IPAA [13,14].
However, although proctocolectomy with IPAA is the recommended treatment, a significant number of patients do not undergo subsequent surgery after STC.Therefore, these patients are fated to live with a definitive stoma while their rectal remnant is still in place.While small single-centre retrospective studies describe rather low rates, between 10% [14] and 30% [15][16][17], of patients having a definitive stoma, two recent national cohort studies in Sweden and England revealed that up to 65% and 70% of patients, respectively, had a definitive stoma after STC with ileostomy for IBD [18,19].In a multicentre study, Aquina et al. reported a similarly low rate of only 32% of patients proceeding to an IPAA [20].
The aim of this study was to assess the rate of definitive stoma and reconstructive surgery after STC in a large national populationbased cohort and identify the risk factors associated with definitive stoma to take stock of the current practices in France and to understand the discrepancies reported in the literature.

Study design and population
We conducted a retrospective observational cohort study using the

Primary endpoint
The primary endpoint was secondary reconstructive surgery with either IPAA or IRA.Patients undergoing IPAA surgery after STC as a two-stage procedure (completion proctectomy and IPAA with no defunctioning ileostomy) were identified with the codes HJFA012 and HHFA031 for open procedures and HJFC023 and HHFA028 for laparoscopic procedures.Patients having a three-stage procedure Crohn's disease, definitive stoma, ulcerative colitis

What does this paper add to the literature?
Rates of definitive stoma after subtotal colectomy for inflammatory bowel disease are around 60%-70%.In this observational population-based study of 1860 patients, we showed that 67% of patients underwent secondary reconstructive surgery, mainly with ileal pouch-anal anastomosis, and only 33% had a definitive stoma.We identified age, Crohn's disease, neoplasia, laparotomy, postoperative complications and a low-volume hospital as risk factors for definitive stoma.
The interval from STC to reconstructive surgery was recorded.
Patients undergoing an abdominoperineal resection with definitive stoma were identified with the codes HHFA03, HHFA029, HJFA005, HJFA007, HJFA014 or HJFA019 without any codes for stoma closure.All codes are summarized in Appendix S1.

Covariate exposures
The following data were extracted and collected from the database: patient-related variables [age, gender, type of IBD, comorbidities, smoking, nutritional status (malnutrition, obesity), social characteristics], perioperative variables (emergency surgery, colorectal neoplasia, time between admission and surgery, surgical procedures, postoperative outcomes, length of stay) and institution-related variables (hospital type, annual caseload).
The Charlson Comorbidity Index (CCI) score (0-37) and French index of social deprivation (FDep) score (increasing with the importance of social deprivation) were calculated for each patient (Appendices S2 and S3).Hospital annual volume was calculated regarding the activity of colorectal surgery as follows: number of colectomies for cancer performed per year per hospital during the study period.Hospital volumes were categorized according to the number of colectomies for cancer as low-volume (<50 colectomies/year), medium-volume (50-150 colectomies/year) or highvolume hospitals (>150 colectomies/year).Statistical significance was considered as p < 0.05.R software was used to perform all analyses [26].Data are reported following STROBE and RECORD recommendations [27,28].

Population characteristics
A total of 1860 patients undergoing a STC for IBD between 2013 and 2021 were included.Demographic and perioperative data are reported in Table 1.The mean age was 45.5 ± 18.9 years at the time of STC.The number of male patients was 1040 (56%).Median follow-up was 30 months (range 10-61 months) after STC.UC was the most frequent disease, and affected 1442 patients (77%), whereas 418 patients (23%) presented with CD.A total of 1261 (68%) patients were malnourished.Three hundred and twenty-one STC procedures TA B L E 1 Demographic and perioperative data for 1860 patients with inflammatory bowel disease who underwent subtotal colectomy (STC).(17%) were performed as an emergency and 68 (4%) for neoplasia; 657 (35%) STCs were performed more than 7 days after the date of admission.The majority of STCs (72%) were performed via a laparoscopic approach.Sixty-nine per cent of patients were operated on in high-volume hospitals, 24% in medium-volume hospitals and 7% in low-volume hospitals.
On univariate analysis (

Risk factors associated with definitive stoma
Figure 3 shows the probability of reconstructive surgery according to hospital volume and surgical approach.In multivariate Cox regression analysis (

DISCUSS ION
This observational population-based study reports the outcome following STC for IBD over a median follow-up of 30 months in a large national cohort.We showed that 67% of patients underwent secondary reconstructive surgery with a mean interval after STC of 7 months.Proctectomy with IPAA was the most common surgical procedure (74%), while 26% had an IRA.Among patients with definitive stoma (33%), 19% had an abdominoperineal resection whereas the majority (81%) did not undergo any further surgical procedure after STC.We identified older age, CD, colorectal neoplasia, postoperative complications after STC and low-volume hospital as risk factors for definitive stoma.STC performed by laparoscopy was significantly associated with a lower risk of definitive stoma.
We observed a 90-day postoperative mortality rate of 5% and a morbidity rate of 47% after STC.These unexpectedly high rates contrast with previous studies from specialized hospitals which describe a mortality rate under 2% [6,29].This notable difference can be explained by the pragmatic nature of our study, reflecting real-life practice, and are distinguished from the experience of tertiary specialized centres alone.This important morbidity and mortality after STC reflect the fragility of IBD patients at the time surgical intervention is required.Indeed, in our cohort more than two-thirds of patients were malnourished and 17% of STCs were performed in an emergency setting.Moreover, we determined that age above 65 years and comorbidities were independent risk factors for postoperative mortality.Interestingly, we observed that 35% of STCs were performed more than 7 days after admission, suggesting a long duration of medical therapy before choosing surgical treatment in cases of acute or emergent colitis: it is known that delayed surgery is associated with increased postoperative complications [30].
A laparoscopic approach seems to play an important role in decreasing postoperative mortality and likely morbidity.Faster return of bowel function, lower risk of wound infection and intra-abdominal abscess and shorter length of stay are among the benefits of laparoscopic STC reported in the literature [31,32].We observed a significant difference in 90-day postoperative mortality between open and laparoscopic approaches (16% vs. 1%).However, these results must be considered with caution, as mortality could also be related to the severity of the underlying condition that explained the choice of open surgery (colonic perforation or important comorbidities contraindicating laparoscopy, for example).Nevertheless, the majority of STCs (72%) were performed via a laparoscopic approach and laparoscopy remained associated with lower odds of mortality after adjustment.
We found that medium-and low-volume hospitals were independent risk factors for postoperative mortality.Kaplan et al. also reported an inverse correlation between postoperative morbidity and mortality after STC and annual hospital volume [33].Hospital volume can influence postoperative outcomes by gathering multidisciplinary management with experienced clinicians and trained surgeons providing intense medical therapy but timely surgery, improved patient selection and high-quality perioperative care.Thus, referring patients for STC, when possible, to more specialized hospitals should be considered.
We chose to focus on the rate of secondary reconstruction after STC, and thus did not include patients undergoing primary reconstructive surgery (IPAA or IRA at the same time of colectomy) in order to elucidate why some patients do not proceed to this recommended secondary procedure and have a definitive stoma.In this study, we observed a large proportion of patients accessing secondary reconstructive surgery (67%), distinct from reports in recent national studies.Nordenvall et al. [18] and Worley et al. [19] stated that only 34% and 20% of patients had secondary restorative surgery in Sweden and England, respectively, while Aquina et al. [20] reported a low rate of 32% of patients being offered a secondary IPAA in a multicentre study conducted in New York State.A higher rate ranging from 66% to 86% is observed in small single-centre retrospective studies carried out in in specialized hospitals [15][16][17].This variable access to secondary reconstructive surgery can be partly explained by two modifiable risk factors: hospital volume and laparoscopic approach.
We observed that low hospital volume was independently associated with the risk of definitive stoma, as has been previously noted in the Swedish cohort and in the Aquina et al. multicentre study [18,20].In our cohort, 69% of patients were operated on in high-volume hospitals.Comparatively, in Sweden only 27% underwent STC in high-volume hospitals and 65% of the procedures Moreover, laparoscopic expertise is required for IPAA surgery as it provides better short-term postoperative outcomes and a limited impact on fertility compared with open surgery in this relatively young population [22,[39][40][41][42]. Considering these arguments, patients should be referred to high-volume hospitals for STC, and where this is not feasible they should be referred after STC for secondary reconstruction.Our study reinforces the need for centralization of IBD surgery to improve patient care and suggests a need for national or international directives to defined expert centres.
We observed that a laparoscopic approach for STC was associated with a higher probability of undergoing reconstructive surgery.
Furthermore, patients who undergo laparoscopic STC progress more rapidly to completion proctectomy with IPAA, and adhesiolysis is performed less often and less extensively when compared with open STC [32,41].Thus, the laparoscopic approach appears to have a significant impact on short-term postoperative outcomes that may lead to faster recovery and a higher probability of achieving a secondary restorative surgery, that may be performed earlier.In our study, 72% of patients had a laparoscopic STC whereas this accounted for less than 2% of Swedish patients, which could partly explain the gap between reconstruction rates.It should be noted that the Nordenvall et al. study [18] was carried out before our study, and that surgery and surgical techniques have evolved between these periods.
Discrepancies in surgical management of IBD between European countries can be described.While only 26% of patients undergoing restorative surgery had an IRA in our cohort, higher rates were observed in Sweden, where it is the preferred reconstruction procedure (54%) [18].IRA can be considered as a reasonable alternative for selected patients, with fewer postoperative complications.However, it is associated with a higher long-term failure rate than IPAA, related to the poorer functional results and neoplasia of the rectum that require regular endoscopic monitoring [14,43,44].For these reasons IRA is not the restorative procedure of choice among French surgeons.The high rates of definitive stoma in English and Swedish series compared with France could also be explained by better stoma acceptance in northern populations, which is illustrated by the relatively high rate of abdominoperineal resections in the management of rectal cancer in these countries [45].However, studies depict satisfactory functional outcomes and quality of life with IPAA, and an improvement in sexuality, work and social function when compared with a definitive stoma, which should encourage IPAA surgery if possible [9,10,46].
Even where IPAA is the recommended surgical treatment, some patients are unsuitable for surgery to restore bowel continuity.
Firstly, we identified several risk factors for definitive stoma related to the frailty of these patients (older age, colorectal neoplasia and postoperative morbidity after STC) that could explain the absence of reconstructive surgery, in conjunction with poor physical condition and/or reduced anal function.Secondly, patients with CD were more likely to have a definitive stoma, potentially due to the impossibility of performing restorative surgery in the presence of severe perianal or small bowel disease, and the higher risk of IPAA failure [47].In a subgroup analysis restricted to patients with UC, the association between patient frailty, hospital volume and surgical approach and the risk of definitive stoma remained evident.
In our cohort, 19% of patients with a definitive stoma required an abdominoperineal resection, suggesting that their disease did not permit the restoration of bowel continuity.Unexpectedly, however, most patients (81%) had a definitive stoma related to the absence of a subsequent surgery after STC.This rate raises questions about patient management after STC.Some patients might not be offered reconstructive surgery because they have no medical follow-up or because they are not referred onwards to surgeons to complete the surgery.
Some might be ineligible or not keen for reconstruction and thus do not proceed to completion proctectomy, leaving them at risk of developing persisting symptoms or neoplasia [48][49][50].Regular endoscopic monitoring of the retained rectum is necessary, but fewer than half of patients are reported to be compliant [9].In long-term studies, 7% to 25% of patients with the rectum in place after STC will require completion proctectomy as a consequence of persistent symptoms or neoplasia [51,52].Even where reconstructive surgery is unlikely, patients should proceed to completion proctectomy, especially when the indication for STC was colorectal neoplasia.
The major strengths of this analytic study are its large size, its long length of follow-up and its population-based nature allowing the inclusion of 1860 patients with limitation of selection bias and reflecting real life practice.This national cohort based on systematic routinely collected data provides accurate information about current practice in France.Nevertheless, several limitations are inherent to the study.
This registry study is dependent on the quality of coding, as incorrect or missing data may lead to selection or misclassification bias.Coding of surgical procedures is expected to be accurate since it is realized with remuneration purposes for surgeons.We also chose to study postoperative mortality as it is an objective and robust outcome.
Our study is limited by the lack of clinical or histological data, such as disease activity, disease duration before surgery or immunosuppressive medical therapies, that could influence postoperative outcomes.
Moreover, technical surgical details such as management of the rectal remnant (rectal stump closed intra-abdominally or double-end ileosigmoidostomy) are not available and could have an impact on restorative surgery.Indeed, double-end ileo-sigmoidostomy seems to be associated with a lower rate of conversion [52].

CON CLUS ION
In this large national population-based study, 67% of patients undergoing STC with ileostomy for IBD subsequently underwent reconstructive surgery.This rate is higher than previously reported in other population-based studies, but a substantial proportion of patients remain with definitive stoma due to the absence of a further surgical procedure after STC (81%).Definitive stoma is associated with older age, Crohn's disease, colorectal neoplasia, postoperative morbidity, laparotomy and low-volume hospital.Patients should be referred to specialized high-volume hospitals for STC and secondary reconstruction.

French
national registry database Programme de Médicalisation des Systèmes d'Information (PMSI).The PMSI routinely collects and gathers standardized administrative and medical information concerning every medical structure (public and private) in France.Public health care services are equally available to the entire French population.Data about demographic characteristics, principal and associated discharge diagnosis codes (based on the International Classification of Diseases 10th edition, ICD-10), medical procedure codes (based on the French national Classification Commune des Actes Médicaux, 11th edition), discharge date, length of stay and hospital identifiers are available for each inpatient admitted.The validity of this prospective coding system has been tested by crossreferencing it with other cohort databases as previously described [21-25].Access to the data was approved and authorized by the Commission Nationale de l'Informatique et des Libertés (CNIL).All patients undergoing STC between 2013 and 2021 were identified with the codes HHFA021 for open and HHFA005 for laparoscopic procedures.Patients were included if the surgery was performed for UC or CD, identified with the codes K50 and K51.Patients younger than 15 years at the time of STC were excluded from the analysis.

F I G U R E 2
Cumulative incidence curve representing time to reconstructive surgery after subtotal colectomy in 1860 patients with inflammatory bowel disease.F I G U R E 1 Flow-chart representing the outcome following 1860 subtotal colectomies for inflammatory bowel disease (IBD) over a median follow-up of 30 months.
described in the Aquina et al. multicentre study were performed in low-volume hospitals.IPAA is a technically demanding surgery requiring experience in colorectal surgery.An inverse correlation between IPAA surgical experience and postoperative morbidity is reported, and it is observed that low institutional volume is associated with higher risk of IPAA failure and reintervention [34-38].

F I G U R E 3
(A) Kaplan-Meier curves stratified by hospital volume of colectomies for cancer per year (low <50; medium 50-150; high >150), representing time to reconstructive surgery after subtotal colectomy in 1860 patients with inflammatory bowel disease.(B) Kaplan-Meier curves stratified by surgical approach (laparoscopy or laparotomy), representing time to reconstructive surgery after subtotal colectomy in 1860 patients with inflammatory bowel disease.

day mortality Univariate analysis Multivariate analysis Yes (n = 100) No (n = 1760) Odds ratio (95% CI) p-value Adjusted odds ratio (95% CI) p-value
TA B L E 2 Univariate and multivariate analysis of 90-day postoperative mortality in 1860 patients who underwent subtotal colectomy (STC) for inflammatory bowel disease.90-Note:FDep indicates French index of social deprivation.Odds ratios indicate relative likelihood for definitive stoma and were adjusted for all variables included.Hospital volume is given in the number of colectomies for cancer per year.
Multivariate analysis of risk factors for definitive stoma in 1860 patients who underwent subtotal colectomy (STC) for inflammatory bowel disease (IBD).
TA B L E 3Note: Hazard ratios were adjusted for all variables included.FDep indicates French index of social deprivation.Hospital volume is given in the number of colectomies for cancer per year.