This review has been created in line with the Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) statement. The protocol was developed and registered to PROSPERO prior to analysis (PROSPERO CRD42023473654).

Studies were eligible if they included adults (≥18 years) with left-sided UC (E1–E2, as defined by the Montreal classification) refractory to medical therapy who underwent surgery. Eligible interventions included TAC or any alternative surgical procedure for left sided colitis. Comparators could include other surgical approaches or standard medical management. Primary outcomes were patient-centred outcomes such as quality of life and satisfaction with surgery. Secondary outcomes included short- and long-term surgical outcomes such as length of hospital stay, morbidity, mortality, recurrence, bowel function, and nutritional status. Exclusion criteria included studies that involved Chron’s disease, pharmacological management only, isolated proctitis, rectal cancer, non-human subjects, or non-English publications.

A comprehensive search was conducted across Ovid MEDLINE, EMBASE, the Cochrane Library, Web of Science, and Scopus. The search covered studies published from 2001 (marking the first use of infliximab for steroid-refractory UC) to 1 November 2023. Only studies in English and involving human participants were included.

A comprehensive literature search was developed with the support of an expert in literature searching. It was conducted across the following databases: Ovid MEDLINE and EMBASE ( Figure 1), the Cochrane Library ( Figure 2), Web of Science ( Figure 3), and Scopus ( Figure 4). The publication period was limited from 2001 (marking the first pilot study of infliximab use for steroid-refractory ulcerative colitis ¹¹ ) to 1 November 2023, when the search was performed. Other restrictions included the English language and studies involving human participants.

Key search terms covered disease characteristics (‘ulcerative colitis’, ‘left-sided colitis’, ‘colitis gravis’, ‘idiopathic proctocolitis’, ‘inflammatory bowel disease’), treatment methods (‘surgical management’, ‘surgical treatment’, ‘surgical intervention’, ‘colectomy’, ‘subtotal colectomy’, ‘loop colostomy’, ‘sigmoidectomy’), and patient outcomes (‘quality of life’, ‘postoperative complications’, ‘surgical outcome’, ‘recurrence’, ‘length of stay’, ‘hospital stay’). The search strategy was tailored to each database to optimise retrieval. Full search strategies are provided in Figures 1– 4.

Inclusion and exclusion criteria were developed using the PICO framework (Population, Intervention, Comparison, Outcomes) and aligned with PRISMA guidelines ¹² to guide the systematic selection of studies. Population (P): adult (>18 years old) patients diagnosed with left-sided colitis (as per Montreal classification ⁴ ) caused by UC, refractory to medical treatment, who underwent surgical management. Intervention (I): standard surgical management of UC patients - TAC. Comparison (C): any other type of surgical management used for left-sided colitis. Primary outcomes (O): patient-focused outcomes (eg., quality of life, satisfaction with surgery). Secondary outcomes included short-term and long-term surgical outcomes (e.g., length of hospital stay (LOS), morbidity rates, mortality rates, recurrence rates, bowel function, nutritional status, etc). Patients who only had pharmacological interventions or had isolated proctitis or rectal cancer were set to be excluded from the study.

A narrative review of those papers that did not fulfil the PICO criteria, but had relevance to answering the key question was also undertaken. For example, if the paper involved surgical interventions on patients with left-sided colitis but in the elective setting (not emergency), we decided to mark such manuscripts and undertake a commentary on their findings.

Screening of the studies (title, abstract) was independently conducted by three independent reviewers (JP, IH, MJ) employing the software Rayyan. ¹³ Duplicate records were subsequently removed. Discrepancies between the studies in the screening stage were discussed and adjudicated with a senior reviewer (GR). In the full-text screening, the process was repeated utilising the same principle where a senior reviewer was involved in resolving the discrepancies. The PRISMA flow chart is available as Figure 5.

Data extraction was planned to be conducted by three independent reviewers (JP, IH, MJ). In the event of discrepancies, a senior reviewer (GR) was to be involved in resolving disagreements through discussion. All studies were to be evaluated for eligibility prior to data extraction. Subsequently, data from eligible studies would have been extracted by two independent reviewers using a predefined Excel form. Extracted data would have included study population. If necessary, missing data would have been sought by contacting study investigators for unreported information or additional details.

Data were collected on study characteristics, patient demographics, disease extent, type of surgical procedure, comparator, and outcomes.

Primary outcomes comprised patient-centred measures such as quality-of-life post-surgery. Secondary outcomes comprised peri-surgical and long-term outcomes including morbidity, mortality, recurrence, and bowel function.

The planned assessment of selected studies was to employ the GRADE system (Grading of Recommendations Assessment, Development, and Evaluation), ¹⁴ which would evaluate study limitations, inconsistency, indirectness, imprecision, and publication bias. Following the same principles for article screening, three reviewers would have applied these criteria to each study and outcome, with another reviewer resolving discrepancies.

For comparative studies, dichotomous data were to be summarised using odds ratios or risk differences, and continuous data using mean differences.

Due to the absence of eligible studies, quantitative synthesis was not performed.

A meta-analysis was planned if sufficient homogeneous data were available. In its absence eligible randomised controlled trials, a narrative synthesis was conducted to summarise available findings and highlight gaps in the literature.

Assessment of reporting bias, including publication bias and selective reporting, was planned using funnel plot analysis and protocol comparison where applicable. This was not performed because no eligible studies met inclusion criteria.

The overall certainty of evidence for each outcome was to be assessed using the GRADE framework. As no studies met inclusion criteria, this assessment was not undertaken.

The study protocol was registered on the PROSPERO website ¹⁵ prior to completing the review (CRD42023473654).

A total of 6,606 records were identified through database searching. After removing duplicates and screening titles and abstracts, 127 full-text articles were assessed for eligibility. No studies met the inclusion criteria for patients with acute left-sided ulcerative colitis undergoing surgery. Six studies examining surgical management of UC in related contexts (such as elective surgery or non-acute settings) were narratively reviewed. The study selection process is summarised in the PRISMA flow diagram ( Figure 5).

Despite a systematic approach exploring as comprehensive a search as possible we were unable to identify any suitable papers that were eligible for inclusion. There were, however, six papers that did explore the surgical management of left-sided colitis in other settings, such as elective presentations or in patients without acute flares ( Table 1). We therefore continued to undertake a narrative review of this literature.

Zaharie et al. ¹⁶ is a retrospective single centre retrospective study that discussed the surgical management of 221 IBD patients. A total of 33 (14.93%) patients underwent surgical management, including 16 UC patients. Patients were grouped into those with (n = 3) or without (n = 13) proctectomy. The commonest indication for surgery was the failure of medical management (n = 7). Authors reported no complications or postoperative morbidity. While surgical management and postoperative outcomes were stratified by disease extent, the study did not compare outcomes for different surgical interventions or whether surgeries were performed for acute colitis, leading to its exclusion ( Table 1).

Gu et al. ¹⁷ conducted a retrospective cohort study on 204 UC patients undergoing laparoscopic colectomy, comparing outcomes between intraperitoneal (n = 99) and subcutaneous (n = 105) placement of the rectal stump. Of these, 104 (55%) had TAC for limited left-sided colitis. Only one (0.5%) postoperative death occurred in the subcutaneous group, and there were no significant differences in reoperation rates or overall morbidity. Again, the study did not provide a specific subgroup analysis for left-sided colitis or focus on outcomes related to acute colitis presentations, limiting its relevance to our review ( Table 1).

Jain et al. ¹⁸ retrospectively assessed colectomy rates in patients with left-sided colitis compared to extensive colitis following an index episode of ASUC in a single centre. Out of a cohort of 2,076 patients with UC, 241 (11.6%) experienced ASUC, with 34 (14.1%) requiring colectomy during the index admission. Among those with ASUC, 85 (35%) presented with left-sided colitis at the time of hospitalisation. Colectomy rates were similar for left-sided colitis (10 [11.9%]) and extensive colitis (24 [15.4%]). No significant differences occurred in long-term outcomes, including colectomy-free survival rate, mortality, readmission rates (due to repeat ASUC), the number of relapses per year or duration of remission between the two groups ( Table 2). This study assessed colectomy rates in patients with an acute UC and provided a comparison of long-term outcomes based on the disease extent. However, it did not compare different surgical interventions and focuses solely on colectomy as an outcome, thus falling outside the scope of our review ( Table 1).

Carpenter et al. ¹⁹ utilized the American College of Surgeons – National Surgical Quality Improvement Program database to compare postoperative outcomes of partial colectomy (PC) and TAC in unmatched (n = 9,888) and matched (n = 1,856; 1:1) cohorts. Post-matching, patients undergoing TAC had higher overall 30-day post operative complications (41.9% versus 36.5%, P = 0.017) and serious complications (37.2% versus 31.5%, P = 0.011). Although TAC patients suffered from more postoperative VTE (6.4% versus 3.6%, P < 0.005), rates of most other individual outcomes were similar between the groups. When restricted to emergency settings, there was no significant difference in outcomes between TAC and PC ( Table 2). Despite reporting the comparison of outcomes between PC and TAC, this study was excluded as it did not stratify outcomes by disease extent and included elective surgeries ( Table 1).

Frontali et al. ²⁰ conducted a multicentred retrospective study reporting the outcomes associated with segmental colectomy in 72 UC patients without active colitis patients. Early (<30 d) postoperative complications occurred in 17 (24%) patients, including three (4%) postoperative deaths. No cases of ASUC were reported during that timeframe. Five patients (7%) experienced early UC flares within three months, all requiring reoperation. Late UC flares occurred in 29 (45%) patients, with 12 (41%) needing surgical intervention. Over a median follow-up of 40 months, 24 (35%) patients underwent re-operation, primarily for refractory colitis (n = 14) ( Table 2). Although disease extent was mentioned, postoperative outcomes were not stratified by the disease extent, the operations were mostly elective, and surgeries were not performed for refractory medical treatment, making this study irrelevant to our review ( Table 1).

McKenna et al. ²¹ evaluated 3,763 UC patients for 30-day major morbidity following semiurgent total proctocolectomy (TPC) with IPAA (n = 101), semiurgent STC (n = 797), and elective TPC with IPAA (n = 2,865). Major morbidity was higher in semiurgent TPC with IPAA (28%) compared to elective TPC with IPAA (22%) (P = 0.04), with the largest differences seen in organ space infection (19% vs 8%, P < 0.01) and postoperative sepsis (16% vs 7%, P < 0.01) ( Table 2). No significant difference was observed in major morbidity between semiurgent TPC with IPAA and semiurgent STC (28% vs 22%, P = 0.19), though organ space infections were more frequent in the TPC group (19% vs 9%, P < 0.01). Although this study provides outcomes based on surgery type in semiurgent settings, it does not stratify results by disease extent, limiting its applicability to our focus on acute left-sided colitis ( Table 1).

No eligible studies addressed surgical management in acute left-sided colitis. The related studies reviewed provided indirect evidence, primarily from elective or non-acute settings. Collectively, they showed similar postoperative outcomes between different colectomy types, with no clear advantage of one surgical approach for left-sided disease.

As no studies met inclusion criteria, formal risk of bias assessment was not undertaken. Given the absence of eligible studies, assessment of reporting bias and grading of the certainty of evidence using GRADE could not be performed.