Impact of anastomotic leakage on long-term prognosis after colorectal cancer surgery

Abstract

Colorectal cancer (CRC) is one of the most common malignancies in the world. Despite significant improvements in surgical technique, postoperative complications still occur in a fair percentage of patients undergoing colorectal surgery. The most feared complication is anastomotic leakage. It negatively affects short-term prognosis, with increased post-operative morbidity and mortality, higher hospitalization time and costs. Moreover, it may require further surgery with the creation of a permanent or temporary stoma. While there is no doubt about the negative impact of anastomotic dehiscence on the short-term prognosis of patients operated on for CRC, still under discussion is its impact on the long-term prognosis. Some authors have described an association between leakage and reduced overall survival, disease-free survival, and increased recurrence, while other Authors have found no real impact of dehiscence on long term prognosis. The purpose of this paper is to review all the literature about the impact of anastomotic dehiscence on long-term prognosis after CRC surgery. The main risk factors of leakage and early detection markers are also summarized.

Core Tip: Colorectal cancer (CRC) is one of the most common malignancies in the world. Despite significant improvements in surgical technique, postoperative complications still occur in a fair percentage of patients undergoing colorectal surgery. The most feared complication is anastomotic leakage. It negatively affects short-term prognosis, with increased post-operative morbidity and mortality, higher hospitalization time and costs. Moreover, it may require further surgery with the creation of a permanent or temporary stoma. While there is no doubt about the negative impact of anastomotic dehiscence on the short-term prognosis of patients operated on for CRC, still under discussion is its impact on the long-term prognosis. Some authors have described an association between leakage and reduced overall survival, disease-free survival, and increased recurrence, while other authors have found no real impact of dehiscence on long term prognosis. The purpose of this paper is to review all the literature about the impact of anastomotic dehiscence on long-term prognosis after CRC surgery. The main risk factors of leakage and early detection markers are also summarized.

INTRODUCTION

Definition, incidence and classification

Anastomotic leakage (AL) is a major cause of postoperative morbidity and mortality after colorectal cancer (CRC) surgery. AL is a defect of the intestinal wall integrity at the colorectal or colo-anal anastomosis site (including suture and staple lines of neorectal reservoirs) leading to a communication between the intra- and extraluminal compartments[1]. However, there are several definitions of AL in literature and most studies define it using clinical signs (pain, fever, tachycardia, peritonitis, purulent or fecal drainage), radiographic findings (fluid and/or gas-containing collections), and/or intraoperative features (peritoneal effusion and ruptured anastomosis)[1,2]. The use of different definitions in clinical studies can partly explain the considerable variations in AL reported rates. The incidence of AL reported in different studies is highly variable (2%-19%) and certainly influenced first of all by the surgeon's experience and the emergency or elective surgical setting. It is also influenced by the site of the anastomosis: It is lowest for ileocolic anastomoses (1%-3%) and highest for coloanal anastomoses (10%-20%)[3-5].

AL has been divided into "early" and "late" depending on whether AL is diagnosed within or after 30 d after surgery[6]. In general, early AL manifests with severe peritonitis and it is mainly related to a technical error in performing the anastomosis, usually due to mal vascularization of the intestinal stumps or tension at the anastomotic site[7]. In contrast, late AL is often associated with long-standing pelvic abscess[8] and is due to preexisting conditions in patients, such as local sepsis, poor nutrition, immunosuppression, morbid obesity, and radiation exposure[9].

AL is also classified according to severity into grade A, B and C. Grade A is represented by AL that does not require active therapeutic intervention, grade B by AL that requires active therapeutic intervention but manageable without re-laparotomy, and grade C by AL that requires re-laparotomy[1].

RISK FACTORS

Several risk factors for anastomotic dehiscence following colorectal surgery have been identified over the years. They can be classified for convenience into preoperative, intraoperative, and postoperative[10].

Preoperative risk factors commonly reported in the literature include male sex[7], obesity[11], tobacco habit, alcohol consumption, an American Society of Anaesthesiologistscore of 3 or higher[6], and prolonged corticosteroid intake[12]. Tumor location, size and stage must be considered among the risk factors. Akiyoshi et al[13] reported that tumor localization in the rectum, rather than the colon, was independently predictive of AL development on multivariate analysis.

The AL rate was 10 times higher (20.6% vs 2.3%) when the anastomotic region was located within 5 cm of the anal verge[14].

Low anterior resection (LAR) involves surgery in an anatomically confined space and when tumor size and/or stage increases, intrapelvic manipulation becomes limited and rectal dissection more challenging. In a series of 154 patients with rectal carcinoma, tumor size ≥ 5 cm in diameter was associated with a 4-fold increased risk of leakage[15]. Zhu et al[16] found that tumors greater than 3 cm in diameter, as well as TNM stage, were independently associated with leakage.

Intraoperative risk factors include: The surgeon experience (and hospital size)[7], the number of linear stapler firings[7], left colic artery ligation[17], emergency surgery (patients with peritonitis and/or bowel obstruction are at higher risk of postoperative adverse events)[18], operative time[19] and blood loss during surgery. Intra-operatively, it is also important to ensure good vascularization of the anastomosed bowel segments. Indocyanine green (ICG) fluorescence angiography may help in this evaluation. In a recent meta-analysis, an incidence of anastomotic dehiscence was observed in 3.8% of cases in the ICG group and 7.8% in the control group in which ICG was not used[20].

Postoperative risk factors are anemia, hypoalbuminemia, and late initiation of enteral nutrition[21].

EARLY DETECTION AND MARKERS

Early detection of AL is crucial to treat patients limiting negative effects. Baeza-Murcia et al[22] analyzed the accuracy of C-reactive protein (CRP) and procalcitonin (PCT) for early detection of AL and have found that CRP is more accurate than PCT on both postoperative day (POD) 3 and 5. According to this study CRP measured on POD 5 is the most useful test for early diagnosis of AL and that values above 9.1 mg/dL are indicative of anastomotic dehiscence.

In a recent meta-analysis by Yeung et al[23] a CRP cutoff level of 14.8 mg/dL at POD 3 had a sensitivity and specificity of 95%, while CRP cut-off levels of 12.3 mg/dL at day 4, 11.5 mg/dL at day 5, 10.5 mg/dL at day 6, and 9.6 mg/dL at day 7 had a sensitivity and specificity of 100% for anastomotic dehiscence.

According to Garcia-Granero et al[24] and El Zaher et al[25], PCT is also a very good predictor of anastomotic dehiscence, particularly from POD 5 or higher. The predictive power of PCT may also be enhanced in combination with CRP or white blood cell, or both (area under the curve 0.92, 0.92, 0.93, respectively)[25]. A recent meta-analysis by Xu et al[26] shows that PCT at POD 3 has potential clinical value in the early diagnosis of AL and has better diagnostic accuracy in patients undergoing laparoscopic surgery. Cut-off values are recommended in the range of 0.7-1.3 ng/mL to ensure accurate diagnosis and safe discharge. However, PCT is a valid predictor only for patients with major clinical losses confirmed by radiology and presenting with severe clinical signs and symptoms that require a change in therapeutic management and in most cases a reintervention. Cousin et al[27] conducted a meta-analysis and concluded that PCT does not add value to CRP in the diagnosis of AL.

It can be said that CRP and PCT at POD 5 have a high negative predictive value, which would allow early and safe discharge.

Tavernier et al[28] considered 5 criteria for safe early discharge after laparoscopic colorectal surgery: A CRP level of less than 15 mg/dL, absence of fever during the entire hospital stay (temperature < 38 °C), return of bowel function (flatus with or without stool), adequate pain control with oral analgesics (pain less than 5 out of 10 on a 10-point visual analog scale) and tolerance of a solid diet. The negative predictive value in ruling out an anastomotic leak was 98.4% for all 5 criteria combined. The false-negative rate was 13.3%.

RELATIONSHIP BETWEEN AL AND SHORT-TERM PROGNOSIS

AL affects the outcome of surgery, worsening the short-term outcomes and increasing the time and cost of hospitalization[29,30]. The mortality related was reported to be between 0.8% and 27%[31]. Mortality was higher after leak from a colonic anastomosis than after leak from a rectal anastomosis (43.8% vs 7.1%)[31]. Bertelsen et al[32] found in a multicenter study a 4-fold increase in 30-d mortality in patients with AL[32]. According to a Cochrane review, AL is associated with a perioperative mortality rate of 2% to 24% and high morbidity, with the risk of a definitive ostomy exceeding 25%[33]. Warps et al[34] found an overall AL rate of 4.8%, ranging from 4.0% (right hemicolectomy) to 15.4% (subtotal colectomy). AL was predominantly managed with reintervention, ranging from 81.2% of cases after transversectomy to 92.4% after sigmoid resection. After reintervention, the highest mortality rates were observed for transversectomy (15.4%) and right hemicolectomy (14.4%) and the lowest for sigmoid resection (5.6%) and subtotal colectomy (5.9%). The intensive care unit admission rate was 62.6% overall (range 56.7%-69.2%) and the stoma rate ranged from 65.5% (right hemicolectomy) to 93.0% (sigmoid resection).

RELATIONSHIP BETWEEN AL AND LONG -TERM PROGNOSIS

While the short-term consequences of AL are well known, its impact on long-term prognosis in CRC patients is still debated.

In the literature, the first authors to concern themselves with outcomes related to anastomotic dehiscence after resective surgery for CRC were Phillips et al[35] and Sauven et al[36]. In both cases, the parameter evaluated was local recurrence (LR). In the first study, AL did not appear among the significant risk factors for recurrence, while in the second, anastomotic dehiscence was associated with an increased rate of LR. In the same years, Amato et al[37] evaluated the association between CRC and AL by focusing exclusively on patients with rectal tumors operated with an anterior resection. In this study, AL did not influence the recurrence rate.

In 1991, Akyol et al[38] performed a study on patients operated for left colon or rectal cancer and demonstrated an important influence of AL on recurrence and cancer-specific survival (CSS) at 24 mo. The independence of the impact of dehiscence on outcomes from tumor stage was highlighted. This was the first study that analyzed local and distant recurrence separately and used multivariate Cox regression.

Two years later, a study published by Fujita et al[39] showed the impact of AL on LR and disease-free survival (DFS). DFS is significantly lower in the AL group for patients with Duke stage A and B cancers but not for C and D. The importance of this work also lies in the separate evaluation of subjects with colon and rectal cancer.

Petersen et al[40] studied the influence of leakage on LR, CSS, overall survival (OS) and postoperative mortality. AL influence only LR and CSS, confirming the previous findings of Akyol et al[38]. Branagan et al[41] reached similar conclusions in 2005. Further studies[42-44] showed a correlation between AL and higher 30-d mortality, lower OS and CSS.

Law et al[45] in 2007 found a significant association between AL and 5-year CSS, 30-d mortality and recurrence (local and systemic).

According to the study by Eberhardt et al[46], AL does not change the risk of recurrence and mortality for colon cancer, whereas it does for rectal cancer. The article also offers an assessment of OS, CSS, LR and overall recurrence for each stage, as well as an analysis of these outcomes at both 1 and 5 years after surgery.

According to Marra et al[47], AL significantly reduces OS without affecting the risk of recurrence, while other studies[48-58] have found an impact of leakage on OS, recurrence, and DFS. However, Katoh et al[50] evaluated only patients with stage II CRC and Breugom et al[54] only patients with stage I-III colon cancer. To be precise, Park et al[56] in 2016 found an effect of AL on OS and DFS only for patients with rectal cancer. Nachiappan et al[52] found a reduction in OS in patients with AL who required reoperation compared with subjects without AL. Ramphal et al[58] demonstrated that LR develops with the same frequency in symptomatic and asymptomatic dehiscence.

Krarup et al[59-60] identified in patients with AL an increase in distant recurrence (DR) and in mortality. However, there was no significant association with LR. Nordholm-Carstensen et al[61] and Ng et al[62] evaluated the impact of AL in patients with stage IV CRC. The 3-year survival rate is affected by dehiscence for both colon (18.7% vs 44.6%) and rectum (53.7% vs 73.3%).

The first meta-analysis on this topic was performed by Mirnezami et al[63] on 22 studies. It reported an association between AL and LR, DR and cancer-specific mortality.

The subsequent meta-analysis by Ha et al[64] evaluated 34 studies and divided the results into two categories. In the first group rectal anastomosis data were analyzed, and AL was associated with increased LR and reduced OS, CSS, and DFS. There were no significant effect on distant recurrence. In the second group colic anastomoses were analyzed and AL was associated with reduced OS and DFS and there was no correlation with local or distant recurrence.

The studies by Sammour et al[65] and Goto et al[66] also analyzed CSS. They showed a significant reduction in 5-year OS for patients with AL, without finding differences in LR, CSS and postoperative mortality (in rectal carcinoma, leakage affects only the latter). The second one documented instead a reduction in OS (80. 8% vs 90.3%) and CSS (89.6% vs 95.1%), an increase in LR and no correlation with distant recurrence.

A subsequent meta-analysis conducted in 2020 by Bashir Mohamed et al[67] demonstrated a lack of significant effect of AL on recurrences, however it reduced OS, DFS and CSS.

Recent articles on this topic were written by Stormark et al[68] and Kryzauskas et al[69]. The former concluded that leakage only after surgery for stage III CRC is able to reduce survival, whereas the latter demonstrated that AL impaired disease-free and OS in patients undergoing sigmoid and rectal surgery.

Regarding rectal cancer alone, the first data of the new millennium showed an increase in LR and a decrease in CSS[70,71]. Subsequent studies can be divided into 3 categories. In the first group, there are studies that supported the absence of an impact of AL on cancer outcomes such as OS, CSS, DFS, LR and DR[72-79]. The second group covers studies defining AL as an independent prognostic factor for reduced OS, CSS, DFS and increased recurrence[80-83]. In the third group, we can place studies[84-88] midway between the first two categories, as the study of Noh et al[88], demonstrating that AL is associated with increased LR and reduced DFS, whereas its relationship with OS and distant recurrence is not significant. These findings were confirmed in a recent study by Peltrini et al[89].

To the above groups, we must also add studies evaluating also perioperative mortality. Ptok et al[90] and Hain et al[91] found an impact of dehiscence on 30-d mortality, DFS, and LR, whereas Eriksen et al[92] and Bertelsen et al[32] found an increase in 30-d mortality, but without a significant increase in LR. Bertelsen et al[32] also noted the lack of reduction in OS and impact on distant recurrence[32].

Lim et al[93] in 2015 classified ALs into 3 categories based on the consequences: (1) Generalized peritonitis; (2) Localized peritonitis with or without abscess; and (3) Fistula. Oncologic outcomes were evaluated separately for each type and reduced OS and LRFS (LR-free survival) were identified. According to Boström et al[94], leaks only impact OS if they require intervention.

In 2022, Dulskas et al[95] evaluated AL in patients undergoing right colectomy for CRC and concluded that AL is a factor that negatively affects long term prognosis. In contrast, a Dutch retrospective study found that disease recurrence is not associated with AL after CRC resection[96].

Koedam et al[97], analyzing data from the COLOR and COLOR II studies, show that ALs after rectal cancer surgery are associated with an increased rate of LR and a decreased DFS at 5-year follow-up. DR and OS are not significantly affected. Regarding colon cancer surgery, no significant effect of AL on long-term oncologic outcomes was observed, presumably because of a relatively low leakage rate. Strengths of this study include the randomized, multicenter design of the two included studies[98,99] and uniform study protocol for perioperative care and follow-up to limit practice variability.

All studies on this topic are summarized in Table 1.

Table 1.

Summary table of all studies reporting on anastomotic leakage and outcomes after colorectal cancer surgery

Ref.	Study	Period	Cancer	Patients	LR	DR	OS	CSS	DFS	30 d mortality	Follow-up (mo)	Stage	Leak's definition	AL rate (%)	LR rate (%)	Multivariate analysis
Phillips et al[35], 1984	PCS	1976-1980	C + R	1627	Yes	No	No	No	No	No	≥ 60	I, II, III	NR	8	14	No
Sauven et al[36], 1989	RCS	1978-1981	C + R	53	Yes	No	No	No	No	No	36	I, II, III, IV	Clin, Rad	19	13	No
Amato et al[37], 1991	PCS	1981-1995	R	78	Yes	No	No	No	No	No	≥ 24	I, II, III	Clin, Rad	17	12	No
Akyol et al[38], 1991	RCS	1985-1989	C + R	167	Yes	Yes	No	Yes	No	No	25	I, II, III, IV	Clin, Rad	19	18	Yes
Fujita et al[39], 1993	PCS	1970-1991	C/R	980	Yes	Yes	No	No	Yes	No	NR	I, II, III, IV	Clin, Rad	3	3	Yes
Pakkastie et al[100], 1995	PCS	1981-1990	R	116	Yes	No	Yes	No	Yes	No	48	I, II, III	Clin, Rad	16	28	No
Petersen et al[40], 1998	RCS	1985-1995	C + R	331	Yes	No	Yes	Yes	No	Yes	32	I, II, III, IV	Clin	8	9	Yes
Merkel et al[70], 2001	RCS	1978-1996	R	814	Yes	No	No	Yes	No	No	90	I, II, III	Clin	11	14	Yes
Bell et al[101], 2003	PCS	1971-1991	R	401	Yes	No	No	No	No	No	≥ 60	I, II, III	Clin, Rad	13	12	Yes
Law et al[102], 2004	PCS	1993-2002	R	622	Yes	No	No	Yes	No	No	39, 6	I, II, III	Clin, Rad, Endo	6	10	Yes
Walker et al[43], 2004	PCS	1971-1999	C + R	1722	No	No	Yes	No	No	No	≥ 60	I, II, III	Clin, Rad	5	NR	Yes
Branagan et al[41], 2005	PCS	1991-1995	C/R	1834	Yes	No	Yes	No	No	No	≥ 60	I, II, III	Clin, Rad	4	10	Yes
Eriksen et al[92], 2005	PCS	1993-1999	R	1958	Yes	No	Yes	No	No	Yes	45	I, II, III	Clin, Rad	12	11	Yes
McArdle et al[42], 2005	PCS	1991-1994	C + R	2235	No	No	Yes	Yes	No	Yes	≥ 60	I, II, III	Clin, Rad	4	NR	Yes
Choi et al[44], 2006	PCS	1996-2004	C + R	1417	No	No	Yes	No	No	No	NR	I, II, III, IV	Clin, Rad	2	NR	Yes
Ptok et al[90], 2007	RCS	2000-2001	R	2044	Yes	No	No	No	Yes	Yes	40	I, II, III	Clin, Rad, Endo	15	6	Yes
Law et al[45], 2007	PCS	1996-2004	C + R	1580	Yes	Yes	No	Yes	No	Yes	46	I, II, III, IV	Clin, Rad	4	6	Yes
Jung et al[71], 2008	RCS	1997-2003	R	1391	No	No	Yes	Yes	No	No	40	I, II, III	Clin, Rad	3	10	No
Lee et al[79], 2008	PCS	1996-2004	R	1278	Yes	No	Yes	No	Yes	No	45	I, II, III, IV	Clin, Rad, Endo	4	NR	Yes
den Dulk et al[87], 2009	RCS	1987-2002	R	2726	Yes	Yes	Yes	Yes	Yes	No	71	I, II, III	Clin, Rad, Endo	10	9	Yes
Eberhardt et al[46], 2009	PCS	1979-2007	C/R	468	Yes	Yes	Yes	Yes	No	No	94	I, II, III	Clin, Rad	33	6	Yes
Marra et al[47], 2009	RCS	1991-2004	C	440	Yes	Yes	Yes	No	No	Yes	63	I, II, III	Clin, Rad	3	6	No
Bertelsen et al[32], 2010	PCS	2001-2004	R	1494	Yes	Yes	Yes	No	No	Yes	45	I, II, III	Clin, Rad, Endo	11	7	Yes
Kube et al[48], 2010	PCS	2000-2004	C	28271	No	No	Yes	No	Yes	Yes	23	NR	Clin, Rad	3	NR	No
Boccola et al[49], 2011	PCS	1984-2004	C + R	1576	No	No	Yes	Yes	Yes	No	67	I, II, III, IV	Clin, Rad	7	NR	Yes
Jörgren et al[72], 2011	PCS	1995-1997	R	250	Yes	Yes	Yes	Yes	No	Yes	≥ 60	I, II, III	Clin, Rad, Endo	9	8	Yes
Katoh et al[50], 2011	RCS	1990-2000	C/R	207	No	No	No	No	Yes	No	116	II	Clin, Rad	6	NR	Yes
Lin et al[80], 2011	PCS	1993-2003	R	999	Yes	Yes	Yes	Yes	Yes	Yes	≥ 60	I, II, III	Clin, Rad	5	5	Yes
Smith et al[73], 2012	RCS	1991-2010	R	1127	Yes	No	Yes	Yes	No	Yes	74	I, II, III	Clin, Rad	4	5	Yes
Smith et al[103], 2013	RCS	1992-2010	R	184	Yes	No	Yes	Yes	No	No	30	IV	Clin, Rad	7	13	Yes
Krarup et al[59], 2014	RCS	2001-2008	C	8589	Yes	Yes	Yes	No	No	No	≥ 60	I, II, III	Clin, Rad	6	10	Yes
Bakker et al[104], 2014	RCS	2009-2011	C	15667	No	No	No	No	No	Yes	NR	I, II, III, IV	Clin, Rad	8	NR	Yes
Jäger et al[81], 2015	RCS	2003-2010	R	108	No	No	Yes	Yes	Yes	No	70	I, II, III	Clin, Rad	18	NR	Yes
Ke et al[78], 2015	RCS	2007-2011	R	653	Yes	Yes	No	No	Yes	No	47	I, II, III, IV	Clin, Rad	6	4	Yes
Ebinger et al[74], 2015	RCS	1991-2010	R	584	Yes	Yes	Yes	Yes	No	No	62	I, II, III	Clin, Rad, Endo	11	17	Yes
Jannasch et al[84], 2015	PCS	2000-2010	R	17867	Yes	No	Yes	No	Yes	No	30	I, II, III	Clin, Rad	12	9	Yes
Nachiappan et al[52], 2015	PCS	2004-2013	C + R	1048	Yes	Yes	Yes	No	Yes	No	40	I, II, III, IV	Clin, Rad	9	2	Yes
Kang et al[82], 2015	RCS	2006-2009	R	1083	Yes	No	Yes	No	Yes	Yes	54	I, II, III	Clin, Rad	6	2	Yes
Kulu et al[85], 2015	RCS	2002-2011	R	570	Yes	No	Yes	No	No	No	56	I, II, III	Clin, Rad, Endo	9	4	Yes
Krarup et al[60], 2015	RCS	2001-2008	C	8597	No	No	No	No	No	Yes	≥ 60	I, II, III	Clin, Rad	6	NR	Yes
Lim et al[93], 2015	RCS	2007-2011	R	2510	No	No	Yes	No	No	No	33	I, II, III, IV	Clin	6	NR	Yes
Kim et al[53], 2015	RCS	2008-2013	C + R	809	Yes	Yes	Yes	No	No	Yes	NR	I, II, III	Clin, Rad	4	4	Yes
Espín et al[77], 2015	RCS	2006-2008	R	1181	Yes	No	Yes	Yes	No	Yes	60	I, II, III	Clin	9	5	Yes
Breugom et al[54], 2016	RCS	2006-2008	C	761	No	No	Yes	No	Yes	No	60	I, II, III	NR	5	NR	Yes
Park et al[56], 2016	RCS	2000-2011	C/R	10477	Yes	Yes	Yes	No	Yes	Yes	45	I, II, III, IV	Clin, Rad	3	2	Yes
Sammour et al[65], 2018	PCS	1988-2015	C/R	4892	Yes	No	Yes	Yes	No	Yes	60	I, II, III, IV	Clin, Rad	4	C = 5/R = 2	Yes
Noh et al[88], 2016	RCS	2006-2012	R	1258	Yes	Yes	Yes	No	Yes	No	50	I, II, III, IV	Clin, Rad	8	5	Yes
Nordholm et al[61], 2017	RCS	2009-2013	C/R	774	No	No	Yes	No	No	Yes	36	IV	Clin, Rad	9	NR	Yes
Goto et al[66], 2017	RCS	2007-2008	C	3364	Yes	Yes	Yes	Yes	No	Yes	96	I, II, III, IV	Clin, Rad	3	1	Yes
Hain et al[91], 2017	RCS	2005-2014	R	428	Yes	No	No	No	No	Yes	40	I, II, III, IV	Clin, Rad	28	8	Yes
Hüttner et al[51], 2018	RCS	2001-2014	C	628	No	No	Yes	No	Yes	No	60	I, II, III	Rad	4	NR	Yes
Voron et al[57], 2019	RCS	1990-2015	C	1025	No	No	Yes	No	Yes	No	60	I, II, III, IV	Clin, Rad	4	NR	Yes
Boström et al[94], 2018	RCS	2007-2016	R	6948	No	No	Yes	No	No	No	60	I, II, III, IV	NR	10	NR	Yes
Ng et al[62], 2018	RCS	2002-2015	C + R	843	No	No	Yes	No	No	Yes	150	I, II, III, IV	Clin, Rad	6	NR	Yes
Ramphal et al[58], 2018	RCS	2005-2015	C + R	1984	Yes	Yes	Yes	No	Yes	Yes	48	I, II, III, IV	Clin, Rad	8	2	Yes
Furnée et al[86], 2019	RCS	2011	R	746	Yes	Yes	Yes	No	Yes	Yes	42	I, II, III	Rad	14	4	Yes
Allaix et al[83], 2020	RCS	1998-2013	R	532	Yes	Yes	Yes	No	Yes	No	80	I, II, III	Clin, Rad	8	6	Yes
Zimmermann et al[55], 2019	RCS	2001-2014	C + R	1122	Yes	Yes	Yes	No	Yes	No	63	I, II, III, IV	NR	8	1	Yes
Jang et al[76], 2019	RCS	2000-2013	R	698	Yes	Yes	Yes	Yes	Yes	No	48	I, II, III	Clin, Rad	7	17	Yes
Crippa et al[75], 2020	RCS	2000-2013	R	787	Yes	No	Yes	Yes	Yes	No	64	I, II, III, IV	Clin, Rad	5	2	Yes
Kryzauskas et al[69], 2020	PCS	2014-2018	C/R	900	No	No	Yes	No	Yes	Yes	NR	I, II, III, IV	Clin, Rad, Endo	C = 5/R = 11	NR	Yes
Dulskas et al[95], 2022	RCS	2014-2018	C	488	No	No	Yes	No	No	No	48	I, II, III, IV	Clin, Rad, Endo	5	NR	Yes
Arron et al[96], 2022	RCS	2008-2018	C/R	88154	No	No	No	Yes	Yes	No	NR	I, II, III, IV	Clin, Rad	C = 5/R = 8	NR	Yes
Koedam et al[97], 2022	RCS	1997-2010	C/R	1832	Yes	Yes	Yes	No	Yes	No	60	I, II, III (No T4)	Clin, Rad	C = 3/R = 11	C = 15/R = 13	Yes
Peltrini et al[89], 2022	RCS	2011-2017	R	367	Yes	Yes	Yes	No	Yes	No	60	I, II, III, IV	Clin, Rad, Endo	17	23	Yes

AL: Anastomotic leakage; LR: Local recurrence; DR: Distant recurrence; OS: Overall survival; CSS: Cancer-specific survival; DFS: Disease-free survival; PCS: Prospective cohort study; RCS: Retrospective cohort study; C: Colon cancer; R: rectal cancer; C + R: Colon and rectal cancer analyzed together; C/R: Colon and rectal cancer analyzed separately; NR: Not reported; Clin: Clinical; Rad: Radiological; Endo: Endoscopic.

CONCLUSION

AL appears to be an independent risk factor influencing long-term oncologic outcomes after rectal cancer surgery. On the other hand, regarding colon cancer, the results are still extremely heterogeneous and unclear. Further studies on patients undergoing resection for CRC are needed to confirm the oncological impact of AL.

Based on these data, we would recommend more frequent follow-up for patients with AL after CRC cancer surgery.