Original article
Risk factors for anastomotic leakage and leak-related mortality after colonic cancer surgery in a nationwide audit I. S. Bakker1 , I. Grossmann2 , D. Henneman3 , K. Havenga1 and T. Wiggers1 Departments of 1 Surgery and 2 Critical Care, University Medical Centre Groningen, University of Groningen, Groningen, and 3 Department of Surgery, Leiden University Medical Centre, Leiden, The Netherlands Correspondence to: Dr I. S. Bakker, Department of Abdominal Surgery, BA11, University Medical Centre Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands (e-mail: [email protected])
Background: Surgical resection with restoration of bowel continuity is the cornerstone of treatment for
patients with colonic cancer. The aim of this study was to identify risk factors for anastomotic leakage (AL) and subsequent death after colonic cancer surgery. Methods: Data were retrieved from the Dutch Surgical Colorectal Audit. Patients undergoing colonic cancer resection with creation of an anastomosis between January 2009 to December 2011 were included. Outcomes were AL requiring reintervention and postoperative mortality following AL. Results: AL occurred in 7·5 per cent of 15 667 patients. Multivariable analyses identified male sex, high American Society of Anesthesiologists (ASA) fitness grade, extensive tumour resection, emergency surgery, and surgical resection types such as transverse resection, left colectomy and subtotal colectomy as independent risk factors for AL. A defunctioning stoma was created in a small group of patients, leading to a lower risk of leakage. The mortality rate was 4·1 per cent overall, and was significantly higher in patients with AL than in those without leakage (16·4 versus 3·1 per cent; P < 0·001). Multivariable analyses identified older age, high ASA grade, high Charlson score and emergency surgery as independent risk factors for death after AL. The adjusted risk of death after AL was twice as high following right compared with left colectomy. Conclusion: The elderly and patients with co-morbidity have a higher risk of death after AL. Accurate preoperative patient selection, intensive postoperative surveillance for AL, and early and aggressive treatment of suspected leakage is important, especially in patients undergoing right colectomy. Paper accepted 7 November 2013 Published online in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.9395
Introduction
Surgical resection is the cornerstone of treatment for patients with colonic cancer. Generally, restoration of bowel continuity with a primary anastomosis is pursued in uncomplicated colonic resections. The most serious complication of colonic surgery with restoration of bowel continuity is anastomotic leakage (AL)1 , which is associated with the possible need for reintervention, increased mortality2,3 and, possibly, worse oncological outcome4,5 . The reported incidence of AL in colonic anastomosis varies between 3 and 6·4 per cent, depending on patient and tumour characteristics, definitions, site of the anastomosis and perhaps surgeon caseload6 – 9 . Several risk factors, including co-morbidity, higher American Society of Anesthesiologists (ASA) fitness grade, stage of disease, type of operation, emergency surgery and intraoperative complications, have all been associated 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
with AL7,10 – 12 . Concentration of surgery in high-volume centres has been considered as a strategy to improve quality of care, surgical outcomes and mortality13,14 . Hospital procedural volume could also be a possible risk factor for AL. Although AL has long been a subject of debate, prediction of the risk of leakage for the individual patient remains difficult. The Dutch Surgical Colorectal Audit (DSCA), a clinical outcome registry in which all Dutch hospitals participate, was initiated in 2009 to monitor and improve outcome of surgical care for patients with colorectal cancer. The DSCA facilitates individual hospitals in quality improvement projects, but is also used to identify treatment and outcome patterns for different patient groups. AL after colorectal resection has been selected as an outcome indicator for surgical quality of care in the DSCA15 . In rectal cancer surgery, the practice of routine construction BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
of a defunctioning stoma may play a large role in measuring this outcome and determining risk factors16 , whereas stoma construction may be of less importance in colonic cancer resections. Among all colorectal surgical procedures, patients undergoing colonic cancer resection may be considered a specifically vulnerable group, being at risk for morbidity and mortality because of advanced age and co-morbidity17 . Risk factors for AL and related postoperative mortality have not yet been investigated in this group. The aim of the present study was to identify risk factors for AL after colonic cancer resection and factors influencing mortality associated with AL using patient information from a national audit database.
425
grade, previous abdominal surgery), tumour factors (stage, location, preoperative tumour complications) and treatment factors (preoperative surgical procedures (stoma or other), type of surgical resection, urgency of surgery, extent of resection, fashioning of a defunctioning stoma), were extracted from the database. Hospitals were categorized as low-volume (50 or fewer), medium-volume (51–100) or high-volume (more than 100) centres, based on the number of patients who underwent surgical treatment of colonic cancer per year for the years 2010 and 2011 (in 2009 not all hospitals completed registration). These categories reflect the present situation in the Netherlands, with 50 per cent of clinics performing between 50 and 100 colonic cancer resections annually22 .
Methods
The data set was retrieved from the DSCA, a web-based national database, in which information on all patients undergoing surgical resection for colorectal cancer is recorded18 . Data on patient and tumour characteristics, diagnostics, treatment and outcome were collected. The data set included information registered from 92 hospitals with nearly 100 per cent concordance on validation against the National Cancer Registry. In 2009, 89 per cent of Dutch hospitals participated, increasing to 99 per cent in 2010 and 201119 . Medical ethics committee approval was not required for this study as all patients and hospital information in the DSCA was deidentified. Individual patient data were collected in the treating hospital and transferred in encrypted form to the DSCA database.
Statistical analysis Univariable analyses were carried out using the χ2 test to examine the association between patient, tumour, treatment and hospital factors and the occurrence of AL and death after AL. Multivariable logistic regression analyses were performed to correct for possible confounders. A manual stepwise model was used, with inclusion of variables with P < 0·050 in univariable analysis. Clinically relevant variables were added to the statistical model. The variables ‘timing of surgery’ (elective/emergency) and ‘preoperative tumour complications’ were assumed to indicate the same clinical situation. To check for collinearity when both variables were incorporated in the model, the variance inflation factor was computed. To check the model, the multivariable analysis was repeated with the outcome AL,
Inclusion and exclusion criteria All Dutch patients who underwent colonic cancer resection in the Netherlands between January 2009 and December 2011 were included in this study. Patients with rectal cancer, those with multiple synchronous tumours, patients without a primary anastomosis and those with an unknown surgical resection type were excluded from analysis. Surgical resections were categorized as ileocaecal resection, right colectomy, transverse resection, left colectomy, sigmoid resection and subtotal colectomy.
DSCA 2009–2011 Resection for primary colorectal cancer n = 27 259 Excluded n = 8557 Patients with rectal cancer n = 7614 Patients with multiple synchronous tumours n = 943 Patients with colonic cancer n = 18 702
Outcomes Primary outcome measures were AL, defined as clinically significant leakage requiring surgical or radiological reintervention, and mortality after AL, defined as death in hospital within 30 days after primary surgery. Potential risk factors for postoperative complications, including patient factors (age, sex, body mass index (BMI), Charlson co-morbidity score20,21 , ASA fitness 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Excluded n = 3035 No anastomosis n = 2581 Unknown/other resection type n = 454 Patients who underwent colonic cancer resection with primary anastomosis n = 15 667 Fig. 1
Study flow chart. DSCA, Dutch Surgical Colorectal Audit
www.bjs.co.uk
BJS 2014; 101: 424–432
426
Table 1
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
Univariable analyses of patient variables associated with anastomotic leakage and with death after anastomotic leakage No. of patients (n = 15 667)
Age (years) < 65 65–80 > 80 Unknown Sex F M BMI (kg/m2 ) < 20 20–25 25–30 > 30 Unknown ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Previous abdominal surgery Yes No
Anastomotic leakage (n = 1176)
P*
Death after anastomotic leakage (n = 193)
0·284 4825 (30·8) 7616 (48·6) 3211 (20·5) 15 (0·1)
386 (8·0) 551 (7·2) 239 (7·4) 0 (0)
7605 (48·5) 8062 (51·5)
477 (6·3) 699 (8·7)
623 (4·0) 4922 (31·4) 4986 (31·8) 2079 (13·3) 3057 (19·5)
54 (8·7) 366 (7·4) 389 (7·8) 169 (8·1) 198 (6·5)
11 638 (74·3) 3713 (23·7) 316 (2·0)
822 (7·1) 340 (9·2) 14 (4·4)
8335 (53·2) 3579 (22·8) 3753 (24·0)
583 (7·0) 285 (8·0) 308 (8·2)
5309 (33·9) 10 358 (66·1)
425 (8·0) 751 (7·3)
P* < 0·001
20 (5·2) 101 (18·3) 72 (30·1) 0 (0) < 0·001
0·783 80 (16·8) 113 (16·2)
0·643
0·535 11 (20) 59 (16·1) 54 (13·9) 29 (17·2) 40 (20·2)
< 0·001
< 0·001 94 (11·4) 96 (28·2) 3 (21)
0·032
< 0·001 56 (9·6) 59 (20·7) 78 (25·3)
0·090
0·065 81 (19·1) 112 (14·9)
Values in parentheses are percentages. BMI, body mass index; ASA, American Society of Anesthesiologists. *χ2 test.
first without ‘preoperative tumour complications’ and including ‘timing of surgery’. The same analysis was then repeated but including ‘preoperative tumour complications’ and without ‘timing of surgery’. Results are reported as odds ratios (ORs) with 95 per cent confidence intervals (c.i.). Two-sided P < 0·050 was considered statistically significant. All data were analysed using PASW Statistics, release 20.0.0.1 (IBM, Armonk, New York, USA). Results
From 2009 to 2011, data from 27 259 patients were included in the DSCA database (Fig. 1). After exclusion of those without colonic cancer, 18 702 patients were eligible for inclusion. Following exclusion of a further 2581 patients without a primary anastomosis and 454 who underwent total colectomy or an unknown resection, 15 667 patients with colonic cancer remained for analysis. Surgical procedures included: 240 ileocaecal resections, 7788 right colectomies, 527 transverse resections, 1601 left colectomies, 5354 sigmoid resections and 157 subtotal colectomies. Surgery was performed in 92 hospitals, with 17·1 per cent of the patients being treated in a low-volume centre, 54·0 per cent in a medium-volume centre and 28·9 per cent in a high-volume centre. 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Anastomotic leakage AL leading to reintervention occurred in 1176 patients (7·5 per cent). The reinterventions were laparotomy (82·1 per cent), laparoscopy (2·8 per cent), radiological drainage (8·2 per cent), and interventions such as drainage of wounds and abscesses (6·9 per cent). During the primary operation, a defunctioning stoma was created in 606 patients (3·9 per cent), usually after sigmoid resection. Of all anastomoses created after sigmoid resection, 8·7 per cent were diverted. There was no difference in AL rate between patients with and without a defunctioning stoma (6·4 versus 7·5 per cent respectively; P = 0·308). The incidence of AL was significantly higher after resection of the transverse colon, left colectomy and subtotal colectomy than after other types of resection. A secondary stoma was created in 805 patients (68·5 per cent) requiring a surgical or radiological reintervention for AL.
Risk factors for anastomotic leakage Tables 1–3 show the results of univariable analyses of possible risk factors for the development of AL. Patient factors associated with an increased risk of AL were male sex, higher ASA grade and higher Charlson score. Among tumour factors, preoperative tumour complications www.bjs.co.uk
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
Table 2
427
Univariable analyses of tumour characteristics associated with anastomotic leakage and with death after anastomotic
leakage
Preop. tumour complications None Perforation Obstruction Blood loss Other Tumour stage 0 1 2 3 4 Unknown Tumour location Caecum Ascending colon Hepatic flexure Transverse colon Splenic flexure Descending colon Sigmoid colon
No. of patients (n = 15 667)
Anastomotic leakage (n = 1176)
11 968 (76·4) 356 (2·3) 1557 (9·9) 951 (6·1) 835 (5·3)
820 (6·9) 43 (12·1) 175 (11·2) 65 (6·8) 73 (8·7)
153 (1·0) 2811 (17·9) 5769 (36·8) 4918 (31·4) 1784 (11·4) 232 (1·5)
13 (8·5) 180 (6·4) 461 (8·0) 376 (7·6) 131 (7·3) 15 (6·5)
3513 (22·4) 3085 (19·7) 1064 (6·8) 1168 (7·5) 448 (2·9) 887 (5·7) 5502 (35·1)
216 (6·1) 199 (6·5) 70 (6·6) 122 (10·4) 54 (12·1) 93 (10·5) 422 (7·7)
P*
Death after anastomotic leakage (n = 193)
< 0·001
P* 0·100
130 (15·9) 8 (19) 30 (17·1) 6 (9) 19 (26) 0·176
0·782 3 (23) 31 (17·2) 67 (14·5) 67 (17·8) 23 (17·6) 2 (13)
< 0·001
0·225 45 (20·8) 35 (17·6) 15 (21) 20 (16·4) 9 (17) 11 (12) 58 (13·7)
Values in parentheses are percentages. *χ2 test. Table 3
Univariable analyses of treatment characteristics associated with anastomotic leakage and with death after anastomotic
leakage
Preop. surgical procedure None Stoma formation Other Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Timing of surgery Elective Emergency Unknown Stoma No stoma Defunctioning stoma Extensive resection No Yes Hospital volume Low Medium High
No. of patients (n = 15 667)
Anastomotic leakage (n = 1176)
15 285 (97·6) 118 (0·8) 264 (1·7)
1133 (7·4) 15 (12·7) 28 (10·6)
7788 (49·7) 240 (1·5) 527 (3·4) 1601 (10·2) 5354 (34·2) 157 (1·0)
495 (6·4) 18 (7·5) 57 (10·8) 172 (10·7) 413 (7·7) 21 (13·4)
13 139 (83·9) 1625 (10·4) 903 (5·8)
925 (7·0) 159 (9·8) 92 (10·2)
P*
Death after anastomotic leakage (n = 193)
0·015
0·383 188 (16·6) 3 (20) 2 (7)
< 0·001
0·029 101 (20·4) 3 (17) 10 (18) 20 (11·6) 58 (14·0) 1 (5)
< 0·001
0·028 145 (15·7) 37 (23·3) 11 (12)
0·308 15 061 (96·1) 606 (3·9)
1137 (7·5) 39 (6·4)
14 216 (90·7) 1451 (9·3)
1025 (7·2) 151 (10·4)
2680 (17·1) 8461 (54·0) 4526 (28·9)
180 (6·7) 653 (7·7) 343 (7·6)
P*
0·860 187 (16·4) 6 (15)
< 0·001
0·321 164 (16·0) 29 (19·2)
0·225
0·162 38 (21·1) 99 (15·2) 56 (16·3)
Values in parentheses are percentages. Hospital volume: low, 50 or fewer patients per year; medium, 51–100 patients per year; high, more than 100 patients per year. *χ2 test.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
BJS 2014; 101: 424–432
428
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
Multivariable logistic regression analysis of risk factors for anastomotic leakage among patients who underwent colonic surgery with a primary colonic anastomosis
Mortality after anastomotic leakage
Table 4
Odds ratio Sex F M ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Preop. surgical procedures None Stoma Other Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Extensive resection No Yes Stoma No stoma Defunctioning stoma Time of surgery Elective Emergency Unknown
P
Age > 80 years ASA grade III–IV Charlson score ≥ 2
1·00 (reference) 1·38 (1·22, 1·56)
< 0·001
Right colectomy
1·00 (reference) 1·26 (1·09, 1·45) 0·59 (0·34, 1·02)
0·002 0·075
Transverse resection
1·00 (reference) 1·11 (0·95, 1·29) 1·10 (0·94, 1·29)
Emergency surgery
Ileocaecal resection
Left colectomy Sigmoid resection
0·178 0·218
Subtotal colectomy Overall rate 0
1·00 (reference) 1·52 (0·87, 2·65) 1·21 (0·81, 1·81)
0·139 0·361
1·00 (reference) 1·13 (0·69, 1·85) 1·69 (1·26, 2·26) 1·69 (1·40, 2·03) 1·28 (1·11, 1·47) 2·28 (1·42, 3·66)
0·630 < 0·001 < 0·001 0·001 0·001
1·00 (reference) 1·43 (1·19, 1·72)
< 0·001
1·00 (reference) 0·68 (0·49, 0·96)
0·026
1·00 (reference) 1·33 (1·11, 1·59) 1·55 (1·23, 1·96)
0·002 < 0·001
Values in parentheses are 95 per cent confidence intervals. ASA, American Society of Anesthesiologists.
(mostly tumour perforation or obstruction) and tumour location were associated with an increased risk of AL. Treatment factors associated with a higher risk of AL were preoperative surgical procedures (such as construction of a defunctioning stoma), extensive resection (resection of other organs during surgery), emergency surgery, and type of resection, especially transverse resection, left colectomy and subtotal colectomy. Multivariable analyses confirmed male sex and a high ASA grade as independent risk factors for AL (Table 4). Treatment factors that remained associated with a higher risk of AL were type of resection (mainly transverse resection, left colectomy and subtotal colectomy compared with right colectomy as reference group), extensive resection and emergency surgery. On clinical grounds, the variable ‘defunctioning stoma’ was added to the model and adjusted data also showed a lower risk of AL in patients 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Anastomotic leakage
5
10
15
20
25
30
35
Rate (%)
Anastomotic leakage and subsequent mortality rates after colonic resection and primary anastomosis among patients with risk factors identified by univariable analysis, and overall rates for the whole cohort. ASA, American Society of Anesthesiologists
Fig. 2
with a defunctioning stoma (OR 0·68, 95 per cent c.i. 0·49 to 0·96). To check for the presence of collinearity between the variables ‘timing of surgery’ and ‘preoperative tumour complications’, the variance inflation factor was computed. The results indicated no collinearity between these variables. Analysis of the multivariable model including ‘preoperative tumour complications’ instead of ‘timing of surgery’ showed a significantly higher risk of AL in patients with preoperative tumour complications, such as perforation (OR 1·68, 1·26 to 2·26) and obstruction (1·62, 1·39 to 2·02).
Mortality Some 648 (4·1 per cent) of the 15 667 patients died within 30 days of surgery, 3·4 per cent after elective surgery and 7·2 per cent after emergency operation (P < 0·001). AL was diagnosed after surgery in 193 (29·8 per cent) of those who died. The mortality rate among patients with AL was significantly higher than in those without leakage (16·4 versus 3·1 per cent; P < 0·001). There was no significant association between the number of patients treated yearly per hospital and mortality after AL (P = 0·162).
Risk factors for death after anastomotic leakage Univariable analyses revealed that patient factors associated with death after AL were older age, high ASA grade and a high Charlson score (Table 1). Surgery in the www.bjs.co.uk
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
429
Multivariable logistic regression analysis of risk factors for death after anastomotic leakage among patients who underwent colonic surgery with a primary colonic anastomosis
Table 5
Age (years) < 65 65–80 > 80 ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Timing of surgery Elective Emergency Unknown
Odds ratio
P
1·00 (reference) 3·15 (1·89, 5·27) 5·16 (2·98, 8·96)
< 0·001 < 0·001
1·00 (reference) 1·77 (1·24, 2·52) 1·89 (0·48, 7·47)
0·002 0·363
1·00 (reference) 1·76 (1·16, 2·69) 2·23 (1·47, 3·37)
0·008 < 0·001
1·00 (reference) 1·00 (0·25, 3·94) 0·80 (0·38, 1·71) 0·54 (0·31, 0·92) 0·75 (0·51, 1·08) 0·28 (0·04, 2·24)
0·998 0·566 0·025 0·124 0·232
1·00 (reference) 1·75 (1·12, 2·73) 0·78 (0·39, 1·57)
0·014 0·483
Values in parentheses are 95 per cent confidence intervals. ASA, American Society of Anesthesiologists.
emergency setting and type of surgical resection were also associated with a higher risk (Table 3). In particular, patients undergoing right colectomy, transverse resection or ileocaecal resection had high mortality rates after AL (Fig. 2). After adjustment for possible confounders, multivariable analyses showed that older age, high ASA grade, higher Charlson score and emergency surgery were independent risk factors for death among patients diagnosed with AL (Table 5). Multivariable analysis also revealed that the risk of death associated with AL was lower after left colectomy compared with other surgical resection types. Discussion
In the present study of risk factors for AL and mortality following colonic cancer resection with an anastomosis, the overall AL rate was 7·5 per cent. The mortality rate was 4·1 per cent overall, and was significantly higher for patients with AL than in those without (16·4 versus 3·1 per cent). The present leak rate is higher than in recent reports (3–6·4 per cent)6 – 8 . This might be attributed to the complete registration in a clinical audit. Reported results from a nationwide study on AL in Denmark also showed a rather high rate of 6·4 per cent7 . Other publications 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
with lower rates usually originated from dedicated centres. Analyses adjusted for confounding factors revealed male sex and high ASA grade as independent risk factors for AL, which is consistent with the literature7,23,24 . Other reported predictors of AL, such as previous abdominal surgery24 and high BMI3,25 , were not confirmed in the present analysis. Another well known risk factor for AL is co-morbidity8,11,24 , reflected in the Charlson score and ASA classification in the present study. Both scores were associated with AL in univariable analyses, and ASA grade remained a significant predictor in multivariable analyses. Adjusted analysis in the present study indicated that treatment factors such as extensive resection and type of surgical resection were independent risk factors for AL. The incidence of AL differs depending on tumour location and subsequent type of resection. Right-sided colectomy is associated with a lower leak rate than leftsided colectomy3,26 , and the AL rate is higher after transverse resection11 . Differences in vascularization of the anastomotic site may explain this as a well vascularized ileal bowel loop is anastomosed to an adequately vascularized transverse colonic loop in right-sided hemicolectomy. However, in transverse or left colonic resection, where the middle colic artery or inferior mesenteric artery is divided, vascularization of the anastomotic site might be compromised27,28 . Another possible explanation is the lack of full mobilization of one or both flexures. The poorer outcome after transverse resection in the present study emphasizes the importance of careful surgical decisionmaking. For patients requiring a transverse resection, an extended colectomy could be a better option. Data on other known risk factors such as weight loss26 , intraoperative complications, duration of operation, blood loss and faecal contamination8,11 were not available in the DSCA database. The overall mortality rate of 4·1 per cent in the present analysis is within the range in population-based studies (3·0–7·4 per cent)6,29,30 . The mortality rate of 16·4 per cent following AL is high, as noted previously (12·0–18·6 per cent)6,11 , and related to old age and comorbidity. The mortality rate after AL is higher after colonic than after rectal resection (0·7–4 per cent)19,31,32 . This might be explained by differences in anatomy. AL after colonic resection often results in a generalized peritonitis, compared with more local, extraperitoneal abscess formation after rectal resection. Leakage scores have been developed for early detection of AL after rectal surgery29,33 . It would be of interest to validate these scores for colonic resections, as the severity of the consequences of leakage from a colonic anastomosis is underestimated. www.bjs.co.uk
BJS 2014; 101: 424–432
430
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
To reduce the incidence of AL or its clinical sequelae, a defunctioning stoma can be constructed. In rectal anastomoses, temporary defunctioning stomas are fashioned to reduce the clinical consequences of leakage10,34 . In the present series most defunctioning stomas were constructed after sigmoid resection. A defunctioning stoma had no impact on the incidence of AL in univariable analysis but, after adjustment for patient and tumour characteristics, multivariable analysis showed a protective effect in terms of AL. Apparently there was good patient selection for faecal diversion, based on the preoperative or intraoperative judgement of the surgeons concerning the risk of AL. Emergency surgery is also considered a risk factor for both AL12,35 and postoperative mortality12,30,36 . Poor general condition and nutritional status is associated with a higher risk of morbidity and mortality in these patients36,37 . In the present study, emergency surgery was identified as an independent risk factor for both AL and mortality following leakage. Further multivariable analysis also confirmed that patients with preoperative complications such as tumour perforation and obstruction had a higher risk of postoperative AL. Emergency surgery is frequently performed during evening and night shifts. Studies from different medical fields have reported worse postoperative outcome after operations performed at these times38 – 40 . Surgical procedures during late shifts are sometimes performed by surgeons with a lower disease-specific caseload. Some studies have suggested that the experience and caseload of the individual surgeon are predictors of postoperative death9,41 . Unfortunately, the DSCA database does not contain data relating to individual surgeons. The strength of this study is that the results are based on a complete and large nationwide data set, which included registered data from all Dutch hospitals performing colorectal surgery. Validation of the registered data against the national cancer registry showed a high concordance. Therefore a valid analysis of colonic cancer surgery in the Netherlands could be made. However, several limitations are worth mentioning. A somewhat heterogeneous study population was analysed. All patients undergoing resection for colonic cancer were included, including those with stage IV disease, who may represent both curative and palliative resections and operations in the emergency setting. Although the analysis of such a complete cohort leads to fair results, the heterogeneity is also accompanied by confounding factors and might lead to bias. To control for bias, multivariable analysis was carried out stratified for time of surgery (elective versus emergency). The main results of analyses for elective and
emergency subgroups did not differ significantly from those of the total study population (data not shown). Therefore, ‘timing of surgery’ was included as a possible risk factor for AL and subsequent death in the multivariable model. Another limitation of this data set is that only patients treated for cancer could be analysed, although there are also benign indications for colonic surgery. Furthermore, only limited intraoperative data were available. The data set contains no information on duration of operation, blood loss or surgical techniques. Intraoperative information could make a valuable contribution to the identification of risk factors for AL. The same also applies to data regarding surgeon caseload. Colonic cancer resections are common surgical procedures. Although the exact mechanism leading to the development of AL is often unknown for the individual patient, and the clinical risk assessment by the operating surgeon is of low predictive value42 , it is important to understand that tumour and treatment factors may play an important role. Mortality rates following AL are high, and determined mainly by patient factors such as age and co-morbidity. Therefore, it is important to monitor patients after surgery according to standardized surveillance protocols, perhaps incorporating colonic leakage scores designed for left-sided colorectal procedures29,33 . Future research should be focused on further evaluation of leakage scores in colonic resections, and on continuous monitoring through clinical audit.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
Acknowledgements
The authors thank M. Fiocco for the statistical guidance and advice in the present study. Disclosure: The authors declare no conflict of interest. References 1 Khan AA, Wheeler JM, Cunningham C, George B, Kettlewell M, Mortensen NJ. The management and outcome of anastomotic leaks in colorectal surgery. Colorectal Dis 2008; 10: 587–592. 2 van Westreenen HL, Ijpma FF, Wevers KP, Afzali H, Patijn GA. Reoperation after colorectal surgery is an independent predictor of the 1-year mortality rate. Dis Colon Rectum 2011; 54: 1438–1442. 3 Buchs NC, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B, Morel P. Incidence, consequences, and risk factors for anastomotic dehiscence after colorectal surgery: a prospective monocentric study. Int J Colorectal Dis 2008; 23: 265–270. 4 Petersen S, Freitag M, Hellmich G, Ludwig K. Anastomotic leakage: impact on local recurrence and survival in surgery of colorectal cancer. Int J Colorectal Dis 1998; 13: 160–163. 5 Law WL, Choi HK, Lee YM, Ho JW, Seto CL. Anastomotic leakage is associated with poor long-term
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
6
7
8
9
10
11
12
13
14
15
16
17
18
outcome in patients after curative colorectal resection for malignancy. J Gastrointest Surg 2007; 11: 8–15. Kube R, Mroczkowski P, Granowski D, Benedix F, Sahm M, Schmidt U et al.; Study Group Quality Assurance in Primary Colorectal Carcinoma. Anastomotic leakage after colon cancer surgery: a predictor of significant morbidity and hospital mortality, and diminished tumour-free survival. Eur J Surg Oncol 2010; 36: 120–124. Krarup PM, Jorgensen LN, Andreasen AH, Harling H; Danish Colorectal Cancer Group. A nationwide study on anastomotic leakage after colonic cancer surgery. Colorectal Dis 2012; 14: 661–667. Leichtle SW, Mouawad NJ, Welch KB, Lampman RM, Cleary RK. Risk factors for anastomotic leakage after colectomy. Dis Colon Rectum 2012; 55: 569–575. Renzulli P, Lowy A, Maibach R, Egeli RA, Metzger U, Laffer UT. The influence of the surgeon’s and the hospital’s caseload on survival and local recurrence after colorectal cancer surgery. Surgery 2006; 139: 296–304. Peeters KC, Tollenaar RA, Marijnen CA, Klein Kranenbarg E, Steup WH, Wiggers T et al.; Dutch Colorectal Cancer Group. Risk factors for anastomotic failure after total mesorectal excision of rectal cancer. Br J Surg 2005; 92: 211–216. Alves A, Panis Y, Trancart D, Regimbeau JM, Pocard M, Valleur P. Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World J Surg 2002; 26: 499–502. Buchs NC, Gervaz P, Bucher P, Huber O, Mentha G, Morel P. Lessons learned from one thousand consecutive colonic resections in a teaching hospital. Swiss Med Wkly 2007; 137: 259–264. Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA 2000; 284: 3028–3035. Mroczkowski P, Kube R, Ptok H, Schmidt U, Hac S, ¨ Kockerling F et al. Low-volume centre vs high-volume: the role of a quality assurance programme in colon cancer surgery. Colorectal Dis 2011; 13: e276–e283. Snijders HS, Henneman D, van Leersum NL, ten Berge M, Fiocco M, Karsten TM et al. Anastomotic leakage as an outcome measure for quality of colorectal cancer surgery. BMJ Qual Saf 2013; 22: 759–767. Snijders HS, van den Broek CB, Wouters MW, Meershoek-Klein Kranenbarg E, Wiggers T, Rutten H et al. An increasing use of defunctioning stomas after low anterior resection for rectal cancer. Is this the way to go? Eur J Surg Oncol 2013; 39: 715–720. Lemmens VE, Janssen-Heijnen ML, Houterman S, Verheij KD, Martijn H, van de Poll-Franse L et al. Which comorbid conditions predict complications after surgery for colorectal cancer? World J Surg 2007; 31: 192–199. Van Leersum NJ, Snijders HS, Henneman D, Kolfschoten NE, Gooiker GA, ten Berge MG et al. The Dutch Surgical Colorectal Audit. Eur J Surg Oncol 2013; 39: 1063–1070.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
431
19 Dutch Institute for Clinical Auditing. Rapportages 2011. http://www.clinicalaudit.nl/jaarrapportage/archief/DICA %20Jaarrapportage%202011.pdf [accessed 26 November 2013]. 20 Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383. 21 Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 1245–1251. 22 Elferink MA, Wouters MW, Krijnen P, Lemmens VE, Jansen-Landheer ML, van de Velde CJ et al. Disparities in quality of care for colon cancer between hospitals in the Netherlands. Eur J Surg Oncol 2010; 36(Suppl 1): S64–S73. 23 Choi HK, Law WL, Ho JW. Leakage after resection and intraperitoneal anastomosis for colorectal malignancy: analysis of risk factors. Dis Colon Rectum 2006; 49: 1719–1725. 24 Lipska MA, Bissett IP, Parry BR, Merrie AE. Anastomotic leakage after lower gastrointestinal anastomosis: men are at a higher risk. ANZ J Surg 2006; 76: 579–585. 25 Biondo S, Par´es D, Kreisler E, Ragu´e JM, Fraccalvieri D, Ruiz AG et al. Anastomotic dehiscence after resection and primary anastomosis in left-sided colonic emergencies. Dis Colon Rectum 2005; 48: 2272–2280. 26 Veyrie N, Ata T, Muscari F, Couchard AC, Msika S, Hay JM et al. Anastomotic leakage after elective right versus left colectomy for cancer: prevalence and independent risk factors. J Am Coll Surg 2007; 205: 785–793. 27 Drummond H. Some points relating to the surgical anatomy of the arterial supply of the large intestine. Proc R Soc Med 1914; 7(Surg Sect): 185–193. 28 Hall NR, Finan PJ, Stephenson BM, Lowndes RH, Young HL. High tie of the inferior mesenteric artery in distal colorectal resections – a safe vascular procedure. Int J Colorectal Dis 1995; 10: 29–32. 29 den Dulk M, Noter SL, Hendriks ER, Brouwers MA, van der Vlies CH, Oostenbroek RJ et al. Improved diagnosis and treatment of anastomotic leakage after colorectal surgery. Eur J Surg Oncol 2009; 35: 420–426. 30 Sjo OH, Larsen S, Lunde OC, Nesbakken A. Short term outcome after emergency and elective surgery for colon cancer. Colorectal Dis 2009; 11: 733–739. 31 den Dulk M, Marijnen CA, Collette L, Putter H, P˚ahlman L, Folkesson J et al. Multicentre analysis of oncological and survival outcomes following anastomotic leakage after rectal cancer surgery. Br J Surg 2009; 96: 1066–1075. 32 Snijders HS, Wouters MW, van Leersum NJ, Kolfschoten NE, Henneman D, de Vries AC et al. Meta-analysis of the risk for anastomotic leakage, the postoperative mortality caused by leakage in relation to the overall postoperative mortality. Eur J Surg Oncol 2012; 38: 1013–1019. 33 Dekker JW, Liefers GJ, de Mol van Otterloo JC, Putter H, Tollenaar RA. Predicting the risk of anastomotic leakage in
www.bjs.co.uk
BJS 2014; 101: 424–432
432
34
35
36
37
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
left-sided colorectal surgery using a colon leakage score. J Surg Res 2011; 166: e27–e34. Shiomi A, Ito M, Saito N, Ohue M, Hirai T, Kubo Y et al. Diverting stoma in rectal cancer surgery. A retrospective study of 329 patients from Japanese cancer centers. Int J Colorectal Dis 2011; 26: 79–87. Kolfschoten NE, Wouters MW, Gooiker GA, Eddes EH, Kievit J, Tollenaar RA et al.; Dutch Surgical Colorectal Audit Group. Nonelective colon cancer resections in elderly patients: results from the Dutch Surgical Colorectal Audit. Dig Surg 2012; 29: 412–419. Zielinski MD, Merchea A, Heller SF, You YN. Emergency management of perforated colon cancers: how aggressive should we be? J Gastrointest Surg 2011; 15: 2232–2238. Lee YM, Law WL, Chu KW, Poon RT. Emergency surgery for obstructing colorectal cancers: a comparison between right-sided and left-sided lesions. J Am Coll Surg 2001; 192: 719–725.
38 Peled Y, Melamed N, Chen R, Pardo J, Ben-Shitrit G, Yogev Y. The effect of time of day on outcome of unscheduled cesarean deliveries. J Matern Fetal Neonatal Med 2011; 24: 1051–1054. 39 Chacko AT, Ramirez MA, Ramappa AJ, Richardson LC, Appleton PT, Rodriguez EK. Does late night hip surgery affect outcome? J Trauma 2011; 71: 447–453. 40 Coumbe A, John R, Kuskowski M, Agirbasli M, McFalls EO, Adabag S. Variation of mortality after coronary artery bypass surgery in relation to hour, day and month of the procedure. BMC Cardiovasc Disord 2011; 11: 63. 41 Schrag D, Panageas KS, Riedel E, Hsieh L, Bach PB, Guillem JG et al. Surgeon volume compared to hospital volume as a predictor of outcome following primary colon cancer resection. J Surg Oncol 2003; 83: 68–78. 42 Karliczek A, Harlaar NJ, Zeebregts CJ, Wiggers T, Baas PC, van Dam GM. Surgeons lack predictive accuracy for anastomotic leakage in gastrointestinal surgery. Int J Colorectal Dis 2009; 24: 569–576.
Editor’s commentary
Anastomotic leakage is one of the most feared complications after gastrointestinal surgery; the rate varies with the type of surgery and the location of the anastomosis. As evident from the two studies in this issue of BJS, laparoscopic bariatric surgery has a considerably lower leak rate (less than 2 per cent) than colonic surgery (less than 8 per cent), yet both can be fatal. While no-one would argue against careful patient selection and vigilant postoperative surveillance, it may be necessary to think outside the box to further reduce and, eventually, prevent this dreaded complication. In the era of personalized medicine, one could envision future staplers with biosensors and feedback systems to optimize closure and ensure decent blood flow rather than the non-standardized and blunt ‘one-size-fits-all’ approach currently available. Or is it the gut microbiome that makes some patients’ suture line prone to delayed healing and leakage? The BJS welcomes novel research that further highlights our understanding of surgical technology and biological processes in this regard. K. Søreide Editor, BJS
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
BJS 2014; 101: 424–432
Risk factors for anastomotic leakage and leak-related mortality after colonic cancer surgery in a nationwide audit I. S. Bakker1 , I. Grossmann2 , D. Henneman3 , K. Havenga1 and T. Wiggers1 Departments of 1 Surgery and 2 Critical Care, University Medical Centre Groningen, University of Groningen, Groningen, and 3 Department of Surgery, Leiden University Medical Centre, Leiden, The Netherlands Correspondence to: Dr I. S. Bakker, Department of Abdominal Surgery, BA11, University Medical Centre Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands (e-mail: [email protected])
Background: Surgical resection with restoration of bowel continuity is the cornerstone of treatment for
patients with colonic cancer. The aim of this study was to identify risk factors for anastomotic leakage (AL) and subsequent death after colonic cancer surgery. Methods: Data were retrieved from the Dutch Surgical Colorectal Audit. Patients undergoing colonic cancer resection with creation of an anastomosis between January 2009 to December 2011 were included. Outcomes were AL requiring reintervention and postoperative mortality following AL. Results: AL occurred in 7·5 per cent of 15 667 patients. Multivariable analyses identified male sex, high American Society of Anesthesiologists (ASA) fitness grade, extensive tumour resection, emergency surgery, and surgical resection types such as transverse resection, left colectomy and subtotal colectomy as independent risk factors for AL. A defunctioning stoma was created in a small group of patients, leading to a lower risk of leakage. The mortality rate was 4·1 per cent overall, and was significantly higher in patients with AL than in those without leakage (16·4 versus 3·1 per cent; P < 0·001). Multivariable analyses identified older age, high ASA grade, high Charlson score and emergency surgery as independent risk factors for death after AL. The adjusted risk of death after AL was twice as high following right compared with left colectomy. Conclusion: The elderly and patients with co-morbidity have a higher risk of death after AL. Accurate preoperative patient selection, intensive postoperative surveillance for AL, and early and aggressive treatment of suspected leakage is important, especially in patients undergoing right colectomy. Paper accepted 7 November 2013 Published online in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.9395
Introduction
Surgical resection is the cornerstone of treatment for patients with colonic cancer. Generally, restoration of bowel continuity with a primary anastomosis is pursued in uncomplicated colonic resections. The most serious complication of colonic surgery with restoration of bowel continuity is anastomotic leakage (AL)1 , which is associated with the possible need for reintervention, increased mortality2,3 and, possibly, worse oncological outcome4,5 . The reported incidence of AL in colonic anastomosis varies between 3 and 6·4 per cent, depending on patient and tumour characteristics, definitions, site of the anastomosis and perhaps surgeon caseload6 – 9 . Several risk factors, including co-morbidity, higher American Society of Anesthesiologists (ASA) fitness grade, stage of disease, type of operation, emergency surgery and intraoperative complications, have all been associated 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
with AL7,10 – 12 . Concentration of surgery in high-volume centres has been considered as a strategy to improve quality of care, surgical outcomes and mortality13,14 . Hospital procedural volume could also be a possible risk factor for AL. Although AL has long been a subject of debate, prediction of the risk of leakage for the individual patient remains difficult. The Dutch Surgical Colorectal Audit (DSCA), a clinical outcome registry in which all Dutch hospitals participate, was initiated in 2009 to monitor and improve outcome of surgical care for patients with colorectal cancer. The DSCA facilitates individual hospitals in quality improvement projects, but is also used to identify treatment and outcome patterns for different patient groups. AL after colorectal resection has been selected as an outcome indicator for surgical quality of care in the DSCA15 . In rectal cancer surgery, the practice of routine construction BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
of a defunctioning stoma may play a large role in measuring this outcome and determining risk factors16 , whereas stoma construction may be of less importance in colonic cancer resections. Among all colorectal surgical procedures, patients undergoing colonic cancer resection may be considered a specifically vulnerable group, being at risk for morbidity and mortality because of advanced age and co-morbidity17 . Risk factors for AL and related postoperative mortality have not yet been investigated in this group. The aim of the present study was to identify risk factors for AL after colonic cancer resection and factors influencing mortality associated with AL using patient information from a national audit database.
425
grade, previous abdominal surgery), tumour factors (stage, location, preoperative tumour complications) and treatment factors (preoperative surgical procedures (stoma or other), type of surgical resection, urgency of surgery, extent of resection, fashioning of a defunctioning stoma), were extracted from the database. Hospitals were categorized as low-volume (50 or fewer), medium-volume (51–100) or high-volume (more than 100) centres, based on the number of patients who underwent surgical treatment of colonic cancer per year for the years 2010 and 2011 (in 2009 not all hospitals completed registration). These categories reflect the present situation in the Netherlands, with 50 per cent of clinics performing between 50 and 100 colonic cancer resections annually22 .
Methods
The data set was retrieved from the DSCA, a web-based national database, in which information on all patients undergoing surgical resection for colorectal cancer is recorded18 . Data on patient and tumour characteristics, diagnostics, treatment and outcome were collected. The data set included information registered from 92 hospitals with nearly 100 per cent concordance on validation against the National Cancer Registry. In 2009, 89 per cent of Dutch hospitals participated, increasing to 99 per cent in 2010 and 201119 . Medical ethics committee approval was not required for this study as all patients and hospital information in the DSCA was deidentified. Individual patient data were collected in the treating hospital and transferred in encrypted form to the DSCA database.
Statistical analysis Univariable analyses were carried out using the χ2 test to examine the association between patient, tumour, treatment and hospital factors and the occurrence of AL and death after AL. Multivariable logistic regression analyses were performed to correct for possible confounders. A manual stepwise model was used, with inclusion of variables with P < 0·050 in univariable analysis. Clinically relevant variables were added to the statistical model. The variables ‘timing of surgery’ (elective/emergency) and ‘preoperative tumour complications’ were assumed to indicate the same clinical situation. To check for collinearity when both variables were incorporated in the model, the variance inflation factor was computed. To check the model, the multivariable analysis was repeated with the outcome AL,
Inclusion and exclusion criteria All Dutch patients who underwent colonic cancer resection in the Netherlands between January 2009 and December 2011 were included in this study. Patients with rectal cancer, those with multiple synchronous tumours, patients without a primary anastomosis and those with an unknown surgical resection type were excluded from analysis. Surgical resections were categorized as ileocaecal resection, right colectomy, transverse resection, left colectomy, sigmoid resection and subtotal colectomy.
DSCA 2009–2011 Resection for primary colorectal cancer n = 27 259 Excluded n = 8557 Patients with rectal cancer n = 7614 Patients with multiple synchronous tumours n = 943 Patients with colonic cancer n = 18 702
Outcomes Primary outcome measures were AL, defined as clinically significant leakage requiring surgical or radiological reintervention, and mortality after AL, defined as death in hospital within 30 days after primary surgery. Potential risk factors for postoperative complications, including patient factors (age, sex, body mass index (BMI), Charlson co-morbidity score20,21 , ASA fitness 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Excluded n = 3035 No anastomosis n = 2581 Unknown/other resection type n = 454 Patients who underwent colonic cancer resection with primary anastomosis n = 15 667 Fig. 1
Study flow chart. DSCA, Dutch Surgical Colorectal Audit
www.bjs.co.uk
BJS 2014; 101: 424–432
426
Table 1
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
Univariable analyses of patient variables associated with anastomotic leakage and with death after anastomotic leakage No. of patients (n = 15 667)
Age (years) < 65 65–80 > 80 Unknown Sex F M BMI (kg/m2 ) < 20 20–25 25–30 > 30 Unknown ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Previous abdominal surgery Yes No
Anastomotic leakage (n = 1176)
P*
Death after anastomotic leakage (n = 193)
0·284 4825 (30·8) 7616 (48·6) 3211 (20·5) 15 (0·1)
386 (8·0) 551 (7·2) 239 (7·4) 0 (0)
7605 (48·5) 8062 (51·5)
477 (6·3) 699 (8·7)
623 (4·0) 4922 (31·4) 4986 (31·8) 2079 (13·3) 3057 (19·5)
54 (8·7) 366 (7·4) 389 (7·8) 169 (8·1) 198 (6·5)
11 638 (74·3) 3713 (23·7) 316 (2·0)
822 (7·1) 340 (9·2) 14 (4·4)
8335 (53·2) 3579 (22·8) 3753 (24·0)
583 (7·0) 285 (8·0) 308 (8·2)
5309 (33·9) 10 358 (66·1)
425 (8·0) 751 (7·3)
P* < 0·001
20 (5·2) 101 (18·3) 72 (30·1) 0 (0) < 0·001
0·783 80 (16·8) 113 (16·2)
0·643
0·535 11 (20) 59 (16·1) 54 (13·9) 29 (17·2) 40 (20·2)
< 0·001
< 0·001 94 (11·4) 96 (28·2) 3 (21)
0·032
< 0·001 56 (9·6) 59 (20·7) 78 (25·3)
0·090
0·065 81 (19·1) 112 (14·9)
Values in parentheses are percentages. BMI, body mass index; ASA, American Society of Anesthesiologists. *χ2 test.
first without ‘preoperative tumour complications’ and including ‘timing of surgery’. The same analysis was then repeated but including ‘preoperative tumour complications’ and without ‘timing of surgery’. Results are reported as odds ratios (ORs) with 95 per cent confidence intervals (c.i.). Two-sided P < 0·050 was considered statistically significant. All data were analysed using PASW Statistics, release 20.0.0.1 (IBM, Armonk, New York, USA). Results
From 2009 to 2011, data from 27 259 patients were included in the DSCA database (Fig. 1). After exclusion of those without colonic cancer, 18 702 patients were eligible for inclusion. Following exclusion of a further 2581 patients without a primary anastomosis and 454 who underwent total colectomy or an unknown resection, 15 667 patients with colonic cancer remained for analysis. Surgical procedures included: 240 ileocaecal resections, 7788 right colectomies, 527 transverse resections, 1601 left colectomies, 5354 sigmoid resections and 157 subtotal colectomies. Surgery was performed in 92 hospitals, with 17·1 per cent of the patients being treated in a low-volume centre, 54·0 per cent in a medium-volume centre and 28·9 per cent in a high-volume centre. 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Anastomotic leakage AL leading to reintervention occurred in 1176 patients (7·5 per cent). The reinterventions were laparotomy (82·1 per cent), laparoscopy (2·8 per cent), radiological drainage (8·2 per cent), and interventions such as drainage of wounds and abscesses (6·9 per cent). During the primary operation, a defunctioning stoma was created in 606 patients (3·9 per cent), usually after sigmoid resection. Of all anastomoses created after sigmoid resection, 8·7 per cent were diverted. There was no difference in AL rate between patients with and without a defunctioning stoma (6·4 versus 7·5 per cent respectively; P = 0·308). The incidence of AL was significantly higher after resection of the transverse colon, left colectomy and subtotal colectomy than after other types of resection. A secondary stoma was created in 805 patients (68·5 per cent) requiring a surgical or radiological reintervention for AL.
Risk factors for anastomotic leakage Tables 1–3 show the results of univariable analyses of possible risk factors for the development of AL. Patient factors associated with an increased risk of AL were male sex, higher ASA grade and higher Charlson score. Among tumour factors, preoperative tumour complications www.bjs.co.uk
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
Table 2
427
Univariable analyses of tumour characteristics associated with anastomotic leakage and with death after anastomotic
leakage
Preop. tumour complications None Perforation Obstruction Blood loss Other Tumour stage 0 1 2 3 4 Unknown Tumour location Caecum Ascending colon Hepatic flexure Transverse colon Splenic flexure Descending colon Sigmoid colon
No. of patients (n = 15 667)
Anastomotic leakage (n = 1176)
11 968 (76·4) 356 (2·3) 1557 (9·9) 951 (6·1) 835 (5·3)
820 (6·9) 43 (12·1) 175 (11·2) 65 (6·8) 73 (8·7)
153 (1·0) 2811 (17·9) 5769 (36·8) 4918 (31·4) 1784 (11·4) 232 (1·5)
13 (8·5) 180 (6·4) 461 (8·0) 376 (7·6) 131 (7·3) 15 (6·5)
3513 (22·4) 3085 (19·7) 1064 (6·8) 1168 (7·5) 448 (2·9) 887 (5·7) 5502 (35·1)
216 (6·1) 199 (6·5) 70 (6·6) 122 (10·4) 54 (12·1) 93 (10·5) 422 (7·7)
P*
Death after anastomotic leakage (n = 193)
< 0·001
P* 0·100
130 (15·9) 8 (19) 30 (17·1) 6 (9) 19 (26) 0·176
0·782 3 (23) 31 (17·2) 67 (14·5) 67 (17·8) 23 (17·6) 2 (13)
< 0·001
0·225 45 (20·8) 35 (17·6) 15 (21) 20 (16·4) 9 (17) 11 (12) 58 (13·7)
Values in parentheses are percentages. *χ2 test. Table 3
Univariable analyses of treatment characteristics associated with anastomotic leakage and with death after anastomotic
leakage
Preop. surgical procedure None Stoma formation Other Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Timing of surgery Elective Emergency Unknown Stoma No stoma Defunctioning stoma Extensive resection No Yes Hospital volume Low Medium High
No. of patients (n = 15 667)
Anastomotic leakage (n = 1176)
15 285 (97·6) 118 (0·8) 264 (1·7)
1133 (7·4) 15 (12·7) 28 (10·6)
7788 (49·7) 240 (1·5) 527 (3·4) 1601 (10·2) 5354 (34·2) 157 (1·0)
495 (6·4) 18 (7·5) 57 (10·8) 172 (10·7) 413 (7·7) 21 (13·4)
13 139 (83·9) 1625 (10·4) 903 (5·8)
925 (7·0) 159 (9·8) 92 (10·2)
P*
Death after anastomotic leakage (n = 193)
0·015
0·383 188 (16·6) 3 (20) 2 (7)
< 0·001
0·029 101 (20·4) 3 (17) 10 (18) 20 (11·6) 58 (14·0) 1 (5)
< 0·001
0·028 145 (15·7) 37 (23·3) 11 (12)
0·308 15 061 (96·1) 606 (3·9)
1137 (7·5) 39 (6·4)
14 216 (90·7) 1451 (9·3)
1025 (7·2) 151 (10·4)
2680 (17·1) 8461 (54·0) 4526 (28·9)
180 (6·7) 653 (7·7) 343 (7·6)
P*
0·860 187 (16·4) 6 (15)
< 0·001
0·321 164 (16·0) 29 (19·2)
0·225
0·162 38 (21·1) 99 (15·2) 56 (16·3)
Values in parentheses are percentages. Hospital volume: low, 50 or fewer patients per year; medium, 51–100 patients per year; high, more than 100 patients per year. *χ2 test.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
BJS 2014; 101: 424–432
428
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
Multivariable logistic regression analysis of risk factors for anastomotic leakage among patients who underwent colonic surgery with a primary colonic anastomosis
Mortality after anastomotic leakage
Table 4
Odds ratio Sex F M ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Preop. surgical procedures None Stoma Other Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Extensive resection No Yes Stoma No stoma Defunctioning stoma Time of surgery Elective Emergency Unknown
P
Age > 80 years ASA grade III–IV Charlson score ≥ 2
1·00 (reference) 1·38 (1·22, 1·56)
< 0·001
Right colectomy
1·00 (reference) 1·26 (1·09, 1·45) 0·59 (0·34, 1·02)
0·002 0·075
Transverse resection
1·00 (reference) 1·11 (0·95, 1·29) 1·10 (0·94, 1·29)
Emergency surgery
Ileocaecal resection
Left colectomy Sigmoid resection
0·178 0·218
Subtotal colectomy Overall rate 0
1·00 (reference) 1·52 (0·87, 2·65) 1·21 (0·81, 1·81)
0·139 0·361
1·00 (reference) 1·13 (0·69, 1·85) 1·69 (1·26, 2·26) 1·69 (1·40, 2·03) 1·28 (1·11, 1·47) 2·28 (1·42, 3·66)
0·630 < 0·001 < 0·001 0·001 0·001
1·00 (reference) 1·43 (1·19, 1·72)
< 0·001
1·00 (reference) 0·68 (0·49, 0·96)
0·026
1·00 (reference) 1·33 (1·11, 1·59) 1·55 (1·23, 1·96)
0·002 < 0·001
Values in parentheses are 95 per cent confidence intervals. ASA, American Society of Anesthesiologists.
(mostly tumour perforation or obstruction) and tumour location were associated with an increased risk of AL. Treatment factors associated with a higher risk of AL were preoperative surgical procedures (such as construction of a defunctioning stoma), extensive resection (resection of other organs during surgery), emergency surgery, and type of resection, especially transverse resection, left colectomy and subtotal colectomy. Multivariable analyses confirmed male sex and a high ASA grade as independent risk factors for AL (Table 4). Treatment factors that remained associated with a higher risk of AL were type of resection (mainly transverse resection, left colectomy and subtotal colectomy compared with right colectomy as reference group), extensive resection and emergency surgery. On clinical grounds, the variable ‘defunctioning stoma’ was added to the model and adjusted data also showed a lower risk of AL in patients 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
Anastomotic leakage
5
10
15
20
25
30
35
Rate (%)
Anastomotic leakage and subsequent mortality rates after colonic resection and primary anastomosis among patients with risk factors identified by univariable analysis, and overall rates for the whole cohort. ASA, American Society of Anesthesiologists
Fig. 2
with a defunctioning stoma (OR 0·68, 95 per cent c.i. 0·49 to 0·96). To check for the presence of collinearity between the variables ‘timing of surgery’ and ‘preoperative tumour complications’, the variance inflation factor was computed. The results indicated no collinearity between these variables. Analysis of the multivariable model including ‘preoperative tumour complications’ instead of ‘timing of surgery’ showed a significantly higher risk of AL in patients with preoperative tumour complications, such as perforation (OR 1·68, 1·26 to 2·26) and obstruction (1·62, 1·39 to 2·02).
Mortality Some 648 (4·1 per cent) of the 15 667 patients died within 30 days of surgery, 3·4 per cent after elective surgery and 7·2 per cent after emergency operation (P < 0·001). AL was diagnosed after surgery in 193 (29·8 per cent) of those who died. The mortality rate among patients with AL was significantly higher than in those without leakage (16·4 versus 3·1 per cent; P < 0·001). There was no significant association between the number of patients treated yearly per hospital and mortality after AL (P = 0·162).
Risk factors for death after anastomotic leakage Univariable analyses revealed that patient factors associated with death after AL were older age, high ASA grade and a high Charlson score (Table 1). Surgery in the www.bjs.co.uk
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
429
Multivariable logistic regression analysis of risk factors for death after anastomotic leakage among patients who underwent colonic surgery with a primary colonic anastomosis
Table 5
Age (years) < 65 65–80 > 80 ASA fitness grade I–II III–IV Unknown Charlson score 0 I ≥ II Surgical resection Right colectomy Ileocaecal resection Transverse resection Left colectomy Sigmoid resection Subtotal colectomy Timing of surgery Elective Emergency Unknown
Odds ratio
P
1·00 (reference) 3·15 (1·89, 5·27) 5·16 (2·98, 8·96)
< 0·001 < 0·001
1·00 (reference) 1·77 (1·24, 2·52) 1·89 (0·48, 7·47)
0·002 0·363
1·00 (reference) 1·76 (1·16, 2·69) 2·23 (1·47, 3·37)
0·008 < 0·001
1·00 (reference) 1·00 (0·25, 3·94) 0·80 (0·38, 1·71) 0·54 (0·31, 0·92) 0·75 (0·51, 1·08) 0·28 (0·04, 2·24)
0·998 0·566 0·025 0·124 0·232
1·00 (reference) 1·75 (1·12, 2·73) 0·78 (0·39, 1·57)
0·014 0·483
Values in parentheses are 95 per cent confidence intervals. ASA, American Society of Anesthesiologists.
emergency setting and type of surgical resection were also associated with a higher risk (Table 3). In particular, patients undergoing right colectomy, transverse resection or ileocaecal resection had high mortality rates after AL (Fig. 2). After adjustment for possible confounders, multivariable analyses showed that older age, high ASA grade, higher Charlson score and emergency surgery were independent risk factors for death among patients diagnosed with AL (Table 5). Multivariable analysis also revealed that the risk of death associated with AL was lower after left colectomy compared with other surgical resection types. Discussion
In the present study of risk factors for AL and mortality following colonic cancer resection with an anastomosis, the overall AL rate was 7·5 per cent. The mortality rate was 4·1 per cent overall, and was significantly higher for patients with AL than in those without (16·4 versus 3·1 per cent). The present leak rate is higher than in recent reports (3–6·4 per cent)6 – 8 . This might be attributed to the complete registration in a clinical audit. Reported results from a nationwide study on AL in Denmark also showed a rather high rate of 6·4 per cent7 . Other publications 2014 BJS Society Ltd Published by John Wiley & Sons Ltd
with lower rates usually originated from dedicated centres. Analyses adjusted for confounding factors revealed male sex and high ASA grade as independent risk factors for AL, which is consistent with the literature7,23,24 . Other reported predictors of AL, such as previous abdominal surgery24 and high BMI3,25 , were not confirmed in the present analysis. Another well known risk factor for AL is co-morbidity8,11,24 , reflected in the Charlson score and ASA classification in the present study. Both scores were associated with AL in univariable analyses, and ASA grade remained a significant predictor in multivariable analyses. Adjusted analysis in the present study indicated that treatment factors such as extensive resection and type of surgical resection were independent risk factors for AL. The incidence of AL differs depending on tumour location and subsequent type of resection. Right-sided colectomy is associated with a lower leak rate than leftsided colectomy3,26 , and the AL rate is higher after transverse resection11 . Differences in vascularization of the anastomotic site may explain this as a well vascularized ileal bowel loop is anastomosed to an adequately vascularized transverse colonic loop in right-sided hemicolectomy. However, in transverse or left colonic resection, where the middle colic artery or inferior mesenteric artery is divided, vascularization of the anastomotic site might be compromised27,28 . Another possible explanation is the lack of full mobilization of one or both flexures. The poorer outcome after transverse resection in the present study emphasizes the importance of careful surgical decisionmaking. For patients requiring a transverse resection, an extended colectomy could be a better option. Data on other known risk factors such as weight loss26 , intraoperative complications, duration of operation, blood loss and faecal contamination8,11 were not available in the DSCA database. The overall mortality rate of 4·1 per cent in the present analysis is within the range in population-based studies (3·0–7·4 per cent)6,29,30 . The mortality rate of 16·4 per cent following AL is high, as noted previously (12·0–18·6 per cent)6,11 , and related to old age and comorbidity. The mortality rate after AL is higher after colonic than after rectal resection (0·7–4 per cent)19,31,32 . This might be explained by differences in anatomy. AL after colonic resection often results in a generalized peritonitis, compared with more local, extraperitoneal abscess formation after rectal resection. Leakage scores have been developed for early detection of AL after rectal surgery29,33 . It would be of interest to validate these scores for colonic resections, as the severity of the consequences of leakage from a colonic anastomosis is underestimated. www.bjs.co.uk
BJS 2014; 101: 424–432
430
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
To reduce the incidence of AL or its clinical sequelae, a defunctioning stoma can be constructed. In rectal anastomoses, temporary defunctioning stomas are fashioned to reduce the clinical consequences of leakage10,34 . In the present series most defunctioning stomas were constructed after sigmoid resection. A defunctioning stoma had no impact on the incidence of AL in univariable analysis but, after adjustment for patient and tumour characteristics, multivariable analysis showed a protective effect in terms of AL. Apparently there was good patient selection for faecal diversion, based on the preoperative or intraoperative judgement of the surgeons concerning the risk of AL. Emergency surgery is also considered a risk factor for both AL12,35 and postoperative mortality12,30,36 . Poor general condition and nutritional status is associated with a higher risk of morbidity and mortality in these patients36,37 . In the present study, emergency surgery was identified as an independent risk factor for both AL and mortality following leakage. Further multivariable analysis also confirmed that patients with preoperative complications such as tumour perforation and obstruction had a higher risk of postoperative AL. Emergency surgery is frequently performed during evening and night shifts. Studies from different medical fields have reported worse postoperative outcome after operations performed at these times38 – 40 . Surgical procedures during late shifts are sometimes performed by surgeons with a lower disease-specific caseload. Some studies have suggested that the experience and caseload of the individual surgeon are predictors of postoperative death9,41 . Unfortunately, the DSCA database does not contain data relating to individual surgeons. The strength of this study is that the results are based on a complete and large nationwide data set, which included registered data from all Dutch hospitals performing colorectal surgery. Validation of the registered data against the national cancer registry showed a high concordance. Therefore a valid analysis of colonic cancer surgery in the Netherlands could be made. However, several limitations are worth mentioning. A somewhat heterogeneous study population was analysed. All patients undergoing resection for colonic cancer were included, including those with stage IV disease, who may represent both curative and palliative resections and operations in the emergency setting. Although the analysis of such a complete cohort leads to fair results, the heterogeneity is also accompanied by confounding factors and might lead to bias. To control for bias, multivariable analysis was carried out stratified for time of surgery (elective versus emergency). The main results of analyses for elective and
emergency subgroups did not differ significantly from those of the total study population (data not shown). Therefore, ‘timing of surgery’ was included as a possible risk factor for AL and subsequent death in the multivariable model. Another limitation of this data set is that only patients treated for cancer could be analysed, although there are also benign indications for colonic surgery. Furthermore, only limited intraoperative data were available. The data set contains no information on duration of operation, blood loss or surgical techniques. Intraoperative information could make a valuable contribution to the identification of risk factors for AL. The same also applies to data regarding surgeon caseload. Colonic cancer resections are common surgical procedures. Although the exact mechanism leading to the development of AL is often unknown for the individual patient, and the clinical risk assessment by the operating surgeon is of low predictive value42 , it is important to understand that tumour and treatment factors may play an important role. Mortality rates following AL are high, and determined mainly by patient factors such as age and co-morbidity. Therefore, it is important to monitor patients after surgery according to standardized surveillance protocols, perhaps incorporating colonic leakage scores designed for left-sided colorectal procedures29,33 . Future research should be focused on further evaluation of leakage scores in colonic resections, and on continuous monitoring through clinical audit.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
Acknowledgements
The authors thank M. Fiocco for the statistical guidance and advice in the present study. Disclosure: The authors declare no conflict of interest. References 1 Khan AA, Wheeler JM, Cunningham C, George B, Kettlewell M, Mortensen NJ. The management and outcome of anastomotic leaks in colorectal surgery. Colorectal Dis 2008; 10: 587–592. 2 van Westreenen HL, Ijpma FF, Wevers KP, Afzali H, Patijn GA. Reoperation after colorectal surgery is an independent predictor of the 1-year mortality rate. Dis Colon Rectum 2011; 54: 1438–1442. 3 Buchs NC, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B, Morel P. Incidence, consequences, and risk factors for anastomotic dehiscence after colorectal surgery: a prospective monocentric study. Int J Colorectal Dis 2008; 23: 265–270. 4 Petersen S, Freitag M, Hellmich G, Ludwig K. Anastomotic leakage: impact on local recurrence and survival in surgery of colorectal cancer. Int J Colorectal Dis 1998; 13: 160–163. 5 Law WL, Choi HK, Lee YM, Ho JW, Seto CL. Anastomotic leakage is associated with poor long-term
BJS 2014; 101: 424–432
Anastomotic leakage after colonic cancer surgery
6
7
8
9
10
11
12
13
14
15
16
17
18
outcome in patients after curative colorectal resection for malignancy. J Gastrointest Surg 2007; 11: 8–15. Kube R, Mroczkowski P, Granowski D, Benedix F, Sahm M, Schmidt U et al.; Study Group Quality Assurance in Primary Colorectal Carcinoma. Anastomotic leakage after colon cancer surgery: a predictor of significant morbidity and hospital mortality, and diminished tumour-free survival. Eur J Surg Oncol 2010; 36: 120–124. Krarup PM, Jorgensen LN, Andreasen AH, Harling H; Danish Colorectal Cancer Group. A nationwide study on anastomotic leakage after colonic cancer surgery. Colorectal Dis 2012; 14: 661–667. Leichtle SW, Mouawad NJ, Welch KB, Lampman RM, Cleary RK. Risk factors for anastomotic leakage after colectomy. Dis Colon Rectum 2012; 55: 569–575. Renzulli P, Lowy A, Maibach R, Egeli RA, Metzger U, Laffer UT. The influence of the surgeon’s and the hospital’s caseload on survival and local recurrence after colorectal cancer surgery. Surgery 2006; 139: 296–304. Peeters KC, Tollenaar RA, Marijnen CA, Klein Kranenbarg E, Steup WH, Wiggers T et al.; Dutch Colorectal Cancer Group. Risk factors for anastomotic failure after total mesorectal excision of rectal cancer. Br J Surg 2005; 92: 211–216. Alves A, Panis Y, Trancart D, Regimbeau JM, Pocard M, Valleur P. Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World J Surg 2002; 26: 499–502. Buchs NC, Gervaz P, Bucher P, Huber O, Mentha G, Morel P. Lessons learned from one thousand consecutive colonic resections in a teaching hospital. Swiss Med Wkly 2007; 137: 259–264. Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA 2000; 284: 3028–3035. Mroczkowski P, Kube R, Ptok H, Schmidt U, Hac S, ¨ Kockerling F et al. Low-volume centre vs high-volume: the role of a quality assurance programme in colon cancer surgery. Colorectal Dis 2011; 13: e276–e283. Snijders HS, Henneman D, van Leersum NL, ten Berge M, Fiocco M, Karsten TM et al. Anastomotic leakage as an outcome measure for quality of colorectal cancer surgery. BMJ Qual Saf 2013; 22: 759–767. Snijders HS, van den Broek CB, Wouters MW, Meershoek-Klein Kranenbarg E, Wiggers T, Rutten H et al. An increasing use of defunctioning stomas after low anterior resection for rectal cancer. Is this the way to go? Eur J Surg Oncol 2013; 39: 715–720. Lemmens VE, Janssen-Heijnen ML, Houterman S, Verheij KD, Martijn H, van de Poll-Franse L et al. Which comorbid conditions predict complications after surgery for colorectal cancer? World J Surg 2007; 31: 192–199. Van Leersum NJ, Snijders HS, Henneman D, Kolfschoten NE, Gooiker GA, ten Berge MG et al. The Dutch Surgical Colorectal Audit. Eur J Surg Oncol 2013; 39: 1063–1070.
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
431
19 Dutch Institute for Clinical Auditing. Rapportages 2011. http://www.clinicalaudit.nl/jaarrapportage/archief/DICA %20Jaarrapportage%202011.pdf [accessed 26 November 2013]. 20 Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383. 21 Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 1245–1251. 22 Elferink MA, Wouters MW, Krijnen P, Lemmens VE, Jansen-Landheer ML, van de Velde CJ et al. Disparities in quality of care for colon cancer between hospitals in the Netherlands. Eur J Surg Oncol 2010; 36(Suppl 1): S64–S73. 23 Choi HK, Law WL, Ho JW. Leakage after resection and intraperitoneal anastomosis for colorectal malignancy: analysis of risk factors. Dis Colon Rectum 2006; 49: 1719–1725. 24 Lipska MA, Bissett IP, Parry BR, Merrie AE. Anastomotic leakage after lower gastrointestinal anastomosis: men are at a higher risk. ANZ J Surg 2006; 76: 579–585. 25 Biondo S, Par´es D, Kreisler E, Ragu´e JM, Fraccalvieri D, Ruiz AG et al. Anastomotic dehiscence after resection and primary anastomosis in left-sided colonic emergencies. Dis Colon Rectum 2005; 48: 2272–2280. 26 Veyrie N, Ata T, Muscari F, Couchard AC, Msika S, Hay JM et al. Anastomotic leakage after elective right versus left colectomy for cancer: prevalence and independent risk factors. J Am Coll Surg 2007; 205: 785–793. 27 Drummond H. Some points relating to the surgical anatomy of the arterial supply of the large intestine. Proc R Soc Med 1914; 7(Surg Sect): 185–193. 28 Hall NR, Finan PJ, Stephenson BM, Lowndes RH, Young HL. High tie of the inferior mesenteric artery in distal colorectal resections – a safe vascular procedure. Int J Colorectal Dis 1995; 10: 29–32. 29 den Dulk M, Noter SL, Hendriks ER, Brouwers MA, van der Vlies CH, Oostenbroek RJ et al. Improved diagnosis and treatment of anastomotic leakage after colorectal surgery. Eur J Surg Oncol 2009; 35: 420–426. 30 Sjo OH, Larsen S, Lunde OC, Nesbakken A. Short term outcome after emergency and elective surgery for colon cancer. Colorectal Dis 2009; 11: 733–739. 31 den Dulk M, Marijnen CA, Collette L, Putter H, P˚ahlman L, Folkesson J et al. Multicentre analysis of oncological and survival outcomes following anastomotic leakage after rectal cancer surgery. Br J Surg 2009; 96: 1066–1075. 32 Snijders HS, Wouters MW, van Leersum NJ, Kolfschoten NE, Henneman D, de Vries AC et al. Meta-analysis of the risk for anastomotic leakage, the postoperative mortality caused by leakage in relation to the overall postoperative mortality. Eur J Surg Oncol 2012; 38: 1013–1019. 33 Dekker JW, Liefers GJ, de Mol van Otterloo JC, Putter H, Tollenaar RA. Predicting the risk of anastomotic leakage in
www.bjs.co.uk
BJS 2014; 101: 424–432
432
34
35
36
37
I. S. Bakker, I. Grossmann, D. Henneman, K. Havenga and T. Wiggers
left-sided colorectal surgery using a colon leakage score. J Surg Res 2011; 166: e27–e34. Shiomi A, Ito M, Saito N, Ohue M, Hirai T, Kubo Y et al. Diverting stoma in rectal cancer surgery. A retrospective study of 329 patients from Japanese cancer centers. Int J Colorectal Dis 2011; 26: 79–87. Kolfschoten NE, Wouters MW, Gooiker GA, Eddes EH, Kievit J, Tollenaar RA et al.; Dutch Surgical Colorectal Audit Group. Nonelective colon cancer resections in elderly patients: results from the Dutch Surgical Colorectal Audit. Dig Surg 2012; 29: 412–419. Zielinski MD, Merchea A, Heller SF, You YN. Emergency management of perforated colon cancers: how aggressive should we be? J Gastrointest Surg 2011; 15: 2232–2238. Lee YM, Law WL, Chu KW, Poon RT. Emergency surgery for obstructing colorectal cancers: a comparison between right-sided and left-sided lesions. J Am Coll Surg 2001; 192: 719–725.
38 Peled Y, Melamed N, Chen R, Pardo J, Ben-Shitrit G, Yogev Y. The effect of time of day on outcome of unscheduled cesarean deliveries. J Matern Fetal Neonatal Med 2011; 24: 1051–1054. 39 Chacko AT, Ramirez MA, Ramappa AJ, Richardson LC, Appleton PT, Rodriguez EK. Does late night hip surgery affect outcome? J Trauma 2011; 71: 447–453. 40 Coumbe A, John R, Kuskowski M, Agirbasli M, McFalls EO, Adabag S. Variation of mortality after coronary artery bypass surgery in relation to hour, day and month of the procedure. BMC Cardiovasc Disord 2011; 11: 63. 41 Schrag D, Panageas KS, Riedel E, Hsieh L, Bach PB, Guillem JG et al. Surgeon volume compared to hospital volume as a predictor of outcome following primary colon cancer resection. J Surg Oncol 2003; 83: 68–78. 42 Karliczek A, Harlaar NJ, Zeebregts CJ, Wiggers T, Baas PC, van Dam GM. Surgeons lack predictive accuracy for anastomotic leakage in gastrointestinal surgery. Int J Colorectal Dis 2009; 24: 569–576.
Editor’s commentary
Anastomotic leakage is one of the most feared complications after gastrointestinal surgery; the rate varies with the type of surgery and the location of the anastomosis. As evident from the two studies in this issue of BJS, laparoscopic bariatric surgery has a considerably lower leak rate (less than 2 per cent) than colonic surgery (less than 8 per cent), yet both can be fatal. While no-one would argue against careful patient selection and vigilant postoperative surveillance, it may be necessary to think outside the box to further reduce and, eventually, prevent this dreaded complication. In the era of personalized medicine, one could envision future staplers with biosensors and feedback systems to optimize closure and ensure decent blood flow rather than the non-standardized and blunt ‘one-size-fits-all’ approach currently available. Or is it the gut microbiome that makes some patients’ suture line prone to delayed healing and leakage? The BJS welcomes novel research that further highlights our understanding of surgical technology and biological processes in this regard. K. Søreide Editor, BJS
2014 BJS Society Ltd Published by John Wiley & Sons Ltd
www.bjs.co.uk
BJS 2014; 101: 424–432
