The American Journal of Surgery (2015) 210, 1-5
Clinical Science
Smoking as dominant risk factor for anastomotic leak after left colon resection Rebeccah B. Baucom, M.D.*, Benjamin K. Poulose, M.D., M.P.H., Alan J. Herline, M.D., Roberta L. Muldoon, M.D., Molly M. Cone, M.D., Timothy M. Geiger, M.D. Division of General Surgery, Vanderbilt University Medical Center, D-5203 Medical Center North, 1161 Medical Center Drive, Nashville, TN 37232, USA KEYWORDS: Smoking; Anastomotic leak; Colectomy; Left colectomy; Risk factors
Abstract BACKGROUND: Some risk factors for anastomotic leak have been identified, but the effect of smoking is unknown. METHODS: This study aimed to evaluate the effect of smoking on clinical leak after left-sided anastomoses. Adult patients who underwent elective left colectomy between January 1, 2008 and December 31, 2012 were included. Those with stomas and inflammatory bowel diseases were excluded. Primary outcome was anastomotic leak requiring percutaneous drainage or operative intervention within 30 days. RESULTS: There were 246 patients included; 56% were female. Most had a diagnosis of diverticular disease (53%) or cancer (37%). Anastomotic leak rate was 6.5% (n 5 16). The rate in smokers was 17% versus 5% in nonsmokers (P 5 .01). Smokers had over 4 times greater chance of leak (odds ratio 4.2, 95% confidence interval 1.3 to 13.5, P 5 .02). CONCLUSION: Smoking is a risk factor for leak after left colectomy. Consideration should be given to delaying elective left colectomy until smoking cessation is achieved. Ó 2015 Elsevier Inc. All rights reserved.
Clinically significant leak occurs in 1% to 12% of colorectal operations and up to 14% after distal colorectal resections.1–5 Anastomotic leak significantly increases morbidity and mortality, with mortality reported between The authors declare no conflicts of interest. Data were collected and managed using REDCap electronic data capture tool, which is supported by the UL1 TR000445 grant from NCATS/ NIH and hosted at Vanderbilt University Medical Center. Presented as a podium presentation at the American Society of Colon and Rectal Surgeons, May 17–21, 2014, Hollywood, Florida. * Corresponding author. Tel.: 11-615-343-5613; fax: 11-615-3439485. E-mail address: [email protected] Manuscript received September 11, 2014; revised manuscript September 30, 2014 0002-9610/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2014.10.033
6% and 13%.2,5,6 Length of hospitalization can increase up to 4 times for those patients with a leak compared with those without a leak.3,7 Although anastomotic leak remains a challenging complication following colon and rectal surgery, solid evidence evaluating specific risk factors for anastomotic leak is lacking. A recent prospective study evaluating risk factors for leak after colorectal surgery found that low anastomosis, increasing number of comorbidities, high ligation of the inferior mesenteric artery, intraoperative complications, and male sex are significant risk factors.2 However, smoking was not specifically evaluated in this study. In 2013, the Surgeon General released an updated report on the health consequences of smoking.8 Important to the colorectal patient population, the report linked smoking with colorectal cancer and inflammatory bowel disease.
2
The American Journal of Surgery, Vol 210, No 1, July 2015
The report demonstrates tobacco’s continued burden to the nation’s health with its clear impact on all-cause mortality in men and women. The annual cost for direct medical care of adults attributable to smoking between the years 2009 and 2012 was $132.5 to 175.9 billion. Smoking is known to cause microvascular disease, tissue ischemia, and hypoxia, which are contributing factors to poor anastomotic healing. The purpose of this study was to evaluate smoking as an independent, modifiable risk factor for clinical leak after left-sided colorectal anastomoses.
reflection (low anterior), at the level of the sigmoid, or above the level of the sigmoid. The primary outcome measure was postoperative clinical anastomotic leak within 30 days of surgery. A clinical leak was defined as air or fluid near the site of the anastomosis, requiring either return to the operating room or percutaneous drainage. Secondary outcome measures included organ space infection and 30-day mortality. The study was performed with approval of the Vanderbilt University Institutional Review Board and Human Research Protection Program.
Patients and Methods
Statistical considerations
Study design and data sources
Descriptive statistics for the overall study population were calculated, as well as for the group of smokers compared with nonsmokers. Chi-square and Fisher’s exact tests were used to compare proportions, as appropriate. Multivariable logistic regression was performed, with covariates determined a priori with a 2-tailed alpha level of .05 determining significance. The number of events limited the inclusion of all possible factors, so those that were known, strong risk factors, were given priority. The final model included smoking and anastomotic level. Additional covariates which were given consideration in the modeling process included sex, body mass index, recent chemotherapy or radiation therapy, diagnosis, and type of leak test performed. Stata version 13.1 (StataCorp LC, College Station, TX) was used for statistical analysis.
A retrospective cohort study was performed including adult patients who underwent left colectomy between January 1, 2008 and December 31, 2012, by one of the 4 board-certified colorectal surgeons at a single, tertiary care center. These patients were identified from the Vanderbilt Procedural Outcomes Databasedan institutional database that includes clinical and administrative information on patients undergoing procedures. Exclusion criteria were postoperative stoma, diagnosis of inflammatory bowel disease, or emergent operation. Variables collected from the Vanderbilt Procedural Outcomes Database included current procedural terminology code, date of operation, age, American Society of Anesthesiologists classification, case level (emergent, urgent, or elective), length of operation, perioperative antibiotic compliance, sex, race, and date of death (if applicable). The electronic medical record was used to determine intraoperative and postoperative details including indication for surgery, level of anastomosis, type of anastomotic leak test performed, body mass index, current smoking status, diagnosis of diabetes mellitus, history of chronic obstructive pulmonary disease, current steroid use, and preoperative chemotherapy or radiation therapy. Smoking status was collected as a dichotomous variable: current smoker or not current smoker. A current smoker was defined as anyone smoking within 30 days of the procedure. The colorectal surgeons at our institution have developed a standardized technique for left colon resections. All patients underwent preoperative mechanical bowel preparation. When feasible, the majority of resections were performed using a laparoscopic approach with a hand-assist port. Anastomoses were preferentially stapled. An intraoperative leak test was routinely performed for left-sided colorectal anastomoses. The type of leak test performed, rigid proctoscopy or flexible sigmoidoscopy, was collected, as this varied by surgeon preference. If a leak was detected intraoperatively, the method of management was also recorded (ie, reinforcing or repairing the anastomosis vs resecting and redoing the anastomosis). The level of anastomosis was classified as below the peritoneal
Results There were 246 patients who met inclusion criteria. The mean age was 58.5 years (613 standard deviation), and 56% (n 5 137) were female. Current smokers made up 12% (n 5 29) of the study population, and smokers were more likely to be male and were younger than nonsmokers. Detailed demographic characteristics of the overall population, smokers, and nonsmokers can be seen in Table 1. Notably, the nonsmoking group tended to be older, female, and with higher American Society of Anesthesiologists classification. Operative details are found in Table 2. In the overall population, 75% (n 5 184) had flexible sigmoidoscopy performed for a leak test. The rate of positive intraoperative leak test was 12.6% (n 5 31), with no difference between rigid or flexible endoscopy. Most commonly, these leaks were managed by oversewing the area and repeating the leak test (n 5 28). In 3 patients, the anastomosis was taken down and redone. The majority of patients, 53% (n 5 130), underwent an operation for diverticulitis, 37% (n 5 90) for malignancy, and 8% (n 5 20) for an endoscopically unresectable polyp. Other indications included endometriosis (n 5 2), volvulus (n 5 2), ischemic colitis (n 5 1), and anastomotic stricture (n 5 1). Smokers were more likely to have a diagnosis of diverticulitis. Of the
R.B. Baucom et al. Table 1
Demographics of left colectomy patients
3
Demographics of left colectomy patients by exposure group
Characteristic Sex, n (%) Male Female Age, years (SD) Body mass index, kg/m2 (SD) ASA classification, n (%) I II III IV Diabetes mellitus, n (%) Recent steroids, n (%)
Overall (n 5 246)
Smokers (n 5 29)
Nonsmokers (n 5 217)
109 137 58.5 28.8
(44) (56) (13) (7)
19 10 52.2 28.0
(66) (34) (11) (7)
90 127 59.4 28.9
(41) (59) (13) (7)
3 115 124 4 34 10
(1) (47) (50) (2) (14) (4)
0 20 8 1 4 1
(0) (69) (28) (3) (14) (3)
3 95 116 3 30 9
(1) (44) (53) (1) (14) (4)
P value .01 .003 .52 .03
1.00 .86
ASA 5 American Society of Anesthesiologists; SD 5 standard deviation.
patients with a diagnosis of malignancy, 3 underwent both chemotherapy and radiation therapy preoperatively (within 3 months of surgery), and 2 received preoperative chemotherapy alone.
stay was 5 days for both smokers and nonsmokers (P 5 .57), if a patient developed a clinical leak requiring intervention, the mean length of stay increased by 3 times to 15 days (P , .001).
Univariate analysis
Multivariable analysis
The clinical leak rate was 6.5% (n 5 16), and smokers were significantly more likely than nonsmokers to develop a leak (17% vs 5%, P 5 .01). The leaks were managed with reoperation in 63% (n 5 10) and percutaneous drainage in 37% (n 5 6), with no significant difference in management between smokers and nonsmokers (P 5 .21). There was no difference in leak rate based on diagnosis of malignancy versus diverticulitis (7.9% vs 5.4%, P 5 .58). Other potential risk factors for leak are shown in Table 3. The rate of organ space infection in smokers was 17% versus 8% in nonsmokers (P 5 .21). Overall mortality was 1.2% (n 5 3) with no statistically significant difference between smokers and nonsmokers. These outcomes are represented graphically in Fig. 1. Although the mean length of
The results of the multivariable model, including odds ratios, confidence intervals, and P values, are shown in Table 4. Smokers had over 4 times the odds of developing a clinical anastomotic leak than nonsmokers (P 5 .02). Having a low anterior anastomosis was associated with 3.5 times the odds of developing a clinical leak (P 5 .02) compared with a sigmoid anastomosis. In constructing this final model, sex was also considered as a covariate. However, addition of sex to the model resulted in noise and a poorly fit model. Additionally, sex was not significant in this model, and smoking remained strongly significant. Given our limited number of events, anastomotic level was felt to be the most appropriate covariate for inclusion, given its strong association with leak risk.
Table 2
Operative details
Characteristic Indication for surgery, n (%) Malignancy Diverticular disease Other Level of anastomosis, n (%) Low anterior Sigmoid Above sigmoid Type of leak test, n (%) Rigid proctoscopy Flexible sigmoidoscopy Positive leak test, n (%) Case length, minutes (SD) SD 5 standard deviation.
Overall (n 5 246)
Smokers (n 5 29)
Nonsmokers (n 5 217)
P value
90 (36) 130 (53) 26 (11)
5 (17) 22 (76) 2 (7)
85 (39) 108 (50) 24 (11)
.03
75 (31) 161 (65) 10 (4)
8 (28) 21 (72) 0 (0)
67 (31) 140 (64) 10 (5)
.43
50 167 29 236
.03
62 184 31 246
(25) (75) (13) (237)
12 17 2 247
(41) (59) (7) (81)
(23) (77) (13) (72)
.32 .34
4
The American Journal of Surgery, Vol 210, No 1, July 2015 Table 3 Rate of clinical leak by sex, disease process, and anastomotic level Characteristic Sex Male Female Diagnosis Diverticular disease Malignancy Anastomotic level Low anterior Sigmoid
Clinical leak rate, %
Table 4 Multivariable logistic regression model predicting leak after left colectomy
P value
8.3 (n 5 9) 5.2 (n 5 7)
.33
5.4 (n 5 7) 7.9 (n 5 7)
.46
12.2 (n 5 9) 4.4 (n 5 7)
.03
Comments Smoking significantly increases the rate of clinical leak requiring intervention after left colectomy and is a modifiable risk factor in the elective setting. The rate of organ space infection was also higher in smokers, although not statistically significant. The risk of leak increased with progressively lower rectal anastomoses. Sharma et al,9 in a study of over 47,000 patients, demonstrated that smoking increases the risk of major complications after all types of colorectal surgery, and smoking appeared to have a dose-dependent effect. The study did not specifically determine correlation with clinical leak, but did include sepsis and organ space infection as significant outcomes. Additionally, a smaller study by Kim et al,1 which included 215 patients who underwent low anterior resection for cancer, demonstrated that a history of heavy smoking was an independent risk factor for anastomotic stricture and leak. A recent systematic review found preoperative smoking to be associated with increased postoperative morbidity, although this study was not specific to colorectal surgery patients and did not include anastomotic leak as a primary outcome.10 One study published in 2003 did not find smoking to be a significant risk factor for anastomotic leak in left colectomy,
Figure 1 Outcomes after left colectomy. The rate of clinical leak was 6.5% (n 5 16), with smokers having significantly more leaks than nonsmokers (17% vs 5%, P 5 .01). The rate of organ space infection was 9% overall, with a rate of 17% in smokers and 8% in nonsmokers (P 5 .12). The 30-day mortality was 1.2% (n 5 3), and was not significantly different between groups.
Variable
Odds ratio
95% Confidence interval
P value
Smoker Lower anastomosis
4.2 3.5
1.3–13.5 1.2–9.6
.02 .02
although the study included only 88 patients and evaluated multiple risk factors for leak.11 Our findings correlate with previous studies indicating an increased risk of leak for lower anastomoses.2–5,12,13 Our findings also correlate with studies demonstrating a significantly increased length of hospital stay after clinical leak.3,7 In our study, length of stay was 3 times longer for patients who developed clinical leak compared with those who did not (P , .001). In addition to the increased cost for each day of hospitalization, these patients had to undergo at least one additional procedure, further increasing costs. A recent study demonstrated that quality of life for patients who experience leak is also affected, as many have a drain or a prolonged ostomy as a result of the complication. The study demonstrated that physical function was negatively affected, as well as emotional and social well-being.14 In light of this, smoking cessation before elective colectomy should be seriously considered. Because our study included only patients without a postoperative stoma, the impact of fecal diversion in these patients is not known, although our data suggest that it should be considered for this subset of patients. Several studies have demonstrated that smoking has an impact on all phases of wound healing through various toxic effects and oxidative stress. A review published by Sørensen15 in 2012 concluded that smoking cessation reverses some but not all of the negative effects on wound healing in surgical patients. Some effects, such as decreased tissue oxygenation, are reversed within hours of smoking cessation, whereas others have a prolonged detrimental effect on fibroblast proliferation and collagen synthesis.16,17 A reduced risk of infectious complications after surgery has been observed clinically and in randomized trials after 4 weeks of smoking cessation.15,18,19 Although the effect of nicotine on wound healing is not fully understood, there is currently no evidence to suggest that nicotine replacement drugs negatively impact wound healing.18,19 No studies have specifically evaluated the effect of smoking on wound healing in the colorectal surgery population, but based on the existing literature, 4 weeks of smoking cessation, with or without a nicotine patch, is likely to be beneficial to reduce wound healing complications. This is certainly an area for further study. Although this is one of the only studies looking specifically at the effect of smoking on clinical leak, it is limited by its retrospective nature and small number of events. The ability to adjust for confounders was limited. When sex, which is a known risk factor for leak, was
R.B. Baucom et al.
Demographics of left colectomy patients
included in the model, smoking remained significant even when sex did not. This supports our finding that smoking is truly associated with an increased risk of clinical leak. The strongest known risk factor was chosen for inclusion in the multivariable model, but larger studies are needed to account for all possible confounders. In our population, the overall rate of positive intraoperative leak tests fell within the 8% to 25% range found in the literature.20,21 However, the data from univariate analysis found that a significantly higher proportion of smokers had rigid proctoscopy performed for intraoperative leak testing, compared with nonsmokers where flexible endoscopy was more frequently used. The impact that the type of leak test performed had on our results is unknown based on the current literature available. Finally, the definition of smoker used in this study was oversimplified and based only on smoking status within 30 days of operation. This does not capture the impact of a previous history of smoking, nor the quantitative effect of total pack-years. Because the definition used for this study included former smokers in the ‘‘nonsmokers’’ group, it was felt that this would bias the study toward finding no association between smoking and leak. However, the fact that our study did find a significant difference suggests that it may even underestimate the effect of smoking on clinical leak.
Conclusions Smoking increases the odds of developing a clinical anastomotic leak requiring intervention by more than 4 times, and consideration should be given to delaying elective left colectomy until smoking cessation is achieved. The existing literature suggests that 4 weeks of smoking cessation, with or without a nicotine patch, decreases the risks of infectious complications postoperatively.15 This finding may have an especially meaningful impact on patients with colorectal cancer and a history of smoking. More detailed studies are needed to determine the optimal length of smoking cessation that decreases the risk of anastomotic leak in patients undergoing left colectomy. At the very least, patients and surgeons should be aware of the impact that smoking has on anastomotic leak following colectomy.
References 1. Kim MJ, Shin R, Oh HK, et al. The impact of heavy smoking on anastomotic leakage and stricture after low anterior resection in rectal cancer patients. World J Surg 2011;35:2806–10. 2. Trencheva K, Morrissey KP, Wells M, et al. Identifying important predictors for anastomotic leak after colon and rectal resection: prospective study on 616 patients. Ann Surg 2013;257:108–13.
5 3. Akiyoshi T, Ueno M, Fukunaga Y, et al. Incidence of and risk factors for anastomotic leakage after laparoscopic anterior resection with intracorporeal rectal transection and double-stapling technique anastomosis for rectal cancer. Am J Surg 2011;202:259–64. 4. Ko¨ckerling F, Rose J, Schneider C, et al. Laparoscopic colorectal anastomosis: risk of postoperative leakage. Results of a multicenter study. Laparoscopic Colorectal Surgery Study Group (LCSSG). Surg Endosc 1999;13:639–44. 5. Karanjia ND, Corder AP, Bearn P, et al. Leakage from stapled low anastomosis after total mesorectal excision for carcinoma of the rectum. Br J Surg 1994;81:1224–6. 6. Buchs NC, Gervaz P, Secic M, et al. Incidence, consequences, and risk factors for anastomotic dehiscence after colorectal surgery: a prospective monocentric study. Int J Colorectal Dis 2008;23:265–70. 7. Alves A, Panis Y, Trancart D, et al. Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World J Surg 2002;26:499–502. 8. Alberg AJ, Shopland DR, Cummings KM. The 2014 Surgeon General’s report: commemorating the 50th Anniversary of the 1964 Report of the Advisory Committee to the US Surgeon General and updating the evidence on the health consequences of cigarette smoking. Am J Epidemiol 2014;179:403–12. 9. Sharma A, Deeb AP, Iannuzzi JC, et al. Tobacco smoking and postoperative outcomes after colorectal surgery. Ann Surg 2013;258: 296–300. 10. Grønkjær M, Eliasen M, Skov-Ettrup LS, et al. Preoperative smoking status and postoperative complications: a systematic review and metaanalysis. Ann Surg 2014;259:52–71. 11. Ma¨kela¨ JT, Kiviniemi H, Laitinen S. Risk factors for anastomotic leakage after left-sided colorectal resection with rectal anastomosis. Dis Colon Rectum 2003;46:653–60. 12. Vignali A, Fazio VW, Lavery IC, et al. Factors associated with the occurrence of leaks in stapled rectal anastomoses: a review of 1,014 patients. J Am Coll Surg 1997;185:105–13. 13. Rose J, Schneider C, Yildirim C, et al. Complications in laparoscopic colorectal surgery: results of a multicentre trial. Tech Coloproctol 2004;8(Suppl 1)):s25–8. 14. Marinatou A, Theodoropoulos GE, Karanika S, et al. Do anastomotic leaks impair postoperative health-related quality of life after rectal cancer surgery? A case-matched study. Dis Colon Rectum 2014;57: 158–66. 15. Sørensen LT. Wound healing and infection in surgery: the pathophysiological impact of smoking, smoking cessation, and nicotine replacement therapy: a systematic review. Ann Surg 2012;255:1069–79. 16. Sørensen LT, Toft BG, Rygaard J, et al. Effect of smoking, smoking cessation, and nicotine patch on wound dimension, vitamin C, and systemic markers of collagen metabolism. Surgery 2010;148:982–90. 17. Sørensen LT, Toft B, Rygaard J, et al. Smoking attenuates wound inflammation and proliferation while smoking cessation restores inflammation but not proliferation. Wound Repair Regen 2010;18: 186–92. 18. Møller AM, Villebro N, Pedersen T, et al. Effect of preoperative smoking intervention on postoperative complications: a randomised clinical trial. Lancet 2002;359:114–7. 19. Sørensen LT, Karlsmark T, Gottrup F. Abstinence from smoking reduces incisional wound infection: a randomized controlled trial. Ann Surg 2003;238:1–5. 20. Beard JD, Nicholson ML, Sayers RD, et al. Intraoperative air testing of colorectal anastomoses: a prospective, randomized trial. Br J Surg 1990;77:1095–7. 21. Ricciardi R, Roberts PL, Marcello PW, et al. Anastomotic leak testing after colorectal resection: what are the data? Arch Surg 2009;144: 407–11; discussion 411–2.
Clinical Science
Smoking as dominant risk factor for anastomotic leak after left colon resection Rebeccah B. Baucom, M.D.*, Benjamin K. Poulose, M.D., M.P.H., Alan J. Herline, M.D., Roberta L. Muldoon, M.D., Molly M. Cone, M.D., Timothy M. Geiger, M.D. Division of General Surgery, Vanderbilt University Medical Center, D-5203 Medical Center North, 1161 Medical Center Drive, Nashville, TN 37232, USA KEYWORDS: Smoking; Anastomotic leak; Colectomy; Left colectomy; Risk factors
Abstract BACKGROUND: Some risk factors for anastomotic leak have been identified, but the effect of smoking is unknown. METHODS: This study aimed to evaluate the effect of smoking on clinical leak after left-sided anastomoses. Adult patients who underwent elective left colectomy between January 1, 2008 and December 31, 2012 were included. Those with stomas and inflammatory bowel diseases were excluded. Primary outcome was anastomotic leak requiring percutaneous drainage or operative intervention within 30 days. RESULTS: There were 246 patients included; 56% were female. Most had a diagnosis of diverticular disease (53%) or cancer (37%). Anastomotic leak rate was 6.5% (n 5 16). The rate in smokers was 17% versus 5% in nonsmokers (P 5 .01). Smokers had over 4 times greater chance of leak (odds ratio 4.2, 95% confidence interval 1.3 to 13.5, P 5 .02). CONCLUSION: Smoking is a risk factor for leak after left colectomy. Consideration should be given to delaying elective left colectomy until smoking cessation is achieved. Ó 2015 Elsevier Inc. All rights reserved.
Clinically significant leak occurs in 1% to 12% of colorectal operations and up to 14% after distal colorectal resections.1–5 Anastomotic leak significantly increases morbidity and mortality, with mortality reported between The authors declare no conflicts of interest. Data were collected and managed using REDCap electronic data capture tool, which is supported by the UL1 TR000445 grant from NCATS/ NIH and hosted at Vanderbilt University Medical Center. Presented as a podium presentation at the American Society of Colon and Rectal Surgeons, May 17–21, 2014, Hollywood, Florida. * Corresponding author. Tel.: 11-615-343-5613; fax: 11-615-3439485. E-mail address: [email protected] Manuscript received September 11, 2014; revised manuscript September 30, 2014 0002-9610/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2014.10.033
6% and 13%.2,5,6 Length of hospitalization can increase up to 4 times for those patients with a leak compared with those without a leak.3,7 Although anastomotic leak remains a challenging complication following colon and rectal surgery, solid evidence evaluating specific risk factors for anastomotic leak is lacking. A recent prospective study evaluating risk factors for leak after colorectal surgery found that low anastomosis, increasing number of comorbidities, high ligation of the inferior mesenteric artery, intraoperative complications, and male sex are significant risk factors.2 However, smoking was not specifically evaluated in this study. In 2013, the Surgeon General released an updated report on the health consequences of smoking.8 Important to the colorectal patient population, the report linked smoking with colorectal cancer and inflammatory bowel disease.
2
The American Journal of Surgery, Vol 210, No 1, July 2015
The report demonstrates tobacco’s continued burden to the nation’s health with its clear impact on all-cause mortality in men and women. The annual cost for direct medical care of adults attributable to smoking between the years 2009 and 2012 was $132.5 to 175.9 billion. Smoking is known to cause microvascular disease, tissue ischemia, and hypoxia, which are contributing factors to poor anastomotic healing. The purpose of this study was to evaluate smoking as an independent, modifiable risk factor for clinical leak after left-sided colorectal anastomoses.
reflection (low anterior), at the level of the sigmoid, or above the level of the sigmoid. The primary outcome measure was postoperative clinical anastomotic leak within 30 days of surgery. A clinical leak was defined as air or fluid near the site of the anastomosis, requiring either return to the operating room or percutaneous drainage. Secondary outcome measures included organ space infection and 30-day mortality. The study was performed with approval of the Vanderbilt University Institutional Review Board and Human Research Protection Program.
Patients and Methods
Statistical considerations
Study design and data sources
Descriptive statistics for the overall study population were calculated, as well as for the group of smokers compared with nonsmokers. Chi-square and Fisher’s exact tests were used to compare proportions, as appropriate. Multivariable logistic regression was performed, with covariates determined a priori with a 2-tailed alpha level of .05 determining significance. The number of events limited the inclusion of all possible factors, so those that were known, strong risk factors, were given priority. The final model included smoking and anastomotic level. Additional covariates which were given consideration in the modeling process included sex, body mass index, recent chemotherapy or radiation therapy, diagnosis, and type of leak test performed. Stata version 13.1 (StataCorp LC, College Station, TX) was used for statistical analysis.
A retrospective cohort study was performed including adult patients who underwent left colectomy between January 1, 2008 and December 31, 2012, by one of the 4 board-certified colorectal surgeons at a single, tertiary care center. These patients were identified from the Vanderbilt Procedural Outcomes Databasedan institutional database that includes clinical and administrative information on patients undergoing procedures. Exclusion criteria were postoperative stoma, diagnosis of inflammatory bowel disease, or emergent operation. Variables collected from the Vanderbilt Procedural Outcomes Database included current procedural terminology code, date of operation, age, American Society of Anesthesiologists classification, case level (emergent, urgent, or elective), length of operation, perioperative antibiotic compliance, sex, race, and date of death (if applicable). The electronic medical record was used to determine intraoperative and postoperative details including indication for surgery, level of anastomosis, type of anastomotic leak test performed, body mass index, current smoking status, diagnosis of diabetes mellitus, history of chronic obstructive pulmonary disease, current steroid use, and preoperative chemotherapy or radiation therapy. Smoking status was collected as a dichotomous variable: current smoker or not current smoker. A current smoker was defined as anyone smoking within 30 days of the procedure. The colorectal surgeons at our institution have developed a standardized technique for left colon resections. All patients underwent preoperative mechanical bowel preparation. When feasible, the majority of resections were performed using a laparoscopic approach with a hand-assist port. Anastomoses were preferentially stapled. An intraoperative leak test was routinely performed for left-sided colorectal anastomoses. The type of leak test performed, rigid proctoscopy or flexible sigmoidoscopy, was collected, as this varied by surgeon preference. If a leak was detected intraoperatively, the method of management was also recorded (ie, reinforcing or repairing the anastomosis vs resecting and redoing the anastomosis). The level of anastomosis was classified as below the peritoneal
Results There were 246 patients who met inclusion criteria. The mean age was 58.5 years (613 standard deviation), and 56% (n 5 137) were female. Current smokers made up 12% (n 5 29) of the study population, and smokers were more likely to be male and were younger than nonsmokers. Detailed demographic characteristics of the overall population, smokers, and nonsmokers can be seen in Table 1. Notably, the nonsmoking group tended to be older, female, and with higher American Society of Anesthesiologists classification. Operative details are found in Table 2. In the overall population, 75% (n 5 184) had flexible sigmoidoscopy performed for a leak test. The rate of positive intraoperative leak test was 12.6% (n 5 31), with no difference between rigid or flexible endoscopy. Most commonly, these leaks were managed by oversewing the area and repeating the leak test (n 5 28). In 3 patients, the anastomosis was taken down and redone. The majority of patients, 53% (n 5 130), underwent an operation for diverticulitis, 37% (n 5 90) for malignancy, and 8% (n 5 20) for an endoscopically unresectable polyp. Other indications included endometriosis (n 5 2), volvulus (n 5 2), ischemic colitis (n 5 1), and anastomotic stricture (n 5 1). Smokers were more likely to have a diagnosis of diverticulitis. Of the
R.B. Baucom et al. Table 1
Demographics of left colectomy patients
3
Demographics of left colectomy patients by exposure group
Characteristic Sex, n (%) Male Female Age, years (SD) Body mass index, kg/m2 (SD) ASA classification, n (%) I II III IV Diabetes mellitus, n (%) Recent steroids, n (%)
Overall (n 5 246)
Smokers (n 5 29)
Nonsmokers (n 5 217)
109 137 58.5 28.8
(44) (56) (13) (7)
19 10 52.2 28.0
(66) (34) (11) (7)
90 127 59.4 28.9
(41) (59) (13) (7)
3 115 124 4 34 10
(1) (47) (50) (2) (14) (4)
0 20 8 1 4 1
(0) (69) (28) (3) (14) (3)
3 95 116 3 30 9
(1) (44) (53) (1) (14) (4)
P value .01 .003 .52 .03
1.00 .86
ASA 5 American Society of Anesthesiologists; SD 5 standard deviation.
patients with a diagnosis of malignancy, 3 underwent both chemotherapy and radiation therapy preoperatively (within 3 months of surgery), and 2 received preoperative chemotherapy alone.
stay was 5 days for both smokers and nonsmokers (P 5 .57), if a patient developed a clinical leak requiring intervention, the mean length of stay increased by 3 times to 15 days (P , .001).
Univariate analysis
Multivariable analysis
The clinical leak rate was 6.5% (n 5 16), and smokers were significantly more likely than nonsmokers to develop a leak (17% vs 5%, P 5 .01). The leaks were managed with reoperation in 63% (n 5 10) and percutaneous drainage in 37% (n 5 6), with no significant difference in management between smokers and nonsmokers (P 5 .21). There was no difference in leak rate based on diagnosis of malignancy versus diverticulitis (7.9% vs 5.4%, P 5 .58). Other potential risk factors for leak are shown in Table 3. The rate of organ space infection in smokers was 17% versus 8% in nonsmokers (P 5 .21). Overall mortality was 1.2% (n 5 3) with no statistically significant difference between smokers and nonsmokers. These outcomes are represented graphically in Fig. 1. Although the mean length of
The results of the multivariable model, including odds ratios, confidence intervals, and P values, are shown in Table 4. Smokers had over 4 times the odds of developing a clinical anastomotic leak than nonsmokers (P 5 .02). Having a low anterior anastomosis was associated with 3.5 times the odds of developing a clinical leak (P 5 .02) compared with a sigmoid anastomosis. In constructing this final model, sex was also considered as a covariate. However, addition of sex to the model resulted in noise and a poorly fit model. Additionally, sex was not significant in this model, and smoking remained strongly significant. Given our limited number of events, anastomotic level was felt to be the most appropriate covariate for inclusion, given its strong association with leak risk.
Table 2
Operative details
Characteristic Indication for surgery, n (%) Malignancy Diverticular disease Other Level of anastomosis, n (%) Low anterior Sigmoid Above sigmoid Type of leak test, n (%) Rigid proctoscopy Flexible sigmoidoscopy Positive leak test, n (%) Case length, minutes (SD) SD 5 standard deviation.
Overall (n 5 246)
Smokers (n 5 29)
Nonsmokers (n 5 217)
P value
90 (36) 130 (53) 26 (11)
5 (17) 22 (76) 2 (7)
85 (39) 108 (50) 24 (11)
.03
75 (31) 161 (65) 10 (4)
8 (28) 21 (72) 0 (0)
67 (31) 140 (64) 10 (5)
.43
50 167 29 236
.03
62 184 31 246
(25) (75) (13) (237)
12 17 2 247
(41) (59) (7) (81)
(23) (77) (13) (72)
.32 .34
4
The American Journal of Surgery, Vol 210, No 1, July 2015 Table 3 Rate of clinical leak by sex, disease process, and anastomotic level Characteristic Sex Male Female Diagnosis Diverticular disease Malignancy Anastomotic level Low anterior Sigmoid
Clinical leak rate, %
Table 4 Multivariable logistic regression model predicting leak after left colectomy
P value
8.3 (n 5 9) 5.2 (n 5 7)
.33
5.4 (n 5 7) 7.9 (n 5 7)
.46
12.2 (n 5 9) 4.4 (n 5 7)
.03
Comments Smoking significantly increases the rate of clinical leak requiring intervention after left colectomy and is a modifiable risk factor in the elective setting. The rate of organ space infection was also higher in smokers, although not statistically significant. The risk of leak increased with progressively lower rectal anastomoses. Sharma et al,9 in a study of over 47,000 patients, demonstrated that smoking increases the risk of major complications after all types of colorectal surgery, and smoking appeared to have a dose-dependent effect. The study did not specifically determine correlation with clinical leak, but did include sepsis and organ space infection as significant outcomes. Additionally, a smaller study by Kim et al,1 which included 215 patients who underwent low anterior resection for cancer, demonstrated that a history of heavy smoking was an independent risk factor for anastomotic stricture and leak. A recent systematic review found preoperative smoking to be associated with increased postoperative morbidity, although this study was not specific to colorectal surgery patients and did not include anastomotic leak as a primary outcome.10 One study published in 2003 did not find smoking to be a significant risk factor for anastomotic leak in left colectomy,
Figure 1 Outcomes after left colectomy. The rate of clinical leak was 6.5% (n 5 16), with smokers having significantly more leaks than nonsmokers (17% vs 5%, P 5 .01). The rate of organ space infection was 9% overall, with a rate of 17% in smokers and 8% in nonsmokers (P 5 .12). The 30-day mortality was 1.2% (n 5 3), and was not significantly different between groups.
Variable
Odds ratio
95% Confidence interval
P value
Smoker Lower anastomosis
4.2 3.5
1.3–13.5 1.2–9.6
.02 .02
although the study included only 88 patients and evaluated multiple risk factors for leak.11 Our findings correlate with previous studies indicating an increased risk of leak for lower anastomoses.2–5,12,13 Our findings also correlate with studies demonstrating a significantly increased length of hospital stay after clinical leak.3,7 In our study, length of stay was 3 times longer for patients who developed clinical leak compared with those who did not (P , .001). In addition to the increased cost for each day of hospitalization, these patients had to undergo at least one additional procedure, further increasing costs. A recent study demonstrated that quality of life for patients who experience leak is also affected, as many have a drain or a prolonged ostomy as a result of the complication. The study demonstrated that physical function was negatively affected, as well as emotional and social well-being.14 In light of this, smoking cessation before elective colectomy should be seriously considered. Because our study included only patients without a postoperative stoma, the impact of fecal diversion in these patients is not known, although our data suggest that it should be considered for this subset of patients. Several studies have demonstrated that smoking has an impact on all phases of wound healing through various toxic effects and oxidative stress. A review published by Sørensen15 in 2012 concluded that smoking cessation reverses some but not all of the negative effects on wound healing in surgical patients. Some effects, such as decreased tissue oxygenation, are reversed within hours of smoking cessation, whereas others have a prolonged detrimental effect on fibroblast proliferation and collagen synthesis.16,17 A reduced risk of infectious complications after surgery has been observed clinically and in randomized trials after 4 weeks of smoking cessation.15,18,19 Although the effect of nicotine on wound healing is not fully understood, there is currently no evidence to suggest that nicotine replacement drugs negatively impact wound healing.18,19 No studies have specifically evaluated the effect of smoking on wound healing in the colorectal surgery population, but based on the existing literature, 4 weeks of smoking cessation, with or without a nicotine patch, is likely to be beneficial to reduce wound healing complications. This is certainly an area for further study. Although this is one of the only studies looking specifically at the effect of smoking on clinical leak, it is limited by its retrospective nature and small number of events. The ability to adjust for confounders was limited. When sex, which is a known risk factor for leak, was
R.B. Baucom et al.
Demographics of left colectomy patients
included in the model, smoking remained significant even when sex did not. This supports our finding that smoking is truly associated with an increased risk of clinical leak. The strongest known risk factor was chosen for inclusion in the multivariable model, but larger studies are needed to account for all possible confounders. In our population, the overall rate of positive intraoperative leak tests fell within the 8% to 25% range found in the literature.20,21 However, the data from univariate analysis found that a significantly higher proportion of smokers had rigid proctoscopy performed for intraoperative leak testing, compared with nonsmokers where flexible endoscopy was more frequently used. The impact that the type of leak test performed had on our results is unknown based on the current literature available. Finally, the definition of smoker used in this study was oversimplified and based only on smoking status within 30 days of operation. This does not capture the impact of a previous history of smoking, nor the quantitative effect of total pack-years. Because the definition used for this study included former smokers in the ‘‘nonsmokers’’ group, it was felt that this would bias the study toward finding no association between smoking and leak. However, the fact that our study did find a significant difference suggests that it may even underestimate the effect of smoking on clinical leak.
Conclusions Smoking increases the odds of developing a clinical anastomotic leak requiring intervention by more than 4 times, and consideration should be given to delaying elective left colectomy until smoking cessation is achieved. The existing literature suggests that 4 weeks of smoking cessation, with or without a nicotine patch, decreases the risks of infectious complications postoperatively.15 This finding may have an especially meaningful impact on patients with colorectal cancer and a history of smoking. More detailed studies are needed to determine the optimal length of smoking cessation that decreases the risk of anastomotic leak in patients undergoing left colectomy. At the very least, patients and surgeons should be aware of the impact that smoking has on anastomotic leak following colectomy.
References 1. Kim MJ, Shin R, Oh HK, et al. The impact of heavy smoking on anastomotic leakage and stricture after low anterior resection in rectal cancer patients. World J Surg 2011;35:2806–10. 2. Trencheva K, Morrissey KP, Wells M, et al. Identifying important predictors for anastomotic leak after colon and rectal resection: prospective study on 616 patients. Ann Surg 2013;257:108–13.
5 3. Akiyoshi T, Ueno M, Fukunaga Y, et al. Incidence of and risk factors for anastomotic leakage after laparoscopic anterior resection with intracorporeal rectal transection and double-stapling technique anastomosis for rectal cancer. Am J Surg 2011;202:259–64. 4. Ko¨ckerling F, Rose J, Schneider C, et al. Laparoscopic colorectal anastomosis: risk of postoperative leakage. Results of a multicenter study. Laparoscopic Colorectal Surgery Study Group (LCSSG). Surg Endosc 1999;13:639–44. 5. Karanjia ND, Corder AP, Bearn P, et al. Leakage from stapled low anastomosis after total mesorectal excision for carcinoma of the rectum. Br J Surg 1994;81:1224–6. 6. Buchs NC, Gervaz P, Secic M, et al. Incidence, consequences, and risk factors for anastomotic dehiscence after colorectal surgery: a prospective monocentric study. Int J Colorectal Dis 2008;23:265–70. 7. Alves A, Panis Y, Trancart D, et al. Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World J Surg 2002;26:499–502. 8. Alberg AJ, Shopland DR, Cummings KM. The 2014 Surgeon General’s report: commemorating the 50th Anniversary of the 1964 Report of the Advisory Committee to the US Surgeon General and updating the evidence on the health consequences of cigarette smoking. Am J Epidemiol 2014;179:403–12. 9. Sharma A, Deeb AP, Iannuzzi JC, et al. Tobacco smoking and postoperative outcomes after colorectal surgery. Ann Surg 2013;258: 296–300. 10. Grønkjær M, Eliasen M, Skov-Ettrup LS, et al. Preoperative smoking status and postoperative complications: a systematic review and metaanalysis. Ann Surg 2014;259:52–71. 11. Ma¨kela¨ JT, Kiviniemi H, Laitinen S. Risk factors for anastomotic leakage after left-sided colorectal resection with rectal anastomosis. Dis Colon Rectum 2003;46:653–60. 12. Vignali A, Fazio VW, Lavery IC, et al. Factors associated with the occurrence of leaks in stapled rectal anastomoses: a review of 1,014 patients. J Am Coll Surg 1997;185:105–13. 13. Rose J, Schneider C, Yildirim C, et al. Complications in laparoscopic colorectal surgery: results of a multicentre trial. Tech Coloproctol 2004;8(Suppl 1)):s25–8. 14. Marinatou A, Theodoropoulos GE, Karanika S, et al. Do anastomotic leaks impair postoperative health-related quality of life after rectal cancer surgery? A case-matched study. Dis Colon Rectum 2014;57: 158–66. 15. Sørensen LT. Wound healing and infection in surgery: the pathophysiological impact of smoking, smoking cessation, and nicotine replacement therapy: a systematic review. Ann Surg 2012;255:1069–79. 16. Sørensen LT, Toft BG, Rygaard J, et al. Effect of smoking, smoking cessation, and nicotine patch on wound dimension, vitamin C, and systemic markers of collagen metabolism. Surgery 2010;148:982–90. 17. Sørensen LT, Toft B, Rygaard J, et al. Smoking attenuates wound inflammation and proliferation while smoking cessation restores inflammation but not proliferation. Wound Repair Regen 2010;18: 186–92. 18. Møller AM, Villebro N, Pedersen T, et al. Effect of preoperative smoking intervention on postoperative complications: a randomised clinical trial. Lancet 2002;359:114–7. 19. Sørensen LT, Karlsmark T, Gottrup F. Abstinence from smoking reduces incisional wound infection: a randomized controlled trial. Ann Surg 2003;238:1–5. 20. Beard JD, Nicholson ML, Sayers RD, et al. Intraoperative air testing of colorectal anastomoses: a prospective, randomized trial. Br J Surg 1990;77:1095–7. 21. Ricciardi R, Roberts PL, Marcello PW, et al. Anastomotic leak testing after colorectal resection: what are the data? Arch Surg 2009;144: 407–11; discussion 411–2.
