Abstract

Objective

Despite being routinely used before elective colorectal surgery in most surgical clinics, mechanical bowel preparation (MBP) remains controversial. This study aimed to investigate postoperative complications and outcomes of right, left, or rectosigmoid resection without MBP.

Material and Methods

Patients who underwent elective colorectal surgery without mechanical bowel preparation and oral antibiotics between January 2011 and December 2021 were included in the study. Patients were categorized according to the side of resection, and these subgroups were compared for anastomotic leakage, surgical site infections (SSI), and overall morbidity measured using the Clavien-Dindo complication grade.

Results

Data of 422 patients were analyzed. Overall anastomotic leakage was found in 14 patients (3.3%), SSI in 46 (10.9%), collection in 14 (3.3%), mortality in 18 (4.3%), and reoperation in 17 (%4) patients. Anastomotic leakage was observed in six (3.9%) in right colectomy, two (1.9%) in left colectomy, and in six (3.7%) patients in the rectosigmoid resection group when the groups were evaluated separately. There was no statistical difference between the groups (p= 0.630). Furthermore, there was no statistical difference between the groups regarding collection and reoperation (p values were p= 0.31, and p= 0.251, respectively).

Conclusion

Study results showed that anastomotic leakage, surgical site infection, intra-abdominal collection, reoperation, and mortality rates were similar to the current literature obtained from the studies with mechanical bowel preparation. In addition, these results were found to be similar according to the resection site.

Introduction

Colorectal surgeons use various protocols for bowel preparation to prevent complications such as anastomotic leakage, intraabdominal abscess, and surgical site infections. These include oral antibiotics, intravenous antibiotics, rectal enemas, oral solutions, and combinations.

Despite having been used for nearly a century to reduce postoperative infectious complications and minimize the contamination of the operation area by reducing the colonic bacterial load (1, 2), the usage of mechanical bowel preparation (MBP) is still questionable and the debate of the usage has been not to be finalized yet. Based on evidence-based studies, three different aspects are formed in clinical practice. Studies conducted in recent years have shown that complications such as anastomotic leakage, surgical site infection (SSI), and intraabdominal abscess are less common in patients with mechanical bowel preparation (3-6). On the other hand, some studies have shown that MBP does not affect postoperative infectious complications and anastomotic leakage rates (7-9). Other studies have paradoxically cited increased rates of infectious complications after MBP and also slower return of bowel function and increased rates of cardiac complications, electrolyte disturbances, and anastomotic leak (10-12).

The first questioning of the necessity and effectiveness of MBP was shown in a study by Hughes in 1972 (13). After this study, many studies have emerged exploring the potential benefits of MBP. However, various studies have shown that the reduction or prevention of SSI, intraabdominal abscess, and anastomotic leakage cannot be prevented after MBP (14, 15). Furthermore, MBP is not recommended before colorectal surgery as it causes various side effects such as bloating, nausea, fatigue, electrolyte imbalance, abdominal discomfort, and perforation, especially in elderly patients (16, 17).

The World Health Organization (WHO) and American Society of Colon and Rectal Surgeons (ASCRS) guidelines recommend oral antibiotics (OA) together with MBP (18, 19). On the other hand, the Enhanced Recovery After Surgery (ERAS) guidelines in elective colorectal surgery assigned the low quality of evidence for MBP with OAs (20). Therefore, the ERAS guidelines still recommend that MBP should not be routinely used in colon surgery. Therefore, this study investigated the postoperative morbidity of resections in different colon regions in patients who had undergone elective colorectal surgery without MBP retrospectively.

The aim of this study was to investigate complications in patients who underwent colonic resection and anastomosis without performing MBP and comparing the outcomes of right, left or rectosigmoid resections with each other. The primary aim was to evaluate anastomotic leakage and surgical site infection in addition to the rate of SSI within 30 days after surgery and subcategories of SSI (superficial incisional, deep incisional, and organ/ space. Secondary aim was to evaluate overall morbidity measured by using Clavien-Dindo complication grade.

Material and Methods

Source of Data and Study Population

We conducted a retrospective single-center cohort study of patients from tertiary centers experienced in colorectal surgery. These patients underwent elective colorectal resections for benign and malignant diseases for ten years from January 2011 to December 2021. The ethics committee approved the study protocol of the university hospital (E-78017789-050.01.04- 1647269/2021/347). A total of 767 consecutive patients were enrolled in this study. Exclusion criteria were accepted as follows: 1) patients who underwent emergent surgery (n= 87), 2) age under 18 years (n= 2), 3) patients with bowel obstruction (n= 14), 4) patients who underwent abdominoperineal resection with end stoma (n= 24), patients who performed laparoscopic surgery (n= 184) and 6) patients with no enough data available in the medical records (n= 34). Finally, four hundred and twenty-two patients who fulfilled the eligibility criteria were included in the study.

Patients were divided into three groups according to lesion localization and resection site. Of these, regions from the ileocolic region to the 2/3 proximal of the transverse colon were included in the right colectomy group; resections from the 1/3 distal part of the transverse colon to the distal sigmoid colon were included in the left colectomy group, and resections from the distal sigmoid colon to the distal rectum were included in the rectosigmoid resection group.

Prophylactic intravenous antibiotic prophylaxis was routinely administered with 1500 mg of cefuroxime and 500 mg of metronidazole 30 minutes before the incision and was terminated on day one postoperatively. It was also repeated when the operation time exceeded four hours, and blood loss exceeded 1.5 liters. In addition, ciprofloxacin 500 mg was administered to patients with penicillin and cephalosporin allergy.

Demographic data (age, sex), ASA scores, transfusion needing, receiving neoadjuvant treatment, comorbidity status (Charlson Comorbity Index), operation indication (malignant causes, benign causes), type of operation, protective ileotomy status and stage of the disease were recorded. In terms of postoperative results, anastomotic leakage, intraabdominal collection, mortality, reoperation and extraintestinal infection were recorded.

The primary and secondary aims of the study are stated in the manuscript. The primary aim was to evaluate the anastomotic leakage and surgical site infection. In addition, the rate of SSI within 30 days after surgery and subcategories of SSI (superficial incisional, deep incisional, and organ/space), as defined by the Centers for Disease Control and Prevention (21). Secondary outcomes included overall morbidity measured using the Clavien-Dindo complication grade (22).

Statistical Methods

Descriptive findings were presented as numbers and percentages for categorical variables and as mean and standart deviation for continuous variables. The Kolmogorov-Smirnov test evaluated the conformity of continuous variables to normal distribution. In the comparisons of groups of three or more, those with normal distribution were analyzed with the ANOVA test, and those which did not show normal distribution were analyzed with the Kruskal Wallis test. Tukey equal variances for those who show equal variances when comparing binary groups and Tamhane’s T2 post-hoc test was applied in those who did not. Pearson’s chi-square test was used to compare categorical variables in independent groups. The exact test was applied in cases that did not meet the Pearson’s chi-square test conditions. Multivariable logistic regression analysis was performed to evaluate the relationship between colon regions and complications with further analysis. The results were evaluated with a 95% confidence interval, with an alpha error of 0.05. Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS) for Windows 25.0 (IBM SPSS Inc., Chicago, IL).

Results

A total of 422 consecutive patients with a mean age of 59.9 ± 13.7 (range, 18-92) years were included in our study. Male/Female ratio was 239 (56.6%)/183 (43.4%). A total of 252 (59.7%) patients were under 65 years, whereas 170 (40.3%) of the patients were 65 years or over. ASA score was I (+43.4%) or II (56.6%) in most of the patients. The Charlson comorbidity index score of the patients was similar in all groups, and the average was 3. Neoadjuvant treatment was applied in 39 (9.2%) the patients. The need for peri/post-operative blood transfusions was seen in 193 (45.7%) patients, while 229 (54.3%) patients did not need any transfusion. Three hundred and sixty-five (86.5%) patients were operated for malignant reasons, and the remaining 57 (13.5%) of the patients were operated for benign reasons.

In the form of reconstruction after resection, colorectal anastomosis was performed in 234 (55.5%) patients, ileocolic anastomosis in 147 (34.8%) patients, and colocolic anastomosis in 41 (9.7%) patients. Protective ileostomy was applied to 108 (25.6%) patients. Clinicopathological and demografic features of the patients were summarized in Table 1.

When the three groups were evaluated in terms of mean age, it was seen that the mean age of the right colon patients was 62.99 ± 13.03, the left colon patients were 56.18 ± 15.44, and the rectum patients were 59.66 ± 12.50 years, and a statistically significant difference was determined (p< 0.001). In group comparisons, it was determined that the group that created a statistically significant difference was the mean age of the right and left colon (p< 0.001).

Postoperative complication rates of the patients were analyzed according to the Clavien-Dindo complication grade and when all groups were evaluated, major complication (3b and above) was seen in 35 (8%) patients. Anastomotic leakage was observed in 14 (3.3%), intra-abdominal collection in 14 (3.3%), reoperation in 17 (4%), wound infection in 46 (10.9%), extraintestinal infection in 65 (15.4%), and mortality in 18 (4.3%) patients. Postoperative infective complications of the patients according to lesion localization are summarized in Table 2.

The patients were divided into three groups according to lesion localization and the resections performed: right colectomy group consisted of 152 (36.02%) patients, whereas left colectomy group included 106 (25.12%) patients, and rectosigmoid resection group had 164 (38.86%) of the patients. Anastomotic leakage was observed in six (3.9%) patients in the right colectomy group, two (1.9%) patients in the left colectomy group, and seven (4.3%) patients in the rectosigmoid resection group. Intraabdominal collection rates were seen in six (3.9%) patients in the right colectomy, one (0.9%) in the left colectomy, and in seven (4.3%) patients in the rectosigmoid resection group. No statistically significant results were found between the three groups (p= 0.31) in terms of intraabdominal collections. Reoperation was seen in nine (5.9%), two (1.9%), four (2.4%) patients in the right colectomy, left colectomy, and rectosigmoid resection groups, respectively. Wound infection was seen in 15 (9.9%), 10 (9.4%) and 21 (12.8%) patients in the right colectomy, left colectomy, and rectosigmoid resection groups, respectively. Extraintestinal infection was seen in 23 (15.1%), 11 (10.4%) and 31 (18.9%) patients in the right colectomy, left colectomy, and rectosigmoid resection groups, respectively. Mortality was observed in 13 (8.6%), one (0.9%), four (2.4%) patients in the right colectomy, left colectomy, and rectosigmoid resection groups, respectively. There was no statistically significant difference between the groups regarding anastomotic leakage, intraabdominal collection, reoperation, wound infection, extraintestinal infection, and p values were 0.093, 0.31, 0.251, 0.612, and 0.234, respectively. Considering the mortality rates, it was found to be higher in the right colectomy group compared to the other groups, and the p value was 0.003. In addition, multivariable logistic regression analysis of the clinicopathological data of the patients according to the lesion localization is shown in Table 3.

When multivariable logistic regression analysis of the colon regions and complications was performed, anastomotic leakage among the patients who underwent rectosigmoid resection (p= 0.196, OR= 0.28, 95% CI for OR= 0.04-1.94) was found when the right colon was taken as a reference. For SSI (p= 0.219, OR= 0.59, 95% CI for OR= 0.25-1.37), complication status (p= 0.054, OR= 0.59, 95% CI for OR= 0.25-1.37), and collection (p= 0.521, OR= 0.57, 95% CI for OR= 0.25-1.37), no statistical difference was observed. In addition, anastomotic leakage (p= 0.462, OR= 0.41, 95% CI for OR= 0.04-4.36) and SSI (p= 0.493, OR= 0.71, 95% CI for OR= 0.27-1.87) compared to the left colon region, again when the right colon is referenced (0.27-1.87), and complication status (p= 0.183, OR= 0.13, 95% CI for OR= 0.01-2.59) and collection (p= 0.559, OR= 2.14, 95% CI for OR= 0.17-27.36) showed no difference. These findings are shown in Table 4.

Discussion

The present study evaluated postoperative outcomes in patients who underwent elective colorectal surgery without mechanical bowel preparation. The role of mechanical bowel preparation in colorectal surgery is still controversial. The negative effect on infection rates, the lack of effectiveness of mechanical preparation, and its use have led to a decrease (23). In line with the evidence of randomized trials and metaanalyses conducted in recent years, it has been understood that mechanical bowel preparation has no benefit on postoperative results (24, 25).

While SSI is 11.4% in colorectal surgery, it varies between 5.4% and 23.2% (26). In the European results, depending on the ERAS protocol, SSI rates of >10% have been observed in patients who did not undergo mechanical bowel preparation (27). In the MOBILE trial investigating mechanical and oral antibiotic bowel preparation (MOABP) versus no bowel preparation (NBP) in the right and left colectomy, subgroup analysis has shown that the rate of SSI in patients who underwent right colectomy was similar in the MOABP and NBP groups, 7% and 10%, 9%, respectively (OR= 0.71, 95% CI= 0.26-1.95; p= 0.510). In addition, SSI has been found at a similar rate in the MOABP and NBP groups who underwent left colectomy and were 6% and 10%, respectively (OR= 0.57, 95% CI= 0.18-1.82; p= 0.338) (28). The SSI rates in the current study were 10.9%. When we evaluated it as a subgroup, the rates of 9.9%, 9.4%, and 12.8% were observed in those who underwent right colectomy, left colectomy, and rectosigmoid resections, respectively.

Anastomotic leakage is among the most important causes of mortality after colorectal surgery. Anastomotic leakage rates reported in colorectal surgery vary between 1.8% and 19% (29). The present study also evaluated the effect of NBP on anastomotic leakage. In a study evaluating patients with and without MBP, anastomotic leakage rates were 2.3% and 2.6%, respectively; and there was no statistical difference (30). In addition, similar results are supported by other studies (31, 32). As demonstrated in a prospective randomized trial, there was no difference in anastomotic leakage between MBP and NBP among 249 patients who underwent rectal surgery. Anastomotic leakage rates were 4.2% and 2.3%, respectively (33). In our study, the rate of anastomotic leakage was 2.8%. In subgroup analysis, it was found as 3.9% in right colon surgery, 1.9% in left colon surgery, and 3.7% in patients with rectosigmoid surgery.

In a meta-analysis evaluating the effect of mechanical bowel preparation on postoperative outcomes in elective colorectal surgery, when MBP was compared with no MBP, there was no difference in the incidence of anastomotic leak (OR = 0.90, 95% CI= 0.74 to 1.10, p= 0.32) also in terms of SSI. When the studies were evaluated, no difference existed between those who underwent MBP and those who did not. Also, intraabdominal collection (OR = 0.86, 95% CI= 0.63 to 1.17, p= 0.34), reoperation (OR= 0.91, 95% CI= 0.75 to 1.12, p= 0.38) and mortality (OR= 0.50, 95% CI= 0.34 to 0.74, p= 0.0005) rates were evaluated in this meta-analysis, and effectiveness of MBP was observed on it (34). In our study, similar to this meta-analysis, when we performed and evaluation according to the resection sites of the colon, the rates of intraabdominal collection (p= 0.31) and reoperation (p= 0.251) were similar, but mortality rates (p= 0.003) were not different from the patients who underwent MBP, unlike this meta-analysis.

When the studies conducted in recent years are evaluated, there is disperancy in meta-analyses on mechanical bowel preparation, oral antibiotic use and IV antibiotic use before elective colorectal surgery. In a meta-analysis involving 5107 patients in 10 randomized controlled trials, patients have been grouped as IV antibiotics only, MBP with IV antibiotics, IV and oral antibiotics and MBP with oral antibiotics. Although there was no difference in terms of anastomotic leakage; SSI was seen to be reduced by more than 50% in patients who did not undergo MBP (35). In another meta-analysis, the analysis included a total of 22 studies involving 8852 patients. Patients were divided into two groups as MBP alone and MBP with oral antibiotics. As a result, the incidence of AL was significantly lower in the group treated with MBP plus OAB compared with MBP alone (OR= 0.43, 95% CI= 0.23-0.81, p= 0.009, I2= 73%). In addition, SSI was significantly lower in the MBP plus oral antibiotics group (OR= 0.38, 95% CI= 0.32-0.46, p< 0.0001, I2= 24%) (36).

The effect of gut microbiota composition on postoperative infectious complications after colorectal surgery has been demonstrated (37). When MBP is combined with oral antibiotics, both the microbiome and pathobionts are affected. MBP with oral antibiotics causes the disruption of the delicate balance between pathogen proliferation and natural suppression by rearrangement of the normal microbiota (38). In addition, the importance of the gut microbiota in its influence on gut sensorimotor function, which is associated with postoperative recovery of gut function, has been demonstrated in recent animal studies (39).

Although this study had several limitations, it also included some powerful features. The first significant limitation was the retrospective and single-center design. Although many studies evaluate the effectiveness of MBP, a vital aspect of the study was that the first study evaluated the outcomes of right and left hemicolectomy and anterior resection without MBP and compared them. Another strength of this study might be the large sample size. In addition, some patients need neoadjuvant chemotherapy, which is expected to increase postoperative complications. Nevertheless, the results of this study showed that using neoadjuvant chemotherapy might not increase postoperative infective complications under the condition of nonmechanical bowel preparations.

Conclusion

Surgical site infections are in an Achilles heel condition after colorectal surgery. Within the framework of the ERAS protocols, mechanical and oral antibiotic bowel preparations have been abandoned for decades. However, the rate of anastomotic leakage, one of the most feared complications after colorectal surgery, has not changed. Contrary to dogma and popular belief, data from patients who did not undergo mechanical bowel preparation were analyzed and discussed with the current literature in this study. Surgical site infection, postoperative mortality, intraabdominal collection rates, and anastomotic leakage were similar.