Visceral obesity is a preoperative risk factor for postoperative ileus after surgery for colorectal cancer: Single‐institution retrospective analysis

Abstract Aim Visceral obesity (VO) reportedly has a stronger association with complications after colorectal surgery than does body mass index. Here, we retrospectively assessed VO as a risk factor for postoperative ileus (POI) after colorectal resection in patients with colorectal cancer. Methods This study included 417 consecutive patients with colorectal cancer who underwent elective surgery at our institute from January 2010 to December 2012. Visceral fat area (VFA) was calculated by image analysis software. VO was defined as VFA ≥100 cm2. We assessed 49 factors, including VO, comorbidities, surgical procedure, and postoperative complications. Data were analyzed using a propensity score‐matching strategy. Results Postoperative ileus occurred in 18 patients (4.3%) from the entire cohort, and in 14 (5.5%) of the 256 matched patients. Multivariate analysis (n = 417 patients) showed that significant risk factors for POI included VO (OR 7.9, 95% confidence interval [CI] 1.9‐32.1, P = .004), open surgery (OR 6.4, 95% CI 1.6‐26.7, P = .010), and pelvic/intra‐abdominal abscess (OR 11.0, 95% CI 1.1‐110.2, P = .041). Propensity score matching showed two independent risk factors in the multivariate analysis: VO (OR 6.2, 95% CI 1.3‐30.4, P = .025) and open surgery (OR 9.1, 95% CI 2.0‐40.5, P = .004). Conclusion Visceral obesity may be an independent risk factor for POI in patients with colorectal cancer.


| INTRODUC TI ON
Despite remarkable progress in the field of colorectal surgery, postoperative complications remain a major problem during the course of colorectal cancer treatment. [1][2][3][4] After colorectal surgery, postoperative ileus (POI) occurs at a frequency of 10%-17%, resulting in longer hospital stays and higher costs. [3][4][5][6][7] POI is characterized by lack of bowel sounds, delayed passage of flatus and stool, abdominal distension, nausea, vomiting, and pain. 5,7 Postoperative complications, including POI, after surgery for rectal cancer are reportedly associated with delays in adjuvant chemotherapy, and patients who receive delayed adjuvant chemotherapy have worse recurrence rates and worse overall survival than patients who receive chemotherapy within 8 weeks of surgery. 8 Preoperative identification of risk factors could enable improved postoperative management for patients at higher risk of complications.
Prior studies have assessed the risk factors for POI, including male gender, peripheral vascular disease, respiratory comorbidity, preoperative albumin, stoma construction, operation lasting over 3 hours, conversion to open surgery, and intra-abdominal surgical site infection. 3,5,9 Compared to open surgery, laparoscopic colorectal surgery is associated with lower incidence of postoperative bowel obstruction. 10 Some studies have reported that body mass index (BMI), which is widely used for the assessment of general obesity, is also a risk factor for POI. 6,11 However, recent findings suggest that, compared to BMI, visceral obesity (VO) is more strongly associated with complications after colorectal surgery. 12,13 In an analysis of 338 consecutive patients with colon cancer, Watanabe et al reported that VO was more strongly related to the incidence of anastomotic leakage and surgical site infection than high BMI. 13 However, to the best of our knowledge, no study has shown that VO is a risk factor for POI after surgery for colorectal cancer.
In the present study, we assessed whether VO is a risk factor for POI among patients with primary colorectal cancer.

| Study population
This study involved patients with primary colorectal cancer who underwent elective surgery at Osaka University Hospital between January 2010 and December 2012. We included patients who underwent one of 12 surgical procedures: ileocecal resection, right hemicolectomy, transverse colectomy, left hemicolectomy, sigmoidectomy, anterior resection, low anterior resection, super-low anterior resection, abdominoperineal resection, intersphincteric resection, Hartmann's operation, or total pelvic exenteration. Patients who underwent subtotal colectomy, total colectomy, or two different procedures (eg, ileocecal resection and sigmoidectomy) were systematically excluded ( Figure 1). All patients provided informed consent, and patient anonymity was preserved. This study was approved by the ethics committee at our institution.
F I G U R E 1 Flow diagram of patient inclusion in the present study. VO, visceral obesity F I G U R E 2 Images from the SYNAPSE VINCENT (Fuji Medical Systems, Tokyo, Japan) 3-D image-analysis system. Visceral fat is coloured in red, and its area was automatically calculated. A, Image from a patient with visceral obesity (visceral fat area = 231 cm 2 ). B, Image from a patient without visceral obesity (visceral fat area = 23 cm 2 )

| Definitions
Postoperative ileus was defined as a Clavien-Dindo grade II or higher ileus within 30 days after surgery. Ileus was diagnosed when patients complained of nausea, vomiting, or abdominal distension, and dilatation of the small bowel was radiologically confirmed without obvious small bowel obstruction. 14 All patients underwent computed tomography (CT) prior to surgery, and visceral fat area (VFA) was calculated at the umbilicus level using the SYNAPSE VINCENT (Fuji Medical Systems, Tokyo, Japan) 3-D image analysis system. Visceral fat, which has a Hounsfield unit threshold of −150 to −30, was coloured red in the images and its area calculated automatically ( Figure 2). VFA ≥100 cm 2 was considered to indicate VO. 12,15,16

| Data collection
We retrospectively collected demographic and clinicopathologi-

| Propensity score matching
We used a propensity score-matching strategy to identify a popula-

| Propensity score matching
To reduce the possibility of selection bias, we conducted propensity score matching. A total of 256 cases were matched (128 cases each).

| Univariate and multivariate analyses of POI risk factors in matched patients
Univariate analysis showed that patients with and without POI had significantly different rates of VO, open surgery, operation time >228.5 minutes, and estimated blood loss >50 mL ( Table 6)

| D ISCUSS I ON
Postoperative ileus is a frequent postoperative complication after colorectal surgery at our institution. POI is reported to result in more extended hospital stays and higher costs, as mentioned earlier.
Therefore, we assessed the risk factors for POI in this setting. To effectively prevent POI, we must identify risk factors that can be assessed prior to surgery. In multivariate analysis including all 417 patients in our study cohort, we found that VO, open surgery, and pelvic/intra-abdominal abscess are independent risk factors for POI.
Of these, VO is a predictive factor that can be preoperatively assessed. Thus, we focused on analyzing VO as a risk factor for POI.
As obesity is associated with both tumour initiation and progression in colorectal cancer, 17,18 it is commonly encountered among patients requiring colorectal surgery for colorectal cancer.
Previous reports have shown that, compared to BMI, VO is more strongly associated with complications after colorectal surgery. 12,13 Among 338 patients who underwent colorectal resection for colorectal cancer, and 75 patients who underwent total gastrectomy for gastric cancer, VO was more strongly related to incidences of anastomotic leakage and surgical site infection than high BMI. 13,19 However, to the best of our knowledge, no previous study has included multivariable analysis to assess VO as a risk factor for POI, and the present study is the first to show that VO may be a risk factor for POI after surgery for colorectal cancer.
Some studies have reported that BMI, which is widely used for the assessment of general obesity, is a risk factor for POI. 6,11 He et al.
carried out a meta-analysis, showing that BMI >30 is a risk factor for POI after laparoscopic colorectal surgery (OR 1.73, P = .02); however, they did not assess VO as a potential risk factor. 11 In contrast, our present findings did not indicate that higher BMI is a risk factor for POI, although we used a cut-off of 25 for higher BMI because only eight of our 417 patients had a BMI >30. This is compatible with other reports indicating that BMI is not a risk factor for POI after colorectal surgery. 3,20 In the present study, VO was an independent risk factor for POI, suggesting that VO is a more accurate predictor of POI after colorectal surgery than BMI. OR for VO and POI was 7.9 in the multivariate analysis, and 6.2 after propensity score matching, indicating that VO may have an impact on postoperative outcome as well as on POI.
Preoperative assessment of VO as a predictor of POI may enable action to be taken before surgery to prevent or reduce POI. we did not assess postoperative levels of IL-1 family members in the patients, but VO can be used as a surrogate marker of these levels.
Fifth, the enhanced recovery after surgery (ERAS) protocol was not introduced at our institution. POI prevention is one of the aims of the ERAS protocol, 28,29 and Ni et al reported that ERAS resulted in shorter average length of postoperative hospital stay, time to first flatus, and time to first defecation than traditional perioperative care after colorectal surgery in 1298 patients in a meta-analysis of randomized controlled trials. 29 They also reported that the overall complication rates were significantly lower with ERAS than with traditional perioperative care. Although, currently, the ERAS protocol is not common in Japan, it may be effective, especially for patients with VO. Despite these limitations, we think that the present study is of great importance because it is the first report of VO as a risk factor for POI after surgery among patients with colorectal cancer.
In conclusion, the present study showed that VO is an independent risk factor for POI after colorectal resection among an entire cohort of 417 patients and among 256 matched patients.
This finding enables the assessment of POI risk before surgery by checking VFA, as VFA >100 cm 2 may indicate a need for perioperative interventions, such as opioid-avoidance strategies and giving prokinetic agents before and after surgery. Furthermore, selectively introducing the ERAS protocol in Japan for patients with VO may be acceptable and effective. The present findings should be verified in a prospective multicenter study with a greater number of patients.

ACK N OWLED G EM ENT
Author T.M. was supported by donations from Kinshukai Medical Corporation.

D I SCLOS U R E
Conflicts of Interest: Authors declare no conflicts of interest for this article.
Author Contribution: All authors are in agreement with the content of the manuscript.