Introduction

As laparoscopic surgery has been penetrated rapidly in the management of colon cancer based on evidence of oncologic safety and advantages of short-term surgical outcomes, it has been also tried for treatment of rectal cancer because of the potential benefits of a minimally invasive approach [1–3]. However, laparoscopic rectal surgery has been regarded as a challenging procedure. A high conversion rate or a considerable rate of circumferential resection margin involvement in a well-designed multicenter clinical trial might reflect the technical complexity of laparoscopic rectal surgery [2]. Postoperative complications delayed patient recovery. Even worse, conversion or anastomotic leakage has been proven to be associated with poor prognosis [3–5]. For this reason, proper patient selection for laparoscopic rectal surgery would be important in efforts to obtain better outcomes.

Obesity has been known as a contributing factor to technical difficulties in the performance of surgery and in postoperative morbidities in colorectal cancer surgery [6]. An excessive amount of visceral fat could be an obstacle to finding and keeping the exact dissection plane during the procedure, especially in the laparoscopic approach [7, 8]. Moreover, the number of retrieved lymph nodes could be affected by obesity [9, 10]. Body mass index (BMI) has been used to define obesity. In western studies, a BMI over 30 kg/m2 is usually defined as obese using the World Health Organization (WHO) classification [7, 11]. However, because the average BMI might be different according to geography [12, 13], the definition of obesity by BMI differs among various ethnic groups in that obesity in Japan is adequately specified as BMI ≥ 25 kg/m2 by the Japan Society for the Study of Obesity [14]. Furthermore, it has been reported that BMI was not always consistent with visceral fat area [12, 14]. Visceral fat has been proposed as an alternative to BMI for a more accurate prediction of surgical complexity and morbidity in the management of colon cancer with laparoscopic technique [15, 16]. In laparoscopic rectal cancer surgery, precise prediction of obesity would be valuable in the determination of the high-risk group for technical difficulty, which could result in the optimization of minimally invasive techniques. However, the significance of visceral fat in surgical outcomes after laparoscopic rectal resection remains inconclusive since there are few reports on the impact of visceral fat obesity (VFO) on surgical morbidity and oncologic outcomes [17].

The aim of this study was to investigate the impact of VFO on early postoperative outcomes and pathological parameters in laparoscopic rectal cancer surgery.

Patients and methods

Eligibility

Between June 2003 and June 2009, a total of 142 patients who underwent laparoscopic surgery with curative intent for adenocarcinoma located within 12 cm from the anal verge were included in the analysis. All included patients underwent abdomino-pelvic computed tomography (APCT) before surgery. No distant metastasis was observed at initial diagnosis. We excluded patients (1) who underwent laparoscopic abdomino-perineal resection or who underwent hand-sewn coloanal anastomosis, (2) for whom visceral fat area could not be calculated from the APCT in the workstation, and (3) who underwent combined resection, which meant that another part of an organ (e.g. gall bladder, ovary, uterus, seminal vesicle, and liver) was resected simultaneously with the main rectal resection.

Preoperative evaluation was performed using a combination of digital rectal examination, rigid sigmoidocopy, colonoscopy, chest X-ray examination, APCT, transrectal ultrasonography, and magnetic resonance imaging.

Surgical procedure

After pneumoperitoneum was accomplished using carbon dioxide, we used either a medial-to-lateral or a lateral-to-medial approach for dissection. The inferior mesenteric artery was ligated at the level of the root (high ligation technique). Splenic flexure mobilization was optional. Total mesorectal excision (TME) was performed for middle or low rectal cancers. TME included the complete removal of the mesorectum circumferentially and distally with dissection to the level of the levator ani muscle. After the dissection was completed, the rectum was transected with an endoscopic linear stapler. The specimen was extracted through a 5–6-cm-sized left lower abdominal incision or low pfannenstiel incision. End-to-end anastomosis was performed using the circular stapler. Air leakage test was performed to confirm anastomosis completeness. Diverting loop ileostomy was applied when assumed necessary by the surgeon depending on the anastomosis status in each patient. Pelvic drain was indwelled routinely [18].

Definitions of perioperative surgical outcomes and data collection

Open procedure was defined as performance of a conventional laparotomy, with an abdominal incision greater in size than that initially planned for specimen delivery [19]. Anastomotic leakage was diagnosed based on clinical signs or image studies. Clinical signs include abdominal pain or fever, discharge of pus or feces through the indwelling drain, local or generalized peritonitis, pus discharge through the anus, etc. Radiologic evaluation was performed to confirm clinical suspicion [18]. Pathologic variables and postoperative outcomes were extracted from our prospectively collected database.

Visceral fat area measurement

All included patients underwent preoperative APCT. A single cross-section image from APCT at the level of the L4–L5 intervertebral space was obtained (Fig. 1) [15, 16, 20–22]. Two experienced radiologists (SE Baek and JS Lim) reviewed all of the APCT studies and measured the total fat area, visceral fat area, and subcutaneous fat area. For measurement of these parameters, the fat margin was manually traced, and calculation of the area was processed on a workstation using imaging software (Aquarius workstation; TeraRecon, San Mateo, CA, USA) (Fig. 1). Radiologists were not informed with regard to clinical data and pathologic results prior to measurement.

Fig. 1
figure 1

Measurement of the visceral fat area. Visceral fat areas were obtained using abdomino-pelvic computed tomography and TeraRecon ™ software. Using one cross-sectional scan at the level of the L4–L5 intervertebral space, intraperitoneal tissue is defined by manually tracing its contour on the scan. The green-colored area is the manually measured visceral fat area and the purple-colored area is the subcutaneous fat area. In this picture, both patients have almost the same body mass index (a 24.79 kg/m2; b 24.8 kg/m2); however, the visceral fat area is significantly different (a 43.3 cm2, b 230.2 cm2)

Full size image

Definition of obesity

VFA ≥130 cm2 was defined as VFO and BMI ≥25 kg/m2 was defined as BMI obesity (BO). This threshold value was adapted from the previous study [16]. Patients were divided into the obese and the non-obese group according to BMI and VFA and early surgical and oncologic parameters were compared.

Statistical analysis

All statistical analyses were performed using SPSS software, version 12.0 (SPSS, Chicago, IL, USA). Categorical variables were analyzed using chi-square or Fisher’s exact test, and continuous variables were analyzed using Student’s t-test. All variables associated with open conversion or less than 12 examined lymph nodes, with p < 0.25 in univariate analysis, were entered into multivariate analysis using logistic regression analysis. Disease-free survival (DFS) was defined from the date of surgery to the date of detection of recurrence, last follow-up, or death. Differences in survival between groups were compared using the Kaplan–Meier method and tested with the log-rank test. A p value of <0.05 was considered to indicate significance.

Results

Demographics

A total of 37 (26.0%) patients were categorized as obese patients by BMI, whereas 29 (20.4%) patients were grouped as obese patients by VFA. Incidence of cardiovascular disease was significantly higher in the obese group using either definition. The VFO group was associated with older age and a higher rate of American Society of Anesthesiologist (ASA) grade more than 2. There was no difference in other parameters between the obese and the non-obese group in either definition (Table 1).

Table 1 Clinicopathological characteristics according to BMI and visceral fat
Full size table

Surgical outcomes

Compared with the non-obese group, the obese group, defined by both VFA and BMI, had a significantly longer operative time. The VFO group experienced significantly higher blood loss during surgery (205.8 ± 257.0 mL vs. 102.5 ± 219.9 mL; p = 0.031), whereas there was no difference when defined by BMI. Time to the first flatus in the VFO was significantly longer compared with that of the non-obese group (mean 3.5 days vs. 2.7 days; p = 0.046), whereas it was not significantly different when classified by BMI. However, the difference did not last in other following parameters such as day to oral intake or hospital stay (Table 2).

Table 2 Comparison of operative outcomes and early recovery parameters between groups according to BMI and visceral fat obesity
Full size table

The VFO group experienced more frequent conversion to laparotomy (17.2% vs. 5.3%; p = 0.047) (Table 3).

Table 3 Reasons for conversion to open laparotomy
Full size table

Multivariate analysis revealed that a previous abdominal surgery history and VFO were independent predictors of open conversion (hazard ratio 6.3; 95% CI 1.7–23.6, p = 0.006; hazard ratio 4.0; 95% CI 1.0–15.5, p = 0.041, respectively) (Table 5).

Postoperative recovery

One postoperative mortality occurred during the study. Although the patient showed good recovery of gastrointestinal function, she failed to extubate several times and died due to pulmonary problems on postoperative day 49. Overall morbidity rate was 23.2%. Regarding operation-related morbidities, there was no significant difference between obese and non-obese patients using either definition (Table 4).

Table 4 Comparison of operation-related morbidity and mortality between BMI and visceral fat obesity
Full size table

Early oncologic outcomes

Regarding oncological parameters, significantly more patients in the VFO group had less than 12 retrieved lymph nodes (65.5% vs. 34.5%; p = 0.002); however, there was no difference between obese and non-obese groups defined by BMI. Multivariate analysis revealed that a tumor size of less than 5 cm, (y)pT stage(pCR or pT1 or pT2), and VFO were independent predictors of inadequate lymph node retrieval (less than 12 retrieved lymph nodes) (Table 5).

Table 5 Multivariate analysis of surgical outcomes
Full size table

Long-term follow-up results

All patients with diverting loop ileostomy underwent ileostomy repair within several months after the initial surgery (median 6, range 2–11 months). After a median follow-up period of 31.4 months (range 1.6–90.4 months), there was no significant difference of 3-year DFS between the obese group and the non-obese group as defined by BMI and VFA (3-year DFS: 82.5% in patients with BMI < 25 kg/m2 vs. 76.9% in patients with BMI ≥ 25 kg/m2, p = 0.390; 82.5% in patients with VFA < 130 cm2 vs. 74.4% in patients with VFA ≥ 130 cm2, p = 0.262) (Fig. 2).

Fig. 2
figure 2

Comparison of disease-free survival between obese group and non-obese group using either definition. a Three-year DFS: 82.5% in patients with BMI < 25 kg/m2 vs. 76.9% in patients with BMI ≥ 25 kg/m2, p = 0.390. b Three-year DFS: 82.5% in patients with VFA < 130 cm2 vs. 74.4% in patients with VFA ≥ 130 cm2, p = 0.262

Full size image

Discussion

The impact of VFO on early surgical or oncological outcomes in patients who have undergone laparoscopic rectal surgery remains controversial. This study shows that VFO can predict technical difficulties in laparoscopic rectal resection more accurately than BMI in terms of increased conversion rate, increased blood loss during surgery, and delayed bowel function recovery. With regard to oncologic outcomes, VFO is associated with a higher proportion of less than 12 retrieved lymph nodes.

Obesity has been known to influence surgical outcomes in open colorectal surgery [6]. With respect to laparoscopic colorectal surgery, the impact of obesity on surgical outcomes is inconclusive yet. Some reports have stated that obesity is associated with prolonged operation time, greater risk of postoperative complications, and higher conversion rate [7, 23]. In contrast, other studies have insisted that obesity did not adversely influence postsurgical outcome [11, 24]. In these studies, obesity was determined by BMI since BMI has been accepted as a determinant of obesity due to its simplicity. It has been advocated that most of the technical difficulties during laparoscopic resection originated from limited exposure or difficult dissection due to an excessive amount of visceral fat [7, 8]. It has been reported that BMI was not always consistent with the visceral fat area [12, 14]. Therefore, several studies have attempted to define obesity using visceral fat area. In laparoscopic colorectal surgery, determination of obesity with visceral fat area was reported as a more precise parameter for the prediction of postoperative complications than BMI [15–17]. In the current study, operation time in the obese group, regardless of the method of determination, was significantly longer than that of the non-obese group. However, it could be found only in the VFO group, not in the BMI obese group, that prolonged operation time was associated with an increased amount of blood loss and significant delay of bowel function recovery. Tumor infiltration/fixation, obesity, and adhesion have been known as the most common reasons for conversion in laparoscopic rectal surgery [3]. In this study, it was confirmed that a previous abdominal surgical history and obesity defined by visceral fat area were independent predictors of open conversion in multivariate analysis. However, there was no significant increase in conversion rate in the obese group defined by BMI. These results strongly indicate that visceral fat obesity was a more reliable factor than BMI in the prediction of surgical complexity in laparoscopic rectal resection.

In terms of postoperative complications, surgical morbidity after laparoscopic sigmoidectomy was reported to be significantly increased in the VFO group compared with the non-obese group [16]. Increased morbidity was derived mainly from different rates of wound infection between the obese group and the non-obese group [16]. Our study did not show any significant difference in overall morbidity rate between the obese and the non-obese groups. Also, the individual complication rate, including wound infection and anastomotic leakage, did not differ between the two groups. In our institute, when keeping the right surgical plane in the usual manner was difficult during the laparoscopic approach, conversion to open surgery without delay was indicated. For this reason, VFO would not result in increased morbidity but in an increased rate of conversion to open surgery.

Increased conversion rate and decreased number of retrieved lymph nodes have been found to be associated with VFO in our study. Conversion to open surgery is not only a reflection of surgical complexity but is also regarded as a cause of poor prognosis [3, 4]. Adverse impacts of conversion, such as delayed patient recovery in early postoperative periods or poor oncologic outcomes, may emphasize the need for adequate patient selection in laparoscopic surgery. Therefore, a history of previous abdominal surgery and VFO rather than BMI should be taken into account in patient selection, especially during the early learning curve periods when performing laparoscopic rectal surgery.

In management of colorectal cancer, an adequate number of retrieved lymph node is important for accurate tumor staging and patient prognosis [25, 26]. Higher BMI has been reported as an obstacle to the performance of proper lymph node dissection and was an independent predictor of disease recurrence in T2/T3 gastric cancers [9]. However, the association of retrieved lymph node numbers with obesity was not clearly identified in colorectal cancer. In the current study, VFO patients had a smaller mean number of retrieved lymph nodes and a higher proportion of less than 12 examined lymph nodes. Additionally, VFO was proven as an independent risk factor for inadequate number of retrieved lymph nodes in multivariate analysis. In contrast, using BMI, there was no difference in the mean numbers of retrieved lymph nodes or in the proportion of inadequate lymph node retrieval between obese and non-obese patients. Various factors are known to affect the number of retrieved lymph nodes in rectal cancer surgery [25, 27]. It has been reported that the most significant reduction of mean retrieved lymph node numbers in rectal cancer was observed in obese patients with short specimen length [10]. However, in our series, no significant difference in specimen length was observed between the groups (data not shown). In another point of view, VFO “per se” might act as an obstacle to a thorough search for lymph nodes. During lymph node detection in pathologic examination, it has been well established that the presence of fat tissue makes lymph node discovery difficult, especially in rectal cancer [28]. Recently, another interesting experimental result showed that visceral fat accumulation caused atrophy of mesenteric lymph nodes by inducing T-cell activation-mediated apoptosis in obese mice [29]. Though direct application of this result to humans may be inappropriate, further investigation will be needed.

In the present study, straight end-to-end anastomosis was the preferred method. It was reported that colonic J-pouch (CJP) reconstruction increased reservoir function and provided better short-term functional results than straight anastomosis [30, 31]. CJP reconstruction after TME has become the preferred method in the west. However, in the long-term follow-up, stool evacuation problems were reported, requiring digital evacuation or the use of laxatives [32]. Besides, CJP reconstruction needs a more extensive dissection in order to completely take down the splenic flexure for tension-free anastomosis and could cause difficulty in secure anastomosis because of a bulky colonic J-pouch [31, 33]. For these reasons, all of the enrolled patients in this study underwent straight end-to-end anastomosis.

It should be mentioned that this study has the limitation of defining obesity using BMI. According to WHO classification, obesity by BMI is defined as over 30 kg/m2, while BMI over 25 kg/m2 is defined as overweight. However, due to the diversity of mean BMI according to geography, the definition of obesity by BMI is inevitably different among various ethnic groups [12]. Asian populations may have greater visceral adiposity; however, BMI is relatively lower than in non-Asian populations [14]. The percentage of the population with BMI over 30 kg/m 2 is no more than 2.0–3.0% in Japan [34]. In our study group, only 3.5% of the enrolled patients are equivalent to the definition of obesity by BMI. For this reason, obesity by BMI would be defined with the cutoff value of 25 kg/m2 in studies with Asian populations [15–17]. Even though our definition of obesity is not relevant to regions of higher BMI, we thought that the contradictive impact of BMI on surgical outcomes for laparoscopic rectal surgery originated mainly from the fact that BMI could not adequately reflect visceral fat adiposity. To determine more reliable parameters presenting visceral fat adiposity, the impact of visceral fat in the field of surgery such as gastric or colorectal operation has been discussed mainly in Asian populations [15–17, 35, 36]. According to previous results on the impact of visceral fat on surgical outcomes, including ours, visceral fat obesity may be a useful tool for prediction of difficulties in general surgery. However, to clarify the conclusion, further studies, especially after including more patients with higher BMI, are warranted.

In this study, we intended to investigate the usefulness of visceral fat obesity for prediction of surgical difficulties. For this purpose, the validity or feasibility of visceral fat measurement should be established. Visceral fat measurement with CT scan might be one of the reliable modalities for use in defining visceral fat obesity. It was reported that visceral fat areas from a single scan obtained at the level of the umbilicus (approximately at the level of L4 and L5) were highly correlated with the total visceral fat volume [20–22]. The visceral fat area was measured manually by a radiologist and then calculated mechanically by a computer system [15, 16]. It was a simple procedure, not a special technique; therefore, we believe that any radiologist could easily perform the VF measurement using CT imaging. However, it might take some time and require special software and equipment. These situations might be the limitation of using VFA in daily clinical work. In one study, waist circumference showed the closest relationship with VFA among convenient anthropometric parameters, such as BMI, waist circumference, waist/hip ratio, and waist/body height ratio [14]. As far as we know, no simple alternative parameter for VFA measurement has been definitely confirmed according to current evidence. This might be another issue for further evaluation. Besides, the cutoff value of 130 cm2 for definition of visceral obesity in this study might be somewhat arbitrary. We expect that additional investigation on this subject could result in a more delicate or reliable parameter on the definition of visceral fat obesity.

In conclusion, VFO can be used as a more accurate predictor of surgical complexity than BMI. VFO should be taken into account especially during the early period of the learning curve when planning laparoscopic rectal resection. Additionally, VFO is associated with a smaller number of retrieved lymph nodes. Therefore, close pathologic inspection using a fat clearing method might be required for VFO patients after laparoscopic rectal surgery, although long-term oncologic impact of the small number of examined lymph nodes in these patients needs to be carefully investigated.