Deep pelvis and low visceral fat mass as risk factors for neurogenic bladder after rectal cancer surgery

Abstract

Background

Postoperative neurogenic bladder (PONB) frequently occurs as a complication after rectal cancer surgery. This study aimed to analyze risk factors for developing PONB after rectal cancer surgery, particularly the association between pelvic anatomy and visceral fat mass.

Methods

We included 138 patients who underwent rectal resection for lower rectal cancer in our department between 2017 and 2021. PONB was defined as the need for urethral catheter reinsertion or oral medication administration for urinary retention after catheter removal with severe NB that required treatment for ≥ 60 days. We obtained visceral fat area (VFA) at the umbilical level based on a CT scan and measured five pelvic dimensions.

Results

Of the 138 patients, 19 developed PONB, with 16 being severe cases. PONB more frequently occurs in patients with a height of < 158 cm, age ≥ 70 years, surgery lasting ≥ 8 h, intraoperative bleeding volume ≥ 150 mL, lateral lymph node dissection, and narrower pelvis. It was more prevalent in cases with low VFA. Conversely, gender, body mass index (BMI), and medical history showed no significant correlations. Multivariate analysis revealed older age, prolonged surgery, and low VFA as independent risk factors for PONB. Independent risk factors for severe PONB included low VFA, older age, prolonged surgery, and deep pelvis.

Conclusion

Lower VFA, older age, and prolonged surgery are independent risk factors for developing PONB. Additionally, a deep pelvis is an independent risk factor for severe PONB. Delicate surgical techniques should consider the risk of nerve injury in cases with low VFA and deep pelvis.

Introduction

Rectal cancer surgery is one of the most challenging procedures of gastrointestinal surgery. This is due to the need to achieve conflicting events simultaneously: the need to operate within a narrow pelvis and to achieve oncologic margins. Rectal resection is associated with early postoperative complications such as anastomotic leakage and surgical site infection. Anastomotic leakage is a severe and potentially fatal complication. On the other hand, there are many complications that persist long after surgery and impair the patient's quality of life, such as anorectal dysfunction and sexual dysfunction. One of these complications that can persist long after surgery is postoperative neurogenic bladder (PONB) [1].

Postoperative neurogenic bladder occurs in 5.5%–16.0% of patients after rectal cancer surgery [1]. Once PONB occurs, the prolonged indwelling of a urinary catheter is frequently required, which significantly reduces the patient’s quality of life. PONB is mainly due to intentional or unintentional nerve injury intraoperatively [2]. A meta-analysis revealed male sex, older age, diabetes mellitus, urological diseases, tumor location in the lower rectum, abdominoperineal resection (APR), laparoscopic surgery, and longer operation time (≥ 4 h) as risk factors for PONB in colorectal surgery [1].

Rectal cancer surgery is a complicated procedure because of the anatomic limitations of the surgical field. In particular, visceral obesity and narrow/deep pelvis have been reported as factors that make rectal cancer surgery technically more difficult [3, 4]. PONB prevalence is expected to increase in anatomically challenging cases because direct nerve damage due to intraoperative manipulation is thought to increase. However, the association between PONB occurrence and such anatomic characteristics remains unclear. Therefore, this study investigated the risk factors and anatomical characteristics, particularly related to PONB development, in rectal cancer surgery.

Materials and methods

Patients

The institutional review boards of Kyoto Prefectural University of Medicine (ERB-C-1178–1) approved this study. This study conducted a database-based retrospective analysis of 138 consecutive patients who underwent primary resection for rectal cancer at Kyoto Prefectural University of Medicine between 2017 and 2021. Overweight patients with a BMI between 25 and 29.9 accounted for 31 cases or 22% of the total. Obesity patients with a BMI over 30 accounted for 5 cases or 3.6% of the total. Rectal cancer was defined as a tumor within 12 cm from the anal verge (AV). The rectum is divided into three sections according to the distance from the AV. AV 8–12 cm sites were defined as high, 4–8 cm sites as middle, and < 4 cm sites as low. The tumor stage was classified according to the eighth tumor node metastasis classification system [5].

Preoperative treatment

Patients who received chemoradiotherapy (CRT) underwent long-course treatment. The patients had preoperative radiation of a total dose of 45 Gy and concomitant chemotherapy with oral TS-1. All patients underwent mechanical bowel preparation with 34 g of magnesium citrate along with 180 mL of water and 75 mg of sodium picosulfate during the daytime the day before surgery. As chemical bowel preparation, 1000 mg of kanamycin and 1000 mg of metronidazole were given in two divided doses after evening meal and before bedtime the day before surgery.

Surgical procedure

Laparotomy, laparoscopy, and robotic surgery were used for the operations. Da Vinci Si/X/Xi was used for robot-assisted surgery. The degree of lymph node dissection was determined by the preoperative tumor staging based on the colorectal cancer treatment guidelines, in addition to lateral lymph node dissection in middle/low tumors with cT3/4 cases [6, 7]. The rectum was divided > 3 cm below the lower edge of the tumor in the case of high rectal cancer and > 2 cm below the lower edge of the tumor in middle/low rectal cancer, except for APR. Diverting ileostomy was constructed in all patients in cases that underwent very low anterior resection (vLAR) or intersphincteric resection (ISR). The creation of ileostomy depended on the surgeons’ discretion in other cases, considering the risk of anastomotic leakage.

Definition of PONB and other complications

The urinary catheter was usually removed 3 days postoperatively or 7 days after lateral lymph node dissection. The urologist was consulted for the urinary retention after catheter removal, and they diagnosed the patient who had the catheter reinserted and was started on medical therapy with PONB. Patients who needed PONB treatments for > 60 days were defined as severe PONB. Postoperative ileus was defined as Clavien–Dindo classification ≥ II among patients who developed vomiting symptoms within 30 days postoperatively and revealed small bowel dilatation on X-ray examination or computed tomography (CT) scan.

Anatomical measurement

We retrospectively obtained patients’ anatomical characteristics, such as visceral fat area (VFA) and pelvic diameters. VFA was measured using a preoperative cross-sectional CT scan obtained at the single level of the umbilicus (Fig. 1a). We determined the adipose tissue amount by setting the attenuation level from − 50 to − 200 HU and then defined VFA by contour tracing from the whole fat region [8]. The VFA was calculated by Ziostation2 (Ziosoft Inc., Tokyo, Japan) automatically. Cut-off values for VFAs associated with PONB occurrence and severe PONB were calculated using receiver operating characteristic (ROC) curve analyses. The threshold values associated with PONB and severe PONB occurrence were 115 cm² and 35 cm², respectively (Supplement Fig. 1).

Fig. 1

Measurement of each anatomical structure. The measurements of visceral fat area and pelvic diameter were shown. a The area of visceral fat area (VFA) was calculated on the umbilical slice in the preoperative computed tomography (CT) image. b Interspinous distance: distance between the tips of the ischial spines. c Intertuberous distance: distance between the lowest points of the ischial tuberosities. d Pelvic inlet (triangle): distance from the superior aspect of the pubic symphysis to the promontory. Pelvic outlet (arrow): distance from the inferior aspect of the pubic symphysis to the tip of the coccyx. Pelvic depth (star): distance from the promontory to the tip of the coccyx

Pelvic diameters were obtained from the five linear distances of CT scan, as in the study of Escal et al. [3]. The interspinous distance was the shortest distance between the tips of the ischial spines (Fig. 1b). The intertuberous distance was the shortest distance between the lowest points of the ischial tuberosities (Fig. 1c). The pelvic inlet, outlet, and depth were the shortest distance between the sacral promontory and pubic symphysis, tip of the coccyx and pubic symphysis, and sacral promontory and tip of the coccyx, respectively (Fig. 1d).

Statistical analysis

We investigated PONB risks regarding clinical characteristics, preoperative treatment, operative outcomes, postoperative course, pathological results, VFA, and pelvimetric information. Continuous variables were presented as median and range, and categorical variables were presented as number and percentage. We used Fisher’s exact test to compare the two groups for the analysis. A p-value of < 0.05 was considered significant. The optimal cut-off values for height, VFA, and pelvimetric factors were determined using ROC curve analysis. We performed multivariate analysis by stepwise logistic regression to determine the most influential predictive factors of PONB. We used JMP version 16.2.0 (SAS Institute, Cary, NC, USA) for all statistical analyses.

Results

Patient characteristics

Table 1 shows the patient characteristics. This study enrolled 138 patients, including 57 men and 81 women. The median age was 67 years, height was 160 cm, and body mass index (BMI) was 22.3 kg/m². Middle rectal cancer was the most prevalent, accounting for 55% of the cases. Of the patients, 11 had previously undergone endoscopic tumor resection and were operated on for the additional resection, 31 had neoadjuvant chemoradiotherapy (CRT), and 3 patients had only neoadjuvant chemotherapy (NAC). The median VFA was 91.2 cm², and Table 1 describes the pelvimetries. LAR/vLAR was the most common type of resection, accounting for 74 cases, and robotic surgery was the most common surgical approach. All cases underwent robotic surgeries since January 2020 except for one case, which converted to laparotomy. The median surgical duration was 375 min, and the estimated blood loss was 41 ml. A total of 6 patients, who were converted to laparotomy because of accidental bleeding or severe intraabdominal adhesion, were included in the laparotomy group because the pelvic manipulations were performed after the conversion.

Table 1 Patient characteristics

Assessment of the difficulty of the surgery

First, we examined the factors that were associated with surgical difficulty in the cohort analysis in this study. We focused on intraoperative blood loss and operative time as indices of surgical difficulty and examined factors related to surgical difficulty in patients who underwent LAR and vLAR. Patients with preoperative CRT, undergoing lateral lymph node dissection, and visceral obesity showed a significantly increased intraoperative blood loss (Table 2). Patients with middle/low rectal tumors, preoperative CRT and lateral lymph node dissection, and visceral obesity demonstrated a significantly prolonged surgical time (Table 3). In particular, visceral obesity was a factor that made surgery more difficult, along with preoperative CRT and lateral lymph node dissection.

Table 2 The association between intraoperative blood loss and clinicopathological factors in cases undergoing LAR/vLAR

Table 3 The association between operative time and clinicopathological factors in cases undergoing LAR/vLAR

Assessment of the risk factor of PONB

A total of 19 patients had PONB (Table 4). The univariate analysis revealed older age (≥ 70 years; P = 0.0057), lower height (≤ 157.6 cm; P = 0.0205), lateral lymph node dissection (P = 0.0131), longer operative time (≥ 480 min; P = 0.0002), intraoperative blood loss of ≥ 150 mL (P = 0.0157), lower VFA (≤ 115 cm²; P = 0.0188), and narrow pelvic inlet (interspinous distance of ≤ 121.3 mm; P = 0.0438) as significantly associated with PONB occurrence. Cases with PONB had significantly lower VFA values than those without PONB (Fig. 2). The logistic regression multivariate analysis revealed that older age (≥ 70 years; P = 0.003, odds ratio [OR] = 8.43, 95% confidence interval [CI]: 2.05–34.7), longer operative time (≥ 480 min; P < 0.001, OR = 19.5, 95% CI: 4.69–81.3), and lower VFA (≤ 115 cm²; P = 0.002, OR = 21.4, 95% CI: 3.24–141) were the independent risk factors of PONB.

Table 4 The association between the occurrence of PONB and clinicopathological factors

Fig. 2

Boxplots of VFA in non-PONB, PONB, and severe PONB groups. The patients with severe PONB were all included in the PONB group

Assessment of the risk factor of severe PONB

A total of 16 patients had severe PONB (Table 5). The univariate analyses revealed that older age (≥ 70 years; P = 0.0352), longer operative time (≥ 480 min; P = 0.0009), intraoperative blood loss of ≥ 150 mL (P = 0.0024), extremely lower VFA (≤ 35cm²; P = 0.0081), narrow pelvic outlet (intertuberous distance of ≤ 89.7 mm; P = 0.0458), and deep pelvis (pelvic depth of ≥ 139.9 mm; P = 0.0469) were significantly associated with the occurrence of severe PONB. Cases with severe PONB had significantly lower VFA than those without PONB, but with no significant difference when compared to cases with mild PONB (Fig. 2). The multivariate analysis revealed that older age (≥ 70 years; P = 0.021, OR = 4.84, 95% CI: 1.26–18.5), longer operative time (≥ 480 min; P = 0.0009, OR = 7.57, 95% CI: 2.12–27.1), extremely lower VFA (≤ 35 cm²; P = 0.0081; OR = 8.80, 95% CI: 2.21–35.0), and deep pelvis (pelvic depth of ≥ 139.9 mm; P = 0.030, OR = 5.01, 95% CI: 1.17–21.4) were the independent risk factors of severe PONB.

Table 5 The association between the occurrence of severe PONB and clinicopathological factors

Assessment of the risk factor of PONB or severe PONB in patients with low VFA or deep pelvis

Finally, we examined which surgical procedures could be risk factors for PONB in cases of low VFA and deep pelvis, which were identified as risk factors for PONB in the present analysis. As a result, no surgical technique was extracted that was significantly associated with PONB development in either low VFA or deep pelvic cases (Supplementary Table 1 and Supplementary Table 2).

Discussion

This study aimed to analyze the risk factors for PONB after rectal cancer surgery focusing on anatomical characteristics. The result revealed low visceral fat with a VFA of < 115 cm² as an independent risk factor for PONB development and very low visceral fat with a VFA of < 35 cm² for severe PONB development. The most exciting finding of this study is that the incidence of PONB was rather more common in patients with less visceral fat despite the greater surgical difficulty in the visceral fat of patients with obesity. This result is contrary to the instinctive impression that many surgeons have that cases with low visceral fat are less prone to nerve damage caused by surgical manipulation because the surgical difficulty is not high.

Many studies have emphasized the importance of measuring pelvic dimensions when planning rectal cancer surgery to assess surgical difficulty. Several previous articles have revealed the association between surgical difficulty and the anatomical factors detected as risk factors of PONB in this study. Escal et al. reported a significant increase in surgical difficulty in cases with shorter intertuberous distances [3]. Yamaoka et al. demonstrated shorter pelvic outlet length, longer sacral depth, and shorter interspinous distance as factors that make the surgery more complex [4]. Sun et al. revealed that shorter interspinous distances could help predict the surgical difficulty of laparoscopic rectal surgery after preoperative CRT [9]. In general, the greater the spatial restriction in the pelvic region, the more frequently the nerves involved in urinary function are injured, thereby increasing PONB. The assessment of surgical difficulty in this cohort identified middle/low rectal tumors, preoperative CRT, and visceral obesity as high-difficulty factors. Any of these factors are considered by many surgeons as variables of high surgical difficulty. However, we extracted not only deep pelvis, which has been reported as a factor for high-difficulty surgery, but also low VFA, as a risk factor for the development of PONB, and the opposite of the results extracted as a factor for high-difficulty surgery in our analysis. Surgical procedures, such as APR and laparoscopic surgery, which had been reported in a previous meta-analysis as risk factors for developing PONB [1], were not extracted as risk factors in this study. Therefore, we focused our analysis on low VFA and deep pelvic cases, which were identified as risk factors for PONB in the current analysis, to remove bias due to VFA and pelvic anatomy. However, surgical procedures were not a risk for developing PONB. This indicates that a low VFA and a deep pelvis are themselves a risk for developing PONB, regardless of the surgical technique. One of the reasons for not identifying the surgical procedure as a risk factor for developing PONB, as previously reported, is the introduction and development of robotic surgery. The advantages of robot-assisted surgery include the following three points [10]. Firstly, the robotic forceps are multijointed and less susceptible to spatial constraints. Secondly, combining magnification effects and high-quality three-dimensional images enables better visualization of intricate anatomical structures. Lastly, suppressing hand tremors provides a stable field of view and precise forceps manipulation. These advantages offered by surgical robotic assistance enable safe and efficient operations even in cases with significant spatial limitations, such as narrow and deep pelvic structures. Kowalewski et al. revealed, in a meta-analysis, a lower incidence of urinary retention after robot-assisted rectal cancer surgery than laparoscopic surgery [11]. Aliyev et al. reported that robot-assisted ISR is associated with better circumferential resection margin, no conversion, and lower postoperative complication rate than laparoscopic ISR [12]. Hence, the characteristics of pelvic anatomy were not detected as the factors determining surgical difficulty. Procedures requiring deep pelvic manipulation, such as APR, were not possibly identified as risk factors for PONB occurrence for the same reasons.

Visceral obesity refers to a type of obesity with high visceral fat content. It is said to be associated with diseases such as diabetes and cerebrovascular disease [13, 14]. Although BMI is widely used as an indicator of obesity, visceral fat has been reported as a reliable indicator for determining rectal cancer surgery difficulty. Watanabe et al. reported that visceral obesity was more highly related to the incidence of overall postoperative complications, including anastomotic leakage, than BMI [15]. Yamaoka et al. revealed the association between high VFA and longer operative time [4]. Heus et al. demonstrated that patients with visceral obesity had more operative blood loss, longer operating time, and more postoperative complications such as surgical site infections [16]. However, these previous reports have not reported on the association between visceral fat and PONB development, which is what gives this study its novelty. Furthermore, the increase in PONB in cases with low visceral fat rather than high visceral fat is contrary to the impression held by many colorectal surgeons. A previous study reported nerve damage as the leading cause of PONB [2]. One possible reason for the increase in nerve damage in low VFA cases is that low visceral fat makes the nerves more susceptible to thermal damage from energy devices. Frequent bleeding from small blood vessels occurs during rectal mesorectal dissection, especially around the neurovascular bundle, leading to the use of energy devices for hemostasis. Ultrasonic coagulating shears and vessel sealing systems have the hemostatic ability, but they have secondary damage to surrounding tissues due to heat diffusion. Recurrent laryngeal nerve preservation is vital in esophageal cancer surgery to prevent vocal cord paralysis. Hirahara et al. recommend sharp maneuvers without electrical current from scissors, particularly in the dissection around recurrent laryngeal lymph nodes [17]. Likewise, using a combination of blunt dissection and cold cutting without an electrical source may be desirable for rectal cancer surgery.

In recent years, studies reporting on the efficacy of neoadjuvant chemotherapy, chemoradiotherapy, and neoadjuvant chemotherapy combined with chemoradiotherapy (total neoadjuvant chemotherapy; TNT) for local control and prognosis improvement of rectal cancer have been increasing. At our institution, preoperative CRT with lateral lymph node irradiation is performed for patients with rectal cancer with cT3 or deeper. Preoperative radiation therapy has been reported as a risk for many complications [18], but to the best of our knowledge, the association between preoperative CRT and PONB has not been reported. The present study revealed no significant association between preoperative CRT and PONB development. Urinary retention can occur in cases of radiation therapy for prostate cancer and uterine cervical cancer when referring to urology or gynecology reports [19, 20]. The main cause of urinary retention in these reports is ureteral and urethral obstructions, which are different from PONB pathogenesis after rectal cancer surgery. To our knowledge, nerve damage related to urinary drainage is not associated with radiation therapy for pelvic organ malignancies. Intraoperative nerve injury is a more likely cause of neurogenic bladder in rectal cancer surgery than irradiation, considering the present analysis. However, careful intraoperative manipulation with attention to nerve injury is necessary in cases after preoperative CRT because preoperative CRT can induce changes, such as tissue edema around the rectum, making it difficult to secure a good surgical field [21].

This study had several limitations. First, this is a retrospective study with a limited scale conducted at a single institution, thereby requiring further research with an increased sample size and studies conducted at other facilities. Second, the presence of a urologist’s diagnosis of the neurogenic bladder was used as an analysis factor because data on residual urine volume in the diagnosis of neurogenic bladder were unavailable. This may be influenced by diagnostic bias. Third, selection bias also existed. The majority of the cohort used in the present analysis was of standard body type, and superobese cases with BMI of > 30 kg/m² were very rare. The incidence of PONB was low in the high VFA group in this analysis, but the surgical difficulty is very high in patients with severe obesity, which may increase the incidence of PONB.

Conclusion

Low VFA is an independent risk factor for the development of PONB in rectal cancer surgery, along with advanced age and prolonged surgery. In particular, low VFA and deep pelvis are independent risk factors for developing severe PONB. In addition to highly difficult surgery, delicate surgical techniques are required in cases of low VFA with attention to PONB development.