Introduction

Urinary (UD) and sexual dysfunction (SD) is a common and serious sequela following total mesorectal excision (TME) for rectal cancer, which can severely impact patients’ quality of life. The rates of UD and SD reach up to 70 and 90%, respectively, with a great variety in range attributed to study design and definition of functional assessment.1–6 Most studies are limited to retrospective analysis, small sample size, missing of baseline data, short-term follow-up, and the use of non-validated instruments. Several predicting factors, for instance, advanced age, tumors located less than 12 cm from the anal verge, radiotherapy, and injury to pelvic autonomic nerves were reported under these circumstances.7–10

The patient desires precise information about the risk of developing such disturbances.11 Only few could recall any specifics regarding their probability. In the context of the complexity of multimodal rectal cancer therapy, a profound knowledge is important for a focused preoperative discussion on functional outcome. Therefore, well-conducted prospective studies concerning long-term dysfunction rates are needed.

This prospective 2-year follow-up study aimed to investigate potential patient-, tumor-, and operation-related risk factors for urogenital dysfunction at specific time intervals after TME in order to rule out their relevance in the long run. The impact of pelvic intraoperative neuromonitoring (pIONM) on functional outcome was analyzed.

Materials and Methods

Patients

Among a consecutive series of 146 prospectively investigated patients undergoing elective TME for primary rectal cancer between January 2008 and August 2014, 85 were included. Of those, 43 underwent pIONM-controlled surgery (within a monocentric clinical trial, “IKONA” ISRCTN06042867).12 There was no randomization. Within this translational research project, the patients needed to undergo a standardized pIONM procedure for accurate data acquisition. Thus, patients were more likely to undergo pIONM if optimum conditions were achieved, such as first operation of the day, performance of total intravenous anesthesia, and no use of a thoracal epidural catheter. Furthermore, the surgical team strived for intraoperative and postoperative support in data analysis by a team of specialized medical engineers and assistants trained in pIONM. For safety reasons, patients with cardiac pacemaker did not undergo pIONM. Excluded were patients undergoing Hartmann’s procedure and those with previous history of open prostatectomy, T4 rectal cancer, postoperative adjuvant chemoradiotherapy (CRT), or missing follow-up on urogenital function. The patients undergoing pIONM within the ongoing prospective randomized controlled multicenter trial “NEUROS” (Clinicaltrials.gov: NCT01585727) were also excluded.

In the present study, all patients underwent standardized nerve-sparing TME with dissection in front of the Denonvilliers’ fascia performed by colorectal surgeons. Those patients with indication for neoadjuvant CRT were treated using 50 Gy in 5 weeks with accompanying chemotherapy followed by surgery 6 to 8 weeks later.

Written informed consent was obtained from all patients. The study was approved by the local Ethics Committee (Ethics Committee of the Medical Association of Rhineland-Palatinate, Germany).

Pelvic Intraoperative Neuromonitoring

The pIONM was performed using the standard methodological setup.12 Repetitive electric stimulations were carried out to map the autonomic nerves at different sites along the pelvic side and above the level of the pelvic floor, ensuring adequate nerve identification and functional verification during the operation. The neuromapping was performed with a hand-guided probe right after posterolateral mesorectal dissection in order to identify the pelvic splanchnic nerves and exposed nervous tissue of the inferior hypogastric plexus. Further stimulations were carried out after lateral and anterolateral dissection. After full mobilization of the rectum, stimulations were additionally performed at the level of the pelvic floor in order to identify nerve branches heading to the internal anal sphincter. Bilateral repetitive stimulations were finally carried out after rectal resection for quality control of the nerve-sparing technique. Currents of 6 mA, frequency of 30 Hz, and monophasic rectangular pulses of 200 μs were used. The electrophysiological stimulation was observed under simultaneous cystomanometry and online processed electromyography of the internal anal sphincter. Signals were continuously visualized on the monitor of the system.

Urogenital Function

Patients were asked to complete the validated International Prostate Symptom Score (IPSS) and Quality of life index (Qol).13 Higher scores indicated lower urinary function and quality of life. New onset of urinary dysfunction was defined by worsening of the IPSS combined with diminished quality of life (minor dysfunction) or the need of long-term catheterization (major dysfunction).

Male sexual function was evaluated using the validated International Index of Erectile Function (IIEF) score, which covers five domains: erectile function, orgasmic function, sexual desire, intercourse satisfaction, and overall satisfaction.14 The questionnaire includes 15 questions, which are each scored from 0 to 5, with lower scores indicating greater dysfunction. The severity of erectile dysfunction (questions 1–5 and 15) can be classified into five categories: no dysfunction (score 26–30), mild (score 22–25), mild to moderate (score 17–21), moderate (score 11–16), and severe dysfunction (score 6–10).15 A reduced score leading to a more severe classification of erectile dysfunction indicated new onset of erectile dysfunction. Female sexual function was evaluated using the Female Sexual Function Index (FSFI), which covers six domains: desire, subjective arousal, lubrication, orgasm, satisfaction, and pain.16 The total score ranges from 2.0 to 36.0 points, with a higher score indicating better sexual function. A score of 26.55 was considered to be the optimal cut-off for differentiating women with and without sexual dysfunction.17 Thus, a postoperative total score that was newly below 26.55 was considered to indicate new onset of female sexual dysfunction with severe dysfunction at values ≤15.

Preoperative urogenital function was compared to the functional outcome at 3 or 6 months after stoma closure (median time interval between TME and stoma closure was 3 months) or at 6 and 9 months after surgery in patients with a permanent stoma. Further follow-ups were performed at 12 and 24 months after surgery.

Statistical Analysis

The data were analyzed using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA). The influence of the predictor variables on the risk of new onset of urogenital dysfunction following surgery was calculated using univariate analysis. Therefore, functional data were transformed into a binary outcome (new onset of dysfunction vs. no new onset of dysfunction). To examine the independent influence of these variables, all variables significantly associated with urogenital dysfunction in the univariate analysis were included in a logistic regression analysis. For comparisons of function between the non-pIONM and pIONM group, the Chi-square test or Mann-Whitney U test was used. Statistical significance was defined as p < 0.05.

Results

Patients

Patients’ demographic, clinical, and histopathological details are shown in Table 1. Baseline data on urogenital function of the non-pIONM and pIONM group are summarized in Table 2.

Table 1 Patients’ characteristics
Full size table
Table 2 Patients’ baseline data of urogenital function
Full size table

No deaths occurred within 30 days following surgery. During the further follow-ups, tumor-related death occurred in two patients with stage IV and in two patients with stage III rectal cancer. Another two patients died of progressive heart failure. Seven patients had a history of pelvic surgery (transurethral resection of the prostate (n = 3); hysterectomy (n = 4)).

Urinary Dysfunction

Of 85 patients, 25 (29%) reported new onset of UD according to the IPSS and Qol after the first follow-up (FU1: 3 months after stoma closure or 6 months after surgery in those with a permanent stoma). Minor UD was reported by 22 patients (26%) and major dysfunction with need for long-term catheterization by 3 (4%) of these. At the second follow-up (FU2: 6 months after stoma closure or 9 months after surgery in those with a permanent stoma), 27 of 84 patients (32%) had newly developed UD, 26 patients (31%) with minor dysfunction, and one (1%) with major dysfunction. At the third follow-up (FU3: 12 months after surgery), 26 of 81 patients (32%) reported new onset of UD with major dysfunction in one (1%) of them. After 2 years (FU4: 24 months after surgery) still 27 of 79 patients (34%) suffered from minor UD and one from major dysfunction.

In the univariate analysis, increased risk for new onset of UD after short-term follow-up was associated with tumor involving the anterior quadrant, tumor location in the middle rectal third, excessive intraoperative blood loss (≥1000 ml), non-performance of pIONM, and a voluminous mesorectum (≥6 cm). At the second follow-up non-performance of pIONM and a voluminous mesorectum remained significant risk factors. At the 1- and 2-year follow-up, neoadjuvant CRT and absence of pIONM were found to significantly increase the risk for new onset of UD (Table 3). In the multivariable logistic regression analysis, all identified factors, except for excessive blood loss and tumor involving the anterior quadrant, remained significant predictors (Table 4). During the follow-ups, the pIONM group had significantly lower rates of newly developed UD than the non-pIONM group (Fig. 1). After the first follow-up, 6 of 43 patients (14%) had newly developed UD in the pIONM group and 23 of 42 (55%) in the non-pIONM group (p = 0.002) and at the second follow-up, 7 of 43 patients (16%) in the pIONM group and 20 of 41 (49%) in the non-pIONM group (p = 0.001). After 1 and 2 years, 6 of 41 (15%) and 8 of 40 patients (20%) undergoing pIONM reported new onset of UD while 20 of 40 (50%) and 20 of 39 patients (51%) undergoing surgery without pIONM had newly developed UD (p = 0.001 and p = 0.004).

Table 3 Univariate analysis with new onset of urinary dysfunction after total mesorectal excision for rectal cancer
Full size table
Table 4 Independent risk factors for postoperative new onset of urinary dysfunction assessed by logistic regression analysis
Full size table
Fig. 1
figure 1

New onset of urinary dysfunction after total mesorectal excision with and without pelvic intraoperative neuromonitoring (pIONM). Urinary functional outcome (minor and major dysfunction) was prospectively assessed at 3 and 6 months after stoma closure or at 6 and 9 months after surgery in patients with a permanent stoma (follow-up (FU) 1 and FU2). Further follow-ups were performed at 12 and 24 months after surgery (FU3 and FU4)

Full size image

Sexual Dysfunction

Of 85 patients, 48 (56%) were sexually active (7 females and 41 males). Of those, 25 (52%) reported new onset of SD after the first follow-up (FU1). Eleven patients (44%) had mild to moderate dysfunction and 14 (56%) severe SD. At the second follow-up (FU2), 30 patients (63%) developed SD. Fifteen of those (50%) reported mild to moderate dysfunction and 15 patients (50%) severe dysfunction. Twelve months after surgery (FU3), 34 of 47 patients (72%) suffered from new onset of SD, 18 patients (53%) with mild to moderate dysfunction, and 16 (47%) with severe dysfunction. After 2 years (FU4), still 33 of the remaining 46 patients (72%) reported of disturbed sexual function with mild to moderate dysfunction in 18 patients (55%) and severe dysfunction in 15 patients (45%).

In the univariate analysis, non-performance of pIONM and a voluminous mesorectum (≥6 cm) were associated with an increased risk for new onset of SD at short-term follow-up. Again at the 1- and 2-year follow-up, neoadjuvant CRT and absence of pIONM were found to significantly increase the risk for SD (Table 5). In the logistic regression analysis, all identified risk factors remained significant predictors (Table 6). At each follow-up, the pIONM group was found to have significantly lower rates of newly developed SD than the non-pIONM group (Fig. 2).

Table 5 Univariate analysis with new onset of sexual dysfunction after total mesorectal excision for rectal cancer
Full size table
Table 6 Independent risk factors for postoperative new onset of sexual dysfunction assessed by logistic regression analysis
Full size table
Fig. 2
figure 2

New onset of sexual dysfunction after total mesorectal excision with and without pelvic intraoperative neuromonitoring (pIONM). Sexual functional outcome (mild to moderate and severe dysfunction) was prospectively assessed at 3 and 6 months after stoma closure or at 6 and 9 months after surgery in patients with a permanent stoma (follow-up (FU) 1 and FU2). Further follow-ups were performed at 12 and 24 months after surgery (FU3 and FU4)

Full size image

After the first follow-up, 10 of 26 patients (38%) had newly developed SD in the pIONM group and 15 of 22 (68%) in the non-pIONM group (p = 0.038). At the second follow-up, 12 of 26 patients (46%) had new onset of SD in the pIONM group and 18 of 22 (82%) in the non-pIONM group (p = 0.011). After 1 and 2 years, 15 of 26 (58%) and 14 of 25 patients (56%) undergoing pIONM reported new onset of SD while in the non-pIONM group 19 of 21 patients (90%) had newly developed SD for both follow-ups (p = 0.013 and p = 0.010).

Discussion

The present prospective study demonstrated that a majority of patients suffered at least from minor/mild to moderate newly developed urogenital dysfunction after TME. New onset of minor UD was reported by 26% and slightly increased to 34% during the further course, while major UD was present in 4% and dropped to 1% after 2 years. New onset of mild to moderate SD occurred in 44% and increased to 55%. Severe SD was present in 56% and decreased to 45%. Studies based on non-validated instruments, however, generally underestimate the dysfunction rates. This is likely to be attributed to the simple format of questions, which do not allow one to distinguish between the varying levels of disturbances. The inconsistency of data is further enhanced by investigating heterogeneous study populations, lack of documentation of preoperative function, and the inclusion of patients who already suffered from severe dysfunction prior to any rectal cancer therapy. In the context of multimodal rectal cancer therapy, the actual results further highlight that even after 2 years a considerable amount of patients still report UD and SD.

In previous studies, several predictors for postoperative urogenital dysfunction were identified on the basis of quantitative surveys. Hendren and colleagues used highly sensitive validated instruments to evaluate sexual function in over 200 rectal cancer patients and demonstrated in a retrospective design that up to approximately 70% of those reported abnormal functional results following treatment.3 According to their multivariate analysis, pelvic radiation was found to have a negative impact on sexual life. This was further supported by the findings of a prospective randomized trial in patients undergoing TME with and without neoadjuvant short-term radiation. On the basis of non-validated questionnaires, the investigators of the Dutch TME Trial revealed preoperative radiotherapy to be an independent predictor for SD.7, 10 The present study demonstrated that the negative effect of preoperative radiation on sexual function is not evident at short course but becomes initially significant 1 year after surgery and remains an independent predictor even after 2 years. According to a previous randomized controlled study, these results might also hold true for patients undergoing preoperative short-term CRT. McLachlan and colleagues compared the effect of neoadjuvant short- and long-term CRT for T3 rectal cancer and showed within a 1-year follow-up that both groups worsened substantially in the postoperative period to the same degree.18

UD after TME was thought to be mainly caused by surgery. The data of the Dutch TME trial demonstrated in a cohort of approximately 800 patients within a 5-year follow-up period that damage to the pelvic autonomic nerves was a significant risk factor for UD, whereas preoperative radiation was not found to be an independent predictor.7 Conversely, Adam and colleagues observed temporary UD in men in relation to neoadjuvant CRT without impact of surgery in a fairly large collective of 169 rectal cancer patients including also T4 cancer and partial mesorectal excision. Their survey was based on validated instruments limited to 1 year.6 The present study further demonstrated, similarly to the results concerning new onset of SD, that pelvic irradiation has a negative impact on UD in the long run, even after 2 years without any significance in the short-term. This is supported by the data of the Stockholm I and II trial, which showed based on non-validated questionnaires that the 65 patients who were randomized to preoperative radiotherapy had significantly worse UD within a mean follow-up of 15 years than those undergoing TME alone.19

The assessment of pelvic autonomic nerve preservation in the Dutch TME trial was carried out based on the surgeons’ intraoperative visual impression. This should be treated with caution and may be a weakness of this study. As stated by the authors themselves, even with accurate neuroanatomical knowledge sparing of this complex neural network is difficult, which suggests that its identification is at least equally difficult. Previous studies based on electrophysiological nerve testing demonstrated that visual assessment of pelvic autonomic nerves is less meaningful.20 A recently published work based on this novel method showed that identification rates of the inferior hypogastric plexus were almost twice as high under electrophysiological assessment compared to those under visual assessment (~80% vs. 45%).21 Thus, early electrophysiological detection of iatrogenic nerve damage may lead to more effective personalized initiation of urologic and proctologic therapies. The pIONM was found to be a reliable tool for quality control of nerve-sparing and could be used for clarification of functional and surgical topography in the minor pelvis in order to guide nerve-sparing.12, 22 A case control study further demonstrated that pIONM is associated with significant lower ano-vesical dysfunctions and a trend towards lower sexual functional disturbances following TME.23 Fang and colleagues reported in a collective of 189 rectal cancer patients that those undergoing rectal excision with electrophysiological nerve testing had significantly lower rates of UD and erectile dysfunction within a follow-up of 6 months after surgery.24 The present study confirmed these results by identifying non-performance of pIONM to be a significant risk factor for UD at any follow-up, even after 2 years. Similarly, nerve-sparing without pIONM was found to be an independent predictor for SD. This is further highlighted by the subgroup analysis, which showed significantly higher rates of newly developed UD and SD in the non-pIONM group (Figs. 1 and 2). In accordance to this, the domains covered by the IIEF and FSFI were also lower.

An expert group of surgeons outlined that surgery is especially challenging in a narrow, deep pelvis and is further complicated in obese males, patients with bulky tumors located less than 12 cm from the anal verge and distorted tissue planes due to neoadjuvant radiotherapy.25 Our multivariate analysis revealed that an excessive volume of mesorectum interferes with the transabdominal nerve-sparing technique by possibly leading to additional narrowness in the minor pelvis. Similarly, a bulky tumor located in the middle rectal third was found to have a negative impact resulting in worse functional outcome. However, these findings do only hold true in the short run.

Based on validated surveys, Duran and colleagues identified in a small collective of 56 rectal cancer patients that either localization of tumor in the middle or lower rectal third is a significant risk factor for new onset of UD, while for SD only the tumor in the distal part was determined to be an effective cause.26 These differences are probably attributed to the small study population and the fact that only patients without neoadjuvant CRT were included for the risk factor analysis. Adam and colleagues further confirmed that a low-lying tumor compared to one in the upper part is a predictive factor for erectile dysfunction, while no significant difference was found for the comparison of tumor location in the middle and lower rectal third.6

The present study is limited to the relatively small sample size, especially with regard to the risk factor analysis. The results of this analysis are of an exploratory nature and thus should be verified in bigger datasets. The strength, however, lies in the prospective design with evaluation of function based on validated instruments at specific time intervals. A selection bias cannot be ruled out, as performance of pIONM was not randomized in this cohort. Furthermore, the patients and the pIONM team were not blinded.

Conclusion

The performance of pIONM during TME was found to be one of the most effective factors reducing the incidence of postoperative UD and SD. When applying this method, the surgeon needs to additionally focus on the detection of pelvic nervous tissue, which probably supports his or her ability to identify nerve branches and thereby preserve function. The negative impact of neoadjuvant CRT is initially evident 1 year after surgery and remains an independent predictor in the long run. Therefore, in order to maintain function, any doctors treating rectal cancer patients should particularly consider the importance of intraoperative protection of the pelvic autonomic nerves and restriction of indication for neoadjuvant CRT where appropriate without compromising oncological result. This is rounded off by better educating patients about posttreatment functional disturbances in order to channel expectations into realistic perspective.