Sexual, urinary, and intestinal dysfunction after rectal surgery: pre-, intra-, and post-operative predictors and trends over time in a single high-volume center

The incidence of long-term complications after rectal surgery varies widely among studies, and data regarding functional sequelae after transanal surgery are lacking. The aim of this study is to describe the incidence and change over time of sexual, urinary, and intestinal dysfunction in a single-center cohort, identifying independent predictors of dysfunction. A retrospective analysis of all rectal resections performed between March 2016 and March 2020 at our institution was conducted. Validated questionnaires were administered to assess post-operative function. Predictors of dysfunction were assessed by univariate and multivariate analysis. Latent class analysis was used to distinguish different risk profile classes. One hundred and forty-five patients were included. Sexual dysfunction at 1 month rose to 37% for both sexes, whereas urinary dysfunction reached 34% in males only. A significant (p < 0.05) improvement in urogenital function was observed between 1 and 6 months only. Intestinal dysfunction increased at 1 month, with no significant improvement between 1 and 12 months. Independent predictors of genitourinary dysfunction were post-operative urinary retention, pelvic collection, and Clavien–Dindo score ≥ III (p < 0.05). Transanal surgery resulted an independent predictor of better function (p < 0.05). Transanal approach, Clavien–Dindo score ≥ III, and anastomotic stenosis were independent predictors of higher LARS scores (p < 0.05). Maximum dysfunction was found at 1 month after surgery. Improvement was earlier for sexual and urinary dysfunction, whereas intestinal dysfunction improved slower and depended on pelvic floor rehabilitation. Transanal approach was protective for urinary and sexual function, although associated with a higher LARS score. Prevention of anastomosis-related complications resulted protective of post-operative function.


Introduction
A multidisciplinary approach to rectal cancer treatment typically yields the best results, especially in terms of recurrence, with 5-year local recurrence rates ranging from 4 to 10% [1][2][3]. TME (total mesorectal excision) as first described by Heald, is recognized as the gold standard of rectal cancer surgery, due to its success in preventing local recurrence and reducing the rate of post-operative dysfunction [4][5][6][7]. Furthermore, TME together with chemoradiotherapy allows for more sphincter-preserving procedures to be carried out as compared to abdominal perineal resection [3,[8][9][10].
Given the improved oncological prognosis afforded by TME and chemoradiotherapy, more attention can be directed toward post-operative function and its impact on quality of life [11]. Bearing in mind too that rectal resections are also performed for benign pathologies such as chronic inflammatory bowel disease or endometriosis, in which patients are often young and with an excellent prognosis, ensuring good post-operative function is of utmost importance [12].
Despite the considerable literature on this topic, data on post-operative function are heterogeneous, and there is a lack of data regarding function following transanal surgery in particular. The aim of this retrospective study is, therefore, to describe the incidence and change over time of sexual, urinary, and intestinal dysfunction after rectal surgery in a single-center population of 145 consecutive patients, identifying independent predictors of dysfunction and risk profiles, so as to optimize patient management.

Population and study design
The study was approved by the hospital ethics committee (Prot. N. 39551) and was written in accordance with "STrengthening the Reporting of Observational Studies in Epidemiology" (STROBE) guidelines.
All patients undergoing rectal resection for malignant or benign pathology between March 2016 and March 2020 at the General Surgery Unit of IRCCS Sacro Cuore Don Calabria in Negrar were included in this study. Exclusion criteria were an age of less than 16 years, local recurrence during follow-up, follow-up < 6 months, previous bladder surgery, death within 90 days of surgery or a lack of response to the administered questionnaires.
Pre-operative demographic and clinical data, surgical and post-operative data were obtained from our prospectively collected database. Surgical procedures were conducted either by laparoscopy or laparotomy, according to the nervesparing technique [6] and included sphincter-preserving resections (with or without colorectal or coloanal anastomosis, and with or without loop ileostomy) and abdominoperineal resections according to Miles. In cases of oncological resection, a high ligation of the inferior mesenteric artery was performed. TME was complete for tumors of the middle or inferior rectum, while in cases of neoplasms of the upper rectum, a partial excision removing 5 cm of mesorectum below the inferior margin of the tumor was performed. In cases of benign pathology, the ligation of the inferior mesenteric artery was below the left colic artery and the resection was close rectal [13]. Anastomosis was arbitrarily considered low, if confection was 5 cm or less from the dentate line, and high if above 5 cm. Complications were classified according to the Clavien-Dindo classification [14].

Questionnaires
Sexual function was assessed in males using the IIEF-15 (International Index of Erectile Function) questionnaire, which investigates five domains of male sexual activity, with scores ranging from 5 (worst score) to 75 (best score); a score of less than 42.5 identifies dysfunction [15,16]. Sexual function in females was assessed with the FSFI (Female Sexual Function Index) questionnaire, which investigates five areas of female sexual activity. Scores can range from 2 (worst score) to 36 (best score), and a score lower than 26.5 is considered pathological [16,17].
Questionnaires were administered before surgery (base line), and postoperatively at 1 month, 3 months, 6 months and, in cases of longer follow-up, at 1 year, as well as at last follow-up, dating back to September 2020. Continence was also assessed at last follow-up using the Wexner questionnaire to evaluate continence for gas, solid stools, and liquid stools, and their impact on daily life. Scores range from 0 (best score) to 10 (worst score), and three degrees can be identified: perfect continence (0-4), moderate incontinence (5)(6)(7)(8)(9), and severe incontinence (> 10) [21].
The EORTC-QLQC30 quality of life assessment questionnaire was administered at last follow-up, with scores ranging from 0 (worst score) to 100 (best score) [22]. Data regarding questionnaires were collected prospectively during outpatient visits, telephone follow-ups or using online questionnaires.

Data analysis plan
Demographic and clinical data were summarized using descriptive statistics, measures of variability and precision, and plots. Statistical tests were used based on the type of variables, test assumptions, and sample dimension. All parameters are reported with 95% confidence intervals (CIs). Statistical models and estimations were adjusted for covariates as necessary.
The three dysfunction rates (sexual, urinary, and intestinal) were calculated as the ratio of number of cases with dysfunction to the number of subjects exposed to the dysfunction. Each rate was reported with a 95% CI. A trend test was used to estimate the trend of the three dysfunction rates at 1, 3, 6, and 12 months after surgery.
Latent class models (LCMs) were used for the LCA. The basis of LCA is that each subject belongs to one of a finite number of classes. In our study, two classes were established for each type of dysfunction (sexual, urinary, and intestinal) that were denoted as being high or low risk of presenting it. Each class has been described by a set of parameters that define the statistical distribution of the outcomes.

3
The conditional probability is that the subject has or does not have the disorder (specificity and sensitivity) and the probability that the condition is present (prevalence). Any observations with missing reference standard results were excluded from the analysis.
Multivariate logistic regression models were used to investigate the presence of potential predictor factors of sexual, urinary, and intestinal dysfunction rate, whereas multivariate linear regression models were used to investigate the presence of potential predictor factors of variation of the sexual, urinary, and intestinal dysfunction score.
STATA vers. 15 (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC) was used for the statistical analysis and a p value of less than 5% was set for statistical significance.

Results
Of the 162 patients that underwent rectal surgery in the considered time period, 150 responded to the administered questionnaires (92.6%). Five other patients were excluded from the study for the following reasons: one died within 90 days of surgery, three experienced local recurrence during followup, and one had a previous radical cystectomy (Fig. 1). The data regarding the final population of 145 patients are summarized in Table 1.

Sexual function
Before surgery, 83.3% of patients reported normal sexual function, while 13.1% patients reported being sexually inactive. The remaining 16.7% presented with pre-operative sexual dysfunction. The rate of sexually inactive patients preoperatively was 8.5% in males and 23.5% in females. The mean pre-operative IIEF score in males was 61.6 (SD 2.0), while in females, the mean pre-operative FSFI score was 30.0 (SD 2.0). Changes in post-operative sexual dysfunction rate were not statistically significant when considering males and females as two separate populations (p = 0.12 and p = 0.10 for males and females, respectively); however, the trend of variation was significant for the study population as a whole (p = 0.03). Moreover, there was a significant decrease in the dysfunction rate between 1 and 6 months for the entire study population (p = 0.02). The peak of dysfunction rate, in our series, was at 1 month after surgery, reaching 38.1% in males and 37.0% in females. The rate of dysfunction before and after surgery was significantly different (p = 0.04), with an overall increase of 9.3% in sexual dysfunction at 12 months after surgery. Graphs of the dysfunction rate at each time point are shown in Fig. 2.
Variation of IIEF-15 and FSFI score over time is shown in Fig. 3a and b. Score variation over time was statistically significant both in males (p = 0.049) and females (p = 0.036). Multivariate analysis revealed independent predictors of a decrease in IIEF-15 score for males at all time points to be: the onset of urinary retention in the immediate post-operative period, and a Clavien-Dindo score ≥ III (p < 0.05 at all time points, see Table 2). Independent predictors of a higher score were an age of less than 70 years, and a low ligation of the inferior mesenteric artery (p < 0.05, see Table 2).
In females, metastatic disease was an independent predictor of a decrease in FSFI score. Among the post-operative factors, the onset of urinary retention, post-operative pelvic collection, reoperation, and anastomotic stenosis were all predictive of a lower FSFI score. No predictors of higher FSFI scores were identified for females. The results of multivariate analysis at each time point are summarized in Table 2 162 Rectal resecƟons  and Table 3. Patients with sexual dysfunction at 1 month postoperatively had a 70% chance of dysfunction which persisted at 6 months.

Urinary function
Before surgery, 93.8% of patients had normal urinary function. Among patients with pre-operative dysfunction, the majority (88.9%) reported only mild dysfunction. The trend of urinary dysfunction rate over time is shown in Fig. 4. Dysfunction rate before and after surgery is significantly different in males, with an increase of 10% at 12 months postoperatively (p = 0.002). The decrease in urinary dysfunction rate was significant in males between 1 and 6 months (p = 0.04), while it was insignificant in females (p = 0.68). Figure 5 shows significant variation of IPSS scores over time for the overall population (p = 0.03). Being male, and the onset of post-operative urinary retention, were independent predictors of a higher IPSS score, while an age of less than 70 years, and having undergone surgery with transanal approach were independent predictors of a lower IPSS score. The results of the multivariate analysis for urinary dysfunction are reported in Table 4. For patients with urinary dysfunction at 1 month, the risk of continued dysfunction at 6 months was 62%.
IMA inferior mesenteric artery, AM anal margin, LOS length of stay *Bibliographic reference [14]

Intestinal function
The rate of patients with LARS-like symptoms in the pre-operative period was 15.2% with 8.3% having severe symptoms. In our study population, half of these patients received a definitive stoma. The post-operative trend of intestinal dysfunction rate was not statistically significant (p = 0.07). As reported in Fig. 6, there was a significant increase (p = 0.005) between the pre-operative and 12-month LARS rate in the whole study population. A transanal surgical approach and the onset of high-grade complications (Clavien-Dindo score ≥ III) were independent predictors   Table 5.
Patients with LARS-like symptoms at 1 month had a 92% chance of symptoms persisting at 6 months. Patients treated with post-operative pelvic floor rehabilitation were 20.0% of the whole population: in this group, the LARS score at all time points was higher by 18, 17, 15, and 10 points at 1, 3, 6, and 12 months compared to patients who did not undergo pelvic floor rehabilitation. However, the decrease in LARS score at each time point among these patients resulted statistically significant (p < 0.001).

Latent class analysis
LCA allowed us to distinguish two classes for each type of dysfunction. Tables 6, 7, and 8 show estimated predicted probabilities with 95% Cis. The considered variables allowed the identification of two phenotypes: one with lower risk of developing dysfunction, denoted as the "low risk" class, and another with characteristics that are associated with a higher risk of developing dysfunction, denoted as the "high risk" class. Regarding sexual dysfunction, the high risk class was defined by surgery with a transabdominal approach and no confection of loop ileostomy. No difference was found regarding complication rates between the two classes. Considering urinary dysfunction, a Clavien-Dindo

Quality of life
Quality of life was assessed at the last follow-up using the EORTC-QLQC30 questionnaire revealing a mean score of 93.25 (SD 8.4) points. After dividing the study population into two groups of patients, one with a follow-up < 12 months and the other with a followup ≥ 12 months, there was no significant difference in quality of life scores apparent between the two groups, respectively, 93.36 (SD 9.85) points for follow-up < 12 months and 93.23 (SD 8.13) points for follow-up ≥ 12 months (p = 0.58). For patients with sexual dysfunction, the mean quality of life score was 90.8 (SD 9.54). Patients with urinary dysfunction alone had a mean score of 88.8 (SD = 10.52), and patients experiencing LARS and incontinence had a mean score of 81 (SD = 7.57). Only five patients reported all three forms of dysfunction simultaneously, having a mean quality of life score of 76.6 (SD 6.95).  Table 5 Predictive factors of LARS score variation in the whole study population (multivariate analysis): pathologies of the higher rectum were associated with a lower LARS score. Transanal approach, anastomotic stenosis, and Clavien-Dindo ≥ III were instead associated with a higher LARS score.

Discussion
Multiple factors can influence sexual, urinary, and intestinal function after rectal surgery. The principal aims of this study were to describe post-operative dysfunction, and to identify independent predictors, to highlight modifiable factors. With the use of LCA, we have grouped patients according to different risk profiles so as to better define therapeutic options.
Our data showed an increase in dysfunction after surgery. The proportion of patients reporting no urogenital dysfunction reduced from 71.0% preoperatively to 58.6% at last follow-up. These data are in line with those reported elsewhere in the literature, where the incidence of post-operative sexual and urinary dysfunction was found to be between 10 and 35% after nerve-sparing surgery [23][24][25]. If we also consider patients with LARS-like symptoms, the percentage of patients without dysfunction drops to 29.6% at last follow-up. These data are also supported by the literature, where the percentage of patients with LARS-like symptoms at 12 months after surgery, or at closure of the protective stoma, reaches approximately 80% [11,26,27].

Sexual and urinary dysfunction
Sexual and urinary dysfunction showed similar trends over time. After a post-operative peak in dysfunction at 1 month, a significant improvement was observed between 1 and 6 months whereby the decrease in dysfunction, as assessed as rate and IIEF-15/FSFI or IPSS score, loses significance. Sexual function certainly has a psychological component, that could partly explain this trend. In the first month after surgery, the proportion of patients that were sexually inactive increased from 13 to 20%, and for females, in particular, sexual inactivity increased from 21 to 35%. Psychological factors potentially influencing sexual behavior could also be supported by the multivariate analysis which found more advanced disease among patients with sexual dysfunction (higher T, metastatic disease) [11,16].
Certain predictors suggest that other factors may contribute to dysfunction, for example, post-operative urinary retention could be an early manifestation of intra-operative nerve damage. In addition, a Clavien-Dindo score ≥ III, the need for reoperation, pelvic collection, and anastomotic stenosis are significant predictors of dysfunction at most time points. In such cases, the effects appear to be a consequence of postoperative complications and could be related to secondary neural damage or distortion of pelvic anatomy. Although anastomotic leak was not found to be a significant variable, these predictors are collateral indicators of anastomotic fistula, and may be a proxy of this complication. Regarding urinary dysfunction, the detrimental role of complications in post-operative function is apparent from LCA. The high risk class featured Clavien-Dindo score ≥ III and pelvic collection among the variables that defined the group.
As has been reported by others, in our series, the presence of ostomy did not correlate significantly with reduced sexual function [16,28,29].
Although the influence of neoadjuvant therapy was not significant in our series, in females at 12 months, adjuvant therapy, and in particular pelvic irradiation, correlated significantly with a decrease in FSFI score. This is in accordance with the literature, which recognizes radiotherapy as a predictor of sexual dysfunction [16,18].
As reported in the literature, the protective effect of youth in males is explained by pre-existing age-related risk factors (prostatic hypertrophy, neural or microvascular subclinical changes) that predispose to both urinary and sexual dysfunction. Microvascular and nerve damage arising from surgery may represent triggers for dysfunction in such cases [12].
Concerning urinary dysfunction, multivariate analysis shows the transanal approach to be a predictor of lower IPSS score. Likewise, the LCA of sexual dysfunction shows that the transanal approach was more frequently undertaken in patients in the low risk class. The superiority of the transanal technique in preserving urinary and sexual function is also confirmed by the literature and is possibly due to more effective magnification of anatomical structures in the difficult to access surgical site of the pelvis [27,30,31].

Intestinal dysfunction and continence
Intestinal dysfunction has a different trend over time compared to urogenital dysfunction. After an initial increase at 1 month after surgery/loop ileostomy closure, the extent of dysfunction plateaus between 1 and 12 months with no significant variation. Multivariate analysis and LCA show transanal surgery to be a significant predictor of a higher LARS score in the high risk group. This is probably related to the greater incidence of low anastomoses, sometimes with partial or total demolition of the internal sphincter, as well as the stretching of the anal sphincter following the insertion of the GelPOINT Path [10,26,30]. Anastomotic leak is also insignificant; however, complications related to anastomotic

3
fistula negatively affect post-operative function, both for intestinal function and continence. Patients with severe dysfunction were referred for pelvic floor rehabilitation, the effects of which were apparent as a significant decrease in mean LARS score at each time point. These findings are consistent with the literature [32].
Intestinal function and continence were found to be the functional aspects with the greatest impact on quality of life, as the mean EORTC-QLQC30 score was lower in patients with LARS-like symptoms and incontinence.
The limitations of this study are mainly due to its retrospective nature; however, standardized questionnaires were used and a relatively short period of time was considered.

Conclusion
In our single-center series of 145 consecutive rectal resections, there is a peak of post-operative dysfunction at 1 month after surgery for sexual, urinary, and intestinal dysfunction. By 6 months after surgery, there was a significant improvement in urogenital dysfunction; however, pre-operative functional levels were not reached. Intestinal dysfunction improvement was slow; however, pelvic floor rehabilitation correlated with significant advances. The transanal technique was protective of urogenital function, allowing a more anatomical dissection at the level of the lower rectum; however, it was related to the onset of LARSlike symptoms and more severe incontinence. Finally, the prevention of anastomosis-related complications appears to be protective in the maintenance of post-operative function.