Impact of Abdominoperineal Resection on Postoperative Male Sexual Function After Minimally Invasive Rectal Cancer Surgery: A Subgroup Analysis From the LANDMARC Study

ABSTRACT

Aim

Using data from the LANDMARC study, we aimed to evaluate the impacts of abdominoperineal resection (APR) on postoperative sexual function in males with rectal cancer undergoing minimally invasive surgery.

Methods

Patients were categorized into the APR (n = 44) and anal‐sparing surgery (AS) groups (n = 329). A detailed subgroup analysis was performed in patients with tumors located < 5 cm from the anal verge (APR, n = 39; AS, n = 59). Sexual function was assessed preoperatively and at 3, 6, and 12 months postoperatively using the International Index of Erectile Function‐5 (IIEF‐5), Erection Hardness Score (EHS), and a customized questionnaire for ejaculatory function. Propensity score–based logistic regression models and longitudinal analyses were applied to address significant baseline imbalances.

Results

In the overall cohort, APR was independently associated with significantly impaired erectile function. Adjusted models showed higher predicted probabilities of erectile dysfunction (ErD) after APR at 3 months (38.7% vs. 16.8) and 12 months (34.7% vs. 13.7%). Longitudinal analyses confirmed increased odds of ErD at 3 months (OR 2.86; 95% CI 1.25–6.53) and 12 months (OR 3.05; 95% CI 1.30–7.14). Potency impairment was also significantly more frequent at 3 months. Ejaculatory dysfunction was more common after APR early postoperatively. In patients with tumors located < 5 cm from the anal verge, APR remained associated with impaired erectile function, whereas ejaculatory function showed partial recovery.

Conclusion

Results suggest that APR is associated with a high risk of postoperative sexual dysfunction, even in minimally invasive surgery. Sexual function preservation is an important consideration when APR is considered.

Abdominoperineal resection (APR) after minimally invasive rectal cancer surgery was associated with significantly higher rates of erectile and early ejaculatory dysfunction compared with anal‐sparing surgery. Although ejaculatory function showed partial recovery over time, erectile dysfunction persisted up to 12 months postoperatively. These findings highlight the need for meticulous nerve preservation and careful surgical decision‐making when considering APR.

1. Introduction

Sexual dysfunction is a major postoperative complication of rectal cancer surgery. This potential complication is important in male patients because postoperative sexual dysfunction has a significant impact on the quality of life [1]. Therefore, reducing its incidence is an important clinical objective when considering operative approaches. The primary cause of sexual dysfunction after rectal surgery is injury to the pelvic autonomic nerves [1, 2]. Careful preservation of the hypogastric and pelvic autonomic nerves is crucial during surgery for rectal cancer. However, the incidence of sexual dysfunction has been reported as 40%–70% [3, 4, 5] and continues to occur at a considerable frequency. Therefore, the establishment of more refined surgical techniques and a deeper understanding of the pelvic anatomy are necessary.

The International Index of Erectile Function‐5 (IIEF‐5) is widely used to provide an overall assessment of sexual function [6]. The index comprises five questions; three of which are based on the assumption that the patient currently engages in sexual activity. Therefore, the IIEF‐5 is not a reliable assessment of patients with limited sexual activity, such as those during the perioperative period or those without a partner. Consequently, these individuals receive a score of zero. In response to these limitations, the current guidelines issued by the American Urological Association, European Association of Urology, and Japanese Society for Sexual Medicine recommend combining the IIEF‐5 and the Erection Hardness Score (EHS) [7, 8, 9]. The EHS assesses erectile hardness and can be applied regardless of sexual activity, potentially offering a more accurate representation of erectile function [10]. However, neither the EHS nor the IIEF‐5 assesses ejaculatory function. For a more comprehensive assessment of sexual function, the IIEF‐5, EHS, and an evaluation of ejaculatory function should be conducted.

The LANDMARC study, conducted in Japan, is a prospective study comparing sexual function outcomes between laparoscopic and robot‐assisted surgeries in patients undergoing rectal cancer surgery [11]. To our knowledge, this is the only prospective study that has combined the assessment of both EHS and ejaculatory function. Although the study results suggested the superiority of robot‐assisted surgery, detailed analyses according to surgical procedure were not performed. Therefore, the effect of each surgical procedure on sexual function remains unclear. Previous reports have indicated a higher incidence of sexual dysfunction after abdominoperineal resection (APR) than that after anal‐sparing surgery [12, 13, 14]. However, most of these studies were conducted during the era in which open surgery was the mainstream. With the widespread use of minimally invasive surgery and an improved understanding of pelvic microanatomy, nerve‐sparing procedures have become increasingly precise. Nevertheless, evidence regarding sexual dysfunction in APR in minimally invasive surgeries remains limited.

Therefore, the present study aimed to reveal the effects of APR on postoperative sexual function using data from the LANDMARC study with a focus on minimally invasive surgical approaches.

2. Methods

2.1. Study Design and Participants

The inclusion and exclusion criteria for the present study have been previously described [11]. Eligible participants were patients aged between 20 and 70 years with pathologically confirmed primary rectal adenocarcinoma in the mid or lower rectum and scheduled to undergo total autonomic nerve‐preserving rectal resection via laparoscopic, robot‐assisted, or transanal approaches. The exclusion criteria included a history of treatment for sexual dysfunction, previous pelvic surgery for malignant disease, clinical stage IV disease, and planned total pelvic exenteration. The study protocol was approved by the Clinical Research Review Board of Yokohama City University (IRB No. B190800044) and the institutional review boards of all participating centers. Written informed consent was obtained from all patients, and the study was registered with the University Hospital Medical Information Network (UMIN000037367).

2.2. Surgical Techniques and Neoadjuvant Therapy

According to the guidelines of the Japanese Society for Cancer of the Colon and Rectum [15], the standard procedure consists of total mesorectal excision with lateral pelvic lymph node dissection, followed by adjuvant chemotherapy. In this study, decisions regarding the administration of neoadjuvant therapy, creation of a diverting stoma, and lateral lymph node dissection were left to the discretion of each participating institution.

2.3. Perioperative Outcomes

Perioperative outcomes were categorized into preoperative, intraoperative, postoperative, and surgeon‐related factors as previously reported. The intraoperative factors included the use of energy devices during rectal dissection. Surgeon‐related variables included the cumulative number of laparoscopic and robotic procedures performed by the operating surgeon and their certification status under the Endoscopic Surgical Skill Qualification System (ESSQS) of the Japan Society for Endoscopic Surgery [16]. Postoperative complications were recorded if they were classified according to the Clavien‐Dindo classification [17] as Grade 2 or higher.

2.4. Assessment of Sexual Function

Sexual function was assessed using the EHS, which evaluates the rigidity of erections on a five‐point scale, and the IIEF‐5, which consists of five questions scored on a five‐point scale for a total score out of 25. Ejaculatory function was assessed using a customized questionnaire based on the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ‐CR38) [11]. This questionnaire asks whether the patient has attempted ejaculation. If so, the patient is asked to assess the degree of difficulty using a four‐point scale as follows: (a) not at all, (b) a little, (c) quite a lot, and (d) very much. Only patients who reported having attempted ejaculation were included in analyses of ejaculatory dysfunction (EjD). The test–retest reliability of the single‐item ejaculatory measure between 3 and 12 months was considered acceptable (Spearman's ρ = 0.7381, p < 0.001).

Changes from baseline were analyzed preoperatively and at 3, 6, and 12 months postoperatively. Post‐treatment responses were considered the baseline for patients who received neoadjuvant therapy. Patients completed the questionnaires during outpatient visits, and responses obtained outside (before or after) a three‐week window were considered invalid.

2.5. Endpoints

The following outcomes were compared among the study cohorts: Erectile dysfunction (ErD), potency impairment, EjD, ejaculatory impairment, IIEF‐5 score change, and ErD per IIEF‐5 change. Regarding the ErD per IIEF‐5 change, as the IIEF‐5 score of 22 points or higher indicates “no erection dysfunction”, we used 21 points as the cutoff value for ErD based on the IIEF‐5. Sexual function parameters were defined as follows, with the primary endpoint being ErD at 12 months postoperatively. Analyses at other time points and of other sexual function parameters were exploratory.

• ErD: Decrease of ≥ 2 points in the EHS score from baseline.

• Potency Impairment: Loss of intercourse‐capable erections, defined as a decline from baseline EHS ≥ 3 to postoperative EHS ≤ 2.

• EjD: Decrease of ≥ 1 point(s) in the ejaculatory score in patients who were able to ejaculate preoperatively.

• Ejaculatory Impairment: Decrease from a baseline ejaculatory status score of 1 to ≥ 2.

• IIEF‐5 score change: Change from baseline score.

• ErD per IIEF‐5 change: Percentage of patients shifting from a sexually active status (IIEF‐5, ≥ 22) to an inactive status (IIEF‐5, ≤ 21).

2.6. Statistical Analysis

Descriptive statistics for continuous variables are presented as means with standard deviations and were compared using a two‐sample t‐test. Categorical variables are presented as proportions and were compared using the chi‐square test. A p of < 0.05 was considered statistically significant, and a p between 0.05 and 0.10 was interpreted as indicating a statistical trend. All statistical analyses were performed using JMP Student Edition 18.2.0 (JMP Statistical Discovery LLC, Cary, NC, USA).

3. Results

3.1. Patient Enrollment

A total of 410 patients from 49 institutions were enrolled between December 2019 and October 2022. After excluding patients who were ineligible (n = 2), had missing preoperative questionnaires (n = 6), were found to have unresectable tumors intraoperatively (n = 1), had a complete response to preoperative therapy (n = 1), or withdrew consent (n = 1), 399 patients were deemed eligible for analysis. For consideration based on surgical procedure, patients who underwent transanal procedures (n = 25) were excluded owing to the presence of only one patient who underwent APR. Patients who underwent total proctocolectomy were also excluded, owing to only one patient having the procedure. The included patients were categorized into two groups: APR (n = 44) and anal‐sparing surgery (AS) (n = 329) (Figure 1).

FIGURE 1.

Patient flow diagram. APR, abdominoperineal resection group; AS, anal‐sparing surgery group; AV, anal verge; Ta, transanal laparoscopic surgery; Lap‐TPC, laparoscopic total proctocolectomy.

3.2. Questionnaire Response Rate

The questionnaire response rate for patients matching the inclusion criteria was 100% preoperatively, 92.0% at 3 months postoperatively (18 uncollected and 12 invalid), 93.3% at 6 months postoperatively (18 uncollected and seven invalid), and 90.9% at 12 months postoperatively (29 uncollected and five invalid). More detailed questionnaire response rates are provided in Supporting Information Table S1.

3.3. Patient Characteristics, Operative Findings, and Pathological Outcomes

A summary of the characteristics, operative findings, and pathological outcomes is presented in Table 1, and the full dataset is provided in Supporting Information Table S2. Baseline imbalances were observed between the groups owing to inherent differences in the surgical procedures. The APR group had tumors significantly closer to the anal verge (AV) than the AS group. Regarding the clinical T (cT) factor, 84.1% of the APR group had cT3–4 tumors compared with 53.5% in the AS group. Consequently, the rate of positive circumferential resection margins was significantly higher in the APR group. Regarding the clinical N (cN) factor, 22.7% of the patients in the APR group had cN3 disease, which was associated with a significantly higher frequency of lateral pelvic lymph node dissection. Moreover, the APR group had significantly longer operative times and greater intraoperative blood loss owing to both surgical and oncological factors.

TABLE 1.

Patient characteristics, operative findings, and pathological outcomes of the entire cohort.

	APR (n = 44)	AS (n = 329)	p
Age (years)	58 ± 8.6	58 ± 52	0.9945
BMI (kg/m2)	24.1 ± 3.5	23.6 ± 21.3	0.3016
ASA			0.7055
1	30 (68.2%)	204 (62%)
2	13 (29.5%)	118 (35.9%)
3	1 (2.3%)	7 (2.1%)
Serum CEA (ng/ml)	8.1 ± 11.1	6.1 ± 12.7	0.3266
Serum Albumin (g/dl)	4.2 ± 0.3	4.2 ± 0.6	0.8935
Serum CRP (mg/dl)	0.2 ± 0.3	0.3 ± 0.8	0.5458
Lower Location of Tumor			< 0.0001
Middle rectum	1 (2.3%)	112 (34%)
Low rectum	27 (61.4%)	212 (64.4%)
Anal canal	16 (36.4%)	5 (1.5%)
Tumor distance from AV (cm)	2.7 ± 1.2	6.9 ± 2.7	< 0.0001
c (yc) T			0.0001
1	2 (4.5%)	66 (20.1%)
2	5 (11.4%)	87 (26.4%)
3	30 (68.2%)	162 (49.2%)
4	7 (15.9%)	14 (4.3%)
c (yc) N			0.0024
0	24 (54.5%)	199 (60.5%)
1	8 (18.2%)	84 (25.5%)
2	2 (4.5%)	30 (9.1%)
3	10 (22.7%)	16 (4.9%)
Neoadjuvant therapy			0.0859
None	30 (68.2%)	279 (84.8%)
RT + CT	1 (2.3%)	3 (0.9%)
RT only	4 (9.1%)	13 (4%)
CT only	9 (20.5%)	34 (10.3%)
Surgical approach			0.3511
Laparoscopy	22 (50%)	140 (42.6%)
Robot	22 (50%)	189 (57.4%)
Operative time (min)	469 ± 146	348 ± 130	< 0.0001
Blood loss (ml)	516 ± 1407	43 ± 78	< 0.0001
Lateral lymph node dissection	21 (47.7%)	88 (26.7%)	0.0575
Conversion	5 (11.4%)	0 (0%)	< 0.0001
Autonomic nerve preservation			0.0829
Complete (AN4)	40 (90.9%)	319 (97%)
Incomplete (AN0–3)	4 (9.1%)	10 (3%)
Energy Device			0.5927
Electrocautery	39 (88.6%)	301 (91.5%)
Ultrasonically activated device	5 (11.4%)	26 (7.9%)
Vessel sealer	0 (0%)	2 (0.6%)
Certification status of main surgeon (ESSQS)			0.64
Yes	39 (88.6%)	299 (90.9%)
No	5 (11.4%)	30 (9.1%)
Previous experience of same procedure			0.5204
≤ 20	8 (18.2%)	50 (15.2%)
≥ 21, ≤ 40	10 (22.7%)	49 (14.9%)
≥ 41, ≤ 60	5 (11.4%)	48 (14.6%)
≥ 61, ≤ 80	1 (2.3%)	26 (7.9%)
≥ 81, ≤ 100	2 (4.5%)	13 (4%)
≥ 101	18 (40.9%)	143 (43.5%)
Postoperative complication
Total	15 (34.1%)	90 (27.4%)	0.3588
Anastomotic leakage	0 (0%)	30 (9.1%)	NE
Ileus	2 (4.5%)	28 (8.5%)	0.3301
Urinary retention	4 (9.1%)	11 (3.3%)	0.1069
Wound infection	2 (4.5%)	4 (1.2%)	0.1615
Abdominal abscess	4 (9.1%)	14 (4.3%)	0.2002
Hemorrhage	0 (0%)	4 (1.2%)	0.3149
Pneumonia	1 (2.3%)	1 (0.3%)	0.1836
Urinary tract infection	2 (4.5%)	1 (0.3%)	0.0246
Tumor diameter (mm)	48.5 ± 24.1	37.7 ± 20.5	0.0017
Harvested lymph nodes	25.4 ± 16.2	22.6 ± 14	0.2308
RM			< 0.0001
RM0	37 (84.1%)	320 (97.3%)
RM1	5 (11.4%)	1 (0.3%)
RMX	2 (4.5%)	7 (2.1%)
Adjuvant therapy	23 (52.3%)	101 (30.7%)	0.0055

Note: Categorical data are presented as n (%), and continuous variables are presented as mean and standard deviation.

Abbreviations: Alb, albumin; AN, preservation of autonomic nerves; APR, abdominoperineal resection group; AS, anal‐sparing surgery group; AV, anal verge; BMI, body mass index; CEA, carcinoembryonic antigen; CRP, C‐reactive protein; CT, chemotherapy; ESSQS, Endoscopic Surgical Skill Qualification System; NE, not estimable; RM, radial margin; RT, radiation therapy.

3.4. Unadjusted Results

Unadjusted results showed significantly higher rates of ErD in the APR group at 3, 6, and 12 months postoperatively than those in the AS group (38.9% vs. 15.9%, 32.5% vs. 16.8%, and 35.1% vs. 13.0%), with no improvement over time in either group (Figure 2A). Similarly, potency impairment was more frequent after APR throughout follow‐up (50.0% vs. 21.3%, 41.4% vs. 23.4%, and 39.3% vs. 18.5%; Figure 2B). EjD occurred more often after APR at 3 and 6 months (69.2% vs. 35.8% and 57.1% vs. 26.3%) but showed partial recovery with no significant difference at 12 months (Figure 2C,D). For ErD per IIEF‐5 change, dysfunction affected approximately 60% of the AS group and over 80% of the APR group, reaching 100% at 3 months in the APR group (Figure 2E). The mean decreases in the IIEF‐5 scores were approximately 4 and 6 points in the AS and APR groups, respectively (Figure 2F).

FIGURE 2.

Unadjusted analysis of total patients matching the inclusion criteria. (A) The incidence rates of ErD. The APR group showed a significantly higher incidence of ErD than the AS group at all time points. Nearly no improvements were observed in either group over time. (B) The incidence rates of potency impairment. Potency impairment showed trends similar to ErD. (C) The incidence rates of EjD. The APR group had significantly higher incidences of EjD than the AS group at 3 and 6 months postoperatively. A trend toward recovery was observed in the APR group, and at 12 months postoperatively, there was no difference between the two groups. (D) The incidence rates of Ej impairment. Ej impairment showed trends similar to EjD. (E) The incidence rates of ErD per IIEF‐5 change. The incidence of ErD per IIEF‐5 score change was approximately 80% in the APR group and approximately 60% in the AS group during the study period. The incidence rate at 3 months postoperatively was significantly higher in the APR group than in the AS group. (F) Change from baseline IIEF‐5 score. The IIEF‐5 score decreased by approximately 6 points in the APR group and by approximately 4 points in the AS group during all time points, and no recovery was observed over time in either group.

APR, abdominoperineal resection group; AS, anal‐sparing surgery group; Ej, ejaculatory; EjD, ejaculatory dysfunction; ErD, erectile dysfunction; IIEF‐5, international index of erectile function‐5.

3.5. Adjusted Analysis Using Propensity Score and Longitudinal Models

Because of substantial baseline imbalances between the groups, propensity score (PS) matching was initially attempted. However, this resulted in a marked reduction in analyzable cases; therefore, PS matching was deemed infeasible. We subsequently considered conventional multivariable regression; however, the APR cohort was small (n = 44); therefore, direct adjustment using all candidate covariates was considered inappropriate owing to a low events‐per‐variable ratio and the associated risk of overfitting. Accordingly, to achieve confounding control with minimal degrees of freedom, we adopted a PS–based strategy. The PS was estimated using logistic regression, including the following covariates: Distance from the AV, cT stage, cN stage, neoadjuvant therapy, lateral pelvic lymph node dissection, blood loss, operative time, conversion, resection margin, autonomic nerve preservation, and surgeon certification status under the ESSQS. The ESSQS was used as a proxy for center and surgeon effects.

The PS and distance from the AV were incorporated into PS‐adjusted logistic regression models for all cross‐sectional analyses. Longitudinal models were additionally analyzed using generalized estimating equations with patient‐level clustering.

3.6. Erectile Dysfunction

The adjusted logistic analyses demonstrated consistently higher rates of ErD in the APR group (Figure 3A). Specifically, the probability of ErD was significantly higher in the APR group at 3 months (38.7% vs. 16.8%, p = 0.0498) and at 12 months (34.7% vs. 13.7%, p = 0.0409). In longitudinal analyses, trends were similar to those observed in adjusted logistic analyses. APR was independently associated with higher odds of ErD at both 3 months (odds ratio [OR] 2.86, 95% confidence interval [CI] 1.25–6.53; p = 0.015) and 12 months (OR 3.05, 95% CI 1.30–7.14; p = 0.013) (Table 2).

FIGURE 3.

Adjusted predicted probabilities in the overall cohort (PS‐adjusted logistic regression). (A–E) The adjusted predicted probabilities of each outcome were estimated using PS‐adjusted logistic regression models. Error bars indicate SDs. (F) Adjusted change from baseline in the IIEF‐5 score. Error bars indicate standard SEs. (A) Erectile dysfunction (ErD). The APR group showed a trend toward consistently higher adjusted ErD incidence rates compared with the AS group at all postoperative time points. (B) Potency impairment. The APR group showed a significantly higher adjusted predicted probability than the AS group at 3 months postoperatively. Differences were observed at 6 and 12 months postoperatively; however, these were not statistically significant. (C) Ejaculatory dysfunction (EjD). The APR group showed a higher tendency for adjusted predicted probability than the AS group at 3 months postoperatively. The probability decreased over time. (D) Ejaculatory impairment. Ejaculatory impairment showed trends similar to those observed for EjD. (E) ErD per IIEF‐5 change. The adjusted predicted probability of ErD per IIEF‐5 changes was approximately 80% in the APR group and 60% in the AS group throughout the study period. (F) Change from baseline IIEF‐5 score. Throughout the study period, both groups showed a decrease of approximately 4 points from baseline without substantial recovery.

APR, abdominoperineal resection group; AS, anal‐sparing surgery group; Ej, ejaculatory; EjD, ejaculatory dysfunction; ErD, erectile dysfunction; IIEF‐5, International Index of Erectile Function–5; PS, propensity score; SD, standard deviation; SE, standard error.

TABLE 2.

Adjusted odds ratio of the entire cohort.

Variable	Adjusted OR	(95% CI)	p
APR vs. AS
ErD
3 months	2.86	(1.25–6.53)	0.0152
6 months	2.01	(0.88–4.58)	0.105
12 months	3.05	(1.3–7.14)	0.0131
Potency impairment
3 months	2.64	(1.09–6.38)	0.0328
6 months	1.61	(0.66–3.94)	0.3012
12 months	1.95	(0.77–4.92)	0.1685
EjD
3 months	2.86	(0.8–10.21)	0.0936
6 months	2.52	(0.78–8.11)	0.1191
12 months	1.37	(0.38–4.93)	0.6346
Ej impairment
3 months	3.55	(0.68–18.45)	0.1016
6 months	2.06	(0.57–7.48)	0.2665
12 months	1.1	(0.3–4.06)	0.8885
ErD per IIEF‐5
3 months	> 1 × 109	(0–NE)	0.008
6 months	2.83	(0.25–31.52)	0.3589
12 months	3.32	(0.3–36.92)	0.285

Abbreviations: APR, abdominoperineal resection group; AS, anal‐sparing surgery group; CI, confidence interval; Ej, ejaculatory; EjD, ejaculatory dysfunction; ErD, erectile dysfunction; IIEF‐5, International Index of Erectile Function–5; NE, not estimable; OR, odds ratio.

In potency impairment, both the adjusted predictive probability model and the longitudinal model showed a significantly higher incidence rate in the APR group at 3 months than that in the AS group (49.0% vs. 21.6%, p = 0.0304 and OR 2.64, 95% CI 1.09–6.38, p = 0.033); however, no significant difference was observed at 12 months (Figure 3B, Table 2).

3.7. Ejaculatory Dysfunction

A trend toward a higher probability of EjD was observed in the APR group at 3 months (66.3% vs. 36.3%, p = 0.062) (Figure 3C). In the longitudinal analysis, a similar trend was observed at 3 months (OR 2.86, 95% CI 0.80–10.21; p = 0.094) (Table 2). However, a trend toward recovery was observed in the APR group, and the difference was no longer significant at 12 months postoperatively. A similar trend was observed for ejaculatory impairment (Figure 3D, Table 2).

3.8. IIEF‐5

Regarding ErD per IIEF‐5 assessment, the adjusted predicted probability throughout the study period was approximately 60% in the AS group and over 80% in the APR group. Notably, the incidence at 3 months postoperatively was 100% in the APR group, which was significantly higher than the 54.4% observed in the AS group (Figure 3E). The longitudinal model also demonstrated significantly increased odds of ErD per IIEF‐5 at 3 months (OR > 1 × 10 [9], p = 0.008) (Table 2). The IIEF‐5 scores decreased by approximately 4 points in both groups, with no recovery trend over time in either group (Figure 3F).

3.9. Detailed Analysis: Tumors < 5 Cm From the AV

For a more detailed analysis, we focused on cases close to the AV. The analysis was conducted on patients with tumors located < 5 cm from the AV for clinical and methodological reasons. Tumors at this level are clinically considered for the indication of APR. In addition, most APR cases in this cohort were clustered within this range. Applying spline modeling would have required extrapolation beyond the observed data, thereby producing unstable estimates. Therefore, a clinically meaningful cutoff (< 5 cm) was used without spline‐based modeling. The patients were categorized into the APR (n = 39) and AS (n = 59) groups.

The analysis was performed using a PS‐adjusted logistic regression model and longitudinal model, as described previously.

3.10. Patient Characteristics, Operative Findings, and Pathological Outcomes

The baseline characteristics, operative findings, and pathological outcomes of this subgroup are shown in Table 3, with detailed data provided in Supporting Information Table S3. As in the full cohort, some baseline imbalances were noted. Although less pronounced, the tumors in the APR group remained significantly closer to the AV. The cT factor profile was similar to that of the full cohort. Notably, in contrast to the previous analysis, the AS group had a higher rate of cN3 (8.5%), eliminating the previous differences in the rate of lateral lymph node dissection. Operative time and blood loss remained significantly higher in the APR group than in the AS group.

TABLE 3.

Patient characteristics, operative findings, and pathological outcomes of the AV < 5 cm cohort.

	APR (n = 39)	AS (n = 59)	p
Age (years)	57.8 ± 8.8	56.3 ± 7.6	0.3794
BMI (kg/m2)	24.1 ± 3.5	23.3 ± 3	0.2302
ASA			0.5054
1	27 (69.2%)	37 (62.7%)
2	12 (30.8%)	22 (37.3%)
3	0 (0%)	0 (0%)
Serum CEA (ng/ml)	7.8 ± 11.3	8.3 ± 17.6	0.8715
Serum Albumin (g/dl)	4.2 ± 0.3	4.3 ± 0.4	0.2619
Serum CRP (mg/dl)	0.2 ± 0.3	0.2 ± 0.3	0.6223
Lower Location of Tumor			0.0001
Middle rectum	1 (2.6%)	4 (6.8%)
Low rectum	36 (92.3%)	55 (93.2%)
Anal canal	2 (5.1%)	0 (0%)
Tumor distance from AV (cm)	2.5 ± 1.1	3.7 ± 0.7	< 0.0001
c (yc) T			< 0.0001
1	1 (2.6%)	16 (27.1%)
2	4 (10.3%)	18 (30.5%)
3	28 (71.8%)	24 (40.7%)
4	6 (15.4%)	1 (1.7%)
c (yc) N			0.1069
0	20 (51.3%)	42 (71.2%)
1	7 (17.9%)	10 (16.9%)
2	2 (5.1%)	2 (3.4%)
3	10 (25.6%)	5 (8.5%)
Neoadjuvant therapy			0.334
None	25 (64.1%)	47 (79.7%)
RT + CT	1 (2.6%)	1 (1.7%)
RT only	4 (10.3%)	2 (3.4%)
CT only	9 (23.1%)	9 (15.3%)
Surgical approach			0.8993
Laparoscopy	18 (46.2%)	28 (47.5%)
Robot	21 (53.8%)	31 (52.5%)
Operative time (min)	455 ± 140	385 ± 130	0.0126
Blood loss (ml)	529 ± 1495	67 ± 117	0.0199
Lateral lymph node dissection	19 (48.7%)	22 (37.3%)	0.2621
Conversion	4 (10.3%)	0 (0%)	0.0057
Autonomic nerve preservation			0.8642
Complete (AN4)	36 (92.3%)	55 (93.2%)
Incomplete (AN0–3)	3 (7.7%)	4 (6.8%)
Energy Device			0.6018
Electrocautery	36 (92.3%)	56 (94.9%)
Ultrasonically activated device	3 (7.7%)	3 (5.1%)
Vessel sealer	0 (0%)	0 (0%)
Certification status of main surgeon (ESSQS)			0.8642
Yes	36 (92.3%)	55 (93.2%)
No	3 (7.7%)	4 (6.8%)
Previous experience of same procedure			0.3029
≤ 20	7 (17.9%)	8 (13.6%)
≥ 21, ≤ 40	7 (17.9%)	5 (8.5%)
≥ 41, ≤ 60	5 (12.8%)	8 (13.6%)
≥ 61, ≤ 80	1 (2.6%)	7 (11.9%)
≥ 81, ≤ 100	2 (5.1%)	1 (1.7%)
≥ 101	17 (43.6%)	30 (50.8%)
Postoperative complication
Total	12 (30.8%)	17 (28.8%)	0.8357
Anastomotic leakage	0 (0%)	4 (6.8%)	NE
Ileus	0 (0%)	4 (6.8%)	0.0411
Urinary retention	3 (7.7%)	2 (3.4%)	0.3499
Wound infection	1 (2.6%)	1 (1.7%)	0.7682
Abdominal abscess	4 (10.3%)	2 (3.4%)	0.1702
Hemorrhage	0 (0%)	0 (0%)	NE
Pneumonia	1 (2.6%)	0 (0%)	0.1728
Urinary tract infection	2 (5.1%)	0 (0%)	0.0528
Tumor diameter (mm)	47 ± 24.5	35.5 ± 19	0.0137
Harvested lymph nodes	24.4 ± 14.4	21.6 ± 12.6	0.3157
RM			0.1672
RM0	33 (84.6%)	55 (93.2%)
RM1	4 (10.3%)	1 (1.7%)
RMX	2 (5.1%)	3 (5.1%)
Adjuvant therapy	17 (43.6%)	30 (50.8%)	0.0003

Note: Categorical data are presented as n (%), and continuous variables are presented as mean and standard deviation.

Abbreviations: Alb, albumin; AN, preservation of autonomic nerves; APR, abdominoperineal resection group; AS, anal‐sparing surgery group; AV, anal verge; BMI, body mass index; CEA, carcinoembryonic antigen; CRP, C‐reactive protein; CT, chemotherapy; ESSQS, Endoscopic Surgical Skill Qualification System; NE, not estimable; RM, radial margin; RT, radiation therapy.

3.11. Erectile Dysfunction

The erectile function outcomes after APR were consistently poor overall, with continuing differences compared with the AS group. Neither group showed a clear or substantial pattern of recovery over time. The adjusted predicted probability of ErD was significantly higher after APR at 3 months (40.8% vs. 16.7%, p = 0.0122) and showed a trend toward a higher probability at both 6 months (31.1% vs. 13.3%, p = 0.0729) and 12 months (37.3% vs. 16.9%, p = 0.0638) (Figure 4A). The adjusted predicted probability of potency impairment was also significantly higher after APR at 3 months (44.9% vs. 20.1%, p = 0.0155) (Figure 4B).

FIGURE 4.

Adjusted predicted probabilities in patients with tumor distance from the AV < 5 cm (PS‐adjusted logistic regression). (A–E) The adjusted predicted probabilities for each outcome in patients with AV < 5 cm. Error bars indicate SDs. (F) The adjusted change from baseline in the IIEF‐5 score. Error bars indicate SEs. (A) Erectile dysfunction (ErD). The APR group showed a trend toward consistently higher adjusted ErD incidence rates compared with the AS group at all postoperative time points. (B) Potency impairment. The APR group showed a significantly higher adjusted predicted probability than the AS group at 3 months postoperatively. The difference in probability decreased over time. (C) Ejaculatory dysfunction (EjD). Higher adjusted predicted probabilities were observed in the APR group at early postoperative time points, with a gradual decrease over time. The difference diminished by 12 months. (D) Ejaculatory impairment. Ejaculatory impairment showed trends similar to those observed for EjD. (E) ErD per IIEF‐5 change. Throughout the study period, the predicted probability of ErD per IIEF‐5 change was approximately 80% or higher in the APR group. The predicted probability at 3 months postoperatively was significantly higher in the APR group than in the AS group. (F) Change from baseline IIEF‐5 score. Throughout the study period, no significant difference was observed between the two groups. Both groups showed decreases from baseline without notable recovery trends.

APR, abdominoperineal resection group; AS, anal‐sparing surgery group; AV, anal verge; Ej, ejaculatory; EjD, ejaculatory dysfunction; ErD, erectile dysfunction; IIEF‐5, International Index of Erectile Function–5; PS, propensity score; SD, standard deviation; SE, standard error.

Consistent with these patterns, the longitudinal analysis demonstrated significantly increased odds of ErD after APR at 3 months (OR 2.86, 95% CI 1.25–6.53; p = 0.015) and at 12 months (OR 3.05, 95% CI 1.30–7.14; p = 0.013). Similarly, potency impairment showed a significant association with APR at 3 months (OR 2.64, 95% CI 1.09–6.38; p = 0.033) (Table 4).

TABLE 4.

Adjusted odds ratio of the AV < 5 cm cohort.

Variable	Adjusted OR	(95% CI)	p
APR vs. AS
ErD
3 months	4.19	(1.41–12.42)	0.0088
6 months	3.44	(1.09–10.87)	0.0325
12 months	3.39	(1.12–10.21)	0.0289
Potency impairment
3 months	4.18	(1.36–12.87)	0.0115
6 months	2.18	(0.71–6.69)	0.1724
12 months	1.77	(0.57–5.49)	0.3252
EjD
3 months	5.55	(1.15–26.81)	0.0277
6 months	2.16	(0.49–9.55)	0.3098
12 months	1.43	(0.29–7.07)	0.6662
Ej impairment
3 months	10.31	(1.4–76.16)	0.0132
6 months	3.29	(0.58–18.61)	0.1728
12 months	2.01	(0.38–10.77)	0.4157
ErD per IIEF‐5
3 months	> 1 × 109	(0–NE)	0.0023
6 months	3.93	(0.22–71.38)	0.3269
12 months	9.79	(0.55–172.88)	0.0828

Abbreviations: APR, abdominoperineal resection group; AS, anal‐sparing surgery group; AV, anal verge; CI, confidence interval; Ej, ejaculatory; EjD, ejaculatory dysfunction; ErD, erectile dysfunction; IIEF‐5, International Index of Erectile Function–5; NE, not estimable; OR, odds ratio.

3.12. Ejaculatory Dysfunction

The ejaculatory function showed a trend of worsening in the early postoperative period following APR but improvement over time thereafter. At 3 months, the adjusted predicted probability of EjD was significantly higher after APR (53.7% vs. 24.3%, p = 0.0208) (Figure 4C). Ejaculatory impairment also showed a significantly higher probability at 3 months (64.8% vs. 28.2%, p = 0.0265), with a trend persisting at 6 months (49.7% vs. 32.3%, p = 0.0758). However, the difference was not significant at 12 months postoperatively (Figure 4D).

In the longitudinal analysis, a trend toward increased odds of EjD after APR was observed at 3 months (OR 2.86, 95% CI 0.80–10.21; p = 0.094). Thereafter, the difference decreased and was no longer observed (Table 4).

3.13. IIEF‐5

For the ErD per IIEF‐5 assessment, the adjusted predicted probability during the study period was more than 80% in the APR group. Conversely, this probability did not reach 80% in the AS group. The adjusted predicted probability was significantly higher after APR at 3 months (100.0% vs. 41.1%, p = 0.0085), after which no significant or trend‐level differences were observed. Similarly, the longitudinal model demonstrated a marked increase in odds at 3 months (OR > 1 × 10 [9]; p = 0.008) (Figure 4E, Table 4).

The change of the IIEF‐5 scores did not show a statistically significant difference at any time point. (Figure 4F).

4. Discussion

This study demonstrated that APR was associated with significantly higher risks of ErD and early postoperative EjD compared with those in AS, even after adjusting for baseline imbalances using a PS‐based approach. Most previous reports comparing anorectal preservation surgery to APR have investigated open surgeries [12, 13, 14]. The occurrence of postoperative sexual dysfunction is associated with pelvic nerve damage, particularly nerve injury on the pelvic floor [18]. Despite the expansion of laparoscopic and robot‐assisted surgery, enabling more precise nerve preservation, APR remains associated with an increased risk of sexual dysfunction.

In the present study, we evaluated the effect of APR on sexual function in the LANDMARC study cohort. The results demonstrated that the incidences of ErD and early postoperative EjD were significantly higher in the APR group than those in the AS group. The initial analysis included patients with upper rectal cancer who were not candidates for APR. Therefore, a more detailed analysis was performed, restricting the cohort to patients with tumors located closer to the AV, which are more likely to be treated with APR. This definition closely matches previous reports evaluating sexual dysfunction in APR [19, 20, 21]. Consistent with the primary analysis, the APR was associated with significantly higher rates of ErD and early postoperative EjD, suggesting that the APR procedure itself may contribute to postoperative sexual dysfunction.

Erectile function is strongly dependent on the cavernous nerve [1, 2], which branches off the pelvic plexus, joins the neurovascular bundles, and runs along the lateral side of the prostate, leading to the cavernous body. Recent anatomical studies have shown that the cavernous nerves also form a hammock‐like distribution on the dorsal aspect of the prostate, running in close proximity to the Denonvilliers' fascia (rectoprostatic fascia) [22, 23, 24]. Reports have shown that preservation of the Denonvilliers' fascia may affect the postoperative sexual function [25]. Moreover, the cavernous nerve passes through the rectourethral muscle into the cavernous body [24, 26, 27, 28]. In minimally invasive surgery, the magnified view enables clear identification and preservation of the Denonvilliers' fascia, and precise dissection of the rectourethral muscle can be performed safely without injuring the urethra or rectum. In contrast, during an APR, the pelvic floor is approached only from the perineal side under direct vision, which creates difficulty in recognizing the detailed structure. Therefore, confirming the degree of preservation of the Denonvilliers' fascia may be challenging for surgeons. Additionally, landmarks for the rectourethral muscle are not as distinct in the perineal approach, complicating the identification and division of the border between the rectourethral muscle and rectum. Although evaluation of the extent of pelvic dissection using the abdominal approach and investigation of the details of the surgical procedure on the perineal approach were not included in this study, the significantly higher incidence of ErD in the APR group suggests that the preservation of the cavernous nerves during the perineal dissection may have been insufficient. In addition, the APR group showed significantly greater blood loss, suggesting the presence of neurovascular bundles and posterior prostate injury. Increased bleeding may have impaired visibility and resulted in poor nerve preservation. Furthermore, the baseline differences between the groups in cT3/T4 stage, tumor size, distance from AV, and lateral pelvic lymph node dissection may have contributed to EjD. In previous reports, the cT3/T4 stage, tumor size, and lateral pelvic lymph node dissection were not considered as risk factors, whereas distance from the AV was suggested as a risk factor [29, 30]. Therefore, we adjusted for these factors using a PS‐based strategy. Nevertheless, APR remained independently associated with erectile dysfunction, suggesting that this result may be associated with the surgical procedure.

Ejaculatory function is mediated by the dorsal penile, pudendal, hypogastric, and lumbar splanchnic nerves [2]. Both low anterior resection and APR involve similar dissection of the hypogastric and lumbar splanchnic nerves using the abdominal approach; therefore, the difference in EjD can be attributed to the management of the dorsal penile and pudendal nerves. The pudendal nerve arises from the pelvic plexus and branches into the dorsal penile, perineal, and inferior rectal nerves [1]. Although the dorsal penile nerve is not directly manipulated when the levator ani muscle is dissected during APR, the central side of the pudendal nerve may be damaged by traction or electrocoagulation [1, 18]. This may cause ejaculation disorders due to damage to the dorsal penile nerve via the pudendal nerve [1], thereby contributing to transient EjD observed in the early postoperative period. The tendency for improvement over time noted in our results is an interesting observation; to our knowledge, no previous studies have reported this finding. Further investigations are needed to elucidate this phenomenon.

This study had several limitations. First, the APR group included many patients with tumors that were cT3/T4 depth, larger sizes, and closer to the AV, which may have influenced surgical manipulation around the prostate. We attempted PS matching to adjust for these factors; however, the analysis was difficult owing to the insufficient number of cases after adjustment. This finding suggests the need for future prospective studies with appropriate adjustments for patient characteristics. Furthermore, given the APR sample size of 44 cases and considering the typical postoperative dysfunction rate of 20%–30% in the AS group, the detectable absolute risk difference at 80% power and a two‐sided α of 0.05 would generally exceed 25%. Accordingly, the study may not have been powered to detect more modest associations. Second, although adjuvant chemotherapy was administered more frequently in the APR group, we did not include this as an adjustment factor. We took this approach owing to the limited number of previous studies investigating adjuvant chemotherapy for postoperative sexual dysfunction following rectal cancer surgery. Thus, such therapy has not previously been reported as a risk factor [30]. However, the potential indirect effects via peripheral neuropathy cannot be completely ruled out, and these effects should be considered when interpreting the results. Third, the extent of the pelvic floor dissection from the abdominal side and the details of the surgical procedure in the perineal approach were not assessed. Minimally invasive surgery may allow better preservation of the Denonvilliers' fascia and rectourethral muscle; however, standardizing surgical quality remains challenging. Robot‐assisted APR preserves these nerves through precise manipulation, potentially reducing sexual dysfunction [20]. Therefore, a thorough understanding of the neuroanatomy involved in erection and ejaculation, and surgical techniques based on that knowledge, is an important consideration for surgeons.

In conclusion, this study suggests that even with minimally invasive surgery, the incidence of sexual dysfunction remains high in patients who underwent APR. These results highlight the importance of understanding the neuroanatomy of the prostate and levator ani and of performing surgical manipulation with attention to preserving sexual function. Additionally, sexual function preservation in patients with rectal cancer should be an important consideration in surgical decision‐making and in determining if APR is the optimal approach.

Author Contributions

Taiki Kajiwara: writing – original draft, data curation, investigation, methodology. Shinobu Ohnuma: investigation, writing – review and editing, supervision. Michiaki Unno: writing – review and editing. Masakatsu Numata: investigation, writing – review and editing, methodology. Tomohiro Yamaguchi: investigation, writing – review and editing. Koya Hida: investigation, writing – review and editing. Tomonori Akagi: investigation, writing – review and editing. Suguru Hasegawa: investigation, writing – review and editing. Akio Shiomi: investigation, writing – review and editing. Ryo Inada: investigation, writing – review and editing. Manabu Shiozawa: investigation, writing – review and editing. Keisuke Kazama: writing – review and editing, investigation. Masanori Hotchi: investigation, writing – review and editing. Masahiko Watanabe: project administration, writing – review and editing. Takeshi Naitoh: project administration, writing – review and editing.

Funding

The authors have nothing to report.

Ethics Statement

All procedures were conducted in accordance with the ethical standards established by the institutions and national ethics committees responsible for human experimentation and in compliance with the 1964 Declaration of Helsinki and subsequent amendments. The study protocol was approved by the Clinical Research Review Board of Yokohama City University (IRB No. B190800044) and the institutional review boards of all participating centers. Written informed consent was obtained from all the patients, and the study was registered with the University Hospital Medical Information Network (UMIN000037367).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Aavailability Sstatement

The data that support the findings of this study are either available within the article, in the supporting information, or from the corresponding author upon reasonable request.

Supporting information

Supporting Information: Table S1. Questionnaire return number and rate.

Supporting Information: Table S2. Full version of patient characteristics, operative findings, and pathological outcomes of the entire cohort.

Supporting Information: Table S3. Full version of patient characteristics, operative findings, and pathological outcomes of AV < 5 cm cohort. AV, anal verge.