The uterosacral ligament suspension is commonly used as a transvaginal approach for repair of apical pelvic organ prolapse. In this procedure, sutures are placed into the uterosacral ligament complex and affixed to the vagina to resuspend a prolapsing vaginal apex. Multiple authors have described technical variations of uterosacral ligament suspension,1–3 and recent studies4,5 have demonstrated that suture placement in a more postero-medial location may decrease the risk of ureteral injury.
In addition to ureteral injury, another notable complication is neuropathic pain occurring in the buttock, thigh, or perineum in up to 4% of patients.6,7 This injury is thought to be related to nerve compression and may resolve spontaneously or require release of the uterosacral ligament suspension suture. Previous anatomic work has demonstrated that uterosacral ligament suspension sutures can entrap sacral nerve roots.8 Involvement of the posterior femoral cutaneous nerve also has been hypothesized.6 This latter hypothesis is based on the lack of motor findings and the unique sensory distribution in patients with neuropathic pain after uterosacral ligament suspension, but posterior femoral cutaneous–nerve involvement has not been clearly demonstrated anatomically.
Therefore, our objective is to identify the retroperitoneal neural structures at risk during vaginal uterosacral ligament suspension suture placement. In addition, we wish to estimate the risk of neural injury based on suture placement technique.
MATERIALS AND METHODS
Uterosacral ligament suspension sutures were placed, followed by retroperitoneal dissection in 10 fresh-frozen (unembalmed) female cadavers obtained from the Duke University Medical Center Anatomic Gifts Program. This study was exempt from institutional review board approval. Cause of death was noted, and cadavers were excluded if there was gross disease preventing placement of uterosacral ligament suspension sutures. All cadavers underwent concomitant placement of a mid-urethral sling. Retropubic dissections were performed for a separate anatomic study. No other prolapse surgeries were performed.
Each cadaver was placed in lithotomy position. The rectum was packed with gauze to facilitate identification of the rectal margins while placing uterosacral ligament suspension sutures. A hysterectomy was performed; in cadavers with previous hysterectomy, the vaginal cuff was opened using a vertical midline incision. The small bowel was packed away and Breisky-Navratil retractors were used to expose the pelvic hollow. A total of six uterosacral sutures (three on each side) were placed transvaginally in each cadaver (Fig. 1). Sutures were placed in the same order each time and all procedures were performed by a right-hand-dominant pelvic reconstructive surgery fellow under direct guidance and assistance from an attending urogynecologist.
Sutures were placed using three previously described techniques.1,2,4 In the first technique (Allis), an Allis clamp was used to apply traction to the vaginal cuff. The strong suspensory tissue of the uterosacral ligament was identified, and a second Allis clamp was placed in the mid-portion of the uterosacral ligament, cephalad to the ischial spine. This clamp then was elevated off the underlying retroperitoneal structures, and a suture was passed directly under the clamp while “tenting” the ligament. In the second technique (deep UR-6), traction again was applied to the distal end of the uterosacral ligament at the vaginal cuff to identify the course of the ligament as it traversed posteromedially toward the sacrum. The packed rectum was retracted medially using a Breisky-Navratil retractor. The ischial spine was palpated with the contralateral index finger, and its location was noted. A suture on a UR-6 needle (26-mm diameter, five-eighths circle) was placed in a “deep” (dorsal and posterior) fashion into the pararectal area of the uterosacral ligament, approximately 1 to 2 cm medial and cephalad to the ischial spine. The third technique (deep CT-1) was identical, but with a larger CT-1 needle (36-mm diameter, half-circle). All three sutures were placed first in the right hemipelvis, and then the same procedures were followed on the left side.
After suture placement, the abdominal cavity was accessed. A pin was placed into the peritoneum at each ischial spine; transperitoneal horizontal and vertical distances of each suture from the ischial spine were measured (Fig. 1). The rectum was tied and the torso was disarticulated from the pelvis. All dissections were performed by an anatomist. The peritoneum was dissected from the underlying connective tissue and the inferior hypogastric plexus was identified. Uterosacral sutures that traversed the ureter or rectum were noted. The pelves then were divided in the midline, and each hemipelvis was dissected separately. Pudendal nerves, posterior femoral cutaneous nerves, and sacral nerve roots were identified. Sutures that grossly entrapped nerves or pelvic floor muscles were noted.
Biopsy specimens were taken of the most dorsal tissue that each suture traversed. Biopsy specimens were sectioned, processed, and stained with hematoxylin and eosin. Biopsy specimens were also immunostained with S100, a nerve-specific antibody. Anti-rabbit immunoglobulin G immunostaining was performed as a negative control for all specimens. Slides were reviewed by a blinded investigator (R.C.B.). Neural tissue was identified and the largest cross-sectional nerve diameter was recorded.
Right- and left-side anatomic distances of sutures from the ischial spine were compared using the Wilcoxon signed rank test for paired data. Comparisons of distances based on technique of suture placement were analyzed using the Kruskal-Wallis test. Histologic nerve diameter was compared with gross findings using the Student's t test. All statistical analyses were performed using SPSS 15.0 for Windows. P<.05 was considered statistically significant.
All cadaveric subjects were white women; median age was 84 years (range 60–98) and median body mass index (BMI) was 22.2 (range 16–29). Body mass index is calculated as weight (kg)/[height (m)]2. Four of 10 subjects (40%) had undergone previous hysterectomy. In one subject with a colostomy and colon cancer, there was nodularity along the right aspect of the rectum, but the remainder of the peritoneum was preserved; therefore, we elected to continue study procedures. Another cadaver had significant adhesions of the rectosigmoid to the pelvis requiring adhesiolysis before placement of uterosacral ligament suspension sutures. The remaining eight cadavers had normal pelvic anatomy.
A total of 60 sutures were placed (six in each cadaver). Median locations of sutures relative to the ischial spine are summarized in Table 1. Sutures placed with the Allis technique were further cephalad relative to the ischial spine on the right compared with the left hemipelvis (3.7 compared with 2.5 cm; P=.03). Sutures placed with the deep techniques were not significantly different in relation to the ischial spine between the right and left sides of the pelvis. Overall, when comparing suture placement techniques (ie, Allis, deep UR-6, and deep CT-1) there were no differences in distances of sutures from the ischial spine (P=.44).
We noted one ureteral injury in which a deep CT-1 suture traversed the right ureter. Notably, this was a cadaver in which adhesiolysis had been performed because of significant rectosigmoid adhesions. Rectal injury occurred in two instances. An Allis suture traversed the right rectal lumen in a cadaver with previous colostomy and rectal nodularity. In addition, a deep CT-1 suture perforated the left rectal lumen in a cadaver with normal pelvic anatomy. In four cadavers, sutures incorporated muscles of the pelvic floor (levator ani, coccygeus, or piriformis). In three cadavers, there was a unilateral deep suture incorporating pelvic floor muscles (one deep UR-6 on the right, one deep CT-1 suture on the right, and another deep CT-1 suture on the left). In one cadaver, pelvic floor muscles were incorporated by bilateral deep UR-6 and deep CT-1 sutures.
In seven cadavers, we identified gross entrapment of portions of sacral nerve roots (Figs. 2 and 3). In two cadavers, bilateral deep UR-6 and deep CT-1 sutures entrapped sacral nerves. In four cadavers, deep UR-6 and deep CT-1 sutures on the right side involved sacral nerves. In one cadaver, only the right-side deep CT-1 suture involved a sacral nerve. There were no instances of sacral nerve involvement with sutures placed using the Allis technique. There was no relationship between neural injury and distances of sutures from the ischial spine. Entrapment was more likely with sutures placed into the right hemipelvis compared with the left (43% compared with 13%; P<.001). When nerve entrapment was noted, the S2 root and S2/S3 trunk were most often incorporated (10 sutures), followed by the S1 root, S3 root, S3/S4 trunk (two sutures each), and the S4 root (one suture). Of the sutures that incorporated muscles of the pelvic floor, four of seven also entrapped sacral nerves.
In two instances, we noted gross entrapment of hypogastric nerve: one in a suture placed with the Allis technique and another with the deep UR-6 technique. We also noted small diameter branches of hypogastric nerves extending through the uterosacral ligaments in all cadaveric specimens. Our gross dissections located the pudendal and posterior femoral cutaneous nerves dorsal and lateral to the ischial spine. Because all of our uterosacral ligament suspension sutures were located medial and slightly cephalad to the ischial spine, there were no instances of pudendal or posterior femoral cutaneous–nerve involvement in any of the 10 cadavers.
S100 immunostaining was performed on biopsy specimens from the deepest (most dorsal) tissue that all sutures traversed. Eight biopsy specimens were unable to be analyzed secondary to tissue autolysis (evidenced by lack of nuclei on hematoxylin and eosin staining). Nerve tissue was present in all 52 analyzed specimens, with the largest cross-sectional diameter ranging from 30 to 1,225 micrometers (Fig. 4). Mean nerve diameter was significantly larger in biopsy specimens in which entrapment was noted grossly (472 micrometers compared with 108 micrometers; P<.001).
We identified a high risk of neural entrapment when placing uterosacral ligament suspension sutures using a deep (dorsal and posterior) technique, regardless of the needle size or curvature. Nerve entrapment was most likely to involve fibers of the S2 sacral nerve root or S2/S3 sacral nerve trunk in the right hemipelvis.
We were able to simulate vaginal uterosacral ligament suspension in unembalmed cadavers using various published techniques. We also were able to compare these techniques regarding risk of neural injury and explore differences between left- and right-side suture placements. Strengths of this study are that we placed sutures in a systematic fashion and corroborated our gross anatomic findings with tissue biopsy specimens. In addition, the multidisciplinary approach utilizing an anatomist and a blinded pathologist affords some additional diagnostic accuracy regarding retroperitoneal and nervous structures.
We found a higher risk of neural entrapment than expected. This is likely related to the fact that our dorsal and posterior sutures were purposefully very deep because we were trying to elucidate which neural structures might be at risk during uterosacral ligament suspension. It is likely that sutures placed during live surgery are not routinely as deep as those placed in this study, even if they are placed in a dorsal and posterior arc. Alternatively, our cadavers were mainly normal weight (median BMI 22.2) and the sacral nerves may be more vulnerable to injury in a subject with less retroperitoneal fat. We did not exclude cadavers with pelvic adhesions or previous pelvic surgery because we were mainly interested in studying retroperitoneal structures. Certainly the presence of adhesions and rectal nodules might have contributed to our observed ureteral and rectal injuries. In addition, the subjects in our study did not necessarily have prolapse, which is a distinct difference between cadavers and live surgical patients. However, whereas our cadaveric subjects may not be fully generalizable to the typical urogynecologic surgical population, we believe that the results regarding the retroperitoneal neural structures are still valid.
One additional limitation is that we did not randomize the order of suture placement. We were unsure whether previously placed sutures would inhibit the ability to ventrally tent the peritoneum, as specified with the Allis technique. Therefore, we elected to place Allis sutures first in all cadavers, followed by deep UR-6 and deep CT-1, respectively. We left sutures untied and maintained each suture off tension when placing successive sutures. In addition, we reassessed anatomic landmarks and palpated the ischial spine before each successive suture placement. Therefore, we believe that each suture was placed in a relatively independent manner.
Flynn et al6 reported a series of 7 of 182 patients who experienced posterior thigh pain in the S2–S3 sensory distribution after undergoing uterosacral ligament suspension. Because of the dermatomal distribution and lack of motor findings, the authors hypothesized that the posterior femoral cutaneous nerve was involved. Our findings corroborate that entrapment is likely to occur to S2 or S3 nerve fibers, but we were unable to demonstrate any sutures involving the posterior femoral cutaneous nerve. Rather, our sutures were placed in areas overlying sacral nerve roots, and the posterior femoral cutaneous nerve was consistently lateral and further dorsal than the uterosacral ligament suspension sutures. Our findings are consistent with other anatomic data revealing sacral nerve roots and trunks in close proximity to vaginal uterosacral ligament suspension sutures.8–10 Collins et al11 performed a study in embalmed cadavers in which uterosacral ligament suspension sutures were placed from an abdominal approach. They also reported the posterior femoral cutaneous nerve to be lateral and dorsal to the uterosacral ligament suspension operative site. In their study, they identified entrapment of a branch of the inferior hypogastric plexus and hypothesized that patients experience referred pain from this visceral plexus. We also noted hypogastric nerve involvement in a few instances and identified small diameter nerves, likely autonomic branches, in all of our tissue biopsy specimens. Although our data do not entirely refute the theory of autonomic nerve compression with referred pain, other groups have also reported perineal and vulvar pain (ie, S3–S4 dermatomes) after uterosacral ligament suspension.7 Our findings of sacral nerve involvement would explain sensory pain in any region from S1–S4 because we demonstrated entrapment of sacral nerve fibers in all of these areas.
In our study, there were significantly more neural injuries in the right hemipelvis compared with the left. In addition, the Allis sutures were approximately 1 cm more cephalad on the right compared with the left. This likely reflects the right-hand dominance of the surgeon, which tends to facilitate a more aggressive grasp of tissue on the right side. In addition, it is easier to expose the right hemipelvis as the rectosigmoid lies slightly to the left of midline, even after deflection with a vaginal retractor. Our results are actually consistent with the clinical findings of Flynn et al6 in which five of seven patients experienced pain on the right side only, with one patient experiencing pain on the left and one reporting bilateral pain.
Our findings indicate that sacral nerve roots or trunks can become entrapped within uterosacral ligament suspension sutures placed in a very dorsal (deep) fashion. However, it is curious that patients with an entrapment syndrome of a large nerve root with mixed motor and sensory components mainly report only sensory pain. It is possible that the amount of crossover among nerve fibers within the sacral plexus can compensate for an injury to a single nerve root or trunk. Alternatively, it may be that the superficial location of sensory fibers within a nerve root makes them more susceptible to direct compression than motor fibers.
It is still unclear whether there is a particular zone of safety with uterosacral ligament suspension that allows the surgeon to obtain good purchase on the uterosacral ligament while avoiding ureteral injury, rectal injury, and nerve entrapment. Aronson et al4 proposed the deep (dorsal and posterior) method of suture placement in efforts to minimize ureteral injury. In our study, this technique was more likely to entrap sacral nerves compared with the technique of elevating the uterosacral ligament off of underlying structures. However, sutures placed with the Allis technique contained mainly peritoneum with a small amount of underlying connective tissue. In contrast, sutures placed using the deep UR-6 or deep CT-1 techniques encompassed a large amount of connective tissue; therefore, they are likely to have greater anchoring strength.
The depth of needle passage with the dorsal and posterior technique appears to be dependent on the surgeon because we were unable to demonstrate differences based on needle size or curvature. In addition, it is unclear whether BMI may play a role in the risk of nerve entrapment. In our heaviest cadaver (with a BMI of 29.2), all three techniques resulted in sutures traversing mainly peritoneum with only a small amount of underlying connective tissue. Because our sample is skewed toward individuals with normal weight, further anatomic data with subjects having higher BMI may be useful.
Finally, many surgeons use a combination of permanent and delayed absorbable suture for uterosacral ligament suspension. In previous case series, patients with neuropathic pain symptoms have experienced prompt resolution of pain with release of uterosacral ligament suspension sutures, when vaginally accessible.6 Surgeons who use the dorsal and posterior technique may wish to consider our findings and the risk of involvement of sacral nerve roots, because permanent sutures are less accessible once the vaginal cuff is closed.
1. Barber MD, Visco AG, Weidner AC, Amundsen CL, Bump RC. Bilateral uterosacral ligament vaginal vault suspension with site-specific endopelvic fascia defect repair for treatment of pelvic organ prolapse. Am J Obstet Gynecol 2000;183:1402–10.
2. Shull BL, Bachofen C, Coates KW, Kuehl TJ. A transvaginal approach to repair of apical and other associated sites of pelvic organ prolapse with uterosacral ligaments. Am J Obstet Gynecol 2000;183:1365–73.
3. Symmonds RE, Williams TJ, Lee RA, Webb MJ. Posthysterectomy enterocele and vaginal vault prolapse. Am J Obstet Gynecol 1981;140:852–9.
4. Aronson MP, Aronson PK, Howard AE, Morse AN, Baker SP, Young SB. Low risk of ureteral obstruction with “deep” (dorsal/posterior) uterosacral ligament suture placement for transvaginal apical suspension. Am J Obstet Gynecol 2005;192:1530–6.
5. Buller JL, Thompson JR, Cundiff GW, Krueger Sullivan L, Schon Ybarra MA, Bent AE. Uterosacral ligament: description of anatomic relationships to optimize surgical safety. Obstet Gynecol 2001;97:873–9.
6. Flynn MK, Weidner AC, Amundsen CL. Sensory nerve injury after uterosacral ligament suspension. Am J Obstet Gynecol 2006;195:1869–72.
7. Lowenstein L, Dooley Y, Kenton K, Mueller E, Brubaker L. Neural pain after uterosacral ligament vaginal suspension. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:109–10.
8. Wieslander CK, Roshanravan SM, Wai CY, Schaffer JI, Corton MM. Uterosacral ligament suspension sutures: anatomic relationships in unembalmed female cadavers. Am J Obstet Gynecol 2007;197:672e1–6.
9. Siddique SA, Gutman RE, Schon Ybarra MA, Rojas F, Handa VL. Relationship of the uterosacral ligament to the sacral plexus and to the pudendal nerve. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:642–5.
10. Schon Ybarra MA, Gutman RE, Rini D, Handa VL. Etiology of post-uterosacral suspension neuropathies. Int Urogynecol J Pelvic Floor Dysfunct 2009;20:1067–71.
© 2010 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
11. Collins SA, Downie SA, Olson TR, Mikhail MS. Nerve injury during uterosacral ligament fixation: a cadaver study. Int Urogynecol J Pelvic Floor Dysfynct 2009;20:505–8.