Vassallo, Brett J. MD; Karram, Mickey M. MD
The scant literature describing the complication of postoperative vaginal stenosis repeatedly claims that it is a “relatively uncommon” finding. In our referral-based practice, we have seen women with this complication at a rate of four to six cases per year. It is our belief that any gynecologic surgeon who operates with significant frequency will either encounter this complication or see patients in referral from other surgeons. It is therefore paramount that one is familiar with the different approaches to this problem.
This complication can occur at any point during a woman's reproductive and postreproductive years. It most often arises in the following settings: postepisiotomy scarring, repair of obstetric laceration, levator-plasty at the time of posterior colporrhaphy, posthysterectomy, or excessive subepithelial plication or trimming of the vaginal mucosa at the time of colporrhaphy. The purpose of this study was to prospectively assess the outcomes of four approaches to the surgical management of iatrogenic vaginal constriction.
MATERIALS AND METHODS
All patients referred to our practice with postoperative vaginal constriction and resultant dyspareunia or apareunia who are able to tolerate placement of a small vaginal dilator are initially managed conservatively with twice-daily vaginal dilation and vaginal estrogen cream, unless medically contraindicated. Between January 1997 and February 2002, 20 women were identified who failed manual dilation and required a surgical correction to treat this complication. Five patients were operated on before beginning the study. Institutional review board approval was obtained to study this problem in 1999, and after consent was obtained from these five individuals, they were asked to return to the office for a review of their history and for a postoperative examination. The remaining 15 patients received both pre- and postoperative evaluations in the form of a thorough history and physical examination, with measurements taken of vaginal length and caliber using standardized vaginal dilators. Three women were treated with Z-plasty. Eight had incision of a vaginal ring or ridge. Eight had vaginal advancement flaps. One underwent placement of a free skin graft. Postoperative vaginal packing with moistened gauze was used in all cases, and gauze was left in place for approximately 24 hours. Daily sitz baths, douches, and postoperative dilators were not recommended. Patients were asked to refrain from sexual intercourse for at least 6 weeks after their surgery.
Z-plasty was employed to treat patients with well-circumscribed constriction rings and stenotic vaginal canals to increase vaginal diameter. It was used in women without extensive scarring, as healthy surrounding tissue is required to create healthy flaps without predisposing the patient to further constriction. Vaginal length was not a significant consideration, as this procedure does little to compromise this parameter. The incision length and angles of the “Z” were determined by the degree of stenosis but most frequently were made 2 cm long at 60° angles from one another (Figure 1). 1,2 The width gained increases as the angles increase. The orientation of the “Z” can be vertical or transverse, depending on the location of the stricture. To perform a vertical Z-plasty, a transverse incision was made through the length of the constrictive scar. The upper arm of the “Z” was then incised from one end of the previously described transverse incision into the vagina. The lower arm was created by extending the other end of the transverse incision toward the perineum. The flaps were then mobilized. Once hemostasis was achieved, the flaps were transposed and sutured into place using fine absorbable suture. For a transverse Z-plasty, vertical incisions were made at the introitus and the flaps created by extending one end of the incision into the vagina and the other onto the labia. This technique has the advantages of widening the introitus without creating a midline scar or compromising vaginal length.2
Incision of vaginal ring or ridge was performed in such a fashion that the band of scar tissue is completely incised(Figure 2). This usually requires cutting all the way to the ischiorectal fat. If inadequate caliber was created after one incision, another was made on the opposite side. Cautery or interrupted ties were then used, if necessary, to obtain hemostasis. The incision was then permitted to granulate and heal by secondary intention. This procedure was selected if extensive scarring prevented the use of a Z-plasty or if concern for vaginal length due to amount of scarring or location of the stricture excluded the use of an advancement flap.
Vaginal advancement flap was performed by making a linear or curvilinear incision over the scar tissue (Figure 3). A flap of unscarred mucosa was created by undermining the healthy mucosa and excising the scar tissue. The freed mucosa, once adequately mobilized, was sutured over the defect, with care being taken not to produce excessive tension at the incision. Hemostasis is important in this procedure, as large hematomas may otherwise result owing to the potential space provided by the advanced flap. Vaginal packing was routinely used for tamponade effect. This procedure was chosen if the stricture was more distal, as tissue mobilization was less likely to compromise vaginal caliber. Preoperative vaginal length was an important consideration, as it is decreased somewhat by this operation.
Although a somewhat more involved operation than the three described above, the free skin graft is also quite simple (Figure 4). Because of its more invasive nature, however, the selection of this operation was limited to patients in whom very extensive scarring and/or previously compromised vaginal length and caliber prevented the use of one of the simpler operations already described. Our graft was composed of dermis and epidermis. It was constructed from an elliptical incision in the skin overlying the patient's iliac crest. Full-thickness grafts are superior to split-thickness grafts in that there is a decreased incidence of postoperative contracture, and because it is not a vascular flap, one need not be as concerned about length–width ratio.3,4 All fat was removed from the donor tissue, and careful attention was paid to hemostasis at the recipient site. A relaxing incision was made in the vagina through the area of stenosis. The graft was then sewn into place with fine, absorbable suture.
Each patient's age, parity, estrogen status, and antecedant procedure were reviewed (Table 1). Mean age and parity were 51.3 years (range, 39–76 years) and 2.4 (range, 0–6), respectively. One woman was nulliparous. In 15 cases (75%), the antecedent procedure involved a posterior colporrhaphy. As inferred above, all patients received a trial of vaginal estrogen cream before surgery; only three (15%) were either premenopausal or receiving hormone replacement therapy upon referral. Mean follow-up was 17 months (range, 3–32 months). Subjective cure was defined as resumption of pain-free vaginal intercourse. Objective (anatomic) cure was defined as resolution of physical examination findings that, upon palpation, reproduced the pain felt during intercourse.
Three women (15%) underwent Z-plasty. All were anatomic and subjective cures. One patient's course was complicated by postoperative bleeding due to a superficial separation of one of the flaps of tissue. This required a return to the operating room within 6 hours of the initial procedure and the placement of two simple, interrupted 3-0 polyglactin 910 sutures. She did not require any blood products.
Eight women (40%) were treated with incision of vaginal constriction ring or ridge. Seven of the eight patients (87.5%) had complete resolution of the objective finding (constriction ring or ridge), and five (62.5%) reported resolution of their subjective complaint of dyspareunia. One patient was noted to have a 2 × 3-cm pararectal hematoma at her 2-week, follow-up visit. This resolved spontaneously. An additional patient in this group was noted to have persistent bleeding from the vaginal incision in the recovery room. This was resolved by application of a pressure packing that was left in place for approximately 12 hours. One of the patients who reported complete resolution of her dyspareunia did have persistent dyschezia. This was present after her initial vaginal procedure and was thought to be some kind of nerve entrapment secondary to scarring from that procedure.
Eight patients (40%) underwent vaginal advancement flap, and six (75%) were objective and subjective cures. No complications were encountered in this group.
The single participant who had the free skin graft showed complete resolution of her subjective complaint, associated with normal vaginal sensation. Similarly, there was complete resolution of her vaginal stricture.
Overall, 85% were cured of the objective finding on examination that was correlated with their subjective complaint of dyspareunia (Table 2). Seventy-five percent reported complete resolution of their symptoms. Four of the five patients with persistent symptoms reported improvement. The overall complication rate was 15%, and all complications were hemorrhagic in nature and resolved with minimal intervention.
Although the prevalence of iatrogenic vaginal constriction is unknown, a small percentage of women do develop this complication after reconstructive surgery. In 15 (75%) of the cases reported here, the antecedent procedure included a posterior repair. Of the patients who did not have posterior repairs, only one had an anterior colporrhaphy. Two of the women had procedures complicated by erosions of foreign materials, and two patients developed dyspareunia after procedures that intentionally induce tissue damage (laser and radiation therapy). Review of available detailed operative reports suggested that the posterior colporrhaphies performed involved plication of levator muscles to the mid-line to effect the reconstruction of the posterior vaginal compartment. Several authors have shown that aggressive plication of the levator muscles as part of a posterior colporrhaphy results in a high rate of dyspareunia.5,6 We have altered our clinical practice because of these findings and now routinely perform site-specific rectocele repairs to avoid the complication of iatrogenic vaginal constriction. Furthermore, some surgeons elect to routinely perform anterior and posterior colporrhaphies together, regardless of the fact that only the anterior or the posterior compartment is lacking support. One must be cautioned against attempting to improve support where no such improvement is needed. Such practice increases the risk of excising too much mucosa or over-aggressively plicating the vaginal subepithelium.
Various clinical criteria were considered in selecting the appropriate corrective operation. Z-plasty was used for midvaginal and introital constrictions. It is most appropriate in situations in which the amount of scar tissue is not great and the depth of scar involvement is not too deep. Larger constrictions require the surgeon to create larger flaps and lead to a greater chance of further postoperative constriction. We employed incision of the ring or ridge in patients with more concentric constrictions and extensive fibrosis. This repair permits mobilization of tissues with the least amount of dissection. It is also appropriate for patients who are greater surgical risks, as it can be performed easily and quickly under local anesthesia. Limitations to vaginal advancement flaps primarily include the operator's ability to mobilize adequate healthy vaginal tissue, exposure, as well as length of the vagina. We have generally employed this procedure in more distal strictures in women who have adequate vaginal length. Free skin grafts have been used when the condition of the mucosa was too poor to permit the usage of one of the less invasive repairs described here. We elected to perform the free skin graft on the one patient in this series because she had received prior radiation therapy. Alternatively, free skin graft would be appropriate when excision of the constricting portion of the vagina leaves a defect too large to close primarily. Because the graft is not on any pedicle, this operation can be used in any part of the vagina, including more proximal areas where it would be difficult to mobilize tissues and/or swing in a flap.
We believe clinical findings should guide the selection of the particular corrective procedure, but as a matter of statistical interest we compared our outcomes using the Pearson χ2 test. There was not a statistically significant difference in cure rates among the four groups (P = .71). We also performed a sample size analysis and determined that, assuming a cure rate of 90%, to detect a 20% difference between the groups with 90% power we would require two individuals in each group. Because of this finding we ran the Pearson χ2 test on the three groups that met this sample size requirement and still failed to detect a statistically significant difference (P = .56).
Z-plasty has been described several times in the literature.1,7 These reports, however, are descriptive and offer no long-term follow-up. Our technique is similar to that described by Lee.2 Although this series of patients only includes three in whom we used this technique, we have been able to document both objective and subjective success with long-term follow-up.
Incision of the vaginal constriction ring or ridge is one of the fastest and simplest approaches to this complication. A similar approach was described in 1977, in which a mid-vaginal constriction ring was incised in a vertical fashion and the defect closed horizontally.8 The modification we employ simplifies this operation even further by avoiding immediate closure of the incision, which may create tension and further stricturing. As demonstrated by our outcome data, it is imperative that meticulous hemostasis be achieved in performing this procedure.
The technique we used in the patient who received the free skin graft is well described.3 This previous study of ten women represents the largest sample and best-described results for a single operation for postoperative vaginal constriction.
Nichols has described how to close rectangular, circular, and triangular defects with various flap procedures.9 We have found it helpful to be familiar with these techniques if the wound does not close easily once the vaginal mucosa has been mobilized.
Two of the five patients who did not achieve complete resumption of normal sexual function were found to have vaginal atrophy on examination and are being treated with estrogen cream. The patient who is still apareunic had a back injury after her vaginal repair that prevented her from having intercourse for approximately 6 months. This likely contributed to the persistence of her symptoms. Two of the women with unresolved dyspareunia only had pain with deep penetration, indicating that their persistent symptoms are more likely due to inadequate vaginal length rather than caliber.
Although the complication of iatrogenic vaginal constriction occurs in several settings, the frequency with which patients are referred to a practice such as ours limits the ability for one group to accumulate data on a large number of women. The small sample size in this study limits one's ability to generalize our success rate and low incidence of complications. Length of follow-up is clearly important in reports of this nature. We aggressively sought face-to-face follow-up with these patients over the time of the study and as a result have more than 1 year follow-up in 70% of the patients, and all patients were followed through the resumption of sexual intercourse. The one patient who did not resume intercourse was followed for 32 months.
Each of the procedures described is a straightforward and valid approach to this sensitive complication. Z-plasty, incision of vaginal ring or ridge, vaginal advancement flap, and free skin graft have each been employed in our practice with success. Regardless of the procedure selected, all patients with iatrogenic vaginal constriction should be given an attempt at vaginal dilation and adequate vaginal estrogenization. The complication of iatrogenic vaginal constriction is probably underreported and much more common than the available literature suggests.