Harmanli, Ozgur H. MD; Khilnani, Rubina MD; Dandolu, Vani MD; Chatwani, Ashwin J. MD
From the Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, Temple University, Philadelphia, Pennsylvania
Reprints are not available. Address correspondence to: Oz Harmanli, MD, Temple University, School of Medicine, Dept. of OB/GYN/RS, 3401 North Broad Street, 7 OPB, Philadelphia, PA 19140; e-mail: firstname.lastname@example.org.
Received May 8, 2004. Received in revised form June 15, 2004. Accepted June 24, 2004.
OBJECTIVE: The purpose of this study was to evaluate the risk factors involved in failed vaginal hysterectomies.
METHODS: Data on all vaginal hysterectomies performed by a single gynecologic surgeon were collected prospectively. Patients requiring pelvic floor repair were excluded. Any procedure converted to the abdominal approach was classified as a failed vaginal hysterectomy and comprised the study group. For every woman who had a failed vaginal hysterectomy, the next 2 women who had successful vaginal hysterectomies immediately after the failed vaginal hysterectomy were taken as controls. Risk factors such as age, parity, body weight, surgical indication, uterine size, presence of leiomyomata in the anterior lower uterine segment, previous pelvic surgeries, abdominopelvic adhesions, location and length of cervix, narrow pubic arch, intraoperative complications such as bleeding requiring transfusion, visceral injury, nulliparity, and adnexal removal were compared between groups.
RESULTS: We compared 25 failed vaginal hysterectomies with 50 controls whose procedures were completed successfully through the vagina. Among all the factors gynecologists can assess preoperatively, only the presence of a narrow pubic arch increased the risk of failure for vaginal hysterectomy (odds ratio [OR] 4.1; 95% confidence interval 1.32–12.69). Intraoperative bleeding with transfusion was also found as an independent cause for conversion to laparotomy (OR 7.37; 95% confidence interval 1.75–31.06).
CONCLUSION: Women with a narrow pubic arch are not good candidates for vaginal hysterectomy. The most common unpredictable cause for conversion to laparotomy from the vaginal approach is intraoperative bleeding requiring transfusion.
LEVEL OF EVIDENCE: II-2
Each year in the United States, approximately 600,000 women undergo hysterectomy, and it is the most common nonpregnancy-related surgical procedure performed.1 The majority of these hysterectomies are performed abdominally, despite numerous published reports supporting the benefits and safety of the vaginal approach (Isik-Akbay EF, Harmanli OH, Panganamamula UR, Akbay M, Gaughan J, Chatwani AJ. Hysterectomy in obese women: a comparison between abdominal and vaginal routes. Oral poster presentation at the 2004 American Urogynecologic Society and Society of Gynecologic Surgeons Joint Scientific Meeting, San Diego, July 29–31, 2004).1–6 The conditions that are often cited as contraindications for vaginal hysterectomy are narrow pubic arch, enlarged uterus, morbid obesity, nulliparity, lack of uterine prolapse, need for oophorectomy, and previous pelvic surgeries.7,8 However, in recent years, all of the above reasons have been challenged with good evidence by gynecologists who have extensive clinical knowledge in vaginal surgery.2–6
At Temple University Hospital many attending physicians consider vaginal hysterectomy for all patients, except for those with advanced stage cancer, severe endometriosis, tuboovarian abscess, undiagnosed adnexal growth, uterine size greater than 24 weeks, or any nongynecologic indication that would necessitate laparotomy. Although most of these procedures are completed successfully through the vaginal route, there are some cases that require conversion to laparotomy due to various factors. The objective of this study was to identify the risk factors associated with failed vaginal hysterectomies and analyze the morbidity when a scheduled vaginal hysterectomy case is later converted to and completed through laparotomy.
MATERIALS AND METHODS
In a prospective cohort study, all women whose hysterectomies were started vaginally but subsequently converted to abdominal route between January 1996 and September 2003 by the same gynecologic surgeon at Temple University Hospital were followed up. During this period, all of the hysterectomy cases of this gynecologist were scheduled vaginally except for severe endometriosis, tuboovarian abscess, adnexal mass, uterine size greater than 24 weeks, or any nongynecologic indication that would necessitate laparotomy. Vaginal hysterectomies performed together with pelvic floor repair were excluded. Controls were the 2 subsequent successful vaginal hysterectomies by the same surgeon immediately after each failed vaginal hysterectomy.
Vaginal hysterectomies were performed under general anesthesia with the patient in steep lithotomy position. Weighted specula and vaginal malleable ribbons were used for retraction. The peritoneum was entered first posteriorly then anteriorly, with sharp dissection. All the uterine ligaments and vascular structures were ligated with absorbable sutures using surgical clamps. The bladder was drained intermittently during the procedures. Standard morcellation techniques were used when necessary. Ovaries and tubes were removed, if accessible, upon patients' request.
Risk factors, such as age, parity, surgical indication, uterine size, presence of leiomyomata in the anterior lower uterine segment, previous pelvic surgeries, abdominopelvic adhesions, location and length of cervix, obesity and morbid obesity, narrow pubic arch, intraoperative complications such as bleeding requiring transfusion, visceral injury, nulliparity, and adnexal removal were compared between the groups. The pubic arch, the angle formed by the inferior pubic rami with pubic symphysis as its vertex was measured using a right-angled cardboard. Any measurement less than 90° was considered narrow. Cervical length 4 cm and greater was classified as elongated.
The study was approved by the institutional review board. Univariate logistic regression analysis was used to estimate the relationship of the variables with the occurrence of a failed vaginal hysterectomy. The odds ratio (OR) and the 95% confidence interval (CI) were calculated for each discrete variable. The statistically significant variables were then combined in a multivariate stepwise logistic regression to ascertain that each variable was indeed statistically significant independent of the other factors. Paired and unpaired Student t tests were also used for statistical analysis. A P value of less than .05 was the significance level used for analysis of all the continuous variables.
Of 356 hysterectomies performed by the same gynecologic surgeon between January 1996 and September 2003, 34 were accomplished by laparotomy, and 6 with laparoscopic assistance. Three hundred sixteen procedures were started vaginally and 291 of them (92.1% of those attempted vaginally) were successfully completed without any laparotomy, giving a successful vaginal hysterectomy rate of 81.7%. Twenty-five (7.9%) women who were converted to laparotomy made the study group. Fifty controls were composed of the 2 subsequent successful vaginal hysterectomies after each failed vaginal hysterectomy. Characteristics of the patients including age, race, parity, body weight, uterine weight, and indication for surgery undergoing vaginal hysterectomy were statistically similar as shown in Table 1.
Among all the presumed risk factors for vaginal hysterectomy, only narrow pubic arch was statistically significant (40% in the failed, compared with 14% in the successful vaginal hysterectomy groups; OR 4.1, 95% CI 1.32–12.69) (Table 2). As for the complications, intraoperative bleeding requiring transfusion was the only one that reached statistical significance (40% in the failed compared with 6% in the successful vaginal hysterectomy groups; OR 7.37, 95% CI 1.75–31.06) (Table 3). Multivariate analysis showed that the contributions of these 2 factors to the outcome were independent.
Anterior location of cervix, long cervix, and anterior location of uterine leiomyomata showed only a trend toward increased prevalence in the failed vaginal hysterectomy group. However, the other variables, namely cystoscopy, previous cesarean section, previous myomectomy, pelvic adhesions, and removal of adnexa, were not significantly associated with failed vaginal hysterectomy.
Evidence-based medicine supports the vaginal approach as the preferred surgical procedure for hysterectomies, even for conditions commonly referred to as relative contraindications, in the hands of skilled gynecologists. Most of this evidence was derived from studies comparing vaginal hysterectomy outcomes for women with and without risk factors such as enlarged uterus, previous pelvic surgery, nulliparity, and obesity.2–6 Our prospective cohort study offers a unique perspective, because it is the first attempt to evaluate failed vaginal hysterectomy in a prospective comparative fashion according to our MEDLINE search of all articles reported in English from January of 1966 to March of 2004 using various combinations of the keywords “vaginal hysterectomy, hysterectomy, pubic arch, subpubic arch, female pelvis, pelvic dimensions, failed vaginal hysterectomy.”
In this study only the presence of a narrow pubic arch increased the risk of failure for vaginal hysterectomy, among all the risk factors gynecologists can assess preoperatively. This finding provided the first scientific evidence to this general recommendation that the pubic arch should not be compromisingly narrow.9
Intraoperative bleeding which required transfusion was also found to be an independent factor for failed vaginal hysterectomies and, therefore, necessitated conversion to laparotomy. It had an even higher odds ratio, which indicates that this complication was the most frequent cause for conversion. It was interesting to see that 2 of 10 (20%) bleeding complications occurred during removal of the ovaries after successful vaginal removal of the uterus (Table 3). These complications might have been avoided had we not attempted oophorectomy in these cases. The other complications such as ureter or bladder injury were not significantly different between the groups (Table 3). Of note, the only ureteral injury reported in this study occurred during the vaginal approach but was not recognized until laparotomy.
Finally, our study indicates that it is prudent to choose abdominal hysterectomy without a trial of vaginal approach in women with narrow pubic arch. Although some of the other factors we studied, such as anterior positioning of the cervix, elongated cervix, and the presence of uterine leiomyomata in the anterior lower segment, did appear more frequently in the failed vaginal hysterectomy group, we were not able to show statistically significant differences between the groups. We believe that bleeding could have been avoided or controlled had there been a wider pubic arch or no leiomyomata in the lower segment or a smaller uterus and so forth. We have to acknowledge that a larger group would improve the power and enable us to elicit significance of some of the other factors analyzed in this study. Thus, we will continue to collect data on all our failed vaginal hysterectomies and also will try to combine data from other practices with similar vaginal hysterectomy rates to delineate the proximate cause or causes for failure to complete the hysterectomy vaginally.
1. Farquhar CM, Steiner CA. Hysterectomy rates in the United States 1990–1997. Obstet Gynecol 2002;99:229–34.
2. Davies A, Vizza E, Bournas N, O'Connor H, Magos A. How to increase the proportion of hysterectomies performed vaginally. Am J Obstet Gynecol 1998;179:1008–12.
3. Varma R, Tahseen S, Lokugamage AU, Kunde D. Vaginal route as the norm when planning hysterectomy for benign conditions: change in practice. Obstet Gynecol 2001;97:613–6.
4. Kovac SR. Hysterectomy outcomes in patients with similar indications. Obstet Gynecol 2000;95:787–93.
5. Benassi L, Rossi T, Kaihura CT, Ricci L, Bedocchi L, Galanti B, et al. Abdominal or vaginal hysterectomy for enlarged uteri: a randomized clinical trial. Am J Obstet Gynecol 2002;187:1561–5.
6. Unger JB, Meeks GR. Vaginal hysterectomy in women with history of previous cesarean delivery. Am J Obstet Gynecol 1998;179:1473–8.
7. Carlson KJ, Nichols DH, Schiff I. Indications for hysterectomy. N Engl J Med 1993;328:856–60.
8. American College of Obstetricians and Gynecologists. Precis IV: an update in obstetrics and gynecology. Washington (DC): The College;1990:197.
9. Grody MHT. Vaginal hysterectomy: The large uterus. J Gynecol Surg 1989;5:301–12.
© 2004 by The American College of Obstetricians and Gynecologists.