There are various indications and approaches for hysterectomy; yet, the difference in long-term risk of subsequent prolapse after surgery is not well studied.
To assess the risk of prolapse after abdominal, vaginal, and laparoscopic or robotic hysterectomy for up to 17 years from surgery.
A retrospective chart review study of women undergoing hysterectomy across all indications (benign and malignant) between 2001 and 2008 was conducted. An equivalent random sample of hysterectomy patients was selected each year. We compared demographic and other surgical characteristics data including age, race, parity, body mass index, indication and year of hysterectomy, blood loss, cervix removal, cuff suspension, and complications using chi-square, Kruskal-Wallis test, and Fisher’s exact across the 3 groups. Presence and treatment of subsequent prolapse (based on patient symptoms, pelvic exam, International Classification of Diseases, Ninth Revision diagnosis, and current procedural terminology pessary or surgical codes) were compared with Kaplan-Meier survival analysis and Cox proportional hazards regression.
Of the 2158 patients, 1459, 375, and 324 underwent open, vaginal, and laparoscopic or robotic hysterectomy, respectively. The vaginal group (56) was older than the abdominal (52) or laparoscopic or robotic (49) groups, with a P value of <.05. Most patients were White with a mean body mass index of 30 kg/m 2 . The main indication was cancer for abdominal (33%) and laparoscopic or robotic hysterectomy (25%) and prolapse for vaginal hysterectomy (60%). Time to prolapse was shortest after vaginal surgery (27 months) and longest after laparoscopic or robotic surgery (71 months). After controlling for confounders, including surgery indication, the hazard ratio for subsequent prolapse was no different among vaginal (hazard ratio=1.36 [0.77–2.45]), laparoscopic or robotic (hazard ratio=1.47 [0.80–2.69]), or open (reference) hysterectomy. Prolapse grade was similar across the 3 groups. About 50% of women with recurrent prolapse received physical therapy, pessary, or surgical treatment.
At the 17-year follow-up, the route of hysterectomy is not associated with a difference in recurrence, grade, or subsequent treatment of prolapse when the indication for hysterectomy is considered. Prolapse, as an indication for hysterectomy, increases risk for recurrence. Women planning a hysterectomy should be counseled appropriately about the risk of subsequent prolapse.
Pelvic organ prolapse (POP) has an overall prevalence of 3% to 6%, and is even more common in older women. With the increase in prevalence of POP, , the need for reconstructive surgery is predicted to increase by 45% over the next 3 decades associated with a predicted rise in costs to exceed $1 billion per year. , POP has a considerable impact on quality of life where patients generally complain of feeling a vaginal bulge and pressure in addition to voiding, defecatory, and sexual dysfunction. Risk factors for POP include increasing age, parity, race and body mass index (BMI).
Why was this study conducted?
There is a lack of robust long-term data on whether differences exist in the incidence of pelvic organ prolapse between different modes of hysterectomy, generating conflicting opinions.
The unadjusted risk of prolapse is highest for vaginal hysterectomy in up to 17-year follow-up (17%); however, the adjusted risk is similar for abdominal, vaginal, and laparoscopic or robotic hysterectomy after controlling for age, parity, body mass index, and year and indication of surgery. About half of women with pelvic organ prolapse after hysterectomy receive treatment.
What does this add to what is known?
Most vaginal hysterectomies are performed for prolapse, which in turn are associated with the highest risk of prolapse recurrence. However, this risk is no different across all modes of hysterectomy when indication (such as prolapse, cancer, or other) is accounted for.
Hysterectomy is the most common major gynecologic surgery in the United States , and is considered to be a potential risk factor for POP with an incidence of postoperative vault prolapse varying from 2% to 43%. , One study estimated an incidence of 6.25% for posthysterectomy vault prolapse requiring surgical correction. In another study, the incidence of prolapse requiring surgical correction after hysterectomy was 1.3 to 4.2 per 1000 women-years. Although the American College of Obstetricians and Gynecologists (ACOG) recommends vaginal apex suspension such as a McCall culdoplasty (MC) to be performed at the time of hysterectomy to reduce risk of subsequent POP, it is not known if all gynecologists at our institution or elsewhere follow this recommendation routinely. Furthermore, little is known about whether different hysterectomy approaches have a different risk factor profile regarding subsequent POP. Moreover, the effectiveness of prophylactic measures at the time of surgery that reduce the risk of POP (such as uterosacral ligament suspension) after different hysterectomy routes is unknown.
The goal of our study was to determine whether there exists a difference in subsequent POP occurrence and treatment after different modes of hysterectomy (abdominal, vaginal, and laparoscopic or robotic), and whether the 3 groups differed by timing of POP occurrence and by indication of hysterectomy.
Material and Methods
This was a retrospective chart review analysis of women who underwent hysterectomy for any indication at a tertiary care hospital in Boston from January 2001 through December 2008 to allow us to have at least 10-year follow-up data from the last year of the study period electronic review of the medical records was completed through the end of 2018 for a total of up to 17 years of follow-up. We included all women who underwent hysterectomy regardless of indication. We excluded women who did not follow-up within our healthcare system after the index surgery. The exposure was defined as hysterectomy (by type) and the primary outcome of interest was defined as symptomatic prolapse in any compartment subsequent to the index surgery. All surgeries between 2001 and 2008 with current procedural terminology (CPT) code for hysterectomy were abstracted from the electronic system, and a random sample of all hysterectomies by route of surgery were included in the analysis.
Specifically, each third medical record number pooled by the system was included in the review. Based on our power calculations discussed later in this section, it was estimated we will have an adequate sample size by following this strategy to answer our study question. During the study period, because the majority of hysterectomies was performed abdominally, open hysterectomies were oversampled compared with the vaginal and laparoscopic or robotic cases to have equivalent and proportional representation. The hysterectomies were conducted by different gynecologists and gynecologic subspecialists with different practice standards with respect to postoperative follow-up care. To simplify, we considered women who had at least 1 gynecologic follow-up exam postoperatively to be eligible for study inclusion. After the first 12 weeks postoperatively, most patients were followed up by their primary care physicians. There was a total of 172 women who did not follow-up within the system or who had incomplete medical records that were excluded from the final analysis.
All charts were thoroughly reviewed from the date of the index surgery till the end of the study period, including all progress notes from primary care physicians, general gynecologists, gynecologic subspecialists, urologists, and colorectal surgeons. Data abstracted from the medical records included: age at hysterectomy, race, parity, BMI, indication for hysterectomy, type of hysterectomy (abdominal, vaginal, laparoscopic or robotic), concomitant surgeries, removal of cervix, intra and perioperative complications, estimated blood loss (EBL), vaginal apex suspension, presence of prolapse after hysterectomy (based on progress notes subsequent to the index surgery), time to prolapse occurrence, type of prolapse (cystocele, rectocele, vault prolapse), grade of prolapse (defined using the Baden-Walker grading system), and treatment of subsequent prolapse (none, pessary, surgery). The presence of POP after hysterectomy was based on documentation in the clinical progress notes (ie, patient’s subjective symptoms and physician’s pelvic examination or diagnosis), pelvic floor physical therapy notes, International Classification of Diseases, Ninth Revision (ICD-9) POP diagnosis codes, and prolapse pessary or CPT codes. When available, POP by compartment was objectively measured using the Baden-Walker system (or inferred from the pelvic exam or POP quantification (POP-Q) exam), because most surgeons performing the hysterectomies were not female pelvic medicine and reconstructive surgeons and as such they did not use the POP-Q system.
The null hypothesis was that there is no difference in the rate of posthysterectomy prolapse among the 3 hysterectomy routes. Considering the incidence of clinically significant posthysterectomy prolapse to be approximately 6.25%, assuming a 10% difference (6.25% vs 16.25%) in prolapse rates between the hysterectomy routes to be clinically relevant, and using an alpha value of 0.05, and a beta value of 80%, we estimated approximately 300 patients per group of hysterectomy are needed. With 8 years of study period, we needed approximately 40 patients each year per group to have a representative sample during the study period and meet the sample size requirements.
To compare patient characteristics and surgical details (eg, EBL, complications, cuff suspension, cervix removed) by type of surgery, we used analysis of variance for normally distributed variables (age, BMI), the Kruskal-Wallis test for nonnormally distributed variables (EBL), chi-square tests for categorical variables, and Fisher’s exact tests for categorical variables with small expected numbers. Among patients who experienced posthysterectomy POP, we compared prolapse type, grade, and treatment method by surgery type. The Kaplan-Meier method was used to estimate time to prolapse curves and log-rank tests were used to compare crude survival distributions. In addition, Cox proportional hazard regression was used to calculate hazard ratios (HR) and 95% confidence intervals (CI) for the associations between type of surgery and posthysterectomy prolapse. To determine what factors might confound the association between surgery type and prolapse, we assessed the change in HRs when each patient demographic or surgical characteristic variable was added to the Cox proportional hazard model individually. In addition, we decided to adjust for age (continuous) and BMI (<25, 25–29.9, 30–34.9, ≥35, missing) a priori. To verify the assumption of proportional hazards, we added an interaction term between log transformed time and each predictor. Interaction terms with P values <.05 indicated nonproportional hazards. A sensitivity analysis was run examining the association between surgery type and prolapse after excluding patients with cancer. All analyses were performed using Statistical Analysis System software version 9.4 (SAS Institute, Cary, NC). The study was approved by Partners Institutional Board Review (2014P001869).
We reviewed 2158 charts of women who underwent hysterectomy between 2001 and 2008 for any indication including 1459 abdominal, 375 vaginal, and 324 laparoscopic or robotic cases with a mean age of 51.9 years, 56.3 years, and 49.7 years, respectively. Over the years of the study period, there was a noticeable decline in open abdominal surgeries and an increase in laparoscopic/robotic surgeries. Most women were multiparous and White, with mean BMI approximately 30 kg/m 2 ( Table 1 ). The most common indication for abdominal hysterectomy was cancer (33%), followed by fibroids (24%). For laparoscopic or robotic surgery, the primary indication was cancer (25%) followed by abnormal uterine bleeding (25%). Prolapse was the indication for the index surgery in 60% of vaginal cases and only 2% of abdominal and laparoscopic or robotic cases. The EBL was lowest for laparoscopic or robotic hysterectomy (median=100 mL), and highest for abdominal surgery (median=250 mL). Retention of cervix (subtotal hysterectomy) was more commonly performed with laparoscopic or robotic hysterectomies (42.9%). Documentation of prophylactic vaginal cuff suspension in the operative note was present only in 10.1% of abdominal hysterectomies and 5.9% of laparoscopic or robotic hysterectomies. The majority of patients in the vaginal group had documentation of cuff suspension in the operative note (73.3%) ( Table 2 ).
|Open (abdominal) hysterectomy |
|Vaginal or lap-assisted vaginal hysterectomy (n=375)||Laparoscopic or robotic hysterectomy (n=324)||P value a|
|Age (y) at the time of hysterectomy, mean (SD)||51.9 (11.7)||56.3 (12.1)||48.8 (10.4)||<.0001|
|Parity, n (%)|
|Nulliparous||350 (27.4)||12 (4.0)||78 (25.0)||<.0001|
|1–2||610 (47.8)||156 (51.5)||151 (48.4)|
|≥3||316 (24.8)||135 (44.5)||83 (26.6)|
|Race, n (%)|
|White||1169 (83.5)||301 (84.3)||263 (84.6)||.86|
|Nonwhite||231 (16.5)||56 (15.7)||48 (15.4)|
|BMI, mean (SD) b||30.2 (8.7)||27.6 (6.1)||29.5 (8.6)||.001|
|Year of hysterectomy, n (%)|
|2001||196 (13.4)||44 (11.7)||5 (1.5)||<.0001|
|2002||190 (13.0)||47 (12.5)||2 (0.6)|
|2003||191 (13.1)||49 (13.1)||0 (0.0)|
|2004||192 (13.2)||38 (10.1)||9 (2.8)|
|2005||193 (13.2)||37 (9.9)||16 (4.9)|
|2006||231 (15.8)||69 (18.4)||106 (32.7)|
|2007||127 (8.7)||38 (10.1)||81 (25.0)|
|2008||139 (9.5)||53 (14.1)||105 (32.4)|
|Open (abdominal) hysterectomy |
|Vaginal or lap-assisted vaginal hysterectomy (n=375)||Laparoscopic or robotic hysterectomy (n=324)||P value a|
|Indication for surgery, n (%)|
|Abnormal uterine bleeding||175 (12.0)||58 (15.5)||77 (24.0)||<.0001|
|Fibroids||356 (24.4)||19 (5.1)||71 (21.9)||<.0001|
|Endometriosis or pelvic pain||176 (12.1)||14 (3.7)||55 (17.0)||<.0001|
|Prolapse||34 (2.3)||226 (60.3)||7 (2.2)||<.0001|
|Cancer||488 (33.4)||29 (7.7)||82 (25.3)||<.0001|
|Cesarian hysterectomy||26 (1.8)||0||0||.0005|
|Gastrointestinal involvement||13 (0.9)||1 (0.3)||0||.14|
|Preneoplastic (EIN, CIN)||55 (3.8)||18 (4.8)||21 (6.5)||.09|
|Ovarian benign||117 (8.0)||2 (0.5)||6 (1.8)||<.0001|
|Prophylactic||19 (1.3)||8 (2.1)||5 (1.5)||0.44|
|EBL, median (IQR) b||250 (150–400)||200 (100–350)||100 (50–200)||<.0001|
|Cervix removed, n (%)|
|Yes||1255 (86.0)||374 (100.0%)||185 (57.1)||<.0001|
|No||204 (14.0)||0 (0%)||139 (42.9)|
|Complications (any), n (%)|
|Any complication||85 (5.8)||13 (3.5)||11 (3.4)||.06|
|None||1374 (94.2)||362 (96.5)||313 (96.6)|
|Complications, n (%)|
|None||1374 (94.2)||362 (96.5)||313 (96.6%)||.02|
|Hemorrhagic||19 (1.3)||1 (0.3)||0 (0)|
|Bladder injury||12 (0.8)||2 (0.5)||2 (0.6)|
|Ureteral injury||4 (0.3)||0 (0)||1 (0.3)|
|Bowel injury||20 (1.4)||0 (0)||0 (0)|
|Cardiopulmonary event||10 (0.7)||1 (0.3)||0 (0)|
|Other||20 (1.4)||9 (2.4)||8 (2.5)|
|Detailed cuff suspension, n (%)|
|Yes||148 (10.1)||275 (73.3)||19 (5.9)||<.0001|
|No||1311 (89.9)||100 (26.7)||305 (94.1)|