An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Katina Robison, MD, Kyle Wohlrab, MD, Vivian Sung, MD, Melissa Clark, PhD, Christine Luis, MA, Christina Raker, ScD, Chanelle Howe, PhD, Robin Cram, MPA, Kerri Bevis, MD, Holly Richter, MD, Elizabeth Lokich, MD, Gena Dunivan, MD, Amy Brown, MD, Elena Tunitsky, MD, Stephanie Wethington, MD, Grace Chen, MD, David Rahn, MD, Matthew Carlson, MD, Carolyn McCourt, MD, Jerry Lowder, MD, John Occhino, MD, and Gretchen Glaser, MD.
Author Information and AffiliationsWomen with endometrial cancer have higher rates of stress urinary incontinence (SUI). Concomitant treatment of endometrial cancer and SUI may improve quality of life (QOL). At the time of endometrial cancer diagnosis, women are evaluated by a gynecologist and/or a gynecologic oncologist and undergo surgery within weeks of their diagnosis. Thus, urinary incontinence could easily be identified at this time, a referral made, and concomitant surgery performed if necessary.
Compare the QOL among women with endometrial intraepithelial neoplasia (EIN) or clinical stage I/II endometrial cancer and SUI who chose to have concomitant surgery (cancer plus SUI surgery) with women who chose cancer surgery alone.
A multicenter prospective cohort study was conducted across 8 US sites. Women with EIN or clinical stage I/II endometrial cancer were screened for SUI symptoms. Those who screened positive were offered a preoperative referral to a urogynecologist and offered all treatment options, including concomitant surgery. All women in the study underwent cancer surgery. However, women were distributed into 1 of 3 groups regarding incontinence treatment: concomitant cancer/SUI surgery, nonsurgical SUI treatment, or no SUI treatment. QOL for all women was assessed at baseline, 6 weeks, 6 months, and 12 months postoperative using the Functional Assessment of Cancer Therapy-Endometrial (FACT-En). This general QOL instrument specifically designed for patients with endometrial cancer does not include any incontinence measures. Multivariable Poisson regression with generalized estimating equations was used to examine the relationship between SUI treatment group and high QOL (FACT-En score greater than overall median score). Adjusted relative risks (RRs) and 95% CIs were reported.
Of the 1322 women screened, 702 (53.1%) screened positive for SUI. A total of 88 women declined participation; 58 were ineligible before signing consent due to language barriers or mental illness, or opting not to have surgery (ie, a hysterectomy). Overall, 556 women were enrolled, 7 were ineligible, 10 withdrew, and 44 were missing baseline data, leaving 495 evaluable participants. Of these 495, 106 (21.4%) women chose concomitant cancer and SUI surgery, 91 (18.4%) chose nonsurgical SUI treatment with cancer surgery, and 298 (60.2%) chose cancer surgery alone. For all groups, median FACT-En scores increased from baseline through 12 months (78 points vs 84 points). Adjusting for demographics, clinical measures, and baseline SUI severity and QOL, concomitant urogynecologic surgery was associated with having QOL (as measured by the Fact-EN) above the overall study population median score compared with nonsurgical SUI treatment (RR, 1.19; 95% CI, 0.97-1.45; P = .093) and with no SUI treatment (RR, 1.20; 95% CI, 1.01-1.43; P = .038). These results were uncertain for the concomitant SUI surgery compared with the nonsurgical SUI treatment, but the point estimate is in the direction of favoring concomitant surgery. However, women in the concomitant surgery group were 20% more likely to have adjusted cancer-related QOL scores above the overall study population median score after surgery, compared with women who had cancer surgery alone (RR, 1.20; 95% CI, 1.01-1.41).
Women with endometrial cancer have high rates of SUI. Women who received treatment of SUI at the time of cancer surgery had a higher cancer-specific QOL compared with women who had no SUI treatment. The impact on cancer-specific QOL was uncertain for the concomitant SUI surgery group compared with the nonsurgical SUI treatment group. Gynecologists and gynecologic oncologists should screen patients for SUI at the time of endometrial cancer diagnosis and discuss the possible benefits of concomitant surgery.
This was a prospective cohort study that intrinsically had confounding and selection bias associated with it. Additionally, more women than expected chose not to have concomitant surgery, so our study was slightly underpowered.
Endometrial cancer is the most common gynecologic malignancy and the fourth most common cancer among women in the United States. In 2020, a total of 65 620 women were diagnosed with endometrial cancer.85 A woman's lifetime risk of developing endometrial cancer is 3%, and the incidence is continuing to rise.1 Overall, 75% of cases are diagnosed at stage I and another 10% at stage II (confined to the uterus). Among women with disease confined to the uterus, there is a 95% five-year survival rate.2 Treatment of early-stage endometrial cancer is primarily surgical, with removal of the uterus, ovaries, and fallopian tubes.3 More than 75% of women with stage I or II disease are considered cured with surgery alone, and many women undergo minimally invasive surgery with a rapid return to their normal activities of daily living. Therefore, it is an ideal time to identify and possibly treat other conditions that might impact a woman's quality of life (QOL) after cancer treatment.3
Female urinary incontinence is most prevalent among middle-aged women, and the prevalence increases with age.4 The fear of urinary leakage may cause a woman to avoid exercise, sexual intercourse, or even social activities.4 Stress urinary incontinence (SUI) is correlated not only with reduced QOL but also with depression.5,6 Endometrial cancer and SUI share common risk factors including obesity, diabetes, and pulmonary disease. Obesity is associated with endometrial cancer with an estimated relative risk (RR) of up to 5 and is also associated with urinary incontinence with an estimated odds ratio (OR) of 4.2.3,7 These common risk factors may explain the higher rates of SUI among women with endometrial cancer compared with the general population. Erekson et al found SUI in 80% of women with endometrial cancer, which is in stark contrast to the 30% to 40% of women who report SUI symptoms in the general population.11 An additional study reported that 49.9% of women with endometrial cancer had SUI before treatment of their cancer.8
Adding to the problem, the treatment of endometrial cancer may also be associated with higher rates of SUI.10 Although the data are conflicting, hysterectomy has been associated with the development of SUI.11-14 Furthermore, approximately 15% to 20% of women diagnosed with early-stage endometrial cancer will undergo some form of radiation therapy. A randomized controlled trial evaluating the use of postoperative radiation therapy (PORTEC-1) among women with endometrial cancer reported an increased rate of urinary urgency and incontinence at the 15-year analysis.15 There has been interest in identifying preventive measures that could maintain pelvic floor function to diminish the added effects of radiation and age after endometrial cancer treatment; to date, none have been studied.16
Although SUI is very common, nearly 40% of women with severe urinary incontinence have not discussed their symptoms with a provider.17 In a large US population study, only 25% of women with urinary incontinence sought care.18 In addition, the typical time that a woman waits before she presents for management of her “bothersome” complaints of incontinence is 1 to 6.5 years.19 Therefore, if SUI screening is not conducted at the time of endometrial cancer diagnosis, a woman may suffer for many more years.
Gynecologic cancers have significant psychosocial morbidity associated with them, and longer survivorship periods are not necessarily associated with reduced distress.27 Those women with poorer physical and mental health before diagnosis will likely experience the greatest emotional distress after diagnosis.28 Treatment for gynecologic cancer often has adverse side effects, including hormonal, physical, and vaginal changes.29,30 Such changes are likely to impair the physiological mechanisms underlying sexual responses, resulting in sexual disruptions and reduced QOL.31 Among women with early-stage endometrial cancer, the median score for the Female Sexual Function Index (FSFI) was 16.6 (range, 0-32.8); scores <26 are diagnostic for sexual dysfunction.32-34
Another important factor influencing both physical and mental health among endometrial cancer survivors is physical activity. Endometrial cancer survivors who have increased levels of physical activity report feeling better physically and emotionally.35 SUI may decrease a woman's level of exercise due to the fear of urinary leakage.4 Treatment of SUI can improve a woman's physical activity.36 Therefore, treating SUI among endometrial cancer survivors who have multiple other risk factors for poor physical activity can not only improve a woman's level of physical activity but also decrease emotional distress.28
SUI and endometrial cancer are often curable with minimally invasive surgery. In 1973, David Nichols popularized the gauze hammock sling technique to cure SUI.22 Midurethral sling procedures have been shown to be highly effective, with continence rates between 85% and 88%.21,22 The intermediate- and long-term results suggest that the 10-year continence rate is similar to the 1-year continence rate.23 In fact, it appears that sling procedures that are effective after 6 months are likely to remain effective for many years.24 In addition, endometrial cancer surgery is now treated predominantly by laparoscopic or robotic-assisted hysterectomy. The Gynecologic Oncology Group (GOG) performed a randomized controlled trial among women with endometrial cancer comparing total laparoscopic hysterectomy with abdominal hysterectomy among women with endometrial cancer and found no difference in survival or outcomes, supporting the safety of laparoscopically staging endometrial cancer.25,26,64 Therefore, it appears feasible to perform minimally invasive procedures for both endometrial cancer and SUI, potentially improving overall QOL.
Unfortunately, it has been reported that bowel and bladder symptoms were only identified by the oncologist in 5% to 15% of survivors of gynecologic cancer, while as many as 23% to 58% of these women self-reported severe bladder and bowel symptoms.33 In one of the few available studies of treatment of bladder symptoms in women with cancer, most women reported that they would undergo treatment for pelvic floor symptoms if this was proposed and their provider felt it was efficacious. Unfortunately, 68% of women with severe incontinence symptoms did not seek help, citing reasons such as “it's not as bad as cancer” or that the provider did not engage them about these symptoms; some women were embarrassed or reluctant to be seen, or they were unaware of possible treatments and adjusted their behavior to make their symptoms bearable.33
Among women without cancer, QOL is improved following SUI treatment.38 This study focused on QOL issues in a cancer population and offered treatment for SUI at the time of cancer diagnosis. However, some women may prefer to have nonsurgical treatment of their SUI or defer treatment of SUI. This study evaluated the impact of concomitant SUI and cancer surgery on QOL, as well as the impact of nonsurgical SUI treatments on QOL. The findings of this study will help providers identify which SUI, cancer, and personal factors are important for women in making decisions about concomitant vs nonconcomitant surgery. Providers can use this information to counsel women about the risks and benefits of surgical options.
Endometrial cancer is the most common gynecologic malignancy in the United States. Most women are diagnosed at an early stage (I or II) resulting in a 5-year survival rate of 95%.85 Treatment is typically surgical, and most women are cured with surgery alone. Therefore, this is an important time to identify and possibly treat factors that can impact a woman's QOL after endometrial cancer treatment.
SUI is the leakage of urine with activities that increase abdominal pressure. SUI occurs frequently from middle age onward and is associated with a reduced QOL and decreased sexual function. Rates of SUI among women with endometrial cancer have been reported as high as 83%.3 SUI is treatable with a midurethral sling placement resulting in the resolution of incontinence in 85% of women.9
At the time of endometrial cancer diagnosis, women are seen by a provider, and most undergo surgery within weeks of their diagnosis. SUI could be identified at that time, and concomitant SUI and cancer surgery could be performed. This could decrease distress and the personal costs associated with SUI, potentially improving QOL. However, currently few providers screen for SUI at the time of endometrial cancer diagnosis, and few patients ask about concomitant surgery. Most of the time, patients with both endometrial cancer and SUI undergo 2 separate surgeries, 1 surgery for their endometrial cancer and a second surgery at some subsequent time for their SUI; or they do not have surgery for SUI.
The purpose of this study was to compare QOL among women with endometrial cancer and SUI who undergo concomitant surgery with women who do not have concomitant surgery (ie, women who have cancer surgery only, with either nonsurgical SUI treatment or no SUI treatment). In addition, this study determined differences in the clinical outcomes of women with SUI who have concomitant surgery compared with those who do not.
Aim 1: Compare QOL scores of women with SUI and endometrial cancer who have concomitant surgery with women who have cancer surgery only.
Aim 2: Compare the sexual function of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Aim 3: Compare the clinical outcomes of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Aim 4: Determine the characteristics of women who choose concomitant surgery compared with those who choose cancer surgery only.
Secondary aim: Compare the rate of midurethral sling complications among women who received radiation therapy after sling placement with women who did not receive radiation therapy after sling placement.
Women who had a history of early-stage endometrial cancer and who answered yes to the screening question, “Do you leak urine when you cough, jump, sneeze, or laugh?” were recruited to participate in four 60- to 90-minute focus groups at Women & Infants Hospital (WIH) and the University of Alabama at Birmingham (UAB) Medical Center. Focus group discussions were moderated by a trained facilitator. The purpose of this proposal was to engage women with endometrial cancer and SUI to refine our larger study design and further discuss the information women would like to know in order to make a decision regarding concomitant SUI and cancer surgery. We used the information obtained in these focus groups to finalize our outcome measures that were used in our large cohort study.
We convened a panel of 17 stakeholders: 5 gynecologic oncologists, 5 urogynecologists, 1 general obstetrician/gynecologist, 1 oncology case manager, and 5 patients. The case manager on the panel is also a survivor of endometrial and ovarian cancer. The panel consisted of 14 women and 3 men. There were 2 African American stakeholders and 1 Asian stakeholder. The group was chosen based on geographic location, medical setting diversity, and interest in the research question. There were clinician representatives from diverse settings, from private practice to large cancer centers in the United States. This diversity helped make the research findings of this study more generalizable. In addition, the panel provided scientific and practice-based feedback in the design and implementation of the study.
During our first few conference calls with our stakeholders, we reviewed the results of our 4 focus groups, discussed questionnaires to be used as measures of primary outcomes, determined tools to eliminate confounding variables, and reviewed data collection forms. Participants at our 4 focus groups discussed a preference for the term “urinary leakage” in lieu of “stress urinary incontinence,” and they expressed an interest in screening for SUI at multiple times during their first oncology visit. Patients in our focus groups asked for this because they felt overwhelmed with the nature of their cancer, and they wanted to have time to think about their incontinence. Therefore, in our cohort study, patients received a phone call before their visit introducing the study and letting them know that they could be eligible for a study, which they would be approached about at their upcoming appointment. The patient also received a screening sheet in the waiting area that contained the screening question, and the patient was able to answer the question by circling “yes” or “no.” The patient was later asked the screening question again by the provider or research assistant (RA).
Patients in our focus groups also voiced their support of the option of concomitant surgeries. Patient stakeholders agreed with participants in the focus groups that being made aware of the opportunity for a urogynecology referral before their initial clinic visit would be beneficial. Stakeholders reviewed the proposed questionnaires to determine the primary outcome variables of the study. It was determined that the Functional Assessment of Cancer Therapy-Endometrial (FACT-En), the FSFI, the Pelvic Floor Distress Inventory (PFDI), the Pelvic Floor Impact Questionnaire (PFIQ), and a reporting form on surgical and clinical outcomes would be used. Stakeholders also discussed the need to evaluate potential confounding variables using the Sandvik Severity Index (SSI) and the Life Orientation Test (LOT). As a group, stakeholders made suggestions to the data collection forms, including questions on finances associated with the surgeries and the timeline for follow-up visits.
Stakeholders suggested ways to improve recruitment at our UAB Medical Center site, which had a slow start to recruitment. One stakeholder suggested that the UAB study team send emails directly to providers to remind them of patients who might be eligible for this study. Other suggestions included talking about the study at Tumor Board and identifying patients to approach in this setting while more actively involving providers in the screening of potential patients. Given to the patient in the waiting area, a screening sheet with the question “Do you ever leak urine when you cough, jump, sneeze, or laugh?” was thought to help identify future patients as well. We also discussed adding more sites to help increase recruitment numbers and meet our recruitment goal. These suggestions greatly improved our recruitment strategies.
The stakeholders in this study played a pivotal role throughout the study. Women from our pilot study who served as patient stakeholders participated in a discussion regarding the pilot study, the outcomes they felt were important to measure, and the ideal time to screen for SUI when diagnosed with endometrial cancer. Women with endometrial cancer and SUI also contributed to the study through participation in focus groups. The clinician stakeholders engaged in the study design; they reviewed the specific aims, defined the clinical outcomes of interest, and discussed their own experience with SUI screening and referrals. All members of the stakeholder advisory panel continued to be engaged in the refinement of the study design, implementation, and dissemination. In addition, the clinician stakeholders contributed to discussions about barriers to screening, referring patients to providers who perform incontinence surgery, and coordinating concomitant surgery so that our findings would be relevant for different types of care settings in which coordinating concomitant surgery would be more vs less difficult. Stakeholders provided useful insight to ensure that gynecologists, gynecologic oncologists, urogynecologists, and urologists from varying types of practices would have the data necessary to make informed decisions about performing concomitant cancer and SUI surgery and suggestions for successful methods of coordinating care.
The proposed study was an observational cohort study to compare the QOL and clinical outcomes among women with endometrial cancer and SUI who have concomitant surgery with women who do not have concomitant surgery (ie, women who have cancer surgery only, with either nonsurgical SUI treatment or no SUI treatment). At the first gynecologic oncology visit, women with documented endometrial cancer or endometrial intraepithelial neoplasia (EIN) were screened for SUI with a single question: “Do you ever leak urine when you cough, sneeze, jump, or laugh?” This screening question was chosen because it has been shown to have high likelihood of identifying SUI. It has also been shown to have high interobserver agreement (κ = 0.8; 95% CI, 0.3-0.9) in diagnosing SUI.2 Women who screened positive for SUI by answering “yes” to the screening question were asked by an RA to participate in the study. Eligible and interested participants were consented. After signing an informed consent, patients completed a validated QOL survey specific to endometrial cancer patients (ie, the FACT-En) and validated incontinence surveys.
All participants were offered referral to urogynecology at their initial visit with their oncologist. If they accepted referral, they were seen by a urogynecologist and completed an SUI work-up within 2 weeks of their initial oncology visit. If the patient desired surgery and was deemed a candidate for a surgical intervention for SUI, concomitant surgery was offered by the urogynecologist. Postoperative follow-up visits for all participants were conducted in the usual clinical fashion at 6 weeks, 6 months, and 12 months with an oncologist. At each of these visits, participants answered the same questionnaires they completed at baseline, with the exception of the FSFI at 6 weeks. We compared each woman's QOL scores before and after treatment across surgery groups to control for individual differences in QOL at baseline (ie, time of diagnosis). Comparing women who chose to have concomitant surgery with those who did not provided us with information about the impact of SUI treatment on an endometrial cancer survivor's overall QOL and clinical outcomes. We followed all enrolled women from the time of diagnosis of endometrial cancer/EIN up to 12 months. We controlled for multiple factors associated with choice of surgical options, including SUI severity and bother scores, optimism, cancer grade and stage, medical comorbidities, and sociodemographic factors.
Women were recruited from 8 sites: (1) WIH in Rhode Island; (2) the Gynecologic Oncology Program at UAB Medical Center; (3) Hartford Hospital and Hospital of Central Connecticut within the Hartford HealthCare (HHC) System in Connecticut; (4) University of New Mexico (UNM) Hospital in Albuquerque; (5) Washington University in St. Louis (WUSTL) Hospital in St. Louis, Missouri; (6) University of Texas Southwestern (UTSW) Medical Center in Dallas; (7) the Mayo Clinic (MAYO) in Rochester, Minnesota; and (8) Johns Hopkins University (JHU) Hospital in Baltimore, Maryland.
Women were eligible if they (1) had clinical stage I or II endometrial cancer (disease confined to the uterus determined by physical examination), endometrial carcinoma, or EIN (also known as complex atypical hyperplasia [CAH]); (2) had a treatment plan of hysterectomy; and (3) answered “yes” to 1 screening question: “Do you ever leak urine when you cough, jump, sneeze, or laugh?” We included women with CAH or EIN because 40% of these women have an endometrial cancer identified at the time of surgery, and the risk factors are the same for EIN as they are for clinical stage I and II endometrial cancers.1 We included only women with EIN or clinical stage I and I cancer for the following reasons: (1) The prognosis for clinical stage III and IV endometrial cancer is much worse and has a high emotional burden associated with the diagnosis; and (2) women with clinical stage III or IV endometrial cancer at the time of diagnosis often do not have surgical treatment for their endometrial cancer and instead have chemotherapy. Women who were thought to have clinical stage I or II at the time of enrollment but were later found to have pathologic stage III or IV endometrial cancer by a computed tomography scan or at the time of surgery remained in the study. We included only women with early clinical stage endometrial cancer because we expected that they would be counseled about their excellent prognosis and would be more likely to enroll in a study focused on improving their QOL after cancer treatment. Finally, we thought that including only women with early clinical stage endometrial cancer would reduce potential confounding bias because some practitioners might feel that treating the patient's SUI would affect cancer treatment type or timing and would discourage them from participating in the study.
Patients who spoke either English or Spanish were eligible to participate in this study. A Spanish version of the consent form was available for patients who spoke only Spanish; it was translated into Spanish by the professional translating services at WIH. Spanish-speaking patients were also provided with Spanish versions of the study surveys. Validated Spanish versions of the study surveys were found online, with the exception of the SSI, which was translated by the hospital's professional translating services because a Spanish version of the survey did not exist. All Spanish versions of the surveys were approved by the IRB. Each site had bilingual staff who spoke Spanish to help answer any questions for patients.
Women were excluded if they had (1) screened negative for SUI, (2) clinical stage III or IV endometrial cancer upon initial physical examination, (3) prior pelvic radiation therapy, (4) vesicovaginal fistula, or (5) urethrovaginal fistula. We excluded women with vesicovaginal fistula or urethrovaginal fistula due to the risk of infection when placing a midurethral sling. Women were not eligible if they were not having a hysterectomy.
For any patient who wished to decline, the reason for declining was captured and documented in our study database. Reasons for decline included that the patient did not want to participate in research, lack of time, reason not given, or other.
Women with SUI and endometrial cancer or EIN had the option to have concomitant cancer and SUI surgery, have nonsurgical treatment of their SUI, or have no treatment of their SUI. Patients answering “yes” to the SUI screening question were offered referral to urogynecology. At the discretion of the urogynecologist, patients were evaluated for urinary leakage by a history of symptoms, physical examination, and, if deemed appropriate, simple cystometrograms and complex multichannel urodynamic testing. Women were counseled and offered treatment for SUI ranging from conservative to concomitant anti-incontinence surgery at the discretion of the urogynecologist. About half of women who deferred concomitant anti-incontinence surgery reported urge-predominant symptoms or “SUI not severe enough” as their reasons (not mutually exclusive). Women with mild symptoms may have been managed conservatively according to the discretion of the consulting urogynecologist. For the present analysis, we defined treatment groups based on the SUI treatment received at the time cancer surgery was completed. This definition accommodated patients for whom definitive cancer surgery treatment was delayed due to advanced disease but who ultimately had concomitant surgery for cancer and SUI. Table 1 provides details on each group definition.
Description of Treatment Comparators.
We had 3 outcome measures: QOL, sexual function, and clinical outcomes (Table 2).
Outcome Measures.
Historically, cancer therapies have focused on curing the cancer. However, as cure rates have improved, we are now faced with the consequences of our treatments and the long-term outcomes of women surviving cancer. Our primary outcome, QOL, was measured using the FACT-En. This is a general QOL instrument specifically designed for patients with endometrial cancer. This instrument has been validated for measuring physical, functional, and emotional well-being.5 The total score is based on 43 items, including physical, social, emotional, and functional subscales, with an additional section of concerns specific to endometrial cancer. Higher scores on this measure indicate a better QOL.
As a second QOL measure, we used the PFIQ-7/Form E, a shortened version of the PFIQ used to assess life impact in women with pelvic floor disorders.39 It includes all of the Incontinence Impact Questionnaire-7 as well as items related to other pelvic floor disorders. The PFIQ-7 consists of 3 scales of 7 questions each taken from the Urinary Impact Questionnaire, the Pelvic Organ Prolapse Impact Questionnaire, and the Colorectal-Anal Impact Questionnaire. The 3 scales are scored from 0 (least impact) to 100 (greatest impact) and an overall summary score (0-300). The PFIQ-7 gives a comprehensive assessment of the effect of pelvic floor disorders on the QOL of women, rather than only assessing 1 aspect of pelvic floor function such as urinary incontinence.
Sexual function was measured using the FSFI/Form F. It is a 19-item instrument assessing 6 domains of sexual function: desire, arousal, lubrication, orgasm, satisfaction, and pain. These domains have ordinal Likert-type response formats and are scored from 0 to 5. The scoring algorithm sums items on each domain/subscale and then scales the sums so that each subscale has a maximum score of 6. The FSFI total score is the sum of the 6 domain/subscale scores and has a maximum score of 36. Higher scores indicate better sexual functioning. A total score of 26 has been validated as a cutoff score for diagnosing female sexual dysfunction.8 The original validation study established the FSFI's reliability and construct validity in women diagnosed with sexual arousal disorder and in women without sexual difficulties.9 The FSFI has also been found to be a reliable and valid instrument for measuring cancer-related female sexual dysfunction among cancer survivors.10
The clinical outcomes we measured are listed in Table 3. These measures were chosen by reviewing the clinical outcomes of interest in prior GOG surgical trials and midurethral sling trials11,12 and were measured using hospital records. There are some data on the clinical outcomes of women without cancer having concomitant hysterectomy for uterine prolapse and a prophylactic midurethral sling placement.13 Women with the prophylactic midurethral sling placement had slightly higher rates of bladder perforation, urinary tract infection, major bleeding complications, and incomplete bladder emptying compared with women having only a hysterectomy, but the decrease in SUI was significant enough for the authors to recommend consideration of prophylactic midurethral sling placement for women having a vaginal hysterectomy for prolapse.13 However, little data are available regarding the clinical outcomes among women with cancer having concomitant surgery. To measure postoperative pain, we also used the modified Brief Pain Inventory (BPI) scale, which contains 3 severity items and 5 interference items. The 3 pain severity items are “worst pain,” “pain on the average,” and “pain right now,” and the 5 interference items are “walking ability,” “mood,” “sleep,” “relations with others,” and “ability to concentrate.” The BPI has been validated among surgical patients with cancer and used in several clinical trials.11,14-17
Primary Clinical Outcomes.
Patient satisfaction and regret regarding treatment decision was measured by using 2 scales: the Satisfaction With Decision (SWD) Scale and the Decision Regret Scale (DRS)/Form H. The SWD is a 6-item scale measuring satisfaction with health care decisions, developed and validated in the context of women making decisions about hormone replacement therapy, and subsequently validated in adults with depression making decisions about treatment. The scale has good internal consistency reliability (α = .85), evidence of construct validity, relevance to designing and assessing patient-centered decision support interventions, and is sensitive to changes in information in trials of decision aids. The scale uses a 1 to 5 rating (1, strongly disagree; 5, strongly agree). Higher scores indicate higher satisfaction with the decision. Decisional regret was measured using the DRS. This 5-item scale is a reliable and valid indicator of health care decision regret at a given point in time, with excellent psychometric properties. In this study, the DRS was used to measure patient regret with the treatment that was initially chosen. Both the SWD and the DRS were used at the 12-month visit.40-42
There are 2 mesh-related complications: sling exposure and sling erosions. Sling exposures typically occur within 6 weeks postoperatively, and there is no increase in erosion rates among women undergoing concomitant hysterectomy and midurethral sling placement.13 There is 1 case report of a sling erosion in a woman with vulvar and anal cancer who had undergone radiation therapy before a midurethral sling placement.18 We defined mesh exposure as presence of the mesh at the suture line during the postoperative period and mesh erosion as the presence of mesh eroding into a surrounding structure (eg, the vagina, urethra, or bladder). At the time of the examination, surgeons noted if there was an exposure or erosion, the location(s), approximate amount of mesh exposed/eroded, and action taken (eg, observation, topical estrogen therapy, surgical revision).
The target sample size was 556 women. This target sample size was based on a continuous FACT-En score as the primary outcome and a comparison of women who did and did not choose concomitant SUI surgery as the exposure variable, assuming that during statistical analyses a linear regression model with generalized estimating equations (GEEs) would be fit. The target sample size calculation was performed using the GEESIZE v.3.1 macro with the specification of a 2-sided α = .05, a β = .2 to correspond to 80% power, an exchangeable working correlation structure (correlation, 0.5), and an approximately 10% increase in the sample size obtained from the GEESIZE v.3.1 macro (ie, 504) to account for potential reduced explanatory power due to covariate adjustment and attrition. The target sample size calculation assumed that half of the eligible women would choose concomitant SUI surgery and half of the eligible women would not choose concomitant SUI surgery. The target sample size calculation also assumed a 5-point difference between women who did and did not choose concomitant SUI surgery in the mean change in FACT-En score from baseline to each specified follow-up visit. Although the minimal clinically important difference for FACT-En has not been established, this effect size approximates the 5% difference in FACT-En score (on a 0-100 scale) considered clinically relevant in a QOL study for patients with endometrial cancer.28 Informed by pilot data, the target sample size also assumed the standard deviation for change in FACT-En score would be 20 points.
Data collectors were trained in person and/or via phone conferences. Manual operating procedure and standard operating procedure documents were created by the main site, at WIH, and used to train each of the sites to ensure data quality and data entry standardization. Data managers at WIH sent data queries weekly to ensure that data were being properly collected and entered.
Phone calls were made by the RA to the patient at 1 week postoperative, as well as during follow-up at 3 months (between their 6-week and 6-month visit) and at 9 months (between their 6- and 12-month visit). These phone calls helped reduce attrition by reminding patients of their upcoming visits, allowed patients to ask any questions, and helped study staff maintain a rapport with the patient over time.
If a participant wished to withdraw from the study, she would indicate this in writing. The research team filed the letter in a locked cabinet in the research office, and the reason for withdrawing was recorded in the database. Participants were reminded that there was no penalty for choosing not to participate or for withdrawing from the study and that their decision to participate did not impact the standard of care they received. Participants were considered lost to follow-up if they did not complete the 12-month visit. They remained in the study even if they did not complete visits at an earlier time point.
Most questionnaires were completed by iPad or email. Only 9 participants opted for the Spanish versions of the surveys. More than 94% of surveys were sufficiently completed for scoring by standard protocols, with the exception of the FSFI (<40% complete). We sampled approximately 20 of each survey with high scores, low scores, or missing items for scoring confirmation by 2 research staff members. Internal consistency by Cronbach α was >.7 for subscales at all time points, with the exception of the PFDI prolapse and bowel subscales (range, 0.67-0.84).
Demographics and clinical confounders are listed in Table 4. SUI severity was measured using the SSI/Form B. The SSI is a validated measure that assesses urinary incontinence severity using 2 questions quantifying how often a woman leaks and the volume of urine leakage with each episode. The scores range from 0 to 12.19 SUI bother was measured using the PFDI/Form C. The PFDI includes 45 questions covering 3 scales: Urinary Distress Inventory (UDI), Pelvic Organ Prolapse Distress Inventory, and Colorectal-Anal Distress Inventory. Each of the 3 scales has subscales scored from 0 (least distress) to 100 (greatest distress).
Potential Covariates and Confounders.
For each aim, relevant descriptive statistics (eg, proportions, medians, and means) were calculated to examine the distribution of demographic factors, clinical characteristics, and incontinence severity at baseline. Some continuous variables, such as body mass index (BMI), were categorized for ease of interpretation. Categorical variable levels with small cell counts were combined to lessen sparse data. Missing data patterns were examined, and potential bias from missingness was minimized using regression adjustment as appropriate. The completeness of baseline and follow-up survey data and attrition due to death or dropout was compared by surgical treatment group, study site, demographics, and clinical variables.
The pool of potential covariates and confounders was narrowed after reviewing the results of aim 4, predictors of urogynecology referral acceptance, alongside clinical and statistical considerations. Clinical measures that were not routinely collected at all study sites, such as performance status, were removed from further evaluation. For variables that appeared redundant (eg, study site and region) or had a high possibility of multicollinearity (eg, multiple incontinence severity scales and subscales), subject matter knowledge was used to select the most appropriate variables for each aim's analysis. Scales measuring SUI severity, impact, or bother were both more strongly related to treatment group in the study population and considered to be more relevant to the main outcomes, compared with other urinary or pelvic floor symptom scales. The UDI stress subscale was selected instead of the SSI based on feedback from the urogynecologists.
Categorical variables were compared by Fisher exact test, and continuous or ordinal variables were compared by Wilcoxon rank sum test. All statistical tests and 95% CIs were 2 sided. SAS v.9.4 (SAS Institute) and STATA v.15 (StataCorp) were used for data analysis.
Compare QOL scores of women with SUI and endometrial cancer who have concomitant surgery with women who have cancer surgery only.
We compared QOL scores between women who received concomitant SUI and cancer surgery and women who received cancer surgery only. A second comparison divided the latter group into nonsurgical SUI treatment and no SUI treatment. The primary QOL outcome, FACT-En score, was examined as a continuous variable at baseline and each postsurgery time point (ie, 6 weeks, 6 months, and 12 months). Mean postoperative scores were compared between SUI treatment groups by fitting a GEE multiple linear regression model. This regression modeling strategy allowed for estimation of between-group mean differences in scores while accounting for within-participant correlation of outcomes. Participants with some missing outcome data are included under the assumption of missing completely at random.43 An independent working correlation matrix was specified for the within-participant correlation in outcomes. The quasi-likelihood under independence model criterion was used to evaluate model fit and effect modification by time. Furthermore, regression adjustment was used to minimize the effect of potential sources (ie, clinical and demographic characteristics measured at baseline) of confounding bias and selection bias due to study attrition. Covariates included demographics, BMI, comorbidities, study site, baseline SUI severity (UDI stress subscale), baseline optimism scores (LOT-Revised [LOT-R]), and cancer treatment as a proxy for disease severity. Baseline FACT-En score was also included to account for baseline differences. Examination of the linear model residuals indicated deviations from multivariate normality (P < .05 by Royston multivariate normality test), suggesting that a normally distributed marginal mean model was not the most appropriate model for FACT-En scores. To contrast high vs low scores for subsequent analyses, postoperative QOL score was dichotomized at the overall study population median value at each time point, and the relationship between SUI treatment and QOL was examined using a GEE modified Poisson regression model. The Urinary Impact Questionnaire, Short Form 7 (UIQ-7) subscale of the PFIQ survey was examined as a secondary QOL outcome. Scores were heavily skewed toward 0 (no urinary impact) at follow-up visits; therefore, scores were dichotomized (>0 vs 0) and examined by GEE modified Poisson regression as described for FACT-En scores. Associations between SUI treatment and QOL measures were summarized by adjusted RRs, with values >1 indicating higher QOL.
Compare the sexual function of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Sexual function was compared between women who received concomitant SUI and cancer surgery and women who received cancer surgery only. A second comparison divided the latter group into nonsurgical SUI treatment and no SUI treatment. FSFI scores could not be computed for half of the women with incomplete items or responses indicating no recent sexual activity. As described by Baser et al,53 the validity of the FSFI for women without recent sexual activity is questionable. Therefore, women without recent sexual activity were included as a separate group. Women with incomplete response were categorized in the “no recent sexual activity” group under the assumption that incomplete responses indicated not applicable responses. To explore FSFI response patterns at baseline, we compared participant characteristics with the Fisher exact test or the Wilcoxon rank sum test. For the longitudinal analysis at 6 and 12 months, the primary outcome was examined as a 3-group variable: (1) no recent sexual activity, (2) sexual dysfunction (FSFI score ≤26.55), and (3) normal sexual function (FSFI >26.55). GEE multinomial logistic regression models were fit with the clinical and demographic covariates described for aim 1.
Compare the clinical outcomes of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Clinical outcomes were compared between women who received concomitant SUI and cancer surgery and women who received cancer surgery only, with and without adjustment for the baseline clinical and demographic variables described under aim 1. Multiple linear regression models were used to estimate the relationship between surgery group and the following outcomes: blood loss, operative time, and pain scores. Blood loss was natural logarithm-transformed to normalize the residuals. Multiple logistic regression was used to examine the relationship between surgery group and hospital stay >1 day. Time from oncology visit to cancer surgery was examined as a measure of treatment delay. A Cox proportional hazards model was used to quantify the relationship between surgery group and days until surgery. Clinical and demographic covariates were included as described for aim 1. Complications, intraoperative injury, transfusion, and urinary retention were too infrequent for multivariable regression modeling. Because this aim examined several outcome variables, the possibility of false positives (type I errors) was minimized by hypothesis testing based on the Benjamini-Hochberg procedure for an α = .05 and the observed P values. This procedure entails ranking the observed P values in ascending order and comparing them with critical values that account for rank, the total number of tests, and the overall false-discovery rate of .05. Observed P values less than these critical values were considered statistically significant.
Determine the characteristics of women who choose concomitant surgery compared with those who choose cancer surgery only.
A prediction model was developed to identify the clinical and demographic characteristics of women with SUI that influence acceptance of SUI evaluation and treatment at the time of endometrial cancer diagnosis. The primary outcome for model development was acceptance of referral to the urogynecology clinic at the time of consent and baseline questionnaire completion. Participants who initially declined referral but later changed their mind were counted as accepting referral as long as their change of mind occurred before surgery. Concomitant SUI surgery vs cancer surgery alone was examined as a secondary outcome. We used multiple logistic regression to identify independent predictors of referral acceptance. Potential predictors and referral acceptance were first compared with the Fisher exact and Wilcoxon rank sum tests. Variables associated with referral acceptance with a P < .1 were considered for the multivariable model. The linearity assumption for continuous predictors was evaluated by the Akaike information criterion (AIC) comparing models with linear terms and spline terms. Restricted quadratic splines (RQS) were created with the RQS macro in SAS.87 Interactions between predictors were assessed by the AIC. For nonlinear predictors, we plotted the relative odds of referral acceptance against predictor values and estimated ORs and 95% CIs for contrasts between specific values of the predictor (eg, UDI stress subscale score). Model discrimination was evaluated by area under the receiver operating characteristic curve, and model calibration was examined by the Hosmer-Lemeshow goodness-of-fit test and plotting expected vs observed probabilities.
Compare the rate of midurethral sling complications among women who received radiation therapy after sling placement with women who did not receive radiation therapy after sling placement.
Exact binomial or Poisson 95% CIs were estimated due to the small sample sizes. To account for differences in the timing of radiation therapy initiation and variable follow-up lengths, we computed person-time separately by radiation therapy status. Person-time with radiation therapy was defined as the days from initiation of the first radiation treatment until the end of follow-up (attrition or end of study). Person-time without radiation therapy was defined as the days from surgery until initiation of the first radiation treatment or the end of follow-up (for those not treated with radiation).
During the first 6 months of the project, before the start of recruitment, we made minor changes to the study protocol with respect to the instruments that the study would use. These changes were based on the feedback we received during our conference calls with our panel of stakeholders. After stakeholders reviewed the proposed questionnaires, we decided to use the PFDI and PFIQ-7 questionnaires, instead of the UDI and Incontinence Impact Questionnaire, to better assess prolapse and to control for prolapse symptoms as a confounder. Later, we added an SWD Scale and DRS to be able to measure patient satisfaction and regret regarding the initial treatment decision. The other minor change we made was removing the FSFI questionnaire from our 6-week visit. Because patients are not sexually active 6 weeks postsurgery, we found that most patients were not answering this questionnaire, as it did not apply to them at that given time point. After discussing it with the research team, we decided to remove this questionnaire from the 6-week time point. The protocol was amended each time to reflect these changes.
Our goal was to enroll a total of 556 patients across all sites within the 3-year time frame of the study. At the beginning of recruitment, WIH's recruitment was a little behind anticipated, which was entirely associated with patient volume at the site. This was lower than expected because fewer gynecologic oncologists than anticipated were in active practice. Also, due to IRB delays and recruitment challenges at both WIH and UAB Medical Center at the beginning of the study, we fell behind our recruitment goal. To keep up with the milestones set for this study and recruitment goals, we added 6 additional sites including HHC, UNM, WUSTL, UTSW, MAYO, and JHU. Adding these sites not only optimized recruitment but increased the overall generalizability of the study as well. We updated the study protocol to reflect the addition of the sites. All changes that were made to the study protocol were approved by the IRB at each site, and our administrative official prepared formal justifications to add the additional sites, which were all approved by PCORI.
The study end points and sample size remained the same throughout the project. The engagement plan was the same as in the original proposal.
A total of 1322 women with newly diagnosed stage I or II endometrial cancer or EIN/CAH were screened for SUI at the 8 study sites (Figure 1). Overall, 702 (53.1%) screened women were positive for SUI, and of these, 642 (91.5%) met all eligibility criteria. Characteristics that differed by eligibility were study site, race/ethnicity, histologic grade, and clinical stage (P < .05; Table 5). A total of 556 eligible women (86.6%) consented and enrolled. The primary reason for declining enrollment was lack of interest in research participation. Feeling too overwhelmed was among the other reasons reported for declining participation. Characteristics that differed by participation were age, study site, and race/ethnicity (P < .05; Table 5).
STROBE Diagram of Participant Flow Through 12 Months.
Screening, Eligibility, and Enrollment Characteristics.
Seven women were found to be ineligible after consent due to metastatic disease (n = 1), hyperplasia without atypia instead of EIN (n = 1), visiting the urogynecology clinic before enrollment (n = 1), or no medical clearance for surgery (n = 4), and another 10 women withdrew for lack of time or not wanting to complete surveys. The main cohort was defined as the remaining 539 women. A total of 252 (46.8%) accepted referral to urogynecology, and 111 of these women ultimately had concomitant SUI and cancer surgery (44.0%). The comparator distribution in the main cohort of 539 was 20.6% with concomitant SUI and cancer surgery, 19.5% with nonsurgical SUI treatment, and 59.9% with no SUI treatment. This distribution was different from the anticipated breakdown of 50% with concomitant surgery.
During follow-up, there were 16 deaths, 5 withdrawals, and 4 discontinuations by study staff for medical and/or cognitive reasons or transfer of care to another institution. Other reasons for missing data were loss to follow-up or missing surveys. Overall, 429 participants attempted surveys at all 3 follow-up visits and were considered to have completed the study (79.6% of 539 in the main cohort). The number analyzed varied by aim-specific criteria but was higher than the number completed due to the use of GEE models that account for within-participant correlated outcomes and can accommodate some missing outcome data.
Demographic, clinical, and patient-reported measures at baseline are presented in Tables 6 to 8. Women who received concomitant SUI surgery more frequently reported Hispanic ethnicity or White race and less frequently reported Black/African American race. Having private insurance, being married or living with a partner, and living close or very far from the clinic were also more common characteristics within the concomitant surgery group. Site A (WIH) made up nearly half of the concomitant SUI surgery group compared with 31% of the cancer surgery only group. Clinical stage I/II and several comorbidities were less commonly identified among the concomitant surgery group, although psychiatric comorbidity was higher (33.35% vs 21.7% for cancer surgery only). Median FACT-En scores for QOL and LOT-R scores for optimism were both high overall and similar across SUI treatment groups. In contrast, measures of SUI severity and impact were noticeably higher for the concomitant SUI surgery group compared with the cancer surgery only group. More than 50% of the concomitant surgery group reported severe or very severe urinary symptoms, compared with <30% of the cancer surgery only group. Median UDI stress subscale scores and UIQ-7 impact scores in the concomitant surgery group were nearly double the median scores in the cancer surgery only group.
Baseline Demographic Variables by SUI Treatment Groups.
Baseline Clinical Characteristics by SUI Treatment Groups.
Baseline QOL, Optimism, and Incontinence Severity Measures.
Compare QOL scores of women with SUI and endometrial cancer who have concomitant surgery with women who have cancer surgery only.
Of the 539 participants in the main cohort, 495 had complete data on baseline covariates of interest and at least 1 postoperative FACT-En total score for analysis. The proportions included by treatment group were 95.5% for concomitant SUI surgery (n = 106/111), 86.7% for nonsurgical SUI treatment (n = 91/105), and 92.3% for no SUI treatment (n = 298/323). Attrition was more common in the nonsurgical SUI treatment group; therefore, baseline covariates potentially related to attrition and missing data were included in the multivariable analysis to minimize the impact of bias. Baseline differences between treatment groups were most notable for race/ethnicity, study site, distance traveled to the clinic/hospital, and incontinence severity measures. FACT-En scores were skewed toward higher values, with no scores during the study <25 and some scores at the maximum value of 100 at every visit. Unadjusted median FACT-En scores increased overall from 78.1 at baseline to 84.2 at 12 months postoperative (Figure 2, Table 9). The concomitant SUI surgery group had increasing median scores over follow-up, whereas median scores for the nonsurgical and no SUI treatment groups remained relatively flat during the postoperative period.
Median FACT-En Scores by SUI Treatment Group Over Time.
FACT-En Scores by Stress Urinary Incontinence Treatment Group Over Time.
Postoperative FACT-En scores were examined using GEE regression models (Table 10). Adjusting for study visit only, the estimated mean FACT-En score for concomitant SUI surgery was approximately 2.5 points higher than the mean score for nonsurgical SUI treatment, no SUI treatment, or both groups combined. Further adjustment for baseline covariates, including FACT-En score, reduced the differences to <2 points (P > .05). Because model residuals deviated from multivariate normality, FACT-En score was also examined as a binary outcome. Scores were dichotomized at the overall follow-up visit-specific medians presented in Table 9 (82.9, 83.1, and 84.2 for 6 weeks, 6 months, and 12 months, respectively). The concomitant SUI surgery group was about 20% more likely (RR, approximately 1.2) to have QOL scores above the overall median score at follow-up visits, compared with the other treatment groups. This association was statistically significant at α = .05 for concomitant SUI surgery vs no SUI treatment (RR, 1.20; P = .038) and vs no concomitant SUI surgery (RR, 1.20; P = .031). Tests for interaction between treatment and time for all models did not indicate effect modification by time (P > .5).
Multivariable Regression of SUI Treatment and FACT-En Scores.
Although the study population was limited to patients with clinical stage I/II cancer or CAH/EIH at diagnosis, 49 patients were found to have pathologic stage III or IV cancer after hysterectomy. Most were upstaged by having a positive lymph node identified pathologically. Adjusting the preceding analyses for pathologic stage III/IV reduced the RR estimates for concomitant SUI surgery and high FACT-En score by <2%.
UIQ-7 scores were examined as a secondary measure of QOL (Table 11). The concomitant SUI surgery group was consistently more likely to have scores indicating zero impact of incontinence across follow-up visits (adjusted RRs, 1.25-1.43), and all of these comparisons were statistically significant at α = .05.
Multivariable Modified Poisson Regression of SUI Treatment and UIQ-7 Values Equal to 0.
Compare the sexual function of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Of the 539 participants in the main cohort, 310 had complete baseline data including FSFI responses, and 278 had FSFI data at 6 and/or 12 months of follow-up. Just under half of women were sexually active at baseline, and, of these, 40% reported normal sexual function (Table 12). Sexually active women and those with normal function were younger and had higher median FACT-En scores. Women with any comorbidity were less likely to be sexually active. Median optimism scores (as measured by the LOT-R) were higher and median UDI stress subscale scores were lower for women reporting normal function vs dysfunction.
Characteristics of Baseline Sexual Activity and Function Measured by the FSFI.
The proportion of sexually inactive women decreased from baseline to 12 months postoperative, while the proportion reporting normal function modestly increased (Table 13). The concomitant surgery group had the highest proportion of inactive women at all time points. This group also had the highest proportion with normal function at baseline yet showed little change over time. In contrast, the frequency of normal function increased over time for the remaining participants. Adjusting for baseline covariates, the concomitant SUI surgery group had lower odds of dysfunction or normal function vs sexual inactivity during follow-up, although the associations were not statistically significant at α = .05 (Table 14).
FSFI Analysis Groups by SUI Treatment Over Time.
Multivariable Multinomial Logistic Regression for Postoperative FSFI.
When sexual activity/function at 12 months was compared with baseline level, the proportion of participants showing improvement was lower for the concomitant SUI surgery group than for the no concomitant SUI surgery group (Figure 3).
Change From Preoperative Sexual Activity and Function to 12 Months Postoperative by SUI Treatment.
Compare the clinical outcomes of women with endometrial cancer and SUI who have concomitant surgery with women who have cancer surgery only.
Among 505 participants with baseline covariate data, we examined clinical outcomes during and shortly after discharge, as described in aim 1. The 2 women who received prolapse surgery only (no SUI surgery) were excluded. Almost 90% of the concomitant SUI surgery group received tension-free vaginal tape slings (Table 15). Most hysterectomies were robotic-assisted in both groups.
Surgical Procedures.
Time until surgery, total estimated blood loss (EBL), and operative time were all greater for the concomitant SUI surgery group (Table 16). Intraoperative injury, retained Foley catheter or intermittent self-catheterization, and experiencing urinary frequency 3 weeks postoperatively were all more common in the concomitant surgery group. Adjusting for multiple comparisons, the differences between surgery groups for time until surgery, total EBL, operative time, and retained Foley catheter remained statistically significant at α = .05. Pain scores were similar in both groups. Regression adjustment did not appreciably change the findings (Table 17). Other complications were not frequently reported and were not examined by regression analysis.
Clinical Outcomes by Surgery Group.
Multivariable Regression Analysis of Concomitant Surgery and Clinical Outcomes.
Determine the characteristics of women who choose concomitant surgery compared with those who choose cancer surgery only.
Of the 556 participants who enrolled in the study, 3.1% (n = 17) were found to be ineligible or withdrew immediately after consent, leaving 539 participants in the main cohort. Of these, 45.3% (n = 252) accepted referral to urogynecology and 51.6% (n = 287) declined referral. After excluding 32 participants (5.9%) with missing baseline clinical, demographic, or survey data, 507 remained for analysis (240 who accepted referral and 267 who declined). Characteristics were similar between women included and excluded due to missing baseline data, with the exception of study site. A higher proportion of excluded participants were from site C (ie, sites other than WIH and UAB Medical Center) compared with those included (71.9% vs 26.4%). Site differences in the timing of the initial visit vs surgery may have led to more incomplete baseline surveys or other patient-reported data. In the analytic study population, 90.4% of women who accepted referral had a visit scheduled, and 49.8% (n = 108) of those received concomitant SUI and cancer surgery.
Race/ethnicity, language, distanced traveled, study site, region, and dyslipidemia or psychiatric comorbidity were all associated with referral acceptance with P < .1 (Table 18). Language and region overlapped with race/ethnicity and site, respectively, and were not further examined. Of the patient-reported measures examined, higher SUI severity measures (SSI, UDI stress subscale, UIQ-7) were observed for participants who accepted referral compared with those who declined (Table 19). Because these measures are correlated, the most specific measure of SUI, the UDI stress subscale, was selected for the multivariable model. FACT-En scores were lower for women who accepted referral, but the association by univariable logistic regression was P > .1.
Baseline Demographics and Clinical Characteristics by Acceptance of Referral to Urogynecology.
Baseline QOL, Optimism, and SUI Severity and Life Impact by Acceptance of Referral to Urogynecology.
The relationship between UDI stress subscale and referral acceptance deviated from linearity to some extent, with RQS terms improving model fit over the linear term (AIC = 634.4 vs 637.94) (Figure 4). UDI stress subscale remained a statistically significant, strong predictor of referral acceptance in the multivariable model (P < .0001) (Table 20). The odds of referral acceptance remained higher for site A (WIH) than for site B (UAB Medical Center) (P =.0080). The model had good discrimination (area under the curve [AUC], 0.775; 95% CI, 0.735-0.815) and calibration (Hosmer-Lemeshow goodness-of-fit P = .343).
UDI Stress Subscale and Referral Acceptance by Univariable Logistic Regression.
Univariable and Multivariable Logistic Regression Models for Acceptance of Referral to Urogynecology.
Predictors of having concomitant SUI and cancer surgery vs cancer surgery alone were similar, with the addition of more comorbid conditions (thyroid disorder: OR, 1.6; P = .049; musculoskeletal condition: OR, 0.47; P = .017) (Table 21). UDI stress subscale was the strongest predictor in the multivariable model (P < .0001). Model fit was also similar (AUC, 0.81; Hosmer-Lemeshow goodness-of-fit test, 0.631).
Multivariable Logistic Regression Model for Concomitant SUI Surgery.
Compare the rate of midurethral sling complications among women who received radiation therapy after sling placement with women who did not receive radiation therapy after sling placement.
A total of 6 adverse events related to sling placement were observed during follow-up (Table 22). One return to the operating room was reported for incomplete emptying. Three mesh exposures in the vagina and 2 erosions in the vagina were also reported. None of the participants with complications received radiation therapy; thus, no comparison of rates was performed. The incidence proportions, rates, and 95% CIs for each complication are in Table 23.
Mesh Complications and Return to Operating Room for Concomitant SUI Surgery Group.
Mesh Complication Incidence Rates for Concomitant SUI Surgery Group by Radiation.
SWD scores at 12 months after surgery were high overall, with a median of 95.8 and a lower quartile of 75 on a 0 to 100 scale. Similarly, decision regret was low with 75% of participants scoring ≤30 on a 0 to 100 scale. Participants who had concomitant SUI and cancer surgery had greater satisfaction and less regret about their decisions than participants with cancer surgery only, but these differences were not statistically significant at α = .05 (Table 24).
Decision Satisfaction and Regret at 12 Months.
Endometrial cancer is the most common gynecologic malignancy in the United States, and our study confirmed that 54% of women have SUI at the time of their endometrial cancer diagnosis. This is the first prospective study evaluating the true prevalence of SUI among women with a new diagnosis of endometrial cancer before hysterectomy, but our SUI rates are consistent with those reported in the literature, ranging from 49.9% to 80%.8,9,67 Additionally, we found that SUI symptoms are important to women with endometrial cancer because nearly half (45%) of all women enrolled in our study opted to see a urogynecologist before having their cancer surgery to discuss the options for treatment of their urinary incontinence. Women in all 3 of our treatment arms had improvement in endometrial cancer-related QOL (measured by FACT-En scores) 12 months after their cancer surgery. However, women in the concomitant surgery group were 20% more likely to have adjusted cancer-related QOL scores above the overall study population median score after surgery, compared with women who had cancer surgery alone (RR, 1.20; 95% CI, 1.01-1.41).
Recently, a study using the American College of Surgeons National Surgical Quality Improvement Program database reported only 2.4% of women undergoing gynecologic cancer surgery had a concomitant urogynecologic procedure.83 In our study, 21% of women opted for concomitant endometrial cancer and SUI surgery, and 45% opted for a urogynecology referral before cancer surgery. This finding supports the importance of screening women for SUI symptoms and providing them with information regarding treatment options for SUI at the time of endometrial cancer diagnosis. Many women currently being treated for endometrial cancer are not being screened for SUI but could benefit from SUI surgery.
There are many possible reasons physicians do not offer concomitant surgery, including possible delay to cancer surgery, possible increased complications with concomitant surgery, and possible operative complications. Consistent with the literature, our study found an increase in EBL during the surgery.83 However, this difference was only 75 mL, which is not clinically significant, and there was no concomitant increase in blood transfusions. There were no other differences seen in intraoperative complications or postoperative complications, including urinary tract infections. Additionally, we found that women who had concomitant surgery waited a median of 22 days from diagnosis until surgery compared with 16 days for women who had cancer surgery alone (P < .0001). However, this is only a 6-day delay, and both groups had surgery well under the 12-week time frame recommended for women with an endometrial cancer diagnosis. Finally, there was a 33-minute increase in operative time for women undergoing concomitant surgery. This is a small increase in anesthesia time compared with what would be needed with an additional surgery.
Finally, there has often been concern with placing a midurethral sling in women who might receive radiation therapy. Although the sling is out of the radiation field, sling erosion is still a concern. We know our study was not powered to look at sling erosion rates because the outcome is rare. However, 15 women in the concomitant group had radiation therapy, and we have not seen any sling erosions to date. We will continue to follow this cohort for 2 years because sling erosions are most common in that time frame.
Although QOL is important in women after endometrial cancer, not all women are interested in the same treatment options. Our study was designed as a cohort study so that we could identify factors that would be predictors of opting for a urogynecology referral. We identified many factors associated with choosing a urogynecology referral, including more severe SUI symptoms, distance <150 miles away from the treating hospital, race/ethnicity, study site region, and dyslipidemia. There were no differences based on cancer characteristics, other medical comorbidities, or other demographics. This information could be helpful to providers when counseling women on whether they would benefit from a preoperative urogynecology referral.
Our study involved 8 sites that were geographically diverse, including sites in the Northeast, South, Midwest, and West. We found no differences in our outcomes based on the study site region. Our study also included clinically diverse study sites, ranging from community hospitals to major academic referral centers. Our results are generalizable to many geographic locations and clinical settings. Our study was not racially/ethnically similar to the United States, because we only had a small percentage of Hispanic participants.
Our study has many limitations. It was a prospective cohort study, which has the risk of confounders, but our analytic methods took into consideration the confounders deemed most important by our team. Residual confounding by unmeasured confounders and measurement error of the confounders examined is possible. Additionally, there is the risk of selection bias inherent in the study design. We again accounted for factors that would impact a woman's choice of treatment groups in our analyses. As noted earlier, the study sample was not racially and ethnically the same as that of the United States, so the generalizability to women of certain racial or ethnic backgrounds is limited. Finally, more women chose cancer surgery alone than anticipated, so our study was slightly underpowered.
Women with both endometrial cancer and SUI often desire additional information about the treatment options for SUI before having their cancer surgery. Women opting for concomitant SUI surgery at the time of their cancer surgery had more severe and bothersome SUI symptoms. Women having concomitant surgery had a better chance of having a higher cancer-related QOL after surgery than women with no SUI treatment. Our study was a large multisite cohort that was geographically diverse. It has the risk of confounders because it was a cohort study, but our analytic methods took into consideration as many confounders as our team could identify. Additionally, more women chose cancer surgery alone than anticipated, so our study was slightly underpowered. However, the study had many strengths as well. Our results are generalizable to multiple clinical settings because our 8 sites were diverse, with some being community hospitals and others academic referral centers. Our study also is relevant to all women with endometrial cancer because we confirmed the high rate of SUI among this patient population as well as patients' significant interest in learning about treatment options for SUI before having cancer surgery. Therefore, we believe that gynecologists and gynecologic oncologists should screen patients for SUI at the time of endometrial cancer diagnosis and discuss the possible benefits of concomitant surgery.
We thank Joan Walker, Antonella Leary, Emmanuel Soultanakis, Amy Brown, Kathleen Schmeler, Gretchen Glaser, Erron Kinsler, Elizabeth R. Mueller, Paul Tulikingas, Eric Hurtado, Blair Washington, Danielle Antosh, Nancy Derby, Cheryl Pierce, Teri Kinder, Elaine Williams, Sinde Wickersham, Donna Macdonald, Christine Luis Marin, Christina Raker, Robin Cram, and Heidy Mendez.
Research reported in this report was funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (#CER-1409-22034). Further information available at: https://www.pcori.org/research-results/2015/comparing-surgeries-women-who-have-both-cancer-uterus-and-bladder-problems
Robison K, Wohlrab K, Sung V, et al. (2020). Comparing Surgeries for Women Who Have Both Cancer of the Uterus and Bladder Problems. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/11.2020.CER.140922034
The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.
This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/
Your browsing activity is empty.
Activity recording is turned off.
See more...
