Hypothesis / aims of study
Endometrial cancer is the most common cancer of the female reproductive system. If diagnosed early, surgical treatment consists of total hysterectomy and bilateral salpingo-oophorectomy (BSO) with a possible need for pelvic lymph node dissection for staging and therapeutic purposes. However, little is known about long-term morbidity when lymph node dissection is combined with robotic-assisted hysterectomy with BSO in relation to pelvic floor dysfunction.
The purpose of this study is to investigate the effects of lymph node dissection at the time of robotic-assisted hysterectomy and BSO on patients’ quality of life in regards to lower urinary tract symptoms. We hypothesized that patients who undergo lymph node dissection for endometrial cancer do not exhibit higher scores-indicative of more distress-in quality of life questionnaires related to urinary incontinence symptoms as compared to patients who do not require lymph node dissection.
Study design, materials and methods
This is a retrospective cohort of the patients who underwent robotic-assisted total laparoscopic hysterectomy (TLH) from June 2010 to March 2013. STROBE guidelines were followed. Gynecologic oncology surgical databases were searched and all patients who underwent robotic-assisted TLH, BSO and lymph node dissection with a final pathology of Stage 1 endometrial carcinoma were identified. These patients constituted the study group. Patients who underwent robotic-assisted TLH and BSO for the same diagnosis were selected as controls. The patients who underwent concomitant anti-incontinence and pelvic reconstructive procedures were excluded from this study.
Power and sample size calculations were based on the two-sample t-test for a retrospective cohort study design with 1:1 match ratio. Sample size was assessed to detect a difference between groups based on the outcome of Incontinence Impact Questionnaire-7 (IIQ-7). It was determined that for the primary outcome IIQ-7 score, group sample sizes of 34 patients each would achieve 81% power to detect a clinically significant difference of 10% in IIQ-7 score.
Two quality of life questionnaires related to lower urinary tract dysfunction were administered to 140 randomly selected patients among the 242 patients who met the study inclusion criteria. The prevalence and nature of urinary dysfunction after surgery was assessed by using the Urinary Distress Inventory-6 (UDI-6). The impact on patients’ quality of life was assessed by using the Incontinence Impact Questionnaire-7 (IIQ-7). Data was collected by means of a postal questionnaire and phone interviews. Cases were matched with controls by age, weight and presence of urinary incontinence risk factors including post-operative vaginal brachytherapy and presence of self-reported urinary incontinence at baseline.
Summary statistics were reported for each outcome, stratified by study group. Departures from normality were assessed by evaluating the distributions for each outcome via utilization of histograms and boxplots. Assuming a normal distribution, the two-sample t-test was used to compare the scores of the two groups. Discrete pre- and peri-operative variables were analyzed using the chi-square test. Simple and multiple linear regressions were carried out to determine the relationship between IIQ-7 score and urinary incontinence risk factors. All analysis was performed using IBM SPSS v.20 software. An alpha of 0.05 was used for statistical significance.
Results
The response rate to the questionnaire was 52.8% including those who responded via telephone survey.
Our study included 74 women for whom IIQ-7 and UDI-6 scores were available. Each group contained 37 subjects. Therefore, we were able to achieve the required sample size for an adequate power analysis. The mean age was 59±11 years and the mean BMI was 39±11 kg/m2. Adjuvant radiation therapy was completed by 23% of women and 28.4% reported the presence of urinary incontinence preoperatively. The above mentioned and additional risk factors for urinary incontinence did not significantly differ between the study and control groups.
Peri-operative data for both groups was collected from hospital records. Groups did not differ in estimated blood loss, uterine weight and postoperative length of stay. As expected, average surgical operative time was longer in the study group (189±63 compared to 134.9±41 minutes, p <.000) . The control group had an average follow-up time of 11.8 months, which was four months shorter than the study group and statistically significant (p=.035).
Preoperatively, 52 of the 74 patients did not have any symptoms of urinary incontinence and the remaining 21 patients reported experiencing some type of urinary incontinence preoperatively. Postoperative urinary incontinence was defined as a self-reported score above 0 in three urinary leakage subscales (questions 2, 3 and 4 of UDI-6 questionnaire). Based on that definition, the total postoperative urinary incontinence rate in the entire cohort of patients was 74.3%; 37 of 52 (71.2%) women reported new onset of urinary incontinence after surgery and 3 out of 21 (14.3%) reported a resolution of preoperative urinary incontinence symptoms. The odds ratio for developing new onset urinary incontinence was 2.4 with 95% CI .62-9.5 (p-value = 0.18).
The primary outcomes of our study were mean IIQ-7 scores for the study and control groups of 14.9±23.3 and 10.5±22.9, respectively (p=0.419, Figure 1). A simple linear regression revealed a significant correlation between postoperative IIQ-7 and UDI-6 scores with a 0.4 point increase in IIQ-7 score and with each 1 point increase in UDI-6 score (Figure 2). All other urinary incontinence risk factors including obesity, age, race and vaginal brachytherapy did not demonstrate any correlation with IIQ-7 score in our multiple linear regression model.
Interpretation of results
The crude rate of urinary incontinence in our study population of women treated for endometrial cancer was similar to previously published studies [1,2]. However, in our patient population, we were not able to demonstrate such high score values indicative of more of an effect on quality of life related to urinary dysfunction as reported in these two studies.
Strengths of this study were the matched design and power analysis for sample size determination. By comparing the two well-characterized groups, we were able to control for pre-existing incontinence, age, obesity, brachytherapy while analyzing the quality-of-life impact using validated measures.
A specific risk factor in relation to the gynecological cancer population is postoperative adjuvant radiation therapy. However, we did not observe a correlation between adjuvant radiation therapy and more severe incontinence symptoms and a higher impact on disease-specific quality of life as reported in the literature [1,2]. This can be related to the fact that postoperative adjuvant radiation therapy for Stage 1 endometrial cancer at our institution is almost exclusively limited to vaginal brachytherapy, which is associated with lower bladder toxicity.
This study has several limitations. Total non-response rate was 47.2%. There was 7.5 years difference between respondents and non-respondents, with non-respondents being significantly older. Because age is one of the strongest predictors of urinary incontinence, it is an important confounder to consider when making comparisons across groups.
Also, due to the retrospective study design we cannot comment on the progression of urinary symptoms after surgery. Our only pre-operative urinary symptoms assessment included self-reported urinary incontinence rates obtained at the time of pre-operative evaluation of the patient.