Immediate versus delayed urinary catheter removal following non‐hysterectomy benign gynaecological laparoscopy: a randomised trial

To compare rates of urinary retention and postoperative urinary tract infection between women with immediate versus women with delayed removal of indwelling catheter following benign non‐hysterectomy gynaecological laparoscopic surgery.


| I N TRODUC TION
Bladder emptying prior to laparoscopic pelvic surgery is common to reduce the risk of urinary tract injury. Perioperative bladder care ranges from preoperative voiding with no catheterization, 1,2 to the use of an indwelling catheter (IDC) and early removal during or immediately following surgery, 3 to the use of IDC and delayed removal after one or more postoperative days. 4 The reported risk of postoperative voiding dysfunction following gynaecological surgery ranges widely, from 3% to 53%. [5][6][7][8] These studies are limited by small sample size and have largely focused on outcomes following hysterectomy or urogynaecological surgery, on an assumption of a greater impact on the bladder. 7,8 Subsequently, protocols for bladder care that do not involve this type of surgery have been instituted largely without randomised data to support such decisions. Delayed postoperative catheter removal may increase the risk of urinary tract infection (UTI), 6,[8][9][10] the rates of which range between 9.9%-21.0% in catheterised patients and 3.8%-12.0% in noncatheterised patients following non-hysterectomy gynaecological laparoscopy. 3,5 Comparison between studies is hindered by the heterogenous nature of surgeries, the differing practices for prophylactic antibiotics and variable definitions of UTI.
Evidence-based guidance for bladder care during and after non-hysterectomy laparoscopic gynaecological surgery is poor, and the optimal duration for catheterisation (if any) and the timing of IDC removal following surgery is uncertain.
Globally, there is considerable variation in practice for length of stay following gynaecological laparoscopy. In one US study, 57% of patients who underwent laparoscopic myomectomy were discharged in <24 h. 11 In contrast, a randomised controlled trial (RCT) found that the median length of stay following simple elective gynaecological laparoscopy in a tertiary hospital in China was 2 days. 12 This variation in the average duration of hospital stay may also reflect variation in postoperative bladder care, as it is likely that most women who undergo earlier discharge have an earlier removal of the catheter. It is unknown whether practice variations in postoperative bladder care contribute to differences in clinical outcomes, specifically in the rates of urinary retention and UTI.
This RCT was designed to address this evidence gap and compare immediate versus delayed removal of the urinary catheter following benign non-hysterectomy laparoscopic gynaecological surgery.

| Study design
Between February 2012 and December 2019, an RCT was conducted for women undergoing benign gynaecological laparoscopy (of any type, other than hysterectomy or pelvic floor surgery). Participants were enrolled from two universityaffiliated hospitals in Sydney, Australia, comparing women with the removal of urinary catheter immediately after surgery (immediate group) with women who underwent urinary catheter removal the following morning (delayed group).

| Participants
Eligible participants were women aged 18 years and over, scheduled for gynaecological laparoscopic surgery, excepting those treated for the suspicion of or confirmed gynaecological malignancy, hysterectomy, pelvic floor surgery or when concomitant bowel surgery was planned. Other preoperative exclusion criteria included pregnancy, current UTI and a lack of fluency in English.

| Preoperative
Demographic and historical data were recorded and study participants completed the validated ICIQ-FLUTS questionnaire (with a score range of 0-56, with a higher score being indicative of poorer urinary outcomes) and the ICIQ-LUTSqol questionnaire (with a score range of 19-266, with a higher score being indicative of poorer quality of life). 13 A preoperative urine dipstick analysis was undertaken, and was required to be negative for nitrites and leukocytes for inclusion.

| Intraoperative
In keeping with standard practice at our institutions, participants in both arms received intravenous (IV) cefazolin 2 g or an appropriate alternative when cephalosporins were contraindicated. The type and duration (in minutes) of surgery were recorded. Where excision of endometriosis was performed, disease stage was recorded in accordance with the revised American Society of Reproductive Medicine (rASRM) classification system, and diagnosis was histologically confirmed. Intraoperative exclusion criteria included surgery confined to diagnostic laparoscopy only and intraoperative injury to the urinary tract.

| Postoperative
Participants were randomised at the completion of surgery (following the last suture) to either immediate or delayed IDC removal by an independent, third-party telephone randomisation service, according to a computer-generated randomisation schedule, in balanced blocks of 20. The block size of 20 was chosen because of the lack of blinding and because larger blocks reduce predictability. In the immediate removal group, the IDC was removed at the completion of the surgery. In the delayed group, the IDC remained in place and was removed at 6:00 am the following day, in keeping with standard local operating policy.
A passive trial of void was completed for all participants, noting the time and volume of each void. The post-void residual (PVR) was measured within 5 min of each void by nursing staff using a transabdominal ultrasound scan of the bladder (CubeScan BioCon-500™; Mcube Technology Co. Ltd, Seoul, Republic of Korea). Participants in the delayed group followed the usual hospital trial-of-void protocol, 14 and were considered to have passed their trial of void after they had voided twice, each time with a volume >150 ml, and the PVR was <100 ml. If the PVR was between 100 and 200 ml, the patient was asked to double void and the bladder scan was repeated. A PVR of >200 ml, an inability to void or a persisting sensation of incomplete emptying were indications for re-catheterisation for participants in the delayed arm. Participants in the immediate group followed a different trial-of-void protocol. This protocol was designed to emulate the practice of discharging patients home after surgery once they are voiding, without reliance on PVR volume. Participants in the immediate group were considered to have passed their trial of void after two voids of >150 ml in volume. PVR was still assessed in this group, but medical staff were blinded to the PVR result unless it was >1000 ml. The level of >1000 ml was selected because of the risk of long-term voiding dysfunction at this level of retention. A PVR of >1000 ml, an inability to void or a persisting sensation of incomplete emptying were indications for re-catheterisation for participants in the immediate arm. As a result of the detailed protocol for participants in both arms, urinary retention in this study was defined as the requirement for any postoperative re-catheterisation.
Intra-and postoperative analgesic requirements were collected and converted to total morphine equivalents (mg). 15 The duration of postoperative hospitalisation was recorded. Participants completed a consecutive 14-day post-discharge bladder diary to record symptoms suggestive of UTI (frequency, dysuria or fever). At 2 weeks after surgery, participants were contacted via email or phone by a health service provider to document any symptoms suggestive of a UTI, if urine microscopy and culture had been performed and if antibiotics had been commenced empirically for UTI. This procedure was repeated at the 6-week follow-up. Participants were diagnosed with a UTI if they had experienced suggestive symptoms and had been treated empirically with antibiotics or had symptoms and had confirmation of UTI on urine culture. A safety monitoring committee of two independent researchers, unaware of group allocation, assessed any adverse events by interim analyses.

| Study outcomes
Primary outcomes were postoperative urinary retention and UTI. Secondary outcomes included hospital readmission, unscheduled presentation to healthcare provider, analgesia requirements, duration of hospital stay and bladder symptoms/quality of life, assessed by validated questionnaires.

| Statistical analyses
Sample size was estimated based on previous studies that had reported postoperative voiding dysfunction at 9.2%, 16 and a postoperative UTI rate of 7.5%. 5 This included surgeries other than non-hysterectomy laparoscopy, but reflects the limited data on the specific prevalence of postoperative retention and UTI. We calculated that to detect a doubling in these rates with 80% power at the 5% significance level, 345 participants would be required in each group. This included an allowance for a 10% withdrawal and loss to follow-up.
Statistical analyses were performed using an intentionto-treat analysis. Data were recorded and transcribed to Stata BE 17 (StataCorp LLC, College Station, TX, USA). 17 A data audit to confirm participant study data for all primary and secondary outcome data prior to statistical analyses was performed. Data normality was assessed by the Kolmogorov-Smirnov method. Hypothesis tests were twosided. The co-primary outcomes of urinary retention and UTI were analysed using the chi-square test. Readmission and unscheduled representation were also analysed using the chi-square test. The duration of hospital stay and intraoperative analgesia was analysed using the independent Student's t-test. Postoperative analgesia was analysed using the Wilcoxon rank sum test. The paired ICIQ-FLUTS/LUTS data were analysed using the Wilcoxon signed rank test.
Missing data were assessed prior to analysis and if determined to be missing at random were managed by deletion, where missing data were <5%. 18 Where missing data were determined to not be missing at random, multiple imputation was undertaken.
A post hoc subgroup analysis, including a regression analysis using an interaction test, was performed to assess the relationship between urinary retention and the covariate of stage of disease.

| Trial participants
Between April 2012 and December 2019, 2147 women were considered for inclusion to the study. Figure 1 presents the CONSORT flow diagram for the study. Patient demographics are presented in Table 1, with no clinically significant differences between the groups with immediate (n = 355) and delayed (n = 338) removal. The most common surgical procedures were treatment for endometriosis and adnexal pathology, commensurate with the specialty area of our team, and the duration of surgery did not differ between the two groups. Voiding data were complete for 100% of study participants, and data on UTI rate, hospital readmission and representation were complete for 83% of participants. Data for preoperative bladder symptoms/quality of life, intra-and postoperative analgesia and duration of hospital admission were complete for 99% of participants. Postoperative bladder symptoms/qualityof-life data were present for 65% of participants.
The co-primary and secondary outcomes are presented in Table 2.
All participants who required re-catheterisation were symptomatic, with either an inability to void (n = 18) or a reported feeling of incomplete emptying (n = 26). Only one participant who required re-catheterisation had a measured residual bladder volume of >1000 ml (allocated to immediate removal), and this patient was one of those who was unable to void. In other words, no participants were diagnosed with urinary retention entirely on bladder scanning criteria without also either reporting being unable to void or incomplete emptying. One participant in the delayed removal group had a prolonged trial of void because of high residual volumes, which resolved itself without catheterisation, and she was not included in the urinary retention group. No participants were discharged home with an IDC.  Although the rate of UTI was higher in the delayed removal group at 7.2% compared with 4.7% in the immediate removal group, and the risk ratio is 0.7 (95% CI 0.3-1.2%), this difference did not reach statistical significance. The risk difference between the groups is 2.4% (95% CI −1.4% to 6.3%).
One participant had both urinary retention and UTI (immediate removal). One participant required a prolonged stay for postoperative fever and was discharged home after inpatient management of UTI (immediate removal). All other participants diagnosed with UTI were managed as outpatients. The observed characteristics of participants who were followed up were compared with participants lost to follow-up (Table S1). Sensitivity analyses using estimated rates of UTI between 2.8% and 10.8% were selected, as these reflected the upper and lower values of the confidence intervals (Table S2).

| Secondary outcomes
There was no difference in analgesia requirements and duration of stay between the groups. In the immediate versus delayed removal groups, no significant difference was demonstrated in hospital readmission rates (3.7% vs 2.5%) or unscheduled re-presentation (9.8% vs 11.9%). There was no significant difference between the two groups for preoperative or postoperative bladder function, as assessed by IFLUTS or quality of life, and no difference within each group over time.

| Subgroup analysis
Seventy-six percent of the study participants underwent surgery for the excision of endometriosis (72.4% as a primary indication for procedure, and another 3.6% in conjunction with other pathology). Staging was available for over 97% of cases, with 25.8% undergoing surgery for stage-1, 38.5% for stage-2, 12.5% for stage-3 and 20.6% for stage-4 disease. Figure 2 demonstrates that although there is a higher rate of retention following the excision of stage-4 disease compared with stage-1 disease, the relationship between retention and endometriosis stage is nonlinear. A post hoc subgroup analysis showed that, apart from stage-2 disease, there was no statistically significant increased risk of retention between the immediate and delayed groups for endometriosis stage ( Table 2). The significance of the finding for stage-2 disease is questionable, given that an interaction test demonstrated no increased risk of retention for the covariate of endometriosis stage, both overall (p = 0.08) and in the treatment groups of immediate IDC removal (p = 0.3) and delayed IDC removal (p = 0.2).
Three participants were unexpectedly found to have cancer or borderline ovarian pathology on histology. Although surgery for malignancy was an exclusion criterion for recruitment, these patients were not excluded from the analysis as the diagnosis was made after the first primary outcome was recorded. None of these patients had urinary retention or postoperative UTI.

| Main findings
In this large RCT of bladder care following non-hysterectomy benign gynaecological laparoscopy, there was a significantly increased rate of urinary retention following immediate, compared with delayed, catheter removal. There was no statistically significant difference in the rate of UTI between the two groups.

| Interpretation
Contributors to urinary retention in surgical patients include urethral oedema, bladder overdistension and surgical denervation. 19 Comparison of urinary retention rates between studies is challenging because of the range of indications for surgery, the variable definitions for urinary T A B L E 1 Demographic and surgical data of study participants. retention and the range of durations of postoperative follow-up. The findings in this study demonstrated a rate of urinary retention lower than the 9.2% reported in a prospective observational study of 284 gynaecological surgery patients that included hysterectomy, by both open and laparoscopic approaches. 16 Although urinary retention rates for laparoscopy versus laparotomy have not been directly compared, the same study noted a 13.7% rate of urinary retention following hysterectomy by laparotomy, compared with 8.3% following hysterectomy by laparoscopy. A simple explanation for this, given the same surgical steps are followed, is the increased analgesia use and delayed postoperative mobilization associated with laparotomy compared with laparoscopy. 20 By limiting the surgical approach to non-hysterectomy laparoscopy, other factors that might contribute to postoperative bladder dysfunction can be considered in this study. Over 32% of participants underwent the excision of moderatesevere endometriosis, with the potential for damaging pelvic nerves involved in bladder function, similar to when extensive dissection is required during hysterectomy. 21 A previous prospective cohort study found that 82.4% of patients reported severe pelvic neurological dysfunction following non-nerve sparing radical excision of pelvic endometriosis with segmental bowel resection. 22 Although this study determined higher rates of retention following the excision of stage-4 (severe) disease, compared with stage-1 (mild) disease, this was neither linear by stage nor statistically different, probably because of T A B L E 2 Primary and secondary outcomes.   Rate Urinary RetenƟon (%) by Stage of Endometriosis All ParƟcipants Immediate Removal Delayed Removal a lack of power for subgroup analysis. There are no RCTs to assess the outcomes on bladder function and quality of life following nerve-sparing compared with non-nerve-sparing surgery. Such trials would pose substantial and possibly prohibitive methodological challenges. Based on this study, patients can be appropriately discharged if they are voiding, have good sensation and feel empty. This study determined that postoperative bladder dysfunction that requires re-catheterisation can be assessed by symptom-guided assessment, rather than trial-of-void measurements. Repeated measuring of voided volumes and bladder scanning in the asymptomatic patient is unnecessary, invasive, time consuming for nursing staff and may confer an increased risk of unnecessary re-catheterisation. Findings from this RCT are in keeping with a 2021 retrospective study, which found no difference in the rates of postoperative voiding dysfunction between liberal versus strict voiding protocols in 115 women following laparoscopic myomectomy, suggesting that a liberal trial-of-void protocol does not increase the risk of urinary retention. 23 A 2021 systematic review of 21 randomised studies, including 40-246 patients each, reported different postoperative trial-of-void techniques following vaginal and laparoscopic surgery, including hysterectomy. 24 The review found no difference in urinary retention outcomes between different methods, even in the studies using a high post-void residual cut-off value, and did not find any one method to be superior. This suggests that although it is often assumed that hysterectomy, urogynaecological and non-urogynaecological surgery affect bladder function differently, postoperative voiding protocols do not necessarily need to be specialised according to the type of surgery.

| Strengths and limitations
The rate of UTI is lower in this study compared with other studies. Participants in this study preoperatively received IV prophylactic antibiotics, in keeping with the usual practice at our institutions. Although this is not universal practice, there are no placebo-controlled double-blinded RCTs in this field. Antibiotic use may have impacted the UTI rates seen in this study; however, any impact would have applied equally to both groups. Post-randomisation dropouts were different between the groups, potentially because the protocol required randomisation at surgical completion. The majority of participants in the delayed arm who were withdrawn had diagnostic laparoscopy only, and it is therefore unlikely that this post-randomisation dropout did impact findings. The difference in trial-of-void protocols between groups was by deliberate design. A review of all notes for participants who were re-catheterised showed that all patients reported an inability to void or a sensation of incomplete voiding. Therefore, all these participants would have met the study criteria for re-catheterisation, regardless of the measured PVR.
The loss to follow-up rate is a major limitation. However, the sensitivity analysis for participants lost to follow-up indicates that the rate of UTI needed to be 10% higher in the delayed compared with the immediate group to reach statistical significance -substantially higher than the observed 2.5% difference between the groups. The loss to follow-up does compromise the power to determine a difference for UTI based on the a priori power calculation. However, a post hoc power calculation estimated that a sample of 2186 participants would be required to determine statistical significance. The local antibiotic policy may contribute to this discrepancy in data, but also calls into question the general guidance that does not recommend antibiotic use for this type of surgery. 25 Other study limitations include the external validity of this study, given the average study participant was aged 34 years, with a BMI of 25 kg/m 2 , nulliparous, white, had few comorbidities and underwent surgery for endometriosis. The study was underpowered for subgroup analysis, and the findings of the post hoc analysis of the relationship between retention and stage of endometriosis should be interpreted on this basis. In addition, the other patient and surgical factors may increase the risk of urinary retention and UTI, apart from immediate or delayed IDC removal.
There is an increasing move towards earlier postoperative discharge, even after hysterectomy. 26 Good postoperative bladder care will be important in this setting. Participants in this study were not subject to enhanced recovery after surgery (ERAS) protocols. A previous study has noted such protocols should be mindful of the perioperative anticholinergic load in patients undergoing benign gynaecologic surgeries, 23 and future research could assess the possible relative risk of postoperative voiding dysfunction.

| CONCLUSION
As there is a move towards earlier postoperative discharge, findings from this study may be used to enhance perioperative counselling and shared decision-making about bladder care following non-hysterectomy gynaecological laparoscopic surgery. There is no reduction in UTI rates with immediate compared with delayed catheter removal. Clinicians who favour the immediate removal of IDC and earlier discharge should be aware that there is an associated increased risk of urinary retention. They can inform their patients that there is approximately a 1 in 12 chance of needing the reinsertion of a urinary catheter to manage retention. It is important to ensure that patients report normal voiding and emptying prior to discharge, to reduce the need for readmission for the management of urinary retention.

AU T HOR C ON T R I BU T ION S
JAA, EN-H and RD were involved in the conception, planning, carrying out the study, and analysing, writing and reviewing the article. LM, SS, SM-L, AB, CM, HW, AA, BK, KN, FGL and CYL were involved in carrying out the study, and analysing, writing and reviewing the article.

AC K NO W L E D GE M E N T S
Open access publishing facilitated by University of New South Wales, as part of the Wiley -University of New South Wales agreement via the Council of Australian University Librarians.

C ON F L IC T OF I N T E R E S T S TAT E M E N T
None declared. Completed disclosure of interests form available to view online as supporting information.

DATA AVA I L A BI L I T Y S TAT E M E N T
The trial data set will be uploaded to a publicly available data repository upon publication. All of the individual participant data collected during the trial will be available, after deidentification. The study protocol, statistical analysis plan, the participant information sheet and the informed consent form will also be available. The data will remain available from time of publication, with no end date. The data will be available to anyone who wishes to access the data, for any purpose and the link to the data repository link will be published.

E T H IC S A PPROVA L
This study received Institutional Review Board ethical approval: HREC 11/080 (HREC/11/POWH/103). Prospective patient consent was obtained for participation in the trial and publication of the article.