Postoperative Urinary Retention and Urinary Tract Infections Predict Midurethral Sling Mesh Complications

Accepted Manuscript Title: Post-Operative Urinary Retention and Urinary Tract Infections Predict Mid-Urethral Sling Mesh Complications Author: Nahid Punjani, Jennifer Winick-Ng, Blayne Welk PII: DOI: Reference:

S0090-4295(16)30732-4 http://dx.doi.org/doi: 10.1016/j.urology.2016.10.019 URL 20089

To appear in:

Urology

Received date: Accepted date:

31-8-2016 11-10-2016

Please cite this article as: Nahid Punjani, Jennifer Winick-Ng, Blayne Welk, Post-Operative Urinary Retention and Urinary Tract Infections Predict Mid-Urethral Sling Mesh Complications, Urology (2016), http://dx.doi.org/doi: 10.1016/j.urology.2016.10.019. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Post-Operative Urinary Retention and Urinary Tract Infections Predict Mid-Urethral Sling Mesh Complications Nahid Punjani1, Jennifer Winick-Ng2, Blayne Welk1,2,3 1Division

of Urology, Department of Surgery, Western University, London, Canada 2Institute for Clinical Evaluative Sciences, Ontario, Canada 3Department of Epidemiology and Biostatistics, Western University, London, Canada

Funding and Disclosures: None Word Count: 2774 Key Words: mid-urethral sling, urinary retention, urinary tract infection, mesh complications Running heading: Retention, urinary infections and mesh complications Corresponding Author: Blayne Welk, MD FRCSC MSc Assistant Professor, Division of Urology Western University Room B4-667 St Joseph's Health Care 268 Grosvenor Street London ON N6A 4V2 Telephone: 519 646-6367 | Fax: 519 646-6037 [email protected] ACKNOWLEDGEMENTS This project was conducted at the Institute for Clinical Evaluative Sciences (ICES) Western Site. The ICES is funded by an annual grant from the Ontario Ministry of Health and Long-term Care. ICES Western is funded by an operating grant from the Academic Medical Organization of Southwestern Ontario. Disclaimer: The opinions, results, and conclusions are those of the authors, and no endorsement by the Institute for Clinical Evaluative Sciences, Ontario Ministry of Health and LongTerm Care, or Academic Medical Organization of Southwestern Ontario is intended or should be inferred. Parts of this material are based on data and information compiled and provided by CIHI. However, the analyses, conclusions, opinions and statements expressed herein are those of the author, and not necessarily those of CIHI.

1

Page 1 of 22

ABSTRACT

Objective: To determine if post-operative urinary retention and urinary tract infections (UTI) were predictors of future mesh complications requiring surgical intervention after midurethral sling (MUS).

Methods: Administrative data in Ontario, Canada between 2002 and 2013 was used to identify all women who underwent a mesh-based MUS. The primary outcome was revision of the transvaginal mesh sling (including mesh erosion/removal, fistula or urethrolysis). Two potential risk factors were analyzed: post-operative retention (within 30 days of procedure) and number of post-operative emergency room visits or hospital admissions for UTI symptoms.

Results: A total of 59,556 women had a MUS, of which 1,598 (2.7%) required revision surgery. Of the 2,025 women who presented to the emergency room or were admitted to hospital for post-operative retention, 212 (10.5%) required operative mesh revision. Of the 11,747 patients who had at least one post-operative UTI, 366 (3.1%) patients required operative mesh revision. In adjusted analysis, post-operative retention was significantly predictive of future reoperation (HR 3.46, 95%CI 2.97-4.02), and this difference persisted when urethrolysis was excluded as a reason for sling revision (HR 3.08, 95% CI 2.62-3.63). Similarly, in adjusted analysis, each additional post-operative hospital visit for UTI symptoms increased the risk for surgical intervention for mesh complications (HR 1.74 95% CI 1.61-1.87).

2

Page 2 of 22

Conclusions: Post-operative urinary retention and hospital presentation for UTI symptoms are associated with an increased risk of reoperation for MUS complications. These patients should be followed and investigated for mesh complications when appropriate.

INTRODUCTION

Midurethral slings (MUS) are considered the gold standard for the treatment of female stress urinary incontinence and their efficacy and safety have been demonstrated in clinical trials involving thousands of patients1,2. All surgical procedures can lead to complications, but the MUS is associated with potential difficulties specific to transvaginal mesh, and this has attracted significant media and regulatory attention2. These problems include urethral mesh erosions and vaginal mesh exposures, as well as post-operative pain and voiding dysfunction, all of which may necessitate surgical removal or revision of the mesh3,4.

Despite the significant discussion around mesh complications there have been few studies examining the predictive ability or the magnitude of risk accompanying postoperative symptoms commonly associated by experts with mesh complications 5,6. Two common features of the clinical history of women with problems from their MUS include a period of post-MUS urinary retention necessitating catheterization, or the new onset of frequent urinary tract infection (UTI) symptoms which may or may not be culture

3

Page 3 of 22

positive7,8. The management of post-MUS urinary retention is inconsistent, ranging from immediate operative loosening to more conservative measures such as long periods of catheterization5,6,9. UTIs in the post-operative period may be a symptom of MUS problems such as mesh erosion, inappropriate sling placement or excessive tensioning, or the misdiagnosis of urinary symptoms related to mesh complications10,11.

The objective of our study was to examine whether post-operative urinary retention and frequent UTI diagnoses in women after a MUS were predictive of future mesh complications requiring surgical intervention.

MATERIALS AND METHODS

Study design and setting

This study was conducted through the Institute for Clinical Evaluative Sciences (ICES) according to a pre-specified protocol. We conducted a population based, retrospective cohort study of all adult women who received a MUS between April 1, 2002 and December 31, 2013 in the province of Ontario, Canada (where the entire population (approximately 13 million people), has access to a universal healthcare system). This cohort includes several different MUS (including retropubic and transobturator slings), and various surgical practice patterns. Approval for research ethics was granted by Sunnybrook Hospital (Toronto, Ontario).

4

Page 4 of 22

Data sources

Four linked databases were used for our study: (1) The Canadian Institute for Health Information’s Discharge Abstract and Same Day Surgery (CIHI-DAD/SDS) databases for hospital-based diagnostic and procedural information, (2) The Ontario Health Insurance Plan (OHIP) for physician claims, (3) The Registered Persons Database (RPDB) for demographic information and (4) The National Ambulatory Care Reporting System (NACRS) for emergency room visits. Patient level data linkage was performed via deterministic linkage using an encrypted version of the provincial personal health identifier and analyzed at ICES. The databases are complete for all study variables (aside from geographic variables, which were missing in <2% of patients). The accuracy and validity of these databases are high, and has been described in our previous work12. The data analyst (JW) had full access to the study data sets.

Patient cohort

Canadian Classification of Health Intervention (CCI) codes from CIHI-DAD/SDS were utilized to identify women who underwent a MUS procedure (Appendix 1). These codes are specific to midurethral slings, and different codes are used for cadaveric or fascial slings. The index date was considered the date of this procedure. Censoring of patients occurred at death, emigration from the province (no healthcare contact for 12 months), or the end of the study period (March 31, 2015). Patients were excluded if they were <18 years of age (n=8), male or missing gender (n=39), were not a resident of Ontario

5

Page 5 of 22

(n=40), had a stress incontinence procedure in the 5 years prior (based on the availability of our administrative data) to the index date (n=1,057), or if their records were missing the institution identification number (n=491).

Primary outcome

Using previous defined CIHI-DAD/SDS codes12 we identified subsequent surgical procedures which were likely the result of a MUS complication: urethrolysis, mesh removal, mesh erosion, or bladder or urethral fistula repair. As we hypothesized a significant (and obvious) association between post-operative retention and future urethrolysis, we subdivided our primary outcome into two sub-outcomes: 1) urethrolysis and 2) mesh removal, erosion or fistula repair in order to better assess the impact of post-operative retention on mesh-specific problems.

Risk factors and covariates

Our a priori primary risk factor of interest was post-operative urinary retention within 30 days of the MUS procedure. This was defined using appropriate CCI codes/Diagnosis codes in CIHI-DAD/SDS and NACRS (Appendix 2). We could only reliably measure this outcome in patients who were in hospital or presented to the emergency department. Patients discharged home with a catheter or those managed in a physician’s office were not captured. In Ontario, women with acute urinary retention will often present to an

6

Page 6 of 22

emergency room for catheterization, as same day access to surgeons or family physicians is often not available.

Our secondary risk factor was post-operative UTIs or UTI-like symptoms which presented to the emergency department or required hospital admission (determined from CIHI-DAD/SDS and NACRS, Appendix 2). This was included as a cumulative count variable (visits within 7 days were considered a single infection). Women with ≥3 UTIs during the observation period were considered as a single group. Post-hoc, we evaluated the impact of outpatient UTI symptoms on the future risk of midurethral sling surgery. Outpatient UTIs were defined as any family physician or specialist visit with the ICD9 code 599.

The additional covariates selected were based on previous work, and have been described and referenced in detail previously12. Briefly, they included age, obesity (BMI >40), diabetes, number of UTIs in the year prior to the MUS, combined hysterectomy, combined prolapse repair, repeated MUS procedures, surgeon volume and specialty, patient’s geographic region, academic or community hospital, comorbidity, socioeconomic status, and healthcare utilization in the year prior to the mesh SUI procedure. We also evaluated concurrent neurogenic disease (including a prior diagnosis of multiple sclerosis, spinal cord injury, Parkinson’s disease, or neurogenic bladder). The independent effects of these covariates have been discussed in detail previously, and were primarily included to adjust for potential confounding with the current risk factors of interest12.

7

Page 7 of 22

Statistical Analysis

We reported medians (interquartile range, IQR) or frequencies (count). Baseline differences of patients were evaluated using standardized differences (SD), which better recognizes potentially meaningful differences (SD >10%) compared to traditional hypothesis testing13. We performed univariate and multivariate Cox proportional hazards regression, accounting for the time-varying nature of the primary and secondary risk factors (SAS 9.4). Results are reported as hazard ratios (HR) with 95% confidence intervals and p-values. A two sided p<0.05 was considered significant.

RESULTS

Baseline Characteristics

We identified 59,556 women who underwent a MUS procedure over the study period. The median age was 52 (IQR 45-63) years. The median follow-up was 5.9 (IQR 3.6-8.6) years, and 2,362 (4.0%) and 1,383 (2.3%) of patients were censored for death and emigration respectively. Demographics of the cohort are shown in Table 1 (and additional demographics are included in Appendix 3). Women who re-presented to hospital with post-operative urinary retention were significantly more likely to have had their surgery completed by a Urologist (54.9% vs 42.2%, SD 26%) than a Gynecologist. A significantly lower portion of patients who experienced post-operative urinary retention

8

Page 8 of 22

underwent a combined hysterectomy (8.4% vs. 13.1%, SD 15%). A total of 1,648/59,566 (2.8%) women underwent multiple midurethral slings during the study period.

Primary Analysis

Overall, 1,598/59,556 (2.7%) women had at least one code representing a MUS problem requiring operative revision. A total of 1,503 women underwent a procedure for mesh removal, erosion or fistula, and 162 underwent isolated urethrolysis (67 women underwent both procedures). Our unadjusted results are shown in Table 2: a total of 2,025/59,556 (3.4%) patients presented to the emergency room or required hospital admission for post-operative urinary retention (the time between sling placement and retention is shown in Appendix 4). Among patients without postoperative retention, 1,386/57,531 (2.4%) had subsequent surgery for mesh complications compared to 212/2025 (10.5%) of those with postoperative urinary retention. This represents an absolute risk increase of 8.1% (95% CI 6.7-9.4%) and an unadjusted HR of 4.43 (95% CI 3.83-5.12). The magnitude of risk was higher among women with retention between 3-30 days after sling insertion (HR 5.41, 95% CI 4.55-6.43) compared to retention 0-2 days after sling insertion (HR 3.24 (95% CI 2.55-4.12). A total of 11,747/59,556 (19.7%) of women presented a total of 22,195 times to an emergency room or hospital after sling insertion for UTI symptoms, and of these women, 366/11,747 (3.1%) underwent a revision for a MUS problem (Table 2); the unadjusted HR was 1.94 (95% CI 1.81-2.07)

9

Page 9 of 22

for each additional UTI (up to a maximum of 3). Both post-operative retention and UTI symptoms were identified in 786/59,556 (1.3%) of the women.

Our unadjusted results when considering the outcomes of urethrolysis and mesh erosion/removal/fistula separately are shown in Table 3. Post-operative urinary retention is strongly predictive of a future urethrolysis (HR 9.67, 95% CI 6.80-13.75), and is also significantly elevated for the other mesh problems (HR 3.92, 95% CI 3.35-4.58).

The results from our multivariable analysis were similar in magnitude to our unadjusted analysis (summarized in Table 4, full results included in Appendix 5). There was a greater than 3-fold increase in mesh problems (HR 3.46, 95% CI 2.97-4.02) among women with post-operative urinary retention. When excluding patients who underwent an isolated urethrolysis, the HR remains similar (3.08, 95% CI 2.62-3.63).

There was a statistically significant increase in the risk of mesh complications as the number of post-operative UTI related visits increased in unadjusted analysis (Table 2), and adjusted analysis (Table 4), and post-operative UTI visits were associated with both urethrolysis and other mesh complications when these outcomes were considered individually. When we compared the number of post-operative UTI visits relative to women with no post-operative UTI visits, there was a sequential increase in the adjusted hazard ratio (1 versus 0 UTIs: HR 2.41, 95%CI 2.07-2.79; 2 versus 0 UTIs: HR 3.32, 95%CI 2.57-4.29; ≥3 versus 0 UTIs: HR 4.01, 95%CI 3.03-5.32). Our post-hoc analysis of outpatient physician visits for UTI symptoms (identified in 25,309/59,556

10

Page 10 of 22

(42.5%) of women) revealed an unadjusted HR of 1.32 (95% CI 1.29-1.35) per outpatient UTI (with those having ≥10 classified in one group), and an adjusted HR of 1.29 (95% CI 1.26-1.31), with results for the separate outcomes of 1) removal/erosion/fistula and 2) urethrolysis both similar in magnitude and significance (Appendix 7).

COMMENT

MUS have become the gold standard for stress incontinence but they are associated with problems which may require surgical revision1. There is limited data evaluating specific presentations which may be associated with these complications, and the magnitude of any association. We found that both post-operative urinary retention (within 30 days of the MUS), as well as a hospital or emergency room visit for UTI symptoms was associated with future mesh problems requiring surgery. Given the intuitive relationship between post-operative urinary retention and urethrolysis, we stratified our outcome to examine only mesh removal, erosion or fistula (excluding isolated urethrolysis), and demonstrated a continued statistically significant association in both of these outcome groups. Consistent with our previous work, low surgeon volume, younger age, multiple mesh slings, and concomitant hysterectomy were associated with future mesh removal or revision12.

Both the management and definition of post-operative urinary retention and voiding dysfunction after a MUS are inconsistent2. A recent survey of both Urologists and

11

Page 11 of 22

Urogynecologists presented with various MUS post-operative complications revealed no clear consensus on management strategies5. In practice, many obstructed women are eventually able to void without using catheters, however upon careful questioning they report significant voiding dysfunction that persisted after the initial period of obvious retention2,5. A careful history can identify new symptoms after the MUS such as prolonged voiding time, positional voiding, abdominal straining, incomplete emptying, and storage symptoms. Authors have advocated for temporizing measures (such as short term indwelling catheterization, or 4-6 weeks of clean intermittent catheterization) or an early urethral diltation2,6. Other authors tend to offer early reoperation (which involves loosening the sling or urethrolysis/sling removal) to avoid the long term consequences of partial obstruction and voiding dysfunction2,9,14,15. Studies have suggested that earlier intervention for iatrogenic urethral obstruction may be associated with better outcomes7. A theoretical etiology for urethral erosion of mesh is excessive tension (which may initially manifest as post-operative urinary retention), and this may explain the significant association we demonstrated between post-operative urinary retention and procedures such as mesh removal and fistula, which for some women were carried out years after MUS placement2,10. Whether early revision of the MUS changes the risk of future mesh complications is unknown, and the optimal timing for reoperation is unclear. In our cohort surgical intervention for MUS problems occurred a median of 14 months after sling insertion, and a significant number were treated years later, suggesting a delayed diagnosis or presentation of the potential MUS problem. Our results confirm the importance of post-operative urinary retention as a predictor of future MUS problems, and these patients should be carefully followed for complications. The

12

Page 12 of 22

risk is higher for patients with retention presenting after the initial 48hrs post-sling, presumably as the anesthesia and narcotics contribute to retention in this very early period.

The sequential increase in the magnitude of risk of MUS revision or removal with each subsequent post-operative hospital or emergency room visit for UTI symptoms (similar to our post-hoc analysis of outpatient UTIs) is one of the factors suggesting a true relationship between exposure and outcome. This aligns with the frustrations described by many women as they attempt to seek treatment for symptoms after a MUS (which are often treated by non-MUS surgeons as UTIs). In this study, women with at least one hospital presentation of UTI symptoms had a prolonged time between MUS placement and revision or removal of the MUS (median 2.6 years). True bacterial UTIs may be a result of voiding dysfunction and altered bladder mechanics or a result of a subtle erosion and colonization of the mesh11. Other causes of UTI symptoms aside from bacterial cystitis (such as overactive bladder symptoms, pelvic floor spasm, and obstruction) could all contribute to these UTI symptoms. Studies have suggested that UTIs may be more frequent with MUS compared to other incontinence procedures such as colposuspension procedures16. However UTIs may also cause mesh complications, as some authors have suggested that early UTIs may actually disrupt tissue healing and predispose patients to future mesh complications11. It is likely that some of these women likely presented with symptoms mimicking UTIs but were actually related to noninfectious problems with their MUS. This study also demonstrates the more frequent contact with the emergency room that a small number of these women with UTI

13

Page 13 of 22

symptoms have after their midurethral sling, and this may be a result of a delay in being reassessed by a surgeon comfortable dealing with midurethral sling complications.

There are a few important implications of this research. Firstly, women who experience post-operative urinary retention or frequent UTI symptoms after a MUS should be followed by their surgeon to assess for both early or late mesh related complications or voiding dysfunction. Even a patient with a “successful” trial of void may be at an increased risk for a mesh related complication in the future. Secondly, women who are assessed for problems after a MUS should be questioned about episodes of UTI symptoms and post-operative urinary retention, and if present these factors should lead to a higher level of suspicion of mesh complications and may warrant further work-up and investigation. Finally, future research should evaluate whether early intervention for post-operative urinary retention (within 24-48hrs) is superior to conservative therapy in terms of patient outcome and future complications, and whether different recommendations should apply for various MUS types such as retropubic or transobturator slings.

There are limitations inherent to our study. Administrative data allows us to evaluate a large population of patients managed in the real world setting and outcomes can be ascertained with only a small number of patients lost to follow-up. However, this means that potentially relevant clinical details (such as type of MUS, degree of post-operative voiding dysfunction, and the exact nature and success of the removal/revision surgery) are not measurable. While our administrative data study demonstrates a tendency

14

Page 14 of 22

towards eventual complications among women with UTI symptoms or post-operative retention, further clinical study is needed on this topic. The risk of retention and voiding dysfunction is higher with retropubic midurethral slings compared to transobturator slings17; early retention after placement of these slings may have different magnitudes of risk for future sling surgery based on the type of midurethral sling. Misclassification of our study outcome is always possible, however given our two risk factors had consistent coding practices throughout the study period, the risk of differential misclassification is low. To maximize the validity of our definition, post-operative retention and UTI symptoms were defined as patients who presented to the emergency department and/or required hospital admission. These definitions likely underestimate the true incidence of these events (as patients managed in the outpatient setting, or by outpatient nursing services would not be captured), however this would tend to bias our results towards the null and would suggest our HR and absolute risk differences actually underestimates the true value. In our sub-analysis we separated our outcome into those patients who specifically underwent a urethrolysis versus surgery for a mesh complication; however, it is likely that some women who underwent mesh removal or revision for another reason also had a degree of obstruction.

CONCLUSION

Following a midurethral sling, post-operative urinary retention (in the first 30 days), and UTI symptoms presenting to a hospital are associated with future sling revision or removal. Physicians should be aware that women with early urinary retention after a

15

Page 15 of 22

midurethral sling appear to have a higher risk of future sling surgery. Women with UTI symptoms who present multiple times should likewise be considered for further evaluated to ensure there are no complications from the prior sling.

16

Page 16 of 22

REFERENCES

1.

Cox A, Herschorn S, Lee L. Surgical management of female SUI: is there a gold

standard? Nat Rev Urol 2013;10:78-89. 2.

Cetinel B, Tarcan T. Management of complications after tension-free midurethral

slings. Korean J Urol 2013;54:651-9. 3.

Forzini T, Viart L, Alezra E, Saint F. [Erosive complications of mid urethral slings

(MUS): 10 years of surgical experience]. Prog Urol 2015;25:240-8. 4.

Lee D, Zimmern PE. Management of complications of mesh surgery. Curr Opin

Urol 2015;25:284-91. 5.

Hashim H, Terry TR. Management of recurrent stress urinary incontinence and

urinary retention following midurethral sling insertion in women. Ann R Coll Surg Engl 2012;94:517-22. 6.

Sokol AI, Jelovsek JE, Walters MD, Paraiso MF, Barber MD. Incidence and

predictors of prolonged urinary retention after TVT with and without concurrent prolapse surgery. Am J Obstet Gynecol 2005;192:1537-43. 7.

Abraham N, Makovey I, King A, Goldman HB, Vasavada S. The effect of time to

release of an obstructing synthetic mid-urethral sling on repeat surgery for stress urinary incontinence. Neurourol Urodyn 2015. 8.

Abbott S, Unger CA, Evans JM, et al. Evaluation and management of

complications from synthetic mesh after pelvic reconstructive surgery: a multicenter study. Am J Obstet Gynecol 2014;210:163 e1-8.

17

Page 17 of 22

9.

Klutke C, Siegel S, Carlin B, Paszkiewicz E, Kirkemo A, Klutke J. Urinary

retention after tension-free vaginal tape procedure: incidence and treatment. Urology 2001;58:697-701. 10.

Sergouniotis F, Jarlshammar B, Larsson PG. Urethral complications after

tension-free vaginal tape procedures: A surgical management case series. World J Nephrol 2015;4:396-405. 11.

Mellano EM, Nakamura LY, Choi JM, et al. The Role of Chronic Mesh Infection in

Delayed-Onset Vaginal Mesh Complications or Recurrent Urinary Tract Infections: Results From Explanted Mesh Cultures. Female Pelvic Med Reconstr Surg 2016. 12.

Welk B, Al-Hothi H, Winick-Ng J. Removal or Revision of Vaginal Mesh Used for

the Treatment of Stress Urinary Incontinence. JAMA Surg 2015;150:1167-75. 13.

Austin PC. The performance of different propensity-score methods for estimating

differences in proportions (risk differences or absolute risk reductions) in observational studies. Stat Med 2010;29:2137-48. 14.

South MM, Wu JM, Webster GD, Weidner AC, Roelands JJ, Amundsen CL.

Early vs late midline sling lysis results in greater improvement in lower urinary tract symptoms. Am J Obstet Gynecol 2009;200:564 e1-5. 15.

Clifton MM, Linder BJ, Lightner DJ, Elliott DS. Risk of repeat anti-incontinence

surgery following sling release: a review of 93 cases. J Urol 2014;191:710-4. 16.

Rehman H, Bezerra CC, Bruschini H, Cody JD. Traditional suburethral sling

operations for urinary incontinence in women. Cochrane Database Syst Rev 2011:CD001754.

18

Page 18 of 22

17.

Richter HE, Albo ME, Zyczynski HM, et al. Retropubic versus transobturator

midurethral slings for stress incontinence. N Engl J Med. 2010;362(22):2066-2076. doi:10.1056/NEJMoa0912658.

Table 1. Cohort Demographics Entire Cohort

Age (median, IQR) Surgeon Volume** Low High Surgeon Type Urologist OB/GYN Unclassified Teaching Hospital Obesity (BMI >40) Diabetes Mellitus Neurogenic Disease UTI in the prior year 0 1 2 3 Procedures Prior Cystoscopy Prior Urodynamics (UDS) Hysterectomy Previous*** Combined Pelvic Organ Prolapse Surgery (POP) Previous*** Combined High Risk Features****

30 Day PostOp Retention

(n=59,556) 52 (45-63)

(n=2,025) 51 (43-61)

No 30 Day Post-Op Retention (n=57,531) 52 (45-63)

SD*

26.8% (15,945) 73.2% (43,611)

30.0% (607) 70.0% (1,418)

26.7% (15,338) 73.3% (42,193)

7% 7%

42.6% (25,379) 56.9% (33,880) 0.5% (297) 25.5% (15,181) 4.7% (2,819) 12.3% (7,341) 0.6% (369)

54.9% (1,112) 44.7% (906) 0.3% (7) 20.4% (414) 3.5% (70) 12.2% (247) 0.9% (19)

42.2% (24,267) 57.3% (32,974) 0.5% (290) 25.7% (14,767) 4.8% (2,749) 12.3% (7,094) 0.6% (350)

26% 25% 3% 13% 7% 0% 3%

95.1% (56,656) 4.2% (2,472) 0.5% (325) 0.2% (103)

92.1% (1,865) 6.0% (121) 1.4% (28) 0.5% (11)

95.2% (54,791) 4.1% (2,351) 0.5% (297) 0.2% (92)

13% 9% 9% 5%

56.6% (33,727) 39.3% (23,394)

62.0% (1,256) 35.1% (711)

56.4% (32,471) 39.4% (22,683)

11% 9%

8.6% (5,145) 12.9% (7,688)

9.4% (191) 8.4% (171)

8.6% (4,954) 13.1% (7,517)

3% 15%

5.7% (3,385) 29.4% (17,510) 0.1% (82)

6.8% (138) 26.8% (542) NR

5.6% (3,247) 29.5% (16,968) NR

5% 6% 3%

14%

19

Page 19 of 22

Further details are provided in Appendix 3. There were differences in fiscal year of cohort entry, rural residence, and Aggregated Diagnostic Group Resource Utilization Bands. There was significant regional variation across the province. *SD = Standardized differences of the mean are less sensitive to large sample sizes; a value >10% is considered a potentially meaningful difference between groups. Due to the large sample size of administrative data studies, p values often identify small, statistically significant differences which are not clinical relevant. **High Surgeon Volume Defined as >75th percentile for the year ***Previous surgery includes the time period of October 1, 1992 until 2 days prior to index date ****High Risk Features includes prior fistula, urethral diverticulum, pelvic radiation or urethral injury

Table 2. Primary Outcomes Based on Primary and Secondary Risk Factors Entire No 30 Day 30 Day No Post-Op Post-Op Cohort Post-Op Post-Op UTI UTI Retention Retention (n=59,556) (n=57,531) (n=2,025) (n=47,809) (n=11,747) All Complications n (%) 1,598 1,386 212 1,232 366 (3.1%) (2.7%) (2.4%) (10.5%) (2.6%) Follow-Up (py) 368,061 355,157 12,904 319,047 49,014 Event Rate (per 4.3 3.9 16.4 3.9 7.5 1000py) Years to Outcome, 1.20 1.31 0.41 0.97 2.60 median (IQR)* (0.38-3.42) (0.45-3.65) (0.12-1.64) (0.35-2.75) (0.90-5.25) Unadjusted Analysis Hazard Ratio 1.00 (ref) 4.43 1.00 (ref) 1.94** (95% CI) (3.83-5.12) (1.81-2.07) p-value

<0.01

<0.01

*The time to outcome is illustrated with boxplots in Appendix 6 of the supplement. **Modelled as a continuous variable with step-wise increases from 0, 1, 2 ≥3 postoperative UTIs. Interpreted as each additional UTI (to a maximum of 3) increases the hazard of future mesh complications by 94%.

20

Page 20 of 22

Table 3. Secondary Outcomes Separated by Removal/Erosion/Fistula and Urethrolysis Based Stratified by Post-Operative Retention and Urinary Tract Infections Entire No 30 Day 30 Day No PostPost-Op Cohort Post-Op Post-Op Op UTI UTI Retention Retention (n=59,556) (n=57,531) (n=2,025) (n=47,809) (n=11,747) Removal/Erosion/ Fistula n (%) 1,503 1,322 181 1,166 337 (2.5%) (2.3%) (8.9%) (2.4%) (2.9%) Follow-Up 368,675 355,584 13,091 319,486 49,189 (person years) Event Rate 4.1 3.7 13.8 3.6 6.9 (per 1000 person-years) Years to 1.25 1.37 0.49 1.00 2.82 Outcome (IQR)* (0.40-3.54) (0.45-3.69) (0.13-1.70) (0.36-2.86) (1.02-5.33) Unadjusted Analysis Hazard Ratio 1.0 (ref) 3.92 1.0 (ref) 1.90** (95% CI) (3.35-4.58) (1.77-2.04) p-value <0.01 <0.01 Urethrolysis n (%) 162 120 42 108 54 (0.3%) (0.2%) (2.1%) (0.2%) (0.5%) Follow-Up 374,288 360,601 13,687 324,125 50,164 (person-years) Event Rate 0.4 0.3 3.1 0.3 1.1 (per 1000 person-years) Years to 0.96 1.34 0.25 0.81 1.81 Outcome (IQR)* (0.31-3.06) (0.57-3.47) (0.10-1.07) (0.24-2.11) (0.36-4.78) Unadjusted Analysis Hazard Ratio 1.0 (ref) 9.67 1.0 (ref) 2.55** (95% CI) (6.80-13.75) (2.12-3.07) p-value <0.01 <0.01 *The time to outcome is illustrated with boxplots in Appendix 6 of the supplement. **Modelled as a continuous variable with step-wise increases from 0, 1, 2 ≥3 postoperative UTIs. Interpreted as each additional UTI (to a maximum of 3) increases the hazard of future mesh complications by 2.5 fold.

21

Page 21 of 22

Table 4. Multivariable Analysis All Complications (n=1,598) Hazard pRatio value (95%CI) Post-Operative Retention Urinary Tract Infection Age (per year increase) Combined POP Combined Hysterectomy Diabetes Mellitus Future Additional MUS Neurogenic Disease Obesity Prior UTI Surgeon Type (ref=Urologist) Gynecologist

3.46 (2.97-4.02) 1.74 (1.61-1.87) 0.99 (0.98-0.99) 1.06 (0.93-1.20) 0.80 (0.67-0.96) 1.03 (0.88-1.20) 3.31 (2.81-3.89) 1.42 (0.89-2.27) 0.83 (0.64-1.07) 1.12 (0.98-1.29)

<0.01 <0.01 <0.01 0.37 0.01 0.70 <0.01 0.14 0.15 0.11

Removal/Erosion/Fistula (n=1,503) Hazard Ratio p-value (95%CI) 3.08 (2.62-3.63) 1.74 (1.61-1.87) 0.99 (0.98-0.99) 1.04 (0.91-1.18) 0.81 (0.67-0.96) 1.02 (0.87-1.20) 3.37 (2.86-3.97) 1.46 (0.90-2.36) 0.77 (0.59-1.02) 1.10 (0.95-1.28)

<0.01 <0.01 <0.01 0.56 0.02 0.83 <0.01 0.12 0.07 0.19

Urethrolysis (n=162) Hazard pRatio value (95%CI) 6.63 (4.54-9.67) 1.99 (1.62-2.44) 1.01 (1.00-1.03) 1.31 (0.89-1.92) 0.86 (0.47-1.59) 1.07 (0.68-1.68) 3.48 (2.10-5.78) 1.33 (0.33-5.44) 1.74 (0.95-3.19) 0.95 (0.64-1.40)

<0.01 <0.01 0.07 0.17 0.63 0.77 <0.01 0.69 0.07 0.78

1.05 0.45 1.12 0.07 0.53 <0.01 (0.93-1.18) (0.99-1.27) (0.36-0.79) Surgeon 0.74 <0.01 0.72 <0.01 0.88 0.51 Volume (High)* (0.66-0.82) (0.65-0.81) (0.61-1.28) Teaching 1.10 0.18 1.07 0.35 1.43 0.13 Hospital (0.96-1.26) (0.93-1.23) (0.90-2.28) Further details are provided in Appendix 4. A higher Aggregated Diagnostic Group Resource Utilization Band was associated with an increased risk for mesh complications, and there was significant regional variation. *Defined as >75th percentile of MUS implanters in the given year

22

Page 22 of 22