Risk factors for 5-year complications after midurethral sling surgery for stress urinary incontinence: a retrospective cohort study from Taiwan

Midurethral sling surgery is the current gold standard worldwide for stress urinary incontinence (SUI) surgery, with over 90% of surgeons worldwide using the midurethral sling for SUI between 2008 and 2018. However, concerns surround mesh-related adverse events associated with the midurethral sling. The decision to use the midurethral sling for surgical treatment has become a challenging one for clinicians, surgeons and patients. We sought to determine the factors for 5-year complications after midurethral sling surgery, to improve the clinical decision-making process. Records were reviewed from a total of 1961 female patients who underwent their first midurethral sling surgery for SUI between 2003 and 2018 at a single teaching hospital in Taiwan. A multivariable Cox proportional hazard model calculated the hazard ratios of risk factors for surgical complications, after adjusting for confounders. Surgical complications (i.e., secondary surgery and urinary retention) occurred in 93 (4.7%) patients within 5 years following the index operations. These patients were more likely to be older, to have a history of menopausal syndrome within 1 year prior to the index operation, a medication history of oral antidiabetic drug use, hormone replacement therapy (HRT), slower average flow rate, and longer voiding time compared with patients without surgical complications. In the multivariate analysis, HRT (adjusted hazard ratio, 1.787; 95% confidence interval, 1.011–3.158, p = 0.04) was significantly associated with surgical complications at 5 years, after adjusting for age, gender, diabetes, menopause syndrome, average flow rate, and sling type. Our findings suggest that a medication history of HRT may be a risk factor associated with surgical complications, especially urinary retention, at 5 years in women undergoing midurethral sling surgery for SUI.


Study population
The China Medical University Hospital-Clinical Research Data Repository (CMUH-CRDR) was established by the Big Data Center of CMUH, which contains the medical records of 2,873,887 patients who sought care in CMUH between 2003 and 2018.The CMUH-CRDR has been described in greater detail in previous articles 19,20 .All patients enrolled in the CMUH-CRDR were followed up until December 31, 2018, or death, whichever occurred earlier.This study was approved with waived informed consent by the Big Data Center of CMUH and the Research Ethical Committee/Institutional Review Board of CMUH (CMUH105-REC3-068 & CMUH111-REC3-138).
In this retrospective cohort study, we identified 2789 patients from the CMUH-CRDR records who underwent 2834 midurethral sling surgery procedures between 2003 and 2018.Miduretheral sling surgeries included in this study were tension free vaginal tape (TVT) or trans-obturator tension free vaginal tape (TVTO), which are the current gold standard surgical treatment 8,9 .Next, we excluded operations without SUI diagnosis records during the index visit or within six months prior to the procedure date, patients whose surgical visits were not in the Department of Obstetrics and Gynecology, and patients who did not have any surgical records or who did not have sling type records on their surgical notes during the index visit (Fig. 1; Supplementary Table 1).The final

Definition of risk factors
Demographic information, comorbidities, medication histories, biochemical profiles, reports of urodynamic testing and surgical procedures were collected from the CMUH-CRDR.Comorbidities of diabetes mellitus, hypertension, cardiovascular diseases, psychosis, menopausal syndrome and constipation were defined by the International Classification of Diseases, 9th/10th Revisions, Clinical Modification (ICD-9-CM and ICD-10-CM) codes, and medication histories for hypertension, diabetes, and hormone replacement therapy (HRT) were recorded within 1 year before the index midurethral sling surgery (Supplementary Table 1).Biochemical profiles of blood glucose, liver function, renal function, and complete blood count were obtained within 1 year prior to the index procedures.Reports of urodynamic testing, such as uroflowmetry information and urethra pressure profiles, were obtained within 3 years prior to the index procedures.Surgical information extracted from the operation database included length of operation, American Society of Anesthesiology (ASA) score, wound contamination classification, anesthesia type, and sling type.

Definition of outcomes
The primary outcome of interest was surgical composite complications (including secondary surgery and urinary retention) within 5 years following the index procedures.Secondary surgery was defined as the patient undergoing another operation (i.e., TVT/TVTO) within 5 years due to recurrent SUI; urinary retention was defined by ICD-9-CM and ICD-10-CM codes (Supplementary Table 1).

Statistical analysis
Continuous variables are presented as medians and interquartile ranges (IQRs); categorical variables are reported as frequencies and proportions (%).The associations between surgical complications and covariates were analyzed using the Wilcoxon rank-sum test (nonparametric) for continuous variables and the chi-square test for categorical variables.
We evaluated the associations of the risk factors with the risk of 5-year surgical complications by using a competing risk analysis with deaths considered censoring events using cause-specific Cox proportional hazards modeling and used age as the time scale.We included age, diabetes mellitus, duration of surgery, menopausal syndrome, HRT, average flow rate (AFR), voiding time and sling type in the Cox model.Subgroup analyses were performed for patients categorized by age groups and diabetes status.The Kaplan-Meier method was employed to generate time-to-event curves throughout the study period, and the log-rank test was used to assess differences between the curves of different subgroups.

Results
We identified 1961 women with SUI who underwent first-time midurethral sling surgery between 2003 and 2018 in CMUH, with a median age of 53.5 years (IQR, 46.4-64.0;Table 1).Of them, 21.8% were concomitant procedure with pelvic floor reconstruction and 16.7% with hysterectomy.Surgical complications (i.e., secondary surgery and urinary retention) occurred in 93 (4.7%) patients within 5 years following the index operations.The mean follow-up time after the miduretheral sling surgery was 1754 days (standard deviation, 340 days).
Patients with surgical complications were more likely to be older (60.3 years vs 53.1 years), to have a history of menopausal syndrome within 1 year prior to the index operation (16.1% vs 9.6%), a medication history of oral antidiabetic drug (OAD; 8.6% vs 3.9%), HRT (32.3% vs 13.7%), slower AFR (8.0 ml/s vs 10.0 ml/s), and longer voiding time (45.4 s vs 33.0 s), compared with patients without 5-year surgical complications.
We used multivariable Cox proportional hazard modeling to investigate the risk factors associated with 5-year surgical complications (Table 2).After including age, diabetes mellitus, duration of surgery, menopausal syndrome, HRT, AFR, and sling type in the full model, only HRT (adjusted HR, 1.787; 95% CI, 1.011-3.158,p = 0.04) remained significantly associated with the risk of 5-year surgical complications.Similarly, when combining menopausal syndrome and HRT as one variable in the model, the effect size stayed significant (aHR, 1.794; 95% CI, 1.033-3.114;Supplementary Table 2).The adjusted time-to-event curves demonstrated that patients with HRT had a significant higher probability of developing 5-year surgical complications than those without menopausal syndrome or HRT (Fig. 2).

Discussion
This large retrospective cohort study found that 4.7% of patients receiving miduretheral sling surgery had unfavorable composite complications (i.e., urinary retention or secondary surgery) within 5 years.History of HRT was the only significant risk factor of 5-year surgical complication, with an almost two-fold increased risk.The    association between HRT remained significant in the subgroup of patients without diabetes.In addition, HRT was significantly associated with increased risk of urinary retention when we did separate analysis on surgical complication.Several interesting findings are worth for discussion and further investigation.

Surgical complications following midurethral sling surgery
A global review of epidemiological research on SUI that included studies published between January 1980 and October 2002 reported a median prevalence of female UI of 27.6%, most commonly caused by stress (50%) 21 .
Population-based prevalence rates of SUI among Chinese women range from 6.7 to 44% 22 , while one Taiwanese study has reported a prevalence of 18.0% based on patients' perceptions 23 .Midurethral sling surgery is the current gold standard worldwide for SUI surgery; however, concerns surround mesh-related adverse events associated with the midurethral sling.One retrospective cohort study that identified 188,454 eligible women who underwent an index sling surgery showed that the 9-year risk of sling revision/removal was relatively low at 3.7%, with a 9-year risk of 1.3% (95% CI, 1.2-1.4)for urinary retention and the majority occurred within 4 years after the surgery 24 .Similarly, in our study, a relatively low proportion of patients (4.7%) developed surgical complications within 5 years of undergoing midurethral surgery.

Risk factors associated with surgical complications
We identified factors that potentially place patients at risk of experiencing surgical complications after midurethral sling surgery and are worthy for being discussed.

Menopausal syndrome
A 2012 review of the literature on the epidemiology of UI in women and the effects of HRT on urinary leakage found that UI was a common symptom during menopause 25 .Decreased estrogen concentrations associated with menopause have been considered to be responsible for the increasing prevalence of SUI in aging women, possibly because vaginal tissue is weaker in postmenopausal women than in premenopausal women and thus becomes a risk factor for the deterioration of continence mechanisms and consequently the efficacy of anti-incontinence surgery 26,27 .Although menopausal syndrome is a recognized risk factor for SUI, scant study evidence describes the effects of menopausal syndrome on surgical outcomes.A retrospective study from Turkey that investigated mesh erosion after tension free vaginal tape (TVT) and transobturator tape (TOT) found that menopausal status was a statistically significant factor in patients with mesh erosion, but was no longer a significant independent risk factor after multivariate analysis 28 .In contrast, our study found that patients who had a history of menopausal syndrome within 1 year prior to the index operation were more likely to develop surgical complications.Moreover, menopausal syndrome or HRT were significantly associated with the risk of 5-year surgical complications (aHR, 1.794; p = 0.0390), although the association was not significant in patients aged ≤ 65 years (aHR, 2.164; p for interaction = 0.3750).Estrogen increases angiogenesis, which is important for nourishing vaginal tissue.However, a negative correlation has been observed between angiogenic activity and mesh-induced inflammation in mice implanted with steroid-coated polyvinylidenfluoride (PVDF) meshes 29 .At menopause, decreased estrogen levels lead to a reduction in angiogenesis and therefore poor nourishment of vaginal tissue, which would increase the possibility of surgical complications.

Hormone replacement therapy
International guidance recommends conservative therapies as the first-line management of women with SUI, such as lifestyle changes and behavioral therapies, weight loss, and supervised pelvic floor muscle training 8,9 .Estrogens are believed to be beneficial in the treatment of SUI; starting estrogen replacement soon after menopause may be effective in preventing or delaying the onset of SUI 27,30 .
HRT is a common medical treatment used to supplement women with hormones that are lost during the menopausal transition.Conventional HRT includes an estrogen and progesterone component to relieve the menopausal syndrome 31 .In our study, patients with diagnosed menopausal syndrome who used HRT within a year prior to midurethral sling surgery had a significantly higher risk of surgical complications, especially Table 3. Adjusted hazard ratios of 5-year surgical complications after midurethral sling surgery, stratified by subgroups of age and diabetes mellitus.CI, confidence interval; HR, hazard ratio; TOT, transobturator tape; TVT-O, tension-free vaginal tape-obturator.*Statistically significant.a P for interaction with hormone replacement therapy was 0.0807.urinary retention, at 5 years following midurethral sling surgery for SUI compared with their counterparts who did not have these risk factors.As patients diagnosed with menopausal syndrome commonly receive HRT to relieve their symptoms, it is difficult to distinguish between the effects of HRT and aging-related menopausal syndrome upon the risk of surgical complications associated with midurethral sling surgery.Although vaginal estrogen may be effective in preventing or delaying SUI, no evidence has shown that oral estrogen has benefits in SUI patients.A retrospective study in 2018 had found that the use of systemic estrogen may increase the SUI risk 32 .Estrogens are known to stimulate collagenase activity, which may lead to degradation of total collagen, especially the most supportive type I collagen, which may be replaced by weaker immature collagen 32 .The lack of mature collagen might lead to inadequate support of the vaginal and pelvic structure, especially postoperatively, increasing the possibility of recurrent SUI.Estrogens may affect continence by several mechanisms, including increasing urethral resistance, raising sensory threshold of the bladder, increasing adrenoreceptor sensitivity in the urethral smooth muscle, and mediating relaxation of the detrusor muscle by adrenoceptor promotion 33,34 .Moreover, estrogen receptors are presented in the squamous epithelium of both proximal and distal urethra, and estrogen supplementary has been shown to improve the maturation index of urethral squamous epithelium 35 .Thus, the HRT application prior to the surgery may potentially exacerbate the continence and eventually cause urine retention.

AFR and voiding time
According to our univariable findings, AFR and voiding times from the urodynamic exam were worse in patients with surgical complications.Preoperative urodynamic evaluations may predict the risk of voiding dysfunction in women with SUI undergoing midurethral sling surgery 36 , although other research has failed to support such evaluations 37 .Another study from Taiwan has found that an abnormal preoperative uroflowmetry pattern and preoperative peak flow < 15 ml/s are risk factors contributing to voiding dysfunction following midurethral sling surgery 38 , and a low flow rate is reported by other researchers to be a risk factor for early voiding dysfunction postoperatively 39 .In our study, all of our patients (with and without surgical complications) had much lower AFR values (8.0 vs 10.0 ml/s, respectively) than the studies mentioned above.The discrepant outcomes may depend upon different criteria used by the various studies to determine low flow rates, as well as the differing severity of SUI in the various study populations.

Study strengths and limitations
The strengths of this study include the careful efforts made to verify the study population to avoid selection bias.Because the sling type data were recorded only in unstructured text and could not be easily identified, an Obstetrics and Gynecology (OBGYN) physician (TFC) reviewed the operation records of all eligible patients to verify that these patients indeed received midurethral sling surgery.Another strength is that this study was able to examine the association between urodynamic parameters and surgical outcomes, as OBGYN physicians in CMUH regularly perform detailed urodynamic evaluations prior to surgery.Interestingly, a Cochrane review has concluded that urodynamic studies may change clinical decision making, but scant evidence shows that these evaluations improve clinical outcomes [40][41][42] .In our unadjusted analysis, we identified that lower average flow rate and longer voiding time were associated with surgical complications, which, if validated in another study, may serve as quantitative indicators for a high risk of surgical complications.
This study has several limitations.First, residual and unmeasured confounders could not be entirely excluded.For example, we may have misclassified patients with menopausal syndrome as being without menopausal syndrome, because we were reliant on ICD records from different doctors and from a retrospective single center database.Therefore, the effects of potential risk factors (such as patient lifestyle factors, physician bias and experience with recording menopausal symptoms) may have been slightly overestimated due to the underestimation of these positive confounders.Second, the standard protocol for follow-up is usually 6 months following surgery; we may have misclassified patients with surgical complications as instead belonging to the group without complications, because these patients may not come back to our hospital for care so were lost to follow-up.Third, the associations found in our study do not guarantee causality.Our findings were not verified in other populations under different healthcare systems and the results should be externally validated to prove generalizability.

Conclusion
Our study found that a medication history of HRT was a risk factor for 5-year complications in women undergoing midurethral sling surgery for SUI.If this association is validated in future studies, it could provide guidance for surgeons, clinicians and patients assessing the risk of surgical complications when considering midurethral sling surgical treatment for SUI.

Table 1 .
Demographic and clinical characteristics of patients with stress urinary incontinence who received surgical treatment.aCategoricalvariables are presented as frequencies (%) and continuous variables are presented as medians (Q1, Q3), if not otherwise specified.bThe5-year surgical complications included secondary surgery and urine retention.cP-valuesare calculated by the Wilcoxon rank-sum test for continuous variables and the chi-square test for categorical variables.dDiabetesmellitus and hypertension were defined as having the respective diagnosis and medication within 1 year before the index date.Cardiovascular diseases, psychosis, menopause syndrome and constipation were defined by the diagnosis within 1 year before the index date.Stroke history was defined as having the diagnosis before the index date.Detailed definitions are listed in the Supplemental Table1.