Risk Factors for Returning to the Operating Room for a Second Surgery After Midurethral Sling for Stress Urinary Incontinence

Abstract

Objectives:

The objective of this study was to identify risk factors for having to return to the operating room for a second surgery after midurethral sling (MUS).

Methods:

We used a case-control design. Cases return to operating room were a composite of 6 surgical complications or recurrent stress urinary incontinence because we believed that women would consider return to the operating room (OR) a similar MUS-related complication regardless of indication. Cases were obtained from Cerner Health Facts database, including 213 hospitals, using current procedural technology codes 57288 (repeat sling), 57287 (sling revision), and 53500 (urethrolysis) for procedures after index MUS. Controls no return to OR were randomly selected in 4:1 ratio from the remaining slings without these procedures. Multivariable regression analysis included all variables with P < 0.10 on univariable analysis.

Results:

Between January 1, 2010, and December 31, 2016, 1247 patients returned to the OR of 17,953 patients who underwent initial MUS (6.9%). After adjusting for confounders, white race (OR, 1.47 [1.20–1.81]), lack of concomitant prolapse surgery (OR, 1.37 [1.18–1.59]), immunosuppressant drugs (OR, 1.27 [1.12–1.45]), and blood thinner use (OR, 1.38 [1.18–1.62]) significantly impacted the odds for returning to the OR. Anticholinergic use and smoking tobacco or marijuana, although significant on univariable analysis, were no longer significant after adjusting for confounders.

Conclusions:

The rate of a second surgery after MUS using a composite outcome, over a 7-year period including multiple diagnoses, is 6.9%. White race, using immunosuppressant drugs, using blood thinners, and not having concomitant prolapse surgery are all risk factors for having second surgery after MUS.

Stress urinary incontinence (SUI) affects up to 25% of women.¹ The most common surgical procedure for SUI is the midurethral sling (MUS) surgery. Compared with other surgical approaches for SUI, midurethral slings are associated with less pain, a greater likelihood of being performed as an outpatient procedure, decreased costs, and quicker return to activities.² Still, approximately 3.5% of patients experience a postoperative complication.³

Despite the success of the MUS, the reoperation rate has been reported as high as 8.6%.⁴ Reoperation can include a sling revision, which usually involves incising, or removing a portion of or all of the mesh, urethrolysis, which usually involves lysis of scar tissue adjacent to the urethra, or a repeat anti-incontinence procedure. Indications for sling revision or urethrolysis include mesh exposure, urinary retention, voiding symptoms, recurrent urinary tract infection (UTI), or pain.⁵ The indication for a repeat anti-incontinence procedure is usually persistent urinary incontinence,^4,6 and most commonly involves a repeat midurethral sling procedure.

Approximately 2.2% to 3.7% will undergo a sling revision.^7–9 This includes approximately 1.2% to 2.3% of patients receiving a MUS who will experience urinary retention of urethral obstruction, requiring surgery.^7,9–11 Approximately 1.0% to 2.5% will experience mesh exposure or erosion requiring surgery,^7,9,11 and 4.1% will experience refractory chronic pain.¹¹ Preexisting voiding dysfunction, age, retropubic sling type, and concurrent prolapse repair are risk factors for sling revision.^7,12 Sling revision usually resolves symptoms associated with mesh exposure, urinary retention, de novo voiding symptoms, and recurrent UTI^12,13 and improves patient satisfaction.^5,14

Approximately 2.4% of women receiving a MUS may end up undergoing a repeat anti-incontinence procedure.¹⁵ Risk factors for a repeat MUS include obesity, mixed urinary incontinence, previous continence surgery, preexisting detrusor overactivity on urodynamics, intrinsic sphincter deficiency, diabetes, lack of urethral hypermobility, and increased pad weight.^6,16,17 Repeat midurethral slings result in subjective improvement of SUI in 54% to 88% of cases.^18–20

Although the MUS surgery has been the subject of many prospective trials, very few were designed to follow complications over a long term follow-up period.¹⁰ Furthermore, studies that have examined return to the operating room for a second surgery have mostly been institution-specific studies,^8,21,22 have examined only repeat MUS anti-incontinence surgery^4,12,15 and only sling revision,⁷ or have included patients who received prolapse mesh without an incontinence procedure.⁹

By understanding the risk factors for a repeat surgery after MUS, we can better counsel our patients who undergo this surgery. Our objective was to use the health facts (HF) database to identify risk factors for having to return to the operating room after MUS for a second surgery related to urinary retention, mesh exposure, voiding symptoms, pain, recurrent UTI, or persistent SUI.

MATERIALS AND METHODS

This is a case-control study using the HF database of women who underwent a second surgery after an initial MUS surgery. The HF database is maintained by Cerner Corporation, an electronic health record vendor and leverages the data generated by health systems to create volumes of information that can be used to improve patient safety and the quality of care. Data in HF are abstracted directly from the electronic medical record from hospitals that have a data use agreement with Cerner.

The current University of Missouri—Kansas City installation of HF has data available through 2016 with almost 64 million unique patients and over 360,429,629 encounters across 863 facilities. The database is longitudinal, allowing for patients to be followed postdischarge if they return to a participating facility.

We identified all women who underwent an isolated MUS surgery between January 1, 2010, and December 31, 2016, using the current procedural technology (CPT) code 57288. Cases return to operating room (+OR) were defined as those who underwent both an initial placement of a MUS AND a repeat surgery any time after the index MUS, within the above date range. Repeat surgeries and their corresponding codes included 57288 (repeat sling operation), 57287 (removal or revision of sling), and 53500 (urethrolysis). Controls no return to OR (NoOR) were defined as those who underwent a MUS but did not undergo 1 of the 3 second urethral surgeries listed previously. Controls were randomly selected on a 4:1 ratio. We included women who either had an isolated MUS or a MUS with concomitant prolapse repair. Our database unfortunately did not include patients who underwent a second surgery at a non-Cerner institution.

All associated diagnoses at the time of the second surgery were then grouped according to the following indication groups: recurrent SUI, pain, mesh complication, urinary retention, recurrent UTI, and bothersome voiding symptoms, such as urinary urgency or frequency. These diagnoses were not mutually exclusive, meaning more than 1 could be present during sling revision. All cases whose diagnoses were not defined or could not be classified into 1 of these groups was classified as “other.”

We compared sociodemographic and other medical diagnoses between cases and controls using Student t test for continuous variables and χ² for categorical variables. We looked for variables for return to the OR that were known risk factors for each of the individual second procedures. For example, we investigated diabetes and immune suppression for mesh exposure and concurrent prolapse surgery for persistent SUI. In some instances, these variables were available and easily identifiable within the database. In other cases, we were required to use a surrogate metric. We additionally performed multivariable logistic regression to identify risk factors for being in the +OR group. Variables with P < 0.10 on the univariable regression were included in multivariable logistic regression analysis. Independent risk factors for second surgery after MUS (+OR) included adjusted odd ratios and 95% confidence intervals. All analyses were performed using SAS software (SAS Institute, Cary, North Carolina.

The University of Missouri Kansas City Institutional Review Board has determined that research using HF data is “Non-human subjects research” (institutional review board protocol ID, 14–567).

RESULTS

We identified 17,953 women who underwent an initial MUS surgery between January 1, 2010, and December 31, 2016. Of these, 1247 or 6.9% returned to the OR for a second surgery (+OR) before December 31, 2016. Table 1 shows the indications and frequencies for the second surgery. In 44.6% of these patients, the indication was either not listed or not defined. The next most common diagnosis was persistent urinary incontinence (32.6%).

TABLE 1.

Number of Cases and Rate of Requiring a Second Surgery, Grouped by Primary Indication

Indication	n	% (Sums to 1247 Undergoing a Second Surgery)	% (Sums to 17,953 Total Midurethral Slings)
Pain*	43	3.4	0.24
Mesh exposure	127	10.1	0.71
Recurrent UTI	8	0.1	0.05
Voiding symptoms†	11	0.1	0.06
Urinary retention	95	7.6	0.53
Persistent SUI	407	32.6	2.27
Unknown	556	44.6	3.10

^*Pain includes pelvic pain, dysuria, and dyspareunia.

^†Voiding symptoms includes urgency, frequency, and nocturia.

Combined with 4989 matched controls, randomly selected within the same date range on a 4:1 ratio (NoOR), the 6236 women in our study had an average age of 55.5 years (interquartile range, 45–66 years), 44.0% were obese (body mass index [BMI], >30), and 2.5% had diabetes with no differences between +OR and NoOR (all P > 0.05). The majority were Caucasian race, but this was higher in +OR than NoOR (90.3% vs 86.4%, P < 0.001). Table 2 shows all sociodemographic and medical variables we considered for our analysis. The 6 variables with P < 0.10 on univariable comparisons were Caucasian race, tobacco and marijuana smoking, lack of concomitant prolapse surgery, use of immunosuppressant drugs, use of blood thinners, and use of anticholinergics. These were included in the final model.

TABLE 2.

Demographic and Bivariate Results

Variable*	Overall N = 6236	OR n = 1247	NoOR n = 4989	P
Age (mean ± SD), y	55.5 (13.2)	55.0 (13.3)	55.7 (13.2)	0.11
Caucasian race	5435 (87.2)	1126 (90.3)	4309 (86.4)	<0.01
Tobacco and/or marijuana use	235 (3.8)	60 (4.8)	175 (3.5)	0.03
History of cancer	4 (0.1)	1 (0.1)	3 (0.1)	0.80
Heart failure or stroke	19 (0.3)	2 (0.2)	15 (0.3)	0.40
Steroid/immune modulating drug use	2576 (41.3)	583 (46.8)	1993 (40.0)	<0.01
Diabetes	156 (2.5)	25 (2.0)	131 (2.6)	0.21
Obesity (BMI >30.0)†	1320 (43.99)	234 (47.3)	1086 (43.3)	0.11
Anticoagulant use	1153 (18.5)	282 (22.6)	871 (17.5)	<0.01
Anticholinergic use	2135 (34.2)	452 (36.3)	1683 (33.7)	0.09
Lack of concomitant Prolapse Surgery	1645 (26.4)	968 (77.6)	3623 (72.6)	<0.01

^*Variables are reported as n (%) unless otherwise noted.

^†Body mass index data available for 3001 subjects.

OR, women who underwent a second surgery after midurethral sling; NoOR, women who only underwent midurethral sling without a second surgery.

In the final multivariable model adjusting for confounders (Table 3), the following variables remained more likely to undergo return to the OR for a second surgery: Caucasian race (OR, 1.47 [1.20–1.81]), use of immunosuppressant drugs (OR, 1.27 [1.12–1.45]), use of blood thinners (OR, 1.38 [1.18–1.62]), and lack of concomitant prolapse surgery (OR, 1.37 [1.18–1.59]). Anticholinergic use and smoking tobacco or marijuana, although significant on univariable analysis, were no longer significant risk factors after adjusting for confounders.

TABLE 3.

Logistic Regression Analyses for Risk Factors of Having a Second Surgery After Midurethral Sling

	Univariable Unadjusted Odds Ratio (95% CI)	Multivariable Adjusted Odds Ratio (95% CI)
Caucasian	1.47 (1.20–1.80)	1.47 (1.20–1.81)
Tobacco and/or marijuana use	1.39 (1.03–1.88)	1.25 (0.92–1.69)
Steroid/immune modulating drug use	1.32 (1.17–1.50)	1.27 (1.12–1.45)
Anticoagulant use	1.38 (1.19–1.61)	1.38 (1.18–1.62)
Anticholinergic use	1.12 (0.98–1.27)	1.00 (0.87–1.15)
Lack of concomitant prolapse surgery	1.32 (1.24–1.52)	1.37 (1.18–1.59)

CI, confidence interval.

DISCUSSION

In this large nationwide population-based study of women who underwent MUS surgery, we found that Caucasian race, use of immunosuppressant drugs, and use of blood thinners were independent risk factors for requiring a second surgery. Concomitant prolapse surgery was protective. Of our patients, 6.9% required a second surgery, whether for urinary retention, mesh exposure, voiding symptoms, pain, recurrent UTI, or persistent SUI. A 6.9% composite rate of a second surgery is consistent with other studies that have found the following individual rates:

• 1% to 2% sling revision for mesh exposure^7,9,11;

• 1% to 2% sling revision for urinary retention^7,9–11;

• 2% to 3% repeat sling for persistent SUI.¹⁵

Our study includes patients with up to 7 years of follow-up; therefore, 6.9% may be a more realistic estimate we can give our patients for their chance of undergoing a second surgery after their first MUS. Wu et al²³ found that 5.2% of the patients in their database had a subsequent surgery for either SUI or prolapse within 5 years of the index MUS.

Our study found that the use of immunosuppressant drugs, including glucocorticoids, was predictive of undergoing a second surgery. After implantation, synthetic mesh invokes a host immune response, involving neutrophils, macrophages, and various cytokines and chemokines.²⁴ It is possible that these medications interfere with this natural immune response and predispose these patients to a second surgery necessary to care for a mesh exposure. Glucocorticoids are known to have a detrimental effect on wound healing.²⁵

Our study also found that the use of blood thinners was predictive of undergoing a second surgery. Anticoagulant medications can result in blood collection at a surgical site, which can promote the growth of bacteria and subsequent abscess formation or even mechanically disrupt the wound.²⁶ It is possible that this would predispose a patient to pain, mesh exposure, or urinary retention, and ultimately a second surgery.

Contrary to other studies that have found concomitant prolapse surgery to be a risk factor for undergoing a sling revision or repeat sling,^7,12,23 our study found this to be protective against these. In perhaps the largest institution-specific case-control study of sling revision, Unger et al⁸ compared 89 patients who underwent sling revision to 178 matched patients who did not have the sling revised. They included patients who underwent a concomitant prolapse repair and found that women who underwent concomitant apical suspension were more likely to need midurethral sling revision. Similar to our study, common indications for sling revision were urinary retention and voiding dysfunction, mesh exposure, and pain.⁸ Our study is unique in that it considered a several indications for a second surgery, whereas the other studies either looked at MUS revision^7,8,12 or repeat MUS.²³

Ours is the largest study we are aware of that looks at risk factors for a composite variable of reoperation for multiple procedures including sling revision, urethrolysis, and repeat MUS. Our decision to examine all patients returning to the OR, regardless of indication or type of repeat surgery was made purposefully. Return to the OR is a patient-centered outcome. We postulate women who undergo MUS consider any reoperation directly related to their MUS as a surgical complication regardless of indication or procedure. Historically, others have looked at these procedures individually. The largest population study looking at MUS revision was Funk et al⁷ who studied more than 188,000 women who underwent an index MUS. Of those women, 3.7% underwent a subsequent MUS revision for either a mesh complication or urinary retention.⁷ The women who underwent sling revision or removal were younger in age and more likely to have undergone a concomitant apical prolapse repair. Suskind et al²⁷ calculated that 6.5 per 100,000 Florida women underwent either a MUS revision or a urethrolysis between 2004 and 2009.

The largest study that looked at reoperation for recurrent SUI was Fialkow et al⁴ who studied 1895 women in the state of Washington who underwent a repeat anti-incontinence procedure out of a population of 41,705 women who underwent an initial anti-incontinence procedure (including predominantly MUS and Burch procedures). Hysterectomy decreased the rate of repeat anti-incontinence surgery.⁴

Strengths and Limitations

The primary strength of our study was the size of the population. The large number and geographical diversity of Cerner hospitals improves the generalizability of our findings. Patient information was collected in a standardized and audited manner. Furthermore, 7 years should capture most second surgeries. Limitations included those inherent in a retrospective study using a hospital-based electronic medical record database, such as susceptibility to information and data collection biases including coding inaccuracies. However, HF has auditing strategies to minimize inaccuracies. After data are loaded, operations staff search for data holes and clean the data by removing duplications, performing date range checks, and standardizing terms by automapping to common nomenclature. Experienced clinical staff review and map unmapped terms. We also recognize that a case-control design can only measure associations and causality cannot be inferred.

One potential area for bias is misclassification of controls. We identified cases as those who had a second operation during the study timeframe. However, even if a control did not have a second operation within the time range, it is possible that they would subsequently undergo one. We considered choosing controls from earlier in the database to allow a longer window for reoperation to have occurred. However, we felt that this risked introducing other biases such as temporal changes in sling products, technique, or patient characteristics that could impact our primary outcome. We feel that this impact is likely minimal because the majority of cases would have still been identified in the study window. This is supported by Linder et al²⁸ who showed that the median time from sling placement to surgery for mesh exposure was 7.6 months. Likewise, other surgeries for recurrent stress incontinence, such as a Burch urethropexy, may not have been identified, further increasing possibility of misclassifying controls. Finally, participants who had surgery outside of the Cerner network would still be eligible for selection as a control because these would not be available within our database. Ultimately, any misclassification is more likely to dilute the significance of potential risk factors, strengthening the validity of the risk factors that we identified. However, it is possible that there are other significant risk factors that were not identified because of this limitation. Our stated reoperation rate of 6.9% may also be an underestimation for these reasons. In defining the cases, we were only able to assess additional procedures that were performed. Within this database, it is impossible to determine if there are subjects who have persistent stress incontinence or other voiding dysfunction who do not ultimately undergo surgery for their complaints.

We were unable to calculate the time to return to surgery for our population because the dates of the surgery were shifted as part of deidentification of the data. This shifting within HF is random and both forward and backwards but still allowed us to find a cohort within our selected date range. Furthermore, we do not have data on patient outcomes after the second surgery. Finally, the HF database was missing indication data on 45% of the patients, limiting any indication-specific conclusions.

In conclusion, the rate of a second surgery after MUS over a 7-year period is 6.9% including sling revision, urethrolysis, or repeat sling procedure. This rate includes second surgeries for recurrent SUI, pain, mesh complication, urinary retention, recurrent UTI, and voiding symptoms. We found that white race, using immunosuppressant drugs, using blood thinners, and not having concomitant prolapse surgery are all risk factors for having a second surgery after MUS.