Mortality and Reoperation Following Midurethral Sling versus Urethral Bulking In Older Women

Plain language summary:

Medicare beneficiaries undergoing bulking for treatment of stress incontinence are sicker, have increased risk of retreatment, and shorter lifespans than those undergoing sling.

Introduction:

Over half of US women aged 40 and over report stress urinary incontinence, a figure that remained stable between 2005 and 2018.¹ Among these millions of women, 13.6% are expected to undergo surgery.² While a number of seminal studies on SUI surgery evaluated participants with relatively low median age^{3 4}, epidemiologic evidence demonstrates a bimodal distribution in prevalence of surgery with peaks at age 46 and age 70–71.² Thus, it is up to the clinician to weigh the surgical risks of complications and reoperation and retreatment for women in older age groups.^{5 6} This older cohort inherently has shorter life expectancy and higher rates of comorbidities.

When comparing midurethral sling with urethral bulking, the evidence examining urethral bulking is far less robust. The gold standard midurethral sling is among the most thoroughly researched of surgical procedures, with well-known risks and an estimated five-year retreatment rate of 3.8%.^{7 8} In contrast, urethral bulking agents may have greater appeal in an older population due to a lower risk of voiding dysfunction and the greater tolerability without anesthesia. However, data are limited regarding the long-term risk of retreatment after bulking agents. In 2017, a Cochrane review commented that bulking’s “durability after one year remains unknown.”⁹ Studies with five-year outcomes often have exceedingly low numbers, with the exception of Brosche et al’s retrospective analysis of 388 women undergoing bulking with polyacrylamide hydrogel. In this study, 32% required repeat bulking and 19% underwent other surgical retreatment within seven years.^10–12

While all surgeons must consider a patient’s suitability for a procedure, it is unclear if and how an older population electing midurethral sling differs from those opting for urethral bulking. We aimed to use a nationally representative dataset of women over 65 to examine differences in comorbidities, retreatment rates, and eventual mortality between these two groups.

Materials and Methods:

We conducted a retrospective cohort study using the 5% limited data set from the Center for Medicare and Medicaid Services (CMS) between 2010 and 2018. This representative nationwide sample of beneficiaries includes demographics, eligibility and claims data of over 1.8 million older adults annually. Unique individuals may be followed longitudinally using deidentified, encrypted numbers. The Institutional Review Boards at both Providence Health and Duke University deemed this study exempt from further review as it included only deidentified data.

Inclusion criteria were: female sex, age 65 or older, having undergone an SUI procedure with either a sling or bulking in the years 2011–2014, with at least one year of continuous prior enrollment without prior SUI procedure. Those undergoing concomitant procedures other than cystoscopy were excluded. This choice was made to ensure a more homogenous sample population, excluding beneficiaries whose SUI procedures were performed for occult SUI, and also to avoid reoperations for postoperative complications which may have been related to a procedure other than the SUI procedure. Procedures were identified in the carrier files using Common Procedural Technology (CPT) codes 57288 (sling operation for stress incontinence, fascia or synthetic) and 51715 (endoscopic injection of implant material into the submucosal tissues of the urethra and/or bladder). Beneficiaries were followed until reoperation for a complication, retreatment for SUI, loss of fee-for-service Medicare, death, or until the end of 2018.

Reoperations for complications included sling removal, urethrolysis, urethro-vaginal or vesico-vaginal fistula closure, incision and drainage of hematoma or abscess, debridement, cystotomy repair, urethral repair, placement of percutaneous nephrostomy tubes, or ureteral reimplantation. Procedural retreatment for SUI recurrence included urethral bulking, sling, or laparoscopic urethropexy (e.g. Burch colpopsuspension, CPT code 51990) or open urethropexy (CPT code 51840, 51841).

Baseline characteristics of the cohort, including age, race, census-region derived from state of residence, procedure setting, length of stay, and comorbidities were collected from the Master Beneficiary Summary file and inpatient revenue and outpatient revenue tables. Procedures performed in an office setting, ambulatory surgery center setting, as well as inpatient or outpatient hospital setting were included. The following comorbidities were evaluated: cardiovascular disease, cardiac arrhythmia, atrial fibrillation, coronary heart disease, deep vein thrombosis, hypertension, chronic pulmonary disease, diabetes, renal disease, cancer, dementia, Alzheimer’s disease, chronic pain, fall-related injury, and urinary tract infection. Comorbid conditions were determined by searching inpatient, outpatient and carrier claims for the one year prior to the index SUI surgical date, using validated algorithms.^{13 14} In addition, we calculated the Charlson Comorbidity Index (CCI), a weighted score of 12 morbidities with a range of 0–24, in which higher scores are associated with postoperative adverse events and mortality within one year. ^{15 16}

Cohorts were stratified by index procedure type, with proportions for categorical variables and means with standard deviations for continuous variables. Group differences were evaluated using chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables. Cumulative incidence for mortality at 30 days and one to five years from the index procedure date were calculated using Kaplan-Meier estimates, and group differences were calculated using the log-rank test. For all other outcomes, the cumulative incidence function was used to account for the competing risk of death, and group differences were evaluated using Gray’s test. The associations between index SUI treatment and all reoperations, retreatment with sling, and retreatment with bulking were evaluated in separate Cox proportional hazards models. In adjusted versions, we controlled for age, race, place of service and comorbidities as possible confounding covariates.

Results:

From 2011 through 2014, we identified 8,347 women within the CMS 5% dataset who underwent a procedure for the treatment of SUI. Of these, 6,418 met the inclusion criteria for age and were 65 years or older. Among them, 549 were excluded due to a lack of Medicare benefits for at least 12 months prior to the index procedure, and an additional 179 were excluded because of SUI surgery in 2010. An additional 89 underwent laparoscopic or open urethropexy and were excluded. Of the remaining 5,601 patients, 1,700 underwent sling and 875 underwent bulking without any concomitant procedures (except cystoscopy), and were included in the analysis. While the most common place for both index procedures was an outpatient hospital (73.8% of slings, 53.0% of bulkings), 25.6% of bulking patients underwent the procedure in the office setting.

The median follow-up time was 5.7 years for patients electing sling and 5.2 years for the patients electing bulking (IQR 4.3–7.0 vs 3.8–6.4, p<0.001). Those undergoing bulking injection were older than those undergoing sling (median 76.7 vs 72.7, p< 0.001). Bulking patients were significantly more likely to have all of the medical comorbidities with the exception of diabetes and chronic pulmonary disease. Patients undergoing bulking scored a mean of 2.4 on the Charlson comorbidity index, compared with 1.9 among sling patients (p<0.001, range 0–15, IQR 3, Table 1).

Table 1.

Baseline characteristics.

Variable	Sling (n=1,700)	Bulking (n=875)	p-value
Age (years), Mean (SD)	72.7 (5.8)	76.7 (7.1)	< 0.001
Race			0.04
Asian	*	*
Black	36 (2.1%)	20 (2.3%)
Hispanic	23 (1.4%)	*
North American Native	*	*
White	1,484 (87.3%)	795 (90.9%)
Unknown/Missing/Other	140 (8.2%)	52 (5.9%)
Region			0.01
Other	123 (7.2%)	48 (5.5%)
Northeast	204 (12.0%)	123 (14.1%)
South	770 (45.3%)	371 (42.4%)
Midwest	381 (22.4%)	183 (20.9%)
West	222 (13.1%)	150 (17.1%)
Place of Service			< 0.001
Office	20 (1.2%)	224 (25.6%)
Inpatient Hospital	235 (13.8%)	57 (6.5%)
Outpatient Hospital	1,254 (73.8%)	464 (53.0%)
Ambulatory Surgical Center	191 (11.2%)	130 (14.9%)
Comorbidities
Cerebrovascular Disease	251 (14.8%)	171 (19.5%)	0.002
Cardiac Arrhythmias	395 (23.2%)	255 (29.1%)	0.001
Atrial Fibrillation	146 (8.6%)	128 (14.6%)	< 0.001
Coronary Heart Disease	440 (25.9%)	273 (31.2%)	0.004
Deep Vein Thrombosis	24 (1.4%)	25 (2.9%)	0.01
Hypertension	1,340 (78.8%)	730 (83.4%)	0.005
Chronic Pulmonary Disease	506 (29.8%)	267 (30.5%)	0.69
Diabetes	501 (29.5%)	267 (30.5%)	0.58
Diabetes with complications	169 (9.9%)	90 (10.3%)	0.78
Renal Disease	148 (8.7%)	144 (16.5%)	< 0.001
Cancer (metastatic or non-metastatic)	197 (11.6%)	124 (14.2%)	0.06
Dementia	33 (1.9%)	37 (4.2%)	< 0.001
Alzheimer’s Disease	15 (0.9%)	22 (2.5%)	< 0.001
Chronic Pain	1,153 (67.8%)	652 (74.5%)	< 0.001
Fall-related injury	410 (24.1%)	259 (29.6%)	0.003
Urinary Tract Infection	766 (45.1%)	525 (60.0%)	< 0.001
Charlson Score, Mean (SD)	1.9 (2.1)	2.4 (2.4)	< 0.001

^*N<11 suppressed in accordance with the CMS small cell size policy

Mortality was significantly different between groups (Figure 1). At two years postoperatively, 3.1% of sling patients and 6.9% of bulking patients had died. By five years, mortality had diverged further to 23.2% of bulking patients and 10.2% of sling patients. When controlling for age, race, and comorbidities, patients who underwent bulking still had a 1.73 greater likelihood of death during the study period compared with patients who underwent sling (95%CI 1.36–2.22, p<0.001).

Figure 1:

Kaplan-Meier curve for mortality by index SUI treatment

Patients undergoing bulking injection are more likely than those undergoing sling to have died within two years of index treatment, a trend which continues to the five year mark.

The most common indication for repeat procedure in both groups was retreatment for SUI, with 3.9% of sling patients and 20.1% of bulking patients being retreated within the first year. By 5 years, 6.7% of sling patients had been retreated, with equivalent numbers choosing repeat sling and bulking injection as second-line treatment. In comparison, 24.6% of bulking patients were retreated within 5 years, overwhelmingly favoring repeat bulking (Figure 2). Reoperation for complications were uncommon in both groups. Among sling patients, the rate of reoperation for complications was 3.4% within 5 years, compared with 1.5% among bulking patients. Median time to any repeat procedure was 5 years in the sling group, as compared to 3.3 years for the bulking group (p<0.001).

Figure 2.

A-B. Retreatment for SUI by index surgery

Retreatment rates for women undergoing midurethral sling as index procedure plateau two years after surgery. Women undergoing bulking are retreated with sling at similar rates, but nearly 30% undergo repeat bulking within five years.

When evaluating variables associated with reoperation or retreatment, urethral bulking was associated with 3.9 times the risk of a second procedure compared to a sling (95% CI 3.12, 4.89). Hypertension was also associated with reoperation or retreatment (HR 1.36, 95%CI: 1.04, 1.78) (Table 2). Racial and ethnic minority women were 37% less likely to undergo reoperation, including retreatment (HR 0.63, 95% CI 0.41, 0.97).

Table 2.

Unadjusted and adjusted hazard ratios for the associations between patient characteristics and retreatment or reoperation

Parameter	Unadjusted,		Adjusted
	HR (95% CI)	p	HR (95% CI)	p
Index Surgery (Ref=Sling)	4.19 (3.46, 5.06)	< 0.001	3.91 (3.12, 4.89)	< 0.001
Demographics
Age	1.03 (1.01, 1.04)	< 0.001	0.99 (0.97, 1.00)	0.11
Racial / ethnic minority groups (Ref=Non-Hispanic white)	0.61 (0.43, 0.86)	0.005	0.63 (0.41, 0.97)	0.04
Comorbidities (Ref=None)
Cerebrovascular Disease	1.25 (0.99, 1.57)	0.06	1.07 (0.83, 1.37)	0.62
Cardiac Arrhythmias	1.07 (0.87, 1.32)	0.53	0.97 (0.78, 1.21)	0.81
Coronary Heart Disease	1.09 (0.89, 1.33)	0.40	0.95 (0.76, 1.19)	0.66
Deep Vein Thrombosis	0.97 (0.48, 1.96)	0.94	0.71 (0.34, 1.50)	0.37
Hypertension	1.43 (1.11, 1.83)	0.005	1.36 (1.04, 1.78)	0.02
Chronic Pulmonary Disease	0.96 (0.78, 1.17)	0.67	0.92 (0.75, 1.14)	0.46
Diabetes	1.02 (0.83, 1.24)	0.87	0.97 (0.79, 1.20)	0.78
Renal Disease	1.32 (1.02, 1.72)	0.04	1.00 (0.75, 1.34)	0.98
Cancer	0.92 (0.70, 1.23)	0.58	0.87 (0.65, 1.15)	0.33
Dementia	1.57 (0.95, 2.58)	0.08	1.35 (0.81, 2.27)	0.25
Chronic Pain	1.24 (1.01, 1.52)	0.04	1.08 (0.87, 1.34)	0.50
Fall-related injury	1.19 (0.98, 1.46)	0.08	1.04 (0.84, 1.28)	0.73
Urinary Tract Infection	1.27 (1.06, 1.52)	0.01	1.01 (0.83, 1.21)	0.95

Secondary analyses examined risks of sling or bulking as procedure for persistent or recurrent SUI. Women who chose sling or bulking as index surgery showed no difference in likelihood of undergoing sling as a second procedure (p=0.21). This remained true when controlling for age, race, and comorbidities. Among comorbidities evaluated, only cancer significantly impacted the likelihood of reoperation with sling, lowering it by 68% (p=0.04). Use of bulking as a retreatment procedure was more likely among those with dementia (HR 1.89, p=0.03) and less likely among racial and ethnic minority women (HR 0.47, p=0.048).

Discussion:

Our study focusing on Medicare beneficiaries with SUI clearly shows that patients who are sicker and have fewer years of life remaining preferentially undergo bulking rather than midurethral sling. Our data suggest that over forty thousand beneficiaries age 65 and over undergo SUI procedures each year in the United States. This population is rarely treated with open or laparoscopic Burch urethropexy, leaving sling and bulking the primary options for patients and their providers. While it is appropriate that patients and their surgeons are considering treatment options with the patient’s comorbidities in mind, our findings cannot assess whether surgeons are weighing a patient’s comorbidities appropriately, excessively, or insufficiently with respect to her surgical risk.

While our study controlled for many comorbidities commonly found among Medicare beneficiaries, a 1.73 greater likelihood of death was seen among women undergoing bulking injection as compared to sling after controlling for age, race, and comorbidities. This increased likelihood may be due to other comorbidities not included in our model, or a more general perception of patient frailty on the part of the surgeon weighing treatment options.¹⁷

We found the cumulative incidence of repeat procedure—including retreatment for SUI and surgical treatment of complications—to be 10.3% among Medicare beneficiaries undergoing sling and 36.1% among those choosing urethral bulking with median 5-year follow-up. Our estimate for the risk of reoperation following sling is consistent with studies on women with younger median age, as Jonnson Funk et al found a 9.0% risk of surgical retreatment and a 3.4% risk of reoperation for complication among sling patients within five years.^{18 19} In contrast, they found a 54.6% risk of repeat procedure within five years of bulking procedure in their younger cohort. Our lower risk of repeat procedure after bulking in a Medicare population could, in theory, be explained by greater efficacy of bulking in older women. We believe it is unlikely that older women are better responders to bulking procedures, particularly in light of evidence demonstrating that they have higher risk of failure with sling procedures.⁵ Rather, we interpret our lower risk of repeat procedure as a lack of desire for additional invasive procedures, as well as a likely increase in willingness to accept ongoing incontinence in an older and sicker population, many of whom consider incontinence a normal part of aging.²⁰

While comorbidities likely led to patients preferentially undergoing bulking for treatment of their SUI, preoperative comorbidities were not otherwise associated with a higher or lower risk of retreatment with either approach. The finding that dementia patients undergoing retreatment were more likely to be retreated with bulking than sling may be due to concerns that anesthesia could exacerbate dementia, or that such patients would have greater difficulty learning intermittent self-catheterization. Of note, some comorbidities, including chronic obstructive pulmonary disease and diabetes, were not associated with a significantly higher chance of undergoing bulking. It is unclear which comorbidities predispose to greater surgical risk in the context of SUI procedures. Future research using outcomes data could allow for development of a risk calculator such as the one devised for colorectal surgery, allowing surgeons to assess surgical risk quantitatively on an individual basis.²¹

The lower risk of reoperation or retreatment among racial and ethnic minority patients was driven by their decreased likelihood of undergoing repeat bulking procedure. While prior studies have demonstrated lower initial treatment rates for SUI among African American and Asian women as compared with non-Hispanic White women, our findings indicate lower retreatment rates as well.^{22 23} Given equivalent Medicare eligibility among all in the cohort and the lack of a biologically plausible reason for racial and ethnic minority women to have a greater success with index treatment, this discrepancy is explained by some combination of greater reluctance to be retreated, decreased offers of retreatment, increased financial burden, or other social determinants of health. Further investigation with a cohort that includes a large number of women in each racial or ethnic subgroup is warranted to explain these findings.

Strengths of this analysis include its large size and nationally representative cross-section of US older adults. The large population allowed us to study a substantial cohort, even after excluding those whose concomitant surgeries were likely to be confounding factors. Furthermore, the eligibility of nearly all people 65 and over for Medicare makes our conclusions generalizable to this age group, without concern for insurance status as a confounder. Finally, the 5-year follow-up and inclusion of mortality data provides insights not previously available for studies of SUI surgery.

Our study has several limitations. Inherent within the analysis of claims data is the inability to account for variables not included in the data, including severity of SUI, body mass index, smoking status, and history of treatment for SUI prior to 2010 or prior to Medicare enrollment. The dataset also does not include information regarding the type of bulking agent used, although the years examined are prior to the Food and Drug Administration’s approval of polyacrylamide hydrogel (PAHG) and so may not be generalizable to patients undergoing bulking injection with this agent. Freitas et al found a 24.6% chance of retreatment after bulking with PAHG, and if similar findings and efficacy were observed in a Medicare population, this would result in a significant improvement for patients.²⁴ In addition, we were unable to control for the use or disuse of anesthesia in the bulking group. While procedure location is presumed to be strongly correlated with anesthesia use, there may have been women receiving bulking injection in a hospital setting without anesthesia and vice versa. Given the study design as a time-to-event analysis, we were unable to capture all reoperations or retreatments in patients who had more than one. This is particularly notable in the case of women undergoing bulking injection, in which repeat bulking may be anticipated. Finally, we are aware that CPT code 57288 represents a heterogenous group of women undergoing both mesh and fascial slings. However, Suskind et al found that among Medicare beneficiaries undergoing sling at an outpatient hospital or ambulatory surgery center, nonmesh sling represented only 6.2% of sling procedures.²⁵

Conclusions:

Patients and their surgeons are considering preoperative comorbidities when deciding between SUI procedures, leading to use of sling disproportionately in healthier women with more remaining years of life. Future evidence should give providers the ability to consider based on data to what extent comorbidities make a patient a poor candidate for an SUI procedure, rather than relying on global impressions of surgical risk.