Transition to surgery after pessary among female Medicare beneficiaries with pelvic organ prolapse

Abstract

Background:

The American College of Obstetricians and Gynecologists recommends offering a vaginal pessary to women seeking treatment of pelvic organ prolapse. However, single-institution series have suggested that a sizable proportion of women fit with a pessary will transition to surgery within the first year.

Objective:

To estimate the proportion of female U.S. Medicare beneficiaries with pelvic organ prolapse who undergo surgery after pessary fitting, to describe the median time to surgery from pessary fitting, and to identify factors associated with the transition from pessary to surgery.

Study Design:

The Medicare 5% Limited Data Set was queried from 2011 to 2016 for women ≥65 years with a diagnosis of prolapse who underwent pessary fitting. Patients with at least 3 years of follow up in the Medicare Data Set were followed longitudinally for the primary outcome of surgery for prolapse. The cumulative incidence of prolapse surgery following index pessary fitting was calculated. The characteristics of women who underwent surgery were compared with those who did not using time-varying Cox regression analysis.

Results:

Among 2,032 women fit with a pessary, 608 underwent surgery within 7 years. The median time to surgery was 496 days (IQR 187–1089 days). The cumulative incidence of prolapse surgery was 12.2% at 1 year and 30.9% at 7 years. After adjusting for covariates, factors significantly associated with the transition to surgery included prior prolapse surgery (aHR 1.50, 1.09–2.07) and a diagnosis of urinary incontinence at the time of pessary fitting (aHR 1.20, 0.62–0.99). Factors associated with a lower hazard of surgery included age (aHR 0.96 per year, 95% CI 0.95–0.97), dual Medicare/Medicaid eligibility (aHR 0.75, 0.56–1.00) and pessary fitting by a non-gynecologist (aHR 0.78, 0.62–0.99).

Conclusion:

In this population of Medicare beneficiaries, within 7 years of pessary fitting, almost one-third of women over 65 underwent surgery for prolapse. These results add to our current understanding of the demographics of pessary use in an older population and may aid in counseling older patients presenting for treatment of symptomatic pelvic organ prolapse.

Condensation

Among 2,032 female Medicare beneficiaries fit with a pessary, 30.9% underwent prolapse surgery within 7 years (median time of 496 days after initial fitting).

Introduction

Pelvic organ prolapse (POP) is common, and the prevalence increases with age.¹ Treatment options for pelvic organ prolapse include expectant management, conservative management with a pessary, or surgical management. The American College of Obstetricians and Gynecologists recommends offering a vaginal pessary to women seeking treatment of pelvic organ prolapse.² However, prior research suggests that pessary fitting is successful for only 73–88% of patients with prolapse.^3,4,5 Additionally, short-term studies suggest discontinuation of pessary use in up to 50% of women by 1 year.^3,5–8 Longitudinal case series^7,9 suggest that a sizable proportion of women fit with a pessary will transition to surgery within the first year. A prior study utilizing Medicare data through 2009 demonstrated that 12% of women with prolapse went onto surgery within one year of being fit with a pessary and 24% by nine years.¹⁰

These published data may not represent the contemporary pessary experience of U.S. women. Specifically, it is unknown at what rate older women seeking pessary care will ultimately transition to surgery for prolapse. We sought to describe the transition from pessary to surgery for a representative sample of older adult women in the U.S., using data from Medicare claims. The objectives of this study are to: estimate the proportion of Medicare beneficiaries with POP who go onto surgery after pessary fitting, describe the median time to surgery from pessary fitting, and describe factors associated with the transition from pessary to surgery.

Materials and Methods

The Medicare 5% Limited Data Set (LDS) was queried from January 1, 2011 to December 31, 2016. This dataset is derived from a random sample of 5% of the Medicare Fee-For-Service population obtained from the Centers for Medicare and Medicaid Services (CMS). This dataset provides data on claims from a variety of settings including inpatient, outpatient, and skilled nursing facilities for more than 98% of Americans aged 65 and over, and withdrawal from the program is rare.^11,12 Criteria for inclusion were women age 65 and older with a diagnosis of prolapse whose first pessary fitting occurred between 2011 and 2016. Prolapse diagnosis was based on International Classification of Diseases (ICD) code (Appendix 1), and pessary fitting was defined by Current Procedural Terminology (CPT) code 57160 or Durable Medical Equipment (DME) code A4561 or A4562. Patients were considered to have undergone pessary fitting if they had CPT code 57160 and/or one of the DME codes. The index date was defined as the date of the pessary fitting code or DME charge. If patients had more than one instance of CPT code 57160 or DME code A4561/A4562, the first instance was used as the index date and subsequent instances categorized as a “re-fitting”. If patients had both CPT 57160 and a relevant DME code, but these codes did not occur on the same date, the date of the DME charge was used as the index date. Patients were excluded if they had a prior diagnosis of genitourinary fistula; genitourinary cancer, including endometrial hyperplasia; or rectal prolapse present at the time of the index date or within one year prior (Appendix 1). These criteria were selected for exclusion as these diagnoses might influence the recommendation for management of pelvic organ prolapse. Patients were also excluded if they did not have at least three years of follow up data in the Medicare dataset after pessary fitting.

Patients were followed longitudinally for the primary outcome of surgery for prolapse, identified using CPT codes (Appendix 1). For women with more than one surgery after pessary fitting, the first occurrence of surgery was used to define the primary outcome. Prolapse surgeries were classified as either reconstructive, obliterative, or hysterectomy-only. Patients with surgical codes indicating both reconstructive and obliterative prolapse surgery on the same date were categorized as having undergone obliterative surgery. Patients undergoing concomitant hysterectomy with codes for prolapse repair were categorized based on their prolapse repair (reconstructive vs obliterative). Due to the nature of the Medicare dataset, a small number of patients had multiple relevant surgical codes coded on different dates but within two calendar days of each other. These surgeries were considered as having occurred simultaneously, and similar rules were applied. However, patients with multiple surgical codes on dates greater than two days apart were considered as having distinct surgeries.

Secondary outcomes included number of pessary re-fittings, new-onset urinary incontinence or rectal prolapse after pessary fitting, and surgeon subspecialty for those who underwent prolapse surgery. Additional covariates were also captured, including age, race, dual Medicare/Medicaid eligibility status, smoking status, obesity, Charlson Comorbidity Index (CCI), geographical region, prolapse type, history of prolapse surgery (occurring within one year look-back), year of index pessary fitting, physician specialty for first pessary fitting, and prior diagnosis of urinary incontinence. Race/ethnicity categories were specified as White, Black, and Other (including Unknown, Other, Asian, Hispanic, and North American Native), due to underrepresentation of nonblack minority groups in the Medicare population, and poor accuracy of race/ethnicity codes for Asian and Hispanic beneficiaries in the Master Beneficiary Summary File from the Centers for Medicare and Medicaid Services.¹³ Dual-eligible patients qualify for Medicaid based on a lower income.¹⁴

Demographic variables were summarized. Cumulative incidence function was used to calculate the cumulative incidence of prolapse surgery over the study time period following index pessary fitting. Crude associations were examined between prolapse surgery during follow-up and secondary outcomes (new onset rectal prolapse and new onset urinary incontinence). Time-varying Cox regression analysis was performed comparing patients that underwent surgery to those that did not to determine factors associated with transition to surgery. We included the following factors in the model: age, smoking status, obesity, race, dual eligibility (Medicare/Medicaid), geographic region, physician who performed first pessary fitting, type of prolapse, prior prolapse surgery, Charlson comorbidity index, and prior urinary incontinence. These covariates were selected a priori as they would be known at the time of pessary fitting and may influence provider recommendation or patient preference for surgical or conservative management. SAS version 9.4 (Cary, NC) was used for analysis. This study was deemed exempt from IRB review.

Results

There were 2,032 patients included in the study cohort who met eligibility criteria and had at least 3 years of follow up in the Medicare dataset (Figure 1). Median duration of follow up for all patients included in the dataset was over 5 years (1,960 days, IQR 1,511.5–2,471 days). Most patients were classified as white (89.7%) with a median age of 75 (IQR 70–81), with about one-third of the sample residing in the South and another third in the Midwest (Table 1). 12.8% of the sample was dually eligible for Medicare and Medicaid. Uterovaginal prolapse was the most common prolapse diagnosis (48.6%), and 4.8% of the sample had prior prolapse surgery within one year of the index date. Most pessary fittings were performed by Obstetrician-Gynecologists (Ob/Gyns). Nearly half of the patients (42.9%) had a diagnosis of urinary incontinence at time of index pessary fitting or in the year prior.

Figure 1.

Flow diagram describing patient inclusion and final sample size attainment. Follow-up is the difference between the last date (e.g. fee-for-service end date, death date, study end date) and the index date (e.g. date of first pessary fitting). LDS, Limited Data Set; FFS, fee for service.

Table 1.

Baseline characteristics of Pessary fitting cohort from 2011–2016

	No surgery	Surgery	p-value
N	1,424	608
Age, years
Median [25th, 75th]	76.0 (71.0, 82.0)	73.0 (69.5, 78.0)	<.0001
Age group, n (%)			<.0001
65–74	617 (43.3)	341 (56.1)
75–84	586 (41.2)	228 (37.5)
85+	221 (15.5)	39 (6.4)
Race, n (%)			0.9428
White	1,275 (89.5)	547 (90.0)
Black	81 (5.7)	34 (5.6)
Other/unknown	68 (4.8)	27 (4.4)
Dual eligibility, n (%)	200 (14.0)	60 (9.9)
Smoking status, n (%)			0.5985
Never	1,309 (91.9)	556 (91.4)
Former	78 (5.5)	39 (6.4)
Current	37 (2.6)	13 (2.1)
Obesity, n (%)	51 (3.6)	19 (3.1)	0.6054
Prolapse diagnosis type, n (%)			0.1566
Utero-vaginal	687 (48.2)	300 (49.3)
Vaginal	470 (33.0)	179 (29.4)
Post-hysterectomy	167 (11.7)	90 (14.8)
Not specified	100 (7.0)	39 (6.4)
Prior prolapse surgery, n (%)	53 (3.7)	45 (7.4)	0.0004
US Census Region, n (%)			0.3453
South	427 (30.0)	196 (32.2)
Midwest	409 (28.7)	187 (30.8)
West	259 (18.2)	97 (16.0)
Northeast	329 (23.1)	128 (21.1)
Physician specialty for first pessary fitting, n (%)			0.0441
Urology	134 (9.4)	51 (8.4)
Obstetrics/Gynecology	1,030 (72.3)	471 (77.5)
Others	260 (18.3)	86 (14.1)
Charlson comorbidity index
Median [25th, 75th]	1.0 (0.0, 3.0)	1.0 (0.0, 2.0)	0.0083
Prior urinary incontinence, n (%)	584 (41.0)	287 (47.2)	0.0098

^*The Other/Unknown race/ethnicity category includes Asian, Hispanic, North American Native, and unreported

The median number of pessary fittings was 1 (IQR 1–3), although 49.2% underwent at least one re-fitting (Table 2). After pessary fitting, 608 women underwent surgery for prolapse during the follow up period. Of those, 410 (67.4%) underwent reconstructive surgeries, 179 (29.4%) underwent obliterative surgeries, and 19 (3.1%) underwent hysterectomy without concomitant prolapse repair. Most surgeries were performed by Ob/Gyns (73.7%). Secondary outcomes, including new-onset urinary incontinence and urinary incontinence, were diagnosed in 523 (25.7%) and 91 (4.5%) of the cohort, respectively, after pessary fitting and were associated with transition to surgery (urinary incontinence: OR 2.90, 95% CI 2.34–3.60; rectal prolapse: OR 2.39, 95% CI 1.57–3.65). 351 patients (17.3%) died during the study period (2011–2019).

Table 2.

Frequencies for secondary outcomes over study period (2011–2019)

N	2,032
Pessary re-fitting, n (%)	1,000 (49.2)
Number of pessary fitting(s)
Median [25th, 75th]	1.0 (1.0, 3.0)
Surgeon specialty, n (%)
Obstetrics/Gynecology	489 (24.1)
Urology	75 (3.7)
Others	44 (2.2)
None	1,424 (70.1)
Route of surgery, n (%)
Obliterative surgery	179 (8.8)
Reconstructive surgery with/without apical support procedure	410 (20.2)
Hysterectomy*	19 (0.9)
No surgery	1,424 (70.1)
Death, n (%)	351 (17.3)
New onset rectal prolapse, n (%)	91 (4.5)
New onset urinary incontinence, n (%)	523 (25.7)

^*Hysterectomy only without concomitant prolapse repair.

The median time from initial pessary fitting to surgery was 496 days (IQR 187–1089 days). The cumulative incidence of prolapse surgery was 12.2% at 1 year and 30.9% at 7 years (Figure 2). Most surgeries occurred within the first 3 years of pessary fitting, but the incidence of prolapse surgery continued to increase with time. This pattern was similar for all routes of surgery.

Figure 2.

Cumulative incidence of surgery for prolapse after initial pessary fitting.

In a multivariable model, characteristics associated with surgery included age, dual Medicare/Medicaid eligibility, prior prolapse surgery, physician specialty performing the first pessary fitting, and a history of urinary incontinence (Table 3). Older patients were less likely to transition to surgery (aHR 0.96, 95% CI 0.95–0.97), and the hazard of undergoing prolapse surgery was 25% lower for patients with dual Medicare/Medicaid eligibility (aHR 0.75, 95% CI 0.56–1.00). Patients who underwent pessary fitting by a provider other than an Ob/Gyn or Urologist had a 21% lower hazard of transitioning to surgery (aHR 0.78, 95% CI 0.62–0.99); Table 3). Patients who previously had prolapse surgery had a 50% higher hazard of undergoing repeat prolapse surgery after pessary fitting (aHR 1.50, 95% CI 1.09–2.07), and patients with a prior diagnosis of urinary incontinence had a 20% higher hazard of transitioning to surgery after pessary fitting (aHR 1.20 95% CI 1.02–1.42). There was no difference in hazard of surgery after pessary fitting based on number of pessary re-fittings or CCI.

Table 3.

Crude and adjusted hazard ratio estimates from prolapse surgery-specific time-varying Cox regression model over 7-year follow-up after index pessary fitting

Variable	Unadjusted Hazard Ratio (95% CI)	p-value	Adjusted Hazard Ratio (95% CI)	p-value
Age, years	0.96 (0.95,0.97)	< .001	0.96 (0.95,0.97)	< .001
Race
White	1.00 [Reference]		1.00 [Reference]
Black	0.99 (0.70,1.41)	.98	1.04 (0.72,1.49)	.83
Other/unknown	0.91 (0.61,1.35)	.64	0.99 (0.66,1.49)	.97
Dual eligibility	0.72 (0.55,0.94)	.02	0.75 (0.56,1.00)	.049
Smoking status
Never	1.00 [Reference]		1.00 [Reference]
Former	1.20 (0.87,1.66)	.27	1.18 (0.85,1.65)	.33
Current	0.90 (0.52,1.56)	.70	0.79 (0.45,1.38)	.41
Obesity	0.96 (0.61,1.52)	.87	0.91 (0.57,1.45)	.69
Prolapse diagnosis type
Uterovaginal	1.00 [Reference]		1.00 [Reference]
Not specified	0.88 (0.63,1.24)	.48	0.93 (0.66,1.31)	.68
Post-hysterectomy	1.19 (0.94,1.51)	.16	1.12 (0.87,1.42)	.38
Vaginal	0.87 (0.72,1.05)	.14	0.86 (0.71,1.04)	.12
Prior prolapse surgery	1.84 (1.36,2.50)	< .001	1.50 (1.09,2.07)	.01
US Census Region
South	1.00 [Reference]		1.00 [Reference]
Midwest	1.02 (0.83,1.25)	.87	1.08 (0.88,1.32)	.49
Northeast	0.86 (0.69,1.08)	.19	0.93 (0.74,1.17)	.53
West	0.83 (0.65,1.06)	.14	0.87 (0.67,1.12)	.28
Year of index pessary fitting
2011	1.00 [Reference]		1.00 [Reference]
2012	1.07 (0.81,1.43)	.63	1.05 (0.79,1.40)	.74
2013	1.17 (0.89,1.54)	.26	1.06 (0.81,1.40)	.67
2014	1.21 (0.92,1.60)	.17	1.08 (0.82,1.43)	.60
2015	0.99 (0.73,1.32)	.92	0.85 (0.63,1.14)	.27
2016	1.05 (0.77,1.42)	.78	0.92 (0.67,1.26)	.60
Physician specialty for first pessary fitting
Obstetrics/Gynecology	1.00 [Reference]		1.00 [Reference]
Urology	0.86 (0.64,1.14)	.30	0.87 (0.65,1.18)	.38
Others*	0.74 (0.58,0.93)	.01	0.78 (0.62,0.99)	.04
Charlson comorbidity index	0.96 (0.92,1.00)	.04	0.97 (0.93,1.01)	.17
Prior urinary incontinence	1.24 (1.05,1.45)	.01	1.20 (1.02,1.42)	.03
Number of pessary re-fittings	0.97 (0.93,1.01)	.16	0.99 (0.95,1.03)	.61

^*Other providers performing first pessary fitting include nurse practitioner (n=100), physician assistant (n=19), Internal Medicine (n=18), Family Practice (n=17), Diagnostic radiology (n=18), missing (n=122), and all other specialties (n=52).Dual eligibility refers to patients eligible for both Medicare and Medicaid.

Comment

Principal Findings

In this population of Medicare beneficiaries, the cumulative incidence of transition from pessary to surgery for pelvic organ prolapse was 30.9% over 7 years, with a median time to surgery of 496 days after initial pessary fitting. Younger women, those with a history of prior prolapse surgery, and those with urinary incontinence at the time of pessary fitting were more likely to transition from pessary use to surgery. Additionally, new-onset urinary incontinence or rectal prolapse were associated with transition to surgery in this sample.

Results in the Context of What is Known

The transition to surgery was highest in the first few years after initial pessary fitting. However, the curve did not plateau over the 7 years of follow up suggesting ongoing transition from pessary to surgery occurring even years from the initial pessary fitting. This pattern of early transition to surgery was also observed among 444 women fit for pessary at the University of North Carolina.⁹ They found that 31% transitioned to surgery and that almost two-thirds of the surgeries occurred in the first year after pessary fitting. Similarly, in a Dutch study of women fit with a pessary,¹⁵ 79 of 335 women (23.6%) switched to surgery within 2 years. Likewise, in an older study utilizing Medicare data, 24% of women initially fit with a pessary ultimately transitioned to surgery over a nine-year period, with half transitioning within the first year.¹⁰ These early transitions to surgery may represent situations in which pessary is elected as a temporizing measure (e.g., among women awaiting definitive surgical management) or early dissatisfaction with pessary use. In the Dutch study, the most common reasons for transition to surgery after pessary fitting were pessary expulsion, inadequate symptom relief, discomfort, and incontinence.¹⁵ Due to the nature of the data source used in our analysis, we were unable to explore the specific reasons for transition to surgery in the Medicare population. We did, however, identify de novo diagnoses of rectal prolapse in 4.5% and urinary incontinence in 25.7% of women after pessary fitting, and these were associated with 2–3-fold higher odds of transitioning to surgery. As these symptoms may be addressed concomitantly with prolapse surgery, the emergence of these symptoms after pessary fitting may have been a significant driver of the transition from pessary to surgery for some patients.

Clinical Implications

This data adds to our current understanding of pessary use in an older population and may aid in counseling patients presenting for treatment of bothersome pelvic organ prolapse about expectations for pessary use. Although there are many factors that ultimately influence the decision to choose a management strategy for pelvic organ prolapse, an improved understanding of the incidence and timing of surgical transition after pessary fitting provides more specific information about “typical” pessary use. This improved understanding can be used to address patient’s goals and expectations, including likelihood of transition to surgery long-term, which is important in the shared decision-making process.

Research Implications

Our findings suggest that prolapse care after pessary fitting may be influenced by social determinants of health. We found that dual Medicare/Medicaid eligibility was associated with a lower hazard of transitioning to surgery after initial pessary fitting for prolapse. Dual eligibility is often utilized as a proxy for socioeconomic status, although dual-eligible patients have also been found to have more chronic medical conditions, with higher average medical risk scores, and are more likely to be institutionalized.^14,17 The lower hazard of transition to surgery in this population may reflect higher medical comorbidity, but it may also reflect underlying socioeconomic disparities in care, either due to financial constraints, limited resources (such as transportation) or access to specialty care. Further research is needed to understand the role of socioeconomic disparities on prolapse management.

Another unanswered question pertains to long term outcomes of women who do not transition to surgery. We were unable to determine whether those women continued pessary use or abandoned use over time. Tracking pessary continuation is difficult in the Medicare dataset due to the lack of a procedure code for ongoing pessary management. As a result, the “no surgery” group includes patients that continued using their pessaries and those who abandoned prolapse treatment, which limits the conclusions we can draw from this group. Investigation into pessary continuation and discontinuation in the Medicare population would further enhance the findings presented here.

Strengths and Limitations

This study has limitations. As an administrative database, the Medicare 5% LDS is subject to the risk of coding errors, which may have affected the observed rates of the procedures in this population, and some variables, including prolapse stage and coital activity, are not available for analysis. We do not have data on the indication for surgery within this population. Some women may have acquired another indication for surgery, particularly hysterectomy, other than prolapse during the study period, which may have overestimated the rate of surgery for prolapse in this sample. These findings may not represent patterns of care in other countries. Also, we elected to exclude women who did not have continuous Medicare data for at least 3 years after pessary fitting to ensure that we would capture surgeries that occurred remote from pessary fitting. Women who died within the first 3 years or were no longer covered under Medicare were therefore excluded. Our results may underestimate the incidence of death in this population or the number of surgeries within the first two years if those patients were no longer covered under Medicare. However, because we restricted the inclusion criteria to those over 65, those women would remain eligible for Medicare and would, therefore, likely be retained in the dataset. Also, because there is not a distinct specialty code for urogynecologists in the Medicare Data Set, we were unable to determine whether transition to surgery is influenced by initial pessary fitting by a subspecialist urogynecologist (compared to a general obstetrician/gynecologist or urologist). Finally, we accommodated a 1-year look-back. Patients may have had a pessary fitting before the 1-year look-back or before turning 65 and becoming eligible for Medicare, and as a result, would not have been captured in our sample.

Our study also has several strengths. The use of the Medicare 5% LDS enabled us to utilize a large sample size of over 2,000 patients for this analysis and follow these patients continuously over time. This dataset represented a longer interval of follow up (7 years) in comparison to prospective studies, which have typically represented less than 3 years of follow up.^9,15 Also, the study population is likely highly generalizable because U.S. adults over 65 years are typically covered by Medicare insurance. In contrast, case series from tertiary centers may be biased by the inclusion of women with access to tertiary care.

Conclusions

Our findings provide valuable information on the demographics of female Medicare beneficiaries who initially choose pessary for prolapse treatment and subsequently transition to surgical management. This may help providers counsel patients on expectations for long-term conservative versus surgical management of pelvic organ prolapse and improve the shared decision making.

AJOG At a Glance:

A. Why was this study conducted?

Treatment options for pelvic organ prolapse include expectant management, conservative management with a pessary, or surgical management. Data on progression from pessary to surgical management among older women are limited.

B. What are the key findings?

Cumulative incidence of transition from pessary to surgery for pelvic organ prolapse was 30.9% over 7 years, with a median time to surgery of 496 days after initial pessary fitting. De novo urinary incontinence and rectal prolapse were associated with increased risk of transition to surgery.

C. What does this study add to what is already known?

Our findings provide valuable information on the demographics of female Medicare beneficiaries who initially choose pessary for prolapse treatment. Transition from pessary to surgery continues, remote from the initial pessary fitting.