Introduction

Ageing is associated with a number of chronic health conditions including an increased incidence and prevalence of overactive bladder (OAB), fecal incontinence (FI), and idiopathic urinary retention (UR) [1]. OAB in particular disproportionately impacts older women, with 45% of women over 65 reporting bothersome OAB at an annual cost of US$ 14 billion [2]. Although less common than OAB, FI impacts nearly 1 out of 10 women and can have an even more negative impact on quality of life [3]. Although these conditions are not life threatening, they are life altering and are associated with increased anxiety, depression, falls and fractures, and social isolation [4, 5].

Since the late 1990s, sacral neuromodulation (SNM) has been a treatment option for refractory OAB and non-obstructive, idiopathic UR, and in 2011, the FDA approved SNM for the treatment of FI [6]. Despite the availability of SNM for the better part of three decades, there are limited epidemiological data describing long-term outcomes following SNM, particularly in a population of older women. Existing literature on long-term outcomes beyond 1 year only capture outcomes on a small number of patients [7,8,9].

Comparative data regarding outcome differences following basic and advanced SNM trials in older women are even sparser. The basic trial, the percutaneous nerve evaluation (PNE), is performed in an office setting and usually involves blind placement of an electrode and evaluation of symptoms for 3–7 days. If the PNE demonstrates more than 50% improvement in symptoms, it is considered successful and followed by a full implant, consisting of placement of a neuro-electrode and implantable pulse generator (IPG) in the operating room. The advanced trial involves the placement of the permanent neuro-electrode in the operating room under fluoroscopic guidance, and evaluation for 1–2 weeks. Similar to PNE, a 50% improvement is considered successful and the patient is brought back to the operating room for IPG placement or removal of the neuro-electrode if unsuccessful. Given that the surgical technique and perioperative care differ significantly between these two approaches, these differences could impact clinical outcomes and device-related reoperations such as neuro-electrode replacement or revision, battery replacement, or revision or removal of the system.

A better understanding of outcomes following SNM implantation for older women will help clinicians to counsel patients about the risks of subsequent surgical reintervention. The expected battery life for the SNM IPG was approximately 5 years, extending up to 8 years for optimally placed leads. Given the expected battery life and that the existing literature had only reported on outcomes up to 3 years following initial implantation, the primary objective of this study was to describe long-term device-related reoperations at 8 years following SNM implantation for women 65 years of age and older for the indications of OAB, FI, and/or UR. We also sought to describe long-term device-related reoperations at 1 and 3 years as well as differences in outcomes between staged and PNE/full implantation at 1, 3, and 8 years. We hypothesized that staged SNM implants would be more common in this older cohort of women and that women undergoing staged SNM implants would have fewer 8-year device-related reoperations.

Materials and Methods

The 2010–2019 Medicare 100% Outpatient Limited Dataset was used to identify women aged 65 years and older who underwent an SNM implantation procedure. The Institutional Review Board deemed this study exempt from review. Given that in the USA, people 65 years of age and older are eligible for Medicare, a federally funded health insurance program, the Medicare 100% Limited Dataset (LDS), which provides detailed longitudinal data for Medicare beneficiaries, is well suited to explore outcomes in older women undergoing SNM procedures. The following Current Procedure Terminology (CPT) codes were used to identify the index SNM implant: 64561 (percutaneous implantation of neurostimulator electrode array), 64581 (incision for implantation of neurostimulator electrode array), and 64590 (insertion or replacement of peripheral IPG). PNE followed by full implant was defined using CPT 64561 followed by 64581 and 64590 (henceforth referred to as “full implant”). Staged implant was defined as CPT 64581 followed by 64590 (henceforth referred to as “staged implant”).

The earliest procedure (staged implant or full implant) was selected. Women were included if the indication for SNM implantation was OAB, FI, or UR using the ICD codes (596.51, 788.31, 788.33, 788.41, 788.63, 788.43, 787.60, 788.20, N32.81, N39.41, N39.46, R35.0, R39.15, R35.1, R15., and R33.). Women who underwent SNM for other indications were excluded. Any IPG implantation (CPT 64590) that was not preceded by CPT 64561 or CPT 64581 was excluded. Women were excluded if they did not have at least 12 months of Part B coverage prior to their procedure and were followed until 8 years after their index procedure, their last date of coverage, or the date of their first adverse event, whichever was earliest.

Demographic and clinical characteristics of the patients included were described. The primary outcome was a composite outcome defined as 8-year device-related reoperations following the initial implant, which included: IPG replacement (CPT 64590); IPG revision or removal (CPT 64595); and neuro-electrode revision or removal (CPT 64585). Other outcomes of interest included mechanical breakdown (T85.111A).

Kaplan–Meier analyses were performed for each of the outcomes. Follow-up time for those without an event was determined as the last date covered in the Master beneficiary denominator file. Women were right censored at the last date of coverage for all time to event analyses if an event had not occurred by that date. Log rank test was performed to compare time-to-event by implant category. Comparisons between the staged and full implant groups were made using t test and Chi-squared test as appropriate. Univariate and multivariate Cox regression was performed to describe the association between the type of implant (staged versus full) and any device-related reoperation, controlling for age, geographic region of residence, and diagnosis.

Results

Between 2010 and 2019, a total of 32,454 women met all the criteria for inclusion in the analysis. Table 1 describes the characteristics of the overall cohort. The mean age was 74 ± 6 years. Over one third of the cohort (47%) resided in the Southern USA, and the large majority (94%) of patients identified as white. The most common diagnosis was OAB (71%), followed by OAB and FI (13%) and FI only (8%). Staged procedures were more commonly performed than full implants, with 60% of the cohort undergoing staged procedures. Patients who underwent staged SNM procedures were more likely to have the diagnosis of FI (p < 0.001).

Table 1 Characteristics of study population
Full size table

Table 2 describes the rates of device-related adverse outcomes in the 8 years following initial implantation including any device-related reoperations such as IPG replacement, neuro-electrode removal or revision and IPG removal or revision. The overall rate of any device-related reoperation was 24%, with a higher rate seen in the full implant cohort compared with the staged cohort (26% vs 22%, p < 0.001). This difference was largely driven by the high rate of IPG replacement seen in the full implant cohort compared with the staged implant cohort (16% versus 10%, p < 0.001).

Table 2 Device-related outcomes through 8 years
Full size table

Table 3 describes the cumulative rate of device-related adverse outcomes at 1, 3, and 8 years. The rate of any device-related reoperation was 9.4% at 1 year, 20% at year 3, and 43% at year 8. The rates of IPG replacement, IPG revision or removal, and neuro-electrode revision or removal steadily increased over the 8-year study period.

Table 3 Cumulative incidence of reoperations at 1, 3, and 8 years
Full size table

Figure 1 depicts the Kaplan–Meier curves for time-to-any adverse event over 8 years comparing stage and full implant approaches. The difference in rate of any reoperation is highest in the first year following initial implantation but steadily declines over the 8-year window. Figure 2 depicts the Kaplan–Meier curves for time-to-IPG replacement over 8 years. Over the course of 8 years, the rate of IPG replacement steadily declines to 24%, and a difference between full and staged implants is noted throughout those 8 years.

Fig. 1
figure 1

Kaplan–Meier time to any device-related reoperation

Full size image
Fig. 2
figure 2

Kaplan–Meier time to implantable pulse generator removal/revision

Full size image

After controlling for age, indication for SNM therapy, and geographic region of residence, staged SNM procedures were associated with a lower likelihood of undergoing any device-related events in the 8 years following initial implantation (adjusted hazard ratio 0.83, 95% CI 0.79–0.87) compared with full procedures (Table 4).

Table 4 Characteristics associated with any reoperation
Full size table

Discussion

This study describes outcomes in female Medicare beneficiaries using the 100% Medicare Limited Dataset between 2010 and 2019. To our knowledge, this study represents the largest cohort of older women—over 32,000 aged 65 years and older—undergoing SNM for the treatment of OAB, FI, or UR. In this cohort, OAB was the most common diagnosis associated with the procedure, and staged procedures were more commonly performed than PNE/full procedures. In the 8 years following the initial device implantation, any device-related reoperation occurred in nearly half of the patients. The most common reoperation was IPG replacement. Although reoperation for IPG replacement may be an expected intervention, it is still a surgical intervention that incurs risk to the patient and is an important outcome to track for patient counseling. Compared with full implants, staged procedures were associated with a 17% lower likelihood of reoperation for any device-related reoperation compared with full implants in the 8 years following initial implantation.

This study found that the risk for surgical reintervention in older women undergoing SNM at 8 years is high, with 42% requiring a device-related reoperation. Our data are novel in that they reflect a steadily increasing rate of reoperation over time. Although systematic reviews have described a wide range of reported rates of surgical reintervention (between 3 and 42%), this study’s rate of reoperations appears to mirror those described in recently published literature. Chughtai et al. have published two studies reporting long-term outcomes following SNM [10]. The first study, published in 2015, described at a mean of 3 years a reoperation rate of 26% for device-related issues in a cohort of male and female Medicare beneficiaries between 2000 and 2010. At 3 years, we found a similar rate of 20% for any device-related reoperation in our cohort. The second study by Chughtai et al., published in 2018, reported a similar cumulative rate of reintervention at 3 years of 26.3% (767 out of 2,920) using a statewide longitudinal procedural database [11]. At 5 years, they noted a steady increase in the rate of reoperation at 38%. The majority of these reinterventions were due to treatment failures and device malfunction. Other studies with shorter follow-up interviews have reported rates of reintervention closer to 30% [12,13,14]. Our study adds to the existing literature and helps to shed light on the rate of reintervention following SNM at even longer-term follow-up, which may pose a unique risk to older women with multiple chronic conditions who are disproportionately impacted by OAB, FI, and UR.

When comparing trial strategies, we found that women who underwent a staged approach for SNM implantation had a significantly lower likelihood of any device-related reoperation in the 8 years following initial implantation than women who underwent PNE followed by a full implantation of the device after controlling for age, indication for procedure, and region of residence. The difference in any device-related complication was likely driven by the higher frequency of replacement of the IPG in the full implantation cohort. The higher frequency of IPG replacement may reflect a shortened battery life or complications associated with the IPG implant, typically discomfort or malpositioning, although we do not have clinical data to support these suppositions. Additionally, with a staged approach, patients undergo a trial based on the permanent neuro-electrode rather than temporary neuro-electrode, and given that the trial neuro-electrode stays in the same location after the trial, patients experience clinical improvement post-IPG implantation that was identical to the trial experience, possibly leading to decreased revisions or removals for abnormal sensations following a full implant. However, surgeons should weigh the risks associated with subjecting older patients to two separate anesthesia events when deciding whether to offer full versus staged procedures.

A strength of this study is the use of the 100% Medicare limited dataset, which is the largest collection of hospital data in the USA and thus represents a comprehensive, geographically diverse cohort. This dataset also allows for long-term longitudinal follow-up of women undergoing SNM implantation. Limitations of this study include those that are inherent to database studies such as its retrospective design. Coding errors in the dataset lead to misclassification bias. Also, we were unable to collect relevant clinical information associated with device-related reoperations, such as pain associated with the device, lack of clinical response of the device, or concern regarding surgical site infection. We were not able to collect information on the clinical practice around these procedures, such as surgeon technique, volume, or experience with the procedure; or administration of perioperative antibiotics. This study did not distinguish patients who had a failed PNE and then underwent a staged procedure either, but future studies may consider exploring factors associated with a failed PNE. Also, given that the follow-up time varied by procedure, this difference introduces potential bias into our findings. Last, given improvements in SNM technology, expected IPG battery life exceeds 15 years, and thus the findings of this study may not be applicable to new generations of SNM; however, we still believe that it is important to investigate differences between expected and actual battery life of the implanted IPGs.

Device-related reoperations are an important clinical outcome that needs to be examined in greater detail as it is a significant contributor to increased health care costs and it poses additional risk to the patient. Future research should examine the role that surgeon experience and surgeon volume play in these device-related reoperations. Also, patient factors that may lead to poor outcomes should be explored in order to better understand those patients who are at a higher risk of requiring a reoperation following SNM implantation for OAB, FI, or UR, particularly in older women.

In a cohort of over 32,000 female Medicare recipients 65 years of age and older who underwent an SNM procedure for the treatment of OAB, FI, or UR, 42% of women required a device-related reoperation in the 8 years following initial implantation, with those undergoing staged implantation having a lower risk of subsequent re-intervention. Additional research is needed to identify and mitigate both patient and surgeon factors that increase a patient’s likelihood of reoperation.