Failures of Sacral Neuromodulation for Incontinence

Journal JAMA Surgery
Authors Chughtai B, Thomas D, Sun T, Sedrakyan A
Year Published 2018
Link to publication


Sacral neuromodulation (SNM) is increasingly used to treat lower urinary tract symptoms and bowel symptoms refractory to conventional therapy.1 The US Food and Drug Administration approved the device in 1997, and the device was modified in 2006, which included a 50% reduction in size and an improvement in leads to decrease lead migration. These changes at least partially explain the more than 3.5-fold increase in the number of SNM procedures in the following 5 years.2 While guidelines based on small clinical studies with highly selected patients report 50% or greater symptom improvement in most patients,1,3,4 an early safety report using real world data found that a large number of reinterventions occurred in the Medicare population.2

The indications for SNM use are growing despite the lack of evidence. Hence, we sought to determine the reintervention rate associated with treatment failures and device malfunctions at 1, 3, and 5 years after surgery. The secondary aim was to evaluate the volume-outcome relationship.


We used data from 2008 to 2015 from the Statewide Planning and Research Cooperative System, with longitudinal records of all procedures conducted in New York. Patients who underwent the SNM second-stage procedure (InterStim; Medtronic) as well as treatment failures and device malfunctions were identified using Current Procedure Terminology, Fourth EditionInternational Classification of Diseases, Ninth Revision, and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes. We evaluated the reintervention (ie, revision, replacement, or removal) rate of SNM within 1 year, 3 years, and 5 years following the initial procedure and analyzed principal diagnoses during reinterventions. We defined high hospital and surgeon volumes based on the upper quartile and used multivariable generalized linear mixed models to account for hospital clustering (Figure). This protocol was reviewed and approved by the institutional review board at Weill Cornel Medical College. Because deidentified information from the Statewide Planning and Research Cooperative System database was used, informed consent was waived.


A total of 4946 patients underwent the SNM procedure between 2008 and 2015. The mean (SD) age was 60.7 (16.2) years, and 3943 patients (79.7%) were female. The primary indication for SNM was wet overactive bladder (n = 2534 [51.2%]). Reinterventions were conducted for 592 of 4313 patients (13.7%) within 1 year, for 767 of 2920 (26.3%) within 3 years, and for 575 of 1514 (38.0%) within 5 years following SNM placement. The most common principal diagnoses for reintervention were related to treatment failure (278 of 592 patients [47.0%]) and device malfunction (280 of 592 [47.3%]) at 1 year. At 5 years, 190 of 575 reinterventions (33.0%) were related to treatment failure and 367 of 575 (63.8%) to device malfunctions (Table). There were no significant differences between high vs low volume hospitals or surgeons.


We found an alarming number of reinterventions, mostly because of device malfunctions, in patients undergoing SNM for all indications. One in 3 patients had some form of reintervention—either revision or removal—within 3 to 5 years of initial device placement, which is in stark contrast to the manufacturer’s reported device survival probability of 97.9% over a 3-year period.5 Almost half of the patients undergoing reintervention needed it within the first year, and hospital and surgeon volume did not influence these results.

Our previous study2 using limited 5% Medicare data reported an 11.3% reintervention rate at 90 days and a 33% rate at 5 years. In a 5-year follow-up study of patients implanted with SNM for overactive bladder, Siegel et al6also reported only a 67% therapeutic success rate. We extend these results in, to our knowledge, the first large-scale, all-inclusive statewide study demonstrating high SNM device failure. We found that device malfunction was the second most common indication for early reinterventions and the predominant indication at 5 years. Moreover, our results demonstrate that even in the hands of high-volume surgeons, the invasive surgical reintervention rate remains very high.

Limitations include generalizability to the entire US population, despite inclusion of all ages and data from all of New York, with its diverse population and practices. Use of billing codes may introduce some misclassification but is a valid method, particularly for reinterventions.


Sacral neuromodulation is associated with a very high rate of failure, as measured by reinterventions, that occur at alarming rates within 5 years of placement. A device registry is urgently needed in this setting to advise stakeholders and assist future innovations.