Reoperation Rates and Risk Factors After Spinal Cord Stimulation Revision Surgery

Abstract

Introduction:

Spinal cord stimulation (SCS) is an effective treatment for patients with refractory chronic pain. Despite its efficacy, rates of reoperation after initial implantation of SCS remain high. While revision rates after index SCS surgeries are well reported, less is known about rates and risk factors associated with repeat reoperations. We sought to evaluate patient, clinical, and surgical characteristics associated with repeat reoperation among patients who underwent an initial SCS revision procedure.

Methods:

We performed a retrospective review of patients who underwent SCS revision surgery performed at a single institution between 2008 and 2022. Patients were stratified by whether they underwent a single revision (SR) or multiple revision (MR) surgeries. Multivariate logistic regression was performed to determine risk factors associated with repeat SCS revision. Kaplan-Meier survival analysis was used to compare rates of devices requiring revision across groups.

Results:

A total of 54 patients underwent an initial SCS revision. Of these, 15 (28%) underwent a second revision. The most common indication for revision surgery was lead migration (65%). No significant differences were observed in age, BMI, comorbidities, lead type, and revision indication among the SR and MR groups. On multivariate adjusted analysis, only cervical lead position was significantly associated with repeat reoperation (OR 7.10, 95% CI [1.14, 44.3], p=0.036). Time to reoperation after a single and multiple revision SCS surgeries did not differ.

Conclusions:

Among patients who undergo SCS reoperation, a substantial portion require additional revisions. Cervical lead placement may be associated with a higher risk of repeat revision surgery compared to thoracic lead positioning. Consideration of lead positioning in the decision to perform and undergo reoperation may therefore result in lower revision rates and improved clinical outcomes among SCS patients with multiple revisions.

Introduction

Chronic pain impacts 30% of the world’s population, resulting in significant work disability, decreased quality of life, and high healthcare costs [1]. In select patients who fail first-line pharmacological treatments, neuromodulation with spinal cord stimulation (SCS) may represent a suitable intervention for intractable chronic pain [2]. SCS is currently utilized for several pain conditions, including refractory low back and leg pain, persistent spinal pain syndrome (PSPS), complex regional pain syndrome (CRPS), and refractory angina pectoris (RAP) [3]. In a randomized clinical trial of 100 patients with PSPS, 48% of patients achieved the primary endpoint at 6 months (≥50% pain relief) compared to 9% of patients receiving conventional medical management [3]. Subsequent prospective studies have reported improvements in pain, quality of life, and patient satisfaction after SCS for PSPS [4,5]. Among patients with CRPS and RAP, studies have also reported favorable outcomes after SCS, resulting in significant gains in functional status and reduction in pain symptoms [6,7]. However, despite its efficacy in select populations, device placement is associated with a high risk of complications [8,9].

Although the complications related to SCS are generally linked to low morbidity, their incidence is common, with implantation generally carrying a risk of complications of 20% to 40% [9–12]. At 12-months, Kumar et al. observed 32% device-related complications after SCS placement for PSPS, with electrode migration (10%) being the most common, followed by infection and wound dehiscence (8%) [3]. Revision surgeries have also been reported in up to one-third of patients implanted with SCS system due to lead migration, lead fractures, infection, generator malfunction, and hardware-related pain [8,13]. With the growing utilization of SCS and technological advances, such as high-frequency stimulation and the development of closed-loop paradigms to improve efficacy, it is critical to understand and inform patients regarding the potential risks associated with SCS surgery [14–16]. While multiple studies have examined the incidence and risk factors of primary revision surgery after SCS, less is known regarding the rates of complications and repeat reoperation after secondary or multiple revisions. In this study, we sought to evaluate rates of repeat revision SCS surgery and examine patient, clinical, and surgical factors associated with repeat reoperation among patients who underwent an initial SCS revision procedure.

Methods

Population and Data Acquisition

We conducted a retrospective review of all patients between 18 and 89 years of age who had previously undergone SCS implantation and who subsequently underwent revision at a single institution during the period beginning in January 2014 through October 2022. All revision surgeries were performed by the authors [AR], while the institution of initial implantation varied. All patient data were tracked through electronic records from the time of initial revision through the end of the observation period for a median duration of 43.0 months (IQR 25.−61.8). Patient Current Procedural Terminology (CPT) codes 63663, 63664, 63685, and 63688 were used to identify index SCS surgery. Device explantations without subsequent replacement and procedures indicated for battery end-of-life replacement were not included in the definition of revision surgery. This study adheres to STROBE reporting guidelines and was approved by the local Institutional Review Board [17].

Predictor and Outcome Variables

Patient demographic, clinical, procedural, and outcome data were abstracted from the electronic medical record. Demographic variables included age, sex, body mass index (BMI), substance use, and psychiatric comorbidities. Substance use included tobacco, alcohol, and other recreational substances. Pre- and postoperative opioid dosages were calculated as morphine milligram equivalents (MME). Initial indication for SCS was recorded and categorized as spinal pain (PSPS, chronic back pain) or other, which included neuropathic pain due to spinal cord injury, brachial plexus injury, post-stroke pain, and other upper extremity neurogenic pain.

Revision indications were identified as lead migration, generator malposition, device failure, infection, and refractory pain. Surgical predictors of repeat reoperation were categorized as follows: revision procedure type (lead revision vs generator revision/other), lead type (percutaneous vs paddle), lead location (thoracic vs cervical), and the occurrence of any perioperative complications (lead fracture, infection, hematoma, and CSF leak). Our primary outcome was to examine risk factors predictive of multiple SCS revisions among patients who underwent an initial primary SCS revision. The intervals between initial and subsequent revisions were calculated according to the recorded date of service for each procedure.

Revision Procedure

In all instances of revision, either the leads, the generator, or both were revised. In the case of paddle leads, a layer of Tisseel fibrin sealant was placed atop the laminotomy site to adhere the electrodes to the dura. Typically, both paddle and percutaneous leads were anchored via anchoring devices, such as the Cinch Lead Anchor (Abbott Laboratories, Abbott Park, IL) or Clik Anchor (Boston Scientific, Marlborough, MA). Leads were secured to the anchors via torque wrench, and the anchors were subsequently sutured to the thoracolumbar fascia overlying the paraspinous muscles with three 2–0 Ti-Cron stitches. Proper placement of leads was confirmed with intraoperative AP and lateral fluoroscopy. A strain relief loop of the electrode was created and held within an epifascial subcutaneous pocket. Occasionally, electrode placement was re-confirmed postoperatively by fluoroscopy.

Statistical Analysis

Patients were stratified into two groups: those who underwent single SCS revision (SR) and those who underwent ≥2 revisions which we termed multiple revision (MR). Descriptive analyses were performed comparing the SR and MR cohorts. Univariate analyses were conducted using t-tests and chi-square tests for continuous and categorical variables, respectively. A backward stepwise multivariate logistic regression with an entry criterion of p < 0.30 was used to evaluate patient, clinical, and surgical risk factors associated with multiple revision. A receiving operating characteristic curve (ROC) was used to evaluate the discriminatory capacity of the regression model, with a ROC curve value of 0.5, indicating no greater discrimination than chance alone. We assessed model calibration using the Hosmer-Lemeshow test with p > 0.05 signifying a good model fit based on established methods [18]. Analyses were performed using Stata BE, version 18 (StataCorp LP). All tests were 2-sided and statistical significance was established with a p < 0.05. Kaplan-Meier survival curves were plotted for the interval between initial implant and first revision and the interval between first revision and second revision. Survival was defined as the time elapsed without need for revision; total explantations and routine battery replacements were not considered revision procedures. No data points were censored, and curve comparison was evaluated by Mantel-Cox log-rank test. Figures were created in GraphPad Prism 9 and SankeyMATIC.

Results

A total of 54 patients underwent SCS revision surgery at the authors’ primary institution during the study period. Among these patients, 39 (72%) received a single revision (SR) and 15 (28%) had multiple revisions (MR) (shown in Table 1). Of the 39 patients undergoing a single revision, 32 (80%) underwent initial implantation at the primary institution. Of the 15 patients undergoing multiple revisions, 13 (87%) underwent initial implantation at the primary institution (Fig 1a,b). Of the 15 patients who received multiple revision surgeries, 13 (87%) underwent two revisions, 1 (7%) underwent three revisions, and 1 (7%) underwent four revisions. Among 261 patients who underwent initial device implantation at the authors’ primary institution during the study period, 42 (16%) underwent revision over the same period. Qualitative review of postoperative notes indicates that among those with two revisions, 8 (62%) achieved adequate pain relief while 5 (38%) did not. Among patients receiving more than two revisions, SCS therapy was ultimately unsuccessful. The overall median age at index revision was 55.5 years, 52% were male, the average BMI was 30.6 kg/m², and 81% were indicated due to PSPS or chronic low back and leg pain. Among the entire cohort, half had a psychiatric comorbidity, 22% were smokers or used other recreational substances, and 46% had no change in their opioid usage after the initial SCS procedure. Among 27 patients with underlying psychiatric conditions, 5 of 9 (56%) patients reporting substance use required multiple revisions, while only 1 of 18 (6%) patients reporting no substance use required multiple revisions (p=0.003). This discrepancy based on substance use was not detected among those without underlying psychiatric co-morbidity, though only 3 patients reported substance use in this category (p=0.19). Forty-one (76%) patients underwent placement with a paddle electrode compared to 24% that had percutaneous leads implanted. With regards to the spinal level, lower thoracic lead placement was most common (74%), compared to cervical positioning (26%).

Table 1.

Univariate analysis of patient demographics and clinical features. All values reported as Number of Patients (%) unless otherwise indicated.

Demographic/Clinical Features	Overall (n=54)	Single Revision (n=39)	Multiple Revision (n=15)	p value
Age mean years (SD)	54 (16)	54 (17)	52 (14)	0.69
Age > 65	15 (28)	13 (33)	2 (13)	0.14
Sex				0.28
Male	28 (52)	22 (56)	6 (40)
Female	26 (48)	17 (44)	9 (60)
BMI				0.82
<30	27 (50)	19 (49)	8 (53)
30 to 34.9	13 (24)	9 (23)	4 (27)
≥ 35	14 (26)	11 (28)	3 (20)
ASA III	28 (52)	20 (51)	8 (53)	0.89
Substance Use	12 (22)	7 (18)	5 (33)	0.22
Psychiatric Comorbidities	27 (50)	21 (54)	6 (40)	0.36
ΔMME				0.94
No change	25 (46)	18 (46)	7 (47)
Increase	16 (30)	12 (31)	4 (27)
Decrease	13 (24)	9 (23)	4 (27)

Figure 1.

(a) Flow chart depicting patient inclusion criteria and breakdown by initial device implantation. (b) Distribution of initial device implantation hospital (OHSU vs external) across single revision (SR) and multiple revision (MR) patients.

The most common indication for any revision was lead migration (65%), followed by device-related pain (13%) (shown in Table 2). Less common indications were refractory pain, device malfunction, lead fracture, and generator malposition. There was a total of four (7%) intraoperative and perioperative complications documented across the 54 initial revision procedures, with 2 in the single revision cohort (intraoperative lead fracture, CSF egress) and 2 in the multiple revision cohort (hypotension, hematoma). Only one infection was identified, resulting in a reoperation. No significant differences in indication for reoperation were observed between both the SR and MR groups. A larger proportion of patients who underwent multiple revisions were <65 years (87% vs 67%, p=0.14), were female (60% vs 40%, p=0.28), and were smokers or used other substances (33% vs 18%, p=0.22). No differences were observed with regards to ASA status or BMI.

Table 2.

Univariate analysis of revision features. All values reported as Number of Patients (%) unless otherwise indicated.

Revision Features	Overall (n=54)	Single Revision (n=39)	Multiple Revision (n=15)	p value
Initial SCS Indication				0.86
Spinal pain	44 (81)	32 (82)	12 (80)
Other	10 (19)	7 (18)	3 (20)
1st Revision Indication				0.65
Lead Migration	35 (65)	26 (67)	9 (60)
Adverse Event/Device Defect	19 (35)	13 (33)	6 (40)
Device discomfort	7 (13)	5 (13)	2 (13)
Refractory pain	5 (9)	3 (8)	2 (13)
Inappropriate device stimulation	3 (6)	3 (8)	0 (0)
Lead fracture	2 (4)	1 (3)	1 (7)
Generator Malposition	1 (2)	1 (3)	0 (0)
Infection	1 (2)	0 (0)	1 (7)
Time to 1st Revision median months (SD)	8.6 (37.8)	10.1 (38.2)	6.3 (38.1)	0.81
Initial Revision Procedure				0.81
Lead Revision	42 (78)	30 (77)	12 (80)
Generator Revision/Other	12 (22)	9 (23)	3 (20)
Lead Type				0.66
Paddle	41 (76)	29 (74.4)	12 (80.0)
Percutaneous	13 (24)	10 (25.6)	3 (20.0)
Lead Location				0.031
Thoracic	40 (74)	32 (82)	8 (53)
Cervical	14 (26)	7 (18)	7 (47)
1st Revision Perioperative Complications	4 (7)	2 (5)	2 (13)	0.30
2nd SCS Revision Indication				-
Adverse Event/Device Defect	-	-	8 (53)
Lead Migration	-	-	7 (47)

There was no significant difference in the likelihood of undergoing at least one revision due to lead migration in those with percutaneous and those with paddle leads (77% vs 66%, p=0.45). Similar rates in the utilization of paddle and percutaneous electrodes were observed between SR and MR cohorts. Notably, the rates of cervical lead placement were significantly higher in the MR group compared to the SR group (47% vs 18%, p=0.031) on univariate analysis. However, there was no significant difference between the rates of multiple revision in cervical paddle and cervical percutaneous leads (45% vs 33%, p=0.70), nor was there a difference in thoracic paddle and thoracic percutaneous leads (15% vs 20%, p=0.70) (Fig 2c).

Figure 2.

Percentage of initial SCS devices requiring revision among patients undergoing at least one revision (solid) and percentage of first revision devices requiring second revision among patients undergoing two or more revisions (dashed) over time.

On adjusted multivariable analysis, cervical lead positioning was the only statistically significant predictor of undergoing multiple revisions (OR 7.10, 95% CI [1.14, 44.3], p=0.036, model AUC=0.84) (shown in Table 3). Female sex (OR 4.32, p=0.08), substance use (OR 4.49, p=0.10), and age below 65 years old (OR 0.16, p=0.09) did not meet statistical significance. The median duration from initial device implantation to first revision for the entire cohort was 8.6 months (IQR 3.4–16.6) and did not significantly differ from the median device survival in the MR cohort (median=6.3, IQR 1.2–18.4, p=0.94), nor across groups (SR median=9.0, IQR 4.6–12.9, p=0.87). Among all patients requiring at least one revision, 69%, 85%, and 89% underwent revision by the end of 1, 2, and 3 years following initial device implantation, respectively. Among patients with multiple revisions, median duration between first and second revision was 9.6 months (IQR 3.5–16.3), and this did not significantly differ from the overall duration to first revision (Fig 3, p=0.91). In this cohort, 60%, 87%, and 93% required secondary revision by the end of 1, 2, and 3 years following first revision, respectively.

Table 3.

Predictors of multiple SCS revisions by multivariate analysis.

Variable	OR (95% CI)	p value
Age 65+	0.16 (0.18–1.36)	0.093
Sex: Female	4.32 (0.84–22.2)	0.080
Initial SCS Indication: Other	0.29 (0.03–2.81)	0.288
Psychiatric Comorbidities	0.37 (0.07–1.96)	0.243
Cervical Lead Location	7.10 (1.13–44.28)	0.036
Lead Type: Paddle	3.01 (0.44–20.89)	0.263
Substance Use	4.49 (0.75–26.76)	0.099

Figure 3.

Revisions stratified by site of lead placement. (a,b) Sankey diagram of first and second revision indications in thoracic and cervical patients (AE/DD=Adverse Event/Device Defect). (c) Distribution of single and multiple revision patients by lead type–paddle vs percutaneous (perc)–for thoracic and cervical leads.

Discussion

In this study, we examined the rates and risk factors associated with repeat revision surgery after initial SCS revision. Over the 9-year period analyzed, 28% of patients with a history of SCS revision required at least one additional revision. This is consistent with previously reported SCS revision rates ranging from 20–30% [19,20]. While patient and clinical factors predicting initial SCS revision have been studied extensively, less is known about the risk factors associated with additional revision. Given that approximately 30% of patients may experience an additional revision should they continue to elect SCS therapy, we sought to identify predictors of repeat revision surgery to inform patients and providers, manage expectations, and aid in meaningful clinical decision-making.

We found that cervical lead placement was significantly associated with an increased likelihood of undergoing more than one revision SCS surgery. Among our cohort, 47% of patients who required repeat SCS revision had electrodes positioned at the cervical level, compared to 18% who only needed a single revision. In contrast to thoracic SCS, literature examining outcomes after cervical SCS is limited. However, cervical SCS has been shown to be safe and effective for the treatment of heterogeneous chronic pain conditions, including PSPS relating to prior cervical neck surgery, CPRS of the upper extremities, and craniofacial pain syndromes [22,23,24]. In their study of 100 patients who underwent permanent cervical SCS, Chivukula et al. reported 24 revision surgeries, 33% of which were related to lead migration [23]. In comparison, lead migration accounted for approximately two-thirds of revisions in our cohort of patients indicated for either thoracic or cervical SCS, with cervical lead placement incurring a higher risk of additional revision procedures. Studies have suggested cervical motion as a possible mechanism for lead migration, and proper mitigation of lead changes may be of greater importance in patients implanted with cervical SCS systems [24,25]. Despite increased mobility of the cervical spine, we did not observe an increased likelihood for lead fracture, though this may be due to insufficient sample size. Among 48 thoracic revisions, 3 (6%) were due to lead fracture, while only 1 instance (5%) of lead fracture was documented among 21 cervical revisions. Of these revisions, 9 (43%) were revised for reasons other than lead migration. 3 of these were due to generator discomfort, 3 were due to persistent pain without evidence of lead migration, 1 was due to lead fracture following a fall, 1 was due to generator malposition, and 1 was due to infection. Adequate lead length, utilization of anchoring devices, placement of anchor tips through fascia, increasing the strength of suture material, and incorporating strain relief loops are all recommended methods to prevent lead migration [24,26]. Despite the adoption of these practices, lead migration remains a significant contributor of revision surgery, and further research is warranted to minimize lead migration, particularly in patients with cervical SCS who have already undergone an initial revision.

In our multivariate model, we did not find other statistically significant predictors of repeat revision SCS, though several variables trended towards significance which warrant mention. We found that patients over 65 years of age were less likely to undergo additional revisions (13% vs 33%). Similarly, Antonovich et al. reported a 27% decrease in revision surgery with a ten-year increase in age, postulating that increased mobility in younger populations may be a mechanism for lead migration requiring revision surgery [19]. The authors also suggested that younger patients who undergo SCS may develop tolerance to therapy more rapidly. In a long-term analysis of SCS efficacy in 252 patients, Dougherty et al. showed that younger age was associated with higher rates of explantation due to insufficient pain relief [27]. It is also possible that older populations are less likely to be recommended for or seek additional revision procedures. However, other studies are mixed, showing minimal effect of age on outcomes [28,29,30] Thus, further investigation is warranted to examine the impact of age on SCS therapy and subsequent revision surgery.

Female sex was also linked with increased revision recurrence. We observed a greater proportion of females in the MR cohort (60% vs 44%). This is consistent with a prospective analysis by Slyer et al. which reported increased explantation rates and increased pain following SCS implantation among female patients [30,31]. However, others have observed sex-based differences in reported efficacy which vary over time and stimulation frequency [32,33]. For example, Kumar et al. reports higher rates of SCS trial success and first year efficacy among females, but greater long-term benefit among males [33]. In such a case, the relative loss of efficacy over time may contribute to the slightly higher rate of multiple revision we observed among females. Perhaps the greater degree of SCS trial and early implantation success encourages females to seek revision for restoration of this benefit if therapeutic effect is diminished over time. Conversely, males who experience milder initial benefit from SCS therapy may be less likely to seek revision if therapeutic benefit is diminished. Although differences in the prevalence and perception of pain across sex are well documented, with a greater prevalence of chronic pain among females, the underlying basis for these differences is unclear [34]. We observed no significant differences in revision indication and BMI across sex, nor were there differences within sex across multiple revision status. Differences in hormone production across sex have been implicated as a physiological driver of this discrepancy, but cultural beliefs on sex and gender roles may exert influence on both patients’ expression of pain along with providers’ approach to care [34]. Further investigation into the interplay between sex and device-related morbidity is warranted. Future efforts should direct specific attention to both psychosocial and physiologic drivers of potential sex-based differences. Given the cognitive and affective components of chronic pain, we also observed trends of revision recurrence associated with both psychiatric comorbidities and substance use. Consistent with prior data, we found a greater rate of substance use in the recurrent revision cohort [35,36,37]. Many previous studies have documented the association of psychiatric comorbidities with worse outcomes, and there is an established understanding that depression and pain reciprocally exacerbate one another [27,38,39]. Accordingly, we report a greater prevalence of comorbidity–predominantly Major Depressive Disorder (MDD) and Generalized Anxiety Disorder–among our population of revision patients as compared to rates documented among the broader population of all individuals who receive SCS implantation (50% vs 30%) [40]. Notably, we observed that among those with underlying psychiatric comorbidity, those reporting substance use were significantly more likely to require multiple revisions than those who did not (56% vs 6%, p=0.0032). While the nature of this interaction is not well understood, this may be due, in part, to the increased potential for maladaptive psychologic behaviors and coping strategies seen in those with comorbid diagnoses, and pre-treatment psychiatric screening is standard [41]. With the mounting body of evidence supporting a biopsychosocial approach to chronic back pain, there is some evidence suggesting that concurrent treatment of psychiatric comorbidities such as MDD may improve SCS efficacy [42]. Given the complex subjective nature of pain and heterogeneous pain conditions treated with neuromodulation, efforts to identify patient and clinical factors that are predictive of successful SCS are difficult and remain mixed [43].

In our cohort, the median time to first SCS revision was 8.6 months with 50% of patients requiring a reoperation within 6 months after the initial revision. Meanwhile, Leplus et al., which similarly excluded explants and routine battery replacements, reported a considerably longer median time of 44.6 months (range 0–135, IQR 50.5) across 116 SCS revision patients [21]. Bir et al. reports a median time to revision of 25.4 months (range 3–145) [44], while Negoita et al. reports 16 (IQR 5.5–69, range 1–213) [13]. Most other reports either do not report median times to revision [45,46] or do not consider revisions independently of explantations [47]. One possible explanation for the abbreviated time to revision is the relatively large proportion of lead migrations we observed, while both Leplus et al. and Bir et al. report hardware malfunctions to be the leading cause for revision (Negoita et al. does not report indications) [13,21,44]. Hayek et al. shows that time to recognition of SCS complication varied significantly by complication type, with lead migration recognized at a considerably earlier median of 4.28 months (95% CI [3.42, 18.14]) relative to hardware malfunctions of the generator and leads at a median of 50.4 and 17.0 months following implantation, respectively (95% CI [10.75, 72.24] and [9.40, 37.22]) [48]. Of note, we report only 1 out of 69 revision procedures (1.4%) indicated due to infection, which is lower than the rate reported in Leplus et al. (18%) but similar to the rate reported in Bir et al. (2.8%) [21, 44]. Meanwhile, the rate of infection across all SCS procedures has been reported at 3.4% – 8.6% [9,48,49]. Many other factors which vary by institution may also influence the time to revision, including follow-up frequency, patient demographics, device and stimulation types, and operating room availability. Based on our study and prior literature, however, patients who undergo revision appear most likely to do so within the first year of device implantation [48].

Lead migrations comprised the majority of initial revision indications (65%) and half of all second revision indications (53%), with 37 of 54 (69%) patients undergoing at least one revision due to lead migration. This is substantially higher than the rates reported by others [21,44]. As such, we had concerns regarding the incidence of lead migration. However, it should be noted that the rates reported here most likely substantially overestimate the true rate of lead migration. Though we do not have access to the overall lead migration rates from the external hospitals at which some of the included patients underwent initial device implantation, we found that a total of 261 patients underwent initial device implantation at the authors’ primary institution over the duration of the study period. Meanwhile, only 37 patients underwent revision due to lead migration over the same period. Therefore, the true rate of lead migration is more likely to be in the range of 10–15%. Rosenow et al., which examined 289 SCS patients who underwent a total 251 revision procedures, reports a comparable rate of lead migration (10%) [50]. Still, 9 of the 35 (26%) patients whose initial revision was due to lead migration required further revision. While 4 of these instances were due to unrelated adverse events or device defects, 5 of 9 (56%) underwent a second revision due to lead migration (Fig 2a,b). Given our concern for repeat lead migrations, we further analyzed these cases. Among the 5 patients with multiple instances of lead migration, all reported excellent relief from SCS stimulator trial and underwent device implantation. However, in each of these cases, patients reported partial or complete loss of benefit within the first three months, and radiography confirmed displacement of the electrode. Of note, 3 of the 5 instances were in cervical patients, with one case noted as having particularly challenging lead placement due to laminectomy defects from prior dorsal root entry zone myelotomy. Among the 2 thoracic patients, one required repeat revision after experiencing loss of coverage following a fall. Though our overall rate of revision due to lead migration is consistent with previous reports, we nonetheless seek to minimize this figure in future operations [50,51]. Proper anchoring of both the leads and the generator are critical to minimizing lead migration. As we see greater revision rates among cervical leads, consideration of shifting the battery to a subclavicular pocket—to minimize the distance between the electrode and battery—may decrease the opportunity for lead disruption. However, even with optimal anchoring and placement, we believe that a certain degree of lead migration is inevitable, as unforeseen falls may eclipse the strength of the underlying fascia to which anchoring devices are tethered. We will continue to monitor institutional rates of lead migration in order to promote long-term benefit from SCS therapy.

Limitations

There are several limitations of this study as a result of its design. This study is vulnerable to the typical limitations of a single-center retrospective analysis including sampling bias, insufficient follow-up, and missing data. In addition, we were likely underpowered to detect statistical significance in several potentially relevant variables described above. Of note, we did not focus on explantation rates or device therapeutic survival, although these are described elsewhere [52]. Duration of postoperative follow-up was not tracked in this study due to heterogeneous documentation of follow-up with patients’ primary pain clinician, though the median duration between initial revision and the conclusion of study observation was 43.0 months (IQR 25.−61.8) and did not vary significantly across groups (p=0.35). Furthermore, we did not report pain scores, as the outcome of interest was the advent of surgical revision and not the degree of pain relief. Consequently, our survival analysis overestimates true device or therapeutic survival, as treatment efficacy is predictably lost prior to the date of revision. Despite this, the percentage of initial SCS devices requiring revision are higher than one might expect (69%, 85%, and 89%, at years 1, 2, and 3, respectively). However, prior survival analysis from this institution which included both revision and non-revision SCS patients showed initial device failure rates of approximately 20%, 25%, and 35%, at 1, 2, and 3 years [52]. Finally, the analysis presented here only considers the subset of patients who required at least one revision. Furthermore, we elected to include all patients who underwent revision at this institution, regardless of the hospital at which the initial device implantation was performed, to reflect de facto clinical practice. Consequently, however, we are unable to make broader claims regarding the failure rates of SCS therapy among all patients who undergo initial SCS implantation, such as the overall rate of revision, and the overall failure rates of percutaneous vs paddle leads, or cervical vs thoracic leads. However, these data have been reported by many other expert surgical centers and meta-analyses [11,50,51]. Instead, we are able to demonstrate that, for a given patient requiring SCS revision, there is no significant difference in the likelihood of revision success based on lead type, and cervical lead revisions are more likely to fail than thoracic leads. Future prospective multicenter studies with larger sample sizes are warranted to better elucidate risk factors associated with revision surgery.

Conclusion

SCS remains a safe and effective procedure for patients with chronic pain. However, rates of complication and revision are persistently high, with one in five patients undergoing at least one revision procedure. In this retrospective analysis of 54 patients who underwent an initial SCS revision, cervical lead placement was associated with a higher risk for additional revision procedures. Lead migration was the most common indication for any revision procedure. Thus, consideration of lead positioning may result in lower revision rates and improved clinical outcomes among SCS patients with multiple revisions.

Data Availability Statement

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author [SK] upon reasonable request.