Article Text
PDF
Abstract
Background and objectives In spinal cord stimulation (SCS) therapy, electricity is the medication delivered to the spinal cord for pain relief. In contrast to conventional medication where dose is determined by desired therapeutic plasma concentration, there is lack of equivalent means of determining dose delivery in SCS. In open-loop (OL) SCS, due to the dynamic nature of the epidural space, the activating electric field delivered is inconsistent at the level of the dorsal columns. Recent Food and Drug Administration guidance suggests accurate and consistent therapy delivered using physiologic closed-loop control (PCLC) devices can minimize underdosage or overdosage and enhance medical care. PCLC-based evoked compound action potential (ECAP)-controlled technology provides the ability to prescribe a precise stimulation dose unique to each patient, continuously measure neural activation, and objectively inform SCS therapy optimization.
Methods Neurophysiological indicator metrics of therapy dose, usage above neural activation threshold, and accuracy of SCS therapy were assessed for relationship with pain reduction in over 600 SCS patients.
Results Significant relationships between objective metrics and pain relief across the patient population are shown, including first evidence for a dose-response relationship in SCS.
Conclusions Higher dose, more time over ECAP threshold, and higher accuracy are associated with better outcomes across patients. There is potential to optimize individual patient outcomes based on unique objective measurable electrophysiological inputs.
- neuromodulation
- spinal cord stimulation
- chronic pain
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, an indication of whether changes were made, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
There is a desire for more objective data guiding dosing guidelines for spinal cord stimulation (SCS).
A dose-response relationship has not been shown in SCS to date.
WHAT THIS STUDY ADDS
We present first evidence for a dose-response relationship in SCS.
We introduce three major objective metrics that can be used to optimize SCS therapy across and within patients.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE, OR POLICY
The metrics presented may be used to inform prescription dosing of SCS, to diagnose causes of suboptimal responses to SCS, and to optimize therapy for SCS patients.
Introduction
Spinal cord stimulators represent a significant advancement in the field of pain management. Developed in the 1960s, these implantable devices offer much-needed quality-of-life improvements to countless individuals suffering from chronic neuropathic pain conditions.1 2 By delivering electrical impulses to the spinal cord, spinal cord stimulation (SCS) modulates pain signals to provide relief from chronic pain.3 Despite alleviating pain for many patients, challenges remain when SCS fails to deliver or sustain sufficient pain relief. Traditional open-loop (OL) SCS technology, regardless of waveform (eg, tonic, burst, or high frequency), remains subjectively programmed, with no objective confirmation that the target nerves responsible for pain inhibition are activated consistently. Lack of consistent and adequate therapy may be responsible for loss of efficacy, the primary reason for explants (44%) with conventional OL SCS systems.4
The historical absence of ongoing objective measures of therapy delivery in SCS has significantly constrained the potential for improving outcomes. Without real-world, out-of-clinic measures, it is difficult to systematically compare the immediate physiological impacts of different interventions, identify the most effective stimulation parameters, and gauge the impact of therapy on pain relief and functional improvement.5 6
The evoked compound action potential (ECAP) is directly related to the number of dorsal column A-beta (Aβ) fibers activated by a stimulus pulse.6 7 The Evoke closed-loop (CL) spinal cord stimulator (Evoke System; Saluda Medical, Macquarie Park, Australia) incorporates physiological closed-loop control (PCLC) technology that measures the ECAP and adapts stimulation levels in real time to adjust for postural and neurological response differences, thereby allowing patients to receive consistent, comfortable therapy.8 9 Moreover, the ECAP-based CL system provides comprehensive objective data to optimize therapy and guide clinical practice.
We introduce three objective ECAP-based SCS therapy indicator metrics (objective neural metrics), explore their associations with therapeutic outcome, and demonstrate their utility in optimizing SCS therapy. These indicator metrics fall into categories of usage, dose, and variability. Device usage has long been measured but not correlated with outcome, signaling that device ‘on’ time does not accurately reflect therapeutic time. We hypothesize that time eliciting a neural response (ECAP) is more influential than time ‘on’. Dose is postulated as an important determinant of SCS success,6 but no dose-response relationship has been established for SCS.5 Using objective, out-of-clinic data from our entire patient population, we here provide first evidence for a dose-response relationship in SCS. Variability is controlled in CL therapy and uncontrolled in OL. A randomized, double-blind controlled trial showed conclusively that CL outperforms OL therapy.8–10 This is likely to be because high stimulation variability in OL is associated with brief periods of uncomfortable overstimulation, leading patients to reduce their therapy intensity and receive subtherapeutic understimulation in the long term.5 6 8
Use of these metrics in SCS mirrors conventional pharmacology. No therapeutic benefit could be reasonably expected in a patient treated with the correct drug but with a dose that is inadequate or highly variable. By the same token, we cannot expect therapeutic outcomes in SCS with inadequate or variable dosing. The metrics proposed here can help identify optimal parameters for patients, thereby guiding practitioners towards better prescriptions for device use across patients and allowing them to monitor therapy on a per-patient basis.
Methods
The patient cohort used in this study comprises patients receiving CL therapy across multiple Food and Drug Administration (FDA)-approved and ethics-approved clinical studies and commercial settings (n=690; January 2017 to June 2023; Evoke System; Saluda Medical, Macquarie Park, Australia). Data were collected in a real-world environment and retrospectively sourced from a global database. The breakdown of patient demographic associated with study and commercial subjects can be found in table 1.
Demographics of 690 unique subjects
The Evoke stimulator minimizes the difference between the measured ECAP amplitude and the target ECAP amplitude by automatically varying the stimulation current amplitude in real time.6 8 10 The system uses a physiologic closed-loop control system (PCLCS,11 per IEC 60601-1-10) to maintain a consistent neural response in which the average error between the prescribed ECAP amplitude target and the measured ECAP amplitude is zero. PCLCS adjustments for postural changes and neurological response changes are both triggered by a change in ECAP amplitude. The PCLCS responds to the summation of both sources of change, with the aim of maintaining consistent overall neural activation.
For all analyses in this paper, objective data were extracted from each subject’s clinic visit corresponding to their maximum percent pain relief at the time of archive (June 2023) (see online supplemental materials for details).
Supplemental material
Associations were investigated for percent usage over ECAP threshold, the dose ratio, and dose accuracy. For the purposes of SCS, the metric of dose is measured physiologically from the ECAP. Usage, traditionally reported as percent of total time that the stimulator is on, is inadequate alone as a variable relating to therapeutic outcome. Dose ratio is a standardized metric of dose obtained by taking the ratio of the median ECAP current over the ECAP threshold current and allows for comparisons of dose to be made across patients.12 The variability metric is defined as the out-of-clinic root mean squared error of the measured dose compared with the target dose. In OL therapy, this error is typically large, as deviations in stimulation delivery from target can occur with even minute physiological movements such as heartbeat and breathing.8 10 In CL therapy, the loop corrects for changes with movement at short latency which results in a small error, and large deviations may indicate poor loop tracking.
Definitions are provided in table 2 and examples are given in figure 1.
Terms and definitions of neural metrics
Supplemental material
Activation plot (AP) (left) and matching out-of-clinic histogram (right) data for two example visits. (A) A typical AP with usage above evoked compound action potential (ECAP) threshold, corresponding to a dose ratio >1 and (B) an example with dose ratio <1. Gray dots and bars in AP represent data means and SD per current. Orange curve overlaid on gray dots indicates the best-fit line for the AP. The green dashed line indicates the ECAP threshold. The blue solid line indicates the median ECAP amplitude. The corresponding dose ratio and percent usage over ECAP threshold are written above the plots.
Associations between outcomes and metrics were investigated for the total percent of time that stimulation was on, percent usage over ECAP threshold, dose ratio, and therapy accuracy. A priori, percent usage over ECAP threshold and dose ratio were assumed to have positive associations with percent pain reduction: more is better. Conversely, based on OL versus CL differences in neural activation and therapy outcomes in the EVOKE trial data, dose accuracy was predicted to have a negative relationship with percent pain reduction: more variability is worse.8–10 These relationships were tested by binning data points into four groups for each category and performing a Jonckheere-Terpstra one-tailed ranked test for trend.13
Additionally, case studies are presented to exemplify how clinicians may use the patient’s objective neural metric data to optimize therapy.
Results
Percent usage over threshold, dose ratio and dose accuracy—all measured out-of-clinic from patients’ at-home data—show a statistically significant relationship with percent pain reduction (p<0.01, p<0.01, and p<0.03, respectively (figure 2). As percent usage over ECAP threshold increased from 20% to 100%, percent pain reduction improved. As dose ratio increased from 0 to 1.6, percent pain reduction improved. Dose accuracy is higher as its associated numeric values of error in µV lessen. As the accuracy increases (corresponding to smaller error numbers), the percent pain reduction increases. For all three metrics, the majority of patients fell into groupings corresponding to better-than-median outcomes (usage over threshold >80%, dose ratio >1.2, and error < ±10µV). Percent time the stimulator was on did not show a statistically significant relationship with outcome (p>0.05, online supplemental figure A1).
Trends in objective neural indicator metrics individually, expressed as metric groupings by percent difference from the median maximal analgesic effect (MAE) percent pain reduction (79%, 80%, and 82% left to right). (A) Usage is positively associated with percent pain reduction. Number of patients in each grouping, left to right: 42, 48, 79, 380. (B) Dose ratio is positively associated with percent pain reduction. Number of patients in each grouping, left to right: 95, 121, 188, 138. (C) Dose accuracy is positively associated with percent pain reduction within the population of patients whose median dose is greater than threshold. Number of patients in each grouping, left to right: 23, 51, 189, 220. Bar heights represent median values. Black bars show SE. Lower subplots show the percentage of patients included in each grouping. ECAP, evoked compound action potential.
Objective neural metrics and patient-reported outcomes were analyzed at patients’ maximal analgesic effect timepoint. To test for the dependence of any metric on the timing of the MAE visit chosen, a Spearman’s rank correlation was performed for each of the metrics in comparison with the days since implant for that visit, with no significant trends identified (p>0.05 for all).
To better understand the relationships between the three objective neural metrics, we took the rank correlation of each pair. The relationships between metrics and distributions of the metric values are shown in figure 3. The Spearman’s r2 value for dose ratio versus percent time above ECAP threshold is 0.81 (p<0.001, figure 3A), for dose ratio versus dose accuracy is 0.16 (p<0.001, figure 3B), and for percent time above ECAP threshold versus dose accuracy is 0.07 (p<0.001, figure 3C).
Scatter plots with associated distribution plots of usage over threshold, dose ratio, and dose accuracy. ECAP, evoked compound action potential.
To illustrate how these three categories of indicator metrics may be useful to guide SCS therapy, we include two retrospective case studies from real patients (figure 4) in which the indicator metrics may have been used to alert the care team to potential targets for intervention in the visits in which the patient was reporting poor pain relief (<50%, red text). Changes to therapy were associated with better outcomes at later timepoints. For case 1, pain relief was low at 1-month post implant, when the patient had an effective dose ratio <1 and corresponding percent utilization over threshold <50%, indicating that they were running below threshold a majority of the time. At later visits, they achieved total pain relief with higher dose and associated higher percent utilization over threshold. In case 2, the low pain relief was associated with suboptimal PCLCS programming, resulting in a loop that was tracking irregularly out of clinic. This device was reprogrammed to deliver more stable neural activation at the 1-month visit, and the patient achieved a high level of pain relief thereafter.
Case studies from real patients with changing pain scores over time. Reported pain relief text is red if <50%, green if ≥50%. The circles in the Neural Metrics section are colored gray if they correspond to pain relief below median, green otherwise. A dose accuracy of ±0 µV relays the error with posture change is equal or less than that of the recording environment at baseline. ECAP, evoked compound action potential.
Discussion
The objective metrics of therapy dose, usage over ECAP threshold, and dose accuracy are inherently sensible from first principles as crucial measurable indicators to evaluate SCS utilization and performance. Patient outcomes improve with higher percent usage over ECAP threshold, higher dose ratio, and higher accuracy (figure 2).
Usage metrics capture time the device is on and usage above ECAP threshold, highlighting the patient’s attempted and practical adherence to treatment. Neural dose ratio provides a normalized assessment of the stimulation dose. Therapy accuracy identifies outlying cases in which PCLCS operation is suboptimal, leading to high variation in neural activation (administered dose) compared with the target; in these cases, reprogramming is warranted. By considering these metrics collectively, healthcare providers gain a comprehensive overview of the patient’s device usage and can identify modifiable factors contributing to inadequate pain relief. This enables targeted interventions to address specific challenges and improve overall pain management.
To our knowledge, the observed relationship between dose ratio and pain reduction across our SCS patient population is the first evidence for a dose-response relationship in SCS for chronic pain. Using the dose ratio as opposed to the raw ECAP dose (µV) measurement allows fair comparison of the amount of activation above threshold across patients and programs, contributing to our ability to observe this relationship.12 Our finding that pain relief increases with higher dose ratios over threshold suggests a dose-response relationship above and beyond the simple delineation of whether dorsal column activation is occurring.
Relationships between dosing levels and patient outcomes may be used to inform a prescriptive level of dose. Dose ratios <1, corresponding to <50% usage over ECAP threshold, provide the least pain relief, likely due to too few consistently stimulated Aβ fibers. A dose ratio of 1.2 and above is likely to provide more pain relief than peri-threshold or subthreshold stimulation levels. Moreover, the dose ratio of 1.2 corresponds with the inflection point past which the usage over threshold levels off to 100% (figure 3). Theoretically, there should be a level above which the patient is overstimulated, and the side effects will outweigh the benefits of higher dosing.6 SCS patients in this study achieved high dose ratio levels while avoiding transient overstimulation events by nature of the CL therapy, which prevents overstimulation events even at high-dose levels. In comparison with OL therapy, CL therapy has improved the three aforementioned metrics greatly: CL patients have been shown to run their device at a higher dose, generally corresponding to higher usage over ECAP threshold (figure 3A), and their dosing is more accurate.10
Without continuous monitoring of a physiological response to stimulation, there is no guarantee that device ‘on’ time reflects the time the device is therapeutic. The ECAP is the result of activation of Aβ fibers, and its absence is indicative of no activation of those fibers.14 We show that the percent of time evoking ECAPs, rather than simply the percent of time stimulating, scales with pain reduction. In cases for which stimulation was subthreshold most of the time, the percent time over ECAP threshold is low even while the total percent time stimulating can be nearly 100%. The percent time stimulating was not associated with outcome (see online supplemental materials for non-significant result), whereas the percent time stimulating over threshold (therapeutic stimulation) was associated positively with outcome. Higher percent pain reduction with higher activation over ECAP threshold indicates that measurable Aβ fiber activation is important for optimal pain relief.
The variability metric serves as a valuable tool for pinpointing outlying data points that require optimization within a PCLCS. This could be secondary to poor loop tracking—for example, due to settings that cause the device to respond too slowly to daily postural changes—or related to other factors such as lead migration. Because this risk factor appears in <15% of subjects, its overall effectiveness as a predictor of outcome is expected to be limited. It is included in the objective neural metrics because the fix can be relatively easy and important: namely, the patient should be reprogrammed with an optimized loop. In the future, it is possible for advances in programming automation to optimize parameters and ensure high CL accuracy.
Recent FDA guidance suggests that PCLC device manufacturers should characterize and consider disturbances affecting patients’ response to the intervention, including interpatient variability in physiological response to stimulation and changes in this response over time due to disease progression or changes in patient sensitivity to stimulation.11 Current solutions for OL therapy optimization include remote monitoring15 and artificial intelligence-based16 17 program changes and are based on SCS device utilization and subjective patient-reported outcomes, respectively; however, these non-PCLCSs cannot objectively evaluate and optimize dosing, automatically account for changes in physiological response to SCS over time, or objectively adjust for interpatient variability in response. The objective neural metrics are consistent with the FDA guidance for PCLCSs for objective therapy optimization and for guiding clinical practice. Explicitly, physiological response to stimulation is characterized in terms of percent time over ECAP threshold; interpatient physiological response variability11 and changes in physiological response over time are characterized using the dose ratio; and variability in activation is characterized using the therapy variability metric.
The objective neural metrics are related, but each remains independent in its meaning. While dose ratio and percent utilization over ECAP threshold are highly correlated, this is only apparent for dose ratios below approximately 1.2, above which the utilization over threshold saturates. The continued trend of higher pain relief with higher dose, even when usage over threshold is near 100% for all subjects, further affirms the benefits of higher dose. The relationship between dose and variability is expected; higher targets invite higher recorded magnitude of deviation from the target due to the growth in magnitude of the ECAP levels recorded.
The objective neural indicator metrics can identify correctable causes of suboptimal pain relief on a per-patient basis. In case examples (figure 4), we demonstrate that by employing the three fundamental metrics of dose, usage, and variability, we can identify underlying factors contributing to heightened pain levels. Through a targeted evaluation, clinicians can identify correctable failure mechanisms responsible for suboptimal pain relief. This approach allows for tailored interventions and personalized pain management strategies. The chosen three metrics are readily understandable and modifiable, and therefore serve as active dials that may be tuned for each patient to achieve optimal therapy. As we learn more about factors that influence outcomes, we may expand the objective neural metrics to include other markers.
Limitations of this study include the retrospective nature of the work, especially with respect to the available data for each metric. Patients were not assigned different dose ratios, percent utilization categories, or variability metrics; rather, data collected at regular visits was analyzed for trends post hoc, taking advantage of natural variability in device utilization. Additional prospective studies would further characterize these relationships. Other caveats include that the data points were taken at different timepoints for each patient, by nature of selecting the maximum pain score visit. Some patients have only been followed for months whereas some have data representing years of device usage. However, the timepoint at which the maximum effect was observed was not significantly correlated with any of the metrics explored, indicating that the trends in this dataset are robust across the extended timeframes that patients obtain pain relief with CL SCS.10 18
With the introduction of objective neural metrics, one can envision a future where clinical decision-making is directly steered by objective, data-driven insights. If a patient returns to clinic with all metrics within expected ranges and complaining of increased pain, this could suggest new pain or lead migration affecting coverage of the therapeutic target. Rather than immediately resorting to reprogramming as is often done today, the clinical team can instead directly explore new causes for the increased pain. In contrast, many patients today are lost to regular clinical follow-up. Consider the ability to monitor objective neural metrics and proactively signal the care team to check in and provide guidance when metrics do not meet expectations for therapeutic levels. Prescribing usage levels will likely enhance pain relief for the population when used alongside individual evaluation, as it does in pharmacology. This new era of SCS monitoring holds great potential for improving therapy in the chronic pain population.
Considering the future of medical research and the evaluation of device effectiveness, it may be beneficial to incorporate a broad range of metrics that encompass both patient characteristics and device usage patterns. By expanding our repertoire to include metrics inherent to the patient, such as neurophysiological measures, pain type and history, and patient demographics, we may gain a more comprehensive understanding of what determines responsiveness to SCS. Additionally, as global adoption of SCS systems with ECAP measurement capabilities increases, we expect the increased volume of data will strengthen the results presented. Inclusion of additional patients enhances the statistical power of analyses and accounts for a greater diversity of demographics, comorbidities, and usage patterns, leading to more robust findings to better guide clinical decision-making and the development of future technologies.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by EVOKE (NCT02924129), ECAP (NCT04319887), AVALON (ACTRN12615000713594), BRIGHTON (ACTRN12618001808235), DR (DR-UK: ISRCTN27710516; DR-NL: NL7889; DR-DE: NCT05272137), DURABILITY (NCT04627974), FRESHWATER (NCT04662905). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
The authors thank Weirong Gei, Abeer Khurram, Angela Leitner, Dave Mugan, Aileen Ouyang, Daniel Parker, Peter Single, and Martin Wong for their thoughtful contributions and technical input affecting this work.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Presented at Interim data from this work were presented at the 2024 North American Neuromodulation Society Meeting in Las Vegas, January 18–21, 2024.
Contributors All authors contributed meaningfully to this manuscript. LM is guarantor of this work.
Funding Funded by Saluda Medical.
Competing interests JP reports research and consulting fees from Saluda Medical during the conduct of the study; consultancy for Abbott, Medtronic, Saluda Medical, Flowonix, SpineThera, Vertos, Vertiflex, SPR Therapeutics, Tersera, Aurora, Spark, Ethos, Biotronik, Mainstay, WISE, Boston Scientific, and Thermaquil outside the submitted work; has received grant and research support from: Abbott, Flowonix, Aurora, Painteq, Ethos, Muse, Boston Scientific, SPR Therapeutics, Mainstay, Vertos, AIS, and Thermaquil outside the submitted work; and is a shareholder of Vertos, SPR Therapeutics, Painteq, Aurora, Spark, Celeri Health, Neural Integrative Solutions, Pacific Research Institute, Thermaquil, and Anesthetic Gas Reclamation. EP has received research support from Mainstay, Medtronic, Neuros Medical, Nevro Corporation, ReNeuron, SPR, and Saluda Medical outside the submitted work, as well as personal fees from Abbott Neuromodulation, Biotronik, Medtronic Neuromodulation, Nalu, Neuros Medical, Nevro, Presidio Medical, Saluda Medical, and Vertos outside the submitted work. She holds stock options from SynerFuse and neuro42. PV is a consultant at Saluda Medical, NaluMedical, Abbott, and Biotronik. JWK is on the advisory board for Abbott Laboratories, Nevro Corporation, Saluda Medical, Medtronic and Boston Scientific. LM, IG and DMK are employed by Saluda Medical. There are no other relationships that might lead to a conflict of interest in the current study.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.