Safety

When choosing a therapy for chronic pain, as physicians we strive to limit the potential for harm to our patients. So there is an impetus to choose medications or interventions which carry risks that are acceptably low to the physician and patient. When viewed in isolation, neuromodulation therapies carry a significant risk of bleeding, infection, or nerve injury, which traditionally placed them as a treatment of last resort ( ).

In 2004, in a review compiling 20 years of data (approximately 2700 patients) involving complications related to SCS, Cameron divided complications into technical (relating to hardware), biologic (effects on host tissue), or “other” ( ). She reported a total technical complication rate of 27.2%, of which 82% were related to lead issues (hardware) including lead migration (13.2%), fracture (9.1%), or hardware malfunction (2.9%). Biologic complications, however, only totaled 4% of overall SCS complications, and included infection (3.4%), cerebrospinal fluid leak (0.3%), and hematoma (0.3%). Other complications such as undesired stimulation amounted to 2.4%, but this number is likely to diminish with the advent of newer paresthesia-free and subthreshold stimulation constructs. Based on these early observations regarding complications of SCS, Kumar et al. proposed practical expert guidelines that would minimize such complications, such as proper lead insertion, positioning, and anchoring techniques ( ). Since these reviews by Cameron and Kumar et al., a decade of innovation and experience has elapsed. A more recent retrospective analysis of 13,774 patients over 10 years reported a much lower total complication rate of 2.8% in the first 90 days, with a significantly higher complication rate (mostly involving wound, pulmonary, or renal issues) in patients receiving surgically placed paddle leads (3.4%) when compared to percutaneous leads (2.2%) ( ). At 2 years, however, patients with percutaneous leads had a significantly higher reoperation rate than those with paddle leads (6.3% vs. 3.5%). It should be noted that in neuromodulation implants, as in any evolving therapy, there is likely significant variation in complication rates among institutions and providers, and we encourage all physicians and institutions to keep track of their personal complication rates to aid in assessing the real safety component of this therapy.

To have practical use in an algorithm, the complication rates cited above must be interpreted in comparison to complications posed by alternative and competing therapies. While complications from neuromodulation therapies typically occur in the first 2–3 months postoperatively, it is important to remember that complications from medications are often delayed and may occur months to years after the start date of the therapy. It should be noted that while short-term use of nonsteroidal antiinflammatory drugs (NSAIDs) is often deemed “low risk,” there is significant long-term morbidity with their chronic use that includes a 17%–31% risk of gastric ulcer formation, leading to 16,500 deaths and more than 100,000 hospitalizations annually in the United States ( ). Similarly, chronic use of opiates over the same time course of SCS use has been associated with significant endocrinopathy ( ), including low testosterone ( ), osteoporosis ( ), and executive cognitive dysfunction ( ). Furthermore, from 1999 to 2014 there have been more than 165,000 deaths related to opiate use in the United States alone ( ).

Complications from commonly used interventional pain procedures range from asymptomatic to life threatening. A single-center prospective analysis over 20 months, combining over 10,000 patients, estimated an overall complication rate of approximately 7% for patients receiving epidural steroid injections (ESI, transforaminal or interlaminar) or percutaneous adhesiolysis ( ). In this study the highest risk for intravascular injection and its potential serious sequellae, including death, was seen for adhesiolysis (11.6%), followed by lumbar transforaminal injections (7.9%) and then cervical interlaminar approaches (4.2%). A more recent metaanalysis suggests that interoperator variability is high, with complication rates for interlaminar and transforminal ESI ranging from 0.4% to 16%, combining complications that include neck pain, back pain, dural puncture, epidural hematoma, and infection, depending on the specific study group and operators involved ( ). Severe complications such as quadriplegia or cardiac arrest after cervical ESI (from presumed intravascular injection) have been reported in isolated cases ( ).

When evaluating the safety and relative risks of neuromodulation procedures for placement into an algorithm of care, we must compare the safety and relative risks of neuromodulation to the risks of alternative therapies such as medications (NSAIDs, opiates), nonsurgical interventions (ESI, adhesolysis, etc.), and surgery (back and neck surgery) over a similar time course of the intervention to build a full picture of the actual safety profile of neuromodulation when compared to other therapies. Increased experience and multidisciplinary dissemination of knowledge of neuromodulation technologies will naturally occur over time, but active engagement with these efforts will help to make implantable devices safer.

Appropriateness

The principle of appropriateness as an evaluation tool refers to the selection of an appropriate therapy or therapies and matching that therapy/therapies to an appropriate particular patient or particular population of patients for inclusion into a medical algorithm for the treatment of chronic pain or other disorders or diseases. While it is paramount to establish the correct diagnosis and indication for a therapy (e.g., neuropathic pain vs. ischemic pain when considering neuromodulation interventions), it is equally important to ensure the absence of medical or psychosocial contraindications for that intervention. Nonappropriate intervention, such as using chronic systemic opioids in a population of addicted patients, using NSAIDS in patients who have peptic ulcer disease, or implanting a neuromodulation device into someone who does not understand how to use the device or is paranoid, is not advised and should be proscribed.

As another example of appropriateness, most insurance companies require a psychological screening prior to implantation of a neuromodulation device because there is substantial evidence that preimplantation psychological factors might serve as negative outcome predictors for neuromodulation therapies. In a retrospective analysis of 83 patients who underwent SCS from 2005 to 2013, Bendinger et al. used univariate regressions and receiver operator characteristic analyses to demonstrate that poorer outcomes (decrease in pain score by <50%) were associated with higher presurgical depression, anxiety, sleep interference, and higher scores on the Pain Catastrophizing Scale ( ). Similarly, a prospective cohort study by Block et al. showed that poorer SCS outcomes, as measured by the Oswestry Disability Index, and poorer patient satisfaction at 5 months were correlated with preimplant elevations in the Minnesota Multiphasic Personality Inventory 2 (MMPI 2-RF), which measures emotional dysfunction, somatic/cognitive complaints, and interpersonal problems ( ). A rich literature compiled over the last 2 decades supports an increased risk of poor outcome of SCS from presurgical depression, somatization, demoralization ( ), tobacco use, and substance abuse, but a review by Fama et al. supports the idea that depressed patients may be made more appropriate for SCS therapy by treatment with antidepressants prior to surgery ( ). From a practical standpoint, it is important to establish that a patient is cognitively capable of operating the technology related to implantable devices, such as a wireless transmitter (or recharger) needed to operate the device.

The question of appropriateness overlaps somewhat with questions of efficacy when considering how early in a patient’s disease course the therapy is offered. While neuromodulation therapies have often been relegated to the end of the treatment continuum, there is substantial evidence that offering modalities such as SCS earlier is not only appropriate but has a higher success rate than waiting until other treatments have been trialed and the patient’s disease has progressed. In a retrospective analysis of 437 patients, Kumar et al. examined delays to accessing SCS and the impact of wait times on success of therapy ( ). These authors stated that “the long term success rate of SCS is inversely proportional to the time interval between the beginning of the chronic pain syndrome and the implantation time.” In particular, the inflection point at which the chance of SCS failure exceeds the chance of SCS success for pain control is approximately 5 years, suggesting that early implantation is most appropriate for chronic pain, and certainly before 5 years from symptom onset.

Recent advances in neurostimulation technology, targets, waveforms, and pulse-generator batteries impose a second layer of consideration. After deciding if neuromodulation is an appropriate therapy for a patient, it is crucial to decide which specific neuromodulation therapy or construct is most appropriate for the patient. At the time of this writing there are six manufacturers offering neurostimulation therapies through a host of different pulse generators, with primary cells versus rechargeable batteries and leads, which provide traditional paresthesia-based (conventional SCS) and newer paresthesia-free systems (burst SCS and KHz-frequency SCS) that target neural structures from the dorsal columns, dorsal root ganglion, and peripheral nerves. In addition, patients and physicians treating pain from progressive diseases such as arthritis or multiple sclerosis may value the option of magnetic resonance imaging (MRI) conditionality when determining the appropriateness of a device, and several manufacturers offer MRI conditionality for their devices. Choosing which therapy/device is appropriate for a particular patient’s diagnosis is paramount to increasing chances of success during a trial and conversion to permanent implant.

The Neuromodulation Appropriateness Consensus Committee of the International Neuromodulation Society has recently reviewed current literature, clinical experience, and expert opinion to determine the utility of various modes and targets of SCS to treat chronic pain, chronic critical limb ischemia, angina, and craniofacial pain ( ). While it is beyond the scope of this chapter to review the details of specific technologies and the evidence for their listed indications, we direct the reader to these articles to identify diagnoses, stimulation parameters, targets, contraindications, and the specifics of implantable constructs relevant to specific diagnoses.

Time to Fiscal Neutrality

Healthcare expenditure in the United States is higher than in most other countries and has grown at a faster pace from 1980 to 2013, almost doubling when expressed as a percentage of gross domestic product (from approximately 8% to 16%). Much of this cost is attributed to spending on newer medical technologies or technological innovation, which often carries higher up-front costs when compared to the up-front costs of conservative medical management (CMM). Immediate or up-front cost of therapy, however, fails to capture the actual dollar amount spent or saved over the lifetime of a given therapy when compared to another. While neuromodulation therapies for pain treatment may initially cost more, they can actually lead to significant cost savings over years of use because of the inherent reduced expenditures on potential medications, physician visits, and other nonsurgical interventions or surgeries. For this reason a robust medical algorithm for chronic pain care must employ the concept of “time to fiscal neutrality,” defined as the cost of a therapy over the lifetime of the therapy when compared to the cost of a comparator over the same time frame and appropriate for a particular disease state. A medical therapy is said to achieve fiscal neutrality when the initial cost of the therapy is neutralized by the cost savings of that therapy when compared to the comparator over time.

In the most simple cases, the up-front cost for a neuromodulation device may be less than for a comparator therapy. North et al. compared the cost of SCS implantation versus repeat back surgery for the treatment of failed back surgery syndrome (FBSS) in a randomized, controlled, cross-over trial, and demonstrated that SCS therapy (mean cost per patient $31,530) was more cost effective on day 1 than repeat operation (mean $38,160) in all analyses presented ( ). In addition to the up-front cost for a neuromodulation implant, we must consider the cost of maintenance of these devices and correction for potential complications such as lead migration, fracture, or battery failure over time (as discussed above).

Kumar et al. have published numerous analyses supporting the cost effectiveness of SCS over months to years compared to alternative therapies. A case-control series in Canada evaluating the mean 5-year cost of SCS in 104 patients demonstrated a cost of $29,123 per SCS patient when compared to CMM cost (including chronic medication management, diagnostic imaging, interventions, nursing, physical therapy, massage, and hospitalization for breakthrough pain) of $38,029 per patient ( ). In this analysis, the cost of SCS implantation and maintenance was greater than CMM for the first 2.5 years (time to fiscal neutrality), after which SCS resulted in cost savings. A similar latency (1–3 years) to fiscal neutrality of SCS implant versus CMM was observed by Taylor and Taylor, largely due to reduced healthcare resource use with SCS, which likely extrapolates to the lifetime of a patient ( ). In the Kumar study of 2002, 88% of SCS patients were able to return to work when compared to 0% for the CMM patients, further contributing to unmeasured cost benefits from SCS. Similar results were found in a prospective study conducted in the United Kingdom and Canada ( ).

When considering IDDS, simulated assessments of cost effectiveness were similar to that of SCS when compared to CMM in multiple studies ( ) for intrathecal opiates ( ) and intrathecal baclofen. Latencies to fiscal neutrality for IDDS when compared to CMM are comparable to SCS at a range of 11–24 months.

Efficacy

In any algorithm designed to choose between medical therapies for chronic pain, or for any disorder or disease, for that matter, a dominant perspective is to offset perceived risks with potential benefits. Such benefits are usually best described in terms of efficacy; the likelihood of desired outcomes for a patient, such as reduction of perceived pain, decreased disability and reliance on medication, and increased occupational and psychosocial function. Importantly, the efficacy of any particular treatment can only be assessed in the context of the specific diagnosis that is being treated to target a putative pain-generating mechanism. While older studies in neuromodulation relied on retrospective analyses to evaluate the efficacy of therapies, newer and more robust studies of neuromodulation rely on prospective randomized controlled clinical trials (RCTs), which remain the gold standard for efficacy/safety trials.

As mentioned above, the Neuromodulation Appropriateness Consensus Committee of the International Neuromodulation Society has published comprehensive up-to-date reviews that comment on evidence for efficacy of neuromodulation therapies for a host of disorders. However, most high-quality RCTs in neuromodulation to date have been published for the indication of FBSS and neuropathic pain syndromes, comparing implantable therapies to CMM or alternative invasive procedures.

In North et al. conducted a prospective RCT in patients with FBSS, comparing the relative efficacy of traditional dorsal column SCS versus reoperation, using desire to cross over to the alternative therapy as the primary endpoint. For the first 27 patients reaching the 6-month cross-over point, SCS showed a statistically significant benefit over reoperation. A more recent RCT by that group ( ), comparing SCS versus reoperation in 42 patients using a 3-year time point, confirmed their original results, showing that 92.9% of SCS patients reported improvement in pain control and usual daily activities, with 73.7% continuing the use of their stimulator. Further, patients undergoing reoperation required significantly increased opiate analgesics for pain control ( ).

In a systematic review and metaanalysis, Taylor and Taylor showed that SCS in patients suffering from refractory neuropathic back and leg pain in the setting of FBSS resulted in significantly improved quality of life (QoL), reduced analgesic consumption, reduced pain, and increased likelihood of return to work ( ). When comparing SCS and CMM versus CMM alone for FBSS in a 6-month prospective RCT (PROCESS study) in 100 patients, Kumar et al. showed that SCS provided superior pain relief and functional improvement (as measured by the Oswestry Disability Index) ( ), which was later extended to a 2-year follow-up.

A common comparison in neuromodulation-related trials is to divide patients into those who receive >50% pain relief from a particular treatment (responders) and those who do not (nonresponders). In the PROCESS study, 47% of SCS patients with refractory leg pain were responders at 2 years as compared to only 7% in the CMM group ( ).

The literature supports an efficacy for SCS systems that has been dubbed “the 50/50 effect”: approximately 50% of patient obtain >50% of pain relief when all devices are considered. However, the latest evidence supporting the efficacy of SCS for axial low back and radicular leg pain comes from a direct head-to-head comparison of a newer paresthesia-free high-frequency 10 kHz SCS system (HF10, Nevro Inc.) with a conventional/traditional paresthesia-based SCS system (Boston Scientific) in the multicenter SENZA trial that included 198 patients ( ). At 3 months, 84.5% of HF10 patients were responders for back pain and 83.1% for leg pain when compared to 43.8% for back pain and 55.5% for leg pain in the traditional SCS group, respectively. These benefits were sustained at 2-year follow-up for both axial back pain (76.5% HF10 patients responding versus 49.3% traditional SCS responding) and leg pain (72.4% HF10 responders vs. 49.3% traditional SCS responders) ( ).

When the individual characteristics for “responders” are evaluated further and parsed out, there is a clear improvement in psychosocial and functional outcomes in SCS for neuropathic pain. A large prospective trial of SCS recently demonstrated improvement in QoL scales, depression, and catastrophizing among 401 subjects of whom 75% were “responders,” with 85% of patients reporting satisfaction with pain control and 73.3% improvement in QoL at 12 months. Small RCTs also bear out the efficacy of occipital nerve stimulators to treat refractory migraine, with a mean of 8.51 reduced headache days per month ( ), and the benefits of SCS to treat refractory angina in exercise duration, angina frequency, and nitrate consumption ( ). While numerous companies have reported pivotal trials in therapies such as dorsal root ganglion stimulation and novel waveforms such as burst protocols, peer-reviewed reporting of RCTs is currently pending ( ).