Neuromodulation interventions use electrical stimulation to modulate neuronal activity and elicit a therapeutic response.19 These procedures are generally adjustable, reversible, and have demonstrated some efficacy in the treatment of several conditions, including chronic pain.26 Despite promising evidence in the field, the general quality of the evidence is weak or moderate.15 Improvements after neuromodulation therapy are typically attributed to the neuromodulation itself, but as the exact mechanisms responsible for improvement are unknown32 and as appropriate control conditions are lacking,21 more careful investigations of the mechanisms underlying these apparent effects are warranted. The advancement of neuromodulation techniques and the advent of new ways of controlling and evaluating neuromodulation trials will potentially help to advance our understanding of the mechanisms underlying neuromodulation.
Placebo controls are used to distinguish the effects of an intervention, i.e., effects specifically related to the active properties of the treatment from effects attributable to the treatment context, such as the ritual of the treatment, the invasiveness of the procedure, the doctor–patient relationship, and the patients’ expectations towards the treatment. An adequate placebo control ensures that the active condition and the placebo condition are indistinguishable, the only difference being the presence of the supposed active component of the treatment, thereby simulating the treatment and its context without delivering the actual treatment. This approach allows for a double-blind administration of the intervention, so that neither the patients nor the health care providers know which treatment is given. Without an adequate placebo control and blinding, it is not possible to determine whether the apparent effect is due to the supposed active treatment component or contextual factors such as high expectations toward treatment.
Strong placebo analgesic effects have repeatedly been demonstrated for pharmacological treatments.7,9,12,45 Factors through which placebo effects may influence treatment effects include enthusiasm about the new treatment, doctor–patient interactions, increased expectations of treatment effects, and decreased negative emotions such as anxiety,45 and these effects may increase with the invasiveness of the procedure.38 Recent meta-analyses of placebo-controlled trials of interventional procedures involving the introduction of medical devices inserted into the body demonstrated that the magnitude of change in the placebo arm is large and longlasting, and it accounts for about 80% of the improvement in the active arm, especially in studies with subjective outcomes such as pain.23,25,58,59
In the field of surgery and neuromodulation, placebo controls are sometimes referred to as sham interventions, but as both concepts involve interventions that simulate the treatment context without delivering actual treatment, they are conceptually similar. There is an ongoing debate on appropriate methods to control and evaluate treatment efficacy and when and how the placebo control should be implemented.21 First, it may be challenging to create appropriate placebo control conditions, which imitate all elements of treatment except for the active component so that blinding is possible.57 Moreover, an objection that is sometimes posited is that the inclusion of placebo controls in surgical trials poses ethical problems because it involves risks for the patient or may not offer the best possible treatment.36 However, not validating interventional procedures in placebo-controlled trials may also pose ethical concerns because patients might undergo invasive and risky treatments without proven efficacy.21,56
Like most medical procedures, neuromodulation involves risks, and it is, therefore, essential to demonstrate that its true efficacy outweighs potential harms. Furthermore, there are relatively high financial costs associated with the neuromodulation device and implantation surgery as well as battery replacement, and any complications or hardware malfunctions that may require additional surgery.34 For a new pharmacological treatment to be approved, it has to produce beneficial effects that outweigh placebo effects, biases, and fluctuations in symptoms; however, these rigorous requirements do not yet apply to treatments involving neuromodulation. Before widening the scope of neuromodulation therapy further, it is critical to evaluate treatment effects adequately.
In this article, we review the current evidence for neuromodulation therapy in chronic pain and reflect upon the challenges this field is facing when securing high-quality evidence.
2. Neuromodulation in pain: current evidence
Analgesic effects of neuromodulation have been demonstrated in patients experiencing chronic pain caused by conditions such as osteoarthritis, back problems, or Parkinson disease.11,18,20,42,47 Procedures of neuromodulation proposed for targeting pain involve both invasive procedures such as deep brain stimulation, spinal cord stimulation, peripheral nerve stimulation, nerve root stimulation, and epidural motor cortex stimulation, and noninvasive procedures such as transcutaneous electrical nerve stimulation, and repetitive transcranial magnetic stimulation.14 Spinal cord stimulation is an example of neuromodulation for chronic pain, mainly for the treatment of peripheral vascular disease, complex regional pain syndrome, and postsurgical chronic back and leg pain.15 In this article, the spinal cord stimulation literature will be used as an example to demonstrate some of the challenges in the field of neuromodulation regarding the inclusion of placebo controls.
Spinal cord stimulation is widely used for analgesia in research and clinical practice. Inspired by the gate control theory of pain,39 the objective of spinal cord stimulation therapy is to inhibit nociceptive transmission in the dorsal horn of the spinal cord. Traditional spinal cord stimulation (<1000 Hz) entails implanting an electrode, which delivers electrical impulses, in the epidural space over the dorsal column structure of the spine.16 Such stimulation often evokes paraesthesiae, an abnormal painless sensation in the target innervation.24 Advancements in neuromodulation technology have allowed for the emergence of high-frequency spinal cord stimulation, which stimulates at 10,000 Hz, and at this frequency, pain relief is obtained without paraesthesiae.17 It has been established that patients with chronic pain probably benefit from traditional low-frequency spinal cord stimulation as well as from high-frequency spinal cord stimulation.3,6,31
Previous research has drawn attention to the quality of evidence underlying neuromodulation, highlighting the need for further preclinical studies elucidating mechanisms, sufficient blinding and controls, and the consideration of placebo effects.13,22 As a Cochrane review on the topic has been withdrawn due to being out of date,37 there is currently a lack of recent reviews on the evidence of spinal cord stimulation for pain. Systematic reviews show that the majority of current evidence of spinal cord stimulation relies on small-scale studies, case studies, or prospective studies with a lack of randomized controlled trials.43,55 Randomized controlled trials in the field often compare different modalities of spinal cord stimulation therapy, eg, various frequencies of stimulation. This is illustrated in a sample identified through PubMed for the purpose of this article (the latest 200 out of 2035 results searching for “spinal cord stimulation” AND “pain,” with no restrictions regarding the year of publication, identified the sixth of May 2019). In this sample, 17 studies included control groups. Fourteen of these used another spinal cord stimulation therapy (eg, other frequencies of stimulation) as a control group, 2 studies compared results to a healthy control group,27,28 and one to medical treatment.49 These control groups may not be sufficient to evaluate treatment efficacy. Medical treatment, or standard care, reflects the course of symptoms without spinal cord stimulation, but cannot control for the placebo effects associated with an invasive technique. Healthy controls are also not an adequate control group, because they do not demonstrate these fluctuations in symptoms. Other types of spinal cord stimulation may act as a control; however, because the working mechanisms underlying this therapy are unclear, comparing 2 neuromodulation therapies may not be sufficient to demonstrate that they are truly effective and placebo-controlled trials will be needed.
In the existing spinal cord stimulation literature, there is a small number of neuromodulation studies that have included placebo procedures, but they have used various stimulation frequencies and produced mixed results regarding the efficacy of active stimulation over placebo.2,33,44,52 In the following examples of placebo-controlled trials, placebo conditions entailed programming the same settings in the spinal cord stimulation device as the active stimulation, but with no amplitude in the placebo treatment, meaning that no electrical impulses were delivered. A small case series study, which included only 4 subjects, found beneficial effects of spinal cord stimulation in comparison with placebo.52 In this study, subjects were assigned to placebo spinal cord stimulation and active high-density spinal cord stimulation (1200 Hz), which involves paraesthesiae-free stimulation through increased frequency and pulse width and reduced amplitude compared with traditional stimulation. While all subjects achieved pain relief as a result of active spinal cord stimulation, one subject also responded well to placebo stimulation. Larger studies have also produced mixed results regarding the efficacy of active stimulation and the most beneficial stimulation frequency. In a double-blind randomized two-period crossover study in 33 patients with chronic back pain, a similar proportion (N = 14/33) of patients responded to high-frequency stimulation (5000 Hz) and placebo condition without stimulation (N = 10/33).44 In a study, placebo stimulation was compared to varying frequencies of stimulation (1200, 3030, and 5882 Hz).2 Although stimulation at 5882 Hz produced statistically significant pain relief compared with both the lower frequencies and placebo stimulation, the lower frequencies of stimulation did not produce any effect beyond placebo.2 By contrast, another trial showed no difference in pain reduction across 4 frequencies of stimulation (40, 500, 1200 Hz, and burst spinal cord stimulation), but found that all spinal cord stimulation frequencies outperformed placebo.33 Thus, although stimulation at 1200 Hz produced analgesic effects beyond placebo in one trial, this effect was absent in another trial. Hence, the efficacy of stimulation frequencies over placebo has varied across studies, thereby questioning the efficacy of active stimulation over placebo. Differences between the trials might have been caused by bias because some patients may have been unblinded either in low-frequency spinal cord stimulation by paraesthesiae or in high-frequency stimulation by the higher number of recharge sessions during active stimulation.2,33,52 Although all trials were set up to be double-blind, only one study tested whether blinding was obtained.44 Thus, it seems that the placebo conditions may have been insufficiently robust in so far as blinding can be questioned, thereby potentially compromising the control for contextual factors and the findings of the studies.
3. Future evidence: possible placebo controls in neuromodulation
For many years, the inclusion of placebo controls in low-frequency spinal cord stimulation trials has been regarded as impossible,53 due to paraesthesiae potentially unblinding patients to intervention allocation. This is not a limitation of high-frequency spinal cord stimulation because it does not evoke paraesthesiae. Recently, a new high-frequency spinal cord stimulation device has been endorsed by the National Institute for Health and Care Excellence.41 This spinal cord stimulation device avoids paraesthesiae and thereby potentially allows for adequate blinding. Although the title “Senza spinal cord stimulation system for delivering HF10 therapy to treat chronic neuropathic pain” of this Medical Technologies Guidance implies generic effectiveness in neuropathic pain, the randomized controlled trial–derived evidence considered was restricted to studies of patients with what was described as chronic back or leg pain or failed back surgery syndrome.1,3–6,10,29,48,50,54 The studies were retrospective analyses of patients with spinal cord stimulation implants,10,50 prospective studies of implants,3,4,48,54 or randomized studies comparing high-frequency spinal cord stimulation with traditional spinal cord stimulation.5,6,29 No studies included control groups other than the active spinal cord stimulation therapy as controls. This approval is one of many examples in the neuromodulation literature of the seemingly underlying assumption that it is not possible or necessary to compare these procedures to placebo controls. However, with the advancement of neuromodulation techniques that enable inclusion of adequate placebo controls, this assumption is increasingly challenged. One ongoing study includes a placebo control group in the evaluation of the high-frequency Senza spinal cord stimulation device51 and this study might be followed by additional placebo-controlled studies.
Placebo effects are typically investigated through comparison of active vs placebo treatment and as illustrated in the surgical literature placebo control of invasive procedures are feasible.58,59 In addition, so-called placebo-like effects8 can be investigated using open vs hidden administration of active treatment. In open–hidden designs, patients only receive active treatment, and therefore no inactive placebo treatment is given. Yet, the active treatment is either administered openly so that the effect of the treatment plus the contextual effect of knowing that an effective treatment is administered is tested, or the active treatment is administered hidden to the patients so that only the effect of the active treatment is measured. This design has been used in a neuromodulation trial of Parkinson disease, in which deep brain stimulation treatment was turned on and off while information about treatment and thus expectations toward the treatment were altered.40 Information about treatment (open condition) enhanced the clinical effect compared with no information about treatment (hidden condition), showing that placebo effects represented 34.8% of the magnitude of the effect of active treatment.40 In another study in patients with Parkinson disease, instead of switching off the stimulation, deep brain stimulation intensity was lowered from 80% to 20%, but this fact was not revealed to the patients. When patients were informed that stimulation was not changed, they showed a significantly better movement velocity of the hand compared with patients who were correctly informed that stimulation was reduced.46 This result has been replicated in related studies.30,35 Similar designs might offer an opportunity to include placebo controls in patients treated with neuromodulation devices. Correspondingly, an open–hidden design could be applied in high-frequency paraesthesiae-free spinal cord stimulation. In both the open condition and the hidden condition, patients would receive active spinal cord stimulation. However, in the open condition, patients would receive positive verbal suggestions about the treatment effect, whereas patients in the hidden condition would receive suggestions that treatment was not given. Including and controlling for a no-treatment group further allows for a more precise estimation of the specific effects of actual treatment and the additional nonspecific effects of the patients’ knowledge of treatment.
4. Conclusions
The evidence underlying neuromodulation techniques in chronic pain conditions is promising, but further research using designs including improved control conditions are warranted. Because biases, spontaneous remission, and placebo effects all seem to contribute to the effects of neuromodulation, future research and trials of neuromodulation would benefit from the inclusion of placebo controls. In this article, we have discussed how the continuous development of neuromodulation technology generates opportunities for better placebo controls and more successful blinding. To move the research field forward, we need to consider which demands we want to make for the evidence of efficacy in these types of procedures in present and future evaluations. This work would further elaborate our understanding of treatment mechanisms as well as optimization of treatment in clinical practice.
Conflict of interest statement
A.S.C. Rice conflicts of interest occurring in the past 24 months: A.S.C. Rice undertakes consultancy and advisory board work for Imperial College Consultants. In the past 24 months, this has included remunerated work for: Pharmanovo, Galapagos, Toray, Quartet, Lateral, Novartis, Pharmaleads, Mundipharma, Orion, Asahi Kasei, and Theranexus. A.S.C. Rice was the owner of share options in Spinifex Pharmaceuticals from which personal benefit accrued upon the acquisition of Spinifex by Novartis in July 2015 and from which future milestone payments may occur. ASCR is named as an inventor on patents: A.S.C. Rice, Vandevoorde S, and Lambert DM Methods using N-(2-propenyl) hexadecanamide and related amides to relieve pain. WO 2005/079771. Okuse K, et al. Methods of treating pain by inhibition of vgf activity EP13702262.0/WO2013 110945. The remaining authors have no conflicts of interest to declare.
References
[1]. Al-Kaisy A, Van Buyten JP, Smet I, Palmisani S, Pang D, Smith T. Sustained effectiveness of 10 kHz high-frequency spinal cord stimulation for patients with chronic, low back pain: 24-month results of a prospective multicenter study. Pain Med 2014;15:347–54.
[2]. Al-Kaisy A, Palmisani S, Pang D, Sanderson K, Wesley S, Tan Y, McCammon S, Trescott A. Prospective, randomized, sham-control, double blind, crossover trial of subthreshold spinal cord stimulation at various kilohertz frequencies in subjects suffering from failed back surgery syndrome (SCS frequency study). Neuromodulation 2018;21:457–65.
[4]. Al-Kaisy A, Palmisani S, Smith TE, Pang D, Lam K, Burgoyne W, Houghton R, Hudson E, Lucas J. 10 kHz high-frequency spinal cord stimulation for chronic axial low back pain in patients with No history of spinal surgery: a preliminary, prospective, open label and proof-of-concept study. Neuromodulation 2017;20:63–70.
[5]. Amirdelfan K, Yu C, Doust MW, Gliner BE, Morgan DM, Kapural L, Vallejo R, Sitzman BT, Yearwood TL, Bundschu R, Yang T, Benyamin R, Burgher AH, Brooks ES, Powell AA, Subbaroyan J. Long-term quality of life improvement for chronic intractable back and leg pain patients using spinal cord stimulation: 12-month results from the SENZA-RCT. Qual Life Res 2018;27:2035–44.
[7]. Atlas LY, Wielgosz J, Whittington RA, Wager TD. Specifying the non-specific factors underlying opioid analgesia: expectancy, attention, and affect. Psychopharmacology (Berl) 2014;231:813–23.
[8]. Benedetti F. Mechanisms of placebo and placebo-related effects across diseases and treatments. Annu Rev Pharmacol Toxicol 2008;48:33–60.
[9]. Bingel U, Wanigasekera V, Wiech K, Mhuircheartaigh RN, Lee MC, Ploner M, Tracey I. The effect of treatment expectation on drug efficacy: imaging the analgesic benefit of the opioid remifentanil. Sci Transl Med 2011;3. doi: 10.1126/scitranslmed.3001244.
[39]. Melzack R, Wall PD. Pain mechanisms: a new theory. Science 1965;150:971–9.
[40]. Mercado R, Constantoyannis C, Mandat T, Kumar A, Schulzer M, Stoessl AJ, Honey CR. Expectation and the placebo effect in Parkinson’s disease patients with subthalamic nucleus deep brain stimulation. Mov Disord 2006;21:1457–61.
[51]. Sham-controlled RCT on 10kHz high-frequency spinal cord stimulation for chronic neuropathic low back pain (Modulate-LBP)—tabular view—ClinicalTrials.gov. Available at: https://clinicaltrials.gov/ct2/show/record/NCT03470766. Accessed April 26, 2019.
[53]. Tesfaye S, Watt J, Benbow SJ, Pang KA, Miles J, MacFarlane IA. Electrical spinal-cord stimulation for painful diabetic peripheral neuropathy. Lancet 1996;348:1698–701.
[54]. Tiede J, Brown L, Gekht G, Vallejo R, Yearwood T, Morgan D. Novel spinal cord stimulation parameters in patients with predominant back pain. Neuromodulation 2013;16:370–5.