Surgical treatment of degenerative conditions of the spine often requires arthrodesis. Fusion procedures are among the most common procedures performed on the spine and were performed at a total hospital cost of $24.3 billion in 2015.1 As a result, strategies to ensure the success of these procedures are of great importance to both patients and the healthcare system. Rates of pseudoarthrosis range from 5% to 35% in the lumbar spine and can be even higher with three or more fused levels.2 Pseudoarthrosis often requires additional fusion procedures or revision of hardware, leading to further costs, procedures, and pain for patients. Despite advancements in surgical techniques and the understanding of risk factors for pseudoarthrosis, the role of electrical stimulation in preventing pseudoarthrosis and improving clinical outcomes remains less clear.
It is understood that bone produces electrical fields when under mechanical stress. Implanted direct current stimulation (DCS) was first applied to spine fusion procedures by Dwyer and Wickham in 1974.3 Noninvasive electrical stimulation has also been an attractive option for promoting fusion due to its ease of use, having been approved for clinical use by the FDA in 1979.4 Despite the use of noninvasive electrical stimulation to promote fusion for the past few decades, evidence to support the use of such devices has been sparse.4,5 In contrast, there is strong evidence to support the use of DCS to promote fusion in clinical as well as preclinical models, which suggests that the underlying biophysical principles of electrical stimulation for spinal fusion are valid.6 Noninvasive electrical stimulation comes with advantages and disadvantages that are distinct from those for implanted direct current stimulators. Noninvasive electrical stimulation has an overall mechanism of action different from that of DCS, although some intracellular signal transduction pathways may ultimately overlap with DCS pathways.7 Even though these stimulators are noninvasive, easily worn, and free from infection risk, the device requires good compliance on the part of the patient as inconsistent use may reduce the effectiveness.8 Patient factors such as age, obesity, and smoking status may also play a role in the effectiveness of noninvasive electrical stimulation.6,9 Factors which may influence patient compliance with these devices include patient discomfort, minor skin rashes, and pain, although these are relatively uncommon and minor complaints.8 As they are easily given to patients in the postoperative period, wearable noninvasive electrical stimulation devices may be overutilized in certain cases, a situation that has important implications for cost and healthcare utilization. Prior meta-analyses of electrical stimulation have pooled DCS along with noninvasive modalities to support the overall use of electrical stimulation, despite the inherent differences between implanted DCS devices and wearable devices for noninvasive electrical stimulation.6,10 In the present systematic review and meta-analysis we aimed to do the following: 1) summarize clinical studies available in the literature that investigated noninvasive electrical stimulation used to augment fusion rates and 2) determine the pooled efficacies of noninvasive electrical stimulation modalities in improving fusion rates. To our knowledge, this is the first meta-analysis of fusion rates after treatment with combined magnetic fields (CMFs) separately from those for pulsed electromagnetic fields (PEMFs).
Noninvasive Electrical Stimulation Modalities
There are three types of noninvasive electrical stimulation: capacitively coupled stimulation (CCS), PEMF, and CMF. CCS involves the use of electrical current passed through capacitive pads placed on the skin, which passes a current at a frequency of 60 kHz for a current density of 5 µA root mean square/cm2 and for a 12-mV root mean square/cm2 field at the vertebral body.11 CCS is thought to promote fusion by causing the opening of voltage-gated Ca2+ channels, leading to an increase in cytosolic Ca2+.7 This causes activation of phospholipase A2 and synthesis of prostaglandin E2, which has been shown to induce proliferation of osteoblast-like cells in vitro.7,12 The cytosolic increase in Ca2+ may also induce osteoblast proliferation via the calmodulin pathway, which is similar to the mechanism by which mechanical stress promotes osteoblast proliferation.7 CCS has been shown to modulate pain in vertebral body fractures, in addition to its use in augmenting fusion rates.13
Both PEMF and CMF are methods of inductive coupling, which involves the use of electromagnetic coils through which a current is passed to generate a magnetic field. This magnetic field induces a secondary electric field at the fusion site.14 CMF differs from PEMF in that it involves a time-varying magnetic field superimposed on a static magnetic field.14–16 Based on in vitro studies, both methods of inductive coupling are thought to release Ca2+ from the endoplasmic reticulum into the cytoplasm, which then activates calmodulin to induce the expression of genes implicated in the differentiation of osteoblasts to promote bone growth and healing.7,17,18 PEMF has been shown to help in pain modulation, likely by acting on inflammatory pathways to suppress the release of inflammatory cytokines.19,20 Reduced analgesic use with PEMF has been demonstrated in breast reconstruction patients postoperatively as well as patients with failed back surgery syndrome.19,20 PEMF has also been shown to induce osteogenesis through upregulation of BMP-2 and BMP-4 in vitro.21 Based on animal studies, CMF has been similarly shown to promote fusion by inducing the expression of osteoinductive factors such as BMP-2 and transforming growth factor (TGF)–β1.22 Computer simulations of CMF have demonstrated adequate magnetic field coverage of lumbar interbody and intertransverse fusions; however, the intensity of the resulting electrical field produced at the fusion sites in the simulation was orders of magnitude lower than would be expected to have a clinically significant effect.23
Methods
PubMed, Embase, and the Cochrane Clinical Trials database were searched according to the following search terms for any field of the text: ((electrical bone growth stimulator) OR (capacitive coupling) OR (capacitively coupled) OR (pulsed electromagnetic field) OR (combined magnetic field) OR (inductive coupling)) AND ((spine fusion*) OR (spinal fusion) OR (cervical fusion*) OR (lumbar fusion*) OR (lumbar spinal fusion*) OR (cervical spinal fusion*) OR (spine arthrodesis) OR (lumbosacral fusion) OR (thoracolumbar fusion) OR (cervicothoracic fusion)). This search identified 103 articles in PubMed, 93 in Embase, and 17 in the Cochrane Clinical Trials database. The studies were then further narrowed down according to the search strategy outlined in Fig. 1. Included studies were peer-reviewed publications in the English language assessing the use of noninvasive electrical stimulation after a fusion procedure for the purpose of augmenting fusion. Noninvasive electrical stimulation was defined as CCS, PEMF, or CMF, with synonymous terms used during the search. Studies utilizing animal models or nonsurgical patients and those without an unstimulated control group were excluded. Data were searched and extracted for analysis by the lead author. Fusion rates at around 12 months postoperatively were used for meta-analysis. Funding sources and definitions of successful fusion were also extracted to compare studies.
Flow diagram of search strategy. The search identified 8 articles from 213 in the initial search. The study by Cheaney et al.15 included both PEMF and CMF.
Meta-analysis was performed using R version 4.1.0 (The R Foundation for Statistical Computing). When comparing fusion rates between studies, rates of successful radiographic fusion as determined by each study were used for comparison. Odds ratios (ORs) were used for measurements of effects on fusion rates with each modality. Dichotomous outcomes of successful or unsuccessful fusion were pooled via the Mantel-Haenszel method. A random-effects meta-analysis model was then used to give a pooled estimate of noninvasive electrical stimulation for fusion rates and clinical outcomes. The random-effects model was chosen over a fixed-effects model due to differences in study design, patient selection, and measurement of outcomes, which may result in significant variations between studies that are not due to chance. Systematic review and meta-analysis was performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.24 Retrospective studies were included in the meta-analysis due to the low overall number of studies and randomized controlled trials (RCTs) identified. Forest plots were used to visualize data. A p value < 0.05 was used to determine significance. Subgroups for analysis were chosen based on patient and surgical factors that may affect rates of pseudoarthrosis or interfere with the effectiveness of electrical stimulation modalities. Subgroup analysis was also performed using a random-effects model with dichotomous outcomes pooled via the Mantel-Haenszel method. Data were extracted according to the information available from the included studies. Risk-of-bias analysis for level I studies was conducted using the Risk of Bias 2 (RoB 2) tool developed by the Cochrane Collaboration.25 Similarly, the Cochrane Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tool was used to determine risk of bias for the nonrandomized studies.26 Studies were assessed for bias using these tools by the lead author and summarized with traffic light plots using the robvis package for R.27 This review has not been registered and no protocol was prepared for this study.
Results
A total of 213 articles were identified in the initial search. Studies which were duplicated, irrelevant, or lacking appropriate controls were excluded (Fig. 1). A total of 8 clinical studies were identified, 1 study of CCS, 6 studies of PEMF, and 2 studies of CMF, with an overall total of 1216 participants.8,9,11,15,16,28–30 The identified studies and their grading according to the system described by Wright et al. are summarized in Table 1.31 A number of additional studies that initially appeared to meet inclusion criteria were excluded due to lack of appropriate control groups or use of a study design in which electrical stimulation was used for purposes other than to promote bony fusion.4,32–34 Pooled fusion rates at around 12 months postoperatively were extracted from each study. Results of fusion rates with CCS, PEMF, and CMF are summarized in Table 2. As there was only 1 study performed which assessed CCS, meta-analysis of fusion rates could not be done for CCS. Publication bias with the use of a funnel plot could not be performed due to the low number of studies identified. A risk-of-bias analysis for the level I studies using the RoB 2 tool is shown in the traffic light plot in Fig. 2. Among the level I studies there was an overall high risk of bias for almost every study, due to multiple common factors. These include insufficient blinding, unclear methods to address the effect of noncompliance, and large proportions of patients being lost to follow-up in some studies. A risk-of-bias analysis was similarly conducted for nonrandomized studies using the ROBINS-I tool and is summarized in the traffic light plot in Fig. 3.
List of all clinical studies evaluating fusion rates in noninvasive electrical stimulation as identified in the search strategy and included in the meta-analysis
| Authors & Year | Study Type | Stimulation Modality | Inclusion Criteria | Fusion Procedure | Primary Outcome | Radiographic Success Definition | No. of Pts | Funding or COIs | Level of Evidence |
|---|---|---|---|---|---|---|---|---|---|
| Cheaney et al., 202015 | Retrospective case-control study | PEMF, CMF | All adults undergoing thoracolumbar fusion for degenerative indications | Any thoracic or lumbar procedure for degenerative pathology | Radiographic fusion on 12-mo CT | Evidence of solid bony growth or lack of bony growth w/o hardware failure determined by blinded investigator & radiologist | 60 | None | III |
| Coric et al., 201830 | Retrospective cohort clinical trial w/ historic controls | PEMF | Adults undergoing ACDF w/ ≥1 of: age ≥65 yrs, required multilevel arthrodesis (≥5 levels), prior failed fusion at any cervical level, habitual nicotine use at time of op, diabetes, & osteoporosis | ACDF | Radiographic fusion at 6 & 12 mos | Presence of continuous bridging bone as determined by unblinded treating surgeon on plain film | 274 | Funded by grant from Orthofix | II |
| Foley et al., 20089 | RCT | PEMF | Adults 18–75 yrs old undergoing ACDF, multilevel fusion, or active smokers w/ radiographically correlating symptomatic radiculopathy | ACDF anywhere from C3 to T1 | Radiographic fusion at 3, 6, & 12 mos | >50% bony bridging w/o radiolucency & >4° motion btwn all fused levels on plain film, assessed by 2 independent surgeons w/ radiologist as reference | 323 | None | I |
| Linovitz et al., 200216 | RCT | CMF | Adults undergoing 1- & 2-level posterolat lumbosacral fusions w/o instrumentation | Primary intertransverse fusion w/o internal fixation of 1–2 vertebral levels btwn L3 & S1 w/in past 30 days | Radiographic fusion at 9 mos on CT scan & plain film | Fusion grading from 0 (no fusion) to 3 (solid fusion) by blinded review panel | 201 | Corporate or industry funds supported work | I |
| Jenis et al., 200028 | RCT | PEMF, DCS | Adults 18–75 yrs old undergoing posterolat fusion | Posterolat lumbar fusion w/ autogenous iliac crest graft, w/ or w/o decompression | Radiographic fusion at 12 mos | Grading from 1 (obvious pseudoarthrosis) to 3 (solid fusion) on plain film by unblinded independent reviewer | 61 | None | I |
| Marks, 200029 | Retrospective cohort study | PEMF | Randomly selected lumbar fusion pts w/ discogenic back pain or failed back op syndrome failing 6 mos of conservative therapy | Lumbar fusion | Radiographic fusion at 6–24 mos | Incorporation of graft, w/o radiolucency btwn graft & vertebral bone & no motion at each level of fusion on plain film | 61 | None | III |
| Goodwin et al., 199911 | RCT | CCS | Adult pts undergoing 1- or 2-level PLIF, ALIF, or posterolat fusion | 1- or 2-level PLIF, ALIF, or posterolat lumbar fusion w/ any type of fixation other than interbody cages & any type of graft material | Combination of radiographic fusion w/in 12 mos & clinical improvement as defined by study | Uninterrupted bilat bony fusion masses on both anteroposterior & lat radiographs, w/o lucency or evidence of screw motion | 179 | None | I |
| Mooney, 19908 | RCT | PEMF | Adult patients undergoing primary lumbar interbody fusion | ALIF or PLIF | Radiographic fusion w/in 12 mos | >50% assimilation of bone on plain film per treating surgeon & confirmed by independent blinded radiologist w/ independent blinded orthopedic surgeon acting as referee | 195 | None | I |
ACDF = anterior cervical discectomy and fusion; ALIF = anterior lumbar interbody fusion; COI = conflict of interest; PLIF = posterior lumbar interbody fusion; pts = patients.
The studies used differing methods of evaluating and defining fusions.
Summary of fusion rates with CCS, PEMF, and CMF
| Study | Pt Group | OR (95% CI) | p Value | Cochran’s Q | |
|---|---|---|---|---|---|
| Treatment | Control | ||||
| CCS | |||||
| Goodwin et al., 1999,11 total | 77/85 (90.6%) | 77/94 (81.6%) | 2.13 (0.87–5.21) | 0.0998 | |
| PEMF | |||||
| Mooney, 19908 | 59/64 | 36/53 | 1.89 (0.36–9.80) | 0.4492 | 21.62 |
| Jenis et al., 200028 | 14/22 | 18/22 | |||
| Marks, 200029 | 41/42 | 10/19 | |||
| Foley et al., 20089 | 116/125 | 104/120 | |||
| Coric et al., 201830 | 201/217 | 76/92 | |||
| Cheaney et al., 202015 | 11/16 | 20/20 | |||
| Total | 442/486 (90.9%) | 264/326 (81.0%) | |||
| PEMF (level I only) | |||||
| Mooney, 19908 | 59/64 | 36/53 | 1.72 (0.39–7.55) | 0.4734 | 8.80 |
| Jenis et al., 200028 | 14/22 | 18/22 | |||
| Foley et al., 20089 | 116/125 | 104/120 | |||
| Total | 189/211 (89.6%) | 158/195 (81.0%) | |||
| CMF | |||||
| Linovitz et al., 200216 | 67/104 | 42/97 | 0.90 (0.07–11.93) | 0.9385 | 3.09 |
| Cheaney et al., 202015 | 21/24 | 20/20 | |||
| Total | 88/128 (68.8%) | 62/117 (53.0%) | |||
Only 1 study was available for fusion rates with CCS; therefore, no meta-analysis could be done and the results of the only available study are summarized above. Meta-analyses of fusion rates with PEMF and CMF are also summarized above. None of the included modalities reached statistical significance for fusion rates.
Risk-of-bias analysis as performed with the Cochrane RoB 2 tool for level I randomized studies. All but 1 of the included studies had an overall high risk of bias. Figure is available in color online only.
Risk-of-bias analysis performed with the Cochrane ROBINS-I tool for nonrandomized studies. Figure is available in color online only.
Fusion Rates With CCS
The single RCT of CCS by Goodwin et al. in 1999 did not identify any difference between treatment and control groups in terms of fusion rates alone.11 The study found a 90.6% fusion rate in the treatment group and 81.6% fusion rate in the control group, with an OR of 2.13, which did not reach statistical significance (95% CI 0.87–5.21, p = 0.0998). Although no meta-analysis could be performed, the results of the study by Goodwin et al. are summarized in Table 2 and Fig. 4A. This study utilized a combined primary endpoint of clinical success and radiographic fusion rather than fusion alone. However, when radiographic fusion was considered separately, there was no difference in fusion rates between treatment and control groups.
A: Rates of radiographic fusion using CCS as demonstrated by Goodwin et al.11 No meta-analysis could be performed because only 1 study of CCS was identified. B: Random-effects meta-analysis of fusion rates with PEMF. C: Random-effects meta-analysis of PEMF with only level I evidence studies D: Random-effects meta-analysis of fusion rates with CMF. MH = Mantel-Haenszel. Figure is available in color online only.
Meta-Analysis of Fusion Rates With PEMF
Meta-analysis of fusion rates with PEMF did not identify any significant differences between stimulated and nonstimulated groups. Three RCTs and 3 retrospective studies were included in the meta-analysis of fusion rates for PEMF, summarized in Table 2 and the forest plot in Fig. 4B. Fusion rates were 90.9% of the pooled treatment group and 81.0% of the pooled control group, with an OR of 1.89, which was not significant (95% CI 0.36–9.80, p = 0.4492). Statistical heterogeneity between the studies was substantial, with an I2 of 77%. Meta-analysis of fusion rates with PEMF using only level I studies was also performed and is shown in Fig. 4C. No significant differences were noted between control and treatment groups in high-quality level I studies, with an OR of 1.72 (95% CI 0.39–7.55, p = 0.4734). The I2 statistic was also 77% for this analysis. Between studies there were significant variations in how the presence of a bony fusion was assessed and defined. While some studies used a grading method to describe the presence and quality of bony fusion, others treated fusion as a binary outcome.
Meta-Analysis of Fusion Rates With CMF
Similarly to PEMF meta-analysis, meta-analysis of fusion rates with CMF also did not identify any significant differences between stimulated and nonstimulated groups. While prior meta-analyses have included CMF in the analysis of PMF, it represents a distinct modality of noninvasive electrical stimulation and is considered separately here. One level I RCT and 1 level III retrospective case-control study were included in the meta-analysis, summarized in Table 2 and Fig. 4D. Fusion rates were 68.8% of the pooled treatment group and 53.0% of the pooled control group, with an OR of 0.90, which was not significant (95% CI 0.07–11.93, p = 0.9385). Heterogeneity between the 2 studies was substantial, with an I2 of 68%. The studies differed on the overall benefit to CMF, with an RCT by Linovitz et al. showing an overall benefit.16 However, when stratified by sex, the benefit of CMF was only seen among women in this study. On the other hand, Cheaney found no difference in fusion rates between CMF and unstimulated control groups.15 The studies used similar methods for assessing the presence and quality of bony union radiographically through a graded scale.15,16
Subgroup Analysis
Although subgroup analysis identified significantly increased fusion rates in some subgroups, it remains unclear whether this is a true effect or the result of small samples given the low number of studies and patients in most subgroups available for subgroup analysis. Subgroup analysis was performed for PEMF to identify any potential subgroups which benefitted from stimulation (Table 3). Subgroups were selected based on availability in the included studies and factors which may influence the fusion rate. Successful fusion was significantly increased with PEMF in cervical, single-level, multilevel, index, allograft, interbody fusion, and noninstrumented patients. However, there was no difference in fusion rates with PEMF among patients who were smokers compared with nonsmokers or among lumbosacral, autograft, revision, posterior fusion, and instrumented patients. Due to the low number of studies and lack of subgroup congruency for equal comparisons between studies, much of the subgroup analysis was dependent on studies by Mooney8 and Marks.29
Subgroup analysis of PEMF
| Study | Pt Group | OR (95% CI) | p Value | Cochran’s Q | |
|---|---|---|---|---|---|
| PEMF Treatment | Control | ||||
| Smokers | |||||
| Mooney, 19908 | 24/27 | 12/20 | 4.84 (0.34–68.26) | 0.2428 | 8.04 |
| Jenis et al., 200028 | 6/12 | 8/13 | |||
| Marks, 200029 | 18/19 | 0/3 | |||
| Total | 48/58 (82.8%) | 20/36 (55.6%) | |||
| Nonsmokers | |||||
| Mooney, 19908 | 35/37 | 24/33 | 3.67 (0.28–47.84) | 0.3205 | 6.36 |
| Jenis et al., 200028 | 7/10 | 8/9 | |||
| Marks, 200029 | 23/23 | 10/16 | |||
| Total | 65/70 (92.9%) | 42/58 (72.4%) | |||
| Cervical | |||||
| Foley et al., 20089 | 116/125 | 104/120 | 2.34 (1.33–4.10) | 0.0030 | 0.25 |
| Coric et al., 201830 | 201/217 | 76/92 | |||
| Total | 317/342 (92.7%) | 180/212 (84.9%) | |||
| Lumbosacral | |||||
| Mooney, 19908 | 59/64 | 36/53 | 3.93 (0.31–50.22) | 0.2924 | 14.68 |
| Jenis et al., 200028 | 14/22 | 18/22 | |||
| Marks, 200029 | 41/42 | 10/19 | |||
| Total | 114/128 (89.1%) | 64/94 (68.1%) | |||
| Single level | |||||
| Mooney, 19908 | 43/46 | 29/40 | 8.69 (1.73–43.70) | 0.0087 | 1.29 |
| Marks, 200029 | 18/18 | 6/12 | |||
| Total | 61/64 (95.3%) | 35/52 (67.3%) | |||
| Multilevel | |||||
| Mooney, 19908 | 16/18 | 7/13 | 9.46 (2.16–41.43) | 0.0029 | 0.34 |
| Marks, 200029 | 23/24 | 4/7 | |||
| Total | 39/42 (92.9%) | 11/20 (55.0%) | |||
| Index op | |||||
| Mooney, 19908 | 59/64 | 36/53 | 16.78 (1.07–263.83) | 0.0448 | 3.22 |
| Marks, 200029 | 38/38 | 6/14 | |||
| Total | 97/102 (95.1%) | 42/67 (62.7%) | |||
| Revision | |||||
| Marks, 200029 | 3/4 | 4/5 | 0.75 (0.03–17.51) | 0.8579 | NA |
| Total | 3/4 (75%) | 4/5 (80%) | |||
| Autograft | |||||
| Mooney, 19908 | 23/25 | 14/19 | 2.87 (0.29–28.88) | 0.3706 | 10.04 |
| Jenis et al., 200028 | 14/22 | 18/22 | |||
| Marks, 200029 | 19/20 | 5/11 | |||
| Total | 56/67 (83.6%) | 37/52 (71.2%) | |||
| Allograft | |||||
| Mooney, 19908 | 25/27 | 16/22 | 2.85 (1.19–6.83) | 0.0191 | 2.27 |
| Marks, 200029 | 11/11 | 4/7 | |||
| Foley et al., 20089 | 116/125 | 104/120 | |||
| Total | 152/163 (93.3%) | 124/149 (83.2%) | |||
| Interbody fusion | |||||
| Mooney, 19908 | 59/60 | 36/53 | 5.69 (1.48–21.84) | 0.0114 | 8.87 |
| Marks, 200029 | 19/20 | 6/14 | |||
| Foley et al., 20089 | 116/125 | 104/120 | |||
| Coric et al., 201830 | 201/217 | 76/92 | |||
| Total | 395/422 (93.6%) | 222/279 (79.6%) | |||
| Posterior fusion | |||||
| Jenis et al., 200028 | 14/22 | 18/22 | 0.39 (0.10–1.56) | 0.1825 | NA |
| Total | 14/22 (63.6%) | 18/22 (81.8%) | |||
| Instrumentation | |||||
| Mooney, 19908 | 44/48 | 28/39 | 1.90 (0.88–4.10) | 0.1025 | 7.35 |
| Jenis et al., 200028 | 14/22 | 18/22 | |||
| Marks, 200029 | 9/10 | 1/1 | |||
| Foley et al., 20089 | 116/125 | 104/120 | |||
| Coric et al., 201830 | 201/217 | 76/92 | |||
| Total | 384/422 (91.0%) | 227/274 (82.8%) | |||
| w/o instrumentation | |||||
| Mooney, 19908 | 15/16 | 8/14 | 21.81 (3.60–132.26) | 0.0008 | 0.85 |
| Marks, 200029 | 32/32 | 9/18 | |||
| Total | 47/48 (97.9%) | 17/32 (53.1%) | |||
NA = not applicable.
Subgroups were selected based on availability in prior studies and possible effect on fusion rates. Boldface type indicates statistical significance.
Discussion
The results of this meta-analysis suggest that noninvasive electrical stimulation does not have a significant effect on fusion rates at 1 year postoperatively. There is still no evidence to suggest that CCS improves fusion rates, and at the time of this report only 1 study had collected these data.11 In comparison, there have been multiple studies done of PEMF for the improvement of fusion rates. However, the present meta-analysis did not identify any significant difference in fusion rates between treatment and control groups for PEMF. This remains true even when only level I evidence is considered. Although subgroup analysis identified cervical, single-level, multilevel, index, allograft, interbody fusion, and noninstrumented patients as having significantly higher rates of fusion when using PEMF, this result was largely based on a small number of patients from studies which are now decades old. It is unclear whether the statistical significance seen in the subgroup analysis is indicative of a true benefit in these subgroups, or whether it is a random result due to the selection of small cross-sections of patients across studies. The very wide confidence intervals seen in most subgroups reflect a great deal of uncertainty; thus, the significant results in the subgroup analysis must be interpreted with caution. As with the PEMF group, there was no statistically significant increase in fusion rates among the CMF treatment group. Although Linovitz et al. found an overall statistically significant benefit in fusion rates with CMF, this was not seen in males when patients were stratified by sex, and the overall benefit of CMF could not be replicated by Cheaney et al.15,16 Prior meta-analyses have found a significant increase in fusion rates with PEMF; however, these meta-analyses did not include the study by Cheaney et al. and grouped CMF with PEMF in their analyses.6,10,15,16 The only study included in this meta-analysis not included in prior analyses is Cheaney et al.6,15
The results presented here should be taken into consideration given the main drawback of noninvasive electrical stimulation: cost. Given the concern over high costs to the healthcare system, the cost of devices such as noninvasive electrical stimulators must be considered along with clinical risks and benefits. In one of the studies included in this meta-analysis, the PEMF and CMF devices came with costs of $4995 and $1800, respectively.15 Another study of costs found that noninvasive electrical stimulation after single-level anterior lumbar interbody fusions was associated with a 27% higher cost per patient, suggesting that patients selected for noninvasive stimulation are prone to higher costs.35 This meta-analysis does not support the routine use of noninvasive electrical stimulation for all patients, and we recommend caution when utilizing these devices given the high costs. Prior meta-analyses have found a statistically significant increase in fusion rates with PEMF; however, these analyses have grouped CMF with PEMF and due to timing could not include the study by Cheaney et al.6,10,15 That this single study by Cheaney et al.15 has changed the outcome of the meta-analysis for fusion rates in PEMF is likely a reflection of the small sample size and low strength of evidence for PEMF. Although the subgroup analysis identified certain groups as having significantly increased fusion rates with PEMF, these results must be interpreted with caution given the small number of studies that could be included for comparison. Additionally, much of the data extracted for subgroup analysis came from some studies which are now more than 20 years old. This must also be considered in addition to the small sample sizes when interpreting the results of subgroup analysis. Some of the factors present in the subgroup analysis, such as graft material, were not variables which were controlled for during the original studies, and patient selection factors may also explain some of the results of the subgroup analyses.
The heterogeneity of the studies included in the meta-analysis may be due to small sample sizes but may also be due to other factors. I2 values of 77% for PEMF and 68% for CMF suggest that a substantial amount of the variation between the various studies may not be due to chance alone and is likely related to differences in study design. This heterogeneity between studies may also explain the insignificance of the results of the present meta-analysis, as significant variations in aspects of the included studies could mask a true treatment effect, if one exists. Factors such as blinding, study design, duration of device use, rates of compliance, and the definition of a successful fusion may all contribute to the heterogeneity between the studies. Similarly, patient selection was also varied as some studies restricted themselves to patients based on instrumentation, approaches, or patient factors for various reasons. We have attempted to account for this with our subgroup analysis, but because data on subgroups were not uniformly reported across studies, much of the subgroup analysis comes from only a small number of patients. Many of the studies included did not report compliance rates with the devices, which has been shown to be related to their purported effectiveness, and may also explain the high rates of patients excluded from the final analysis in multiple studies.8 Similarly, the duration and manner in which the devices were used was not always the same between studies in the PEMF group, although it was similar between the 2 CMF studies.8,15,16 In particular the duration of PEMF use has varied from 2 to 8 hours per day for 3–6 months for the included studies.8,9,15,28–30 The CMF devices were used 30 minutes daily, and the CCS device was intended to be used continuously. The significant heterogeneity and high risk of bias among these studies make it difficult to draw strong conclusions. However, along with the lack of strong evidence from the meta-analysis, this supports our conclusion that evidence for noninvasive electrical stimulation is insufficient to justify its routine use.
The studies included in the meta-analysis have also taken place over a span of 30 years, during which time new technologies and techniques for fusion have emerged alongside noninvasive electrical stimulation to improve fusion rates. These advancements may influence the effectiveness of electrical stimulation or may be so effective at promoting fusion themselves that the effect of electrical stimulation becomes small or negligible, especially when assessed with small or statistically underpowered studies. It would be particularly valuable to know the interactions of noninvasive electrical stimulation with biologics and biologically active implants. However, there are no studies that we have identified that could clarify this.
Future studies should aim to develop more robust evidence for the use of noninvasive electrical stimulation. Further studies of CCS and fusion rates are especially needed as only 1 such study has been done and is more than 20 years old. More data with respect to subgroups of patients who may benefit from noninvasive electrical stimulation are also needed. Factors such as smoking, instrumentation, and graft material have been explored in previous meta-analyses, but as with the present subgroup analysis, it is unclear whether these subanalyses are indicative of clinically significant differences between treatment and control groups due to small sample sizes.6,36 Further studies should also aim to delineate the role of patient factors such as smoking, age, diabetes, revision surgery, history of failed fusions, and BMI on the effectiveness of noninvasive electrical stimulation. Reporting of subgroup outcomes will play an important role in this process by allowing more robust meta-analyses in the future. Many of the studies included here did not report subgroup results, which led to their exclusion from subgroup analysis despite the original studies collecting this information. Additionally, patient-reported outcome measures and pain measures would facilitate future analyses and comparisons of clinical outcomes as well as costs. Grouping future studies based on duration and intensity of stimulation may help to discern whether any effect is present at all. The present meta-analysis included only studies of degenerative spine patients, but information on the effect of noninvasive electrical stimulation on fusion for other indications may also be useful.
Despite the lack of evidence for noninvasive electrical stimulation in promoting bony fusion, the significant results in the subgroup analyses suggest that the principles underlying electrical stimulation to promote fusion may be sound overall and that noninvasive electrical stimulation therefore merits further study. The achievement of stable arthrodesis is the goal of fusion procedures, and technologies to augment fusion rates are needed despite the advances of past decades. However, without strong evidence to support the use of noninvasive electrical stimulation, skepticism must be exercised. Despite the low risk of harm and ease of use of noninvasive electrical stimulation, the findings of this review do not support their routine use. However, noninvasive electrical stimulation may be used with appropriate caution at the discretion of individual surgeons for patients in whom there is a strong concern for pseudoarthrosis.
Study Limitations
Despite the inclusion of multiple RCTs and retrospective studies, this meta-analysis has limitations. The overall number of patients and studies included are low, although this is a limitation of the existing literature rather than the methods used. When pooling outcomes, the studies included used different definitions of successful fusion, which may be a potential source of bias in this meta-analysis. This meta-analysis included retrospective studies in addition to RCTs, and although some of the retrospective studies included were of high quality, they suffer from their inherent lack of randomization and blinding.
Conclusions
In conclusion, this meta-analysis did not identify any significant improvement in fusion rates at 12 months postoperatively with any modality of noninvasive electrical stimulation. The overall body of evidence to support the use of noninvasive electrical stimulation is small and centers around only a handful of small RCTs and retrospective studies. Despite the significant results seen in the subgroup analysis, these analyses were based on a small number of studies and require further investigation of patient risk factors and surgical factors which may influence the effectiveness of noninvasive electrical stimulation. Further studies of clinical outcomes and fusion rates among all modalities of noninvasive electrical stimulation would help further delineate which patients, if any, benefit from the use of these devices. In the absence of strong clinical evidence of the effectiveness of noninvasive electrical stimulation, this treatment should be used with caution in light of the costs involved.
Disclosures
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
Author Contributions
Conception and design: Plummer. Acquisition of data: Matur. Analysis and interpretation of data: Matur. Drafting the article: Matur. Critically revising the article: Plummer, Mejia-Munne, Tabbosha, Virojanapa, Nasser, Cheng. Reviewed submitted version of manuscript: all authors. Statistical analysis: Matur. Study supervision: Virojanapa, Nasser, Cheng.
References
- 1↑
Martin BI, Mirza SK, Spina N, Spiker WR, Lawrence B, Brodke DS. Trends in lumbar fusion procedure rates and associated hospital costs for degenerative spinal diseases in the United States, 2004 to 2015.Spine (Phila Pa 1976).2019;44(5):369–376.
- 2↑
Chun DS, Baker KC, Hsu WK. Lumbar pseudarthrosis: a review of current diagnosis and treatment. Neurosurg Focus. 2015;39(4):E10.
- 4↑
Simmons JW. Treatment of failed posterior lumbar interbody fusion (PLIF) of the spine with pulsing electromagnetic fields. Clin Orthop Relat Res. 1985;193):127–132.
- 5↑
Park P, Lau D, Brodt ED, Dettori JR. Electrical stimulation to enhance spinal fusion: a systematic review. Evid Based Spine Care J. 2014;5(2):87–94.
- 6↑
Cottrill E, Pennington Z, Ahmed AK, et al. The effect of electrical stimulation therapies on spinal fusion: a cross-disciplinary systematic review and meta-analysis of the preclinical and clinical data. J Neurosurg Spine. 2020;32(1):106–126.
- 7↑
Brighton CT, Wang W, Seldes R, Zhang G, Pollack SR. Signal transduction in electrically stimulated bone cells. J Bone Joint Surg Am. 2001;83(10):1514–1523.
- 8↑
Mooney V. A randomized double-blind prospective study of the efficacy of pulsed electromagnetic fields for interbody lumbar fusions. Spine (Phila Pa 1976). 1990;15(7):708–712.
- 9↑
Foley KT, Mroz TE, Arnold PM, et al. Randomized, prospective, and controlled clinical trial of pulsed electromagnetic field stimulation for cervical fusion. Spine J. 2008;8(3):436–442.
- 10↑
Tian NF, Wu YS, Zhang XL, Mao FM, Xu HZ, Chi YL. Efficacy of electrical stimulation for spinal fusion: a meta-analysis of fusion rate. Spine J. 2013;13(10):1238–1243.
- 11↑
Goodwin CB, Brighton CT, Guyer RD, Johnson JR, Light KI, Yuan HA. A double-blind study of capacitively coupled electrical stimulation as an adjunct to lumbar spinal fusions. Spine (Phila Pa 1976). 1999;24(13):1349–1357.
- 12↑
Feyen JHM, Di Bon A, van der Plas A, Löwik CWGM, Nijweide PJ. Effects of exogenous prostanoids on the proliferation of osteoblast-like cells in vitro. Prostaglandins. 1985;30(5):827–840.
- 13↑
Rossini M, Viapiana O, Gatti D, de Terlizzi F, Adami S. Capacitively coupled electric field for pain relief in patients with vertebral fractures and chronic pain. Clin Orthop Relat Res. 2010;468(3):735–740.
- 14↑
Gan JC, Glazer PA. Electrical stimulation therapies for spinal fusions: current concepts. Eur Spine J. 2006;15(9):1301–1311.
- 15↑
Cheaney B II, El Hashemi M, Obayashi J, Than KD. Combined magnetic field results in higher fusion rates than pulsed electromagnetic field bone stimulation after thoracolumbar fusion surgery. J Clin Neurosci. 2020;74:115–119.
- 16↑
Linovitz RJ, Pathria M, Bernhardt M, et al. Combined magnetic fields accelerate and increase spine fusion: a double-blind, randomized, placebo controlled study. Spine (Phila Pa 1976). 2002;27(13):1383–1389.
- 17↑
Zayzafoon M. Calcium/calmodulin signaling controls osteoblast growth and differentiation. J Cell Biochem. 2006;97(1):56–70.
- 18↑
Choi YH, Choi JH, Oh JW, Lee KY. Calmodulin-dependent kinase II regulates osteoblast differentiation through regulation of Osterix. Biochem Biophys Res Commun. 2013;432(2):248–255.
- 19↑
Rohde CH, Taylor EM, Alonso A, Ascherman JA, Hardy KL, Pilla AA. Pulsed electromagnetic fields reduce postoperative interleukin-1β, pain, and inflammation: a double-blind, placebo-controlled study in TRAM flap breast reconstruction patients. Plast Reconstr Surg. 2015;135(5):808e–817e.
- 20↑
Harper WL, Schmidt WK, Kubat NJ, Isenberg RA. An open-label pilot study of pulsed electromagnetic field therapy in the treatment of failed back surgery syndrome pain. Int Med Case Rep J. 2014;8:13–22.
- 21↑
Bodamyali T, Bhatt B, Hughes FJ, et al. Pulsed electromagnetic fields simultaneously induce osteogenesis and upregulate transcription of bone morphogenetic proteins 2 and 4 in rat osteoblasts in vitro. Biochem Biophys Res Commun. 1998;250(2):458–461.
- 22↑
Zhuo X, Li C, Li B, et al. Effects of combined magnetic fields treatment and nano-hydroxyapatite coating on porous biphasic calcium phosphate bone graft in rabbit spinal fusion model. Spine (Phila Pa 1976). 2018;43(11):E625–E633.
- 23↑
Stippick TW, Sheller MR. Combined magnetic fields provide robust coverage for interbody and posterolateral lumbar spinal fusion sites. Med Biol Eng Comput. 2016;54(1):113–122.
- 24↑
Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.
- 25↑
Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:I4898.
- 26↑
Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919.
- 27↑
McGuinness LA, Higgins JPT. Risk-Of-Bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments. Res Synth Methods. 2021;12(1):55–61.
- 28↑
Jenis LG, An HS, Stein R, Young B. Prospective comparison of the effect of direct current electrical stimulation and pulsed electromagnetic fields on instrumented posterolateral lumbar arthrodesis. J Spinal Disord. 2000;13(4):290–296.
- 29↑
Marks RA. Spine fusion for discogenic low back pain: outcomes in patients treated with or without pulsed electromagnetic field stimulation. Adv Ther. 2000;17(2):57–67.
- 30↑
Coric D, Bullard DE, Patel VV, et al. Pulsed electromagnetic field stimulation may improve fusion rates in cervical arthrodesis in high-risk populations. Bone Joint Res. 2018;7(2):124–130.
- 31↑
Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am. 2003;85(1):1–3.
- 32↑
Patel VV, Billys J, Okonkwo DO, He DY, Ryaby JT, Radcliff K. Three- and 4-level lumbar arthrodesis using adjunctive pulsed electromagnetic field stimulation: a multicenter retrospective evaluation of fusion rates and a review of the literature. Int J Spine Surg. 2021;15(2):228–233.
- 33
Massari L, Brodano GB, Setti S, et al. Does capacitively coupled electric fields stimulation improve clinical outcomes after instrumented spinal fusion? A multicentered randomized, prospective, double-blind, placebo-controlled trial. Int J Spine Surg. 2020;14(6):936–943.
- 34↑
Bose B. Outcomes after posterolateral lumbar fusion with instrumentation in patients treated with adjunctive pulsed electromagnetic field stimulation. Adv Ther. 2001;18(1):12–20.
- 35↑
D’Oro A, Buser Z, Brodke DS, et al. Trends and costs of external electrical bone stimulators and grafting materials in anterior lumbar interbody fusion. Asian Spine J. 2018;12(6):973–980.
- 36↑
Akhter S, Qureshi AR, Aleem I, et al. Efficacy of electrical stimulation for spinal fusion: a systematic review and meta-analysis of randomized controlled trials. Sci Rep. 2020;10(1):4568.
