Experimental Evidence

The biochemical basis of the cellular response to electrical fields has been elucidated and nicely summarized by Brighton.¹⁵ Cultured bone cell DNA production increases in response to capacitive coupling, inductive coupling, or combined electromagnetic fields. These electrical fields act at cell membrane receptors, causing rapid signal transduction through PI3K and mTOR proteins.^16,17 This cellular cascade activates specific protein kinases, which tend to increase intracellular calcium concentrations, increase cytoskeletal calmodulin activation, and ultimately increase cellular proliferation. This growth and differentiation of osteoblasts are the desired effects in spinal fusion. It should be noted that the mTOR cellular cascade acts through different pathways than those activated by bone morphogenetic proteins (BMPs),¹⁸ although electrical stimulation has been shown to up-regulate multiple osteoinductive BMPs.^19,20

Much of the initial work in evaluating the use of external stimulation for the augmentation of a failing arthrodesis was performed on long-bone models. The relative success of these models was the impetus for the transfer of such technologies to use in the different arena of failed spinal fusions. Initial studies of spinal stimulation for lumbar fusions, however, were performed on non–weight-bearing models (i.e., quadrupeds, without axial forces experienced through the spine), undergoing primary spinal fusions, usually without instrumentation. There were no controls for these variables.

Nerubay et al.²¹ studied implanted direct current stimulation (DCS) of primary dorsal spinal fusions in 30 pigs and reported a customized nonparametric score improvement of radiographic and histologic criteria 2 months after the operation. Conversely, Kahanovitz et al.²² investigated the use of DCS in primary lumbar spinal fusions in dogs. They found no advantage for electrical stimulation in the radiologic or histologic indices for fusion until 12 weeks after the operation using an internal control. However, each facet joint involved was considered a statistical sample; when the animals were considered separately, there were no statistical differences at any time for fusion success. This same group reported the use of PEMF on a similar canine model.² They were unable to demonstrate radiologic or histologic differences between experimental and control groups at 6 and 12 weeks after operation. In another report, this group observed that by 12 and 15 weeks after surgery, there were no histologic or radiologic differences in dogs receiving primary spinal fusions using instrumentation and PEMF.²³

France et al.²⁴ showed the dose-dependent efficacy of subcutaneous DCS for primary, noninstrumented lumbar intertransverse process fusions in a controlled rabbit model. In a similar model, DCS increased fusion rates in a dose-dependent manner while employing hydroxyapatite as a bone graft substitute, as compared with nonstimulated rabbits with autologous bone.²⁵ The same team also demonstrated that in a similar model, DCS did not provide benefit to a nicotine-dosed group.²⁶ DCS also was studied in a randomized, controlled, and blinded (for outcome measure) sheep experiment, used in conjunction with titanium interbody fusion cages packed with autograft.²⁷ A dose-dependent rate of fusion was seen in this study and in another similar study in a dog model.²⁸

The use of various electrical stimulating devices for spinal fusion in animal models has met with mixed results. However, no animal studies have addressed the situation of failed fusion. Furthermore, the spinal biomechanics of quadrupeds are not identical to those in upright primates, whose weight bearing has direct influence on fusion success.³ Neither invasive nor external spinal stimulation has been studied adequately in an appropriate upright animal model for failed fusion of the lumbosacral spine. The use of instrumentation is a complicating factor for which there are no controlled studies with electrical stimulation in animals. Furthermore, the application of animal data to human use is controversial, because spinal fusion rates differ among species.²⁵ Thus, there is insufficient experimental evidence in animal studies to support the use of augmentative electrical stimulation for reoperation for failed spinal fusion.²³

Clinical Evidence

Early studies concerning the use of electrical stimulation centered on nonspinal applications, both for primary surgeries and for nonunions. Earlier papers dealing with electrical stimulation for spinal fusion dealt mainly with mixed populations (e.g., primary and reoperation for pseudarthrosis). More recent trends have demonstrated the use of external electrical stimulation versus implanted stimulators. There are no published studies specifically dealing with reoperation for failed lumbar spinal fusion that prospectively control for the type of surgery, construct variety, instrumentation, and patient status (e.g., weight, smoking status, steroid therapy).^10,29–31 Only one referenced study indicates a statistically significant improvement in the fusion rate for failed spinal fusion with electrical stimulation.¹² The 33 patients in that study were instrumented and were not completely randomized versus controls for other factors. None of the other studies demonstrated an improvement in the fusion rate or time to fusion in the subpopulation of failed dorsal spinal fusions. In fact, one author reported that electrical stimulation had an adverse effect on the time to fusion and stated that it “is not advocated for use in any type of spine fusion.”³ Thus, although growing, the clinical evidence in favor of electrical stimulation for the augmentation of failed dorsal spinal fusions is meager and requires further controlled investigation.¹

Multiple clinical reports have indicated the efficacy of external capacitively coupled electrical stimulation (CCES) in primary spinal fusions.²⁴ In a study of 179 patients (reduced from 337 due to withdrawals and noncompliance), Goodwin et al. did not control for variables, such as a concomitant disease (e.g., diabetes), institution, bone graft materials used for fusion, and internal fixation strategies.³² There was no statistical significance for CCES efficacy in the smoking or previous surgery cohorts for this study. Jenis et al.³³ determined that neither DCS nor PEMF was able to statistically improve the fusion rate or clinical outcome of instrumented dorsolateral arthrodeses.

A retrospectively random review of patients to analyze the efficacy of PEMF to augment primary spinal fusions was undertaken to measure clinical and radiologic outcome.³⁴ A statistically significant difference was found, with 97.6% fusion success for the stimulated group, and only a 52.6% fusion rate for the control group. However, there was a bias toward using PEMF on smokers, on multilevel fusions, and in conjunction with the allograft. Surgical approaches included ventral interbody and dorsolateral approaches, or both, thereby adding further variables to the analysis of these results. The number of patients (61) probably was too small to make meaningful conclusions from the clinical outcomes. With regard to reoperation for failed spinal fusion, the review found that “the low number of patients undergoing repeat fusion precluded specific conclusions or recommendations” for the use of PEMF.

Some prospective well-designed studies in which combined magnetic fields are evaluated for primary spinal fusions have been published.³⁵ In a multicenter, randomized, double-blind, placebo-controlled trial of 201 noninstrumented fusions, Linovitz et al. demonstrated a statistically significant increase in the fusion rate at 9 months after surgery with use of a noninvasive device. When stratified by sex, this significant difference remained only for women. There were only 28 patients who smoked in the entire starting population of 243, so no conclusions can be drawn for this subgroup. Furthermore, no conclusions could be drawn with regard to reoperation for failed spinal fusion. A study by Kucharzyk dealt with a high-risk pool of patients in whom implanted fusion stimulators provided a statistically significant benefit for radiographic and clinical fusion.³⁶ Although interesting, these results could not be used to draw conclusions concerning use of electrical stimulation for failed spinal fusion. Similar findings in a retrospective study were published by Silver,³⁷ who included primary and revision fusion cases, although not as a salvage technique. Fusion results were similar in the two groups in this particular study.

In 2008, Foley et al. prospectively studied PEMFs for higher-risk primary cervical fusion (i.e., smoking or multilevel), showing an improvement in fusion rate with stimulation at the 6-month, but not at the 12-month, mark from surgery. Prior surgery was an exclusion criterion.³⁸

In recent years, several studies of the use of PEMFs for failed spinal fusion have been done. Simmons et al.³⁹ studied patients with pseudarthrosis 9 months after initial surgery, using PEMF as a “salvage” technique. They found that such electrical stimulation (at least 2 hours per day) was effective in these cases for both dorsolateral and interbody fusion success, without any statistical differences when comparing smoking status, allograft versus autograft, instrumentation, or multilevel fusion. They reported 67% fusion success as a salvage in 100 patients in multiple sites. Such results are not inconsistent with the orthopaedic literature for nonspinal union salvages.^5,40,41

Thus, although there is mounting evidence to support the use of electromagnetic fields for primary lumbar fusions, no prospective controlled study of recalcitrant subgroups such as patients with pseudarthroses has been performed. There appears to be a trend toward fusion in the face of electrical stimulation, although the clinical evidence for use of stimulation as a standard for failed fusion has not yet been produced. The work of Simmons et al. shows efforts in the direction of demonstrating electrical stimulation as a standard treatment for failed dorsal lumbar fusion. However, a prospective, controlled, and blinded study of electrical stimulation for failed lumbar fusion is still needed.

Discussion

The lack of upright animal models underscores the importance of understanding the biomechanics involved in bone fusion.² The weight-bearing forces acting on spinal fusions in the upright position are not equivalent to those in other positions, as has been demonstrated in nonspine models.⁴⁰ Additionally, the pathology, fusion type, and use of instrumentation affect the biomechanics of a spinal fusion construct.^2,6,12,29 Fixation alone has been shown to enhance spinal fusion.^2,3,12 These variables have not been consistently controlled in the animal and patient studies published on the use of electrical stimulation for spinal fusion.^1,3,7 Furthermore, the experimental and clinical evidence supporting stimulation for bone fusion for pseudarthrosis is based mainly on long-bone and other orthopaedic data. This appendicular physiology exhibits completely different biomechanical dynamics and differs from the situation of spinal fusion, where the bones are not “normally” physiologically fused bones. In other words, it is still necessary to fuse a fracture of a long bone that has failed to heal, whereas spinal segments were not meant to be connected in an immobile fashion.

Many factors affect the spinal fusion success rate, despite the method of fusion surgery, including patient health, primary pathology, weight, weight-bearing status, steroid and nonsteroidal anti-inflammatory drug (NSAID) use, PEMF compliance, and smoking.³ None of the literature studies concerning electrical stimulation for failed spinal fusions were designed with complete control for these parameters, although multivariate analyses were performed in some.^10–12 Many authors agree that there is a paucity of appropriate controlled patient studies for electrical stimulation in spinal fusion and that further well-controlled prospective studies are warranted.^{1,3,7,10,42,43} However, there is ongoing improvement in the availability of clinical studies for the use of spinal stimulation for primary lumbar fusions and one noncontrolled study concerning the use of such stimulation for failed lumbar fusion. The reader should be careful not to extrapolate those conclusions automatically to the case of spinal pseudarthrosis, which is a more difficult and recalcitrant physiologic problem.

Furthermore, patients at high risk for fusion failure may be destined for failure despite the best surgical techniques, with or without electromagnetic augmentation.¹¹ These patients’ pseudarthroses are hypovascular and fibrotic, leading to a trend toward less successful subsequent reoperation. This positive feedback cycle toward negative outcome has not been borne out in the present literature analysis. Also, time to fusion was not measured in all studies, and the natural history of spinal fusion may be different in various patient populations for which the aforementioned study variables were not completely controlled.^10,31 These interrelated factors must be evaluated independently along with electrical stimulation to provide adequate evidence for their effective use as an augmentative device for failed spinal fusion. One author astutely reported that “good surgical technique remains the most important influence on surgical outcome.”¹²

In addition to the paucity of animal and human evidence in favor of using electrical stimulation to augment spinal fusion for pseudarthrosis, there are numerous disadvantages and potential complications. The additional time required for surgery for implantation of invasive electrical stimulators is significant and increases perioperative morbidity. Surgical use of internal stimulators provides an additional foreign body, which may increase the incidence of postoperative infection, particularly in high-risk patients with pseudarthrosis such as smokers and steroid users. The invasive stimulators would thus potentially require removal, adding further surgical risks and costs to the patient’s burdens.¹¹ Implanted leads have been shown to break and corrode on occasion.^8,44