Introduction

Detailed descriptions of the enteric nervous system began with research performed during the nineteenth century. Leopold Auerbach, Georg Meissner, Santiago Ramon y Cajal, and Camillo Golgi are just a few of the scientists who laid the foundations for our current understanding of the enteric nervous system. However, it was two British scientists, William Bayliss and Ernest Starling, who first described the functional implications of the enteric nervous system and have been credited with introducing the field of neurogastroenterology [1, 2]. Their landmark publication appeared in the Journal of Physiology in May of 1899. The opening line of this manuscript remains as true today as it did at the time of publication: “On no subject in physiology do we meet with so many discrepancies of fact and opinion as in that of the physiology of the intestinal movements” [3]. Despite a century of continued research and progress, aspects of enteric neurophysiology and pathophysiology remain incompletely understood and hotly debated. Gastroparesis and gastric electrical stimulation (GES) for the treatment of gastroparesis are no exceptions.

We assume that an understanding of normal physiology and pathophysiology are necessary before discovering viable treatments for diseases. However, the history of medicine is ripe with instances in which successful treatments are discovered before a full understanding of the pathophysiology and treatment mechanism(s) have been achieved. Such is the case for gastroparesis, as well as one of its treatment options, GES. In this chapter, we will examine what is known and what remains uncertain regarding the mechanism and clinical effects of GES.

Gastric Electrical Stimulation Device

There is only one Food and Drug Administration (FDA) approved device for the purpose of GES. The device is called the Enterra Therapy System and is manufactured by Medtronic Inc. On September 23, 1999, the Enterra Therapy System was granted Humanitarian Use Device (HUD) designation (HUD #990014) by the FDA. Currently, HUD designation may be granted by the FDA when a medical device is intended to treat patients for a condition that affects fewer than 8000 individuals in the United States annually. However, at the time of HUD designation for the Enterra Therapy System, the FDA HUD designation required that the device be intended to treat patients for a condition that affects fewer than 4000 individuals in the United States annually.

On March 30, 2000, the Enterra Therapy System was granted a Humanitarian Device Exemption (HDE) . When the HDE is granted, a medical device is exempt from the usual FDA requirements demonstrating efficacy for its intended purpose. However, approval does require evidence that the device does not pose unreasonable risk of injury or illness and that the potential health benefit outweighs the risks. This FDA designation is, therefore, offered in instances where it is difficult to gather enough clinical evidence to meet the usual FDA standards due to the relatively low incidence of disease. Devices with the HDE may only be used under an Institutional Review Board (IRB) approval and only for the FDA-approved indication. In the case of the Enterra Therapy System, the FDA-approved indication is “for the treatment of chronic, intractable (drug refractory) nausea and vomiting secondary to gastroparesis of diabetic or idiopathic etiology.”

Gastroparesis Pathophysiology

Gastroparesis is a chronic clinical condition with a pathophysiologic mechanism that is incompletely defined. Gastroparesis has been defined as “a chronic symptomatic disorder of the stomach characterized by delayed emptying without evidence of mechanical obstruction” [4]. Absent from this accepted definition is a description of the pathophysiologic mechanism for delayed gastric emptying. Although diabetes has been identified as one of the causes of gastroparesis, there are many patients with gastroparesis in whom the cause is uncertain and the condition is labeled idiopathic gastroparesis. There are a variety of methods that have been utilized to assess the presence and severity of gastroparesis, including patient-reported symptoms, gastric function, cellular changes in the stomach and pylorus, gastric myoelectric changes, and response to treatment. Some intriguing associations have been identified, but the interplay between cellular defects, myoelectric alterations, gastric function, symptoms, and response to treatment is poorly understood and remains incompletely described. The diagnosis, therefore, remains based on the presence of typical gastroparesis symptoms in addition to gastric dysfunction (delayed gastric emptying) in the absence of mechanical obstruction.

Understanding gastric motility begins with one necessary consideration: gastric motility is exceedingly complex. On a functional level, gastric motility involves many unique coordinated sensory and motor functions. Motor functions within the organ include receptive relaxation, fundic contraction, antral peristalsis, pyloric function, and antropyloroduodenal coordination. These actions are not necessarily continuous and must be controlled and coordinated at appropriate times. There are a variety of cellular gastric motility mediators involved in these coordinated actions, including smooth muscle, myenteric plexuses (enteric or intrinsic nervous system), interstitial cells of Cajal (ICC) , and autonomic (extrinsic) nervous system. There are also both excitatory and inhibitory stimuli, such as acetylcholine, VIP, carbon monoxide, Substance P, and nitric oxide [5, 6]. The normal physiologic impact of these gastric motility mediators can be influenced by other local factors, such as the enteric microbiome, stromal fibrosis, macrophages, and the infiltration of inflammatory cells [7–9]. When the cellular mediators of gastric motility are disrupted, gastric dysrhythmias can occur, as can dysfunction in gastric accommodation. These abnormalities result in dysfunction and/or discoordination of gastric motility, which subsequently produces a myriad of symptoms associated with gastroparesis. Complicating this story is that multiple mechanisms of cellular injury likely exist. This is suggested by the fact that there are variable histologic findings and there is variable treatment efficacy between diabetic and idiopathic gastroparesis.

Investigators studying disorders of gastrointestinal motility have recently focused much of their attention on ICC. Many believe that the ICC represent the most likely cellular source of gastric dysfunction causing gastroparesis. The ICC are considered gastric pacemakers and are responsible for generating and propagating the gastric electrical slow waves [7]. ICC branches form gap junctions with smooth muscle cells. The slow waves occur at a frequency of three cycles per minute and produce phasic coordinated smooth muscle contractions that propagate from the gastric body to the pylorus. These antral contractions serve to grind and mix [10].

This depletion of ICC has also been associated with the disruption gastric slow waves, resulting in gastric dysrhythmias [6, 7, 11]. The most commonly identified is tachygastria [12]. Disruptions of both slow-wave conduction and propagation have been associated with gastroparesis [7].

Recent research has identified histologic changes associated with gastroparesis that implicate ICC. Full thickness gastric biopsies have shown a depletion of ICC associated with both diabetic and idiopathic gastroparesis [13, 14]. This loss of ICC was also shown to be associated with delayed gastric emptying in patients with diabetic gastroparesis. However, these same studies found inconsistencies regarding the role of ICC in gastroparesis. For example, ICC were not associated with symptom severity in diabetic gastroparesis. Additionally, the loss of ICC was not associated with delayed gastric emptying or symptom severity in idiopathic gastroparesis. Other studies have not shown any association between ICC loss and gastric emptying or symptom severity [11]. These findings suggest that ICC may have a role, but there are likely other players involved as well.

ICC depletion has been identified by several research groups, but there are other cellular changes that may also alter ICC-mediated slow-wave initiation and propagation. These include thickened basal lamina around the smooth muscles and nerves, stromal fibrosis, and enteric nerve damage or loss [14]. The cellular injury pattern also seems to differ between patients with diabetic versus idiopathic gastroparesis, lending more evidence that the mechanism of injury is likely different between the two conditions [14]. There is a small but growing body of evidence suggesting that these cellular changes may be immune mediated [5, 8].

The clinical severity of idiopathic gastroparesis has been shown to be associated with myenteric plexus immune infiltrate [5]. ICC loss and stromal fibrosis within the smooth muscle layer have been documented in both the gastric antrum and pylorus, suggesting that both gastric and pyloric functions are likely impaired and contributing to gastroparesis. The degree of ICC loss and the presence of fibrosis in the pylorus have recently been identified as being associated with gastroparesis symptom severity. In fact, the degree of ICC loss (70.5% of patients with gastroparesis) and fibrosis (82% of patients with gastroparesis) was more prominent in the pylorus than the gastric antrum [12, 15]. This suggests that pyloric dysfunction may be underappreciated as a contributor to symptoms of gastroparesis.

As an additional layer of complexity, full-thickness gastric biopsy specimens have shown a lack of antiinflammatory macrophages and/or a lack of protective enzymes against oxidative stress, which are expressed by macrophages [8, 9, 16, 17]. Although inflammatory mediators appear to play a role in the pathogenesis of gastroparesis, the interplay between unique or excessive inflammatory insult versus loss of the native antiinflammatory mechanisms has not been delineated.

Although the details remain elusive, a framework for the pathophysiology of gastroparesis appears to be emerging. A pathologic immune-mediated response is triggered against the gastrointestinal myoelectric network. Cellular damage disrupts the generation, propagation, and conduction of electrical activity, which subsequently impairs gastric function. Cellular damage may also distort gastric sensory function. Sensorimotor dysfunction ultimately produces a variable constitution and severity of symptoms.

GES Mechanism of Action

There have been two proposed methods for gastric electrical stimulation. One method has been termed gastric pacing . This type of GES has never been approved for clinical treatment. The method involves gastric electrical stimulation that mimics the native gastric slow waves, which are generated and propagated by the ICC. Entrainment of slow wave requires high energy, synchronized, low frequency cycles of long pulses (10–600 ms). These parameters result in gastric smooth muscle contractions. Several barriers have emerged regarding the application of this GES method. The energy needed to generate these waves, especially with a single-channel pacing method , is quite high and requires an external power source that has limited the ability to utilize an implantable generator. Additionally, multiple precisely located leads are required to generate an appropriate wave that propagates in the correct direction and for an adequate distance. Initial studies in canines and humans have shown entrainment of slow waves, normalization of gastric dysrhythmias, and improvement in gastric emptying [18]. In the few human studies, there was also a reduction of gastroparesis symptoms. Although this type of GES is not in clinical use, it remains under investigation as some researchers believe there may be treatment potential if the described barriers can be overcome.

The other type of GES therapy is also termed neurostimulation . The Enterra Therapy System described above is the only approved form of GES for the treatment of chronic, intractable (drug refractory) nausea and vomiting secondary to gastroparesis of diabetic or idiopathic etiology. The system produces low-energy, high-frequency, pulses of short duration (several hundred microseconds) [18]. This form of GES is termed high-frequency neurostimulation because the frequency is four times the intrinsic slow-wave frequency (12 cycles per minute compared to three cycles per minute in the case of native gastric slow waves). High-frequency GES does not entrain gastric slow waves [19] and therefore is not expected to result in dramatic gastric emptying improvements. The mechanism for high-frequency neurostimulation is not completely understood.

Some have suggested that high-frequency neurostimulation is able to modulate gastric slow waves and/or correct gastric dysrhythmias, but recent studies have not supported this theory. Initial studies suggesting the enhancement of slow-wave amplitude and propagation velocity have largely been performed with less accurate recording methods. Recent advanced high-resolution mapping data failed to show an impact of high-frequency GES on slow-wave modulation or gastric dysrhythmias [19]. It has become clear that high-frequency neurostimulation does not alter gastric muscular contraction and has no effect on slow-wave dysrhythmias.

If high-frequency neurostimulation does not seem to dramatically change gastric emptying, then how do patients benefit from this treatment? It appears that high-frequency neurostimulation alters visceral vagal afferent and efferent pathways. Spectral analysis of the heart rate variability has shown increased vagal activity, both afferent and efferent. Additionally, FDG-PET imaging indicates increased neural activity in the thalamus and caudate nuclei in response to high-frequency neurostimulation. It is postulated that neurostimulation increases vagal afferent activity, which then has an inhibitory influence on the nausea and vomiting control centers in the central nervous system. This stimulation may also have an impact on vagal efferent activity, resulting in increased gastric accommodation. It is uncertain if the stimulation of vagal afferent pathways increases the symptom threshold for gastric volume and pressure changes or if the stimulation of vagal efferent pathways enhances gastric accommodation, resulting in symptom reduction [20].

Some initial studies have shown that patients with only mild-to-moderate loss of ICC show a better symptom response to GES than patients with severe ICC loss. Additionally, patients with mild-to-moderate loss of ICC were more likely to show improved gastric emptying (23%) than those with severe loss of ICC (0%), suggesting that ICC counts may be important predictors of treatment response to GES [12].

Indications

The FDA-approved indications for GES are clearly stated. GES is approved for “the treatment of chronic, intractable (drug refractory) nausea and vomiting secondary to gastroparesis of diabetic or idiopathic etiology.” The patient must have diabetic or idiopathic gastroparesis, they must have chronic nausea and vomiting, and the symptoms must be intractable (refractory to medical management).

Although the indications are clear, patient selection for GES is quite complex. This mainly stems from the fact that the diagnosis of gastroparesis can be difficult to secure. At a minimum, the diagnosis of gastroparesis requires that a patient displays symptoms of gastroparesis, such as nausea and vomiting. Additionally, diagnosis requires objective evidence of delayed gastric emptying during an appropriately performed gastric emptying study. Appropriate diagnosis of gastroparesis, however, goes far beyond this. If providers were to only utilize these nonspecific symptoms and the gastric emptying study for diagnosis, then overdiagnosis of gastroparesis would invariably occur. Symptoms associated with gastroparesis are nonspecific and are associated with a variety of other gastrointestinal diseases, eating disorders, medication side effects, and other neuropsychiatric disorders. Unfortunately, reliable objective testing for gastroparesis does not exist. Although delayed gastric emptying may occur in the presence of these symptoms, it may not be the cause of these symptoms. Additionally, the degree of delayed gastric emptying does not correlate with the severity of gastroparesis symptoms. These patients, therefore, pose a significant diagnostic challenge.

Other causes of the symptom profile should be thoroughly investigated before settling on a diagnosis of gastroparesis. Upper endoscopy, upper gastrointestinal contrast study and computed tomography imaging should be utilized at the discretion of the treating physician to rule out mechanical obstruction and to evaluate for an alternative symptom source. Additionally, laboratory testing should be completed to evaluate for hypothyroidism and for diabetes. If diabetes is a known diagnosis, then glycemic control should be evaluated. Achieving glycemic control can result in improved gastroparesis symptoms. However, sometimes gastroparesis can be the underlying reason for poor glycemic control due to vomiting and irregular eating habits. The treating physician should consider eating disorders as a possible source of symptoms. Additionally , chronic narcotic use, cannabinoid hyperemesis syndrome, and cyclic vomiting syndrome should be on the differential as well.

Another criterion cited in the FDA indications is that the symptoms should be intractable (refractory to medical management). Attempts at modifying dietary and eating behavior should be documented, as should attempts to control symptoms with antiemetic medications and motility agents. Failure of medications may be due to poor efficacy or intolerable side effects.

The Enterra Therapy System is not magnetic resonance imaging (MRI) compatible. If the patient has an anticipated need for an MRI, this therapy should not be offered. Diathermy (e.g., shortwave diathermy, microwave diathermy, or therapeutic ultrasound diathermy) is another contraindication to implantation of the Enterra Therapy System. Energy can be transferred through the components, which can cause damage to the device or damage to tissue, resulting in severe injury or death.

Implantation

The Enterra system consists of a battery pack and two neurostimulation leads. It is implanted by a trained physician, and the procedure is performed under a general anesthetic. The procedure can be performed utilizing an open or minimally invasive approach. The leads are implanted onto the stomach 10 cm proximal to the pylorus along the greater curvature. The leads should be 1 cm apart in a parallel fashion. The depth of the leads should be into the muscularis propria of the gastric wall. The leads should not penetrate the gastric mucosa, and this should be confirmed with endoscopy at the time of implantation. The leads are then brought out through the anterior abdominal wall and tunneled into a subcutaneous pocket on the chest or abdomen. The leads are then connected to the battery pack, and the battery pack is sutured into place. The device is then turned on and programmed using a remote external programming device. Patients can potentially be sent home the same day. However, anesthesia and postoperative pain medications can result in a temporary worsening of gastroparesis symptoms, so some patients may end up staying in the hospital for several days.

Results

To date, there have been three randomized controlled trials (RCT) evaluating the use of GES for the treatment of severe medically refractory gastroparesis. The initial RCT on this topic was termed the Worldwide Anti-Vomiting Electrical Stimulation Study (WAVESS) [21]. This study involved the implantation of the GES device into 33 patients with gastroparesis (17 diabetic, 16 idiopathic). Patients were then randomized to having the device either turned ON or OFF for one month. Patients from each group then crossed over for one month. All patients then entered an open label portion of the study where the device was turned ON for the next ten months. During the blinded phase of the trial, patients with the GES device turned ON versus OFF had a 50% decrease in weekly vomiting frequency (WVF) but no change in total symptom score (TSS) or health-related quality of life score (HQOL). During the open label phase of the study compared to baseline, patients reported 72% reduction in WVF and significant improvements in both TSS and HQOL. In summary, the results did not show a major difference between the ON and OFF groups during the blinded portion of the study; however, patients reported significant improvements over baseline during the open label portion. Some interpreted these results as evidence of potential for a placebo effect or regression to the mean. Others argue that the study design did not allow for maximal effect of the device during the crossover phase and that placebo effect would be unlikely to result in durable results at 12 months after implantation.

Authors of two subsequent RCTs , one for patients with diabetic gastroparesis and one for patients with idiopathic gastroparesis, attempted to remedy some of the design flaws of the original RCT [22, 23]. In these identically designed trials, all patients had the GES device turned ON for six weeks after implantation. Following the initial six weeks with the device ON, patients were randomized in a blinded fashion to having the device ON or OFF for three months. A crossover then occurred for three months. This was followed by the open label phase, which included all patients having the device ON for four months. In both studies, the patients had a significant reduction in WVF and TSS, as well as an increase in HQOL from baseline following implantation and during the final open label phase. However, there were no significant differences in these measures between the ON and OFF groups during the blinded portion of the study. Supporters of GES suggested that the initial six weeks of having the device ON resulted in the entrainment of the central control mechanisms for nausea and vomiting, thus preventing a significant recurrence of symptoms during the three-month OFF phase. They also pointed out that patients had significant and sustained improvements in symptoms following GES. Detractors continue to point out the fact that once the device has been implanted, no study has shown a difference in outcomes between patients randomized to having the device ON versus OFF.

Several meta-analyses have been conducted that include the available RCT, as well as many prospective open label cohorts. All of these analyses found improvement in TSS, nausea, vomiting, and HQOL with the open label cohorts but no significant differences between blinded groups [24–27]. This has been the source of debate regarding GES for gastroparesis. We are left with opposing data that have not yet been reconciled and have become a source of debate regarding the efficacy of GES for the treatment of gastroparesis. On the one hand, patients randomized to blinded groups having the GES device ON versus OFF show no difference in outcomes. On the other hand, open label trials uniformly show significant and durable improvements in symptoms and quality of life following GES.

There are several other unique outcomes that have been studied following GES for the treatment of gastroparesis. When examining healthcare utilization, prospective cohorts have shown reduced number of hospitalizations, fewer total hospital days, and few emergency department visits in patients with gastroparesis, who were managed with GES compared to pharmacotherapy alone [28–30]. Others have examined this from the standpoint of healthcare costs. In these studies, GES resulted in a reduction of healthcare costs to the point that it was shown to be cost-effective after two years [31].

GES has also shown great benefit in many patients with gastroparesis who require enteral or parenteral nutritional support. In one prospective cohort of patients who required enteral feedings via feeding tube (n = 99), 89% were able to eliminate the need for enteral feedings after the initiation of GES [30]. In one meta-analysis , the initiation of GES in patients requiring nutritional support (tube feed and/or TPN) resulted in 78% of them eliminating any form of supplemental nutrition [26].

In diabetic patients who have gastroparesis, blood sugar control can be extremely difficult. The gastrointestinal symptoms associated with gastroparesis (especially vomiting) can make glucose control difficult by interrupting regular and predictable eating behaviors. Several studies have shown that GES results in improvements in diabetic control. In one cohort of 17 patients with gastroparesis and diabetes, initiation of GES was associated with an improvement of mean glycohemoglobin from a baseline of 8.6% to 6.2% and 6.5% after 6 and 12 months of GES [32]. Another cohort showed a durable improvement in glycohemoglobin three years after the initiation of GES from a baseline of 9.5% to 7.9% [33].

Nearly all studies of GES for the treatment of gastroparesis show significant improvement in symptom control; however, the impact of GES on gastric emptying appears to be much less predictable and much more controversial. Although several studies have shown GES to be associated with significant improvements in gastric emptying, other studies show little to no improvements in this measure. Even when improvements in gastric emptying were noted, the mean gastric retention remained severely abnormal. Additionally, improvements in gastric emptying have not necessarily shown any correlation with symptom improvement [34]. In fact, many patients with significant improvement in symptoms and quality of life show little to no improvement in gastric emptying. These findings seem to support the idea of GES having a centrally acting mechanism rather than a mechanism that accelerates gastric emptying.

Predictive Factors

Several investigators have searched for patient factors associated with successful response to GES for gastroparesis. Although subsets of patients with diabetic or idiopathic gastroparesis have shown significant improvement in symptoms with the initiation of GES, several investigators have shown that diabetic gastroparesis was associated with a higher clinical response rate [30, 35–37].

Additionally, primary symptoms of nausea and vomiting seem to be associated with more favorable clinical response than abdominal pain [35]. That is not to say that abdominal pain does not respond to GES as several studies have shown a significant improvement in abdominal pain in response to GES [27, 38, 39]. It is simply that improvement in symptoms of abdominal pain seems to improve to a lesser degree and in fewer patients than other symptoms. Finally, several studies have shown that preoperative use of narcotic pain medication is associated with worse response to GES in most symptom domains [35, 38].

Complications

GES carries a defined risk profile of complications and adverse events that require device removal. These events include device infection, device migration, erosion of stimulator leads into the gastric mucosa, bowel obstruction, pain, and system malfunction. The rate of device removal for these types of events varies in the literature, but meta-analyses have put this number around 9% [26, 27, 30]. Another reason for device removal is lack of efficacy. Patients with inadequate symptom relief may elect to have the device removed and pursue continued medical management, or they may have the device removed have a different surgical intervention for gastroparesis, such as pyloroplasty, endoscopic pyloromyotomy, or roux-en-y bypass with or without remnant gastrectomy. This has been reported at a rate of 7% with a mean time to removal of 36 months following initial placement [40]. A more predictable need for reoperation is for battery pack replacement. The implanted battery pack has a limited life span, which has been reported to be up to ten years; however, battery life is dependent on the device settings. Once the battery wears out, an operation is required to replace the subcutaneous battery pack. What has become clear is that there is significant chance of needing another operation when selecting GES, and this should be disclosed to patients considering this option. The rate of complications requiring device removal combined with the treatment failure rate is only one part of the discussion. The other is disclosure that if the device is effective and complication free, the patient will likely need an operation for battery replacement sometime within the next decade.

Conclusions

The pathophysiology of gastroparesis and the mechanism of GES are both poorly understood. Further understanding of these pathways should shed light on our understanding of this disease and the optimal treatment approaches. The use of GES for the treatment of severe, medically refractory gastroparesis results in significant and durable improvements in gastroparesis symptoms and health-related quality of life. In addition, GES appears to be associated with reduced health care utilization, improved diabetic control, and significant reduction on the need for supplemental nutrition. There remains uncertainty in reconciling these impressive results of patient benefit with the lack of difference in outcomes between blinded groups with the device ON versus OFF. Given the reported benefits, GES remains a viable option for patients with chronic, intractable (drug refractory) nausea and vomiting secondary to gastroparesis of diabetic or idiopathic etiology when initiated under the HDE protocol with IRB approval. Patients and providers should consider device removals due to complications, treatment failures, or battery expiration as there is a fairly substantial reoperation rate that must be understood.