Address for Correspondence Dr. Martin P. Mintchev, Department of Electrical and Computer Engineering, University of Calgary, 2500 University Drive, NW, Calgary, AB, Canada T2N1N4. Tel: 403 220 5309; fax: 403 282 6855; e-mail: firstname.lastname@example.org
Background Gastric electrical stimulation (GES) is an avenue for treating gastroparesis and obesity by controlling gastric motility using electrically mediated gastric contractions. Neural gastrointestinal electrical stimulation (NGES) is a GES modality capable of producing strong lumen-occluding local gastric contractions. Conversely, EnterraTM Therapy, a commercial implantable gastric electrical stimulator, has been utilized to treat symptoms of gastroparesis, but its nominal electrical parameters are not capable of generating lumen-occluding contractions. However, comparative studies between these two stimulation modalities are lacking.
Methods Strain gauge transducers complemented by endoscopic monitoring have been utilized to register gastric contractions invoked with NGES and Enterra neurostimulators in four acute dogs. Mucosal and serosal electrode implantations, ‘nominal’ and ‘maximum’ electrical parameters, and longitudinal and transverse electrode placements have been tested with each neurostimulator type.
Key Results Strong lumen-occluding, circumferential contractions were induced with a wide variety of NGES parameters utilizing both transverse and longitudinal electrode configurations from the serosal side of the stomach. Similarly, local gastric contractions were observed with the Enterra neurostimulator programmed at its ‘maximum’ electrical parameters but only when utilizing transverse serosal electrode implantation. Under ‘maximum’ electrical parameters Enterra was not capable of producing registerable gastric contractions with longitudinally implanted serosal electrodes. Mucosal electrode implantations did not result in GES-invoked gastric contractions in both stimulation modalities.
Conclusions & Inferences Enterra Therapy is capable of producing gastric contractions under ‘maximum’ parameters and transverse electrode configuration. Neural gastrointestinal electrical stimulation produces stronger, lumen-occluding contractions under a wider range of electrode configurations and parameters.
Gastroparesis, obesity, and gastric electrical stimulation (GES)
Gastroparesis is a stomach disorder in which emptying of content is abnormally delayed.1 Associated symptoms may include nausea, discomfort, fullness, early satiety, and bloating.1,2 Among the available approaches to treat this disease, gastric electrical stimulation (GES) has been proposed as a promising alternative to treat the symptoms and effects of gastroparesis. Three separate GES modalities have been proposed, each utilizing electrical pulses aimed at achieving three distinct results: (i) slow wave entrainment,2 (ii) low-energy, high-frequency stimulation (12 cycles-per-minute trains of 0.1 s duration, with 14 Hz embedded pulses in each train),3 which is believed to override the afferent vagal transmission,1 and (iii) induction of microprocessor-controlled distally propagated antral muscle contractions promoting gastric emptying.4–6
Separately, GES has been suggested as a treatment option for addressing the worldwide epidemy of obesity7 by inducing retrograde antral contractions that delay gastric emptying resulting in early satiety, thus promoting the reduction of the volume of food intake, consequently leading to a controllable weight loss.8–12
EnterraTM Therapy (Medtronic, Minneapolis, MN, USA) is a pacemaker-type implantable neurostimulator that has been shown in several studies to reduce nausea and vomiting associated with gastroparesis.13,14 However, due to equivocal results in the only fully published randomized controlled trial,14 it has not obtained full FDA approval but is available as a humanitarian use device. Typical therapy electrical parameter settings include frequencies of 14 Hz, higher than the intrinsic human gastric slow wave (2–4 cycles-per-minute), low energy levels (330 μs pulse width) and voltage sets to generate currents of 5 mA.15 Despite its effects on some of the symptoms of gastroparesis, no strong gastric contractions have been achieved with this GES modality and its mechanism of action remains unclear.16
Neural gastrointestinal electric stimulation (NGES), possibly the most promising GES modality,17,18 has been proven to produce strong lumen-occluding retrograde or antegrade contractions capable of moving gastric content.5,6 Neural gastrointestinal electric stimulation is based on high-energy (up to 20 V peak-to-peak), high-frequency (over 40 Hz) bipolar voltage waveforms which are mediated via cholinergic neurotransmitters and cause repeatable local circumferential contractions.4 The production of sequential contractions opens clear possibilities for the treatment of both gastroparesis and obesity.19 Unfortunately, comparative gastric motility studies of EnterraTM Therapy, GES, and NGES are completely lacking, both on experimental animals and on humans.
The aim of the present study is to compare the capabilities of invoking gastric contractility of the EnterraTM Therapy implantable neurostimulator and a custom-made NGES implantable stimulator utilizing a wide range of stimulating parameters and various electrode configurations both from the mucosal and the serosal side of the stomach in acute canine models. Although of limited number of animals and in an acute experimental setup, the study can be regarded as the first step toward similar comparative studies on chronic animals and humans.
Materials and methods
Four adult mongrel dogs, two male and two female (32.8 kg, 30.0 kg, 20 kg and 19.3 kg) underwent abdominal surgery under sodium thiopental induction (Thiotal 15 mg kg−1 intravenous (i.v.), Vetoquinol Canada, Lavaltrie, QC, Canada) and were maintained with 1–3% isoflurane inhaled (Halocarbon Laboratories, River Edge, NJ, USA). Gastroscopy for mucosal electrode placement, internal gastric motility recordings, and images was performed with a Pentax EG 3430K flexible endoscope (Hoya Corporation, Tokyo, Japan). At the end of the experiment the four animals were euthanized with an i.v. injection of Euthanyl, 480 mg 4.5 kg−1 (Bimeda-MTC Animal Health Inc., Cambridge, ON, Canada). This research was approved by the Life and Environmental Sciences Animal Care Committee, University of Calgary, Calgary, AB, Canada.
Electrode placement and configuration
Medtronic’s 6416-200 cm temporary cardiac pacing leads (Medtronic, Minneapolis, MN, USA) were utilized for the electrical stimulation. These leads are used experimentally and clinically for gastric mucosal stimulation applications.20 Each pacing lead, L1 and L2, comprises of two electrodes spaced 1 cm apart. One electrode is shaped as a fixed helix tip which screws into the gastric wall, while the second ring-type electrode makes contact with gastric mucosa by securing it with metallic clips. Although different leads are used in clinical serosal implantation applications, the Medtronic 6416 leads were used in identical configuration to replicate that approach, with the lead tunneled subserosally, the helix screwed into the gastric wall and the second lead affixed in place with surgical thread.
Electrode sets were implanted on both serosal and mucosal sides of the stomach. On both sides of the stomach (serosal and mucosal) two electrode configurations were used.
The first was a modified Enterra longitudinal single-lead implantation in which both lead electrodes were implanted.21 In the present study, the implantation was performed 2–3 cm from the pylorus on the lesser curvature of the stomach (Fig. 1A). Although standard Enterra electrode implantations are 9.5–10.5 cm from the pylorus and on the greater curvature of the stomach,14 we aimed at implanting the electrodes in a way that permitted comparative Enterra-NGES study, assuming that: (i) in the entire antrum the production of contractions resulting from electrical stimulation depends on the electrode positioning and the inter-electrode distance, not on the electrode location per se,5,6 and (ii) the production, or the lack of local distal contractions in the pre-pyloric region is of pivotal importance not only for propulsive3,22,23 but also for retrograde control.11,19
The second configuration consisted of a transverse electrode implantation which is standard for NGES and was placed 2–3 cm proximally from the pylorus both on the lesser and on the greater curvatures. For this transverse serosal electrode implantation two separate Medtronic’s 6416 electrode leads were used, where only the fixed helix tip of each one was implanted while the ring-type electrode was isolated and not utilized (Fig. 1B). Transverse electrode implantation has been shown to be pivotal for producing strong, local, circumferential, lumen-occluding contractions using NGES most likely due to the fact that the depolarization wave propagates much faster circumferentially than longitudinally.24 Therefore, transverse positioning of the NGES stimulating electrodes most likely facilitates such rapid circumferential depolarization resulting in the production of local circumferential contractions which has only limited scope in its longitudinal vicinity.
For the mucosal implantations, the electrode leads were guided endoscopically through the esophagus and out the mouth to connect to the stimulator, but otherwise the same electrode combinations and locations were followed as in the serosal implantations.
Polarity, order, and connection security were carefully observed and monitored throughout the experiments. To insure that the electrodes were implanted correctly and no short or open circuit conditions were present, the impedance measurement functions available in both the Enterra Therapy device and the custom-made NGES unit were utilized. All Enterra configurations had tissue impedance within the suggested ranges and mucosal and transverse configurations had impedances ranging from 784 Ohm to 1603 Ohm, which was deemed a sufficient indication of correct and secure electrode implantation.
Neurostimulators and electrical parameters
EnterraTM Therapy Model 3116 neurostimulator (Medtronic) was tested with two sets of electrical parameters (Table 1). A frequency of 14 Hz, pulse width of 330 μs, 0.1 s ON-time and 5.0 s OFF-time were considered the ‘nominal’ parameters for Enterra gastroparesis treatment (Fig. 1C).22 To have a proper comparison with the NGES neurostimulator, the voltage amplitude was set at 10 V. The second set of electrical parameters used for the Enterra neurostimulator has been dubbed ‘maximum’ and has been considered as the closest to the NGES parameters known to produce contractions.5,6 Enterra Therapy Model 3116 allows a maximum of 450 μs pulse width with the remaining parameters being configurable to resemble the NGES ones (see Table 1).
*5 V amplitude utilized as nominal electrical parameter for NGES longitudinal implantation.
Nominal electrical parameters
Maximum electrical parameters
A custom-designed NGES implantable-size neurostimulator was utilized to produce standard trains of high-energy, high-frequency; bipolar stimulating waveforms (Fig. 1D). Similar to the Enterra device, the NGES neurostimulator had wireless transmitting/receiving capabilities and was monitored through a custom-made PC software interface. Nominal parameters that have been proven to produce contractions were utilized including a frequency of 50 Hz, pulse width of 10 ms, 6.0 s ON-time, 15.0 s OFF-time, and 7-V amplitude.5 In the case of longitudinally implanted electrodes, 5-V amplitude was utilized (see Table 1).
Three 90W24 strain gauge transducers (RB Products, Stillwater, MN, USA) designed for gastrointestinal (GI) motility studies were surgically sutured on the serosal side of the stomach.25 These transducers measured the relative force exerted on them by gastric contractions and the resulting proportional signals were amplified by a custom-made bridge amplifier and subsequently digitized by a data acquisition system based on PCMCIA card DAQCard-AI-16XE-50 (National Instruments, Austin, TX, USA), combined with a custom-designed signal processing and visualization software. As the measured force was relative to the initial tension of the transducer at implantation, and this initial tension was not known, the force measurements were expressed in relative units (r.u.) and only comparative force dynamics was studied, with the NGES maximum contraction strength considered 100% when an invoked contraction was observed under a given electrode configuration and stimulation pattern. The first transducer (FT1, Fig. 1) was placed at the same distance from the pylorus (2–3 cm) as the stimulating electrodes and was positioned as close as possible to them. The remaining force transducers, FT2 and FT3, were implanted proximally along the gastric axis with a separation of approximately 3 cm between them (see Fig. 1) to monitor retrograde propagation of contractions, if any. Antegrade propagation of contractions was seldomly observed using NGES, if at all, and for this reason pivotal concept of NGES has been the sequential antegrade propagation of contractions by applying time-shifted electrical stimuli to sets of transversely implanted electrode sets in sequential fashion starting from the most proximal one.4–6 For this reason, we did not research antegrade propagation of contractions any further. With the advent of modern microcontroller techniques and wireless control, the production of local contractions is completely sufficient for controlling gastric motility, as adding more channels to obtain such contractions at various locations and synchronizing them in time does not present any technological challenge whatsoever.6
The experiment was divided in two parts. The first part evaluated longitudinal electrode implantations on both serosal and mucosal sides of the stomach. The second part evaluated invoked gastric contractility with transverse electrode implantations. For each combination, several sets of electrical parameters were tested (see Table 1). Each combination was repeated at least five times to confirm results. Gastric contractility was measured by the time plots generated by the force transducers, with qualitative assessment by endoscopic video from the gastric lumen.
Longitudinal electrode implantations for Enterra Therapy
Force transducer recordings for the standard single-channel Enterra Therapy longitudinal implantation are shown in Fig. 2. Under nominal stimulation parameters (Fig. 2A and B) neither serosal nor mucosal stimulation showed evidence of contractions. Periodic waveforms seen on the force transducer recordings were of respiratory frequency, demonstrating the excellent sensitivity of the contractile monitoring. The force transducer scale (in relative units) typically shows contractions above 0.1 r.u. The results were expected, as contractions with Enterra-based neurostimulation had not been reported under nominal electrical parameters.26 Under maximum electrical parameters, weak (<0.2 r.u.), local, non-lumen occluding contractions in the vicinity of the electrodes were evidenced when stimulating from the serosal side (Fig. 2C). No contractions were observed from the mucosal side (Fig. 2D).
Longitudinal electrode implantations for NGES
Neural gastrointestinal electrical stimulation utilizes high energy bipolar waveforms (up to 20 V peak-to-peak and 10 ms pulse width) to induce gastric contractions. Lumen occlusion was previously observed using transverse electrode configurations.19 When electrodes were implanted longitudinally at a distance of 1 cm between them, the effective tissue impedance was lower compared to a transverse implantation. For this reason, a reduced 5 V maximum voltage was utilized to limit the current flow and avoid any tissue damage. All other electrical parameters were not modified. Fig. 2E depicts the contractile effect of single-channel longitudinal electrode implantation utilizing NGES waveforms. Fairly strong contraction (above 0.5 r.u.) is evidenced in FT1. The effect of the contractions lasted for approximately 10 s. Furthermore, these induced contractions affected FT2 as they propagated retrogradely with a diminished strength. No effect of such contractions was recorded by FT3, indicating a limited scope of the retrograde propagation. Similar to the Enterra case, no contractions were observed with NGES stimulation on the mucosal side using longitudinal implantations (Fig. 2F).
Transverse electrode implantations for EnterraTM Therapy
Transverse electrode implantations are standard for NGES,4 but for the first time Enterra nominal parameters were tested under such electrode configurations. Fig. 3A depicts the effect of a single-channel transverse serosal implantation and, as can be observed, no contractions were generated. Identical results were obtained when attempting stimulation from the mucosal side (Fig. 3B). After this attempt, ‘maximum’ parameters were applied by re-programming the Enterra neurostimulator. FT1 in Fig. 3C clearly illustrates that contractions did occur. Relatively strong contractions reaching 0.6 r.u. took place immediately after the stimulation was delivered. Proximal propagation of these invoked contractions was less evident. FT2 did not register definitive contractions, and none was recorded by FT3. The situation changed when mucosal side stimulation was administered, again with ‘maximum’ parameters. As seen in Fig. 3D, no contractions occured and this is yet another evidence for the lack of contraction effectiveness from the mucosal side observed in the previous tests.
Transverse electrode implantations for NGES
As mentioned, transverse serosal electrode implantations in NGES have been considered standard,4 but in this study we performed it for comparative purposes. The results are depicted in Fig. 3. Stimulation results using NEP from the serosal side are presented in Fig. 3E. Similar to previously reported results,25,27 strong local contractions occurred in the vicinity of the stimulating electrodes. FT1 read the strongest contraction and it was partially reflected by FT2, indicating a limited retrograde propagation. The effect diminished proximally and no readings were observed by FT3. Once again, mucosal stimulation produced no effect on the contractility of the tissue (Fig. 3F). When ‘maximum’ parameters were used for serosal stimulation, the response was the strongest compared to all previous tests (Fig. 3G). The contractile force increased up to 1 r.u. (full scale) in FT1. A strong contractile effect can also be seen on FT2. However, the contractions were not reflected by the most proximal FT3. Mucosal implantation stimulation response is depicted in Fig. 3H. As it had been the case with all previous mucosal tests, no contractile effect was observed.
A summary of the results is presented in Table 2. In each electrode configuration and stimulation modality, the averaged NGES maximum contractions (if present) were normalized to 100%.
Table 2. Contraction strength summary in percent (100%: strongest contraction). Note that stimulations with maximum NGES parameters for the longitudinal implantations were not performed to avoid burning of the muscle due to the short (1 cm) interelectrode distance
Several studies have demonstrated the effectiveness of NGES to produce lumen-occluding local gastric contractions.4–6,25,27 The sequential generation of such contractions has been shown to control the movement of content through the stomach by either propelling it distally or in a retrograde fashion. These capabilities could be applied for the treatment of gastroparesis or obesity.4–6,11,19 The present study has confirmed the production of local gastric contractions utilizing high-frequency high-energy NGES electrical parameters under transverse serosal implantations, but additionally, it has shown that NGES is capable of producing strong local contractions utilizing longitudinal implantations (1 cm separation between electrodes). This opens the possibility of applying NGES therapy using available and tested electrodes in longitudinal configurations.
A previous study3 has reported an increased mechanical activity in the stomach resulting from the utilization of what later became known as ‘nominal’ Enterra parameters. In the present study, we could not confirm these findings under any electrode configurations (see Table 2). However, Enterra Therapy has never been tested under a combination of electrical parameters as close as possible to the ones utilized by NGES in combination with transverse electrode implantations. This combination has been applied in the present study and has demonstrated that it is possible to repeatedly generate contractions using the Enterra implantable device. This opens the possibility of utilizing a clinically available neurostimulator not only for the treatment of symptoms of nausea and vomiting, but as a potential controller of gastric content movement as well, providing it has at least two transverse electrode channels capable of being sequentially programmed at maximal Enterra parameters in either retrograde or antegrade timing sequence.
However, it has to be noted that the present study is of a limited number of animals in an acute canine model. Much more research is needed in chronic animals and on humans to confirm these findings. Certainly, a study of gastric emptying and/or gastric content retainment using multichannel Enterra Therapy with transverse electrode implantations, and multichannel NGES in chronic human experiments would be pivotal in elucidating the real capabilities and limitations of both technologies as gastric motility controllers.
Serosal electrode placement
Optimal serosal NGES electrode placement has been investigated by several previous studies.4–6 Transverse electrode placement was determined to be the best electrode configuration to produce lumen-occluding gastric contractions. These contractions occur locally and circumferentially, which can then be combined with multisite, multichannel sequential stimulation to mimic gastric peristalsis and to control content movement. Enterra Therapy has traditionally used serosally placed longitudinally implanted electrodes to achieve gastroparesis symptom relief.22,28,29 The use of transverse implantations in the present study has demonstrated the capability of the Enterra neurostimulator to produce contractions with its ‘maximum’ electrical parameters. The use of such parameters under longitudinal electrode implantations did not yield strong contractions (see Fig. 2C), confirming the advantage of transverse electrode implantations. Serosal-side implantations also demonstrated the suitability of Medtronic’s 6416 electrodes for short term gastric stimulation from the serosal side.
Mucosal electrode placement
No combination of electrode implantations, neurostimulators, or electrical parameters yielded any gastric contractions from the mucosal side. This was despite optimal conditions for lead placement. Possible reasons for the lack of tissue response may be the inability of the helix-tip electrode to penetrate far enough through the mucosa into the gastric muscle layers, or alternatively the ring-type electrode may not have reliably apposed the mucosa, despite the mucosal clips in this region. This brings into question the clinical utility of this approach as a temporary stimulation method in gastroparesis patients using this lead. As Medtronic’s 6416 leads are intended for temporary intracardiac pacing, an electrode design geared specifically toward mucosal gastric implantation will have to be devised to insure contact with the muscularis. Once such electrode is tested, the effectiveness of GES from the mucosal side will have to be re-examined.
Impact of electrical parameters
Electrical parameter selection plays a fundamental role in achieving effective and controllable gastric contractions by GES. High-energy, high-frequency NGES electrical parameters have been tested over the years and proven to produce strong lumen-occluding contractions.18 When working at a frequency of 50 Hz, on-time duration of 6 s, voltages of up to 20 V peak-to-peak, and a pulse width of 10 ms, the energy delivery to the gastric tissue is comparably larger to the energy delivery of the nominal Enterra parameters (See Table 1). This is not necessarily a drawback for the Enterra Therapy as applied in the present study, because smooth muscle fatigue observed with the abrupt, high-energy NGES11 could probably be avoided, or at least minimized using the significantly gentler Enterra parameters in transverse electrode configurations. This, in fact, can enhance the utility of the Enterra Therapy from providing purely antiemetic effect,26 to adding content-propulsion and gastric motility enhancement for the treatment of gastroparesis, or content-retainment and gastric motility impediment for the treatment of obesity. Conversely, the study not only confirms the ability of NGES to act as a controller of GI motility, but also demonstrates that longitudinal serosal electrode implantations utilizing NGES parameters can produce curvature-localized contractions depending on the area of implantation (lesser or greater curvature).
After testing the Enterra neurostimulator with NGES-like parameters it becomes clear that the frequency (50 Hz) plays a major role in determining if a contraction would occur and its strength. The fact that the maximum Enterra pulse width (450 μs) is roughly 20 times shorter than the NGES pulse width did not preclude the occurrence of contractions. Another notable difference is the fact that Enterra electrical waveforms are almost monopolar and the NGES waveforms are fully bipolar. On the other hand, NGES is capable of producing much stronger propagating contractions due to longer pulse widths compared with Enterra. This is shown in the case of longitudinal stimulation, where NGES was capable of delivering strong contractions despite the fact that the maximum amplitude voltage was lowered to 5 V. Enterra was not capable of delivering a convincing contraction under the same electrode placement despite setting the maximum voltage amplitude to 10 V.
Although performed on a limited number of animals in an acute experimentation setup, the present study clearly demonstrates that local gastric contractions can be generated with the Enterra Therapy neurostimulator using NGES-like electrical parameters and transverse serosal electrode implantations. Evidence has been presented as well that strong contractions can be induced with NGES therapy utilizing longitudinal serosal electrode implantations. Suitable electrodes have to be developed to invoke contractions by mucosal-side electrical stimulation both with NGES and Enterra Therapy. Further studies are needed, first on chronic animals, and later on humans, to confirm these findings in the long-term.
It is important to note that the NGES portion of the present study was performed utilizing a custom-designed, wirelessly programmable, implantable neurostimulator, which in the process of testing performed excellently and can be considered ready for future long-term chronic experiments both on animals and on humans.
This work was supported in part by Natural Sciences and Engineering Research Council of Canada.