背景与目的 部分胃轻瘫患者对常规饮食调整和药物治疗无效，属于难治性胃轻瘫。胃电刺激（gastric electrical stimulation，GES）通过产生电脉冲刺激胃窦收缩，改善患者症状，提高生活质量，但其作用机制尚未得出一致结论。有研究发现GES能够影响胃排空速度，神经系统可能参与其作用机制，但缺乏GES对摄食的影响以及对肠神经系统（enteric nervous system，ENS）特异性神经元参与机制的深入探究。本研究通过在健康比格犬体内植入国内研发的胃电刺激装置，研究高频短脉冲刺激对其摄食、饮水、体重、血清果糖胺和胃排空的影响，并监测不良事件，评估其安全性，进一步探究GES对胃肌间神经丛特异性神经元的影响。

材料与方法 采用随机交叉设计，4只健康成年比格犬随机分为A组和B组，均经小开腹手术植入电刺激装置，手术恢复后进行二轮刺激，分别为真刺激和假刺激，每轮刺激时间为4周。在每轮电刺激开始前、刺激期间测量并记录实验犬的饮水量、摄食量和体重，以刺激期间饮水量、摄食量与刺激前的基线饮水量、摄食量比值—相对饮水量、相对摄食量表示刺激期间饮水量、摄食量的变化；每轮刺激前和刺激结束时检测血清果糖胺，并进行核素胃排空检查；两轮刺激结束时分别经手术取胃壁全层活检标本，制作石蜡切片，采用间接免疫双重染色法检测胃窦肌间神经丛神经节内神经元总数、乙酰胆碱酯酶免疫反应（ChAT-IR）阳性神经元、血管活性肠肽（VIP）IR阳性神经元、神经元一氧化氮合酶（nNOS）IR阳性神经元、钙视网膜蛋白（calretinin）IR阳性神经元等特异性神经元。比较胃电真刺激和假刺激对上述研究指标的影响。研究中监测手术相关不良事件和胃电刺激装置植入后的安全性。研究结束后开腹检查刺激装置在位情况。

结果 4只实验犬植入电刺激装置手术顺利，术后恢复良好。在胃电刺激第3周和第4周，真刺激组相对摄食量显著高于假刺激组（1.19±0.26比1.05±0.20，P=0.039；1.26±0.26比1.01±0.20，P=0.001）；胃电刺激第4周真刺激组相对饮水量显著高于假刺激组（1.27±0.25比1.07±0.17，P=0.001）；真刺激组体重增长中位数明显高于假刺激组（0.61kg 比0.36kg，P=0.034）。血清果糖胺水平在真刺激组和假刺激组间无显著差异（P=0.909）。真刺激组胃电刺激4周后2小时胃排空率中位数65%（46%,76%）与刺激前52%（34%,78%）比较无显著差异（P =1）。

胃电真刺激4周后，犬胃窦肌间神经丛神经节平均神经元数与假刺激比较无显著差异；真刺激组nNOS-IR阳性神经元比例（34.9±1.7 %）明显高于假刺激组（27.8±1.9%，P＝0.004）；真刺激组VIP-IR阳性神经元比例低于假刺激组（62.5±1.5 % 比65.0±1.8 %，P＝0.019）；二组ChAT-IR阳性神经元和calretinin-IR阳性神经元比例无显著差异（分别为66.4±1.0 %比68.4±1.3 %，P＝0.060；27.4±2.4 % 比29.7±2.7 %，P=0.192）。

整个手术和试验期间，实验犬未发生腹腔、脉冲发生器囊袋和皮肤感染等不良事件。1只试验犬的胃体电极阻抗在术后第6周明显超出正常范围，术中检查证实一根电极发生移位。

结论 国产植入式胃电刺激装置在胃窦采用高频短脉冲刺激4周能显著增加健康比格犬的相对摄食量、相对饮水量和体重，但对血清果糖胺水平和2小时胃排空率无显著影响；胃电刺激后胃窦肌间神经丛nNOS-IR阳性神经元比例明显增加，提示抑制性神经元可能参与了GES效应机制。植入式胃电刺激装置整体安全性好。

Background and Objectives  Some patients with gastroparesis are refractory to conventional dietary modification and pharmacological treatment. Gastric electrical stimulation (GES) improves patients' symptoms and quality of life by generating electrical pulses to stimulate gastric antrum contraction, but its mechanism of action has not been consistently concluded. Some studies have found that GES can affect the gastric emptying and that the nervous system may be involved in its mechanism of action, but the effect of GES on food intake and the involvement of specific neuron to the enteric nervous system (ENS) are lacking. In this study, we investigated the effects of high-frequency short-pulse stimulation on food intake, water intake, body weight, serum fructosamine and gastric emptying by implanting a domestically developed gastric electrical stimulation device in healthy Beagle dogs, and monitored adverse events to assess its safety and further investigate the effects of GES on specific neurons of the gastric myenteric plexus.

Materials and Methods  Four healthy Beagle dogs were randomly divided into groups A and B. All dogs were implanted with electrical stimulation devices after a minor laparotomy, and two rounds of stimulation were performed after recovery from surgery, one for real stimulation and one for sham stimulation, each for 4 weeks. The water intake, food intake and body weight of the dogs were measured and recorded before and during each round of stimulation, and the data during the stimulation period were expressed as the ratio of water intake and food intake to the baseline water intake and food intake before stimulation-relative water intake and relative food intake; serum fructosamine was measured before and at the end of each round of stimulation, and nuclear scintillation scanning for gastric emptying was performed. At the end of each round of stimulation, the whole gastric wall biopsy specimens were taken surgically, and the paraffin sections of tissue were prepared for indirect immunofluorescence double staining. The effects of true and sham stimulation were compared. Monitor procedure-related adverse events and the safety of the gastric electrical stimulation device after implantation during the study. At the end of the study, the device's position was checked by laparotomy.

Results  Four Beagle dogs implanted with electrical stimulation devices underwent successful surgery and recovered well. At weeks 3 and 4 of gastric electrical stimulation, the relative food intake was significantly higher in the true stimulation group than in the sham stimulation group (1.19±0.26 vs 1.05±0.20, P=0.039; 1.26±0.26 vs 1.01±0.20, P=0.001); the relative water intake was significantly higher in the true stimulation group than in the sham stimulation group at week 4 of gastric electrical stimulation (1.27±0.25 vs 1.07±0.17, P=0.001); median weight gain in the real stimulation group was significantly higher than that in the sham stimulation group (0.61 kg vs 0.36 kg P=0.034). Serum fructosamine levels were not significantly different between the true stimulation and sham stimulation groups (P=0.909). There was no significant difference in the median 2-hour gastric emptying rate of 65% (46%,76%) after 4 weeks of gastric electrical stimulation in the true stimulation group compared with 52% (34%,78%) before stimulation (P=1).

After 4 weeks of true gastric electrical stimulation, the mean number of neurons in the ganglia of the myenteric plexus of the canine gastric antrum was not significantly different compared with sham stimulation; the proportion of nNOS immunoreactivity (nNOS-IR) positive neurons was significantly higher in the true stimulation group (34.9±1.7%) than in the sham stimulation group (27.8±1.9%, P=0.004); the proportion of VIP-IR positive neurons were lower than the sham group (62.5±1.5% vs 65.0±1.8%, P=0.019); there was no significant difference in the proportion of ChAT-IR positive neurons and calretinin-IR positive neurons (66.4±1.0% vs 68.4±1.3%, P=0.060; 27.4±2.4% vs 29.7±2.7%, P=0.192 respectively).

No adverse events such as abdominal, pulse generator pouch or skin infections occurred in the experimental dogs throughout the surgery and trial period. The resistance of gastric body electrode was significantly outside the normal range at the 6th postoperative week in one dog, and the displacement of one electrode was confirmed after a laparotomy examination.

Conclusions  The domestically developed implantable gastric electrical stimulation device significantly increased the relative food intake, relative water intake and body weight of healthy Beagle dogs after 4 weeks of high-frequency short-pulse stimulation of the gastric antrum, but had no significant effect on serum fructosamine levels and 2-hour gastric emptying rate; the proportion of nNOS-IR positive neurons in the myenteric plexus of the gastric antrum increased significantly after gastric electrical stimulation, suggesting that inhibitory neurons may be involved in the GES effect mechanism. The overall safety of the implantable gastric electrical stimulation device was good.