Competing Interest There are no potential conflicts of interest in the preparation of this paper.
Diabetic gastroparesis—Backwards and forwards
Article first published online: 4 JAN 2011
© 2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd
Journal of Gastroenterology and Hepatology
Special Issue: Silver Jubilee Supplement: Celebrating 25 years of JGH
Volume 26, Issue Supplement s1, pages 46–57, January 2011
How to Cite
Chang, J., Rayner, C. K., Jones, K. L. and Horowitz, M. (2011), Diabetic gastroparesis—Backwards and forwards. Journal of Gastroenterology and Hepatology, 26: 46–57. doi: 10.1111/j.1440-1746.2010.06573.x
- Issue published online: 4 JAN 2011
- Article first published online: 4 JAN 2011
- blood glucose;
- gastric emptying;
- incretin hormones
Diabetic gastroparesis was once thought to be rare, associated with a poor prognosis, and to affect only patients with type 1 diabetes and irreversible autonomic neuropathy. A landmark study conducted by Horowitz et al. and published in JGH in 1986 paved the way for further studies to examine the pathophysiology, natural history and prognosis of diabetic gastroparesis, as well as its optimal management. This review summarizes the developments in knowledge gained over the last ∼25 years that have led to understanding about normal and disordered gastric emptying in diabetes, with a particular emphasis on the inter-relationship between the rate of gastric emptying and the regulation of blood glucose.
Yogi Berra, the former American baseball player and manager, well known for his malapropisms, once suggested: “It's tough to make predictions, especially about the future.” Diabetic gastroparesis was once thought to be a rare condition, afflicting patients with longstanding type 1 (and not type 2) diabetes, associated with a poor prognosis, predictable on the basis of upper gastrointestinal symptoms, and solely attributable to irreversible autonomic (vagal) neuropathy.1 The study reported in 19862 represented the first comprehensive evaluation of gastric emptying in type 1 diabetes and stimulated a substantial redefinition of these concepts. The study capitalised on the availability of radionuclide techniques to quantify gastric emptying, and assessment of autonomic function using standardised cardiovascular reflex tests. By monitoring blood glucose concentrations during the measurements of gastric emptying, the potential impact of acute changes in the blood glucose concentration on gastric emptying was evaluated; upper gastrointestinal symptoms were assessed by a standardised questionnaire. The study also provided information about esophageal motility in diabetes—esophageal transit of a radioisotopically labelled bolus was measured and shown to be delayed in 42% of cases. This review focuses on the substantial advances in knowledge gained during the subsequent ∼25 years relating to normal and disordered gastric emptying in diabetes, with particular emphasis on the impact of gastric emptying on the regulation of blood glucose.
Physiology of gastric emptying
It is now recognised that normal gastric emptying is dependent on the coordinated activity of the proximal and distal stomach, pylorus and the upper small intestine. It has been known since 18933 that in the fasting state, gastric motility undergoes a cyclical pattern, termed the “migrating motor complex”. This consists of phase I (motor quiescence, ∼40 min), phase II (irregular contractions, ∼50 min) and phase III (regular contractions at 3 per minute for ∼5–10 min).4 Large, indigestible solid particles are usually emptied from the stomach into the small intestine during phase III and, accordingly, absent or disordered phase III activity has the potential to result in gastric “bezoar”. Fasting motility is converted promptly to a postprandial pattern following meal consumption, with irregular antral contractions and an increase in tonic and phasic pyloric pressures.5
The proximal stomach initially relaxes to “accommodate” a meal while the antrum grinds solid food into particles <2 mm in size, and pumps chyme into the duodenum against pyloric resistance in a predominantly pulsatile manner. Contractions of the antrum and pylorus are controlled by electrical slow waves generated by the so-called interstitial cells of Cajal (ICC). These are specialized pacemaker cells that initiate approximately three slow waves per minute in the stomach. Gastric emptying of nutrients occurs at an overall rate of 1–4 kcal/min primarily as a result of the interaction with receptors in the small intestine which generate inhibitory neurohumoral responses. The latter are mediated, at least in part, by cholecystokinin (CKK),6 glucagon-like peptide-1 (GLP-1)7 and peptide YY (PYY),4 and are dependent on the length, and region, of small intestine exposed.8 Solids and liquids have different patterns of emptying. Solids empty in an overall linear pattern after an initial lag phase, while liquid emptying does not usually exhibit a lag phase and slows from an exponential to a linear pattern as the caloric content increases.9 The lag phase for solids reflects the time taken for meal redistribution from the proximal to the distal stomach and the grinding of solids into small particles by the antrum. When liquids and solids are consumed together, liquids empty preferentially.
Prevalence (and natural history) of disordered gastric emptying in diabetes
Gastroparesis refers to abnormal gastroduodenal motility characterized by delayed gastric emptying in the absence of mechanical obstruction. The etiology is multifactorial and it is now recognised that diabetes is probably the most common cause. Gastric retention in diabetes was first noted by Boas in 1925,10 with subsequent radiological findings by Ferroir in 193711 noting that the stomach motor responses in diabetics are weaker than normal—“contractions are slow, lack vigour and die out quickly”.11,12 The first detailed description of the association between delayed gastric emptying and diabetes was by Rundles in 1945, who reported that gastric emptying of barium was abnormally slow in 5 of 35 type 1 patients with peripheral neuropathy.1 In 1958, Kassander named the condition “gastroparesis diabeticorum” and commented that this syndrome was “more often overlooked than diagnosed”.13
While the prevalence of gastroparesis remains uncertain because of the lack of population-based studies, cross-sectional studies, which for the main part have employed radioisotopic methods, indicate that gastric emptying is abnormally delayed in 30–50% of outpatients with longstanding type 1 (as reported in the original study of 45 patients)2 and type 2 diabetes.14,15 This prevalence was clearly underestimated in early studies, which employed less sensitive diagnostic methods to quantify gastric emptying. The reported prevalence is highest when gastric emptying of both solids and nutrient-containing liquids is quantified, either concurrently or separately, reflecting the relatively poor correlation between gastric emptying of solids and liquids in diabetes.16,17 Symptoms attributable to gastroparesis are reported in 5–12% of patients with diabetes in the community, but much higher rates are evident in patients evaluated in tertiary referral centres.18 Gastric emptying is not infrequently abnormally rapid in both type 1 and 2 diabetes.19
In the study reported in 1986, the patients were selected at random from an outpatient setting, and only patients with type 1 diabetes were included. While blood glucose levels were monitored, they were not stabilised. A subsequent study, which evaluated a cohort of 20 unselected outpatients with longstanding type 2 diabetes,14 indicated that the prevalence of gastroparesis was comparable to that observed in type 1 patients. Given that acute hyperglycaemia slows gastric emptying (discussed under “Pathogenesis—Impact of Glycaemia”), the reported prevalence of gastroparesis in both studies14,20 probably represents an overestimate. Data from these studies allowed subsequent evaluation of the impact of both upper gastrointestinal symptoms and gastroparesis on mortality21 and the natural history of delayed gastric emptying in diabetes.22 The prognosis of diabetic gastroparesis had hitherto been assumed to be poor, however, when 20 subjects from the original cohort were re-evaluated after a mean period of 12 years, there was no major deterioration in either the rate of gastric emptying, or symptoms over this time period.22 While there was a deterioration in cardiovascular autonomic nerve function, there was a concomitant improvement in glycemic control, as assessed by glycosylated haemoglobin,22 (attributable to the increased attention given to the achievement of tight blood glucose control subsequent to the outcome of the DCCT study), which may potentially account for the lack of change in gastric emptying. Further studies are indicated.
Diagnosis of disordered gastric emptying
The decision of when to evaluate patients with diabetes for disordered gastric emptying is not straightforward. While upper gastrointestinal symptoms occur frequently, the original2,14 and subsequent22 studies have established that they are not strongly predictive of delayed gastric emptying, contrary to what was thought previously.13 Furthermore, some patients with markedly delayed gastric emptying are asymptomatic. In any patient with diabetes who presents with upper gastrointestinal symptoms suggestive of delayed gastric emptying, reversible causes of gastroparesis must be excluded after endoscopy has been performed (Table 1). The diagnosis of gastroparesis is usually based on the presence of upper gastrointestinal symptoms in combination with objective evidence of delayed gastric emptying. The latter should ideally be measured during euglycemia, or at least with the blood glucose >4 mmol/L and ≤10 mmol/L, given the effect of hyperglycemia to slow emptying. Medications that may influence gastric emptying should ideally be withdrawn for 48–72 h prior to the test (or for the half-life of the drug)23 and smoking, which has been shown to slow gastric emptying, should be avoided on the day of investigation.24
|Drugs (eg, opiates, anticholinergics, levodopa, calcium-channel blockers, beta-blockers, octreotide, alcohol, nicotine, cannabis)|
|Electrolyte or metabolic disturbance (hyperglycemia, hypokalemia)|
|Viral illness (gastroenteritis, Herpes zoster infection)|
|Idiopathic (including functional dyspepsia)|
|Surgery (including vagotomy, and heart and/or lung transplantation)|
|Gastroesophageal reflux disease|
|Connective tissue diseases (eg. systemic sclerosis, dermatomyositis or polymyositis, systemic lupus erythematosis, amyloidosis)|
|Endocrine or metabolic disturbance (hypothyroidism or hyperthyroidism, Addison's disease, porphyria, chronic liver or renal failure)|
|Chronic idiopathic intestinal pseudo-obstruction|
|Neuromuscular conditions (central nervous system disease including stroke. trauma, or tumour; brainstem or spinal cord lesions, Parkinson's disease, autonomic degeneration, myotonic and muscular dystrophy)|
|Infection (HIV, Chagas' disease)|
There are various methods of assessing gastric emptying, but scintigraphy, which is non-invasive and reproducible, remains the most sensitive and accurate method and is the “gold standard” technique. Intragastric distribution of solid and/or liquid meal components, which is frequently abnormal in diabetic patients17 can also be evaluated with scintigraphy. In an effort to standardize the test meal and technique between various centres, a recent consensus statement recommends the use of a low fat, egg white meal labelled with 99mTc-sulfur colloid,25 with measurement of gastric emptying for 4 h. Despite this recommendation, scintigraphy is still not well standardized. Low nutrient liquids should not be used to quantify gastric emptying for diagnostic purposes since they do not stimulate small intestinal feedback mechanisms which retard gastric emptying. Contrary to what is generally assumed, there is little, if any, evidence that the use of high nutrient liquid, or semi-solid, meals is inferior to solids. Moreover, the concurrent measurement of solid and nutrient liquid emptying adds diagnostic value, since, as shown in the original study, the relationship between gastric emptying of solids and nutrient liquids is poor in diabetes.20 If carbohydrate is included in the meal the relationship between glycemic response and the rate of gastric emptying can be evaluated.
Another non-invasive method for assessing gastric emptying is the stable isotope breath test. This uses 13C-acetate or 13C-octanoate as a label and, in contrast to scintigraphy, does not involve exposure to ionising radiation. It has good reproducibility and the results have been reported to correlate well with scintigraphy, with a sensitivity and specificity of 86% and 80%, respectively, for the presence of delayed gastric emptying,26 including in a diabetic population. Following ingestion, the labelled meal passes through the stomach to the small intestine, where the 13C-acetate or 13C-octanoate is absorbed, metabolized into 13CO2 in the liver and exhaled via the breath.13 CO2 in breath samples is analyzed by mass spectrometry. While this technique has advantages over scintigraphy, information relating to the validity of breath tests in patients with markedly delayed gastric emptying is limited.
Transabdominal ultrasound is a simple, non-invasive, inexpensive and convenient method to assess gastric distension, antral contractility, transpyloric flow and gastric emptying and is uniquely able to measure the latter three parameters simultaneously.18 However, the necessity for considerable expertise, and technical limitations of obesity and abdominal gas, restrict its widespread use. While 2-dimensional ultrasonography provides an indirect measure of gastric emptying which is determined by changes in antral area over time,27 the more recently applied 3-dimensional ultrasonography has the capacity to provide comprehensive imaging of the stomach, including information about intragastric meal distribution. It has also been validated against scintigraphy to measure gastric emptying in both healthy subjects and patients with diabetic gastroparesis.28–31
Magnetic resonance imaging (MRI) has also been used to measure gastric emptying and motility with excellent reproducibility.18 However, its use is limited to research purposes because of its high cost and limited availability.
Barium meal, involving a non-nutrient contrast load, has no role in quantifying gastric emptying and its use is limited to excluding mucosal lesions or obstructions. The paracetamol (acetaminophen) absorption test as a simple bedside test is limited to evaluation of the emptying of liquids and is not recommended as a diagnostic tool as its accuracy is variable at best.32
Swallowed capsule telemetry (“SmartPill”) employs an indigestible capsule that has the capacity to measure intraluminal pH and pressure as the capsule travels through the digestive tract to determine the gastric emptying rate. The pressure measurements also provide information about the motor function of the stomach, small intestine and colon.33 This method has been reported to correlate relatively well with scintigraphy with good sensitivity (82%) and specificity (83%), but has not been used widely. Emptying of the capsule presumably usually occurs after that of digestible meal components.
Electrogastrography measures the frequency of the gastric slow wave (∼3 cycles/min) using surface electrodes attached to the skin of the epigastrium.34 While it is clear that abnormalities in gastric electrical activity, particularly tachygastria, occur frequently in diabetic gastroparesis and may be induced by hyperglycemia,35 the relationship is not sufficiently strong to be of diagnostic value.
Antropyloroduodenal manometry, using a water-perfused or solid-state catheter to measure intraluminal pressures in the stomach, pylorus, and small intestine, is only available in a few centres and remains primarily a research tool.
Pathogenesis—impact of glycemia
The pathogenesis of diabetic gastroparesis is now recognized to be complex and multifactorial; there has been recent awareness of defects in various interacting cell types, in addition to the more established roles of autonomic neuropathy and acute hyperglycemia.
The similarity in gastrointestinal symptoms experienced by surgically vagotomised patients and patients with longstanding diabetes led to the initial concept that irreversible vagal damage underlies disordered gastric emptying in diabetes.1 Due to the difficulties of assessing gastrointestinal autonomic function directly, evaluation of cardiovascular autonomic function has been employed widely as a surrogate marker for the function of the abdominal vagus.36 Though the initial2,14 and subsequent22 studies established that the prevalence of disordered gastric emptying is higher in those patients with cardiovascular autonomic neuropathy, the relationship between disordered gastric emptying and abnormal cardiovascular autonomic function is relatively weak 16,37
Diabetic gastroparesis is associated with heterogeneous motor dysfunctions, including “incoordination” of the motor activity of the proximal stomach, antrum, pylorus and duodenum.38 Data from the National Institutes of Health (NIH)-funded Gastroparesis Clinical Research Consortium, based in the USA, have contributed substantially to knowledge of the role of cellular defects in the pathogenesis of gastroparesis. Recent insights gained from animal and human gastric tissue indicate a heterogeneous pathological picture, with abnormalities in multiple, interacting cell types, including decreased numbers of ICC,39,40 deficiencies of inhibitory neurotransmission,39,40 reduced numbers of extrinsic autonomic neurons,41 smooth muscle fibrosis39and abnormalities in the function of immune cells.42 Loss/dysfunction of ICC appears to be central to the pathogenesis of diabetic gastroparesis.43 In animal models and humans with diabetic gastroparesis, a reduction in intraneuronal levels of nitric oxide, an important enteric neurotransmitter, has been observed, reflecting loss of neuronal nitric oxide synthase (nNOS) expression within the myenteric neurons and, potentially, inhibition of nNOS by advanced glycation products.44 Heme-oxygenase-1, the enzyme which gives rise to carbon monoxide (CO), which protects the ICC from oxidative stress, has recently been shown to be reduced in non-obese diabetic (NOD) mice with delayed gastric emptying.45 Administration of hemin, which increases the expression of hem-oxygenase-1,42,45 and administration of CO,46 reversed the loss of ICC with normalization of delayed gastric emptying. Hemin also increases plasma levels of heme-oxygenase-1 when given intravenously to healthy humans47 and may, accordingly, have a therapeutic role.
In the initial study, while there was no significant correlation between plasma glucose levels and the rate of gastric emptying, gastric emptying of liquids and the lag phase for solids were slower when the mean plasma glucose was >15mmol/L. It was subsequently established, using the glucose “clamp” technique, that acute variations in blood glucose impact significantly on gastric emptying in both healthy and diabetic subjects,48 with marked hyperglycemia (blood glucose ∼15mmol/L) delaying gastric emptying of solids and liquids substantially.49 Gastric emptying is also slower when the blood glucose is at the upper end of the physiological postprandial range (∼8mmol/L), when compared to a blood glucose of ∼4mmol/L, in both healthy subjects and patients with uncomplicated type 1 diabetes.50 The mechanisms by which acute hyperglycemia slows gastric emptying include suppression of antral contractions,48 increased pyloric contractions,48 proximal stomach relaxation48 and induction of gastric electrical dysrhythmias.35 In the initial study, the duration of the lag phase for solids was apparently related to chronic blood glucose control, as assessed by glycated hemoglobin, but the relevance of long-term glycemia to the pathogenesis of gastroparesis remains uncertain.
In contrast to the effects of acute hyperglycemia, insulin-induced hypoglycemia accelerates gastric emptying in healthy subjects,51 patients with uncomplicated type 1 diabetes52 and in type 1 diabetics with gastroparesis.53 Such enhanced gastric emptying probably serves as a counter-regulatory mechanism to hasten the delivery of nutrients for absorption.
Significance of upper gastrointestinal symptoms in diabetes and their etiology
The prevalence of upper gastrointestinal symptoms such as nausea, vomiting, early satiety, postprandial fullness, bloating and abdominal pain, is higher in patients with both type 1 and 2 diabetes in comparison to the general population.54,55 What is contentious is the magnitude of this difference. It has been shown that gastrointestinal symptoms in patients with diabetes impact negatively on health-related quality of life, and assessment of these symptoms should take into account potential psychological/psychiatric factors, along with other variables such as age, gender, body weight and use of drugs such as nicotine and alcohol.56 Subsequent to the recognition that the relationship between upper gastrointestinal symptoms and the rate of gastric emptying is weak, studies have focused on other potential causes for inducting symptoms.16,57 In some patients, there is an increased perception of gastric distension, implicating the role of visceral hypersensitivity in the etiology of symptoms.58–60 Acute hyperglycemia has been shown to increase the perception of gastrointestinal sensations (e.g. nausea), and fullness induced by gastric or duodenal distension, or small intestinal nutrient infusion, is greater during hyperglycaemia (blood glucose level ≥ 11mmol/L) when compared to euglycaemia.48,61,62 In diabetic patients, the perception of postprandial fullness is greater as the blood glucose increases.48,63
Impact of gastric emptying on incretin hormones and glycaemia
In the original study, the potential relationship between glycaemic response to the test meal and the rate of gastric emptying did not receive close attention and the study design was less than optimal to evaluate this. It is now recognised that the rate of gastric emptying impacts on blood glucose and this issue has assumed increasing importance. The presence of nutrients in the small intestine stimulates the release of so-called “incretin” hormones, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) that stimulate insulin secretion64 and are responsible for ∼70% of the postprandial insulin response in healthy humans.65 The “incretin effect” refers to the substantially greater insulin response to an oral, when compared to an isoglycemic, intravenous glucose load. GIP is secreted primarily from the proximal small intestine, and GLP-1 predominantly from the distal small intestine and colon.64 Exogenous66 and endogenous7 GLP-1 slows gastric emptying and decreases glucagon secretion in a glucose-dependent manner, whilst GIP has no effect on gastric emptying and may stimulate glucagon levels.67 In healthy subjects, exogenous GLP-1 slows gastric emptying with subsequent attenuation in postprandial insulin secretion.68 In type 2 diabetes, the incretin effect is reduced, probably representing an epiphenomenon69 due to the inability of GIP to augment insulin secretion, partly attributable to hyperglycemia, whilst the effects of GLP-1 are intact.70,71
Glycated haemoglobin, which is used widely as a measure of “overall” glycemic control, is influenced by both fasting and postprandial glucose levels, with a greater contribution of the latter, especially as glycemic control improves.72 This is not surprising—as emptying of nutrients from the stomach occurs at an overall rate of ∼1–4 kcal/min in health (and frequently slower than this in diabetes), only a few hours each day, prior to breakfast, are truly reflective of the “fasting” glycemic state. Thus, the management of postprandial blood glucose excursions has in recent years attracted increasing interest.72,73
Postprandial glycemia is potentially influenced by several factors, including preprandial glycemia, the carbohydrate content of a meal, the rate of small intestinal delivery and absorption of nutrients, insulin and glucagon secretion and peripheral insulin sensitivity. While the relative contribution of these factors is variable, it is now appreciated that gastric emptying accounts for at least a third of the variance in peak postprandial levels after oral glucose in both healthy subjects74 and patients with type 12 and type 2 diabetes.57 In type 1 patients with gastroparesis, less insulin is initially required to maintain euglycemia postprandially when compared to those with normal gastric emptying.75 Gastric emptying also accounts for a substantial amount of variation in glycemic response to carbohydrate of variable glycemic indices.48
What has only recently been appreciated is that the relationship of glycemia with small intestinal glucose delivery is non-linear, as evidenced by the glycemic response to intraduodenal infusion of glucose at rates within the normal range for gastric emptying in both healthy76 and type 2 diabetic subjects.77 At an intraduodenal glucose infusion rate of 1kcal/min, there is only a modest elevation in blood glucose, but a substantial elevation in blood glucose occurs in response to an infusion rate of 2 kcal/min. However there is minimal further increase when the rate is increased to 4 kcal/min (Fig. 1).76 These discrepant blood glucose responses are likely to reflect the substantially increased plasma insulin response to the 4 kcal/min infusion, which is probably accounted for by incretin hormone secretion.76 At 1 kcal/min, there is minimal, transient, stimulation of GLP-1 compared with sustained elevation of GIP. In contrast at 4 kcal/min, there is a substantial increase in GLP-1 secretion with further increase in GIP.64,76 Thus, the marked increase in insulin secretion at higher rates of intraduodenal glucose infusion is likely to be attributable to GLP-1,64 secretion of which increases in a non-linear fashion whilst GIP rises linearly.78
In both healthy and type 2 diabetic subjects, an initially more rapid delivery of glucose to the small intestine results in higher GIP, GLP-1 and insulin responses in comparison to constant delivery of an identical glucose load (Fig. 2).79 However this early insulin response is unable to compensate for the greater amount of absorbed glucose, so that there is no improvement in overall glycemic control, rendering this an unsuitable therapeutic strategy.79,80
Modulation of gastric emptying to improve glycemic control
The novel insights relating to the impact of gastric emptying on glycemia have stimulated the development of dietary and pharmacological strategies to improve glycemic control by modulating gastric emptying. Such strategies differ between type 1 and those type 2 diabetic patients who are using insulin, as opposed to those with type 2 diabetes who are treated with oral hypoglycemic agents and/or lifestyle modifications. In the former, treatment would aim to coordinate the delivery of nutrients with insulin delivery—potentially by either slowing or accelerating gastric emptying—but it is essential that gastric emptying is predictable to achieve a more stable glycemic profile with less fluctuation. Thus in a select group of insulin-treated patients with recurrent postprandial hypoglycemia, delayed gastric emptying can potentially be the cause of low blood glucose levels, and drugs which accelerate emptying may be of therapeutic benefit in these individuals. Certainly, measurement of gastric emptying is indicated in patients with potential “gastric” hypoglycemia.81 In contrast, in type 2 patients who are not on insulin, a slower rate of nutrient delivery would be beneficial given the delay in insulin release and/or insulin resistance.
Non-pharmacological approaches for the management of type 2 diabetes include dietary strategies to slow gastric emptying by increasing dietary fibre,82 addition of guar gum83 and, more recently, the use of fat84,85 or protein “preloads” taken before a meal.86 The rationale of the latter strategy is to slow gastric emptying by stimulating small intestinal neurohumoral feedback mechanisms and stimulate the release of GIP and GLP-1 before the main meal.84,86 Fat, a potent inhibitor of gastric emptying, when consumed in small amounts before or with a meal, was shown to slow gastric emptying of other meal components and thus minimize the postprandial rise in blood glucose.85 However only a modest suppression of the peak postprandial blood glucose level was observed,84 as opposed to the effects of an acute whey protein preload,86 which in addition to delaying gastric emptying and stimulating GIP and GLP-1, also increases insulin secretion markedly, possibly via amino acids (Fig. 3).86
Pharmacological agents known to modify gastric emptying have been shown to affect glycemic control acutely in patients with type 1 and 2 diabetes, including prokinetics and agents which slow emptying. There is evidence that erythromycin, in addition to accelerating gastric emptying as a result of its motilin agonist properties, may stimulate insulin secretion, thus improving glycemic control in type 2 diabetes.87,88
Pramlintide, an amylin analogue, slows gastric emptying in healthy subjects89 and in type 1 and 2 diabetes,90 and its long term use is associated with an improvement in glycemic control.91,92 GLP-1 analogues, such as exenatide and liraglutide, are now used therapeutically in the management of type 2 diabetes. They may augment the postprandial insulin response, as well as suppressing glucagon secretion and appetite. However, the main mechanism leading to the reduction in postprandial glycemic excursions, at least in the case of exogenous GLP-1, or its analogues such as exenatide, may be via retardation of gastric emptying68,93 with a significant correlation between the magnitude of the slowing of gastric emptying and the pre-existing rate of emptying i.e. the magnitude of the reduction in glycemic excursions is less when there is pre-existing delay in gastric emptying.93 It has been suggested that there may be tachyphylaxis with long term use of GLP-1 analogues, resulting in diminution of their effects in slowing of gastric emptying.94
Dipeptidyl-peptidase-4 (DPP-4) inhibitors increase plasma concentration of active GLP-1 and thus would be expected to slow gastric emptying, but the data to date are inconsistent and any effect on gastric emptying appears to be modest.94 This may potentially be accounted for by the effects of DPP-4 inhibitors on other gut hormones, such as PYY or ghrelin, which neutralize the effect of active GLP-1 elevation.95
Management of symptomatic gastroparesis
The management of patients with symptomatic diabetic gastroparesis should focus on the relief of gastrointestinal symptoms, improvement in nutritional status, and optimization of glycemic control. The latter is, of course, pivotal to a reduction in the risk of development, and progression, of micro- and macrovascular complications. Patients with type 2 diabetes may need insulin therapy in place of, or in addition to, oral hypoglycaemic agents, and type 1 patients may benefit from insulin pump therapy.96 Dietary recommendations include increasing the liquid content of meals, restricting fat and fibre intake, and eating a vitamised diet with small, frequent (4–6 per day) meals,97 as well as avoiding alcohol, but none of these measures have been evaluated formally so their use is empirical.
At present, prokinetic agents, including metoclopramide, erythromycin and domperidone, form the mainstay of treatment. These drugs accelerate gastric emptying by increasing antral contractility and improving the organisation of gastropyloroduodenal motility.98 The acceleration of gastric emptying by prokinetics is greater when the emptying at baseline is more delayed and the effect is attenuated during acute hyperglycemia.99 In a systematic analysis of clinical trials of prokinetics, erythromycin appeared to be superior in accelerating gastric emptying and in relieving symptoms,99 but its long term efficacy is limited by tachyphylaxis due to the down regulation of the motilin receptors, gastrointestinal adverse effects and, possibly, an increased risk of cardiac death.
Metoclopramide, when administered subcutaneously, appears to generate plasma concentrations comparable to those achieved via the intravenous route and is an option for those who cannot tolerate oral medications. Central nervous system adverse effects are common and irreversible tardive dyskinesia is a rare complication with its use. The US Food and Drug Administration (FDA) has recently issued a “black box” warning in relation to the latter. Metoclopramide appears to be less effective than cisapride, which has been largely withdrawn from clinical use due to its capacity to prolong the QT interval and lead to ventricular arrhythmias.100 Domperidone is also effective at relieving symptoms whilst not crossing the blood-brain barrier in significant quantities and may now be regarded as the current “first-line” agent. Several drugs, including the motilin agonist, mitemcinal,101 ghrelin and ghrelin receptor agonists,102,103 5-HT4-receptor agonists and the muscarinic antagonist, acotiamide18 are being investigated for their potential use.
A number of non-pharmacological treatments for diabetic gastroparesis have been explored. Intrapyloric botulinum toxin has been shown in randomized, controlled trials to have little, if any, effect to improve gastric emptying or symptoms104,105 despite promising data in earlier, uncontrolled studies.106,107 Gastric electrical stimulation (GES) employs the use of electrodes implanted in the smooth muscle layer of the gastric wall, which are connected to a subcutaneously located pulse generator. Two types of stimulation have been evaluated in humans, one using low frequency, long duration pulses at, or just above, the frequency of gastric slow wave of 3 pulses per minute, and the other using high frequency, short duration, pulses at about four times the slow wave frequency (12 per minute).108 The latter mode is commercially available as the Enterra device and benefits have been reported in several uncontrolled case series.109–111 However, a recent double-blind trial with GES in diabetic gastroparesis showed initial improvement in the run-in “on” phase, but no significant difference when the subsequent phase was randomized to “on” or “off”;112 this indicates the need for further evaluation before GES can be recommended. Benefits of surgical therapy for intractable gastroparesis remain uncertain as case series have been uncontrolled and involve small numbers.113,114 Uncontrolled observations have also been made of the benefit of pancreatic transplantation on gastric emptying.115
In summary, the search for more effective treatments for diabetic gastroparesis represents an area of major research activity as therapy remains suboptimal.
There have been major advances in knowledge about diabetic gastroparesis, of which a number were stimulated by the publication of the pivotal JGH paper in 1986.20 The results have allowed a longitudinal evaluation of the prognosis and natural history of diabetic gastroparesis.21,22 While numerous novel diagnostic and therapeutic strategies have been evaluated and implemented, there is still much to be understood about this complex and beguiling disorder, which is now recognised to be inextricably linked to glycemic control. To quote Yogi Berra again, “The future ain't what it used to be.”
- 4Gastric function. In: HorowitzM, SamsonM, eds. Gastrointestinal Function in Diabetes Mellitus. Chichester: John Wiley & Sons Ltd, 2004; 117–76., , et al.
- 10Diseases of the Stomach, 9th edn. Leipzig: Georg Thieme, 1925..
- 11The Diabetic Stomach. Thesis in Medicine. Paris, France, 1937..
- 33American Neurogastroenterology and Motility Society consensus statement on intraluminal measurement of gastrointestinal and colonic motility in clinical practice. Neurogastroenterol. Motil. 2008; 20: 1269–82., , et al.
- 56Impact of chronic gastrointestinal symptoms in diabetes mellitus on health-related quality of life. Am. J. Gastroenterol. 2001; 96: 71–6., , et al.Direct Link:
- 60Proximal gastric compliance and perception of distension in type 1 diabetes mellitus: effects of hyperglycemia. Am. J. Gastroenterol. 2000; 95: 1175–83., , , , , .Direct Link:
- 66Gastric emptying, glucose responses, and insulin secretion after a liquid test meal: effects of exogenous glucagon-like peptide-1 (GLP-1)-(7-36) amide in type 2 (noninsulin-dependent) diabetic patients. J. Clin. Endocrinol. Metab. 1996; 81: 327–32., , , , , .
- 77Effects of variations in duodenal glucose load on glycaemia, insulin and GLP-1 in type 2 diabetes (Abstract). Diabetologia 2009; 52: S105., , et al.
- 95The effect of dipeptidyl peptidase-4 inhibition on gastric volume, satiation and enteroendocrine secretion in type 2 diabetes: a double-blind, placebo-controlled crossover study. Clin. Endocrinol. (Oxf) 2008; 69: 737–44., , et al.
- 105Botulinum toxin A for the treatment of delayed gastric emptying. Am. J. Gastroenterol. 2008; 103: 416–23., , , , .Direct Link:
- 106Botulinum toxin for the treatment of gastroparesis: a preliminary report. Am. J. Gastroenterol. 2002; 97: 1548–52., , , .Direct Link:
- 107Treatment of idiopathic gastroparesis with injection of botulinum toxin into the pyloric sphincter muscle. Am. J. Gastroenterol. 2002; 97: 1653–60., , et al.Direct Link:
- 114A systematic review of surgical therapy for gastroparesis. Am. J. Gastroenterol. 2003; 98: 2122–9., .Direct Link: