Portions of this work were presented at Digestive Diseases Week 2000 and published as an abstract; Gastroenterology 2000; 118: A3175.
Visceral hypersensitivity and impaired accommodation in refractory diabetic gastroparesis
Article first published online: 13 FEB 2008
© 2008 The Authors. Journal compilation © 2008 Blackwell Publishing Ltd
Neurogastroenterology & Motility
Volume 20, Issue 6, pages 635–642, June 2008
How to Cite
Kumar, A., Attaluri, A., Hashmi, S., Schulze, K. S. and Rao, S. S. C. (2008), Visceral hypersensitivity and impaired accommodation in refractory diabetic gastroparesis. Neurogastroenterology & Motility, 20: 635–642. doi: 10.1111/j.1365-2982.2008.01081.x
- Issue published online: 13 FEB 2008
- Article first published online: 13 FEB 2008
- Received: 9 May 2007 Accepted for publication: 27 December 2007
- gastric motility;
- visceral hypersensitivity
Abstract The pathophysiology of persistent gastrointestinal (GI) symptoms in patients with diabetic gastroparesis is poorly understood. Our aim was to evaluate gastric sensation and accommodation to a meal in patients with diabetic gastroparesis and refractory symptoms. We performed intermittent, phasic balloon distensions of the stomach using a gastric barostat device in 18 patients with diabetes and gastroparesis unresponsive to prokinetic therapy and in 13 healthy volunteers. We assessed the biomechanical, sensory and accommodation responses of the stomach, during fasting and after liquid meal. During balloon distension, the sensory thresholds for discomfort were lower (P < 0.02) in patients with diabetes than those in controls, in both the fasting and the postprandial states. The accommodation response to a meal was significantly impaired (P = 0.01) in patients with diabetes when compared to controls, although fasting gastric tone was similar (P = 0.08). Patients with diabetic gastroparesis and refractory GI symptoms demonstrate sensori-motor dysfunction of the stomach, comprising either impaired accommodation, gastric hypersensitivity or both. An objective evaluation of these biomechanical and sensory properties may provide valuable mechanistic insights that could guide therapy.
Gastrointestinal (GI) symptoms are not only common in patients with diabetes,1–5 but can have a profound impact on health-related quality of life6 and can cause significant morbidity.7 In a large population-based survey, the prevalence of dyspeptic symptoms in patients with diabetes was 18%,3 and up to 50% in the previous month when compared to 38% in controls.5
Although delayed gastric emptying is commonly felt to be an important mechanism for GI symptoms in patients with diabetes, it has been reported in only 25–55% of patients with either type 1 or type 2 diabetes mellitus.8–12 Furthermore, most studies have found a poor correlation between dyspeptic symptoms and delayed gastric emptying.11,13–16 This is not surprising, given that emptying is only one of the many complex functions rendered by the stomach.17 Thus, delayed gastric emptying is only one of the many features of diabetic gastropathy, and by itself may not explain the pathogenesis of dyspeptic symptoms in patients with diabetes.
Recent studies suggest that diabetic patients may have altered gastric sensation and/or accommodation responses of the stomach,22–25 although the evidence is conflicting.26 Furthermore, there is no information regarding sensory thresholds or accommodation responses in gastroparetics with persistent dyspeptic symptoms.
Traditionally, treatment of diabetic gastroparesis consists of good glycaemic control, along with dietary modification (small, low fibre and low fat meals) and prokinetic/anti-emetic medications (metoclopramide, erythromycin and domperidone).18–21 But these approaches are only modestly effective and many patients continue to experience persistent dyspeptic symptoms. Management of such patients has remained a challenge in part because of poor understanding of the pathophysiology of their symptoms. A better understanding of the underlying mechanisms in patients with diabetic gastroparesis may lead to more selective and pathophysiologically based approaches to therapy.
Symptoms such as early satiety, postprandial fullness, abdominal discomfort and nausea may be a consequence of altered sensori-motor function of the stomach. We hypothesized that patients with diabetic gastroparesis and refractory GI symptoms will demonstrate gastric hypersensitivity and impaired accommodation response to a meal. Thus, our aim was to investigate the sensory and biomechanical properties of the stomach in response to both mechanical distension with a barostat and a nutrient meal, in a cohort of patients with diabetic gastroparesis and refractory upper GI symptoms.
Materials and methods
Patients with diabetic gastroparesis (>18 years of age) referred to the Gastroenterology service at a tertiary care centre with chronic (>1 year) upper GI symptoms, and a lack of clinical response to prokinetic therapy (≥4 weeks), were eligible for enrolment into the study. Patients were also required to have had a recent upper endoscopy with no erosive disease, ulcers or bezoars in the upper GI tract. Patients were mostly referred from our Diabetes Clinic. Lack of response to prokinetic medications was defined as persistent upper GI symptoms despite daily therapy for ≥12 weeks. Prokinetic therapy was discontinued at least 1 week prior to the gastric barostat study. Patients were advised that the purpose of the gastric barostat study was to investigate the possible mechanisms for their persistent symptoms. Patients were allowed to use prescription anti-emetics (for example: promethazine and ondansetron) up to 24 h before the gastric emptying or the barostat study, but all other medications that could affect motility such as metoclopramide, erythromycin or opioids were discontinued at least 1 week prior to the study.
Healthy volunteers were also recruited using posters and advertisements in the hospital and the University campus. Volunteers were required to be free from any GI or other systemic ailments and not using any medications. Eligible diabetic patients and healthy volunteers were enrolled after obtaining written informed consent. This study was approved by the Institutional Review Board of the University of Iowa.
A symptom questionnaire, modified from the Nepean Dyspepsia Index (NDI),22 was developed and administered to the patients with diabetes prior to the study. This enquired about the occurrence of the following five symptoms in the preceding 2 weeks: postprandial fullness, nausea, vomiting, abdominal pain and bloating. Each symptom was assessed for its frequency, intensity, and duration and was rated on a 0–3 scale; 0 – no symptoms, 1 – mild, 2 – moderate and 3 – severe symptoms. A cumulative score was calculated for each symptom along with a total score for all symptoms. On this scale, the total scores for each symptom and cumulative scores could range from 0 to 9 and 0 to 45 respectively. Additionally, unintentional weight loss in the preceding 6 months was also enquired. A comprehensive medical history, including the use of prokinetic medications, was obtained.
A 6 mm diameter, double lumen plastic-probe containing a 10-cm long, highly compliant balloon was used (MUI Scientific, Toronto, ON, Canada). The capacity of the balloon was 600 mL. The balloon was connected to a barostat (GMB Distender II; G&J Electronics Inc., Toronto, ON, Canada). The assembly incorporated a manometry catheter with two antral pressure sensors. The balloon volume and pressure data from the barostat were fed to an IBM-PC computer via an interface (Golden Gate; G & J Electronics Inc.) that enabled simultaneous display of antral pressure activity together with balloon and pressure data.
Subjects were asked to discontinue any medications that could affect GI motility, at least 24 h prior to the study. Solid meal gastric emptying was evaluated by scintigraphy after ingestion of TC-99M sulphur colloid with one and half eggs, one slice of toast and 120 cc of water. At least 1 week after the gastric emptying test, subjects presented to our lab after an overnight fast for the barostat study. Patients were instructed to not take their insulin and/or oral hypoglycaemic medications on the morning of the test. Firstly, the oropharynx was sprayed with a local anaesthetic, pontocaine (Abbott Laboratories, North Chicago, IL, USA). Next, the probe with the balloon was placed through the mouth into the stomach. The subjects were asked to sit on a high chair in the upright position. The pattern of antral pressure activity was used to confirm the location of the two pressure sensors in the antrum.17 If unsure, fluoroscopy was used to confirm the probe placement. The balloon was distended with 250 mL of air and the probe was slowly retracted until a ‘tug’ was felt signalling that the proximal end of the balloon was located in the fundus.23,25 The balloon was deflated and the probe was anchored at the cheek with tape.
After a rest period of 15 min, the balloon was distended by 1 mmHg increments to assess the minimum distending pressure (MDP), which was defined as the pressure at which respiratory excursions were clearly discernable on the tracing. The intra-operating pressure (IOP) was set at a value of 2 mmHg above the MDP.25 Subsequently, resting (fasting) gastric tone was recorded for a baseline period of 20 min at IOP. Next, isobaric balloon distensions were performed at 3 mmHg increments up to a maximum of 21 mmHg above the IOP. Each distension was maintained for 3 min followed by a period of rest for 3 min. Balloon volumes were recorded continuously at each level of distention. Thirty seconds after each distension, the subject was asked to rate their sensory perception on a scale of 0–6 (0 = no sensation, 1 = vague perception of mild sensation, 2 = definite perception of mild sensation, 3 = vague perception of moderate sensation, 4 = definite perception of moderate discomfort, 5 = severe, but tolerable discomfort and 6 = intolerable pain).25 If the subject reported discomfort at two incremental distensions or pain at any one level of distension, further balloon distensions were discontinued. After a 20-min rest period, a test meal consisting of 350 mL liquid nutrient meal (Boost Plus®; Novartis Nutrition Corporation, Fremont, MI, USA) that provided 360 kcal, (carbohydrates 45 g, fats 14 g and protein 14 g) was given to each subject and they were asked to consume this within 10 min. Baseline recording of postprandial balloon volume was performed for 20 min at IOP. Subsequently, serial balloon distensions were performed and balloon volumes and perception scores were recorded.
Gastrointestinal symptoms Baseline scores for dyspeptic symptoms such as postprandial fullness, nausea, vomiting, abdominal pain and bloating, as reported by patients on the prestudy questionnaire were summated to provide a cumulative score of upper GI symptoms.
- 1Sensory thresholds: We measured the pressure thresholds for discomfort during intragastric balloon distention, in both the fasting and the postprandial state. Thresholds for discomfort were defined as the level of intraballoon pressures at which subjects first reported a perception score of 4. The ‘normal’ threshold for discomfort was defined as the lower limit of the normal range as calculated in the control group (≤5th percentile).
- 2Perception scores: Sensory perception scores were recorded at baseline and at each balloon distension level.
- 1Fasting and postprandial volumes: Baseline volume was defined as the average balloon volume recorded over 20 min of baseline gastric tone assessment at baseline, and was derived by dividing the area under the curve (AUC) for the volume tracing at baseline by the time in seconds. After the test meal, postprandial volumes were similarly calculated.
- 2Gastric compliance: Intraballoon volume responses to stepwise pressure increments in the fasting state were recorded and the compliance was calculated as a ratio of the change in balloon volume and change in pressure (dV/dP).
- 3Accommodation responses: Volume responses to the meal were calculated as the difference between postprandial and fasting baseline volumes for each subject. Data are expressed in millilitres (mL) ± standard error of mean (SEMs).
Mean values were compared using two tailed Student’s t-test with Welch’s correction for unequal variances, one way analysis of variance (anova), or the Kruskall–Wallis test, where appropriate, using a commercially available software package (prism 3.0; GraphPad Software, Inc., San Diego, CA, USA). Correlations were assessed by calculating Spearman’s correlation coefficients for nonparametric data. A P-value <0.05 was considered significant. Sample size analysis was not performed prior to the study.
Subject characteristics and demographics
Eighteen ambulatory out-patients with diabetic gastroparesis (m/f = 4/14; mean age 39.3; range 25–59 years) and 13 healthy volunteers (m/f = 7/6; mean age 32.4; range 18–54 years) participated in the study. The mean ages of these groups were not different (P = 0.1). All patients with diabetes had significantly delayed gastric emptying (% retention at 120 min: 65.1 ± 14.2%, normal range in our centre: 0–50%, mean 25%). Fourteen patients (77%) had type I diabetes and four (23%) had type II diabetes. All patients with diabetes were receiving insulin and additionally two patients were on oral hypoglycaemics. Patients had long standing diabetes (23.3 ± 6.4 years) and the HbA1C was reasonably good [mean (±SEM) 7.2 ± 1.6 mg dL−1], probably because the patients were under regular follow-up in the university Diabetes Clinic. Seventeen patients (94%) had documented peripheral neuropathy, 12 (66%) had nephropathy (proteinuria and/or elevated plasma creatinine) and 10 (55%) had retinopathy. Formal testing for autonomic neuropathy was not performed. The mean BMI in patients with diabetes was (28.9 kg cm−2) with a range of (24–36 kg cm−2). All patients had previously received specific therapy for their gastroparesis and GI symptoms with prokinetic agents for at least 4 weeks: 12 (67%) received metoclopramide, four (22%) received cisapride, one (5.5%) received erythromycin and one (5.5%) received tegaserod, without clinical improvement.
Blood glucose levels were monitored and documented by the patients on the day of study but were not regulated, as this was not part of the study protocol. On retrospective chart review, we identified that 10 patients had a serum glucose level recorded on the day or within 7 days prior to the study date. The mean (±SEM) serum glucose was 161 ± 46 mg dL−1, with a range of 64–231 mg dL−1.
All patients with diabetes reported moderately severe GI symptoms over the preceding 3 months, with nausea and postprandial fullness being the most prominent symptoms. The mean symptom scores (±SEM) were as follows: nausea 6.8 ± 1.1, postprandial fullness 5.7 ± 0.9, vomiting 5.6 ± 1.2, abdominal pain 4.3 ± 0.9 and bloating 4.1 ± 1.3. Two subjects (11%) reported the weight loss of over 10 lb in the previous 6 months.
Fasting state All subjects tolerated the balloon distensions without adverse events. The lower limit of the normal range of threshold for discomfort (5th–95th percentile in the control group) was 12 mmHg (Fig. 1).
Sensory thresholds: During phasic distension, when compared to controls, the diabetic group demonstrated lower pressure thresholds for discomfort (P = 0.02) (Fig. 1). However, while 10 of 18 (55%) patients with diabetes demonstrated hypersensitivity (fasting sensory thresholds below 12 mmHg), eight of 18 (45%) patients with diabetes had thresholds within the normal range, overlapping with the control group (Fig. 1).
Perception scores: When compared to controls, patients with diabetes had higher perception scores in the fasting state, both at baseline (0.89 ± 0.2 vs 0; P = 0.01) and at each pressure level during phasic balloon distentions (P = 0.001). For example, at 9 mmHg balloon distension pressure, the mean (±SEM) perception scores were 3.1 ± 1.1 vs 1.6 ± 0.9 respectively.
Postprandial state All controls tolerated the test meal and postprandial phasic balloon distensions. However, eight of 18 could not tolerate the test meal due to severe nausea, vomiting and discomfort and consequently could not undergo the postprandial portion of the study.
Sensory thresholds: In controls, when compared to the preprandial balloon distensions, there was no change in the pressure thresholds for discomfort after a liquid meal. However, all patients with diabetes who tolerated the meal (10/10) showed significantly lower pressure threshold for discomfort (P < 0.01, Fig. 1).
Perception scores: When compared to controls, postprandial perception scores were significantly higher in the patients with diabetes, both at baseline (3.5 ± 1.9 vs 1.8 ± 0.6; P = 0.02) and at each pressure level during phasic balloon distentions (P < 0.05). For example, at 9 mmHg balloon distension pressure, the mean (±SEM) perception scores were 4.4 ± 2.9 vs 1.7 ± 1.1 respectively.
Fasting and postprandial volumes At baseline, in the fasting state, the gastric balloon volumes in the patients with diabetes were higher when compared to controls although the difference was not significant (103.9 ± 34.5 vs 79.3 ± 24.6, P = 0.4). Following a test meal, we could assess the gastric volume changes in 10 of 18 patients with diabetes. In these subjects, when compared to controls, the balloon volumes at baseline were significantly lower (100.6 ± 26.5 vs 244.9 ± 58.1, P = 0.01).
Gastric compliance Gastric compliance (dV/dP) was assessed during balloon distensions in the fasting state and was not significantly different between the controls and patients with diabetes (18.9 and 21.3 mL mmHg−1 respectively, P = 0.8).
Accommodation responses We defined impaired accommodation as a volume response to test meal outside the ‘normal’ range (between 5th and 95th percentile) obtained in our control group. The lower limit (5th percentile) of the normal range of volume change was 76 cc and this was used as the empiric cut-off point (Fig. 2A). Using these criteria, nine of 10 patients with diabetes demonstrated an impaired accommodation response to a liquid meal (Fig. 2A). The controls demonstrated rapid changes in intraballoon volumes over 20 min at baseline after ingestion of a test meal. Patients with diabetes demonstrated lower volume variations, probably as a result of impaired accommodation and delayed emptying of the liquid meal (Fig. 2B).
In the patients with diabetes, there was no significant correlation between sensory thresholds and individual upper GI symptoms (Spearman’s r-values between −0.22 and −0.5, all P-values >0.1). Similarly, the postmeal accommodation response (volume change) did not correlate with individual symptoms (Spearman’s r-values between −0.01 and 0.4, all P-values >0.1).
In this study, we evaluated gastric sensation and accommodation responses to a meal in a cohort of patients with diabetes with gastroparesis and refractory upper GI symptoms. We found that patients with diabetes demonstrated gastric hypersensitivity, especially in the postprandial state, and that the postprandial accommodation response was impaired in almost all patients with diabetes.
We found that 55% of our cohort of patients with diabetes demonstrated gastric hypersensitivity in the fasting state. Most importantly, in the postprandial state, almost all patients with diabetes demonstrated hypersensitivity. Previous studies on gastric sensation in patients with diabetes have reported inconsistent results. Samsom et al.26 showed that during ‘isobaric’ balloon distention in the fasting state, patients with diabetes had higher gastric compliance and higher symptom scores than those of healthy volunteers. In contrast, Rayner et al.27 found no such difference in the euglycaemic state, although sensation scores were higher during ‘isovolumetric’ balloon distensions, and gastric compliance was higher in the hyperglycaemic state. Also, the accommodation response to a liquid meal in patients with diabetes was impaired in some28,29 but not all studies.30 We believe that these discrepancies may be due to a subgroup of patients with diabetes with normal gastric sensation in the fasting state or due to methodological differences. Interestingly, hypersensitive and normosensitive patients with diabetes in our group had similar glycaemic control. Also, the duration of diabetes, prevalence of diabetic complications and other comorbidities, prevalence of upper GI symptoms and the degree of postprandial accommodation responses were similar in the two groups. We believe that the underlying sensory dysfunction may contribute to the persistent postprandial bloating, nausea and discomfort in these patients, although the relationship is not linear, possibly due to differences in central processing of gut sensations between patients.31
The fasting gastric tone in our patients with diabetes was not significantly different from controls, whereas the postprandial gastric tone (accommodation response) was significantly impaired. Differences in gastric emptying could have partially accounted for these observations. This confirms and extends previous observations that the accommodation response to a meal is impaired in patients with diabetes.24,25 We did not find any correlation between the baseline symptoms (nausea, vomiting, abdominal pain, postprandial fullness and bloating) and either gastric hypersensitivity or impaired accommodation, but this could be due to a type II error given the small sample size.
We did not study the patients with diabetes with upper GI symptoms and normal gastric emptying; therefore our study cannot explain the pathophysiology of symptoms in patients with diabetes with or without gastroparesis. Also, the diagnosis of gastroparesis was based on measurement of gastric emptying following a solid and not liquid meal. Additionally, reflex motor changes, especially in the antrum, were not assessed.23,32 Because we did not assess blood glucose during the procedure, the potential effect of altered glycaemia on sensori-motor function could not be assessed and is a limitation of our study. Several studies11,33,34 have shown that blood sugar concentrations can affect GI symptoms and motor function. It is unknown whether postprandial hypersensitivity observed in our study is related to the meal itself or the abnormal regulation of blood glucose following the meal. However, most of our patients were under the care of the Diabetes Clinic and had reasonably well-controlled diabetes and on stable therapy. The available data on serum glucose levels within the week prior to the study, along with the HbA1C levels, suggest that overall glycaemic control was reasonably good. Therefore, it would appear that their daily symptoms are less likely due to hyperglycaemia-induced changes in gastric accommodation and sensation. More appropriate study design including monitoring glucose values during the test, and possibly the use of glucose clamping as well as the study of diabetic patients without delayed gastric emptying may help to determine if these abnormalities in visceral sensitivity and accommodation are due to the gastroparetic state and/or hyperglycaemia.
Our findings may have implications for clinical practice. In patients with diabetic gastroparesis and symptoms refractory to standard therapy with prokinetic agents, physiological testing may identify altered sensory and accommodation responses. A better understanding of these underlying mechanisms may allow clinicians to provide appropriate pharmacological therapy. Drugs that modulate gastric perception and accommodation, such as tricyclic agents,35 selective serotonin reuptake inhibitors36 and other serotonergic agents such as Tegaserod37 are currently available for clinical use and several classes are under investigation.38 Further prospective studies are needed to assess whether a treatment approach that is specifically targeted at an underlying mechanism can improve clinical outcomes in patients with diabetic gastroparesis, similar to the discovery and use of glucagon-like peptide compounds in the treatment of diabetes.39,40
In conclusion, we found that patients with diabetes with gastroparesis and refractory upper GI symptoms demonstrated gastric hypersensitivity, impaired accommodation or both dysfunctions, confirming our hypothesis. These pathophysiological dysfunctions were seen in addition to delayed gastric emptying. Therefore, in selected patients with diabetic gastroparesis and persistent symptoms, formal testing of gastric sensation and accommodation may be useful in revealing underlying pathophysiology.
This research was supported in part by an American College of Gastroenterology Clinical Research Grant, and Grant RR00059 from the General Clinical Research Centers Program and Grant R01DK57100-03, National Institutes of Health.
The authors would like to thank William Sivitz, MD and Joseph Dillon, MD for referring patients, Bridget Zimmerman, PhD for statistical support; Dr Xing Zhao and Mr Bernie Hayek for technical assistance; and Ms Heidi Vekemans for secretarial assistance.
- 17Normal gastrointestinal motility of stomach and duodenum. In: KumarD, WingateDL, eds. An Illustrated Guide to Gastrointestinal Motility, 2nd edn. London, UK: Churchill Livingstone, 1993: 373–92., .
- 27Proximal gastric compliance and perception of distension in type 1 diabetes mellitus: effects of hyperglycemia. Am J Gastroenterol 2000; 95: 1175–83., , , , , .Direct Link:
- 35Effect of amitriptyline on symptoms, sleep, and visceral perception in patients with functional dyspepsia. Am J Gastroenterol 1998; 93: 160–5., , , , , .Direct Link: