How To Test For Gastric Emptying

Transcription

1 Neurogastroenterol Motil (2006) 18, doi: /j x Relationships between hypoglycaemia and gastric emptying abnormalities in insulin-treated diabetic patients 1 J.LYSY,E.ISRAELI,N.STRAUSS-LIVIATAN&E.GOLDIN Department of Gastroenterology, Hadassah University Hospital, Hebrew University School of Medicine, Jerusalem, Israel Abstract We hypothesize that hypoglycaemia in insulin-treated diabetic patients may result from gastric emptying abnormalities causing insulin and food absorption mismatching. We tested gastric emptying in insulin-treated diabetic patients with unexplained hypoglycaemia and without dyspepsia and in diabetic patients without hypoglycaemia, prospectively. Thirty-one diabetic patients with unexplained hypoglycaemic events within 2 h of insulin injection and 18 insulin-treated diabetic patients without hypoglycaemic events underwent gastric emptying breath tests, glycaemic control and autonomic nerve function. Gastric emptying tests were abnormal in 26 (83.9%) and in four (22.2%) patients with and without hypoglycaemia, respectively (P < 0.001). Gastric emptying was significantly slower in hypoglycaemic diabetic patients (t 1/ ± 74.1 vs 77.8 ± 23.3 and t lag 95.8 ± 80.3 vs ± min, P < for both comparisons; t-tests). A significant association between hypoglycaemic patients and abnormal values of t 1/2 and t lag was found (P < 0.001). Gastric emptying abnormalities were more frequent in hypoglycaemic patients. We suggest gastric emptying tests for diabetic patients with unexplained hypoglycaemic events. Keywords breath test, diabetes, gastric emptying, hypoglycaemia. Address for correspondence Prof Eran Goldin, MD, Department of Gastroenterology, Hadassah University Hospital, Jerusalem 91120, Israel. Tel: ; fax: ; goldin@cc.huji.ac.il Received: 19 December 2005 Accepted for publication: 3 April The study was presented in a preliminary form at American Motility Society Conference, Texas, USA, INTRODUCTION While hypoglycaemic episodes in diabetic patients are rarely fatal, they can be associated with serious clinical sequelae. 1,2 Hypoglycaemia in insulin-treated diabetic patients is associated with significant cost to the healthcare system. 3 The existence of hypoglycaemia generally precludes maintenance of euglycaemia over a lifetime in diabetic patients and thus prevents the long-term benefits of euglycaemia. 4 Hypoglycaemia is a frequent and important clinical problem in diabetic patients, especially in patients receiving intensive insulin treatment. 5 Each year, according to the UK Prospective Diabetes Study (UKPDS) Group, approximately 3% of patients treated with insulin experienced a severe hypoglycaemic episode and 40% had a hypoglycaemic episode of any severity. 6 Although some episodes of hypoglycaemia can be related to patient s non-compliance or are exercise induced, many of them are still qualified as unexplained hypoglycaemic events. Thus, treatment and especially prevention of hypoglycaemia are major clinical challenges. Gastric emptying delay has been documented to be common as follows: in type 1 diabetes, delayed emptying has been identified in 27 58% of cases; 7,8 in longstanding type 2 diabetes, its prevalence is approximately 30%; 7,9 similarly, in a cohort of unselected diabetic patients, delayed gastric emptying was found to be around 28%. 10 Slower oesophageal transit and delayed gastric emptying was found in 40% of randomly selected type 1 and type 2 diabetic patients. 11 Keshavarizian et al. found frequent (27%) delayed gastric emptying in insulin-requiring diabetic patients, which were often asymptomatic. 12 Lyrenas et al. evaluated ÔunstableÕ type 1 diabetic patients and found significant gastric emptying delay in 44% of their patients. 13 It was suggested that the delayed absorption of nutrients by the small intestine (due to the impaired gastric emptying), may result in time mismatching Journal compilation Ó 2006 Blackwell Publishing Ltd 433

2 J. Lysy et al. Neurogastroenterology and Motility between blood glucose and insulin peaks, making the blood glucose regulation difficult, and thus becoming a major obstacle to glycaemic control. 14 Similarly, hypoglycaemia may result from time mismatching between food absorption and insulin peak. On the other hand, upper gastrointestinal symptoms seem to be a poor predictor of delayed gastric emptying in type 1 diabetic patients 8 and in unselected diabetic patients. 9 Furthermore, it is documented that even severe gastroparesis can be asymptomatic. 15 It is generally assumed that unlike some other diabetic complications (e.g. retinopathy, nephropathy), there is no clinical reason to treat asymptomatic gastric motility disturbances in diabetes. 16 Clinicians considered unexplained hypoglycaemia in diabetic patients without upper gastrointestinal symptoms as a possible indicator of gastric emptying disturbance, but until now there was a lack of data to prove this hypothesis. Herein, based on our prospective gastric emptying studies in insulin-treated patients with and without hypoglycaemic events, we hypothesize that hypoglycaemic events in diabetic patients may be the consequence of gastric emptying abnormalities despite the lack of any upper gastrointestinal symptoms. MATERIALS AND METHODS Patients Thirty-one consecutive insulin-treated diabetic ambulatory patients (16 type 1 diabetic patients and 15 type 2 diabetic patients) of at least 18 years of age were enrolled (15 men and 16 women, mean age ± 13 SD). All the patients were followed up at the different diabetic clinics in Jerusalem, had frequent (more than once a week) and symptomatic hypoglycaemic events (blood glucose 3.3 mmol L )1 and typical symptoms of hypoglycaemia, i.e. sweating, palpitations, hunger, blurred vision, restlessness and anxiety) and at least one event of severe hypoglycaemia (event requiring i.v. glucose or glucagon injection) in the last 3 months preceding the study. Insulin regimens: all the patients were treated with a multiple-dose regimen, consisting of injections of rapid-acting insulin before meals and intermediate-acting insulin once or twice daily. Four patients received long-acting insulin Glargine (Lantus, Aventis Pharma Deutschland GmbH, Frankfurt, Germany). Medications: eight patients were currently using ACE inhibitors and three patients were currently using cholesterol-lowering statins. Another 34 patients with known risk factors for hypoglycaemia were excluded from the study (e.g. non-compliant patients, patients receiving drugs that could potentiate hypoglycaemia, those with renal and liver failure or patients with exercise-induced hypoglycaemia). Another 18 insulin-treated diabetic patients (six type 1 and 12 type 2 diabetic patients) without hypoglycaemic events serving as the control group (13 men and five women, mean age ± 13 SD) agreed to participate in the study. Insulin regimens: 10 patients received premixed insulin twice daily, eight patients were treated with multiple injections consisting of rapid or shortacting insulin before meals and intermediate-acting insulin once or twice daily. Medications: six patients were currently using ACE inhibitors, one patient was currently using disothiazide and four patients were currently using cholesterol-lowering statins. Patients with short-duration diabetes (less than 3 years) and well-controlled diabetes (HbA 1c < 6.5 %) were excluded. They underwent the same motility tests. In all patients, medical history was negative for gastrointestinal symptoms, abdominal surgery and any other systemic diseases. Patients were questioned by a physician regarding upper abdominal discomfort in the last 3 months preceding the study (i.e. postprandial fullness, bloating, epigastric discomfort, early satiety, belching after meals, postprandial nausea, vomiting or epigastric pain). Hypoglycaemic patients and controls were free of any medication that could affect gastrointestinal motility. All the candidates for the study (patients and controls) were instructed to complete a diary for 1 month prior to their participation in the study, indicating symptomatic hypoglycaemic events of less than 3.3 mmol L )1. Venous blood samples for blood glucose concentration and HbA 1c were taken before ingestion of the test meal. The body mass index was checked on the day of the breath test. Autonomic neuropathy was assessed by a standard series of cardiovascular reflex tests. Parasympathetic function was evaluated by the variation (RR interval) of the heart during deep breathing. Sympathetic function was assessed by the fall in systolic blood pressure in response to standing. 17 The protocol of the study was accepted by the ethical committee of the hospital and a written consent approval form was signed by patients recruited to the study. ASSESSMENT OF GASTRIC EMPTYING BY BREATH TEST Gastric emptying rate breath testing is based on measuring changes in 13 CO 2 / 12 CO 2 isotope ratios in the patient s exhaled breath which result from 434 Journal compilation Ó 2006 Blackwell Publishing Ltd

3 Volume 18, Number 6, June 2006 Hypoglycaemia and gastric emptying disorders in insulin-treated diabetes absorption and metabolism of a test meal marker with 13 C labelling as it exits in the stomach. Specifically, after a baseline collection of the patient s breath, a standard meal using 100 mg 13 C- labelled octanoic acid dissolved in a scrambled egg with a pita bread was used (250 kcal). The patients and controls were tested for 4 h in the sitting position in the morning after an overnight fast. 18 Insulin was injected before the test-meal, fasting and 2 h glucose levels were measured. Gastric emptying rate was analysed by a continuous breath test in real time using the BreathID Ò system (Oridion BreathID Ltd, Jerusalem, Israel). The patients and controls were connected to the device via nasal cannula. The test meal (TM) was processed and emptied by the stomach, and after absorption and metabolism, the 13 C was derived and exhaled in the breath as CO 2. The device continuously measures the ratio of 13 C and 12 CinCO 2 of breath exhaled using molecular correlation spectroscopy TM (MCS) by means of absorption of this radiation. 19 The BreathID device measures the ratio, which is normalized for patient weight, height and 13 C substrate and dose, and provides percentage dose recovery (PDR) and cumulative percentage dose recovery (CPDR). It then calculates t 1/2, t lag and gastric emptying coefficient (GEC) based on analysis according to a non-linear model to obtain t lag, t 1/2 and GEC as described by Ghoos et al. 18 The measurements were carried out for 4 h for study purposes, although final t lag, t 1/2 and GEC values were usually obtained at an earlier stage. Normal ranges of gastric emptying parameters were defined as follows: t 1/ min, t lag min and GEC (delayed emptying <2.95). These values were based on Ghoos et al. 18,20 ±1 STD. We used 1 STD (and/or 90 percentile) to be able to detect patients even with slightly impaired motility and not just those with significant impaired motility. Statistical methods Data for analysis consisted of a group of 31 diabetic patients with hypoglycaemic events and 18 diabetic patients without hypoglycaemic events. Variables considered in analysis were patient s general characteristics (gender, age and BMI), patient s disease characteristics (diabetes type, diabetic duration, fasting blood glucose, HbA 1c and neuropathy) and data of the three gastric emptying parameters (t 1/2, t lag and GEC). In addition, five dichotomous gastric emptying variables were defined on the basis of the existing ones: the first three (1) impaired t 1/2, (2) impaired t lag and (3) impaired GEC were defined to equal 1 if the corresponding variable had abnormal values and 0 otherwise. The last two variables were defined as follows: (4) Ôimpaired t 1/2 or t lag Õ equalled to 1 if either t 1/2 or t lag were abnormal, and 0 otherwise, and (5) Ôimpaired any of the threeõ was defined to equal 1 if any of t 1/2, t lag or GEC had abnormal value and 0 otherwise. To analyse what factors are related to hypoglycaemic patients, we performed a two-level analysis. Univariate tests (i) t-tests were used to compare hypoglycaemic and non-hypoglycaemic patients over the following continuous variables: age, BMI, diabetic duration, fasting blood glucose, HbA 1c and the five gastric emptying rate variables; and (ii) chi-squared test of 2 2 tables was used to compare the association between hypoglycaemic patients and diabetes type, gender, neuropathy and the Ôimpaired t 1/2 Õ, Ôimpaired t lag,õ impaired GEC, Ôimpaired t 1/2 or t lag Õ and Ôimpaired any of the threeõ. A multivariate model We used a logistic regression model to determine the factors, which simultaneously influence hypoglycaemic events in diabetic patients. The dependent variable was a dichotomous one, equal to 1 for hypoglycaemic patient and 0 for a non-hypoglycaemic patient. Explanatory variables included variables describing patient and disease and variables concerning gastric emptying measurements. Our aim was to test the significance of gastric emptying variables while controlling for patient and disease variables that can affect the presence of hypoglycaemic events in diabetic patients. The model was estimated using the maximum likelihood estimators method, and a forward stepwise method was used to select significant variables. The model chi-squared statistic determined if the overall model was statistically significant. The Ôpercentage correct predictionsõ statistic determined how well the estimated model could perform in predicting hypoglycaemic patients (yes/no) by given values of the explanatory variables. Exp (B) is the exponent of the coefficient in the logistic model and can be interpreted as the contribution effect of the relevant variable to the odds ratio. Diabetic complications To analyse complications we used, for each type of complication (autonomic, peripheral neuropathy, nephropathy and retinopathy) a model of logistic regression with stepwise selection method and with the following explanatory variables: hypoglycaemic/control group, HbA 1c, diabetic duration, BMI and age. Journal compilation Ó 2006 Blackwell Publishing Ltd 435

4 J. Lysy et al. Neurogastroenterology and Motility RESULTS A total of 49 patients were included. Thirty-one patients with frequent hypoglycaemic events were selected for the study group and 18 patients without hypoglycaemic events served as the control group. Sixteen type 1 diabetic patients and 15 type 2 diabetic patients were in the hypoglycaemic group, and six type 1 diabetic patients and 12 type 2 diabetic patients were in the control group. The hypoglycaemic patients had an average of 9.7 ± 3 (range 6 19) episodes of hypoglycaemia (blood glucose 3.3 mmol L )1 and typical symptoms of hypoglycaemia) during the month preceding the study, and 1.9 ± 1 (range 1 5) episodes of severe hypoglycaemia (events requiring i.v. glucose or glucagon injection) in the 3 months preceding the study. More than 80% of the hypoglycaemic events were in the postprandial period within 2 h following preprandial fast- or short-acting insulin injections, 15% were nocturnal events and 4% not related to postprandial period. Eighteen control patients did not have any symptomatic event of hypoglycaemia; however, of five of them measured during the month preceding the study, few (less than four) had non-symptomatic low blood glucose levels of less than 4.5 mmol L )1 but above 3.3 mmol L )1. All subjects tolerated the study well, and none became hypoglycaemic during the gastric emptying measurement. Analysis of factors that are related to hypoglycaemic patients Univariate analysis Patient and disease variables: Univariate analysis revealed that the characteristics of the diabetic patients with hypoglycaemia and the diabetic patients without hypoglycaemia did not differ regarding diabetics type, levels of HbA 1c, fasting blood glucose, diabetic duration, autonomic neuropathy and gender. Two hours postprandial glucose was tested during the gastric emptying test (comparing mean values of postprandial glucose between the two groups) and this yielded nearly significant lower means for the hypoglycaemic group (P ¼ 0.069). Younger patients and patients with lower BMI belonged to the hypoglycaemic group and this was statistically significant (P < 0.01 for both variables) (Table 1). Gastric emptying variables: Twenty-six (83.9 %) of the hypoglycaemic patients had abnormal gastric emptying rate (25 delayed and one accelerated), compared with four (22.2%) patients in the non-hypoglycaemic group (three delayed and one accelerated, P < 0.001). Hypoglycaemic patients displayed significantly higher levels of t 1/2 and t lag than non-hypoglycaemic Table 1 Patient s general and disease characteristics t-tests results Characteristics Hypoglycaemia n Mean SD Age (P < 0.01) Non-hypo Hypo BMI (P < 0.01) Non-hypo Hypo HbA1 c (P ¼ 0.23, ns) Non-hypo Hypo Diabetic duration Non-hypo (years) (P ¼ 0.63, ns) Hypo Fasting blood glucose (mg dl )1 )(P ¼ 0.50, ns) patients ( ± min vs ± and ± min vs ± 16.95, respectively, P < for both parameters; t-tests). These patients had significantly lower GEC values (2.62 ± 0.79 vs 3.27 ± 0.48, P ¼ 0.001; t-test) (Figs 1 and 2). Similarly, for all gastric emptying dichotomous variables, chi-squared statistics indicated significantly that a higher proportion of hypoglycaemic patients had abnormal values of the five gastric emptying parameters: impaired t 1/2 (P < 0.001), impaired t lag (P < 0.001), impaired GEC t 1/ Non-hypo Hypo GEC 4.00 Figure 1 Individual data for gastric emptying coefficient (GEC) and t 1/2. Control Hypo 436 Journal compilation Ó 2006 Blackwell Publishing Ltd

5 Volume 18, Number 6, June 2006 Hypoglycaemia and gastric emptying disorders in insulin-treated diabetes 95% CI T 1/ (P ¼ 0.028), impaired t 1/2 or t lag (P < 0.001) and impaired any of the three (P < 0.001). Correlations between glucose control on the day of the study (fast blood glucose and 2 h postprandial glucose) and breath test variables were all non-significant for both hypoglycaemic and control patients. Comparing gastric emptying variables between the autonomic neuropathy yes/no for each of hypoglycaemic/control groups yielded a non-significant difference. 95% CI T lag 95% CI GEC Control Control Hypo Hypo Multivariate analysis Estimating the model of logistic regression to explain factors associated with hypoglycaemia given the explanatory variables was performed by using the Ôforward stepwise methodõ with the above patient/disease and gastric emptying variables serving as explanatory variables. This method yielded the following results (Table 2): in the first step, the variable Ôimpaired t 1/2 or t lag Õ appeared as the best explanatory variable (P < 0.001). In the second step, age of patient was added to the model as the second variable (P-value for age ¼ and for Ôimpaired t 1/2 or t lag ¼ 0.01Õ). When these two variables were in the model, no other variable could contribute significantly to the model. Chi-squared statistic for the estimated model was 23.8 with 2 df; overall percentage of cases correctly classified according to this model to hypoglycaemic group or nonhypoglycaemic group was 83.7 (87.1% of hypoglycaemic patients and 77.8% of non-hypoglycaemic patients). The sign of the coefficient of Ôimpaired t 1/2 or t lag Õ in the estimated logistic regression was positive, which implied that impaired gastric emptying is positively associated with hypoglycaemia. In addition, as the coefficient of patient s age in the logistic regression was negative, it followed that controlling for impaired gastric emptying, younger age was associated with hypoglycaemia more than older age. Diabetic complications: Logistic regression models for each type of complication, with hypoglycaemic/ control group, HbA 1c, diabetic duration, BMI and age as Table 2 Results of estimated logistic regression of hypoglycaemia Control Hypo Figure 2 The 95% confidence interval for t 1/2, t lag and gastric emptying coefficient (GEC). Normal t 1/2 values, min; normal values of t lag, min; normal GEC values, (delayed emptying <2.95). B coefficient in estimated model SE P-value Exp (B) Step 1 Impaired t 1/2 or t lag Constant ) Step 2 (final model) Age ) Impaired t 1/2 or t lag Constant Journal compilation Ó 2006 Blackwell Publishing Ltd 437

6 J. Lysy et al. Neurogastroenterology and Motility explanatory variables yielded the following results: for autonomic and peripheral neuropathy, all variables were non-significant. HbA 1c and age were significant explaining nephropathy (P ¼ and 0.04 respectively) with higher values associated with nephropathy. The model for retinopathy yielded HbA 1c and diabetic duration as significant variables (P ¼ and respectively). Higher values were associated with retinopathy. DISCUSSION Here we present a group of insulin-treated diabetic patients with frequent and occasionally severe episodes of hypoglycaemia. Patients with known risk factors, such as intake of drugs that can cause or potentiate hypoglycaemia, severe renal and heart failure were excluded. Patients with exercise-induced hypoglycaemia 21 and non-compliant patients were also not included. However, patients with multiple insulin injections were included. All recruited patients with and without hypoglycaemia had no gastrointestinal symptoms. Several studies showed no correlations between upper gastrointestinal symptoms and gastric emptying rate abnormalities in diabetic patients. 8,10,15 Gastric emptying rate is slower during hyperglycaemia and accelerated during hypoglycaemia. 22,23 Moreover, increase in blood glucose within the normal postprandial range slows gastric emptying. 24 It was demonstrated that insulin requirement after test meal was significantly lower in type 1 diabetic patients with gastroparesis. 25 Another study showed that an increase in the rate of duodenal glucose entry has a profound (increase) effect on subsequent glycaemic, insulin and incretin response. 26 We studied prospectively 49 diabetic patients, 31 of whom had apparently unexplained hypoglycaemic episodes. Eighteen insulin-treated diabetic patients not optimally controlled and without gastrointestinal symptoms served as the control group (HbA 1c of controls was not statistically different from the study group). Gastric emptying rate abnormalities were significantly more frequent among diabetic patients with hypoglycaemia compared to diabetic patients without hypoglycaemia (P < 0.001). Accelerated gastric emptying was infrequent in both hypoglycaemic (one in 31) and control patients (one in 18). t 1/2 and t lag values were significantly longer in the hypoglycaemic patients (P < for both parameters), and GEC was significantly lower in hypoglycaemic patients. Moreover, we found a significant relationship between hypoglycaemic patients and the severity of gastric emptying delay. We observed an increase in the proportion of hypoglycaemic patients as the t 1/2 values increased. On the other hand, no correlation was found in our study between autonomic neuropathy and impaired gastric emptying. Some studies did not find a relationship between the rate of gastric emptying and autonomic neuropathy In other studies a relationship was found Younger age was found to be a risk factor for hypoglycaemia, but the significance (P ¼ 0.049) was weaker compared to the significance of impaired gastric emptying and hypoglycaemia. Lower basal metabolic index was also more prevalent among hypoglycaemic patients. We estimate that the more pronounced gastric emptying delay in our younger patients combined with the more dynamic and physically active life conducted by younger people, rendered them more prone to hypoglycaemia. Potential therapeutic targets (i) Prokinetics: domperidone 20 mg t.i.d. half an hour before meals was tried in 10 hypoglycaemic patients. In the first 2 weeks of treatment, improvement in the frequency and severity of the hypoglycaemic events was noted. However, after 1 month no significant improvement in gastric emptying parameters and hypoglycaemic events was noted, probably because of tolerance to the drug. Overall, available prokinetic agents in diabetic gastropathy seem at best to provide only modest efficacy. 7,16,33 Perhaps more effective prokinetics will be available in the future. (ii) Change of insulin type and timing schedule: a treatment trial in hypoglycaemic patients with impaired gastric emptying was performed, based on insulin lispro injections min after each meal combined with once-a-day administration of Glargine (Lantus) as basal treatment usually at bedtime. 34 We have preliminary encouraging results showing a trend for decrease in the frequency and severity of hypoglycaemic events. Normal gastric emptying was found in some of our patients with frequent hypoglycaemic events. This observation may suggest that other factors besides gastric emptying rate abnormalities may play a role in the aetiology of hypoglycaemia in some of our patients. All the patients with hypoglycaemic events were on multiple-injection regimens. The desire to achieve a good metabolic control by using multipleinjection regimens is a known risk factor to hypoglycaemia. 6,35 On the other hand, although some of our patients from the control group were found to have gastric emptying delay, they did not suffer from hypoglycaemic events. This can be explained by a more effective and prompt counter-regulatory response in some patients which may reduce the risk of hypoglycaemia. 36 However, the statistically significant 438 Journal compilation Ó 2006 Blackwell Publishing Ltd

7 Volume 18, Number 6, June 2006 Hypoglycaemia and gastric emptying disorders in insulin-treated diabetes increase in gastric emptying rate abnormalities in diabetic patients with frequent postprandial hypoglycaemic events compared with controls is suggestive that gastric emptying rate abnormalities do play an aetiological role in causing hypoglycaemia. CONCLUSION The increased frequency and severity of gastric emptying abnormalities in diabetic patients with unexplained hypoglycaemia suggest a causative role between gastric emptying abnormalities and hypoglycaemic events. Lack of gastrointestinal symptoms does not exclude severe gastric emptying abnormalities. Therefore, we propose that gastric emptying rate should be tested in diabetic patients with hypoglycaemic events, regardless of the absence of gastrointestinal symptoms. ACKNOWLEDGMENTS The study was supported by a grant from the Ministry of Health of Israel Chief Scientist. REFERENCES 1 Chelliah A, Burge MR. Hypoglycemia in elderly patients with diabetes mellitus: causes and strategies for prevention. Drugs Aging 2004; 21: Hershey T, Lillie R, Sadler M et al. A prospective study of severe hypoglycemia and long-term spatial memory in children with type 1 diabetes. Pediatr Diabetes 2004; 5: Heaton A, Martin S, Brelje T. The economic effect of hypoglycemia in a health plan. Manag Care Interface 2003; 16: Davis S, Alonso MD. Hypoglycemia as a barrier to glycemic control. J Diabetes Complications 2004; 18: The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications of insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: UK prospective Diabetes Study (UKPDS) Group. Intensive blood glucose-control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS33). Lancet 1998; 352: Rabine JC, Barnett JL. Management of the patient with gastroparesis. J Clin Gastroenterol 2001; 32: Jones KL, Russ A, Stevens JE et al. Predictors of delayed gastric emptying in diabetes. Diabetes Care 2001; 24: Horowitz M, Harding PE, Maddox AF et al. Gastric and oesophageal emptying in patients with type 2 (non-insulin dependent) diabetes mellitus. Diabetologia 1989; 32: Samson M, Vermeijden JR, Smout AJ et al. Prevalence of delayed gastric emptying in diabetic patients and relation to dyspeptic symptoms: a prospective study in unselected diabetic patients. Diabetes care 2003; 26: Horowitz M, Maddox AF, Wishart JM et al. Relationship between oesophageal transit and solid and liquid gastric emptying in diabetes mellitus. Eur J Nucl Med 1991; 18: Keshavarzian A, Iber FL, Vaeth J. Gastric emptying in patients with insulin requiring diabetes mellitus. Am J Gastroenterol 1987; 82: Lyrenas EB, Olsson EH, Arvidsson UC et al. Prevalence and determinants of solid and liquid gastric emptying in unstable type 1 diabetes. Relationship to post prandial blood glucose concentration. Diabetes Care 1997; 20: Mantides A, Stefanides G, Kioulanis J et al. Cutaneous electrogastrography for the assessment of gastric myoelectrical activity in type 1 diabetes mellitus. Am J Gastroenterol 1997; 92: Verme GN, Sninsky CA. Diabetes and the gastrointestinal tract. Gastroenterol Clin North Am 1998; 27: Talley NJ. Diabetic gastropathy and prokinetics. Am J Gastroenterol 2003; 98: Stacher G, Lenglinger J, Bergman H et al. Impaired gastric emptying and altered intragastric meal distribution in diabetes mellitus related to autonomic neuropathy? Dig Dis Sci 2003; 48: Ghoos YF, Maes BD, Geypens BJ et al. Measurement of gastric emptying of solids by means of a carbon-labeled octanoic acid breath test. Gastroenterology 1993; 104: Shirin H. Evaluation on novel continuous real time 13 C urea breath analyzer for Helicobacter pylori. Aliment Pharmacol Ther 2001; 15: Maes BD, Ghoos YF, Hiele MI, Rutgeers PJ. Gastric emptying rate of solids in patients with nonulcer dyspepsia. Dig Dis Sci 1997; 42: Ertl AC, Davis SN. Evidence for a vicious cycle of exercise and hypoglycemia in type 1 diabetes mellitus. Diabetes Metab Rev 2004; 20: Rayner CK, Samson M, Jones KL, Horowitz M. Relationship of upper gastrointestinal motor and sensory function with glycemic control. Diabetes Care 2001; 24: Fraser RJ, Horowitz M, Maddox AF et al. Hyperglycaemia slows gastric emptying in type 1 (insulin dependent) diabetes mellitus. Diabetologia 1990; 33: Schvarcz E, Palmer M, Aman J et al. Physiological hyperglycemia slows gastric emptying in normal subjects and patients with insulin-dependent diabetes mellitus. Gastroenterology 1997; 113: Ishii M, Nakamura T, Kasai F et al. Altered postprandial insulin requirement in IDDM patients with gastroparesis. Diabetes Care 1994; 17: O Donovan GG, Doran S, Feinle-Bisset C et al. Effect of variations in small intestinal glucose delivery on plasma glucose, insulin and incretin hormones in healthy subjects and type 2 diabetes. J Clin Endocrinol Metab 2004; 89: Annese V, Bassoti G, Caruso N et al. Gastrointestinal motor dysfunction, symptoms and neuropathy in noninsulin-dependent (type 2) diabetes mellitus. J Clin Gastroenterol 1999; 29: De Block CE, De leeuvw IH, Pelckmans PA, Callens D, Maday E, Van Gaal LF. Delayed gastric emptying and Journal compilation Ó 2006 Blackwell Publishing Ltd 439

8 J. Lysy et al. Neurogastroenterology and Motility gastric autoimmunity in type 1 diabetes. Diabetes Care 2002; 25: Russo A, Stevens JE, Chen R et al. Insulin induced hypoglycemia accelerates gastric emptying of solids and liquids in long-standing type 1 diabetes. J Clin Endocrinol Metab 2005; 90: Ziegler D, Schadewaldt P, Pour Mirza A et al. [13] C octanoic acid breath test for non-invasive assessment of gastric emptying in diabetic patients: validation and relationship to gastric symptoms and cardiovascular autonomic function. Diabetologia 1996; 39: Lehmann R, Borovicka J, Kunz P et al. Evaluation of delayed gastric emptying in diabetic patients with autonomic neuropathy by a new magnetic resonance imaging technique and radio opaque markers. Diabetes Care 1996; 19: Stacher G, Lenglinger J, Bergmann H et al. Impaired gastric emptying and altered intragastric meal distribution in diabetes mellitus related to autonomic neuropathy. Dig Dis Sci 2003; 48: Rohtashav D, Richter J. Erythromycin in the short- and long-term control of dyspepsia symptoms in patients with gastroparesis. J Clin Gastroenterol 2004; 38: Murphy NP, Keane SM, Ong KK et al. Randomized crossover trial of insulin glargine plus lispro or NPH insulin plus regular human insulin in adolescents with type 1 diabetes on intensive insulin regimens. Diabetes Care 2003; 26: Tsui E, Barnie A, Ross S et al. Intensive insulin therapy with insulin lispro: a randomized trial of continuous subcutaneous insulin infusion versus multiple daily insulin injection. Diabetes Care 2001; 24: Bolli GB, Fanelli CG. Physiology of glucose counterregulation to hypoglycemia. Endocrinol Metab Clin North Am 1999; 28: , v. 440 Journal compilation Ó 2006 Blackwell Publishing Ltd