Digestive Diseases and Sciences

, Volume 54, Issue 10, pp 2161–2166

Ghrelin and Obestatin Levels in Type 2 Diabetic Patients With and Without Delayed Gastric Emptying


    • Department of Medicine 1, Division of Endocrinology and MetabolismFriedrich-Alexander University Erlangen-Nuremberg
  • Corinna Koebnick
    • Department of Research and EvaluationSouthern California Permanente Medical Group
  • Atingwa M. Tasi
    • Department of SurgeryKlinikum Lippe-Detmold
  • Eckhart Georg Hahn
    • Department of Medicine 1Friedrich-Alexander University Erlangen-Nuremberg
  • Peter C. Konturek
    • Department of Medicine 1Friedrich-Alexander University Erlangen-Nuremberg
Original Article

DOI: 10.1007/s10620-008-0622-2

Cite this article as:
Harsch, I.A., Koebnick, C., Tasi, A.M. et al. Dig Dis Sci (2009) 54: 2161. doi:10.1007/s10620-008-0622-2


Alterations in the neurohumoral regulation of the upper intestine may change rhythmicity and pattern of ghrelin and obestatin, the latter presumably antagonizing ghrelin effects. Five nongastroparetic diabetic patients and five with gastroparesis were investigated. Over 390 min including breakfast and lunch, ghrelin was significantly lower in patients with gastroparesis compared with in those without (P = 0.015). Ghrelin subsequent to lunch decreased significantly (P = 0.011) in patients without gastroparesis, but not in gastroparetic patients (P = 0.669). Obestatin was similar in both groups and unchanged. No significant differences in ghrelin-to-obestatin ratio were observed (P = 0.530). Loss of rhythmicity in the ghrelin levels of gastroparetic diabetics highlights the importance of integrity of the neurohumoral-intestinal axis. Stable diurnal obestatin levels do not support the concept of interaction between ghrelin and obestatin in terms of regulation of food intake and gastric emptying.


GhrelinObestatinDiabetes mellitusGastroparesis


Ghrelin is a hormone implicated in hunger and long-term regulation of body weight. Plasma ghrelin levels rise shortly before meals and fall shortly after a meal; furthermore, ghrelin levels peak between 00:00 and 02:00 h [1]. Ghrelin is primarily secreted by the stomach [2], but has also been detected at lower levels in other tissues, including the upper intestine [3].

If these anatomical structures are compromised, e.g., after gastric bypass, meal-related fluctuations and diurnal rhythm of ghrelin secretion are altered; for example, after Roux-en-Y gastric bypass, a decrease in ghrelin levels has been reported by some [1] but not all studies [4]. Data on changes in ghrelin release after bariatric surgery procedures such as gastric bypass are also conflicting, and understanding of the underlying mechanisms remains incomplete [5]. It has been hypothesized that insulin resistance before and after surgery may be the driving factor for changes in ghrelin levels [5].

The aim of this study was to determine if alterations in the neurohumoral regulation of the upper intestine do also change the rhythmicity and pattern of ghrelin secretion. Such alterations in the neurohumoral regulation are well known in patients with diabetes mellitus and diabetic neuropathy [6, 7]. Thus, we chose to investigate patients with type 2 diabetes mellitus without gastroparesis and type 2 diabetic patients with confirmed gastroparesis as a biomodel of impaired neurohumoral regulation of gastric function.

Obestatin has recently been discovered in the rat stomach. It is a novel 23-amino-acid peptide derived from the processing of the ghrelin gene. The peptide name was in keeping with its initially reported actions to suppress food intake and digestive motility and to antagonize ghrelin’s stimulatory effect through interaction with the orphan GPR-39 receptor. Obestatin has been claimed to be a functional opponent of ghrelin [8].

To date, this concept is more and more under discussion; e.g., in a mouse and a rat model, intestinal and fundic smooth muscle strips did not respond to obestatin either in the absence or in the presence of electrical field stimulation. Obestatin also did not inhibit fasting-induced hyperphagia [9]. In another study, ghrelin receptor function was studied using a GTPgammaS assay and transfected cells. Obestatin did not modulate the recombinant ghrelin receptor function. Ghrelin did increase gastric emptying, whereas obestatin had no effects [10]. Ghrelin also increases gastric emptying in diabetic mice with gastroparesis [11].

This growing body of data from animal models suggests that peripheral obestatin does not play an important role in regulation of gastric emptying, nor does it support the concept that obestatin is a physiological opponent of ghrelin with respect to gastric emptying. However, circulating preprandial ghrelin-to-obestatin ratio is elevated in human obesity [12]. These findings suggest that high preprandial ghrelin-to-obestatin ratio may be involved in the etiology and pathophysiology of obesity.

Considering these contradictory findings, we decided to determine peripheral serum levels of obestatin too, and to try to clarify possible interactions of ghrelin and obestatin during daytime in the aforementioned patient groups.


Participant Characteristics

We studied five patients with type 2 diabetes mellitus without diabetic gastroparesis and five type 2 diabetic patients with confirmed diabetic gastroparesis.

In the patient group with gastroparesis, mean duration of disease was 11.5 ± 9.8 years (range 1.5–27 years). Four of the patients were treated with insulin, one additionally with metformin, and one patient was treated with metformin and glimepiride. Four of the patients were treated with antihypertensive drugs, one with a statin. Retinopathy was detected in two patients. In the patient group without gastroparesis, four of the patients were treated with insulin, three additionally with metformin, and one patient was treated with repaglinide. Three of the patients were treated with antihypertensive drugs, one with a statin. Mean duration of disease was 5.5 ± 3.6 years (P = 0.234 compared with the gastroparetic group). Retinopathy was detected in two patients. Further characteristics of the patients are given in Table 1. Exclusion criteria were: history of recurrent ulcers or gastric tumors (since scars and fibrous tissue from ulcers and tumors may block the outlet of the stomach and mimic gastroparesis); drugs altering stomach emptying such as tricyclic antidepressants, metoclopramide or erythromycin; calcium blockers such as diltiazem, nifedipine or l-dopa; history of previous stomach surgery; hypothyroidism, anorexia, or bulimia; neurologic or brain disorders such as Parkinson’s disease, strokes, or brain injury; and history of lupus erythematosus and sclerodermia.
Table 1

Patient characteristics


Diabetic gastroparesis patients

Diabetic controls

P value









Age (years)

53.4 ± 5.5

61.2 ± 15.0


Body mass index (kg/m2)

28.9 ± 9.6

29.9 ± 5.5


HbA1c (%)

9.4 ± 2.5

10.8 ± 2.2


Fasting ghrelin (ng/l)

179 ± 56

249 ± 60


Fasting obestatin (ng/l)

252 ± 87

257 ± 23


Insulin dosage (sampling day) (IE/day)

50.4 ± 18.4

46.0 ± 13.9


Half time of gastric emptying (min)

191 ± 36

67 ± 12


Values are mean ± standard deviation (SD)

The study was approved by the Ethics Committee of the University Erlangen-Nuremberg, and each patient gave written informed consent.

Gastrointestinal Symptoms

The patients were asked for usual symptoms of gastroparesis such as feeling of fullness after only a few bites of food, bloating, excessive belching, and nausea. These symptoms were assessed by a questionnaire as described by DeBlock et al. [6].

Autonomic Nerve Function

All patients underwent neurological examination and autonomic function tests. Possible presence of peripheral neuropathy was investigated by standard clinical examinations (monofilament test, peripheral reflexes, Rydel-Seiffer tuning fork, temperature discrimination). Furthermore, if the patients had not already undergone subtle testing for autonomic neuropathy according to international standards during the last 12 months, the patients underwent autonomic function tests as described by DeBlock et al. [6]. Parasympathetic function was evaluated by variation of heart rate during deep breathing and immediate heart rate response to standing and the Valsalva maneuver. Sympathetic function was assessed by fall in systolic blood pressure in response to standing. Classification was as follows: borderline cardiovascular autonomic neuropathy (CAN) (one heart rate test abnormal), definite CAN (≥2 heart rate tests abnormal), severe CAN (all heart rate and blood pressure tests abnormal).

Gastric Emptying

In all patients in our study, after overnight fast, a gastric emptying test was performed to confirm the diagnosis. In this test, a pancake was labeled with 80 MBq/2.2 mCi Tc-Sn colloid. A scanner placed ventral to the standing patient’s thorax and abdomen monitored deglutition, and esophageal and gastric transport. Passage time >120 min was considered pathologic. The half-time of gastric emptying is given in Table 1. As for glycemic control, on the day before blood sampling as well as on the sampling day, the patients were treated with insulin alone. They were not treated with oral antidiabetic drugs to minimize possible effects on gastric emptying and ghrelin/obestatin levels. The insulin dosage on the sampling day is given in Table 1. To eliminate the effect of hyperglycemia on gastric emptying, blood glucose levels between 80 and 120 mg/dl (4.44–6.66 mmol/l) were regarded as acceptable and the patients received the standard morning insulin dose about 30 min before the pancake meal. At 30-min intervals, capillary glucose levels were measured by fingerstick (Accu Check Roche Diagnostics GmbH) by the patients. If glycemia exceeded 126 mg/dl (7 mmol/l), intravenous short-acting insulin was given according to a protocol as given by DeBlock et al. [6].

Blood Samples/Meals

Breakfast and lunch were served as part of a standard diet for diabetic patients and varied according to the days of hospital admission. The breakfast usually consisted of bread, butter, cheese, sausage, jam, yogurt, and fruits and had a maximum of 2 MJ. Lunch consisted of meat, vegetables, potatoes, pasta or rice and had a maximum of 2.5 MJ. The energy composition of both meals was similar to a standard diet (50% of energy as carbohydrate, 37% of energy as fat, 13% of energy as protein, and 1.1 g/MJ dietary fiber).

Blood samples were taken at six different time points: fasting (07:30–07:45), 1 h after breakfast (09:00), 2 h after breakfast (10:00), before lunch (12:00), 1 h after lunch (13:30), and 2 h after lunch (14:30).

The samples were collected in ethylenediamine tetraacetate-coated polypropylene tubes kept on ice, centrifuged immediately at 3,000 rpm for 20 min at 0°C, and the clear plasma supernatant was then stored until plasma active ghrelin and obestatin levels were measured using radioimmunoassays (RIAs) (Ghrelin active, Linco; Obestatin (human & monkey), Phoenix Pharmaceuticals). All plasma samples were analyzed in duplicate and in the same batch.

Statistical Analysis

All statistical analyses were performed with the statistical package SPSS 16.0 (SPSS Inc., Chicago, IL). Data are given as mean ± SD in text and tables; range is provided for baseline characteristics. Data in graphs are given as mean ± standard error (SE).

Linear mixed-effects modeling was used to estimate the effects of gastroparesis on ghrelin and obestatin levels. The intercept represents the fasting levels of ghrelin and obestatin. For all models, fixed effects included were patient group (gastroparesis versus diabetic); all models were performed with and without adjustment for age and body mass index (BMI). The effect was tested for significant fixed effects on the intercept and slope. Fasting levels of ghrelin and obestatin were compared between diabetic patients with and without gastroparesis by independent-sample Student’s t-test. Single time points (before and after meal initiation) were compared by paired Student’s t-test. The differences were considered significant at P < 0.05 for all analyses.


Diabetic patients with and without gastroparesis were comparable in terms of BMI, insulin dosage, and hemoglobin A1c (HbA1c) (Table 1). All patients with gastroparesis were also considered to have autonomic neuropathy after autonomic nerve function testing (definite CAN in three cases, severe CAN in two cases). In the nongastroparetic group, two patients had borderline CAN. Although not statistically significant, diabetic patients without gastroparesis were slightly older than those with gastroparesis.

The values of ghrelin and obestatin throughout the sampling period are given in detail in Table 2. Fasting acylated ghrelin was similar in diabetic patients with and without gastroparesis (P = 0.096). Breakfast did not initiate a significant drop in acylated ghrelin in either study group (time 0 versus time 60: P = 0.133 for patients without gastroparesis, 0.966 for patients with gastroparesis).
Table 2

Crude and adjusted plasma ghrelin and obestatin

Time (min)

Diabetic gastroparesis patients

Diabetic controls

P value

Active ghrelin (ng/l)


179 ± 25

249 ± 27



178 ± 19

210 ± 18



162 ± 12

222 ± 34



193 ± 19

276 ± 24



200 ± 20

205 ± 25



214 ± 27

234 ± 20


Obestatin (ng/l)


252 ± 39

257 ± 10



250 ± 48

243 ± 16



245 ± 50

243 ± 22



244 ± 47

253 ± 25



233 ± 43

237 ± 22



242 ± 46

249 ± 22


Data are mean and SE. P values are based on Student’s t-test

Acylated ghrelin before lunch, however, was lower in diabetic patients with gastroparesis (mean ± SD, 276 ± 55) compared with in diabetic patients without gastroparesis (193 ± 43, P = 0.029). Acylated ghrelin subsequent to lunch decreased significantly by 70 ± 36 pg/l (P = 0.011) in diabetic patients without gastroparesis, whereas no significant change in acylated ghrelin was observed in diabetic patients with gastroparesis (P = 0.669).

Over the time course of 390 min, including breakfast and lunch consumption, acylated ghrelin was significantly lower in diabetic patients with gastroparesis compared with in those without gastroparesis (Fig. 1) without (P = 0.015) and with (P < 0.001) adjustment for age and BMI.
Fig. 1

Active (acylated) ghrelin and obestatin levels in diabetic patients with (n = 5) and without gastroparesis (n = 5). Over the course of 390 min, active ghrelin (P < 0.001) but not obestatin levels (P = 0.911) were significantly different after adjustment for BMI and age. *Significant difference (P < 0.05)

Obestatin was similar between diabetic patients with and without gastroparesis and did not change over the time course of 390 min including breakfast and lunch consumption (P = 0.911). Comparably, no significant differences in acylated ghrelin-to-obestatin ratio between diabetic patients with and without gastroparesis were observed (P = 0.530).


Our data do not support the concept of an interaction between acylated ghrelin and obestatin in terms of regulation of food intake/gastric emptying in diabetic patients. Consistent with the findings reported in the literature, acylated ghrelin increased preprandially and was suppressed after meal intake in the diabetic patients without gastroparesis [13]. Interestingly, rhythmicity was abrogated in our subjects with neuropathy and gastroparesis.

These findings are in accordance with a study of Gaddipati et al. [14]. Those authors studied the ghrelin and pancreatic polypeptide responses in healthy subjects and patients with diabetic, idiopathic, or postsurgical gastroparesis after sham feeding as a measure of vagal integrity. The authors reported that sham feeding was characterized by an increase in pancreatic polypeptide and ghrelin in healthy controls and in patients with idiopathic gastroparesis. The changes in pancreatic polypeptide and ghrelin levels in patients with diabetic and postsurgical gastroparesis were significantly less than those in normal subjects. Furthermore, the systemic ghrelin concentrations increased with sham feeding in normal subjects and patients with idiopathic gastroparesis but not in diabetic or postsurgical gastroparesis. These findings highlight the importance of vagal nerve function in the regulation of ghrelin levels. It is not clear what precautions those authors took in their study to erase or minimize the possible effects of antidiabetic therapy, as well as concomitant medication on gastric emptying and on ghrelin/obestatin. We tried to minimize such possible effects by withdrawing oral antidiabetic drugs the day before sampling and on the sampling day. As for possible effects on gastric emptying, we could not withdraw antihypertensive premedication due to ethical considerations. However, although effects of these drugs cannot be completely excluded, they do not seem to play a major role, and the withdrawal of antihypertensives is not recommended in recent consensus recommendations for gastric emptying scintigraphy [15].

Interestingly, in a recent review [16] the authors reported that, depending upon the experimental condition, ghrelin as well as obestatin have been reported to be able to either inhibit or stimulate insulin secretion in animals and humans. In the case of ghrelin, most available data suggest a negative association between systemic ghrelin and insulin levels. Given the nature of biological feedback systems, effects of the (exogenously administered) insulin cannot be excluded. However, since all patients were insulin-treated with comparable doses to achieve euglycemia (Table 1), a possible effect would only account for a systematic error in our investigation.

For obestatin, we did not observe any diurnal or postprandial rhythmicity in either patient group. Furthermore, there was no interaction detectable with the ghrelin levels. This opposes the concept of a direct interaction between ghrelin and obestatin but does not exclude long-term effects of obestatin on food intake and energy regulation. The low patient number in our study restricts a correlation analysis between ghrelin and obestatin and a 24-h analysis would have been desirable. However, Fig. 1 shows a low degree of interindividual variability between the hormone levels in our two homogenous patient groups. Thus, we regard our conclusions justified.

It has recently been demonstrated that ghrelin stimulates gastrointestinal motility in healthy subjects [17, 18]. Given neuropathy present in diabetic subjects, these mechanisms seem to be defective in such individuals. Interestingly, plasma levels of ghrelin in gastroparetic diabetic patients are lower than those of the nongastroparetic group. We hypothesize that the lower ghrelin levels in the gastroparetic group are a direct consequence of the defective innervation of their stomach, since there are otherwise no salient differences between the two groups, for example, in terms of medication and comorbidities. This could be among the reasons why gastroparetic patients do not only report symptoms such as nausea, abdominal bloating/fullness, vomiting, abdominal pain, dysphagia, heartburn, and acid regurgitation, but also early satiety [3]. Ueno et al. also reported lower ghrelin levels in patients with diabetic complications. In patients with diabetic triopathy, plasma ghrelin concentrations were significantly lower than in patients without diabetic complications [19]. Basal ghrelin levels are usually lower in obesity [20], but this cannot account for the lower basal ghrelin levels in our gastroparetic group, since the two patient groups in our study are well matched for BMI.

In contrast, we did not observe a decrease in postprandial obestatin levels as recently reported by Guo et al. [12]. This study group investigated 16 obese and 14 normal-weight healthy individuals and the situation may be different in diabetic individuals. Those authors also implicated increased preprandial ghrelin-to-obestatin ratio as involved in etiology of obesity, a concept challenged by the observation that this also occurs in anorexia nervosa [21].

Our patient group with long-standing diabetes mellitus can be considered highly insulin resistant. The observations concerning the postprandial course of obestatin are compatible with a recent Austrian [22] study that investigated the association between insulin resistance and plasma concentrations of obestatin and ghrelin in nondiabetic individuals with high and low degree of insulin insensitivity. After a clamp insulin infusion, which is somehow comparable to the physiological response to food intake, plasma obestatin decreased to approximately 81% of basal values in insulin-sensitive persons but not in insulin-resistant individuals. The regulatory function of obestatin in terms of food regulation are still a matter of debate. Other functions of obestatin were recently demonstrated: Granata et al. [23] investigated the effect of obestatin on survival of beta-cells and human pancreatic islets and the underlying signaling pathways in cell cultures. Their results suggested that obestatin promotes beta-cell and human islet cell survival and stimulates expression of main regulatory beta-cell genes. Furthermore, in a rat model, a sleep-promoting effect of centrally administered obestatin was demonstrated [24], as well as an improvement in memory retention (using two different paradigms), and anxiety-like behavior (plus maze test), after icv obestatin injection [25]. Furthermore, Camina et al. have evaluated the effect of obestatin on cell proliferation in primary cultures of human retinal epithelial cells (hRPE cells). The results showed that this peptide induced, in a dose-dependent manner, cell proliferation by mitogen-activated protein kinase/extracellular signal-regulated kinase 1/2 phosphorylation [26].

These findings, although still derived from animal models and cell cultures, suggest that obestatin also exerts important other functions in the organism. With our data not supporting the concept of interaction between ghrelin and obestatin in terms of food intake and energy regulation, at least in diabetic patients, it will be a fascinating future issue to clarify what the main effects of obestatin are.

Copyright information

© Springer Science+Business Media, LLC 2008