| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Section of Endocrinology and Metabolism, Department of Internal Medicine, Endocrinology and Metabolism and Biochemistry (S.C., A.M., L.P., C.C., F.S., C.F., F.P.), and Department of Surgery and Bioengineering (G.C.), University of Siena, 53100 Siena, Italy
Address all correspondence and requests for reprints to: Furio Pacini, Section of Endocrinology, Department of Internal Medicine, Endocrinology and Metabolism and Biochemistry, University of Siena, Viale Bracci 1, 53100 Siena, Italy. E-mail: pacini8{at}unisi.it.
| Abstract |
|---|
|
|
|---|
Methods: We studied 233 patients with autoimmune gastritis and 211 control subjects. All patients and control subjects were screened for circulating parietal cell antibodies (PCAs) and were tested for serum ghrelin, gastrin, pepsinogen I and II, and anti-Helicobacter pylori antibody levels. A total of 52 patients and 28 control subjects underwent a gastric endoscopy.
Results: In PCA/positive patients, mean (±SD) serum ghrelin levels were significantly lower (238 ± 107 pmol/liter), and mean (±SD) serum gastrin levels were significantly higher (81.2 ± 128.3 ng/ml), with respect to PCA/negative patients (282 ± 104 pmol/liter and 20.7 ± 13.3 ng/ml, respectively; P < 0.0001). Serum ghrelin and gastrin levels were inversely correlated (P = 0.004). A total of 40 patients had ABG documented by the gastric biopsy (90% in PCA/positive patients and 10% in PCA/negative patients). The receiver operating characteristic curve analysis revealed that a cutoff value for serum ghrelin of 188 pmol/liter was associated with the highest sensitivity and specificity (97.3 and 100%, respectively) in detecting gastric atrophy and was superior to gastrin (P = 0.012), PCA (P = 0.002), and the pepsinogen I/II ratio (P = 0.016) measurements.
Conclusions: Our study demonstrates that ghrelin secretion is negatively affected by autoimmune gastritis, and its serum level represents the most sensitive and specific noninvasive marker for selecting patients at high risk for ABG.
| Introduction |
|---|
|
|
|---|
Ghrelin is a novel 28-amino acid peptide isolated from the stomach of rats and humans (22, 23). This peptide is an endogenous ligand for the GH secretagogue receptor, and it is involved in the regulation of the GH, food intake, fat storage, and body weight (24, 25, 26). The principal source of ghrelin synthesis is in the endocrine cells (X/A-like cells) of the oxyntic gland-containing portion of the stomach (23, 27). It is possible that persistent chronic damage of the gastric mucosa might affect ghrelin production. Indeed, atrophic gastritis, induced by H. pylori infection, has been associated with impaired gastric ghrelin production and the reduced concentration of serum ghrelin (28, 29, 30, 31, 32, 33, 34). The distribution of ghrelin-immunoreactive cells, in the gastric oxyntic glands, is in close contact with the parietal cells (35), the target of PCAs in patients with autoimmune gastritis. Based on this we suggest that such autoimmune damage may affect the production and release of ghrelin.
To investigate this possibility, we designed the present study aimed at determining serum levels of ghrelin in a population of PCA/positive patients and to assess the diagnostic accuracy of serum ghrelin measurement in predicting the presence of gastric atrophy in a noninvasive way.
| Patients and Methods |
|---|
|
|
|---|
We studied 233 patients (mean age ± SD, 58 ± 13 yr, range 23–82; female to male ratio 7.2:1) with elevated circulating PCA levels. Autoimmune thyroiditis, documented by typical thyroid sonographic changes and elevated levels of serum thyroid peroxidase and/or antithyroglobulin antibodies, was found in 174 (74.6%) patients. In addition, we studied 211 control subjects (mean age ± SD, 49.4 ± 16 yr, range 23–78; female to male ratio 7:1) seen in our outpatient clinic for unconfirmed suspicion of thyroid disorders or with autoimmune thyroiditis, all with normal levels of serum PCAs. In particular, 140 had no thyroid or other organ-specific autoimmune disorders, and 71 had autoimmune thyroiditis. These patients were allowed to enter the control group because the study group (PCA/positive) had a significant proportion of subjects with autoimmune thyroiditis. All patients and control subjects with autoimmune thyroiditis were either euthyroid or hypothyroid in replacement therapy using L-T4.
In all subjects, we measured serum ghrelin, gastrin, PG I, PG II, and anti-H. pylori antibodies.
The exclusion criteria for the study were: a history of recent intake (within 3 months) of drugs interfering with gastric function; previous therapy for H. pylori infection; a history of previous gastric surgery; and malignancy and major chronic diseases, alcohol abuse, and body mass index (BMI) more than 30 kg/m2.
Informed consent for the study was obtained from all patients.
Hormonal assays
PCAs were determined in the serum using an immunoenzymatic assay that recognized the
- and ß-subunits of pork H+/K+ATPase (CHEMACTIL, Angri, Italy) with a normal upper limit (mean + 2 SD values of control subjects) of 14 U/ml.
Serum ghrelin was tested by RIA after overnight fasting (Mediagnost, Reutlingen, Germany), and the normal range given by the manufacturer is 180–420 pmol/liter. We determined to recalculate the normal range in our population of normal subjects and found a range that overlapped that of the manufacturer, and we selected the lower limit as the mean – 2 SD values (188 pmol/liter).
Serum gastrin levels were tested by RIA after overnight fasting (Euro-Diagnostica AB, Malmö, Sweden) with a normal upper limit (mean+ 2 SD values of control subjects) of 39.3 pg/ml. Serum PG I and II were tested by RIA after overnight fasting (Epitope Diagnostics, San Diego, CA) with a normal lower limit (mean – 2 SD values of control subjects) of 45 and 6 ng/ml, respectively. The normal range of the PG I/II ratio in our control subjects is 4 to 25.
IgM class antibodies against H. pylori were detected with an ELISA (DRG, Marburg, Germany): an H. pylori IgM index of one or more, calculated as the ratio between the OD of the serum sample divided by the OD of a standard calibrator, was considered as seropositive.
Free T3 and free T4 were measured by chemiluminescent assay (Beckman Coulter, Inc., Fullerton, CA), and serum TSH was determined by chemiluminescent assay (Euro DPC Ltd., Llanberies, Gwynedd, UK); serum antithyroglobulin and thyroid peroxidase antibodies were determined by chemiluminescent assay (Euro DPC Ltd.).
Blood samples were taken between 0800 and 1000 h, after overnight fasting; the serum was immediately separated by centrifugation and stored in 0.5 ml aliquots at –20 C until tested.
A thyroid ultrasound was performed using a color Doppler apparatus (Technos MP; Esaote Biomedica, Florence, Italy) with a 7.5-MHz linear transducer to confirm the typical features of autoimmune thyroiditis.
Endoscopy and biopsy
A subgroup of 52 PCA/positive patients and 28 PCA/negative subjects accepted to undergo gastric endoscopy. In the study group, the indication was limited to the presence of positive PCAs with or without dyspeptic symptoms, dysphagia, unexplained loss of weight, anemia, thoracic palpitations, and atypical chest pain. In the control group, the indication for endoscopy was only the presence of the aforementioned symptoms, and they were present in 13% of the total. Three gastric biopsies were taken from the lesser curve of the antrum 3-cm proximal to the pylorus, and three biopsies were taken from the midpoint of the greater curvature, using standard biopsy forceps. The biopsy specimens were stained with hematoxylin and eosin, Alcian blue periodic acid-Schiff, and modified Giemsa stains. The specimens were examined in a blinded manner by a pathologist, and scored to indicate absent, mild, moderate, and severe, including those with metaplasia, atrophy in accordance with the Sydney system (36).
Statistical analysis
Because serum ghrelin, gastrin, and PG values do not conform to a normal distribution, statistical analysis was performed using nonparametric tests. The Mann-Whitney U rank test provided group comparison, and Spearmans rank test was used to correlate the variables. Fischers exact test was applied to evaluate differences in categorical data frequencies. Cutoff values between different markers of gastric damage were defined using receiver operating characteristic (ROC) curves computed by an appropriate software package MedSnap version 2.0 (Butterworth-Heineman, Surrey, UK). This software also supplied pairwise comparison of areas under ROC curves (AUCs) and gave a 95% confidence interval (CI) of AUC. A P value less then 0.05 was considered statistically significant.
Confirmatory data. An additional group of 24 patients and control subjects whose data were collected after completion of the cohort study was enrolled to confirm and reinforce the validity of the ROC curve analysis. These testing subjects did not differ in age, gender, BMI, and thyroid function to the original study subjects. The same statistical analysis was performed on this testing data group.
| Results |
|---|
|
|
|---|
PCA/positive and PCA/negative subjects (Table 1
) did not differ in age, gender, BMI, and thyroid function. By definition PCA levels were elevated in the PCA/positive group. A second test for PCAs was available in nearly 80% of our study group and controls. The results of the second test confirmed the previous measurement in all patients.
|
|
A PG I/II ratio less than three is usually considered a sensitive marker of ABG. As shown in Fig. 1C
, the PG I/II ratio was less than three in 24.2% of PCA/positive and 16.9% of PCA/negative patients (P = 0.04).
To understand the possible role of an active H. pylori infection on autoimmune gastric damage, we screened for the presence of anti-H. pylori antibodies. The prevalence of H. pylori infection was found equally in the two groups, 11% in PCA/negative and 17.4% in PCA/positive patients (P = 0.1). When we excluded subjects with H. pylori antibodies, the significant decrease of serum ghrelin observed in PCA/positive patients was maintained (Fig. 1D
), suggesting that H. pylori infection was not responsible for the reduction of serum ghrelin. The same conclusion applied also to the changes observed for serum gastrin and the serum PG I/II ratio (data not shown).
Ghrelin levels and thyroid function
Serum ghrelin levels were not significantly different in PCA/positive patients with or without autoimmune thyroiditis (238 ± 107 vs. 238 ± 108 pmol/liter; P = 0.9) and in PCA/negative patients with or without autoimmune thyroiditis (284 ± 104.5 vs. 282.9 ± 82.5 pmol/liter; P = 0.7).
Furthermore, in PCA/positive patients with autoimmune thyroiditis, serum ghrelin levels were not correlated to the levels of antithyroid antibodies (data not shown).
In patients with autoimmune thyroiditis, again no difference was found between serum ghrelin levels in euthyroid patients and in hypothyroid patients on L-T4 replacement therapy (257 ± 113 vs. 275 ± 178 pmol/liter; P = 0.2).
Hematological findings
A complete blood cell count was available in 110 PCA/positive and 98 PCA/negative patients. Increased mean corpuscular volume (>95 fl) was found in 21 of 110 PCA/positive and three of 98 PCA/negative patients. Mean corpuscular volume was significantly higher (P = 0.003) in PCA/positive patients (90 ± 7 fl) with respect to PCA/negative patients (86 ± 3 fl). In addition, an inverse correlation was found between mean corpuscular volume values and serum ghrelin levels (P = 0.03) in PCA/positive patients.
Histological findings
A gastroscopy with multiple biopsies was performed in a total of 80 subjects (52 PCA/positive and 28 PCA/negative). A total of 40 patients had ABG (90% PCA/positive and 10% PCA/negative). In the PCA/positive group, there was no strict correlation between symptoms and gastric damage; in fact, only 30% of the patients with documented gastric atrophy presented symptoms. On the contrary, in PCA/negative patients, findings of gastric atrophy were always associated with specific symptoms. However, the presence of symptoms was not always associated with gastric damage. As expected, gastric atrophy was significantly associated with hypergastrinemia (Fig. 2B
) and low serum ghrelin levels (Fig. 2A
). Gastric atrophy was also associated with a lower PG I/II ratio, although not with statistical significance (Fig. 2C
).
|
|
|
|
ROC analysis on the 24 testing patients confirmed the results of the main study group. Ghrelin AUC was equal to 0.95 that is within the 95% CI of original patients. A cutoff value of 187 pmol/liter was confirmed to give the best sensitivity (95%) and specificity (100%) in diagnosing atrophic gastritis.
| Discussion |
|---|
|
|
|---|
As in previous studies (7, 8, 9, 10, 11, 12, 13), we found that high levels of serum gastrin and a low PG I/II ratio are associated with the presence of positive/PCA and gastric damage. The original finding of our study is that low levels of serum ghrelin are also associated with positive/PCA and represent an additional marker of gastric damage in autoimmune gastritis. Of particular interest is the fact that the measurement of serum ghrelin levels is the most sensitive and specific tool in predicting gastric damage, significantly superior to the measurements of PCA levels, the PG I/II ratio, and serum gastrin.
Analysis of the ROC curve proved the high accuracy of the serum ghrelin measurement in recognizing patients at high risk for clinical gastric damage. Thus, ghrelin represents the most specific and sensitive noninvasive marker among those studied.
The occurrence of impaired ghrelin secretion by the X/A-like cells in patients with PCAs may be explained by a close linkage between ghrelin secreting cells and parietal cells in the oxyntic glands, as demonstrated by immunohistochemical studies in humans and Mongolian gerbils (32, 35).
The presence of H. pylori infection and autoimmune thyroiditis might be an interfering factor in our study population. We excluded this possibility by measuring anti-H. pylori antibodies and demonstrating that in PCA/positive patients, serum ghrelin changes were independent of a concomitant H. pylori infection.
In addition, we demonstrated that serum ghrelin was not affected by the presence of autoimmune thyroiditis, antithyroid antibody levels, or thyroid function. In particular we were not able to confirm recent data reporting a reduction of serum ghrelin levels in hypothyroid patients on replacement L-T4 therapy (37).
In conclusion, our study demonstrates that serum ghrelin levels are negatively affected by autoimmune gastritis, and represent the most sensitive and specific noninvasive markers for selecting those patients at high risk for having gastric damage that need to undergo further investigation, including gastric biopsy.
|
| Acknowledgments |
|---|
| Footnotes |
|---|
Disclosure Information: The authors have nothing to disclose.
First Published Online August 21, 2007
Abbreviations: ABG, Atrophic body gastritis; AUC, area under receiver operating characteristic curve; BMI, body mass index; CI, confidence interval; PCA, parietal cell antibody; PG, pepsinogen; ROC, receiver operating characteristic.
Received May 2, 2007.
Accepted August 9, 2007.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
K. E L McColl, H. Watabe, and M. H Derakhshan Role of gastric atrophy in mediating negative association between Helicobacter pylori infection and reflux oesophagitis, Barrett's oesophagus and oesophageal adenocarcinoma Gut, June 1, 2008; 57(6): 721 - 723. [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |