SUBJECTS AND METHODS

Twelve healthy subjects were recruited for the study. They were H pylori positive by serology and the ¹³carbon urea breath test. They were studied before and a minimum of six weeks after cure of H pylori. During each study, they were admitted onto a research ward at 0700 hours after an overnight fast. An indwelling catheter was inserted into a forearm vein. Blood (10 ml, to which 40 IU of aprotonin were added) was sampled hourly between 0800 and 1300 hours, stored on ice during collection, centrifuged, plasma separated, aliquoted into polypropylene tubes, and stored at −20°C for later analysis. Aliquots were number coded to ensure that the assays were performed blind.

The study conditions were identical on each study day—that is, before and after H pylori cure. During each study, subjects ate a standard breakfast at 0815 hours, which consisted of a glass of fruit juice, a bowl of cereal with milk, two hard boiled eggs, two slices of toast with butter, and two cups of coffee. At 1045 hours they had two cups of tea and two biscuits.

Twenty four hour intragastric acidity was also measured by aspiration of 2–5 ml aliquots of gastric juice via a 10 Fr nasogastric tube hourly between 0800 hours and 0800 hours the next day.

Hormone assays

Plasma immunoreactive ghrelin concentrations were measured in duplicate using a commercial radioimmunoassay (Phoenix Pharmaceuticals, Belmont, California, USA). The intra-assay coefficient of variation (CV) was 8.1% and the interassay CV, 9%. The range of detection was 10–1280 pg/ml.

Plasma leptin concentrations were measured using a commercially available leptin radioimmunoassay kit (Linco, USA). The intra-assay CV was 4.9% and the interassay CV 5.3%. The range of detection was 0–100 ng/ml

Plasma gastrin concentrations were performed by an inhouse radioimmunoassay (Professor Bloom’s laboratory, Hammersmith Hospital, Du Cane Road, London, UK). The reference range for this assay in humans is 0–40 pmol/l.

H pylori cure

Immediately after the first study, subjects received a triple regimen, consisting of ranitidine bismuth citrate, 400 mg twice daily, tetracycline hydrochloride 500 mg twice daily, and clarithromycin 500 mg twice daily, for a total of 7 days. The ¹³carbon urea breath test was repeated a minimum of four weeks after the end of therapy.

Statistical analysis

Statistical analysis was performed using Analyse-it for Microsoft Excel (Leeds, UK; see http://www.analyse-it.com). Pairwise comparisons were performed using non-parametric statistical tests (Wilcoxon’s rank sum test). Group data were expressed as median and 95% confidence intervals (95% CI). Correlation was by Spearman’s rank coefficient test.

Ethical considerations

The Coventry and Warwickshire Research Ethics Committee approved this study. Subjects gave written informed consent.

RESULTS

Ten of the 12 subjects completed the study (males, n=6; females, n=4); median age 36 years (range 27–43) and median body mass index (BMI) 25.8 kg/m² (range 21–32). Median BMI remained unchanged at the end of the study. H pylori was cured in all 10 subjects.

Figure 1 shows a six hour profile of hourly median plasma ghrelin before and after H pylori cure for the 10 subjects. The profile of plasma ghrelin is as reported by others^18,21,22; highest concentration after an overnight fast, decreasing following breakfast, and increasing in anticipation of lunch. After H pylori cure, plasma ghrelin concentrations were raised above the highest concentration recorded before the cure, except at 1100 hours. Figure 2 shows that six hour integrated plasma ghrelin increased in all 10 subjects after H pylori cure. Median (95% CI) integrated plasma ghrelin increased from 1160.5 (765.5–1451) pg/ml×h to 1910.4 (1675.6–2395.6) pg/ml×h (p=0.002, Wilcoxon’s rank sum test).

Figure 3 shows a six hour plasma profile of hourly median plasma leptin before and after H pylori cure for the 10 subjects. Both profiles are similar; typically flat and unaffected by meals. Figure 4 illustrates six hour integrated plasma leptin for all subjects before and after H pylori cure. Median (95% CI) integrated plasma leptin was 23.66 (7.33–128.22) ng/ml×h before H pylori cure and remained unchanged at 26.75 (7.43–72.81) ng/ml×h after H pylori cure (p=0.375, Wilcoxon’s rank sum test).

Figure 5 shows the relationship between six hour integrated ghrelin and leptin with six hour integrated gastrin and 24 hour integrated intragastric acidity. The graph shows percentage change in each parameter after Helicobacter pylori cure. After cure, ghrelin and intragastric acidity increased significantly by 75% (p=0.002) and 14% (p=0.006), respectively. Furthermore, a significant positive correlation was noted between ghrelin and intragastric acidity (r_s=0.44, p=0.05, Spearman’s rank correlation).

After cure, leptin and gastrin decreased non-significantly by 11% and 30%, respectively.

DISCUSSION

In the present study, we have demonstrated for the first time that after successful cure of H pylori in healthy asymptomatic subjects, the median plasma ghrelin concentration increased at each time point throughout the six hour period. In common with previous investigators,^21,22 we noted that plasma ghrelin concentrations were highest before a meal, falling shortly after, suggesting that it has a role in meal time hunger. Furthermore, like nutrients such as glucose and amino acids and hormones such as cholecystokinin, ghrelin may play a role in meal initiation. In a randomised double blind crossover study, ghrelin was shown to acutely enhance appetite and increase food intake in healthy human subjects.²⁶ Moreover, our observation that integrated plasma ghrelin was increased by 75% (fig 5) six weeks after cure of H pylori lends support to the view that ghrelin could be involved in the long term regulation of body weight.^15,16 In contrast with ghrelin, the six hour profile of plasma leptin was not significantly changed. Although H pylori modifies leptin expression in the gastric mucosa²⁵ it does not appear to affect plasma leptin concentrations, making it unlikely that any association of H pylori and appetite is mediated through this polypeptide.

The aetiology of common obesity is multifactorial and includes changing lifestyle and diet. However, in rodents, administration of ghrelin leads to an increase in food intake and body weight,¹⁶ but also decreases the catabolism of fat.¹⁵ It is therefore not inconceivable, given our findings, that subjects cured of H pylori may potentially experience an increase in appetite and in the long term gain weight; an observation particularly relevant in Western populations where the incidence of H pylori is decreasing but that of obesity is increasing reciprocally. This is supported by a study which found that cure of H pylori improved nutritional parameters in patients; body weight, plasma lipid, and albumin concentration increased significantly.²⁷ In our study, we also observed a modest but significant increase in intragastric acidity after cure of H pylori, which could contribute to GORD. Therefore, increasing obesity and GORD would account for the increased incidence of Barrett’s oesophagus and OA, and ghrelin could be the missing link that explains the relative rarity of H pylori among patients with Barrett’s oesophagus and OA and the apparent “protective” effect of H pylori against OA.^1,3

It should be noted that there are no studies in adults illustrating whether H pylori positive subjects have a lower BMI than H pylori negative controls. Of interest however are data to suggest that H pylori positive children have a high incidence of growth retardation.^28–33 In this context, our findings may be particularly relevant, given that ghrelin stimulates growth hormone secretion in rodents and humans.²⁰ Also, children with H pylori may have relatively low ghrelin concentrations, contributing towards growth retardation. Thus our hypothesis suggests that H pylori positive children may fail to thrive, have a poor appetite, and decreased food intake. They grow up into underweight adults, some of whom will develop peptic ulcer disease, decreasing their food intake further. Finally, a small proportion will develop distal gastric cancer but very few will suffer heartburn or die of OA. This is the typical pattern of foregut disease seen in developing countries with a very high prevalence of H pylori. Conversely, in the developed world, children with no H pylori and relatively high concentrations of plasma ghrelin will grow to their full potential, with a good appetite and increased food intake. They may eventually grow into overweight adults, develop GORD and Barrett’s oesophagus, and in some cases OA. General factors such as lifestyle, diet, and genes would modify these patterns of foregut disease. Specific factors that could contribute to carcinogenesis at the oesophagogastric junction include the direct effect at that site of nitric oxide derived from dietary nitrate.³⁴

The mechanism by which H pylori infection leads to a reduction in plasma ghrelin concentrations remains to be elucidated. There may be a direct effect of H pylori associated inflammation on ghrelin producing cells within oxyntic glands. One possibility is that gastrin may inhibit the release of ghrelin and vice versa. Plasma gastrin decreases following H pylori cure in healthy subjects.^35–38 In our study, we did not observe a significant decrease in gastrin as we believe that the post cure study was undertaken too soon to observe complete resolution of H pylori associated hypergastrinaemia. In a recent observation in human subjects, intravenous ghrelin produced a non-significant fall in gastrin, similar to our findings.³⁹ The observed increase in intragastric acidity could contribute to GORD after H pylori cure, and may reflect recovery of parietal cells after resolution of H pylori gastritis. When administered centrally or peripherally to rodents, ghrelin stimulates gastric acid secretion.⁴⁰ Furthermore, a study in rodents has found ghrelin receptors on gastric vagal afferents which could initiate reflex gastric acid secretion.⁴¹ Based on these observations, it is possible that the increased intragastric acidity noted by us may be mediated by ghrelin either via a central pathway or locally given the close juxtaposition of parietal and ghrelin cells in the stomach. This hypothesis is supported by our observation of a significant positive correlation between plasma ghrelin and intragastric acidity.

The limitations of our study need to be addressed. Even though there was a substantial increase in ghrelin in all patients, the study was open and longitudinal in nature, with each subject acting as their own control. It is therefore important for a placebo controlled study not only to confirm our findings but also, if possible, to provide immunohistochemical data on gastric biopsy specimens before and after H pylori cure. Ghrelin administration in rodents causes weight gain via appetite stimulation and reduced fat oxidation^15,16; we did not note any change in BMI. With hindsight, we feel that waist circumference and waist to hip ratio may have provided a better picture of fat mass than BMI,⁴² given the short duration of the study. However, compared with rodents, not much is known of the relationship between ghrelin and the time course of change in body mass and body composition in humans. Certainly, given our findings, a case control study is needed to confirm whether there are differences in body composition between H pylori positive and negative subjects. Another approach would be to measure changes in fat mass, and omental and retroperitoneal fat pads, using sequential dual energy x ray absorptiometry before and after H pylori cure, with long term follow up.

In conclusion, we found that cure of H pylori increased plasma ghrelin in healthy asymptomatic subjects, which in turn may lead to increased appetite and weight gain and contribute to the increasing obesity seen in Western populations where the prevalence of H pylori is low. This study supports the hypothesis that ghrelin may be the missing link between decreasing H pylori, increasing obesity, GORD, and oesophageal adenocarcinoma that is observed in Western populations.