Heightened visceroperception and a decreased duodenal motor response to intraduodenal acid infusion have been reported in functional dyspepsia.
Heightened visceroperception and a decreased duodenal motor response to intraduodenal acid infusion have been reported in functional dyspepsia.
To investigate the effect of treatment with a proton pump inhibitor on sensorimotor impairment in 19 patients with functional dyspepsia.
Patients were assigned double-blind to pantoprazole (n=10) or placebo (n=9) treatment for 2 weeks. Antropyloroduodenal manometry was performed before and after treatment, using a 21-channel catheter, and the responses to intraduodenal infusion of 5 mL of saline and acid were assessed. Nausea, fullness and epigastric pain were scored before and after each infusion.
Acid induced a modest duodenal motor response and suppression of antral pressure waves, not altered by either treatment. However, acid evoked isolated pyloric pressure waves after pantoprazole treatment (P < 0.02), and not after placebo. Saline induced no motor response. Acid (not saline) induced nausea, both before and after treatment in both groups (all P < 0.05). Subgroup analysis of the seven acid-hypersensitive patients (37%) showed a tendency towards a decrease in nausea in all four pantoprazole-treated patients (P=0.07), in contrast to the three placebo-treated patients (P=1.0).
In functional dyspepsia, pantoprazole influenced the acid-induced duodenogastric feedback mechanism, but not the impaired duodenal motor response. Duodenal acid hypersensitivity was decreased to some extent.
Recently, duodenal acid hypersensitivity has been recognized as one of the factors in the pathophysiology of functional dyspepsia.1 In a subset of patients with functional dyspepsia, intraduodenal infusion of a small amount of acid (5 mL, 0.1 M HCl) induced nausea. In addition, the duodenal motor response to acid was decreased in patients with functional dyspepsia. These sensorimotor alterations were found to be chemospecific; they did not occur during infusion of physiological amounts of saline or lipid of similar osmolality and volume.2
Although it has been shown that gastric acid secretion is normal in patients with functional dyspepsia,3 a subset of these patients benefits from strong acid suppression by a proton pump inhibitor.4 Acid secretion inhibitory drugs are therefore widely prescribed to patients with functional dyspepsia all over the world, but the underlying mechanisms of their effect are unknown. This effect might be brought about by the elimination of a noxious stimulus, but a more prolonged beneficial effect of acid inhibition might be caused by a reduction of duodenal mucosal sensitivity for acid by a process of desensitization. It has been shown that the oesophageal mucosa can be sensitized by acid infusions.5 Similarly, a hypersensitive duodenal mucosa could be one that has become sensitized by prolonged acid exposure. Finally, acid suppression might have an effect on the impaired duodenal motor response to acid, thus normalizing acid clearance and exposure to acid of the duodenum.
This study therefore aimed to assess whether treatment with a proton pump inhibitor affects duodenal perception and antropyloroduodenal motor response to acid infusion in functional dyspepsia. In this study, pantoprazole was used, a proton pump inhibitor that has been proven to suppress gastric acid secretion profoundly at a daily dose of 40 mg.6, 7 To conduct this study, a 21-channel manometric catheter assembly was used, enabling the organization of antropyloroduodenal motility to be investigated in great detail.
Twenty patients with functional dyspepsia (13 women, seven men; mean age, 43 years; range, 21–63 years) were enrolled in the study. All patients had suffered from severe dyspeptic symptoms, such as nausea, vomiting, epigastric pain, belching, fullness or early satiety, for at least 3 months, not necessarily consecutive, within the preceding 12 months.8 The severity of the symptoms was scored according to the following grading system: 0=none, 1=mild (symptoms can be ignored), 2=moderate (symptoms cannot be ignored, but do not influence daily activity), 3=severe (symptoms influence daily activity). Only patients with a mean score above 1.5 were included. The symptom scores (mean ± S.E.M.) of these 20 patients were: 2.6 ± 0.2 for nausea, 1.8 ± 0.3 for vomiting, 2.7 ± 0.1 for epigastric pain, 1.1 ± 0.2 for belching, 2.4 ± 0.2 for fullness and 2.4 ± 0.2 for early satiety. Organic disease was ruled out by means of physical examination, upper gastrointestinal endoscopy and upper abdominal ultrasonography. All patients were negative for Helicobacter pylori at the time of the study, tested by means of a negative rapid urease test result (Hpfast, Camphill, PA, USA) or a negative H. pylori serology test. Patients had no history of gastrointestinal surgery and had not used medications known to affect gastrointestinal motility, gastric acid secretion or (visceral) perception within 2 weeks prior to the study or during the study period. Six of the 20 patients were regular smokers, smoking less than 20 cigarettes per day. Five patients consumed alcohol with a maximum of 20 g per day. All patients refrained from the use of alcohol and tobacco for a minimum of 24 h prior to each experiment.
Written informed consent was obtained from all patients, and the study protocol was approved by the Ethics Committee of the University Medical Center, Utrecht.
Patients were assigned in a double-blind randomized fashion to receive pantoprazole, 40 mg once daily, or matching placebo for a period of 2 weeks. Identical experiments were carried out twice in each patient, before and at the end of the treatment period. Compliance was verified by tablet count at the end of the study, and patients were allowed to take antacids (Maalox sachets) if required until 2 days before the second experiment.
After an overnight fast, the manometric catheter was introduced transnasally and positioned across the pylorus under fluoroscopic control. The position of the catheter was monitored by measurement of the antroduodenal transmucosal potential difference (see below). Subjects were in a supine position with the head of the bed tilted at an angle of 30° during the experiment. After a stabilization period of at least 10 min, the experiment was started during a well-defined phase II.
Infusions of 5 mL acid (0.1 M HCl solution) and 5 mL saline (0.9% NaCl) were administered over 1 min via an infusion port in the duodenal bulb and followed by a washout period of at least 15 min. Each infusion was performed twice, and was administered in randomized order and in a double-blind fashion.
A 21-channel, water-perfused silicone rubber catheter (outside diameter, 4 mm; length, 225 cm; channel diameter, 0.4 mm) was used to record antropyloroduodenal pressures, to measure the transmucosal potential difference and for infusions (Dentsleeve Pty Ltd, Belair, Australia). The assembly incorporated a transpyloric sleeve sensor (Psl), 4 cm in length, with three sideholes spaced along the sleeve (P1–P3, spaced at 1-cm intervals), four antral sideholes (A1–A4, spaced at 1-cm intervals) and 12 duodenal sideholes (D1–D12, spaced at 1.5-cm intervals). D1 was located 2 cm from the mid-pylorus; consequently, D12 was located 18.5 cm distally to the pylorus. An infusion port was located at 4.25 cm from the mid-pyloric region, between D2 and D3.
During the study, the position of the catheter was monitored by measurement of the transmucosal potential difference via two sideholes spaced at an interval of 4 cm (A4, the most distal antral sidehole, and D1, the most proximal duodenal sidehole) using established criteria.9 A disposable Ag/AgCl electrode, attached to the forearm, was used as the reference electrode. Pressures from the perfused sideholes were recorded via external transducers (Abbott, Sligo, Ireland). Pressure and transmucosal potential difference data were stored in digital format in two 12-channel dataloggers (Medical Measurement Systems, Enschede, The Netherlands) with a memory capacity of 4 Mb each, using a sample frequency of 4 Hz for pressure and 1 Hz for transmucosal potential difference signals. At the end of the study, the data were transferred to the hard disk of a computer for subsequent analysis.
Analysis was performed on 5-min recording periods pre- and post-infusion. The infusion period of 1 min was included in the 5-min post-infusion period. Data from pre-infusion periods were averaged for each subject. Antral pressure waves and isolated pyloric pressure waves were analysed visually. Interdigestive phases and isolated pyloric pressure waves were recognized according to established manometric criteria.10 A locally developed computer program was used for the calculation of numbers and the spatiotemporal characteristics of duodenal pressure waves. The algorithms used have been described in detail previously.11, 12 In short, for each sidehole, the number of antegrade propagated pressure wave sequences, originating from this location, were calculated. An antegrade propagated pressure wave sequence was defined as a pressure wave propagating over at least two recording sites (i.e. a distance of at least 1.5 cm). Antegrade propagated pressure wave sequences were divided into six categories according to the distance of propagation: over two sites (1.5 cm), over three sites (3 cm), over four sites (4.5 cm), over five sites (6 cm), over six sites (7.5 cm) and a cumulative group over 7–12 sites (9–16.5 cm).
During the study, the sensations of nausea, epigastric pain and fullness were scored using a 100-mm visual analogue scale ranging from no sensations to sensations as bad as can be.1, 13 These sensations were scored before, 1 and 5 min after each infusion. To make subgroup analysis feasible, duodenal hypersensitivity was arbitrarily defined as a minimal increase of 10 mm on the visual analogue scale after infusion of acid or saline.
The motor and sensory responses to infusion were assessed by subtracting pre- from post-infusion variables. The effect of treatment was assessed by subtracting before-treatment variables from after-treatment variables.
Repeated measures analysis of variance (RM-ANOVA) or the paired Student’s t-test was performed to test for spatial and infusion-induced differences in manometric variables obtained from the total of 12 duodenal sideholes, or from smaller clusters of sideholes. The unpaired t-test or RM-ANOVA was also used for comparisons between the two treatment groups, by defining a group variable. Motility data are shown as mean values ± S.E.M.
Because the sensation scores were not normally distributed, these data were compared using non-parametric tests. The Wilcoxon signed rank test was used for paired intraindividual comparisons, and the Mann–Whitney U-test for comparisons between the two treatment groups. Sensation scores are presented as median values and interquartile ranges.
Statistical significance required P < 0.05.
One patient, who was treated with placebo, was excluded from further analysis because of a technical failure while performing the experiment. The remaining 19 patients (10 pantoprazole, nine placebo) completed the study. Compliance among these patients was excellent (100%), as assessed by tablet counts. The experiments were well tolerated by all subjects. The catheter position measured using the transmucosal potential difference was correct for 97% of the entire study time.
Before treatment (n=19). During the pre-infusion periods, most duodenal antegrade propagated pressure wave sequences propagated over short distances (47% over 1.5 cm, 20% over 3 cm, 10% over 4.5 cm, 9% over 6 cm, 5% over 7.5 cm, 10% over 9–16.5 cm).
Saline infusion did not significantly affect the numbers of duodenal pressure waves, antegrade propagated pressure wave sequences and isolated pyloric pressure waves before treatment with pantoprazole or placebo (Table 1).
Acid infusion induced a small increase in the number of duodenal pressure waves (P=0.10) (Table 1). This acid-induced increase in pressure waves mainly occurred around the infusion site, i.e. in sideholes D1–D3 (pantoprazole group, P=0.04; placebo group, P=0.03). Also, the number of duodenal antegrade propagated pressure wave sequences increased after acid infusion (P < 0.05) (Table 1). No significant increases in antegrade propagated pressure wave sequences were observed when subdivided according to propagation distance. In the antrum, acid strongly suppressed the number of antegrade propagated pressure wave sequences (P < 0.001). The number of isolated pyloric pressure waves after acid infusion did not differ compared to that pre-infusion.
Neither pantoprazole treatment nor placebo resulted in a significant change in the numbers of duodenal pressure waves or antegrade propagated pressure wave sequences, both during pre-infusion periods and after acid or saline infusion (Table 2). In addition, no differences in the distribution of duodenal antegrade propagated pressure wave sequences according to propagation distance were observed after treatment. The number of duodenal antegrade propagated pressure wave sequences after acid infusion increased, similar to before treatment (Table 2). Antral antegrade propagated pressure wave sequences were suppressed by acid infusion in both treatment groups, similar to before treatment (Table 2, Figure 1). However, in contrast to the manometric recordings before treatment, a significant increase in isolated pyloric pressure waves induced by acid infusion was observed after pantoprazole treatment (before treatment, 1.0 ± 0.3 vs. 1.1 ± 0.4 (pre-infusion vs. acid); after treatment, 1.4 ± 0.4 vs. 5.1 ± 1.3 (pre-infusion vs. acid, P < 0.02); before vs. after treatment with pantoprazole, P=0.02) (Table 2, Figure 1). Comparison of the two treatment groups showed no differences between the antroduodenal motor responses.
The sensation scores at the start of the experiments were no different before and after treatment or among the two groups.
Acid infusion increased the sensation of nausea 1 min after infusion (P < 0.05), both before and after treatment in both groups (after vs. before treatment and when comparing both groups, P=N.S.) (Figure 2A). However, when focusing on the acid-hypersensitive patients (pantoprazole, n=4; placebo, n=3), a decrease in nausea after pantoprazole treatment was observed in all four patients, although not statistically significant (pantoprazole, from 27.3 ± 3.2 to 12.8 ± 6.7, P=0.07; placebo, from 16.3 ± 4.1 to 16.7 ± 5.2, P=1.0; comparing both groups, P=0.11). Acid infusion did not affect the sensations of fullness and epigastric pain.
Saline infusion did not induce any sensations, nor was there a change after treatment, or a difference in response among the two groups (Figure 2B).
Gastric acid suppression through proton pump inhibition has been found to be efficacious in a subset of patients with functional dyspepsia,4 and acid secretion inhibitory drugs are widely prescribed in these patients. To our knowledge, this study is the first to investigate the effects of treatment with an acid-suppressing drug on pathophysiological factors that are assumed to underlie this functional ailment. Many of these factors have only been recently explored and recognized and are heterogeneously distributed.1, 2, 14–16 Whereas no hypersensitivity to gastric acid could be demonstrated in patients with functional dyspepsia,17 recent studies have shown that duodenal hypersensitivity to acid exists in a subset of these patients.1, 2 In this study, we chose to investigate the effect of treatment with a proton pump inhibitor on acid-induced sensations and motor responses, not on dyspeptic symptoms in daily life. This approach was chosen to further assess the role of duodenal acid hypersensitivity in patients with functional dyspepsia.
This study in patients with functional dyspepsia shows that: (i) the previously described decreased duodenal motor response to acid infusion2 is not affected by pantoprazole treatment; (ii) intraduodenal acid induces an increase in isolated pyloric pressure waves only after pantoprazole treatment; and (iii) hypersensitivity to intraduodenal acid persists after pantoprazole treatment, but tends to decrease in the acid-hypersensitive patients.
The antropyloroduodenal motor response to intraduodenal acid infusion can be divided into two parts. First, there is the local effect of acid in the duodenum, with the largest increase in pressure waves found around and distal to the infusion site.12 A previously performed study showed that this duodenal motor response is decreased in patients with functional dyspepsia, which was confirmed in the present study.1, 2 The second part involves an enterogastric feedback mechanism causing a marked suppression of propagated waves from the antrum and stimulation of isolated pyloric pressure waves, and is unaffected in patients with functional dyspepsia.2 This reflex mechanism serves to regulate gastric emptying and has also been described after intraduodenal infusion of lipids, glucose, amino acids and hyperosmolar solutions.12, 18 Acid has been shown to induce isolated pyloric pressure waves when infused intraduodenally in supraphysiological quantities,19 but, in this study, 5 mL of acid did not induce isolated pyloric pressure waves before treatment, as also shown previously in healthy subjects.12 However, after treatment with pantoprazole, isolated pyloric pressure waves were induced by intraduodenal acid. The cause of this exaggerated pyloric motor response is unknown, but it confirms our previous findings that the duodenogastric feedback mechanism in patients with functional dyspepsia is intact.2 As the impaired duodenal motor response is fast and spatiotemporally related to the infusion site, the most likely location for the responsible afferent nerve dysfunction would be the local enteric nerve system.2 Apparently, the mucosal receptors involved are unaffected by the period of acid depletion (e.g. up-regulated), or acid exerts its effect directly on free nerve endings in the duodenal mucosa.
Although proton pump inhibitor treatment did not affect the impaired duodenal motor response to acid, a trend towards a reduction in duodenal hypersensitivity to acid was observed. Statistical significance was not reached, probably due to a type 2 error (too few subjects). Hypersensitivity may be the result of peripheral sensitization of primary afferents after local tissue injury by noxious stimuli (e.g. acid), leading to the recruitment of ‘silent’ nociceptors and ultimately central sensitization.20 As hypothesized previously, a decreased motor response and subsequent decreased clearance of acid from the proximal duodenum,1 leading to longer mucosal acid exposure, may lead to this process of peripheral sensitization. Treatment with pantoprazole might therefore lead to desensitization of the mucosa. The remaining hypersensitivity after treatment could be explained by the existence of a ‘memory’ of the initiating peripheral stimulus,20 which might subside further after a longer treatment course. However, this hypothesis is contradicted by the observation that patients with acid hypersensitivity (n=7) are not always the same as those with an impaired motor response to acid (i.e. of the 10 patients with the poorest motor response to acid, only three were acid-hypersensitive). Thus, it is unlikely that duodenal acid hypersensitivity is primarily caused by an impaired duodenal motor response to acid infusion. We postulate that a defect in neural control of the gut is responsible for both the impaired motor response and the heightened acid sensitivity in the duodenum. This is analogous to the previous speculations by Malagelada,21 when discussing the concomitant existence of impaired reflex activity and mechanohypersensitivity.22 According to this theory, patients with functional dyspepsia may express either visceral hypersensitivity, gut dysmotility, or both, depending on the predominance of one or the other. In this way, some patients develop acid hypersensitivity independent of the duration of acid exposure, due to individual susceptibility to sensitization, e.g. via a relative loss of descending bulbospinal inhibitory modulation.20 An alternative explanation would be that different motor and/or sensory impairments coexist in the heterogeneous group of patients with functional dyspepsia.
In summary, the combination of an unaffected impaired duodenal motor response and an intact duodenogastric response to acid, during a period of gastric acid inhibition, is indicative for the presence of a localized (enteric nervous system?) defect in the duodenum of patients with functional dyspepsia. Acid inhibition may cause duodenal hypersensitivity to decrease, probably through a process of desensitization. However, hypersensitivity does not seem to be related to the decreased duodenal motor activity, which is suggestive for individual susceptibility to sensitization, rather than a role for decreased duodenal acid clearance. Future studies, preferably in a selection of acid-hypersensitive patients, are required to assess whether the observed effects are of clinical importance, and thus warrant maintenance therapy with a proton pump inhibitor in patients with functional dyspepsia.
This study was supported by Byk Nederland b.v., Zwanenburg, The Netherlands. Dr M. Samsom is a fellow of the Royal Netherlands Society of Arts and Sciences. The authors thank Dr J. A. J. Faber of the Center for Biostatistics for his help in performing the statistical analyses.