These authors contributed equally to this study.
Intersubject and intrasubject variability of gastric volumes in response to isocaloric liquid meals in functional dyspepsia and health2
Article first published online: 21 JUN 2007
Neurogastroenterology & Motility
Volume 19, Issue 7, pages 553–561, July 2007
How to Cite
Fruehauf, H., Goetze, O., Steingoetter, A., Kwiatek, M., Boesiger, P., Thumshirn, M., Schwizer, W. and Fried, M. (2007), Intersubject and intrasubject variability of gastric volumes in response to isocaloric liquid meals in functional dyspepsia and health. Neurogastroenterology & Motility, 19: 553–561. doi: 10.1111/j.1365-2982.2007.00904.x
Parts of this work were presented at the 19th Symposium on Neurogastroenterology and Motility, Cambridge 2004 and at the DDW 2005, Chicago.
- Issue published online: 21 JUN 2007
- Article first published online: 21 JUN 2007
- Received: 12 September 2006 Accepted for publication: 25 December 2006
- functional dyspepsia;
- gastric emptying;
- gastric volume response;
- magnetic resonance imaging;
Abstract Gastric emptying (GE) has a considerable variability, but data on reproducibility of gastric volume measurements are sparse. We aimed to study the reproducibility of postprandial gastric volume responses and GE using magnetic resonance imaging (MRI) in healthy controls (HC) and patients with functional dyspepsia (FD). Eight HC and eight FD patients underwent a MRI study on two occasions. MR images were acquired in seated position before and up to 120 min after liquid meal administration (200 mL, 300 kcal). Fasting (V0), initial postprandial stomach volumes (V1), volume changes (V1 − V0) and meal emptying half-times () were determined. Intersubject and intrasubject coefficients of variation (CVinter, CVintra) and Pearson's correlation coefficients (r) were calculated. on both occasions were (mean ± SD) 113 ± 28 and 121 ± 30 min in HC (ns) and 127 ± 31 and 128 ± 37 min in FD (ns), respectively. In HC, CVinter, CVintra, r were 31%, 23%, 0.49 for V0; 13%, 7%, 0.68 for V1; 10%, 4%, 0.71 for V1 − V0 and 25%, 7%, 0.90 for . In FD these parameters were for V0: 42%, 41%, −0.06; for V1: 18%, 10%, 0.40; for V1 − V0: 20%, 14%, 0.74 and for : 26%, 10%, 0.84. The stomach accommodates to a given meal volume, resulting in similar and reproducible postprandial volumes within- and between-subjects. MRI provides reproducible measurements of gastric volume responses in health and disease.
Ingestion of a meal results in a vagally mediated increase in gastric volume, compliance and a reduction in gastric tone. This adaptation facilitates the ingestion of large volumes of solid or liquid meals without inducing symptoms or the vomiting reflex. Failure of gastric responses to food ingestion or reduced postprandial relaxation of the stomach (impaired ‘accommodation’) has been demonstrated in a significant proportion (40%) of patients with functional dyspepsia (FD), particularly in those with symptoms of bloating, distension, nausea, early satiety and weight loss.1–7
A variety of techniques have been utilized to evaluate the gastric adaptation process.8–10 Magnetic resonance imaging (MRI) allows the non-invasive evaluation of the entire gastric motor function, i.e. the simultaneous assessment of both, changes in total gastric volume and meal emptying.11,12 The technique is operator-independent, not associated with ionizing radiation, and can differentiate between an ingested meal (solid and liquid phases), gastric secretion and intragastric gas.13,14 It has been validated and contrasted against barostat15,16 and has been proposed for measurements of gastric volume responses and gastric emptying (GE).11,12,16–19
Current volumetric imaging techniques such as MRI can reflect in a non-invasive approach the physiology of the postprandial gastric adaptation process and have the potential to be integrated in the clinical management of patients with functional gastrointestinal (GI) disorders.7,12,13,20–23 Further use of these imaging modalities requires the evaluation of the day-to-day variability of gastric volume responses to meal ingestion in health and in dyspeptic patients and under therapeutic intervention.9 However, data on the reproducibility of gastric volume measurements, i.e. total stomach volume, intragastric meal volume, secretion or gas, are sparse9 and in particular, there are no data on the contribution of intragastric air volume changes to total stomach volume changes in health and FD patients.
Therefore, the aims of the current study were (i) to compare gastric volume responses to ingestion of a standard 200 mL nutrient liquid meal on two occasions in healthy controls (HC) and patients with FD and (ii) to document the interindividual and intraindividual reproducibility of gastric volume measurements and GE using a validated MRI technique.
Material and methods
Eight healthy volunteers (five male, mean age 29 years, BMI 19.3–23.9 kg m−2) and eight patients with FD with early satiety as a predominant symptom3 (five male, mean age 34 years, BMI 18.0–24.7 kg m−2) participated in the study. All subjects were screened for GI symptoms prior to inclusion and had no evidence of organic GI disease by investigations including previous upper endoscopy, abdominal ultrasound, laboratory tests and a full physical examination. Diagnosis of FD was based on symptoms referable to the upper gut according to the Rome II criteria.24 Patient exclusion criteria were (i) predominant reflux symptoms postprandially, (ii) history of diabetes mellitus, (iii) prior abdominal surgery except appendectomy, cholecystectomy, hysterectomy or hernia repair and (iv) use of any drugs or supplements known to influence GI motor function or nutrient metabolism within 1 week prior to the start of the study.
None of the HCs had GI symptoms or a history of other diseases, drug allergies or previous abdominal surgery, except appendectomy, nor were they taking any regular medication apart from oral contraceptives or the occasional use of analgesics. Females with inadequate contraception, breast feeding or pregnancy were excluded from the study. Patients with FD and HC were matched for gender, age, weight, height and BMI (Table 1) and all subjects had a similar dietary history within 2 weeks prior to the study.25
|HC (n = 8)||FD (n = 8)|
|Age (years)||29 ± 5||34 ± 13|
|Weight (kg)||65 ± 6||67 ± 15|
|Height (cm)||174 ± 6||173 ± 15|
|BMI (kg m−2)||21.4 ± 1.6||22.3 ± 2.2|
Written informed consent was obtained from each participant prior to entry into the study. The study was carried out according to Good Clinical Practice and the Declaration of Helsinki. The study protocol was approved by the local Ethics Committee of the Department of Internal Medicine at the University Hospital Zurich, Switzerland.
The study was designed as a prospective controlled trial. All tests were performed after an 8-h overnight fast and in a relaxed sitting position. Physical activity was restricted during the test. Each subject underwent MRI of the gastric region prior to and after administration of a liquid meal on two different morning sessions, separated by 1 week.
The meal consisted of 200 mL Ensure®plus Vanilla (300 kcal, 53% carbohydrate, 32% fat, 15% protein, caloric density 1.5 kcal mL−1; Abbott AG, Baar, Switzerland) and was labelled with 0.5 mmol L−1 Gd-DOTA (Dotarem®, Guerbet, Roissy CdG Cedex, France) to enhance image contrast. Meals were administered per orally via a straw within a maximum time period of 5 min.
Magnetic resonance imaging
Data on gastric volume responses to meal ingestion and emptying were obtained using an open configuration MRI system (0.5 T, Signa SP/l, GE, Milwaukee, WI, USA) in an upright seated, i.e. the physiological position, as the intragastric distribution of a liquid meal affects gastric motor activity and GE. A volume scan, covering the complete gastric region was performed to assess fasting stomach volume. Postprandial volume scans were performed immediately after meal ingestion and then every 3 min until 15 min, every 10 min until 60 min, and finally every 15 min until 120 min. A 2D fast multiphase spoiled gradient echo technique and a standard send/receive abdominal surface coil (body flex) was used for excitation and image acquisition. Total scan time of a volume scan was 44 s, divided into two breath holds of 22 s to minimize motion artefacts. MRI sequence parameters were as follows: 20 sagittal image planes; slice thickness, 10 mm; receive bandwidth, 12,5 kHz; repetition time, 170 ms; echo time, 7.5 ms, flip angle, 60%; field of view, 350 mm; matrix, 256 × 160 pixel.
Magnetic resonance image processing was performed using a semiautomated segmentation algorithm based on in-house written software. The stomach contour was identified and manually outlined in each sagittal MRI of a volume scan (Fig. 1). Multiplying the sum of the segmented areas within a volume scan by the slice thickness resulted in the total stomach volume. The intragastric meal volume was identified by distinct positive contrast within the segmented stomach volume and computed using an intensity threshold method. The evaluation of GE half-time was carried out by nonlinear regression analysis of the normalized volume plots of intragastric meal volumes to a power exponential equation with Vnormalized = 2 as described by Elashoff et al.26 (: meal emptying half-time in min and β: mathematical parameter to be determined by nonlinear regression analysis).
The interindividual coefficient of variation (CVinter) for fasting stomach volume (V0), initial postprandial stomach volume (V1), change in stomach volume (V1 − V0) and in HC and FD patients was computed by dividing the standard deviation of the 16 measurements (eight subjects on two occasions) by their mean. The intraindividual coefficient of variation (CVintra) for V0, V1, V1 − V0 and was determined for each participant in both groups by dividing the individual standard deviation of the measurements on both days by the mean values of the two measurements. Coefficients of variation were expressed as percentage.
Bland–Altman plots27 were used to display the data on repeated studies and to characterize the CV with the repeated measurements. These graphs show the differences between the individual measurements on both days plotted against the averages of the two measurements, which represent the best estimates of the individual ‘true’ value. After testing all data for deviation from normal distribution by the Kolmogorov–Smirnov test intraindividual MRI measurements on both occasions were correlated with Pearson's correlation analysis; providing the Pearson's correlation coefficient (r) and the coefficient of determination (r2).
Paired t-test compared the within-subject variation of V0, V1, V1 − V0 and obtained on both days and unpaired t-test compared the data between groups. Demographic data, stomach volumes and meal emptying half-times are presented as mean ± SD. A P-value of <0.05 was considered statistically significant. All statistical calculations and graphical analysis were performed using standard software (SPSS® for Windows 10.0.7, SPSS Inc., Chicago, IL, USA; GraphPad Prism for Windows 4.02, GraphPad Software Inc., San Diego, CA, USA).
Fasting and initial postprandial stomach volumes and meal emptying half-times
There was no significant difference in mean stomach volumes (V0, V1) and (P > 0.05) between HC and patients with FD or within groups on both days (Table 2). V1 − V0 was different on the two study days in patients with FD only (Table 2).
|Day 1||Day 2||Day 1||Day 2|
|V0 (mL)||153 ± 41||129 ± 46||134 ± 49||112 ± 54|
|V1 (mL)||355 ± 42||340 ± 49||322 ± 62||333 ± 57|
|V1 − V0 (mL)||203 ± 21||211 ± 20||188 ± 35*||221 ± 41|
|(min)||113 ± 28||121 ± 30||127 ± 31||128 ± 37|
Reproducibility of stomach volumes and meal emptying half-times
For V0, there was no correlation between measurements on the two occasions in both groups (HC: r = 0.49, P = 0.219; FD: r = −0.06, P = 0.887; Fig. 2A). For V1, there was a correlation between measurements on both days in HC (r = 0.68), but this could be due to random sampling (P = 0.065; Fig. 2B). No correlation was found in the group of patients with FD (r = 0.40, P = 0.329). In contrast, there was a significant correlation in both groups between the measurements on both occasions for change in stomach volume (V1 − V0; HC: r = 0.71, P < 0.05; FD: r = 0.74, P < 0.05; Fig. 2C) and (HC: r = 0.90, P < 0.01; FD: r = 0.84, P < 0.01; Fig. 3A).
Fasting stomach volumes showed a considerable intersubject and day-to-day variability within subjects (Table 3). After meal ingestion, however, there was a high intraindividual reproducibility for V1, V1 − V0 and . In the patient group, the intraindividual CV (mean and range) for all parameters were higher compared with controls, despite overall similar mean values. CVinter were higher than CVintra and higher in FD patients compared with HC (Table 3).
|CVinter||CVintra (range)||r2||CVinter||CVintra (range)||r2|
|V0||31||23 (8–50)||0.24||42||41 (4–102)||0.01|
|V1||13||7 (0–14)||0.46||18||10 (1–36)||0.16|
|V1 − V0||10||4 (0–12)||0.50||20||14 (4–20)||0.55|
|25||7 (0–17)||0.80||26||10 (1–27)||0.70|
In HC and FD all individual data of all parameters (V0, V1, V1 − V0 and ) were within the limits of agreement (95% CI of mean difference in Bland–Altman plots, Figs 3B and 4). In both groups, the limits of agreement decreased postprandially when compared with fasting conditions (HC fasting vs HC postprandial: 88 vs 74 mL, for FD: 151 vs 130 mL). The limits of agreement for change in stomach volume (V1 − V0) were even smaller (HC 32 mL, FD 55 mL; Fig. 4C). Moreover, the limits of agreement were always smaller in HC when compared to patients with FD for all parameters (V0: 88 vs 151 mL; V1: 74 vs 130 mL; V1 − V0: 32 vs 55 mL; : 27 vs 40 min), despite similar overall mean values between groups.
Postprandial change in intragastric air volume
There was a positive correlation between postprandial change in intragastric air volume and change in total stomach volume in HC (r = 0.900, P < 0.001) and FD (r = 0.825, P < 0.001) on both days (Fig. 5A). A lower increase in total stomach volume after meal ingestion was associated with a less pronounced increase or even a reduction in intragastric air volume (Fig. 5B).
This study evaluates for the first time the intersubject and day-to-day variability of postprandial gastric volume responses to isocaloric liquid meals and meal emptying half-times in healthy volunteers and patients with FD, using MRI.
We have demonstrated that in fasting condition stomach volumes show a considerable intersubject and day-to-day variability within subjects (up to 42% in this study), which are slightly higher in FD patients. Despite a standardized fasting period of 8 h prior to the test, this variability may be attributed to image acquisition during different phases of the myoelectrical activity of the fasting stomach. However, immediately after meal ingestion, the CV for intersubject and intrasubject variability in HC and patients with FD were markedly lower, providing an excellent intraindividual reproducibility of stomach responses to meal administration. Mean stomach volumes (V0, V1) and on both study days did not differ between both groups or within groups on both occasions. These findings indicate that the stomach accommodates to a given meal volume, resulting in similar postprandial stomach volumes, irrespectively of day-to-day varying fasting stomach volumes. This is reflected by the excellent day-to-day reproducibility of change in stomach volumes especially within healthy subjects (CVintra 4%). In contrast, in patients with FD the initial change in stomach volume V1 − V0 was higher on day 2 when compared with day 1. This finding is consistent with a higher variability for the postprandial stomach volume response within patients (CVintra 14%) and may represent the functional correlate of impaired accommodation reported in FD.3
Interestingly, we could demonstrate in both groups on both days a strong and highly significant correlation between postprandial change in total stomach volume and change in intragastric air volume following meal ingestion. In some subjects, the intragastric storage capacity for the meal was partly provided by replacement of some fasting intragastric air, resulting in a lower expansion of total stomach volume than the administered meal volume would have suggested. This observation suggests that the gastric volume response to meal ingestion is highly regulated on an individual basis, either by ‘active’ changes of gastric tone and hence intragastric pressure, or ‘passively’ by venting during transient lower oesophageal sphincter relaxations, through the belch reflex or changes in the shape of the stomach, e.g. flattening of the His angle, enabling fundic air to escape during meal ingestion. The reduction in air volume might be facilitated by the upright body position during all our experiments.
The observed gastric volume response results in similar postprandial stomach volumes on different occasions within subjects, which provide an individually defined and highly reproducible starting point for the initiation of the GE process. This is the basis for an excellent individual day-to-day reproducibility of meal emptying half-times in HC and patients with FD (CVintra 7% and 10%, respectively). A mechanistic explanation for this finding might be that during the accommodation process, the day-to-day varying fasting stomach volumes are balanced by expanding or reducing intragastric air volumes according to a preset gastric tone, e.g. by belching or emptying into the duodenum. However, only further studies, combining high-resolution imaging techniques such as MRI with concomitant sensitive intragastric and intrasphincteric pressure monitoring can elucidate these regulative processes by simultaneous assessment of total stomach volume, gastric content, intragastric air and intragastric (air) pressure.
Hitherto, the simultaneous analysis of total stomach volumes and intragastric meal and air volumes has not been studied with other techniques and can only be achieved using validated non-invasive gastric MRI systems.9,12,28,29 To our knowledge, there has been no variability studies conducted on gastric volumes using MRI in HCs and in patients with FD. In particular, there are no data on the relative contribution of changes in intragastric air volume to changes in total stomach volume during postprandial gastric relaxation. However, data on the MRI reproducibility of gastric volume measurements are important, as there is growing recognition of the need to measure gastric volume responses to meal ingestion non-invasively to obtain an improved understanding of the physiology of the gastric accommodation process and of the evaluation of the associated mechanisms behind symptom generation in health and in disease.9,30
The only studies evaluating the reproducibility of total gastric volume measurements so far used SPECT imaging and were performed in healthy volunteers.31,32 The reported interindividual and intraindividual CV for the postprandial change in stomach volume were up to 14%, which is higher than reported in our study. These differences may be due to technical limitations of the SPECT technique and highlight the methodological advantages of MRI for the analysis of gastric volume responses and emptying. Compared with SPECT, MRI is able to detect fasting and postprandial intragastric fluid and air volume and thus allows the precise and direct assessment of the actual postprandial meal (fluid) volume increase and its relationship to the observed stomach relaxation. Furthermore, the short (22 s) and frequent (initially every 3 min) high-resolution (1.4 × 1.6 × 10 mm3) volume acquisition provides more accurate and immediate information on the initial stomach and meal volumes and the emptying process.17 In addition, the interindividual and intraindividual variability of meal emptying half-times in our MRI study is less than that seen in most GE measurements by scintigraphy,33–40 ultrasound41 or 13C acetate breath test.30 These report a day-to-day variability of 6–41% (in patients up to 73%) and an interindividual CV of 14–61% (in patients up to 84%), depending on meal composition and applied measurement technique.30,31,33–40,42–46 This may also be the consequence of the higher accuracy and frequency of single volume measurements by MRI. These lower variances favour the use of MRI in assessing postprandial gastric volume responses to meal ingestion in health and disease and in treatment studies and allow to detect small volume differences in small sample sizes.
In conclusion, we have shown that the gastric volume response to meal ingestion is a highly regulated process leading to an excellent reproducibility of postprandial stomach volumes and meal emptying half-times. This finding endorses the use of MR imaging as a valid tool in clinical practice and for research purposes in healthy subjects and different patient groups, when repeated measurements of both, gastric volume responses to meal ingestion and GE half-times are needed, or drugs that modulate gastric tone, volume or emptying are tested on different occasions.
This study was supported by grant No. 3100-067071 from the Swiss National Science Foundation. OG is supported by Deutsche Forschungsgemeinschaft (Gö 1358/1-1). We thank Ms Bernadette Stutz and Mr Karl Treiber for their excellent technical assistance.
- 4The proximal stomach and postprandial symptoms in functional dyspeptics. Am J Gastroenterol 2002; 97: 40–8., , , , .Direct Link:
- 7Noninvasive measurement of gastric accommodation in patients with idiopathic nonulcer dyspepsia. Am J Gastroenterol 2001; 96: 3099–105., , et al.Direct Link:
- 8Novel testing of human gastric motor and sensory functions: rationale, methods, and potential applications in clinical practice. Am J Gastroenterol 2000; 95: 3365–73., , .Direct Link:
- 18How reliable is 13C-acetate breath testing for measuring accelerated gastric emptying of different isocaloric test meals in functional dyspepsia and health? A comparison with MRI. Gastroenterology 2005; 128 (Suppl. 2): A545., , et al.
- 19Inter- and intrasubject variability of total gastric volumes in response to isocaloric liquid meals in functional dyspepsia and health. Gastroenterology 2005; 128 (Suppl. 2): A545., , et al.
- 33Reproducibility and simplification of 13C-octanoic acid breath test for gastric emptying of solids. Am J Gastroenterol 1998; 93: 92–8., , , , , .Direct Link: