- Email: [email protected]
Simultaneous measurement of gastric emptying of a solid meal by ultrasound and by scintigraphy
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 1999 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 94, No. 10, 1999 ISSN 0002-9270/99/$20.00 PII S0002-9270(99)00485-2
Simultaneous Measurement of Gastric Emptying of a Solid Meal by Ultrasound and by Scintigraphy L. Benini, M.D., C. Sembenini, M.D., R. C. Heading, M.D., P. G. Giorgetti, M.D., S. Montemezzi, M.D., M. Zamboni, M.D., P. Di Benedetto, M.D., F. Brighenti, Ph.D., and I. Vantini, M.D. Department of Gastroenterology, Rehabilitation Hospital of Valeggio sM, University of Verona, Verona, Italy; Department of Medicine, Royal Infirmary, Edinburgh, Scotland; Service of Nuclear Medicine and Service of Radiology, General Hospital, Verona, Italy; and Department of Alimentary and Microbiological Sciences and Technologies, University of Milan, Milan, Italy
OBJECTIVE: Although ultrasonic imaging may represent a valid alternative to scintigraphy for measurement of gastric emptying, most studies comparing the two methods have been carried out with liquid meals. The aim of this study was to compare scintigraphic and ultrasonographic measurements of gastric emptying of a solid meal in healthy subjects and in patients with possible delay in emptying. METHODS: Nineteen subjects were studied: five controls, six patients with gastroesophageal reflux, and eight patients with dysmotility-like dyspepsia. Gastric emptying was measured by both scintigraphy and ultrasonography after ingestion of an 800-calorie solid, realistic meal containing 99mTclabeled chicken liver. Scintigraphic measurements were made every 15 min for 6 h, and ultrasonic imaging of antral sections was undertaken every 15 min for the first 1 h and every 30 min thereafter. Total emptying times were calculated independently using the two methods, and the emptying patterns recorded by the two methods were compared. RESULTS: Maximal antral dilation occurred 30 min (range 0 –90 min) after the end of the meal and persisted until 96 ⫾ 42 min, by which time gastric radioactivity had decreased from its maximum by 43% ⫾ 23%. From this time on, the antral cross-sectional area returned toward the basal value, declining faster than the gastric counts recorded by scintigraphy. Total emptying times measured by ultrasound and by scintigraphy were in good agreement in all subjects, with a mean difference of only 4.5 min (limits of agreement, ⫺17.1 to 21.6 min). CONCLUSIONS: Ultrasonographic measurement of antral cross-sectional area provides a valid alternative to scintigraphy for the measurement of total gastric emptying of a solid meal. It is less reliable if other parameters of gastric emptying such as T1/2 are required. (Am J Gastroenterol 1999;94:2861–2865. © 1999 by Am. Coll. of Gastroenterology)
INTRODUCTION Ultrasound has been suggested as a valid alternative to scintigraphy for measurement of gastric emptying (1–5). The ultrasonographic method cannot discriminate between emptying rates of the liquid and solid phases of a meal, but offers some practical advantages over scintigraphy. For example, the necessary equipment is widely available and its use incurs only modest running costs. In most hospitals, access to ultrasonography disrupts routine clinical work less than the need for a gamma camera. Additionally, the lack of radiation hazard makes ultrasound especially suitable for repeated measurements, including research studies in healthy volunteers (1, 2). Preliminary suggestions of an acceptable concordance between the results of scintigraphic and ultrasonographic methods (2, 3) were followed by specific studies of liquid meals only (4 – 6); at present, the method is considered as a screening test to select patients who should undergo scintigraphic measurements (7–9). In particular, because the antrum exhibits a greater postprandial enlargement in dyspeptic patients than in controls (8, 10), a demonstration of good agreement between ultrasonographic and scintigraphic emptying measurements in normals cannot be assumed to apply in patients. The aim of our study was to compare scintigraphic measurements of gastric emptying of a solid meal with ultrasonographic measurements of antral emptying both in healthy subjects and in patients in whom delayed gastric emptying might be present. To optimize the comparison, the two methods of emptying measurement were applied simultaneously.
MATERIALS AND METHODS Ultrasonographic measurement of antral diameters and scintigraphic measurement of gastric emptying of 99mTc-labeled chicken liver were carried out simultaneously in 20 subjects (14 male, six female) aged 39 – 63 yr. Five were healthy members of our medical staff (controls, C), six were patients
2862
Benini et al.
with pathological gastroesophageal reflux (GER), and nine were patients with dysmotility-like dyspepsia (DYS). One dyspeptic patient had to be excluded from the study for lack of a hepatic acoustic window, which made ultrasound imaging unsatisfactory. Only a glass of tea with 1 teaspoonful of sugar was allowed for breakfast on the day of the study. At 1 PM, subjects were given the standard, two course meal that we previously described for ultrasound studies on gastric emptying (60 g of “macaroni alla bolognese” with 70 g of meat sauce and then 50 g of ham, 50 g of soft fatty cheese, one roll, and 250 ml of water; 800 Kcal, 15% from protein, 45% from fat, and 40% from carbohydrates) (11, 12). The only difference from previous studies was the substitution of 20 g of ham by 50 g of chicken liver. The raw liver was repeatedly injected in different sites with a total of 37 MBq of 99mTc colloid, cut in cubes of approximately 1 cm3 and then fried to a firm consistency. Both the liver and the oil used to fry it were then added to the meat sauce of the pasta. Walking in quiet surroundings was permitted between measurements. Ultrasonographic and scintigraphic measurements were always performed within a couple of minutes of each other, with patients sitting in an armchair tilted at an angle of 45 degrees backward. Scintigraphy After ingestion of the meal, radioactivity was measured from the anterior aspect of the abdomen for 1 min every 15 min up to h 6, or until radioactivity had reverted to the basal value. Measurements were corrected for decay according to the formula A ⫽ a ⫹ (a ⫻ 0.00192 ⫻ t), where A represents the value corrected for decay, a represents the measured value, and t represents the time (min) elapsed after the first postprandial measurement. The gastric region of interest was drawn visually on the computer display of the first frame and then replicated on later images. First frame activity in this region was taken as 100%. The residual radioactivity at each subsequent time was then plotted against time using an electronic spreadsheet (Excel 7.0 for Windows 95, Microsoft, Redmond, WA), and the tendency line for their relationship automatically calculated (moving mean of two points). The scintigraphic half time (T1/2) and total emptying time were identified from the points where this tendency line crossed the 50% and 3% values, respectively. The 3% value was chosen to allow for any nonspecific binding of tracer to the gastric wall. Ultrasound Real time ultrasonography (Toshiba Diagnostic Equipment model SSA-220A, fitted with a 3.75-MHz angular probe) was used to measure the diameters of the gastric antrum in the sagittal plane passing through the aorta. Measurements were made before the meal (basal), immediately after ingestion (time 0), then at 15-min intervals for the first 1 h and at 30-min intervals thereafter. The mean of three readings was calculated at each time during interperistaltic relaxation. The antral section was calculated using the formula
AJG – Vol. 94, No. 10, 1999
S ⫽ d1 ⫻ d2 ⫻ /4, where S represents the antral crosssectional area and d1 and d2 the measured diameters. The antral cross-sectional area was then plotted against the time. Visual inspection of the completed plot readily allowed identification of the initial period, when values of antral cross-sectional area were increasing or stable, and its distinction from the rapid antral emptying phase, when areas progressively declined. All the measurements made in this latter phase were used to calculate the regression line for antral emptying. The ultrasonographic half emptying time (T1/2) and total emptying time were identified from the points where this tendency line crossed the 50% and basal values, respectively. The percent residual distension of the antrum was calculated by relating the increase above basal value of the cross-sectional area at each time to the maximal postprandial increase above basal value observed in that subject. Statistical Methods Results are given as means ⫾ 1 SD. The statistical significance of the difference was calculated using Student’s t test for nonpaired data and analysis of variance. Pearson’s r coefficient between the results of the two methods is shown for descriptive purposes. The agreement between the two methods is better expressed by calculating the difference between their results and then plotting the difference against the mean values obtained by scintigraphy and by ultrasound. This obviates any possible effect of the measured values on the magnitude of the difference and gives an indication of the magnitude of potential error incurred by exchanging the new method for the old “reference” one. The mathematical and statistical bases for this evaluation are described by Bland and Altman (13). The limits of agreement were defined as the mean ⫾ 2 SD of the difference between the two methods.
RESULTS No difference was found in the basal and maximal postprandial antral sections as measured by ultrasound in the three groups of subjects (C 319 ⫾ 92 and 1667 ⫾ 425 mm2; GER 462 ⫾ 185 and 1693 ⫾ 444 mm2; DYS 403 ⫾ 117 and 1619 ⫾ 266 mm2, respectively). Figure 1 shows the change with time of the scintigraphic counts in the gastric area and of ultrasonographic antral cross-sectional areas, each expressed as percentage of the maximal value found in each subject. The scintigraphic counts attained their maximum immediately after the meal and, in most cases, began to fall thereafter. A scintigraphic “lag-phase” was evident only in three patients. In contrast, maximal antral dilation occurred at a mean of 30 min after the end of the meal (range 0 –90 min) and persisted until 96 ⫾ 42 min (echographic “lag phase”). At this time, the activity in the gastric region of interest had decreased by 43% ⫾ 24%. From this time onward, the antral cross-sectional area decreased progressively, regaining the basal value at the same time as scin-
AJG – October, 1999
Figure 1. Scintigraphic counts in the gastric area and ultrasonographic antral cross-sectional areas, each expressed as a percentage of the maximal value found in each patient, plotted against time after meal ingestion. Vertical bars represent 1 SD.
tigraphy. In the figure, mean antral cross-sectional areas do not reach 100% at any time because the maximum (100%) value was reached in each subject at different times. Figure 2 shows the radioactivity within the gastric region of interest when distension of the antrum was 50%, 30%, 20%, and 10% of the maximum. It is evident that scintigraphic retention varied widely in the different subjects, for any given ultrasonographic value. Additionally, it is clear that the proportion of the radioactive tracer remaining
Figure 2. Radioactivity in the gastric area of interest when the residual distension of the antrum is reduced to 50%, 30%, 20%, or 10% of the maximum. The columns represent mean values. Vertical bars represent 1 SD.
Measurement of Gastric Emptying by US and Scintigraphy
2863
Figure 3. Correlation between total emptying times measured by scintigraphy and by ultrasonographic measurement of the antral diameters in controls (triangles), in patients with dysmotility-like dyspepsia (squares) and with gastroesophageal reflux (open circles).
within the stomach was consistently less than suggested from the ultrasonographic measurement of antral diameters. The discrepancy was largest in the early phase of the study and progressively decreased. Figure 3 shows the relationship between the results of the two methods when total emptying times (i.e., times necessary for return of the stomach to the basal condition) were considered. Only in one patient did radioactive counts and antral diameters fail to return to basal values by h 6 and total emptying time had to be estimated by extrapolation. A good linear correlation between the two methods was found (y ⫽ 22 ⫹ 0.91x; R2 ⫽ 0.92), with no obvious differences among the three groups of subjects. Concordance between the two measurements was much poorer when half emptying times were considered. The time necessary for a 50% reduction of the postprandial antral dilation, as measured by ultrasound (164 ⫾ 34 min), was much longer than the T1/2 of scintigraphy (112 ⫾ 35 min). Interestingly, there was also a poor correlation between the scintigraphic T1/2 and the total emptying time, whether measured by ultrasound (y ⫽ 197 ⫹ 0.77x; R2 ⫽ 0.58) or by scintigraphy (y ⫽ 201 ⫹ 0.77x min; R2 ⫽ 0.54) (Fig. 4A and 4B). The difference in total emptying time between controls and patients (grouping together GER and DYS) is significant both for scintigraphy (C 265 ⫾ 19 min; GER ⫹ DYS 301 ⫾ 41 min, p ⬍ 0.02) and for ultrasound (259 ⫾ 23 vs 296 ⫾ 37 min respectively, p ⬍ 0.03). The difference plot (i.e., the difference between the results of ultrasonography and scintigraphy plotted against the mean result of the two methods) showed a mean difference of 4.5 min (limits of agreement, ⫺17.1 to 21.6 min); the
2864
Benini et al.
Figure 4. Correlation between scintigraphic half emptying times and total emptying times as measured by ultrasound (A) and by scintigraphy (B). (Symbols as in Figure 3.)
difference was ⬍20 min in all but one subject. There was no relationship between the mean emptying time as measured by the two methods and the value of their difference. Finally, there was no evidence that the magnitude of the difference varied in the three groups of subjects (analysis of variance), although the relatively small number of subjects means that this can be no more than a tentative conclusion.
DISCUSSION Ultrasound has been suggested as a cheap and valid alternative to scintigraphy for the measurement of gastric emptying. The necessary equipment is widely available and the test does not cause radiation exposure, allowing repeated measurements to be carried out when the effects of drugs or therapeutic procedures are to be evaluated (14, 15). Although gastric dysmotility problems are frequent in clinical practice, the measurement of gastric emptying by ultrasound is restricted to a few centers. An explanation for this may be the methodological difficulty of the procedure, as the posterior wall of the stomach may be difficult to identify after the ingestion of solids. Restricting ultrasonographic studies to liquid meals seems reasonable (16) but, in fact, consistent results have been reported with solid meals when experienced operators are involved (11, 17). The complex relationship between antral cross-sectional area and intragastric volume has caused difficulty for investigators using ultrasound for gastric emptying measurements and also when attempts are made to reconcile the findings reported by different investigators. Although a linear relationship has been suggested between the volume of ingested liquid and antral size (17), other evidence indicates that emptying of liquids relates to the content of the proximal stomach and that, after liquid ingestion, the content of the distal stomach initially increases, then remains relatively constant for a period and decreases thereafter (5, 18). Factors favoring antral distension (such as gravity and fundic
AJG – Vol. 94, No. 10, 1999
tone) or contraction (such as vagally mediated fundic adaptive relaxation and phasic antral contractions) modify antral size for any given intragastric volume. The modification of antral size by the caloric value of test meals may explain the good agreement between ultrasound and scintigraphy for gastric emptying of water, but not of milk (4). The variables become even more important after solid meals because of the important role of the fundus in solid meal accommodation. For example, when two of our volunteers ate progressively increasing volumes (100 to 900 ml) of chocolate pudding on different days, antral cross-sectional areas measured 0 and 30 min after completion of the meal increased linearly with volume to a maximum of 500 ml, but no further increase was seen with the larger meals (unpublished data). The time required for the antrum to return to baseline (“total emptying time”), as measured by ultrasound, can be divided into two periods. In the first, the antral cross-sectional areas progressively increase toward a maximum, and then remain fairly constant (described here as the “ultrasonographic lag phase”); in the second, the cross-sectional areas gradually decrease to their basal value. Although this behavior resembles the scintigraphic sequence of solid emptying, it relates to different aspects of emptying. The “lag phase” on ultrasound clearly does not represent the period before emptying begins, as the simultaneous scintigraphic study has shown that a substantial proportion of the meal is leaving the stomach. When antral size begins to decrease, 40% of the meal has emptied. From this time on, the antral cross-sectional area falls fairly quickly, regaining the basal value at the same time as scintigraphic total gastric emptying. The difference between the two tests (⬍10% in most patients) suggests that this use of ultrasonography produces valid emptying measurements. The different time course of the decrease of radioactivity and of antral size after a solid meal is similar to results described in abstract form by other investigators, who either measured the half emptying time (19), or visually compared the residual curves (20). The difference between ultrasound and scintigraphy in the early phases of our study may even be an underestimate of the true difference, because the use of anterior-only acquired scintigraphic counts may overestimate the lag phase as a result of anterior movement of the meal (21). We believe that any error was small, because a detectable scintigraphic lag phase was observed in only three of our subjects. In conclusion, ultrasonographic measurement of antral diameters offers a valid alternative to scintigraphy for the measurement of gastric emptying of solid meals if used to determine total emptying time. It is unlikely to be satisfactory if other gastric emptying measurements such as T1/2 are desired. The accuracy of the method seems as good in patients with dysmotility-like dyspepsia or gastroesophageal reflux as in healthy normal subjects.
AJG – October, 1999
Measurement of Gastric Emptying by US and Scintigraphy
ACKNOWLEDGMENT This work was supported by grant 95.00773.PF41 from the Italian National Research Concil (CNR)-Targeted Project “Prevention and Control of Disease Factors”, Subproject “Alimentation”. Reprint requests and correspondence: Luigi Benini, M.D., Department of Gastroenterology, Rehabilitation Hospital, 37067 Valeggio sM, Verona, Italy. Received Aug. 26, 1997; accepted Apr. 30, 1999.
REFERENCES 1. Bateman DN, Whittingham TA. Measurement of gastric emptying by real time ultrasound. Gut 1982;23:524 –7. 2. Holt S, Cervantes J, Wilkinson AA, et al. Measurement of gastric emptying rate in humans by real-time ultrasound. Gastroenterology 1986;90:918 –23. 3. Bolondi L, Bortolotti M, Santi M, et al. Measurement of gastric emptying time by real-time ultrasonography. Gastroenterology 1985;89:752–9. 4. Marzio L, Giacobbe A, Conoscitore P, et al. Evaluation of the use of ultrasonography in the study of liquid gastric emptying. Am J Gastroenterol 1989;84:496 –500. 5. Hveem K, Jones KL, Chatterton BE, et al. Scintigraphic measurement of gastric emptying and ultrasonographic assessment of antral area: relation to appetite. Gut 1996;38:816 –21. 6. Mamtora H, Thompson DG. Gastric ultrasound. In: Read NW, ed. Gastrointestinal motility: Which test? London: Wrightson Biomedical Publishing, 1989:99 –104. 7. Heading RC, Bolondi L, Camilleri M, et al. Gastric emptying. Gastroenterol Int 1992;5:203–15. 8. Troncon LEA, Bennet RJM, Ahluwalia NK, et al. Abnormal intragastric distribution of food during gastric emptying in functional dyspepsia patients. Gut 1994;35:327–32. 9. Ahluwalia NK, Thompson DG. Ultrasonography. In: Kumar
10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
21.
2865
D, Wingate D, eds. An illustrated guide to gastrointestinal motility. Second edition. Edinburgh: Churchill Livingstone, 1993:242–55. Hausken T, Berstad A. Effect of nitric oxide on antral motility and symptoms in patients with functional dyspepsia. Scand J Gastroenterol 1994;29:23– 8. Benini L, Castellani G, Sembenini C, et al. Gastric emptying of solid meals in achalasic patients after successful pneumatic dilatation of the cardia. Dig Dis Sci 1994;39:733–7. Benini L, Castellani G, Bardelli E, et al. Omeprazole causes a delay in gastric emptying of digestible meals. Dig Dis Sci 1996;41:469 –74. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. Benini L, Brighenti F, Castellani G, et al. Gastric emptying of solids is markedly delayed when meals are fried. Dig Dis Sci 1994;39:2288 –94. Benini L, Castellani G, Brighenti F, et al. Gastric emptying of a solid meal is accelerated by the removal of dietary fibre naturally present in food. Gut 1995;36:825–30. Irvine JE, Tougas G, Lappalainen R, et al. Reliability and interobserver variability of ultrasonographic measurement of gastric emptying rate. Dig Dis Sci 1993;38:803–10. Ricci R, Bontempo I, Corazziari E, et al. Real time ultrasonography of the gastric antrum. Gut 1993;34:173– 6. Collins PJ, Houghton LA, Read NW, et al. Role of proximal and distal stomach in mixed solid and liquid meal emptying. Gut 1991;32:615–9. Bolondi L, Santi V, Bortolotti M, et al. Correlation between scintigraphic and ultrasonographic assessment of gastric emptying. Gastroenterology 1986;90:1349 (abstract). Tani H, Haruma K, Hata J, et al. Ultrasonographic assessment of gastric emptying rate of solids: A comparison of ultrasonography and scintigraphy. Gastroenterology 1996;110:768 (abstract). Tothill P, McLoughlin GP, Heading RC. Techniques and errors in scintigraphic measurements of gastric emptying. J Nucl Med 1978;19:256 – 61.
Vol. 94, No. 10, 1999 ISSN 0002-9270/99/$20.00 PII S0002-9270(99)00485-2
Simultaneous Measurement of Gastric Emptying of a Solid Meal by Ultrasound and by Scintigraphy L. Benini, M.D., C. Sembenini, M.D., R. C. Heading, M.D., P. G. Giorgetti, M.D., S. Montemezzi, M.D., M. Zamboni, M.D., P. Di Benedetto, M.D., F. Brighenti, Ph.D., and I. Vantini, M.D. Department of Gastroenterology, Rehabilitation Hospital of Valeggio sM, University of Verona, Verona, Italy; Department of Medicine, Royal Infirmary, Edinburgh, Scotland; Service of Nuclear Medicine and Service of Radiology, General Hospital, Verona, Italy; and Department of Alimentary and Microbiological Sciences and Technologies, University of Milan, Milan, Italy
OBJECTIVE: Although ultrasonic imaging may represent a valid alternative to scintigraphy for measurement of gastric emptying, most studies comparing the two methods have been carried out with liquid meals. The aim of this study was to compare scintigraphic and ultrasonographic measurements of gastric emptying of a solid meal in healthy subjects and in patients with possible delay in emptying. METHODS: Nineteen subjects were studied: five controls, six patients with gastroesophageal reflux, and eight patients with dysmotility-like dyspepsia. Gastric emptying was measured by both scintigraphy and ultrasonography after ingestion of an 800-calorie solid, realistic meal containing 99mTclabeled chicken liver. Scintigraphic measurements were made every 15 min for 6 h, and ultrasonic imaging of antral sections was undertaken every 15 min for the first 1 h and every 30 min thereafter. Total emptying times were calculated independently using the two methods, and the emptying patterns recorded by the two methods were compared. RESULTS: Maximal antral dilation occurred 30 min (range 0 –90 min) after the end of the meal and persisted until 96 ⫾ 42 min, by which time gastric radioactivity had decreased from its maximum by 43% ⫾ 23%. From this time on, the antral cross-sectional area returned toward the basal value, declining faster than the gastric counts recorded by scintigraphy. Total emptying times measured by ultrasound and by scintigraphy were in good agreement in all subjects, with a mean difference of only 4.5 min (limits of agreement, ⫺17.1 to 21.6 min). CONCLUSIONS: Ultrasonographic measurement of antral cross-sectional area provides a valid alternative to scintigraphy for the measurement of total gastric emptying of a solid meal. It is less reliable if other parameters of gastric emptying such as T1/2 are required. (Am J Gastroenterol 1999;94:2861–2865. © 1999 by Am. Coll. of Gastroenterology)
INTRODUCTION Ultrasound has been suggested as a valid alternative to scintigraphy for measurement of gastric emptying (1–5). The ultrasonographic method cannot discriminate between emptying rates of the liquid and solid phases of a meal, but offers some practical advantages over scintigraphy. For example, the necessary equipment is widely available and its use incurs only modest running costs. In most hospitals, access to ultrasonography disrupts routine clinical work less than the need for a gamma camera. Additionally, the lack of radiation hazard makes ultrasound especially suitable for repeated measurements, including research studies in healthy volunteers (1, 2). Preliminary suggestions of an acceptable concordance between the results of scintigraphic and ultrasonographic methods (2, 3) were followed by specific studies of liquid meals only (4 – 6); at present, the method is considered as a screening test to select patients who should undergo scintigraphic measurements (7–9). In particular, because the antrum exhibits a greater postprandial enlargement in dyspeptic patients than in controls (8, 10), a demonstration of good agreement between ultrasonographic and scintigraphic emptying measurements in normals cannot be assumed to apply in patients. The aim of our study was to compare scintigraphic measurements of gastric emptying of a solid meal with ultrasonographic measurements of antral emptying both in healthy subjects and in patients in whom delayed gastric emptying might be present. To optimize the comparison, the two methods of emptying measurement were applied simultaneously.
MATERIALS AND METHODS Ultrasonographic measurement of antral diameters and scintigraphic measurement of gastric emptying of 99mTc-labeled chicken liver were carried out simultaneously in 20 subjects (14 male, six female) aged 39 – 63 yr. Five were healthy members of our medical staff (controls, C), six were patients
2862
Benini et al.
with pathological gastroesophageal reflux (GER), and nine were patients with dysmotility-like dyspepsia (DYS). One dyspeptic patient had to be excluded from the study for lack of a hepatic acoustic window, which made ultrasound imaging unsatisfactory. Only a glass of tea with 1 teaspoonful of sugar was allowed for breakfast on the day of the study. At 1 PM, subjects were given the standard, two course meal that we previously described for ultrasound studies on gastric emptying (60 g of “macaroni alla bolognese” with 70 g of meat sauce and then 50 g of ham, 50 g of soft fatty cheese, one roll, and 250 ml of water; 800 Kcal, 15% from protein, 45% from fat, and 40% from carbohydrates) (11, 12). The only difference from previous studies was the substitution of 20 g of ham by 50 g of chicken liver. The raw liver was repeatedly injected in different sites with a total of 37 MBq of 99mTc colloid, cut in cubes of approximately 1 cm3 and then fried to a firm consistency. Both the liver and the oil used to fry it were then added to the meat sauce of the pasta. Walking in quiet surroundings was permitted between measurements. Ultrasonographic and scintigraphic measurements were always performed within a couple of minutes of each other, with patients sitting in an armchair tilted at an angle of 45 degrees backward. Scintigraphy After ingestion of the meal, radioactivity was measured from the anterior aspect of the abdomen for 1 min every 15 min up to h 6, or until radioactivity had reverted to the basal value. Measurements were corrected for decay according to the formula A ⫽ a ⫹ (a ⫻ 0.00192 ⫻ t), where A represents the value corrected for decay, a represents the measured value, and t represents the time (min) elapsed after the first postprandial measurement. The gastric region of interest was drawn visually on the computer display of the first frame and then replicated on later images. First frame activity in this region was taken as 100%. The residual radioactivity at each subsequent time was then plotted against time using an electronic spreadsheet (Excel 7.0 for Windows 95, Microsoft, Redmond, WA), and the tendency line for their relationship automatically calculated (moving mean of two points). The scintigraphic half time (T1/2) and total emptying time were identified from the points where this tendency line crossed the 50% and 3% values, respectively. The 3% value was chosen to allow for any nonspecific binding of tracer to the gastric wall. Ultrasound Real time ultrasonography (Toshiba Diagnostic Equipment model SSA-220A, fitted with a 3.75-MHz angular probe) was used to measure the diameters of the gastric antrum in the sagittal plane passing through the aorta. Measurements were made before the meal (basal), immediately after ingestion (time 0), then at 15-min intervals for the first 1 h and at 30-min intervals thereafter. The mean of three readings was calculated at each time during interperistaltic relaxation. The antral section was calculated using the formula
AJG – Vol. 94, No. 10, 1999
S ⫽ d1 ⫻ d2 ⫻ /4, where S represents the antral crosssectional area and d1 and d2 the measured diameters. The antral cross-sectional area was then plotted against the time. Visual inspection of the completed plot readily allowed identification of the initial period, when values of antral cross-sectional area were increasing or stable, and its distinction from the rapid antral emptying phase, when areas progressively declined. All the measurements made in this latter phase were used to calculate the regression line for antral emptying. The ultrasonographic half emptying time (T1/2) and total emptying time were identified from the points where this tendency line crossed the 50% and basal values, respectively. The percent residual distension of the antrum was calculated by relating the increase above basal value of the cross-sectional area at each time to the maximal postprandial increase above basal value observed in that subject. Statistical Methods Results are given as means ⫾ 1 SD. The statistical significance of the difference was calculated using Student’s t test for nonpaired data and analysis of variance. Pearson’s r coefficient between the results of the two methods is shown for descriptive purposes. The agreement between the two methods is better expressed by calculating the difference between their results and then plotting the difference against the mean values obtained by scintigraphy and by ultrasound. This obviates any possible effect of the measured values on the magnitude of the difference and gives an indication of the magnitude of potential error incurred by exchanging the new method for the old “reference” one. The mathematical and statistical bases for this evaluation are described by Bland and Altman (13). The limits of agreement were defined as the mean ⫾ 2 SD of the difference between the two methods.
RESULTS No difference was found in the basal and maximal postprandial antral sections as measured by ultrasound in the three groups of subjects (C 319 ⫾ 92 and 1667 ⫾ 425 mm2; GER 462 ⫾ 185 and 1693 ⫾ 444 mm2; DYS 403 ⫾ 117 and 1619 ⫾ 266 mm2, respectively). Figure 1 shows the change with time of the scintigraphic counts in the gastric area and of ultrasonographic antral cross-sectional areas, each expressed as percentage of the maximal value found in each subject. The scintigraphic counts attained their maximum immediately after the meal and, in most cases, began to fall thereafter. A scintigraphic “lag-phase” was evident only in three patients. In contrast, maximal antral dilation occurred at a mean of 30 min after the end of the meal (range 0 –90 min) and persisted until 96 ⫾ 42 min (echographic “lag phase”). At this time, the activity in the gastric region of interest had decreased by 43% ⫾ 24%. From this time onward, the antral cross-sectional area decreased progressively, regaining the basal value at the same time as scin-
AJG – October, 1999
Figure 1. Scintigraphic counts in the gastric area and ultrasonographic antral cross-sectional areas, each expressed as a percentage of the maximal value found in each patient, plotted against time after meal ingestion. Vertical bars represent 1 SD.
tigraphy. In the figure, mean antral cross-sectional areas do not reach 100% at any time because the maximum (100%) value was reached in each subject at different times. Figure 2 shows the radioactivity within the gastric region of interest when distension of the antrum was 50%, 30%, 20%, and 10% of the maximum. It is evident that scintigraphic retention varied widely in the different subjects, for any given ultrasonographic value. Additionally, it is clear that the proportion of the radioactive tracer remaining
Figure 2. Radioactivity in the gastric area of interest when the residual distension of the antrum is reduced to 50%, 30%, 20%, or 10% of the maximum. The columns represent mean values. Vertical bars represent 1 SD.
Measurement of Gastric Emptying by US and Scintigraphy
2863
Figure 3. Correlation between total emptying times measured by scintigraphy and by ultrasonographic measurement of the antral diameters in controls (triangles), in patients with dysmotility-like dyspepsia (squares) and with gastroesophageal reflux (open circles).
within the stomach was consistently less than suggested from the ultrasonographic measurement of antral diameters. The discrepancy was largest in the early phase of the study and progressively decreased. Figure 3 shows the relationship between the results of the two methods when total emptying times (i.e., times necessary for return of the stomach to the basal condition) were considered. Only in one patient did radioactive counts and antral diameters fail to return to basal values by h 6 and total emptying time had to be estimated by extrapolation. A good linear correlation between the two methods was found (y ⫽ 22 ⫹ 0.91x; R2 ⫽ 0.92), with no obvious differences among the three groups of subjects. Concordance between the two measurements was much poorer when half emptying times were considered. The time necessary for a 50% reduction of the postprandial antral dilation, as measured by ultrasound (164 ⫾ 34 min), was much longer than the T1/2 of scintigraphy (112 ⫾ 35 min). Interestingly, there was also a poor correlation between the scintigraphic T1/2 and the total emptying time, whether measured by ultrasound (y ⫽ 197 ⫹ 0.77x; R2 ⫽ 0.58) or by scintigraphy (y ⫽ 201 ⫹ 0.77x min; R2 ⫽ 0.54) (Fig. 4A and 4B). The difference in total emptying time between controls and patients (grouping together GER and DYS) is significant both for scintigraphy (C 265 ⫾ 19 min; GER ⫹ DYS 301 ⫾ 41 min, p ⬍ 0.02) and for ultrasound (259 ⫾ 23 vs 296 ⫾ 37 min respectively, p ⬍ 0.03). The difference plot (i.e., the difference between the results of ultrasonography and scintigraphy plotted against the mean result of the two methods) showed a mean difference of 4.5 min (limits of agreement, ⫺17.1 to 21.6 min); the
2864
Benini et al.
Figure 4. Correlation between scintigraphic half emptying times and total emptying times as measured by ultrasound (A) and by scintigraphy (B). (Symbols as in Figure 3.)
difference was ⬍20 min in all but one subject. There was no relationship between the mean emptying time as measured by the two methods and the value of their difference. Finally, there was no evidence that the magnitude of the difference varied in the three groups of subjects (analysis of variance), although the relatively small number of subjects means that this can be no more than a tentative conclusion.
DISCUSSION Ultrasound has been suggested as a cheap and valid alternative to scintigraphy for the measurement of gastric emptying. The necessary equipment is widely available and the test does not cause radiation exposure, allowing repeated measurements to be carried out when the effects of drugs or therapeutic procedures are to be evaluated (14, 15). Although gastric dysmotility problems are frequent in clinical practice, the measurement of gastric emptying by ultrasound is restricted to a few centers. An explanation for this may be the methodological difficulty of the procedure, as the posterior wall of the stomach may be difficult to identify after the ingestion of solids. Restricting ultrasonographic studies to liquid meals seems reasonable (16) but, in fact, consistent results have been reported with solid meals when experienced operators are involved (11, 17). The complex relationship between antral cross-sectional area and intragastric volume has caused difficulty for investigators using ultrasound for gastric emptying measurements and also when attempts are made to reconcile the findings reported by different investigators. Although a linear relationship has been suggested between the volume of ingested liquid and antral size (17), other evidence indicates that emptying of liquids relates to the content of the proximal stomach and that, after liquid ingestion, the content of the distal stomach initially increases, then remains relatively constant for a period and decreases thereafter (5, 18). Factors favoring antral distension (such as gravity and fundic
AJG – Vol. 94, No. 10, 1999
tone) or contraction (such as vagally mediated fundic adaptive relaxation and phasic antral contractions) modify antral size for any given intragastric volume. The modification of antral size by the caloric value of test meals may explain the good agreement between ultrasound and scintigraphy for gastric emptying of water, but not of milk (4). The variables become even more important after solid meals because of the important role of the fundus in solid meal accommodation. For example, when two of our volunteers ate progressively increasing volumes (100 to 900 ml) of chocolate pudding on different days, antral cross-sectional areas measured 0 and 30 min after completion of the meal increased linearly with volume to a maximum of 500 ml, but no further increase was seen with the larger meals (unpublished data). The time required for the antrum to return to baseline (“total emptying time”), as measured by ultrasound, can be divided into two periods. In the first, the antral cross-sectional areas progressively increase toward a maximum, and then remain fairly constant (described here as the “ultrasonographic lag phase”); in the second, the cross-sectional areas gradually decrease to their basal value. Although this behavior resembles the scintigraphic sequence of solid emptying, it relates to different aspects of emptying. The “lag phase” on ultrasound clearly does not represent the period before emptying begins, as the simultaneous scintigraphic study has shown that a substantial proportion of the meal is leaving the stomach. When antral size begins to decrease, 40% of the meal has emptied. From this time on, the antral cross-sectional area falls fairly quickly, regaining the basal value at the same time as scintigraphic total gastric emptying. The difference between the two tests (⬍10% in most patients) suggests that this use of ultrasonography produces valid emptying measurements. The different time course of the decrease of radioactivity and of antral size after a solid meal is similar to results described in abstract form by other investigators, who either measured the half emptying time (19), or visually compared the residual curves (20). The difference between ultrasound and scintigraphy in the early phases of our study may even be an underestimate of the true difference, because the use of anterior-only acquired scintigraphic counts may overestimate the lag phase as a result of anterior movement of the meal (21). We believe that any error was small, because a detectable scintigraphic lag phase was observed in only three of our subjects. In conclusion, ultrasonographic measurement of antral diameters offers a valid alternative to scintigraphy for the measurement of gastric emptying of solid meals if used to determine total emptying time. It is unlikely to be satisfactory if other gastric emptying measurements such as T1/2 are desired. The accuracy of the method seems as good in patients with dysmotility-like dyspepsia or gastroesophageal reflux as in healthy normal subjects.
AJG – October, 1999
Measurement of Gastric Emptying by US and Scintigraphy
ACKNOWLEDGMENT This work was supported by grant 95.00773.PF41 from the Italian National Research Concil (CNR)-Targeted Project “Prevention and Control of Disease Factors”, Subproject “Alimentation”. Reprint requests and correspondence: Luigi Benini, M.D., Department of Gastroenterology, Rehabilitation Hospital, 37067 Valeggio sM, Verona, Italy. Received Aug. 26, 1997; accepted Apr. 30, 1999.
REFERENCES 1. Bateman DN, Whittingham TA. Measurement of gastric emptying by real time ultrasound. Gut 1982;23:524 –7. 2. Holt S, Cervantes J, Wilkinson AA, et al. Measurement of gastric emptying rate in humans by real-time ultrasound. Gastroenterology 1986;90:918 –23. 3. Bolondi L, Bortolotti M, Santi M, et al. Measurement of gastric emptying time by real-time ultrasonography. Gastroenterology 1985;89:752–9. 4. Marzio L, Giacobbe A, Conoscitore P, et al. Evaluation of the use of ultrasonography in the study of liquid gastric emptying. Am J Gastroenterol 1989;84:496 –500. 5. Hveem K, Jones KL, Chatterton BE, et al. Scintigraphic measurement of gastric emptying and ultrasonographic assessment of antral area: relation to appetite. Gut 1996;38:816 –21. 6. Mamtora H, Thompson DG. Gastric ultrasound. In: Read NW, ed. Gastrointestinal motility: Which test? London: Wrightson Biomedical Publishing, 1989:99 –104. 7. Heading RC, Bolondi L, Camilleri M, et al. Gastric emptying. Gastroenterol Int 1992;5:203–15. 8. Troncon LEA, Bennet RJM, Ahluwalia NK, et al. Abnormal intragastric distribution of food during gastric emptying in functional dyspepsia patients. Gut 1994;35:327–32. 9. Ahluwalia NK, Thompson DG. Ultrasonography. In: Kumar
10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
21.
2865
D, Wingate D, eds. An illustrated guide to gastrointestinal motility. Second edition. Edinburgh: Churchill Livingstone, 1993:242–55. Hausken T, Berstad A. Effect of nitric oxide on antral motility and symptoms in patients with functional dyspepsia. Scand J Gastroenterol 1994;29:23– 8. Benini L, Castellani G, Sembenini C, et al. Gastric emptying of solid meals in achalasic patients after successful pneumatic dilatation of the cardia. Dig Dis Sci 1994;39:733–7. Benini L, Castellani G, Bardelli E, et al. Omeprazole causes a delay in gastric emptying of digestible meals. Dig Dis Sci 1996;41:469 –74. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. Benini L, Brighenti F, Castellani G, et al. Gastric emptying of solids is markedly delayed when meals are fried. Dig Dis Sci 1994;39:2288 –94. Benini L, Castellani G, Brighenti F, et al. Gastric emptying of a solid meal is accelerated by the removal of dietary fibre naturally present in food. Gut 1995;36:825–30. Irvine JE, Tougas G, Lappalainen R, et al. Reliability and interobserver variability of ultrasonographic measurement of gastric emptying rate. Dig Dis Sci 1993;38:803–10. Ricci R, Bontempo I, Corazziari E, et al. Real time ultrasonography of the gastric antrum. Gut 1993;34:173– 6. Collins PJ, Houghton LA, Read NW, et al. Role of proximal and distal stomach in mixed solid and liquid meal emptying. Gut 1991;32:615–9. Bolondi L, Santi V, Bortolotti M, et al. Correlation between scintigraphic and ultrasonographic assessment of gastric emptying. Gastroenterology 1986;90:1349 (abstract). Tani H, Haruma K, Hata J, et al. Ultrasonographic assessment of gastric emptying rate of solids: A comparison of ultrasonography and scintigraphy. Gastroenterology 1996;110:768 (abstract). Tothill P, McLoughlin GP, Heading RC. Techniques and errors in scintigraphic measurements of gastric emptying. J Nucl Med 1978;19:256 – 61.