Standardization of a simplified scintigraphic methodology for the assessment of gastric emptying in a multicenter setting

THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2000 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.

Vol. 95, No. 1, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(99)00762-5

Standardization of a Simplified Scintigraphic Methodology for the Assessment of Gastric Emptying in a Multicenter Setting Gervais Tougas, M.D.C.M., F.R.C.P.C., Ying Chen, M.Sc., Geoffrey Coates, M.B.B.S., F.R.C.P.C., William Paterson, M.D., F.R.C.P.C., Christian Dallaire, M.D., F.R.C.P.C., Pierre Pare´, M.D., F.R.C.P.C., Michel Boivin, M.D., F.R.C.P.C., Alain Watier, M.D., F.R.C.P.C., Sandra Daniels, M.Sc., and Nicholas Diamant, M.D., F.R.C.P.C. McMaster University, Hamilton, Ontario; Queen’s University, Kingston, Ontario; Universite´ Laval, Que´bec, Que´bec; Universite´ de Montre´al, Montre´al, Que´bec; Universite´ de Sherbrooke, Sherbrooke, Que´bec; Astra Pharma, Mississauga, Ontario; and University of Toronto, Toronto, Ontario, Canada

OBJECTIVE: Scintigraphy remains the gold standard to study gastric emptying. The technique is onerous and normal values vary between centers. Standardized protocols, although desirable, are not presently available. We validated a simplified scintigraphic protocol in a multicenter setting. METHODS: In 69 healthy volunteers from seven Canadian institutions, gastric emptying of a standard meal (99mTclabeled beef liver) was assessed by scintigraphy every 10 min for 1 h, then every 20 min for the next 2 h. Gastric retention was fitted to a power exponential model, Propt ⫽ {-(␬t)␤ } with Propt ⫽ proportion of retention at time t, either using all 13 time intervals (conventional technique) or using measurements at 0, 1, 2, and 3 h (simplified technique). RESULTS: The power exponential model yielded identical emptying curves and T 1⁄2 values with both techniques. Gastric emptying was more rapid in men than in women ⬍ 35 yr (p ⬍ 0.01) and in younger than in older men (p ⬍ 0.005). Gastric emptying was slower in women from Que´bec than in women from Ontario (p ⬍ 0.04). Gastric retention was similar at 1, 2, and 3 h among the seven centers. Gastric emptying of a beef liver meal was slower than that of a low fat egg substitute (p ⬍ 0.03). CONCLUSIONS: A simpler scintigraphic approach, using four rather than 13 samples, provides results comparable to those of the conventional technique. This simpler approach provides an economical, yet accurate, alternative to the techniques presently used and is applicable to a multicenter setting. (Am J Gastroenterol 2000;95:78 – 86. © 2000 by Am. Coll. of Gastroenterology)

INTRODUCTION Gastroparesis is usually defined as a symptomatic delay in gastric emptying in the absence of any definable mechanical obstruction (1). Delayed gastric emptying is one of the factors implicated in the pathogenesis of several common

GI conditions and symptoms, including gastroesophageal reflux disease (GERD), and dyspepsia (2, 3). Furthermore, symptoms usually associated with gastroparesis, such as nausea, vomiting, early satiety, bloating, and belching, can also be the result of several other mechanisms and often correlate poorly with delayed gastric emptying (4). Finally, it is often difficult to identify gastroparesis on clinical rounds alone (4). Therefore, it is important to establish whether or not there is an actual delay in gastric emptying, as prokinetic drugs may be useful in patients who have a prolonged gastric emptying but may be of limited efficacy in those with normal, or even rapid, gastric emptying. Gastric scintigraphy is the most commonly used test for the assessment of gastric emptying and, although several other methods have been proposed as alternatives, scintigraphy remains the gold standard. However, a standardized approach to the scintigraphic assessment of gastric emptying is still lacking, and techniques can vary substantially between individual centers. In a Canadian survey of gastric scintigraphy, a wide variability was found in the actual technique as well as in the values believed to represent the normal range (5). Thus, comparisons between gastric emptying studies obtained in different institutions and pooling of results in the context of clinical trials are difficult. In addition, as presently conducted, scintigraphy is costly, labor intensive, and time consuming, both for the patients and for the personnel involved. The aims of the present study were 1) to establish a standardized yet simple scintigraphic protocol for the assessment of gastric emptying in seven academic Canadian institutions, and 2) to determine what constitutes normal emptying of a standardized meal. We have also attempted to simplify data analysis and the requirement for frequent sampling to make the technique more widely usable. Finally, we examined whether there are regional differences in gastric emptying between residents of French- and Englishspeaking Canada.

AJG – January, 2000

MATERIALS AND METHODS Subjects A total of 69 healthy volunteers (33 male and 36 female; mean age ⫽ 39 yr, range 20 – 63 yr) were recruited from seven different Canadian institutions. The centers involved were Hamilton (McMaster University Medical Centre), Toronto (The Toronto Hospital, Western Division), Kingston (Hotel-Dieu Hospital), Montre´al (Hoˆpital Saint-Luc), Sherbrooke (Center Hospitalier Universitaire de Sherbrooke), Que´bec (Hoˆtel-Dieu de Que´bec and Hoˆpital Saint-Franc¸ois d’Assise). All subjects were in good health and without GI, neurological, endocrine, renal, hepatic, pulmonary, or cardiovascular disease. They had no prior GI surgery, with the exception of appendectomy. In addition, none of the subjects had any symptoms suggestive of delayed gastric emptying. The subjects were currently not taking any medication and had not been smoking for ⱖ14 days. Female participants were either postmenopausal or had had a negative pregnancy test within 3 days of the study. Signed, informed consent was obtained from each participant. The study was approved by the ethical review board of every institution involved. Test Meal The standard meal used in every center consisted of a slice of beef liver (57 g) mixed into 225 g of Campbell’s chicken stew (225 g). The caloric value of the meal was 275.5 kcal (22.8% carbohydrate, 43.7% protein, and 33.5% fat). The meal was prepared de novo on the morning of the study. Using a microwave oven, liver and stew were first separately thawed but not cooked. Then 18.5 Mbp (0.5 mCi) of 99m Tc sulfur colloid were drawn up into a tuberculin syringe and made up to 1 ml using normal saline. Afterward, the isotope was injected into the slice of liver, using 10 discrete injections so as to cover as much of the total surface area as possible. Immediately afterward, the liver was fried in an electric, Teflon-coated frying pan for approximately 1 min on each side. It was then removed from the pan, cut into square pieces of approximately 1 cm2, and then returned to the frying pan and fried until thoroughly cooked (5 min more). At that time, the thawed chicken stew was added to the pan and the entire mixture heated. The meal was then taken to the scintigraphy room and the fasting volunteer instructed to consume it within 10 min. In two centers (Hamilton and Montre´al), a low fat meal using a synthetic egg-substitute was also used to assess gastric emptying. This meal consisted of a scrambled egg substitute (120 g EggBeater, equivalent to the volume of two large eggs), two slices of bread, strawberry jam (30 g), and drinking water (120 ml). The egg substitute was first radio-labeled using 1 mCi (37 Mbp) of 99mTc-labeled sulfur colloid, which was mixed into the liquid egg substitute preparation. The labeled mixture was then either scrambled in a nonstick frying pan for 4 –5 min or cooked in a micro-

Simplified Scintigraphy for Gastric Emptying Assessment

79

wave oven for 1–2 min, stirring the mixture once during cooking. The meal was then ingested within 10 min. The caloric value of the low fat meal was 255 Kcal; it contained 2% fat, with 72% carbohydrate 24% proteins and 2% fiber). Scintigraphy To avoid possible diurnal variations, all studies began between 09:00 and 10:00, to finish between 12:00 and 13:00. Anterior and posterior images of the gastric region were taken within 1 min of the completion of the meal (defined as time 0, T0). Subsequent images were obtained every 10 min during the first 1 h and every 20 min afterward until 3 h had elapsed. For every time point, both 1-min anterior and posterior images were obtained in the 140 KeV 99Tc peak with a 20% window. All images were collected with the volunteers standing and leaning right up against the camera. Between imaging periods, the subjects were allowed to sit upright in an adjacent waiting area. This frees the gamma camera for most of the time. Analysis of Gastric Emptying For each image taken, the gastric region was drawn using a manually operated cursor. The geometric mean was calculated as the square root of the product of the anterior and posterior counts corrected for time decay. Several parameters were individually calculated for every subject. These include calculation of percent gastric retention at 13 data collecting times, gastric retention curves spanning the entire period from the end of the meal (T0) to 180 min later, and estimation of lag phase and T 1⁄2. Gastric retention curve were generated using a power exponential model, Propt ⫽ {⫺(␬t)␤}, where Propt is the proportion of retention at time t. Lag phase was defined a priori as the period extending from the end of the meal to the time at which 5% of the original gastric content has left the stomach, and T 1⁄2 as the time to emptying of 50% of the original gastric content. We compared two different approaches to generate gastric retention curves: in the first instance we used the conventional technique, and all data samples from each time interval (every 10 min in the first 1 h, and every 20 min thereafter) used to determine lag phase and T 1⁄2. In the second instance we used a simplified method in which only the gastric counts at T0 and 60, 120 and 180 min were used to generate the gastric emptying curve and subsequently calculate lag phase and T 1⁄2. Statistics As most of the measured data were not normally distributed, results are primarily reported as medians, with 90th and 95th percentiles, and plots presented as box and whiskers plots. The central horizontal line of the box represents the median and the bottom and top edges the 25th and 75th percentiles, respectively. The whiskers extend from the box upward and downward to the 10th and 90th percentiles, respectively. Data that are normally distributed are reported as mean ⫾ 2 SD (T 1⁄2 in particular). When appropriate, comparisons by

80

Tougas et al.

AJG – Vol. 95, No. 1, 2000

Table 1. Descriptive Statistics of Gastric Emptying of a Healthy Volunteers (N ⫽ 69) % Gastric Retention 1h 2h 3h

99m

Tc-Labeled Beef Liver Meal

Median

1st Quartile

3rd Quartile

77 42 ⫾ 36 17

70 29 10

85 53 27

Skewness

Probability of Normal Distribution

90th Percentile

95th Percentile

⫺1.29 ⫺0.12 0.33

⬍0.0001 ⬍0.5 ⬍0.06

92 67 34

95 70 42

* As data for % retention at 2 h are normally distributed, mean ⫾ SD are presented.

groups were performed using repeated analysis of variance (ANOVA) based on rank, two sample t test, Kruskal-Wallis, and Wilcoxon rank-sum tests. Differences are considered significant only at p ⬍ 0.05.

RESULTS Sample Population: Demographic Characteristics A total of 69 subjects (range 20 – 63 yr) were studied. There were 33 men and 36 women. In all, 29 subjects (15 men and 14 women) were ⬍35 yr of age, and 40 subjects (18 men and 22 women) were ⱖ35 yr. Each of the seven centers contributed 10 subjects, with the exception of one center, which recruited nine individuals. There were a total of 29 subjects from Ontario and 40 from Que´bec. Gastric Emptying of a Liver Meal: Descriptive Values Table 1 shows the median measured percent gastric retention at 1, 2, and 3 h. The data are not normally distributed (Fig. 1A, B, and C), particularly at 1 h (the probability of normal distribution p ⬍ 0.0001); consequently, the median and 90th or 95th percentile, rather than the mean and SD, are used to determine normal values. At 1, 2, and 3 h, the 90th percentile values for percent gastric retention are 92%, 67%, and 34%, respectively. The 95th percentile values are 95%, 70%, and 42%, respectively, for the same intervals. Figure 2 shows the estimated gastric emptying curves generated by the power exponential model (Propt ⫽ {⫺(␬t)␤}) using median ␬ and ␤ values derived from the entire database. Both the simplified method, using only four data points (0, 60, 120, and 180 min), and the conventional method, which uses a total of 13 data points, provide essentially identical results. The median lag phase and mean T 1⁄2, as determined by the conventional method, were 25 min (5th and 95th percentile ⫽ 7 and 55 min) and 107 min (normal range, defined as mean ⫾ 2 SD ⫽ 46 –169 min), respectively. The results are almost identical when only four data samples are used, with the median lag phase being 24 min (5th and 95th percentiles ⫽ 7 and 54 min) and the mean T 1⁄2 being 107 min (normal range, defined as mean ⫾ 2 SD ⫽ 46 ⫺ 169 min). The calculated ␬ and ␤ were 0.0077 and 1.80 with the conventional method and were 0.0077 and 1.85 with the simplified scintigraphic method. All data are summarized in Table 2 and Table 3.

Effect of Gender and Age on Gastric Emptying Although there are no overall significant differences between the gastric emptying of men and women, there are significant differences between men and women ⬍35 yr of age. Men exhibit a significantly faster gastric emptying than women ⬍35 yr of age (p ⬍ 0.01, Fig. 3A). This is reflected in longer mean T 1⁄2 in women than in men ⬍ 35 yr (118 and 91 min, respectively, p ⬍ 0.04). However there are no differences in gastric emptying between men and women ⱖ35 yr (p ⬍ 0.4) (Fig. 3B). As shown in Figure 4A, younger men (⬍35 yr) empty their stomach more rapidly than men ⬎35 yr (p ⬍ 0.005 overall, p ⬍ 0.02, 0.02 and 0.01 at 1, 2, and 3 h, respectively). Correlation between percent residual and age of men was significant at 2 and 3 h (p ⬍ 0.02) and nearly significant

Figure 1. Overall distribution of percent gastric retention of a beef liver meal at 1, 2, and 3 h in 69 subjects. The distribution of the data are markedly skewed at 1 and 3 h, but less so at 2 h.

AJG – January, 2000

Simplified Scintigraphy for Gastric Emptying Assessment

81

than in those located in English Canada (p ⬍ 0.04 overall). Median lag phase and mean T 1⁄2 were, respectively, 20 and 96 min in Ontario and 27 and 116 min in Que´bec (p ⬍ 0.009). This interprovincial difference is entirely attributable to a significantly slower gastric emptying in women from Que´bec (French Canada) compared to gastric emptying in women from Ontario (English Canada, p ⬍ 0.04 overall, p ⬍ 0.02, 0.04 and 0.08 at 60, 120 and 180 min, see Fig. 6A). There were no interprovincial differences in the gastric emptying among men (p ⬍ 0.4 overall) (Fig. 6B). Comparison Between Gastric Emptying of a Tc-Labeled Beef Liver Meal and a Low Fat Meal Using 99mTc-Labeled Egg Substitute (Eggbeater) In two centers (Hamilton and Montre´al), we further compared the values obtained using 99mTc-labeled beef liver and a low fat meal using 99mTc-labeled synthetic egg preparation (EggBeater). The protocols used were identical in both studies except for the preparation and nature of the meal. The demographic characteristics of the participants were comparable. The two meal have a very similar caloric content (liver meal ⫽ 275.5 kcal, low fat meal ⫽ 255 kcal). The beef liver meal has higher fat content and different composition than the low fat meal (liver meal ⫽ 33.5% fat, 22.8% carbohydrate, and 43.7% protein; low fat meal ⫽ 2% fat, 72% carbohydrate, 24% protein, plus 2% fiber). Gastric emptying of a low fat meal was significantly more rapid than gastric emptying of a beef liver meal which contains more fat (overall, p ⬍ 0.03; p ⬍ 0.05 and 0.007 at 2 and 3 h, respectively, Fig. 7). Median lag phase and mean T 1⁄2 were 17 min and 81 min, respectively, when using the low fat egg substitute meal compared to 18 and 96 min with the beef liver meal (p ⬍ 0.02). Although most subjects found the low fat meal easier to ingest, both meals were well tolerated and could be ingested easily by everyone within the allotted time. Technical staff involved generally found preparation of the low fat meal easier than the preparation of the beef liver meal, which requires an electric frying pan and freezer facilities to store the frozen constituents. Cleaning of the cooking utensils was also mentioned as an inconvenient feature of the liver meal. 99m

Figure 2. Overall estimated percent gastric retention curves generated using a power exponential model, either with frequent imaging (dashed line, median ␬ ⫽ 0.0077 and median ␤ ⫽ 1.80), or only with imaging at 0, 60, 120, and 180 min (solid line, median ␬ ⫽ 0.0077 and median ␤ ⫽ 1.85). The power exponential model is defined as propt ⫽ {-(kt)␤} (see Materials and Methods).

at 1 h (p ⬍ 0.07). The calculated median lag phase and mean T 1⁄2 were 21 and 91 min, respectively, in men ⬍35 yr and 29 and 115 min in men ⱖ35 yr (p ⬍ 0.05). There were no comparable differences between women ⬍35 yr and ⬎35 yr of age (p ⬍ 0.8 overall) (Fig. 4B). Interinstitutional and Interprovincial Variability As can be seen in Figure 5, there were no overall differences in gastric emptying among participants from the seven different centers (p ⬍ 0.2). However, when grouped according the province of origin, there were significant differences between the four centers from Que´bec (⬎90% French Canadian origin) and the three centers from Ontario (⬍5% of French Canadian origin). Overall, gastric emptying was significantly slower in the centers located in French Canada

DISCUSSION A precise definition of what constitutes normal gastric emptying remains elusive. A direct relationship is often missing

Table 2. Calculated Parameters Using Power Exponential Model Based on Four Data Points

␬ ␤ T50 (min)* Lag phase (min)

Median (1st to 3rd Quartile)

Probability of Normal Distribution

Skewness

10th to 90th Percentile

5th to 95th Percentile

0.0077 (0.0065–0.0090) 1.85 (1.51–2.21) 107 (46–169) 24 (18–38)

⬍0.0001 ⬍0.004 ⱕ0.5 ⬍0.002

2.36 1.09 ⫺0.17 0.90

0.0056–0.0119 1.29–2.53 68–151 13–49

0.0052–0.0156 1.26–2.89 53–160 7–54

* As data for T50 were normally distributed, mean ⫾ 2 SD are reported.

82

Tougas et al.

AJG – Vol. 95, No. 1, 2000

Table 3. Calculated Parameters Using Power Exponential Model Based on 13 Data Points

␬ ␤ T50 (min)* Lag phase (min)

Median (1st to 3rd Quartile)

Probability of Normal Distribution

Skewness

10th to 90th Percentile

5th to 95th Percentile

0.0077 (0.0064–0.0090) 1.80 (1.48–2.14) 107 (46–169) 25 (17–36)

⬍0.0001 ⬍0.0001 ⬍0.8 ⬍0.0001

2.58 1.38 ⫺0.13 1.18

0.0056–0.0118 1.29–2.64 71–149 12–49

0.0051–0.0164 1.21–2.91 46–158 7–55

* As data for T50 were normally distributed, mean ⫾ SD reported.

between symptoms and evidence of delayed gastric emptying (6) (7). Patients with seemingly delayed gastric emptying are often completely asymptomatic, whereas others, with ostensibly normal gastric emptying, are very symptomatic. The usual symptoms associated with delayed gastric emptying (nausea, vomiting, bloating, early satiety, excessive belching, and regurgitation) can result from not only from gastroparesis but from other GI factors as well as from several metabolic or neurological causes (8). In these conditions, there is often no alteration of gastric emptying. It is therefore clinically important to have methods capable of objectively assessing gastric emptying in patients in whom it may be contributing to symptoms, to better determine the respective contributions of gastroparesis and other factors to the symptoms reported. An objective and reproducible assessment of gastric emptying would also be very useful in clinical trials examining the effects and clinical utility of prokinetic compounds on gastric emptying and symptoms. Disorders of gastric emptying being relatively uncommon, therapeutic trials studying the efficacy of new drugs and approaches in those conditions usually require the involvement of several centers for patient recruitment. At the present time, the techniques, meals, and methods of analysis used for the scintigraphic assessment of gastric emptying vary widely among institutions (9 –13), and large multicenter trials with standardized techniques are very difficult to conduct because of this lack of standardization. As the data are generated differently at

different institutions, pooling of results is problematic if not impossible. Our study is the first multicenter study to use a standardized methodology in seven different institutions to characterize gastric emptying, in this case of healthy Canadian adults of both genders. Our results demonstrate that, by using a standardized methodology, data from different centers can in fact be pooled. These seven centers collectively generated a database of 69 healthy volunteers, allowing us to statistically determine the normal range in a healthy adult Canadian population of different ages. Furthermore, the large sample size also provided us with the statistical power to examine the data set in a number of different ways and to see if the procedure could actually be simplified and more easily used in the clinical arena. Simplification has practical implications. Presently, gastric emptying scintigraphy requires frequent, if not continuous, sampling for 2– 4 h, thereby monopolizing the use of a gamma camera for almost the entire period (14, 15). This is something that is increasingly difficult in an era of decreasing resources and accessibility. In the clinical setting, identification of delayed gastric emptying is the necessary clinical endpoint. Therefore, our approach was designed to define normal gastric emptying of a standardized meal and to identify subjects with delayed gastric emptying. In this context, the ideal measure is one that will focus on the demonstration of an abnormal gastric retention of a meal for a prolonged period. The measure-

Figure 3. Box whiskers plots of the male and female percent gastric retention at 1, 2, and 3 h, in subjects aged ⬍35 yr (A) and in subjects ⱖ35 yr (B). The central horizontal line of the box represents the median and the bottom and top edges the 25th and 75th percentiles, respectively. The whiskers extend from the box upward and downward to the 10th and 90th percentiles, respectively. There was a clear gender effect in younger persons, p ⬍ 0.01, which was not present among older individuals (p ⬍ 0.4).

AJG – January, 2000

Simplified Scintigraphy for Gastric Emptying Assessment

83

Figure 4. Box whiskers plots of the percent gastric retention at 1, 2, and 3 h among groups separated according to age (⬍35 yr vs ⱖ35 yr) and gender. The central horizontal line of the box represents the median and the bottom and top edges the 25th and 75th percentiles, respectively.The whiskers extend from the box upward and downward to the 10th and 90th percentiles, respectively. A: The significant difference in percent gastric retention at 1, 2, and 3 h between younger (⬍35 yr) and older men is shown. B: Percent gastric retention at 1, 2, and 3 h between younger and older women. There was a significant overall age effect in men (p ⬍ 0.005) but not in women (p ⬍ 0.8).

ment of percent gastric retention after 2, 3, or even 4 h is probably a more useful index (16, 17). However, measures such as T 1⁄2 or the lag phase may be more important than percent gastric retention over time to assess gastric function in relation to symptoms such as early satiety and fullness, or in conditions such as functional dyspepsia. In these conditions, it is the initial part of the postprandial period that seems to be disordered (18). Several aspects of the data deserve further attention. Gastric emptying data are often described in terms of mean and SD (5, 12). This approach implies that the data are normally distributed, which is clearly not the case based on our analysis (see Fig. 1). A more appropriate approach would be to use the median and a range defined as either between the 5th and 95th percentiles or between the 10th and 90th percentiles to define what constitutes normal emptying. If such an approach is used, delayed gastric emptying would

Figure 5. Box whisker plots of the percent gastric retention at 1, 2, and 3 h in all seven individual centers. The central horizontal line of the box represents the median and the bottom and top edges the 25th and 75th percentiles, respectively. The whiskers extend from the box upward and downward to the 10th and 90th percentiles, respectively. There was no center effect at 1, 2, or 3 h (p ⬍ 0.2).

be defined by a gastric retention of ⬎34% (90th percentile) or ⬎42% (95th percentile) of a standard beef liver meal at 3 h. These values are greater than those generally reported. A more commonly used measure of gastric emptying is the T50 or T 1⁄2 emptying time, the estimated time required for the stomach to empty one-half of the original meal (19). As our data indicate, these calculated values are more normally distributed, and the mean ⫾ standard deviations is probably appropriate to define the normal range. However, in a recent study examining the performance of scintigraphic gastric emptying studies in 173 different Canadian hospitals, abnormal emptying was defined in a number of ways, with widely different values, based on several different criteria to describe the normal range (5). These criteria sometimes, but not always, included the mean ⫾ 1 or 2 SD. It ranged from as low as 36 min to as high as 125 min as the upper limit of normal for T 1⁄2. On the basis of our own data, generated among seven different centers, these previous values seem to be underestimations and likely produce a high incidence of false-positive results. Based on the present database, T 1⁄2 values ⬍135 min (mean and 2 SD) should be considered normal with a low fat meal (an egg substitute), whereas a liver- based meal (with a higher fat content) would require T 1⁄2 values ⬎169 min for delayed gastric emptying to be demonstrated. Poitras et al. (20) have previously reported comparable values using similar meals in another series of healthy volunteers examined in two of the institutions included in the present study. In addition, using a low fat meal indentical to the meal used in the present study, we have previously reported nearly identical results in a large international multicenter study involving 11 institutions in four different countries (21). If T 1⁄2 is used to assess gastric emptying in that previous study, delayed gastric emptying would be defined by a T 1⁄2 ⬎141 min, a figure that is almost identical to that reported in the present study (135 min).

84

Tougas et al.

AJG – Vol. 95, No. 1, 2000

Figure 6. Calculated power exponential model of gastric emptying of a beef liver meal in subject from Ontario (English Canadians, solid line) and from Que´bec (French Canadians, dashed line). A: Power exponential curves for women, ␬ ⫽ 0.0084 and ␤ ⫽ 1.72 among women from Ontario. ␬ ⫽ 0.0069 and ␤ ⫽ 1.95 among women from Que´bec (p ⬍ 0.04). B: Power exponential curves in men; ␬ ⫽ 0.0078 and ␤ ⫽ 1.70 among men from Ontario. ␬ ⫽ 0.0079 and ␤ ⫽ 1.95 among men from Que´bec (p ⬍ 0.4).

Performance of gastric scintigraphy and analysis of the data are cumbersome with the conventional methodology, because of the necessity to obtain frequent (if not continu-

Figure 7. Box whisker plots of the percent gastric retention at 1, 2, and 3 h of a beef liver meal (Beef) and a low fat meal (Egg), using EggBeater. The central horizontal line of the box represents the median and the bottom and top edges the 25th and 75th percentiles, respectively. The whiskers extend from the box upward and downward to the 10th and 90th percentiles, respectively. There was a significant overall difference between the two meals (p ⬍ 0.03), with the gastric retention of the beef liver meal being greater than that of the egg substitute meal.

ous) imaging, to get a precise representation of the actual emptying curve, and a reasonable estimate of parameters such as T 1⁄2, and lag phase (10, 11, 15). Less frequent imaging would go a long way in improving the clinical utility and the applicability of the technique. Using our large database, we have demonstrated that a much simpler approach to the analysis of scintigraphic data for the determination of gastric emptying is feasible and reliable. The simplified power exponential model makes it possible to describe describe and predict accurately the emptying behavior of the stomach after a standardized meal, using only four scintigraphic readings rather than the 13 readings presently necessary. Our simplified approach provides nearly identical values for the estimation of lag phase, and T 1⁄2 to those obtained with the much more labor-intensive conventional methods. The simplified approach only requires hourly scintigraphic sampling (T0, then at 60, 120, and 180 min), does not monopolize use of scintigraphic equipment for the entire period of the study, and therefore allows the use of the equipment for other purposes between readings. In an era of cost constraints and increasing pressure toward a better, if not optimal, use of equipment, this type of approach certainly has its appeal. Gastric emptying has been shown to vary depending on

AJG – January, 2000

the age and gender of the subjects (22, 23). It tends to be more rapid in younger subjects and in men. In the present study, we find that in men, gastric emptying is more rapid in younger (⬍35 yr) than older (ⱖ35 yr) individuals, but this age effect was not seen in women. Others have previously reported similar effects of age on gastric emptying, but not specifically in men (24 –26). However different meals, with lower fat content, were used. Horowitz et al. (26) further argued that, although the age-related difference in gastric emptying is statistically significant, it was unlikely to be clinically significant. Our present results would tend to support this assertion, as we found such a large variance and a substantial overlap in what constitutes normal emptying in both age groups. Furthermore, Madsen has also questioned the effect of age on gastric emptying, having found it to be similar across a wide age range (27). Men have been shown to empty their stomachs more rapidly than women (22, 28 –30). The present study supports this observation, but only in younger men and women. Interestingly, there were no detectable differences between men and women ⬎35 yr. The difference in gastric emptying could be entirely attributed to the difference observed in subjects ⬍35 yr. This is in keeping with several previous reports including the study of Knight et al. (31), who have reported a similar difference in gastric emptying between young women and young men. When a beef liver meal is compared with a low fat alternative such as the egg substitute used in the present study, we find that the latter is emptied faster and more completely at 3 h. In addition, the low fat meal was believed to be more palatable and was actually preferred by most subjects. Still, both the low fat meal and the beef liver meal were generally well tolerated. Although the higher fat content of the beef liver meal is likely to result in a slower gastric emptying, it is also possible that the different consistency of the beef liver and the low fat meal may also contribute to the observed difference. Although there was no difference in gastric emptying between men from Ontario and Que´bec, there was a surprising but significant difference between women of the two provinces, gastric emptying being much slower in Que´bec women. The difference cannot be explained on the basis of age. However, we are unable to confirm whether differences in body mass index (BMI) could account for the difference (29). In the present study, premenopausal women were not studied at a specific time in their menstrual cycle. Therefore, an involuntary but systematic bias resulting in the subjects being studied at different times of their cycle in the three English Canadian and the four French Canadian centers cannot be completely ruled out. However, on statistical grounds alone, this is very unlikely. Cultural or dietary factors could also be involved, but further studies will be needed to confirm and possibly to explain the reasons behind this surprising observation. In summary, accurate scintigraphic estimation of the gastric emptying of a solid meal can be obtained with a sim-

Simplified Scintigraphy for Gastric Emptying Assessment

85

plified approach based on a power exponential model using only four scintigraphic scans over 3 h. The method compares favorably with the more labor intensive conventional approach. The adoption of standardized methods would facilitate comparability of results between institutions and the conduct of multicenter trials. There are, however, significant differences between men and women and between younger and older individuals. There may be ethnic or cultural factors involved in the differences observed between individuals originating from French- and Englishspeaking Canada, although this original observation remains to be confirmed.

ACKNOWLEDGMENT This study was supported in part through an unrestricted grant from Astra Ha¨ssle AB, Sweden. Reprint requests and correspondence: Gervais Tougas, M.D.C.M., F.R.C.P.C., Division of Gastroenterology, Digestive Disease Research Unit, Room 3N5D, McMaster University Medical Centre, 1200 Main Street West, Hamilton, ON, Canada L8N 3Z5. Received Apr. 20, 1999; accepted Sep. 7, 1999.

REFERENCES 1. Moore JG. Gastroparesis: Pathogenesis and evaluation. In: Champion MC, Orr WC, ed. Evolving concepts in GI motility, 1st ed. Oxford: Blackwell Science, 1996:87–107. 2. McCallum RW, Berkowitz DM, Lerner E. Gastric emptying in patients with gastroesophageal reflux. Gastroenterology 1981; 80:285–91. 3. Malagelada J-R. GI motor disturbances in functional dyspepsia. Scand J Gastroenterol 1991;26(suppl)182:29 –32. 4. Stanghellini V, Tosetti C, Paternico` A, et al. Risk indicators of delayed gastric emptying of solids in patients with functional dyspepsia. Gastroenterology 1996;110:1036 – 42. 5. House A, Champion MC, Chamberlain M. National survey of radionuclide gastric emptying studies. Can J Gastroenterol 1997;11:317–21. 6. Lin HC, Hasler WL. Disorders of gastric emptying. In: Yamada T, ed. Textbook of gastroenterology, 2nd ed. Philadelphia: JB Lippincott, 1995:1318 – 46. 7. Horowitz M, Wishart JM, Jones KL, et al. Gastric emptying in diabetes: An overview. Diabet Med 1996;13:S16-22 8. Malagelada JR. The stomach: Clinical management. In: Phillips SF, Wingate DL, ed. Functional disorders of the gut, 1st ed. London: Churchill Livingstone, 1998:215–39. 9. Moore JG, Christian PE, Taylor AT, et al. Gastric emptying measurements: Delayed and complex emptying patterns without appropriate correction. J Nucl Med 1985;26:1206 –10. 10. Katz N, Toney MO, Heironimus JDI, et al. Gastric emptying: comparison of anterior only and geometric mean correction methods employing static and dynamic imaging. Clin Nucl Med 1994;19:396 – 400. 11. Christian PE, Datz FL, Moore JG. Confirmation of short solid-food lag phase by continuous monitoring of gastric emptying. J Nucl Med 1991;32:1349 –52. 12. Elashoff JD, Reedy TJ, Meyer JH. Analysis of gastric emptying data. Gastroenterology 1982;83:1306 –12. 13. Dugas MC, Schade RR, Lhotsky D, et al. Comparison of

86

14. 15. 16. 17. 18. 19.

20. 21.

Tougas et al.

methods for analyzing gastric isotopic gastric emptying. Am J Physiol 1982;243:G237-42. Christian PE, Datz FL, Sorenson JA, et al. Technical factors in gastric emptying studies. J Nucl Med 1983;24:264 – 8. Christian PE, Datz FL, Moore JG. Technical considerations in radionuclide gastric emptying studies. J Nucl Med Tech 1987; 15:200 –7. Camilleri M, Zinsmeister AR. Toward a relatively inexpensive, noninvasive, accurate test for colonic motility disorders. Gastroenterology 1992;103:36 – 42. Camilleri M, Zinsmeister AR, Greydanus MP, et al. Toward a less costly but accurate test of gastric emptying and small bowel transit. Dig Dis Sci 1991;36:609 –15. Gilja OH, Hausken T, Wilhelmsen I, et al. Impaired accommodation of proximal stomach to a meal in functional dyspepsia. Dig Dis Sci 1996;41:689 –96. Heading RC, Tothill P, McLoughlin GP, et al. Gastric emptying rate measurement in man. A double isotope scanning technique for simultaneous study of liquid and solid components of a meal. Gastroenterology 1976;71:45–50. Poitras P, Picard M, De´ry R, et al. Evaluation of gastric emptying function in clinical practice. Dig Dis Sci 1997;42: 2183–9. McMaster University, University of Kansas, University of Nebraska, Digestive Disease Specialists, University of Tennessee, University of Go¨teborg, University of Florida, Hershey Medical Center, Albert Einstein College of Medicine, University of Bologna, Bakken Research Center. Gastric emptying of a low fat meal: Establishment of international control values. Gastroenterology 1998;114:A755 (abstract).

AJG – Vol. 95, No. 1, 2000

22. Datz FL, Christian PE, Moore J. Gender-related differences in gastric emptying. J N Med 1987;28:1204 –7. 23. Hutson WR, Roehrkasse RL, Wald A. Influence of gender and menopause on gastric emptying and motility. Gastroenterology 1989;96:11–7. 24. Moore JG, Tweedy C, Christian PE, et al. Effect of age on gastric emptying of liquid-solid meals in man. Dig Dis Sci 1983;28:340 – 4. 25. Kao CH, Lai TL, Wang SJ, et al. Influence of age on gastric emptying in healthy Chinese. Clin Nucl Med 1994;19:401– 4. 26. Horowitz M, Maddern GJ, Chatterton BE, et al. Changes in gastric emptying rates with age. Clin Sci 1984;67:213– 8. 27. Madsen JL. Effects of gender, age, and body mass index on GI transit times. Dig Dis Sci 1992;37:1548 –53. 28. Gryba¨ck P, Naslund E, Hellstrom PM, et al. Gastric emptying of solids in humans: Improved evaluation by Kaplan-Meier plots with special reference to obesity and gender. Eur J Nucl Med 1996;23:1562–7. 29. Hermansson G, Sivertsson R. Gender-related differences in gastric emptying rate of solid meals. Dig Dis Sci 1996;41: 1994 – 8. 30. Bennink R, Peeters M, Van den Maegdenbergh V, et al. Comparison of total and compartmental gastric emptying and antral motility between healthy men and women. Eur J Nucl Med 1998;25:1293–9. 31. Knight LC, Parkman HP, Brown KL, et al. Delayed gastric emptying and decreased antral contractility in normal premenopausal women compared with men. Am J Gastroenterol 1997;92:968 –75.