Solid phase gastric emptying times in the dog measured by 13C-sodium-acetate breath test and 99mTechnetium radioscintigraphy

Summary Objective: The aim of the study was to assess solid gastric emptying via non-invasive 13C-sodium acetate breath test and compare this technique to 99mTechnetium scintigraphy in 12 healthy adult dogs. Material and methods: The dogs were fed a test meal containing either 100 mg 13C-sodium acetate or 150–250 MBq 99mTechnetium albumin colloid. Breath test and scintigraphy were performed on two consecutive days; this set of procedures was repeated in all dogs. Breath samples and scintigrams were obtained at baseline and every 15 minutes for 4 hours, then every 30 minutes for another 2 hours. 25%, 50% and 75% gastric emptying times for breath test (Gt25%b, Gt50%b, Gt75%b) and scintigraphy (Gt25%s, Gt50%s, Gt75%s) were calculated and compared. Results: The mean (± SD) Gt50%b and Gt50%s were 165 (± 28.1) and 71 (± 16.6) minutes, respectively. There was a significant correlation at all three gastric emptying times between breath test and scintigraphy. Conclusion and clinical significance: While gastric emptying times between both methods varied considerably, both methods correlated significantly showing that the gastric emptying breath test can be used to assess gastric emptying times in dogs.

Gastric emptying is also known to change depending on a variety of endocrine, exocrine, neurogenic and dietetic influences. In the dog, idiopathic gastric dilatation and volvolus (GDV) plays an important role, as this disease occurs frequently in large and giant breeds and is associated with high mortality (around 30%) (12).
However, it remains unclear, if impaired gastric emptying is a cause or consequence of GDV, especially as abdominal surgery itself can cause disturbance of gastric motility (19). Other diseases associated with delayed gastric emptying are only recognized in a small number of patients [for example pyloric hypertrophy (36), gastric ulcerations (10), fibrosing leiomyositis (21), dysautonomia (37)] or available data are derived from experimental studies [e. g. induced endotoxaemia (7) and experimental diabetes mellitus (33)]. In addition, the dog is commonly used as an animal model to assess drug efficacy on gastric emptying and to evaluate the influence of various physiological, environmental and pathological parameters (33) or surgical procedures (1). The assessment of gastric emptying is therefore important.
Unfortunately, radioscintigraphy, the gold standard method to measure gastric emptying in humans and animals is not widely available due to scarcity of equipment and usage of radioactivity. Easier to use non-invasive methods, such as 13 C-breath tests have been developed for humans (11) and applied in a wide variety of situations (5,15). The same approach has been used in several animal species (39). So far 13 C-breath tests have been compared extensively to scintigraphy in humans (32,40) and in a single study in horses (31), but not in the dog. As a tracer, 13 C-octanoic acid has been used most frequently. However, this substance has several limitations in animals, e. g. its unpleasant taste. Since it is a fatty acid, some influence on gastric emptying cannot be fully excluded (30). In a very recent study, in which different compositions of the test meal were administered with different concentrations of octanoic acid, it could be shown that there is a dose-dependent delay to 13 CO 2 elimination when the rate of 13 CO 2 entering the body was higher than the respiratory elimination capacity (13). Other markers, such as 13 C-sodium acetate, may lack these problems and have been used successfully in humans (17).
The aim of this study was thus to establish the 13 C-sodium-acetate breath test ( 13 C-SABT) to measure solid phase gastric emptying times in dogs and compare the results with 99m Tc radioscintigraphy.

Animals
Twelve clinically healthy dogs privately owned by personnel of the small animal clinic were included in the study. Written owner consent as well as governmental ethical approval was obtained. There were three female intact, five female spayed, three male neutered dogs and one male intact dog. The dogs had a median body weight of 19.7 kg (range 9.2-38.1 kg) and a median age of 6.5 years (range 1-12 years). Most dogs (n = 6) were of mixed breed, other breeds represented included one of the following: Beagle, Doberman Pinscher, Fox Terrier, Labrador Retriever, Old German Shepherd and West Highland White Terrier.
Exclusion criteria were medication during the preceding 12 months (except immunization, flea products or anthelmintics) and special dietary requirements. The dogs were treated with fenbendazole (50 mg/kg over 5 days) 1-3 weeks prior to the first test. Basal and 2 hour post prandial serum bile acid concentrations after ingestion of the breath test meal had to be normal to ensure normal hepatic function.

Diet
The test meal consisted of commercially available canned dog food 1 with about 50 kcal ME/kg BW 0,75 . Prior to feeding the diet was labelled with either 100 mg of 13 C-sodium acetate 2 for breath test analysis or 150 to 250 MBq of 99m Tc 3 mixed with an albumin colloid solution 4 for scintigraphy. The test meal was blenderized for 3 minutes. For scintigraphy, homogenous spread of the tracer was assured for every test meal by placing the food on the gamma camera before offering it to the test dogs (data not shown).

C sodium acetate breath test
The first breath sample was collected immediately prior to the ingestion of the test meal (baseline sample). Afterwards, breath samples were collected every 15 minutes for 4 hours and then every 30 minutes for additional 2 hours. Samples were collected with a commercially available anaesthetic face mask for small animals that fitted tightly around the dog's muzzle. The mask was connected to a one-way valve and attached to a breath reservoir bag 5 . Dogs breathed normally until the reservoir bag was filled with exhaled air. Upon filling, the bag was immediately sealed, and breath samples were analyzed within half an hour.
The 12 CO 2 / 13 CO 2 ratio was measured for each breath sample via non-dispersive infrared spectroscopy in an automated analyser 6 ; the difference of the 12 CO 2 / 13 CO 2 ratio for each sample in comparison to the baseline value was expressed as delta over baseline (DOB). The area under the curve (AUC) of the DOB values for every single breath test was calculated. Percentiles of 25%, 50% and 75% (= quartiles) of the total AUC were calculated and the respective Gt 25%b , Gt 50%b and Gt 75%b determined. tral images were taken immediately after the intake of the test meal and then every 15 minutes for 4 hours and every 30 minutes for the following 2-4 hours. If the stomach still contained radioactivity after 6 hours, a final image was taken after a total of 8 hours post feeding. If the stomach appeared completely empty at an earlier time point, image acquisition was aborted. Scintigramms were acquired in 256 × 256 pixel matrices. Pixel size was 2.27 mm, resulting in a field of view of 540 mm. Acquisition time was 60 seconds for each image during which the animal was not allowed to change position. If the dog moved during the imaging sequence another picture was obtained immediately afterwards.
The area of the stomach (region of interest) was manually drawn on the scintigraphic images and all pixels within this region were counted automatically. Subsequently, the decay-corrected radioactive counts were plotted against time. Calculation of Gt 25%s , Gt 50%s and Gt 75%s was done by calculating the AUC for every single test equally to the breath test analysis.

Experimental protocol
Following an overnight fast, the test meal containing either 13 C-sodium acetate or 99m Tc colloid was given to the dogs at 8 am. The dogs had free access to water. All but three dogs ingested the test meals rapidly and completely; in these three dogs the blenderized food was given by syringe.
In all dogs, scintigraphic and breath test measurements were done twice on consecutive days in an alternating order. The gap between these two sets of measurements varied between 2 weeks and 3 months.

Statistical analysis
Statistical analysis as well as the estimation of the quartiles of the total AUC and the corresponding emptying times, representing non-parametric estimation methods, were done with the statistical program package BMDP © (Statistical Solutions 2008). Because the DOB-values and the radioactivity were nearly log-normal distributed, they were described by geometric mean and geometric standard deviation. Within-dog and between-dog variation were calculated as standard deviation by a hierarchical two way analysis of variance (ANOVA) with random effects. Correlation between mean gastric emptying times of scintigraphy and breath test was calculated using the product moment correlation coefficient and tested by the adequate t-statistic. A p-value less or equal 0.05 was considered to show statistical significance.

Breath test
In all dogs a rapid increase of DOB in expired air was noticed during the first hour of breath analysis (ǠFig. 1). At the end of the col-lection period (360 minutes in five and 480 minutes in 19 tests) basal levels of DOB were reached in all but one dog. In this dog the DOB level after 480 minutes was still nearly 50% of the peak value. No further breath test sample was obtained in this dog. Mean (± SD) Gt 25%b, Gt 50%b and Gt 75%b were 104 (± 18.9), 165 (± 28.1) and 232 (± 36.1) minutes, respectively (ǠFig. 2). The coefficients of variation between dogs ranged between 8.9 and 14% and within dogs between 11.2 and 12.9%.

Scintigraphy
Overall, gastric emptying showed two different patterns. In three of 24 scintigraphies a linear pattern could be observed. This was preceded in one case by a lag phase of 15 minutes. In 21 scintigraphies a sigmoidal shape of the gastric emptying graph was seen (ǠFig. 3); 18 of them were preceded by a lag phase of 15 to 108 minutes (median 41 minutes). In two scintigraphies the occurrence of a lag phase could not be assessed because the graphs showed an irregular course. The gastric emptying pattern was similar between the two sets of scintigraphic studies in eight dogs.

Comparison of both methods
The correlation coefficients between breath test and scintigraphy were statistically significant at all time points (ǠTable 1).

Discussion
This is the first study assessing canine gastric emptying with 13 C-SABT. So far, breath tests to assess canine gastric emptying have not been compared to scintigraphy. In humans, the 13 C-SABT has been used successfully to measure gastric emptying (17). In the present study, collection of breath samples was done up to 6 hours post feeding, because this has shown to produce a better accuracy than collecting for shorter periods of time (3). Sample intervals were relatively short, but might be prolonged in the future to make the test easier to perform. The NDIRS-method for measuring the 12 C/ 13 C-ratio in canine breath is simple and has been shown to produce reliable and reproducible results (24). Unfortunately, high levels of exhaled CO 2 are necessary to analyse the samples. Thus, alternative measuring techniques, for example Laser Assisted Ratio Analyzer System, may be more useful in panting dogs or in smaller animal species.
The median G 50%b was 104 minutes, a time similar to published results (4,39). Unfortunately, there was a wide range from 32 to 312 minutes; however, the within-and between-dog variation was relatively low, indicating acceptable reproducibility. Only the absolute variation within dogs at 75% gastric emptying was higher (30.0 minutes). Interestingly, others have found high variability of gastric emptying as well, both between dogs as well as in the same dog when analysed repetitively (28). As such, it is difficult to assess canine solid phase gastric emptying via breath test analysis.
Gastric radioscintigraphy is -since its first use in 1966 (16)considered the gold standard method to assess gastric emptying. In human medicine, all new methods should be compared with scintigraphy (25). The Society of Nuclear Medicine defined a standard protocol, and reference ranges based on large study groups are available (8). This is not the case in veterinary medicine and no standard protocols are recognized. Nonetheless, scintigraphy has been used to assess gastric emptying in various animal species, including dogs, cats, horses, rats, monkeys and pigs (38). 99m Technetium (Tc) has been used most frequently in animals [bound to pertechnetate (18), DPTA (6), sulphur colloid (18), tin (9), albumin colloid (38), disofenin (14), mebrofenin (2) or resin beads (20)]. Although some authors recommend the usage of mebrofenin or disofenin in small animals based on efficiency-studies (18), the present study indicates that 99m Tc albumin colloid is also acceptable. It disperses homogeneously in the test meal as was shown in every single meal before its ingestion (data not shown).
Individual variation between scintigraphic measurements was relatively large in the examined group of dogs, giving a coefficient of variation of about 22%. There was no significant difference in this variation between the different emptying quartiles. There are Mean (± SD) gastric emptying times at 25, 50 and 75% emptying measured by 13 C-sodium acetate breath test and 99m Tc radioscintigraphy in 12 dogs  several possible reasons for this bigger coefficient of variation compared to the breath test. Peristaltic movement of the food in the stomach during early digestion can increase the measured radioactivity. When the gastric content moves closer to the camera, gastric emptying may be underestimated. This effect could be observed in the majority of dogs in this study and is also seen in human patients (27). In later phases of digestion, superimposition of the region of interest (stomach) with the small intestines or colon can lead to difficulties in reading the scintigramms accurately. Using not only ventrodorsal but also lateral imaging might circumvent some of these problems. However, superimposition of the upper small intestine or the colon over the stomach most likely nullifies this advantage. Thus, scintigraphy itself has its limitations and might not be as useful in dogs as has been postulated. In addition, the clinical application of scintigraphy in animals will always be limited by the radiation hazard and the availability and expense of this method. But, because of the limited numbers of dogs used in the present study, no final statement about the usefulness of scintigraphy is possible and further studies are necessary to approve or confute these results.
Whilst absolute gastric emptying times vary markedly between animals, correlation of mean gastric emptying times between scintigraphy and breath test was good at all time points. However, reference values depend heavily on the method employed, since scintigraphic gastric emptying times are less than half the time compared to breath test results. This is probably due to the fact that scintigraphy is a more direct measurement than the breath test, where metabolism of the tracer substance is needed. This theory is underlined by several publications from human studies, where similar problems are encountered and the fact if breath test and scintigraphy can be compared appropriately at all has been discussed vividly in recent years (35). These problems may be exaggerated in a clinical setting, when gastric emptying time of an individual dog is assessed; as it could be shown that individual variations can be marked (withindog variance) not only between gastric emptying times, but also between emptying patterns. This is contradictory to other studies, where minimal intraindividual changes were found (23). It is still possible that these changes are linked to breed, sex, body weight or other unknown factors, but in the present study the number of dogs was too small to correlate emptying values to these parameters. In other studies however, correlation of body weight with gastric emptying was not a consistent finding (4).
Feeding increases total body CO 2 (26) and thus might have an effect on the 13 C-SABT. On the other hand, respiratory problems or acid-base disturbances might have an effect on CO 2 -excretion as well. While all dogs were clinically healthy, they were given about half their daily caloric requirements as test meal and this might affect gastric emptying times. Firstly, since a 12 CO 2 / 13 CO 2 ratio was measured, increased CO 2 levels will unlikely be affected and secondly, absolute emptying times can only be compared between identical methods and gastric emptying calculations. Therefore increased or decreased CO 2 -levels seem of minor consequence when using the 13 C-SABT. It is a limitation of the present study that scintigraphic and breath test measurements were not performed simultaneously, which would theoretically minimize the influence of day-to-day variance in gastric emptying. However, in previous studies it could be shown that results of both methods are reproducible even if performed on separate days (27) and thus can still be interpreted as adequate.
In human medicine, the use of correction factors and the right way of calculating comparable values is discussed vividly. For example a discussion arose if the "corrected" half emptying time used by the first study describing gastric emptying breath tests in humans (11) is a useful parameter at all (29). This study (11) suggested a linear-regression based approach and a significant correlation between half-emptying times measured via breath test and scintigraphy could be found, which has since then be demonstrated in other studies as well (40). Recently, two new mathematical models to achieve a better comparability between scintigraphy and breath test were presented for humans (3), and were also proposed for the dog (34). To extrapolate other mathematical models used in human medicine to explain variations between the two methods is difficult, as they mostly rely on the different metabolism of octanoic acid, which was not used in this study (22). At this moment multiple generalized linear models have been reported in humans and it is difficult to select the most appropriate for any species (35).
Finding a right mathematical model in the present study was a challenge, and several other models (i. e. nonlinear regression) failed to give adequate results. In addition, between-dog variance seems to be even larger than within-dog variation. This makes interpretation of gastric emptying times in an individual dog in a clinical setting extremely difficult. Further studies with larger study groups are necessary to assess accuracy of the available methods and mathematical models in the dog and to identify causes for the individual variations. In addition, correlation of both methods in dogs where gastric emptying times are delayed due to different underlying diseases have to be evaluated, as correlation may either improve or deteriorate in pathophysiological conditions. A different mathematical analysis may be developed in the future so that a correlation between breath test and scintigraphy can be measured more accurately.

Conclusion for practice
In conclusion, 13 C-SABT can be used to measure gastric emptying times in the dog. As indicated by other studies, gastric half emptying time (Gt 50% ) seems to be the most reliable parameter. 25% emptying is probably too early in the emptying process to be of value in diagnosing delayed gastric empting, and individual variation may increase at later time points (75% emptying) when using breath test analysis. However, individual variations do occur and have to be taken into account when interpreting gastric emptying results. Inferring results of this study to a clinical situation may be challenging, and further research in this field is necessary before the breath test can be used in a wider clinical context to assess gastric emptying of individual patients.