University of Birmingham In vivo models to evaluate ingestible devices

Evaluation of orally ingestible devices is critical to optimize their performance early in development. Using animals as a pre-clinical tool can provide useful information on functionality, yet it is important to recognize that animal gastrointestinal physiology, pathophysiology and anatomy can differ to that in humans and that the most suitable species needs to be selected to inform the evaluation. There has been a move towards in vitro and in silico models rather than animal models in line with the 3Rs (Replacement, Reduction and Reﬁnement) as well as the better control and reproducibility associated with these systems. However, there are still instances where animal models provide the greatest understanding. This paper provides an overview of key aspects of human gastrointestinal anatomy and physiology and compares parameters to those reported in animal species. The value of each species can be determined based upon the parameter of interest from the ingested device when considering the use of pre-clinical animal testing.


Introduction
Animals are used as pre-clinical models to evaluate the therapeutic index, toxicity and linear/non-linear pharmacokinetics of a drug, guiding human dose and formulation development. In general, small animals (rats, mice, guinea pigs and rabbits) are suitable for determining bioavailability and the mechanism of drug absorption, typically, from suspensions or solutions, while larger animals (dogs, pigs and monkeys) are used to evaluate absorption from formulations.
However, due to anatomical, physiological and metabolic differences between animals and humans, no single animal species provides an adequate physiological paradigm for the human environment [1]. Instead, the selection of the appropriate animal model used to extrapolate drug behaviour to humans depends on the mechanism of absorption and/or metabolism of the drug. Hence, microbiota composition will be the main factor to be considered when an animal model is selected to assess drug metabolism, including potential enterohepatic circulation of its metabolites, and/or the performance of microbial-triggered dosage forms. However, the animal model might differ when transit time is the important factor for an extended/sustained release formulation. This is also applied when the osmotic pressure of the lumen is the main factor that will affect the performance of an osmotic pump formulation.
A similar approach should be followed for the development of ingestible devices. For instance, considering the current size of ingestible devices, small animals (for example, rats) are unsuitable. The diameter of the small intestine in rats is 0.3-0.5 cm [2], whereas the diameter of different ingestible devices ranges between 1.08 and 1.3 cm [3]. In addition to any difficulty in swallowing the device or fatal damage to the GI tract of the animal due to such a large size, the device will alter motility and hence transit times and not be representative of the human in vivo environment. However, advances in technology permit smaller devices and as such small animals may have value in their evaluation [4]. Currently, wireless capsule technology has been validated for gastric and small intestinal transit times in dogs [5][6][7][8], pigs [9], rabbits [10] and horses [11], but only single studies have used this technology for determining colonic transit times in dogs [6] and rabbits [10].
When considering ingestible devices for screening and diagnosis, the animal model suitable for developing the proper wireless device will be based on similarities in disease pathophysiology to human beings. The use of the most suitable model is critical to extrapolate the pre-clinical data to inform the design of the device and to enable translation into clinical use. Key factors to consider in the applicability of animal models in the evaluation of ingestible devices for screening and diagnosis include the anatomical shape of the organ, the motility and the transit time.
This review will describe the latest information on human anatomy and physiology as the basis against which the animal models are compared. The paper is subdivided into sections to represent the stomach, small intestine and large intestine.

The stomach
The stomach is the first region within the GI tract where the ingested device will reside for a period of time. An understanding of the gastric environment is important to consider how the fluid volume; composition; motility and transit may affect the device under test.

Human gastric volume
The fluid volume within the stomach has previously been reviewed [12], however the total volume of the stomach or gastric capacity is different to the residual fluid within the human stomach. A summary of data on gastric capacity measurements in healthy adults is presented in Table 1.
There is some controversy in the literature about whether BMI is linked to gastric capacity; in some studies the correlation is significant yet the Pearson correlation coefficient was <0.3 [13].

Animal stomach anatomy
When evaluating an ingestible device, the macroenvironment within the gastrointestinal tract is of importance; this can be related to the physical dimensions of the product under test to ensure that this can be accommodated within the stomach of the animal. The amount of fluid as well as total stomach capacity can be of interest to determine the dosing conditions for the ingested device. Other features of the stomach are of importance depending upon the function of the device under test, for example, mucus thickness. The relative shape of the stomach as well as the dimensions of the pyloric sphincter will affect the transit through the stomach of the animal in comparison to a human. Much of this data is summarized in Table 2 which compares the gastric anatomical and physiological conditions in relevant animal species to the human adult stomach.

Gastric contents
The composition of human gastric fluids has previously been reviewed [12], however further detail on the gastric pH is of relevance here as this can be a critical factor to replicate when evaluating ingestible devices; particularly those that may be sensitive to prolonged exposure at low pH.

Human gastric pH
The harsh acidic conditions within the stomach are well known. There is extensive data on gastric pH throughout the literature. Table 3 shows the mean (and standard deviation of) gastric pH reported previously for both fed and fasted participants. The methodology used to assess pH and calorific load of the food ingested is also presented. The highest pH reported is in the fed state; typically, the peak pH is early in the post ingestion phase. As well as maximum pH values, the dynamic change in pH is of interest in simulating the stomach as well as regional differences in pH values. Often the percentage of time where the pH is greater than 4 is reported as this links to control in gastric reflux where a target for therapy is gastric pH > 4 [23]. Despite average pH values being recorded it is known that there are ''spikes" in pH that occur and that the pH is not constant within the stomach; it is also known that there may be regions within the stomach where pH is lower due to the proximity to sources of acid secretion.

Gastric pH in animals
When comparing the gastric pH between humans and animals it is important to consider both mean and time-based profiles of Table 2 Summary of anatomical and physiological aspects of animal and human stomach. Images are adapted from [1]; not to scale.

Species
Fasted fluid volume (mL) Capacity [1,17] Mucus thickness (um) Pyloric sphincter diameter (mm) Shape Mouse 0.09 ± 0.07 [22] 0.37 [17] 0.25 [19] Guinea pig 234-473 [17] 5 [19] Table 3 Mean and standard deviation pH values recorded in healthy adults from the literature. Table adapted and updated from [39]. *Mean and standard deviation (SD) were estimated and reported from median and interquartile range (IQR) or from median and range, original calculations performed by [39]. pH in both a fasted and fed state. As stated for the human data, the gastric pH is not constant but is subject to some variation and spikes in pH values. Similar profiles have been reported in animal models, of which the pH data is presented in Table 4. The table has included pH measurements collected using a range of devices, including ingestible devices; the data in this table can be compared to the data from humans presented in Table 3. A comparison of the gastric pH in humans and dogs was presented by Koziolek et al (2019) following ingestion of a meal. It was noted that the timing of the meal relative to the medicine is important when matching the pH profile in the stomach in dogs compared to humans to provide the best correlation [49]. The pig model showed duodeno-gastric reflux prior to gastric emptying where there were spikes of higher pH within the stomach in the three hours before gastric emptying [42]; this coupled with the much longer gastric residence time in pigs suggests that a dog is a better model of the gastric environment.
A comparison of pH data in the stomach of animals is presented in Figs. 1 and 2. Due to many papers showing total GI pH, it can be complicated to extract the actual gastric pH from certain studies. Hence, only data from pigs and dogs are available for the fasted state; however, the range of human values was presented in Table 3.
The variability in gastric pH is large in pigs and dogs yet this is likely to reflect the conditions in humans. In the fed state it should be noted that the gastric pH returns to low levels in pigs; this may be an artefact of the normalised x axis for time as gastric emptying was significantly longer in pigs compared to either humans or dogs.
Gastric pH is subject to variability influenced by diet and it is possible to use food or drugs to alter the pH within the stomach of animals and humans if required.

Gastric emptying
There is a significant difference between the emptying of liquid food and the emptying of solid food, and the emptying rate of liquids is faster than that of solids. The time taken to empty 50% of the   ingested contents (t 1/2 ) has often been used to describe gastric emptying rate for the purpose of comparison.

Gastric liquid emptying rate in humans
Crude measurements of gastric emptying were conducted in two males in 1977 by aspiration of the residual volume in the stomach following administration of a 250 mL volume via an intragastric tube [51]. The relative rate of emptying of water was compared to that of a suspension of barium sulphate in two healthy individuals where both fluids emptied at a similar rate; just 10% of the ingested volume remained after 15 min [51]. However, it is recognised that the presence of the gastric tube may affect physiology and gastric emptying. The use of ultrasound to measure gastric emptying was reported as a non-invasive technique that was used to measure gastric emptying of 500 mL of orange cordial [52]. Magnetic resonance imaging (MRI) was first reported as a tool to measure gastric emptying in 1992 [53]. The study used a liquid test meal consisting of 500 mL of a 10% glucose solution containing contrast medium and 99m Tc-DTPA to compare the MRI data to the detection of the fluid in the duodenum via a luminal duodenal tube [53].
Gastric emptying in relation to drug absorption has often considered in clinical scenarios where medicines are taken with a glass of water; this may be more relevant when considering ingestible medical devices. The gastric emptying time (gastric half-life) for a 240 mL volume of water when ingested following a 10 h fast was reported to be 13.1 ± 1 min [54]. There was an initial increase to 242 mL after 2 min, but this declined rapidly to less than 50 mL by 45 min [54]. This finding was consistent with another study where the initial volume rose to 270 ± 20 mL and subsequently decreased over time and at 90 min, when a second 240 mL volume of water was ingested, volume rose again to 251 ± 38 mL [55]. The gastric emptying rate constant for water was found to be between 0.10 and 0.15 min À1 [55]. Gastric volumes and gastric emptying rates of water when administered 30 min after three different meals were compared in a single center, 4-way crossover study [55]. A fasted control group was compared to groups fed a highfat solid meal, a high-fat homogenous meal and a low-fat solid meal. The calorific content of the high-and low-fat meals was matched [55]. Shortly after consumption, gastric volumes were recorded to be 434 ± 31 mL after the high-fat meal, 336 ± 17 mL after the high-fat homogeneous meal, and 546 ± 48 mL after the low-fat solid meal [55]. The gastric emptying rate constants of water (k e ) were significantly higher in the case of the solid meals (>1.5 min À1 ) compared to the semisolid homogeneous meal (<0.1 min À1 ) [55]. This study suggests that the presence of food affects the gastric emptying of liquids and confirms the existence of the Magenstrasse (stomach road) phenomenon for three different meals [55]. Fig. 3 compares the gastric emptying rate of liquids in human from published data, highlighting the variability in gastric emptying and how this can depend upon the composition and volume of fluid ingested, as well as the technique used for the measurement.

Gastric solid emptying rate in humans
Gastric emptying of solids shows a biphasic pattern. In the first phase almost nothing is emptied, it is the delayed phase, where the solids are manipulated to form smaller particles and moved to the distal part of the stomach. The second phase is the linear emptying phase through the pylorus where small particles (less than 2-3 mm in size) are emptied in the digestive period that lasts 2-3 h after a meal. However, the stomach retains larger food particles that escape mincing during the digestive period, and then forcefully empties them into the small bowel during the interdigestive period.
Gastric emptying of four homogenised meals (500 mL volume) of varying viscosity and nutrient content was evaluated using MRI [57] where it was found that increasing the nutrient content delayed gastric emptying, whereas increasing viscosity had a smaller effect. This was in agreement with a previous study that reported that the rate of emptying of food materials was based on calorific load with emptying controlled at rate (1-4 kCal min À1 ) [58].

Gastric transit time in humans
Gastric transit time is influenced by many factors including fed status, caloric content of meal as well as measurement technique [59]. A summary of gastric transit times as a function of fasted / fed state and measurement technique for a range of oral solid dosage forms is presented in Table 5. The use of the solid dosage forms together with their size and density is of interest for ingested medical devices.

Gastric residence time in animals
A comparative gastric residence time can be of great importance in the evaluation of ingestible devices. It has been reported that the gastric residence time in pigs far exceeds that reported in adults and as such the power supply for ingested devices can limit their functionality when evaluated in pig models. Table 6 shows the residence time of solid units as measured in animal models to provide an indication of the relative gastric residence time.
The size shape and density of the ingested material has been shown to influence the gastric emptying time in rats [85]; dogs [86] and humans [20]. The relative gastric transit needs to be considered when using an animal to evaluate an ingested device.

Pyloric sphincter in animals
The pyloric sphincter can control transit from the stomach to the small intestine, in humans it has been reported that particles larger than 2 mm diameter are unable to pass through the pyloric sphincter where the diameter is 2.5-5 cm [19]. The diameter of the pyloric sphincter in a range of animals has been reported with values of 1-2 cm for a Beagle dog; 8 mm for a rabbit; 5 mm for a guinea pig and estimated values of 0.25 mm for a mouse and 2 mm for a rat [19].

Gastric pressures/motility
Normal motility in the human adult stomach differs in the fasted and fed state to account for the function of the stomach. There is also a difference in motility within the stomach depending upon the anatomy, the proximal part acts as a reservoir and the antral part as a pump. The reservoir part of the stomach can expand upon the ingestion of food without a significant increase in intraluminal pressure. The antral part of the stomach exhibits phasic contractions to mix and breakdown food prior to passage to the small intestine via the pyloric sphincter.
The migrating motor complex (MMC) describes the contractile activity in the stomach (and remainder of the intestine). The total duration of the MMC in humans has been reported to be 172 ± 76 min [88]. The MMC consists of three phases: phase I is a period of no contractions; phase II is irregular contractions and phase III is regular high amplitude contractions; defined as regular contractions for at least 2 min with 2-3 contractions per minute [89]. The ingestion of food can interrupt the MMC cycle.
The dominant phase of the MMC in the stomach was phase I which was present for more than half of the duration of the MMC [88]. The duration of phase III activity in the stomach was reported to be 3.6 ± 1.1 min and 3.8 ± 1.7 min in the proximal and distal positions within the stomach in fasted adults with mean frequencies of 2.7 ± 0.3 and 2.9 ± 0.2 respectively [88]. The mean amplitude of a phase III contraction is 75 mg Hg in the stomach [89].

Gastric pressures and motility in animals
Pressures reported in the stomach have included values of up to 800 mbar in dogs (typical range 500-800 mbar) [49]; were up to 402 mbar in the stomach with an average of 349 ± 84 mbar under fasted conditions, and 250 ± 103 mbar in the fed state in pigs (Landrace) [42].
The strength of gastric contractions in dogs measured using an ingested sensor reported maximum amplitude contractions ranging from 13 to 94 mmHg with a frequency of 3.7 contractions per minute [8].
The gastric emptying force has been compared in dogs and humans using pressure sensitive tablets where the area areanormalized gastric emptying force (dynes cm À2 ) was 606 and 3858 in fasted humans and dogs with values of 962 and 3639 dynes cm À2 for fed humans and dogs respectively [90]. Other work has compared the mechanical destructive force in the stomach of dogs and humans with values of 3.2 N and 1.9 N being reported respectively [91].
Further work is required to compare data on gastric pressures and motility in animal species and humans to better understand   the similarities and differences to enable selection of the most appropriate species.

The small intestine (SI)
The human small intestine extends from the pylorus at the exit of the stomach to the beginning of the colon. The duodenum is the first part of the small intestine, with an average length of 25 cm, approximately in line with the length of its Latin definition ''12 fingers" [92]. The jejunum and ileum make up the remainder of the small intestine and are reported to measure 5-8 m in length, where the jejunum represents the proximal 40% and the ileum the distal 60% [93,94]. However, the length of the small intestine varies with the degree of longitudinal smooth muscle contraction and the method used for assessment [95,96]. A comparison of the dimensions of the small intestine between humans and animals is provided in Table 7.
Arrangement is less meaningful for SI compared to the stomach as it is the overall length and diameter that are of interest. The length of the small intestine relative to the size of the animal has previously been explored to link the absorptive capacity to the nutrient requirements. The overall absorptive capacity of the small intestine is also a function of the surface area; the villi and microvilli in humans greatly extends the surface area compared to the basic anatomical length and diameter.

Small intestinal pH
There is extensive data on human small intestinal pH within the literature [59]. Table 8 shows the mean (and standard deviation) small intestinal pH reported previously in the literature for both fed and fasted participants. The methodology used to assess pH and calorific load of the food ingested is also presented. Often the trigger to signal that the ingested device is within the small intestine is a change in pH and the pH changes with location. As with the stomach the pH profile is not constant, fluctuating between pH6 to pH 7-8 [99] although fewer spikes in pH are observed in the small intestine compared to the stomach.
Data on small intestinal pH is complex as it relies on accurate measurement of transit from the stomach into the small intestine. Data is available on the pH values in the small intestine of pigs in both the fasted and fed state [42], however due to the long gastric emptying time of >8 h it is questionable whether the fed state affects small intestinal pH in pigs. Overall pH values ranged from pH 6-8 in the small intestine of a Landrace pig [42].
Small intestinal pH in animals has been measured using ingested devices; the transit from stomach to small intestine is often noted as a change in pH and thus the recording is initiated at this point. A summary of pH data reported from the small intestine of animals is presented in Table 9.    Table 8 Mean and standard deviation pH values recorded in healthy adults from the literature. Table adapted and updated from [17]. *Mean and standard deviation (SD) were estimated and reported from median and interquartile range (IQR) or from median and range, original calculations performed by [39]. The pH in the small intestine showed a narrower distribution compared to that in the stomach; most likely due the reflux that is observed in the stomach and not observed within the small intestine. Most of the animal data shows similar pH values to the human although the data from mice and rats is lower compared to other animals and humans.

Small intestinal transit time in humans
The transit through the small intestine can be complex to measure as invasive methods are required to define the start and end points. Techniques used to measure small intestinal transit range from imaging techniques to ingestion of devices that report the pH over time and can thus be used to measure the total small intestinal transit time. A summary of the literature reports on small intestinal transit time together with the method used to measure this in humans is presented in Table 10. The data shows the shortest transit times are less than 2 h and the longest are close to 5 h. The method used to measure transit as well as dimensions and density of the object under test can influence the transit through the small intestine.

Small intestinal transit time in animals
Data on small intestinal transit can be difficult to measure in small animals. A summary of the small intestinal transit times from a range of animals is presented in Table 11.

Small intestinal contents
The composition of human small intestinal fluids has previously been reviewed [12], however a comparison of key parameters to   [41] those identified in animal intestinal fluids is presented in Table 12.
It should be noted that there are simulated small intestinal fluids reported for rats [119]; dogs [120] and pigs [121]. A comparison of bile and phospholipid concentrations in the mid-jejunum of dogs and humans following a 500 mL meal of Ensure Plus Ò was 18 mM and 11.8 mM for the bile and 19.4 mM and 4.31 mM for phospholipids [123].

Small intestinal pressures/motility
The MMC cycle in humans usually starts in the stomach but there are reports that the phase III contractions can start within the proximal small intestine [88]. The velocity of propagation of phase III of the MMC was reported to decrease progressively from the proximal duodenum to the distal jejunum from 11.4 ± 4.7 in the duodenum to 9.3 ± 5 at the junction of the duodenum and jejunum and 7.4 ± 5.8 cm min À1 in the jejunum (in humans) [88]. The duration of the phase III activity increased from the proximal duodenum to the distal jejunum from 5.4 ± 3 to 9.2 ± 9 min yet the frequency decreased from 11.7 ± 0.5 to 10.6 ± 1 [88]. The mean amplitude of a phase III contraction is 33 mg Hg in the duodenum in humans [89].
Motility within the small intestine of animals has been measured using a range of techniques including air-filled and water filled catheter systems in dogs [125]; telemetric capsules [42,49]; ex vivo cannulated intestinal segments [126].
Propagation velocities of ileal contractions in dogs of 0.76 cm s À1 have been reported with the frequency of contractions being 17.8 contractions per minute [127].

The colon
The human colon is the last site of absorption for undigested food. It is comprised of the caecum as well as ascending, transverse, descending and sigmoid colon and lastly, the rectum. The human colon has four bends. The first is located between the ascending and transverse sections (the so-called hepatic flexure), secondly between the transverse and descending colon (termed the splenic flexure) and the remaining two are found at the beginning and the end of the sigmoid colon.
The main functions of the colon are the absorption of water, electrolytes and nutrients from partially digested food as well as storage and distal propulsion of faecal matter until defecation. Every region of the human colon has a different role, e.g., right colon (i.e. caecum, ascending and mid-transverse absorbs more water and electrolytes than the left colon (i.e. descending, sigmoid and anal canal) [131].

Colon anatomy
The colon anatomy in humans and animals is compared in detail in Table 13. This includes details on the dimensions and arrangement of the colon as well as the fluid volumes and composition within the colon.
The colons of rabbits and pigs share anatomical similarities with the human colon, in terms of haustra and semilunar folds, making these two species suitable to explore transit times in the colon [1,155]. However, there are still significant interspecies dif-ferences, e.g., pigs and humans, as pigs' colon constitutes a spiral arrangement and possesses a much larger cecum [155].

Colonic pressure/motility
Recent studies have used telemetric technology to obtain dynamic pH and pressure profiles in animals [42,49]. Aside from the anatomical similarities, pigs showed higher pressure amplitudes in fasted (Fig. 4A) and fed (Fig. 4B) state compared to humans. On the other hand, dogs showed close to human pressure profiles with low variability; probably this can be attributed to the short colon in dogs. It should be pointed out that the direct comparison of animals and humans should be performed with caution as the location of the SmartPill Ò capsule is unknown and hence the pressure and pH registration cannot be linked to a specific colonic region [69]. Although this technology provides information for the luminal environment of the whole colon, the colonic regional residence times are not known in order to understand for how long a dosage form or device will be exposed under specific conditions. For instance, studies using the magnetic pill as tracker of colonic motility in healthy humans showed longer transit times in the proximal part of the colon (ascending and transverse) compared to the distal part [156]. If this is the case for animals (there are no animal studies using magnetic capsule), then it is important to know the location of the wireless pH/pressure sensor to inform fields such as formulation development.
Pigs have shown a ''parabolic" pressure profile (Fig. 4) compared to the highly variable but fairly constant range of pressures across the human colon. In the human colon, pressure ranges between 1.1 and 231 mbar in fasted and 0.55-168 mbar in fed state. In the dog colon, pressure ranges between 0.5 and 98 mbar in fasted state. The pressure amplitudes in the pig colon range between 2.1 and 305 and 1-256 mbar in the fasted and fed state respectively. Thus, a dosage form targeting the colon will be exposed to a significantly heterogeneous environment in the pig colon, unless it spends most of the time within the proximal part in which the pressure amplitudes have the same range as humans (see data points for the pig that are lying within the first 0-0.1 normalised time (h) in Fig. 4, these show similarity to those in humans and dogs). However, this is more of a speculation rather than a solid assumption as knowledge of regional pressure amplitudes is required.

Colonic pH
With regards to pH profile, pigs showed high variability with values between pH 5.5 and pH 10 in fasted (Fig. 5A) and 5.4 to 8.2 in fed (Fig. 5B) state. Interestingly, pig colonic pH profile was highly variable in fasted but not that in fed which was exactly the opposite for humans. Dogs showed pH values between 5 and 8.3 in fasted state (Fig. 5A). In addition, dogs showed constant pH for relatively long periods which has not been observed in a distinctive manner in humans and pigs. This is probably related to the shorter canine colon, implying that the SmartPill Ò might have passed from the proximal colon to the rectum in quite a short time and remained there until defecation [49]. This might also explain the relatively low variability in the observed pressure profiles (Fig. 4A).
These recent studies revealed an important gap on how the characterization of the GI tract in preclinical species has been translated/extrapolated to humans and/or used to develop in vitro models and that a single attribute (e.g., anatomy or a single value of luminal pH) is not enough to assess the suitability of an animal.
Saying this, an urgent area is the development of suitable wireless technologies to expand the dynamic characterization of the colon in small animal models (i.e., rats, mouse and guinea pig). This will allow to withdraw conclusions for the suitability of the small animals and potentially to redesign e.g., the simulation of the media to properly reflect physiological conditions. Beside any fundamental difference in physiology and anatomy between animal models and humans, a harmonization of the study protocols in terms of caloric intake, volume, viscosity should also be considered.
The differences in pH and pressure amplitudes between animals and humans as well as between the two states for each species, are related to the underlying regional secretions, absorption rates, microbiota and more importantly to motility patterns.
Clinical studies increasingly require adequate animal models to develop and assess new therapies. Thus, changes in colonic motility of animals under different large bowel disorders should be equivalent to humans. However, this requires a rational use of terminology to describe those motility patterns that are equivalent to those observed in humans. Recently, there is an effort to harmonize the terminology. Direct comparison of the colonic motility patterns between human and animals is not easy.

Evaluation of colon transit in humans and animals
A comparison of colonic transit in humans and animals is presented in Table 14.

Radiopaque markers
Radiopaque markers (RMs) have been used to measure segmental and total colonic transit times. Different approaches to measure colonic transit with RMs have been described elsewhere [157][158][159][160]. The minimum number of markers required to be ingested daily should be at least 10-12 for reporting reasonably accurate colonic transit time in days or hours [161].
Depending on the distribution of markers, normal and abnormal functionality of the regions of the colon can be assessed. For instance, accumulation of markers in the rectosigmoid area might imply functional outlet obstruction, whilst distribution along the region suggests slow-transit constipation. However, this is not a diagnostic method because delayed rectosigmoid transit resulting from pelvic floor dyssynergia may inhibit proximal colonic transit and result in widespread distribution of markers.
The radiopaque marker is a well-established method, and it is readily available and reasonably inexpensive. RMs are the reference standard for the evaluation of colonic transit times in clinical practice [162].

Colonic scintigraphy
Scintigraphy (SC) is a relatively safe, non-invasive method and correlates with RMs. SC provides kinetics of ascending colon (AC) emptying and total colon transit [163]. The most common method is for subjects to ingest a pH-sensitive coated capsule containing indium-111-labeled activated charcoal particles after fasting overnight. The coated capsule dissolves in the terminal ileum and releases the radiolabelled tracer into the lumen [164]. Alternatively, a whole-gut transit test, including colonic transit, can be performed using a radiolabeled solid-liquid meal [165]. Then, repeated images of 2 min duration are acquired with a gamma camera at several time intervals after meal ingestion to appraise colonic transit [166]. The data is presented as a geometric mean (weighted average) of radioactivity based either on 5 [166] or 7 regions [165]. The recording times required to fully assess functionality of the colon are at 24, 48, and 72 h.
Application of SC in patients with diarrhoea or constipation, normal values, coefficients of variation performance characteristics and differences between slow colonic transit constipation and defecation disorders as well as relationship between colonic transit summaries of prolonged recordings and bowel function symptoms have been well documented [167][168][169][170].
High intraindividual coefficients of variation has been observed between different measuring times, e.g., 31% at 24 h vs 27% at 48 h over a period of less than 3 weeks, and 38% at 24 h vs 30% at 48 h over a median interval of 2 years [167]. This variation reflects the underlying physiological, natural variation in colonic motility, stool frequency and consistency.

Use of wireless pH capsule to measure colonic transit
The arrival of the wireless motility capsule (WMC) into the cecum is determined by a sharp decrease of pH which is more than one unit and lasts for at least 5 min; with this method, the colonic transit time is defined as the time from entry into the cecum to the time the WMC passes out of the colon with the sudden loss of pressure recordings and a decrease in temperature. The colonic transit time measured in 87 healthy and 78 constipated subjects showed a median value of 21.7 h (IQR, 15.5-37.3 h; 95th percentile, 59 h) in healthy subjects and 46.7 h (IQR, 24.0-91.9 h) in constipated patients [174]. A significant correlation between the WMC and the percentage of radiopaque markers retained on fifth day was found (r = 0.69) in constipated patients studied simultaneously with both methods [174].
High agreement of 87% for classifying subjects as having slow or normal colonic transit was found in a large multicenter study of 158 patients with constipation, using WMC method [175]. Moreover, characterization of colonic pressure activity in healthy controls and in constipated patients has been performed using WMC technology [176].
The advantages of the WMC to assess colonic transit times over the previous techniques are no radiation exposure, its non-invasive nature and allowing for whole gut and regional transit. However, distinguishing the entry into the cecum by the pH drop, is sometimes difficult. Moreover, the method is more expensive than other methods, like RMs and SC, to measure colonic or whole-gut transit. Moreover, this method does not provide information about the location of the capsule within the different regions of the small intestine and the colon.

Use of (electro)magnetic capsules to measure colonic transit
The use of electromagnetic capsules, like the electromagnetic Motilis 3D-Transit system, to measure colonic transit has been reviewed elsewhere [177]. The advantage of this method is the spatiotemporal resolution that allows assessment of regional colonic transit times which is not possible with other ambulatory methods, such as the wireless motility capsule. Although this technology has been used in several human studies, including patients, animal studies are needed to provide information about the regional and total colonic transit times to allow comparison between different species.
The first feasibility study using a magnetic capsule was conducted by Stathopoulos et al [178] in healthy subjects. The authors showed that it was possible to obtain a 3D configuration of the GI tract as well as dynamic measurements of the internal gut environment via magnet displacements, such as velocity and transit times. Few years later Worsøe et al [179] validated the technology against PillCam. The authors demonstrated that it could distinguish the between the contraction frequencies of the small intestine in the fasted and fed state.
Further improvements are required to make this technology suitable for commercial use and to decrease the data losses during recording. The latter occurring due to loss of transmission signal and poor recording quality.

Conclusions
An extensive review on animals and human physiology and the methods used to analysis different parameters was performed. There are considerable anatomical, physiological and pathophysiological differences between animal models and human. Thus, there is no single, ''gold-standard" animal model that can reproduce all the aspects and be used to extrapolate findings to human physiology. In addition, the analytical/monitoring method used to obtain information about GI tract environment, further complicating the extrapolation from animal to humans. However, significant improvements in tracking systems have permitted increased understanding of human GI tract environment. Nevertheless, the animal studies are limited, especially in small animals due to incompatibility issues, e.g., size of capsules. There is a catch 22 whereby small ingestible devices are required to characterise the GI environment in small animals that can subsequently permit evaluation of newer devices.
In terms of the most suitable animal model to use to evaluate an ingested device, this still relates to the primary function of the device under test to ensure that the study aims are met. However, it is important to recognise limitations in certain animal models for certain aspects of GI environment and their ability to replicate the human.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.