Abstract

Ingestible sensors are potentially a powerful tool for monitoring human health. Sensors have been developed that can, for example, provide pH and pressure readings or monitor medication, but capsules that can provide key information about the chemical composition of the gut are still not available. Here we report a human pilot trial of an ingestible electronic capsule that can sense oxygen, hydrogen, and carbon dioxide. The capsule uses a combination of thermal conductivity and semiconducting sensors, and their selectivity and sensitivity to different gases is controlled by adjusting the heating elements of the sensors. Gas profiles of the subjects were obtained while modulating gut microbial fermentative activities by altering their intake of dietary fibre. Ultrasound imaging confirmed that the oxygen-equivalent concentration profile could be used as an accurate marker for the location of the capsule. In a crossover study, variations of fibre intake were found to be associated with differing small intestinal and colonic transit times, and gut fermentation. Regional fermentation patterns could be defined via hydrogen gas profiles. Our gas capsule offers an accurate and safe tool for monitoring the effects of diet of individuals, and has the potential to be used as a diagnostic tool for the gut.

Main

The development of ingestible sensors offers significant opportunities in medical diagnostics and the monitoring of the human body1. In contrast to wearable sensors, which are mostly limited to contact with the skin, ingestible sensors can be immersed in the gut, an environment in which the concentrations of chemicals exchanged by our body are high1,2. However, the field of ingestible sensors remains a relatively underdeveloped area of technology, though advances in electronics and sensors now provides a base for creating reliable, safe, low-cost and durable ingestible chemical sensors. Commercial ingestible sensors are currently mainly limited to pH and pressure profilers, medication monitoring tools, and esophageal optical coherence tomography monitoring using tethered capsules1,3,4. In order to develop powerful ingestible capsules that can sense the chemical components of the gut, a first step is to identify the right target analytes of the gut, as certain gas constituents have been identified as efficient biomarkers that contain a plethora of information about the health of our bodies5,6,7,8.

The main intestinal gases are hydrogen, carbon dioxide, nitrogen, and oxygen (and also methane in some people), as well as several odoriferous gas and vapour species9,10,11. These gases originate partly from swallowed air, though their most significant sources are endogenous chemical conversions, enzymatic perturbations and, above all, the metabolic activity of the intestinal microbiota upon their interaction with unabsorbed food content12,13.

The gases of the gut have been shown to be of significant value in understanding the pathogenesis of disorders of the gut and in diagnostics6,14,15. Gas production from bacterial fermentation is likely to induce abdominal symptoms via luminal distension in patients with irritable bowel syndrome (IBS) and visceral hypersensitivity16. Carbohydrate malabsorption17 and small intestine bacterial overgrowth (SIBO) are also frequently diagnosed from H2 measurements18. These gases have also been demonstrated to be effectively modulated by diet7. Because diet can quickly and reproducibly modulate the microbiome of the gut, intestinal gas profiles would also be expected to be altered at the same rate and repeatability19. Being able to accurately measure the gastrointestinal gases should provide unique insight into the functionality of microbiota, and may enable the development of new diagnostic, therapeutic and monitoring procedures. Food intake has been strongly linked to gas production and an accurate gas-sensing device could help in designing functional foods, leading to better colonic health and understanding of the of food on our bodies1,5,12,20,21.

A variety of methods have previously been used to measure and assess gases of the gut. These include flatus analysis, tube insertion, whole body calorimetry, in vitro incubation of faecal samples and breath tests, each of which has its own advantages and drawbacks8,11,14,22,23. The methods can be highly invasive (tube insertion), inconvenient (flatus analysis, tube insertion, whole body calorimetry) and so far, apart from tube insertion, appear to be unreliable15,24. The most common is breath tests, which, while providing valuable information, suffer from interpretative difficulties mainly associated with not knowing the regional origin of the gas production and also from the very low signal-to-noise ratio of the output (gases are only measured in part-per-million concentrations).

In proof-of-concept experiments, we previously demonstrated that gas concentrations and their modulation along the gastrointestinal tract in pigs can be assessed by gas sensing capsules8,25,26. These capsules measured gut H2 (with a cross talk to CH4, CO2) and showed patterns similar to previous post mortem measurements in animals27. Here we report a human pilot trial of more advanced gas capsules, which incorporate electronic components for measuring H2, CO2 and O2, and can assess gas profile changes in response to dietary alterations28. The gas selectivity and sensitivity are obtained by modulating the heating elements of the sensors and by continuous sampling at different operating temperatures. Gas sensors can operate at both physisorptive (low temperatures) and chemisorptive (high temperatures) modes that significantly widen the gas response spectrum29,30. Additionally, this allows the extraction of gas constituents in both aerobic and anaerobic segments of the gut, which is impossible with current commercial sensors. The heat modulation also permits the use of small sensors that can be incorporated in the limited size of the capsule (in comparison to optical gas sensors, which require large spaces) and allows the regeneration of sensors by heating the sensitive elements for continuous measurements on the order of several days, equalling the capsule’s lifetime in the body.

Researches have previously shown that the microbial community of the intestine can be rapidly and reproducibly modulated by food19. Our data are complemented by the analysis of the microbial community of the faecal samples and metabolomic analysis of faecal short chain fatty acids (SCFA). The focus of our discussion is on the O2, H2 and, to some extent, CO2 profiles. O2 is chosen due to the fact that different segments of the gut have very different O2 concentration levels. Furthermore, and because the movement of the capsules through the gut is governed by the type of dietary intake1, we examined whether the O2 profile can be used to identify the location of the capsule and the speed of food passage through each segment. This process of localizing the position of the capsule in the gastrointestinal tract is benchmarked with ultrasound31.

To investigate fermentation of the food intake in the gut both CO2 and H2 profiles were obtained. However, the H2 profiles are of more interest here, as CO2 profile can still be interfered with the respiratory production. H2 plays an important role in understanding the microbial fermentation of the food in the gut as it appears in their metabolic pathways10. Gut fermentation is the anaerobic process by which most small bowel and colonic microbes gain energy from unabsorbed food. From previous flatus and measurements in vitro, it is known that H2 excretion varies markedly with different food substrates24,32,33. H2 by-production is critical for initiating and continuous fermentation, while excessive H2 is thermodynamically counterproductive, restricting further fermentation. This is naturally mitigated as H2 concentration is regulated by its simultaneous oxidation, which is conducted by three main groups of H2-utilizing microbes: reductive acetogens, methanogens and sulphate-reducing bacteria10. These microbes, together with flatus and breath excretion, dynamically reduce H2 concentration. Overall, the first step to understand the food fermentation in the gut is measuring the dynamics of H2 in situ, which has so far not been possible. The capabilities of our gas capsule in measuring H2 are explored in this work through modulating the dietary fibre content (excluding readily fermentable carbohydrates) of the food intake of the subjects in various scenarios.

Description of capsule technology

In comparison to the capsules that we used for animal trials8,25,26, the sensing capsules for humans were different in size and density, and also included intelligent processing units that could differentiate between the target gases with high selectivity. These capsules were made according to the dimensions for 000 human capsule standard (26 mm length and 9.8 mm external diameter) (Fig. 1). A non-transparent, polyethylene shell housed the internal electronic components. The polymer shell was machined in two pieces, and these top and bottom claddings were sealed together using a bio-compatible adhesive. The capsule was non-transparent as volunteers showed hesitation in swallowing capsules with transparent covers that they could see the electronic circuits inside. The capsules included sensors for CO2, H2 and O2 gases that operate in various aerobic and anaerobic conditions, a temperature sensor, a microcontroller, a transmission system (433 MHz), and button-size silver oxide batteries. A combination of thermal conductivity and semiconducting sensors, with an extraction algorithm, were used for generating the gas profiles and extract the gas concentration in both aerobic and anaerobic segments of the gut. Capsule accuracy for measuring H2 and O2 was better than 0.2%, and for CO2 it was 1%. The key technological differences between our human gas sensing capsules, and those used for animal trials on pigs8,25,26 are as follows: the implementation of an advanced gas profile extraction algorithm based on heat modulation that allows us to separate H2 and CO2 gas profiles with much higher accuracy; inclusion of an oxygen sensor for locating the capsule in different gut segments; and inclusion of a temperature sensor to measure the core body temperature and show the excretion of the of capsule out of the body of volunteers (when the temperature drops below 35 °C). The capsules also incorporated membranes with embedded nanomaterials that allow for the fast diffusion of dissolved gases, while efficiently blocking liquid34,35,36.

Fig. 1: Illustrations and photo of the capsule.
Fig. 1

a, Three-dimensional rendering of the human gas sensing capsule sliced to demonstrate the internal components. b, Photo of the packaged gas capsule and receiver device. c, Dimensions of the capsule. d, Schematic representation of intestinal gas penetration through the gas permeative membrane, reaching the sensing elements. e, Basic circuit diagram and operation of gas sensors in the capsule.

Small pocket-size receivers were developed that could be readily carried by volunteers (Fig. 1b). Information about the concentration of these gases and temperature records are continuously transmitted from the capsule to a monitor that displays the profiles in real-time on a mobile phone (Bluetooth communications between the pocket size receiver and mobile phone). Information was coded and transmitted to a hand-held monitor every 5 minutes. In order to prolong the lifetime of the capsules to more than 4 days for gas sensing, the microcontroller puts the electronic components inside the capsule to sleep mode, while the capsule is in idle mode. The temperature sensor and transmission circuit can operate for nearly 30 days, as they only require micro-amp level currents, providing an important safety feature to monitor the excretion of the capsule out of volunteer’s bodies. More advanced methods such as energy harvesting from the liquid of the gut can be used for prolonging the lifetime of the capsule even further37.

The capsules were fully calibrated prior to trials to a range of gas mixture, pressure, humidity and temperature. They were also tested to ensure no defects in sealing and ensure capsule integrity. No capsule failure was observed during the human trials. Occasional loss of data transmission occurred, but this did not impact the continuity of any of the gas and temperature profiles. The capsules were collected after their full transit and excretion from the volunteers’ bodies and all fully inspected. No damage was observed.

Capsule location and initial fermentation evaluations

Ultrasound was used for evaluating the relationship between the location of the gas capsule and changes in the gas profiles. A volunteer consumed relatively higher-than-average fibre content (approximately 33 g/day) for 24 h prior to ingestion of the gas capsule. The full detail of the diet’s nutritional value is presented in Supplementary Information. After an overnight fast, the capsule was swallowed. The fasting continued for 4 hours and then returned to the previous diet until the capsule was excreted. The gas profiles were continuously monitored and sonographic imaging was performed every 20 minutes.

The profile regions associated with the intestinal locations are presented in Fig. 2, while the actual ultrasound images are shown in the Supplementary Information. The capsule was identified within the stomach, jejunum, distal ileum and right colon as a hyperechoic object with dimensions corresponding to the capsule and an acoustic shadow (see Supplementary Information). The gas profiles (Fig. 2) and the alterations were as follows.

Fig. 2: Gas and temperature profiles and their association with the capsule location obtained using ultrasound.
Fig. 2

H2, O2-equivalent, CO2 and temperature profiles for the test associating the capsule location with gas profiles and initial evaluations. In addition to the gas profiles, temperature changes also correspond to the daily activities of the volunteer such as resting and sleeping. The identity of the gastrointestinal segment containing the capsule was determined by its location within the abdomen in relation to other organs, appearance of the bowel wall, for example, thickness and fold pattern, presence of peristalsis, and luminal contents (see Supplementary Information).

The non-specific oxygen gas sensor can sense any oxidative reagent. As such, we call the measurement “O2-equivalent” gas profile as its changes are due to the contribution from any oxidizing reagent in the environment. Three stepped changes in the O2-equivalent gas profile were observed. The first was associated with the transition from the stomach to the small intestine. The O2-equivalent concentration in the stomach liquid increased to >65% with reference to that of the atmospheric concentration (21%). In the step from the stomach to proximal small intestine, conditions remained relatively aerobic. The duodenum was not visible on ultrasound, but the capsule appeared to move rapidly into the jejunum. The second step was in the small intestine, where the O2-equivalent concentration dropped to <5% within 45 mins; this equated sonographically to movement into the ileum. The O2-equivalent concentration fell to <1% in the colon, although the exact timing of movement into the caecum was not seen sonographically. However, the capsule was located in the ascending colon shortly after.

The concentration of H2 was <0.25% in the stomach, and increased to >0.6% in the jejunum and to ~2% in the distal ileum. Its concentration peaked at 3.5% in the proximal colon and then decreased to 2% at the distal colon before capsule excretion. There are conspicuous changes in the production rates in the H2 gas profile which are closely associated with the steps seen in the O2-equivalent profile.

The concentration of CO2 in the gastric liquid was ~5% and increased after moving from the stomach. It peaked at ~30% in the proximal colon and then reduced significantly by ~25% afterwards. It increased again before the excretion of the capsule.

In this trial, the capsule left the body after 20 hours. Using the landmarks as defined above, the capsule stayed in the stomach for 4.5 hours. Transit through the small intestine took 2.5 hours.

Cross-over study at high and low dietary fibre intake

Trials were conducted based on the modulation of the fibre content in the diet to identify the effect of such diets on the gas profiles as an initial benchmark. It has been suggested that such differences in diet modulate the microbial community that should appear in the analysis of the microbial community and metabolites from faecal samples38. However, such modulations are only conspicuous at dietary changes that are considered outside the daily norm.

The initial experiment was a cross-over study performed in a single healthy male subject who consumed a diet with a markedly increased in fibre (on average 50 g/d) for 2 days prior to and during the passage of the capsule. After 2 weeks, a diet relatively low in fibre (approximately 15 g/d) was consumed for 2 days prior and during the passage of the capsule. On the third day after taking the capsule, dietary fibre was reintroduced as the capsule had not been passed.

On the very high fibre diet, the capsule resided in the stomach for 12 hours, 7 hours in the small intestine and 4 hours in the large bowel. The O2-equivalent concentration in the stomach liquid increased to >50% with reference to the atmospheric concentration before the transition to the small intestine. As in the previous study, steps were seen in the transition through the jejunum, ileum and colon. However, the dissolved oxygen content remained >10% even in the distal colon before the capsule left the body. Interestingly, the H2 concentration increased to 0.25% in the stomach after taking lunch that included lentils and chickpeas. A simultaneous drop in the O2-equivalent (~4%) was also seen after lunch. H2 concentrations only increased in the colon to <0.4% as a sign that fermentation is occurring in the colon. The CO2 content measured was ~15% when it left the stomach, fell in the small intestine to ~10% and then increased to ~20% in colon.

On the low fibre diet, the capsule remained in the stomach for ~13 hours. Small intestinal transit was 5 and a half hours and it remained in the large intestine for 54 hours. On the evening of the fourth day after taking the capsule, dietary fibre (43 g/d) was reintroduced and the capsule left the body 36 hours after this intervention, as shown in Fig. 3. The gastric O2-equivalent concentration was >21% for the duration of the capsule’s time there. As before, step transitions were observed through the jejunum, ileum and then distal colon. The O2-equivalent concentration in entering the colon remained at >10% and subsequently dropped to <3%. However, the day after dietary fibre was reintroduced, colonic conditions became highly anaerobic. The concentration of H2 remained in the order of 0.05–0.2% in the stomach and increased anytime food was taken. H2 content increased to as large as 0.4% upon the passage through the small intestine. After entering the colon, H2 concentrations decreased to 0% in a few hours. Within 12 hours of resumption of a high-fibre intake, the H2 concentration rose to >1.5% in the colon, indicative of active fermentation. The concentration of CO2 in the stomach was ~10%, dropped to ~5% during passage through the small intestine and increased in the colon to 25% coinciding with 0% for H2 and rose in concert with the rise in H2 concentration after reintroduction of dietary fibre.

Fig. 3: Outcomes from the cross-over study.
Fig. 3

a, Gas profiles for the volunteer on high fibre diet (stool collections were conducted at −2 and +21 h with reference to the capsule intake at 8 am). b, Gas profiles from the same volunteer initially on low fibre diet and then shifting to high fibre diet. Dark blue arrows identify the times for major meals and bright blue arrows the times for snacks (see Supplementary Information for details) while the capsule resided in the stomach (stool collections were conducted at +21 and +69 h with reference to the capsule intake at 8 am)). In a and b, blue is likely associate to the jejunum region, while pink is the transition through ileum (similar to Fig. 2). c, Heatmap of the relative abundance of the community members (operational taxonomic unit; OTU) from faecal samples taken during the cross-over study. Each column represents an OTU clustered at 97% identity. Only OTUs present at ≥5% relative abundance in at least one sample are shown.

In order to corroborate changes in fermentation as shown by the gas profiles, faecal SCFA and microbiota were assessed at times shown in Table 1 and Fig. 3c. During the initial very high-fibre dietary period, faecal SCFA were only minimally higher than during the low-fibre period. However, there was a marked increase following introduction of increased fibre during the low-fibre arm (Table 1). The hierarchical analysis of the microbial community of faecal samples is shown in Fig. 3c. Continuation of the high fibre diet from the 2nd to 3rd day – when the O2-equivalent concentration increased - significantly decreased the abundance of Erysipelotrichaceae family and Bacteroides spp., while the abundances of Peptostreptococcaceae and Christensenellaceae families were increased (Fig. 3c and Supplementary Information). A change in microbiota was also observed after shifting to the high fibre diet (day 3–4) from the low fibre diet (day 2). The abundance of Lachnospiraceae family and Roseburia spp. were increased.

Table 1 Metabolomic analysis of short-chain fatty acids in faeces (in µmol/L) from a healthy male associated with changes in the intake of dietary fibre for the cross-over study

Repeatability tests at high and low fibre diet

Four healthy volunteers were recruited by advertising. To ensure that the volunteers were comfortable, the difference between the dietary fibre content was less than the previous trial. The subjects were provided with all their food for 24 hours prior to and during the capsule study. Two (#1 and #2 in Fig. 4) received a diet high in fermentable fibre (male 36 g/day and female 33 g/day) and two (#3 and #4 in Fig. 4) a diet lower in fermentable fibre (male 23 g/day and female 22 g/day).

Fig. 4: Gas profiles obtained from repeatability tests.
Fig. 4

For volunteers #1 and #2 (P07 and P09, respectively – both female) who are on a low fibre diet, only one step in the O2 equivalent graphs are seen, which is associated with transition through the small intestine (the region is shown in pink). For volunteers under a high fibre diet (volunteers #3 (female) and #4 (male) which are P01 and P05, respectively), a two-step transition in the small intestine area is seen (similar to Fig. 2) that is likely associated with the transition from the stomach to jejunum (blue region) and from the ilium to colon (pink region). The colonic telemetry regions have been cut down to make the graphs 15 and 24 h for low and high and fibre groups, respectively, to facilitate the direct comparison.

The gas profiles together with the temperature profiles are presented in Fig. 4. Total gut transit time was >25 those on the high fibre and >20 h on the lower fibre diet. The patterns of O2-equivalent, H2 and CO2 concentrations were very similar for those on low fibre diet. Passage of the capsule through the small intestine was rapid (<50 minutes) in those on the lower fibre diet and O2-equivalent down-step in mid-small intestine was not seen. In contrast, small intestinal transit was several hours in both subjects on the high-fibre diet and a clear step in the O2-equivalent content was observed. H2 concentrations increased along the small intestine, being particularly evident in the ileum with the high fibre diet, and further increased in the colon, although the magnitude of increase varied across the individuals. The CO2 profiles in the small intestine followed a similar pattern with reduction in the distal small intestine, but, in the colon, a progressive increase in the CO2 content occurred during passage through the colon, but a fall was observed in the third subject on a high fibre diet.

For the repeatability tests, community analysis of the faecal samples were also conducted (Supplementary Information) and provided no discernible trend in the abundance change of any microbial species.

Analysis of gas profiles

This in-human pilot study of the fully electronic-component-based gas-sensing capsules shows their capability to obtain O2-equivalent, H2, CO2 and temperature profiles, and to transmit signals to a portable receiver successfully for surveying gas concentrations along the entire gastrointestinal tract. The capsules were easily swallowed and well tolerated (see Methods) and the tests were conducted when the dietary fibre (excluding readily fermentable carbohydrates) of the meals were modulated.

An essential part of any telemetric device in the gastrointestinal tract is to have a means of locating its anatomical position along each segment of the gut. This was achieved using an O2 gas sensor that was non-specifically responsive to all oxidizing gases and vapours. We have proposed that changes in the O2-equivalent concentrations, given the understanding that the luminal environment becomes more anaerobic along the gut. There were steps in O2-equivalent concentrations as anticipated from the stomach to small intestine and from the ileum to caecum. These transitions were validated using intestinal ultrasound. An additional step was observed in most of the studies within the small intestine, except when intestinal transit was very rapid in the small intestine. The extra step most likely represents the capsule passing from the jejunum to ileum where bacterial density increases. Breath testing is currently a commonly-used diagnostic method in gut disorders including in diagnosis of malabsorption of carbohydrates, SIBO and IBS. However, one of the reasons that the accuracy of breath testing is compromised is due to the lack of information regarding the point of production of the gas in the gut. The localization capability of the gas sensing capsule sufficiently addresses this need.

The O2-equivalent gas profile in the stomach region showed an increase in magnitude to the levels above that of dissolved O2 in aqueous solution at 37 °C. So it was likely that the origin could be an oxidizing gas rather than O2, affecting the non-specific O2 sensor. To understand the origin of this phenomenon, in vitro simulations of the mouth and stomach phase were conducted with gastric liquid based on the recipe by Minekus et al. (see Supplementary Information) and gas profiles were obtained39. The simulation outcomes showed that the presence of K2HPO4 in the gastric juice is the likely source of the oxidizing agent. The gastrointestinal tract enzymes also mitigated the agent’s effect. It is likely then that the flow of enzymes after taking each meal (Fig. 3a) is responsible for reducing the O2-equivalent concentration in the current studies.

An initial cross-over study was performed in which a very high fibre diet was compared to one that was much lower and changes in gas profiles provided unique data. First, colonic fermentation was minimally stimulated by the very high fibre arm with only a small increase in CO2 concentration in the colon, minimal increase in H2 concentration and some change to SCFA concentrations. Furthermore, microbial profiles changed to a pattern not seen when fermentation was stimulated in the second part of the cross-over. Indeed, the decreased abundance of Bacteroides spp.40 and increased abundance of Erysipelotrichaceae family have been associated with poor gut health41. Secondly, the colon did not become anaerobic as anticipated. Thirdly, the subject developed some abdominal discomfort in response to very high fibre content diet. The concomitant unexpected lack of fermentation and bacterial profile together with failure of the colon to become anaerobic suggested that the microbiota were unable to ferment the whole fibre content possibly due to excessive trapped oxygen in the fibre, affecting the activity of anaerobic bacteria. Furthermore, the diet induced gut discomfort was a possible manifestation of oxidative stress in the colon42.

In the second arm of the cross-over trial, the same subject took a diet that was low in fibre. The gas profiles were more akin to that expected with evidence of fermentation (decrease in H2 concentration together with anaerobic conditions in the colon). Also consistent with low fibre intake, the frequency of bowel movements decreased. The subsequent increased intake of fibre to moderate levels in response to failure to pass the capsule was then associated with increased colonic H2 and CO2 concentrations 11 hours after recommencement of fibre, a pattern indicative of increased fermentation. Faecal indices confirmed the increase in fermentation with fibre-related increases in SCFA concentrations and changes in the microbial communities, as demonstrated by the increase in the strongly butyrate-producing Roseburia species43. The furthest shift of microbiota community in the hierarchical chart (Fig. 3c) was generated by the lowest fibre diet. However, after reverting to a higher fibre diet (4th day low to high fibre), the microbial community more closely resembled that of the baseline period of high fibre diet (2nd day high fibre).

A pilot repeatability assessment of the gas profiles measured by the capsules was conducted using a four-subject trial on diets that varied in fibre content. The O2-equivalent profiles from the capsules showed repeatability between subjects on the same diets, especially for volunteers on the low fibre diet. The lower fibre diet was associated with rapid small intestinal transit, such that the step in the small bowel was not observed, as for the first subject on a high-fibre diet. The notable difference in small intestinal transit are likely to be related to the changed motility in responses to the effects of soluble fibre and changes in the concentration of absorbable carbohydrates as previously reported44,45. An important observation in the repeatability pilot trial was the short transition time of the capsules for the low fibre diet case and prolonged small bowel transition in response to higher fibre diet. Despite the small number of samples, the outcomes were encouragingly consistent with past direct observations using motility measurements46 and indirect measurements using breath test47. As there is a strong correlation between the passage of capsule and dietary intake passage (Kuo et al. reported that there is a correlation of r = 0.73 between the two events)48, this observation is of significance for dietary and medical considerations. Such changes in small intestinal transit are currently utilized in slowing absorption of carbohydrates with soluble fibres to reduce the glycaemic index of foods and can also enhance the delivery of drugs to the colon49. In all four subjects, increased CO2 and H2 concentrations were noted in the distal small intestine and especially in the proximal colon, consistent with fermentation of carbohydrates.

As noted above, the gas-sensing capsule appears to be a valuable means for measuring the gastric, small and large intestinal transit time of solid food. This relies on the measurement of O2 concentration rather than pH, which is used in conventional smart capsules. The main advantage of O2 measurement is that oxygen sensors are linear devices that are inherently less noisy than the logarithmic base pH sensors. Reduced H2 provided evidence for subdued fermentation in the colon that is also in agreement with previous reports on its importance in colonic homeostasis and energy extraction/disposal processes10.

Another possible observation was that there was no discernible pattern in the microbial community structure from the faecal samples, in the high fibre and low fibre cohorts, based on the pilot repeatability tests. In contrast, the gas sensing capsule was able to clearly discern the differences between the two cohorts using the passage rate of the digesta and fermentation states. This observation however should be further endorsed using a much larger number of volunteers.

Conclusions

Gas-sensing capsules were successfully used to evaluate small and large intestinal transit times in a human pilot trial. Modulation of dietary fibre and, therefore, fermentative substrates in general produced changes in H2 and CO2 gas profiles consistent with past indirect measurements of gas production in the terminal ileum and colon. The capsule could document inter-individual differences in fermentative patterns, which should be of significant value in the individualization of drug disposition or the use of dietary manipulations. One scenario associated with very high fibre intake produced paradoxical results that are unlikely to be an artefact, but rather highlight the value of measuring gas production at its source, which could provide a new approach to understanding intestinal physiology and its food modulation. Our pilot trial illustrated the significant potential role for electronic-based gas-sensing capsules in understanding functional aspects of the intestine and its microbiota in health and in response to dietary changes. The capsules provide a potentially powerful diagnostic technique, and could offer unique insights into the effects of diet and medical supplements. This might also translate into a monitoring tool that can be used to help develop individualized diets.

Methods

Operation of the capsule

The semiconducting metal oxide based sensor used is non-specific and responsive to all oxidizing gases and vapours. This sensor is responsible for measuring the concentration of oxidizing reagents in different segments of the gut. The sensor baseline also changes in the presence of H2 gas. The thermal conductivity sensor, which is known not to be O2 sensitive, is used for H2 measurements.

A pulse is applied to the heater element of the sensors. Semiconducting and thermal conductivity sensors outputs are sampled both when the heater element is heating up and while it is cooling down. At different heating temperatures, semiconducting and thermal conductivity sensing elements become more or less sensitive to different target gases.

The process for measuring an unknown gas first requires the pre-calibration of the sensors to the selected gases. The basic steps of the modelling process are as follows: apply a known gas to the sensor; operate the sensor heater element with a specific wave shape and record the sensor’s resistivity in time; and generate a mathematical model such as those by Bermak et al.50 and Kermit and Tomic51 for extracting information about the gas constituents using the recorded calibration data.

Once an adequate model has been generated, the sensor is used for measuring unknown gases in the capsule. This process is as follows: expose the sensor to the gases of the gut; operate the sensor heater element and record its resistivity in time using an applied wave shape similar to the calibration process; and estimate the absolute concentration of the unknown gas from the model. See Supplementary Information for more details.


Tolerability and safety

The exclusion criteria applied to the healthy volunteers is presented in the Supplementary Information. All healthy participants were tested with patency capsules prior to the trials and they were excluded if the patency capsule gut transit exceeded 48 hours. All capsules were easily swallowed and none were retained. No abnormal capsule retention was observed during the trials (for all 7 tests), which is also in accordance with the previous reports that capsule retention in healthy volunteers is not common. In a report regarding smart capsules tested on 6,000 healthy volunteers only 5 underwent endoscopic intervention after prolonged retention52. The retention chance in patients with gut disorders is in the order of ~1%53. It is important to consider that our capsule is 000 human standard, while many commercial smart capsules are significantly larger in size. In the cross-over study, two adverse effects were recorded that were likely due to the diet taken and not the capsules. Abdominal discomfort was reported by the first subject on the very high-fibre diet due to the known effect of excessive fibre intake54. This resolved after the bowel movement of the volunteer on the third day. The second was a mild constipation effect (between +21 and +69 h Fig. 3) on the low-fibre diet by the same subject (no stool sample could be collected between +21 and +69 h due to constipation of the volunteer) which is also a very recognized symptom of lack of fibre. This resolved following increasing the fibre intake after three days. No symptoms were seen during the repeatability tests as the fibre content of the diet was kept within the recommended daily doses. All capsules remained fully operational during their transit along the gut and excretion.


Study design and procedures

Three trials were conducted to evaluate the performance of capsules. The study protocols were approved by the Nepean Ethics HREC Reference No.: 15/64 - HREC /15/NEPEAN/128. This is a multi-institutional approval that allows the trials to be conducted at any registered site across Australia on healthy volunteers. The inclusion criteria for trial participants are presented in Supplementary Information.


Evaluation of gas profiles to determine localization in the gastrointestinal tract

Localization of the gas capsule was performed using ultrasound. For detail refer to Supplementary Information.


Benchmarking with high and low-fibre diets (cross-over study)

The volunteer was a non-obese 44 year-old male. Weighed-food records were kept enabling fibre content of the diets to be determined using FoodWorks software (Xyris Software Pty Ltd, Brisbane, QLD, Australia, version 8) utilizing the NUTTTAB database.


Comparison of diets differing in fibre content on gas profiles

Four healthy volunteers were studied under controlled dietary conditions to examine the repeatability of profiles generated for high and low fibre diets. The amount of fibre in the diets was calculated from weighed-food intake records. One was male and 3 female, and their ages ranged between 29 and 41 years. Two were randomly allocated to a diet high in fibre and two low in fibre. The full detail of the diets’ nutritional values that were provided is presented in the Supplementary Information.


Faecal sampling and analysis

Faecal samples were collected from participant in the cross-over study at times indicated in the Results section. For analysis of the microbial community, the samples were collected with swabs and processed according to a protocol defined by the Australian Centre for Ecogenomics (Brisbane, Australia). Faecal material was collected using 4N6FLOQ Swabs (Copan) and processed as detailed in Supplementary Information. Faeces were also collected for metabolomic analysis and kept at −20 °C until processed and analysed using nuclear magnetic resonance spectroscopy (details in Supplementary Information).


Life Sciences Reporting Summary

Further information on experimental design is available in the Life Sciences Reporting Summary.


Data availability

The data that support the plots within this paper and other findings of this study are available from the corresponding author upon reasonable request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Kalantar-Zadeh, K., Ha, N., Ou, J. Z. & Berean, K. J. Ingestible Sensors. ACS Sens. 2, 468–483 (2017).

  2. 2.

    Mowat, A. M. & Agace, W. W. Regional specialization within the intestinal immune system. Nat. Rev. Immunol. 14, 667–685 (2014).

  3. 3.

    Bettinger, C. J. Materials advances for next-generation ingestible electronic medical devices. Trends Biotechnol. 33, 575–585 (2015).

  4. 4.

    Gora, M. J. et al. Tethered capsule endomicroscopy enables less invasive imaging of gastrointestinal tract microstructure. Nat. Med. 19, 238–240 (2013).

  5. 5.

    Gibson, G. R., Probert, H. M., Van Loo, J., Rastall, R. A. & Roberfroid, M. B. Dietary modulation of the human colonic microbiota: updating the concept of prebiotics. Nutr. Res. Rev. 17, 259–275 (2004).

  6. 6.

    Mathur, R. et al. Methane and hydrogen positivity on breath test is associated with greater body mass index and body fat. J. Clin. Endocrinol. Metab. 98, E698–E702 (2013).

  7. 7.

    Ong, D. K. et al. Manipulation of dietary short chain carbohydrates alters the pattern of gas production and genesis of symptoms in irritable bowel syndrome. J. Gastroenterol. Hepatol. 25, 1366–1373 (2010).

  8. 8.

    Ou, J. Z. et al. Human intestinal gas measurement systems: in vitro fermentation and gas capsules. Trends Biotechnol. 33, 208–213 (2015).

  9. 9.

    Levitt, M. D. & Bond, J. H. Volume, composition, and source of intestinal gas. Gastroenterol 59, 921–929 (1970).

  10. 10.

    Carbonero, F., Benefiel, A. C. & Gaskins, H. R. Contributions of the microbial hydrogen economy to colonic homeostasis. Nat. Rev. Gastroenterol. Hepatol. 9, 504–518 (2012).

  11. 11.

    Levitt, M. D. Volume and composition of human intestinal gas determined by means of and intestinal washout technic. N. Engl. J. Med. 284, 1394–1398 (1971).

  12. 12.

    Nicholson, J. K. et al. Host-gut microbiota metabolic interactions. Science 336, 1262–1267 (2012).

  13. 13.

    Tuohy, K. M., Probert, H. M., Smejkal, C. W. & Gibson, G. R. Using probiotics and prebiotics to improve gut health. Drug Discov. Today 8, 692–700 (2003).

  14. 14.

    King, T. S., Elia, M. & Hunter, J. O. Abnormal colonic fermentation in irritable bowel syndrome. Lancet 352, 1187–1189 (1998).

  15. 15.

    Braden, B., Lembcke, B., Kuker, W. & Caspary, W. F. C-13-breath tests: Current state of the art and future directions. Dig. Liver Dis. 39, 795–805 (2007).

  16. 16.

    Major, G. et al. Colon hypersensitivity to distension, rather than excessive gas production, produces carbohydrate-related symptoms in individuals with irritable bowel syndrome. Gastroenterol 152, 124–133 (2017).

  17. 17.

    Rumessen, J. J. & Gudmandhoyer, E. Functional bowel-disease - malabsorption and abdominal distress after ingestion of fructose, sorbitol, and fructose-sorbitol mixtures. Gastroenterol 95, 694–700 (1988).

  18. 18.

    Shin, W. Medical applications of breath hydrogen measurements. Anal. Bioanal. Chem. 406, 3931–3939 (2014).

  19. 19.

    David, L. A. et al. Diet rapidly and reproducibly alters the human gut microbiome. Nature 505, 559–563 (2014).

  20. 20.

    Gibson, G. R. & Roberfroid, M. B. Dietary modulation of the human colonic microbiota - introducing the concept of prebiotics. J. Nutr. 125, 1401–1412 (1995).

  21. 21.

    Roberfroid, M. et al. Prebiotic effects: metabolic and health benefits. Br. J. Nutr. 104, S1–S63 (2010).

  22. 22.

    Suarez, F., Furne, J., Springfield, J. & Levitt, M. Insights into human colonic physiology obtained from the study of flatus composition. Am. J. Physiol. Gastrointest. Liver Physiol. 272, G1028–G1033 (1997).

  23. 23.

    Rotbart, A. et al. Designing an in-vitro gas profiling system for human faecal samples. Sens. Actuat. B Chem. 238, 754–764 (2017).

  24. 24.

    Tomlin, J., Lowis, C. & Read, N. W. Investigation of normal flatus production in healthy-volunteers. Gut 32, 665–669 (1991).

  25. 25.

    Kalantar-Zadeh, K. et al. Intestinal gas capsules: a proof-of-concept demonstration. Gastroenterol 150, 37–39 (2016).

  26. 26.

    Ou, J. Z. et al. Potential of in vivo real-time gastric gas profiling: a pilot evaluation of heat-stress and modulating dietary cinnamon effect in an animal model. Sci. Rep. 6, 33387 (2016).

  27. 27.

    Jensen, B. B. & Jorgensen, H. Effect of dietary fiber on microbial activity and microbial gas-production in various regions of the gastrointestinal-tract of pigs. Appl. Environ. Microbiol. 60, 1897–1904 (1994).

  28. 28.

    Walsh, C. J., Guinane, C. M., O’Toole, P. W. & Cotter, P. D. Beneficial modulation of the gut microbiota. FEBS Lett. 588, 4120–4130 (2014).

  29. 29.

    Barsan, N. & Weimar, U. Conduction model of metal oxide gas sensors. J. Electroceram. 7, 143–167 (2001).

  30. 30.

    Ou, J. Z. et al. Physisorption-Based Charge Transfer in Two-Dimensional SnS2 for Selective and Reversible NO2 Gas Sensing. ACS Nano 9, 10313–10323 (2015).

  31. 31.

    Kobayashi, Y. et al. Sonographic detection of a patency capsule prior to capsule endoscopy: case report. J. Clin. Ultrasound 42, 554–556 (2014).

  32. 32.

    Steggerda, F. R. Gastrointestinal gas following food consumption. Ann. NY Acad. Sci. 150, 57–66 (1968).

  33. 33.

    Lemarchand, L., Wilkens, L. R., Harwood, P. & Cooney, R. V. Use of breath hydrogen and methane as markers of colonic fermentation in epidemiologic studies - circadian patterns of excretion. Environ. Health Perspect. 98, 199–202 (1992).

  34. 34.

    Berean, K. J. et al. 2D MoS2 PDMS Nanocomposites for NO2 Separation. Small 11, 5035–5040 (2015).

  35. 35.

    Berean, K. J. et al. Enhanced gas permeation through graphene nanocomposites. J. Phys. Chem. C. 119, 13700–13712 (2015).

  36. 36.

    Nour, M. et al. Silver nanoparticle/PDMS nanocomposite catalytic membranes for H2S gas removal. J. Membr. Sci. 470, 346–355 (2014).

  37. 37.

    Nadeau, P. et al. Prolonged energy harvesting for ingestible devices. Nat. Biomed. Eng. 1, 0022 (2017).

  38. 38.

    Fernandes, J., Su, W., Rahat-Rozenbloom, S., Wolever, T. M. S. & Comelli, E. M. Adiposity, gut microbiota and faecal short chain fatty acids are linked in adult humans. Nutr. Diabetes 4, e121 (2014).

  39. 39.

    Minekus, M. et al. A standardised static in vitro digestion method suitable for food - an international consensus. Food Funct. 5, 1113–1124 (2014).

  40. 40.

    Guarner, F. & Malagelada, J. R. Gut flora in health and disease. Lancet 361, 512–519 (2003).

  41. 41.

    Henao-Mejia, J. et al. Inflammasome-mediated dysbiosis regulates progression of NAFLD and obesity. Nature 482, 179–185 (2012).

  42. 42.

    Bhattacharyya, A., Chattopadhyay, R., Mitra, S. & Crowe, S. E. Oxidative stress: an essential factor in the pathogenesis of gastrointestinal mucosal diseases. Physiol. Rev. 94, 329–354 (2014).

  43. 43.

    Aminov, R. I. et al. Molecular diversity, cultivation, and improved detection by fluorescent in situ hybridization of a dominant group of human gut bacteria related to Roseburia spp. or Eubacterium rectale. Appl. Environ. Microbiol. 72, 6371–6376 (2006).

  44. 44.

    Wilfart, A., Montagne, L., Simmins, H., Noblet, J. & van Milgen, J. Digesta transit in different segments of the gastrointestinal tract of pigs as affected by insoluble fibre supplied by wheat bran. Br. J. Nutr. 98, 54–62 (2007).

  45. 45.

    Roberfroid, M. Dietry fiber, inulin, and oligofructose - a review comparing their physiological effects. Crit. Rev. Food Sci. Nutr. 33, 103–148 (1993).

  46. 46.

    vonSchonfeld, J., Evans, D. F. & Wingate, D. L. Effect of viscous fiber (Guar) on postprandial motor activity in human small bowel. Dig. Dis. Sci. 42, 1613–1617 (1997).

  47. 47.

    Simren, M. & Stotzer, P. O. Use and abuse of hydrogen breath tests. Gut 55, 297–303 (2006).

  48. 48.

    Kuo, B. et al. Generalized transit delay on wireless motility capsule testing in patients with clinical suspicion of gastroparesis, small intestinal dysmotility, or slow transit constipation. Dig. Dis. Sci. 56, 2928–2938 (2011).

  49. 49.

    Jenkins, D. J. A., Kendall, C. W. C., Axelsen, M., Augustin, L. S. A. & Vuksan, V. Viscous and nonviscous fibres, nonabsorbable and low glycaemic index carbohydrates, blood lipids and coronary heart disease. Curr. Opin. Lipidol. 11, 49–56 (2000).

  50. 50.

    Bermak, A., Belhouari, S. B., Shi, M. & Martinez, D. in Encyclopedia of Sensors Vol. 10 (eds C. A. Grimes, E. C. Dickey, & M. V. Pishko) 1-17 (American Scientifc Publishers, 2006).

  51. 51.

    Kermit, M. & Tomic, O. Independent component analysis applied on gas sensor array measurement data. IEEE Sens. J. 3, 218–228 (2003).

  52. 52.

    Tran, K., Brun, R. & Kuo, B. Evaluation of regional and whole gut motility using the wireless motility capsule: relevance in clinical practice. Ther. Adv. Gastroenterol. 5, 249–260 (2012).

  53. 53.

    Höög, C. M. et al. Capsule retentions and incomplete capsule endoscopy examinations: An analysis of 2300 examinations. Gastroenterol. Res. Pract. 2012, 518718 (2012).

  54. 54.

    High Fiber Diet (University of Michigan, Michigan Medicine, USA, 2011). http://www.med.umich.edu/1libr/MBCP/HighFiberDiet.pdf.

Download references

Acknowledgements

The authors acknowledge the Australian Centre for Ecogenomics for sequencing and bioinformatics advice, Queensland, Australia. The authors also thank the National Health and Medical Research Council (NHMRC), Australia and Department of Business, Australia for the financial support of the project via a Development grant and an Acceleration Commercialisation (AC) grant, respectively.

Author information

Affiliations

  1. School of Engineering, RMIT University, Melbourne, Australia

    • Kourosh Kalantar-Zadeh
    • , Kyle J. Berean
    • , Nam Ha
    • , Adam F. Chrimes
    • , Kai Xu
    • , Jian Zhen Ou
    •  & Naresh Pillai
  2. School of Science, RMIT University, Melbourne, Australia

    • Danilla Grando
    • , Jos L. Campbell
    •  & Robert Brkljača
  3. Department of Gastroenterology, Alfred Hospital, Monash University, Melbourne, Australia

    • Kirstin M. Taylor
    • , Rebecca E. Burgell
    • , Chu K. Yao
    • , Jane G. Muir
    •  & Peter R. Gibson
  4. Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia

    • Stephanie A. Ward
  5. CSIRO, Agriculture and Food, St Lucia, Brisbane, Australia

    • Chris S. McSweeney

Authors

  1. Search for Kourosh Kalantar-Zadeh in:

  2. Search for Kyle J. Berean in:

  3. Search for Nam Ha in:

  4. Search for Adam F. Chrimes in:

  5. Search for Kai Xu in:

  6. Search for Danilla Grando in:

  7. Search for Jian Zhen Ou in:

  8. Search for Naresh Pillai in:

  9. Search for Jos L. Campbell in:

  10. Search for Robert Brkljača in:

  11. Search for Kirstin M. Taylor in:

  12. Search for Rebecca E. Burgell in:

  13. Search for Chu K. Yao in:

  14. Search for Stephanie A. Ward in:

  15. Search for Chris S. McSweeney in:

  16. Search for Jane G. Muir in:

  17. Search for Peter R. Gibson in:

Contributions

K.K.-Z. and P.R.G. initiated the concept. K.K.-Z., J.Z.O. and K.J.B. designed the trial. N.H. designed and fabricated the capsules with some help from A.F.C. and K.X. K.K.-Z., K.J.B., C.K.Y., D.G., J.Z.O., K.M.T., R.E.B., P.R.G. and J.G.M. organized the human trials on volunteers. D.G. and R.B. carried out the metabolomics analysis. K.K.-Z., N.P., J.Z.O. and J.L.C. conducted the in vitro tests. All authors participated in analysis of data and authorship of the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to Kourosh Kalantar-Zadeh or Peter R. Gibson.

Supplementary information

  1. Supplementary Information

    Supplementary Figures 1–8, Supplementary Tables 1–4, Supplementary Notes, Supplementary Methods, and Supplementary References.

  2. Life Sciences Reporting Summary

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41928-017-0004-x

Further reading