SEARCH

SEARCH BY CITATION

Keywords:

  • colonic motility;
  • fermentation;
  • starch

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Background  In healthy humans, up to 30 g of daily ingested starch escape small intestinal digestion, and are fermented in the colon. This physiological starch malabsorption could modify colonic motility through metabolites such as short-chain fatty acids produced by fermentation.

Methods  Ten healthy volunteers swallowed a probe, consisting of an infusion catheter, six perfused catheters and a balloon connected to a barostat. On two consecutive days colonic motility was recorded in fasting subjects in the basal state (1 h), and then during (3 h), and after (2 h) the intracolonic infusion of 750 mL of isoosmotic and isovolumetric solutions containing sodium chloride with or without 15 g wheat starch. We determined (i) the volume of hydrogen and methane exhaled in breath, (ii) a global motility index and the number of high amplitude propagated contractions (HAPCs), and (iii) the mean balloon volume, reflecting the tonic motor activity.

Key Results  [median (IQR)] Compared to the basal period, colonic infusion of starch or saline did not modify the colonic motility index and tone. However, the number of HAPCs was significantly higher during and after infusion of starch than of saline [4.5 (2.75–6.5) vs 0.96 (0–2.66)/5 h, starch vs saline respectively; = 0.011].

Conclusions & Inferences  In healthy humans, colonic fermentation of a physiological malabsorbed amount of starch has no effect on the tonic and phasic colonic motor activities, but produces a significant increase in the number of HAPCs. This may participate in the physiological propulsion of colonic contents.


Abbreviations:
anova

analysis of variance

HAPC

high amplitude propagated contraction

ppm

parts per million

SCFA

short-chain fatty acid

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Carbohydrates which escape digestion in the human small intestine undergo bacterial fermentation in the colon. This results in the colonic production of short-chain fatty acids (SCFAs), gases (hydrogen, H2; carbon dioxide, CO2; and sometimes methane, CH4) and in a variable decrease in the colonic pH.1,2 Intracolonic fermentation may also modulate the postprandial release of various hormones,3,4 and could inhibit biliary5 and gastric6 secretions. It produces a gastric relaxation,7 could delay gastric emptying4 and may influence lower esophageal sphincter function.8 Finally, previous studies have suggested that colonic fermentation could also modify colonic motility.9,10 We recently confirmed these results by showing in healthy volunteers that 10–15 g ingestion or intracolonic infusion of lactulose produces a tonic contraction in the colon associated with an acceleration of ascending colon emptying whereas the non-fermentable laxative polyethylene glycol 4000 had no effect.11

Colonic motor activity in humans can be recorded by electromyography, manometry, and with the electronic barostat. These methods require usually the placement of recording probes through the anus with the aid of colonoscopy after prior colonic cleansing. We have previously developed a method to study the motor activity in the unprepared colon.12 The recording probe is ingested by healthy volunteers and migrates into the colon. Because this method does not require prior colonic cleansing, it preserves colonic microbiota and thus colonic fermentation.

Under physiological conditions, a part of the ingested starch escapes digestion in the normal human small intestine. Various methods have been used to determine the amount of malabsorbed starch.13–16 Although quantitative assessment varies with the method used, up to 30 g of the daily ingested amount of starch can be delivered to the colon, depending on the total amount of starch ingested, botanical source, technological treatments and the way of consumption.13–15,17–20

The aim of our study was to compare the effect of colonic infusion of starch to the effect of saline infusion on colonic phasic and tonic motor activity in healthy volunteers. Fermentation of starch was confirmed by monitoring the excretion of H2 and CH4 in breath. Our hypothesis was that starch fermentation could have an effect on colonic motility in healthy humans.

Subjects and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Participants

Ten healthy volunteers (6 males, 4 females, aged 22–28 years) were included after they gave informed consent to the protocol, which had been approved by the Ethics Committee of Saint-Louis Hospital (Paris, France). All were required to have no gastrointestinal (GI) symptom, a normal bowel habit, a normal physical examination, no previous GI surgery (except for appendectomy) and none was taking medications other than oral contraceptives. None had received antibiotic therapy within the previous 2 months.

Recording assembly

We used a multilumen tube assembly (1 cm OD, 3.5 m long) with six manometric catheters, a radiopaque infusion catheter and a latex balloon in addition to a cylindrical non-compliant barostat bag (10 cm longitudinal axis, 9.4 cm diameter, 450 mL capacity) made of ultrathin polyethylene. The latex balloon (5 cm axis) was placed at the tip of the tube, 24 cm caudad to the bag of the barostat. It contained 25 g of mercury and could be filled in with air to facilitate the progression of the tube. The six manometric ports were located 2, 7 and 12 cm distal from and proximal to the barostat bag. The infusion catheter opening was located 17 cm proximal to the barostat bag. Manometric catheters were connected to a low-compliance perfusion system (0.1 mL min−1 flow rate), and phasic activity was recorded by external pressure transducers connected to a data processing system as did pressure and volume of the barostat (Synetics; ABS, Saint-Dié, France). The barostat bag system was used as previously described.21 Validation experiments were performed in vitro to demonstrate that the length of the tube did not modify the technical properties of the electronic barostat. For each experiment the minimal distending pressure needed to overcome the intra-abdominal pressure was determined.22,23 Then basal pressure to record intrabag volume variations was defined as a pressure level 2 mmHg above the minimal distending pressure.23

Experimental procedure

Participants were intubated by mouth. The progression of the tube through the small bowel and the colon was aided by inflating the distal balloon, and monitored by fluoroscopic control. When the infusion catheter was located in the cecum, the distal balloon was deflated. The duration of the progression was 24–48 h. The exact location of the catheter was checked by fluoroscopy before and after each experiment. Recordings were performed on two consecutive days. Low-residue meals were served to the volunteers during the progression of the tube and on the two experimental days. On each experimental day, the barostat bag was unfolded by inflation with 100 mL air and thereafter completely deflated and connected to the barostat. Basal colonic intrabag volume and phasic activity were recorded for 1 h in the fasting state.

Then a 750 mL solution of distilled water with or without 15 g wheat starch was infused in a random order using the infusion catheter at a constant flow rate of 4 mL min−1 corresponding to the postprandial ileocecal flow rate in normal subjects.9,10 In both solutions sodium chloride was added in order to reach isotonicity (300 mosmol L−1) and the pH was adjusted to 7.0. Raw native wheat starch, extracted from flour by soaking in cold water (Roquette Frères, Lestrem, France), was used. Volume variations and phasic activity were continuously recorded during (3 h) and after (2 h) the infusion. During the recordings, subjects continued to fast and not to smoke, a glass of water was allowed and they were asked to avoid any movement and not to sleep. Symptoms including borborygmi, bloating, abdominal pain and flatus were collected hourly and graded (weak, moderate or high). The occurrence of symptoms was reported by pressing a button in order to mark the event on motility recording. Abdominal girth was determined with a flexible measuring tape hourly for 6 h. End-expiratory samples were collected into plastic syringes every 30 min before and for 5 h after the beginning of starch infusion. When the latex balloon was spontaneously expelled through the anus, it was then taped to the thigh, and the proximal part of the manometric probe was cut. Then the volunteer had to wait the natural expulsion of the entire probe. This procedure was painless and did not require sedation.

Data analysis

In end-expiratory samples H2 and CH4 were measured by gas chromatography (Quintron Instrument Co. Inc., Milwaukee, WI, USA). Subjects were considered as CH4 excretors when breath CH4 concentrations were at least two parts per million (ppm) above room atmosphere on three consecutive breath samples.24 The excess volumes of H2 and CH4 excreted in breath were determined by integrating the areas under the H2 and CH4 concentration curves after subtracting the lowest concentration of H2 and CH4 measured in the basal state; tidal volumes were determined from the Radford nomogram25 and data were expressed in mL.26 As methanogenic bacteria from the human colon consume 4 mol of H2 to form 1 mol of CH4 (4H2 + CO2[RIGHTWARDS ARROW]CH4 + 2H2O), an excess volume of H2 equivalent (H2E) was therefore calculated for methanogenic subjects as [H2E (mL) = 4 × CH4 excretion (mL) + H2 excretion (mL)], allowing both the H2 used in methanogenesis and the H2 excreted to be summed.27 Thus the calculation of the mean excess volume of H2 excreted used the values of H2 volume in non-methanogenic subjects and of H2E in methanogenic subjects. Change in abdominal girth was evaluated by determining the maximal change in abdominal girth compared to the basal value.

Motility recordings were analyzed in a blinded fashion. Phasic pressure activity was expressed as a motility index (area under the curve of phasic contractions) calculated for 1-hour periods. A visual analysis was initially performed to exclude motion artifacts, baseline drift was adjusted for motility index analysis. We calculated the mean motility index for all proximal and distal catheters. The mean motility index of the basal period recorded during the fasting state represented the basal motility index and variations in motility indexes were expressed as a percentage of this value. High amplitude propagated contractions (HAPCs) were defined as single pressure waves with an amplitude of at least 60 mmHg and that rapidly propagated through the colon on at least three ports.28,29

Intrabag volume variations were visually analyzed. After deleting motion artifacts, phasic volume events defined as rapid changes in the baseline volume ≥10% occurring at a frequency of 1–4 per minute were excluded from the barostat volume analysis according to Von der Ohe et al.30 A ten-minute baseline volumes were averaged over 1-hour periods. In each subject, the mean baseline volume during the 1-hour fasting period represented the basal volume of the colonic segment occupied by the barostat bag. The mean intrabag volumes for the five successive 1-hour periods following the beginning of intracolonic infusion were expressed as a percentage of the 1-hour basal value to correct for interindividual variations in baseline volumes reflecting differences in colonic diameters.

We considered that a motor event coincided with a symptom when it occurred within 5 min of the beginning of the symptom.

Statistical analysis

Results are expressed as median (IQR). Two-way analysis of variance (anova) with treatment and time was used to compare hourly frequencies of each symptom and changes induced in colonic volume, motility index and number of HAPCs. The median total number of HAPCs and the median maximal change in abdominal girth during and after saline and starch infusions were compared with the Wilcoxon’s test. The non-parametric Spearman correlation test was used to study the relationships between different variables. A P value <0.05 was considered to be significant. SPSS version 14 (SPSS Inc, Chicago, IL, USA) was used for statistics.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Location of the barostat bag and of the infusion port

The location of the barostat bag was the ascending or transverse colon (proximal colon) in five cases of saline infusion and six cases of starch infusion, and the descending or sigmoid colon (distal colon) in five cases of saline infusion and four cases of starch infusion. The radiopaque infusion catheter located 17 cm proximal to the barostat bag was in the terminal ileum, ascending or transverse colon. The most distal location was the splenic flexure in one subject receiving starch and in one subject receiving saline. Therefore, all starch infusion occurred in the distal ileum, ascending or transverse colon.

Symptoms and breath excretion of H2 and CH4

Eight of ten participants reported brief symptoms such as borborygmi (= 39) and flatus (= 23) during or after intracolonic infusion of both starch and saline. Bloating was reported in eight of the ten volunteers as a long lasting (more than 10 min) symptom. No volunteer had abdominal pain or mentioned the urge to defecate during starch or saline infusion. Symptoms were always felt as weak and comparable in frequency between starch and saline infusion (= 0.45).

The median maximal change in abdominal girth was 1.4 (0–2.4) cm after starch infusion and 2.5 (0.3–3) cm after saline infusion (= 0.23). The median excess volume of H2E excreted in breath during and after starch infusion was 191.9 (133.9–224.2) mL (Fig. 1). Seven of the ten volunteers receiving starch were CH4 excretors. In four of the ten participants, breath CH4 excretion but not H2 excretion increased during and after starch infusion. In these four subjects, the infusion catheter was located in the mid-transverse (= 3) or splenic flexure (= 1). The preservation of colonic microbiota was confirmed in all of them, since a subsequent lactulose breath test (performed 3 days later) showed an increase in breath H2 > 10 ppm after the ingestion of 20 g lactulose.

image

Figure 1.  Effect of starch infusion on the hourly excess volume of H2 equivalent excreted in breath. Median values are expressed in mL per hour.

Download figure to PowerPoint

Phasic motor activity

Motility indexes during the first hour of recording (basal period) on days 1 and 2 of the experimental procedure were not statistically different [11513 (9123–15887) on day 1 vs 12097 (10353–15744) on day 2; = 0.48]. Saline and starch infusions did not significantly modify the colonic motility indexes compared to the basal value (= 0.8) (Fig. 2). Numbers of HAPCs during the first hour of recording (basal period) was not statistically different on day 1 and 2 of the experimental procedure with 0 (0–2) vs 0 (0–1.25) HAPC/1 h on day 1 and 2, respectively (= 0.71). Compared to saline infusion, starch infusion significantly increased the mean number of HAPCs per hour (= 0.02; anova) (Fig. 3) and the median total number of HAPCs recorded during the 5 h following the beginning of intracolonic infusions [4.5 (2.75–6.5) vs 0.96 (0–2.66)/5 h, starch vs saline respectively; = 0.011]. The increase in HAPCs was found when starch infusion occurred in the terminal ileum or ascending colon as well as in the transverse colon (data shown in Supporting Information, Figures S1 and S2).Characteristics (amplitude, propagation) of HAPCs recorded during and after starch infusion were identical to those recorded during the basal period. The total number of HAPCs increased in all but one subject and did not correlate with the excess volume of H2E equivalent excreted in breath (= 0.3). Retrograde propagating activity was not recorded during the experiments. Figure 4 shows an example of colonic phasic motor activity recorded 1 h after the beginning of starch infusion.

image

Figure 2.  Effect of saline (inline image) and starch (inline image) infusions on the colonic motility index recorded by manometric catheters. Values are expressed as percentages of basal values (means ± SEM). Note that saline and starch infusion did not significantly modify the colonic motility index.

Download figure to PowerPoint

image

Figure 3.  Effect of saline (black columns) and starch (white columns) infusions on the mean number of HAPCs per hour (h) (means ± SEM).

Download figure to PowerPoint

image

Figure 4.  Recordings of phasic motor activity and intrabag volume during starch infusion. Tracings 1–6 refer to perfused catheters located 2, 7 and 12 cm proximal to and distal from the barostat bag. Tracing 7 is the barostat bag volume recording. The 2 arrows show the occurrence of HAPCs.

Download figure to PowerPoint

Tonic motor activity

During the fasting state, the median volumes of the barostat bag were 181.0 (151.5–226.7) mL before saline infusion and 178.8 (139.1–214.6) mL before starch infusion (= 0.6). Compared to basal values, intrabag volumes were not significantly modified during and after starch and saline infusions (= 0.6) (Fig. 5). Figure 4 shows an example of colonic tonic motor activity recorded 1 h after the beginning of starch infusion.

image

Figure 5.  Effect of saline (inline image) and starch (inline image) infusion on the intrabag volume recorded from the barostat. Values are expressed as percentages of basal values (means ± SEM). Note that saline and starch infusion did not significantly modify the colonic tonic motor activity.

Download figure to PowerPoint

Relationship between symptoms and motor events

Twenty six of the 62 reported brief symptoms (16 borborygmi and 10 flatus) coincided in time with HAPCs and 43% of HAPCs were associated with a symptom occurrence. There was no evident relationship between a precise symptom and a motor event. Bloating was a long lasting symptom that could not be related to any phasic or tonic motor events.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Our study found that colonic fermentation of 15 g starch produces an increase in the number of HAPCs in the colon of healthy volunteers. In a previous study we have shown that in physiological conditions the human colonic microbiota is able to catabolize in totality a larger amount—50 g—of the same raw native wheat starch as we used in the present experiment when infused into the cecum of healthy volunteers.31 From other experiments we know that lower amounts (25 g raw wheat starch) could be also fermented when infused in the transverse colon.32 In the present study, intracolonic infusion of 15 g starch produced an increase in H2E excretion in breath which was maximal the third hour after the beginning of starch infusion. Four of ten subjects had an increase in breath CH4 but not in breath H2. As the methanogenic microorganisms are mainly present in the distal colon,24,33 the fermentation of starch in the distal colon could explain that the H2 produced was preferentially consumed by the methanogenic bacterial flora, leading to a high production of CH4 without breath excretion of H2. As a matter of fact in these four volunteers, the infusion catheter was located in the mid-transverse or at the splenic flexure and the preservation of the colonic microbiota was confirmed by a subsequent lactulose breath H2 test showing an increase in breath H2 in all four subjects. According to these results and previous experiments31–33, we can reasonably assume that the infused starch was fermented in all subjects, although we did not measure precisely if starch fermentation was complete.

Intracolonic infusion of 15 g of starch as well as saline solution produced weak symptoms similar in intensity and frequency. The feeling of symptoms in healthy subjects even in the control period may be related to experimental conditions, notably the intubation, or be due to the special attention of volunteers to record symptoms. As the abdominal girth increased during starch and saline infusions, felt symptoms may also be related to the infused volume which was the same with starch and saline solutions. Symptoms did not coincide in time with a specific motor event. We found that colonic motor activity changes produced by starch fermentation consisted in an increase in the number of HAPCs. High amplitude propagated contractions are a short lasting motor event that cannot explain long lasting symptoms such as bloating. Only 26 of 62 reported symptoms as borborygmi and flatus (borborygmi in 16 cases, flatus in 10 cases) coincided in time with HAPCs and less than half of HAPCs were associated with a symptom occurrence.

Therefore, although starch infusion increased the number of HAPCs, symptoms were not significantly enhanced; this may be however related to the limited number of volunteers enrolled in the study. Due to the low number of volunteers, the analysis of association between symptoms and motor events was made by symptom and not by patient. However, none of the volunteers seemed to have an association between a symptom and a motor event.

The increase in HAPCs occurred mainly during the first 3 h after the beginning of starch infusion. During that time the excreted volume of H2E increased in breath, showing that the bacterial fermentation of starch was taken place resulting in acidification of the colonic contents, SCFAs and gas production. The increase in HAPCs was not due to an order effect because the order of starch and saline infusions was randomised and because the motility indexes and number of HAPCs were not different during the basal hour of recording on day 1 and 2 of the experimental procedure. The increase in HAPCs was not related to the volume infused because the same volume of saline did not stimulate the number of HAPCs. It may be argued that saline solution was absorbed by the colon faster than the isoosmotic, isovolumetric starch solution; however, this is not likely as starch is fermented producing SCFAs which are rapidly absorbed;34 moreover, SCFA absorption stimulates water and sodium absorption.34 The increase in HAPCs could be explained firstly by the acidification of colonic contents. We have shown previously that the colonic fermentation of 50 g of the same starch resulted in the decrease in human cecal pH 31 and acidification of the colonic contents by intraluminal infusion of an acidic solution has been shown to stimulate the colonic contractions in guinea-pigs.35 Secondly, it could be related to the increase in colonic SCFA concentrations resulting from starch fermentation.31 The effects of SCFAs on colonic motility are still debated. At low doses, they have a stimulating effect in guinea-pig and rat isolated colon.35,36 Conversely, at high doses, intracolonic infusion of SCFAs has been shown to produce an inhibition of the colonic motility in rats 37,38 but not in dogs.39 Most studies have used a mixture of SCFAs including acetic, propionic, and butyric acids. Lactic acid has also been shown to have an inhibitory effect at high doses in sheep,40 but this result remains to be confirmed in other species.35,36 In any case, starch fermentation by the human colonic microbiota does not result in a high production of lactic acid.31 Thirdly, the stimulation of HAPCs could also be related to the colonic distension resulting from the production of gas during starch fermentation. This has been suggested by Bassotti et al.41 who have shown that colonic distension of the human colon with an intraluminal balloon produces colonic propagating contractions. However, these contractions were qualitatively different from the spontaneously occurring HAPCs and we did not find any correlation between the volume of H2E expired in breath and the magnitude of the increased number of HAPCs. Lastly, the mechanism of the increase in HAPCs induced by starch fermentation could be due to the release of a GI regulatory peptides induced by colonic fermentation such as peptide YY or enteroglucagon.3,4

Regional differences in colonic motility have been described in healthy volunteers.42 With our technique preserving the colonic flora, the exact location of the probe depended on probe migration and could not be controlled. In our study some volunteers had recording in the proximal and others in the distal colon. Statistical comparison between the proximal and distal colon was not performed because it was not planned and because of the small number of volunteers in each group. However, separated analysis of the effect of starch and saline infusion on the number of HAPCs per hour recorded in the proximal vs distal colon suggests that the effect of starch occurred in both part of the colon, with an apparent lower magnitude in the distal colon. Comparisons are however difficult. Indeed, the increase in HAPCs during fermentation of starch was variable among subjects: as the same subject was not perfused with starch in both the proximal and potentially distal colon, difference in the increasing effect on HAPCs may be due to individual differences. The different magnitude in the increasing effect on HAPC number may also be due to regional difference along the colon in the ability to trigger HAPCs during fermentation or finally to difference in the degree or kinetic of starch fermentation between both parts of the colon. We have previously shown that intracolonic fermentation of 15 g lactulose produces a tonic contraction of the colon associated with an acceleration of the transit in the ascending colon but does not increase the number of HAPCs11. Such discrepancy between the disaccharide lactulose and polysaccharide raw wheat starch could be explained by differences in fermentation kinetic and end-products,1 that may differently influence colonic motor activity. Experiments with fecal slurries in vitro as well as with breath H2in vivo have shown that sugars of small molecular size are very rapidly fermented compared with long-chain carbohydrates.43,44

The stimulation of HAPCs by the fermentation of starch could participate in the postprandial increase in HAPCs occurring within 2 h after a meal.45 This increase could indeed result from the fermentation of ileal residue from previous meal propelled into the cecum immediately after meal ingestion.46,47 The increase in the number of HAPCs could accelerate the intestinal transit because of their propulsive effect. Simultaneous recordings of colonic motility and transit have found previously that half of the HAPCs are associated with antegrad movements of the colonic contents and 10% with a movement of all isotopes from the right colon to the rectosigmoid.48 In addition Spiller et al.49 found that almost all HAPCs stimulated by intracolonic oleic acid infusion had a propulsive effect on colonic contents. Thus, we can speculate that the physiological malabsorption of ingested starch by enhancing the number of HAPCs participates in the physiological stimulation of the colonic transit.

Author Contribution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

PJ was involved in the design of the study, manometric recordings, manuscript writing, and accepts full responsibility for the conduct of the study. J-MS was involved in the statistical analysis, manuscript writing. BC was involved in the manometric recordings. ML was involved in the design of the study. RJ was involved in the manometric recordings. BF was involved in the design of the study, manuscript writing. All authors were involved in approving the final draft.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Author Contribution
  8. Conflict of Interest
  9. Disclosure
  10. References
  11. Supporting Information

Figure S1. Mean number of HAPCs/hour recorded in the proximal colon during and after starch (white column) or saline (black column) infusion in the terminal ileum or ascending colon.

Figure S2. Mean number of HAPCs/hour recorded in the distal colon during and after starch (white column) or saline (black column) infusion in the transverse colon.

FilenameFormatSizeDescription
NMO_1652_sm_Figs-S1-S2.doc114KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.