1Gastric mechanics were investigated by categorizing the temporal and spatial patterning of pressure waves associated with individual gastric contractions.
2In twelve healthy volunteers, intraluminal pressures were monitored from nine side hole recording points spaced at 1.5 cm intervals along the antrum, pylorus and duodenum.
3Pressure wave sequences that occurred during phase II fasting contractions (n= 221) and after food (n= 778) were evaluated.
4The most common pattern of pressure wave onset along the antrum was a variable combination of antegrade, synchronous and retrograde propagation between side hole pairs. This variable pattern accounted for 42% of sequences after food, and 34% during fasting (P < 0.05). Other common pressure wave sequence patterns were: purely antegrade - 29% after food and 42% during fasting (P < 0.05); purely synchronous - 23% fed and 17% fasting; and purely retrograde - 6% fed and 8% fasting. The length of sequences was shorter after food (P < 0.05). Some sequences ‘skipped’ individual recording points.
5The spatial patterning of gastric pressure wave sequences is diverse, and may explain the differing mechanical outcomes among individual gastric contractions.
6Better understanding of gastric mechanics may be gained from temporally precise correlations of luminal flows and pressures and gastric wall motion during individual gastric contraction sequences.
As has previously been demonstrated in the oesophagus (Brasseur, 1993; Ren et al. 1993) the spatio-temporal patterns of intraluminal pressure are likely to be important determinants of the mechanical outcome of individual antropyloric contraction sequences. It has been assumed (Meyer, 1987), but never tested by direct measurement, that the aborad propagation of the gastric slow wave pacemaker discharge and any associated muscular contraction in humans results in a consistently aborad pattern of lumen approximation in the distal stomach, which would be reflected in an aborad sequencing of the timing of onset of antropyloric pressure waves. Animal studies, however, have shown that there are substantial differences in the spatial patterning of pressure wave onsets in the antrum and pylorus amongst individual contraction cycles (Anvari et al. 1995). These differences are a plausible basis for the different mechanical results of individual antral contractions (Cannon, 1898; Malbert, Anvari, Jamieson & Dent, 1992).
In this study we have evaluated the spatio-temporal sequencing of pressure wave onset resulting from antral and pyloric contractions in healthy volunteers during fasting and after a meal, in order to determine the range of variation of spatial patterning of lumen approximation amongst individual antropyloric contraction sequences in humans.
Studies were performed on twelve healthy volunteers (10 male, 2 female; age range, 21–52 years; median, 27 years). No subject had a previous history of gastrointestinal disease nor was taking any medication. Two of the volunteers had previously participated in research studies involving nasogastric intubation. Smoking was prohibited from the night before the study. The protocol was approved by the Human Ethics Committee of the Royal Adelaide Hospital and each subject gave written informed consent.
Each study commenced at about 10.30 h following an overnight fast. All measurements were made with the subject supine. The manometric assembly was passed through an anaesthetized nostril and positioned across the pylorus with the aid of feedback from dual point transmucosal potential difference (TMPD) measurements (Heddle et al. 1988a). The manometric assembly was positioned so that the antral TMPD was equal to or more negative than -20 mV, the duodenal TMPD equal to or more positive than -15 mV, and the difference between the two readings was at least 15 mV (Heddle et al. 1988a; Sun et al. 1995). This ensured that the sleeve sensor was astride the pylorus. After the manometric assembly was positioned correctly, a recording of phase II activity of the interdigestive motor complex was made for 15–30 min. This pattern of motor activity was denned as a rate of more than 1, but less than 3, antropyloric pressure waves per minute (Sun et al. 1995). Following completion of the fasting recording, a meal of 100 g cooked minced beef and 150 ml of 10% dextrose (Collins, Horowitz & Chatterton, 1988) was given at about 13.00 h and consumed over a 5 min period. Postprandial recordings were then made for 3 h.
Recording of pressure waves.
The ten-lumen manometric assembly was similar to that used by us in previous studies. It incorporated a 4.5 cm sleeve sensor in parallel with an array of side holes spaced at 1.5 cm intervals (Sun et al. 1995). The side holes at each end of the sleeve recorded intraluminal pressure and TMPD simultaneously (Fig. 1). The side holes along the sleeve allowed discrimination between pressure waves isolated to the pylorus (IPPWs; Edelbroek et al. 1994; Sun et al. 1995) and short antropyloric pressure waves (APPWs). Manometric side holes sampled gastric antral pressures 7.5, 6.0, 4.5, 3.0 and 1.5 cm orad to the antral TMPD port. All manometric lumina were perfused with degassed distilled water at 0.3 ml min−1 except the TMPD lumina, which were perfused with degassed normal saline at the same rate. Manometric pressures were detected by transducers (Transpac, Abbot Critical Care Systems, North Chicago, IL, USA) coupled to a sixteen-channel Polygraf (Synectics, Stockholm, Sweden). The amplified signals were digitized at ten samples per second (NB-MI016, National Instruments, TX, USA). Digitized values were stored in an Apple Macintosh IIci computer using software (MAD 16, C. Malbert/Royal Adelaide Hospital/Synectics) based on Lab View (National Instruments). The data were then analysed off-line on the computer screen by two investigators (W.M.S. and G.S.H.), using Acqknowledge (Biopac Systems Inc., Goleta, CA, USA).
Analysis of pressure waves
Recordings were only analysed for the periods when the TMPD measurements satisfied the criteria given above for correct position of the catheter.
This analysis had five distinct steps which allowed recognition and classification of the spatial patterns of the pressure wave sequences. Firstly, gastric pressure waves were identified. Recordings from individual side holes were examined for pressure waves that were greater than 10 mmHg above end-expiratory gastric pressure, lasted between 5 and 25 s, and were not ascribable to passive intra-gastric transmission of changes of intraperitoneal pressure due to events such as straining or respiration (Heddle, Dent, Toouli & Read, 1988b).
Secondly, the location of the side hole recording point on which a pressure wave was seen was defined as precisely as possible relative to the pylorus. This was done on the basis of the pressure wave patterns recorded by the side hole array that spanned the distal antrum, pylorus and duodenum. The use of dual channel TMPD recording (see above) ensured that the side hole chain could be kept consistently astride the pylorus. Judgements of the position of the pylorus within the span of the 4.5 cm long sleeve were updated continuously, to allow for the mobility of the pylorus relative to the recording assembly. The side hole closest to the transition of motility from antral to duodenal patterns within the span of the sleeve was defined as the most distal antral side hole. When an individual pressure wave sequence lacked features that allowed this judgement, the pyloric position for that wave was taken as that which had been defined for the last pressure wave sequence with diagnostic features.
Thirdly, pressure waves were defined as temporally related or unrelated to pressure waves recorded from other antral manometric side holes. Temporal relationships were defined solely from the time of onset of the pressure wave, which was taken as the time at which the tangent of the major upstroke of the pressure wave bisected the intragastric pressure baseline (Dodds et al. 1972). The time of occurrence of the peak of the pressure wave was therefore not considered in this analysis. The judgement as to whether waves recorded by different side holes were associated in time started by identifying the most orad antral side hole in which a wave was identified. This was defined as the reference side hole for a particular sequence. Any pressure wave recorded by the adjacent aborad side hole was scored as associated if its onset was within 5 s before to 10 s after that of the reference side hole. This analysis then proceeded down the chain of side holes, using the same time window. In some clusters of pressure waves, one or more side holes were skipped. When one side hole was skipped, a pressure wave that occurred in the next aborad side hole was defined as associated if its onset occurred between 6 s before and 11s after the onset of the pressure wave in the orad side hole. If two side holes were skipped, the time window relative to the wave recorded at the orad side hole was 7 s before to 12 s after.
When a cluster of pressure waves recorded on two or more side holes were defined as temporally related, it was defined as a pressure wave sequence. Once a pressure wave recorded on a particular channel was classified as belonging to a sequence, it was excluded from any further definition of relationship to another pressure wave. Pressure wave sequences could therefore include any combination of side holes and did not necessarily have to reach the most distal antral side hole or pylorus. Any pressure wave that did not fall within the definitions of association was categorized as an isolated pressure wave.
Fourthly, the spatial patterning of the time of onset of individual pressure waves that made up pressure wave sequences was defined. This analysis was started from the side hole that recorded the wave with the earliest time of onset for each pressure wave sequence, regardless of its position in the antropyloric region (Figs 1 and 2). The direction of propagation within each pressure wave sequence was then determined. The relationship between two side holes was scored as synchronous (s) if the time difference between the onset of an adjacent pair of waves was 1 s or less (Fig. 2). If this interval was more than 1 s the pattern of propagation was scored as either antegrade (a) or retrograde (r), according to the relative position of the recording points. Thus, a sequence of waves recorded over five side holes could have been scored aaaa, aasa, assr, aass, or any other of the eighty-one combinations of the three lower case letters that coded for antegrade, retrograde or synchronous time relationships between each side hole. Pressure waves were scored as isolated antral pressure waves (IAPWs) if no pressure waves occurred in any other antral side hole within 5 s before to 10 s after the onset of the wave being evaluated (Fig. 2). Isolated pyloric pressure waves (IPPWs) were defined as pressure waves that were detected by the sleeve sensor and no more than one side hole within the sleeve length (Heddle et al. 1988b). To be scored as IPPWs, the onset of these waves had to occur in the absence of an associated pressure wave of any magnitude ascribable to either gastric or duodenal contraction from 5 s before to 10 s after the onset of the wave being evaluated.
Fifthly, the patterns of pressure wave sequences were grouped into major categories. This was necessary because of the very wide variation of patterns of extent and propagation. This simplified description of the spatial pattern of pressure wave sequences does not classify extent, and merely describes the patterns of propagation that occurred within the sequence. Capital letters are used for this condensed classification. Thus, an aaaa or aa sequence was coded as A, an aasa sequence as AS, a sara sequence as SAR, and so on. The order of the capital letters coded for the order of first occurrence of the three propagation categories from proximal to distal antrum.
Data were analysed using analysis of variance (ANOVA). For the postprandial data, consecutive 60 min time intervals were compared. Fisher's exact test was employed to compare the proportion of particular wave sequence patterns amongst subjects. P < 0.05 was considered significant in all analyses.
Fasting pressure wave sequences
The recording assembly was correctly positioned on the basis of TMPD criteria for 100% of the recording time, which totalled 303 min. During this time, thirty-nine isolated pyloric, and sixty-nine isolated antral pressure waves were recorded. The isolated antral pressure waves tended to be seen in a particular side hole in a given individual.
Spatio-temporal patterning was analysed in the 221 pressure wave sequences that involved at least two manometric side holes. Of these pressure wave sequences, ninety-three (42%) were completely antegrade (category A) - that is, they showed a sequential onset time of pressure waves along the antrum. The second most common broad category was the seventy-five (34%) sequences which had a mixture of antegrade, retrograde or synchronous wave onsets in different parts of the antrum (categories AS, AR, SA and other combinations; Figs 2 and 3). The other common patterns of onset along the antrum were thirty-six (17%) purely synchronous sequences (category S) and seventeen (8%) purely retrograde sequences (category R; Figs 2 and 3). The percentages of the different major categories of waves are shown in Fig. 3.
The spatial extent over which each classification of the pressure wave sequence could be identified is shown in Fig. 4. The majority (63 %) of completely antegrade pressure wave sequences commenced from one of the three most proximal antral side holes. Most of the pressure wave sequences extended to the pyloric side hole. Of the completely synchronous pressure sequences, 75% extended to the pyloric side hole, and 12% to the side hole just aborad to the pylorus (Fig. 4). Pressure wave sequences that were initially antegrade and then either synchronous or retrograde, had an extent similar to completely synchronous sequences. The majority (97%) of retrograde pressure wave sequences had a short span commencing in the distal antrum or pylorus, usually at either the last antral side hole or the pyloric side hole.
Postprandial pressure wave sequences
The manometric assembly was correctly positioned for 95% of total postprandial recording time, which amounted to 2057 min. There were 409 isolated pyloric and 318 isolated antral pressure waves.
The 778 pressure wave sequences that involved at least two nianometric side holes were analysed for propagation patterns. A purely antegrade pattern of wave onset occurred in only 228 (29%) of the recorded sequences (category A). There were 333 (42%) sequences that had complex spatial patterns of pressure wave onset time (categories AS, AR, SA and others combined; Fig. 3). Purely synchronous sequences (category S) accounted for 181 (23%) and retrograde sequences (category R) for forty-six (6%) of those recorded.
Comparison of patterns of fasting and fed pressure wave sequences
Purely antegrade pressure wave sequences were less frequent (P < 0.05) after the meal than when fasting (Fig. 3).
The proportions of purely retrograde and synchronous pressure wave sequences did not differ significantly between the postprandial and fasted states (Fig. 3). The proportion of sequences with complex patterns of pressure wave onset time were similar before and after a meal except that antegrade-synchronous sequences (category AS) were more frequent (P < 0.05) after the meal.
The spatial extent of sequences varied between the fed and fasted states. In the fed state, almost all antegrade sequences commenced in the most distal antral side holes, and had a smaller extent than those during fasting (P < 0.05; Fig. 4). Similarly, more AS and AR sequences commenced at the distal antral side hole (P < 0.05; Fig. 4). Almost all (87%) of the AS and AR sequences extended to the pyloric side hole.
Although the number of isolated antral pressure waves was similar before and after a meal, more of these occurred at the most distal antral side holes postprandially (P < 0.05) whereas during fasting, the majority (P < 0.05) occurred at the most proximal antral side holes. The number of isolated pyloric pressure waves was greater (P < 0.05) after the meal than when fasting.
To our knowledge this is the first study which has evaluated the spatial and temporal organization of the pressure waves generated by antropyloric phasic contraction sequences in humans. Our observations indicate that individual antropyloric contractions result in diverse spatial patterns of antral pressure waves. In particular, consistently aborally directed pressure waves accounted for only a minority of the sequences recorded both during fasting and after food. While there were some differences in the proportions of different patterns of pressure wave sequences between the fasting and the fed states, there was a more impressive difference in the extent of antral pressure wave sequences, with a greater length of antrum showing pressure waves during fasting. The description of the spatio-temporal patterning of antropyloric pressure wave sequences in our study complements previous fluoroscopic and manometric studies of the spatial patterns of isolated pyloric pressure waves (Heddle et al. 1988b; Tougas, Anvari, Dent, Somers, Richards & Stevenson, 1992). The concept that variation in the timing of closure of the pylorus could be a major determinant of the mechanical outcome of antropyloric contractions is not novel, having been proposed by Cannon on the basis of purely fluoroscopic observations (Cannon, 1898). We have also shown previously in the pig that the considerable variations in the timing of pyloric closure in relation to antral pressure waves are associated with substantial differences in patterns of transpyloric flow (Anvari et al. 1995).
We do not believe that the pressure wave patterns we have recorded reflect mechanical distortions of the manometric assembly. Manometric assemblies have been observed in the antropyloric region of the stomach of dogs and humans using fluoroscopic (Carlson, Code & Nelson, 1966) and ultrasound (Hveem, Sun, Horowitz & Dent, 1995) techniques and no deformity has been seen, even during vigorous contractions; in our studies the position of the manometric assembly relative to the pylorus was also monitored by dual point transmucosal potential difference recordings (Heddle et al. 1988a), and the position of the assembly stabilized by a ‘tail’ that passed around the duodenal loop. It should also be recognized that perfused side hole manometric assemblies tolerate very acute angles of flexion without an effect on baseline pressure, and are substantially superior to most intraluminal transducer assemblies in this regard.
The variation of spatio-temporal patterning of antropyloric luminal approximation that we have observed in the present study may appear at first to be difficult to reconcile with the orderly aboral progression of antral contraction which is determined by propagation of the gastric slow wave (Szurszewski, 1987). This is not a real inconsistency, as there is a very important difference between the time of onset of a contraction, and the time of onset of a localized intraluminal pressure wave generated by luminal approximation (Dent, Sun & Anvari, 1994). A local pressure wave is only detectable when the contraction-induced antral muscle shortening is close to maximal, so that the lumen at that site is approaching occlusion. Thus, there is a variable time lag between contraction onset and luminal approximation (Brasseur, 1993). If the variation in the time of onset of pressure waves is occurring in the face of a stereotyped aborad progression of the onset time of antral contractions, the observed diverse sequencing of lumen approximation for a particular recording position in the antrum is difficult to explain on any basis other than modulation of the vigour of contraction of individual antral segments. In this way, aboral progression of contraction onset is compatible even with a retrograde sequencing of luminal approximation. This proposal implies a sophisticated segmental control of at least the strength of antral contractions, and there is evidence that local neural and central nervous system controls of antral contraction are capable of this (Dent et al. 1994). The degree of antral filling would also be expected to influence the mechanics of gastric contraction. The present study showed that the extent of antropyloric pressure waves after the meal was less for all patterns of sequences. We cannot determine how much of this difference was directly due to passive alteration of mechanics by antral filling, as opposed to feedback modulation of gastric contractions by small intestinal receptors (Horowitz et al. 1994). Furthermore, while there is undoubtedly a direct effect of the degree of antral filling on the mechanics of antropyloric contractions and the associated pressure sequences, this interaction should not be taken to exclude the simultaneous existence of (presumably) neural modulation of patterns of antropyloric luminal approximation, independently from any direct effect from the degree of antral filling.
It was necessary for us to devise analysis criteria that would permit description of the spatio-temporal patterning of antropyloric pressure wave sequences. The criteria were chosen in the light of knowledge about gastric motor function, with the assistance of insights into the mechanics of oesophageal motor function, and the hydraulics of luminal flows. The judgement whether pressure waves recorded in different channels were associated or not (that is whether they formed a sequence) used relatively arbitrary values, but these were themselves based on the normal frequency and pattern of propagation of the gastric pacemaker potential in humans. The differences of pressure wave onset times between recording side holes that we used for distinguishing synchronous from antegrade and retrograde pressure wave propagation proved practical, but whether these criteria are optimal in defining mechanical significance cannot be judged from the available data. Our analysis also required that we develop a precise terminology for the events that we analysed. For reasons discussed above, it is important to note that ‘contraction’ and ‘pressure wave’ describe different mechanical events which may have variable relative timings and even frequencies, so these should not be used as synonyms. For the oesophagus there is a well developed concept of a temporally related sequence of pressure waves that occurs along the length of the oesophagus (Brasseur, 1993). This concept is less well defined for the antropyloric region. The pressure patterns associated with antropyloric contractions make it more difficult to define what is a sequence in the antropyloric region, especially when contractions are not being measured directly. Our intention was to define a pressure wave sequence as a group of pressure waves recorded at multiple points along the antropyloric region that could be reasonably attributed to the progression of a single contraction along this region. We have used the term ‘luminal approximation’ to describe the period during which a contraction of the antropyloric region results in a pressure wave with amplitude or time values that are unique to one recording point. This term embraces the period when a contraction is displacing fluid from around the recording side hole, and when the contraction has actually occluded the lumen at the level of the recording side hole. Mechanical analysis predicts that both of these situations are associated with local increases of pressure, a prediction that has been confirmed by concurrent manometry and fluoroscopy in the oesophagus (Brasseur, 1993).
We believe that the variations in the space/time patterning of antropyloric pressures that we have observed are mechanically significant. These patterns can only be attributed to differing spatio-temporal sequencing of active antral luminal approximation. The physical laws that govern the relationship between movement of liquids and pressure gradients predict that propagating pressure waves of detectable amplitude will not be evident until the lumen is significantly narrowed (Brasseur, 1993). The observation that individual phasic antropyloric contractions have a range of outcomes on transpyloric flow, which vary from being a high stroke volume to pure retropulsion without any forward flow within a single gastric contraction (Malbert & Mathis, 1994; Anvari et al. 1995), is consistent with the variability in patterns of lumen occlusion which we have demonstrated. It is likely that complex common cavity pressures (Anvari et al. 1995) occur in the distal stomach due to trapping and compression of the lumen content, but these were not evaluated in our study.
Critical evaluation of our hypothesis demands further studies to correlate spatial patterns of antropyloric pressures with patterns of luminal flow, but it should be recognized that observations on the oesophagus have shown that the spatio-temporal pattern of onset of lumen approximation is the primary determinant of the luminal movement of liquid (Brasseur, 1987; Ren et al. 1993). We elected not to analyse the timing and magnitude of the peak of antropyloric pressure waves, in the belief that these variables are likely to be of less importance in determining antropyloric mechanics than the sequencing of pressure wave onset times (Ren et al. 1993; Dent et al. 1994). Our thinking on this has also been guided by studies in the oesophagus that show that the timing and magnitude of the peak of the peristaltic pressure wave have relatively little impact on patterns of liquid oesophageal transit, provided the peristaltic wave reaches a threshold of amplitude (Kahrilas, Clouse & Hogan, 1994).
Understanding of the functional significance of the diverse patterns of antropyloric pressures observed in the present study is a substantial challenge that is likely to be advanced by spatially and temporally precise simultaneous monitoring of luminal flows and pressure, preferably with concurrent detailed observations of gastric wall motion. Gastric contractions and their associated pressure sequences need to be studied individually at first, given the diversity of observed patterns, but the findings of the present study suggest that it may be possible to group gastric pressure wave sequences into several major categories that would predict patterns of transpyloric flow. Ultimately, such an analysis may advance the understanding of the mechanics of abnormal gastric emptying (Horowitz & Dent, 1991).
This work was supported by the National Health and Medical Research Council of Australia.