Spatio-temporal analysis reveals aberrant linkage among sequential propagating pressure wave sequences in patients with symptomatically defined obstructed defecation

Authors


  • Supported by: NH&MRC Australia.

  • This work has been presented in abstract form at the International Motility Society, Jeju, Korea, September 2007. Dinning PG, Szczesniak MM, Cook IJ. Aberrant spatio-temporal linkage among sequential colonic propagating sequences exists in severe constipation. Neurogastroenterol Motil 2007; 19: A71 (Poster presentation).

Dr Phil Dinning, Department of Gastroenterology, St George Hospital, Kogarah, NSW, 2217, Australia.
Tel: +61 2 93502817; fax: +61 2 93503993; e-mail: p.dinning@unsw.edu.au

Abstract

Abstract  Available evidence implicates abnormal colonic contractility in patients suffering from constipation. Traditional analysis of colonic manometry focuses on the frequency, extent and amplitude of propagating sequences (PS). We tested the hypotheses that the spatio-temporal linkage among sequential PSs exists throughout the healthy human colon and is disrupted during constipation. In eight patients with severe constipation and eight healthy volunteers, we recorded colonic pressures from 16 regions (caecum–rectum) for 24 h. Sequential PSs were regionally linked if the two PSs originated from different colonic regions but the segments of colon traversed by each PS overlapped. In order to determine whether this linkage occurred by chance, a computer program was used to randomly rearrange all PSs in time. Data were re-analysed to compare regional linkage between randomly re-ordered PSs (expected) and the natural distribution of PSs (observed). In controls the observed regional linkage (82.5 ± 9.0%) was significantly greater than the expected value (60.5 ± 4.3%; P = 0.0001). In patients the observed and expected regional linkage did not differ. The (observed − expected) delta value of regional linkage in controls was significantly greater than in patients (21.7 ± 8.5%vs−2.3 ± 7.0%; P = 0.01). Regional linkage among sequential PSs in the healthy colon appears to be a real phenomenon and this linkage is lost in patients with constipation. Regional linkage may be important for normal colonic transit and loss of linkage might have pathophysiological relevance to and provide a useful biomarker of severe constipation.

Introduction

One of the most readily apparent manometric colonic motor patterns is the propagating pressure wave sequence (PS). This motor pattern is an essential component of normal colonic transit1,2 and stool expulsion.3,4 Available evidence implicates abnormalities in the frequency, amplitude and extent of propagation of these motor patterns in patients with severe constipation.4–7 However, these traditional criteria, used to differentiate health from constipation, are likely to only partly explain the delayed transit and or disordered defecation displayed by these patients. For example, when compared with controls, patients with slow-transit constipation demonstrate fewer high-amplitude PSs (absent in 30%)5,6,8,9 and yet many such patients still generate a normal frequency of PSs of lesser amplitude.10 Indeed meal-stimulated colonic movement of isotope is preserved in at least 50% of individuals who lack high-amplitude PSs11 and propagating sequences with quite low amplitudes are still capable of moving content in the healthy colon.1,2

We have previously reported colonic pressure patterns in constipated patients with symptomatically defined obstructed defecation.4 These patients demonstrated a normal, or in some cases increased, frequency of PSs and high-amplitude PSs and yet a majority of these patients had delayed left colonic transit. Interestingly these patients also displayed a disruption of the normal, and very obvious, stereotypical pattern of linkage among sequential PSs in the lead up to defecation.3,4 The presence of a similar linkage among sequential PSs at times other than the immediate pre-defecatory period is not immediately apparent and has not been examined systematically in health or constipation. It is very difficult to appreciate the phenomenon of any potential linkage among sequential colonic motor events occurring outside the pre-defecatory phase because, in contrast to the closely spaced grouping of PSs seen immediately prior to defecation, minutes or hours may elapse between sequential PSs at other times of the day. Recently, we developed a technique which enabled as to condense and display each PS from 24 h of pan-colonic pressure recordings in a 3D, spatio-temporal colour plot.12 This condensed representation of colonic pressure patterns is a powerful tool with which to explore the linkage, if any that may exist among sequential PSs throughout the day, as well as any disturbance in the phenomenon in patients with severe constipation.

The aim of this study was to examine the presence, if any, of the prevalence and extent of linkage among sequential propagating pressure wave sequences (PS) in the healthy and constipated human colon. We hypothesized that linkage among sequential propagating sequences: (i) exists throughout the normal healthy colon throughout the day; and (ii) is disrupted or absent in the constipated colon.

Methods

Patients and controls

We analysed colonic pressure data from eight healthy volunteers (three female; mean age: 24 ± 2 years) and from eight patients with severe constipation and symptomatically defined obstructed defecation (six female, mean age 50 ± 13 years). Data from these patients and controls formed part of a larger cohort from which data regarding regional variation, frequency, extent and amplitude of PSs had already been published.4 For this study we have included only those subjects for which we had obtained pan-colonic (cecum-rectum) manometry for the entire 24 hr recording period. Pan-colonic manometry was essential for the linkage analysis outlined below.

The patients included in that study were selected when three of the following symptoms were present: (i) an inability to initiate defecation following the urge to do so, or difficulty with stool evacuation; (ii) excessive straining at stool more than 25% of the time or self-digitation to facilitate defecation more than 25% of the time; (iii) a feeling of incomplete evacuation after defecation.4,13–18 All patients who met the symptomatic criteria underwent a physical and rectal examination, rectal sensory assessment by simple volumetric balloon distension,19 evacuation proctography, and measurement of colonic transit by isotope scintigraphy.20 Patients were excluded from consideration for study if they had anal stenosis, or any major mechanical obstructive features (i.e. large rectocoele with confirmed retention of contrast, full rectal prolapse or occlusive intra-anal intussusception). Delayed transit was defined as greater than 9% isotope retention at 72 h.20 All the patients included in this study demonstrated delayed colonic transit. The main site of retention was in the left colon in four patients (splenic flexure to recto-sigmoid) and rectosigmoid delay was displayed in the remaining four. All patients had normal ano-rectal manometric studies, with no evidence of paradoxical sphincter contraction.4 There was no significant difference in rectal sensory thresholds between patients and controls.19

All participants had given written, informed consent and the studies were approved by the Human Ethics Committees of the South Eastern Area Health Service, Sydney and the University of New South Wales.

Colonic manometric technique and experimental protocol

The study design and recording techniques have been described previously in detail.3,21 Briefly, we used a 21-lumen (16 recording sideholes) 4.5 m long extruded silicone-perfused manometric assembly (Dentsleeve, Wayville, SA, Australia), with an overall diameter of 3.5 mm. Each recording lumina had an internal diameter of 0.5 mm with an inter-sidehole distance of 7.5 cm. A silicon balloon at the tip of the catheter could be inflated and deflated with water through the core channel (internal diameter 0.8 mm). The catheter was rendered radio-opaque with a barium core. In healthy volunteers and patients, the catheter was placed using the technique of naso-colonic intubation. The selection of this antegrade technique instead of the more commonly used colonoscopically assisted technique has been detailed in a previous publication.4,21 A detailed explanation of the nasocolonic placement has also been described previously.4,21 Twenty-four hour colonic manometry was recorded in a private room for a minimum of 24 h and all participants ate calorie-controlled meals.4,21

Data analysis

Definition and identification of PS  For the purpose of analysis the colon was divided into 16 regions (region 1 = cecum, region 4 = hepatic flexure, region 8 = splenic flexure, region 12 = proximal sigmoid colon, region 16 = rectum). Recording sideholes were assigned to the colonic region within which they lay. Allocation of sideholes to colonic regions was confirmed by X-ray and has been detailed previously.3,21 Definitions of propagating activity have been described in detail previously.3,21 Briefly, a PS was defined as an array of three or more pressure waves recorded in adjacent recording sites in which the conduction velocity between wave onset within that sequence lay between 0.2 and 12 cm s−1. Propagating sequences were further qualified by the terms ‘antegrade’ or ‘retrograde’, depending upon the direction of propagation. A propagating sequence was classified as a high amplitude propagating sequence (HAPS) if the amplitude of at least one component propagating pressure wave was >116 mmHg (derived from normal mid-colonic mean amplitude + 2SD).21 Propagating sequences which propagated more than 30 cm are highly associated with luminal propulsion and defection in health.1,3 Hence, PSs with component pressure waves identified in five or more colonic regions were also flagged and their frequency was compared between patients and controls.

Definition of regional linkage among PS  The next stage of analysis was to examine the relationship among sequential PSs throughout the colon over the 24-h period. We have coined the term ‘regional linkage’ to help define this relationship. A PS was deemed regionally linked to the PS immediately preceding it if the two PSs originated from different colonic regions but the segments of colon traversed by each PS overlapped (Fig. 1). Propagating sequences that propagate in an antegrade direction were analysed independently of those that propagated in a retrograde direction. Regional linkage was first defined between sequential antegrade PSs and then between sequential retrograde PSs. Finally regional linkage between all sequential PSs, regardless of polarity was determined.

Figure 1.

 Examples of regional linkage between sequential propagating sequences in the (A) observed data and (B–D) expected (re-ordered) data. The ridges within each map represent individual antegrade propagating sequences. The start of each ridge indicates the site of origin and the time of day the PS occurred. The length of the ridge indicates the extent of propagation. The shading of green within the ridge indicates the amplitude of the component pressure waves. A ✓ indicates regional linkage between sequential PSs, a × indicates a break in regional linkage. In the observed data (top) the PSs are numbers from 1 to 13 and 92% of these PSs are marked as regionally linked. Maintaining the site of origin and extent of propagation, the PSs are randomly re-ordered throughout time. Regionally, linkage in the resultant expected data (B, C and D) is reduced to 75%, 67% and 50%, respectively.

To determine whether the time interval between sequential PSs influenced regional linkage, we batched candidate PSs according to five time epochs: (A) 0–15 min; (B) >15 to ≤30 min; (C) >30 to ≤60 min; (D) >60 min; (E) regional linkage without regard to time interval.

Regional shift in the site of origin of PS  We determined the extent of the regional shift in the site of origin from one PS to next as follows. The first PS in each manometric trace was labelled the ‘primary PS’. The distance, measured in colonic regions (see section ‘Definition and identification of PS’ above), between the site of origin of the primary PS and the site of origin of the subsequent PS was measured. Moving chronologically through the manometric trace, each subsequent PS became the primary PS and was related to the PS immediately succeeding it throughout the entire 24-h period. A positive regional shift in site of origin of PS had a possible score between 1 and 13 colonic regions and indicated an orad movement in the site of origin between the primary PS and the subsequent PS (Fig. 1). An aborad shift in site of origin of adjacent PSs had a possible score between −1 and −13 (Fig. 1). Primary and subsequent PSs originating from the same colonic region had a zero score. The frequency distribution curve of regional shift between sequential PSs was plotted for both patients and controls.

Does regional linkage occur by chance?  The possibility exists that adjacent PSs may simply overlap and have no systematic or physiological relationship to each other. Hence, in this part of the analysis we tested the hypothesis that the spatial and temporal distribution of PSs occurred purely by chance. While maintaining the original site of origin and extent of propagation of each individual PS, a program written within our department in Microsoft Excel XP (Microsoft Corporation) and Visual Basic 6.3 (Microsoft Corporation) (available upon request) redistributed the time of onset of each PS randomly throughout the 24-h period (Fig. 1). The data for each subject were reordered on 50 occasions. Each set of reordered data was analysed, as described above, to test whether the expected (reordered PSs) regional linkage and frequency distribution of regional shift between sequential PSs differed from the observed (real) values.

Statistical analysis

Comparisons between the observed and expected regional linkage within patients and within controls were performed using chi-squared analysis. Comparisons between the delta values of the regional linkage (observed values − expected values) between controls and patients were performed using the Wilcoxon signed ranks test. The same non-parametric test was used to examine the differences between the frequency of antegrade and retrograde PSs between patients and controls. A repeat-measures anova was used to compare regional linkage in different time epochs between patients and control. Chi-squared analysis was used to determine whether sequential PS originating from the same colonic region differed between patients and controls. Kolmogorov–Smirnov comparison of two data sets was used to determine any significant difference in the delta distribution of regional shift between sequential propagating sequences between patients and control.

Results

Propagating sequence characteristics

The frequency of PSs propagating in an antegrade direction was higher in patients (154 ± 91 PS/24 h) than it was in controls (65 ± 8 PS/24 h; P = 0.04; Figs 2 and 3). The extent of propagation of PSs that extended through three to five colonic regions (≈15–30 cm) was significantly greater in patients (131 ± 82 PS extend extent <30 cm/24 h) than in controls (42 ± 14; P = 0.02). However, the frequency of PSs extending >30 cm was similar between the two groups (controls 30 ± 8 vs patients 25 ± 14; NS).

Figure 2.

 Space–time pressure mapping of colonic propagating sequences (PS). Antegrade PSs (green) originate at the orad end of the ridge and retrograde PSs (red) originate at the anal end of the retrograde ridge. A 24 h pan-colonic map of PS activity is shown in health (Top A) and a patient (Top B). A 6 h time window has been selected from each 24-h maps (Bottom A & B). The white hatched arrows join the site of origin of regionally linked propagating sequences. An increased frequency of both antegrade and retrograde propagating sequences is clearly visible in the patient spatiotemporal maps. However, the regional linkage between propagating sequences is severely disrupted.

Figure 3.

 Average frequency (vertical bars) and extent of propagation (horizontal bars) of propagating sequences originating from each of the defined colonic regions in healthy controls and patients. The frequency of antegrade propagating sequences is significantly greater in patients than controls (P = 0.04). However, the extent of propagation of PSs originating in the ascending or transverse colon is significantly reduced in patients (P < 0.001).

The frequency of PSs propagating in an retrograde direction was also higher in patients (104 ± 120 PS/24 h) than it was in controls (17 ± 10 PS/24 h; P = 0.01; Fig. 2). However, unlike antegrade PSs, the majority (93%) of all retrograde PSs in patients and control extended between three and five colonic regions (≈15–30 cm).

The HAPS frequency was similar between the groups (patients, 12 ± 9 HAPS/24 h vs control 9 ± 6 HAPS/24 h). However, the HAPSs in patients were of a significantly shorter extent (36 ± 14 cm vs 55 ± 15 cm; P = 0.04) and lower amplitude (78 ± 24 mmHg vs 102 ± 16 mmHg; P = 0.04).

Regional linkage among PS

Antegrade PS  In controls the observed prevalence of regional linkage (original data) (82 ± 9%) was significantly greater than the expected (re-ordered data) prevalence (60 ± 4%; P = 0.0001). This indicates that the phenomenon of regional linkage in healthy controls does not simply occur by chance. In contrast, the observed regional linkage in patients (41 ± 8%) did not differ from the expected prevalence of linkage (44 ± 6%) indicating that the temporal distribution of PSs in patients may be simply random. The delta value of linked PSs in controls (observed − expected) was significantly greater than that measured in patients (21 ± 9%vs−2 ± 7%; P = 0.008). This indicates that regional linkage among adjacent PSs occurs significantly less frequently in patients when compared with controls.

Regional linkage among PSs in controls was not dependent upon the time interval between sequential PSs. Similarly, the diminished regional linkage in patients was not influenced by time at any of the time epochs examined. Across all time intervals regional linkage was significantly more prevalent in controls when compared with patients (P < 0.001).

Retrograde PSs   In controls the observed prevalence of regional linkage among retrograde PSs (36 ± 21%) did not differ from the expected value (36 ± 16%). Similarly, no differences were detected in the patient population between observed (42 ± 8%) and expected (43 ± 6%) values. The prevalence of regional linkage among retrograde PSs did not differ between controls and patients. This indicates that regional linkage between sequential retrograde PSs in controls and patients may be random.

Regional linkage between antegrade and retrograde PSs   The relationships between antegrade and retrograde PSs did not differ between controls and patients or between observed and expected data values in either group.

Regional shift in site of origin between sequential PSs

Antegrade PSs  In controls, primary and subsequent PSs originated from the same colonic region in 6% of occasions while the expected prevalence of this occurrence was 13% (P = 0.01), indicating that this phenomenon was not a random occurrence (Fig. 4A). In contrast, sequential PSs originated from the same colonic region in patients in 20% of occasions which was significantly greater than the expected 13% (P = 0.02; Fig. 4B).

Figure 4.

 The distribution of regional shift between sequential propagating sequences (PS) in (A) healthy controls and (B) patients. A regional shift of 0 indicates sequential PSs originating from the same colonic region. A positive regional shift indicates an orad move and a negative regional shift an aborad move, in the site of origin between sequential PSs. In controls a delta of 0 was significantly (P = 0.01) less likely to occur in the observed compared with the control expected data. In contrast the observed data in patients was significantly (P = 0.02) more likely to record a delta of 0 than in the patient expected data. (C) The delta distribution (observed − expected) of the regional shift in site of origin of sequential PSs differed significantly between controls and patients (P = 0.006). The control data display a bimodal distribution with peak frequencies at a regional shift of +2 and −2, indicative of a subtle shift in the site of origin between sequential PSs.

The expected and observed frequency distribution of the regional shift in the site of origin between sequential PSs was plotted for both controls and patients (Fig. 4A,B). The delta (observed − expected) distribution curve of regional shift in site of origin between sequential PSs in controls was then compared with the delta distribution curve in patients (Fig. 4C). These two distributions differed significantly from one another (Kolmogorov–Smirnov, P = 0.006). The delta curve in controls indicates a bimodal distribution with peak frequency at ±2 colonic regions. This suggests in health, but not in constipated patients, that there is a subtle shift in either an orad or aborad direction between sequential PSs.

Retrograde PSs  In the control group, on 31% of occasions, primary and subsequent retrograde PSs originated from the same colonic region. This did not differ from the 24% scored in the expected data (Fig. 4A). In patients sequential PSs originating from the same colonic region was significantly more likely to occur in the observed data in comparison with the expected data (22%vs 13%; P = 0.02). As with the patient antegrade data this indicates that that sequential retrograde PSs originating from the same colonic region occurred more frequently than would be expected by chance. The delta frequency distribution curves between controls and patients were comparable. Therefore, the apparent regional shift that exists between sequential antegrade PSs is not evident between sequential retrograde PSs.

Discussion

This study confirms the spatio-temporal linkage among sequential antegrade but not retrograde PSs in the healthy colon and that this physiological linkage is lost in severe constipation. The regional linkage pattern in health is characterized by a shift in the colonic site of origin from one PS to the next and the segments of colon traversed by sequential PSs overlap. In contrast, coordination between sequential antegrade PSs in this group of patients with constipation, with symptoms of obstructed defecation and delayed colonic transit, is severely impaired. This observation holds true even though these patients demonstrate an increased frequency of PSs when compared with healthy controls. The constipated colon demonstrates a spatio-temporal patterning in which numerous sequential PSs originate from the same colonic region. That is, the likelihood of both a spatial shift in the site of origin and of overlap in the segments of the colon transversed by those PSs is reduced.

The pathophysiological mechanisms for delayed transit in constipation syndromes are only partially understood. It is known that PSs are responsible for stepwise movement of colonic content over short or long distances.1,2 In the lead up to actual defecation, there is an obvious spatio-temporal linkage among multiple PSs.3,4 This observation, together with the remarkable lack of linkage among sequential PSs in the present study, suggests that the loss of this spatio-temporal linkage may be a significant contributor to the delayed colonic transit as well-disordered stool expulsion in these patients.4

The cause of the breakdown in regional linkage in these patients is unknown. It could be argued that diminished regional linkage is a secondary phenomenon to the retention of stool. This however is unlikely because we have recorded the same breakdown in regional linkage in patients suffering from constipation, who have undergone full colonic bowel preparation for retrograde placement of manometric catheters and therefore had an empty colon at the time of recording (P. G. Dinning, N. Zarate, S. M. Scott and I. J. Cook, unpublished data). It is also possible that the diminished linkage simply represents a manifestation of the reduced extent of propagation of the PSs in the colon of these patients. This too is unlikely because the randomized data in health demonstrated a significant reduction in linkage.

It is more likely that the current findings are compatible with the hypothesis that defective transmission within enteric neural pathway(s) disturbs the normal coordination of regional linkage among propagating pressure wave sequences. Identification of these pathways and their control are difficult. An enteric occult reflex has been recently described.22 This reflex suggests that colonic elongation, resulting from the movement of content into a colonic region, triggers nitric oxide release from descending interneurones, which in turn inhibits colonic propulsive activity. As the inhibitory effects of nitric oxide have been demonstrated in the healthy human proximal colon in vivo,23 colonic elongation may help to explain part of the neural mechanism underlying the regional linkage phenomena in health. For example a PS may move content in an aborad direction, thereby resulting in colonic elongation in that part of the colon and inhibiting further PS activity. In our patients with delayed transit, colonic elongation might be expected in left colon because of the retained feces, which in turn could result in colonic inhibition. However, as the converse is true it may suggest impairment of this enteric occult reflex.

A postulate defect existing within the enteric interneuronal pathways responsible for coordinating colonic motor activity has also been recently described in patients with obstructed defecation. Bassotti et al. have shown that the overall population of glial cells were decreased in both the colonic myenteric and submucosal plexus of these patients.24 Glial cells are thought to be involved with enteric neurotransmission,25,26 the homeostasis of enteric neurones27 and their disruption, in mice, can lead to changes in neurochemical coding of enteric neurones resulting in intestinal dysmotility.

These patients were recruited for our original study4 before the publication of Rome II criteria. At that time we selected our patients on the basis of symptoms that were then believed to be compatible with obstructed defecation14–17,28,29 It has been shown subsequent to the commencement of our study that such symptoms are poor predictors of physiological measures of rectal evacuation disorders.30,31 Indeed there is a substantial body of data now showing that obstructed defecation and slow transit constipation cannot be discriminated on symptoms alone.32–37 The patients included in this study also had delayed transit, it could therefore be argued that they actually represent a population of slow-transit constipation. In short all our patients satisfy the Rome II criteria for functional constipation38 and the symptoms of outlet delay but did not satisfy the Rome II criteria for pelvic dyssynergia. However, the fundamental observation in the present study was that all controls and none of the patients demonstrated regional linkage. Hence, these study populations, irrespective of what label we attach to it, have demonstrated a homogeneous motor abnormality, which might be the measure for delayed transit and/or disordered defecation. Further prospective evaluation of this phenomenon in carefully selected and defined subsets of constipation is needed to determine its potential value.

Ancillary