Women with diarrhoea-predominant irritable bowel syndrome show an increased pressure response to 35% carbon dioxide stress challenge

Abstract

The responses to inhalation of 35% carbon dioxide (CO₂) as a stressor were compared in female irritable bowel syndrome (IBS) patients and healthy controls to assess potential differences in cardiovascular, neuroendocrine and behavioural responses to stress. A total of 22 women (12 patients with ROME II defined diarrhoea-predominant IBS and 10 aged-matched controls) were challenged with a single vital capacity breath of 35% CO₂ (with 65% oxygen). Beat-to-beat blood pressure and heart rate were recorded prior to, during and after the inhalation. Serum cortisol concentration and behavioural ratings were measured pre- and post-inhalation. A typical pattern of responses to CO₂ was observed, characterised by a reduction in heart rate and increases in serum cortisol and anxiogenic symptoms; however, these responses did not differ between groups. Both groups also demonstrated an increase in systolic blood pressure; however, this response was significantly enhanced in IBS patients compared to healthy controls (P < 0.05). These findings demonstrate that females with diarrhoea-predominant IBS have an exaggerated pressor response to 35% CO₂ stress challenge, suggesting a more stress-responsive sympathetic nervous system.

• Irritable bowel syndrome
• stress
• autonomic dysfunction
• carbon dioxide
• inhalation
• mood

Introduction

Irritable bowel syndrome (IBS) is often considered to be a maladaptive response to stress and these patients are seen as more stress sensitive as well as having increased levels of anxiety and depression. Studying responses of IBS patients to stress is difficult since potential stressors are either too aversive or do not reliably provoke the physiological sympathetic–adrenal–medullary (SAM), parasympathetic nervous system (PNS) and hypothalamic–pituitary–adrenal (HPA) axis and behavioural responses (e.g. gastrointestinal and anxiety symptoms) that typify a stress reaction.

Hypercapnia represents a threat to homeostasis and carbon dioxide (CO₂) inhalation has been used extensively in research on anxiety disorders to explore behavioural responses to stress (Griez et al. 1990, Perna et al. 1995, Papp et al. 1997). More recently CO₂ inhalation has been used as a novel and pragmatic model for acute stress induction since it reliably and robustly activates all of the key systems involved in the stress response (Kaye et al. 2004a,b). Specifically, a single breath of a 35% CO₂/65% O₂ mixture modestly increases anxiety and activates the SAM (increases in circulating noradrenaline level and blood pressure), PNS (pressor independent bradycardia) and HPA (increases in cortisol secretion) systems, and these responses are stable over a 4-to-6 week period (Wetherell et al. 2006). CO₂ inhalation is safe, reliable and easy to administer and, therefore, provides a useful tool for investigating stress-related dysregulation in IBS.

We present a controlled study comparing responses to this challenge between diarrhoea-predominant female IBS sufferers and healthy controls to test the hypothesis that this patient group has an exaggerated stress response. We restricted this study to female patients only in order to remove gender as a possible source of variance in this small study.

Method

Participants

All participants gave written informed consent and the local research ethics committee approved the study in accordance with the Declaration of Helsinki. Twelve female patients were recruited from a gastroenterology clinic at the Bristol Royal Infirmary with a current diagnosis of diarrhoea-predominant IBS as defined by the ROME II criteria (Drossman et al. 2000), and free from any other gastroenterological condition. Ten healthy age and sex-matched controls were recruited from staff and students at the University of Bristol. All potential participants underwent a full clinical examination to exclude significant medical conditions. Exclusion criteria for all participants were a personal history of hypertension or psychiatric illness, pregnancy, lactation, asthma or other respiratory illness, and any neurological disorder. In addition, control participants were excluded if they had a personal history of gastrointestinal disorders. All participants were free of any prescribed medication other than the contraceptive pill and were asked to refrain from smoking, eating and drinking for 1 hour prior to arrival at the laboratory.

Test day protocol

Participants were individually tested between 0900 and 1200 h and at least 1 hour after waking. On arrival at the laboratory a 20 G intravenous cannula was inserted into the antecubital vein of the non-dominant arm and kept patent using 0.9% saline infusion. After 20-minutes rest, participants practiced the inhalation procedure with the researcher. The inhalation procedure was performed as described in Wetherell et al. (2006). Briefly, an analogue flow meter (Ohmeda, USA) and a 10L Douglas bag (Harvard Apparatus) containing 35% CO₂ and 65% O₂ (BOC gases, UK) was attached to a silicone mouthpiece via a 3-way stopcock (Hans-Rudolf, Germany). During the practice breaths the Douglas bag was detached, the participant's nose was occluded using a nose clip and the mouthpiece inserted. Participants were then asked to inhale a full breath, exhale fully and then take a fast full breath. This breath was held for 4 seconds after which time the participant exhaled and the mouthpiece and nose clip were removed. During this practice period, inspired vital capacity (VC) was determined by flow meter and then projected onto a monitor to allow the participant a direct visual reference for the inhalation volume required for their test breath. For the test breath, the Douglas bag containing the CO₂/O₂ mixture was attached and the inhalation procedure detailed above was repeated. During the test breath, participants were asked to watch the monitor to ensure that they inhaled the pre-determined VC. The test breath was considered adequate if it exceeded 80% of the practice breath. In all participants, test breaths equalled or exceeded the VC determined during the practice breath.

Physiological parameters

Continuous pulse and blood pressure recordings were made with a Finapres recorder (FMS, Netherlands) commencing 15 minutes prior to and continuing for 5 minutes following inhalation of the CO₂ mixture. Blood samples were taken for cortisol assay 1 minute prior to inhalation and then 10, 20 and 30 minutes post-inhalation using pre-chilled SST tubes. After collection, samples were centrifuged at 4°C, serum was separated and then stored at − 20°C until they were assayed. Cortisol assays were performed using a direct radioimmunoassay (RIA) technique, which utilized radiolabelled ¹²⁵iodine. The intra- and interassay coefficients of variation were < 5% and < 7.1% respectively across the standard range 0.5–200 μg/100 ml.

Somatic and psychological symptoms

At the beginning of the study participants completed the panic symptom inventory (PSI) (Coupland et al. 1997) and the state form of the Spielberger State and Trait Anxiety Inventories (SSAI) (Spielberger et al. 1983). Prior to inhalation, participants rated a number of IBS-related symptoms (lower abdominal discomfort, abdominal bloating, stomach distension, gaseousness, indigestion, nausea and the need to open bowels) and psychological and somatic symptoms (anxiety, nervous, tense, fear, heart racing) using 100 mm visual analogue scales (VAS), where 0 = not at all and 100 = most ever. The VAS, SSAI and the PSI were assessed again immediately post-inhalation with participants reporting on how they felt during the peak of the CO₂ effect and finally at 30 minutes post-inhalation with subjects reporting how they felt at that time. The psychological and somatic VAS measures have been previously used to assess symptoms following CO₂ inhalation (Kaye et al. 2004a, Kaye et al. 2004b) and demonstrate significant test-retest reliability over a 4–6 week period (Wetherell et al. 2006). The IBS VAS measures have previously been used to assess stress-related changes in IBS symptoms and clearly differentiate between IBS patients and healthy controls on the basis of IBS symptom severity (Shufflebotham et al. 2006).

Statistical analysis

All parameters were analysed using a two (time: pre versus post) by two (group: IBS versus control) ANOVA. The means of the VAS scores for the IBS symptoms were used as a measure of IBS symptom severity. For all VAS items (subjective responses and IBS symptom severity) the pre- and post-inhalation scores were compared. Following Kaye et al. (2004a,b) and Wetherell et al. (2006), individual maximal responses were calculated for all biological parameters in order to account for individual differences in the time taken to mount a stress response. For heart rate and blood pressure the maximal response was calculated from the mean of five beats within the 100 beats post CO₂ exposure and compared with the mean of the 25–30th beat prior to inhalation. For serum cortisol the maximal post-inhalation response was compared with pre CO₂ levels. In all analyses, P < 0.05 was considered significant.

Results

Sample characteristics

All participants were female, and the groups did not differ on mean age (IBS = 34.4 years [SD = 12.3], Controls = 31.2 years [SD = 8.5]). None of the participants developed any adverse effects post-inhalation. Pre- and post-inhalation scores for IBS severity and systolic blood pressure are presented in Figure 1 and 2, all other measures are presented in Table I.

Figure 1 VAS IBS symptoms for IBS patients and volunteers pre- and post 35% CO₂ inhalation. IBS, Irritable Bowel Syndrome; Pre, Before CO₂ inhalation, Post, After CO₂ inhalation. Pre-post inhalation X group: F_(1,20) = 5.8, P = 0.026. Values are means ± SEM IBS group, n = 12; control group, n = 10.

Figure 2 Systolic blood pressure for IBS patients and controls pre and post CO₂ inhalation. IBS, Irritable Bowel Syndrome; Pre, Before CO₂ inhalation, Post, After CO₂ inhalation; SBP, Systolic Blood Pressure, mmHg. Pre- Post-inhalation: (F_(1,20) = 37.3, P < 0.001); Pre-Post inhalation X Group, (F_(1,20) = 5.17, P = 0.034). Values are means ± SEM IBS group, n = 12; Control group, n = 10.

Table I.  Variables pre- and post CO₂ inhalation not different between IBS patients and controls.

		IBS (N = 12)	Control (N = 10)	Time	Time × Group
SSAI	Pre	34.08 (1.9)	29.7 (1.7)	P < 0.001	NS
	Post	29.7 (4.5)	48.2 (4.2)
PSI	Pre	4.8 (1.9)	3.4 (1.6)	P < 0.001	NS
	Post	23.5 (5.5)	23.8 (4.3)
VAS anxiety	Pre	13.7 (2.8)	12.3 (3.1)	P < 0.001	NS
	Post	48.8 (6.9)	38.0 (6.1)
VAS nervous	Pre	18.6 (6.3)	16.0 (4.5)	P = 0.017	NS
	Post	42.6 (6.8)	28.5 (5.3)
VAS tense	Pre	21.1 (5.5)	16.0 (4.8)	P = 0.018	NS
	Post	40.0 (6.9)	30.0 (5.6)
HR (bpm)	Pre	82.5 (6.3)	92.4 (7.0)	P < 0.001	NS
	Post	62.9 (3.6)	60.7 (4.5)
Serum cortisol (μg/100 ml)	Pre	55.8 (6.3)^*	55.9 (5.6)	P = 0.046	NS
	Post	63.0 (5.6)^*	63.6 (7.5)

IBS, Irritable Bowel Syndrome; Pre, Before CO₂ inhalation, Post, After CO₂ inhalation; SSAI, Spielberger State Anxiety Inventory; PSI, Panic Symptom Inventory; VAS, Visual Analogue Scale; HR, Heart Rate, bpm, beats per minute.

^*n = 11.

IBS symptoms

As expected, patients reported more IBS symptoms compared to controls at baseline. A significant time by group interaction was revealed (F_(1,20) = 5.8, P = 0.026), which reflected a reduction in reported severity of symptoms in IBS patients post CO₂ inhalation in comparison with no change in reported severity in the controls (Figure 1). Inspection of the means for the individual symptoms found a reduction in reported severity in abdominal bloating, need to open bowels, lower abdominal discomfort, stomach distension and indigestion (data not shown).

Subjective response

CO₂ inhalation significantly increased anxiety as measured by the SSAI (F_(1,20) = 30.4, P < 0.001) and the PSI (F _(1,20) = 32.2, P < 0.001) and the VAS items of anxiety (F_(1,20) = 38.6, P < 0.001), fear (F_(1,20) = 16.6, P = 0.001), nervous (F_(1,20) = 6.8, P = 0.017) and tense (F_(1,20) = 6.6, P = 0.018). There was no difference between IBS patients and control participants in any of these measures.

Cardiovascular response

Inhalation of CO₂ resulted in a significant increase in systolic blood pressure (F_(1,20) = 37.3, P ( 0.001) and a significant decrease in heart rate (Table I; F_(1,20) = 75.4, P < 0.001) in patients and controls. For systolic blood pressure (Figure 2), a significant time by group interaction was observed with IBS patients demonstrating a significantly greater increase in blood pressure compared to controls (F_(1,20) = 5.17, P = 0.034). There was no difference in the mean decrease in heart rate post CO₂ exposure between the two groups.

Cortisol response

Cortisol data were not available for one IBS patient at any time point. Therefore, the analysis was based on 11 IBS patients. CO₂ inhalation elicited significant increases in serum cortisol concentration in patients and controls (Table I; F_(1,19) = 4.6, P = 0.046); however, there were no differences in reactivity between the groups.

Discussion

Although IBS is considered a stress responsive disorder it has been difficult to establish the exact nature of any abnormalities in the stress response. The latter comprises cardiovascular, neuroendocrine and behavioural components and it is important to establish which, if any, of these components is reacting differently in IBS patients versus controls. Psychological stressors (Stroop test, dichotomous listening) have been shown to alter visceral sensitivity in IBS patients (Posserud et al. 2004, Murray et al. 2004); however, a variety of stressors including the Stroop test (Levine et al. 1997), mental exercises (Elsenbruch et al. 2001a), and public speaking (Elsenbruch et al. 2006) have failed to provoke any differences in physiological stress responses between IBS patients and healthy controls. The current findings provide further evidence that inhalation of 35% CO₂ reliably activates all components of the stress response and thus represents a useful tool for exploring these responses in health and disease. This is the first study to assess responses to inhalation of 35% CO₂ in IBS. We found that a single breath of 35% CO₂ activated key elements of the stress response as evidenced by increased subjective anxiety, increased serum cortisol level and a vagal response (bradycardia) to a similar extent in both IBS and healthy control participants. Although a systolic pressor response was also induced in both groups, there was a significantly greater response in IBS patients representing twice the increase in blood pressure post CO₂ inhalation compared with that seen in healthy controls.

Autonomic dysfunction, also thought to be a potential etiological factor in IBS, has previously been investigated using spectral analysis of heart rate variability. Studies have shown that IBS patients have increased sympathetic activity at baseline (Karling et al. 1998, Orr et al. 2000), and after sympathetic stimulation with orthostatic testing (Waring et al. 2004), and visceral distension (Tillisch et al. 2005). The sympathetic system is important in coordinating the initial physiological changes associated with a stress response. In the upper gastrointestinal system SAM activation can result in reduced contractions and secretions leading to increased bloating and lack of appetite. Paradoxically, in the lower gastro-intestinal (GI) tract, adrenergic stimulation can result in increased contractions and secretions and thus result in increased motility and abdominal pain and these symptoms can occur as part of a normal stress response.

The observed pressor response is likely mediated by peripheral vasoconstriction induced by noradrenaline (Kaye et al. 2004a,b); however, it is possible that the pressor response may be a result of changes in the depth and rate of respiration. Previous studies (Ford et al. 1995, Bharucha et al. 1996) have demonstrated that hypocapnic hyperventilation alters heart rate variability and systolic blood pressure and also increases colonic tone in healthy individuals. The current study did not assess respiration rate, and therefore it is not clear whether this would differ between IBS patients and controls. However, an excessive sympathetic response as suggested by this study could help explain the increased lower gastrointestinal symptoms experienced by this group of IBS patients.

Bradycardic responses to CO₂ inhalation in the IBS patients did not differ from controls in our study, suggesting normal functioning of this component of the response to CO₂. The bradycardic response to 35% CO₂ inhalation is considered to result from a direct effect of CO₂ on vagal medullary nuclei and as such may be a marker for parasympathetic outflow (Kaye et al. 2004a,b). It should be noted, however, that as changes in heart rate are subject to sympathetic and parasympathetic co-activation, measures of heart rate variability, rather than heart rate alone, would provide more conclusive evidence regarding the parasympathetic response to CO₂ inhalation.

The paradoxical results of studies assessing autonomic nervous system functioning in response to challenge in IBS clearly need further elucidation. The nature of the challenge is probably critical, and some stressors may activate pain pathways that may influence the results. Gender differences may also be relevant. We chose to recruit only female patients in order to remove sex as a source of variability given the small sample size in the study. A recent study has reported altered autonomic responses to stress in IBS patients with this effect being greater in males (Tillisch et al. 2005). In future studies it would be important to test responses of male patients to the CO₂ inhalation to further elucidate sex differences in response to stress in IBS patients. Altered sympatho-vagal balance has also been demonstrated in different IBS subgroups with increased sympathetic activity and decreased parasympathetic activity reported in diarrhoea- and constipation predominant patients respectively (Heitkemper et al. 1998, Aggarwal et al. 1994, Heitkemper et al. 2001, Elsenbruch and Orr 2001b). For this reason we thought it important in a study of this size to recruit only diarrhoea-predominant IBS patients, so minimizing any effect of potential differences in autonomic response between these subgroups. In future studies it would be important to determine how constipation-predominant patients respond to this challenge, and determine whether there are different responses between the two subgroups.

Regarding the HPA axis, baseline differences in cortisol levels have been found in several studies of IBS patients (Dinan et al. 2006, Ehlert et al. 2005, Heitkemper et al. 1996, Patacchioli et al. 2001), although these results have sometimes been conflicting and difficult to reproduce (Elsenbruch and Orr 2001b, Bohmelt et al. 2005). In the present study there were no baseline differences in serum cortisol levels between patient and control groups. CO₂ inhalation did provoke a significant cortisol rise in patients and controls but there were no differences in reactivity between the groups suggesting that this element of the neuroendocrine stress response is functioning normally in diarrhoea-predominant IBS. However, although all testing took place at least 1 hour post-waking, the precise time of waking was not recorded. As cortisol secretion shows a marked diurnal pattern characterised by an increase immediately post-waking and a gradual decline throughout the day (Kirschbaum and Hellhammer 2000), it is possible that at the time of testing, individuals were at differing points on their diurnal profile, which may have affected subsequent reactivity. Future studies should, therefore, include time of waking in analyses to reduce this source of inter-individual variation.

As expected, IBS patients had more GI and depressive symptoms at baseline than the controls. After inhalation of the CO₂ mixture there was a significant increase in anxiety using SSAI, PSI and visual analogue scores, but paradoxically there was a significant reduction in GI symptoms at the same time (Figure 1). One explanation for this apparently counterintuitive finding is that activation of the sympathetic nervous system leads directly to a reduction in GI symptoms, however, the mechanism by which this reduction might occur is unclear. Alternatively, as CO₂ inhalation provokes significant, but transient increases in anxiety there may be a temporary exchange between anxiety and GI symptoms. That is, higher anxiety or lowered mood may temporarily distract patients from their GI symptoms. Consistent with this hypothesis, we have recently published a study in which increasing serotonin availability through tryptophan ingestion increased gastrointestinal symptoms but reduced anxiety in IBS patients (Shufflebotham et al. 2006). Serotonin may also be a key link between the pressor effect of CO₂ inhalation seen in our patients and the changes in gastrointestinal and emotional symptoms. Evidence for this comes from a recent study showing that depleting serotonin via the acute tryptophan depletion paradigm enhances pressor responses to an anxiogenic challenge in anxious patients (Davies et al. 2006). Given this, in further studies we aim to assess whether SSRI treatment of patients with IBS normalizes the pressor effect in this paradigm and whether this correlates with symptomatic improvement.

In summary, we have shown that a single VC inhalation of 35% CO₂ results in activation of all key components of the stress response (SAM, PNS, HPA and behavioural responses) in women. Simultaneous investigation of these components allows for a more accurate study of the factors that modulate the stress response. In this study 35% CO₂ inhalation resulted in an exaggerated systolic pressor response in diarrhoea-predominant IBS patients, which may reflect dysregulation in the autonomic nervous system. An abnormal response to stress with increased sympathetic tone could account for the increased incidence of adrenergic symptoms seen in this patient group. It would be important to see whether effective treatment modulates this excessive pressor response, whether there are gender differences in this effect and whether a similar pattern of reactivity is observed in patients with constipation-predominant IBS.

Role of funding source

Avon and Wiltshire Partnership NHS Trust R and D funded this study. The sponsors had no role in study design, data collection, data analysis, data interpretation or writing of the report.

Conflict of interest

We declare that we have no conflict of interest.

Acknowledgements

M Franklin performed the cortisol assays. Thank you to A Rich for technical support.

Additional information

Notes on contributors

Jonathan Shufflebotham

Contributors J Shufflebotham instigated the study. J Shufflebotham, M Wetherell, and J Potokar designed the study. J Shufflebotham recruited patients from C Probert's clinic. J Shufflebotham, M Wetherell ran the study days and collected data. J Shufflebotham, M Wetherell, D Hince, S Hood, S Lightman, J Potokar, D Nutt, and C Probert contributed to the analysis. J Shufflebotham, M Wetherell, J Potokar, D Hince S, Hood, S Lightman, D Nutt, and C Probert contributed to the final manuscript, which was approved by all authors.