Pharmacological management of constipation
Dr Anton V Emmanuel BSc, MD, FRCP, Senior Lecturer in Gastroenterology, Physiology Unit, University College Hospital, 235 Euston Road, London NW1 2BU, UK.
Tel: 0207 380 9311; fax: 0207 380 9239; e-mail: firstname.lastname@example.org
Abstract The approach of this review is to give a pragmatic approach to using laxatives, based on a combination of what is known about mechanism of action and the available literature on evidence.
In constipated patients, laxatives are the most commonly prescribed pharmacological agents, and they are classically classified as belonging to four main types: bulking agents, stool softeners, stimulant laxatives and osmotically active compounds. Robust clinical trial evidence is difficult to determine with a condition as chronic and subjective as constipation.
However, meta-analyses and reviews over the last few years have highlighted that there are some proven agents.1–3 The corollary of this is that there are a number of older agents whose efficacy has never been demonstrated to the standards of modern clinical trial design, and we would be ill-advised to ignore the potential role of these drugs.4,5
Bulking agents According to their FDA approval, these agents are mostly indicated for patients with episodic constipation when stools are lacking in water content. They are primarily organic polysaccharides that act by encouraging water retention in the stool. A number of them (methylcellulose, psyllium) also undergo bacterial fermentation which may enhance this effect; unfortunately this also predisposes to the frequent side effect of bloating and flatulence. Much rarer side-effects of bolus obstruction of the oesophagus or colon have been reported.1,6 Reflecting the poverty of the literature and the modest effect of these agents, two systematic reviews1,6 have yielded conflicting conclusions. The Tramonte et al. review1 compared only bulking agents with placebo, showing an increase in stool frequency. By contrast, the Jones et al. meta-analysis6 compared stimulants, osmotic and bulking agents, showing no efficacy for the latter over placebo.
Psyllium. Psyllium is the husk of the seed of Plantago ovata. Two7,8 of three placebo-controlled trials7–9 have shown that stool frequency, consistency and ease of evacuation are improved by the agent. Two other studies have compared psyllium with lactulose,10,11 showing that compared to baseline, the bulking agent results in fewer hard stools10 and greater stool frequency;11 the magnitude of effect of psyllium and lactulose was similar.10,11 However, all but one study8 was of less than 4 weeks duration, a significant factor given that it is suggested that it takes 14 days for these drugs to exert an effect.
Bran. A placebo-controlled trial of bran12 failed to show an improvement over placebo in a cross-over design. Two other low-quality randomized studies showed a minimal13 or absent14 effect on stool frequency and moisture.
Methylcellulose. One study15 comparing various doses of methylcellulose with psyllium showed that whilst both agents exerted a modest effect in terms of stool frequency and consistency, there was no difference between agents. The effect of the drug was more marked in healthy controls than those with chronic constipation (CC).
Calcium polycarbophil. One, nursing home-based cross-over study (compared with psyllium) showed that whilst these patients preferred the calcium salt to psyllium, there was no effect on stool frequency or ease of defecation.16
Stool softeners These are indicated for occasional constipation, and often as an adjuvant to a bulking or stimulant agent. They are anionic surfactants, with an emulsifying and wetting action. Their effect depends on the strength of their action on the surface of the stool with a generally modest effect. Stool softeners are well-tolerated.
Docusate. Two placebo-controlled,17,18 and one psyllium-controlled19 randomized studies have addressed the effect of docusate in patients with CC. The 1968 study of sodium docusate17 showed significant improvements in stool frequency and overall subjective assessment, whilst the later study of calcium docusate18 showed no advantages over placebo. The comparison of sodium docusate with psyllium19 showed that the bulking agent had a greater effect on stool frequency, although this study was limited by only looking at a 14-day treatment window. One further single-blind study20 underlined the modest effect of stool softeners: neither of two doses of sodium docusate improved stool frequency, whilst calcium docusate was shown to have a slight increase in stool frequency compared to the placebo run-in period.
Stimulant laxatives Although widely used, as over-the-counter and prescribed agents, there is an extremely limited evidence base supporting the use of stimulants in CC. They are either naturally occurring agents (such as senna and cascara) or phenolphthalein analogues (such as bisacodyl). They are hydrolysed in the gut (by either enterocyte enzymes or colonic flora) and act by stimulating peristalsis, sensory nerve endings (hence the frequent side effect of colic) and possibly interfering with electrolyte flux to inhibit water absorption. Stimulant agents are taken up by enterocytes and hence cause melanosis coli. Whilst classical thinking was that they caused a ‘cathartic colon’,21 there is no evidence that currently available stimulants do this.22 There are no placebo-controlled studies of efficacy.
Nevertheless, stimulants tend to be used for occasional constipation as a night-time dose prior to a morning bowel action, and have a particular (if unproven) place in neurogenic bowel dysfunction.23
Senna. Two non-placebo controlled studies are found in the literature.24,25 The earlier one compared senna with sodium picosulfate (an osmotic agent), and found that whilst there was no difference in stool frequency between the two agents, the osmotic agent was more likely to cause loose stools.24 The second trial compared senna liquid with bran, showing equal effect of both in terms of stool frequency and consistency.25
Bisacodyl. A cross-over trial of bisacodyl and bisoxatin (a stool softener) represents the only available randomized clinical trial and that showed no difference between the agents.26
Osmotic laxatives This class of laxatives comprises both inorganic salts (magnesium compounds) and organic alcohols or sugars [such as lactulose, polyethylene glycol (PEG)]. They act by generating an osmotic gradient which encourages water retention in the lumen. As such they have the potential to be dose titrated according to the stool output, and are usually used for both chronic and occasional constipation. The nature of their mechanism of action means that a number of electrolyte abnormalities may occur, but the commonest adverse effect relates to difficulty getting titration correct (namely diarrhoea). Abdominal bloating, nausea and flatulence are common, but these drugs are often used in slow transit, and have a particular place in megacolon.27
Magnesium salts. A cross-over study of magnesium hydroxide vs laxamucil (a bulking agent) showed a benefit with the osmotic agent, such that as required use of extra laxative doses were reduced and spontaneous stool frequency was increased with magnesium hydroxide.28 No trials are available for magnesium sulphate (Epsom salts).
Lactulose. Two placebo-controlled trials have shown that although both placebo and lactulose improved stool frequency and consistency compared to baseline, this was significantly greater with lactulose.29,30 This efficacy was seen across a range of doses from 15 to 60 mL per day.
Polyethylene glycol. Recently, PEG has been subject to a larger number of clinical trials than other agents: there are five placebo-controlled studies31–35 and two lactulose-controlled ones.36,37 Of the five placebo-controlled studies, three are parallel design and two crossover. Stool frequency – and when assessed, stool consistency – was improved by PEG compared to placebo in all these studies.31–35 In the lactulose-controlled studies, one showed improved stool frequency and ease of defecation with PEG over lactulose37 whilst the other showed equal effect with both agents.36 Diarrhoea was the commonest adverse effect, occurring in as many as 40% in one of the trials. Electrolyte abnormalities, again, may occur – but tend to be rare except in the elderly.
Critical appraisal of RCTs
Despite the fact that constipation can be a chronic disabling disorder accompanied, as we now realize, in the most severely affected individuals, by very marked impairments in quality of life (QOL) and social functioning,38 relatively few well-conducted clinical trials have been performed. Indeed, until very recently, when the pharmaceutical industry became aware of both the high prevalence of these conditions and of the glaring unmet therapeutic needs that existed throughout the functional gastrointestinal disorders, including CC, few trials of any quality were performed in CC. This drought of new compounds is vividly illustrated by the fact that over the past 17 years only four pharmaceuticals have been approved in the US by the FDA for constipation, and one, tegaserod, has already been withdrawn because of safety concerns. The reasons for this paucity of research are explained by the range of problems that confront the clinical trialist who sets out to study this disorder and the clinician or clinical scientist who attempts to interpret the results of their research and translate them into clinical practice. Each of these issues will now be dealt with in the following sections.
Constipation is a symptom which means different things to different individuals; to some it refers to stool frequency, for others, to stool form and consistency, difficulty with defecation or a sense of incomplete evacuation. How are these symptoms norm referenced by the patient? To their own prior experience or to what they expect the norm to be? Furthermore, as is the case with any purely symptom-based disorder, there are the vagaries of recall bias which may be refuted by prospective stool diaries.39 Up until fairly recently and reflecting the physician’s perspective, constipation was viewed, for the purposes of patient selection in clinical trials, as simply an issue of stool frequency; we now realize that frequency of defecation is not the most bothersome symptom from the constipated patient’s perspective.40,41 Patients are usually more concerned about straining, hard stools and difficult defecation. Appreciative of these patient concerns and seeking to design trials around end points which have real clinical meaning, investigators have, more recently turned to definitions that are more encompassing of these variable expressions of constipation.
Modifications of the Rome II or III definitions of functional constipation have been especially popular.42 Although moving in the right direction, this approach still does not satisfy the rigors of the standards for a patient-reported outcome measure recently promulgated by the FDA for use in functional gastrointestinal disorders.43
Constipation is not a single homogenous entity but includes those with idiopathic slow transit constipation, others with an evacuation disorder (ED) which itself comprises a number of sub-types, a further group where these two pathophysiological entities co-exist or overlap, and a final group in whom constipation is a component of the irritable bowel syndrome (IBS). As it is assumed that the therapeutic approach to these groups will vary, it would appear desirable to study these separately or, at the very least, to stratify for these groups in designing a trial. This is not an easy task, not only because of the frequency of overlap between the groups, but also because of the difficulty of predicting pathophysiology on the basis of symptoms44 or tests.45 It should come as no surprise, therefore, that inclusion criteria for clinical trials have adopted ‘looser’ and more inclusive symptom-based criteria such as that described above; this approach not only simplifies life for the investigator but avoids the spectre (from the perspective of the pharmaceutical industry) of a label which limits the use of particular medication to those patients who have a particular pathophysiology underlying their constipation. The down side of this approach is that the patient population in most trials is inevitably heterogeneous, a factor which, no doubt, contributes to lower than expected success rates for an agent.
The overlap with IBS raises a further challenge; in clinical practice, differentiating IBS-C from CC is often one of semantics, representing parts of the same spectrum. It is interesting to note that a number of recently approved drugs for CC have also been approved for IBS-C. While it may seem churlish to criticize studies from some decades ago, it needs to be stressed that most of the older literature (which forms the ‘scientific’ basis for our continued use of laxatives, as described above) is, quite simply, bad, in terms of trial design. Several basic principles of trial design, now regarded as de rigeur, were absent: randomization, blinding, placebo control, adequate description of entry criteria and outcomes, to name but a few. Even recent studies are not above criticism. For example, some have employed a cross-over design, a feature regarded as inappropriate for use in functional disorders, while other studies vary considerably in overall duration (commonly somewhere from 2 to 12 weeks) and in the length and conduct of run-in and wash-out periods. A major lacuna in the CC literature is the dearth of long-term studies; CC is a chronic condition yet there are few controlled studies beyond 12 weeks!
There is, then the issue of comparator. In general, it is clinically relevant to compare the efficacy of novel agents to a well-chosen active comparator drug. However, most recent high quality studies of putative enterokinetics in CC have been placebo controlled and have not compared the new agent with ‘standard therapy’ (i.e. a laxative), although it needs to be conceded that, in some instances, only those who had failed laxative therapy were included.
Given the considerable gulf in cost that exists between these remedies and the more recently developed prescription drugs, it seems inevitable that regulatory agencies, mindful of potential costs to governments and the individual patient, will soon demand that new chemical entities be compared to a laxative. However, comparator controlled trials in CC are specifically problematic for a number of reasons: currently available laxatives, at least in Europe, are generally limited in their indication and established duration of use. In addition, many patients with CC do not have an adequate long-term response to existing laxatives. Enrolling such patients in a trial comparing a novel agent against such a laxative may thus be considered inappropriate. PEG could be considered as an active comparator in constipation trials, as its efficacy in the short-term relief of constipation has been well documented, and some long-term studies are now also available. Indeed, two studies comparing PEG with lactulose37 and tegaserod46 are among the few attempts to make clinically relevant comparisons. However, using PEG as a comparator is difficult with regards to blinding. Use of PEG requires the consumption of large volumes of fluid. Matching this fluid consumption with the investigational drug could also potentially impact on its pharmacokinetic and pharmacodynamic properties as compared with its normal clinical use without a large fluid load. Furthermore, the taste and appearance of PEG are distinctive, such that it may not be feasible to make appropriate as placebo.
Two other issues in trial design deserve consideration: the use of rescue medication and the timing of invasive investigations prior to trial commencement. There is no standardization of approach to rescue medications in CC. Again, given its demonstrated efficacy in good quality trials, and the fact that it is, in general, well-tolerated and widely available, a PEG-based laxative appears a reasonable option. Furthermore, and regardless of the choice of rescue laxative, its use should be an important secondary outcome variable in CC trials. It is well known to clinicians that the performance of a colonoscopy or any procedure that requires bowel cleansing often leads to prolonged periods of remission among CC patients, it is critical that this ‘honeymoon’ be accounted for at the time of randomization.
Optimization of study design and valid outcome measures
It follows from the discussions above of the very definition of constipation and patient selection, that the definition and validation of clinically meaningful outcome measures and thus study end points represents a major challenge. End points such as bowel frequency, laxative use, patient-reported changes in stool consistency and ‘overall’ symptom improvement were variably recorded. Considerable efforts have been exerted of late to produce outcome measures that are more meaningful, and which encompass the totality of the patient experience with constipation.
Some progress has been made: from bowel movements per week we have advanced through spontaneous bowel movements47 to spontaneous complete bowel movements (SCBM).48,49 To account for the range of CC symptoms, composite outcome measures such as the Patient Assessment of Constipation Assessment Questionnaire (PAC-SYM) have been developed.50 The PAC-SYM contains 12 items assigned to three subscales (stool symptoms, rectal symptoms and abdominal symptoms) and has been shown to be internally consistent, valid, and responsive to change.51
Given the appropriate preoccupation in the functional bowel world with impact on QOL, a variety of measures of QOL have been included as secondary outcomes in more recent studies the ability of a particular intervention to improve QOL, especially when measured by a CC-specific instrument such as the Patient Assessment of Constipation Quality of Life Questionnaire (PAC-QOL),52 being regarded as adding additional credibility to the reported outcome. The PAC-QOL was designed to provide a comprehensive assessment of the burden of constipation on patients’ everyday functioning and well-being, and has been shown in a multinational study to be internally consistent, reproducible, valid, and responsive to improvements over time.52
It is obvious, but essential, to recognize that a composite score or a measure of QOL should be given in the context of results being clinically meaningful. For example, the frequently used achievement of three or more SCBM/week as a clinically valid endpoint, still only represents the lower end of the normal range. In designing new standards of outcome measures and study endpoints, there is a great and urgent need to clearly define the magnitude of an effect that is truly meaningful to a patient.
Despite its impact on social functioning and personal life, as expressed on any number of QOL scales, CC is a benign disorder and, as such, is one in which regulatory agencies retain a very low threshold for acceptance of risk. This poses another major hurdle for drug development, especially given the recent case of tegaserod,53 where the more rare events only came to light in post-marketing surveillance. More rigorous pre-clinical testing, especially in relation to cardiovascular toxicity is now required, and clinical trials in more diverse patient groups and, not just the ultra-fit, may soon be mandated.
Evolving and future medical treatments
The normal gut flora (microbiota): an essential factor in digestive health The human gastrointestinal microflora is a complex ecosystem of approximately 300–500 bacterial species comprising nearly two million genes (the ‘microbiome’). When disturbed, the flora has a remarkable capacity to restore itself and to return to exactly the same state as it was before.54 Because of the normal peristaltic motility of the intestine and the antimicrobial effects of gastric acid, the stomach and proximal small intestine contain relatively small numbers of bacteria in healthy subjects; on crossing into the colon, the bacterial concentration and variety of the enteric flora changes dramatically. Here, concentrations as high as 1012 CFU mL−1 may be found; comprised mainly of anaerobes, such as bacteroides, porphyromonas, bifidobacterium, lactobacillus and clostridium, with anaerobic bacteria outnumbering aerobic bacteria by a factor of 100–1000 : 1.55 However, conventional culture-based approaches to the enumeration and identification of the bacterial species greatly underestimate both the number and diversity of the flora, and the full definition of the normal human microbiota and its variations in health must await detailed and painstaking molecular studies. The normal enteric bacterial flora influences a variety of intestinal functions directly relevant to motor function and the pathogenesis of constipation.56 Firstly, unabsorbed dietary sugars are salvaged by bacterial disaccharidases, converted into short-chain fatty acids (SCFAs) and used as an energy source by the colonic mucosa. Secondly, the colonic flora is essential to the deconjugation of primary bile acids which could significantly influence stool form and consistency. Finally, certain members of the commensal flora have been shown to stimulate gut motility, and the production of neuropeptides and gases, capable of modulating gut nerve and muscle function, has been identified. However, data are scant as to whether constipation is associated with an alteration in the colonic flora, although there is a limited literature to suggest some disturbances in the flora in both children57 and adults.58
Mining the microbiota The concept of a pharmabiotic was introduced to attempt to embrace the various ways in which the microbiota could be manipulated to confer a clinical benefit to the host.59 A therapeutic benefit could be conferred through supplementation with commensal organisms (the probiotic approach), through selectively stimulating the growth of certain commensal species (the prebiotic approach) or by extracting and purifying components or products of commensal flora which are biologically active.
Probiotics are defined as live organisms that, when ingested in adequate amounts, exert a health benefit to the host.60 The most commonly used probiotics are lactic acid bacteria and non-pathogenic yeasts. The many products that are found in health food stores and supermarket shelves which include the term probiotic in their label often fail to fulfil the definition provided above:
- 1they may not contain live organisms or have not been adequately tested to ensure that the organisms will survive in the conditions (e.g. room temperature) or for the length of time (days, weeks, or months) that is claimed;
- 2they may not confer health benefit because, either they have never been tested in man, or because what tests have been preformed have been inadequate or even negative;
- 3they may contain organisms (including pathogens) that they are not supposed to contain.61
Although probiotics have been proposed for use in inflammatory, infectious, neoplastic and allergic disorders, the ideal probiotic strain for many of these indications has yet to be defined, although progress continues in this area. While probiotic ‘cocktails’ have also been advocated to maximize effect, it needs to be noted that some probiotic combinations have been shown to prove antagonistic, rather than synergistic, in certain situations.
Prebiotics are defined as non-digestible, but fermentable, foods that beneficially affect the host by selectively stimulating the growth and activity of one species or a limited number of species of bacteria in the colon.62 Compared with probiotics, which introduce exogenous bacteria into the human colon, prebiotics stimulate the preferential growth of a limited number of health-promoting commensal flora already residing in the colon, especially lactobacilli and bifidobacteria. Their relative efficacies will depend, in part, on the initial concentrations of the probiotic component of the flora and on luminal pH. The only prebiotics for which sufficient data have been generated to allow an evaluation of their possible classification as functional food ingredients are the inulin-type fructans [which are linked by β (2-1)bonds that limit their digestion by upper intestinal enzymes] and fructooligosaccharides.
Both are present in significant amounts in many edible fruits and vegetables, including wheat, onion, chicory, garlic, leeks, artichokes and bananas. Because of their chemical structure, prebiotics are not absorbed in the small intestine but are fermented, in the colon, by endogenous bacteria to energy and metabolic substrates, with lactic and short-chain carboxylic acids as end products of the fermentation. Most of the evidence regarding the potential health benefits of prebiotics is derived from experimental animal studies and human trials in small numbers of subjects; there are insufficient, prospective, adequately powered studies in gastrointestinal disease to permit definitive conclusions to be drawn.
Synbiotics, defined as a combination of a probiotic and a prebiotic, aim to increase the survival and activity of proven probiotics in vivo and stimulating indigenous bifidobacteria and lactobacilli. Again, data for efficacy in human disease is scanty.
The potential importance of bacterial metabolic products, such as SCFAs and deconjugated bile salt salts to the pathogenesis or alleviation of diarrhoea has already been mentioned. One can also envision how changes in bacterial fermentation rates or in the relative production of gases, such as hydrogen, methane or sulphide could affect symptomatology.
The identification of the elaboration of a number of neuropeptides and the important inhibitory neurotransmitter nitric oxide by certain commensal bacteria63 suggests new mechanisms whereby the flora could influence colonic motor function.
Prebiotics and probiotics in constipation While probiotics and, to a lesser extent, prebiotics have been extensively studied in IBS,64,65 data on the impact of probiotics on constipation is lacking. A number of studies have demonstrated the ability of Bifidobacterium animalis (also referred to as Bifidobacterium lactis DN) to shorten colonic transit in healthy women,66 the elderly67 and in subjects with IBS.68 These effects were observed only if live organisms were used and were most obvious if colon transit time was slower at baseline. Of note, when studied, these effects on (particularly right colon) transit were noted to be independent of changes in faecal mass or bile acid content, suggesting a direct effect on colonic motility.68 A symbiotic combination of Lactobacillus rhamnosus, B. lactis and inulin prebiotic has been shown to stimulate small bowel motility.69
While few, if any studies, have examined the effect of one of the classical prebiotics (e.g. inulin, fructose oligosaccharides) in constipation, it must be remembered that the main contributor to the increase in stool bulk that accompanies the ingestion of bran or fibre supplements comes from bacteria; i.e. these agents are acting like prebiotics to promote bacterial proliferation. Furthermore, lactulose, a very commonly used laxative, has been shown to exert a truly prebiotic effect by promoting the growth of a number of strains of bifidobacteria in human subjects.70
Very few double-blind placebo controlled trials of probiotics in acute or CC per se (i.e. other that in association with IBS) is available. Among the few positive randomized controlled trials (RCT) are those of Koebnick and colleagues which documented a positive benefit for a probiotic beverage containing Lactobacillus casei Shirota and of Yang and colleagues which reported benefits from B. lactis DN-173010, in patients with CC.71,72 In their study using lactobacillus GG as an adjunct to lactulose, Banaszkiewicz and Szajewska found no additional benefit from the probiotic.73 Other studies which were uncontrolled, or in which the probiotic was combined with some other form of therapy, reported variable benefits for bifidobacteria, lactobacilli and proprionibacteria and infusions of faecal suspensions.74–80
Some support for the use of probiotics in constipation comes from IBS studies and, especially those that provided results for constipated IBS subjects.64,65,81–83 In accordance with the aforementioned effects on colonic transit, Bifidobacterium animalis has been shown to significantly increase stool frequency among IBS subjects with baseline constipation (as defined by stool frequency of less than 3 per week).84 Another strain Bifidobacterium infantis 35624 was shown to normalize stool consistency (as assessed by the Bristol stool scale) and reduce straining among constipated IBS subjects, without an apparent effect on stool frequency.85,86
5 HT4 agonists
The 5-HT4 receptor is considered to have a particularly important role, both physiologically and pathophysiologically, in the regulation of GI function.87–89 Activation of neuronal 5-HT4 receptors results in prokinetic activity throughout the GI tract,90 and triggers the release of neurotransmitters from the enteric nerves resulting in increased contractility and stimulation of the peristaltic reflex.88,91–93 Pro-secretory effects occur with release of chloride and bicarbonate being observed from duodenal, colonic or jejunal epithelial cells.94–96 Notably, the precise role of the 5-HT4 receptor in visceral sensitivity remains unclear. A number of 5-HT4 receptor agonists are currently under investigation with the goal to avoid the cardiac side effects seen with tegaserod and cisapride.
Tegaserod Tegaserod, an aminoguanidine indole compound, is a partial 5-HT4 agonist. In animal models, tegaserod acts as a motility-enhancing agent, exerting activity throughout the gastrointestinal tract.97,98 Tegaserod has also been shown to significantly accelerate colonic transit in healthy volunteers and in patients with constipation-predominant IBS-C.99,100 In addition, tegaserod may attenuate the visceromotor response evoked by colorectal distension in control animals,101 and this may occur in healthy human subjects.102 Several large randomized, double-blind, placebo-controlled trials of oral tegaserod performed in patients with IBS-C103–106 have shown superiority for tegaserod over placebo, as evaluated by the subjects’ global assessment of overall relief, and secondary end points (i.e., abdominal pain, bowel frequency and consistency). In a recent Cochrane review on tegaserod,107 the relative risk (RR) of being a responder in terms of global relief of GI symptoms during the last 4 weeks of treatment was significantly higher with tegaserod compared with placebo in patients with C-IBS; the RR of being a responder in terms of SCBM/week with tegaserod 12 mg was 1.54 (95% CI 1.35–1.75), compared with placebo where it was 0.6 (95% CI 0.42–0.78). Two large randomized clinical trials108,109 have been performed to evaluate the efficacy of tegaserod in the management of CC. They showed that constipated patients younger than 65 years randomized to tegaserod had 16–19% greater SCBMs compared to placebo.
Following these two large studies, several other relatively small studies46,110–112 have been performed. On Chan et al.110 evaluated the effect of tegaserod in 216 patients with CC from a Chinese population, and showed that the tegaserod group induced a significantly higher response rate in terms of the mean increase of SCBM >1/week (48%) compared with placebo group (29%). In another Chinese study, Lin et al.111 reported that tegaserod (48%) was more effective vs placebo (35%) as measured by the mean increase of more than one SCBM/week. Fred et al. evaluated the effect of tegaserod in 322 male patients with CC, and also reported that increases of at least one SCBM/week occurred more often on tegaserod (41%) compared with placebo (29%). Finally, Di Palma et al.46 performed a randomized, open-labelled, parallel, multicenter study comparing tegaserod and PEG laxative in 237 patients with CC. In this study, treatment success was defined differently; the primary end point was no longer fulfilling modified ROME criteria for constipation for 50% or more of their treatment weeks. This study demonstrated that PEG was more efficacious than tegaserod in treating constipation over a 4-week period; the primary end point was reached significantly more often with PEG (50%) vs tegaserod (31%). However this study has some limitations, including its open label design and lack of a baseline observation period. Tegaserod was more commonly associated with an increase in diarrhoea compared with placebo (6.6%vs 3.0%); more concerning was a possible association with the development of ischemic colitis.93,107,113 However, the most clinically important possible adverse events with tegaserod have been reported to be coronary and cerebrovascular events (with a 0.01% incidence). For these reasons, tegaserod was suspended in the US.53
Prucalopride Prucalopride, a dihydrobenzofurancarboxamide, is a highly selective, high-affinity 5-HT4 receptor agonist with enterokinetic properties.114 It differs from other 5-HT3 receptor agonists, such as cisapride, tegaserod, mosapride, and renzapride, which interact in part with one or more other receptors [5-HT3, 5-HT1B, and the human ether-a-go-go-related protein (hERG) channel]. In vivo studies have shown that prucalopride increases the velocity of coordinated colonic propulsion115 and triggers the peristaltic reflex in animal models.98 Prucalopride has also been shown to enhance colonic motility and transit.116–119
With regard to clinical trials of prucalopride in CC, Sloots et al.119 studied 28 patients, and showed that prucalopride enhanced colonic transit time, and increased the number of SCBM and SBM per week, and improving individual constipation symptoms. Coremans et al.120 evaluated the efficacy of prucalopride in 53 patients with CC, and showed that prucalopride was significantly more effective than placebo in softening stools, and decreasing straining and time to first bowel movement. After these early studies, three multicenter phase III studies121–123 were performed simultaneously between 1998 and 1999.
Camilleri et al.121 in the United States reported that prucalopride increased the number of the SCBM per week (47% in prucalopride vs 29% in placebo groups), and reduced the severity of symptoms in patients with CC. In another international multicenter, placebo-controlled, phase III trial, Tack et al.122 showed that prucalopride also significantly and consistently improved bowel function including SCBMs/wk, straining, stool consistency, laxative use, time to first bowel movement, and constipation severity, and satisfaction in chronically constipated patients. Finally Quigley et al.123 also reported the similar efficacy of prucalopride in 641 patients with CC compared to placebo. They showed that significantly more patients taking prucalopride 2 or 4 mg (24%) than placebo (12%) achieved the primary efficacy end point (3 or more SCBMs per week), or an increase of more than one SCBMs per week (43% and 47%vs 28% respectively). The most frequently reported adverse events of prucalopride were headache, nausea, abdominal pain and diarrhoea, which were mainly reported on the first day of treatment. There were no clinically relevant adverse events reported on prucalopride. However, there is still a theoretical concern about the potential cardiac risks associated with this drug.
Renzapride and other 5-HT4 agonists Renzapride is a mixed 5-HT4 agonist and 5-HT3 antagonist that has a stimulatory effect on gastrointestinal motility and transit.124 Renzapride has shown promising effects on gastrointestinal transit and relief of symptoms of constipation in patients with constipation-predominant IBS.125–128 Camilleri et al.126 reported that renzapride induces a clinically significant dose-related acceleration of colonic transit which was associated with improvement of bowel function in female IBS-C. Tack et al.125 also reported similar effects of renzapride on enhancing gut transit and improvement of symptoms (reducing abdominal pain, increasing the number of pain-free days, and improving stool consistency) in patients with IBS-C. In a multicenter, randomized, placebo-controlled, double-blind study in a primary healthcare setting, George et al.127 reported among 510 patients with IBS-C that renzapride 4 mg improved the frequency of bowel movements, and provided a positive effect on abdominal pain and/or discomfort in the post hoc analysis in women, although the primary endpoint did not show statistical significance and overall the results were disappointing.
Other 5HT4 agonists, such as mosapride,129 TD-5108,130,131 and ATI-7505132 are currently under development in CC.88 In addition, MKC-733 is a 5-HT3 receptor partial agonist which was shown to stimulate motility in the upper gastrointestinal tract.133 In a single-blind dose-escalation study (placebo, 0.2 and 0.5 mg b.i.d.) in 14 patients with CC, MKC-733 improved stool frequency, sensation of incomplete evacuation and gastrointestinal symptoms. This was associated with enhanced transit as evaluated by radio-opaque marker technique.134
Opioid antagonists such as methylnaltrexone and alvimopan have been investigated for opiate-induced constipation and postsurgical ileus.135,136 Alvimopan, a peripherally acting mu-opioid antagonist that does not cross the blood–brain barrier, has been investigated in patients with opiate-induced constipation.136 In a healthy volunteer study, alvimopan reversed the prolongation of orocaecal transit time produced by oral loperamide,137 morphine138 and codeine.139 In a recent randomized, double-blind, placebo-controlled study in 522 patients with opioid induced constipation, Webster et al.136 showed that alvimopan increased the mean spontaneous bowel movement frequency compared to placebo, and improved symptoms such as straining, stool consistency, incomplete evacuation, abdominal bloating/discomfort and decreased appetite.
Few studies have tested the efficacy of alvimopan in patients with idiopathic CC. Gonenne et al.139 showed that alvimopan significantly accelerated colonic transit compared to placebo in healthy volunteers. In addition Garnett et al.140 evaluated the effect of alvimopan in 23 patients with CC. They showed that alvimopan decreased colonic transit time, and improved bowel movement frequency, stool hardness, straining, discomfort and satisfaction with bowel movements in patients with CC vs placebo. In another study of 217 patients with chronic idiopathic constipation, however, Kelleher et al.141 failed to show any differences in mean weekly SCBM frequency between alvimopan and placebo. Furthermore, alvimopan did not benefit other bowel symptoms including the degree of straining or stool consistency, although alvimopan was generally well-tolerated, with a safety profile similar to placebo. Overall, further studies are needed to establish the place of alvimopan in chronic non-opioid induced constipation. Methylnaltrexone is the first quaternary ammonium opioid receptor antagonist that does not cross the blood–brain barrier in human,142,143 and has been introduced for the treatment of opiate-induced constipation.144,145 In healthy volunteers, methylnaltrexone reversed the morphine-induced delay in both gastric emptying and orocaecal transit time without affecting analgesia.146–148 Yuan et al.145 evaluated the effect of methylnaltrexone in 22 patients with constipation due to chronic methadone use. In this small pilot study, they showed that all 11 subjects after intravenous methylnaltrexone administration defecated, compared to no one after placebo administration. Recently, a relatively large phase III study was performed evaluating the treatment effect of methylnaltrexone in 133 patients with opiate-induced constipation.144 Given as a subcutaneous injection, they observed that 48% of patients in the methylnaltrexone group had defecation within 4 h after the first dose, as compared with 15% in the placebo group. Further, 52% defecated without the use of a rescue laxative within 4 h after two or more of the first four doses, compared with 8% in the placebo group. There are no studies that have tested the efficacy of methylnaltrexone in patients with idiopathic CC.
Finally, in the latest Cochrane review on mu-opioid antagonists for opioid-induced bowel dysfunction,149 alvimopan and methylnaltrexone both appear to show promise in treating in OBD. However, they concluded further data will be required to fully assess their place in therapy.
Lubiprostone, a bicyclic fatty acid derived from prostaglandin E1, is a chloride channel activator which stimulates intestinal fluid secretion, was approved by the U.S. FDA in 2006 for the treatment of chronic idiopathic constipation in adults. Lubiprostone acts on the enterocytes from the luminal side and is not systemically absorbed, which is an attractive mode of action from a safety perspective. This is associated with faster colonic transit and softer stool consistency.150,151 The drug is not currently approved in Europe. In RCTs, lubiprostone was superior to placebo in improving the frequency of spontaneous bowel movements, stool consistency and straining. The most frequently reported side-effects with lubiprostone are diarrhoea and especially nausea, usually transient, which was reported in up to 30% of patients.47,48
Linaclotide is a 14-amino acid oligopeptide that activates the luminal receptor guanylate cyclase-C on enterocytes, and leads to increased chloride and bicarbonate secretion into the intestinal lumen through a rise in cyclic guanosine monophosphate. Linaclotide was shown to accelerate ascending colon transit and to improve stool pattern and consistency in women with IBS-C.152 In CC, linaclotide was shown to improve the frequency of SCBM, straining and stool consistency.49,153
Neurotrophin-3 stimulates the development, growth and function of the nervous system and enhances colonic transit.154 In a phase II randomized, double-blind, placebo-controlled trial, subcutaneous injection of neurotrophin-3, three times per week, significantly increased the frequency of SCBMs and improved other measures of constipation.155 To date, no further development of neurotrophins for the treatment of CC seems to have occurred.
Issues in future drug development
As discussed above, a long-standing area of controversy is the need for comparator drugs in CC trials. Another issue is that many recent studies with novel agents for the treatment of CC instructed physicians to exclude patients with a known ED. On the one hand, this is a group of patients that may respond well to biofeedback therapy,40 and on the other hand it seems logical that these patients may be less responsive to agents that act on colonic motility or secretion. Both concerns, whether this is a population that is less responsive to agents targeting the colon, and whether there are more suitable ways to recognize these patients using for instance a simple questionnaire, will need to be addressed for future studies.
A final unresolved issue is the overlap between CC and IBS-C. It is clear that a large grey area of overlap between both conditions exists, and that agents that are effective in one condition are often also efficacious in the other indication. How to handle IBS-C overlap when recruiting patients for clinical trials in CC is unclear. It might be useful to investigate whether CC patients who also fulfil Rome IBS criteria have a different symptomatic response than those without a concomitant IBS diagnosis.
Areas for future research
- 1Improve clinical trial design.
- 2Improve symptom measures as efficacy outcomes.
- 3Include quality of life measures as efficacy outcomes.
- 4Include comparator laxative (such as PEG) rather than placebo in new drug studies.
- 5Undertake more studies of probiotics – looking at specific probiotic classes and probiotic mixtures.
- 6Undertake longer trials of medication for this chronic condition.
Dr Talley wishes to acknowledge the input provided by Rok Seon Choung MD for his contribution to this chapter.
Conflicts of Interest
EQ has worked as a consultant for, Procter and Gamble, Salix, Sucampao/Takeda, Movetis, Boehringer Ingleheim, Schering Plough, McNeil and Ironwood. He has been a speaker/teacher for and received Honoraria from, Norgine, Danone and Yakult. He has also received research support from Procter and Gamble.
AE has received consulting/speaker fees from the following companies in the last 2 years: Abbott Laboratories, Astra-Zeneca, Coloplast, Ferring, Pfizer and Reckitt-Benckiser.
NT has been a consultant for the following companies: AccreditEd, Addex Pharma, Annanberg Center, AryX, Astellas Pharma, AstraZeneca, Callisto PharmY, Centocor, Conexus, Dynogen, Eisai/MGI Pharma, Elsevier, F-Network, Ferring Pharmaceuticals, Gilead, Interactive Forum, In2Med, Ironwood, Johnson & Johnson, Lexicon, McNeil, Medscape, Meritage Pharma, Metabolic Pharma, Microbia, NicOx, Novartis, Oakstone Publishing, Pharma Frontiers, Proctor & Gamble, Optum HC, Salix, Sanofi-Aventis, SK Life Sciences, Spire, Steigerwald, Theravance, The Journal of Medicine, XenoPort, and Wyeth. He has also received financial support from Novartis, GlaxoSmithKline, Dynogen and Tioga.
JT has been an advisor for AGI Therapeutics, Movetis, Sucampo, Pfizer, Procter and Gamble and Theravance, and has been a speaker for Movetis.