Management of faecal incontinence and constipation in adults with central neurological diseases

Abstract

Background

People with central neurological disease or injury have a much higher risk of both faecal incontinence and constipation than the general population. There is often a fine line between the two symptoms, with any management intended to ameliorate one risking precipitating the other. Bowel problems are observed to be the cause of much anxiety and may reduce quality of life in these people. Current bowel management is largely empirical, with a limited research base. This is an update of a Cochrane review first published in 2001 and subsequently updated in 2003 and 2006. The review is relevant to individuals with any disease directly and chronically affecting the central nervous system (post‐traumatic, degenerative, ischaemic or neoplastic), such as multiple sclerosis, spinal cord injury, cerebrovascular disease, Parkinson's disease and Alzheimer's disease.

Objectives

To determine the effects of management strategies for faecal incontinence and constipation in people with a neurological disease or injury affecting the central nervous system.

Search methods

We searched the Cochrane Incontinence Group Trials Register (searched 8 June 2012), which includes searches of the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and MEDLINE In‐Process as well as handsearching of journals and conference proceedings; and all reference lists of relevant articles.

Selection criteria

Randomised and quasi‐randomised trials evaluating any type of conservative or surgical intervention for the management of faecal incontinence and constipation in people with central neurological disease or injury were selected. Specific therapies for the treatment of neurological diseases that indirectly affect bowel dysfunction were also considered.

Data collection and analysis

At least two review authors independently assessed the risk of bias of eligible trials and independently extracted data from the included trials using a range of pre‐specified outcome measures.

Main results

Twenty trials involving 902 people were included.

Oral medications

There was evidence from individual small trials that people with Parkinson's disease had a statistically significant improvement in the number of bowel motions or successful bowel care routines per week when fibre (psyllium) (mean difference (MD) ‐2.2 bowel motions, 95% confidence interval (CI) ‐3.3 to ‐1.4) or oral laxative (isosmotic macrogol electrolyte solution) (MD 2.9 bowel motions per week, 95% CI 1.48 to 4.32) are used compared with placebo. One trial in people with spinal cord injury showed statistically significant improvement in total bowel care time comparing intramuscular neostigmine‐glycopyrrolate (anticholinesterase plus an anticholinergic drug) with placebo (MD 23.3 minutes, 95% CI 4.68 to 41.92).

Five studies reported the use of cisapride and tegaserod in people with spinal cord injuries or Parkinson's disease. These drugs have since been withdrawn from the market due to adverse effects; as they are no longer available they have been removed from this review.

Rectal stimulants

One small trial in people with spinal cord injuries compared two bisacodyl suppositories, one polyethylene glycol‐based (PGB) and one hydrogenated vegetable oil‐based (HVB). The trial found that the PGB bisacodyl suppository significantly reduced the mean defaecation period (PGB 20 minutes versus HVB 36 minutes, P < 0.03) and mean total time for bowel care (PGB 43 minutes versus HVB 74.5 minutes, P < 0.01) compared with the HVB bisacodyl suppository.

Physical interventions

There was evidence from one small trial with 31 participants that abdominal massage statistically improved the number of bowel motions in people who had a stroke compared with no massage (MD 1.7 bowel motions per week, 95% CI 2.22 to 1.18). A small feasibility trial including 30 individuals with multiple sclerosis also found evidence to support the use of abdominal massage. Constipation scores were statistically better with the abdominal massage during treatment although this was not supported by a change in outcome measures (for example the neurogenic bowel dysfunction score).

One small trial in people with spinal cord injury showed statistically significant improvement in total bowel care time using electrical stimulation of abdominal muscles compared with no electrical stimulation (MD 29.3 minutes, 95% CI 7.35 to 51.25).

There was evidence from one trial with a low risk of bias that for people with spinal cord injury transanal irrigation, compared against conservative bowel care, statistically improved constipation scores, neurogenic bowel dysfunction score, faecal incontinence score and total time for bowel care (MD 27.4 minutes, 95% CI 7.96 to 46.84). Patients were also more satisfied with this method.

Other interventions

In one trial in stroke patients, there appeared to be a short term benefit (less than six months) to patients in terms of the number of bowel motions per week with a one‐off educational intervention from nurses (a structured nurse assessment leading to targeted education versus routine care), but this did not persist at 12 months. A trial in individuals with spinal cord injury found that a stepwise protocol did not reduce the need for oral laxatives and manual evacuation of stool.

Finally, one further trial reported in abstract form showed that oral carbonated water (rather than tap water) improved constipation scores in people who had had a stroke.

Authors' conclusions

There is still remarkably little research on this common and, to patients, very significant issue of bowel management. The available evidence is almost uniformly of low methodological quality. The clinical significance of some of the research findings presented here is difficult to interpret, not least because each intervention has only been addressed in individual trials, against control rather than compared against each other, and the interventions are very different from each other.

There was very limited evidence from individual trials in favour of a bulk‐forming laxative (psyllium), an isosmotic macrogol laxative, abdominal massage, electrical stimulation and an anticholinesterase‐anticholinergic drug combination (neostigmine‐glycopyrrolate) compared to no treatment or controls. There was also evidence in favour of transanal irrigation (compared to conservative management), oral carbonated (rather than tap) water and abdominal massage with lifestyle advice (compared to lifestyle advice alone). However, these findings need to be confirmed by larger well‐designed controlled trials which should include evaluation of the acceptability of the intervention to patients and the effect on their quality of life.

Plain language summary

Management of faecal incontinence and constipation in adults with central nervous system diseases

Individuals with central nervous system disease or injury have a much higher risk of loss of bowel control and severe constipation than other people. This is called neurogenic bowel dysfunction (NBD). It can be very difficult to treat constipation without causing bowel leakage, or to prevent bowel leakage without causing constipation. The time spent on emptying the bowel is nearly always much greater for these individuals. Bowel problems like this cause a lot of anxiety and distress and can reduce the quality of life of those who suffer them. This review of research about NBD could be of interest to individuals with any damage to the central nervous system caused by disease or injury, or present at birth, which has a long term effect on how their bowel works.

While there is a great deal of information on the causes of NBD, there are few studies that focus on how to practically manage the problem. Currently the usual advice is to have a good fluid intake, a balanced diet, as much physical exercise as is practical and a regular planned bowel routine. A bowel routine may include use of oral laxative medicines, suppositories or enemas; abdominal massage; digital rectal stimulation and digital evacuation of stool. The steps used will depend on the needs of each individual and some degree of trial and error is usually needed to achieve a satisfactory routine.

Only research studies where participants were allocated to either the control group (who received either no intervention or usual care) or the treatment group by chance (called randomisation) were included in this review as these studies provide the most reliable evidence. Fifteen new studies have been added in this update. Five have been removed because the drugs the studies reported on (cisapride and tegaserod) have been withdrawn from the market and are no longer available. Most of the 20 randomised studies in this review included very small numbers of participants and the study reports did not always give the information needed to be sure that the study findings were reliable.

Some oral laxatives were found to improve bowel function including psyllium, a stool‐bulking laxative (one study), and an isosmotic macrogol (one study), which were both studied in individuals with Parkinson's disease. Some suppositories and micro‐enemas used to help the bowel to open produced faster results than others (three studies) and the timing of suppository use may affect the response of the bowel (one study). Digital evacuation of stool may be more effective than oral or rectal medication (one study). The use of transanal irrigation in individuals with spinal cord injury improved bowel control, constipation and quality of life measures (one study). Three studies found that abdominal massage was helpful in reducing constipation. One study found that patients may benefit from even one educational contact with a nurse.

This review shows that there is still remarkably little research on this common problem which is so important to patients. The research evidence found by the review is generally very poor because the way the studies were carried out and reported means that the results are not reliable. It is not possible to make recommendations for care based on these studies. Managing NBD will continue to rely on trial and error until more high quality studies with larger numbers of participants and which examine the most important aspects of this problem are carried out.

Background

The focus of this review is the management of faecal incontinence and constipation in people with neurological disease or injury affecting the central nervous system.

Description of the condition

Loss of voluntary control over bowel function and constipation frequently accompany central nervous system (CNS) conditions. People need to be able to control their bowel function to enable them to participate in all areas of adult life. Working, studying, leisure activities, travelling, and developing and maintaining personal relationships may all be affected by loss of voluntary bowel control. However, it is not simply loss of voluntary control that impacts on the lives of affected individuals. Even in individuals who rarely experience frank incontinence, fear of incontinence may result in reduced social activity and isolation (Emmanuel 2010; Byrne 2002; Correa 2000; Glickman 1996; Stiens 1997). Additionally, neurogenic bowel is associated with haemorrhoids, abdominal pain, faecal impaction, rectal bleeding, rectal prolapse, anal fissure, abdominal bloating, nausea, autonomic dysreflexia and prolonged evacuation, which all impair quality of life (Correa 2000; De Looze 1998; Han 1998; Harari 1997; Lynch 2001; Menter 1997).

Hospitalisation for impaction, megacolon, constipation and volvulus is more than twice as frequent in adults with neurological diseases than in individuals without neurogenic bowel dysfunction (Sonnenberg 1994). The often time‐consuming and arduous active management of the neurogenic bowel that is required to avoid constipation and incontinence also diminishes quality of life. The issue of incontinence and its management impacts on partners and other family members, and hospital and community health services.

With the increase in the length of survival after the development or onset of some neurogenic conditions, the use of healthcare resources is likely to increase in the coming years. While effective management may have a significant cost, both to the individual in terms of impact on his or her ability to engage with their community and to the community health and social services, ineffective management may have even higher costs in terms of reduced integration, increased carer attendance, laundry costs and ultimately stoma surgery or admission to a residential care setting with significant incident and ongoing costs.

Trauma and disease of the CNS affect bowel function in a number of ways depending on the location and severity of the damage. This damage may result in partial or complete loss of sensory and voluntary motor function in the ano‐rectum and the pelvic floor. Normal modulation of the intrinsic nervous system of the bowel by the CNS is disrupted resulting in altered blood flow, enzyme secretion, and water and electrolyte transport in the gastrointestinal tract. While the intrinsic system itself usually remains intact, except in Parkinson's disease, the lack of CNS modulation results in dysmotility, increased colonic transit time and constipation. The broader effects of disability, including loss of mobility, spasticity, muscle weakness, fatigue, loss of independence in toileting, in some cases related to cognitive impairment and the use of drugs whose actions may adversely affect bowel function, all contribute to neurogenic bowel dysfunction.

The prevalence of bowel dysfunction in people suffering from neurological diseases is much higher than in the general population. While not all people with neurological conditions experience neurogenic bowel dysfunction, it is a major problem for many.

Estimates of the prevalence vary but faecal incontinence has been reported in:

• up to 70% of people with multiple sclerosis,

• up to 75% of people with spinal cord injury (Glickman 1996; Harari 1997; Krogh 1997; Menter 1997),

• 56% of individuals with cerebral palsy (Turk 1997),

• up to 68% of individuals with spina bifida (Malone 1994; McDonnell 2000; Verhoef 2005),

• up to 23% of people after cerebrovascular accidents (Brocklehurst 1985; Harari 2003; Nakayama 1997), and

• up to 24% of individuals with Parkinson’s disease (Sakikibara 2010).

Moreover, complaints of constipation are common in:

• up to 80% of spinal cord injured people (Coggrave 2009; Glickman 1996; Krogh 1997; Menter 1997),

• up to 70% of people with multiple sclerosis (Chia 1995; Hinds 1990), and

• between 20% to 81% of people with Parkinson's disease (Quigley 1996; Sakikibara 2010; Stocchi 2000).

Faecal incontinence and constipation frequently coexist in people with central neurological diseases. The broad range of reported prevalences in the literature is due to the varying definitions used for faecal incontinence and constipation and to the selection of differing populations with a range of disabilities.

Despite the significance of faecal incontinence and constipation to individuals who suffer from central neurological conditions, high quality research evidence to support neurogenic bowel management remains scanty. Consequently the development of bowel management programmes for this patient group continues to rely on tradition and experience, or trial and error.

Description of the intervention

An effective bowel management programme for a person with a neurogenic bowel involves the modulation of stool consistency, promotion of stool transit through the bowel, and effective reflex or mechanical evacuation of stool from the rectum at an appropriate time and place (Coggrave 2012b). Such care is pre‐emptive; bowel function is manipulated so that effective evacuation of the rectum occurs at a pre‐specified and predictable time, when appropriate resources such as carer assistance are available. By emptying the bowel at a chosen time incontinence is avoided; and regular emptying reduces the risk of impaction of stool due to constipation.

The strategy most commonly employed, therefore, is the establishment of a routine. Postal questionnaire data suggest an association between irregular or infrequent bowel management and constipation (Coggrave 2009). Bowel care is conducted at a set frequency and time of day and may include stimulation of the gastrocolic reflex, abdominal massage, rectal stimulants, digital rectal stimulation to produce reflex evacuation of stool, and digital removal of stool (Krassioukov 2010; Paris 2011). Oral laxatives, which include osmotic, bulk‐forming and stimulant varieties may be recommended if required. It is sometimes suggested that senna use may lead to reduced colonic activity. While robust evidence is lacking to support this concern, long term use of senna should be avoided if possible (Gattuso 1994). Individuals are often advised to modulate intake of dietary fibre and to take adequate fluids in order to influence stool consistency. However, there is scant evidence of the effectiveness of any of these interventions in individuals with any neurogenic disease. This broad approach is referred to as conservative bowel management.

Each individual should have a bowel programme which is designed to meet their specific requirements, based on an assessment of their needs and abilities, social setting and the availability of resources, particularly carers (Coggrave 2012b). An effective programme may be difficult to achieve and may prove excessively time consuming, with obvious effects on quality of life.

Even when a programme is successful in avoiding incontinence and constipation, individuals may still be plagued with other symptoms including abdominal distension, pain and ano‐rectal conditions such as haemorrhoids, fissures and rectal ulcers. Those with spinal cord injuries at or above the sixth thoracic vertebra are also vulnerable to autonomic dysreflexia, an abnormal sympathetic nervous system response to a noxious stimulus resulting in rapidly rising blood pressure with accompanying risk of brain haemorrhage and death; the response may be provoked by bowel management problems (Kavchak‐Keyes 2000). The bowel programme should be reviewed regularly to ensure that it continues to meet the needs of the individual (Coggrave 2012b).

Where conservative methods of management are unsuccessful the choices remain limited. Biofeedback has been found to be of benefit in some individuals with multiple sclerosis (Wiesel 2000). The use of gut motility stimulants in patients with neurogenic bowel dysfunction remains largely unexplored. Irrigation of the bowel may be undertaken via the anus, or through a surgically constructed irrigation stoma. While the latter has been most commonly reported in children (Dey 2003; Driver 1998), transanal irrigation, facilitated by the recent development of more sophisticated equipment, has been reported with increasing frequency in adults. Outcomes have been broadly positive though reliable identification of those individuals likely to benefit remains elusive. The significant number of people who stop using these methods over the long term merits further exploration (Faaborg 2010).

Sacral nerve stimulation, or surgically implanted neuro‐modulation, has been reported in case series to improve neurogenic bowel dysfunction in individuals with incomplete cauda equina (Gstaltner 2008) and incomplete spinal lesions (Lombardi 2010). However, the evidence is very limited and further study is required. Sacral anterior root stimulation techniques produce improvements in bowel dysfunction (Valles 2009), although to date such devices are usually implanted primarily to address bladder dysfunction and their impact on bowel function is a useful by‐product (Creasey 2001).

The use of colostomy in this context remains limited, but a number of cross‐sectional studies conducted with spinal cord injured cohorts suggest that this approach can be beneficial (Branagan 2003; Coggrave 2012a), and indeed colostomy should be considered sooner than it is currently in the course of the individual’s journey with neurogenic bowel dysfunction.

How the intervention might work

Neurogenic bowel management is comprised of a range of interventions from which a selection is usually combined in an individually designed programme or routine intended to modulate stool form or to prompt planned pre‐emptive bowel evacuation, or both. The programme is repeated regularly, with sufficient frequency to prevent constipation and faecal incontinence.

Hence continence may be achieved through conservative methods such as emptying the bowel at a chosen time when resources for the activity are available (such as the availability of carers), thus minimising the risk that the bowel will open at another inopportune time resulting in faecal incontinence. To facilitate the success of this approach, stool form is modulated through dietary manipulation or use of oral laxatives depending on the type of bowel dysfunction (reflex or areflexic) and to promote the chosen method of evacuation (a firm stool is required for manual evacuation, while a soft formed stool is needed for stimulated reflex evacuation with pharmaceutical or digital rectal stimulation).

Bowel irrigation is thought to both stimulate colonic reflex activity and to act as a mechanical bowel washout. Irrigation must conform to the need for regularity and frequency, described above. Transanal irrigation of the bowel involves passing water into the bowel via the anus in a large enough volume to reach beyond the rectum (Coggrave 2012b). An antegrade continence enema uses a surgically created irrigation stoma constructed from the appendix or a segment of caecum to allow infusion of water for irrigation of the colon from the caecum towards the rectum. Rectal stimulation, manual evacuation of stool, and oral laxatives may also be required.

Prucalopride is a selective, specific 5‐hydroxytryptamine4 (5‐HT4) agonist targeting 5‐HT or serotonin‐4 receptors, which are involved in initiating peristalsis. It increases colonic motility, hastens transit, reduces symptoms and increases satisfaction in individuals with severe chronic constipation of non‐neurogenic origin (Camilleri 2008; Tack 2009).

The precise mechanism by which sacral neuromodulation techniques can improve both faecal incontinence and constipation (and bladder function and neuropathic pain) has yet to be identified. Following placement of temporary wires and a successful trial period of two weeks with an external stimulator, permanent electrodes are placed on a sacral nerve (usually S2 or S3) under radiographic control and the stimulator is implanted. The implant provides continuous low level electrical impulses to the sacral plexus, influencing the anal canal, colon and pelvic floor; however, consistent changes in anal resting pressure and other manometric outcome measures have not been reported in the literature. How sacral neuromodulation impacts on or integrates with a bowel management programme has not been studied.

A sacral anterior root stimulator (SARS) is implanted primarily for neurogenic bladder dysfunction after spinal cord injury but it has been reported in case series to have beneficial effects on bowel dysfunction for many individuals (Binnie 1991; Creasey 2001; Kachourbos 2000; Liu 2005). A stimulator is implanted and during open surgery electrodes are placed on the second, third and fourth sacral anterior nerve roots. The stimulator is triggered by a handheld external device and high voltage, short‐lived stimulation can be applied several times daily to empty the bladder. The colon is stimulated simultaneously, resulting in increased colonic activity, reduced constipation, and sometimes defaecation during stimulation (Chia 1996; Valles 2009). Implantation remains rare; just seven of more than 1330 respondents to a recent spinal cord injury bowel management postal questionnaire reported using a SARS for bowel management (Coggrave 2009).

Intestinal stoma formation results in the diversion of faeces away from the rectum to a receptacle attached to the abdomen. The person no longer needs to evacuate stool from the rectum, resulting in much reduced expenditure of time on managing bowel dysfunction, eradication of faecal incontinence, and promotion of independence due to the location of the stoma on the abdomen. Constipation may persist, with many individuals requiring continued use of laxatives.

Why it is important to do this review

Despite the significance of faecal incontinence and constipation to individuals who suffer from central neurological conditions, research evidence to support neurogenic bowel management remains scanty. Consequently the development of bowel management programmes for this patient group continues to rely on ritual and experience, trial and error.

Objectives

To determine the effects of management strategies for faecal incontinence and constipation in people with a neurological disease or injury affecting the central nervous system.

To compare alternative management strategies for faecal incontinence and constipation in people with neurological diseases involving the central nervous system. The following comparisons were made:
 1. a particular treatment versus placebo (or no active treatment);
 2. a particular treatment versus another treatment;
 3. a particular combination of treatments versus another approach to treatment.

Methods

Criteria for considering studies for this review

Types of studies

All randomised or quasi‐randomised controlled trials.

Types of participants

People 18 years of age or more with a central neurological disease and symptoms of faecal incontinence or constipation.

Neurological diseases are defined here as any disease with a pathological process directly and chronically affecting the central nervous system (post‐traumatic, degenerative, ischaemic or neoplastic), such as multiple sclerosis, spinal cord injury, cerebrovascular diseases, Parkinson's disease and Alzheimer's disease. Acute conditions with no chronic legacy are not included, nor are diseases affecting only the peripheral nervous system, whether involving the motor, sensory or autonomic nerves (for example pelvic floor neuropathy).

Faecal incontinence is defined as recurrent uncontrolled, involuntary, or unplanned passage of faecal material or flatus, for at least one month (Whitehead 1999). It causes staining, soiling or leakage onto underwear or a pad.

Constipation is defined as having two or more of the following: straining during at least 25% of defecations, lumpy or hard stools, sensation of incomplete evacuation, sensation of ano‐rectal obstruction or blockage, or  use of manual manoeuvres to facilitate evacuation in at least 25% of defecations (for example digital evacuation, support of the pelvic floor), and fewer than three defecations a week. Loose stools should rarely be present without the use of laxatives. These criteria should have been fulfilled for the last three months prior to diagnosis, with symptom onset at least six months prior to diagnosis (Longstreth 2006). Small, unsatisfactory, inefficient or absent bowel movements during two or more attempts at toileting are also considered here as constipation.

Types of interventions

Any types of conservative management, anal irrigation or surgical management of bowel dysfunction in people with central neurological diseases were considered.

Conservative measures included dietary manipulation, oral drug or rectal evacuation therapy (for example enemas or suppositories), bowel training programmes, and assistive techniques (for example abdominal massage, Valsalva manoeuvre, digital rectal stimulation, manual evacuation). Combinations of these interventions in specific bowel programmes were also considered. Bowel irrigation per anus was included under conservative management.

Surgical measures included antegrade colonic enema (Malone procedure), percutaneous endoscopic colostomy (PEC), stoma formation, sphincter reconstruction, and sacral nerve stimulation.

Types of outcome measures

There is a range of perspectives from which the outcome of bowel management can be assessed, and this is reflected in the wide range of measures listed below. It is not known which outcomes have greater clinical relevance hence all outcome measures reported in the relevant literature were included.

1. Symptoms and clinical endpoints

• Number of people failing to achieve full continence (worse, unchanged or improved)

• Number of people with no improvement in incontinence status (worse or unchanged)

• Number of incontinence episodes per day or per week

• Number of pad or underwear changes required per day

• Number of people failing to achieve bowel evacuation at defecation (worse, unchanged or improved)

• Number of bowel motions or successes of a bowel care routine per day or per week

• Rate of improvement in bowel evacuation (self rated)

• Ease of evacuation (scored)

• Total time spent on bowel care

• Time from bowel stimulation to first flatus (minutes)

• Time from first flatus to passage of stool (minutes)

• Time from administration of medication to start of bowel movement ‐ defecation response (minutes)

• Total time of defecation (minutes)

• Consistency of stool (scored)

• Stool weight

• Number of people needing additional measures or drugs for bowel management

• Number of people with adverse effects or complications

2. Satisfaction with bowel management

• Number of people satisfied with their bowel management

• Overall evaluation of bowel management (good, poor or questionable)

• Rate of improvement in independence

• Carer satisfaction (scored)

3. Health status measures

• Health‐related quality of life measures, generic, e.g. Short Form‐36 (Ware 1993)

• Health‐related quality of life measures, condition‐specific, e.g. neurogenic bowel dysfunction score (Christensen 2006)

• Psychological health measures e.g. Hospital Anxiety and Depression Scale (Zigmond 1983)

• Activities of daily living measures e.g. Activity of Daily Life (Mahoney 1965)

4. Colonic and ano‐rectal physiological measurements

• Colonic transit time (radio‐opaque markers or scintigraphic assessment)

• Oro‐caecal transit time (hydrogen (H₂) breath test)

• Oro‐anal transit time (carmine red)

• Manometric resting pressure (cm of water)

• Pressure rise and squeeze increment on voluntary contraction (pressure or electromyography (EMG))

• Duration of pressure rise on voluntary contraction (pressure or EMG)

• Rectal sensory thresholds (by balloon distension or electrical stimulation)

• Vector symmetry index

• Saline retention test

• Staining test (present, absent)

5. Health economics

• Costs of interventions

• Resource implications of differences in outcome

• Measures of cost effectiveness

6. Other outcome measures

• Any other outcome measure later judged important by the reviewers

Note: ano‐rectal physiological measurements may not be directly comparable between different trials due to differences in equipment and the methodology used in different units. There are no objective measures of faecal incontinence. Reductions in patients' symptoms must be recognised as primary endpoints, and physiological measurements as surrogate endpoints.

Search methods for identification of studies

We did not impose any language or other restrictions on any of the searches described below.

Electronic searches

This review has drawn on the search strategy developed for the Cochrane Incontinence Group. Relevant trials were identified from the Group's Specialised Register of controlled trials which is described under the Group's module in The Cochrane Library. The register contains trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, MEDLINE In‐Process, CINAHL, and handsearching of journals and conference proceedings. The date of the most recent search of the register for this review was 8 June 2012.
 The trials in the Incontinence Group Specialised Register are also partially contained in CENTRAL.

The terms used to search the Incontinence Group Specialised Register are given below:

({design.cct*} OR (design.rct*})
 AND
 {topic.faecal*}
 (All searches were of the keywords field of Reference Manager 12, Thomson Reuters).

Additional searches were performed for the previous version of this review, these are detailed in Appendix 1.

Searching other resources

Additionally, all reference lists of identified trials were checked for other possible trials.

Data collection and analysis

Selection of studies

All three review authors examined all the citations and abstracts derived from the electronic search strategy and independently selected trials to be included in the review. Reports of potentially relevant trials were retrieved in full. Review authors were not blind to the names of the trial authors, institutions or journals. Any disagreement about any trial's inclusion was resolved by discussion.

Data extraction and management

At least two review authors independently extracted data from the included trials. Results were compared between review authors and any difference of opinion was settled by discussion. Where data may have been collected but not reported, further information was sought from the trial investigators.

Assessment of risk of bias in included studies

The risk of bias of eligible trials was assessed independently by both authors in terms of quality of random allocation concealment, number of withdrawals and dropouts, whether data were analysed on an intention‐to‐treat basis, and whether the participants and the outcome assessors were blind to the treatment provided. Where the method of allocation concealment was not clearly reported the trial authors were contacted for clarification. All disagreements were resolved by discussion. Studies were excluded if they were not randomised or quasi‐randomised controlled trials, or if they made comparisons other than those specified. These studies are listed in the table "Characteristics of excluded studies".

Measures of treatment effect

Data were analysed using the statistical package in Review Manager. For dichotomous outcomes, risk ratios (RR) and 95% confidence intervals (CI) were derived for each study. For continuous data, mean differences (MD) were calculated for each study. When appropriate, the results of included trials were combined for each outcome. For dichotomous outcomes, pooled RR and 95% CI were calculated using a fixed‐effect model. For continuous data, pooled weighted mean differences (WMD) and 95% CIs were estimated. Where an increase in outcome was beneficial (for example increased number of bowel motions) these data were recorded as a negative value to ensure consistency of the graphical results in RevMan.

Crossover trials were combined in a meta‐analysis using the generic inverse variance method. Correlations between repeat outcome measurements on the same patient were calculated, where possible, from the paired t‐test result or from any available raw data. When correlations could not be calculated, they were imputed for each outcome using the lowest estimate among the other trials. If no comparative estimate was available for an outcome, a correlation of 0.5 was imputed. Sensitivity analyses were undertaken to investigate the robustness of results to the imputed correlation value. The generic inverse variance was calculated when two or more trials reported data on a single outcome. Where there were data from only one crossover study the data were not entered into the analysis but reported narratively in the review.

Results

Description of studies

Results of the search

The literature search produced 230 records to screen, from which 58 full‐text articles were further assessed for eligibility. Twenty randomised trials met the inclusion criteria for this review. The flow of literature through the assessment process is shown in the PRISMA diagram in Figure 1.

1.

PRISMA study flow diagram.

Included studies

Trials and participants

Twenty randomised trials including 902 participants met the inclusion criteria for this review:

• eight trials were performed in people with spinal cord injuries (Christensen 2006; Coggrave 2010; Cornell 1973; House 1997; Korsten 2004; Korsten 2005; Krogh 2002; Rosman 2008),

• two in people with Parkinson's disease (Ashraf 1997; Zangaglia 2007),

• two in people with multiple sclerosis (McClurg 2011; Medaer 1999),

• two in institutionalised people with a mix of conditions including cerebral palsy, multiple sclerosis, brain injury and spinal cord injury (Dahl 2005; Emly 1998), and

• six in people with stroke (Harari 2004; Huang 2002; Jeon 2005; Mun 2011; Venn 1992; Wang 2008).

In five trials attempts to control concomitant treatment other than for the intervention under investigation were lacking or inadequate (Ashraf 1997; Cornell 1973; Harari 2004; Huang 2002; Venn 1992).

Settings

Eleven of the included trials were carried out in a single centre, two were conducted at two centres (Krogh 1997; Zangaglia 2007), one recruited from three centres (Emly 1998), one from four centres (Harari 2004) and one from five centres (Christensen 2006). The number of centres was not stated in four trials (Dahl 2005; Medaer 1999; Rosman 2008; Wang 2008).

Interventions and outcome measures

Drugs

Four trials investigated drugs used to stimulate gastrointestinal motility; three in spinal cord injury (SCI) individuals (Korsten 2005; Krogh 2002; Rosman 2008) and one in individuals with multiple sclerosis (MS) (Medaer 1999).

Prucalopride

Two trials were found of prucalopride, a systemic enterokinetic drug which is a selective and specific 5‐HT4 agonist targeting serotonin‐4 receptors involved in initiating peristalsis. Medaer and colleagues (Medaer 1999) reported a double‐blind, randomised placebo‐controlled phase II trial of prucalopride in patients with MS. The number of participants was not stated. Following a four week run‐in period, session 1 of the study examined tolerability and safety with 1 mg of prucalopride or placebo once daily for 4 weeks; session 2 examined efficacy with 2 mg prucalopride or placebo daily for 4 weeks.

Krogh and colleagues (Krogh 2002) reported a double‐blind placebo‐controlled phase II study of prucalopride. The study included 23 patients in a dose escalation study. A four week run‐in period was followed by random allocation to placebo, prucalopride 1 mg or 2 mg groups for 4 weeks. The participants in the 2 mg prucalopride group were older at injury and older in age than the other groups. Bowel management in the run‐in period and its outcomes were not described. Participants were required not to alter their diet and to refrain from the use of laxatives except in clearly defined circumstances and to a stated protocol.

Neostigmine‐glycopyrrolate

Two trials investigating the effects of neostigmine‐glycopyrrolate on bowel evacuation were reported (Korsten 2005; Rosman 2008). Neostigmine is an anticholinesterase drug which increases parasympathetic stimulation of the bowel. Glycopyrrolate, a selective anticholinergic agent, was added because it might antagonize the untoward side effects of neostigmine on the heart without interfering with its actions on the bowel.

Korsten 2005 reported evacuation of a radio‐opaque oatmeal and barium mix in response to intravenous administration of 2 mg neostigmine or 2 mg neostigmine plus 0.4 glycopyrrolate or normal saline, administered in random order once in each participant on a different day, in 13 participants. No run‐in period was reported. Individuals undertook their usual bowel care prior to the intervention and a 20 minute equilibration period was allowed following insertion of the porridge stool prior to delivery of the injection.

Rosman 2008 reported a study in which individuals received 2 mg neostigmine plus 0.4 glycopyrrolate intramuscularly for one week and normal saline as placebo intramuscularly for one week in random order, with a week of washout between the phases. A total of 21 episodes of bowel care were recorded in 7 participants. Participants continued their usual bowel care throughout. No preparation, run‐in or follow‐up was reported.

Outcome measures for these trials included use of bowel diaries; frequency, ease and duration of evacuation; evacuation of stool or not; stool consistency; modality of bowel evacuation; frequency of use of  digital rectal stimulation; digital evacuation of stool; suppositories and laxatives; time from suppository insertion to bowel movement; symptoms; side effects; gastric emptying; total and segmental colonic, oro‐caecal and oro‐anal transit time using hydrogen breath test, radio‐opaque or carmine red markers; and presence of faecal impaction. One trial (Korsten 2005) used additional outcomes: amount of a 200 ml barium oatmeal paste evacuated from rectum, seen radiographically 30 minutes after infusion; time from injection to completion of bowel care; to first flatus; to beginning of bowel movement; to end of bowel movement; and total time for bowel movement including clean up.

Oral laxatives

Three trials evaluated oral laxatives, two in individuals with Parkinson's disease (Ashraf 1997; Zangaglia 2007) and one in a mixed sample including 15 wheelchair bound institutionalised adults with MS (7), cerebral palsy (4), SCI (2) and brain injury (2) (Dahl 2005). The study by Ashraf and colleagues (Ashraf 1997) was a placebo‐controlled study of psyllium, a soluble fibre bulk‐forming laxative aimed at absorbing water and increasing bulk to encourage normal peristalsis and bowel motility. Seven out of 12 potential participants who self reported constipation in the preceding six months were selected for inclusion if during the final week of a four week run‐in period three or fewer bowel movements were reported on a standardised diary. Three people were allocated to psyllium 5.1 g daily, four to placebo for eight weeks followed by a four week washout period. Dietary habits were neither monitored nor modified during the treatment period. In particular, no attempt was made to control fibre intake.

Zangaglia and colleagues (Zangaglia 2007) randomised 57 participants to 8 weeks of isosmotic macrogol electrolyte solution (7.3 g in water) versus placebo (flavoured maltodextrine), both dissolved in 250 ml water twice daily. The length of the laxative free baseline was not specified. The macrogol dose could be varied between one and three sachets daily depending on stool form; patients were asked to keep their diet and fluid intake unchanged.

The study by Dahl and co‐workers (Dahl 2005) included participants with and without dysphagia who were given 15 g daily of an inulin‐fortified beverage versus standard modified starch‐thickened beverages for three weeks.

Outcome measures for these trials included stool frequency, symptoms related to bowel function, total stool weight, colon transit time, ano‐rectal manometry, frequency of enema or other rectal laxative administration, straining, stool consistency, diary cards and numbers of unused sachets of medication.

Chinese medicines

A trial of a Chinese medication was conducted by Huang and colleagues (Huang 2002). They compared a Chinese mixture "qirong runchang oral liquid" with a different type of Chinese medicine, "yuchi runchang ointment", in 120 stroke patients. The trial mixture was received by 90 individuals, the control or alternative treatment by 30 individuals; they were randomised 3:1. The treatment was continued for two weeks and the response to treatment was evaluated at the end of treatment and two weeks later. Possible outcomes were defined as: "cure" where stool consistency and frequency was "normal" or returned to premorbid levels, "good" where there was significant improvement in stool consistency and frequency, "partial" where there was some improvement in frequency or consistency, or "no response".

Due to difficulties with obtaining a complete translation it was not possible to fully evaluate the methods or statistical methods of the Huang study. It was not clear if the control group was actually receiving standard care; they may have received an alternative therapy. Few outcome measures were used and the main rating scale was unvalidated. Both the treatment period and follow‐up were of short duration. There was considerable variation in time since the stroke (range three days to 40 years).

Rectal stimulants

One trial (House 1997), involving 15 people with SCIs, compared two rectal stimulant suppositories: a hydrogenated vegetable oil‐based bisacodyl suppository (HVB) and a polyethylene glycol‐based bisacodyl suppository (PGB). The analysis used timed events that divided the bowel care sessions into discrete intervals. The analysis also compared the trial groups in respect to: the need to use digital rectal stimulation, incontinence, and stool quantity. The trial included 10 people with upper motor neurone spinal cord lesions, with both subacute and chronic conditions, and with complete and incomplete spinal cord lesions.

Descriptions of withdrawals and dropouts were lacking. Investigators did not routinely record whether stool was present in the rectum. The involvement of a number of different nurses to administer the bowel programme might have introduced a source of variation in the delivery of care as well as in the recording and interpretation of findings. In addition, 10 individuals were treated with a docusate and glycerine mini‐enema; this non‐randomised study arm will not be further discussed in this review.

Bowel programmes

Morning versus evening and mandatory versus optional suppository use

Venn and co‐workers (Venn 1992) compared four bowel programmes based on the use of suppositories and scheduled bowel care in consecutive individuals admitted to a rehabilitation unit following stroke. Forty‐six of all 58 stroke patients admitted in the study period were assigned to one of four protocols: 1) mandatory morning suppository; 2) optional morning suppository if the patient had not had a bowel movement within the previous four hours; 3) mandatory evening suppository; 4) optional evening suppository if the participant had not had a bowel movement within the previous four hours.

The outcome was evaluated using a study‐specific unvalidated efficiency rating (Venn 1992). This was determined by how long it took for the individual to achieve a bowel movement at the scheduled time for five consecutive days; the fewer days required, the higher the rating. The report did not include details of the characteristics of participants (for example age distribution, sex ratio, onset of disease, stage and severity of stroke, gastrointestinal symptoms, concomitant treatments). The type of suppository used was not described.

Stepwise protocol of least (for example massage) to most invasive (for example manual evacuation) interventions versus usual care

Coggrave and colleagues (2010) randomised 68 individuals with chronic SCI to usual care or to a stepwise protocol which focused on greater use of less invasive interventions (for example massage) to reduce the need for oral laxatives and more invasive interventions (for example manual evacuation of stool), and to improve bowel management outcomes.

The outcome measures employed in this trial included time from bowel stimulation to evacuation, frequency of non‐response to rectal stimulation, duration of bowel care, level of intervention required to instigate and complete evacuation, and frequency of faecal incontinence.

Physical interventions

Abdominal massage

Three trials evaluated the impact of abdominal massage.

Emly and colleagues (Emly 1998) randomly allocated 32 individuals with cerebral palsy or genetic conditions associated with abnormal muscle tone and learning disability to one of two groups. Each group received both the intervention and undertook usual care as a control phase, one group receiving the trial intervention first, the other receiving it second. A five day washout period was undertaken prior to participation in the study. The treatment and control phases were separated by a one week washout period. During washout periods all laxatives were withdrawn but enemas could be administered, if required, during all stages of the study.

McClurg et al (McClurg 2011) randomised 30 individuals with MS to receive either weekly visits from a physiotherapist or weekly visits from a physiotherapist plus specific instruction on use of abdominal massage. The physiotherapy visit included discussion of clinical symptoms and lifestyle advice about diet, fluids, toilet positioning and exercise to control for the effect of contact with a healthcare professional. In addition, the intervention group received instruction in abdominal massage.

Jeon and co‐workers (Jeon 2005) explored the benefits of abdominal meridian massage in 31 individuals with stroke; the intervention group received massage but no further details were available regarding the methods used in the study.

Outcome measures for these trials included total and segmental transit time, frequency of evacuation, stool size and consistency, a constipation assessment scale, use of enemas, Neurogenic Bowel Dysfunction score, MS impact score, bowel diaries and assessment of participant well‐being.

Transanal irrigation

Transanal irrigation in individuals with SCI was explored by Christensen and colleagues (Christensen 2006). Eighty‐seven individuals were randomised to either 10 weeks of transanal irrigation using a specially designed system with a pump and self retaining balloon catheter (Peristeen, Coloplast, Denmark), used independently after training by a nurse, or to conservative bowel management. Participants undertook bowel care every day or every two days. Follow‐up was undertaken weekly by a researcher not involved in recruitment and the teaching of participants. Data from the last four weeks were used in the analysis of the study findings.

Outcomes included the Cleveland Clinic constipation score, St Marks faecal incontinence score,  neurogenic bowel dysfunction score, modified American Society of Colorectal Surgeons Faecal Incontinence score, bowel function, influence of bowel function and management on daily activities and quality of life, duration of bowel care, and frequency of urinary tract infections.

Irritant‐contact medication versus stimulant bowel medication versus mechanical evacuation procedure with tap water enema

Cornell et al (Cornell 1973) compared three different bowel programmes, all of which included an oral wetting agent (dioctyl), amongst 60 people with a recent SCI:

1. irritant contact bowel medication (Bisacodyl) orally and rectally;

2. stimulant bowel medication orally and rectally (anthracene, senna, glycosides); and

3. a mechanical evacuation procedure with a tap water enema, plus oral stimulant laxatives and oral bulking agent (psyllium and methylcellulose).

Very little information about the presence or absence of constipation and faecal incontinence prior to the study was provided. Evaluation of the bowel programme was performed subjectively by rehabilitation nurses. Data on patients' satisfaction were not collected. Nurses had the freedom to modify the bowel programme for participants if it was "ineffective"; no indication of the type of modifications made is given and hence the degree of compliance with the allocated protocol is unclear. The carmine red total gut transit time test was reported for only 14 participants. The reported levels of SCI indicated that remaining bowel function would have varied among participants. The participants also varied in the density of their injury (that is whether or not they had complete loss of sensory and motor function). The responses of these subgroups were not reported. There was no analysis of withdrawals and dropouts. There was no run‐in period but data collection was conducted over a prolonged period for those who continued with their protocol.

The outcome measures that were employed in this trial included time from bowel stimulation to evacuation, frequency of non‐response to rectal stimulation, duration of bowel care, level of intervention required to instigate and complete evacuation, and frequency of faecal incontinence.

Other interventions

Electrical neuromuscular stimulation

Korsten and co‐workers (Korsten 2004) reported a study of the effect of neuromuscular stimulation of the abdominal wall on bowel function in individuals with SCI. Eight participants used an abdominal belt embedded with electrodes wrapped around the abdomen at the level of the umbilicus during six bowel management episodes. Use of the belt was integrated with the individual's existing bowel regimen but was randomly activated in three of the six uses. The analysis considered time from activation to first stool and total time from activation to completion of bowel care confirmed by digital rectal examination.

Electroacupuncture versus drug (cisapride)

Electroacupuncture for constipation in individuals with stroke was investigated by Wang and colleagues (Wang 2008), in 80 individuals. Participants were randomly assigned to receive electroacupuncture or cisapride 10 mg three times a day for 14 days. The effects were evaluated using a cumulative score of symptoms and an effectiveness rate, neither of which were defined in the English language abstract.

Carbonated water versus tap water

Mun and colleagues (Mun 2011) randomly allocated 34 bed‐bound older individuals with stroke to receive either tap water or carbonated water orally for two weeks. Outcome measures included laxative use, frequency of defecation and constipation symptoms.

Structured nurse assessment leading to targeted education

Harari and colleagues (Harari 2004) reported a trial involving 146 stroke patients comparing a structured nurse assessment leading to targeted education versus routine care. The intervention consisted of a one‐off structured nursing assessment followed by targeted patient and carer education and treatment recommendations according to a pre‐established protocol. A diagnostic summary and treatment recommendations were also provided to the treating physician or general practitioner (GP). The intervention that was tested included treatment recommendations based on an unvalidated protocol, and the targeted education or booklet was not described. The nurse conducting the intervention was not a specialist though she had received training and was supported by a medical colleague during the study.

The methodology did not allow the comparison of groups for important variables: physiological bowel status, diet, drug use, etc. The numbers of individuals allocated to the various management pathways in the treatment protocol were not given, nor was it stated whether participants continued with the recommended management protocol subsequent to the assessment or whether individual treatment pathways were equally effective.

One outcome measure was the percentage of bowel movements graded as normal by respondents; the grading system was not defined. Other outcome measures included the self reported number of bowel movements per week, number of bowel movements rated as normal, episodes of faecal incontinence, quality of life measures, and treatment and resource use.

Further details of the included trials are presented in the Characteristics of included studies table. There were no trials that evaluated surgery.

Excluded studies

Reasons for studies being excluded included non‐randomisation, withdrawal of the treatment (cisapride and tegaserod), and non‐neurogenic populations. Full details are included in the Characteristics of excluded studies table.

Risk of bias in included studies

The findings regarding risk of bias are summarised in Figure 2 and Figure 3.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation and allocation concealment were judged as low risk for four trials (Christensen 2006; Coggrave 2010; Harari 2004; McClurg 2011).

Two trials (House 1997; Rosman 2008) were unclear as to the method of random sequence generation but had adequate allocation concealment.

Data from four trials (Huang 2002; Jeon 2005; Mun 2011; Wang 2008) were obtained from the English abstracts but information about randomisation was not available.

The other 10 trials (Ashraf 1997; Cornell 1973; Dahl 2005; Emly 1998; Korsten 2004; Korsten 2005; Krogh 2002; Medaer 1999; Venn 1992; Zangaglia 2007) failed to mention the random sequence generation method or any method of allocation concealment.

Blinding

Five trials were reported as double‐blinded (Dahl 2005; House 1997; Krogh 2002; Medaer 1999; Mun 2011).

Participants were reported as blinded in three trials (Korsten 2004; Korsten 2005; Zangaglia 2007).

The outcome assessor was reported as blinded in three trials (Christensen 2006; Emly 1998; McClurg 2011).

Three trials (Huang 2002; Jeon 2005; Wang 2008) required translation for further clarification.

The remaining six trials (Ashraf 1997; Coggrave 2010; Cornell 1973; Harari 2004; Rosman 2008; Venn 1992) either could not use blinding due to the nature of the intervention or made no mention of blinding.

Incomplete outcome data

Eleven trials were judged to be at low risk of attrition bias because of few or no dropouts (Ashraf 1997; Christensen 2006; Cornell 1973; Dahl 2005; Emly 1998; House 1997; Korsten 2004; Korsten 2005; Krogh 2002; McClurg 2011; Rosman 2008).

In one trial (Harari 2004) the numbers of individuals allocated to the various management pathways in the treatment protocol were not given, nor was it stated whether participants continued with the recommended management protocol subsequent to the assessment, nor whether individual treatment pathways were equally effective.

Four trials (Huang 2002; Jeon 2005; Mun 2011; Wang 2008) required translation for further clarification.

Three trials used an intention‐to‐treat analysis (Christensen 2006; Coggrave 2010; Harari 2004).

Zangaglia 2007 was considered to be at high risk of bias.

In two trials (Medaer 1999; Venn 1992) attrition bias could not be assessed as descriptions of withdrawals and dropouts were not given.

Effects of interventions

In most of the trials the characteristics of the participants and the choice of outcome measures were too dissimilar for a quantitative synthesis of results to be appropriate. Confidence intervals were wide because of the small sample sizes.

1. Active treatment versus no treatment or placebo

Drugs versus no treatment

Prucalopride versus placebo

In a trial of prucalopride in 23 SCI patients (Krogh 2002), the study report stated there was a statistically significant median increase in weekly bowel movements of 0.6 over four weeks (95% CI 0.2 to 1.2) and a significant median decrease in colonic transit time after four weeks (‐38.5 hours, 95% CI ‐80 to ‐5) in the 2 mg group compared to the run‐in period. There was no change in other bowel care related parameters such as duration of bowel care, use of anal stimulation or digital evacuation, faecal incontinence or unproductive toilet visits. Changes in laxative and enema use were noted but large variations in use during run‐in made the data difficult to interpret. Mild to moderate adverse events were reported by seven out of eight and six out of eight patients in the 1 mg and 2 mg groups, respectively, during treatment. They included gastrointestinal disorders (flatulence, abdominal pain and diarrhoea) and headache. Five patients discontinued treatment due to adverse events.

A second study (Medaer 1999), reported in abstract form on only 11 patients with MS, stated that the evaluation of constipation severity improved on both 1 mg and 2 mg prucalopride. Diarrhoea was reported as an adverse event but the numbers in each group were not stated so a comparison could not be made.

Neostigmine‐glycopyrrolate versus placebo

One crossover trial (Korsten 2005) studied SCI patients and compared neostigmine‐glycopyrrolate with placebo (saline); however, the data were not suitable for analysis in RevMan. The trial report stated that prompt bowel evacuation was reported for the drug infusion compared with placebo; the volume of simulated stool expelled on a proctogram was rated as greater with the drug combination than with saline and was reported as a statistically significant difference. Side effects were reported in 92% of participants; in the drug arm of the trial facial fasciculations and abdominal cramps were experienced by most patients with injury below T10.

One small crossover study (Rosman 2008) with only seven SCI participants, analysed using the generic inverse variance method, showed a statistically significant reduction in total bowel evacuation time favouring the neostigmine‐glycopyrrolate group (MD ‐23.30 minutes, 95% CI ‐41.92 to ‐4.68) (Analysis 1.5). The study report stated that participants in both groups reduced their use of enemas, with a slight decrease in heart rate.

1.5. Analysis.

Comparison 1 Active treatment versus no treatment or placebo, Outcome 5 total bowel care time.

Isosmotic oral laxative versus placebo

One trial of 57 patients with Parkinson's disease compared isosmotic macrogol versus placebo (Zangaglia 2007). Macrogol was administered for eight weeks and the trial showed a statistically significant result favouring macrogol for improved frequency of bowel movements (MD ‐2.90, 95% CI ‐4.32 to ‐1.48) (Analysis 1.1) and numbers not responding to treatment (RR 0.29, 95% CI 0.11 to 0.72) (Analysis 1.4).

1.1. Analysis.

Comparison 1 Active treatment versus no treatment or placebo, Outcome 1 Number of bowel motions or successful bowel care routine per week.

1.4. Analysis.

Comparison 1 Active treatment versus no treatment or placebo, Outcome 4 Numbers NOT responding to treatment.

Bulk‐forming laxative versus placebo

One trial compared psyllium with placebo in seven people with Parkinson's disease (Ashraf 1997). Raw data were not reported for the number of bowel motions per week. Mean values were approximated from graphical presentations of data and therefore may not be completely accurate. Eight weeks of psyllium was associated with a significantly increased number of bowel motions (5.7 versus 3.5 mean stools per week, WMD ‐2.20, 95% CI ‐3.00 to ‐1.40) (Analysis 1.1) and stool weight (1300 versus 820 mean g/week, WMD ‐480.00, 95% CI ‐935.29 to ‐24.71) (Analysis 1.3). Colonic transit time was not altered (65 versus 58 mean hours, WMD ‐7.00, 95% CI ‐24.67 to 10.67) (Analysis 1.2). Side effects were not reported.

1.3. Analysis.

Comparison 1 Active treatment versus no treatment or placebo, Outcome 3 Stool weight (g).

1.2. Analysis.

Comparison 1 Active treatment versus no treatment or placebo, Outcome 2 Colonic transit time (hours).

Physical intervention versus no treatment

One crossover study in patients with SCI (Korsten 2004) compared electrical stimulation (abdominal belt with embedded electrodes) with no electrical stimulation. The trialists demonstrated a statistically significant result favouring electrical stimulation for total bowel care time (MD ‐29.30, 95% CI ‐51.25 to ‐7.35) (Analysis 1.5). The study also reported a reduced time to first stool.

One study (Jeon 2005) in stroke patients comparing massage with no massage showed a statistically significant result favouring massage for the number of bowel motions per week at three weeks (MD ‐1.70, 95% CI ‐2.22 to ‐1.18) (Analysis 1.1).

2. One treatment versus another treatment

Twelve trials were included in this comparison (Christensen 2006; Coggrave 2010; Cornell 1973; Dahl 2005; Emly 1998; Harari 2004; House 1997; Huang 2002; McClurg 2011; Mun 2011; Venn 1992; Wang 2008).

One drug versus another drug

Polyethylene glycol‐based (PGB) bisacodyl suppository versus hydrogenated vegetable (HVB) bisacodyl suppository

House and colleagues (House 1997) compared two types of bisacodyl suppositories in 15 patients with SCI. The study reported mean values with no measures of variance, hence the data could not be analysed in RevMan.

The trialists reported that bowel care took less time when it was initiated with a PGB suppository as compared to a HVB suppository (43 versus 74.5 minutes; P < 0.01). The effectiveness was mainly explained by a decrease in the mean time between insertion of stimulant and first flatus (15 versus 36 minutes; P < 0.026). No significant differences were reported in mean times of stool flow (HVB 6.7 minutes versus PGB 5.5 minutes). The total defecation period was shorter with PGB (20 versus 36 minutes; P < 0.037). No differences were detected in number of digital stimulations per bowel care procedure (HVB 4.4 versus PGB 4.1), time per digital stimulation (HVB 107 versus PGB 40 seconds) and total time spent performing digital stimulations during bowel care (HVB 10.0 versus PGB 2.7 minutes), although the trial was underpowered to detect these differences. The number of episodes of faecal incontinence and stool volume produced were not significantly different between the two suppositories.

One type of Chinese medicine versus another

Huang and colleagues (Huang 2002) compared a group of 90 stroke patients taking one kind of Chinese medicine (QiRong intestines‐moist oral liquid) with a smaller control group taking another Chinese remedy (YiQi intestines‐moist oral ointment). The intervention group improved more than the control group: 28/90, 31% of the study group were "cured" compared to 5/30, 16.7% in controls; 35/90, 38% of the intervention group had a "good response" compared to 9/30, 33% of controls; and 18/90, 20% of the intervention group had a "partial response" compared to 12/30, 40% of controls;no response was seen in 9 and 4 respectively (P < 0.05). The overall response rate was reported as 90% for the intervention group and 87% for the control group. Responses did not vary between individuals with stroke of different aetiologies. Two weeks after finishing the treatment, 7 of 25 cured intervention patients (23%) had developed constipation again. Follow‐up for the control group was not reported.

One bowel programme versus another bowel programme

Morning versus evening and mandatory versus optional suppository use

The trial amongst stroke patients comparing four programmes (Venn 1992) showed that 85% of all the participants that were successfully studied achieved effective bowel training within a month. The paper reported that people assigned to morning schedules achieved effectiveness earlier than those assigned to evening schedules, but the data could not be analysed in RevMan. A participant who achieved effectiveness in five days was given an efficiency rating of 16. The efficiency rating dropped by one for each additional two days required to establish effectiveness. No differences were noted between the use of mandatory or optional suppositories. When the timing for bowel management matched premorbid bowel habit, individuals were significantly more likely to achieve continence. Study results were difficult to interpret. Attempts to contact the trial authors were unsuccessful.

Irritant laxatives versus stimulant laxatives versus mechanical evacuation with tap water enema

In one trial (Cornell 1973), the authors reported that mechanical evacuation (tap water enema) without oral stimulant laxatives was associated with the most effective bowel control, using the Mann‐Whitney U test, in terms of time required for evacuation (range 10 to 14 minutes) compared with both the irritant‐ and the stimulant‐medication groups (range 67 to 102 minutes), but no measures of variance were given. Results favoured this method over the first nine weeks of the rehabilitation phase but the data did not reach statistical significance (Analysis 2.2). There were fewer accidental bowel evacuations and a smaller number of participants who reported accidents after mechanical evacuation during five out of the six weeks of treatment but this did not achieve statistical significance at nine weeks (Analysis 2.6). The number of participants with unsuccessful evacuations was smallest in the tap water enema protocol group; there was no difference between the two laxative groups. Twenty‐six participants (of 60 initially) remained in the study at its completion: nine in the enema group, four in the stimulant group and 13 in the contact‐irritant group.

2.2. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 2 Number of people with no evacuation after stimulation (poor response).

2.6. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 6 Number of people with faecal incontinence.

Transanal irrigation plus conservative bowel management versus conservative bowel management alone

One study conducted in people with SCI (Christensen 2006) compared transanal irrigation with conservative bowel management. Transanal irrigation provided statistically significant benefits compared to the conservative bowel programme for:

• constipation scores (Analysis 2.3),

• incontinence (Analysis 2.7),

• satisfaction scores (Analysis 2.5), and

• total time for bowel care (Analysis 2.8).

2.3. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 3 Constipation score (various).

2.7. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 7 Faecal incontinence score.

2.5. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 5 Satisfaction with bowel care.

2.8. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 8 Total time for bowel care.

A supplementary report stated that this resulted in a lower total cost to society (Christensen 2009).

Stepwise protocol of least (for example massage) to most invasive (for example manual evacuation) interventions versus usual care

One small trial (Coggrave 2010) reported the outcome differences between a stepwise protocol and usual individual bowel care in people with SCI. The duration of bowel care was consistently longer in the stepwise intervention group than in the control group (the mean for each week ranged from 48 to 67 minutes in the stepwise group versus 32 to 37 minutes in the control group, but this was only statistically significant in week six). Time to first stool was also longer in the intervention group (the mean for each week ranged from 15 to 28 minutes versus 10 to 14 minutes). Faecal incontinence was more frequent in the stepwise protocol group (Coggrave 2010).

Systematic use of less invasive interventions (gastrocolic reflex, abdominal massage, perianal digitation, ano‐rectal stimulation, glycerin suppositories) did not reduce the need for more invasive interventions (rectal medications, manual removal of stool, oral laxatives). There were no data suitable for analysis in RevMan and the study was underpowered.

Abdominal massage plus lifestyle advice versus lifestyle advice alone

Abdominal massage plus lifestyle advice was compared to lifestyle advice alone in patients with MS (McClurg 2011). At the end of the four week treatment period, the constipation score was statistically significantly improved (lower) in the massage group (MD ‐2.53, 95% CI ‐4.99 to ‐0.07) (Analysis 2.3) but the neurogenic bowel dysfunction score was not (Analysis 2.4). The patients reverted to their normal (baseline) constipation rates when treatment stopped after four weeks. The paper reported that there was no difference in the quality of life score.

2.4. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 4 Neurogenic bowel dysfunction score.

Structured nurse assessment leading to targeted education versus routine care

A study of 146 stroke patients by Harari and colleagues (Harari 2004) compared the outcomes of a single structured nurse assessment leading to targeted education versus routine care. There appeared to be a short term benefit (less than six months) for patients in terms of the number of bowel motions per week, but this did not persist at 12 months (Analysis 2.1). Nor was there a difference in the number of incontinent episodes per week at 12 months (Analysis 2.9).

2.1. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 1 Number of bowel motions or successful bowel care routines per week.

2.9. Analysis.

Comparison 2 One treatment versus another treatment, Outcome 9 Number of incontinence episodes between bowel cares, per day or per w.

Using a prospective one week stool diary, the percentage of bowel movements per week graded as "normal" by the patients in the intervention group was significantly higher compared to the control group at one, six and 12 months (for example at 6 months, 72% versus 55%; P = 0.027). Uncomfortable bowel movements as reported by the patients were significantly fewer in the intervention group at one month, and the mean number of bowel movements per week was significantly higher in the intervention group (5.2 versus 3.6; P = 0.005) at one and six months. At 12 months, those in the intervention group were more likely to be modifying their diets (odds ratio (OR) 3.1, 95% CI 1.2 to 8.0) and fluid intake (OR 4.2, 95% CI 1.4 to 12.2) to control their bowels, to have visited their GP for their bowel problem (OR 5.0, 95% CI 1.4 to 17.5) and to be taking two or more laxatives (OR 6.57, 95% CI 1.3 to 33.1; P = 0.02) than control group members.

Drug treatment versus bowel programme

No trials were found.

One type of diet or fibre versus another type of diet or fibre

One crossover study (Dahl 2005) compared an inulin‐fortified beverage with a standard modified starch‐thickened beverage in wheelchair‐bound institutionalised adults. The authors reported better subjective assessment of bowel movement volume over 21 days with inulin. The study also reported that stool frequency did not change. The report stated a statistically significant difference in enema administration, which decreased by 30% with inulin supplementation. It was not possible to analyse the data in RevMan.

One study (Mun 2011) compared oral carbonated water with tap water in stroke patients. The analysis comparing the two groups showed at 14 days a statistically significantly lower (better) constipation score favouring carbonated water (‐4.05, 95% BI ‐6.47 to ‐1.63) (Analysis 2.3) but there was no statistically significant difference in frequency of defecation at 8 to 14 days (Analysis 2.1).

Physical treatment versus drug treatment

One trial evaluated electroacupuncture versus cisapride 10 mg three times per day in 80 patients after stroke (Wang 2008). The paper reported that electroacupuncture was significantly better than cisapride in the cumulative score of symptoms (92.5% versus 72.5%; P < 0.05) (Wang 2008) but there were no other available data.

Note: cisapride has subsequently been withdrawn from the market.

One small crossover study (Emly 1998) compared abdominal massage performed five days a week versus the usual laxative regimen in 32 profoundly disabled institutionalised adults over a period of seven weeks. The study reported no difference in transit time nor in secondary outcome measures between the groups, but the data were not suitable for analysis in RevMan.

3. A particular combination of treatments versus another approach to treatment

No trials were found.

Discussion

Summary of main results

There is an indication from small single trials that some interventions may produce positive outcomes in certain study participants. However, due to the small number of trials and the poor methodological quality of most of the identified evidence, it is not possible to draw any reliable conclusions or to predict whether future large, well‐designed trials will confirm current findings.

The current evidence from small trials suggests that participants have a statistically significant increase (improvement) in the number of bowel motions or successful bowel care routines when bulk‐forming laxatives (Ashraf 1997) or isosmotic macrogol solution (Zangaglia 2007) are used as the intervention compared with no treatment or placebo.

There is evidence from one trial (Christensen 2006) that transanal irrigation improves constipation score, neurogenic bowel dysfunction score, faecal incontinence score and total time for bowel care in individuals with chronic spinal cord injury, and patients were also more satisfied with this method. Another older trial supported the advantages of tap water enemas over laxatives (Cornell 1973) but the data were not suitable for analysis in RevMan.

Three trials showed statistically significant improvement (reduction) in total bowel care time with polyethylene glycol‐based bisacodyl suppositories (House 1997), neostigmine‐glycopyrrolate (Rosman 2008) and electrical stimulation (Korsten 2004) when compared with placebo or no treatment. There was some evidence from Mun 2011 suggesting that carbonated water improved constipation scores. Finally, there was some evidence from three small trials to support the use of abdominal massage (McClurg 2011; Jeon 2005; Emly 1998).

Timing evacuation to fit in with premorbid bowel habits is considered to contribute to the efficacy of a bowel management routine, though with little supporting evidence. The study by Venn and colleagues (Venn 1992) offers some evidence, although their adoption of an unvalidated "efficiency rating" makes the results of the study difficult to interpret. This remains the only study to address this issue. They explain their finding that a morning routine was more successful than an evening one by reference to stimulation of reflexes related to eating breakfast (gastrocolic reflex), although some trials have suggested that this may be diminished or absent in some neurogenic bowel conditions (Aaronson 1985).

Overall completeness and applicability of evidence

This review identified only small trials of almost uniformly poor methodological quality; the evidence base for this area of care is very scanty. Although this review now includes 20 trials, the bowel management interventions, populations, comparators and outcome measures were so varied that it was impossible to pool data and hence to draw strong conclusions.

In most of the trials, the authors did not report what "usual care" comprised, which made it difficult to classify and therefore generalise the data. There was little information on the acceptability of the interventions to participants and seldom on the impact on their quality of life, which might arguably be the most important factor for this group of patients.

For example, the value of using drugs that reduce transit time and may therefore produce a softer stool needs careful consideration in the bowel management of the neurologically disabled; a softer stool is more difficult to control and is often associated with increased faecal incontinence. An increase in the frequency of bowel evacuation may be equally unwelcome in this population due to the increased use of time and other resources, including carers' time, which it implies.

There is a need for trials to use outcome measures that reflect the wider implications of neurogenic disability and are specific to this patient group. In the trial conducted by Cornell (Cornell 1973) a mechanical method (tap water enema) was the most effective management method in rehabilitating spinal cord injured individuals; but as the authors pointed out, this approach to bowel care was not acceptable to injured individuals and their families for use at home. This underlines the importance of collaborative development of bowel management programmes which are acceptable to the people who are going to use them and those who may be involved in their implementation. It also highlights the different requirements of care in the hospital and home setting, which should be borne in mind when researching management methods.

Nurses are the healthcare professionals who deal most intimately with the problems of bowel management. They are involved in hands‐on care and in supporting, educating and advising patients and their families and carers in the long term. Nurse‐led care has been found to be effective in the treatment of non‐neurogenic faecal incontinence in a hospital outpatient setting (Norton 2001). While the impact of the intervention in the study by Harari and colleagues (Harari 2004) was relatively short lived, it suggests that even a one‐off intervention involving face‐to‐face nurse contact can be effective in improving outcomes in patients with bowel dysfunction. The role of the nurse in assessment, information giving and empowerment of individuals with neurogenic bowel dysfunction merits further investigation.

Rectal stimulation is an important part of the bowel management process for many individuals with neurogenic bowel dysfunction, providing a trigger for bowel evacuation at a chosen time. This can be achieved by mechanical or pharmaceutical means. Despite the frequent clinical use of suppositories and enemas for this purpose, research evidence for the efficacy of the various products and approaches is seriously lacking. The small study reported by House and colleagues (House 1997) needs to be repeated, with a larger sample and a range of other products.

Good evidence for the benefits of transanal irrigation in individuals with chronic spinal cord injury and problematic bowel management was provided in a well‐designed and appropriately powered randomised controlled trial (Christensen 2006). Further work is needed to explore the benefits in newly injured individuals.

Quality of the evidence

The evidence base for management of neurogenic bowel dysfunction across all aetiologies is extremely poor. In most of the trials the characteristics of participants and the choice of outcome measures were too dissimilar for a quantitative synthesis of results to be appropriate. Confidence intervals were wide because of the small sample sizes.

Most interventions were only addressed in single small studies. However, the risk of bias was judged to be low in two of the trials which favoured an active intervention (transanal irrigation: Christensen 2006; and abdominal massage: McClurg 2011).

Agreements and disagreements with other studies or reviews

Oral laxatives are commonly used by individuals with neurogenic bowel dysfunction despite the lack of evidence to support their prescription. In the able‐bodied, prucalopride has been reported to reduce colonic transit time, improve stool consistency and increase frequency of bowel evacuation in chronically constipated and older people (Coremans 2003; Emmanuel 2002; Quigley 2009; Sloots 2002). However, the use of prucalopride in people with spinal cord injuries, reported by Krogh and colleagues (Krogh 2002), did not show any statistically significant improvement in clinically relevant parameters such as duration of bowel care, reduction in episodes of faecal incontinence or the need for digital evacuation of stool. The rate of side effects from the drug was high even if the events were not serious.

While Huang and colleagues (Huang 2002) do not propose a mode of action for their herbal remedy, it seems likely that increased motility may at least in part account for its efficacy. The use of another herbal remedy is reported by Suehiro and colleagues (Suehiro 2005). They found that Dai‐ken‐chu‐to and Keish‐bukuryo‐gan given to patients from the first day post‐colorectal resection improved bowel motility and reduced length of hospital stay. Publications in western journals of studies of Chinese or herbal medicine are very infrequent, and obtaining translations of papers written in Chinese is difficult. However, the report by Huang and Li (Huang 2002) suggests that there may be much to learn from this traditional approach to health care.

Modulation of stool consistency is the purpose of bulk‐forming laxatives such as ispaghula husk, psyllium, methylcellulose and sterculia. Ispaghula husk has been reported to improve evacuation parameters in a sample of healthy volunteers (Davies 1998). Its efficacy has also been reported in a study of people with simple constipation (Dettmar 1998). Psyllium has been reported to be effective in the management of simple (McRorie 1998) and chronic (Ashraf 1997) constipation and in constipation among older people (Cheskin 1995). However, some studies have suggested that the neurogenic bowel does not respond in the same way to fibre supplementation (Cameron 1996; Menardo 1987). The very small study of psyllium by Ashraf and colleagues (Ashraf 1997) remains the only reported work conducted in a neurologically impaired population and the role of fibre supplementation in neurogenic bowel management is yet to be determined.

Authors' conclusions

Implications for practice.

While the evidence reviewed here is generally of poor quality, and few studies were identified, there was some limited evidence from individual trials in favour of bulk‐forming laxatives (psyllium), isosmotic macrogol laxative, abdominal massage, electrical stimulation and an anticholinesterase and anticholinergic drug combination compared to no treatment or controls. There was also evidence in favour of transanal irrigation (compared to conservative management), oral carbonated (rather than tap) water and abdominal massage with lifestyle advice (compared to lifestyle advice alone).

Implications for research.

There is a need for rigorous randomised trials using clinically relevant outcome measures specific to this patient group, and particularly taking account of personal dignity, acceptability and effect on quality of life. The variety of studies included in the review underscores the complexity of neurogenic bowel management. In this area, where effective management rather than cure is the aim, the significance of the nursing contribution in establishing and maintaining effective care needs further evaluation. Despite the addition of 15 new trials since the last publication of this review, trials are mostly small and evidence for neurogenic bowel management remains poor.

Further trials are required in all aspects of this care including the role of fibre and diet, the use and long term implications of laxatives, the use of digital rectal stimulation and manual evacuation, and the relative benefits of rectal stimulants (suppositories and enemas) and facilitative techniques such as abdominal massage and positioning. Studies to determine selection criteria for transanal irrigation, use in patients with subacute spinal cord injury, and the reason for decline in usage over time are also needed.

In the absence of reliable information about what is important to patients regarding the outcomes of bowel management, very limited validated outcome measures, and with little evidence to support any intervention, it is not possible to prioritise where experimental research activity should be focused.

The development of condition‐specific outcome measures, particularly for quality of life, will benefit the objective evaluation of interventions in the future. The development and adoption of agreed definitions of faecal incontinence and constipation in individuals with neurogenic bowel dysfunction would also greatly improve comparability of studies and facilitate meta‐analysis of the findings of smaller trials.

The duration and type of the underlying neurological condition may also be significant in the response of people to interventions, whether due to continuing physical changes related to the condition (improvement or deterioration) or due to prior bowel management interventions that is stimulant laxatives. Harari and colleagues (Harari 2003) reported that one third of stroke patients who had faecal incontinence at three months were continent at one year. It may be appropriate to select participants according to the duration of their neurological condition. A fuller description of participants and interventions is needed.

No randomised controlled studies have been conducted of surgical interventions for neurogenic bowel management, although a study based on a registry of sacral nerve stimulation in incomplete spinal cord injured patients found they benefited from the intervention (Jarrett 2005). Further studies on both sacral anterior root and sacral nerve stimulation are required. Research into the low uptake of functional electrical stimulation implants for continence, to date, might be of value in shaping future developments.

The content and methods of delivery of assessment and teaching also require formal evaluation, as do the optimal frequency and content of routine assessments required to maintain efficacy and avoid complications.

The development of an effective bowel management programme requires a number of inputs which include oral and rectal laxatives (stimulants, osmotics, softeners and bulkers), physical interventions such as anorectal stimulation, digital evacuation, enema or transanal irrigation, abdominal massage, and the adoption of a routine (of intervention and timing). These interventions function against a backdrop of other variables such as diet and fluid intake, level of activity, posture adopted for bowel evacuation, premorbid bowel habit, concomitant medications, psychological factors and the pathology of the specific neurogenic condition. The reduced mobility and other physical disabilities imposed by the underlying condition must also be considered. All of these factors interact and are variable between individuals; however, a clearer understanding of the contribution of each factor to the bowel management process will improve the ability to establish effective bowel care with a smaller degree of "trial and error" than is currently the case.

The development of bowel management protocols continues to be empirical, and achievement of an effective programme will involve trial and error. It is widely accepted that the development of interventions and bowel management programmes should be based on a comprehensive, individualised, patient‐centred assessment and use interventions of proven effectiveness. Such an assessment should include the physical abilities of the individual, availability of care, and the social setting as well as clinical issues. The chosen lifestyle and preferences of the individual should be central to planned care. The education of patients and their carers should go hand in hand with the planning of the bowel programme. The bowel programme should be reviewed regularly.

What's new

Date	Event	Description
21 November 2013	New citation required but conclusions have not changed	15 trials added: Christensen 2006; Coggrave 2010; Dahl 2005; Emly 1998; Jeon 2005; Korsten 2004; Korsten 2005; McClurg 2011; Medaer 1999; Mun 2011; Rajendran 1992a; Rosman 2008; Sullivan 2006a; Wang 2008; Zangaglia 2007
21 November 2013	New search has been performed	15 trials added: Christensen 2006; Coggrave 2010; Dahl 2005; Emly 1998; Jeon 2005; Korsten 2004; Korsten 2005; McClurg 2011; Medaer 1999; Mun 2011; Rajendran 1992a; Rosman 2008; Sullivan 2006a; Wang 2008; Zangaglia 2007

History

Protocol first published: Issue 2, 2000
 Review first published: Issue 4, 2001

Date	Event	Description
9 October 2008	Amended	Converted to new review format.
22 February 2006	New citation required and conclusions have changed	Substantive amendment. Three studies added, Huang 2002, Harari 2004, Krogh 2002 Two review authors (CN, MC) assessed the eligibility of three trials which resulted in the inclusion of all three into the review. Both reviewers interpreted the results and contributed to the writing of the review.
1 August 2003	New search has been performed	Issue 3,2003. Two reviewers (MC, CN) assessed the list of eligible trials and their quality leading to one extra excluded trial (Kim 1991). In the original review, one review author (P. Wiesel) wrote the initial protocol for the review. Two reviewers (P. Wiesel, C. Norton) selected studies to be included in the review. All three review authors (P. Wiesel, C. Norton, M. Brazzelli) independently assessed the quality of eligible studies and extracted data from identified trials. All three reviewers interpreted the results and contributed to the writing of the final version of the review.

Appendices

Appendix 1. Additional searches performed for the previous version of this review

Additional searches were performed by the lead author for the previous version of this review, these are detailed below.

Electronic databases 
 The Cochrane Central Register of Controlled Trials (Issue 2, 2005), MEDLINE (January 1966 to May 2005), and EMBASE (January 1998 to May 2005) were searched in May 2005. The neurological search terms used were: central nervous system diseases, brain diseases, movement disorders, spinal cord diseases, demyelinating diseases, neurodegenerative diseases, neuromuscular diseases.

Data and analyses

Comparison 1. Active treatment versus no treatment or placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of bowel motions or successful bowel care routine per week	3		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 diet / fibre vs usual care / no treatment	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 massage versus usual care / no treatment	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 laxative versus placebo	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Colonic transit time (hours)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1 diet / fibre vs usual care / no treatment	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Stool weight (g)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1 diet / fibre vs usual care / no treatment	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Numbers NOT responding to treatment	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
4.1 laxative versus placebo	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 total bowel care time	2		Mean Difference (Fixed, 95% CI)	Totals not selected
5.1 neostigmine‐glycopyrrolate vs placebo	1		Mean Difference (Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 electrical stimulation vs no electrical stimulation	1		Mean Difference (Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 2. One treatment versus another treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of bowel motions or successful bowel care routines per week	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 assessment and education versus usual care	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 carbonated water versus tap water	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Number of people with no evacuation after stimulation (poor response)	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2.1 irritant‐contact laxative versus stimulant laxative	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 irritant‐contact laxative versus tap water enema	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 stimulant laxative versus tap water enema	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Constipation score (various)	3		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1 transanal irrigation versus usual bowel programme	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 abdominal massage + lifestyle versus lifestyle alone	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 carbonated water versus tap water	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Neurogenic bowel dysfunction score	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
4.1 transanal irrigation versus usual bowel programme	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 abdominal massage + lifestyle versus lifestyle alone	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Satisfaction with bowel care	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
5.1 transanal irrigation versus usual bowel programme	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Number of people with faecal incontinence	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
6.1 irritant‐contact laxative versus stimulant laxative	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 irritant‐contact laxative versus tap water enema	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.3 stimulant laxative versus tap water enema	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Faecal incontinence score	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
7.1 transanal irrigation versus usual bowel programme	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 Total time for bowel care	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
8.1 transanal irrigation versus usual bowel programme	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9 Number of incontinence episodes between bowel cares, per day or per w	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
9.1 stepwise assessment and education versus usual care	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Ashraf 1997.

Methods	Randomised single‐blind trial Intention‐to‐treat analysis: no dropouts
Participants	Setting and country: Movement Disorders Clinic, USA Inclusion criteria: 7 Parkinson's disease patients Exclusion criteria: not mentioned Age: mean age 66 years, range 54‐80 Sex: 3 women, 4 men Other characteristics: reporting fewer than 3 stools during the final week of a 4‐week run‐in period. Duration of Parkinson's disease range 4‐14 years and mean duration of constipation was 9.4 years
Interventions	Psyllium (5.1 g daily) versus placebo during an 8‐week period followed by a 4‐week washout period. Four patients received placebo and 3 psyllium
Outcomes	Stool frequency and symptoms related to bowel function (stool consistency, straining effort, occurrence of anal pain, presence of a sensation of incomplete evacuation). Each symptom was scored on a visual analogue scale that ranged from 1 to 7. Total stool weight during the final week of each phase. Colon transit time and ano‐rectal manometry (basal anal sphincter pressure profile, the squeeze response, the recto‐anal inhibitory reflex, the threshold for rectal sensation) at entry, at the end of the 4th, 12th and 16th weeks
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	no mention of method
Allocation concealment (selection bias)	Unclear risk	no mention of method
Blinding (performance bias and detection bias) All outcomes	Unclear risk	no mention of blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	all 7 completed and reported

Christensen 2006.

Methods	Randomised controlled trial Intention‐to‐treat analysis: yes
Participants	Setting and country: 5 spinal cord injury centres in European countries (Denmark, Italy, UK, Germany, Sweden) Inclusion criteria: 87 patients with spinal cord injury and neurogenic bowel dysfunction. SCI at any level; at least 3 months after injury and at least one of the following symptoms: 30 minutes + on bowel management; episodes of FI once or more per month; symptoms of autonomic dysreflexia before or during defecation; abdominal discomfort before or during defecation Exclusion criteria: coexiting major physical problems; using irrigation already; bowel obstruction or inflammatory bowel disease; cerebral palsy or stroke; multiple sclerosis; diabetic polyneuropathy; previous major abdominal or perineal surgery; pregnancy or lactation; spinal shock; mental instability; 5mg+ prednisolone; implant for sacral nerve stimulation Age: mean 47.5 years (transanal); 50.6 years (conservative) Sex: female 25; male 62 Other characteristics: details given of ASIA scores and functional level
Interventions	10 weeks of transanal irrigation (Peristeen, Coloplast, Denmark) using a specially designed system with a pump and self retaining balloon, used independently after training by a nurse, compared with conservative bowel management (best supportive bowel care using a guideline without using irrigation). Both each day or 2 days
Outcomes	Primary endpoint: Cleveland Clinic constipation score and St Marks faecal incontinence Secondary outcomes: neurological bowel dysfunction score; modified American Society of Colorectal Surgeons FI score; bowel function; influence on daily activities; general satisfaction; impact on quality of life; urinary tract infection; time for bowel care
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	computer generated sequence: block randomisation by centre
Allocation concealment (selection bias)	Low risk	sealed numbered envelopes
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not possible for participants but independent observer "who had not participated in the training" for outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	few missing data; "intention to treat analysis with baseline values carried forward for patients who did not complete"

Coggrave 2010.

Methods	Randomised controlled trial Intention‐to‐treat analysis: yes
Participants	Setting and country: spinal cord injury centre in Stoke Mandeville, UK Inclusion criteria: 68 traumatic or non‐traumatic spinal cord injury. Any ASIA grade, discharged from rehabilitation one year or more Exclusion criteria: pregnancy, current or past bowel disease or major surgery, stoma Age: intervention 49.5 years (range 24‐73), control 47 years (range 27‐62) Sex: women 23, men 45 Other characteristics: flaccid bowel 9, reflex bowel 59
Interventions	Intervention: stepwise protocol of increasingly invasive evacuation methods (from massage to manual evacuation) Control: usual care in terms of type, number and order of interventions to achieve evacuation
Outcomes	Primary: duration of bowel care and level of intervention required to complete evacuation Secondary: time to stool, level of intervention at which evacuation began, frequency of faecal incontinence
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	computer generated block randomisation for subjects to each block
Allocation concealment (selection bias)	Low risk	sealed opaque consecutively numbered envelopes, administered by someone not involved in study
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not possible to blind participants or care givers. No mention of outcome assessor blinding
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	intention‐to‐treat analyses, 10 discontinued: 4 control, 6 intervention

Cornell 1973.

Methods	Randomised trial Intention‐to‐treat analysis: no.
Participants	Setting and country: Comprehensive Rehabilitation Institute, USA Inclusion criteria: 60 newly admitted SCI inpatients Exclusion criteria: patients re‐admitted or admitted 6 months after the SCI Age: mean 27.2, range 13 to 58 years Sex: 8 women, 52 men Other characteristics: mean duration of SCI 43.9 days
Interventions	Patients randomly assigned to one of 3 bowel programmes: A) irritant‐contact laxative medication, oral and rectal (bisacodyl, Dulcolax), with oral wetting agent (dioctyl) B) stimulant laxative medication, orally and rectally, with oral wetting agent (dioctyl) C) mechanical evacuation with tap water (500 ml) instilled into the rectum, and oral wetting agent (dioctyl) and oral bulking agents (psyllium and methylcellulose) Longitudinal assessment of 6 periods were made: first week of admission, second week, week 3, week 6, week 9 and week of discharge
Outcomes	Duration from stimulation until evacuation occurred Number of unsuccessful attempts at evacuation after stimulation (data from week 9 used) Number of unstimulated or accidental evacuations (faecal incontinence) (data from week 9 used) Staining test as oro‐anal transit indicator (carmine test) Global nursing evaluation of the bowel programme rated as "good to fair", "poor" and "questionable"
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"randomly assigned"
Allocation concealment (selection bias)	Unclear risk	no description
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned, blinding of intervention not possible
Incomplete outcome data (attrition bias) All outcomes	High risk	participants dropped from the study if they "did not have an optimal programme" (A: 7/20, B: 16/20 C: 11/20)

Dahl 2005.

Methods	Randomised double‐blind crossover study Intention‐to‐treat analysis: not mentioned
Participants	Setting and country: Canada, setting not specified Inclusion criteria: 15 wheelchair bound institutionalised adults, multiple sclerosis 7, cerebral palsy 4, spinal cord injury 2, brain injury 2 Exclusion criteria: Age: under 60 years of age Sex: not stated Other characteristics: 10 with dysphagia and 5 without dysphagia
Interventions	15g daily inulin‐fortified beverage versus standard modified starch‐thickened beverages for 3 weeks each
Outcomes	Chart record of bowel frequency, enema administration, urine samples for calcium, sodium and potassium
Notes	No washout period between crossover
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Low risk	double‐blind but no details given
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	blood results on 12/15; other results appear complete

Emly 1998.

Methods	Randomised crossover study Intention‐to‐treat analysis: no
Participants	Setting and country: UK single long stay residential setting and 2 group homes Inclusion criteria: 32 profoundly disabled institutionalised adults with cerebral palsy or genetic conditions associated with abnormal muscle tone regular use of laxatives or enemas for at least 12 months Exclusion criteria: none stated Age: mean 42.25 years (massage); 43.53 years (laxatives) Sex: 14 men, 18 women Other characteristics: 71% non‐ambulant and totally physically dependent
Interventions	Abdominal massage 5 times a week by physiotherapist or nurse for 20 minutes for 7 weeks versus usual laxative ± enema as a rescue for 7 weeks then crossover. All laxative medication was stopped for both groups for washout period
Outcomes	15 days baseline and 10 days after treatment used for assessment. Gastrointestional and segmental transit time; stool frequency; stool size and consistency; need for enemas; assessment of patient well‐being
Notes	7 day washout period between phases with no laxatives or massage
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Low risk	radiologists interpreting transit time were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	follow‐up for 30/32

Harari 2004.

Methods	Randomised trial Intention‐to‐treat analysis: yes
Participants	Setting and country: 3 stroke rehabilitation units and community in UK Inclusion criteria: 146 participants, 122 community and 24 stroke rehabilitation inpatients. Had a stroke between 1 month and 4 years previously. Patients screened by questionnaire to identify bowel dysfunction (constipation or faecal incontinence) according to pre‐set definitions Exclusion criteria: people reporting acute diarrhoea or colonic disease other than diverticular disease Age: intervention 72.2 ± 10.2 years; control 72.9 ± 9.6 years Sex: 59 women, 87 men Other characteristics: pre‐stroke self reported constipation and faecal incontinence were comparable to similarly aged populations
Interventions	Intervention: one‐off assessment by a nurse leading to 1. targeted patient and carer education; 2. provision of a booklet; 3. diagnostic summary and treatment recommendations sent to patient's GP and ward physician if in hospital Control: routine care
Outcomes	Primary outcome was bowel movements per week. Secondary outcomes were percentage of bowel movements graded as normal by patients and number of episodes of faecal incontinence all measured by postal prospective diary at 1, 3, 6 and 12 months. Other outcomes were bowel related symptoms, stool diary, visual analogue scores for severity rating, quality of life (bowel related and SF‐12) and self reported treatment and resource use
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	external process using computer generated numbers
Allocation concealment (selection bias)	Low risk	closed envelopes
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	24% dropout at 6 months: 51/73 in usual care and 55/73 in intervention group

House 1997.

Methods	Randomised double‐blind controlled trial Intention‐to‐treat analysis: no, but no dropouts
Participants	Setting and country: Physical Medicine and Rehabilitation Center, USA Inclusion criteria: 15 SCI inpatients; duration of SCI range from 3 months to 45 years. 3 months or longer since SCI, injury above T12, lack of anal sensation, lack of voluntary anal sphincter contraction, stable bowel program, absence of known gastrointestinal disease Exclusion criteria: not mentioned Age: age range 26 to 61 years Sex: not mentioned Other characteristics: subject history reviewed for consistency in bowel management
Interventions	Comparison of two bowel programmes based on the use of different rectal stimulants: 1) hydrogenated vegetable oil‐based bisacodyl suppository (HVB) 2) polyethylene glycol‐based bisacodyl suppository (PGB) Randomised sequence for each of the six scheduled bowel care sessions
Outcomes	Time to flatus, time to stool flow, defecation duration, total bowel care duration Duration and frequency of digital stimulation and manual evacuation Stool amount (0 ‐ none, 1 ‐ minimal, 2 ‐ small, 3 ‐ moderate, 4 ‐ large, 5 ‐ very large) Number of episodes of faecal incontinence
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	reported as predetermined
Allocation concealment (selection bias)	Low risk	central pharmacy allocated
Blinding (performance bias and detection bias) All outcomes	Low risk	reported as double‐blind
Incomplete outcome data (attrition bias) All outcomes	Low risk	all 15 completed

Huang 2002.

Methods	Randomised controlled trial Intention‐to‐treat analysis: not stated
Participants	Setting and country: outpatients and inpatients of a single hospital Inclusion criteria: 120 stroke patients, 90 patients in treatment group, 30 controls. Patients with a history of stroke evidenced by CT/MRI scan and constipation with bowel opening less frequently than every 72 hours and hard stool causing straining during defecation or difficulty in defecation Exclusion criteria: not stated Age: treatment group ‐ average 59 years ± 7 years, control group 60 years ± 9 years Sex: treatment group ‐ 41 male, 49 female, control group 13 male, 17 female Other characteristics: other Chinese medicine criteria also applied, other organic causes of constipation excluded
Interventions	A (90): Quiong Runchang oral liquid ‐ a Chinese herbal medicine for intervention group B (30): Yichi Runchang ointment for controls. Treatment for 2 weeks
Outcomes	Definitions: Cure ‐ normal stool consistency and frequency or a return to pre‐morbid level and all symptoms subsided. Good response ‐ significant improvement (near normal) in frequency of defecation and consistency of stool or slightly hard stool and frequency more than once in 72 hours and almost complete subsidence of symptoms of constipation. Partial response ‐ frequency of stool more than or equal to once a day or improvement in consistency of stool and improvement of symptoms of constipation. No response ‐ reaching none of the above
Notes	Abstract only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	randomised 3:1
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	intention‐to‐treat analysis, dropouts not mentioned

Jeon 2005.

Methods	Randomised controlled trial Intention‐to‐treat analysis: not stated
Participants	Setting and country: one centre, Korea Inclusion criteria: 31 stroke patients Rome II criteria for constipation. 21 more than one year after stroke Exclusion criteria: Age: median 63 years Sex: 13 men, 18 women Other characteristics: also used a constipation assessment for participant selection
Interventions	Abdominal meridian massage versus no massage
Outcomes	Frequency of defecation, constipation assessment scale
Notes	Abstract only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	not mentioned

Korsten 2004.

Methods	Randomised crossover design Intention‐to‐treat analysis: data on all 8 participants
Participants	Setting and country: USA Inclusion criteria: 8 spinal cord injured patients, 6 tetraplegic, mean duration of injury 13 years (range 2 ‐33), fewer than 2 bowel motions per week all with stable bowel regimes Exclusion criteria: Age: average 48 ± 14 years Sex: all men Other characteristics: none stated
Interventions	Abdominal belt with embedded electrodes applied at level of umbilicus used during bowel care, randomly activated or not activated, during 6 sessions of bowel care over 2 weeks
Outcomes	Time to first stool, time for total bowel care
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	no mention
Allocation concealment (selection bias)	Unclear risk	no mention
Blinding (performance bias and detection bias) All outcomes	Low risk	participants were blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	data on all 8 participants

Korsten 2005.

Methods	Randomised crossover blinded study Intention‐to‐treat analysis: all subjects completed
Participants	Setting and country: Veterans centre USA Inclusion criteria: 13 spinal cord injury patients, 5 quadriplegics Exclusion criteria: cardiac or renal disease Age: mean 46, range 25‐69 years Sex: not mentioned Other characteristics: mean duration of injury 14 years, range 1‐31; bowel care required 1‐2 hours at the time of the study
Interventions	Intavenous infusion of neostigmine 2mg versus 2mg neostigmine plus 0.4mg glycopyrrolate versus normal saline
Outcomes	Amount of a 200ml barium oatmeal paste evacuated from rectum seen radiographically 30 minutes after infusion (blinded assessor, graded 0‐4), blood pressure and pulse rate, total and central airway resistances
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Unclear risk	subjects blinded as to which infusion receiving
Incomplete outcome data (attrition bias) All outcomes	Low risk	all subjects completed

Krogh 2002.

Methods	Double‐blind placebo‐controlled, pilot, phase II, dose escalation study Intention‐to‐treat analysis: no
Participants	Setting and country: two Danish centres. Inclusion criteria: 23 randomised ‐ 7 received placebo, 8 prucalopride 1 mg and 8 prucalopride 2 mg. SCI, more than 6 months since spinal cord injury; history of constipation subsequent to injury defined as 2 or fewer spontaneous bowel movements per week or needing manual assistance with evacuation in more than 50% of bowel movements per week Exclusion criteria: cauda equina lesions, drug induced constipation, organic causes of constipation, impaired renal function Age: 18‐60 years Sex: 5 women, 16 men Other characteristics: if participants reported constipation pre‐spinal cord injury, developed constipation aged >50 years or had macroscopic blood in their faeces underwent colonoscopy
Interventions	Study consisted of 2 sessions each preceded by a 4 week run‐in period when patients' own bowel care recorded. For 1 group this was followed by random allocation to placebo or prucalopride 1 mg for 4 weeks. When safety of this dose established another group was randomised to placebo or prucalopride 2 mg for 4 weeks
Outcomes	Bowel diary including time of evacuation, use of anal stimulation, stool evacuated or not, stool consistency, use of digital evacuation, time taken for evacuation, presence of faecal impaction. Patient‐based outcomes were changes in frequency and consistency of bowel evacuation, abdominal pain and distention, incontinence, laxative use, time to first bowel movement, time taken for defecation. Investigator outcomes were global assessment of treatment efficacy on a 5 point Likert scale and total colonic transit time
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Low risk	double‐blind
Incomplete outcome data (attrition bias) All outcomes	Low risk	7/8 completed 2mg intervention, all others completed

McClurg 2011.

Methods	Randomised study Intention‐to‐treat analysis: no
Participants	Setting and country: multicentre, UK Inclusion criteria: 30 people with confirmed multiple sclerosis, recruited via newsletter, websites, MS groups and day centres, Rome 11 criteria for constipation Exclusion criteria: history of bowel disease Age: mean 55 years (SD 13, range 34‐83) Sex: 18 women, 12 men Other characteristics: EDSS score 2.5‐6; 10% could walk unaided, 40% wheelchair bound
Interventions	Abdominal massage daily by participant or carer (with supporting DVD) + lifestyle advice versus lifestyle advice alone. 4 weeks intervention period. Both groups received a weekly visit from a physiotherapist
Outcomes	Measured before intervention and at 4 weeks and 8 weeks after end of intervention Primary: constipation score Secondary: Neurogenic Bowel Dysfunction score, MS impact score; Qualiveen questionnaire for bladder function, 7 day bowel diary
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	web‐based system
Allocation concealment (selection bias)	Low risk	web‐based system
Blinding (performance bias and detection bias) All outcomes	Low risk	research assistant analysed data and was blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	incomplete data on one control patient at 8 weeks

Medaer 1999.

Methods	Double‐blind placebo‐controlled pilot phase 2 study Intention‐to‐treat analysis: no
Participants	Setting and country: Belgium, site/s not specified Inclusion criteria: 11 patients with multiple sclerosis Exclusion criteria: not mentioned Age: not mentioned Sex: not mentioned Other characteristics: none stated
Interventions	4 week run‐in. Prucalopride 1 mg for 4 weeks then 2mg for 4 weeks before breakfast, versus placebo
Outcomes	Stool frequency, number of spontaneous bowel motions per week, median time to first bowel movement after first treatment, unproductive toilet trips, patient evaluation of constipation severity visual analogue score, side effects
Notes	Abstract only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	no mention
Allocation concealment (selection bias)	Unclear risk	no mention
Blinding (performance bias and detection bias) All outcomes	Unclear risk	states double‐blind but no detail
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	not mentioned

Mun 2011.

Methods	Randomised controlled double‐blind study Intention‐to‐treat analysis: not clear, 6 dropouts
Participants	Setting and country: one centre in Korea Inclusion criteria: 34 bedridden constipated older people following a stroke Exclusion criteria: none stated Age: treatment group 67.0 ± 6.0 years, controls 66.7 ± 5.4 years Sex: male 24, female 10 Other characteristics: none stated
Interventions	Oral carbonated water versus oral tap water for two weeks
Outcomes	Laxative use, frequency of defecation and constipation symptoms
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	random allocation
Blinding (performance bias and detection bias) All outcomes	Unclear risk	double‐blind
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	6 dropouts

Rosman 2008.

Methods	Randomised crossover study Intention‐to‐treat analysis: all 7 participants reported
Participants	Setting and country: USA, sites not specified Inclusion criteria: 7 spinal cord injured patients with documented defecation problems; mean duration of injury 15.9 years range 3‐27, 4 cervical, 3 thoracic Exclusion criteria: contraindications to neostigmine Age: mean age 47 years, range 30‐56 Sex: not mentioned Other characteristics: 5 Afro‐American, 2 Hispanic
Interventions	2mg neostigmine and 0.4 mg glycopyrrolate intramuscularly for 1 week, 1 week washout, 1 week placebo, in random order plus usual bowel care. Fleet enema administered after 60 minutes if no evacuation
Outcomes	Time from injection to completion of bowel care, time to first flatus, time to beginning of bowel movement, time to end of bowel movement, total time for bowel movement including clean up, side effects
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Low risk	order determined by sealed random assignment
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Low risk	outcomes for all 7 participants reported

Venn 1992.

Methods	Randomised double‐blind controlled trial Intention‐to‐treat analysis: no
Participants	Setting and country: Hospital Rehabilitation Unit, USA Inclusion criteria: 58 stroke patients entered the trial. 46 patients were analysed Exclusion criteria: not mentioned Age:mean age 72 years, range 36 to 90 Sex: not mentioned Other characteristics: all participants were actively rehabilitating, were given a diet gradually increasing in fibre and encouraged to maintain a high fluid intake
Interventions	Comparison of 4 suppository‐based bowel programmes during study period (1 month): (1) mandatory morning suppository (2) optional morning suppository (3) mandatory evening suppository (4) optional evening suppository
Outcomes	Number of subjects with effective bowel programme defined as a bowel movement within the scheduled time for 5 consecutive days. Efficiency rating determined by the number of days needed to achieve effectiveness. Score starts at 16 and dropped by 1 for each 2 additional days required to establish effectiveness
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	46/58 were analysed: not clear which group

Wang 2008.

Methods	Randomised controlled trial Intention‐to‐treat analysis: not mentioned
Participants	Setting and country: China, site/s not specified Inclusion criteria: 80 patients with stroke Exclusion criteria: Age: not mentioned Sex: not mentioned Other characteristics: none stated
Interventions	Abdominal electroacupuncture 30 minutes daily for 2 seven day courses (40 patients) versus 10mg oral cisapride three times a day for two 7 day courses (40 patients)
Outcomes	Cumulative scores of symptoms, total effective rate (definition unclear)
Notes	Abstract only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	random allocation
Allocation concealment (selection bias)	Unclear risk	method not described
Blinding (performance bias and detection bias) All outcomes	Unclear risk	not mentioned
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	not mentioned

Zangaglia 2007.

Methods	Randomised placebo‐controlled trial Intention to treat analysis: no
Participants	Setting and country: two Italian centres Inclusion criteria: 57 Parkinson's disease patients, with constipation on Rome ii criteria Exclusion criteria: pregnant or not using effective contraception, use of drugs affecting gastrointestinal motility, previous gastrointestinal surgery, other causes of constipation Age: mean age 71 years SD ± 6.5 years Sex: 34 male, 23 female Other characteristics: mean disease duration 6.4 years
Interventions	Laxative free baseline followed by 8 weeks isosmotic macrogol electrolyte solution (7.3g in water) versus placebo (flavoured maltodextrine) both dissolved in 250ml water twice daily. Dose could be varied between 1 and 3 sachets daily depending on stool form, patients were asked to keep diet and fluid intake unchanged
Outcomes	Predominant symptom, other symptoms, stool frequency (number of evacuations per week), straining (visual analogue scale), stool consistency (4 point scale), use of rectal laxative as a rescue, diary cards, unused sachets of medication
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	not mentioned
Allocation concealment (selection bias)	Unclear risk	not mentioned
Blinding (performance bias and detection bias) All outcomes	Low risk	participants were blinded
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	9 out of 29 withdrew from intervention group compared with 5 out of 28 from placebo

SCI: spinal cord injury

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Amir 1998	Non‐randomised trial Sequential prospective comparison of 4 bowel care regimens (bisacodyl suppository, glycerin suppository, mineral oil enema, docusate sodium mini‐enema) in 7 SCI subjects. Docusate mini‐enema significantly reduced colonic transit time. Docusate and mineral oil enemas both significantly shortened colonic transit time when compared to bisacodyl or glycerin suppositories. Bowel evacuation time was least for docusate mini‐enemas, which also scored best in subjective evaluation scoring for difficulty with evacuation
Badiali 1997	Non‐randomised trial Prospective comparison in 10 SCI subjects of a 4 weeks step‐by‐step protocol based on a high residue diet, a standardised water intake, and on the use of a sequential schedule of evacuating stimuli. When compared to basal period subjects showed a shortened total gastrointestinal transit time and a decreased need for laxatives
Badiali 1991	Intervention drug withdrawn from market
Barnes 1985	Randomised single blind therapy Non‐specific intervention on bowel function. No proper outcome assessment. Expanded Disability Status Scale (EDSS) is inaccurate, non‐quantitative, and does not allow differentiation between bladder and bowel symptoms. Chronic MS patients assigned to hyperbaric oxygen daily for 20 exposures showed no objective benefit and very little subjective benefit. Improvement was stated as an increase of a point on EDSS for bowel/bladder. This occurred in significantly more patients in the hyperbaric‐oxygen group (12 of 51) than in the control group (3 of 47). This benefit was not sustained at 1 year (Barnes 1987)
Binnie 1988	Non‐randomised trial Prospective comparison in 10 SCI subjects of cisapride versus baseline
Binnie 1991	Non‐randomised trial Prospective comparison of 10 SCI subjects implanted with an electronic Brindley stimulator, 7 non‐implanted SCI subjects and 10 healthy volunteers. Implanted subjects had a significant increase in frequency of defecation compared to the non‐implanted group. A comparison made before and after implantation would have been much more appropriate. This promising technique has not yet been evaluated in randomised controlled trials (Chia 1996; MacDonagh 1990; Varma 1986)
Bliss 2011	Non‐neurogenic population
Bond 2005	Only 4 out of 27 individuals randomised to the intervention had central neurological conditions and it was not possible to identify their contribution to the data
Carter 1988	Stroke patients without identified bowel dysfunction at recruitment and no bowel‐specific outcome measure, post hoc analysis
Chapman 1974	Non‐randomised trial with non‐neurological patients
de Both 1992	Intervention drug withdrawn from market
Dunn 1994	Non‐randomised trial Sequential prospective comparison of a bisacodyl mini‐enema (Therevac) with a bisacodyl suppository in 14 SCI inpatients. Ten patients completed the evaluation and experienced a decrease in mean rectal evacuation times with Therevac and a shortening of the time needed for bowel care
Fang 2003	Stroke patients without identified bowel dysfunction at baseline and no bowel specific outcome measures
Frisbie 1997	Non‐randomised trial Sequential prospective comparison of a polyethylene glycol bisacodyl‐based suppository (PGB) with a hydrogenated vegetable oil‐based bisacodyl suppository (HVB) in 19 SCI patients. Patients experienced a reduction of time for complete bowel evacuation with PGB versus HVB
Frost 1993	Non‐randomised trial Transcutaneous sacral electro‐stimulation applied in 7 SCI subjects significantly increased the number of colonic spikes waves. However, no change in the time required to initiate a bowel movement or to complete bowel emptying was noted
Geders 1995	Intervention drug withdrawn from the market
Heymen 2005	Non‐neurogenic population
Kim 1991	Continent and incontinent patients not reported separately. No primary bowel outcome measures reported
Latimer 1984	Only 2 out of 8 patients with faecal incontinence were neurological (SCI and MS)
Loening‐Baucke 1988	Efficacy of biofeedback for faecal incontinence in 12 children with myelomeningocele. Three of eight patients in the biofeedback group and three of the four given conventional treatment alone reported greater than or equal to 75% improvement in frequency of soiling 1 year later. Biofeedback did not improve anal squeeze and rectal sensation
Markwell 2006	Non‐neurogenic population
Marshall 1997	Randomised controlled trial in children Transcutaneous pre‐sacral electro‐stimulation in 50 children with spina bifida showed a trend in improvement of episodes of spontaneous normal defecation
Munchiando 1993	Non‐randomised trial Every‐other‐day digital stimulation (DS) was compared in 23 cerebrovascular accident patients with daily DS in 25 other patients. More patients performing daily DS established a regular bowel programme than those performing every‐other‐day DS, although establishing a regular bowel programme took longer in the daily DS group
Petajan 1996	Not a specific intervention for bowel function, with no result for bowel function alone (EDSS rates bladder and bowel together) Randomised trial. No outcome assessment. EDSS is inaccurate and non‐quantitative for bowel function. 21 MS patients randomly assigned to a training exercise and compared with 25 control patients non‐exercising. Exercise consisted of 15 weeks aerobic fitness training, including maximal aerobic capacity (VO2max) and isometric strength. The training resulted in significantly improved cardiovascular fitness and muscular strength for the exercise group. Exercising subjects experienced significant improvement in physical function as well as in well‐being (anger, depression, fatigue). The exercise intervention resulted in significant improvement in EDSS bladder and bowel function
Rajendran 1992	Intervention drug withdrawn from the market
Rao 2005	Non‐neurogenic population
Rao 2010	Non‐neurogenic population
Stiens 1995	Randomised comparison of a hydrogenated vegetable oil‐based bisacodyl (HVB) suppository with polyethylene glycol‐based bisacodyl suppositories (PGB) in a single SCI subject. The average total bowel care time was significantly shorter with PGB
Stiens 1998	Non‐randomised trial Open label crossover controlled trial comparison in 14 SCI inpatients of a hydrogenated vegetable oil‐based (HVB) and a polyethylene glycol‐based (PGB) bisacodyl suppository. PGB significantly shortened the total bowel care time
Sullivan 2006	Intervention drug withdrawn from the market
Tekeoglu 1998	Stroke patients without identified bowel problems at baseline and no bowel‐specific outcome measures
Van Winckel 2005	Children
Weinland 2010	Functional bowel disorders not neurogenic

Differences between protocol and review

Five studies related to the use of the prokinetic medications included in previous reviews have been removed. Cisapride has been withdrawn from sale worldwide since the conduct of these studies. Since the drug was licensed in 1988 there have been 386 reports of serious ventricular arrhythmias (125 fatal) associated with its use and 50 reports of sudden unexplained death (Ferriman 2002). Tegaserod was also withdrawn in 2007 due to concerns regarding adverse cardiovascular effects.

Contributions of authors

For this update (2013) all three review authors assessed the eligibility of the studies which resulted in the inclusion of a further 15 trials in the review. Five trials dealing with drugs that have been withdrawn from the market were excluded. The Protocol was also updated. All reviewers interpreted the results and contributed to the writing of the review.

Sources of support

Internal sources

• Chief Scientist Office, Scottish Executive Health Department, UK.

External sources

• The National Institute for Health Research, UK.

  The National Institute for Health Research (NIHR) is the largest single funder of the Cochrane Incontinence Group

• Department of Internal Medicine, University Hospital (CHUV), Lausanne, Switzerland.

• NHS National R&D Programme for People with Physical and Complex Disabilities, UK.

• Swiss Paraplegics Foundation, Basel, Switzerland.

• Action Reserach, UK.

Declarations of interest

CN has consulted for Coloplast and Shire

MC has consulted for Coloplast

New search for studies and content updated (no change to conclusions)