Sacral nerve stimulation for faecal incontinence and constipation in adults

Mohamed A Thaha; Amin A Abukar; Noel N Thin; Anthony Ramsanahie; Charles H Knowles

doi:10.1002/14651858.CD004464.pub3

. 2015 Aug 24;2015(8):CD004464. doi: 10.1002/14651858.CD004464.pub3

Sacral nerve stimulation for faecal incontinence and constipation in adults

Mohamed A Thaha ^1,^✉, Amin A Abukar ¹, Noel N Thin ¹, Anthony Ramsanahie ¹, Charles H Knowles ¹

Editor: Cochrane Incontinence Group

PMCID: PMC9208727 PMID: 26299888

Abstract

Background

Faecal incontinence (FI) and constipation are both socially‐embarrassing and physically‐disabling conditions that impair quality of life. For both, surgery may be required in a minority of people when more conservative measures fail. However, the invasiveness and irreversible nature of direct surgery on bowel and sphincter muscles, poor long‐term outcomes and well‐established compIications makes such procedures unappealing for these benign conditions. A less‐invasive surgical option to treat faecal incontinence and constipation is direct, low‐voltage stimulation of the sacral nerve roots, termed sacral nerve stimulation (SNS). SNS has become the first line surgical treatment for FI in people failing conservative therapies. Its value in the treatment of constipation is less clear.

Objectives

To assess the effects of sacral nerve stimulation using implanted electrodes for the treatment of faecal incontinence and constipation in adults.

Search methods

We searched the Cochrane Incontinence Group Specialised Register, which contains trials identified from the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, MEDLINE In‐Process, ClinicalTrials.gov, the World Health Organization (WHO) ICTRP and handsearched journals and conference proceedings (searched 5 February 2015), EMBASE (1 January 1947 to 2015 Week 5), and the reference lists of retrieved relevant articles.

Selection criteria

All randomised or quasi‐randomised trials assessing the effects of SNS for faecal incontinence or constipation in adults.

Data collection and analysis

Two review authors independently screened the search results, assessed the methodological quality of the included trials, and undertook data extraction.

Main results

Six crossover trials and two parallel group trials were included.

Six trials assessed the effects of SNS for FI. In the parallel group trial conducted by Tjandra, 53 participants with severe FI in the SNS group experienced fewer episodes of faecal incontinence compared to the control group who received optimal medical therapy (mean difference (MD) −5.20, 95% confidence interval (CI) −9.15 to −1.25 at 3 months; MD −6.30, 95% CI −10.34 to −2.26 at 12 months). Adverse events were reported in a proportion of participants: pain at implant site (6%), seroma (2%) and excessive tingling in the vaginal region (9%).

In the parallel group trial carried out by Thin, 15 participants with FI in the SNS group experienced fewer episodes of FI compared with the percutaneous tibial nerve stimulation (PTNS) group (MD −3.00, 95% CI −6.61 to 0.61 at 3 months; MD −3.20, 95% CI −7.14 to 0.74 at 12 months). Adverse events were reported in three participants: mild ipsilateral leg pain during temporary testing (n = 1); and stimulator‐site pain following insertion of neurostimulator (n = 2).

In the crossover trial by Leroi 7 of 34 recruited participants were excluded from the crossover due mainly to complications or immediate device failure. Twenty‐four of the remaining 27 participants while still blinded chose the period of stimulation they had preferred. Outcomes were reported separately for 19 participants who preferred the 'on' and five who preferred the 'off' period. For the group of 19, the median (range) episodes of faecal incontinence per week fell from 1.7 (0 to 9) during the 'off' period to 0.7 (0 to 5) during the 'on' period; for the group of five, however, the median (range) rose from 1.7 (0 to 11) during the 'off' period compared with 3.7 (0 to 11) during the 'on' period. Four of 27 participants experienced an adverse event resulting in removal of the stimulator.

In the crossover trial by Sørensen and colleagues, participants did not experience any FI episodes in either the one‐week ‘on’ or ‘off’ periods.

In the crossover trial by Vaizey, participants reported an average of six, and one, episodes of faecal incontinence per week during the 'off' and 'on' periods respectively in two participants with FI. Neither study reported adverse events.

In the crossover trial by Kahlke, 14 participants with FI experienced significantly lower episodes of FI per week during the stimulator 'on' (1 (SD, 1.7)) compared with the 'off' period (8.4 (SD, 8.7)). Adverse events reported include: haematoma formation (n = 3); misplacement of tined lead (1); and pain at stimulator site (n = 1).

Two trials assessed SNS for constipation. In the Kenefick trial, the two participants experienced an average of two bowel movements per week during the 'off' crossover period, compared with five during the 'on' period. Abdominal pain and bloating occurred 79% of the time during the 'off' period compared with 33% during the 'on' period. No adverse events occurred. In contrast, in the trial by Dinning with 59 participants, SNS did not improve frequency of bowel movements and 73 adverse events were reported, which included pain at site of the implanted pulse generator (32), wound infection (12), and urological (17) events.

Authors' conclusions

The limited evidence from the included trials suggests that SNS can improve continence in a proportion of patients with faecal incontinence. However, SNS did not improve symptoms in patients with constipation. In addition, adverse events occurred in some patients where these were reported. Rigorous high quality randomised trials are needed to allow the effects of SNS for these conditions to be assessed with more certainty.

Plain language summary

Sacral nerve stimulation for treating faecal incontinence and constipation in adults

Importance of the review/Background: Faecal incontinence occurs when a person passes stools without the usual control. Constipation is harder to define but generally describes a situation in which a person feels that their bowel opening is unsatisfactory (usually a combination of difficulty or infrequency of passing stools). Both conditions can severely affect people's quality of life. There are many non‐surgical treatments for both conditions but occasionally surgery is required when other options fail. Sacral nerve stimulation (SNS) is a relatively new treatment for these conditions. It involves implanting a battery‐powered stimulator unit in the buttock. This is connected to electrodes which rest on the nerves in the lower spine. The stimulator then continuously sends impulses to the nerves and muscles that control the bowel and anus. Initially, a temporary electrode lead is connected to a portable battery unit outside the body. If symptoms are improved enough, this is replaced by the implanted battery.

Main findings: This review evaluated the published evidence for the use of SNS for patients with faecal incontinence or constipation from six trials of SNS for faecal incontinence (219 participants) and two trials of SNS for constipation (61 participants). Two of the faecal incontinence trials had a 'parallel group design', which means that one group of participants received SNS and the other control group did not receive SNS throughout the trial. The remaining six trials had a 'crossover design', in which the participants experienced equal periods with stimulation 'off' then 'on', or vice versa. The level of stimulation was such that participants could not tell whether the system was 'on' or 'off'.

SNS for faecal incontinence: In the two 'parallel group' trials, 53 and 15 participants with faecal incontinence who were in the SNS group experienced fewer episodes of faecal incontinence compared to the control group at 3 and 12 months. In the first crossover trial, 24 participants who completed the trial chose the period of stimulation they had preferred while still unaware whether this was 'on' or 'off'. Nineteen participants who preferred the 'on' period experienced 59% fewer episodes of FI per week during the 'on' period, and 5 participants who preferred the 'off' period experienced 118% more episodes of FI per week. In the second crossover trial, the participants did not experience episodes of FI during either the 'on' or the 'off' periods. In the third trial, participants experienced 83% fewer episodes of faecal incontinence during the 'on' compared with the 'off' period. In the fourth crossover trial participants experienced 88% fewer episodes of faecal incontinence during the 'on' period compared with the 'off' period.

SNS for faecal incontinence—adverse effects: Not all trials reported adverse effects after SNS. The two 'parallel group' trials reported only minor complications, in 10% of SNS participants in the first study, and in 3 participants in the second study. In the first crossover study 7 out of 34 participants were excluded from crossover due mainly to complications. Four out of 27 participants with an implanted system in this study experienced a problem that led to the device being removed. The participants in the fourth crossover trial experienced some complications with the SNS implanted electrode such as pain (one person), misplacement of the tined lead (one person) and haematoma (swelling containing blood) (three people).

SNS for constipation: In one trial assessing SNS for constipation, two participants reported an increase of 150% in the frequency of passing stools per week, and time with abdominal pain and swelling went down from 79% during the 'off' period to 33% during the 'on' period. However, in the much larger second trial assessing SNS for constipation, in 59 participants SNS did not improve frequency of bowel movements.

Limitations of the review: The limited evidence suggests that SNS can improve continence in some people with faecal incontinence. SNS did not improve symptoms in patients with constipation. Larger, good‐quality trials are needed to provide more reliable evidence on the effectiveness of SNS for these two conditions.

Background

Description of the condition

The act of defaecation is dependent on the co‐ordinated functions of the colon, rectum and anus. Considering the complexity of neuromuscular (sensory and motor) functions required to achieve planned, conscious, and effective defaecation (Scott 2011), it is no surprise that disturbances to perceived ‘normal’ function occur commonly at all stages of life. Clinically, such problems principally encompass presentations with faecal incontinence or constipation. Although usually described separately (as in this review), it should be noted that considerable overlap between these symptoms occurs at all ages of presentation (Burgell 2012; Rao 2004a).

Faecal incontinence

Faecal incontinence (FI) is a socially‐embarrassing and physically‐disabling condition. It may be defined as the uncontrolled loss of faeces (liquid or solid) from the bowel. It may occur passively (without the person affected being aware of passing faeces); or be preceded by urgency (a sense of an urgent need to defaecate); or both, commonly termed 'mixed faecal incontinence'. A further term—‘post‐defaecatory seepage or leakage’—is also sometimes used to denote passive loss of small amounts of faeces retained after incomplete evacuation (Rao 2004b).

Faecal incontinence is a common problem, although the true prevalence is hard to determine due to under‐reporting (Johanson 1996; NICE 2007). US population surveys suggest the prevalence of FI ranges anywhere from 2% to 17% affecting both men and women (Nelson 2004; Peery 2012; Whitehead 2009). In the UK, it is estimated that up to 10% of adults experience involuntary loss of solid or liquid stool causing physical, psychological and social disability; 0.5% to 1% of adults experience regular faecal incontinence that severely affects their quality of life (NICE 2007; Norton 2007). Although awareness of the health burden from faecal incontinence has increased in recent years, it still largely remains a taboo problem (Johanson 1996). Considering the higher prevalence with advancing age (Brown 2010), FI may be an even greater problem in the future. The prevalence of FI approaches 50% in nursing homes (Nelson 1998).

Faecal incontinence may result from:

damage to the anal sphincter mechanism (either from direct trauma or damage to its nerve supply);
age‐related degeneration of the sphincter;
spinal injury;
other neurological causes; or
non‐sphincter causes (e.g. diarrhoea, dementia) (Lunniss 2004).

The well‐documented association between FI and birth injuries has led to the belief that pudendal nerve injury or direct damage to the anal sphincter complex is the main pathophysiology (Sultan 1993; Sultan 1997a; Sultan 1997b). In truth, most FI is caused by a complex interplay of pathophysiological factors including aberrant anorectal sensation and colorectal motility, as well as structural changes or damage to the pelvic floor and sphincter complex. Indeed, women who sustain obstetric trauma often do not present with symptoms of FI until up to 30 years postpartum (Lunniss 2004; Rao 2004a).

With such aetiological and pathophysiological heterogeneity, FI remains a challenging condition to directly address by any single treatment. Conservative measures include:

dietary modification;
bowel retraining;
anti‐diarrhoeal medication (Ehrenpreis 2007; Omar 2013);
biofeedback (Enck 2009; Norton 2003; Norton 2004); and
supportive devices such as absorbent padding or plugs.

However, such measures have, at best, a lasting success in only approximately 50% of patients (Otto 2010); thus surgical treatment is often considered. The latter may include:

bulking agents (Graf 2011);
direct anal sphincter repair (Malouf 2000);
artificial bowel sphincter (Altomare 2004); and
dynamic graciloplasty (Baeten 1991).

All of these operations have well‐established complications and high long‐term failure rates (Altomare 2004; Malouf 2000). While several newer treatments, such as autologous muscle cell therapy, magnetic sphincter augmentation and sphincter bulking agents, are at an early clinical evaluation stage (Carr 2013; Lehur 2010; Ratto 2011), many patients may still resort to a stoma (usually colostomy) for significant intractable symptoms.

Constipation

Constipation is common in adults and children and up to 20% of the population report this symptom depending on definitions used (2% to 28% adults; 0.7% to 30% children) (Sonnenberg 1989; Stewart 1999; van den Berg 2006), with a higher prevalence in women (McCrea 2009; Mugie 2011; Sonnenberg 1989), and the elderly (Gallegos‐Orozco 2012; Norton 2006). Chronic constipation (CC), usually defined as more than six months of symptoms, is less common (Probert 1995), but results in half a million UK GP consultations per annum. A proportion of the population suffer symptoms which are both chronic and more disabling (probably about 1% to 2% of the population) (Cook 2009). Such patients, who are most often female (Knowles 2003), are usually referred to secondary care with many progressing to tertiary specialist investigation. Patient dissatisfaction is high in this group: nearly 80% feel that laxative therapy is unsatisfactory (Wald 2008), and the effect of symptoms on measured QOL is significant (Irvine 2002). Chronic constipation consumes significant healthcare resources: in the US in 2012, a primary complaint of constipation was responsible for 3.2 million physician visits (Peery 2012), resulting in (direct and indirect) costs of $1.7 billion. In the UK, it is estimated that 10% of district nursing time is spent on constipation (Poulton 1999), and the annual spend on laxatives exceeds £80 million, with £17.4 million prescribed in 2012 (HSCIC 2013).

Management of chronic constipation is a major problem due to its high prevalence and lack of widespread specialist expertise. In general, a step‐wise approach is undertaken, with first‐line conservative treatment such as lifestyle advice and laxatives (NHS 2012), followed by nurse‐led bowel retraining programs, often including focused biofeedback (Woodward 2014), and psychosocial support. Although these treatments may improve symptoms in more than half of patients, they are very poorly standardised and are not universally successful. While some hope has been offered by a range of new prokinetic and pro‐secretory drugs (Camilleri 2008; Johanson 2008), patients with intractable symptoms and impaired QOL may be offered a range of costly, irreversible surgical interventions with unpredictable results (Knowles 1999; Knowles 2009), sometimes resulting in major adverse events or a permanent stoma.

Description of the intervention

Neuromodulation is one of the fastest growing areas of medicine and may be defined as a technology that impacts upon neural interfaces to produce benefit. The concept of electrical stimulation of the pelvic floor may be traced back to the early 19^th century with several methods of direct neuromuscular stimulation developed subsequently (Hopkinson 1966). In the last 20 years, a group of treatments have been developed that employ chronic, low‐voltage electrical stimulation to recruit residual function of pelvic organs by direct or indirect stimulation of the sacral spinal nerves (Matzel 1995; Matzel 1990). Such developments, which in general have evolved from the urology field (Brindley 1974; Tanagho 1989), have the potential to bridge the gap between conservative treatments and potentially hazardous direct surgery to the bowel or anal sphincter. An attraction of these treatments is that, unlike direct surgery to the sphincter or colon, they have the potential to modify all aspects of the co‐ordinated neuromuscular functions required for defaecation (Carrington 2011), and on this basis may benefit both FI and chronic constipation. Further, neuromodulation has the intrinsic qualities of dose variation and reversibility which are rarely possible with other surgical therapies. The most established of these treatments is Sacral Nerve Stimulation (SNS).

SNS involves direct, chronic, low‐voltage electrical stimulation of the sacral nerve roots by the siting of an electrode via a sacral foramen (S3 is the optimal site for most patients). Whilst there has been an evolution of systems over time (MacDonagh 1990; Matzel 1990), in its most common current form SNS utilises a percutaneously‐sited, commercially‐manufactured quadripolar electrode lead system connected to an implanted pulse generator (InterStim® Therapy, Medtronic, Inc., Minneapolis, US) buried in the subcutaneous fat of the buttock. The Medtronic system also allows for a temporary percutaneous nerve evaluation (PNE) phase which allows the patient to trial the impact of sacral nerve stimulation on their lifestyle and test the feasibility of success before the more expensive permanent stimulator is implanted.

SNS techniques are well described in the literature, with agreement that this can be carried out effectively and safely under both general anaesthetic, and local anaesthetic with or without sedation (Mitchell 2011); with minimal reported morbidity (overall complications rates of 5% to 26% (Tjandra 2004)); and no reported mortality (Wong 2011).

How the intervention might work

Understanding of the mechanism of action of SNS is universally acknowledged to be an important knowledge gap limiting potential for patient selection, technology and procedural optimisation that could focus therapy, improve outcomes and reduce healthcare costs (Carrington 2014). Traditional understanding of the pathophysiology and surgical management of FI held that sphincter ‘barrier’ had primacy. It is now clear that whilst sphincter disruption (resulting from, for example, obstetric injuries) is still relevant to the development of FI, it is only one factor in complex defaecatory dysfunction that involves alteration in unconscious anorectal and pelvic reflexes and conscious modulation by the central nervous system.

The importance of sensory dysfunction on both urinary and bowel control is being increasingly appreciated and there is strong evolving evidence (in man (Knowles 2012); and experimental animals (Evers 2014; Griffin 2011)) that the mechanism of action of SNS results primarily from modulation of afferent nerve activity rather than motor effects on the anal sphincter itself (Carrington 2014; Duelund‐Jakobsen 2013; Patton 2013). The effects of SNS on colonic motor activity have also been studied. Data suggest that SNS, but not sham stimulation, increased the frequency of retrograde propagated sequences throughout the colon in participants with FI (Patton 2013) i.e. effects anticipated to delay colonic transit. In contrast SNS also increased colonic propagating sequences in participants with slow‐transit constipation (Dinning 2012). This disparity of effect may reflect baseline differences of colonic motor activity, however further study is required to explain these seemingly counter‐intuitive findings.

Why it is important to do this review

SNS is now considered the first‐line surgical intervention for people with FI for whom conservative treatments have failed (NICE 2007). Over 100 published case series (of over 2000 participants), including some quite large prospective studies (Melenhorst 2007; Wexner 2010), attest to the general success of this approach, and SNS has regulatory approval in most major health systems (NHS 2013; NICE 2007). Recent systematic review data show favourable mid‐ and long‐term positive outcomes for SNS of approximately 80% based on a greater than 50% reduction in FI episodes, although this figure is reduced to approximately 60% when results are reanalysed using all available participants who start therapy as the denominator (comparable with intention‐to‐treat principles) (Thin 2013).

In comparison with FI, the role of SNS in treating chronic constipation is less well established. However, it carries the same attractions: namely the avoidance of potentially hazardous direct surgery to the colon or pelvic floor. Unfortunately, favourable results of a European experience of SNS on 62 participants (63% successful) with chronic constipation, (of mixed or specific pathophysiology (Kamm 2010 and Knowles 2012 respectively)) have not been replicated by others (Holzer 2008; Maeda 2010; Vitton 2009). On this basis SNS has not yet, in general, been commissioned as a treatment for constipation in most health systems.

Despite the implicit advantages and reported success of SNS over the last 20 years (especially for FI), one major drawback remains: the initial costs of SNS are still considerable both in terms of direct equipment cost and indirect hospital admission costs (Dudding 2008). On this basis, health practitioners need to be reassured of the efficacy of SNS therapy. This systematic review updates the available evidence from randomised controlled trials.

Objectives

To assess the effects of sacral nerve stimulation using implanted electrodes for the treatment of faecal incontinence and constipation in adults.

The following comparisons were made:

Sacral nerve stimulation (implanted) versus control or sham stimulation
Sacral nerve stimulation (implanted) versus another active treatment

Methods

Criteria for considering studies for this review

Types of studies

We included randomised, quasi‐randomised and crossover trials. Crossover trials have been identified with the suffix '#'.

Types of participants

Adults with faecal incontinence or constipation, including functional, structural and neurological causes.

Types of interventions

One arm of the trial had to use a surgically implanted device to provide sacral nerve stimulation.

Comparators considered included:

mock, sham or placebo treatment; and
any alternative active intervention considered appropriate by the trialists, such as dynamic graciloplasty, artificial bowel sphincter implants, stoma formation, absorbent pads, anal plugs, and physical or behavioural therapies.

Trials in which both arms used active implanted sacral nerve stimulation were not included in this review. Newer treatment modalities including magnetic and transcutaneous stimulation were considered.

Types of outcome measures

Primary outcomes

Faecal incontinence

The primary outcomes were: (a) cure or improvement of incontinence; and (b) quality of life. Outcomes were considered in the following categories:

(1) Faecal incontinence (number cured or improved, episodes of faecal incontinence, urgency, ability to defer defaecation, use of pads, use of anal plugs, incontinence score, and need for further treatment such as medication or surgery).

(2) Quality of life (generic and condition‐specific).

Constipation

The primary outcomes were: (a) cure or improvement of constipation; and (b) quality of life. Outcomes were considered in the following categories:

(1) Constipation (number cured or improved, frequency of bowel movements, abdominal pain/bloating, constipation score, and need for further treatment such as medication or surgery).

(2) Quality of life (generic and condition‐specific).

Secondary outcomes

Faecal incontinence

Outcomes were considered in the following categories:

(1) Surrogate measures, such as anorectal manometry (resting pressure, maximum squeeze pressure, rectal sensory threshold to balloon distention, sensation of urgency to balloon distention, and maximum tolerated rectal volume to balloon distention).

(2) Adverse effects (infection or pain or both at the implantation site; displacement of the electrodes; technical failure requiring removal or change in urinary function, or both).

(3) Health economics (costs, resource implications and cost‐effectiveness or cost utility evaluation).

(4) Other outcomes (other outcome measures quoted by trial authors and judged to be important by the authors of this review).

Constipation

Outcomes were considered in the following categories:

(2) Adverse effects (infection or pain or both at the implantation site; displacement of the electrodes; technical failure requiring removal or change in urinary function, or both).

(3) Health economics (costs, resource implications, and cost‐effectiveness or cost utility evaluation).

(4) Other outcomes (other outcome measures quoted by trial authors and judged to be important by the authors of this review).

Search methods for identification of studies

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

Electronic searches

This review drew on the search strategy developed for the Incontinence Group as a whole. Relevant trials were identified from the Cochrane Incontinence Group Specialised Trials Register. The methods used to derive this, including the search strategy, are 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, ClinicalTrials.gov, WHO ICTRP and hand searching of journals and conference proceedings.  The Incontinence Group Specialised Register was searched using the Group's own keyword system; the search terms used are given in Appendix 1.

Date of the most recent search of the register for this review: 5 February 2015.

The majority of the trials in the Incontinence Group's Specialised Register are also contained in CENTRAL.

Also Embase (1 January 1947 to 2015 Week 5) was searched on 5 February 2015 using the search strategy given in Appendix 2. Only the years 2010 to 2015 (inclusive) were searched as these years were not covered by the Cochrane Collaboration's centralised search of Embase for CENTRAL at this time.

Searching other resources

All reference lists of identified trials were searched.

Data collection and analysis

Analyses of crossover trials were based on data available from included relevant trials. Meta‐analysis could not be performed because of variation in reported outcomes.

Selection of studies

Two review authors (AA, MAT) independently evaluated reports of all potentially eligible studies for inclusion. Any disagreements were resolved by discussion or arbitration by the review group at consensus meetings.

Data extraction and management

Two review authors (AA, MAT) independently undertook data extraction for the five new trials that were added (Dinning 2015#; Kahlke 2015#; Sørensen 2010#; Thin 2015; Tjandra 2008); and checked and updated the information from the previously included trials, in particular 'Risk of bias' assessment. We planned to seek clarification from the trialists where data were collected but not reported, or reported in a form unsuitable for the review. Included data were processed as described in the Cochrane Handbook for Systematic Reviews of Interventions (http://www.cochrane.org/resources/handbook/hbook.htm) (Higgins 2011). Any disagreements were resolved by discussion or arbitration by the review authors. Data from crossover were entered into Other Data Tables.

Assessment of risk of bias in included studies

Two review authors (AA, MAT) independently assessed the methodological quality of all the included trials using the Cochrane Incontinence Group's assessment criteria. Studies were not excluded from the review on the basis of methodological quality. Any disagreements were resolved by discussion or arbitration by all the review authors.

Measures of treatment effect

We planned to combine data, if this was considered appropriate, by applying fixed‐effect methods, with relative risk used for dichotomous outcomes and weighted mean difference used for continuous outcomes.

Dealing with missing data

We sought clarification on missing data from the trialists and, when available, they were included in the final abstraction and analysis.

Assessment of heterogeneity

If visual inspection or the Chi² test had suggested heterogeneity at the 10% level then we planned to explore the reasons for this. If the level of heterogeneity was serious enough to affect the validity of the results then applying a random‐effects method would be considered. Alternatively, if it was considered inappropriate to combine data from the included trials we planned instead to present a narrative synthesis of the results of the primary studies. In the event it was not possible to find any data suitable for meta‐analyses, and all the data are presented in 'Other Data' tables.

Assessment of reporting biases

All included trials were formally assessed for methodological quality, including for selective reporting using appropriate tools in Cochrane's Review Manager (RevMan) software. The methodological quality of included trials are reported in 'Risk of bias' figure and summary.

Data synthesis

We planned to combine data by applying fixed‐effect methods when appropriate.

Subgroup analysis and investigation of heterogeneity

If the evidence had allowed, we planned to undertake sub‐group analysis on the following categories of participants:  (1) participants with faecal urgency;  (2) participants with structural defects of the anal sphincter;  (3) participants with partial spinal cord injury; and  (4) participants with central neurological disease.

Sensitivity analysis

We planned to use sensitivity analysis to test the effects of included studies of poor methodological quality, if appropriate.

Results

Description of studies

Please see 'Characteristics of included studies' and 'Characteristics of excluded studies' tables for further details.

Results of the search

A total of 197 records, identified by the literature search, were screened and 16 full‐text articles of potentially eligible studies were obtained. There were 10 reports of 8 included studies and 3 reports of 3 excluded studies. Additionally there were three reports of three ongoing studies, details of which can be found in the Characteristics of ongoing studies table. Figure 1 illustrates the process of the literature search and selection of studies for the update of this review.

Included studies

Eight potentially relevant trials identified by the search strategy were included (Dinning 2015#; Kahlke 2015#; Kenefick 2002#; Leroi 2005#; Sørensen 2010#; Thin 2015; Tjandra 2008; Vaizey 2000#).

Six trials assessed the effects of sacral nerve stimulation (SNS) for faecal incontinence (Kahlke 2015#; Leroi 2005#; Sørensen 2010#; Thin 2015; Tjandra 2008; Vaizey 2000#).

Two trials assessed the effects of SNS for constipation (Dinning 2015#; Kenefick 2002#).

Design

Six trials had a crossover design (Dinning 2015#; Kahlke 2015#; Kenefick 2002#; Leroi 2005#; Sørensen 2010#; Vaizey 2000#), while two had a parallel‐group design (Thin 2015; Tjandra 2008).

In the study carried out by Thin 2015 patients were allocated to receive either SNS or percutaneous tibial nerve stimulation (PTNS). In the SNS group participants underwent a trial with temporary SNS for two weeks; those who exhibited a good response (at least 50% reduction in FI) had permanent SNS implantation. In the trial conducted by Tjandra 2008 all participants attended multidisciplinary pelvic floor clinic and were randomised to sacral nerve stimulation (SNS group) or optimal medical therapy (control group). In the SNS group all participants were tested for a minimum of seven days and participants who exhibited a good response underwent permanent implantation with a quadripolar electrode for the 12 months of follow up. The electrode combination with best perception of muscle contraction of perineum and anal sphincter with least voltage was selected for permanent stimulation.

Moreover, in the trials conducted by Kenefick 2002# and Vaizey 2000#, participants underwent two two‐week intervals with subsensory stimulation either 'on' or 'off'. There was no interval between the treatment periods.

Similarly, in the trial conducted by Sørensen 2010#, participants underwent two one‐week intervals with stimulation either ‘on’ or ‘off’. There was no interval between the treatment periods.

In the trial carried out by Dinning 2015#, participants underwent three weeks of temporary peripheral nerve stimulation, then all patients underwent permanent SNS electrode implantation. Participants were then randomised to receive subsensory stimulation or sham for three weeks; and after a two‐week washout period the group that received sham stimulation now received subsensory stimulation for three weeks and vice versa. Then, following a further two‐week washout period, participants were re‐randomised to receive either sham or suprasensory stimulation for three weeks. Participants then underwent a further two‐week washout period, after which participants who received sham stimulation now underwent suprasensory stimulation for three weeks and vice versa.

In the trial conducted by Kahlke 2015#, after permanent implantation 31 patients had a minimum of 3 months with the stimulator turned on. Then participants were invited to take part in the crossover study. Sixteen participants were enrolled, who were subsequently randomised in a crossover design to 'on' or 'off' stimulation, each period lasting three weeks. At the end of each three‐week period the stimulator was programmed to the opposite mode ('off' or 'on'). While still blinded at the end of the crossover period participants selected the period of stimulation they had preferred ('off' or 'on'). The mode of stimulation corresponding to the preferred period ('off' or 'on') was continued for a further three months.

In the trial conducted by Leroi 2005#, after definitive 'permanent' implantation each participants had a one‐ to three‐month phase when the stimulator was turned 'on' (postimplantation period), to optimise the effectiveness of stimulation by determining the most effective parameters of stimulation (choice of stimulation electrodes and intensity of stimulation). At the end of the postimplantation period, participants were randomised in a crossover design to 'on' or 'off' stimulation for a two‐month period. At the end of the first month, the neurostimulator was programmed to the opposite mode ('off' or 'on'), and monitoring continued for a second month. There was no interval between the two treatment periods. At the end of the crossover period, while still blinded, the participants chose the period of stimulation ('on' or 'off') they had preferred. The mode of stimulation corresponding to the selected period ('on' or 'off') was then continued for a three‐month follow‐up period. If the participant could not choose between one of the two periods, the stimulator was turned 'on'.

Sample sizes

In total the trials enrolled 280 people, of whom 177 received a definitive implant. The trials included:

Faecal incontinence:

Vaizey 2000# included two participants;
Kahlke 2015# enrolled 16 participants, all received a permanent SNS implant;
Leroi 2005# enrolled 34 participants, of whom 27 went on to receive a permanent implant;
Sørensen 2010# enrolled 7 participants;
Thin 2015 enrolled 40 participants, of whom 16 received a permanent SNS implant;
Tjandra 2008 enrolled 120 participants, of whom 53 received a permanent implant.

Constipation

Dinning 2015# enrolled 59 participants of whom 55 had permanent implant; and
Kenefick 2002# included two participants;

Setting

Two trials were carried out in the same hospital centre (St. Mark's Hospital, Harrow); neither reported the time period during which the trial was undertaken (Kenefick 2002#; Vaizey 2000#). The third trial was carried out in a single centre located in Germany (Kahlke 2015#). The fourth trial was a multicentre trial (number of centres not stated), set in France, which took place during the period February 2000 to February 2003 (Leroi 2005#). The fifth trial was a single‐centre trial set in Denmark (Sørensen 2010#). The sixth trial was a two‐centre trial set in two London hospitals (Royal London Hospital and University College London Hospital, UK) (Thin 2015). The seventh trial was a single‐centre trial set in Australia (Tjandra 2008). The eighth trial was a two‐centre trial conducted in Australia (Dinning 2015#).

Participants

Of 280 participants enrolled, 246 (87.9%) were women.

Compliance and treatment received

In the trial carried out by Thin and colleagues, 8 of the 23 participants in the SNS group did not receive permanent implantation (Thin 2015). The reasons cited include: intercurrent illness requiring urgent medical management (n = 1); symptoms resolved (n = 1); declined further participation (n = 2); failed temporary SNS (n = 3); withdrawn from treatment for urgent investigation of a new pararectal cyst (n = 1). In the trial conducted by Tjandra and colleagues 7 of the 60 participants in the SNS arm did not receive permanent implantation (Tjandra 2008). The reasons were: cannulation of foramen with electrode was not achieved due to previous back surgery, which required bone graft from the sacral area (n = 1); insufficient therapeutic response with peripheral nerve evaluation (n = 5); and concern participants might require magnetic resonance imaging of brain post excision of meningioma (n = 1).

In the trial carried out by Kahlke 2015#, out of the 16 participants enrolled, 2 discontinued; the reasons were (i) participant turned stimulator 'on' with handheld device because she had become completely incontinent (n = 1); (ii) participant fell on lower back leading to breakage of permanent lead (n = 1).

In the trial conducted by Leroi 2005# 10 of 34 participants prematurely discontinued the trial. The reasons for discontinuation were: device‐related adverse events (n = 4); protocol violation (n = 3); insufficient therapeutic response (n = 1); no return to follow‐up (n = 1); and adverse event (stroke) not related to SNS (n = 1).

In the trial conducted by Dinning 2015# four participants did not receive permanent SNS implant due to depression (n = 1); too far to travel (n = 2); and could not commit (n = 1). Of those who received permanent implant 2 participants discontinued the study due to pregnancy (n = 1); and severe wound infection (n = 1).

Age

The age of both participants in the trial conducted by Kenefick 2002# was 36 years, while in the trial conducted by Vaizey 2000# the participants were aged 65 and 61 years. The mean age of the participants in the trial carried out by Kahlke 2015# was 55.5 years (standard deviation (SD) = 11.8). The median age of the participants in the trial conducted by Leroi 2005# was 57 years (range 33 to 73 years). In the trial conducted by Sørensen 2010# the median age of participants was 67 years (range 60 to 87 years). In the trial conducted by Thin 2015 the mean age of participants was 59 years (SD = 13 in the SNS group and 11 in the PTNS group). In the trial conducted by Tjandra 2008 the median age of the participants was 63.9 years (SD = 13.2) in the SNS arm and 63 years (SD = 12.1) in the control group. In the trial conducted by Dinning 2015# the median age was 42 years (range 19 to 74 years).

Cause/type of faecal incontinence or constipation

The causes of faecal incontinence in the two participants in Vaizey's trial were, respectively, scleroderma and idiopathic degeneration of the internal anal sphincter (Vaizey 2000#). The cause of faecal incontinence cited in the trial conducted by Kahlke 2015# were idiopathic (n = 8), anorectal surgery (n = 4), and neurogenic (n = 4). In the trial conducted by Leroi 2005# the majority of participants (22 of 34, 65%) suffered from urge incontinence, while four suffered from passive incontinence and eight from mixed incontinence. The causes of incontinence were reported as idiopathic (n = 18); pudendal neuropathy (n = 14); postoperative internal anal sphincter fragmentation (n = 1); and primary internal anal sphincter degeneration (n = 1) (Leroi 2005#). In the trial conducted by Thin 2015 participants suffered from urge and passive incontinence. In the trial conducted by Tjandra 2008 all participants had both passive and urge incontinence. The participants in the trial conducted by Kenefick 2002# suffered from severe resistant idiopathic constipation that had failed maximal conservative treatment. The participants in the trial conducted by Dinning 2015# suffered severe refractory idiopathic slow transit constipation. The cause of faecal incontinence in the trial conducted by Sørensen 2010# was not stated. (See 'Characteristics of included studies' table for details of the inclusion/exclusion criteria applied by each trial).

Interventions

The intervention in all eight trials consisted of SNS via a permanently‐implanted stimulator (Medtronic InterStim®).

In the trials conducted by Kenefick 2002#, Sørensen 2010#, and Vaizey 2000#, the participants had received permanent implants respectively 12 months, 12 months and 9 months previously.

Kenefick 2002# reported that prior to permanent implantation, both participants had undergone a successful three‐week trial of temporary percutaneous stimulation; whereas Thin 2015, Tjandra 2008, and Leroi 2005# reported that before permanent implantation, participants underwent temporary percutaneous stimulation for, on average, 14 days, 10 days and between 8 and 15 days respectively to assess their response to treatment.
Similarly, Kahlke 2015# reported that before permanent implantation, participants underwent temporary percutaneous stimulation for a period between 14 and 20 days to assess their response to treatment.
In the trial conducted by Dinning 2015#, following 3 weeks of temporary peripheral nerve evaluation all participants underwent permanent SNS electrode implantation.
In the trial conducted by Thin 2015, participants underwent a trial of temporary SNS and participants with a good response (at least 50% reduction in episodes of FI) progressed to permanent stimulation.
Also, in the trial conducted by Leroi 2005#, the decision to progress from temporary to permanent stimulation was made on the basis of at least a 50% reduction in the number of episodes of incontinence per week, or a 50% reduction in the number of faecal urgency episodes per week, or both.
Similarly, in the trial conducted by Tjandra 2008 the decision to progress to permanent stimulation was based on a good response, defined as a 50% or greater reduction in faecal incontinence episodes per week or a 50% or greater reduction in number of days with faecal incontinence per week.

In the trial conducted by Kahlke 2015# and Leroi 2005# the permanent pulse generator was placed in the upper outer part of the buttocks. Similarly, in the trial conducted by Dinning 2015# and Tjandra 2008 the permanent pulse generator was placed in the gluteal area. In the trial conducted by Vaizey 2000# it was placed in the abdominal wall. Kenefick 2002# and Sørensen 2010# did not report the location of the pulse generator.

Duration of treatment and follow up

The length of time between the participants receiving a permanently‐implanted stimulator and the start of the crossover period varied from 3 weeks (Dinning 2015#), around 3 months (Kahlke 2015#; Leroi 2005#), through 9 months (Vaizey 2000#), to 12 months (Kenefick 2002#). In the trial carried out by Thin 2015 the follow‐up time for the SNS and PTNS groups was six months. In the trial conducted by Tjandra 2008, the follow up time for the SNS group and the control group was 12 months.

In two trials the crossover period consisted of two two‐week periods when the stimulator was either 'on' or 'off', with no follow‐up (Kenefick 2002#; Vaizey 2000#). Similarly, in the third trial the crossover period consisted of two one‐week periods when the stimulator was either 'on' or 'off' with no follow‐up (Sørensen 2010#); while in the fourth, larger, trial the crossover period consisted of two one‐month periods when the stimulator was 'on' or 'off', with a three‐month follow‐up (Leroi 2005#). In the fifth trial (Kahlke 2015#), the crossover period consisted of two three‐week periods when the stimulator was 'off' or 'on', with a three‐month follow‐up. In the sixth trial (Dinning 2015#), the crossover period consisted, in the suprasensory phase, of two three‐week periods where the stimulator was on (suprasensory) or off (sham); long‐term follow‐up is still ongoing.

Outcomes

The trials reported a variety of outcome measures.

Vaizey 2000# reported that participants used a diary to record episodes of faecal incontinence for liquid or solid stool; anal manometry was undertaken; and health status was assessed both before permanent implantation and again at the beginning of the trial, using an SF‐36 questionnaire. In terms of anal manometry, a stationary pull‐through method was used with an eight‐channel perfused system, and the squeeze pressure was measured as the incremental rise. Rectal sensation was tested using balloon distension with air (Vaizey 2000#).

Leroi 2005# reported that participants also used a diary, recording episodes of faecal incontinence, faecal urgency, delay in postponing defaecation, and bowel movements. Severity of incontinence was graded by the Cleveland Clinic Incontinence Scoring System, while quality of life was assessed with the French version of the American Society of Colon and Rectal Surgeons' (ASCRS) Fecal Incontinence Quality of Life Scale (FIQL). Anal manometry was also undertaken (Leroi 2005#). Measurements of maximum resting pressure and maximum squeeze pressure were recorded at baseline, at the end of each crossover period, and at follow‐up; while rectal sensation to balloon distension with air was recorded at baseline and at the end of the follow‐up period (Leroi 2005#).

Kahlke 2015# reported outcomes that included frequency of bowel movements, frequency of faecal incontinence, and the Cleveland Clinic Incontinence Score (CCIS).

In the trial conducted by Dinning 2015# participants used a bowel diary (stool frequency per week; bowel motion per week; feeling of complete evacuation per week; straining per week; laxative‐free days per week; soft stool per week; normal stool per week; hard stool per week; pain score weekly average; bloating score weekly average; weekly global satisfaction score; weekly bothersome score). Quality of life was documented using the SF‐36 questionnaire.

Also, in the trial conducted by Thin 2015 participants utilised a bowel diary to document the number and type of incontinence episodes per week. Additionally, symptom severity was assessed through the use of Cleveland Clinic Incontinence Score and quality of life was measured through use of Faecal Incontinence Quality of Life Scale (FIQL), Short Form 36 (SF‐36) and EQ‐5D.

Similarly, in the trial conducted by Tjandra 2008 participants used a bowel diary to record the number of incontinent episodes per week; days with incontinence per week; days with staining per week; and days with pads per week. Severity was scored by the Wexner incontinence score (Cleveland Clinic Incontinence Scoring System); and quality of life was evaluated by ASCRS's FIQL, and the standard Short Form‐12 health survey quality of life questionnaire (SF‐12) (Tjandra 2008). Measurements of anal manometry (resting and squeeze pressure) were undertaken at baseline for both SNS and control group, while further anal manometry measurements were only carried out for the SNS group during peripheral nerve evaluation and at 3‐, 6‐ and 12‐month follow‐up (Tjandra 2008).

In the trial conducted by Sørensen 2010#, outcomes were reported using bowel diaries and anal manometry.

In the trial conducted by Kenefick 2002#, outcomes included diary‐recorded bowel movements; the Wexner Constipation Score; a symptom analogue score; anal manometry; and quality of life. In terms of anal manometry, resting pressure was the maximal anal resting pressure using a stationary pull‐through technique, and squeeze pressure was the maximal increment above resting pressure. Threshold, urge and maximal tolerated volume were to rectal balloon distension with air. Anal and rectal electrosensation measured threshold sensation with increasing amplitude of stimulation, recorded with a 1 cm bipolar ring electrode (5 Hz, 0.1 ms and 10 Hz, 0.5 ms respectively) (Kenefick 2002#).

Timing of outcome measures

Four trials reported results at the end of the crossover period (Dinning 2015#; Kenefick 2002#; Sørensen 2010#; Vaizey 2000#), while the trials by Kahlke 2015# and Leroi 2005# also reported results at the end of a three‐month follow‐up after the crossover period ended. The trial by Thin 2015 reported results at baseline prior to intervention, at 3 months and 6 months follow up. Tjandra 2008 reported results at the end of a 12‐month follow‐up.

Excluded studies

See 'Characteristics of excluded studies' table.

Risk of bias in included studies

The results of the methodological quality assessment along with justifications are covered in the 'Risk of bias' tables (see 'Characteristics of included studies' table) for each trial and are illustrated in Figure 2 and Figure 3.

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

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

Allocation

The information is provided in the 'Characteristics of included studies' table and presented visually in Figure 2 and Figure 3.

Random number generation

The method of randomisation was unclear in the trial conducted by Tjandra 2008. Thin 2015 randomised patients remotely using a statistician uninvolved in recruitment: requests were made by and actioned by email. Dinning 2015# utilised an off‐site (NHMRC) clinical trials centre automated phone link. Leroi 2005# and Kahlke 2015# randomised participants by using a random numbers table. Similarly, Sørensen 2010# randomised participants using shuffling of envelopes but did not indicate how the envelopes were filled. Kenefick 2002# and Vaizey 2000# provided no details of the method for deciding which participants were initially allocated to stimulator 'off' or 'on'.

Concealment of allocation

Five of the trials did not explicitly state that allocation was concealed (Kahlke 2015#; Kenefick 2002#; Leroi 2005#; Sørensen 2010#; Vaizey 2000#). In the trial by Thin 2015 and Dinning 2015# randomisation and allocation was made remotely, hence low risk of bias. The Tjandra 2008 trial, which used sealed envelopes, reported that the quality of allocation concealment was at low risk of bias.

Blinding

In the trial carried out by Thin 2015, the outcome assessor was blinded to participant allocation. The participants and the outcome assessor were not blinded to the allocation in the trial conducted by Tjandra 2008. In the remaining six trials the participants and outcome assessors were blinded to the allocation (Dinning 2015#; Kahlke 2015#; Kenefick 2002#; Leroi 2005#; Sørensen 2010#; Vaizey 2000#).

Incomplete outcome data

There were no withdrawals in three trials (Kenefick 2002#; Sørensen 2010#; Vaizey 2000#). However, in the trial conducted by Thin 2015, 9 participants out of the 40 enrolled withdrew from the trial. Likewise, in the trial carried out by Leroi 2005#, 10 people out of the 34 enrolled prematurely discontinued the trial. Furthermore, in the trial conducted by Kahlke 2015#, 2 participants out of the 16 enrolled discontinued the trial. Similarly, in the trial conducted by Tjandra 2008 seven participants in the SNS group prematurely discontinued. In the trials conducted by Kenefick 2002# and Vaizey 2000#, most of the outcome results were reported for everyone who entered the trials. As the participants were analysed in the groups to which they were originally allocated, the analysis was effectively intention‐to‐treat. In the trial carried out by Dinning 2015#, out of the 59 participants enrolled, 6 withdrew. The reasons for discontinuation were depression (n = 1), too far to travel (n = 2), could not commit (n = 1), pregnancy (n = 1) and severe wound infection (n = 1).

Selective reporting

Five trials were at low risk of bias as they reported all outcomes for everyone (Dinning 2015#; Kahlke 2015#; Leroi 2005#; Tjandra 2008; Thin 2015); the others trials were unclear (see 'Risk of bias' tables Figure 2; Figure 3).

Other potential sources of bias

None of the trials reported other sources of bias and so were judged to be at low risk on this domain.

Effects of interventions

The eight trials, which comprised two parallel group trials (Thin 2015 and Tjandra 2008) and six crossover trials (Dinning 2015#, Kahlke 2015#, Kenefick 2002#, Leroi 2005#, Sørensen 2010#, and Vaizey 2000#), included 280 participants, of whom 177 received a permanently‐implanted stimulator.

Faecal incontinence

Six of these trials investigated the effects of SNS for faecal incontinence (Kahlke 2015#; Leroi 2005#; Sørensen 2010#; Thin 2015; Tjandra 2008; Vaizey 2000#).

1. Clinical outcomes (number cured and improved, episodes of faecal incontinence, ability to defer defaecation, urgency, incontinence score)

SNS versus control (medical therapy)

In the trial by Tjandra 2008 the SNS group experienced fewer episodes of faecal incontinence compared to the control group who received optimal medical therapy

both at 3 months (MD −5.20, 95% CI −9.15 to −1.25, Analysis 1.1.1)
and 12 months (MD −6.30, 95% CI −10.34 to −2.26, Analysis 1.1.2),

1.1 — Comparison 1 Faecal incontinence: parallel group data SNS vs medical therapy, Outcome 1 Episodes of faecal incontinence per week.

In addition, SNS was better than control in terms of:

Number of days of pad use per week (MD −1.40, 95% CI −2.59 to −0.21 at 3 months, Analysis 1.2.1; though this was no longer statistically significant at 12 months, MD −1, 95% CI −2.13 to 0.13, Analysis 1.2.2);
Cleveland Clinic Incontinence Score (MD −11, 95% CI −11.60 to −10.40 at 3 months; and MD −12.90 to −13.58 to −12.22, Analysis 1.5).

1.2 — Comparison 1 Faecal incontinence: parallel group data SNS vs medical therapy, Outcome 2 Days using pads per week.

1.5 — Comparison 1 Faecal incontinence: parallel group data SNS vs medical therapy, Outcome 5 Cleveland Clinic Incontinence Score.

Additionally, none of the participants in the SNS group had worsening of faecal incontinence, with 25 participants (47.2%) achieving continence, but the data for the control group were not reported (Tjandra 2008).

SNS versus PTNS

In the trial carried out by Thin 2015 the SNS group experienced fewer episodes of faecal incontinence compared to the PTNS group:

both at 3 months (MD −3.00, 95% CI −6.61 to 0.61, Analysis 2.1.1)
and 6 months (MD −3.20, 95% CI −7.14 to 0.74, Analysis 2.1.2).

2.1 — Comparison 2 Faecal Incontinence: parallel group data SNS vs PTNS, Outcome 1 Episodes of faecal incontinence per week.

Additionally, SNS was better than PTNS with regards to Cleveland Clinic Incontinence Score:

both at 3 months (MD −1.70, 95% CI −5.14 to 1.74, Analysis 2.2.1)
and 6 months (MD −3.00, 95% CI −6.74 to 0.74, Analysis 2.2.2).

2.2 — Comparison 2 Faecal Incontinence: parallel group data SNS vs PTNS, Outcome 2 Cleveland Clinic Incontinence Score.

Additionally, at the six months follow‐up three participants in the SNS group and PTNS group reported no episodes of faecal incontinence. Moreover, in the SNS group 10 participants had a 50% or greater reduction in episodes of faecal incontinence compared with 7 participants in the PTNS group (Thin 2015).

Faecal incontinence: crossover trials

In the trial conducted by Vaizey 2000#, during the period when the stimulator was 'off' the participants experienced an average of six episodes of faecal incontinence per week, compared with an average of one episode per week when the stimulator was 'on' (Analysis 3.2). SNS resulted in one participant being cured and the other experiencing a reduction in the number of episodes of faecal incontinence per week from 10 when the stimulator was 'off' to one when it was 'on' (Analysis 3.1).

3.2. Analysis.

Comparison 3 Faecal incontinence: crossover trials, Outcome 2 Episodes of faecal incontinence per week.

Episodes of faecal incontinence per week
Study	Group	Measure	Patients	Baseline	Three months	'Off' period	'On' period	Change (%)	Follow‐up	Notes
Kahlke 2015#		Mean (SD)	14	18 (19.6)		8.4 (8.7)	1 (1.7)	‐7.4 ( ‐ 88%)	0.3 (0.5)	p<0.05
Kahlke 2015#
Leroi 2005#	Group of 19 who chose 'on' following the crossover period	Median (range)	Baseline: 16  3 months: 16  'Off' period: 19  'On' period: 19  Follow‐up: 18	3.5 (0 to 16)	0.3 (0 to 3)	1.7 (0 to 9)	0.7 (0 to 5)	‐1 (‐59%)	0.5 (0 to 11)	1. P < 0.05: 'off' versus 'on' period; 3 months, 'off' period, 'on' period, follow‐up period versus baseline.
Leroi 2005#	Group of 5 who chose 'off' following the crossover period	Median (range)	Baseline: 5  3 months: 4  'Off' period: 5  'On' period: 5  Follow‐up: 5	7 (0 to 11)	1.9 (1 to 10)	1.7 (0 to 11)	3.7 (0 to 11)	2 (118%)	3.5 (0 to 10)
Sørensen 2010#		Mean	7			0	0	0
Sørensen 2010#
Thin 2015	PTNS group	Mean±SD	16	10.6 ± 11.2	5.8 ± 6.9			‐4.3 (‐41%)	6.3 ± 6.9	Follow up was at 6 months. % change is reported as the difference between baseline and at 6 months.
Thin 2015	SNS group	Mean±SD	15	12.1 ± 13.7	2.8 ± 2.5			‐9 (‐74%)	3.1 ± 4.0	Follow up was at 6 months. % change is reported as the difference between baseline and at 6 months.
Tjandra 2008	Control group	Mean±SD	60	9.2±13.4	8.1±14.1			0.2 (2%)	9.4±11.8	P > 0.05 at 3 months and 12 months compared with baseline. Follow‐up was at 12 months. % change is reported as the difference between baseline and at 12 months.
Tjandra 2008	SNS group	Mean±SD	53	9.5±12.8	2.9±6.3			‐6.4 (‐67%)	3.1±10.1	P < 0.0001 at all time points compared with baseline. Follow‐up was at 12 months. % change is reported as the difference between baseline and at 12 months.
Vaizey 2000#		Mean (range)	2	10 (5 and 15)		6 (2 and 10)	1 (0 and 1)	‐5 (‐83%)
Vaizey 2000#

Participants cured and improved on treatment
Study	Cured	%	Improved	%
Leroi 2005#	5/19	26	17/19	89
Thin 2015	3/15	20	10/15	67
Tjandra 2008	22/53	41.5	53/53	100
Vaizey 2000#	1/2	50	2/2	100

Bloating score (weekly average)
Study	Measure	Patient	Baseline	Off period	On period	hange(%)	Notes
Dinning 2015#	Mean (SD)	53	1.6 (0.6)	1.2 (0.7)	1.2 (0.7)	‐25%	Off period is sham, and on period is suprasensory stimulation

Adverse events
Study	Type	Number of adverse events
Dinning 2015#	Implanted pulse generator site (IPG) related pain	32
Dinning 2015#	Wound infection	12
Dinning 2015#	Leg pain/discomfort	4
Dinning 2015#	Abdominal pain/discomfort	3
Dinning 2015#	Urological	17
Dinning 2015#	Tachycardia	1
Dinning 2015#	Headache	2
Dinning 2015#	Altered mood	2

Study id	Measure	Participants	Baseline	Follow‐up	p‐value	Change (%)	Notes
Ganio 2002	Mean (range)	31	7.5 (1 to 11)	0.15 (0 to 2)	Not reported	−7.35 (−98%)
Jarrett 2004	Median (range)	46	7.5 (1 to 78)	1.00 (0 to 39)	< 0.0001	−6.50 (−87%)
Leroi 2001	Mean (SD)	4	3.0 (2.7)	0.50 (0.6)	Not reported	−2.50 (−83%)
Matzel 2003	Median	16	40%	0%	< 0.001		Matzel and colleagues reported the percentage of bowel movements that were faecally incontinent
Rosen 2001	Median (range)	16	2.0 (1 to 5)	0.67 (0 to 1.67)	Not reported	−1.33 (−67%)
Uludag 2002	Mean	34	8.66	0.67	< 0.01	−7.99 (−92%)
MDT‐301	Mean (SD)	37;33	16.4 (19.3)	2.7 (4.8)	< 0.001	−13.70 (−84%)	The number of participants was 37 at baseline and 33 at follow‐up

Date	Event	Description
17 August 2015	New search has been performed	Five new trials were added in this update (Sørensen 2010#; Tjandra 2008;Dinning 2015#; Kahlke 2015#; Thin 2015). Risk of bias assessment was redone of all included trials.
17 August 2015	New citation required but conclusions have not changed	Five new trials were added in this update (Sørensen 2010#; Tjandra 2008;Dinning 2015#; Kahlke 2015#; Thin 2015). Risk of bias assessment was redone of all included trials.

Date	Event	Description
9 October 2008	Amended	Converted to new review format.
23 May 2007	New citation required and conclusions have changed	Substantive amendment
10 January 2006	New search has been performed	Minor update

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Episodes of faecal incontinence per week	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Days using pads per week	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Days with incontinence per week	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Days with faecal staining per week	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
4.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Cleveland Clinic Incontinence Score	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
5.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 SF12 Physical	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 SF12 Mental	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
7.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 FIQL Lifestyle	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
8.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9 FIQL Coping/behaviour	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
9.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
10 FIQL Depression/self perception	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
10.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
10.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11 FIQL Embarrassment	1	Mean Difference (IV, Fixed, 95% CI)	Totals not selected
11.1 At 3 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
11.2 At 12 months	1	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	Statistical method	Effect size
1 Participants cured and improved on treatment	Other data	No numeric data
2 Episodes of faecal incontinence per week	Other data	No numeric data
3 Ability to defer defaecation (minutes)	Other data	No numeric data
4 Episodes of urgency per week	Other data	No numeric data
5 Days with pads per week	Other data	No numeric data
6 Cleveland Clinic Incontinence Score	Other data	No numeric data
7 Quality of life ‐ SF36	Other data	No numeric data
8 Quality of life ‐ SF12	Other data	No numeric data
9 Quality of life ‐ modified ASCRS	Other data	No numeric data
10 Quality of life ‐ ASCRS FIQL	Other data	No numeric data
11 Anorectal manometry	Other data	No numeric data
11.1 Maximal resting pressure (cm H20)	Other data	No numeric data
11.2 Maximal squeeze pressure (cm H20)	Other data	No numeric data
11.3 Squeeze pressure duration (seconds)	Other data	No numeric data
11.4 Rectal volume sensation ‐ threshold (ml)	Other data	No numeric data
11.5 Rectal volume sensation ‐ urge (ml)	Other data	No numeric data
11.6 Rectal volume sensation ‐ maximal tolerated (ml)	Other data	No numeric data
11.7 Rectal volume sensation ‐ constant sensation volume (ml)	Other data	No numeric data
12 Bowel movements per week	Other data	No numeric data
13 Soiling/week	Other data	No numeric data
14 Days with staining per week	Other data	No numeric data
15 Days with incontinence per week	Other data	No numeric data
16 Passive leakage/week	Other data	No numeric data

Quality of life ‐ SF12
Study	Group	Scale	Measure	Patients	Baseline	3 months	12 months	Notes
Tjandra 2008	Control	Physical health	Mean±SD	60	39.29±12.12	41.5±9.89	40.5±10.2	P > 0.05 at 3 months and 12 months compared with baseline.
Tjandra 2008		Mental health	Mean±SD	60	45.38±12.32	47.82±10.66	48.22±10.12	P > 0.05 at 3 months and 12 months compared with baseline.
Tjandra 2008	SNS	Physical health	Mean±SD	53	39.81±11.14	43.18±11.68	42.22±9.25	P > 0.025 at all time points.
Tjandra 2008		Mental health	Mean±SD	53	45.25±11.09	50.16±10.41	49.22±10.88	P > 0.025 at 12 months. However, p = 0.005 at 3 months.

Quality of life ‐ ASCRS FIQL
Study	Group	Period	Measure	Patients	Lifestyle	Coping/behaviour	Depression/self‐perception	Embarrassment	Notes
Tjandra 2008	Control	Baseline	Mean±SD	60	2.26±0.98	1.79±0.82	2.59±0.72	1.81±0.52
Tjandra 2008		3 months	Mean±SD	60	2.12±0.91	1.85±0.92	2.68±0.65	1.7±0.67	P > 0.05 at 3 months and 12 months compared with baseline.
Tjandra 2008		12 months	Mean±SD	60	2.31±0.89	1.86±0.88	2.64±0.84	1.78±0.61	P > 0.05 at 3 months and 12 months compared with baseline.
Tjandra 2008	SNS	Baseline	Mean±SD	53	2.39±0.99	1.89±0.82	2.65±0.84	1.93±0.78
Tjandra 2008		3 months	Mean±SD	53	3.34±0.72	2.87±0.8	3.31±0.77	2.89±0.85	P < 0.0001 at 3 months and 12 months compared with baseline.
Tjandra 2008		12 months	Mean±SD	53	3.31±0.72	2.68±0.87	3.25±0.8	2.76±0.94	P < 0.0001 at 3 months and 12 months compared with baseline.

Quality of life ‐ SF36
Study	Group	Measure	Role‐emotional	General health	Mental health	Bodily pain	Physical functioning	Role‐physical	Social function	Vitality
Thin 2015	SNS (3 months) : n=15	Mean (95% CI)	11.1 (‐19.8 to 42.1)	‐6.8 (‐20.2 to 6.7)	‐2.7 (‐13.8 to 8.4)	2.7 (‐15.7 to 21.0)	‐9.8 (‐24.2 to 4.7)	26.7 (3.6 to 49.8)	19.1 (3.7 to 34.6)	2.7 (‐8.9 to 14.3)
Thin 2015	PTNS (3 months) n=1 6	Mean (95% CI)	20 . 0 (−15 . 4 to 55 . 4)	‐4.0 (‐12.1 to 4.1)	8.3 (1.2 to 15.3)	‐5.5 (‐20.5 to 9.5)	9.1 (‐3.4 to 21.7)	6.7 (‐23.6 to 36.9)	13.3 (1.2 to 25.5)	2.0 (‐3.7 to 7.7)
Thin 2015	SNS ( 6 months) n=15	Mean (95% CI)	17.8 (‐12.5 to 48.1 )	‐ 4.8 (‐15.4 to 5.8)	1 .7 (‐12.2 to 15.5)	7.2 (‐12.7 to 27.0 )	‐9.1 (‐25.4 to 7.3)	25.0 (‐5.5 to 55.5)	17.5 (‐0.4 to 35.4)	8.7 (‐4.3 to 21 .6 )
Thin 2015	PTNS ( 6 months) n=16	Mean (95% CI)	16.7 (‐0.5 t o 33.8 )	‐7.5 (‐14.7 to ‐0.3)	10.5 (5.1 to 15.9)	0.9 ( ‐8.8 to 10.6 )	9.1 (‐3.3 to 21.4)	0.0 (‐22.3 to 22.3)	10.9 (‐0.7 to 22.6)	0.0 (‐8.9 to 8.9)
Vaizey 2000#	Baseline	Mean (range)	33 (only recorded for one participant)	79 (72 to 86)	70 (64 to 76)	52 (30 to 74)	71 (70 to 72)	38 (0 to 75)	44 (12 to 75)	40 (10 to 70)
Vaizey 2000#	Nine months	Mean (range)	83 (66 to 100)	92 (both 92)	80 (72 to 88)	92 (84 to 100)	83 (80 to 85)	100 (both 100)	88 (75 to 100)	75 (70 to 80)

Quality of life ‐ SF36
Study	Group	Measure	Role ‐ emotional	General health	Mental health	Bodily pain	Physical functioning	Role‐physical	Social function	Vitality
Dinning 2015#	Baseline (n=59)	Mean (SD)	44 (13)	42 (12)	46 (13)	43 (9)	49 (10)	44 (12)	42 (12)	44 (11)
Dinning 2015#	Sham/Off (n=53)	Mean (SD)	46 (13)	45 (11)	47 (11)	42 (11)	49 (9)	43 (11)	42 (12)	45 (12)
Dinning 2015#	Suprasensory/On (n=53)	Mean (SD)	49 (8)	46 (10)	48 (10)	45 (9)	51 (8)	48 (10)	46 (11)	46 (10)

Outcome or subgroup title	Statistical method	Effect size
1 Bowel movements per week	Other data	No numeric data
2 Time with abdominal pain and bloating (%)	Other data	No numeric data
3 Wexner Constipation Score	Other data	No numeric data
4 Symptom Analogue Score	Other data	No numeric data
5 Anorectal manometry	Other data	No numeric data
5.1 Maximal resting pressure (cm H₂0)	Other data	No numeric data
5.2 Maximal squeeze pressure (cm H₂0)	Other data	No numeric data
5.3 Rectal volume sensation ‐ threshold (ml air)	Other data	No numeric data
5.4 Rectal volume sensation ‐ urge (ml air)	Other data	No numeric data
5.5 Rectal volume sensation ‐ maximal tolerated (ml air)	Other data	No numeric data
6 Stool frequency (per week)	Other data	No numeric data
7 Pain score (weekly average)	Other data	No numeric data
8 Bloating score (weekly average)	Other data	No numeric data
9 Quality of life ‐ SF36	Other data	No numeric data
10 Adverse events	Other data	No numeric data

Methods	Randomised crossover trial  Blinding: participant and investigator.  Follow‐up: 3 weeks (long‐term follow‐up not reported)  Setting: two centres in Sydney, Australia.  Intention‐to‐treat analysis: yes Withdrawals: 6 participants
Participants	Enrolled: 59  Median age (range): 42 (19 to 74)  Gender: male (4); female (55).  Duration of symptoms: constipation > 10 years (43), 5 to 10 years (7), and 2 to 5 years (9)  Received permanent implant: 55  Lost to follow‐up: 2  Inclusion criteria: All of the following: i) A complete bowel movement on less than 3 days per week, for at least 2 of 3 weeks of the surveillance phase (determined from daily stool diaries). ii) Aged 18 to 75 yrs; iii) Scintigraphically confirmed slow colonic transit defined as colonic isotope retention > 20% at 96 h (laboratory upper limit of normal = 1% at 96 h; iv) Normal anorectal manometry, with no evidence of dyssynergia and confirmed ability to expel rectal balloon; v) With no radiographic evidence of functional (e.g. pelvic floor dyssynergia) or anatomical (e.g. significant rectocele with retention of contrast, occluding intussusception) impediment to the expulsion of the radio‐opaque contrast; vi) Failed symptomatic response to standard therapies including osmotic agents, faecal emollients, laxatives, dietary modification and exercise; and vii) A normal colonoscopy within 5 years Exclusion criteria: Any one of the following; i) Metabolic, neurogenic or endocrine disorder(s) known to cause constipation; ii) Drugs listing constipation as a potential side effect; or iii) Non‐English speaker or an inability to provide informed consent. Any one of the following; i) Prior abdominal radiotherapy; ii) Prior abdominal surgery (except cholecystectomy, appendicectomy, inguinal hernia repair, splenectomy, fundoplication; oophorectomy or hysterectomy); iii) Current or planned pregnancy; iv) Co‐morbidity considered to put the patient at risk from surgical electrode implantation; or v) History of malignancy.
Interventions	All participants underwent peripheral nerve evaluation. Then participants received permanent electrode implantation. Intervention: 'suprasensory' stimulation for 3 weeks Control: Sham stimulator off for 3 weeks. Wash out period: 2 weeks.
Outcomes	Bowel diary (stool frequency per week; bowel motion per week; feeling of complete evacuation per week; straining per week; laxative free days per week; soft stool per week; normal stool per week; hard stool per week; pain score weekly average; bloating score weekly average; weekly global satisfaction score; weekly bothersome score). Quality of life SF‐36.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Utilised an off‐site NHMRC clinical trials centre automated phone link.
Allocation concealment (selection bias)	Low risk	Utilised an off‐site NHMRC clinical trials centre automated phone link.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants and personnel blinded.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Investigator blinded.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Out of the 59 participants enrolled, 6 withdrew. The reasons for discontinuation were depression (n = 1), too far to travel (n = 2), could not commit (n = 1), pregnancy (n = 1) and severe wound infection (n = 1).
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	Low risk	None identified

Methods	Randomised parallel‐arm  Blinding: Investigator  Follow‐up: 6 months  Setting: two centres, London, UK  Intention‐to‐treat analysis: yes
Participants	Enrolled: 40  Age: 59 years (mean age)  Gender: male (1); female (39).  Duration of symptoms: not stated  Inclusion criteria: Age > 18 years Meets NICE criteria (2007) for symptom severity and failure of previous conservative therapy Exclusion criteria: Inability to provide informed consent for the research study Severe concomitant medical condition precluding randomization to operative treatment Neurological diseases, such as diabetic neuropathy, multiple sclerosis and progressing Parkinson’s disease Other medical conditions precluding stimulation, such as bleeding disorders, certain cardiac pacemakers, peripheral vascular disease Congenital anorectal anomalies or absence of native rectum as a result of surgery Present evidence of external full‐thickness rectal prolapse. Previous rectal surgery (rectopexy/resection) < 12 months ago Stoma in situ Chronic bowel disease, such as inflammatory bowel disease, chronic uncontrolled diarrhoea Anatomical limitations that would prevent successful placement of electrodes Pregnancy or intention to become pregnant. Previous experience of SNS or PTNS
Interventions	SNS vs percutaenous tibial nerve stimulation (PTNS)
Outcomes	FI episodes/week (total, urge and passive); symptom severity scoring with Cleveland Clinic Incontinence Score (CCIS); quality of life measurements using EQ‐5D, FIQL and SF‐36.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization sequence generated remotely by a statistician uninvolved in recruitment, and requests were made and actioned by e‐mail.
Allocation concealment (selection bias)	Low risk	Randomization sequence generated remotely by a statistician uninvolved in recruitment, and requests were made and actioned by e‐mail.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Investigator blinded but unclear whether patients were blinded.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Investigator blinded but unclear whether patients were blinded.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Out of 23 participants randomised to the SNS group, 4 withdrew prior to intervention (intercurrent disease n = 1; symptoms resolved n = 1; declined further participation n = 2). Then a further 3 participants failed temporary SNS. Another participant who passed the temporary SNS withdrew from treatment owing to urgent investigation for new pararectal cyst. Moreover, in the PTNS group 1 participant withdrew before intervention (symptoms resolved).
Selective reporting (reporting bias)	Low risk	Pre‐specified outcomes in study protocol reported.
Other bias	Low risk	None identified

Methods	Randomised controlled trial, parallel groups  Blinding: no  Follow‐up: 12 months Withdrawals: 7  Setting: single centre, Australia  Intention‐to‐treat analysis: no
Participants	Enrolled: 120  Age: 63.9 (±13.2) in SNS group. 63 (±12.1) in the control group.  Gender: male (9); female (111).  Duration of symptoms: not mentioned  Received permanent implant: 53  Lost to follow‐up: none  Inclusion criteria: involuntary passage of solid or liquid stool at least once per week; refractory to medical therapy and pelvic floor exercises; and aged 35 to 86 years.  Exclusion criteria: rectal prolapse, inflammatory bowel disease, congenital anorectal malformation, neurologic disorders such as Parkinson’s disease, multiple sclerosis, spinal cord injury, stoma in situ, pregnancy, external anal sphincter defect of more than 120° of the circumference, bleeding diathesis, and mental or physical disability precluding adherence to study protocol.
Interventions	A (53): SNS (Peripheral nerve evaluation for at least 7 days; then permanent implantation in participants with ≥ 50% reduction in faecal incontinence episodes per week or ≥ 50% reduction in number of days with faecal incontinence per week based on the two‐week bowel diary) B (60): control group (optimal medical therapy). Optimal medical therapy includes bulking agents, pelvic floor exercises with a team of dedicated physiotherapists, and dietary management on fluid and fibre with a team of dieticians.
Outcomes	Anorectal physiology, severity of incontinence, two‐week bowel diary (number of incontinent episodes per week, days with incontinence per week, days with staining per week and days with pads per week), faecal incontinence quality of life. Continuous data: all mean (SD) N Episodes of faecal incontinence per week at 3 months: A 2.9 (6.3) 53, B 8.1 (14.1) 60 Episodes of faecal incontinence per week at 12 months: A 3.1 (10.1) 53, B 9.4 (11.8) 60 Days with pads per week at 3 months: A 1.6 (2.6) 53, B 3 (3.8) 60 Days with pads per week at 12 months: A 2.2 (3) 53, B 3.2 (3.1) 60 Days with incontinence per week at 3 months: A 1 (1.7) 53, B 2.9 (2.4) 60 Days with incontinence per week at 12 months: A 1 (1.7) 53, B 3.1 (3.1) 60 Days with staining per week at 3 months: A 1.3 (1.7) 53, B 4.5 (2.1) 60 Days with staining per week at 12 months: A 1.4 (2) 53, B 4.5 (2.3) 60 Cleveland Clinic Incontinence Score [higher = worse] at 3 months: A 1.1 (1) 53, B 12.1 (2.1) 60 Cleveland Clinic Incontinence Score at 12 months: A 1.2 (1.8)53, B 14.1 (1.9) 60 Quality of Life SF12 [higher = better]: SF12 Physical at 3 months: A 43.18 (11.68) 53, B 41.5 (9.89) 60 SF12 Physical at 12 months: A 42.22 (9.25) 53, B 40.5 (10.2) 60 SF12 Mental at 3 months: A 50.16 (10.41) 53, B 47.82 (10.66) 60 SF12 Mental at 12 months: A 49.22 (10.88) 53, B 48.22 (10.12) 60 Quality of Life ASCRS Faecal Incontinence Quality of Life Index (FIQL) [higher = better]: Lifestyle at 3 months: A 3.34 (0.72) 53, B 2.12 (0.91) 60 Lifestyle at 12 months: A3.31 (0.72) 53, B 2.31 (0.89) 60 Coping at 3 months: A 2.87 (0.8) 53, B 1.85 (0.92) 60 Coping at 12 months : A 2.68 (0.87) 53, B 1.86 (0.88) 60 Depression/self perception at 3 months: A 3.31 (0.77) 53, B 2.68 (0.65) 60 Depression/self perception at 12 months: A 3.25 (0.8) 53, B 2.64 (0.84) 60 Embarrassment at 3 months: A 2.89 (0.85) 53, B 1.7 (0.67) 60 Embarrassment at 12 months: A 2.76 (0.94) 53, B 1.78 (0.61) 60 Anal manometry measures only reported in SNS Group A at 3 months Ability to defer defecation and worsening of faecal incontinence only reported in SNS Group A at 3 months Adverse effects only reported for SNS group: implantation site pain (6 %) particularly in slimmer participants, seroma (2%) which required percutaneous aspiration and excessive tingling in the vaginal region (9%).
Notes	Severity of incontinence was assessed by the Wexner's incontinence score. The score ranged from 0 (normal continence) to 20 (maximum incontinence).  Quality of life was assessed using Fecal Incontinence Quality of Life index (FIQL) of The American Society of Colon and Rectal Surgeons; and the standard short form‐12 health survey quality of life questionnaire (SF‐12).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation was performed from the central registry by using sealed envelopes
Allocation concealment (selection bias)	Low risk	Randomisation was performed from the central registry by using sealed envelopes
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Participants could not be not blinded and the control group did not have sham stimulation
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Unblinded assessment for both control and experimental group
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	7 participants in the SNS group stopped early but have been counted in the denominators
Selective reporting (reporting bias)	Low risk	Data from all participants reported
Other bias	Low risk	None identified

Methods	Randomised crossover trial.  Blinding: participant and investigator.  Follow‐up: the study consisted of two 2‐week intervals with subsensory stimulation either 'on' or 'off'  Setting: single centre, UK  Withdrawals: none  Intention‐to‐treat analysis: yes
Participants	Enrolled: 2  Age: 65 and 61 years  Gender: female (2).  Duration of symptoms: 2.5 years and 3 years.  Inclusion criteria: passive faecal incontinence; intact external sphincter; informed consent given  Exclusion criteria: not stated
Interventions	To implant the unilateral electrode (Medtronic InterStim model 3080), the sacral nerve root (usually S3) that produces the maximal anal response was identified via percutaneous needle stimulation. An incision over the sacrum allows access to the sacral foramen. The permanent electrode was inserted directly and secured to the sacral periosteum after checking its correct placement by stimulation. A connecting lead (model 7495) was then tunnelled to the anterior abdominal wall to be connected to the stimulator. The Implantable Pulse Generator (model 3023) is programmable using telemetry. The voltage required for stimulation was between 0.5 and 2 volts at a frequency of 15 pulses per second and a pulse width of 210 μs. A Intervention: stimulator 'on' for two weeks with subsensory stimulation B Control: stimulator 'off' for two weeks No treatment‐free period between the weeks
Outcomes	Episodes of faecal incontinence for liquid or solid stool; anal manometry; psychological assessment; quality of life.
Notes	Quality of life was assessed with the SF‐36 instrument, score 0 (poor) to 100 (excellent), reporting the domains of role‐emotional, general health, mental health, bodily pain, physical functioning, role‐physical, social function, and vitality.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear.
Allocation concealment (selection bias)	Unclear risk	Unclear.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants and outcome assessors were blinded to the allocation.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Participants and outcome assessors were blinded to the allocation.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Results were reported for everyone who entered the studies and the participants were analysed in the groups to which they were originally allocated, effectively resulting in an intention‐to‐treat analysis.
Selective reporting (reporting bias)	Unclear risk	Psychological assessment not reported in the results.
Other bias	Low risk	None identified.

Study	Reason for exclusion
Dinning 2007	No comparator used and study is a trial of wave measurements during the temporary stimulation phase
Dinning 2013	Only reported proportion of participants who satisfied primary outcome measure. Secondary outcomes (pain, bloating, laxative‐free days) and longer term responses of primary outcomes are not available. Author was contacted.
Knowles 2012	No report on results of permanent SNS stimulation.

Trial name or title	Sacral nerve stimulation or anal bulking therapy for faecal incontinence ‐ a comparative study
Methods	Randomised controlled trial, parallel groups  Blinding: unknown  Follow‐up: 12‐months. Withdrawals: unknown  Setting: multi‐centre, Sweden  Intention‐to‐treat analysis: unknown
Participants	100
Interventions	SNS vs. anal bulking therapy
Outcomes	Primary outcomes: proportion of participants responding with > 50% reduction in the number of faecal incontinence episodes compared to baseline, one year after randomisation. Secondary outcomes: change in number of faecal incontinence episodes, change in deferring time, change in incontinence score, change in quality of life, and adverse events.
Starting date	01/03/2013
Contact information	Dr Lars Börjesson lars.g.borjesson@vgregion.se Sahlgrenska University Hospital  Department of Surgery  Gothenburg  416 85  Sweden
Notes	Recruitment end date: 01/03/2015

Trial name or title	SaFaRI: Sacral nerve stimulation versus the FENIX TM magnetic sphincter augmentation for adult faecal incontinence: a randomised Investigation
Methods	Randomised controlled trial, parallel groups  Blinding: unknown  Follow‐up: 2 weeks; 6, 12, and 18 months. Withdrawals: unknown  Setting: multi‐centre (at least 20 centres), UK  Intention‐to‐treat analysis: yes
Participants	350
Interventions	FENIXTM magnetic anal sphincter (MAS) vs. SNS
Outcomes	Primary outcomes: proportion of patients with device in‐situ at 18 months, and proportion of patients reporting > 50% improvement in continence scores. Secondary outcomes: complications, quality of life, and cost‐effectiveness.
Starting date	01/01/2015
Contact information	Primary contact: Professor David Jayne. Julie Croft: safari@leeds.ac.uk
Notes	Funded by NIHR HTA; co‐ordinated by the Clinical Trials Research Unit, University of Leeds.

Trial name or title	Sacral nerve stimulation versus the FENIXTM magnetic anal sphincter for adult faecal incontinence: A randomised investigation (SaFaRI).
Methods	Randomised controlled trial, parallel groups  Blinding: unknown  Follow‐up: 2 weeks; 6, 12, and 18 months. Withdrawals: unknown  Setting: multi‐centre (at least 20 centres), UK  Intention‐to‐treat analysis: yes
Participants	350
Interventions	FENIXTM magnetic anal sphincter (MAS) vs. SNS
Outcomes	Primary outcomes: proportion of patients with device in‐situ at 18 months, and proportion of patients reporting > 50% improvement in continence scores. Secondary outcomes: complications, quality of life, and cost‐effectiveness.
Starting date	01/01/2015
Contact information	Primary contact: Professor David Jayne. Julie Croft: safari@leeds.ac.uk
Notes	Funded by NIHR HTA; co‐ordinated by the Clinical Trials Research Unit, University of Leeds. ISRCTN16077538 DOI 10.1186/ISRCTN16077538 http://www.isrctn.com/ISRCTN16077538

PERMALINK

Sacral nerve stimulation for faecal incontinence and constipation in adults

Mohamed A Thaha

Amin A Abukar

Noel N Thin

Anthony Ramsanahie

Charles H Knowles

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Faecal incontinence

Constipation

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Faecal incontinence

Constipation

Secondary outcomes

Faecal incontinence

Constipation

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Design

Sample sizes

Faecal incontinence:

Constipation

Setting

Participants

Compliance and treatment received

Age

Cause/type of faecal incontinence or constipation

Interventions

Duration of treatment and follow up

Outcomes

Timing of outcome measures

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Random number generation

Concealment of allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions