Surgery for constipation: systematic review and practice recommendations

To assess the outcomes of rectal excisional procedures in adults with chronic constipation.


Introduction Background and procedural variations
Chronic constipation is related to an inability to evacuate the rectum in over half of all adults presenting for specialist advice. This phenomenon, which may be variably described as obstructed defaecation or rectal evacuation disorder (and many other terms) is characterized by excessive straining, the feeling of incomplete evacuation, post-defaecatory seepage and often mucous discharge and pelvic pain [1]. In some of these patients, there is clinical and radiological (usually proctographic) evidence of a dynamic structural abnormality, leading to physical impediment to emptying during defaecation. The most common abnormalities are rectocoele and/or intussusception. Theoretically, these anatomical variants could lead to the features of obstructed defaecation by a process of loss of force vector (ballooning of the rectum into a rectocoele or invagination of the rectum into an intussusception, rather than evacuation of stool on straining) or mucosal obstruction (in the case of an intussusception) [1]. Correction of these variants can be carried out by surgically excising the redundant rectal wall, i.e. that ballooning out or prolapsing in, thus restoring 'normal' anatomy.

Techniques
The stapled transanal rectal resection (STARR) procedure uses two circular procedure for prolapsing haemorrhoids (PPH)01 stapling devices (Ethicon Endosurgery, INC., Cincinnati, Ohio, USA), or at times, two circular PPH03 devices. They are used sequentially to remove the anterior rectal wall prolapse and then the posterior rectal wall prolapse. These full thickness rectal wall excisions result in defects that are anastomosed with staples contained within the device. The STARR procedure was derived from the technique of stapled haemorrhoidopexy and was introduced into clinical practice by Antonio Longo using the PPH01 device [2]. Initially, most STARR procedures were performed with PPH01; however, subsequently the PPH03 device was developed, essentially for haemorrhoidopexies, to reduce intra-operative staple line bleeding. The two devices, PPH01 and PPH03, differ only with respect to the closed staple height, which varies between 1-2.5 mm and 0.75-1.5 mm, respectively. The PPH03 device is no longer licensed for use for STARR because of concerns that the staple line height was insufficient for the thickness of tissue that was resected and anastomosed. As with all new techniques, in order to ensure that safety and efficacy concerns were addressed, a number of initiatives were put in place by the manufacturer. The procedure could only be performed by trained surgeons who then mentored other surgeons. In the UK, the PPH01 device is no longer used for haemorrhoidopexies, and consequently Ethicon will only permit the sale of this device to institutions with surgeons trained in the STARR technique. The European STARR registry was established through collaboration between Ethicon Endosurgery and the colorectal societies in the UK, France, Germany and Italy. Outcomes from this registry have been published and, in response, the National Institute for Health and Care Excellence (NICE) issued interventional procedure guidance in 2010. The NICE guidance concluded that 'the current evidence on safety and efficacy of stapled transanal rectal resection (STARR) for obstructed defaecation syndrome (ODS) is adequate . . . . . . the procedure may therefore be used with normal arrangements for clinical governance, consent and audit.' [www.nice.org.uk/ ipg351].
Despite its proven safety and efficacy, the STARR procedure still had limitations. These were primarily related to the stapling device itself. The two biggest concerns were that the rectal resection was performed 'blind' within the low rectum and that the volume of the resected tissue was entirely dependent upon the capacity of the stapling device housing, rather than the extent of the prolapse. As such, the surgeon was unable to modify the extent of resection in relation to the size of the prolapse. Consequently, a new stapling device, the Contour Transtar curved cutter stapler (Ethicon Endosurgery, INC.), was introduced [3] in order to allow a tailored correction of the prolapse by removing more tissue. It would also allow this resection to be performed under direct vision. The Contour Transtar stapler utilises replaceable staple cartridges allowing multiple firings with a single device.
Although STARR has been heavily popularized, it is not the first procedure to use of rectal excision as a means of treating prolapse. Edward Delorme originally described his procedure for full-thickness external rectal prolapse in 1900 [4]. There have been several modifications, one being an intra-anal Delorme's procedure for the treatment of recto-anal intussusception [5]. Unlike STARR and Contour Transtar, there is no full-thickness rectal wall excision and only the redundant mucosa is excised. The mucosa is stripped cephalad, and the underlying muscle is plicated together with sutures to concertina the prolapse.

Scope
The purpose of this review was to assess the efficacy and harms of rectal excisional procedures for internal prolapse and/or rectocoele in adults. Procedures considered beyond the scope of systematic review included: (i) those where 'tightening' is effected without excision (covered in rectovaginal reinforcement systematic review); (ii) those where only mucosa is excised (mucosectomy) and there is no plication of the muscular wall; (iii) those where a mucosectomy and plication are limited to the anterior wall (covered in rectovaginal reinforcement systematic review); (iv) those where the whole rectum is resected rather than a component of the wall, i.e. as occasionally performed for megarectum [6]. Studies where outcomes could not be segregated by eligible procedure were also excluded, due to a mixed patient population with internal and external rectal prolapse, mixed indications, including numerous pelvic floor abnormalities or limited postoperative outcomes, transanal endocopic microsurgery procedures, technical reports, or cost analyses only.
2000 and 2014 contributed to the systematic review, providing data on outcomes in a total of 8340 patients (range 20-2224 patients per study) ( Table 1). Specific exclusions after full-text review (and after exclusion of non-English language publications) included nine studies where the population sample was confirmed to be < 20 patients, eight where follow-up was < 12 months; five studies with out of scope procedures [8][9][10][11], three studies where data were considered duplicate [12,13], and three where outcomes could not be segregated by eligible procedure. Other exclusion criteria were: indication not constipation (n = 1) and lack of extractable outcome data (n = 5).
The quality of studies varied. The 47 included studies (Table 1) included 42 observational studies and five randomised controlled trials (RCTs). The latter included three good quality RCTs (level IB) with low levels of predicted bias, and two with less well described methodology (level IIB). The remaining 42 observational studies encompassed several good quality prospective cohort studies with low levels of bias including four prospectively maintained funded registries (level IIB). Other studies were a mix of prospective and retrospective case series. Mean study follow-up was 1.9 years (range 1.0-5.5 years); 36 studies originated from European centres, three from the USA and eight from other countries.

Perioperative data
Perioperative data were reported by 47 studies reporting 55 procedures (Tables 2 and 3). Measures of variation of continuous measures included ranges or standard deviations but were inconsistently reported. Average procedural duration and length of stay (LOS) varied between procedures but design heterogeneity, small numbers of studies and large range of values precluded any clear conclusions ( Table 3). The overall average duration of procedures reported by cohorts was 44 (range 23-95) min, and the overall average LOS was 3.0 (range 1.0-8.0) days (Figs 2 and 3). In one RCT of 100 patients (STARR vs Contour Transtar) [37], Contour Transtar took significantly longer to perform (52.2 vs 42.4 minutes P = 0.008). However this reduction was not apparent more broadly within the pooled findings (Fig. 2). While LOS for intra-anal Delorme's procedure was shorter (2 days) this was based on only one study reporting this outcome in 34 patients.   Summary evidence statements: perioperative data 1 The average duration of procedures was 44 min, although this ranged widely between studies from 23 to 95 min (level II). 2 The average length of stay was 3 days although this ranged widely between studies from 1 to 8 days (level II). 3 There was inadequate evidence to determine variations in procedural duration or length of stay by type of procedure (level IV).

Harms Perioperative complications
Five measures of perioperative harm were reported by a majority of studies within the review. In general, there was considerable heterogeneity between cohort findings. This heterogeneity may have reflected (for example) differing inclusion, procedural content, context of care, or thresholds or conventions for recording complications.
Overall procedural complication rates varied between cohorts from 0% to 61% Fig. 4. Random effects metaanalysis found the overall complication rate to be 16.9% (95% CI: 12.7-21.5%), I 2 = 93%. The Contour Transtar procedure reported a lower overall complication rate of 8.9% (95% CI: 5.1-13.5%), I 2 = 70%, although cohort findings within this category were heterogeneous and non-comparative. The two RCTs (IB) comparing STARR to Contour Transtar reported no difference in perioperative complication rates [37,41], although these included only 161 patients in total and six events. A more generalisable estimate of the overall complication rate may come from the European STARR registry which reported an overall morbidity rate of 36% from 2838 patients [24]. The inclusion criteria for morbidity in this registry incorporated urgency (20%), but this was not universally applied in most other studies, explaining its discrepancy from the pooled findings.
Rates of post-operative bleeding requiring treatment varied between cohorts from 0% to 19% (Fig. 5). Random effects meta-analysis found the pooled bleed rate to be 1.6% (95% CI: 0.9-2.5%), I 2 = 63%. Regarding intra-operative bleeding from staple lines, results from one prospective cohort study (IIB) showed that less intraoperative staple line sutures were required with PPH03 compared with PPH01 [18]. Details of treatments provided (e.g. transfusion or re-operation) were unclear in many instances.
Sepsis rates varied from 0% to 16%, but were consistently low, occurring in no patients in 78% of cohorts reported. Random effects meta-analysis found the sepsis rate to be 0.2% (95% CI: 0.0-0.7%), I 2 = 63%. Septic complications were mostly related to urinary tract infections, C. difficile infection, or more rarely, pneumonia. Sepsis rates for Delorme's procedure appeared high at 6.2% (95% CI: 0.0-19.4%), I 2 = 82%, but findings from the three studies were highly heterogeneous and the only moderately large study [49] reported a 1.2% sepsis rate. The rate of post-operative anastomotic dehiscence (wound rupture) varied between cohorts from 0% to 13% but was consistently low, occurring in no patients in 76% of cohorts reported. Random effects meta-analysis found the pooled rate of anastomotic dehiscence to be 0.3% (95% CI: 0.0-0.8%), I 2 = 62%.  There was inadequate evidence to select between procedures on the basis of individual complications. No deaths were reported in any cohort of patients reporting mortality, which included a total of 5896 patients ( Table 3). Incidences of serious complications were reported but these were rare. These included: one case of sepsis with retropneumoperitoneum [19], two cases of intra-operatively recognized rectal perforation requiring diverting colostomy [32,35], one case of rectovaginal abscess requiring diversion [50], one case of suture line disruption requiring diversion [30], and one case of rectal necrosis requiring a colostomy [24]. There were no cases of inadvertent small bowel injury as a result of peritoneal inclusion with anterior rectal wall excision and enterocoele excision. This is despite the fact that in one study [42], 84 of 101 patients who had Contour Transtar, and 14 of 69 patients who had PPH01, had peritoneum present in the resected tissue.

Long-term adverse outcomes
Although obstructed defecation is a benign condition, it may have a significant impact on a patient's quality of life. Studies have consistently reported a number of long-term conditions and symptoms occurring after stapled rectal excision procedures that may have a further negative impact upon quality of life. Measures reported by more than half of cohorts include: rectal stenosis (83% of cohorts), recto-vaginal fistula (76%), pain/proctalgia (70%), and urgency after defaecation (72%); those less consistently reported include recurrent prolapse (46%) and dyspareunia (33%) ( Table 4).
Random effects meta-analysis found that reported rectal stenosis rates were 0.2% (95% CI: 0.0-0.6%), I 2 = 30%, although individual cohorts varied from 0.0% to 7.4%, with no stenosis reported in 67% of cohorts. Recto-vaginal fistula was a very rare outcome, occurring in just 3 of 4851 patients (0.062%) studied. Rectal pain or proctalgia lasting > 6 months post-procedure was reported by 0.7% of patients (95% CI: 0.1-1.6%), I 2 = 79%, although individual cohorts varied from 0% to 17%, with no proctalgia reported in 53% of cohorts. Similarly to pooled findings, there was no difference in pain comparing STARR and Contour Transtar in the two level 1B RCTs [37,41]; at 36 months, the incidence of pain was significantly less compared with pre- operative values in both groups [37]. The two studies with the highest reported incidence of proctalgia were both PCHs (IIB). At 38 AE 18 months post-STARR (n = 30), 17% complained of intermittent anal pain, however despite the fact that pre-operative Cleveland Clinic Constipation scores were recorded, there was no analysis about de novo pain [8]. At 12 months post-STARR (n = 2838), 7.1% complained of persistent pain [24]. Again there was no analysis regarding the de novo nature of this pain.
Urgency of defecation, at least one year post-procedure, was reported by 5.2% of patients (95% CI: 2.7-8.2%), I 2 = 92%, although findings were heterogeneous and individual cohorts varied from 0% to 34% (Fig. 6). The European STARR registry reported higher urgency rates of 20% at 12 months [24] and the German STARR registry reported rates of 25% at 12 months [30]. Both groups of authors pointed to poor recording of pre-operative urgency symptoms, relying on the subset scores from patients' Symptom Severity Score (SSS). Schwander et al. [30] suggest that this high incidence reflected numbers with new onset urgency with a score of > 1 in SSS, 5.5% had score > 3, and 8% of patients observed a reduction in urgency compared with pre-op. Jayne et al. [24] acknowledged in a subsequent reply to an invited expert's analysis of their paper, that de novo urgency was not recorded in their registry but 20% of patients reported this as a complication. Analysis of the subset question in their SSS suggested that 39.9% of patients experienced urgency pre-operatively and this fell to 26.8% at 12 months post-STARR. The highest reported incidence of urgency came from an RCT (IB) comparing STARR/ PPH01 (34% at 3 years) with Contour Transtar (14% at 3 years) (P = 0.035) [37]. Urgency was reported as being de novo in both groups. However, the text in the results reports that the symptoms resolved in all but one patient (it is probable that this refers to tenesmus which is reported for one patient post-op). The other RCT (level IB) comparing STARR/PPH01 with Contour Transtar showed no difference in urgency rates after 24 months [41]. Summary evidence statements: harms 1 Evidence is largely drawn from observational studies and comparisons. Findings were heterogeneous, making estimates tentative and imprecise (level IV). 2 Overall procedural complication rates ranged from 0% to 61%. However, these complications may typically occur in about 13-22% of procedures (level II). 3 The Contour Transtar procedure may feature a lower overall complication rate although this needs to be confirmed with better research (level IV). 4 Post-operative bleeding requiring treatment may typically occur in 1-3% of patients (level II).
5 Post-operative sepsis and anastomotic dehiscence are rare complications rare typically occurring in less than 1% of patients (level II). 6 Serious acute post-operative complications are very rare occurring in about one in a thousand patients (level II). 7 In the longer term (12 months or more), rectal stenosis is a rare complication typically occurring in less than 1% of patients (level II). 8 The most common longer term adverse outcome is urgency of defaecation, typically occurring in up to 10% of patients (level II). 9 Longer term pain is experienced typically by less than 2% of patients (level II).  10 Rectovaginal fistula is a very rare longer term complication, occurring in about one in about one in every one thousand six hundred patients (level II). 11 There was insufficient evidence to establish whether specific post-operative or longer term complications varied between procedures (level IV). 12 In all studies where mortality was recorded, there were no deaths in a total of 5896 patients studied (level II).

Efficacy
Measurement of clinical outcome was consistently recorded in many studies using a variety of the available subjective summative scoring instruments for constipation symptoms (Table 6). These included the Cleveland Clinic Constipation Score (21 studies), obstructed defecation syndrome (ODS) score (30 studies), Symptom Severity Score (SSS) (eight studies). The Longo scoring system for ODS was reported for 56% of cohorts. Reduction in ODS occurred in 72.2% of patients (95% CI: 68.5-75.8%), I 2 = 83% (Fig. 7). There was considerable heterogeneity between findings, which varied from 53-91% and may reflect multiple causes. Although findings are imprecise, there was no robust evidence that one procedure produced greater score reductions than another. Improvement in constipation was reported in all studies where it was recorded. In many publications regarding surgical treatment of constipation, success has been defined as > 50% improvement in objective scores. A total of 18/21 studies reported > 50% improvement in CCS; all studies (30/30) reported > 50% improvement in ODS; 5/8 studies reported > 50% improvement in SSS. Significant improvement in ODS was seen in both arms of the two RCTs (IB) comparing STARR (PPH01) and Contour Transtar [37,41] and in the other level IB RCT comparing STARR with PPH01 and mucosectomy (with PPH01) together with levatoroplasty [15]. Whilst this  improvement was maintained at three years, without difference between procedures in one trial [37], in the other at 24 months the improvement was maintained in the Contour Transtar group but not in the STARR group and there was a significant difference in scores between the two groups [41]. In the other two RCTs of lesser quality (IIB), STARR (PPH01) was shown to be significantly better for functional outcome compared with biofeedback (although there was > 50% attrition rate in the biofeedback group) [21], and STARR (PPH01) was shown to give similar functional results compared with Intra-anal Delorme's with levatoroplasty [57]. In other non-randomised (level IIB) comparisons of STARR (PPH01) vs Contour Transtar [8,27,43] there was no difference in functional outcome. One prospective cohort study (level IIB) compared PPH01 with PPH03 [18] and showed no difference in functional outcome, and a further compared STARR with macrogol therapy [8]. There was no difference in outcome when looking at response rates, however it was unclear if groups were matched and no report on different laxative use was included in inclusion criteria. One would normally assume that macrogol or other medical management would have been tried before selecting a patient for a rectal excisional procedure.
The effect of time post-procedure and efficacy was examined in the four level I/IIB studies with a followup of > 30 months [8,37,53,57]. Three studies showed greater than 75% reduction in subjective scores [8,37,57] and the other greater than 50% reduction, regardless of the procedure provided.
Global 'success' or 'satisfaction' ratings (GSR) were obtained via a variety of methods in 28 of 55 cohorts (Table 5), where 'satisfied' or 'very satisfied', 'good', 'very good' and 'excellent' were interpreted as positive outcomes. Further studies also reported individual symptoms. No study reported acquiring data objectively using personnel not involved in the surgical care of the patient or data collection blind to intervention status. Most reports assert that the majority of patients undergoing PPH01, PPH03, Contour Transtar, and intra-anal Delorme's procedure were satisfied. Random effects meta-analysis found the global satisfaction rating to be 76.3% (95% CI: 72.8-79.5%), I 2 = 59% (Fig. 8). There was considerable heterogeneity between findings, which varied from 51% to 90% and may reflect multiple causes. Although findings are imprecise there was no evidence that one procedure produced higher levels of satisfaction than another.
Disease-specific and generic QoL scoring instruments were used less often including PAC QoL (11 studies, data not shown), EQ-VAS (six studies, data not shown), EQ-5D (two studies, data not shown), FIQL (one study, data not shown). Using such instruments, improvements in QoL were associated with symptom improvements in most studies except for Ribaric et al. [56] who showed non-significant improvements in EQ-5D and EQ VAS in 100 patients 12 months after Contour Transtar and Madbouly et al. [29] who, reporting on 46 patients after PPH01, showed significant improvements in PAC QoL at 18 months but not at 42 months.
The aim of rectal excision procedures is to enable normal evacuation and assessment of constipation symptoms is therefore the most important outcome. However, many patients also suffer from incontinence, either as a result of neurogenic/traumatic sphincter pathologies or from the effects of a high grade recto-anal intussusception. Instrumented excisional procedures are postulated to increase the risk of faecal incontinence by their very nature. Indeed one of the exclusion criteria in many studies was greater than minor faecal incontinence (Table S1). Objective scores (Wexner or St Marks incontinence score, FISI) when reported (Table 6) showed either no change or reduction in scores other than in one study where an increase was reported [47]. In this study after PPH01, the median Wexner incontinence score rose from 3.3 to 5.5 after 2-3 years. It was reported that a third of patients developed incontinence symptoms 1-4 years after the procedure.
While anatomical outcome is necessarily only a surrogate of clinical outcomes, recurrent prolapse was reported for less than half of cohorts and occurred in 4.3% of patients (95% CI: 2.0-7.3%), I 2 = 78%. Again findings were heterogeneous and there was no robust evidence that one procedure was associated with higher long term adverse outcomes than another. One RCT comparing STARR with Contour Transtar reported that at three years post-procedure 12% of STARR patients had a clinical recurrence compared with none of the Contour Transtar patients (P = 0.035) [37]. The weight of resected tissue was significantly greater in the Transtar group.

Summary evidence statements: efficacy
1 Data on efficacy were inconsistently measured and findings heterogeneous, making estimates tentative and imprecise (level IV). 2 Although inconsistent, patient global satisfaction ratings typically suggest (at least) satisfactory outcome in about 73-80% of patients (level II). 3 Although inconsistent, a reduction of 53-91% in Longo scoring system for obstructive defecation syndrome occurred in about 68-76% of patients (level II).  4 Evidence for longer term benefits persisting from procedures is sparse but suggests improvements may be maintained (level II). 5 Findings from efficacy measures provided insufficient evidence to recommend one type of procedure over another (level IV). 6 The use of Contour Transtar increases the amount of resected tissue and may reduce recurrence rates although this requires further study (level IV). 7 There is no evidence that rectal excisional procedures contribute towards post-operative faecal incontinence (level II). 8 Recurrent prolapse occurred in 4.3% of patients considering all studies (level IV); rates were lower (1.7%) in higher quality studies (level II).

Variation in outcomes by level of evidence
Meta-analytic findings are reported for each of the outcomes reported, subdivided by level of evidence (

Patient selection
Patient selection is generally perceived to be essential when choosing a surgical approach. Whilst rectal excision procedures may be efficient at correcting normal anatomy, there may be many underlying functional and organic pathologies that mean surgery is unsuccessful at 'curing' the patient [59]. It is advised when contemplating a surgical approach to exhaust first all forms of conservative or medical management. Forty-one out of 55 cohorts studied highlight the fact that all patients had undergone a period of conservative management ( Table 7). Many of the current studies reference inclusion and exclusion criteria based upon a previous consensus conference [60], namely that patients should be selected on the basis of recognized symptoms of ODS with evidence of anatomical defects on defaecography (rectocoele and/or internal rectal prolapse) and adequate anal sphincter function (at least assessed by digital rectal  (Table 8). In general when selecting patients for functional surgery it has also been advised to select patients with a high symptom load [61]. A criticism made to the authors of the European STARR Registry (n = 2224) was that the median ODS score was low (15.7 out of a possible 40) and that the range was high (1-31) [24]. The two level I RCTs reported inclusion criteria of an ODS score > 12 in one [41] and 15 in the other [37]. The majority of studies reported an inclusion criterion of rectal intussusception and/or rectocoele, (38 out of 55 study groups). Thirty-two out of the 55 study groups reported a rectocoele > 2 cm, and in 26 of these, more than 75% of patients had a rectocele > 2 cm. Thirtyseven out of the 55 study groups reported on recto-rectal or recto-anal intussusception, and in 28 of these more than 60% of patients had an intussusception. Multiple classifications were used for diagnosing an intussusception including Pescatori [62], Shorvon [63], and Oxford [64]. Other studies stated that inclusion required a rectal intussusception greater than 10 mm [41] or recto-anal intussusception > 10 mm [31]. One study reported on outcome at five years following PPH01 for patients with a symptomatic rectocoele when the authors specifically excluded intussusception [32]. This study found an 80% satisfaction rate at one year, which remained at five years. It was not possible to draw conclusions regarding efficacy according to either the presence of an intussusception (whichever grade/stage) or rectocoele size.
It is pertinent to discuss exclusion criteria as this may influence the surgeon in their choice of procedure (reinforcement/suspension/excision). Reference was again made in many studies to the consensus statement [60] and a decision making algorithm for STARR [65] when defining exclusion criteria. Specifically, external rectal prolapse, solitary rectal ulcer syndrome (SRUS), anatomical enterocele, and 'significant gynaecological pelvic floor abnormality requiring treatment' were stated as exclusions. Less than 50% of the reviewed studies excluded external rectal prolapse or other pelvic organ prolapse, and no study excluded SRUS (Table S1). Only one study graded pelvic organ prolapse (rectocoele, cystocoele, vault, uterine) with a standardized grading system e.g. POP-Q (data not shown) [26]. Whilst it is sensible to exclude patients with external rectal prolapse, no conclusion can be made regarding excluding patients with SRUS, or concomitant pelvic organ prolapse.
Anatomical enterocoele is generally regarded as a contraindication to an anterior rectal wall excision [60,65], however many studies did not exclude this (Table S1). Irrespective of this, there were no cases of iatrogenic small bowel injury reported. One study examined the outcome with respect to harms in patients with a functional enterocoele, and those without [42]. All patients had either PPH01 or Contour Transtar and outcomes with respect to harms were similar.
Summary evidence statements: patient selection 1 Patient selection, although perceived as vital in predicting outcome, is inconsistently documented and poorly informed by current evidence (level IV). 2 Patients with at least three ODS symptoms together with a rectocoele with or without an intussusception, who have failed conservative management may benefit from a rectal excisional procedure (level II). 3 It is not possible to advise on excluding patients with concomitant pelvic organ prolapse or SRUS when considering a rectal excisional procedure (level IV).

Conclusions
A systematic review of evidence for the perioperative and long term benefits and harms of rectal excisional procedures identified a modest number of high quality studies together with a larger number of observational lesser quality studies. Whilst the evidence is superior to the level of evidence for other procedures for constipation it is still rather poor in the field of medicine systematic reviews. Notwithstanding this, certain conclusions can be drawn: 1 In the presence of a rectocoele with or without an intussusception, together with at least three symptoms of ODS, a rectal excisional procedure may benefit the patient after all conservative measures have been exhausted. 2 Reliance on proving efficacy using GSR and unvalidated scoring systems (Longo) is not satisfactory and greater emphasis should be placed in future studies on the use of disease-specific and generic QoL scoring instruments. 3 It is not possible to advise which excisional technique is superior from the point of view of efficacy, perioperative variables, or harms (peri-operative or long term adverse outcomes). Future study is required. 4 Both short-and long-term harms may have been exaggerated in previous reports. Rectal excisional procedures appear to be safe with little major morbidity. Rectal urgency appears to be a problem, but findings are inconsistent because of poor reporting. 5 Despite evidence of mixed grade (IB, IIB and IV) being available for rectal excision procedures, there was little evidence that outcomes varied by grade of evidence. Recurrent prolapse was the only adverse outcome where findings varied by grade of evidence and thus may be a chance finding in the absence of any systematic pattern across outcomes.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Figure S1. Forest plot showing rates of sepsis by operation type. Figure S2. Forest plot showing rates of anastomotic dehiscence by operation type.