Surgery for constipation: systematic review and practice recommendations

To assess the outcomes of rectal suspension procedures (forms of rectopexy) in adults with chronic constipation.


Introduction Background and procedural variations
Constipation, in a proportion of patients and in the broad sense of the term, is related to an inability to evacuate the rectum. This obstructed defaecation or rectal evacuation disorder 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 proctographic evidence of a rectocoele and/or intussusception. These anatomical variants are considered to cause 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]. It follows that clinical resolution of symptoms could be achieved by restoration of normal anatomy by surgery. Resuspension of the rectum aims to hitch the prolapsing or redundant rectal wall thus straightening the intussusception and/or effacing the rectocoele. This concept while anatomically rational remains clinically controversial for a number of reasons. First, such anatomical variants are common and are often found in healthy individuals with no symptoms of obstructed defaecation [2]. Secondly, resuspension operations when employed to patients with full thickness rectal prolapse, may themselves cause increasing constipatory symptoms [3]. Such procedures include posterior rectopexy [4]. The potential for worsening constipation is thought to relate to fibrosis caused by insertion of foreign material or mobilization of the lateral ligaments of the rectum. These ligaments contain nerves to the rectal wall and the resultant denervation may be the cause. In the process of developing alternative resuspending procedures, surgeons have attempted to limit the effect of the foreign material by using sutures only [5], added a resection of the sigmoid colon to the rectopexy [6][7][8] or more recently, limiting the dissection of the rectum to the ventral surface by supporting the rectum with mesh [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23]. In addition, laparoscopy has become the favoured approach procedurally, not only allowing a more rapid recovery but also easing access to, and visibility in the pelvis.

Scope
The purpose of the overall CapaCiTY review process is to assess the efficacy and harms of surgical procedures for chronic constipation in adults. Thus, the aim of this review is to assess the outcomes of rectal suspension procedures in adults presenting with chronic constipation symptoms. In effect, this is however limited to patients with obstructed defaecation and internal prolapse (intussusception). Procedures considered beyond the scope of systematic review included rectal excisional procedures, e.g. STARR [9], rectal reinforcement procedures, e.g. transanal/transperineal repair of rectocele [10], and uncommon variant of suspension procedures, e.g. laparoscopic promonto-fixation [11]. 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 [12][13][14][15][16][17][18][19], mixed indications including numerous pelvic floor abnormalities [20] or limited postoperative outcomes [21].

Perioperative data
Perioperative data were reported by all 18 studies ( Table 2). Reporting of procedure duration was inconsistent but median procedural duration for laparoscopic ventral mesh rectopexy (LVMR) was 159 (range 75-198) min; for robotic ventral mesh rectopexy (RVMR), 205 (range 191-218) min; for laparoscopic resection rectopexy (LRR), 123 min (one study) [45,46]. Although robotic procedures appeared to take longer, substantial non-reporting of other procedures precluded a clear finding. The two papers on RVMR were from the same centre. It is interesting to note a decrease in duration of operation, which may indicate a learning curve. Conversion to laparotomy was rare (median 2%, range 0-8%) ( Table 2), with the most common reason being adhesions. The median length of stay (LOS) was similar for procedures: LVMR, median 3.3 (range 1.0-7.1) days; RVMR, median 4.3 (range 4.0-4.6) days (data from one centre via two reports) [45,46]; LRR, 4 days (data from one study) [7]. LOS possibly reflected local policy rather than clinical need, since day case procedures have been shown to be feasible [59,60]. The reason to keep patients in hospital for up to 1 week was not documented. Only one paper commented on LOS after open rectopexy (OR) (8.5 days) [54].
Summary evidence statements: perioperative data 1 Procedures are reported to take from 1.5 to 3.5 h, with consequent typical LOS of 4-5 days (level IV). 2 There was no clear variation between procedures in perioperative measures, although non-reporting by studies may have concealed differences (level IV).

Harms
There was a considerable heterogeneity in surgical morbidity reported as well as in overall procedural  Table 2 Perioperative data by procedure.  complication rates (Fig. 2), with individual study rates varying from 0.0% to 23.5% (Table 2). Such heterogeneity may reflect different inclusion, thresholds or conventions for recording complications. Complications typically occurred in about 5-15% of patients. Pooled findings suggest that LRR might be associated with higher morbidity (total complications 15% for LRR vs 10% LVMR) although the findings were not statistically significant (Z-test, P = 0.30), and absolute patient numbers were small for LRR. The majority of complications were minor and included urinary tract infections (the most common reported), wound infections, haematoma formation, persistent pain and urinary retention. There were some more serious complications including port-site hernia, small bowel obstruction (usually after conversion but also related to mesh or suture adhesions), osteomyelitis and bladder injury (often when associated to bladder prolapse surgery). Specific mesh complication rates were rare, with only five occurrences after 939 procedures (0.53%). Overall, procedures were safe: conversion to laparotomy was rare (median 2%, range 0-8%) ( Table 2), with the most common reason being adhesions; stoma was only reported in one study; no perioperative deaths were reported. Two open rectopexy procedures (posterior mesh) were described, but data concerning post-operative complications were limited. There was no mortality recorded after any resuspension procedures.

Summary evidence statements: harms
1 Data on harms were inconsistently reported and heterogeneous, making estimates of harm tentative and imprecise (level IV).   [48], Cleveland Clinic Incontinence score [46,49,56], Faecal Incontinence Severity Index (FISI) [47,48,[51][52][53][56][57][58] and St Marks Incontinence score [48]. Global 'success' or 'satisfaction' ratings (GSR) were obtained via a variety of methods in 7 studies (where 'satisfied' or 'very satisfied', 'good', 'very good', and 'excellent' were interpreted as a positive outcome or overall improvement). 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 blinded to intervention status. Average reported studies follow-up was 31 months (range 12-72 months).
Accepting these methodological limitations, several reports assert that most patients undergoing rectal suspension procedures were satisfied. Meta-analysis of studies reporting a summary measure found considerable heterogeneity, which may reflect variation in measurements, patients or procedures. Overall improvement (a good or satisfactory outcome) was reported in 83% (95%CI: 74-91%, I 2 = 77%) of cases, based on 328 patients (Table 3; Fig. 3). Similar levels of improvement were recorded for LVMR and OR; only one small study reported improvement after LRR, and data were not available for RVMR.
The initial aim of 'suspension' procedures is to treat symptoms. Functional assessment of constipation is therefore the most important outcome. However, many patients also suffer from incontinence, typically post-defaecatory seepage. The various scoring instruments and functional outcomes employed are reported in Table 4. Generally, measures are too sparsely reported to be informative. For LVMR, Cleveland Clinic Constipation score improved from a median of 14 (range 7-18) to a median of 5 (range 4-7) in 6 studies providing pre-and post-operative data. Improvement in constipation was highly heterogeneous and only reported in a minority of studies, varying from 20% to 97%. By pooling data for LVMR, the reported improvement in constipation was 86% (95%CI: 20-97%).
While the clinical outcome has primacy, the most immediate visible consequence of surgery is to correct anatomy. Therefore, an assessment of anatomical recurrence is also important (although necessarily representing only a surrogate outcome). Anatomical recurrence rates varied between 0 to 21% (Fig. 4), but typically occurred in 2-7% of patients in most studies. Functional outcome data on robotic surgery and LRR were rarely available, but again anatomical correction was very likely achieved with both procedures. No conclusions about functional or anatomical outcomes could be made for the other rectopexy procedures.

Patient selection
Patient selection is perceived by many experts as extremely important when choosing the surgical approach. Whilst these procedures may be efficient at correcting normal anatomy (median 95%, range 79-100%), many underlying functional and organic pathologies may jeopardize the success of surgery in the attempt of 'curing' the patient [61]. Fifteen of 18 papers highlight the fact that all patients had undergone a period of conservative management. Other than this common feature, selection was inconsistent. Even the diagnosis of abnormal anatomy varied throughout the literature. Studies described interventions for patients with: ungraded intussusception [7,54]; 'rectoanal' intussusception [6,47]; 'high grade' intussusception [57]; 'grade 3 or 4 0 intussusception [48,49,51,53,56]; 'anterior or circumferential' intussusception [4]; rectocoele +/intussusception [52,54,58] or +/cystocoele [13]; complex rectocoele of above 2-3 cm [50]; multi-compartment pelvic floor disorders [46]; solitary rectal ulcer syndrome (SRUS) [34,55]. Thus, it was difficult to draw any conclusions as to which group could benefit from intervention. When summarising the data, the most common theme regarding patient selection is a high grade intussusception (i.e. rectoanal or Oxford grade ≥ 3). Table 5 lists the papers where this inclusion criterion has been adopted and one of the primary indications along with a summary of the outcome measures reported (if given in more than one paper). The conclusions from this sub-analysis resemble those described in the whole review. SRUS deserves specific mention as two papers included patients specifically diagnosed with this condition [34,55]. Patients report passage of mucus and bloody liquid on defaecation, with an ulcer seen within the rectum. Treatment is conservative, initially using biofeedback and behavioral intervention. A proportion of patients present an element of internal intussusception, which may reflect the ulcerated area as the apex of the intussusception, repetitively traumatised with straining. The surgical correction of a prolapse (when    2 One common indication appears to be high grade rectal intussusception (level IV). 3 For high grade intussusception, LVMR, RVMR and resection rectopexy typically correct anatomy in about 80-100% of cases (level IV). 4 If SRUS is associated with prolapse, a LVMR typically results in healing of the ulcer in around 80% of patients (level IV).

Discussion
A systematic review of evidence for the perioperative and long terms benefits and harms of rectal suspension procedures identified no high quality studies. The evidence base is characterised by observational studies of variable and often uncertain methodological quality. Definitions are poor, e.g. grading of complications was inconsistent. Future studies should provide robust and comparative evidence for clinicians to support patient decision making, in terms both of the incremental benefits and harms of suspension procedures. A Clavien-Dindo (or equivalent) classification is essential. Greater understanding is required of the mediating effects of prognostic factors particularly preoperative definition of both functional and radiological parameters that impact upon treatment success. Relevant to future research would be to define a minimum set of outcomes for reporting future studies. Finally, and most obviously, the evidence base requires urgent augmentation with some high quality studies focused on having at least one well powered randomized controlled trial to inform future clinical decision making.
9 Biviano I et al. Comparative outcome of stapled trans-anal