Laparoscopic surgery has been used over the past decade with resection of the colon and rectum for benign and malignant conditions. The technical limitations of specimen retrieval and bowel anastomosis, and port-site recurrence in cancer surgery have limited the impact of laparoscopy on colorectal surgery. This is despite numerous early case studies and small randomized controlled trials combined with ongoing multicentre trials1,2. Full-thickness rectal prolapse repaired by fixation rectopexy is well suited to a minimally invasive laparoscopic approach as no specimen is removed and no anastomosis is necessary37. An initial historical controlled trial suggested that laparoscopic rectopexy was superior to open rectopexy with respect to pain and return to normal physiological function3.
The objectives of this study were to determine whether a laparoscopic approach to rectopexy has advantages over the traditional open approach in either subjective clinical outcome or objective stress response. The study was designed to assess large differences in perioperative outcomes (using a small sample size with good internal validity and standardization of therapy) rather than any small differences in longer-term outcomes such as recurrent prolapse, postoperative constipation and faecal incontinence. To assess these endpoints adequately would require a multicentre study (difficult in terms of feasibility, poor internal but good external validity)8. Differences in event rates with respect to these were assessed in the present study but no differences were detected since the commencement of laparoscopic rectopexy in 1994.
| Patients and methods
From December 1996 to December 1999 all patients with full-thickness rectal prolapse with or without faecal incontinence referred to one teaching hospital department of colorectal surgery were assessed for eligibility. Patients were excluded if concomitant gynaecological procedures were planned, rectopexy had been performed previously or a large irreducible prolapse was present which would be better repaired by perineal proctosigmoidectomy. Patients with obstructive defaecation and concomitant rectocele or anal mucosal prolapse were offered a Delorme procedure. Patients with intractable constipation preceding the development of the rectal prolapse were advised to have a resection rectopexy. Constipation developing in conjunction with the rectal prolapse was not an exclusion criterion nor was previous lower abdominal surgery. Institutional ethics committee approval was obtained for this trial.
The surgical technique has been described previously; however, modifications in the staple driver, and improvement in bowel grasping forceps and surgical technique have reduced the size and number of operating ports from five to three3. Standardization of the surgical technique was agreed and achieved largely as 85 per cent of enrolled patients were under the care of one surgeon (M.J.S.). Full rectal mobilization to the anorectal junction was performed with a posterior incomplete mesh wrap. The mesh was secured to the sacral promontory with a single spiked chromium staple (12–8720; Smith & Nephew Richards, Memphis, Tennessee, USA). The mesh was secured to the lateral rectum using hernia staples3.
Subjective clinical outcomes
Patients entering the trial agreed to a standardized clinical pathway (CP) to minimize patient or surgeon bias in favour of either group. The clinical pathway had four subjective outcomes (CP1–4) per patient that were documented by a blinded assessor. They included oral fluids (CP1) and full mobilization (CP2) on day 1 and, if tolerated, a light diet (CP3) on day 2. The patients themselves determined the outcome variables, not the clinicians. The urinary catheter was left in for 48 h to measure total catecholamine excretion and removed on the morning of day 3. Patients were advised that they would be discharged after day 3 but before the end of day 5 (CP4). A specific clinical reason was required for discharge after this date.
Patients were assessed by three gastrointestinal clinical nurse specialists who were trained to assess the trial outcomes before the trial and were independent of the colorectal ward. The assessors were blinded to the operative procedure and the patient was instructed not to inform the assessors. To enhance blinding the clinical notes were marked only as 'laparoscopic rectopexy trial patients'. Patients were assessed by the blinded assessor before operation, at 4, 24 and 48 h after operation, and then daily until discharge. Visual analogue scales for pain and mobility, patient-controlled morphine requirements911, respiratory function12, oral intake and morbidity were assessed at each visit and blood samples were taken.
Twenty-four-hour urine collections were performed to measure catecholamine excretion (adrenaline and noradrenaline)13. Preadmission serum tests were performed for fasting cortisol, interleukin 6, C-reactive protein and erythrocyte sedimentation rate (ESR)1315. These were repeated 4, 24 and 48 h after operation. Serum cortisol and interleukin 6 are representative of the acute-phase response (4–24 h), and ESR and C-reactive protein of a delayed stress response (48 h). The respiratory function tests comprised measurement of forced expired volume, vital capacity, and inspiratory and expiratory wedge pressures12. The latter two tests best assess the ventilation of the lung bases. The results were measured as change from the individual patient's preoperative values.
Theatre utilization times were measured prospectively by the research assistant. Two measurements were made: the time from entering the anaesthetic bay to leaving the theatre and the time from the first incision to the last stitch. The 30-day morbidity and mortality rates were recorded prospectively for both groups by the blinded assessor. All potential morbidities (for example fever) were documented. A final determination and definition of morbidities was performed at the conclusion of the study by a panel of clinicians blinded to the operative group. Morbidities were classified as all documented events (for example fever unknown cause), all definable morbidities (for example urinary tract infection) and major morbidities (defined as those requiring active intervention and delaying discharge, such as pulmonary oedema and pneumonia).
Statistical analysis was performed by a statistician to whom the operative group was identified only as group A or B. Unblinding was performed only after completion of all analyses. Analysis was performed using the Wilcoxon rank sum test, the Student t test for continuous variables, and ?2 test and Fisher's exact test (less than five in one or more cells) for proportions. Sample size calculations were based on the power to detect only clinically significant differences of subjective outcomes (i.e. proportions ranging from 20 to 40 per cent, a = 0·05, ß = 0·80, sample size 17–24 in each group)16.
Of 101 patients with a full-thickness rectal prolapse seen during the accrual of the study, 40 were randomized to the trial. Only three patients deemed by the surgeon to be suitable for rectopexy refused the trial. Of the 40 patients, one patient referred from an outside institution who was randomized to open rectopexy then refused any surgery and was excluded from the analysis. One patient with a full understanding of the randomization process refused open surgery after randomization into this group. The trial management committee met and elected to perform a laparoscopic rectopexy and include the clinical data in the open group (intention-to-treat analysis).
Thus 39 of the 40 patients randomized completed the trial. No patient required intraoperative conversion from a laparoscopic to an open procedure. There were no significant differences in duration of prolapse, previous abdominal surgery, age, gender, duration or severity of incontinence, or preoperative constipation scores in the two cohorts. The mean total operating room time and surgical time were just over 50 min longer for the laparoscopic group (P < 0·01) (Table 1). The mean day of discharge following operation was significantly earlier in the laparoscopic group (3·9 versus 6·6; P < 0·01). After a mean follow-up of 24 months there has been one recurrence in the open group and none in the laparoscopic group.
In the laparoscopic group 75 per cent of the clinical objectives of early recovery (CP1–4) were fulfilled compared with only 37 per cent in the open group (60 of 80 versus 28 of 76; P < 0·01) (Table 2). Nineteen of 20 patients in the laparoscopic group were discharged by day 5 compared with nine of 19 in the open group (P < 0·01). Sixteen patients in the laparoscopic group tolerated solid diet by day 2 compared with only three in the open group (P < 0·01), even though both were offered fluids and diet at the same time. More patients in the laparoscopic group tolerated fluids on day 1 (17 of 20 versus 12 of 19) and were fully mobile on the first postoperative day (eight of 20 versus four of 19) but these results were not statistically different.
The objective measures of stress response to surgery all favoured the laparoscopic group. The haemoglobin level at 48 h had fallen more in the open group (1·2 versus 1·9 g per cent; P = 0·02). The increase in the acute-phase reactant interleukin 6 was greater but not significantly different in the open group at 4 h (11 versus 111 pg/ml) and 24 h (21 versus 44 pg/ml), and equivalent at 48 h (13 versus 12 pg/ml). The increase in serum cortisol was significantly greater in the open group at 4 h (343 versus 582 nmol/l; P = 0·03), 24 h (79 versus 387 nmol/l; P < 0·01) and at 48 h (24 versus 437 nmol/l; P = 0·03). Similarly, significant differences were found in C-reactive protein at 24 h (44 versus 77 mg/l; P = 0·03) and 48 h (70 versus 112 mg/ml; P = 0·05) favouring the laparoscopic group. There were no significant differences in ESR at 4, 24 or 48 h.
The amount of catecholamine recovered in the 24-h urine collection was significantly lower in the laparoscopic group (adrenaline 38·1 versus 72·7 nmol/l (P < 0·01); noradrenaline 118 versus 281 nmol/l (P < 0·01)). No differences were found between the two groups in respiratory function over the first 48 h.
There were significant differences in morbidity between the two groups. All potential adverse events occurred more often in the open group (six of 20 versus 14 of 19; P < 0·01). Defined morbidities (three of 20 versus nine of 19; P = 0·03) and major morbidities (none of 20 versus four of 19; P < 0·05) were significantly fewer in the laparoscopic group. The major morbidities in the open group were cardiorespiratory. The only defined morbidities in the laparoscopic group were atelectasis (one), urine retention (one) and wound infection (one).
Morphine requirements were lower in the laparoscopic group at each of the points of measurement but the differences were not significant. However, the mean(s.d.) total dose of morphine over 5 days was significantly less in the laparoscopic group than in the open group (73(48) versus 113(94) mg; P = 0·03). Significant differences in mean scores in favour of the laparoscopic group were recorded in nine of 16 analogue scores of pain and mobility with trends to significance in a further three (total 12 of 16). All were in favour of the laparoscopic group.
There were significant improvements in continence scores in each group but no differences between the groups (Table 3). There was no significant increase in constipation scores in either group and no difference between the groups. The mean visual analogue score for cosmesis was better in the laparoscopic group but not significantly so (0·8 versus 2·5; P = 0·10).
This small randomized clinical trial has shown that a rectopexy performed laparoscopically results in less postoperative pain, an earlier return of alimentation and earlier discharge from hospital, at the expense of a longer operation. Blinded assessors and an agreed clinical pathway were used to improve the quality of subjective data. The subjective outcomes in this study coincide with the objective measures of stress response which adds weight to their validity. Apart from operating time there was no subjective or objective measure favouring the open group. That so many results between the two groups were significant with the small sample size suggests a large group difference.
There are no large randomized controlled trials comparing different operations for rectal prolapse and only two published trials have compared laparoscopically assisted colorectal procedures. The first trial, by Milsom et al.1, demonstrated an improved return of bowel function and respiratory function in the laparoscopic group but no decrease in the length of stay (6 versus 7 days). They also found a 75-min increase in operating time1. The second trial, by Stage et al.2, demonstrated no difference in respiratory function, but decreased pain scores and morphine requirements. They also noted a smaller objective stress response in terms of interleukin 6 and C-reactive protein as well as a significantly shorter length of stay (5 versus 8 days), but an increased operating time of 55 min2. A third non-randomized controlled trial that assessed interleukin 6 and cortisol stress responses in laparoscopic 'assisted' colorectal resections demonstrated a similar lowering effect of laparoscopy on interleukin but not serum cortisol levels15.
The data from this randomized controlled trial confirm and augment the differences detected in an original historical controlled trial3. With the introduction of a clinical pathway in the current study early alimentation and discharge have improved in both groups compared with the historical controls. None the less there is a relative difference strongly in favour of the laparoscopic group. The difference in operating time has been reduced from 70 min in the previous study to 50 min in the current study. Including this trial, the evidence to date from controlled trials suggests that there is a benefit in terms of the immediate objective stress response but no immediate objective respiratory difference. Combined with the decreased stress response, less ileus, less pain, greater mobility and a decreased morphine requirement, there is less respiratory morbidity and an earlier discharge for patients treated laparoscopically. The increased theatre utilization (50 min) and disposable costs will most likely be offset by both the shorter hospital stay and less need for treatment of minor and major complications. A cost-effectiveness analysis will be the focus of a further evaluation of the data collected from this study.
The limitations of this study include the small numbers of patients and the fact that the technique was performed by one specialized surgical unit and largely by one surgeon. The wide application of the technique still remains to be tested. However, when abdominal rectopexy is chosen as the procedure to correct rectal prolapse, a laparoscopic approach has short-term benefit in terms of return to normal diet and mobility, earlier discharge from hospital and less morbidity. These findings are paralleled by a reduced neuroendocrine and immunological stress response. No long-term differences in constipation, recurrent prolapse or improvement in continence scores between open and laparoscopic approaches have been identified. Interestingly, these data suggest that there was underlying constipation before operation that was not increased significantly by either rectopexy method (Table 3). This has brought into question the suggestion from retrospective studies that the major drawback to rectopexy compared with alternative approaches is an increase in constipation.
While long-term studies are required, the data indicate that a laparoscopic technique may be beneficial to patients with rectal prolapse for whom rectopexy is proposed.