Article Text
Abstract
Objective: To treat anal incontinence due to obstetric external anal sphincter disruption via injection of autologous myoblast cells.
Design: Observational pilot study.
Setting: University hospital and district hospital
Patients: 10 women suffering from anal incontinence due to obstetric anal sphincter injury, refractory to conventional non-surgical therapy.
Interventions: Autologous myoblasts were cultured from a pectoralis muscle biopsy, harvested, and injected into the external anal sphincter defect using direct ultrasound guidance.
Main outcome measures: Wexner incontinence score, anal squeeze pressures, and quality of life 12 months after injection. Safety and technical feasibility.
Results: The procedure was well tolerated and no adverse events were observed. At 12 months the Wexner incontince score had decreased by a mean of 13.7 units (95% CI, −16.3 to −11.2), anal squeeze pressures were unchanged, and overall quality of life scores improved by a median of 30 points (95% CI, 25 to 42). Anal squeeze pressures did rise significantly at 1 month and 6 months post-injection (p = 0.03).
Conclusions: Injection of autologous myoblasts is safe, well tolerated, and significantly improves symptoms of anal incontinence due to obstetric anal sphincter trauma.
Statistics from Altmetric.com
Anal incontinence is common and significantly diminishes quality of life.1 In women, the commonest cause is believed to be obstetric trauma and ultrasound studies have demonstrated clinically unsuspected anal sphincter disruption (eg, a third degree tear) following vaginal delivery in up to one-third of women.2 3 4 Symptoms of anal incontinence to gas and/or solids can appear immediately after delivery or develop several years later, due to the cumulative negative effects of ageing and the menopause on the anal sphincters.5 Conservative treatments such as biofeedback and modification of stool consistency may be ineffective,6 so that ultimately surgical intervention is often required. Surgery aims to restore mechanical integrity to the sphincter ring, normally by suturing together the ends of the disrupted external anal sphincter, but symptomatic benefit may be disappointing and is often unsustained, with failure rates of up to 86% by 5 years.7 8 9 Even today, permanent colostomy is still ultimately required in some patients. Non-surgical alternatives have included injection of bulking agents such as silicone into the anal sphincter complex,10 but results have been variable, possibly because injected materials migrate, necessitating repeated treatments.11
Autologous myoblasts and fibroblasts injected into the urinary sphincter in animals and humans have been used successfully for the treatment of stress urinary incontinence since these cells become new, functioning muscular tissue following injection.12 13 14 Furthermore, using autologous cells avoids adverse events related to injection of foreign material while negating ethical issues related to the use of embryonic stem cells. For the first time, we have applied this approach to anal incontinence, administering autologous myoblasts to patients whose anal sphincters had been disrupted by obstetric trauma. We used ultrasound guidance to direct the cells into the fibrotic sphincter defect, hypothesising that the engineered cells would integrate into their surroundings and restore functionality to the damaged striated muscle, in a relatively non-invasive fashion. This article is the first description of the technical feasibility, acceptability, and clinical effects of this procedure in 10 women who were followed up for 1 year.
Methods
Patients
Participants were recruited prospectively from a pelvic-floor outpatient clinic that kept a registry of women with third- and fourth-degree obstetric tears. Consecutive women with symptoms of anal incontinence but who had thus far refused surgical treatment were contacted by telephone by the principal investigator (AF) and invited to participate following a brief description of the intended procedure. If agreeable, they were then invited to attend an outpatient clinic, at which time the procedure was explained in detail and inclusion criteria checked. No women approached refused to participate. Symptomatic inclusion criteria were severe anal incontinence to gas and solids defined as a Wexner incontinence score15 of 9 or greater. This was confirmed by incontinence diaries kept by participants for 2 weeks before definitive recruitment. A Rockwood quality of life score16 was also kept over this time. Ten women were selected and all provided written, informed consent. Their mean age was 38.3 years, range 25–66 years (table 1).
Their general health was otherwise good and proctoscopic examination was normal. All women had tried conservative treatments for anal incontinence without significant symptomatic benefit and none had undergone anal sphincter surgery previously, either for the present or previous conditions. Symptom duration ranged from 1 to 30 years (table 1).
Diagnostic procedures
Anal endosonography was performed in all women by a single female operator (AF), who was experienced in the technique, having performed at least 1500 procedures previously. A B-K Medical Scanner, Type 2101 Falcon, fitted with a 2050 endoprobe was used (B-K Medical, Herlev, Denmark). Patients were examined in the supine position using a standard technique.17 A three-dimensional (3D) dataset that encompassed the entire anal canal length was obtained.18 All women had sonographic evidence of an anterior defect of the external anal sphincter typical of obstetric injury following vaginal delivery. Defects were defined as a sonographic discontinuity of the external sphincter ring and their radial extent was noted by the investigator and described by reference to hours on a clock face. Longitudinal extent was defined by dividing the anal canal into thirds (upper, mid, and lower) and any concomitant defect of the smooth muscle internal sphincter was also noted.
Anorectal manometry was performed in all women with a water-perfused 5 mm, six-channel catheter (Samsung, Sync Master 700 TFT) and Polygram software (Synetics Medical, Enfield, UK). Manometry data were recorded at stationary 1 cm increments and also continuously, using an automatic puller.19 Pudendal nerve terminal motor latency (PNTML) was measured with a Keypoint electromyography device (Dantec, Scovlunde, Denmark) and a St. Mark’s pudendal nerve electrode (Medtronic, Shoreview, Minnesota, USA).20 Neuropathy was defined by latencies >2.3 ms (table 1). Rectal sensation and PNTML was normal in all patients.
Myoblast culture
A striated muscle biopsy approximately 1 cm3 in size was taken from the pectoralis muscle by the principal investigator (AF) under sterile operating theatre conditions and using 1% local anaesthetic applied subcutaneously in order not to disrupt the underlying muscle tissue. A 1 cm incision was made in the anterior axillary line (so that the scar could not been seen easily afterwards), the muscle biopsy excised using scissors and forceps, and placed in a prepared tube for transportation. Two hundred and fifty millilitres of venous blood for preparation of autologous serum was taken from the antecubital fossa at the same sitting. Both samples were taken immediately to a GMP production facility cooled at 2–10°C in a validated shipping container (Innovacell Biotechnologie, Innsbruck, Austria) where striated muscle was separated from any non-muscular tissue by eye using sterile scalpels, and then fragmented. The muscle was then immersed in a digestion medium containing a collagenase-1 and the cells subsequently sedimented, resuspended and plated.
Cells were stained at the time of primary culture, last passage, and harvest for CD56. Aliquots from cultures were stained with the Isotype- PE control (IgG1-PE, Ref A07796; Beckman Coulter, Brea, California, USA) and with CD56-PE (CD56-PE, Ref A07788; Beckman Coulter) and then analysed in a fluorescence activated cell sorter (EPICS XL) that detected phycoerythrin emitted fluorescence. At harvest all cultures displayed high CD56 expression (mean value 76.09% of total events; SEM 13.21). Aliquots from the culture were also grown to near confluency in small Petri dishes and subsequently washed twice with phosphate buffered saline, which was aspirated and replaced by skeletal muscle cell differentiation medium (Cambrex, Charles City, Iowa, USA). After 3–5 days all cultures demonstrated myotube formation (a unique characteristic of myoblasts) on microscopy. Petri dishes containing myotubes obtained from the myoblast fusion assay were then subjected to immunohistology for desmin staining, performed using a commercially available kit (Sigma Aldrich, Saint Louis, Missouri, USA). All cultures demonstrated high desmin expression.
At harvest (mean cultivation time 39 days, SEM 4.32) cells were washed and pelleted in a concentration of 20.16 million cells (SEM 1.91)/ml medium with 10% autologous serum. At this time aliquots of the cells were checked for foci formation with soft agar assays and for chromosomal aberrations with standard karyotyping in order to exclude malignancy. Viability was assessed by fluorescence activated cell sorting based on 7AAD staining: All cultures showed high viability (mean 98.14% (SEM 1.48)). An aliquot of suspension was also transferred to a packed cell volume measurement tube (TPP, Trasadingen, Switzerland) and centrifuged for 7 min at 12 413 g to determine packed cell volume. Cell counts via a Neubauer chamber were correlated with packed cell volume.
All procedures were carried out under clean room conditions by a GMP certified laboratory. In process sterility testing and sterility testing on the cell culture supernatant at harvest were performed according to the European Pharmacopoeia and were negative on all cultures.
Patient preparation and myoblast injection
Because animal21 and human22 studies have shown that electrical stimulation enhances integration of myoblasts into host tissue, all 10 women underwent an anal electrical stimulation programme before implantation via an anal plug fitted with two ring electrodes. Batteries were contained within an external handset, which also turned the unit on and off and set the stimulation time and electrical frequency. Women were asked to select the maximum frequency that did not cause discomfort, and to use the unit for approximately 15 min each day for 10 weeks prior to implantation (and also 28 days after). Some patients noted an itching sensation in the anus or sometimes the perineum during stimulation but the procedure was generally well-tolerated by all subjects. To exclude a therapeutic effect based on electrical stimulation alone, patient work-up was repeated immediately before myoblast injection.
Myoblast injection was performed by the principal investigator (AF) with the patient supine and under general anaesthesia to avoid movement during injection. Following a single dose of intravenous antibiotics, myoblasts were injected under direct ultrasound guidance using a 2202 scanner and type 8848 10 MHz transducer with a specifically designed injection device (B-K Medical, Herlev, Denmark). The transducer allowed simultaneous biplane scanning so that implantation could be visualised in a 65 mm +30° trapezoid sagittal plane while simultaneously viewing needle ingress in a 180° transverse plane. The tip of the transducer was lubricated with sterile ultrasound gel, covered with a sterile condom, lubricated again and inserted into the anal canal. The entire anal canal length was scanned before injection and the defect in the external anal sphincter visualised sonographically. Individual injections of 0.5 ml were extended in a semicircular array directly into the external anal sphincter ends and the scar tissue lying between these. In most cases the anterior arc covered was approximately 9 o’clock to 3 o’clock. Care was taken not to inject into the longitudinal muscle, internal anal sphincter or subepithelium. Twelve to fourteen individual injections were administered to each patient. The injection procedure took 25 min in the first patient but diminished to an average of 11 min for the last seven women.
Outcome measures
All women continued their pre-operative symptom diaries post-operatively and noted the number of daily bowel movements, weekly episodes of incontinence to stool, Wexner Continence Grading Score,15 and the Rockwood Faecal Incontinence Quality of Life Scale16 for 1 year post-injection. At 1, 6 and 12 months following injection, anal endosonography and anorectal physiology procedures were repeated by the principal investigator in all women and data recorded for comparison with pre-operative values. The ultrasound data from pre-implantation, and 1, 6 and 12 months post-implantation were presented to the original operator (AF) on a PC display by one of the co-investigators (PF). Patient and scan order were randomised and the date of the scan removed from the image. Sonographic thickness of the external sphincter, internal sphincter, and longitudinal layer was then measured at 3, 6, and 9 o’clock and averaged.
We were also concerned with safety, acceptability and technical feasibility. Discomfort following the procedure was assessed using visual analogue scales and the biopsy and injection sites were inspected for haematoma or infection. Adverse events following general anaesthetic were sought for as per usual practice. Venous blood was taken pre- and post-procedure to monitor any deviation from normal parameters.
Data analysis
Our main objective was to compare measured parameters pre-injection and at 12 months post-injection. Because there was no significant improvement in any parameter when measured after electrical stimulation (but before implantation) the pre- and post-stimulation values were averaged for each parameter and this single pre-implantation value was used in all subsequent comparisons. We also investigated the change in parameters over the time of the study, pre-implantation and at 1, 6 and 12 months post-implantation. Examination of the differences in values over time indicated that all excepting the quality of life scores were approximately normally distributed, and change over time was investigated using repeated measures ANOVA. Quality of life scores over time were evaluated using Friedman’s test. Statistical significance was assigned at the 5% level, and 95% confidence intervals calculated. An independent statistician performed the analysis.
Results
Adverse events
All women recorded a visual analogue scale pain value of 0 the day following the procedure and all were discharged home that day. There were no subsequent complications in any subject. All post-operative blood samples were normal.
Symptoms of anal incontinence
A comparison of parameters measured pre-implantation and at 1 year is shown in table 2, together with an analysis of change in parameters over the time course of the study.
These data indicate a significant decrease in the Wexner score and the number of bowel movements per day, both at 1 year and over the time course of the study. The Wexner score decreased by a mean of 14 units (p<0.001, fig 1), while the number of bowel movements decreased by an average of 0.4 units (p = 0.02). We correlated the pre-treatment Wexner incontinence score with both mean and maximum squeeze and resting pressures pre-implantation and also at 12 months post-implantation but could find no significant association.
Anorectal physiology
There was no significant change in mean or maximum anal squeeze pressures at 12 months (table 2). Both maximum and mean resting pressure were significantly decreased at 1 year (table 2). There was a reduction of 6 units over the year for mean resting pressure, while there was a reduction of over 7 units for maximum resting pressure. There was no change in the anal pressure zone. When the change in values over time was investigated, maximum anal squeeze pressure increased significantly over the study period, notably at the 1- and 6-month assessments, but had returned to baseline at 12 months (table 2, fig 2). The significant decrease in mean and maximal anal resting pressure appeared to occur between the 6- and 12-month assessment (table 2, fig 3). There was no strong evidence of an overall effect of time upon the remaining variables.
Quality of life
A comparison of overall and constituent quality of life scores measured pre-implantation and at 1 year is shown in table 3, together with an analysis of change in parameters over the time course of the study.
These data indicated there was a significant change in the majority of variables both at 1 year and over time. Overall quality of life was significantly increased at 12 months by a median of 30 units (p = 0.005, fig 4). All individual components also increased over time, excepting the scores for depression, which showed no significant change. There was also a significant change in episodes of incontinence, with a decrease in the values from pre-treatment to 12 months post-treatment. The median reduction was eight episodes per week over this time (p = 0.005). At their 1 year follow-up, all patients stated that their symptoms had improved and considered the treatment very successful.
Ultrasound data
Endosonographic measurements of anal sphincter layers pre-implantation and at 1, 6 and 12 months post-implantation are shown in table 4.
There was no significant change in the thickness of any anal sphincter component over the course of the study. At the time of performing the scan, the sonographer did have a subjective impression of slightly increased echogenicity 1 month post-injection in the region of the injection site, with some blurring of adjacent tissue boundaries, which was hypothesised to represent possibly a sequel to needling and injection (fig 5).
Discussion
The results of conventional treatment for severe anal incontinence are often disappointing. Anterior overlapping external sphincter repair is the standard surgical therapy but a review of this procedure at a median of 77 months in 47 women with obstetric trauma concluded that continence rarely returns to normal, with many residual symptoms, and that results deteriorate with time.8 Another evaluation of 86 patients similarly found that over half still had symptoms at 3 months; only 21 were continent at 40 months.9 A long-term study of 62 patients found that only 45% were satisfied by surgery.23 Less invasive procedures have been attempted, notably perianal injection of various bulking agents. However, injection of foreign material may induce granuloma formation24 whereas some autologous materials, such as fat, may embolise.25 Perianal silicone is perhaps the most popular agent, but long-term data are lacking. A recent publication of six patients with a median follow-up of 61 months found variable clinical benefit.11
Embryonal stem cells have the potential to differentiate into a variety of different mature cells, but their use has attracted ethical debate. The use of autologous cells avoids adverse events related to foreign material and bypasses ethical concerns. Injection of autologous cells has recently been used to treat stress urinary incontinence in women. A study of 20 women experiencing stress urinary incontinence found that 18 were continent 1 year following imjection.12 Other authors described eight women, five of whom were improved at 1 year, with improvement onset between 3 and 8 months.26 It seemed natural to us to attempt this procedure in the anus. For the first time, we describe the technical feasibility and clinical effects of autologous myoblast injection to treat anal incontinence due to obstetric trauma. Injection was very well tolerated by all 10 volunteers and we could find no evidence of immediate or delayed side effects. Because the procedure is novel, we were cautious and performed injection under general anaesthesia and observed patients overnight. Over the course of the study, the mean injection time fell to 11 min and the very low pain scores achieved led us to conclude that implantation will ultimately be performed as a day-case, comparable to prostate biopsy for example.
We were able to demonstrate a highly significant reduction in incontinence scores and bowel movements over the study period, and quality of life improved markedly. Faecal incontinence causes considerable anxiety and patients may severely restrict their activities accordingly.1 An effective treatment should therefore be reflected by an improved quality of life. We used a sensitive and previously validated quality of life instrument that has been designed specifically to address the personal implications of faecal incontinence.16 Following myoblast injection we observed significant and sustained improvements in the quality of life experienced by participants, both overall and for individual components of lifestyle, coping and embarrassment.
We were unable, however, to demonstrate any sustained physiological change to account for the improvements we observed. We hypothesised that the implanted myoblasts would integrate into the damaged external sphincter and improve its functional integrity. The external anal sphincter is under voluntary control, composed of striated muscle, and its function is reflected by anal canal mean and maximal squeeze pressures.26 We found significant improvements in external sphincter function at 1 and 6 months but these had returned to baseline by 1 year. In contrast, the internal anal sphincter is composed of smooth muscle and is involuntary, functional activity being reflected by anal canal mean and maximum resting pressures.27 We expected to find no improvement in these pressures; indeed resting pressures had fallen significantly by the time of the 1 year assessment.
Our study does have limitations. In particular, the discrepancy between markedly improved symptoms/quality of life and an unsustained improvement in external sphincter pressures raise the possibility of a placebo effect. However, this is a small pilot study and the effect size will need to be relatively large in order to reach significance. We were able to demonstrate improved external sphincter squeeze pressures at 1 and 6 months following implantation and it is possible that the quality of life scores reflected a persistent reaction to this. Patients were also aware that they were in an experimental study of a novel therapy for anal incontinence and it is possible that this alone might result in an improved quality of life. A randomised controlled study that used sham injection, of saline for example, would account for any placebo effect and also allow us to investigate the effect that needling and injection alone have on the anal tissues. However, we believe that further studies in human subject should be preceded by bench studies that investigate the ultimate distribution of injected cells, and whether they remain viable. Preclinical investigations in the urinary sphincter suggest that injection of autologous myoblasts leads to cell engraftment but reported survival rates for the implanted cells vary, possibly due to different implantation modalities and subsequent analytical methods. For example, it is presently unclear whether the myoblasts establish functional connections with nerves. Indeed, in patients where low external sphincter pressures are due to neuropathy in addition to laceration, it is unclear whether innervation can ever be achieved. It is possible that raised external sphincter pressures at 1 month were due to the bulking effects of the injection (6–7 ml were administered), but the observed benefits at 6 months due to this are unlikely because only 0.33 ml of solid cell mass was administered and the suspension would have been resorbed by this time. We used electrical stimulation before and after myoblast injection because stimulation appears to enhance integration and physiological adaptation of the implanted cells.21 22 We found that electrical stimulation alone neither reduced incontinence episodes nor improved anal sphincter pressures, again implicating the implanted cell as the cause of symptomatic improvement. A recent article described injection of bone marrow-derived mesenchymal stem cells into the anal sphincters of rats that had been subject to sphincterotomy.28 Histological examination at 30 days showed new muscle fibres and sphincter contractility was improved when compared to those rats that had undergone saline injection.28 A further study of cryo-injured rats injected with muscle-derived stem cells found that while contractility was not significantly different from uninjected controls, the authors were able to identify differentiated muscle at the injection site.29 Clearly, many issues relating to cell survival, distribution, integration and functional contribution need further study.13
In summary, we have used injection of autologous myoblasts to significantly improve symptoms of anal incontinence due to obstetric trauma. We found the procedure feasible and safe, without significant side effects. If the beneficial results we observed are sustained in controlled and dose-finding studies, then this procedure holds considerable promise as a relatively non-invasive treatment for a common and distressing clinical problem.
Acknowledgments
The authors are grateful to P Bassett for the statistical analysis.
REFERENCES
Footnotes
The views expressed in this publication are those of the authors and not necessarily those of the UK Department of Health.
Funding This study was funded by the Medical University of Graz (Investgator time and clinical facilities for implantation) and University College London (Investigator time). Innovacell Biotechnology cultured the myoblast cells free of charge but had no role in the study design, patient recruitment, data collection, data analysis, data interpretation, or writing of the manuscript. Statistical analysis was performed by an independent biostatistician who is not employed by the corporate entity. The guarantors (AF and SH) had full access to all study data and had final responsibility for the decision to submit for publication. Additionally, a proportion of this work was undertaken at UCLH/UCL, which receives funding from the UK Department of Health’s NIHR Comprehensive Biomedical Research Centre funding scheme.
Competing interests WS is an employee of Innovacell Biotechnology, Innsbruck, Austria, who provided the autologous myoblasts described in the manuscript. Innovacell had no role in the study design, patient recruitment, data collection, data analysis, data interpretation, or manuscript writing. The other authors have no conflicts of interest to declare.
Ethics approval The study protocol was approved by the Ethics Committee of the Medical Universities of Innsbruck and Graz, and registered with the European Union Drug Regulatory Authorities: EudraCT-Nr.2005-003759-11.
Provenance and Peer review Not commissioned; externally peer reviewed.