Conservative treatment for urinary incontinence in Men After Prostate Surgery (MAPS): two parallel randomised controlled trials

Urinary incontinence after prostate surgery and effect on well-being

Urinary incontinence is a debilitating condition that can be an iatrogenic consequence of prostate surgery. 1 The effect of urinary incontinence on quality of life can be profound. The economic costs can be personal (such as the need to use pads or devices and the deleterious effect on quality of life) and societal (use of health services and the need for residential or nursing home care). The effect on quality of life of urinary incontinence is greater than that of erectile dysfunction, another possible iatrogenic consequence of prostate surgery. 2

Continence mechanisms in men

Urinary continence in men is achieved by the interaction of anatomical structures (bladder, urinary sphincter, urethra and the pelvic floor muscles; Figure 1) and neurological control. Continence is maintained by contraction of the sphincter and pelvic floor muscles and relaxation of the bladder muscle (detrusor muscle), while controlled, appropriate urination requires the relaxation of the sphincter and pelvic floor muscles at the same time as the bladder muscle contracts. This process is under neurological control. Failure in either muscle or neurological function or both will result in incontinence or urinary retention.

FIGURE 1.

Anatomy of the pelvic organs in men (see Appendix 4.3).

The male urinary sphincter may be divided into two functionally separate units, the proximal sphincter (nearest the bladder, consisting of the bladder neck, prostate and the portion of the urethra that passes through the prostate) and the distal sphincter (further away from the bladder, just below the prostate at the level of the pelvic floor muscles). The pelvic floor muscles contribute to the ability of the distal sphincter to keep the urethra closed.

Radical prostatectomy physically disrupts the integrity of both the muscles and the nerves, thus resulting in urinary incontinence. The proximal sphincter is removed during prostatectomy, and may also be damaged by radiation to the prostate. After prostatectomy, men achieve continence using the distal sphincter and the pelvic floor muscles that surround it. Thus, following prostatectomy, continence depends on the distal sphincter mechanism, which includes soft tissue supporting structures, the muscles of both the sphincter and the pelvic floor, and their intact innervation (both autonomic via the pelvic nerve and somatic via the pudendal nerve). In addition, the disruption of the nerve supply to the penis can interfere with normal erectile and hence sexual function.

Transurethral surgery for benign prostatic hypertrophy, while theoretically not disrupting the distal sphincter or the nerve supply to the pelvic floor muscles, does remove the proximal urethral sphincter. Such sphincter injury can result in incontinence.

Thus, urinary incontinence can be an iatrogenic outcome of prostate surgery. However, incontinence may also result from bladder dysfunction, which may persist from before surgery or be of new onset. Before surgery, men with benign prostatic hypertrophy have difficulty in emptying their bladders owing to bladder outlet obstruction. This may manifest as the lower urinary tract symptoms of frequency, urgency and urgency urinary incontinence. Detrusor overactivity may be demonstrated by urodynamic studies. While these symptoms are relieved by prostatectomy in over 75% of men,3,4 residual overactivity or incontinence may be accounted for by a variety of mechanisms, including persistent obstruction (38%), impaired contractility of the bladder detrusor muscle (25%) and sphincter deficiency (8%). 3

For a detailed description of the continence mechanisms and incontinence resulting from prostatic disease and its treatment see Koelbl et al.,5 pp. 299–308.

Definition of urinary incontinence

The MAPS study has used methods, definitions and units that conform to the standards jointly recommended by the International Continence Society and the International Urogynaecology Association, except where specifically noted. 6 These replace those formerly in use. 7

Urinary incontinence is defined as the ‘complaint of involuntary loss of urine’. 6 This can be subcategorised as follows:

Stress urinary incontinence (SUI) Complaint of involuntary loss of urine on effort or physical exertion (e.g. sporting activities) or on sneezing or coughing.

Urgency urinary incontinence (UUI) Complaint of involuntary loss of urine associated with urgency.

Mixed urinary incontinence (MUI) Complaint of involuntary loss of urine associated with urgency and also with effort or physical exertion or on sneezing or coughing.

Postmicturition leakage Complaint of a further involuntary passage of urine following the completion of micturition.

Urgency Complaint of a sudden compelling desire to pass urine that is difficult to defer.

In the MAPS study these were differentiated according to the men’s responses to questionnaires. Men could also categorise their incontinence as ‘other’ if they were incontinent under other circumstances, but we did not ask for clarification of the type of incontinence.

Incontinence can be further categorised according to the results of urodynamic studies (cystometry). These are:

Urodynamic stress incontinence (USI) Involuntary leakage of urine during filling cystometry, associated with increased abdominal pressure, in the absence of a detrusor contraction.

Detrusor overactivity (DO) Involuntary detrusor contractions occurring during filling cystometry. The symptoms of urgency and/or urgency incontinence may or may not occur.

However, we did not require the type of incontinence to be defined using urodynamics in MAPS.

Extent of problem in the UK

The number of men undergoing surgery for prostate disease is changing: in 2000–1, the number of TURPs in NHS hospitals in England was just under 30,000, while there were about 2000 open excisions of prostate (of which the majority would have been for prostate cancer). 16 By 2008–9, the number of TURPs had fallen to just over 25,000 (of which 2700 were with laser), while over 4000 open operations were performed. Thus, prostate surgery and its sequelae represent a considerable use of health resources and a health burden to men.

Existing guidelines

For men who have undergone prostate cancer treatment, the current National Institute for Health and Clinical Excellence (NICE) guidelines acknowledge that urinary incontinence has been reported as a result, most especially stress incontinence (which is mentioned as either temporary or permanent). NICE highlights that incontinence may be a problem after brachytherapy and external beam radiotherapy, as well as in those men who have also had a TURP.

NICE guidelines highlight some of the treatments available to men, including physical (pelvic floor muscle re-education, bladder retraining), medical (drug therapy) or surgical (injection of bulking agents, artificial urinary sphincters or perineal sling) interventions, and they give the following recommendations for urinary incontinence management following prostate cancer treatment. 17

Current recommendations from NICE

• Men experiencing troublesome urinary symptoms before treatment (of their prostate problem) should be offered a urological assessment.

• Men undergoing treatment for prostate cancer should be warned of the likely effects of the treatment on their urinary function.

• Health-care professionals should ensure that men with troublesome urinary symptoms after treatment should have access to specialist continence services for assessment, diagnosis and conservative treatment. This may include coping strategies, along with pelvic floor muscle re-education, bladder retraining and pharmacotherapy.

• Health-care professionals should refer men with intractable stress incontinence to a specialist surgeon for consideration of an artificial urinary sphincter.

• The injection of bulking agents into the distal urinary sphincter is not recommended to treat stress incontinence after prostate surgery in men.

No guidelines could be found for the treatment of urinary incontinence associated with either benign prostatic hypertrophy or TURP.

Treatment options

Treatment options for men with urinary incontinence after prostate surgery include:

• containment using continence products, including absorbent products, sheaths, urine drainage bags, mechanical devices such as penile occlusive devices or clamps, and catheters (see Cottenden et al. 18 for a comprehensive review)

• conservative options such as advice to modify lifestyle factors and pelvic floor muscle training (PFMT) (see Hay-Smith et al. 19 for a comprehensive review)

• surgery using injectable urethral bulking agents, a male sling, an adjustable balloon device or an artificial urinary sphincter or, as a last resort, creation of a catheterisable continent stoma by bladder neck closure or urinary diversion into a rectal reservoir or ileocaecal pouch with a catheterisable stoma (see Herschorn et al. 20 for a comprehensive review).

However, few of these options are supported by reliable research evidence.

Inclusion criteria

• Urinary incontinence at 6 weeks after prostate surgery (incontinence was defined as a response indicating a loss of urine to either of two questions in the screening questionnaire: ‘how often do you leak urine?’ and ‘how much urine do you leak?’.

• Full informed consent.

• Ability to comply with intervention.

Exclusion criteria

• Formal referral for physiotherapy or teaching PFMT related to prostate surgery.

• Radiotherapy planned or given during the first 3 months after surgery for men with prostate cancer.

• Transurethral/endoscopic resection of prostate carried out as palliation for outflow obstruction in advanced prostate cancer (known as ‘channel TURP’).

• Inability to complete study questionnaires.

Men with prostate cancer diagnosed at transurethral resection of the prostate

The literature suggested that approximately 15% of men have incidental prostate cancer when the prostatic chips removed at TURP are examined for pathology. 25 Within MAPS, men were still considered eligible for randomisation if the initial management plan did not include formal treatment (a wait and see policy). If the cancer was identified before randomisation, and either radiotherapy or radical prostatectomy was planned within the following 3 months, the man was not eligible for the TURP trial. However, men who were not randomised but subsequently readmitted for radical prostatectomy were eligible to be recruited as new participants to the radical prostatectomy group (after signing a new consent form and completing a new screening questionnaire after surgery). If cancer was diagnosed after randomisation, the men remained in the group to which they had been allocated even if radiotherapy or radical prostatectomy was carried out subsequently. These men could still have the MAPS intervention, if the timing of the new treatment allowed, and were followed up as per the MAPS protocol.

Thus, the MAPS study consisted of two stages: stage 1, the screening survey (used to identify eligible men), and stage 2, the two RCTs.

Withdrawal

Men were free to withdraw from the study at any point without giving a reason. Verbal consent was obtained from men who initially agreed to enter the trial, but later decided to withdraw, to enable relevant data to be retained or collected through central NHS resources.

Randomisation and allocation to group

When the baseline questionnaire and the consent form were received, the Aberdeen MAPS study office randomised the man to the intervention or standard care group.

Randomisation was by computer allocation using the randomisation service of the Centre for Healthcare Randomised Trials (CHaRT, in the Health Services Research Unit, University of Aberdeen). Allocation was stratified by type of operation (radical prostatectomy or TURP) and minimised using centre, age and pre-existing urinary incontinence. The process was independent of all clinical collaborators.

The study office informed all men of their allocation by post. All groups received a lifestyle advice leaflet (see Appendix 4.2). For men allocated to the intervention group, the study office arranged for the local therapist (physiotherapist or continence nurse) to send out the necessary appointments. A letter and GP information sheet were sent to each participant’s GP. Copies of the GP’s letter and the consent form were sent to the hospital urological consultant for filing in the man’s hospital notes.

A flow chart summarising the trial recruitment processes and procedures is shown in Figure 2.

FIGURE 2.

Flow chart of the trial recruitment processes and procedures. Q, questionnaire.

Intervention arm

The men in the intervention group attended for a MAPS therapist assessment of their symptoms after randomisation (see Chapters 6 and 11). The first appointment, for an hour, consisted of assessment and training, including customised goal setting for home practice of exercises. The men then attended a further three appointments, each lasting approximately 45 minutes, at around 2, 6 and 12 weeks after the first appointment. They were taught PFMT, with bladder training for men with urgency or urge incontinence. 26 This was supplemented by a booklet containing reminder instructions for PFMT and bladder training (see Pelvic Floor Exercises for Men Taking Part in the MAPS Study; Appendix 4.3). Men also received the lifestyle advice leaflet sent to the men in the standard care arm (see Appendix 4.2).

Biofeedback using digital anal examination was used to teach correct contraction technique and to monitor the strength of contractions. Although biofeedback used for diagnosis or training (repetitive exercising with machine-led feedback on the effectiveness of contractions) was not used routinely in the trial, therapists could use this at their discretion in individual cases. Further details of the intervention are given in Chapter 3.

Control arm

Men in the control group received standard care and a booklet containing supportive lifestyle advice only (without reference to PFMT) by post after randomisation (see Appendix 4.2). Men did not receive any formal assessment or treatment but were able to access usual care and routine NHS services if they felt they needed help with incontinence. This could include written advice if this was part of routine hospital care (such as leaflets containing instructions on PFMT).

Both arms

Use of NHS services, use of pads and practice of PFMT were documented in both groups using information from questionnaires (see Appendix 3.4) and 3-day urinary diaries (see Appendix 3.5) issued at baseline and 3, 6, 9 and 12 months after randomisation.

Questionnaires (see Appendices 3.1–3.4)

Men were sent postal questionnaires at baseline and 3, 6, 9 and 12 months. An additional health-care unit cost questionnaire was sent only at 6 months. The short questionnaires at 3 and 9 months contained only brief urinary incontinence, exercise and health-care utilisation questions.

Urinary diaries (see Appendix 3.5)

Men were asked to keep diaries at baseline and 3, 6, 9 and 12 months after randomisation, kept for 3 days at each time period.

Data processing

Data from the various sources outlined above were sent to the study office in Aberdeen for processing. Staff in the study office carried out extensive range and consistency checks to enhance the quality and accuracy of the data. Essential missing data were sought from the recruitment officers at the centres, or the men, by post, telephone or email as appropriate.

Outcomes

The primary clinical outcome was subjective report of urinary continence at 12 months. 27

Incontinence was defined as a response indicating a loss of urine to either of two questions in the screening questionnaire, ICI-SF: ‘how often do you leak urine?’ and ‘how much urine do you leak?’

The primary measure of cost-effectiveness was incremental cost per quality-adjusted life-year (QALY).

Secondary outcome measures were as follows.

Trial analyses

The principal comparisons in each trial were between men allocated to active therapy (up to four visits to a therapist plus the lifestyle advice leaflet) and men allocated to the control group (lifestyle advice leaflet only). The two populations of men (having radical prostatectomy or TURP) were analysed as separate trials. The primary outcome measure (urinary incontinence at 12 months) and secondary outcome measures were analysed using general linear models that adjusted for the minimisation covariates (age and pre-existing urinary incontinence) and, when possible, the baseline measure of the outcome. For the primary outcome only, unadjusted analyses were also reported. All analyses used 95% CIs. For the binary outcomes, a Poisson link function was used to estimate relative risks (instead of estimating odds ratios from a logistic model) and robust standard errors were used to estimate the CIs. 28 For illustrative purposes, the relative risk of the primary outcome was also transformed to a risk difference.

The primary statistical analysis was based on all men as randomised, irrespective of subsequent compliance with the treatment allocated (intention to treat). The intention-to-treat approach gives the least biased estimate of effectiveness of the two interventions. Given that it was likely that some of the participants would not attend the therapy sessions (e.g. because they were continent), a secondary comparison was conducted to estimate the efficacy of the treatment received (i.e. what is the effect if the participants actually received the treatment they were allocated to?). The so-called ‘per-protocol’ approach for estimating efficacy of treatment, in which compliers with treatment in each group are compared with each other, can have substantial selection bias. A more robust method is to use a latent variable approach. 29 We used the method of adjusted treatment received as described by Nagelkerke et al. 30 The method used a two-stage least squares approach, whereby treatment received and the residuals from that model were used as an independent variable in a second model together with the treatment received to estimate the effects on the primary outcome.

Missing items in the health-related outcome measures were treated as per the instructions for that particular measure. No further imputation for missing values was undertaken. The ways in which the data were analysed were prespecified in the statistical analysis plan, which was agreed in advance with the MAPS Trial Steering Committee.

Timing and frequency of analyses

A single principal analysis was carried out at 15 months after the last man was recruited. The Data Monitoring Committee considered confidential interim analyses of data on three occasions during the data collection period (January 2006 – 31 randomised to radical prostatectomy and 48 randomised to TURP; January 2007 – 180 randomised to radical prostatectomy and 200 randomised to TURP; January 2008 – 297 randomised to radical prostatectomy and 364 randomised to TURP). The Data Monitoring Committee did not recommend any amendments to the protocol on any occasion.

Planned secondary subgroup analyses

Subgroup analyses (separately for the two populations) explored the effect on urinary incontinence at 12 months after randomisation of:

1. pre-existing urinary incontinence (before prostate surgery)

2. age (up to 60 years, 61 years and over for radical prostatectomy; up to 70 years, 71 years and over for TURP)

3. body mass index (BMI) (up to 30 kg/m², 30–34.9 kg/m², 35 kg/m² or greater)

4. type of incontinence at trial entry (SUI, UUI, MUI, postmicturition leakage)

5. other morbidity

6. type of therapist (physiotherapist or nurse)

7. centres with and without biofeedback machines.

Stricter levels of statistical significance (2p < 0.01) were sought, reflecting the exploratory nature of these analyses.

Ancillary analyses

Introduction

The economic evaluation was based on a within-trial analysis at 12 months after recruitment for men with urinary incontinence 6 weeks after radical prostatectomy or TURP. The question addressed was: what is the clinical effectiveness and cost-effectiveness of active conservative treatment delivered by a specialist continence physiotherapist or a specialist continence nurse compared with usual management? The perspective of the study was based on a societal viewpoint and included both the costs of the health service provider (the NHS) and those of the patients.

Measurement of resource use

The use of health services as a consequence of being incontinent was recorded prospectively for every participant in the study. Resource utilisation data were collected using questionnaires and urinary diaries. These data were collected using questionnaires sent to the participants at baseline, and 3, 6, 9 and 12 months. Resource utilisation data collected also included the intervention, i.e. the number of visits to the therapists, who were either specialist continence physiotherapists or continence nurse specialists. According to the protocol, PFMT intervention comprised four sessions. Details of the intervention are provided in Chapter 3. The first session of PFMT was 1 hour and the other three sessions were approximately 45 minutes each. Each session was conducted in a hospital department. The consumables required per session were gloves, K-Y Jelly, wipes and paper towels. No additional resources were required for the biofeedback as no equipment was used; verbal biofeedback was used to teach the men how to contract their muscles optimally and advise them on improvement from previous appointments.

Primary care and outpatient resource use included visits to the GP as well as to the outpatient department. The number of GP visits and the contact (doctor or nurse) were obtained from the 3-, 9- and 12-month follow-up questionnaires. Number of outpatient visits was obtained from the 3-, 6-, 9- and 12-month follow-up questionnaires. For the length of stay, the number of days the men were admitted was recorded. Other resource use included the number and type of drugs the patients were prescribed for their incontinence problems, the number of pads used and, finally, the number of bed and chair protectors used. The data reported by the patients were used to calculate the average and total resource use per patient.

The information generated from these questionnaires entailed manipulation of the data to perform the comparative analysis. Details of methods used to estimate resource use collected are included in Table 3.

Identification of unit costs

As described above, costs focused on the direct health service costs associated with each treatment. Unit cost data were extracted from the literature or from relevant sources such as manufacturers’ price lists (British National Formulary, BNF)31 and NHS Reference Costs. 32 The year of the cost data is 2008 and the currency is pounds sterling (£).

The costs of the intervention included the cost of PFMT sessions. These comprised the costs of the staff involved, consumables and overheads. The costs of producing the leaflets for the trial were not included in the analysis as all the men in the trial received leaflets. Men in both groups received a booklet containing supportive lifestyle advice (without reference to PFMT) by post after randomisation (see Appendix 4.2). Men in the intervention group also received a MAPS pelvic floor exercise leaflet (see Appendix 4.3) from the therapist at the first visit. The booklet aimed both to support and to reinforce the anatomy teaching received during MAPS therapy appointments, as well as the exercise programme that had been set. It was therefore assumed that the costs would be the same for both groups. The cost of training that therapists received (1-day course) was included in the intervention costs because it was low and was not likely to impact on the overall costs.

The cost of the follow-up management comprised the cost per visit to both primary (GP and nurse appointments) and secondary (outpatient appointments and number of inpatient days) health-care providers. Unit costs for GP’s visits were obtained from the Personal Social Services Research Unit (PSSRU) unit costs of community care. 33 Unit costs for outpatient services were obtained from the Scottish Health Service Costs (SHCS) (Information and Statistics Division website34) for the primary analysis and the national reference costs in a sensitivity analysis (Department of Health website35). The inpatient costs were those of the wards to which the men were admitted.

Other costs considered included containment products. These comprised all the products that participants used, such as absorbent pads, penile collecting sheaths, bladder catheters and bed and chair protectors. The unit costs of these items were taken from the providers of these items or from the NHS suppliers, where available. The unit cost of sheaths is based on a weekly cost of sheaths, estimated assuming that one sheath is used each day, the reusable leg bag is used for 3 days and one night bag is used each night. The unit costs of the catheter were based on the assumption that the catheter was used over a 3-month period, and similar assumptions were made for the leg and night bags. The unit cost of the medications was taken from the BNF,31 and the cost per patient in terms of medication use was calculated by multiplying the unit cost by each number of units consumed for each patient. The costs considered were those of the drugs, not the prescription charges. Table 4 provides a summary of the unit costs for the resources used.

The data describing the resource utilisation of participants were combined with estimates of unit costs for each of the areas of management considered. This allowed for estimation of total cost for each participant, as well as the average cost for each area of resource utilisation and average total cost. The results are reported in Chapters 8 and 13.

Participant costs of urinary incontinence

As the perspective of the study was the NHS and patient, those costs borne by participants and their families were also considered. Participants’ resource use was taken as time taken to access services (e.g. attend GP, physiotherapist, outpatient or inpatient appointments), travel costs and the time taken off usual activities owing to poor health. Similar costs were included for spouses, relatives or friends who accompanied them to their appointments. Travel costs to patients and their families were based on actual fares when public transport was used and published mileage rates in the case of those who used their own vehicles (HM Revenue and Customs website36). These data were collected through postal questionnaires administered at 12 months.

In the case of patients who would have been engaged in employed work, the value of their time was taken as the gross average full-time wage rate for men (Office for National Statistics website37). The value for those who were not in formal employment was based on 57% of the average national rate and 43% for those who may have been involved in leisure activities. 38 The costs of friends/relatives accompanying patients to hospital were estimated in the same way. These unit time costs, measured in terms of their natural and monetary terms, were combined with estimates of number of health-care contacts derived from the health-care utilisation questions. Self-purchased health care included items such as pads bought by the participant, prescription medicines and over-the-counter medications. Information about these was collected through the health-care utilisation questions. Patients’ time and travel costs were based on the information collected, and are described in Table 5.

Quality of life

Effectiveness within the trial was measured in terms of QALYs and subjective continence at 12 months (assessed using data from the ICI-SF). Quality of life data were collected at baseline and 6 and 12 months. This was generated using generic health status measurement tools, the EQ-5D and SF-12, included in the questionnaires. The EQ-5D measure divides health status into five dimensions (mobility, self-care, usual activities, pain/discomfort and anxiety/depression). Each of these dimensions has three levels, therefore 243 possible health states exist. 39 Responses of the patient EQ-5D questionnaires were transformed using a standard algorithm to produce a health state utility at each time point for each patient. The utility scores obtained at baseline and 6 and 12 months were used to estimate the mean QALY score for each group. The estimation of QALYs took account of the mortality of study participants. Participants who died within the study follow-up were assigned a zero utility weight from their death until the end of the study follow-up. QALYs before death were estimated using linear extrapolation between the QALY scores at baseline and all available EQ-5D scores up to death.

As described below in the section on the sensitivity analysis, the responses from the SF-12 questionnaire were also used as the basis of QALYs, and were mapped on to the existing Short Form questionnaire-6 Dimensions (SF-6D) measure using the algorithm by Brazier et al. 40 to allow utility values to be estimated for each time point. These utility scores were transformed into QALYs using the methods described above to provide an alternative measure of QALYs for each patient.

Incremental cost-effectiveness

Data collected on costs and effects of the interventions were combined to obtain an incremental cost-effectiveness ratio (ICER). This was performed by calculating the mean difference in costs between the interventions and control groups over the difference in effect between the interventions and control groups. This gives us the cost per additional QALY gained for the new interventions relative to standard practice.

The primary analysis was based on the 1-year follow-up of the trial and the outcome was the incremental cost per QALY. This outcome was chosen to reflect a societal decision-making perspective. The results are presented as point estimates of mean incremental costs, proportion of men continent, QALYs and cost per QALY. Measures of variance were based upon bootstrapped estimates of costs, QALYs and incremental cost per QALY. Incremental cost-effectiveness data are presented in terms of cost-effectiveness acceptability curves (CEACs).

Data analysis (economics)

As data were collected over a 1-year period, discounting was not carried out. The numbers of missing data for each variable used in the analyses of cost were quite low, and data that were missing were considered to be missing completely at random. Data reported as mean costs for both cases and controls were derived for each item of resource use and then compared using unpaired t-tests and linear regression adjusted for baseline values. As the data were not normally distributed, non-parametric bootstrapping was used to estimate confidence limits around the difference in cost for each area of resource use and total costs.

Assessment of pelvic floor strength

During each therapy appointment, pelvic floor muscle contraction strength was evaluated by a digital anal examination using the Oxford score (graded 0–5). 45 An additional grade (6) was added to define a very strong anal squeeze. 46 The new grading system was used to assess separately the strength of both the external anal sphincter and the deeper puborectalis muscle (taken to represent the pelvic floor muscles). The external anal sphincter was assessed at 1–2 cm from the anal meatus, and the puborectalis at 3–4 cm from the anal meatus.

Verbal biofeedback from this examination was used to teach the men how to contract their muscles optimally, and advise them on improvement from previous appointments. At each assessment, the maximum duration of each contraction was timed by counting.

Pelvic floor muscle therapy regimen

PFMT was aimed at improving the strength of the pelvic floor muscles to allow effective contraction during exertion to prevent urinary leakage. PFMT consisted primarily of three maximum-strength contractions with a 10-second break between each one, practised in three positions (lying, sitting and standing) twice daily (see Appendix 4.3). Targets were set for the duration of each contraction, up to a maximum of 10 seconds, and revised in successive appointments if progress had been made. In addition, men were taught to carry out a sustained submaximal contraction of the pelvic floor muscles during walking and to perform a strong contraction before and during any event that might cause leakage, such as coughing or rising from sitting (‘the knack’). 47 Men were advised to eliminate urine remaining in the bulbar urethra by using a strong contraction after urination was finished, in order to prevent postmicturition dribble. 42 Contracting the pelvic floor muscles during sexual activity was also recommended to achieve, maintain or improve erectile strength.

Bladder training/urge suppression

Men with urgency or UUI were taught urge suppression techniques in order to avoid rushing to the toilet when the bladder was starting to contract (see Appendix 4.3). Fluid advice, including avoiding or reducing caffeine, was also offered.

Written supplementary guidance

The MAPS PFMT leaflet (see Appendix 4.3) aimed both to support and to reinforce the anatomy teaching received during appointments, as well as the exercise programme that had been set. To maximise understanding, careful consideration was given to the language and terminology used in this leaflet, taking into account the sensitive nature of incontinence and erectile dysfunction. The use of medical and anatomical terms was minimised in favour of a plain English approach (‘urine leakage’ for incontinence). 48

Drafts of the leaflets were reviewed for face and content validity by lay persons and health-care professionals with knowledge of men’s health and continence issues.

Rationale for performing a digital anal examination

A digital anal examination was undertaken to assess the strength and endurance of, firstly, the external anal sphincter and, secondly, the puborectalis muscles. Wyndaele and Van Eetvelde49 demonstrated the reproducibility of assessing the puborectalis by anal assessment using grades 0–5. By assessing puborectalis muscle strength, the strength of the surrounding pelvic floor muscles would also be graded. The muscles were graded from 0 to 6, with 0 being ‘no flicker or contraction’ and 6 being ‘very strong, unable to withdraw finger. 42 Repeating the examination at subsequent visits enabled therapists to provide verbal feedback to men that their exercises were effective in building up muscle strength and to monitor progress.

Rationale for asking men to perform pelvic floor exercises in three positions

Pelvic floor muscles support the abdominal contents and prevent urinary leakage. The three positions provided a graded method of increasing the effect of gravity, in order to provide extra muscle work load. The pelvic floor muscles were recruited initially in a lying position, without the effect of gravity. As strengthening occurred, pelvic floor muscles were subject to a higher load by recruiting them in a sitting position, where the downwards descent of the pelvic floor would be partly prevented by the seat of the chair. A greater load would be placed on the pelvic floor during standing, when gravitational forces opposed the elevation of the pelvic floor during exercise. MAPS adopted this regimen for the intervention supported by evidence from four previously documented trials, which found it to be convenient, acceptable and comfortable for patients. 24,50–52 It was believed that men needed to be able to tighten their pelvic floor muscles in a number of positions, so that they could recruit them speedily during coughing and sneezing.

Rationale for performing three pelvic floor muscle contractions

The PFMT programme was aimed at increasing pelvic floor muscle strength in order to counteract increases of abdominal pressure during exertion. Based on clinical research of quadriceps strengthening using a progressive resistance machine, repeated computerised readings showed that the first contraction gave the patient the feel of the movement but failed to achieve maximum power. The second contraction attained maximum power, whilst the third failed to reach maximum power owing to fatigue. 53 Kegel54 stated that maximum power was a key element to gaining increased muscle strength. These principles informed the PFMT programme, considering that the maximum power of pelvic floor contraction would be attained using three muscle contractions in each position held for up to 10 seconds. The target was individually adjusted as performance improved.

Rationale for performing the regimen twice a day

Kegel54 recommended 300–400 pelvic floor muscle contractions a day to treat SUI in women.However, clinical practice has shown that patients find this level of commitment to be too arduous, resulting in attrition and demotivation. The principles of muscle building show that it is the quality of the contraction that is more important than the quantity. 53,55

The MAPS intervention was therefore designed to provide targets that were achievable in order to motivate men to maintain the regimen within the constraints of the protocol. In a previous trial,42 55 men were asked to perform their exercise sets only twice a day. After 3 months, all except one (who had severe back pain) showed a statistically significant increase in pelvic floor muscle strength. We therefore felt that this regimen had a proven ability to increase pelvic floor muscle strength.

Rationale for contracting the muscles as strongly as possible

The pelvic floor muscles consist of two-thirds slow-twitch continually tonic muscle fibres and one-third fast-twitch muscle fibres, which can be speedily recruited when extra support is needed during activities that increase intra-abdominal pressure. 56 Both fibre types are recruited during maximum contraction of the pelvic floor muscles. In order to achieve an increase in muscle bulk, the MAPS intervention used maximum voluntary effort, which was expected to result in the hypertrophy of muscles and an increase in local blood supply. 53,55

Rationale for functional use of muscles

Pelvic floor muscles need to be recruited to prevent leakage of urine during activities that increase intra-abdominal pressure. ‘The Knack’ is the technique, or learned skill, of tightening just before and during these activities. 47 Owing to its significant role in contributing to continence, teaching of ‘the Knack’ was therefore included as an element in the MAPS intervention.

Reasons for increasing pelvic floor muscle endurance

Slow-twitch muscle fibres fulfil a number of important functions: pelvic floor support, bladder and bowel control, sexual activity, posture and respiration. The upright posture stimulates the pelvic floor reflex, which results in contraction of the slow-twitch fibres in response to the weight of the abdominal contents. 57 In order to meet this demand, the pelvic floor muscles need to have sufficient muscle endurance to prevent urinary leakage. By encouraging the patient to tighten the pelvic floor muscles slightly during walking (as taught in the MAPS intervention), a functional method of potentially increasing the use of slow-twitch fibres and hence muscle endurance was achieved.

Rationale for tightening the pelvic floor muscles after urinating

One of the superficial pelvic floor muscles, the bulbocavernosus muscle, encircles the proximal 50% of the penis and tightens by reflex action at the end of micturition to facilitate emptying of the bulbar portion of the urethra. 58 Teaching men to contract their pelvic floor muscles strongly after they have completed micturition will result in the recruitment of the bulbocavernosus muscle along with the other pelvic floor muscles. 42 This muscle contraction will then facilitate the evacuation of residual urine from the bulbar urethra. This may restore or develop the reflex postvoid milking mechanism identified by Wille et al. 58 and termed the ‘urethrocavernosus reflex’ by Shafik and El-Sibai. 59 Thus, as an additional strategy to attain continence, participants were taught to perform consciously a pelvic floor muscle contraction immediately after micturition.

Rationale for tightening the pelvic floor muscles during sexual activity

The superficial bulbocavernosus and ischiocavernosus muscles are active during penile erection. 60 The bulb of the penis sits on the inferior aspect of the deeper layer of the pelvic floor muscles, which form a firm base for the erect penis. The bulbocavernosus muscle prevents blood from escaping through the deep dorsal vein during an erection. One study has shown that pelvic floor exercises can restore erectile function in 40% of men and improve it in a further 36%. 42 As this is another potential benefit of PFMT, it was decided that it would be appropriate to include erectile function in the MAPS intervention materials and be measured as a secondary outcome. However, it is not yet clear whether men will benefit after radical prostatectomy as the amount and degree of nerve damage caused by surgery is likely to be variable. Erectile function was a secondary outcome of the study.

Rationale for choice of urge suppression techniques

A detrusor contraction produces a desire (urge) to empty the bladder. If urgency sensations cannot be overcome, urinary incontinence may occur. The resulting fear of leakage can cause anxiety, breath-holding and descent of the diaphragm, which, coupled with abdominal muscle contraction, can produce early inappropriate micturition. A retrospective study in women has reported that effective urge suppression techniques include keeping calm, sitting down or standing still and waiting 1 minute until the initial urge sensation disappears. 46 PFMT can be used to strengthen the pelvic floor musculature and, together with urge suppression techniques, can help to restore bladder control.

Rationale for giving fluid, dietary and lifestyle advice

All men received fluid, dietary and relevant lifestyle advice as part of the therapy appointments, supplemented by written information (see Appendix 4.2). Advice included information that reducing fluid intake (underdrinking) to avoid leakage may lead to urinary tract infections, constipation and dehydration. 61 Conversely, drinking excessive amounts (in the belief that this is beneficial for health) may have adverse effects such as an increased risk of leakage. 62 However, men experiencing nocturia were advised that avoiding fluids 2 hours before bedtime may be helpful.

Drinks containing caffeine or alcohol may cause increased risk of urgency and men were advised to reduce or avoid them. 61 Anecdotal evidence has shown that certain foods (e.g. onions, spicy foods and curries) can cause increased gut peristalsis, which may also have an effect on the bladder, causing it to be overactive and contractile. Other risk factors for an overactive bladder were highlighted, including the effect of constipation, smoking and obesity. 63 Information on all these elements was included in the MAPS lifestyle advice leaflet (see Appendix 4.2).

Rationale for four appointments in 12 weeks

The value of psychological support for men following radical prostatectomy has been stressed in the literature,52 as has the intrinsic value of therapist contact in order to maintain patient motivation. 64 Within MAPS, therefore, a schedule of four appointments (at baseline, 2 weeks, 6 weeks and 12 weeks) was considered sufficient to monitor postsurgical muscle strength development and maintain motivation but not be too burdensome on patients or costly in terms of resources (chiefly therapist time).

Pelvic floor muscle strength can improve over a 3-month period of PFMT. 42 Men in the study received four appointments and were encouraged to continue their exercise regimen for life, with particular emphasis on functional work (e.g. contracting during activity or counteracting increases in intra-abdominal pressure by use of ‘the Knack’). A previous trial by van Kampen et al. 24 using pelvic floor exercises and functional use of these muscles showed significant reduction in urinary incontinence at 1, 6, and 12 months after radical prostatectomy, demonstrating that improvement was maintained while men continued to perform their exercises.

Pelvic floor muscle training

Men in the intervention group were instructed:

• to carry out three maximum pelvic floor contractions in three positions (lying supine with knees bent and feet on the couch, sitting with knees apart and standing with feet apart) twice per day

• to ‘lift’ their pelvic floors slightly while walking

• to tighten their pelvic muscles before activities that might cause them to leak, such as coughing

• to tighten after urinating to eliminate the last few drops.

Biofeedback

Biofeedback involved monitoring the strength of a pelvic floor contraction (by digital anal assessment) and verbally relaying the information back to the men in order to confirm that they were performing contractions correctly (lifting up in a cranial direction) and to inform them when they were increasing the strength or duration of their contractions. Therapists were trained to consistently grade the pelvic floor muscle strength and endurance by digital anal assessment at each session. The findings were used to set progressive targets for the men. Treatment was therefore individualised and could be progressively increased for each man.

If it was felt clinically indicated, in addition to digital anal assessment, therapists used machine-mediated biofeedback with an anal biofeedback probe in centres where this was available (see Table 7), both for diagnosis and for teaching of correct muscle contraction.

Bladder training/urge suppression

BT/US involved advice to gradually delay urination (by pelvic floor muscle contraction or calming/distracting activities) to teach the bladder to hold increasing volumes of urine. Men were instructed to relax for 1 minute when they first felt an ‘urge’, then walk calmly to the toilet or delay urination until the next ‘urge’.

Written information

All participants received the lifestyle advice booklet (see Appendix 4.2), and those in the intervention group also received the pelvic floor exercise booklet (see Appendix 4.3).

Ensuring standardisation of intervention

All staff delivering the intervention received exactly the same training in order to ensure consistency of their method of teaching and delivery of the PFMT, BT/US and biofeedback. Both specialist continence physiotherapists and specialist continence nurses were eligible to deliver the intervention, thus increasing the generalisability of the trial.

The therapists recorded their assessments and treatment programmes on standard study forms (see Appendix 4.1). Data were stored locally in case notes but collected and analysed centrally (see Chapters 6 and 11).

Centre screening and recruitment

Table 8 shows the number of men approached in each centre, and how many were eventually eligible for screening and randomisation. In total, 780 men having radical prostatectomy and 2836 men who had TURP were sent a screening questionnaire, 742 and 2590 responded, and 429 and 442 were eventually randomised. However, 18 of the ‘radical’ men from one centre were subsequently excluded after randomisation (postrandomisation exclusion) as therapy was not available during some of the period of screening in that centre, leaving 411 randomised to the radical prostatectomy trial.

Recruitment of men to the randomised controlled trial

Each man who indicated on his screening questionnaire that he had urinary incontinence was sent a baseline questionnaire and contacted around a week later by a dedicated recruitment officer based at the MAPS study office in Aberdeen. Eligibility was confirmed and, upon receipt of the signed RCT consent form, men were randomised to the intervention or standard care groups. Figures 4 and 5 show the flow of the number of patients who were approached to take part in the screening through to randomisation into the radical prostatectomy and TURP trials respectively.

FIGURE 4.

Flow chart of men from operation to recruitment to RCT: radical prostatectomy. a, Postrandomisation exclusion: therapy was not available during some of the period of screening in one centre (18 men).

FIGURE 5.

Flow chart of men from operation to recruitment to RCT: TURP.

Incidence of urinary incontinence

The therapists asked the men at each visit to rate their incontinence (in the previous week). This allowed the therapists to monitor the change in reported incontinence. They used the same form of question as the questionnaires, based on the ICI-SF instrument, which were also used to measure the primary outcome. The proportion of men with incontinence fell from 92% to 73%, while the mean ICI score decreased (improved) from around 8 at the start of treatment to around 4 afterwards (Table 20 and Figure 7).

FIGURE 7.

Trend analysis of the proportion (%) of men incontinent at each visit: men having had radical prostatectomy. ICI-Q score: 0 = none, 21 = maximum (worst) score. Derived from questions 1–3 of the ICIQ-UI Short Form Questionnaire.

Type of urinary incontinence during therapy period

The distribution of type of incontinence reported by the men did not vary with time across the therapy visits (Table 21 and Figure 8) except that the proportion with stress incontinence alone decreased slightly (from 84% to 72%), and postmicturition leakage also decreased (from 63% to 25%).

FIGURE 8.

Type of incontinence at each visit and change over time: men having had radical prostatectomy.

Examination and functional use of pelvic floor muscles

Therapists examined men at each visit to assess skin damage, skin infection, ability to tighten the anus, perform penile retraction and testicular lift, evidence of leakage on coughing and (for those who did leak) ability to prevent leakage on coughing. Very few men showed evidence of skin damage or infection (data not shown).

Four different aspects of functional use of pelvic floor muscles were assessed (ability to tighten anus; ability to perform penile retraction; leakage on coughing; ability to prevent leakage on coughing). While most men (around 95%) were able to contract well enough to tighten the anal sphincter at least a little from baseline onwards, the proportion able to demonstrate a testicular lift increased slightly with time (from 80% to 90%; Figure 12). The proportion who leaked when coughing decreased from about 18% to 9% during the therapy period. Around 80% of these men were able to contract their pelvic floor muscles sufficiently to prevent leakage when coughing at the first visit, and this improved only slightly to around 85% by the fourth visit.

FIGURE 12.

Aspects of ability to use pelvic floor muscles at each visit and change over time: men having had radical prostatectomy.

Use of machine-led biofeedback

Biofeedback was available in 13 of 34 MAPS centres, and was used clinically for MAPS men in five of them (see Table 7). Therapists would have liked access to this facility in four other centres where biofeedback was not available. Biofeedback can be used in two ways:

• to feed back information to men that they are actually performing a correct pelvic floor contraction, and at what strength

• as part of a repetitive training regimen when men are asked to use the machine to enable them to monitor their exercise function for a period of time (such as 20 minutes).

It was unclear which type of biofeedback was practised in the centres where this was available, but therapists from five centres recorded its use in 16 men (see Table 7) from the radical prostatectomy group (the number at each time point is shown in Table 25). In some cases men may have preferred anal examination using a machine rather than digital examination by the therapist for teaching of correct contractions. As it was most often used on the first visit, this suggests that suggests that biofeedback was used in a diagnostic rather than in a training capacity.

Urinary outcomes

Urinary incontinence was also measured at 3, 6 and 9 months after randomisation together with other urinary outcomes. Table 29a describes the various urinary outcomes at each follow-up and Table 29b shows the formal statistical testing of the differences at each time point. Figure 14 is a pictorial representation of the percentage of men incontinent at each follow-up and Figure 15 shows the change in mean ICI-score over time. The data show that there were no statistically significant differences between the intervention and control groups at any of the time points in terms of urinary incontinence and the other urinary outcomes measured.

FIGURE 14.

Per cent of men incontinent at each time point.

FIGURE 15.

Mean ICI-Q score at each time point. ICI-Q score: 0 = none, 21 = maximum (worst) score. Derived from questions 1–3 of the ICIQ-UI Short Form Questionnaire.

Type of incontinence

Table 30 and Figure 16 show the type of incontinence at baseline, and at 6 and 12 months after randomisation. Men could report more than one type of incontinence. At all time points the majority (70%) of men had (any) stress incontinence: this did not vary much after the first 6 months, and the proportions were not significantly different between the intervention and control groups (see Table 30 and Figure 16a). Around half of the men had urgency or mixed incontinence. The proportions of men with other types of urinary incontinence (urgency, Figure 16b; mixed, Figure 16c; postmicturition leakage, Figure 16d; and other types of incontinence) decreased over the first 6 months, but there was little further improvement, or difference between the groups.

FIGURE 16.

Type of incontinence at baseline and at 6 and 12 months after randomisation. Men could report more than one type of incontinence. (a) Stress urinary incontinence. (b) Urgency urinary incontinence. (c) Mixed urinary incontinence (proportion of men reporting both stress and urgency urinary incontinence). (d) Postmicturition leakage (defined as ‘urine leaking when urination finished’).

Use of aids or protection for urinary incontinence

Table 31a shows the men’s use of aids to protect them from urinary leakage: this did not differ significantly according to the randomised groups at any of the follow-up time points. Table 31b presents the statistical analyses of these outcomes. About 40% of the men were still using pads at 12 months to protect themselves from leakage accidents, although in some cases this might have been more of a precaution than because they actually leaked.

Bowel function

In addition to urinary outcomes, men were also asked to describe some aspects of bowel function. Few men (< 10%) had faecal incontinence or constipation by the end of follow-up at 12 months, although rather more reported faecal urgency occasionally or more often. Table 32 and Figure 17 show that there were no differences in any aspect of bowel function between the men in the randomised groups.

FIGURE 17.

Type of bowel problems at baseline and 6 and 12 months.

Sexual function

Table 33 compares the men in the randomised groups in terms of sexual function outcomes. Over 90% of men had normal erectile function before their operation. Although around one-third had an active sex life at 12 months, the majority said that this was less satisfactory than before their operation. There were, however, no differences at 12 months according to the randomised groups in terms of the proportion of men with an active sex life [RR 0.94 (95% CI 0.73 to 1.22); adjusted for age, urinary incontinence before surgery and baseline value, p = 0.661] or the proportion of men who rated their sex life as worse after the operation [RR 0.79 (95% CI 0.47 to 1.34); p = 0.391].

Table 34 compares the randomised groups in terms of problems with sexual function. There were no significant differences in sexual function outcomes between the intervention and control groups (Table 34 and Figure 18). Only about 20% of the men were able to achieve a normal erection or one with slightly reduced stiffness by 12 months after surgery, and almost all reported a lack of semen or no ejaculation. Of those men, few reported more than slight pain. Around 60% of men used drugs and about 20% used a vacuum device to help with sexual function. Only 20% reported urinary incontinence during intercourse.

FIGURE 18.

Quality of erectile function at 12 months after prostate surgery: Radical prostatectomy.

Quality of life

General health outcomes were measured using the EQ-5D and SF-12, the latter subdivided into role – mental (SF-12M) and role – physical (SF-12P) scores. The slight increase in the scores over time represents recovery from the operation but there were no differences between the randomised groups at any time point in EQ-5D or SF-12 scores (Table 35 and Figure 19).

FIGURE 19.

Graphical representations of EQ-5D and SF-12 scores over time. SF-12M, SF-12 role – mental; SF-12P, SF-12 role – physical.

Pelvic floor muscle training

All men were asked to report on their practice of carrying out pelvic floor exercises at baseline and 6 and 12 months after randomisation. Initially a high proportion of men (over 80%) reported practising them, of whom around 80% carried them out every day. As this occurred before randomisation, it must reflect the high profile given to PFMT in the standard care of men after radical prostate surgery.

The prevalence of exercising in the control group had fallen by 6 months (to 62%), while men in the therapy group were more likely than controls still to be performing exercises (83%). This difference was maintained at 12 months: a significantly higher proportion of men from the intervention group were carrying out any PFMT at 12 months after randomisation (67%) compared with those in the control group (50%; see Table 36). Significantly more were practising for at least 3–4 days each week in the intervention group than in the control group (56% vs 36%). Although men in the intervention group were performing fewer daily contractions by 12 months than they had previously, the difference between randomised groups at this time point did not reach statistical significance (mean contractions 11.7 intervention vs 19.4 control) (see Table 36). This is likely to reflect the taught exercise regimen in the intervention group (aiming for 18 strong contractions every day) compared with recommendations from NICE,71,72 which are to perform eight contractions three times a day.

Men in the intervention group were also significantly more likely to perform contractions while walking. Differences between groups in performing a pelvic floor muscle contraction prior to increases in intra-abdominal pressure such as coughing or lifting (also known as ‘the Knack’) did not quite reach statistical significance (Table 36). There were no differences between the groups in terms of ‘the Knack’ reducing or stopping urinary leakage (although this analysis includes only the men using this technique).

Lifestyle outcomes

Men were also advised, in the lifestyle advice leaflet sent to both groups but reinforced by the therapists in the intervention group, about the benefits of general health strategies such as taking more exercise. There were few differences between the groups in terms of other types of exercise practised (Table 37) and no statistically significant difference in the proportion taking general exercise [RR 0.94 (95% CI 0.87 to 1.02); p = 0.151].

Finally, men in both groups were given (via the lifestyle advice leaflet) other general advice on lifestyle changes they could make that might help both with incontinence and with general health. Again, this advice was reinforced by the therapists for men in the intervention group. There were few differences between the groups in terms of changes made to lifestyle factors (Table 38).

Resource use

Table 40 details the average total resource use for the intervention and subsequent use of health services over the 12-month follow-up period after randomisation. The pattern of resource use was similar across both groups and there were few statistically significant differences between the two groups. In addition to differences in costs of the intervention (incurred only by those randomised to the intervention group), these costs included the use of pads paid for by the NHS, incontinence-related GP doctor and nurse visits, hospital physiotherapist visits and the number of days that participants were off work. The use of resources in each of these areas was higher for the control group, apart from the subsequent number of hospital physiotherapy visits, which was higher for the intervention group. A detailed description of the use of NHS and private provider health services is provided in Appendix 5.

Costs

Societal perspective

The cost-effectiveness of the intervention compared with control is dependent on whether the differences in QALYs are considered to be important to people with incontinence. Taking the mean difference in the total societal costs from Table 42 (–£588) and the mean difference in QALYs from Table 43 (–0.002) it can be seen that the intervention is, on average, more effective and less costly (Table 44).

Uncertainty around the estimates of QALYs and costs was derived using 1000 bootstrap simulations. The bootstrap estimates in Figure 21 indicate that in most of the instances the intervention group had lower costs than the control; however, there was a relatively wide distribution in the difference in QALYs and costs.

FIGURE 21.

Representation of the uncertainty in differential mean costs and QALYs: societal perspective (radical prostatectomy).

At a cost-effectiveness threshold of £20,000 per QALY, the intervention has a likelihood of 89% of being cost-effective, and at a threshold of £30,000 per QALY the intervention is 84% likely to be cost-effective (Figure 22). However, these results have to be interpreted cautiously as they are driven almost entirely by differences in time away from usual activities.

FIGURE 22.

Cost-effectiveness acceptability curve for intervention versus control: societal perspective (radical prostatectomy).

NHS perspective

Taking the mean difference in total NHS costs from Table 42 (£181) and the mean difference in the QALY estimate after adjusting for baseline EQ-5D scores from Table 43, the mean incremental cost per QALY is £90,510 (Table 45).

As with the societal perspective, bootstrap simulations were undertaken to estimate the uncertainty around the benefit and costs. The bootstrap estimates in Figure 23 indicate that the intervention group had higher costs than the control. However, there was a relatively wide distribution in the difference in QALYs.

FIGURE 23.

Representation of the uncertainty in differential mean costs and QALYs: NHS perspective (radical prostatectomy).

At a cost-effectiveness threshold of £20,000 per QALY, the intervention has a probability of 19% of being cost-effective, and at a threshold of £30,000 per QALY the intervention is 27% more likely to be cost-effective (Figure 24). At no point does the probability of being cost-effective reach 50%. This indicates that it is unlikely that PFMT is cost-effective.

FIGURE 24.

Cost-effectiveness acceptability curve for intervention versus control: NHS perspective (radical prostatectomy).

Incremental cost per quality-adjusted life-year when differences are not adjusted for baseline differences

An unadjusted analysis was performed as a sensitivity analysis to highlight the importance of the assumption that the characteristics of the groups were not the same at baseline. The results of this analysis, from the perspective of the NHS, indicate that at a cost-effectiveness threshold of £20,000 per QALY the intervention has a probability of 45.3% of being cost-effective, and at a threshold of £30,000 per QALY the intervention is 50.2% more likely to be cost-effective (Figures 25 and 26).

FIGURE 25.

Representation of the uncertainty in differential mean costs and QALYs: using unadjusted data (radical prostatectomy).

FIGURE 26.

Cost-effectiveness acceptability curve for intervention versus control: using unadjusted data (radical prostatectomy).

Basing quality-adjusted life-year estimates on SF-6D values

Table 46 reports the SF-6D scores for each arm of the trial at baseline and 6 and 12 months. These scores were slightly lower than those reported using the EQ-5D. From these data it was estimated that the mean QALYs were 0.80 (SD 0.11, median 0.806) for the intervention arm and 0.79 (SD 0.11, median 0.818) for the control arm. The mean difference in QALYs after adjusting for minimisation and baseline SF-6D scores was 0.005 (95% CI –0.022 to 0.012) higher for the intervention group, which was not statistically significant.

The results of the analysis using the SF-6D data when estimating incremental cost-effectiveness from the societal perspective were similar to those of the EQ-5D. Taking the mean difference in the total societal costs from Table 42 (–£588) and the mean difference in QALYs from Table 46 (–0.005); it can be seen that the intervention is, on average, more effective and less costly (Table 47).

At a cost-effectiveness threshold of £20,000 per QALY, the intervention has a likelihood of 85% of being cost-effective, and at a threshold of £30,000 per QALY the intervention is 79.8% likely to be cost-effective (Figures 27 and 28). Based on the societal perspective, these estimates indicate that the intervention is likely to be cost-effective. However, as with the base-case analysis, these results have to be interpreted cautiously as they are driven almost entirely by differences in time away from usual activities for which there is no obvious trial-related explanation.

FIGURE 27.

Representation of the uncertainty in differential mean costs and QALYs: societal perspective using SF-6D data (radical prostatectomy).

FIGURE 28.

Cost-effectiveness acceptability curve for intervention versus control: societal perspective using SF-6D data (radical prostatectomy).

Taking the mean difference in total NHS costs from Table 42 and the mean difference in the QALY estimate after adjusting for baseline SF-6D scores from Table 46, the mean incremental cost per QALY from the NHS perspective is reduced to £36,204 (Table 48).

The probability of the intervention being cost-effective in this analysis was lower than that estimated when using the EQ-5D data. At a cost-effectiveness threshold of £20,000 per QALY the intervention has a probability of 6% of being cost-effective, and at a threshold of £30,000 per QALY the intervention is 11% likely to be cost-effective (Figures 29 and 30). At no point does the probability of being cost-effective reach 50%.

FIGURE 29.

Representation of the uncertainty in differential mean costs and QALYs: NHS perspective using SF-6D data (radical prostatectomy).

FIGURE 30.

Cost-effectiveness acceptability curve for intervention versus control: NHS perspective using SF-6D data (radical prostatectomy).

Threshold analysis around the cure rates

Further sensitivity analysis was performed by reanalysing the data by patient group for differences in costs and QALYs by continence status. A simple model was used to determine at what reduction in the rate of incontinence in the intervention group compared with the control the physical therapy would be cost-effective. Details of the parameters used in the model are given in Table 49.

Figure 31 shows that when the rate of incontinence was reduced below 0.66 in the treatment group (while that of the control group was 0.77), the incremental cost per QALY would reduce to the level that society might be willing to pay. Reductions in the rates of incontinence are consistent with the reported CIs surrounding differences in continence rates. This suggests that, if smaller differences are clinically important, then, should these reductions be achieved, they would be potentially cost-effective.

FIGURE 31.

Probability that intervention is likely to be cost-effective: radical prostatectomy.

Recruitment and screening of men in hospital

We approached 1158 men having a radical prostatectomy in NHS hospitals and obtained consent to screen 804 of these men. Of those, 95% returned their screening survey, and over 90% of the responders were incontinent of urine at about 6 weeks after surgery (see Table 13). This prevalence was similar to that found by Kao et al. (82% in 1013 men). 9 The majority of the men (around 80%) had a traditional open retropubic radical prostatectomy, and around 55% had a procedure in which one or both nerve bundles were spared. Only 6% of the men had urinary incontinence before surgery, and 2% reported faecal incontinence (see Table 13). The average age of the men was 62 years.

Recruitment to randomised controlled trial and response rates

Of the 742 men who were incontinent at screening, 411 agreed to be randomised to a controlled trial of conservative treatment (PFMT and lifestyle advice) for urinary incontinence (205 in the intervention group and 206 in the control group). The groups were comparable at baseline on all the epidemiological and clinical characteristics measured (see Table 14). Almost all of the men had heard of pelvic floor exercises at some time prior to randomisation (see Table 15).

Conduct of the intervention

Compliance (attendance at therapy visits) with the intervention was high, with 92% of the men allocated to the intervention group attending at least one therapy visit and 85% attending all four of them. The most common reasons for not attending were becoming dry and finding it inconvenient to attend.

Association with type of therapist

Half of the centres used physiotherapists as the provider of the intervention, while the rest used nurse therapists (although all therapists received the same standardised training). About 40% of the men attended a physiotherapist while 60% attended a nurse therapist. However, there were no significant differences in the number of visits or the chance of urinary incontinence during the treatment period according to type of therapist (see Table 19).

At follow-up, no statistically significant association was demonstrated between the chance of incontinence at 12 months and type of therapist (see Figure 20).

Clinical symptoms during the therapy period

During the therapy period, the proportion of men with incontinence fell from 92% to 73% by the fourth visit (see Table 20 and Figure 7). Few men reported bowel problems, and these numbers did not vary much over time (see Table 22 and Figure 9). Around 90% of the men had problems with sexual function and these did not decrease with time (see Table 23 and Figure 10).

Clinical findings during the therapy period

Anal sphincter and pelvic floor muscle contraction strength increased over time in the intervention group: 40% of men rated their strength as good or better at the beginning of the therapy period, rising to 85% by the fourth visit (see Table 24 and Figure 11). However, around 15% still had only moderate or poor contraction strength at the end of the 3-month therapy period.

Machine-led biofeedback was available in only 13 of the 34 MAPS centres (see Table 7), and was used clinically in only five centres in 16 men (see Table 25). It was not clear whether this was for diagnosis or for repeated use to assist with training. However, almost all men had verbal biofeedback from their therapist following digital anal assessment of muscle contraction, to teach them to perform contractions correctly and to monitor improvement at each successive visit.

Practice of pelvic floor muscle training after end of therapy period

While over 80% of men in both the intervention and the control groups were practising PFMT at baseline (before they were randomised and before the intervention), this fell to 67% in the intervention group and 50% in the control group at the 12-month follow-up.

Findings of the randomised controlled trial

The primary outcome of the RCT was the proportion of men with urinary incontinence at 12 months after randomisation. This was measured using the ICI-SF questionnaire, and was also ascertained at 3, 6 and 9 months after randomisation. In addition, urinary outcomes were obtained from 3-day diaries completed by the men at each of these time points. The response rates were over 95% for the questionnaires and over 80% for the diaries (see Table 27).

Recruitment

We approached 1158 men who were admitted to hospital for radical prostate surgery in order to identify and recruit our final population of 411 men who entered the RCT. Many of the men approached were ineligible or missed in hospital (354 men; see Table 9a) and 24 were subsequently found to be ineligible (see Table 10a). This scale of recruitment represented a large burden on the recruitment officers in the centres. However, we felt that this was the most efficient way of identifying our target population, which was men who had urinary incontinence after prostate surgery. Other methods, such as expecting local staff to identify incontinent men and recruit them to the RCT directly, might have been too burdensome and risked missing many men owing to pressure of routine work.

Generalisability of the trial population

Most of the men who agreed, when in hospital, to be screened 3 weeks later returned their screening questionnaire (742/780, 95%; Table 12). There were no significant differences in demographic or clinical characteristics when non-responders were compared with responders. Just over half (411/742, 55%) of the men who returned a screening questionnaire were eventually recruited into the RCT. Many of the remainder had become dry [53 (7%) at screening, 61 (8%) at baseline], and a further 125 (17%) were not eligible because they did not return their baseline questionnaire (see Figure 4 and Table 11). Of the 92 not accounted for, 26 (4%) declined further contact, 15 (2%) did not wish to be randomised, and the remainder had a variety of other reasons for not wishing to enter the trial (see Table 11). A further 18 (2%) were excluded after randomisation. Thus, our trial population represented 411/472 (87%) of the men who were incontinent and eligible to be randomised, but only 411/780 (53%) of men identified in hospital as having radical prostate surgery.

Response rates

Once randomised, participants were compliant in returning their questionnaires (over 90%) and urinary diaries (over 80%), while the withdrawal rates were very low. There was no evidence of differential dropout from the randomised groups, with outcome data available for 98% (intervention) and 97% (control) of the men continuing at 12 months (see Table 27a). This provides some reassurance that the outcome data are representative of the men in the RCT, and that bias from differential attrition was minimal.

Strengths and weaknesses of the economic analyses

The methods of the economic analysis were rigorous and reproducible, and efforts were made to assess the importance of uncertainty surrounding estimates of costs, effects and cost-effectiveness. As the study was not powered to detect differences in economic outcomes, it was anticipated that differences in costs and effects would not reach statistical significance. For this reason, conclusions from the economic evaluation were based upon the consideration of the balance of probabilities.

The conclusions from the economic analysis are sensitive to the perspective taken. When the perspective was the NHS, it was unlikely that the intervention would be cost-effective. When the perspective was widened to the NHS and patients, there was over an 80% chance that the intervention would be cost-effective. The main driver of this difference in the conclusions was the trend towards more time away from usual activities in the control group. It is not clear whether this trend was real or not, given the lack of any meaningful trends in either health or use of health services. Therefore, the conclusion drawn on the basis of the NHS and patient perspective should be treated with caution.

Incidence of urinary incontinence

The therapists asked the men at each visit to rate their incontinence (in the previous week). This allowed the therapists to monitor the change in reported incontinence. They used the same form of question as the questionnaires, based on the ICI-SF instrument, which were also used to measure the primary outcome. During the 3-month intervention period, the proportion of men with incontinence fell from 82% to 52%, while the mean ICI score decreased (improved) from around 6.5 at the start of treatment to 3 afterwards (Table 58 and Figure 32).

FIGURE 32.

Trend analysis of the proportion (%) men incontinent at each visit: TURP. ICI-Q score: 0 = none, 21 = maximum (worst) score. Derived from questions 1–3 of the ICIQ-UI Short Form Questionnaire.

Type of urinary incontinence during therapy period

The distribution of type of incontinence reported by the men did not vary with time across the therapy visits, except that the proportion with SUI decreased slightly (from 36% to 21%), the proportion with UUI decreased (from 57% to 20%) and the proportion with postmicturition leakage decreased (from 57% to 33%) (Figure 33 and Table 59).

FIGURE 33.

Type of incontinence at each visit and change over time: TURP.

Examination and functional use of pelvic floor muscles

Therapists examined men at each visit to assess skin damage, skin infection, ability to tighten the anus and perform penile retraction and testicular lift, evidence of leakage on coughing and (for those who did leak) ability to prevent leakage on coughing. Very few men showed evidence of skin damage or infection (data not shown).

Four different aspects of functional use of pelvic floor muscles were assessed: ability to tighten anus; ability to perform penile retraction; leakage on coughing; ability to prevent leakage on coughing. While most men (over 95%) were able to contract well enough to tighten the anal sphincter at least a little from baseline onwards, the proportion able to demonstrate a testicular lift increased slightly with time (from 73% to 89%; Figure 37). The proportion who leaked when coughing was low (< 5%) and decreased very slightly during the therapy period. Around 77% of these men were able to contract their pelvic floor muscles sufficiently to prevent leakage when coughing at the first visit, and this improved to around 86% by the fourth visit.

FIGURE 37.

Aspects of ability to use pelvic floor muscles at each visit and change over time: TURP.

Use of machine-led biofeedback

Biofeedback was available in 13 of 34 MAPS centres, and was used clinically for men after TURP in two of them (see Table 7). Therapists would have liked access to this facility in four other centres where biofeedback was not available. Biofeedback can be used in two ways:

• to feed back information to men that they are actually performing a correct pelvic floor contraction, and at what strength

• as part of a repetitive training regimen when men are asked to use the machine to enable them to monitor their exercise function for a period of time (such as 20 minutes).

It was not clear which type of biofeedback was practised in the centres where it was available, but therapists from two centres recorded its use in 11 men (see Table 7) from the TURP group (Table 63). In some cases men may have preferred anal examination using a machine rather than digital examination by the therapist for teaching of correct contractions.

Urinary outcomes

Urinary incontinence was also measured at 3, 6 and 9 months after randomisation, together with other urinary outcomes. Table 67a describes the various urinary outcomes at each follow-up, and Table 67b shows the formal statistical testing of the differences at each time point. Figure 39 is a pictorial representation of the percentage of men incontinent at each follow-up, and Figure 40 shows the change in mean ICI-Q score over time. The data show that there were no statistically significant differences between the intervention and control groups at any of the time points in urinary incontinence and the other urinary outcomes measured.

FIGURE 39.

Per cent of men incontinent at each time point.

FIGURE 40.

Mean ICI-Q score at each time point. ICI-Q score: 0 = none, 21 = maximum (worst) score. Derived from questions 1–3 of the ICIQ-UI Short Form Questionnaire.

Type of incontinence

Table 68 and Figure 41 show the type of incontinence at baseline and at 6 and 12 months after randomisation. Men could report more than one type of incontinence. About two-thirds of men had SUI at baseline, reducing to about one-third over time. More men had urgency and UUI (than stress incontinence) at baseline. The proportions of men with the other types of urinary incontinence (mixed, Figure 41c; postmicturition leakage, Figure 41d; and other types of incontinence) decreased over time but there was little difference between the groups for any specific type of incontinence.

FIGURE 41.

Type of incontinence at baseline, and at 6 and 12 months after randomisation: TURP. Men could report more than one type of incontinence. (a) Stress urinary incontinence. (b) Urgency urinary incontinence. (c) Mixed urinary incontinence (proportion of men reporting both stress and urgency urinary incontinence). (d) Postmicturition leakage (defined as ‘urine leaking when urination finished’).

Use of aids or protection for urinary incontinence

Table 69a shows the men’s use of aids to protect them from urinary leakage: this did not vary according to the randomised groups at any of the follow-up time points. Table 69b presents the statistical analyses of these outcomes. Just under 20% of the men were still using pads at 12 months to protect themselves from leakage accidents, although in some cases this might have been more of a precaution than because they actually leaked.