Pelvic floor exercises to improve recovery and quality of life after rectal cancer surgery: plan for a clinical trial

Randomization will be performed 3 months after the LAR in patients without stoma construction, and in case of a temporary stoma, 6 weeks after stoma closure. The coordinating investigator will conduct the randomization procedure, using Castor EDC (www.castoredc.com↗). Stratification will be done in variable blocks. It will be stratified for gender and radiation because of frequent sphincter function impairment due to birth trauma in women and the association between radiation and sphincter and bowel dysfunction. The coordinating investigator will be responsible for communication with the pelvic floor specialized physiotherapists and will inform them about referral and the allocated intervention.

The physiotherapists will be informed about the pre- and postoperative medical history of the patient. Surgeons and physiotherapists are blinded in terms of outcomes of questionnaires taken before the start of PFR (measurement M2, Fig. 1). Complete blinding of the allocated intervention for patients and the participating physiotherapists is impossible. Unblinding will not occur, since the design is open label. The coordinating investigator, who will also be involved in data analysis, is not blinded for the allocation since he will inform the physiotherapists about referral and the allocated intervention. Additional data analysists will be blinded for allocation.

The primary outcome is based on the Wexner score. The dependent variable is the severity of FI measured by this score. The independent variable is the intervention or the standard treatment. The difference between the intervention and control group will be analyzed by an analysis of covariance (ANCOVA), with the baseline measurement as covariate. This test is preferred over the T-test since this analysis reduces the within-group error variance. Therefore, the precision of the treatment estimate is increased, and the length of the confidence interval is reduced [18]. The M3 measurement will be valid as the primary outcome. The M4 measurement, which is used to assess sustainability of results, will be influenced by the patients of the control group who will choose to undergo PFR after finishing their primary endpoint measurements (after 3 months of standard treatment, i.e., in case of severe FI). For statistical analysis, the effect between the intervention group and control group will be evaluated with and without this group to measure the influence of this group on the M4 measurement. The same analysis of covariance (ANCOVA) will be used.

The dependent variable is the degree of QoL/anorectal outcome. The independent variable is the intervention or the standard treatment. The difference between the effect in the intervention group and the control group will be analyzed by analysis of covariance (ANCOVA) as well. No additional analyses, such as subgroup analysis, are written out in advance.

Primary study outcome will be based on intention to treat analysis. In addition, per protocol analysis will be performed. This analysis will be done to avoid the effects of crossover, protocol violation, and dropout.

The anonymized data will be stored in Castor EDC. Only the coordinating and principal investigators will have access to the key of the code that matches participants to study data. Most patients will fill in their questionnaires themselves online. Only those without email will fill in the paper version of the questionnaires.

No specific plans, besides an active attitude towards the participating centrums, are made to promote participant retention and complete follow-up. If patients fail to start/continue their allocated intervention (Pelvic Floor Rehabilitation, since failure of control group is not possible), data will be collected according to the study protocol for use in per protocol analysis.

The patient information will include the medical history, gender, age, tumor height (MRI and scoping), height and type of the anastomosis, type of surgery (laparoscopic, open or robotic), construction of a temporary stoma, peri- and postoperative complications, time of surgery, blood loss, time of hospitalization, and clinical and pathological TNM staging.

This questionnaire incorporated various Rome IV criteria and scoring tools for severity of constipation and fecal incontinence. Overall reproducibility of the Groningen DeFeC questionnaire is acceptable and its validity is good [19]. This makes it a feasible screening tool for defecation disorders. The primary outcome, the Wexner score, will be derived from this questionnaire. The distribution of subtypes and symptoms of fecal incontinence in the general Dutch population has already been investigated with the use of this questionnaire [20].

The Wexner score allows the severity of fecal incontinence to be assessed using five questions (Table 2). The minimal score patients can obtain is 0 (continent), and the maximum score is 20 (highest severity of incontinence). This score describes the type and frequency of the incontinence and impact on daily life. No data exist about the internal consistency and the criteria/content validity of this score. The intraclass correlation coefficient (ICC) of the Wexner score is good to excellent (ICC 0.75). The construct validity is R(Pearson) = 0.78 (correlation with Vaizey score) [21]. The Wexner score has an important association with the Fecal Incontinence Quality of Life score (r = − 0.45), but a weak correlation with changes in the EQ-5D has been described [22].

The LARS score is a validated scoring system for bowel dysfunction after low anterior resection for colorectal cancer. The Dutch version was recently validated [23]. This score covers the five most bothersome issues in terms of prevalence and impact of QoL, namely incontinence for flatus and liquid stool, frequency, clustering, and urgency. Scores range from 0 to 42, with subdivisions in three categories: no (0–20), minor (21–29), and major LARS (30–42). A statistically significant association exists between a higher LARS score and an impaired QoL. The test-retest reliability of the LARS score is good, with an interclass correlation coefficient of 0.79 [23].

The FIQL score is a condition-specific questionnaire on quality of life and consists of four multi-item subscales: lifestyle (10 items), coping/behavior (nine items), depression/self-perception (seven items), and embarrassment (three items), for a total of 29 items [24]. Each item has four answer opportunities, with a score of 1 to 4, where 1 indicates a low quality of life and 4 a high quality of life.

The internal consistency (Cronbach’s α) is good (subscale 1, 0.96; subscale 2, 0.96; subscale 3, 0.88; and subscale 4, 0.80). The total FIQL has an adequate internal and external responsiveness (standardized response mean = 0.5, r = − 0.48, and area under the curve = 0.765). The intraclass correlation coefficient is good to excellent: 0.80 (embarrassment) to 0.93 (lifestyle) [25].

Quality of life will be measured by the EORTC Colorectal Quality of Life Questionnaire QLQ-CR29. The ICC of the QLQ-CR29 is good to excellent (Intraclass Correlation Coefficient: 0.78) [26, 27]. The Internal consistency of this questionnaire reaches the criterion of 0.70 [26].

This log provides insight into the defecation pattern, type of consistency (Bristol Stool Form Scale) [28], feeling of urge to defecate, episodes of fecal incontinence, and the use of diapers. Also, the use of medication will be noted. The patient will keep up this log 5 days per week after LAR or temporary stoma from the moment of randomization until the end of the intervention (3 months). This log will be used to allow discussion about the number of treatment sessions that will be needed to obtain an effect of PFR. Fecal incontinence occurs often in combination with incontinence for urine. In case of urine incontinence, the patient has to keep up the severity and episodes of this urinary incontinence. The incontinence will namely influence the costs (i.e., use of diapers).

The cost analysis exists of two main parts. First, at the patient level, the volume of care will be measured prospectively over the time path of the study using patient-level diaries (on fecal incontinence-related health care consumption), completed, if necessary, by data from the patient training facility’s administration system. The diary will be developed in a way that it structures and makes uniform the health care consumption of a fecal incontinent patient for this particular target population. Second, per item of health care consumption, the standard cost will be determined using the guideline for performing economic evaluations (ZINL, 2015), which will be completed with the total costs of items via activity-based costing. Productivity losses for patients will be assessed using a patient-based diary complemented with an interview on a 3-month recall basis between the researcher and the patient (at baseline and at the 3- and 6-month follow-ups). The friction cost-method will be applied following the Dutch guidelines (ZINL, 2015). In addition, travel time to the training site and related costs will be considered (also on the basis of the diary). Differences in costs between both groups will be evaluated using regression-based techniques.

The effect analysis adheres to the design of the randomized controlled trial and measures at baseline and at fixed points along time path/follow-up of the clinical trial (see design clinical trial). For measurements on the quality of the health status of the patients, a validated so-called health-related quality of life (HR-QoL) instrument—the EuroQol-5D (EQ-5D) [29]—will be used. This HR-QoL instrument will be completed by the patients and is available in a validated Dutch translation. The EQ-5D-5 L which is used in this study is a generic HR-QoL instrument comprising five domains: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [30, 31]. The EQ-5D-5 L index is obtained by applying predetermined weights to the five domains. This index gives a societal-based global quantification of the patient’s health status on a scale ranging from 0 (death) to 1 (perfect health). Patients will also be asked to rate their overall HR-QoL on a visual analog scale (EQ- 5D VAS) consisting of a vertical line ranging from 0 (worst imaginable health status) to 100 (best imaginable). For the EQ-5D-3 L, the version used before the introduction of the EQ-5D-5 L, a weak correlation was known to exist with changes on the Wexner score [22]. The same can be expected from the EQ-5D-5 L.

BIA will be conducted to assess how healthcare budgets will be influenced when offering the standardized pelvic floor rehabilitation program. This program is anticipated to allow savings of €191 in direct medical cost and €280 in productivity gains for the average patient on a yearly basis. The patient-based budget impact will be extrapolated to the population level. Although the BIA relies heavily on the findings from the economic evaluation described above, it will be conducted from various perspectives, such as from the broader perspective of the public purse down to the narrow healthcare perspective. To that end, use will be made of a health economic (decision analytical) model in which uncertainty will be taken into account. Deterministic uncertainty concerning BIA input, such as perspective, pricing parameters, time horizon, uptake, etc., will be dealt with by generating the budget impact as a series of sensitivity analyses covering a relevant range of costs. In general, for the BIA we adhere to the new guideline for performing economic evaluations in health care (ZINL, 2015).

Patients in the control group will get the standard treatment that currently is used in daily practice. This standard treatment for postoperative FI consists of the prescription of bulking agents (i.e., Metamucil, psyllium fibers, Volcolon, or Normacol). Bulking agents can be beneficial to the consistency of the stool and therefore diminish soiling problems. These bulking agents may be used once or twice per day depending on the severity of FI. Standard oncologic follow-up after LAR at the surgical outpatient clinic will be provided. In case of severe FI and/or failing of standard treatment, cross over to the intervention group will be allowed. After finishing their allocated control group period, these patients with severe complaints might wish to undergo PFR. This will be allowed, and these patients will receive PFR according to the study protocol. After 12 sessions, patients will be asked to fill in another questionnaire, which will allow them to evaluate whether the length of the post-operative period after which PFR has been applied, influenced the patient’s outcomes.

Treatment of patients in the intervention group will consist of four modalities of pelvic floor rehabilitation in addition to the standard treatment:

1) pelvic floor muscle training to a) increase the maximum strength of muscle contraction, b) extend the time of muscle contraction, and c) improve the coordination of contraction of the pelvic floor muscles; 2) biofeedback, which is a behavioristic therapy that allows the patient to become conscious of the contraction and relaxation of the pelvic floor muscles and uses an anal electromyography probe; 3) electrostimulation, which can improve the effectiveness of the contraction strength of the pelvic floor muscles and uses the same anal electromyography probe as used for biofeedback; and 4) training with a rectal balloon to simulate the urgency to defecate, which is a method that enables the patients to train to retain stool, enabling the patient to retain a larger stool in the rectum.

The intervention will be performed by specialized pelvic floor physiotherapists, registered in the Dutch KNGF-NVFB register. These physiotherapists will be chosen based on the location of their practices. For the comfort of patients included in this study, the close location of the practices to the participating hospitals is important. Therefore both hospital-based and private practices were allowed to participate. The physiotherapists will be obliged to attend a course during which the treatment protocol and the case report form will be explained. The coordinating investigator will guide this course and will train the physiotherapists. During this training session, the correct way of digital rectal examination to assess the function of the pelvic floor, the use of biofeedback, electrostimulation, and the rectal balloon will be explained and illustrated using videos. An internationally accepted protocol, The Pelvic Floor Assessment Protocol, will be used for the digital rectal examination performed by the physiotherapist (International Continence Society 2006, Messelink et al.). The physiotherapist will also be instructed on how to use the case report form and report any adverse events during the 12 sessions of PFR. Since a patient’s adherence to PFR is a common problem, an instruction on how to enlarge this adherence to therapy will be provided [21, 22]. The complete treatment trajectory for both the control and the intervention group will be recorded in a standardized protocol, which has already been developed by clinical experts in the field [8].

Patients will start with the intervention trajectory 3 months after LAR or 6 weeks after stoma closure. During the next 3 months, they will undergo 12 treatment sessions, once per week. The first session will take 45 min, and the following sessions 30–35 min. During every treatment, the physiotherapist will complete a case report form. This case report form contains all information concerning pelvic floor muscle training, biofeedback, electrostimulation, peri-anal examination, and digital rectal examination. Digital rectal examination of pelvic floor functionality will be used to assess the ability to consciously contract the pelvic floor muscles and to quantify the strength of the contraction [8, 32]. The use of medication according the standard treatment will be reported also in the case report form. All patients randomized to the intervention group will be instructed by their physiotherapist on how to exercise the pelvic floor. The patients will be thought to selectively generate voluntary contractions of the puborectal muscle and the external anal sphincter and also how to relax these muscles and to avoid co-contractions of other muscles. The pelvic floor has to satisfy requirements of maximal strength, progression of the duration of the strength, and progression in timing and coordination of the contraction [8, 32]. Additionally, folders will be provided, with illustrations and short descriptions of exercises, which have to be performed at home three times per day, preferably at fixed time points [8, 32]. The success of PFR depends greatly on the motivation, willingness, and self-discipline of patients to exercise at home [8, 32].