Neurourology and Urodynamics
Pelvic Floor Muscle Training for Bowel Dysfunction Following Colorectal Cancer Surgery: A Systematic Review 1
Kuan-Yin Lin,1* Catherine L Granger,1,2 Linda Denehy,1 and Helena C Frawley1,3 Department of Physiotherapy, School of Health Sciences, The University of Melbourne, Melbourne, Australia 2 Department of Physiotherapy, Royal Melbourne Hospital, Melbourne, Australia 3 Allied Health Research, Cabrini Health, Melbourne, Australia
Aims: To identify, evaluate and synthesize the evidence examining the effectiveness of pelvic floor muscle training (PFMT) on bowel dysfunction in patients who have undergone colorectal cancer surgery. Methods: Eight electronic databases (MEDLINE 1950–2014; CINAHL 1982–2014; EMBASE 1980–2014; Scopus 1823–2014; PsycINFO 1806–2014; Web of Science 1970–2014; Cochrane Library 2014; PEDro 1999–2014) were systematically searched in March 2014. Reference lists of identified articles were cross referenced and hand searched. Randomized controlled trials, cohort studies and case series were included if they investigated the effects of conservative treatments, including PFMT on bowel function in patients with colorectal cancer following surgery. Two reviewers independently assessed the risk of bias of studies using the Newcastle-Ottawa Scale (NOS). Results: Six prospective non-randomized studies and two retrospective studies were included. The mean (SD) NOS risk of bias score was 4.9 (1.2) out of 9; studies were limited by a lack of non-exposed cohort, lack of independent blinded assessment, heterogeneous treatment protocols, and lack of long-term follow-up. The majority of studies reported significant improvements in stool frequency, incontinence episodes, severity of fecal incontinence, and health-related quality of life (HRQoL) after PFMT. Meta-analysis was not possible due to lack of randomized controlled trials. Conclusions: Pelvic floor muscle training for patients following surgery for colorectal cancer appears to be associated with improvements in bowel function and HRQoL. Results from nonrandomized studies are promising but randomized controlled trials with sufficient power are needed to confirm the effectiveness of PFMT in this population. Neurourol. Urodynam. # 2014 Wiley Periodicals, Inc. Key words: biofeedback; bowel function; colorectal neoplasms; pelvic floor; physiotherapy; rehabilitation
INTRODUCTION
Colorectal cancer (CRC) is the third most common cancer in males and the second most common cancer in females worldwide, with an estimation of 1.3 million new cases diagnosed each year.1 Colorectal cancer is responsible for 8.5% of all cancer deaths yearly.1 The five-year survival for CRC is around 65%, indicating that most CRC patients today can expect to live many years after their diagnosis.2,3 Surgery is usually the preferred treatment for patients with CRC, with approximately 54% to 85% of patients receiving surgery as part of their treatment.4 Several surgical procedures are used for CRC (the choice of which depends on the cancer site, stage, and patient age). For colon cancer, these include open colectomy, polypectomy, and local excision; for rectal cancer: polypectomy, local excision, local transanal resection, transanal endoscopic microsurgery, total mesorectal excision, low anterior resection, ultra-low anterior resection, proctectomy, and abdominoperineal resection.5,6 Radiotherapy and chemotherapy are used in conjunction with surgery for patients with advanced CRC.6 Although more patients with CRC live longer after diagnosis due to early detection and improved treatments, surgical procedures and chemoradiation for CRC may impact on normal pelvic floor function through damage to muscular, neural, and fascial structures.7–9 Post-operatively patients frequently complain of impaired bowel function, such as incomplete evacuation, excessive flatus, fecal urgency, straining at stool, perianal soreness or itching, bloating, and fecal incontinence.10 Among all bowel dysfunction, fecal incontinence in particular is a prevalent symptom after CRC surgery with a reported incidence from 3.2% to 79.3%.11 The wide range in incidence of fecal #
2014 Wiley Periodicals, Inc.
incontinence may be explained by differences in definition and research methodology. Although one previous study has shown that bowel dysfunction such as urgency and tenesmus may resolve by 12 months postoperatively, fecal incontinence increases significantly after one year.12 As a severely distressing and embarrassing symptom, bowel dysfunction may significantly impact on patients’ health-related quality of life (HRQoL) and cause considerable burden to patients, the healthcare system, and society. A wide range of physical therapies (e.g., pelvic floor muscle training [PFMT], weighted vaginal cones, biofeedback, electrical stimulation) have been used in the treatment of bowel dysfunction.13,14 Of these, PFMT is a highly recommended conservative management in the treatment of fecal incontinence in non-cancer populations.15–17 Previous narrative
Abbreviations: CINAHL, Cumulative Index to Nursing and Allied Health Literature; CRC, colorectal cancer; EMBASE, Excerpta Medica Database; FACT-C, Functional Assessment of Cancer Therapy – Colorectal; FIQL, Fecal Incontinence Quality of Life; HRQoL, health-related quality of life; MCIS, Modified Cleveland Incontinence Score; MEDLINE, Medical Literature Analysis and Retrieval System Online; NOS, Newcastle-Ottawa Scale; PEDro, Physiotherapy Evidence Database; PFM, pelvic floor muscle; PFMT, pelvic floor muscle training; PRISMA, Preferred Reporting Items for Systematic Reviews; SF-36, Short Form 36 questionnaire; SPSS, Statistical Product and Service Solutions; WIS, Wexner Incontinence Scale. Research conducted at: University of Melbourne Potential conflicts of interest: Nothing to disclose. *Correspondence to: Kuan-Yin Lin, MSc, PT, Alan Gilbert Building, Level 7, 161 Barry Street, Carlton 3053 Victoria, Australia. E-mail: [email protected] Received 20 March 2014; Accepted 16 June 2014 Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/nau.22654
2
Lin et al.
literature reviews have also suggested that pelvic floor exercise with biofeedback may be tried as a treatment option to improve muscular strength of the external anal sphincter, rectal sensitivity, and pelvic floor muscle (PFM) coordination for cancer patients with pelvic floor or bowel symptoms following surgery.14,18,19 While a recent systematic review reported that conservative therapies such as pelvic floor re-education and colonic irrigation can improve anorectal function in patients undergoing rectal resection for cancer and non-cancer diagnoses,20 an important literature gap still exists in terms of the benefit of PMFT in patients specially with CRC. Individuals with CRC behave differently to those undergoing colorectal surgery for non-cancer diagnoses in terms of symptoms, surgical approaches, and side effects of chemotherapy and radiotherapy. Therefore, the aim of this review was to systematically investigate the effect of PFMT on bowel dysfunction in patients after surgery for CRC. MATERIALS AND METHODS
This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines. No protocol for this review has been registered or published.21 Search Strategy
Electronic databases, MEDLINE (1950–2014), CINAHL (1982– 2014), EMBASE (1980–2014), Scopus (1823–2014), PsycINFO (1806–2014), Web of Science (1970–2014), Cochrane Library (2014), and Physiotherapy Evidence Database (1999–2014) were searched in March 2014 by one reviewer (KYL) using different combinations of the following terms: pelvic floor muscle training; rehabilitation; physical therapy modalities; biofeedback psychology; pelvic floor exercise; kegel; biofeedback; bowel function; defecation; fecal incontinence; gastrointestinal function; total mesorectal excision; sphincter-saving surgery; anterior resection; abdominoperineal resection; colectomy; colorectal surgery; sphincter-preserving surgery; colorectal neoplasms; rectal neoplasms; colonic neoplasms. An example of search strategy is presented in Appendix 1. Electronic databases were accessed via The University of Melbourne, Australia, with the last search run on March 10, 2014. No restrictions on the publication date were imposed. The search was limited to articles written in English. The reference lists of identified articles were also hand searched.
study to have CRC, or data to be reported separately for the cancer participants. Types of intervention. Studies which included PFMT with or without biofeedback were eligible. This included all types of visual, sensory, or auditory biofeedback. According to the 5th International Consultation on Incontinence, PFMT (also known as Kegel exercise training) for treating fecal incontinence involves the contraction of PFM including the external anal sphincter and puborectalis.17 For the purpose of this review, PFMT/Kegel exercise and anal sphincter exercise will be considered together with the use of biofeedback for the purpose of PFM strengthening (PFMT combined with anal canal pressure or intra-anal electromyograph feedback device), sensory training (introducing a balloon-tipped catheter into the rectum), or coordination training (employing a pressure transducer to provide simultaneous feedback in intra-rectal and intra-anal pressure during PFM contraction) made explicit.17,23 Studies that used solely electrical stimulation were not eligible for inclusion. Types of outcomes. Studies were eligible for inclusion if their primary outcome was bowel function measured by a patientreported outcome measure (such as a bowel diary, questionnaire, patient report) or anal manometry. Two reviewers (KYL and CG) independently assessed studies based on titles and abstracts for inclusion. Full texts of potentially relevant studies were retrieved and reviewed by the two reviewers to evaluate the eligibility. Disagreements could be resolved by a third independent reviewer (HF) although this was not required. Authors were contacted by e-mail if more information was needed to screen an article (this occurred on one occasion). Data Collection Process
An electronic data collection form was specifically developed and utilised. Data were extracted from the included studies by one reviewer (KYL) and crossed-checked by a second independent reviewer (CG). Disagreements were resolved through discussion or with the use of a third reviewer (HF). Data items extracted from the included studies included the first author’s name; year published; number of participants; gender; age; type of surgery; adjuvant therapy; type, duration, frequency, and length of intervention program; and results of reviewed studies. Risk of Bias
Study Selection
Studies were eligible if they met the following inclusion criteria: Types of studies. Quantitative study designs including randomized controlled trials (RCTs), cohort studies and case series were eligible for inclusion. Case studies with only one participant were not eligible. Only studies published in a peer reviewed journal were eligible. Types of participants. Participants of any age, who had undergone surgery for colon or rectal cancer, were eligible to be included in this review. Colorectal cancer was defined as cancer that forms in the tissues of the colon or the rectum, and participants with any stage of CRC were eligible.22 Given the limited number of studies available on this topic, we included studies with mixed cohorts of participants with cancer and non-cancer diagnoses who had undergone colorectal surgery; for inclusion, we required at least 50% of the participants in the Neurourology and Urodynamics DOI 10.1002/nau
As non-randomized studies were the primary study design found, the risk of bias was assessed by the two independent reviewers using the Newcastle-Ottawa Scale (NOS) for nonrandomized studies (Supplemental section). A third review was available if any disagreements arose between the two reviewers (although this was not required), and a consensus score/point was assigned after discussion. Studies were also ranked on the National Health and Medical Research Council Hierarchy of Evidence Scale.24 The scoring criteria used for synthesis of bias risk of included studies were reported as ‘‘poor’’ to ‘‘excellent’’ based on the score calculated in the NOS. Statistics
Kappa statistics and percentage agreement were calculated to establish agreement between reviewers for study selection and risk of bias using SPSS for Windows statistical software
Pelvic Floor Muscle Training in Bowel Cancer 25
package (SPSS Inc., Version 21, Chicago, IL). Values of kappa greater than 0.8 reflect ‘‘excellent’’ agreement.26 Mean and variance of pre- and post- intervention assessments were obtained directly from the study results or calculated using the following equations when not reported: 27,28 Standard deviation ¼ Range=4 Standard deviation ¼ Standard error of the mean ðsquare root of the number of samplesÞ: RESULTS Study Selection
A total of 51 studies were identified following the electronic database search, and an additional four studies were identified though searching reference lists. After exclusion of duplicate records and assessment of titles and abstracts, 10 studies met inclusion criteria and were further assessed with full text. The author of one potentially eligible study29 was contacted to clarify whether the study had included any participants with CRC; the study was subsequently excluded because the diagnosis of participants was unknown. The study of Hwang et al. was excluded because only three of the fourteen participants (21%) in their study had a diagnosis of rectal cancer.30 After assessment of the eligibility of full-text articles, eight studies were finally included for review. Figure 1 shows the flow diagram of study selection. Kappa statistics for agreement between the two independent reviewers on title/ abstracts and full text were 0.94 (percentage agreement, 98.0%) and 1.000 (percentage agreement, 100%), respectively.25
Study Characteristics Methods. No randomized controlled trials were found. Six prospective non-randomized studies31–36 and two retrospective review studies37,38 were included (Table I). Participants’ characteristics. The characteristics of the participants are presented in Table II. A total of 374 participants were included in the studies. Six of eight studies reported data on cancer patients only.34–39 The remaining two studies32,33 included 54% (7/13) and 91% (10/11) cancer patients, respectively. The mean (SD) age of the participants across all studies ranged from 55.0 (11.3) to 67.0 (8.8) years. Of the eight studies, two included patients who had undergone low anterior resection,33,39 three following sphincter-saving surgery,34,35,38 and three receiving mixed types of surgery.32,36,37 Of the three studies with mixed types of surgery, one included patients who had undergone colectomy for colon cancer.37 The majority of studies (75%) included participants who had undergone chemotherapy and/or radiotherapy.32–34,36,38,39 Intervention. The types, duration, and frequency of PFMT programs varied among all studies (Table III). One study used PFMT without biofeedback.35 Six of eight studies provided PFMT with different types of biofeedback: EMG biofeedback for strengthening;39 manometric pressure biofeedback for coordination training;32–34,37 and manometric biofeedback for coordination training, sensory training, and strength training.38 One study used a multimodal rehabilitative program including PFMT with manometric biofeedback for strengthening.36 Length of intervention ranged from three weeks to 20 months. The frequency of outpatient sessions ranged from daily to once weekly, and that of home-based sessions ranged from 3–4 times daily to once daily. In most studies, home Neurourology and Urodynamics DOI 10.1002/nau
3
practice was encouraged via a formal protocol35–39 or with general encouragement.32,33 Outcome measures. All included studies evaluated patientreported measures of bowel function in terms of stool frequency, incontinence episodes, and severity of fecal incontinence. Seven of eight studies assessed stool frequency using a bowel diary,32,33,37 patient report,35,36,38 or gastrointestinal questionnaires.34 Three studies used a bowel diary32,33,37 to assess number of incontinence episodes. Five studies included severity of fecal incontinence as a patient-reported outcome.34,36–39 However, different fecal incontinence severity scales including Modified Cleveland Incontinence Score (MCIS)39 and Wexner Incontinence Scale (WIS)34,36–38 were used. In addition to the patient-reported measures of bowel function, six studies used anorectal manometry to measure anorectal physiological function.32,33,36–39 Three studies examined HRQoL after CRC surgery.34,35,37 The HRQoL was measured by the Functional Assessment of Cancer Therapy – Colorectal (FACT-C) questionnaire (studies n ¼ 1),35 Short Form 36 questionnaire (SF-36) (studies n ¼ 1),34 and Rockwood Fecal Incontinence Quality of Life Scale (FIQL) (studies n ¼ 2).34,37 Risk of Bias and Level of Evidence
Risk of bias of studies is presented in Table I. The agreement using kappa statistics between the two independent reviewers on risk of bias assessed using the NOS was 0.972 (percentage agreement, 98.6%). Minor disagreements were resolved with consensus on 100% of occasions, and on all occasions, scores between reviewers deviated by only one point. The results of risk of bias assessment showed a mean total score of 4.9 (SD 1.2, range 3 to 7). Although most criteria within participant selection and outcome measure categories were satisfied, the majority of studies were limited by lack of a non-exposed cohort, lack of independent blinded assessment, and/or inadequate longer-term follow-up.32–39 Five studies also scored zero points (out of possible two points) in the category of comparability of study group.32,35–38 At best, the level of evidence was III-2 (range III-2 to IV) (Table I). Outcomes Patient-reported outcomes measures. Study results on patient-reported measures of bowel function are presented in Table IV. Six studies showed improvement in stool frequency after intervention,32–34,36–38 and four were statistically significant (P < 0.05).32,34,37,38 Liu et al. reported no significant differences in stool frequency between intervention and control groups after home-based PFMT.35 Only one study reported the long-term follow-up results of stool frequency, which increased at two years after the final treatment session.37 Of the three studies that assessed incontinence episodes, two showed a significant decrease in incontinence episodes (P < 0.05) after PFMT using manometric biofeedback.32,33 Bartlett et al. reported no significant difference in incontinence episodes after PFMT using manometric pressure biofeedback.37 Five studies used MCIS or WIS to measure severity of fecal incontinence.34,36–39 A higher score on MCIS indicates better continence; whereas a higher WIS score indicates worse continence. In the study by Allgayer et al., the MCIS significantly increased (P < 0.001) indicating improved incontinence in 95 participants after a 3-week intensive PFMT.39 The results of studies using WIS also showed a significant improvement in severity of fecal incontinence after intervention (P < 0.05),36–38 except for Laforest et al., who reported no
4
Lin et al.
Fig. 1. PRISMA flow diagram of study selection process.
significant difference in WIS score between intervention and control groups.34 Two studies demonstrated conflicting longterm follow-up results of severity of fecal incontinence.37,39 One study reported maintenance of effects of PFMT with intra-anal EMG biofeedback after 1 year,39 and the other study reported increased severity at 2 years following last treatment session of PFMT using manometric biofeedback, although the severity remained better than at baseline.37 Manometric measures. Six studies used anorectal manometry to measure anorectal physiological function.32,33,36–39 The majority reported no significant differences in mean resting pressure,32,33,36,37,39 maximum squeeze pressure,32,33,36,37,39 maximal tolerable volume,32,33,36,37 volume at initial sensation,32,33,36,37,39 physiological length,32,33,39 and compliance32,33 Neurourology and Urodynamics DOI 10.1002/nau
in patients after intervention. Only one study by Kim et al. showed significant differences in mean resting pressure (P ¼ 0.01), maximum squeeze pressure (P ¼ 0.006), and maximum tolerable volume (P ¼ 0.003) in patients with anterior resection after PFMT with manometric biofeedback38 (Supplemental Table I). Health-related quality of life (HRQoL). Three studies examined HRQoL after CRC surgery.34,35,37 Two studies demonstrated significant improvement in some domains of HRQoL,34,37 and one found a significant difference in total score of HRQoL (P ¼ 0.038) measured using FACT-C after intervention.35 Although not statistically significant, Bartlett et al. reported further improvement in FIQL scores from the final session at two-year follow-up.
Neurourology and Urodynamics DOI 10.1002/nau
Italy
Taiwan
France
Country
Korea
Prospective case control Prospective cohort Prospective cohort
Design
Prospective case series Retrospective case series Prospective case series Prospective case series Retrospective case series
Design
2) Representativeness of the cases
1) Is the case definition adequate?
3) Selection of Controls
3) Ascertainment of exposure
Selection (0–4)
1) Representativeness of the exposed cohort
4) Definition of Controls
() Criteria not satisfied, high risk of bias.
( ) Criteria satisfied.
NOS score
1) Comparability of cases and controls on the basis of the design or analysis
1) Assessment of outcome
1) Ascertainment of exposure
1) Comparability of cohorts on the basis of the design or analysis
Comparability (0–2)
NOS score
Comparability (0–2)
4) Demonstration that outcome of interest was not present at start of study
Abbreviation: NOS, Newcastle-Ottawa Scale; NHMRC, National Health and Medical Research Council, Australia.
Pucciani et al. 2008
Laforest et al. 2012 Liu et al. 2011
Author and year
Kim et al. 2011
Singapore
Ho et al. 1997
Australia
Singapore
Germany
Country
Allgayer et al. 2005 Bartlett et al. 2011 Ho et al. 1996
Author and year
2) Selection of the non exposed cohort
Selection (0–4)
TABLE I. Risk of Bias in Reviewed Observational Trials Using the Newcastle-Ottawa Scale
2) Same method of ascertainment for cases and control
Outcome (0–3)
3) Adequacy of follow up of cohorts
3) Non-Response rate
2) Was follow-up long enough for outcomes to occur
Outcome (0–3)
4
5
7
Total (0–9)
4
6
5
3
5
Total (0–9)
III-2
III-2
III-2
NHMRC grade of evidence
IV
IV
IV
IV
IV
NHMRC grade of evidence
Pelvic Floor Muscle Training in Bowel Cancer 5
6
Lin et al.
TABLE II. Participants in Reviewed Studies
Author and year
Number of participants, (n) %
Prospective non-randomized trials Allgayer et al., 95 2005 Irradiated patients (41) 43.2% Non-irradiated patients (54) 56.8% Ho et al., 1996 13
Irradiated: 28/13
Low anterior resection (95) 100%
Post-op RT (41) 43.2%
Fecal incontinence
Non-irradiated: 33/21
Irradiated: 58.5 11.3 (range: 31.0–76.0) Non-irradiated: 67.0 8.8 (range: 48.0–83.0)
10/3
62.1 16.6
Anterior resection (7) 53.8%
Post-op RT (4) 30.8%
Frequent bowel movement and/episode of incontinence
Post-op RT (2) 18.2%
Incontinence or constipation
Laparoscopic sphincter-saving TME (46) 100%
Pre-op RCT (32) 69.6%
Not specified
Anal sphincter preserving surgery (22) 100%
NR
Frequent defecation and fecal incontinence
Pre-op RT (19) 21.6%
Incontinence
Type of surgery, (n) %
Total colectomy (6) 46.2% Low anterior resection (11) 100%
64.8 10.9
Rehabilitation: 11/11
Rehabilitation: 55.0 11.3 (range: 33.0–78.0)
Rehabilitation group (22) 47.8% Control group (24) 52.2% 22
Control: 15/9
Control: 60.0 11.3 (range: 35.0–80.0)
Exercise: 4/7
Exercise: 55.3 14.3 (range: 27.0–82.0)
Exercise group (11) 50.0% Non-exercise group (11) 50.0% 98
Non-exercise: 8/3
Non-exercise: 65.7 8.7 (range: 45.0–73.0)
Patients: 34/54
Incontinent patients (88) 89.8% Healthy control (10) 10.2%
Healthy control: 4/6
Patients: 59.6 6.8 (range: 46.0–73.0) Healthy control: 57.4 6.3 (range: 44.0–69.0)
Low anterior rectal resection (69) 78.4% Straight coloanal anastomosis (19) 21.5%
Retrospective studies Bartlett et al., 2011
19
10/9
64.1 (95% CI: 47.0–81.3)
Kim et al., 2011
70
49/21
58.1 10.1 (range: 31.0–79.0)
Anterior resection (3) 15.8% Ultra-low anterior resection (10) 52.6% Segmental colectomy (2) 10.5% Proctocolectomy (4) 21.1% Sphincter-saving surgery with TME (70) 100%
Laforest et al., 2012
Liu et al., 2011
Pucciani et al., 2008
11 Incontinence (6) 54.5% Constipation (5) 45.5% 46
Symptoms at baseline
Mean age, year SD
5/6
Ho et al., 1997
Adjuvant therapy, (n) %
Gender, M/F
Post-op RT (34) 38.6%
NR
Bowel dysfunction
Pre-op RT (30) 42.9%
Anterior resection syndrome
Post-op RT (19) 27.1% Pre-op CT (1) 1.4% Post-op CT (25) 35.7% Pre-op plus post-op CT (31) 44.3%
Abbreviations: M, male; F, female; post-op, postoperative; pre-op, preoperative; RT, radiotherapy; CT, chemotherapy; RCT, radiochemotherapy; TME, total mesorectal excision; NR, not reported; CI, confidence interval.
Value calculated (not reported in original study).
DISCUSSION
The aim of our review was to systematically evaluate the evidence for the effect of PFMT on bowel dysfunction in patients who have undergone CRC surgery. No randomized controlled trials were identified in this review, which reflects the urgent need for RCTs in this field to provide high levels of evidence to direct PFMT for patients suffering from this highly distressing problem. Despite the paucity of RCTs and the inherent methodological shortcomings of the included observational studies, this systematic review found that PFMT may improve the patient-reported measures of bowel function and the HRQoL of colon as well as rectal cancer patients following surgery. Neurourology and Urodynamics DOI 10.1002/nau
The findings in this review are in line with a previous systematic review on treatment options for improving anorectal function after rectal surgery for patients with or without cancer.20 The previous review examined fifteen studies, which included heterogeneous treatment options, such as pelvic floor re-education (seven out of fifteen studies), colonic irrigation, and sacral nerve stimulation. Although the review reported beneficial effects of conservative therapies on anorectal function after rectal resection, the papers reviewed were of limited methodological quality with heterogeneous patient populations, such as patients with rectal cancer, cervical cancer, and undefined pathology.20 The findings of this systematic review on specific bowel cancer populations are consistent with the 5th International
Pelvic Floor Muscle Training in Bowel Cancer
7
TABLE III. Intervention Programs in Reviewed Studies
Author and year
Intervention program
Prospective non-randomized trials Allgayer et al., 2005 Outpatient intensive PFMT with intraanal EMG biofeedback for strengthening in addition to routine rehabilitation program (information, psychological support, and light aerobic exercise) Ho et al., 1996 Outpatient PFMT using manometric pressure biofeedback for coordination training Ho et al., 1997 Outpatient PFMT using manometric pressure biofeedback for coordination training Laforest et al., 2012 Outpatient PFMT using manometric pressure biofeedback for coordination training Liu et al., 2011 Home-based PFMT and 4 ways to strengthen pelvic muscles (raising low back while lying in bed, standing with back against a wall, standing on tiptoes with support of chair back, and holding a rubber ball between legs)
Pucciani et al., 2008
Retrospective studies Bartlett et al., 2011
Kim et al., 2011
Multimodal rehabilitative program (pelviperineal kinesitherapy [outpatient], PFMT with manometric biofeedback for strengthening [home-based], volumetric rehabilitation [homebased], and electrostimulation [home-based])
Duration (min)
Frequency (per week)
Length
Total number of sessions
Home practice (Y/N)
30–40 (PFMT)
Daily
3 weeks
NR
Y, PFMT with biofeedback, 1-hour daily
60
1
NR
4
Y, encouraged
60
1
NR
4
Y, encouraged
60
1
NR
15
NR
10
21–28 (3–4/day)
Mean 20.0 (SD 4.7) months
NR
Y, PFMT
7 (pelviperineal kinesitherapy)
Y, PFMT with biofeedback and electrostimulation
Y, individually prescribed regimen of relaxation, muscle squeezes, and evacuation techniques, daily, 4 weeks Y, not specified
20 (PFMT with biofeedback)
2 (pelviperineal kinesitherapy)
Range: 12.0–29.0 months 4 weeks (PFMT with biofeedback)
14 (PFMT with biofeedback, 2/day) 14 (volumetric rehabilitation, 2/day) Daily (electrostimulation)
12 weeks (electrostimulation)
Outpatient PFMT using manometric pressure biofeedback for coordination training
60–90
Daily
Mean 7 weeks
4–5
Outpatient PFMT with manometric biofeedback for coordination training, sensory training, and strength training.
NR
1
10 weeks
NR
Abbreviation: EMG, electromyography; Y, yes; N, no; NR, not reported; PRS, perineal rectosigmoidectomy; JP, J pouch.
Each rehabilitative technique was provided based on participant’s manometric reports. Only twelve participants followed all four rehabilitative techniques.
Consultation on Incontinence, which recommends PFMT as an early intervention in the treatment of fecal incontinence in general adult populations.17 The results of the present study demonstrate a modest reduction in stool frequency and incontinence episodes measured by patient bowel diary or patient report after PFMT. However, most studies that showed significant effects were underpowered with small numbers of participants or used a retrospective study design; hence, it is possible that the positive effects of PFMT on patient-reported measure of bowel function are the results of natural recovery or the Hawthorne effect.40 Furthermore, the effectiveness of PFMT for the severity of incontinence in CRC patients after surgery in this study was consistent with the result of the systematic review conducted by Maris et al.;20 however, our study found that PFMT may be beneficial to certain patient population - CRC Neurourology and Urodynamics DOI 10.1002/nau
patients. Only two studies included in our review had evaluated whether the improvements in the patient-reported measures of bowel function were maintained over time, and the results were conflicting. Randomized controlled trials with longitudinal follow-ups are urgently needed to investigate if the long-term effects of PFMT for patients after CRC surgery are maintained. In contrast to the beneficial effects of PFMT on the patientreported measures of bowel function, all studies except one retrospective study reported no significant changes in manometric measures following interventions.32,33,36,37,39 The inconsistent findings between patient-reported measures and objective manometric measures of bowel function were consistent with previous studies, which demonstrated no correlation between anorectal manometry and severity of fecal
8
Lin et al.
TABLE IV. Results: Patient-Reported Bowel Function Pre- and Post-Intervention Pre-intervention Mean SD
Post-intervention Mean SD
Patient bowel diary
Median 5.0 (IQR 2.9–8.6)
Median 2.9 (IQR 1.9–4.4)
7 6
Patient bowel diary Patient bowel diary
8.7 5.6 6.2 5.1
6
Patient bowel diary
5
Patient bowel diary
37.3 31.1 (per week) 3.0 1.1 (per week)
58 8 4 70
Patient Patient Patient Patient
22
Gastrointestinal standardized questionnaire Gastrointestinal standardized questionnaire
Author and year
N
Daily stool frequency Bartlett et al., 2011
19
Ho et al., 1996 Anterior resection Total colectomy Ho et al., 1997 Incontinence Constipation Kim et al., 2011 fecal incontinence incomplete evacuation frequent defecation All Laforest et al., 2012 Rehabilitation
Control
Pucciani et al., 2008 Low anterior rectal resection Straight coloanal anastomosis Daily incontinence episodes Bartlett et al., 2011
Ho et al., 1996 Anterior resection Total colectomy Ho et al., 1997 Incontinence Constipation Severity of fecal incontinence Allgayer et al., 2005 Irradiated Non-irradiated Bartlett et al., 2011 Kim et al., 2011 fecal incontinence incomplete evacuation frequent defecation All Laforest et al., 2012 Rehabilitation Control Pucciani et al., 2008 Low anterior rectal resection Straight coloanal anastomosis All
24
Measure
10.1 4.4 4.1 2.7 8.3 2.1 9.4 4.5
report report report report
P-value
0.003
4.6 3.2 3.3 3.9
Pelvic Floor Muscle Training for Bowel Dysfunction Following Colorectal Cancer Surgery: A Systematic Review 1
Kuan-Yin Lin,1* Catherine L Granger,1,2 Linda Denehy,1 and Helena C Frawley1,3 Department of Physiotherapy, School of Health Sciences, The University of Melbourne, Melbourne, Australia 2 Department of Physiotherapy, Royal Melbourne Hospital, Melbourne, Australia 3 Allied Health Research, Cabrini Health, Melbourne, Australia
Aims: To identify, evaluate and synthesize the evidence examining the effectiveness of pelvic floor muscle training (PFMT) on bowel dysfunction in patients who have undergone colorectal cancer surgery. Methods: Eight electronic databases (MEDLINE 1950–2014; CINAHL 1982–2014; EMBASE 1980–2014; Scopus 1823–2014; PsycINFO 1806–2014; Web of Science 1970–2014; Cochrane Library 2014; PEDro 1999–2014) were systematically searched in March 2014. Reference lists of identified articles were cross referenced and hand searched. Randomized controlled trials, cohort studies and case series were included if they investigated the effects of conservative treatments, including PFMT on bowel function in patients with colorectal cancer following surgery. Two reviewers independently assessed the risk of bias of studies using the Newcastle-Ottawa Scale (NOS). Results: Six prospective non-randomized studies and two retrospective studies were included. The mean (SD) NOS risk of bias score was 4.9 (1.2) out of 9; studies were limited by a lack of non-exposed cohort, lack of independent blinded assessment, heterogeneous treatment protocols, and lack of long-term follow-up. The majority of studies reported significant improvements in stool frequency, incontinence episodes, severity of fecal incontinence, and health-related quality of life (HRQoL) after PFMT. Meta-analysis was not possible due to lack of randomized controlled trials. Conclusions: Pelvic floor muscle training for patients following surgery for colorectal cancer appears to be associated with improvements in bowel function and HRQoL. Results from nonrandomized studies are promising but randomized controlled trials with sufficient power are needed to confirm the effectiveness of PFMT in this population. Neurourol. Urodynam. # 2014 Wiley Periodicals, Inc. Key words: biofeedback; bowel function; colorectal neoplasms; pelvic floor; physiotherapy; rehabilitation
INTRODUCTION
Colorectal cancer (CRC) is the third most common cancer in males and the second most common cancer in females worldwide, with an estimation of 1.3 million new cases diagnosed each year.1 Colorectal cancer is responsible for 8.5% of all cancer deaths yearly.1 The five-year survival for CRC is around 65%, indicating that most CRC patients today can expect to live many years after their diagnosis.2,3 Surgery is usually the preferred treatment for patients with CRC, with approximately 54% to 85% of patients receiving surgery as part of their treatment.4 Several surgical procedures are used for CRC (the choice of which depends on the cancer site, stage, and patient age). For colon cancer, these include open colectomy, polypectomy, and local excision; for rectal cancer: polypectomy, local excision, local transanal resection, transanal endoscopic microsurgery, total mesorectal excision, low anterior resection, ultra-low anterior resection, proctectomy, and abdominoperineal resection.5,6 Radiotherapy and chemotherapy are used in conjunction with surgery for patients with advanced CRC.6 Although more patients with CRC live longer after diagnosis due to early detection and improved treatments, surgical procedures and chemoradiation for CRC may impact on normal pelvic floor function through damage to muscular, neural, and fascial structures.7–9 Post-operatively patients frequently complain of impaired bowel function, such as incomplete evacuation, excessive flatus, fecal urgency, straining at stool, perianal soreness or itching, bloating, and fecal incontinence.10 Among all bowel dysfunction, fecal incontinence in particular is a prevalent symptom after CRC surgery with a reported incidence from 3.2% to 79.3%.11 The wide range in incidence of fecal #
2014 Wiley Periodicals, Inc.
incontinence may be explained by differences in definition and research methodology. Although one previous study has shown that bowel dysfunction such as urgency and tenesmus may resolve by 12 months postoperatively, fecal incontinence increases significantly after one year.12 As a severely distressing and embarrassing symptom, bowel dysfunction may significantly impact on patients’ health-related quality of life (HRQoL) and cause considerable burden to patients, the healthcare system, and society. A wide range of physical therapies (e.g., pelvic floor muscle training [PFMT], weighted vaginal cones, biofeedback, electrical stimulation) have been used in the treatment of bowel dysfunction.13,14 Of these, PFMT is a highly recommended conservative management in the treatment of fecal incontinence in non-cancer populations.15–17 Previous narrative
Abbreviations: CINAHL, Cumulative Index to Nursing and Allied Health Literature; CRC, colorectal cancer; EMBASE, Excerpta Medica Database; FACT-C, Functional Assessment of Cancer Therapy – Colorectal; FIQL, Fecal Incontinence Quality of Life; HRQoL, health-related quality of life; MCIS, Modified Cleveland Incontinence Score; MEDLINE, Medical Literature Analysis and Retrieval System Online; NOS, Newcastle-Ottawa Scale; PEDro, Physiotherapy Evidence Database; PFM, pelvic floor muscle; PFMT, pelvic floor muscle training; PRISMA, Preferred Reporting Items for Systematic Reviews; SF-36, Short Form 36 questionnaire; SPSS, Statistical Product and Service Solutions; WIS, Wexner Incontinence Scale. Research conducted at: University of Melbourne Potential conflicts of interest: Nothing to disclose. *Correspondence to: Kuan-Yin Lin, MSc, PT, Alan Gilbert Building, Level 7, 161 Barry Street, Carlton 3053 Victoria, Australia. E-mail: [email protected] Received 20 March 2014; Accepted 16 June 2014 Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/nau.22654
2
Lin et al.
literature reviews have also suggested that pelvic floor exercise with biofeedback may be tried as a treatment option to improve muscular strength of the external anal sphincter, rectal sensitivity, and pelvic floor muscle (PFM) coordination for cancer patients with pelvic floor or bowel symptoms following surgery.14,18,19 While a recent systematic review reported that conservative therapies such as pelvic floor re-education and colonic irrigation can improve anorectal function in patients undergoing rectal resection for cancer and non-cancer diagnoses,20 an important literature gap still exists in terms of the benefit of PMFT in patients specially with CRC. Individuals with CRC behave differently to those undergoing colorectal surgery for non-cancer diagnoses in terms of symptoms, surgical approaches, and side effects of chemotherapy and radiotherapy. Therefore, the aim of this review was to systematically investigate the effect of PFMT on bowel dysfunction in patients after surgery for CRC. MATERIALS AND METHODS
This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines. No protocol for this review has been registered or published.21 Search Strategy
Electronic databases, MEDLINE (1950–2014), CINAHL (1982– 2014), EMBASE (1980–2014), Scopus (1823–2014), PsycINFO (1806–2014), Web of Science (1970–2014), Cochrane Library (2014), and Physiotherapy Evidence Database (1999–2014) were searched in March 2014 by one reviewer (KYL) using different combinations of the following terms: pelvic floor muscle training; rehabilitation; physical therapy modalities; biofeedback psychology; pelvic floor exercise; kegel; biofeedback; bowel function; defecation; fecal incontinence; gastrointestinal function; total mesorectal excision; sphincter-saving surgery; anterior resection; abdominoperineal resection; colectomy; colorectal surgery; sphincter-preserving surgery; colorectal neoplasms; rectal neoplasms; colonic neoplasms. An example of search strategy is presented in Appendix 1. Electronic databases were accessed via The University of Melbourne, Australia, with the last search run on March 10, 2014. No restrictions on the publication date were imposed. The search was limited to articles written in English. The reference lists of identified articles were also hand searched.
study to have CRC, or data to be reported separately for the cancer participants. Types of intervention. Studies which included PFMT with or without biofeedback were eligible. This included all types of visual, sensory, or auditory biofeedback. According to the 5th International Consultation on Incontinence, PFMT (also known as Kegel exercise training) for treating fecal incontinence involves the contraction of PFM including the external anal sphincter and puborectalis.17 For the purpose of this review, PFMT/Kegel exercise and anal sphincter exercise will be considered together with the use of biofeedback for the purpose of PFM strengthening (PFMT combined with anal canal pressure or intra-anal electromyograph feedback device), sensory training (introducing a balloon-tipped catheter into the rectum), or coordination training (employing a pressure transducer to provide simultaneous feedback in intra-rectal and intra-anal pressure during PFM contraction) made explicit.17,23 Studies that used solely electrical stimulation were not eligible for inclusion. Types of outcomes. Studies were eligible for inclusion if their primary outcome was bowel function measured by a patientreported outcome measure (such as a bowel diary, questionnaire, patient report) or anal manometry. Two reviewers (KYL and CG) independently assessed studies based on titles and abstracts for inclusion. Full texts of potentially relevant studies were retrieved and reviewed by the two reviewers to evaluate the eligibility. Disagreements could be resolved by a third independent reviewer (HF) although this was not required. Authors were contacted by e-mail if more information was needed to screen an article (this occurred on one occasion). Data Collection Process
An electronic data collection form was specifically developed and utilised. Data were extracted from the included studies by one reviewer (KYL) and crossed-checked by a second independent reviewer (CG). Disagreements were resolved through discussion or with the use of a third reviewer (HF). Data items extracted from the included studies included the first author’s name; year published; number of participants; gender; age; type of surgery; adjuvant therapy; type, duration, frequency, and length of intervention program; and results of reviewed studies. Risk of Bias
Study Selection
Studies were eligible if they met the following inclusion criteria: Types of studies. Quantitative study designs including randomized controlled trials (RCTs), cohort studies and case series were eligible for inclusion. Case studies with only one participant were not eligible. Only studies published in a peer reviewed journal were eligible. Types of participants. Participants of any age, who had undergone surgery for colon or rectal cancer, were eligible to be included in this review. Colorectal cancer was defined as cancer that forms in the tissues of the colon or the rectum, and participants with any stage of CRC were eligible.22 Given the limited number of studies available on this topic, we included studies with mixed cohorts of participants with cancer and non-cancer diagnoses who had undergone colorectal surgery; for inclusion, we required at least 50% of the participants in the Neurourology and Urodynamics DOI 10.1002/nau
As non-randomized studies were the primary study design found, the risk of bias was assessed by the two independent reviewers using the Newcastle-Ottawa Scale (NOS) for nonrandomized studies (Supplemental section). A third review was available if any disagreements arose between the two reviewers (although this was not required), and a consensus score/point was assigned after discussion. Studies were also ranked on the National Health and Medical Research Council Hierarchy of Evidence Scale.24 The scoring criteria used for synthesis of bias risk of included studies were reported as ‘‘poor’’ to ‘‘excellent’’ based on the score calculated in the NOS. Statistics
Kappa statistics and percentage agreement were calculated to establish agreement between reviewers for study selection and risk of bias using SPSS for Windows statistical software
Pelvic Floor Muscle Training in Bowel Cancer 25
package (SPSS Inc., Version 21, Chicago, IL). Values of kappa greater than 0.8 reflect ‘‘excellent’’ agreement.26 Mean and variance of pre- and post- intervention assessments were obtained directly from the study results or calculated using the following equations when not reported: 27,28 Standard deviation ¼ Range=4 Standard deviation ¼ Standard error of the mean ðsquare root of the number of samplesÞ: RESULTS Study Selection
A total of 51 studies were identified following the electronic database search, and an additional four studies were identified though searching reference lists. After exclusion of duplicate records and assessment of titles and abstracts, 10 studies met inclusion criteria and were further assessed with full text. The author of one potentially eligible study29 was contacted to clarify whether the study had included any participants with CRC; the study was subsequently excluded because the diagnosis of participants was unknown. The study of Hwang et al. was excluded because only three of the fourteen participants (21%) in their study had a diagnosis of rectal cancer.30 After assessment of the eligibility of full-text articles, eight studies were finally included for review. Figure 1 shows the flow diagram of study selection. Kappa statistics for agreement between the two independent reviewers on title/ abstracts and full text were 0.94 (percentage agreement, 98.0%) and 1.000 (percentage agreement, 100%), respectively.25
Study Characteristics Methods. No randomized controlled trials were found. Six prospective non-randomized studies31–36 and two retrospective review studies37,38 were included (Table I). Participants’ characteristics. The characteristics of the participants are presented in Table II. A total of 374 participants were included in the studies. Six of eight studies reported data on cancer patients only.34–39 The remaining two studies32,33 included 54% (7/13) and 91% (10/11) cancer patients, respectively. The mean (SD) age of the participants across all studies ranged from 55.0 (11.3) to 67.0 (8.8) years. Of the eight studies, two included patients who had undergone low anterior resection,33,39 three following sphincter-saving surgery,34,35,38 and three receiving mixed types of surgery.32,36,37 Of the three studies with mixed types of surgery, one included patients who had undergone colectomy for colon cancer.37 The majority of studies (75%) included participants who had undergone chemotherapy and/or radiotherapy.32–34,36,38,39 Intervention. The types, duration, and frequency of PFMT programs varied among all studies (Table III). One study used PFMT without biofeedback.35 Six of eight studies provided PFMT with different types of biofeedback: EMG biofeedback for strengthening;39 manometric pressure biofeedback for coordination training;32–34,37 and manometric biofeedback for coordination training, sensory training, and strength training.38 One study used a multimodal rehabilitative program including PFMT with manometric biofeedback for strengthening.36 Length of intervention ranged from three weeks to 20 months. The frequency of outpatient sessions ranged from daily to once weekly, and that of home-based sessions ranged from 3–4 times daily to once daily. In most studies, home Neurourology and Urodynamics DOI 10.1002/nau
3
practice was encouraged via a formal protocol35–39 or with general encouragement.32,33 Outcome measures. All included studies evaluated patientreported measures of bowel function in terms of stool frequency, incontinence episodes, and severity of fecal incontinence. Seven of eight studies assessed stool frequency using a bowel diary,32,33,37 patient report,35,36,38 or gastrointestinal questionnaires.34 Three studies used a bowel diary32,33,37 to assess number of incontinence episodes. Five studies included severity of fecal incontinence as a patient-reported outcome.34,36–39 However, different fecal incontinence severity scales including Modified Cleveland Incontinence Score (MCIS)39 and Wexner Incontinence Scale (WIS)34,36–38 were used. In addition to the patient-reported measures of bowel function, six studies used anorectal manometry to measure anorectal physiological function.32,33,36–39 Three studies examined HRQoL after CRC surgery.34,35,37 The HRQoL was measured by the Functional Assessment of Cancer Therapy – Colorectal (FACT-C) questionnaire (studies n ¼ 1),35 Short Form 36 questionnaire (SF-36) (studies n ¼ 1),34 and Rockwood Fecal Incontinence Quality of Life Scale (FIQL) (studies n ¼ 2).34,37 Risk of Bias and Level of Evidence
Risk of bias of studies is presented in Table I. The agreement using kappa statistics between the two independent reviewers on risk of bias assessed using the NOS was 0.972 (percentage agreement, 98.6%). Minor disagreements were resolved with consensus on 100% of occasions, and on all occasions, scores between reviewers deviated by only one point. The results of risk of bias assessment showed a mean total score of 4.9 (SD 1.2, range 3 to 7). Although most criteria within participant selection and outcome measure categories were satisfied, the majority of studies were limited by lack of a non-exposed cohort, lack of independent blinded assessment, and/or inadequate longer-term follow-up.32–39 Five studies also scored zero points (out of possible two points) in the category of comparability of study group.32,35–38 At best, the level of evidence was III-2 (range III-2 to IV) (Table I). Outcomes Patient-reported outcomes measures. Study results on patient-reported measures of bowel function are presented in Table IV. Six studies showed improvement in stool frequency after intervention,32–34,36–38 and four were statistically significant (P < 0.05).32,34,37,38 Liu et al. reported no significant differences in stool frequency between intervention and control groups after home-based PFMT.35 Only one study reported the long-term follow-up results of stool frequency, which increased at two years after the final treatment session.37 Of the three studies that assessed incontinence episodes, two showed a significant decrease in incontinence episodes (P < 0.05) after PFMT using manometric biofeedback.32,33 Bartlett et al. reported no significant difference in incontinence episodes after PFMT using manometric pressure biofeedback.37 Five studies used MCIS or WIS to measure severity of fecal incontinence.34,36–39 A higher score on MCIS indicates better continence; whereas a higher WIS score indicates worse continence. In the study by Allgayer et al., the MCIS significantly increased (P < 0.001) indicating improved incontinence in 95 participants after a 3-week intensive PFMT.39 The results of studies using WIS also showed a significant improvement in severity of fecal incontinence after intervention (P < 0.05),36–38 except for Laforest et al., who reported no
4
Lin et al.
Fig. 1. PRISMA flow diagram of study selection process.
significant difference in WIS score between intervention and control groups.34 Two studies demonstrated conflicting longterm follow-up results of severity of fecal incontinence.37,39 One study reported maintenance of effects of PFMT with intra-anal EMG biofeedback after 1 year,39 and the other study reported increased severity at 2 years following last treatment session of PFMT using manometric biofeedback, although the severity remained better than at baseline.37 Manometric measures. Six studies used anorectal manometry to measure anorectal physiological function.32,33,36–39 The majority reported no significant differences in mean resting pressure,32,33,36,37,39 maximum squeeze pressure,32,33,36,37,39 maximal tolerable volume,32,33,36,37 volume at initial sensation,32,33,36,37,39 physiological length,32,33,39 and compliance32,33 Neurourology and Urodynamics DOI 10.1002/nau
in patients after intervention. Only one study by Kim et al. showed significant differences in mean resting pressure (P ¼ 0.01), maximum squeeze pressure (P ¼ 0.006), and maximum tolerable volume (P ¼ 0.003) in patients with anterior resection after PFMT with manometric biofeedback38 (Supplemental Table I). Health-related quality of life (HRQoL). Three studies examined HRQoL after CRC surgery.34,35,37 Two studies demonstrated significant improvement in some domains of HRQoL,34,37 and one found a significant difference in total score of HRQoL (P ¼ 0.038) measured using FACT-C after intervention.35 Although not statistically significant, Bartlett et al. reported further improvement in FIQL scores from the final session at two-year follow-up.
Neurourology and Urodynamics DOI 10.1002/nau
Italy
Taiwan
France
Country
Korea
Prospective case control Prospective cohort Prospective cohort
Design
Prospective case series Retrospective case series Prospective case series Prospective case series Retrospective case series
Design
2) Representativeness of the cases
1) Is the case definition adequate?
3) Selection of Controls
3) Ascertainment of exposure
Selection (0–4)
1) Representativeness of the exposed cohort
4) Definition of Controls
() Criteria not satisfied, high risk of bias.
( ) Criteria satisfied.
NOS score
1) Comparability of cases and controls on the basis of the design or analysis
1) Assessment of outcome
1) Ascertainment of exposure
1) Comparability of cohorts on the basis of the design or analysis
Comparability (0–2)
NOS score
Comparability (0–2)
4) Demonstration that outcome of interest was not present at start of study
Abbreviation: NOS, Newcastle-Ottawa Scale; NHMRC, National Health and Medical Research Council, Australia.
Pucciani et al. 2008
Laforest et al. 2012 Liu et al. 2011
Author and year
Kim et al. 2011
Singapore
Ho et al. 1997
Australia
Singapore
Germany
Country
Allgayer et al. 2005 Bartlett et al. 2011 Ho et al. 1996
Author and year
2) Selection of the non exposed cohort
Selection (0–4)
TABLE I. Risk of Bias in Reviewed Observational Trials Using the Newcastle-Ottawa Scale
2) Same method of ascertainment for cases and control
Outcome (0–3)
3) Adequacy of follow up of cohorts
3) Non-Response rate
2) Was follow-up long enough for outcomes to occur
Outcome (0–3)
4
5
7
Total (0–9)
4
6
5
3
5
Total (0–9)
III-2
III-2
III-2
NHMRC grade of evidence
IV
IV
IV
IV
IV
NHMRC grade of evidence
Pelvic Floor Muscle Training in Bowel Cancer 5
6
Lin et al.
TABLE II. Participants in Reviewed Studies
Author and year
Number of participants, (n) %
Prospective non-randomized trials Allgayer et al., 95 2005 Irradiated patients (41) 43.2% Non-irradiated patients (54) 56.8% Ho et al., 1996 13
Irradiated: 28/13
Low anterior resection (95) 100%
Post-op RT (41) 43.2%
Fecal incontinence
Non-irradiated: 33/21
Irradiated: 58.5 11.3 (range: 31.0–76.0) Non-irradiated: 67.0 8.8 (range: 48.0–83.0)
10/3
62.1 16.6
Anterior resection (7) 53.8%
Post-op RT (4) 30.8%
Frequent bowel movement and/episode of incontinence
Post-op RT (2) 18.2%
Incontinence or constipation
Laparoscopic sphincter-saving TME (46) 100%
Pre-op RCT (32) 69.6%
Not specified
Anal sphincter preserving surgery (22) 100%
NR
Frequent defecation and fecal incontinence
Pre-op RT (19) 21.6%
Incontinence
Type of surgery, (n) %
Total colectomy (6) 46.2% Low anterior resection (11) 100%
64.8 10.9
Rehabilitation: 11/11
Rehabilitation: 55.0 11.3 (range: 33.0–78.0)
Rehabilitation group (22) 47.8% Control group (24) 52.2% 22
Control: 15/9
Control: 60.0 11.3 (range: 35.0–80.0)
Exercise: 4/7
Exercise: 55.3 14.3 (range: 27.0–82.0)
Exercise group (11) 50.0% Non-exercise group (11) 50.0% 98
Non-exercise: 8/3
Non-exercise: 65.7 8.7 (range: 45.0–73.0)
Patients: 34/54
Incontinent patients (88) 89.8% Healthy control (10) 10.2%
Healthy control: 4/6
Patients: 59.6 6.8 (range: 46.0–73.0) Healthy control: 57.4 6.3 (range: 44.0–69.0)
Low anterior rectal resection (69) 78.4% Straight coloanal anastomosis (19) 21.5%
Retrospective studies Bartlett et al., 2011
19
10/9
64.1 (95% CI: 47.0–81.3)
Kim et al., 2011
70
49/21
58.1 10.1 (range: 31.0–79.0)
Anterior resection (3) 15.8% Ultra-low anterior resection (10) 52.6% Segmental colectomy (2) 10.5% Proctocolectomy (4) 21.1% Sphincter-saving surgery with TME (70) 100%
Laforest et al., 2012
Liu et al., 2011
Pucciani et al., 2008
11 Incontinence (6) 54.5% Constipation (5) 45.5% 46
Symptoms at baseline
Mean age, year SD
5/6
Ho et al., 1997
Adjuvant therapy, (n) %
Gender, M/F
Post-op RT (34) 38.6%
NR
Bowel dysfunction
Pre-op RT (30) 42.9%
Anterior resection syndrome
Post-op RT (19) 27.1% Pre-op CT (1) 1.4% Post-op CT (25) 35.7% Pre-op plus post-op CT (31) 44.3%
Abbreviations: M, male; F, female; post-op, postoperative; pre-op, preoperative; RT, radiotherapy; CT, chemotherapy; RCT, radiochemotherapy; TME, total mesorectal excision; NR, not reported; CI, confidence interval.
Value calculated (not reported in original study).
DISCUSSION
The aim of our review was to systematically evaluate the evidence for the effect of PFMT on bowel dysfunction in patients who have undergone CRC surgery. No randomized controlled trials were identified in this review, which reflects the urgent need for RCTs in this field to provide high levels of evidence to direct PFMT for patients suffering from this highly distressing problem. Despite the paucity of RCTs and the inherent methodological shortcomings of the included observational studies, this systematic review found that PFMT may improve the patient-reported measures of bowel function and the HRQoL of colon as well as rectal cancer patients following surgery. Neurourology and Urodynamics DOI 10.1002/nau
The findings in this review are in line with a previous systematic review on treatment options for improving anorectal function after rectal surgery for patients with or without cancer.20 The previous review examined fifteen studies, which included heterogeneous treatment options, such as pelvic floor re-education (seven out of fifteen studies), colonic irrigation, and sacral nerve stimulation. Although the review reported beneficial effects of conservative therapies on anorectal function after rectal resection, the papers reviewed were of limited methodological quality with heterogeneous patient populations, such as patients with rectal cancer, cervical cancer, and undefined pathology.20 The findings of this systematic review on specific bowel cancer populations are consistent with the 5th International
Pelvic Floor Muscle Training in Bowel Cancer
7
TABLE III. Intervention Programs in Reviewed Studies
Author and year
Intervention program
Prospective non-randomized trials Allgayer et al., 2005 Outpatient intensive PFMT with intraanal EMG biofeedback for strengthening in addition to routine rehabilitation program (information, psychological support, and light aerobic exercise) Ho et al., 1996 Outpatient PFMT using manometric pressure biofeedback for coordination training Ho et al., 1997 Outpatient PFMT using manometric pressure biofeedback for coordination training Laforest et al., 2012 Outpatient PFMT using manometric pressure biofeedback for coordination training Liu et al., 2011 Home-based PFMT and 4 ways to strengthen pelvic muscles (raising low back while lying in bed, standing with back against a wall, standing on tiptoes with support of chair back, and holding a rubber ball between legs)
Pucciani et al., 2008
Retrospective studies Bartlett et al., 2011
Kim et al., 2011
Multimodal rehabilitative program (pelviperineal kinesitherapy [outpatient], PFMT with manometric biofeedback for strengthening [home-based], volumetric rehabilitation [homebased], and electrostimulation [home-based])
Duration (min)
Frequency (per week)
Length
Total number of sessions
Home practice (Y/N)
30–40 (PFMT)
Daily
3 weeks
NR
Y, PFMT with biofeedback, 1-hour daily
60
1
NR
4
Y, encouraged
60
1
NR
4
Y, encouraged
60
1
NR
15
NR
10
21–28 (3–4/day)
Mean 20.0 (SD 4.7) months
NR
Y, PFMT
7 (pelviperineal kinesitherapy)
Y, PFMT with biofeedback and electrostimulation
Y, individually prescribed regimen of relaxation, muscle squeezes, and evacuation techniques, daily, 4 weeks Y, not specified
20 (PFMT with biofeedback)
2 (pelviperineal kinesitherapy)
Range: 12.0–29.0 months 4 weeks (PFMT with biofeedback)
14 (PFMT with biofeedback, 2/day) 14 (volumetric rehabilitation, 2/day) Daily (electrostimulation)
12 weeks (electrostimulation)
Outpatient PFMT using manometric pressure biofeedback for coordination training
60–90
Daily
Mean 7 weeks
4–5
Outpatient PFMT with manometric biofeedback for coordination training, sensory training, and strength training.
NR
1
10 weeks
NR
Abbreviation: EMG, electromyography; Y, yes; N, no; NR, not reported; PRS, perineal rectosigmoidectomy; JP, J pouch.
Each rehabilitative technique was provided based on participant’s manometric reports. Only twelve participants followed all four rehabilitative techniques.
Consultation on Incontinence, which recommends PFMT as an early intervention in the treatment of fecal incontinence in general adult populations.17 The results of the present study demonstrate a modest reduction in stool frequency and incontinence episodes measured by patient bowel diary or patient report after PFMT. However, most studies that showed significant effects were underpowered with small numbers of participants or used a retrospective study design; hence, it is possible that the positive effects of PFMT on patient-reported measure of bowel function are the results of natural recovery or the Hawthorne effect.40 Furthermore, the effectiveness of PFMT for the severity of incontinence in CRC patients after surgery in this study was consistent with the result of the systematic review conducted by Maris et al.;20 however, our study found that PFMT may be beneficial to certain patient population - CRC Neurourology and Urodynamics DOI 10.1002/nau
patients. Only two studies included in our review had evaluated whether the improvements in the patient-reported measures of bowel function were maintained over time, and the results were conflicting. Randomized controlled trials with longitudinal follow-ups are urgently needed to investigate if the long-term effects of PFMT for patients after CRC surgery are maintained. In contrast to the beneficial effects of PFMT on the patientreported measures of bowel function, all studies except one retrospective study reported no significant changes in manometric measures following interventions.32,33,36,37,39 The inconsistent findings between patient-reported measures and objective manometric measures of bowel function were consistent with previous studies, which demonstrated no correlation between anorectal manometry and severity of fecal
8
Lin et al.
TABLE IV. Results: Patient-Reported Bowel Function Pre- and Post-Intervention Pre-intervention Mean SD
Post-intervention Mean SD
Patient bowel diary
Median 5.0 (IQR 2.9–8.6)
Median 2.9 (IQR 1.9–4.4)
7 6
Patient bowel diary Patient bowel diary
8.7 5.6 6.2 5.1
6
Patient bowel diary
5
Patient bowel diary
37.3 31.1 (per week) 3.0 1.1 (per week)
58 8 4 70
Patient Patient Patient Patient
22
Gastrointestinal standardized questionnaire Gastrointestinal standardized questionnaire
Author and year
N
Daily stool frequency Bartlett et al., 2011
19
Ho et al., 1996 Anterior resection Total colectomy Ho et al., 1997 Incontinence Constipation Kim et al., 2011 fecal incontinence incomplete evacuation frequent defecation All Laforest et al., 2012 Rehabilitation
Control
Pucciani et al., 2008 Low anterior rectal resection Straight coloanal anastomosis Daily incontinence episodes Bartlett et al., 2011
Ho et al., 1996 Anterior resection Total colectomy Ho et al., 1997 Incontinence Constipation Severity of fecal incontinence Allgayer et al., 2005 Irradiated Non-irradiated Bartlett et al., 2011 Kim et al., 2011 fecal incontinence incomplete evacuation frequent defecation All Laforest et al., 2012 Rehabilitation Control Pucciani et al., 2008 Low anterior rectal resection Straight coloanal anastomosis All
24
Measure
10.1 4.4 4.1 2.7 8.3 2.1 9.4 4.5
report report report report
P-value
0.003
4.6 3.2 3.3 3.9
