Biofeedback Treatment of Constipation A Critical Review Steve Heymen, M.S., Kenneth R. Jones, Ph.D., Yolanda Scarlett, M.D., William E. Whitehead, Ph.D. From the Center for Functional Gastrointestinal and Motility Disorders, Department of Medicine, The University of North Carolina, Chapel Hill, North Carolina PURPOSE: This review was designed to 1) critically examine the research design used in investigations of biofeedback for pelvic floor dyssynergia, 2) compare the various biofeedback treatment protocols for pelvic floor dyssynergia-type constipation used in this research, 3) identify factors that influence treatment outcome, and 4) identify goals for future biofeedback research for pelvic floor dyssynergia. METHODS: A comprehensive review of both the pediatric and adult research from 1970 to 2002 on “biofeedback for constipation” was conducted using a Medline search in all languages. Only prospective studies including five or more subjects that described the treatment protocol were included. In addition, a meta-analysis of these studies was performed to compare the outcome of different biofeedback protocols for treating constipation. RESULTS: Thirtyeight studies were reviewed, and sample size, treatment protocol, outcome rates, number of sessions, and etiology are shown in a table. Ten studies using a parallel treatment design were reviewed in detail, including seven that randomized subjects to treatment groups. A meta-analysis (weighted by subjects) was performed to compare the results of two treatment protocols prevalent in the literature. The mean success rate of studies using pressure biofeedback (78 percent) was superior (P ⫽ 0.018) to the mean success rate for studies using electromyography biofeedback (70 percent). However, the mean success rates comparing studies using intra-anal electromyography sensors to studies using perianal electromyography sensors were 69 and 72 percent, respectively, indicating no advantages for one type of electromyography protocol over the other (P ⫽ 0.428). In addition to the varied protocols and instrumentation used, there also are inconsistencies in the literature regarding the severity and etiology of symptoms, patient selection criteria, and the definition of a successful outcome. Finally, no anatomic, physiologic, or demographic variables were identified that would assist in predicting successful outcome. Having significant psychological symptoms was identified as a factor that may influence treatment outcome, but this requires further study. CONCLUSION: Although most studies report positive results using biofeedback to treat constipation, quality research is lacking. Specific recommendations are made for future investigations to 1) improve experimental design, 2) clearly define outcome measures, 3) identify the etiology and severity of symptoms, 4) determine which treatment protocol and which component of treatment is most effective for different types of Preparation of this manuscript was supported, in part, by National Institutes of Health grant #R01DK57048. Address reprint requests to Mr. Heymen: CB# 7080, 777 BurnettWomack Building, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.
subjects, 5) systematically explore the role of psychopathology in this population, 6) use an adequate sample size that allows for meaningful analysis, and 7) include long-term follow-up data. [Key words: Biofeedback; Constipation; Outlet obstruction; Pelvic floor dyssynergia; Electromyography; Manometry] Heymen S, Jones KR, Scarlett Y, Whitehead WE. Biofeedback treatment of constipation: a critical review. Dis Colon Rectum 2003;46:1208–1217.
C
onstipation is a common disorder, occurring in an estimated 4 percent of United States adults,1 with pelvic floor dyssynergia-type constipation (PFD) making up an estimated 25 to 50 percent of this group.2,3 Preston and Lennard-Jones2 first described the association of PFD with constipation, and subsequent investigators have confirmed their observation using a variety of synonymous terms, such as anismus, spastic pelvic floor syndrome, outlet obstruction, and paradoxical puborectalis contractions.4–6 During the act of defecation, the puborectalis sling muscle and the external anal sphincter should relax to permit defecation. This can be demonstrated by recording electromyographic activity or anal canal pressure from pelvic floor muscles during attempts to defecate. However, some chronically constipated patients inappropriately contract or fail to relax the external anal sphincter and puborectalis muscles. This uncoordinated effort obstructs defecation. Some studies suggest that the finding of PFD varies from one occasion of testing to another7 and is less likely to be seen at home when ambulatory monitors are used to record the response to straining. PFD also is observed in some asymptomatic controls and fecalincontinent patients,8 leading some investigators to question whether this is a distinct abnormality causing constipation.7–9 However, most clinicians and researchers believe that although current diagnostic criteria10 can lead to a false-positive diagnosis, a subgroup of patients exists whose symptoms of chronic constipation and/or fecal impaction occur as a result of inability to relax the pelvic floor when straining to defecate.11
1208
Vol. 46, No. 9
BIOFEEDBACK TREATMENT OF CONSTIPATION
Although constipation is not life threatening, it does have an adverse effect on quality of life and is associated with significant morbidity and costs (e.g., cost of care and work absenteeism). In a United States Householder Survey of functional gastrointestinal disorders, 9 percent of constipation subjects reported being too sick to go to work.12 Patients with symptoms of difficult defecation showed significant impairments on the Health Survey Questionnaire (SF-36)13 scales for “bodily pain,” “role physical” (limitations in ability to work or perform usual physical activities), and “general health,” even when the possible mediating effects of neuroticism are statistically controlled.14 The estimated cost of laxative use alone is more than 4 million dollars annually in the United States.15 Furthermore, when traditional treatments are unsuccessful, symptoms often worsen over time. When patients do not respond to conservative interventions such as dietary recommendations, bowel scheduling, and medications, biofeedback is frequently used to treat PFD-type constipation. Because uncontrolled trials suggest that biofeedback is associated with outcomes as good as medical management or surgery, and because it has a low incidence of adverse effects, biofeedback often is recommended as the first-line treatment for patients with PFD.16,17 Biofeedback studies for treating PFD have been reviewed by others.18,19 Those reviews suggest that approximately two-thirds of adult patients with pelvic floor dyssynergia benefit from biofeedback training. However, despite more than 20 years of positive results that are reported in the literature, the lack of adequately controlled trials has limited the widespread use of what seems to be a promising treatment. This review was designed to 1) critically examine the research designs used in investigations of biofeedback for PFD, 2) compare the various biofeedback treatment protocols for PFD-type constipation used in this research, 3) identify factors that influence treatment outcome, and 4) identify goals for future research of biofeedback for PFD.
METHODS A comprehensive review of both the pediatric and adult research from 1970 to 2002 on “biofeedback for constipation” was conducted using a MEDLINE search in all languages. Only prospective studies including five or more subjects that described the treatment protocol were included. In addition, a meta-analysis
1209
of these studies was performed to compare the outcome of different biofeedback protocols for treating constipation.
RESULTS Thirty-eight studies met the criteria for inclusion in this review. In addition to a description of experimental design, information regarding sample size, etiology, number of treatment sessions, and success rates are presented in Table 1.
Research Design An evaluation of the research design used in the 38 studies under review revealed that only ten (six pediatric and four adult) used any degree of experimental control. In addition to weak experimental design, many of these studies failed to use consistent criteria for subject selection and rarely included an adequate sample size to allow for meaningful statistical analysis.
Pediatric Studies Of the 11 pediatric studies reviewed, 5 were uncontrolled and 6 used a parallel research design. Unfortunately, only three of these studies randomly assigned subjects to treatment.47,49,56 All six of the controlled studies compared biofeedback to traditional medical/behavior management, such as laxative use and toileting schedules. In the three randomized studies, no significant difference between biofeedback and medical management was observed. In the only pediatric study that included a sample size that was sufficient to allow for meaningful statistical analysis,49 van der Plas et al.49 compared a combination of anal canal pressure and perianal electromyography (EMG) biofeedback to traditional medical management in 71 subjects. No significant difference in treatment outcome was found comparing these protocols. Unfortunately, the method for determining outcome is not described. In addition, PFD was identified in only 40 percent of subjects. The etiology of the remaining subject’s constipation symptoms is unclear. In the second randomized study, which also included subjects with mixed etiology, Wald et al.56 found a trend for a subgroup of subjects with PFD to do better with pressure biofeedback than with mineral oil. However, there were no significant differ-
HEYMEN ET AL
1210
Dis Colon Rectum, September 2003
Table 1. Biofeedback Review: Constipation (1980 –2002) Study Adult studies Dailianas et al.20 Weisel et al.21 McKee et al.22 Rhee et al.23 Heymen et al.24 Karlbom et al.25 Rao et al.26 Patankar et al.27 Reiger et al.28 Glia et al.29 Ko et al.30 Ho and Tan31 Ho et al.32 Koutsomanis et al.33 Hull et al.34 Papachrysostomous and Smith35 Koutsomanis et al.36 Bleijenberg and Kuijpers37 Wexner et al.38 Turnbull and Rituo39 Fleshman et al.40 Kawimbe et al.41 Lestar et al.42 Dahl et al.43 Emery et al.44 Bleijenberg and Kuijpers45 Weber et al.46 Pediatric studies Nolan et al.47 Ponticelli et al.48 van der Plas et al.49 Cox et al.50 Beninga et al.51 Loening-Baucke52 Loening-Baucke53 Keren et al.54 Veyrac et al.55 Wald et al.56 Wald et al.56 Olness et al.57
Date Sample Design Treatment 2000 2000 1999 1999 1999 1997 1997 1997 1997 1997 1997 1997 1996 1995 1995 1994 1994 1994 1992 1992 1992 1991 1991 1991 1988 1987 1987
11 13 30 45 36 28 25 86 19 26 17 5 62 60 12 22 20 20 18 7 9 15 16 14 65 10 42
u u u u r,p u u u u r,p u u u r,p u u u r,p u u u u u u u u u
P E*,B P,B E E/X/Y/Z E P E E E*/P,B E P P E*/B E*,S H,E,B E E/B E P,G E,B E,H P,B E,S *E E,B P
Outcome (%)
Sessions
64 38 30 69 ? 43 92 73 11 90 and 60 76 80 90 69 and 64 82 86 85 73 and 22 89 86 89 INDEX HT 3 wks 56 93 80 70 78 and 0
2 3.5 3.5 8 ? 8 6 8 6 ⬍11 4 4 4 3&2 1 to 3 HT 36d 4 8 9 4.5 12 PFD 1 5 ? 10 4
Dx OB PFD OB AN PFD PFD OB PFD CONST PFD PFD PO OB PFD OB,PO OB PFD PFD PFD CONST PFD PFD PFD PFD PFD RAIR, PFD
1998 29 r,p *E/M 43 and 53 3.5 PFD 1998 22 p P,B/M 60 and 58 varied” SB 1996 71 r,p E*,P/M 39 and 19 5 ENC 1994 26 p E*/M 88 and 60 2.5 PFD 1993 29 u E*,P 55 5 PFD, SD 1991 38 u E*,P 37 ⬍7 PFD 1990 41 p E*,P/M 55 and 5 ⬍7 PFD 1988 12 u P 100 4 PFD 1987 12 u P 83 1 OB 1987 18 r,p P/M 67 and 33 4 PFD 1987 ? r,p P/M 40 and 71 4 No PFD 1980 50 u P 94 3.5 ENC Dx ⫽ diagnosis; u ⫽ uncontrolled; r ⫽ randomized; p ⫽ parallel; E ⫽ electromyography, intra-anal; E* ⫽ electromyography, perianal; S ⫽ sensory; P ⫽ pressure; B ⫽ pass balloon; M ⫽ medical management; H ⫽ home trainers; G ⫽ gen relax tr.; X ⫽ E&S; Y ⫽ E&H; Z ⫽ E&S&H; OB ⫽ obstructed defecation; PFD ⫽ pelvic floor dysenergia; CONST ⫽ constipated; SD ⫽ intrarectal sensory defecit; RAIR ⫽ rectoanal inhibitory reflex; ENC ⫽ encopresis; PO ⫽ postoperative; SB ⫽ spina bifida; AN ⫽ anismus.
ences when evaluating the entire group. All subjects in the third randomized trial were diagnosed with PFD.47 In that study Nolan et al.47 found no advantages for perianal EMG biofeedback over medical management. Unfortunately, neither Wald et al.56 nor Nolan et al.47 had sufficient sample size to allow for meaningful analysis.
The quality of the research design used in the three remaining parallel studies varies considerably.48,50,53 Although, none of these studies randomly assigned subjects to treatment, Cox et al.50 provided some degree of experimental control by matching subjects by age and gender. However, Loening-Baucke53 compared subjects in the biofeedback treatment group to
Vol. 46, No. 9
BIOFEEDBACK TREATMENT OF CONSTIPATION
subjects from a previous study who were treated with medical management, making it difficult to compare the treatments. Not only did Ponticelli et al.48 fail to identify how subjects were assigned to treatment, the criteria for determining outcome were confusing.48 Finally, all three of these studies had insufficient sample size to provide reliable conclusions regarding outcome. A positive aspect of these three nonrandomized studies was that, like Nolan et al.,47 these investigators48,50,53 used consistent subject selection. The etiology of constipation symptoms was consistent within each study, in which children with PFD50,53 or children with a subtype of spina bifida48 were treated. Two of these studies found perianal EMG biofeedback to be superior to medical management53 or behavior management50 protocols. Although the remaining pediatric investigations were uncontrolled, ⬎50 percent of the pediatric investigations used a research design with some degree of experimental control. It is encouraging that all of the pediatric studies published since 1994 used parallel designs, perhaps indicating a positive trend toward more quality research in the area of biofeedback for children with PFD.
Adult Studies In contrast to the high percentage of controlled studies found in the pediatric literature, ⬍15 percent (4 of 27) adult studies) used parallel designs.24,29,33,37 Two studies compared EMG biofeedback to simulated defecation of a balloon protocols,33,37 and two studies compared different biofeedback techniques.24,29 All four of the parallel-design studies randomly assigned subjects to treatment. It should be noted that three of these uncontrolled adult studies included some pediatric subjects. Unfortunately, as with the pediatric literature, only one well-controlled adult study had a sample size that was sufficient to provide meaningful statistical conclusions. In this study, Koutsomanis et al.33compared perianal EMG biofeedback and balloon-defecation training to balloon-defecation training alone and found no significant difference between treatments in 60 subjects. The Koutsomanis et al.33 study also is one of the few investigations20,26 to address the question of adequate propulsive force during attempts to defecate. In some cases, it may be that subjects are not pushing adequately to generate propulsive force on the rectum. Although the problem of excessive force
1211
is more often seen in these patients, inadequate propulsive force may contribute to unsatisfying defecation attempts. Possible shortcomings of the Koutsomanis trial include a crossover to alternate treatment after only two unsuccessful sessions and the fact that only 75 percent of subjects had PFD identified as the etiology of their constipation. In a randomized study of 20 adult subjects, all with PFD, Bleijenberg and Kuijpers37 found intra-anal EMG biofeedback to be superior to balloon-defecation training. Ninety percent compared with 60 percent of subjects improved, respectively. Although the sample size is too small to draw reliable conclusions regarding treatment outcome comparisons, subjects who failed balloon-defecation training were then given biofeedback training, yielding an 80 percent success rate. Finally, two randomized studies directly compared different biofeedback protocols.24,29 All subjects in both studies were identified as having PFD. Heymen et al.24 compared intra-anal EMG biofeedback to a combination of EMG and intrarectal balloon-distention training, EMG and home trainers, or a combination of all three techniques. Although all groups showed significant improvement, no significant differences among treatment strategies were found. Glia et al.29 found perianal EMG biofeedback to be superior to pressure biofeedback combined with balloondefecation training. However, as with most of the controlled studies reviewed, neither Glia et al.29 nor Heymen et al.24 had sufficient sample size to provide a meaningful analysis. The experimental designs of the majority of biofeedback studies for adult and pediatric subjects with PFD were uncontrolled, within-subjects designs. This design fails to account for a myriad of alternative explanations that may affect the results of these investigations. Most of the studies that used parallel designs, including those using randomization to treatment, failed to have adequate sample size to allow for meaningful analysis. The two studies that used a randomized parallel design with adequate power, studied subjects with inconsistent or unidentified etiology, limiting the conclusions that can be drawn from their results.
BIOFEEDBACK TREATMENT PROTOCOLS Biofeedback training for constipation is directed at coordinating pelvic floor muscle (PFM) relaxation with a Valsalva maneuver (a downward intra-abdom-
1212
HEYMEN ET AL
Dis Colon Rectum, September 2003
inal pressure to generate propulsive force). Biofeedback instrumentation for treating PFD primarily falls into two categories: EMG or the use of an anal canal pressure-feedback device. Most studies in the last ten years have used pelvic floor EMG instead of anal canal pressure for biofeedback training. However, only one study has attempted to systematically compare the two types of biofeedback.29 Unfortunately, the sample size was too small for reliable comparisons of these two treatment protocols.29 To address this question, a meta-analysis to compare the treatment outcome of studies using EMG vs. studies using pressure biofeedback was conducted. The number of subjects in each study weighted this analysis. The results of this meta-analysis should be interpreted with caution. The heterogeneity of the populations, treatment methods, as well as outcome measures of these investigations makes a direct comparison of success rates difficult. For example the success rates reported in each study were determined by the subject’s, or the subject’s parent’s, satisfaction with treatment. In addition, because negative results are less likely to be published, a meta-analysis of the published literature, such as this one, may overestimate the true benefit of biofeedback. Only large, controlled trials can correct for this bias. This analysis is intended as a general overview of treatment outcome that has been published in the literature. The results of this meta-analysis revealed that the outcome of studies using pressure biofeedback was superior when compared with the outcome of studies using EMG biofeedback (chi-squared [1] ⫽ 5.597; n ⫽ 717; P ⫽ 0.018). Pressure biofeedback training (Fig. 1) was used in 13 studies20,22,26,31,32,39,42,46,48,54–57 (5 pediatric48,54–57) with a mean success rate of 78 percent
(215 of 275 subjects improved). Eighteen studies21,23,25,27,28,30,33–38,40,43–45,47,50, (two pediatric47,50) used EMG biofeedback (Fig. 2) with a mean success rate of 70 percent (310 of 442 subjects improved). Studies in which biofeedback techniques were directly compared,24,29 combined,49,51–53 or the number of successful subjects was not reported41 were not included in the meta-analysis. A second meta-analysis was conducted to evaluate two different methods for providing EMG biofeedback. Eleven of the EMG studies used intra-anal EMG sensors23,25,27,28,30,36–38,40,43,45 with a mean improvement rate of 69 percent (192 of 279 subjects improved) and seven studies used perianal EMG feedback21,33–35,44,47,50 with a mean success rate of 72 percent (118 of 163 subjects improved). This analysis revealed that there were no significant differences in the outcomes of studies using one type of EMG technique over the other (chi-squared [1] ⫽ 0.628; n ⫽ 442; P ⫽ 0.428). Overall, the meta-analyses show success rates ranging from 69 to 78 percent, regardless of which protocol or what instrumentation was used. This mirrors success rates found in earlier reviews conducted by Enck18 in 1993 and Rao et al.19 in 1997, in which 7 and 20 studies were reviewed, respectively. However, without controlled trials using direct comparisons, the optimal protocol for subjects with PFD remains unclear.
Figure 1. Outcome of studies using pressure biofeedback showing percentage of patients who improved.
Figure 2. Outcome of studies using electromyogram biofeedback showing percentage of patients who improved.
FACTORS THAT INFLUENCE TREATMENT OUTCOME In addition to treatment strategies and instrumentation, patient characteristics may influence the outcome of biofeedback treatments for PFD. Attempts
Vol. 46, No. 9
BIOFEEDBACK TREATMENT OF CONSTIPATION
have been made to determine whether demographic, physiological, anatomical, or psychological variables have any influence on the outcome of biofeedback therapy for subjects with PFD. Furthermore, some investigators recognize the importance of the relationship between the subject and the therapist as affecting treatment outcome. Anatomic, Physiologic, and Demographic Factors. Several investigators56,58–61 have questioned the usefulness of diagnostic manometric evaluations, having found no predictive value associated with treatment outcome. For example, in a retrospective study, Chiotakakou-Faliakou et al.62 found that anorectal testing did not predict outcome of biofeedback therapy in a review of 100 constipated subjects. They further recommend that pelvic floor abnormalities and colonic transit studies should not influence subject selection criteria. Similarly, Loening-Baucke53 and Dahl et al.43 reported that the ability to defecate an air-filled balloon does not correlate well with a successful treatment outcome. In another retrospective review of 194 patients with constipation, Gilliland et al.63 found that neither age, gender, duration of symptoms, nor the presence of rectal pain correlated with the outcome of biofeedback therapy. In addition, none of the physiologic indices, such as pelvic floor resting pressure, rectal capacity, sensory threshold, or length of high-pressure zone, had any relation to treatment outcome. Nor were anatomic abnormalities, such as the presence of a rectocele, intussusception, or a fixed or dynamic pelvic floor decent, associated with outcome of biofeedback therapy. The only predictor of a successful outcome, Gilliland reported, was the number of sessions attended (five or more) and whether the therapist discharged the patient (63 percent success rate) rather than the patient terminating treatment prematurely (25 percent success rate).63 To date, researchers have been unable to identify any physiologic, anatomic, or demographic variables that influence treatment outcome. Many investigators do suggest, however, that psychopathology may influence treatment outcome. Psychological Factors. Guerrero and Cavender64 identified “. . .depressed self-esteem, loss of coping skills, regression, school phobia, truancy, obsessive compulsive disorder and pseudoseizures,” as problems associated with children diagnosed with encopresis. These investigators challenged care providers to resist a narrow, medical focus on this complex psychosocial disorder.64
1213
Anxiety and/or psychological stress may contribute to the development of PFD by increasing the level of skeletal muscle tension. Burnett et al.14 reported significantly higher scores on the SCL-90R65 scales for anxiety, depression, hostility, interpersonal sensitivity, obsessive compulsive traits, phobic anxiety, and somatization for adult subjects with constipation. Using the Minnesota Multiphasic Personality Inventor® (University of Minnesota Regents, Minneapolis, MN), Heymen et al.66 found that adult subjects with PFDtype constipation and subjects with rectal pain showed a tendency to use somatization as a defense mechanism to manage psychologic distress. This pattern was not seen in a comparison group of subjects with fecal incontinence. Devroede et al.67 and Kumar et al.68 suggest that there may be a psychosomatic basis for chronic idiopathic constipation, including subjects with PFD. For example, Nehra et al.69 reported that 65 percent of constipated subjects in their study were diagnosed with various psychological disorders. Similar studies reviewed by Whitehead70 also showed elevated levels of psychological distress in patients with symptoms of difficult defecation. In an interesting study by McKee et al.,22 the authors contended that the constipated patients who benefit most from biofeedback have psychological difficulties as the predominant cause of their problem. Van der Plas et al.49 also reported the presence of psychosocial problems in a subgroup of encopretic children. They further state that “the relation between treatment success and improved behavioral functioning demonstrates that encopresis has an etiological role in the occurrence and maintenance of behavior problems in children with encopresis.” There are other studies, however, that suggest that constipated children with pelvic floor dyssynergia do not have any more behavior problems than children without pelvic floor dyssynergia.49,56,71,72 Alternatively, some investigators73,74 suggest that psychopathology may be a consequence rather than a cause of constipation. Even in subjects whose constipation was attributed to slow colonic transit, without PFD, Dykes et al.74 found that 60 percent of subjects had a concurrent affective disorder, 66 percent reported having a previous affective disorder, and 33 percent had “distorted attitudes about food.” Such a high incidence of psychopathology in subjects whose constipation was believed to be because of a neurologic deficit in the colon suggests that psychopathology may be
1214
HEYMEN ET AL
the result of suffering from chronic constipation. This finding led Dykes et al.74 to recommend research comparing the effectiveness of biofeedback to psychotherapy for subjects with constipation. Furthermore, inclusion of psychologic treatment for subjects with constipation is frequently recommended in the literature.75–77 Several studies reported that subjects with constipation often have a history of sexual abuse. Leroi et al.,78 for example, found a greater incidence of sexual abuse in women who had pelvic floor dyssynergia compared with women without PFD. These authors speculated that after the trauma of sexual abuse, any sensation of rectal fullness may trigger a memory of the original trauma and may lead to an involuntary contraction of the pelvic floor muscles. Devroede et al.67 and Kumar et al.68 also reported that many of their constipated subjects had a history of sexual or physical abuse. Lennard-Jones77 recommended supportive psychotherapy, in addition to biofeedback, for those constipation subjects who have been sexually abused. Establishing an effective psychotherapeutic relationship may be critical for these subjects to discuss their experiences openly with the clinician to facilitate treatment. Ponticelli et al.48 stressed the importance of the “specialist” in providing treatment. In particular, they pointed to the role of motivation as a predictor in determining outcome. Koutsomanis et al.33 also emphasized the importance of the rapport between the therapist and subject. It is likely that rapport between the therapist and subject significantly affects the subject’s motivation and the quality of the therapeutic relationship. Furthermore, Jones et al.79 stated, “the success of biofeedback training depends on the skill of the trainer/ therapist.” Unfortunately, at this time, there is no identified standard for training biofeedback clinicians to treat pelvic floor disorders. Whether psychopathology is caused by constipation or simply associated with constipation in some way, successfully treating many of these individuals may require adjunctive psychotherapy to achieve optimal results. Well-designed research investigating the role of psychopathology in these patients is necessary to tailor treatment strategies for children and adults suffering from this chronic condition. Special training for the therapist or the inclusion of a psychologist as part of the treatment team may improve the outcome for this population.
Dis Colon Rectum, September 2003
DISCUSSION Although the success rates for biofeedback treatment of PFD that are reported in the literature are impressive, these results should be considered cautiously for the following reasons. Only 10 of 38 studies reviewed used parallel group designs. Seven of those studies used random assignment to treatment, but only two had sufficient statistical power.33,49 Although these two studies (one pediatric49 and one adult33) were well designed, they included subjects with mixed etiology,33,49 a crossover to alternate treatment after only two unsuccessful sessions,33 or poorly described outcome criteria,49 which limit the conclusions offered. The remaining 28 studies reviewed used a within-subjects study design. In most cases these subjects had failed medical management for constipation, and it was assumed that the improvement that occurred was a result of the biofeedback treatment. However, no attempt was made in these studies to compare treatments or to control for placebo or nonspecific effects. Protocols for treating subjects with PFD were divided into two categories: EMG and pressure biofeedback. Only one study directly compared these protocols. However, a meta-analysis comparing the success rates of studies using EMG to studies using pressure biofeedback demonstrated a statistically significant advantage for pressure biofeedback. In both protocols, the meta-analysis yielded success rates of more than two-thirds of subjects. There seem to be no physiologic, anatomic, or demographic variables that influence treatment outcome or predict the type of patient who will benefit from biofeedback therapy for constipation. The number of treatment sessions was identified as a possible variable-effecting outcome, but this requires further investigation. Although psychologic and behavior problems often are observed in subjects with constipation, there is question as to whether this is a cause or consequence of being constipated. Providing psychological counseling is recommended by several researchers and may prove helpful, especially in cases in which patients with PFD do not respond to biofeedback training. Biofeedback is not readily available to patients suffering from PFD-type constipation in most areas of the United States. Recently the Centers for Medicare and Medicaid ruling (October 6, 2000) recommended biofeedback therapy for patients with urinary incontinence, leading Medicare and other third-party pay-
Vol. 46, No. 9
BIOFEEDBACK TREATMENT OF CONSTIPATION
ers to begin reimbursing patients for this treatment. To date no such endorsement has occurred for patients suffering from fecal incontinence or constipation. Several factors may have thwarted reimbursement for biofeedback therapy for constipation, thereby limiting its widespread use. These factors include 1) the lack of adequately controlled trials of sufficient sample size that demonstrate biofeedback’s effectiveness, 2) no uniform protocols established for conducting biofeedback training, 3) no consensus on criteria that should be used to select subjects for treatment, 4) no consensus on how to define treatment outcome, 5) uncertainty as to which components of treatment are most effective, and 6) limited data on the long-term efficacy of biofeedback for PFD.
GOALS FOR FUTURE RESEARCH Controlled, randomized studies are needed that clearly identify the etiology and severity of the constipation, define outcome measures, and identify the most efficacious treatment protocol for different types of patients. It also is imperative that future studies look at long-term follow-up measures and include adequate numbers of subjects to allow for meaningful statistical analysis. In addition, the role of psychopathology should be systematically explored. Although promising results are reported in the literature, studies meeting the above-mentioned criteria are necessary to confirm the efficacy of biofeedback for constipation. Until the efficacy of biofeedback for subjects with PFD is confirmed in well-controlled investigations, it is unlikely that this promising treatment will become available to the substantial numbers of individuals suffering from this challenging disorder.
REFERENCES 1. Harari D, Gurwitz JH, Avom J, Bohn R, Minaker KL. Bowel habit in relation to age and gender. Findings from the National Health Interview Survey and clinical implications. Arch Intern Med 1996;56:315–20. 2. Preston DM, Lennard-Jones JE. Anismus in chronic constipation. Dig Dis Sci 1985;30:413– 8. 3. Wald A, Caruana BJ, Freimanis MG, Bauman DH, Hinds JP. Contributions of evacuation proctography and anorectal manometry to the evaluation of adults with constipation and defecatory difficulty. Dig Dis Sci 1990;35: 481–7. 4. Kuijpers HC, Bleijenberg G. The spastic pelvic floor syndrome: a cause of constipation. Dis Colon Rectum 1985;28:669 –72.
1215
5. Heymen S, Wexner SD. Biofeedback for constipation. In: Smith LE, ed. A practical guide to anorectal testing. 2nd ed. New York: Igaku-Shoin, 1995:261–70. 6. Roberts JP, Womack NR, Hallan RI, Thorpe AC, Williams NS. Evidence from dynamic integrated proctography to redefine anismus. Br J Surg 1992;79:1213–5. 7. Duthie GS, Bartolo DC. Anismus, the cause of constipation? Results of investigation and treatment. World J Surg 1992;16:831–5. 8. Schouten WR, Briel JW, Auwerda JJ, et al. Anismus: fact or fiction? Dis Colon Rectum 1997;40:1033– 41. 9. Jones PN, Lubowski DZ, Swash M, Henry MM. Is paradoxical contraction of puborectalis muscle of functional importance? Dis Colon Rectum 1987;30:667–70. 10. Whitehead WE, Wald A, Diamant NE, Enck P, Pemberton JH, Rao SS. Functional disorders of the anus and rectum. In: Drossman DA, Corazziari E, Talley NJ, Thompson WG, Whitehead WE, eds. Rome II: The functional gastrointestinal disorders. 2nd ed. McLean, VA: Degnon Associates, 2000:483–532. 11. Whitehead WE, Wald A, Diamant NE, Enck P, Pemberton JH, Rao SS. Functional disorders of the anus and rectum. Gut 45(Suppl 2):55–9, 1999. 12. Drossman DA, Li Z, Andruzzi E, et al. U.S. Householder Survey of functional gastrointestinal disorders. Dig Dis Sci 1993;38:1569 – 80. 13. Ware JE, Sherbourne CD. The MOS 36-item short form Health Survey (SF-36): I. Conceptual framework and item selection. Med Care 1992;30:473– 83. 14. Burnett C, Whitehead WE, Drossman D. Psychological distress and impaired quality of life in patients with functional anorectal disorders. Gastroenterology 1998; 14:A729. 15. Lahr CJ. Evaluation and treatment of incontinence. Practical Gastroenterol 1988;12:27–35. 16. Barnett JL, Raper SE. Anorectal diseases. In: Yamada T, Alpers DH, Owyang C, Powell DW, Silverstein FE, eds. Textbook of gastroenterology. New York: JB Lippincott, 1991:1813–32. 17. Diamant NE, Kamm MA, Wald A, Whitehead WE. AGA technical review on anorectal testing techniques. Gastroenterology 1999;116:735– 60. 18. Enck P. Biofeedback training in disordered defecation: a critical review. Dig Dis Sci 1993;38:1953– 60. 19. Rao SS, Enck P, Loening-Baucke V. Biofeedback therapy for defecation disorders. Dig Dis Sci 1997;15(Suppl 1):78 –92. 20. Dailianas A, Skandalis N, Rimikis MN, Koutsomanis D, Kardasi M, Archimandritis A. Pelvic floor study in patients with obstructed defecation: influence of biofeedback. J Clin Gastroenterol 2000;30:176 – 80. 21. Weisel PH, Norton C, Roy A, Storrie JB, Bowers J, Kamm MA. Gut focused behaviour treatment (biofeedback) for constipation and faecal incontinence in multiple sclerosis. J Neurol Neurosurg Psychiatry 2000;69:240 –3.
1216
HEYMEN ET AL
22. McKee RF, McEnroe L, Anderson JH, Finlay IG. Identification of patients likely to benefit from biofeedback for outlet obstruction constipation. Br J Surg 1999;86: 355–9. 23. Rhee PL, Choi MS, Kim YH, et al. An increased rectal maximum tolerable volume and long anal canal are associated with poor short-term response to biofeedback therapy for patients with anismus with decreased bowel frequency and normal colonic transit time. Dis Colon Rectum 2000;43:1405–11. 24. Heymen S, Wexner S, Vickers D, Nogueras J, Weiss E, Pikarsky A. Prospective randomized trial comparing four biofeedback techniques for patients with constipation. Dis Colon Rectum 1999;42:1388 –93. 25. Karlbom U, Hallden M, Eeg-Olofsson KE, Pahlman L, Graf W. Results of biofeedback in constipated patients: a prospective study. Dis Colon Rectum 1997;40: 1149 –55. 26. Rao SS, Welcher KD, Pelsing RE. Effects of biofeedback on anorectal function in obstructed defecation. Dig Dis Sci 1997;42:2197–205. 27. Patankar SK, Ferrara A, Levy JR, Larach SW, Williamson PR, Perozo SE. Biofeedback in colorectal practice: a multicenter, statewide, three-year experience. Dis Colon Rectum 1997;40:827–31. 28. Rieger NA, Wattchow DA, Sarre RG, et al. Prospective study of biofeedback for treatment of constipation. Dis Colon Rectum 1997;40:1143– 8. 29. Glia A, Glyin M, Gullberg K, Lindberg G. Biofeedback retraining in patients with functional constipation and paradoxical puborectalis contraction: comparison of anal manometry and sphincter electromyography for feedback. Dis Colon Rectum 1997;40:889 –95. 30. Ko CL, Tong J, Lehman RE, Shelton AA, Schrock TR, Weldon ML. Biofeedback is effective therapy for fecal incontinence and constipation. Arch Surg 1997;132: 829 –33. 31. Ho Y, Tan M. Biofeedback for bowel dysfunction following low anterior resection. Ann Aced Med Singapore 1997;26:299 –302. 32. Ho YH, Tan M, Goh HS. Clinical and physiologic effects of biofeedback in outlet obstruction constipation. Dis Colon Rectum 1996;39:5, 520 – 4. 33. Koutsomanis D, Lennard-Jones JE, Roy AJ, Kamm MA. Controlled randomized trial of visual biofeedback versus muscle training without a visual display for intractable constipation. Gut 1995;37:95–9. 34. Hull TL, Fazio VW, Schroeder T. Paradoxical contraction in patients after pelvic pouch construction. Dis Colon Rectum 1995;38:1144 – 6. 35. Papachrysostomou M, Smith AN. Effects of biofeedback on obstructive defecation: reconditioning of the defecation reflex? Gut 1994;35:252– 6. 36. Koutsomanis D, Lennard-Jones JE, Kamm MA. Prospective study of biofeedback treatment for patients with
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
Dis Colon Rectum, September 2003
slow and normal transit constipation. Eur J Gastroenterol Hepatol 1994;6:131–7. Bleijenberg G, Kuijpers HC. Biofeedback treatment of constipation: a comparison of two methods. Am J Gastroenterol 1994;89:1021– 6. Wexner SD, Cheape JD, Jorge JMN, Heymen S, Jagelman DG. Prospective assessment of biofeedback for the treatment of paradoxical puborectalis contraction. Dis Colon Rectum 1992;35:2, 145–50. Turnbull GK, Ritvo PG. Anal sphincter biofeedback relaxation treatment for women with intractable constipation symptoms. Dis Colon Rectum 1992;35:530 – 6. Fleshman JW, Dreznik Z, Meyer K, Fry RD, Carney R, Kodner IJ. Outpatient protocol for biofeedback therapy of pelvic floor outlet obstruction. Dis Colon Rectum 1992;35:1–7. Kawimbe BM, Papachrysostomou M, Binnie NR, et al. Outlet obstruction constipation (anismus) managed by biofeedback. Gut 1991;32:1175–9. Lestar B, Penninckx F, Kerremans R. Biofeedback defecation training for anismus. Int J Colorect Dis 1991; 202–7. Dahl J, Lindquist BL, Tysk C, Leissner P, Philipson L, Jarnerot G. Behavioral medicine treatment in chronic constipation with paradoxical anal sphincter contraction. Dis Colon Rectum 1991;34:769 –76. Emery Y, Descos L, Meunier P, Louis D, Valancogne G, Weil G. Terminal constipation with abdominopelvic asynchronism: analysis of etiological, clinical and manometric findings, and the results of biofeedback therapy. Gastroenterol Clin Biol 1988;12:6 –11. Bleijenberg G, Kuijpers HC. Treatment of the spastic pelvic floor syndrome with biofeedback. Dis Colon Rectum 1987;30:108 –11. Weber J, Ducrotte PH, Touchais JY, Roussignol C, Denis PH. Biofeedback training for constipation in adults and children. Dis Colon Rectum 1987;30:11, 844 – 6. Nolan T, Catto-Smith T, Coffey C, Wells J. Randomized controlled trial of biofeedback training in persistent encopresis with anismus. Arch Dis Child 1998;79:131–5. Ponticelli A, Iacobelli BD, Silveri M, Broggi G, Rivosechi M, De Gennaro M. Colorectal dysfunction and faecal incontinence in children with spina bifida. Br J Urol 1998;81(Suppl 3):117–9. van der Plas RN, Benninga MA, Redekop WK, Taminiau JA, Buller HA. Randomized trial of biofeedback training for encopresis. Arch Dis Child 1996;75:367–74. Cox DJ, Sutphen J, Borowitz S, Dickens MN, Singles J, Whitehead WE. Simple electromyographic biofeedback treatment for chronic pediatric constipation/encopresis: preliminary report. Biofeedback Self-Regulation 1994; 19:41–50. Benninga MA, Buller HA, Taminiau JA. Biofeedback training in chronic constipation. Arch Dis Child 1993; 68:126 –9.
Vol. 46, No. 9
BIOFEEDBACK TREATMENT OF CONSTIPATION
52. Loening-Baucke V. Persistence of chronic constipation in children after biofeedback treatment. Dig Dis Sci 1991;36:153– 60. 53. Loening-Baucke V. Modulation of abnormal defecation dynamics by biofeedback treatment in chronically constipated children with encopresis. J Pediatrics 1990;116: 214 –22. 54. Keren S, Wagner Y, Heldenbert D, Golan M. Studies of manometric abnormalities of the rectoanal region during defecation in constipated and soiling children: modification through biofeedback therapy. Am J Gastroenterol 1988;83:827–31. 55. Veyrac M, Granel D, Parelon G, Michel H. Evaluation of biofeedback in childhood idiopathic constipation. Pediatric 1987;42:719 –21. 56. Wald A, Chandra R, Gabel S, Chiponis D. Evaluation of biofeedback in childhood encopresis. J Pediatr Gastroenterol Nutr 1987;6:554 – 8. 57. Olness K, McFarland FA, Piper J. Biofeedback: a new modality in management of children with fecal soiling. J Pediatrics 1980;96:505–9. 58. Loening-Baucke V. Factors determining outcome in children with chronic constipation and faecal soiling. Gut 1989;30:999 –1006. 59. Rao SS. The technical aspects of biofeedback therapy for defecation disorders. J Clin Gastroenterol 2000;6: 96 –103. 60. Sangwan YP, Coller JA, Barrett RC, Roberts PL, Murray JJ, Schoetz DJ Jr. Can manometric parameters predict response to biofeedback therapy in fecal incontinence? Dis Colon Rectum 1995;38:1021–5. 61. Beninga MA, Buller, HA, Staalman CR, et al. Defaecation disorders in children, colonic transit time versus the BARR-Score. Eur J Pediatr 1995;154:277– 84. 62. Chiotakakou-Faliakou E, Kamm MA, Roy AJ, Storrie JB, Turner IC. Biofeedback provides long term benefit for patients with intractable, slow and normal transit constipation. Gut 1998;42:517–21. 63. Gilliland R, Heymen S, Altomare DF, Park UC, Vickers D, Wexner SD. Outcome and predictors of success of biofeedback for constipation. Br J Surg 1997;84:1123– 6. 64. Guerrero RA, Cavender CP. Constipation: physical and psychological sequelae. Pediatr Ann 1999;28:312– 6. 65. Derogatis L. SCL-90-R: administration, scoring, and procedures manual. Minneapolis: National Computer Systems, 1994.
1217
66. Heymen S, Wexner SD, Gulledge AD. MMPI assessment of patients with functional bowel disorders. Dis Colon Rectum 1993;36:593– 6. 67. Devroede G, Girard G, Bouchoucha M, et al. Idiopathic constipation by colonic dysfunction. Relationship with personality and anxiety. Dig Dis Sci 1989;34:9, 1428 –33. 68. Kumar D, Bartolo DC, Devroede G, et al. Symposium on constipation. Int J Colorect Dis 1992;7:47– 67. 69. Nehra V, Bruce B, Rath-Harvey DM, Pemberton JH, Camilleri M. Psychological disorders in patients with evacuation disorders and constipation in a tertiary practice. Am J Gastroenterol 2000;95:1755– 8. 70. Whitehead WE. Illness behaviour. In: Kamm NA, Lennard-Jones JE, eds. Constipation. Petersfield: Wrightson Biomedical Publishing, 1993:95–100. 71. Friman PC, Mathews JR, Finney JW, Christopherson ER, Leibowitz JM. Do encopretic children have clinically significant behavior problems? Pediatrics 1988;82: 407–9. 72. Loening-Baucke V, Cruikshank B, Savage C. Defecation dynamics and behavior profiles in encopretic children. Pediatrics 1987;80:672–9. 73. Keck JO, Staniunas RJ, Coller JA, et al. Biofeedback training is useful in fecal incontinence but disappointing in constipation. Dis Colon Rectum 1994;37:1271– 6. 74. Dykes S, Smilgin-Humphreys S, Bass C. Chronic idiopathic constipation: a psychological enquiry. Eur J Gastroenterol Hepatol 2001;1:39 – 44. 75. Strickland M, Heymen S. Psychiatric treatment of constipation. In: Wexner SD, Bartolo DC, eds. Constipation: etiology, evaluation and management. London: Butterworth Heinemann, 1995:251– 61. 76. Kuijpers HC. Application of the colorectal laboratory in diagnosis and treatment of functional constipation. Dis Colon Rectum 1990;33:35–9. 77. Lennard-Jones JE. Clinical management of constipation. Pharmacology 1993;47(Suppl 1):216 –23. 78. Leroi AM, Berkelmahs I, Denis P, Hemond M, Devroede G. Anismus as a marker of sexual abuse. Consequences of abuse on anorectal motility. Dig Dis Sci 1995;40: 1411– 6. 79. Jones KR, Heymen S, Whitehead WE. Biofeedback for anorectal disorders. In: Drutz HP, Hershorn S, Diamant NE, eds. Urogynaecology and reconstructive pelvic surgery. London: Springer-Verlag, 2003:313–25.