• anorectal manometry;
  • balloon expulsion testing;
  • defecography;
  • pelvic floor dyssynergia


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information


Dyssynergic defecation (DD) results from inadequate relaxation of the pelvic floor on attempted defecation. The prevalence of DD in patients with chronic constipation (CC) is not certain. Aims of this study are to estimate the prevalence of abnormal findings associated with DD across testing modalities in patients referred for physiological testing for CC.


Systematic search of MEDLINE, EMBASE and PUBMED databases were conducted. We included full manuscripts reporting DD prevalence in CC, and specific findings at pelvic floor diagnostic tests. Random effects models were used to calculate pooled DD prevalences (with 95% CI) according to individual tests and specific findings.

Key Results

A total of 79 studies on 7581 CC patients were included. The median prevalence of any single abnormal finding associated with DD was 37.2%, ranging from 14.9% (95% CI 7.9–26.3) for absent opening of the anorectal angle (ARA) on defecography to 52.9% (95% CI 44.3–61.3) for a dyssynergic pattern on ultrasound. The prevalence of a dyssynergic pattern on manometry was 47.7% (95% CI 39.5–56.1). The prevalence of DD was similar across specialty and geographic area as well as when restricting to studies using Rome criteria to define constipation.

Conclusions & Inferences

Dyssynergic defecation is highly prevalent in CC and is commonly detected across testing modalities, type of patient referred, and geographical regions. We believe that the lower prevalence of findings associated with DD by defecography supports use of manometry and balloon expulsion testing as an initial evaluation for CC.


anorectal manometry


chronic constipation


confidence interval


dyssynergic defecation




magnetic resonance imaging




  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

Chronic constipation (CC) is a prevalent disorder, which according to recent estimates, has a global prevalence of 14% (95% CI: 12–17%) and is associated with significant cost and health care utilization.[1, 2] Subtypes of CC include normal and slow-transit constipation and dyssynergic defecation (DD), which refers to the paradoxical contraction or inadequate relaxation of the pelvic floor on attempted defecation.[3] As symptoms do not reliably discriminate between subtypes of CC, diagnostic tests are frequently required.[3] Distinguishing patients with DD from other subtypes of CC is important clinically as they may be less likely to respond to traditional medical and dietary therapies and may be more likely to respond to biofeedback therapy.[4-6]

The Rome Criteria require a combination of two abnormal dynamic tests of the pelvic floor on attempted defecation [i.e., impaired evacuation on balloon expulsion or defecography, and inappropriate contraction of the pelvic floor muscles or incomplete relaxation of the anal sphincter on manometry, electromyography (EMG), or imaging) to diagnose DD (Table 1).[7] The clinical utility and application of the Rome Criteria for DD have not been well evaluated and the individual pelvic floor dynamic tests lack rigorous evaluation of their testing characteristics, such as test-retest reliability, validity, accuracy, and clinical utility.[1, 8]

Table 1. Rome 3 diagnostic criteria* for functional defecation disorders
  1. *Criteria fulfilled for the last 3 months with symptom onset at least 6 months prior to diagnosis.

  2. **Diagnostic criteria for functional constipation:

  3. 1. Must include two or more of the following: (a) Straining during at least 25% of defecations, (b) Lumpy or hard stools at least 25% of defecations, (c) Sensation of incomplete evacuation at least 25% of defecations, (d) Sensation of anorectal obstruction/blockage at least 25% of defecations, (e) Manual maneuvers to facilitate at least 25% of defecations (e.g., digital evacuation, support of the pelvic floor), (f) Fewer than three defecations per week.

  4. 2. Loose stools are rarely present without the use of laxatives.

  5. 3. There are insufficient criteria for IBS.

1. The patient must satisfy diagnostic criteria for functional constipation**
2. During repeated attempts to defecate must have at least two of the following:
a. Evidence of impaired evacuation, based on balloon expulsion test or imaging
b. Inappropriate contraction of the pelvic floor muscles (i.e., anal sphincter or puborectalis) or less than 20% relaxation of basal resting sphincter pressure by manometry, imaging, or EMG
c. Inadequate propulsive forces assessed by manometry or imaging

We conducted a systematic review of studies that included patients with CC, who underwent a variety of pelvic floor function testing to determine the prevalence and potential predictors of abnormal findings.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

Search strategy

We conducted a comprehensive literature search of MEDLINE and EMBASE databases (1966–2011). The search terms included combinations of the following keywords: pelvic floor imaging, balloon expulsion test, anorectal manometry, defecography, magnetic resonance, ultrasound (US), electromyography (EMG), constipation, anismus, obstructed defecation, dyssynergia, dyssynergic defecation, defecatory disorders, pelvic floor dysfunction, pelvic floor dyssynergia (PFD), and gastrointestinal motility. A manual search of the references listed by studies retrieved from the online databases and from previously published systematic reviews was also performed to identify additional studies of interest. The search was limited a priori to studies that were published as a full text papers in English.

Selection of studies

Details of inclusion criteria and variables assessed and extracted are outlined in Table 2. We included studies in which unselected patients with constipation were evaluated with one or more diagnostic test. Specific inclusion criteria were as follows: (i) patients were unselected (did not have an established diagnosis of DD or slow-transit constipation), (ii) criteria for a positive finding consistent with DD were defined by the author, and (iii) the prevalence of the positive finding was reported.

Table 2. Inclusion criteria and variables assessed and recorded for each study
  1. MRI, magnetic resonance imaging; EMG, electromyography; DD, dyssynergic defecation; PR, puborectalis.

Study features

Performance of one or more among:

● Anorectal manometry

● Balloon expulsion test

● Defecography

 ○ Conventional

 ○ MRI

 ○ Ultrasound


Population (N > 20) included unselected patients with constipation

Raw data available from manuscript on:

● Specific criteria for dyssynergic defecation, pelvic floor dyssynergia or anismus

● The prevalence of the finding

Study population

Mean age

Gender distribution

Criteria for constipation (though studies not using specific criteria were included in the analysis)

Geographic region

Department of corresponding author of study where the population was evaluated (Gastroenterology, Surgery, Radiology)

Test characteristics

Criteria for positivity

Prevalence of one or more of eight specific findings:

1. Anorectal manometry: a pattern consistent with DD (studies reporting only pressures were not included)

2. Defecography (conventional or MRI) or ultrasound showing:

a. Paradoxical motion of the PR

b. Impaired relaxation of the PR

c. Prominent or persistent impression of the PR

d. Failure to open the anorectal angle

e. Impaired or failed evacuation of contrast

f. Excessive perineal descent

3. EMG: PR activity present or increased

Prevalence of positive findings on colonic transit and digital rectal exam where reported

Data extraction

Two investigators (FC, EV) extracted the data. Discrepancies regarding individual study inclusion, data extraction, and interpretation were resolved by consensus prior to the final analysis. Details of the extracted data are presented in Table 2 and included demographic data, test modality, specific finding, and prevalence of the finding.

Meta-analysis and subgroup analyses

All statistical analyses were performed using Comprehensive Meta-Analysis 2.2 (Biostat, Englewood, NJ, USA). The prevalence and 95% CI of DD was calculated for each finding in each study. Statistical heterogeneity across the various studies was then tested with the use of Q-statistic.[9] A P value <0.10 indicated a significant statistical heterogeneity across studies, and a random effects model was used.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

Eligible studies

The meta-analysis flow is shown in Fig. 1. A total of 1857 records were screened, and after 401 full text articles were reviewed, 82 articles that met the search criteria were included.[10-91]


Figure 1. Meta-analysis flow.

Download figure to PowerPoint

Characteristics of included studies

The 82 included articles were published from 1986 to 2011 (Table 3). After careful review and after contacting authors, it was determined that the 82 articles comprised 79 unique patient cohorts. The majority (51.9%) of cohorts were studied by investigators from surgical departments, followed by gastroenterology (GI) (38.0%) and radiology (10.1%). 55.7% of cohorts were studied in Europe, 25.3% in the United States and 8.9% in Asia with the remaining 10.1% conducted in South America, Australia, and Africa. The combined mean age of the patients was 51.2 years and 81.1% were women.

Table 3. Characteristics and findings of included studies
Study nameMean ageSpecialty, regionPercent with findings associated with DD by test modality (specific criteria used)
  1. SD, standard deviation; SEM, standard error of the mean; ARM, anorectal manometry; US, ultrasound; EMG, electromyography; MRI, magnetic resonance imaging; GI, gastroenterology; Surg., surgery; EU, Europe; US, United States; Af., Africa; S.Am, South America; Defecography: ARA, anorectal angle; PR-P, puborectalis paradoxical; PR-NR, puborectalis non-relaxing, CE-I, contrast evacuation impaired; CE-F, contrast evacuation failed; PI-PR, prominent impression of the puborectalis; PD, excessive perineal descent; EMG: I, activity increased; I/NC, activity increased or no change; P, activity present; DD, dyssynergic defecation.

Agachan[10]65.5 (12–95)Surg., US  

38 (PD)

34 (PRNR)

Alstrup[11]49.5 (18–85)Surg., EU  27 (CE-F)37 (I/NC)  
Bannister[12]32 (14–53)GI, EU 21 (50/5)    
Barnes[13]Not availableGI, EU 84 (50)    
Barthet[14]51 (30–74)GI, EU  67 (PD)   
Bartolo[15]56 (25–74)Surg., EU  22(PI-PR/ARA)   
Bennani[16]45 (SD 16)GI, Africa47     
Bharucha[17]40 (SEM 2.1)GI, US 44 (unable)   46 (MRI:PR-NR)
Boonjunwetwat[18]50.1 (22–86)Rad., Asia  

65 (PD)

4 (ARA)

Bordeianou[19]50.5 (SD 16.4)Surg., US 47 (60/5)51 (PR/EAS-NR+CE-I)52 (P)  
Bouchoucha[20]37 (SD 15)GI, EU40     
Brusciano[21]51.7 (19–81)Surg., EU60 56 (PRNR) 48 (PR-NR)71 (DRE)
Brusciano[22]51.3 (21–71)Surg., EU68 69 (PRNR) 50 (PR-NR)77 (DRE)
Choi[23]61 (20–81)Rad., EU  47 (PR-NR)   
Dailianas[24]40.2 (15–75)GI, EU  8 (PR-P)58 (I)  
Fink[25]41.3 (15–75)Surg., Aus   21 (P)  
Ger[26]60 (18–84)Surg., US63 

40 (CE-I)

36 (ARA/CE-I)

7 (CE-F)

38 (I/NC)  
Gladman[27]Not availableSurg., EU  

11 (PR-NR)

5 (CE-I)

Glia[28, 29]52 (17–79)Surg., EU 23 (60 mL)

44 (CE-I/F)

34 (ARA/PI-PR)

10 (PD)

31 (P)  
Gosselink[30]50 (18–75)Surg., EU  24 (ARA)   
Guo[31]44 (16–82)GI, Asia51    63 (DRE)
Habib[32]40.5 (20–69)GI, EU  

53 (CE-I)

38 (PD)

19 (ARA/EAS-P)

34 (I)  
Halligan[33]Not availableRad., EU 53 (2.5 cm)    
Halligan[34]Not availableRad., EU  35 (PR-P)   
Johansson[35]45 (20–73)Surg., EU  15 (PI-PR)60 (I)  
Jones[36]46 (17–74)GI, EU 63 (50 mL)34 (PD)76 (I)  
Jorge[37]59 (12–83)Surg., US  38 (PR-P+PR-NR+CE-I)36 (I)  
Karasick[38]47 (18–79)Rad., US  

11 (PR-NR)

6 (PD)

Karlbom[39]49 (22–87)Surg., EU  14 (PI-PR)   
Karlbom[40-42]51 (20–93)Surg., EU 24 (40 mL per 3 min)25 (PI-PR+/ARA) (data from 1999)

83 (P)

24 (I)

Katsinelos[43]42 (29–53)GI, EU 21 (50 mL per 5 min)    
Kuijpers[44]42 (15–82)Surg., EU  74 (CE-I)   
Lee[45]42 (19–79)Surg., Asia  

40 (PI-PR)

29 (PD)

Liu[46]45 (28–72)Surg., Asia  5 (PR-P)   
Martellucci[47]59 (29–83)Surg., EU  6 (PI-PR) 13 (ARA) 
Mertz[48]51.1 (SD 17.3)GI, US61  59 (I)  
Mibu[49]46 (20–78)Surg., Asia 24

52 (CE-F)

30 (PD)

Miller[50]34.3 (16–60)Surg., EU  58 (CE-I)   
Minguez[51]38.4 (SD 15)GI, EU2225 (60 mL per 1 min)18 (PR-NR/ARA/CE-I)   
Mohammed[52]53 (19–83)GI, EU  66 (CE-I)   
Murad-Regadas[53]47.7 (24–79)Surg., S Am  30 (PR-NR) 27 (ARA) 
Murad-Regadas[54]52.7 (SD 15)Surg., US  

54 (PD)

27 (PR-NR/P)

Murad-Regadas[55]51.4 (13–91)Surg., S Am    40 (ARA) 
Murad-Regadas[56]50 (24–78)Surg., S Am39   47 (DYS) 
Murad-Regadas[57]43 (23–74)Surg., EU  

55 (PR-P)

41 (PD)

 59 (EAS-P) 
Nielsen[58]61 (24–85)Rad., EU  9 (PR-NR)   
Oncu[59]Not availableGI, EU 33 (5 min)    
Pfeifer[60]72 (15–86)Surg., US  

32 (CE-I)

23 (PR-NR)

6 (ARA)

Privitera[61]52 (12–83)Surg., EU3773 (5 min)    
Prokesch[62]44 (20–82)Rad., EU  13 (PR-NR)   
Rantis[63]54 (21–81)Surg., US  8 (PR-P)6 (P)  
Rao[64]44.5 (16–81)GI, US48     
Rao[65]44 (16–81)GI, US5157 (50 mL per 5 min)    
Rao[66]53 (SD 18)GI, US70     
Raza[67]44 (SEM 1.4)GI, US4538 (50 mL per 2 min)    
Regadas[68]53.4 (26–77)Surg., S Am  22 (ARA) 30 (ARA) 
Reiner[69]48 (20–76)Rad., EU     42 (MR:PR-P)
Roberts[70]43 (19–77)Surg., EU  30 (CE-I)13 (CE-F)

42 (P)

37 (I)

Roe[71]48 (21–73)Surg, EU 32 (50 mL)    
Roe[72]Not availableSurg., EU  9 (PI-PR)   
Schouten[73]51.5 (19–87)Surg., EU 72 (urge, 20s)18 (ARA)56 (I/NC)  
Seidl[74]57 (20–87)GI, EU  21 (CE-I)   
Shafik[75]48.2 (24–70)Surg., Af   63 (I)  
Shouler[76]28 (18–52)Surg., EU 40 (150 mL)32 (ARA)36 (I)  
Siproudhis[77]41 (SEM 2.3)GI, EU34 

59 (PD)


Spazzafumo[78]51 (SD 17)GI, EU  

58 (PD)

53 (CE-I)

38 (ARA)

Sunderland[79]47 (13–88)Surg., EU  

21 (PD)

18 (CE-F)

6 (PR-P)

2 (ARA)

Takao[80]63 (17–86)Surg., US  

49 (CE-I)

30 (PD)

Tantiphlachiva[81]41 (SD 15.3)GI, US8845 (50 mL, 1 min)   73 (DRE)
Tjandra[82]40 (24–77)Surg., Aus   46 (I)  
Tomita[83]46.5 (15–73)Surg., Asia  

19 (PD)

16 (PR-NR/P)

Touchais[84]47 (25–80)GI, EU  56 (PD)   
Van Outryve[85]53.1 (18–82)GI, EU    80 (PR-P) 
Voderholzer[86]52.9 (SEM 1.6)GI, EU41     
Wald[87]43 (16–78)GI, US20     
Wald[88]46 (24–84)GI, US31 

25 (CE-I)

19 (PR-P)

Wald[89]45.7 (19–80)GI, US21 29 (CE-F)   
Xiong[90]35.6 (19–72)GI, Asia   75 (P)  
Yeh[91]62.1 (SD 17.9)Surg., US  22 (PR-P + CE-I)33 (I/NC)  

The data from the majority of the studies (78.5%) were collected prospectively. Two studies (2.5%) were clinical trials where baseline characteristics were reported. The remaining 19.0% were conducted retrospectively.

Rome I,[92] II[93] or III[7] criteria for functional constipation were used for the diagnosis of chronic constipation in 17 studies. In 32 studies, there was no further qualification of symptoms beyond the diagnosis of constipation. Detail of patient characteristics is presented in the online supplement.

The prevalence of each of eight specific findings (Table 2) ranged from 14.9% to 52.9%. The median prevalence was 37.2% and the trimmed mean excluding the highest and lowest prevalence was 34.6%.


Twenty studies on ARM findings were included. These studies were published between 1989 and 2010 and included 1784 patients with CC. Half of these studies were conducted in the United States and 35.0% in Europe. 75.0% were from GI departments. There were no studies using high-resolution manometry. The pooled prevalence (95% CI) for a dyssynergic pattern (paradoxical contraction or inadequate relaxation of the anal sphincter on attempted defecation) was 47.7% (39.5–56.1).

Balloon Expulsion

Nineteen studies reported results from balloon expulsion testing (BET). These studies were published from 1985 to 2011 and included 1,654 patients with CC. 68.4% of patients were from Europe and 26.3% from the United States. GI and surgery specialists authored 53.6% and 42.1% of the studies, respectively. The test was performed in the seated position in 13 studies. The pooled prevalence (95% CI) for failed or impaired balloon expulsion by any criteria was 42.5% (34.2–51.3). When limited to the seven studies requiring the patient to be unable to expel the balloon after 5 min, the prevalence (95% CI) was 43.1% (30.3–56.8). When the analysis was repeated including only studies that performed the test with the patient seated on a commode, the pooled prevalence was 40.3% (30.3–51.1) for any criteria and 37.1% (22.3–54.9) for criteria requiring >5 min.


Fifty-two studies reported results from defecography. Studies were published from 1986 to 2011 and included 5267 patients with CC. The majority were conducted in Europe (64.2%), and by investigators in surgical departments (66.0%). These studies were analyzed according to specific positive finding and the pooled prevalence ranged from 14.9% (7.9–26.3) for absent opening of the ARA to 37.4% (27.2–48.9) for excessive perineal descent. We also reported pooled prevalence combining different criteria. Because several studies reported more than one specific finding, we performed this analysis twice including either the finding with the highest or the lowest prevalence from each study.


Twenty studies, published from 1986 to 2011, evaluated EMG including 1507 patients. Fifteen studies (75.0%) were from surgical departments and 50.0% were conducted in Europe. The pooled prevalence (95% CI) for increased activity of the puborectalis (PR) and either present or increased activity were 44.0% (39.8–57.8) and 47.7% (39.2–56.3), respectively.

Magnetic resonance imaging (MRI)

Two studies evaluated MRI in a combined 100 patients with CC. One study was conducted in the United States from a GI department[17] and the other from a radiology department in Europe.[69] The pooled prevalence of a non-relaxing or paradoxical PR on MRI was 44.0% (34.6–53.9).


A total of 814 patients in nine studies were evaluated by US between 2001 and 2011. Five studies were published by one group in Brazil and the rest were published in Europe. One study was published from a GI department and the rest were from surgical departments. The pooled prevalence ranged from 27.9% (17.9–40.7) for abnormalities of the ARA alone to 52.9% (44.3–61.3) for either dyssynergia or abnormalities of the ARA.

Sensitivity and subgroup analyses

The prevalence of positive findings varied considerably when analyses were repeated for each specialty and for studies using Rome criteria for inclusion. There was a trend for prevalence to be higher in studies published from surgical departments, but there was considerable overlap in the 95% CI. For example, the prevalence (95% CI) of DD by manometry was 45.5% (35.4–56.0) and 54.0% (41.6–65.9) for studies published from GI (N = 15) and surgery (N = 5), respectively. For the finding of a non-relaxing PR on defecography, prevalence was again highest among studies published from surgical departments at 27% (95% CI: 18.9–37.1). Prevalence (95% CI) in studies from radiology and GI departments was 20.4% (9.4–38.6) and 32.1% (15.2–55.4), respectively. Results also were not affected by alternative exclusion of two studies that may have included an overlapping patient population[39, 41] or when analyses were repeated including only studies with either prospective or retrospective designs (data not shown).

Comparison of tests within populations and concurrent positivity

Rates of concurrent positive test results and prevalence by combination of different test modalities are shown in Table 4. There was not enough consistency among studies to perform a meta-analysis.

Table 4. Rates of concurrent positivity and combined prevalences
RefTestATestB% Positive for TestB among those positive for TestA B+/A+ (%)% Positive for TestA among those positive for TestB B+/A+ (%)% Both A and B positive
  1. ARM, anorectal manometry; BET, balloon expulsion testing; US, ultrasound; EMG, electromyography; MRI, magnetic resonance imaging; GI, gastroenterology; Surg., surgery; EU, Europe; US, United States; Af., Africa; S.Am, South America; Defecography: ARA, anorectal angle; PR-P, puborectalis paradoxical; PR-NR, puborectalis non-relaxing; CE-I, contrast evacuation impaired; CE-F, contrast evacuation failed; PI-PR, prominent impression of the puborectalis; PD, excessive perineal descent; EMG: (I), activity increased; (I/NC), activity increased or no change; (P), activity present.

[65] ARMBET16/18 (88.9)16/20 (80.0)16/35 (45.7)
[66] ARMBET42/70 (60.0) 42/100 (42.0)
[22] ARMUS39/57 (68.4) 39/84 (46.4)
[56] ARMUS16/19 (84.2) 16/49 (32.7)
[21] ARMUS (PR-NR) 39/44 (88.6)39/92 (42.4)
[47] ARMUS (ARA) 5/7 (71.4)5//54 (9.3)
[47] ARMDEF (PI-PR) 2/3 (66.7)2/54 (3.7)
[88] ARMDEF (PR-P)4/11 (36.4)4/7 (57.1)4/36 (11.1)
[27] ARMDEF (PR-NR) 24/56 (42.9)24/508 (4.7)
[27] ARMDEF (CE-I) 15/24 (62.5)15/508 (3.0)
[66] ARMDEF (CE-I)26/70 (37.1) 26/100 (26.0)
[17] ARMMRI (PD) 7/11 (63.6)7/52 (13.5)
[19] DEF (PR-NR/EAS-NR+CE-I)BET33/63 (52.4)33/58 (56.9)33/123 (26.8)
[21] DEF (PR-NR)US (PR-NR)22/24 (91.7)22/22 (100.0)22/43 (51.2)
[68] DEF (ARA)US (ARA)16/19 (91.7)16/26 (61.5)16/86 (18.6)
[57] DEF (PR-P)US (EAS-P)14/16 (87.5)14/17 (82.4)14/29 (48.3)
[11] DEF (CE-F)EMG (PR-I/NC)5/8 (62.5)5/11 (45.5)5/30 (16.7)
[19] DEF (PR-NR/EAS-NR+CE-I)EMG (PR-P)31/63 (49.2)31/64 (48.4)31/122 (25.4)
[26] DEF (ARA/CE-I)EMG (PR-I/NC)19/42 (45.2)19/44 (43.2)19/116 (16.4)
[29] DEF (CE-I/F)EMG (PR-P)33/59 (55.9)33/42 (78.4)33/134 (24.6)
[32] DEF (ARA/anal canal <10 mm)EMG (PR-I)11/22 (50.0)11/11 (100.0)11/32 (34.4)
[32] DEF (PD)EMG (PR-I) 4/12 (33.3)4/32 (12.5)
[32] DEF (CE-I)EMG (PR-I) 8/17 (47.1)8/32 (25.0)
[37] DEF (PR-P+NR+CE-I)EMG (PR-I)28/42 (66.7)28/40 (70.0)28/112 (25.0)
[70] DEF (CE-I)EMG (PR-I)14/21 (66.7)14/26 (53.8)14/71 (19.7)
[91] DEF (PR-P+CE-I)EMG (PR-I)7/58 (12.1)7/21 (33.3)7/64 (10.9)
[10] DEF (PR-NR)EMG (PR-NR)115/167 (68.9) 115/487 (23.6)
[43] BETEMG (EAS-I) 4/4 (100.0)4/19 (21.1)
[36] BETEMG (PR-I)18/22 (81.8)18/20 (90.0)18/32 (56.3)

The prevalence of positive BET in patients with dyssynergia on ARM (concurrent positive BET) was reported in two studies (combined N of 135) from the same center, a referral GI practice in the United States. In one study, which was a prospective evaluation of patients undergoing biofeedback therapy that reported baseline characteristics, 88.9% subjects with positive ARM failed to expel the balloon,[65] while in the other, also a prospective study, the rate of concurrent positive BET was 60.0%.[66] The prevalence of having both dyssynergia on ARM and failure to expel the balloon was 45.7% and 42.0%.

Four studies evaluated concurrent positivity of US and ARM. These were conducted by three groups, all in surgical departments, one in Brazil[56] and two in Europe[21, 22, 47] and included a total of 279 patients. Concurrent positivity for US among those with positive ARM ranged from 68.4% to 88.6% and the median prevalence of combined positivity (positive on both tests) was 37.6% (range 9.3–46.4).

Four studies (three from GI departments in the United States and one from a surgical department in Europe), compared defecography with ARM in a total of 698 patients. The rate of positive defecography among those with a dyssynergic pattern on ARM (concurrent positive defecography) was reported in two studies.[66, 88] The prevalence of a dyssynergic pattern on ARM was 30.6% in one study[88] and 70.0% in the other[66]; however, the rate of concurrent positive defecography was similar at 36.4% (paradoxical motion of the PR) and 37.1% (impaired contrast evacuation), respectively. For the three studies that reported rates of positive ARM among those with positive defecography (concurrent positive ARM), the median rate of concurrent positivity was 59.8% (range: 42.9–66.7). The median prevalence of combined positive ARM and defecography was 4.7% (range: 3.0–26.0).

In the three studies that compared defecography with US, agreement was high with 87.5–91.7% of patients positive by defecography having the corresponding findings on US as well (concurrent positive US). These evaluated 158 patients and were conducted by two groups from surgical departments; one in Europe,[21, 22] one in Brazil.[56, 68]

Ten studies compared defecography with EMG in a total of 1168 patients. One study was from a GI department in the United States, five were from surgical departments in the United States and three were from surgical departments in Europe. The median prevalence of positive EMG among patients with positive findings on defecography (concurrent positive EMG) was 55.9% (range: 12.1–68.9). For positive defecography among those with positive findings on EMG (concurrent positive defecography), the median prevalence was 51.1% (range: 33.3–100.0). Median prevalence of having both tests positive was 23.6% (range: 10.9–34.4).


Higher mean age at study level was predictive of increased prevalence of abnormal BET (slope 0.030, intercept −1.76, P 0.0021), manometry (slope 0.032, intercept −1.51, P 0.00017), and defecography (all tests) with the highest prevalence from each study included (slope 0.00799, intercept −1.07, P 0.049), but it was nonsignificant with the lowest prevalence finding (slope: −0.0062, intercept −0.48, P 0.22). Nine studies were excluded from the meta-regression because no mean age was reported.[13, 22, 27, 33, 34, 59, 65, 72, 78]


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

In this meta-analysis of 79 studies evaluating 7581 patients with CC, the overall prevalence of any single abnormal dynamic pelvic floor test (i.e., anorectal manometry, defecography, MRI defecography) ranged from 14.9% to 52.9% with a median of 37.2%, and the prevalence for manometry was 47.7% (95% CI 39.5–56.1). The prevalence of an abnormal test tended to be lower in studies that evaluated patients by defecography (23.7–32.6%, using all defecography findings, ranging from the lowest to highest prevalences of specific findings in individual studies). The prevalence of an abnormal test was consistent across specialty and geographic area as well as when restricting to studies using Rome criteria.

The significance of this article lies in the synthesis of a body of data that has been historically difficult to synthesize. The challenges to the interpretation of the published data on the prevalence of DD and the diagnostic utility of available modalities are multiple and include the following: various and changing diagnostic criteria, lack of established testing protocols and definitions of positive results, and the fact that publications describe potentially heterogeneous populations, from surgical and GI practices. These circumstances have led the authors of a prior systematic review on diagnostic tests for DD to not proceed to a formal meta-analysis of data.[8]

As an example of the difficulty in comparing studies in this meta-analysis, three studies reporting the prevalence of DD may have used three separate criteria for positivity, such as ‘failure to open the ARA,’ ‘prominent impression of the PR muscle,’ or ‘paradoxical motion of the PR muscle.’ Furthermore, some studies also required a degree of impaired contrast evacuation in addition to one of the above findings to reach a diagnosis of DD. These apparently smaller details are among the many between-study inconsistencies in positive criteria that account for the variability in the prevalence of DD. This meta-analysis attempted to tease out criteria differences between studies.

In the prior systematic review on this topic,[8] heterogeneity of the studies and settings limited data pooling. While there have been a significant number of studies published since the prior review, there is still too much variability in methodology between the studies to have them pooled solely based on the presence or absence of DD. While this variability in the diagnostic strategies used, populations studied, and criteria used to interpret test results limits the ability to compare the results among studies, especially with regard to the prevalence of the DD, a main strength of this review is the focus on the prevalence of specific findings on these tests. By extracting data and grouping studies by specific findings, we provided pooled estimates of the prevalence of each finding, providing a prevalence map of individual abnormalities of pathophysiological relevance in CC.

We believe that these findings also add to our understanding of the epidemiology and evaluation of constipation, by determining the prevalence of findings associated with DD across time, geographical regions, disease definition, and specialty setting.

Caution should be exerted when attempted to extrapolate results. While we have shown the prevalences according to tests side-by-side (Fig. 2), one must remind the reader that this can only loosely be interpreted as a direct comparison between tests, as the prevalence was estimated in different populations for each test.


Figure 2. Pooled prevalence and 95% CI for specific findings according to test. ARA, anorectal angle; PR, puborectalis; EMG, electromyography; MRI, magnetic resonance imaging.

Download figure to PowerPoint

The relatively lower prevalence yielded by defecography which was apparent among pooled results (Fig. 2) was also true in the majority of studies (60%) evaluating the same patients with defecography and either manometry or BET. The reason behind the lower prevalence of DD by defecography could relate to the increased detection of structural abnormalities, though the presence of structural abnormalities and DD should not be mutually exclusive by definition.

The clinical utility of the increased sensitivity of defecography for structural abnormalities deserves further investigation, as structural abnormalities may be sequelae of longstanding constipation rather than causes of it, and more importantly, they do not appear to interfere with the success of biofeedback therapy. In one study examining the predictors of response to biofeedback, among 151 patients who underwent defecography, dynamic perineal descent, anterior or posterior rectocoele and intussusception were present in 100, 32, and 15 out of 151 patients, respectively, and none of these findings influenced the success of biofeedback treatment for constipation.[94] Another prospective study of 173 patients undergoing biofeedback therapy for chronic constipation found similarly that structural abnormalities (e.g., rectocele, sigmoidocele, intussusception) coexisting with dyssynergia did not impact the rate of response to therapy, which was 55% based patient report of whether or not symptoms were improved.[95]

There is conflicting evidence regarding whether the presence of dyssynergia predicts a response to biofeedback therapy. One study reported a large difference in outcome based on the presence of PFD, which was defined as dyssynergia on ARM and failure to expel a 50 cc balloon in 5 min, with 76% success in those with PFD and 8% in those without.[96] However, in a study of 100 patients, the 60% with dyssynergia on EMG were not more likely to report improvement in constipation following biofeedback therapy (overall response rate was 57%).[97] Similar results were reported in another study in which 65% of participants had dyssynergia by EMG, and this finding did not affect response to biofeedback therapy, which resulted in improvement in 59% of the 49 patients (63% among those with dyssynergia on EMG and 50% among those without).[98] These conflicting results may be indicative of the value of combining ARM with BET, but also highlight potential discrepancies between ARM and EMG in their reliability as predictive tests for treatment response. These contradictory findings could also be attributed to variable severity of pelvic floor abnormalities and to inconsistencies the criteria used for abnormal test results. For example, in a study of 148 patients, the response to biofeedback was better in patients with ‘partial’ relaxation of the anal canal on ARM than in those with no relaxation (78% vs 51%).[99] It is possible that those with partial relaxation may have been classified as not having dyssynergia in other studies.

This meta-analysis has several limitations. First, as we were interested in the prevalence of positive results for individual tests, studies not reporting data from individual tests were excluded. In some cases, this led to the exclusion of studies reporting the prevalence of DD based on a combination of tests. Because current guidelines recommend a combination of tests in the evaluation of DD, these combined prevalences remain of clinical relevance. Regrettably, of the included studies that did report prevalence based on more than one test (Table 4), there was not enough consistency across these studies to pool their data with meta-analysis methods. A second limitation of this study is that, in spite of our data extraction protocol designed to minimize heterogeneity at multiple levels, there was still likely variability in the performance of tests and interpretation of results. We provided the most exhaustive detail of pooled prevalences according to individual test findings and interpretations, yet we could not entirely abate the impact of such variability on our estimates.

A third limitation is intrinsic to most, if not all tests assessing defecatory mechanisms, which rely on artificial simulation in the lab of a physiological activity which is of the most personal and private in nature. In addition, tests performed in the left-lateral position may not faithfully reflect the dynamics of defecation act, which is naturally performed in the sitting position.

Despite the limitations, this study provides evidence that DD is prevalent across referral specialties and geographical regions. Our results, in context of the existing data on predictors of response to biofeedback, support the use of ARM with BET as an initial investigation for CC. Defecography, US, and EMG may provide more detailed information about the anatomy and physiology of the pelvic floor, but this information should not influence the decision to refer for biofeedback, a therapy that is effective even among unselected populations with constipation.

In conclusion, across testing modalities, geographical regions and specialties, a significant proportion of patients with CC show features of DD on diagnostic testing. This represents a sizable patient population for whom we lack evidence-based guidelines for evaluation and management. There is a need for prospective studies designed to evaluate the utility and cost-effectiveness of different approaches to evaluation and management of constipation as well as to determine predictors of response to biofeedback therapy.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

Elizabeth J. Videlock: nothing to disclose. Anthony Lembo: consultant for Ironwood/Forest. Filippo Cremonini: Ironwood/Forest: speaker, advisory board member.

Author Contribution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

EJV took care of study concept and design, acquisition of data, analysis and interpretation of data, drafting of the manuscript, did critical revision of the manuscript for important intellectual content, statistical analysis; AL contributed to study concept and design, did critical revision of the manuscript for important intellectual content; FC took care of the study concept and design, acquisition of data, analysis and performed interpretation of data, did critical revision of the manuscript for important intellectual content, statistical analysis.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contribution
  10. References
  11. Supporting Information
nmo12096-sup-0001-DDesupp.docWord document91KData S1. PubMed Search (((pelvic floor imaging) OR (balloon expulsion test) OR (anorectal manometry) OR (defecography) OR (magnetic resonance) OR (ultrasound) OR (electromyography)) AND ((constipation OR anismus OR obstructed defecation OR dyssynergia OR dyssynergic defecation OR defecatory disorders OR pelvic floor dysfunction OR pelvic floor dyssynergia) OR ((‘Rectal Diseases’[Mesh]) OR ((‘Pelvic Floor’[Mesh]) OR ‘Constipation’[Mesh])) OR (gastrointestinal motility))).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.