Reference

Study type

Characteristics

Intervention (I)

Controls (C)

Outcome measures and follow-up time

Results

Quality

assessment study

Altringer 1995

Design

Prospective cohort

N = 62

Country

USA

Aim of the study:

to determine the accuracy of physical examination (as judged by four-contrast

defecography) for women with pelvic floor relaxation disorders.

Inclusion criteria

women with vaginal prolapse and/or pelvic

pain evaluated

at the Center for Pelvic Floor Disorders, Rush-

Presbyterian-St. Luke's MedicalCenter, Chicago, Illinois.

Exclusion criteria:

Intervention

Fourcontrast

defecography

N = 62

Controls

No

Primary outcomes:

Comparison physical examination and defeco

Secundary outcomes:

Follow up:

No

Primary outcomes:

The physical examination diagnosis was changed in 46 patients (75 percent).

Twenty-six percent of presumed cystoceles, 36 percent of enteroceles, and 25 percent of rectoceles were not present on defecography. In contrast, when

physical examination was negative for these defects, 63 percent of patients were found to have cystoceles, 46 percent to have enteroceles, and 73 percent to

have rectoceles on four-contrast defecography.

For surgically significant posterior vaginal eversions (Grade 2 or 3), physical examination was accurate in only 61 percent; in the remainder, diagnosis was either

changed completely or additional unsuspected abnormalities were found

Clearly defined groups:

Yes. Prolapse defects were graded as follows: Grade 1, above the introitus; Grade 2, extending to the introitus; and Grade 3, below the introitus. A rectocele was defined as any distortion or bulging of the anterior rectal wall, with simultaneous deviation of the posterior vagina. An enterocele or sigmoidocele was present if corresponding portion of bowel descended at least 5 cm below a line between the symphysis pubis and coccyx

Selection bias:

Yes, no control group

Method of assessing the outcome

appropriately:

Yes

Blinding:

The radiologist was not informed of the physical examination diagnosis.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Cappabianca 2011

Design

Prospective cohort study

N = 1142

Country

Italy

Aim of the study:

to compare the

diagnostic efficacy of dynamic MR defecography (MR-D)

with entero-colpo-cysto-defecography (ECCD) in the assessment

of midline pelvic floor hernias (MPH) in female

pelvic floor disorders.

Inclusion criteria

- From August 2004 to August 2010, female patients who

required ECCD for the evaluation of pelvic floor disorders

were introduced in a prospective database, patients with ECCD findings of

MPH were asked to undergo MR-D.

Exclusion criteria:

Intervention

MR defecography (MR-D)

N = 1142

Controls

entero-colpo-cysto-defecography (ECCD)

N = 1142

Primary outcomes:

Sensitivity and specificity

Secundary outcomes:

Follow up:

NA

Primary outcomes:

Of 3,006 female patients who had undergone ECCD for the instrumental evaluation of pelvic floor disorder, 1,160 (38.5%) patients showed an MPH at ECCD. Out of these, 1,142 (98.4%) (mean age, 52.47±13.71) accepted to undergo MR-D and constituted the object of analysis. Among the enrolled patients, 722 (63.2%) complained of obstructed defecation syndrome, whereas 371 (32.4%) complained of anal and/or urinary incontinence.

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

From August 2004 to August 2010, female patients who required ECCD for the evaluation of pelvic floor disorders were introduced in a prospective database, investigated

about their clinical history, and considered for the enrolment in this study. All the patients with ECCD findings of MPH were asked to undergo MR-D. The diagnosis and measurement were then

independently performed by two of the authors

with 20 years of experience; in case of disagreement, the final diagnosis was made in consensus.

According to Yang's classification [12], the limits of normal descent with maximal strain are 1.0 cm below the PCL for the bladder base, 1.0 cm above for the vaginal cuff or lower end of the cervix, and 2.5 cm below for the rectal area. The rectocele was defined as an out-pouching of the anterior rectal wall occurring during evacuation or straining. Enterocele, sigmoidocele, and omentocele represent the herniation below the proximal (apical) one third of the vagina of the peritoneal

sac containing ileal loops, part of the sigmoid, or peritoneal fat, respectively. If the small bowel, the

peritoneal fat, or the sigmoid colon entered the Retzius' or Douglas' space, they were identified as Retzius' and Douglas' hernias, respectively; if they entered the vaginal fornix posteriorly, causing a complete eversion of the vaginal wall, an elytrocele was recognized (posterior vaginal hernia) In the same way, if they entered the rectum anteriorly, leading to a rectal wall eversion, an edrocele was detected. On evaluation of ECCD, the diagnosis of an enterocele/ sigmoidocele/omentocele was made if the picture obtained during evacuation compared with that during rest showed

an increase in the distance between the vagina and rectum. This expansion should extend below the PCL reference line and show a sagittal diameter of more than 2 cm. On MR-D, the relationship between the lowest point of the peritoneal border line and the PCL was assessed. A descent of parts of the peritoneal content below this line

and the identification of herniated contents allowed the distinction in enterocele, sigmoidocele, and omentocele

Blinding:

All personal information was removed from the radiological images. The diagnosis and  measurement were then independently performed by two of the authors

Selective loss to follow up:

Of 1,160 patients  who showed an MPH at ECCD, 1,142 (98.4%) accepted to undergo MR-D and constituted the object of analysis.

Identification confounders and correction in analysis:

No

Funding:

Not reported

Dvorkin 2004

Design

Prospective cohort study

N = 10 (6 women)

Country

UK

Aim of the study:

To determine

whether open-magnet magnetic resonance (MR) defaecography

could provide more useful clinical information

than evacuation proctography (EP) alone in the evaluation

of a cohort of patients with full-thickness rectal

intussusception and could assist in decisions concerning

management.

Inclusion criteria

- patients with symptomatic circumferential rectal intussusception

diagnosed both clinically (using rigid sigmoidoscopy)

and on EP

Exclusion criteria:

Intervention

MR-defaecography

N = 10

Controls

evacuation proctography (EP)

N = 10

Primary outcomes:

Secundary outcomes:

Follow up:

NA

Primary outcomes:

Ten patients (4 male; median age 43 years, range 30–65 years) with symptomatic circumferential rectal intussusceptions diagnosed both clinically (using rigid sigmoidoscopy) and on EP, underwent open-magnet MR defaecography.

Median time between EP and MR defaecography was 9.5 months (range 3–32 months). No patient underwent treatment for rectal intussusceptions during this time.

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Evacuation proctography was performed by standard technique with no bowel preparation prior to the investigation. MR defaecography was performed in a 0.5T open configuration MR system (Signa SPIO, General Electric Company, MilwaukeeIL, USA). No bowel preparation was used and patients were instructed to empty their bladders and rectum prior to the procedure.

A rectocoele was defined as an anterior rectal wall protrusion beyond a line drawn along the main anterior rectal wall axis, and if present, its lateral and vertical dimensions were recorded.

Descent was defined as the difference in anorectal

junction (ARJ) position between the resting and straining positions. Based on values obtained

in a study of normal asymptomatic volunteers, significant bladder and vaginal descent on MR defaecography were defined as greater than 3 cm, while significant ARJ descent was defined as greater than 2 cm

Blinding:

Videotaped images from EP and cine loops from the MR defaecography were analysed by two of the authors, one of whom was blinded to the results of the patient’s history and physical examinations.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Eltik 2005

Design

Prospective cohort study

N = 76

Country

Turkey

Aim of the study:

To investigate the efficacy of the magnetic resonance fluoroscopy in the diagnosis and staging of the

pelvic prolapse (PP).

Inclusion criteria

- patients with urinary incontinence, dysuria, pollakiuria, dyspareunia,

defecation problems, and a feeling of a mass in the vagina

and who had positive physical examination signs for PP

Exclusion criteria:

Intervention

Women with signs and symptoms suggesting

PP who underwent a series of fast MR-fluoroscopy images.

N = 46

Controls

women who underwent vaginal exam and shown not have pelvic prolapsed who underwent a series of fast MR-fluoroscopy images.

N = 30

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

Patients were between 21–55 years old (mean: 35±9). All patients were parous (range: one–three; median parity: 2). The mean age of the control group was

31 ± 8 years (21–35).

Of the 46 women, 36 had a cystocele, 25 had a rectocele, and 28 had an uterine prolapses. We found seven of 30 control subjects to have PP by MRfluoroscopy

technique. Distribution of these seven subjects was: four cystocele, one uterine prolapse, one cystocele and uterine prolapses, and one rectocele.

There was significant statistical correlation between physical examination signs and MR-fluoroscopy results in both patients and control groups for all of three compartments (P < 0.01, <0.01 and <0.05, respectively).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Gynecologic examinations were done by an experienced gynecologist. PP was separately evaluated as anterior (cystocele), medium (uterocele), and posterior (rectocele). The

degree of PP was determined according to Baden’s classification. 0.3 T open MR (Hitachi, AIRIS, Tokyo, Japan) was used. Measurements were separately done for three compartments on fluoroscopic images and T2-W FSE images, which were taken at rest. According to these measurements we used modified Sigh et al.’s

MR classification for PP and used this new classification (table 2)

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Kamal 2013

Design

Prospective case-control study

N = 30

Country

Egypt

Aim of the study:

To determine the role of magnetic resonance imaging (MRI) in surgical planning

for females with pelvic organ prolapsed (POP) and to determine the clinical utility of MR

imaging in predicting successful surgical repair.

Inclusion criteria

- women aged 35–68 (mean 55 years) with symptoms and/or physical findings indicating pelvic organ prolapse who were seen in Royal Commission Hospital, EL Gubail Industrial City between Feb 2009 and October 2011.

-  controls: asymptomatic nulliparous female volunteers

matched for age, body mass index, and menopausal status

Exclusion criteria:

Intervention

patients with different varieties of pelvic floor dysfunction

N = 15

Controls

nulliparous

females

N = 15

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

                After surgery all values were significantly modified (P < 0.01). A significant difference was seen between pre and post surgery measurements with respect to the distance descended (P < 0.01). All the patients except two were continent in the postoperative period.

                Preoperative MRI examination revealed 40 pelvic defects compared to 30 defects on preoperative physical exam. On preoperative MRI, vaginal vault prolapse (n = 5) or uterovaginal prolapse (n= 7), 12 cystocele and 2 cystourethrocele, 7 enterocele,

and 3 rectocele were observed. Combined defects in all the three compartments were observed in ten patients.

                Pearson’s correlation of paired samples compared the operative findings with the preoperative pelvic examination and with the preoperative MRI findings. A significant correlation was found between the preoperative pelvic examination and the operative findings for cystocele (r =0.61, P < 0.005), vaginal cuff prolapse (r= 0.75, P <0.008).and rectocele (r = 72, p0.008). The preoperative pelvic examination did not correlate with the operative findings of uterine prolapse and enterocele. On the other hand, the comparison of the MRI and operative findings showed a significant correlation in all types of prolapse, except rectocele, as follows: cystourethrocele, R =0.68 and P < 0.001; enterocele, r = 0.78 and P < .001; vaginal cuff prolapse, r= 0.75 and P < 0.008; and, uterine prolapse, r= 0.78 and P < 0.014.

                Dynamic imaging lead to changes in the initial operative plan in five (40%) of patients. Such alterations included three patients with chronic constipation and chronic lower abdominal pain in whom dynamic MR demonstrated enterocele one after vault prolapse and two after vaginal hysterectomy. All were managed by culdoplasty with posterior colporraphy of all facial defects adjacent to sigmoid and rectum for adequate repair. The operative plan was also altered in a 67-year-old female with urinary and fecal incontinence and a large cystocele. On dynamic MR, she was found to have compression of the rectum secondary to massive descent of the bladder, uterus and cervix. Her fecal incontinence was secondary to compression

of the rectum by the prolapsing anterior viscera, which then resulted in fecal evacuation. This patient underwent total vaginal hysterectomy, culdoplasty, and anterior and posterior colporrhaphy and remains continent of both stool and urine.

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

An experienced gynecologist evaluated all patients by means of detailed medical and surgical history, documentation of complaints, and physical examination. Compartment defects and specific organ prolapses were recorded but not graded.

MR images were acquired using the (General Electric, Milwaukee, WI) GE Horizon 1.5-T.

Preoperative and (6– 12 months), postoperative examination with MRI for assessing the efficacy of surgical repair was performed for (group 1),

Volunteers in group 2 were examined only once. A bulge of more than 3 cm between the extended

line of the anterior border of the anal canal and the tip of the rectal descensus constituted a rectocele (3,4). A wide rectovaginal space or deep pouch of Douglas with the peritoneal contents

with or without small bowel loops beyond the

pubosacral line was considered enterocele.

Blinding:

The interpretation of the images was performed by an experienced radiologist unaware of the result of clinical evaluation

Selective loss to follow up:

No

Identification confounders and correction in analysis:

No

Funding:

Not reported

Kelvin 1992

Design

Prospective cohort study

N = 74

Country

USA

Aim of the study:

to evaluate the contribution

of evacuation proctography (EP) to the investigation of

patients with pelvic prolapse and, in

particular, the role of EP in the assessment

of rectoceles and enteroceles.

Inclusion criteria

- Between April 1990 and June 1991, 74

consecutive women with pelvic prolapse

were referred for EP by one urogynecologist.

Exclusion criteria:

Intervention

EP in women with pelvic prolapse

N = 74

Controls

No

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

                The mean age of the patients was 58 years (range, 29-78 years). Sixtyone (82%) were multiparous, and 47(64%) had previously undergone hysterectomy. In 28 (38%), one or more prior surgical procedures for pelvic prolapse (other than hysterectomy) had been performed, including surgery for a cystocele and/or stress urinary incontinence in 22 and for rectocele repair in 11. All except four of the patients had rectal symptoms: 61 (82%) complained of difficulty

with rectal evacuation manifested by constipation, straining to defecate, or a

sensation of incomplete rectal emptying. Eighteen patients used digital pressure in

their vagina or rectum or on their perineum to aid rectal emptying. Nineteen

patients gave a history of anal incontinence.

                A rectocele was found at physical examination or EP in 73 (99%) of the

74 patients: Of these 73 rectoceles, six (8%) were not detected at physical examination and three (4%) were not seen at EP. The six rectoceles missed at physical examination were all small, and none of these were repaired surgically.

                An enterocele was found at physical examination or EP in 14 (19%) of the patients, all of whom had a coexistent rectocele. Seven (50%) of these 14 enteroceles were not detected at physical examination, and three (21%) were not seen at EP. Eleven of the 14 patients with an enterocele subsequently underwent surgery, including six of the seven patients in whom the enterocele was missed at physical examination.

                Four sigmoidoceles were found at EP, none of which were recognized

clinically. Two of the four sigmoidoceles were associated with an enterocele, and in all four patients there was a coexistent rectocele. Two of these four patients subsequently underwent surgery.

Clearly defined groups:

Yes

Selection bias:

No

Method of assessing the outcome

appropriately:

Physical examination of the pelvis was performed by the urogynecologist or his Fellow. The technique of EP was a modification of that described by Shorvon et al.

At EP, an enterocele was suspected if the upper vagina showed separation by 2 cm or more from the adjacent rectum on straining downward at the end of rectal evacuation. Rectoceles were classifled

as small ( < 2 cm in depth), moderate sized (2-4 cm in depth), or large ( > 4 cm in depth). A sigmoidocele was considered to be present if sigmoid colon was found to descend below this line into the rectovaginal space.

Blinding:

Not reported

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Lopez 2002

Design

Cohort study

N = 25

Country

Sweden

Aim of the study:

to prospectively

evaluate pre- and postoperative findings of cystodefecoperitoneography

(CDP) and to correlate the findings to

the clinical examination in patients with genital prolapse.

Inclusion criteria

- patients with a clinically diagnosed

symptomatic genital prolapse

Exclusion criteria:

- Patients with concomitant urinary incontinence

Intervention

cystodefecoperitoneography

N = 25

Controls

No

Primary outcomes:

Secundary outcomes:

Follow up:

9 months post-operatively

Primary outcomes:

                Mean age was 59 (range 29–76) years. The surgical procedures were planned according to the preoperative clinical findings. In 2 patients an enterocele was diagnosed during surgery and then repaired.

                Rectocele: preoperatively, a rectocele was diagnosed in 24 patients at clinical examination and in 21 patients at CDP. In 3 patients a rectocele was diagnosed only on clinical examination. In these 3 patients CDP demonstrated a peritoneocele with bowel, i.e. an enterocele, and a cystocele (patient 2), a cystocele and a peritoneocele (patient 9), a cystocele and a rectocele at the start which disappeared at maximum straining when a peritoneocele pushed the rectum posteriorly (patient 10) (Table 2). Postoperatively, a rectocele was not diagnosed in any patient at clinical examination and in 4 patients at radiologic investigation.

                Enterocele: preoperatively, an enterocele was diagnosed in 2 patients

at clinical examination. At CDP a peritoneocele with bowel, i.e. an enterocele, was diagnosed in 6 patients. A peritoneocele without bowel was diagnosed in another 3

patients. Postoperatively, an enterocele was diagnosed in 1 patient at clinical examination. At CDP a peritoneocele with bowel, i.e. an enterocele, was diagnosed in 3 patients. A peritoneocele without bowel was diagnosed in another 4 patients.

Clearly defined groups:

Yes

Selection bias:

Not expected, 25 of 29 patients with a clinically diagnosed symptomatic genital prolapse agreed to participate in the study.

Method of assessing the outcome

appropriately:

patients were investigated both preand

postoperatively with a standardized questionnaire,

clinical examination and cystodefecoperitoneography (CDP), including contrast medium in the rectum, vagina, bladder, small bowel and peritoneal cavity.

Defecoperitoneography was performed according to the techniques described by Bremmer et al.

A rectocele was defined as an outpocketing of

the anterior rectal wall and the posterior vaginal wall into the lumen of the vagina. The size of the rectocele was measured as the distance between the extended line of the anterior border of the anal canal and the tip of the rectocele. A peritoneocele in the pelvis was defined as an extension of the pouch of Douglas below the upper third of the vagina. An enterocele was defined as bowel present in the peritoneocele.

Selective loss to follow up:

No

Identification confounders and correction in analysis:

No

Funding:

This study was supported by grants from Kjell

and Ma¨rta Beijers Stiftelse, Marianne and Marcus Wallenbergs Stiftelse, Bert von Kantzows Research Foundation, Karolinska Institutet Research Funds and Fo¨renade Liv Mutual Group Live Insurance Company, Stockholm, Sweden.

Murad 2012

Design

Prospective cohort study

N = 29

Country

Brazil

Aim of the study:

to describe a novel

three-dimensional dynamic anorectal ultrasonography

technique (dynamic 3-DAUS) for assessment of perineal

descent (PD) and establishment of normal range values,

comparing it with defaecography. Secondarily, the study

compares the ability of the two techniques to identify

various pelvic floor dysfunctions.

Inclusion criteria

- Patients with obstructed defaecation disorder

Exclusion criteria:

Intervention

dynamic 3-DAUS

N = 29

Controls

defaecography

N = 29

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

                Between March 2008 and February 2009, 29 female patients (mean age 43 years, range 23–74) with a mean validated Wexner constipation score of 10 (range 8–18) were prospectively enrolled in the study. Fifteen patients had undergone at least one vaginal delivery.

                Both techniques identified five patients without rectocele, two with Grade I rectocele, 10 with Grade II rectocele and nine with Grade III rectocele. Diagnosis was discordant in three cases with Grade II rectocele on defaecography:

on dynamic 3-DAUS, two of these had Grade III rectocele and one had none. The index of agreement between defaecography and dynamic 3-DAUS was high with regard to patients without rectocele and Grade I rectocele (kappa 0.89 and 1.00; 95% CI 0.53–1.00 and 0.64–1.00, respectively) (P < 0.001), and agreement was substantial for Grade II and Grade III rectocele (kappa 0.72 and 0.77; 95% CI 0.37–1.00 and 0.41–1.00, respectively) (P < 0.001) (Table 3).

                Rectal intussusception was identified in six patients on defaecography. These were confirmed on dynamic 3-DAUS in addition to the identification of another seven cases (40%), indicating moderate agreement between the methods (kappa 0.46; 95% CI 0.12–0.79) (P < 0.007) (Table 4).

                Grade III enterocele ⁄ sigmoidocele was identified in one patient on defaecography. This was confirmed on dynamic 3-DAUS in addition to the identification of another case, indicating substantial agreement between

the methods (kappa 0.65; 95% CI 0.31–0.99) (P < 0.001).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

The subjects were given a complete proctological examination and underwent defaecography

and dynamic 3-DAUS performed by different

examiners for the evaluation of pelvic floor dysfunctions.

Anterior rectocele was defined as a herniation of the anterior rectal wall protruding into the posterior vaginal wall during straining. (Grade I ≤ 2.0 cm; Grade II 2.0–4.0 cm; Grade III > 4.0 cm). Sigmoidocele ⁄ enterocele was defined as a herniation of the peritoneum (containing the sigmoid colon or small bowel) into the pelvis. It was considered significant when the loop of sigmoid ⁄ bowel extended below the ischiococcygeal line (Grade III).

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Pilkington 2012

Design

Prospective cohort study

N = 42

Country

UK

Aim of the study:

To investigates whether there are measurable differences

between Barium proctography (BaP) and Magnetic resonance (MR) proctography. Patient preference

for the tests was also investigated.

Inclusion criteria

- Referred for proctography as part of

routine NHS management

- Patient gives informed written consent

- Patient is > 18 years old

Exclusion criteria:

- Patient incompetent to give informed consent

- Claustrophobia or unable to tolerate MRI

- Contraindications to MRI such as pacemaker,

high body mass index

- Patient unable to lie flat

Intervention

Magnetic resonance (MR) proctography

N = 42

Controls

Barium proctography (BaP)

N = 42

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

The mean age of participants was 59 years with a range of 37–76 years.

                Rectoceles were extremely common and present in almost all participants on both BaP and MR proctography. The measure of agreement on presence of rectocele was substantial (kappa = 0.690). The mean rectocele size was 3.1 cm (SD 1.4) on BaP and 2.9 cm (SD 1.6) on MR proctography. The mean difference in rectocele size measured on BaP and MR proctography was 0.20 cm (95% CI –0.23, 0.63; P = 0.35).

                The measurement of agreement between BaP and MR proctography for rectal intussusception was fair (kappa = 0.209). However, MR proctography missed 31% (11 ⁄ 35) of cases detected on BaP compared with 8% (2 ⁄ 26) missed by BaP.

                The patient questionnaire was administered to 42 participants

and 25 participants completed it for both BaP and MR proctography. All participants who completed the questionnaire for both tests agreed to have either test repeated if necessary except for one participant who did not want the MR

proctogram repeated. The tests were reported as equal in two (6%) participants. A preference was reported in 32 (94%) participants. MR proctography was the preferred test for 21 (62%) participants and BaP was the preferred test for 11 (32%) participants.

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

During BaP, the rectum was filled with contrast (barium paste).

The technique for MR proctography was similar to

BaP in that the participant had contrast (ultrasound gel) placed in the rectum. The MRI scanner had a 1 T magnet (Phillips Intera).

A record was made of the presence or absence of rectocele, complete rectocele emptying, rectal intussusception, complete rectal emptying,

anismus, mucosal prolapse, enterocele, uterovaginal prolapse and cystocele. Rectal intussusception was classified according to the Oxford Radiological Grading of Rectal Intussusception. Rectocele size was measured

as the maximum length from an extended anterior wall of the anal canal. The distance from the ARJ to the PCL was measured during rest, squeeze and evacuation. The ARA was measured at the intersection between a line along the posterior wall of the rectum and a line along the central axis of the anal canal.

Blinding

At the time of reporting, each radiologist was blinded to the results of the other proctogram.

Selective loss to follow up:

Between 8 May 2008 and 11 December 2009 this study invited 216 patients (202 women and 14 men) to participate. At the appointment for BaP, 71 participants were recruited and 42 of these completed the study by attending for MR proctography. The remaining 29 patients withdrew from the study.

Identification confounders and correction in analysis:

No

Funding:

This project was funded by a Research Bursary from the Bowel Disease Research Foundation.

Regadas 2011

Design

Prospective cohort study

N = 86

Country

Brazil, USA, Venezuela

Aim of the study:

to validate the

effectiveness of echodefecography compared with

defecography in the assessment of anorectal dysfunctions

related to obstructed defecation.

Inclusion criteria

- From January 2009 through October 2009, women presenting

with obstructed defecation symptoms at 6 centers for colorectal surgery

Exclusion criteria:

Patients with previous anorectal and vaginal surgery, fecal incontinence, and/or previous anorectal radiation

Intervention

Echo-defecography

N = 86

Controls

Defecography

N = 86

Primary outcomes:

kappa statistic

in the evaluation of rectocele,

intussusception, anismus, and grade III enterocele

Secundary outcomes:

Follow up:

Primary outcomes:

                The median validated Wexner constipation score was 13.4 (range, 6–23). The median age was 53.4 (range, 26–77) years. Among the patients, 16 (18.6%) were nulliparous, 40 (46.5%) had had vaginal deliveries, and 30 (34.9%) had undergone Cesarean section.

                Rectocele was identified in 80 patients by defecography and in 76 patients by echodefecography. Agreement for individual rectocele grades is shown in Table 1. Overall, agreement between the 2 methods was classified as substantial

(kappa = 0.61; 95% CI = 0.48–0.73).

                Defecography identified rectal intussusception in 42 patients, with echodefecography identifying 37 of these cases, plus 4 additional cases of intussusceptions. Absence of intussusception was noted by both methods in

40 patients, with agreement classified as substantial (kappa = 0.79; 95% CI = 0.57–1.0).

                Defecography identified 3 cases of grade I, 8 cases of grade II, and 9 cases of grade III enterocele. Echodefecography found 8 of the 9 cases of grade III enterocele identified by defecography, and found an additional subject with grade III enterocele, for a total of 9 grade III enteroceles (Table 6). Agreement for grade III enterocele was classified as almost perfect (kappa = 0.87; 95% CI = 0.66 –1.0).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Patients were evaluated with a clinical examination consisting of a full proctologic evaluation, followed by defecography and echodefecography

performed by different examiners across the 6

centers.

Rectocele, defined as outpouching of the anterior upper anal canal or rectal wall occurring during straining, was measured perpendicularly

to the expected contour of the anterior rectal

wall and classified as grade I (<2.0 cm), grade II (2.0 – 4.0 cm), or grade III (>4.0 cm). Intussusception was defined as invagination of the rectal wall occurring during straining but not passing through the anal canal. Enterocele and sigmoidocele were diagnosed as herniations

of the peritoneum (containing the small bowel or

sigmoid colon) into the pelvis.

Echodefecography was performed with a

Pro Focus 3-dimensional ultrasound scanner (B-K

Medical, Herlev, Denmark) using a 2050 endoprobe with 55-second proximal-to-distal 6.0-cm automatic scanning, a frequency range of 10 MHz to 16 MHz, and a focal distance of 2.8 cm to 6.4 cm.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Steensma 2010

Design

Prospective cohort study

N = 75

Country

Netherlands

Aim of the study:

To compare evacuation proctography (EP) with

3D transperineal

Ultrasound (3DTPUS) in diagnosing posterior compartment prolapse.

Inclusion criteria

- All women with symptoms related to posterior compartment

prolapse referred to our tertiary pelvic floor unit

Exclusion criteria:

Intervention

3D transperineal

Ultrasound

N = 75

Controls

evacuation proctography (EP)

N = 75

Primary outcomes:

Cohen’s kappa index

Secundary outcomes:

Follow up:

Primary outcomes:

                The median age was 59 years (range: 22–83 years). A previous hysterectomy had been carried out in 31 (41%) women and a previous pelvic organ prolapse repair in 37 (49%). Median parity was 2 (range: 0–10 vaginal deliveries; four nulliparous women).

                Clinical examination showed absence of prolapse in 16 (21%) patients, stage 1 prolapse in 27 (36%), stage 2 in 21 (28%) and stage 3 in 11 (15%). Evacuation proctography revealed in 7 (44%) out of the 16 patients without

prolapse during clinical examination some form of posterior compartment prolapse. 3D ultrasound revealed some form of prolapse in six (38%) of these patients. The

clinical diagnosis of posterior compartment prolapsed (stage 1–3) was confirmed by EP and 3D ultrasound in 83% and 81% respectively.

                Of the nine grade 1 enteroceles detected by EP, seven (78%) were not

detected by ultrasound. However, only two out of 16 grade 2 and 3 enteroceles detected with EP were not detected with transperineal ultrasound, resulting in a

Cohen’s kappa value of 0.77 for detection of grade 2 and 3 enteroceles. Furthermore, 3D ultrasound revealed an enterocele in two other patients, which was not diagnosed by EP. Overall, we found a sensitivity of 64% and a specificity of 96% for the detection of all enteroceles with 3D ultrasound.

                A Cohen’s kappa coefficient of 0.55 was found for the detection of rectocele. The mean rectocele depth was 2.82 cm at EP and 2.32 cm with ultrasound

(r = 0.47, P < 0,05). Eight rectoceles detected by EP were not detected by ultrasound; however, nine rectoceles detected by ultrasound were not found by EP. Of all grade 2 and 3 rectoceles, 87% was confirmed by 3D ultrasound in contrast to only 25% of all grade 1 rectoceles. The sensitivity and specificity of diagnosing

rectocele with 3D ultrasound in general was respectively 78% and 77%.

                In the detection of intussusception a Cohen’s Kappa Index of 0.21 was found with no differences for grade 1 and grade 2 intussusceptions. Overall, the detection rate of intussusceptions with 3D ultrasound was low, 21 (78%) of the 27 intussusceptions detected with EP were not confirmed by 3D ultrasound. The sensitivity and specificity of diagnosing intussusception with transperineal

ultrasound were 22% and 96% respectively.

                Sixty-four (85%) patients responded to the questionnaire. EP caused significantly more discomfort for the patients (P < 0.001). VAS scores concerning discomfort for EP were rated at median 4,0 ⁄ 10 (range: 0–10), whereas VAS scores for 3D ultrasound were rated at median 1.0 ⁄ 10 (range: 0–9). EP was a less tolerated

investigation compared with 3DTPUS in 87% of the patients, while in 8% the 3DTPUS was less tolerated (P < 0.001). Significantly, more patients preferred to

have a repeat ultrasound than an evacuation proctogram (P < 0.003).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Patients were interviewed using a standardized questionnaire, concerning medical history, urinary function, pelvic discomfort and bowel function. A clinical examination was performed according to the International Continence Society guidelines, using the POP-Q system. The patients underwent standardized EP and dynamic 3DTPUS, with a maximum interval of 6 months.

3D transperineal ultrasound was performed using a GE Kretz Voluson 730 expert system (GE Healthcare, Clinical Systems, Hoevelaken, the Netherlands), using an abdominal 4–8 MHz transducer.

Blinding

All video files were analysed by one colorectal surgeon (WRS), blinded against all clinical data and the results of 3DTPUS.

Selective loss to follow up:

NA

Identification confounders and correction in analysis:

No

Funding:

Not reported

Weemhoff 2013

Design

Prospective cohort study

N = 50

Country

Netherlands

Aim of the study:

to determine the

level of agreement between transperineal ultrasound and evacuation proctography. Secondarily, the diagnostic test characteristics of transperineal ultrasound for diagnosing enteroceles, rectoceles, and intussusceptions

were estimated.

Inclusion criteria

- Women with

complaints of fecal incontinence or obstructed defecation

visiting the tertiary care colorectal pelvic floor unit and who were scheduled to undergo a diagnostic endoanal

ultrasonography and evacuation proctography

Exclusion criteria:

age <18 years, legally

incapable patients, and persons who were not able to understand the information given

Intervention

transperineal ultrasonography,

N = 50

Controls

evacuation proctography

and

endoanal ultrasonography

N = 50

Primary outcomes:

Sensitivity, specificity,

positive (PPV) and negative predictive value, as well as

the positive and negative likelihood ratio

Secundary outcomes:

Follow up:

Primary outcomes:

The mean age was 59 years (range, 28–95). Eighty-two percent (82 %) of women had fecal incontinence, and 16 % had complaints of obstructed defecation.

The interobserver agreement of the two ultrasonographers assessed as the quadratic weighted kappa of diagnosing enteroceles was 0.44 (95 % CI 0.08, 0.80); that of diagnosing rectoceles was 0.51 (95 % CI 0.27, 0.75), and that of diagnosing

intussusceptions was 0.23 (95 % CI −0.22, 0.69).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Transperineal ultrasound was performed using a GE Kretz Voluson 730 expert system. For transperineal ultrasonography a 4- to 8-mHz transabdominal curved 2D transducer was used.

Enterocele was defined as a herniation of the peritoneal cavity with abdominal contents between the rectum and vagina. Grade 1 enterocele extended into the upper half of the vagina. Grade 2 enterocele reached the perineum, and grade 3 enterocele protruded out of the vagina. Rectocele was defined as a bulging of the anterior rectal

wall into the vagina. The maximum depth of the bulging rectocele was measured perpendicular to the expected contour of the anterior rectal wall. Grade 1 was a rectocele with a depth below 2 cm. Grade 2 rectocele had a depth between 2 and 4 cm. In a grade 3 rectocele, the depth of the bulge

exceeded 4 cm.

Intussusception was defined as invagination of the proximal rectal wall during defecation. The presence of an intussusceptions was dichotomized (present, absent).

Rectoceles with a depth of 2 cm or more were classified as clinically relevant

Blinding:

The observers were blinded to other results (symptoms, physical examination, and other imaging studies). Disagreement between assessors was solved at a consensus meeting.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

None