Endoscopy 2018; 50(07): 671-683
DOI: 10.1055/s-0043-125207
Original article
© Georg Thieme Verlag KG Stuttgart · New York

Meta-analysis of randomized controlled trials challenging the usefulness of purgative preparation before small-bowel video capsule endoscopy

Paraskevas Gkolfakis
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital, Athens, Greece
,
Georgios Tziatzios
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital, Athens, Greece
,
George D. Dimitriadis
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital, Athens, Greece
,
Konstantinos Triantafyllou
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital, Athens, Greece
› Author Affiliations
Further Information

Corresponding author

Konstantinos Triantafyllou, MD, PhD
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital
Rimini 1
12462, Athens
Greece   
Fax: +30-21-05326454   

Publication History

submitted 19 September 2017

accepted after revision 03 December 2017

Publication Date:
06 February 2018 (eFirst)

 

Abstract

Background The usefulness of purgative preparation before small-bowel video capsule endoscopy is controversial. We aimed to examine the effect of purgative preparation on small-bowel video capsule endoscopy outcomes.

Methods We performed literature searches in MEDLINE and the Cochrane library for randomized controlled trials evaluating the effect of purgative preparation (polyethylene glycol, sodium phosphate, others) vs. clear-liquid diet/fasting in patients undergoing small-bowel capsule endoscopy. Meta-analysis outcomes included the examination’s diagnostic yield, small-bowel mucosal visualization quality, the examination’s completion rate, and gastric and small-bowel transit times. The effect size on study outcomes was calculated using a fixed- or random-effect model, as appropriate, and is shown as the risk ratio (RR) with 95 % confidence interval (CI).

Results We identified 12 eligible trials with 17 sets of data including 1221 subjects. Significant heterogeneity was detected with no evidence of publication bias. As compared with clear-liquid diet, purgative bowel preparation did not increase capsule endoscopy diagnostic yield (RR 1.17 [95 %CI 0.97 to 1.40]; P = 0.11). Neither the small-bowel mucosal visualization quality (RR 1.14 [95 %CI 0.96 to 1.35]; P = 0.15) nor completion rate for the examination (RR 0.99 [95 %CI 0.95 to 1.04]; P = 0.76) significantly improved after purgative preparation. Purgatives also had no effect on video capsule endoscopy gastric and small-bowel transit times.

Conclusions Our analysis challenges the usefulness of purgative preparation for improving the diagnostic yield of small-bowel video capsule endoscopy and the quality of small-bowel mucosal visualization.


#

Introduction

Meticulous mucosal visualization is a prerequisite for the unequivocal assessment of patients undergoing small-bowel video capsule endoscopy (VCE). However, intraluminal presence of food debris, air bubbles, bile or mucus often significantly decreases the diagnostic accuracy of the examination. Although the device’s manufacturer proposes a preparation regimen of 12 hours of clear liquids and/or fasting overnight, several studies evaluating different agents, including purgative bowel preparation with polyethylene glycol (PEG) and sodium phosphate (Na-P), and antifoaming agents such as simethicone and prokinetics, have been conducted.

To date, six meta-analyses have tried to pool the results of the aforementioned studies with outcomes that have been difficult to interpret because of the inclusion of different study designs, variable preparation protocols, and inconsistent endpoints [1] [2] [3] [4] [5] [6]. Because purgative bowel preparation is the most widely used preparation in clinical practice today and in order to overcome the inconveniences that the previous meta-analyses introduced, we decided to exclusively meta-analyze randomized controlled studies evaluating the effects of purgative bowel preparation – given before capsule ingestion – on the main outcomes of VCE.


#

Methods

Data identification and extraction

A thorough computer-assisted search of the medical literature was conducted across MEDLINE and the Cochrane Central Register of Clinical Trials databases from January 2001 to July 2017 combining the terms “capsule endoscopy” and “bowel preparation.” Both terms were searched as medical subject headings (MeSH) and free-text terms. The complete search strategy can be found in [Appendix e1] (available online). We further hand searched all references from review articles, editorials, and all original studies to identify additional citations fulfilling the inclusion criteria. Data were extracted in a predefined worksheet by two of the authors (P.G. and G.T.) working in parallel but independently.

Data extracted included: name of first author, year of publication, country of origin, study design, number of participants, population characteristics, regimen of bowel preparation provided and its comparator, device used, diagnostic yield of the examination, small-bowel mucosal visualization quality (SBVQ), gastric transit time (GTT), small-bowel transit time (SBTT), and number of completed (capsule entered the cecum) examinations. Data were extracted as originally presented or following appropriate calculations, if applicable. For studies with missing or unavailable data, an attempt to contact the corresponding author to provide additional information was made. Any disagreements between the two authors were settled by discussion with the senior author (K.T.).

The search was initially performed on 12 February 2017, and repeated twice on 2 May, and 31 July 2017. Our study complies with PRISMA guidelines [7], see [Appendix  e2] (available online).


#

Study endpoints

The primary endpoint of our meta-analysis was to explore the effect of small-bowel preparation with cathartics compared with clear-liquid diet or fasting overnight on the diagnostic yield of VCE.

Secondary endpoints included the effects of purgative bowel preparation on SBVQ, VCE completion rate, GTT, and SBTT.


#

Definitions

Diagnostic yield

This is defined as the likelihood that a test or procedure will provide the information needed to establish a diagnosis. In the field of VCE, the definition of a positive study incorporates any examination with findings that explain a patient’s complaints or symptoms, as well as any examination with findings that lead to modification of a patient’s management.


#

Small-bowel mucosal visualization quality

Investigators have used various definitions to determine SBVQ. Because no official consensus exists, we arbitrarily bisected the proposed “qualities of visualization” into two broad groups: adequate and inadequate. The “adequate” group consisted of a quality of mucosal visualization characterized as adequate, excellent, or good by the authors; visualization characterized as inadequate, moderate, fair, or bad comprised the “inadequate” group. A study was considered eligible to be included in SBVQ analysis only if it reported data for the visualization of the entire small bowel. For studies that traced SBVQ using a score that could not be dichotomized and consequently homogenized in the adequate vs. inadequate preparation, it was pre-decided that these would be excluded from the meta-analysis of the respective endpoint.


#

Completion rate

An examination was considered to have been completed when the capsule reached the cecum.


#

Gastric transit time

The time needed for the capsule to reach the duodenal bulb.


#

Small-bowel transit time

The time needed for the capsule to traverse the entire small bowel and reach the cecum.


#
#

Inclusion criteria

Inclusion criteria were determined before the beginning of the literature search. A study was considered eligible for inclusion in the meta-analysis if it met the following criteria: (i) published as a full article in English; (ii) randomized controlled design with adult participants; (iii) contained raw data regarding at least one of the five endpoints; (iv) purgative bowel preparation regimen given before capsule ingestion; (v) anti-foaming (e. g. simethicone) and prokinetic agents excluded; (vi) same type of capsule used for all arms in the study.


#

Quality assessment of studies

Cochrane collaboration’s risk of bias assessment tool [8] evaluates six domains of potential source of bias namely: random sequence generation (selection bias); allocation concealment (selection bias); blinding of participants and personnel (performance bias); blinding of outcome assessment (detection bias); incomplete outcome data (attrition bias); and selective reporting (reporting bias). Authors reviewed and judged studies according to each domain as having a high, low, or unclear risk of bias, providing at the same time support for their judgement. In the absence of a validated quality scale for VCE studies, we modified the performance bias domain of the tool in order to judge only personnel blinding, because VCE studies evaluating bowel preparation cannot be double blinded.

Two authors (P.G. and G.T.) independently assessed all of the studies for risk of bias and any discrepancies were resolved after discussion with the senior author (K.T.).


#

Statistical analysis

Extracted data were entered in a suitable form (number of events, means, medians, etc.) and meta-analyzed using the statistical software Review Manager (RevMan 5.3.5. Copenhagen, Denmark: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

For diagnostic yield, SBVQ, and completion rate, the risk ratios (RRs) and their 95 % confidence intervals (95 %CIs) were calculated. For continuous data in our study – GTT and SBTT – the inverse variance statistical method was used and mean difference with 95 %CI was calculated. Where necessary, mean value and standard deviation (SD) were calculated, taking into account the median and variance, as proposed by Hozo et al. [9].

All meta-analytic outcomes were further compared using either the fixed-effects model (Mantel and Haenszel method) or the random-effects model (DerSimonian and Laird method) in the absence or presence of significant heterogeneity, respectively. Finally, appropriate forest plots were created to visually display the results.


#

Assessment of heterogeneity and sensitivity analyses

Heterogeneity among studies was measured using I 2, with lower values representing lower levels of heterogeneity. In case of significant heterogeneity (P < 0.1), we examined potential excessive influence of a sole study in our results. Sensitivity analysis was therefore performed by repeating the meta-analysis excluding one study at a time to assess the effect of exclusion in the heterogeneity’s significance. For each outcome, we performed additional predefined subgroup analyses for each type of purgative regimen.


#

Publication bias

In order to estimate potential publication bias, funnel plots were constructed by plotting the log odds ratios (ORs) vs. precision of individual studies per outcome and visually assessing these for symmetry [10].


#
#

Results

Descriptive assessment and study characteristics

The search strategy yielded 288 citations from the two databases. No other article was identified from the searching of reference lists. After preliminary title and abstract review, 260 articles were excluded as non-relevant and/or duplicates; therefore, 28 articles were further assessed for eligibility. Among these, 16 were deemed to be ineligible for a variety of reasons, which left 12 articles [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] that met all of the eligibility criteria to be included in the meta-analysis. The flow chart describing the process is depicted in  [Fig.1].

Zoom Image
Fig. 1 Flowchart showing study inclusion and exclusion.

The 12 meta-analyzed studies included 17 sets of data that compared VCE outcomes in 736 and 485 individuals who had received purgative bowel preparation and clear liquids/fasting, respectively.

Among the retrieved sets of data, 10 evaluated PEG [11] [12] [13] [14] [15] [16] [17] [18], four were for regimens based on Na-P [13] [18] [20] [21] (in one study [20], Na-P was combined with bisacodyl), while magnesium-based regimens were evaluated in three studies [11] [19] [22]. More precisely, magnesium citrate (MgC) was evaluated as the sole preparation regimen [19], and in combination with sodium picosulfate (P/MgC) [11], while magnesium carbonate/anhydrous citric acid with the anthraquinone stimulant laxative senna was administered in one study [22]. In two studies, a small identical amount of simethicone was used as complementary therapy in both arms of the study [11] [22]. Because simethicone was not comparatively evaluated, we decided to include these two studies (three sets of data) in the meta-analysis.

Eight studies were conducted in Western countries [11] [13] [14] [15] [16] [20] [21] [22] (seven of these in Europe) and four in the East [12] [17] [18] [19]; there were only two multicenter studies [18] [21].

Nine studies reported on the device used and participants received M2A capsules (Given Imaging Ltd., Yoqneam, Israel) in five studies [13] [15] [16] [17] [18], PillCam SB1 (Given Imaging Ltd.) in three [19] [21] [22], and PillCam SB2 (Given Imaging Ltd.) in one study [14].

 [Table1] presents the main characteristics of the included studies.

Table 1

Characteristics of the included studies.

Author
(Year)

Country of origin

Capsule model

Study design

Type of small-bowel preparation

Participants, n

Outcomes reported

Purgative

Comparator

Purgative

Comparator

DY

SBVQ[*]

CR

GTT

SBTT

Viazis (2004) [16]

Greece

M2A

Prospective, randomized, controlled, single-blinded, single center

2 L PEG 16 hours before VCE, clear liquids the day before VCE, overnight fasting

Clear liquids the day before VCE and overnight fasting

40

40

Van Tuyl (2007) [15]

The Netherlands

M2A

Prospective, randomized, controlled, single-blinded, single center

A) 1 L PEG the evening before VCE, overnight fasting;
B) 2 L PEG the evening before VCE, overnight fasting

Clear liquids for 12 hours the day before VCE and overnight fasting

A) 30
B) 30

30

Wei
(2008) [17]

China

M2A

Prospective, randomized, controlled, single-blinded, single center

1 L PEG 12 hours before VCE

1 L clear water 12 hours before VCE, then fasting

30

30

Franke (2008) [20]

Germany

N/A

Prospective, randomized, controlled, single blinded, single center

12 hours of fasting followed by 30 mL
Na-P 3 hours before VCE and 20 mg bisacodyl just before VCE ingestion

At least 8 hours of fasting before, water allowed till 2 hours before VCE

26

26

Lapalus (2008) [21]

France

Pillcam SB1

Prospective, randomized, controlled, single-blinded, multicenter

45 mL Na-P and 2 L clear liquids the evening before VCE, plus 45 mL Na-P the morning of the procedure

Clear liquids the evening before VCE followed by 8 hours fasting

63

64

Wi
(2009) [18]

Korea

M2A

Prospective, randomized, controlled, single-blinded, multicenter

A) 45 mL Na-P at 15:00 h and 45 mL Na-P at 19:00 h on the day before capsule ingestion, followed by 2 L of clear liquids till midnight;
B) 2 L PEG 16 hours before capsule ingestion

12 hours of fasting

A) 45
B) 45

44

Postgate (2009) [22]

United Kingdom

Pillcam SB1

Prospective, randomized, controlled, single-blinded, single center

Two packets of senna at 14:00 h and two packets magnesium carbonate/anhydrous citric acid (one at 14:00 h and one at 18:00 h) on the day before capsule ingestion. VCE ingested with 0.5 mL add-on simethicone

Clear liquids on the evening before capsule ingestion followed by 10 hours of fasting. VCE ingested with 0.5 mL add-on simethicone

40

38

Park (2011) [12]

Korea

N/A

Prospective, randomized, controlled, single center

A) Overnight fasting and 2 L PEG 4 hours before capsule ingestion;
B) Overnight fasting and 4 L PEG 4 hours before capsule ingestion

12 hours of fasting

A) 20
B) 25

23

Pons Beltran (2011) [13]

Spain

M2A

Prospective, randomized, controlled, single-blinded, multicenter

A) 90 mL Na-P and 4 L clear liquid 24 hours before VCE, followed by 8 hours of fasting;
B) 4 L PEG 24 hours before VCE, followed by 8 hours of fasting

4 L of clear liquids during 24 hours before VCE, followed by 8 hours of fasting

A) 89
B) 92

92

Ninomiya (2012) [19]

Japan

Pillcam SB1

Prospective, randomized, controlled, single-blinded, single center

34 g MgC 12 hours before VCE, then fasting

12 hours of fasting

22

22

Rosa (2013) [14]

Portugal

Pillcam SB2

Prospective, randomized, controlled, single-blinded, single center

24 hours of liquid diet, followed by 2 L PEG the evening before capsule ingestion

24 hours of liquid diet, followed by overnight fasting

20

20

Hookey (2017) [11]

Canada

N/A

Prospective, randomized, controlled, single-blinded, single center

A) 2 L PEG the evening before VCE and overnight fasting;
B) One sachet P/MgC at 16:00 h and one sachet P/MgC at 21:00 h the day before VCE, and overnight fasting. Patients received 80 mg simethicone 10 minutes before capsule ingestion

12 hours of clear liquids the day before and overnight fasting. Patients received 80 mg simethicone 10 minutes before capsule ingestion

A) 58
B) 61

56

VCE, video capsule endoscopy; PEG, polyethylene glycol; Na-P, sodium phosphate; MgC, magnesium citrate; P/MgC, sodium picosulfate plus magnesium citrate; DY, diagnostic yield; SBVQ, small-bowel visualization quality; CR, completion rate; GTT, gastric transit time; SBTT, small-bowel transit time; N/A, not available; • outcome reported; – outcome not reported.

* Applicable if adequate vs. inadequate classification was possible and data for evaluation of the whole small bowel were available.



#

Quality assessment

[Fig. 2] summarizes the assessment of per-study risk of bias according to Cochrane collaboration’s risk of bias assessment tool. Exact judgement per study and per quality domain can be found in [Appendix  e3] (available online). It is remarkable that in four [14] [18] [20] [21] and seven [12] [14] [15] [17] [18] [20] [22] of the studies detailed methods of randomization and allocation concealment were not described (uncertain selection bias).

Zoom Image
Fig. 2 Illustration of the risk of bias for the 12 included studies. + low risk of bias; − high risk of bias; ? unknown risk of bias.

#

Primary endpoint

VCE diagnostic yield

There were 10 studies [11] [12] [13] [14] [15] [16] [18] [19] [21] [22] with 15 data sets that investigated the effect of cathartics on the diagnostic yield of VCE. Overall, 677 and 673 patients received purgative bowel preparation and clear liquids/fasting, respectively. Using the random-effects model because of the presence of significant heterogeneity (I 2 = 41 %; P = 0.05), no statistically significant difference was detected regarding the diagnostic yield between the two arms, overall. There were 289 (42.7 %) and 245 (36.4 %) positive studies in the cathartics and in the clear liquids/fasting arms, respectively (overall RR [95 %CI] was 1.17 [0.97 to 1.40]; P = 0.11) ([Fig. 3a]). We did not detect any publication bias ([Fig. 3b]).

Zoom Image
Fig. 3 Results for studies that assessed the diagnostic yield of small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MgC, magnesium citrate; RR, risk ratio.

Sensitivity analysis by excluding one study every time revealed that there were two studies [16] [22] responsible for the significant heterogeneity. However, exclusion of only one [22] of these two studies changed the significance of our results in slight favor of purgative bowel preparation overall (RR 1.23 [95 %CI 1.07 to 1.40]; P = 0.002).

Overlapping confidence intervals ( [Fig.3a]) and testing for subgroup differences (χ2 = 2.32, df = 2; P = 0.31) did not indicate differences in the per-regimen subgroup analysis.


#
#

Secondary endpoints

Small-bowel mucosal visualization quality

We retrieved data regarding SBVQ from six studies [11] [13] [14] [16] [17] [18] (nine sets of data) including 478 patients prepared with cathartics and 474 receiving only clear liquids/fasting. Six more studies [12] [15] [19] [20] [21] [22] provided data regarding SBVQ, but were not included in the meta-analysis because they either evaluated SBVQ using a composite score that could not be dichotomized or reported segmental outcomes only ([Table e2]; available online).

Tablee 2

Studies with mucosal visualization scores that could not be dichotomized.

Study (Year)

Reported score

Reason study could not be dichotomized

Park (2011) [12]

Scoring system composed of two visual properties (proportion of mucosa visualized and extent of obscuration by bubbles, debris and bile etc.). Each property was scored using a 4-step scale (0 to 3). Images were manually selected at 5-minute intervals and mean scores for each property were calculated. The average of two means was considered the representative score for each patient. Using ROC curve values above 2.25 were considered as adequate cleansing

Authors report median scores (+/- interquartile range [IQR]) per group of participants and per segment. Score per patient not reported

Van Tuyl (2007) [15]

Small-intestine transit time divided into quartiles. Segments of 10 minutes were examined at the beginning and at the end of each quartile. Visualization classified into six categories depending on the percentage of the mucosal surface that could be visualized per quartile. Visualization of 75 % or more of the mucosa was considered good visibility

Percentages of patients with good visibility are presented per group and per segment. An average rating of visibility per patient is not provided. A score per patient for the entire small bowel is not reported

Lapalus (2008) [21]

Two scores (ranging from 1 to 4) were used to evaluate the bowel cleanliness and the mucosal visibility. The final score of quality was the mean of the two above values as provided by the two blinded investigators at each participating center

Median scores ( + /- IQR) and percentages of patients with good and excellent quality of preparation are reported per segment. Scores or percentages of patients are not reported for the entire small bowel

Franke (2008) [20]

Mucosal visibility was assessed using a scale from 1 (excellent) to 4 (poor). Two 1-hour periods were evaluated; one in the proximal and one in the distal small bowel

Median score of small-bowel mucosal visibility is presented per group and per segment (proximal vs. distal small bowel)

Ninomiya (2012) [19]

Three effect scores were used. Residue elimination effect score (1 to 4 points), intestinal juice clarity score (1 to 4 points), and froth reduction effect score (1 to 4 points). Small bowel was divided into upper, middle, and lower sections; images recorded for each of these segments were used to evaluate the aforementioned scores. Cleansing effect was calculated by summing the three scores for all three small-bowel sections

Authors do not provide any cut-off above which the mucosal visibility could be considered adequate according to the score they used. Accordingly, the number of patients per group whose preparation could be considered adequate is not reported. Scores are presented as mean +/- standard deviation either for the entire small bowel or per segment

Postgate (2009) [22]

A 5-point scale (0 – 4) based on the percentage of the capsule image that was unimpaired by debris or dark luminal fluid (80 % – 100 % visible, score 4; 60 % – 80 % visible, score 3; etc.) was used. An average score was estimated for 5-minute segments of video, assessed at capsule entry into the proximal duodenum, and evenly spaced for every 10 % of SBTT thereafter, with the final segment in the terminal ileum. Accordingly, a maximum score of 44 was possible

Authors do not provide any cut-off above which the mucosal visibility could be considered adequate according to the score they used. Accordingly, the number of patients per group whose preparation could be considered adequate is not reported. Scores are presented as median +/- IQR either for the entire small bowel or per segment

ROC, receiver operating characteristic; SBTT, small-bowel transit time.

Three of the included studies [16] [17] [18] evaluated the adequateness of SBVQ according to the scoring system of Viazis et al. [16]. Small-bowel cleansing was considered “adequate” if the objective score (the percentage of SBTT during which less than 25 % of the intestinal mucosa was clean) was less than 10 % and “inadequate” if it was more than 10 %. The rest of the studies used either a 4 – or 5-point scale (ranging from inadequate to excellent mucosal visualization) using either qualitative [13] or quantitative criteria [11] [14]. As predefined, we homogenized the available data by grouping “good” and “excellent” mucosal visualization along with “adequate.” All other characterizations were grouped as “inadequate.”

After data homogenization, no difference was detected regarding SBVQ between the two arms. Overall, adequate mucosal visualization was achieved in 300 of the patients who received purgatives (62.7 %) and in 268 of those receiving only clear liquids/fasting (56.5 %). Significant heterogeneity was present (I 2 = 59 %; P = 0.01) and the effect size of the analysis ([Fig. 4 a]) did not indicate any favor of purgative bowel preparation over clear liquids/fasting (RR 1.14 [95 %CI 0.96 to 1.35]; P = 0.15). No publication bias was detected ( [Fig. 4 b]).

Zoom Image
Fig. 4 Results for studies that assessed small-bowel mucosal visualization quality with video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgC, magnesium citrate; RR, risk ratio.

The significance of analysis results did not change by excluding one study at a time.

No differences were detected in the per-regimen subgroup analysis, overlapping confidence intervals ([Fig. 4 a]), or test for subgroup differences (χ2 = 5.61, df = 2; P = 0.06).


#

Completion Rate

A total of 11 studies [11] [12] [13] [14] [15] [16] [17] [18] [19] [21] [22] with 16 sets of data provided data that the examination was completed in 711 and 709 patients administered purgative bowel preparation and clear liquids/fasting, respectively. Overall, we detected no heterogeneity (I 2 = 0 %; P = 0.96) and no difference in the completion rate of the examination between patients receiving purgatives (588 /711 [82.7 %]) and those receiving clear liquids/fasting (590 /709 [83.2 %]); RR 0.99 [95 %CI 0.95 to 1.04]; P = 0.76) ([Fig. e 5 a]; available online).

Zoom Image
Fig. e5 Results for studies that assessed completion rate of small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MgC, magnesium citrate; RR, risk ratio.

Visual inspection of the funnel plot ([Fig. e 5 b]) did not reveal publication bias. The per-regimen subgroup analysis did not reveal any difference among the used regimens ([Fig. e 5 a]).


#

Gastric transit time

There were 13 sets of data from nine studies [12] [13] [14] [15] [16] [18] [20] [21] [22] that reported on GTT. Overall, GTT was evaluated in 559 and 558 patients prepared with purgative bowel preparation and clear liquids/fasting, respectively. For two studies [12] [22], mean (SD) GTTs were calculated using the provided median, range, and sample size. Data were characterized from significant heterogeneity (I 2 = 93 %; P < 0.001). Analysis revealed no difference in GTT between the two groups (mean difference = −2.71 [95 %CI −7.87 to 2.46]; P = 0.30) ( [Fig. e 6 a]; available online) with no evidence of publication bias ([Fig. e 6 b]).

Zoom Image
Fig. e6 Results for studies that assessed gastric transit time for small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MD, mean difference.

The significance of heterogeneity was not attributed to any of the meta-analyzed studies in the step-by-step sensitivity analysis.

The significance detected in the test for per-regimen subgroup differences ( [Fig. e 6 a]) should be interpreted cautiously as it may simply reflect a lack of information rather than a different effect.


#

Small-bowel transit time

SBTT was reported in 10 studies [12] [13] [14] [15] [16] [17] [18] [20] [21] [22] (14 sets of data). SBTT was recorded in 581 and 584 patients prepared with purgative bowel preparation and clear liquids/fasting, respectively. For two of the studies [12] [22], we used median, range, and sample size to calculate the mean (SD) SBTT according to Hozo et al. [9]. Significant heterogeneity (I 2 = 95 %; P < 0.001) was present and analysis did not reveal any difference in SBTT between the two arms (mean difference = −8.35 [95 %CI −20.41 to 3.70], Z = 1.36; P = 0.17) ( [Fig. e 7 a]; available online). No publication bias was identified ( [Fig. e 7 b]).

Zoom Image
Fig. e7 Results for studies that assessed small-bowel transit time for small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MD, mean difference.

During the step-by-step sensitivity analysis, we did not detect any study that if excluded would diminish the heterogeneity’s significance.

The per-regimen subgroup analysis did not reveal any difference among the used regimens ([Fig. e 7 a]).


#
#
#

Discussion

In our center and in contrast to the recommendation of the manufacturer, we instruct patients to consume a clear-liquid diet on the day before their examination but they additionally receive 2 L of PEG on the evening before the procedure. However, the usefulness of purgative bowel preparation remains controversial.

In an era of financial barriers [23], the diagnostic yield of the examination remains the most important outcome of small-bowel VCE. In the presence of significant heterogeneity, our meta-analysis of 12 randomized controlled trials, involving more than 1200 subjects, questions the efficacy of purgative preparation compared with clear liquids and/or fasting on the diagnostic yield of small-bowel VCE. To our great surprise, the significant heterogeneity disappeared after exclusion of a single study (with only 78 participants) [22] and repetition of the analysis favored administration of purgative with regard to the diagnostic yield of VCE. One might assume that the extreme and unusually low diagnostic yield, especially in the purgative group of the aforementioned study, (5 /39 [12.8 %] and 13 /37 [35.1 %] in the purgative bowel preparation and clear liquids/fasting groups, respectively) could be responsible for the difference that was noticed.

Our primary endpoint result is not in accordance with the results of four previous meta-analyses [1] [2] [3] [4] ([Table 3]) that revealed a significant benefit of active bowel preparation (pooled ORs ranging from 1.68 to 1.88); the sensitivity analysis in the most recent of these meta-analyses [4] showed that the benefit of the purgative bowel preparation was attributed merely to a single study [16]. In contrast, our findings are in line with the two most recently published meta-analyses [5] [6], which also failed to detect any benefit of active bowel preparation on the diagnostic yield of the examination giving ORs (95 %CI) of 1.55 (0.79 to 3.04) and 1.11 (0.85 to 1.44), respectively.

Table 3

Meta-analyses evaluating small-bowel preparation and video capsule endoscopy (VCE) outcomes.

Author (Year)

Intervention

Outcomes

Number of studies

Type of studies

Number of patients

Regimens evaluated

Quality assessment

Sensitivity analyses

Niv
(2008) [1]

Preparation of any kind vs. no preparation

SBVQ; CR; GTT; SBTT

8

2 RCTs;
2 prospective non-randomized;
4 retrospective

437

PEG; Na-P; simethicone; erythromycin

None mentioned

None performed

Rokkas (2009) [2]

PBP vs. CL/F

Primary: DY and SBVQ;
Secondary: CR, GTT, and SBTT

12

6 RCTs;
6 retrospective

1162

PEG; Na-P

Self-proposed score:
1) Study design: prospective (1 point), retrospective/cohort (0);
2) Number of examiners: > 1 (1), one (0);
3) Examiners blinded to preparation: yes (1), no (0);
4) Grades for overall bowel cleansing: ≥ 3 (1), < 3 (0);
5) Entire small-bowel evaluation: yes (1), no (0)

1) Step-by-step one study removed if heterogeneity present;
2) Full paper vs. abstracts;
3) Prospective vs. retrospective;
4) PEG vs. Na-P

Belsey (2012) [3]

Impact of PBP, antifoaming agents and their combination on VCE outcomes

SBVQ; DY

8

8 RCTs

746

PEG; Na-P; simethicone; bisacodyl

No specific score or tool mentioned

PEG vs. Na-P

Kotwal (2014) [4]

Impact of PBP, antifoaming agents, their combination or prokinetics on VCE outcomes

SBVQ; DY; CR

15

15 RCTs

1468

PEG; Na-P; simethicone; mannitol; dimethylpolysiloxane; erythromycin; mosapride; metoclopramide (either alone or as different combinations)

Modified Jadad Scale (maximum of 9 points):
Randomization (2 points)
Blinding procedures (4 points)
Extent of attrition (3 points)

Step-by-step one study removed if heterogeneity present

Wu (2017) [5]

Network meta-analysis for the efficacy of bowel preparation with different doses of PEG vs. CL/F

Primary: SBVQ;
Secondary: DY and CR

9

9 RCTs

982

PEG

Cochrane Collaboration’s tool

None performed

Yung (2017) [6]

Laxatives vs. no laxatives

Primary: DY;
Secondary: SBVQ and CR

40

24 RCTs;
5 prospective non-randomized;
11 retrospective

6565

PEG; Na-P; MgC; mannitol

Score as proposed by Rokkas et al:
1) Study design: prospective (1 point), retrospective/cohort (0)
2) Number of examiners: > 1 (1), one (0)
3) Examiners blinded to preparation: yes (1), no (0)
4) Grades for overall bowel cleansing: ≥ 3 (1), < 3 (0);
5) Entire small-bowel evaluated: yes (1), no (0)
High quality: 4/5 and above;
Moderate quality: 3/5;
Low quality: 2/5 and below

1) Repetition of meta-analysis with exclusion of outliers if heterogeneity present;
2) Type of laxatives used;
3) Use of simethicone and/or prokinetics;
4) Timing of administration of laxatives;
5) Large studies only ( ≥ 30 VCEs in both laxative and control groups)

CL/F, clear liquids and/or fasting; CR, completion rate; DY, diagnostic yield; GTT, gastric transit time; MgC, magnesium citrate; Na-P, sodium phosphate; PBP, purgative bowel preparation; PEG, polyethylene glycol; RCT, randomized controlled trial; SBTT, small-bowel transit time; SBVQ, small-bowel visualization quality.

While all of these meta-analyses – ours included – suffer from significant heterogeneity, the discrepant results among them could be attributed to different methodological design, type of included studies, small sample sizes, and variable endpoints. In particular, the first meta-analysis included prospective and retrospective studies for all types of preparation [1], while the two subsequent studies examined the role of purgative bowel preparation regardless of the study design [2] [6]. Two more studies meta-analyzed only randomized controlled trials [3] [4], investigating the role of purgatives, antifoaming agents, prokinetics and/or their combinations on VCE outcomes, while the most recent network meta-analysis [5] – including both retrospective and prospective studies – tried to identify the most efficient PEG schema.

The “immaturity” of physicians in interpreting VCE findings during the first years of VCE utilization, as shown recently in a multicenter retrospective study that revealed that the rate of positive studies progressively declined in a period between 2002 and 2014 [23], and the lack of standardized definitions for findings comprise other potential explanations of the discrepant meta-analysis results with regarding to the diagnostic yield of VCE.

Our analysis of six RCTs enrolling more than 950 individuals failed to confirm the findings of Rokkas et al. [2] who meta-analyzed RCTs and cohort studies with 653 patients and showed that administration of cathartics resulted in improved small-bowel mucosal inspection. Our findings are consistent with the largest meta-analysis so far [6], which regardless of the design of studies showed no benefit of purgative administration on SBVQ, with the exception of a minimal SBVQ improvement in patients receiving Na-P or simethicone alongside their preparation [6]; this finding was not replicated in our per-regimen sensitivity analysis.

Our results are in line with previous meta-analytic findings [2] [4] [5] [6] showing that the rate of complete examinations does not increase after the administration of purgatives. Moreover, as previously reported [2], the GTT and SBTT are not accelerated by purgative bowel preparation. Nevertheless, the issues of accelerating the passage of the capsule in the small bowel and of increasing the completion rate of VCE in order to potentially improve the diagnostic yield of the examination may be out of date given the advent of new-generation capsule endoscopes with longer battery life, the use of which has a significant impact on VCE completion rate [24] [25]. In addition, the latest generation capsules (PillCam SB3; Given Imaging Ltd.) offer not only prolonged video recording (up to 12 hours), but also advanced optics, with better image resolution and a variable adaptive frame rate (ranging from 2 to 6 frames per second). It is obvious that these characteristics could affect not only completion rate but also small-bowel mucosal visualization and the diagnostic yield of the examination.

The major strength of this systematic review and meta-analysis is the inclusion of RCTs evaluating the effect of purgative bowel preparation only before capsule ingestion on the outcomes of VCE. In this way, we abolished any potential bias attributable to study design (retrospective studies, cohort studies missing comparator arms, etc.) and to non-purgative preparation agents (anti-foaming agents and/or prokinetics) that might have influenced the outcomes of previous meta-analyses. Moreover, rigorous quality assessment of the studies, absence of publication bias, and sensitivity and subgroup analyses applied to minimize heterogeneity comprise additional advantages of our study.

The main limitation of our meta-analysis is the significant heterogeneity among the included studies. In an effort to blunt this problem, we performed sensitivity analyses, but heterogeneity originates from multiple and diverse sources that are difficult to deal with. First, the heterogeneous mixture of populations and indications that might be related to different VCE outcomes. Of the 12 included studies, seven originated from the West, while there were only two multicenter studies and three of the studies enrolled less than 30 individuals per arm.

Second, there is an absence of standardized and validated definitions for diagnostic yield and SBVQ. Although diagnostic yield is more or less “self-defined,” several authors classify their VCE findings as “positive,” “relative,” or even “suspicious,” and the classification proposed by Saurin et al. [26], although adapted in some studies, has never been validated. Regarding SBVQ, multiple simple or composite scores, as well as the arbitrary judgments of authors, have been used. In addition, we were unable to reclaim SBVQ data from six of the 12 included studies, because the authors provided evidence in a very heterogeneous manner.

Finally, the fact that studies with different purgative regimens have been included in our meta-analysis might be considered a limitation of our work, although it does reflect clinical practice; we tried to eliminate this issue by performing per-regimen subgroup analysis.

In conclusion, our study challenges the usefulness of purgative bowel preparation before small-bowel VCE. Performing VCE without any prior bowel preparation could lead – from a patient’s perspective – to a more convenient examination and at the same time save some costs for the healthcare system. However, significant heterogeneity among meta-analyzed studies owing to the lack of standardized definitions, different patients demographics, different study indications, unjustified sample sizes, and use of older generation capsules may weaken a recommendation for abandoning purgative bowel preparation for VCE.


#

Appendix e1 Search strategy

Search strategy; MEDLINE database search strategy.

PubMed

Step

Search strategy

Found

Time

#1

Search (("Capsule Endoscopy"[Mesh] OR ("capsule endoscopy"[MeSH Terms] OR ("capsule"[All Fields] AND "endoscopy"[All Fields]) OR "capsule endoscopy"[All Fields]))) Filters: Publication date from 2001 /01 /01 to 2017 /07 /31

4195

16:31:30

#2

Search ((("Cathartics"[Mesh] OR preparation[All Fields]) OR (purgativ[All Fields] OR purgativa[All Fields] OR purgativas[All Fields] OR purgative[All Fields] OR purgativelor[All Fields] OR purgativen[All Fields] OR purgatives[All Fields] OR purgativna[All Fields] OR purgativni[All Fields] OR purgativnih[All Fields] OR purgativnim[All Fields] OR purgativo[All Fields]))) Filters: Publication date from 2001 /01 /01 to 2017 /07 /31

163129

16:31:52

#3

#1 AND #2

210

16:32:11

Cochrane Central Register of Clinical Trials database search strategy.

ID

Search hits

Found

#1

MeSH descriptor: [Capsule Endoscopy] explode all trees

154

#2

MeSH descriptor: [Cathartics] explode all trees

659

#3

bowel preparation: ti, ab, kw and capsule endoscopy (Word variations have been searched)

78

#4

#1 and #2

15

#5

#3 or #4

78


#

Appendix e2 PRISMA 2009 checklist

Section/topic

#

Checklist item

Reported in the following section

Title

Title

1

Identify the report as a systematic review, meta-analysis, or both

Title

Abstract

Structured summary

2

Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number

Abstract

Introduction

Rationale

3

Describe the rationale for the review in the context of what is already known

Introduction, paragraphs 1 and 2

Objectives

4

Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS)

Introduction, paragraph 2

Methods

Protocol and registration

5

Indicate if a review protocol exists, if and where it can be accessed (e. g. web address), and, if available, provide registration information including registration number

Eligibility criteria

6

Specify study characteristics (e. g. PICOS, length of follow-up) and report characteristics (e. g. years considered, language, publication status) used as criteria for eligibility, giving rationale

Data identification and extraction; Inclusion criteria

Information sources

7

Describe all information sources (e. g. databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched

Data identification and extraction

Search

8

Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated

[Appendix e1 ]

Study selection

9

State the process for selecting studies (i. e. screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis)

Data identification and extraction

Data collection process

10

Describe method of data extraction from reports (e. g. piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators

Data identification and extraction

Data items

11

List and define all variables for which data were sought (e. g. PICOS, funding sources) and any assumptions and simplifications made

Data identification and extraction; Study endpoints

Risk of bias in individual studies

12

Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis

Quality assessment of studies

Summary measures

13

State the principal summary measures (e. g. risk ratio, difference in means)

Statistical analysis

Synthesis of results

14

Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e. g. I 2) for each meta-analysis

Statistical analysis; Assessment of heterogeneity and sensitivity analysis

Risk of bias across studies

15

Specify any assessment of risk of bias that may affect the cumulative evidence (e. g. publication bias, selective reporting within studies)

Publication bias

Additional analyses

16

Describe methods of additional analyses (e. g. sensitivity or subgroup analyses, meta-regression), if done, indicating which were prespecified

Assessment of heterogeneity and sensitivity analysis

Results

Study selection

17

Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram

Descriptive assessment and study characteristics (see also [ Fig. 1 ])

Study characteristics

18

For each study, present characteristics for which data were extracted (e. g. study size, PICOS, follow-up period) and provide the citations

Descriptive assessment and study characteristics (see also [ Table 1 ])

Risk of bias within studies

19

Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12)

Quality assessment (see also [ Fig. 2 ]; [ Appendix e3 ])

Results of individual studies

20

For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot

Primary endpoint; Secondary endpoints

Synthesis of results

21

Present results of each meta-analysis done, including confidence intervals and measures of consistency

Primary endpoint; Secondary endpoints

Risk of bias across studies

22

Present results of any assessment of risk of bias across studies (see Item 15)

Primary endpoint; Secondary endpoints

Additional analysis

23

Give results of additional analyses, if done (e. g. sensitivity or subgroup analyses, meta-regression [see Item 16])

Primary endpoint; Secondary endpoints

Discussion

Summary of evidence

24

Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e. g. healthcare providers, users, and policy makers)

Discussion, paragraphs 1 and 2

Limitations

25

Discuss limitations at study and outcome level (e. g. risk of bias), and at review-level (e. g. incomplete retrieval of identified research, reporting bias)

Discussion, paragraphs 8 – 10

Conclusions

26

Provide a general interpretation of the results in the context of other evidence, and implications for future research

Discussion, paragraphs 3 – 6 and 11

Funding

Funding

27

Describe sources of funding for the systematic review and other support (e. g. supply of data); role of funders for the systematic review

NA

PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analysis; PICOS, population, intervention, comparator, outcome; NA, not applicable.


#

Appendix e3 Risk of bias assessment for each included study, with authors’ judgement and support for judgement.

Bias

Authorsʼ judgement

Support for judgement

Franke (2008) [20]

Random sequence generation (selection bias)

Unclear risk

Authors mention the randomized design of the study but the exact method of randomization is not described

Allocation concealment (selection bias)

Unclear risk

The enrolled patients were randomly allocated to one of two groups, but whether the list was concealed is not clarified

Blinding of participants and personnel (performance bias)

Low risk

Personnel blinded, but no blinding of participants; outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

All investigators were blinded to preparation provided

Incomplete outcome data (attrition bias)

Low risk

No missing outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Hookey (2017) [11]

Random sequence generation (selection bias)

Low risk

Subjects were randomized to one of the three treatment groups via consecutively numbered opaque envelopes and random computer-generated numbers prepared by an independent biostatistician

Allocation concealment (selection bias)

Low risk

Patients were allocated in randomly ordered permuted blocks of sizes 3, 6, and 9 without stratification

Blinding of participants and personnel (performance bias)

Low risk

The clinical research assistant performed randomization and preparation instructions after the clinician had left the room; the patients were asked not to reveal to the clinician which preparation they were assigned during future contact

Blinding of outcome assessment (detection bias)

Low risk

The capsule examinations were not reviewed for study purposes until after the last patient had completed the study; the reviewers were blinded regarding preparation

Incomplete outcome data (attrition bias)

Low risk

Reasons for missing outcome data unlikely to be related to true outcome; missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; missing data have been imputed using appropriate methods

Selective reporting (reporting bias)

Low risk

The study protocol is available and all of the study’s prespecified outcomes that are of interest in the review have been reported in the prespecified way

Lapalus (2008) [21]

Random sequence generation (selection bias)

Unclear risk

Authors mention the randomized design of the study and the use of envelopes; however, the exact method of randomization is not described

Allocation concealment (selection bias)

Low risk

All patients eligible to enter the study were randomly allocated into two groups by using sealed envelopes

Blinding of participants and personnel (performance bias)

Low risk

Each group was instructed to follow a different preparation regimen; however, this had no effect on outcome; personnel evaluating outcome were blinded to the provided regimen

Blinding of outcome assessment (detection bias)

Low risk

Sixteen investigators independently evaluated the VCE images; they were blinded to which group each patient was randomized

Incomplete outcome data (attrition bias)

Low risk

Missing outcome data unlikely to be related to true outcome and are balanced across intervention groups

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Ninomiya (2012) [19]

Random sequence generation (selection bias)

Low risk

Authors mention the randomized design of the study and the use of envelopes; however, the exact method of randomization is not described

Allocation concealment (selection bias)

Low risk

The subjects were randomly allocated to two groups using opaque envelopes

Blinding of participants and personnel (performance bias)

Low risk

No blinding of participants; personnel blinded; outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

All video capsule endoscopy results were blind evaluated by an endoscopist who was not informed of the group allocation of subjects

Incomplete outcome data (attrition bias)

Low risk

Missing outcome data unlikely to be related to true outcome and are balanced across intervention groups

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Park (2011) [12]

Random sequence generation (selection bias)

Low risk

The enrolled patients were randomly allocated to one of the three groups by use of a computer-generated list

Allocation concealment (selection bias)

Unclear risk

The enrolled patients were randomly allocated to one of the three groups by use of a computer-generated list, but whether the list was concealed is not clarified

Blinding of participants and personnel (performance bias)

High risk

No blinding of participants; it is not clear if personnel were blinded; outcome is likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

High risk

Authors do not mention the readers’ characteristics nor if they were blinded to the provided regimen

Incomplete outcome data (attrition bias)

Low risk

No missing outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Pons Beltran (2011) [13]

Random sequence generation (selection bias)

Low risk

273 patients were included in the data analysis and were randomized into three groups using an information-technology system

Allocation concealment (selection bias)

Low risk

An information technology system was used to allocate the patients

Blinding of participants and personnel (performance bias)

Low risk

No blinding of either participants or personnel but outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

All researchers were blind to the group to which patients belonged

Incomplete outcome data (attrition bias)

Low risk

No incomplete outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Postgate (2009) [22]

Random sequence generation (selection bias)

Low risk

Patients were prospectively randomized by using a computer-generated random number sequence

Allocation concealment (selection bias)

Unclear risk

Author mention the randomized character of the study but they do not clarify the allocation method

Blinding of participants and personnel (performance bias)

Low risk

No blinding of participants; personnel blinded; outcome is not likely to be influenced by lack of participants blinding

Blinding of outcome assessment (detection bias)

Low risk

Endoscopist evaluating the study outcome were unaware of the type of bowel preparation

Incomplete outcome data (attrition bias)

Low risk

No incomplete outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Rosa (2013) [14]

Random sequence generation (selection bias)

Unclear risk

Authors mention the randomized character of the study but do not clarify either the way of randomization or the allocation method

Allocation concealment (selection bias)

Unclear risk

Authors mention the randomized character of the study but do not clarify either the way of randomization or the allocation method

Blinding of participants and personnel (performance bias)

Low risk

No blinding of participants; personnel blinded; outcome is not likely to be influenced by lack of participants blinding

Blinding of outcome assessment (detection bias)

Low risk

Endoscopists evaluating the study outcome were unaware of the type of bowel preparation

Incomplete outcome data (attrition bias)

Low risk

No incomplete outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Van Tuyl (2007) [15]

Random sequence generation (selection bias)

Low risk

Patients were randomized by one of the investigators to one of the three preparation regimens by consecutive assignment using a computer-generated list of random numbers

Allocation concealment (selection bias)

Unclear risk

No data provided about concealment of the list of random numbers

Blinding of participants and personnel (performance bias)

Low risk

No blinding of either participants or personnel but outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

Each video capsule endoscopy file was reviewed by two independent investigators without knowledge of the preparation regimen

Incomplete outcome data (attrition bias)

Low risk

No incomplete outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Viazis (2004) [16]

Random sequence generation (selection bias)

Low risk

By using the sealed envelope technique, all patients were randomly allocated to two groups; a table of random numbers was used (one to one proportion and a blocked design of four)

Allocation concealment (selection bias)

Low risk

By using the sealed envelope technique, all patients were randomly allocated to two groups; a table of random numbers was used (one to one proportion and a blocked design of four)

Blinding of participants and personnel (performance bias)

Low risk

Personnel blinded, but no blinding of participants; outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

Three investigators independently evaluated the capsule endoscopy images; they were blinded as to which group each patient had been randomized

Incomplete outcome data (attrition bias)

Low risk

Missing outcome data balanced in numbers across intervention groups

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Wei (2008) [17]

Random sequence generation (selection bias)

Low risk

At least two expert endoscopists who were unaware of the type of bowel preparation evaluated the capsule endoscopy findings

Allocation concealment (selection bias)

Unclear risk

No data regarding allocation concealment

Blinding of participants and personnel (performance bias)

Low risk

Personnel blinded, participants not blinded; outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

One endoscopist who was unaware of the type of bowel preparation evaluated the capsule endoscopy findings

Incomplete outcome data (attrition bias)

Low risk

No missing outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes

Wi (2009) [18]

Random sequence generation (selection bias)

Unclear risk

The patients were assigned to groups A, B, or C using randomization lists

Allocation concealment (selection bias)

Unclear risk

No data regarding concealment of the lists

Blinding of participants and personnel (performance bias)

Low risk

Participants not blinded, but examiners blinded; outcome is not likely to be influenced by lack of blinding

Blinding of outcome assessment (detection bias)

Low risk

At least two expert endoscopists who were unaware of the type of bowel preparation evaluated the capsule endoscopy findings

Incomplete outcome data (attrition bias)

Low risk

No missing outcome data

Selective reporting (reporting bias)

Low risk

Protocol not available but the published reports include all expected outcomes


#
#

Competing interests

None.

Acknowledgments

The authors would like to thank Dr. Lawrence Hookey for kindly providing raw data from his original study regarding the quality of small-bowel mucosal visualization. We also thank Mr. Marios Detsis for reviewing the analysis and providing statistical expertise.

Appendix e1 – e3, Table e2, Figs. e5 – e7


Corresponding author

Konstantinos Triantafyllou, MD, PhD
Hepatogastroenterology Unit, Second Department of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center, Medical School, National and Kapodistrian University of Athens, “Attikon” University General Hospital
Rimini 1
12462, Athens
Greece   
Fax: +30-21-05326454   


Zoom Image
Fig. 1 Flowchart showing study inclusion and exclusion.
Zoom Image
Fig. 2 Illustration of the risk of bias for the 12 included studies. + low risk of bias; − high risk of bias; ? unknown risk of bias.
Zoom Image
Fig. 3 Results for studies that assessed the diagnostic yield of small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MgC, magnesium citrate; RR, risk ratio.
Zoom Image
Fig. 4 Results for studies that assessed small-bowel mucosal visualization quality with video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgC, magnesium citrate; RR, risk ratio.
Zoom Image
Fig. e5 Results for studies that assessed completion rate of small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MgC, magnesium citrate; RR, risk ratio.
Zoom Image
Fig. e6 Results for studies that assessed gastric transit time for small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MD, mean difference.
Zoom Image
Fig. e7 Results for studies that assessed small-bowel transit time for small-bowel video capsule endoscopy shown as: a a forest plot; b a funnel plot. CI, confidence interval; PEG, polyethylene glycol; Na-P, sodium phosphate; MgCO3, magnesium carbonate; MD, mean difference.