| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, (G1-215), Academic Medical Center, University of
Amsterdam, Meibergdreef 9, 1100 DD, Amsterdam, The Netherlands.
2 Department of Radiology, Onze Lieve Vrouwe Gasthuis, Amsterdam, The
Netherlands.
Received September 9, 2007;
accepted after revision January 27, 2008.
E-Z-Em provided the barium contrast agent (Tagitol V) without cost for this
study. The authors retained control of all aspects of the study, including
study design, analysis, and reporting.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Forty consecutive increased-risk patients were randomized. Group 1 received orally 20 mg of bisacodyl, group 2 received 30 mg of bisacodyl, group 3 received 20 mg of bisacodyl and 8.2 g of magnesium citrate, and group 4 received 30 mg of bisacodyl and 16.4 g of magnesium citrate. All patients used a 2-day low-fiber diet and received diatrizoate meglumine and barium for fecal tagging. One reviewer blindly scored subjective image quality (fecal tagging, amount of residual feces [liquid or solid], luminal distention, and image readability) on a 5- to 6-point scale using a 2D review technique. The mean and SD of attenuation of tagging were measured as well as the relative SD as a measure of homogeneity. Furthermore, patient acceptance (burden related to diarrhea, abdominal pain, flatulence, and overall burden) was evaluated. Ordinal regression, generalized estimating equations, and parametric and nonparametric tests were used for analysis.
RESULTS. Image readability was evaluated as good or excellent in all examinations except one in group 2 (nondiagnostic) and two in group 3 (moderate). Group 2 contained more feces than group 4 (p = 0.04). With regard to mean attenuation and homogeneity of tagging, no significant differences were observed between groups. Group 4 experienced more severe diarrhea than groups 1 and 2 and higher overall burden than groups 1 and 3 (p < 0.042).
CONCLUSION. The mildest preparation with 20 mg of bisacodyl provided good image quality of CTC images. Increasing the amount of laxatives did not improve image quality or tagging characteristics but was associated with a lower patient acceptance.
Keywords: bowel preparation • CT colonography • fecal tagging • image quality • patient acceptance
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
For labeling fecal material, three types of tagging agents are available: barium and nonionic and ionic iodinated contrast agents. Barium is traditionally used for solid fecal matter tagging and iodinated contrast agents are used to tag residual fluid. Furthermore, a variety of mild laxatives, for example, bisacodyl sodium or magnesium citrate [13, 14], can be added to reduce the amount of fecal matter in the colon. At present, no consensus exists about which contrast agent should be used and whether mild cathartics should be added to the tagging regimen, and if so, in what dose [19].
We hypothesized that a higher level of catharsis would improve image quality but reduce patient acceptance. Our objective was to determine the optimal dosage of laxatives for CTC with limited bowel preparation with regard to both image quality and patient acceptance. Therefore, we compared image quality and patient acceptance between four regimens with increasing levels of mild catharsis, using bisacodyl and magnesium citrate as laxative agents. Although some studies have compared image quality of CTC with and without laxatives [20, 21], to our knowledge, our study is the first that has investigated the effect of different amounts of mild laxatives in CTC.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Bowel Preparation
A low-fiber diet was prescribed for all patients for 2 days before the CTC
examination. No specific meal kit was used. Through a customized computer
program, patients were randomized into one of four groups using a
randomization block design (Table
1). Preparations contained solely orally administered bisacodyl or
bisacodyl in combination with magnesium citrate (LoSo Prep, E-Z-EM) as
laxative agents, and the extent of catharsis gradually increased from group 1
to group 4. We did not include a laxative-free regimen because such regimens
have been shown to result in insufficient image quality
[20,
21]. The most extensive
preparation (group 4) contained a combination of bisacodyl and magnesium
citrate in a dosage that is commercially available as a preparation kit for CT
colonography (LoSo Prep) and is considered a mild preparation
[13,
14]. Tagging consisted of
barium sulfate, 40% weight per volume (Tagitol V, E-Z-EM) and diatrizoate
meglumine with an iodine concentration of 200 mg/mL. Patients were not
informed of the content of the other regimens.
|
CT Colonography
The CTC examination was performed in the supine and prone positions on an
Mx8000 4-MDCT scanner (Philips Medical Systems) with the following parameters:
120 kV; collimation, 4 x 2.5 mm; rotation time, 0.75 second; pitch,
1.25; slice thickness, 3.2 mm; and reconstruction interval, 1.6 mm. Patients
with a circumference < 103 cm were scanned with 50 mAs (n = 15);
patients with a circumference > 103 cm, with 70 mAs (n = 25). A
total of 20 mg of butyl scopolamine bromide (Buscopan, Boehringer Ingelheim)
(n = 22) or, if contraindicated, 1 mg of glucagon hydrochloride
(GlucaGen, Novo Nordisk) (n = 17) was administered IV immediately
before scanning. In one patient neither was administered because of
contraindications. On average, 4 L of carbon dioxide (CO2) was
insufflated via an automatic insufflator (PROTOCO2L,
E-Z-EM) to achieve adequate distention. The time that patients spent in the CT
room was recorded with a stopwatch.
Conventional Colonoscopy
Colonoscopy was performed 4–30 days (mean, 17 days) after CTC. Bowel
preparation for colonoscopy consisted of 3 (n = 2), 4 (n =
36), or 6 L (n = 2) of polyethylene glycol–electrolyte solu
tion (Klean-Prep, Helsinn Birex Pharma ceuticals) administered the day before
the exam ination. Patients received midazolam (n = 28), fentanyl
(n = 20), or alfentanil (Rapifen, Astra Zeneca) (n = 6) on
request. Buscopan was ad ministered IV to all patients. The colonoscopy was
performed either by an experienced gastro enterologist or sur geon or by a
resident under the direct supervision of a staff member. During the
examination, a research nurse involved in our study was present.
Outcome Parameters
Subjective image quality—The CTC data were evaluated by a
research fellow in CTC who was blinded to all clinical data, including the
bowel preparation used. The reviewer had a prior experience of 150 nontagged
cathartic-prepared CTC examinations with colonoscopic verific ation. A primary
2D evaluation technique (axial views) was applied using multiplanar
reformatted (MPR) images and 3D endoluminal views for problem solving
(ViewForum 5.1, Philips Medical Systems). No electronic cleansing software was
applied.
The reviewer filled out a standardized questionnaire for every examination with regard to image quality. Table 2 displays the scoring system used by the reviewer. The image quality parameters of fecal tagging (Fig. 1A, 1B, 1C, 1D), luminal distention, amount of fecal residue (combined estimation for liquid and solid feces), and image readability were scored by the reviewer for the complete (supine and prone together) examination. This evaluation was con sidered the subjective image quality assessment on a per-patient basis.
|
|
|
|
|
The same image quality parameters were again evaluated on a per-segment basis. To this end, the colon was divided in six segments: cecum; ascending, transverse, descending, and sigmoid colon; and rectum. Fecal tagging and the amount of fecal material were evaluated only with the patient in the supine position because we presumed these parameters would not be influenced by a position change. However, because alteration of the position of the patient might substantially influence dis tention of different colonic segments [22, 23], luminal distention and segmental image readability were again evaluated with the patient in the prone position. Furthermore, to examine the effect of increasing catharsis on the consistency of fecal material, the reviewer scored the ratio between feces with a solid consistency and feces with a liquid consistency. This was performed on a persegment basis with the patient in the supine position. A 5-point scale was applied: 1, 0–20% solid; 2, 21–40% solid; 3, 41–60% solid; 4, 61–80% solid; and 5, 81–100% solid. If no stool was present in a segment, no evaluation of consistency was perform ed for that particular segment.
Numeric evaluation of degree and homogeneity of tagging—To evaluate the quality of fecal tagging, the mean and SD of attenuation of tagged material (in Hounsfield units) were measured on seven separate slices that were randomly selected by a computer program (Windows Excel 2002, Microsoft) for every patient. To this end, a resident in the fourth year of training, who was blinded as to the preparation used, placed a region of interest (ROI) of at least 60 mm2 in the area that contained the most fecal material on the randomly picked axial slices. The resident recorded whether the ROI was placed in liquid stool, stool adherent to the colon wall, or solid stool. The mean and SD of attenuation values within the ROIs were then calculated by our CTC software (ViewForum 5.1) (Fig. 2). If not enough fecal material was present to draw an ROI, the randomization program provided another slice for evaluation. The SD of attenuation values of tagged material is a measure of homogeneity. However, because the mean attenuation of tagging can vary considerably among patients, we considered the relative SD (SD / mean) to be a better measure of homogeneity. After all, by adapting the window width when viewing images, tagging with a larger mean attenuation value and a large SD may have a similar visual appearance to tagging with a smaller attenuation value and a proportionally smaller SD. Furthermore, separate assessments were performed for solid, adherent, and liquid feces.
|
Interpretation time—The time needed to interpret a complete examination for polyp detection, excluding report time and image quality evaluation, was recorded with a stopwatch by the reviewer. Because of the limited number of patients and the expected low prevalence of polyps in our cohort, performance characteristics were not part of our study design and are not discussed in this article.
Patient experience—Patients were asked to fill out a questionnaire on the day of the CTC examination with regard to the burden and side effects of the CTC preparation—that is, flatulence, abdominal pain, and diarrhea—as well as the discomfort caused by the intake of contrast material and laxative agents on a 5-point scale (Table 2). To understand if the everyday bowel habits of patients were of influence on image quality parameters, patients were asked to fill out their normal frequency of defecation: 1, more than one de fecation per day; 2, once per day; 3, once per 2 days; 4, once in 3 days; 5, less than one defecation per 3 days; and 6, less than one defecation per 5 days. On the day of the colonoscopy, patients filled out a questionnaire with regard to the burden associated with the bowel preparation for colon oscopy (Table 2).
Patient preference—Five weeks after colonoscopy, patients were sent a questionnaire in which they were asked which preparation had been most burdensome (CTC or conventional colonoscopy) and which examination (CTC or conventional colonoscopy) they would prefer in the future, and they were asked to indicate the most burdensome event of both examinations (Table 2). The patient preference and experience questionnaires were designed by the department of social medicine and had previously been used in two studies that were performed in our institution [7, 24].
Statistical Analysis
Subjective image quality—With regard to the per-patient
analysis, possible differences in image quality parameters between the four
preparations were assessed using ordinal regression analysis. In this
analysis, first the preparation with the highest regression coefficient was
determined; this was subsequently used as the reference group.
With regard to the per-segment analysis, ordinal regression analysis was applied using generalized estimating equations (GEE) to revise the data clustering and dependency [25]. This was done because more than one segment was obtained from each patient. The link function was set at log link, and an independent working correction matrix was used. Furthermore, associations between normal defecation frequency and subjective image quality parameters were tested with the chisquare test. If significant associations were present, additional ordinal regression analyses using GEE for revis ing the patient's normal defecation frequency were performed. Associations between the subjective image quality parameters (e.g., association be tween fecal tagging and the amount of residual feces) within each group were assessed using the chi-square test.
Numeric evaluation of degree and homogeneity of tagging—Because more than one measurement of Hounsfield units and SD was obtained from each patient, linear regression analysis was applied using GEE to revise the data clustering and dependency. Furthermore, ROIs were placed at random and were not evenly distributed among segments. This might have resulted in differences in measured homogeneity because of the physio logic variation of mean tagging attenuation within a given patient due to the normal dehydrating action of the colon. Therefore, the GEE analysis was adjusted to correct for this segmental dis tribution. Furthermore, because liquid material tends to be more homogeneously tagged than solid material, the GEE analysis was adjusted to correct for stool consistency (solid, adherent, or liquid feces). For each group, estimates of means with corresponding standard errors could be calculated from the results (intercept and slopes) obtained by the analysis.
Patient experience and preference—Differences in patient experience (burden) were analyzed using ordinal regression analysis and the preparation with the lowest regression coefficient (most burdensome) as the reference group. The chi-square test was used to test for significant differences in patients' indications of the most burdensome preparation and examination (CTC or conventional colonoscopy) between groups. Patient preference for either CTC or colonoscopy was tested using the chi-square test after the data were first dichotomized as preference for CTC versus preference for colonoscopy.
Interpretation time—Differences in interpretation times for the different preparations were tested for significance using the independent samples Student's t test. Statistical analyses were performed with SPSS version 12.0.2 for Windows (SPSS) and SAS version 8.02 for Windows (SAS Institute). The Proc Genmod command was used to apply the generalized estimating equations. A p value of < 0.05 was considered to indicate a statistically significant difference.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Image Quality Results
Subjective image quality—On a per-patient basis (n
= 40), no significant differences were found between groups with regard to
image quality. All examinations were evaluated as good (diagnostic for lesions
6 mm) or excellent (diagnostic with no limitations) image readability.
Only one examination (group 2) was of nondiagnostic value (because of poor
distention) and two examinations (group 3) were of moderate image readability
(because of moderate fecal tagging) (Table
3). Within each group, a higher degree of homogeneity was
significantly associated with better image readability (p <
0.006). For group 2, poorer distention was associated with decreased image
readability (p = 0.002).
|
|
|
|
|
|
Numeric evaluation of degree and homogeneity of tagging—The
degree of tagging, expressed as the mean attenuation of tagged material,
decreased when more laxatives were used
(Table 4); however, this
decrease was not statistically significant (all p 0.253). A
gradual decrease in mean SD of pixel values was also found from group 1 to
group 4. This decline, however, was not significant between groups (all
p 0.067). Furthermore, the relative SD (SD / mean) did not
differ significantly among groups (all p 0.157), indicating no
difference in homogeneity of tagging among groups.
|
Regardless of the preparation, the relative SD was significantly higher for solid feces (0.22) versus adherent (0.16) or liquid (0.14) feces, showing that tagging was less homogeneous for solid feces than for adherent stool (p = 0.008) or liquid feces (p = 0.001) (Table 5).
|
Interpretation time—The mean interpretation time for group 1
was 16 minutes, for group 2 was 24 minutes, for group 3 was 26 minutes, and
for group 4 was 21 minutes (all p 0.107).
Patient Experience
All patients experienced diarrhea except for three patients in group 2. The
burden of diarrhea was evaluated as none or mild by six patients (55%) in
group 1, six patients (60%) in group 2, four patients (40%) in group 3, and
one patient (11%) in group 4. Diarrhea was considered significantly more
burdensome in group 4 than in groups 1 (p = 0.042) and 2 (p
= 0.031) but not compared with group 3 (p = 0.179). With regard to
abdominal pain and flatulence, no significant differences were found among
groups. With regard to abdominal pain (n = 37), one patient in group
4 had severe abdominal pain, and two patients (groups 2 and 4) had moderate
abdominal pain. The remaining patients had little (n = 6) or no
(n = 28) pain. With regard to flatulence (n = 37), one
patient in group 4 experienced a severe burden, one patient in group 2 had a
moderate burden, and the other patients experienced little (n = 5) or
no (n = 30) burden.
The total burden of the bowel preparation was rated as none or mild by all patients (100%) in group 1, nine (90%) in group 2, eight (80%) in group 3, and five (55%) in group 4—a significantly higher total burden in group 4 than in groups 1 (p = 0.002) and 3 (p = 0.02). No significant difference (p = 0.082) was found between groups 2 and 4, most likely because one patient in group 2 rated the CTC bowel preparation as extremely burdensome. This patient stated, before the CTC preparation, an explicit preference for the colonoscopy preparation because in his opinion that preparation was simpler and shorter. Excluding this patient, a significantly lower overall burden (p = 0.04) was found for group 2 in comparison with group 4. Most patients experienced no or only a mild burden with regard to the intake of barium (32/34), diatrizoate meglumine (36/40), bisacodyl (34/37), or magnesium citrate (9/15), with no significant differences between the groups.
Patient Preferences
Except four patients (one in group 2 and three in group 3), all patients
found colonoscopy a more burdensome examination than CTC. If patients were
asked what examination they would prefer in the future, most patients (30/40)
indicated a preference for CTC, two patients were indifferent, and eight
patients preferred conventional colonoscopy
(Fig. 5). The reasons patients
preferred colonoscopy were: direct polypectomy (n = 4), shorter
preparation time (n = 2), false-positive CTC (n = 1), and no
particular reason (n = 1). No significant differences in pref erence
were observed among the groups.
|
All (100%) patients in group 1, nine (90%) patients in group 2, eight (80%) in group 3, and seven (78%) patients in group 4 preferred CTC bowel preparation over the polyethylene glycol preparation, with no significant differences among groups (p = 0.054). In all groups, most patients evaluated the bowel preparation for colonoscopy (n = 30) or colonoscopy itself (n = 5) as the most burdensome event, with no significant differences among the groups (Table 6).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In two previous studies, the same or a higher dose of magnesium citrate (without bisacodyl) as used in our group 4, was used; both reported significantly better image quality of CTC images compared with images made without the use of laxatives [20, 21]. Magnesium citrate liquefies residual feces and decreases the amount of remaining feces. This is associated with better readability [21] and can be expected to lead to increased homogeneity of tagged material. In concordance with these studies, we expected that image quality would gradually improve in our study and be best with group 4.
In our study, however, a higher level of catharsis resulted in neither
better image quality nor improvement of tagging characteristics (subjective or
numeric). Our data showed that residual feces in group 4 was more liquefied
compared with the other groups (all p 
0.004), but the relative
SD (SD / mean) did not decrease, indicating that homogeneity did not improve.
A decline, although not significant (p 0.253), of the mean
attenuation of the tagged material from group 1 to group 4 was observed that
we did not anticipate. The lower attenuation values in groups with more
catharsis might be explained because lower concentrations of contrast agent
were present as a result of dilution caused by the hyperosmotic effect of
magnesium citrate.
Regardless of the preparation, solid feces was less homogeneously tagged than liquid stool. However, we believe that homogeneity of solid feces was still satisfactory in all groups. Attenuation of solid feces was high (group 1, 1,030 H; group 2, 634 H; and group 3, 485 H) with relatively small SDs (215, 140, and 74, respectively) and nearly constant relative SDs (0.18, 0.21, and 0.23, respectively). As a result, the lowest attenuation value of tagged solid material was still well above soft-tissue density in all groups.
Most segments (88%) contained no or only small amounts of feces. Furthermore, no decrease in the amount of fecal material was observed using a higher dose of laxatives. This is in contrast with the results of Dachman et al. [21], who found no feces in only 65% of segments using magnesium citrate. That we used a combination of barium and diatrizoate meglumine for tagging purposes instead of only barium might explain this difference. Probably the laxative side effect of diatrizoate meglumine combined with (small amounts of) laxatives caused relatively clean colons with good homogeneity of tagged feces.
Increasing the amounts of laxatives did not improve subjective image quality but did increase the burden perceived by patients. In our study, most patients experienced diarrhea, regardless of the preparation used. These observations are not in line with the previously discussed study by Zalis et al. [20]. In that study, only one of 22 patients reported mild transient diarrhea after preparation with magnesium citrate. Two important differences between the studies may explain this discrepancy. First, we used ionic iodinated contrast material, which is known to have a strong osmotic effect, instead of a nonionic contrast agent. Second, in our study, all patients received bisacodyl, whereas in the Zalis et al. study, no bisacodyl was administered at all. It is likely that a combination of both factors led to some degree of diarrhea. Although diarrhea was present in all preparations, the burden of diarrhea and the overall burden of the preparation significantly increased with higher doses of laxatives. In fact, one could argue whether laxatives should be added at all to the preparation if ionic contrast agents are used in high concentrations. At the time of the writing of this article, we no longer add any form of laxatives to the preparation, and we believe this has not impaired the image quality of CTC.
An interval of 5 weeks was applied after colonoscopy to ask patients about their preference because the memory of adverse reactions may decrease over time [7, 8, 24]. The majority of patients (75%) preferred CTC as a colorectal examination and considered the polyethylene glycol–electrolyte solution for colonoscopy the most unpleasant factor (75%). In our institution, this bowel preparation is used by the gastroenterologists for all patients because it is safe to use and rigorously cleans the colon. Other bowel preparations for colonoscopy, such as an osmotic laxative ([sodium phosphates solution] Phospho-soda, Fleet), are considered more patient-friendly [26, 27]. Had these been used in our study, they might have caused a shift in patient preference toward colonoscopy. A disadvantage of Phospho-soda is that it can cause electrolyte imbalance and is contraindicated in patients with congestive heart failure or renal failure [28, 29].
Several limitations of our study must be considered. A limited cohort of 40 patients, randomized into four groups, participated in this study. Despite the relatively low numbers of patients per group, we did find a significantly higher patient burden with increasing laxative dosage. With regard to image quality, no trend of improvement in quality was observed with more catharsis. However, in group 2, segments contained significantly more feces compared with group 4. This was mainly attributed to the fact that two of the three patients in our study with a relatively low defecation frequency in normal life were placed in group 2. Despite more feces present in the colon, image readability of the examinations was not significantly affected. Although our data are limited, this might suggest that with adequate tagging, patients with bowel habits of one defecation in 3–5 days can be prepared with minimal amounts of laxatives without affecting image quality. Furthermore, all patients without diarrhea (n = 3) were in group 2. Patients were not asked about compliance with the preparation. Because noncompliance could have contributed to more residual feces in group 2 but could also provide a possible explanation for the moderate image quality in two patients in group 3, we considered that a limitation of our study.
Furthermore, only one observer subjectively evaluated all data on image
quality. It is possible that another observer would rate the data differently.
Another possible qualifier is that we did not include a full catharsis regimen
as a reference standard to which the studied preparations were compared.
However, we believe that for this study a clinically relevant scoring system
was constructed with regard to image quality of the examinations: using
classifications as not diagnostic, diagnostic for all lesions, diagnostic for
lesions 6 mm, or only diagnostic for lesions 10 mm. The experience
of the reviewer before our study consisted of 150 cathartic CTC examinations,
and consequently image quality of these examinations served as a reference
standard for the reviewer. With regard to patient experience, the bowel
preparation used for colonoscopy in our study was a full cathartic bowel
preparation using a polyethylene glycol–electrolyte solution that can
also be used for CTC. Most patients in all groups indicated the colonoscopy
preparation as more burdensome when compared with the CTC bowel
preparation.
We used an iodine-based contrast medium for fecal tagging in addition to barium. Some investigators prefer using only barium because of the side effects and possible adverse reactions to iodine-based contrast agents [10]. However, we believe that a combination of both contrast agents resulted in adequate tagging of both solid and liquid feces as stated in other studies [2, 29]. Although the laxative side effect of an iodine-based contrast agent will probably increase the patient burden in some way, an advantage is that CTC images are easier to interpret and a possible cleansing algorithm might be more effective [20, 30]. Finally, in our study we focused on image quality. Polyp conspicuity was not investigated. So far, one larger study without catharsis has reported an excellent sensitivity of 90% and specificity of 92% for patients with polyps of any size [18]. Although further research is warranted, these results underscore our findings that good image quality can be obtained with small amounts of laxatives.
We conclude that CTC with limited bowel preparation, using barium and ionic iodinated contrast agents for fecal tagging, requires only minimal doses of laxatives—in our study only 20 mg of bisacodyl—to obtain good image quality and minimize patient burden. This is important because a mild bowel preparation will undoubtedly increase patient willingness to participate in a screening program.
Acknowledgments
We thank Karin Horsthuis for critical review of this manuscript and Henk W.
Venema for his help with the data analysis and for critical review of this
manuscript.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
