Comparison of 3% sorbitol vs psyllium fibre as oral contrast agents in MR enterography

Abstract

Objective:

To compare the degree of small bowel distension achieved by 3% sorbitol, a high osmolarity solution, and a psyllium-based bulk fibre as oral contrast agents (OCAs) in MR enterography (MRE).

Methods:

This retrospective study was approved by our institutional review board. A total of 45 consecutive normal MRE examinations (sorbitol, n = 20; psyllium, n = 25) were reviewed. The patients received either 1.5 l of 3% sorbitol or 2 l of 1.6 g kg⁻¹ psyllium prior to imaging. Quantitative small bowel distension measurements were taken in five segments: proximal jejunum, distal jejunum, proximal ileum, distal ileum and terminal ileum by two independent radiologists. Distension in these five segments was also qualitatively graded from 0 (very poor) to 4 (excellent) by two additional independent radiologists. Statistical analysis comparing the groups and assessing agreement included intraclass coefficients, Student's t-test and Mann–Whitney U-test.

Results:

Small bowel distension was not significantly different in any of the five small bowel segments between the use of sorbitol and psyllium as OCAs in both the qualitative (p = 0.338–0.908) and quantitative assessments (p = 0.083–0.856). The mean bowel distension achieved was 20.1 ± 2.2 mm for sorbitol and 19.8 ± 2.5 mm for psyllium (p = 0.722). Visualization of the ileum was good or excellent in 65% of the examinations in both groups.

Conclusion:

Sorbitol and psyllium are not significantly different at distending the small bowel and both may be used as OCAs for MRE studies.

Advances in knowledge:

This is the first study to directly compare the degree of distension in MRE between these two common, readily available and inexpensive OCAs.

MR enterography (MRE) is an imaging technique used to evaluate the small bowel for diagnoses that include inflammatory bowel disease and neoplasms. Sufficiently distending the small bowel with an oral contrast agent (OCA) is critical, as collapsed loops may hide or mimic abnormalities.

The OCAs employed in MRE can be classified as positive, negative or biphasic contrast agents according to their signal properties on T₁ and T₂ weighted imaging. Biphasic contrast agents typically provide low signal intensity on T₁ weighted imaging and high signal intensity on T₂ weighted imaging (T2W), and are most commonly used today because of their favourable signal characteristics for detecting bowel pathology. Water is a biphasic contrast agent, but its rapid physiological absorption from the small bowel necessitates the use of additives. These additives are often highly osmolar agents or non-osmotic agents that form hydrogels that trap water molecules in the bowel lumen. Combinations of highly osmolar agents and non-osmotic agents are also frequently used, as in the case of sorbitol and locust bean gum.^1,2

Ideally, the chosen contrast agent should (i) provide adequate distension of the entire small bowel, (ii) provide uniform and homogenous opacification of the lumen, (iii) provide good delineation of the bowel wall, (iv) be below cost and easy availability, (v) have easy storage and (vi) have no serious adverse side effects.³

Currently, VoLumen® (E-Z-EM Inc., Westbury, NY), a biphasic OCA comprising 0.1% w/v ultra-low-dose barium with sorbitol, simethicone and natural gums, is commonly used for MRE examinations in the USA. However, VoLumen is not easily available for sale in Canada, the UK and Europe. Instead, our institution (St Michael's Hospital, Toronto, ON) used commercially available Metamucil® (Proctor & Gamble, Cincinnati, OH), a non-osmotic bulk fibre laxative containing psyllium seed husk, as a biphasic OCA for routine MRE studies from 5 January 2010 to 20 January 2013. On 21 January 2013, this was switched to the highly osmolar agent, 3% sorbitol, for all routine cases. This switch was prompted after a patient expressed concern about whether psyllium could result in small bowel obstruction.

The purpose of our study was to retrospectively compare psyllium and 3% sorbitol as OCAs for MRE using the outcome measure of small bowel distension. On the basis of past literature, we hypothesized that sorbitol would provide better distension than psyllium.⁴

METHODS AND MATERIALS

Study population

This retrospective study was approved by our institutional review board with a waiver of written informed consent. Adult patients who underwent MRE with either 3% sorbitol or psyllium as an OCA with a normal MRE result were included. All patients who had been referred to our department for MRE for assessment of suspected bowel pathology from 5 January 2010 to 26 June 2013 were reviewed.

We performed an a priori sample size estimation (power, 0.80; α, 0.05) by using data from previous studies that examined bowel diameter and had used either sorbitol⁵ or psyllium.⁶ In order to detect a 20% difference in distension between the OCAs, while assuming a standard deviation of 3.5, we calculated that a sample size of 20 in each cohort would be sufficient.

20 patients who received 3% sorbitol after 21 January 2013 and 25 patients who received psyllium before 21 January 2013 were included in this study. A total of 78 consecutive patients who received 3% sorbitol after 21 January 2013 and 66 consecutive patients who received psyllium before 21 January 2013 were reviewed to achieve this cohort size prior to applying exclusion criteria. Exclusion criteria were (1) patients with documented small bowel pathology or previous bowel surgery on MRE (sorbitol, 55; psyllium, 38), (2) an inability to drink the full volume of the OCA (sorbitol, 2; psyllium, 1), (3) patients who vomited before imaging was completed (sorbitol, 1; psyllium, 1) and (4) examinations with severe motion artefacts (psyllium, 1). Patients with extraenteric pathology but normal small bowel findings were not excluded.

Surgical, endoscopic, pathological or clinical follow-up data were reviewed where available for all 45 patients to confirm that the normal MRE examination corresponded to no small bowel disease.

Image acquisition

All patients were instructed to abstain from solid food after 17h00 the night before the examination. Patients receiving sorbitol were instructed to drink 500 ml of 3% sorbitol 45, 30 and 15 min prior to the start of image acquisition. Patients receiving psyllium were instructed to drink 500 ml of 1.6 g kg⁻¹ psyllium in 2.0 l of water every 30 min for 2 h prior to the start of image acquisition, with 10 mg of metoclopramide given per os 2 h before imaging and an additional 10 mg given 1 h before imaging.

Images were obtained with a 1.5-T Philips Achieva® MR scanner (Philips Healthcare, Best, Netherlands) or a 3.0-T Magnetom® Skyra, MR scanner (Siemens Medical Solutions, Erlangen, Germany). The same MRE protocol was used for both cohorts. A multichannel phased array surface coil was used on both scanners. All studies included intravenous (i.v.) administration of gadobenate dimeglumine (MultiHance®; Bracco, Milan, Italy), a gadolinium-based contrast agent, at a dose of 0.1 mmol kg⁻¹ and flow rate of 2 ml s⁻¹ via a power injector and patients were placed head first in the scanner in a prone position. All studies also included i.v. administration of butylscopolamine as an antiperistaltic agent by a radiologist. 20 mg intramuscular and 20 mg i.v. were administered at the commencement of scanning to the sorbitol cohort and 20 mg i.v. only to the psyllium fibre cohort. This difference in butylscopolamine doses between cohorts was because of our clinical experience with sorbitol that 20 mg was insufficient to arrest bowel motion. Although multiple sequences were acquired, only the coronal and axial T2W fat-suppressed (FS) single shot fast-spin echo (SSFSE) breath-hold (BH) imaging was used for the study analysis in all patients. The standard imaging parameters for two-dimensional axial/coronal images included a repetition time of 455/1030 ms, echo time of 80/80 ms, matrix size of approximately 320 × 214/268 × 236, slice thickness of 7/5 mm and BH acquisitions of approximately 20 s. Parallel imaging with an acceleration factor of two was also utilized.

Image analysis

The degree of small bowel distension was assessed both quantitatively and qualitatively.

Two board-certified radiologists with subspeciality training and 13 combined years' of practice in abdominal MRI independently obtained quantitative bowel distension measurements for all patients. Measurements were taken in five anatomic locations: proximal jejunum, distal jejunum, proximal ileum, distal ileum and terminal ileum. The reviewers were instructed to use the entire image set of the coronal and axial T₂ FS sequences (Figures 1 and 2) to identify the maximally distended bowel loop for each of the five segments. This bowel segment was then measured perpendicular to the long axis of the lumen from the outer margins of the bowel wall. Thus, ten measurements were taken for each patient between the two readers.

Figure 1.

Coronal T₂ fat-saturated image after small bowel distension using 1.5 l of 3% sorbitol solution. This case was given an overall distension score of 3 (good) by both radiologists in the qualitative analysis. This score is not a reflection of the single image provided but of the entire axial and coronal series that was given to the radiologists for grading.

Figure 2.

Coronal T₂ fat-saturated image after small bowel distension using 2.0 l of 1.6 g kg⁻¹ psyllium in water solution. This case was given an overall distension score of 3 (good) by both radiologists in the qualitative analysis. This score is not a reflection of the single image provided but of the entire axial and coronal series that was given to the radiologists for grading.

A qualitative assessment was also conducted in order to evaluate the overall distension of a segment. For this assessment, all MR examinations were independently analysed by two separate board-certified radiologists with 5 years' of combined practice in abdominal MRI. These radiologists were not a part of the quantitative assessment. The radiologists graded the distension of each bowel segment on a five-point scale as validated in previous studies: 0, very poor distension; 1, poor distension; 2, fair distension; 3, good distension; and 4, excellent distension.^1,2,7 In addition, the overall distension of the entire small bowel was graded on the same five-point scale. The reviewers were instructed to use the entire image set of the coronal and axial T₂ SSFSE sequences (Figures 1 and 2) as part of their assessment.

The images for all four readers were reviewed on a picture archiving and communication system workstation. All four reviewers were blinded to the clinical diagnosis and patient demographics, as well as the other readers' assessments.

Statistical analysis

Interobserver variability and agreement for distension measurements was assessed using the intraclass correlation coefficient (ICC). Two-way mixed model ICCs with measures of absolute agreement were interpreted as follows: 0–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial and >0.80, almost perfect.⁸

Substantial agreement was required in order to take the mean value of the two readers for comparative analyses. The normality of the quantitative and qualitative data was assessed using the Shapiro–Wilk test before determining which statistical test to use. The mean bowel distension for patients receiving psyllium and sorbitol was compared using Student's independent samples t-test on a segment-by-segment basis. The mean qualitative measurements of distension for each segment and overall distension were compared at a 95% confidence level by using the Mann–Whitney U-test. Differences were considered significant if p < 0.05. All statistical analyses were performed using SPSS® v. 22.0 software (SPSS Inc., Chicago, IL).

RESULTS

The sorbitol cohort (n = 20) consisted of 17 females and 3 males, whereas the psyllium cohort (n = 25) consisted of 19 females and 6 males. The mean age of all patients was 40.7 years (range, 18–85 years) and was not significantly different between the cohorts (p = 0.606). Interobserver agreement was substantial for all analysed bowel segments (ICC >0.61). Both the qualitative and quantitative data were normally distributed as per the Shapiro–Wilk test (p > 0.05).

To confirm that the cohorts consisted of normal small bowel patients, we found surgical, endoscopic, pathological or clinical follow-up data in 38/45 patients. In 37/38 patients, the follow-up data showed normal small bowel by surgical, endoscopic, pathological or clinical assessment. In the remaining 1/38 patients, there was mention of stable Crohn's symptoms (with no histopathology confirmation of Crohn's) in an ambulatory consult 3 months earlier than the MRE examination. In this patient, we performed a consensus read of the MRE by four radiologists following small bowel distention data collection. This determined that the small bowel was normal on the MRE. For the remaining 7/45 patients, a consensus read of the MRE was performed by all 4 radiologists who assessed images following data collection and determined that the small bowel was normal on all 7 MRE examinations.

Qualitative analysis

The non-parametric qualitative data were compared using the Mann–Whitney U-test rather than the independent samples Student's t-test. The qualitative assessment, which describes the overall distension of the small bowel segment rather than maximal bowel distension, did not show a statistically significant difference between the two cohorts (p = 0.338–0.908; Table 1 and Figure 3). The qualitative grading of distension yielded a mean score of all segments and overall distension grading of 2.3 ± 0.7 for the sorbitol group and 2.3 ± 0.5 for the psyllium group (p = 0.855). The distension of the ileum was graded as “good” or “excellent” in 65% of all examinations in both the sorbitol and psyllium cohorts.

Table 1.

Qualitative results: bowel distension graded from 0 (very poor) to 4 (excellent)

Bowel segment	ICC	Sorbitol		Psyllium		p-value
		Median	IQR	Median	IQR
Proximal jejunum	0.799	1.00	0.50	1.00	1.25	0.484
Distal jejunum	0.771	1.75	1.75	2.00	1.50	0.908
Proximal ileum	0.646	2.75	1.50	3.00	0.50	0.833
Distal ileum	0.675	3.00	1.00	3.00	0.50	0.338
Terminal ileum	0.711	2.50	1.00	3.00	1.75	0.389
Overall	0.853	2.75	1.00	2.50	1.00	0.530

ICC, intraclass correlation coefficient; IQR, interquartile range.

ICC two-way mixed model with measures of absolute agreement and 95% confidence intervals. p-values were calculated using the Mann–Whitney U-test. Equal variances were not assumed.

Figure 3.

Qualitative assessment of bowel distension.

Quantitative analysis

The mean bowel distension across all five bowel segments for the sorbitol and psyllium cohorts, respectively, was 20.1 ± 2.2 mm [range, 13.0–27.0 mm; confidence interval (CI), 19.1–21.1] and 19.8 ± 2.5 mm (range, 11.0–33.0 mm; CI, 18.8–20.8). This difference was not statistically significant (p = 0.722). The mean bowel distension for each of the five individual small bowel segments was also not statistically significantly different (p = 0.083–0.856) between the two cohorts (Table 2 and Figure 4).

Table 2.

Quantitative results: bowel distension measurements in millimetres

Bowel segment	ICC	Sorbitol (mm ± SD)	Psyllium (mm ± SD)	p-value
Proximal jejunum	0.774	18.5 ± 3.3	19.3 ± 4.2	0.492
Distal jejunum	0.737	22.0 ± 3.3	22.6 ± 4.4	0.587
Proximal ileum	0.827	21.8 ± 3.1	20.7 ± 3.5	0.287
Distal ileum	0.747	21.8 ± 2.9	20.3 ± 2.4	0.083
Terminal ileum	0.641	16.4 ± 2.5	16.2 ± 2.5	0.856

ICC, intraclass correlation coefficient; SD, standard deviation.

ICC two-way mixed model with measures of absolute agreement and 95% confidence intervals. p-values were calculated using statistical one-way analyses. Equal variances were not assumed.

Figure 4.

Quantitative assessment of bowel distension. Error bars = standard error of the mean.

DISCUSSION

Prior to cross-sectional imaging, radiological investigation of the small bowel relied upon barium small bowel follow-through studies. These studies provided visualization of the bowel lumen, but their interpretation was difficult when bowel loops overlapped, and they did not provide extraluminal visualization. Today, CT and MRI are commonly used for detailed small bowel assessment, including the lumen, mucosal lining and wall, while allowing for extraluminal visualization. The past decade has brought technological improvements in MRI that have largely addressed issues such as lengthy examination time and bowel motion-related artefacts. As a result, MRE is now becoming the method of choice for assessment of small bowel for inflammatory disease.^3,9 Importantly, MRE provides a diagnostic accuracy that is comparable to CT, colonoscopy and surgical pathology.^10–13 Adequate bowel distension of the small bowel is essential for this diagnostic accuracy and hence the use of OCAs.

There have been many OCAs proposed for use in MRE. Water is a biphasic contrast agent that has been used in the past;^4,6,14 however, it is rapidly absorbed and does not provide adequate distension of distal ileal loops.^6,15,16 Various compounds have been added to water to increase its retention in the lumen. These include highly osmolar agents such as polyethylene glycol,^14,16,17 mannitol^1,2,4,6 and sorbitol,^1,2,18 as well as non-osmotic agents that can form hydrogels, such as locust bean gum,^1,2,4,6 methylcellulose¹⁶ and psyllium seed husk.^4,6,19

There are currently two approaches to oral contrast administration: MR enteroclysis and MRE.³ In MR enteroclysis, a nasojejunal tube is used to administer contrast in a controlled fashion directly into the small bowel. Placement of the nasojejunal tube requires fluoroscopy that exposes the patient to radiation and may require the patient to be moved between different sections of a medical imaging department. MRE is a less invasive procedure in that the patient simply drinks 1–2 l of contrast medium. For this reason, patients prefer to undergo MRE over MR enteroclysis.²⁰ However, MRE is less consistent in distending the small bowel than enteroclysis and is more reliant on using an OCA with effective distension properties.³

Despite the increasing use of MRE, there is still no universal consensus on the preferred OCA. Some commonly used highly osmolar contrast agents include polyethylene glycol, mannitol, sorbitol and barium sulphate. A mixture of highly osmolar agents and non-osmotic agents is also often used, as in the case of sorbitol and locust bean gum. VoLumen is becoming the mainstay contrast agent used in the USA and also contains a mixture of agents, including sorbitol and barium sulphate in addition to the non-osmotic agent, xanthene gum. VoLumen is not readily commercially available in many countries outside of the USA.

Our institution has used psyllium, a non-osmotic biphasic OCA containing ispaghula or psyllium seed husk, for routine MRE examinations for 3 years until 2013, before using 3% sorbitol. The active portion of psyllium seed husk is a highly branched arabinoxylan consisting of copolymers of two pentose sugars.²¹ Unlike the arabinoxylans found in cereal grains, psyllium seed husk is not extensively fermented by colonic bacteria.²¹ Psyllium seed husk has a laxative effect in low doses but, interestingly, has an antidiarrhoeal effect at higher doses.²²

The use of psyllium in small bowel cross-sectional imaging was first demonstrated by Vogel et al²³ in 1996, where psyllium was mixed with barium solution for CT enterography (CTE). Psyllium seed husk was later implemented in MRE in 2001 by Patak et al.¹⁹ Subsequent MRE and CTE studies provided conflicting results where two MRE studies found psyllium to have poor small bowel distension,^4,6 whereas two CTE studies argued that it provided good distension.^24,25 The benefits of using psyllium in MRE are that is it is biphasic, inexpensive, commercially available and easily stored, and has a favourable taste. Side effects of psyllium include nausea, diarrhoea and bloating. In addition, there have been two reported cases of small bowel obstruction using psyllium.^24,25 The initial motivator to switch MRE OCAs at our institution occurred after a patient with a history of bowel obstruction expressed a desire not to use psyllium after reading about its potential risks.

Sorbitol is a sugar alcohol that is often used as an alternative sweetener in food products. It is slowly metabolized and absorbed in the small bowel and the majority is able to reach the large bowel where it is broken down by bacterial flora into lactulose and eventually short chain fatty acids and gases.²⁶ Unlike mannitol, a structurally similar hexose sugar that is also used as an OCA, it does not form explosive gases in the colon.²⁷ Owing to its hyperosmolarity, sorbitol is able to draw water molecules into the bowel and exert a laxative effect in high doses.

Several comparative studies have shown sorbitol to provide excellent distension, both on its own and when mixed with other additives, including barium sulphate and locust bean gum.^1,18 There have also been studies attempting to optimize bowel distension using sorbitol without decreasing patient tolerability with different volumes and concentrations.² Benefits of sorbitol are that it has been shown to provide good distension and is biphasic, inexpensive, easy to store and commercially available in many countries. While sorbitol shares many of the same side effects as psyllium, there are no reported cases of bowel obstruction with its use.

On the basis of this previous literature and our anecdotal experience, we hypothesized that sorbitol would be more effective at bowel distension than psyllium.^1,4 However, our results indicate that there is no statistically significant difference between psyllium and sorbitol in both the quantitative (p = 0.338–0.908) and qualitative (p = 0.083–0.856) assessments of bowel distension. A similar profile of distension (Figures 3 and 4) was found for both OCAs in the five small bowel segments investigated with distension of the ileum favoured over the jejunum. Maximum bowel distension was achieved in our study in the distal ileum by the qualitative assessment. Distension of the ileum is more desirable than the jejunum because the ileum is most often the site of Crohn's disease.³ Agreement between the blinded independent observers was substantial for all bowel segments.

The ability of psyllium to provide comparable distension to sorbitol may be explained by the dose of psyllium used in our study. Previously, the largest reported dose of psyllium used for small bowel imaging was 0.8 g kg⁻¹,^28,29 which is half the dose used in our study. The use of this high dose of psyllium may carry the risk of bowel obstruction in patients with pre-existing small bowel strictures.³⁰ This is demonstrated by two cases of obstruction in a combined cohort of 68 patients with Crohn's disease undergoing CTE with psyllium.^24,25,30 More studies are required before the risk of obstruction in MRE can be established. Nonetheless, there have been no cases of small bowel obstruction using sorbitol, and it is able to provide the same degree of distension. As a result of this, we recommend sorbitol over Metamucil for use in MRE examinations.

Our study had limitations. First, our sample size was small and study design was retrospective. Second, we used normal patients who are not fully representative of the population undergoing routine MRE examinations at our institution, and distension values may differ with patients having small bowel pathology. Third, by choosing to examine all coronal and axial images instead of a single coronal image for the bowel distension measurements, it was challenging to achieve high inter-rater reliability because of the difficulty in defining landmarks for which to segregate the five segments of small bowel chosen. Fourth, the two cohorts had a difference in volume of OCA administered: 2.0 l of psyllium was administered, whereas only 1.5 l of 3% sorbitol was administered. This limitation arose from selecting the 1.5-l sorbitol regimen after examining “best practices” in other institutions, since we wished to create a diagnostic protocol that would result in high patient compliance as quickly as possible. Future studies in this area could potentially add to this area by assessing different volumes of the same OCA and/or the exact same volumes of different OCAs. In addition, assessment of the patients' perspectives of different OCA regimens may be useful and could not be assessed in this study. Fifth, the two cohorts had one additional difference in technique: the doses of butylscopolamine used to arrest bowel motion. This difference in dose between cohorts is unlikely to have made a difference in distension, and, anecdotally, the readers observed no difference in bowel motion artefact between cohorts.

In conclusion, our study demonstrated that there is no significant difference in bowel distension between cohorts of patients receiving psyllium and 3% sorbitol as OCAs for MRE. Thus, we recommend the use of 3% sorbitol over psyllium because it likely has a decreased risk of small bowel obstruction.

FUNDING

This article was funded by the University of Toronto and Department of Medical Imaging, St Michael's Hospital, Toronto, Canada.