Digestive Diseases and Sciences

, Volume 58, Issue 4, pp 926–935

A Validated Bowel-Preparation Tolerability Questionnaire and Assessment of Three Commonly Used Bowel-Cleansing Agents

Authors

    • Centre for Inflammatory Bowel DiseasesFremantle Hospital
    • The School of Medicine and Pharmacology, Fremantle HospitalUniversity of Western Australia
  • R. P. Willert
    • Centre for Inflammatory Bowel DiseasesFremantle Hospital
  • K. Murray
    • Statistical Consulting Group, School of Mathematics and StatisticsUniversity of Western Australia
Original Article

DOI: 10.1007/s10620-012-2449-0

Cite this article as:
Lawrance, I.C., Willert, R.P. & Murray, K. Dig Dis Sci (2013) 58: 926. doi:10.1007/s10620-012-2449-0

Abstract

Background and Study Aims

Bowel-cleansing studies are frequently underpowered, poorly designed, and with subjective assessments. Consensus on tolerability of the bowel-cleansing agents is thus lacking. This study developed and validated a bowel-preparation tolerability questionnaire and used it to assess the tolerability of three bowel-cleansing agents, sodium phosphate (NaP), polyethylene glycol (PEG), and sodium picosulphate (Pico), in a prospective randomized single-blinded trial of ambulatory patients.

Patients and Methods

The bowel-preparation tolerability questionnaire was validated in 125 consecutive patients and then bowel-preparation agent tolerability was assessed in 634 patients in a prospective randomized single-blinded trial.

Results

The questionnaire’s internal consistency was satisfactory with good to excellent “test–retest” reliability for aggregate tolerability and visual analogue scores. Validity assessment confirmed it as reliable and accurate. Of 634 patients, 97.8 % took >75 % of the allocated preparation and 98.9 % completed the questionnaire. Overall, Pico was better tolerated than PEG (p < 0.001) and NaP (p < 0.001). NaP was better tolerated than PEG (p < 0.001). Regardless of the bowel-preparation agent used, males tolerated them better than females (p = 0.009) as did patients having their procedure in the AM. Older patients, however, tolerated all preparations better than younger patients (p = 0.006).

Conclusions

This study used the first validated bowel-preparation tolerability questionnaire and identified that age, sex, and procedure time all impacted tolerability. Overall, Pico was best tolerated, but PEG’s tolerability in patients ≥60 years was equal to that of Pico and NaP, suggesting that PEG can be recommended for older patients to avoid the electrolyte disturbances associated with the osmotic preparations.

Keywords

FleetPicoprepColonlytelyColonoscopyBowel preparationAssessment

Introduction

Colonoscopy is common, but for adequate colonic assessment, bowel-cleansing preparations are required. Poor preparation reduces the detection rate of colonic lesions and results in 10–20 % of colonoscopies failing [1]. This can lead to repeated examinations, longer waiting lists, and increased healthcare costs. Patient tolerability of the bowel-preparation agents is central to why patients do not take the bowel preparation as prescribed, resulting in poor bowel preparation and suboptimal visualization of the colonic mucosa.

Over 80 trials have investigated colon-cleansing preparations, however, almost all have small patient numbers with poor study design and ill-defined outcome measures [2]. Only 14 prospective randomized studies have included over 100 patients in each study arm [315] and only our recent study [14] has utilized a validated bowel-preparation scoring system. None of the studies to date have used a validated patient tolerability questionnaire resulting in the several meta-analyses being unable to achieve a consensus as to which bowel-preparation agent is best tolerated [1618].

Both bowel preparation tolerability and safety are important. The osmotic agents are associated with electrolyte disturbances, so that they need to be used with care [2, 1618]. Beazely identified 11 patients who died as a direct consequence of NaP, 26 suffering renal failure, and six requiring dialysis. Unfortunately, the total number of patients receiving NaP was unknown, thus the relative risk was also unknown [16]. Despite this, patients at greater risk of electrolyte disorders should not routinely receive NaP and it has been withdrawn from the US market. Risks are also associated with the tablet form of NaP, as the FDA warns that the tablets should be used with caution in patients over 55 years. The use of NaP is also associated with colonic mucosal aphthous ulceration [14, 16, 19, 20] and the finding that both NaP and Pico have a significantly increased rate of mucosal inflammation compared to PEG [14] may impact their use in patients with inflammatory bowel disease (IBD) where postcolonoscopic flaring could be provoked by these changes [21].

It has been suggested that PEG is not well tolerated, primarily due to the volume required [10] as low-volume PEG solutions have better tolerability with equivalent bowel cleansing [22, 23] and splitting the PEG dose may also improve its tolerability [15, 24]. Our study was designed to clear some of the confusion surrounding the tolerability of three of the most commonly used bowel-preparation agents used for colonoscopy using in a prospective randomized study.

Methods

The bowel-preparation tolerability questionnaire was developed and validated in 125 consecutive patients. This was followed by a prospective randomized single-blinded trial comparing sodium phosphate (NaP), polyethylene glycol (PEG), and sodium picosulphate (Pico). This patient cohort was the same as used in our previous study investigating mucosa cleaning efficacy [14]. In short, 634 patients were randomized to receive one of the agents. All patients were consecutive community-based ambulatory patients at Kaleeya Hospital, Western Australia. Excluded patients were confirmed or suspected IBD, current non-steroidal anti-inflammatory medication (excluding low-dose aspirin), heart failure (NYHA > 2), renal failure (GFR < 30) [2], over 75 years [2]. Patients were randomized (Generator Pro 1.69, Segobit Software) as previously described [14], to received PEG, NaP, or Pico. Patient demographics were recorded.

Aims

Primary Aim

The primary aim was to determine if there were any differences in the overall tolerability of the three agents determined by the aggregate tolerability score (ATS - sum of all individual components) and visual analogue score (VAS).

Secondary Aims

The secondary aims were to identify:

  1. 1.

    if bowel-preparation tolerability is affected by sex, age, and time of colonoscopy,

     
  2. 2.

    if the individual components of the ATS are affected by sex, age, and time of colonoscopy,

     
  3. 3.

    if the patients missed work due to the bowel preparation used,

     
  4. 4.

    and if the patients were prepared to take the same bowel preparation again.

     

Ethical Approval and Patient consent

Ethical approval was obtained from the Southern Metropolitan Area Health Service Human Research Ethics Committee (SMAHSRC). As all three bowel-preparation agents were considered as standard agents for bowel cleansing, patient consent was not deemed necessary. Written consent was also considered unnecessary prior to patients completing the tolerability questionnaire. All patients were free to refuse to complete the questionnaire, but none of the patients in this consecutive cohort did so.

Patient Preparation Instructions

All patients were contacted by phone by a trained endoscopy nurse to discuss diet and bowel-preparation instructions prior to commencing the preparation. In line with the routine practice at our institution, all patients commenced a minimal residue diet 2 days prior to the colonoscopic procedure and were restricted to clear fluids the day prior to colonoscopy. The bowel-preparation agents NaP and Pico were taken as a split course separated by 12 h. Instructions on taking the individual bowel preparation were as previously published [14].

Validation of the Preparation the Bowel-Preparation Tolerability Questionnaire

On the day of colonoscopy, prior to endoscopist review and colonoscopy, all patients were asked to complete the questionnaire by the admitting nurse without help (Fig. 1). Just prior to discharge, the patients were again requested to complete a second questionnaire without reference to the first. Thirty-six patients completed a second questionnaire that was identical to the pre-colonoscopic questionnaire and 89 completed ones that had the question order altered.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2449-0/MediaObjects/10620_2012_2449_Fig1_HTML.gif
Fig. 1

Patient tolerability questionnaire

Preparation Tolerability Assessment

An additional 634 consecutive patients were assessed. On the day of colonoscopy, prior to endoscopist review and colonoscopy, patients assessed the colon-cleansing agent tolerability via a nurse-administered questionnaire as described above.

Statistics

The patient tolerability questionnaire was validated by:

Internal Consistency

The extent to which items within a dimension correlate with each other by calculation of Cronbach’s alpha, a method based on correlations between items.

Test–Retest Reliability

Bland–Altman plots assessed the distribution of differences in continuous and aggregated scores with the differences plotted against the average. The overall mean and variance of differences was calculated, with 95 % tolerance intervals constructed around the mean by assuming a normal distribution [25]. The test and retest scores were assumed to be from the same distribution when the differences have a mean of zero, and 95 % of the differences lie within the 95 % tolerance limits. Intraclass correlation coefficients (ICCs) were calculated to give an additional quantification to this reliability aspect. For the individual component scores, Cohen’s weighted kappa was used to assess how well the scores agreed.

Validity

Kruskal–Wallis analyses tested whether the individual tolerability scores (ITS) related to the VAS score. We assessed how a patient’s willingness to use the same bowel preparation again correlated with the ITS and the ATS using Chi-square tests.

Tolerability Analysis

Tolerability analysis used ordinal logistic regression for the nine individual aspects. Potential predictors included sex, age group, bowel preparation, and time of procedure, together with two-way interactions. Results are presented as odds ratios. Tests for proportional odds used the score test with assumptions satisfied. Due to the distribution of two aspects (vomiting and headache), an additional grouping of the response variable was made for analysis. This grouping combined ratings of 1/2 and 3/4 to generate a 3-point instead of a 5-point scale. The overall ATS and visual analogue score (VAS) were analyzed using general linear models with square root and arcsine square root transformations, respectively. Whether patients would take the bowel preparation again, and if patients missed work were analyzed using logistic regression.

The number of patients in each treatment group had previously been calculated to determine differences in acute mucosal inflammation and ulceration induced by the three different bowel-preparation agents. Due to the differences in assumed ulceration rates, different patient numbers were required in each patient group [14]. With 150 patients in each group and a type I error of 0.05 at 80 % power significance, however, would detect with a minimum of a 15 % difference between the tolerability of the bowel-preparation agents.

Results

Validation of the Questionnaire

Internal Consistency

The bowel tolerability questionnaire was validated with 125 consecutive patients. The internal consistency of the questionnaire assessed by the Cronbach’s alpha value was 0.78 and deemed as satisfactory [25].

Test–Retest Reliability

The test–retest reliability for each of the nine individual components, along with the aggregate of these components (aggregate tolerability score) and the VAS, are presented in Table 1. The ratings ranged from 0.53 to 0.97 with all demonstrating at least a “moderate” correlation. The ATS correlation was “good” and the VAS correlation was “excellent”.
Table 1

Questionnaire validation

 

Kappa/ICC

Interpretation

Unpleasant taste

0.53

Moderate

Excessive thirst

0.59

Moderate

Nausea

0.60

Moderate

Vomiting

0.64

Good

Bloating

0.60

Moderate

Abdominal/cramps

0.54

Moderate

Headache

0.63

Good

Dizziness

0.69

Good

Sleep disturbance

0.57

Moderate

Aggregate

0.71

Good

Visual analogue

0.97

Excellent

Bland–Altman plots assessing the kappa and ICC scores for the nine individual components of the questionnaire along with the aggregate of these nine components and the visual analogue score. Kappa and ICC values; ≤0.20 poor, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 good, and >0.81 excellent

When assessing the ATS, six of 125 values lay outside the limits of agreement, but this was considered as acceptable. The limits ranged from −8 to 11 suggesting a large range despite the average difference being close to zero. On assessment of the VAS, nine out of 125 fell outside the limits of agreement. This was slightly higher than expected with seven of these having lower scores on the second assessment. There was, however, no systematic bias with the average difference close to zero and the upper and lower bounds within reasonable limits.

Validity

It was expected that the individual components of the questionnaire would relate to the ATS and thus those patients scoring higher (tolerated worse) on the individual components would also score the VAS lower (tolerated worse). Kruskal–Wallis tests demonstrated that for each component there was a statistically significant correlation between VAS and each ITS. As individual components increased, there was also a trend downwards in the VAS in all but one component (abdominal pain/cramps).

How each component related to “willingness to use the same bowel preparation again” correlated with the ITS and the ATS was assessed using Chi-squared tests. For each component (except abdominal pain/cramps and sleep disturbance), a statistically significant association was observed between the component and willingness to take the preparation again. There was noted to be a general trend downwards in the proportion of patients prepared to take the same preparation as the ATS increased.

Bowel-Preparation Tolerability

Patient Characteristics

A total of 634 consecutive patients were randomized. Table 2 summarizes patient sex and age with the primary colonoscopic indication. There were no significant differences detected between any of these parameters and the bowel preparation taken.
Table 2

Demographic data

 

Total

PEG

NaP

Pico

Patients, n (%)

634 (100)

284 (45)

179 (28)

171 (27)

Age (mean ± SD)

53.6 ± 12.6

53.2 ± 13.1

52.7 ± 12.1

55.3 ± 12.6

 <40 years, n (%)

81 (13)

41 (14)

22 (12)

18 (11)

 40–49 years, n (%)

137 (22)

58 (20)

43 (24)

36 (21)

 50–59 years, n (%)

191 (30)

85 (30)

56 (31)

50 (29)

 60–69 years, n (%)

168(27)

76 (27)

48 (27)

44 (26)

 70–74 years, n (%)

57 (9)

24 (9)

10 (6)

23 (14)

Gender: male, n (%)

292 (46)

127 (45)

83 (46)

82 (48)

Primary indication, n (%)

    

 Abdominal pain

56 (9)

23 (8)

15 (8)

18 (11)

 Altered bowel habits

124 (20)

66 (23)

34 (19)

24 (14)

 Bleeding/anemia

266 (42)

124 (44)

71 (40)

71 (42)

 Screening

170(27)

60 (21)

56 (31)

54 (32)

 Weight loss/miscellaneous

18 (3)

11 (4)

3 (2)

4 (2)

Sex, age, and primary indication for colonoscopy distributed by bowel preparation

Preparation Tolerability

All patients confirmed adherence to preparation instructions. Of 284 patients allocated PEG, 255 (89.8 %) took the full preparation with 277 (97.5 %) taking 75 % or more (>3 l). The remaining seven patients took 50 % (2 l). Of the 179 patients allocated NaP, 172 (96.1 %) took the full preparation. The remaining seven patients took 50 %. All 171 (100 %) patients allocated Pico took the full preparation. A total of 627 (98.9) patients fully completed the preparation tolerability questionnaire; PEG, 281 (98.9 %); NaP, 178 (99.4 %); Pico, 168 (98.2 %). Using a straightforward cross-tabulation with Fisher’s exact test, a difference in patient percentages completing the different bowel preparations was identified (p = 0.022).

Primary Aim

The ATS for PEG was significantly worse than NaP (8.0 vs. 6.7, p = 0.013) and Pico (8.0 vs. 5.2, p < 0.0001), while NaP was worse than Pico (6.7 vs. 5.2, p = 0.0035). Analysis on square-root transformed data, however, identified that the differences were dependent on sex and age (p = 0.009). A difference was also observed using the unpaired t test on the VAS ratings with a worse VAS for patients taking PEG than NaP (63.8 vs. 72.4, p = 0.0004) and Pico (63.8 vs. 81.2, p < 0.0001), while NaP was worse when compared to Pico (72.4 vs. 81.2, p = 0.0003). Analysis on the arcsine square root transformed response noted significant effects with age (p = 0.006) and bowel preparation (p < 0.001) but significant interactions were not present. Sex had a borderline significant effect (p = 0.050).

Secondary Aims

Controlling for sex, age, and procedure time (Table 3), Pico scored better than PEG for taste, nausea, vomiting, and bloating (all p < 0.001). NaP scored better than PEG for taste and bloating (both p < 0.001). Pico, however, scored better than NaP for taste, nausea, vomiting, and bloating (all p < 0.001). A significant interaction, however, was identified between sleep disturbance, procedure time (AM vs. PM) and bowel preparation used. For patients having an AM procedure, sleep disturbance was less for PEG than Pico (odds ratio (OR) = 1.5, 95 % CI 1.03–2.3) while NaP caused less than PEG (OR = 2.2, 95 % CI 1.4–3.5), and Pico (OR = 3.4, 95 % CI 2.0–5.7). For patients having a PM procedure, no significant differences were detected between Pico and PEG or NaP and Pico, however, patients taking PEG scored sleep disturbance better than NaP (OR = 0.43, 95 % CI 0.19–0.95).
Table 3

Bowel-preparation tolerability

 

PM vs. AM

Female vs. male

Pico vs. NaP

Pico vs. PEG

NaP vs. PEG

Unpleasant taste

 OR

 CI

 p value

NS

1.782

1.333–2.381

0.001

0.365

0.245–0.543

<0.001

0.140

0.096–0.205

<0.001

0.385

0.273–0.542

<0.001

Excessive thirst

 OR

 CI

 p value

1.570

1.094–2.253

0.0144

1.665

1.229–2.254

0.001

NS

NS

NS

Nausea

 OR

 CI

 p value

NS

3.186

2.266–4.478

<0.001

0.354

0.221–0.567

<0.001

0.320

0.207–0.495

<0.001

0.904

0.626–1.306

0.5901

Vomiting*

 OR

 CI

 p value

NS

4.123

2.110–8.056

<0.0001

0.316

0.112–0.894

0.03

0.187

0.072–0.490

0.0006

0.593

0.316–1.113

0.1038

Bloating

 OR

 CI

 p value

NS

1.868

1.351–2.583

0.0002

0.437

0.273–0.699

0.0006

0.234

0.152–0.360

<0.0001

0.536

0.371–0.774

0.0009

Abdominal pain/cramps

 OR

 CI

 p value

1.775

1.203–2.619

<0.0001

1.731

1.236–2.424

0.001

NS

NS

NS

Headache

 OR

 CI

 p value

NS

2.716

1.983–3.719

<0.0001

NS

NS

NS

Dizziness

 ORP  CI

 p value

NS

1.873

1.264–0.775

0.002

NS

NS

NS

Sleep disturbance

 OR

 CI

 p=

Interaction

1.975

1.430–2.728

<0.0001

Interaction with time of procedure

OR odds ratio, CI 95 % confidence intervals and p values are presented by time of procedure, sex, and bowel preparation

0 = nil, 1 = very mild, 2 = mild, 3 = moderate, and 4 = severe, NS not significant

Effect of Sex and Time of Procedure

Regardless of the bowel preparation agent used, females (n = 342) scored all nine ATS aspects worse than males (n = 292) (Fig. 2). These were significant for taste, thirst, nausea (all p < 0.001), vomiting (p = 0.03), bloating (p = 0.0006), headache (p < 0.0001), dizziness (p = 0.0018), and sleep disturbance (p < 0.0001). Regardless of the agent used, an AM procedure (n = 505) gave better scores for thirst (p = 0.014) and pain/cramping (p = 0.004) compared with PM procedures (n = 129) (Table 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2449-0/MediaObjects/10620_2012_2449_Fig2_HTML.gif
Fig. 2

Scoring of individual aspects of the bowel-preparation tolerability. Average overall tolerability scores for each bowel preparation and sex. Regardless of the bowel-preparation agent used, females (n = 342) scored all nine of the bowel tolerability aspects higher (worse) than males (n = 292) with the p values added for each tolerability aspect

Effect of Age

Regardless of the bowel preparation, patient age impacted ratings for taste, nausea, vomiting, bloating, and pain/cramps (Table 4). Patients <40 years scored them significantly higher than any other age group. As patient age increased, tolerability was better. The ATSs were significantly different between the bowel preparations but these were dependent on age (interaction p = 0.006; Fig. 3). NaP was less well tolerated than Pico in the 40–49 years (p = 0.009), 50–59 years (p = 0.045) and 70–75 years (p = 0.009) age groups. PEG was less well tolerated than Pico in patients aged <40 years (p = 0.004), 40–49 years (p = 0.002), 50–59 years (p < 0.001), and NaP in the 50–59 years (p < 0.001) age group. Older patients, however, tolerated all the preparations better than younger patients (p = 0.006). Of interest was that patients ≥60 years (n = 225) did not demonstrate any statistical differences in the ATS between PEG and NaP, or PEG and Pico.
Table 4

Bowel-preparation tolerability

 

Age group (years)

<40 vs. 40–49

<40 vs. 50–59

<40 vs. 60–69

<40 vs. 70–74

Unpleasant taste

 OR

 CI

 p value

1.719

1.047–2.822

0.0321

2.247

1.400–3.605

0.0008

2.007

1.242–3.244

0.0045

4.065

2.132–7.750

<0.0001

Excessive thirst

 OR

 CI

 p value

NS

NS

NS

NS

Nausea

 OR

 CI

 p value

2.192

1.298–3.703

0.0034

2.372

1.441–3.904

0.0007

4.093

2.413–6.944

<0.0001

4.428

2.137–9.175

<0.0001

Vomiting*

 OR

 CI

 p value

2.128

0.969–4.674

0.0599

3.013

1.371–6.623

0.0060

3.949

1.710–9.120

0.0013

3.432

1.049–11.228

0.0414

Bloating

 OR

 CI

 p value

1.396

0.821–2.373

0.2179

1.132

0.688–1.862

0.6268

2.304

1.359–3.906

0.0019

2.420

1.180–4.962

0.0159

Abdominal pain/cramps

 OR

 CI

 p value

2.279

1.343–3.865

0.0023

2.391

1.449–3.945

0.0006

4.265

2.488–7.312

<0.0001

4.476

2.151–9.316

<0.0001

Headache

 OR

 CI

 p value

NS

NS

NS

NS

Dizziness

 ORP CI

 p value

NS

NS

NS

NS

Sleep disturbance

 OR

 CI

 p value

NS

NS

NS

NS

OR odds ratio, CI 95 % confidence intervals

0 = nil, 1 = very mild, 2 = mild, 3 = moderate, and 4 = severe, NS not significant

https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2449-0/MediaObjects/10620_2012_2449_Fig3_HTML.gif
Fig. 3

Overall bowel-preparation tolerability. a Average overall tolerability scores for each bowel preparation used (±SEM). b Statistical significance identified by age group between the preparations used (p value). Using square-root transformed data tolerability scores, a significant interaction between age and bowel preparation (p = 0.006) was observed when all nine tolerability aspects were combined with the patient age group, with older patients tolerating all the bowel preparations better than younger patients. PEG was significantly less well tolerated than Pico in patients aged <60 years but no statistical differences were detected between PEG and either NaP or Pico in patients older than 60 years. NS not significant

The VAS also demonstrated that tolerability was different between the bowel preparations (p < 0.001), but this was dependant on the age group with the younger patients tolerating all preparations, regardless of type, worse than older patients (p = 0.039). Females also tolerated all preparations, regardless of type, worse than males (p = 0.009).

Correlation of Tolerability with Mucosal Cleansing

The ATS and VAS ratings (Table 5) determined if these impacted the mucosal cleaning for each group as previously published [14]. Once the response was adjusted for age, sex, time of procedure, and preparation, there was no impact of the tolerability scores on mucosal cleaning as the patients still took the bowel preparation as prescribed despite poorer tolerability.
Table 5

Correlation between tolerability and mucosal cleansing

 

Rectosigmoid

Mid colon

Right colon

Total score

ATS

 n

CC

p value

605

−0.00948

0.8160

603

0.02700

0.5081

603

0.03131

0.4428

602

0.01601

0.6950

VAS

 n

 CC

 p value

618

0.01803

0.6543

619

−0.04556

0.2585

617

−0.07658

0.0573

616

−0.04797

0.2345

CC correlation coefficient, VAS visual analogue score

Missing Work and Reuse of the Same Preparation

Missing work was assessed in patients <70 years as no patients ≥70 years in this cohort were still working. A total of 412 patients were analyzed (186 PEG, 123 NaP, 103 Pico). Only patient age had an impact on missing work (p = 0.018) with older patients less likely than younger to miss work. No differences were detected between the bowel preparations (Table 6).
Table 6

Correlation with bowel-preparation tolerability and mucosal cleansing at colonoscopy

 

Rectosigmoid

Mid colon

Right colon

Total score

VAS

 n

 CC

 p value

605

−0.00948

0.8160

603

0.02700

0.5081

603

0.03131

0.4428

602

0.01601

0.6950

 n

 CC

 p value

618

0.01803

0.6543

619

−0.04556

0.2585

617

−0.07658

0.0573

616

−0.04797

0.2345

CC correlation coefficient

There were 94.5 % of the patients that responded to the question about taking the same preparation again. Sex (p < 0.001), age (p = 0.048), and bowel preparation (p < 0.001) impacted patient willingness to retake the same preparation. Males (264 of 295; 89.4 %) were more likely than females (161 of 34; 75.7 %) to be willing to take the same bowel preparation again (OR = 3.15, 95 % CI 1.921–5.175), while patients <40 years (60 of 87; 69 %) were less likely to use the same preparation again regardless of the agent initially used compared to any other age group (40–50 years, 87.9 %; 50–60 years, 81.2 %; 60–70 years, 84.4 %; >75 years, 81 %). Adjusting for age and sex, patients were more willing to take Pico for a subsequent procedure (Pico [154 of 165, 93.3 %] vs. PEG [214 of 297, 72.1 %], OR = 5.947, 95 % CI 2.932–12.061, p < 0.001; Pico vs. NaP [157 of 178, 88.2 %], OR = 2.013, 95 % CI 0.905–4.479, p = 0.087) and more likely to take NaP than PEG (NaP vs. PEG, OR = 2.955, 95 % CI 1.712–5.100, p = 0.001). A second analysis assessing each tolerability variable was undertaken and demonstrated that taste (p < 0.001), nausea (p < 0.001), and sleep disturbance (p < 0.0090) impacted willingness to take the same bowel preparation again.

Discussion

Colonoscopy is extremely common. Many patients complain about the bowel preparation, while endoscopists complain if the bowel is not clean. Unfortunately, the literature contains numerous underpowered studies. Our study is the largest prospective randomized controlled trial in a community setting, and the first to use an objective validated tolerability assessment tool to examine three commonly used bowel-preparation agents in Australia. All our patients were ambulatory, confirmed compliance to dietary instructions, and 98 % took 75 % or more of the preparation agent. Data collection was extremely good, with 99 % questionnaire completion.

Questionnaire validation identified a satisfactory internal consistency, with a good to excellent test–retest reliability. The questionnaire identified patients who tolerated the bowel preparation less well, were less likely to want to take the same preparation again, and that ITS were associated with a downward trend in the VAS as predicted. The questionnaire was confirmed as a reliable and accurate method as assessing patient tolerability.

In line with previous studies, PEG was least well tolerated, while NaP was less well tolerated than Pico [10, 26]. Due to the study size, however, we also identified differences due to patient age, sex, and time of the colonoscopic procedure. Contrary to expectations, females rated all nine tolerability aspects worse than males regardless of the preparation received. Tolerability by age identified that younger patients rated all the bowels preparations worse than the older patients, but any proposed explanations for these differences would be purely speculative.

Splitting the PEG preparation was not undertaken in the study. Splitting the dose has been suggested to improve a patient’s willingness to reuse the same preparation regime [24], with superior palatability, but without any differences in overall tolerability [15]. Neither of the studies investigating split dosing, however, used a validated tolerability questionnaire nor did they stratify by patient age. Our study, however, included 225 patients ≥60 years and PEG’s tolerability, without dose splitting, was observed to be the same as Pico and NaP despite the greater fluid volume. This is a welcome finding as use of osmotic agents in patients >75 years is associated with adverse outcomes [2] but concerns over PEG’s tolerability can limit its use. Although patients >75 years were not included in this study, it would appear that the patients ≥60 years can tolerate the PEG volumes and avoid the risks associated with osmotic agents. The generalized acceptability of this preparation in the older age group will undoubtedly be further improved by wider use of the recently developed low-volume PEG [22] and if splitting the dose of the 4L PEG solution also improves its tolerability, then this can also enhance patient acceptability.

In summary, this is the largest prospective randomized single-blinded study yet to assess bowel-cleansing preparation tolerability, and the first to use a validated questionnaire to evaluate the three major classes of bowel-cleansing preparations. Due to the large patient numbers, we identified that patient age, sex, and procedure time all impacted tolerability. Pico was tolerated best, but against expectations, PEG’s tolerability in older patients was equal to that of NaP and Pico. Its use in the older patient population can thus be recommended and will avoid the electrolyte disturbances not infrequently observed with the use of the osmotic agents.

Conflict of interest

None.

Copyright information

© Springer Science+Business Media New York 2012