CTT was measured on days 4 and 7 using multiple radioopaque markers. Bowel preparation state was predicted by measuring markers on days 4 and 7. A single tablet of Kolomark™ (M. I. Tech, Pyeongtaek, Gyeonggi, Korea) containing 20 small rings was taken with water once per day for 3 days at 9 AM, and simple abdominal radiographs were taken at 9 AM on days 4 and 7. The right colon was identified on simple abdominal radiographs using two guidelines: one that connected all the spinous processes of the vertebrae and another that crossed from the 5th lumbar vertebral body to the right pelvic outlet. To identify the left colon, a line was drawn to connect all of the spinous processes, and the second line was drawn from the 5th lumbar vertebral body to the left anterior superior iliac spine. The portion of the colon within this plane was determined to be the left colon. A plane below these two lines, composed of a line that connected the right pelvic outlet and a second line that crossed the left anterior superior iliac spine, was determined to be the location of the rectum and sigmoid colon (Fig. 1).¹⁷ CTT was assessed by counting the number of radio-opaque markers on a plain radiogram on day 4. These numbers were multiplied by 1.2 and added to the number of radio-opaque markers counted on day 7.¹⁸ An average CTT has not been clearly determined for the Korean population, and previous studies have used various standards. However, recent studies have settled on a mean CTT of 30 hours.^5,19,20

Based on the results of these studies, we classified patients with a CTT of >30 hours as the slow transit time (STT) group, and those with a CTT of <30 hours as the normal transit time (NTT) group. Furthermore, no standard transit time has been determined for cases of pelvic outlet obstruction, which delays passage through the sigmoid colon and rectum. Based on the mean CTT in each segment, we used 20 hours as the mean right and left CTT, and 10 hours as the mean transit time for the rectum and sigmoid colon. Slow transit constipation was diagnosed in people with a >20-hour delay in transit time in both right and left colons. Those with a transit time postponed by >10 hours in the rectum and sigmoid colon were defined as having pelvic outlet obstruction.

All patients were instructed to avoid fiber-rich foods, fruits with seeds, and some grains for 3 days. They also took 4 L of PEG as a split dose: 2 L of solution in the evening before the procedure, and 2 L of solution the next morning at 8 am. Colonoscopies were performed within 5-7 hours after the end of bowel preparation. We enrolled patients who had finished their bowel preparation on schedule. To determine a bowel preparation scale without bias, all recorded colonoscopic images were reevaluated by one experienced physician (>5,000 colonoscopies). The medical record of each patient was also reviewed, and previous medical histories, previous colonoscopy reports, and pathologic findings were investigated.

The Boston Bowel Preparation Scale (BBPS) was used to evaluate the extent of bowel preparation. The scale splits the colon into three segments: the right, left, and transverse colons. Bowel preparation was measured on a scale from 0 to 3 in each segment. A part of the colon in which the mucosa cannot be observed on colonoscopy was given 0 points. When the mucosa could be clearly observed throughout the colon, the segment was given 3 points. Scores from each segment were added together, and the mean score was 6. Using a score of 6 as a reference, patients who received >6 points were classified as having proper bowel preparation, and patients who had <6 points in total or <2 points for one of any three segments were classified as having inadequate bowel preparation.²⁰

SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses. Demographics and medical histories were assessed using a descriptive statistics method. The frequencies and ratios of categorical variables were determined and analyzed using Pearson's chi-square or Fisher's exact test. The mean, minimal, and maximal values of continuous variables were described, and values were analyzed using Student's t-test. A P-value <0.05 was considered statistically significant. Pearson's correlation coefficient was used to analyze the correlations between BBPS and CTT. A receiver operator characteristic (ROC) curve was drawn to determine which BBPS scores could best predict bowel preparation quality.

We analyzed a total of 160 patients with chronic constipation. There were 82 patients in the STT group (≥30-hour CTT), and 78 patients were classified in the NTT group (<30-hour CTT). Sixty-six percent of all patients were female, but no statistical differences in sex distributions (percentages of female patients) were observed between the two groups (STT vs. NTT, 63% vs. 69%; P=0.437). The mean patient age was 62±14 years. There were no differences in patient age between the two groups (STT vs. NTT, 64±15 years vs. 61±13 years; P=0.150). The proportions of patients with hypertension (STT vs. NTT, 27% vs. 33%; P=0.370) or diabetes (STT vs. NTT, 18% vs. 15%; P=0.623) determined from medical histories were not significantly different between the two groups. The proportions patients with neurovascular diseases including Parkinson's disease (STT vs. NTT, 10% vs. 4%; P=0.212), thyroid diseases including thyroid cancer (STT vs. NTT, 2% vs. 5%; P=0.434), and chronic obstructive pulmonary disease (STT vs. NTT, 5% vs. 1%; P=0.368), and history of abdominal surgery did not show any statistically significant differences between the two groups. Adenoma was identified in 32 patients (20%), but adenoma detection rate showed no significant difference between the two groups (STT vs. NTT, 21% vs. 19%; P=0.812). During colonoscopy, colon cancer was found in one patient from each group (Table 1).

With reference values set to a mean CTT of 30 hours and a mean BBPS score of 6, the proportions of patients with a BBPS <6, reflecting inadequate bowel preparation, were 10.3% and 31.7% in the NTT and STT groups, respectively. The rate was 3-times higher in the STT group, and this difference was statistically significant (P=0.001) (Fig. 2). There was also a significant difference in the mean values of BBPS scores between the two groups (STT vs. NTT, 6.1±1.5 vs. 7.0±1.2; P<0.001). Although age, sex, history of diabetes, neurovascular diseases including Parkinson's, thyroid diseases, and CTT (reference, 30 hours) were found to be related to inadequate bowel preparation in other studies, this study showed that CTT was the only factor associated with inadequate bowel preparation(OR, 0.261; 95% CI, 0.107-0.634; P=0.003). Using ROC curve analysis, we found that a cut-off CTT of 37 hours could predict inadequate bowel preparation with 0.735 specificity and 0.643 sensitivity. This method was a more effective predictor of bowel preparation than CTT (area under the ROC curve: 0.676) (Fig. 2A). When the reference value for CTT was changed to 37 hours, the inadequate bowel preparation rates were 10.0% in the NTT group and 35.7% in the STT group, representing a larger difference (Fig. 3).

Subgroup analysis was performed in the STT group. Those with a >20-hour delay in the transit times of the right and left colons were classified as the slow transit constipation group, and those with a >10-hour delay in the rectum and sigmoid CTT were classified as the pelvic outlet obstruction group. The slow transit constipation group included 82 patients, while the pelvic outlet obstruction group contained 57 patients. The number of patients with both slow transit constipation and pelvic outlet obstruction was 46. Because most pelvic outlet obstruction patients had mixed-type constipation, 11 were diagnosed with pelvic outlet obstruction only. All patients were then further divided into three groups: group A (only slow CTT), group B (only pelvic outlet obstruction), and group C (mixed constipation types). When each group was compared with the NTT group, the relationship between CTT and inadequate bowel preparation revealed statistically significant differences for groups A and C, which included patients with slow transit constipation (group A vs. NTT group, 26.0% vs. 9.4%; P=0.027) (group C vs. NTT group, 30.4% vs. 9.4%; P=0.008). However, group B showed no statistically significant relationship between inadequate bowel preparation and CTT (group B vs. NTT group, 18.0% vs. 9.4%; P=0.593).