Although tapentadol and oxycodone both increase colonic volume, tapentadol treatment resulted in softer stools and less constipation: a mechanistic study in healthy volunteers

Study overview

This study was a randomized, double-blind, placebo-controlled, cross-over study, where 21 healthy male volunteers were administered tapentadol (Palexia®, 50 mg extended release), oxycodone (OxyContin®, 10 mg extended release) or placebo in three study periods of 14 days. Tablets were administered once orally on days 1 and 14 (morning) and twice orally on days 2–13 (morning and evening). Tablets were identical in appearance and were manufactured at the Hospital Pharmacy, Aarhus University Hospital, Aarhus, Denmark. The tablets were encapsulated in DPcaps® (red color, size AA, 13.07–14.44×9.39 mm, Capsugel®) that are shell capsules of hard gelatin that hides the appearance of the original tablets. The drug release properties of the tablets were not affected as previous shown in vitro and in vivo [10], [, 11]. The wash-out period between treatments was at least 7 days, where the effect of tapentadol and oxycodone has worn off in line with the halftime of the active components [12], [, 13]. The dosages of tapentadol 50 mg and oxycodone 10 mg were believed to be equianalgesic as previous shown in a phase II trial by Stegmann et al. and a phase III trial by Hartrick et al. [5], [, 14]. Included subjects were randomized to the order of treatment into one block with six possible sequences (randomization.com), see detailed overview in Figure 1. Mirror randomization was performed in case of participant drop out. The research was carried out at the research facilities of Mech-Sense, Aalborg University Hospital, Departments of Gastroenterology & Radiology, Aalborg, Denmark. Included subjects did not have current GI symptoms or history of GI diseases. They underwent an MRI examination of the abdomen at study day 1 (pre-treatment) and at study day 14 (post-treatment). Subjects were asked to fast 6 h before MRI, and no instructions on toilet visits were given. The clinical trial was approved by The North Denmark Region Committee on Health Research Ethics and the Danish Health and Medicines Authority (reference numbers: N-20170009 and 2017041794). The clinical trial was registered at www.clinicaltrialsregister.eu (EudraCT number: 2017-000141-52).

Figure 1:

An overview of the study flow. One drop-out participant was replaced using mirror-randomization.

Data were analyzed according to intention-to-treat. Oxy, oxycodone; Tap, tapentadol; Pla, placebo; MRI, Magnetic resonance imaging.

The data presented in the current study were secondary end-points obtained as part of a larger study protocol. The primary end-point is not included in this manuscript and has been submitted for publication elsewhere.

Magnetic resonance imaging

Abdominal MRI scans were carried out using a 1.5 T GE Discovery MR450 System (GE Healthcare, Milwaukee, WI). One coronal T2-weighted single-shot fast spin-echo scan was performed with the following settings: Repetition time (TR)=715 msec, echo time (TE)=93 msec, flip angle=90°, field of view=480×480 mm, matrix size=512×512 pixels, acceleration factor=2, in-plane resolution=0.9375×0.9375 mm, 30–40 slices, and a slice thickness of 4 mm. The images were obtained in two breath-holds lasting approximately 20 s each.

Data analysis

Image segmentation was performed using a semi-automatic segmentation platform (Matlab version R2016a, MathWorks, Natick, MA, USA) [15]. The final segmented volume included the outer surface of the colon (and hence the total colon volume including bowel wall, faeces and gas). The segmentation was divided in the four colonic segments: ascending colon/cecum, transverse colon, descending colon, and sigmoid colon/rectum, see Figure 2. The analysis procedure was that the observer outlines a coarse region of interest encapsulating the colon on each coronal image where the colon was visible. The software allows the observer to scroll through images and add or delete regions of interest after which the software automatically uses a clustering algorithm (k-means initiated with three random cluster centres) to discard non-colon voxels based on their relative MRI signal intensities. The observer can afterwards correct failed classifications of areas. The classified colonic volume was then quantified (voxel width×height×slice thickness). A single observer (KKJ, trained radiographer) did all segmentations. Segmentations were approved by another observer (EBM). The observers were blinded to study treatment during analysis.

Figure 2:

Coronal T2-weighted magnetic resonance image of the abdomen. (A) Segmented colonic segments are superimposed on the image; yellow, ascending colon and cecum; red, transverse colon; green, descending colon; blue, sigmoid colon and rectum. (B) Three-dimensional model of the segmented colon.

Questionnaires

Study subjects were asked to fill in three questionnaires for evaluation of side-effects during treatments. (1) A modified version of the Gastrointestinal Symptom Rating Scale (GSRS) were filled in at first (pre-treatment) and last study day, where 11 side-effects were rated on a 7-points Likert scale. (2) The Bristol stool scale for evaluation of spontaneous bowel movement, including frequency and stool consistency rated on a scale from 1 to 7, where 1=very hard stool, 3–4=normal, and 7=diarrhoea [16], (Bristol stool scale data has been submitted for publication elsewhere). (3) Treatment side-effects were rated at first (pre-treatment), fourth, and last study day, on a 4-points Likert scale, where 0=no, 1=mild, 2=moderate, 3=severe and 4=unbearable.

For simplification of analysis, data from the side-effect questionnaires were divided into three sub-categories (see Table 1) including (1) GI-related questions, (2) central nervous system (CNS)-related questions, and (3) other questions. Redundant questions were left out from analysis.

Subjects were also asked to fill out the Subjective Opiate Withdrawal Scale questionnaire three days after each treatment period [17]. A total score of 1–10 means mild withdrawal, a score of 11–20 means moderate withdrawal, and a score of 21–30 means severe withdrawal.

Pain tolerance threshold

The efficacy of the administered dosages of tapentadol, oxycodone and placebo were tested in an experimental pain tolerance test using a handheld pressure algometer (Type 2, Somedic production AB, Sweden) with a probe of 1 cm². The pain tolerance threshold was found during a continuous and increasing pressure (30 kPa/s) on the thigh 15 cm proximal to the patella. The test was performed at day 1, day 4 and day 14.

Statistical analysis

All data are presented as means ± standard deviation or medians with interquartile range (IQR) according to data distribution. The effect of treatments was analysed on baseline-corrected colonic volume, i.e. the result on day 14 minus the result on day 1, in order to minimize physiological fluctuations. Analysis was performed using repeated measures mixed models with the two factors treatment and colonic segment and adjusted for multiple comparisons with Bonferroni post hoc tests. Differences between study periods were investigated for evaluation of potential carry-over effects using the same type of repeated measures mixed models. Questionnaire scores were tested for each question with Kruskal-Wallis one-way analysis of variance and adjusted for multiplicity using Bonferroni correction. Correlations between questionnaires (global score for each questionnaire) and total colonic volume change were tested for each treatment with Spearman correlation analyses. The pain tolerance threshold was tested using a repeated measures one-way ANOVA. p-values less than 0.05 were considered statistically significant. Statistical analysis was performed with STATA, version 15 (StataCorp LP, College Station, Texas, USA).

Study subjects

The 21 enrolled healthy male subjects had a mean ( ± standard deviation) age of 28.0 ± 9.0 years, mean height of 182.2 ± 5.7 cm, and a mean BMI of 23.9 ± 2.7 kg/m². All the 21 subjects successfully completed the imaging protocol; thus 126 imaging sets were included in analysis.

Colonic volume

Difference between total colonic volumes obtained on study days 1 and 14 are shown in Figure 3. Oxycodone treatment increased total colonic volume change compared to placebo (p=0.002). Tapentadol treatment likewise increased total colonic volume change compared to placebo (p=0.003), while no difference was observed between oxycodone and tapentadol (p=0.9). The post hoc analysis revealed no segmental differences, see Figure 3. There were no differences between colonic volumes obtained in the three study periods (when data was in chronological order of study visits before stratification into treatments).

Figure 3:

Boxplots of change in colonic volume from day 1 to day 14 during treatment with oxycodone, tapentadol or placebo.

(A) Total colon, (B) ascending colon and cecum, (C) transverse colon, (D) descending colon, (E) sigmoid colon and rectum.

Treatment side-effects

Table 1 summarizes the number of subjects that reported side-effects with the GSRS and side-effects questionnaires. The severity of reported side-effects can be seen in Supplementary data. At day 14 of oxycodone treatment, subjects reported increased GI-related side-effects including constipation and straining during defecation as compared to both tapentadol and placebo (both p≤0.007). Moderate or severe constipation symptoms (2 or 3 on the Likert scale) were reported of 11 subjects (52%) during oxycodone treatment, 1 subject (5%) during tapentadol treatment and 0 during placebo. Subjects likewise reported increased tired/drowsy (p≤0.010) at the last study day during oxycodone treatment compared to both tapentadol and placebo. Moderate or severe nausea were reported by 7 subjects (33%) on the last study day (all 7 during oxycodone), and 2 subjects (10%) experienced vomiting (1 during oxycodone and 1 during placebo). No differences were observed on the baseline data or between tapentadol treatment and placebo.

No subjects scored more than 10 on the Subjective Opiate Withdrawal Scale questionnaire, meaning that no subjects experienced more than mild withdrawal symptoms after the opioid treatments.

Bristol stool scale

The mean weekly number of spontaneous bowel movements and mean Bristol stool scale are summarized in Table 2. Compared to placebo, oxycodone treatment decreased the number of bowel movements the first week (p<0.001) and decreased mean stool consistency both in the first week (p<0.001) and in the second week (p=0.005). Compared to placebo, tapentadol treatment did not change the number of bowel movements or stool consistency (all p>0.6). Compared to tapentadol, oxycodone treatment decreased the number of bowel movements in the first week (p=0.02) and decreased mean stool consistency in the first week (p<0.001) and the second week (p=0.001). The bowel movement frequency for each individual subject is illustrated in Figure 4 and the frequency and distribution of the different stool patterns are visualised in Figure 5. The frequency of spontaneous bowel movements and the Bristol stool scale were not affected by which order treatments were administered.

Figure 4:

Line plots of the number of spontaneous bowel movements during treatment with tapentadol, oxycodone and placebo.

(A) Spontaneous bowel movements during week 1. (B) Spontaneous bowel movements during week 2.

Figure 5:

Bristol stool scale for 14 days of treatment with tapentadol, oxycodone and placebo.

Data is presented in mean ± standard error of the mean (SEM). The data distribution is shown in the panel to the right.

Pain tolerance threshold

An overview of the pain tolerance thresholds during the three treatments can be seen in Table 3. There were no differences in the pain tolerance thresholds during any of the treatments (p=0.35).

Associations between colonic volume change and side-effects

An association between change in total colonic volume and mean stool Bristol stool scale was observed during tapentadol treatment (Spearman’s Rho=0.51, p=0.019), meaning that a colonic volume increase was related to more watery stool. No other associations were found (all Spearman’s Rho< ± 0.26; all p>0.255).

Study limitations

The data presented in this manuscript were secondary end-points obtained as part of a larger study protocol in the randomized trial. The study may be underpowered for evaluation of colonic volume, although the included sample size is similar or higher compared to other studies of colonic volume [7], [18], [23].

Baseline-corrected measurements were used in analysis to account for a large variation in the baseline measurements. The high physiological variation in colonic volume between study periods implies that the colonic volume is dependent on several factors including e.g. diet or toilet visit prior to assessment [7], [, 9]. The randomized study design minimizes the effect of this bias. In future studies, control of diet and even toilet visits (if possible) may minimize physiological variation in data and make interpretation of any intervention effect easier. Lastly, the findings in the healthy volunteers may not translate directly into patients having comorbidities, taking different medications or having higher tolerance of opioid usage due to longer treatment periods.