Randomised clinical trial: the effects of daikenchuto, TU-100, on gastrointestinal and colonic transit, anorectal and bowel function in female patients with functional constipation


  • J. Iturrino,

    1. Clinical Enteric Neuroscience Translational and Epidemiological Research (CENTER), College of Medicine, Mayo Clinic, Rochester, MN, USA
    2. Division of Gastroenterology and Hepatology, College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author
  • M. Camilleri,

    Corresponding author
    1. Division of Gastroenterology and Hepatology, College of Medicine, Mayo Clinic, Rochester, MN, USA
    • Clinical Enteric Neuroscience Translational and Epidemiological Research (CENTER), College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author
  • B. S. Wong,

    1. Clinical Enteric Neuroscience Translational and Epidemiological Research (CENTER), College of Medicine, Mayo Clinic, Rochester, MN, USA
    2. Division of Gastroenterology and Hepatology, College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author
  • S. J. Linker Nord,

    1. Clinical Enteric Neuroscience Translational and Epidemiological Research (CENTER), College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author
  • D. Burton,

    1. Clinical Enteric Neuroscience Translational and Epidemiological Research (CENTER), College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author
  • A. R. Zinsmeister

    1. Department of Health Sciences Research, Division of Biomedical Statistics and Informatics, College of Medicine, Mayo Clinic, Rochester, MN, USA
    Search for more papers by this author

Correspondence to:

Dr M. Camilleri, Mayo Clinic, 200 First St. S.W., Charlton Bldg., Rm. 8-110, Rochester, MN 55905, USA.

E-mail: camilleri.michael@mayo.edu



Daikenchuto, a Japanese herbal medicine used for post-operative ileus and constipation, dose dependently stimulates gastrointestinal (GI) motility and decreases rectal compliance and sensation. Effects of TU-100 (commercial form of daikenchuto) in adults with constipation are unknown.


To compare the effects of oral TU-100, 2.5 g t.d.s. or 5 g t.d.s. and placebo t.d.s. on GI and colonic transit (CT), rectal compliance (RC) and sensation thresholds (RST), anal sphincter pressures (ASP) and bowel function in female patients with functional constipation (FC).


We conducted a single-centre, randomised, parallel-group, double-blind, pharmacodynamic study; 45 female patients with FC without evidence of rectal evacuation disorder were assigned to 28 days' treatment with oral placebo or TU-100 (Tsumura USA, Princeton, NJ, USA). Demographic data and CT were measured at baseline and randomisation stratified by baseline CT (GC> or <1.9) and by BMI (<25 or ≥25 kg/m2). At the end of treatment period, we measured GI and CT by scintigraphy, RST and RC by barostat, ASP by manometry, psychosensory sensations, bowel function by daily diary and quality of life (QOL). The study had power to detect effect sizes of 33% (CT), 40% (RC) and 46% (RST). Statistical analysis included BMI as covariate.


TU-100 had no significant effects on GI or CT, RC, ASP, recto-anal pressure difference, or RST. The 5 g t.d.s./day dose was associated with lower RST for first sensation and gas (unadjusted P: 0.045 and 0.024 respectively). There were no treatment effects on psychosensory symptoms, stool frequency, stool consistency or QOL.


Mechanisms underlying the observed clinical benefit of TU-100 remain unclear (ClinicalTrials.gov NCT01139216).


Complementary and alternative medicine (CAM) is a combination of healing philosophies and therapies that is not commonly used in conventional medicine.[1] Over the last decade, the US population has shown a substantial and steady use of CAM. It is estimated that 38.3% of adults in the United States use some form of CAM; among these CAM therapies, natural products (17.7%) are the second most commonly used, after prayer.[2] In 2007, there was expenditure of $14.8 billion to purchase natural products such as herbal medicines.[3] Although herbal medicines have been widely use to treat gastrointestinal (GI) symptoms,[4] there are few studies documenting the pharmacological effects of these agents.

Daikenchuto, a Japanese herbal medicine, which consists of an extract powder from dried Sichuan pepper, processed ginger, ginseng and maltose powder, has been used for the treatment of paralytic ileus and radiation-induced enteritis due to its possible prokinetic effect.[5-7]

In animal studies, intraduodenal and intrajejunal administration of TU-100 (a modern herbal product of daikenchuto manufactured in the granule form) dose dependently increased motility of the duodenum, proximal jejunum and distal jejunum in conscious dogs.[8] In vivo and in vitro experiments have shown that the prokinetic effect of daikenchuto after intestinal dysmotility, induced laparoscopically or chemically, was inhibited by atropine or a 5-HT4 antagonist, suggesting that daikenchuto modulates cholinergic and serotonergic mechanisms.[9-11] Daikenchuto also stimulated acetylcholine release in pig ileal smooth muscle.[12]

In a study in healthy adult humans, TU-100 appeared to be safe and well tolerated and demonstrated prokinetic effects on ascending colon transit emptying.[13] In another study, daikenchuto decreased the sensation thresholds (implying enhanced sensation) and rectal compliance in children with severe constipation; the authors attributed the potential beneficial clinical effect on constipation to decrease in rectal reservoir volume and increased intestinal peristalsis.[14]

Given the known delay in colonic transit in patients with constipation,[13] the promising prokinetic effects seen with TU-100 in healthy adults and the lack of data from TU-100 on rectal sensation and compliance in adults, we aimed to evaluate the effects of TU-100 on gastrointestinal and colonic transit, rectal sensation and compliance, bowel function and quality of life in female adults with functional constipation.

Materials And Methods

Study design, randomisation and allocation to treatment

We performed a single-centre, randomised, three-group parallel fashion, double-blinded, placebo-controlled, dose–response, pharmacodynamic study evaluating the effects of orally administered TU-100 on rectal compliance, rectal sensation, gastrointestinal and colonic transit in female patients (aged between 18 and 65 years) with functional constipation without evidence of evacuation disorder. Randomisation was stratified based on BMI (<25 or ≥25 kg/m2) and baseline geometric centre at 24 h (GC24, >1.90 or <1.90). Cato Research, a contract research organisation (CRO), was responsible for providing the schedule of medication doses, which were randomly assigned in fixed block sizes according to a schedule provided by an independent study statistician. The three study medications were identical in appearance and were prepared and dispensed by the Mayo Clinic Research Pharmacy. The clinical investigators, study personnel and participants were blinded to the treatment assignments. Allocation was concealed and medication was administered in double-blinded fashion.


Participants were enrolled by one study coordinator between October 2010 and November 2012. The study was approved by the Mayo Clinic Institutional Review Board and was registered at ClinicalTrials.gov (NCT01139216). Data were collected in the Clinical Research Unit at Mayo Clinic in Rochester, MN.

Female adults with functional constipation were recruited from the local community by public advertisement or targeted mailing. After informed consent was signed, candidates who met the eligibility criteria for the study underwent a complete history and physical examination. All candidates were screened by means of a short Bowel Disease Questionnaire (BDQ) to confirm the presence of functional constipation by Rome III criteria.[15] Participants were also required to have no history of anal digitation or perineal or vaginal pressure to facilitate rectal evacuation, as well as a normal rectal examination within the last 2 years by one of the study physicians or in the screening visit for the study. As a result of this evaluation, seven patients were excluded from participation. In the remaining participants, examination documented normal perineal descent, absence of high resting anal sphincter tone, and absence of paradoxical contraction of the pelvic floor during simulation of evacuation of the examining finger. Therefore, none of the participants underwent anorectal manometry and balloon expulsion as part of the study. We excluded candidates with structural or metabolic conditions that affect the gastrointestinal system or with functional gastrointestinal disorders other than constipation; history of allergic reactions to egg, ginseng, ginger and Sichuan pepper or lactose intolerance; history of uncontrolled medical conditions (e.g. significant cardiovascular, respiratory, renal, hepatic, haematological, neurological or psychiatric diseases); or current use of any medication that could potentially alter GI transit. Tricyclic antidepressants, selective serotonin reuptake inhibitors (SSRI) and antidepressants were permissible at low and stable doses. Analgesics and alcohol in moderation were permissible except for the 48 h before transit, anorectal manometry and rectal barostat studies. A urine pregnancy test was performed in all women of child-bearing potential within 48 h of transit studies. Women who were surgically sterilised or postmenopausal were exempt from pregnancy testing. All participants signed informed consent.

Study medication

Dosage of 2.5 g or 5 g, three times a day (t.d.s.), of TU-100 or matching placebo was administered orally as a solution immediately before meals for 28 consecutive days. Each dose was dissolved in 30 mL of warm water immediately prior to consumption and swallowed over a maximum of 1 min. All packets containing the study medication or placebo were identical. The packets were distributed in boxes and refills were given during scheduled visits.

Study design

Female patients with functional constipation were informed about the study and screened for eligibility at Visit 1 (Figure 1). The consent form was reviewed and signed by the patient and the study coordinator. A brief interview and physical exam were performed by a study physician. A short BDQ questionnaire was filled out and the participants received a bowel diary, which they completed throughout the study. At this time point, blood and urine samples were obtained for routine laboratory safety tests.

Figure 1.

Experimental protocol: Participants were enrolled in a 6- to 8-week trial including 2 weeks of baseline observation to record bowel habit in a diary and baseline colonic transit. Patients were randomised to one of three treatment groups which included placebo. During the next 4 weeks, they completed daily bowel function diary and, in the last 3 days of treatment, they underwent studies of gastrointestinal and colonic transit, rectal compliance and sensation, and anal sphincter function.

About 2 weeks after Visit 1, eligible patients returned for Visit 2, which consisted of an abbreviated baseline measurement of colonic transit by scintigraphy, with images obtained only at 4 and 24 h after 111In capsule ingestion. Based on previous scintigraphic studies,[13, 16] a geometric centre of ≤2.33 at 24 h was required for randomisation to study medication. The participants who qualified after scintigraphic study were randomised to placebo or one of two doses of TU-100, 2.5 g or 5 g t.d.s. Randomisation was stratified to ensure equal distribution of patients with baseline colon transit GC24> or <1.9, and BMI> or <25 kg/m2.

Participants returned for Visit 3 within 2 days of the previous visit to receive the study medication to take at home for approximately 14 days until their next visit, and they completed the Patient Assessment of Constipation Quality of Life (PAC-QOL) Questionnaire.[17] At Visit 4, participants received a new supply of the study medication to take at home until Visit 5, which was approximately 12 days later. Blood and urine samples were obtained for routine laboratory safety tests.

About 2 weeks later, the participants returned for Visits 5–7 for the poststudy medication transit scintigraphy test performed over a 48-h period. Study medication was administered at the Clinical Research Unit (CRU) to ensure the correct time of dosing during the last days of the scintigraphy test. During Visit 7, patients again completed the PAC-QOL Questionnaire and underwent a physical exam. After the 48-h scan, anorectal manometry and rectal barostat procedures were performed. Within 7–10 days of Visit 7, participants returned for Visit 8 to provide blood and urine samples for routine laboratory safety tests and for a final study interview.

GI and colonic transit measurements

The gastrointestinal and colonic transit measurements consisted of the ingestion of a radiolabelled capsule containing 111In absorbed on activated charcoal particles, followed by ingestion of a meal labelled with 99mTc-sulphur colloid. Abdominal images were taken at specific time points after ingestion of the radiolabelled materials to assess colonic transit, all as previously described.[16, 18-22] An abbreviated version of the established scintigraphic method was used to measure GI and colonic transit at baseline. Participants ingested a radiolabelled 111In capsule after overnight fast. Ingestion of a methacrylate-coated, delayed-release capsule delivered 111In absorbed on activated charcoal particles to the colon. Participants returned 24 h after ingesting the capsule for colonic transit scans.

The full end of treatment transit test was performed at 28 days of study medication. After an overnight fast of at least 8 h, participants ingested the radiolabelled (111In) capsule. After the capsule emptied from the stomach, documented by its position relative to radioisotopic markers placed on the anterior iliac crests, or 1 h after capsule ingestion (whichever was shorter), a radiolabelled meal consisting of 350 kcal (99mTc-sulphur colloid scrambled eggs, one slice of whole-wheat bread and one glass of skim milk) was ingested to measure gastric and small bowel transit. Subsequently, participants ingested a standardised chicken breast lunch (445 kcal) and a roast beef sandwich snack (387 kcal) about 4 and 8 h, respectively, after the first gamma camera image was obtained. Abdominal images (anterior and posterior) were obtained every hour for the first 6 h, and at 8, 24 and 48 h. The images at 24 and 48 h were taken before breakfast after a minimum of 8 h of fasting on the second and third day of imaging.

Transit data analysis

A variable region of interest program was used to quantify the counts in the stomach and in each of the following four colonic regions: ascending (AC), transverse (TC), descending (DC) and combined sigmoid and rectum (RS). These counts were corrected for isotope decay, tissue attenuation and downscatter of 111In counts in the 99mTc window.[16, 18] The gastric emptying T1/2 was estimated by linear interpolation of the hourly images, which has been validated to be accurate.[23] The proportion of 99mTc reaching the colon at 6 h was also estimated as a measure of orocecal transit (as a surrogate for small bowel transit).

Geometric centre (GC) of colonic counts for solids at 4, 8 and 24 h was estimated using geometric mean of anterior and posterior counts in AC, TC, DC, RS and stool (weighted by factors of 1, 2, 3, 4 and 5, respectively, to estimate the geometric centre). The T1/2 of AC emptying was also estimated by plotting the activity-time curve for the per cent residing in the AC; linear interpolation was used to connect the points.

Rectal compliance and sensation thresholds

After completion of the 48-h transit measurements, all participants received a bowel preparation [Fleet Enema (C. B. Fleet Co., Lynchburg, VA, USA), self-administered at least 30 min before test]. A rectal catheter with a polyethylene bag attached (length 22 cm; capacity 600 mL) was inserted into the rectum, approximately 10 cm from the anal verge. To decrease the effects of abdominal viscera on the balloon volume, the patients were placed in a semi-prone position and the foot end of the bed elevated 15°. The intra-rectal balloon was unfolded by inflation with 75 mL of air and later completely deflated. After a 5-min recovery period, the catheter was connected to a barostat (G&J Electronics Inc., Toronto, Ontario, Canada) and an initial ‘conditioning’ distension of the rectum was performed in which the pressure was increased from zero mmHg in steps of 4 mmHg for 15 s per step until 20 mmHg was reached. This ‘conditioning’ distension to 20 mmHg renders subsequent assessments of compliance and perception more reproducible.[24] The bag was then deflated to zero mmHg and participants were allowed to rest for 5 min before proceeding to the ascending method of limits.

Ascending method of limits: compliance and sensory thresholds

Rectal sensory thresholds and compliance were measured by ramp inflation, increasing in steps of 4 mmHg for 1 min per step from zero to a maximum of 60 mmHg. Thresholds for first sensation, gas, urgency and pain were known by the participants pressing a button at the distension pressure at which sensations were perceived. Ramp inflation was terminated as soon as the participant reported the first sensation of pain.[25]

Anal sphincter pressures and rectoanal pressure difference

With the participant in the left lateral position, we used high resolution manometry to measure the anal sphincter pressure at rest and during squeeze, and the rectoanal pressure difference during straining to simulate defecation.

Assessment of stool frequency and consistency

During the entire study, participants completed a daily diary to assess stool frequency and consistency, ease of passage and sense of complete evacuation using the Bristol Stool Form Scale.[26] The diary was dispensed at Visit 1 and collected at the conclusion of the study.

Study endpoints

The primary endpoints were colonic geometric centre at 24 h (GC24) and ascending colon emptying half-time (AC T1/2). Secondary endpoints were colonic geometric center at 4 h (GC4) and 48 h (GC48), colonic filling at 6 h, gastric emptying half-time of solids (GE T1/2), rectal compliance (Pr1/2, where a lower estimated Pr1/2 implies greater compliance and vice versa), rectal sensation thresholds (gas, urgency to defecate and pain), anal sphincter pressure at rest, anal sphincter pressure during squeeze, recto-anal pressure difference during straining to simulate defecation, stool frequency, stool consistency, proportion of incomplete bowel movements and the Patient Assessment of Constipation Quality of Life (PAC-QOL) scores.

Sample size assessment and statistical analysis

The three treatment groups were compared using an analysis of covariance incorporating relevant covariates in the model for specific endpoints (e.g. baseline GC at 24 h for CT, pre-treatment bowel diary responses for daily diary endpoints). Data were summarised as mean ±S.E.M. Table S1 in the Supportive Information section summarises data from previous studies in similar subjects for the primary response measures and uses the (relative) coefficient of variation as a percentage (CV%) to estimate the effect size detectable with 80% power, based on a two-sample t-test at a two-sided alpha level of 0.05, assuming 15 patients per group. It was anticipated that the analysis of covariance (ancova) would provide 80% power to detect similar (pairwise) differences using a pooled estimate of variation across all three groups and potentially even smaller effect sizes by adjusting for important covariates. Based on data acquired in the laboratory, the following effect sizes in the primary endpoints were estimated to be detectable: overall colonic transit at 24 h (33%) and AC emptying T1/2 (48%). This magnitude of change is clinically important.


Participant characteristics

Sixty-nine female patients were initially screened and 24 were excluded from the protocol because: (i) baseline colonic transit GC24 was >2.33; (ii) there was evidence of obstructive defecation; (iii) there were laboratory abnormalities; or (iv) there were scheduling conflicts. Consequently, 15 participants were randomised to each treatment arm. All participants completed the study (Figure 2). There were no clinically importance differences in age, BMI or baseline geometric centre at 24 h (GC24) between the three groups. Baseline demographic characteristics are summarised in Table 1. Compliance with the study medication was nearly complete (% total doses: placebo 96%, TU-100 2.5 g t.d.s. 97%, TU-100 5 g t.d.s. 97%; P > 0.05).

Table 1. Participants' baseline characteristics and overall treatment effects of TU-100 on gastrointestinal and colonic transit
ResponsePlaceboTU-100 2.5 g t.d.s. TU-100 5 g t.d.s.
  1. GC24, geometric centre 24 h; GC48, geometric centre 48 h.

  2. There were no significant differences between the three treatment groups at baseline and no significant overall or dose-related effects were observed for all the transit parameters corresponding to gastric, small bowel and colonic (proximal and overall) transit.

  3. Data are expressed as mean ± S.E.M.; P > 0.05.

N 151515
Age, years44.0 ± 2.243.9 ± 2.243.0 ± 3.1
Body Mass Index, kg/m226.7 ± 1.125.0 ± 0.925.6 ± 1.7
Baseline colonic transit GC24 >1.9445
Baseline colonic transit GC24 <1.7655
Gastric emptying of solids Tt1/2 min120.7 ± 6.0130.5 ± 4.7139.2 ± 14.3
Colonic filling at 6 h, %48.4 ± 7.957.6 ± 6.656.3 ± 7.9
Ascending colon Tt1/2 h19.5 ± 1.818.6 ± 1.920.8 ± 2.8
Colonic transit GC241.93 ± 0.21.77 ± 0.12.02 ± 0.2
Colonic transit GC482.97 ± 0.32.63 ± 0.23.27 ± 0.2
Figure 2.

Study flow chart shows numbers assessed for eligibility to participate, patients excluded and allocated to intervention and analysis. Note that there were no drop-outs from the study and all participants' data were included in the intent-to-treat analysis.

Effects of TU-100 on gastric emptying, colonic filling and ascending colon emptying

The effects of treatment on gastric, small bowel and ascending colonic transit are shown in Table 1. There were no significant overall treatment effects of TU-100 on gastric emptying T1/2 or colonic filling at 6 h or ascending colon emptying T1/2 (Table 1).

Effect of TU-100 on overall colonic transit

The effects of treatment on colonic transit are shown in Table 1. Overall colonic transit at 24 and 48 h was not statistically different among the three treatment groups. In addition, in a post hoc analysis of the subgroup with baseline colonic transit GC24 ≥1.7 (involving 9–10 patients in each treatment group), there were no significant treatment effects identified (data not shown).

Effects of TU-100 on rectal sensation and compliance

Overall, rectal sensation thresholds for first sensation (P = 0.113), gas (P = 0.064), urge (P = 0.301) and pain (P = 0.721) were not significantly different between groups. However, the 5 g t.d.s./day dose was associated with lower sensation thresholds (implying increased sensation) for first sensation and gas (unadjusted P-values: 0.045 and 0.024, respectively). There were no treatment effects on psychosensory symptoms (data not shown). Results are summarised in Table 2.

Table 2. Overall treatment effects of TU-100 on rectal sensation and compliance and anal sphincter pressures
ResponsePlaceboTU-100 2.5 g t.d.s. TU-100 5 g t.d.s.
  1. ASP, anal sphincter pressure.

  2. No significant overall or dose-related effects were observed for all the parameters measured.

  3. Data are expressed as mean ± S.E.M.; P > 0.05.

N 151515
Rectal compliance (Pr½, mmHg)17.4 ± 1.515.2 ± 1.216.1 ± 1.2
Rectal first sensation threshold, mmHg10.0 ± 1.65.3 ± 1.24.8 ± 1.0
Rectal gas threshold, mmHg19.7 ± 2.811.2 ± 1.911.7 ± 1.7
Rectal urge threshold, mmHg23.3 ± 3.315.7 ± 1.616.5 ± 1.9
Rectal pain threshold, mmHg37.7 ± 3.334.1 ± 3.235.2 ± 2.7
Maximum resting ASP, mmHg78.1 ± 9.090.2 ± 11.9106.8 ± 7.3
Maximum squeeze ASP, mmHg155.4 ± 19.4161.7 ± 20.1168.3 ± 13.8
Recto-anal pressure difference−22.8 ± 9.3−49.7 ± 9.2−47.8 ± 6.7

Effects of TU-100 on bowel function and quality of life

Overall, there were no posttreatment effects (Table 3) on average number of daily bowel movements, number of complete bowel movements, stool consistency assessed by Bristol Stool Scale (1–7) or mean proportion of incomplete bowel movements among groups. Similarly, no differences in overall quality of life or quality of life indices for physical discomfort, psychosocial discomfort, satisfaction or worries/concerns were observed between groups. Results are summarised in Table S2 in the Supporting Information section.

Table 3. Effects of TU-100 on number and consistency of daily bowel movements
ResponsePlaceboTU-100 2.5 g t.d.s. TU-100 5 g t.d.s.
  1. BM, bowel movements.

  2. No significant overall or dose-related effects were observed for all the bowel functions assessed.

  3. Data expressed as mean ±S.E.M.; P > 0.05.

  4. a

     Bristol Scale form 1–7.

  5. b

     0,1 scale.

N 151515
Number of BM0.96 ± 0.11.17 ± 0.21.13 ± 0.1
Number of complete BM0.57 ± 0.10.46 ± 0.10.49 ± 0.1
Consistency of BMa3.13 ± 0.23.04 ± 0.33.19 ± 0.3
Mean proportion of incomplete BMb0.39 ± 0.10.52 ± 0.10.55 ± 0.1

Safety and adverse events

TU-100 was safe and well tolerated throughout the study. The most common adverse events recorded in at least two participants in any treatment group were headache, abdominal pain, abdominal cramping, gas production and backache. Overall, there were no significant differences between groups for any of the observed adverse events. Results are summarised in Table S3 in the Supporting Information section.


In this randomised, double-blind, placebo-controlled study, we sought to address the effects of TU-100 on gastrointestinal motor function in female patients with constipation. The study met target for accrual and had adequate power to detect clinically relevant differences between treatment groups; no significant treatment effects were observed for TU-100 on any of the outcomes under study.

Previous literature suggested TU-100 as a potential intestinal prokinetic. Our group observed a promotility effect in small intestine and ascending colon transit in healthy adults treated for 5 days with TU-100 (13). Such a promotility effect was also reproduced in dogs, distal to and at the sites of intraluminal administration,[8] and in guinea pigs.[12, 27, 28] However, in prior human studies, no effects were demonstrated on gastric emptying in healthy adults[13] or on stool frequency or consistency in patients with severe constipation.[29]

Based on the apparent pharmacological action of daikenchuto on serotonergic and cholinergic mechanisms, its extensive use in Japan for relief of abdominal bloating, constipation and intestinal pseudo-obstruction, as well as its favourable safety profile, we sought to further investigate the effects of TU-100 in patients with functional constipation by measuring geometric centre of colonic transit at 24 h and ascending colon emptying half-time. We took the additional precaution of excluding patients with clinical evidence of a rectal evacuation disorder. Delayed emptying of ascending colon transit is observed in patients with severe constipation[30] and in patients with constipation-predominant irritable bowel syndrome (IBS-C).[31] Our current findings do not confirm our prior observation in a study of TU-100 in healthy adults, in which the same medication at a dose of 2.5 g t.d.s decreased ascending colon half-time.[13] It is conceivable that the lack of response in patients with chronic constipation is due to known deficits in colonic innervation or interstitial cells of Cajal (ICCs) in patients with slow transit constipation.[32, 33] Animal models lacking ICCs have reduced gastrointestinal smooth muscle responses to acetylcholine released by excitatory motor neurons.[34] Daikenchuto's actions are mediated through cholinergic pathways, which may be less responsive in patients with slow transit constipation compared with healthy individuals. We considered that the lack of significant effects in our study could be explained by one of five main reasons: lack of power (type II error), placebo effect, low sensitivity of our data collection methods, tachyphylaxis or a true effect.

First, a post hoc power analysis showed that the coefficient of variation (COV) between groups was nearly the same as our predicted COV that led to the proposed sample size of 15 individuals per group; this suggests that lack of power is an unlikely explanation of the lack of response.

Second, we did not observe an important effect of placebo. Thus, the mean values of colonic GC24 at pre-treatment were 1.76, 1.77, 1.84 for the placebo, 2.5 g t.d.s. and 5 g t.d.s. groups respectively. The corresponding (mean) deltas, posttreatment minus pre-treatment for GC24 were, respectively, 0.17, 0.00 and 0.18. The delta of 0.17 GC units at 24 h observed in the placebo group is consistent with the intra-individual variation in colonic transit in patients with irritable bowel syndrome. Thus, Deiteren et al. showed that the mean difference between replicate transit assessments, performed within a 3-week period, was 0.08 ± 0.11 for GC24.[35]

Third, our data collection methods have been extensively studied, standardised and validated for studies evaluating gastrointestinal transit physiology and collection of bowel diaries, as summarised elsewhere[36]; hence, we believe that a low sensitivity of our assessments does not play a role in the observed findings of this study.

Fourth, tachyphylaxis may be a reason to explain the lack of drug effect in our study. Other studies have shown the acute beneficial effect of daikenchuto in patients with post-operative ileus, or the prokinetic effects on ascending colon transit emptying in healthy adults after a shorter period of only 5 days of treatment, during which time tachyphylaxis may not have occurred.

Finally, poor medication compliance was unlikely, as all patients recorded the time of the administration of the study drug daily in a diary, and during each follow-up visit, the participants were specifically asked about compliance with medications and were required to provide the empty packets to maintain a count, which resulted in nearly complete compliance.

Consistent with our previous findings in healthy adults,[13] this study did not show statistically significant changes in stool frequency or consistency, or quality of life in patients with chronic constipation. However, this study was not powered to detect a change in stool frequency or consistency, or quality of life. Other IBS symptoms such as pain and bloating were not addressed in our study. There is an ongoing clinical trial to address the effect of TU-100 in functional constipation using the Constipation Severity Instrument (CSI, NCT01348152).[37]

In addition, our study appraised rectal sensation thresholds, but not sensory ratings in response to barostat-controlled rectal distensions. The association of effects on rectal sensation thresholds with prediction of effects on visceral hypersensitivity is imperfect, with higher coefficients of variation with estimates by the ascending limits of threshold for sensation method compared with sensory ratings in response to standardised distension pressures.[38] It is possible that the ascending limits of threshold for sensation method may be associated with response bias, where participants report their sensation at each of the escalating pressure levels. Thus, the current observations do not support a potential benefit of the drug on sensory mechanisms. In fact, the 5 g t.d.s. dose was associated with lower sensation thresholds for first sensation and gas (unadjusted P-values: 0.045 and 0.024, respectively), which suggests an increase in rectal sensation.

Our findings are limited to female patients with functional constipation, and future studies of TU-100 should include sensation ratings in response to distension and address whether it is beneficial for patients with IBS symptoms of pain and/or abdominal bloating or for patients with documented visceral hypersensitivity.

In summary, in this pharmacodynamic study, there were no significant treatment effects of TU-100 at 2.5 g t.d.s. or 5 g t.d.s dosages on gastrointestinal or colonic transit, rectal compliance, thresholds for sensation and anal sphincter pressures. The mechanisms underlying the clinically observed benefit of this Japanese traditional medication remain unclear.


Guarantor of the article: Dr Michael Camilleri takes responsibility for the integrity of the work as a whole, from inception to published article.

Author contributions: M. Camilleri contributed to the study conceptualisation, writing of protocol and manuscript; J. Iturrino is a research fellow who contributed to the recruitment, writing protocol and manuscript; B. S. Wong is a research fellow who contributed to the recruitment, writing protocol and manuscript; S. J. Linker Nord performed the participant recruitment and motility studies; D. Burton performed motility studies; A. R. Zinsmeister contributed to the database management and statistical analysis. All authors approved the final version of the manuscript.


Declaration of personal interests: In the past 2 years, Dr Camilleri has consulted for Salix, Ikaria, Ferring, NPS Pharmaceuticals, Shire, Theravance, Ironwood, Ocera, Tranzyme, AstraZeneca and Novartis.

Declaration of funding interests: This study was funded in full by Tsumura USA (Princeton, NJ, USA). The study was conducted in the Clinical Research Unit of the Mayo Clinic CTSA (grant UL1 TR000135). This study is registered in ClinicalTrials.gov with the identifier NCT01139216.