Randomized clinical trial: A phase 2b controlled study of the efficacy and safety of trazpiroben (TAK‐906) for idiopathic or diabetic gastroparesis

Previous clinical studies of trazpiroben, a dopamine D2/D3 receptor antagonist for long‐term treatment of moderate‐to‐severe idiopathic and diabetic gastroparesis, have shown improved symptoms of fullness. This study assessed trazpiroben efficacy, safety, and tolerability in adults with idiopathic and diabetic gastroparesis versus placebo.


| INTRODUC TI ON
Gastroparesis, a chronic gastric motility disorder, is characterized by delayed gastric emptying without mechanical obstruction.[5] Despite the substantial patient burden associated with gastroparesis, therapies for the disease are limited.Treatment options include dietary therapy, antiemetics, dopamine-2 (D 2 ) receptor antagonists, and 5-HT 4 receptor agonists. 4,6D 2 receptor antagonists offer benefit in gastroparesis through muscular D 2 receptor blockade in the stomach, preventing the direct relaxant effect of dopamine and promoting gastric motility. 7,8D 2 receptor antagonist therapies may also alleviate nausea via action on the chemoreceptor trigger zone. 9toclopramide is a peripherally and centrally acting D 2 receptor antagonist providing relief of nausea and vomiting owing to its central effects in the chemoreceptor trigger zone, and also accelerating gastric emptying by blocking dopamine receptors in the stomach and proximal duodenum, thus facilitating additional acetylcholine release and accentuating upper gut motility. 8,10,11Domperidone is a peripherally acting D 2 receptor antagonist that decreases nausea and increases gastric emptying rates; it is approved by the European Medicines Agency for short-term use to treat nausea and vomiting, but is not approved by the US Food and Drug Administration (FDA). 4,12[15] Trazpiroben (previously known as TAK-906 or ATC-1906M) is a D 2 /D 3 receptor antagonist.Previous in vitro and animal studies demonstrated that trazpiroben selectively antagonizes D 2 /D 3 receptors with limited central nervous system penetrance, and also shows low affinity for the hERG potassium channel. 16,179][20][21][22] A phase 2a study (NCT03268941) assessed the safety profile, pharmacokinetics (PK), and pharmacodynamics of trazpiroben in patients with gastroparesis, with pharmacodynamics evaluated by the effect of trazpiroben on prolactin levels and gastric function. 23This phase 2a study showed significantly improved volume-to-fullness following the nutrient drink test, and improved American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary (ANMS GCSI-DD) symptom scores with trazpiroben, further supporting its development to treat gastroparesis. 23e phase 2b trial described herein assessed trazpiroben efficacy in adults with IG and DG versus placebo at 12 weeks, in addition to evaluating its safety and tolerability.The study also evaluated trazpiroben PK in participants with gastroparesis, and treatment effect on patient-reported outcomes (e.g., symptoms and healthrelated quality of life [HRQoL] vs. placebo).

| Study design
This was a global, multicenter, randomized, double-blind, placebocontrolled, parallel-group, phase 2b study in adults with symptomatic IG or DG conducted at 73 study sites: 45 in the USA, 9 in Europe, and 19 in Japan (Figure 1).
The study included a screening period (≤5 weeks), during which gastric motility and gastroparesis symptoms were assessed, and a 12week treatment period.A 12-week treatment period was selected in line with FDA guidance recommending a minimum 12-week duration to assess treatment effect and collect safety data. 24Following the first visit during the screening period, participants taking prescribed gastrointestinal motility medications were instructed to stop these therapies during a 2-week washout period, before undergoing a gastric emptying breath test (GEBT) performed at the second visit.The GEBT was required for all participants prior to a protocol amendment in 2020; post-amendment, a GEBT was not required for participants with previously confirmed delayed gastric emptying by an accepted diagnostic test.If required to rule out gastric outlet obstruction, an upper endoscopy was performed during screening.At the third visit, participants were provided with and trained to use the electronic patient-reported outcomes (ePRO) hand-held device, which delivered the ANMS GCSI-DD in an electronic daily diary format for participant completion.A 2-week run-in symptom assessment period was then conducted before randomization to allow participants to become proficient at completing the ePRO tool to record daily symptoms, and to collect baseline symptom data for assessment of participant suitability for randomization.During treatment, participants were required to complete the ANMS GCSI-DD each day and to return for site visits at weeks 4, 8, and 12. Follow-up safety phone calls were then conducted 30 days after the last treatment dose.The

Key Points
• Trazpiroben (previously known as TAK-906 or ATC-1906M) is a dopamine D2/D3 receptor antagonist that has previously been investigated for long-term treatment of moderate-to-severe idiopathic and diabetic gastroparesis, with studies showing improved symptoms of fullness.
• In this study, there was no clinically meaningful difference in efficacy between trazpiroben and placebo in treating gastroparesis, based on change from baseline in weekly average of the American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary composite score at Week 12.
• Trazpiroben was well tolerated with no new safety concerns identified, strengthening evidence supporting its favorable safety profile.

| Participant population
Adults aged 18-85 years with a body mass index of ≥18 to ≤40 kg/ m 2 and with symptomatic IG or DG were eligible for inclusion.
Participants were required to have: (1) gastroparesis symptoms for ≥3 months before screening, as assessed by a physician; (2) delayed gastric emptying confirmed by an accepted diagnostic testing method (either GEBT, gastric emptying scintigraphy, or wireless motility capsule) documented either before or during screening; and (3) an average ANMS GCSI-DD composite score (described under the Study outcome measures section) of ≥2 during the 7 days before randomization, and symptomatic nausea prior to randomization (either a nausea subscale score ≥2 on at least 4 of the 7 days, or an average nausea subscale score ≥2 during the 7 days).Participants were excluded if their predominant symptom was epigastric pain, diffuse abdominal pain, or pain associated with bowel movement, as F I G U R E 1 Study design.a Randomization into the trazpiroben 5 mg arm was discontinued in July 2020 to improve operational feasibility in the setting of the COVID-19 pandemic, and to reduce the number of participants exposed to a potentially minimally efficacious dose that likely approaches the lower end of the dose response curve.Participants attended the study site for a screening/consent visit (visit 1, day −35) when written informed consent was obtained and their general study eligibility was reviewed.After signing, participants had a blood sample taken to assess their laboratory values.Participants discontinued all excluded medications and returned to the site 3-17 days later for a 4-h 13 C-spirulina GEBT (visit 2, day −21).Participants who already had confirmed delayed gastric emptying were not required to undergo the visit 2 GEBT and proceeded directly to visit 3 for instruction on use of the ePRO tool (the ANMS GCSI-DD and a bloating severity scale).Participants not taking any medications at screening that required washout may have attended the clinic for visit 2 after laboratory results were reviewed by the investigator.Participants then completed visit 3 (day −14) in the 7 days after the GEBT visit to confirm their eligibility based on gastric emptying criteria.At visit 3, participants with confirmed delayed gastric emptying were instructed on using the ePRO tool.Participants recorded their symptoms once daily (in the evening) and returned to the clinic approximately 2 weeks later for assessment of their gastroparesis symptom eligibility based on the ANMS GCSI-DD (visit 4, randomization).Participants took their first dose of study drug in clinic during the randomization visit (day 1).On days when participants attended clinic (visits 5, 6, and 7), the morning dose was taken in clinic.Electrocardiogram measures, clinical laboratory parameters, vital signs, and adverse events were assessed, and blood sampling for evaluation of trazpiroben and prolactin pharmacokinetics were performed at visits 5, 6, and 7. ANMS GCSI-DD, American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary; EOT, end of treatment; ePRO, electronic patient-reported outcomes; GEBT, gastric emptying breath test.determined by the investigator.Participants also needed to comply with ANMS GCSI-DD completion, defined as ≥80% diary completions during the 2-week run-in symptom assessment period.For those with DG, a glycosylated hemoglobin (HbA1c; measured in percentage of total hemoglobin) level of ≤11% was required before randomization, and blood glucose was monitored throughout the study using this measure.Full inclusion and exclusion criteria are provided in Table S1.
Rescue medication use during the 12-week treatment period was monitored, given the potential impact on symptoms, and limited to twice daily for ≤3 days per week.Ondansetron 4 mg taken orally was the preferred rescue medication, but alternatives such as diphenhydramine, dimenhydrinate, or promethazine could be provided.If a participant used the maximum permitted rescue medication for two consecutive weeks, they were discontinued from the study owing to lack of study drug efficacy.

| Study outcome measures
The ANMS GCSI-DD 25 composite score was used to assess the primary efficacy endpoint, with ANMS GCSI-DD total and core symptom scores also used in this analysis.Patient symptom severity and HRQoL were assessed using the patient assessment of upper gastrointestinal disorders-symptom severity index (PAGI-SYM) 26 and patient assessment of upper gastrointestinal quality-of-life index (PAGI-Q oL), 27 as well as the 12-item short-form health survey (SF-12). 28All measures are described in Table 1.

| Efficacy endpoints
The

| Safety and tolerability analysis
Treatment-emergent adverse events (TEAEs), vital signs, 12-lead electrocardiograms, and clinical laboratory parameters (hematology, clinical chemistry, and urinalysis) were assessed as part of the safety analysis.To this end, TEAEs were assessed throughout the study and other safety measures were evaluated at day 1, week 4, week 8, and week 12.

| Pharmacokinetic and pharmacodynamic analysis
Plasma samples to determine trazpiroben and prolactin levels were taken at day 1, week 4, week 8, and week 12. Concentration data were used to characterize trazpiroben PK in this population using a population PK (popPK) model, and to conduct an exploratory exposure-response analysis using the week 12 ANMS GCSI-DD composite score as the response variable (Text S2).Prolactin levels, as a marker of peripheral D 2 receptor antagonism, were measured to determine target engagement.

| Health-related quality of life analysis
To assess participant HRQoL, changes from baseline to week 12 in PAGI-QoL total score and SF-12 subscale scores were assessed.

| Statistical methods and analysis
Categorical data were summarized as the number and percentage of participants in each category.Continuous data were descriptively summarized.All statistical analyses were performed using SAS System version 9.4 or a later version, and change from baseline in weekly ANMS GCSI-DD composite score was analyzed using a mixed-effects model for repeated measures (MMRM).
The model used treatment, week, participant disease population (IG, DG), and treatment-by-week interaction as fixed effects, and baseline and baseline score by-week interaction as covariates, with an unstructured variance covariance matrix.The estimates of the least-squares mean (LSM) difference between trazpiroben dose groups and placebo at each week and their 90% confidence intervals were reported.In particular, the LSM differences between trazpiroben dose groups and placebo at week 12 were used to evaluate treatment effects.Estimates of the LSM difference between trazpiroben arms and placebo at each week and their 90% confidence intervals were reported, with significance defined as one-sided p < 0.05.Additional details of handling of missing data are provided in Text S3.

| Sample size determination
A standard deviation (SD) of 1.25, a difference of 0.625 between trazpiroben and placebo in the change from baseline in the ANMS GCSI-DD composite score at week 12, and a 15% dropout rate were assumed, based on existing literature regarding gastroparesis trials using this measure. 23,25Given these assumptions, a total of 205 participants (25 in the trazpiroben 5 mg arm following protocol amendment in response to the COVID-19 pandemic, and 60 in each of the other study arms) was calculated to be sufficient to achieve approximately 80% power to detect the treatment effect of 0.625 at a 1-sided 5% significance level.• Consists of four patient-reported cardinal symptoms (nausea, early satiety, postprandial fullness, and upper abdominal pain), and is calculated by summing scores for these symptoms and then dividing by four • For each symptom, the peak intensity in the previous 24 h is reported using a five-point Likert scale, with higher scores reflecting more severe symptoms • Also includes number of vomiting episodes during the past 24 h as a continuous variable The ANMS GCSI-DD was completed by participants using a diary ePRO tool and was collected daily throughout the entire treatment period ANMS GCSI-DD 25 core symptom score • Includes scores for all five symptoms (four cardinal symptoms included in the composite score, plus number of vomiting episodes) • For each symptom, the peak intensity in the previous 24 h is reported using a five-point Likert scale, with higher scores reflecting more severe symptoms • Also includes number of vomiting episodes during the past 24 h as a continuous variable (capped at four episodes) ANMS GCSI-DD 25 total score median, greater than the median), and geographical region (North America, Europe, Asia).

| Pre-specified sensitivity analyses for the primary efficacy endpoint
Overall, three sensitivity analyses were performed: two were conducted to deal with participants receiving rescue therapy during the treatment period, and one to deal with those with missing values.
Regarding rescue therapy, one sensitivity analysis was conducted by considering ANMS GCSI-DD scores reported on and after start of rescue medication through 48 h after the end date of rescue medication as missing at random (MAR).Rescue medication was assumed to have no effect on ANMS GCSI-DD scores 48 h after the end date of its administration.Another sensitivity analysis was performed by imputing of ANMS GCSI-DD scores reported on and after start of rescue medication through 48 h after the end date of rescue medication by considering the nausea symptom level as the worst score.
Rescue medication was assumed to primarily affect nausea or vomiting symptoms and to have no effect on these symptoms 48 h after the end date of rescue medication.For participants with missing values, a pattern mixture model with the control-based imputation was applied to handle missing values as a sensitivity analysis of the MAR assumption in the primary analysis previously described.

| Post hoc subgroup analysis of PAGI-SYM nausea/vomiting score
A post hoc subgroup analysis of the PAGI-SYM nausea/vomiting score by participant disease population (IG and DG) was also conducted using a similar MMRM to the primary efficacy analysis.

| Participant demographics and baseline characteristics
Eligible participants were enrolled and followed up from October For the 242 participants included for analysis, mean (SD) age was 55.7 (14.2) years, and the majority were female (75.6%;Table 2).
Just over half of participants (54.5%) had received a prior diagnosis of gastroparesis.During the treatment period, 40.1% of participants received at least one dose of rescue medication, most commonly ondansetron (35.1% of participants).

| Efficacy analysis
For the primary efficacy analysis, no significant differences in respectively; Table 3).A high placebo response (PR) was observed.
Absolute weekly composite scores decreased over study for all treatment groups with no meaningful differences noted between groups (Figure 3).Given that the primary endpoint was not met, all presented p values are at the nominal level.
When considering the secondary efficacy analysis, no clinically meaningful difference between study arms was observed for mean change from baseline in ANMS GCSI-DD weekly nausea symptom score, or for other individual weekly symptom scores (Figure 4).
Similarly, no meaningful difference in treatment effect was seen for change from baseline in weekly ANMS GCSI-DD total scores (Table S2), overall severity of gastroparesis scores or bloating sever-

| Pharmacokinetic analysis
PopPK modeling and exposure-response analysis results are provided in Text S5, Table S4, Figures S1, and S2.Serum prolactin levels increased in response to trazpiroben treatment with higher prolactin levels observed with 25 and 50 mg dosing than in the 5 mg and placebo arms (Figure 5).).Similarly, no significant differences were seen for SF-12 physical or mental component summary scores between participants receiving trazpiroben or placebo (Table S5).Note: Change from baseline in weekly ANMS GCSI-DD composite score was analyzed using a mixed-effects model for repeated measures.Estimates of the LSM difference between trazpiroben arms and placebo at each week and their 90% CIs were reported, with significance defined as one-sided p < 0.05.

| Health-related quality of life analysis
Abbreviations: ANMS GCSI-DD, American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary; CI, confidence interval; LSM, least-squares mean; SD, standard deviation; SE, standard error.S6).In addition, the IG subgroup comprised a higher proportion of females (81.1%) and had a lower diabetes with available data).For patients with diabetes receiving placebo, mean (SD) baseline HbA1c levels were 7.04% (1.42%) compared with 7.33% (1.48%) at week 12.For all patients receiving trazpiroben, mean (SD) HbA1c levels at baseline were 7.05% (1.54%) Trends in weekly nausea scores were consistent with weekly composite scores for both disease populations, with slight improvements observed in patients with IG but not those with DG.

| Subgroup demographics and pre-specified analyses for the primary efficacy endpoint
For participants receiving trazpiroben, reductions in weekly ANMS GCSI-DD composite scores at week 12 were slightly higher in women (5, 25, and 50 mg: LSM [SE] −1.17 Note: prolactin was used as a measure of trazpiroben target engagement.LLN, lower limit of normal; ULN, upper limit of normal.S7).No significant differences were observed with respect to baseline composite scores below or equal to the median or greater than the median when investigating treatment effect, although participants receiving trazpiroben with baseline ANMS GCSI-DD scores below or equal to the median showed smaller composite score reductions from baseline to week 12 versus placebo (Table S7).For European participants, a reduction in composite score from baseline to week 12 was observed compared with placebo with trazpiroben 25 mg (difference in LSM [SE] vs. placebo −0.99 [0.53];Table S7).

| Pre-specified sensitivity analyses for the primary efficacy endpoint
The results of the sensitivity analyses dealing with rescue medication use and the handling of missing data were consistent with the main efficacy analysis, with no statistically significant differences in changes from baseline at week 12 in weekly ANMS GCSI-DD composite scores observed between placebo and trazpiroben treatment arms.

| DISCUSS ION
In this phase 2b study of trazpiroben in adults with symptomatic IG or DG, the primary and secondary efficacy endpoints were not met.
Improvement in both primary and secondary outcome measures A high PR was observed in this study, which refers to a therapeutic outcome following an inactive treatment. 29While ANMS GCSI-DD results did not indicate any substantial difference between placebo and trazpiroben treatment, total PAGI-SYM scores showed a small improvement versus placebo at weeks 4 and 8 in the trazpiroben 50 mg group (one-sided p < 0.05).However, no considerable difference in PAGI-SYM results between trazpiroben and placebo was present at week 12 (one-sided p > 0.05).GEBT t 1/2 also showed small numeric improvements in change from baseline to week 12 with trazpiroben 25 and 50 mg compared with placebo, but the effect on gastric emptying was unlikely to be clinically meaningful.Similarly, a high PR was observed throughout other subgroup analyses.Thus, selected exploratory outcome variables suggest some clinical improvement with trazpiroben.
In trials showing high PR, it can be more difficult to successfully differentiate between efficacious and non-efficacious treatments. 29,306][37] The observed PR was also consistent with previous studies in functional dyspepsia and other functional gastrointestinal disorders, with a prior review reporting a mean PR of 40% across 45 trials in these conditions. 38,39The design and interpretation of results from clinical trials in gastroparesis are frequently complicated by this PR, which is difficult to mitigate, and patient-reported outcomes are typically associated with higher PR rates than biomarker endpoints. 39Across gastrointestinal diseases, spontaneous waxing, and waning of symptoms is estimated to explain up to 50% of the PR, adding further complexity in evaluating treatment efficacy and potentially leading to overestimation of PR rates. 39bgroup analysis of ANMS GCSI-DD data by disease population indicated a trend towards efficacy of trazpiroben compared with placebo in participants with IG, but not in those with DG.This difference appeared to be driven by both the small improvement observed in participants with IG, as well as the larger PR observed in participants with DG versus those with IG.Similar findings were observed in a phase 2b trial of velusetrag in patients with gastroparesis, with a correspondingly smaller treatment effect noted in those with DG. 40 The prokinetic prucalopride has also failed to show meaningful benefit versus placebo in patients with DG. 41 The reasons for the greater PR observed in patients with DG in the aforementioned trials are unclear, but are likely to be multifactorial. 39Participants with DG were also observed to have a greater PR versus those with IG when considering PAGI-SYM total scores, as well as in a post hoc analysis assessing the nausea/vomiting score component of PAGI-SYM.
Given the mechanism of action of trazpiroben, it is notable that no significant difference was seen in ANMS GCSI-DD nausea symptom scores between trazpiroben and placebo groups, which may also reflect the high PR observed in this study. | The findings of this study, and the role played by PR, raise several important considerations regarding trial design and selection of participants for future clinical studies in gastroparesis.Early identification of participants most prone to PR can be challenging and, while no specific strategy has been consistently demonstrated to reduce the risk of a high PR, some study design characteristics may be considered to reduce inclusion of those most susceptible.An extension of the baseline period could potentially reduce PR, by allowing evaluation of symptom stability when assessing participants for inclusion, as well as longer periods between study visits.However, in a clinical trial in patients with irritable bowel syndrome employing a single-blinded run-in period and 5-week intervals between study visits, a high PR was observed despite the design features intended to mitigate this effect. 42Patient training to allow reliable assessment and recording of the ANMS GCSI-DD measure using an ePRO tool and reducing the patient burden of daily data entry may also help to lessen PR, in addition to proactively identifying patients likely to respond regardless of treatment, managing patients' expectations, and improved management and recording of patient rescue medication during future trials.Patients with more severe symptoms could also be considered for inclusion, given that low symptom severity at baseline has been noted as a predictor of higher PR in prior trials in functional gastrointestinal disease; this could also be considered during the design of study inclusion criteria. 29,39While few predictors are known to consistently determine the size of the PR, aspects such as the duration of the run-in period, moderate versus severe symptom burden, duration of gastroparesis symptoms, and previous patient treatment history may be taken into account during future study design.Consideration of these factors may aid mitigation of the PR effect and allow clearer evaluation of treatment efficacy in this disease area.
In addition, use of a robust, validated endpoint is important because patient-reported outcomes usually produce higher PR rates than objective laboratory measures such as specific biomarkers. 39ta from this analysis were used to perform a comprehensive psychometric evaluation of the ANMS GCSI-DD, used as the primary efficacy endpoint measure in the current study, along with trial exit interviews that aimed to assess participant perception of meaningful change with respect to disease symptoms. 25,43This psychometric analysis provided additional validation that the ANMS GCSI-DD represents a fit-for-purpose, reliable, and well-defined measure of overall symptom severity in patients with gastroparesis as previously shown. 25mitations of this study must be considered.A high PR rate was observed during the primary and secondary efficacy analyses, probably leading to the associated endpoints not being met.The

| 3 of 15 TACK
et al. total study duration was approximately 20 weeks from the start of the screening period to the post-treatment final safety follow-up phone call.Participants were enrolled by clinicians at study sites and randomized 1:1:1:1 into one of four arms: oral placebo or trazpiroben 5, 25, or 50 mg twice daily (once in the morning and once in the evening, approximately 1 h before meals).Randomization was stratified by IG and DG.Owing to the COVID-19 pandemic, the study protocol was amended and randomization into the trazpiroben 5 mg arm was discontinued on July 13, 2020.This change was implemented to improve operational feasibility and reduce the number of participants exposed to a potentially minimally efficacious dose.Following this amendment, eligible participants were randomized 1:1:1 into the placebo or trazpiroben 25 mg or 50 mg group.Details of blinding are provided in Text S1.This study was conducted in accordance with the Declaration of Helsinki, the International Conference on Harmonisation Good Clinical Practice recommendations, and all applicable regulations.All participants agreed to participate in the study by providing written informed consent.The study is registered with the ClincialTrials.gov registry (https://clini caltr ials.gov/ct2/show/NCT03544229; ClinicalTrials.govIdentifier: NCT03544229).
primary efficacy endpoint was change in weekly mean ANMS GCSI-DD composite score (an average of nausea, early satiety, upper abdominal pain, and postprandial fullness severity scores) from baseline to week 12 of the treatment period.Secondary efficacy endpoints included change from baseline to week 12 in ANMS GCSI-DD symptom scores for nausea, early satiety, postprandial fullness, and upper abdominal pain, as well as recorded vomiting frequency.Changes from baseline to week 12 in ANMS GCSI-DD overall severity of gastroparesis symptoms and ANMS GCSI-DD total scores, plus bloating severity scale score and PAGI-SYM total score, were also assessed.Additional secondary efficacy endpoints were the proportion of participants with a ≥50% reduction from baseline in ANMS GCSI-DD composite score at week 12, and the proportion of symptomatic weeks (weeks with mean composite ANMS GCSI-DD symptom score ≥2 [assessed as worse than mild] during 12 weeks of treatment).Additional efficacy endpoints included the percentage of symptomatic days, when symptoms were assessed as ANMS GCSI-DD score ≥2 (worse than mild) during 12 weeks of treatment, and the proportion of participants with a ≥20% reduction in ANMS GCSI-DD composite score from baseline in the last 6 consecutive weeks of the study.Change from baseline to week 12 in ANMS GCSI-DD core symptom score (nausea, early satiety, upper abdominal pain, postprandial fullness, and vomiting) and nauseafree days were also measured, in addition to change from baseline to week 4 and week 12 in GEBT time to gastric half-emptying (GEBT t 1/2 ).

2. 8 . 2 |
Pre-specified subgroup analyses for the primary efficacy endpoint Subgroup analyses were performed using a similar MMRM to the primary efficacy analysis, and treatment groups were compared within each subgroup.Treatment groups were compared by sex (female, male), disease population (IG, DG), age (≤65 years, >65 years [>65 years represents the threshold at which patients were considered geriatric]), baseline composite score (below or equal to the TA B L E 1 Study outcome measures.Outcome measure Details of measure When measure was assessed ANMS GCSI-DD 25 composite score

Figure 2 )
Figure2).The study finished at the end of planned follow-up for the final participant.
change from baseline in weekly ANMS GCSI-DD composite score at week 12 were seen between placebo (LSM [standard error; SE] −1.19 [0.12]) and any trazpiroben dosing group (trazpiroben 5, 25, and 50 mg: LSM [SE] −1.11 [0.22], −1.17 [0.12], and −1.21 [0.12], scale scores.Details of PAGI-SYM total scores are provided in Text S4.Total PAGI-SYM scores showed small numerical improvements versus placebo at weeks 4 and 8 in the trazpiroben 50 mg group (LSM [SE] −0.31 [0.16] and −0.29 [0.16], respectively; onesided p < 0.05).However, no considerable difference in PAGI-SYM score between trazpiroben and placebo was present at week 12 (one-sided p > 0.05).For additional efficacy endpoints evaluated during the study, no meaningful differences were observed between treatment groups in: the proportion of participants with a ≥ 50% reduction from baseline in weekly ANMS GCSI-DD composite score at week 12; the proportion of symptomatic weeks during 12 weeks of treatment; the percentage of symptomatic days with more than "mild" symptoms using the ANMS GCSI-DD during 12 weeks of treatment; or the percentage of participants with a ≥ 20% reduction from baseline in the ANMS GCSI-DD composite score in the last six consecutive study weeks.Additionally, no treatment effect was noted with respect to change from baseline to week 12 in weekly ANMS GCSI-DD core symptom scores.Nausea-free days increased from baseline (trazpiroben 5, 25, and 50 mg: mean [SD] 0.

F I G U R E 2
No significant difference was observed for the change from baseline to week 12 in PAGI-QoL total scores between participants receiving trazpiroben (5, 25, and 50 mg: LSM [SE] −0.53 [0.21], −0.77 [0.11], and −0.74 [0.10], respectively) or placebo (LSM [SE] −0.66 [0.11]).With respect to PAGI-QoL psychological wellbeing subscale scores, change Participant flow.aOne participant was screened twice with two different participant identification numbers and was mistakenly not identified during electronic data capture, because a rescreening with a prior participant identification number was noted; therefore, the total number of participants screened was 795 rather than 796.The participant was randomized following the second screening.b Percentages are based on the number of participants with a screening failure.c Percentages are based on all participants in the randomized set within each column.

TA B L E 2
Overall participant demographics.

For
participants with IG (n = 127) and DG (n = 115), those with IG were slightly younger (mean [SD] age: 52.2 [15.0] years) than those with DG (59.5 [12.3] years; Table mean (SD) body mass index(27.2[5.5] kg/m 2 ) than the DG subgroup (69.6% female and mean[SD] body mass index of 30.4[4.6] kg/m 2 ).For participants with diabetes, observed HbA1c levels did not substantially change over the course of the study (mean [SD] change from baseline at week 12: 0.15% [0.78%] across all individuals with F I G U R E 3 ANMS GCSI-DD weekly composite scores over time.(A) Change from baseline in ANMS GCSI-DD Weekly Composite Score, (B) Absolute ANMS GCSI-DD score.The ANMS GCSI-DD composite score can range from 0 to 4, with higher scores reflecting greater symptom severity.ANMS GCSI-DD, American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary; LS, least-squares; SD, standard deviation; SE, standard error. (A) Change from baseline in ANMS GCSI-DD Weekly Composite Score (B) Absolute ANMS GCSI-from baseline in ANMS GCSI-DD nausea score (LS mean, SE) from baseline in ANMS GCSI-DD early satiety score (LS mean, SE) in ANMS GCSI-DD postprandial fullness score (LS mean, SE) in ANMS GCSI-DD upper abdominal pain score (LS mean, SE)

F I G U R E 4 F I G U R E 5
[0.26], −1.19 [0.14], and −1.21 [0.15], respectively) than in men (5, 25, and 50 mg: LSM [SE] −0.97 [0.46], −1.15 [0.28], and −1.15 [0.19], respectively), although there was no significant difference compared with placebo observed for either sex.Difference in age did not appear to lead to substantial differences compared with placebo for changes from baseline to week 12; although for participants aged >65 years, smaller reductions in ANMS GCSI-DD composite score from baseline to week 12 were observed versus placebo for trazpiroben 25 and 50 mg (difference in LSM [SE] vs. placebo 0.66 [0.29] and 0.24 [0.30], respectively; Change from baseline in ANMS GCSI-DD weekly symptom scores over the study period for (A) Nausea, (B) Early satiety, (C) Postprandial fullness, (D) Upper abdominal pain, and (E) Recorded vomiting frequency.The ANMS GCSI-DD symptom scores can range from 0 to 4, with higher scores reflecting greater symptom severity.An ANMS GCSI-DD recorded vomiting frequency greater than or equal to 60 per day was deemed clinically implausible and set to missing for analysis.ANMS GCSI-DD, American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary; LS, least-squares; SE, standard error.Prolactin levels during the treatment period.(A) Prolactin concentration, (B) Change from baseline in prolactin concentration.
Post hoc subgroup analysis of PAGI-SYM nausea/vomiting score Details of the PAGI-SYM nausea/vomiting score for different treatment groups and disease subgroups are provided in Text S4, Tables S8 and S9.Participants with IG showed reductions in the PAGI-SYM nausea/vomiting subscale score with trazpiroben treatment (mean [SD] reductions of −0.82 [0.35; one-sided p = 0.01] and −0.62 [0.34; one-sided p = 0.04], respectively, for trazpiroben 25 and 50 mg doses vs. placebo), whereas individuals with DG did not.

over the 12 -
week course of the study was observed in all study arms (including placebo), although the magnitude of improvement was similar among participants receiving placebo or active treatment.Overall, trazpiroben was well tolerated by study participants, with no new safety concerns identified.No significant differences were noted in changes in HRQoL scores across groups, although some changes from baseline versus placebo were observed at earlier timepoints in the study (e.g., at week 4) in a few PAGI-QoL domain subscales (e.g., trazpiroben 50 mg for the relationship and psychological wellbeing subscales, data not shown).Following administration of trazpiroben, a D 2 /D 3 receptor antagonist, increases in prolactin levels observed with trazpiroben 25 and 50 mg doses were consistent with dopamine antagonism suggesting target engagement, and PK data showed that adequate trazpiroben concentrations were attained in these treatment groups.Although acceptable target engagement was seen with trazpiroben 25 and 50 mg doses, the ANMS GCSI-DD composite score showed no clear efficacy signal.
study was not powered to detect efficacy in subgroup analyses, although some nominal trends were visible.Recruitment into the trazpiroben 5 mg arm was discontinued owing to the COVID-19 pandemic to reduce exposure of participants to a potentially minimally efficacious dose and to manage the likelihood of statistical, clinical, and operational success of this study.The study protocol was also amended to no longer require a GEBT prior to screening if patients had a prior delayed gastric emptying test, although some studies have suggested that gastric emptying may change over time in patients with gastroparesis.44 Nonetheless, this study highlights several important considerations regarding participant selection and study design that may help to mitigate the PR as far as possible, to gain a clearer picture of treatment efficacy in gastroparesis in future investigations.In conclusion, in participants with moderate-to-severe IG or DG, trazpiroben did not show meaningful efficacy in the treatment of gastroparesis versus placebo.However, trazpiroben was well tolerated with no new safety concerns identified, adding to prior literature supporting its favorable safety profile.While the primary endpoint was not met, some exploratory measures indicated improvement in symptoms of nausea and vomiting, particularly in patients with IG.From the current study there arise several key suggestions for design of future clinical trials and selection of participants to help minimize high PR, such as extension of the study baseline period and requirements for more severe disease as part of patient eligibility criteria.AUTH O R CO NTR I B UTI O N S Henry P. Parkman acts as the guarantor for this manuscript.Jan Tack, Richard McCallum, Braden Kuo, Susanna Y. Huh, Yaozhu J. Chen, and Henry P. Parkman were involved in study conceptualization, and Yaozhu J. Chen was involved in curation of the study data.Susanna Y. Huh, Yanwei Zhang, Yaozhu J. Chen, and Shailly Mehrotra performed the formal data analysis, and Braden Kuo, Susanna Y. Huh, Yanwei Zhang, Yaozhu J. Chen, and Shailly Mehrotra were involved in development or design of study methodology.Susanna Y. Huh performed project administration activities for the study.Yaozhu J. Chen was involved in validation of the study, and Jan Tack, Richard McCallum, Susanna Y. Huh, and Henry P. Parkman were involved in visualization of the manuscript.Susanna Y. Huh was involved in preparing the original draft, and all authors were involved in the review and editing of the manuscript.FU N D I N G I N FO R M ATI O NNo funding declared.ACK N OWLED G M ENTSThe authors thank the patients who participated in the trial, their caregivers, and the trial investigators and staff, members of the data and safety monitoring board, and members of the clinical trial team.We also thank Irene Sarosiek as Research Supervisor for the Texas Tech El Paso Team, and Theresa White, Min Jung Yoon, Yueping Zhu, Decai Deng, Jin Qian, and Kirtana Sripathi of Takeda Development Center Americas, Inc. for contributing their expertize and support.All authors had access to the data supporting this publication, and approved the final version of the article for submission.This study was sponsored by Takeda Development Center Americas, Inc. Medical writing assistance was provided by Alexandra Kisbey-Ascott of Oxford PharmaGenesis, Oxford, UK and was supported by Takeda Development Center Americas, Inc.
• Includes scores for all five symptoms (four cardinal symptoms included in the composite score, plus number of vomiting episodes) plus a bloating severity scale score • For each symptom, the peak intensity in the previous 24 h is reported using a five-point Likert scale, with higher scores reflecting more severe symptoms • Also includes number of vomiting episodes during the past 24 h as a continuous variable (capped at four episodes) 28GI-SYM26measureUses a six-point Likert response scale to measure symptom severity in patients with upper gastrointestinal disorders, with a recall period of 2 weeks.Higher PAGI-SYM scores indicate more severe symptoms PAGI-SYM and PAGI-QoL scores were measured at day 1 and weeks 4, 8, and 12PAGI-Q oL27measure Uses a six-point Likert response scale to measure HRQoL in patients with upper gastrointestinal disorders, with a recall period of 2 weeks. HiI-QoL scores reflect better quality of life SF-1228measure Used to assess self-reported general HRQoL (1-week recall was used in this study).Two SF-12 summary scores (mental and physical component summary scores) were reported (overall measure scored 0-100).Higher scores reflect better health status and higher quality of life SF-12 scores were assessed at day 1 and week 12Abbreviations: ANMS GCSI-DD, American Neurogastroenterology and Motility Society Gastroparesis Cardinal Symptom Index-Daily Diary; ePRO, electronic patient-reported outcomes; HRQoL, health-related quality of life; PAGI-QoL, patient assessment of upper gastrointestinal quality-of-life index; PAGI-SYM, patient assessment of upper gastrointestinal disorders-symptom severity index; SF-12, 12-item short-form health survey.