Phase 3 Study of TAK-438 10 mg in the Treatment of Non-Erosive Gastroesophageal Reflux Disease (NERD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase 3 Study of TAK-438 10 mg in the Treatment of Non-Erosive Gastroesophageal Reflux Disease (NERD) (NCT02954848), a randomized phase 3 study. The primary outcome is Percentage of Days Without Symptoms of Heartburn.

2. Clinical Question

3. Trial Design

The registry describes NCT02954848 as a randomized, double-blind, placebo-controlled, multicenter Phase 3 study with a parallel-group design. It was completed. The stated enrollment was 484. The study began on 2016-11-15, and primary completion was recorded as 2018-02-25. Takeda is listed as the sponsor.

The intervention arms were placebo and TAK-438 10 mg. Both groups received placebo-matching tablets once daily after breakfast during a 1-week run-in period. In the treatment period, the placebo group continued placebo-matching tablets for up to 4 weeks, while the active-treatment group received TAK-438 10 mg tablets for up to 4 weeks. The registry lists quadruple masking. It does not identify the masked roles in the information available for this account.

The primary outcome was the percentage of days without heartburn symptoms, assessed up to Week 4. The registry says symptoms were recorded in daily participant diaries, including whether heartburn was present and its severity. The primary analysis population was the full analysis set, defined in the record as randomized participants who received at least one dose of study drug. The posted arm counts for this endpoint were 245.0 for placebo and 238.0 for TAK-438 10 mg.

The registry lists 27 outcome measures and 51 analyses, and it posts 24 secondary outcomes. The listed secondary outcomes include symptom-free days, cumulative improvement, and symptom severity, with some analyses stratified by response or endoscopic findings. The registry record does not report the randomization procedure, allocation concealment details, a sample-size justification, or the full prespecified statistical analysis plan.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants with a diagnosis of non-erosive gastroesophageal reflux disease (NERD) were enrolled and received one tablet of placebo orally, once daily in a run-in period. After that, the participants who met all the entry criteria were randomized in 1:1 ratio to receive TAK-438 10 mg or TAK-438 placebo-matching tablets.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set (FAS) included participants who were randomized and received at least one dose of the study drug.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Randomized Set included all participants who were randomized in the study.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Days Without Symptoms of Heartburn, measured over Up to Week 4.

The primary endpoint was the percentage of treatment-period days without heartburn symptoms, assessed up to Week 4. Participants recorded symptom presence and severity in daily diaries. The stated severity scale ranged from 0 to 4: 0 meant no symptom, 1 meant symptoms did not hinder daily activities, and 2, 3, and 4 represented mild, moderate, and severe symptoms. Higher scores indicated greater severity. For the endpoint, days without heartburn were divided by days in the treatment period to derive each participant’s percentage.

This endpoint expresses symptom experience across treatment days rather than recording only a single symptom assessment. A diary can capture changes over time and distinguish days without symptoms from days when symptoms were present. The registry reports a median and full range for each treatment group. Such summaries describe the distribution of observed percentages, while the inferential comparison is reported separately.

Interpretation depends on how diary completion, incomplete treatment periods, and missing entries were handled. Those details are not reported in the registry record. The record also does not state a threshold for a clinically meaningful difference in symptom-free days. Therefore, the endpoint can be described as a diary-based measure of heartburn-free treatment days, but the clinical importance of a particular numerical contrast cannot be determined from the posted endpoint description alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a percentage derived from daily records. For each participant, the registry describes dividing the number of treatment days without heartburn by the number of days in the treatment period. The underlying diary recorded symptom presence and severity. This produces a bounded measure: the posted full range runs from 0.0 to 100.0. A bounded percentage cannot take values outside its defined limits, and its distribution need not be symmetric. The registry reports a median and full range rather than a mean and standard deviation.

Why report a median?

The median is the value at the centre of an ordered set of observations. It is less influenced by unusually high or low values than the mean. That can be useful when observations are skewed or when a measure is bounded. The full range gives the smallest and largest reported values, but it does not show how concentrated observations are near the centre or whether there are clusters or gaps. The posted summaries therefore give a limited description of the distribution.

The arm summaries were 61.5 for placebo and 72.55 for TAK-438 10 mg, with a full range of 0.0 to 100.0 in each group. The inferential result is separately reported as a median difference in final values. It should not be assumed that subtracting the two group summaries reproduces the sponsor’s analysis: the registry reports a Wilcoxon rank-sum test, and it does not provide the individual observations or further calculation details needed to reproduce its estimate.

What does the Wilcoxon rank-sum test assess?

In general, the Wilcoxon rank-sum test compares the ordering of observations between independent groups. Rather than requiring a comparison of group means, it uses the ranks of observations across groups. It can be useful when the outcome distribution does not meet assumptions commonly associated with a mean-based comparison. Its interpretation is not automatically a test of medians alone: a median-shift interpretation requires additional assumptions about the shapes of the group distributions. The registry names the test and reports a median difference, but does not describe its assumptions or the detailed estimation procedure.

For this outcome, the posted analysis gives a median difference in final values of 3.800, a 95% confidence interval from 0.000 to 10.600, and p = 0.064. The confidence interval represents a range of values compatible with the data under the method’s assumptions and repeated-sampling interpretation. It is not a range containing a specified proportion of individual participant outcomes. Because its lower limit is 0.000, the interval includes zero. Its width also shows that the estimated contrast is uncertain. The registry does not give a clinically important threshold against which to assess that interval.

How should the p value be read?

In general, a p value measures how unusual data at least as incompatible with a null hypothesis as the observed data would be, assuming that null model and the test procedure. It does not measure the probability that the treatment has no effect, and it does not quantify the size or clinical importance of a difference. The reported p value is 0.064. A decision about statistical significance depends on a prespecified threshold and analysis plan; those details are not reported in the registry record for this endpoint. The sample-size planning assumptions elsewhere in the record mention a two-sided significance level of 0.05, but that planning statement does not establish the primary outcome’s full analysis plan or decision rule.

What does the confidence interval add?

In general, a confidence interval communicates uncertainty around an estimate and is more informative than a single p value because it gives a range of plausible effect sizes under the chosen method. Here, the reported 95% confidence interval is from 0.000 to 10.600. The interval’s lower limit is at zero, while its upper limit allows a positive contrast. Readers should consider both the estimate and the interval, as well as the outcome scale and clinical context. The registry does not report a threshold that would define a meaningful improvement in the percentage of days without heartburn.

All statistical methods depend on assumptions. For rank-based methods, the ordering of observations and independence between participants matter; interpreting a location shift specifically also depends on the distributions being suitably comparable in shape. The registry does not report diagnostics or an explanation of these assumptions. Nor does it report whether any adjustment was made for baseline outcome values or other covariates. The posted result should therefore be understood as the reported unadjusted group comparison unless further analysis details are available in the registry record.

The record includes planning illustrations based on fractions of the observed effect, with equal allocation, a normal approximation, and a two-sided significance level of 0.05. These are planning assumptions, not evidence about the trial’s achieved power or its original design rationale. The registry does not report the sample-size justification used when the trial was designed. A planning illustration based on an observed effect should not be confused with an observed treatment effect or with a retrospective claim that the study had adequate power.

10. Analyses Reported by the Sponsor

The posted primary analysis is the registry’s reported Wilcoxon rank-sum comparison of final values. It gives a median difference of 3.800, a 95% confidence interval from 0.000 to 10.600, and p = 0.064. The group summaries are 61.5 for placebo and 72.55 for TAK-438 10 mg. These summaries and the test result are different statistical descriptions and should be read together rather than treated as interchangeable.

The registry record does not report an independent recalculation of the primary analysis, and the participant-level data needed to reproduce the test are not included in the registry record described for this analysis. There is no recorded discrepancy between a recalculation and the sponsor’s result. Accordingly, the sponsor-reported analysis is the available inferential result; no recalculated estimate or p value should be substituted for it.

The interval reaches 0.000 at its lower limit and extends to 10.600. This indicates uncertainty that includes zero under the reported interval method. The p value is 0.064. The registry does not report the complete prespecified statistical plan or the decision threshold for the primary comparison, so the reported p value should not be converted into a definitive claim about the trial’s conclusion. It is also not a measure of clinical importance. The record does not report a clinically important difference threshold for this endpoint.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 27 outcome measures, 51 analyses, and 24 posted secondary outcomes. Several listed secondary outcomes concern heartburn symptoms, symptom-free days, cumulative improvement, and symptom severity. Some are defined within subgroups based on response or endoscopic findings. Multiple outcomes and subgroup analyses raise a general interpretive issue: when many statistical comparisons are made, the chance of obtaining apparently notable results by chance can increase. Methods such as prespecifying a testing hierarchy or adjusting for multiplicity can address this issue, but the registry record does not report whether such methods were planned or used.

The primary outcome’s reported analysis is a Wilcoxon rank-sum test. The record does not report the complete prespecified analysis plan, including a testing hierarchy, multiplicity adjustment, or the status of the posted analyses as confirmatory or exploratory. The secondary outcomes should therefore be interpreted with caution, particularly where they involve subgroups. The registry listing alone does not establish how those analyses were prioritized or whether their results were adjusted for multiple comparisons.

The endpoint depended on daily diaries and the treatment-period duration. The registry defines the full analysis set as randomized participants who received at least one dose of study drug. It does not report how missing diary entries were handled, what constituted sufficient diary completion, how withdrawals were treated in the endpoint analysis, or whether sensitivity analyses explored alternative assumptions about missing observations. These details matter because missing symptom records can affect a participant’s calculated percentage and can affect the comparison if missingness differs between groups or relates to symptoms.

The posted primary endpoint counts are 245.0 for placebo and 238.0 for TAK-438 10 mg. The registry record does not provide further information in the registry record about exclusions from the analysis set, data completeness, or the handling of participants with incomplete follow-up. The posted analysis can be reported as defined, but the impact of missing data and analysis-population choices cannot be assessed from the registry record alone.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: colitis ulcerative (0.4%).

The registry reports serious adverse events for 1 participant in the placebo group and 0 participants in the TAK-438 10 mg group. The corresponding groups at risk were 245 and 238. The record also reports other adverse events affecting 12 participants receiving placebo and 10 receiving TAK-438 10 mg. No deaths were reported in either group. The listed top serious adverse event term is “Colitis ulcerative,” with 0.4% shown; the record does not identify an arm for that term in the information.

These figures are counts reported by the registry, not a complete description of the safety profile. They do not indicate the duration, severity beyond the serious-event classification, relatedness to study drug, or clinical course of each event. The record does not report further details about event definitions, collection procedures, or follow-up for adverse events. The figures should not be read as proving that a treatment is safe or unsafe, or as establishing a causal relationship between treatment and an event.

When event counts are small, the observed numbers can provide limited precision about the frequency of uncommon harms. A zero reported count does not establish that an event cannot occur. The posted serious and other adverse-event summaries are best interpreted as the events recorded during this study, alongside the limited information about exposure and event detail in the registry record.

13. What the Result Does, and Does Not, Mean

The result describes a randomized comparison of TAK-438 10 mg and placebo on a diary-based measure: the percentage of treatment days without heartburn symptoms, assessed up to Week 4. The posted group summaries were 72.55 for TAK-438 10 mg and 61.5 for placebo. The reported inferential comparison was a Wilcoxon rank-sum test, with a median difference in final values of 3.800 and a 95% confidence interval from 0.000 to 10.600. The reported p value was 0.064.

These statistics do not, by themselves, answer whether any observed contrast is clinically important. The confidence interval includes zero at its lower limit, and the registry does not state a threshold for a clinically meaningful difference in symptom-free days. The p value does not measure the probability that the treatment is ineffective, nor does it describe the size of a difference. It is one part of the reported statistical evidence and must be interpreted in the context of the analysis plan, which is not fully reported in the registry record.

The endpoint captures symptom-free treatment days, not all dimensions of gastroesophageal reflux disease or longer-term outcomes. The posted primary endpoint covers up to Week 4. The registry’s brief description states that the study evaluated efficacy and safety, but the primary result does not establish effects beyond the measured endpoint and timeframe. The record lists additional secondary outcomes, but their presence does not resolve questions about multiplicity, missing diary data, or the clinical meaning of the primary contrast.

The trial’s randomized, double-blind, placebo-controlled design supports a structured comparison between assigned groups. However, interpretation still depends on implementation and analysis details, including handling of incomplete diaries and the prespecified testing plan. Those details are not reported in the registry record. The appropriate conclusion is therefore limited: the posted results give the observed group summaries and a rank-based comparison with substantial uncertainty, without establishing a definitive clinical interpretation on their own.

14. Why This Trial Matters Statistically

This trial illustrates why clinical-trial results are best read through both the endpoint definition and the analysis used. A diary-derived percentage of days without heartburn compresses repeated symptom records into a participant-level summary. The median and full range describe aspects of those summaries, while the rank-sum test compares the ordering of observations between groups. Those quantities answer related but distinct statistical questions.

The reported median difference, confidence interval, and p value also show why a single statistic is not enough. The interval extends from 0.000 to 10.600, so it communicates uncertainty that reaches zero as well as values above zero. The p value, 0.064, is not a measure of clinical importance and cannot replace the estimate or interval. Because the registry does not report a clinically important threshold or the full prespecified testing plan, interpretation should remain tied to what was actually posted.

The study also highlights the importance of transparent diary and missing-data methods. A percentage based on treatment days depends on the available daily records and the denominator used. The registry describes the endpoint calculation but does not report how incomplete diaries were handled. That gap limits independent assessment of how the endpoint was constructed for participants with missing information.

Finally, a Phase 3 designation and randomized, double-blind, placebo-controlled design describe important features of the trial, but they do not eliminate uncertainty about estimates or replace reporting of analysis details. The statistical value of a registry result depends on knowing the outcome definition, population, comparison method, uncertainty, and limits of the available record. This trial provides those elements in part, while leaving aspects of the statistical plan and data handling unreported.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the sample-size justification used in the original trial design. Although planning illustrations are based on the observed effect and specified assumptions, such illustrations are not evidence of the trial’s originally planned power and should not be interpreted as observed power. The record also does not report whether interim analyses were planned or conducted.

The primary outcome was derived from daily participant diaries, but the registry does not report the rules for incomplete diary records, missing days, withdrawals, or incomplete treatment periods. These decisions can affect the calculated percentage of days without symptoms. Without those details or sensitivity analyses, the influence of missing data cannot be assessed from the posted result.

The primary comparison is reported as a Wilcoxon rank-sum test with a median difference in final values. The registry does not report further details of the estimation procedure, assumptions checks, covariate adjustment, or a complete prespecified statistical analysis plan. It also does not provide a clinically important difference threshold for interpreting the endpoint. The reported confidence interval and p value can be described, but their implications for a prespecified decision criterion are not fully assessable.

The registry lists multiple outcomes and analyses, including subgroup-based secondary outcomes, but does not report whether multiplicity was addressed or how the analyses were prioritized. This limits interpretation of secondary findings and any comparisons selected from the wider set of analyses. The serious and other adverse-event summaries also lack detail about event duration, severity, relatedness, and follow-up. These limitations do not negate the posted findings; they define what can and cannot be concluded from the registry record alone.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Phase 3 Study of TAK-438 10 mg in the Treatment of Non-Erosive Gastroesophageal Reflux Disease (NERD) — NCT02954848.
• Registry record last updated 2019-08-02; status last verified 2019-06.
• Results first posted 2019-08-02.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT02954848 was a completed, randomized, double-blind, placebo-controlled, multicenter Phase 3 parallel-group study sponsored by Takeda. The stated enrollment was 484. Participants received placebo or TAK-438 10 mg after a 1-week placebo run-in, with treatment planned for up to 4 weeks.

The primary endpoint was the percentage of treatment days without heartburn symptoms, based on participant diaries, in the full analysis set. Posted group summaries were 61.5 for placebo and 72.55 for TAK-438 10 mg. The reported Wilcoxon rank-sum analysis gave a median difference in final values of 3.800, a 95% confidence interval from 0.000 to 10.600, and p = 0.064.

The registry lists 27 outcome measures, 51 analyses, and 24 posted secondary outcomes. It reports serious adverse events affecting 1 participant in the placebo group and 0 in the TAK-438 10 mg group, other adverse events affecting 12 and 10 participants respectively, and no deaths in either group. Further details about missing diary data, multiplicity, and the full statistical plan are not reported in the registry record.