Post-marketing Clinical Study of Rebamipide in Patients With Gastric Ulcer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Post-marketing Clinical Study of Rebamipide in Patients With Gastric Ulcer (NCT00233389), a randomized phase 4 study. The primary outcome is Gastric Ulcer Healing Rate (Number of Subjects Whose Gastric Ulcer Was Healed/Number of Subjects Evaluated x 100) at Week 8.

2. Clinical Question

The clinical question is focused on continued treatment after eradication therapy, rather than on comparing eradication therapy with no eradication therapy. Both groups are described as receiving Helicobacter pylori eradication therapy; the contrasting treatment is rebamipide or placebo. The study summary states that the trial examined the efficacy of continued rebamipide administration in patients with Helicobacter pylori-positive active gastric ulcer.

The outcome is a short-term, endoscopically assessed healing status. It does not describe symptom relief, later ulcer recurrence, or other outcomes. The primary endpoint is reported at Week 8, and no secondary outcomes are posted in the registry record. The available results therefore address the defined Week 8 healing question and do not establish how the groups compared on outcomes that were not posted.

The analysis population is defined by receipt of the specified eradication therapy and at least one dose of the investigational product. That definition is narrower than simply describing everyone enrolled, and the registry record does not explain how participants who did not meet those criteria were handled. The posted arm results are 70.13% for rebamipide and 60.54% for placebo, with uncertainty described by 95% confidence intervals. The group comparison and its uncertainty are considered in the interpretation of the result.

3. Trial Design

This was a randomized, double-blind, parallel-group phase 4 study. Its title describes a post-marketing clinical study of rebamipide in patients with gastric ulcer, and the official title specifies investigation of gastric ulcer healing with continued dosing following Helicobacter pylori eradication therapy. The registry record lists Otsuka Pharmaceutical Co., Ltd. as sponsor and the study status as completed.

The registry record lists enrollment of 301. The posted outcome groups are rebamipide, with 154 participants, and placebo, with 147 participants. These are the group denominators shown for the primary comparison. The treatment allocation is identified as randomized, but the sequence-generation procedure, allocation concealment, and any stratification factors are not reported in the registry record. Double masking is stated; which parties were masked and how masking was maintained are not reported.

The study used a parallel-group model, so the comparison is between participants assigned to different treatment groups rather than a within-participant comparison. The trial is described as placebo-controlled, with eradication therapy in the context of continued rebamipide or placebo. The specific eradication regimen, rebamipide dose, treatment schedule, adherence procedures, and criteria for stopping treatment are not reported in the registry record.

The registry record lists a study start of 2004-04-30 and primary completion on 2005-09-01. The primary outcome was assessed at Week 8. The endpoint description says healing was determined by endoscopic assessment of ulcer stage, using the Sakita-Miwa Classification, and classifies S1 or S2 as healed. The registry record does not report further details about endoscopy procedures, assessor training, or adjudication.

The registry record does not report the trial's original sample-size justification or the statistical analysis plan. It also does not report interim analyses or a planned analysis threshold for the primary comparison. Planning illustrations based on the observed effect are separate from the trial's original design assumptions and should not be read as a description of how enrollment was chosen.

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: Following the eradication therapy period, 154 subjects in the rebamipide group and 147 subjects in the placebo group, excluding withdrawals during the eradication therapy period, received the investigational medicinal product (IMP) (rebamipide or placebo) for the treatment of gastric ulcer.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Efficacy Analysis Set: Subjects who received H. pylori eradication therapy as specified and at least one dose of the IMP.

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: Full Efficacy Analysis Set: Subjects who received H. pylori eradication therapy as specified and at least one dose of the IMP.

6. Primary Endpoint

The registry lists the primary outcome as Gastric Ulcer Healing Rate (Number of Subjects Whose Gastric Ulcer Was Healed/Number of Subjects Evaluated x 100) at Week 8, measured over Week 8.

The primary endpoint is the percentage of participants whose gastric ulcer was judged healed at Week 8. Healing was defined as an endoscopic ulcer stage of S1 or S2 in the Sakita-Miwa Classification. The endpoint description says that the percentage was calculated by group among participants in the Full Efficacy Analysis Set, defined as those who received Helicobacter pylori eradication therapy as specified and at least one dose of the investigational product.

This is a binary outcome at a specified assessment time: each participant is classified as healed or not healed for the endpoint. A percentage expresses how common the defined event was in each group. The posted percentages are 70.13% for rebamipide and 60.54% for placebo. The comparison also gives the event counts: 108 of 154 in the rebamipide group and 89 of 147 in the placebo group.

In general, an objective assessment tied to a prespecified classification can make an endpoint easier to interpret than an outcome with no explicit definition. Endoscopic assessment also connects the endpoint to the recorded ulcer stage rather than relying only on a participant's report of symptoms. However, the registry record does not report details about the assessment process, whether assessors were masked, or how ambiguous classifications were resolved.

The endpoint captures healing status at Week 8, not the time taken to heal or the durability of healing after that visit. It does not, by itself, describe symptoms, quality of life, recurrence, or outcomes beyond the stated assessment. No secondary outcome measures are posted in the registry record, so the reported primary endpoint is the available basis for describing the treatment-group comparison.

7. Results Posted in the Registry

• Rebamipide: 70.1% (108/154) versus 60.5% (89/147) on control; risk difference 9.6 percentage points (95% CI -1.1 to 20.3), risk ratio 1.16 (0.98–1.37); unadjusted two-proportion z-test p = 0.080.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the healing percentage represent?

The primary outcome is binary because each participant is classified as healed or not healed at Week 8 according to the stated endoscopic definition. The group percentage describes the proportion classified as healed among the participants evaluated for the endpoint. The posted percentages are 70.13% for rebamipide and 60.54% for placebo. The corresponding event counts are 108 of 154 and 89 of 147, respectively. These are descriptive summaries; the comparison between groups is a separate inferential step.

In general, binary outcomes can be summarized using an absolute difference between proportions or relative measures such as a risk ratio and odds ratio. Each answers a different descriptive question. An absolute difference describes the separation between the group proportions on the probability scale. A risk ratio compares the proportions as a ratio, while an odds ratio compares the odds of the event. These scales can produce different-looking numerical estimates without representing different endpoint data.

How should the risk difference be read?

The posted risk-difference estimate is 0.096, with a 95% confidence interval from -0.011 to 0.203. The sign indicates the direction under the comparison as reported, with the estimate favoring a higher observed proportion in the rebamipide group. The interval gives a range of values compatible with the data under the method's assumptions. Because the range includes zero, it includes no group difference on the risk-difference scale. The interval also extends to differences in both directions, so the data do not pin down the direction and magnitude with high precision.

The relative measures are a risk ratio of 1.158, with a confidence interval from 0.981 to 1.368, and an odds ratio of 1.53, with a confidence interval from 0.949 to 2.468. In general, the no-association value is one for ratio measures. Both posted intervals include that value. The odds ratio is not the same thing as a risk ratio, especially when an outcome is common; it should not be described as a direct percentage increase in the probability of healing.

What does the p-value add?

The posted unadjusted p-value is 0.08. In general, a p-value summarizes how compatible the observed data are with a null hypothesis under a specified statistical model. It does not measure the probability that a treatment effect is real, and it does not describe the size or practical importance of an effect. The estimate and confidence interval convey the direction, scale, and precision of the comparison more directly. The registry record does not report the original analysis plan, the test used, or its planned decision threshold, so the p-value should be interpreted as a posted unadjusted result rather than as a complete account of the analysis.

Why do the analysis population and design matter?

The trial is listed as randomized, double-blind, and parallel-group. Random allocation can, in general, help balance prognostic factors across groups, while masking can reduce some forms of influence on treatment and outcome assessment. Those design labels alone do not establish that allocation was concealed or that every assessment remained masked; procedural details are not reported in the registry record. The endpoint is defined for the Full Efficacy Analysis Set, consisting of participants who received eradication therapy as specified and at least one dose of the investigational product. The registry record does not report a separate intention-to-treat result or explain how participants outside that analysis set were handled.

For binary outcomes, the precision of a group percentage depends on the number of participants and the observed event frequency. A confidence interval represents sampling uncertainty under its statistical assumptions; it does not account automatically for bias, outcome misclassification, missing data, or limitations in the trial conduct. The registry record does not report the method used to construct the primary confidence intervals or provide enough detail to assess those assumptions fully.

What is not specified?

The registry record does not report the statistical analysis plan, the primary test procedure, covariate adjustment, handling of missing outcome assessments, or interim analyses. It lists one outcome measure and no posted secondary outcome measures. A multiplicity strategy is not reported. Without these details, the posted comparison can be described, but it cannot be fully checked against a prespecified statistical plan. This distinction matters: a registry result can supply useful estimates while leaving important analytical choices undocumented.

General statistical interpretation should keep the scale of each estimate clear. The healing percentages are observed group summaries; the risk difference and ratio measures are comparative summaries; the confidence intervals express uncertainty; and the p-value is a model-based summary of compatibility with a null hypothesis. None of these alone addresses clinical importance, generalizability, bias, or outcomes that were not measured or posted.

10. Confidence Intervals and Precision

The healing-rate estimates are 70.13% for rebamipide, with a 95% confidence interval from 62.9% to 77.36%, and 60.54% for placebo, with a 95% confidence interval from 52.64% to 68.45%. These intervals describe uncertainty around each group's estimated healing percentage. Their overlap is not, by itself, a formal test of the between-group difference; the direct comparison is better represented by the posted risk difference and its own interval.

The estimated risk difference is 0.096, with a 95% confidence interval from -0.011 to 0.203. This interval includes zero and spans values on either side of no difference. It therefore does not identify a single, precise value for the treatment-group contrast. The posted risk-ratio interval, from 0.981 to 1.368, and odds-ratio interval, from 0.949 to 2.468, also include their respective no-association values.

In general, a confidence interval is a measure of statistical precision under the method's assumptions, not a guarantee that the true value lies within the particular interval or a range of likely individual outcomes. Wider intervals indicate less precision than narrower intervals on the same scale. The interpretation also depends on the analysis being appropriate and on the data being measured and handled as assumed. Confidence intervals do not automatically account for missing outcomes, bias, or unreported analytical choices.

The registry record does not report the interval-construction method or the original statistical analysis plan. The posted intervals can still be read as showing that the estimates are uncertain, and that the data are compatible with a range of possible group differences. The observed percentages alone should not be treated as exact treatment effects.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrations based on the observed effect and are not the trial's original design assumptions, which are not reported in the registry record. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario targets the stated fraction of the observed effect.

For planning based on 100% of the observed effect, the illustration gives 386 participants per arm for 80% power and 516 participants per arm for 90% power. Planning for 75% of the observed effect gives 698 participants per arm for 80% power and 934 participants per arm for 90% power. Planning for 50% of the observed effect gives 1,593 participants per arm for 80% power and 2,132 participants per arm for 90% power.

In plain terms, the 80% power column describes the illustrative number of participants per arm needed to have an 80% chance of detecting the chosen effect under the stated assumptions, if that effect and the model assumptions were correct. The 90% power column uses a higher target chance of detection, so its planning counts are larger. Smaller effects need more participants because a smaller separation between groups is harder to distinguish from random variation.

These values are planning illustrations, not observed power and not a reconstruction of the completed trial's sample-size decision. They use the observed effect as a starting point, so they should be treated cautiously: an observed estimate is uncertain, and a different true effect would imply different planning requirements. The trial's original sample-size justification is not reported in the registry record.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.605; treatment proportion p₁ = 0.701; average p̄ = 0.653.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6730 = 1.3191.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6696 = 0.5635.
4. n = (1.3191 + 0.5635)² / (0.0959)² = 385.7, rounded up to 386 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.605442176870748                          # control proportion
p1 <- 0.701298701298701                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(233389)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 386 (80% power), 516 (90% power)
 75% of observed effect: n per arm = 698 (80% power), 934 (90% power)
 50% of observed effect: n per arm = 1,593 (80% power), 2,132 (90% power)
power.prop.test: 385.7 
Simulated power at n = 386 per arm: 0.804 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one outcome measure and no posted secondary outcome measures. A multiplicity strategy is not reported. When several outcomes, time points, or treatment comparisons are tested, the chance of at least one seemingly unusual result can increase unless the analysis accounts for the number and structure of tests. That general concern cannot be quantified for this study from the registry record, because a broader testing plan is not reported.

The record posts one primary endpoint at Week 8 and a group comparison, including an unadjusted p-value of 0.08. It does not report whether that comparison was the sole planned test, whether there were additional analyses that were not posted, or whether any adjustment was planned. The absence of posted secondary measures should not be taken as proof that no other analyses were conducted.

The endpoint population is defined as the Full Efficacy Analysis Set: participants who received Helicobacter pylori eradication therapy as specified and at least one dose of the investigational product. This definition identifies who was included in the stated endpoint analysis, but the registry record does not report how that population was formed from all enrolled participants or whether a separate intention-to-treat analysis was conducted.

Missing outcome handling is not reported in the registry record. It does not state how participants without an evaluable Week 8 endoscopy were treated, whether missing outcomes were imputed, or whether sensitivity analyses were conducted. These details matter because the endpoint is a participant-level healed/not-healed classification, and exclusions or missing assessments can affect the resulting group percentages.

The statistical analysis plan, interim analyses, covariate adjustment, and procedures for checking the analysis assumptions are also not reported. Without those details, the posted estimates remain interpretable as results for the defined endpoint and analysis population, but the extent to which they reflect all planned analyses and decisions cannot be assessed from the registry record.

14. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: constipation (0.7%), agranulocytosis (0.6%), thrombocytopenia (0.6%), cerebellar haemorrhage (0.6%).

The serious adverse-event figures are counts with the number at risk shown for each group. In the rebamipide group, 2 serious adverse events are reported among 154 participants at risk; in the placebo group, 1 is reported among 147 participants at risk. The record reports 0 deaths in each group. These are descriptive counts, not a formal comparative safety analysis.

For other adverse events, the record reports 85 among 154 participants at risk in the rebamipide group and 75 among 147 in the placebo group. These counts should be read separately from the serious adverse-event counts. The registry record does not report event-specific timing, duration, severity definitions beyond the serious classification, or statistical comparisons between groups.

The listed serious-event terms are Constipation (0.7%), Agranulocytosis (0.6%), Thrombocytopenia (0.6%), and Cerebellar haemorrhage (0.6%). The registry record does not make clear in this listing how each term maps to a treatment group or whether the percentages refer to participants or events. The counts and terms should therefore be read as the posted safety summaries, not as evidence of a comparative risk estimate. Sparse serious-event counts also limit how precisely group differences can be characterized.

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

The posted results show Week 8 healing rates of 70.13% with rebamipide and 60.54% with placebo in the defined Full Efficacy Analysis Set. The risk-difference estimate is 0.096, with a 95% confidence interval from -0.011 to 0.203, and the unadjusted p-value is 0.08. The point estimate is in the direction of a higher observed healing proportion in the rebamipide group, but the interval includes no difference and allows a range of possible contrasts. The data therefore do not precisely determine the size or direction of the underlying group difference.

This is a statistical description of the posted trial outcome, not a treatment recommendation. It does not establish what would happen for an individual participant or in other groups of patients. The result applies to the stated endpoint, population, and Week 8 assessment; it does not address other clinical outcomes or later time points.

The record describes randomized, double-blind, placebo-controlled treatment following Helicobacter pylori eradication therapy. However, important details about the allocation process, masking procedures, treatment regimen, missing outcomes, and analysis plan are not reported. Those omissions limit assessment of possible bias and of whether all relevant analysis choices are represented in the posted result.

The safety summaries report serious adverse events in both groups and zero deaths in each. They are counts, not a complete comparative safety analysis, and the registry record does not provide enough detail to establish how the listed serious-event terms relate to groups. The overall meaning is therefore bounded: the record offers a Week 8 healing comparison and descriptive adverse-event summaries, with substantial uncertainty around the treatment contrast and limitations in the available methodological detail.

16. Why This Trial Matters Statistically

This study is statistically instructive because its main result combines an observable difference in group percentages with uncertainty that remains substantial. The healing percentage is 70.13% for rebamipide and 60.54% for placebo, while the risk-difference interval extends from -0.011 to 0.203. Reading the point estimate alone would omit the uncertainty; reading only the p-value would omit both the scale of the estimate and the range of compatible differences.

The case also illustrates why the effect measure matters. The risk difference is on an absolute scale, while the risk ratio and odds ratio are relative measures. Each conveys a different description of the same binary endpoint. Reporting the measures together can help readers understand the result without treating one scale as a substitute for all others.

The endpoint is clinically concrete in its definition: Week 8 endoscopic assessment, with healing classified as S1 or S2. At the same time, the registry record does not report the assessment procedures or handling of missing endoscopies. Statistical precision is only one part of evidence quality; outcome measurement and analysis decisions also affect interpretation.

Finally, the planning illustrations show the dependence of sample-size requirements on the assumed effect. The displayed scenarios use fractions of the observed effect and stated planning assumptions, not the original trial design rationale. They illustrate why a study's estimated effect should not automatically be treated as a fixed planning target, and why the registry's unreported original justification cannot be inferred from post hoc planning calculations.

17. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The registry record does not report the original sample-size justification, statistical analysis plan, planned significance threshold for the primary comparison, or interim analyses. It also does not describe allocation concealment, the randomization sequence, or the details of double masking. These omissions make it difficult to assess whether the posted analysis follows a prespecified plan and how safeguards against bias were implemented.

The primary endpoint depends on endoscopic classification at Week 8, but details of the assessment process, assessor masking, and adjudication are not reported. Missing outcome handling is also not reported. The Full Efficacy Analysis Set is defined by receipt of specified eradication therapy and at least one dose of the investigational product, but the registry record does not describe exclusions, a separate intention-to-treat analysis, or sensitivity analyses.

The registry record posts no secondary outcome measures. It therefore does not provide a broader set of outcomes or follow-up results with which to assess the Week 8 endpoint. The posted result is limited to the specified healing outcome and does not address durability, recurrence, or other unposted outcomes.

The comparison is statistically uncertain: the risk-difference confidence interval includes zero, and the relative-measure intervals include their respective no-association values. The registry record does not report a formal comparative safety analysis, and the mapping of the listed serious adverse-event terms to treatment groups is unclear. These limitations mean the results should be interpreted as a bounded summary of the posted endpoint and safety counts, rather than a complete account of efficacy, safety, or trial conduct.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Post-marketing Clinical Study of Rebamipide in Patients With Gastric Ulcer — NCT00233389.
• Registry record last updated 2021-06-04; status last verified 2021-05.
• Results first posted 2021-06-04.
• 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.

21. Record Summary

NCT00233389 is a completed, randomized, double-blind, parallel-group phase 4 post-marketing study sponsored by Otsuka Pharmaceutical Co., Ltd. The registry record lists enrollment of 301, a start date of 2004-04-30, and primary completion on 2005-09-01. It examined continued rebamipide versus placebo following Helicobacter pylori eradication therapy in participants with active gastric ulcer.

The primary outcome was the percentage with endoscopically assessed healing at Week 8, defined as stage S1 or S2 in the Sakita-Miwa Classification, among the Full Efficacy Analysis Set. Posted healing rates were 70.13% for rebamipide and 60.54% for placebo. The risk-difference estimate was 0.096, with a 95% confidence interval from -0.011 to 0.203 and an unadjusted p-value of 0.08.

The registry record lists no posted secondary outcome measures. It reports serious adverse-event counts of 2 among 154 for rebamipide and 1 among 147 for placebo, with 0 deaths in each group. Important details about the statistical plan, missing outcomes, and safety-event attribution are not reported in the registry record.