A Study of Gefapixant (MK-7264) in Japanese Adult Participants With Refractory or Unexplained Chronic Cough (MK-7264-038): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Gefapixant (MK-7264) in Japanese Adult Participants With Refractory or Unexplained Chronic Cough (MK-7264-038) (NCT03696108), a randomized phase 3 study. The primary outcome is Number of Participants Who Experienced at Least One Adverse Event (AE).

2. Clinical Question

This question is about the observed frequency of participants with at least one adverse event, not the total number of events. A participant with several adverse events still contributes once to this participant-count endpoint. The registry defines an adverse event as an untoward medical occurrence temporally associated with study intervention, whether or not considered related to that intervention. The definition includes an unfavorable or unintended sign, symptom, or disease, as well as an abnormal laboratory finding.

The registry also lists changes in Leicester Cough Questionnaire total score as secondary outcomes. Those outcomes concern cough-related quality of life and are distinct from the primary safety endpoint. Their posted results are not included in the primary comparison. Consequently, the central statistical question is whether the observed counts and their uncertainty give a clear basis for distinguishing the adverse-event experience between the two active-dose groups.

3. Trial Design

NCT03696108 is a completed Phase 3 randomized, double-blind, parallel-group clinical study sponsored by Merck Sharp & Dohme LLC. It enrolled Japanese adult participants with refractory or unexplained chronic cough. The registry lists an enrollment of 175. The study start date is 2018-10-31, and primary completion is recorded as 2020-10-07.

The two outcome arms are gefapixant 15 mg twice daily and gefapixant 45 mg twice daily. The arm descriptions state that each group received its assigned gefapixant tablet with a tablet intended to match the other dose, during the 52-week study period. The posted primary-outcome counts are 79 participants with at least one adverse event among 84 in the 15 mg group, and 82 among 85 in the 45 mg group. These are the denominators and counts listed for the outcome; the registry record does not explain how they relate to the enrollment figure.

Random allocation and double masking are listed, but the details of the allocation procedure and the people masked are not reported in the registry record. The parallel design indicates that participants were assigned to a dose group rather than described as switching between the listed groups. The primary outcome timeframe is up to 54 weeks, while the arm descriptions specify a 52-week study period. The registry record does not explain the difference between these timeframe descriptions.

The primary endpoint is a participant count for any adverse event, in the population of randomized participants who received at least one dose of study treatment. The record lists three secondary outcomes concerning Leicester Cough Questionnaire change and a threshold change in that score. The posted primary results compare the two active-dose groups, not gefapixant with an untreated or placebo-only outcome arm. That distinction matters when interpreting the scope of the comparison.

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: A total of 175 participants were originally enrolled in the study. 6 participants at one site were excluded from all analyses, including disposition, due to good clinical practice (GCP) noncompliance at the study site.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants who received at least one dose of study treatment.

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.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Experienced at Least One Adverse Event (AE), measured over Up to 54 Weeks.

The primary endpoint counts participants who experienced at least one adverse event, with follow-up described as up to 54 weeks. It is a participant-level count: the outcome does not count every event separately, and a participant with more than one adverse event remains a participant with at least one event for this endpoint. The registry’s definition includes an untoward medical occurrence temporally associated with study intervention, regardless of whether it is considered related to the intervention. It also includes abnormal laboratory findings and new or worsened disease.

This endpoint is used because the stated primary objective is to evaluate the safety of the two gefapixant doses. Counting participants with at least one event provides a direct description of how many participants in each group had any recorded adverse event during the specified follow-up. It does not, by itself, describe the number, duration, severity, seriousness, or causal attribution of all events. Those are different features of safety data and require their own definitions and analyses.

The registry lists 79 of 84 participants in the 15 mg group and 82 of 85 in the 45 mg group as having at least one adverse event. The primary-outcome population is described as randomized participants who received at least one dose. The registry record does not report additional details about handling participants who discontinued treatment or had incomplete follow-up. Therefore, the count should be read as the posted result for the stated endpoint and population, not as a complete account of every aspect of safety.

7. Results Posted in the Registry

• Gefapixant 45 mg: 96.5% (82/85) versus 94.0% (79/84) on control; risk difference 2.4 percentage points (95% CI -4.0 to 8.8), risk ratio 1.03 (0.96–1.10); unadjusted two-proportion z-test p = 0.458.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each participant either experienced at least one adverse event during the stated timeframe or did not. The posted values are counts of participants with an event and the number of participants in each arm. In general, a binary outcome can be described as a count, a proportion, or a contrast between group proportions. The primary outcome itself is a count endpoint, while the registry comparison reports a risk difference.

What does the risk difference describe?

A risk difference subtracts one group’s event proportion from the other group’s event proportion, using the direction specified for the comparison. It is an absolute contrast, so it expresses how far apart the group proportions are rather than how many times larger one is relative to the other. The registry’s estimate is positive for the 45 mg versus 15 mg comparison, consistent with the observed event proportion being higher in the 45 mg group. The numeric estimate is not repeated because the registry value has more than three decimal places.

Other measures can describe the same binary comparison from different perspectives. A risk ratio compares proportions multiplicatively, while an odds ratio compares event odds. Those measures can look different from an absolute risk difference, especially when events are common. The registry lists risk-ratio and odds-ratio estimates, but the main interpretation remains the risk difference because it is the comparison metric identified for the posted primary result. The registry record does not state why that metric was selected.

What does the p-value contribute?

The unadjusted p-value is 0.458. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It is not the probability that the null hypothesis is true, and it is not a measure of the size or clinical importance of a group difference. A p-value also does not resolve uncertainty about bias, incomplete follow-up, or the appropriateness of the analysis model. The registry record does not report the statistical test used to produce this p-value, so the test’s assumptions cannot be checked from the posted record.

How should the confidence interval be read?

A confidence interval gives a range of estimates that reflects sampling uncertainty under the interval method and its assumptions. For the posted risk difference, the interval includes zero because its lower limit is negative and its upper limit is positive. In general, zero represents no absolute difference between group proportions. An interval that crosses zero therefore does not rule out no difference at the confidence level used. It also does not demonstrate that the groups are equivalent: equivalence requires a prespecified margin and an analysis designed to assess whether the full interval lies within that margin. Such a margin and an equivalence analysis are not reported in the registry record.

The confidence interval’s width matters as well as whether it crosses zero. A broad interval permits a wider range of plausible effects than a narrow interval. The posted interval leaves appreciable uncertainty about the magnitude and direction of the comparison. The registry record does not identify the confidence level or the method used to calculate it, limiting a fuller evaluation of its statistical properties.

Why does event frequency matter?

When a binary outcome is common, a small absolute difference in proportions can coexist with risk ratios or odds ratios that tell a different-looking numerical story. More generally, the statistical scale should match the question: absolute differences are often useful for understanding the separation between observed group proportions, while relative measures express proportional change. These are complementary descriptions, not interchangeable answers. In this trial, adverse events were frequent in both groups, so the interpretation should attend to the absolute comparison and the uncertainty rather than relying on a single metric.

For count data with greater variability than a simple model expects, over-dispersion can make standard errors too small if not accounted for. That general concern is most relevant when modeling event totals or rates, rather than a participant-level indicator of any event. The primary endpoint counts participants with at least one event, not the total number of recurrent events. The registry record does not report a model for recurrent-event counts or an over-dispersion adjustment, so neither can be inferred from the posted endpoint.

What can be assessed from the registry analysis record?

The registry lists no analysis entries for this outcome. It does post a primary comparison and an unadjusted p-value, but does not describe the statistical test, adjustment variables, confidence-interval method, or analysis model. The record also does not report whether the primary comparison followed a prespecified plan or whether any covariate adjustment was intended. Without those details, the result can be described but the analysis cannot be fully reproduced or audited from the registry record alone.

Randomization can, in general, help balance measured and unmeasured characteristics across groups in expectation, but it does not remove uncertainty in an individual study or repair missing outcome data automatically. Double masking can reduce some forms of performance and assessment bias, but the registry record does not specify who was masked or how masking was maintained. These design labels are useful context, not proof that all potential biases were prevented.

The participant count endpoint also differs from an event-level analysis. If a participant experiences several events, the primary endpoint still counts that participant once. The posted primary result therefore cannot reveal the total burden of events or whether repeat events were concentrated in particular participants. Additional safety summaries may address some aspects, but event definitions and analyses beyond those explicitly described are not reported in the registry record.

10. Confidence Intervals and Precision

The risk-difference confidence interval has a lower limit below zero and an upper limit above zero. In general, an interval crossing zero indicates that the data are compatible with no absolute difference under the interval’s assumptions, while also allowing differences in either direction. It is not proof that there is no difference. Nor is it an equivalence finding: that would require an explicit equivalence margin and an appropriate analysis, neither of which is reported in the registry record.

The interval’s breadth signals limited precision about the size of the group contrast. The point estimate is positive, following the higher observed event proportion in the 45 mg group, but the interval does not pin down the magnitude or direction with confidence. The unadjusted p-value of 0.458 is consistent with substantial uncertainty; it should not be read as a measure of the probability that the groups are identical. The registry record does not report the confidence level or the method used to construct the interval, so those aspects of precision cannot be independently evaluated.

Precision is affected by the number of participants with outcome information, the frequency of the outcome, and the analysis method. The posted denominators are 84 and 85, and the outcome counts are high in both arms. The registry record lists an enrollment of 175 but does not explain the relationship between that enrollment and the outcome denominators. It also does not report missing-data handling or whether every randomized participant contributed to this endpoint. These omissions limit assessment of how much information the comparison contains and whether incomplete follow-up could have affected it.

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 sample-size display is a planning illustration based on the observed risk-difference effect, not the original trial design. For this binary outcome, the figures are participants per arm, not total events. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect.

For planning based on 100% of the observed effect, the display gives 1,207 participants per arm for 80% power and 1,615 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 2,276 participants per arm for 80% power and 3,046 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 5,412 participants per arm for 80% power and 7,245 participants per arm for 90% power.

In general, power is the chance that a planned study will reject the null hypothesis under a specified assumed effect and statistical model. The 80% column corresponds to a design intended to detect the stated effect with 80% power under the listed assumptions; the 90% column describes a higher target for detecting that effect. For a fixed effect assumption, reaching the higher target generally requires more participants. Smaller effects also require more participants because they are harder to distinguish from random variation.

The curve represents the same planning relationship across assumed effect sizes; it should be read as an illustration conditional on its assumptions, not as a description of this trial’s observed evidence. These figures are not observed power. The trial’s own original sample-size assumptions and justification are 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 = 0.940; treatment proportion p₁ = 0.965; average p̄ = 0.953.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3005 = 0.5890.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3000 = 0.2525.
4. n = (0.5890 + 0.2525)² / (0.0242)² = 1,206.4, rounded up to 1,207 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 <- 0.94047619047619                          # control proportion
p1 <- 0.964705882352941                          # 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(3696108)
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 = 1,207 (80% power), 1,615 (90% power)
 75% of observed effect: n per arm = 2,276 (80% power), 3,046 (90% power)
 50% of observed effect: n per arm = 5,412 (80% power), 7,245 (90% power)
power.prop.test: 1206.4 
Simulated power at n = 1,207 per arm: 0.807 (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 lists three secondary outcomes: change from baseline in Leicester Cough Questionnaire total score at Week 12, change from baseline in the total score, and the percentage of participants with a change of at least 1.3 points. It also lists five outcome measures overall, but does not post analyses for these secondary outcomes in the record. Their estimates, uncertainty, and statistical significance therefore cannot be assessed from the posted primary comparison.

When several outcomes or comparisons are tested, the chance of at least one apparently notable result can increase. In general, multiplicity procedures can control that risk by adjusting significance thresholds or reported inference. Whether any adjustment was planned or performed for this study is not reported in the registry record. The primary comparison’s p-value is explicitly unadjusted; it should not be treated as multiplicity-adjusted evidence. The registry does not report a hierarchy of outcomes, a testing sequence, or a family of hypotheses for the listed secondary outcomes.

The registry record does not report a missing-data strategy, rules for handling treatment discontinuation, or the method for dealing with incomplete questionnaires. It does not describe imputation, sensitivity analyses, or the timing of outcome ascertainment for participants with incomplete follow-up. These details matter because the primary outcome population is defined as randomized participants who received at least one dose, while the posted outcome denominators are 84 and 85 and registered enrollment is 175. The record does not reconcile those figures.

The primary endpoint population is stated, but other analysis populations, such as an intention-to-treat analysis set, are not described. Details about protocol deviations and exclusions are also not reported in the registry record. Consequently, it is not possible to determine whether the posted outcome represents every randomized participant or how results might change under alternative missing-data assumptions. These are reporting limitations; they do not establish that the original study lacked a statistical plan.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (3.5%), asthma (2.4%), bronchitis (1.2%), pneumonia bacterial (1.2%), large intestine polyp (1.2%).

The registry lists serious adverse events for 2 participants in the gefapixant 15 mg group, among 84 at risk, and for 10 participants in the gefapixant 45 mg group, among 85 at risk. It lists 0 deaths in each group. These figures describe the serious-event summary and are separate from the primary endpoint, which counts participants with at least one adverse event of any type.

The record also lists 73 participants with other adverse events in the 15 mg group and 76 in the 45 mg group. Serious-event terms appearing in the registry include Pneumonia at 3.5%, Asthma at 2.4%, and Bronchitis, Pneumonia bacterial, Large intestine polyp, and Hepatic function abnormal at 1.2% each. The registry summary does not clarify the denominator or whether these listed term percentages are specific to a group, so they should not be interpreted as group-specific rates.

These counts alone do not establish that an event was caused by treatment, and the primary endpoint’s definition expressly includes events whether or not considered related to study intervention. Definitions of seriousness, severity, relatedness, and event coding details are not reported in the registry record. The figures should therefore be read as posted safety summaries, not as a complete account of clinical context or causation.

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

The posted primary outcome shows that at least one adverse event was recorded for 79 of 84 participants in the 15 mg group and 82 of 85 participants in the 45 mg group. The observed proportion was higher in the 45 mg group, and the registry’s risk-difference estimate is positive. However, the confidence interval crosses zero and the unadjusted p-value is 0.458. The comparison is therefore uncertain about the size and direction of a population-level difference.

This result does not establish that the dose groups have identical safety profiles. Failure to distinguish a difference statistically is not the same as demonstrating equivalence. It also does not establish that one dose causes more adverse events, because the comparison is based on observed counts and the endpoint includes events whether or not they were considered related to treatment. The registry record does not report the analysis method in enough detail to evaluate all assumptions or potential sources of bias.

The outcome records participants with at least one event, not the total number of adverse events, their severity, duration, or clinical consequences. Serious-event counts are reported separately and include 2 participants in the 15 mg group and 10 in the 45 mg group; the record lists no deaths. Those figures need their own context and do not replace the primary endpoint. The serious-event term percentages are not clearly linked to a particular group in the registry summary.

The comparison is between two active gefapixant dose groups. It cannot answer how either dose compares with a placebo-only group, because such an outcome arm is not listed in the posted primary results. The registry also lists secondary cough-questionnaire outcomes, but their results are not included in this comparison. Accordingly, the statistical meaning is limited to the posted adverse-event summaries and their reported uncertainty, rather than broader claims about efficacy, causation, or the complete safety profile.

16. Why This Trial Matters Statistically

This trial illustrates why a difference in observed counts should be considered alongside uncertainty. The 45 mg group has a higher posted participant count with at least one adverse event than the 15 mg group, but the risk-difference interval crosses zero and the unadjusted p-value is 0.458. A point estimate alone can make a comparison seem more decisive than the data warrant. The interval communicates that the estimated contrast remains imprecise.

It also demonstrates the importance of defining the endpoint precisely. “At least one adverse event” is a participant-level measure, not a count of all events. It answers how many participants experienced an event, but it does not show whether events were repeated, how severe they were, or how they affected participants. The separate serious-event summary provides another perspective, but it does not resolve the limitations of the primary endpoint or establish causality.

The study design and comparator also shape the conclusion. Randomization and double masking are relevant design features, yet the posted primary comparison is between two active doses rather than against a placebo-only group. The results should not be stretched beyond that contrast. In addition, the registry record does not report the test method, missing-data strategy, multiplicity plan, or the relationship between enrollment and the outcome denominators. These details are important for judging reproducibility and the robustness of conclusions.

Finally, the planning illustration shows how assumptions about the effect to detect and desired power influence prospective sample-size requirements. It is not evidence of the trial’s power or a substitute for its original design rationale, which is not reported in the registry record. Keeping planning calculations separate from observed results helps prevent a common interpretive error: treating a study’s post hoc sample-size illustration as proof that its findings are conclusive.

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.

The registry record has important limits for interpreting this particular comparison. Although it reports randomized, double-blind, parallel-group design, it does not describe the randomization procedure, masking roles, allocation concealment, or how masking was maintained. Those details are not available for assessing how the design was implemented.

The registered enrollment is 175, while the primary-outcome denominators are 84 and 85. The registry record does not explain the difference, state whether every randomized participant received treatment, or describe participant disposition by group. It identifies the primary endpoint population as randomized participants who received at least one dose, but does not report how missing outcome information or treatment discontinuation was handled. The timeframe is described as up to 54 weeks for the primary endpoint and as a 52-week study period in the arm descriptions; the record does not explain the distinction.

Analysis reporting is limited. The registry posts a risk difference, confidence interval, and unadjusted p-value, but does not identify the test, interval method, confidence level, model, covariate adjustments, or prespecified analysis plan. It does not report multiplicity procedures for the listed secondary outcomes. No secondary outcome results are included in the primary comparison.

The safety summaries also lack context needed for a fuller interpretation. The record does not provide event-level detail, seriousness criteria, severity grading, relatedness judgments, exposure-adjusted rates, or a clear group-specific context for the listed serious-event term percentages. These gaps do not show that such information was absent from the full study documentation; they are limitations of what the registry record reports. The comparison therefore supports a description of posted counts and their uncertainty, not a complete assessment of safety or causal attribution.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Study of Gefapixant (MK-7264) in Japanese Adult Participants With Refractory or Unexplained Chronic Cough (MK-7264-038) — NCT03696108.
• Registry record last updated 2021-11-01; status last verified 2021-08.
• Results first posted 2021-11-01.
• 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

NCT03696108 is a completed Phase 3 randomized, double-blind, parallel-group study of gefapixant in Japanese adults with refractory or unexplained chronic cough. The registry lists enrollment of 175 and a primary objective of evaluating the safety of gefapixant 15 mg and 45 mg twice daily.

The primary endpoint is the number of participants with at least one adverse event during up to 54 weeks. The posted counts are 79 of 84 in the 15 mg group and 82 of 85 in the 45 mg group. The comparison’s risk-difference interval crosses zero, and the unadjusted p-value is 0.458. The result is imprecise and does not establish equivalence or causation.

The record also lists serious adverse events, other adverse events, deaths, and secondary Leicester Cough Questionnaire outcomes. Analysis methods, missing-data handling, and multiplicity procedures are not reported in the registry record.