Research on Nicorandil Treatment of Patients Diagnosed as CHD (Coronary Heart Disease) With Stable Angina: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Research on Nicorandil Treatment of Patients Diagnosed as CHD (Coronary Heart Disease) With Stable Angina (NCT01396395), a randomized phase 4 study. The primary outcome is Number of Myocardial Ischemia Attacks in 24 Hours.

2. Clinical Question

The clinical question is whether adding nicorandil to standard treatment is associated with a different number of myocardial ischemia attacks over 24 hours at week 12, compared with continuing standard treatment alone, among people with stable angina and coronary heart disease. The endpoint is based on attacks identified during 24-hour Holter monitoring. It is a measure of ischemic events recorded during a defined monitoring period, rather than a direct measure of symptoms, exercise capacity, or longer-term clinical events.

The comparison concerns two treatment strategies: standard treatment plus nicorandil and standard treatment alone. The registry describes random allocation and a parallel-group design, but no masking. Participants in the nicorandil group were to receive nicorandil alongside standard treatment for 12 weeks. The control group was to remain on standard treatment. The record says nitrates and beta blockers were to be maintained at stable doses.

3. Trial Design

The registry describes a phase 4, randomized, open-label, parallel-group, multicenter study. It lists an enrollment of 402 and a completed status. The record gives a start date of 2011-09 and a primary completion date of 2014-05. The sponsor is Merck KGaA, Darmstadt, Germany.

Participants with stable angina who were receiving standard treatment were assigned to standard treatment plus nicorandil or to standard treatment alone. The nicorandil regimen was 5 mg orally three times daily for 12 weeks. The control group was to continue standard treatment. The record says nitrates and beta blockers were to be kept at stable doses; other drugs not affecting the primary endpoint could be adjusted at investigators’ discretion. The description of the standard-treatment arm does not provide further detail.

Because the trial was open-label, treatment assignment was not masked. The registry does not report whether endpoint assessment or data analysis was masked. It identifies the primary endpoint as the number of myocardial ischemia attacks in 24 hours at week 12, evaluated using Holter monitoring. The registry’s outcome description supplies the electrocardiographic criteria used to identify an attack.

For the primary outcome, the record describes the full analysis set as randomized participants who received at least one dose of study treatment. It also specifies that the reported analyzed number is the number evaluable for the outcome. The group counts in the posted primary results are 141 for standard treatment plus nicorandil and 150 for standard treatment. Reasons for differences between enrollment, the analysis population, and outcome-evaluable counts are not reported in the registry record.

The original sample-size justification, planned handling of missing data, and detailed statistical analysis plan are not reported in the registry record. Planning illustrations based on observed effects are separate from the trial’s original design assumptions.

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: 4 subjects randomized to standard+nicorandil group were included in standard group for safety analysis as they did not receive nicorandil. 1 subject randomized to standard group was included in standard+nicorandil group as nicorandil was received. Thus, safety set have 197 and 205 in Standard+nicorandil and standard group, respectively.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS included all randomized subjects who received at least one dose of study treatment. N (number of subjects analyzed) signifies the number of subjects evaluable for this outcome measure.

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 analysis set (FAS) included all randomized subjects who received at least one dose of study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Number of Myocardial Ischemia Attacks in 24 Hours, measured over At Week 12.

The primary endpoint is the number of myocardial ischemia attacks recorded over 24 hours at week 12. Holter monitoring continuously records an electrocardiogram during the monitoring period. The registry defines an attack using ST-segment changes measured relative to the J point or baseline, with the specified change lasting at least 1 minute and at least 1 minute between attacks. The registry record describes the threshold as a horizontal or downward-sloping ST segment at least 0.1 mV below the reference.

This endpoint is used to quantify episodes of electrocardiographic ischemia during a fixed observation period. Counting attacks gives a frequency measure, while the accompanying monitoring criteria define which episodes contribute to that count. The endpoint therefore differs from a simple report of chest pain: it is based on recorded electrical changes meeting the stated criteria.

The posted results report a mean number of attacks per 24 hours for each group. A mean gives the average count in the analyzed group, but by itself does not show how counts are distributed across participants. The standard deviations in the registry indicate substantial variability in the observed counts relative to the group means. That variability is relevant when interpreting the precision of a comparison between group averages.

The endpoint is a measure at week 12, not a direct estimate of the effect on all possible manifestations of coronary disease or on outcomes beyond the monitoring period. The registry lists other outcome measures, but those do not change what the primary endpoint itself measures.

7. Results Posted in the Registry

• Standard Treatment: mean 3.80 versus 2.30 on control; difference 1.50 (95% CI -1.70 to 4.70), standardized effect d = 0.11; z-test p = 0.359.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline in Total Myocardial Ischemic Burden at Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Standard Treatment: mean 29.26 versus -38.63 on control; difference 67.89 (95% CI -49.47 to 185.25), standardized effect d = 0.14; z-test p = 0.257.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Standard Treatment and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Maximum ST-depression at Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Standard Treatment: mean 0.36 versus 0.25 on control; difference 0.11 (95% CI -0.73 to 0.95), standardized effect d = 0.03; z-test p = 0.798.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Standard Treatment and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Longest Duration of ST Segment Depression at Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Standard Treatment: mean 3.40 versus -3.80 on control; difference 7.20 (95% CI 0.09 to 14.31), standardized effect d = 0.24; z-test p = 0.047.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Standard Treatment than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists secondary outcomes spanning ischemic burden, ST-segment changes, heart-rate variability, arrhythmia, and a six-minute walk test measure. Fifteen secondary outcomes are posted. The available continuous-outcome comparisons do not show a uniform pattern: some estimates are close to zero, while another has a confidence interval that narrowly excludes zero and its unadjusted p-value is 0.047. The posted results therefore vary across outcome types and do not form a consistently aligned set of findings.

These secondary measures provide context about related aspects of ischemia and cardiac electrical activity, but they are not interchangeable with the primary count of attacks during Holter monitoring. Differences in measurement scale, variability, and endpoint definition mean that a result for one secondary measure does not confirm or overturn the primary endpoint result. The secondary results also cannot resolve the discrepancy between the sponsor’s Poisson models and the unadjusted primary mean comparison.

Unadjusted secondary results carry less weight when many outcomes are examined. As a general statistical principle, repeated testing increases the chance that at least one result appears unusual by chance, even if there is no corresponding underlying difference. The registry lists 16 outcome measures, but does not report a multiplicity-adjustment procedure or a prespecified hierarchy for interpreting secondary outcomes. Therefore, the isolated unadjusted result should be treated as exploratory rather than as definitive evidence on its own.

The registry record does not report the full results for every listed secondary outcome in the information available for this analysis. The overall pattern can be described only for the posted comparisons. It should not be taken as a complete account of every planned outcome or as proof that unposted outcomes were null.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change From Baseline in Total Myocardial Ischemic Burden at Week 12
• Change From Baseline in Maximum ST-depression at Week 12
• Change From Baseline in Longest Duration of ST Segment Depression at Week 12
• Percentage of Subjects Experienced Ischemic Heart Attack During the Six-minute Walk Test
• Heart Rate Variability (HRV) Rate: Time Domain
• Heart Rate Variability (HRV) Rate: Frequency Domain Power-24 Hour
• Number of Arrhythmia Occurred Within 24 Hours
• ECG QT Dispersion

9. Statistical Interpretation

10. Statistical Methods Explained

Several statistical ideas help distinguish the posted summaries from the sponsor’s reported analysis. The group summaries describe the observed data, the recalculation compares group means using those summaries, and the Poisson regression results come from the trial’s own reported modeling. These approaches answer related but not necessarily identical statistical questions.

What does a mean difference describe?

A mean is the arithmetic average of the recorded count across participants in a group. The primary results report means of 2.3 and 3.8 attacks per 24 hours. The unadjusted recalculation expresses the contrast as standard treatment minus standard treatment plus nicorandil, giving a mean difference of 1.5. A positive value in that stated contrast means the posted mean is higher in the standard-treatment group. This contrast does not adjust for other variables or account for any model structure beyond the summary-statistic calculation.

The recalculated confidence interval runs from -1.704 to 4.704. Because zero represents no difference in means, an interval that includes zero is compatible with no difference as well as with differences in either direction. The associated unadjusted p-value is 0.359. This calculation is a simple unadjusted reanalysis of the posted summaries; it is not a replacement for the trial’s reported analysis.

Why use a count model?

Poisson regression is a general statistical method for modeling event counts in relation to explanatory variables. It commonly expresses a group contrast as a ratio rather than a difference. A ratio of 1 represents equal modeled rates under the model’s comparison; values on either side represent relative differences according to the model’s group coding. The registry reports a ratio of 0.503 for both the non-calibration and calibration analyses. It does not provide enough detail to establish all aspects of the model’s coding and implementation from the registry record alone.

A basic Poisson model makes assumptions about how counts vary around their expected values. Over-dispersion occurs when observed count variability is greater than the model allows. In general, if over-dispersion is not handled appropriately, standard errors and confidence intervals can be too narrow, making the apparent precision misleading. The primary outcome’s posted standard deviations are 8.46 and 18.02, while the corresponding means are 2.3 and 3.8. This large spread relative to the means makes the treatment of variability an important issue to consider. The registry labels one analysis as a calibration model, but does not explain its calibration method, the model diagnostics, or whether over-dispersion was assessed.

It would be inappropriate to assume that the two Poisson analyses use particular corrections or that the calibration model resolves a specific diagnostic problem. Such details are not reported in the registry record. The reported ratios and intervals can be described as they stand, while their implementation cannot be fully audited from the posted summary alone.

Why do the two approaches differ?

A comparison of means and a regression model can differ because they use different assumptions and may target different summaries of the data. The mean-difference recalculation uses the posted group means, standard deviations, and analyzed counts. The sponsor’s analyses use Poisson regression and report ratios. The registry does not provide participant-level data or enough model details to reproduce the regression estimates from the summary statistics.

In this trial, the unadjusted mean-difference interval includes zero and its p-value is 0.359, while both sponsor-reported ratio intervals exclude 1 and their reported p-values are below 0.05. This is a genuine discrepancy in statistical conclusions, not merely a difference in how the same number is displayed. Because the sponsor’s own analysis takes precedence when the posted recalculation conflicts with it, the regression results should be identified as the trial’s reported analysis and the recalculation as an unadjusted comparison of summaries.

What does a p-value tell the reader?

As a general statistical concept, a p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. The p-values reported for the sponsor’s analyses are tied to those models and their assumptions. The recalculated p-value is tied to the unadjusted mean comparison. They should not be treated as interchangeable.

The registry does not report whether the analyses were adjusted for multiple outcome testing, whether there were interim analyses, or the full prespecified statistical analysis plan. As a result, the intended hierarchy among outcomes and the scope of any confirmatory interpretation cannot be established from the registry record.

11. Confidence Intervals and Precision

A confidence interval describes a range of estimates compatible with the observed data under the assumptions of the analysis. Its width communicates precision: a narrower interval indicates less statistical uncertainty about the estimated contrast, while a wider interval indicates more. An interval is not a range that contains every participant’s result, and it does not account for limitations that the statistical model leaves unaddressed.

For the unadjusted mean difference in the primary endpoint, the estimate is 1.5 attacks per 24 hours and the 95% confidence interval is -1.704 to 4.704. The interval spans zero and includes values on both sides of no mean difference. Its width therefore leaves considerable uncertainty about the magnitude and direction of the mean contrast. The associated p-value is 0.359.

The sponsor’s non-calibration Poisson regression gives a ratio of 0.503 with a 95% confidence interval of 0.435 to 0.581. Its calibration-model result is also 0.503, with a 95% confidence interval of 0.301 to 0.840. Both intervals exclude 1, the no-ratio-difference value. The calibration-model interval is wider, indicating less precision in that reported ratio estimate than in the non-calibration result. The estimates are model-based and should not be directly substituted for the mean difference.

These intervals depend on the assumptions of their respective analyses. The registry provides no participant-level data, model diagnostics, or complete specifications that would allow independent assessment of those assumptions. The open-label design, outcome-evaluable analysis population, and unreported missing-data approach are also relevant to how much confidence readers place in the estimates. Statistical precision is only one part of evaluating evidence; it does not establish clinical importance or remove possible sources of bias.

12. 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 figures are planning illustrations based on the observed effect, not the trial’s original design calculation. For these continuous-outcome comparisons, the figures refer to participants per arm. They are not total numbers across the trial and do not refer to total events. The original sample-size justification and its assumptions 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 asks how many participants per arm would be needed to detect a specified fraction of the observed effect. The rows describe planning for 100%, 75%, or 50% of the observed effect. The columns for 80% and 90% power represent different chances, under those planning assumptions, of detecting the specified effect if it were present. Higher planned power generally requires more participants for the same effect size.

Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants per arm. For example, the primary endpoint illustration gives 1,410 participants per arm for 80% power when planning for 100% of the observed effect, and 5,640 participants per arm when planning for 50% of the observed effect. The corresponding 90% power illustrations are 1,888 and 7,550 participants per arm. These are hypothetical planning values, not observed trial outcomes.

This is not observed power. It is a forward-looking illustration based on the observed effect and specified assumptions, and it should not be used to judge whether the completed trial had an adequate design without the original design rationale and prespecified assumptions.

13. 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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 1.500; pooled standard deviation σ = 14.216.
2. n = 2 × (1.96 + 0.842)² × 14.216² / 1.500² = 1,409.9, rounded up to 1,410 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
delta <- 1.5                      # observed difference in means
sigma <- 14.2157575711532                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1396395)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 1,410 (80% power), 1,888 (90% power)
 75% of observed effect: n per arm = 2,507 (80% power), 3,356 (90% power)
 50% of observed effect: n per arm = 5,640 (80% power), 7,550 (90% power)
power.t.test: 1410.9 
Simulated power at n = 1,410 per arm: 0.802 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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.

14. Analyses Reported by the Sponsor

The registry’s reported primary analyses are Poisson regression models. The non-calibration analysis reports a ratio of 0.503, with a 95% confidence interval of 0.435 to 0.581 and p < 0.001. The calibration model reports the same ratio estimate, with a 95% confidence interval of 0.301 to 0.840 and p = 0.009. Both reported confidence intervals exclude 1.

The simple unadjusted recalculation instead compares the posted group means. It gives a mean difference of 1.5 attacks per 24 hours, with a 95% confidence interval of -1.704 to 4.704 and p = 0.359. That interval includes zero. The two approaches therefore lead to different statistical conclusions: the sponsor-reported models indicate a ratio differing from 1, while the unadjusted mean comparison does not establish a difference from zero.

The sponsor’s own analysis takes precedence over the recalculation. The recalculation is useful as a transparent comparison of the posted means, but it does not reproduce the Poisson regression. The registry does not report enough detail about model specification, group coding, or calibration to fully reconcile the results. Readers should keep the model-based result and the simple mean comparison distinct rather than treating one as a restatement of the other.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when statistical tests are conducted for several outcomes, comparisons, or analysis choices. Even if each test uses the same nominal significance threshold, conducting many tests increases the chance of at least one apparently notable result arising by chance. Common approaches include adjusting the threshold or controlling an error rate across a prespecified family of tests. The appropriate approach depends on which outcomes and comparisons were planned as confirmatory.

The registry lists 16 outcome measures and 15 posted secondary outcomes. It does not report a multiplicity-adjustment method, a testing hierarchy, or which secondary outcomes were intended to be confirmatory. As a result, the unadjusted secondary p-values should not be interpreted as if they were necessarily protected against the risks of multiple testing. The isolated secondary p-value of 0.047 is especially difficult to assess without knowing the prespecified analysis plan and the family of tests to which it belongs.

The registry reports two Poisson regression analyses for the primary endpoint, one labeled non-calibration and one labeled calibration model. It does not explain how these analyses were related in the prespecified plan, whether one was primary and the other a sensitivity analysis, or how their results were intended to be reconciled. The full statistical analysis plan and any interim analyses are not reported in the registry record.

Missing-data handling is also not reported. The primary-outcome description says the analysis population included randomized participants who received at least one dose, and that the analyzed number is the number evaluable for the outcome. That definition does not explain why some participants were not evaluable, how missing measurements were treated, or whether any imputation or sensitivity analyses were used. The record also does not report whether the analysis followed all randomized participants regardless of treatment received.

These omissions limit assessment of the robustness of the reported estimates and the intended strength of inference. They do not show that a particular method was or was not used; those details are simply not reported in the registry record.

16. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: sudden death (0.5%), angina (0.5%), supraventricular tachycardia (0.5%), unstable angina (0.5%), coronary artery disease (0.5%).

The registry reports serious adverse events for 9 of 197 participants at risk in the standard-treatment-plus-nicorandil group and 7 of 205 participants at risk in the standard-treatment group. It also reports other adverse events affecting 12 of 197 participants at risk in the nicorandil group and 1 of 205 in the standard-treatment group. These are event counts and reported at-risk group sizes; they should be read as safety summaries, not as adjusted comparisons.

The listed serious adverse event terms each have a reported frequency of 0.5%: sudden death, angina, supraventricular tachycardia, unstable angina, coronary artery disease, and peripheral edema. The registry listing does not assign those terms to a particular arm in the information, so an arm-specific interpretation of each term is not supported by the record. Death counts are not reported in the registry record.

Safety counts need context, including how events were defined, collected, and assessed for relatedness. Those details are not reported in the registry record. The group counts alone do not establish that a treatment caused or prevented an event, and they do not provide a complete assessment of safety.

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

The results describe a comparison of week 12 Holter-recorded myocardial ischemia attack counts in a randomized study of standard treatment plus nicorandil versus standard treatment alone. The sponsor’s reported Poisson regression analyses give a ratio of 0.503, with confidence intervals excluding 1. The simple unadjusted recalculation of the posted means gives a mean difference of 1.5 attacks per 24 hours, but its confidence interval includes zero and its p-value is 0.359. The sponsor’s own analysis takes precedence where these results conflict, while the discrepancy remains important for transparent interpretation.

The primary endpoint is a defined electrocardiographic count over a limited monitoring period. It does not, by itself, establish an effect on symptoms, quality of life, heart attacks, survival, or outcomes beyond week 12. The listed secondary outcomes address other measures, but the registry does not report a multiplicity strategy, and secondary results require caution.

The findings do not establish that the treatment is beneficial or ineffective for every person with stable angina. Statistical significance is not the same as clinical importance, and an estimate from a model does not remove uncertainty about the model assumptions. The open-label design and unreported details about missing data, analysis procedures, and outcome assessment further limit what can be concluded from the registry record alone.

The appropriate reading is therefore narrow: the posted data and sponsor-reported models concern a specific endpoint, comparison, and follow-up time. They do not answer broader treatment questions, and the registry record does not provide all information needed to independently verify or fully contextualize the analyses.

18. Why This Trial Matters Statistically

This trial illustrates why statistical conclusions depend on the analysis as well as the observed group summaries. The posted means and standard deviations permit a simple unadjusted comparison, while the sponsor’s reported Poisson regressions analyze the count outcome using a different model and report ratios. In this case, those approaches do not lead to the same inferential conclusion. Keeping the estimands and assumptions distinct is essential for accurate interpretation.

The endpoint is a count, and the reported standard deviations are large relative to the group means. That feature makes the modeling of variability consequential. The registry names a calibration model but does not explain its construction or provide diagnostics. This leaves readers able to report the sponsor’s numerical results, but unable to fully audit the model choices from the registry summary.

The study also highlights the difference between a primary outcome and a collection of secondary outcomes. The primary endpoint has a defined monitoring procedure, while the secondary outcomes cover several related measurements. Without a reported multiplicity plan, isolated secondary p-values should be interpreted cautiously. The available information does not establish how the outcome hierarchy was intended to guide conclusions.

Finally, the planning illustrations show how sample-size requirements depend on the effect a future study aims to detect and the desired power. They are not evidence about the power of the completed trial. The original design assumptions are not reported, so the illustrations cannot substitute for the trial’s own sample-size rationale. Together, these issues make the record useful for teaching careful distinctions among descriptive summaries, model-based inference, precision, and design planning.

19. 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 leaves important details unavailable for assessing the primary analysis. It reports the names and results of two Poisson regression approaches but does not give a complete model specification, group coding, calibration procedure, or model diagnostics. The posted summary statistics do not permit independent reconstruction of those regressions. The difference between the sponsor’s findings and the unadjusted mean comparison therefore cannot be fully explained from the registry record.

The study was open-label, and the registry does not report whether endpoint assessment or statistical analysis was masked. The primary endpoint relies on Holter interpretation using specified electrocardiographic criteria, but information about assessor procedures and reproducibility is not reported. The registry describes the analysis population as randomized participants who received at least one dose and says the analyzed count represents those evaluable. Reasons for non-evaluability and the handling of missing measurements are not reported.

The record does not report the original sample-size calculation, its assumptions, or the full statistical analysis plan. It also does not report whether interim analyses occurred, how multiple outcomes were handled, or what hierarchy was planned across outcomes. These omissions make it difficult to assess prespecification and the confirmatory status of the secondary findings.

Only the posted outcome summaries can be interpreted. The record does not provide participant-level distributions, so the shape of the count data, the influence of unusually high counts, and the relationship between repeated measurements cannot be examined. The results are also specific to the reported week 12 endpoint and do not supply evidence about outcomes beyond the period described.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Research on Nicorandil Treatment of Patients Diagnosed as CHD (Coronary Heart Disease) With Stable Angina — NCT01396395.
• Registry record last updated 2016-04-14; status last verified 2016-03.
• Results first posted 2016-04-14.
• 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.

23. Record Summary

This completed phase 4 study compared standard treatment plus nicorandil with standard treatment alone in people with stable angina and coronary heart disease. It was randomized, open-label, parallel-group, and multicenter, with an enrollment of 402. The intervention was nicorandil 5 mg three times daily for 12 weeks alongside standard treatment.

The primary endpoint was the number of myocardial ischemia attacks in 24 hours at week 12, assessed by Holter monitoring. The sponsor’s Poisson regression analyses report a ratio of 0.503, while a simple unadjusted recalculation of the posted group means gives a mean difference of 1.5 with a confidence interval including zero. Because the results conflict, the sponsor’s own analysis takes precedence. The registry does not report enough methodological detail to fully reconcile the analyses or assess missing-data and multiplicity procedures.