Clinical Study to Evaluate the Efficacy and Safety of Givinostat in Ambulant Patients With Duchenne Muscular Dystrophy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Clinical Study to Evaluate the Efficacy and Safety of Givinostat in Ambulant Patients With Duchenne Muscular Dystrophy (NCT02851797), a randomized phase 3 study. The primary outcome is Mean Change From Baseline in 4 Standard Stairs (4SC) Climb After 18 Months of Treatment.

2. Clinical Question

3. Trial Design

The registry describes a randomized, double-blind, placebo-controlled, multicentre, parallel-group phase 3 study. It lists quadruple masking and an enrollment of 179. The study is marked completed. The recorded start date is 2017-06-06, and the primary completion date is 2022-02-22. Italfarmaco is listed as sponsor.

Participants were assigned to givinostat or placebo. Both treatments were oral suspensions at a concentration of 10 mg/mL, taken twice daily. The primary objective was to establish the effects of chronic givinostat versus placebo over 18 months on slowing disease progression in ambulant subjects with Duchenne muscular dystrophy. Secondary objectives included safety and tolerability, pharmacokinetics, quality of life, and activities of daily living.

The primary outcome was the mean change from baseline in the time to climb 4 standard stairs after 18 months. The registry also lists outcomes concerning time to rise from the floor, the six-minute walking test, the North Star Ambulatory Assessment, loss of function, muscle strength, muscle fat fraction, adverse events, and acceptability or palatability. The record lists eight secondary outcomes as posted and nine outcome measures overall.

The primary endpoint results list 81 in the givinostat group and 39 in the placebo group. Those analysis counts are not the same as the safety risk counts, which are reported separately. The registry does not report the original sample-size justification, the allocation ratio planned in the protocol, or the detailed procedures used to maintain masking. The listed planning assumptions for a recalculated sample-size illustration should not be mistaken for 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT analysis set: the intent-to-treat (ITT) analysis set included all subjects who were randomised, received at least one dose of study drug, and had at least 1 non-missing post-baseline 4SC measure or missing post-baseline 4SC measure due to being either non-ambulatory or otherwise physically unable to perform the assessment, irrespective of any deviation from the protocol or premature discontinuation.

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: Number of Subjects Enrolled in Each Country (Overall Population).The all-enrolled subjects set included all subjects who were randomised to study treatment. The overall all-enrolled subjects set included all subjects in both the target and off-target populations (Group A + Group B) who were randomised to study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change From Baseline in 4 Standard Stairs (4SC) Climb After 18 Months of Treatment, measured over Baseline and 18 months.

The primary endpoint is the mean change from baseline in the time needed to climb 4 standard stairs, measured in seconds, after 18 months of treatment. Stair climbing is a functional task: it captures an aspect of mobility that can be assessed repeatedly using a timed performance test. The registry describes it as a functional test of stair-climbing ability, administered in a standardized manner by qualified functional evaluators who were separate from site personnel reviewing safety results.

The direction of the measure matters. A lower time represents a better outcome, so a negative difference in change between givinostat and placebo can indicate a relative advantage for givinostat if the comparison is expressed as givinostat minus placebo. The endpoint is a continuous measure: values can vary across a range, rather than simply recording whether an event did or did not occur. A mean summarizes the average change in each group, while the comparison between groups addresses whether their average changes differ.

Baseline was defined using the randomization assessment unless that measurement was missing, in which case the last nonmissing value recorded before or on the date of first study treatment was used. The registry states that the primary measure was evaluated at baseline and 18 months. The posted results do not include enough information to reconstruct every participant's trajectory or to establish how all missing observations were handled.

7. Results Posted in the Registry

• Givinostat: mean 1.27 versus 1.48 on control; difference -0.21 (95% CI -0.35 to -0.07), standardized effect d = -0.58; z-test p = 0.003.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous, and the registry describes the arm-level values as least-squares means with standard errors. A continuous measure records quantities such as time, where differences between values are meaningful. A mean summarizes the average in a group. A standard error describes uncertainty in an estimated mean; it is not the same as the spread of individual participants' measurements.

What does the simple recalculation represent?

The recalculation compares the posted givinostat and placebo means without adjustment for other variables. It reports a mean difference of -0.21 seconds, an interval from -0.348 to -0.072, and an unadjusted p-value of 0.003. The standard deviations used for this calculation are approximate and were derived from the posted standard errors. It is therefore a summary-statistic recalculation, not a recreation of the trial's full participant-level analysis. The registry record does not state the confidence level for this recalculated interval.

A difference of means is interpreted in the units of the endpoint, here seconds. Given the endpoint's direction, a lower time is better. The sign of the estimate depends on the order of subtraction; the posted comparison is expressed as givinostat minus placebo. The interval conveys the range of values compatible with the calculation under its assumptions. An interval that does not include zero is consistent with a difference under that calculation, but this does not resolve the discrepancy with the sponsor's adjusted result.

What is ANCOVA doing?

The sponsor's reported method is ANCOVA, or analysis of covariance. In general, ANCOVA compares groups while adjusting for specified covariates. Adjustment can account for measured baseline characteristics or other variables included in the model, and can produce an estimate different from an unadjusted comparison. The registry reports a generalized least squares mean ratio of 0.86, a 95% confidence interval from 0.745 to 0.989, and p = 0.035. The covariates, model details, and exact estimation procedures are not reported in the registry record, so the analysis cannot be independently reconstructed from the posted summary.

The sponsor's reported ratio and the simple recalculated difference are different estimands: a ratio is not measured in seconds, whereas a difference in times is. Their numerical values and intervals should not be compared as if they were on a common scale. Because the registry marks a discrepancy between the sponsor result and the recalculation, the sponsor's own analysis takes precedence.

What does the p-value say?

As a general statistical concept, a p-value describes how unusual data at least as incompatible with a specified null model would be, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, nor a measure of the size or practical importance of an effect. The unadjusted p-value of 0.003 belongs to the recalculation. The reported p = 0.035 belongs to the sponsor's ANCOVA. They answer questions under different analyses and should remain clearly distinguished.

Why can over-dispersion matter?

As a general concept, over-dispersion means that variability in observed data is greater than a model's assumed variability. When a model underestimates variation, standard errors can be too small and confidence intervals can be too narrow. This issue is especially relevant to count outcomes, but it can also arise in other model settings. The primary endpoint is a continuous timed measure, and the registry does not report an over-dispersion assessment or a correction of that kind. It would be inappropriate to assume one was needed or applied.

How should the analysis population be understood?

The registry describes the primary analysis set as an intent-to-treat set comprising randomized subjects who received at least one dose and had at least one nonmissing post-baseline 4SC measure, with an exception for specified missing measures related to being non-ambulatory or otherwise physically unable. The full wording of the population definition is not available in the record. This description is not enough to infer a conventional analysis of every randomized participant regardless of treatment received, and the detailed handling of missing data is not reported.

The registry reports 81 for the givinostat arm and 39 for placebo in the primary endpoint results. These are endpoint analysis counts. Safety uses separate risk counts. Differences in reported counts can reflect different analysis sets or outcome availability, but the registry does not give enough detail to identify the specific reasons. The enrollment figure should not be substituted for the endpoint-specific analysis counts.

In general, an analysis plan can specify the primary model, covariates, missing-data rules, and any adjustments for multiple outcomes or interim looks. Those details matter because they influence how estimates and uncertainty are interpreted. The statistical plan, missing-data methods, and multiplicity procedures are not reported in the registry record. No protocol details should be inferred from the fact that ANCOVA is named.

10. Confidence Intervals and Precision

A confidence interval expresses uncertainty around an estimate under the assumptions of the analysis that produced it. It shows a range of values compatible with the data and model; it does not guarantee that the true effect lies inside the interval, and it is not a range containing most individual participant outcomes.

The sponsor's ANCOVA reports a generalized least squares mean ratio of 0.86, with a 95% confidence interval from 0.745 to 0.989. The interval lies below 1, but its upper endpoint is near 1, so the estimate's direction is more certain than its exact magnitude. The width reflects uncertainty under the sponsor's model. The registry does not give enough model detail to assess how covariate choices or missing-data handling affected this precision.

The summary-statistic recalculation reports a mean difference of -0.21 seconds, with an interval from -0.348 to -0.072. This is an unadjusted result on the time scale, not the ratio scale used by the sponsor. The recalculated standard deviations are approximate, and the analysis differs from the sponsor's ANCOVA. The interval should therefore be read as precision for that simple recalculation rather than as a substitute for the sponsor's reported interval.

Precision is distinct from clinical importance. Even a relatively narrow interval may leave uncertainty about how meaningful a difference is to patients, while a result with an interval excluding a null value does not describe variation in individual outcomes. The registry does not report a threshold for a clinically important change in this endpoint, so the posted estimates cannot be classified against such a threshold using the record alone.

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 illustration relates hypothetical planning numbers to fractions of the observed effect. The rows describe planning for 100%, 75%, and 50% of the observed effect. The figures are participants per arm, not total participants or total events. For planning for 100% of the observed effect, the illustration gives 47 participants per arm for 80% power and 63 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 83 participants per arm for 80% power and 111 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 186 participants per arm for 80% power and 249 participants per arm for 90% power.

In plain language, the 80% power column describes the per-arm sample size that the stated calculation estimates would give an 80% chance of detecting the specified effect under its assumptions. The 90% power column gives the per-arm size for a higher chance of detecting that effect under those same assumptions. These are planning illustrations, not observed power.

The 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 the stated fraction of the observed effect. The illustration is calculated from the observed effect and is not the trial's original design justification. The original design assumptions and sample-size justification are not reported in the registry record.

Smaller target effects require more participants because they are harder to distinguish from random variation. The displayed curve should be read as a planning aid showing this relationship, not as evidence about the trial's actual power or an account of how the trial was designed. The note that time-to-event calculations use total events does not apply to this continuous primary outcome; its planning figures are per arm.

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

Worked calculation for this trial

1. Observed difference Δ = -0.210; pooled standard deviation σ = 0.361.
2. n = 2 × (1.96 + 0.842)² × 0.361² / -0.210² = 46.3, rounded up to 47 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 <- -0.21                      # observed difference in means
sigma <- 0.36071313547152                     # 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(2851797)
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 = 47 (80% power), 63 (90% power)
 75% of observed effect: n per arm = 83 (80% power), 111 (90% power)
 50% of observed effect: n per arm = 186 (80% power), 249 (90% power)
power.t.test: 47.3 
Simulated power at n = 47 per arm: 0.798 (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.

13. Analyses Reported by the Sponsor

The sponsor's own primary analysis is reported as ANCOVA, with a generalized least squares mean ratio of 0.86, a 95% confidence interval from 0.745 to 0.989, and p = 0.035. This is the trial's reported analysis and takes precedence over the simple recalculation.

The recalculation uses the posted arm-level summary statistics to produce an unadjusted difference of -0.21 seconds, with an interval from -0.348 to -0.072 and p = 0.003. It is not adjusted as the sponsor analysis is, uses a different effect scale, and relies on approximate standard deviations. The two estimates therefore are not directly interchangeable.

The numerical discrepancy is recorded, but the registry does not explain its cause. The covariates and full model specification for the sponsor's ANCOVA are not reported, so the calculation cannot be reproduced from the registry summary alone. The sponsor result should be the principal reference for the reported primary finding; the recalculation is useful as a transparent, unadjusted comparison of the posted group summaries, not as a replacement analysis.

14. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a trial evaluates multiple outcomes, time points, or statistical comparisons. In general, repeated testing can increase the chance of finding at least one result that appears statistically notable by chance. The registry lists one primary endpoint and several secondary outcome areas, including timed function tests, walking performance, assessment scores, muscle measures, adverse events, and suspension acceptability. It does not report whether or how the analyses were adjusted for multiple comparisons, or which secondary outcomes were designated in a formal testing hierarchy.

The registry lists eight posted secondary outcomes and nine outcome measures overall. This does not establish that every outcome was tested in the same analysis population or under a single multiplicity procedure. The statistical analysis plan, multiplicity adjustment, and any interim analyses are not reported in the registry record. Consequently, the primary result should be distinguished from secondary findings, and a collection of nominal p-values should not be interpreted as if it were protected by a known familywise error procedure.

Missing data can also affect estimates and their precision. For the 4SC endpoint, the registry's population description refers to post-baseline measurements and includes an exception for certain cases where a measure was missing because a subject was non-ambulatory or otherwise physically unable. The wording available in the record is incomplete. The specific method for imputing, excluding, or otherwise accounting for missing observations is not reported. It is also not reported whether sensitivity analyses were used to test assumptions about missingness.

The endpoint results list 81 in the givinostat group and 39 in the placebo group, while the safety risk counts are reported separately. These values should be understood in the context of their stated analyses rather than treated as interchangeable denominators. The registry does not explain all differences in the analysis sets or give enough information to reconstruct participant flow through each outcome analysis.

These reporting gaps limit independent assessment of robustness. They do not prove that a particular method was or was not used; they mean only that the method is not reported in the registry record. Interpretation should therefore keep the sponsor's stated primary analysis distinct from unadjusted recalculations and avoid making claims about unreported multiplicity or missing-data procedures.

15. 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: gastroenteritis (3.3%), gastroenteritis viral (0.8%), influenza (0.8%), abdominal pain (0.8%), vomiting (0.8%).

Safety risk counts are reported separately from the primary endpoint analysis counts. In the givinostat group, 8 of 118 at risk experienced a serious adverse event; in the placebo group, 2 of 61 at risk did so. Other adverse events were reported for 112 of 118 at risk in the givinostat group and 57 of 61 at risk in the placebo group. The registry lists 0 deaths in each group.

The serious adverse event terms with the highest posted percentages include gastroenteritis at 3.3%, and gastroenteritis viral, influenza, abdominal pain, vomiting, and chest pain at 0.8% each. These percentages describe the listed event terms; the registry summary does not provide enough context to infer duration, recurrence, exposure-adjusted rates, or causal relationship. The counts and terms should not be treated as a complete account of each participant's safety experience.

These figures are descriptive. They do not by themselves show whether an event was caused by treatment, and comparisons are limited by the reported risk counts and absence of further detail in the registry record. Safety results address a different question from the primary functional endpoint and should be interpreted independently.

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

The posted sponsor analysis reports a primary endpoint result favoring a lower generalized least squares mean ratio for givinostat relative to placebo, with a 95% confidence interval from 0.745 to 0.989 and p = 0.035. The endpoint is time to climb 4 standard stairs, for which a shorter time is better. The reported interval is below 1 under the sponsor's ratio-based analysis. This describes the result of that model in this trial, not a guarantee of the same result in every patient or setting.

The simple unadjusted recalculation also has a negative difference in seconds, but it is not the sponsor's model and should not replace it. The discrepancy matters: the sponsor result is the primary reference, while the recalculation is an approximate comparison based on posted summaries. Neither the p-value nor the confidence interval alone establishes the practical importance of the effect. The registry does not report a clinically important threshold for the 4SC measure, and the full model and missing-data procedures are not reported.

The result does not show that every participant improved, that the effect was uniform, or that changes in stair climbing necessarily correspond to changes in other functional outcomes. The registry lists secondary measures, but their presence alone does not establish results for them. The findings also do not settle long-term outcomes beyond the stated 18-month treatment period or establish generalizability beyond the studied ambulant DMD population.

Safety counts should be considered alongside, but separately from, the efficacy endpoint. Serious adverse events were recorded in both groups, and no deaths are listed. These summaries do not establish causation or provide a full characterization of safety. In general, trial results are evidence about group outcomes under the trial's conditions; they are not individualized predictions or medical advice.

17. Why This Trial Matters Statistically

This trial is statistically informative because it combines a randomized, placebo-controlled, parallel-group design with a functional outcome that is directly measured as time. Random allocation supports a between-group comparison, while masking can reduce some forms of measurement and expectation bias. The registry records the study as double blind and quadruple masked, although it does not specify which roles were masked.

The example also illustrates why the analysis method matters. A simple comparison of arm-level means produces a difference in seconds, while the sponsor's ANCOVA reports a ratio after adjustment. These are not the same estimand, and their p-values should not be read as competing votes. The sponsor's own analysis takes precedence, while the recalculation offers a limited, transparent check on the posted summaries. The discrepancy highlights the importance of reporting model details and analysis populations clearly.

Precision is also central to interpretation. The sponsor's interval gives a range for the estimated ratio and shows that the result is not known as a single exact value. However, statistical evidence against a null value does not determine whether the size of the difference is important to patients. The registry does not report a clinically important threshold, full covariate details, or missing-data methods.

Finally, the sample-size illustration demonstrates how planning depends on the effect one aims to detect: smaller effects require more participants under the stated assumptions. Those calculations are illustrations based on the observed effect, not the original trial design and not observed power. Together, these features make the record a useful case for separating design, estimation, precision, significance, planning, and clinical meaning.

18. 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 summary does not include the full statistical analysis plan. It names ANCOVA for the sponsor's primary analysis but does not report the covariates, detailed model specification, or procedures used to handle missing observations. The primary analysis population description is truncated, so its full inclusion rules cannot be verified from the registry record.

The posted comparison also has a documented discrepancy between the sponsor's analysis and the unadjusted recalculation. The recalculation depends on approximate standard deviations derived from standard errors and does not reproduce the sponsor's adjusted model. Its interval and p-value should therefore be treated as a descriptive recalculation rather than an independent confirmation. The sponsor analysis takes precedence, but the registry does not give enough detail to explain the numerical difference.

The primary endpoint results list 81 for givinostat and 39 for placebo, while the adverse-event risk groups list 118 and 61. These counts belong to different reported contexts; the record does not explain the participant flow or the full reasons for differences. Enrollment is listed as 179, but it should not be used as a substitute for the endpoint-specific counts.

The record lists multiple secondary outcome areas but does not report multiplicity procedures or a testing hierarchy. It also does not report the trial's original sample-size justification, interim analyses, or a clinically important threshold for the 4SC outcome. The planning illustration is based on the observed effect and stated assumptions, so it cannot fill those gaps.

Safety information is summarized through counts and selected serious adverse event terms. The registry summary does not provide all details needed to assess timing, duration, exposure, recurrence, relatedness, or comparative rates. These limitations narrow what can be concluded from the posted record; they do not establish that unreported procedures or analyses were not performed.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Clinical Study to Evaluate the Efficacy and Safety of Givinostat in Ambulant Patients With Duchenne Muscular Dystrophy — NCT02851797.
• Registry record last updated 2023-02-02; status last verified 2023-01.
• Results first posted 2023-02-02.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The registry records NCT02851797 as a completed phase 3 randomized, double-blind, placebo-controlled, multicentre, parallel-group study sponsored by Italfarmaco. It enrolled ambulant patients with Duchenne muscular dystrophy and compared givinostat oral suspension with placebo oral suspension, both given twice daily. The primary endpoint was mean change from baseline in the time to climb 4 standard stairs after 18 months.

The sponsor's reported ANCOVA gives a generalized least squares mean ratio of 0.86, with a 95% confidence interval from 0.745 to 0.989 and p = 0.035. An unadjusted recalculation is also posted, but the sponsor analysis takes precedence. Safety summaries report serious adverse events for 8 of 118 at risk with givinostat and 2 of 61 at risk with placebo, with 0 deaths listed in each group. Important details about the model, missing data, multiplicity, and original sample-size justification are not reported in the registry record.