Safety and Efficacy of EN3835 in Participants With Frozen Shoulder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy of EN3835 in Participants With Frozen Shoulder (NCT04496167), a randomized phase 2 study. The primary outcome is Change From Baseline in the Adapted American Shoulder and Elbow Surgeons Shoulder Assessment Form (ASES) Composite Score for the Affected Shoulder at Day 95.

2. Clinical Question

3. Trial Design

This was a Phase 2 randomized, double-blind, placebo-controlled study with a parallel-group design. The registry lists 198 enrolled participants and identifies EN3835 and placebo as the two interventions. The EN3835 arm description says “up to 1.74 mg”; the placebo arm is described as placebo. The registry record does not report the treatment schedule in the intervention details summarized for this trial.

The condition under study was adhesive capsulitis of the shoulder, or frozen shoulder. The primary endpoint was change from baseline in the adapted American Shoulder and Elbow Surgeons Shoulder Assessment Form composite score for the affected shoulder at Day 95. The endpoint description specifies a modified intent-to-treat population that required receipt of at least 1 injection, a valid baseline score, and at least 1 valid score after injection. This population definition is narrower than simply describing all enrolled participants.

The registry gives the start date as 2020-07-29, the primary completion date as 2022-04-14, and the status as completed. It lists 31 outcome measures and 30 posted secondary outcomes, including changes in passive range of motion for several shoulder movements. The registry record does not report the original sample-size justification, the planned duration of treatment, or further details of the randomization and masking procedures in the trial details summarized for this analysis. Those details should not be inferred from the general design labels.

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: Modified intent-to-treat population included all randomized participants who received at least 1 injection of study treatment and who had a valid baseline adapted ASES composite score in the affected shoulder and at least 1 valid adapted ASES composite score after the injection of the study treatment and had both Baseline and Day 95 measurements. Here, overall number of participants analyzed = participants 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: All randomized participants who received at least 1 injection of study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in the Adapted American Shoulder and Elbow Surgeons Shoulder Assessment Form (ASES) Composite Score for the Affected Shoulder at Day 95, measured over Baseline, Day 95.

The primary endpoint is the change from baseline to Day 95 in the adapted ASES composite score for the affected shoulder. It is a participant-reported measure combining pain and function. Its pain component is based on a single 11-point rating that is transformed to a score from 0, worst pain, to 50, no pain. Its function component covers 10 activities, with responses transformed to a score from 0, no function, to 50, full function. The composite is the sum of these components, ranging from 0, worst pain, to 100, least pain. Positive change from baseline means improvement in pain and/or shoulder function.

This kind of endpoint is useful for describing a patient’s reported symptoms and ability to perform activities, rather than relying only on a physical measurement. The composite also brings pain and function into one outcome, with each contributing 50% of the composite score. That design means a change in the total can reflect pain, function, or both; the composite alone does not show which component accounts for a change.

The reported arm means for change were 41.35 for EN3835 and 41.16 for placebo. The sponsor’s primary comparison was a least-squares mean treatment difference. The registry describes the analysis population and score construction, but the record does not report a threshold for interpreting a change as clinically important. Statistical estimates should therefore be read as score differences and uncertainty, not as proof of a particular level of patient-perceived benefit.

7. Results Posted in the Registry

• EN3835: mean 41.35 versus 41.16 on control; difference 0.19 (95% CI -6.56 to 6.94), standardized effect d = 0.01; z-test p = 0.956.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a score on a scale, and the registry identifies the measure as a mean with standard deviation. A mean describes the central value of a group’s scores, while a standard deviation describes how spread out individual scores are around that mean. In general, a group mean alone does not show the range of individual responses. The endpoint is change from baseline, so the reported means summarize changes rather than the Day 95 score levels themselves.

What does the treatment difference represent?

The sponsor reports a least-squares mean treatment difference. In general, a treatment difference expresses the estimated contrast between groups, with the sign indicating which group has the higher estimated change according to the stated comparison. A least-squares mean is a model-based estimate; it is not necessarily identical to subtracting the two raw group means. The registry record does not report the model covariates or other detailed specifications needed to reproduce that estimate.

The reported sponsor estimate is 1.386, with a 95% confidence interval from -4.804 to 7.577. In general, a confidence interval represents uncertainty around an estimate under the assumptions of the analysis. It gives a range of values compatible with the data and method, rather than a guarantee that the true value lies inside it. Because this interval includes zero, the comparison does not rule out no average difference. Its span also allows estimates in both directions.

What does MMRM mean?

The sponsor analysis is labeled MMRM, commonly used as an abbreviation for a mixed model for repeated measures. In general, repeated-measures models analyze outcome observations collected over time while accounting for the fact that measurements from the same participant are related. They can estimate group contrasts using the observations included in the model, subject to the model’s assumptions. The registry record does not report the specific repeated-measures structure, covariance assumptions, included visits, or handling of missing observations, so those details cannot be inferred from the method label alone.

The sponsor describes the reported parameter as a least-squares mean treatment difference. The registry entries identify the endpoint timeframe as baseline and Day 95, but do not give the model’s full specification. Accordingly, the reported value should be presented as the sponsor’s model-based result, not as a value that can be independently reconstructed from the arm summaries.

How does the unadjusted re-calculation differ?

The re-calculation uses the posted means, standard deviations, and analysis group sizes to estimate an unadjusted difference between groups. It gives a mean difference of 0.19, a confidence interval from -6.56 to 6.94, and p = 0.956. This is a summary-statistic comparison; it does not incorporate a model adjustment. It is useful as a transparent check on the magnitude and uncertainty implied by the posted summaries, but it does not replace the sponsor’s primary analysis.

In general, standard deviations describe variation among participant scores, whereas standard errors describe uncertainty in an estimated mean. The posted arm standard deviations are 22.43 and 21.94. The re-calculation uses a pooled standard deviation of 22.19. These quantities have different interpretations: the standard deviation concerns spread among observations, while the standard error concerns precision of an estimate. They should not be read as interchangeable measures.

What does a p-value say?

In general, a p-value measures how unusual data at least as inconsistent with a specified null hypothesis would be if that hypothesis and the test assumptions were true. It is not the probability that the null hypothesis is true, nor is it a measure of treatment importance. The sponsor reported p = 0.659; the unadjusted re-calculation gives p = 0.956. Both values are consistent with the confidence intervals including zero. Neither value demonstrates equivalence or proves that there is no difference.

The trial has multiple listed outcome measures, including 30 posted secondary outcomes. In general, testing multiple outcomes raises a multiplicity issue because the chance of at least one apparently noteworthy result can increase as more comparisons are made. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting secondary outcomes. The primary comparison can be described on its own, but any broader interpretation across outcomes would require those additional details.

Finally, all interpretations depend on the population included in the analysis and the treatment of missing observations. The registry describes the primary population as modified intent-to-treat, requiring at least 1 injection and valid baseline and post-injection scores. The record does not report the number excluded by this definition or detailed missing-data procedures. General statistical methods cannot fill in those trial-specific details.

10. Confidence Intervals and Precision

The sponsor’s 95% confidence interval for the primary treatment difference is -4.804 to 7.577 around an estimate of 1.386. The interval spans zero, so the data are compatible with no average difference under the reported analysis. They are also compatible with values on either side of zero within the stated limits. That uncertainty is important: the point estimate alone is not a complete description of the result.

In general, a confidence interval conveys precision by showing the range of estimates compatible with the data and analysis assumptions. A narrower interval indicates greater precision than a wider interval on the same scale, while an interval that crosses zero does not establish the absence of an effect. The interval does not give the probability that any one value is correct, and its interpretation depends on the model and analysis population.

The unadjusted re-calculation gives an estimate of 0.19 with a confidence interval from -6.56 to 6.94. This interval also crosses zero and includes values in both directions. It is a separate calculation based on summary statistics, not the sponsor’s MMRM estimate. Differences in the estimates and intervals reflect the fact that the calculations use different approaches; the registry does not provide enough model detail to attribute the difference to any one feature.

The endpoint scale runs from 0 to 100, but the registry record does not report a clinically important difference threshold. Thus, statistical precision can be described from the estimates and intervals, while the importance of particular score differences cannot be settled from these registry entries alone. Both intervals leave substantial uncertainty about the size and direction of the underlying average difference.

11. Planning a Similar Trial: Sample Size

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

The planning figures are participants per arm, not total events. They are illustrative calculations based on the observed effect, not a record of the trial’s original design justification. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set as the stated fraction of the observed effect.

The column for 80% power gives the planned number of participants per arm for a design that would have an 80% probability of detecting the specified effect under the stated assumptions, if that effect and the assumptions were correct. The 90% power column describes the corresponding planning target for a 90% probability. These are prospective design illustrations, not probabilities that this completed trial detected an effect.

For planning around 100% of the observed effect, the figures are 214,140 participants per arm for 80% power and 286,673 participants per arm for 90% power. For 75% of the observed effect, they are 380,694 and 509,641 participants per arm, respectively. For 50% of the observed effect, they are 856,560 and 1,146,691 participants per arm, respectively.

The figures rise as the effect used for planning becomes smaller: a smaller difference is harder to distinguish from random variation, so more participants are needed under the same assumptions. The observed effect used for this illustration is very small, which is why these planning figures are very large. This is not observed power; it is a sample-size planning illustration and should not be interpreted as a retrospective assessment of the trial’s power. The registry record does not report the assumptions used in the trial’s own original design.

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 = 0.190; pooled standard deviation σ = 22.191.
2. n = 2 × (1.96 + 0.842)² × 22.191² / 0.190² = 214,139.9, rounded up to 214,140 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 <- 0.190000000000005                      # observed difference in means
sigma <- 22.1913400869271                     # 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(4496167)
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 = 214,140 (80% power), 286,673 (90% power)
 75% of observed effect: n per arm = 380,694 (80% power), 509,641 (90% power)
 50% of observed effect: n per arm = 856,560 (80% power), 1,146,691 (90% power)
power.t.test: 214140.8 
Simulated power at n = 214,140 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 reported primary analysis is an MMRM estimate of the least-squares mean treatment difference: 1.386, with a 95% confidence interval from -4.804 to 7.577 and p = 0.659. This is the trial’s reported model-based comparison and should be distinguished from the unadjusted re-calculation.

The re-calculation uses posted group summary statistics and gives a mean difference of 0.19, with a confidence interval from -6.56 to 6.94 and p = 0.956. It is a simple, unadjusted comparison rather than a reproduction of the sponsor’s MMRM. The registry record does not provide the model details needed to recreate the sponsor estimate from the summary data.

The estimates are not identical, but both are close to zero relative to the uncertainty represented by their intervals, and both intervals include zero. Since the discrepancy flag is false, there is no flagged disagreement requiring the sponsor analysis to take precedence under the stated reporting convention. The sponsor’s analysis remains the reported primary analysis; the re-calculation is best read as a complementary check based on posted summaries, not as a replacement result.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists 31 outcome measures and 30 posted secondary outcomes. Those secondary outcomes include changes in passive range of motion for forward flexion, internal rotation, external rotation, abduction, and shoulder extension in the affected shoulder, as well as comparisons of selected movements with the contralateral shoulder at baseline. The primary endpoint is the adapted ASES composite score at Day 95.

When a trial assesses multiple outcomes, multiplicity is a statistical consideration. In general, the more comparisons made, the greater the opportunity for one result to appear notable by chance. A prespecified hierarchy or adjustment procedure can help guide interpretation. A multiplicity adjustment, testing hierarchy, or other strategy for interpreting the outcomes is not reported in the registry record. The posted primary p-value should therefore be understood as the reported primary comparison, without assuming how secondary comparisons were controlled.

The primary analysis population is described as modified intent-to-treat. It includes randomized participants who received at least 1 injection, had a valid baseline adapted ASES composite score, and had at least 1 valid post-injection score. The registry does not report how many randomized participants did not meet these criteria, or how missing primary endpoint values were handled. The term modified intent-to-treat should not be taken to mean every randomized participant contributed to the endpoint analysis.

The reported method is MMRM, but the registry record does not specify the model’s covariance structure, included visits, covariates, assumptions about missingness, or sensitivity analyses. These details matter because model estimates depend on the analysis specification and the observations included. Their absence does not invalidate the posted result, but it limits independent assessment of the model and of the consequences of missing data. No interim-analysis plan or original statistical analysis plan is reported in the registry record.

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: fall (1.1%), depression (1.0%).

The serious adverse event figures are reported separately from the primary efficacy analysis. The registry lists 1 serious adverse event among 101 participants at risk in the EN3835 group and 1 among 94 at risk in the placebo group. It reports 0 deaths in each group. These are counts and denominators as recorded; they are not the primary endpoint analysis group sizes.

The record also lists other adverse events in 69 participants at risk in the EN3835 group and 22 in the placebo group. These figures should not be confused with the serious adverse event counts. The record’s top serious-event terms are Fall, 1.1%, and Depression, 1.0%; the registry entries do not specify which group each term corresponds to.

These descriptive figures do not establish whether an event was caused by study treatment, and they do not provide a complete account of event timing, duration, or severity. The registry record does not report a statistical comparison for these safety counts in the registry record. The figures can be read as the posted event counts and denominators, without inferring causality or making a treatment recommendation.

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

The primary endpoint results show a small estimated difference between groups relative to the uncertainty in the comparison. The sponsor’s MMRM estimate was 1.386, with a 95% confidence interval from -4.804 to 7.577 and p = 0.659. The unadjusted re-calculation was 0.19, with a confidence interval from -6.56 to 6.94 and p = 0.956. Both intervals include zero, so neither comparison rules out no average difference.

This does not establish that EN3835 and placebo are equivalent. Equivalence requires a suitable design and a prespecified margin, neither of which is reported in the registry record. Nor does a non-significant p-value prove that there is no difference. The data and analyses leave uncertainty about the direction and size of the average contrast.

The endpoint combines pain and function into one participant-reported score. A change in the composite may reflect either component or both, and the composite result does not identify which aspect accounts for the group means. The registry record does not report a threshold for a clinically important score difference, so it is not possible to classify a particular estimate as important or unimportant on that basis.

The result pertains to the endpoint, analysis population, and timeframe stated in the registry: the modified intent-to-treat population and change in adapted ASES composite score at Day 95. It should not be generalized to unreported time points, other measures, or populations not described by this analysis. The listed secondary outcomes may address additional aspects of shoulder movement, but their results are not included in the registry record for this interpretation.

Safety counts are also descriptive and do not establish causation. Taken together, the registry supports a careful statistical description of the reported primary comparison and adverse-event counts, but not a broader claim about treatment benefit, absence of effect, equivalence, or causality.

17. Why This Trial Matters Statistically

This trial illustrates why a point estimate should be read together with its interval and analysis method. The sponsor’s reported estimate and the unadjusted re-calculation are both close to zero, but they come from different analyses. The sponsor result is model-based, while the re-calculation uses summary statistics. Treating them as identical would obscure the role of the analysis method; treating the re-calculation as a replacement would misrepresent the reported primary analysis.

The confidence intervals are central to understanding the evidence. Each includes zero and also permits differences in either direction. That pattern does not prove equivalence or absence of an effect; it shows that the estimates are uncertain enough that a range of possibilities remains compatible with the data and analysis. A p-value alone would not convey this full range.

The result also demonstrates how a small observed difference can produce very large sample-size planning illustrations when calculations are based on that effect. Those figures are not retrospective power and are not the trial’s original design justification. They make clear why planning assumptions and the effect considered meaningful need to be stated in advance; the registry record does not report the trial’s original sample-size assumptions.

Finally, interpreting a patient-reported composite requires attention to its construction, analysis population, and missing-data procedures. The endpoint combines pain and function, while the population definition requires valid baseline and post-injection scores. The registry provides some of these details but not the full model or missing-data specifications. These limits shape what can be concluded from the posted result.

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 record gives the primary endpoint definition and the sponsor’s MMRM result, but not the full model specification. Covariates, repeated-measures structure, covariance assumptions, and details of missing-data handling are not reported in the registry record. This limits independent reconstruction and assessment of the sponsor analysis from the posted information.

The primary analysis population is modified intent-to-treat, with requirements for treatment receipt and valid baseline and post-injection scores. The record does not report how many participants were excluded from that endpoint population or why, nor does it describe sensitivity analyses for missing observations. The reported endpoint group sizes are 84.0 and 82.0, while safety denominators are different; these populations should not be conflated.

The registry lists 30 posted secondary outcomes but does not report a multiplicity strategy in the registry record. Their results are not available in the information used for this interpretation, so they cannot support conclusions about other shoulder measurements. The registry record also does not report a clinically important difference threshold for the adapted ASES composite.

Other trial details are also limited in the record summarized: the EN3835 arm is described as up to 1.74 mg, but the intervention description does not specify a schedule; and further details of allocation and masking procedures are not reported. The original sample-size justification and trial-specific planning assumptions are not reported in the registry record. These gaps limit assessment of design choices without changing the meaning of the posted primary estimate and its uncertainty.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Safety and Efficacy of EN3835 in Participants With Frozen Shoulder — NCT04496167.
• Registry record last updated 2023-09-13; status last verified 2023-08.
• Results first posted 2023-09-13.
• 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

NCT04496167 is a completed Phase 2 randomized, double-blind, placebo-controlled parallel-group study of EN3835 for adhesive capsulitis, or frozen shoulder. The registry lists 198 enrolled participants. The primary endpoint was change from baseline to Day 95 in the adapted ASES composite score for the affected shoulder.

The sponsor reported an MMRM least-squares mean treatment difference of 1.386, with a 95% confidence interval from -4.804 to 7.577 and p = 0.659. An unadjusted re-calculation from the posted summaries gave a mean difference of 0.19, with a confidence interval from -6.56 to 6.94 and p = 0.956. Both intervals include zero.

The record lists 30 posted secondary outcomes and reports serious adverse events in 1 participant in each group, with 0 deaths in either group. Details of the original sample-size rationale, model specification, and missing-data procedures are not reported in the registry record.