Study Evaluating the Safety and Efficacy of AR-15512: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Evaluating the Safety and Efficacy of AR-15512 (NCT05285644), a randomized phase 3 study. The primary outcome is Percentage of Subjects Who Achieved Equal to or Greater Than 10 Millimeter Increase From Pre-drop at Baseline to Post-drop on Day 14 in Study Eye Unanesthetized Schirmer Score.

2. Clinical Question

The primary outcome is a responder measure: each participant is classified according to whether the specified increase was achieved. The registry reports responder percentages for both groups and a sponsor analysis of the difference in percentages. This makes the central contrast an absolute difference in the proportion meeting the threshold, rather than a comparison of average Schirmer scores.

The primary test concerns the Day 14 post-drop assessment. Other recorded assessments include symptom measures and Schirmer measurements at different times, but they address related questions rather than replacing the stated primary endpoint. The findings therefore need to be read in the context of this particular threshold, eye, and assessment time.

3. Trial Design

COMET-2 was a Phase 3, randomized, double-masked, parallel-group, vehicle-controlled study of AR-15512 for dry eye disease. The registry record reports enrollment of 465 participants and lists Aerie Pharmaceuticals as the sponsor. The study status is completed. It began on 2022-05-09, and the primary completion date was 2023-07-24.

Participants were assigned to 0.003% AR-15512 or AR-15512 vehicle. In both groups, the regimen was one drop in each eye twice a day for 90 days. The record describes screening at Day -14, baseline at Day 1, and subsequent visits at Day 7, Day 14, Day 28, and Day 90, which was the study exit visit. Individual participation was described as approximately 15 weeks. The study summary describes conduct at approximately 20 sites in the United States.

For the primary outcome, the registry reports 230 participants in the AR-15512 group and 235 in the vehicle group. These are the group sizes shown for the primary analysis. The primary endpoint was the percentage achieving at least a 10-millimeter increase in study-eye unanesthetized Schirmer score from the Day 1 pre-drop measurement to the Day 14 post-drop measurement. The analysis population is described as the intention-to-treat population comprising all randomized participants, with multiple imputation used for the primary analysis.

The record does not report the trial’s original sample-size justification, its planned statistical assumptions, or a detailed account of any interim analyses. Those matters cannot be inferred from the enrollment or the posted outcome results. The sample-size illustrations elsewhere on this page are calculations based on the observed effect, not documentation of the study’s original design.

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: This reporting group includes all randomized participants.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population comprising all randomized subjects. This analysis uses multiple imputation methodology.

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

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Subjects Who Achieved Equal to or Greater Than 10 Millimeter Increase From Pre-drop at Baseline to Post-drop on Day 14 in Study Eye Unanesthetized Schirmer Score, measured over Baseline (Day 1) pre-drop; Day 14 post-drop.

The primary endpoint is the percentage of participants who achieved an increase of at least 10 millimeters in the study-eye unanesthetized Schirmer score between the Day 1 pre-drop measurement and the Day 14 post-drop measurement. The Schirmer test uses filter paper placed on the lower eyelid to record tear production. The registry describes a scale from 0 millimeters, indicating no tear production, to 35 millimeters, the maximum. An increased score is identified in the record as a positive outcome.

This endpoint turns a measured change into a binary responder classification. A participant either reached the specified increase or did not. The result is then summarized as a percentage in each group, making it possible to compare the proportion meeting the threshold in the AR-15512 and vehicle groups. This is different from comparing average score changes: a group’s responder percentage does not show how far above or below the threshold each participant’s measurement fell.

A threshold-based outcome can be useful when the trial question is framed around the proportion of participants reaching a specified level of change. Its interpretation is specific to the chosen threshold, measurement procedure, eye, and time point. The endpoint alone does not describe all aspects of tear production, symptoms, or longer-term outcomes. The registry record identifies one study eye as contributing data to this analysis and states that the primary analysis used the intention-to-treat population and multiple imputation.

7. Results Posted in the Registry

• 0.003% AR-15512: 42.6% (98/230) versus 8.1% (19/235) on control; risk difference 34.5 percentage points (95% CI 27.2 to 41.8), risk ratio 5.27 (3.34–8.32); unadjusted two-proportion z-test p = <0.001.

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.

Least Squares Mean Change From Baseline in Global Symptom Assessment iN Dry Eye (SANDE) Score on Day 28

• 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 errors.
• Result:
  • 0.003% AR-15512: mean -19.70 versus -14.70 on control; difference -5.00 (95% CI -9.42 to -0.58), standardized effect d = -0.21; z-test p = 0.027.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for 0.003% AR-15512 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Least Squares Mean Change From Pre-drop Baseline in Unanesthetized Schirmer Score on Post-drop Day 14 (Study Eye)

• 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 errors.
• Result:
  • 0.003% AR-15512: mean 9.80 versus 2.60 on control; difference 7.20 (95% CI 5.69 to 8.71), standardized effect d = 0.87; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for 0.003% AR-15512 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Subjects Who Achieved Equal to or Greater Than 10 Millimeter Increase From Pre-drop at Baseline to Post-drop on Day 1 in Study Eye Unanesthetized Schirmer Score

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • 0.003% AR-15512: 36.1% (83/230) versus 9.4% (22/235) on control; risk difference 26.7 percentage points (95% CI 19.5 to 34.0), risk ratio 3.85 (2.50–5.95); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for 0.003% AR-15512 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary results cover both symptom assessment and Schirmer measurements. Across the displayed results, the between-group contrasts generally point in the same direction as the primary comparison: the AR-15512 group has a more favorable reported change or a higher responder proportion than the vehicle group. This offers supporting context across related measures, but the secondary endpoints are not interchangeable with the primary Day 14 responder outcome.

Symptom scores and tear-production measurements capture different aspects of the study question. A pattern across them can help readers consider whether the primary finding sits alongside other measured outcomes, but it does not show that every endpoint changed to the same extent or that the endpoints measure a single underlying construct. Each has its own measurement scale, time point, and analysis.

The registry record lists secondary analyses and reports inferential results for some of them. Their p-values should be treated with more caution than the primary result because the registry does not report a multiplicity-adjustment strategy. When multiple outcomes and assessment times are examined, the chance of at least one apparently notable result can increase if each is assessed separately without adjustment. This is a general concern about multiple testing, not evidence that any particular posted finding is false.

Secondary results are therefore best read as supportive or exploratory context unless a prespecified testing hierarchy or other multiplicity procedure is documented. Such a procedure is not reported in the registry record. The secondary pattern does not replace the primary endpoint, and it should not be used to claim that all symptoms or measures improved, that an effect persisted beyond the recorded assessments, or that the result applies to outcomes not measured in the trial.

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

• Least Squares Mean Change From Baseline in Global Symptom Assessment iN Dry Eye (SANDE) Score on Day 28
• Least Squares Mean Change From Pre-drop Baseline in Unanesthetized Schirmer Score on Post-drop Day 14 (Study Eye)
• Percentage of Subjects Who Achieved Equal to or Greater Than 10 Millimeter Increase From Pre-drop at Baseline to Post-drop on Day 1 in Study Eye Unanesthetized Schirmer Score
• Least Squares Mean Change From Pre-drop Baseline in Unanesthetized Schirmer Score on Post-drop Day 1 (Study Eye)
• Percentage of Subjects Who Achieved Equal to or Greater Than 10 Millimeter Increase From Pre-drop at Baseline to Post-drop on Day 90 in Study Eye Unanesthetized Schirmer Score
• Least Squares Mean Change From Pre-drop Baseline in Unanesthetized Schirmer Score on Post-drop Day 90 (Study Eye)
• Least Squares Mean Change From Baseline in Global SANDE Score on Day 90
• Least Squares Mean Change From Baseline in SANDE Frequency Score on Day 90

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary: participants are counted as responders if their study-eye Schirmer score increased by at least 10 millimeters from the Day 1 pre-drop measurement to the Day 14 post-drop measurement. The group summaries are percentages. For this kind of outcome, a difference in percentages describes the absolute separation between the groups in the proportion meeting the criterion. It is not the same quantity as a relative risk or an odds ratio, which express a comparison on different scales.

How to read the reported test

The sponsor reports a Pearson’s Chi-squared test for the primary comparison, with the parameter described as difference in percentages. In general, a chi-squared test assesses whether observed categorical counts are compatible with no association between group and outcome under the test’s assumptions. Here, the categories are responder and nonresponder, and the groups are AR-15512 and vehicle. The test result is reported alongside the estimated difference and its confidence interval, so the size and precision of the contrast can be considered as well as the evidence against a no-difference hypothesis.

A p-value summarizes how unusual results at least this incompatible with the null hypothesis would be under that hypothesis and the test assumptions. It does not give the probability that the null hypothesis is correct. It also does not indicate whether a result is important to participants or whether it applies beyond the study setting. The primary p-value in the sponsor analysis is reported as p < 0.001.

What the estimate means

The reported difference is an absolute contrast in percentages. It directly describes how much the responder proportion differed between the randomized groups according to the sponsor analysis. The registry also presents a recalculated unadjusted risk difference from posted summary statistics. Such a recalculation can help readers see whether a simple comparison is in the same general range as the reported result, but it does not reproduce all possible features of the trial analysis.

For a binary outcome, the observed percentage depends on both the number of responders and the number of participants included in the analysis. The record describes the primary population as all randomized participants and says multiple imputation was used. It does not report sufficient detail in the registry record to reconstruct the imputation procedure, its assumptions, or the sensitivity analyses used to evaluate missing outcomes.

Why the confidence interval matters

A confidence interval expresses uncertainty around an estimated contrast. In general, a narrower interval indicates greater precision, while a wider interval indicates more uncertainty about the magnitude. The sponsor’s 95% interval for the primary difference is reported from 26.9 to 42.0. Under the analysis assumptions, values within this interval are compatible with the observed data at the stated confidence level. The interval is not a guarantee that the true effect lies within its limits, and it does not account for every possible source of bias or model uncertainty.

The interval and p-value are related but serve different descriptive purposes. The p-value summarizes evidence against a specified null value; the interval shows a range of plausible values for the estimated contrast. Neither should be interpreted without the endpoint definition and trial design. An interval for a responder difference does not provide an interval for an average Schirmer change or for a symptom score.

Reported and recalculated analyses

The registry lists the sponsor’s primary method as Pearson’s Chi-squared test and reports its difference in percentages, 95% confidence interval, and p-value. The posted recalculation is a simple, unadjusted calculation from summary statistics. It is not a replacement for the sponsor’s own analysis. For continuous secondary outcomes, the registry lists ANCOVA as the sponsor method; the additional recalculations are based on posted group summaries and are not adjusted analyses.

These distinctions matter because analyses may differ in their treatment of baseline measurements, missing observations, and other model terms. The registry record does not provide a full statistical analysis plan or all model specifications. Readers should therefore identify which result comes from the reported trial analysis and which is a separate, simpler recalculation. For the primary endpoint, the sponsor’s reported result is the principal inferential result.

What is not specified

The registry record does not report the original sample-size justification, the full prespecified statistical plan, or whether interim analyses were conducted. It also does not describe a multiplicity-control strategy for the secondary outcomes. Those details should not be inferred from the posted p-values, the observed enrollment, or the sample-size illustrations. General statistical interpretation can explain what a method does, but it cannot fill in trial-specific procedures that are not reported.

Randomization and double masking are design features intended to support a fair comparison and reduce opportunities for knowledge of assignment to influence assessment. Their presence does not eliminate every possible source of uncertainty. The result remains conditional on the study population, outcome measurement, analysis population, handling of missing data, and execution of the trial. The registry summary provides some of this information, but not every detail necessary for a full independent audit.

11. Confidence Intervals and Precision

The sponsor reports the primary difference in percentages as 34.4, with a 95% confidence interval from 26.9 to 42.0. The interval is entirely above zero, consistent with the reported p-value of p < 0.001 for the comparison. Its width conveys uncertainty about the size of the between-group contrast: the estimate is not known with exactness, even though the result is statistically distinguishable from no difference under the reported analysis.

In general, a confidence interval gives a range of values that are reasonably compatible with the data under the model and its assumptions. A 95% interval does not mean there is a 95% probability that this particular interval contains the true value. It also does not cover uncertainty from every possible source, such as departures from model assumptions or limitations in how the trial was conducted.

The interval applies to the difference in the percentage meeting the stated responder threshold. It is not an interval for an average Schirmer score change, an individual participant’s response, or any symptom outcome. Its interpretation is bounded by the endpoint’s definition: study eye, at least 10 millimeters of increase, and the Day 14 post-drop assessment.

The registry also displays an unadjusted recalculation based on the posted group summaries. Its estimate and interval are on a different numerical scale from the sponsor’s reported percentage-point presentation, so they should be read as the recalculation is labeled rather than substituted for the sponsor’s result. The sponsor’s reported analysis remains the trial analysis reported in the registry. Neither the interval nor the p-value establishes clinical importance; those are separate judgments not determined by statistical precision alone.

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 table and curve are planning illustrations based on the observed effect, not a reconstruction of COMET-2’s original design. For this binary endpoint, the listed values are participants per arm. They are not total events. The assumptions used for these illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each row uses the stated fraction of the observed effect as the effect to detect.

The 80% power column gives the estimated participants per arm for a hypothetical study to have an 80% chance of detecting the specified effect at the stated significance level, if the planning assumptions and effect are correct. The 90% power column gives the corresponding planning size for a 90% chance. Power is a design property under assumed conditions; it is not a measure of the probability that a particular trial result is true.

Rows that plan for a smaller fraction of the observed effect require more participants per arm in the illustrations. This is because a smaller difference is harder to distinguish from random variation when other assumptions are held constant. The table therefore shows how sample-size requirements change when the assumed effect is more conservative than the observed result. It should not be read as evidence that the observed effect will recur in another study.

These numbers are not observed power. They are hypothetical planning estimates derived from the observed effect and the stated assumptions. The trial’s own original design assumptions and sample-size justification are not reported in the registry record. The illustration cannot tell readers what assumptions the investigators used when planning COMET-2, nor can it establish whether the original study was adequately powered for every endpoint.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.081; treatment proportion p₁ = 0.426; average p̄ = 0.253.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6152 = 1.2057.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5647 = 0.4752.
4. n = (1.2057 + 0.4752)² / (0.3452)² = 23.7, rounded up to 24 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.0808510638297872                          # control proportion
p1 <- 0.426086956521739                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

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

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(5285644)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 24 (80% power), 32 (90% power)
 75% of observed effect: n per arm = 38 (80% power), 50 (90% power)
 50% of observed effect: n per arm = 73 (80% power), 97 (90% power)
power.prop.test: 23.7 
Simulated power at n = 24 per arm: 0.838 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis uses Pearson’s Chi-squared test and gives a difference in percentages of 34.4, a 95% confidence interval from 26.9 to 42.0, and p < 0.001. The registry also displays a simple unadjusted recalculation from the posted group summaries. The recalculated estimate and interval are broadly consistent with the sponsor’s reported contrast, while reflecting the calculation method used for the reanalysis.

The sponsor result should be distinguished from the recalculation. The sponsor’s figure is the reported trial analysis; the recalculation is a separate, simple comparison and does not incorporate any analysis features not represented in the posted summaries. The registry identifies the primary population as the intention-to-treat population and states that multiple imputation was used, but it does not provide enough detail to reproduce that procedure or all aspects of the sponsor’s analysis.

The record does not flag a discrepancy between the sponsor result and the recalculation. Even so, agreement in broad direction or scale does not make the methods identical. The sponsor’s reported analysis is the appropriate reference for what the trial reported, while the recalculation offers a transparent supplementary check based on summary statistics. Neither result alone addresses clinical importance, missing-data assumptions beyond those reported, or the possible impact of analyses not documented in the registry.

15. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is clearly identified in the registry record, and the primary analysis population is described as all randomized participants. The record also states that multiple imputation was used for the primary analysis. These details help define the reported result, but do not constitute a complete statistical analysis plan. The full plan, including the planned handling of all endpoints and any formal testing sequence, is not reported in the registry record.

Multiplicity arises when investigators assess several outcomes, time points, or contrasts. In general, if each is tested separately at a conventional significance threshold, the chance of at least one apparently strong result can rise across the collection of tests. Statistical approaches to address this include adjustment procedures or a prespecified hierarchy, but the registry does not report whether such an approach was used for these secondary analyses. The posted secondary p-values should therefore not be treated as though multiplicity control is established.

Missing data are another consideration. The primary analysis is described as using multiple imputation, a general approach that fills in missing outcome values using a statistical model and then combines results across completed datasets. The registry record does not report the imputation model, the assumptions about why data were missing, the amount or pattern of missing primary outcome data, or sensitivity analyses. Those omissions limit an independent assessment of how robust the primary result is to alternative missing-data assumptions.

The population is described as the intention-to-treat population comprising all randomized participants. That description is useful, but the registry does not report every detail of how participants with missing measurements contributed to each analysis. It also does not report the original sample-size justification, the complete prespecified analysis plan, or whether interim analyses were conducted. These details cannot be inferred from the number enrolled or from the reported endpoint results.

For these reasons, the primary result can be interpreted using the methods explicitly reported, while acknowledging that important analysis details are absent from the registry record. Secondary findings are more tentative when the handling of multiplicity is not documented. Neither missing documentation nor an unreported procedure should be mistaken for evidence that a procedure was or was not used.

16. 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: cartilage injury (0.4%), coronary artery occlusion (0.4%), constipation (0.4%), haemorrhoidal haemorrhage (0.4%), volvulus (0.4%).

The registry separates serious adverse events by treatment group and by ocular or nonocular classification. In the 0.003% AR-15512 ocular group, 0 serious adverse events are reported among 230 participants at risk; the nonocular group reports 1 among 230. In the vehicle ocular group, 0 are reported among 235 participants at risk; the nonocular group reports 6 among 235. The record also lists 1 serious adverse event in the pretreatment group, where 465 participants were at risk. Deaths are reported as 0 in each listed group.

These are event counts, not conclusions about cause. The registry figures do not establish that an event was caused by study treatment, and the group summaries do not show the full clinical circumstances or timing for each event. The serious-event terms listed in the record should be read as reported events, not as evidence of a treatment relationship. The safety information is descriptive and should be interpreted alongside the way the registry grouped and counted events.

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

The primary result describes a between-group difference in the percentage of participants who met a specific Day 14 Schirmer responder criterion. It does not describe every participant’s score, the average change in tear production, or the entire course of dry eye disease. It also does not directly measure every symptom that may matter to participants. Other posted outcomes add context, but each answers a distinct question.

A statistically significant result is not, by itself, a judgment about clinical importance. The p-value and confidence interval describe statistical evidence and precision under the reported analysis; they do not establish whether the size of the difference is meaningful to patients, clinicians, or decision-makers. The registry record does not provide a threshold for clinical importance for this endpoint.

The result applies to the participants, regimen, comparator, endpoint, and assessment time studied in COMET-2. It should not be extended automatically to other populations, dosing schedules, comparators, or durations. The study’s primary completion date and listed visit schedule define the period represented by the posted results; the record does not establish persistence beyond the measured assessments.

The result also does not prove that the treatment caused every observed group difference in all circumstances. Randomization and double masking support a controlled comparison, but interpretation still depends on how the study was conducted, how missing data were handled, and whether the analysis assumptions were appropriate. Some of those details are not reported in the registry record.

In short, the posted data show a higher reported proportion meeting the specified responder threshold in the AR-15512 group than in the vehicle group, with a sponsor-reported interval above zero. That finding is specific and statistically clear for the primary analysis, but it is not a complete account of symptoms, individual experience, safety, clinical importance, or outcomes beyond those recorded in the trial.

18. Why This Trial Matters Statistically

This trial illustrates why endpoint definition matters in clinical research. A responder endpoint translates a continuous measurement into a threshold-based classification. That can make the result easy to describe as a proportion, but it also discards information about how much each participant’s measurement changed and how close nonresponders came to the threshold. The primary contrast should therefore be read as a difference in responder proportions, not as a full description of the underlying score distribution.

The comparison also demonstrates the value of reporting both an estimate and a confidence interval. A p-value alone would not communicate the range of plausible magnitudes associated with the primary contrast. The reported interval helps readers assess precision, while the endpoint definition and study design determine what the interval refers to. Statistical significance does not substitute for an assessment of clinical meaning.

COMET-2 also provides an example of the distinction between a sponsor’s reported analysis and a simple recalculation from group summaries. Recalculations can be useful for transparency and orientation, but they may not reproduce the trial’s analysis choices. Clear labeling helps keep the sponsor’s inferential result distinct from supplementary calculations.

Finally, the record highlights the importance of documenting analysis decisions. The primary population and use of multiple imputation are reported, but the registry does not report the full imputation details, multiplicity strategy, original sample-size justification, or whether interim analyses were conducted. These omissions constrain how completely readers can evaluate robustness. Statistical interpretation depends not only on the numerical result, but also on knowing how the result was planned, generated, and analyzed.

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

Several trial-specific limitations follow from what is and is not reported in the registry record. The primary outcome is a threshold-based responder measure at a single post-drop assessment. It does not show the distribution of individual Schirmer changes, and the registry’s summary does not provide a full account of how close nonresponders were to the threshold. The analysis uses one study eye, so the primary result is not a direct comparison of outcomes from both eyes.

The primary analysis is described as intention-to-treat and as using multiple imputation, but the imputation model, missing-data patterns, assumptions, and sensitivity analyses are not reported. Without these details, readers cannot independently assess how the handling of missing measurements might influence the result. The original sample-size justification and detailed statistical analysis plan are also not reported.

The registry does not report whether multiplicity was addressed for the secondary outcomes, or whether interim analyses were conducted. This limits the strength of conclusions that can be drawn from secondary p-values and prevents a full reconstruction of the testing strategy. The posted secondary findings should not be assumed to have the same confirmatory status as the primary endpoint.

Safety summaries are presented as event counts grouped by ocular and nonocular classification. The record’s summary does not establish event causation, and counts alone do not describe severity, duration, timing, or the full clinical context. They should not be interpreted as proof of a treatment relationship or as a complete safety assessment.

Finally, the interpretation is bounded by the enrolled dry eye disease population, the specified vehicle comparator, the twice-daily 90-day regimen, and the recorded assessment schedule. The registry summary does not establish how findings would generalize beyond those study conditions. These limitations do not erase the reported primary comparison, but they define the scope and certainty of what can be concluded from the posted information.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study Evaluating the Safety and Efficacy of AR-15512 — NCT05285644.
• Registry record last updated 2025-07-23; status last verified 2025-07.
• Results first posted 2025-07-23.
• 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

COMET-2 was a completed Phase 3, randomized, double-masked, parallel-group, vehicle-controlled study of 0.003% AR-15512 in participants with dry eye disease. The registry reports enrollment of 465 participants. Both groups used one drop in each eye twice a day for 90 days, and the study summary describes visits through Day 90.

The primary endpoint was the percentage achieving at least a 10-millimeter increase in study-eye unanesthetized Schirmer score from the Day 1 pre-drop assessment to Day 14 post-drop. The sponsor reported a difference in percentages of 34.4, with a 95% confidence interval from 26.9 to 42.0 and p < 0.001. The analysis population was the intention-to-treat population, with multiple imputation stated for the primary analysis.

Secondary results provide additional symptom and Schirmer context, while the registry does not report a multiplicity strategy or the full statistical plan. The sample-size illustrations are based on the observed effect and stated planning assumptions, not the trial’s original design. Safety event counts are descriptive and do not establish causation.