A Safety Study of of AG-920 Sterile Topical Ophthalmic Solution: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Safety Study of of AG-920 Sterile Topical Ophthalmic Solution (NCT05501860), a randomized phase 3 study. The primary outcome is Number of Participants With Treatment Emergent Adverse Events (TEAEs).

2. Clinical Question

The central question is whether the number of participants with treatment-emergent adverse events differed between healthy subjects assigned to a single topical administration of AG-920 and those assigned to placebo. The primary endpoint counts participants with at least one treatment-emergent adverse event, rather than counting every event experienced by each participant. The registry record reports 40 participants with events in the AG-920 group and 13 in the placebo group, alongside group sizes of 166 and 83.

This is a safety question, not a test of clinical benefit. The secondary outcome listed in the registry record concerns mean change from baseline in corneal endothelial cell density. A numerical result for that outcome is not reported in the registry record.

3. Trial Design

The registry record describes a completed Phase 3, randomized, placebo-controlled, parallel-group study of AG-920 in healthy subjects. It reports enrollment of 249, a start date of 2022-07-30, and a primary completion date of 2022-12-07. The two arms were AG-920 and placebo, with posted group counts of 166 and 83, respectively. The intervention is described as a single topical ocular administration.

The official title calls the study double-masked, while the registry design fields classify masking as triple. The brief summary also describes the study as double-masked. The record does not explain this difference in wording, so the exact masking roles should not be inferred from those labels alone. Randomized allocation and a parallel-group model are recorded. Details of the randomization sequence, allocation concealment, and which groups of people were masked are not reported in the registry record.

The study’s stated purpose was to evaluate ocular safety compared with placebo. The primary outcome is the number of participants with treatment-emergent adverse events. The timeframe is from Day 1, the treatment day, through 4 days following treatment; participants in specular microscopy follow-up are covered through 90 days following treatment. The registry also lists mean change from baseline in corneal endothelial cell density as a secondary outcome.

The record identifies the sponsor as American Genomics, LLC and the condition as local anesthesia. Although the trial is marked completed, that status does not establish that every planned analysis or outcome is posted. The original sample-size rationale, the trial’s own statistical analysis plan, and any protocol amendments are not reported in the registry record.

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

Analysis population for the primary outcome, as defined in the registry: The safety population is defined as all subjects which were randomized and who received at least one drop of the dose (2 drops) of the study medication. The ITT population is defined as all subjects who were randomized to treatment and have received at least one dose (two drops) of the study medication. In this study, the ITT and safety populations were found to be identical.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Treatment Emergent Adverse Events (TEAEs), measured over From Day 1 (treatment day) to 4 days following treatment day; 90 days following treatment for subjects participating in specular microscopy..

The primary endpoint is the number of participants with treatment-emergent adverse events, abbreviated as TEAEs in the registry record. It counts participants who experienced an event, not the total number of event episodes. A participant with more than one event would still contribute once to a count of participants with an event. The record says TEAEs were to be summarized by treatment group using frequency and percent for each system organ class and preferred term within each class. A breakdown by those categories is not reported in the registry record.

The specified timeframe begins on Day 1, the treatment day, and continues through 4 days following treatment. For subjects participating in specular microscopy, the timeframe extends to 90 days following treatment. The registry record does not state how many subjects took part in that follow-up, or whether all events were assessed using identical visit schedules.

Counting participants with adverse events is a straightforward way to describe how many people had at least one recorded event during the defined observation period. It is useful in a safety-focused study because the participant, rather than each event entry, is the unit of the endpoint. The count alone does not describe event severity, duration, causal attribution, or clinical importance. Those features require additional information, and a detailed breakdown is not reported in the registry record.

The stated safety population includes randomized subjects who received at least one drop of the study medication, specified as a dose of 2 drops. The endpoint should therefore be interpreted in the context of the population and follow-up described by the registry, rather than as a complete account of every possible safety outcome.

7. Results Posted in the Registry

• AG-920: 24.1% (40/166) versus 15.7% (13/83) on control; risk difference 8.4 percentage points (95% CI -1.7 to 18.6), risk ratio 1.54 (0.87–2.71); unadjusted two-proportion z-test p = 0.125.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each person is classified according to whether at least one treatment-emergent adverse event was recorded during the specified follow-up. The posted group summaries give the number of participants with an event and the number of participants in each group. This makes the outcome different from a count of all event episodes. General statistical interpretation must keep that distinction clear because participants can experience multiple events, while the endpoint counts each participant only once.

What does a risk difference describe?

A risk difference compares the proportion with an event in one group with the proportion in the other group by subtraction. It is expressed on an absolute scale: it describes the separation in event proportions rather than their relative size. The registry comparison reports a positive risk difference for AG-920 versus placebo. Its confidence interval includes zero, meaning that the range of values compatible with the data includes no absolute difference as well as values on either side of no difference. The posted estimate and interval are unadjusted, simple comparisons based on the group summaries.

The registry also presents a risk ratio and an odds ratio. A risk ratio compares the event proportions as a ratio; a value of one represents equal proportions. An odds ratio compares event odds, rather than proportions, and one likewise represents equal odds. Both posted point estimates are greater than one, and both confidence intervals include one. These relative measures offer another scale for the comparison, but they do not remove the uncertainty evident in the intervals. Odds and risks are distinct quantities, so an odds ratio should not be described as though it were a risk ratio.

How should the p-value be read?

The registry reports an unadjusted p-value of 0.125 for the comparison. In general, a p-value measures how unusual the observed data, or data more extreme under the test, would be if the null hypothesis and the test’s assumptions held. It is not the probability that the treatments are equal, and it does not quantify the size or importance of an effect. The registry does not report the full testing procedure, so a more detailed account of its assumptions cannot be reconstructed from the record. The label unadjusted indicates that the posted p-value is not presented as adjusted for other variables or for multiple comparisons.

Confidence intervals and p-values answer related but different questions. A confidence interval communicates a range of estimates compatible with the data under the method used, while the p-value summarizes evidence against a specified null under a test. For this comparison, the risk-difference interval spans zero, and the relative-measure intervals span one. Those features align with the unadjusted p-value in indicating uncertainty about the direction and size of a group difference. None of these summaries by itself establishes causation or clinical importance.

Why does the endpoint use participant counts?

Using the number of participants with at least one event gives each participant a single contribution to the primary binary outcome. This avoids allowing one person with multiple events to contribute repeatedly to that particular count. It also means the endpoint does not capture the total burden of events. Counts by event type, severity, duration, and recurrence would be needed to describe those dimensions, and such details are not reported in the registry record.

General statistical methods for event counts sometimes require attention to over-dispersion, which occurs when variability in counts exceeds what a simple count model expects. That concern is relevant to analyses of event totals, but the primary endpoint posted is the number of participants with any event, not a repeated-event total. The record does not report a model for event counts or a method for handling recurrent events. No specific over-dispersion adjustment can therefore be attributed to this trial.

The outcome description states that TEAEs were to be summarized by treatment group and by system organ class and preferred term. The registry record posts aggregate participant counts for the primary endpoint, but does not report those detailed category-level results. It also does not report a numerical secondary-outcome analysis for mean change from baseline in corneal endothelial cell density. As a result, interpretation of the posted primary comparison should remain focused on the aggregate participant-level endpoint rather than implying detail that is not recorded.

Population definitions matter because estimates depend on who is included. The safety population is defined as randomized subjects who received at least one drop of the study medication, described as a dose of 2 drops. The record also gives an ITT definition based on randomized subjects who received at least one dose, described as two drops. These descriptions do not establish how missing outcome information was handled, or whether any randomized participants were excluded from the posted primary comparison. Those analysis details are not reported in the registry record.

10. Confidence Intervals and Precision

The reported confidence intervals express uncertainty around the estimated group comparisons. For the risk difference, the interval extends from a negative value to a positive value, so it includes zero. For the relative measures, the confidence intervals include one, the value that represents no relative difference. These intervals are broad enough to include more than one plausible interpretation of the direction and size of the comparison.

Precision is influenced by the amount of information in the data, including the number of participants and the number of observed outcomes. The posted groups include 166 participants assigned to AG-920 and 83 assigned to placebo. The registry record does not report a precision target or a sample-size justification for the original design. The unequal group sizes are part of the observed study record; the reasons for that allocation are not reported.

A confidence interval is not a range that guarantees the true value lies inside it. In general, it is a range produced by a method that has a stated long-run coverage property when its assumptions hold. It describes statistical uncertainty, not uncertainty from every possible source, such as measurement limitations, missing outcomes, or incomplete reporting. The registry record does not provide enough detail to assess all such sources.

The interval crossing the null value should not be read as proof that the groups are identical. Rather, the posted data do not precisely distinguish no difference from a range of possible differences under the reported unadjusted comparison. Conversely, a point estimate in one direction does not by itself establish that the difference is reliable or clinically important. Clinical interpretation would also require event details and context that are not reported in the registry record.

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 illustrations based on the observed effect, rather than the trial’s original design calculation. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario targets the stated fraction of the observed effect. The original trial’s own sample-size assumptions and justification are not reported in the registry record.

For planning around 100% of the observed effect, the illustration gives 351 participants per arm for 80% power and 469 participants per arm for 90% power. For planning around 75% of the observed effect, it gives 599 participants per arm for 80% power and 801 participants per arm for 90% power. For planning around 50% of the observed effect, it gives 1,289 participants per arm for 80% power and 1,725 participants per arm for 90% power.

In plain language, the 80% power column describes the illustrated number of participants per group for a future study designed to have an 80% chance of detecting the target effect under the stated assumptions, if that effect and those assumptions hold. The 90% power column describes the corresponding planning target for a higher chance of detection. These are planning illustrations, not observed power for the completed trial. Smaller target effects require more participants in these scenarios because a subtler difference is harder to distinguish from sampling variation.

The calculations use equal allocation even though the posted arm counts are not equal. They should therefore be read as standardized illustrations, not as a reconstruction of the trial’s enrollment plan. They also depend on the observed effect used as the reference, which is uncertain. A planning calculation cannot establish that a future study will achieve its target if the actual effect, outcome frequency, or analysis assumptions differ.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.157; treatment proportion p₁ = 0.241; average p̄ = 0.199.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5644 = 1.1062.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5612 = 0.4724.
4. n = (1.1062 + 0.4724)² / (0.0843)² = 350.3, rounded up to 351 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 <- 0.156626506024096                          # control proportion
p1 <- 0.240963855421687                          # 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(5501860)
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 = 351 (80% power), 469 (90% power)
 75% of observed effect: n per arm = 599 (80% power), 801 (90% power)
 50% of observed effect: n per arm = 1,289 (80% power), 1,725 (90% power)
power.prop.test: 350.3 
Simulated power at n = 351 per arm: 0.800 (20000 replicates)

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

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome and one secondary outcome, mean change from baseline in corneal endothelial cell density after treatment with AG-920. A numerical result for that secondary outcome is not reported in the registry record. The number of outcome measures is recorded as 2, while the record reports no posted analyses. The primary group comparison is available as a registry summary, but a detailed statistical analysis plan is not reported in the registry record.

When several outcomes or several comparisons are tested, the chance of obtaining at least one apparently notable result can increase. General approaches to multiplicity include defining a primary outcome in advance, adjusting tests, or treating secondary analyses as exploratory. The registry does not report whether such a strategy was planned or used. No claim can therefore be made about multiplicity control for this trial. The unadjusted p-value should be understood as the posted unadjusted comparison, not as evidence that a family of tests was controlled.

Missing-data handling is also not reported in the registry record. The record does not state how participants with incomplete follow-up, unavailable event assessments, or missing secondary outcome measurements were treated in the analyses. It does not provide the number of participants with missing primary endpoint information or describe any imputation method. Without those details, the effect of missing observations on the posted comparison cannot be assessed.

The safety population is defined as randomized subjects who received at least one drop of the study medication, specified as 2 drops. The ITT population is described as randomized subjects who received at least one dose, also described as two drops. The record does not explain whether the same population was used for every outcome, how deviations from the intended treatment were handled, or whether analysis populations differed from the posted arm counts. Those details are not reported in the registry record.

The endpoint description says events were to be summarized by system organ class and preferred term, but the registry does not provide those detailed results. It also does not state whether any interim analyses were planned or conducted. The absence of these details limits reconstruction of the analysis strategy and prevents assessment of how the posted comparison relates to any unreported analyses.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event summary lists 0 serious adverse events among 166 participants at risk in the AG-920 group and 0 among 83 participants at risk in the placebo group. The death count is 0 in each group. These are the serious-event and death figures posted in the registry record; they do not mean that no non-serious events occurred.

The same summary lists other adverse events affecting 38 participants in the AG-920 group and 10 in the placebo group. These counts are separate from the primary TEAE counts of 40 and 13. The registry does not explain whether the categories, timeframes, or definitions are identical, or whether participants overlap between the summaries. The figures should therefore be read as separately reported safety results, not combined or treated as interchangeable.

A zero count for serious events in these groups is a statement about the events recorded during the reported study follow-up. It cannot establish that serious events are impossible, and it does not resolve uncertainty about less common outcomes. The registry record lists no serious adverse-event terms. Details about severity grading, relatedness to treatment, and event narratives are not reported in the registry record.

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

The posted data describe a safety comparison in healthy subjects who received a single topical ocular administration of AG-920 or placebo. The primary endpoint counts participants with at least one treatment-emergent adverse event. The reported participant counts are 40 in the AG-920 group and 13 in the placebo group, with group sizes of 166 and 83. The unadjusted comparison is uncertain: its confidence interval includes the null value, and the posted unadjusted p-value is 0.125.

These results do not show that the groups had identical event experience. An interval that includes the null can reflect limited precision as well as compatibility with no difference. Nor does the direction of the point estimate alone establish a dependable treatment effect. The registry comparison is best interpreted as an estimate with uncertainty, not as a definitive conclusion about the presence, absence, or importance of a difference.

The safety summaries add that no serious adverse events or deaths were reported in either group, while other adverse events were recorded for 38 participants in the AG-920 group and 10 in the placebo group. Those categories should remain distinct from the primary TEAE count because their relationship is not explained in the registry record. The aggregate counts do not reveal the event types, severity, duration, or attribution for individual cases.

The study’s stated purpose was ocular safety, but the posted information does not answer every safety question. It is limited to the study population, administration, endpoints, and observation period described in the registry. Generalizing to repeated administration, other populations, or longer observation is not supported by these data alone. The listed secondary outcome concerning corneal endothelial cell density has no numerical result reported in the registry record.

Accordingly, the result is a description of observed participant counts and an imprecise unadjusted comparison. It is not a complete clinical benefit-risk assessment and does not establish whether AG-920 should be used. Such conclusions require evidence beyond the posted summary, including fuller outcome details and relevant clinical context.

16. Why This Trial Matters Statistically

This trial illustrates why a safety result needs more than a count in each group. The primary endpoint counts participants with at least one treatment-emergent adverse event, while separate safety summaries report serious events, deaths, and other adverse events. These measures describe different aspects of safety and should not be collapsed into a single total without definitions showing that they can be combined.

The comparison also illustrates the role of uncertainty. The point estimate favors a higher event proportion in the AG-920 group, but the risk-difference interval crosses zero, and the relative-measure intervals include one. The unadjusted p-value is 0.125. Reporting the estimate together with its interval gives a fuller account than treating the participant counts as a conclusive difference. It helps readers distinguish the direction of the observed result from the precision with which that direction is estimated.

Population and follow-up definitions matter in safety research. The study was conducted in healthy subjects and involved a single topical administration. The outcome timeframe covers the treatment day through 4 days following treatment, with follow-up to 90 days for subjects participating in specular microscopy. Those features define the setting to which the posted summaries apply. They do not automatically describe other populations or patterns of use.

Finally, registry reporting makes the limits of interpretation visible. A secondary outcome is listed but no numerical result is reported; detailed TEAE categories, missing-data methods, and multiplicity procedures are also not reported in the registry record. Clear distinction between observed counts, uncertainty, and unreported analysis details is essential for a transparent statistical reading of a safety study.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The study population is described as healthy subjects. Demographic characteristics and detailed eligibility criteria are not reported in the registry record, limiting assessment of how the participants compare with other populations. The study evaluates a single topical ocular administration, so the posted results do not describe repeated administration. The registry record does not provide evidence about whether the observed pattern would persist under other dosing schedules.

The primary outcome has a defined but limited observation window: from Day 1 through 4 days following treatment, with follow-up to 90 days following treatment for subjects participating in specular microscopy. The number of subjects in that longer follow-up group and the event ascertainment procedures are not reported in the registry record. A limited timeframe may not capture events occurring later, and the registry summary does not allow that possibility to be assessed.

The primary TEAE counts are aggregate participant counts. A breakdown by system organ class and preferred term, despite being described as part of the planned summary, is not reported in the registry record. Details of severity, seriousness within the primary endpoint, attribution, duration, and recurrent events are also not reported. The separate serious and other adverse-event summary has no detailed event terms, and its relationship to the primary TEAE endpoint is not explained.

The registry record does not report the original sample-size justification, missing-data procedures, multiplicity strategy, interim-analysis plan, or a full statistical analysis plan. It also does not provide a numerical result for the listed endothelial-cell-density outcome. These gaps limit assessment of the planned analysis, the handling of incomplete observations, and the breadth of the reported safety evidence. The posted unadjusted comparison should therefore be interpreted as a summary with substantial limits on detail and generalizability.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Safety Study of of AG-920 Sterile Topical Ophthalmic Solution — NCT05501860.
• Registry record last updated 2024-10-09; status last verified 2024-04.
• Results first posted 2024-10-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

NCT05501860 is a completed Phase 3 randomized, placebo-controlled, parallel-group study sponsored by American Genomics, LLC. It enrolled 249 healthy subjects and evaluated a single topical ocular administration of AG-920 against placebo. The registry reports a start date of 2022-07-30 and a primary completion date of 2022-12-07. Its title and brief summary describe double masking, while the design fields classify masking as triple.

The primary outcome was the number of participants with treatment-emergent adverse events. The posted counts were 40 of 166 for AG-920 and 13 of 83 for placebo; the unadjusted p-value was 0.125, and the risk-difference confidence interval crossed zero. The serious adverse-event and death counts were 0 in both groups. One secondary outcome, mean change from baseline in corneal endothelial cell density, is listed, but a numerical result is not reported in the registry record.