Safety and Tolerability Study of Suprachoroidal Injection of CLS-AX Following Anti-VEGF Therapy in Neovascular AMD: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Tolerability Study of Suprachoroidal Injection of CLS-AX Following Anti-VEGF Therapy in Neovascular AMD (NCT04626128), a non-randomized phase 1, phase 2 study. The primary outcome is Number of Participants With Treatment-emergent Adverse Events (TEAEs).

2. Clinical Question

This is an early-phase safety and tolerability question, rather than a randomized estimate of treatment efficacy. The registry describes CLS-AX administration after intravitreal anti-VEGF therapy in people with neovascular age-related macular degeneration. The primary endpoint counts participants with treatment-emergent adverse events. A participant count is a direct way to describe how many people in the defined safety population had an event, but it does not by itself describe event severity, duration, recurrence, or the relationship of each event to treatment.

The registry also lists outcomes concerning pre-injection intraocular pressure, additional intravitreal aflibercept injections and qualification for those injections, central subfield thickness, best corrected visual acuity, and maximum plasma concentration of axitinib. These outcomes address a range of safety, treatment-use, eye-measurement, and pharmacokinetic questions. Their results are not reported in the registry record, so they cannot be interpreted alongside the adverse-event counts.

Because the cohorts were sequential and there was no randomized control group, differences in counts between cohorts cannot isolate a dose effect. The study’s question is therefore best understood as descriptive: what safety events were recorded among the participants in each dose cohort during the specified observation period?

3. Trial Design

The registry lists OASIS as an open-label Phase 1/2a study of the safety and tolerability of suprachoroidally administered CLS-AX following intravitreal anti-VEGF therapy in subjects with neovascular age-related macular degeneration. The study is marked completed, with enrollment of 27. Its recorded start date is 2020-12-15, and its primary completion date is 2022-10-13.

Allocation was non-randomized, masking was none, and the study model was sequential. The registry describes four dose cohorts: Cohort 1 (Low Dose), 0.03 mg; Cohort 2 (Low-mid Dose), 0.10 mg; Cohort 3 (High-mid Dose), 0.50 mg; and Cohort 4 (High Dose), 1.0 mg. The description attached to Cohort 4 calls its 1.0 mg dose “high-mid,” while the cohort title calls it “High Dose.” This wording difference is present in the registry record.

A sequential dose-escalation structure means that cohorts are considered in an ordered sequence rather than assigned by random allocation to concurrent treatment groups. In general, such a design can describe safety observations at successive dose levels, but it does not provide the protection against selection differences that randomization is intended to offer. Open-label status means the study was not masked; the record does not describe additional procedures to limit the influence of that knowledge on observation or reporting.

The registry lists CLS-AX and Anti-VEGF as drug interventions and describes CLS-AX use following intravitreal anti-VEGF therapy. The detailed anti-VEGF regimen, the timing between therapies, and any other dosing procedures are not reported in the registry record. Nor is a separate anti-VEGF-only arm described. Therefore, the design does not support a direct randomized comparison of CLS-AX with anti-VEGF alone.

The stated primary endpoint covers treatment-emergent adverse events from Day 1 to Week 12. Although the registry names the study as a safety and tolerability investigation, detailed protocol procedures, dose-escalation decision rules, and the statistical analysis plan are not reported in the registry record. The design and endpoint descriptions should be interpreted at that level of detail, without assuming unlisted procedures.

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: Defined as all enrolled participants who were administered CLS-AX, and from whom at least one post-Baseline safety measurement was obtained.

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 Day 1 to Week 12.

The primary endpoint is the number of participants with treatment-emergent adverse events (TEAEs), measured from Day 1 to Week 12. The registry describes the events as those reported between administration of CLS-AX and study exit. The population is defined as all enrolled participants who received CLS-AX and had at least one post-Baseline safety measurement.

A participant-count endpoint records whether participants experienced an event during the specified period. In general, this is a useful descriptive measure in an early-phase safety study because it communicates how many people had one or more events, rather than only listing event reports without a participant-level context. It is not a measure of how many separate events each affected participant experienced. Nor does the count alone indicate event severity, duration, clinical importance, or whether investigators considered an event related to study treatment.

The observation window matters: the endpoint refers to events in the stated interval, not all possible events over a participant’s lifetime or after study exit. The registry record does not provide further detail about event coding, attribution, recurrence handling, or how incomplete follow-up was addressed. Those details are not reported in the registry record.

The adverse-event summary separately reports serious events, other adverse events, and deaths by cohort. These categories help describe the safety observations but should not be treated as substitutes for a complete account of the primary TEAE endpoint. In particular, the record does not list a statistical comparison or an effect estimate for the primary endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record identifies a non-randomized, open-label, sequential study and a participant-count primary endpoint. It does not list statistical analyses, and detailed statistical methods are not reported in the registry record. The following explains the general statistical ideas relevant to reading these results; it does not imply that every method described was used in OASIS.

What does a participant count represent?

A participant-count endpoint records the number of people who experienced the outcome, according to the endpoint definition. Here, the primary outcome is the number of participants with treatment-emergent adverse events during the stated Day 1 to Week 12 period. The unit is participants, not separate event episodes. If one participant experienced multiple events, the participant-count endpoint would still concern whether that person had an event; the registry’s endpoint description does not specify how repeated events were handled in any event-level analysis.

Counts are straightforward to communicate, but a count without its denominator is difficult to interpret. The adverse-event summary gives participants at risk for each category and cohort. Those denominators provide context for the reported counts. A count describes the observed group; it is not automatically an estimate of the probability of an event in another population. Generalization requires attention to how participants were selected, how completely they were followed, and whether the observed participants resemble the population of interest.

Why does the design affect interpretation?

Random allocation is used in general to make treatment groups comparable, on average, with respect to measured and unmeasured factors. This study is recorded as non-randomized, with no masking and a sequential model. Its cohorts therefore should not be treated as exchangeable arms in a randomized experiment. Participants enrolled at different points in a sequential study may differ for reasons unrelated to dose, and care or observation may also change over time. The registry does not report whether or how such factors were measured or adjusted for.

In general, a randomized concurrent comparison can help separate a treatment contrast from background events and other time-related influences. No randomized comparator is described in this record. Thus, the cohort counts can describe safety observations at the listed dose levels, but do not by themselves estimate what would have happened to comparable participants without CLS-AX. A comparison among sequential cohorts would also be vulnerable to confounding, meaning a difference could reflect factors other than the dose level.

What does an uncertainty interval add?

In general, a confidence interval gives a range of values that are compatible with the observed data under a specified statistical procedure. It conveys the uncertainty in an estimate, which a single count or point estimate cannot show. For event counts from limited cohorts, plausible underlying event frequencies can span a wide range. Confidence intervals do not correct weaknesses in study design, establish causation, or guarantee that a parameter lies in the particular calculated interval. No confidence intervals for the adverse-event counts are reported in the registry record.

A hypothesis test, in general, evaluates data relative to a specified null hypothesis and analysis plan. A p-value is not the probability that a treatment works, nor a direct measure of effect size or clinical importance. The registry record lists no analyses for these outcomes, and it does not report p-values or a prespecified test for a cohort contrast. No inferential conclusion should be inferred from the counts alone.

Why can event counts be difficult to compare?

Adverse-event data can involve different event types, severities, timing, recurrence, and judgments about seriousness or attribution. A participant-level count can conceal that variation. In general, over-dispersion occurs when observed variability is greater than a statistical model assumes. It can matter for analyses of event rates or repeated events because a model that assumes too little variability may understate uncertainty. The registry does not report a repeated-event analysis or a model for over-dispersion, so this is a general consideration rather than a description of the study’s analysis.

The record reports zero serious adverse events in each cohort and provides counts of participants with other adverse events. These observed zeros are not equivalent to a zero underlying risk. With finite observation and limited cohort sizes, events that did not occur in the observed participants may still occur in a wider population. In general, uncertainty methods can help express that limitation, but no such estimates are reported.

How should planning illustrations be understood?

Any sample-size planning illustration is distinct from the number actually observed and from a trial’s original design justification. The assumptions described for the planning illustration are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is expressed as a fraction of an observed effect. The illustration is not the trial’s own design, and it should not be read as observed power. The record does not report the original sample-size justification. Because this study is non-randomized and sequential, an equal-allocation, two-arm planning setup is not a description of its actual cohort structure.

In general, power is a design-stage probability of detecting a specified effect under stated assumptions, not a property that can be recovered as a meaningful explanation of a completed study’s observed result. A planning calculation based on an observed effect can be especially dependent on that effect estimate. It does not establish that the observed effect is reliable or that a future study will reproduce it. The record does not report the effect estimate or detailed planning outputs needed to assess such illustrations in greater depth.

The study’s statistical analysis plan, including methods for handling missing outcomes, covariate adjustment, uncertainty estimation, and any interim review, is not reported in the registry record. Accordingly, interpretation should stay close to the endpoint definitions and observed cohort counts rather than presuming a statistical procedure that is not described.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Cohort 1 (Low Dose)", "Cohort 2 (Low-mid Dose)", "Cohort 3 (High-mid Dose)", "Cohort 4 (High Dose)")
n      <- c(6, 5, 8, 8)
events <- c(1, 2, 1, 3)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Cohort 1 (Low Dose)  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9
ARM  Cohort 2 (Low-mid Dose)  5  2  40.0  5.3  85.3  11.8  76.9  11.8  77.7
ARM  Cohort 3 (High-mid Dose)  8  1  12.5  0.3  52.7  2.2  47.1  2.8  48.2
ARM  Cohort 4 (High Dose)  8  3  37.5  8.5  75.5  13.7  69.4  13.7  70.1

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary safety endpoint and several secondary outcomes: change in pre-injection intraocular pressure, additional intravitreal aflibercept injections and qualification to receive them, change in central subfield thickness, change in best corrected visual acuity letter score, and maximum plasma concentration of axitinib. The registry record does not report results for those secondary outcomes in the information. Their presence in the outcome list should not be confused with evidence that a particular result was observed or statistically tested.

Multiplicity arises in general when several outcomes, comparisons, time points, or subgroup analyses are tested. With many tests, the chance of at least one apparently notable finding can increase unless the analysis plan addresses that issue. Whether and how multiplicity was handled in OASIS is not reported in the registry record. No analysis results or multiplicity adjustment are listed, so it is not possible to assess a family-wise error approach, a hierarchical testing strategy, or another prespecified method.

The primary safety population is defined as all enrolled participants who were administered CLS-AX and had at least one post-Baseline safety measurement. That definition clarifies who qualifies for the primary endpoint population, but it does not describe how participants without a post-Baseline measurement were handled in every outcome analysis. A separate intention-to-treat population is not described in the registry record.

Missing-data methods, the amount of missing information, and reasons for missing measurements are not reported in the registry record. In general, missingness can affect estimates if participants with unavailable measurements differ systematically from those with observed measurements. Complete-case summaries, imputation, and sensitivity analyses make different assumptions; no particular approach should be attributed to this study without documentation. Interim analyses and stopping rules are also not reported in the registry record.

12. 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 0% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse-event summary reports 0 affected participants in Cohort 1 (Low Dose), among 6 at risk; 0 in Cohort 2 (Low-mid Dose), among 5 at risk; 0 in Cohort 3 (High-mid Dose), among 8 at risk; and 0 in Cohort 4 (High Dose), among 8 at risk. The record lists no top serious adverse event terms, and deaths are reported as 0 in each cohort.

For other adverse events, the corresponding affected-participant counts are 1 of 6 in Cohort 1, 2 of 5 in Cohort 2, 1 of 8 in Cohort 3, and 3 of 8 in Cohort 4. These are counts, not comparisons adjusted for differences between cohorts. Because allocation was non-randomized and sequential, the pattern should not be interpreted as evidence that the dose level caused a difference in event occurrence.

In general, a reported zero means no event of that category was recorded among the participants at risk in the stated summary; it does not demonstrate that the underlying risk is zero. The record’s counts also do not establish event relatedness, duration, recurrence, or severity for other adverse events. The registry does not report additional details about individual event narratives or how attribution was assessed. The figures are therefore best read as descriptive safety observations within this study, not as a guarantee about events outside its observed participants or timeframe.

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

The registry record supports a narrow description of the study: OASIS was a completed, open-label, non-randomized, sequential Phase 1/2a study evaluating safety and tolerability of suprachoroidal CLS-AX following intravitreal anti-VEGF therapy in neovascular age-related macular degeneration. It lists four dose cohorts and reports cohort-level counts for serious and other adverse events, as well as deaths. The serious-event and death counts are zero in each cohort; other adverse-event counts vary across cohorts.

These observations do not, by themselves, show that CLS-AX caused the events or that it prevented them. Nor do they establish that one dose is safer or less safe than another. The study lacks randomized allocation, and the registry does not describe a concurrent comparator group. Cohort differences may therefore reflect factors other than dose. The listed counts also do not provide a complete account of the severity, timing, recurrence, or attribution of every event.

The study is not a randomized efficacy comparison. The registry lists outcomes involving eye measurements, visual acuity, additional aflibercept treatment, and axitinib concentration, but results for those outcomes are not reported in the registry record. The adverse-event information should not be used to infer changes in vision, retinal anatomy, need for other treatment, or drug exposure.

In general, early-phase studies can contribute descriptive information about tolerability and guide questions for later research. Such contribution is different from demonstrating a treatment effect in a controlled comparison. The registry record does not report a statistical analysis plan, inferential comparisons, confidence intervals, or detailed missing-data methods. Without those elements, precision and robustness cannot be assessed beyond the reported counts.

The most defensible conclusion is limited to what was recorded in these cohorts during the study’s stated safety observation period. The data neither prove safety in a broader population nor establish efficacy. They should be interpreted alongside the design limitations and the incomplete reporting of outcome analyses, rather than as a definitive comparative assessment.

14. Why This Trial Matters Statistically

This trial is statistically notable for the distinction between a safety description and a causal comparison. The registry asks how many participants experienced treatment-emergent adverse events after CLS-AX administration, with safety observations grouped by sequential dose cohort. That is a meaningful descriptive question, but the design does not include randomized allocation or a concurrent control group. The resulting counts cannot isolate the effect of CLS-AX from background events or differences between cohorts.

The study also illustrates why a denominator and a clearly defined population matter. The primary safety population is specified as enrolled participants who received CLS-AX and had at least one post-Baseline safety measurement. The adverse-event summary gives the participants at risk for each cohort and category, allowing the event counts to be read in context. Even with denominators, however, limited cohort sizes leave substantial uncertainty about how often events might occur beyond the participants observed.

Another statistical lesson is that zero observed serious adverse events is not a proof of zero risk. A finite study can describe what occurred during its observation period; it cannot rule out events that were not observed. Uncertainty estimates can make that limitation more explicit, but none are reported for these counts in the registry record.

Finally, the record lists multiple secondary outcomes but does not report their results or statistical analyses in the information. This underscores the value of transparent reporting of outcome definitions, analysis populations, missing data, multiplicity decisions, and uncertainty estimates. Without those details, readers should avoid treating a list of planned or named outcomes as a set of established findings. The principal statistical contribution of the reported information is therefore descriptive, bounded by its study design and reporting.

15. 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 most direct limitation is the non-randomized, sequential design. The four dose cohorts are not randomized comparison groups, and no concurrent control arm is described. Differences in observed event counts cannot be attributed to dose without assumptions that the registry record does not establish. The open-label design also means masking was absent; the record does not describe procedures that might reduce the influence of awareness on reporting or assessment.

The cohort sizes listed in the adverse-event summary are limited, which restricts precision for uncommon events. The absence of recorded serious events in a cohort does not rule out a nonzero underlying occurrence in a broader group. No confidence intervals or other uncertainty estimates for the event counts are reported. The registry record also does not provide detailed event-level information about severity, recurrence, timing, or treatment attribution.

The primary endpoint has a stated safety population and timeframe, but the record does not report detailed rules for event coding, handling repeated events, or accounting for incomplete follow-up. Missing-data methods and reasons for unavailable measurements are not reported in the registry record. The statistical analysis plan, interim analyses, stopping rules, and original sample-size justification are also not reported.

Although secondary outcomes are listed, their results are not reported in the registry record. No inferential analyses are listed. This limits assessment of the eye measurements, visual acuity, additional aflibercept use, and plasma concentration outcomes, as well as any multiplicity strategy. The anti-VEGF regimen and details of its timing relative to CLS-AX are not reported. Finally, the Cohort 4 label and dose description use different dose-level wording: the title says “High Dose,” while the description calls the 1.0 mg dose “high-mid.”

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Safety and Tolerability Study of Suprachoroidal Injection of CLS-AX Following Anti-VEGF Therapy in Neovascular AMD — NCT04626128.
• Registry record last updated 2023-09-18; status last verified 2023-08.
• Results first posted 2023-08-21.
• 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.

19. Record Summary

OASIS (NCT04626128) is recorded as a completed, open-label, non-randomized, sequential Phase 1/2a study of suprachoroidal CLS-AX following intravitreal anti-VEGF therapy in neovascular age-related macular degeneration. Enrollment is listed as 27. The recorded start date is 2020-12-15, and primary completion is listed as 2022-10-13.

The primary endpoint is the number of participants with treatment-emergent adverse events from Day 1 to Week 12. Four sequential cohorts are listed at doses of 0.03 mg, 0.10 mg, 0.50 mg, and 1.0 mg. The adverse-event summary reports zero serious adverse events and zero deaths in each cohort, alongside cohort-specific counts of participants with other adverse events.

The registry record does not report inferential analyses, confidence intervals, detailed missing-data methods, or results for the listed secondary outcomes. Because allocation was non-randomized and no concurrent control group is described, the cohort counts are descriptive and do not establish comparative treatment effects.