Study to Evaluate the Safety and Efficacy of Epi-on Corneal Cross-linking in Eyes With Progressive Keratoconus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate the Safety and Efficacy of Epi-on Corneal Cross-linking in Eyes With Progressive Keratoconus (NCT03442751), a randomized phase 3 study. The primary outcome is Difference Between Treatment Groups in the Change From Baseline to Month 6 in Kmax.

2. Clinical Question

The registry reports a between-group ANCOVA estimate of -1.0 diopters, with a 95% confidence interval from -1.5 to -0.4 and p < 0.001. The estimate is negative, while the posted mean changes are -0.3 in the treatment group and 0.6 in the sham group. These figures describe the recorded outcome; they do not by themselves establish how the finding applies to people outside the study or address outcomes beyond those reported.

3. Trial Design

The registry identifies this as a Phase III, randomized, double-masked, parallel-group study of epithelium-on corneal collagen cross-linking in eyes with progressive keratoconus. Enrollment is listed as 201, and the study status is completed. The recorded start date is 2018-04-06 and the primary completion date is 2019-11-01.

The treatment group received Test Article A and Test Article B and was exposed to a cross-linking dose of UVA light generated by the KXL medical device system. The sham control group received placebo and a mock dose of UVA light from the same named device system. The registry describes the treatment as epithelium-on, meaning the formulation was intended to allow riboflavin to penetrate without removing the corneal epithelium.

The primary outcome was the between-group difference in change from baseline to Month 6 in Kmax, measured in diopters. The registry also lists a secondary outcome comparing change in Kmax at Month 12. The posted primary result describes an intent-to-treat analysis set that included randomized study eyes with at least one post-treatment follow-up efficacy assessment; missing post-baseline Kmax data were imputed using multiple imputation.

The arm-level outcome counts are 156 for the epithelium-on CXL group and 83 for the sham group. These counts do not match the registry's enrollment figure, and the record does not explain the relationship between the enrollment count and the analysis counts. Randomization procedures, masking procedures in practice, allocation ratio, study sites, eligibility details, and participant characteristics are not reported in the registry record. Nor are reasons for the difference in listed counts reported.

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: 280 eyes of 201 subjects were randomized in a 2:1 treatment allocation. Of the 280 randomized eyes, 279 were treated: 189 eyes of 156 subjects received CXL treatment \& 90 eyes of 83 subjects received Sham/Control, for a total of 239 treatments. Of the 201 unique subjects, 118 received CXL only (in 1 or both eyes), 45 received Sham/Control only (in 1 or both eyes), and 38 subjects received CXL in 1 eye and Sham/Control in the other eye.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat analysis set which included all randomized study eyes that had at least one post-treatment follow-up efficacy assessment. Missing post-baseline Kmax data were imputed using multiple imputation.

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: Subjects with progressive keratoconus in at least one eye (study eye). Subjects may contribute more than one eye if both eyes qualify. Eyes vs. subjects were randomized to the treatment arms.

6. Primary Endpoint

The registry lists the primary outcome as Difference Between Treatment Groups in the Change From Baseline to Month 6 in Kmax, measured over 6 months.

Kmax is the maximum corneal curvature measure, in diopters. The primary endpoint compares the change in Kmax from baseline to Month 6 between the epithelium-on CXL group and the sham control group. The trial description frames this measure as relevant to the aim of impeding keratoconus progression and/or reducing corneal curvature.

Change from baseline focuses the comparison on how the recorded measure shifted over follow-up, rather than on the Month 6 value alone. A between-group contrast then asks whether the average change differed between assigned groups. In this result, each group’s posted value is a mean change, while the registry’s reported primary analysis is a net mean difference between groups.

As a general statistical point, an endpoint is most informative when its definition, time point, analysis population, and handling of missing measurements are clear. The registry identifies the time point as 6 months and the measure as Kmax; it also says that missing post-baseline Kmax data were imputed using multiple imputation in the described analysis population. Details of the measurement procedure and the specific covariates used in the analysis are not reported in the registry record.

Kmax is one outcome measure. The registry also lists a Month 12 Kmax change outcome, but the primary result concerns Month 6. Neither this endpoint nor its statistical contrast alone describes every aspect of safety or every possible clinical outcome.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is reported as an analysis of covariance, or ANCOVA, with the parameter described as a net mean difference. ANCOVA is a general method for comparing group outcomes while accounting for one or more covariates. Which covariates were included in this trial’s model is not reported in the registry record, so the analysis should be described by the method named in the record without assuming a particular adjustment set.

What does a mean difference describe?

Kmax is a continuous measure, and the registry describes the group values as means. A mean difference expresses the estimated separation between the groups on the outcome scale. Here the reported net mean difference is -1.0 diopters. Its sign indicates the direction of the contrast as recorded; the estimate should be interpreted alongside the group labels, the outcome definition, and the confidence interval. It is not a count of participants whose Kmax improved, nor does it describe the distribution of individual responses.

How should the confidence interval be read?

The reported 95% confidence interval extends from -1.5 to -0.4 diopters. In general, a confidence interval reflects the precision of an estimate under the model and its assumptions. A narrower interval indicates less statistical uncertainty than a wider interval on the same scale. The interval is not a guarantee that the true value lies within its limits, and it does not indicate how much outcomes vary between individuals. The posted result’s interval does not include zero, which is consistent with the reported p-value for a test of no between-group difference under the analysis.

What does the p-value contribute?

The reported p-value is < 0.001. In general, a p-value quantifies how unusual data at least as discrepant as those observed would be under a specified null hypothesis and the model assumptions. It does not establish that the null hypothesis is true or false, give the probability that the result occurred by chance, or quantify the importance of the estimated difference. Its interpretation depends on the analysis being specified and performed appropriately. The registry identifies ANCOVA but does not report all details needed to reconstruct the analysis independently.

Why do the arm means and the adjusted contrast differ in form?

The posted group means are -0.3 diopters for the treatment group and 0.6 diopters for the sham group. The primary reported estimate is an ANCOVA net mean difference of -1.0 diopters. A model-based adjusted contrast need not be identical to a simple subtraction of displayed group means, because adjustment can change the estimate and the displayed values may not incorporate the model. The registry does not provide model covariates or additional detail on the calculation. No separate unadjusted re-calculation is reported in the registry record for this continuous two-arm comparison.

Why does over-dispersion matter in general?

Over-dispersion is a general concern for count outcomes when variation exceeds what a chosen probability model expects. It can make standard errors too small if the extra variation is ignored. It is not a reported feature of this primary endpoint: Kmax is a continuous measure, and the registry reports a mean-based ANCOVA analysis. No over-dispersion assessment is reported in the registry record.

How does missing-data handling affect interpretation?

The described analysis population included randomized study eyes with at least one post-treatment follow-up efficacy assessment. The record says missing post-baseline Kmax data were imputed using multiple imputation. In general, multiple imputation uses a statistical model to represent uncertainty about missing values rather than filling every missing value with one fixed number. Its validity depends on the imputation model and assumptions about why data are missing. The imputation variables, number of imputations, assumptions, and sensitivity checks are not reported in the registry record.

Randomization and masking are recorded as design features, but the record does not describe their operational details. Nor does it report a full statistical analysis plan, the covariates in ANCOVA, or an interim-analysis plan. These omissions limit independent assessment of how the reported analysis was specified, while leaving the reported estimate, interval, and p-value as the registry’s stated primary result.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparison as an ANCOVA estimate: a net mean difference of -1.0 diopters, with a 95% confidence interval from -1.5 to -0.4 and p < 0.001. This is the trial’s posted analysis and is the result to use when describing the sponsor-reported primary finding.

The group-level values are -0.3 diopters for the epithelium-on CXL group and 0.6 diopters for the sham group. Those are descriptive mean changes, whereas the reported primary contrast comes from ANCOVA. A simple unadjusted comparison of displayed summary values would not necessarily reproduce a model-based estimate. In general, adjusted and unadjusted estimates can differ because they answer related but distinct statistical questions.

No separate unadjusted recalculated estimate is reported in the registry record for this continuous two-arm comparison. The registry’s reported ANCOVA result should therefore not be conflated with an independently recalculated estimate. The registry does not give the ANCOVA covariates or enough model detail to reconstruct the sponsor analysis from the posted summary statistics alone. The reported result remains the most direct account of the trial’s own analysis, while its precision and interpretation are conveyed by the stated confidence interval and p-value.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome: the between-group difference in change from baseline to Month 6 in Kmax. It also lists a secondary Kmax outcome at Month 12. The record identifies two outcome measures and two analyses, but does not report a multiplicity adjustment, a hierarchy for testing outcomes, or a prespecified strategy for controlling the chance of false-positive findings across multiple tests. Whether any such approach was specified elsewhere is not reported in the registry record.

Multiplicity matters in general when several outcomes, time points, groups, or analyses are examined. Repeated testing can increase the chance that at least one result appears statistically notable even when the tested effects are absent. The primary result should be distinguished from the listed secondary outcome, and the absence of a reported multiplicity method means the record does not support a detailed account of how the study handled that issue. A result for the Month 12 secondary outcome is not reported in the registry record.

The stated primary analysis population included all randomized study eyes with at least one post-treatment follow-up efficacy assessment. Missing post-baseline Kmax data were imputed using multiple imputation. This is important because the result is not described as a complete-case analysis alone, but the registry does not provide the amount or pattern of missing data, the imputation model, the assumptions used, or sensitivity analyses.

The record also does not report whether interim analyses were planned or conducted, or provide a full statistical analysis plan. These elements should not be inferred from the final p-value or from the Phase III designation. Without further detail, readers can identify the stated endpoint, analysis population, and imputation approach, but cannot determine how multiplicity, interim monitoring, or alternative missing-data scenarios were addressed.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: uterine leiomyoma (2.6%), urinary tract infection (0.8%), kidney infection (0.8%), depression suicidal (0.8%), cataract subcapsular (0.5%).

The registry reports serious adverse event counts using separate at-risk groups. In the ocular treatment group, 1 serious adverse event is reported among 189 at risk; in the ocular sham group, 0 are reported among 90 at risk. For non-ocular events, the epithelium-on group lists 2 among 118 at risk, the sham group lists 0 among 45 at risk, and the combined epithelium-on and sham group lists 1 among 38 at risk. Deaths are recorded as 0 in each listed group.

These are counts within the categories and risk sets named in the registry. The non-ocular categories include a combined group, so they should not be treated as mutually exclusive treatment arms or added to make a trial-wide total. The registry does not explain the allocation or overlap behind these safety groupings.

The listed serious-event terms include Uterine leiomyoma at 2.6%, Urinary tract infection at 0.8%, Kidney infection at 0.8%, Depression suicidal at 0.8%, Cataract subcapsular at 0.5%, and Retinal detachment at 0.5%. The registry does not provide enough detail in the registry record to determine the relationship of each term to treatment, timing, severity beyond its listed classification, or causal attribution. The serious-event figures describe reported events; they are not by themselves a complete assessment of comparative safety.

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

The primary posted result is a between-group difference in change in Kmax at Month 6. The registry reports a net mean difference of -1.0 diopters, with a 95% confidence interval from -1.5 to -0.4 and p < 0.001. The group means are -0.3 diopters for epithelium-on CXL and 0.6 diopters for sham. In statistical terms, the estimated contrast is negative and its reported confidence interval lies below zero.

This finding does not describe how every eye changed. A group mean can conceal variation among individual outcomes, and the registry does not report the distribution of individual changes in the registry record. The confidence interval describes uncertainty in the estimated group contrast under the analysis, not a prediction interval for individual eyes. The p-value does not measure clinical importance or the probability that the treatment effect is real.

The result is limited to the recorded primary endpoint and time point. A Month 12 Kmax outcome is listed, but its result is not reported in the registry record. The record also does not report other measures needed to characterize the full clinical experience, such as longer-term outcomes or broader functional outcomes. The posted Kmax contrast alone cannot establish durability or answer every question about safety.

The analysis population and handling of missing data also matter. The primary analysis set included randomized study eyes with at least one post-treatment follow-up efficacy assessment, and missing post-baseline Kmax values were imputed using multiple imputation. The record does not detail the missing-data pattern or sensitivity analyses. In addition, the efficacy group counts and the enrollment figure are different as recorded, with no explanation of their relationship.

Accordingly, the result should be read as evidence about the registry’s specified Month 6 Kmax comparison under its reported ANCOVA analysis. It is not a substitute for the full protocol, analysis plan, complete clinical report, or assessment of outcomes not reported in the registry record. It also should not be generalized beyond the studied setting without information about participant characteristics and applicability.

14. Why This Trial Matters Statistically

This trial is statistically informative because it reports a continuous endpoint as both group mean changes and a between-group estimate with a confidence interval. Those summaries allow readers to distinguish the direction of the estimated contrast from its precision. The reported ANCOVA result is -1.0 diopters, with a 95% confidence interval from -1.5 to -0.4 and p < 0.001.

The randomized, double-masked, parallel-group design is relevant to how a treatment comparison is interpreted. In general, randomization is intended to reduce systematic differences between assigned groups, while masking can limit some forms of influence on treatment delivery or outcome assessment. The registry identifies these design features but does not describe their implementation, so the label alone cannot answer how well these protections operated.

The endpoint also illustrates why an analysis result should be read together with its definition and population. Change in Kmax at Month 6 is a specific measure at a specific time point. The stated analysis set required at least one post-treatment follow-up efficacy assessment, and missing post-baseline Kmax data were imputed using multiple imputation. Both details help define what the estimate represents, while missing methodological detail leaves uncertainty about robustness.

Finally, the record makes clear that different parts of a trial can use different denominators and groupings. Enrollment is listed as 201, primary outcome counts are 156 and 83, and safety summaries use their own at-risk counts. The record does not explain how those figures relate. Careful statistical reading avoids treating them as if they were one common analysis population or combining safety categories without justification.

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.

Several trial-specific details needed to assess the primary result are not reported in the registry record. These include the ANCOVA covariates, randomization and masking procedures, participant eligibility criteria, baseline characteristics, and the detailed measurement process for Kmax. Without these details, readers cannot fully assess comparability, applicability, or reproduce the reported model.

The registry lists enrollment as 201, while the primary outcome group counts are 156 and 83. The reason for these different figures is not reported in the registry record. The primary analysis population is described as randomized study eyes with at least one post-treatment follow-up efficacy assessment, but the record does not provide a participant flow or explain exclusions and missing follow-up. It also does not report the number or pattern of missing Kmax observations.

Multiple imputation is identified, but its model, assumptions, implementation details, and sensitivity analyses are not reported in the registry record. The record also does not describe a multiplicity strategy for the Month 6 primary endpoint and Month 12 secondary outcome, or state whether interim analyses took place. These omissions limit assessment of the full analysis plan.

The safety results use distinct ocular and non-ocular categories, including a combined non-ocular group. The registry entries do not explain overlap or attribution, and they do not provide detailed event timing or causal assessments. A comprehensive comparison of safety cannot be made from the listed serious-event counts alone.

Finally, only the Month 6 primary result is included in the posted outcome details. The Month 12 result and other information about longer-term outcomes are not reported in the registry record. These limitations constrain interpretation to the reported endpoint, population, and follow-up, rather than supporting conclusions about outcomes not recorded.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study to Evaluate the Safety and Efficacy of Epi-on Corneal Cross-linking in Eyes With Progressive Keratoconus — NCT03442751.
• Registry record last updated 2024-01-30; status last verified 2024-01.
• Results first posted 2024-01-30.
• 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

The registry describes a completed Phase III, randomized, double-masked, parallel-group study of epithelium-on corneal collagen cross-linking in eyes with progressive keratoconus. Enrollment is listed as 201. The treatment used Test Article A, Test Article B, and a cross-linking dose of UVA light; the sham group received placebo and a mock UVA dose.

The primary endpoint was the between-group difference in change from baseline to Month 6 in Kmax. The registry reports an ANCOVA net mean difference of -1.0 diopters, with a 95% confidence interval from -1.5 to -0.4 and p < 0.001. Group mean changes were -0.3 in the treatment group and 0.6 in the sham group.

The described analysis population included randomized eyes with at least one post-treatment follow-up efficacy assessment, and missing post-baseline Kmax values were imputed using multiple imputation. The record also lists a Month 12 Kmax outcome, but its result is not reported in the registry record. Details of covariate adjustment, multiplicity, and the relationship between enrollment and outcome counts are not reported.