ESBA105 in Patients With Severe Dry Eye: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for ESBA105 in Patients With Severe Dry Eye (NCT01338610), a randomized phase 2 study. The primary outcome is Visual Analog Scale (VAS) Global Ocular Discomfort Score, Area Under the Curve, Day 0 to Day 28.

2. Clinical Question

3. Trial Design

This was a randomized, double-masked, parallel phase 2 trial of ESBA105 in patients with severe dry eye. The registry lists Alcon Research as sponsor and the status as completed. The listed study start was 2011-06, and primary completion was 2012-02. Enrollment was 334. The registry record does not report the original sample-size justification or the design assumptions used to choose enrollment.

The active intervention was ESBA105 ophthalmic solution; the comparator was ESBA105 vehicle. Both arms were described as 1 drop in each eye 3 times per day for 4 weeks. Random allocation and double masking are stated in the record, but details of the allocation procedure and masking implementation are not reported in the registry record. The parallel model indicates that the comparison was between separately assigned groups rather than a crossover comparison.

The primary outcome population included randomized subjects who received at least one administration of study medication. The primary outcome lists 56.0 for ESBA105 and 29.0 for vehicle. Although total enrollment is reported as 334, the record does not explain how that figure relates to these outcome counts or provide a participant flow account in the material summarized by the registry result.

The primary endpoint was a Visual Analog Scale global ocular discomfort score area under the curve from Day 0 to Day 28. Subjects assessed frequency and severity at Day 0 before treatment and daily thereafter for 28 days. A mixed model repeated measure approach was used to handle missing data during the randomized treatment period. Further details of the model specification, covariates, and analysis implementation 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.

Before assignment: Of the 334 enrolled, 210 subjects did not qualify for Run-In and were exited from the study without exposure to product. Of the 124 entering Run-In, 39 did not qualify for treatment. The 85 subjects qualifying for treatment were randomized 2:1 to receive either ESBA105 or Vehicle.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All subjects randomized to treatment and receiving at least 1 administration of study medication (intent-to-treat). Mixed model repeated measure (MMRM) approach was used to handle missing data during randomized treatment period.

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: Baseline data is presented for all subjects exposed to product.

6. Primary Endpoint

The registry lists the primary outcome as Visual Analog Scale (VAS) Global Ocular Discomfort Score, Area Under the Curve, Day 0 to Day 28, measured over Up to 28 days.

The primary endpoint was the Visual Analog Scale (VAS) Global Ocular Discomfort Score area under the curve from Day 0 to Day 28, reported in units on a scale x days. Subjects used an electronic VAS to rate both the frequency and severity of ocular discomfort. Frequency ranged from 0 (rarely) to 100 (all the time), while severity ranged from 0 (very mildly uncomfortable) to 100 (very severely uncomfortable). The global score combined the frequency and severity ratings and ranged from 0 to 100.

Assessments were recorded at Day 0 before treatment and daily thereafter for 28 days. An area-under-the-curve endpoint combines the repeated scores across the stated period into a summary measure, rather than relying on a single daily assessment. In general, this can represent the overall pattern of reported discomfort over time, while also making the result dependent on the scale and time period used. Because higher scores reflect more frequent or more severe discomfort, a lower area-under-the-curve value corresponds to lower reported discomfort on this measure.

The reported least-squares means were 63.88 for ESBA105 and 70.22 for vehicle. These values are expressed in scale x days, not as a single-day VAS rating. The registry record does not report additional secondary outcomes.

7. Results Posted in the Registry

• ESBA105: mean 63.88 versus 70.22 on control; difference -6.34 (95% CI -13.11 to 0.42), standardized effect d = -0.38; z-test p = 0.066.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous measure, and its reported parameter is a least-squares mean. The comparison is therefore expressed as a difference between group means on the area-under-the-curve scale. The reported estimate is -6.342, with ESBA105 subtracted from the vehicle comparison in the sense that the estimate is ESBA105 minus vehicle. A negative value indicates a lower estimated score in the ESBA105 group. Because the underlying scale reflects discomfort frequency and severity, lower scores point toward less reported discomfort, without by themselves establishing a treatment effect.

What does a least-squares mean represent?

In general, a least-squares mean is a model-based estimate of a group mean. It can account for features included in the statistical model, unlike a simple arithmetic average calculated directly from raw observations. The registry identifies the endpoint parameter as a least-squares mean and says that a mixed model repeated measure approach was used for missing data. However, the covariates, model terms, covariance structure, and other implementation details are not reported in the registry record. It is therefore not possible to describe which adjustments, if any, shaped the posted estimates.

Why use a repeated-measures model?

In general, repeated-measures methods are used when outcomes are recorded for the same participants on multiple occasions. They can account for the fact that observations from one participant are related, rather than treating every rating as independent. The registry says the eVAS was assessed at Day 0 and daily thereafter for 28 days and specifies a mixed model repeated measure approach to handle missing data during the randomized treatment period. The record does not report the exact model specification or assumptions governing the relationship among repeated scores.

Missing-data methods matter because the analysis may depend on how much outcome information is absent and why it is absent. A model-based method can use available repeated observations under its assumptions, but it cannot guarantee that missingness has no effect on the result. The registry record does not report the amount, timing, or reasons for missing measurements, nor does it describe sensitivity analyses. Those omissions limit assessment of how robust the estimate is to missing-data assumptions.

How should the standard errors be read?

The group values are accompanied by standard errors of 2.483 for ESBA105 and 2.398 for vehicle. In general, a standard error describes uncertainty in an estimated quantity, such as a mean. It is not the same as the standard deviation, which describes variation among individual measurements. The posted standard errors concern the estimated group means; they do not show the full distribution of participants’ discomfort scores. The registry record does not provide a standard deviation for the primary endpoint.

What does the mean difference add?

The mean difference directly expresses the estimated contrast between treatment groups in the endpoint’s units. Here, the estimate is -6.342 units on a scale x days. The confidence interval ranges from -13.11 to a positive upper bound of 0.424, and the unadjusted p-value is 0.066. In general, the confidence interval conveys a range of values compatible with the data and the analysis assumptions. Since the interval includes zero, the posted result does not exclude a null difference. The p-value describes how unusual the observed result, or a more extreme result, would be under a null model; it is not the probability that the null hypothesis is true.

The registry record does not report whether the p-value or confidence interval reflects any multiplicity adjustment. No additional primary outcome analysis is posted in the record. These methods details should be kept distinct from the numerical result: the estimate describes the observed model-based contrast, while the uncertainty and analysis assumptions determine how strongly that contrast can be interpreted.

10. Confidence Intervals and Precision

The estimated mean difference was -6.342 units on a scale x days, with a confidence interval extending from -13.11 to a positive upper bound of 0.424. This interval is important because it shows the range of effects compatible with the posted estimate and its uncertainty under the analysis. It includes zero, so the result does not rule out no difference between groups. It also includes negative values, which correspond to lower scores with ESBA105, and a positive upper bound, which allows for a small difference in the opposite direction.

A confidence interval is not a guarantee that the true effect lies within its endpoints, nor does it state the probability that any one value is correct. In general, it reflects the uncertainty produced by the data and statistical model under repeated use of the same method. The confidence level is not reported in the registry record, so it should not be assigned a specific level based on assumption.

The group estimates were 63.88 for ESBA105 and 70.22 for vehicle, with standard errors of 2.483 and 2.398, respectively. Those standard errors describe precision of estimated means, not variability among individual participants. The confidence interval for the difference is more directly relevant to the group contrast. Its inclusion of zero and its span across negative and positive values indicate that the estimated direction is uncertain. The registry record does not report a clinical-importance threshold, so statistical precision cannot be translated into a judgment about clinical importance.

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, not the original trial’s design calculation. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to a stated fraction of the observed effect. The registry record does not report the trial’s own original sample-size assumptions or justification.

For planning for 100% of the observed effect, the figures are 112 participants per arm for 80% power and 149 participants per arm for 90% power. For planning for 75% of the observed effect, they are 198 participants per arm for 80% power and 265 participants per arm for 90% power. For planning for 50% of the observed effect, they are 446 participants per arm for 80% power and 596 participants per arm for 90% power.

In general, the 80% power column describes the planned number per arm under the stated assumptions for a design with an 80% chance of detecting the specified effect if that effect is the true one. The 90% column describes the corresponding planning target for a 90% chance. Smaller effects require more participants because they are harder to distinguish from random variation. These planning figures are not observed power; they are hypothetical sample-size illustrations, not a retrospective measure of the trial’s success.

12. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -6.342; pooled standard deviation σ = 16.884.
2. n = 2 × (1.96 + 0.842)² × 16.884² / -6.342² = 111.3, rounded up to 112 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -6.34209999999999                      # observed difference in means
sigma <- 16.8842820021965                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1338610)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 112 (80% power), 149 (90% power)
 75% of observed effect: n per arm = 198 (80% power), 265 (90% power)
 50% of observed effect: n per arm = 446 (80% power), 596 (90% power)
power.t.test: 112.2 
Simulated power at n = 112 per arm: 0.801 (20000 replicates)

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

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome measure and no secondary outcomes. It does not report additional outcome analyses. No multiplicity adjustment or strategy for controlling the chance of false-positive findings across multiple comparisons is reported in the registry record. The posted p-value is described as unadjusted in the registry results, so it should not be interpreted as evidence that multiplicity was addressed.

Multiplicity can arise when a study evaluates several endpoints, time points, treatment contrasts, or analysis choices. In general, testing more hypotheses can increase the chance that at least one result appears statistically notable by chance. The registry record does not describe a prespecified testing hierarchy, a correction procedure, or interim analyses. The absence of those details in the record does not establish what was or was not done beyond what is reported; it means the multiplicity strategy cannot be assessed from this registry information.

The primary outcome population was defined as all subjects randomized to treatment who received at least one administration of study medication. A mixed model repeated measure approach was used to handle missing data during the randomized treatment period. The registry record does not report the number of missing assessments, their distribution across groups, reasons for missingness, or a sensitivity analysis testing alternative missing-data assumptions. It also does not provide a full account of exclusions from the endpoint analysis.

These limits matter because missing measurements and analysis choices can affect the estimated group difference. In general, models for repeated measurements rely on assumptions about the observed data and how missing observations relate to unobserved values. The record identifies the method but not enough implementation detail to evaluate those assumptions. No additional analysis population definitions are reported, so the posted result should be understood in the specific population definition stated for the primary outcome.

14. Safety Results

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

Other (non-serious) adverse events

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

The registry lists serious adverse event figures for the run-in group and both randomized treatment groups. There were 0 serious adverse events among the 124 participants at risk in the run-in group, 0 among the 56 participants at risk in the ESBA105 group, and 0 among the 29 participants at risk in the vehicle group. The registry also reports 0 other adverse events for each of those groups. These are the recorded counts in the registry, not estimates of comparative risk.

The record lists no serious adverse event terms. Death counts are not reported in the registry record. A zero count in the posted safety figures means no events were recorded in that category for the stated group; it does not establish that an event is impossible or that the treatments have identical safety profiles. The registry record does not provide further detail on event definitions, follow-up beyond the stated reporting, or comparative safety analyses.

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

The primary result is an estimated lower global ocular discomfort score area under the curve with ESBA105 than with vehicle: 63.88 versus 70.22. The mean difference was -6.342 units on a scale x days. Because the underlying score combines frequency and severity of discomfort, the negative estimate is in the direction of lower reported discomfort with ESBA105. This describes the numerical result, not proof that ESBA105 caused a reduction.

The confidence interval ranges from -13.11 to a positive upper bound of 0.424 and includes zero. The unadjusted p-value was 0.066. Thus the posted result does not exclude no group difference, and uncertainty remains about both the size and direction of the underlying contrast. The registry record does not report a clinical-importance threshold, so it cannot establish whether a difference of this magnitude would be meaningful to patients.

In general, a trial estimate is not interchangeable with a definitive claim about benefit or lack of benefit. An interval that includes zero does not show that the treatments are the same; it indicates that the observed data are compatible with no difference as well as with a range of other effects. Likewise, the point estimate alone does not establish a reliable effect. Interpretation is limited by the posted analysis details, including the absence of reported information on missing-data patterns and the model’s full specification.

The outcome reflects subjects’ reported ocular discomfort during the stated measurement period. It does not answer questions about outcomes that are not posted in the registry record. The record does not report secondary outcomes. The safety figures list zero serious adverse events in the reported groups, but deaths are not reported in the registry record and the safety information does not establish that no risk exists. Overall, the result should be read as an uncertain estimate for a specific self-reported endpoint and trial population, rather than as a general conclusion about treatment effects.

16. Why This Trial Matters Statistically

This trial illustrates why the direction of an estimated difference and the precision around it need to be considered together. The ESBA105 group had a lower posted mean area-under-the-curve score than the vehicle group, and the mean difference was negative. Yet the confidence interval includes zero and the unadjusted p-value was 0.066. A point estimate that favors one group is not, by itself, enough to establish a clear between-group difference.

The endpoint also demonstrates how a repeated self-reported measure can be summarized over time. The global ocular discomfort score combined frequency and severity ratings, while the area-under-the-curve measure represented assessments across Day 0 to Day 28. Such a summary is useful for describing an outcome over a period, but its interpretation remains tied to the instrument, scale, and time window. The result is in units on a scale x days and should not be mistaken for a single-day rating.

The registry identifies use of a mixed model repeated measure approach for missing data, which makes missingness and model assumptions relevant to interpretation. The record does not report the amount or pattern of missing data or the model’s full details. It also does not report the original sample-size justification. The planning figures based on the observed effect are therefore illustrations, not evidence about the original design rationale or the information available when the trial was planned.

Finally, the posted outcome population counts differ from the registry’s total enrollment, and the record does not explain that relationship. Recognizing what is reported and what remains undocumented is part of sound statistical reading: it keeps the interpretation focused on the actual endpoint estimate without implying more certainty or context than the registry record supports.

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 primary outcome is based on self-reported ocular discomfort, with frequency and severity combined into a global score. The record describes the instrument and range but does not report a threshold for clinical importance. Consequently, the magnitude of the estimated contrast cannot be classified against a trial-specific clinical benchmark.

The registry reports enrollment of 334, whereas the primary outcome lists 56.0 for ESBA105 and 29.0 for vehicle. The record does not explain the relationship between overall enrollment and these outcome counts or provide a participant flow account in the information reported for this outcome. Reasons for any difference are not reported in the registry record and should not be inferred.

Although a mixed model repeated measure approach was used to handle missing data during the randomized treatment period, the amount and pattern of missing observations and the model specification are not reported. The record also does not describe sensitivity analyses for missing-data assumptions. These omissions make it difficult to assess how dependent the estimate may be on the analysis approach.

The registry record does not report the original sample-size justification, additional secondary outcomes, or details of any multiplicity strategy. The confidence interval includes zero, and the unadjusted p-value was 0.066. Taken together, the result remains uncertain; neither the point estimate nor the planning illustrations resolve that uncertainty. The posted safety figures also have limits: deaths are not reported in the registry record, and zero recorded serious adverse events should not be treated as evidence of no possible safety risk.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: ESBA105 in Patients With Severe Dry Eye — NCT01338610.
• Registry record last updated 2013-04-19; status last verified 2013-04.
• Results first posted 2013-04-19.
• 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

The registry record describes a completed, randomized, double-masked, parallel phase 2 trial sponsored by Alcon Research. The study evaluated ESBA105 ophthalmic solution against ESBA105 vehicle in patients with severe dry eye. Registry enrollment was 334; the listed study start was 2011-06 and primary completion was 2012-02.

The primary outcome was the Visual Analog Scale Global Ocular Discomfort Score area under the curve from Day 0 to Day 28. The posted least-squares means were 63.88 for ESBA105 and 70.22 for vehicle, with a mean difference of -6.342. The confidence interval included zero, and the unadjusted p-value was 0.066. A mixed model repeated measure approach was used to handle missing data, but further model and missingness details are not reported in the registry record.

No secondary outcomes are reported. The safety record lists zero serious adverse events in each reported group; deaths are not reported in the registry record.