A Phase 3 Prospective, Randomized, Double-Masked, 12-Week, Parallel Group Study In Pediatric Subjects With Glaucoma.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 3 Prospective, Randomized, Double-Masked, 12-Week, Parallel Group Study In Pediatric Subjects With Glaucoma. (NCT00716859), a randomized phase 3 study. The primary outcome is Reduction From Baseline in Mean IOP at Week 12, Last Observation Carried Forward (LOCF).

2. Clinical Question

The trial addressed a comparative question about intraocular pressure (IOP), a continuous measure. It did not define its primary question as a comparison of the proportion of participants reaching a particular pressure threshold. Several secondary outcomes also concerned IOP reductions or mean IOP at earlier or later visits, and one concerned the proportion reaching at least a 15% reduction at both Weeks 4 and 12. The registry record does not report results for that percentage outcome among the posted secondary results.

The primary population is narrower than everyone enrolled: the registry record describes a Per Protocol analysis with eligibility conditions tied to protocol violations, medication exposure, and Week 1 measurements. The posted primary arm summaries include 54 timolol participants and 53 latanoprost participants. These are analysis counts, not the overall enrollment count. The record does not report the reasons that participants were excluded from this primary analysis or the number excluded for each reason.

3. Trial Design

This was a Phase 3, prospective, randomized, parallel-group study with a planned treatment period of 12 weeks. The registry record marks the study completed and gives a total enrollment of 139. It lists the study start as 2008-07 and primary completion as 2009-11. The condition listed is glaucoma, and the population described in the brief summary is pediatric subjects aged 18 years or under.

Participants were assigned to latanoprost or timolol. The brief summary describes once-daily dosing for latanoprost 0.005% ophthalmic solution and twice-daily dosing for timolol 0.5%. The registry record does not give further arm-level dosing instructions in its arm descriptions. The study title describes it as double-masked, while the registry masking classification is quadruple. The record does not explain how masking was implemented or which study roles were masked, so those labels cannot be reconciled from the registry entries alone.

The primary outcome was reduction from baseline in mean IOP at Week 12 using last observation carried forward (LOCF). The primary population was Per Protocol rather than all enrolled participants. This analysis included participants with no major protocol violations who received at least 1 week of medication and had Week 1 IOP measurements. The primary result was reported as a least-squares mean with standard error for each treatment group, alongside a net mean difference and a 95% confidence interval.

The registry record lists 10 outcome measures and 5 analyses. Among the secondary outcomes are changes in mean IOP at Weeks 1, 4, and 12, mean IOP at those visits and baseline, and the percentage with at least a 15% reduction at both Weeks 4 and 12. The record does not report the original sample-size justification, the allocation ratio details beyond the two-arm comparison, interim analyses, or the full statistical analysis plan. A Phase 3 label and randomized design describe important features of the trial, but do not fill in those unreported design and analysis details.

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: Randomization was stratified by age, diagnosis (congenital glaucoma \[PCG\] or non-congenital glaucoma \[non-PCG\], and intraocular pressure \[IOP\]) of the study eye at baseline.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Per Protocol (PP) Population: participants with no major protocol violations who received at least 1 week of study medication and had at least Week 1 IOP measurements. LOCF.

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 Reduction From Baseline in Mean IOP at Week 12, Last Observation Carried Forward (LOCF), measured over Baseline, Week 12.

The primary endpoint is reduction from baseline in mean intraocular pressure at Week 12, measured in mmHg and analyzed using LOCF. The registry description defines reduction as baseline IOP minus Week 12 IOP, so a larger positive reduction corresponds to a larger decrease in the measured pressure. This is a continuous outcome: each participant contributes a pressure measurement, and the group result summarizes those measurements rather than counting participants who cross a threshold.

IOP was measured using one of three methods: Goldmann applanation tonometry, preferred when feasible; Perkins tonometry; or TonoPen. Measurements were taken twice. If the readings were within 2 mmHg of each other, their mean was recorded; otherwise a third measurement was taken and the median recorded. This procedure describes how the endpoint measurement was obtained and summarized at a time point.

A continuous endpoint can retain information about the size of changes across participants, whereas a binary threshold outcome distinguishes only whether each participant meets a specified cutoff. The primary endpoint therefore addresses the average magnitude of IOP reduction, not how many participants achieved a particular reduction. Its Week 12 timing aligns with the study's 12-week treatment period. The registry record does not report why this endpoint was selected over other IOP measures or how the endpoint was used in the original sample-size justification.

LOCF uses the last available observation to represent a later measurement when the Week 12 value is unavailable. The record identifies this approach for the primary outcome but does not give the number of participants whose Week 12 values were carried forward, the missingness pattern, or details of any sensitivity analyses. Those omissions matter because an imputed value is not the same as a measurement obtained at Week 12.

7. Results Posted in the Registry

• Latanoprost: mean 7.18 versus 5.72 on control; difference 1.46 (95% CI -0.79 to 3.71), standardized effect d = 0.25; z-test p = 0.202.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Reduction From Baseline in Mean IOP at Week 1

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Latanoprost: mean 6.70 versus 6.02 on control; difference 0.68 (95% CI -1.63 to 2.99), standardized effect d = 0.11; z-test p = 0.565.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Latanoprost and the control. The interval gives the range of effects that remain compatible with the data.

Reduction From Baseline in Mean IOP at Week 4

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Latanoprost: mean 6.99 versus 5.37 on control; difference 1.62 (95% CI -0.96 to 4.20), standardized effect d = 0.25; z-test p = 0.218.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Latanoprost and the control. The interval gives the range of effects that remain compatible with the data.

Reduction From Baseline in Mean IOP at Week 12 (Observed)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Latanoprost: mean 7.75 versus 6.96 on control; difference 0.79 (95% CI -1.07 to 2.65), standardized effect d = 0.18; z-test p = 0.404.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Latanoprost and the control. The interval gives the range of effects that remain compatible with the data.

Across the posted continuous secondary IOP outcomes, the estimated differences generally point toward a larger mean reduction with latanoprost than with timolol. The confidence intervals for the listed comparisons include zero, so those results do not provide a precise separation between the groups. The secondary outcomes span earlier assessments and an observed Week 12 assessment, and they are related to the primary question because they use IOP levels or reductions measured at other time points.

The overall pattern is broadly consistent in direction with the primary point estimate, but secondary results do not independently resolve the uncertainty in the primary comparison. Different time points can have different numbers of participants contributing measurements, and the registry counts vary across the posted analyses. Changes in analysis population, missing observations, or measurement timing can affect how closely a secondary estimate corresponds to the primary LOCF result. The registry record does not report enough detail to attribute the differences across visits to any particular cause.

The Week 12 observed secondary comparison is not interchangeable with the primary LOCF outcome. Its reported net difference is 0.408 mmHg, whereas the simple recalculation from the observed arm summaries gives 0.79 mmHg. Its reported confidence interval runs from -1.10 to 2.67. The registry record does not specify the analysis method, so the source of the difference between the posted comparison and the summary-based recalculation is not reported.

There are multiple secondary outcomes and time points. Unless analyses are arranged in a prespecified hierarchy or adjusted for multiplicity, interpreting each unadjusted comparison as confirmatory can inflate the chance of a false-positive finding across the set. The registry record does not report whether multiplicity was addressed. Secondary results therefore carry less weight than a prespecified primary analysis and are best treated as contextual evidence about consistency over time, not as a series of independent confirmations.

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

• Reduction From Baseline in Mean IOP at Week 1
• Reduction From Baseline in Mean IOP at Week 4
• Reduction From Baseline in Mean IOP at Week 12 (Observed)
• Mean IOP at Baseline
• Mean IOP at Week 1
• Mean IOP at Week 4
• Mean IOP at Week 12
• Percentage of Participants With Greater Than or Equal to (≥) 15% IOP Reduction From Baseline at Both Weeks 4 and 12

9. Statistical Interpretation

10. Statistical Methods Explained

The endpoint is continuous because it measures a quantity on a numeric scale: reduction in IOP, in mmHg. A group mean summarizes the average observed reduction, while a difference in means compares the group averages. In the primary results, the reported group means are least-squares means, a model-based form of group average. The registry record does not identify the model used to obtain them, so the adjustment variables and model structure are not reported in the registry record.

What does the mean difference represent?

The posted primary net mean difference is 1.46 mmHg, with the comparison defined as latanoprost relative to timolol. A positive value indicates a larger mean reduction in the latanoprost group for this subtraction order. The difference is a contrast between groups, not a change within one individual and not the percentage of participants who improved. It also does not show how much individual participants varied around their group average.

Standard error describes the estimated uncertainty of a group mean. The primary summaries report a standard error of 0.81 for each arm. It is not the same as a standard deviation: a standard deviation describes variation among participant measurements, while a standard error describes uncertainty in an estimated mean. The record does not report a standard deviation for the primary endpoint.

How should a confidence interval be read?

A confidence interval gives a range of values compatible with the data and the statistical procedure used, at the stated confidence level. The primary 95% interval runs from -0.81 to 3.74 mmHg. A range that crosses zero includes no difference between the group means. It also includes positive and negative differences, so the data do not sharply determine the direction or magnitude of the comparative effect. The interval is not a range containing 95% of individual participants' measurements, nor does it mean there is a 95% probability that the fixed true difference lies within this particular interval.

The simple recalculation uses the posted arm summaries to form an unadjusted comparison. It gives a mean difference of 1.46 and p = 0.202 for the primary outcome. Such a calculation does not reproduce a model-based analysis unless its assumptions and inputs match the original procedure. The posted primary estimate and interval are the sponsor's reported analysis; the recalculation is useful as a basic comparison of summaries, not as a replacement for that analysis. The discrepancy flag is false for the primary result.

What does LOCF do?

Last observation carried forward substitutes a participant's last available measurement for a missing later measurement. In this trial, the primary endpoint description specifies LOCF for Week 12. This approach makes an analysis possible when a later observation is absent, but it assumes that the last observed value is an adequate stand-in for the unobserved Week 12 value. That assumption may not hold for every participant. The record does not report how many Week 12 measurements were missing, the reasons for missingness, or sensitivity analyses using alternative approaches.

The population definition also affects interpretation. The Per Protocol population excluded participants with major protocol violations and required at least 1 week of study medication and Week 1 IOP measurements. These criteria focus the estimate on participants meeting those conditions, rather than preserving everyone assigned at random. The record does not report a primary intention-to-treat estimate. Exclusions after randomization can weaken the protection random assignment provides against differences between groups, depending on who is excluded and why; those details are not reported.

Why can over-dispersion matter?

For a continuous endpoint, greater variability among participants makes a group mean less precise when the sample size is held fixed. Over-dispersion is a general term for variability greater than a model assumes. If a model understates variability, its standard errors can be too small and its confidence intervals too narrow. This is a general statistical consideration, not a finding reported for this trial. The registry does not provide enough information to assess over-dispersion or the model assumptions used for the least-squares means.

Likewise, the use of different tonometry methods could be a source of measurement variation in general, although the registry lists the measurement procedures and does not report a comparison of method-specific results. It would not be justified to infer that method choice biased the treatment comparison. The record also does not report whether measurement method differed by group or how method use was distributed.

Why do secondary comparisons need caution?

Multiple endpoints and visit-specific comparisons create opportunities for apparently notable results even when each individual comparison is evaluated at its own significance threshold. Multiplicity adjustment can control the chance of false-positive findings across a family of tests, but the registry record does not report whether any adjustment was planned or used. Secondary comparisons should therefore be interpreted as supportive and exploratory unless a prespecified analysis hierarchy and multiplicity procedure are documented. The observed secondary results do not replace the stated primary endpoint.

Finally, a p-value is a measure of how unusual data at least as incompatible with a null hypothesis would be under a specified statistical model. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. The recalculated p-value is unadjusted and based on summary statistics. It should be read alongside the estimated difference and interval, not as a standalone verdict. The registry record does not report the sponsor's p-value for the primary comparison.

11. Confidence Intervals and Precision

The primary 95% confidence interval for the net mean difference extends from -0.81 to 3.74 mmHg. The estimate is 1.46 mmHg. The interval is wide enough to include no difference and differences in either direction. Its lower endpoint is below zero, while its upper endpoint is above zero; therefore, the primary estimate does not precisely settle the comparative direction. The interval gives more information than the point estimate alone because it conveys the range of values compatible with the analysis.

Precision is influenced by sample size, variability, and the analysis method. The primary Per Protocol summaries include 54 timolol participants and 53 latanoprost participants, with standard errors of 0.81 in both groups. Those counts are the analysis sample sizes, not the total enrollment of 139. The registry record does not report the standard deviation for the primary outcome or enough detail to reconstruct the sponsor's model-based interval independently.

The simple unadjusted recalculation also estimates a mean difference of 1.46, with p = 0.202. That p-value is not a confidence interval and does not convey the range of plausible effect sizes. The sponsor's reported interval is the appropriate posted interval to interpret for the primary analysis. Since it crosses zero, the data remain compatible with no difference under that analysis, while also allowing a positive difference as large as the interval's upper endpoint.

Confidence intervals do not address every uncertainty. They reflect sampling uncertainty under the analysis assumptions; they do not automatically account for missing-data assumptions, outcome measurement variation, selection into the Per Protocol population, or unreported protocol decisions. The primary endpoint uses LOCF, and the registry does not report the number of carried-forward measurements or alternative missing-data analyses. Those sources of uncertainty cannot be quantified from the posted primary summary. The interval should therefore be read as precision for the stated analysis, not as a complete measure of all uncertainty about the trial's findings.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrations derived from the observed effect, not the trial's original design calculation. The assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set in turn to 100%, 75%, or 50% of the observed effect. The registry record does not report the original sample-size justification or its assumptions.

The 80% power column means the estimated number of participants per arm that, under those planning assumptions, would give an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column uses a higher target chance of detection and therefore calls for more participants per arm for the same effect. For the primary endpoint, planning for 100% of the observed effect gives 259 participants per arm for 80% power and 347 per arm for 90% power. Planning for 75% gives 460 per arm and 616 per arm, respectively; planning for 50% gives 1,035 per arm and 1,385 per arm.

Smaller effects are harder to distinguish from random variation, so the planning illustration requires more participants per arm as the target effect decreases. These figures are not a claim about how well the completed trial was powered for its original objective, and they do not show the probability that the trial's result is true. This is not observed power. It is a forward-looking planning illustration based on the observed estimate, which may itself be uncertain. The trial's own design assumptions, including its intended effect and any allowance for attrition, are not reported in the registry record.

13. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = 1.460; pooled standard deviation σ = 5.925.
2. n = 2 × (1.96 + 0.842)² × 5.925² / 1.460² = 258.5, rounded up to 259 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 <- 1.46                      # observed difference in means
sigma <- 5.92490289251345                     # 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(716859)
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 = 259 (80% power), 347 (90% power)
 75% of observed effect: n per arm = 460 (80% power), 616 (90% power)
 50% of observed effect: n per arm = 1,035 (80% power), 1,385 (90% power)
power.t.test: 259.5 
Simulated power at n = 259 per arm: 0.800 (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.

14. Analyses Reported by the Sponsor

For the primary outcome, the sponsor-reported net mean difference is 1.46 mmHg, with a 95% confidence interval from -0.81 to 3.74. The simple unadjusted recalculation from the posted arm summaries also gives a mean difference of 1.46. The recalculated p-value is 0.202, but the registry does not report a sponsor p-value for this comparison. The posted confidence interval should be used to describe the sponsor's primary result; the recalculation is a check based on summary values, not the trial's full analysis.

The registry labels the primary reported analysis as a legacy non-inferiority or equivalence analysis, but it does not report the margin, the justification for that margin, or the statistical method used. A confidence interval that includes zero does not by itself establish non-inferiority or equivalence. Such conclusions depend on a prespecified margin and a suitable analysis. Because those details are not reported, the legacy label alone is insufficient to determine whether either criterion was met.

The discrepancy flag for the primary result is false, and the sponsor estimate agrees with the unadjusted recalculation at the displayed precision. The registry does not give enough methodological detail to explain every aspect of the reported model-based analysis. Where a posted sponsor analysis and a simple recalculation differ, the sponsor's own analysis takes precedence when the discrepancy flag is true; for this primary result, that condition does not apply.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists multiple outcomes and analyses, including repeated IOP measurements at several visits and a percentage-based outcome requiring a specified reduction at two visits. When many outcomes or contrasts are examined, the chance of at least one apparently unusual result can increase if each is judged using an unadjusted threshold. Statistical procedures such as family-wise error control or a prespecified testing hierarchy can address this issue, but the registry record does not report whether such a procedure was planned or used.

The posted secondary comparisons should consequently be interpreted with caution. Their confidence intervals and recalculated p-values are not adjusted for a family of tests. The primary result remains the central comparison, but the record does not provide the complete analysis plan needed to determine which outcomes were confirmatory, how testing order was specified, or whether outcome reporting followed the original plan. The registry does not report the analysis plan or any interim analyses.

Missing data are also relevant because the primary endpoint uses LOCF. The registry record does not report how many primary Week 12 measurements were unavailable, why they were unavailable, or whether the pattern differed by treatment group. It also does not describe sensitivity analyses or alternative missing-data methods. LOCF replaces a missing later measurement with an earlier one; the resulting estimate depends on whether that substitution reasonably represents the unobserved Week 12 value. Without the missingness details, the robustness of the primary result to that assumption cannot be assessed.

The primary analysis population was Per Protocol, with participants required to have no major protocol violations, at least 1 week of study medication, and Week 1 IOP measurements. The registry record does not report a matching intention-to-treat result or a participant flow accounting for exclusions from the primary population. A Per Protocol analysis can be useful for describing outcomes among participants meeting those criteria, but exclusions after randomization may affect comparability. The size and direction of any such effect cannot be inferred from the posted results.

The registry also does not report the original sample-size justification, the protocol's detailed statistical analysis plan, or the rationale for the primary population. These missing details limit assessment of whether the analysis followed prespecified choices and whether the study was designed around a clinically meaningful difference. They are uncertainties about reporting and analysis, not evidence that a particular method was or was not used.

16. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: lens dislocation, study eye (1.5%), bronchopneumonia (1.5%), gastroenteritis (1.5%), pneumonia viral (1.5%), developmental glaucoma, both eyes (1.4%).

The registry reports serious adverse events in 7 of 69 participants at risk in the timolol group and 2 of 68 at risk in the latanoprost group. Other adverse events are reported in 20 of 69 and 12 of 68 participants, respectively. These are counts of participants with events and the reported group denominators; they are not the primary efficacy analysis counts, which come from the Per Protocol population.

The listed serious adverse-event terms include lens dislocation, study eye; bronchopneumonia; gastroenteritis; and pneumonia viral, each reported as 1.5%. Developmental glaucoma, both eyes, and angle closure glaucoma, both eyes, are each listed as 1.4%. The registry record does not give further detail about the timing, severity, relatedness, or participant-level overlap of these terms. The displayed terms should not be read as a complete account of all event types or as evidence that a treatment caused an event.

Deaths are not reported in the registry record. The event counts describe what was posted during the trial; without additional information on event definitions, follow-up, and attribution, they do not establish a comparative safety profile. The study's 12-week duration also limits what these figures can say about longer-term safety.

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

The primary estimate describes a numerically larger mean IOP reduction in the latanoprost group than in the timolol group at Week 12: 7.18 mmHg compared with 5.72 mmHg. The reported net difference is 1.46 mmHg, with a 95% confidence interval from -0.81 to 3.74. Because zero lies within that interval, the posted primary analysis does not precisely distinguish a difference from no difference. The interval also leaves uncertainty about the size and direction of the comparative effect.

This result does not establish that the treatments are equivalent or non-inferior. The registry classifies the primary analysis with a legacy non-inferiority or equivalence label, but it does not report a margin or its justification. Those are essential to interpret such a design claim. Nor does a confidence interval crossing zero prove that the groups are the same; it indicates that the data and analysis remain compatible with a range of differences that includes zero.

The result applies to the stated primary analysis: participants in the Per Protocol population, with the Week 12 outcome handled using LOCF. The record does not report a primary intention-to-treat analysis, how many values were carried forward, or sensitivity analyses. It is therefore not possible to infer that the same estimate would be obtained under other populations or missing-data methods.

The findings concern a continuous IOP measure over the trial's 12-week period. They do not establish effects on outcomes that were not reported, longer-term outcomes, or the proportion of participants who achieved a clinically meaningful response. The record does not report a threshold for a clinically important between-group difference. It also does not support medical advice or a treatment choice for an individual.

Finally, the adverse-event counts are descriptive and have no reported attribution details. They should not be combined with the efficacy estimate to make a broad claim about overall benefit or harm. The most defensible interpretation stays close to the posted evidence: the primary group summaries favor a greater mean reduction with latanoprost numerically, while the interval leaves substantial uncertainty about the between-group difference.

18. Why This Trial Matters Statistically

This trial illustrates why a treatment comparison should be read through both its point estimate and its uncertainty. A positive difference in group means may look directionally clear, but the confidence interval indicates how imprecisely the comparison is estimated. Here, the primary interval includes zero and spans values on both sides of it. That limits what can be concluded from the point estimate alone.

It also shows why an endpoint definition matters. The primary result concerns mean reduction from baseline at Week 12, not a threshold response or a long-term outcome. The registry's measurement description, including alternative tonometry methods and repeat-reading rules, defines how the continuous endpoint was obtained. The use of LOCF and a Per Protocol population further defines the estimand being summarized; the result is not automatically a statement about every randomized participant.

The record highlights the distinction between a sponsor-reported comparison and a simple recalculation. For the primary outcome, the reported estimate and summary-based recalculation agree at the displayed precision, while the recalculation supplies an unadjusted p-value absent from the sponsor report. A calculation from aggregate summaries cannot replace an analysis whose model and assumptions are not fully reported. This distinction is especially important for secondary outcomes where the posted comparison and recalculated summary contrast differ.

Planning illustrations based on an observed effect also demonstrate how strongly required sample size depends on the effect one aims to detect. Smaller target effects require more participants per arm under the stated assumptions. Such illustrations answer a hypothetical planning question; they do not recover the trial's original design rationale or provide observed power.

More broadly, the registry record shows why missing statistical details affect interpretation. The primary analysis label references non-inferiority or equivalence, yet the margin is not reported. Multiplicity procedures, missing-data sensitivity analyses, and original sample-size assumptions are also not reported. Careful statistical reading distinguishes what the estimates show from what cannot be established without those details.

19. Limitations

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

The primary analysis is limited to the Per Protocol population rather than a reported all-randomized analysis. Its eligibility conditions include no major protocol violations, at least 1 week of study medication, and Week 1 IOP measurements. The registry record does not report how many participants failed these conditions or why. This limits assessment of selection after randomization and of how the primary population relates to all enrolled participants.

The primary outcome uses LOCF, but the record does not report the amount or pattern of missing Week 12 data, reasons for missingness, or sensitivity analyses. Since LOCF substitutes an earlier measurement for a missing later one, its validity depends on how well that value represents Week 12. The impact of this assumption cannot be quantified from the posted summary.

There is a mismatch in masking terminology: the title says double-masked, while the registry's masking classification is quadruple. The record does not describe which roles were masked or how masking was maintained. This is a reporting limitation; it does not establish that masking failed or that it was implemented in a particular way.

The record does not report the model underlying the least-squares means, the original sample-size justification, the detailed statistical analysis plan, an interim-analysis plan, or a non-inferiority/equivalence margin. It also does not report whether multiplicity adjustment was used. These omissions limit evaluation of the prespecification, model assumptions, and strength of confirmatory claims.

For the observed Week 12 secondary outcome, the sponsor-reported estimate is 0.408 mmHg, while the simple recalculation from the group summaries is 0.79 mmHg. The registry does not identify the analysis method needed to explain the difference. More generally, the trial lasted 12 weeks, so the posted results do not characterize longer-term outcomes. These limitations constrain interpretation; they do not in themselves determine the direction of any bias.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Phase 3 Prospective, Randomized, Double-Masked, 12-Week, Parallel Group Study In Pediatric Subjects With Glaucoma. — NCT00716859.
• Registry record last updated 2021-02-03; status last verified 2011-01.
• Results first posted 2010-12-02.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

The registry describes a completed Phase 3, prospective, randomized, parallel-group study comparing latanoprost 0.005% once daily with timolol 0.5% twice daily in pediatric subjects aged 18 years or under with glaucoma. Total enrollment was 139, and the planned treatment period was 12 weeks. The study title says double-masked, while the registry classification says quadruple; the masking details are not reported.

The primary endpoint was reduction from baseline in mean IOP at Week 12 using LOCF in the Per Protocol population. The reported mean reductions were 5.72 mmHg for timolol and 7.18 mmHg for latanoprost. The net difference was 1.46 mmHg, with a 95% confidence interval from -0.81 to 3.74. This interval includes zero, so the primary comparison is uncertain about the presence and size of a between-group difference.

The registry does not report the original sample-size justification, analysis plan details, non-inferiority or equivalence margin, or missing-data sensitivity analyses. Serious adverse events were reported for 7 of 69 timolol participants and 2 of 68 latanoprost participants; deaths are not reported in the registry record.