Study of Fixed-Dose Combinations of Trabodenoson and Latanoprost in Adults With OHT or POAG: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Fixed-Dose Combinations of Trabodenoson and Latanoprost in Adults With OHT or POAG (NCT02829996), a randomized phase 2 study. The primary outcome is Mean Intraocular Pressure (IOP).

2. Clinical Question

3. Trial Design

The registry record describes a phase II, randomized, parallel-group study sponsored by Inotek Pharmaceuticals Corporation. It lists quadruple masking, while the brief study summary describes the study as double-masked. The record does not explain the difference between those masking labels. The study is marked completed, with enrollment of 201 and five treatment arms: three trabodenoson–latanoprost combinations and two latanoprost-only groups.

Before treatment, the study summary describes screening and washout of prohibited medications when washout was needed, including routine glaucoma medications. During a placebo run-in, placebo was applied to both eyes twice daily. The treatment period lasted 8 weeks, and study drops were administered in the morning and evening. Each participant was assigned 4 weeks of morning and 4 weeks of evening dosing in a masked manner; placebo was used alongside active product to maintain masking. The summary then describes an observation period of approximately 7 days without study drops.

The stated study purpose was to assess the overall benefit/risk profile of binocular topical application of different trabodenoson doses combined with latanoprost at different concentrations. The primary endpoint is identified as intent to treat, but the registry record does not report the number of participants included in that analysis or explain how the analysis population was constructed. The original sample-size justification, planned statistical model, and planned analysis details are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat

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 Mean Intraocular Pressure (IOP), measured over Two Months.

The primary endpoint is mean daily change from diurnal baseline in intraocular pressure (IOP), measured in mm of mercury at two months. IOP is a continuous measurement, so the endpoint is expressed as a group mean rather than as a count of participants meeting a threshold. A change from baseline describes how the measurement at follow-up differs from the baseline measurement; negative mean changes in the posted results indicate values below baseline on average.

General statistical context: a mean condenses the measurements in a group into a measure of its central level. The standard error (SE) describes the estimated sampling uncertainty of that mean. It is not the standard deviation of individual measurements; it is about the precision of the estimated group mean. The registry reports each arm’s mean change and SE, but does not provide arm-level confidence limits.

Using a common continuous endpoint across treatment groups makes it possible to compare average changes in IOP. The reported endpoint is specifically a daily change from diurnal baseline, rather than a single named time-of-day value. The record does not provide further detail about how the daily value was derived from measurements across the day, or how missing endpoint measurements were handled. The endpoint is listed for the intent-to-treat population and a two-month timeframe.

7. Results Posted in the Registry

• Trabodenoson 3.0% / Latanoprost 0.005% QD: mean -6.80 versus -6.40 on control; difference -0.40 (95% CI -1.85 to 1.05), standardized effect d = -0.12; z-test p = 0.588.
• Trabodenoson 6.0% / Latanoprost 0.0025% QD: mean -6.40 versus -6.40 on control; difference 0.00 (95% CI -1.45 to 1.45), standardized effect d = 0.00; z-test p = 1.000.
• Latanoprost 0.005% QD: mean -7.10 versus -6.40 on control; difference -0.70 (95% CI -2.00 to 0.60), standardized effect d = -0.24; z-test p = 0.292.
• Latanoprost 0.0025% QD: mean -5.90 versus -6.40 on control; difference 0.50 (95% CI -0.90 to 1.90), standardized effect d = 0.15; z-test p = 0.484.

8. Statistical Interpretation

9. Statistical Methods Explained

What do the arm means describe?

The primary outcome is continuous: daily change from diurnal baseline in IOP. For each treatment group, the registry reports a mean and a standard error. The mean describes the group’s average change; the SE describes the precision of that estimated average. The SE is not a description of how much individual participants’ measurements vary. The record does not report the standard deviation of the individual endpoint values directly.

The arm means are -6.4 for trabodenoson 6.0% / latanoprost 0.005%, -6.8 for trabodenoson 3.0% / latanoprost 0.005%, -6.4 for trabodenoson 6.0% / latanoprost 0.003%, -7.1 for latanoprost 0.005%, and -5.9 for latanoprost 0.003%. Each value is in mm of mercury, and the endpoint timeframe is two months. General statistical context: an average is a useful summary for a continuous outcome, but it does not reveal the full distribution of participant-level measurements or whether response patterns differed across participants.

How are the group differences formed?

The posted comparison values are unadjusted recalculations from the arm summary statistics. They express the mean in a comparison group relative to the mean in the reference group, trabodenoson 6.0% / latanoprost 0.005%. For example, the comparison for trabodenoson 3.0% / latanoprost 0.005% is -0.4, corresponding to the posted group means -6.8 and -6.4. The comparison for trabodenoson 6.0% / latanoprost 0.003% is 0.0, matching the two posted means of -6.4. These are descriptive differences in group averages, not adjusted estimates from a model reported by the sponsor.

The registry record reports no sponsor analysis entries for the outcomes. The calculated comparisons should therefore be read as unadjusted comparisons based on the summaries, rather than as a substitute for an unreported prespecified analysis. The registry’s discrepancy flag is false; there is no posted discrepancy identified between these recalculations and a sponsor result. If a sponsor analysis were available and differed from a recalculation, the sponsor’s own analysis would take precedence.

What do confidence intervals and p-values contribute?

A confidence interval gives a range of values compatible with the estimate under the calculation’s assumptions and indicates how precisely the difference is estimated. It is not a range containing a known proportion of individual participant outcomes, nor does it give the probability that the true difference lies inside that particular interval. All four posted comparison intervals include zero. This means the summaries do not exclude zero difference at the interval’s stated confidence level; it does not establish that the groups are equivalent.

The accompanying p-values are unadjusted: 0.588, 1.0, 0.292, and 0.484 for the three combination/latanoprost comparisons and the latanoprost 0.003% comparison, respectively. A p-value describes how unusual the observed result, or one more extreme, would be under a specified null model. It is not the probability that the null hypothesis is true, and it does not describe clinical importance. The registry record does not report a plan for addressing the multiple comparisons.

Why can over-dispersion matter?

General statistical context: over-dispersion occurs when outcome variation is greater than a model assumes. If a model understates variability, standard errors can be too small and confidence intervals too narrow, making evidence appear more precise than it is. This concept is especially relevant to count or event outcomes and to models whose variance assumptions may not fit the data. The primary endpoint in this study is continuous, and the posted summaries use means and standard errors. The registry record does not report a model-based analysis or an assessment of over-dispersion, so it is not possible to say whether such an issue was assessed in the sponsor’s analysis.

What is not described?

The registry record does not report the original statistical analysis plan, the method used to calculate the sponsor’s primary analysis, covariate adjustment, handling of repeated measurements, missing-data methods, or the original sample-size justification. It also does not report whether there were interim analyses. These details matter because analytic choices can affect the estimate and its uncertainty. The posted summaries support descriptive interpretation and simple unadjusted recalculation, but they do not supply the detail needed to reconstruct a full prespecified analysis.

10. Confidence Intervals and Precision

Precision is best assessed by considering the confidence interval alongside the estimated difference, rather than by looking only at the point estimate or p-value. The recalculated unadjusted intervals are -1.846 to 1.046 for trabodenoson 3.0% / latanoprost 0.005%, -1.446 to 1.446 for trabodenoson 6.0% / latanoprost 0.003%, -2.003 to 0.603 for latanoprost 0.005%, and -0.901 to 1.901 for latanoprost 0.003%, each compared with trabodenoson 6.0% / latanoprost 0.005%. The intervals all include zero.

These intervals show that the estimated differences are not precise enough to rule out zero or all differences in either direction within the reported ranges. For instance, the interval for latanoprost 0.005% extends from -2.003 to 0.603; the interval for latanoprost 0.003% extends from -0.901 to 1.901. The posted estimates should not be read as demonstrating that the groups have identical average changes. Demonstrating equivalence would require a prespecified equivalence margin and an appropriate analysis, neither of which is reported in the registry record.

The arm means have reported standard errors between 0.46 and 0.56, and endpoint group counts between 39 and 42. The registry does not give arm-level confidence intervals for the means, only their SEs. The intervals for the differences are recalculated from posted summary statistics and are unadjusted. They do not account for any multiplicity adjustment, and no such adjustment plan is reported. Greater precision would generally require less uncertainty in the estimated contrast; general statistical context is that this can depend on group size, outcome variability, and study design. The registry record does not provide sufficient detail to determine how the sponsor’s planned analysis handled those factors.

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 sample-size figures are planning illustrations based on the observed effect, not the original trial’s design calculation. The listed figures are participants per arm, not total participants or total events. They apply to the comparisons for which a planning illustration is posted. The 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 the stated fraction of the observed effect. The original sample-size assumptions used to design the trial are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the assumed effect under those planning assumptions, if that effect and the model assumptions are correct. The 90% power column gives the corresponding per-arm planning number for a 90% chance. These are design calculations, not descriptions of the probability that this completed study found a true effect. This is not observed power.

Rows planning for 100% of the observed effect use the full observed difference as the assumed effect. Rows planning for 75% or 50% of the observed effect assume progressively smaller differences. Smaller effects are harder to distinguish from random variation, so the illustrations require more participants per arm. For example, for the trabodenoson 3.0% / latanoprost 0.005% comparison, the posted planning values rise from 1,041 participants per arm for 80% power at 100% of the observed effect to 1,851 at 75% and 4,164 at 50%. The corresponding values for 90% power are 1,394, 2,477, and 5,574. These figures are conditional illustrations, not recommendations or a retrospective judgment of the study design.

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 = -0.400; pooled standard deviation σ = 3.257.
2. n = 2 × (1.96 + 0.842)² × 3.257² / -0.400² = 1,040.8, rounded up to 1,041 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 <- -0.399999999999999                      # observed difference in means
sigma <- 3.25699247773157                     # 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(2829996)
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 = 1,041 (80% power), 1,394 (90% power)
 75% of observed effect: n per arm = 1,851 (80% power), 2,477 (90% power)
 50% of observed effect: n per arm = 4,164 (80% power), 5,574 (90% power)
power.t.test: 1041.7 
Simulated power at n = 1,041 per arm: 0.798 (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 posted secondary outcomes. It also lists zero outcome analyses. This means the record does not provide a sponsor analysis result for the primary endpoint in the analysis entries; the comparisons discussed in the posted results are unadjusted recalculations from arm-level summaries. The record does not describe the sponsor’s analysis population beyond identifying the endpoint population as intent to treat.

Multiplicity arises when several hypotheses or comparisons are examined. If each is assessed separately without adjustment, the chance of at least one result appearing statistically notable by chance can increase. Here, four unadjusted comparisons are reported against the reference group. Whether the original study protocol designated these comparisons, how it prioritized them, or whether any adjustment was planned is not reported in the registry record. The unadjusted p-values should therefore not be interpreted as if a multiplicity procedure had been applied.

The record also does not report how missing endpoint measurements were handled, whether all randomized participants contributed endpoint data, or whether any sensitivity analyses were performed. It does not state whether analysis used a repeated-measures model or adjusted for baseline values. Such information can matter for interpretation of a continuous outcome collected during a treatment period. The designation intent to treat identifies the stated analysis population label, but by itself does not establish how missing observations or deviations from assigned treatment were handled.

These reporting gaps do not show that an analysis or a plan was absent; they show only that those details are not reported in the registry record. The available results support unadjusted descriptive comparisons, while questions about multiplicity control, missing-data assumptions, and the prespecified statistical approach remain unresolved from the posted record.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 3% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse event counts separately for each treatment group. The serious-event count is 0 in each of the five groups: trabodenoson 6.0% / latanoprost 0.005%, trabodenoson 3.0% / latanoprost 0.005%, trabodenoson 6.0% / latanoprost 0.003%, latanoprost 0.005%, and latanoprost 0.003%. The deaths count is also 0 in every group. The reported numbers at risk for the serious-event summaries are 39, 39, 40, 41, and 42, respectively.

Zero reported serious events means none were recorded in these groups during the registry’s reported safety period; it does not establish that the risk is zero. The registry record does not provide serious-event terms, because none are listed, or detailed information about event ascertainment and follow-up duration for these summaries. It reports other adverse events affecting 9, 5, 6, 3, and 5 participants in the same arm order. Those counts are distinct from the serious-event counts and should not be treated as a measure of severity or as a comparison adjusted for differences in exposure. These descriptive counts alone do not establish a causal safety difference between groups.

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

The posted results describe average changes in IOP over two months in five randomized treatment groups. The arm means are relatively close in the displayed summaries, and the unadjusted recalculated differences compared with trabodenoson 6.0% / latanoprost 0.005% range from -0.7 to 0.5. The associated confidence intervals include zero and also include values on either side of zero. The results therefore leave uncertainty about the direction and size of the differences between groups.

These estimates do not establish equivalence or show that the regimens have the same effect. Nor do the unadjusted p-values establish that a difference is absent. A claim of equivalence would require an explicit margin and a design and analysis suited to that question; the registry record does not report those details. The registry also does not post sponsor analyses for the outcome, so the comparisons are recalculated unadjusted contrasts based on the available summary statistics, not adjusted treatment effects from a reported sponsor model.

The figures describe an average continuous outcome, not the experience of every participant. They do not show how responses varied among participants, and the mean alone does not determine the distribution of individual changes. The outcome is measured at two months; the record does not report longer-term IOP results. Safety summaries report no serious adverse events or deaths in the listed groups, but finite observation and the absence of detailed safety analyses limit what those counts can establish.

General statistical context: a statistically nonsignificant result is not proof of no effect, while a difference in sample means is not by itself proof of a causal or clinically important advantage. Interpretation depends on the planned comparisons, analysis method, outcome variability, missing data, and context for judging meaningful differences. Several of those details are not reported in the registry record. The findings should therefore be read as posted descriptive outcomes with substantial uncertainty around comparative estimates, not as definitive evidence for or against any regimen.

16. Why This Trial Matters Statistically

This study illustrates how a randomized multi-arm design can compare several formulations against active comparators while measuring a continuous clinical outcome. The common endpoint, mean change from diurnal baseline in IOP, permits direct description of each arm’s average change and unadjusted comparison with a reference group. Randomization is relevant because it is intended to support fair group comparisons, though the registry record does not provide enough information to evaluate balance or the details of the analysis.

Statistically, the central feature is not simply that the group means differ. The uncertainty around each contrast matters. All four recalculated confidence intervals include zero, and each is broad enough to cover differences in either direction. The point estimates alone could encourage an overly definite reading; the intervals make clear that the posted summaries do not tightly determine the comparative effects.

The study also highlights the distinction between an observed result and a planning calculation. The sample-size illustrations use observed effects and explicit assumptions to show how the number of participants per arm would change under different target effects and power levels. They are not observed power and do not reveal the original rationale for enrollment. Because smaller target effects require larger planning numbers, these calculations also show why a study can yield imprecise comparisons even when average changes are measured in each arm.

Finally, the record’s limited reporting of analysis methods, missing-data handling, and multiplicity makes transparent separation of reported summaries from recalculation particularly important. A careful reading can identify what the data describe while keeping unreported protocol details and sponsor analyses distinct from the calculations available in the registry record.

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 registry record has important limits for interpretation of this trial’s comparative results. It reports arm means, standard errors, endpoint group counts, and unadjusted recalculated contrasts, but no sponsor outcome analyses. The original statistical analysis plan, covariate adjustment, repeated-measures approach, handling of missing values, and sensitivity analyses are not reported in the registry record. The endpoint is labeled intent to treat, but the record does not explain how participants with missing data or treatment deviations were handled.

There are multiple treatment groups and four reported unadjusted comparisons against one reference group. A multiplicity plan is not reported. The registry also does not describe a prespecified equivalence or noninferiority margin, so the confidence intervals cannot be interpreted as establishing equivalence. The original sample-size justification and planned detectable effect are not reported; the available sample-size figures are illustrations based on observed effects and should not be used to infer the study’s original design rationale.

The primary endpoint is reported at two months, and the record does not provide longer-term outcome summaries. Arm-level standard errors convey uncertainty about means, but the record does not report the distribution of individual responses. The registry’s masking labels differ between its design information and brief summary, which describes the study as double-masked; an explanation of the difference is not reported. Safety information is limited to counts of participants affected by other adverse events and serious adverse events, with no serious events or deaths listed. Further detail on safety definitions, event ascertainment, and follow-up for those summaries is not reported.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Study of Fixed-Dose Combinations of Trabodenoson and Latanoprost in Adults With OHT or POAG — NCT02829996.
• Registry record last updated 2018-02-15; status last verified 2018-01.
• Results first posted 2018-02-15.
• 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

NCT02829996 is a completed phase II randomized parallel-group study of fixed-dose trabodenoson and latanoprost combinations in adults with ocular hypertension or primary open-angle glaucoma. The registry reports enrollment of 201 and five arms, including three combination regimens and two latanoprost-only comparators. The primary endpoint is mean daily change from diurnal baseline in IOP at two months, for the intent-to-treat population.

Posted arm means range from -5.9 to -7.1 mm of mercury. Four unadjusted recalculated comparisons with trabodenoson 6.0% / latanoprost 0.005% have confidence intervals that include zero. The registry reports no serious adverse events or deaths in any listed arm. It does not report sponsor outcome analyses, a multiplicity plan, missing-data methods, or the original sample-size justification. The comparative results are therefore best read as imprecise descriptive estimates with important analysis details not reported in the registry record.