Phase IIa, Randomized, Double Blinded, Placebo Controlled, Dose Finding Study for TLC599 in OA Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Phase IIa, Randomized, Double Blinded, Placebo Controlled, Dose Finding Study for TLC599 in OA Patients (NCT03005873), a randomized phase 2 study. The primary outcome is Change From Baseline in WOMAC Pain Subscale at Week 12.

2. Clinical Question

3. Trial Design

This was a Phase IIa, randomized, double-blind, placebo-controlled, parallel-group dose-finding study of TLC599 in people with osteoarthritis of the knee. The official title describes single-dose administration. The registry lists three groups: TLC599 LD, TLC599 HD, and placebo. The LD formulation is described as 12 mg DSP with 100 µmol PL in 1.0 mL; the HD formulation as 18 mg DSP with 150 µmol PL in 1.5 mL; and placebo as 1.5 mL normal saline.

The registry lists an enrollment of 76. For the primary outcome, the listed group sizes are 26 in the LD group, 24 in the HD group, and 25 in the placebo group. Enrollment and the group counts are distinct registry entries and should not be treated as interchangeable descriptions of the outcome analysis population.

The study started on 2017-03-16, had primary completion on 2018-07-04, and is recorded as completed. The sponsor was Taiwan Liposome Company. Randomization, double masking, and a parallel model describe important structural features, but the registry record does not report the allocation sequence method, concealment procedures, or who specifically was masked. It also does not report the original design assumptions or sample-size justification. Those details cannot be reconstructed from the outcome results.

The primary outcome was change from baseline in WOMAC Pain subscale score at Week 12. The posted group summaries are least-squares means with standard errors. Missing data were imputed, but the imputation method and the analysis population are not reported in the registry record. The record lists one outcome measure and no posted secondary outcomes. It does not provide a separate sponsor-reported contrast analysis in the results.

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:

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: One patient from the Placebo group was enrolled but withdrawn from the analysis population due to ineligibility.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in WOMAC Pain Subscale at Week 12, measured over Baseline, Week 12.

The primary endpoint was change from baseline in the WOMAC Pain subscale at Week 12. WOMAC is the Western Ontario and McMaster Universities Osteoarthritis Index. Its pain subscale has five items, each ranging from 0 to 4; the subscale total is normalized and expressed on a 0 to 4 scale. Higher scores represent higher pain and a worse outcome, while lower scores represent lower pain and a better outcome.

A change score expresses how the measured outcome at follow-up differs from its baseline value. For this endpoint, negative changes indicate decreases in pain score. The posted least-squares mean changes were -0.83 for TLC599 LD, -0.64 for TLC599 HD, and -0.47 for placebo. Those summaries describe change on the normalized score scale, not the proportion of participants who improved.

WOMAC pain is used to quantify a patient-reported symptom on a defined scale, allowing group-level changes to be compared across randomized arms. The registry record does not report a threshold for interpreting a change as clinically important, nor does it describe the precise normalization procedure beyond stating that the score is expressed on a 0 to 4 scale. Missing data were imputed; the method used is not reported in the registry record.

7. Results Posted in the Registry

• TLC599 LD Group: mean -0.83 versus -0.47 on control; difference -0.36 (95% CI -0.61 to -0.11), standardized effect d = -0.78; z-test p = 0.005.
• TLC599 HD Group: mean -0.64 versus -0.47 on control; difference -0.17 (95% CI -0.43 to 0.09), standardized effect d = -0.37; z-test p = 0.196.

8. Statistical Interpretation

9. Statistical Methods Explained

The endpoint results are reported as least-squares means with standard errors. A least-squares mean is a model-based estimate of a group mean; it may account for terms in a statistical model, unlike a simple arithmetic average. The registry record does not describe the model specification, its covariates, or the precise adjustment used to obtain the estimates, so the least-squares label alone does not identify how the estimates were produced.

What does a standard error describe?

A standard error describes the estimated sampling uncertainty of a statistic, such as a mean. It is not the spread of individual participant scores. A small standard error indicates that the estimated group mean is relatively precise under the analysis assumptions; it does not mean that participants had similar scores. The posted standard errors are 0.09 for LD, 0.094 for HD, and 0.092 for placebo. These values accompany the group means and should be read as uncertainty around those estimated means.

A standard deviation, by contrast, describes dispersion among individual observations. The registry record does not post standard deviations for the primary outcome. The pooled standard deviations used for the simple recalculations are approximations derived from the posted standard errors and group sizes. They are not direct participant-level summaries, and the recalculated effect-size quantities therefore depend on assumptions beyond the posted least-squares means.

How are the group contrasts interpreted?

A mean difference compares the average change in one group with the average change in another. The recalculated contrasts use placebo as the reference and subtract the placebo estimate from each TLC599 estimate. A negative difference means the active-group mean change is more negative, or a larger decrease in score, than the placebo mean change. This direction is consistent with the scale definition that lower pain scores are better, but it does not by itself establish a causal or clinically important effect.

The recalculated intervals and p-values are simple, unadjusted comparisons based on posted summary statistics. They do not necessarily reproduce the trial's intended analysis. The record does not report a sponsor-led contrast model, its covariate adjustment, or a multiplicity procedure. If a registry record contained a conflicting sponsor analysis, that analysis would take precedence; the recorded discrepancy status is false.

What does an interval add?

A confidence interval gives a range of values reasonably compatible with the data and the assumptions used to calculate it. It conveys uncertainty around an estimate rather than only naming a single value. The LD interval is entirely below zero, while the HD interval spans zero. This distinction is about the unadjusted recalculations and should not be interpreted as a definitive classification of the clinical effects. The registry record does not state the confidence level for the sponsor's intended analysis.

In general, intervals become narrower when estimates are more precise and wider when uncertainty is greater. They are influenced by variability, group size, and the analysis model. An interval is not a range containing a fixed proportion of individual outcomes, nor does it provide the probability that a particular effect is true.

Why can over-dispersion matter?

General statistical explanation: over-dispersion occurs when the variability in observed data exceeds what a model assumes. If a model understates variability, standard errors can be too small and intervals too narrow. That concern is especially relevant for outcomes whose distributions have more variability than a simple model allows. The primary endpoint is continuous, and the registry record does not report an over-dispersion assessment or a model diagnostic. It would therefore be speculative to claim that over-dispersion affected these results.

Randomization supports comparisons between groups by helping balance factors that could otherwise influence outcomes. Double masking can reduce some risks of expectation or assessment bias. These are general advantages, not proof that bias was eliminated in this trial. The registry record does not describe the randomization procedure, masking implementation, adherence, or other conduct details needed to assess those issues.

What is known about missing data and analysis populations?

The endpoint description states that missing data were imputed. Imputation fills in missing outcome values using a specified method, and the choice of method can affect estimates and uncertainty. The registry record does not report the method, assumptions, sensitivity analyses, or the number of missing observations. The analyzed population definition is also not reported in the registry record. These omissions limit independent assessment of how incomplete follow-up was handled.

The record lists one primary outcome measure and no posted secondary outcome results. It does not report whether the primary analysis followed an intention-to-treat principle, a per-protocol approach, or another analysis population definition. No assumptions about those choices should be inferred from the presence of randomized groups or from the available group sizes.

10. Confidence Intervals and Precision

The posted least-squares mean changes have standard errors of 0.09 for TLC599 LD, 0.094 for TLC599 HD, and 0.092 for placebo. Standard errors describe uncertainty in the estimated group means, not the range of individual participant responses. The registry record does not provide sponsor-reported confidence intervals for these means.

For the unadjusted LD-versus-placebo recalculation, the estimated mean difference is -0.36 and the interval runs from -0.612 to -0.108. The interval is entirely below zero, indicating that the recalculation is compatible with a more negative mean change in LD than placebo across its interval. For HD versus placebo, the estimate is -0.17 and the interval runs from -0.428 to 0.088. This range includes zero and is compatible with no mean difference as well as with differences in either direction. These intervals are rounded to no more than three decimal places.

Intervals describe uncertainty under the method and assumptions used to calculate them. They do not indicate the distribution of participant-level changes, and they do not account for any unreported modeling choices or multiplicity adjustment. The recalculated intervals are based on posted summary statistics and are not identified as sponsor-generated intervals. The registry record does not report a confidence level for the sponsor's intended analysis, so these intervals should be treated as approximate and unadjusted rather than as a complete account of the trial's inferential plan.

The p-values associated with the same recalculations are 0.005 for LD and 0.196 for HD. The p-values and intervals answer related but not identical presentation questions: the interval emphasizes a range of effect estimates, while the p-value assesses compatibility with a null value under a statistical model. Neither conveys clinical importance. The record does not give a minimum clinically important difference for the WOMAC pain scale.

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 unadjusted effect, not the trial's original design assumptions. The registry record does not report the original sample-size justification. These illustrations use 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 figures are participants per arm, not total events.

For LD versus placebo, planning for 100% of the observed effect gives 26 participants per arm for 80% power and 35 participants per arm for 90% power. Planning for 75% of the observed effect gives 46 participants per arm for 80% power and 61 participants per arm for 90% power. Planning for 50% of the observed effect gives 103 participants per arm for 80% power and 137 participants per arm for 90% power.

For HD versus placebo, planning for 100% of the observed effect gives 116 participants per arm for 80% power and 155 participants per arm for 90% power. Planning for 75% of the observed effect gives 205 participants per arm for 80% power and 274 participants per arm for 90% power. Planning for 50% of the observed effect gives 461 participants per arm for 80% power and 617 participants per arm for 90% power.

In plain terms, the 80% column gives the estimated per-arm size for a hypothetical study designed to have an 80% chance of detecting the chosen effect under the listed assumptions. The 90% column makes the analogous calculation for a 90% chance. Smaller target effects require more participants, as reflected by the larger figures when the assumed effect is a smaller fraction of the observed one. These are planning calculations, not observed power for the completed trial.

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 = -0.360; pooled standard deviation σ = 0.459.
2. n = 2 × (1.96 + 0.842)² × 0.459² / -0.360² = 25.6, rounded up to 26 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 <- -0.36                      # observed difference in means
sigma <- 0.459445095651665                     # 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(3005873)
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 = 26 (80% power), 35 (90% power)
 75% of observed effect: n per arm = 46 (80% power), 61 (90% power)
 50% of observed effect: n per arm = 103 (80% power), 137 (90% power)
power.t.test: 26.6 
Simulated power at n = 26 per arm: 0.792 (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: change from baseline in WOMAC Pain subscale score at Week 12. It does not list posted secondary outcome results. The primary endpoint has two active-versus-placebo comparisons, one for each TLC599 dose group. The unadjusted recalculations do not account for the possibility of making more than one comparison. The registry record does not report whether the trial's intended analysis applied a multiplicity adjustment, used a testing hierarchy, or specified another approach to controlling error across comparisons. Consequently, the recalculated p-values should not be read as adjusted evidence across the dose comparisons.

Missing data were imputed, according to the endpoint description. The imputation method, assumptions about why outcomes were missing, and any sensitivity analyses are not reported in the registry record. These details matter because different approaches to missing outcomes can produce different estimates and different levels of uncertainty. The group sizes listed for the outcome do not reveal how many outcomes were missing or how imputation affected the estimates.

The analysis population is not reported in the registry record. In particular, the record does not state whether the primary analysis included all randomized participants, only participants with observed outcomes, or a defined subset. It also does not report the handling of protocol deviations or treatment discontinuation. These details should not be inferred from the randomized design or the posted outcome counts.

The statistical model used for the least-squares means is not described beyond the parameter label. Covariates, model assumptions, and model diagnostics are not reported. The registry record also does not report an interim analysis plan or a broader statistical analysis plan. These omissions limit the ability to determine whether the simple recalculated comparisons match the trial's prespecified analysis.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: clear cell renal cell carcinoma (3.8%).

For serious adverse events, the registry lists 1 affected participant in the TLC599 LD group, 0 in the TLC599 HD group, and 0 in the placebo group. Deaths are listed as 0 in each group. The serious adverse-event counts at risk are 26 for LD, 24 for HD, and 25 for placebo. These are group-specific figures; they should not be combined into a total or interpreted as rates without additional information about event definitions, follow-up, and exposure.

The registry also lists other adverse events for 18 participants in LD, 20 in HD, and 17 in placebo, with the corresponding groups at risk shown as 26, 24, and 25. The most frequently listed serious-event term is Clear cell renal cell carcinoma, reported as 3.8%. The record does not provide enough detail in these figures to infer event timing, duration, severity beyond the serious-event classification, or causal relation to study treatment. Counts from a trial of this size are descriptive and do not establish comparative safety.

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

The results describe changes in a patient-reported pain score from baseline to Week 12 in three randomized groups. The posted mean changes were -0.83 for TLC599 LD, -0.64 for TLC599 HD, and -0.47 for placebo. Since lower WOMAC Pain scores represent less pain, each negative group estimate indicates a decrease in the score over the assessment period. The unadjusted recalculated differences versus placebo were -0.36 for LD and -0.17 for HD.

The recalculated interval for LD-versus-placebo lies below zero, while the interval for HD-versus-placebo includes zero. These findings describe the uncertainty in simple comparisons based on summary statistics; they are not a replacement for a reported prespecified analysis. The record does not report an adjustment for the two comparisons, and it does not specify a threshold for a clinically important change. Therefore, statistical compatibility with a difference should not be equated with a clinically meaningful benefit.

The results do not establish how individual participants experienced changes, whether the group differences persisted beyond Week 12, or whether any difference was caused by treatment. The record does not provide additional time points, secondary outcome results, analysis-population details, or the imputation method. Those missing elements restrict interpretation of durability, robustness, and generalizability.

The group means also do not establish that the higher TLC599 dose produced a larger effect than the lower dose. The LD estimate is more negative than the HD estimate, but a dose-response conclusion would require an appropriate comparison and analysis, which are not reported in the registry record. The study is described as dose-finding, but that label alone does not establish a dose pattern.

Safety counts are a separate part of the evidence. They are descriptive and do not determine whether the pain-score changes are meaningful. Together, the outcome summaries and safety entries provide a limited account of the trial results, with important uncertainty about analysis methods and missing data.

16. Why This Trial Matters Statistically

This trial is statistically informative because it places two dose groups alongside a placebo group and reports a continuous symptom outcome. Such a design can describe group-level changes and estimate contrasts against a concurrent comparator. The randomized, double-blind, parallel structure is relevant to interpreting those comparisons, although the registry record does not describe how randomization and masking were implemented.

The two recalculated contrasts illustrate why an estimate should be read with its uncertainty. The LD-versus-placebo estimate is -0.36, with an interval below zero, while the HD-versus-placebo estimate is -0.17, with an interval that includes zero. A point estimate alone would conceal this difference in precision. The intervals and unadjusted p-values still do not address the clinical importance of the score changes or any unreported multiplicity procedure.

The trial also illustrates the limits of registry-level outcome summaries. Least-squares means and standard errors are reported, but the model details, analysis population, imputation method, and original sample-size justification are not. Without these details, independent readers cannot fully reconstruct or verify the intended analysis. The recalculated planning illustrations can show how hypothetical sample requirements change when the target effect changes, but they do not reveal the trial's actual design rationale or its observed power.

Finally, dose-finding interpretation requires more than noting that two doses were studied. The posted means do not show a monotonic pattern across the groups, and no dose-response analysis is reported. The trial therefore offers a useful example of why dose comparisons, uncertainty, multiplicity, and missing-data methods must be considered together rather than inferred from the study label or the point estimates alone.

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 endpoint is reported as a least-squares mean with a standard error, but the model specification and adjustment variables are not reported in the registry record. The recalculated contrasts are unadjusted and depend on summary statistics rather than participant-level data. They may not reproduce the intended trial analysis.

The record states that missing data were imputed but does not identify the method, assumptions, or sensitivity analyses. It does not define the primary analysis population or explain how protocol deviations and treatment discontinuation were handled. These gaps make it difficult to assess the robustness of the posted means.

The record does not report the original sample-size justification, the prespecified statistical analysis plan, or a multiplicity procedure for the dose comparisons. The displayed planning calculations are based on observed effects and should not be treated as evidence of the trial's design rationale. No clinically important difference for the WOMAC scale is reported, limiting interpretation of the size of the score changes.

The results cover the primary endpoint at Week 12; additional outcome results are not posted in the registry record. The record does not provide longer-term pain outcomes or details sufficient to assess the durability of change. It also does not describe randomization procedures, masking implementation, adherence, or the timing and context of adverse events. These limitations constrain assessment of bias, generalizability, and comparative safety.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Phase IIa, Randomized, Double Blinded, Placebo Controlled, Dose Finding Study for TLC599 in OA Patients — NCT03005873.
• Registry record last updated 2024-04-22; status last verified 2024-04.
• Results first posted 2024-04-22.
• 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

NCT03005873 is a completed Phase IIa randomized, double-blind, placebo-controlled, parallel-group dose-finding study of TLC599 in people with osteoarthritis of the knee. The sponsor was Taiwan Liposome Company. The study started on 2017-03-16 and reached primary completion on 2018-07-04. Enrollment was 76.

The primary outcome was change from baseline in WOMAC Pain subscale score at Week 12, reported as least-squares mean with standard error. Posted changes were -0.83 for TLC599 LD, -0.64 for TLC599 HD, and -0.47 for placebo. Simple unadjusted recalculations estimated differences versus placebo of -0.36 for LD and -0.17 for HD. Missing data were imputed, but the method and analysis population are not reported in the registry record.

Serious adverse events were reported for 1 participant in LD and none in HD or placebo; deaths were 0 in each group. The record lists one outcome measure and no posted secondary outcome results. The original sample-size justification, model details, multiplicity approach, and sponsor-reported contrast analysis are not reported in the registry record.