A Study to Evaluate 3 Dose Levels of Luvadaxistat of Adults With Negative Symptoms of Schizophrenia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate 3 Dose Levels of Luvadaxistat of Adults With Negative Symptoms of Schizophrenia (NCT03382639), a randomized phase 2 study. The primary outcome is Change From Baseline on the PANSS NSFS at Week 12.

2. Clinical Question

3. Trial Design

This completed Phase 2 study was randomized, double-blind, placebo-controlled, and parallel-group. It evaluated three dose levels of luvadaxistat as adjunctive treatment for adults with negative symptoms of schizophrenia. Participants were assigned to luvadaxistat 50 mg, 125 mg, or 500 mg, or to matching placebo; each was taken orally once daily for 12 weeks, from Day 1 to Day 84. The stated enrollment was 256.

The study start date was 2018-01-04, and primary completion was 2020-12-29. The registry record lists Neurocrine Biosciences as sponsor and the study status as completed. These record details describe the trial’s administrative design and conduct; they do not establish how many participants contributed to every endpoint or visit.

The primary outcome was change from baseline in PANSS Negative Symptom Factor Score at Week 12. The registry defines the FAS as randomized participants who received at least 1 dose of double-blind study treatment. Outcome counts were 87 for placebo, 58 for luvadaxistat 50 mg, 56 for luvadaxistat 125 mg, and 55 for luvadaxistat 500 mg. These are the counts attached to the posted primary results.

The registry reports a mixed model for repeated measures for the primary endpoint and gives sponsor comparisons labelled Mixed Models Analysis. Further details of the original protocol’s allocation procedure, sample-size justification, model specification, covariates, missing-data handling, and interim analyses are not reported in the registry record. The planned design should therefore be distinguished from assumptions used in any separate planning illustration.

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: The study consisted of a screening period of up to 28 days, a 14-day single-blind placebo run-in period, a 12-week double-blind treatment period and a safety follow-up visit. Randomization was stratified by age at screening (\<35 and ≥35 years). The allocation ratio was 2:2:2:3 to the 3 luvadaxistat groups and placebo group, respectively.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The FAS included all randomized participants who received at least 1 dose of double-blind study treatment.

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: The Full Analysis Set (FAS) will included all participants who were randomized and received at least 1 dose of the study drug during the Double-Blind Treatment Period.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline on the PANSS NSFS at Week 12, measured over Baseline and Week 12.

The primary endpoint was change from baseline in the PANSS Negative Symptom Factor Score at Week 12. PANSS assesses positive symptoms, negative symptoms, and general psychopathology associated with schizophrenia. It contains 30 items, each rated from 1, symptom not present, to 7, symptoms extremely severe. The Negative Symptom Factor Score consists of 7 items and ranges from 7 to 49; a higher score indicates greater severity.

Baseline was defined as the last observed value before the first dose of study treatment. The registry reports least squares mean change from baseline, estimated using a mixed model for repeated measures. A negative change indicates improvement on this score. The primary comparison asks whether the mean change differed between each luvadaxistat dose group and placebo at Week 12.

A negative-symptom measure is relevant to the stated study question because the trial was designed to evaluate treatment added to existing care in adults identified as having negative symptoms. The score provides a defined, repeated measure rather than relying on a broad impression alone. As a continuous endpoint, it can describe both the direction and magnitude of average change, while uncertainty measures indicate how precisely group differences are estimated. The registry’s endpoint description does not establish that changes in the score necessarily correspond to changes in daily functioning or other outcomes.

7. Results Posted in the Registry

• Luvadaxistat 50 mg: mean -3.30 versus -3.10 on control; difference -0.20 (95% CI -1.54 to 1.14), standardized effect d = -0.05; z-test p = 0.769.
• Luvadaxistat 125 mg: mean -3.40 versus -3.10 on control; difference -0.30 (95% CI -1.62 to 1.02), standardized effect d = -0.08; z-test p = 0.656.
• Luvadaxistat 500 mg: mean -2.50 versus -3.10 on control; difference 0.60 (95% CI -0.78 to 1.98), standardized effect d = 0.15; z-test p = 0.394.

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.

Change From Baseline on the Brief Negative Symptom Scale (BNSS) Total Score (12-item) at Week 12

• 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:
  • Luvadaxistat 50 mg: mean -8.90 versus -7.10 on control; difference -1.80 (95% CI -4.94 to 1.34), standardized effect d = -0.19; z-test p = 0.262.
  • Luvadaxistat 125 mg: mean -8.00 versus -7.10 on control; difference -0.90 (95% CI -4.04 to 2.24), standardized effect d = -0.09; z-test p = 0.575.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Luvadaxistat 50 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Luvadaxistat 125 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline on the Brief Assessment of Cognition in Schizophrenia (BACS) Composite Score at Week 12

• 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:
  • Luvadaxistat 50 mg: mean 4.60 versus 2.30 on control; difference 2.30 (95% CI -0.20 to 4.80), standardized effect d = 0.30; z-test p = 0.072.
  • Luvadaxistat 125 mg: mean 3.50 versus 2.30 on control; difference 1.20 (95% CI -1.29 to 3.69), standardized effect d = 0.16; z-test p = 0.344.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Luvadaxistat 50 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Luvadaxistat 125 mg and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline on the Schizophrenia Cognition Rating Scale (SCoRS) Interviewer Total Score at Week 12

• 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:
  • Luvadaxistat 50 mg: mean -3.80 versus -1.60 on control; difference -2.20 (95% CI -4.19 to -0.21), standardized effect d = -0.36; z-test p = 0.030.
  • Luvadaxistat 125 mg: mean -2.30 versus -1.60 on control; difference -0.70 (95% CI -2.69 to 1.29), standardized effect d = -0.12; z-test p = 0.490.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Luvadaxistat 50 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Luvadaxistat 125 mg and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broadened assessment beyond the primary negative-symptom factor score. They included negative symptoms measured by the Brief Negative Symptom Scale, cognition measured by the Brief Assessment of Cognition in Schizophrenia, clinician-rated severity and improvement, interviewer-rated cognition, and additional PANSS measures. The registry lists outcomes at Week 4, Week 8, and Week 12, as well as several measures at Week 12.

The detailed numerical secondary results include the BNSS total score, BACS composite score, and SCoRS interviewer total score. Across these measures, group patterns are not uniform: some posted treatment-group changes are more favorable in direction than placebo, while other contrasts are smaller or point in the opposite direction. The recalculated intervals for these comparisons generally retain substantial uncertainty. The pattern does not provide a consistent basis for overriding the primary result or establishing a clear dose-related effect.

Numerical results for the other listed secondary outcomes are not reported in the registry record. Their inclusion in the outcome list indicates what was assessed, not what the study found. The registry also does not provide a single integrated analysis across all secondary measures.

Unadjusted secondary comparisons carry less weight because multiple outcomes and comparisons create more opportunities for chance findings. Unless multiplicity is addressed through a prespecified plan, isolated nominal p-values should be treated as exploratory rather than as independent confirmation. The registry record does not report whether a multiplicity adjustment was used. The secondary findings are therefore best viewed as context for the primary endpoint and as potential inputs to further study, not as a replacement for the primary analysis.

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

• Change From Baseline on the PANSS NSFS at Week 4 and Week 8
• Change From Baseline on the Brief Negative Symptom Scale (BNSS) Total Score (12-item) at Week 12
• Change From Baseline on the Brief Assessment of Cognition in Schizophrenia (BACS) Composite Score at Week 12
• Change From Baseline on the Clinical Global Impression-Schizophrenia Severity (CGI-SCH-S) Negative Symptoms Score at Week 12
• Clinical Global Impression Schizophrenia Improvement (CGI-SCH-I) Negative Symptoms Score at Week 12
• Change From Baseline on the Schizophrenia Cognition Rating Scale (SCoRS) Interviewer Total Score at Week 12
• Change From Baseline on the PANSS Total Score at Week 12
• Change From Baseline on the PANSS Positive Symptom Factor Score (PSFS) at Week 12

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is continuous: it is a score measured on a numeric scale, and the posted group results are least squares mean changes. A group mean summarizes the average modeled change, while a between-group mean difference expresses how far apart the modeled averages are. For this endpoint, the interpretation of direction depends on the scale’s definition: negative change from baseline indicates improvement, but a treatment-minus-placebo difference must be read in relation to the reported group values and the stated contrast.

What the repeated-measures model does

A mixed model for repeated measures, or MMRM, is a general approach for analyzing outcomes collected at multiple visits. It uses the pattern of measurements over time to estimate group means and contrasts at specified visits. Modeling repeated observations can account for the fact that measurements from one participant are related to each other. The registry identifies MMRM for the primary outcome and gives the Week 12 result; it does not report the model’s full specification, such as its covariance structure, covariates, or exact handling of missing observations. Those details should not be inferred.

Least squares means are model-based averages, often adjusted for terms included in the model. They are not necessarily identical to simple arithmetic averages of the raw observations. The sponsor’s reported contrasts are labelled LS Mean Difference and come from Mixed Models Analysis. A separate recalculation from posted group summaries is an unadjusted approximation; it does not reproduce the repeated-measures analysis.

What the standard error conveys

The group results report standard errors. A standard error describes the statistical uncertainty in an estimated mean: smaller standard errors generally indicate a more precise estimate, while larger ones indicate greater uncertainty. It is not the same as the spread of individual participants’ scores. The posted group standard errors are 0.44 for placebo, 0.52 for luvadaxistat 50 mg, 0.51 for 125 mg, and 0.55 for 500 mg.

The recalculation uses the posted standard errors as the basis for approximate variability estimates, as indicated by the recalculation method. It then estimates an unadjusted mean difference and its confidence interval. Because this reconstruction relies on summary statistics and does not incorporate the sponsor’s repeated-measures model, its values can differ from the reported model-based estimates. The sponsor’s own analysis takes precedence when describing the trial’s primary analysis.

How to read a confidence interval and p-value

A confidence interval gives a range of values compatible with the observed estimate under the method used. Its width communicates precision: wider intervals leave more uncertainty about the size and direction of the underlying difference. For the sponsor’s primary comparisons, all reported 95% intervals include zero. Zero represents no difference between the compared group means on the contrast scale. An interval containing zero does not prove no effect; it indicates that the data and method do not sharply distinguish the estimate from zero at that confidence level.

A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The sponsor’s p-values are linked to the model-based comparisons. The recalculated p-values are unadjusted and based on approximate summary-statistic calculations. They should not be confused with the p-values from the sponsor’s analysis or treated as confirmation of its model.

Why over-dispersion and variability matter

In general, over-dispersion means that observed variability is greater than a model assumes. When variability is underestimated, standard errors may be too small and confidence intervals too narrow; when it is modeled appropriately, uncertainty better reflects the data. The registry record does not identify an over-dispersion issue or report a separate adjustment for it. This is a general statistical consideration, not a claim about this trial.

For a continuous endpoint, variability affects how precisely group differences can be estimated. A small estimated difference may be compatible with a broad range of true differences when uncertainty is substantial. In this trial, the confidence intervals around the sponsor’s comparisons span values on either side of zero, so their interpretation should emphasize the range of uncertainty rather than the point estimates alone.

Multiplicity and analysis choices

Multiple comparisons raise the chance of observing a nominally notable result by chance if many hypotheses are tested. The primary endpoint includes three dose-versus-placebo comparisons; secondary outcomes also cover several measures and visits. Whether and how the original study controlled multiplicity is not reported in the registry record. Accordingly, the posted p-values should be read as reported, without assuming a family-wise adjustment or an ordered testing procedure.

Similarly, the registry does not report the trial’s original sample-size justification or interim analysis plan. The sample-size illustrations associated with observed effects are separate planning calculations, not evidence about the design choices made before enrollment. They do not alter the interpretation of the completed study’s results.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data estimate a group difference. The sponsor reports 95% confidence intervals for the Week 12 primary comparisons. For 50 mg, the interval runs from -1.4 to 1.2 around an LS mean difference of -0.1. For 125 mg, it runs from -1.5 to 1.1 around -0.2. For 500 mg, it runs from -0.7 to 1.9 around 0.6. Each interval crosses zero, so each remains compatible with no average difference under the reported model, as well as with differences in either direction.

The intervals are not narrow enough to establish that the true between-group difference is exactly zero or to rule out all potentially relevant differences. The registry record does not state a threshold for a clinically important change, so the clinical importance of the range of values cannot be determined from these results alone. A point estimate is the center of the reported comparison, not a guarantee of the underlying effect.

The recalculated intervals are also broad: -1.535 to 1.135 for 50 mg, -1.62 to 1.02 for 125 mg, and -0.780 to 1.98 for 500 mg. They are unadjusted approximations from posted summary statistics, not the sponsor’s MMRM intervals. Their agreement in general direction and broad overlap with zero supports a cautious reading, but the recalculations do not supersede the sponsor analysis.

In general, confidence intervals become narrower when an estimate is more precise and wider when data leave more uncertainty. Precision depends on factors such as sample size, variability, and the analysis method. The group counts and standard errors help describe the information behind the estimates, but they do not make an interval conclusive. Here the main statistical message is uncertainty around modest point estimates, not a precise estimate of a nonzero treatment difference.

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 sample-size figures are planning illustrations expressed as participants per arm, not total events. They use the observed effect as a starting point and show the illustrative sample size needed for a stated fraction of that effect. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The original trial’s own sample-size justification is not reported in the registry record.

The 80% power column means the illustrated design would have an 80% chance of detecting the specified effect under the calculation assumptions if that effect were the true one. The 90% power column applies the same interpretation at a higher target chance of detection. These are design-planning concepts; neither column describes the probability that the completed trial found a real effect.

Rows for planning for 100%, 75%, and 50% of the observed effect illustrate how the assumed effect size changes the planning requirement. Smaller effects need more participants to detect with the same target power, all else equal. The figures are calculated separately for each dose-versus-placebo comparison and are not the original study’s planned enrollment targets.

The calculations are not observed power. They are forward-looking planning illustrations based on effects estimated in the completed trial. Because observed effects are uncertain, these figures should not be treated as a definitive recommendation for a future study or as evidence that the trial was adequately or inadequately planned.

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 = -0.200; pooled standard deviation σ = 4.047.
2. n = 2 × (1.96 + 0.842)² × 4.047² / -0.200² = 6,428.6, rounded up to 6,429 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 <- -0.2                      # observed difference in means
sigma <- 4.04732286550287                     # 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(3382639)
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 = 6,429 (80% power), 8,607 (90% power)
 75% of observed effect: n per arm = 11,429 (80% power), 15,300 (90% power)
 50% of observed effect: n per arm = 25,715 (80% power), 34,425 (90% power)
power.t.test: 6429.5 
Simulated power at n = 6,429 per arm: 0.801 (20000 replicates)

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

14. Analyses Reported by the Sponsor

The sponsor’s reported primary results come from Mixed Models Analysis and are presented as LS mean differences with 95% confidence intervals and p-values. The recalculated comparisons use posted group means and standard errors to form simple unadjusted estimates. These methods answer related questions but are not identical: the recalculations cannot reproduce the repeated-measures model or any model features not reported in the registry record.

For the 50 mg comparison, the sponsor reports -0.1 with a 95% confidence interval of -1.4 to 1.2 and p = 0.426; the recalculation is -0.2 with an interval of -1.535 to 1.135 and p = 0.769. For 125 mg, the sponsor reports -0.2, an interval of -1.5 to 1.1, and p = 0.725; the recalculation is -0.3, an interval of -1.62 to 1.02, and p = 0.656. For 500 mg, the sponsor reports 0.6, an interval of -0.7 to 1.9, and p = 0.808; the recalculation is 0.6, an interval of -0.780 to 1.98, and p = 0.394.

The registry record does not flag a discrepancy. The primary interpretation should therefore follow the sponsor’s model-based analysis, while the recalculation serves as a transparent approximation from posted summaries. Neither set of results should be read as proof of equivalence.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when investigators test several hypotheses, such as multiple dose groups against one comparator or several secondary outcomes. Even when each test uses a conventional significance threshold, a larger set of tests increases the chance that at least one result appears notable by chance. This issue is relevant because the primary analysis includes three dose-versus-placebo comparisons and the study lists numerous secondary measures and time points.

The registry record does not report a multiplicity-control method, such as an adjustment to the testing threshold or a hierarchical testing sequence. It also does not state whether secondary analyses were designated as confirmatory or exploratory. Therefore, the p-values should be read as the values reported for individual comparisons; no correction should be presumed. This limitation is particularly important when interpreting isolated secondary results.

The registry defines the FAS as randomized participants who received at least 1 dose of double-blind study treatment. It identifies this population for the primary results, but the record does not provide a detailed account of how missing outcome observations were handled in the MMRM. The model name alone does not establish the assumptions used for missing data, the amount or pattern of missingness, or any sensitivity analyses. These details are not reported in the registry record.

The record also does not report the original sample-size justification, the full prespecified statistical analysis plan, interim analyses, or whether any analysis changes were made. These omissions limit assessment of how closely the posted analyses correspond to the original plan. They do not invalidate the reported results, but they prevent a complete audit of the choices governing estimation, missing data, and multiple testing.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: schizophrenia (1.8%), psychotic disorder (1.1%), rhabdomyolysis (1.1%), covid-19 (1.1%), tooth abscess (1.1%).

The posted serious adverse event counts were 4 participants in the placebo group, none in the luvadaxistat 50 mg group, 1 in the 125 mg group, and none in the 500 mg group. The corresponding groups at risk were 87, 58, 56, and 55. The registry reports no deaths in any group. These are counts from the posted safety results and should be interpreted with the group sizes in mind.

Other adverse events were reported for 4 participants receiving placebo, none receiving 50 mg, 5 receiving 125 mg, and 2 receiving 500 mg. The registry also lists serious-event terms including schizophrenia, psychotic disorder, rhabdomyolysis, COVID-19, and tooth abscess, with posted percentages of 1.8%, 1.1%, 1.1%, 1.1%, and 1.1%, respectively.

These counts describe reported events in this study; they do not by themselves establish whether an event was caused by treatment. The registry record does not provide enough detail to assess event timing, duration, severity beyond the serious-event classification, or individual clinical context. Sparse counts also limit comparisons among groups. The figures should be read descriptively, not as proof of relative safety or risk.

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

The primary result is a comparison of average PANSS Negative Symptom Factor Score change at Week 12 between each luvadaxistat dose group and placebo. The sponsor’s model-based estimates are close to zero, and all reported 95% confidence intervals include zero. The posted results therefore leave uncertainty about the direction and size of the average difference. They do not establish equivalence, because no equivalence margin is reported and the intervals include a range of possible differences.

The endpoint is a symptom score. It is not, by itself, a direct measure of every aspect of participants’ lives, and the registry record does not establish how score changes relate to functioning or other outcomes. The score’s stated range and direction help interpret the measurement, but they do not supply a threshold for what would be clinically important. Such a threshold is not reported in the registry record.

The results also do not settle questions about other doses, longer treatment, different populations, or treatment outside the study setting. The study evaluated the stated dose levels over 12 weeks in adults with schizophrenia and negative symptoms. Generalizing beyond that population, treatment period, and design requires evidence not contained in these results.

The secondary measures and safety counts offer additional context, but their interpretation is limited by multiple comparisons, uncertainty, and missing reporting details. The posted results can describe what was estimated and how precisely, but they do not prove a mechanism, a clinically meaningful benefit, or a lack of effect. No treatment recommendation follows from these statistical results alone.

18. Why This Trial Matters Statistically

This trial illustrates the distinction between a within-group change and a treatment-versus-comparator effect. All four primary groups had negative mean changes, but the relevant question is whether their changes differed. Comparing the treatment groups with placebo, rather than interpreting change in isolation, is essential when improvement may occur in both groups during a study.

It also illustrates why point estimates, confidence intervals, and analysis methods should be considered together. The sponsor’s estimates are modest and their intervals include zero, leaving appreciable uncertainty. The model-based analysis uses repeated observations, while a simple reconstruction from posted summaries is only an approximation. Reporting both the source analysis and the limits of a recalculation helps readers distinguish the trial’s reported inference from an independent arithmetic check.

The study has three dose comparisons and multiple secondary outcomes, making multiplicity relevant. Without a reported adjustment plan, isolated secondary findings deserve caution. Likewise, the planning illustrations based on observed effects are not the original study’s design rationale and are not observed power. Keeping these concepts separate helps prevent a retrospective calculation from being mistaken for evidence about the trial’s planning adequacy.

Finally, the record demonstrates that statistical interpretation depends on what is reported as well as what is estimated. Missing detail about the statistical plan, multiplicity, and missing-data handling limits assessment of robustness. A clear account of those boundaries is part of interpreting clinical-trial results responsibly.

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.

Several trial-specific limitations affect interpretation. The registry record identifies the primary endpoint analysis as MMRM but does not report the full model specification, including covariates, covariance structure, or detailed missing-data methods. It does not report the original sample-size justification, interim analyses, or the statistical plan for multiplicity. Those gaps make it difficult to reconstruct the complete prespecified analytical framework.

The registry defines the FAS, but does not provide detailed participant flow or explain whether all randomized participants contributed to every visit and outcome. Outcome group counts are available for the primary endpoint, while a full account of missing measurements and reasons for missingness is not reported. The extent to which missing-data assumptions might affect estimates cannot therefore be assessed from the registry record.

The recalculated comparisons are unadjusted approximations based on posted summary statistics. They do not reproduce the sponsor’s repeated-measures model, and the estimated variability is approximate. They are useful as a cross-check but should not be treated as an alternative definitive analysis. The sponsor’s reported analysis takes precedence.

There are multiple dose comparisons and several secondary outcomes, but multiplicity control is not reported. Numerical results are also not reported for every outcome listed in the registry. Safety results are available as counts, but sparse counts and limited event detail constrain comparative interpretation. Finally, the registry does not report a threshold for clinical importance on the primary score, limiting assessment of whether a difference of a particular size would matter clinically.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate 3 Dose Levels of Luvadaxistat of Adults With Negative Symptoms of Schizophrenia — NCT03382639.
• Registry record last updated 2024-02-28; status last verified 2024-01.
• Results first posted 2024-02-28.
• 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

NCT03382639 was a completed Phase 2 randomized, double-blind, placebo-controlled, parallel study of adjunctive luvadaxistat in adults with negative symptoms of schizophrenia. It evaluated oral once-daily doses of 50 mg, 125 mg, and 500 mg against matching placebo for 12 weeks.

The primary endpoint was change from baseline in PANSS Negative Symptom Factor Score at Week 12, analyzed using a mixed model for repeated measures. The sponsor’s reported dose-versus-placebo confidence intervals all included zero. The registry record reports the study dates, enrollment, primary outcome counts, and selected secondary and safety results, but does not report several details needed to assess the full statistical plan, multiplicity, and missing-data handling.