Vilazodone for Corticosteroid-Induced Memory Impairment: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Vilazodone for Corticosteroid-Induced Memory Impairment (NCT01828515), a randomized phase 2 study. The primary outcome is Change From Baseline RAVLT (Rey Auditory Verbal Learning Test) Total T-Score at Day 19.

2. Clinical Question

The registry describes a phase 2 randomized, triple-masked crossover study. Its stated purpose was to examine whether vilazodone attenuates corticosteroid-related memory and mood effects on the human hippocampus. The posted primary outcome is a memory test score, not a direct measure of hippocampal activity or mood. The interpretation therefore needs to stay tied to the recorded RAVLT outcome and the comparison of the two medication conditions.

3. Trial Design

The registry record describes a phase 2, randomized, triple-masked crossover trial with an enrollment of 24 and completed status. The sponsor was the University of Texas Southwestern Medical Center. The record lists a start of 2012-12 and primary completion of 2014-04. These dates describe the registry record and do not supply further details about recruitment or follow-up.

A crossover design exposes participants to more than one condition in sequence, so each participant may contribute outcome information under both conditions. The registry lists two sequences: vilazodone and hydrocortisone followed by placebo and hydrocortisone, and the reverse order. It records a 23 day medication washout between periods. The full arm-description text is incomplete in the registry record, so details beyond those legible in the descriptions are not assumed.

For the vilazodone period, the record lists titration to 10 mg for 7 days, 20 mg for 7 days, and 40 mg for 5 days, over 19 days. Hydrocortisone is listed at 160 mg for 4 days, on days 16–19. The placebo period is described as daily placebo for 19 days before the same hydrocortisone schedule. These are the recorded regimens, not a basis for inferring adherence or exposure for each participant.

The primary outcome summaries list 14 observations for each condition. Because the study is a crossover, those condition-specific counts should not be read as separate groups of 14 people without further information. The relationship between the outcome counts and the 24 enrolled participants is not explained in the registry record. Randomization details, the method used to generate or conceal the sequence, and the identities included in triple masking are not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: Healthy control participants without any history of major psychiatric illness, drug/alcohol abuse/dependence, history of neurological disorders, or significant medical conditions were included.

First Intervention (19 Days)

Washout (23 Days)

Second Intervention (19 Days)

Analysis population for the primary outcome, as defined in the registry: All participants who received at least one dose of each intervention and completed all study visits were included in the efficacy analysis.

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 Change From Baseline RAVLT (Rey Auditory Verbal Learning Test) Total T-Score at Day 19, measured over Baseline and Day 19.

The primary endpoint is change from baseline in the Rey Auditory Verbal Learning Test (RAVLT) total T-score at day 19. The record describes the RAVLT as a measure of verbal or declarative learning and memory. In the test, participants hear 15 nouns over five consecutive trials and attempt free recall after each trial. A separate interference list is followed by recall, and recall of the first list is assessed immediately after the interference list and after a 20-minute delay.

The reported score is based on the number of words correct across trials 1–5, converted to a standardized T-score. The registry says equivalent alternative word lists were used to reduce practice or learning effects from repeated administration. The stated scale has M=50 and SD=10, with scores interpreted relative to age- and sex-based norms. Higher scores indicate better performance.

Crucially, the record defines the reported change as baseline score minus day 19 score. A negative change therefore means the day 19 score was higher than baseline on this measure; a positive change means the day 19 score was lower. This direction is important when reading the negative mean in the vilazodone condition. The endpoint is used to quantify test performance over the treatment period; it does not by itself establish a mechanism or represent every aspect of memory or mood.

7. Results Posted in the Registry

• Vilazodone and Hydrocortisone: mean -3.00 versus 0.00 on control; difference -3.00 (95% CI -10.76 to 4.76), standardized effect d = -0.29; z-test p = 0.449.

8. Statistical Interpretation

9. Statistical Methods Explained

The outcome is continuous: each participant's RAVLT result is represented by a numerical T-score, and the registry reports mean change and standard deviation. A mean describes the center of a set of scores, while a standard deviation describes how dispersed those scores are around the mean. General statistical interpretation: a mean is a summary and does not imply that all participants had that value or changed in the same direction.

What does the change score represent?

The registry defines change as baseline minus day 19. This subtraction order determines the sign. In general, a negative score under this convention means the later score is larger than the baseline score; a positive score means the later score is smaller. The posted mean changes are -3.0 in the vilazodone-and-hydrocortisone condition and 0.0 in the placebo-and-hydrocortisone condition. The comparison is between these condition means, not between the number of participants who improved or worsened.

Using a change score can describe how test results shifted over a period rather than relying only on the final measurement. The registry does not report whether baseline values were also included as covariates in a model, or whether another outcome model was used. Nor does it describe whether the reported means account for the crossover pairing. Those analysis details are not reported in the registry record, so the comparison should be described using the posted estimand and summary statistics rather than an assumed model.

How should the mean difference be read?

The posted mean difference is -3.0 T-score points, with the vilazodone-and-hydrocortisone condition first and placebo-and-hydrocortisone second. In general, a mean difference is the subtraction of one condition's mean from the other's. Here the negative direction indicates that the mean change score was lower in the vilazodone condition under the record's baseline-minus-day-19 convention. That sign does not mean that the day 19 memory score was necessarily lower in that condition: because the change is defined in the reverse order, the negative mean change corresponds to a higher day 19 score than baseline on average.

A mean difference condenses the comparison into one number, but it does not show the distribution of individual responses. It also does not establish that the difference is practically important. A clinical or scientific threshold for importance is not reported in the registry record. The displayed estimate should therefore be considered alongside its uncertainty and the way the study analyzed its crossover observations.

What does the confidence interval contribute?

A confidence interval gives a range of estimates reasonably compatible with the data and the method used to calculate it. General statistical explanation: a 95% confidence procedure is designed so that, over repeated samples under its assumptions, intervals constructed by that procedure would cover the target parameter at the stated rate. A single interval is not a probability statement that the true value has a 95% chance of being inside its endpoints.

The posted interval for the mean difference extends from -10.76 to 4.759. Because it includes zero, the values compatible with the interval include no mean difference as well as differences in either direction. The interval is not a range of individual participants' scores. Its validity depends on the statistical method and assumptions used, and the registry record does not describe enough about the analysis to assess those assumptions fully.

How should the p-value be interpreted?

The posted unadjusted p-value is 0.449. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and statistical model. It is not the probability that the null hypothesis is true, and it is not a measure of the size or importance of an effect. The record does not report a detailed hypothesis-testing plan or the model underlying this p-value. It is therefore appropriate to describe it as the posted unadjusted value, without inferring additional tests or adjustments.

The p-value and confidence interval address related but distinct aspects of uncertainty. The p-value summarizes evidence against a null value under a test; the interval shows a range of effect estimates compatible with the analysis. Here the interval includes zero and the p-value is 0.449. Neither result proves equivalence or rules out potentially meaningful differences.

Why does the crossover structure matter?

In a crossover trial, participants receive conditions in sequence, and measurements from the same participant are generally related. General statistical explanation: accounting for within-person correlation can affect the uncertainty of a comparison; treating repeated observations as independent when they are related may misrepresent precision. Period effects, sequence effects, and carryover are also considerations in crossover analysis. The record lists a 23 day medication washout, but does not report whether the analysis tested or adjusted for period, sequence, or carryover, or how participant-level pairing was incorporated. No conclusion about those features should be inferred from the presence of a washout alone.

The comparisons list 14 observations for each condition, while the overall enrollment is 24. In a crossover, condition-level counts need not represent independent groups, because one participant can contribute data in both conditions. The record does not explain how many participants contributed complete paired outcomes or whether any observations were excluded from the analysis beyond the stated efficacy-population definition.

What do the spread statistics describe?

The reported standard deviations are 12.48 for vilazodone and hydrocortisone and 7.98 for placebo and hydrocortisone. The comparison also reports a pooled standard deviation of 10.47. General statistical explanation: a larger standard deviation indicates more variation in observed scores, while a pooled standard deviation combines information about spread across conditions according to a method. It is not itself an estimate of the mean difference. The record's calculation method for the pooled value is not described in detail.

Over-dispersion is a general concern when the variability in data exceeds what an assumed statistical model expects. It can make standard errors too small and confidence intervals too narrow if the model fails to account for that variability. The posted summaries alone do not establish whether over-dispersion occurred here, and the registry does not report a diagnostic assessment. For a continuous change score, the key point is that spread and the analysis model both affect precision; the posted standard deviations do not replace the uncertainty interval for the comparison.

Overall, the registry provides a continuous outcome, condition means, standard deviations, a mean difference, confidence limits, and an unadjusted p-value. It does not report a complete statistical analysis plan, model specification, or detailed handling of crossover structure and missing measurements. Interpretation should remain anchored to the posted estimate and its uncertainty, while recognizing that those analysis details are unavailable.

10. Confidence Intervals and Precision

The estimated mean difference is -3.0 T-score points, and its 95% confidence interval ranges from -10.76 to 4.759. The interval includes zero, so the posted data are compatible with no mean difference as well as differences in either direction. It also includes values more negative than the point estimate and values above zero. The confidence interval therefore communicates substantial uncertainty about the direction and size of the condition contrast.

In general, a confidence interval describes uncertainty about a population-level estimate under the analysis method; it is not a band containing most individual participants' results. The interval's width reflects both the variability in the outcome and the information available for the estimate, along with the calculation method. The registry lists 14 observations for each condition and standard deviations of 12.48 and 7.98. In a crossover study, however, the relation between the two condition observations within participants matters for precision. The registry record does not state how that relation was incorporated into the confidence interval.

The interval should not be interpreted as proof of equivalence or as proof that any effect is absent. Equivalence requires a prespecified margin and a suitable analysis; neither is reported in the registry record. Nor is a threshold for a clinically or scientifically important score difference reported. The confidence limits thus help show the imprecision of the posted estimate but do not independently determine whether values within the range are important.

The p-value of 0.449 is consistent with the interval including zero, but it does not quantify the probability that the null is true. The record gives no adjusted comparison or additional inferential analyses. The most defensible precision statement is that the estimated difference is uncertain, and the interval permits a range of contrasts rather than pinning down one direction or magnitude.

11. Planning a Similar Trial: Sample Size

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

The planning figures are participants per arm, not total events. This is a crossover study, but the planning illustration is explicitly described in arm-based terms; it should not be mistaken for the enrolled sample or for the number who completed the study. The registry's own original sample-size assumptions and justification are not reported in the registry record.

The illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its starting point, then considers detecting 100%, 75%, or 50% of that observed effect. The 80% power column means the number of participants per arm estimated to give an 80% chance of detecting the assumed effect under those planning assumptions, if that effect is present and the assumptions hold. The 90% power column uses the same interpretation but targets a 90% chance.

For 100% of the observed effect, the illustration gives 192 participants per arm for 80% power and 257 participants per arm for 90% power. For 75% of the observed effect, it gives 341 participants per arm for 80% power and 456 participants per arm for 90% power. For 50% of the observed effect, it gives 766 participants per arm for 80% power and 1,025 participants per arm for 90% power. Smaller effects require more participants under these assumptions because they are harder to distinguish from random variation.

These are planning illustrations computed from the observed effect, not the trial's own design calculations and not observed power. They do not show the probability that this completed trial succeeded, nor do they account for unreported design choices or analysis methods. They are sensitive to the observed effect used as the starting point, so they should be read as conditional planning examples rather than definitive sample-size recommendations.

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 = -3.000; pooled standard deviation σ = 10.475.
2. n = 2 × (1.96 + 0.842)² × 10.475² / -3.000² = 191.4, rounded up to 192 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 <- -3                      # observed difference in means
sigma <- 10.4745119218033                     # 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(1828515)
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 = 192 (80% power), 257 (90% power)
 75% of observed effect: n per arm = 341 (80% power), 456 (90% power)
 50% of observed effect: n per arm = 766 (80% power), 1,025 (90% power)
power.t.test: 192.3 
Simulated power at n = 192 per arm: 0.804 (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 lists one primary outcome measure and no posted secondary outcome measures. It lists zero analyses. Those entries do not establish that only one statistical test was conducted; a full statistical analysis plan and a complete account of analyses are not reported in the registry record. The record also does not report whether there were interim analyses, whether any outcome definitions changed, or whether adjustments for multiple testing were planned or performed.

Multiplicity arises when several hypotheses or outcomes are tested. General statistical explanation: the more tests conducted, the greater the chance that at least one result appears unusual by chance if each test is interpreted without accounting for the set of tests. The registry information is not sufficient to assess the full testing family for this study. The posted p-value is explicitly unadjusted and should not be represented as adjusted for multiple comparisons.

The efficacy population is described as all participants who received at least one dose of each intervention and completed all study visits. This definition identifies a selected analysis population; it is not a statement that all 24 enrolled participants contributed complete outcome data. The condition summaries list 14 observations each, but participant flow, reasons for missing outcomes, and reconciliation with enrollment are not reported in the registry record.

Missing-data handling is also not reported. The record does not say whether missing scores were imputed, whether incomplete participants were excluded, or whether sensitivity analyses assessed alternative assumptions about missingness. It does not describe an intention-to-treat analysis. These omissions matter because an analysis limited to participants who completed all visits can differ from one that includes all randomized participants, particularly if completion is related to outcomes or treatment experience. Whether that occurred cannot be determined from the record.

For the crossover design, the record does not explain how the statistical analysis accounted for repeated measurements within participants, nor whether period, sequence, or carryover effects were considered. Those omissions limit assessment of the model behind the comparison and its uncertainty. They do not invalidate the posted values by themselves, but they mean the numerical results cannot be fully evaluated from registry information alone.

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 non-serious adverse event terms

The adverse-event record lists serious adverse events for 0 of 24 participants in the vilazodone-and-hydrocortisone group and 0 of 24 in the placebo-and-hydrocortisone group. No serious adverse-event terms are listed. Deaths are recorded as 0 in both groups.

These are counts for the groups and at-risk totals as reported in the registry. Zero recorded serious events means none were reported in these entries; it does not establish that the risk of a serious event is zero. The record does not provide additional information about event definitions, ascertainment procedures, timing, or whether participants contributed to both crossover conditions for safety reporting. The counts should not be interpreted as independent samples without further design-specific detail.

The record separately lists other adverse events for 2 of 24 participants in the vilazodone-and-hydrocortisone group and 0 of 24 in the placebo-and-hydrocortisone group. These are not serious-event counts and should not be conflated with them. The registry does not provide terms for those other events in the listed information. Adverse-event counts are descriptive and do not by themselves establish a causal relationship to an intervention.

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

The posted result is a comparison of mean change in RAVLT total T-score at day 19 between vilazodone plus hydrocortisone and placebo plus hydrocortisone. The mean changes were -3.0 and 0.0, respectively, and the estimated mean difference was -3.0. Under the record's baseline-minus-day-19 definition, a negative mean change corresponds to a higher score at day 19 than at baseline on average. The confidence interval, -10.76 to 4.759, includes zero, and the unadjusted p-value is 0.449. The data therefore do not provide a clear distinction between the condition means in this posted comparison.

This result does not show that the conditions are equivalent. No equivalence margin is reported, and the interval includes a range of possible differences. It also does not prove that an effect is absent: a nonsignificant result can reflect uncertainty as well as a small effect. The confidence interval is the more informative description of how broadly the estimated contrast could vary under the analysis, although the underlying analysis details are not fully reported.

The result applies to the recorded endpoint and analysis population, not to every aspect of memory or mood. The RAVLT is a verbal learning and memory measure, and the record describes a norm-referenced T-score. This outcome is not a direct readout of hippocampal activity. The trial summary's purpose mentions mood and hippocampal effects, but no posted primary result for mood or a direct hippocampal measure is included in the registry record described in the registry record.

The population was described as healthy controls, and both compared conditions included hydrocortisone. The comparison therefore does not answer what would happen without hydrocortisone, or establish a result for people with a diagnosed memory disorder or other populations. The record also does not support conclusions about other doses, durations, or treatment settings. Its completed status indicates the study is recorded as completed, not that every scientific question raised by its purpose has been resolved.

Finally, safety counts are limited descriptive information. No serious adverse events or deaths are listed in either condition's record, but the available counts do not prove zero risk. The findings should be read as the outcome and safety information recorded for this small crossover study, with the statistical and reporting limitations described elsewhere.

16. Why This Trial Matters Statistically

This trial is statistically instructive because its crossover design makes the unit of comparison more subtle than a simple parallel-group comparison. Participants can contribute data under both conditions, so repeated observations from the same person may be related. How that pairing is handled can influence uncertainty. The registry provides condition-level summaries and a comparison, but does not describe the model well enough to establish how within-person correlation, period, sequence, or potential carryover were addressed.

The study also illustrates why an estimate should be read together with its confidence interval. The mean difference is -3.0, but the interval extends from -10.76 to 4.759 and includes zero. A single point estimate can give a misleading impression of certainty if its uncertainty is not considered. The interval makes clear that the posted result does not narrowly identify one direction or magnitude of difference.

The change-score convention is another important interpretive detail. Because the registry defines change as baseline minus day 19, the negative mean in one condition points to a higher day 19 score relative to baseline, not a decline as might be assumed under the opposite subtraction order. Stating the direction explicitly prevents a sign from being misread.

Finally, the planning illustration shows how sample-size requirements depend on the effect targeted and the assumptions used. Smaller target effects require more participants in the displayed scenarios. Those planning figures are not a substitute for the original protocol rationale, which is not reported in the registry record, and they are not evidence about the trial's observed power. Together, these features make the record a useful example of separating design, estimand, uncertainty, and planning assumptions when interpreting a clinical result.

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 leaves several trial-specific uncertainties. Although it describes a randomized, triple-masked crossover design, it does not report sequence-generation or allocation-concealment methods, explain who was masked, or provide the full protocol. The arm descriptions are incomplete in the record, limiting what can be established about the intended procedure beyond the legible schedule details.

The enrollment is 24, while the primary outcome summaries list 14 observations for each condition. The efficacy analysis is defined as including participants who received at least one dose of each intervention and completed all study visits, but participant flow and reasons for any exclusions or missing outcomes are not reported. The record does not explain how these numbers relate to the enrolled group or whether the condition observations form complete pairs.

The statistical model and analysis plan are not reported. In particular, the record does not say how the crossover pairing was handled, or whether period, sequence, or carryover effects were assessed. A 23 day washout is recorded, but the record does not provide evidence about whether it was sufficient to prevent carryover. No conclusion about actual carryover or its absence can be drawn from that schedule alone.

The posted comparison is unadjusted, and the record does not describe the broader testing plan, multiplicity handling, interim analyses, missing-data methods, or sensitivity analyses. It also does not report a clinically or scientifically meaningful difference threshold. The study concerns a specific RAVLT outcome in healthy controls under two hydrocortisone-containing conditions, so its scope is limited to that comparison and recorded population. These limitations qualify interpretation of the estimate; they do not change the numerical result recorded for the comparison.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Vilazodone for Corticosteroid-Induced Memory Impairment — NCT01828515.
• Registry record last updated 2019-04-11; status last verified 2019-03.
• Results first posted 2018-07-05.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes NCT01828515, a completed phase 2 randomized, triple-masked crossover study sponsored by the University of Texas Southwestern Medical Center. Enrollment was 24, and the listed population was healthy controls. The two conditions paired vilazodone plus hydrocortisone or placebo plus hydrocortisone, with a 23 day medication washout between periods recorded in the arm descriptions.

The primary outcome was change from baseline in RAVLT total T-score at day 19, defined as baseline minus day 19. The posted mean changes were -3.0 and 0.0, with a mean difference of -3.0 and a 95% confidence interval from -10.76 to 4.759; the unadjusted p-value was 0.449. The interval includes zero. The record lists zero serious adverse events and zero deaths in both conditions. It does not report enough analysis detail to determine how crossover pairing, missing data, or other design features were handled.