A Study to Assess the Safety and Efficacy of Levoketoconazole in the Treatment of Endogenous Cushing's Syndrome.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Assess the Safety and Efficacy of Levoketoconazole in the Treatment of Endogenous Cushing's Syndrome. (NCT03277690), a randomized phase 3 study. The primary outcome is Number of Subjects With Loss of Therapeutic Response to Levoketoconazole Upon Withdrawing to Placebo Compared With the Proportion of Subjects With Loss of Therapeutic Response Upon Continuing Treatment With Levoketoconazole..

2. Clinical Question

This is a withdrawal question in a population previously exposed to treatment, rather than a comparison of starting treatment with no prior treatment. The posted outcome counts participants who met the registry definition of loss of response; it does not report a continuous measure of cortisol change as the primary result. The comparison therefore concerns the occurrence of a defined event during the randomized phase. The registry reports 21 participants with loss of response in the placebo group and 9 in the levoketoconazole group. Those counts describe the posted result for the randomized withdrawal phase and should be interpreted in the context of the selected study population and its outcome definition.

3. Trial Design

NCT03277690 is recorded as a completed phase 3 randomized parallel study of levoketoconazole for endogenous Cushing's syndrome. The registry lists enrollment of 84 participants, with a start date of 2017-09-26 and primary completion on 2020-06-30. Cortendo AB is listed as the sponsor. The official study title describes a double-blind, placebo-controlled, randomized withdrawal following open-label therapy. The brief description also characterizes the study as a withdrawal and rescue or restoration study.

Participants first received single-arm open-label levoketoconazole, according to the study description, and the randomized portion compared continued levoketoconazole with withdrawal to placebo. The registry lists the masking as quadruple, while the study title and brief description use the term double-blind. The record does not explain which study roles were masked or how these masking descriptions relate to one another. In the randomized withdrawal phase, the arm descriptions assign levoketoconazole, at a dose up to 1200 mg, or placebo. The restoration phase descriptions pair the two treatments in opposite sequences.

The primary result is reported for the Intent to Treat population. Its arm counts are 22 for levoketoconazole and 22 for placebo. These are the group sizes for the posted primary outcome result, not a description of how the total enrollment was distributed across every stage of the study. The record reports one outcome measure and zero analyses. It does not report a statistical analysis plan, the original sample-size justification, or enough detail to reconstruct the flow of participants through open-label treatment, withdrawal, and restoration. Those details are not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: Variable therapeutic response results in customized therapeutic doses (TD). The study design included an initial Titration-maintenance phase to identify a TD for each subject. It included a requirement for maintenance at TD for at least 4 weeks prior to entering the Randomized-withdrawal (RW) phase. Then a 1:1 RW to placebo was used to establish efficacy of the active therapy. 5 additional subjects began directly in RW after completing the parent study without requiring re-titration to TD.

Dose Titration and Maintenance

Reasons for not completing this period:

Randomized Withdrawal

Reasons for not completing this period:

Restoration

Analysis population for the primary outcome, as defined in the registry: Intent to Treat (ITT)

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 Number of Subjects With Loss of Therapeutic Response to Levoketoconazole Upon Withdrawing to Placebo Compared With the Proportion of Subjects With Loss of Therapeutic Response Upon Continuing Treatment With Levoketoconazole., measured over max. 9.5 weeks.

The primary endpoint counts participants with loss of therapeutic response to levoketoconazole after withdrawal to placebo, compared with the proportion who lost response while continuing levoketoconazole. The registry defines the outcome using the mean urinary free cortisol, or mUFC, calculated from three 24-hour urinary free cortisol measurements at a visit from RW1 through RW5. Loss of response was recorded if mUFC was above 1.5 times the upper limit of normal for the central laboratory reference range, or if mUFC was more than 40% above the RW0 baseline when that baseline was above the upper limit of normal, or if an early rescue criterion was met.

The outcome is binary at the participant level: each participant is counted as having or not having the defined event. A binary endpoint can express whether a specified clinical or laboratory threshold was crossed, rather than the size of a continuous change. Here, the criteria combine cortisol thresholds with an early rescue condition, so the event count reflects the registry's operational definition rather than a single cortisol value. The endpoint is used to compare the frequency of loss of response during withdrawal with the frequency during continued treatment. Its timeframe was a maximum of 9.5 weeks. The record does not provide further detail about the early rescue criterion or report additional primary endpoint components.

7. Results Posted in the Registry

• Levoketoconazole: 40.9% (9/22) versus 95.5% (21/22) on control; risk difference -54.5 percentage points (95% CI -76.9 to -32.2), risk ratio 0.43 (0.26–0.71); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The posted primary outcome is a binary participant count: participants are classified according to whether they met the loss-of-therapeutic-response definition. For the placebo group, 21 of 22 participants had the event; for the levoketoconazole group, 9 of 22 did. The primary population is identified as Intent to Treat. The registry presents a risk difference, risk ratio, odds ratio, and unadjusted p-value for the comparison. These are distinct statistical summaries of the same two-group outcome, not four separate outcome measurements.

What does a risk difference describe?

General explanation: a risk difference subtracts the event proportion in one group from the event proportion in the other, using a specified comparison order. It is an absolute measure, expressed on the probability scale. The reported risk difference is -0.545, with an interval from -0.769 to -0.322, for levoketoconazole compared with placebo. Its negative sign indicates that the event proportion was lower in the levoketoconazole group under this ordering. The interval expresses uncertainty around the estimated contrast. The registry record does not state the interval's confidence level or the method used to calculate it.

How do relative measures differ?

General explanation: a risk ratio compares the event probabilities between groups by division; a value of 1 represents equal probabilities. An odds ratio compares event odds rather than probabilities, and a value of 1 represents equal odds. Because odds and probabilities are not interchangeable, an odds ratio can look numerically farther from 1 than a risk ratio, especially when events are common. In this study the reported risk ratio is 0.429, with an interval from 0.257 to 0.714. The reported odds ratio is 0.033, with an interval from 0.004 to 0.291. The registry does not name the calculation methods used for these estimates or intervals.

What does the p-value add?

The registry reports an unadjusted p-value of <0.001. General explanation: a p-value evaluates compatibility between observed data and a null hypothesis under a particular statistical test and its assumptions. Smaller p-values indicate that data such as these would be less compatible with that null model. They do not quantify the probability that the null is true, the probability that the result is due to chance, or the magnitude of an effect. Because the test is not identified in the registry record, the test-specific assumptions cannot be assessed from the posted information. The p-value is unadjusted, and the record does not describe a multiplicity procedure.

What can be said about the analysis population?

The outcome is designated for the Intent to Treat population, and the reported randomized withdrawal groups each contain 22 participants. General explanation: an Intent to Treat analysis is ordinarily intended to preserve randomized group assignment in the comparison, including when participants do not follow the assigned regimen. However, the registry record does not specify how participants with missing measurements, early rescue, or other departures were classified for this endpoint. It also does not state whether all randomized participants contributed a complete endpoint assessment. The posted counts should therefore be read as the recorded results, without assuming an unreported missing-data procedure.

What does the endpoint definition mean for analysis?

The event is based on mUFC from three 24-hour urinary free cortisol measurements, at any eligible visit from RW1 through RW5, and can also be triggered by an early rescue criterion. General explanation: defining a binary endpoint from thresholds turns potentially varying measurements into an event count. That makes group frequencies straightforward to compare, but it can discard information about how far measurements lie from a threshold or how they change within a participant. The record gives the threshold conditions but does not report the handling of incomplete sets of measurements, repeated qualifying visits, or the operational details of early rescue. Those analysis choices are not reported in the registry record.

The study is described as randomized and parallel, with withdrawal to placebo compared against continuation of levoketoconazole after open-label therapy. Random assignment can support a group comparison when the analysis follows the design and its assumptions, but the record alone does not establish that every analysis condition was met. The registry reports zero analyses and does not give a detailed model specification, a prespecified testing hierarchy, covariates, or a statistical analysis plan. General statistical interpretation must therefore remain at the level of the reported counts and measures rather than attributing a particular model, test, or adjustment to the study.

General explanation: over-dispersion is a concern when outcome variability exceeds what a chosen model assumes. For a simple binary outcome summarized as participant counts, the relevant uncertainty still depends on the sampling and analysis assumptions; for clustered or repeated observations, treating observations as independent can understate uncertainty. The registry's primary result is a participant-level count, not a model of the repeated cortisol measurements. It does not describe a model for those measurements, a cluster adjustment, or an over-dispersion assessment. No conclusion about such procedures can be drawn from the registry record.

10. Confidence Intervals and Precision

The risk difference estimate is -0.545, with an interval from -0.769 to -0.322. The risk ratio is 0.429, with an interval from 0.257 to 0.714, and the odds ratio is 0.033, with an interval from 0.004 to 0.291. These intervals convey uncertainty around their respective estimates. They do not express the same scale: the risk difference is an absolute contrast, while the risk ratio and odds ratio are relative measures. Each should be interpreted with its own estimate and interval rather than comparing the interval widths directly across measures.

General explanation: a confidence interval is a range produced by a statistical procedure intended to reflect uncertainty in an estimate under specified assumptions. A narrower interval indicates greater precision on that measure's scale, while a wider interval indicates less precision. An interval is not a guarantee that the true value lies within its particular endpoints, and it does not show the range of individual participant outcomes. The registry record does not report the confidence level or calculation method for the posted intervals, so those details cannot be specified.

The primary outcome groups each include 22 participants, and the estimates are based on 21 events in the placebo group and 9 in the levoketoconazole group. The estimates therefore reflect the observed counts in this randomized withdrawal phase, not a larger or longer-term population by direct measurement. The intervals offer an indication of uncertainty around the comparison, but their interpretation depends on methods not detailed in the registry record. The unadjusted p-value is <0.001; it is a separate summary and does not replace the estimates or their intervals.

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. They are illustrations computed from the observed effect, with two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is a stated fraction of the observed effect. These are planning assumptions for the illustration; the trial's original sample-size assumptions and justification are not reported in the registry record.

For planning around 100% of the observed effect, the illustration gives 11 participants per arm for 80% power and 14 participants per arm for 90% power. For 75% of the observed effect, it gives 17 participants per arm for 80% power and 22 participants per arm for 90% power. For 50% of the observed effect, it gives 31 participants per arm for 80% power and 40 participants per arm for 90% power. The 80% column means the planned design would have an 80% chance of detecting the specified effect under the stated assumptions if that effect and the model assumptions were correct. The 90% column expresses the same idea at a higher target chance of detection.

Smaller effects require more participants in these illustrations: the figures rise as the assumed effect moves from the full observed effect toward a smaller fraction. These numbers are not observed power and do not establish the study's original design rationale. They are planning illustrations based on the observed result. The registry record does not report whether the original trial design used these assumptions or any other sample-size method.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.955; treatment proportion p₁ = 0.409; average p̄ = 0.682.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6587 = 1.2910.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5340 = 0.4494.
4. n = (1.2910 + 0.4494)² / (-0.5455)² = 10.2, rounded up to 11 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
p0 <- 0.954545454545455                          # control proportion
p1 <- 0.409090909090909                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(3277690)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 11 (80% power), 14 (90% power)
 75% of observed effect: n per arm = 17 (80% power), 22 (90% power)
 50% of observed effect: n per arm = 31 (80% power), 40 (90% power)
power.prop.test: 10.2 
Simulated power at n = 11 per arm: 0.852 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 outcome measure and zero analyses. It does not list secondary outcomes. The primary outcome is therefore the only posted outcome measure in this record, but the record does not establish the full set of outcomes or comparisons that may have been specified in the protocol or analysis plan. The registry record does not report a multiplicity adjustment, a hierarchy of hypotheses, or whether any interim analyses were planned or conducted.

The reported p-value is unadjusted. General explanation: when many outcomes or comparisons are tested, the chance of obtaining at least one small p-value by chance can increase. Multiplicity procedures can address that issue, but the appropriate procedure depends on the prespecified testing structure. Since that structure and any adjustment are not reported in the registry record, no multiplicity correction can be inferred from the posted result. The unadjusted p-value should be described as such, not treated as evidence that other possible comparisons were accounted for.

The primary population is labeled Intent to Treat. The registry does not explain whether the analysis included every randomized participant or how missing outcome measurements were handled. It also does not state the handling of participants who met an early rescue criterion, despite rescue being included in the endpoint definition. Methods for imputation, exclusion, censoring, or sensitivity analyses are not reported in the registry record. These omissions matter because the event classification depends on repeated cortisol measurements and may also depend on a rescue event.

The original statistical analysis plan, sample-size justification, interim analysis plans, and detailed analysis model are not reported in the registry record. As a result, it is not possible from the posted information to assess whether the analysis matched a prespecified plan or to identify adjustments beyond the unadjusted result. The posted group counts and estimates remain interpretable as registry results, but a fuller assessment of multiplicity and missing-data handling requires information not reported in the registry record.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hypokalemia (4.5%), abdominal pain (2.4%), liver disorder (2.4%), limb injury (1.2%), hepatic enzyme increased (1.2%).

In the randomized withdrawal phase, the registry reports 1 serious adverse event among 22 participants in the levoketoconazole group and 0 among 22 in the placebo group. Other adverse events are reported for 11 of 22 participants in the levoketoconazole group and 10 of 22 in the placebo group. No deaths are reported for either withdrawal-phase group. In the safety population, 13 of 84 participants had a serious adverse event, 73 of 84 had another adverse event, and no deaths were reported.

The registry lists the serious adverse event terms Hypokalemia at 4.5%, Abdominal pain at 2.4%, Liver disorder at 2.4%, Limb Injury at 1.2%, Hepatic enzyme increased at 1.2%, and Myocardial infarction at 1.2%. The record does not specify the denominators or additional detail for these term-level percentages in this summary. The safety population covers the full reported population, whereas the withdrawal-phase figures concern the randomized phase; the groups should not be treated as interchangeable. Counts describe recorded events and do not by themselves establish cause, severity beyond the serious classification, or comparative risk.

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

The posted primary result shows that loss of therapeutic response, as defined by the registry, was recorded for 21 of 22 participants assigned to placebo during randomized withdrawal and 9 of 22 participants assigned to continue levoketoconazole. The reported risk difference is -0.545, with an interval from -0.769 to -0.322, and the unadjusted p-value is <0.001. These statistics describe a contrast in recorded event frequency in this randomized withdrawal phase. The interval communicates uncertainty about the estimate, while the p-value addresses compatibility with a null comparison under an unspecified test.

The result does not answer every question about levoketoconazole. The participants had first received open-label therapy, so the comparison concerns withdrawal versus continuation among previously treated participants. It is not a direct comparison of initiating levoketoconazole against placebo in an untreated population. The endpoint is a composite operational definition involving urinary free cortisol thresholds and early rescue, and the timeframe was a maximum of 9.5 weeks. The result therefore does not describe outcomes beyond that period or separately quantify each component of the endpoint.

The primary counts also do not establish how missing measurements were handled, whether the posted methods followed a prespecified plan, or how other outcomes would compare. No secondary outcomes or analyses are posted in the registry record, and detailed safety analysis methods are not reported. Adverse event counts and term-level percentages should be read as recorded safety information, not as proof of causation. The statistics describe what was reported for this study and should not be converted into individual treatment advice or a broad conclusion about all patients with endogenous Cushing's syndrome.

16. Why This Trial Matters Statistically

This trial's statistical question is shaped by its randomized withdrawal design. Rather than comparing initial treatment with placebo in all enrolled participants, the study randomized participants after open-label levoketoconazole therapy to continued treatment or withdrawal to placebo. This design makes the clinical comparison specific: whether the defined loss-of-response event was recorded during withdrawal versus continued treatment among participants who reached the randomized phase.

The binary endpoint turns multiple urinary free cortisol measurements and an early rescue condition into a participant-level event count. That gives a clear frequency comparison, while also making interpretation dependent on the threshold definition and the handling of repeated measurements and rescue. The registry reports the primary outcome counts and several effect measures, which show the same comparison on absolute and relative scales. The posted interval is important because it makes the uncertainty around the risk difference visible rather than relying on a significance statement alone.

For biostatistical interpretation, the record also illustrates the limits of registry reporting. The primary population and endpoint definition are stated, but the record does not provide the analysis test, interval method, missing-data procedures, multiplicity approach, or original sample-size rationale. The unadjusted p-value can be reported, but the absence of those details restricts independent assessment of the inferential process. The study's randomized structure and posted estimates are informative, while careful interpretation requires keeping the design, outcome definition, time window, and reporting gaps in view.

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 posted result has a short, design-specific scope. It concerns participants previously treated with open-label levoketoconazole who entered the randomized withdrawal phase; the record does not show that the primary result applies to participants who had not received prior treatment. The primary outcome timeframe is a maximum of 9.5 weeks, so the result does not describe longer-term event patterns. The registry does not provide the number who entered each stage of open-label, withdrawal, and restoration, beyond the overall enrollment and the posted primary outcome arm counts.

The endpoint combines cortisol threshold criteria with an early rescue criterion. The detailed rescue rule is not reported in the registry record, nor are procedures for incomplete or repeated measurements, or the relationship between rescue and event classification. Missing-data methods and sensitivity analyses are also not reported. Those gaps limit assessment of how robust the counts are to different outcome-handling choices.

The record reports an unadjusted p-value but does not identify the test, interval calculation method, confidence level, multiplicity approach, or detailed statistical plan. The original sample-size assumptions are not reported; the posted planning illustrations are based on the observed effect and are not evidence of the trial's own design assumptions. The masking entry and the study descriptions use different terms, and the registry does not explain the roles masked. Finally, zero analyses and no secondary outcomes are posted, limiting the range of results that can be assessed from this registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Study to Assess the Safety and Efficacy of Levoketoconazole in the Treatment of Endogenous Cushing's Syndrome. — NCT03277690.
• Registry record last updated 2022-11-08; status last verified 2022-10.
• Results first posted 2022-11-08.
• 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

NCT03277690 is a completed phase 3 randomized parallel study of levoketoconazole in endogenous Cushing's syndrome, sponsored by Cortendo AB. The registry lists enrollment of 84, a start date of 2017-09-26, and primary completion on 2020-06-30. Its design is described as open-label therapy followed by a double-blind randomized withdrawal and rescue or restoration study; masking is recorded as quadruple.

The primary outcome counts participants with loss of therapeutic response during withdrawal to placebo compared with continued levoketoconazole. In the Intent to Treat randomized withdrawal result, the event counts were 21 of 22 for placebo and 9 of 22 for levoketoconazole. The reported risk difference was -0.545, with an interval from -0.769 to -0.322, and the unadjusted p-value was <0.001. The registry reports one outcome measure and zero analyses. Details of the statistical plan, missing-data methods, and original sample-size justification are not reported in the registry record.