Melatonin Versus Placebo in Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Melatonin Versus Placebo in Breast Cancer (NCT01805089), a randomized early/phase 1 study. The primary outcome is Absolute Plasma Estradiol Levels After 4 Month Course of Melatonin or Placebo.

2. Clinical Question

3. Trial Design

The registry describes a randomized, triple-masked, parallel-group trial. Its recorded phase is EARLY PHASE1, and its status is COMPLETED. The study was sponsored by Dana-Farber Cancer Institute. The registry lists an enrollment of 95; the posted primary outcome results include 43 participants in the melatonin group and 43 in the placebo group. The relationship between enrollment and the outcome-specific analyzed groups is not reported in the registry record.

The intervention was melatonin 3 mg taken orally once per day at or around 9:00pm. The placebo was also taken orally once per day at or around 9:00pm. The trial record gives a start of 2006-10 and a primary completion of 2009-07. Those dates describe the registry record and do not establish when any particular participant began or completed treatment.

The primary outcome is absolute plasma estradiol after a 4 month course. The registry says only the 4-month level is provided; baseline estradiol values and change-from-baseline results are not included in the posted primary outcome. A secondary outcome listed in the record concerns change in mood, sleep quality, and menopausal symptoms from baseline to 4 months. The registry record does not report detailed eligibility criteria, the recruitment setting, treatment adherence, or the analysis population definition.

The record identifies a Wilcoxon (Mann-Whitney) comparison for the primary outcome. It does not report the protocol’s original sample-size assumptions, allocation ratio details beyond the randomized parallel design, or other planned analysis procedures. Those omissions limit what can be concluded about how closely the posted results follow the prespecified design.

4. Participant Flow and Accounting

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

Overall Study

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 Absolute Plasma Estradiol Levels After 4 Month Course of Melatonin or Placebo, measured over 4 months.

The primary endpoint is absolute plasma estradiol, measured in pg/ml after a 4 month course of melatonin or placebo. The posted group summaries are means with standard deviations: the melatonin group has a mean of 4.59 and standard deviation of 3.42, while the placebo group has a mean of 3.39 and standard deviation of 2.25. The registry describes the endpoint as the 4-month level and explicitly notes that only this level is provided. It does not post baseline levels alongside the result.

Estradiol is an estrogen hormone. The study summary says the research was intended to determine whether nightly melatonin could affect estrogen or IGF blood levels, which may influence breast cancer risk. This states the study’s rationale; it does not establish that a change in estradiol predicts a change in cancer outcomes. The posted endpoint is a laboratory measure, not a direct measure of recurrence, survival, or another cancer outcome.

In general, a continuous endpoint such as a hormone concentration can be summarized by its mean and the spread of individual values around that mean. A post-treatment level comparison describes group values at the specified follow-up. Without posted baseline results or a change analysis, it cannot by itself show how much estradiol changed within either group over the treatment period.

7. Results Posted in the Registry

• Melatonin: mean 4.59 versus 3.39 on control; difference 1.20 (95% CI -0.02 to 2.42), standardized effect d = 0.41; z-test p = 0.055.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a Wilcoxon (Mann-Whitney) method for comparing melatonin and placebo on the primary endpoint. It also posts group means and standard deviations, from which a separate unadjusted comparison of means has been recalculated. These methods answer related but distinct statistical questions, so their results need not match exactly.

What does the Wilcoxon (Mann-Whitney) test compare?

In general, the Wilcoxon (Mann-Whitney) test is a rank-based method. It orders observed values and compares how the ranks are distributed across groups. It does not directly test the difference between arithmetic means in the way a conventional mean-comparison test does. Its interpretation as a general shift in location relies on assumptions about the shapes of the group distributions; if their shapes differ, the rank comparison can reflect more than a simple location shift. The registry does not report diagnostic checks or additional details about the test’s assumptions.

The sponsor’s posted result is p <0.05, with no estimate or confidence interval reported alongside it. A p-value summarizes how compatible the data are with a specified null model under the test’s assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. A p-value alone also does not reveal the range of effect sizes compatible with the observed data.

What does the mean-based recalculation represent?

In general, a difference in means compares the average outcome in one group with the average in the other. The recalculated contrast is oriented as melatonin minus placebo and uses the posted group means of 4.59 and 3.39 pg/ml. The resulting unadjusted mean difference is 1.2 pg/ml. The calculation uses the posted summary statistics rather than participant-level observations, and it does not adjust for baseline values or other variables. The registry does not report an adjusted mean comparison.

The standard deviation describes the spread of observations within a group; it is not the uncertainty around the group mean. The posted standard deviations are 3.42 in the melatonin group and 2.25 in the placebo group. The recalculation uses information about group sizes and variability to estimate uncertainty in the difference in means. It is an unadjusted summary-statistic calculation, not a reconstruction of the sponsor’s rank-based analysis.

Why can the two analyses yield different results?

General statistical reasoning explains why: a rank-based test and a mean-based test use the data differently, and they target different features of the distributions. A mean can be influenced by unusually high or low observations, while ranks depend on relative ordering. The registry does not provide individual observations, so it is not possible to examine whether distribution shape or extreme values explain the difference between the reported p-values. The sponsor’s own analysis remains the reported trial analysis; the recalculation offers a separate view based on the posted means and standard deviations.

What does a confidence interval add?

In general, a confidence interval gives a range of values compatible with the data and the method used to calculate it. The recalculated interval for the mean difference is -0.024 to 2.424 pg/ml. Because it contains zero, it does not rule out a zero difference in means at the conventional two-sided 0.05 level under that calculation. The interval is not a range containing a fixed proportion of individual estradiol measurements, and it is not a guarantee that the true effect lies within its endpoints.

The methods recorded do not resolve all questions needed for a full statistical assessment. The registry record does not report whether the primary analysis was prespecified in a protocol, whether it was adjusted for baseline estradiol, how missing outcome data were handled, or which participants were included in the sponsor’s test. It also does not report a multiplicity procedure for the several outcomes described in the study summary. These omissions make it important to distinguish the specific posted result from broader conclusions about the study’s overall evidence.

10. Confidence Intervals and Precision

The recalculated mean difference is 1.2 pg/ml for melatonin minus placebo. Its confidence interval is -0.024 to 2.424 pg/ml. The lower endpoint is slightly below zero and the upper endpoint is above zero, so the interval includes no difference in means. It also permits positive differences within the interval. This is uncertainty about the group mean contrast, not a description of the range of values among individual participants.

In general, a confidence interval conveys both the estimated direction and the precision of an effect estimate. A narrower interval indicates greater precision under the calculation’s assumptions; a wider interval allows a broader range of compatible effects. The interval should be read with the underlying summary-statistic method in mind. It is a recalculation from posted group summaries, not an interval reported with the sponsor’s Wilcoxon analysis.

The recalculated unadjusted p-value is 0.055. Together with the interval, it indicates that the mean-based calculation does not clearly distinguish the observed mean difference from zero at the conventional two-sided 0.05 threshold. This does not invalidate the sponsor’s reported rank-based p <0.05: the methods differ, and the registry does not supply the sponsor’s effect estimate or confidence interval. The registry record does not report additional analyses that would clarify the sensitivity of either result to missing data, baseline adjustment, or distributional features.

Precision is also limited by what is posted. The record gives outcome-specific group sizes, means, and standard deviations, but does not provide participant-level values or baseline estradiol. Consequently, the interval describes uncertainty for the recalculated 4-month mean contrast only; it does not quantify uncertainty in change from baseline or in any clinical outcome not reported.

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 illustrate the group size required under the stated calculation assumptions when planning for 100%, 75%, or 50% of the observed effect. For 100% of the observed effect, the estimates are 92 participants per arm for 80% power and 123 participants per arm for 90% power. For 75% of the observed effect, they are 163 and 218 participants per arm, respectively. For 50% of the observed effect, they are 366 and 490 participants per arm, respectively.

The calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. In plain language, the 80% power column gives the estimated number of participants per arm for a design with an 80% chance of detecting that assumed effect under the calculation’s conditions. The 90% power column gives the corresponding estimate for a 90% chance. Higher target power requires more participants under these assumptions, and smaller effects require more participants than larger effects.

These are planning illustrations based on the observed effect, not the trial’s original sample-size rationale; the original design assumptions are not reported in the registry record. This is not observed power. The estimates do not establish what the completed trial was designed to detect, and they should not be read as a retrospective judgment about the trial’s result.

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 = 1.200; pooled standard deviation σ = 2.895.
2. n = 2 × (1.96 + 0.842)² × 2.895² / 1.200² = 91.3, rounded up to 92 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 <- 1.2                      # observed difference in means
sigma <- 2.89472796649357                     # 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(1805089)
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 = 92 (80% power), 123 (90% power)
 75% of observed effect: n per arm = 163 (80% power), 218 (90% power)
 50% of observed effect: n per arm = 366 (80% power), 490 (90% power)
power.t.test: 92.3 
Simulated power at n = 92 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.

13. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis uses a Wilcoxon (Mann-Whitney) test and gives p <0.05. The registry record does not report a sponsor-derived effect estimate or confidence interval for that analysis. The recalculated result instead compares the posted group means: it gives a melatonin-minus-placebo difference of 1.2 pg/ml, a confidence interval from -0.024 to 2.424, and an unadjusted p-value of 0.055.

These results should be kept distinct. The Wilcoxon test is rank based, whereas the recalculation is a comparison of arithmetic means from summary statistics. They do not necessarily test the same feature of the data, and one should not be substituted for the other. The sponsor’s own analysis takes precedence as the reported trial analysis; the recalculation is an additional, unadjusted reading of the posted summaries. The registry record does not report enough detail to reproduce the sponsor’s test or determine why its p-value differs from the mean-based recalculation.

14. Multiplicity, Missing Data and Analysis Populations

The study summary describes several areas of measurement, including estrogen and IGF levels and effects on sleep, hot flashes, and mood. The primary outcome posted is absolute plasma estradiol at 4 months. The registry lists a secondary outcome concerning change in mood, sleep quality, and menopausal symptoms from baseline to 4 months. It records three outcome measures and one posted secondary outcome. Detailed results for the other measures are not reported in the registry record.

When a study examines multiple outcomes or makes multiple statistical comparisons, the chance of obtaining at least one nominally notable result can increase. In general, a multiplicity strategy may control how such results are interpreted, for example by adjusting significance thresholds or identifying a single primary hypothesis. The registry record does not report whether a multiplicity adjustment was planned or used, how the outcome hierarchy was defined, or whether any analyses were exploratory. The sponsor’s primary p-value should therefore be understood as the posted result for its reported test, not as evidence that all outcomes were accounted for in a broader testing strategy.

Missing-data handling is also not reported in the registry record. Although group sizes are posted for the primary outcome, the record does not explain whether those measurements include all randomized participants, how absent 4-month values were treated, or whether any imputation or sensitivity analysis was used. The analysis population is not specified. The relationship between the registry enrollment figure and the primary-outcome analysis groups is not explained, so it should not be inferred from the posted numbers.

The record does not provide a statistical analysis plan, baseline estradiol summaries, adjusted analyses, or details about protocol deviations. These omissions matter because analysis choices can affect both the estimate and its uncertainty. They also limit evaluation of whether the reported test and the summary-statistic recalculation address a prespecified primary question. The absence of those details is a reporting limitation, not proof that a particular procedure was or was not used.

15. Safety Results

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

Other (non-serious) adverse events

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

Most frequent non-serious adverse event terms

The registry records serious adverse events for 0 participants in the melatonin group among 48 at risk and 0 participants in the placebo group among 47 at risk. No serious adverse event terms are listed. For other adverse events, the record reports 7 affected participants among 48 at risk in the melatonin group and 1 affected participant among 47 at risk in the placebo group. These are the posted counts; they should not be read as a complete description of event severity, timing, or causality.

The record does not report deaths for either group. It also does not provide event descriptions for the other adverse events, information about how events were collected, or an assessment of whether events were considered related to treatment. The at-risk counts for the adverse-event reporting are not the same as a statement about who contributed to the primary estradiol analysis. The registry’s safety figures are descriptive counts and, without further detail, do not establish comparative risk or explain individual participants’ experiences.

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

The posted data show mean plasma estradiol levels of 4.59 pg/ml in the melatonin group and 3.39 pg/ml in the placebo group after 4 months. The mean-based recalculation estimates a difference of 1.2 pg/ml, with a confidence interval from -0.024 to 2.424. That interval includes zero, so this calculation leaves uncertainty about the direction and size of the mean difference. The sponsor separately reports a Wilcoxon (Mann-Whitney) p-value of <0.05; that rank-based result should be distinguished from the recalculated mean comparison.

The result does not by itself show whether melatonin changes breast cancer risk, recurrence, or survival. Estradiol is the posted primary endpoint, and the study summary gives a biological rationale for measuring hormone levels. A laboratory measure and a clinical cancer outcome are not interchangeable. The registry record does not report results for IGF, nor detailed results for the listed secondary outcome on mood, sleep quality, and menopausal symptoms.

The 4-month endpoint is an absolute level, not a reported change from baseline. Without baseline results, the posted means do not show the amount of within-group change or whether baseline levels were similar. The registry also does not report how missing values were handled or define the analysis population. These gaps limit the interpretation of both the effect estimate and its precision.

The study is recorded as completed, but the posted results do not settle every question raised by its summary. They provide a specific hormone-level comparison and a sponsor-reported rank-based test, while leaving uncertainty about the mean difference and about several design and analysis details. The findings should be interpreted within that scope rather than extended to outcomes not reported in the registry record.

17. Why This Trial Matters Statistically

This trial illustrates why the statistical method matters when reading a clinical result. The sponsor’s reported Wilcoxon (Mann-Whitney) comparison yields p <0.05, while the unadjusted recalculation from the posted means and standard deviations yields p = 0.055 and a confidence interval that includes zero. These are not necessarily contradictory findings: the methods use the data differently and do not target precisely the same feature of the outcome distribution. Reporting an effect estimate and its uncertainty alongside a test result would make the magnitude and precision easier to assess.

The trial also shows the limits of interpreting a post-treatment biomarker level without posted baseline data. The endpoint is absolute estradiol at 4 months, but no baseline estradiol summary or change analysis is included. A reader can compare the posted group means at follow-up, but cannot infer the amount of change within either group from those values alone.

Finally, the record’s concise outcome reporting leaves unanswered questions about analysis populations, missing data, and multiplicity across outcomes. These are central to judging how robust a result may be, particularly when a study includes multiple measures. The posted information supports a bounded statistical reading of the primary hormone result, not a broader conclusion about outcomes that were not reported.

18. 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 reports only the 4-month estradiol level for the primary outcome and explicitly does not provide baseline levels. As a result, the posted summaries cannot describe within-group change or show whether the groups had similar estradiol levels before treatment. The primary outcome’s population field is blank, and the analysis population definition is not reported in the registry record.

The reported sponsor analysis is a Wilcoxon (Mann-Whitney) test with p <0.05, but the record gives no sponsor-reported effect estimate or confidence interval. The mean-based recalculation is therefore not a full reproduction of that analysis. Participant-level data and distribution summaries are not posted, preventing examination of the observations or the assumptions relevant to either method.

The registry does not report missing-data procedures, baseline adjustment, protocol deviations, the original sample-size justification, or the statistical analysis plan. It lists a secondary outcome on mood, sleep quality, and menopausal symptoms, but detailed results are not reported. The study summary mentions IGF, sleep, hot flashes, and mood, while the primary results discussed are limited to estradiol. The record does not establish how these outcomes were prioritized or whether multiplicity was addressed.

The adverse-event record provides counts but not event descriptions, relatedness assessments, collection procedures, or death counts. These limitations restrict interpretation of safety as well as efficacy-related outcomes. They do not establish that unreported analyses or procedures were absent; they mean those details are not reported in the registry record.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Melatonin Versus Placebo in Breast Cancer — NCT01805089.
• Registry record last updated 2015-03-23; status last verified 2015-02.
• Results first posted 2015-02-23.
• 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.

22. Record Summary

The registry record describes a completed, randomized, triple-masked, parallel-group trial of oral melatonin 3 mg versus placebo in breast cancer survivors. Both were taken once per day at or around 9:00pm. The listed primary endpoint is absolute plasma estradiol at 4 months. The posted means are 4.59 pg/ml for melatonin and 3.39 pg/ml for placebo.

The sponsor reports a Wilcoxon (Mann-Whitney) p-value of <0.05. A separate unadjusted recalculation of the mean difference is 1.2 pg/ml, with a confidence interval from -0.024 to 2.424 and p = 0.055. The record does not report baseline estradiol, the analysis population, missing-data handling, or detailed results for the listed secondary outcome. The registry’s findings therefore support a focused reading of the posted 4-month hormone comparison, with important uncertainty about analysis details and outcomes not reported.