Physiological Effects of New Polyphenol-enriched Foods in Humans: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Physiological Effects of New Polyphenol-enriched Foods in Humans (NCT01288859), a randomized na study. The primary outcome is Serum Polyphenol Concentrations Over 24h From Food Consumption.

2. Clinical Question

The clinical question concerns measurable serum exposure after consumption of foods containing free or encapsulated polyphenols, rather than a clinical health outcome. The registry’s brief summary frames the study as an evaluation of bioavailability and describes encapsulation as a possible way to improve it; those statements describe the study rationale, not a conclusion from these results. The posted primary outcome is an AUC measure in nmol*h/L, based on serum concentrations observed at the stated post-dose times. The reported arm means differ in magnitude, and the unadjusted recalculations compare each intervention summary with the control summary. Because this is a crossover trial, comparisons may involve repeated observations from participants, but the registry record does not report the participant-level data or analysis details needed to assess that dependence. The findings therefore describe the posted serum AUC summaries and the precision of simple recalculations, not proof of a health benefit or a definitive ranking of food products.

3. Trial Design

The registry identifies a randomized, single-masked crossover trial titled “Physiological Effects of New Polyphenol-enriched Foods in Humans.” It lists the study condition as healthy and the phase as not applicable. The trial is marked completed. Its recorded start is 2010-12, and its primary completion is 2011-02. Federico II University is listed as sponsor.

The six conditions in the outcome results are free curcumin, encapsulated curcumin, encapsulated curcumin with PQG, control cream, free cocoa polyphenols, and encapsulated cocoa polyphenols. The intervention descriptions identify free curcumin as bread added with free curcumin and define PQG as Piperine, Quercetin and Genistein. The intervention list also calls the comparator a control nut cream. Further details about the foods and their preparation are not reported in the registry record.

A crossover model means, in general, that participants receive more than one study condition, so observations may be related within participant. The record does not report the order or sequence of conditions, the washout interval, how masking was implemented, or whether every participant contributed data to every condition. It records 10 participants in each arm-level outcome summary; that does not establish the details of individual participation across periods.

The registry describes the sample-size rationale as power calculations derived from a previous study. It states that, at α = 0.05 with power of 80%, 8 subjects would allow detection of a 20% difference in serum and urinary concentrations of parental compounds, glucuronides and phenolic acids. More detail about that earlier study or the calculation is not reported in the registry record. The primary outcome is serum polyphenol AUC over 24 hours. The registry lists three outcome measures and three analyses, but details of the other outcomes 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: Subjects were enrolled after signing a Consent Form and were randomized to the treatments. All subjects enrolled completed the study.

Free Then Encapsulated Curcumin (1 Day)

Washout Period of 1 Week

Encapsulated Then Free Curcumin (1 Day)

Washout Period of 1 Week

Analysis population for the primary outcome, as defined in the registry: The number of subjects was based on power calculations derived from our previous study. We calculated that, at α = 0.05 with a power of 80%, 8 subjects would allow us to detect a 20% difference in serum and urinary concentrations of parental compounds, glucuronides and phenolic acids.

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 Serum Polyphenol Concentrations Over 24h From Food Consumption, measured over 0, 0.5, 1, 2, 4, 6, and 24 hours post-dose.

The primary endpoint is the area under the serum concentration-time curve for parent polyphenols from 0 to 24 hours after food consumption. Concentrations were recorded at 0, 0.5, 1, 2, 4, 6, and 24 hours post-dose. The registry says the AUC was calculated using a trapezoidal rule applied to concentration-time curves. Its unit is nmol*h/L, and results are reported as means with standard errors.

As a general statistical concept, AUC combines information across measurements over time into a single summary of overall measured exposure during the specified observation window. It can be useful when a question concerns the concentration profile over time rather than a single time point. In this trial, the endpoint is intended to inform evaluation of polyphenol bioavailability, consistent with the stated study rationale. It remains a measure of serum exposure to parent compounds; it does not itself measure health benefits, long-term effects, or concentrations of every metabolite mentioned in the sample-size rationale.

The arm summaries report a mean AUC of 0.01 for control cream, 1.7 for free curcumin, 12.1 for encapsulated curcumin, 7.7 for encapsulated curcumin with PQG, 7.3 for free cocoa polyphenols, and 0.5 for encapsulated cocoa polyphenols. The standard errors are reported alongside these means. The record does not state whether the AUC calculation or analysis incorporated any additional adjustments.

7. Results Posted in the Registry

• Free Curcumin: mean 1.70 versus 0.01 on control; difference 1.69 (95% CI 1.49 to 1.89), standardized effect d = 7.56; z-test p = <0.001.
• Encapsulated Curcumin: mean 12.10 versus 0.01 on control; difference 12.09 (95% CI 9.93 to 14.25), standardized effect d = 4.92; z-test p = <0.001.
• Encapsulated Curcumin + PQG: mean 7.70 versus 0.01 on control; difference 7.69 (95% CI 4.16 to 11.22), standardized effect d = 1.91; z-test p = <0.001.
• Free Cocoa Polyphenols: mean 7.30 versus 0.01 on control; difference 7.29 (95% CI 5.53 to 9.05), standardized effect d = 3.62; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean difference describe?

For each intervention, the recalculation expresses the difference between that arm’s posted mean AUC and the control mean AUC. A positive estimate means the intervention’s posted mean is higher than the control’s posted mean on the recorded outcome scale. For example, the recalculated mean difference for encapsulated curcumin is 12.09 nmol*h/L, with a 95% confidence interval from 9.934 to 14.25. The corresponding estimate for encapsulated cocoa polyphenols is 0.49, with a 95% confidence interval from 0.294 to 0.686. These are simple unadjusted calculations from the posted arm summaries; they are not estimates from the trial’s reported ANOVA.

As a general statistical explanation, a mean difference keeps the comparison on the original measurement scale. It is distinct from a standardized effect, which expresses a difference relative to a measure of variability. Neither form of effect size establishes clinical importance on its own. The registry record does not define a clinically meaningful difference for this AUC outcome.

What do the standard errors indicate?

The arm results report means and standard errors. A standard error describes the estimated sampling uncertainty of a mean, under the assumptions of its calculation; it is not the spread of individual participant values. The standard errors reported for the arms are 0.1 for free curcumin, 1.1 for encapsulated curcumin, 1.8 for encapsulated curcumin with PQG, 0.0 for control cream, 0.9 for free cocoa polyphenols, and 0.1 for encapsulated cocoa polyphenols. These values are reported as rounded in the registry record. In particular, a displayed standard error of 0.0 should not be read as proof that all control observations were identical or that there was no uncertainty.

As a general concept, uncertainty in a difference depends on the uncertainty in both means and, for repeated measurements, on how the observations are related. In a crossover study, the same participant may contribute observations under multiple conditions. A paired analysis can use that within-participant relationship. The record does not include participant-level data or the covariance between condition measurements, so the recalculations cannot use that information. This is a material distinction when reading confidence intervals and p-values from simple summaries.

How should the confidence intervals and p-values be read?

The recalculated 95% confidence intervals give ranges of values compatible with the estimated mean differences under the recalculation’s statistical assumptions. A confidence interval is not a probability statement that the true value lies in the particular interval, nor does it capture every source of uncertainty, such as uncertainty caused by design details that are not reported. The recalculated p-values are unadjusted and are reported as <0.001 for each intervention-versus-control comparison. A p-value describes the compatibility of data with a specified null model under the test assumptions; it is not the probability that the null hypothesis is true, nor a measure of effect importance.

The registry’s own analysis is described as ANOVA across all six groups, with mean difference in final values, a reported value of 1, and p = 0.02. Its type is labeled non-inferiority or equivalence. No confidence interval is reported for it, and the registry record does not describe the margin or the detailed model. It cannot be reconstructed as a direct equivalent of the recalculated intervention-versus-control estimates. The record also does not report whether that analysis accounts for crossover periods or within-participant correlation.

Why does the crossover model matter?

As a general statistical explanation, crossover designs can control for stable differences between participants because participants receive multiple conditions. They also require suitable handling of repeated measurements and may be affected by sequence, period, or carryover effects. These are design and analysis considerations, not findings about this trial. The registry record identifies the model as crossover but does not report sequence allocation, period-specific results, a washout interval, or how such effects were addressed. Without these details, readers cannot determine whether the reported ANOVA or the summary-based recalculations reflect the intended crossover analysis.

General statistical reasoning also explains why over-dispersion matters in other settings: when outcome variability exceeds what a chosen model assumes, standard errors can be too small and inference can become overconfident. The posted endpoint is continuous, and the record gives standard errors rather than individual values, so over-dispersion cannot be assessed from the available summaries. No claim about over-dispersion in this trial is warranted.

What does multiplicity change?

Each recalculated p-value addresses an individual comparison with control. When multiple comparisons are examined, the chance of at least one apparently unusual result under a collection of null hypotheses can increase. A multiplicity adjustment is one approach to controlling that issue. The registry record does not report whether the sponsor adjusted for multiple comparisons, and the recalculated p-values are unadjusted. Their small displayed values should therefore be read in the context of multiple intervention comparisons and the missing analysis details, not as a substitute for a prespecified family-wise or other multiplicity plan.

The sample-size planning figures are illustrative calculations based on observed effects, not results from the trial’s original planning process. They use a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect set to a stated fraction of the observed effect. As a general concept, power is the probability of detecting an effect of a specified size under the assumed model if that effect is present. These planning calculations are not observed power and do not validate the trial’s analysis.

10. Confidence Intervals and Precision

The recalculated mean differences are accompanied by 95% confidence intervals. For free curcumin the estimate is 1.69, with an interval from 1.494 to 1.886. For encapsulated curcumin it is 12.09, with an interval from 9.934 to 14.25. For encapsulated curcumin with PQG it is 7.69, with an interval from 4.162 to 11.22. For free cocoa polyphenols it is 7.29, with an interval from 5.526 to 9.054. For encapsulated cocoa polyphenols it is 0.49, with an interval from 0.294 to 0.686. Each interval is a summary of statistical uncertainty under the recalculation’s assumptions.

As a general explanation, narrower confidence intervals indicate greater precision for the estimated quantity under the model, while wider intervals indicate more uncertainty. Precision is not the same as accuracy, and an interval does not account for limitations that the statistical model omits. Here, the intervals are based on arm-level means and standard errors. The registry identifies a crossover design, but the summaries do not report the within-participant covariance needed to reconstruct a paired comparison. The recalculated intervals therefore do not establish the precision of an analysis that appropriately models the crossover data.

The control mean is 0.01 with a standard error reported as 0.0. Since the values are rounded in the registry record, that displayed standard error does not show the underlying unrounded uncertainty. The recalculated intervals are likewise presented at the precision already recorded; no extra digits should be inferred. The sponsor’s reported ANOVA has no confidence interval in the registry record, so its precision cannot be evaluated from an interval estimate.

All five recalculated intervals lie above zero. That is a feature of these particular unadjusted calculations, not a guarantee that the same inference would result from a model accounting for crossover structure or multiplicity. Statistical precision also does not determine whether a difference matters clinically. The record does not report a threshold for a meaningful AUC difference.

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 expressed as participants per arm, not as total event counts. They apply to a comparison between two arms under the stated planning assumptions: two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at 100%, 75%, or 50% of the observed effect for that comparison. The listed values are planning illustrations computed from the observed effect, not observed power.

The 80% power column gives the number of participants per arm used in the calculation for an 80% chance of detecting the specified effect under those assumptions, if that effect is present. The 90% power column gives the corresponding planning count for a 90% chance. These are hypothetical design calculations, not counts observed in the trial. The figures can be very small when the calculation uses a large observed effect, as in some of the comparisons. They should not be read as evidence that the trial’s actual design was adequate, especially because the calculations use an observed effect and the registry identifies a crossover model.

The entries illustrate the expected direction of the relationship between effect size and required sample size: planning for 75% or 50% of an observed effect generally requires more participants than planning for the full observed effect. Smaller effects are harder to distinguish from random variation, so more participants are generally needed to attain the same target power, all else equal. These illustrations do not reproduce the original study design. The registry separately states that its own sample-size rationale was based on prior work and described 8 subjects as sufficient, at α = 0.05 and power of 80%, to detect a 20% difference in specified serum and urinary concentrations. Further details of that original calculation are not reported in the registry record.

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 = 1.690; pooled standard deviation σ = 0.224.
2. n = 2 × (1.96 + 0.842)² × 0.224² / 1.690² = 0.3, rounded up to 1 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 <- 1.69                      # observed difference in means
sigma <- 0.223606797749979                     # 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(1288859)
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 = 1 (80% power), 1 (90% power)
 75% of observed effect: n per arm = 1 (80% power), 1 (90% power)
 50% of observed effect: n per arm = 2 (80% power), 2 (90% power)
power.t.test: 1.8 
Simulated power at n = 1 per arm: NA (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 registry reports a sponsor analysis using ANOVA across the six listed groups. Its parameter is described as mean difference in final values, its reported value is 1, its p-value is 0.02, and its analysis type is labeled non-inferiority or equivalence. The registry record does not provide a confidence interval, a non-inferiority or equivalence margin, or enough detail to identify a specific pairwise contrast represented by the reported value.

The recalculated results instead compare each intervention’s posted mean AUC with the control mean using the arm-level summaries. They give five unadjusted mean differences, confidence intervals, and p-values reported as <0.001. Because the reported analysis and the recalculations are not described as the same estimand or model, they should not be treated as competing estimates of one clearly defined contrast. The reported ANOVA result does not supply the missing details needed to map it directly to each comparison.

The registry record does not flag a discrepancy between the posted analysis and the recalculations. The recalculations remain secondary, unadjusted summaries and do not replace the sponsor’s analysis. Neither result resolves the absence of crossover-specific model details, multiplicity information, or a reported confidence interval for the sponsor analysis.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists three outcome measures and three analyses, while the posted primary outcome is serum polyphenol AUC over 24 hours. It does not report the complete set of secondary outcome results; the number of secondary outcomes posted is recorded as 0. The detailed analysis plan, definitions of analysis populations, and rules for handling missing observations are not reported in the registry record.

Multiplicity is relevant because the recalculations make five comparisons with the control. Each reported recalculated p-value is unadjusted and is <0.001. As a general statistical concept, testing multiple hypotheses can increase the chance of a false positive across the collection of tests if each is interpreted separately at the same threshold. Study plans may address this by identifying a primary comparison, using a hierarchical testing approach, or adjusting the threshold. Which approach, if any, was planned or used for the sponsor’s analyses is not reported in the registry record. The unadjusted recalculations should be interpreted accordingly.

The record identifies randomized allocation, single masking, and a crossover model, but does not describe the analysis population, exclusions, missing-data method, period effects, sequence effects, carryover assessment, or the model used to account for repeated participant measurements. These omissions matter because missing measurements and within-participant dependence can affect estimates and their uncertainty. The reported arm summaries do not show how many observations were missing at particular time points or whether all participants contributed to each AUC summary.

The listed sponsor analysis is an ANOVA across all six groups, described as mean difference in final values and labeled non-inferiority or equivalence. Its statistical margin, confidence interval, and detailed model are not reported. The record also does not state whether the sponsor analysis included all randomized participants, only participants with complete data, or another analysis population. Consequently, the posted results do not allow a full audit of the analysis strategy or a determination of how missing data and multiple comparisons were handled.

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 0% reporting threshold.

The registry reports zero serious adverse events in each listed intervention group. For each group, the number at risk for serious adverse events is 10 and the number affected is 0. It also reports zero other adverse events in each group, with 10 at risk and 0 affected. The deaths fields are not populated; deaths are not reported in the registry record.

As a general statistical point, a count of zero observed events does not establish that an event cannot occur or that an intervention is safe in every setting. Adverse-event counts are observations from the recorded study groups and should be read alongside the number at risk, the duration and completeness of follow-up, and the way events were collected. Details about those follow-up and ascertainment procedures are not reported in the registry record.

The serious adverse event figures are not evidence about long-term safety, uncommon events, or outcomes outside the trial conditions. The record concerns healthy participants and does not provide a detailed account of event definitions or the safety-analysis method. These counts describe what was recorded, without supporting broader safety conclusions.

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

The posted results show that the reported mean serum AUC for parent polyphenols differed across the listed food conditions. The arm-level means range from 0.01 for control cream to 12.1 for encapsulated curcumin; the other intervention means are 1.7 for free curcumin, 7.7 for encapsulated curcumin with PQG, 7.3 for free cocoa polyphenols, and 0.5 for encapsulated cocoa polyphenols. These values describe measured serum exposure over the specified 24-hour period, in nmol*h/L.

The recalculated comparisons with control have positive mean differences and 95% confidence intervals above zero, with unadjusted p-values of <0.001. Those results are simple calculations from the posted summaries. They do not reproduce the registry’s ANOVA analysis and do not account for the participant-level correlations that may arise in a crossover design. The registry does not report enough detail to determine how those correlations, period effects, sequence, or missing observations were handled in the sponsor’s analysis.

The endpoint is a pharmacokinetic-style measure of serum exposure, not a direct measure of clinical benefit. The study rationale discusses bioavailability, but the record does not report health outcomes that would show whether a difference in serum AUC leads to a health effect. Nor does the endpoint establish effects beyond the measured observation window or the specific food conditions studied.

The study is described as involving healthy participants, with enrollment recorded as 10 and 10 participants in each arm-level summary. This record does not establish results for other populations or settings. The absence of reported adverse events in the listed groups is also not proof that risks are absent. Overall, the data support description of the posted AUC summaries and their unadjusted recalculated contrasts; they do not by themselves establish clinical importance, generalizability, or a definitive comparison under a fully documented crossover analysis.

17. Why This Trial Matters Statistically

This trial raises several useful statistical issues in a compact setting. First, it combines a continuous concentration-time outcome with an AUC summary. AUC compresses measurements taken over time into a single measure, which can make conditions easier to compare, but it also means that the result depends on how the concentration profile is measured and integrated. The registry specifies the sampling times and a trapezoidal calculation, giving readers a basic account of the endpoint’s construction.

Second, the stated crossover model matters for interpreting uncertainty. Repeated observations from the same participant can be related, and methods that use that relationship may differ from calculations based on separate group summaries. The registry posts means and standard errors but not participant-level data or the covariance across conditions. That makes it possible to calculate simple comparisons, but not to reconstruct a fully specified crossover model from the record.

Third, there is a meaningful distinction between the sponsor’s reported ANOVA and the unadjusted recalculations. The registry’s reported analysis has a p-value of 0.02, while each recalculated comparison with control has a p-value of <0.001. The analyses are not documented as testing the same contrast or using the same model. Keeping their interpretations separate is more informative than treating the numbers as contradictory.

Finally, the example illustrates that statistical evidence and clinical meaning are different questions. A serum AUC difference can inform evaluation of exposure, while the record does not report clinical outcomes or a threshold for a meaningful AUC difference. The small participant counts, missing crossover-analysis details, and unreported multiplicity and missing-data procedures limit what can be inferred. These are general lessons about how to read trial summaries, not claims about effects beyond the posted outcome.

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.

Several trial-specific limitations constrain interpretation. The registry records enrollment of 10, and each arm-level outcome summary reports 10 participants. The record does not describe participant characteristics beyond the healthy condition, limiting assessment of how the results relate to other populations. It also does not report individual participant measurements, which prevents independent review of the concentration-time profiles or AUC calculations.

The crossover model is stated, but key design details are not reported in the registry record. These include treatment sequence, period structure, washout interval, and whether carryover or period effects were considered. The identity of the person or group masked under the single-masked design is also not reported. Without these details, the relationship between the design and the sponsor’s ANOVA cannot be fully assessed.

The endpoint is a serum AUC of parent polyphenols over 24 hours. The record does not provide a clinical outcome, a clinically meaningful AUC threshold, or results beyond that window. It therefore cannot establish whether the measured exposure differences correspond to health effects. The intervention descriptions are also limited: free curcumin is specified as added to bread, while other food-matrix details are not reported in the registry record. The control is called control cream in the results and control nut cream in the intervention list.

Further limitations include the absence of a reported confidence interval for the sponsor’s ANOVA, an unreported analysis population and missing-data method, and no reported multiplicity strategy. The recalculated contrasts use arm summaries and do not incorporate within-participant covariance. The reported zero adverse-event counts do not resolve uncertainties about event ascertainment or follow-up details. These limitations do not negate the posted measurements, but they restrict conclusions about precision, reproducibility, comparative effects, and generalizability.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Physiological Effects of New Polyphenol-enriched Foods in Humans — NCT01288859.
• Registry record last updated 2012-08-01; status last verified 2012-07.
• Results first posted 2012-08-01.
• 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 describes a completed randomized, single-masked crossover study of polyphenol-enriched foods in healthy participants, sponsored by Federico II University. Enrollment is recorded as 10. The primary outcome is mean serum AUC of parent polyphenols from 0 to 24 hours after food consumption, calculated by the trapezoidal rule and reported in nmol*h/L with standard errors.

The arm summaries report mean AUC values of 1.7 for free curcumin, 12.1 for encapsulated curcumin, 7.7 for encapsulated curcumin with PQG, 0.01 for control cream, 7.3 for free cocoa polyphenols, and 0.5 for encapsulated cocoa polyphenols. The sponsor reports an ANOVA result with p = 0.02. Unadjusted recalculations compare each intervention with control and report p-values of <0.001, but cannot incorporate unreported crossover-level data or analysis details.

The record does not report the full analysis plan, multiplicity procedure, missing-data method, or key crossover details. The outcomes describe serum exposure; they do not establish clinical benefit.