Efficacy Study of Activated Prothrombin Complex for Prevention of Bleeds in Hemophilia A With Inhibitors: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy Study of Activated Prothrombin Complex for Prevention of Bleeds in Hemophilia A With Inhibitors (NCT00221195), a randomized phase 2, phase 3 study. The primary outcome is Number of Bleeds During 6 Month Treatment Period.

2. Clinical Question

3. Trial Design

This was a randomized crossover study listed as Phase 2 and Phase 3. The registry lists an enrollment of 34 and identifies Tulane University School of Medicine as the sponsor. The study began in 2003-06, had primary completion in 2009-12, and is recorded as completed. Masking was none.

Participants were assigned to one of two treatment orders. In the “On-demand first” order, participants received six months of on-demand study drug followed by six months of prophylaxis. In the “Prophylaxis first” order, participants received six months of prophylaxis followed by six months of on-demand treatment. The adverse-event record also lists a washout period. Its length and the precise timing of assessments are not reported in the registry record.

A crossover design exposes participants to both treatment approaches at different times. In general, this can allow comparisons that account for stable differences between participants, provided the analysis uses the within-participant structure appropriately. The registry’s posted outcome summaries do not identify sequence-specific results or explain whether the analysis accounted for pairing, treatment order, or possible period effects.

The primary outcome measure was the number of bleeds during a six-month treatment period. The registry reports 26 participants for each treatment-period outcome group. It does not explain why this number differs from the enrollment of 34, how many participants completed both periods, or how participants were handled if they did not contribute outcome data for both periods. The original design assumptions, including any planned sample-size justification, 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: Patients were randomized to one of two 6 month treatment periods. Following the initial treatment period each subject had a 3 month wash-out period during which time they used on-demand therapy to treat bleeding. Following the 3 month wash-out period each patient was crossed-over to the other 6 month treatment period.

Therapy 1 (Months 1-6)

Reasons for not completing this period:

Washout Period (Months 7-9)

Reasons for not completing this period:

Therapy 2 (Months 10-15)

Reasons for not completing this period:

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 Bleeds During 6 Month Treatment Period, measured over 6 months.

The primary endpoint is the number of bleeds during a six-month treatment period. It is a count of bleeding events over a defined period, summarized in the registry as a mean with a standard deviation. The posted mean was 13.1 bleeds during the on-demand period and 5.0 during the prophylaxis period. The corresponding standard deviations were 7.1 and 5.0.

Counting bleeds is directly related to the study’s stated objective: assessing whether prophylactic therapy was associated with a reduction in the number of bleeding events. Using the same six-month duration for both approaches makes the treatment-period summaries interpretable on a common time basis. In general, a mean describes the average count in a group, while a standard deviation describes how spread out individual counts are around that average. Neither summary shows the full distribution of bleed counts or how individual participants’ counts changed between periods.

The record identifies one primary outcome measure and one posted analysis. It does not provide a more detailed definition of which events qualified as bleeds, how events were confirmed, or how recurrent events were counted. Those endpoint details are not reported in the registry record.

7. Results Posted in the Registry

• Bleeds During the Prophylaxis Period: mean 5.00 versus 13.10 on control; difference -8.10 (95% CI -11.44 to -4.76), standardized effect d = -1.32; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods help distinguish what the observed summaries say from what can be inferred about the treatment comparison. The registry reports a mean and standard deviation for each treatment-period group, a sponsor analysis using Wilcoxon (Mann-Whitney), and a separate unadjusted recalculation based on summary statistics. These are related but not interchangeable descriptions of the data.

What do the means and standard deviations describe?

A mean is the arithmetic average of the recorded bleed counts in a group. The on-demand period mean was 13.1 and its standard deviation was 7.1. The prophylaxis period mean was 5.0 and its standard deviation was 5.0. A standard deviation describes the spread of observations around a mean; it is not the uncertainty in the mean itself, and it is not a range containing all participants’ results. In general, count outcomes can be unevenly distributed, with some participants having substantially more events than others. The registry does not provide the individual counts or distributional summaries that would show the shape of the data.

What does the sponsor-reported test assess?

The reported method is Wilcoxon (Mann-Whitney), a rank-based test. In general, this test compares the relative ordering of observations between groups rather than directly testing a difference between arithmetic means. Its interpretation depends on the study design and the structure of the observations. The registry gives p < 0.001, but does not report the test statistic, a test-based effect estimate, or a test-based confidence interval.

The crossover design makes the dependence between treatment-period observations especially relevant. If each participant contributed data under both approaches, those observations are paired. A method that treats groups as independent may not use that pairing, while a paired method compares participants’ outcomes across their own periods. The registry does not report whether the sponsor’s analysis accounted for pairing, or whether the outcome groups represent the same participants. It is therefore not possible to determine from the posted description whether the reported test was adapted to the crossover structure.

What is the unadjusted recalculation?

The unadjusted recalculation compares the posted group means. It gives a mean difference of -8.1 for prophylaxis relative to on-demand treatment, with an interval from -11.44 to -4.762. The standard error summaries are 1.392 for the on-demand mean and 0.981 for the prophylaxis mean. These calculations use the aggregate outcome summaries and do not incorporate participant-level pairing, sequence, or period information. The recalculated result is therefore not a substitute for a crossover analysis using the individual records.

In general, a confidence interval describes a range of values compatible with the data and the assumptions of the calculation. A narrower interval indicates greater precision under those assumptions; it does not address bias from design or analysis choices. Here, the recalculated interval is for an unadjusted mean comparison, and the registry does not provide a corresponding interval from the sponsor’s rank-based analysis.

Why can count data require care?

General statistical consideration: event counts can have more variability than a simple model assumes, a feature often called over-dispersion. This can occur when event rates differ across participants or when counts cluster. A model that does not account for such extra variability may estimate uncertainty too narrowly. The registry does not report a count-regression model, an over-dispersion assessment, or another model-based analysis. The posted standard deviations give some information about spread, but the aggregate summaries are not enough to assess the distribution or choose a count model.

The reported p-value and the recalculated interval answer different questions and rest on different methods. The p-value is the sponsor-reported test result. The interval and mean difference are unadjusted recalculations from the posted summaries. Reading them together can clarify the observed scale and statistical evidence, but neither resolves missing details about pairing, period effects, carryover, or which participants contributed to both treatment-period results.

10. Confidence Intervals and Precision

The unadjusted recalculated mean difference is -8.1 bleeds for prophylaxis relative to on-demand treatment, with an interval from -11.44 to -4.762. The point estimate summarizes the difference between the posted means; the interval expresses uncertainty around that unadjusted comparison under its calculation assumptions. It excludes zero, while the sponsor-reported test result is p < 0.001. The registry does not report a confidence interval for the sponsor’s Wilcoxon (Mann-Whitney) analysis.

In general, a confidence interval is a range of estimates compatible with the observed data under a specified statistical method. It is not a guarantee that the true effect lies within the interval, and it does not account for sources of uncertainty omitted from the method. Here, the recalculation uses group-level means and variability and does not reflect the crossover pairing or adjust for treatment order or period. Its precision should therefore be interpreted as precision for that unadjusted comparison, not necessarily for a design-adjusted treatment effect.

The standard deviations of 7.1 and 5.0 indicate substantial variation in the posted bleed counts relative to their respective means. However, standard deviations do not reveal the full distribution, and the aggregate report does not show participant-level changes. The number contributing to each outcome group is 26; the registry does not explain missing outcome data or whether the same participants contribute to both groups. These reporting gaps limit assessment of the effective information available for a paired comparison.

The interval gives more information about the range and scale of the unadjusted mean difference than a p-value alone, but it cannot resolve design and reporting uncertainties. A paired estimate and its uncertainty, together with a clear account of the analysis population, are not reported in the registry record.

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 based on the observed effect, rather than the trial’s original sample-size plan. The calculation assumes a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It considers effects set at 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect.

The 80% power column gives the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect, if the assumptions hold. The 90% power column gives the corresponding number per arm for a 90% chance. These are planning targets for a hypothetical study, not statements about the probability that the completed study’s result is correct.

For 100% of the observed effect, the illustration gives 10 participants per arm for 80% power and 13 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 37 participants per arm for 80% power and 49 participants per arm for 90% power. The pattern reflects a general principle: smaller effects are harder to distinguish from random variation, so detecting them requires more participants.

These figures are not observed power. They are calculated from the observed effect as planning illustrations, with the stated assumptions. They do not establish what sample size the investigators originally planned, and they do not incorporate unreported features of the crossover design or any paired analysis. The trial’s own original design assumptions and sample-size justification 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 = -8.100; pooled standard deviation σ = 6.140.
2. n = 2 × (1.96 + 0.842)² × 6.140² / -8.100² = 9.0, rounded up to 10 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 <- -8.1                      # observed difference in means
sigma <- 6.140439723668                     # 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(221195)
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 = 10 (80% power), 13 (90% power)
 75% of observed effect: n per arm = 17 (80% power), 22 (90% power)
 50% of observed effect: n per arm = 37 (80% power), 49 (90% power)
power.t.test: 10.1 
Simulated power at n = 10 per arm: 0.795 (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’s reported analysis is Wilcoxon (Mann-Whitney), with p < 0.001. It does not give a test statistic, an effect estimate from that test, or its confidence interval. The unadjusted recalculation is a separate comparison of the posted means: it estimates a mean difference of -8.1 for prophylaxis relative to on-demand treatment, with an interval from -11.44 to -4.762.

The two results should not be described as if they were the same analysis. The sponsor’s test is rank-based, whereas the recalculated estimate is a difference between arithmetic means. The registry does not report whether the sponsor’s test accounted for the paired structure expected when participants receive both treatments, or whether it incorporated treatment sequence or period.

The registry’s reported p-value and the recalculated mean comparison point in the same direction in the posted summaries, but the recalculation does not replace the sponsor’s analysis. The registry does not identify a discrepancy between the two. Because the methods and estimands differ, the mean difference and interval should be presented as supplementary unadjusted calculations, not as a correction to the reported test.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and one posted analysis, and it lists no secondary outcome results. It does not report whether additional outcomes or analyses were planned, whether any interim analyses were conducted, or whether the primary analysis was prespecified in a statistical analysis plan. Those details are not reported in the registry record.

Multiplicity arises when several statistical tests or outcomes are examined: the more opportunities there are to find a small p-value, the greater the chance of a false positive across the collection of tests. General statistical practice may address this through prespecified outcome hierarchies or adjustments, but the registry does not describe any multiplicity strategy. Since only one primary result is posted, the record does not allow assessment of a broader set of analyses that may or may not have been conducted.

Missing-data handling is also not described. Enrollment was 34, while each posted treatment-period outcome group has 26 participants. The registry does not state why these counts differ, whether participants withdrew, how many completed both treatment periods, or how missing bleed records were handled. It does not report an intention-to-treat or per-protocol analysis population. These details matter because excluding participants or periods with incomplete data can affect both the estimated treatment difference and its uncertainty.

In a crossover study, the analysis population and the handling of incomplete treatment sequences are particularly important. The registry does not say whether participants with results from only one period were included, whether the analysis required complete paired data, or how sequence and period were handled. It also does not report whether potential carryover was assessed. These omissions prevent a full evaluation of whether the posted analysis corresponds to the randomized comparison and how robust the result is to missing observations.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: muscle hemorrhage (6.9%), catheter-site infection (6.5%), cerebral hemorrhage (3.4%), subdural hematoma (3.4%), chest pain (3.2%).

The registry lists serious adverse events for three periods. During the on-demand period, 3 of 31 participants at risk were affected by a serious adverse event. During the prophylaxis period, 4 of 31 were affected. During the washout period, 4 of 29 were affected. These are period-specific counts and should not be added together as if they represented unique participants across the study; the same participant could contribute to more than one period.

The record lists serious adverse-event terms including Muscle Hemorrhage (6.9%), Catheter-site Infection (6.5%), Cerebral Hemorrhage (3.4%), Subdural Hematoma (3.4%), Chest Pain (3.2%), and Pain (3.2%). These percentages are presented as recorded and are not recalculated. The registry does not provide enough detail in these summary figures to determine the timing, duration, outcome, or relationship to treatment for each event.

Deaths are not reported in the registry record. Serious adverse events and other adverse events are distinct categories; the figures above describe those recorded as serious and do not provide a complete account of all possible harms. These data are descriptive and do not by themselves establish whether an event was caused by a treatment or whether the periods differ in safety.

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

The posted results show a lower mean number of bleeds during the prophylaxis period than during the on-demand period: 5.0 compared with 13.1 over six months. The sponsor-reported Wilcoxon (Mann-Whitney) analysis gives p < 0.001. An unadjusted recalculation gives a mean difference of -8.1, with an interval from -11.44 to -4.762. Together, these summaries describe a difference in the observed treatment-period results.

They do not, by themselves, show the participant-level pattern of change. The design is crossover, but the registry does not report whether the analysis used paired data, whether the same participants contributed to both outcome groups, or how treatment sequence and period were handled. The outcome-group counts of 26 are smaller than the enrollment of 34, and the reason for this is not reported in the registry record. These details limit how fully the posted group comparison can be interpreted as a crossover treatment effect.

The result also does not explain why the observed difference occurred, establish that the same difference would appear in another population, or quantify the effect of unreported design and analysis choices. The recalculated interval concerns an unadjusted mean comparison; it is not a substitute for a design-aware estimate. The reported p-value is evidence from the stated test, but does not measure the size or clinical importance of a difference.

The study’s safety results are recorded separately. Period-specific serious adverse-event counts do not establish causation and should not be interpreted as a complete safety comparison. Secondary outcome results are not posted. Overall, the registry documents lower observed mean bleed counts during prophylaxis and a statistically significant sponsor-reported test, while leaving important details of the crossover analysis, missing data, and outcome reporting unclear.

17. Why This Trial Matters Statistically

This trial illustrates why the study design and the analysis method need to be read together. In a crossover study, participants receive both approaches in different periods. The analysis can potentially use within-participant comparisons, but only if the data and method preserve that pairing. The registry posts treatment-period group means and a Wilcoxon (Mann-Whitney) result without explaining how the crossover structure was incorporated.

It also shows why different statistical summaries should not be treated as interchangeable. A rank-based test and a difference between arithmetic means answer related but distinct questions. The sponsor-reported test gives p < 0.001, while the unadjusted recalculation describes the mean difference as -8.1 with an interval from -11.44 to -4.762. Reporting both can be informative, but it is essential to identify which result came from the registry analysis and which is an unadjusted recalculation.

The discrepancy between enrollment of 34 and outcome groups of 26 highlights the importance of transparent participant flow and analysis populations. Without an account of missing periods and the participants contributing to each summary, readers cannot determine whether the posted comparison uses complete paired observations. The registry also does not describe missing-data methods, sequence or period adjustments, or assessment of carryover.

Finally, planning illustrations based on an observed effect are not the same as the study’s original design justification. Their assumptions can show how the required sample size changes when the target effect changes, but they do not recover the investigators’ planning process. These distinctions are central to interpreting trials with repeated treatment periods and aggregate outcome reporting.

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 main limitation in interpreting this trial is the limited description of the crossover analysis. Although participants were randomized to treatment order and crossed between on-demand treatment and prophylaxis, the registry does not report whether the primary comparison accounted for paired observations, sequence, period, or carryover. The sponsor-reported Wilcoxon (Mann-Whitney) method is named, but its implementation for this design is not explained.

The registry lists enrollment of 34 but reports 26 participants in each treatment-period outcome group. It does not explain the difference, state whether the same participants contributed to both groups, or describe how incomplete treatment periods and missing bleed records were handled. There is no reported analysis-population definition or participant-flow explanation in the posted results.

The outcome is summarized using means and standard deviations, but individual bleed counts and distributional details are not available in the registry record. This limits assessment of skewness, unusually high counts, and the appropriateness of a mean-based comparison. The recalculated mean difference and interval are unadjusted and do not incorporate crossover pairing.

The registry does not provide a detailed definition of a bleed, a statistical analysis plan, a sample-size justification, interim-analysis information, or a multiplicity strategy. It lists no secondary outcome results. Safety reporting is period-based and does not provide event-level detail or deaths; deaths are not reported in the registry record. These reporting limitations restrict assessment of both the efficacy analysis and the completeness of the safety picture.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Efficacy Study of Activated Prothrombin Complex for Prevention of Bleeds in Hemophilia A With Inhibitors — NCT00221195.
• Registry record last updated 2020-03-02; status last verified 2020-02.
• Results first posted 2017-02-06.
• 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, unmasked crossover study listed as Phase 2 and Phase 3, sponsored by Tulane University School of Medicine. Enrollment was 34. Participants received six months of on-demand study drug and six months of prophylaxis in randomized order, with a washout period also listed in the adverse-event record.

The primary outcome was the number of bleeds during a six-month treatment period. The posted means were 13.1 during on-demand treatment and 5.0 during prophylaxis, with 26 participants in each outcome group. The sponsor-reported Wilcoxon (Mann-Whitney) analysis gave p < 0.001. The unadjusted recalculated mean difference was -8.1, with an interval from -11.44 to -4.762.

The registry does not explain the difference between enrollment and the outcome-group counts or describe how pairing, missing data, sequence, or period were handled in the analysis. These limitations are important when interpreting the posted results from a crossover trial.