Apixaban for the Treatment of Venous Thromboembolism in Patients With Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Apixaban for the Treatment of Venous Thromboembolism in Patients With Cancer (NCT03045406), a randomized phase 3 study. The primary outcome is Recurrent Venous Thromboembolism.

2. Clinical Question

The outcome combined several kinds of recurrent events: proximal lower-limb deep vein thrombosis, whether symptomatic or unsuspected; symptomatic upper-limb deep vein thrombosis; and pulmonary embolism, whether symptomatic or unsuspected. The registry record does not report separate counts for these components, so the reported comparison concerns the composite rather than any single event type. The record also does not report the detailed criteria used to confirm events or the schedule and procedures for outcome assessment.

3. Trial Design

The registry describes a Phase 3, randomized, parallel-group trial comparing apixaban with dalteparin in patients with cancer and venous thromboembolism. Enrollment is listed as 1170. The outcome data give denominators of 576 for apixaban and 579 for dalteparin; the registry record does not explain the difference between the enrollment figure and those outcome denominators. The design is described as open, with blinded endpoint assessment in the official title. At the same time, the registry masking entry is none, and further operational details of endpoint blinding are not reported in the registry record.

The treatment schedules differed in route and dosing. Apixaban was administered orally, with a 10 mg twice-daily initial period for 7 days followed by 5 mg twice daily through six months. Dalteparin was administered subcutaneously once daily, initially at 200 IU/kg for 1 month and then at 150 IU/kg for 5 months. The registry lists Fadoi Foundation, Italy, as sponsor and gives the status as completed. The listed start date is 2017-04-13 and the primary completion date is 2019-06-01.

The primary outcome was recurrent venous thromboembolism during the 6-month timeframe. The registry reports one outcome measure and no posted secondary outcomes. It does not report the allocation ratio, randomization sequence method, concealment procedures, site structure, eligibility details, adherence, treatment discontinuation, or follow-up completeness. Although the official title names a prospective randomized open blinded end-point study, the record does not provide enough detail to assess how endpoint review was organized or whether assessors remained unaware of treatment assignment. These omissions limit interpretation of design implementation; they do not change the event counts and comparison that are posted.

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.

Population: 9 and 6 patients for the Apixaban and Dalteparin group, respectively, did not received the assigned treatment.

6. Primary Endpoint

The registry lists the primary outcome as Recurrent Venous Thromboembolism, measured over 6 months.

The primary endpoint counted participants with objectively confirmed recurrent venous thromboembolism during the 6-month study period. It was a composite outcome: proximal deep vein thrombosis in the lower limbs, symptomatic or unsuspected; symptomatic deep vein thrombosis in the upper limb; and pulmonary embolism, symptomatic or unsuspected. The registry reports counts of participants with an event, rather than a time-to-event summary or a count of every recurrence.

A composite endpoint can capture clinically relevant events across more than one manifestation of the condition and provide a single primary outcome for comparison. Its overall result, however, does not show whether the component event types contributed similarly. Component-specific counts and estimates are not reported in the registry record. The record also does not report whether multiple events in one participant were counted once or describe the event-adjudication criteria beyond stating that recurrence was objectively confirmed.

The posted endpoint counts are 32 participants among 576 in the apixaban group and 46 among 579 in the dalteparin group. They describe the observed number of participants meeting the composite definition. They do not alone account for the uncertainty in the between-group contrast, differences in follow-up, or the clinical implications of particular component events. The comparison estimate and its confidence interval are needed to understand how precisely the registry results distinguish possible group differences.

7. Results Posted in the Registry

• Dalteparin: 7.9% (46/579) versus 5.6% (32/576) on control; risk difference 2.4 percentage points (95% CI -0.5 to 5.3), risk ratio 1.43 (0.92–2.21); unadjusted two-proportion z-test p = 0.106.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry outcome is binary: each participant is classified according to whether recurrent venous thromboembolism occurred during the specified timeframe. The posted arm values are counts of participants with an event, alongside the number of participants in each arm. The comparison therefore concerns event proportions across groups, although no additional participant-level data are available in the registry record.

What does the risk difference mean?

In general, a risk difference subtracts the event proportion in one group from the event proportion in the other. The reported contrast is dalteparin minus apixaban: 0.024. A value of zero would correspond to equal event proportions, while positive and negative values indicate different directions under this subtraction order. The interval from -0.005 to 0.053 represents uncertainty around that estimate. Because zero falls within the interval, the results do not clearly distinguish a nonzero difference from no difference.

How do relative measures differ?

In general, a risk ratio compares event proportions by division, and a value of 1 corresponds to equal proportions. An odds ratio compares the odds of an event, rather than the proportions themselves; its no-difference value is also 1. The reported risk ratio was 1.43, with an interval from 0.924 to 2.212. The reported odds ratio was 1.467, with an interval from 0.92 to 2.34. Both intervals include their no-difference value. These relative measures and the risk difference express the comparison on different scales, so their numerical values should not be treated as interchangeable.

What does the p-value add?

In general, a p-value describes how unusual data at least as incompatible with a specified null model would be if that model were true. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. The unadjusted p-value is 0.106. Considered alongside the confidence intervals, it indicates that the posted comparison does not provide a precise statistical distinction from no difference. The registry does not report a formal analysis plan or a prespecified significance threshold for the trial, so the p-value should not be presented as proof of a particular clinical conclusion.

Why does the analysis scale matter?

General statistical interpretation depends on the outcome definition and the estimand, meaning the precise treatment contrast the analysis seeks to describe. Here, the endpoint counts participants with at least one qualifying composite event over 6 months. The record does not give enough detail to determine how intercurrent events, repeated recurrences, incomplete follow-up, or treatment changes were handled. Nor does it report adjusted results. The reported risk difference and relative measures should therefore be read as the posted comparison, not as evidence about unreported analyses.

General statistical methods distinguish a confidence interval from a test of a single null value. The interval displays a range of values reasonably compatible with the data under the method used, while a p-value summarizes compatibility with a particular null hypothesis. They are related but answer different presentation needs: the interval conveys the range and direction of plausible effects, and the p-value gives a measure tied to a null model. Neither removes uncertainty or determines clinical importance.

General considerations also include outcome ascertainment, missingness, and over-dispersion. Over-dispersion arises when variability is greater than a statistical model expects; it can make uncertainty estimates too narrow if the model does not accommodate that extra variability. For a participant-level binary endpoint, the registry record does not report a model diagnostic or an over-dispersion adjustment, and no such issue can be assessed from the posted summary counts alone. More broadly, the record does not state the exact method used to form the confidence intervals or calculate the comparison estimates. The method details should therefore not be inferred from the numerical results.

The registry reports no posted analyses beyond the outcome summary and comparison information. There is no reported covariate adjustment, interim analysis, subgroup analysis, missing-data procedure, or analysis-population definition. This does not establish that such procedures were absent from the full trial conduct; it means they are not reported in the registry record. The numerical comparison is best read as a summary of the reported endpoint data with the stated uncertainty, not as a complete account of the trial’s statistical plan.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.024, and its confidence interval extends from -0.005 to 0.053. The interval includes zero, the value corresponding to no risk difference. It therefore allows for a small difference in the negative direction, no difference, and a positive difference of greater size than the point estimate. The point estimate alone is not a precise description of the possible underlying contrast.

Relative measures also show uncertainty. The risk ratio is 1.43, with a confidence interval from 0.924 to 2.212; the odds ratio is 1.467, with a confidence interval from 0.92 to 2.34. Each interval includes 1, the no-difference value on its respective scale. These intervals span materially different possible relative effects, so the point estimates should not be read as exact or definitive.

In general, a confidence interval is a range produced by a statistical method to represent uncertainty in an estimate. It does not guarantee that the true effect lies in that particular interval, and it is not a range of individual participant outcomes. Its width reflects the precision of the estimate under the method used and the available data. The registry record does not state the exact interval method, which limits assessment of its construction.

The unadjusted p-value is 0.106. Taken together, the intervals and p-value indicate that the posted comparison does not precisely resolve whether the groups differed. The record does not report a prespecified threshold for interpreting the trial analysis, so statistical uncertainty should not be converted into a claim of equivalence or a clinical conclusion.

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 displayed planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. The calculation is a planning illustration based on the observed effect, not the original trial design. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The trial’s own original sample-size assumptions are not reported in the registry record.

For planning at 100% of the observed effect, the illustration gives 1,730 participants per arm for 80% power and 2,316 participants per arm for 90% power. Planning for 75% of the observed effect gives 2,950 participants per arm for 80% power and 3,948 participants per arm for 90% power. Planning for 50% of the observed effect gives 6,351 participants per arm for 80% power and 8,502 participants per arm for 90% power.

In general, power is the chance that a study design will detect an effect of a specified size under its assumptions, if that effect is the true one. The 80% and 90% columns describe different target chances for a future design to detect the selected effect; they are not results from this completed trial. Smaller effects require more participants to detect under the same planning assumptions, which is why the planning figures rise as the assumed effect becomes smaller. These figures are not observed power.

The calculations use the observed effect as a planning reference, so they are sensitive to uncertainty in that estimate. The confidence interval indicates that the effect itself is imprecisely estimated. A planning illustration should therefore be read as conditional on its assumptions and chosen effect size, not as a guarantee that a future study will achieve its target or as evidence that the completed study was adequately designed.

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 = 0.056; treatment proportion p₁ = 0.079; average p̄ = 0.068.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3548 = 0.6954.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3544 = 0.2983.
4. n = (0.6954 + 0.2983)² / (0.0239)² = 1,729.8, rounded up to 1,730 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 <- 0.0555555555555556                          # control proportion
p1 <- 0.079447322970639                          # 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(3045406)
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 = 1,730 (80% power), 2,316 (90% power)
 75% of observed effect: n per arm = 2,950 (80% power), 3,948 (90% power)
 50% of observed effect: n per arm = 6,351 (80% power), 8,502 (90% power)
power.prop.test: 1729.8 
Simulated power at n = 1,730 per arm: 0.803 (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, the primary efficacy endpoint, and zero posted analyses. It reports no secondary outcomes. The available record therefore does not show a family of outcome tests or analyses for which a multiplicity adjustment can be assessed. It also does not report the trial’s prespecified statistical analysis plan, endpoint hierarchy, testing sequence, or any adjustment for multiple comparisons. The absence of those details in the registry record is not proof that the full trial protocol did not address them.

Multiplicity is a general statistical concern when many outcomes, time points, subgroups, or treatment comparisons are tested. As the number of analyses grows, the chance of obtaining an apparently unusual result by chance can increase. Whether and how to adjust depends on the planned questions and decision framework. Because the registry does not report additional analyses or the plan, no conclusion about multiplicity control can be drawn from the posted results.

Missing-data handling and analysis populations are also not reported in the registry record. The outcome denominators are 576 for apixaban and 579 for dalteparin, while enrollment is listed as 1170. The record does not explain the relationship among those figures, identify exclusions from the outcome analysis, report losses to follow-up, or describe how missing endpoint status was handled. It does not state whether the results use an intention-to-treat, modified intention-to-treat, per-protocol, or other analysis population.

In general, missing outcome information can affect a comparison if participants without observed status differ from those with complete information. The direction and extent of any such effect depend on why data are missing and on the analysis approach. Those features cannot be assessed from the registry summary. It also does not report protocol deviations, treatment adherence, treatment discontinuation, or crossover. These omissions mean the posted comparison should be interpreted as the registry reports it, without assuming a particular analysis population or missing-data procedure.

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: neoplasm progression (13.3%), condition aggravated (3.8%), death (3.3%), pulmonary embolism (2.6%), dyspnoea (2.6%).

The serious adverse event summary reports 255 affected participants among 576 at risk in the apixaban group and 276 among 579 at risk in the dalteparin group. The corresponding counts for other adverse events are 89 among 576 and 74 among 579. Death counts are reported as 135 for apixaban and 153 for dalteparin. These are separate registry summaries; the record does not provide enough detail to infer how the categories overlap or to attribute individual events to treatment.

The listed serious adverse event terms and percentages are Neoplasm Progression, 13.3%; Condition Aggravated, 3.8%; Death, 3.3%; Pulmonary Embolism, 2.6%; Dyspnoea, 2.6%; and General Physical Health Deterioration, 2.2%. The denominator and whether these percentages apply across both groups or another population are not reported in the registry record. The terms are event listings, not comparative treatment-effect estimates.

These summaries should not be used to claim that one treatment is safer. The registry record does not report event definitions, severity breakdowns beyond the serious-event grouping, exposure-adjusted rates, timing, causality assessments, or a formal comparative safety analysis. Counts also do not show whether a participant experienced more than one event. They are descriptive safety information, not a substitute for a complete safety assessment.

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

The posted results show 32 participants with recurrent venous thromboembolism in the apixaban group and 46 in the dalteparin group, with denominators of 576 and 579. The reported risk-difference estimate is 0.024 for dalteparin minus apixaban. The confidence interval, from -0.005 to 0.053, includes zero and spans both directions of difference. The risk-ratio and odds-ratio intervals also include their respective no-difference values, and the unadjusted p-value is 0.106. In statistical terms, these results leave meaningful uncertainty about the comparative effect.

The result does not establish equivalence between the treatments. An interval that includes no difference can also include nonzero effects; equivalence requires a suitable prespecified margin and analysis, neither of which is reported in the registry record. Nor does the result prove that there is no effect. Conversely, the numerical difference in the observed counts does not by itself establish a reliable treatment advantage or disadvantage.

The endpoint is a composite of several types of recurrent venous thromboembolism. The registry result does not identify which component contributed to the overall counts or whether the components had similar patterns. It also does not provide subgroup estimates, adjusted analyses, or details about missing outcomes and analysis populations. The result therefore cannot answer unreported component-specific or subgroup questions.

The safety summaries are descriptive and do not establish a comparative safety profile. In particular, the reported adverse-event counts and serious-event terms lack several details needed to interpret differences as treatment-related risks. The efficacy and safety summaries should be read as registry-record information rather than as clinical advice. Individual treatment decisions require evidence and circumstances beyond these posted aggregate results.

Overall, the registry record documents a randomized comparison and reports an imprecise primary endpoint estimate. It is informative about the observed counts and the uncertainty attached to the posted comparison, but it does not settle the clinical question, demonstrate equivalence, or support conclusions beyond the trial population and endpoint details that are reported.

16. Why This Trial Matters Statistically

This trial is statistically informative because it places a randomized comparison of two treatment strategies within a defined clinical question: recurrent venous thromboembolism in patients with cancer over 6 months. Random allocation is intended, in general, to support a fair comparison by balancing prognostic factors across groups on average. The registry record identifies the trial as randomized and parallel-group, but it does not report the sequence-generation or allocation-concealment procedures needed to assess those processes in detail.

The results also illustrate why event counts and point estimates need to be read with uncertainty. The observed counts differ, and the risk-difference estimate has a positive point value for dalteparin minus apixaban. Yet its confidence interval includes zero and spans both directions. Reporting the interval alongside the estimate communicates more than a simple statement about which arm has the larger count: it shows how uncertain the estimated contrast remains.

The composite endpoint is another important statistical feature. Combining several manifestations can produce a single primary measure, but the meaning of the composite depends on its components and their contribution to the total. Since component-specific results are not reported, the overall estimate cannot answer whether the event types behaved similarly. This is a reminder that a compact endpoint can still contain clinically distinct outcomes.

Finally, the registry record demonstrates the limits of registry-level interpretation. The posted outcome and comparison provide useful numerical information, while analysis plans, missing-data methods, analysis populations, and detailed safety methods are not reported. Statistical interpretation requires both an estimate and context about how it was produced. The result is most useful when its uncertainty and reporting limits remain visible rather than being reduced to a simple claim that one group did better.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report detailed eligibility criteria, cancer types, baseline characteristics, or the distribution of venous thromboembolism at enrollment. It therefore does not allow assessment of how closely the enrolled population matches a particular clinical subgroup. The record lists 1170 enrolled, while the primary outcome denominators are 576 and 579; it does not explain the relationship between these figures or identify any exclusions from the outcome summaries.

Endpoint information is limited to the composite definition and the 6-month timeframe. Component-specific event counts, adjudication procedures, follow-up completeness, and the handling of recurrent events are not reported. The record also does not provide adherence, discontinuation, crossover, or protocol-deviation information. These details would help readers assess the interpretation of the observed counts and whether the comparison reflects the intended treatment assignment throughout follow-up.

The statistical methods are incompletely described. No trial-specific statistical analysis plan, model details, confidence-interval method, adjustment strategy, interim analysis information, or analysis-population definition is reported. The registry lists zero analyses and no secondary outcomes. It is therefore not possible to assess whether the posted comparison was the sole planned analysis, how missing outcome data were handled, or whether any prespecified sensitivity analyses were conducted.

The safety record provides counts and a list of serious adverse event terms, but does not report the denominator for the listed percentages, event timing, exposure-adjusted measures, causality assessment, or a formal comparative analysis. The estimates for the primary outcome are also imprecise, with confidence intervals that include their no-difference values. These limitations constrain what can be concluded from the registry results; they do not imply that unreported procedures or results were absent from the full trial documentation.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Apixaban for the Treatment of Venous Thromboembolism in Patients With Cancer — NCT03045406.
• Registry record last updated 2025-06-24; status last verified 2025-06.
• Results first posted 2025-06-18.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes NCT03045406, a completed Phase 3 randomized parallel-group trial of apixaban versus dalteparin for venous thromboembolism in patients with cancer. Enrollment is listed as 1170. The primary endpoint was objectively confirmed recurrent venous thromboembolism over 6 months, recorded as participants with an event.

The outcome counts were 32 of 576 for apixaban and 46 of 579 for dalteparin. The reported risk difference, dalteparin minus apixaban, was 0.024, with a 95% confidence interval from -0.005 to 0.053; the unadjusted p-value was 0.106. Relative-effect intervals also included their no-difference values.

The registry record does not report secondary outcomes or additional analyses, and several details of analysis populations, missing data, endpoint procedures, and safety comparisons are not available. The posted findings describe the observed endpoint and its uncertainty but do not establish equivalence or a definitive comparative treatment effect.