Prospective Genotyping For Total Hip or Knee Replacement Patients Receiving Warfarin (Coumadin): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prospective Genotyping For Total Hip or Knee Replacement Patients Receiving Warfarin (Coumadin) (NCT00634907), a randomized na study. The primary outcome is The Number of Participants With Adverse Events Associated With Warfarin Anticoagulation Following Total Hip and Total Knee Replacement.

2. Clinical Question

The registry describes the study as an evaluation of whether genetic testing could improve warfarin initiation compared with usual care. The reported primary outcome is a participant count, not a measure of the time until an event or the total number of events. The comparison therefore concerns whether participants experienced at least one event meeting the listed definition during the stated follow-up period.

This is a question about the effect of a dosing strategy in the trial setting, not a comparison of genotyping alone with no genotyping. Both arms received warfarin dosing as part of postoperative care; the distinction was whether dosing used a formula incorporating genetic test results or followed usual care. The registry record does not report the eligibility criteria in enough detail to establish how closely the enrolled population represents other surgical settings.

3. Trial Design

The registry identifies this as a randomized, parallel-group trial with no masking. Participants were assigned to pharmacogenetic-based warfarin dosing or usual-care warfarin dosing. The trial is listed as completed, with enrollment of 263. The posted primary-outcome arm counts are 114 in the genotype arm and 115 in the control arm. The registry record does not explain the difference between overall enrollment and the arm counts shown for this outcome.

The trial concerned patients undergoing total hip or knee replacement who received postoperative warfarin thromboprophylaxis. In the genotype arm, warfarin dosing was based on a formula incorporating genetic testing results. The comparison arm is described as usual care. The registry notes that warfarin administration and the duration of prophylaxis were not specific to the study; further details of the regimens are not reported in the registry record.

The study is listed with phase NA. The record gives a start date of 2006-09 and a primary completion date of 2008-10. It lists venous thromboembolism and bleeding as conditions. The primary outcome was assessed at 90 days post surgery and counted participants with an adverse event associated with warfarin anticoagulation. This outcome included major bleeding, minor bleeding, symptomatic deep vein thrombosis, and pulmonary embolism.

Three secondary outcomes are listed, all concerning the percentage of INR determinations in therapeutic, subtherapeutic, or supratherapeutic ranges. The registry record does not report the detailed analysis plan, the original sample-size justification, the assignment procedure, or how missing outcome information was handled. It also does not describe masking beyond listing the study as having none. These gaps limit what can be concluded about the planned conduct and analysis beyond the design features explicitly recorded.

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: Exclusion Criteria: 1. Blood transfusion in previous two weeks 2. Participant is already taking warfarin 3. Pre-operative INR \> 4.0 4. Pre-operative bilirubin \> 2.4 mg/dL 5. Current active cancer diagnosis with ongoing treatment 6. Concomitant medications known to exert a major interaction with warfarin

Overall Study

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 The Number of Participants With Adverse Events Associated With Warfarin Anticoagulation Following Total Hip and Total Knee Replacement, measured over 90 days post surgery.

The primary endpoint is the number of participants with adverse events associated with warfarin anticoagulation following total hip or total knee replacement. The assessment window is 90 days post surgery. The registry defines a qualifying event as major bleeding, minor bleeding, symptomatic deep vein thrombosis, or pulmonary embolism.

Major bleeding includes fatal bleeding, bleeding into a critical organ, or bleeding requiring hospital admission. Minor bleeding is clinically overt bleeding that does not meet the criteria for major bleeding. Deep vein thrombosis must be symptomatic under the stated definition. These descriptions specify the kinds of events counted, although the registry record does not report further adjudication procedures or how events were verified.

Counting participants with an event gives an interpretable summary of how many people in each arm experienced at least one qualifying event during follow-up. It does not show the number of repeated events per participant, the timing of events, or how long participants remained event-free. Nor does the combined endpoint distinguish the relative contribution of bleeding and thromboembolic events to the overall comparison. The registry lists each component in its definition, but the posted primary result is an overall participant count.

As a general statistical point, a prespecified clinical endpoint is useful because it gives the comparison a defined outcome and observation window. Here, the registry record identifies the endpoint and timeframe, but does not report whether this was the original protocol’s sole primary endpoint or how component events were handled analytically beyond the stated definition.

7. Results Posted in the Registry

• 2 Genotype Arm: 1.8% (2/114) versus 3.5% (4/115) on control; risk difference -1.7 percentage points (95% CI -5.8 to 2.4), risk ratio 0.50 (0.09–2.70); unadjusted two-proportion z-test p = 0.414.

8. Statistical Interpretation

9. Statistical Methods Explained

The posted comparison is between the genotype arm and the control arm, with the primary endpoint expressed as participants who experienced a qualifying event. For a binary outcome of this kind, each participant is classified according to whether an event was recorded during the specified follow-up. The arm-level counts can then be summarized as risks and compared using an absolute measure such as a risk difference or relative measures such as a risk ratio or odds ratio.

What does a risk difference describe?

A risk difference compares the event risk in one arm with the event risk in the other, using the order specified for the comparison. A negative estimate indicates a lower observed risk in the genotype arm than in the control arm. Unlike a relative measure, a risk difference is expressed on an absolute scale. The posted estimate and interval should be read together: the estimate gives the central result, while the interval communicates how much uncertainty remains around it.

The registry also posts relative-effect measures. A risk ratio compares the risks between arms as a ratio, while an odds ratio compares the odds. These measures answer related but distinct questions, and their numerical values are not directly interchangeable with an absolute risk difference. With sparse outcomes, relative measures may be particularly sensitive to the small number of observed events. The posted values are estimates from the data, not guarantees about the effect in a broader population.

What does the p-value contribute?

The posted unadjusted p-value summarizes how inconsistent the observed data are with a specified null model of no arm difference, under the assumptions of the test used. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. A p-value that is not small does not prove equivalence, particularly when event counts are limited and estimates are imprecise.

The registry record does not describe the test used to produce the posted p-value or provide a detailed statistical analysis plan. It does not report whether analyses were adjusted for covariates, whether a particular analysis population was prespecified, or whether sensitivity analyses were planned. Those aspects should not be inferred from the outcome table.

Why do sparse counts matter?

When events are uncommon, the estimate can be strongly influenced by a small number of participants. This makes the observed difference less stable and typically leaves substantial uncertainty about the underlying effect. The endpoint is a participant count, so it does not account for multiple events in the same person as separate observations. That is suitable for the stated endpoint, but it means the result does not describe recurrent event burden.

The endpoint also combines several event types. A combined outcome can summarize the overall occurrence of clinically defined events, but it may obscure differences among its components. The registry record lists major bleeding, minor bleeding, symptomatic deep vein thrombosis, and pulmonary embolism, but does not give component-specific primary-outcome counts. Interpretation of the overall result should therefore stay at the level of the combined endpoint.

What do the secondary outcomes measure?

The listed secondary outcomes concern the percentage of INR determinations in specified ranges: therapeutic (INR 1.8-2.9), subtherapeutic (INR less than 1.8), and supratherapeutic (INR greater than 2.9). These are based on determinations rather than the primary endpoint’s participant-level event count. A participant may contribute more than one determination, so the unit of analysis and the handling of repeated measurements matter. The detailed analysis methods for these outcomes are not reported in the registry record.

As a general statistical point, repeated measurements from the same participant may be correlated rather than independent. Analyses that treat all determinations as independent can misstate uncertainty if that correlation is not handled. The registry does not report whether a repeated-measures method or another approach was used, so no specific method should be attributed to the study.

What is not documented?

The registry record reports nine outcome measures and three secondary outcomes, but no analysis entries. It does not report the original sample-size calculation, the planned significance threshold, interim analyses, stopping rules, missing-data strategy, or methods for defining analysis populations. The absence of these details in the registry should not be taken as evidence that a particular procedure was or was not used in the trial. It means only that these methods cannot be assessed from the record.

General statistical interpretation separates the estimate, its uncertainty, and the assumptions used to obtain them. The estimate gives the observed direction and magnitude; the interval shows a range of values consistent with the data under the method used; and the p-value addresses compatibility with a null model. Here, sparse events and undocumented analysis details counsel against treating any single statistic as definitive.

10. Confidence Intervals and Precision

The posted confidence interval around the risk difference is central to understanding precision. Its bounds lie on opposite sides of zero, so the interval includes both a difference favoring the genotype arm and a difference favoring the control arm. This means the trial data do not pin down the direction of the underlying difference with precision. The upper bound is reported as 0.024; the lower bound is negative.

A confidence interval is a range produced by a statistical procedure to express uncertainty in an estimate. In general, a narrower interval indicates greater precision, while a wider interval indicates that a broader range of effects remains compatible with the data. The registry record does not report the confidence level or the method used to calculate this interval, so those details cannot be assumed.

The observed event counts are 2 and 4, and the endpoint arm sizes are 114 and 115. Such limited event information constrains precision: the data contain relatively little information about the difference in underlying event risks. The point estimate is therefore best read alongside the interval, rather than as a standalone forecast.

The confidence interval should not be interpreted as proving that every value within it is equally plausible, nor that values outside it are impossible. It is a summary tied to a statistical model and procedure. Since the record does not specify those details, the interpretation should remain cautious and descriptive.

In general, a nonsignificant p-value and a confidence interval spanning zero do not demonstrate equivalence. Establishing equivalence or noninferiority requires an appropriate design and prespecified margin, neither of which is reported in the registry record. The current result is best described as an uncertain estimate from a trial with few observed primary events.

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 sample-size table and curve are planning illustrations based on the observed effect, not a report of the trial’s original design calculation. For this binary endpoint, the displayed figures are participants per arm, not total events. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row uses the stated fraction of the observed effect as the effect to detect.

The rows describe planning for 100%, 75%, and 50% of the observed effect. The table gives the number of participants per arm for 80% power and for 90% power under those assumptions. In plain terms, the 80% column describes a planning target with an 80% chance of detecting the specified effect if that effect is truly present and the assumptions hold; the 90% column uses a higher chance of detection and therefore requires more participants. These are prospective planning quantities, not estimates of what the completed trial achieved.

Smaller effects are harder to distinguish from random variation, so planning to detect a smaller fraction of the observed effect requires more participants per arm. The curve visualizes the same relationship across the displayed planning scenarios. It should be read as a guide to how required sample size changes with the target effect, rather than as a forecast of the trial’s result.

This is not observed power. The registry record does not report the trial’s original sample-size justification or its original design assumptions. The planning assumptions are only those used for the illustration, and they do not establish what the investigators planned before enrollment.

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 = 0.035; treatment proportion p₁ = 0.018; average p̄ = 0.026.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.2257 = 0.4424.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2254 = 0.1897.
4. n = (0.4424 + 0.1897)² / (-0.0172)² = 1,344.7, rounded up to 1,345 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 <- 0.0347826086956522                          # control proportion
p1 <- 0.0175438596491228                          # 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(634907)
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 = 1,345 (80% power), 1,800 (90% power)
 75% of observed effect: n per arm = 2,583 (80% power), 3,458 (90% power)
 50% of observed effect: n per arm = 6,242 (80% power), 8,355 (90% power)
power.prop.test: 1344.7 
Simulated power at n = 1,345 per arm: 0.813 (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 lists three secondary outcomes: the percentage of INR determinations in the therapeutic range, the percentage below that range, and the percentage above it. It reports nine outcome measures overall and no analysis entries. The record does not describe how the primary and secondary outcomes were prioritized in the analysis, whether any multiplicity adjustment was planned, or how conclusions across several outcomes were to be interpreted.

As a general statistical matter, testing several outcomes can increase the chance of observing at least one apparently notable result by chance. Adjustment methods or a prespecified hierarchy can address that concern, but the registry record does not report whether either approach was used. No assumption about the presence or absence of adjustment should be made from the outcome listing alone.

Missing-data handling is also not reported in the registry record. The record does not explain whether every enrolled participant contributed to the primary analysis, how unavailable outcome information was treated, or whether any participants were excluded. The endpoint arm counts are listed as 114 and 115, while overall enrollment is 263; the reason for this difference is not reported in the registry record. The counts should not be treated as proof of a particular exclusion or missing-data process.

The analysis populations are not specified. The record does not say whether the primary analysis followed intention-to-treat, per-protocol, or another definition. It also does not report the use of covariate adjustment, subgroup analyses, sensitivity analyses, or interim analyses. These details matter because different analysis choices can affect both the estimate and its uncertainty.

These omissions limit reproducibility and make it difficult to assess whether the posted comparison follows a prespecified plan. They do not establish that the trial lacked such a plan; they indicate that the registry record does not document it. The most defensible reading is to describe the posted results as they appear, while avoiding claims about unreported methods.

14. 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: venous thromboembolic event or bleed (3.5%).

The registry’s serious adverse event summary lists 2 affected participants among 114 at risk in the genotype arm and 4 among 115 at risk in the control arm. The most frequent serious-event term shown is Venous thromboembolic event or bleed, reported as 3.5%. These figures are separate safety summaries and should not be assumed to reproduce every detail of the primary endpoint analysis.

The record also lists other adverse events affecting 1 participant in the genotype arm and 6 in the control arm, with the same arm-specific numbers at risk. The death counts are not reported in the registry record. The serious-event term is a grouped label; the registry does not provide a breakdown of its components.

Safety summaries depend on their definitions, observation periods, and reporting rules. The registry record does not give further detail about event attribution, severity classification procedures, or how the summaries relate to the primary endpoint’s 90-day window. These counts describe the posted safety information but do not by themselves establish a causal relationship with either dosing strategy.

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

The result describes the recorded primary endpoint in this randomized comparison: 2 participants in the genotype arm and 4 in the usual-care control arm experienced a qualifying event, with arm sizes of 114 and 115. The posted risk-difference estimate is negative, and the interval spans zero. The unadjusted p-value is 0.414. Taken together, these results show an observed difference but leave considerable uncertainty about the underlying difference.

The comparison does not establish that the two dosing strategies are equivalent. It also does not establish that one strategy reduces or increases adverse events in the broader population of people undergoing joint replacement. The sparse event counts and interval spanning both directions make strong claims about the direction or size of an effect unwarranted.

The outcome combines bleeding and thromboembolic events. The overall result cannot show which event types contributed to the observed counts, and it does not describe repeated events or the timing of events. The secondary INR outcomes address a different unit of observation, namely determinations, and the registry record does not provide their detailed analysis methods.

General statistical interpretation distinguishes absence of strong evidence for a difference from evidence that there is no difference. This trial’s posted comparison is not a demonstration of no effect. A finding of equivalence or noninferiority would require a defined margin and an analysis designed to evaluate it; the registry record does not report such a design.

These data should be understood as one trial’s recorded findings under its study conditions and follow-up period. They do not settle how genetic testing should be used in clinical practice, and this page does not offer medical advice. The registry record does not report enough about eligibility, dosing details, missing data, or analysis populations to support broader conclusions beyond the posted comparison.

16. Why This Trial Matters Statistically

This trial raises a statistical question that is common in clinical research: how to interpret an apparently different outcome count when the event is uncommon. The genotype arm had fewer observed participants with a primary event than the control arm, but the counts are small and the confidence interval spans zero. Looking only at the direction of the point estimate would overstate what the data establish.

The record also illustrates why absolute and relative effect summaries serve different purposes. A risk difference expresses an arm contrast on an absolute scale, while a risk ratio or odds ratio expresses a relative contrast. Both can help describe a result, but neither removes the uncertainty caused by limited event information. Readers should consider the estimates together with their intervals and the event counts that underlie them.

Endpoint construction matters as well. Combining bleeding and thromboembolic events can provide an overall summary of adverse events, but the aggregate result does not identify which components drove the comparison. Meanwhile, the secondary INR outcomes concern repeated determinations rather than simply whether a participant experienced an event. Those different outcome structures require methods suited to their units of analysis; the registry record does not document the methods used for the secondary outcomes.

Finally, this example shows the distinction between a trial’s original design and a later planning illustration based on its observed effect. The registry does not report the original sample-size justification. The posted planning table is therefore useful for understanding how target effects relate to participant requirements under stated assumptions, but it cannot be used to reconstruct the trial’s intended design or to describe its achieved power.

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 primary outcome has few observed events: 2 participants in the genotype arm and 4 in the control arm. This limits the information available to estimate the difference and is reflected in the interval spanning zero. The result is consequently imprecise, and small changes in the observed counts could have an appreciable effect on the summary.

The primary endpoint combines major bleeding, minor bleeding, symptomatic deep vein thrombosis, and pulmonary embolism. The registry record does not report component-specific counts for this endpoint, so the overall comparison cannot show whether one event type predominated or whether the components differed in their patterns.

Overall enrollment is 263, whereas the arm counts shown for the primary outcome are 114 and 115. The reason for the difference is not reported in the registry record. The record also does not describe missing outcomes, exclusions, analysis populations, or the treatment of participants without complete follow-up. This prevents a clear assessment of how the outcome counts relate to the full enrolled group.

The study is listed as randomized and unmasked, but the registry record does not give detailed information about the randomization procedure or implementation of dosing. It also does not report the original sample-size justification, analysis plan, adjustment methods, or any interim analyses. The usual-care dosing details and the exact genetic dosing formula are not fully described.

These are limitations of what can be assessed from the registry record. They do not show that unreported procedures were not performed. They mean that interpretation should remain anchored to the posted design and results, without filling in details absent from the record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Prospective Genotyping For Total Hip or Knee Replacement Patients Receiving Warfarin (Coumadin) — NCT00634907.
• Registry record last updated 2017-06-08; status last verified 2017-04.
• Results first posted 2017-06-08.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes a completed, randomized, parallel-group trial with no masking, comparing pharmacogenetic-based warfarin dosing with usual-care dosing after total hip or knee replacement. Enrollment is listed as 263. The primary outcome counts participants with warfarin-associated adverse events within 90 days post surgery, including specified bleeding and thromboembolic events.

The posted primary counts are 2 of 114 participants in the genotype arm and 4 of 115 in the control arm. The unadjusted p-value is 0.414, and the risk-difference interval spans zero. The outcome counts are sparse, and the comparison remains uncertain.

Three secondary outcomes concern INR determinations. The registry record does not report the original sample-size justification, detailed analysis plan, missing-data handling, or analysis populations. The sample-size display is a planning illustration based on the observed effect, not the trial’s original design calculation.