D-dimer to Predict Recurrence in Patients With Multiple Episodes of Venous Thromboembolism: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for D-dimer to Predict Recurrence in Patients With Multiple Episodes of Venous Thromboembolism (NCT00428441), a non-randomized na study. The primary outcome is Recurrent Deep Vein Thrombosis or Pulmonary Embolism in Patients With Persistently Negative D-dimer Levels.

2. Clinical Question

The study asks whether persistently negative D-dimer levels can help identify patients with multiple previous episodes of venous thromboembolism who have a lower risk of another event. Its stated aim is to assess the negative predictive value of D-dimer in this patient group and evaluate the clinical utility of that approach. The primary outcome is recurrent deep vein thrombosis or pulmonary embolism among patients with persistently negative D-dimer levels, assessed over 1 year. The outcome description also includes objectively documented superficial vein thrombosis.

The question arises in the context of uncertainty about the optimal duration of oral anticoagulant therapy after recurrent venous thromboembolism. The registry background describes existing recommendations for treatment longer than 12 months as based on inadequate evidence and notes that the recommendation grade is low. It also refers to prior work on D-dimer after treatment withdrawal in patients with a first VTE episode. The present study extends the question to patients with multiple previous events; it should not be treated as a randomized comparison of alternative treatment strategies.

The registry record does not report the D-dimer threshold, the testing schedule, the criteria for persistent negativity, or a numerical estimate of negative predictive value. It also does not report the number of recurrent events for the primary outcome. Consequently, the question and intended endpoint can be described, but the posted record does not establish how well D-dimer predicted recurrence in this cohort.

3. Trial Design

The registry describes a prospective cohort study with a single-group model, non-randomized allocation, and no masking. The study is not a randomized trial comparing treatment assignments. The registry lists 73 participants enrolled, the condition as venous thromboembolism, and Warfarin as a drug intervention. These details describe the record but do not establish that every enrolled participant received the same treatment or followed the same treatment course.

The study began in 2007-05, had primary completion in 2010-06, and is recorded as terminated. The reason for termination is not reported in the registry record. The phase is listed as not applicable. The registry title and stated aim focus on using D-dimer to assess recurrence risk among patients with multiple previous VTE events, rather than evaluating a randomized treatment contrast.

In a cohort design, groups or risk categories are observed rather than assigned at random. That can be appropriate for describing prognosis, but differences between observed groups may also reflect characteristics other than the marker under study. General statistical interpretation therefore depends on knowing how participants were selected, how D-dimer status was defined, and how follow-up and outcome assessment were conducted. Those procedural details are not reported in the registry record.

The registry identifies the primary outcome as recurrent deep vein thrombosis or pulmonary embolism in participants with persistently negative D-dimer levels, with a timeframe of 1 year. It also names secondary outcomes concerning recurrence and major bleeding among patients who resumed oral anticoagulant therapy, along with mortality. The record does not describe the treatment-resumption criteria or the comparator structure for those outcomes. Since the design is listed as single-group and non-randomized, those secondary outcomes should not be read as randomized comparisons.

The registry gives an enrollment count of 73, while the adverse-event record lists 75 participants at risk in its D-dimer Negative group. The record does not explain this difference. Enrollment and safety denominators are not interchangeable by default, so the discrepancy should be retained as an unresolved reporting feature rather than reconciled by assumption.

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

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 Recurrent Deep Vein Thrombosis or Pulmonary Embolism in Patients With Persistently Negative D-dimer Levels, measured over 1 year.

The primary endpoint is titled recurrent deep vein thrombosis or pulmonary embolism in patients with persistently negative D-dimer levels. It is recorded as a participant count, with a timeframe of 1 year. The outcome description specifies objectively documented deep vein thrombosis, pulmonary embolism, and superficial vein thrombosis. The title and description do not align perfectly: the title names deep vein thrombosis or pulmonary embolism, while the description additionally includes superficial vein thrombosis. The registry record does not clarify how that wording difference was handled in outcome classification.

A recurrence endpoint is relevant to the study question because the stated aim is to assess whether negative D-dimer status predicts subsequent VTE. For prognostic questions, an endpoint needs a defined event and follow-up period so that the observed event experience can be related to the marker status. Objective documentation can reduce ambiguity in deciding whether an event occurred, although the registry does not report the specific diagnostic criteria or adjudication procedures used.

The registry lists the primary endpoint and its timeframe but does not provide an outcome count, rate, negative predictive value, effect estimate, confidence interval, or statistical test. It also does not report the threshold or timing used to define persistent D-dimer negativity. Therefore the endpoint's intended meaning is identifiable, but its observed result and measurement details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas help distinguish the study question from the results needed to answer it. The following explanations are general. The registry record identifies a prospective cohort design and a primary endpoint, but does not report a statistical analysis plan or analysis results.

Prediction and negative predictive value

Negative predictive value describes the proportion of people classified as test-negative who do not experience the outcome over the defined follow-up period. It is conditional on being in the test-negative group. It is not the same quantity as sensitivity, specificity, or the probability of a negative test among people who remain event-free. Predictive values can also vary with the frequency of the outcome in the population being studied. Thus, an estimate in patients with multiple previous VTE events would pertain to the population and follow-up represented in this cohort; it should not automatically be carried over to other clinical settings.

To estimate negative predictive value, an analysis needs an operational definition of a negative D-dimer result, an outcome definition, and a specified follow-up window. The record gives a 1 year timeframe and describes persistent negative D-dimer levels, but does not report the assay threshold, test schedule, or rule for determining persistence. It also does not report the number of people in the relevant D-dimer category or the number experiencing recurrence. The predictive value therefore cannot be reconstructed from the registry entry.

Confidence intervals and precision

A confidence interval communicates uncertainty around an estimated quantity. In general, a narrower interval indicates greater precision under the assumptions of the statistical method, while a wider interval indicates more uncertainty. The interval does not guarantee that the estimate is unbiased, and its interpretation depends on the design, measurement quality, and model assumptions. A point estimate alone can conceal substantial uncertainty, especially when event information is limited.

No confidence interval or point estimate for recurrence or negative predictive value is reported in the registry record. Precision cannot therefore be assessed from the posted results. The enrollment count alone does not determine precision: the number of outcome events, the number of participants actually contributing follow-up, and the analysis approach also matter. Those details are not reported.

Observed counts and rates

The primary endpoint is expressed as a number of participants with an event, rather than a continuous measurement. A count is directly interpretable as the number meeting the outcome definition, but a count by itself does not give the frequency of events unless the relevant denominator is also known. For a time-to-event analysis, follow-up duration and censoring would also affect interpretation. Although the timeframe is recorded as 1 year, the registry does not describe follow-up completeness, losses to follow-up, or censoring.

The primary endpoint description mentions objective documentation of deep vein thrombosis, pulmonary embolism, and superficial vein thrombosis. Consistent outcome ascertainment is important because misclassification can distort estimated event frequency and predictive performance. The registry record does not report the diagnostic criteria, blinded adjudication, or procedures for resolving uncertain events.

Why non-randomized design matters

In a randomized trial, random assignment is intended to balance prognostic factors across treatment groups on average. In an observational cohort, participants are not randomly allocated to a marker category or treatment course. D-dimer status and decisions to resume anticoagulant therapy may be related to patient characteristics or clinical circumstances. As a general matter, such relationships can complicate interpretation of associations. A study can still investigate prognosis, but the analysis should make clear which population is being described and how potential confounding and selection are addressed.

The registry specifies non-randomized allocation and a single-group model. It does not report adjusted analyses, covariates, subgroup methods, or a comparison strategy. The planning assumptions refer to equal allocation between two arms, but that is an assumption for a planning illustration, not a description of the recorded study design. No randomized two-arm comparison should be inferred from that assumption.

Over-dispersion and count models

General explanation: over-dispersion occurs when the variability in count data exceeds what a chosen statistical model expects. If a model assumes that counts vary less than they actually do, standard errors can be too small and confidence intervals can be too narrow. Analysts may use a model that accommodates extra variability, but the appropriate choice depends on the outcome structure and study design. The registry does not report a count model, a dispersion assessment, or any model-based results. This concept cannot be used to infer which method was used.

Planning calculations are not study findings

A sample-size calculation is a planning exercise based on specified assumptions about the effect to detect, variability or event frequency, significance level, allocation, and statistical method. A planning illustration based on an observed effect does not report the trial's observed power. Observed power is not a substitute for an effect estimate with a confidence interval, and it does not show that a result is precise or unbiased.

The assumptions recorded for the planning illustration are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect to detect expressed as a stated fraction of the observed effect. The record says that time-to-event calculations use a total event count and other calculations use participants per arm. These are planning conventions. The study itself is listed as non-randomized and single-group, and its own sample-size justification is not reported in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("D-dimer Negative")
n      <- c(75)
events <- c(9)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  D-dimer Negative  75  9  12.0  5.6  21.6  6.4  21.3  6.5  21.3

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary outcome and secondary outcomes concerning recurrent deep vein thrombosis or pulmonary embolism among patients who resumed oral anticoagulant therapy, major bleeding among patients who resumed therapy, and mortality. It records five outcome measures overall, but the available outcome descriptions do not provide a complete account of every measure or its analysis. The registry records 0 analyses. No statistical analysis plan is reported.

Multiplicity arises when several outcomes, comparisons, time points, or subgroups are tested. As a general statistical matter, conducting many tests can increase the chance of at least one apparently unusual result, unless the analysis plan accounts for the testing structure. The registry record does not report whether any multiplicity strategy was planned, whether the primary outcome was prioritized in formal testing, or how secondary outcomes were intended to be interpreted. Without such information and without analysis results, no claim about multiplicity-adjusted inference is possible.

Missing-data handling is also not reported in the registry record. It does not describe missing D-dimer measurements, incomplete follow-up, withdrawals, censoring, or imputation. The enrolled count does not establish how many participants contributed to each outcome analysis. Nor does the record specify an intention-to-treat, per-protocol, or other analysis population. In a prospective cohort, those definitions still matter: the included participants and handling of incomplete observations determine the population to which an estimate applies.

The registry record does not report how treatment resumption affected outcome classification or analysis. That is particularly relevant for secondary outcomes restricted to patients who resumed anticoagulant therapy, because the record does not describe the criteria for resumption or the size of that subgroup. These omissions prevent assessment of the analysis populations, missing-data bias, or the statistical interpretation of secondary outcomes.

12. 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 1% reporting threshold.

The adverse-event record identifies a D-dimer Negative group with 75 participants at risk for serious adverse events and 0 affected. It also lists 75 at risk for other adverse events and 0 affected. The deaths field is null, which means a death count is not reported in that field; it should not be read as a reported zero. No serious adverse-event terms are listed.

These are registry entries, not a complete account of all possible safety outcomes. The record does not report event definitions, collection procedures, observation duration for adverse events, or a detailed breakdown by event type. The zero entries indicate that no affected participants are recorded in those fields; they do not establish that every possible event was assessed or that no event occurred outside the recorded scope.

The adverse-event denominator of 75 does not match the enrollment count of 73. The registry does not explain this difference, so the safety figures should be interpreted using their stated denominator while retaining the discrepancy as unresolved. The record also lists secondary outcomes for major bleeding among patients who resumed oral anticoagulant therapy, but does not provide a result for that outcome. The general adverse-event entries cannot be substituted for the missing major-bleeding outcome result.

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

The registry record establishes what the study set out to examine: whether persistently negative D-dimer levels could inform assessment of recurrence in patients with multiple previous VTE events. It identifies a primary endpoint involving recurrent deep vein thrombosis or pulmonary embolism, a 1 year timeframe, and an outcome description that also includes superficial vein thrombosis. These details establish the intended question, not its answer.

The record does not report the primary outcome count, negative predictive value, recurrence rate, effect estimate, confidence interval, or p-value. It therefore does not show the observed risk associated with negative D-dimer in this cohort or how precisely that risk was estimated. No conclusion about the predictive performance of D-dimer can be drawn from the outcome title alone.

The study is prospective, non-randomized, and single-group. It can be understood as a cohort investigation rather than a randomized comparison of treatment strategies. Warfarin is listed as an intervention, and some secondary outcomes refer to participants who resumed oral anticoagulant therapy, but treatment details and resumption criteria are not reported. No causal treatment effect should be inferred from the design description.

The registry's serious and other adverse-event fields record 0 affected among 75 participants at risk in each field. The deaths field is null, and the record does not provide a result for the separate major-bleeding outcome. These safety entries should be read narrowly: they describe what was entered in those registry fields and do not supply missing primary outcome data. The enrollment count of 73 and safety denominator of 75 are not reconciled in the record.

In short, the registry describes a relevant prognostic question and its intended endpoint, but not the findings needed to evaluate it. The result does not establish whether negative D-dimer was or was not informative, nor does it establish the safety or effectiveness of any treatment strategy. Further details about outcome results, measurement methods, follow-up, and statistical analyses are not reported in the registry record.

14. Why This Trial Matters Statistically

This study is statistically interesting because it shifts a D-dimer prognostic question from patients with a first VTE episode to patients described as having multiple previous events. Prognostic performance is population-specific: a marker's predictive value depends in part on the underlying outcome frequency and on how the target population is selected. The registry's stated aim therefore addresses a distinct clinical population rather than simply repeating a question in an interchangeable group.

The study also illustrates the difference between a registered research question and an assessable result. A clear endpoint title and timeframe are useful, but prediction requires observed outcomes and enough information to judge uncertainty. The registry does not report the primary event count, predictive-value estimate, or confidence interval. Without those, readers cannot assess either the magnitude or precision of the proposed prognostic finding.

Its non-randomized, single-group structure is important to interpretation. A cohort can describe outcome experience across observed marker categories, but treatment decisions and participant characteristics may be associated with both marker status and subsequent events. The registry does not report adjustment methods or a complete analysis plan. As a result, even an association—if one were reported—would need to be interpreted with attention to selection, confounding, and how the marker and outcome were measured.

The record also demonstrates why denominators and outcome definitions need to be checked carefully. It lists 73 enrolled participants but 75 at risk in the adverse-event group, and the primary endpoint title and description differ on whether superficial vein thrombosis is included. Neither point can be resolved from the registry record. These ambiguities do not determine the study's findings, but they limit transparency about who contributed to each measure and what counted as an event.

Finally, the presence of planning assumptions does not fill gaps in the study's own design rationale or results. A planning illustration is not evidence of observed power and cannot replace outcome estimates with uncertainty intervals. The statistical value of the record is therefore chiefly in documenting the intended question and highlighting what information would be needed for a full evaluation.

15. 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 central limitation for interpretation is that primary outcome results are not reported in the registry record. There is no recurrent-event count, negative predictive value, effect estimate, confidence interval, or p-value. The registry records 0 analyses, so the posted information cannot establish the observed predictive performance of D-dimer in the cohort.

Important operational details are also absent. The record does not report the D-dimer threshold, assay, testing schedule, or criteria for persistent negativity. It does not specify diagnostic criteria or adjudication procedures for the primary outcome, and the endpoint title differs from its description regarding superficial vein thrombosis. Follow-up completeness, losses to follow-up, and censoring are not described.

The design is non-randomized and single-group. The registry does not report adjustment for prognostic characteristics, a prespecified comparison strategy, subgroup methods, or a statistical analysis plan. It also does not explain the treatment regimen or criteria for resuming oral anticoagulant therapy, even though secondary outcomes refer to patients who resumed treatment. The resulting treatment and selection processes cannot be assessed from the record.

The registry gives an enrollment count of 73 but lists 75 participants at risk in each adverse-event field for the D-dimer Negative group. It does not reconcile these figures. The deaths field is null, and results for the named major-bleeding and mortality outcomes are not reported. The record also lists five outcome measures overall without detailing all of them in the available outcome descriptions.

The study is marked terminated, but the reason is not reported. The registry does not provide its own sample-size justification, missing-data procedures, analysis populations, or multiplicity approach. These limitations are specific to what can be established from this record; they should not be mistaken for evidence that a particular method was or was not used outside the information recorded.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: D-dimer to Predict Recurrence in Patients With Multiple Episodes of Venous Thromboembolism — NCT00428441.
• Registry record last updated 2011-09-26; status last verified 2011-08.
• Results first posted 2011-09-26.
• 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.

19. Record Summary

NCT00428441 is a prospective cohort study titled “D-dimer to Predict Recurrence in Patients With Multiple Episodes of Venous Thromboembolism.” The official title describes assessment of the negative predictive value of D-dimer in patients with multiple previous events. The registry lists enrollment of 73, a non-randomized allocation, a single-group model, no masking, and Warfarin as a drug intervention.

The study began in 2007-05, reached primary completion in 2010-06, and is recorded as terminated. The primary endpoint concerns recurrent deep vein thrombosis or pulmonary embolism in participants with persistently negative D-dimer levels over 1 year; the description also includes superficial vein thrombosis. Named secondary outcomes concern recurrence and major bleeding among patients who resumed oral anticoagulant therapy, and mortality.

The registry record does not report primary outcome results, statistical analyses, the reason for termination, or the study's own sample-size justification. Its adverse-event fields list 0 affected among 75 at risk for serious and other events, while the deaths field is null. The difference between enrollment and the adverse-event denominator is not explained.