Clinical Evidence for Rebound Hypercoagulability After Discontinuing Oral Anticoagulants for Venous Thromboembolism

David Keith Cundiff, MD


Medscape J Med. 2008;10(11):258 

In This Article


A MEDLINE search (July 2008) for appropriate RCTs had the following string of search terms: "(vitamin K antagonist or vitamin K inhibitor or warfarin or acenocoumarol or dicumarol or fluindione or ximelagatran) AND (venous thromboembolism or pulmonary embolism or pulmonary emboli or deep venous thrombosis) AND (randomized controlled trial)." There were no date restrictions. Only reports in the English language could be considered. Studies were also included from a Cochrane meta-analysis of RCTs of trials comparing different durations of OAC treatments,[8] and the reference list from this meta-analysis and from published RCTs of VTE treatment.

Trials were excluded if:

  1. No event data were reported for more than 2 months after discontinuation of OACs;

  2. The timing of VTE recurrences was not reported;

  3. A surrogate rather than a clinical endpoint was used (eg, positive Doppler leg scan in an asymptomatic patient) rather than a clinical recurrence (or extension);

  4. The report duplicated an included study;

  5. The study was not an RCT;

  6. The RCT did not involve treatment of VTE;

  7. OAC treatment was not a focus of the study; or

  8. The report was in a language other than English.

Failure to have objective tests confirming the diagnoses of VTE for all subjects was not considered an exclusion. However, such trials were separately analyzed to see whether they affected the overall results.

Data on VTE recurrences (1) while on OACs, (2) in the 2 months after discontinuing OACs, and (3) subsequently were extracted from published RCTs of OAC treatment of VTE. The rate of major hemorrhage per month of OAC treatment, if available, was also assessed.

To estimate the number of subject-months at risk for each phase of a trial, the numbers of subjects who (1) began the trial, (2) finished the anticoagulation period, (3) finished 2 months after anticoagulation cessation, and (4) completed the follow-up time were extracted from the publication. For each trial phase (ie, anticoagulation, 2-month rebound period, and post rebound until the end), the numbers of subjects participating in the trial at the beginning and end of the period were summed and the result divided by 2. That result was multiplied by the number of months in the period. For instance, if 100 subjects entered the trial and 80 completed the 6-month period of anticoagulation, the months at risk for adverse events (VTE recurrences and major bleeding) would be 540 months ([100 + 80]/2 = 90; 90 x 6 months = 540 months).

The VTE recurrences possibly attributable to rebound hypercoagulability due to OAC withdrawal were estimated by subtracting the rate of VTE recurrences after the 2-month rebound period from the rate during the 2-month rebound period.

With the extracted data, comparisons were made of the following:

  • VTE recurrences in the 2 months after discontinuing OACs vs subsequently, also broken down by durations of OAC treatment ( ≤ 3 months and ≥ 6 months);

  • In the RCTs with evaluable data from subjects receiving shorter vs longer durations of OACs, overall VTE recurrence rates in shorter vs longer duration OAC treatment arms of RCTs; and

  • In the RCTs with evaluable data from subjects receiving shorter vs longer durations of OACs, overall rate of adverse events (VTE recurrences plus major hemorrhages) in shorter vs longer duration OAC treatment arms of RCTs.

Rates were calculated as (1) VTE recurrences and major bleeds while taking OACs per subject and (2) VTE recurrences and major bleeds per month of OAC treatment. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated with a chi-square function to make the comparisons. For the VTE recurrences in the 2 months post OACs vs the monthly rate from > 2 months until the end of the trial, a subgroup analysis of studies of similar OAC duration (ie, ≤ 3 months vs ≥ 6 months) was used as a stratification variable in the primary analysis of the combined studies to determine whether duration of OAC treatment affected outcomes. A fixed-effects approach was used, and statistical heterogeneity across all the trials was tested with the I2 statistic. Funnel plots were used to test for possible publication bias.

RevMan 5.1.14 from the Cochrane Collaboration was used to perform chi-square comparisons of the RCT study arms, forest plots, funnel plots, and heterogeneity assessments.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.