Heparin and warfarin anticoagulation intensity as predictors of recurrence after deep vein thrombosis or pulmonary embolism: a population-based cohort study

Heparin is recommended as the initial treatment for acute deep vein thrombosis (DVT) and pulmonary embolism (PE)¹  because such therapy improves survival after PE,²  and reduces asymptomatic DVT extension and possibly 3-month VTE recurrence.³  Activated partial thromboplastin time (APTT) monitoring and heparin dose adjustment to rapidly achieve and maintain an APTT “therapeutic” range corresponding to a plasma heparin level of 0.3-0.7 anti-X_a U/mL is also recommended.^1,4  The College of American Pathologists, the British Committee for Standards in Haematology, and the American College of Chest Physicians (ACCP) have recommended procedures to identify the laboratory-specific APTT therapeutic range,^1,4-6  and heparin therapy nomograms have been developed to achieve this heparin level.⁷ 

Despite these recommendations, several important issues regarding unfractionated heparin (UFH) therapy remain unresolved, including the minimal time to achieve a therapeutic APTT after starting heparin therapy,^7-11  the optimal intensity of heparin therapy,^7,10,12,13  the required duration of heparin and warfarin therapy “overlap,”^14-16  and the effect of these anticoagulation measures on VTE recurrence after controlling for other predictors of recurrence.^17,18  These issues remain clinically relevant because UFH continues to be preferred over low-molecular-weight heparin therapy for patients with acute PE, high bleeding risk, or impaired renal function.^19,20  Moreover, the Joint Commission performance measures on heparin anticoagulation management,²¹  based largely on ACCP recommendations, prompted many hospitals to establish time-consuming and costly pharmacy- and/or nurse-staffed services solely devoted to heparin monitoring and dose adjustment. Whereas a warfarin anticoagulation intensity international normalized ratio (INR) = 2.0-3.0 is superior to lower intensity (ie, INR = 1.5-1.9) or no anticoagulation as secondary prophylaxis for idiopathic VTE,^22,23  only 2 small studies have addressed the optimal intensity of warfarin anticoagulation for acute VTE therapy: The first reported no INR,²⁴  and the second was limited to idiopathic VTE.²⁵  Therefore, the most appropriate intensity of warfarin anticoagulation for overall acute VTE therapy remains uncertain.

To address these important gaps in knowledge, we performed a large, population-based cohort study to test heparin anticoagulation intensity^4,12  as a predictor of 180-day VTE recurrence after controlling for other baseline characteristics previously identified as independent predictors of VTE recurrence.¹⁷  In addition, we tested time from symptomatic VTE onset to start of heparin therapy, rapid achievement of a therapeutic APTT, and the duration of overlapping heparin and warfarin therapy as predictors of recurrence. Finally, we explored the effect of a higher proportion of time spent above several different “therapeutic” thresholds of heparin and warfarin anticoagulation on VTE recurrence.

Using the unique resources of the Rochester Epidemiology Project (REP),²⁶  we identified the inception cohort of all Olmsted County, MN residents with incident DVT and PE over the 35-year period 1966-2000, as described previously.^27,28  The REP medical records linkage system affords access to comprehensive details regarding all medical care provided to local residents for their entire period of residence in the community. The REP exists because: (1) Olmsted County is isolated from other urban centers, (2) unit patient medical records that combine all inpatient and outpatient data for each resident have been preserved since 1910, and (3) indexes to diagnoses, surgical procedures, and test results have provided access to the patients of interest since 1935. The result is linkage of individual level medical data from all sources of medical care available to and used by the population of Olmsted County, thus ensuring complete ascertainment of all clinically recognized VTE events and outcomes.

For this study, we restricted our analyses to residents with incident VTE over the 14-year period 1984-1997 who lived at least 1 day after the incident VTE event (N = 1166). We started the study time period in 1984 because laboratory data were retrievable electronically beginning in 1983. We ended the time period in 1997 because standard heparin was solely used as initial VTE therapy until 1998 (when low-molecular-weight heparin therapy became available) and our study aim was to address the effect of UFH therapy. We followed each patient from the onset of incident VTE symptoms or signs forward in time using their complete (inpatient and outpatient) medical records in the community for first DVT or PE recurrence as defined previously.^17,28  Recurrent VTE was defined as venous thrombosis of a site that was either previously uninvolved or had interval documentation of incident DVT or PE resolution. For deceased patients, all death certificates and autopsy reports were reviewed regardless of the location at death. The study was approved by the Mayo Clinic and the Olmsted Medical Center institutional review boards.

Using explicit criteria, trained and experienced nurse abstractors reviewed all medical records in the community for consenting cases from date first seen by an REP health-care provider until the earliest of death, date of last medical record follow-up, or 2000 as performed previously.^29,30  Data were collected on date and type of incident VTE, baseline characteristics, dates of heparin and warfarin initiation and completion, total daily heparin dose, inferior vena cava (IVC) filter placement, date and type of first VTE recurrence, bleeding events, and vital status at last clinical contact, as described in detail elsewhere.^17,29,30  Baseline characteristics included hospitalization for surgery (general, orthopedic or gynecologic surgery, neurosurgery) requiring anesthesia (general or neuraxial); hospitalization for acute medical illness; nursing home confinement; trauma and/or fracture requiring hospital admission (major fracture or severe soft tissue injury); active cancer (excluding nonmelanoma skin cancer) with or without chemotherapy (cytotoxic or immunosuppressive therapy for malignancy, excluding tamoxifen); serious neurologic disease (stroke or other disease affecting the nervous system with associated extremity paresis or acute stroke with extremity paresis requiring hospitalization within the previous 3 months); for women only: pregnancy or postpartum (within 3 months of delivery) at the time of the incident event, oral contraceptive use, and hormone therapy (estrogen or progesterone); congestive heart failure; chronic lung disease (chronic obstructive pulmonary disease, emphysema, chronic bronchitis, bronchiectasis, interstitial lung disease, pulmonary hypertension, and asthma included only if documented evidence of fixed airflow obstruction); chronic liver disease (including active hepatitis within the previous 3 months); and chronic renal disease (physician's diagnosis and creatinine > 175 μmol/L [2 mg/dL] for at least 3 months, or nephrotic syndrome). The characteristics related to hospitalization for surgery or acute medical illness, nursing home confinement, trauma/fracture, pregnancy/postpartum state, and oral contraceptive or hormone therapy were recorded as present only if documented within the 3 months before the VTE event. Active cancer had to be documented in the 3 months before or after the incident VTE event. All other characteristics were recorded as present if documented any time before the incident VTE event. Major bleeding was defined as CNS, intraocular, mediastinal, pericardial, or retroperitoneal bleeding or visible bleeding with a ≥ 20 g/L hemoglobin decrement or ≥ 2 blood product units transfusion. We also retrieved all complete blood count (CBC), APTT, and prothrombin time (PT)/INR values from the Mayo Clinic Laboratory Information System.

Over the course of the 14-year study period, the 2 successive APTT reagent and coagulometer instrument combinations used were highly correlated (r² = 0.949), as were APTT reagent lot-to-lot comparisons. The target APTT was derived by comparison with chromogenic anti-X_a levels, as recommended previously.^4-6  An APTT threshold of ≥ 58 seconds (corresponding to a plasma heparin level ≥ 0.3 anti-X_a U/mL) was used in the primary analysis of heparin therapy on VTE recurrence, and thresholds of ≥ 40, ≥ 70 or ≥ 90 seconds (corresponding to plasma heparin levels ≥ 0.2, 0.5, or 0.9 anti-X_a U/mL, respectively) were tested in secondary analyses.

Descriptive statistics were used to summarize the demographic and clinical characteristics of the cohort, including counts and percentages for categorical data and Kaplan-Meier rates for “survival” outcomes (cumulative incidence) such as VTE recurrence and major bleeding. For continuous data, means and standard deviations and/or medians along with ranges or interquartile ranges are reported. For formal comparisons of 2 or more groups, such as subjects who did versus did not receive any heparin during their 180-day follow-up, ANOVA methods (t test if only 2 groups) and χ² tests were used for continuous and categorical variables, respectively.

Using Cox proportional hazards (PH) modeling, we tested demographic, baseline, and time-dependent characteristics as potential predictors of the rate of VTE recurrence from 1-180 days. Demographic and baseline characteristics included patient age and body mass index (BMI) at the incident VTE event and sex, active cancer, neurologic disease with leg paresis, and neurosurgery within 3 months before the incident VTE event.¹⁷  In addition, we created a baseline characteristic termed “idiopathic VTE,” as defined previously.³⁰ 

A time-dependent structure of the data was used to account for variables that changed over time, including hemoglobin, platelet count, APTT, PT, and INR. For hemoglobin and platelet count, the daily average was used if assayed more than once per day. In addition, the presence or absence of heparin or warfarin therapy were assessed as time-dependent covariates, as were time from onset of VTE symptoms to start of heparin therapy and duration of overlapping therapy with heparin and warfarin. Using all daily measurements, we calculated the cumulative proportions of time spent with an APTT ≥ 58, ≥ 40, ≥ 70, and ≥ 90 seconds during heparin treatment, and the cumulative proportions of time spent with an INR ≥ 1.5 and ≥ 2.0 during warfarin treatment. A complete description of the time dependent variables is provided in the supplemental Appendix (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

All variables were initially tested for an association with rate of VTE recurrence in univariate Cox PH models. Those variables demonstrating a univariate association with at least marginal significance (P < .10) were included in a multivariable model, with the exception of idiopathic VTE because of collinearity with other covariates included in the idiopathic VTE definition.

Absolute risk reduction was estimated by comparing: (1) the recurrence risk (number of events or rate) derived from applying the fitted Cox PH regression model to the observed data, with (2) the recurrence risk derived from applying the same fitted model to a hypothetical dataset reflecting the assumption that continual APTT monitoring and heparin dose adjustment would lead to a proportion of time in the “therapeutic” APTT range (APTT ≥ 40 or ≥ 58 seconds) that was no less than the median proportion of time in therapeutic APTT range observed in our cohort. The hypothetical estimates for the rate and number of recurrences under this simulated scenario were derived and then subtracted from the original set of estimates to provide the absolute risk reduction. We calculated the number needed to treat (NNT) to prevent one additional VTE recurrence as the inverse of the absolute risk reduction. Similar analyses were performed for the duration of overlapping heparin and warfarin therapy (in days). Technical details of the analyses are provided in the supplemental Appendix.

Finally, we used Cox PH regression to determine whether the time-dependent measures of heparin and warfarin anticoagulation intensity were associated with major bleeding from 1-180 days. However, because of the low number of major bleeding events, these analyses were limited to univariate modeling and were considered exploratory.

All analyses were performed using SAS Version 8.2 (SAS Institute). P < .05 was considered statistically significant.

Over the 14-year period 1984-1997, 1353 Olmsted County residents developed an incident DVT and/or PE, 1166 of whom lived for at least 1 day after their incident VTE onset and were included in the analyses. The demographic and baseline characteristics (including laboratory data) of these 1166 patients are shown in Table 1. Of these 1166 patients, 1023 (88%) were objectively diagnosed, whereas 1025 (88%) and 987 (85%), respectively, received heparin therapy (subcutaneous heparin, n = 16) and warfarin therapy. The proportion of patients on heparin and/or warfarin therapy by day from the incident VTE event is shown in Figure 1. The 141 patients not receiving heparin were more likely to have neurologic disease, congestive heart failure, or recent neurosurgery and were less likely to have received warfarin. Of these 141 patients, 31 received an IVC filter. Of the entire cohort, 1159 (99%) completed follow-up to 180 days; only 7 patients were lost to follow-up before 180 days.

The median duration of heparin therapy was 6 days, and 76% of patients received ≥ 5 days of heparin therapy (Table 2). The median duration of heparin therapy with an APTT ≥ 58 seconds was nearly 4 days, and 33% of patients received heparin therapy with an APTT ≥ 58 seconds for at least 5 days; heparin durations with an APTT ≥ 40, ≥ 70 and ≥ 90 seconds are shown in Table 2. The median time from VTE onset to start of heparin therapy was 1.4 days (range 0-30 days), and of overlapping heparin and warfarin was 4.0 days (mean ± SD = 4.0 ± 3.0; range: 0-14 days; interquartile range: 2-6 days). Both the time interval from start of heparin to start of warfarin and the total duration of heparin decreased in later calendar years (Spearman correlation coefficients = −0.23 and −0.13, respectively; P < .001 for both).

Over 5461 person-years of follow-up, 254 (22%) patients developed recurrent VTE; 85 (8%) events occurred within 180 days and served as outcomes in this analysis. The 14-, 90-, 180- and 14- to 180-day cumulative incidence rates of VTE recurrence are shown in Figure 2A. Of the VTE recurrences, 31 were PE ± DVT, 53 were DVT alone, and 1 was chronic thromboembolic pulmonary hypertension. The 2-week case fatality rates for recurrent DVT alone and recurrent PE ± DVT were 2% and 11%, respectively.

In univariate analyses, the duration of time from VTE symptom or sign onset to start of heparin therapy, as well as the time-dependent variable “having started heparin therapy,” were not associated with recurrence (Table 3). However, rapidly achieving an APTT ≥ 58 or ≥ 40 seconds and increasing proportions of time on heparin with an APTT ≥ 58 seconds or ≥ 40 seconds were associated with significantly reduced hazards of recurrence. Among patients with available heparin dose data, the mean initial heparin dose was significantly greater among those with VTE recurrence within 14 days (n = 14) compared with a random sample of those without early recurrence (n = 84; 33 390 ± 7529 vs 26 083 ± 10 859 U/d; P = .02), and not significantly different from the calculated dose (n = 14; 35 705 ± 8855; P = .48 from paired t test) using a weight-based heparin dosing nomogram.⁷  Whereas the duration of heparin and warfarin overlap was not associated with recurrence, the proportions of time on warfarin with an INR either ≥ 1.5 or ≥ 2.0 as well as the time-dependent variable of being on warfarin therapy were all significantly associated with reduced hazards of recurrence. In contrast, IVC filter placement was associated with a significantly increased hazard of recurrence. Of the 17 patients with recurrent VTE after IVC filter placement, 5 (29%) had recurrent PE ± DVT.

In multivariable analyses including all variables that were at least marginally significant (P < .10) in univariate analyses, rapidly achieving an APTT ≥ 58 seconds was associated with a reduced hazard of recurrence, but the proportion of time on heparin with an APTT ≥ 58 seconds (per 10% increase) was not associated with recurrence (Table 4). Conversely, in a separate multivariable analysis using a lower APTT threshold, rapidly achieving an APTT ≥ 40 seconds was not associated with recurrence, but the proportion of time on heparin with an APTT ≥ 40 seconds (per 10% increase) was associated with a reduced hazard of recurrence (Table 4). A higher proportion of time on warfarin therapy with an INR ≥ 2.0 significantly reduced the hazard of recurrence, whereas the proportion of time on warfarin with 1.5 ≤ INR < 2.0 was not associated with recurrence; IVC filter placement nonsignificantly increased the hazard of VTE recurrence (Table 4).

In univariate analyses of previously identified predictors of VTE recurrence,¹⁷  the hazards of 180-day VTE recurrence were significantly increased for active cancer, neurologic disease with leg paresis, and recent neurosurgery (and correspondingly reduced for idiopathic VTE), and were marginally decreased for low BMI (Table 3); patient age and sex were not predictors of recurrence in this updated inception cohort. The 180-day cumulative incidence of recurrent VTE after idiopathic incident VTE was 4% (n = 13) compared with 16% (n = 32) after active cancer-related incident VTE. Adjusting for warfarin therapy did not account for the reduced hazard of recurrence associated with idiopathic VTE (compared with VTE associated with active cancer, neurologic disease, or other causes), nor did adjusting for rapidly achieving a therapeutic APTT (data not shown). In univariate analysis, a higher daily mean hemoglobin level was significantly associated (hazard ratio [HR] = 0.87 per g/L; P = .02) with a lower recurrence rate. Adjusting for warfarin therapy slightly attenuated this effect (HR = 0.90; P = .09). In the multivariable analysis that incorporated anticoagulation variables, active cancer was the only one of these baseline characteristics that was independently associated with an increased hazard of 180-day VTE recurrence (Table 4); BMI, leg paresis, and neurosurgery were not significant predictors of recurrence after controlling for measures of heparin and warfarin therapy. Analyses restricted to patients with objectively diagnosed incident VTE yielded essentially the same results (data not shown).

Thirty-one (3%) patients had a bleeding episode within 180 days of the incident VTE event; of these, 12 (1%) were major bleeding events (Figure 2B). The incidence of major bleeding was highest within the first 14 days (16.2 per 100 person-years), and progressively decreased to 4.4 and 2.5 per 100 person-years by 90 and 180 days, respectively. The 14- and 30-day case fatality rates after a major bleeding event were 8% (95% confidence interval [CI]: 1%-49%) and 25% (95% CI: 9%-59%), respectively. In univariate analyses, the hazard of major bleeding was not significantly increased for proportion of time on heparin with an APTT ≥ 40 seconds; however, the hazards were marginally and significantly increased for proportion of time on heparin with an APTT ≥ 58 seconds (per 10% increase, HR = 1.18; 95% CI: 0.97, 1.45; P = .101) and ≥ 70 seconds (per 10% increase, HR = 1.21; 95% CI: 1.01, 1.44; P = .034), respectively. In similar analyses, the hazards of major bleeding were not significantly increased for the proportions of time on warfarin with an INR ≥ 1.5 or ≥ 2.0, respectively; however, the hazards were significantly increased for a higher daily mean prothrombin time (HR = 1.04; 95% CI: 1.00, 1.08; P = .036) and for a higher daily mean INR (HR = 1.21; 95% CI: 1.09, 1.35; P < .001) while on warfarin.

To further explore the potential effect of APTT monitoring and heparin dose adjustment on VTE recurrence, we estimated the absolute reduction in risk of recurrence that would be expected if the proportion of time spent with an APTT at: (1) ≥ 40 seconds or (2) ≥ 58 seconds was maintained, in each individual, at or above the respective cohort medians for these parameters (92% for ≥ 40 seconds and 63% for ≥ 58 seconds). For the ≥ 40-second threshold, using the estimated β coefficient, the model predicted an absolute risk reduction of 1.4% in the180-day VTE recurrence rate, resulting in 12 fewer recurrent VTE events (NNT = 71). If the upper and lower confidence limits on the β coefficients are used, then 4 or 22 fewer events, respectively, are predicted. For the same calculation applied to the ≥ 58-second threshold, the predicted reduction (which was not statistically significant) would have been 7 fewer events (point estimate; absolute risk reduction = 0.7%; NNT = 141). If the upper and lower confidence limits on the β coefficients are used, then 9 additional or 19 fewer events, respectively, are predicted. To further explore the potential effect of duration of overlapping heparin and warfarin therapy on VTE recurrence, we estimated the absolute reduction in risk of recurrence that would have occurred if the overlap duration was maintained, in each individual, at or above the cohort median 5-day duration. Using the estimated β coefficient, the risk reduction in the model-predicted 180-day VTE recurrence rate would be 0.1%, resulting in 1 fewer recurrent VTE events (NNT = 883). If the upper and lower confidence limits on the β coefficients are used, then 4 fewer and 2 more events, respectively, are predicted.

In this population-based inception cohort, in which all incident VTE cases were included and the data on initial heparin and subsequent warfarin therapy, laboratory data, and the outcomes of VTE are complete, our observed VTE recurrence and bleeding rates were similar to a contemporary study of UFH and warfarin therapy for acute VTE.³¹  Our most noteworthy findings, however, were related to the effect of heparin anticoagulation measures on VTE recurrence. Specifically, we found that a greater proportion of time spent on heparin but with a lower intensity of anticoagulation (ie, ≥ 0.2 anti-X_a U/mL) provided a significantly reduced hazard of VTE recurrence. In contrast, the proportion of time spent above the currently recommended intensity of heparin anticoagulation was not significantly associated with fewer events. Moreover, achieving a lower intensity of anticoagulation for the duration of heparin therapy is feasible in practice, as reflected by the respective 92% and 100% median and upper quartile proportions of time with an APTT ≥ 40 seconds while on heparin among our cohort of community patients. Current recommendations for a “therapeutic” APTT corresponding to a heparin level of 0.3-0.7 anti-X_a U/mL are largely extrapolated from animal model data,^32-34  and a single small cohort study of 234 patients (162 patients with mainly clinically diagnosed VTE) who were treated with IV UFH (5000 U IV bolus followed by initial dose of 24 000 U/24 hours).^3,4,12  APTT monitoring and UFH dose adjustment to maintain the APTT between 1.5 and 2.5 times that of the controls was recommended but not enforced. VTE recurrence was clinically diagnosed in 5 patients within 12 days of beginning UFH, but only 4 of these patients were actually being treated for VTE. No information was provided about whether these VTE patients were being treated for incident or recurrent VTE, and the investigators did not control for other characteristics that are predictors of VTE recurrence (eg, cancer).

Rapidly achieving an APTT ≥ 0.3 anti-X_a U/mL significantly protected against 180-day VTE recurrence. Whereas initial studies found an association between an early subtherapeutic APTT and VTE recurrence,^8,10  2 meta-analyses of randomized trials failed to support this observation.^9,11  Given the often prolonged delay by patients in seeking medical attention for symptoms of VTE, it seems biologically implausible to suggest that achieving a certain APTT level within the next 24 hours affects VTE recurrence. Indeed, we found no significant relationship between the time from onset of VTE symptoms to start of heparin therapy and the hazard of recurrence. Moreover, the failure to rapidly achieve a target APTT did not appear to reflect heparin underdosing; the initial heparin dose in a subset of patients with heparin dose information available was significantly higher for those patients with (n = 14) compared with those without (n = 84) early recurrence, and was also not significantly different from a calculated dose using a weight-based heparin dosing nomogram.⁷  Therefore, we believe that these patients were relatively heparin resistant, possibly because of high factor VIII activity,^35,36  which is a predictor of VTE recurrence.³⁷ 

The duration of overlapping heparin and warfarin therapy was not a predictor of VTE recurrence. Given the narrow CIs around our estimated HR of recurrence per day of overlap, it is unlikely that a longer duration of heparin/warfarin overlap provides an important reduction in VTE recurrence risk, and our simulated analyses support this conclusion. The recommendation that heparin and warfarin therapy be overlapped for at least 5 days and until the INR is ≥ 2.0 for at least 24 hours¹  is largely extrapolated from animal model studies.^38,39  Two small clinical trials showed that 5-7 and 10-14 days of IV heparin therapy were similarly effective for acute proximal DVT,^14,15  but no clinical trials tested the current recommendation for a 4- to 5-day heparin/warfarin overlap. In a large cohort study, the 6-month incidence of recurrent VTE did not differ for lengths of initial hospitalization for acute DVT ranging from 3-10 days.¹⁶ 

Our finding that a higher proportion of time on warfarin with an INR ≥ 2.0 was associated with a significantly reduced hazard of VTE recurrence is consistent with studies of acute therapy for idiopathic VTE²³  and secondary prophylaxis against recurrent VTE.²⁵  Our finding of an increased hazard of recurrence associated with IVC filter placement, although not statistically significant, is supported by a previous study.⁴⁰  Moreover, almost 1/3 of patients with VTE recurrence after IVC filter placement had recurrent PE. These findings suggest that IVC filter placement is inadequate VTE therapy and that anticoagulation should be started as soon as the bleeding risk is acceptable.

In an earlier analysis of the 1719 Olmsted County residents who survived at least 1 day after an incident VTE over the 25-year period 1966-1990, we identified increasing patient age and BMI, active cancer, neurologic disease with leg paresis, and recent neurosurgery as independent predictors of VTE recurrence out to 10 years.¹⁷  However, we were unable to control for the intensity of heparin and warfarin anticoagulation because APTT and INR data were not easily retrievable. In this update of our Olmsted County VTE inception cohort through 1997, we found that, after controlling for heparin and warfarin therapy, only active cancer was an independent predictor of 180-day VTE recurrence, with an approximately 3-fold increased hazard of recurrence. These findings corroborate studies showing a high rate of warfarin failure among active cancer patients.^41,42  Neither sex nor idiopathic VTE were independent predictors of 180-day VTE recurrence. Sex was not an independent predictor of recurrence in our previous study,¹⁷  a finding that was recently confirmed,⁴³  and idiopathic VTE was not a predictor of recurrence during initial anticoagulation therapy in previous studies.^42,44 

The number of major bleeding events (n = 12) was too few for meaningful multivariable analysis. However, our exploratory findings of: (1) no significant association of major bleeding with the proportion of time on heparin with an APTT ≥ 40 seconds, (2) a marginally significant association with the proportion of time with an APTT ≥ 58 seconds (heparin level 0.3 anti-X_a U/mL), and (3) a significant association with the proportion of time with APTT ≥ 70 seconds (heparin levels of 0.5 anti-X_a U/mL) lend further support to a lower intensity of heparin anticoagulation as therapy for acute VTE. Moreover, whereas a higher mean daily prothrombin time/INR on warfarin is associated with a significantly increased hazard of major bleeding, the hazard is not significantly increased for proportion of time on warfarin with an INR ≥ 2.0. We interpret this finding to suggest that the immediate daily prothrombin time/INR is more relevant to a warfarin-associated bleeding complication than the cumulative time on warfarin with an INR ≥ 2.0. We believe that this further supports a standard intensity of warfarin anticoagulation as therapy for acute VTE. These exploratory results require confirmation in future studies.

Our study has several important strengths. The population-based study design insured that the entire spectrum of VTE disease occurring in the community was included, so our results are generalizable to populations of similar demography and baseline characteristics. We accurately separated incident from recurrent VTE events, used an unambiguous definition of VTE recurrence, and our cohort follow up to 6 months was virtually complete. Our sample size was relatively large and the observed VTE recurrence rate was comparable to contemporary studies. The APTT and PT/INR assays and assay instruments were standardized across the entire study time frame, and we had access to all laboratory results in the analyses. We controlled for all known baseline and time-dependent characteristics that could potentially affect VTE recurrence, and we used a “carry-forward” method for calculation of proportion of time in therapeutic range that avoids the use of future information to bias present data, as is the case with linear interpolation.⁴⁵  However, our study also has important limitations. Because of our observational cohort study design, we could not ensure equal and random patient allocation to differing durations and intensities of heparin and warfarin anticoagulation or to differing durations of overlapping heparin and warfarin therapy. However, the frequency distribution of these characteristics in the cohort was sufficiently broad that we had adequate power to assess their potential effects on VTE recurrence.

Our findings have several implications. First, if providers choose to monitor and adjust the heparin dose according to the APTT despite a heparin dose ≥ 30 000 U/d, a lower intensity of heparin anticoagulation (≥ 0.2 anti-X_a U/mL) may be adequate. However, as has been suggested by others^1,18  and shown in one clinical trial,³¹  with a heparin dose ≥ 30 000 U/d, routine APTT (or anti-X_a⁴⁶ ) monitoring and heparin dose adjustment may be unnecessary because the median proportion of time with an APTT ≥ 0.2 anti-X_a U/mL was > 90%. Second, initiating heparin and warfarin concurrently and stopping the heparin when the INR is ≥ 2.0 regardless of the duration of overlapping heparin and warfarin therapy appears to be as effective as the recommended 4-5 days of heparin/warfarin overlap. If confirmed by focused clinical trial results, these 2 practice changes have the potential to result in important cost savings by reducing the duration of hospitalization and avoiding the need for costly heparin monitoring programs. Moreover, the duration of heparin exposure likely will be shortened, reducing the risk of heparin-induced thrombocytopenia.⁴⁷  Finally, in accordance with current recommendations,¹  VTE patients with active cancer should remain on anticoagulation (preferably low-molecular-weight heparin⁴¹ ) for at least 6 months, and probably as long as the cancer remains active.

This study was funded in part by grants from the National Institutes of Health (HL66216 and AG034676), the US Public Health Service, and the Mayo Foundation.

National Institutes of Health

Contribution: J.A.H. designed and performed the research, collected the data, analyzed and interpreted the data, and wrote the manuscript; B.D.L. performed the statistical analyses and contributed to writing the manuscript; T.M.P. designed and performed the research, collected the data, performed the statistical analyses, and contributed to writing the manuscript; K.R.B. directed the statistical analyses and contributed to writing the manuscript; A.A.A. analyzed and interpreted the data and contributed to writing the manuscript; and L.J.M. contributed to the research design, analyzed and interpreted the data, and contributed to writing the manuscript.

Conflict-of-interest disclosure: J.A.H. has served on advisory boards for which he received honoraria. The remaining authors declare no competing financial interests.

Correspondence: John A. Heit, MD, Stabile 6-Hematology Research, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: heit.john@mayo.edu.