The American Journal of Medicine
Volume 118, Issue 6 , Pages 625-635, June 2005

Oral anticoagulation strategies after a first idiopathic venous thromboembolic event

  • Drahomir Aujesky, MD, MSc

      Affiliations

    • Division of General Internal Medicine, Department of Medicine, University of Pittsburgh, Pennsylvania
    • VA Center for Health Equity Research and Promotion and VA Pittsburgh Healthcare System, Pennsylvania
    • Corresponding Author InformationRequests for reprints should be addressed to Drahomir Aujesky, MD, MSc, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System, University Drive C, Building #28, Suite 1A129, Pittsburgh, PA 15240
    • ,
  • Kenneth J. Smith, MD

      Affiliations

    • Division of General Internal Medicine, Department of Medicine, University of Pittsburgh, Pennsylvania
    • Section of Decision Sciences and Clinical Systems Modeling, Department of Medicine, University of Pittsburgh, Pennsylvania
    • ,
  • Mark S. Roberts, MD, MPP

      Affiliations

    • Division of General Internal Medicine, Department of Medicine, University of Pittsburgh, Pennsylvania
    • Section of Decision Sciences and Clinical Systems Modeling, Department of Medicine, University of Pittsburgh, Pennsylvania

Article Outline

Abstract 

Purpose

The optimal duration and intensity of warfarin therapy after a first idiopathic venous thromboembolic event are uncertain. We used decision analysis to evaluate clinical and economic outcomes of different anticoagulation strategies with warfarin.

Methods

We built a Markov model to assess 6 strategies to treat 40- to 80-year-old men and women after their first idiopathic venous thromboembolic event: 3-month, 6-month, 12-month, 24-month, and unlimited-duration conventional-intensity anticoagulation (International Normalized Ratio, 2–3) and unlimited-duration low-intensity anticoagulation (International Normalized Ratio, 1.5-2). The model incorporated age- and sex-specific clinical parameters, utilities, and costs. Using a societal perspective, we compared strategies based on quality-adjusted life-years (QALYs), lifetime costs, and incremental cost-effectiveness ratios.

Results

In our baseline analysis, incremental cost-effectiveness ratios were lower in younger patients and in men, reflecting the higher bleeding risk at older ages, and the lower risk of recurrence among women. Based on a willingness-to-pay of <$50000/QALY, the 24-month strategy was most cost-effective in 40-year-old men ($48805/QALY), while the 6-month strategy was preferred in 40-year-old women ($35977/QALY) and 60-year-old men ($29878/QALY). In patients aged ≥80 years, 3-month anticoagulation was less costly and more effective than other strategies. Cost-effectiveness results were influenced by the risks associated with recurrent venous thromboembolism, the major bleeding risk of conventional-intensity anticoagulation and the disutility of taking warfarin.

Conclusion

Longer initial conventional-intensity anticoagulation is cost-effective in younger patients while 3 months of anticoagulation is preferred in elderly patients. Patient age, sex, clinical factors, and patient preferences should be incorporated into medical decision making when selecting an appropriate anticoagulation strategy.

Keywords:  Warfarin , Idiopathic venous thromboembolism , Effectiveness , Cost effectiveness

 

Idiopathic venous thromboembolism, defined as deep vein thrombosis or pulmonary embolism of unknown cause, carries a persistent risk of recurrence.1 Warfarin at a target International Normalized Ratio (INR) of 2–3 (the conventional intensity) is highly effective in preventing recurrence but can cause major bleeding.2, 3 Recent clinical trials demonstrated that increasing the anticoagulation duration from 3 months to 6 or 12 months does not reduce the long-term risk of recurrence in patients with idiopathic venous thromboembolism, suggesting that unlimited-duration anticoagulation is needed to prevent recurrent venous thromboembolism.4, 5, 6 In addition, trials of low-intensity anticoagulation (target INR 1.5–2) are less efficacious in preventing recurrent venous thromboembolism events than is long-term conventional-intensity anticoagulation without lowering the risk of bleeding.7, 8

At present, the optimal duration and intensity of anticoagulation after a first idiopathic venous thromboembolic event are uncertain, with tradeoffs between the risk of recurrence and the risks, costs, and inconvenience of anticoagulation complicating medical decision making. Reflecting this uncertainty, recommendations for anticoagulation after a first idiopathic venous thromboembolic event vary from 3 months to unlimited-duration.2, 9, 10 We therefore constructed a decision analytic model to estimate the clinical and economic outcomes of 6 validated anticoagulation strategies: 3-month, 6-month, 12-month, 24-month, and unlimited-duration treatment with conventional-intensity anticoagulation and unlimited-duration treatment with low-intensity anticoagulation. Because the risk of bleeding increases with age, and women with idiopathic venous thromboembolism may have a substantially lower risk of recurrence than men,11, 12, 13, 14 we performed separate analyses by age and sex.

Back to Article Outline

Methods 

We performed a cost-effectiveness analysis by developing a Markov model that simulated transitions between several states:15 well with or without oral anticoagulation; recurrent deep vein thrombosis or pulmonary embolism; severe post-thrombotic syndrome; minor and major bleeding complications; permanent disability after hemorrhagic stroke; vena cava filter treatment; or death from recurrent pulmonary embolism, major bleeding, filter implantation, or natural causes. We followed the recommendations of the Panel on Cost-Effectiveness in Health and Medicine by taking the societal perspective and discounting future costs and benefits at an annual rate of 3%.16 We constructed our model using DATA Professional software (TreeAge, Williamstown, MA).

Clinical problem 

The hypothetical cohorts comprised 40- to 80-year-old men and women 3 months after a first episode of idiopathic deep vein thrombosis or pulmonary embolism. We assumed that all patients had successfully completed a standard 3-month course of conventional-intensity anticoagulation before they entered our model.

Anticoagulation strategies 

Patients underwent 1 of 6 anticoagulation strategies: discontinuation of anticoagulation (ie, the 3-month strategy); conventional-intensity anticoagulation for 3 additional months (ie, the 6-month strategy); conventional-intensity anticoagulation for 9 additional months (ie, the 12-month strategy); conventional-intensity anticoagulation for 21 additional months (ie, the 24-month strategy); unlimited-duration conventional-intensity anticoagulation; and unlimited-duration low-intensity anticoagulation. Patients were followed clinically, with regular INR checks for patients receiving anticoagulants.

Clinical assumptions 

First, the model assumed a high early risk of recurrence for venous thromboembolism during the first year after the discontinuation of anticoagulation regardless of anticoagulation duration and a lower, constant late risk of recurrence.4, 5, 6, 7, 10, 13, 17, 18 Second, recurrent venous thromboembolism was always followed by unlimited-duration conventional-intensity anticoagulation. Third, the model assumed a baseline risk of bleeding even without anticoagulation: the risk was augmented based on anticoagulation intensity and patient age. Fourth, because of the high risk of recurrent hemorrhage when warfarin therapy is reinstituted after major bleeding, a major bleeding episode always led to permanent discontinuation of anticoagulation.19 If venous thromboembolism recurred after a major bleeding episode, an inferior vena cava filter was implanted to prevent pulmonary embolism.2 Fifth, the model assumed that patients who experienced intracerebral bleeding became permanently disabled and died at high rates.20, 21 Sixth, if venous thromboembolism recurred while the patient was receiving conventional-intensity warfarin, 2 scenarios were possible. If venous thromboembolism recurred when the INR was in the therapeutic range (2–3), anticoagulation therapy was continued and a vena cava filter was implanted. If the disease recurred when the INR was subtherapeutic (<2), anticoagulation therapy was continued and no filter was implanted. The model assumed that filter implantation permanently increased the risk of deep vein thrombosis.22 Finally, all patients were at risk of developing a severe post-thrombotic syndrome, defined by venous leg ulcers. This risk was increased in patients with recurrent deep vein thrombosis.1

Parameter values 

We abstracted probabilities for risks of recurrent venous thromboembolism, bleeding complications, and vena cava filter implantation from the peer-reviewed literature (Table 1). We used quality of life measures (Table 2) and cost data (Table 3) from a variety of published data sources. We modeled life expectancy using sex- and age-specific survival data from 2000 US life tables.23

Table 1. Estimates for clinical parameters
Clinical parametersBaseline estimate (range)References or sources
Venous thromboembolism
Risk of recurrence with conventional-intensity anticoagulation, %/year0.7(0.7–1.3)(5, 8, 17)
Relative risk of low- vs. conventional-intensity anticoagulation2.8(1.1–7)(8)
Risk of recurrence after anticoagulation is stopped
Early risk (year 1), %10(7–27)(4, 5, 6, 7, 10, 13, 17, 18)
Late risk (>year 1), %/year5(2–8)(4, 5, 6, 7, 13, 17, 18)
Relative risk of recurrence of women vs men0.7(0.3–1)(7, 13, 14, 17, 24, 25)
Likelihood of venous thromboembolism recurring as pulmonary embolism, %32(25–38)(4, 5, 6, 7, 8, 17)
Case-fatality rate of pulmonary embolism, %16(7–29)(5, 6, 7, 8, 17)
Lifetime risk of severe post-thrombotic syndrome, %9.3(5.6–12.8)(1)
Relative risk following recurrent after deep vein thrombosis6.4(3.1–13.3)(1)
Outpatient treatment for deep vein thrombosis, %30(0–100)Estimate
Warfarin therapy
Risk of major bleeding with conventional-intensity anticoagulation, %/year (8, 11, 12, 26)
40–49 years1.0(0.5–1.5)
50–59 years1.5(0.8–2.3)
60–69 years2.3(1.5–3.4)
70–79 years3.4(2.3–5.1)
≥80 years5.1(3.4–7.6)
Risk of major bleeding without warfarin, %/year0.4(0–1)(7)
Risk of minor bleeding with conventional-intensity anticoagulation, %/year7.7(2.9–9.6)(8, 17)
Risk of minor bleeding without warfarin, %/year1.4(1.0–6.7)(7, 17)
Relative bleeding risk for low- vs. conventional-intensity anticoagulation1(0.4–3)(8)
Case-fatality rate of major bleeding, %13(9–17)(48)
Hemorrhagic stroke, %9(6–13)(48)
Mortality following hemorrhagic stroke, %/year18(17–52)(20, 21)
Likelihood of subtherapeutic International Normalized Ratio, %28(14–47)(49)
Vena cava filter
Efficacy in preventing pulmonary embolism, %78(10–95)(22)
Case-fatality rate of filter placement, %0.2(0–0.4)(50)
Table 2. Estimates for quality-of-life measures
Quality-of-life measuresBaseline estimate (range)References or sourcesNotes
Permanent disability following hemorrhagic stroke0.60(0.02–1)(51)
Warfarin therapy0.99(0.92–1)(52, 53)
Severe post-thrombotic syndrome0.95(0.79–1)(51)
Disutilities for acute complications, days⁎
Recurrent pulmonary embolism6.7(3.4–10.1)(30)Assumes 6.7 inpatient days lost because of pulmonary embolism
Recurrent deep vein thrombosis5.5(2.8–8.3)(30)Assumes 5.5 inpatient or outpatient days lost because of deep vein thrombosis
Minor bleeding1(0.5–1.5)(28)Assumes 1 outpatient day lost because of minor bleeding
Noncerebral major bleeding
Bleeding while hospitalized2.4(1.2–3.6)(30)Assumes 2.4 additional inpatient days lost because of upper gastrointestinal bleeding
Bleeding after discharge4.8(2.4–7.2)(30)Assumes 4.8 inpatient days lost because of upper gastrointestinal bleeding

⁎ Disutilities for acute complications are expressed as days lost from quality-adjusted life expectancy because of hospitalization/disease. The number of days lost because of hospitalization/disease is equal to the average length of hospital stay for a given illness in 2001 or 1 outpatient day for minor bleeding.

Table 3. Cost estimates
Costs (2002 US dollars)Baseline estimate (range)References or sourcesNotes
Anticoagulation with warfarin, per year1272(636–1908)(32, 33, 34)Average Medicare reimbursement for 4 office visits per year and for 2 INR measurements per month; costs for warfarin, patient time⁎ and transportation†
Follow-up without anticoagulation, per year264(132–396)(33, 34)Average Medicare reimbursement for 4 office visits per year; costs for patient time⁎ and transportation†
Pulmonary embolism treatment7102(3551–10653)(30, 33)Average Medicare reimbursement for diagnosis-related group 078 (pulmonary embolism) for 6.7 hospital days, physician visits, and helical CT interpretation
Deep vein thrombosis outpatient treatment722(361–1083)(32, 33, 34)Average Medicare reimbursement for office visits, duplex examination, and INR measurements; costs for a 5-day course of enoxaparin, patient time,⁎ and transportation†
Deep vein thrombosis inpatient treatment3653(1827–5480)(30, 33)Average Medicare reimbursement for diagnosis-related group 128 (deep vein thrombophlebitis) for 5.5 hospital days, physician visits, and duplex examination
Minor bleeding outpatient treatment73(37–110)(33, 34)Average Medicare reimbursement for 1 office visit and 1 INR measurement; costs for patient time⁎ and transportation†
Cost of noncerebral major bleeding
Bleeding while hospitalized2747(1374–4121)(30, 33)Average Medicare reimbursement for diagnosis-related group 174 (upper gastrointestinal bleeding) for 2.4 additional hospital days, physician visits, and upper endoscopy
Bleeding after discharge5317(2659–7976)(30, 33)Average Medicare reimbursement for diagnosis-related group 174 (upper gastrointestinal bleeding) for 4.8 hospital days, physician visits, and upper endoscopy
Cost of hemorrhagic stroke
Stroke while hospitalized10748(5374–16122)(30, 33, 54)Average Medicare reimbursement for diagnosis-related group 014 (cerebrovascular afflictions excluding ischemia) for 3 additional hospital days, physician visits, cerebral CT interpretation, and stroke rehabilitation
Stroke after discharge13781(6891–20672)(30, 33, 54)Average Medicare reimbursement for diagnosis-related group 014 (cerebrovascular afflictions excluding ischemia) for 5.9 hospital days, physician visits, cerebral CT interpretation, and stroke rehabilitation
Care post-hemorrhagic stroke, per year12420(6210–18630)(44)
Vena cava filter placement482(241–723)(33)Average Medicare reimbursement
Severe postthrombotic syndrome, per year1839(920–2760)(55)
Death from any cause5000(2500–7500)Estimate

INR = International Normalized Ratio; CT = computed tomography.

⁎ Based on the average hourly wage of a US nonfarm worker in 2002.

† Estimated at $15.

Clinical parameters 

Clinical trials and prospective cohort studies showed an initially high risk of recurrence for idiopathic venous thromboembolism after anticoagulant discontinuation (“catch-up phenomenon”), leading to a similar long-term incidence of recurrent venous thromboembolism, regardless of anticoagulation duration.4, 5, 6, 7, 10, 13, 17, 18 Based on these studies, we modeled 2 levels of risk of recurrence: a high early risk of recurrence of 10% during the first year after anticoagulant discontinuation, then a late risk of recurrence of 5%/year thereafter (Table 1). A large prospective cohort study suggested that women with idiopathic venous thromboembolism have a substantially lower recurrence rate than men (relative risk = 0.3).13 Because this relative risk of recurrence was somewhat higher in other studies,7, 14, 17, 24, 25 we assumed a relative risk of recurrence in women of 0.7 and varied this parameter widely in sensitivity analysis. Because age is not an independent predictor of recurrence in patients with idiopathic venous thromboembolism,7, 13, 17, 24, 25 we assumed the same risk of recurrence across all age groups. We used clinical trials of patients with idiopathic venous thromboembolism to estimate the risk of recurrence in patients taking conventional-intensity anticoagulation as well as those taking low- versus conventional-intensity anticoagulation.5, 8, 17

We modeled recurrent deep vein thrombosis and pulmonary embolism separately, because there are differences in their management and prognosis. We pooled data from clinical trials of idiopathic venous thromboembolism to estimate the likelihood of venous thromboembolism recurring as pulmonary embolism and its case-fatality rate.4, 5, 6, 7, 8, 17 These trials had follow-up rates close to 100%, included both patients with an initial episode of pulmonary embolism and deep vein thrombosis, and reported not only patients who died from objectively confirmed pulmonary embolism but also patients who died from possible pulmonary embolism (eg, sudden death). Thus, it is unlikely that our analysis underestimated the complication rates of recurrent pulmonary embolism.

We obtained age-specific risk of major and minor bleeding associated with conventional-intensity anticoagulation and the risk of bleeding without anticoagulation from clinical trials and other studies.7, 8, 11, 12, 26 To estimate the relative risk of bleeding associated with low-intensity versus conventional-intensity anticoagulation, we used data from the only trial that directly compared the safety of these two treatments in patients with venous thromboembolism (relative risk = 1).8

In patients with major bleeding episodes, we assumed that the frequency of fatal bleeding and hemorrhagic stroke was constant, regardless of anticoagulation intensity.

Quality-of-life measures (utilities) 

We adjusted for quality of life by using estimates of health state utilities (Table 2). Utilities for health states are scaled from 0 (death) to 1 (perfect health). We calculated quality-adjusted life-years (QALYs) by multiplying the time spent in each health state and its corresponding utility. We used the same utility for low-intensity and conventional-intensity therapy. Decreases in utility because of acute complications (noncerebral major bleeding, deep vein thrombosis, pulmonary embolism) were expressed as days of utility lost because of hospitalization.27, 28, 29 To estimate the number of days lost because of hospitalization, we used the mean length of hospital stay for each illness, based on national average length of stay data for each diagnosis-related group.30 For patients with deep vein thrombosis, the same disutility was applied regardless of whether patients were treated in the hospital or received outpatient care (30% of patients). Patients with minor bleeding lost 1 outpatient day because of disease. For patients without complications, we used age- and sex-specific utility values.31

Costs 

Drug costs were the average wholesale prices based on the 2002 Red Book (Table 3).32 Because no evidence exists that low-intensity anticoagulation requires less intensive monitoring than conventional-intensity therapy with warfarin, we assumed the same costs for both anticoagulation intensities. We used 2002 Medicare reimbursement data to estimate costs for hospitalization, office visits, INR measurements, and medical procedures.33 If noncerebral major bleeding or hemorrhagic stroke occurred during hospitalization for venous thromboembolism, patients incurred the costs for 2.4 or 3 additional hospital days (half the average length of stay for these complications). If major bleeding occurred after hospital discharge or in patients with deep vein thrombosis treated at home, patients incurred the full costs of hospitalization for this complication. The occurrence of minor bleeding led to an office visit and an additional INR measurement.

Indirect costs included costs for time spent obtaining care and for transportation.34 Patients spent an estimated 1 hour of time and $15 in transportation costs per office visit and INR measurement. Lost wages attributable to hospitalization, disability, or premature death were excluded, as recommended by the Panel on Cost-Effectiveness in Health and Medicine.16 All costs were adjusted to 2002 US dollars by using the medical the Consumer Price Index.34

Analyses 

For each treatment strategy, our model calculated the unadjusted life expectancy, QALYs, and lifetime costs. We compared treatment strategies using the incremental cost-effectiveness ratio, defined as the extra cost of a specific strategy divided by its extra clinical benefit in QALYs, compared with the next least expensive strategy. Although there is no absolute threshold for cost-effectiveness, ratios <$50 000/QALY gained are usually considered cost-effective.35 We conducted one-way sensitivity analyses on all variables to assess the effect of varying baseline estimates on cost-effectiveness, using either ranges suggested in the literature or by adding or subtractisuggested in the literatureng 50% from baseline estimates. Variables that influenced cost-effectiveness results were selected for probabilistic sensitivity analyses, with parameters varied simultaneously over triangular probability distributions.36 Values from each probability distribution were randomly selected during each of 1000 Monte Carlo iterations. We calculated the net monetary benefit for each strategy for societal willingness-to-pay thresholds of $50000 and $100000/QALY and determined the proportion of iterations for which a given strategy resulted in the highest net monetary benefit.37

Back to Article Outline

Results 

Baseline analysis 

Unlimited-duration conventional-intensity anticoagulation was the most effective strategy in 40-year-old men (16.648 QALYs) (Table 4). The most effective strategies in the other groups were: 24-month anticoagulation in 40-year-old women (17.551 QALYs), 12-month anticoagulation in 60-year-old men (10.093 QALYs), 6-month anticoagulation in 40-year-old women (11.338 QALYs), and 3-month anticoagulation in 80-year-old men (4.487 QALYs) and 80-year-old women (5.182 QALYs). Overall, all conventional-intensity anticoagulation strategies were similarly effective across all age groups, with the difference between the least and most effective never greater than 53 quality-adjusted life-days. Unlimited-duration low-intensity anticoagulation was always the least effective option. Effectiveness results were similar for unadjusted life expectancy. Three-month anticoagulation was always the least expensive strategy while unlimited-duration low-intensity anticoagulation cost the most.

Table 4. Results of baseline analysis by age and sex
Anticoagulation StrategyLife expectancy (years)⁎QALY⁎Cost ($)⁎Incremental cost($) per QALY
Men
40 years
1. 3-month strategy20.52216.56623740Lowest cost reference strategy
2. 6-month conventional-intensity strategy20.53216.5722381611168
3. 12-month conventional-intensity strategy20.54016.5772405348534
4. 24-month conventional-intensity strategy20.55716.5872451448805
5. Unlimited-duration conventional-intensity strategy20.72916.64832615132396
6. Unlimited-duration low-intensity strategy20.57616.52733020More costly and less effective than strategy 5
60 years
1. 3-month strategy13.84210.08817172Lowest cost reference strategy
2. 6-month conventional-intensity strategy13.84910.0921728429878
3. 12-month conventional-intensity strategy13.85410.09317567225654
4. 24-month conventional-intensity strategy13.85610.08918123More costly and less effective than strategy 3
5. Unlimited-duration conventional-intensity strategy13.85810.04024314More costly and less effective than strategy 3
6. Unlimited-duration low-intensity strategy13.7889.98924640More costly and less effective than strategy 3
80 years
1. 3-month strategy6.5634.48710579Lowest cost reference strategy
2. 6-month conventional-intensity strategy6.5644.48610764More costly and less effective than strategy 1
3. 12-month conventional-intensity strategy6.5634.48211132More costly and less effective than strategy 1
4. 24-month conventional-intensity strategy6.5564.47111844More costly and less effective than strategy 1
5. Unlimited-duration conventional-intensity strategy6.5384.43215094More costly and less effective than strategy 1
6. Unlimited-duration low-intensity strategy6.5214.41915286More costly and less effective than strategy 1
Women
40 years
1. 3-month strategy21.88817.54622589Lowest cost reference strategy
2. 6-month conventional-intensity strategy21.89417.5492271135977
3. 12-month conventional-intensity strategy21.89817.55023008Higher incremental cost than strategy 4
4. 24-month conventional-intensity strategy21.90617.55123582512337
5. Unlimited-duration conventional-intensity strategy21.94117.48933846More costly and less effective than strategy 4
6. Unlimited-duration low-intensity strategy21.77217.35634252More costly and less effective than strategy 4
60 years
1. 3-month strategy15.34711.33516903Lowest cost reference strategy
2. 6-month conventional-intensity strategy15.35011.33817060155033
3. 12-month conventional-intensity strategy15.35111.33517394More costly and less effective than strategy 2
4. 24-month conventional-intensity strategy15.34311.32418065More costly and less effective than strategy 2
5. Unlimited-duration conventional-intensity strategy15.25311.19425932More costly and less effective than strategy 2
6. Unlimited-duration low-intensity strategy15.17211.13426272More costly and less effective than strategy 2
80 years
1. 3-month strategy7.4665.18210602Lowest cost reference strategy
2. 6-month conventional-intensity strategy7.4645.17910828More costly and less effective than strategy 1
3. 12-month conventional-intensity strategy7.4605.17311250More costly and less effective than strategy 1
4. 24-month conventional-intensity strategy7.4455.15512071More costly and less effective than strategy 1
5. Unlimited-duration conventional-intensity strategy7.3975.08716251More costly and less effective than strategy 1
6. Unlimited-duration low-intensity strategy7.3765.07116461More costly and less effective than strategy 1

QALY = quality-adjusted life-year.

⁎ Discounted at 3% per year.

When we estimated the incremental cost-effectiveness ratios for the 6 strategies, cost-effectiveness ratios were lower in younger patients and in men, reflecting the higher bleeding risk at older ages, and the lower risk of recurrence among women. The most effective, economically reasonable strategies were the 24-month strategy in 40-year-old men, the 6-month strategy in 40-year-old women and 60-year-old men, and the 3-month strategy in 60-year-old women and in all patients aged ≥80 years. For 40- and 60-year-olds, more effective strategies were substantially more expensive, while for 80-year-olds other strategies were less effective.

Sensitivity analyses 

In one-way sensitivity analyses, variables that increased recurrent venous thromboembolism risk or decreased major bleeding risk from conventional-intensity anticoagulation favored anticoagulation strategies of extended duration. Incremental cost-effectiveness for unlimited-duration conventional-intensity anticoagulation decreased <$50000/QALY in 40-year-old men if the annual risk of recurrence was >6%, if the case fatality for pulmonary embolism was >22%, or if the annual risk of major bleeding was <0.7%. In 40-year-old women, unlimited-duration conventional-intensity anticoagulation was never preferred but the 24-month strategy became favorable if the case fatality of pulmonary embolism was >26% or the annual recurrence rate was >7%. In women aged ≥60 years, anticoagulation beyond a duration of 6 months was never preferred. In men aged ≥60 years, anticoagulation for more than 12 months was never favored. Conversely, parameters that decreased venous thromboembolism risk or increased major bleeding risk favored the 3-month strategy, which was preferred in 40-year-old men if the case-fatality of pulmonary embolism was <8% and in 40-year-old women if the recurrence rate was <5%, the frequency of recurring pulmonary embolism was <29%, the case-fatality of pulmonary embolism was <15%, or the annual risk of major bleeding was >1.1%.

Results were also sensitive to the decrement of quality of life (ie, disutility) associated with warfarin therapy. If the disutility of taking warfarin was high (0.08), the 3-month strategy was less costly and more effective than other strategies in all patients.

Unlimited-duration conventional-intensity anticoagulation was generally preferred over the unlimited-duration low-intensity anticoagulation strategy; the low-intensity strategy was favored in 40-year-old men only if the relative bleeding risk of low versus conventional-intensity anticoagulation was 0.4, the lower limit of the tested range.

If the relative risk of recurrence for women versus men was <0.45, 3-month anticoagulation was favored among women of all age groups. When we assumed that men and women had the same risk of recurrent venous thromboembolism (ie, relative risk = 1), cost-effectiveness results were very similar for both sexes across all ages.

Results were also sensitive to the discount rate used; with no discounting, the unadjusted life expectancy was 34.664 life-years, the quality-adjusted life expectancy was 27.105 QALYs, and the lifetime costs were $44909 in 40-year-old men who received 24-month conventional-intensity anticoagulation.

When influential parameters were simultaneously varied in probabilistic sensitivity analyses and societal willingness-to-pay was either $50000 or $100000/QALY gained (Table 5), the 6-month strategy was most likely to be cost effective in 40-year-olds and in 60-year-old men while the 3-month strategy was most favored in 60-year-old women and in all patients aged ≥80 years. Overall, these results confirmed the robustness of our baseline findings in all patient groups except for 40-year-old men. This discrepancy between baseline and probabilistic sensitivity analysis results suggests that baseline values in 40-year-old men are more sensitive to multiway variation of parameter values.

Table 5. Results of probabilistic sensitivity analysis by age and sex⁎
Anticoagulation Strategy40 years60 years80 years
MenWomenMenWomenMenWomen
Likelihood of being most cost-effective, %
Willingness-To-Pay = $50000/QALY
3-month conventional-intensity strategy54125668997
6-month conventional-intensity strategy62496432113
12-month conventional-intensity strategy--71--
24-month conventional-intensity strategy17731--
Unlimited-duration conventional-intensity strategy1531---
Willingness-To-Pay = $100000/QALY
3-month conventional-intensity strategy43418588195
6-month conventional-intensity strategy49496037185
12-month conventional-intensity strategy--1231-
24-month conventional-intensity strategy21951--
Unlimited-duration conventional-intensity strategy25741--

QALY = quality-adjusted life-year.

⁎ The likelihood of being most cost-effective corresponds to the proportion of 1000 Monte Carlo iterations for which a given strategy resulted in the highest net monetary benefit at each willingness-to-pay threshold. The unlimited-duration low-intensity strategy is not shown because its likelihood of being most cost-effective was always less than 2%.

Back to Article Outline

Discussion 

Our results suggest that cost-effectiveness of anticoagulation strategies for treating patients with a first idiopathic venous thromboembolic event depend on age, sex, and a number of clinical factors. Because the bleeding risk increases with age, the patient’s age at the time of the initial venous thromboembolic event must be taken into consideration before selecting an initial anticoagulation strategy. In 40-year-old men, 24 months of conventional-intensity anticoagulation appeared to be cost effective because the persisting risk of recurrence, including the possibility of premature death due to recurrent pulmonary embolism, outweighs risks, disutility, and costs of taking warfarin. In 80-year-old patients with an annual major bleeding risk of 5.1%,26 3 months of conventional-intensity anticoagulation was the best option. In all other patients, 6- month conventional-intensity anticoagulation was preferred in most scenarios.

Our results were influenced by several assumptions. First, we assumed that women had a lower risk of recurrence than men,13, 14 resulting in the preference for shorter anticoagulation durations for women at comparable ages. Because this assumption needs confirmation in further studies, we varied the relative risk of recurrence of women versus men widely in sensitivity analysis. If women and men had equal recurrence rates, cost-effectiveness results were very similar for both sexes across all ages. Second, our results were driven by the assumption of an increased early risk of recurrence after the discontinuation of anticoagulation, regardless of anticoagulation duration.10 This effect largely counterbalances the initial benefit in prevented venous thromboembolic events associated with the 12- or 24-month strategy and results in only marginally better effectiveness than the 3- or 6-month strategy. However, this assumption is supported by studies in which patients with idiopathic venous thromboembolism who received 3 to 24 months of conventional-intensity anticoagulation had a high initial risk of recurrence after the discontinuation of anticoagulation, resulting in a similar long-term cumulative incidence of recurrent venous thromboembolism.4, 5, 6, 7, 10, 13, 17, 18 Thus, extending the duration of anticoagulation from 3–6 months to 12–24-months appears to postpone but not prevent recurrence.38, 39 Third, based on the only trial that compared low- with conventional-intensity anticoagulation in patients with venous thromboembolism, we assumed low-intensity anticoagulation to be less efficacious than conventional-intensity anticoagulation at the same risk of bleeding.8 Thus, unlimited-duration low-intensity anticoagulation was clinically and economically the least attractive option in most cases.

Clinical factors that influenced our results were the late risk of recurrence, the case-fatality of pulmonary embolism, and the major bleeding risk of conventional-intensity anticoagulation. Currently, no predictive model exists that reliably identifies patients with idiopathic venous thromboembolism who are at high-risk of recurrence after the discontinuation of anticoagulation. However, patients with pulmonary embolism as their initial thromboembolic manifestation are likely to develop pulmonary embolism again if thromboembolism recurs, with a case fatality rate of 26%.40 Thus, in younger patients with pulmonary embolism, unlimited-duration conventional-intensity anticoagulation may be potentially indicated, especially if their bleeding risk is low (eg, based on the Outpatient Bleeding Risk Index).41, 42 Finally, the disutility of taking warfarin had a major impact on our results. Strategies were more effective than 3-month anticoagulation only if warfarin treatment was well accepted, underscoring the importance of patient preference in medical decision making.

Our findings are consistent with results from a recent cost-effectiveness analysis that found 18 to 36 months of conventional-intensity anticoagulation to be cost effective in patients with idiopathic deep vein thrombosis when the recurrence rates were high.29 Earlier (cost-) effectiveness studies comparing different anticoagulation strategies (with or without testing for hypercoagulable states) in patients with venous thromboembolism are difficult to integrate with our analysis because these studies relied on less recent estimates for venous thromboembolism recurrence.43, 44, 45, 46

Our model has several limitations. First, the model does not consider relatively rare events, such as the occurrence of chronic thromboembolic pulmonary hypertension,47 because this outcome was not reported in clinical trials comparing anticoagulation strategies. Second, although long-term outcome data are not available for idiopathic venous thromboembolism, the model assumes the late risk of recurrence to be constant. A decrease in recurrence over time would favor anticoagulation strategies of shorter duration. Finally, we modeled the decrease of quality of life due to acute complications as days of utility lost because of hospitalization. Thus, we may have overestimated quality of life in cases in which acute complications were followed by a prolonged recovery period.

Our results suggest that longer initial conventional-intensity anticoagulation is cost effective in younger patients with a first idiopathic venous thromboembolic event while 3 months of anticoagulation is preferred in elderly patients. Patient age, sex, clinical factors, and patient preferences should be incorporated into medical decision making when selecting an anticoagulation strategy.

Grant support 

Dr. Aujesky was supported by the Swiss Foundation for Grants in Medicine and Biology and the Swiss Medical Association.

Back to Article Outline

Acknowledgments 

We are indebted to Clive Kearon, MB, PhD, for his helpful suggestions regarding recurrence of venous thromboembolism, and to Kevin L Kraemer, MD, MSc, for his comments on the analysis and for his review of the manuscript.

Back to Article Outline

References 

  1. Prandoni P , Lensing AW , Cogo A , et al.   The long-term clinical course of acute deep venous thrombosis . Ann Intern Med. . 1996;125:1–7
  2. Buller HR , Agnelli G , Hull RD , et al.   Antithrombotic therapy for venous thromboembolic disease (the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy) . Chest . 2004;126:401S–428S
  3. Levine MN , Raskob G , Beyth RJ , et al.   Hemorrhagic complications of anticoagulant treatment (the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy) . Chest . 2004;126:287S–310S
  4. Pinede L , Ninet J , Duhaut P , et al.   Comparison of 3 and 6 months of oral anticoagulant therapy after a first episode of proximal deep vein thrombosis or pulmonary embolism and comparison of 6 and 12 weeks of therapy after isolated calf deep vein thrombosis . Circulation . 2001;103:2453–2460
  5. Agnelli G , Prandoni P , Santamaria MG , et al.  Warfarin Optimal Duration Italian Trial Investigators   Three months versus one year of oral anticoagulant therapy for idiopathic deep venous thrombosis . N Engl J Med. . 2001;345:165–169
  6. Agnelli G , Prandoni P , Becattini C , et al.   Extended oral anticoagulant therapy after a first episode of pulmonary embolism . Ann Intern Med. . 2003;139:19–25
  7. Ridker PM , Goldhaber SZ , Danielson E , et al.   Long-term, low-intensity warfarin therapy for the prevention of recurrent venous thromboembolism . N Engl J Med. . 2003;348:1425–1434
  8. Kearon C , Ginsberg JS , Kovacs MJ , et al.   Comparison of low-intensity warfarin therapy with conventional-intensity warfarin therapy for long-term prevention of recurrent venous thromboembolism . N Engl J Med. . 2003;349:631–639
  9. British Thoracic Society guidelines for the management of suspected acute pulmonary embolism . Thorax . 2003;58:470–483
  10. Kearon C . Long-term management of patients after venous thromboembolism . Circulation . 2004;110:110–118
  11. Van der Meer FJ , Rosendaal FR , Vandenbroucke JP , Briet E . Bleeding complications in oral anticoagulant therapy. An analysis of risk factors . Arch Intern Med. . 1993;153:1557–1562
  12. Van der Meer FJ , Rosendaal FR , Vandenbroucke JP , Briet E . Assessment of a bleeding risk index in two cohorts of patients treated with oral anticoagulants . Thromb Haemost . 1996;76:12–16
  13. Kyrle PA , Minar E , Bialonczyk C , et al.   The risk of recurrent venous thromboembolism in men and women . N Engl J Med. . 2004;350:2558–2563
  14. Schulman S , Wahlander K , Lundstrom T , Clason SB , Eriksson H , Investigators TI . Secondary prevention of venous thromboembolism with the oral direct thrombin inhibitor ximelagatran . N Engl J Med. . 2003;349:1713–1721
  15. Sonnenberg FA , Beck JR . Markov models in medical decision making (a practical guide) . Med Decis Making . 1993;13:322–338
  16. Weinstein MC , Siegel JE , Gold MR , et al.   Recommendations of the Panel on Cost-effectiveness in Health and Medicine . JAMA . 1996;276:1253–1258
  17. Kearon C , Gent M , Hirsh J , et al.   A comparison of three months of anticoagulation with extended anticoagulation for a first episode of idiopathic venous thromboembolism . N Engl J Med. . 1999;340:901–907
  18. Kearon C . Rate of recurrent venous thromboembolism (VTE) after completing two years of anticoagulation for a first episode of idiopathic VTE . [Abstract] Thromb Haemost. . 2001;7(Suppl.):P794
  19. White RH , McKittrick T , Takakuwa J , et al.  National Consortium of Anticoagulation Clinics   Management and prognosis of life-threatening bleeding during warfarin therapy . Arch Intern Med. . 1996;156:1197–1201
  20. Collins TC , Petersen NJ , Menke TJ , et al.   Short-term, intermediate-term, and long-term mortality in patients hospitalized for stroke . J Clin Epidemiol. . 2003;56:81–87
  21. Rosand J , Eckman MH , Knudsen KA , et al.   The effect of warfarin and intensity of anticoagulation on outcome of intracerebral hemorrhage . Arch Intern Med. . 2004;164:880–884
  22. Decousus H , Leizorovicz A , Parent F , et al.  Prevention du Risque d’Embolie Pulmonaire par Interruption Cave Study Group   A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis . N Engl J Med. . 1998;338:409–415
  23. Arias E . United States life tables, 2000. National vital statistics reports; vol 51, no 3 . Hyattsville, MD: National Center for Health Statistics; 2002;
  24. Lindmarker P , Schulman S , Sten-Linder M , et al.   The risk of recurrent venous thromboembolism in carriers and non-carriers of the G1691A allele in the coagulation factor V gene and the G20210A allele in the prothrombin gene. DURAC Trial Study Group. Duration of Anticoagulation . Thromb Haemost. . 1999;81:684–689
  25. Cushman M , Tsai AW , White RH , et al.   Deep vein thrombosis and pulmonary embolism in two cohorts (the longitudinal investigation of thromboembolism etiology) . Am J Med. . 2004;117:19–25
  26. Pengo V , Legnani C , Noventa F , Palareti G . Oral anticoagulant therapy in patients with nonrheumatic atrial fibrillation and risk of bleeding (a multicenter inception cohort study) . Thromb Haemost. . 2001;85:418–422
  27. Beck JR , Pauker SG , Gottlieb JE , et al.   A convenient approximation of life expectancy (the “DEALE”). II. Use in medical decision-making . Am J Med. . 1982;73(6):889–897
  28. Gould MK , Dembitzer AD , Sanders GD , Garber AM . Low-molecular-weight heparins compared with unfractionated heparin for treatment of acute deep venous thrombosis. a cost-effectiveness analysis . Ann Intern Med. . 1999;130:789–799
  29. Auerbach AD , Sanders GD , Hambleton J . Cost-effectiveness of testing for hypercoagulability and effects on treatment strategies in patients with deep vein thrombosis . Am J Med. . 2004;116:816–828
  30. From the Medicare Provider Analysis and Review (MEDPAR) of Short-Stay Hospitals Web site. Available at: http://cms.hhs.gov/statistics/medpar/default.asp. Accessed September 30, 2004.
  31. Gold MR , Franks P , McCoy KI , Fryback DG . Toward consistency in cost-utility analyses (using national measures to create condition-specific values) . Med Care . 1998;36:778–792
  32. Drug Topics Red Book . Montvale, NY: Medical Economics; 2002;
  33. From the Centers for Medicare & Medicaid Services web site. Available at: http://cms.hhs.gov/providers/pufdownload/default.asp. Accessed September 30, 2004.
  34. From the U.S. Bureau of Labor Statistics Web site. Available at: http://www.bls.gov/data/. Accessed September 30, 2004.
  35. Hirth RA , Chernew ME , Miller E , et al.   Willingness to pay for a quality-adjusted life year (in search of a standard) . Med Decis Making . 2000;20(3):332–342
  36. Doubilet P , Begg CB , Weinstein MC , et al.   Probabilistic sensitivity analysis using Monte Carlo simulation (a practical approach) . Med Decis Making . 1985;5:157–177
  37. Stinnett AA , Mullahy J . Net health benefits (a new framework for the analysis of uncertainty in cost-effectiveness analysis) . Med Decis Making . 1998;18:S68–S80
  38. Van Dongen CJ , Vink R , Hutten BA , et al.   The incidence of recurrent venous thromboembolism after treatment with vitamin K antagonists in relation to time since first event (a meta-analysis) . Arch Intern Med. . 2003;163:1285–1293
  39. Frazee LA , Chomo DL . Duration of anticoagulant therapy after initial idiopathic venous thromboembolism . Ann Pharmacother. . 2003;37:1489–1496
  40. Douketis JD , Kearon C , Bates S , et al.   Risk of fatal pulmonary embolism in patients with treated venous thromboembolism . JAMA . 1998;279:458–462
  41. Beyth RJ , Quinn LM , Landefeld CS . Prospective evaluation of an index for predicting the risk of major bleeding in outpatients treated with warfarin . Am J Med. . 1998;105:91–99
  42. Wells PS , Forgie MA , Simms M , et al.   The outpatient bleeding risk index (validation of a tool for predicting bleeding rates in patients treated for deep venous thrombosis and pulmonary embolism) . Arch Intern Med. . 2003;163:917–920
  43. Sarasin FP , Bounameaux H . Decision analysis model of prolonged oral anticoagulant treatment in factor V Leiden carriers with first episode of deep vein thrombosis . BMJ . 1998;316(7125):95–99
  44. Eckman MH , Singh SK , Erban JK , Kao G . Testing for factor V Leiden in patients with pulmonary or venous thromboembolism (a cost-effectiveness analysis) . Med Decis Making . 2002;22:108–124
  45. Marchetti M , Pistorio A , Barosi G . Extended anticoagulation for prevention of recurrent venous thromboembolism in carriers of factor V Leiden—cost-effectiveness analysis . Thromb Haemost. . 2000;84(5):752–757
  46. Van den Belt AG , Hutten BA , Prins MH , Bossuy PM . Duration of oral anticoagulant treatment in patients with venous thromboembolism and a deficiency of antithrombin, protein C or protein S—a decision analysis . Thromb Haemost. . 2000;84:758–763
  47. Pengo V , Lensing AW , Prins MH , et al.   Incidence of chronic thromboembolic pulmonary hypertension after pulmonary embolism . N Engl J Med. . 2004;350:2257–2264
  48. Linkins LA , Choi PT , Douketis JD . Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism (a meta-analysis) . Ann Intern Med. . 2003;139:893–900
  49. Luk C , Wells PS , Anderson D , Kovacs MJ . Extended outpatient therapy with low molecular weight heparin for the treatment of recurrent venous thromboembolism despite warfarin therapy . Am J Med. . 2001;111:270–273
  50. Becker DM , Philbrick JT , Selby JB . Inferior vena cava filters. Indications, safety, effectiveness . Arch Intern Med. . 1992;152:1985–1994
  51. Lenert LA , Soetikno RM . Automated computer interviews to elicit utilities (potential applications in the treatment of deep venous thrombosis) . J Am Med Inform Assoc. . 1997;4:49–56
  52. Gage BF , Cardinalli AB , Albers GW , Owens DK . Cost-effectiveness of warfarin and aspirin for prophylaxis of stroke in patients with nonvalvular atrial fibrillation . JAMA . 1995;274:1839–1845
  53. Marchetti M , Pistorio A , Barone M , et al.   Low-molecular-weight heparin versus warfarin for secondary prophylaxis of venous thromboembolism (a cost-effectiveness analysis) . Am J Med. . 2001;111:130–139
  54. Teng J , Mayo NE , Latimer E , et al.   Costs and caregiver consequences of early supported discharge for stroke patients . Stroke . 2003;34:528–536
  55. Caprini JA , Botteman MF , Stephens JM , et al.   Economic burden of long-term complications of deep vein thrombosis after total hip replacement surgery in the United States . Value Health . 2003;6:59–74

PII: S0002-9343(05)00137-3

doi:10.1016/j.amjmed.2005.02.018

The American Journal of Medicine
Volume 118, Issue 6 , Pages 625-635, June 2005

Article Tools