PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Arch Intern Med. Author manuscript; available in PMC 2009 October 16.
Published in final edited form as:
PMCID: PMC2762782
NIHMSID: NIHMS97801

Outcomes after Deep Vein Thrombosis and Pulmonary Embolism in the Community: The Worcester Venous Thromboembolism Study

Abstract

Background

Despite advances in the management of deep vein thrombosis (DVT) and pulmonary embolism (PE), there are relatively few contemporary data describing and comparing outcomes in patients with these common conditions from a more generalizable community-based perspective. The purpose of this study was to measure and compare clinical characteristics and outcomes of patients with validated symptomatic PE and isolated DVT in a New England community.

Methods

The medical records of residents from the Worcester (MA) area with ICD-9 codes consistent with possible venous thromboembolism (VTE) during 1999, 2001, and 2003 were independently validated and reviewed by trained abstractors.

Results

Patients presenting with PE or isolated DVT experienced similar rates of subsequent PE, overall venous thromboembolism (VTE), and major bleeding during 3-year follow-up (5.9% vs. 5.1%, 15% vs. 17.9%, 15.6% vs. 12.4%, respectively). Mortality was significantly increased at 1-month follow-up in patients initially presenting with PE (13.0% vs. 5.4%) - this difference persisted at 3 years (35.3% vs. 29.6%). Patients whose course was complicated by major bleeding were more likely to suffer recurrent VTE or to die at 3 years than those without these complications.

Conclusions

Patients presenting with PE had similar rates of subsequent PE or recurrent VTE as patients with isolated DVT. However, rates of recurrent VTE and major bleeding following DVT and PE remain unacceptably high in the community setting. Efforts remain needed to identify patients most at risk for VTE-associated complications and development of better anticoagulation strategies conducive to long-term use in the community setting.

Introduction

Despite advances in the medical management of deep vein thrombosis (DVT) and pulmonary embolism (PE) over the last decade, there are relatively few contemporary data describing and comparing outcomes in patients with these common conditions from a more generalizable community-based perspective. The objectives of the Worcester Venous Thromboembolism (VTE) study are to provide contemporary population-based data about the clinical epidemiology of DVT and PE, its management, and associated outcomes (1). The primary objective of this report is to provide data on VTE recurrence, major bleeding, and mortality rates of patients with validated DVT and PE in the “real-world” setting.

Deep vein thrombosis and PE are generally considered to be manifestations of the same disease process. Indeed, a number of studies have suggested that a substantial proportion of patients with clinically recognized DVT, but without symptoms suggestive of PE, will have radiological evidence of PE (2-4). Nevertheless, other studies have suggested that patients with symptomatic PE have a higher risk of subsequent PE or fatal PE, and/or a worse long-term survival, than patients with isolated DVT (5-9). Unfortunately, most prior studies examining the natural history of PE and DVT were of small sample size, were conducted more than a decade ago, or relied on administrative databases for case ascertainment. We hypothesized that while patients presenting with symptomatic PE in the community setting would be older, have more comorbidities, and have higher mortality than their counterparts with isolated DVT have, recurrent rates of PE and overall VTE would not significantly differ between these two patient subsets. Therefore, a secondary objective of this investigation was to compare patient characteristics, management practices, and short and long-term outcomes in residents of the Worcester (MA) metropolitan area with independently validated DVT and PE.

Methods

Computerized printouts of all greater Worcester residents with healthcare system encounters in which any of 34 ICD-9 diagnosis codes possibly consistent with VTE had been listed in 1999, 2001, or 2003 were obtained from each of the 12 hospitals serving the Worcester metropolitan area (2000 census = 478,000) (see Appendix A) (1). These data queries were not limited to hospital discharge diagnoses but also encompassed all outpatient, emergency department, radiology, and laboratory encounters.

Trained data abstractors reviewed the medical records of all identified patients meeting our geographic inclusion criteria. Validation and characterization of each case of VTE as being definite, probable, possible, or absent was performed by trained abstractors using pre-specified criteria (see Appendix B) (1, 10). For purposes of this study, all possible, probable, and definite cases of VTE were utilized. Potential cases of VTE recurrence were classified using similar criteria as that employed for incident cases and were reviewed and validated by the principal investigator (FAS). Definite or probable recurrence of VTE required the new occurrence of thrombosis in a previously uninvolved venous or pulmonary segment.

Data collection

Information was collected about patients' demographic and clinical characteristics, diagnostic test results and management practices, and short and long-term outcomes through the review of hospital and ambulatory medical records. Medical history variables defined as “recent” were those occurring in the 3 months prior to the diagnosis of VTE. Major bleeding was defined as any bleeding episode requiring transfusion, resulting in hospitalization, stroke, or myocardial infarction, or causing death.

Short and long-term rates of an initial recurrence of VTE and first episodes of major bleeding were determined through the complete review of subsequent medical records at the same hospital site as the index event as well as through the screening of medical records from the other participating hospital sites. Mortality data were obtained through hospital record reviews and review of death certificates at the Massachusetts Division of Vital Statistics. Follow-up data were available on our 3 patient cohorts for a minimum of 1 year and a maximum of 3 years.

Data analysis

Differences in the distribution of demographic and clinical characteristics, as well as short and long-term outcomes, between patients with PE versus isolated DVT were examined using chi-square tests of statistical significance for categorical variables and t-tests for continuous variables. Cumulative incidence rates of VTE recurrence, PE recurrence, and major bleeding (censoring subjects at the time of death) and all-cause mortality were estimated using the life-table method.

Cox regression analyses were constructed in order to evaluate whether type of VTE on presentation (PE with/without DVT or isolated DVT) was associated with specific outcomes, including occurrence of new PE, recurrent VTE, major bleeding, and all-cause mortality. All variables listed in Table 1, with the exception of body mass index (BMI) were considered as potential covariates. Since BMI data were missing on approximately a third of patients, this variable was not included in our final regression analyses. Repeat analyses performed including BMI did not significantly alter our study findings (data not shown). The occurrence of recurrent VTE and major bleeding were included in these analyses as time-dependent variables in order to assess the impact of these complications on our principal study outcomes. Candidate variables possibly associated with the outcomes of interest (p<0.25 after univariate analysis) were included in each multivariate regression model. Variables with a p-value >0.05 were eliminated in a stepwise fashion so that only variables with a statistically significant association with the outcome of interest were included in the final regression model(s).

Table 1
Demographic and Clinical Characteristics of Patients According to Venous Thromboembolism Site

Results

The study sample consisted of 1,691 men and women from the Worcester metropolitan area with a validated episode of acute VTE (both hospitalized and non-hospitalized cases) during the 3 study years of 1999, 2001, and 2003. Of these, 549 patients were diagnosed with PE (with or without DVT) (32%), while the remainder (n = 1142) were diagnosed with isolated DVT. The mean age of the study sample was 65 years, 54% were women, and 91% were Caucasian.

Characteristics of Patients with PE versus Isolated DVT

Patients with PE were more likely to be obese than patients with isolated DVT (Table 1). Patients with PE were more likely to have been recently (< 3 months) hospitalized, undergone surgery, admitted to an intensive care unit or intubated, have had an infection, or have experienced congestive heart failure than patients with isolated DVT. They were also more likely to have developed their episode of VTE during hospitalization for a non-VTE related condition (Table 1).

Treatment Practices

Patients with PE were more likely to be initially treated with a low-molecular weight heparin alone or in combination with unfractionated heparin, to have an inferior vena cava (IVC) filter placed during hospitalization, or to be started on warfarin therapy during the initial hospital encounter. Patients with PE were less likely to already have an IVC filter in place at the time of presentation (Table 2).

Table 2
Acute Treatment Strategies in Patients According to Site of Venous Thromboembolism

Outcomes

Among the 549 patients presenting with PE, 31 (5.7%) had recurrent PE, 75 (13.7%) had recurrent VTE, 82 (14.9%) experienced a major bleeding episode, and 226 (41.7%) died over our follow-up period. Among 1142 patients presenting with isolated DVT, 64 (5.6%) had subsequent PE, 217 (19%) had recurrent VTE, 146 (12.8%) experienced a major bleeding episode, and 411 (36.0%) died over our follow-up period.

Recurrent Thromboembolic Events

Cumulative incidence rates of recurrent VTE (DVT or PE) were similar between patients presenting with PE versus isolated DVT at 30 days (3.8% vs. 4.8%), 1 year (10.7% vs. 12.0%), and 3 years (15.0 % vs. 17.9%) (Table 3). After Cox regression analysis, presentation with PE (as compared to isolated DVT) was not significantly associated with an increased risk of recurrent VTE at 30 days, 1 year, or 3 years (Table 4). However, occurrence of a bleeding complication after initial VTE (HR 2.94, 95% CI 2.00, 4.33) was associated with a significantly increased risk of a recurrent episode of VTE at 3-year follow-up.

Table 3
Life-table Estimates of Cumulative Incidence Rates of Selected Outcomes According to Site of Venous Thromboembolism
Table 4
The Association between Presentation with Initial Pulmonary Embolism (versus Isolated Deep Vein Thrombosis) on Subsequent Venous Thromboembolism Associated Complications and Mortality*

Patients presenting with PE initially had essentially similar cumulative incidence rates of new PE at 30 days (1.1% vs. 1.5%), 1 year (3.6% vs. 3.6%), and 3 years (5.9% vs. 5.1%) compared to patients with isolated DVT (Table 3). After Cox regression analysis, presentation with PE (as compared to isolated DVT) was not significantly associated with increased risk of subsequent PE at 30 days, 1 year, or 3 years. Occurrence of major bleeding during follow-up was not associated with an increased risk of subsequent PE.

Major Bleeding

Rates of major bleeding after 30 days (9.4 % vs. 6.5%), 1 year (11.6% vs. 10.3%), and 3 years (15.6% vs. 12.4%) were similar between patients presenting with PE versus isolated DVT. After Cox regression analysis, presentation with PE was not significantly associated with major bleeding at 30-day, 1 year, or 3 years follow-up. Patients experiencing a recurrent VTE event were more likely to experience bleeding at the time of our 3-year follow-up (HR 2.80, 95% CI 1.63, 4.83).

Mortality

The observed all-cause mortality was higher at 30 days (13.0% vs. 5.4%), 1 year 26.0% vs. 20.3%), and 3 years (35.3% vs. 29.6%) for patients with PE compared to those with isolated DVT (table 3). Most of this increase in mortality occurred in the first 30 days. After Cox analysis, presentation with PE was significantly associated with increased 30-day (HR 3.23, 95% C.I. 2.18, 4.79), one-year (HR 1.52, 95% CI 1.20, 1.93), and 3-year mortality (HR 1.43, 95% CI 1.16, 1.76). Patients experiencing a recurrent episode of VTE were not at significantly increased risk of 3-year mortality (HR 1.18, 95% CI 0.93, 1.50). Occurrence of a major bleeding complication was associated with an increase in 3-year death rates (HR 1.36, 95% C.I. 1.07, 1.73).

Discussion

The results of our population-based study in residents of an entire New England metropolitan area provide contemporary insights into the clinical profile and outcomes of patients presenting with symptomatic PE and isolated DVT.

Recurrent Venous Thromboembolism

Rates of recurrent VTE in our community-based study did not differ between our two comparison cohorts but were relatively high - approximately 4% and 11% of all patients had developed a recurrent episode of VTE at 1 month and 1 year, respectively. Not surprisingly, these rates are higher than those reported in recent clinical trials of anticoagulant treatment in patients with VTE in whom study inclusion criteria are narrowly defined and therapy is carefully specified and monitored (11, 12). Disappointingly, the rates of recurrent VTE observed in our study do not differ significantly from those reported by Prandoni and colleagues in their landmark study of the natural history of VTE more than a decade ago (13). In their study of 355 consecutive patients referred to a tertiary care center with a first symptomatic VTE, approximately 10% of their study sample had experienced a recurrent episode of VTE at 1 year. Similarly, in a population-based cohort study of 1,719 residents of the Rochester, MN, area with a first episode of VTE between 1966 and 1990, the cumulative incidence of recurrent VTE at 30-days and 1-year was 5.2% and 12.9%, respectively (14). Although one must be cautious in comparing outcome data from different observational studies, the present findings suggest that VTE recurrence rates in the community have not improved appreciably over the last decade.

Recurrent PE

A number of studies have suggested that patients presenting with PE are at increased risk for recurrent PE as compared to patients with isolated DVT (7, 9, 15). In an analysis of the California Patient Data Set, which included information from more than 70,000 adult patients hospitalized with VTE between 1991-1996, initial presentation with PE was the strongest predictor of recurrent VTE manifesting as re-hospitalization for PE (7). In a prospective study of 436 patients with unprovoked VTE from 1992 to 2002, patients presenting with PE had an approximately 4-fold increased risk of recurrent PE compared with those presenting with isolated DVT over a 30-month follow-up (9). These findings have led some to suggest that DVT and PE are distinct clinical entities with different natural histories.

These findings are at odds with our observations that suggest that patients with PE and isolated DVT may have a similar VTE-specific prognosis. Our findings are consistent with prior studies that have suggested that up to half of patients with DVT who are asymptomatic for PE will have radiological evidence of PE (2-4). In the community setting, patients with DVT who are asymptomatic for PE rarely undergo further diagnostic testing for PE. Indeed, only 6.8% of patients with isolated lower extremity symptoms in our study also underwent diagnostic testing for possible PE (data not shown). As such, there may have been a potential for misclassification bias in our study in that patients with symptomatic DVT may also have had undiagnosed PE. Although we cannot account for this bias, this misclassification is what occurs in the real world clinical setting.

We also acknowledge that physicians in the community may be anticoagulating patients with PE for longer durations than patients with isolated DVT. This may have decreased the recurrence rates of VTE, and possibly of PE, in patients with an initial episode of PE. This, in turn, would mask any difference in recurrence rates in these two patient cohorts.

Finally, we cannot exclude the possibility that some of the excess mortality observed in patients presenting with PE in our study were secondary to fatal (undiagnosed) PE recurrences. In a literature review of 25 studies of VTE treatment, patients presenting with PE had higher rates of fatal PE during anticoagulant therapy than those with isolated DVT (1.5% vs. 0.4%) (8). Interestingly, there was no significant difference in the occurrence rates of fatal PE between these cohorts following anticoagulant therapy.

Further study of the subsequent risk of clinically significant PE in patients presenting with symptomatic PE versus isolated DVT remains warranted. At present, the American College of Chest Physician treatment guidelines for VTE still consider DVT and PE as manifestations of the same disease process and recommend similar durations of anticoagulation for both (16).

Major Bleeding Episodes

As opposed to data derived from recent randomized clinical trials of VTE treatment in which the occurrence of major bleeding episodes has been shown to be ≤ 4% at 3 to 6 months (11, 12), the cumulative incidence of major bleeding in our population at 1 month was approximately twice this rate. In addition, the occurrence of major bleeding following VTE was associated with a 3-fold increased risk of recurrent VTE and a 30% increased risk of mortality at 3 years.

These data suggest a parallel to research currently being conducted in patients with acute coronary syndromes (ACS). There is growing recognition that the occurrence of major bleeding following treatment of patients with ACS is associated with significantly worse outcomes (17, 18). Accordingly, there has been a considerable amount of contemporary research aimed at improving our ability to identify patients with an ACS who are at increased risk for major bleeding, clarifying the impact of selected antithrombotic therapies on serious bleeding, and minimizing the impact of bleeding on mortality when it does occur (19, 20). Further study of patient, clinical, and treatment characteristics associated with VTE-related bleeding are needed if we are to improve the short, as well as long-term, outcomes associated with this prevalent condition.

Mortality

Patients presenting with PE had significantly higher death rates from all causes than those presenting with isolated DVT, particularly during the first 30 days of follow-up. While much of this increase in mortality may be due to the increased prevalence of severe concomitant illnesses in patients with PE, some of this increased risk may be attributable to fatal (but undiagnosed) recurrent PE. The short-term mortality associated with clinically recognized PE in our study is similar to that observed in the International Cooperative Pulmonary Embolism Registry in which 13% of 2,048 patients with PE diagnosed in the mid-1990's (and not treated with thrombolysis) had died at 3 months (21). These observations suggest that there remains substantial room for improvement in the acute detection and management of PE in the community setting.

Approximately 1 in 4 patients with clinically recognized VTE had died within the first year after their initial diagnosis in our population-based study. This represents only a slight improvement in total VTE-associated mortality compared to the original Worcester VTE study conducted in the 1980s, in which nearly one-third of patients with validated VTE had died at 1 year (22). Much of this substantial mortality likely stems from the advanced age and serious comorbidities typically seen in patients with VTE in the community setting. These data highlight the poor prognosis of this group as a whole and reminds us of the difficulties inherent in treating this condition in the real-world setting.

Study limitations

Similar to the design and conduct of any observational study, the present investigation has several limitations. Although we conducted a broad screening for all possible cases of VTE in the greater Worcester population, we cannot claim complete case ascertainment of index VTE events, episodes of VTE recurrence, or episodes of major bleeding. Most notably, we will not have captured events occurring in greater Worcester residents who sought care at hospitals outside of this metropolitan area. We do not collect data on anticoagulation management practices occurring after hospital discharge. As such, we cannot comment on the impact of use (or non-use) of anticoagulation on our observed study outcomes. Finally, due to the low autopsy rates during the period under study, we are unable to estimate the rates of fatal PE. Therefore, we can only comment on complication rates and mortality associated with clinically recognized VTE.

Conclusions

Our study provides insights into the clinical profile of patients with PE or isolated DVT as well as their short and long-term outcomes within this well-defined, large, and representative New England community. Patients who present with PE had similar rates of VTE recurrence or subsequent PE as patients with DVT. Nevertheless, short-term case-fatality rates associated with PE, as well as rates of recurrent VTE and major bleeding following an initial episode of DVT or PE, remain unacceptably high. Although recent randomized clinical trials have resulted in important advances in the management of VTE, our data suggest that, at least in the community setting, there remains considerable room for improvement in the treatment of these high-risk patients.

Acknowledgments

This study was made possible by the cooperation of administrators, physicians, and medical records personnel in 12 central Massachusetts hospitals.

Supported by a grant from the National Heart, Lung, and Blood Institute (R01-HL70283)

Funding/Support: This study was supported by a grant from the National Heart, Lung, and Blood Institute (R01-HL70283). Dr. Spencer also has a Career Investigator Award from the Heart and Stroke Foundation of Canada.

Footnotes

Author contributions: Dr. Frederick Spencer had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Spencer, Gore, Goldberg

Acquisition of data: Spencer, Emery

Analysis and interpretation of data: Spencer, Lessard, Goldberg

Drafting of the manuscript: Spencer, Gore, Douketis, Goldberg

Critical revision of the manuscript for important intellectual content: Spencer, Gore, Douketis, Emery, Lessard, Goldberg

Statistical expertise: Spencer, Lessard, Goldberg

Administrative, technical, or material support: Spencer, Goldberg

Obtained funding: Spencer, Goldberg

Study supervision: Spencer, Emery, Goldberg

.

Financial Disclosures: None reported

Role of the Sponsor: The sponsors had no role in the design and conduct of this study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript.

References

1. Spencer FA, Emery C, Lessard D, Anderson F, Emani S, Aragam J, Becker RC, Goldberg RJ. The Worcester Venous Thromboembolism study. A population-based study of the clinical epidemiology of venous thromboembolism. J Gen Intern Med. 2006;21:722–27. [PMC free article] [PubMed]
2. Moser KM, Fedullo PF, LittleJohn JK, et al. Frequent asymptomatic pulmonary embolism in patients with deep venous thrombosis. JAMA. 1994;27:223–225. [PubMed]
3. Nielsen HK, Husted SE, Krusell LR, et al. Silent pulmonary embolism in patients with deep venous thrombosis. Incidence and fate in a randomized, controlled trial of anticoagulation versus no anticoagulation. J Intern Med. 1994;235:457–461. [PubMed]
4. Huisman MV, Buller HR, ten Cate J, et al. Unexpected high prevalence of silent pulmonary embolism in patients with deep venous thrombosis. Chest. 1989;95:498–502. [PubMed]
5. Kucher N, Tapson VF, Goldhaber SZ, DVT FREE Steering Committee Risk factors associated with symptomatic pulmonary embolism in a large cohort of deep vein thrombosis patients. Thromb Haemost. 2005;93:494–98. [PubMed]
6. Monreal M, Barba R, Tolosa C, Tiberio G, Todoli JA, Sanperiz AL, RIETE Investigators Deep vein thrombosis and pulmonary embolism. The same disease? Pathophysiol Haemost Thromb. 2006;35:133–35. [PubMed]
7. Murin S, Romano PS, White RH. Comparison of outcomes after hospitalization for deep venous thrombosis or pulmonary embolism. Thromb Haemost. 2002;88:407–414. [PubMed]
8. Douketis JD, Kearon C, Bates S, Duku EK, Ginsberg JS. Risk of fatal pulmonary embolism in patients with treated venous thromboembolism. JAMA. 1998;279:458–62. [PubMed]
9. Eichinger S, Weltermann A, Minar E, Stain M, Schonauer V, Schneider B, Kyrle PA. Symptomatic pulmonary embolism and the risk of recurrent venous thromboembolism. Arch Intern Med. 2004;164:92–96. [PubMed]
10. Silverstein MD, Heit JA, Mohr DN, Petterson TM, O'Fallon WM, Melton LJ. Trends in the incidence of deep vein thrombosis and pulmonary embolism. A 25-year population-based study. Arch Intern Med. 1998;158:585–593. [PubMed]
11. Fiessinger JN, Huisman MV, Davidson BL, Bounameaux H, Francis CW, Eriksson H, et al. Ximelagatran vs. low-molecular-weight heparin and warfarin for the treatment of deep vein thrombosis. A randomized trial. JAMA. 2005;293:681–689. [PubMed]
12. Buller HR, Davidson BL, Decousus H, Gallus A, Gent M, Piovella F, et al. Matisse Investigators Fondaparinux or enoxaparin for the initial treatment of symptomatic deep vein thrombosis. A randomized trial. Ann Intern Med. 2004;140:867–873. [PubMed]
13. Prandoni P, Lensing AW, Cogo A, Cuppini S, Villalta S, Carta M. The long-term clinical course of acute deep venous thrombosis. Ann Intern Med. 1996;125:1–7. [PubMed]
14. Heit JA, Mohr DN, Silverstein MD, Petterson TM, O'Fallon WM, Melton JM. Predictors of recurrence after deep vein thrombosis and pulmonary embolism. A population-based cohort study. Arch Intern Med. 2000;160:761–768. [PubMed]
15. The Columbus Investigators. Low-molecular-weight heparin in the treatment of patients with venous thromboembolism. N Engl J Med. 1997;337:657–662. [PubMed]
16. Geerts WH, Pineo GF, Heit JA, Bergqvist D, Lassen MR, Colwell CW, et al. Prevention of venous thromboembolism: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest. 2004;126:338S–400S. [PubMed]
17. Eikelboom JW, Mehta SR, Anand SS, Xie C, Fox KA, Yusuf S. Adverse impact of bleeding on prognosis in patients with acute coronary syndromes. Circ. 2006;114:774–782. [PubMed]
18. Rao SV, O'Grady K, Pieper KS, Granger CB, Newby LK, Van de Werf F, et al. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol. 2005;96:1200–1206. [PubMed]
19. Moscucci M, Fox KAA, Cannon CP, Klein W, Lopez-Sendon J, Montalescot G, et al. Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE) Eur Heart J. 2003;24:1815–1823. [PubMed]
20. Alexander KP, Chen AY, Roe MT, Newby LK, Gibson CM, Allen-LaPointe NM, et al. CRUSADE Investigators Excess dosing of antiplatelet and antithrombin agents in the treatment of non-ST-segment elevation acute coronary syndromes. JAMA. 2005;294:3108–3116. [PubMed]
21. Goldhaber SZ, Visani L, De Rosa M. ICOPER. Acute pulmonary embolism: clinical outcomes in the International Cooperative Pulmonary Embolism Registry (ICOPER) Lancet. 1999;353:1386–1389. [PubMed]
22. Anderson FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester DVT Study. Arch Intern Med. 1991;151:933–938. [PubMed]