Jacobs Journal of Hematology

Anticoagulant Treatment after VTE in the Netherlands: A Retrospective Cohort Study

*Irene Bezemer
Department Of Hematology, Pharmo Institute For Drug Outcomes Research, Netherlands

*Corresponding Author:
Irene Bezemer
Department Of Hematology, Pharmo Institute For Drug Outcomes Research, Netherlands

Published on: 2018-11-09


Background: The aim of the study was to describe initial anticoagulant treatment after venous thromboembolism (VTE) and relate this to the underlying risk factors, guidelines and recurrence rates.
Methods: From the PHARMO Database Network, patients with deep venous thrombosis (DVT) or pulmonary embolism (PE) in 2007-2011 were identified for whom GP records and out-patient pharmacy dispensing data was available. Cancer, other risk factors and duration of low-molecular-weight heparin (LMWH) and/or vitamin K antagonists (VKA) within 90 days of diagnosis and recurrence of VTE were assessed.
Results: The study cohort included 1,581 VTE patients: 1,053 with DVT and 528 with PE. For 70-86% of the VTE patients, anticoagulant dispensing was observed within 90 days. The median duration of the period covered by anticoagulant dispensings among patients for whom both LMWH and VKA dispensings were observed was 3.5 months among patients with provoked VTE and 3.7 months among patients with unprovoked VTE. In these cohorts the observed median duration of (initial) LMWH treatment was 0.4 months (about 12 days) for patients who also received VKA. Recurrent VTE occurred in general after discontinuation of anticoagulant treatment. Longer treatment durations were observed among patients without recurrence.
Conclusions: Treatment after VTE as captured in observational healthcare data generally follows the Dutch guidelines. However, many patients received LMWH dispensing covering periods longer than the recommended 5-10 days. Furthermore, among patients with a VTE recurrence shorter duration of anticoagulant treatment was observed compared to patients without a recurrence.


Anticoagulant Treatment; Low-Molecular-Weight Heparin; Venous Thromboembolism; Vitamin K Antagonist


Deep vein thrombosis (DVT) and its potentially fatal acute complication, pulmonary embolism (PE), are both manifestations of the same disease process of venous thromboembolism (VTE). Each year, approximately 1-3 per 1000 individuals present with acute symptomatic VTE. However, incidence estimates are highly influenced by study design, data collection and differences in healthcare settings between countries. Pathophysiological and epidemiological findings have identified several factors which contribute to an increased risk for VTE. A major risk factor for VTE is cancer. Furthermore, genetic risk factors as well as surgical procedures, hospital admission, trauma, fracture, oral contraceptives and pregnancy can provoke VTE events.

In the Netherlands, patients with VTE-like complaints in general first contact their general practitioner (GP) who will coordinate diagnosis and treatment, or visit the emergency room. As the GP has a gatekeeper role, all information about diagnosis and treatment is reported back to the GP office. Hospital admission is in general not necessary for DVT [10]. Adequate initial anticoagulant therapy of DVT can prevent subsequent PE and death. In the Netherlands, International Normalized Ratio (INR) monitoring and dosing of vitamin K antagonists (VKA) takes place at anticoagulation clinics outside the hospitals. The Dutch 2008 primary care guidelines for VTE treatment advise subcutaneous low-molecular-weight heparin (LMWH) for initial treatment (at least five days until INR is stable) and VKA treatment for the longer term in order to prevent further development of thrombosis (initial treatment) as well as recurrence (longer term treatment) [10, 11]. VKA treatment is recommended for three months for DVT patients with a clear transient trigger, and six months for unprovoked thrombosis [12]. For patients with cancer treatment with LMWH for at least six months is recommended, or as long as the cancer is active. However, clinical practice does not always follow clinical guidelines. The objective of this study was to describe initial anticoagulant treatment after VTE and relate this to the underlying risk factors, guidelines and recurrence rates.


Data source

Data were obtained from the PHARMO Database Network. This network links drug dispensing records to hospital discharge records and other data sources such as general practitioner (GP) data [13, 14]. GP data, out-patient pharmacy as well as hospitalization records are available for approximately 500,000 residents in defined areas of the Netherlands. 

Study design

In this historical cohort study, all patients with a first DVT or PE (no previous DVT or PE diagnosis in the database) between January 1, 2007 and December 31, 2011 were selected from the PHARMO GP Database. Patients who had both DVT and PE were classified as PE. The date of the first VTE diagnosis was defined as the index date. Patients were followed until end of registration in the database or end of the study period (31 December 2011), whichever came first. The actual study cohort comprised VTE patients for whom also out-patient pharmacy data was available in order to optimize coverage of anticoagulant drug dispensing. VTE was classified by provoking factor: ‘cancer’, ‘provoked’ or ‘unprovoked’. Cancer was defined as a GP-recorded or primary hospital discharge diagnosis of cancer (excluding non-melanoma skin cancer), or radiation or chemotherapy between three months before and three months after the first VTE. Patients were assigned to the provoked group if the patient was hospitalised, underwent a surgical procedure, had a trauma or fracture in the three months before VTE, used oral or injectable oestrogens in the six months before VTE, was pregnant in the 9 months before VTE or delivered between three months before and nine months after VTE, or had non-active cancer (diagnosis more than three months before VTE). The remaining VTE patients were regarded as unprovoked VTE.

Outcome definitions

Initial anticoagulant treatment was defined as an out-patient dispensing of LMWH (ATC code B01AB) or VKA (ATC code B01AA) within 90 days of the index date. New oral anticoagulants were not yet licensed for VTE treatment in the Netherlands and were also not observed in the study cohort. Because in-patient treatment is not captured in the PHARMO Out-patient Pharmacy Database, treatment was tracked from after discharge for patients hospitalized at the time of the first VTE diagnosis. Episodes of uninterrupted use were constructed for subsequent dispensing of LMWHs and for VKAs. The duration of each LMWH dispensing was estimated from the number of units prescribed per day and the number of units dispensed. The duration of each VKA dispensing could not be calculated from the number of units prescribed per day because dosing took place at the anticoagulation clinic which is not recorded in the Out-patient Pharmacy Database. Therefore, the duration was estimated from the number of days between dispensing weighted by the number of units dispensed. The minimum duration of one VKA dispensing was set at 7 days. Use of LMWH or VKA was considered uninterrupted if the number of days until a subsequent dispensing did not exceed 2 times the estimated duration of the dispensing. Persistence was defined as the number of days of uninterrupted use of LMWH, VKA or LMWH/VKA (either), following first use on or after cohort entry. Note that ‘use’ actually refers to the period covered by anticoagulant dispensing.

Recurrent VTE was defined as a diagnosis of DVT or PE after the first VTE diagnosis. Recurrent events were determined for a subcohort of patients with at least one year database follow-up after the first VTE. This period bridges most of the treatment durations, at least for patients who do not need long-term treatment and so recurrent VTE could be studied in relation to duration of treatment. Only the first recurrent event of a patient was assessed. Recurrent events were identified in the GP records as well as in hospital discharge records. To limit the risk of reporting repeated entries of the first VTE, only events occurring at least 30 days after the first event were included, or after 7 days when the new entry was a different VTE type (PE after DVT or vice versa). The number of recurrent events was stratified by group of DVT/PE and cancer/provoked/unprovoked cohorts.

Data analysis

Descriptive characteristics are presented for DVT and PE patients, stratified by cohort (cancer, provoked, unprovoked). Continuous variables are presented as mean with standard deviation (SD) and/or median with interquartile range (IQR). Categorical variables are presented as counts and percentages. The incidence of VTE in the GP setting was calculated as the number of patients with a first diagnosis during the study period (numerator) per 1,000 person-years. The denominator was the number of accumulated person-years ‘at risk’ in the source population during the study period, i.e. from start of the study period until occurrence of VTE or end of the study period. The following treatment characteristics were determined: number (%) of patients starting anticoagulant treatment within or after 90 days of the index date (type of anticoagulant treatment within 90 days (LMWH and VKA, LMWH only or VKA only)) and duration of treatment, and number of patients with no anticoagulant treatment during follow-up. Kaplan-Meier curves were used to represent duration of anticoagulant treatment of DVT and PE patients stratified by cohort. For recurrent VTE, the number and percentage of patients with recurrence, the timing of recurrence relative to anticoagulant treatment (during or after) and the median (IQR) duration of initial treatment are presented.All data were analysed using SAS programs organized within SAS Enterprise Guide version 4.3 (SAS Institute Inc., Cary, NC, USA) and conducted under Windows using SAS version 9.2.


General characteristics and incidence of VTE

The study cohort included 1,581 VTE patients: 1,053 (67%) with DVT and 528 (33%) with PE. The incidence of VTE in the source population (linked GP and pharmacy data) was 1.10 per 1,000 person-years (95% CI 1.05-1.16); 0.73 per 1,000 person-years (95% CI 0.69-0.78) for DVT and 0.37 per 1,000 person-years (95% CI 0.34-0.40) for PE. A cancer diagnosis was identified for 109 DVT patients (10%), 403 patients (38%) had a provoking risk factor and for 621 (59%) no risk factor was identified (‘unprovoked’). Among PE patients, 87 (16%) had a cancer diagnosis, 225 (43%) a provoking risk factor and 283 (54%) had unprovoked PE. Proportions of patients with provoking factors were similar for DVT and PE patients; non-active cancer (19 and 22%), recently hospitalized (12 and 13%), use of oestrogens (8 and 10%), recent surgery (7 and 8%), pregnancy/delivery (2%) and recent trauma (2 and 1%). Note that the cancer and provoked cohorts overlap as patients could have cancer as well as another risk factor. The mean age (±SD) was 60 (±18) years for provoked DVT or PE and 61 (±16) years for unprovoked DVT or PE (Table 1). Patients with cancer were 68 (±13) years at DVT or PE diagnosis. In the provoked cohort most patients were female (64%) as pregnancy and use of oestrogens was part of the cohort definition.

Anticoagulant treatment of VTE

Among 109 DVT patients with cancer, 92 (84%) had a dispensing record of LMWH or VKA within 90 days of DVT diagnosis (Table 2). Corresponding figures were 332 of 403 (82%) for provoked DVT and 434 of 621 (70%) for unprovoked DVT, and results were similar for PE patients: 75 of 87 (86%) in the cancer cohort, 190 of 225 (84%) in the provoked PE cohort and 217 of 283 (77%) in the unprovoked PE cohort. Not all patients had records of both VKA and LMWH dispensings. More VKA only and relatively shorter treatment periods were observed among unprovoked VTE compared to provoked VTE (see below) and among patients with PE compared to DVT, which indicates missing information on initial in-hospital treatment. In the overall cancer cohort, both VKA and LMWH dispensing were observed for 30% of patients, LMWH but no VKA dispensing were observed for 36% and only VKA dispensing were observed for 20%. In the overall provoked cohort 42% had records of LMWH as well as VKA dispensing, 17% of LMWH only and 24% of VKA only. In the overall unprovoked cohort these respective percentages were 39%, 3% and 30%.

Persistence with anticoagulant treatment was in general higher among PE patients than among DVT patients and highest in the unprovoked cohorts. The median duration of anticoagulant treatment in the group with both LMWH and VKA dispensing within 90 days of the index date was 1.2 months among DVT patients with cancer, 3.8 months among patients with provoked DVT and 3.6 months among patients with unprovoked DVT (Table 2). Corresponding figures were 2.1, 2.1.

                                                                               Table 1. General characteristics of patients diagnosed with DVT or PE, stratified by cohort.

                                                                               Table 2. Anticoagulant treatment of first DVT or PE, stratified by cohort.

and 5.2 months for PE patients. The median duration of LMWH treatment was 0.4 months (about 12 days) for these patients. Persistence was similar for patients with only VKA dispensings within 90 days. The median duration of LMWH treatment in the group without VKA dispensing was 2.4 months among DVT patients with cancer, 1.8 months in the provoked cohort and 1.2 months in the unprovoked cohort. Among PE patients, median durations of LMWH treatment were 5.5, 3.5 and 1.0 months in the respective cohorts.

                                                                Table 3.VTE recurrence after first DVT or PE in relation to anticoagulant treatment, stratified by cohort, among patients with at least one year follow-up


This study describes initial anticoagulant treatment after a first DVT or PE in the Netherlands during 2007-2011 related to the underlying risk factors, guidelines and recurrence rates. In this population based cohort study the incidence of DVT and PE in primary care corresponded well with previous studies in the Netherlands [15, 16]. VTE patients were classified in cohorts of underlying risk factors. Proportions of patients with cancer were 10% for DVT and 16% among patients with PE. Some previously reported cancer proportions were higher than in our study [5,17]. One of the reasons might be limited sensitivity in our database due to underreporting of cancer diagnoses in hospital and GP registrations. Furthermore, studies differed in study definitions and study design. In other previous reports comparable proportions of patients with VTE and cancer and with provoked VTE were found [18, 19]. For up to one third of the patients and mainly among unprovoked VTE patients, no dispensing record of either LMWH or VKA was observed within 90 days of diagnosis. Some of the patients who did not have a record of anticoagulant treatment may actually have missed treatment, but in a healthcare database study there are several reasons why treatment may go undetected. First, the date of diagnosis, which may in our database also be the date when the GP recorded the VTE diagnosis, might have been misclassified and the time window of observation therefore not correct. Second, patients may have obtained their drugs from another source than the out-patient pharmacy, such as a hospital pharmacy. Among PE patients, for example, a larger proportion had VKA dispensings but no LMWH dispensings, as compared to DVT patients. These patients likely received LMWH during hospital admission in the first week after diagnosis and in many cases will have completed LMWH bridging to VKA treatment prior to discharge. Third, some patients with superficial or distal VT, which may be untreated, were coded (misclassified) as DVT.

The guidelines at the time of the study advised treatment for three months for DVT patients with a clear trigger, and six months for unprovoked thrombosis[12]. In our study, we found that patients with provoked DVT or PE were treated for about 2 to 4 months which is according to the guidelines. The treatment duration of PE was likely underestimated as information on initial in-hospital treatment was not collected. In addition, the more recent guidelines [20] make the distinction between 3 months and long-term treatment instead of 6 months. For unprovoked DVT and PE, the observed persistence with anticoagulant treatment when including VKA treatment was slightly shorter than recommended: about 3.5 to 5 months. However, as for PE overall, it should be noted that because information on initial in-hospital treatment was not collected, actual treatment durations, in particular LMWH treatment, were likely longer. For patients with VTE without cancer, LMWH is advised only for initial treatment until INR is stable. In the provoked and unprovoked VTE cohorts the observed median duration of LMWH treatment was 0.4 months (about 12 days) for patients who also received VKA, which means that one of every two patients was treated with LMWH for more than two weeks outside of the hospital. The duration of LMWH treatment among patients who were hospitalized after diagnosis may be underestimated as mentioned before. On the other hand, out-patient treatment duration may have been overestimated because not all medication that was dispensed was actually used, i.e. ‘use’ in this study actually refers to the period covered by anticoagulant dispensings.

Among patients for whom only LMWH dispensing was observed in the first 90 days, the median treatment duration was 1.1 months for unprovoked VTE and 2.3 months for provoked VTE. Why we did not observe VKA treatment is unclear. Part of this will be explained by the overlap between the cancer cohort and provoked cohort; 55% of provoked patients were also in the cancer cohort and treated accordingly. Treatment with LMWH for at least six months is recommended for cancer patients. In all subgroups, however the median duration of LMWH treatment was shorter (0.5 to 5.5 months), which might be due to the short follow-up among cancer patients. Recurrent VTE occurred in general after discontinuation of anticoagulant treatment. When comparing patients with recurrence after treatment and patients without recurrence, longer treatment durations were observed among patients without recurrence in all cohorts. This could be an indication that the treatment duration after a VTE should be longer than observed in real-life. However, the analysis was performed on a subcohort with few patients and limited follow-up and must be interpreted cautiously. Our study was designed to collect information during a fiveyear observational period on first VTE, VTE treatment and recurrent VTE from GP data, outpatient pharmacy data and hospitalization records. The PHARMO GP Database is a large data source with ‘real-life data’ which is routinely collected. The GP acts as a gatekeeper in the Dutch healthcare system and all VTE diagnoses should be captured in the GP records. However, as the GP information system was designed for clinical practice, some information relevant for our research may not be available from the GP records. In this study for example, the VTE incidence and recurrence might be underestimated as fatal PE may go undetected. Furthermore, assumptions had to be made for the treatment duration, as the dosing of VKA takes place at the anticoagulation clinics, and only out-patient treatment was collected in the database.


In conclusion, treatment after VTE as captured in observational healthcare data generally follows the Dutch guidelines, except for the duration of LMWH treatment, which was longer than recommended. The observation of shorter anticoagulant treatment among patients with recurrent VTE compared to patients with no recurrence may indicate that anticoagulant treatment was not optimal for these patients.

Ethics approval and consent to participate

Not applicable

Consent for publication

Not applicable

Availability of data and materials

Data will not be shared, due to privacy regulations.

Competing interests

Kerstin Folkerts and Luke Bamber are employees of Bayer HealthCare Pharmaceuticals. Irene Bezemer, Elsa van den Berg, Fernie Penning- van Beest and Ron Herings are employees of the PHARMO Institute. This independent research institute performs financially supported studies for government and related healthcare authorities and several pharmaceutical companies.


This study was funded by Bayer HealthCare Pharmaceuticals / Bayer Pharma AG.

Authors’ contributions

IDB, EJB, KF and LB were involved in the study design. IDB and EJB performed the analysis and wrote the first draft of the manuscript. All authors were involved in critical review and subsequent amendment of the manuscript.


1. Day ISCfWT, Day ISCfWT. Thrombosis: A major contributor to global disease burden. Thrombosis research. 2014.

2. Silverstein MD, Heit JA, Mohr DN, Petterson TM, O’Fallon WM et al. 3rd. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study. Archives of internal medicine. 1998, 158(6): 585-593.

3. Oger E. Incidence of venous thromboembolism: a community-based study in Western France. EPI-GETBP Study Group. Groupe d’Etude de la Thrombose de Bretagne Occidentale. Thrombosis and haemostasis. 2000, 83(5): n657-660.

4. Nordstrom M, Lindblad B, Bergqvist D, Kjellstrom T. A prospective study of the incidence of deep-vein thrombosis within a defined urban population. Journal of internal medicine. 1992, 232(2): 155-160.

5. Cushman M, Tsai AW, White RH, Heckbert SR, Rosamond WD et al. Deep vein thrombosis and pulmonary embolism in two cohorts: the longitudinal investigation of thromboembolism etiology. The American journal of medicine. 2004, 117(1): 19-25.

6. Anderson FA Jr, Spencer FA. Risk factors for venous thromboembolism. Circulation. 2003, 107(23): 9-16.

7. Heit JA, O’Fallon WM, Petterson TM, Lohse CM, Silverstein MD et al. Relative impact of risk factors for deep vein thrombosis and pulmonary embolism: a population-based study. Archives of internal medicine. 2002, 162(11): 1245-1248.

8. Heit JA. Venous thromboembolism: disease burden, outcomes and risk factors. Journal of thrombosis and haemostasis : JTH. 2005, 3(8): 1611-1617.

9. Previtali E, Bucciarelli P, Passamonti SM, Martinelli I. Risk factors for venous and arterial thrombosis. Blood transfusion Trasfusione del sangue. 2011, 9(2): 120-138.

10. Oudega R VWH, Stoffers HEJH, Sival PPE, Schure RI, Delemarre J et al. NHG-Standaard Diepe veneuze trombose. Huisarts Wet 2008, 51(1): 24-37.

11. Buller HR, Agnelli G, Hull RD, Hyers TM, Prins MH et al. Antithrombotic therapy for venous thromboembolic disease: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest. 2004, 126(3): 401S-428S.

12. CBO KvdG. Richtlijn Diagnostiek, Preventie en Behandeling van Veneuze Trombo-embolie en Secundaire Preventie Arteriële Trombose. 2008.

13. Herings RMC, Pedersen L. Pharmacy -based Medical Record Linkage Systems. Strom BL, Kimmel S.E., Hennessy, S., editor. Pharmacoepidemiology, 5th Edition. 5th Edition ed: John Wiley & Sons, Ltd. 2012, 270-286.

14. van Herk-Sukel MP, Lemmens VE, Poll-Franse LV, Herings RM, Coebergh JW. Record linkage for pharmacoepidemiological studies in cancer patients. Pharmacoepidemiology and drug safety. 2012, 21(1): 94-103.

15. Van de Lisdonk EH VdBW, Lagro-Janssen ALM. Ziekten in de huisartspraktijk. 4th ed. Maarssen: Elsevier Gezondheidszorg, 2003.

16. Van der Linden MW WG, De Bakker DH, Schellevis FG. Tweede Nationale Studie naar ziekten en verrichtingen in de huisartspraktijk: klachten en aandoeningen in de bevolking en in de huisartspraktijk.2004.

17. Tagalakis V, Patenaude V, Kahn SR, Suissa S. Incidence of and Mortality from Venous Thromboembolism in a Real-World Population: The Q-VTE Study Cohort. The American journal of medicine. 2013, 126(99): 832.e13-21.

18. Blom JW, Doggen CJ, Osanto S, Rosendaal FR. Malignancies, prothrombotic mutations, and the risk of venous thrombosis. JAMA : the journal of the American Medical Association. 2005, 293(6): 715-722.

19. Martinez C, Cohen AT, Bamber L, Rietbrock S. Epidemiology of first and recurrent venous thromboembolism: A population-based cohort study in patients without active cancer. Thrombosis and haemostasis. 2014, 112(2): 255-263.

20. Kearon C, Kahn SR, Agnelli G, Goldhaber S, Raskob GE et al. Antithrombotic therapy for venous thromboembolic disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008: 133(6): 454S-545S.