Rate of venous thromboembolismby cancer type compared to the general population using multiple linked databases

  • A.J. W
  • J. W
  • T.R. C
 et al. 
  • 4


    Mendeley users who have this article in their library.
  • N/A


    Citations of this article.


Introduction: Strategies to prevent venous thromboembolism (VTE) are critical in reducing the impact of this important complication in cancer patients. However, few large scale studies detail which types of cancer present the highest rate of VTE compared with the general population. Aim: We aimed to quantify the rate of VTE prospectively for different cancer types compared to the general population. Patients: We used the UK General Practice Research Database and the recently linked Hospital Episode Statistics data to identify 151,267 incident cancer patients and 704,544 controls frequency matched for age. Death certificate data were available from the Office for National Statistics. We identified people with a first VTE during the study period on the basis of a medical code for VTE in either general practice or hospital records if supported by evidence of anticoagulant, or if the patient died of VTE (regardless of anticoagulation status). Absolute rates were calculated, while Cox proportional hazards models (age and sex adjusted) were used to calculate relative rates. Results: Mean duration of follow-up per person was 4.0 years, while there were 4,755 (3.1{%}) cancer patients and 6,161 (0.87{%}) controls with a post diagnosis VTE event. Overall the absolute rate of VTE was 0.94{%} per year (95{%} confidence interval (CI) 0.91-0.96) in cancer patients, while for controls the rate was 0.22{%} (CI 0.21-0.22). After adjustment for age and sex, this resulted in a hazard ratio (HR) of 3.8 (CI 3.6-3.9) for cancer overall. While nearly all cancer types presented a higher rate of VTE than controls, there was a dramatic variation in the rates between cancer types (see Fig. 1). Highest relative rates were seen in pancreatic (HR=14.3, CI 12.0-17.1), ovarian (HR 13.3 CI 11.5-15.2), brain (HR=10.9, CI 9.1-13.0) and oesophageal cancers (HR=10.9, CI 9.3-12.7). By contrast, there was no increase in rate in people with laryngeal cancers compared to the general population (HR=1.1, CI 0.5-2.3). Summary: For the first time our study provides comprehensive ascertainment of VTEs by cancer type through combining primary care, secondary care and cause of death data for a defined population. Furthermore the large size of our study enables more precise estimates to be made for the current rates of VTE than previous studies. We have found similar overall absolute (Figure presented) and relative rates to those previously reported [1], however we have also shown that that this rate varies by over 10 fold by cancer type. This will have important implications for the appropriate selection of patients for VTE prophylaxis.

Author-supplied keywords

  • *data base
  • *hemostasis
  • *neoplasm
  • *population
  • *thrombosis
  • United Kingdom
  • anticoagulant agent
  • anticoagulation
  • brain
  • cancer patient
  • cause of death
  • confidence interval
  • death certificate
  • diagnosis
  • follow up
  • general practice
  • hazard ratio
  • hospital
  • human
  • larynx cancer
  • medical record
  • patient
  • primary medical care
  • prophylaxis
  • proportional hazards model
  • statistics
  • venous thromboembolism

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document


  • Walker A.J.

  • West J.

  • Card T.R.

  • Crooks C.

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free