Article Text


OP0109 Higher incidence of thromboembolic events among patients with rheumatoid arthritis vs. osteoarthritis, and vs. no arthritis, in the general practice research database (gprd)
  1. DJ Watson,
  2. T Rhodes
  1. Epidemiology, Merck Research Laboratories, West Point, USA


Background Persons with Rheumatoid arthritis (RA) have been noted to have increased all-cause and cardiovascular (CV) disease mortality relative to those without RA. Data on mortality in patients with RA vs those with osteoarthritis (OA), and data on nonfatal + fatal CV incidence in patients with RA relative to those with OA and those with no arthritis are lacking.

Objectives To compare the incidence of all cause mortality, myocardial infarction (MI), sudden death (SD), cerebrovascular events (stroke; excluding transient ischaemic attack), CV death (fatal MI, sudden death, fatal stroke), and all thromboembolic events in RA patients compared with patients with OA but not RA, and compared with those with no arthritis, in the GPRD.

Methods The GPRD is a general practice research database that currently contains over 35 million patient years of data from the UK and represents 6% of general practice patients in England and Wales. This was an observational cohort study of patients >40 years old from GPRD practices in which >80% of patients had recorded visits. Patients with history of MI or stroke prior to beginning of follow-up were excluded. Person-time was calculated for each patient by gender, age (5 year intervals) and diagnosis (RA; OA but not RA, and no arthritis). The rate/1000 person-years was computed for each gender, age, and diagnosis stratum. For each patient, only the first incident endpoint was counted and attributed to the stratum in which it occurred. Gender and age-adjusted incidence rates and ratios of rates of each endpoint during person-time with RA vs that with OA but not RA, and vs that with no arthritis were computed using Poisson regression.

Results 594 practices contributed 6,533,474 patients to the analysis. The incidence of RA was 0.68/1000 patient years in men and 1.33/1000 patient years in women. Age-adjusted rates of all endpoints were significantly increased for both men and women with RA compared to those with OA but not RA (except for sudden death, which was significantly increased in women only). Age- and gender-adjusted rate ratios (95% CI) were: all-cause mortality 1.72 (1.67, 1.78); MI 1.32 (1.24, 1.41); SD 1.33 (1.09, 1.63); stroke 1.26 (1.18, 1.34); CV death 1.41 (1.24, 1.61); and all thromboembolic events 1.31 (1.25, 1.39). Similar results were seen when patients with RA were compared with those with no arthritis: age- and gender-adjusted rate ratios (95% CI) with RA vs no arthritis were: all-cause mortality 1.60 (1.55, 1.64); MI 1.55 (1.46, 1.65); SD 1.30 (1.07, 1.58); stroke 1.37 (1.29, 1.45); CV death 1.45 (1.35, 1.57); and all thromboembolic events 1.47 (1.41, 1.54).

Conclusion In this study patients with a diagnosis of RA were 70% more likely to die and 30–40% more likely to suffer an acute major thromboembolic event compared with patients with a diagnosis of OA but not RA. In addition, patients with a diagnosis of RA were 60% more likely to die, and 30–50% more likely to suffer an acute major thromboembolic event compared with patients with no arthritis diagnosis.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.