Background

A recent hospital-based study suggested a 10-fold increased risk of pulmonary embolism among individuals with systemic lupus erythematosus (SLE) in the year following hospital admission. It is unknown whether the risk is similar among the nonhospitalized SLE population. We estimated the risk of incident pulmonary embolism (PE) and deep venous thrombosis (DVT) events, as well as the associated time trend, among incident cases of SLE compared with general population controls using physician billing and hospitalization data for the entire province of British Columbia, Canada (~5 million).

Methods

Our data included all visits to health professionals and hospital admissions covered under the province's universal healthcare plan from 1 January 1990 until 31 December 2007 for all individuals ≥18 years of age. We conducted a matched cohort study among patients meeting the following criteria: ≥18 years of age, and new diagnosis of SLE based on the following algorithm: one ICD code for SLE on rheumatologist visit billing data or on hospitalization data, or two ICD codes for SLE at least 2 months and no more than 2 years apart on a physician visit by a nonrheumatologist. Controls were selected from the general population, on a 10:1 ratio for each case, matched by birth year, sex and calendar year of exposure. The outcomes, PE and DVT, we identified based on one ICD code for PE in hospitalization data; and one ICD code for DVT in either outpatient or hospitalization data. We estimated relative risks (RRs) of PE and DVT in SLE cases compared with matched general population controls, after adjusting for age, sex, comorbidities, trauma, fracture, surgery, and hospitalizations.

Results

Among 5,156 individuals with SLE, 54 developed PE and 92 developed DVT. Compared with age-matched, sex-matched, and entry-time-matched controls (n = 51,560), the RRs were 4.9 (95% CI = 3.4 to 6.8) for PE and 4.5 (95% CI = 3.5 to 5.7) for DVT. These RRs attenuated slightly after adjusting for covariates, but remained significant (Table 1). When we evaluated the impact of follow-up time, the RRs for PE, DVT and PE or DVT in SLE patients as compared with non-SLE cases were the largest in the first year (Table 2). The estimates decreased over time and were not significant after 5 years of follow-up with the exception of DVT.

Table 1
Full size table
Table 2 Incidence rates and relative risks for PE, DVT and PE or DVT in patients with SLE during follow-up
Full size table

Conclusion

This is the first population-based study assessing the risk of PE and DVT in patients with SLE. These findings support increased monitoring of venous thromboembolic complications and risk factors in SLE patients, especially during the first year after disease onset.