PARIS – Patients with Sjögren’s syndrome appear to have a higher risk of myocardial infarction, but not stroke or transient ischemic attack, than does the general population, based on an analysis of administrative health data from British Columbia.
"Our findings support monitoring for coronary artery disease in addition to management and modification of cardiovascular disease risk factors to reduce the risk of MI in Sjögren’s syndrome patients," Dr. Marko Yurkovich said at the annual European Congress of Rheumatology.
Dr. Marko Yurkovich
An increased risk of MI and cerebrovascular accident (CVA) is already well known in more common rheumatic diseases such as rheumatoid arthritis and systemic lupus erythematosus, but the risk for either MI or CVA at the general population level in patients with Sjögren’s syndrome remains unknown, said Dr. Yurkovich, an internal medicine resident at the University of British Columbia, Vancouver.
The investigators used administrative health data from all people in British Columbia who had an outpatient visit or hospitalization during 1990-2010 to locate those who were older than 18 years and met the study’s criteria for a diagnosis of Sjögren’s syndrome. The criteria were two or more ICD-9-CM codes for the disease 2 or more months apart but within a 2-year period that were made by nonrheumatologists or one or more ICD-9-CM codes for the disease that were made by a rheumatologist or during a hospitalization.
The study excluded patients who had diagnoses of other inflammatory arthritides after the Sjögren’s diagnosis or cases in which an initial diagnosis of Sjögren’s by a nonrheumatologist was not confirmed by a rheumatologist at a later visit. MI and CVA diagnoses were defined from ICD-9-CM codes from hospital or death certificates.
"This diagnostic algorithm has been validated in a Canadian context and shown to have quite a high specificity and sensitivity, both over 95%," Dr. Yurkovich said.
In 1,176 patients with Sjögren’s syndrome, the investigators identified 28 MIs for an incidence rate of 7.7 per 1,000 person-years, compared with 138 MIs in 11,879 control individuals from the general population matched for birth year, sex, and calendar year of exposure. Patients with Sjögren’s had a significantly elevated risk for MI, compared with controls, with an incidence rate ratio of 2.19 (95% confidence interval, 1.40-3.31) and a multivariate relative risk of 2.36 (95% CI, 1.48-3.78), based on incidence rates of 7.7 per 1,000 person-years in patients with Sjögren’s and 3.4 per 1,000 person-years in controls. The relative risk for MI did not differ when the results were stratified by sex.
However, risk of incident CVA was not significantly higher among patients with Sjögren’s than in the controls, based on an incidence rate of 5.1 per 1,000 person-years among patients and 3.4 per 1,000 person-years in controls that yielded an incidence rate ratio of 1.49 and multivariable relative risk of 1.64.
The multivariable analysis controlled for angina, chronic obstructive pulmonary disease, obesity, Charlson Comorbidity Index, the number of hospitalizations in the year before the index date, and number of medications, including oral glucocorticoids, cardiovascular drugs, antidiabetic medication, hormone replacement therapy, contraceptives, fibrates, statins, NSAIDs, and COX-2 inhibitors.
"We found that the risk of MI was even further elevated within 1 year following initial diagnosis," with an incidence rate ratio of 3.6 before 1 year of follow-up, which decreased to 1.7 during 1-5 years of follow-up and 1.9 and after 5 years, Dr. Yurkovich said. He suggested that the increased risk of MI in the first year might occur because the acute inflammatory state in Sjögren’s is highest at the time of diagnosis or because the most susceptible patients had an MI early on and were no longer included in subsequent analyses.
The results remained statistically significant for MI and not for CVA after the investigators performed sensitivity analyses for unmeasured confounding variables using hypothetical low and high prevalences of 10% and 20% and low and high strengths of association in terms of odds ratios (1.3 and 3.0).
A member of the audience wondered whether antiphospholipid antibodies could have contributed to the increased risk for MI because patients with Sjögren’s syndrome typically don’t use medications such as glucocorticoids that increase atherosclerosis, and inflammation is not typically elevated in the condition, with normal C-reactive protein levels. Dr. Yurkovich replied that the cohort used glucocorticoids at a significantly higher rate than did the control group (20% vs. 4%), perhaps because they were put on them initially around the time of diagnosis, but nevertheless, the medication was one of the variables they adjusted for in the multivariable analysis. He noted that the investigators did not have antiphospholipid antibody data but they hope to obtain those and other autoantibody profiles for subsequent analyses.