Inflammation secondary to active rheumatoid arthritis (RA) may be an independent risk factor for cardiovascular disease (CVD). Theoretically, by way of decreasing inflammation, RA pharmacotherapeutics may effectively reduce CVD risk. Evidence for CVD and mortality risk reduction has been published for RA patients treated with methotrexate, however; no previous studies have investigated whether the same may be true for RA patients treated with biologics. Here, Jacobsson et al (J Rheum 2005;32(7):1213) investigate incident CVD events in RA patients treated with TNF inhibitors.
Methods:
CVD related deaths and hospitalizations occurring between February 1, 1999 and December 31, 2001 were determined in a population-based registry of TNF inhibitor treated RA patients in southern Sweden with no prior hospitalizations for CVD events. This was compared to the CVD related deaths and hospitalizations occurring between July 1, 1997 and December 31, 2001 in a geographically similar population-based cohort of RA patients without prior hospitalizations for CVD events who were not receiving treatment with TNF inhibitors.
Results:
531 TNF inhibitor treated patients with a total of 656 person-years of follow-up were compared to 543 non-TNF inhibitor treated patients with a total of 2067 person years of follow-up. 91 patients were included in both groups as their TNF inhibitor treatment status changed over the study period. At the time of inclusion in either cohort, TNF inhibitor treated patients tended to be younger than non-TNF inhibitor treated patients (median age 55 vs. 61 years, respectively). TNF inhibitor treated patients tended to demonstrate higher measures of RA disease activity and severity (i.e. HAQ, pain, patient global assessment), a greater proportion of patients receiving glucocorticoids (75% vs. 22%, respectively), and a slightly greater proportion of patients with comorbid diabetes and COPD than non-TNF inhibitor treated patients. Median RA disease duration was similar for TNF inhibitor treated and untreated patients (11 12 years). The prevalence of traditional CVD risk factors (e.g. hypertension, smoking, etc) was not known in either group.
| TNF treated (n=531) |
Non-TNF treated (n=543) |
|
| CVD Events, number: type | n=13 (MI=6, CVA=4, other =3) |
n=85 (MI=33, CVA=15, CHF=12, ruptured AAA=2, other=23) |
| Incidence per 1000 person-yrs (age-sex adjusted) | 14.0 (95% CI 5.7-22.4) |
35.4 (95% CI 16.5-54.4) |
| Standardized Morbidity Ratio (SMR) for incident CVD relative to Malmo, Sweden | 1.57 (95%CI 0.72-1.57) |
2.28 (95% CI 1.79-2.77) |
In models individually adjusting for confounders significantly associated with incident CVD (i.e. HAQ > 1.38, patient global assessment > 52mm, previous DMARD treatment with > 3 DMARDs), treatment with a TNF inhibitor was associated with an approximately 50% reduction in risk of incident CVD events (RR 0.43 0.46; 95% CI 0.24-0.87). Similar results were obtained when only the period of temporal overlap between the two groups was considered, as well as when follow-up in the two groups was restricted to 1 and 2 years, and when pulmonary vascular disease and CHF were excluded from the definition of CVD events.
Conclusion:
After controlling for RA disease severity, TNF inhibitor treatment was associated with a significant reduction in the risk of new CVD events compared to RA patients not treated with TNF inhibitors. RA patients treated with TNF inhibitors demonstrated incident CVD morbidity in a range compatible with the background incidence of CVD morbidity in the local community.
Editorial Comment:
This report is the first population-based investigation to establish a link between TNF inhibitor use and CVD risk reduction in RA. It is important to note that, after adjusting for confounders that are related to both CVD risk and the likelihood that an RA patient would be prescribed a TNF inhibitor (e.g. elevated pain, swelling and functional disability), prescription of TNF inhibitors was associated with a substantial decrease in incident CVD in what would be considered a highly at-risk population. Nevertheless, some aspects of the study design may call into question the absolute magnitude of the results as reported, such as the high number of patients accounted for in both the case and comparison groups and the relatively short follow-up periods for both groups. Finally, conventional CVD risk factors, which were not available from the databases utilized, may account for some of the differences noted between the two groups.
This study by itself can provide little insight into the mechanism behind the reduction in incident CVD risk observed with TNF inhibitor therapy; although possibilities include a reduction in atherogenic systemic inflammatory mediators, favorable effects on lipid profile, improvement in physical activity due to a decrease in joint pain and swelling, and the ability to reduce or discontinue glucocorticoids, among others. Any, or all, of these mechanisms may play a role in the decrease in CVD events observed and will require additional study to elucidate the relative importance of each.
