Background/Purpose:  Rheumatoid arthritis (RA) has been established as a risk factor for cardiovascular disease (CVD), with systemic inflammation being linked with atherosclerosis. Traditional CVD risk indices like the Framingham Risk Score (FRS) underestimate CVD risk in RA. RA patients that develop CVD often have more inflammation yet lower levels of atherogenic cholesterol in what is labelled “the lipid paradox”. Additionally, evidence suggests RA disease activity may modify the cardioprotective function (versus the absolute value) of cardioprotective high density lipoprotein (HDL). We explore the link between CVD, disease activity, lipid levels, and apoA1 (an atheroprotective HDL subcomponent) in an early inflammatory arthritis (EIA) cohort.

Methods:  Data was obtained from subjects followed in a single centre EIA cohort with systematic collection of disease activity measures, comorbidities and serum lipids. FRS scores were calculated for each patient, and new CVD (myocardial infarction, unstable angina, stroke) confirmed by chart review. Annual lipid profiles (total cholesterol (TChol), triglycerides (TG), low density lipoprotein (LDL) and HDL) were correlated with disease activity (DASESR283variable) using Spearman’s rho (first profile obtained after median (IQR) of 25 (0, 40) months follow-up). ApoA1 was measured by ELISA in serial samples from 22 patients matched for age and DAS28ER3variable (5 with new CVD) and differences in apoA1, lipid levels, and DAS28ESR-3variable over time in both non-CVD and CVD patients compared using non-parametric tests and logistic regression.

Results:  278 patients with EIA (<12months symptoms at baseline) and no prior CVD history (mean(SD) age 49(15) years, 75% female, 67% RA) were followed for a median (IQR) of 67 (25, 102) months. Considering all visits, there was modest negative correlation between DAS28ESR-3variable and HDL levels (r=-0.12, p<0.001), a trend for negative correlation with TChol level (r=-0.07 p=0.03), but no correlation for LDL and TG. CRP correlated with TChol (r=-0.12 p<0.0001), TG (r=0.14 p<0.0001), and HDL (r=-0.31 p<0.0001). A good or moderate EULAR treatment response was achieved by 70% of the cohort at last visit and there was greater increase in LDL levels in these individuals than those with no response (mean(SD) +0.24(0.7) vs +0.01(0.6), p=0.02). Five subjects developed CVD after a median (IQR) of 64(42, 104) months. FRS increased from baseline to last visit (median (IQR) 4.4(1.2, 9.1) vs 5.1(2.1,10.8) p<0.0001) however there was no difference in FRS at baseline, last visit or change in FRS between those with or without CVD. Baseline high FRS (FRS in fourth quartile) was not significant in cox-proportional hazards modelling of CVD-free survival. ApoA1 levels modestly fluctuated over time, but did not correlate with HDL or other lipids nor associate with CVD.

Conclusion:  In EIA, high disease activity is associated with lower levels of pathogenic lipids, supporting the lipid paradox. Neither Framingham risk scores nor apoA1 levels adequately associate with CVD risk in these EIA patients. Alternate measures of lipid metabolism may be more predictive of CVD.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/exploring-the-link-between-ra-disease-activity-lipid-levels-and-cardiovascular-disease-in-an-early-inflammatory-arthritis-cohort/