FRI0137

RISK FACTORS FOR MAJOR ADVERSE CARDIOVASCULAR (CV) EVENTS (MACE) IN RHEUMATOID ARTHRITIS (RA) PATIENTS TREATED WITH TOCILIZUMAB (TCZ)

V. U. Rao^{1 2}A. Pavlov³M. Klearman^2,*D. Musselman²J. T. Giles⁴J. M. Bathon⁴N. Sattar⁵J. S. Lee⁶

¹Indiana Heart Physicians, Indianapolis, ²Genentech, S San Francisco, United States, ³Everest, Toronto, Canada, ⁴Columbia U, New York, United States, ⁵U Glasgow, Glasgow, United Kingdom, ⁶Roche, Nutley, United States

Background: The increased incidence of MACE in RA cannot be explained by traditional CV risk factors alone.¹

Objectives: To examine associations of lipids, inflammation and RA disease activity with risk for MACE in TCZ-treated RA pts.

Methods: Pooled data of TCZ-exposed pts from phase 3 trials and long-term extensions were included in post hoc analyses. 50 pts with MACE (non-fatal myocardial infarction, non-fatal stroke or CV death) and 3,936 pts without MACE were identified via blinded, independent adjudicator. Cox proportional hazards models were used to evaluate the associations of various demographic, laboratory and disease measures (at baseline, wk 24 and change from baseline to wk 24) with time to future MACE. Simple univariate and multivariate models were estimated, and resulting hazard ratios with 95% CIs were used to assess predictors.

Results: For MACE during TCZ exposure, univariate analysis showed baseline age, history of cardiac disorders, statin use, TJC, SJC, DAS28, total cholesterol (TC)/HDL ratio, ApoB and ApoB/ApoA1 ratio as predictive (all p<0.05). Age, history of cardiac disorders, DAS28 and TC/HDL ratio were joint predictors in multivariate models. Neither reduction in inflammation (CRP, ESR) nor lipid increase at wk 24 was statistically significantly associated with future MACE. Cox models using single predictors and adjusting for age and baseline values showed change in DAS28, area under the curve (AUC) of DAS28, EULAR response, TJC and SJC at wk 24 were statistically significant predictors of future MACE (Table). Greater reductions in DAS28, SJC and TJC from baseline to wk 24 were inversely associated with MACE (Table).

Table. Associations between RA Parameters and Future MACE^a
Assessments, wk 24	Age-Adjusted Hazard Ratio (95% CI)	p
SJC (28)^b	1.092 (1.051, 1.136)	<0.0001
TJC (28)	1.054 (1.018, 1.092)	0.0034
DAS28	1.351 (1.118, 1.633)	0.0018
EULAR (Good vs. No Response)	0.295 (0.122, 0.714)	0.0067
AUC of DAS28 to wk 24, score-years	1.714 (1.018, 2.886)	0.0426
Assessments, change from baseline to wk 24^c
SJC	1.084 (1.034, 1.136)	0.0008
TJC	1.043 (1.001, 1.086)	0.0439
DAS28	1.293 (1.043, 1.602)	0.0191
^aHR estimated from Cox model with stated variable, age and other adjusters as noted; no imputation of missing values except last-observation-carried-forward for missing SJC or TJC at wk 24. ^bInterpretation of HR: After adjustment for age, one additional SJC at wk 24 was associated with a 9% increased hazard for MACE. ^cAlso adjusted for value of predictor; HR >1 indicates pts with less improvement were at higher risk for MACE.

Conclusions: Among traditional CV risk factors and disease activity measures, age, DAS28, TC/HDL and history of cardiac disorders were independently associated with MACE in RA pts subsequently treated with TCZ. Risk for MACE was broadly linked to elevated baseline disease activity and less robust therapeutic response at wk 24, whereas no statistically significant association between lipid change and MACE risk was observed. These findings suggest that mitigating CV risk in RA pts may require a multifaceted approach, including effective disease activity control.

References: 1. Gonzalez A et al. Ann Rheum Dis 2008;67:64-9.

Disclosure of Interest: V. Rao Employee of: Roche/GenentechA. Pavlov Consultant for: Everest Clinical Research Services, Inc.M. Klearman Employee of: Roche/GenentechD. Musselman Employee of: Roche/GenentechJ. Giles Consultant for: Roche, UCB, RegeneronJ. Bathon: None DeclaredN. Sattar Grant / Research support from: Roche Consultant for: Roche, Pfizer Speakers Bureau: RocheJ. Lee Employee of: Roche