Abstract 3358: Multibiomarker Approach Predicts Adverse Cardiovascular Outcomes in Women Evaluated for Suspected Ischemia without Obstructive CAD Better than Traditional Risk Predictors: Results from the NHLBI-sponsored Women’s Ischemia Syndrome Evaluation (WISE)
Background Inflammatory markers and low hemoglobin (Hgb) levels predict adverse cardiovascular (CV) outcomes in selected populations. The evaluation of traditional risk predictors in this context is indicated. We investigated the association between these biomarker [high sensitivity C-reactive protein (hs-CRP), interleukin-6 (IL-6), serum amyloid A (SAA) and Hgb] levels and adverse CV outcomes in women without obstructive coronary artery disease (CAD).
Results A total of 381 women with non-obstructive CAD (<50% diameter stenosis in all coronary arteries) of which 61% (232/381) had no coronary obstruction ≥20% were studied. Over 5 ± 2.4 years, adverse CV outcomes (death, nonfatal MI, nonfatal stroke, hospitalization for CHF or other vascular event) occurred in 18% (69/381) and 5.3% (20/381) died. Women with no abnormal biomarkers had fewer adverse CV outcomes (8.7%) compared to those with one (17.7%), two (17.5%), three (32.1%), or four (41.2%) abnormal biomarkers (p<0.0001 for trend). Women with no abnormal biomarkers also had fewer deaths from any cause (<1%) than women with one (2.6%), two (8.8%), or ≥three (12.3%) abnormal biomarkers (p<0.0001 for trend). By multivariate Cox proportional hazards analysis (table⇓), women with one, two, and ≥ three abnormal biomarkers were increasingly more likely to have adverse CV outcomes than women with no abnormal biomarkers. The biomarker panel provided independent predictive information over and above the Framingham Risk Score.
Conclusions Among women with suspected myocardial ischemia without obstructive CAD, a multi-biomarker approach utilizing inflammatory marker and Hgb levels better predicted adverse outcome risk and all-cause mortality than any one marker alone. The biomarkers added to the prognostic information that could be obtained from traditional CV risk predictors like the Framingham Risk Score.