Limited sex-specific performance of the GRACE 2.0 score to predict reinfarction or death in NSTEMI patients during pro-inflammatory states

Abstract

Abstract Background Non-ST segment elevation myocardial infarction (NSTEMI) is particularly common in female patients with acute coronary syndrome (ACS), as women tend to present at older age and with higher comorbidity burden than men. High-sensitive C-reactive protein (hs-CRP) and growth differentiation factor 15 (GDF-15) plasma levels are elevated in NSTEMI patients and mirror inflammatory states, while conferring poor prognosis following the acute event. The GRACE 2.0 score is an established tool to stratify ACS patients according to their risk and to guide treatment, however, its sex-specific performance has not been systematically investigated. Purpose We sought to investigate the sex-specific performance of GRACE 2.0 to predict reinfarction or death within one year in NSTEMI patients in the presence or absence of inflammatory states. Methods Between 2009 and 2019, 2111 NSTEMI patients were enrolled in the multicentre prospective SPUM-ACS cohort of which 2001 completed follow-up at one year. Hs-CRP and GDF-15 plasma levels at presentation were measured by immunoassay and risks for MI or all-cause death within one year were estimated by GRACE 2.0. Discrimination was evaluated by the area under the receiver operating curve (AUC) and compared between both sexes using the De-Long's test. Results A total of 411 (20.5%) women and 1590 (79.5%) men with NSTEMI were included in the study. At presentation, women were older (69.8±11.8 years vs. 63.6±12.0 years, P<0.01), more likely to have a history of systemic inflammatory disease (4.6% vs. 2.1%, P<0.01) and had higher GDF-15 levels (median, 1325 vs. 1093 pg/ml, P<0.01) compared to men. In the entire cohort, the performance of GRACE 2.0 to predict the composite of MI or death was similar in both sexes (AUC: 0.66 vs. 0.74, P=0.24). However, score performance in patients with hs-CRP≥2 mg/l was significantly worse in women than in men (AUC: 0.61 vs. 0.76, P<0.05). These findings were recapitulated in patients with plasma GDF-15≥1133 pg/ml (AUC: 0.61 vs. 0.75, P<0.05) suggesting that GRACE 2.0 performs poorly in female NSTEMI patients with excessive inflammation at presentation. Intriguingly, opposite findings were observed in patients with hs-CRP<2 mg/l (AUC: 0.81 vs. 0.58, P<0.05). Also, GDF-15 plasma levels correlated with hs-CRP at presentation (Rs=0.35, P<0.01) and predicted the composite of MI or death well in men, while the discriminative ability was surprisingly poor in women (0.80 vs. 0.64, P<0.05), results similarly observed with hs-CRP. Conclusions Our results suggest that the established risk prediction model GRACE 2.0 lacks discriminative power in female NSTEMI patients during high inflammatory states. In addition, inflammatory biomarkers such as GDF-15 show sex-specific differences in their predictive utility of adverse events which should be appreciated by future risk-scoring algorithms to overcome inequalities in certain patient populations presenting with ACS. Funding Acknowledgement Type of funding sources: Other. Main funding source(s): Swiss National Science FoundationFoundation for Cardiovascular Research