The ACC/AHA criteria for identification of very high-risk CVD patients have limited discriminative power
Identification of vascular patients at very high risk for recurrent cardiovascular events: validation of the current ACC/AHA very high risk criteriaLiterature - van den Berg MJ, Bhatt DL, Kappelle LJ, et al. - Eur Heart J. 2017; published online ahead of print
- In the SMART cohort, the incidence of recurrent MACE was on average 2.4/100 person years (PY) (95% CI 2.3-2.5/100 PY) and 57% of the SMART population met the ACC/AHA criteria of VHR (incidence 2.7/100 PY, 95% CI 2.5–2.9/100 PY).
- In the REACH registry, the incidence for recurrent MACE was 5.1/100 PY (95% CI 5.0–5.3/100 PY) and 64% of the REACH participants met the AHA/ACC criteria for VHR.
- Patients who did not meet the ACC/AHA criteria for VHR, had a MACE incidence of 2.0/100 PY in SMART and 3.9/100 PY in REACH.
- There was a statistically significant difference in incidence rates between patients meeting the ACC/AHA VHR criteria and those who did not, in both cohorts.
- The highest incidence for recurrent MACE were found in patients with an eGFR <45mL/min/1.73m2 (incidence in SMART 6.9/100 PY, 95% CI 5.8-8.1/100 PY and incidence in REACH 9.4/100 PY, 95% CI 8.8–10.1/100 PY), as well as in patients with polyvascular disease (incidence 5.0 in SMART and 7.7/100 PY in REACH).
- The ACC/AHA VHR criteria had poor to moderate discrimination in identifying patients who developed a recurrent CV event during follow-up. C-statistics were 0.54 (95% CI 0.52–0.55) in SMART and 0.53 (95% CI 0.52–0.53) in REACH.
- The single risk factors polyvascular disease and age >70 led to improved re-classification of patients at very high risk of a recurrent event during follow-up, when compared with the ACC/AHA VHR criteria.
- The proportion of patients with a predicted 10-year risk ≥30% was much higher in patients aged ≥70 years (68% in SMART and 58% in REACH) when compared to the overall study populations (25% in SMART and 30% in REACH).
- In patients at very high risk according to the ACC/AHA criteria, the proportion of patients with >30% 10-year risk was 29% in SMART and 33% REACH, while in patients who were not at VHR according to the ACC/AHA, this was 19% both in SMART and REACH.
In the SMART study and the REACH registry, single very high risk factors or simple criteria, such as the ACC/AHA VHR criteria, had limited discriminative power to identify patients at highest risk of recurrent MACE. These findings suggest that there is a need to improve the identification of very high-risk patients in the secondary CVD prevention.