Direct comparison of very early AMI rule-out algorithms

Abstract

Abstract Introduction Early rule-out of AMI may have a considerable impact on healthcare spending. Numerous algorithms for early rule-out of AMI has been suggested within the last year. However, most algorithms are limited by the need for consecutive, timely separated in-hospital blood samples, prolonging length of hospital stay. Very early rule-out algorithms, providing necessary biomarker results shortly after arrival to hospital or even before, could reduce the burden on EDs and cardiology departments. Many algorithms have been evaluated, validated and compared in several studies; however, often in different subpopulations of larger studies. Reported differences in diagnostic performance of these algorithms may very likely have been caused by differences between the tested patient subgroups. Therefore, a direct comparison of very early rule-out algorithms in a single cohort with all necessary information available for all algorithms is needed. Purpose In this study we aimed to compare the ESC 0h/1h algorithm with ten rule-out algorithms, including the ESC 0h/3h algorithm, an in-hospital Dual-Marker strategy (DMS) (combining hs-cTnT with copeptin), and seven very early algorithms (potential rule-out at admission) for rule-out of AMI. Four algorithms are strictly prehospital and three combines prehospital and in-hospital blood samples, see table 1 for description of evaluated algorithms. Methods The diagnostic performance of the rule-out algorithms was compared in patients with suspected AMI from a randomized, controlled, multicenter trial. We only included patients who had all required information for each diagnostic algorithm available. AMI was adjudicated by at least two cardiologists. Results We included 1.601 patients, see table 2. Of these 136 (8.5%) had type 1 AMI. We found that 7 of 11 algorithms, including the present ECS-recommended 0h/1h algorithm, performed with acceptable sensitivities above 98% and specificities between 13.5 and 52.0%. The ESC 0h/3h algorithm had a unacceptably low sensitivity of 87.5%. Four of the very early rule-out algorithms (the HEART score, the Modified prehospital HEART score, the Modified prehospital DMS, and the Modified prehospital/in-hospital DMS) all performed with excellent sensitivity of 100% and thereby missing no AMIs. Of these, the HEART score derivates ruled out 13.5–16.2% of pts. without AMI, while the two DMS derivates ruled out 14.3–27.4% of pts without AMI. Compared with the ESC 0h/1h algorithm the very early rule out algorithms provide the necessary biomarker results at least 1 hour earlier, and when well-performing POCT analyses becomes available for copeptin and troponin analysis, even before arrival to hospital. Conclusions Four very early rule out algorithms performed excellent in the rule out of AMI, with 100% sensitivities and specificities of up to 27.4%. This enables safe rule out of AMI shortly after arrival to hospital, and in future, potentially already in the ambulance. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): National budget only - Danish Heart Foundation and Independent Research Fund Denmark