Clin Res Cardiol (2021)
DOI DOI https://doi.org/10.1007/s00392-021-01843-w
Application of the sex-specific 99th percentile using four high-sensitivity troponin assays
J. Lehmacher1, J. Neumann1, N. A. Sörensen1, A. Goßling1, T. Hartikainen1, P. Haller1, T. Zeller1, S. Blankenberg1, D. Westermann1, für die Studiengruppe: BACC
1Klinik und Poliklinik für Kardiologie, Universitäres Herz- und Gefäßzentrum Hamburg GmbH, Hamburg;
Background: Current guidelines recommend using high-sensitivity troponin assays for diagnostic workup in patients with suspected myocardial infarction (MI). The 99th percentile is used as upper reference limit for diagnosis of MI.  Several validated high-sensitivity troponin assays are commercially available and the 99th percentile shows a wide inter-assay variation. Additionally, gender-specific 99th percentiles have been recommended. We aimed to compare the diagnostic performance of the sex-specific 99th percentile using four high-sensitivity troponin (hs-Tn) assays in a large population of patients with suspected MI.

Methods: We measured concentrations of hs-Tn at presentation and after 3 hours in 2,718 patients presenting to the emergency department of the University Hospital Hamburg-Eppendorf with suspected MI, using Roche hs-TnT, Siemens-, Abbott- and Mitsubishi hs-TnI assays. The sex-specific 99th percentile was used as cut-off concentration for each assay. Final diagnosis was made independently by two cardiologists, considering all available clinical data but unaware of hs-TnI concentrations. Diagnoses were adjudicated according to 4th Universal Definition of MI. Parameters for diagnostic performance were calculated, applying the ESC 0/3h algorithm, as well as receiver-operating-characteristic (ROC) -curves with corresponding area under the curve (AUC).

Results: A total of 368 patients were diagnosed as having NSTEMI. Median age was 64 years and 64.3% of the population were male. Hypertension was present in 65.5% of patients and 34.8% had previously known hyperlipoproteinemia. Roche hs-TnT provided high sensitivity [98.9% (95% CI 97.2% - 99.7%)], corresponding to a very high negative predictive value (NPV) [99.7% (95%CI 99.3 - 99.9)]. Hs-TnI assays showed lower, but still high sensitivity and NPV, with sensitivity constantly above 85% and NPV above 96% respectively. Considering specificity and PPV, Siemens hs-TnI performed best among all assays, with specificity of 91.3% (95% CI 89.9 - 92.5) and PPV of 61.7% (95% CI 56.9 - 66.5). Comparing hs-TnT with the three hs-TnI assays, Roche hs-TnT showed highest sensitivity and NPV, at the cost of specificity and PPV, which was higher in each hs-cTnI assay used (Table 1). AUC was high among the 4 assays, with AUC in Siemens- and Mitsubishi hs-TnI being significantly greater than Roche hs-TnT (p < 0.001; p = 0.0051) (Figure 1).

Conclusion: Application of sex-specific 99th percentile for diagnosis of MI using the ESC 0h/3h algorithm for four high-sensitivity troponin assays resulted in good diagnostic performance with high sensitivity and specificity for each assay.

Table 1: Application of the sex-specific 99th percentile using the 0/3h ESC algorithm

Assay

Sensitivity (%)

Specificity (%)

NPV (%)

PPV (%)

Roche hs-TnT

98.9 (97.2 - 99.7)

76.2 (74.4 - 78.0)

99.7 (99.3 - 99.9)

40.4 (37.1 - 43.8)

Abbott hs-TnI

88.0 (83.7 - 91.5)

87.3 (85.7 - 88.9)

97.7 (96.8 - 98.4)

53.5 (48.9 - 58.0)

Siemens hs-TnI

86.0 (81.5 - 89.7)

91.3 (89.9 - 92.5)

97.5 (96.6 - 98.2)

61.7 (56.9 - 66.5)

Mitsubishi hs-TnI

85.6 (77.6 - 91.5)

87.2 (84.1 - 89.8)

96.8 (94.8 - 98.1)

56.9 (49.0 - 64.5)

https://dgk.org/kongress_programme/jt2021/aP623.html