The optimal cutoff values of high-sensitivity troponin and NT-proBNP for the risk-stratification of patients with acute pulmonary embolism
OA Version
Citation
Abstract
BACKGROUND: Current guidelines recommend using high-sensitivity troponin T (HST) to risk-stratify hemodynamically stable patients with acute pulmonary embolism (PE). N-terminal prohormone of brain natriuretic peptide (NT-proBNP) can also be used in the risk-stratification of acute PE patients, but there are no established cutoff values for either of these assays to identify PE patients at risk for clinical decompensation.
AIMS: To determine the optimal cutoff of HST and NT-proBNP for the risk-stratification of acute PE.
METHODS: We retrospectively analyzed data from a prospectively collected, single-center registry of patients evaluated by our Pulmonary Embolism Response Team (PERT). This registry is IRB approved with a waiver of informed consent. We included patients with a confirmed PE who had either an HST value (Elecsys Troponin T Gen 5 STAT, Roche Diagnostics) or NT-proBNP value (Elecsys proBNP II STAT, Roche Diagnostics) measured clinically. We used the highest HST and NT-proBNP value within 24 hours of PERT activation. We excluded patients with unstable vital signs on presentation. Our primary outcome was a composite of death, thrombolysis, thrombectomy, endotracheal intubation, extracorporeal membrane oxygenation (ECMO), or ICU admission, within 7 days. We calculated the area under the ROC curve for HST and NT-proBNP and established the optimal cutoff using the distance from (0,1). We performed subgroup analyses limited to patients with intermediate-risk PE.
RESULTS: We screened 815 PE patients from 10/2012 – 06/2021 with an NT-proBNP value measured, of whom 727 were hemodynamically stable. The high-sensitivity troponin assay was first implemented at MGH in April 2018; therefore, we screened 368 consecutive PE patients from 04/2018 – 06/2021, of whom 243 were hemodynamically stable. Of those, 234 (96%) had HST values measured. For both the HST and NT-proBNP populations, the mean age was 62 ± 17 years and 53% were male (Tables 8-9). Overall, the AUC for HST was 0.64 (95% CI, 0.56-0.71) with an optimal cutoff of 46 ng/L (Figure 6a), corresponding to sensitivity 59% (95% CI, 49%-69%) and specificity 61% (95% CI, 53%-69%). The AUC for NT-proBNP overall was 0.56 (95% CI, 0.51-0.61) with an optimal cutoff of 1092 pg/mL (Figure 7a), corresponding to sensitivity 53% (95% CI, 45%-61%) and specificity 59% (95% CI, 55%-63%). Among intermediate-risk PE, the AUC for HST was 0.60 (95% CI, 0.52-0.69) with an optimal cutoff of 46 ng/L (Figure 6b), sensitivity 62% (95% CI, 50%-73%) and specificity 53% (95% CI, 44%-63%); and the AUC for NT-proBNP was 0.56 (95% CI, 0.48-0.64) with an optimal cutoff of 1158 pg/mL (Figure 7b), sensitivity 57% (95% CI, 46%-68%) and specificity 56% (95% CI, 47-64%).
CONCLUSION: We identified an optimal cutoff for HST of 46 ng/L and for NT-proBNP of 1092 pg/mL, though the AUC suggests low to moderate performance for the risk stratification of hemodynamically stable PE.