Combined Testing of High-Sensitivity Troponin T and Copeptin on Presentation at Prespecified Cutoffs Improves Rapid Rule-Out of Non–ST-Segment Elevation Myocardial Infarction

BACKGROUND

Two recent clinical trials showed that adding copeptin to a conventional cardiac troponin assay improved diagnostic performance for patients with chest pain early after symptom onset. We prospectively tested whether copeptin adds information to that provided by a high-sensitivity cardiac troponin T (hscTnT) assay in the early evaluation of patients with suspected acute myocardial infarction, particularly non–ST-segment elevation myocardial infarction (non-STEMI).

METHODS

We enrolled 503 patients with suspected acute coronary syndrome and onset of chest pain occurring within the previous 12 h. Copeptin was measured on presentation, and hscTnT was measured serially at baseline and after 3 and 6 h. We used ROC curve analysis and likelihood ratio χ² statistics for nested models. Diagnostic sensitivities, specificities, positive predictive values (PPVs), and negative predictive values (NPVs) were calculated for admission values of copeptin alone, hscTnT alone, and the combination of both markers.

RESULTS

For ruling out non-STEMI (after excluding STEMI), an hscTnT concentration <14 ng/L (99th percentile) plus a copeptin concentration <14 pmol/L yielded a diagnostic sensitivity of 97.7% (95% CI, 91.9%–99.7%), an NPV of 99.03% (95% CI, 96.6%–99.9%), a diagnostic specificity of 55.9% (95% CI, 50.6%–61.0%), and a PPV of 34.4% (95% CI, 28.5%–40.7%). ROC curve analysis of the continuous biomarker values on admission demonstrated no added value of using this marker combination for ruling out non-STEMI when hscTnT was used as the standard for diagnosing non-STEMI.

CONCLUSIONS

A strategy using copeptin with hscTnT at prespecified cutoffs improves the ruling out of non-STEMI, compared with using hscTnT alone; thus, this strategy could help to obviate a prolonged stay in the emergency department.

Current cardiac troponin assays are limited by both analytical and diagnostic sensitivity problems at presentation. The ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction therefore recommends measurement of cardiac troponin upon presentation and after 6–9 h (1). Until now, biomarkers claiming to detect myocardial ischemia earlier have not demonstrated an added benefit, compared with the measurement of cardiac troponin alone (2). Recently, 2 clinical trials showed that adding copeptin to conventional cardiac troponin assays improved the diagnostic performance for patients with chest pain, particularly early after the onset of symptoms (3, 4). A negative result for both tests provided a remarkable negative predictive value (NPV)² that was helpful for rapidly ruling out acute myocardial infarction (AMI). Recently, high-sensitivity cardiac troponin assays were shown to be superior to conventional assays for earlier and more accurate diagnosis of AMI (5, 6); however, ST-segment elevation myocardial infarction (STEMI) is not a condition for which biomarker testing is recommended. In this investigation, therefore, we tested whether copeptin improved the early rule-out of non-STEMI when the biochemical standard of AMI diagnosis is a high-sensitivity cardiac troponin T (hscTnT) assay.

From November 16, 2009, to July 22, 2010, we enrolled patients who presented to the Chest Pain Unit of the University of Heidelberg with onset of chest pain within the previous 12 h. Patients who had trauma or major surgery within the previous 4 months, pregnancy, anemia, terminal kidney failure, an age <18 years, or lack of consent were excluded. AMI was diagnosed according to the criteria of the universal definition (1). Non-STEMI was defined as an hscTnT value ≥99th percentile that showed an increase and/or decrease of ≥20% within 6 h after admission. Unstable angina was diagnosed if acute coronary syndrome was suspected clinically along with an hscTnT value <99th percentile occurring during serial sampling, with or without electrocardiogram (ECG) changes. Patients who had an hscTnT value <99th percentile on presentation and who underwent percutaneous coronary intervention before collection of the next blood sample were excluded. All patients without a diagnosis of acute coronary syndrome were classified as having noncardiac chest pain. To avoid overdiagnosis of AMI, we had 2 cardiologists readjudicate the diagnoses.

Troponin was measured with the hscTnT assay (Roche Diagnostics), which is commercially available in Germany but is not yet available in the US. The limit of blank (3 ng/L) and the limit of detection (5 ng/L) were determined in accordance with CLSI guideline EP17-A. The interassay CV was 8% at 10 ng/L and 2.5% at 100 ng/L. The intraassay CV was 5% at 10 ng/L and 1% at 100 ng/L (7). The BRAHMS Copeptin KRYPTOR assay is an automated immunofluorescence assay that measures copeptin, the C-terminal fragment of the arginine vasopressin precursor (8). The limit of detection and the limit of quantification, which is defined as the lowest concentration of analyte that can be quantitatively measured with suitable imprecision and accuracy, have been determined to be 4.8 pmol/L and 14.1 pmol/L, respectively, in accordance with CLSI guideline EP17-A.

Logistic regression was used to combine hscTnT and copeptin concentrations for diagnosing AMI and for diagnosing non-STEMI after excluding STEMI. The predictive value of each model was assessed by the likelihood ratio χ² statistic for nested models. ROC curves were constructed to assess the sensitivity and specificity of both measurements.

Positive predictive values (PPVs), NPVs, diagnostic sensitivity, and diagnostic specificity, as well as the odds ratio for the target marker combination, were assessed by first applying a marker-specific cutoff value, both for AMI and for non-STEMI after the exclusion of STEMI patients. Cutoffs were prespecified at 14 ng/L (99th percentile) for hscTnT and at 14 pmol/L (limit of quantification) for copeptin.

We enrolled 503 patients [mean (SD) age, 63 (16) years; female patients, 37%; median GRACE (Global Registry of Acute Coronary Events) score, 92 points (range, 7–214 points; 58 cases (11.5%) with a GRACE score >140 points]. The time of presentation after the onset of symptoms was 0–3 h for 229 patients (45.5%), 3–6 h for 98 patients (19.5%), 6–12 h for 87 patients (17.3%), and >12 h for 76 patients (15.1%). Information on symptom onset was missing for 13 patients (2.6%). The symptoms leading to admission included chest pain in 89.3% of patients (n=449), diffuse abdominothoracic symptoms (n=41), dyspnea (n=9), epigastric pain (n=2), and other symptoms (n=2).

Upon presentation, 78 of the patients (15.5%) had ST-segment elevations, 52 patients (10.5%) had ST-segment depression, and 303 patients (60.2%) had a normal ECG.

At discharge, AMI was diagnosed in 136 patients—49 patients (9.7%) with a STEMI and 87 patients (17.3%) with a non-STEMI. A diagnosis of unstable angina or noncardiac chest pain was made in 390 cases. Coronary angiography was performed in 250 patients (49.7%), and 135 patients (54%) underwent a percutaneous coronary intervention.

The median (25th–75th percentile) values for copeptin and hscTnT concentrations at presentation were 8 pmol/L (2.5–22.1 pmol/L) and 13 ng/L (4.3–43.5 ng/L), respectively.

Logistic regression analysis demonstrated a significant association (P < 0.0001) of both markers with AMI and an additional value of copeptin with hscTnT (added χ² =12.3; P =0.0004). An ROC curve analysis of the continuous biomarker values demonstrated a modest performance for copeptin alone [area under the ROC curve (AUC), 0.697; 95% CI, 0.655–0.737)], excellent performance for hscTnT (AUC, 0.903; 95% CI, 0.874–0.927), and a small but significantly higher AUC (P =0.0016) for the combination of hscTnT and copeptin (AUC, 0.917; 95% CI, 0.889–0.940) compared with hscTnT alone. For patients presenting within 6 h after chest pain onset (n=340; AMI, n=87), copeptin alone performed better than in the overall study cohort (AUC, 0.717; 95% CI, 0.666–0.764), and hscTnT alone performed worse (AUC, 0.872; 95% CI, 0.832–0.906). Conversely, for patients who presented ≥6 h after the onset of symptoms (n=163; AMI, n=49), the performance for hscTnT alone was excellent (AUC, 0.953; 95% CI, 0.908–0.980), whereas the AUC for copeptin alone was 0.688 (95% CI, 0.611–0.758). Given that STEMI is not a condition for which biomarker testing is recommended, we repeated the analysis for non-STEMI patients after excluding the STEMI patients. This analysis demonstrated no added benefit of using copeptin in combination with hscTnT, compared with hscTnT alone (data not shown).

When we applied the prespecified cutoffs, 261 patients had hscTnT concentrations <14 ng/L. Of these patients, 208 consistently showed copeptin concentrations <14 pmol/L. Two patients (copeptin concentration of 11 pmol/L in both cases) with non-STEMI had a normal ECG but showed substantial increases in the hscTnT concentration after 3 h (11 to 24 ng/L and 12 to 26 ng/L). The diagnosis was confirmed by coronary angiography, and both of these patients underwent a percutaneous coronary intervention.

Diagnostic sensitivities, diagnostic specificities, NPVs, and PPVs for the use of baseline hscTnT values <14 ng/L alone to rule out AMI were 91.9% (95% CI, 86.0%–95.9%), 68.1% (95% CI, 63.1%–72.9%), 95.8% (95% CI, 92.6%–97.9%), and 51.7% [95% CI, 45.2%–58.1%; odds ratio, 24.3 (95% CI, 12.6–46.7)], respectively. A combination of an hscTnT value <14 ng/L and a copeptin value <14 pmol/L correctly ruled out AMI with a diagnostic sensitivity of 97.8% (95% CI, 93.7%–99.5%), an NPV of 98.6% (95% CI, 95.8%–99.7%), a diagnostic specificity of 55.9% (95% CI, 50.6%–61.0%), and a PPV of 45.1% [95% CI, 39.3%–51.0%; odds ratio, 56.1 (95% CI, 17.5–179.4)].

After the exclusion of STEMI patients, a combination of an hscTnT value <14 ng/L and a copeptin value <14 pmol/L correctly ruled out non-STEMI with a diagnostic sensitivity of 97.7% (95% CI, 91.9%–99.7%), an NPV of 99.03% (95% CI, 96.6%–99.9%), a diagnostic specificity of 55.9% (95% CI, 50.6%–61.0%), and a PPV of 34.4% [95% CI, 28.5%–40.7%; odds ratio, 53.8 (95% CI, 13.0–221.9)]. On the basis of these cutoffs, 207 patients (41.1% of the total population) could potentially be ruled out for AMI at baseline.

Our study confirms that the use of the combination of copeptin and hscTnT at prespecified cutoffs (i.e., hscTnT at the 99th percentile and copeptin at <14 pmol/L) improves the rule-out of non-STEMI at presentation, with an NPV of 99.03%. In contrast, using a combination of biomarkers as continuous values did not add to the excellent NPV of hscTnT alone for ruling out non-STEMI.

Our findings confirm the results for 487 consecutive patients obtained by Reichlin et al. (3), who found that the diagnostic performance of a combination of the cTnT concentration obtained with a conventional fourth-generation assay plus the copeptin concentration at initial presentation was significantly better for ruling out MI than cTnT alone (AUC, 0.97 vs 0.86; P < 0.001), with an NPV of 99.7%. Consistent with these results are those for 1386 patients reported by Keller et al. (4), who found that the combination of the copeptin concentration and the cTnT concentration at admission obtained with a fourth-generation assay improved the c-statistic from 0.84 for cTnT alone to 0.93 in the overall population (P < 0.001), with an NPV of 94.8%. Our findings add the important information that copeptin improves the rapid rule-out of non-STEMI, even if the hscTnT assay is used as the diagnostic standard of MI. When STEMI diagnoses are excluded, however, an added benefit for the rapid rule-out of non-STEMI is seen only when the copeptin and hscTnT concentrations are used at prespecified cutoffs, not as continuous values. Because the universal definition of MI recommends that cardiac troponin be measured at presentation and after 6–9 h (1), a strategy of combining the diagnostically more sensitive cardiac troponin with copeptin might obviate a prolonged stay in the emergency department for up to 40% of all patients. Currently, this strategy is being tested in a multicenter trial for patients who present with the onset of chest pain within 12 h and have an hscTnT value <99th percentile.

This study has several limitations. First, we cannot provide information on kinetic changes in copeptin, nor can we provide data on the impact of hscTnT or copeptin results on outcomes. The performance of copeptin alone and in combination with hscTnT could have been improved by using the 95th percentile cutoff (9.8 pmol/L), which would yield an NPV of 100%; however, the imprecision of the presently available assay prohibits the use of a cutoff <14 pmol/L because the limit of quantification of this copeptin assay is 14.1 pmol/L. Second, biomarker testing is not recommended for the STEMI population because the diagnosis is made by ECG. Whereas others (4, 5) evaluated the use of combined biomarker testing for the diagnosis or rapid rule-out of MI, we tested performance for the entire AMI cohort and for non-STEMI patients after excluding STEMI patients. Third, the hscTnT assay formed part of the reference method for AMI diagnosis. Therefore, the diagnostic sensitivities and specificities are subject to an incorporation bias that overestimates the hscTnT concentration and underestimates the copeptin concentration. Incorporation bias cannot be eliminated in this kind of study, but it may be reduced by using different troponin assays (3–6). Using an analytically less sensitive assay may cause underdiagnosis and undertreatment, however. Strategies for limiting the overestimation of clinical specificity include strict adherence to the clinical criteria and consideration of the kinetic changes recommended by the universal infarct definition.

2 Nonstandard abbreviations:

 
• NPV

  negative predictive value

 
• AMI

  acute myocardial infarction

 
• STEMI

  ST-segment elevation myocardial infarction

 
• hscTnT

  high-sensitivity cardiac troponin T (assay)

 
• ECG

  electrocardiogram

 
• PPV

  positive predictive value

 
• GRACE

  Global Registry of Acute Coronary Events

 
• AUC

  area under the ROC curve

Author Contributions:All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures or Potential Conflicts of Interest:Upon manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest:

Employment or Leadership: None declared.

Consultant or Advisory Role: E. Giannitsis, Roche Diagnostics, AstraZeneca, Bayer Schering, and BRAHMS Biomarkers; H.A. Katus, AstraZeneca and Bayer Vital.

Stock Ownership: None declared.

Honoraria: E. Giannitsis, Roche Diagnostics, AstraZeneca, Siemens Healthcare, BRAHMS Biomarkers, and Mitsubishi Chemicals; H.A. Katus, Roche Diagnostics, AstraZeneca, and Bayer Vital.

Research Funding: E. Giannitsis, Roche Diagnostics, Mitsubishi Chemicals, Siemens Healthcare, and BRAHMS Biomarkers (Clinical Diagnostics Division, Thermo Fisher Scientific); H.A. Katus, Roche Diagnostics, Daiichi-Sankyo, and AstraZeneca.

Expert Testimony: None declared.

Other Remuneration: H.A. Katus developed the cTnT assay and holds a patent jointly with Roche Diagnostics.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.

Acknowledgments

We thank Oliver Hartmann from BRAHMS Biomarkers for review and provision of statistical assistance. Clinical trial NCT00953251 (TnThs Predicting Evolving Non-STEMI).