Copeptin—Marker of Acute Myocardial Infarction

Acute chest pain is one of the most frequent chief complaints in internal emergency medicine. In the CHARITEM study, which analyzed data of 34,333 internal emergency patients of a German University hospital, 11.5 % (n = 3,954) of the patients presented with a leading symptom of chest pain. Of these, 47.5 % (n = 1,879) were admitted to the hospital, but only half of them where diagnosed with an ACS, and only 21.4 % (n = 402) had a final diagnosis of AMI. This means for the 3,954 patients presenting with chest pain, the frequency of AMI was as low as 10.2 % [4•]. In a study by Goodacre et al. [5] from the UK, 6 % (n = 6.957) of all emergency patients presented with chest pain, of which 11.0 % had ECG evidence of ACS and 34.5 % had clinically diagnosed ACS. In patients admitted to the hospital (n = 4,438), 16.1 % had ECG evidence of ACS and 48.6 % had clinically diagnosed ACS. In the USA, chest pain was the second leading reason (5.4 %) for emergency department (ED) visits in 2010. A principal hospital discharge diagnosis of ischemic heart disease was assigned to 2.9 % of all ED visits [6].

Even though not all chest pain patients will cause the treating physician to suspect ACS and even though ACS also needs to be suspected in many patients with symptoms other than chest pain, these numbers emphasize the clinical need for an early and fast rule-out of AMI. Under the current standard process of evaluating these patients, a minimum stay of between 3 and 9 h is mandatory. Given that the vast number of these stays are ultimately unnecessary, personnel and space resources are wasted in the ED and patients are left to wait until the possible diagnosis of a “heart attack” has dissolved. In times of increasing ED crowding, which has been shown to negatively impact patient outcome [7], this process is basically unacceptable.

The two hallmark articles on copeptin in patients with suspected ACS were published by Reichlin et al. [1] and Keller et al. [8] (Table 1).

Reichlin et al. [1] first hypothesized “that the combination of a marker of cardiac necrosis, such as troponin, with a pathophysiologically different biomarker reflecting acute endogenous stress, such as copeptin, might allow for a rapid and accurate rule out of AMI already at initial presentation without serial blood sampling.” To do so, they enrolled 492 patients with symptoms suggestive of AMI in an ED in Switzerland and collected blood samples at presentation and after 3 h and 6–9 h, which were tested for troponin T and copeptin. They showed that copeptin levels were significantly higher in patients with AMI than in patients with other diagnoses, including unstable angina. Copeptin levels were particularly high in AMI-patients who tested negative for troponin T values at presentation and in patients who presented within 0–4 h after the onset of symptoms. With use of cutoff values of 0.01 μg/l or less for troponin T and less than 14 pmol/l for copeptin, the combined use of these two markers at presentation yielded a negative predictive value (NPV) for the diagnosis of non-ST-segment-elevation myocardial infarction (NSTEMI) of 99.7 %.

Keller et al. [8] confirmed these findings in a chest pain unit (CPU) population of 1,386 patients with chest pain or equivalent symptoms, where the marker combination at the 99th percentile cutoff for sensitive troponin I and 9.8 pmol/l for copeptin showed an NPV of 99 % (95 % confidence interval 96.6–99.9 %). Additionally, they reported the results of serial blood sampling in a subgroup of patients who presented within 2 h after symptom onset to illustrate the complementary kinetics of troponin T, the concentration of which increased 6 h after admission, and copeptin, the concentration of which decreased during the first 6 h after a peak at presentation, in patients with AMI.

Since then, many articles have been published, partly confirming and partly contradicting these findings. To be able to correctly judge these publications, a number of factors have to be considered.

Keller et al. [8] evaluated different potential cutoff values for copeptin in a large reference population (n = 5,000), where the 99th percentile cutoff value was 18.9 pmol/l, the 97.5th percentile value was 13 pmol/l, and the 95th percentile value was9.8 pmol/l. Most clinical studies used a copeptin cutoff of 14 pmol/l. This cutoff was driven by the first copeptin assay available, which did not allow quantification of copeptin below this value [lower detection limit 4.8 pmol/l, functional assay sensitivity (lowest value with an interassay coefficient of variation below 20 %) below 12 pmol/l, limit of quantification 14.1 pmol/l] (manufacturer’s data). In 2011, an ultrasensitive copeptin assay was released, exhibiting a lower detection limit of less than 1 pmol/l, a functional assay sensitivity of less than 2 pmol/l, and a value for a coefficient of variation of 10 % of 2.5 pmol/l (manufacturer’s data). Mainly because 14 pmol/l was used in the first publications, this cutoff is still used in most studies. The lower the cutoff value though, the higher is the NPV for the diagnosis of AMI. To minimize the number of false-negative patients, a 95th percentile value (10 pmol/l) might be advisable. Table 2 shows a number of copeptin studies where different cutoff values were used, for comparison.

Although sensitivity and specificity are independent of the prevalence of the disease, both the positive predictive value and the NPV change with different disease prevalence. The NPV decreases with increasing prevalence, whereas the positive predictive value increases. In most rule-out studies on copeptin and troponin, the NPV is the primary measure of interest, determining success or failure of the new concept. Table 1 provides a list of studies with their respective AMI prevalence and the NPV calculated for this population.

Even though the data are difficult to compare for the multitude of factors influencing the study results, the data reflect a relatively low NPV in cohorts with an AMI prevalence above 20 % (Table 3).

For the same reason, a test can exhibit very different predictive values when it is administered to patients at different levels of risk (Table 4). Again, in a high-risk population the NPV is lower, whereas it is higher in a low-risk population. The marker combination of troponin and copeptin has been shown to achieve the best results in patients at low- to intermediate risk of AMI. Bohyn et al. [19••] tested a rule-out strategy using copeptin, troponin, and the Global Registry of Acute Coronary Events (GRACE) score. Here, both markers had to test negative and the GRACE score had to be below 108 points. With this concept, the NPV was 99 % (95 % confidence interval 94–100 %).

Two studies have shown that the pretest probability can also be determined by the judgment of the treating physician.

Chenevier-Gobeaux et al. [20••] divided their cohort into three pre-test-probability groups (low, medium, or high), as assessed by the treating ED physician after the first clinical evaluation but before the biomarker results were available. The NPVs were 100 (78-100)% in high-risk patients, 98 (87-100)% in medium-risk patients, and 100 (95-100)% in low-risk patients. In the CHOPIN study [22••], the treating ED physicians were asked to judge the likelihood of ACS and AMI on a visual analogue scale. The NPV for patients with intermediate risk was 99.6 % and for patients with low risk was 99.8 %;the NPV for the entire cohort was 99.2 %.

The concentration of copeptin is known to increase with the acute event and to then decrease rapidly to normal values within hours [8]. Thus, the time point of copeptin testing is crucial for this concept. In studies with most patients presenting late after the onset of symptoms, copeptin testing is unlikely to provide added value to troponin testing, as most patients will already be troponin-positive at admission to the ED or CPU [23•]. This was taken to an extreme in a study by Karakas et al. [24•] where copeptin was measured at a median of 4.3 h after presentation to the ED in a study which was primarily set up to evaluate CT angiography in patients with suspected ACS.

It is important to note that in late presenters the new strategy of a single copeptin–troponin measurement at presentation does not harm the patient, as the concentration of troponin will at this stage be increased in patients with NSTEMI and discharge of false-negative patients is thus unlikely.

Some studies have evaluated the diagnostic performance of copeptin and troponin for the diagnosis of ACS rather than the diagnosis of AMI [24•, 25•, 26•]. Reichlin et al. [1] clearly showed that the concentration of copeptin is not increased in patients with unstable angina.

The combined marker strategy also seems to perform better in cohorts with NSTEMI patients as compared with all AMIs [27•]. Nevertheless, given that the diagnosis in ST-segment-elevation myocardial infarction (STEMI) is based on ECG rather than biomarker results, STEMI patients have often been excluded when the diagnostic performance of these markers has been analyzed. Specifically, in unclear situations when patients present early, there is also a potential benefit of copeptin testing for patients who are finally categorized as having STEMI.

All the aforementioned studies are observational studies with retrospective copeptin measurement, where the copeptin value did not change patient care. Recently, the BIC-8 study, the first interventional, randomized process trial, evaluating the early-rule-out strategy in clinical practice, has been published [28••]. Low- to intermediate-risk patients with suspected ACS (n = 902) were randomized into either the standard group, receiving standard diagnostic workup and care, or the copeptin group, where further care depended on the copeptin value. In this group, copeptin-positive patients were considered higher risk and were admitted for standard workup, whereas copeptin-negative patients were considered low risk and were discharged to ambulant care, including a visit with a resident cardiologist within three working days. Importantly, the ultimate decision to discharge or admit a patient was left to the discretion of the treating physician on the basis on his/her clinical workup.

In this multicenter, international study, the major adverse cardiac event proportion at 30 days was not higher in the copeptin group (5.17 %) as compared with the standard group (5.19 %), suggesting safety comparable to that of the current standard process. Secondary end point analysis showed that patients in the copeptin group were discharged directly from the ED/CPU more often (67.6 % in the copeptin group vs 12 % in the standard group) and earlier (median length of stay for patients with AMI exclusion 4 h in the copeptin group and 7 h in the standard group), suggesting an effectiveness benefit of the new process.