Introduction

Acute dyspnea is a frequent clinical presentation in the emergency department (ED). Cardiac and pulmonary disorders account for more than 75% of patients presenting with acute dyspnea to the ED [1, 2]. The identification of acute dyspneic patients at highest risk for death, particularly regarding short-term mortality remains a challenge. Patient history and physical examination remain the cornerstone of clinical evaluation [3], while disease specific scoring tools [4, 5] and biomarkers such as natriuretic peptides have been introduced to assist the clinician in the diagnostic and prognostic assessment [6–10].

The arginin-vasopressin system plays a crucial role in the regulation of the individual endogenous stress response [11]. Levels of arginin-vasopressin have been shown to be elevated in heart failure [12] and in different states of shock [13], but investigation of the vasopressin system was limited so far due to the fact that vasopressin is unstable (half-life 5 to 15 minutes) and largely attached to platelets [14, 15]. Copeptin, the c-terminal part of the vasopressin prohormone, is secreted stoichiometrically with vasopressin from the neurohypophysis and is much more stable, thus overcoming the limitations and difficulties assessing the arginin-vasopressin-system [16]. Recently, several studies investigated the prognostic role of Copeptin in various diseases [17–23], but little is known about the prognostic value of Copeptin in a typical ED population, for example, the patient group admitted with acute dyspnea. In clinical practice, the identification of dyspneic patients at highest risk for adverse outcome remains challenging. Therefore, we tested the prognostic value of Copeptin together with established markers such as BNP and NT-proBNP in an effort to better understand the role of Copeptin in this setting.

Materials and methods

Study population

The study population consisted of unselected patients presenting to the emergency department of the University Hospital of Basel, Switzerland, with a chief complaint of acute dyspnea. From April 2006 to March 2007, 292 patients (out of 327 patients screened) were prospectively enrolled. Exclusion criteria were age younger than 18 years, an obvious traumatic cause of dyspnea and patients on haemodialysis. 287 of the 292 patients had complete copeptin data at presentation and were considered as the study population. The study was carried out according to the principles of the Declaration of Helsinki and approved by the local ethics committee. Written informed consent was obtained from all participating patients.

Clinical evaluation and follow-up

Patients underwent an initial clinical assessment including clinical history, physical examination, electrocardiogram, pulse oximetry, blood tests including BNP, and chest X-ray. Echocardiography, pulmonary function tests and other diagnostic tests like CT-angiography were performed according to the treating physician. CT-angiography was the imaging modality of choice in patients with suspected pulmonary embolism. To assess the dyspnea severity we used the NYHA (New York Heart Association) classification with NYHA II as 'dyspnea while walking up a slight incline', III as 'dyspnea while walking on level ground' and IV as 'dyspnea at rest'.

Two independent internists blinded to Copeptin reviewed all medical records including BNP levels and independently classified the patient's primary diagnosis into seven categories: acute decompensated heart failure (ADHF), acute exacerbation of chronic obstructive pulmonary disease, pneumonia, acute complications of malignancy, acute pulmonary embolism, hyperventilation, and others. In the event of diagnostic disagreement among the internist reviewers, they were asked to meet to come to a common conclusion. In the event that they were unable to come to a common conclusion, a third-party internist adjudicator was asked to review the data and determine which diagnosis was the most accurate.

The endpoint of the present study was defined as 30-day all-cause mortality. Each patient was contacted for follow-up, via telephone, by a single trained researcher after 30 days. Regarding mortality data, referring physicians were contacted or the administrative databases of respective hometowns were reviewed, if necessary. Of note, one patient was lost to follow-up, so mortality analyses were performed in 286 patients.

Laboratory measurements

Blood samples for determination of Copeptin, BNP and NT-proBNP were collected at presentation into tubes containing potassium EDTA. After centrifugation, samples were frozen at -80°C until assayed in a blinded fashion in a single batch using a novel commercial sandwich immunoluminometric assay (B.R.A.H.M.S LUMItest CT-proAVP, BRAHMS AG, Hennigsdorf/Berlin, Germany) as described in detail elsewhere [16]. Since this initial publication, the assay was modified as follows: The capture antibody was replaced by a murine monoclonal antibody directed to amino acids 137 to 144 (GPAGAL) of pro-Arginin-Vasopressin. This modification improved the sensitivity of the assay. The lower detection limit was 0.4 pmol/L and the functional assay sensitivity (< 20% inter assay CV) was <1 pmol/L. Median Copeptin levels in 200 healthy individuals was 3.7 pmol/l and the 97.5th percentile was 16.4 pmol/L. NT-proBNP levels were determined in a blinded fashion by a quantitative electrochemiluminescence immunoassay with CVs claimed by the manufacturer were 1.8% to 2.7% and 2.35% to 3.2% for within-run and total imprecision, respectively (Elecsys proBNP, Roche Diagnostics AG, Zug, Switzerland) [24] and BNP was measured by a microparticle enzyme immunoassay at the hospital laboratory with a CVs claimed by the manufacturer of 4.3% to 6.3% and 6.5% to 9.4% for within-run and total imprecision, respectively. (AxSym, Abbott Laboratories, Abbott Park, IL, USA) [25].

Statistical analysis

Continuous variables are presented as mean ±SD or median (with interquartile range), and categorical variables as numbers and percentages. Univariate data on demographic and clinical features were compared by Mann-Whitney U-test or Fisher's exact test as appropriate. Correlations among continuous variables were assessed by the Spearman rank-correlation coefficient. Plasma levels of Copeptin, NT-proBNP and BNP were log-transformed to achieve a normal distribution. Thus, hazard ratios refer to a 10-fold rise in the levels of these markers. Receiver operating characteristic (ROC) curves were utilized to evaluate the accuracy of Copeptin, NT-proBNP and BNP to predict death. Areas under the curve (AUCs) were calculated for all markers. AUCs were compared according to the method by Hanley and McNeil [26]. We calculated the Net reclassification improvement (NRI) per Pencina et al.[27]. This index sums up the difference in proportions of patients that are reclassified to a higher or lower risk group predicted in terms of their actual outcome by adding Copeptin to the existing model with NT-proBNP and confounders. Cox regression analysis was assessed by univariate and multivariate analysis to identify independent predictors of outcome. In multivariate analysis, each of the three biomarkers (BNP, NT-proBNP and Copeptin) was included to the adjusted model including age, gender, a history of heart failure, glomerular filtration rate, diabetes and systolic blood pressure. In the next step, NT-proBNP or Copeptin was added to the adjusted model as dependent variable to examine the independent value of each biomarker. The Kaplan-Meier cumulative survival curves were compared by the log-rank test. Glomerular filtration rate was calculated using the abbreviated Modification of Diet in Renal Disease (MDRD) formula. Data were statistically analysed with SPSS 15.0 software (SPSS Inc., Chicago, IL, USA) and the MedCalc 9.3.9.0 package (MedCalc Software, Mariakerke, Belgium). All probabilities were two tailed and P < 0.05 was regarded as significant.

Results

Patient characteristics

The baseline characteristics of the 287 patients presenting with acute dyspnea are described in Table 1. Overall, mean age was 74 ± 12 years (median 77 years, interquartile range (IQR) 68 to 83 years), 52% were men and 80% were in NYHA functional class III and IV. The primary diagnosis was ADHF in 154 (54%) patients, acute exacerbation of chronic obstructive pulmonary disease in 57 (20%) patients, pneumonia in 32 (11%) patients, acute pulmonary embolism in 8 (3%) patients, acute complications of malignancy in 7 (2%) patients, hyperventilation in 5 (2%) patients, and other causes such as interstitial lung disease, asthma, or bronchitis in 24 (8%) patients. Differences between patients with ADHF versus patients without ADHF are depicted in Table 1.

Table 1 Baseline characteristics divided in patients with and without acute decompensated heart failure (ADHF)
Full size table

Copeptin levels and prognostic value of Copeptin on short-term outcome

The median Copeptin concentration was 21 pmol/l (IQR 8 to 52 pmol/l) in all patients. Concentrations of Copeptin were significantly higher in those dyspneic patients with ADHF versus those without (34 pmol/l, IQR 13 to 71 pmol/l vs.11 pmol/l, IQR 6 to 31 pmol/l; P < 0.0001). There was only a modest correlation between concentrations of log-transformed Copeptin and BNP (r = 0.42, P < 0.001) or NT-proBNP (r = 0.53, P < 0.001).

At 30 days, 29 patients (10.1%) had died. Non-survivors had significantly higher Copeptin levels than survivors (108 pmol/l, IQR 37 to 197 pmol/l) vs. 18 pmol/l, IQR 7 to 43 pmol/l; P < 0.0001). Among those subjects with ADHF, non-survivors (n = 21) had higher Copeptin levels than survivors (n = 133) (140 pmol/l, IQR 60 to 236 pmol/l vs. 28 pmol/l, IQR 11 to 55 pmol/l; P < 0.0001) and also in patients without ADHF the levels of Copeptin were higher in non-survivors (n = 8) than in survivors (n = 125) (61 pmol/l, IQR 18 to 112 pmol/l vs. 10 pmol/l, IQR 6 to 30 pmol/l; P = 0.007; Figure 1).

Figure 1
figure 1

Copeptin levels according to survival in patients with and without acute decompensated heart failure (ADHF).

Full size image

In Figure 2 the AUC to predict mortality are illustrated for Copeptin, NT-proBNP and BNP. The ROC analyses demonstrated an AUC of 0.83 (95% confidence interval (CI) 0.76 to 0.90) for Copeptin to predict 30-day mortality, with an optimal cut-point of 59 pmol/l. NT-proBNP had an AUC of 0.76 (95% CI 0.67 to 0.84) and BNP of 0.63 (95% CI 0.53 to 0.74) for 30-day mortality. Copeptin had a significantly higher AUC compared with BNP (P = 0.002) but not compared to NT-proBNP (P = 0.21). In patients with ADHF the AUC were 0.84 (95% CI 0.76 to 0.92) for Copeptin, 0.72 (95% CI 0.59 to 0.84) for NT-proBNP and 0.55 (95% CI 0.40 to 0.69) for BNP (Copeptin vs. NT-proBNP P = 0.098; Copeptin vs. BNP P < 0.001).

Figure 2
figure 2

Diagnostic accuracy for for Copeptin, NT-proBNP and BNP to predict 30-day mortality. AUC: Area under the receiver operating characteristic curve; NT-proBNP: N-terminal pro B-type natriuretic peptide; BNP: B-type natriuretic peptide; ADHF: acute decompensated heart failure

Full size image

Kaplan-Meier curves showed that patients in the highest quartile (Q4) had a significantly increased mortality compared with the other quartiles (Q1 to Q3) (Q1 = 1.2%, Q2 = 3.2%, Q3 = 6.8%, Q4 = 30%; P < 0.0001) (Figure 3). This pattern of increased mortality according to quartiles remained true for patients with ADHF but not for patients without ADHF (P < 0.0001 and P = 0.121, respectively). Patients in the highest quartile more often required admission to the hospital (99% vs. 83% in Q1 to Q3; P < 0.001), more often required admission to the intensive care unit (ICU) (16% vs. 6% Q1 to Q3, P = 0.011) and had higher in-hospital mortality (19% vs. 3% in Q1 to Q3, P < 0.0001).

Figure 3
figure 3

Kaplan-Meier curves demonstrating survival over time according to quartiles of Copeptin at baseline in all patients.

Full size image

Univariate Cox regression analysis showed that plasma levels of Copeptin, NT-proBNP, glomerular filtration rate and systolic blood pressure were predictors of 30-day mortality in all patients, regardless of whether they had ADHF or not. Copeptin was the strongest predictor of mortality in all patients (HR 5.28, 95% CI 3.08 to 9.06, P < 0.0001) and in patients with ADHF (HR 5.36, 95% CI 2.82 to 10.19, P < 0.001). BNP showed only a limited prognostic value in all patients and no value in the subgroups (Table 2).

Table 2 Univariate Cox regression analysis for 30-day mortality
Full size table

Table 3 shows the multivariate Cox regression analysis after adjustment for common cardiovascular risk factors (age, gender, history of heart failure, glomerular filtration rate, diabetes, systolic blood pressure). Copeptin remained the strongest independent predictor for 30-day mortality in all patients (HR 4.58, 95% CI 2.29 to 9.13; P < 0.001), especially in a patient with ADHF (HR 6.51, 95% CI 2.83 to 14.95; P < 0.0001). Even when NT-proBNP was included in this model, Copeptin kept its prognostic value in all patients (HR 3.88, 95% CI 1.94 to 7.77; P = 0.0001) and even more in patients with ADHF (HR 5.99, 95% CI 2.55 to 14.07; P < 0.0001). NT-proBNP was not significantly associated with mortality in patients with ADHF after including Copeptin in the model (HR 2.78, 95% CI 0.78 to 10.60; P = 0.11). The addition of Copeptin to the adjusted model including NT-proBNP resulted in reclassification of 37% of the patients. The NRI for events was 0.21 and the NRI for non-events was 0.16, achieving an NRI for the entire study cohort of 0.37 at 30 days (P < 0.001). Overall, 13 patients (30%) of those who experienced events were reclassified as high risk, and an additional 64 patients (26%) without events were reclassified as low risk.

Table 3 Multivariate Cox regression analysis for 30-day mortality
Full size table

Combined prognostic value of Copeptin and NT-proBNP

Mortality rates in all dyspneic patients as a function of Copeptin and NT-proBNP concentrations were examined and showed in Figure 4. Those with low Copeptin concentrations (lowest three quartiles) had lower rates of death in the first 30 days, irrespective of NT-proBNP concentration. Those patients with elevations (highest quartile) in both, Copeptin and NT-proBNP had the highest rates of death (15 out of 37 patients, 40.5%) at 30 days. Considering those patients with ADHF, similar to the group as a whole, the same relationship between Copeptin, NT-proBNP and outcome was observed. Those patients with low rates of Copeptin had low rates of death, irrespective of NT-proBNP levels at 30 days. In those patients with an elevation of both, Copeptin and NT-proBNP, the mortality rate was 42.9%, with 12 deaths out of 28 patients at 30 days.

Figure 4
figure 4

30-day mortality as a function of Copeptin and NT-proBNP concentrations in all patients (a) and in patients with ADHF (b). NT-proBNP: N-terminal pro B-type natriuretic peptide; ADHF: acute decompensated heart failure.

Full size image

Discussion

The early risk stratification of patients with acute dyspnea admitted to the ED is an unmet clinical need to improve the patient care in the first days of hospitalisation. Therefore, we investigated the role of Copeptin to predict short-term mortality in patients presenting with acute dyspnea to the ED. To the best of our knowledge, this is the first study analysing Copeptin plasma levels and their additive value to natriuretic peptides in a broader patient population like the typical ED population admitted with acute dyspnea.

We report five major findings. First, Copeptin levels were significantly higher in patients with ADHF than in patients with other diagnoses responsible for acute dyspnea. Second, Copeptin was significantly higher in non-survivors compared to survivors at 30 days, regardless of whether ADHF was present or not. Third, Copeptin is a new promising prognostic marker for short-term mortality independent of natriuretic peptide levels in patients with acute dyspnea and even more in patients with ADHF. Fourth, patients in the highest quartile of Copeptin levels had the highest mortality rate. Fifth, and of most importance, because both Copeptin and NT-proBNP were independent predictors of death, we also showed that the elevation of both markers was associated with the highest rates of death at 30 days in the entire patient cohort as well as in patients with ADHF. Furthermore, patients with low Copeptin levels had an excellent short-term prognosis even if NT-proBNP levels were high. These findings have major clinical implications. In clinical practice, the identification of dyspneic patients at highest risk for adverse outcomes remains difficult and largely depends on the underlying cause. Adding to this complexity is the fact that acute dyspnea is often due to multiple reasons like cardiac, pulmonary or inflammatory causes. Specific markers of the cardiovascular system like natriuretic peptides may therefore not be ideal to predict the outcome of patients with acute dyspnea in the ED.

Katan et al. showed in a small study that Copeptin is a good marker of the individual stress level comparing three groups of patients with increasing stress levels (healthy controls without apparent stress, hospitalized medical patients with moderate stress and surgical patients 30 minutes after extubation, with maximal stress) [28]. Taking this into account, it seems reasonable that patients with acute dyspnea and apparently in a stress situation could have higher Copeptin levels.

Recent studies by our group and others have suggested Copeptin and, therefore, the vasopressin system to be major determinants of outcome in patients where dyspnea is often the major symptom such as in acute myocardial infarction, in chronic heart failure and even in patients with community-acquired pneumonia and exacerbated chronic obstructive pulmonary disease [17–22, 29].

Copeptin release in acutely dyspneic patients is most likely related to three possible mechanisms. First, vasopressin release in heart failure is mainly driven by arterial underfilling caused by cardiac output failure, which activates the baroreceptors in the carotid sinus and the aortic arch. Volume overload and hyponatremia might also stimulate the release of vasopressin [30, 31]. Second, vasopressin has been shown to have vasoconstrictive effects, which may correlate to the hypoxia induced-vasoconstriction in severe chronic obstructive pulmonary disease [32, 33]. Increased concentrations of vasopressin may compensate for vasopressin (V1) receptor down-regulation following exposure to sustained hypoxemia [34]. Third, Copeptin is significantly increased in bacterial infection and febrile conditions [13, 35]. In the present study, the main causes of acute dyspnea were ADHF (54%), acute exacerbation of chronic obstructive pulmonary disease (20%) and pneumonia (11%). Therefore, the above mentioned mechanisms of Copeptin release reflect the broad spectrum of our acutely dyspneic patients and our findings corroborate that Copeptin may be a new promising candidate for the prediction of short-term mortality in this patient population. One could argue that Copeptin is another unspecific marker of inflammation. We, therefore, also compared the prognostic value of Copeptin with CRP and found that Copeptin was superior to CRP, an established inflammatory and also prognostic marker (AUC 0.83 (95% CI 0.76 to 0.90) vs. 0.71 (0.63 to 0.80), P = 0.04). Further studies should investigate whether Copeptin might help physicians tailor the therapy in view of the relative risk and allocate resources accordingly and whether this risk-stratification guided strategy might affect outcome. Additionally, it should be investigated whether serial copeptin testing further improves the risk stratification of acute dyspneic patients.

There are several limitations to our study. First, data derived from a single-centre study always need to be replicated in larger multicentre studies. However, our cohort is representative because patient characteristics are comparable to multicentre studies of acute dyspnea [1, 36]. Second, we assessed all-cause mortality because classification of death in clinical practice can sometimes be difficult and unreliable [37]. However, exact numbers of all different causes of death could have provided more interesting insights into the pathophysiological role of the biomarkers. And finally, the diagnosis of ADHF remains challenging even with the use of BNP and the diagnostic classification was possibly not 100% accurate. It is possible that some patients have had latent or mild heart failure, even though they were classified in the 'no ADHF'-group because other information (for example, a history of chronic obstructive pulmonary disease, pulmonary infiltrates on chest x-ray) may have suggested that the diagnosis was not ADHF. However, our classification was obtained by experienced physicians, suggesting that our results are valid.

Conclusions

Our study suggests that Copeptin alone or combined with NT-proBNP has a potential to assist clinicians in risk stratifying patients presenting with acute dyspnea regarding short-term mortality.

Key messages

  • In patients with acute dyspnea, Copeptin levels are elevated in non-survivors compared to survivors.

  • Copeptin is a new promising prognostic marker for short-term mortality independently of natriuretic peptide levels.

  • The elevation of both Copeptin and NT-proBNP was associated with the highest rates of death at 30 days.

  • Patients with low Copeptin levels had an excellent short-term prognosis even if NT-proBNP levels were high.