Severe postoperative cardiovascular failure occurs in 3 to 5% of patients undergoing cardiac surgery using cardiopulmonary bypass (CPB) 1 2 . In some cases, a post cardiotomy vasodilatory shock characterized by a decreased vascular tone has been reported that might be explained by inadequate plasma concentrations of the vasoconstrictor hormone arginine vasopressin (AVP) 3 4 5 . Accordingly, first clinical reports showed beneficial effects of AVP treatment in cardiac surgery patients with severe vasodilatory shock 6 7 that correlates with results observed in septic shock where a deficiency in vasopressin has been described 8 9 . However, why some patients are prone to develop a deficiency in vasopressin after cardiac surgery, and whether postoperative vasodilatation without shock (vasoplegia) is associated with vasopressin deficiency, remain unclear.

Generation of copeptin correlates with the release of vasopressin, since both are derived from the same precursor. Furthermore, the stability of the precursor protein was found to be high and fully suitable for routine purposes. Copeptin is therefore easier to measure than vasopressin, and could be used as a marker of vasopressin release 5 10 . High copeptin concentration has been observed in chronic heart failure and acute myocardial infarction 11 12 . There are robust data showing that vasopressin is related to heart failure severity and outcome 13 . A prognosis-predictive value of copeptin has also been shown in critically ill patients 10 11 14 15 , in coronary artery disease 16 and in advanced heart failure 17 .

We hypothesized that patients who develop post cardiac surgery vasodilation (PCSV) may have an activation of the vasopressin system before surgery that favors a relative deficit in vasopressin after surgery. We have analyzed the kinetics of plasma concentrations of copeptin and AVP before, during and just after cardiac surgery in patients undergoing elective uncomplicated cardiac surgery, and compared patients with PCVS syndrome with the other patients.

Sixty-four patients were included within 2 months, of which 10 were treated with norepinephrine in the ICU. Preoperative data for the two groups are summarized in Table 1. Patients who experienced PCSV were more often in New York Heart Association functional class IV (50% vs. 7.4%, respectively; P = 0.01), had lower preoperative left ventricle ejection fraction and natremia, were more frequently treated with diuretics, and had a higher copeptin levels before CPB than controls. The area under the receiver-operator characteristic curve for preoperative copeptin concentration was 0.86 ± 0.04 (95% confidence interval = 0.78 to 0.94; P < 0.001) (Figure 1). The best predictive value for preoperative copeptin plasma concentration was 9.43 pmol/l with a sensitivity of 90% and a specificity of 77% (positive predictive value = 43%, negative predictive value = 98%).

Surgery procedure details are reported in Table 2. PCSV patients had longer CPB (191 minutes (150 to 280) vs. 103 minutes (87 to 159), P < 0.001) and clamping times (75.5 minutes (58 to 104) vs. 138.5 minutes (115 to 192), P < 0.001) than controls, but no significant hemodynamic differences during surgery and CPB. However, they more often needed inotropic support with moderate doses of dobutamine to achieve adequate cardiac output after CPB weaning (8/10 vs. 4/54, P < 0.001) (Table 2). The AVP plasma concentration increased significantly between T0 and CPB in controls (P < 0.016) but not in PCSV patients. Copeptin plasma levels did not increase significantly in either group between T0 and CPB but were significantly higher in the PCSV patients than in controls.

After surgery, PCSV patients were treated with a moderate dose of norepinephrine (range doses from 0.1 to 1.1 mg/hour). The PCSV patients have a longer extubation time (24 hours (16 to 48) vs. 6 hours (4 to 9.25), P < 0.001) and were more frequently treated with dobutamine (range doses from 0 to 10 μg/kg/minute) than other patients (Table 3). The total fluid volume load during surgery and in the first 8 postoperative hours was not significantly different between PCSV patients and other patients (3,250 ml (2,750 to 4,000) vs. 3,000 ml (2,500 to 4,000), respectively; P = 0.43). While PCSV patients remained intubated for a longer time than other patients, the propofol cumulative dose over the first 8 hours was not significantly different from other patients (350 mg (380 to 520) vs. 215 mg (155 to 375), respectively; P = 0.16). The AVP plasma concentration increase was statistically significant between T0 and H8 and between CPB and H8 in the control group (P < 0.001), and between T0 and H8 in the PCSV group (P = 0.01), but was significantly lower at H8 in PCSV patients than in controls (Table 3). The plasma copeptin concentration increased significantly between T0 and H8 and between CPB and H8 in both the control and PCSV groups (P < 0.001).

All patients were discharged from the hospital within a few days without serious complications.

The Spearman rank-order correlation test to assess the relation between AVP and copeptin plasma concentrations was performed on the base of 191 pairs. The correlation is significant (r = 0.76, P < 0.001) (Figure 2).

The present study shows that preoperative copeptin plasma concentration is predictive of PCSV. Increased preoperative copeptin could be a marker of vasopressin system stimulation before surgery, with a relative AVP system deficiency after surgery.

Our results are in agreement with previous publications where low AVP plasma concentrations have been observed in refractory vasodilatory shock when compared with cardiogenic shock shortly after CPB for cardiac surgery 6 . These results extend the likelihood of a vasopressin deficiency to vasodilatation without shock. We did not observe norepinephrine resistance in this small series of patients; however, the so-called catecholamine resistance varies widely between studies and between patients within studies 4 6 7 . As an example, in a series of 40 patients treated with vasopressin for vasodilatory shock the mean administered dose of norepinephrine was 14.4 ± 13.7 μg/minute 6 . The wide standard deviation suggests some patients were considered catecholamine resistant even when a rather small dose of norepinephrine was infused.

The AVP plasma concentration increase during cardiac surgery performed under CPB has been described for years 18 19 . The AVP peak has been related to arterial hypotension during CPB and to postoperative stimulation of AVP release by inflammatory mediators 20 21 . In the present study, AVP and copeptin plasma concentrations increase significantly in both groups. The AVP response to cardiac surgery resembles the postoperative course of other stress hormones, so higher AVP plasma concentrations would have been expected in patients with hypotension when compared with patients with normal blood pressure 6 22 . Although more patients in the norepinephrine-treated group were receiving dobutamine, which may have contributed to the vasodilatation, AVP levels are lower in these patients than expected in cases of significant hypotension (MAP < 60 mmHg) 6 . Similar inadequacy between plasma concentrations of AVP and low blood pressure has been documented in septic shock, where there is some evidence that depletion of endogenous AVP stores follows vigorous stimulation of AVP release from the neurohypophysis 7 8 23 .

A chronic stimulation of the vasopressin system that may contribute to AVP store depletion has been demonstrated in chronic heart failure 11 12 13 . High copeptin plasma levels have been observed in chronic heart failure and acute myocardial infarction; vasopressin production is related to the severity of heart failure 11 12 13 . Increased plasma levels of copeptin have been linked to excess mortality in chronic heart failure, and this link is observed irrespective of the clinical signs of severity of the disease. In the present study, PCSV patients had lower preoperative left ventricle ejection fraction and hyponatremia, both relevant markers of heart failure. Copeptin plasma concentration is higher in these patients than in controls; copeptin plasma levels are in the range of values already reported in New York Heart Association functional class III and IV patients 11 . Interestingly, the median preoperative copeptin value with a significant prognostic value (9.43 pmol/l) is double the median copeptin value of the normal population, as reported in the literature 12 , and also is similar to the control population of the study. The lower response of vasopressin after surgery while the preoperative copeptin concentration was increased may be thus accounted for by a depletion of endogenous AVP stores, as a result of sustained preoperative vasopressin system stimulation.

Although significant, the correlation between AVP and copeptin appears rather weak (r = 0.76), probably because direct determination of AVP is liable to generate inconsistencies (short half-life of AVP, 90% of which bound to blood platelets, and temperature sensitivity of the blood sample). These effects have hindered the use of AVP dosage in routine diagnostics. The correlation between AVP and copeptin plasma concentration observed in the study is nevertheless similar to that reported in other studies 5 14 . Generation of copeptin correlates with the release of vasopressin, since they are derived from the same precursor. Copeptin is a stable fragment of the AVP precursor pre-pro-vasopressin and is released together with AVP in a 1:1 stochiometric pattern 24 . The advantages of copeptin are its long stability and that it can be quickly and reliably measured in unprocessed plasma or serum. Copeptin may therefore be used as a surrogate parameter of vasopressin system stimulation 25 .

The small number of patients in the present study is a limitation that should be taken into account for interpreting negative results such as the absence of difference in copeptin plasma concentration between PCSV patients and the others after surgery during the AVP peak. The similar copeptin and AVP variations and a marked increase in copeptin after surgery in controls when compared with PCSV patients, however, are in agreement with AVP kinetics, and eventually AVP deficiency in PCVS patients. The plasma AVP measurement is known to be affected by platelet counts 5 . Nevertheless, the significant difference in AVP plasma concentration between the two groups after surgery (twofold higher in untreated patients than in PCSV patients) cannot be accounted for by the slight difference in platelet counts. Moreover, the increase in median copeptin plasma concentration after surgery, which is not influenced by platelet counts, is 10-fold higher in controls than in PCSV patients.

This preliminary study suggests that patients who experience vasodilatation after cardiac surgery may have a preoperative AVP system stimulation that results in a relative postoperative deficiency. Future trials are needed to confirm the possible relationship between the sustained preoperative AVP stimulation, as assessed by copeptin, and AVP store depletion.

• Copeptin plasma concentration is increased before CPB in patients experiencing PCSV.

• Copeptin is a marker of cardiac failure intensity, which may be useful to identify patients at risk of post cardiac surgery vasoplegia.

• Increased stimulation of the vasopressin system before surgery may contribute to vasopressin deficiency after surgery.

AVP: arginine vasopressin; CPB: cardiopulmonary bypass; H8: eighth postoperative hour; MAP: mean arterial pressure; PCSV: post cardiac surgery vasodilatation; T0: before cardiopulmonary bypass.

JS and NGM are employed by Thermo Fisher Scientific, the company that developed and patented the copeptin assay. The other authors declare that they have no competing interests.

PHC conceived of the study and helped to draft the manuscript. CB carried out the patient data collection and helped to draft the manuscript. JS and NGM performed the copeptin measurements and drafted the manuscript. BA participated in the design of the study and helped to draft the manuscript. GG carried out the immunoassays of vasopressin measurements. All authors read and approved the final manuscript for publication.