In this prospective observational cohort study, elevated copeptin measured preoperatively, immediately after surgery, and on POD 2–4 were all associated with 12-month mortality and morbidity. Elevated copeptin measured preoperatively and on POD 2–4 remained significantly associated when adjusting for other variables. Immediate postoperative copeptin was modestly associated with PMI, while preoperative copeptin was not. Other risk factors associated with MACCE and/or all-cause mortality were a higher RCRI, BMI, surgical risk, and preoperative hemoglobin. Body mass index, preoperative hemoglobin, and the need for continuous intraoperative norepinephrine were also associated with PMI.

Our study shows that there was an association between preoperatively elevated copeptin and MACCE and/or all-cause mortality at 12 months. Prior studies have shown that elevated preoperative copeptin values are associated with poor postoperative outcomes.9,21,22

In terms of mortality and MACCE, Jarai et al.21 examined 198 patients undergoing vascular surgery and concluded that a preoperative copeptin threshold of 14 pmol·L−1 was associated with a two-year incidence of MACCE with an HR of 2.8 (95% CI, 1.5 to 5.5) when accounting for type of surgery, history of myocardial infarction, preoperative cardiac troponin, and preoperative N-terminal pro-B-type natriuretic peptide (NT-proBNP). Investigating the incremental value of copeptin in addition to NT-proBNP, subgroup analyses revealed that especially patients at low estimated risk according to plasma NT–pro-BNP levels were at significantly higher risk for worse outcomes with higher copeptin levels (HR, 5.983; 95% CI, not reported). Schrimpf et al.22 examined 477 patients undergoing vascular surgery and found preoperative copeptin to be associated with 30-day MACCE (n = 41) defined as cardiac death (n = 1) or perioperative myocardial infarction, or elevation of cardiac troponin with measurements prompted by clinical suspicion (n = 46, subdivision unclear). The breakdown of MACCE supports an association with myocardial injury, but not necessarily with mortality. Preoperative copeptin concentrations were higher in patients with MACCE than in those without (median [range], 18.89 [≤ 4.8–180.7] vs 9.75 pmol·L−1 [≤ 4.8–321.6]; P < 0.001; 95% CI not reported). The area under the receiver operating characteristic of postoperative copeptin and the RCRI for predicting MACCE was 0.752 (95% CI not reported) (RCRI only, 0.714; DeLong P = 0.04; 95% CI not reported).

In terms of myocardial injury, an earlier study9 of 190 adult patients undergoing elevated risk noncardiac surgery (general, orthopedic, and vascular), showed that preoperative copeptin concentrations greater than 10 pmol·L−1 were associated with significantly higher rates of myocardial injury with an OR of 4.67 (95% CI, 2.06 to 11.19) when adjusted for age, sex, and the RCRI. Even when additionally accounting for (NT-proBNP), copeptin remained similarly predictive of myocardial injury. This study, which used a slightly higher cut-off (14 pmol·L−1 as employed in the Jarai study) in another population but largely at the same institution, and with a more stringent definition of ischemia, failed to find as association of preoperative copeptin with PMI.

In comparison with other studies, the patients in our study had no significant differences in the baseline characteristics. Unfortunately, the baseline characteristics in the other studies were not as detailed as in the current study, so a lot of factors could not be compared. For example, the mean age of participants in the studies was between 69 and 73 yr, most of the population were male, and the operations lengths were more or less the same. The study of Mauermann et al. included significantly more patients with a history of coronary heart disease and in the study of Schrimpf et al. the mean preoperative creatinine clearance was lower compared with the others. Our study was more heterogenous in terms of the types of surgery included.9,21,22

Taken together with the above studies, our study suggests that preoperative copeptin is indeed associated with midterm cardiac events in a wide variety of patients. Although only an explorative outcome, the lack of a difference in 30-day survival for elevated preoperative copeptin may be due to a limited number of events and/or may suggest that copeptin is a marker of underlying disease and disease severity rather than of postoperative risk. This interpretation may also explain why preoperative copeptin in this study was not associated with PMI (defined as a delta increase in this study, i.e., a dynamic increase from surgery), while previous studies had found preoperative copeptin to be associated with myocardial injury after noncardiac surgery (defined as an elevated postoperative troponin regardless of preoperative measurements, i.e., also including patients with troponinemia, i.e., a baseline state).

Unlike preoperative measurements, uncertainty remains regarding the value and timing of postoperative copeptin in predicting adverse events. Only the study by Schrimpf et al.22 examined postoperative copeptin concentrations (on POD 1) in vascular patients. They found copeptin concentrations to be higher on POD 1 than preoperatively (median [IQR], 23.55 [11.50–59.9] vs 10.16 pmol·L−1 [5.67–18.07], 95% CI not reported), which is also supported by our data. The results of this study also confirm that copeptin levels generally return to near preoperative levels by POD 2–4.

In terms of mortality and MACCE, this study also shows that implementing the preoperative copeptin cut-off of 14 pmol·L−1 on POD 2–4 may be justified. For simplicity, we used the preoperative cut-off in this study, which was very similar to our Youden’s optimum for copeptin on POD 2–4 and all-cause mortality and/or MACCE. In addition, while the association with MACCE and/or all-cause mortality at 12 months was similar between preoperative and postoperative measurements, only postoperative copeptin was associated with MACCE and/or all-cause mortality during the first 30 postoperative days. Although there were a limited number of events, this finding may also underscore that postoperative measurements dynamically reflect changes occurring during the perioperative period (e.g., blood loss, fluid administration, surgical time, use of vasoactive agents, transfusion, hypotension, osmolality,12,37,38 physiologic stress,16 etc.).

In terms of myocardial injury, we had previously measured copeptin and cardiac troponin at 0 hr, 2 hr, 4 hr, 6 hr, 8 hr, and on POD 1 and 3 in a small pilot study.20 This pilot study suggested copeptin immediately after surgery to be most promising for identifying myocardial injury. As suggested by the pilot, we found a statistically significantly higher OR for PMI in immediate postoperative copeptin, but not for preoperative copeptin. Nevertheless, the magnitude of effect was modest and based on an internally derived cut-off, which may have led to overfitting.

This study clearly shows that copeptin measured preoperatively, immediately postoperatively and on POD 2–4 may be used to risk stratify a broad group of patients for MACCE-free survival at 12 months. This association remained robust when adjusting for a number of variables, particularly for preoperative copeptin and copeptin on POD 2–4. Interestingly, postoperative measurements were associated with 30-day MACCE-free survival, suggesting that these may better reflect perioperative factors, while preoperative copeptin may better mirror baseline risk. Additionally, we showed that the preoperative cut-off of 14 pmol·L−1 may be justified for postoperative measurements on POD 2–4.

No robust association between perioperative copeptin and PMI was found, suggesting that PMI may be a different rather than a surrogate endpoint for all-cause mortality and/or MACCE. The clinical utility of immediate postoperative copeptin for predicting myocardial injury was limited in this study.

While our study had several strengths (size, mixed population, high degree of completeness, hard outcomes), it also has some important limitations.

First, despite being a prospective study with the largest sample and event sizes examining copeptin, data were only collected in two centres and extrapolation to other settings may be difficult. As a result, findings should be extrapolated to other patient populations with caution and further studies are needed to confirm our findings.

Second, some other useful indicators of surgical stress (such as C-reactive protein or the neutrophil-lymphocyte-ratio) as well as some other biomarkers of cardiac stress (NT-proBNP) were not uniformly collected. Nevertheless, our intention was to evaluate the prognostic value of copeptin in addition to routinely measured clinical values.

Third, our model was powered to adjust for the RCRI and up to two other variables (models 1A, 2A, and 3A), which seemed reasonable given the clinical value of the RCRI as in line with previous copeptin publications. Nonetheless, the association was robust even when adding a number of additional variables (models 1B, 2B, and 3B).

Fourth, the timing of postoperative copeptin in this study (POD 2–4) was broadly defined as we wanted to attain the last reasonable measurement prior to discharge rather than define a specific timepoint.

The results of the present prospective observational cohort study suggest that perioperative copeptin concentrations can help identify patients at risk for all-cause mortality and/or MACCE. Other identified risk factors were revised cardiac risk index, body mass index, surgical risk, and preoperative hemoglobin.