Is There an Upper Limit to Cardiopulmonary Bypass Times?

Background: There are no safe operations in cardiac surgery. Every operation can possibly go wrong. We therefore retrospectively evaluated all cardiac operations lasting more than 300 minutes of bypass time at our institution to evaluate outcome and factors relevant for perioperative mortality and morbidity. Methods: We retrospectively included patients receiving cardiac operations or operations at the great vessels with cardiopulmonary bypass times above 300 minutes operated from 1/1/1996 until 12/1/2012 in our study. Patients receiving lung or heart or combined heart and lung transplantations were excluded from our study. 240 patients were included in our study. CPB times, clamp times and operation times were 356.53 ± 55.06 min, 166.18± 65.95 min, 500.47± 96.56 min respectively. Euro score of patients was 14.92 ± 15.35 (range 0.64-79.48). Results: Intraoperative and in-hospital mortality was 11.7% (n=28) and 32.9% (n=79), respectively. Overall mortality was 50.4%. Complication rates were high. Stroke, postoperative dialysis, re-thoracotomy rates were 11.1% (n=24), 35.9% (n=78) and 30.4% (n=66), respectively. Sex, age, infectious endocarditis, need for re-thoracotomy, CABG, aortic clamp times and postoperative dialysis predicted overall mortality in the multivariate analysis. CPB times and operation times were no independent predictors for overall mortality in this collective. In the patients collective excluding the intraoperative deaths, multivariate analysis revealed postoperative lactate levels, amylase levels, and intraoperative need for thrombcyte concentrates and ECMO support to be predictors of mortality. The introduction of reliable ECMO support (general availability starting in 2009) resulted in a signifi cant reduction of intraoperative mortality and overall mortality (p<0.001). Conclusion: Very long CPB times due to intraoperative encountered complications can occur at any given Euroscore. They are associated with a high mortality and morbidity, however even bypass times of over 500 minutes can be survived. The introduction of the ECMO reduced intraoperative mortality however, had no impact on in-hospital mortality. Research Article Is There an Upper Limit to Cardiopulmonary Bypass Times? Saad Rustum*, Felix Fleissner, Erik Beckmann, Fabio Ius, Mathias Wilhelmi, Serghei Cebotari, Axel Haverich and Issam Ismail Division of Cardiac, Thoracic, Transplantation and Vascular Surgery, Hannover Medical School, Hannover, Germany Dates: Received: 20 January, 2017; Accepted: 03 February, 2017; Published: 06 February, 2017 *Corresponding author: Saad Rustum, MD, CarlNeuberg Str.1, 30625 Hannover, Germany, Tel: +495115326589, Fax: +495115325404; E-mail: Felix Fleissner, Division of Cardiac, Thoracic, Transplantation and Vascular Surgery, Hannover Medical School, Hannover, Germany,


Introduction
Cardiac surgery is, despite all advances, still accompanied by a signifi cant mortality and morbidity [1]. If catastrophic events in cardiac surgery happen, eg., Bleeding or impossibility of weaning of CPB, very long bypass times can occur. Despite improvements in bypass safety, prolonged CPB times are associated with increased mortality and morbidity [2]. In the end, if there is no other choice, the cardiac surgeon has to rely on long bypass times to fi nish the operation. Hence we investigated the outcome of patients at our clinic who received cardiac surgery or surgery of the great vessels with a CPB time over 300 minutes. The goal was to identify prognostic markers for survival, identify perioperative morbidities and identify the upper limit to CPB times.

Methods
We retrospectively included adult patients receiving cardiac operations or operations of the great vessels with cardiopulmonary bypass times above 300 minutes operated from 1/1/1996 until 12/1/2012 in our study. Patients receiving lung or heart or combined heart and lung transplantations were excluded from our study. 240 patients were included in our study. Euroscore of patients was 14

Statistical analysis
Primary end-points were intra-operative mortality and overall mortality. Data analysis was performed using SPSS 22.0 (IBM, NY, USA) and the Categorical and continuous variables were summarized as percentages and means ± standard deviation (SD) or median with interquartile range (IQR), respectively. The independent-samples Student's t-test or the non-parametric Mann-Whitney U-test and the  2 or Fisher's exact tests were used for group comparisons of continuous and categorical variables, respectively. Two-tailed P-values ≤0.05 were considered signifi cant. Multivariate analysis, using a forward stepwise logistic regression model, was performed to identify independent risk factors for intraoperative mortality and overall mortality. Model calibration was evaluated using the Hosmer-Lemeshow (H-L) goodness-of-fi t test. A Cox regression analysis was performed to identify independent risk factors for mortality conditioned to hospital discharge. Results are reported as odds ratios (ORs) with a 95% confi dence interval (CI) and corresponding p-value.
In the patients collective excluding the intraoperative deaths, multivariate analysis revealed postoperative lactate levels, amylase levels, intraoperative need for thrombcyte concentrates and ECMO support to be predictors of mortality ( Establishment of reliable ECMO support may result in reduced intraoperative deaths. To further evaluate the impact of ECMO support on our patients collective, two groups were formed, Group 1 before establishment of a 24/7 availability of ECMO and Group 2 after the general availability of ECMO support, beginning after 2009. Between the years 1996 and 12/2008, 9 (14%) patients received ECMO support perioperativly, whereas after general availability after 2009, 22 patients (26.2%) (p<0.001) received ECMO support. The two groups were comparable in indications for surgery with a similar Euroscore (Table 7).

Discussion
Prolonged CPB times are associated with an increase in   mortality and morbidity [2]. Even in isolated aortic valve surgery and CPB times above 90 minutes excluded, CPB time remains a signifi cant factor for blood loss, ICU and hospital length of stay as well as in-hospital mortality [3]. The patient's collective in this study was predominantly sick, with high rates of acute aortic dissections, endocarditis and re-operations. These patients often require long CPB times since intraoperative problems are often encountered. However, even patients with a predicted low Euroscore sometimes require long CPB times due to unexpected intraoperative complications. At CPB times above 300 minutes, the CPB time does not longer predict the overall outcome. The longest CPB time survived in our cohort was above 500 minutes. This indicates that CPB time should not hinder the cardiac surgeon to continue surgery, since these patients requiring such excessive CPB times still have a chance of survival. The introduction of the ECMO and general availability lead to a signifi cant reduction in intraoperative mortality, however, it had no effect on overall in-hospital mortality. Venous arterial ECMO therapy enables support of unstable patients and critically ill patients in cardiogenic shock and allows temporary hemodynamic stabilization with improvement of end-organ function [4]. Most patients after long CPB times require hemodynamic stabilization as well as respiratory support due to oxygenatory failure; therefore a venous/venous ECMO is not the preferred option in these patients [5,6].
As expected, the patients with such complicated operations and therefore long CPB times are prone for complications [7,8].
The re-thoracotomy rate was above 30% (excluding patients with intraoperative death). Kidney function was affected in almost all patients with a high rate of requirement for dialysis postoperatively [9]. Euroscore failed to predict mortality in this group [6]. Nonetheless, the mortality in this patient collective well exceeded the predicted Euroscore due to a negative selection of only those patients with long CPB times. This might in fact be due to rather unexpected complications encountered such as the rare entity of iatrogenic aortic dissections and alike. CABG was an independent risk factor for intraoperative and overall death. This is most likely due to rescue attempts during cardiac  surgery, performing CABG as a bailout if the weaning from CBP is unsuccessful and myocardial ischemia has to be expected.
Re-sternotomy is usually due to technical issues regarding the operation or coagulopathy. In our study, longer CPB times were associated with an increased need for blood transfusions. This underlines the hypothesis, that the longer CPB times, the longer contact between foreign surfaces of the extracorporeal circuit resulting in increased severity of systemic infl ammatory response. However, CRP as a marker for systemic infl ammation was not increased with longer CPB times [10].
Longer cross clamp times are associated with a great risk of myocardial ischemia [2]. However, in our long CPB times, we were not able to identify CPB or cross clamp time as a signifi cant factor determining creatinine kinase muscle-brain isoenzyme (CK-MB), creatinine kinase (CK) or myoglobine release a surrogate marker for myocardial ischemia (p=0.573 CK-MB, p=0.112 CK, p=0.626 myoglobine). Perioperative myocardial infarction remains frequent in cardiac surgery with an incidence from 8% to 35% [11]. This wide variation is due to criteria and tests employed by different investigators institutions. However, the pathogene1perioperative myocardial infarction is not well understood. Inadequate myocardial protection during cardiopulmonary bypass (CPB), perioperative vasospasm, and atheroembolism are some of the current theories [12,13]. Increased duration of CPB and aortic occlusion are associated with a higher incidence of myocardial infarction [2].
As expected there were several early and independent markers to predict postoperative death. These were all markers of kidney and liver function such as lactate levels, amylase. Therefore, one can conclude that the damage to the organs is done already intraoperativly affecting the overall survival in our group.

Conclusion
Every cardiac surgery can possibly go wrong, even if the Euroscore predicts a low morbidity and mortality. If intraoperative encountered problems occur and long CPB times are needed to handle them, even very long CPB times can be survived, however the mortality and morbidity is high. The introduction of ECMO support had a great impact on intraoperative mortality, however the overall mortality seems to be unaffected by ECMO support in this collective.

Limitations
Limitations to this study is the retrospective data evaluation, the lack of a control group and the long time course over which operational techniques and CBP support have changed signifi cantly. However, a long term follow up would be possibly biased by the very heterogenous patient collective. We assume that within a longer time frame, the underlying disease of the patients would make a comparison of mid to long term survival very diffi cult.