A 20-year multicentre outcome analysis of salvage mechanical circulatory support for refractory cardiogenic shock after cardiac surgery

Background Refractory post-cardiotomy cardiogenic shock (PCCS) is a relatively rare phenomenon that can lead to rapid multi-organ dysfunction syndrome and is almost invariably fatal without advanced mechanical circulatory support (AMCS), namely extra-corporeal membrane oxygenation (ECMO) or ventricular assist devices (VAD). In this multicentre observational study we retrospectively analyzed the outcomes of salvage venoarterial ECMO (VA ECMO) and VAD for refractory PCCS in the 3 adult cardiothoracic surgery centres in Scotland over a 20-year period. Methods The data was obtained through the Edinburgh, Glasgow and Aberdeen cardiac surgery databases. Our inclusion criteria included any adult patient from April 1995 to April 2015 who had received salvage VA ECMO or VAD for PCCS refractory to intra-aortic balloon pump (IABP) and maximal inotropic support following adult cardiac surgery. Results A total of 27 patients met the inclusion criteria. Age range was 34–83 years (median 51 years). There was a large male predominance (n = 23, 85 %). Overall 23 patients (85 %) received VA ECMO of which 14 (61 %) had central ECMO and 9 (39 %) had peripheral ECMO. Four patients (15 %) were treated with short-term VAD (BiVAD = 1, RVAD = 1 and LVAD = 2). The most common procedure-related complication was major haemorrhage (n = 10). Renal failure requiring renal replacement therapy (n = 7), fatal stroke (n = 5), septic shock (n = 2), and a pseudo-aneurysm at the femoral artery cannulation site (n = 1) were also observed. Overall survival to hospital discharge was 40.7 %. All survivors were NYHA class I-II at 12 months’ follow-up. Conclusion AMCS for refractory PCCS carries a survival benefit and achieves acceptable functional recovery despite a significant complication rate.


Background
Cardiogenic shock following cardiac surgery can affect as many as 2-6 % of patients undergoing routine surgical coronary revascularization or valve surgery [1][2][3][4]. Although the majority of these patients respond to inotropic support and/or intra-aortic balloon pump counter pulsation (IABP) support, 0.5-1.5 % of patients demonstrate a rapid and progressive decline in their haemodynamic parameters in the immediate aftermath of cardiopulmonary bypass [5]. The occurrence of post-cardiotomy cardiogenic shock (PCCS) can be unpredictable and can occur in patients with normal preoperative myocardial function as well as those with pre-existing impaired function [6]. Refractory PCCS leads to vital organ hypoperfusion and is almost universally fatal [4,[7][8][9] without the use of advanced mechanical circulatory support (AMCS) devices such as extracorporeal membrane oxygenation (ECMO) or ventricular assist devices (VAD).
In our previous study we looked at the outcomes of AMCS utilization at the Edinburgh heart center's cardiothoracic surgery department (a non-transplant, intermediate-sized, adult cardiothoracic surgery centre) in Scotland [10]. This current multicentre observational study aims to consolidate our previous findings and looks at the 20-year outcomes of AMCS utilization to salvage refractory PCCS patients in all the 3 cardiothoracic surgery centres in Scotland.

Methods
Scottish adult cardiothoracic surgical services are provided by three regional centres covering a population of 5.2 million individuals [11]. The relevant data was collected from the databases of the Royal Infirmary of Edinburgh (surgical case load ≈ 900/year), the Golden Jubilee National Hospital in Glasgow (surgical case load ≈ 1300/year), and the Aberdeen Royal Infirmary (surgical case load ≈ 500/year). Our inclusion criteria included any adult patient from April 1995 to April 2015 who had received salvage VA ECMO or VAD for PCCS refractory to IABP and inotropic support following adult cardiac surgery. We acquired information regarding the patients' 12 month follow-up status by accessing the cardiology follow-up clinic letters on the TrakCare R system in Edinburgh, the AMCS database in Glasgow, and through making direct enquiries with the surgeons involved in the long-term outcomes of the patients in Aberdeen via email and telephone communications.
The AMCS devices utilised at the Royal Infirmary of Edinburgh over the defined study period were Levitronix R CentriMag II for ECMO and Medtronic Bio-Medicus R 560 for short-term VAD support. Over the same time period, the AMCS devices used at the Golden Jubilee National Hospital in Glasgow and the Aberdeen Royal Infirmary cardiac surgical units was the CentriMag device for both VA ECMO and short-term VAD support.

Results
A total of 28 patients met the inclusion criteria with one patient excluded due to lack of recorded information in the TrakCare R database regarding the type of AMCS support used, any potential complications and the short and the long-term outcomes of this individual. Overall, 16 patients from the Royal Infirmary of Edinburgh met the inclusion criteria, 8 patients from the Golden Jubilee National Hospital in Glasgow and 3 patients from Aberdeen Royal Infirmary cardiothoracic surgery unit. The reason why more cases belonged to Edinburgh rather than Glasgow, despite the latter being a larger unit, was because AMCS was rarely used to salvage refractory PCCS patients in the west of Scotland prior to 2007 (the year of the merger between Glasgow Royal Infirmary and the Glasgow Western Infirmary forming the Golden Jubilee National Hospital).
The most common cause of death (COD) was refractory biventricular failure that failed to recover sufficiently to allow weaning from AMCS (22.2 %, Fig. 2). In these patients care was withdrawn. One patient died due to a combination of biventricular failure and haemorrhagic shock and another patient died from a combination of biventricular failure and septic shock whilst on VA ECMO. The survival rate to hospital discharge was 40.7 % (Fig. 3). The follow-up data showed that the survivors were all NYHA class I-II functional status at 12 months.

Statistical analysis
Statistical analysis was performed using the Fisher's exact and Pearson's chi 2 tests. Univariate analysis was performed. Table 2 demonstrates the baseline statistics data and the analytical methods used in this study.

Discussion
Our study demonstrates that AMCS used for the treatment of refractory PCCS can lead to good outcomes for a significant number of patients, with 40.7 % surviving to hospital discharge and all surviving patients were graded as either NYHA class I or II at 12 months' postdischarge. Without AMCS, it is likely that the vast majority of these patients would have died. Ours is also the first multi-centre study of its kind to emerge from the UK and one of the few studies to examine functional outcomes post AMCS utilisation for refractory PCCS.
Recent evidence has demonstrated that modern, continuous-flow AMCS devices, such as the CentriMag R that was used in our centres, can lead to improved survival in patients with PCCS [12][13][14]. In the largest cohort, Hernandez et al. [3] collated data from 5735 patients who underwent salvage VAD for refractory PCCS. They reported a 54.1 % survival rate to hospital discharge and concluded that VAD is a valuable, life-saving therapeutic manoeuvre. By comparison, the survival rate in our study was lower but firm conclusions are difficult given the low number of patients in our cohort. However, other smaller studies (relative to the Hernandez study) [5,[15][16][17][18] all using either ECMO or VAD for refractory PCCS, reported less impressive survival to hospital discharge rates of 24.8 %-37 % and a 5 year survival of 13.7 %-16.9 %. Unfortunately, we do not have long-term survival data as many of the survivors were ultimately discharged from the outpatient clinics when no further medical or surgical interventions were required, hence longer term follow up data post out-patient clinic discharge had not been recorded in the database.
We identified advanced age to be a factor leading to an adverse outcome, although again, owing to our smaller numbers, this did not reach statistical significance. Most (64 %) of the survivors were under 60 years of age. Furthermore, the emergent nature of surgery and pre-existing, preoperative severe left ventricular impairment were also identified as probable factors leading to an adverse outcome.
Evidence suggests that early device implantation [6] and appropriate patient selection through a multidisciplinary team approach is paramount to an optimal outcome [10]. There are no national or local protocols for identifying suitable patients for AMCS with refractory PCCS in Scotland: instead, decisions are based on a case-by-case assessment involving a multidisciplinary team (cardiac surgeon, department head, anaesthetist, and perfusionist) in each of the three hospital sites. We continue to believe that this is the best approach to patient selection rather than a standardised algorithmic approach because it ensures an ethically appropriate decision for the patient whilst optimising the cost-benefit equation. The decision regarding when to initiate AMCS support was made for most patients whilst in theatre in those whom weaning from CPB was not possible, although a few were commenced AMCS whilst in ICU. The time to AMCS and how this correlates to survival is an important variable that regrettably was not consistently recorded in our patient cohort.
AMCS devices are expensive [9,19,20] and this, coupled with a potentially prolonged length of stay in ICU, means that cost is an important factor in the decision-making process, particularly within the UK NHS. Indeed, decision-makers have opted to centralise AMCS funding to a restricted number of the larger cardiothoracic centres [21], invariably depriving other units of this potentially life-saving resource. Understandably, this has led to expressions of consternation [21]. In our cohort, the longest duration on AMCS was 33 days (patient 7). This patient was successfully weaned from VA ECMO but died whilst in critical care from a stroke, which may have been a complication from AMCS employment.
The NYHA functional outcomes for our patients were also very positive. Unfortunately, many previous AMCS studies for refractory PCCS do not report such findings, although we did identify two studies, each with similar outcomes to ours. Ko et al. [17] detailed a cohort of 76 patients undergoing ECMO support for refractory PCCS. They reported that all survivors were of NYHA classes I or II at 32 +/− 22 month follow-up. Pennington et al. [15] reported on refractory PCCS support with VAD and found that all survivors were "leading active lives". In 72.7 % of their survivors, ejection fraction had normalized on follow-up echocardiography.
Clearly, given that we only identified 27 patients undergoing AMCS over a 20-year period, and despite our pooled hospital case volume, we acknowledge that    the Scottish approach to institution of AMCS for refractory PCCS has been relatively conservative. This can partly be explained by the fact that salvage AMCS was not employed in the west of Scotland until 2007. Also, our general approach to institution of AMCS dictates that such modalities are instituted only if there is a reversible cause of the cardiogenic shock, which is reflected by our reasonable survival rate. Other possible reasons for underutilization may include: scarcity of resources, prohibitive costs, and lack of consistent evidence for the benefit of AMCS. The decision to institute AMCS must also be balanced with due consideration of the associated risks of this invasive modality, many of which are potentially life-threatening. Common device-related complications include: haemorrhage, thrombus formation and embolization, stroke, device-related infection, limb ischaemia, and multi-organ dysfunction syndrome/failure  [1,2,15,17,22,23]. In our cohort, the most common procedure-related complication was major haemorrhage. Renal failure requiring renal replacement therapy, stroke, and peripheral limb ischaemia also occurred with comparable rates to previous studies.
Given the scarcity of donor hearts in the UK, research continues to focus on implantable AMCS devices as a bridge to recovery, bridge to transplant, or as destination therapy [19]. However, none of our patients were transplanted during the study period and none had implantable long-term VADs.
Finally, this study is limited by the small number of subjects (as previously discussed) and its retrospective nature. It nevertheless reaffirms the findings of our Score not available 4 5

Conclusions
AMCS devices can be used to salvage a significant proportion of patients with refractory PCCS who would otherwise not survive. These patients are also likely to enjoy a reasonable quality of life. However, ACMS devices are associated with high rates of severe, systemic and device-related complications as well as being costly. Multidisciplinary teams experienced with patient selection and decisionmaking are imperative to help ensure appropriate use of AMCS and the best patient outcomes.