Morbidity and Mortality Analysis of Patients with Mechanical Circulatory Support or Replacement Systems VA-ECMO, Impella CP and 'ECMELLA': A 9-Year Analysis at the University hospital Kiel

Background  
Despite advances in intensive care, cardiogenic shock remains highly lethal. Mechanical circulatory support (MCS) systems such as VA-ECMO, Impella, and their combination (ECMELLA) are increasingly used in therapy-refractory circulatory failure. Given limited resources and varying patient profiles, selecting the appropriate device is crucial. 

Methods  
This retrospective study analyzed 307 patients (see table I: Baseline data) treated with VA-ECMO, Impella, or ECMELLA at the University Medical Center Schleswig-Holstein, Kiel, between 2015 and 2024. Data included demographics, comorbidities, laboratory values, resuscitation status, and MCS type. The primary endpoint was 30-day mortality; long-term survival was assessed via official registry inquiries. 

Results  
Within 30-days, 243 patients (77.9%) died. The highest 30-day mortality was seen in the ECMELLA group (84.1%), followed by VA-ECMO (82.5%) and Impella (73.3%). 69 patients survived beyond 30 days, 44 (14.3%) were alive at last follow-up on October 31, 2024. Median survival time across all groups was 1 day, highlighting extreme early mortality. Mean survival time was longest with Impella (945.9 days), followed by VA-ECMO (439.6 days) and ECMELLA (218.6 days). Kaplan-Meier analysis showed steep early mortality in VA-ECMO and ECMELLA groups. In pairwise comparison there was no survival advantage for any system we used, e.g. Impella vs VA-ECMO (p = 0.05) or vs ECMELLA (p = 0,05), nor VA-ECMO vs. ECMELLA (p = 0.59) (see figure 1: Survival Analysis ). Age was the strongest independent predictor of mortality, with each additional year increasing the odds of 30-day death by 2.1% (see figure 2: Logistic Regression age and mortality). In both Impella and VA-ECMO patients, elevated lactate, low pH, and higher age were significantly associated with mortality (p < 0.05). Each 1 mmol/L rise in lactate increased the odds of death by 15%, while a 0.1 unit drop in pH markedly elevated mortality risk (see figure 3: Logistic Regression lactate and mortality and figure 4: Logistic Regression ph and mortality). Troponin and other markers (NT-proBNP, albumin, bilirubin, creatinine) showed no consistent or statistically significant associations. 

Conclusion
This long-term study registered all patients treated with MCS at the university hospital Kiel, department of cardiology from 2015-2024. Due to access to the residents’ registration office, we were able to obtain mortality beyond 30-days with long term follow up. First, our data shows that demographic data, such as age and laboratory parameters including lactate and pH are robust predictors of mortality in patients with MCS therapy. Second, those parameters were predictive in 30-day mortality and even beyond. Age emerged as the strongest predictor of mortality. Overall median survival was 1 day, indicating a disproportional high mortality which warrants further attention. Third, LV venting was not associated with lower mortality.


Table1: Baseline characteristics 

1 Mean (SD); n / N (%) 
2 Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test  

Figure 1: Survival Analysis 

Figure 2: Logistic Regression age and mortality 
Figure 3: Logistic Regression lactate and mortality

Figure 4: Logistic Regression ph and mortality