Survival of patients on ECMO has remained stable in every population. Laboratory values predictors of survival are required to improve patient care.Materials and methods
Clinical Looking Glass software was used to assess Electronic Medical Records (EMRs) of patients at Albert Einstein College of Medicine, Montefiore Medical Center (2007–2014).Results
Our population comprises of 166 adults and was divided in survivors and non-survivors, within 30 days. Indications for ECMO were cardiac (65%), respiratory (25%) and infectious diseases (< 10%). Eighty six patients (51.8%) survived the procedure. Gender, body weight, ejection fraction, diastolic blood pressure, and socio-economic status did not differ among survivors and non-survivors. In contrast, younger patients (45 yo vs 55 yo, p = 0.0001) and higher systolic blood pressure (115 mm Hg vs 103 mm Hg, p = 0.025) have favorable outcome. Univariate analysis shows that pre-cannulation values for creatinine (p = 0.0003), chloride (p = 0.009), bicarbonate (p = 0.015) and pH (p = 0.03) have prognostic value. Post-cannulation aPTT, pH, platelet and lymphocyte counts also have discriminative power. Notably, multiple logistic regressions for Multivariate Analysis identified chloride (OR 1.07; 95% CI 1.02–1.13; p = 0.004), pH (OR 3.35; 95% CI 1.89–5.9; p < 0.0001) and aPTT (OR 0.98; 95% CI 0.976–0.998; p = 0.024) as independent risk factors for 30-day mortality. These results imply that pre-existing renal conditions and hemostatic dysregulation contribute to poor outcome. Finally, patients on VV-ECMO have increase odds of survival (OR 1.88; 95% CI 1.06–3.34; p = 0.029).Conclusions
Laboratory markers identified herein may guide the management of patients on ECMO.