RESUMEN
Objectives: The purpose of this study was to develop and validate a machine learning prediction model for the prognosis of non-cardiogenic out-of-hospital cardiac arrest (OHCA) with an initial non-shockable rhythm. Design: Data were obtained from a nationwide OHCA registry in Japan. Overall, 222,056 patients with OHCA and an initial non-shockable rhythm were identified from the registry in 2016 and 2017. Patients aged <18 years and OHCA caused by cardiogenic origin, cancer, and external factors were excluded. Finally, 58,854 participants were included. Methods: Patients were classified into the training dataset (n=29,304, data from 2016) and the test dataset (n=29,550, data from 2017). The training dataset was used to train and develop the machine learning model, and the test dataset was used for internal validation. We selected XGBoost as the machine learning classifier. The primary outcome was the poor prognosis defined as cerebral performance category of 3-5 at 1 month. Eleven prehospital variables were selected as outcome predictors. Results: In validation, the machine learning model predicted the primary outcome with an accuracy of 90.8% [95% confidence interval (CI): 90.5-91.2], a sensitivity of 91.4% [CI: 90.7-91.4], a specificity of 74.1% [CI: 69.2-78.6], and an area under the receiver operating characteristic value of 0.89 [0.87-0.92]. The important features for model development were the prehospital return of spontaneous circulation, prehospital adrenaline administration, and initial electrical rhythm. Conclusions: We developed a favorable machine learning model to predict the prognosis of non-cardiogenic OHCA with an initial non-shockable rhythm in the early stage of resuscitation.
RESUMEN
BACKGROUND: Cardiac surgery is performed worldwide, and acute kidney injury (AKI) following cardiac surgery is a risk factor for mortality. However, the optimal blood pressure target to prevent AKI after cardiac surgery remains unclear. We aimed to investigate whether relative hypotension and other hemodynamic parameters after cardiac surgery are associated with subsequent AKI progression. METHODS: We retrospectively enrolled adult patients admitted to 14 intensive care units after elective cardiac surgery between January and December 2018. We defined mean perfusion pressure (MPP) as the difference between mean arterial pressure (MAP) and central venous pressure (CVP). The main exposure variables were time-weighted-average MPP-deficit (i.e., the percentage difference between preoperative and postoperative MPP) and time spent with MPP-deficit > 20% within the first 24 h. We defined other pressure-related hemodynamic parameters during the initial 24 h as exploratory exposure variables. The primary outcome was AKI progression, defined as one or more AKI stages using Kidney Disease: Improving Global Outcomes' creatinine and urine output criteria between 24 and 72 h. We used multivariable logistic regression analyses to assess the association between the exposure variables and AKI progression. RESULTS: Among the 746 patients enrolled, the median time-weighted-average MPP-deficit was 20% [interquartile range (IQR): 10-27%], and the median duration with MPP-deficit > 20% was 12 h (IQR: 3-20 h). One-hundred-and-twenty patients (16.1%) experienced AKI progression. In the multivariable analyses, time-weighted-average MPP-deficit or time spent with MPP-deficit > 20% was not associated with AKI progression [odds ratio (OR): 1.01, 95% confidence interval (95% CI): 0.99-1.03]. Likewise, time spent with MPP-deficit > 20% was not associated with AKI progression (OR: 1.01, 95% CI 0.99-1.04). Among exploratory exposure variables, time-weighted-average CVP, time-weighted-average MPP, and time spent with MPP < 60 mmHg were associated with AKI progression (OR: 1.12, 95% CI 1.05-1.20; OR: 0.97, 95% CI 0.94-0.99; OR: 1.03, 95% CI 1.00-1.06, respectively). CONCLUSIONS: Although higher CVP and lower MPP were associated with AKI progression, relative hypotension was not associated with AKI progression in patients after cardiac surgery. However, these findings were based on exploratory investigation, and further studies for validating them are required. Trial Registration UMIN-CTR, https://www.umin.ac.jp/ctr/index-j.htm , UMIN000037074.