Role of risk scoring systems in predicting life expectancy after carotid endarterectomy in asymptomatic patients.

OBJECTIVE: The aim of this study is to compare and to test the performance of all available risk scoring systems (RSSs) designed to predict long-term survival rate in asymptomatic candidate patients for carotid endarterectomy (CEA) for significant carotid artery stenosis. METHODS: Data on asymptomatic patients who underwent CEA in three high-volume centers were prospectively recorded. Through literature research using Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations, six RSSs were identified for the intent of the study. Primary endpoints were 3- and 5-year survival rates after CEA. All items used as variables to compose multiple RSSs were applied to every patient in the study population. The 3- and 5-year mortality prediction rates for each score were assessed by sensitivity, specificity, and predictive negative and positive value calculation, as well as univariable Cox proportional hazard models with the Harrell C index. RESULTS: During the study period, 825 CEAs in 825 asymptomatic patients were analyzed. All items used in RSSs were available in the dataset, with some concerns regarding their definition and application among RSSs. The 3- and 5-year survival rates of the study cohort were 94.5% and 90.3%, respectively. Among the six RSSs analyzed, no RSS demonstrated optimal results in terms of mortality rate prediction accuracy, although some scores had good diagnostic and risk of death precision. CONCLUSIONS: RSSs, when used alone, fail to optimally detect postoperative life expectancy in asymptomatic CEA patient candidates. Further prospective controlled studies are needed to compose and validate RSSs with better calibration to predict outcomes.

Derivation and validation of the clinical prediction model for COVID-19.

The epidemic phase of Coronavirus disease 2019 (COVID-19) made the Worldwide health system struggle against a severe interstitial pneumonia requiring high-intensity care settings for respiratory failure. A rationalisation of resources and a specific treatment path were necessary. The study suggests a predictive model drawing on clinical data gathered by 119 consecutive patients with laboratory-confirmed COVID-19 admitted in Busto Arsizio hospital. We derived a score that identifies the risk of clinical evolution and in-hospital mortality clustering patients into four groups. The study outcomes have been compared across the derivation and validation samples. The prediction rule is based on eight simple patient characteristics that were independently associated with study outcomes. It is able to stratify COVID-19 patients into four severity classes, with in-hospital mortality rates of 0% in group 1, 6-12.5% in group 2, 7-20% in group 3 and 60-86% in group 4 across the derivation and validation sample. The prediction model derived in this study identifies COVID-19 patients with low risk of in-hospital mortality and ICU admission. The prediction model that the study presents identifies COVID-19 patients with low risk of in-hospital mortality and admission to ICU. Moreover, it establishes an intermediate portion of patients that should be treated accurately in order to avoid an unfavourable clinical evolution. A further validation of the model is important before its implementation as a decision-making tool to guide the initial management of patients.