A predictive score for COVID-19 diagnosis using clinical, laboratory and chest image data

Abstract Objectives Differential diagnosis of COVID-19 includes a broad range of conditions. Prioritizing containment efforts, protective personal equipment and testing can be challenging. Our aim was to develop a tool to identify patients with higher probability of COVID-19 diagnosis at admission. Methods This cross-sectional study analyzed data from 100 patients admitted with suspected COVID-19. Predictive models of COVID-19 diagnosis were performed based on radiology, clinical and laboratory findings; bootstrapping was performed in order to account for overfitting. Results A total of 29% of patients tested positive for SARS-CoV-2. Variables associated with COVID-19 diagnosis in multivariate analysis were leukocyte count ≤7.7 × 103 mm-3, LDH >273 U/L, and chest radiographic abnormality. A predictive score was built for COVID-19 diagnosis, with an area under ROC curve of 0.847 (95% CI 0.77-0.92), 96% sensitivity and 73.5% specificity. After bootstrapping, the corrected AUC for this model was 0.827 (95% CI 0.75-0.90). Conclusions Considering unavailability of RT-PCR at some centers, as well as its questionable early sensitivity, other tools might be used in order to identify patients who should be prioritized for testing, re-testing and admission to isolated wards. We propose a predictive score that can be easily applied in clinical practice. This score is yet to be validated in larger populations.

Análise de Escores de Risco para Predição de Mortalidade em Pacientes Submetidos à Cirurgia Cardíaca por Endocardite / Analysis of Risk Scores to Predict Mortality in Patients Undergoing Cardiac Surgery for Endocarditis

Resumo Fundamento Escores de risco estão disponíveis para uso na prática clínica diária, mas saber qual deles escolher é ainda incerto. Objetivos Avaliar o EuroSCORE logístico, o EuroSCORE II e os escores específicos para endocardite infecciosa STS-IE, PALSUSE, AEPEI, EndoSCORE e RISK-E na predição de mortalidade hospitalar de pacientes submetidos à cirurgia cardíaca por endocardite ativa em um hospital terciário de ensino do sul do Brasil. Métodos Estudo de coorte retrospectivo incluindo todos os pacientes com idade ≥ 18 anos submetidos à cirurgia cardíaca por endocardite ativa no centro do estudo entre 2007 e 2016. Foram realizadas análises de calibração (razão de mortalidade observada/esperada, O/E) e de discriminação (área sob a curva ROC, ASC), sendo a comparação das ASC realizada pelo teste de DeLong. P < 0,05 foi considerado estatisticamente significativo Resultados Foram incluídos 107 pacientes, sendo a mortalidade hospitalar de 29,0% (IC95%: 20.4-37.6%). A melhor razão de mortalidade O/E foi obtida pelo escore PALSUSE (1,01, IC95%: 0,70-1,42), seguido pelo EuroSCORE logístico (1,3, IC95%: 0,92-1,87). O EuroSCORE logístico apresentou o maior poder discriminatório (ASC 0,77), significativamente superior ao EuroSCORE II (p = 0,03), STS-IE (p = 0,03), PALSUSE (p = 0,03), AEPEI (p = 0,03) e RISK-E (p = 0,02). Conclusões Apesar da disponibilidade dos recentes escores específicos, o EuroSCORE logístico foi o melhor preditor de mortalidade em nossa coorte, considerando-se análise de calibração (mortalidade O/E: 1,3) e de discriminação (ASC 0,77). A validação local dos escores específicos é necessária para uma melhor avaliação do risco cirúrgico. (Arq Bras Cardiol. 2020; 114(3):518-524)

Abstract Background Risk scores are available for use in daily clinical practice, but knowing which one to choose is still fraught with uncertainty. Objectives To assess the logistic EuroSCORE, EuroSCORE II, and the infective endocarditis (IE)-specific scores STS-IE, PALSUSE, AEPEI, EndoSCORE and RISK-E, as predictors of hospital mortality in patients undergoing cardiac surgery for active IE at a tertiary teaching hospital in Southern Brazil. Methods Retrospective cohort study including all patients aged ≥ 18 years who underwent cardiac surgery for active IE at the study facility from 2007-2016. The scores were assessed by calibration evaluation (observed/expected [O/E] mortality ratio) and discrimination (area under the ROC curve [AUC]). Comparison of AUC was performed by the DeLong test. A p < 0.05 was considered statistically significant. Results A total of 107 patients were included. Overall hospital mortality was 29.0% (95%CI: 20.4-37.6%). The best O/E mortality ratio was achieved by the PALSUSE score (1.01, 95%CI: 0.70-1.42), followed by the logistic EuroSCORE (1.3, 95%CI: 0.92-1.87). The logistic EuroSCORE had the highest discriminatory power (AUC 0.77), which was significantly superior to EuroSCORE II (p = 0.03), STS-IE (p = 0.03), PALSUSE (p = 0.03), AEPEI (p = 0.03), and RISK-E (p = 0.02). Conclusions Despite the availability of recent IE-specific scores, and considering the trade-off between the indexes, the logistic EuroSCORE seemed to be the best predictor of mortality risk in our cohort, taking calibration (O/E mortality ratio: 1.3) and discrimination (AUC 0.77) into account. Local validation of IE-specific scores is needed to better assess preoperative surgical risk. (Arq Bras Cardiol. 2020; 114(3):518-524)