Predição de choque séptico e hipovolêmico em pacientes de unidade de terapia intensiva com o uso de machine learning / Prediction of septic and hypovolemic shock in intensive care unit patients using machine learning

RESUMO Objetivo: Criar e validar um modelo de predição de choque séptico ou hipovolêmico a partir de variáveis de fácil obtenção coletadas na admissão de pacientes internados em uma unidade de terapia intensiva. Métodos: Estudo de modelagem preditiva com dados de coorte concorrente realizada em um hospital do interior do nordeste brasileiro. Foram incluídos pacientes com 18 anos ou mais sem uso de droga vasoativa no dia da admissão e que foram internados entre novembro de 2020 e julho de 2021. Foram testados os algoritmos de classificação do tipo Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost para a construção do modelo. O método de validação utilizado foi o k-fold cross validation. As métricas de avaliação utilizadas foram recall, precisão e área sob a curva Receiver Operating Characteristic. Resultados: Foram utilizados 720 pacientes para criação e validação do modelo. Os modelos apresentaram alta capacidade preditiva com área sob a curva Receiver Operating Characteristic de 0,979; 0,999; 0,980; 0,998 e 1,00 para os algoritmos de Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost, respectivamente. Conclusão: O modelo preditivo criado e validado apresentou elevada capacidade de predição do choque séptico e hipovolêmico desde o momento da admissão de pacientes na unidade de terapia intensiva.

ABSTRACT Objective: To create and validate a model for predicting septic or hypovolemic shock from easily obtainable variables collected from patients at admission to an intensive care unit. Methods: A predictive modeling study with concurrent cohort data was conducted in a hospital in the interior of northeastern Brazil. Patients aged 18 years or older who were not using vasoactive drugs on the day of admission and were hospitalized from November 2020 to July 2021 were included. The Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost classification algorithms were tested for use in building the model. The validation method used was k-fold cross validation. The evaluation metrics used were recall, precision and area under the Receiver Operating Characteristic curve. Results: A total of 720 patients were used to create and validate the model. The models showed high predictive capacity with areas under the Receiver Operating Characteristic curve of 0.979; 0.999; 0.980; 0.998 and 1.00 for the Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost algorithms, respectively. Conclusion: The predictive model created and validated showed a high ability to predict septic and hypovolemic shock from the time of admission of patients to the intensive care unit.

Prediction of septic and hypovolemic shock in intensive care unit patients using machine learning. / Predição de choque séptico e hipovolêmico em pacientes de unidade de terapia intensiva com o uso de machine learning.

OBJECTIVE: To create and validate a model for predicting septic or hypovolemic shock from easily obtainable variables collected from patients at admission to an intensive care unit. METHODS: A predictive modeling study with concurrent cohort data was conducted in a hospital in the interior of northeastern Brazil. Patients aged 18 years or older who were not using vasoactive drugs on the day of admission and were hospitalized from November 2020 to July 2021 were included. The Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost classification algorithms were tested for use in building the model. The validation method used was k-fold cross validation. The evaluation metrics used were recall, precision and area under the Receiver Operating Characteristic curve. RESULTS: A total of 720 patients were used to create and validate the model. The models showed high predictive capacity with areas under the Receiver Operating Characteristic curve of 0.979; 0.999; 0.980; 0.998 and 1.00 for the Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost algorithms, respectively. CONCLUSION: The predictive model created and validated showed a high ability to predict septic and hypovolemic shock from the time of admission of patients to the intensive care unit.

OBJETIVO: Criar e validar um modelo de predição de choque séptico ou hipovolêmico a partir de variáveis de fácil obtenção coletadas na admissão de pacientes internados em uma unidade de terapia intensiva. MÉTODOS: Estudo de modelagem preditiva com dados de coorte concorrente realizada em um hospital do interior do nordeste brasileiro. Foram incluídos pacientes com 18 anos ou mais sem uso de droga vasoativa no dia da admissão e que foram internados entre novembro de 2020 e julho de 2021. Foram testados os algoritmos de classificação do tipo Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost para a construção do modelo. O método de validação utilizado foi o k-fold cross validation. As métricas de avaliação utilizadas foram recall, precisão e área sob a curva Receiver Operating Characteristic. RESULTADOS: Foram utilizados 720 pacientes para criação e validação do modelo. Os modelos apresentaram alta capacidade preditiva com área sob a curva Receiver Operating Characteristic de 0,979; 0,999; 0,980; 0,998 e 1,00 para os algoritmos de Decision Tree, Random Forest, AdaBoost, Gradient Boosting e XGBoost, respectivamente. CONCLUSÃO: O modelo preditivo criado e validado apresentou elevada capacidade de predição do choque séptico e hipovolêmico desde o momento da admissão de pacientes na unidade de terapia intensiva.

Effect of Slower vs Faster Intravenous Fluid Bolus Rates on Mortality in Critically Ill Patients: The BaSICS Randomized Clinical Trial.

Importance: Slower intravenous fluid infusion rates could reduce the formation of tissue edema and organ dysfunction in critically ill patients; however, there are no data to support different infusion rates during fluid challenges for important outcomes such as mortality. Objective: To determine the effect of a slower infusion rate vs control infusion rate on 90-day survival in patients in the intensive care unit (ICU). Design, Setting, and Participants: Unblinded randomized factorial clinical trial in 75 ICUs in Brazil, involving 11 052 patients requiring at least 1 fluid challenge and with 1 risk factor for worse outcomes were randomized from May 29, 2017, to March 2, 2020. Follow-up was concluded on October 29, 2020. Patients were randomized to 2 different infusion rates (reported in this article) and 2 different fluid types (balanced fluids or saline, reported separately). Interventions: Patients were randomized to receive fluid challenges at 2 different infusion rates; 5538 to the slower rate (333 mL/h) and 5514 to the control group (999 mL/h). Patients were also randomized to receive balanced solution or 0.9% saline using a factorial design. Main Outcomes and Measures: The primary end point was 90-day survival. Results: Of all randomized patients, 10 520 (95.2%) were analyzed (mean age, 61.1 years [SD, 17.0 years]; 44.2% were women) after excluding duplicates and consent withdrawals. Patients assigned to the slower rate received a mean of 1162 mL on the first day vs 1252 mL for the control group. By day 90, 1406 of 5276 patients (26.6%) in the slower rate group had died vs 1414 of 5244 (27.0%) in the control group (adjusted hazard ratio, 1.03; 95% CI, 0.96-1.11; P = .46). There was no significant interaction between fluid type and infusion rate (P = .98). Conclusions and Relevance: Among patients in the intensive care unit requiring fluid challenges, infusing at a slower rate compared with a faster rate did not reduce 90-day mortality. These findings do not support the use of a slower infusion rate. Trial Registration: ClinicalTrials.gov Identifier: NCT02875873.

Dureza Superficial de Resinas Compostas Nanoparticuladas e Bulk Fill Expostas a Suplementos Nutricionais / Surface Hardness of Composite Resins Nanoparticulate and Bulk Fill Exposed to Nutritional Supplements

Objetivo: Avaliar a microdureza superficial de duas resinas compostas após imersão controlada a dois suplementos nutricionais à base de carboidratos, desenvolvidos para indivíduos que praticam atividade esportiva. Materiais e Métodos: Utilizou-se as resinas compostas Filtek Z350 XT (3M ESPE); Filtek bulk Fill (3M ESPE) ambas na cor A3. Foram confeccionados 60 corpos- -de-prova, 30 para cada material restaurador, os quais foram divididos nos seguintes grupos: Grupo 1 - controle; Grupo 2 ­ imersão em 20 mL de Gatorade (AmBev) durante 5 minutos 3 vezes por dia; Grupo 3 - imersão em 30 g de gel de carboidrato Carbo Gel (Vox Suplementos) durante 5 minutos 3 vezes por dia. Em todos os grupos o período de reposição foi de 15 dias. Durante os intervalos, as resinas ficaram armazenadas em saliva artificial a 37º C. Foi realizado o teste de microdureza Vickers com base na norma ISO 4049 com o auxílio de um Microdurometro Vickers - Buehler - Lake Bluff, Illinois/USA (MMT-3) com a carga de 50 gf por 30 segundos, foram realizadas 5 endentações por corpo-de-prova, uma no centro e quatro nas periferias. A média dos valores de dureza foi comparada usando o teste ANOVA e o pós-teste de Tukey, em nível de significância de 5,0%. Resultados: Todos os corpos de prova que ficaram imersos em Gatorade e Gel de carboidrato apresentaram valores de microdureza estatisticamente menores, quando comparados aos grupos controle. Conclusão: A imersão de resinas compostas em suplementos alimentares Gatorade (Ambev) e Carbo Gel (Vox Suplementos) diminui significativamente sua dureza superficial.

Objective: To evaluate the surface microhardness of two composite resins when exposed to two nutritional supplements based on carbohydrates developed for sports activity. Materials and Methods: Filtek Z350 XT composite resins (3M ESPE) were used; Filtek bulk Fill (3M ESPE) both in color A3. Sixty specimens were prepared, 30 of which were for each restorative material, divided into the following groups: Group 1 (control); Group 2 - immersion in 20 mL of Gatorade (AmBev) for 5 minutes 3 times per day; Group 3 - immersion in 30 g Carbo Gel Carbohydrate Gel (Vox Supplements) for 5 minutes 3 times a day, the process was repeated for 15 days. The Vickers micro hardness test was performed based on ISO 4049 with the aid of a Vickers - Buehler - Lake Bluff, Illinois / USA (MMT - 3) Microduometer with a load of 50 gF for 30 seconds. Performing 5 indentations per body of evidence, one in the center and four in the peripheries. Data were submitted to analysis of statistical variance ANOVA, followed by Tukey test at a level of 5.0% of significance. Results: All specimens that were immersed in Gatorade and Carbohydrate Gel had statistically lower microhardness values when compared to the control group. Conclusions: The use of Gatorade (Ambev) and Carbo Gel supplements (Vox Supplements) decreased the surface hardness of the studied resins.

Root dentine demineralisation according to frequency of sucrose exposure using high-fluoride dentifrice.

OBJECTIVE: To evaluate root dentine demineralisation, biomass and loosely bound fluoride (CaF2 ) concentration according to different frequencies of sucrose exposure using a high-fluoride dentifrice. BACKGROUND: Although high-fluoride dentifrice has been recommended to arrest root dentine lesions, it is not clear whether it can protect dentine from increased frequencies of sucrose exposure. METHODS: An in situ, crossover, split-mouth study was conducted in 3 phases with 7 days each, in which 10 volunteers used a palatal device containing 4 bovine root dentine slabs (2 on each side) with predetermined initial hardness. Cariogenic challenge consisted in dripping a 20% sucrose solution 0 (control), 2, 4, 6, 8 or 10 times/d in each block. Volunteers used high-fluoride dentifrice (NaF, 5000 µg F/g) 3 times/d. After each phase, final hardness was measured and the percentage of surface hardness loss (%SHL) calculated. Also, biomass and CaF2 concentration on dentine were determined. The data were processed and analysed by ANOVA and Tukey test with significance level set at 5%. The relationship between the variables was analysed by linear regression and Pearson correlation (r). RESULTS: %SHL and biomass were significantly greater than control for sucrose frequencies higher than 6 times/d (P < 0.001), while CaF2 concentration decreased from sucrose frequency higher than 2 times/d (P < 0.001). Regression analysis data showed a significant linear fit between sucrose exposure frequency and the studied variables with a strong correlation (r) for %SHL and CaF2 and moderate for biomass (P < 0.05). CONCLUSION: High-fluoride dentifrice is able to reduce root dentine demineralisation if sucrose consumption is not higher than 6 times/d.