RESUMEN

BACKGROUND: Hemophagocytic Lymphohistiocytosis (HLH) carries a high mortality rate. Current existing risk-evaluation methodologies fall short and improved predictive methods are needed. This study aimed to forecast 30-day mortality in adult HLH patients using 11 distinct machine learning (ML) algorithms. METHODS: A retrospective analysis on 431 adult HLH patients from January 2015 to September 2021 was conducted. Feature selection was executed using the least absolute shrinkage and selection operator. We employed 11 ML algorithms to create prediction models. The area under the curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, F1 score, calibration curve and decision curve analysis were used to evaluate these models. We assessed feature importance using the SHapley Additive exPlanation (SHAP) approach. RESULTS: Seven independent predictors emerged as the most valuable features. An AUC between 0.65 and 1.00 was noted among the eleven ML algorithms. The gradient boosting decision tree (GBDT) algorithms demonstrated the most optimal performance (1.00 in the training cohort and 0.80 in the validation cohort). By employing the SHAP method, we identified the variables that contributed to the model and their correlation with 30-day mortality. The AUC of the GBDT algorithms was the highest when using the top 4 (ferritin, UREA, age and thrombin time (TT)) features, reaching 0.99 in the training cohort and 0.83 in the validation cohort. Additionally, we developed a web-based calculator to estimate the risk of 30-day mortality. CONCLUSIONS: With GBDT algorithms applied to laboratory data, accurate prediction of 30-day mortality is achievable. Integrating these algorithms into clinical practice could potentially improve 30-day outcomes.

Asunto(s)

RESUMEN

BACKGROUND: The purpose of the investigation was to assess the value of post-operation platelet to creatinine ratio (PCR) in predicting in-hospital mortality among patients with acute type A aortic dissection (TAAAD). METHODS: A retrospective study was carried out from January 2017 to December 2019. The best cutoff value of post-operation PCR was assessed by receiver operating characteristic (ROC) curve. Patients were divided into survivors and nonsurvivors. Univariate and multivariate logistic analyses were carried out to identify independent risk factors influencing in-hospital mortality. RESULTS: A total of 171 patients were included in this investigation, with an in-hospital mortality rate of 18.1%. The optimal cut-off value of post-operation PCR was 0.7242 (area under the ROC curve (AUC): 0.798, 95% confidence interval (CI) 0.730-0.856, p < 0.001), and the sensitivity and specificity were 74.2% and 74.3%. The levels of post-operation PCR were lower in nonsurvivors than in survivors (0.56 ± 0.33 vs. 1.50 ± 1.36, p < 0.001). Multivariate logistic regression analysis displayed that post-operation PCR was positively related to in-hospital survivors when confounding factors were adjusted (HR = 8.850, 95% CI = 2.611-30.303, p < 0.001). CONCLUSIONS: Post-operative PCR is a readily accessible and cost-effective biomarker that is independently associated with in-hospital mortality in TAAAD patients. Furthermore, it exhibits superior performance in predicting patient outcomes following surgery.

Asunto(s)

RESUMEN

Objectives: Neonatal sepsis is one of the leading causes of neonatal death. The aim of this study was to evaluate the value of neutrophil - to - monocyte ratio (NMR) in predicting mortality in neonatal sepsis. Methods: In this present retrospective study, a total of 134 neonates with sepsis were included. Baseline laboratory parameters were collected. The best cutoff value of NMR was determined by receiver operating characteristic (ROC) curve. Univariate and multivariate analysis were carried out to survey the predict value of NMR. Results: The results showed that NMR in non-survival group was significantly higher than that in survival group. Results from multivariate analysis showed that high NMR was an independent risk factor for neonatal sepsis (Hazard ratio (HR): 7.519, p = 0.001). ROC displayed that the area under curve (AUC) of NMR was 0.740, sensitivity and specificity of NMR were 80% and 65.8% when 7.65 was selected. Conclusions: NMR could be a promising prognostic factor for neonatal sepsis.

RESUMEN

OBJECTIVE: Acute type A aortic dissection (AAD) is a serious life-threatening cardiovascular emergency with high in-hospital mortality without aggressive clinical treatment. The study intended to identify the relationship between platelet (PLT) to platelet distribution width (PDW) ratio (PPR) and in-hospital mortality in postoperative patients with type A AAD. METHODS: A total of 171 type A AAD patients were recruited in this retrospective study from January 2017 to December 2019. Receiver operating characteristics (ROC) were exploited to determine the best cut-off value of PPR, and then patients were sub-grouped into the low-PPR group and high-PPR group, according to the optimal value of PPR. Finally, univariate, and multivariate analyses were carried out to examine the prognostic value of PPR. RESULTS: The value of PPR was 9.76, and the mortality was statistically higher in the low-PPR group than in the high-PPR group (29.1% vs. 6.0%, P < 0.01). The area under the ROC curve (AUC) of PPR was 0.724 (95% CI, 0.633-0.815; P < 0.001) with a 56.4% sensitivity and 80.6% specificity. Multivariate analysis showed that serum PPR was an independent factor associated with in-hospital mortality (hazard ratio (HR): 1.151; 95% confidence interval (CI): 1.035 -1.297; P = 0.010). CONCLUSION: Serum PPR can be used as an independent predictor of in-hospital mortality in postoperative patients with type A AAD.

Asunto(s)

RESUMEN

Background: The prognosis of Infective endocarditis (IE) is poor, and we conducted this investigation to evaluate the worth of admission lymphocyte-to-white blood cell ratio (LWR) for prediction of short-term outcome in IE patients. Methods: We retrospectively assessed the medical records of 147 IE patients from January 2017 to December 2019. Patients were divided into the survivor group and nonsurvivor group. Univariate and multivariate analyses were applied to estimate the independent factors contribution to in-hospital death, and receiver-operator characteristic (ROC) curve was utilized to check the performance. Results: The levels of LWR (0.17 ± 0.08 vs. 0.10 ± 0.06) were significantly increased among the survivor group compared with the nonsurvivor group (P = 0.001). Multivariate analysis displayed that LWR (hazard ratio (HR): 1.755, 1.304-2.362, P < 0.001) was not interfered by other confounding factors for early death. Moreover, ROC analysis suggested that LWR (cutoff value = 0.10) performed the best among assessed indexes for the forecast of primary outcome (area under curve (AUC) = 0.750, 95% confidence interval (CI) = 0.634-0.867, P < 0.001, sensitivity = 70.0%, specificity = 76.4%), and the proportion of in-hospital mortality was remarkably inferior in patients with LWR > 0.10 than in those with LWR ≤ 0.10. (5.83% vs. 31.8%, P < 0.001). Conclusions: LMR is an independent, simple, universal, inexpensive, and reliable prognostic parameter to identify high-risk IE patients for in-hospital mortality.

Asunto(s)

RESUMEN

BACKGROUND: Lymphocyte and plateletcrit (PCT) as proportions of routine complete blood count tests, have been studied as simple biomarkers for inflammatory diseases. The aim of our study was to investigate whether blood routine parameters, especially platelet parameters could be a useful tool to distinguish Adult onset Still's disease (AOSD) from sepsis. METHODS: We retrospectively reviewed 58 patients with AOSD and 55 sepsis patients diagnosed at the First Affiliated Hospital of Nanjing Medical University between January, 2015 to December 2018. Laboratory data including ferritin, blood routine parameters and C­reactive protein (CRP) level were collected, and the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte (PLR) were calculated. RESULTS: The results showed that AOSD patients showed higher ferritin, lymphocyte and PCT (all P < 0.01) and these factors are independent risk factors for predicting AOSD. In receiver operating characteristic (ROC) curve analysis of LY, PCT and ferritin for distinguish of AOSD, the area under the curve (AUC) was 0.676 (0.576-0.777); 0.706 (95% CI = 0.596-0.816); 0.715 (0.617-0.814). Meanwhile, the AUC of the combination of lymphocyte, PCT and ferritin was 0.836 (0.737-0.909) with sensitivity 67.3, specificity 92.3, and the difference was significant. CONCLUSIONS: Thus we suggest that lymphocyte, PCT may be a useful tool to make a distinction between AOSD and sepsis, as supplementary biomarkers to ferritin.

Asunto(s)

RESUMEN

BACKGROUND: The aim of this study is to establish a clinical diagnosis model as a new evaluation indicator for the differentiation of adult-onset Still's disease (AOSD) and other fever of unknown origin disease (FUO). METHODS: This is an observational case-control study between January 2010 and December 2018. Laboratory parameters of AOSD group (N=91), FUO group (N=89) and control group (N=81) including procalcitonin, C-reactive protein (CRP), ferritin, leukocyte, lymphocyte, neutrophil, lymphocyte proportion, neutrophil proportion, red blood cell distribution width (RDW), platelet and platelet parameters were collected. Descriptive statistics and logistic regression were performed to establish a model based on these laboratory variables. RESULTS: After univariate screening, the variables including CRP, leukocyte, neutrophil, lymphocyte proportion, neutrophil proportion, ferritin and mean platelet volume (MPV) showed significant difference between AOSD and FUO groups, then a stepwise regression analysis was performed to establish a model based on these screened variables, at last ferritin, neutrophil and MPV were significantly different in the model. The results suggested that the higher value of ferritin and neutrophil, the lower value of MPV in the model indicated the higher risk to diagnose AOSD. Area under the curve (AUC) of the model was 0.909 (95% CI: 0.855-0.947), which showed high differential diagnostic value (sensitivity: 86.6%, specificity: 82.0%). CONCLUSIONS: The diagnosis model of AOSD and other FUO was established, with an outstanding performance for differential diagnosis.

Asunto(s)