Background To develop a machine learning model for predicting acute respiratory distress syndrome (ARDS) events through commonly available parameters, including baseline characteristics and clinical and laboratory parameters. Methods A secondary analysis of a multi-centre prospective observational cohort study from five hospitals in Beijing, China, was conducted from January 1, 2011, to August 31, 2014. A total of 296 patients at risk for developing ARDS admitted to medical intensive care units (ICUs) were included. We applied a random forest approach to identify the best set of predictors out of 42 variables measured on day 1 of admission. Results All patients were randomly divided into training (80%) and testing (20%) sets. Additionally, these patients were followed daily and assessed according to the Berlin definition. The model obtained an average area under the receiver operating characteristic (ROC) curve (AUC) of 0.82 and yielded a predictive accuracy of 83%. For the first time, four new biomarkers were included in the model: decreased minimum haematocrit, glucose, and sodium and increased minimum white blood cell (WBC) count. Conclusions This newly established machine learning-based model shows good predictive ability in Chinese patients with ARDS. External validation studies are necessary to confirm the generalisability of our approach across populations and treatment practices.

BackgroundEarly detection of the Acute Respiratory Distress Syndrome (ARDS) has the potential to improvethe prognosis of critically ill patients admitted to the intensive care unit (ICU). However, no reliable biomarkers are currently available for accurate early detection of ARDS in patients with predisposing conditions.ObjectivesThis study examined risk factors and biomarkers for ARDS development and mortality in two prospective cohort studies.MethodsWe examined clinical risk factors for ARDS in a cohort of 178 patients in Beijing, China who were admitted to the ICU and were at high risk for ARDS. Identified biomarkers were then replicated in a second cohort of1,878 patients in Boston, USA.ResultsOf 178 patients recruited from participating hospitals in Beijing, 75 developed ARDS. After multivariate adjustment, sepsis (odds ratio [OR]:5.58, 95% CI: 1.70–18.3), pulmonary injury (OR: 3.22; 95% CI: 1.60–6.47), and thrombocytopenia, defined as platelet count <80×103/µL, (OR: 2.67; 95% CI: 1.27–5.62)were significantly associated with increased risk of developing ARDS. Thrombocytopenia was also associated with increased mortality in patients who developed ARDS (adjusted hazard ratio [AHR]: 1.38, 95% CI: 1.07–1.57) but not in those who did not develop ARDS(AHR: 1.25, 95% CI: 0.96–1.62). The presence of both thrombocytopenia and ARDS substantially increased 60-daymortality. Sensitivity analyses showed that a platelet count of <100×103/µLin combination with ARDS provide the highest prognostic value for mortality. These associations were replicated in the cohort of US patients.ConclusionsThis study of ICU patients in both China and US showed that thrombocytopenia is associated with an increased risk of ARDS and platelet count in combination with ARDS had a high predictive value for patient mortality.

BackgroundTumor necrosis factor-α (TNF-α) may play an important role in the recalcitrant inflammatory and hyperproliferative dermatosis of psoriasis, and there may be a relationship between TNF-α polymorphisms and psoriasis risk. MethodsWe performed a meta-analysis to evaluate the associations between TNF-α polymorphisms and psoriasis. Electronic searches of Pubmed, Embase, and Web of Science were performed for all publications on the associations between TNF-α polymorphisms and psoriasis through September 26, 2012. The pooled odds ratios (ORs) with their 95% confidence interval (95%CIs) were calculated to assess the associations.ResultsSixteen case-control studies with a total of 2,253 psoriasis cases and 1,947 controls on TNF-α 308 G/A polymorphism and fourteen studies on TNF-α 238 G/A polymorphism with 2,104 cases and 1,838 controls were finally included into the meta-analysis. Overall, TNF-α 308 G/A polymorphism was significantly associated with decreased risk of psoriasis under three genetic comparison models (for A versus G: fixed-effects OR 0.71, 95%CI 0.62-0.82, P < 0.001; for AG versus GG: fixed-effects OR 0.67, 95%CI 0.57-0.78, P < 0.001; for AA/AG versus GG: fixed-effects OR 0.67, 95%CI 0.58-0.78, P < 0.001). In addition, TNF-α 238 G/A polymorphism was associated with increased risk of psoriasis under three genetic models (for A versus G: fixed-effects OR 2.46, 95%CI 2.04-2.96, P < 0.001; for AG versus GG: fixed-effects OR 2.69, 95%CI 2.20-3.28, P < 0.001; for AA/AG versus GG: fixed-effects OR 2.68, 95%CI 2.20-3.26, P < 0.001). Subgroup analysis by ethnicity identified a significant association between TNF-α 308 G/A polymorphism and decreased risk of psoriasis in both Caucasians and Asians and a significant association between TNF-α 238 G/A polymorphism and increased risk of psoriasis in Caucasians.ConclusionsThe meta-analysis suggests that TNF-α 308 G/A polymorphism is associated with decreased risk of psoriasis, while TNF-α 238 G/A is associated with increased risk of psoriasis.