Predicting 30-Day Readmissions in Patients With Heart Failure Using Administrative Data: A Machine Learning Approach.

BACKGROUND: We sought to develop machine learning (ML) models trained on administrative data which predict risk of readmission in patients with heart failure and to evaluate and compare the ML model with the currently used LaCE score using clinically informative metrics. METHODS AND RESULTS: This prognostic study was conducted in Alberta, Canada, on 9845 patients with confirmed heart failure admitted to hospital between 2012 and 2019. The outcome was unplanned all-cause hospital readmission within 30 days of discharge. We used 80% of the data for the ML model development and 20% for independent validation. We reported, using the validation set, c-statistics (area under the receiver operating characteristic curves)and performance metrics (likelihood ratio, positive predictive values) for the XGBoost model and a modified LaCE score within their respective predictive thresholds. Boosted tree-based classifiers had higher area under the receiver operating characteristic curves (0.65 for XGBoost) compared with others (0.58 for neural networks) and 0.57 for the modified LaCE. Within the predicted threshold range of the XGBoost classifier, the positive likelihood ratio was 1.00 at the low end of predicted risk and 6.12 at the high end, resulting in a positive predictive value (post-test probability) range of 21%-62%; the pretest probability of readmission was 20.9% using prevalence. The corresponding positive likelihood ratios and positive predictive values across LaCE score thresholds were 1.00-1.20 and 21%-24%, respectively. CONCLUSIONS: Despite predicting readmissions better than the LaCE, even the best ML model trained on administrative health data (XGBoost) did not provide substantially informative prediction performance as it only generated a moderate shift from pre to post-test probability. Health systems wishing to deploy such a tool should consider training ML models with additional data. Adding other techniques like natural language processing, along with ML, to use other clinical information (like chart notes) might improve prediction performance.

Baseline serum interleukin-6 to interleukin-2 ratio is associated with the response to seasonal trivalent influenza vaccine in solid organ transplant recipients.

BACKGROUND: The analysis of pre- and post-vaccination B-cell-associated cytokines might be useful in predicting the immunogenicity of seasonal trivalent influenza vaccine (TIV) in solid organ transplant (SOT) recipients. METHODS: We performed a subanalysis of a clinical trial that compared the safety and efficacy of high-dose intradermal (ID) versus intramuscular (IM) TIV in SOT recipients. Serum levels of selected cytokines (interferon [IFN]-γ, interleukin [IL]-2, IL-4, IL-5, IL-6, IL-12 and IL-21, and tumor necrosis factor [TNF]-α) were measured pre- and one month post-vaccination in 155 patients (with 84 and 71 receiving the ID and IM vaccines, respectively). Cytokine profiles were compared according to vaccine response (seroconversion [≥4-fold increase in hemagglutination inhibition antibody titers] to ≥1 influenza vaccine antigen). RESULTS: Mean baseline IL-6 levels were higher (1.20 versus 0.65pg/mL; P-value=0.021) and IL-2 levels were lower (0.01 versus 0.50pg/mL; P-value=0.051) in patients achieving vaccine response. After adjusting for clinical variables, baseline IL-6/IL-2 ratio remained predictive of vaccine response (odds ratio per 10-unit increment: 1.06; 95% confidence interval: 1.02-1.10; P-value=0.002). Vaccination induced an increase in TNF-α (P-value <0.0001) and a decrease in IL-5 levels (P-value=0.0007). There were no significant differences in cytokine kinetics between vaccine responders and non-responders. Mean baseline TNF-α levels were higher in patients experiencing moderate-to-severe adverse events after vaccination (1.93 versus 1.72pg/mL; P-value=0.009). CONCLUSIONS: Baseline serum IL-6 and IL-2 levels, two cytokines that modulate the role of CD4(+) T follicular helper cells and the terminal differentiation of B-cells, predict vaccine response in SOT recipients.

Mineralocorticoid receptor antagonists attenuate pulmonary inflammation and bleomycin-evoked fibrosis in rodent models.

Accumulating evidence indicates protective actions of mineralocorticoid antagonists (MR antagonists) on cardiovascular pathology, which includes blunting vascular inflammation and myocardial fibrosis. We examined the anti-inflammatory and anti-fibrotic potential of MR antagonists in rodent respiratory models. In an ovalbumin allergic and challenged Brown Norway rat model, the total cell count in nasal lavage was 29,348 ± 5451, which was blocked by spironolactone (0.3-60 mg/kg, p.o.) and eplerenone (0.3-30 mg/kg, p.o.). We also found that MR antagonists attenuated pulmonary inflammation in the Brown Norway rat. A series of experiments were conducted to determine the actions of MR blockade in acute/chronic lung injury models. (1) Ex vivo lung slice rat experiments found that eplerenone (0.01 and 10 µM) and spironolactone (10 µM) diminished lung hydroxyproline concentrations by 55 ± 5, 122 ± 9, and 83 ± 8%. (2) In in vivo studies, MR antagonists attenuated the increases in bronchioalveolar lavage (BAL) neutrophils and macrophages caused by lung bleomycin exposure. In separate studies, bleomycin (4.0 U/kg, i.t.) increased lung levels of hydroxyproline by approximately 155%, which was blocked by spironolactone (10-60 mg/kg, p.o.). In a rat Lipopolysaccharide (LPS) model, spironolactone inhibited acute increases in BAL cytokines with moderate effects on neutrophils. Finally, we found that chronic LPS exposure significantly increased end expiratory lung and decreased lung elastance in the mouse. These functional effects of chronic LPS were improved by MR antagonists. Our results demonstrate that MR antagonists have significant pharmacological actions in the respiratory system.